CORNELL UNIVERSITY LIBRARY BOUGHT WITH THE INCOME OF THE SAGE ENDOWMENT FUND GIVEN IN 1891 BY HENRY WILLIAMS SAGE I-. ^«» .52^J*" University Library TA 403.M65 '*''?iiiii™iiS!iii*i°''*'™''''"'i ••'^'f manutae 3 1924 021 906 239 All books are subject to recall after two weeks. Olin/Kroch Library DATE DUE ^twll J k7T7f"* ■t^lW GAYLORD PRINTED IN U.S.A. Cornell University Library The original of tiiis book is in tine Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924021906239 MATERIALS or CONSTRCCTION THEIR MANUFACTURE, PROPERTIES, AND USES BY ADELBERT P. MILLS, M.S. (C.E.) Assistant Professor of Materials, College of Civil Engineering, Cornell University. Associate Member of The American Society of Civil Engineers, Member of American Society for Testing Materials FIRST EDITION TOTAL ISSUE, SIX THOUSAND NEW YORK JOHN WILEY & SONS, Inc. London: CHAPMAN & HALL, Limited 1915 '■! 1 X TA /1-45€(olt5 ' Copyright, 1915 BY ADELBERT P. MILLS s^«. PRESS 0^ BRAUNWORTH & CO. BOOK MANUFACTURERS BROOKLYN, N, Y. . V3 PREFACE This work is an outgrowth of certain lectures and notes which have been used in the author's classes in the College of Civil Engineering, Cornell University, for the past several years. Its preparation was under- taken to meet the need which was felt for a general text-book covering the manufacture, properties, and uses of the more common materials of engineering construction in a comparatively concise and thoroughly modem manner. Although this book is intended primarily for use as a text-book of somewhat elementary character and is not a treatise exhaustively covering the very broad field of " Materials of Construction," the treat- ment has been made more detailed in some respects than may be nec- essary for class-room purposes, and its applications as a general ref- erence work thereby broadened. The treatment of the various classes of materials considered follows a general systematic form which has been made uniform throughout so far as has been found practicable. The consideration of each material or class of materials is prefaced by a discussion of its ordinary appli- cations in engineering construction, followed by a study of its manu- facture or natural occurrence, and concluded by a discussion of physical and mechanical properties in their relation to its uses. As a result of the author's experience in the teaching of this subject, the properties exhibited by a given material are, for the most part, considered as dependent phenomena closely, related to certain more or less variable factors connected witti the process of manufacture, natural occurrence, and conditions of service or testing, and not as independent qualities inherent in that material. It has been considered advisable to avoid the inclusion of tabulations of investigational data whenever the data could be presented graphically by curves or diagrams; dis- cussions of conflicting empirical data on points admitting of controversy have been reduced to a minimum; and an effort has been made to present iv PREFACE the material in a definite, concrete form, the necessity for the exercise of discriminative judgment upon the part of the student being obviated by conclusions drawn by the author, even though it is recognized that in so doing errors of judgment may be made, and the criticism of those who object to any form of dogmatic statement is invited The subject of testing materials has not been covered, except in so far as methods of testing are inseparable from discuosiono of the properties of materials revealed by laboratory tests. It is the author's conviction that this subject can be handled only in the laboratory itself, and the place for such material is therefore in a laboratory manual. It is assumed, however, that a laboratory course in testing materials will invariably parallel and supplement the text-book course in the study of materials. The author cannot make a pretense of being a specialist in all of the fields which are covered in the various chapters of this book, and this work is therefore to a very large degree a compilation of data and opinion from a great many different sources. The author takes pleasure in acknowledging his great indebtedness to the large number of engineers and manufac- turers who have privately or by their writings contributed much to make up this volume. A large number of technical books which are devoted to the consideration of some part of the ground covered by this text have been frequently consulted and freely used. An effort has been made to always acknowledge the source of information so obtained, and if any error of omission has been committed in this respect, it has been committed inadvertently, not by intention. The following well-known text-books and reference works have been most frequently used : " Cements, I.imes, and Plasters," by E. C. Eckel; " Portland Cement," by R; K. Meade; " Manufacture of Portland Cement," by R.. C. H. West; " Masonry Construction," by I. O. Baker; " Stone for Building and Decoration," by G. P. Merrill; " Building Stones and Clay Prod- ucts " and " Economic Geology," by Heinrich Rics; " The Blast Furnace and the Manufacture of Pig Iron," by Robert Forsythe; " The Met- allurgy of Iron and Steel," by Bradley Stoughton; " The Metallurgy of Steel " and " Iron, Steel, and Other Alloys," by H. M. Howe; " The Manufacture and Properties of Iron and Steel," l)y H. H. Campbell; " Iron and Steel," by H. P. Tiemann; " Modern Iron Foundry Practice,"' by G. R. Bale; " Cast Iron," by W. J. Keep; " The Production of Malleable Castings," by Richard Moldenke; " The Corrosion and Pres- ervation of Iron and Steel," by A. S. Cushman and H. A. Gardner; "The Metallography of Iron and Steel," by Albert Sauveur; and " Economic Woods of the United States " and " The Mechanical' Prop- erties of Wood," by S. J. Record. PREFACE V The following periodicals and publications of various societies have also been frequently consulted : " Engineering News," " Engineering Record," " Metallurgical and Chemical Engineering," Proceedings of the American Society for Testing Materials, Proceedings of the International Association for Testing Materials, Transactions of the American Society of Civil Engineers, Proceedings of the Institution of Mechanical Engineers, Journal of the Iron and Steel Institute, " Tests of Metals," published annually by the U. S. War Department, " Mineral Resources," pubUshed annually by the U. S. Geological Survey, Reports of the various State Geological Surveys, the pubUcations of the Forestry Division of the U. S. Depart- ment of Agriculture, and the publications of the U. S. Bureau of Standards. Adelbert P. Mills. Ithaca, N. Y. February 20, 1915. TABLE OF CONTENTS PART I THE MATERIALS OF MASONRY CONSTRUCTION CEMENTING MATERIALS PAGE CHAPTER I. GYPSUM PLASTERS 1-18 GENERAL efinition and Classification, 1. Application in Arts and Industries, 2. Struc- tural Uses, 3. MANUFACTURE 3T)sam Rocks, 4. Theory ol Calcination, 5. Practice of Calcination, 6. Addi- tions Subsequent to Calcination, 7. PROPERTIES AND USES OF GYPSUM PLASTERS stting and Hardening. Action of Accelerators and Retarders, 8. Strength of Plasters, Tensile Strength, Compressive Strength, Adhesive Strength, 9. Relative Applicability of Various Gjrpsum Plasters to Structural Uses, 10. Production, Value, and Uses of Gypsum Products, 11. CHAPTER II. QUICKLIME 19-49 GENERAL lefinition and Classification, 12. The Place of Lime among Cementing Materials, 13. Application in the Arts and Industries, 14. Structural Uses, 15. MANUFACTURE OF LIME imestone Rocks, 16. Theory of Calcination, 17. Practice of Calcination, Intermittent Kilns, Continuous Kilns, 18. Treatment Subsequent to Cal- cination, 19. viii TABLE OF CONTENTS PROPERTIES AND USES OF QUICKLIMES PAGE Classification of Limes by Uses. Required Qualifications of Each Class, 20. Chemical Composition, 2L Hydration or Slaking. Rate of Hydration, 22. Setting and Hardening. Accompanying Phenomena, 23. Plasticity. Sand- carrying Capacity and Yield, 24. Waste, 25. Hardness, Time of Setting, and Shrinkage, 26. Crushing and Tensile Strength of Lime Mortars; Ten- sile Strength, High-Calcium and Dolomitic Lime Mortar; Compressive Strength, High-Calcium and Dolomitic Lime Mortar; Effect of Temperature of Calcination; Effect of Method of Slaking; Effect of Character of Sand; Effect of Size of Test Specimens, 27. Relative Applicability of Various Limes to Special Uses, 28. Production, Value, and Uses of Lime, 29. CHAPTER III. HYDRATED LIME 50-58 Definition, 30. Process of Manufacture, 31. Properties and Uses, 32. Hydrated Lime vs. Quicklime, 33. Special High Alumina Hydrated Lime, 34. Prop- erties and Uses of High Alumina Lime, 34. Production, Value, and Uses of Hydrated Lime, 36. CHAPTER IV. HYDRAULIC LIME AND GRAPPIER CEMENTS 59-69 HYDRAULIC CEMENTING MATERIALS IN GENERAL Introduction, 37. Classification of Hydraulic Cementing Materials. The Hy- draulic Index and the Cementation Index, 38. HYDRAULIC LIMES. GENERAL Definition and Classification, 39. The Hydraulic Lime Industry. Uses of the Hydraulic limes, 40. MANUFACTURE OF HYDRAULIC LIMES Hydraulic Limestones, 41. Calcination, 42. Slaking ahd Subsequent Treat- ment, 43. PROPERTIES OF HYDRAULIC LIMES Composition, 44. Physical Properties, 45. Tensile and Compressive Strength, 46. Hydraulic Limes in Construction, 47. CHAPTER V. PUZZOLAN CEMENTS. SLAG CEMENTS 70-78 PUZZOLAN CEMENTS General, 48. Definition of Puzzolans and Puzzolan Cement, 49. Naturg,! Puzzo- lanic Materials, 50. Manufacture of Puzzolan Cements from Natural Mate- rials, 51. Properties and Uses of Natural Puzzolan Cements, 52. TABLE OF CONTENTS ix SLAG CEMENTS • PAGE Definition, 53. Blast-Furnace Slag. Required Composition and Physical Con- dition, 54. Manufacture of Slag Cements, 55. Properties and Uses of Slag Cements, 56. CHAPTER VI. NATURAL CEMENTS 79-89 GENERAL Definition. Distinction between Natural and Portland Cements, 57. Natural Cement as a Structural Material, 58. MANUFACTURE OF NATURAL CEMENTS Natural Cement Rocks, 59. Theory of Calcination, 60. Practice of Calcina- tion, 61. The Clinker, 62. Free Lime, 63. Grinding and Packing, 64. Manufacturing Costs, 65. PROPERTIES AND USES OF NATURAL CEMENTS Chemical Composition, Constitution, 66. Specific Gravity, 67. Time of Setting, 68. Fineness, 69. Tensile Strength, 70. Compressive Strength. 71. Modu- lus of Elasticity, 72. Where Natural Cement May be Used, 73. Pro- duction and Value of Natural Cement, 74. Status of the Industry, 75. CHAPTER VIL PORTLAND CEMENT 90-171 GENERAL Historical, 76. Definition of Portland Cement, 77. Portland Cement as a Struc- tural Material, 78. PORTLAND CEMENT MANUFACTURE Raw Materials, 79. Limestone, 80. Argillaceous Limestone, Cement Rook, 81. Marl, 82. Clays, Shales and Slates, 83. Alkali Waste, 84. Blast Furnace Slag, 85. Proportioning the Raw Materials, 86. Control of the Mixture During Operation of Plant, 87. Treatment of Materials Preliminary to Calcination, 88. Theory of Calcination, 89. The Dry Process Quarrying, Crushing and Drying the Rock, 90. Grinding, Mixing and Pulver- izing the Raw Materials, 91. Burning the Cement Mixture, 92. Treatment of the Clinker, Cooling, Grinding, and Pulverizing, 93. Addition of Retarder, 94. Storing and Packing, 95. The Wet Process The Wet Process of Manufacture, Using Marl and Clay or Shale, 96. The Wet Process, Using Chalk and Clay, 97. Cement Mill Equipment Equipment of Raw Mill, 99. Ceirient Kihis, 100. CUnker Cooling Equipment. Equipment of Finishing Mill, 101. Cost of Manufacture, 102. Production of Portland Cement, 103. X TABLE OF CONTENTS PROPERTIES AND USES OF PORTLAND CEMENT ^^ General, 104. Composition of Portland Cement, 105. The Constitution of Portland Cement, 106. Setting and Hardening, 107. Specific Gravity Significance, 108. Specification and Result of Tests, 109. Influence of Thor- oughness of Burning upon Specific Gravity, 110. Influence of Adulteration upon Specific Gravity, 111. Influence of Seasoning upon Specific Gravity, 112. Summary and Conclusions, 113. Fineness of Grinding Significance, 114. Specification and Results of Tests, 115. Influence upon Soundness, 116. Influence upon Setting Time, 117. Influence upon Neat and Mortar Strength, 118. Summary and Conclusions^ 119. Time of Setting Significance, 120. Specification and Results of TestS; 121. Influence of Tem- perature, 122. Influence of the Percentage of Water Used to Gauge the Cement, 123. Influence of Sulphates, 124. Influence of Seasoning, 125. Summary and Conclusions, 126. Soundness Significance, 127. Specification, 128. Influence of Seasoning, 129. Influence of Fineness, 130. Effect of Sulphates, 131. Summary and Conclusions, 132. Tensile Strength Signiflcance, 133. Specification and Results of Tests, Neat Cement, 134. Influ- ence upon Lime Proportion upon Tensile Strength, 135. Influence of Tempera- ture of Burning upon Tensile Strength, 136. Influenceof Fineness of Grinding upon Tensile Strength, 137. Tensile Strength of Sand-cement Mortars Significance, 138. Specification, 139. Standard Sand, 140. Effect of Fineness of Sand upon Mortar Strength, 141. Relation between Density of Mortars and Tensile Strength, 142. Influence of Mica in Sand upon Tensile Strength of Mortars, 143. Influence of Cleanliness of Sand upon Strength, 144. Effect of Addition of Hydrated Lime to Cement and Mortars, 145. Relation be- tween Tensile and Compressive Strength, 146. Tensile Strength and Sound- ness, 147. Relation between Neat and Mortar Strength, 148. Rate of Increase in Tensile Strength and Loss of Strength Observed in Long-time Tests, 149. Summary, Tensile Strength, 150. Compressive Strength. Modulus op Elasticity. Shearing Strength Significance, 151. Specification and Results of Tests, 152. Modulus of Elasticity of Cement and Mortars, 153. Shearing Strength of Cement and Mortars, 154. Adhesive Strength. Abrasive Resistance. Permeability and Absorp- tion. Expansion and Contraction Adhesion to Steel, 155. Adhesion to Brick, 156. Adhesion of Mortar to Various Materials, 157. Abrasive Resistance of Cement and Mortars, 158. Per- meability and Absorptive Properties of Cement Mortar, 159. The Expansion and Contraction of Cement Mortars, 160. TABLE OF CONTENTS xi PAGE CHAPTER VIII. CONCRETE 172-201 GENERAL Concrete as a Structural Material, 161. CONCIIETE MATERIALS The Cement Selection of Cement, 162. Storage of Cement, 163. Inspection and Testing, 164 Sand fob Concrete Aggregate Granulometric Composition, 165. Shape of Sand Grains, 166. Foreign Matter in Sand, 167. Voids in Sand, 168. Broken Stone or Gravel Aggregate Gravel vs. Broken Stone, 169. Crushing and Screening Stone, 170. Mechanical Analysis of Stone, 171. Size and Shape of Fragments of Stone, 172. Voids in Stone or Gravel, 173. THE MAKING OF CONCRETE Proportioning Concrete Importance of Proper Proportioning, 174. Theory of Proper Proportioning, 175. Proportion in Practice, 176. Ingredients Required per Cubic Yard of Con- crete 177. Mixing Concrete Hand vs. Machine Mixing, 178. Method of Mixing by Hand, 179. Mixing Machines and Machine Mixing, 180. Deposition op Concrete Transportation and Deposition in Forms, 182. Consistency, Ramming or Pud- dling, 183. Bonding to Old Work, 184. Facing of Walls, 185. Depositing under Water, 186. THE MAKING OF CONCRETE UNDER SPECIAL CONDITIONS . Laying Concrete in Freezing Weather Effect of Low Temperatures, 187. Methods of Concreting in Freezing Weather, 188. Concrete in Sea Water. Effect op Alkali on Concrete Action of Sea Water on Concrete, 189. Expedients Adopted to Prevent Injury by Sea Water, 190. Effect of Alkali on Concrete, 191. Concrete Where Watertightness is Required Proportioning the Mixture, 192. Thickness Required for Watertightness, 193. Use of Waterproofing Compounds, 194. Layers of Waterproof Material, 195. Surface Treatment for Waterproofing, 196. xii TABLE OF CONTENTS PROPERTIES OF CONCRETE ^^^^ Cotnpressivd Strength, 197. Tensile Strength, 198. Transverse Strength, 199. Shearing Strength, 200. Elastic Properties, 201. Modulus of Elasticity, 202. Elastic Limit, 203. Stress-Strain Curves, 204. Coefficient of Expansion, 205. Contraction and Expansion of Concrete, 206. Weight of Concrete, 207. Adhesion to Steel, 208. Ratio Ec/Es, 209, Fire-resistant Properties of Concrete, 210. Protection of Steel from Corrosion, 211. Working Stresses and Factor of Safety, 212. NON-CEMENTING MASONRY MATERIALS CHAPTER IX. BUILDING STONES AND STONE MASONRY 201-221 BUILDING STONES GENERAL Stone *s a Structural Material, 213. Classification of Rocks, 214. STONE QUARRYING AND CUTTING Methods of Quarrjdng, 215. Stone Cutting, 216. PROPERTIES OF BUILDING STONES General Description of Stones Granite, 217. Gneiss, 218. Limestones, 219. Crystalline Limestone or Marble, 220. Compact Common Limestones, 221. Sandstones, 222. ' Slates, 223. Physical and Mechanical Properties op Building Stones . Selection of Building Stone, 224. Properties of Various Stones, 225. STONE MASONRY Classification, 226. Compressive Strength, 227. Allowable Loads on Stone Masonry, 228. CHAPTER X. BRICKS AND OTHER CLAY PRODUCTS 222-256 GENERAL Clay Products as Structural Materials, 229. MANUFACTURE OF BUILDING BRICKS Kinds of Clay. Their Use in Brick Making, 231. Influence of Kind of Clay upon Character of the Brick, 232. Hand Processes op Manupacturb Preparation of the Clay, 233. Pugging, 234. Molding, 236. Drying, 236. Pressing, 237. TABLE OF CONTENTS xiii Machine Processes PAGE General, 238. Soft-mud Process, 239. Stiff-mud Process, 240. Dry Cl^y Proc- ess, 241. Kilns and Burning, 242. Sorting and Classification. Uses of Various Grades, 243. MANUFACTURE OF SAND-LIME BRICK General, 244. The Sand, 245. The Lime. Quality and Quantity Required, 246. Preparation of the Sand, 247. Preparation of the Lime, 248. Mixing, 249. Pressing the Bgck, 250. Hardening, 251. Uses of Sand-Lime Brick, 252. MANUFACTURE OF PAVING BRICK General, 253. The Clay, 254. Molding and Drying, 255. Burning, Annealing and Sorting, 256. MANUFACTURE OF FIREBRICK General, 257. Acid Bricks, 258. Basic Brick, 259. Neutral Firebricks, 260. TERRA COTTA General, 261. Decorative Terra Cotta. Terra Cotta Lumber, Building Blocks, and Fireproofing, 262. ROOFING, WALL, AND FLOOR TILES Roofing Tiles, 263. Wall Tile, 264. Floor Tile, 265. DRAIN TILE AND SEWER PIPE Drain Tilef, 266. Sewer Pipe, 267. PROPERTIES OF BRICKS OF ALL CLASSES Crushing Strength, 268. Absorbing Power, 269. Transverse Strength, 270. Shearing Strength, 271. Modulus of Elasticity, 272. BRICK MASONRY General, 271. The Mortar and the Joints, 274. Laying the Brick, 276. STRENGTH OF BRICK MASONRY General, 277. Strength of Brick Masonry Shown by Tests, 278. Pressures Allowed in Practice, 279. xiv TABLE OF CONTENTS PART II THE FERROUS METALS PAGE CHAPTER XI. PIG IRON 257-293 GENERAL Historical, 280. Iron and Steel in Construction, 281. General Classification of Iron and Steel, 282. THE RAW MATERIALS OF THE IRON INDUSTRY Ores or Ikon General, 283. " Hematite, 284. Limonite, 285. Magnetite, 286. Iron Carbonate, 287. Extent of Ore Production in the United States, 288. Ore Mining and Transportation, 289. Special Preliminary Treatment of Ores, 290. The Flux Necessity of Use of Flux, 292. Fluxes Used, 293. The Fuel General 294. Coal, 295. Coke, 296. Charcoal, 297. Coke Manufacture, 298. Charcoal Manufacture, 299. Relative Use of Different Fuels, 300. REDUCTION OF IRON ORES . Manufacture of Pig Iron General, 301. The Blast-furnace Process in General, 302. The Blast Furnace and Its Mechanical Equipment, 303. Hot-blast Stoves, 304. The Blowing Engines, 305. Drying the Blast, 306. The Functions op the Blast FuhNACE General, 307. Deoxidation of the Iron Ore, 308. Carburization of the Iron, 309. Melting the Iron, 310. Conversion of Gangue to Fusible Slag, 311. Separation of Iron and Slag, 312. Operation op the Blast Furnace Starting the Furnace, 313. Mechanical Control of Furnace and Accessories, 314. Metallurgical Control of Furnace, 315. Action within the Furnace, 316. Handling the Products, 317. Electric Reduction op Iron Ores General Considerations, 318. The Electric Furnace, 319. Quality of the Product 320. ' THE USES OF PIG IRON Classification of Pig Irons, 321- The Uses of Pig Iron, 322. Production of Pie Iron, 323 * TABLE OF CONTENTS xv PAGE CHAPTER XII. CAST IRON 294-332 INTRODUCTORY, PIG IRON PRODUCTS CLASSIFIED Pig-iron Products, 324. Cast Iron as a Material of Engineering Construction, 325. THE REMELTING OF PIG IRON Iron Melting in General, 326. The Materials Used Foundry Pig Iron, 327. Scrap Iron, 328. The Flux, 329. The Fuel, 330. The Furnace The Cupola Furnace and Its Equipment, 231. The Reverberatory or Air Furnace, 332. Relative Use of Cupola and Air Furnace, 333 Operation op Cupola Starting the Furnace, 334. Charging, 335. Action within the Furnace, 336. Chemical Changes, 337. Duration of the Cupola Run, 338. Tapping Out and Stopping In, 339. Operation op Air Furnace The Charge, 340. Control of Melting, 341. Advantages and Disadvantages, 342. IRON FOUNDING Iron Founding in General, 343. Molds and Molding, 344. Green-sand Molding, 345. Patterns and Cores, 346. Making a Mold in Green-sand, 347. Dry- sand Molds, 348. Loam Molds, 349. Chilled Castings, 350. Pouring the Iron, 351. Cleaning the Castings, 352. PROPERTIES OF CAST IRON Constitution Essential Constituents of Cast Iron, 353. Carbon in Cast Iron, 354. Gray Cast Iron, 355. White Cast Iron, 356. Mottled Cast Iron, 357. Silicon in Cast Iron, 358. Sulphur in Cast Iron, 359. Phosphorus in Cast Iron, 360. Manganese in Cast Iron, 361. Behavior op Iron in Cooling Shrinkage, 362. Checking, 363. Segregation, 364. Chilling, 365. PHYSICAL AND MECHANICAL PROPERTIES Hardness, 366. Tensile Strength. Compressive Strength. Cross-breaking Strength. Tensile Strength in General, 367. Influence of Form of Carbon, 368. Influence of Metalloids and Rate of Cooling upon Strength, 369. Stress-Strain Diagram for Cast Iron, 370. Specification and Allowable Stress, 371. Compressive Strength of Cast Iron, 372. Allowable Stress in Compression, 373. Cross- breaking Strength, Modulus of Rupture, 374. Specification and Allowable Cross-bending Stress, 375. xvi TABLE OF CONTENTS PAGE CHAPTER XIII. MALLEABLE CAST IRON. . 333-340 GENERAL Definition of MaUeable Cast Iron, 376. Malleable Cast Iron as a Material of Engineering Construction, 377. MANUFACTURE OF MALLEABLE CAST IRON The Material Used, 378. The Furnace, 379. Melting Malleable-iron Mixtures, 380. Molding Methods for Malleable Castings, 381. Pouring the Castings, 382. Subsequent Treatment of the Castings, 383. The Anneahng Process, 384. Treatment of Annealed Castings, 385. PROPERTIES AND USES OF MALLEABLE CASTINGS Chemical Composition and Constitution, 386. Physical Properties. Tensile Strength and Ductility, 387. Transverse Strength, 388. Tougbness and Shock Resistance, 389. Uses of Malleable Castings, 390. CHAPTER XIV. WROUGHT IRON 341-359 GENERAL Historical, 391. Definition of Wrought Iron, 392. Wrought Iron as a Mate- rial of Engineering Construction, 393. MANUFACTURE OF WROUGHT IRON The Wet-puddling Process The Puddling Process in General, 394. The Iron Used, 395. The PuddUng Furnace, 396. Preparation of Furnace for Charging, The Fettling, 397. Furnace Operation, Chemical and Physical Changes, 398. Removal of Slag, Squeezing or Shingling, 399. Rolling-mill Operations, 400. Mechanical ■"uddling, 401. Wrought Iron from Scrap, 402. PROPERTIES AND USES OF WROUGHT IRON Composition and Constitution, 403. Classes of Wrought Iron, 404. Tensile Strength and Elongation, 405. Relationship between Tensile Properties of Wrought Iron and Reduction in Rolling, 406. Effect of Previous Strain- ing or Cold Working upon Tensile Properties, 407. Heat Treatment and Crystalline Structure, 408. Compressive Strength of Wrought Iron, 409. Shearing Strength of Wrought Iron, 410. The Welding of Wrought Iron, 411. CHAPTER XV. STEEL 360-499 GENERAL Definition, 412. Classifications of Steels, 413. Steel as a Material of Engineei^ ing Construction, 414. MANUFACTURE OF STEELS General Steel-making Processes, 415. TABLE OF CONTENTS xvii Cabbubization of Wkougiit Iron The Cementation Process, 416. The Crucible Process General, 417. The Coke-Fumace or Melting Hole, 418. The Gas-fired Regenera- tive Furance, 419. The Charge of the Crucible, 420. Operation of Process, 421. Grades of Crucible Steel, 422. Cost of Crucible Steels, 423. The Bessemer Process Historical, 424. The Bessemer Process in General, 425. The Pig Iron Used, 426. The Bessemer Converter and Other Equipment of the Bessemer Plant, 427. Operation of Process, 428. Chemistry of Process, 429. Heat Development and Utilization, 430. Recarburizers and Reoarburizing, 431. Deoxidation, 432. Casting the Ingots, 433. The Basic Bessemer Process General, 434. The Basic Converter, 435. The Pig Iron Used, 436. Operation of the Basic Process, Chemistry of Process, 437. Recarburization, 438. Comparison of Acid and Basic Bessemer Processes, 439. The Open-hearth Process General Historical, 440. The Open-hearth Process in General, 441. The Open-hearth Plant, 442. The Furnace and Its Operation General Features of the Regenerative Furnace, 443. Construction of Open-hearth Furnace, 444. Stationary vs. Tilting Furnaces, 445. Life of Furnace and Repairs, 446. The Furnace Fuel Natural-gas Fuel, 447. Producer Gas, 448. The Basic Open-hearth Process General, 449. The Furnace Charge, Operating Practice, 450. Chemistry of the Basic Process, 451. Recarburization, 452. Pouring the Ingots, 453. The Acid Open-hearth Process General, 454. The Furnace Charge, 455. Chemistry of Acid Process, 456. Recarburizing, 457. Special Open-hearth Processes The Duplex Processes, 458. The Talbot Process. The Monell Process, 459. Electric Refining of Steel Electric Refining Processes in General, 460. Types of Electric Refining Furnaces, 461. Open-arc Furnaces. The Stassano Furnace, 462. Arc-resistance Fur- naces. The H^roult, Girod, and Keller Furnaces, 462. Induction Furnaces. The KjelUn and RochUng-Rodenhauser Furnaces, 464. AppUcations and Limitations of Electric Furnaces and Electric Refining Processes, 465. xviii TABLE OF CONTENTS Rolling Mill Operations. The Finishing op Steel PAGE Reheating Necessity for Reheating, 466. Reheating Furnaces, Practice of Reheating, 467. Eolling General, 468. Rolling Mills, 469. Finishing Steel by Steam Hammer and by Presses Forging under the Steam Hammer, 471. Forging by Means of Presses, 472. Defects in Ingots and TpEiR Correction Blow-holes, Piping, Ingotism, and Segregation, 473. The Heat Treatment op Steel Practice of Hardening, Tempering, Annealing and Case Hardening General, 474. Hardening Steel, 475. Tempering Steels, 476. Annealiiig Steels, 477. Oase-Hardening Steels, 478. THE PROPERTIES AND USES OF STEELS Structure and Constitution The Constituents of Steels, 479. Compounds and Solid Solutions. Eutectics, 480. Phenomena of Slow Cooling op Iron-carbon Alloys Freezing of Iron-carbon Alloys, 481 . Changes in Cooling below the Freezing-Point, • 482. Phenomena op Rapid Cooling Followed by Various Reheating Treatments Effect of Various Heat Treatments upon the Structure and Constitution of Steels, 483. The Physical Properties op Steels Grades of Steel and General Properties, 484. Tensile and Compressive Strength in General, 485. Behavior of Steel under Stress in General, 486. Effect of Carbon upon Physical Properties, 487. Effect of Heat Treatment and Mechanical Working, 488. Effect of SiUcon, Sulphur, Phosphorus, and Manganese, 489. Shearing Strength Direct Shear. Torsion, 490. Trans- verse Strength, Flexure and Deflection, 491 . Effect of Combmed Stresses upon Elastic Properties, 492. Hardness of Steels, The Brinell Method, The Cone Test, The Shore Scleroscope, The Bauer Drill Test, 493. Ductility of Steels as Indicated by Cold Bending, 494. Behavior of Steel under Impact, Shock Resistance, 495. Behavior of Steels under Repeated and Alternating Stresses, Fatigue, 496. The Magnetic Properties of Steel, Permeability, Hystereses, Eddy Current Loss, Core Loss, Relation between Magnetic Properties and Chemical Composition, Relation between Magnetic Properties and Tem- perature, Relation between Magnetic Properties and Mechanical and Ther- mal Treatment, Relation of Magnetic to Mechanical Properties, 497. The Corrosion of Iron and Steel, The Carbonic Acid Theory, The Hydrogen Peroxide Theory, The Electrolytic Theory of Corrosion, 498. Allowable Unit Stresses for Steels, 499. Production of Steel in the United States Statistics, 500. TABLE OF CONTENTS xix PAGE CHAPTER XVI. THE SPECIAL ALLOY STEELS 500-540 General. .Definition and Classification of Alloy Steels, 501. TERNARY ALLOYS Nickel Steel — Critical Changes and Irreversible Transformations; Tensile Prop- erties, Low-Carbon Steels, Medium-carbon Steels, High-Carbon Steels; Relation of Tensile Strength and Yield Point to Carbon Content and Heat Treatment; Relation of Ductility to Carbon Content and Heat Treatment; Resistance to Alternating Stresses; Impact Strength; Magnetic Properties; CorrodibiUty; Manufacture and Uses, 502. Manganese Steel— Structure and Constitution; Tensile Properties; Manufacture and Uses, 503. Chrome Steel — Structure and Constitution; Thermal Critical Points; Tensile Proper- ties; CorrodibiUty; Manufacture and Uses, 504. Tungsten Steel — Structure and Constitution; Tensile Properties; Manufacture and Uses, 505. Molyb- denum Steel — Structure and Constitution; Tensile Properties; Ductility as Indicated by Cold Bending; Alternating Stress Resistance; Brinell Hardness; Manufacture and Uses, 506. Silicon Steel — Tensile Properties; Manufacture and Uses, 507. Vanadium Steels — Structure and Constitution; Tensile Prop- erties; Manufacture and Uses, 508. QUATERNARY ALLOYS The Quaternary Alloys — Nickel-Manganese Steel; Nickel-Chromium Steel; Nickel-Tungsten Steels; Nickel-Vanadium Steels; Nickel-Silicon Steels; Chromium-Manganese Steels; Manganese-Silicon Steels; Tungsten-chromium Steels, 509. PART III THE NON-FERROUS METALS AND ALLOYS AND TIMBER CHAPTER XVII. THE NON-FERROUS METALS AND ALLOYS 541-604 THE PURE METALS GENERAL The Non-Ferrous Metals of Industrial Importance, 510. COPPER General, Classification of Commercial Forms of Copper, 511. Occurrence in Nature, Ores of Copper, 512. The Extraction of Copper from its Ores, Roasting, Smelting and Converting, Pjrrite Smelting, Alternate Oxidation and Reduction, Refining of BUster or Coarse Copper, Fire Refining, 513. The Properties and Uses of Copper, Electrical Resistivity, Tensile Properties, Uses of Copper, 514. Production of Copper, Statistics of Copper Production, 515. X3C TABLE OF CONTENTS ZINC General, Commercial Forms of Zinc, 516. Occurrence in Nature, Ores of Zinc, 517. Extraction of Zinc from its Ores, Preliminary Treatment, Concentration, Calcination and Roasting, Distillation and Condensation, Refining Crade Spelter, 518. Properties and Uses of Zinc, 519. Statistics of Zinc Produc- tion, 520. LEAD General, Commercial Forms of Lead, 521. Occurrence in Nature, 522. Extrac- tion of Lead from its Ore, 523. Properties and Uses of Lead, 524. Statistics of Lead Production, 525. TIN General, Commercial Forms of Tin, 526. Occurrence in Nature, Tin Ore, 527. Extraction of Tin from its Ore, 528. Properties and Uses of Tin, 529. Statistics of Tin Production and Consumption, 530. ALUMINUM General, Commercial Forms of Aluminum, 531. Occurrence in Nature, 532. Extraction of Aluminum, 533. Properties and Uses of Aluminum, 534. Statistics of Production and Consumption of Aluminum, 535. NICKEL General, Commercial Forms of Nickel, 536. Occurrence in Nature, 537. Extrac- tion of Nickel from its Ores, the Smelting of Sulphur Compounds, the Extraction of Nickel from its Silicates, 538. The Properties and Uses of Nickel, 539. Statistics of Nickel Production and Consumption, 540. THE NON-FERROUS ALLOYS GENERAL The Non-Ferrous Alloys in General, 541. COPPER-TIN ALLOYS. BRONZES Ordinary Bronzes, 542. Special Bronzes, Copper-Tin Bronzes, Copper-Tin-Lead Bronzes, Phosphor Bronze, Manganese Bronze, Aluminum Bronze, Vanadium Bronze, Nickel Bronzes, 543. COPPER-ZINC ALLOYS. BRASSES Ordinary Brasses, 544. Special Brasses, Copper-Zinc-Lead Brasses, Copper- Zinc-Aluminum Brasses, Copper-Zinc-Manganese Alloys, Manganese Brasses, Copper-Zinc-Iron Alloys, Other Special Brasses, 545. BINARY ALLOYS OF COPPER OTHER THAN BRONZES AND BRASSES Copper-Aluminum Alloys, 546. Binary Alloys of Copper with Manganese, Phos- phorus, Silicon, etc., 547. ALLOYS OF ZINC, LEAD, TIN, ALUMINUM, AND NICKEL Binary Alloys of Zinc, 548. Binary Alloys of Lead, 549. Binary Alloys of Tin, 550. Aluminum Alloys, 551. Alloys of Nickel, German Silver, 552. Special Bearing or Anti-friction Metals, 553. TABLE OF CONTENTS xxi PAGE CHAPTER XVIII. TIMBER 605-658 GENERAL Timber as a Material of Engineering Construction, 554. TIMBER WOODS, GROWTH AND STRUCTURAL CHARACTERISTICS Classes of Trees, 555. Exogenous Trees Conifers, 556. Broadleaf Trees, 557. Endogenous Trees Endogenous Trees. Monocotyledons, 558. Exogenous Growth of Wood Pith, Wood, and Bark, 659. Primary Wood, Cambium, and Secondary Wood, 560. Structural Elements of Wood, 561. Rays, Resin Ducts, and Pith Flecks, 562. Annual Growth Rings, Spring and Summer Wood, 563. Sap- wood and Heartwood, 564. Endogenous Growth op Wood Endogenous Growth, 565. Structure op Wood op Exogens Structure of Wood of Conifers, 566. Structure of Wood of Broadleaf Trees, 567. PHYSICAL CHARCTERISTICS OF WOOD Grain and Texture of Wood, 568. Color and Odor, 569. Density and Weight, 570. Moisture Content of Wood, 571. Seasoning of Timber, 572. Shrink- age, Warping and Checking in Drying, 573. MECHANICAL PROPERTIES OF WOOD General, 574. Tensile Strength, 575. Compressive Strength, 576. Cross- breaking Strength and Stiffness, 577. Moisture and Strength, 578 Weight and Strength, 579. Rate of Growth, Proportion of Summer Wood, and Strength, 580. The Time Factor in Tests of Timber, 581. Tabulation of Mechanical Properties of Structural Timbers, 582. Factor of Safety and Working Stresses, 583. Durability and Decay of Timber, 584. Preservation of Timber, The Pressure Processes, The Non-pressure Processes, The Empty Cell Non-pressure Processes, Superficial Treatments, Effect of Preservative Treatments upon Strength of Timber, 585. MATERIALS OF CONSTRUCTION PART I THE MATERIALS OF MASONRY CONSTRUCTION CEMENTING MATERIALS CHAPTER I GYPSUM PLASTERS GENERAL 1. Definition and Classification. Gypsum plasters comprise all that class of plastering and cementing materials obtained by the partial or complete dehydration of relatively pure or impure natural gypsum, and to which certain materials which serve as retarders or hardeners, or which impart greater plasticity to the product, may or may not have been added during or after calcination. The usual classification of gypsum plasters follows: Classification of Plasters * (a) Produced by the incomplete dehydration of gypsum, the cal- cination being carried on at a temperature not exceeding 190° C. (1) Plaster of Paris, produced by the calcination of a pure gypsum, no foreign materials being added either during or after calcination. (2) Cement Plaster (often called Patent or Hard Wall Plaster), produced by the calcination of a gypsum containing certain natural * Eckel, " Cements, Limes, and Plasters." 2 MATEEIALS OF CONSTRUCTION impurities, or by the addition to a calcined pure gypsum of certain materials which serve to retard the set or render more plastic the product. (b) Produced by the complete dehydration of gypsum, the cal- cination being carried on at temperatures exceeding 190° C. (3) Flooring Plaster, produced by the calcination of a pure gypsum. (4) Hard Finish Plaster, produced by the calcination at a red heat or over, of gypsum to which certain substances (usually alum or borax) have been added. 2. Application in the Arts and Industries. Raw gypsum, ground but not calcined, is used extensively as a soil amendment and as a slight necessary adulterant to retard the setting of Portland and natural cements. It is also used as an ingredient of some oil paints, many water paints, wall tints and calcimines, in some dry colors and crayons, and in imita- tion ivory, meerschaum, etc. A pure white even-grained variety called alabaster is considerably used for sculptural decorations, art work, etc. Calcined plasters are used as a retarder in cements, as a casting plaster for stereotype molds, pottery, rubber stamps, etc., as dental plaster, as- a cement to bed plate glass during grinding, as an ingredient of many patented cements, and as a seal or cement used in the construction of innumerable small instruments, electric fittings, etc. 3. Structural Uses. Gypsum plasters of one variety or another are used structurally as interior wall plasters, both for mortar and hard- finish coats, as stucco wall coverings and architectural ornamentation of exteriors, as molded hollow blocks and tiles for interior partition walls, as plaster-board wall covering (laminated with thin sheets of card- board or wood), for flooring plasters, and for numerous minor structural purposes. MANUFACTURE OF GYPSUM PLASTERS 4. Gypsum Rocks. Pure gypsum is a hydrous lime sulphate (CaS04+2H20), the composition of which by weight is: f Lime Sulphate (CaS04) I ^fl^^f)'- ' ;_■ ' " '^^-^^ ] 79 1% \ f y *j [ Sulphur tnoxide (SO3). .46.5 J ■" i Water (H2O) • 20.9 100 0% Natural deposits of gypsum are practically never pure, however, the lime sulphate being adulterated with silica, alumina, iron oxide, calcium carbonate, and magnesium carbonate. The total of all impurities amounts to from a few tenths of 1 per cent up to a maximum of 5 or 6 per cent. GYPSUM PLASTERS 3 The physical form of a natural gypsum is usually that of a massive rock formation. It does occur also as an earthy gypsum or gypsite, and as gypsum sands in some localities in the southwestern part of the United "States. Alabaster is a specially pure white massive gypsum of very even texture and fine grain, and selenite is a white semi-transparent crystalline gypsum which occurs only in relatively small deposits in massive gypsum. 5. Theory of Calcination. If pure gypsum be subjected to any temperature exceeding 100° C. by more than a very slight amount, but not exceeding 190° C, a certain portion of the water of combination is driven off. This portion is definitely fixed at three-fourths of that originally present, i.e., 1^ molecules, providing that the heating be con- tinued for a sufficient period of time to accomplish the ultimate extent of dehydration possible without heating above 190° C. The resultant product is called Plaster of Paris (CaS04-l-l/2H20). Since the time required for the process is directly dependent upon the temperature maintained, it is the quite universal practice to maintain this temper- ature near the highest possible limit, thus effecting an economy in both time and fuel. Thus it happens that plaster of Paris and cement or hard wall plasters are rarely calcined at temperatures outside the limits 140° to 180° C. Plaster of Paris readily recombines with water to form gypsum, hardening in a very few minutes. If the gypsum be calcined at temperatures much above 190° C. it loses all of its water of combination, becoming an anhydrous sulphate of lime (CaS04) . AH temperatures exceeding that required for complete dehydration result in some impairment, temporary or permanent, of the capacity of the plaster for .recombination with water, the extent of the injury being dependent upon the intensity and the time-extent of the heating and upon the state of subdivision of the material. It is not at all difficult to render the plaster totally incapable of recombining with water (dead-burnt plaster), the product being valueless as a cementing material. If, however, the material is calcined in a lumpy condition the temperature not exceeding 400° to 500° C, and not prolonged beyond three or four hours, the principal effect upon the products is a great retard- ation of the rate of setting and hardening, an hydrate being ultimately formed, in the course of days or weeks, instead of minutes, which' greatly exceeds ordinary plaster in hardness and strength. These principles govern the manufacture of flooring plasters and hard-finish plasters, though there are differences in the methods employed in producing these two varieties of hard-burned plasters which will be later noted. — "6. Practice of Calcination. Plaster of Paris and cement or hard wall plasters are made in a practically identical manner, the distinct 4 MATEEIALS OF CONSTRUCTION properties of the two materials being due to the use of pure gypsum in the one case, and impure gypsum or adulterated pure gypsum in the other. Two operations only are involved in the process of manufacture: crushing and grinding, and calcination. Three-fourths of the plaster plants in the United States use gypsum rock alone, about one-sixth of the plants use earthy gypsum or gypsite, and the balance use both rock and earthy gypsum. Rock gypsum is first crushed in a gyratory * or jaw * crusher to fragments about 1 inch in diameter. Intermediate reduction to about the size of peas is accomplished in a pot * crusher or cracker,* and final pulverization in a buhr-mill,* a rock-emery mill,* or other type of finishing mill. Gypsite requires no preliminary crushing, but is handled by the finishing mill alone, and material destined to be calcined in a rotary cylinder calciner is not pulverized until after cal- cination. Ninety-five per cent of the plaster made in the United States is cal- cined in kettles of the general type of that shown in Fig. 1. The kettle is a cylindrical steel vessel, 8 to 10 feet in diameter and 6 to 9 feet high, mounted upon a masonry foundation. The bottom is convex, rising about 1 foot in the center, and made of cast iron. A masonry wall encloses the steel cylinder, leaving an open annular space between for the circulation of heat. A fire is maintained on grates below the kettle, and the heated gases pass through ports into the open annular space, then in horizontal flues through the kettle and out through, the stack. A central vertical shaft propels paddles just above the bottom, thus keep- ing the material agitated and preventing the burning out of the bottom. The charge, consisting of 7 to 10 tons of ground gypsum, is delivered by a chute to the charging door provided in the sheet-iron cap of the kettle. Heat is gradually applied as the charge is slowly fed in and, as the temperature rises after charging is complete, the contents boil vio- lently until the mechanically held water is driven oH. Boiling is renewed again when the water of combination begins to be driven off, and con- tinues until the end of the process. The maximum temperature attained in a period of two to three hours does not usually exceed 190° C. (It may be considerably higher in the case of the less pure earthy gypsum.) The point at which the process is complete is sometimes determined by careful observation of temperatures throughout, but is often a matter of judgment upon the part of an experienced calciner. The kettle is discharged by blowing out through a small gate in the lower part of the side of the shell. The hot material runs like water into a masonry pit provided for the purpose and, after cooling therein, is elevated to screens * These various types of crushing and grinding machinery are described and illus- trated on pages 119 to 122 incl. of this volume. GYPSUM PLASTERS 5 which remove a small percentage of over-size material which must be ground. The rotary cylinder type of calciner, Fig. 2, is a late development in the plaster industry, but is doubtless destined to replace the less eco- Horizontal Sections. Vertical Sections. Fig. 1. — Four Flue Plaster Calcining Kettle. nomical kettle. When the rotary calciner is employed the raw material is used in the condition in which it comes from the cracker, and feed- ing from the supply bin is continuous. The cylinder is set on a slight incline, and the lumpy material fed in at the upper end gradually traverses its length as the cylinder slowly rotates, is discharged at the lower end, and enters calcining bins which are lined with non-absorptive paving 6 MATERIALS OF CONSTRUCTION brick, and from which outside air is excluded. The cylinder is enclosed within a brick chamber which includes a firebox at one end, and a considerable air space is provided between the steel shell and the masonry. The hot furnace gases are drawn into the brick chamber by a fan and there mixed with sufficient air, admitted from the outside, to secure the desired temperature. These gases are drawn into the cylinder through hooded openings provided at intervals, and pass through it m a reverse direction to that taken by the wet material. A dust chamber provided between the cylinder and the stack catches the most finely ground plaster which has been carried off in suspension by the gases. Fig. 2. — Rotary Cylinder Type of Plaster Calciner. The heat attained in the rotary calciner is from 200° to 300° C, but time does not suffice for the complete dehydration of the gypsum, and the removal of combined water is completed in the calcining bins through the agency of the residual heat of the material itself. After about thirty-six hours the process will have been completed. Air inlets are then opened and the contents of the bin are rapidly cooled. By providing four bins for each calciner the process is made to be continu- ous; while one bin is being charged, the process of calcination is being completed in the second and third bins, and the fourth is being discharged. The product of the calcining bins is conveyed to finishing mills and there pulverized to the form of the marketable article. GYPSUM PLASTERS 7 Many cement or hard wall plasters are made direct from earthy gypsum or gypsite, which is often found to contain a suitable percentage of foreign material of such character that no corrective material need be added either to retard the set or impart plasticity. Flooring plaster is produced by the calcination of a pure gypsum in a lump form in a vertical separate-feed kiln which differs little from the separate-feed kiln used for the calcination of lime.* The fuel-, burned on grates outside the kiln, does not come in contact with the gypsum, but the hot gaseous products of combustion pass directly through it, heating it to a temperature of from 400° to 500° C. Higher temperatures, or prolongation of heating beyond three or four hours, ruin the plaster by robbing it of its setting properties as above noted. Fine pulveriza- tion of the plaster must follow calcination. The best-known variety of hard-finish plaster is the so-called Keene's cement. This plaster is produced by the double calcination of a very pure gypsum at a red heat, alum being added between the two burnings. The kiln used resembles very closely the mixed-feed lime kiln herein- after described, t After the lump gypsum has been calcined at a red heat, the resulting anhydrous lime sulphate is immersed in a 10 per cent alum solution, then recal.cined, and finally pulverized in a finishing mill. -— ~ 7. Additions Subsequent to Calcination. Plaster of Paris is never adulterated in any way . during manufacture, but cement or hard wall plasters often require the addition of a retarder to. render them suf- ficiently slow setting to adapt them to structural use. The retarders commonly used are unconsolidated organic materials, such as glue, saw- dust, blood, packing-house tankage, etc. As a rule the amount of retarder required does not exceed 0.1 to 0.2 per cent. Certain very impure gypsums produce a plaster which is too slow setting, or some- times, as in the case of dental plaster, extreme rapidity of set is required. In such instances the addition of an accelerator is necessary. The materials used for this purpose are crystalline salts, common salt (NaCl) being one of the best. Gypsum plasters destined for use as wall plasters must usually have their plasticity enhanced by the addition of some material such as clay or hydrated lime; through the agency of which the naturally " short," non-plastic material is greatly improved in working quaUties and sand- carrying capacity. With the exception of those plasters which are made from earthy gypsum and which naturally contain 20 per cent or more of clay, it is therefore the usual practice to add about 15 per cent of hydrated lime or, less frequently, clay, to the calcined plaster. Greater cohesiveness is also usually imparted to wall plasters by the addition * See Fig. 12. t See Fig. 11. 8 MATERIALS OF CONSTRUCTION of finely picked hair or shredded wood fiber. Not more than 1| to 3 pounds of hair or 75 to 150 pounds of wood fiber per ton of calcined plaster is required. No additions are made to flooring plaster subsequent to calcmation. Keene's cement is treated with an alum bath as above noted. Mack's cement, another variety of hard-finish plaster, is made by the addition to dehydrated gypsum (flooring plaster) of 0.4 per cent of calcined sodium sulphate (Na2S04), or potassium sulphate (K2S04). PROPERTIES AND USES OF GYPSUM PLASTERS 8. Setting and Hardening.* Action of Accelerators and Retarders. The setting of plaster of Paris and other gypsum plasters is a simple process of recombination of the partially or totally dehydrated lime sulphate with water to reform hydrated lime sulphate or gypsum. A pure plaster of Paris sets in from five to fifteen minutes after the addi- tion of water. Plasters made from impure gypsum are less quick setting, requiring from one to two hours, and the completely dehydrated classes of plasters are slow setting, whether adulterated or not, because of the action of high temperatures in retarding the rate of setting. The ulti- mate degree of hardness attained by impure cement or hard wall plasters greatly exceeds that of pure plaster of Paris, and the hard-burned plasters are hardest of all. The action of accelerators and retarders, according to Dr. Grimsley,t is based upon the asserted fact that crystalline growth cannot occur unless there are present some crystals which have escaped dehydration during burning, and which act as promoters of hydration, inducing neigh- boring molecules of lime sulphate to assume the crystalline form by the taking up of water of crystallization. The theory further involves the acceptance of the hypothesis that the progressive hardening and gain in strength is due to the formation of a mass of interlocking crystals. The action of accelerators and retarders is therefore explained by assuming that " any substance (added to the water with which the calcined plaster is mixed or to the dry plaster) which will keep the molecules apart or from too close contact will retard the setting. Such substances are dirt or organic matter that is not of a crystaUine character." On the other hand, mineral salts induce crystaUization because of their crystaUine character. * By the term " setting " we mean the initial loss of plasticity, while " harden- ing " means the subsequent gain in ability to resist indentation or abrasion which accompanies the gain in cohesiveness. t Reports, Geological Survey of Kansas, Vol. 5, p. 167. GYPSUM PLASTERS 9 9. Strength of Plasters. Trade conditions in the plaster industry- have not yet reached the point where contracts for plasters are made con- tingent upon their satisfactorily meeting the requirements imposed by a sei-ies of physical and mechanical tests, as is the case in the Portland cement industry. In consequence, scientific investigation of the prop- erties of plasters has been undertaken to a very limited extent only, and the value of the scanty data which bear upon the question, and which are available mainly in the pubUcations of the various State geological surveys, is impaired by the fact that, methods of examination not having been standardized, each investigator has applied whatever tests he may have deemed advisable, though all have followed more or less closely the methods of cement testing, which have been fairly well standardized. e 8 IV 13 Tensile Strength The tensile strength of plasters is dependent upon so many con- siderations incidental to methods and conditions of testing that little data as to tensile strength in general will herein be intro- duced. Discussion of the subject will largely be devoted to a con- sideration of the effect of various factors upon tensile strength as shown by compara- tive tests made in a similar manner under identical exterior con- ditions. Tensile strengths of a number of varie- ties of plasters are shown by Fig. 3, which is a summary of tests made in the laboratory of Washburn College under the direction of Dr. G. P. Grim- sley of the University Geological Survey of Kansas. * Briquettes of the standard form used for cement were molded one at a time from a plaster- water mixture of rather stiff consistency. Storage throughout the period of testing was in the air of the laboratory, and breakage was accomplished in one of the common types of cement briquette-testing machines. The percentage of water used was such a one as was found for each particular * Geological Survey of Michigan, Vol. 9, pp. 168-169. "* ~" — — — — — — — a ^ U P as er iL a ^ -^ ' p£V4 ^000 ^ L. 9 ^ fir IV ill Ph ^ -J .^ Ph t'i iBt, 3r =1 __^ \ _ ■~. u2. jii ' — — c TENSILE:STRENGTH OP NEAT WA1.L PLASTERS o S ^200 Age in Weeks Fig. 3. 10 MATERIALS OF CONSTRUCTION plaster to be sufficient to moisten the plaster so that, when- struck with a trowel, a moist surface would appear but no water be caused to stand on the surface. This amount varied for different plasters between the limits of 26.3 and 40.0 per cent. The curves are averages of the results obtained in from 3 to 6 tests of each brand of plaster. One brand of plaster of Paris was tested, two brands of hard wall plaster made from gypsite, three brands of hard wall plaster made from rock gypsum, and one brand of Keene's cement. The relative strengths of the various plasters shown by the curves should not be accepted as conclusive evidence that plaster of Paris is superior in strength to hard wall plasters, and that plasters made from earthy gypsum excel those made from rock gypsum. Other tests do not always confirm this showing, and the variation on the results of indi- vidual tests is too great, and the number of tests too small to make the results of these tests altogether dependable. The curves do illustrate, however, certain characteristics of plasters which have been universally observed, viz.: (1) The gain in strength of neat plasters is very rapid for the first few days only: (2) The maximum tensile strength of neat plasters is attained in a period of from two to four weeks, after which retrogression in strength invariably occurs except in the case of hard-finish plas- ters. The actual values of neat tensile strength shown by these plasters (about 400 to 500 pounds per square inch at one week and 400 to 600 pounds per square inch at one month) are probably 50 to 100 per cent higher than the values usually attainable in practice, because of the lower water percentages and the greater degree of compacting utilized by the investigator in comparison with conditions possible in construc- tion work. The comparative strength of neat hard wall plaster and various sand mortars is shown by the curves of Fig. 4 which constitute a summary of a series of tests made by Professor A. Marston of Iowa State College.* These curves represent the average results obtained in tests of two brands of hard wall plaster, from 3 to 5 separate tests having been made on each brand. It appears from the curves that the tensile strength of a mortar contammg 1 part of sand to 1 part of plaster is about 86 per cent of the strength of neat plaster; that mortar containing 2 parts of sand to 1 part of plaster possesses 50 to 60 per cent of the strength of neat plaster; and that mortar contaming three parts of sand to 1 part of plaster possesses * Iowa Geological Survey, Vol. 12. GYPSUM PLASTERS 11 1 N pat ^ -r ^ _ _ _ 1:1 _ _ _ _ _ J _ i L ^ =- = = - - — - - ~ - - - - - 1 V - = , — ^ - __ _. - — ' 1 f 1 ,3 „ „ < " "■ — = ~ 1 COMPARATIVE TENSILE STRENGTH OF NEAT PLASTER i VARIOUS MORTARS - M 1 1 1 1 1 L 5 1 . 6 Age 7 8 la Weeka 9 10 11 12 13 about 33 per cent of the strength of neat plaster. The fact must not be overlooked in this connection that the strength of mortar mixtures is largely dependent upon the character of the sand used, the prevail- ing size, granulomet- ric composition, clean- liness, sharpness, etc.* a One other consider- » ation which consider- | ably affects the sand- f™ carrying capacity of a § plaster is the degree |,2oo of fineness possessed | by that plaster. This « fact is illustrated by g Fig. 5, which affords a comparison of the strength of a 1:2 mortar, the plaster Fig. 4. being used as mark- eted, and a similarly proportioned mortar, the plaster used being only that portion of the original sample which has been passed by a sieve of 100 meshes per lineal inch. These curves represent the average of from 3 to 5 tests upon each of two brands of plaster. The curves are based upon data secured by Professor Marston in the series of tests cited in the preceding paragraph. It appears that an advantage in tensile strength in mortar mixtures of about 15 per cent would be secured were all of the plaster fine enough to pass the 100-mesh sieve. The advantage of increased fineness is therefore sufficient to be taken account of, but it is far less marked in plasters than it is in Portland cement, as will be later noted. Indeed, it is not improbable * See discussion of sands for mortars, pages 151 to 154 incl. ~ — — — — — — ~ c crSOO g _ _ _ _ :i L ttl EJ rlt uS i£t ~ ■" 1 K' / £ "100 COMPARATIVE TENSILE STRENGTH OF ORDINARY r:2-pLA?TER MORTAR &. 1:2 MORTAR MADE BY USING PORTION OF PLASTER WHJCH PASSES #100 SIEVE CSSje passed f 100 Sievs) d H Ml 1 1 1 1 1 1 1 1 S (j 1 I 2 1 28 AgelnDay^ Fig. 5. 12 MATERIALS OF CONSTRUCTION that the superiority of the finer particles of plaster over the unsifted article is due simply to the fact that the sifting process has excluded a certain amount of insufficiently calcined material. (It will be recalled that the process of calcination results in the removal of the water by the formation of steam, the expansive force of which is responsible for the fine disintegration of the product. All material which fails to pass the 100-mesh sieve, and a considerable portion of that which does pass this sieve, is probably not a true half-hydrate but contains a certain amount of excess water which impairs its ability to set and harden.) The effect of the presence of excess moisture during the period of hardening is shown by the curves of Fig. 6. These curves represent the average results obtained in two tests of each of nine brands of plaster tested by Professor Marston. The fact is strikingly brought out that the average plaster, which after twenty - eight days in air shows a gain in tensile strength of 102 per cent over its one-day strength, absolutely does not gain at all when stored in water twenty-seven days after having hardened in air for twenty-four hours. This fact lends weight to the hypothesis that the retrogression in strength characteristic of all plasters after about one month in air is partially due to the assimilation of excess moisture from the atmosphere. The fact must not be lost sight of, however, that increase in hardness and strength is accompanied by increased stiffness and brittleness. This condition results in increased diffi- culty in so applying the load in testing as to secure a practically uniform distribution of same on the minimum cross-section of a briquette. With a less stiff and brittle material deformation of the specimen to equalize the stress distribution will have occurred long before rupture, but if such deformation does not occur, or if it is not sufficient to accomplish the end desired, the specimen is subjected to some degree of bending and breaks under an average stress far below the actually existent stress in the extreme fiber on the side where rupture started. It is undoubtedly a fact that the apparent retrogression in tensile strength of practically all 380 — [ ~1 — 1 — [ — — — — — ~" _^ ^ — — — _ — ^ _ 340 — — ~ ^ __ am — — — ^ _ •^300 — ~ ^'i _ ^280 __ — \p '"i ' _ ^3*260 ~ " > 0-240 — ~ -O220 — ~ ~ ~ y^ ' _ §200 l2l80 t - -^ _ _ _ St )re dJ lV ^ eT ^ - - ^100 "140 - - ^ - — - — — — — — — — — ~ 1 SqIOO EFFECT OF MOISTURE UPON TENSILE STRENGTH OF PLASTERS Each the Ave. of 18 Tests, 2 Tests made on each of 9 Brands of Plaster # '^^ __ "^ 20 — — ~ 1 M 1 1 1 1 1 1 1 1 1 1 1 ' — 7 1 4 n 1 JA Age In Days Fig. 6. GYPSUM PLASTERS 13 cements and plasters, after hardening has progressed beyond a certain point, is in large part due to increased stiffness and brittleness and to the deficiencies of the test above noted. ■ cr ^ ft.' t; |2S00 ^ — — ~ ~ — — ' — — — / (0 3 ' 12000 ?. CO ? ^ ^ ■ ■*~ — r ~" ^ COMPRESSIVE STRENGTH OF CEMENT-pLASTER AND MORTAR D 1000 _ _ 1 1 1 1 1 1 1 1 1 M _ Compressive Strength The compressive strength of plasters is dependent upon practically the same factors as the tensile strength, but the experimental determina- tion of compressive strength is fraught with so great difficulties, the results influenced to so large an .extent by the personal factor in making and testing specimens, that sources of satis- factory experimental data are almost en- tirely lacking. Fig. 7 presents the results of a fimited series of compression tests of cement plas- ters and 1 : 2 mortars made upon prisms 2 by 2 inches in section and 4 inches high, by the Wyoming Geo- logical Survey.* The curves are averages of four tests upon each of two brands of hard wall plaster. A maxi- mum average neat ..strength of 2565 pounds per square inch and a maximum average mortar strength of 1595 pounds per square inch are shown at the age of one month. Retrogression between the ages of one month and six months, however, is only sfight and one of the two brands averaged even showed a considerable gain in strength during this period. The very slight amount of retrogression in compressive strength, as compared with that observed above in the discussion of ten- sile strength, appears to be in a measure corroborative of the explanation of the causes of retrogression in tensile strength above advanced. There is no reason why the effect of moisture absorbed should not be operative in the case of compressive tests as well as in tensile tests, but increased stiffness and brittleness would not affect compressive strength, because the bearing surfaces of compressive specimens can easily be made to receive a practically uniform load distribution. * Geological Survey of Michigan, Vol. 9, p. 176. 10 12 U IG IS 20 22 24 26 Age In ■Weeks Fig. 7. 14 MATERIALS OF CONSTEUCTION The effect of the addition of retarders upon the compressive strength of cement plasters is shown by Fig. 8, which has been plotted from the published tests of Professors Slosson and Moudy of the Wyommg Agn- cultural and Mechanical College.* The tests were made upon 2-inch cubes, each of the curves of Fig. 8 repre- senting the results obtained upon one brand of plaster. No data have been found to establish the facts as to the effect of retarders upon ten- sile strength. From the analogy to the case of the relative injurious effect of or- PercentageKetarder gauic matter Upon the YiQ, 8. compressive and the tensile strength of Portland cement, it may be stated, however, that it is improbable that the injurious effect of retarders upon tensile strength of plasters is as marked as it is in the case of compressive strength. — 1 — 1 _ — — — ~ — — ' fl »,<-», ~ ' ~ ■s '" o — ~ ~ __ ui ■^ rt a 1 ~ ~~ '" ~ r~ .^ B rnr (J ~ ~~ *~ -, - _ ~" — ^IbOU — ~ ~ C3 ^ — — ~ ~ s — ■^ -B P' r/ m ~ " n L ^ C3 _ — ~ ~ "' ^ •a 1000 ^ _ ~ ~ ~ "" £ _ ~^ EFF.E.CTOF ADDITION OF RETARDERS UPON COMPRESSIVE STRENGTH OF CEMENT PLASTERS o 1 1 M 1 M M 1 1 bOU n ,_ n 1 ■i Adhesive Strength The adhesion of plaster to various surfaces like that of stone, brick or plaster is a property of this class of building material which should be considered of greatest importance. Very little experimental data bear- ing upon the subject are available, however. The results of a limited series of tests made by Dr. Grimsley t are summarized by curves I and II, Fig. 9. In these tests plaster half-bri- quettes which had been broken at the center and whose fractured surface was fairly smooth were placed in molds and the remaining half filled with new plaster after the surface of the hardened plaster had been thoroughly wetted. The curves average the results from three to nine tests of one brand of cement plaster. It was observed in connection with these tests that, when the surface of the old plaster was not wet, there was scarcely any trace of adhesion. * Tenth Annual Report, Wyoming Agricultural and Mechanical College. t Geological Survey of Michigan, Vol. 9, p. 179. GYPSUM PLASTERS 15 The result of a similar series of tests made by Professor Marston are summarized by curves III and IV, Fig. 9. The test specimens were made "by taking pieces of No. 2 paving-brick from Des Moines and grinding them on the emery wheel so as to make approximately 1-inoh cubes. Each cube had one face carefully trued to give a cross-section exactly 1 inch square. These pieces of paving-brick were placed in the cement briquette molds with this .true surface exactly at the middle of the mold. The plaster or stucco was placed to fill the other half of the mold, while the half in which the piece of brick was placed was filled with neat Port- land cement mortar." No statement is made as to whether the sur- face of the paving brick was wet before placing the mortar. The curve III repre- senting the adhesion of plaster of Paris to paving brick is the average of three tests on three different brands, while the curve IV representing the adhesion of cement plaster to paving brick is the average of eight tests on eight differ- ent brands. Tests of a similar series of specimens which were stored in water after the first twenty-four hours showed that the excess of water during the period of hardening reduced the adhesive strength about one-third. 10, Relative Applicability of Various Gypsum Plasters to Structural Uses. Plaster of Paris, because of the extremely rapid set which espe^ cially fits it for various special uses as a casting plaster, etc., finds very little application as a material of engineering construction. Almost its only structural use is in the form of molded ornaments of " stucco " which serve as architectural adornment of buildings. Stucco is less clear white than pure plaster of Paris, and is less finely divided. Plaster of Paris is not adapted for use as either a wall plaster or a mortar for masonry 300 — — — — — — — — — — — — — — — ^ " 2S0 V ^ '' ^ ^ ^' s A r5 ir T? / ^ / '■^ p. / - m / n — — .3 ^ sW r 3150 I ^ / > S / ^ r / ^ pi Later of Pa>i8 to Pa viig Br ich TO /' / r^ Ce m^nt PI iBter to Pa vmff Jri ck -lioo / A i / y'. f, ■< JA / 50 / ADHESION OF PLASTER OF PARJ6 AND CEMENT PLASTER TO GROUND SURFACE OF PA*VING BRrCK AND TO FRACTURED SURFACE OF PLASTER Age In Weeks Fig. 9. 10 U 12 13 16 MATEEIALS OF CONSTRUCTION construction unless additions be made to retard its set and make it more workable. If this is done, however, it is no longer called plaster ot Paris, but becomes a cement or hard wall plaster. Cement or hard wall plasters find their principal applications as wall plasters with which a certain amount of hair, wood fiber, or asbestos fiber, together with hydrated lime or clay, has been mixed at the place of manufacture. Gypsum wall plasters possess a number of advantages over ordinary lime plasters, but also suffer by comparison with lime plasters in certain other respects. Among the advantages of gypsum plasters, the most important one is the fact that the material comes upon the work ready to be simply mixed with sand and water and immediately applied to the lath, whereas the lime requires careful slaking and should be allowed to season for at least a day, and preferably for a much longer period, before being made up in a mortar and applied to the walls. Secondly, gypsum plasters set more rapidly and dry out in a much shorter time than do lime plasters, thus often avoiding a delay in the completion of the interior finish of buildings. On the other hand, no gypsum product makes as plastic and smooth working a plaster as does the best Ume, provided the latter is properly slaked, seasoned, and mixed. Lime plasters excel in sand-carrying capacity, making it possible to use three or four parts of sand to one of lime for the first or " scratch " coat on walls, and two parts of sand to one of lime for the second or " brown " coat, whereas it is inadvisable with most gypsum plasters to use mortar mixtures leaner than 1 : 2 for either scratch or brown coat on account of their poor spreading qualities. The finish or " skim " coat is a neat or nearly neat paste in either case. This condition tends to make lime plaster cheaper, other things being equal. It is strongly claimed by those interested in the exploitation of gypsum plasters that this class of wall finish is less injured by moisture; less subject to volumetric change and consequent cracking due to variation in either moisture content or temperature; that it does not affect coloring matter incorporated in the mix as do some limes; that it excels Ume plaster as a fire retardant especially when mixed with asbestos fiber; and that it is a superior material from the standpoint of cleanliness and sanitation on account of the smooth polished surface that can be given to it. All of these latter claims are denied by those interested in the market- ing of lime and, in addition, the counter claims are made that gypsum plasters cause wood lath to swell and buckle; that they corrode metal lath; that they are more resonant and less sound-proof; and that they produce an undesirably hard and brittle wall surface. GYPSUM PLASTERS 17 It will be apparent that these conflicting claims are inspired by the intense trade rivalry that exists between the gypsum interests and the lime interests. For the most part the claims of neither are supported by dependable data, and it is an indubitable fact that wall surfaces may be and often have been built of each material which are entirely satisfactory in every way. The structural qualities of each material are largely dependent upon the way it is used on the work, and the choice of one or the other should depend upon whether the conditions of the work and the experience of the builder are the most favorable to the production of the best results that can be obtained with one material or the other. Among the appUcations of hard wall plaster other than as a plaster, the following are worthy of brief notice: Mixed with finely ground cinders and water to form a fluid inix, hard wall plaster is poured in forms for floor panels of fire proof buildings. Mixed with sawdust it is molded into blocks which may be nailed in place as a wall finish, and, without the sawdust it may be molded into solid or hollow building blocks and tiles for the construction of partition walls and floors. A recent special application of cement plaster is in the construction of so-called " plaster-board," wherein the plaster is laminated with thin layers of card-board or wood in sheets which are ready to be nailed to the studding of partition walls. The wall surface thus provided is sub- sequently plastered, as .a rule, but may be finished in panels with no additional plaster by simply covering the joints with wood strips. Flooring plaster is, as the name implies, intended primarily as a sur- face finish for floors. It must be protected from moisture while setting, and must dry evenly to avoid the formation of cracks. After standing about twelve hours it is pounded with wooden mallets and smoothed with trowels. Flooring plasters have been used considerably in Germany, but have never been made or used in the United States. Hard-finish plasters find their principal application as wall plasters and as floor surface, one of the most common applications being as an imitation of tiling or marble for floors and wainscoting in hospitals, lavatories, etc. Keene's cement * is, perhaps, the best known variety of hard-finish plaster. Its set is extremely slow, and it gains in strength very gradually, but ultimately attains a great degree of hardness and a strength exceeding that of any ordinary gypsum plaster. It may be regaged with water after having become partially set, and will then take its set and harden just as satisfactorily, apparently, as if the process of hardening had not been interrupted. * The original patent on " Keene's cement '' has long since expired and many plasters made both in Europe and in the United States are marketed under this gen- eric name. 18 MATERIALS OF CONSTRUCTION 11. Production, Value, and Uses of Gypsum Products. Statistics of the gypsum and plaster industry are presented in the table below. These data have been abstracted from the reports of the U. S. Geological Survey, published in the annual volume entitled "Mineral Resources of the United States." STATISTICS OF THE GYPSUM AND PLASTER INDUSTRY Year No. of mills reported. . . . 1910 (82) 1911 (78) 1912 (76) 1913 (67) Quantity, Short Tons. Value per Ton. Quantity, Short Tons. Value per Ton. Quantity, Short Tons. Value per Ton. Quantity, Short Tons. Value per Ton. Sold without calcining: For Portland cement. . Aa land plaster For other purposes. . . . 334,815 53,815 33,199 $1.56 2.05 1.10 327,953 52,880 6,647 $1.48 1.85 1.13 382,952 53,065 5,591 SI. 33 2.02 1.26 408,221 54,815 100 SI. 47 1.75 2.00 Total 421,829 1.59 7.00 3.78 1.83 2,65 3.70 387,480 1.52 441,608 1.41 463,136 1.51 Sold calcined: For dental plaster For plaster of Paris, and wall plaster, in- cluding K e e n e ' 3 cement To glass factories For Portland cement 115 1,483,046 15,943 84,565 413 1,523,263 33,472 41,270 6.32 3.73 2.40 2.70 3.67 3,190 1,678,417 24,159 "25,908 4.88 3.46 2.18 2.55 861 1,680,157 10,942 81,889 4.84 3.49 1.99 2.36 Total 1,583,669 1,598,418 1,731,674 3.43 1,773,849 3 43 Grand total 2,005,498 3.25 1,985,898 3.25 2,173,282 3.02 2,236,985 3.03 CHAPTER II QUICKLIME GENERAL 12. Definition and Classification.* Quicklime is the name applied to the common or commercial form of calcium oxide (CaO), obtained by the calcination below the sintering point of a stone in which the pre- dominating constituent is calcium carbonate (CaCOs), often replaced, however, to a greater or less degree by magnesium carbonate (MgCO.s)) this product being one which will slake on the addition of water. Hydrated lime is absolutely the same material as quicklime except that it does not possess the power of slaking, since it has been chemically satisfied with water during manufacture. Quicklime is divided into two general grades: (a) Selected. A well-burned lime, picked free from ashes, core, clinker, or other foreign material, containing not less than 90 per cent of calcium and magnesium oxides and not more than 3 per cent of carbon dioxide. (5) Run-of-kiln. A well-burned lime without selection, con- taining not less than 85 per cent of calcium and magnesium oxides and not more than 5 per cent of carbon dioxide. According to the physical form of the material, quicklime is mar- keted as: (a) Liump Lime. The size in which it comes from the kiln. (6) Pulverized Lime. Lump lime materially reduced in size by mechanical means. According to relative content of calcium oxide and magnesium oxide, quicklimes are divided into four types: (a) High^caldum,. Quicklime containing 90 per cent or over of calcium oxide. (Sometimes termed " rich," " fat," or " caustic " lime.) * The classification is that proposed June, 1914, by Committee C-7 of the Am. See. Test. Materials. 19 20 MATERIALS OF CONSTRUCTION (b) Calcium. Quicklime containing not less than 85 per cent and not more than 90 per cent of calcium oxide. (c) Magnesian. QuickUme containing between 10 and 25 per cent of magnesium oxide. id) Dolomitic. Quicklime containing over 25 per cent of magnesium oxide. All of the above limes are sometimes classed as " white limes " in contradistinction to the "gray limes": hydraulic Ume, derived from limestones containing from 5 to 20 per cent of clayey and sandy material (silica and alumina), and natural cements, made from limestones contain- ing from 15 to 35 per cent of clayey and sandy material. 13. The Place of Lime* Among Cementing Materials. Cementing materials used structurally may be divided into two general classes, as non-hydraulic and hydraulic. Lime is the most common non-hydrauUc cementing material, and the only one of any commercial importance except the gypsum plasters discussed in Chapter I. The hydraulic cements include hydraulic limes, grappier cements, natural cements, Portland cement, and puzzolanas. Each of the hydraulic cements will be treated in the chapters which follow. 14. Application in the Arts and Industries. Although the appli- cations of lime in the various arts and industries are innumerable, the most important ones are briefly mentioned in the following list: Natural cement manufacture, added prior to final grinding in hydrated form, beneficial to strength and smooth working qualities. High-calcium lime is necessary. Sand-lime brick manufacture, mixed in hydrated form with sand, and subsequently molded into bricks which are hardened by high pressure steam. High calcium lime is necessary. Gypsum plaster manufacture, added subsequent to calcination in hydrated form, beneficial from the standpoint of working qualities. Either high calcium Ume or magnesian lime satisfactory. Slag cement and puzzolan cement, mixed in hydrated form with granulated blast- furnace slag or puzzolan and subsequently ground to form an inferior grade of hydraulic cement. High calcium lime necessary. Iron and steel manufacture, used as raw limestone or lump lime as a flux in blast furnaces and steel furnaces to render fusible the mechanical impurities and furnish a liquid slag to carry off both mechanical and chemical impurities. Either high-cal- cium or magnesian limestone satisfactory. Glass manufacture, used in the form of ground quicklime m combination with sand and sodium carbonate as the base of all glass mixtures. Magnesian lime usually preferred. Paper manufacture, in soda process quicklime is used to recover caustic soda from liquor in which wood has been cooked to make pulp; in sulphite process either raw * The term " lime " when used herein is intended to include both quicklime and hydrated lime. QUICKLIME 21 limestone or slaked lime in form of milk of lime is used with sulphur dioxide to form bisulphites which are held in solution by excess SO2 thus constituting a liquor in which wood is cooked to make pulp. Magnesian limes preferable. Sugar manufacture, used in the form of the purest calcium oxide obtainable to remove impurities in juice of beet or cane, the Hme being subsequently liberated from sugar by introduction of CO2. High calcium limestone is burned by sugar manu- facturer who uses both CaO and COj. Soda ash manufacture, used as ground lime to recover ammonia used in the proc- ess. High calcium lime preferred. Bleaching powder manufacture, used as ground lime as absorbent for chlorine gas. Purest high calcium lime required. Illuminating gas manufacture, gas distilled from coal is freed from CO2, H2SO4, etc., by passing through moist slaked lime. Ammonia is also freed from by-prod- ucts by action of lime which takes up acids, heavy oils, etc. High calcium lime required. Woodraltohol, acetic acid, and acetone derivation, the pyroligneous acid first derived by distillation of wood is redistilled in excess of lime, producing " gray acetate of Ume " and freeing the wood alcohol. The latter is purified by redistillation with lime. Acetone is obtained from the acetate by dry distillation, and acetic acid by dis- tillation after treating with sulphuric acid. High calcium lime desirable. Calcium carbide manufacture, a mixture of lime and coke is heated to an extreme degree in the electric furnace. The product is the source of acetylene. Purest high calcium Ume required. Glycerine derivation, the treatment of fats with quicklime results in the libera^ tion of glycerine and the formation of lime-soaps by combination of the lime with the organic acids. Either high calpium or magnesian lime satisfactory. Heavy lubricants, Ume-soaps derived as above noted, mixed with mineral oils, form lubricants satisfactory for use with heavy machinery or at high temperature. Waterproofing compounds, lime-soaps derived as above noted, or as a by-product of soap manufacture, form the base of most so-called active waterproofing compounds, used especially in cement mortars and concretes to reduce permeabiUty. Cold-water paints, mixtures of hydrated lime, hme carbonate, or air-slaked lime with pigments and casein form the bulk of that class of paints which are not pre- pared for use by mixing with oils. Magnesian limes preferred. Tanning industries, soaking of hides in a Ume-water bath loosens the hair so that it can be scraped off. High calcium limes preferred. Water purificatiwi, " temporary hardness " of water, caused by presence of CaCOa held in solution by excess CO2, is removed by treatment with hydrated lime, which takes up CO2, forming additional CaCOa which, together with that originally present, precipitates out. High calcium lime preferred. Agricultural uses, used in connection with some organic fertilizer as a soil amend- ment to convert nitrogen of the air into nitrates available as plant food. A definite ratio of magnesium oxide to calcium oxide is desirable for a given soil. A liTn»-water is also used for spraying purposes as an insecticide. 15. Structural Uses. The applications of lime as a structural mate- rial are very few in number compared to its applications in various manu- facturing and chemical processes. The total amount used structurally, however, far exceeds the amount used for any single industrial purpose 22 MATERIALS OF CONSTRUCTION and, indeed, amounts to two-fifths of the total amount of lime marketed for all purposes. By far the largest portion of the lime used as a building material is used in combination with sand as a mortar for use in laying brick and stone masonry. Another large portion is used as a wall plaster, and the greater part of the balance is used in gaging cement mortars either for the purpose of increasing their plasticity, thus making them easier to work, or to reduce their permeability. A small amount of lime is used struc- turally as a whitewash. In the present chapter the use of lime in mortars and plasters will be alone considered. Its use in connection with cements will be dis- cussed in the chapters devoted to the properties and uses of natural and Portland cement. MANUFACTURE OF LIME 16. Limestone Rocks. An ideal, pure limestone consists entirely of calcium carbonate (CaCOs) which, at a temperature of 898° C. or over, becomes dissociated, the carbon dioxide (CO2) being driven off as a gas, leaving behind a white solid, calcium oxide or quicklime (CaO). As the original calcium carbonate consisted of 56.1 parts by weight of CaO to 44.0 parts of CO2, the theoretical proportion of quicklime obtainable by the calcination of limestone will be 56.0 per cent by weight. In practice, the proportion of quicklime obtainable will always fall below this theoretical limit on account of the inevitable presence of impurities in the limestone and the imperfections of the process of calcination. Limestones encountered in practice depart more or less from this theoretical composition. A part of the lime is almost invariably replaced by a certain percentage of magnesia (MgO), making the stone to a greater or less extent a magnesian limestone. In addition to magnesia, silica, iron oxide, and alumina are usually present to an appreciable extent, and, to a slight extent, sulphur and alkalies. 17. Theory of Calcination. The burning or calcination of lime is a very simple chemical process known, in its essentials at least, since very ancient times. Three objects are accomplished by burning the limestone: (1) The water in the stone is evaporated. (2) The limestone is heated to the requisite temperature for chemical dissociation. (3) The carbon dioxide is driven off as a gas, leaving the oxides of calcium and niagnesium. The evaporation of any water present in the stone means that a certain portion of the heat supplied during calcination does not directly QUICKLIME 23 assist in the dissociation of the carbonates. This heat does not mean a thermal loss, however, because the presence of the water, and the steam generated from it, facilitates the dissociation process, as will be herein- after shown. The temperature of dissociation of pure calcium carbonate at a pres- sure of one atmosphere has been determined by Johnson * to be 8i98° C, and the corresponding temperature for magnesium carbonate is generally stated to be between 550° and 750° C. If the magnesium occurs in the stone not as a simple carbonate, MgCOs, but in combination with the cal- cium carbonate as dolomite, CaCOs, MgCOs (as may possibly be the case), the temperature of dissociation cannot be stated precisely. The studies of Professor Bleininger and Mr. Emley t tend to show that the maximum calcination temperature for high-calcium lime low in impurities is not below 1300° C; that for high-calcium lime high in impurities the maximum temperature is from 1050° to 1200° C, and that for magnesian lime low in impurities the maximimi temperature is 900° to 1050° C. The minimum dissociation temperature of calcium carbonate was found to be remarkably constant for most stones in the neighborhood of 880° C, and the corresponding minimum dissociation temperature for mag- nesium carbonate was found to be about 750° C, although this point was found to be more variable than that for the calcium carbonate. • Theoretically, all limestones could be properly burned at a temperature of about 880° C, provided sufficient time be allowed. In practice, however, the maximum kiln output is always striven for, and, since the rate of heating is directly proportional to the temperature of the kiln, the maximimi kiln output is obtained by burning at the highest temper- ature practicable. The highest temperature practicable is not the highest temperature attainable, however, because the increased activity of the impurities with rising temperature set a very definite maximmn limit, the exceeding of which means serious injury to the quality of thejime produced. The impimties in limestones are all essentially acids which readily combine with the basic lime and magnesia. The silica forms silicates of lime and magnesia, the alumina forms aluminates, and the iron oxide forms ferrites. These siKcates, aluminates, and ferrites are all readily fusible compared with pure lime or magnesia, and the result of their formation with increasing temperature is a softening of the portion of the stone thus rendered fusible at the practical dissociation temperature, and its softening closes up the pores and envelopes the particles of Ume * Journal Amer. Chem. Soc, 1910, p. 938. t The Burning Temperature of Limestone, Trans. Nat. Lime Manf's Assoc, Feb., 1911, pp. 68-78. 24 MATERIALS OF CONSTEUCTION with a slag-like coating which causes the quickUme to slake with dif- ficulty, and which therefore greatly injures the commercial value of the kiln product. The softening of the mass also results in compacting the material, thereby further injuring its quaUties by mechanically rendering it less porous. This behavior on the part of the impurities explains the existence of maximum practical temperature hmits in burning impure lunestones, exceeding which causes marked loss of power of the quicklime to com- bine with water in slaking, whereas it is practicaUy impossible to over- burn pure limestone. The physical character of the Umestone has an important bearing upon the burning temperature, quite aside from the question of chemical composition. A naturally coarse, porous stone is much more rapidly acted upon by heat than a dense, finely crystalline stone, and in conse- quence may be burned more rapidly and at a lower temperature. Small pieces of stone may also theoretically be burned more readily than large stones. In practice, however, large sizes seem to be preferred by the practical lime manufacturer, the common practice being to use what are called " one-ma;n-stone," measuring about 8 to 10 inches in diameter. One other consideration affecting calcination temperatures which cannot be overlooked is the fact that the temperature required to effect calcination at atmospheric pressure need not be attained under actual working conditions, for the simple reason that provision is made for suf- ficient draft in the kiln to prevent the internal gas pressure ever becoming so much as one atmosphere. The necessity for provision for carrying off the carbon-dioxide gas evolved during calcination has just been indicated. The necessary cal- cination temperature being directly dependent upon the pressure, it would be very uneconomical of fuel to allow the gas to accumulate. It is possible, in fact, when the pressure is allowed to increase without a proportional increase in temperature, to actually reverse the process of dissociation and cause the carbon-dioxide to recombine with Ume and magnesia and reform their carbonate. This phenomenon is known technically as " recarbonating." (Recarbonating is especially apt to occur in mixed-feed kilns and also in separate-feed kilns if fuel is allowed to fall over the bridge wall into the cooler.) Various expedients are adopted in practice to maintain the gas pressure at as low a point as possible, the most common one being the introduction of jets of steam or water into the hottest portion of the kiln. Upon contact with the glowing carbon of the fuel, the water is robbed of its oxygen, a certain amount of heat being required to effect the chemical decomposition. The hydro- gen travels toward the stack end of the kiln until it encounters a tempera- QITICELIME 25 ture at which recombination with oxygen is possible. The reformation of water vapor is attended by the evolution of the same amount of heat as was utiUzed in effecting dissociation of the original water vapor, result- ing in increasing the draught and decreasing the gas pressure. This practice is attended by another thermal advantage in that it tends to equalize the heat distribution throughout the ipaterial in the kiln. Heat is taken from the overheated portion of the kiln and restored where most needed. Increased kiln eflBciency is therefore attained. 18. Practice of Cal- cination. The types of kilns employed in lime / ^fj^'[^,'fM|it burning may be briefly / &i'/i1//i//;w )M descnbed as follows: Intermittent Kilns. An early form of intermittent kiln consisted simply of a dome-like structure crudely constructed by using the larger blocks of the stone to be burned. The balance of the limestone was piled on top and a wood fire started underneath. A bright red heat having been obtained throughout the mass of stone, this temperature was main- tained for a period of three or foiir days, when, the mass having become soft, the fire was allowed to go out and, after coding, the lime was removed, the structure being demolished in so doing. Permanent intermittent kilns, often called " pot kilns " (Fig. 10), built of stone with a firebrick hning and provided with a grate upon which the fuel is placed, grad- ually replaced the original type. These kUns are provided with an arched opening at the bottom through which fuel is introduced and the burnt lime removed. The fuel consumption of even this type of intermittent kiln is very excessive, due to the neces- sitj' of heating the entire mass of masonry, as well as the contents, to the required tem- perature of calcination each time the kiln is charged. In addition, the product is never uniformly burned. Continuous Kilns, (a) The Vertical Kiln toith Mixei Feed. In this kiln (Fig. 11), theiuel (bituminous coal) and the limestone are charged in alternate layers, the lime being removed at the bottom while fresh fuel and limestone are charged in at the top. The fuel consumption usually amounts to from 15 to 25 per cent of the Fig. 10. — Pot Kiln for Lime Burning. 26 MATERIALS OF CONSTRUCTION weight of lime produced. Few mixed-feed kilns are now employed in the American ^'"^l^^Thfvertical Kiln with Separate Feed. The kilns of this type (f J . 12)/|;i!^ designed that the fuel and limestone do not come in "o-^t^"*; *\V"f.,^\^V Th^ in separate fireplaces either set in the wall of the kiln or ""t^'^^^*^^, „!' '^bustion limestone therefore comes in contact only with the hot gaseous products of «™««°°- The relative advantages of types (a) and (b), which are ^l^J"^* ^^i^^J^^'^Laner in the United States, may be summarized as follows; The --d-fe^^Mx. are d.^per more economical of fuel, and somewhat more rapid in operation. On the other hand, the separate- feed kilns yield a lime somewhat less discolored by contact with the fuel, the Ume is free from the fuel' ash, which is not easily separated out, and the danger of some part of the lime being imper- fectly calcined, owing to a coating of fuel clinker on the limaps, is obviated. From 75 to 80 per cent of the output of the mixed- feed kiln is marketable as well-burned, clean, white lime as compared with 90 per cent obtainable from separate-feed kilns. The present tendency among American lime maniifacturers seems to be in the direction of building large separate- feed kilns having a capac- ity of 60 to 60 tons per day of twenty-four hours, or 10 or possibly 15 tons Fig. 11. — Aalborg Kiln for Lime Burning. instead of the usual type of kiln with a capacity of 8 per day. I (c) Ring or Chamber Kilns. The ring kiln has been used quite extensively in Germany and other parts of Europe, but has never come into favor in the United States. The commonest type of ring kiln, the Hoffman kiln (Fig. 13), consists of a series of chambers arranged around a central stack. Each chamber is connected by flues with the stack and with each of the two adjoining chambers, and each flue is provided with a damper by means of which the passage may be closed. The chambers may be charged with limestone and fuel in the shape of fine coal, and one chamber is fired. The dampers are now so set that the hot gases of the burning chamber must traVerse all the other chambers before passing to the stack, so utilizing the heat of the gases in QUICKLIME 27 preheating the contents of the chambers not yet fired. When the calcination of the limestone in the first chamber is complete, the second chamber is fired and the first chamber is temporarily cut out of the circuit until it has been discharged and recharged, when it becomes the last chamber in the circuit. The operation is continued in this manner indefinitely. The Hoffman kiln is an improvement both in econ- omy and in quality of prod- uct upon the mixed-feed vertical kiln, but its su- periority over the most modern American separate- feed kiln is doubtful. In common German practice the fuel consumption amounts to from 20 to 22 per cent of the Ume pro- duced, while with careful management the percentage may be kept still lower. (d) Rotary Kilns. Ro- tary kilns, resembUng in construction the rotary cement kilns described in the chapter on Portland cement (page 123), have been applied to the calcina- tion of lime to a limited extent. They are subject to the disadvantage of re- quiring that the stone be finely crushed prior to the calcination, and the product is consequently so finely di- vided that it is not market- able as lump lime, but can be sold only after grinding or hydrating, as either ground or hydrated lime. Fig. 12. — Keystone Separate-feed Lime Kiln. The fuels used in lime burning are wood, bituminous coal, and pro- ducer gas. Wood fuel possesses a distinct advantage over coal because of its greater moisture content, which results in longer flames and con- sequent better heat distribution throughout the mass of stone. Coal would never be used as a kiln fuel were it not for the impossibility of pro- curing wood fuel in many districts. The short-flaming hot coal fire results in overheating of the portion of the kiln near the firebox, while the material shortly above the fire zone receives far too Uttle heat. The 28 MATEEIALS OF CONSTRXTCTION use of producer gas as a fuel for lime burning is a late development both in Europe and the United States, its installation having thus far been largely confined to the larger and more modern plants. The necessary installation consists simply in the replacement of the firebox of the separate feed kiln by gas producers and burners. The gas and air ports are so placed in the walls of the kiln that the combustion takes place within the kiln chamber itself. The relative advantages of producer gas and coal and wood as a fuel are not yet perfectly established. It is to be expected, however, that while the fuel economy may be slight, there should be some increase in capac- ity and a decrease in labor costs. The defects of coal as a kiln fuel may be in a measure overcome by the adoption of some expedient which makes possible longer flames or better heat distribution. A forced draught may be obtained by blowing steam in through the grates. The moisture thus introduced is beneficial in that it cools the over- heated portion of the charge, mechanically trans- fers heat to the kiln, and, for other reasons indica- ted above, lowers the cal- cination temperature and equalizes the distribution of heat. The action of the kiln and its thermal eflftciency are further improved, if an exhaust fan is also provided to draw off the burned gases and prevent the possibility of recarbonation. A practical difficulty which attends the use of steam in lime burning, and which has limited the extent of its application, is the fact that the amount of moisture used must be care- fully regulated, a thing not easy to accomplish. Another method of combining induced and forced draught is that known as the " Eldred process," which consists essentially in exhausting Fig. 13.— Hoffmann Ring Kiln. QUICKLIME 29 the kiln by a fan and forcing a portion of the burned gases thus derived back under the firebox grates, together with a supply of fresh air. The fire is cooled by dilution of the air with carbon dioxide. In addition, carbon dioxide transfers heat to the stock in the kiln mechanically, and probably also does so chemically, in a measure, a part of the dioxide being reduced in the firebox to the monoxide, which latter gas burns to the dioxide again in the kiln, thus transferring heat in a manner similar in some respects to the action of steam'. Carbon dioxide does not lower the calcination temperature as does steam, however, and the merits of the Eldred process in general still constitute a much-debated question among lime manufacturers. A lime kiln in operation always contains three classes of material: (a) Stone undergoing preliminary heating through the agency of the escaping hot products of combustion: (6) Stone undergoing dissocia- tion through the agency of the direct heat of the fuel: (c) Calcined lime which accumulates in the lower portion of the kiln and is withdrawn in part from time to time. The total amount of lime present in the kiln cannot be drawn at one time, because, aside from the desirability of letting it cool in the kiln itself, enough must remain at all times to fill the " cooler" (that portion of the kiln below the level of the fuel grate), thus preventing unburned stone from sinking below the level of the zone in which it is subjected to the action of the flames of the fi]*s. The operation of drawing consists simply in opening a draw-door provided in the lower part of the cooler, poking the lime loose with a bar, and allowing it to fall into a barrow or car placed underneath. This operation is repeated at intervals of from one to eight hours (usually from four to six hours) . 19. Treatment Subsequent to Calcination. Lime drawn cold from the kiln is immediately ready to be marketed as lump lime. When the kiln is not provided with a cooler, however, or when drawn from the cooler while still very hot, it is necessary to spread it out on a cooling floor or leave it standing in fireproof containers for a few hours before taking it to storage, packing house, or cars. Underburned and over- burned material is easily recognized by its appearance, and is sorted out by hand while drawing it into the barrows, or while it lies on the cooling floor. Many consumers of lump lime formerly required that it be screened to remove the fine material, their belief being that the fine- ness of the lime indicated that it had become air-slaked. This practice was wholly unjustifiable, and based upon a fallacy now commonly recognized as such. An increasing proportion of the lime intended for use in the building trades, especially as hydrated lime, as well as for agricultural purposes 30 MATEEIALS OF CONSTRUCTION and many applications in the ^rts and industries, is now ground before being marketed. Ground lime is prepared by running the cooled lime through a crusher and then a pulverizing mill, which reduces it to a size sufficient to pass a sieve of 80 meshes per lineal inch. The product is then barreled or sent to the hydrator. Ground Ume used structurally as quicklime is seldom crushed to a size less than about i inch. " Hydrated lime " has come to be regarded as a special product, and its preparation will be separately discussed in the following chapter. PROPERTIES AND USES OF QUICKLIMES 20. Classification of Lunes by Uses. Required Qualifications of Each Class.* Limes may be conveniently classified according to the purposes for which they are used, as (1) agricultural limes, (2) chemical hmes, (3) building limes, and (4) finishing Umes. For agricultural limes only the chemical composition is of importance; for limes used in the chemical industries the chemical composition and the rate of hydration are important; building limes must be satisfactory as regards sand-carrying capacity, yield of hme-paste per unit weight of lime, crushing strength, and tensile strength; and finishing limes must be satisfactory with respect to rate of hydration, plasticity, sand-carrying capacity, color, yield, waste (i.e., that percentage of a lime made up into a putty which caimot be washed through a 20-mesh sieve by a stream of water having a moderate pressure), hardness, time of setting, and shrinkage. 21. Chemical Composition. The approximate chemical composition of limes of various classes has been indicated in Art. 12 above, wherein limes were classified and graded according to their content of calcium oxide, magnesium oxide, and carbon dioxide. The following table, which has been abstracted from one published by Mr. Emley of the U. S. Bureau of Standards,! illustrates the range of composition found for limes of various classes upon analysis of thirty-seven samples of quick- limes and hydrates coming from thirty-two different States in the United States: * This classification is that proposed by W. E. Emley of the Pittsburgh Branch of the U. S. Bureau of Standards. See Trans. Nat. Lime Manfr's. Assoc, Jan., 1913, pp. 77-100. * Trans. Nat. Lime Manfr's. Assoc, Jan., 1913, pp. 77-100. QUICKLIME 31 ANALYSIS OF THIRTY-SEVEN QUICKLIMES AND HYDRATES Summary Class of Lime. High Calcium Quicklimea. Calcium and Magnesian Quicklimes. Dolomitic Quicklimes. Compo- nent. Min. % Max. % Ave. of (10) % Min. % Max. % Ave. of (6) % Min. % Max. % Ave. of (2) % SiOj. . . . 0.33 2.20 0.81 0.66 9.001 3.12 0.14 1.59 0.87 FesOa. . . 0.08 0.43 0.23 0.17 0.59 0.41 0.19 0.39 0.29 AUOs... 0.02 0.42 0.22 0.18 2.57' 0.93 0.14 0.49 0.32 CaO.... 91.37 98.08 94.98 78.59 84.81 81.42 55.89 64.45 60.13 MgO.... 0.17 4.55 1.39 1.03 16.83 9.26 31.61 40.62 36.12 HjO.... . 0.36 3.45 1.66 0.63 12.422 4.18 0.55 1.56 1.06 COs 0.20 1.84 0.83 0.24 1.94 0.18 0.35 3.01 1.68 Class of Lime. High Calcium Hydrates. Dolomitic Hydrates. Compo- nent. Min. % Max. % Ave. of (U) % Min. % Max. % Ave. of (8) % SiO,. . . , FfeOs... AW,... CaO.... MgO.... H,0.... CO2 0.42 0.10 0.23 66.71 0.28 18.94 0.57 5.031 1.01 1.05 73.45 4.11 24.98 6.25' 1.60 0.33 0.55 70.90 1.68 22.45 2.52 0.20 0.15 0.20 40.06 23.10 14.24 0.79 3.781 0.81 3.501 53.20 34.02 21.37 9.81» 1.82 0.35 1.26 47.06 29.92 16.77 2.49 ' Excessively high in acid impurities. : incipient air-slaking shows. » Excessively carbonated. 22. Hydration or Slaking. Rate of Hydration. Quicklime intended for use in mortars for masonry construction or as a wall plaster must first be prepared for mixing with water to form a lime paste by being slaked. The hydration or slaking of quicklime consists simply in the addition of sufficient water for the formation of calcium hydroxide, the operation being represented by the formula: :aO+H20 = Ca(OH)2, 75.7 +24.3= 100 (parts by weight). If the quicklime were absolutely free from impurities the amount of water required for complete slaking would equal 32.1 per cent by weight of the quicklime, but the fact that the quicklime is always impure to a greater or lesser degree makes the amount of water actually required 32 MATERIALS OF CONSTEUCTION less than the above percentage. The formation of lime hydrate . is attended by the evolution of considerable heat and an expansion to about 2k or 3 times its former volume. Magnesian quicklimes, and particularly dolomitic quicklimes, slake more slowly than high calcium limes, and the slaking is attended with the evolution of much less heat and far less expan- sion. The common practice in slaking quicklime for building mortara is to add much more water than is actually required, a practice which with even ordinary care should insure thorough slaking, though it may possibly be at the expense of some loss in mortar strength. The fact that quicklime used in the building trades is so commonly slaked care- lessly by unskilled labor is largely responsible for the demand of late years for gypsum plasters and ready slaked lime or " hydrated-lime." Lime intended for use in a mortar is usually slaked in a mortar mix- ing box, the mixture being stirred as water is added until a thin paste or " putty " has been formed. The putty is thereupon covered with sand to protect it from the action of the air, and is kept so covered until all has been gradually used by mixing with sand to make mortar. Lime paste or putty designed for use as a plaster was once commonly placed in barrels, covered with an excess of water, and allowed to season for several weeks or even months before being used. The present prac- tice is to greatly shorten the period of seasoning, though the quality of the plaster from the standpoint of workability is greatly injured by the common practice of seasoning overnight or for a few hours only. The reaction involved in the hydration of quicklime may result in the production of either crystalline calcium hydroxide or colloidal * calcium hydroxide, the relative quantity of one or the other being depend- ent upon the time afforded for the reaction. Crystals of calcium hydroxide form and grow slowly, whereas the colloidal hydroxide forms with great rapidity. Consequently, the more rapid the reaction is made to be, the greater the proportion of colloidal hydroxide. The reaction may be most readily hastened by using warm water in slaking. A preponder- ance of colloidal hydroxide is eminently desirable from the standpoint of the mason who judges a mortar by its plasticity or spreading qualities, its yield, and its sand-carrying capacity. The hydration of high-calcium quicklimes is attended by great danger of " burning," due to too great a rise in temperature. The exact nature of this chemical phenomenon is not understood but the product " burned lime," appears to be chemically inert and is useless in a mortar or plaster. Burning may best be avoided by securing an intimate contact between * A colloid is a non-crystalline hydrate which possesses the physical character- istics of a gelatine, and which upon desiccation, may acquire a strong binding powei like glue. QUICKLIME 33. every particle of lime and the water. Great watchfulness and continuous stirring of the mixture is tterefore necessary. No danger of burning attends the slaking of most magnesian and. all dolomitic quicklimes. On the contrary, the danger in this case is that the quicklime may never be properly slaked before being used. All magnesian quicklimes slake very slowly and, if the temperature of cal- cination has been very high, hydration may scarcely be possible at all.. Dr. Campbell * concluded, as the outcome of a study of this question, that magnesium oxide will combine with water with reasonable rapidity only when it has been burned at a temperature below 1100° C. The phenomenon known as " air slaking" of quicklime is deserving of mention in this connection, although as will be sho^n directly, " air- slaked " lime is a very different thing from the ordinary slaked lime. Quicklime exposed to the air absorbs moisture and becomes slaked lime, the expansion accompanying hydration causing the lumps to fall into a more or less fine powder. Immediately, the slaked lime is attacked by the carbon dioxide of the air, the water is replaced by CO3, and the resulting product is simply powdered calcium carbonate, CaCOs. The term " air-slaked " will be seen to be a very misleading one, for the quick- lime has not simply been slaked by the moisture in the air, but has been ruined completely as a cementing material by the taking up carbon dioxide while in a loose powdered state. . The fact that quicklime does " air slake " when exposed to the air is not an unmitigated evil, however, because it renders possible the storage and even the shipment of ground lime without its being contained in tight bags or barrels. The quicklime at the surface becomes " air-slaked," but by so doing it immediately forms a film which protects the bulk of the material to which moisture and CO2 cannot gain access. The experimental determination of the rate of hydration of quick- limes is accomplished by observation of the increase in temperature accom- panying hydration. The method employed by Mr. Emley consisted in obssrving the temperature increase of water in a special bomb calorim- eter during the period of hydration of a sample of quicklime enclosed within the bomb with a measured quantity of water. The rate of hydra- tion is really the slope of the time-temperature curve, but is most con- veniently expressed by the time required to attain maximum temperature. The rates of hydration of various quicklimes were found by Mr. Emley to be dependent, first, upon the physical character of the material, finely divided or porous quickhmes being more quickly hydrating because of their greater accessibility to water; second, upon the chemical com- position of the quicklimes, high-calcium quicklimes being more quickly '* Journal Ind. and Ehg. Chemistry, Vol. 1, p. 665, 34 MATERIALS OF CONSTRUCTION hydrating than magnesian or dolomitic quicklimes, and pure limes of either class more quickly hydrating than impure ones, because the rate of hydration is largely dependent upon the proportion of calcium oxide — the active constituent— present; third, upon the temperature of burning of the quicklime, any underburned quicklime having little ability to hydrate, and overburned limes behaving similarly owing to the influence of impurities. The relative rapidity with which pure and impure high-calcium and dolomitic quickUmes hydrate is forcibly illustrated by the curves of Fig. 14. The slopes of the curves indicate the relative rates of hydra- tion, the maximum ordinates indicate the relative amounts of heat gen- erated in hydrating, and the corresponding abscissse indicate the time required for complete hydration. It will be noted that the rate of hydrations of the high- calcium quicklime is about twice that of the impure high-cal- cium quicklime, more than twice that of the pure dolomitic quick- Ume, and fully three times that of the im- pure dolomitic quick- lime. It is also ap- parent that the heat generated during hy- dration is much greater for the high-calcium and the pure quicklimes, that the dolomitic quickhme requires twice as long to become completely hydrated as does the high-calcium, and lastly, impure quicklimes require several times as long to hydrate as do the corresponding pure quicklimes. All of the quicklimes of Fig. 14 were burned at the same temperature, 1050° C. That the temperature of burning is a factor even more influen- tial than the composition is shown by the curves of Fig. 15, which illus- trate the relationship between burning temperature and the ability of quicklimes to hydrate. It is noticeable that the lack of abiUty to hydrate through underburning is common to all limes, since more or less carbon dioxide still remains in combination with the lime, but that the loss of ability to hydrate, due to overburning, is noticeable at lower temperatures, and is much more marked at higher temperatures, if the quicklime is — — "~" ^ ~~" r / __ ^ £ ^ X ■s / Is 1' r- '0 " f. <>^ W 1. S / 1 -f te im te 1 / n /' k ol — o 1 / / ^ — Vio lor s^ Si / V ^ tQ " 1 Ij r ^ RATES OF HYDRATION OF QUICKLIMES (AU burned at 1050" C.) - III / r 5 10 15 20 Time la Minutes since lQtroduction_oi Bime Fig. 14. QUICKLIME 35 J — ■~v ' / \ / s / -/ ~, N 1 / s »; N ^ / 1 N ^ d 1 \ n b- ' \ X^ n J V s s 1 s s s / ^'t ■f' a y' >■ S>1 / .^ ^ — !-> :"0 re D^loi^it ) / ^ i? fliT fe^gU^ ■is. 1 r a. RELATION BETWEEN TEMPERATURE OF BURNINO ^ p- ^ - OF VARIOUS LIMES AND THEIR ABILITY TO HYDRATE _ _ 900° 1000° 1100° 1200' a^emperature at which Lime was JBunoed 0. Fig. 15. impure than if the quicklime is pure. This is due to the fact that the impurities may at high temperatures combine with lime to form com- pounds which will not hydrate, and also because the compounds formed tend to vitrify and close the pores of the lime, thus preventing free access of water. 23. Setting and Hardening. Accom- panying Phenomena. The setting of lime and lime mortar is a simple chemical proc- ess involving essen- tially only the evapo- ration of the large excess of water used in forming the lime paste, followed by the gradual replacement of the water of the hydroxide by carbon dioxide in the atmosphere, causing the lime hydrate to revert to the original calcium carbonate. Dry carbon dioxide will not react with dry hydrated lime, and it is there- fore necessary that excess moisture be present. Recent studies made at the Bureau of Standards by Messrs. W.E.EmleyandS.E. Young * have demon- strated the interesting fact that the loss of water during setting is not a continuous process, but that a reversal of the process occurs in the course of a week after mixing, and the percentage of water thereupon increases for several days or a week, after which it again decreases slowly. The data upon which this statement is based are presented by Figs. 16 and * Proc. Nat. Lime Manfr's. Assoc, 1913, pp. 254-267. -PROGRESS OFCA'RBONATION DRYINQ-OUT, n GATN IN COMPRESSIVE STRENGTH OF , HIGH CALCIUM LIME MORTAR 32 30 o £28 18 r' jtn he nn lix Id ^^' 1 fl J 1^* Sa24 , =22 ■Slao "43 16 1 t '\ / ■ ^ y '^1 \ / v V .C V n V M 1^ — - -, .^ ^1 « *> r" V s J / (*"' y ^ rb ka se- li o- ,,» S« sj ^ S/i A ^-^ y ^1 Y \ tA gSil2 ££io ,•■0 8 •8 we 4 .s y S / ■^ U .?= / c :f / "^ ?f ><* / / f > ^ 1 hh V B.2, s f/ ^t .^ ' / vtS/ 1 .^"Z ^> A f> ' V y\ 1^ r u r r r e- ^ c- f- ^ t" TENSrLE STRENGTH 1:3 LIME MORTAR DOLOMITIC AND HIGH CALCIUM ^ i'.c RA TE 3 A NO QL tc - - / ^A r •^. / j^/ ' tMi CI 'S M Au; n.( (Ju ck Li Tie _ ~^ TENSILE STRENGTH 1:6 LIME MORTAR HIGH CALCIUM ANO DOLOMITIC t nc Rfl TE A ND Q LJIC KL Mi 1 2 3 i l> 6 7 8 10 11 12 Age la Months Fig. 18. 5 6 7 8 9 10 11 12 Age In Months Fig. 19. as compared with the 60 pounds per square inch shown by the high-cal- cium lime mortars. The same condition is evidenced no less strikingly by the tests of mortars of 1 : 5 mix, Fig. 19. In this case the average strength of the dolomitic lime mortar is 124 pounds per square inch at one year, as compared with 39 pounds per square inch for the high-calcium lime mortars. Comparing the relative sand-carrying capacities of the two classes of lime, using as a criterion the relative tensile strengths of the 1 : 3 and 1 : 5 mortars in each case,t we find that they are almost equally eflScient except in the case of the strengths at one year, wherein the * See specifications for standard Ottawa sand, Chap. VII, p. 151. t The term " sand-carrying capacity " is here used not in its strict proper sense, referring only to the working qualities of the mortar. 42 MATERIALS OF CONSTRUCTION dolomitic lime far excels the high-calcium lime. (See also Fig. 23 and accompanying discussion.) In concluding the discussion of these tests of tensile strength of lime mortars, attention is called to the fact that the number of different limes tested, as well as the number of specimens actually broken, is far too limited to afford reliable data as to the actual strength which a given lime mortar may be expected to develop at a given age. In fact, the tests herein averaged showed a great variability in strength of individual specimens of the same mortar. It is believed, however, that the facts emphasized as to relative strengths of high-calcium and dolomitic mortars are dependable and typical of practically all limes. This conclusion is supported by the fact that practically all tests of the two classes of limes, the results of which have been published, show about the same degree of superiority in tensile strength of dolomitic lines compared with that of high-calcium lime as is shown by the tests herein quoted.* Compressive Strength, High-calcium and Dolomitic Lime Mortabs The compressive strength of the same specimens of mortars whose tensile strength is reported above is represented by curves I and II of Figs. 20 and 21. The compressive test specimens were half briquettes which had previously been broken In tension. The compressive strength values are doubtless somewhat higher than would have been obtained with cubical specimens, because the ratio of cross-sectional area to height was greater than it would be for cubes. However, it must be remembered that mortars are never subjected to compression except in situations where the bearing area is very great in comparison with the thickness of the mortar bed. It is fortunate, however, that the specimens were small, since the larger the specimens, the more imperfect will be the hardening in the interior. A further advantage in the use of half bri- quettes as compressive specimens, moreover, is the fact that a direct comparison between tensile and compressive strength of the same mortar is thereby afforded. The striking feature of these curves is again the fact that the com- pressive strength of dolomitic lime mortar is practically double that of high-calcium lime mortar. In the 1 : 3 mix. Fig. 20, the dolomitic lime shows an average of 357 pounds per square inch at one year, compared with 232 pounds per square inch shown by the high-calcium lime; and in the 1 : 5 mix. Fig. 21, the corresponding figures are 528 pounds per *See Municipal Engineering, Vol. 28, p. 6; also Engineering News, Vol. 51, June 9, 1904, and Trans. Nat. Lime Manfr's. Assoc, 1913, pp. 77-103. QUICKLIME 43 square inch for the dolomitic lime, and 251 pounds per square inch for the high-calcium. The condition as to the relative sand-carrying capacity * of the two classes of limes, as measured by the relative compressive strengths of the 1 : 3 and 1 : 5 mortars, is practically identical with the condition found in the case of the tensile strength tests. In both instances the sand- carrying capacity of the dolomitic lime equals or exceeds that of the high-calcium lime. (See also Fig. 23 and accompanying discussion.) M M M M - COMPRESSIVE STRENGTH 1 ;3 LIME MORTARS HIGH CALCIUM AND DOLOMITIC ,1050 f/ .' ^^^ o" k I*™ ^ 1 0.750 3700 §650 U H ig liC aU i'V H) dratii dl An e) vn r f '} USOO g«0 ^100 m350 ^soo ^250 a200 |l50 f s 1 t J I n olo m tic (Q uit kl Lin e) 11 — "~ i/ \ / k ■rf V> P 1 u t- iS CTt m m 1 I L L L L L 123166789 10 11 12 Age In Months Fig. 20. 123466789 10 11 12 Age in Months Fig. 21. Effect of Impueities on Propekties The principal impurities in limestones have been stated to be silica, iron oxide, and alumina, all of which may under certain conditions com- bine with lime and magnesia to form complex compounds which would profoundly affect the properties of lime if present in sufficient amounts. As a matter of fact, however, the maximum content of these constituents usually allowed in the finished lime is so small (not over 5 per cent) that their effect can never be very pronounced. In addition, the temperature attained in calcination is usually not sufficient to effect the formation * See footnote, p. 41. 44 MATERIALS OF CONSTRUCTION of compounds of lime with silica, iron oxide, and alumina, and these latter constituents therefore exist unchanged in the calcined lime and exert no influence on strength other than as diluents. The U. S. Bureau of Standards has made an incomplete study of the effect of various impurities upon the properties of lime, and while no definite report has been made, the following general conclusions have been drawn by Mr. Emley, Assistant Chemist: * " The presence of small amounts of sihca tends to decrease the plas- ticity, sand-carrying capacity, and yield of lime, but has no apparent effect upon its hardness or strength. The same may be said of iron, except that large amounts (25 per cent) show a marked increase in both strength and hardness. Alumina increases all the factors above men- tioned arid also improves the color, so that its presence even in large amounts is very desirable. On the contrary, gypsum shows detrimental effects even when only 1 per cent is present. Kaolin seems to act in a manner similar to silica and iron." The actual import of the conclusion expressed by the investigator just quoted must not be exaggerated. It is not stated that the silica^ or the iron, or the alumina content of commercial limes is under any circumstances a large factor in strength or other mechanical properties of that lime; it is simply stated that when quantities of these substances far exceeding the Umits of possible composition of ordinary Umes are com- bined with lime, the effect upon mechanical properties is a very notable one. When these, adulterants are present in such proportions, indeed, we no longer have an ordinary lime to consider, but instead a very dif- ferent material which partakes of the nature of the hydraulic limes and cements, which owe their superior mechanical properties solely to their comparatively large content of argillaceous materials, the very con- stituents whose influence upon lines we are now discussing — ^the silica, the iron oxide and the alumina, f Finally, it must be stated that, what- ever the extent of the contamination of the lime by these argillaceous materials, the latter would not be materially effective in so far as increas- ing the strength of the Ume is concerned, were not the calcination tem- perature considerably higher than that necessary for proper calcination of the lime. This circumstance explains the fact that it is not easy to overburn a very pure limestone, but an impure hmestone may be over- burned because a temperature is attained at which combination is effected between lime and sihca and iron oxide and alumina, the resulting com- * Technologic Paper No. 16, Bureau of Standards, Feb. 1913.- t The compounds formed by these constituents in hydraulic cements are inert, just as they are in hard-burned lime, until the clinker formed in burning is pulverized to an impalpably fine powder. • QUICKLIME 45 pounds being of such a nature that they greatly impair the ability to slake, the quick-setting properties, and the strength of the lime, as above noted. Effect of Temperature of Calcination The temperature of calcination affects the properties of lime to a marked degree, as has been noted in Arts. 22 and 25 above. It may be so low that dissociation of the carbonates is incomplete on the one hand, causing loss of power to hydrate, or so high that the combination of the lime with the impurities is effected on the other hand, again causing loss of power to hydrate and also causing an increase in the waste. In the first event the Ume is under-burned, in the second it is over-burned, and as has been shown, everburning is practically impossible unless the limestone contains impurities. Both the overburned and underburned material can easily be recognized in the kiln product, moreover, and should be sorted out before the lime is marketed. The underburned material can be recalcined, but the overburned material must be thrown away. It appears, therefore, that (provided the lump lime is properly sorted) the only effect of improper regulation of calcination temperatures is loss of manufacturing efficiency. It is to the interest of the manufacturer to so burn the limestone as to obtain a maximum amount of marketable lime, but the engineer or architect who uses the lime is not concerned beyond the point of being sure that the underburned or overburned material has been rejected by the manufacturer. Effect of Method of Slaking The statement has often been made that lime slaked, as is usually the case on construction work, by the use of a large excess of water, will produce a mortar of lower strength than the same Ume slaked with little or no excess of water. It appears, however, that this statement may be an error, due perhaps to the fact that an increased amount of water used in slaking is apt to mean an increased percentage of water in mortar mixtures. In any event it is an undoubted fact that, under the condi- • tions of slaking and mixing which usually obtain on construction work, it is far safer to instruct the mason's tender to use such an excess of water as to produce a plastic putty or even a paste, thus practically elimi- nating the possibility of incomplete hydration, than to attempt to secure a physically dry hydrate or even one containing only a small excess of water through the instrumentality of an unskilled laborer. 46 MATEEIALS OF CONSTRUCTION An investigation of the effect of various methods of slaking made by Mr. H. Burchartz * resulted as follows, the tests being made upon a lime- sand mortar of 1 : 3 mix at the age of one year: Method of Slaking. Compressive Strength. Kg. per Sq.crn. Lbs. per Sq.in. Water added all at ourr no stirrinff 26.0 28.1 32.1 370 400 Slaked with larce excess of water 457 480 MO d"» "MO ^400 u 380 psco .3 340 RELATION BETWEEN SIZE OF SAND AND STRENGTH OF LIME MORTARS Effect of Character of Sand The influence of the size of the sand grains upon the strength of lime mortars is illustrated by Fig. 22, which represents the results obtained in tests of mortars made at the Bureau of Standards by Mr. W. E. Emley and Mr. S. E. Young.f The specimens used in these tests were 2-inch cubes, ninety days old, containing 1 part by weight of , quickUme to SJ parts sand. Run-of-bank Ottawa sand was obtained and separated by screen- ing into 4 sizes: 20^0, 30-40, 40-60, and 60-80 mesh. A Merri- mac River sand was also obtained and screened to procure a sand of 10-20 mesh size. The curves represent the average of five tests of each mortar for each class of lime. It will be noted that by far the strongest mortars are produced by the fine sand, the compressive strength being roughly an hyperbolic function of the average diameter of the sand grains. These results substantiate the generally accepted belief that fine sand IS best for lime mortars, and should be especially noted in view of the fact * Luftkalke und Luttkalk-mortel. t Proc. Nat. Manfr's. Assoc, 1913, pp. 254-257. .005 .010 .015 ,020 .035 .030 .035 .040 .045 Size of Sand Grains In Inches Fig. 22. QUICKLIME 47 that coarse sand or sand of a composition well graded from fine to coarse has been shown to yield the strongest mortars in the case of both natural and Portland cements. Effect of Size of Test Specimens The setting and hardening of lime mortars has been explained as a process of gradual replacement of. the water of hydration by carbon dioxide in the atmosphere. Since the successful accom- plishment of the process is dependent upon the action of the air, the hardening process begins at the surface of the exposed mortar and proceeds inward very slowly. (Emley * found that a 6-inch cube, the outermost portion of which possessed 17.5 per cent of CO2 and 19.8 per cent of H2O after nine months, possessed 1.2 per cent of CO2 and 31.1 per cent of H2O, at the center, showing that setting and hardening had scarcely begun 3 inches below the surface.) The gain in me- chanical strength parallels the gain in hardness, and the aver- age unit strength of a specimen of large cross-sectional dimen- sions at a given age will there- fore never equal that of a smaller or thinner specimen in which the ratio of exposed surface to volume is greater. This fact is forcibly illustrated by the curves of Fig. 23, which have been plotted from the data of a series of tests made by Mr. M. Gary.f The various mortar mixtures were made from high-calcium lime and a natural sand. Slaking, mixing, and molding were done by hand; all specimens were cubical, either 1.26 inches or 2.80 inches on a side; the * Loc. cit., p. 259, t M. Gary, Erhartung von Kalkmorteln. " Bau Materialienkunde," Vol. 5, 1900, pp. 216-217. EFFECT OF THICKriESS OF CUBICAL SPECIMEN UPON COMPRESSIVE STRENGTH OF LIME MORTARS fi'f \\f. 1:5 .lOLUa — - ' ^ ^ ^ 2.80 < :ube_ , u li — ' i 6'( ■^ e^ - 1 / 1.' a \J 3 ovwi. '- ' gm - ft 300 - f aT a, n'( \\ e VH — ■§ f , ^ ' ^-inn (• t r w 5i tat R l_ - n..m Mol-t at ..- ^m S i' ffi 2^ >o ;u ^ 1 m 1 \4 3 "^ ^■ ^ =• y • rs'os?- — ■ 200 - 13 —■ iSO-tuie- r ii . ^. ^^^ — ' -— xoo 'f r /3Mo. 10 Days Age Fig. 23. 48 MATERIALS OF CONSTRUCTION slaked lime was seasoned one week before molding, and all specimens were stored in air. Each curve is the average of about four compressive tests at each age. A remarkably constant relation is shown for all mortar mixtures between the strength of the small cubes and that of the larger cubes, that of the latter always amounting to from 40 to 70 per cent (usually about 50 to 60 per cent) of that of the former. It is interesting to note further that the relation is nearly constant at all ages except that the pro- portionate gain in strength of the larger cubes between the ages of one and three months is less than it is at other periods. 28. Relative Applicability of Various Limes to Special Uses. High- calcium, magnesium, and dolomitic limes each possess certain special advantages as building materials. Strong preferences exist among plas- terers and masons, but these preferences are not always well founded and are not the same in different sections of the country. In the following enumeration of the principal structural uses of limes an effort has been made to indicate the special reasons why one class of lime may be superior to another for the purpose indicated. The limitations of each hme can- not be defined by hard-and-fast rules, however, and divergence from the practices noted will doubtless be frequently encountered on construc- tion work. Plasters. The plasterer has usually a strong preference for mag- nesian or even dolomitic limes for both the " scratch " coat, the " brown " coat, and the finish, or " skim " coat, on plastered walls. This is due to the fact that +hese classes of limes work much more smoothly under the trowel than do high-calcium limes, and also to the fact that the more slowly setting properties of these limes make it possible for him to spread a larger surface in one operation before stopping to complete the surface treatment. (Leveling up and then roughening with a brush or scratching with a trowel , or other implement, in the case of the first coat, and using the float to produce a smooth surface in the case of the brown coat and the finishing coat.) Another advantage of magnesian and dolomitic limes is the lessened danger of the development of " lime-pops " caused by the late hydration of small particles of lime. The increased hardness of the wall surface secured with dolomitic limes is also usually considered an advantage. One possible additional reason for the preference felt by the plasterer for magnesian or dolomitic hmes is the fact that he is usually given a richer mortar than he would be given if high-calcium lime were used. This is due to the fact that magnesian limes will not carry as much sand as high-calcium limes and the temptation for the mason's tender to turn out a too heavily sanded mixture is not so great. QUICKLIME 49 Mortar for Ordinary Brick Masonry. The bricklayer usually pre- fers the magnesian or dolomitic lime mortar for the same reason that the plasterers do, i.e., its more slowly setting properties and greater plasticity make it possible for him to spread with less effort a greater surface with mortar, and he is consequently able to lay a greater number of bricks in one operation. The cooler limes are also preferred because of the lessened danger of encountering an irregular and lumpy mortar caused by unskillful slaking and consequent burning of high-calcium lime mortar. Mortar for Face Brick. Dolomitic lime is invariably preferred for laying dry-pressed or face brick. This class of bricks is commonly laid with joints not exceeding | inch in thickness, and it is absolutely neces- sary that the mortar be very plastic and very slow setting in order that such thin joints may be secured, and the bricks at the same time be accu- rately placed in proper alignment. Mortar for Stone Masonry. High-calcium lime mortars are usually preferred by the stone mason because of their more quickly setting properties, which constitute a distinct advantage. They become stiff enough to carry the load put upon them without deformation in a very short time, thus facilitating the proper setting of heavy stones. 29. Production, Value and Uses of Lime. Statistics of the lime industry are presented in the table below. These data have been abstracted from the reports of the U. S. Geological Survey, published annually in " Mineral Resources of the United States." STATISTICS OF LIME INDUSTRY Year. 1910. 1911. 1912. 1913 No. plants operating . (1125) (1139) t (1017) (1023) Use. Quantity, Short Tons. Value per Ton. Quantity, Short Tons. Value per Ton. Quantity, Short Tons. Value per Ton. Quantity, Short Tons. Value per Ton. Building lime Chemical works . . , . . Paper mills . . 1,722,488 182,043 286,922 29,421 28,921 585,876 496,930 149,179 $4.26 3.82 3.76 8.14 4.62 2.93 4.27 3.70 1,488,567 256,215 286,485 36,424 30,167 596,664 531,249 167,144 S4.54 3.65 3.87 6.65 4.59 2.87 4.15 3.55 1,556,446 282,984 290,347 30,988 40,595 604,607 560,186 157,843 $4.22 3.50 3.81 6.01 4.40 3.06 4.40 3.79 1,358,099 388,369 284,090 32,236 49,591 590,229 692,265 200,511 $4.43 3.45 4 18 Sugar factories Tanneries 6.72 4 38 Fertilizer 3.05 Dealers, uses not stated 4.56 Other uses * 3.61 Total 3,481,780 3.99 3,392,915 4.03 3,529,462 3.96 3,595,390 4,97 * Includes lime for sand-lime brick, slag cement, alkali works, steel works, glass works, smelters, sheep dipping, disinfectant, manufacture of soap, cyanide plants, glue factories, purification of water, etc. CHAPTER III HYDRATED LIME 30. Definition. " Hydrated lime " is that article of commerce which is produced by the hydration, at the place of manufacture, of ordinary quicklime, the process being so conducted that excess water is largely evaporated by the heat generated, and resulting in the production of a physically dry powder from which impurities or lime existing in the anhydrous form may be removed by air separation or screening.* 31. Process of Manufacture. Three stages of manufacture character- ize all processes of preparation of hydrated hme, whether patented or not. These are: (1) The lump quickhme must be crushed to a fairly small size. (2) The crushed material must be thoroughly mixed with a sufficient quantity of water. (3) The slaked lime must, by air separation, screening, or other- wise, be separated from lumps of unhydrated lime and impurities, or the entire mass must be finely pulverized, thus leaving the mate- rial which made up the lumps in the finished product. Crushing. The degree, of crushing employed at various hydrated lime plants varies greatly. In some plants the quicklime is simply crushed to a 1-inch size; in others, and this is the more common practice, the quicklime is crushed in a pot or Sturtevant crusher to a size of | inch or under. A few plants, after crushing the quicklime, pulverize it so that the greater portion will pass a 50-mesh sieve. Mixing with Watefr. A great number of processes, and several dis- tinct types of machines are used for hydrating quicklime. The only two methods extensively used, however, are the batch process, using a machine of which the Clyde hydrator is typical, and the continuous process, in which a machine of which the Kritzer hydrator is typical *The proposed standard specifications of the Am. Soc. for Test. Mat. define hydrated lime simply as "a dry flocculent powder resulting from the hydration of quicklime." 50 HYDRATED LIME 51 1 ■ ■-" fr' / *. _ ■■■■ p.. -" i?r.v^f^^^f^ Fig. 24.— Clyde Hydrator. is used. Either process may give entire satisfaction providing only that the plant operation is properly supervised and the character of the product kept under strict chemical control. The Clyde batch process hydrator (Fig. 24), is a batch machine in which a given quantity of quicklime (usually 1 ton) is placed, and the proper quantity of water added by means of a spray. The machine itself consists of a revolving pan provided with plows, arranged in a horizon- tal spiral, which stir up and mix the water and lime. The water is weighed and added in a predetermined amount. The hydrated lime is scraped from the pan through an open- ing in the center into a hopper below the hydrator. The Kritzer continuous process hydrator (Fig. 25), consists of a number of cyUn- ders, arranged one above the other, which are provided with screw-conveyors revolv- ing around a central shaft. The quicklime is fed into the upper cylinder in a continuous stream and here water is sprayed upon it, the amount being regulated by valves. The moist lime is gradually worked by the conveyors through the upper cylinder into the lower ones, and by the time it is discharged from the lowest cylinder it is entirely hydrated. The spray of water is admitted through a stack located near the point of entrance of the lime to the upper cylinder. This stack serves as an outlet for both the dust and the steam which are unavoidable incidental products of the process of hydration. The water descending over a series of baffle plates in the stack absorbs the dust and condenses the steam. It therefore returns the dust to the hydrator and enters the upper cylinder of the hydrator hot. The advantages claimed for this type of hydrator over the batch machine, especially for treatment of hot limes, are those accruing from the introduction of water already hot, and the introduction of both quicklime and water in small constantly flowing streams, thus- affording an opportunity for rapid and thorough mixing. On the other hand, the possibility of accurately gauging the proportions of water and quicklime is a great advantage claimed for the batch type of hydrator. Removal of Lumps of Unhydrated Material. Owing to the increase in volume which accompanies slaking, the lumps of hme fall into powder during the process. Any impurities in the lime will not slake, however, and therefore will remain with any imperfectly hydrated lime as lumps which can be removed from the finished product by screening or by other means such as the method of air separation now commonly used. The form of screen usually employed consists of a wire netting 52 MATERIALS OF CONSTEUCTIOISl stretched on an inclined frame which is mechanically agitated as the material traverses its length. The usual size of the mesh is from 35 to 50 meshes per Unear inch. The whole structure must be enclosed to keep the dust in. 'iiiiii iiiigiiiiiiiiiiiiiiiiiifriiiiiiiiiii iiiii Air separation systems usually involve the use of a Raymond impact mill or similar device in which the hydrate is subjected to the beating action of rapidly revolving blades. The material thereupon encounters a current of air by which the fine material is carried off in suspension, while the larger particles settle out. The fine material is subsequently HYDRATED LIME 53 deposited in a chamber provided for the purpose in the air duct, the precipitation being effected by the reduction in velocity of the air current caused by the enlargement of the air duct. The air circuit is a closed one, the same air being used over again, and it is therefore a dust-proof device. A regulation of the size of the particles retained is very easily attained by merely varying the velocity of the air current through the speed control of the fan. A certain few hydrates, especially those which are offered on the market as finishing limes, are not subjected to either screening or air separation, but are pulverized just as they come from the hydrator in a Fuller mill or Bonnot mill. By this method the imperfectly hydrated lime, impurities, etc., find their way into the finished product. The tailings from the screen, or material rejected by the air separator, usually contain a considerable proportion of unslaked lime. It is there- fore possible to regrind this material and market it as agricultural lime. Hydrated hme is usually packed in burlap or duck bags containing 100 pounds, or in paper bags containing 40 pounds. 32. Properties and Uses. Hydrated lime and ordinary lime which has been properly slaked on the work are, of course, exactly the same material, and therefore should have identical physical properties. As a matter of fact, however, the usual experience has been that stronger and more quickly setting mortars, and ones which shrink less upon setting and hardening, are derived from hydrated limes than from ordinary quicklime. On the other hand, mortars prepared from hydrates are vastly inferior to those prepared from quicklimes from the standpoint of plas- ticity, sand-carrying capacity and yield, except in the comparatively rare instances in which the hydrated lime paste is allowed a period of seasoning before being used. The tensile strength of hydrated lime mortars, both high-calcium and dolomitic, is represented by curves III and IV of Figs. 18 and 19, previously referred to in Art. 27. The mortars whose properties are repre- sented by these curves are made from precisely the same limes as those represented by curves I and II of the same figures, the only difference being that in the one case ordinary lump lime was hydrated by the experi- menter, in the other the lime was procured from the same manufacturer in the hydrated form. The proportions used for the test of Fig. 18 were 1 part by weight of hydrated lime to 3 parts by weight of Ottawa sand, and for the tests of Fig. 19, 1 part by weight of hydrated hme to 5 parts by weight of sand. The test specimens of hydrated lime mortar therefore contained a relatively lower percentage of CaO than did the corresponding specimens made from quicklime, since a considerable amount of chemically combined water was included in the weight of the hydrated hme. 54 MATEEIALS OF CONSTEUCTION A comparison of curves III and IV of both Figs. 18 and 19 shows a certain degree of superiority of dolomitic hydrated lime mortar over high- calcium hydrated lime mortar at most ages, the advantage of the former becoming apparent only after the expiration of several months, however, and being more pronounced in the case of 1 : 5 mortar than in the 1 : 3 mortar. Curves I and III, also curves II and IV of these figures afford a direct comparison between the tensile strengths of the hydrated lime mortars and the corresponding quickUme mortars. The following con- clusions appear to be warranted by the test results : (o) High-calcium hydrated lime mortar is about 100 per cent stronger in tension than the corresponding quicklime mortar; (b) Dolomotic hydrated lime mortar and dolomitic quickUme mortar appear not to differ materially in tensile strength except at early ages, and the great superiority of the hydrate noted in the case of high-calcium limes appears not to be characteristic of the dolomitic limes at any age or for any mix. This condition in a measure supports the belief that the superiority of hydrated lime mortars is largely due to the more perfect hydration. Natur- ally this factor would be less operative in the case of the dolo- mitic limes, whose hydration is attended with far less difficulty than in the case with high-calcium limes. The compressive strength of the same specimens of hydrated lime mortars is represented by curves III and IV of Figs. 20 and 21; a com- ( parison of curves III and IV in each case reveals the same situation as to the relative compressive strengths of high-calcium and dolomitic lune mortars as has been noted in the discussion of tensile strength. The dolomitic limes show a less rapid early gain in strength, but ultimately surpass the high-calcium limes by a considerable amount. Comparing the relative compressive strengths of the hydrated lime mortars and the corresponding quicklime mortars (curve I vs. curve III, also curve II vs. curve IV in each case), the following conclusions may be again drawn : (a) High-calcium hydrated Ijme mortar is from 100 to 200 per cent stronger in compression than the corresponding quick- lime mortar: (b) Dolomitic hydrated lime mortar is considerably stronger in compression than the corresponding quickUme mortar, but the superiority of the former is usually less marked than in 'the case of the high-calcium Umes. HYDRATED LIME 55 33. Hydrated Lime vs. Quicklime. The superiority of hydrated lime over the ordinary lump lime from the standpoint of the mortar strength developed has been indicated in the preceding article. The advantage accruing from the use of a material which need only be mixed with water, instead of being slaked upon the work with the attendant danger of burn- ing or incomplete hydration, has also been noted above. There remain, however, certain advantages and disadvantages to be derived from the use of the commercial hydrate which will be considered briefly. Hydrated lime can be more conveniently handled than lump hme because of its powdered condition, and can safely be stored or shipped by rail or water in cloth or paper bags or even in bulk. On the other hand lump lime deteriorates rapidly in storage or transportation, through air-slaking, is considered an unsafe commodity to carry by water, and wherever kept always constitutes a fire hazard. Ground quicklime could be handled in much the same way as is hydrated lime, and perhaps would be so handled were the fact generally recognized that the extent of air slaking is dependent largely upon the fineness of the material and consequent accessibility to the action of the air, rather than upon the quality inherent in the quicklime or in the hydrate. A prejudice against ground lime, founded upon its resemblance to air-slaked lime, is still operative, however, in reducing its use as a structural material. One property of hydrated lime which often constitutes an advantage on construction work is the fact that it is ready to be immediately incor- porated with sand and water to form mortar, whereas ordinary lime must be allowed to season for from one day to several weeks after being slaked, thus probably causing delay. The fact that hydrated lime is a physically dry material is an advantage in mixing mortars. The dry hydrate can be mixed with sand much more easily than can a lime paste or putty, and a more thoroughly homogene- ous mixture is obtained before the excess water is added to make a plas- tic mortar. On the other hand, mortars prepared from hydrated lime are very " short " and non-plastic, the volume of lime paste derived, i.e., the yield, is small, as compared with that obtainable from quicklimes, and the sand-carrying capacity is low. The lack of plasticity of lime hydrate mortars is alone sufficient to condemn the material from the stand- point of the practical mason and plasterer. Some hydrates are, indeed, so lacking in colloidal properties that they are absolutely gritty. As a result there is almost no market for hydrated lime as a plastering or finishing hme, with the exception of certain hydrates of special character which are produced in one district of limited extent. The greatest value of hydrated lime as a structural material is as an 56 MATERIALS OF CONSTRUCTION ingredient of natural and Portland cement mortars and concretes. So used, it effects marked improvement in quality, especially m that greater plasticity characterizes the-wet mixture. Mortars and concretes in which a small proportion of hydrated lime has been incorporated are also improved from the standpoint of impermeability and homogeneity of structure. The use of hydrated lime in combination with cements will be discussed in the chapters devoted to the properties and uses of natural and Portland cement. 34. Special High Alumina Hydrated Lime. The unsatisfactoriness of hydrated lime as a plastering material on account of the non-plastic, poor working mortars formed, and the slowness with which plastered walls dry out and harden, has led to many attempts to so modify the character- istics of the material as to correct these defects and enable it to compete with gypsum plasters as an easily applied, ready-to-use wall plaster. These efforts have been far from uniformly successful, but, in at least one instance, a material has been produced whose performances have made a very favorable impression upon the architect, builders, and artisans, and the material has come to be a fairly well-known article of commerce. This product, marketed under the trade name " Alca Lime," is made by the incorporation with selected hydrated Ume of about 15 per cent of a patented calcium aluminate compound which is derived as a slag from a blast furnace. This compound is not a normal blast furnace slag, but is derived from a furnace whose charge is so regulated that a slag high in alumina and relatively low in silica is produced. Its com- position is about 25 to 35 per cent alumina, 20 per cent silica, and 35 to 40 per cent lime and magnesia. The slag is granulated by running the molten material into water or by causing it to encounter a jet of steam or water, and the material is subsequently ground to such a degree of fineness that not more than about 25 per cent will be retained on a 200-mesh sieve. The lime manufacturer secures this material in a pulverized con- dition (from a proprietary company which holds the patents) and mixes it with his finished hydrated lime. An ordinary mixer is often used, but greater uniformity of composition is attainable if the two materials are together passed through a mill designed for fine pulverization of cement and other similar materials. Hair is subsequently added to the mixture in case it is intended to be used as a first plaster or stucco coat upon lath. Otherwise it is marketed without an admixture of hair. 35. Properties and Uses of High Alumina Lime. The physical and mechanical properties of Alca lime have been little investigated by parties other than those directly interested in the production of the material. In the absence of authentic published results of tests, there- fore, the discussion of the properties of the material will be confined to a HYDRATED LIME 57 consideration of its performances in use, as noted by observers who have had experience with it on actual construction work. The fact should not be overlooked in this cormection that this practical test must ultimately fix the place which any material shall come to hold as a material of construction, Alca lime, considered from the point of view of the physical chemist, is a distinctly different material from ordinary hydrated lime, and par- takes in a measure of the nature of the hydraulic limes and puzzolan cements considered in the chapters which follow. Its setting is not wholly a process of replacement of the water of the hydrate by carbon dioxide, but is a complex and not perfectly understood phenomenon in which the aluminate present plays an important part. Reasoning from the partial analogy to the constitution of Portland cement, it appears probable that the initial setting is due to the hydration of the alumi- nate, resulting in the formation of amorphous hydrated tricalcium alu- minate, and possibly amorphous hydrated alumina. These amorphous compounds tend to render the mixture more plastic than is a mixture of water with hydrated lime alone, because of the fact that the dry-slaked hydrate is in part at least in a crystalline form, which causes it to " work short." The presence of the amorphous hydrate may also have a tendency to cause the crystalline lime hydrate to change over to the amorphous hydrated form, with consequent improvement in plasticity. The amor- phous hydrated aluminate may possibly crystallize during the subsequent period of hardening, but there appears to be no good reason for believing that the hardening of the material is not due almost wholly to the reear- bonation of the lime, as in the case of ordinary limes. Alca lime comes upon the work ready to be mixed with sand and water and be used almost immediately. Best results are obtained if the sand and hme are first mixed dry, then combined with about 16 per cent of water, thoroughly mixed, and allowed to stand for not less than one hour before being used. Where conditions permit allowing the mix to stand overnight, an advantage is gained by so doing in that a very smooth working plastic mortar is thus derived. Alca lime has given satisfaction as an interior plastering material used on wood or metal lath, terra cotta, or brick, and as a mortar material for laying brick or stone masonry. For use as the first coat of plaster or stucco it may be obtained from the manufacturer already mixed with hair. The brown coat of plastered walls should be applied before the scratch coat is quite dry, and an interval or two or three days in good drying weather allowed before the finish coat is applied. The material has excellent sand-carrying capacity, proportions of 1 part by weight of Alca to 3 or 3^ parts by weight of sand being recommended for first- coat work on wood or metal lath, and 1 part Alca to 5 parts sand by 58 MATEEIALS OF CONSTRUCTION weight for first or second coats on brick, terra cotta, stone, etc. For second-coat work on lath 4 or i^ parts by weight of sand may be used with 1 part of lime. Alca lime may also be used as a skim coat except where a pure white finish is required, its color being a gray-white. Alca-lime plasters are well liked by the mason, who finds their working quahties to be equaled only by well-seasoned lump-Ume plasters. Prob- ably no other plastering material is so well adapted for all sorts of construction work— interior plastering, outside stucco, and mortar for brick, terra cotta, or stone masonry. It is also used in combination • with Portland or natural cement, with all the advantages which accrue from the similar use of ordinary hydrated lime. So used it may in fact excel hydrated hme by still further improving the working quali- ties of the mortar and rendering it still denser, thus tending to produce a very impervious mortar or concrete. 36. Production, Value and Uses of Hydrated Lime. The uses of hydrated lime structurally and in various arts and industries have been indicated in the discussion of the uses of ordinary lime in Chapter II. No statistics are available as to the amount of hydrated lime used for various individual purposes, but the amoimts so used have been included in the individual items and the totals given in the table at the end of Chapter II. The production and the average value of the total amount of hydrated lime used for all purposes in the United States are indicated by the following data abstracted from " Mineral Resources of the United States." STATISTICS OF HYDRATED LIME INDUSTRY . Year. 1910. 1911. 1912. 1913. Quantity, Short Tons. Value per Ton. Quantity, Short Tons. Value per Ton. Quantity, Short Tons. Value per Ton. Quantity, Short Tons. Value per Ton. 320,819 $4.02 304,593 $4.50 416,890 $4.39 493,269 $4.47 CHAPTER IV HYDRAULIC LIME AND GEAPPIER CEMENTS HYDRAULIC CEMENTING MATERIALS IN GENERAL 37. Introductoiy. The cementing materials heretofore considered have all belonged to the class of non-hydraulic cementing materials; all have been very simple both in composition and chemical action. The hydraulic cementing materials, however, comprise a class of products of very complex and somewhat variable composition and constitution, products, moreover, whose physical characteristics are not definitely fixed by chemical composition, and whose actual constitution is still im- perfectly understood. These materials all possess in common one property known as " hydrauUcity," i.e., the abiUty to set and harden under water. In com- position they agree to the extent that they all consist essentially of lime, silica, and alumina, or of lime and magnesia, silica, and alumina andiron oxide (i.e., magnesia may replace a part of the Ume, and iron oxide a part of the alumina). The hydraulic cementing materials include hydraulic limes, grappier cements, puzzolan cements, natural cements, and Port- land cements. 38; Classification of Hydraulic Cementing Materials. The Hydrau- lic Index and the Cementation Index. The degree to which the property of hydraulicity is developed by Umes and cements is dependent almost wholly upon the character of the compounds -of lime and magnesia with silica, alumina, and iron formed during manufacture. The chemical composition of the raw materials and the details of the manufacturing processes are influential factors, but are operative, in an indirect manner, only in so far as they affect the ultimate , constitution of the finished product. Formerly it was considered possible to classify cements accurately according to their relative hydrauUcity by the use of a criterion called the " hydraulic index," which is simply the ratio of silica and alumina to lime, expressed as follows : per cent silica-)- per cent alumina Hydraulic index = -t. • per cent lime 59 60 MATERIALS OF CONSTEUCTION The values obtained by the use of this formula permit the tabulation of limes and cements by their hydraulic indices as follows: ,. * Hydraulic Index. Product. 0.00-0. 10 common lime. 0. 10-0. 20 -. feebly hydraulic lime. 0.20-0.40 eminently hydrauhc lime. 0.40-0.60 Portland cement. 0.60-1.50 natural cements. 1 . 50-3 .00 weak natural cements. 3 . 00. : . . . • puzzolans. This attempt to devise a formula which express in scientific terms the relative hydraulic values of hmes and cements has been unsuccessful, primarily because it fails to take account of the character of the compounds formed during manufacture, i.e., the constitution of the product, and secondarily because it assumes silica and alumina to be quantitatively equivalent in chemical action and fails entirely to take account of the action of both magnesia and iron oxide. (See discussion of constitution of cement, Art. 106.) Eckel has made an effort to avoid the deficiences and errors involved in the hydraulic index by the proposal of a new criterion of hydraulicity which he terms the " cementation index." The new rule proposed is based upon the assumption that the essential constituents of hydraulic cementing, materials are: Tricalcium silicate, SCaO, Si02 ; Trimagnesium silicate, BMgO, Si02; Dicalcium aluminate, 2CaO, AI2O3; Dimagnesium aluminate, 2MgO, AI2O3; Dicalcium ferrite, 2CaO, Fe203; Dimagnesium ferrite, 2MgO, Fe203. This assumption is primarily based upon the conclusion previously announced by S. B. and W. B. Newberry, as the result of extensive synthetic investigations of the constitution of Portland cement, that the essential constituents of cement are tricalcium sihcate, 3CaO, Si02, and the dicalcium aluminate, 2CaO, AI2O3. Eckel has extended the applica- tion of the Newberrys' conclusions by assuming " that magnesia is, molecule for molecule, equivalent to lime in its action " and " that iron oxide is, molecule for molecule, equivalent to alumina." * Eckel, " Cements, Limes and Plasters," p. 169. HYDRAULIC LIME AND GEAPPIEE CEMENTS 61 Taking into account relative weights, the ratio of the lime to each of the other constituents will be expressed as follows: 3CaOH-SiO2 = 3(56.1)H-(60.3) =2.8 2CaOH-Al2O3 = 2(56.1)-^-(102.2) = l.l 2CaO-r-Fe2O3 = 2(56.1)-^(159.7)=0.7 CaO-^MgO = (56.1) H- (40.3) = 1.4 and the form of the cementation index is : 2.8(%Si02) + 1.1 (%Al203) +0.7(%Fe2O3) %CaO+1.4(%MgO) = 1. If the value of the cementation index falls below 1.0, the cement must necessarily contain free lime or magnesia; when it rises above 1.0 the lime-magnesia content is lower than the theoretical maximum. If the cementation index be used in classifying the various hydraulic ofementing materials, the ranking of the latter according to hydrauhcity is as follows: Cementation Index. Product. . 00-0 .30 common lime. 0.30-0.70 feebly hydrauhc lime. . 70-1 .10 , eminently hydraulic lime. 1.00-1.20 Portland cement. 1 . 15-1 . 60 * natural cements. 1 . 60-1 .90 puzzolans (slag cements) . The cementation index affords a better basis for the classification of hydrauhc cementing materials than does the hydrauhc index, but it cannot be the sole basis of classification, because it is rigidly fixed by an assumption as to the essential constitution of all this class of materials. If it could be granted that the essential constituents of hydraulic cements are tricalcium silicate and dicalcium aluminate (with the corresponding compounds formed by the replacement of lime by magnesia, or alumina by iron oxide), the fact must still be recognized that the extent to which the lime combines with sihca, alumina, etc., during manufacture to form these compounds is dependent upon a number of variable factors, among which the temperature of burning is the most influential. (As a matter of fact we now know that the characteristic aluminate is the tricalcio not the dicalcic aluminate.) * A few natural cements show indices below 1.15 or above 1.60. 62 MATEEIALS OF CONSTRUCTION It must be concluded that the value of the cementation index as a basis of classification of hydraulic cements is rather limited. It has a very distinct value, however, as a basis of determination of the proper proportions of the raw materials used in making artificial cements, such as Portland cement and slag cements. It is now commonly used for this purpose by the majority of the cement chemists, and doubtless will continue in favor until the time when a more general understanding of the constitution of cements justifies a change in manufacturing practice. In view of the present unsatisfactory state of common knowledge of cement chemistry the classification of hydraulic cementing materials herein utilized will be one which is based upon physical properties and method of manufacture, rather than upon chemical characteristics. The following classification is a modification of one proposed by Mathews and Grasty.* (1) Common Lime. Lime made by burning relatively pure Ume- stone at a very low temperature, the product being one which slakes when mixed with water and which possesses no hydraulic properties. (2) Hydraulic Limes. Limes made by burning slightly argillaceous Umestones at a low temperature, the product being one which will slake slowly but which at the same time possesses feebly hydraulic properties. (3) Natural Cements. Cements made by burning distinctly argil- laceous limestones at a comparatively high temperature, the product being one which will not slake, but which when ground possesses hydraulic properties. (4) Portland Cement. Cement made by burning an artificial mixture of argillaceous and calcareous materials at the temperature of incipient vitrification, the product being one which will not slake but which when ground possesses marked hydrauhc properties. (5). Puzzolan or Slag Cements. Cements made by incorporating slaked hme with granulated blast furnace slag or a natural puzzolanic material such as volcanic ash without subsequent burning, the product being one which will not slake, but which when ground possesses hydraulic properties. hydraulic limes General 39. Definition and Classification. " The hydraulic limes include all those cementing materials (made by burning silicious or argillaceous limestones) whose clinker after calcination contains so large a percentage * E. B. Mathews and J. S. Grasty, " The Limestones of Maryland." HYDRAULIC LIME AND GRAPPIER CEMENTS 63 of lime silicate (with or without lime aluminates and ferrites) as to give hydraulic properties to the product, but which at the same time contain normally so much free lime that the mass of clinker will slake on the addition of water." * It will be seen that the hydraulic limes occupy an intermediate posi- tion between the common limes and the more complex cements. Accord- ing to their content of clayey matter they may be classified as " feebly " or " eminently hydraulic limes," the latter including " grappier cements." Both are feebly hydraulic as compared with good natural cements or with Portland cement. The feebly hydraulic limes, i.e., those whose cementation index lies between 0.30 and 0.70, are occasionally treated with sulphuric acid or plaster of Paris to produce a secondary product called " selenitic lime " which will be briefly mentioned hereinafter. Eminently hydraulic limes, i.e., those whose cementation index lies between 0.70 and 1.10, produce during manufacture a by-product called " grappiers," which are lumps of under-burned and over-burned material which do not slake. Finely ground grappiers constitute the cement called " grappier cement." 40. The Hydraulic Lime Industry. Uses of the Hydraulic Limes. The hydraulic lime industry is one of considerable importance in Europe, especially at Tiel, France, where the largest and best-known plants are located, and in England, where considerable quantities of feebly hydraulic limes are produced. In the United States no effort has ever been made to introduce hydraulic lime manufacture, not because of any lack of the proper raw materials, but because of the abundance of materials suitable for the manufacture of natural cements and Portland cement, with which the relatively feebly hydrauHc and weak hydraulic limes and grappier cements caimot compete as a structural material. A number of hydraulic limes and grappier cements possess, how- ever, a certain value for architectural purposes by reason of their light color, and the further fact that they are " non-staining cements," i.e., they contain so slight an amount of iron and soluble salts that they do not stain masonry. A considerable amount of hydraulic lime and grap- pier cement is therefore annually imported and used for purposes of architectural decoration. Manufacture of Htdkaulic Limes 41. Hydraulic Limestones. The ideal hydraulic limestone rock should have such a composition that, after all the silica has combined with lime during calcination, suflBcient free lime remains to disintegrate * Eckel, " Cements, Limes and Plasters,'' p. 172. 64 MATERIALS OB^ CONSTRUCTION the kiln product by the expansive force set up when it is slaked. In practice, however, not all the silica will combine with lime, and m order to avoid an excess of uncombined lime it is necessary that the content of the rock be lower than is theoretically desirable. The inevitable replacement of a part of the siUca by alumina and iron oxide is also a consideration which causes departure in practice from theory, because these latter constituents act as fluxes, facihtating the formation of lime silicate, and also themselves combine with lime to form aluminates and ferrites of lime. The limestones actually used in the manufacture of eminently hydraulic limes usually contain from 40 to 50 per cent of lime and mag- nesia (the latter rarely exceeding 1.0 per cent), 7 to 17 per cent of siUca, and usually not to exceed 1.0 per cent of alumina and iron oxide. (Occa- sionally the latter constituents amount to as much as 10.0 per cent.) Feebly hydraulic limes are usually made from limestones contain- ing from 45 to 50 per cent of lime and magnesia, 4 to 10 per cent of silica and 4 to 8 per cent of alumina and iron oxide. The difference in composition of limestones suitable for the manu- facture of eminently hydraulic limes, and those which produce feebly hydrauhc limes, is evidently a matter of higher silicon content and lower alumina and iron oxide content in the limestones which yield the product of greatest hydrauUcity. The percentage of total carbonates is about 70 to 80 per cent for eminently hydraulic limestones, and 80 to 90 per cent for feebly hydraulic limestones. 42. Calcination. The burning of hydraulic limes is accomplished in continuous kilns like common lime kilns. The operations involved in the process of calcination are, in fact, praXitically identical for common lime and for hydraulic lime except that the temperature required is somewhat higher in the latter case. (Not less than 1000° C.) The proper temperature of calcination is directly dependent upon the composition of the rock and therefore bears a direct relation to the value of the cementa- tion, index. If the cementation index is less than 0.70, as is the case with the feebly hydraulic limes, it is practically impossible to avoid an excess of free lime, and all limes of this class do contain a large amount of uncom- bined Ume. As the value of the index becomes more than 0.70 there is less and less danger of an excess of free lime in the product until with values approaching 1.00 the difficulty is to obtain a product with suf- ficient free lime to slake properly and produce a finely disintegrated product. 43. Slaking and Subsequent Treatment. The theory of slaking of hydraulic lime differs from the slaking of common quicklime in no respect except that in the former case the quicklime, which will slake, is intimately HYDRAULIC LIME AND GRAPPIER CEMENTS 65 associated in lumps with lime silicate, unburned or at least under-burned irinestbne,"and some^ aluihinate and ferrites, none of which can be" slaked. The expansion of the quicklime in slaking, however, breaks up the entire mass into a fine powder which will consist principally of lime silicate together with one-fourth to one-third as much hydrated lime. The slaking of hydraulic lime was at one time commonly done by the purchaser upon the work. It is now, however, practically the uni- versal practice to slake the lime at the place of manufacture. The lump lime as drawn from the kiln is spread out in thin layers and sprinkled lightly with water. It is then shoveled into heaps or bins where it is allowed to remain for about ten days till the slaking is completed by the aid of the steam generated. In order that the product may be a fine, dry powder, the slaking must be done carefully and with just the right amount of water. A certain proportion of the kiln product either does not contain suf- ficient lime or is not sufficiently burned to be slaked upon the addition of water, and consequently remains as hard lumps called " grappiers." The slaked material is therefore passed over screens of about 50 meshes per linear inch, which reject all of the larger particles. This rejected material (grappiers), is valuable or not according to whether it con- sists principally of lumps of Ume silicate on the one hand, or worthless raw or unburned limestone on the other hand. As a rule all of the grappiers are finely ground under millstones (no selection of good and bad being practicable), and a certain proportion is added to the lime, thereby in- creasing its hydraulicity in proportion to the amount of lime silicate present. The ground grappiers are also separately marketed as a special cement known as " grappier cement." It is practicable to market the ground grappiers as grappier cement only when the proportion of lime silicate is high, producing therefore a cement which approaches natural cement in its properties, with the added advantage of being a " non-staining," very light-colored cement. - -Selenitic limes are made from English feebly hydraulic limes by- the addition of not more than 5 per cent of ground plaster of Paris to calcined hydraulic lime. The mixture is subsequently finely ground. Selenitic lime is sometimes called " Scott's cement " after the inventor and patentee of the process. Properties of Hydraulic Limes 44. Composition. The composition of various hydraulic limes is indicated by the following table: 66 MATEEIALS OF CONSTEUCTION ANALYSES -OF HYDRAULIC LIMES SiOj. Al20j+Fe20j. CaO +MgO. Num. Ave. Ave 10.99 16.05 7.60 I 24.40 34.58 14.60 22-26 25.68 27.50 22.15 Feebly Hydraulic Limes.' 8.84 79.55 12.77 84.55 5.14 72.98 Eminently Hydraulic Limes.^ 10.90 60.03 16.97 65.29 3.15 41.93 9-13 47-57 Grappier Cement.' 6.11 59.13 11.40 63.84 2.90 53.66 5 Max Min Ave Max 43 Min Usual limits 6 • Summary of S analyses quoted by Ecliel, " Cements, Limes, and Plasters." 2 Summary of 43 analyses quoted by Tetmajer, " Methode und Resulte der Priif ung Hydrau- lischen Bindemittel," pp. 140-143. (The author has omitted analyses Nos. 30, 39, 40, and 41, which are abnormal, from the summary.) ■ Summary of six analyses quoted by Candlot, " Cementes et Chaux Hydraullques," p. 178. 45. Physical Properties. The physical properties of hydraulic limes are indicated by the following data from the works of Tetmajer and Cand- lot previously cited. TETMAJER'S TESTS OF EMINENTLY HYDHAULIC LIMES' (Average op 22 Limes) Specific Gravity. Weight, Lbs. per Cu. Ft. Ignition Loss. Setting Time. Fineness, per cent Retained on Sieve of Meshes per Lin. In. Loose. Well Shal^en. Per cent. Initial. Hrs. Min. Final. Hrs. Min. 76 180 Ave Max Min.... 2.766 2.97 2.59 49.0 55.2 39.7 71.5 86.4 52.0 12.14 19.14 5.63 3 :02 11 :00 :03 54 : 18 144 :00 7 :00 13.39 21.8 4.0 CANDLOT'S TESTS OF GRAPPIER CEMENTS.' (Average op 7 Cements) Ave Max . . . Min. . . 53.7 57.4 47.5 12 15 10 24 00 00 4,14 16.00 0.0 17.57 36.0 9.5 1 Tetmajer, pp. 376-378. » Candlot, p. 179, HYDRAULIC LIME AND GRAPPIEE CEMENTS 67 iS'ieo ^160 m uo a 130 t uo i 100 Pi go a 10 I 60 " 60 V a 40 I 30 H 20 10 TCNSILE GTR&NQTM OF EEEUCl' tlYpriAULtC LIMES AND SELENITIC LIMES GHANT'8 TESTS ENGLISH LIMES 1 i G SA> A >JD LIA 8L IM :S , -^ Si S^LEJIITjC .rKES *X > S^ ^ «-.. t ^ ^^"s..:^ < i "V lajw;^ * k Ai^'* Si'i <« ?C '« \^r,N\^ % nv L_ , ■Jli -^ JiS . ut FroportloDS of Mixture, Fig. 26. 46. Tensile and Compressive Strength. The tensile and compressive strength of feebly hydraulic lime mortars are indicated by the curves of Figs. 26 and 27 which have been plotted from tests made by John Grant in England about 1880.* The gray lime mentioned is a typical English feebly hydraulic lime; the lias limes are somewhat more silicious limes, and there- fore less feebly hydraulic ; and the selenitic hmes noted are ones prepared from the same gray and has limes. Each of the curves of Fig. 26 repre- sents the average of five tests of tensile specimens whose minimum cross- section was made 1| inches square. The curves of Fig. 27 represent the average of 10 compressive tests of 6-inch mortar cubes, the aggregate used being a sand and gravel mixture. These tests were made long before meth- ods of testing had been standardized, even for cement, and conse- quently the values shown are of little value except that they show that tensile specimens stored in water are about 60 per cent stronger than similar specimens stored in air, and that the addition of lime sulphate to the calcined limes, thus producing selentic limes, practically doubles the strength. (All of the tests of Mr. Grant were made at the age of one year.) At the first glance the strength values shown by these curves appear * Minutes of Proceedings of Civil Engineers, Vol. 62, p. 165. Sioo- sS76 "360- &326- . 300 - '&226 - g SOO- S176- 1 1 G lA f * ID LIA «L IM =8 \ 1. SE LE m c LiM E8 V \ f ^ r^ti \. J^ > % \ V \ 1^ V? . N"* ^ >-. s j^ ^ 'h Sm ^ 1 'J. ,1 I iai\ I m ff ^> Tl ^ f \ ' ^ % >M. ^ <" Ai r) ^ _ mo Proportions ol Mixture. Fig. 27. 68 MATERIALS OF CONSTRUCTION excessively low. The mortar mixtures, especially those used in the compressive tests, are, however, extremely lean ones, and if the pro- portions were expressed by weight instead of by volume, as is the present practice, the mixtures would appear even leaner to one accustomed to theTiseof the terml': 3, 1 : 4, etc., meaning proportions by weight. The strength of eminently hydraulic limes in neat and mortar mix- tures is indicated by the curves of Fig. 28, which are based upon Tet- majer's tests of the limes whose physical properties have been tabulated in Art. 45 above. The author nas summarized Tetmajer's. results and expressed the values in Enghsh units in the following table: TETMAJER'S TESTS OF EMINENTLY HYDRAULIC LIMES Tension, Air Storage 1 ;3 Neat. .... 7 days. 28 days: 84 days. 210 days. 1 year. 2 years. 7 days. 28 days. 84 days. 210 days. 1 year. 2 years. Ave Max Min No. testa. . 100.7 236.1 43.7 (8) 184.5 385.5 82.5 (8) 211.9 36.7.0 76.8 (8) '400.2 .526.3 230.4 C3) 306.5 614.4 155.0 (8) 287.0 344.2 183.5 (4) 92.4 123.7 75.4 (6) 184.4 250.3 91.0 (6) 307.5 406.8 79.7 (6) 433.8 493.5 293.0 (5) 541.9 779.4 143.7 (6) 553.3 (1) Tension, Water Storage Ave Max Min No. testa. 56.8 108.5 210.9 304.3 316.2 325.6 66.0 129.8 235.6 308.6 352.2 95.3 174.9 342.8 465.1 463.7 409.6 108.1 179.2 345.6 421.0 439.5 32.7 46.9 102.4 166.4 213.3 268.8 29.9 82.5 153.6 253.6 300.1 (16) (22) (22) (17) (22) (12) (8) (8) (8) (7) (8) 460.8 (1) Compression, Air Storage Ave Max Min No. tests. 511.0 981.7 196.2 (8) 900.3 1779 . 398.3 (8) 1184.7 2201.5 632.9 (8) 2409.1 2903 . 3 1763.4 (3) 1834.0 3552.9 1016.9 (8) 1709.6 2200.5 1112.2 (4) 455.1 776.5 190.6 (6) 1026.2 1415.2 791.2 (6) 1884.0 2580.0 1103.6 (6) 2769 . 7 3590.0 1947.2 (5) 2918.3 4147.4 1514.7 (6) 3213.2 (1) Compression, Water Stoeage Ave.. Max. Min. 342.5 709.7 123.7 676.9 1109.4 300.1 1189.9 2511.5 587.4 1806.7 3648.5 933.0 No. tests (19) (20) (22) (17) (22) (12) (7) (8) 1924.9 3625.6 1110.7 2116.9 2988.4 1318.4 330.5 634.4 146.5 720.4 995.6 448.0 1543.1 2508.9 816.4 (8) 2579.6 3285.4 1479.3 (7) 3026.3 3642.6 19995.6 (8) 4083.4 (1) The curves indicate that the strength of hydraulic lime and hydraulic lime mortars is about one-half in the case of tensile strength, and about one-fourth in case of compressive strength, that of average Portland cement. The rate of gain in strength is very slow, however, and the maximum strength is not attained in a period short of one year. A com- parison of the neat strength with the strength of 1 : 3 mortar indicates that the sand-carrying capacity is excellent, the mortar strength averag- HYDRAULIC LIME AND GEAPPIER CEMENTS 69 M I I I I I III I II I STRENGTH OF EMINENTLY HYCRAULIC LIMES Tetmajer's Tests (French and Qarmap Emipently Hydraulic Limee)'^ ing from two-thirds to three-fourths the neat strength. Specimens stored in air appear to gain in strength somewhat more rapidly at first, as a rule, than do specimens stored in water after three days in air. The subsequent gain in strength of the water-stored specimens is usually slightly more rapid, however, so that at the end of one year the strength of the two sets of specimens is nearly equal. One interesting feature of these tests is the apparent fact that hydrau- lic limes are only about five times as strong in compression as they are in tension. With Portland and nat- ural cements this ratio is at least 8, which indicates that the inferiority of hydraulic limes to Portland or even natural ce- ments is much more pronounced than would be indicated by a comparison of the tensile strengths of similar mixtures. (Relative tensile strengths are unfortunately, and quite without justification, often made the sole basis of comparison of cements.) The eminently hydraulic limes are, however, immensely superior in strength to the feebly hydraulic limes and to ordinary non-hydraulic limes. 47. Hydraulic Limes in Construction. Hydraulic limes were at one time considerably used by engineers for various structural purposes. This was, however, principally before the advent of the greatly superior natural cements, and now the latter product is being almost entirely re- placed as an hydraulic cementing material by Portland cement. Hydraulic limes are not suitable for use on subaqueous construction, in spite of their designation; they are too slow setting to render their use on general construction work convenient and practical, and their comparative weakness makes - competition with natural and Portland cement absolutely impossible. Their only present value, as has been previously stated, is that of an architectural decorative material rathfer than a material for engineering construction. t MoS 3466789 10 U121231 TDays Age in Months Age in Months Fig. 28. CHAPTER V PUZZOLAN CEMENTS.* SLAG CEMENTS PUZZOLAN CEMENTS 48. General. The oldest known hydraulic mortars were undoubt- edly made by the incorporation of a volcanic tufa with slaked lime and sand. Thus were produced the cements extensively used by the Romans and other ancient peoples in the construction of many engineering struc- tures, a few of which remain to-day in a remarkable state of preservation. The activity of the volcanic material depends upon the presence of weakly acid silico-aluminates which combine more or less readily with lime hydrate at atmospheric temperatures. If the silica were more strongly acid its ability to combine with the base would be increased and the product would be one of superior hydraulic properties. This prin- ciple explains the fact that lightly burned clay will react with lime to form an hydrauUc cement, and a weakly hydraulic mortar has even been produced by simply mixing brick dust with hydrated lime.f 49. Definition of Puzzolans and Puzzolan Cement. The natural or artificial materials which contain a sufficiently large percentage of available silica to combine with lime hydrate and form a cement possess- ing hydraulic properties are known as " puzzolans " or " puzzuolanas." Puzzolan cements include all that class of hydraulic cementing mate- rials which are made by the incorporation of natural or artificial puzzolans with hydrated lime without subsequent calcination. Puzzolan cements have never attained any great commercial impor- tance in the United States. A small amount of this cement has been pro- duced for a great many years, however, and a considerably larger amount * The term "puzzolan," which is commonly applied to this class of cements by American authorities, is a corruption of the name " puzzuolana " which refers to the class of volcanic material first utilized as a hydraulic cementing material at the town of Puzzuoli, near Naples. Continental writers call these cements " puzzuolana " cements. t Bleininger, Geological Survey of Ohio, Fourth Series, Bull. 3, p. 25. 70 PUZZpLAN CEMENTS. SLAG CEMENTS 71 is produced in Europe. Nowhere does the importance of the industry compare at all with that of the Portland, or even the natural cement industry. The true puzzolan cement, often called " slag cement," made by incorporating blast-furnace slag with hydrated lime without subsequent burning, must not be confused with the Portland cement, which is made by finely pulverizing the clinker derived by the calcination of an inti- mate mixture of the same materials. This class of Portland cement is one of considerable commercial importance and will be later discussed under the head of Portland cement. 60. Natural Puzzolanic Materials. All natural puzzolanic materials of any commercial importance as cement materials are of the same general character and have a like geological origin. All are direct products of volcanic action and the commercial puzzolanic materials are invariably derived from deposits of more or less agglomerated volcanic ash. The natural puzzolanic materials and the principal artificial puzzolanic material, granulated blast-furnace slag, have their origin in practically the same processes. In both cases a more or less finely divided siliceous material is derived by the sudden cooling and ejection into air or immer- sion in water of a fused silico-aluminous material.* The only natural puzzolanic material of commercial importance as cement materials are " puzzuolana," " trass," " santorin " and " tuff" or "tufa." The latter material, as will be later noted, has been used in making only one brand of cement, and in that case it is not used in con- junction with lime, but is finely groujid and mixed with an equal amount of ordinary Portland cement and the resulting blend subsequently finely pulverized to produce a " tufa-Portland " cement, which is neither a true puzzolan nor a true Portland. Puzzuolana was first obtained in Italy by early Greek colonists who opened workings at the town Puzzuoli in the vicinity of Naples. The material was subsequently exploited by the Romans and is still an article of Italian commerce. Puzzuolana is also derived from several localities in southeastern France and from the Azores. The material derived from the two latter sources is now held in higher esteem than the original Italian puzzuolana. Trass is an earthy or consolidated pumiceous dust mixed with fragments of pumice and other minerals. It is derived from the valley of the Rhine in Prussia. Santorin is an unconsolidated volcanic ash derived from the island of Santorin or Thera in the southeastern part of the Grecian Archi- pelago. * Eckel, " Cements, Limes and Mortars," p. 633. 72 MATERIALS OF CONSTEUCTION The following average analyses of natural puzzolanic materials are quoted by Eckel : * AVERAGE ANALYSES 3F NATURAL PUZZOLANIC MATERIALS Puzzuo- lana, Italy. Puzzuo- lana,' France. Puzzuo- lana, Azores. Trass, Germany. Santorin. Average of All Puzzolans. Number of analyses. . SiOz 9 50. C3 15.55 14.41 7.30 1.96 0.63 5.O0 7 41.91 16.16 19.30 6.93 1.37 6.16 7.89 3 57.78 15.15 10.37 2.84 1.63 4.52 7.61 11 63.78 17.38 6.89 3.89 1.17 6.82 9.22 1 66.37 13.72 4.31 2.98 1.29 7.05 4.06 31 61.08 AI2O3 16.30 FejOa 11.13 CaO 6.46 MgO K2O, NajO.. H2O 1.50 6.21 V 64 If these analyses be compared with the analyses of hydraulic limes given on page 06, the necessity for the addition of lime to form an hydrau- lic cement will be at once apparent. The puzzolans contain twice as high percentages of silica, alumina, and iron oxide as. do the eminently hydraulic Umes, and yet contain only about one-seventh as much lime and magnesia. ^ 51. Mantifacture of Puzzolan Cements from Natural Materials. The preparation of puzzolan cements is a very simple mechanical process involving nothing more than the screening, mixing, and grinding of the two constituents employed. Most deposits of natural puzzolans are subject to great variation in the quahty of the material, therefore necessitating careful selection or sorting of the quarry product to' exclude inferior material. Screen- ing of the puzzolan is usually necessary in order to exclude undesirable adulterants and, in the case of the Itahan puzzuolana, roasting at mode- rate temperatures is occasionally resorted to in order to increase its hydraulic properties. Puzzolanic rock is, in .European practice, not usually mixed with hydrated lime at the place of manufacture, but is simply pulverized and marketed as a material to be incorporated with hydrated lime paste and sand where used in construction work. The proportions of the mix- tures used in mortars are not fixed, but the proportion of Ume usually amounts to from one-third to one-half the proportion of puzzolan, and the puzzolan cement thus formed may be combined with any proportion of sand up to about three times the proportion of cement. A less commonly followed practice, but one which should result in the production of a more homogeneous and satisfactory material, consists * Eckel, " Cements, Limes and Plasters," p. 638. PUZZOLAN CEMENTS. SLAG CEMENTS 73 in the grinding at the place of manufacture of a mixture of the puzzolanic rock and dry hydrated lime. Cement made by this method is, of course, ready for immediate use, and is simply mixed with water and sand on the work, just as is the case with hydrated lime or hydraulic lime. 52. Properties and Uses of Natural Puzzolan Cements. Puzzolan cements made from natural materials hold so unimportant a place among structural materials that little study of their physical characteristics has been made, and no data obtained in recent investigations of their properties have been given pubUcity. The tensile and compressive strengths of three puzzolan cements tested by Dr. Boehme * are presented by the curves of Fig. 29. All specimens were composed of one part by weight of puzzolan cement to three parts' by weight of sand. Similar tests were made on speci- mens hardened in air m JJJ and other specimens hardened in water, the strength being deter- minedfor eachatseven days and at twenty- . eight days. In every | case the specimens | ^ stored in water were found to be stronger than thecorresponding air-stored specimens. If these curves be compared with those of Fig. 28, which repre- sent the strengths of eminently hydraulic limes, only those curves which represent strengths of 1 : 3 mortars being compared, it appears that puzzolans hardened in air are about equal in tensile strength to eminently hydraulic limes similarly hardened, but that puzzolans hardened in water are about three times as strong in tension as the hydraulic limes hard- ened under the same conditions. In compressive strength the puzzolans appear to be from two to five times as strong as hydraulic limes. Average 1 : 3 Portland cement mortars excel . these puzzolans by at least 30 per cent in tensile strength, but show little or no advantage over the puzzolans in compressive strength at these early periods. Puzzolan cements produced from natural materials were in the days of the Roman Empire the only known hydraulic cementing 'materials, and the excellence of some of this cement is attested by the many sur- * Mittheilungen aus den K. technischen Versuchsanstalten zu Berlin, 1890, p. 256. Age In Days Age In Days Fig. 29. 74 MATERIALS OF CONSTRUCTION viving examples of early Roman construction in concrete. As an article of present day commerce, however, its importance is almost negligible, and in the United States the small amount of puzzolan cement used in construction is almost entirely derived from the treatment of blast fur- nace slag and therefore comes under the head of slag cement. SLAG CEMENTS 63. Definition. Slag cement may be defined to be an intimate mechan- ical mixture of granulated blast-furnace slag of suitable chemical com- position, with hydrated lime, the materials having been finely pulverized before, during, or after mixing. No calcining of the mixture is practiced, and the product is not to be confused with the true Portland cement which is produced by the calcination and subsequent pulverization of a properly proportioned mixture of blast-furnace slag and raw limestone. 64. Blast-furnace Slag. Required Composition and Physical Con- ditions. Blast-furnace slags, such as are suitable for use in slag cements, are fusible lime silicates derived as waste products from the operation of blast furnaces in smelting iron from its ores. The slag is formed for the most part by the combination in the furnace of the earthy part of the ore, i.e., the " gangue," with lime from limestone which is charged with the ore as a flux. Slags are produced in many metallurgical processes other than the reduction of iron ores. Only the latter process, however, is capable of producing the very basic slag required for cement manu- facture. The required composition of the slag according to American .practice is within the following approximate limits: CaO, 48 to 50 per cent; Si02, 32 to 35 per cent; AI2O3, 12 to 16 per cent, MgO, FeaOs, S, etc., 2 to 5 per cent. (The latter constituents are merely incidental, not required.) In the practice of many European plants the proportion of alumina is somewhat above this range and the proportion of silica somewhat lower. Slag, as derived from the blast furnace and allowed to cool slowly, is a very dense and hard material, and has, moreover, such a chemical con- stitution that even when pulverized it does not combine energetically with water, nor exhibit more than very feebly hydraulic properties. If the molten slag is cooled very rapidly, however, by the use of cold water, it becomes broken up into small porous particles which can be economic- ally handled by the pulverizing machinery. Two important chemical effects are also attained by the process of granulating the slag; the slag is rendered more hydraulic, thus provid- ing a stronger cement, and the content of undesirable sulphides always PUZZOLAN CEMENTS. SLAG CEMENTS 75 encountered in slags is reduced by the formation and passing off of hydro- gen disulphide. The remarkable effect of granulating the slag upon the hydraulic properties of cements made therefrom is illustrated by Figs. 30 and 31, which are based upon the studies of M. Prost.* Fig. 30 brings out the fact that the tensile strength at one year is more than tripled by granulation of the slag, while the compressive strength, as shown by Fig. 31, is increased from five to seven-fold by granulation. At ages under one year these ratios are still larger. — ~^ EFFECT OF GRANUUTING SLAG ON TENSILE STRENGTH OF SLAG CEMENT .680 1 M^ > joi ' — t^ — / » •I Eoi ^ ^UI. ted / ^ f t^ IE 55 - — «5aO / / " ■^ b» — s iia ^ / gtoo / g MO «200 s ^ s a r" S5 a lot 1 OB kv hr — " ^ 10 y * ai V ^ — ' {^ _ _ 210 Age in Bays Fig. 30. EFFECT OF GRANULATING SLAG ON COMPRESSIVE STRENGTH OF SLAG CEMENT iSosoo .■^ "eooo C-, [«» o- — — * £6600 3 A ^ ' r" > a 1800 O4400 ^-1000 / u xe- iroi luI tr^ — ( / SlE «a •o- _ -/ / A _ _ IllR «a __ _ _ •S3800 §3200 Sssoo - /- - — = = — - .- n^ SIS ,1 »1600 S ..< <»< g 800 „ m ^ - - - - ^ t_ =,» ^ L j_ _ 310 Age in Days Fig. 31. 55. Manufacture of Slag Cements. The process of manufacture of slag comments involves the following operations: granulation of the slag, drying the slag sand, preparation of the hydrated hme, proportion- ing the mixture, mixing and grinding. The granulation of the slag may be accomplished by a jet of high- pressure steam which the stream of slag encounters as it issues from the furnace. Air jets have also been similarly used, but both of these methods have, been generally replaced by an arrangement whereby the stream of slag from the furnace falls into a trough containing a rapidly flowing stream of cold water. This method is a very satisfactory one in that it produces a slag-sand which is quite porous and friable and which is in excellent condition from the point of view of chemical composition. * Frost, M. A., Annales des Mines, 8th series. Vol. 16, p. 158. f 6 MATEEIALS OF CONSTEUCTION It is charged with much moisture, however, and must therefore be dried before being mixed with the hydrated lime. The slag as it comes from the bins into which it has been discharged by the granulating device, carries from 15 to 45 per cent of water. The removal of this water is accomphshed through the agency of heat in either a rotary cylinder similar to those used in drying cement rock (see Fig. 41o), or in a vertical shaft dryer, into which the material is introduced at the top and dried by ascending hot furnace gases as it traverses a series of inclined baffle plates toward the base. Rotary dryers are used exclu- sively in the United States. The hydrated lime used in slag cement in American practice is a very pure high-calcium lime, the magnesia content being in fact only a few tenths of 1 per cent*. In a few instances, in European practice particularly, a semi-hydraulic lime is used whose content of sihca, alumina, and iron oxide combined may amount to as much as 6 or 8 per cent. The quicklime is commonly hydrated at the cement mill. The proportioning of the mixture of slag and lime is usually done according to some established standard found by experiment at each plant to produce the most satisfactory cement with the materials in hand. The standard proportions of slag and lime vary therefore at different plants from as low as 25 pounds of lime, to as high as 45 pounds of lime, for 100 pounds of slag sand. The proportioning might of course be done by appUcation of the cementation index as suggested by Eckel, the slag and the Ume having first been analyzed. Eckel states that the values of the cementation index found on examination to represent the propor- tions used at various plants ranges from 1.6 to 1.9, 1.7 being a very satisfactory value. (The method of calculating the mixture by use of the cementation index is explained in detail in Art. 86.) Practice as to method of grinding and mixing the slag and lime is subject to great variation. The best practice probably consists in grind- ing the slag sand in a ball mill or other type of grinder suitable for the intermediate reduction of cement materials, adding the hydrated lime, and accomplishing the mixing and final pulverization of the mixture simultaneously in a tube mill or other type of cement-finishing mill. Slag cements are normally more slowly setting than Portland cements and on this account are often treated with some class of puzzolanic mate- rial which will hasten the setting. Materials so used are burned clay, high-alumina slags, certain active forms of silica, and in the patented " Whiting " process, caustic soda, sodium chloride or potash. The amount of such materials added as an accelerator does not usually exceed about 3 per cent by weight, and unless the material is to qualify under the usual specifications for Portland cement no accelerator should be PUZZOLAN CEMENTS. SLAG CEMENTS 77 necessary. Any addition to the cement made for the purpose of regulating the set must of course be made prior to final pulverization. 56. Properties and Uses of Slag Cements. The usual range of com- position of American slag cements is indicated by the following summary of analyses of 5 different cements quoted by Eckel. Si02. Ab03+Fe203+FeO CaO. MgO.. s. COz+HjQ. 27.2-31.0 11.1-14.2 50.3-51.8 1.4-3.4 0.15-1.42 2.6-5.3 The specific gravity of slag cements usually ranges between 2.7 and 2.85, which fact affords a means of distinguishing slag cements from natural cements, which, rarely fall below 2.9, and Portland cements, which must under the standard specifications be not less than 3.1. The fineness of grinding practiced in making slag cements is about equal to that commonly attained in grinding Portland cements. From 1 to 8 percent is retained on a sieve of 100 meshes per lineal inch, and from 10 to 25 per cent on a 200-mesh sieve. As above noted, the setting of slag cements is somewhat slower than that of Portland cements . This is more particularly true of European ce- ments, however, as most American slag cements have had their set arti- ficially regulated by ad- ditions made prior to grinding, and they are in consequence able to meet the specifications for setting time of Portland cement, i.e., initial set occurs in not less than thirty minutes, . and final set in from one to ten hours. The tensile strength of American slag cements is shown by the curves of Fig. 32, which are based upon tests made by Professor W. K. Hatt at Purdue University.* It will be noted that the tensile strength of 1 : 1 mortar is nearly equal to that of the neat cement, whereas the 1 : 3 mortar develops only about 40 per cent of the neat strength. These curves * Engineering Record, Vol. 43, pp. 196-197. _ II J 1 II TENSILE STRENGTH OF SLAG CEMENTS W.K. Hatt Each result is the average of 30 teats H u'bUU t, 660 1-— __^ __ ~ ~ o 5 480 a «0 o 400 ~ ^ -^ v-i ^ -^ ri r" w-* " / / 1 si ^^ 1 i;3' - 1 o „ m I f ~ __, ;_ — "** ^ 80 / iO J _ _ _ _ _ Age in Days Fig. 32. 78 MATERIALS OF CONSTRUCTION W.K. Hatt Curves represent average of 5 tests of each of two brands of cement except that 90-day results represent tests of only one brand s t- 5600 rSiaoo 04000 A V^ ii-^.^, x" ■^ 1^ ^ k '- ■^ f ^*K 1 :=:5 Xv '^ 1 2800 /. -^ C^ ^ ,? ^ ^ r^ ^- ^- 1 — ' ^ " ^ ^ ^ ■'I- __ - ~J ^ / ^ ^ ^- -, i -^ f,3 e^ _ f" -. — _. ,. .. 1 100 n ^ -h =-< — 1,3 Al^ ^ 1 1 , average the results obtained in tests of four brands of slag cement. Crushed quartz sand between the 20 and the 30-mesh sieves was used. The results. of compressive tests of two of these same cements, also reported by Professor Hatt, are shown by the curves of Fig. 33. Com- parative curves are given for two series of specimens, one of which was allowed to harden in air, the other in water. From a comparison of these curves it appears that slag cement mortars harden and gain in strength most rapidly when stored in water. More extensive investigation of this point revealed the fact that the tensile strength of neat slag cement specimens hard- ened in air averages about 77 per cent of the strength of similar specimens hardened in water. For 1 : 1 mor- tars this ratio averaged 78 per cent, and for 1 : 3 mortars its value was 74 per cent. Comparing compressive strengths, these ratios were found to average 81 per cent for neat specimens, 90 per cent for 1 : 1 mortar, and 96 per cent for 1 : 3 mortar. The tensile strengths of these cements and mortars are little more than half the average strengths of Portland cements and mortars, and the compressive strengths are about one-third the strength of Portland. It will be noted by comparing Figs. 32 and 33 with Figs. 30 and 31 that these American slag cements fall considerably below the strength values found by Prost using European slag cements. The uses to which slag, cement may be put are usually limited to unimportant structures or to work requiring large masses of concrete masonry where weight and bulk are more important than great strength. It IS seldom used on structures above the foundations, and never used on comparatively hght reinforced concrete construction. The industry IS one of declining importance, as is shown by the fact that the annual production of slag cement, which reached its highest point (557 252 barrels) in 1907, had fallen to 107,313 barrels in 1913, only four pllnts being then in operation. For every pound of slag cement produced or used in the Umted States, more than 1000 pounds of Portland cement is made and used, 60 Age in Days Fig. 33. CHAPTER VI NATURAL CEMENTS GENERAL 57. Definition. Distinction between Natural and Portland Cements. Natural cements have been classified above (Art. 28) as cements which are made by burning distinctly argillaceous limestones at a relatively high temperature, the product being one which will not slake with water, but which possesses the hydraulic properties after grinding. It .now becomes necessary to define natural cements in terms of such restricted meaning that there may exist no possibility of confusing natural cements with any other class of cementitious materials. Natural cement may be defined as the finely pulverized product resulting from the calcination of an argillaceous limestone at a tempera- ture sufficient to drive off the carbon dioxide gas and also decompose the clay and effect the formation of alumihates, ferrites and silicates. The definition given by the American Society for Testing Materials differs from the above definition in that it prescribes that the calcination shall be effected " at a temperature only sufficient to drive off the car- bonic-acid gas." The latter definition means the exclusion of practically every natural cement made in this country, for only a very few natural cements are or can be successfully burned at a temperature not exceeding that necessary to drive off the, carbonic-acid gas. The distinctions between natural and Portland cements may be sum- marized as follows : Natural Cements. Portland Cement. Raw material Calcination Temperature . Chemical Composition . . . Color Specific gravity Rate of setting Natural argillo-calcareous rock Low Variable, not under control Usually yellow to brown 2,7to3.1 Normally rapid Low Artificial argillo-calcareous mixture Relatively high Controllable within narrow limits Usually blue-gray , . 3.1 to3.2 Relatively slow Relatively high Strength 79 80 MATERIALS OF CONSTRUCTION 58. Natural Cement as a Structural Material. Natural cement is used structurally as the cementing ingredient of concretes and, in com- bination with sand, as a mortar for laying brick and stone masonry, and as an outside wall plaster. A certain amount is also used as an addition to lime mortars for the purpose of increasing the strength thereof. In all of these applications it comes into direct competition with Portland cement, with respect to which it often suffers by comparison. In con- sequence of a growing. distrust of natural cement and the improvement in the quality of Portland cement produced during the last fifteen years, accompanied by the decrease in cost of the latter, natural cement has found itself gradually being crowded from the field of construction mate- rials. Because of the declining importance of the natural cement industry and the greatly reduced extent to which its use is permitted on masonry construction work, far less detailed consideration will herein be given to its manufacture and properties than will be accorded Portland cement. MANUFACTURE OF NATURAL CEMENTS 59. Natural Cement Rocks. Natural cements are invariably made by the calcination of a natural clayey limestone carrying from 13 to 35 per cent of clayey material, 10 to 20 per cent of the clayey material being silica, and the balance alumma and iron oxide. The hydraulic properties of the cement are entirely due to the presence of this clayey material. The calcium carbonate of the limestone may be and very commonly is replaced to a considerable degree by magnesium carbonate, resulting in the replacement of lime by magnesia to a corresponding degree in the manufactured product. This latter fact is without great significance since, so far as the hydraulic properties of the natural cement are concerned, lime and magnesia may be regarded as almost exactly interchangeable. Argillaceous limestones of the composition required for the manu- facture of natural cement are widely distributed, there being in fact hardly a state in the United States where such rock is not found. In spite of this wide geographical distribution, however, there are only comparatively few districts where the natural cement industry has ever become commercially successful. This fact is due to the necessity for certam commercial advantages dependent upon the character and loca- tion of the quarries. Among these requisites may be mentioned a rock of at least fairly uniform composition, a favorable location of the rock beds for cheap extraction of the cement rock, cheap fuel, a good local market for the product, and good transportation facilities. ' NATUEAL CEMENTS 81 60. Theory of Calcination. From the composition of the raw mate- rial given above it will be seen that the rock as it is charged into" the kiln consists essentially of lime and magnesium carbonate with a more or less definite percentage of clayey matter. The chemical changes that take place during calcination may be briefly mentioned as follows: Water mechanically held by the rock is driven off at temperatures little above 100° C; at a temperature of about 750° to 800° C. mag- nesium carbonate is dissociated, the carbon dioxide being driven off, leaving the magnesia; at a temperature of 900° C. or less the lime car- bonate is similarly dissociated, leaving lime; at a temperature of 900° to 1000° C. the clay is decomposed and the formation of aluminates and ferrites of hme and magnesia is effected; lastly, if the temperature is carried to 1100° to 1300° C, as it usually is, siUcates of lime and magnesia are formed. If the rock could be perfectly burned the loss in burning (the difference in weight of raw material and marketable product) would correspond exactly to the percentage of carbon dioxide -[-water in the rock. Practically, the losses will exceed this amount by about 25 per cent on account of under-burning and over-burning of a portion of the material. 61. Practice of Calcination. The Kiln. The type of kiln almost exclusively used in the United States' natural cement industry is of the continuous vertical mixed feed type, the rock and fuel being either mixed or charged in alternate layers. For present purposes a description of one t3T)e of kiln of modern design will be given as typical of the kilns used in present Ameri- can practice. The Campbell kiln (Fig. 34) consists of an inner cylindrical kiln of masonry, lined with firebrick and enclosed by a sheet^steel cylinder separated from the masonry by a thick layer of clay. The interior diameter of the kiln is 11 feet from a point about 9 feet above the pot to a point about 8 feet below the top. The kiln gradually narrows below this zone to a diameter of about 6 feet 7 inches at the top of the FRONT ELEVATION SECTION Fig. 34. — Campbell Kiln for Natural Cement. 82 MATERIALS OF CONSTRUCTION pot, and above this zone to a diameter of about 9 feet. The height of the kiln above the top of the pot is about 28 feet. The pot is an open grating of iron in the form of an in- verted cone about 3 feet 6 inches in diameter at the bottom where the calcined mate- rial is withdrawn. The capacity of such a kiln is from 125 to 150 barrels (265 poimds) per day. Other types of kiln in common use in America differ from the Campbell kiln in only one essential detail— the replacement of the iron kUn-pot by a masonry pot. The Campbell is somewhat more economical to build and more convenient in operation. The Fuel. The fuel used in natural-cement burning is bituminous coal. The fuel consumption varies greatly, because differences in com- position of the rock cause variation in the required calcination tem- perature. The average fuel consumption is about 12 pounds of coal per 100 pounds of cement produced, the variation being from a minimum of about 6.5 pounds to a maximum of 18 pounds. 62. The Clinker. The output of a natural-cement kiln includes three classes of material, hard-burned clinker, soft porous moderately- burned material, and practically unburned material. The existence of these three classes of calcined material is due to three facts: variation in the composition of the rock charged, variation in the management of the kiln, and variation in the degree of heat to which the material in the kiln has been subjected. The recognition of the presence of these three classes of material is necessary, because with some rocks the best cement is derived from the soft moderately burned material, and in many cases the best cement is a mixture of both materials. Unburned material is valueless. 63. Free Lime. With cement rock high in lime and magnesia the possible range of temperature for proper calcination is shortened, with con- sequent greater possibihty of dangerously high percentages of free lime. The presence of this free lime may be neutraUzed by some method of slaking the lime. This is commonly done by sprinkling or steaming the unground clinker, and in some cases by simply aerating the clinker. This expedient decreases grinding costs, since the slaking of the lime will to some extent disintegrate the clinker. 64. Grinding and Packing. In the early days of the manufacture of natural cement in America the universal practice was roughly to crush the cUnker in some simple form of crusher, followed by fine grind- ing between millstones. This equipment has now been largely super- seded by grinders such as the Sturtevant " rock-emery " mill (Fig. 35), the Cummings grinder (Fig. 36), and various types of grinders and pul- verizers commonly used in Portland cement manufacture, such as the gyratory crusher, the Griffin mill, the Huntington mill, the kominuter, the ball mill, the tube mill, the Fuller-Lehigh mill, etc. NATURAL CEMENTS 83 Fig. 35. — Sturtevant Rock-emery Mill. In the Sturtevant " rock-emery " mill the ordinary mill stones are replaced by " rock-emery " stones, which consist of a cylindrical shell of steel enclosing the abra- sive. The center of this stone is simply a disc of buhrstone surrounded by a zone set with slabs of " rock-emery " cemented by metal poured while molten. Radial strips of buhr- stone are set to continue the furrows of the central buhrstone to the rim of the wheel. The Cummings grinder crushes material between two vertical chilled-iron discs, one of whichis stationary, while the other is revolved at highspeed. The faces of both discs are cut into a series of bands and fur- rows to facilitate the rapid dis- tegration of the materials. The construction and operation of the mills com- monly used in the Portland cement industry will be illustrated and discussed in Chapter VII, and therefore need not be further mentioned here. The packing, of natural cement differs in no respect from the packing of Portland cement, to be discussed later. The standard of the Americal Society for Testing Ma- terials prescribes packing in bags of 94 pounds net weight, three bags to the barrel. Practice varies in respect to packing weights, however; some companies have adopted 300 pounds net weight for their standard barrel, some 280 pounds, and still others 265 pounds. 65. Manufacturing Costs. Manufacturing costs in the natural cement industry will vary greatly, depending upon the nature and avail- ability of raw material, the cost of fuel, the mechanical equipment of the plant, and the cost of labor. In general the manufacturing costs will Fig. 36. — Cummings Grinder. 84 MATERIALS OF CONSTRUCTION vary between about 20 cents per barrel under most favorable circum- stances, and 50 cents per barrel under very unfavorable conditions. PROPERTIES AND USES OF NATURAL CEMENTS 66. Chemical Composition. Constitution. The average composi- tion of various natural cements is indicated by the following summary, which shows the range in composition found upon averaging a number of analyses of each of six well-known American natural cements. (Each figure represents an average for one brand, and individual samples of each brand will be found to be outside the average limits stated.*) SiOj. AbOi. FejQ,. CaO. MgO. 22,3-29.0% 5.2-8.8% 1.4-3.2% 31.0-57.6% 1.4-21.5% In addition to these principal constituents all natural cements contain varying small amounts of alkalies, sulphur trioxide, carbon dioxide, and water. It will be apparent from the above summary that the composition of natural cements is extremely variable, and since wide variations in composition are accompanied by great differences in mechanical proper- ties, this circumstances is in no small degree responsible for the wide- spread distrust of natural cements. The typical constitution of natural cements does not differ materially from that of Portland cements, and discussion of the matter will therefore be taken up in Chapter VII. 67. Specific Gravity. The specific gravity of natural cements is •as above noted sUghtly below that of Portland cements. The standard specifications of the Am. Soc. for Testing Materials make no stipulation concerning specific gravity. Eckel quotes many tests of the more prom- inent brands which show a range of from 2.70 to 3.17 with an average of 2.96. 68. Time of Setting. Natural cements normally are quick setting as compared with Portland cement. The standard specifications prescribe that the initial set shall take place in not less than ten minutes, and hard set in not less than thirty minutes or more than three hours. The time of setting of natural cemettts is materially affected by aera- * These data are abstracted from a large number of analyses of individual cements of SIX different brands, quoted by Eckel, " Cements, Limes and Plasters," pp. 204-213. NATURAL CEMENTS 85 tion. Sabin * found with ten natural cements of widely varying rates of setting, nineteen days aeration retarded the initial set on the average about 43 per cent, the cements naturally least quickly setting being retarded the most. The final set after the same aeration was found to have been retarded on the average 3.6 per cent, the more quickly set- ting cements being retarded the most. The addition of gypsum or plaster of Paris also has a very marked effect in retarding the set of natural cements, the degree of retardation with a given percentage of gypsum being largely dependent upon the chem- ical composition of the particular cement used. Sabin experimented with two brands of natural cement to which varsang percentages of gypsum were added. Both brands showed a maximum retardation in initial set (180 per cent and 225 per cent respectively) when 2 per cent gypsum had been added, while the maximum retardation in final set ,in each case (24 per cent and 275 per cent respectively) was experienced when 3 per cent of gypsum was employed. 69. Fineness. The important bearing of fineness of grinding upon the strength and other properties of cement was formerly overlooked for the most part by manufacturers of natural cement. In recent years, however, they have been compelled to look about for means of improving their product in order to be able to compete at all with Portland cements. With the awakening has come the more general adoption of modern types of grinding machinery and a greatly improved degree of fineness of grinding. Formerly, the grinding was rarely carried further than was necessary to pass 95 per cent of the material through a 30-mesh sieve, which was, indeed, all that most specifications required. Now the general practice is to meet and even exceed the requirements of the standard specifications of the Am. Soc. for Testing Materials which permit a maximum residue of 30 per cent on the 200-mesh sieve and 10 per cent on the 100-mesh sieve. 70. Tensile Strength. In tensile strength natural cements vary quite as much as they do in other physical properties. This variation is found not only in comparing cements from different localities, but even in comparing samples taken at different times from the output of any one locality. The only general statement that may be made concern- ing their strength is that natural cements rarely show more than half the tensile strength of Portland cement of the same age. This is true not only of neat cement, but also of mortars of all proportions. Tests made by Sabin f upon . ten representative brands of natural cement have been averaged and the curves of Fig. 37 plotted therefrom. One notable * Sabin, " Report, Chief of Engineers," 1895, p. 2937. t Sabin, loc. cit., p. 2937. 86 MATERIALS OF CONSTRUCTION feature of these tests is the fact that practically no retrogression is en- countered even after two years, and many other tests of natural cements establish the fact that natural cements do not normally show retrogression. This is of great signifi- cance in view of the fact that the ma- jority of Portland cements show retro- gression before the six-month period has expired, and practi- cally all Portlands re- trogress between the six-months' and the eight-months' periods. The standard specifications of the Am. Soc. for Testing Materials fix the following minimum requirements for tensile strength: ~ — — — — ^ ~ .J __ ^- alOO 7 rrf ^ H ffnr tnT _,- ^'^'S A A >■ ^ "- ~~ ^ /■ y^ L _,_ S ^ nnn J /" — "~ " ■ g §150 ') r / r- TENSILE STRENGTH OF NATURAL CEMENT Average of 10 brands CSabin) 1 / V 1 2 3 6 12 mpresaive Strength Poundfl per Sq.In, Ape n M< ntba .^ — - "' M \ pt** ,- — -" ,^ "' COMPRESSIVE STRENGTH OF NATURAL CEMENTS Average of 8-9 brands (1 brand only at 90 days) - _-:i c i- ^ ,- "' 40 tat ^ ' ■" o- " 1 RichardBoa _ Age in Days Fig. 37. Neat Cement, Lbs. per Sq. In. 1 : 3 Standard Mortar. Lbs. per Sq. In. 24 hrs. in moist air 24 hrs. in moist air, 6 days in water. . 24 hrs. in most air, 27 days in water . 50 125 It is further stipulated that the cement shall show no retrogression in strength within the periods specified. 71. Compressive Strength. The above remarks concerning the vari- ability of tensile strength of natural cements apply equally well to com- pressive strength. The average compressive strengths of a number of natural cements tested by Clifford Richardson are shown by the curves of Fig. 37.* In general it may be stated that the compressive strength of natural cement, neat or mortars, rarely exceeds one-third that of Portland cement in similar mixtures. The ratio of compressive strength to tensile strength for cements and mortars is in the neighborhood of 5 to 6 for neat cement and 3.5 to 5 for the mortars. These values are scarcely more than one-half the ratios found to obtain with Portland cement, f * Brickbuilder, Vol. 6, p. 253. t See Report, Chief of Engineers, U. S. A., 1896, p. 2872. NATUEAL CEMENTS 87 The standard specifications make no mention of compressive strength. 72. Modulus of Elasticity. The ratio of stress to strain for cement and mortars cannot be considered exactly constant even for small loads. The age of the mortar when tested also has a considerable effect upon the results obtained, the progressive change in the value of the stress- strain ratio with increasing loads being considerably less marked as the age increases. The following tabulation giving a summary of tests made upon 12-inch cubes at the Watertown Arsenal * shows the progressive change in the modulus with increased range of load: MODULUS OF ELASTICITY, LBS. PER SQ. IN. Range of Load. 100-600 100-1000 100-2000 1 : 1 mortar 1 : 2 mortar 1 : 3 mortar 2,147,000 1,324,000 955,000 . 1,709,000 1,266,000 804,000 1,081,000 1,042,000 (In using the term " modulus of elasticity " referring to the behavior of an imperfectly elastic material under load, we mean simply the quo- tient obtained by dividing any stress increment by the strain increment which accompanies it. This is admittedly an incorrect use of the term, but is nevertheless a common practice.) 73. Where Natural Cement May be Used. When economy is effected thereby, natural cement may be substituted for Portland cement in mortars and concrete for dry heavy foundations where the stresses en- countered will never be high and will not be imposed for several months after the placing of the concrete, for backing or filling in massive masonry in dry situations where weight and mass are more es sential than strength, for lajong brick and stone masonry subjected only to fight loads and not exposed, for pavement foundations, for sidewalks, etc. It should not be used in exposed situations, should not be placed under water, it is unsuited for use in marine construction, and should not be used in build- ing construction above the foundation. 74. Production and Value of Natural Cement. Statistics of the natural cement industry in the United States are presented in the table below which has been taken from " Mineral Resources of the United States." As a matter of convenience, and to afford direct comparisons the statistics of the Portland cement industry and the puzzolan cement industry are also presented at this point. * Tests of Metals, 1902. 88 MATERIALS OF CONSTRUCTION PRODUCTION OF CEMENT IN THE UNITED STATES Natural. Portland. Fuzzolan. Year. Barrels Ave Price Barrels Ave. Price Barrels Ave Price 265 Lbs. per Bbl. 380 Lbs. per Bbl. 330 Lbs. per Bbl. 1870-1880 22,000,000 1880 2,030,000 42,000? 3.00? 1885 4,000,000 0.80 150,000 1.95 1890 7,082,204 0.69 335,500 2.09 1896 7,741,077 990,324 1.60 1896 7,970,450 1,543,023 1.57 12,265 1897 8,311,688 2,677,775 1.61 48,329 1898 8,418,924 ' 0.'47' 3,692,284 1.62 150,895 1.50 1899 9,868,179 0.52 5,652,266 1.43 336,000 1.47 1900 8,383,519 0.45 8,482,020 1.09 446,609 1.27 1901 7,084,823 0.43 12,711,225 0.99 272,689 0.73 1902 8,044,305 0.50 17.230,644 1.21 478,555 0.81 1903 7,030,271 0.50 22,342,973 1.24 625,896 1.03 1904 4,866,331 0.50 26,505,881 0.88 303,045 0.75 1905 4,473,049 0.54 35,246,812 0.94 382,447 0.71 1906 4,055,797 0.60 46,463,424 1.13 • 481,224 0.86 1907 2,887,700 0.51 48,785,390 1.11 657,262 0.79 1908 1,686,862 0.49 51,072,612 0.85 151,451 0.63 1909 1,537,638 0.42 64,991,431 0.81 160,646 0.62 1910 1,139,239 0.42 76,549,951 0.89 95,951 0.66 1911 926.091 0.41 78,528,637 0.84 93,230 0.83 1912 821,231 0.45 82,438,096 0.81 91,864 0.84' . 1913 744,658 0.46 92,097,131 1.01 107,313 0.91 (Prices stated are f.o.b. at the works and do not include cost of barrels or bags.) 75. Status of the Industry. Reference to the foregoing table will show that natural cement reached its maximum output in 1899, 9,868,179 barrels. Beginning with 1904, the industry has shown marked and con- tinuous decline in production each year, and its production for 1913, 744,658 barrels, is the lowest on record since 1870. At the same time the value of the product has depreciated even more rapidly than the quantity produced, the average net price being 46 cents per barrel in 1913, as compared with 52 cents in 1899, and 60 cents in 1906. _ One of the primary reasons for the decHne of the natural cement industry IS suggested by a comparison of the production of natural cement during the last twenty years with the production of Portland cement for the same period. It will be seen that the period of decline for the natural cenient industry is practically coincident with a period of phenom- enal growth m the Portland cement industry, this growth being accom- Hn «i^ ''iQio^ in average price of Portland cement from $2.09 in 1890 to $0.81 in 1912 and 1.01 in 1913. NATURAL CEMENTS 89 Taking into account the fact that the average natural cement barrel contains 265 pounds, as compared with 380 pounds for Portland cement, it appears that the relative cost of natural and Portland cement is less in favor of the former from the standpoint of the consumer than would at first appear, the cost of natural being 75 per cent of the cost of the Portland pound for pound. The relative cost of natural and Portland cement mortars and concretes is also dependent upon the relative pro- portion of cement necessary in each case. Since the average natural cement possesses not more than half the strength of the Portland cement in neat or mortar mixtures, and an even lower ratio for concretes, it will be evident that the mixture required will be much richer to attain a given strength if natural cement be used. In general, the cost of the additional amount of natural cement required will, therefore, more than offset the advantage due to difference in unit price. One consideration favoring the selection of natural cement in preference to Portland cement for cer- tain classes or mortar should be mentioned in this connection. A mortar of 1 : 2 mix natural cement may be required, for instance, in the construc- tion of certain classes of brick or even stone masonry where mixture of 1 : 4 Portland cement will suffice. The advantage in cost will perhaps be slight either way, the richer natural cement mortar will, however, be preferable to the Portland cement mortar with its heavy burden of sand, because the former will work much more easily under the trowel and thereby enable the mason to increase materially the speed with which the work can be done. In the above discussion we have neglected to take account of the cost of transportation of the cement, always a large factor in total cost. It is very possible that this factor may entirely change the relative cost of the two classes of cement in a given district. Frequently it has hap- pened that Portland cement has required long transportation to a market in the vicinity of natural cement plants. Under these circumstances competition is impossible on the class of work to which natural cement is adapted. This consideration accounts also for the fact that for many years the market for each natural cement has been a local one. CHAPTER VII PORTLAND CEMENT GENERAL 76. Historical. The use of a cementitious material composed of calcareous and argillaceous substances so far antedates authentic his- tory that we have no knowledge when or by whom it was discovered or first employed. It was used by the ancient Egyptians in the con- struction of portions of the pyramids, which have endured for more than 4000 years. The Romans constructed many aqueducts, walls, and build- ings of cement concrete, some of which are still existent, though they are in some instances known to have been built as early as the third centuiy before the Christian Era. The pools of King Solomon near Jerusalem were built of concrete and still furnish water for the city. The Colosseum of Rome has concrete foundations; the Pantheon of Rome has a dome 142 feet in diameter built mainly of concrete and still in perfect condition after 1900 years. In many other old-world countries there exist many examples of the ancient use of concrete. The lookout towers of Ireland, for mstance, supposed to have been built by the Druids more than 1000 years ago, are made of hydrauhc cement concrete. In the new world concrete was used by the Peruvians in the days of Incas, and in North America the Mound-Builders, a race believed to have been existent 11,000 years ago, made utensils of artificial stone. Through the period of stagnation known in history as the Middle Ages the making of cement and concrete appears to have become a lost art. Not until the eighteenth century was cement again made or con- crete used. To an English engineer, John Smeaton, who was commis- sioned to build the Eddystone hghthouse, must be credited the redis- covery of a process of making hydraulic cement. Finding lime mortar unsatisfactory for his purposes, Smeaton began in 1756 a series of experi- ments in the course of which he discovered that the calcination of a clayey limestone yielded a product superior in every way to ordinary lime and possessmg the property of setting under water. Smeaton used, how- ever, only that part of the lunestone which yielded a product which would slake with the addition of water. His discovery was therefore 90 POETLAND CEMENT 91 an hydraulic lime. Forty years later, in 1796, Joseph Parker of Kent County, England, took out a paterit for the process of manufacture of a cement which he called " Roman cement," and which he made by burning and subsequently grinding argillo-calcareous nodules called " septaria," the composition of which resembles many of the natural cement rocks of the present day. Parker obtained the nodules principally from the shores of the Isle of Sheppy, where they were washed up after a storm. This cement, the precurser of modern natural cements, came rapidly into favor in England. In 1824, Joseph Aspdin, a brick mason of Leeds, England, was granted a patent on a method of manufacture of a cement for which he proposed the name " Portland cement," because of a real or fancied resemblance of the concrete made therefrom to the natural oolitic limestone so exten- sively quarried for building purposes at Portland, England. Aspdin proposed to make his cement from the dust of roads repaired with lime- stone, or from limestone itself combined with clay, by burning the mixture and subsequently grinding the product. Aspdin is usually given the credit for the invention of Portland cement, though it is doubtful whether he carried his investigations any further than his predecessors had done. The specifications in his patent failed to state either the relative propor- tions of clayey and calcareous material necessary, or the required degree of calcination of the mixture. In the United States the cement industry began with the discovery in 1818 of a natural cement rock in Madison County, N. Y., by Mr. Canvass White, an engineer on the construction of the Erie canal. The natural cement made and patented by White was used in large quan- tities in the construction of the canal. Within a few years after the dis- covery of natural cement rock in Madison Co., N. Y., other deposits of cement rock were found, and the manufacture of natural cement began near Louisville, Ky., near Hancock, Md., at Utica, 111., Akron, N. Y., at Balcony Falls, Va., at Siegfried, Pa., at Rosendale, N. Y., at Ft. Scott, Kan., at Buffalo, N. Y., at Milwaukee, Wis., and in many other places, all of which have at some time supported a natural cement industry of some importance. The first Portland cement manufactured in the United States was made at Coplay, Pa., in 1875, by Mr. David 0. Saylor, who had begun in 1872 to experiment with the making of Portland cement from the rock used by the Coplay Cement Co. (of which he was president) for the manufacture of natural cement. Saylor was at last successful in making a Portland cement comparable with the foreign Portlands by cal- cining at a high temperature a mixture of the argillaceous limestone rock with a comparatively pure limestone rock. During the next five 92 MATERIALS OF CONSTRUCTION years several Portland cement plants were put in operation in the United States, many of which were commercial failures. Between the years 1880 and 1890 the Portland cement industry slowly grew, but it was not until 1895, with the introduction of coal burning in the rotary kiln, that there began the phenomenally rapid growth which has since characterized the industry. 77. Definition of Portland Cement. In spite of a wide variation in the composition and character of the raw materials used in the manufacture of Portland cement the composition of the product derived therefrom varies only within quite narrow limits, and it would therefore seem a simple matter to define that product in a manner universally acceptable. Thus far, however, no definition has found wide acceptance. By common agreement any definition must prescribe a material consisting essentially of lime, silica, and alumina, properly proportioned, burned to the point of incipient fusion, and finely pulverized. It is neces- sary, however, both in the interests of the consumer and the manufacturer, to restrict the application of the term to only those cements made in a fairly definite manner, from a somewhat restricted class of raw materials, the product having a composition within certain practical limits. Some widely used definitions, notably that proposed as the American standard in the specifications for cement adopted by the American Society for Testing Materials, 1909, have been very loosely constructed. The trend, however, of national societies and associations is now clearly in the direction of a very restricted definition, a fact attributable, per- haps, to the quite general desire to prevent the application of the term " Portland cement " to cements made by burning a natural rock without previous artificial mixing or grinding, whatever the subsequent treatment, also to exclude cements made by mixing Portland cement with pulver- ized blast furnace slag, and those made by mixing natural cement with Portland cement. With the above considerations in view the following definition of Portland cement is proposed. The definition is a modification of that proposed by the American Society for Testing Materials, its provisions being, however, more neariy in accord with the definition adopted as the German Standard by the Association of German Portland Cement Manu- facturers, March 16, 1910. Definition. The term " Portland cement " is applied to the finely pulverized product resulting from the calcination to incipient fusion of an intimate artificial mixture of argillaceous and calcareous materials, this product to contain not less than 1.7 parts by weight of lime to 1 part by weight of silica + alumina + ferric oxide, not more than 4 per cent of magnesia, nor more than 1.75 per cent of anhydrous sulphuric acid, POETLAND CEMENT 93 and to which no addition greater than 3 per cent shall have been made subsequent to calcination. Explanation. The high content of lime in Portland cement necessi- tates the intimate mixing of the raw materials in quite exact propor- tions obtainable only in an artificial way under chemical control. Burn- ing to incipient fusion (sintering), insures the high density so essential to Portland cement. The provision allowing a maximum magnesia content of 4 per cent presupposes that this magnesia content be taken into account in adding the lime. An excess of magnesia causes unsound- ness of cement. The allowance of a maximum sulphuric acid content of 1.75 per cent is necessary because of the unavoidable presence of sul- phurous compounds in the raw material and in the fuel. In addition, sulphur compounds are necessarily introduced by the addition of gypsum or plaster of Paris to regulate the time of setting. The provision Umit- ing the additions made subsequent to calcination to 3 per cent is designed primarily to prevent the possibility of additions made solely for the pur- pose of increasing the weight. 78. Portland Cement as a Structural Material. The use of Port- land cement as a material of engineering construction is so universal that little need here be said concerning the class of construction work on which it may be used. As an ingredient of concretes and mortars it is by far the most important of all masonry materials used in modern engi- neering construction. As monolothic concrete it is used in all types of massive masonry works, such as foundations and footings, piers and abutments, dams, retaining walls, pavements and roadways, etc. When reinforced with steel it is used for framework, walls, floors, and roofs of buildings, for arch and girder bridges, for tunnels, subways, conduits, and innumerable other purposes. In combination with sand alone it is used as mortar for laying brick or stone work, and as a plaster or stucco it is applied to either exterior or interior walls upon a base of terta cotta, brick or metal lath. In general cement holds rank as a structural material second only to steel and possibly timber. (Although an enormous amount of timber is used structurally, a comparatively small amount is used on important permanent work whose features possess any real interest from the strictly engineering point of view.) One great difference exists, however, as to the conditions under which cement is used as compared with steel. The importance of this con- sideration cannot be over-emphasized. Steel is delivered upon the work as a finished material, manufactured under standardized conditions, every step in the metallurgical processes involved in its production, and the mechanical operations of its fabrication having been most carefully 94 MATERIALS OF CONSTRUCTION supervised. Upon the work the fabricated units are simply assembled, this operation also being done by well-standardized methods by work- men specially trained in doing this one class of work. Cement, on the other hand, although now manufactured under fairly well-standardized conditions, usually with competent mill supervision, is received upon the work as one ingredient only of a structural mate- rial, concrete or mortar, which is built in place. The materials which are combined to form concrete or mortar, i.e., sand and crushed stone or gravel, are too often used without careful examination or selection, and ■ — most important of all — the mixing and the deposition of the material in place is as a rule done by unskilled laborers, often without competent supervision. These conditions are largely responsible for the fact that Portland cement concrete and mortar are not as reliable materials as is steel, and accounts for the fact that good practice does not permit its use on impor- tant work without the closest supervision at all times by the engineer in charge or his representative. Many a failure of concrete structures has been attributed to faulty cement, which should have been attributed to poor sand or stone, or to a foreman who considers that " anyone can mix and place concrete." PORTLAND CEMENT MANUFACTURE 79. Raw Materials. The essential constituents of Portland cement are as above stated hme, silica, and alumina. (The place occupied by iron oxide, magnesia, etc., in the constitution of cement will for the present be neglected.) With the exception of lime these substances are found free in nature, but not, however, in a form practicable for use in cement manufacture. Lime is always used in the form of a carbonate, and silica and alumina in the form of clay, shale, or slate. Eckel* makes the following classification of the raw materials: Calcareous.- Argillo-calcareous. Argillaceous (CaCOa over 75 %) (CaCOa = 40-75 %) (CaCOs under 40 %) Pure hmestone. Clayey Hmestone. Slate. Hard Pure chalk. Clayey chalk. Shale. Soft ' Pure marl. Clayey marl. Clay. Unconsolidated. Alkali waste. Blast furnace slag. Unconsolidated. The combination of the materials in any two of these groups which will give a mixture of proper composition might be used as the raw mate- * "Cements, Limes and Plasters," p. 301. PORTLAND CEMENT 95 rial for Portland cement. The only combinations, however, which have been used in this country are in order of their present importance: (1) Argillaceous limestone (cement rock) and pure limestone. (2) Marl and clay or shale. (3) Limestone and shale or clay. (4) Blast furnace slag and limestone. (5) Chalk and clay. (6) Alkali waste and clay. While proper chemical composition and physical character are the primary requisites for the raw materials chosen, many other considera- tions assume great importance when the factors contributory to the eco- nomical manufacture of Portland cement are considered. Among these may be mentioned the ■ availability of the deposits, the location with respect to a market and transportation facilities, and lastly the location with respect to fuel supplies. 80. Limestone. Limestones occur widely distributed throughout the country. When pure, limestone forms the mineral calcite (CaCOs), and all limestones consist essentially of calcium carbonate combined with more or less impurities. The principal foreign elements commonly found are magnesia, silica, iron, alkalies, and sulphur. Magnesia in the form of carbonate of magnesia occurs very commonly in limestone, but, since the effort is always made to keep the magnesia content as low as possible in Portland cement, a limestone containing much over 5 per cent of carbonate of magnesia will be unsuited for use. Silica may be present either alone or in combination with alumina. When alone it may occur as flint in pebbles or beds, or less commonly, in the form of mica, hornblende, or other complex silicates. The silica does not readily combine with Ume in the kiln, and more than a very small amount of silica renders a limestone unfit for use. Silica combined with alumina is a very common impurity in limestone and such limestones are of great value to the cement manufacturer. Compounds of siHca and alumina readily combine with lime in the kiln and the argillaceous limestones are therefore among the most important classes of raw mate- rials for the manufacture of Portland cement. Iron occurs usually as either the oxide (Fe203) or sulphide (FeS2), and less commonly as a carbonate or sihcate. Except as a sulphide the iron forms a useful flux, aiding the combination of lime and silica in the kiln. As a sulphide it is injurious and to be avoided if in amounts over 2 to 3 per cent. The alkaUes, soda and potash, commonly occur in limestones in small percentages. Unless present in quantities over about 5 per cent, in which 96 MATEEIALS OF CONSTRUCTION event they may be carried over into the cement with harmful results, they will be largely driven off in the kiln with no consequent effect upon the cement. Sulphur may be present as iron pyrite or as hme sulphate. In either case its presence is extremely injurious and not over 1 to 2 per cent can be tolerated. The approximate range of composition of limestones used in American Portland cement manufacture is indicated by the following summary: Component. Approximate Range. Per cent. Usual Percentage. CaCOa SiOj AljOa+FejOa. MgCOs 88.0-98.0 0,3- 8.0 0.2- 2.1 0.2- 4.2 93.0-97.0 0,4- 3.0 0.5- 1.3 1.0- 2.5 81. Argillaceous Limestone, Cement Rock. The term " cement rock " is technically applied to a limestone containing about 68 to 72 per cent of hme carbonate, 18 to 27 per cent clayey matter, and not over 5 per cent of magnesium carbonate. Such a rock is found in many districts of the country, but has been largely used for the manufacture of Port- land cement only in the Lehigh district of eastern Pennsylvania and west- ern New Jersey, a territory 4 miles wide and 50 miles long. This dis- trict at one time (1899) produced three-fourths of the Portland cement manufactured in the United States and even now (1913) produces nearly one-third of the entire output (29.5 per cent). The cement rock is a dark gray to black, slatey limestone, softer than pure Umestone and consequently more easily ground. As a rule the cement rock must be mixed with a comparatively pure hmestone in small percentages. In a very few cases, however, the cement rock contains an excess of calcareous material, necessitating the admixture of shale or clay with the cement rock. 82. Marl. Marls are deposits of comparatively pure carbonate of Hme found in beds of existing or extinct lakes. The manner of forma- tion of marl beds is a matter of some dispute but, whether due to deposi- tion of hme carbonate by purely physical and chemical agencies, or partly through the agency of certain vegetable and animal life, a deposit of lime carbonate in a soft, friable form, usually in a finely granular state, is formed. Organic matter, clay and carbonate of magnesia are the principal impurities found in marls, with sometimes small amounts of sulphur in combination with organic matter, iron, or hme. Marls usually analyze about 90 to 97 per cent CaCOs and MgCOs, less than PORTLAND CEMENT 97 1 per cent Si02, less than 1 per, cent AI2O3 and Fe203 combined, the bal- ance being made up of small amounts of organic matter, SO3, etc. When used in the manufacture of Portland cement marls usually require the addition of from 15 to 20 per cent clay. The large percentage of water (often 50 to 60 per cent) usually present in the marl upon arrival at the plant is disregarded in the above statement of composition. 83. Clays, Shales and Slates. Clays, shales, and slate may in gen- eral be considered of the same ultimate composition, differing only in the degree of consolidation. All clays are formed from the debris result- ing from the decay of rocks, and hence they will differ greatly in composi- tion and physical character. Clays left where rock disintegrates are called residual clays, when transported and deposited by stream action they are sedimentary clays, and when they are deposited by glacial action they are glaciali clays. The different classes of clays differ in composition owing to differences in the manner of their deposition. Residual clays are apt to contain fragments of quartz, flint, or lime carbonate, depending upon the character of the rock disintegrated; sedimentary clays in their long water transportation usually have lost all their coarser material and so form a fine-grained homogeneous product; the glacial clays show even less homogeneity than the residual clays and are apt to contain much sand, gravel and pebbles. Absolutely pure clay is hydrated silicate of alumina or kaolin (AI2O3, 2Si02, 2H2O). Such a clay is not available for cement manufacture but it is imperative that the clay be as free as possible from gravel and sand. The proportion of silica should not be less than 55 to 65 per cent, and the combined amount of alumina and iron oxide should be between one-third and one-half the amount of silica (22 to 27 per cent). The presence of gypsum or pyrite in the clay is injurious, and magnesia and alkalies should not be present in quantities exceeding about 3 per cent. Shales are simply clays hardened by pressure, but they have almost invariably been formed from deposits of sedimentary clay and so do not show the irregularities in composition common to most residual or glacial clays. Shales are preferable to soft clay for mixing with limestone because, on account of the similarity in physical character, segregation of the two is less likely to take place. They also carry less water and therefore require less drying. Clay, upon the other hand, is better adapted to use with marl. The slates are clays which through heavy pressure have solidified in a markedly laminated structure and acquired the property of splitting readily into thin sheets. The slates find only a limited application in the manufacture of Portland cement and then only as a utiUzation of the waste from roofing slate works. 98 MATERIALS OF CONSTRUCTION 84. Alkali Waste. The precipitated calcium carbonate obtained from the manufacture of caustic soda by the Leblanc process has been used in Europe for the manufacture of Portland cement. This waste often carries high percentages of sulphur in the form of sulphides, however, and is thereby rendered unfit for use as a Portland cement material. The waste from alkali works using the ammonia process is a very pure precipitated Ume together with some lime-hydrate. The sulphur content is usually very low, making this waste a material superior to that derived by the Leblanc process. No Portland cement is made in the United States at the present time (1914) using any class of alkaU waste as a raw material, and in view of the availability of better suited mate- rials, it is unlikely that use of this waste will again be considered. 85. Blast-furnace Slag. Three classes of cement which must not be confused are made with blast-furnace slag as one of the ingredients. One is the slag or puzzolan cement made by grinding blast-furnace slag with hydrated Ume without subsequent calcination; a second is a true Portland cement made by mixing limestone and slag, grinding the mix- ture very finely, and calcining the product as in the usual process of Port- land cement manufacture; the third is the German " Iron-Portland " cement made by grinding finely together 70 per cent of true Portland cement and 30 per cent of granulated blast-furnace slag. Blast-furnace slag is a fusible silicate formed during the smelting of iron ore by the combination of the fluxing material with the " gangue " of the ore. The slags used in cement manufacture are those of strongly basic character, the higher the lime the better. The following analysis is typical of the slags used in the manufacture of cements of the second class above given. Silica 33.10 Iron -oxide and alumina 12 . 60 Lime ; 49.93 Magnesia . 2 . 45 There is a slight chemical or thermal advantage in the use of slag as a cement material owing to the fact that the lime is present as the oxide (CaO), instead of as the carbonate (CaCOs), meaning therefore a saving of fuel in the kiln. On the other hand this advantage is partially offset by the fact that the granulation of the slag by running the molten material into water results in the absorption of from 15 to 45 per cent of water, which must subsequently be driven off, thus increasing the fuel consumption. 86. Proportioning the Raw Materials. The definition of Portland cement above given (Art. 77) declares it to be a material containing " not less than 1.7 parts by weight of Ume to 1 part by weight of sihca-|- PORTLAND CEMENT . 99 alumina+iron oxide, not more than 4 per cent of magnesia, nor more than . 1.75 per cent of anhydrous sulphuric acid." The combining of the raw materials in such a manner as to achieve the desired ratio of calcareous to argillaceous materials is not, however, entirely a simple matter, for the reason that Portland cement after calcination is not a mixture of lime and clayey materials, but is what may be termed a " solid solution " of a number of components including silicates and aluminates of Ume, but no free lime. The minimum ratio of lime to clayey material pre- scribed in the definition very roughly expresses the relative proportions in which the two classes of material are understood to combine, and the actual proportions of two given materials which will produce a satis- factory cement can only be determined upon the basis of a knowledge of the detailed composition of each of the component raw materials, and a further knowledge of the compounds which will be formed during cal- cination, i.e., the constitution of finished cement. It is easy, of course, to analyze each of the constituent raw materials, but the constitution of cement, so far as the practice of cement chemists is concerned, is unknown. What is actually done under these circum- stances is to proceed upon the basis of an assumption as to what the essential constituents of cement are, the assumption being as a rule based upon the investigations of cement constitution made by M. Le Chatelier and the Messrs. Newberry many years ago. The most recent studies of this problem tend to show that neither Le Chatelier nor the New- berrys were entirely correct in their conclusions, but experience has showji that the methods of proportioning which are based upon these conclu- sions will produce an excellent cement and, that being the case, there is no immediate prospect of any change being made in the practice of cement chemists, regardless of the results of more modern studies of the problem of cement constitution.* The essential facts as to the constitution of cement obtained as the result of the studies of Messrs. Newberry have been briefly noted above (Art. 38). The essential constituents being considered to be tricalcium silicate and dicalcium aluminate, the proportion of lime to silica and alumina is expi'essed by the following rule, called " Newberry's Rule for Proportioning ": Max. lime = 2.8(%Si02) -hLl(%Al203). Eckel's modification of this rule, taking account of the magnesia and the iron oxide, has also been stated and explained above (Art. 38), the form of Eckel's rule called the " cementation index " being: 2.8( %Si02)-H.l(%Al203)+Q.7(%Fe203) ^ ^ %CaO-M.4(%MgO) *See discussion of " Constitution of Portland Cement," in Art. 106, Chapter VII. 100 MATERIALS OF CONSTRUCTION A value of the cementation index below 1 necessarily means an excess of lime or magnesia in the cement, i.e., free lime or magnesia is present. In the practical application of this rule factory chemists aim to attain a composition whose cementation index is shghtly above 1.0, the pro- portion of the limestone used being reduced for the sake of safety about 10 per cent below the maximum called for by the rule. The following example illustrates the application of the rule to the determination of the correct proportions of untried raw materials: We assume that the raw materials are of the following composition: SiOj. AI2O1. FeaOa. CaO. MgO. SOa. Water, CO;, etc. Clay Limestone . 61.92 1.54 16.58 0.39 28 04 2.01 54.97 1.68 0.52 0.02 0.04 10.61 41.50 (1) Clay: Silica X 2.8 = 61 .92 X 2.8 = 173.38 AluminaXl. 1 = 16.58X1.1= 18.24 Iron oxide X 0.7= 7.28X0.7= 5.10 196.72 LimeX1.0 = 2.0lX1.0= 2.01 MagnesiaX1.4 = 1.58X1.4= 2.21 196.72- (2) Limestone: 4.22 -4.22 = 192.50 Silica X 2.8 =1.54X2.8= 4.31 Alumina XI. 1=0.39X1,1= 0.43 Iron oxideX0.7 = 1.04X0.7= 0.73 5.47 Lime X 1.0 = 54.97X1.0= 54.97 Magnesia X 1.4= 0.52X1.4= 0.73 55.70 55.70-5.47= 50.23 (3) Proportion required: , , , , , , 192.50 -=-50.23 = 3.83 parts of limestone to 1 part of clay by weight. Since, as above noted, ideal conditions of grindmg and burnmg cannot be attained in practice, it is customary to reduce the theoretically correct percentage of limestone about 10 per PORTLAND CEMENT 101 cent for safety. Therefore 3.83-0.38=3.45 parts of limestone to 1 part of clay to be actually used. The application of a rule based upon complete analyses of the raw materials is not necessary once a plant is well established with fairly uniform raw materials. Usually a fixed standard total percentage of carbonate (CaCOs and MgCOs) is found by experience with any given raw materials to give a satisfactory mixture, and this standard is there- after maintained as long as the raw materials remain unchanged. Some- times instead of a fixed lime standard, the ratio of total carbonates to total insoluble matter is similarly used. 87. Control of the Mixture During Operation of Plant. The ideal method of control consists in the analysis of both raw materials at the plant before grinding, grinding and mixing according to these analyses, analyzing the mixture as a check, and correcting the mix by the additicc of the constituent required before calcination. In practice, cheaper and quicker methods are adopted as a rule. Either the analysis of the raw material is entirely depended upon and no subsequent effort made to check and correct the mix, or the raw materials are ground and mixed in approximately correct proportions without analysis, and the ground mixture analyzed and then corrected by the addition of the material found deficient. 88. Treatment of Materials Preliminary to Calcination. The cal- cination of Portland cement materials must invariably be preceded by two processes which may or may not be distinct one from the other: (1) the reduction of the materials to an impalpable powder; (2) the intimate mixing of the materials in proper proportions. Often the mix- ing of the materials is accomplished simultaneously with the final pulver- ization and after the preliminary grinding. The treatment of the raw materials before calcination follows, in general, one of two possible processes; (a) The dry process; (6) the wet process. The use of the latter process obtains only when the raw mate- rials consist either of marl and clay or chalky hmestone and clay, and con- stitutes an almost negligible portion of the cement industry in the United States. 89. Theory of Calcination. The principal chemical objects accom- plished by calcination of Portland cement mixtures are, in the approx- imate order of their sequence: (1) the evaporation of water, (2) the dis- sociation of carbonates of lime and magnesia, (3) the expulsion of the alkalies, (4) the oxidation of ferrous to ferric oxides, and (5) the com- bination of Hme and magnesia with- silica, alumina, and ferric oxide to form the silicates, aluminates, and ferrites, which make up the consti- tution of Portland cement. 102 MATERIALS OF CONSTRUCTION Incidentally, contact of the materials with the fuel ash, the kiln lining, and the kiln gases results in the addition of clayey constituents, silica, alumina, and iron oxide, thus very slightly reducing the proportion of lime and magnesia in the finished product. Most of the moisture is driven off at temperatures only slightly ex- ceeding 100° C. Lime carbonate is dissociated at temperatures somewhat above 900° C, and magnesium carbonate at temperatures probably between 800° C. and 900° C. The formation of silicates, alummates, and ferrites does not take place at temperatures below about 1100° C, and for most commercial cement mixtures the attainment of a tempera- ture of about 1550° C. has been found necessary in order to insure the combination of practically all of the lime with the clayey constituents. The Dry Pbocess 90. Quarrying, Crushing and Dr3ring the Rock. Ninety-five per cent of the material used in Portland cement manufacture is obtained by Fig. 38.— Jaw Crusher. quarrying (the handling of marl being excluded from consideration.) The first step in quarrying operations consists in stripping off the sur- face soil. Where the depth of stripping is not too great it may be done without power by scrapers. Where the stripping means the removal of a deep soil cover heavy excavators or hydraulic methods may be employed. PORTLAND CEMENT 103 Quarries are usually opened on a side hill, the rock is blasted down in benches, reduced to manageable size, and removed in small cars running on movable tracks. The steam shovel is often used to load the blasted material upon the cars, and where comparatively soft material is encountered it may be depended upon to excavate the material without blasting. Mining, the obtaining of material by underground work- ings in shafts or tunnels, is rarely employed in the obtain- ing of cement materials be- cause of the excessive cost as compared with quarrying. Occasionally, however, a valu- able stratum of cement rock. Fig. 39.— Gyratory Crusher, limestone, or shale may be overlaid by a thick stratum of other material, making underground working cheaper than stripping and quarrying. Mining has one advantage over Fig. 40.— Edison Roll Crusher. open quarrying in that it is not affected by the adverse weather conditions which usually make quarrying impossible for at least one-fourth of the time. 104 MATERIALS OF CONSTEUCTION The raw materials employed in the dry process are in general in the form of more or less compact rock, either cement rock and limestone or limestone and shale. A few isolated plants where marl and clay or shale, or blast-furnace slag and limestone, are employed, constitute exceptions to the general rule. SECTION "A-A'^ Fig. 41. — Ruggles-Coles Dryer. Preliminary reduction is usually accomplished in a crusher of either the jaw type, Fig. 38, or, much more commonly, one of the gyratory type, Fig. 39. A roll crusher, Fig. 40, is occasionally used. It is in many cases found economical to use more than one size of pre- liminary crusher, the smaller machine taking care of the smaller-sized quarry rock and the over-sized material from the larger machine. Fig. 41a.— Rotary Dryer. The degree to which the preliminary crushing is carried is quite vari- able m the practice of different plants. The general practice, how- ever, IS to reduce the rock only to a size that will pass a 2- or 2|-inch rmg. The presence of moisture in the rock as it comes from the quarry very much impairs the efficiency of grinding and pulverizing machinery. It IS therefore necessary to dry the rock after crushing. PORTLAND CEMENT 105 The type of dryer used almost exclusively in cement mills is the rotary dryer, Figs. 41 and 41a. From the dryers the material is conveyed to the raw grinding mill where it is ground, mixed, and finely pulverized. 91. Grinding, Mixing and Pulverizing the Raw Materials. The further reduction of raw material after drying is almost* invariably car- ried on in two stages. In the first stage the materials, either mixed or 106 MATEEIALS OF CONSTRUCTION separately, are ground to a size varying in the practice of different plants from i inch down to ^V inch or less. In the second stage the mixture of the constituent materials is pulverized to the final degree of fineness required. The choice of the type of grinding and pulverizmg machmery Fig. 43. — Kominuter. adopted is largely dependent upon the age of the plant, the character of the raw materials, and the' type of equipment originally installed. Most American plants operating on the dry process use the ball mill, Fig. 42, or the kominuter. Fig. 43, for the initial grinding of the raw material, in conjunction with the tube mill, Fig. 44, or FuUer-Lehigh Fig. 44.— Tube Mill. mill, Fig. 45, for final pulverization. The Griffin mill. Figs. 46 and 47, is to a lesser extent used for final pulverization, and the Huntington mill, Fig. 48, is similarly used in one large plant. The Williams mill and several types of rolls and disintegrators are sometimes used for raw grind- ing of certain classes of material, and the Raymond pulverizer is occasion- ally used for final pulverization. PORTLAND CEMENT 107 As noted above, mixing of the two classes of raw material may take place at any one of several points in the process of preparing the material for calcination. The choice depends largely upon the relative physical character of the two classes of material and the degree of uniformity in composition of the materials. Two rock ores such as limestone and shale, or limestone and cement rock, may run so nearly constant in composi- tion that the chemist's analyses made in the quarry maj' be trusted, and the mixture may be propor- tioned by weight either just before, or immediately after crushing, with- out fear of segregation of the two rocks in subsequent handling. Such a combination as a hmestone and a clay would give difficulty by segregation, by reason of their differing physical characters, if mixed before being reduced to a finely divided state. Probably the point in the proc- ess where mixing is most often accomplished is immediately after initial grinding and before pulver- ization, an analysis of the materials having been made after grinding. In the extremely finely divided state attained in the process of pulverization a very intimate mix- ing without danger of segregation is accomplished. Very often some type of me- chanical weighuig and proportioning machine is installed between the stages of initial and final grinding; Conveyors discharge the two mate- rials into two hopper scales. Fig. 49, which are so devised that the con- tents of the two hoppers discharge simultaneously into a common hopper when the weight of material in each hopper has reached a predetermined weight dependent upon the chemist's determination of the proper pro- portions of the two materials. Probably no other factor contributory to the production of a satis- factory cement holds so important a plaice as does the degree of fine- FiG. 45.— FuUer-Lehigh Mill. ,108 MATEKIALS OF CONSTEUCTION Fig. 46.— Griffin Mill. allowed, and the amount of surface exposed or tho state of subdivision of the constituent materi- als. Fineness of grind- ing will therefore lead to economy in calcination, i since either the duration j or the temperature of t burning will be lessened by increased fineness of grinding. On the other hand, if the temperature or duration of burning are not increased to com- pensate for lack of fine grinding the production of a relatively homogene- ous product is impossible. There will be very im- perfect diffusion between ness attained in grinding the raw mix. Tbe bearing of this factor is felt both in the con- sideration of economy of op- eration of the plant and in the consideration of the quality of the product. Since the temperature of calcination is simply a sin- tering temperature, and not suflBcient to fuse the mixture and so produce a homogenous product, diffusion between the lime and alumina and silica must take place at a tem- perature usually not exceeding about 1600° C. The amount of diffusion in a soUd is de- pendent upon three factors, the temperature, the time ^ Fig. 47,— Bradley Mill. POETLAND CEMENT 109 the constituents of the mix, resulting in the production of a cement lacking volume con- stancy. The actual degree of fineness attained in practice is somewhat dependent upon the character of the mate- rials used and other local factors. The mix- tures of cement rock with a relatively low proportion of limestone, for instance, require less Fig. 48.— Huntington Mill. fine grinding than do the mixtures of two pure classes of material such as a pure limestone and a clay or shale. In general the degree of fineness required in most instances is such that not less than 95 per cent will pass the 100-mesh sieve, and as much as 98 per cent through the 100-mesh sieve is preferable. 92. Burning the Ce- ment Mixture. For present purposes the calcination of the ce- ment mixture may be considered to be always accomplished in the rotary type of cement kiln. Fig. 50. Differ- ences in operating methods necessitated by the use of kilns other than the rotary type will be later noted briefly in connection with descriptions of various types of kilns. Raw material is dis- FiG. 49. — Tandem-automatic Weighing Machine. charged into the kiln no MATERIALS OF CONSTRUCTION from thfe supply bins either through an inclined spout, Fig. 51, or a water-jacketed screw conveyor, Fig. 52, running through the stack flue. Usually the feeding device is belt-connected to the kiln drive so that the Fig. 50.— Detail Construction of a 100-foot Kiln. feeding starts and stops with the kiln. When wet slurry is burned in the kiln it is pumped in from a tank below, provision being made to make the rate of supply uniform. Fig. 51.— Kiln Feed by Spout. Coal used for kiln burning is usually gas slack and should contain as httle ash as possible, 25 per cent being about the maximum allowable. Coal is usually crushed in rolls or pot crushers, ball or Williams mill, dried in a rotary dryer, and pulverized in a tube mill, Fuller-Lehigh mill, Raymond mill or Griffin mill. PORTLAND CEMENT 111 The pulverized coal is delivered by conveyors from the coal-grinding mill to bins located above and behind the burner end of the kiln. A screw conveyor, Fig. 53, usually carries the coal from the supply bin to a point where it falls into an air injector, where it en- counters an air blast which conveys it through a pipe to a nozzle which projects a foot or more into the kiln. The air thus introduced by the blower is only about one- fourth that required for combustion, a large amount being drawn in by the natural draft of the kiln stack through the opening in the hood where the clinker es- capes. The pressure of air used is sometimes very high — 60 to 80 pounds per square inch — in which event only 7 to 10 per cent of the air required will be so intro- Screw Conveyor Fig. 52. — Kiln Feed by Jacketed Conveyer. Fig. 53.— Coal Feed for Kiln. 112 MATEEIALS OF CONSTRUCTION Ft.i 8 12 16 20 24 28 32 36 40 44 48 52 56 60 1 60 Ft. Total Length ol Klin Dischar^ End duced. Natural-draft systems, whereby the coal is made to fall in a thin sheet across a slit in the end of the kiln, being carried in by the air sucked in by the natural stack draft, have also been introduced. Crude oil, the use of which first made the operation of the rotary kiln a commercial success, is now used only in certain districts where its cost remains low while coal is high. When burned in the rotary kiln it is sprayed in by a blast of air from blowers or air compressors. In 6rder to distribute properly the heat in the kiln two or more oil burners are used. Natural gas is now used in only a few plants, situated for the most part in the State of Kansas. As the sup- ply gradually diminishes it is being replaced by oil or coal. When burned in the kiln it is introduced in a manner entirely similar to the use of oil, a portion of the air required for combustion being used to inject the gas. The operation of the rotary kilns requires at all times the attendance of a skilled burner who may look after from two to four kilns. The fuel supply and the speed or rotation, and hence the temperature of the kiln and degree of burning of the clinker, are under the direct control of the burner. The heat is judged simply by the incandescence of the interior (viewed through darkened glasses) and the degree of burning of the clinker, as well as the temperature found by experience to be best under any special plant conditions, are maintained as uniformly as possible. Proper burning is determined by the color and appearance of the clmker, the properly burned clinker being a greenish black in color, having a vitreous luster and showing bright- glistening specks when just cooled. The lumps are, in the main, from a size of a waUiut down. Under- Chemical Changes in 60-Ft. Kiln Fig. 64. PORTLAND CEMENT 113 burned clinker is brown or has brown centers, and lacks the luster of well- burned clinker. Over-burned clinker has hard brown centers. Over- burned clinker is probably not injurious except for very low lime cements, but over-burning means a fuel waste, and the grinding expense will be increased owing to the greater hardness of over-burned clinker. The nature and sequence of the chemical changes which take place in the kiln has been noted above. Art. 89. Fig. 54 graphically represents the nature of the changes during burning, as determined by experiments made by Wm. B. Newberry, who analyzed the contents of a 6 by 60- foot kiln during a temporary shut-down. Similar experiments made on longer kilns show that the extra length accomplishes practically nothing except the driving off of the moisture in the raw material and heating up the material to the dissociation temperature of the carbonates, heat that would otherwise be carried off by the flue gases being thus utilized. Under good average conditions an 8 by 100-foot kiln should turn' out about 600 barrels of cement per day with a fuel consumption of about 80 pounds of coal per barrel. A 9 by 150-foot kiln should turn out about 750 barrels per day. An idea of the sources of heat and the utilization of same in the rotary kiln is conveyed by the following approjdmate estimate: SOUBCES OF HEAT Heat from combustion of coal 71.8 -84.7 per cent Heat from mix, coal, or air 0.0 -10.8 per ctot Heat from chemical combination 9.0 -17.4 per cent HEAT UTILIZATION Heat used in evaporating water 0.15- 1.7 per cent Heat used in dissociation of sulphates 0.33- 0.76 per cent Heat used in dissociation of carbonates 19.6 -25.0 per cent HEAT LOSSES Heat lost with hot clinker 10.7 -15.5 per cent Heat lost with stack gases and flue-dust 43 . 6 -51 . 4 per cent Heat lost by radiation and combustion 11.2 -15.4 per cent The low heat efficiency attained in the rotary-kiln calcination of cement has led to a quite general effort to devise means of utihzation of the waste heat. Heat lost by radiation cannot by any means be utilized, but may only be reduced as much as possible by the placing of a poor heat conductor between the kiln lining and the shell. Heat carried away from the kiln in the stack gases may be reduced in some degree by lengthening the kiln, thereby utilizing a part of this heat in drying and preheating the raw material in the upper part of the kiln. 114 MATERIALS OF CONSTRUCTION The utilization of the heat carried by the stack gases has been tried out in two ways: the first and most common method is by using these gas^s for the drying of the raw material in rotary dryers, the second method consists in passing the gases through vertical water-tube boilers, thus generating steam for the operation of the power plant. Both methods have been fairly successful and promise in time to be more generally adopted. The principal difficulty encountered is the presence of large quantities of dust in the kiln gases, which makes impossible most ordinary methods of heat regeneration. The utilization of the heat carried off by the hot clinker is quite common in German practice and is becoming more so in this country. In general the method apphed consists in drawing the air used for combustion in the kiln through the hot clinker in a rotary cooler, thus cooling the clinker and preheating the air. 93. Treatment of the Clinker, Cooling, Grinding, and Pulverizing. The clinker as it issues from the kiln is very hot, and must be reduced to a suitable temperature before being ground. Occasionally it is the prac- tice of cement plants to allow the clinker to cool in piles outside the mill. More generally, however, it has been found advisable to adopt some type of mechanical cooling device which may or may not provide for the recovery and utihzation of the heat carried off by the hot clinker. It is the usual practice to grind the clinker by the same type of grind- ing machinery used in the raw-material mill, the grinding and pulverizing being practically invariably done in two separate stages. The principal systems are therefore: 1. Ball mill, kominuter, or rolls, followed by a FuUer-Lehigh mill. 2. Ball mill, kominuter. Griffin mill, Kent mill, Sturtevant mill, or Huntington mill followed by tube mill. 3. Series of rolls. 94. Addition of Retafder. The clinker produced in the rotary kiln process makes a cement which is naturally very quick setting because of its high lime content. In order, therefore, to retard its set sufficiently to enable it to meet commercial requirements, it is the universal practice to add sulphate of lime either before grinding or between the stages of grinding and pulverizing the clinker. The retarder was at one time added in the form of plaster of Paris (CaS04 -|- IH2O) . Now, however, the universal practice (with the exception of a few plants which add plaster of Paris after pulverizing the clinker) is to add raw gypsum (CaS04-|-2H20). The gypsum is obtained in the form of lumps crushed to pass a 1-inch ring and is added to the clinker either by hand or by mechanical weigh- ing devices, the quantity used being, as a rule, about 2 per cent, and never exceeding 3 per cent. The retarding agent is the sulphur trioxide POETLAND CEMENT 115 present, whether plaster or gypsum be used, and the more frequent choice of the latter is due entirely to the advantage in cost of crude gyp- sum over that of the dehydrated form. 95. Storing and Packing. The finished cement is stored in stock- houses containing bins holding from 1000 to 5000 barrels each. The bins usually discharge to hoppers and screw conveyors in tunnels below each row of bins. Cement is either packed in wooden barrels containing 380 pounds, or in cloth or paper bags containing 95 pounds net. Packing is commonly done by some type of semi-automatic machine (Fig. 55), which fills, weighs, and seals the barrels or bag. Cloth bags are now almost exclusively used. The Wet Peocess 96. The Wet Process of Manufacture, Using Marl and Clay or Shale. Iii American practice only one class of raw materials is handled by the wet process — mixtures of marl with clay or shale. In English and German practice chalk and clay are often handled by wet-process methods. In general the mate- rials so handled are those of a soft physical character, and ' which in their natural state carry high percentages of moisture. Marl is usually not only saturated with moisture, but is often covered with water to a considerable depth. Under such circumstances it is obtained by dredging. Usually the excavator is carried on a barge which floats in a channel formed by the dredge in its progress through the basin. The material excavated may be loaded by the dredge upon cars running on tracks alongside the dredged channel, and thence be carried to the mill, ■ or in some cases the material may be loaded by the excavator into hop- pers which feed a pug mill, wherein the marl is mixed with additional water so that it may be pumped to the mill through a pipe line. The marl usually reaches the plant in the shape of a thin mud containing about 50 per cent water. It is passed through a separator to remove stones, roots, etc., and then stored in large cylindrical tanks. Fig. 55. — ^Automatic Packing Machine. 116 MATERIALS OF CONSTRUCTION The clay upon arrival at the plant is often dried in order to facihtate the determination of the correct proportion to be added to the marl. It is then ground in an edge runner mill (Fig. 56), between mill-stones, or in a disintegrator. Fig. 56.— Edge Runner Mill. From the storage tank the marl is pumped either into a measuring tank or a weighing hopper. The ground clay is delivered to bins above, and added to the marl in the proportion determined by analysis of the materials. 3H O— q FiQ. 57.— Pug Mill. The mixture is now discharged into a pug mill, Fig. 57, which consists simply of a horizontal cylinder within which two shafts provided with steel propelling blades rotate. The mixture is churned up by the revolv- ing blades, thoroughly mixed, some additional water being usually PORTLAND CEMENT 117 admitted, and discharged ini,c large vats or dosage tanks where it is sam- pled, analyzed, and the mixture corrected by the addition of the correct amount of clay and marl. In order to prevent any part of the mixture from settling, it is necessary to provide these tanks with revolving arms with paddles which keep the mass constantly agitated. The slurry is now passed on to final grinding, which is usually done either in a wet-emery mill or in a wet- tube mill. The output of the finishing mill is conveyed to supply tanks for the kihis and is charged in without any previous drying. It usually contains, therefore, from 60 to 65 per cent of water which must be evaporated in the upper part of the rotary kiln. This practice necessarily increases the fuel consumption of the kiln, but has been found to be more practical than the drying of the mixture in a rotary dryer prior to calcination. The production of Portland cement by the wet process^ using marl and clay, has steadily decreased for several years, many plants finding it advantageous to substitute limestone for marl and use the dry process. 97. The Wet Process, Using Chalk and Clay. Th^ chalks and clays utihzed in Germany and, more particularly, in England, for the manu- facture of Portland cement, usually contain from 20 to 30 per cent water, and the favorite method of reduction of the material is by the use of a series of wash mills (Fig. 58) . The clay and the chalk are sometimes separately ground in wash mills which subsequently discharge into a common wash mill. Particu- larly is this the case if one material is considerably more refractory than the other. Commonly, however, both materials are discharged into one mill in the desired proportion. Gratings placed at integrals around the periphery of the pits allow the slurry to be washed gradually from one mill to the next. Three or four mills are commonly used, the material becoming finer with each successive pass. The quantity of water used in the operation of the wash mill is some- times only sufiicient to make a thick slurry, 40 to 45 per cent water; in other cases it is sufficient to make a very thin slurry, carrying about 80 per cent water. When the thin slurry is used settling basins are usually provided from which the water can be drawn off at intervals, the mass remaining being usually dug out and briquetted for use in stationary kilns. The wash mill operated with 40 to 45 per cent water leaves the mate- rial in 'a state of subdivision sufficient to pass about 90 per cent through a 100-mesh sieve, the residue consisting largely of comparatively large nibs of calcium carbonate which, if not removed or further ground, would mean the presence of free hme in the manufactured product. It is there- fore the common practice in the more modern European plants using the 118 MATERIALS OF CONSTRUCTION wash mill, either to separate out these large particles by use of a sifting device or to further reduce the material by grinding with either mill stones, emery mills, or wet-tube mills, the latter practice being the more commonly employed. Fig. 58.— Wash Mill. Semi-dry Processes 98. Semi-dry Processes. In a few instances both here and in Europe the wet-slurry process of handling raw materials has been applied to dry raw materials such as limestone and shale, etc. PORTLAND CEMENT 119 The pulverized raw materials are passed to a pug mill, the mix cor- rected in dosage tanks, and the wet slurry fed to rotary kilns, as in the wet process. The only justification for such treatment is the availability of cheap kibi fuel such as crude oil or natural gas. Under such condi- tions some saving may be effected, since an intimate mixture of the con- stituents is procured without such fine grinding as is required for the dry process. Where coal fuel must be used, the saving in the grinding depart- ment is more than offset by the extra fuel requirement of the kiln for drying the slurry, and the only advantage remaining lies in the fact that as in the wet process, the composition of the mix is under better control than in the dry process. Cement Mill Equipment 99. Equipment of Raw Mill. Rock Crushers. Crushers employed in the initial reduction of the quarry output are either jaw crushers or rotary gyratory crushers. In the first type of crusher (Fig. 38) the reduction is accompUshed between a fixed jaw and a hinged reciprocating jaw. In the latter type (Fig. 39) the material falls within a hopper-shaped chamber Uned with concave corrugated plates mounted con- centrically with the crushing head. The latter is a corrugated steel cone widening toward the bottom, where the annular space between the cone hopper and the crush- ing head is only large enough to permit the passage of material of a predetermined size. The crushing head is rotated through an eccentric drive which imparts to it an oscillating motion and thereby lessens the liability of choking. A third class of crusher used for primary reduction consists of two horizontal rolls made to rotate in opposite directions and provided on their periphery with corrugations or large teeth. The best-known crushers of this class are the Edison rolls (Fig. 40), the cylindrical surfaces of which consist of studded plates which grip and crush the rock. In the more recently designed Edison rolls there are provided at two diametrically opposite points on one roll a row of " sluggers," or unusually high knobs, which shatter large rocks sufficiently to bring their size within the angle of grip of the rolls. The capacity of the larger sized Edison rolls is so great that quarry stone up to 7 feet in thickness may be handled. Dryers. The type of dryer used in cement miUs almost exclusively is the rotary dryer. 1 his device (Fig. 41a) consists of a cyUnder, 5 to 7.5 feet in diameter, and 40 to 60 feet long, set on a shght inclination, and rotated slowly on roller bearings. The material fed in at the upper end passes slowly through to the discharge end by gravity. Often the interior wall of the dryer is provided with angle irons bolted on parallel to the axis of the cylinder for the sake of facilitating the drying process by Ufting and dropping the material as the cylinder revolves. The dried material escapes from the lower end of the cylinder into a conveyor or elevator whereby it is transported to the grinding mills. The rotary dryer is heated either by the direct combustion of fuel in a furnace located at the lower end of the cylinder, or by the waste gases from the cement kilns. In either case the hot gases pass through the length of the dryer, escaping up a stack set at the upper end. The use of pulverized coal in a manner similar to its use in the rotary kiln has lately been introduced. 120 MATERIALS OF CONSTRUCTION For drying some classes of material which normally carry very high percentages of water, double-heating dryers are sometimes used. The Ruggles Coles Dryer (Fig. 41) is perhaps the besHinown dryer of this type. It consists of two concentric cylin- ders fastened together and revolved on roller bearings. The inner cylinder connects with a furnace at the upper end and is open at the lower end. A fan blower is used to exhaust the gases from the annular outer chamber at the upper end, the path of the gases being therefore down through the central flue and back through the space between the cylinders. The material to be dried is fed in between the cylinders at the upper end, and is caught up by flights on the inside of the outer cylinder and dropped on the hot inner shell. As the cylinder revolves this action is repeated, the drying being further hastened by the hot gases passing through the outer chamber. The fuel consumption of the rotary dryer is greatly dependent upon the amount of water carried by the material and the character of that material. Marls, because of their high content of organic matter, are especially retentive of moisture. In gen- eral the fuel requirements of the dryer vary from about 1 pound of coal per 5 pounds of water evaporated, under the most unfavorable conditions, to about 1 pound of coal per 7 or 8 pounds of water evaporated under favorable conditions. Grinding Machinery. Centrifugal Grinders and Ring Roll Mills. The Griffin mill (Fig. 46) consists essentially of a steel ring or die against the inside surface of which a heavy steel crushing roll, mounted on a pendulum suspended from a universal joint, is made to roll by centrifugal force. The pendulum is rotated by a pulley at the top and the crushing roll is provided with plows which throw back into the grinding zone the material which has passed the roll into the pit below. The fan mounted upon the pendulum shaft above the roll facilitates the passage of the sufficiently fine material through the annular screen placed just above the grinding zone. The Griffin mill will turn out 8 to 10 barrels of raw material per hour ground to pass about 95 per cent through the 100-mesh sieve. The three-roll Griffin mill or Bradley mill (Fig. 47) is a modification of the usual Griffin mill, resembling it in every respect except that the single crushing roll is replaced by three rolls. The Huntington mill (Fig. 48) is somewhat similar to the three-roll Griffin mill. It is used only in the plants of one large manufacturer. Three heavy crushing rolls are freely suspended from a circular revolving spider. As the head revolves the rolls swing outward through the action of centrifugal force, pressing against an annular die-ring of steel. On raw material the output of the Huntington miU will be from 15 to 25 barrels per hour ground to pass 93 per cent through a 100-mesh sieve. The Raymond impact pulverizer resembles the Bradley and Huntington mills except in its detail design. The Kent mill (Fig. 59) and the Maxecon mill, which latter is simply an improved machme employing exactly the principles of operation of the Kent mill, are ring- roll mills which in some respects resemble the Griffin and Huntington mills, but which employ powerful springs instead of centrifugal force for the grinding pressure. Three convex-faced roUs are mounted on equidistant horizontal axes within the grinding nng. One roll is power driven, and this rotates the ring by traction, the latter in turn rotating the other two rolls. The output of the mill is separated by an auxiliary silting mechanism, the mill itself having no separating device, and the coarser portion IS returned for further grinding. .u J!l^ St"rtevant ring-roll mill operates in a manner similar to the Kent mill except that the ^le is positively driven and not through traction by one of the rolls Ihe class of mill in which a number of balls are driven around a horizontal annular die IS a special type of ring-roll mill. The FuUer-Lehigh mUl (Fig. 45) is perhaps the PORTLAND CEMENT 121 Adjusting. Feed Hoppec bestrknown mill of this class. Four balls, weighing 100 pounds each, are impelled around the grinding ring at a speed of 600 feet per minute, by arms secured to a verti- cal spindle. They press against the grinding ring through centrifugal force. The material is fed into the center of the mill, forced by a fan into the grinding cham- ber, and the fine material expelled through sieves protected by grids. The output of the mill on raw material is from 20 to 30 barrels per hour ground to pass 95 per cent through a 100-mesh sieve. An unusually high proportion of this material will pass a 200-mesh sieve. Rotary Attrition Mills. Ball Mills, Kominuters, and Tube Mills. The ball mill (Fig. 42) is a short drum rotated at a rate of 20 to 30 revolutions per minute about a horizontal axis, the interior surface being so formed that it consists of a series of steps. The drum contains a number of large steel balls which roll and fall from step to step, thereby crushing and grinding the material fed in through one axis. The stepped, plates are perforated, and the material, when sufficiently ground, passes through the perfora- tions onto a screen and finally onto a second screen. The particles too coarse to pass the screens are re- turned to the grinding chamber through the openings between the overlapping grinding plates. Ma- terial which passes the screens will escape into the hopper below and thence to the conveyors. The out- put is 15 to 24 barrels per hour, to pass a 20-mesh screen. The kominuter (Fig. 43) is a modification of the ball mill, the principal point of difference being that the grinding plates are not stepped or perforated and thej material must pass the length of the drum to an outlet provided at the further end, where it drops on a screen outside the grinding plates. These screens being conicaUy shaped, slope toward the inlet end, and discharge the 'over-sized mate- rial back into the drum at the center of the inlet end. The material sufficiently ground escapes in the same manner as in the case of the ball mill. The kominuter is capable of grinding 20 to 30 barrels of raw material per hour to a size to pass a 20-mesh screen. The tube mill (Fig. 44) resembles the ball mill and kominuter in that it is a rotating cylinder partially filled with an abrasive agent. The tube mill is, however, much longer than the ball mill or kominuter. It is usually employed for final reduction of the product of mills of another type. The tube mill consists of a steel cylinder, 5 to 6 feet in diameter, and 20 to 22 feet long, lined with hardened plates or hard natural stone and mounted on trunnions about which it is slowly revolved at a rate of 25 to 30 revolutions per minute. The abrasive charge usually consists of flint pebbles which fill it approximately to the axis of rotation. The material is fed in through a hollow axis at one end, and either escapes in a similar manner at the opposite end through a hollow axis and screens, or is allowed Any OrQund ocSloox Fig. 59.— Kent Mill. 122 MATEEIALS OF CONSTRUCTION to escape through gratings at the periphery of the discharge end. Steel balls are occasionally substituted for flint pebbles as the abrasive agent. The fineness of grind- ing is regulated by the rate of feed, and the capacity for a 5 by 22 foot mill is about 12 to 20 barrels per hour ground to pass 95 to 98 per cent through a 100-mesh sieve. The wet tube mill differs in no essential respect from the dry-tube mill except that the discharge is usually through a screened drum located axially, rather than through perforations located on the periphery of the discharge end. The Edge Runner Mill. The edge runner mill (Fig. 56) consists of a cas1>iron shallow pan in which a pair of wide-tired cast-iron rollers revolve by traction as the pan is rotated, or else the central vertical shaft, which carried the rollers on cranks at right angles to its axis, is itself rotated, the pan being stationary. This latter type is less common. The edge runner mill may be arranged to work with either wet or dry material provided only that the material be of a not too refractory nature. The material is fed into the pan in front of one of the rolls and, after passing the roll, is diverted by scrapers to the perforated plates which surround the pan, then by scrapers again diverted to the pan in front of the other roll, and so the process continues, the size of the output being dependent upon the perforations of the plate, usually being not much under J inch, though sometimes under ot inch. The output of i-inch material for a 5-foot roll machine working on shale is about 6 tons per hour. The miUstones, the rock-emery mill, and the other types of grinding machinery used in the wet process with marl and clay have been described above. (Art. 64.) 100. Cement Kilns. Early Kilns. The earKest type of kiln used for cement burning both in Europe and America was an upright stationary-shaft kiln called a dome kiln or "bottle" kiln (Fig. 60), which closely resembled the kilns used for lime burning. The fuel and the raw slurry molded into bricks were charged in at the top and the clinker was withdrawn at the bottom, the process being inter- mittent. The chnker required hand sorting to separate out the over- and under-burned material. The fuel con- sumption was high, and the demand for hand labor in molding the brick and sorting the clinker, excessive. The first efforts to improve the effi- ciency of the dome kiln were in the direc- tion of utilization of the hot waste gases of the kiln for the drying and preheating of the raw material. The Johnston kiln attempted to make this saving by a simple arrangement whereby the waste gases of the kiln were led through a long flue within which the briquetted material was sorted and dried. The kihi was still wasteful of fuel m consequence of the fact that its operation was intermittent. The Hoffman or ring kiln (Fig. 13), which has been described in connection with the discussion of lime burning, was next utiUzed in the cement industry. It is very economical of fuel, but requires an excessive amount of skilled hand labor in molding the bricks and charging the kiln. The Hoffman kiln is still in operation in a few Get- man cement plants in districts where labor cost is very low and fuel cost very high. Fig. 60.— Dome Kiln. PORTLAND CEMENT 123 The Dietsch shaft kilns (Fig. 61) and the Schofer Icilns (Fig. 62) consist of a long vertical flue terminating in a stack, the lower portion of which serves as a preheating chamber for the dried slurry briquettes which are charged with the fuel. A chamber below the preheating chamber serves as a combustion chamber, and the lower portion serves as a cooling chamber for the clinker, the incoming air for combustion being heated thereby. The shaft kilns are the most economical of fuel of any type of kiln, but they all suffer under the disadvantage of requiring the briquetting of the raw material and the -i Fig. 61.— Dietsch Shaft Kiln. Fig. 62.— Schofer Kiln. sorting out of the under-burned and over-burned cUnker by hand. The fuel consump- tion is only about 45 pounds of coal per barrel, but the labor cost is excessive. The shaft kiln is still in use to a limited extent in Europe, but even in Germany, where economic conditions are most favorable to their use, the ring kiln and the shaft kiln hav.e been replaced in two-thirds of the plants by the rotary kiln. The Rotary Kiln. The modern rotary kiln (Fig. 50) consists of a cylindrical steel shell lined with refractory material and supported on rollers, its axis being slightly inclined from the horizontal. It is revolved on its bearings by a variable speed trans- mission; its upper end projects into the chimney foundation, which also serves as a dust trap; its lower end enters a movable hood adapted to the discharge of clinker and the entry of fuel and air. Pulverized coal injected by an air blast is most commonly 124 MATERIALS OF CONSTRUCTION used as fuel, although crude oil is used in plants whose output total.about 15 per cent of the output of the United States, and natural gas fuel is used in the production of about 5 per cent of the total. The first rotary kiln was patented in 1885 in England by Ernest Ransom. It was first made a commercial success in this country, however, the fuel used being petroleum. The use of powdered coal as a fuel was introduced in 1895, from which time the remark- ably rapid growth of the industry dates. Rotary kilns in use to-day vary in size from 60 feet long and 5 feet in diameter to 240 feet long and 12 feet in diameter. Present practice, however, favors the use of the longer kilns, and few are now installed less than 125 feet long by 8 feet in diameter, and 150 feet long by 10 feet in diameter seems to be about the present standard size for new installations. The kiln is inclined at a pitch varying from j to I inch per foot of length, and is supported by rollers bearing on wide steel tires at from two to five points in the kiln length, the number of supports depending upon this length. Two or more horizontal thrust rollers keep the inclined kiln in place on its bearings. In general the rate of rotation is from one-half to two- thirds of a revolution per minute, though in some cases the rate is as low as one-eighth revolution per minute, and in other cases as high as one or even, in rare cases, two revolutions per minute. The rate of revolution depends upon the slope of the kiln and the nature of the raw material. The lining of the kiln is usually a highly refractory magnesite or bauxite brick, but in a few instances linings of fire clay or of brick made from Portland cement clinker and Portland cement have been used. The thickness of the lining is variable, being usually from 9 to 12 inches at the lower and hotter end, and 4 to 6 inches at the upper end. In some kilns burning wet slurry the upper part of the kiln is left unlined. The brick are keyed to fit the circle of the kiln, and are held in place by the heavy angle iron which encircles the inside of each end of the kiln. The hood at the lower end of the kiln is usually mounted on a movable carriage, and the masonry of its front wall is provided with two openings, one for the entrance of the burner apparatus, the other for observa- tion of the operation of the kiln and the insertion of bars for breaking up rings of clinker and making minor re- pairs to the lining. The lower part of the hood is left partly open to allow the escape of the clinker, which is either diverted to one side into cars, or falls directly onto conveyers. This opening also serves as an air inlet. 101. Clinker Cooling Equipment. Equipment of Finish- ing Mill. Many of the older plants use an upright gravity cooler (Fig. 63) consisting of a cylinder about 8 feet m diameter and 35 feet high, provided with baffle plates and shelves. The hot clinker is discharged into an elevator pit, where it is sprinkled with water; the elevator discharges it into a hopper at the top of the cooler, whence it falls down over the series of baffles, where it encounters air currents introduced through a perforated vertical pipe in the center of the cooler. A hopper at the base discharges the cooled clinker to a conveyor. Fig. 63. Vertical Clinker Cooler. PORTLAND CEMENT 125 More modem mills employ a rotary cooler (Fig. 64) which consists of a steel cylin- drical shell mounted on rollers in a slightly inclined position and rotated slowly in a manner similar to the operation of the rotary dryer. The rotary cooler is often mounted as an under-cooler directly below the kiln for convenience in direct discharging of the clinker from the kiln to the cooler, the cooUng air being drawn into the kiln. The cooler lining consists usually of cast-iron plates bolted to the shell, and shelves are usually provided as in the dryer for the purpose of lifting the material. Water to help cool the clinker is sometimes introduced at the upper end of the cooler and per- forations in the shell during the lower 3 feet of its length,. IJ to 2 inches in diam'eter, screen the ncaterial roughly. An angle iron with the flange projecting inward at Fig. 64. — Rotary Coolers Mounted beneath Kilns. the discharge end of the cooler forces the material to escape only through the perforations. Sometimes the rotary cooler is mounted separately from the kiln, and in such an event it is not usually used to preheat the air for the kiln, but the current of air is drawn through by a stack or blown through by a fan. The grinding equipment of the finishing mill is, as above noted, usually a duplicate of that used in the raw material mill. The reason for this circumstance is the fact that if the same types of machines are used in both mills only one line of repair parts need be carried. 102. Cost of Manufacture. Manufacturing costs are dependent upon many factors, among which may be mentioned the cost of labor, the character and availability of the raw materials, the plant equipment and management, and the plant location with respect to transportation facilities. Cost figures are naturally difficult to obtain for any individual plant. In general, however, it may be said that the cost per barrel 126 MATERIALS OF OONSTEUCTION at the plant is rarely below 55 to 60 cents per barrel, and in many plants such a figure cannot be approached. 103. Production of Portland Cement. The statistics quoted m Art. 74 give the production of Portland cement in the United States in the last forty years. It will be noted that the industry showed a iah but not remarkable growth from its commencement until 1895, when coal burning in the rotary kiln became commercially successful. From that time the growth of the industry has been uniformly rapid and con- tinuous, with the exception of the years affected by the financial crisis of 1907.' For the past few years the growth has averaged from 15 to 20 per cent per year, a rate of increase that no other industry .of equal mag- nitude can approach. PROPERTIES AND USES OF PORTLAND CEMENT 104. General. The value of cement as a structural material depends primarily upon its mechanical strength when hardened. The condi- tions met with in practical use are, however, necessarily so variable as to exclude the possibility of the establishment of standards directly based upon practical experience. The establishment of the existing standards for physical and mechan- ical properties has for its object the fixing of values of certain properties, readily determined in the laboratory, for cements found satisfactory in practice, in order that inferiority in any particular cement may be detected by deviation from such standards. Therefore the results of all laboratory tests of cement cannot be considered to represent the properties of the material under the condi- tions in which the cement is used in structures, but merely the properties shown under certain standardized conditions, the quantitative results obtained having only a relative value. The physical properties so utilized for comparative purposes are: specific gravity, fineness, time of setting, and soundness. The mechan- ical properties similarly used are the tensile and compressive strength in neat cement mixtures and in sand mortars. 105. Composition of Portland Cement. A study of a large number of analyses of commercial cements reveals the fact that the range in possible composition is comparatively limited. For the purpose of showing graphically the extent of the possible variation in composition of Portland cements, Figs. 65 and 66 are introduced. These figures are constructed upon the basis of the following table, which is a summary of detailed analyses of eighty samples of American Portland cement (representing forty-three brands) quoted by Eckel. PORTLAND CEMENT 127 Component. SiOz AlA FezOs CaO MgO CaO+l-4(MgO) Alk SO3 Ave. of 80 Samples. Per cent. Actual. 21.83 7.43 3.31 62.57 1.85 65.14 1.15 1.33 Equiv. (22.34) (7.61) (3.39) (66.66) Ave. of ,5 Highest Brands. Per cent. Actual. 24.21 9.54 4.99 65.06 3.34 68.27 1.93 2.39 Equiv. (24.29) (9.69) (5.12) (68.40) Ave. of 5 Lowest Brands. Per cent. Actual. 19.69 4.67 1.95 59.22 0.45 61.77 0.55 0.39 Equiv. (20.49) (4.79) (2.00) (64.80) Of the compounds included in the above table the alkalies and anhydrous sulphuric acid may be omitted in the discussion of the com- position of cement, their amounts being small and, moreover, some- what arbitrarily limited by practical considerations. Magnesia cannot be neglected, but by the consensus of opinions of most authorities we are safe in assuming that magnesia may be considered equivalent to lime in its action, due allowance being made for the differ- ence in their combining weights. (The atomic weight of Ca being 40.1, Mg, 24.3, and 0, 16, the relative combining weights will be CaO -f- MgO = 56.1-T-40.3 = 1.4.) Figures expressing lime -|- equivalent magnesia have therefore been added in the above table. We have, therefore, four constituents of cements to consider — silica, alumina, iron oxide, and lime -f equivalent magnesia. If, now, we con- sider the average percentages of all constituents, we find they do not total 100 per cent, and the original figures are therefore corrected on the basis SiO2-l-Al2O3H-Fe2O3-FCaO-|-1.4MgO = 100 per cent, and have been placed in parentheses in the column adjoining the original figures. In the second division of the table are given the average percentages of each constituent for the five brands averaging highest for that partic- ular constituent. Similarly in the third division are given the percentages for the five brands averaging lowest for each particular constituent. The figures placed in parentheses in the second and third divisions of the table have been computed from the complete analyses of each of the five cements whose average is expressed by one of the original figures. The manner of derivation of the corrected percentages may best be explained by an example: For instance, the figure 24.21 in the second division of the table represents the average percentage of silica for the five cements which ranked highest in silica. The average amounts of AI2O3, Fe203, CaO, and 1.4 MgO in these same five cements were similarly computed, and 128 MATERIALS OF C0NSTRI3 CTION AloO, the percentage of silica corrected on the basis of the total amount of SiO2+Al2O3+Fe2O3+CaO + 1.4MgO = 100 per cent. The corrected per- centage of silica was thus found to be 24.29 per cent. By similar computations the remaining figures in this column and the last column of the table were obtained. The construction of the diagrams of Figs. 65 and 66 is dependent upon the geometrical principle that the simi of the normals from any point within an equilateral tetrahedron to the four sides is equal to the normal from a vertex to the opposite side. We des- ignate the four vertices to represent 100 per cent CaO + 1.4MgO, 100 per cent Si02, 100 per cent AI2O3, and 100 per cent Fe203, respectively. Then any chosen point within the tetrahedron represents a compound or mixture having certain definite percentages of siOj each of the four con- stituents (i.e., the per- centage of Ume will be represented by the ratio of the normal from the chosen point upon the Al203-Si02- Fe203 - plane to the altitude of the tetrahedron, the percent- age of Fe203 by .the ratio of the normal upon the CaO-Al203-Si02-plane to the altitude, etc.) The maximum and minimum percentages for each con- stituent as given in the Fig. 65. — Composition of Portland Cement. AlsOl ■Composition of Portland Cement. PORTLAND CEMENT 129 above table are therefore represented in the diagram by two planes,, each being parallel to the side of the tetrahedron opposite the vertex which represents the constituent in question. The sohd body represented in perspective in Fig. 65 and in plan in Fig. 66 is the portion of the tetrahedron CaO, AI2O3, Si02, Fe203 occupied by the field of Portland cement. 106. The Constitution of Portland Cement. Chemical analysis suf- fices for the determination of what chemical elements are present in cement and their relative percentages. It will not, however, ascertain in what manner these elements are combined, that is, the constitution of the cement. The solution of the problem of the constitution of Portland cement has been the goal of much scientific research for many years past, but only during the last few years has the investigation been pursued along lines which could promise ultimate solution. Chemical analysis being of no avail, there remain two possible methods of investigation — micro- scopic and synthetic. Microscopic investigation of cement clinker was first attempted suc- cessfully by H. Le Chatelier (1883), who applied to the study of cement clinker the methods employed in petrography in the study of rocks. Sections of clinker were reduced to such a degree of thinness by careful grinding that they became sufficiently transparent for microscopic exam- ination by transmitted light. A later modification of this method more easily employed consists in polishing a single surface only of the specimen, developing the structure by etching with a suitable reagent, and examining by reflected light. A third method consists in examining powdered clinker by transmitted light, the powder being held in some transparent medium. Le Chatelier and subsequently Tornebohm independently discovered four different kinds of crystals by microscopic examination of clinker. These were named by the latter Alit, Belit, Celit, and FeUt. Clifford Richardson, who has made a similar study of clinker by microscopic methods, has thus summarized the facts established: " Alit is the preponderating element and consists of colorless crystals of rather strong refractive power, but of weak double refraction. By this he means that alit in polarized light between crossed Nicol prisms has insufficient optical activity to produce more than weak bluish gray interference colors. " Celit is recognized by its deep color, brownish orange. It fills the interstices between the other constituents, being the magma or liquid of lowest freezing-point out of which the alit is separated. It is strongly double refractive, that is to say, gives brilliant colors when examined between crossed Nicol prisms. 130 MATERIALS OF CONSTEUCTION " Belit is recognized by its dirty green and somewhat muddy color, and by its brilliant interference , colors. It is biaxial, and of high index of refraction. It forms small round grains of no recognized crystalline character. " Felit is colorless. Its index of refraction is nearly the same as that of belit and it is strongly double refractive. It occurs in the form of round grains, often in elongated form, but without crystaUine outUne. Felit may be entirely wanting. " Besides these materials an amorphous isotropic mass was detected by Tomebohm and Le Chatelier. It has a very high refractive index. " Tomebohm adds the important fact that a cement 4 per cent richer in lime than usual consists almost entirely of alit and ceUt." Synthetic investigations are not studies of actual kiln-clinker, but are studies of the product of burning definite artificial trial mixtures of pure lime, silica, alumina, etc. A complete synthetic investigation involves the determination of all the possible compounds, and eutectics (mix- tures of two components in the particular proportion resulting in a mix- ture which exhibits a lower melting-point than other neighboring mix- tures) which they form with one another in the two component systems (as CaO — Si02, CaO— AI2O3, etc.), and further, in the three and four component systems. Le Chatelier and the Messrs. S. B. and W. B. Newberry made syn- thetic investigations so important that their work is still largely the foun- dation upon which the factory control of cement mixtures is based. Clifford Richardson later made synthetic studies of great importance, but the most valuable contributions to the knowledge of the constitution of Portland cement made during the last few years have been the results of the research recently completed by the physical chemists, physicists, and optical mineralogists connected with the Geophysical Laboratory of the Carnegie Institution in Washington. Le Chatelier as a result of his investigations concluded that the essen- tial constituents of Portland cement are: (1) Tricalcium silicate, 3CaO, Si02; (2) tricalcium aluminate, SCaO, AI2O3; and (3) " a fusible calcium silico-aluminate whose chief function is that of a flux during burning to promote the necessary chemical reactions." As has been noted above (Arts. 38 and 86), the Messrs. Newberry's conclusions as to the essential constituents of cement differed from those of Le Chatelier in that they believed that alumina occurred as the dicalcic aluminate, not as tricalcic aluminate. They agreed with Le Chatelier as to the occurrence of the tricalcic silicate as the other essential con- stituent. , The studies of the constitution of cement clinker carried on in the PORTLAND CEMENT 131 Came^e Geophysical Laboratory at Washington by Messrs. A. L. Day, E. S. Shepherd, G. A. Rankin, and F. E. Wright during the last few years constitute, as above noted, the most important contribution to our knowl- edge of the subject. Three lines of investigation have been followed in the researches made in the Geophysical Laboratory: chemical, thermal, and optical. They are not independent, but overlap to some extent, the data from one serv- ing to supplement or corroborate the results obtained by another method. The behavior of lime, silica, and alumina was first exhaustively studied in the two-component systems lime-alumina, lime-silica, and silica-alumina, after which the investigation of the three-component system, lime-silica-alumina, was undertaken. In the study of the two- component systems no trace of the tricalcic silicate (3CaO, Si02), con- sidered by all previous investigators to be an important constituent of clinker, was found. In the more recent work in the ternary system it has been found that a small addition of alumina brings out the tricalcium silicate, which appears to be a compound of very peculiar properties, becoming dissociated into hme and the orthosiUcate below the melting temperature. In the presence of alumina, however, the fusion tem- perature is sufficiently reduced so that the compound can crystallize out of the melt and may therefore form an important part of cement clinker. The work of the Geophysical Laboratory has established the possi- bility of the existence of five classes of cement clinkers:* I. II. Typical. III. IV. V. Uncertain. CaO 3CaO, AW, 3CaO, SiOz 3CaO, SiOj 2CaO, SiOj 3CaO, AI2O3 2CaO, SiOz 3CaO, AI2O3 5CaO, 3AI2O3 2CaO, Si02 CaO, AUO3 6CaO, 3AI2O3 2CaO, Si02 CaO, AI2O8 2CaO, AI2O3, SiOz The four-component series, lime — silica — alumina — iron oxide, has not. yet been investigated, nor has the truth of the assumption made by Eckel and others that magnesia may replace an equivalent amount of lime been established. A single magnesium aluminate was found to exist, having the formula MgO, AI2O3. No information was obtained bearing upon the question of the possible magnesium silicates. Magnesium riietasilicates and orthosihcates, MgO, SiOa and 2MgO, Si02, are known to exist, however. * Amer. Jour. Sci., Vol. 22, p. 301; Amer, Journ. Sci., Vol. 28, p. 293, 132 MATEEIALS OF CONSTRUCTION Messrs. Helpert and Kohlmeyer * have by a synthetic study of the binary series, lime— ferric oxide, established the existence of calcium ferrite, CaO, FegOs, and the tricalcium ferrite, 3CaO, FeaOs- The former is formed with difficulty and is devoid of hydraulic properties; the latter is hydraulic. It will be noted that the typical constitution as determined by the researches at the Geophysical Laboratory is in agreement with the con- clusions of Le Chatelier except that the beta orthosiUcate replaces Le ChateUer's "fusible calcium silico-aluminate." The dicalcic aluminate of the Newberrys was found to be non-existent. The conclusions above quoted from the reports of Messrs. Day, Shepherd, Rankin, and Wright have been recently verified (1914) by Messrs. Rankin and Wright, f and have been independently verified by Messrs. Klein and PhiUips of the Pittsburgh Laboratory of the U. S. Bureau of Standards. As a result of the quite general acceptance which has been accorded the reports from the Geophysical Laboratory, it may be asserted with a considerable degree of positiveness that the typical constitution of Port-- land cement is: Tricalcic silicate, 3CaO, Si02. Tricalcic aluminate, 3CaO, AI2O3. Beta orthosilicate, 2CaO, Si02. Under certain conditions the orthosilicate may be lacking and free lime present, or, either the tricalcium silicate or aluminate or both may be lacking, and the 5 : 3 calcium aluminate or the monocalcic aluminate or both present. 1C7. Setting and Hardening. When a true hydraulic cement is gauged with sufficient water and then left undisturbed it soon loses its plas- ticity and finally reaches a state when its form cannot be changed without producing rupture. This change in condition is known as the " setting " of cement and has usually been considered somewhat distinct from " hardening." Setting usually takes place in a few hours or even minutes, while hardening may proceed for months or even years. The successive theories which have been held regarding the setting of cement have been closely allied with the successive theories which have been held regarding the constitution of cement. *S. Helnert and E. Kohlmeyer, Ber. deut. chem. Ges., Vol. 42, p. 4581; Metal- lurgie. Vol. 7, pp. 193, 225. t Messrs. Rankin and Wright have not yet published, their 'final report. The above statement is based upon assurances privately communicftted to the author by Mr. Rankin, PORTLAND CE|tIENT 133 Le Chatelier considered that the tricalcic aluminate of lime in contact with water first became hydrated, the resulting crystallization explaining the initial setting of the cement. He ascribed the later hardening to the decomposition of the tricalcic silicate in contact with water resulting in the formation of hydrated monocalcic silicate crystals and calcium hydrate crystals, the progressive hardening being due to the growth and interlocking of the crystals. Richardson later modified this theory in important particulars. According to his views the formation of Ume silicates and aluminates during clinkering is only a convenient way of securing indirectly a very active lime hydrate which is itself the real cementing material. " On the addition of water to a stable system made up of the solid solutions which composed Portland cement a new component is intro- duced which immediately results in a lack of equilibrium, which is only brought about again by the liberation of free lime. This free lime, the moment it is Uberated, is in solution in the water, but owing to the rapid- ity with which it is liberated from the aluminate, the water soon becomes supersaturated with calcic hydrate, and the latter crystallizes out in a network of crystals which binds the particles of undecomposed Port- land cement together. From the characteristics of the silicates and aluminates it is evident that the latter are acted upon much more rapidly than the silicates, and it is to the crystallization of the lime from the aluminates that the first or initial set must be attributed. Subsequent hardening is due to the slower liberation of lime from the silicates. If the hme is liberated more rapidly than is possible for it to crystallize out from the water, expansion ensues and the cement is not volume constant. " The strength of the Portland cement after setting is due entirely to the crystallization of calcium hydrate under certain favorable condi- tions, and not at all to the crystallization of the silicates or the aluminates, since in this act of hydration nothing can take place which would tend to bind these sihcates and aluminates together." * The recent development of interest in the so-called " colloidal theory " of the late Dr. William Michaelis t offers another explanation of the setting and 'hardening of cement. According to Michaelis, " the calcar- eous hydraulic cements owe their setting mainly to the formation and desiccation of colloidal hydrated aluminum silicates, and their later hardening to the formation of colloids, calcium silicate and calcium hydroxide, the latter in the crystalline form." * Clifford Richardson, Eng. News, Vol. 53, p. 84 (190.-V t See Translation of Michaelis' latest work published by " Cement and Engi- neering News," 1909. 134 MATERIALS OF CONSTRUCTION Michaelis supports his theory by the following argument based on volume constancy: The action of lime on silica and alumina in the pres- ence of water leads to the formation of a gelatinous mass of variable composition. The absorption of water by the colloidal mass is accompanied by a great increase in the volume of the individual particles of the cement, but not an increase in the volume of cement and water. If cement were a crystalhne substance a certain definite volume^ would be assumed upon the taking up of the water of crystaUization, which volume would be thereafter constant except for minute changes due principally to capillarity, as in the case of natural stones. Experi- ments and practical experience ■ have abundantly shown, however, that cement and cement mortar are not volume constant, but that they invar- iably shrink upon setting and hardening in air and expand upon setting and hardening in water; phenomena that are not characteristic of crystal- line substances. Many experiments have shown that the coarser particles of cement constitute simply so much inert matter. It is known that even the sieve having 300 meshes per lineal inch is not fine enough to pass only active material, which latter can be separated out only by some suspension method which retains only the finest impalpable powder or " flour." Experiments made by Messrs. H. S. Spackman and R. W. Leshe strongly support the above statements. Cement that had passed the 200-mesh sieve was separated into three parts by shaking it in kerosene and allowing it to settle. The portion which settled in less than thirty seconds was only slightly acted upon by water after even two years. The portion which settled in from thirty seconds to one minute was only acted upon by water in from three to four months, and was hydrated only in part then. The portion which remained in suspension more than one minute was acted upon immediately, swelling up and forming a very voluminous jelly. Still other evidence might be cited in support of the theory that the setting of cement is due to the formation of a colloidal mass by the addi- tion of water to the finest portion of the ground chnker which, as the only active part of the cement, binds together the coarser inert particles of clinker as a sort of " mineral glue," which has little strength or hardness in itself, but has great binding power, which is more or less cumulative in its action when formed in contact with inert matter such as the coarser particles of clinker or with sand. The " colloidal theory " has gained wide recognition and much support in the last few years. Really intelligent scientific study of the problem of the hydration PORTLAND CEMENT 135 of Portland cement has become possible only since the determination of the constitution of cement has been made by the Geophysical Laboratory. Such a study has recently (1914) been completed by Messrs. A. A. Klein, petrographer, and A. J. Phillips, former assistant chemist, of the Pitts- burgh Laboratory of the U. S. Bureau of Standards.* Hydration experiments were made upon the following compounds which had been found to be constituents of cement by the Geophysical Laboratory: Monocalcium aluminate (CaO, AI2O3). Monocalcium silicate (CaO, Si02). 5 : 3 calcium aluminate (5CaO, SAI2O3). Beta orthosilicate (2CaO, Si02). Tricalcium aluminate (3CaO, AI2O3). Gamma orthosihcate (2CaO, Si02). Tricalcium siUcate (3CaO, Si02). Limes burned at different temperatures, and ground to various degrees of fineness were also experimented upon, and the following commercial cements — a high silica cement, a low silica cement, a high iron cement, and a high magnesia cement. The tests consisted of (1) hydration on microscopic slides with water without access of air, (2) hydration with superheated steam in a cylinder, (3) hydration in an autoclave, and (4) molding with limited quantities of water, approximating those used in normal consistency mixes. Lime- water and plaster of Paris solution were also used as hydrating mediums. Petrographic methods were employed to determine the hydration proc- esses and final products. Tricalcium aluminate hydrated with' restricted amounts of water formed amorphous hydrated tricalcium aluminate very quickly, this amorphous form slowly changing to the crystalline form. With a large excess of water crystalline hydrated tricalcium aluminate was formed. " Monocalcium aluminate and 5 : 3 calcium aluminate split off amorphous hydrated alumina and form the crystalline hydrated tri- calcium aluminate." The hydration of the aluminates in lime-water re-^ealed no new products, but in plaster solution the additional compound tricalcium sulpho-aluminate, 3CaO, AI2O3, 3CaS04, a;H20, was observed. This * The following discussion 01 the problem of cement hydration is based upon an as yet unpublished abstract of the report which will later be pubUshed by the Govern- ment. This abstract has been privately communicated to the author by Mr. P. H. Bates of the Bureau of Standards, 136 MATERIALS OF CONSTRUCTION compound is crystalline in form and identical for all three aluminates. Its formation is only incidental in the retardation of the initial set caused by gypsum. The high temperature and pressure of the autoclave de- stroys it. <' Burned lime hydrates with an excess of water to either the crystalUne or amorphous form of lime hydrate. A preponderance of the former is produced where the hme is coarse and high burned, while the formation of the amorphous form is favored by fine grinding and low burning. " The monocalcium silicate and the gamma orthosilicate do not hydrate, while the beta form of the latter hydrates but shghtly with water after long periods. Lime-water and plaster solution do not mate- rially increase hydration, whereas a solution of the calcium aluminate gives the maximum hydration and best-appearing test pieces. The twenty-eight-day test pieces of beta orthosilicate and the aluminates, while exhibiting fairly good rigidity, have by no means the strength of corresponding neat cement briquettes. The aluminates are completely hydrated, but the beta orthosilicate shows only a comparatively shght hydration. The hydration product of the silicate is amorphous hydrated orthosilicate, there being no lime hydrate split off and no needles of hydrated monocalcium silicate formed, as noted by others. " The tricalcium sihcate hydrates readily and quickly with all con- centrations of water, the products of hydration being crystallized lime I and amorphous hydrated orthosilicate. Molded specimens set hard in five hours and show no disintegration after twenty-eight days in water. It has no favorable effect on the hydration of beta orthosilicate. Mix- tures of it and the aluminates show first the beginning of hydration of the aluminates followed shortly by the hydration of the silicate. Molded specimens of these are dense, hard, and strong, comparing very favor- ably with neat cement briquettes. " On the hydration of cement, the first constituent to react is the aluminate, with the formation of amorphous hydrated tricalcium alu- minate, with or without amorphous hydrated alumina. The sulpho- aluminate crystals are also formed and the low-burned or finely ground lime hydrates. This occurs within a few hours after the cement is gauged. The next compound to hydrate is the tricalcium silicate. This commences within twenty-four hours and is generally completely hydrated within seven days. Between seven and twenty-eight days, the amorphous aluminate commences to crystallize and the beta orthosilicate, the least reactive compound, begins to hydrate. The twenty-four-hour strengths are due mainly to the hydration of the aluminates and of any fine-grained, low-burned lime present. The large increase in strength between twenty- PORTLAND CEMENT 137 four hours and seven days is due mainly to tlie tricalcium silicate hydra- tion. 'The increase between seven and twenty-eight days is due to the hydration of the beta orthosilicate. Where there is a decrease in strength during this period it is due to the hydration of very high-burned free lime as in very high-burned, high-limed cements, or to the crystallization of the aluminates, as in high-alumina cements. The iron compounds in a cement are resistive to hydration. It does not form crystalline hydration products, but occurs as a rust-like material. " The initial set of cement is affected by the action of small amounts of electrolytes in retarding coagulation of the aluminate material. With a limited amount of water, such as used in normal consistency mixes, the aluminates coagulate and separate from supersaturated solutions as amorphous bodies, the rate of coagulation being affected by such small quantities of electrolyte as to nullify the possibility of the reaction being solely a chemical one. " Failure of cement in accelerated tests is due to the growth of large lime hydrate crystals. The disrupting action results from the pressure caused by growing crystals. Cement 3 will fail in the boiling test which contain lime sufficiently fine and high burned, so that during boiling it hydrates and crystallizes. The growth of crystals is sufiicient to cause disintegration. When a cement passes the boiling test but not the auto- clave test, it contains lime so coarse or high burned as not to hydrate in the boiUng test, but only in the autoclave, due to the high temperature and pressure employed. Some cements will pass either test only after aging. In this case aeration with insufficient water to allow solution and crystallization causes the lime to hydrate as amorphous hydrate, and in the accelerated tests there is no crystallization and no disin- tegration. " The reactions when cement is subjected to the autoclave test are not abnormal. The disintegration action attributed to the crystallization of the sulpho-aluminate has been exaggerated." (Note that the arguments of MichaeUs in support of the " colloidal theory " of setting and hardening do not conflict in any way with the explanation of the phenomena above given. Note also that the authors above quoted carefully avoid the use of the inexact term " colloid," which is so often used by MichaeHs.) Specific Gravity 108. Significance. Until lately the significance of the specific gravity of a cement was considered to be its usefulness " in detecting adulteration and under-burning." More recently, investigations have conclusively 138 MATERIALS OF CONSTRUCTION shown that the principal factor influencing specific gravity is the degree of seasoning of cement; that specific gravity tests will not detect under- burning; and that tests will only detect adulteration in the case of a few classes of adulterants, and even then only when the adulteration is very considerable. Its importance may, therefore, be considered only very limited 109. Specification and Results of Tests.* " The specific gravity of the cement thoroughly dried at 100° C. shall not be less than 3.10. Should the test of cement as received fall below this requirement a second test may be made upon a sample ignited at a low red heat. The loss in weight of the ignited cement shall not exceed 4 per cent." Experiments made at six different mills by members of the Associa- tion of American Cement Manufacturers gave an average of 3.14 for the specific gravity of the under-burned cements, and 3.18 for that of the hard-burned ones. Few freshly made American Portlands will be found outside these limits except under special circumstances as indicated in the discussion that follows : 110. Influence of Thoroughness of Burning on Specific Gravity. Until a few years ago, as above noted, the specific gravity of a cement was considered a valuable indication of under-burning. Experiments have shown, however, in every instance that the specific gravity of under- burned cements is only very slightly below that of normally burned cement and still, in the main, well within the specified limit. Experi- ments made by Meade f upon clinker with varying degrees of burning gave the following results: 1. Very soft under-burned clinker 3.208 2. Slightly under-burned clinker 3.222 3. Normally-burned clinker 3 . 214 4. Very hard-burned clinker 3 . 234 These values all appear very high in consequence of the fact that the clinker was ground and tested as quickly as possible after burning, thus preventing the usual lowering in specific gravity by the absorption of carbon dioxide and water from the air. 111. Influence of Adulteration on Specific Gravity. The substances most readily available and practicable for adulteration of Portlaiid cements in this country are natural cements, limestones, clay, slaked lime, slag, sand, and natural volcanic tufa. * Specifications for cement herein cited are, unless otherwise noted, those of the American Society for Testing Materials adopted Aug. 16, 1909. t Meade, " Portland Cement," p. 380. PORTLAND CEMENf 13d Natural cements will of course lower the specific gravity of the blend in direct proportion to the specific gravity of the natural cement used. Assuming an average natural cement to have a specific gravity of 2.95, and an average Portland 3.15, it will appear that a blend would have to contain about 25 per cent natural cement before the specification would bar its use. Cement rock and limestones having specific gravities in the neighbor- hood of 2.7 and 2.8 could not possibly be used without detection in per- centages exceeding 10 to 15 per cent. Clay or slaked lime would similarly be detected jf used in more than very small amounts. Slag, having a specific gravity of about 3.0, can be used in very large amounts without detection through excessive lowering of the specific gravity of the blend. Sand and volcanic tufa have, so far as we are aware, never been used except when the fact of the product being a blend is not concealed. The foimer, being more expensive to manufacture than pure Portland, must be able to show properties better than pure Portland in order to find a market. The latter has only been made use of to our knowledge in the plant operated by the city of Los Angeles in connection with the con- struction of the Los Angeles aqueduct. The specific gravity of an ignited sample of cement is invariably higher than that of the original sample, so that the retest provided for in the second clause of the specification will give no indication regarding adul- teration. The loss on ignition will serve to detect the presence of an adulterant only in the event of adulteration by addition of raw materials after calcination, the ignition loss of the other adulterants being in them- selves low. 112. Influence of Seasoning on Specific Gravity. It is a well-recog- nized fact that the absorption of carbon dioxide and water from the air, which begins as soon as the clinker leaves the kiln and continues as the clinker is ground and the cement stored, results in a material lowering of the specific gravity of cement. This process results in the formation of calcium carbonate, whose specific gravity is 2.70, and calcium hydroxide with a specific gravity of 2.08. If, for example, 2 per. cent of water is absorbed, the specific gravity might be lowered from 3.125 to about 3.06, and if a hke percentage of CO2 were absorbed, the specific gravity might be lowered from 3.125 to about 2.92. Meade * quotes the following tests, which show plainly the effect of seasoning upon specific gravity of cement. It will be noted that in each case the loss due to seasoning is regained upon ignition. * Meade, loo. cit., p. 382. 140 MATERIALS or CONSTRUCTION EFFECT OF SEASONING UPON SPECIFIC GRAVITY OF CEMENT Sample No. Condition. Dried at 100°. Not Dried. Dried at 100°- Not Dried. Dried at 100° Not Dried. Freshly made . . After 28 days. . After 6 months. Ignited After 6 months, Sample No 3.19 16 I 3.11 13 I 3.08 3.18 3.21 3.18 I 3 3.09 I 3 3.21 3.16 12 04 14 12 3.10 3.08 3.18 Condition. Dried at 100°. Not Dried. Dried at 100°. Not Dried. Ave. Dried at 100°. Not Dried. Freshly made . . After 28 days . . After 6 months Ignited After 6 months 3.15 3.12 I 3.09 3.09 I 3.03 3.15 3.20 3.14 I 3.08 3.09 I 3.04 3.19 3.182 3.148 I 3.100 3.104 I 3.054 3.182 113. Summary and Conclusions. Low specific gravity may be caused by adulteration in large amounts; it is not indicative of under-burning; it is indicative of the degree of seasoning. Since seasoning is in general considered desirable for all Portland Qements and absolutely necessary for some, it is not advisable to reject any cement upon the basis of failure to come up to the specified value of specific gravity unless the history of the cement and its manufacture is known. Fineness of Grinding 114. Significance. As above noted, the fact that the coarser par- ticles in cement constitute so much inert matter has long been recognized. Acceptance of the explanation by the Bureau of Standards of setting and hardening gives the importance of extremely fine grinding an entirely new significance. In general, " the more finely cement is pulverized, all other conditions being the same, the more sand it will carry and produce a mortar of a given strength." No other one detail in the manufacture of Portland cement has so great an influence upon all the properties of the product, and it seems reasonable to expect that the future improvement of Portland cement will PORTLAND CEMENT 141 be largely dependent upon the perfection of pulverizing machinery that will without excessive cost produce a cement having a maximum amount of extremely fine impalpable particles. 115. Specification and Results of Tests. " It shall leave by weight a residue of not more than 8 per cent on the No. 100, and not more than 25 per cent on the No. 200 sieve." Recent improvements in pulverizing machinery have made a great improvement in the fineness of commercial cements, and most of those now marketed in this country will leave a residue not exceeding 2 to 5 per cent on the No. 100, and 10 to 15 or 20 per cent on the No. 200 sieve. A great need is felt for the adoption of some standard method of determining the proportion of extremely fine particles or flour in cement. Many methods of separation by suspension in a liquid or in air have been brought out, but none has been standardized and each gives results not in accord with those of another method. It is impossible at this time, therefore, to determine except in a general way the effect of the pro- portion of material finer than the No. 200 sieve upon the properties of cement, or the actual proportions of material of a given size under the No. 200 sieve in commercial cements. * 116. Influence upon Soundness. It has usually been considered that increased fine grinding is operative in improving the soundness of a cement. The extent of the improvement is slight, however, and not such as to justify extremely fine grinding for the sake of improved sound- ness alone. Meade cites experiments which show that in some cases an unsound cement is made sound by grinding to an impalpable powder; in the majority of cases, however, soundness was attained only by seasoning following fine grinding, making it appear that the beneficial effect of fine grinding is only indirect, in that it affords additional opportunity for seasoning. 117. Influence upon Setting Time. In general, increased fineness of grinding has the effect of making a cement more quick setting. It appears that high-alumina — and consequently low-lime — cements have their set- ting time most affected by fine grinding, so that in general the higher a cement is in hme, the more finely it may be ground without reducing the setting time to too great an extent. Extremely fine grinding, or separation of the cement so that the flour is obtained, produces a cement whose set is almost instantaneous. These facts are exactly in accordance with the conclusion above quoted from the report of the Bureau of Standards upon the hydration of cement. * A method of determining the proportion of extremely fine particles in cement has recently been developed at the U. S. Bureau of Standards, and is now undergoing standardization (1914). 142 MATERIALS OF CONSTRUCTION Fig. 67, which is based upon data given by Meade,* illustrates well the effect of increased fineness upon the setting time. This curve averages the results obtained in tests of eight different cements, each being ground to six different degrees of fineness. 118. Influence upon Neat and Mortar Strength. Many experiments have shown conclusively that increased fine grinding of cement is not only not beneficial to neat strength, but even lowers the neat strength. On the other hand the sand-carrying capacity and mortar strength of a cement is very considerably increased by finer grinding. m 250 210 230 ~ — ~ ~ ~ — IKFLUENCE OF FINENESS OF GRINDING UPON SETTING TIME OF CEMENT (Meade) Curve ie average of tests of 8 cements 210 200 ■^ Ififl ^ ^ ■S140 " ISO S, Vn ■^ 120 ^ a 110 tw 100 \ \ S 90 s ^ H 70 s \ 50 ^ \ v A S 10 _^ _,_ _i_ 200-M«ih 03.9^paa3Dg 00-Mm1i 2ao.M<«h 2«l.UMh 37.1«puA.B 0O.M«h 95% piuudne 200.Ue5h 99.0^ piMiDg OO.M«.h 100^ pnmoE 1 2IN-UMh 100 !S paring lOO-Mwh r-t ^ \ <*. d M I \A ^ 1 X is«« ^m I sj He. at J :^o . |- 1 y-N Q. %>. 1 \ f" ^ f 600 1 i > ^ -^ ^ 1 1 -> 4-. .-p iSsoo I \ 1 1- pt» r . f^ -jj /■^ I 1 1 rf / «.„„ 1 ^•?l r. y. V% 1 / 1 t^ 1 J jlDe Up. P— -r c N !S 1 3; t\*s.^r^<. [T 1 / S^ i^ fr m P *. r* ^ r" H INFLUENCE OF FINENESS UPON NEAT AND MORTAR STRENGTH , if Me «^, ) _L 1 1 6 80 86 00 96 100 Perceot Passing a 200M:esh Sieve Fig. 67. 28 17 28 17 Age in Days Fig. 68. These facts again are entirely in accord with the Bureau of Standards' explanation ,of setting and hardening. Amorphous hydrates have Uttle mechanical strength in themselves, and the more finely a cement is ground the more nearly will the active particles constitute the entire mass of the cement to the exclusion of coarser inert matter. The amorphous hydrates have, however, a great binding power, and will form a strong mass if inert particles are present to be cemented together. These inert particles may either be the coarser portion of the clinker, or sand added in a mortar. In either case the greater the actual amount of impalpable powder present, the greater the quantity of inert matter that may be added and still produce a cement or mortar of a given strength. * Meade, loc. cit., p. 399. PORTLAND CEMENT 143 Fig. 68 shows admirably the effect of differing degrees of the fine grinding of the same cement, upon the neat and the mortar strength.* 119. Summary and Conclusions. The maximum degree of fineness compatible with reasonable manufacturing costs is desirable. The strength except in neat tests is greatly increased and the sand-carrying capacity and mortar strength very materially increased by increased fineness of grinding; the time of setting is materially shortened, but this effect may be lessened and injury in this repsect prevented by making the cement as high in lime as is possible without endangering soundness by the presence of excess free lime. Time of Setting 120. Significance. The rapidity with which a cement sets is simply a criterion by which the suitability of a cement for use under given con- ditions may be established. Absolutely no analogy can be traced between the rapidity with which a cement sets and the strength it will ultimately develop. A cement to be used in submarine construction, for instance, should be quick setting, while a cement to be used under circumstances where rapid handling and deposition in the forms without delay is impossible, should be slow setting. 121. Specification and Results of Tests. " It shall develop initial set in not less than thirty minutes, and must develop hard set in not less than one hour, nor more than ten hours." So many factors influence the time of setting, such as temperature, amount of water used in gauging, presence of sulphates, etc., that no general statement may be made as to the rate of setting of Portland cements. 122. Influence of Temperature. In general, the higher the tempera- ture the shorter the setting time will be. The diagram of Fig. 69, taken from Tetmajer's " Communications," Vol. 6, shows this effect very markedly. Many other series of tests, more recently made than Tetmajer's, agree very well with the older tests, and the latter are quoted here because they are more comprehensive than most later tests. (It should be noted that the cement used by Tetmajer was rather slower setting than the normal.) 123. Influence of the Percentage of Water Used to Gauge the Cement. The percentage of water used to gauge cement influences its setting time to a very marked degree, a wet mix setting much more slowly than a dry mix. It is on this account that tests of setting time are always made with a paste possessing a standard degree of plasticity (i.e., a normal consist- * Meade, loc. cit., p. 401. 144 MATERIALS OF CONSTEUCTION S35 30 O £26 S20 I 15 H 10 'E — — r- — — — — ~ "■ ~ ^ - *" •— * \ 1 EFFECT OF TEMPERATURE UPON . SETTING TIME OF PORTLAND CpHEf^T MORTAR (Tetmajer's "ComnmnicationB") ' (Vol.VI) 1 1 \ ' j s \ \ A *s. ^ ^ fe \ \ k s M \ •s s f- ■v & % ^ > ^ ^t " ^ -^ _^ .^ |~" ■*- -, _ ~" - J ^ ^ ±1 L 8 10 12 14 16 18 lime of Setilng-Hours FlO. 69. ancy mix). Fig. 70, plotted from tests quoted by Meade, shows the char- acteristic effect of variations in the amount of water used m gaugmg upon the time of setting. The degree of humidity of the air similarly affects the setting time and, for the sake of uni- formity in testing, the determination should al- ways be made in a moist closet. 124. Influence of Sul- phates. As noted in a previous chapter in dis- cussing the manufacture of Portland cement, the addition of lime sulphate to the clinker before grinding is absolutely necessary in order to re- tard the set sufficiently to pass the requirements of commercial use. The addition of plaster of Paris or gypsum up to 1 J to 3 per cent retards the set, and further addi- tions beyond this point of maximum retardation has the opposite effect. 125. Influence of Seasoning. The effect of seasoning upon the set- ting time of cement is the source of consider- able difficulty for the cement manufacturer. Freshly made cement which is found to be slow setting is frequent- ly found after a few weeks' seasoning to have become quick setting. Conversely, some cem- ents, originally quick setting, become slow setting after seasoning. The latter case is usually less serious than the former, as the cement will probably still be suitable for marketing. In the case of the cement which becomes quick setting upon storage the cause of the difficulty is apt to lie in the composition of the cement and may usually be remedied by increasing the lime content. ~ ^ " - ■" ~~ ~ ~ ~ r — ~ 24 p' ^ / ^ / r"' u^ 1 / y f y' f / £20 'A / o^> } H f^ u i EFFECT OF PERCENTAGE OF W.ATER UPON .SETTING TIWE - r *^1G / / CMeade p.416>. 10 f ^ i 14 \ 1 3 1. 5 ' 8 Time of SettiDg-Hours Fig. 70. PORTLAND CEMENT 145 126. Summary and Conclusions. The time of set is often an important consideration in the choice of a cement for a particular purpose, and on account of the effect of storage upon the setting time the test should preferably be made after dehvery of the cement on the work. The effect of temperature and the percentage of water used in mixing as well as the humidity of the air, is so marked that the determination of the setting time must always be made with extreme care under standardized con- ditions. Soundness 127. Significance. Soundness in a cement impUes the absence of those qualities which tend to destroy the strength and durabiUty of a cement. The importance of soundness is second to that of no other property of cement. If a cement is ultimately unable to withstand the disintegrating influence of the medium, air or water, in which it is placed, the development of a high degree of strength at the ages usually tested means less than nothing. Unsoundness is manifested by a lack of constancy of volume, dis- integration being caused almost entirely by expansion occurring after the cement has set and acquired a certain degree of inelasticity. Since any amorphous hydrate shrinks during drying and expands when wetted, it is evident that this behavior on the part of the amorphous hydrated constituents of gauged cements must cause shrinkage of neat cement in air and expansion in water. These changes in volume are very much lessened by mixing with inert material as in sand mortar, the degree of volume change being dependent upon the richness of the mortar. The consequence of the desiccation of the cement is the appearance of fine hair-cracks on the surface of cement or rich mortar used as a plaster or top coat. These fine hair-cracks should not be taken to be an indication of defective cement, but their appearance simply is an indication of the use of too rich a mixture. The subject of the expansion and contraction of cement and mortar from the above causes will be discussed further hereinafter. Unsoundness, as the term is commonly used, is not caused by the phenomena above discussed, but is due to disruptive action caused by crystallization of certain of its constituents, pressure being exerted by growing crystals. The principal constituent so involved is the lime present in the free state. The presence of free lime in the cement may be due to an excess of lime in the composition of the cement; failure to calcine at a temperature sufficient to combine all the Ume present with the sihca and alumina; or 146 MATEEIALS OF CONSTRUCTION failure to grind the raw materials sufficiently fine and mix sufficiently well for the hme to enter combination. If this lime were simply present in its usual condition as an amorphous substance it could never be held respon- sible for the phenomena exhibited by an unsound cement, because it would become hydrated immediately upon contact with the gauging water and its expansion would be harmless because it would have taken place before setting had begun. Two explanations have been advanced to account for the slowness with which hydration does take place: one theory sup- poses that the free amorphous lime particles become coated with clinker and are so protected and hydration delayed; the other theory is based upon the fact discovered by Day, Shepherd, and Wright in the researches conducted at the Geophysical Laboratory, that loose powdered lime, if kept for any length of time at 1400° C, agglomerates to form crystals which grow rapidly. These finely powdered crystals when mixed with water appear at first to be inert, but after a time the hydration takes place with explosive violence. Both theories appear to have some verification in the observed behavior of unsound cement, and it is probable that both free amorphous lime and free crystalline lime are present. The presence of excess dehydrated magnesia may less frequently be the cause of unsoundness. In this event unsoundness will be observed only after a much longer period, since magnesia which has been highly heated remains inert for a long time before undergoing hydration. A long series of experiments following the discovery that the cement involved in several disastrous failures of structures in Europe were very highly magnesian, proved that the addition to cement of highly calcined magnesia produced enormous expansion, while fight calcined magnesia was without influence on soundness because of the rapidity with which it became hydrated. In recognition of the effect of strongly calcined magnesia upon the sound- ness of cement the standard specifications of all countries fimit the amount of magnesia, the English to 3 per cent, the United States to 4 per cent, and the German to 5 per cent. The presence of excess sulphates is also thought to be the cause of unsoundness in some few cases. The expansion is not due in this case to the hydration of lime sulphate, but is attributed to the formation of cal- cium-sulpho-aluminate, which is dangerous only in large quantities. The standard specifications of England limit the SO3 to 2.75 per cent, the German to 2.5 per cent, and the United States to 1.75 per cent. 128. Specification. " Pats of neat cement about 3 inches in diam- eter, i inch thick at the center, and tapering to a thin edge, shall be kept in moist air for a period of twenty-four hours. " (a) A pat is then kept in air at a normal temperature and observed at intervals for at least twenty-eight days. PORTLAND CEMENT 147 " (b) Another pat is kept in water maintained as near 70° F. as practicable, and observed at intervals for at least twenty- eight days. " (c) A third pat is exposed in any convenient way in an atmos- phere of steam, above boiling water, in a loosely closed vessel for five hours. " These pats to satisfactorily pass the requirements shall remain firm and hard, and show no signs of distortion, checking, cracking, or disin- tegrating." 129. Influence of Seasoning. Since, as above noted, unsoundness is primarily due to presence of free lime, either amorphus or crystalline, in cement, exposure to the air will in time produce soundness by the conver- sion of the lime into carbonate of lime or lime hydrate through the agency of the carbon dioxide and moisture always present in the air. Meade quotes tests made on several samples of cement originally unsound, rendered sound by simple aeration in paper bags in the laboratory. Several cements which checked more or less badly when tested freshly ground, became sound upon seasoning from one day to two weeks. One which partly disintegrated when tested freshly ground, became sound after twenty-eight days' aeration, and another which entirely disintegrated when first tested, became sound after ninety days' aeration. 130. Influence of Fineness. As noted above, free amorphous lime may probably be so coated with clinker as to be protected from aeration to a large extent, and similarly protected from the water used in gauging for some time, with resultant ultimate unsoundness if such particles of clinker-coated lime are not broken up by fine grinding. Increased fine grinding therefore will often promote soundness in a cement that will not season sound without fine grinding, but the fine grinding should prob- ably be regarded as beneficial only indirectly, in that it makes possible more complete aeration. 131. Effect of Sulphates. The primary effect of sulphates added to cements is greatly to delay the time of setting. Their addition in small percentages does result in appreciable improvement in soundness, as determined in the steam test, but it is probable that here again the effect is only indirectly due to the sulphates, the set being delayed by the latter till the free lime has become hydrated with no resultant evidence of un- soundness. This view is strengthened by the fact that many cements which successfully pass the steam test fail in the air test. The possible effect of excess sulphates prejudicial to soundness has been discussed above. 132. Summary and Conclusions. Soundness is the one most essen- tial property of cement. Its absence is manifested by cracking and disintegration after cement has set, due to the disruptive force caused 148 MATEEIALS OF CONSTRUCTION principally by the expansion of certain of its constituents, notably free lime, upon becoming hydrated in the crystalline form. Soundness is promoted by thorough seasoning, by fine grinding of the raw material and the clinker, by keeping the magnesia content low, and by not exceeding the content of sulphates necessary to retard the set sufficiently. Tensile Strength 133. Significance. The tensile strength of cement is in itself of very little importance, because cements are rarely depended upon to with- stand tensile stresses. The significance of tensile strength as revealed by laboratory tests is therefore limited entirely to the degree of accuracy in the assumption that there exists a fixed relation between tensile strength and compressive strength, soundness, and other properties which contribute to the satisfactoriness with which it meets the demands put upon it as a material of construction. The assumption is also often made that the tensile strength of neat cement bears a more or less close relation to the strength of mortars under the same character of stress. It will be shown that a relation between tensile strength and com- pressive strength can be estabUshed, but that it is by no means a constant relation at all ages, and that it also varies greatly with different cements, and with different mixtures; that there is absolutely no ground for con- cluding that a high tensile test of cement at the usual age of testing indicates soundness; often, in fact, it affords ground for the opposite conclusion; and that the neat tensile strength is no indication of the strength of the cement in mortars and concrete. In view of the above considerations there remains but one defense of the use of, tensile strength test results as a criterion of suitability— the fact that tensile tests are easily made with inexpensive equipment, as compared with the requirements for compressive strength tests. Even this consideration does not justify the dependence that is commonly placed upon the neat tensile strength, since the infinitely more important mortar tests, involving the determination of the power of the cement to adhere to the surfaces of foreign particles, may be made with equal facility. The latest specifications of the American Society of Civil Engineers* include specifications for the making of compressive strength tests. The tensile strength test has been included, however, both for neat cement and mortar mixtures. The new German specifications adopted by the Association of German Portland Cement Manufacturers, March, 1910, have abandoned the neat tests .altogether, and have substituted corn- Final Report of the Special Committee on Tests of Cement." Transactions, Am, Soc. C.E., Vol. 75, p. 665, Dec, 1912. PORTLAND CEMENT 149 pressive tests for tensile tests, except that the making of tensile mortar tests on the work is provided for if required for the sake of expediency, no rehance being placed upon these field tests, however, in case of dispute. 134. Specification and Results of Tests. Neat Cement. " The min- imum requirements for tensile strength for briquettes 1 inch square in cross-section shall be as follows, and the cement shall show no retro- gression in strength within the periods specified: " Age. strength. 24 hrs. in most air 175 pounds 7 days (24 hrs. in moist air, 6 days in water) 600 pounds 28 days (24 hrs. in moist air, 27 days in water) 600 pounds Tests of tensile strength are dependent upon a great many factors which influence the accuracy of the results. Perhaps the greatest of these is the personal equation, which plays an important part in every operation in mak- ing tests of tensile strength. In addition to the disturbing personal factor, tensile strength is greatly influenced by the amount and the temperature of gauging water, the method of mixing and molding in general, the tempera- ture and humidity of the air, the form of bri- quette, the design of the testing machine and the grips, and the manner of storage of the briquettes prior to testing. All of these factors have to do Tvith testing operations. The actual tensile strength, the condi- tions of testing being considered constant, is dependent primarily upon the composition of the cement, particularly the content of bases, alumina, and iron oxide; secondarily, upon the temperature of burning, and the fineness of grinding. Fig. 71 shows the average tensile strength, neat, of seven representative brands of Portland cement tested at the Struc- tural Materials Laboratory in St. Louis,* 1905-07. 135. Influence of Lime Proportion on Tensile Strength. An increase in the proportion of hme yields a stronger cement until a point is passed * Bulletin 331, U. S. Geol. Survey. - y\ f — -o^ ^ % ' ^ D- ( Y '^ -S-800 / 4 - - rENSIlE.STBENGTH OF 7 PORTLAKD CEMENTS NEAT (Eaeh ReBult the Ave, of 30 Tests) (gurve is Average of 310 Tests) Struct.Materials Lab. Tests i,^ Ph L p c tfj 1 ' 200 j_ _ _ _ -1- L L L L L LI T as ISO Age Id DayB Fig. 71. 150 MATERIALS OF CONSTRUCTION beyond which further additions mean the presence of excess free lime with consequent unsoundness. Cement prepared in the modern rotary kiln is capable of holding a rather larger proportion of lime with safety than that made by the older processes where neither so high nor so uniformly distributed temperatures were attainable. High-limed cements which are still perfectly sound are slow setting, but attain their maximum strength early, sometimes showing retrogression in tensile strength after as short a period as seven days and usually showing retrogression after longer periods. This retrogression is not noted, however, in compression tests, and it appears possible that the low tensile test results ace due to the brittle- ness of such cements and the difficulties in the way of so gripping the briquettes as to make the stress a pure tensile stress. 136. Influence of Temperature of Burning on Tensile Strength. The temperature of burning affects the tensile strength chiefly indirectly, in that the temperature of burning must be high for high-Umed cements and therefore high for cements of high tensile strength. Under-burning will result in weak cement, but over-burning probably has no injurious effects so far as the quality of the product is concerned, though it means a lack of economy in kiln fuel and extra expense in grinding. 137. Influence of Fineness of Grinding upon Tensile Strength. This question has been discussed above under the head of " Fineness, — Influence upon strength and sand-carrying capacity," where it is shown that, owing to the amorphous hydrate formation in gauging and the nature of these hydrates, increased fine grinding is not beneficial to neat tensile strength, but materially improves the strength of mortars in that. the additional opportunity for cementing action afforded by fine grinding greatly increases the cohesive power of a cement. Tensile Strength of Sand-cement Mortars 138. Significance. The remarks made above regarding the significance of tensile tests of neat cement largely cover the subject of tensile tests of mortars. If tensile tests of cements must be depended upon, the bri- quettes should at least be mortar briquettes, since in the neat tests one of the most important properties of cement, the ability to adhere to the surfaces of foreign particles and masses, is not determined. The remarks made above concerning tensile vs. compressive tests of neat cement apply equally well to the consideration of mortars, as do also the remarks concerning the factors influencing the results of tensile tests, and the discussion of the influence of the hme ratio, temperature of burning, and fineness of grinding upon tensile strength. PORTLAND CEMENT 151 139. Specification. " The minimum requirements for tensile strength for briquettes 1 inch square in cross-section shall be as follows, and the cement shall show no retrogression in strength within the periods specified." ONE PART CEMENT, THREE PARTS STANDARD OTTAWA SAND Age. Strength. 7 days (1 day in moist air, 6 days in water) 200 pounds 28 days (1 day in moist air, 27 days in water) 275 pounds , 140. Standard Sand. The following specification has been adopted (February, 1912) by the American Society of Civil Engineers. Its provisions are, however, practically identical with the standard specifica- tion in force since 1904, when the use of crushed quartz was abandoned. " The sand to be used should be natural sand from Ottawa, 111., screened to pass a No. 20 sieve and retained on a No. 30 sieve. The sieves should be at least 8 inches in diameter; the wire cloth should be of brass wire and should conform to the following requirements: No. of Sieve. Diameter of Wire. Inch. Meahes per Linear Inch. Warp. Woof. 20 30 0.016 to 0.017 0.011 to ,0.012 19.5 to 20.5 29. 5 to 30.5 19 to 21 28.5 to 31.5 " Sand which has passed the No. 20 sieve is standard when not more than 5 grams passes the No. 30 sieve in one minute of continuous sifting of a 500-gram sample." The average tensile strength in 1 : 3 stand- ard mortar of the same seven brands of ce- ment whose average neat strength was noted in Art. 134 and Fig. 71, is shown by the curve of Fig. 72, which has been plotted from data ob- tained at the Structural Materials Laboratory in Fig. 72. St. Louis. 141. Effect of Fineness of Sand upon Mortar Strength. The size and granulometric composition of a sand is well known to have a marked '" "" ^ "" S ~ ~ \ — ■*" ■* ■* / s ■^ 'v 460 / ^ ^ "^.400 ? / ^ ^ * ■^ >350 ' -0 152 MATEEIALS OF CONSTKUCTION 100 EFFECT OF FINENESS OF SAND UPON TENSILE STRENGTH 1 :1 MORTAR (Tests of R.F.DaviB)'- 180 Ago ui Days Fig. 73. influence upon the strength of mortars. It will later be shown that the strength of mortars is directly proportional to the density for a mortar of given volumetric composition. The relation between fineness of sand and tejisile strength of mortars is excellently shown in Figs. 73, 74 and 75, which are plotted from tests made on 1:1, 1 : 2, and 1 : 3 mortars in the labo- ratories of the College of Civil Engineering, Cornell University, by Dr. R. P. Davis.* For these tests a beach sand was used, composed of nearly pure quartz with slight traces of mica. The various sizes of sand ex- perimented with were syn- thetically prepared from the natural sizes of the above- mentioned sand. Mortars were made of fifteen different granulometric compositions, but to avoid confusion on the curve sheets the results obtained with the three natural sizes and some of the blends of equal amounts of various sizes have not been included on the curve sheets. It will be noted that for the three mortars tested the results are remarkably aUke. In every case the sand between the No. 20 and No. 30 sieves gave a mortar of maximum strength at all ages except the short-time tests and the three-months' test in the case of the 1:3 mortar. The sand of all sizes finer than that passing the No. 10 sieve ranked second except for the short- time tests, and exceeded the 20 to 30 sand at three months in the 1 : 3 mortar. The blend of equal amounts of 10 to 20 and 20 to 30 sand ranked third and (except for the 1 : 1 mortar where the 30 to 40 sand *Dr. R. P. Davis, "ComeU Civdl Engineer," Vol. 19, pp. 114-124, Jan.. 1911. EFFECT OF FINENESS OF 8AN0 UPON TENSILE STRENGTH 1 :2 MORTAR (Tests of R. P. Davis) ISO AgQ in J>a;s Fig. 74. PORTLAND CEMENT 153 ^ '~ — ■~" 10_ 7 1 "~ ~~ ■~ ~~ * so. w /. < — 1— _ ^ - in^: ^ ^ (20-30) (^ ^^^ =S = #t ^m _ y ^ ^ -(^-i-)- _ _ _ "TiVr- 4 200 — w / ^ UO-iO' — — ~ ~ zz. — =- = (3Uj|o) - - ^240 lu — UoM = u\/l '? (50-G(J) .. ^ =^ =q h*= UJ ■y =*n rjTjlL ■S"" / i I ^ — ' — ' — Tsr kb 1 |200 / 1— ■if\^ W (i ^ is CO 180 1 EFFECT OF FINENESS OF SAND UPON TENSILE STRENGTH 1 13 MORTAR (tests of R.P.DAVIS) 1 §160 ^140 1 1 180 Age in Days Fig. 75. excelled it) the 10 to 20 sand ranked forth. All of the finer si?es of sand ranked lower in the order of their fineness. 142. Relation be- tween Density of Mortars and Tensile Strength. The series of tests above quoted affords important sup- port to the theory generally held that the tensile strength of a mortar of given volumetric composi- tion is proportional to the density of the mortar. Figs. 76 and 77 show the fairly constant relation between density and mortar strength at all ages for 1 : 2 and 1 : 3 mortars, established by Davis' tests. Fig. 78 shows the relation between den- sity and average mortar strength at the four periods tested for 1:1, 1 : 2, and 1 : 3 mortars. 143. Influence of Mica in Sand upon Tensile Strength of Mortars. Davis in the above-mentioned arti- cles quotes tests made by Mr. W. N. Willis * showing the detrimental effect of the presence of even small percentages of mica added to 1 : 3 mortars made with standard Otta- wa sand. This information is pre- sented by the curves of Fig. 79. " This loss of strength, which is very considerable even if there is only a slight amount of mica present, averaging about 25 per cent with only 2J per cent of mica, is due to two causes : first, because of its irregular shape the percentage of voids is very large, and second, on account of its smooth surface good bonding will not obtain.", * W. N. Willis, " Engineering News," Vol. 49, p. 145. " ■ T 1 1 1 1 1 1 M 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 IT ■ ■ - - RELATtON OF DENSITy TO TENSILE STRENGTH - - 1 :2 MORTAR (Testa of R.P. Davie) o j3 ::::::::;: ific X 70 72 Density 74 76 Fig. 76. 154 MATERIALS OF CONSTRUCTION 144. Influence of Cleanliness of Sand upon Strength. Sand used for structural purposes is never really clean unless washed on the work, and — m in r I'l 1 1 1 1 1 1 1 - - RELATION OF DENSITY :. i:3h TO TENSILE eTRENGTH - - ORTAR : : ^^1.1 . . . J_ ^ ^ J« 1, . . 250 if H 1 1 00 68 70 72 74 76 Density Fig. 77. -■ 1 1 1 1 1 1 1 1 1 1 f 1 1 i '1 rrr -- RELATION OF DENSITY TO TENSILE STRENGTH LJ- AVEBAGE FOR ALL AGES Pf 1 :::::::::::::::: ^ 300 1 M M M M 1 1 M 1 1 M Mill 70 72 74 76 Density Fig. 78. the importance of this factor is only just coming to be fully recognized in concrete construction. The character of the impurities in sand is more important than the amount. If largely vegetable loam, the danger is in direct pro- portion to the amount of organic matter in the loam or silt. Ex- periments made by Mr. Sanford E. Thompson * led to the conclusion that to be injurious the organic matter must constitute over 10 per cent of the silt and at the same time over 0.1 Tf , , . .,. . per cent of the sand. If the impurities in sand are not of a vegetable loam nature, they are of a mineral character, as clay. The effect of clay is dependent entirety * Sanford E. Thompson, Trans. Am. Soc. C.E., Vol. 65. 6 10 15 20 Percentage of Mica in Terms of Weiglit of Sand Fig. 79. PORTLAND CEMENT 155 upon its state of subdivision and the uniformity with which it is distrib- uted through the sand. The effect is probably not purely physical, but due rather to the distinctly colloidal properties of the clay. In most laboratory tests the addition of clay has been found beneficial to strength, or harmless up to a certain limit. This hmit has .usually been found to be from 5 to 10 per cent of the sand, the beneficial effect not being very marked, however. The above results attained in laboratory tests should not be taken to indicate that similar percentages of clay will be beneficial or harmless in naturally clayey sands used in practice for concrete or mortars. The manner of distribution and degree of fineness of the clay will be the deter- mining factors, and the amount permissi- ble will in general not approach the above limits. 145. Effect of Ad- dition of Hydrated Lime to Cement and Mortars. The addi- tion of hydrated lime in small percentages to cement and mor- tars usually has a very appreciable effect upon the resultant tensile strength shown by tests. The curves of Fig. 80 well illustrate the injurious effect in the case of neat cement and the slightly beneficial effect, in amounts up to 15 per cent, in the case of 1 : 3 mortar using standard Ottawa sand. These curves are based upon tests made by Mr. Harry Gardner.* Investigations of the effect of additions of lime hydrate upon the strength of cement mortars have not invariably shown that additional mortar strength is attained by so doing. Tests made by Mr. H. S. Spack- man f indicate that the addition or substitution of 10 per cent of hydrated lime has no marked effect upon either the tensile or compressive strength of mortars. The idea has been advanced by Mr. Spackman that the addition of the hydrate has a beneficial effect pn the strength of mortars mixed and useH on the work, which is not exhibited by laboratory testa. The hydrate * Harry Gardner, Eng. Rec, Vol. 64, p. 309. t H. S. Spackman, "Concrete Cement Age," Vol. 4, p. 112. r n 850 ON TENSILE STRENGTH OF PORTLAND CEMENT AND MORTARS, ( Gardner) ^ ""ITSO s^ > \ \ \ s ■^ ^ 1- P. GOO ^*550 N S <; -Jt'. ., \ \ f^- •^ k, •^ fia "> ^^ L^ 3 450 g'400 £350 w, , ^ ^ jr ^ ^ 1 r~ -J ^ L, _ _ _ ^^ ^ =& E^ ^ ^ _ -- ::= _ _ msoo c- ^ =1 =- -^ ■= n -^ Z ^ - - "200 "150 ■^ _, ^ Laj. L. -_ ' "^ r- — — 50 % Lime Hydrate 5 10 15 20 a5 30 i Cement 96 90 86 SO 75 70 Fig. 80. 156 MATERIALS OF CONSTRUCTION does not by chemical combination aiTect the process of hardening and gaining strength, but its use is alone justified by the mechanical work it performs during the few moments between gauging and final placing in the work — the period of mixing and handling the mixture. The addi- tion of hydrate makes a fat, viscous mortar in which the sand and cement will not separate to as great an extent as they will when Portland cement is used with sand alone. This tends toward the production of a mixture of greater uniformity and with less voids, therefore securing a mortar of more uniform strength. Under the ideal conditions of mixing and molding in the laboratory this mechanical advantage would not be noted. Aside from the effect of hydrated lime additions upon the strength of cement mortars, the practice of making such additions is often justi- fied by the advantage derived from the standpoint of permeabiUty. Hydrated hme is an excellent waterproofing substance for incorporation in mortars and concrete. Such additions also produce mortar and con- crete which show less expansion and contraction with alternate increase and decrease of moisture content. 146. Relation between Tensile and Compressive Strength. The existence of a more or less definite and constant relation between the ten- sile strength and the compressive strength of cement has been often asserted, and the truth of the assertion widely accepted. In fact such a relation is the only basis of justification of the adoption of tensile strength as a criterion of the suitability of a material for use in situations where it will be subjected only to compressive stress. In order to study the question and make a comparison between the results obtained with representative modern American Portland cements, tested in accordance with the present standards of testing, and the oldei studies based upon tests with European cements made twenty years ago by the methods then in vogue, the tests made in the Structural Materials Laboratory at St. Louis, Mo., in 1905-1907 * have been selected as the most comprehensive series of tests recently made. Seven prominent representative brands of American Portland cement were selectedfor these tests, and the manner of testing was in accord with the present standard methods (1914). Each brand of cement was tested in tension and in compression at ages up to one year, ten tests of each kind being made for each age with each cement, both neat and in standard 1 : 3 mortar. The tension tests were made with standard 1-inch square briquettes in a semi-automatic briquette-testing machine, and the compressive tests were made upon 2-inch cubes in hand-operated screw-testing machines. The sand used in the mortars was standard * Bulletin 331, U. S. Geol. Survey. PORTLAND CEMENT 157 — — ~~ — — — "" ~ ~" "■ ■~ "■ ■" ^ ^ " fl 16 .a *^ -^ ' ^ ^ ^ 1 S r ^ o .^ m ^ •^ ft y^ ^ ^ ^ TT ^ Y o '-' / RATIO OF COMPRESSIVE TO TEHSILE STRENGTH 1 NEAT'pORTLABiO CEMENT (Bach Eeafilt.iB Avera^ of SO Teste of One Brind) (CUrve is Average of '7 Brands) - |e " 4 1 t 7 ' jn 90 i 80 27 1 360 Age Id Days Fig. 81. Ottawa sand. From the data tabulated in the report above mentioned the curves shown in Figs. 81 and 82 have been plotted. A study of these curves shows that while the average curve represent- ing the ratio of com- pressive to tensile strength for the 210 tests of the seven brands is a smooth curve, showing a rapid increase in the ratio up to twenty-eight days, and a fairly uni- form increase between twenty-eight days and one year for both neat and 1 : 3 mixtures, the individual brands depart widely from the average. At seven and twenty-eight days this individual departure from the average amounts to about 12 per cent each way for neat cement. Individual brands among the 1 : 3 mortars depart from the average ratio almost 40 per cent at twenty - eight days, and 15 to 20 per cent at one year. Comparing the curve for neat cement with the curve for 1:3 mortar we find that the ratio is very different in the two cases, the ratio for neat cement having a value at any period about 1.8 times that found in the case of the 1 : 3 standard mortars at the corre- sponding period. It appears therefore that tensile strength of neat cement mixtures are not comparable with compressive strength upon the same basis with 1 : 3 mortars. A similar series of tests made with a blend of the above seven brands 18 ~ ^ ~~ ^ "" tc RATIO OF COMPRESSIVE TO T^NSIUE STHENpTH 1:8 STANDARD fORTUAND OElrfENT MORTAR (Each Heenlt i& Averaffe of 30 lEeste of Diie Brand} (Carve is Average of 7 Brands) fS f f S §10 8 £ ■a > S \ *; S . ^ 3 ! 8 — ' 1- is "~ ^ -8- « 4 S' o i i 1 7 21 S 90 A ge 1 I >aj 8 -70 36 a Fig. 82. 158 MATERIALS OF CONSTEUCTION 18 ~ s ^16 |14 iio H 8 -- s ^ " § -' . ^ t- ^ » n a RATIO OF' COMPRESSIVE TO TENSILE STRENGTH BLEND OF 7 PORTLAND CEMENTS NEAT CEMENT (each result the AVER4GE OF 30 TESTS) (CURVE IS AVERAGE OF 240 TESTS) i .2 6 1., ^l 8 U 12 U 1(J 20 r 24 28 32 36 of cements and twenty-two different natural sands was also made at the Structural Materials Laboratory and reported in the above-mentioned paper. The tabulated data there presented have been summarized, and the ratios of compres- sive to tensile strength plotted in Figs. 83 and 84. In Fig. 83 each result is the average of thirty tests, the curve therefore representing the mean of 240 tests of each kind. Fig. 84 represents the values obtained in 1:3 mortar mixtures of the above blend of seven cements with twenty-two different sands. Since the results for each sand, if individually plotted, would cause confusion on the diagram, only the average for all the mortars, the average for the five mortars showing the highest value of the ratio, and the average for the five mortars showing the lowest value of the ratio are plotted. The value of the ratio found for indi- vidual mortars in many cases far ex- ceeded the departure of the high- and low- Age in Dayjs Fig. 83. ~*~ "■ "■ ~- — — — — — r |12 olO §8 H 6 _ ■a -. — - — — _ — — — _ _ _ _ _ ~ — — — ::; ^ A» = Jf^ )- Hii hf st- - - - ~ - - ■s ^ ^ t- — — — A-v e> 3^ B. Sa id _ _ _ n n _ ^ u 1 ^ — "* — — — — — — — - -^ - a o T ^ — — L =, A V€ o \h Lt =-1 L B.t, _ _ _ _ _ ^ / A Ti' — ~ — — — - - — -^ H - - - - O 4 / Kfl 1 lO OF COMPRESSIVE TO TENSILE STR ~ 1 :3 MORTAR USING BLEND OF 7 PORTLAND CEMENTS AND 53 & ^ _ __ 22 NATURAL SANDS • 1 ... — — — — 4 i 8 9 12 » Ifi 20 u 28 32 » 36 Fig. 84. average curves from the average curve for all the mortars. The curves of Figs. 83 and 84 are here introduced simply to further support the opinion that the ratio of compressive to tensile strength of cement mixtures, neat and mortar, is dependent upon so many factors that the tensile strength cannot in general be taken to be more than a very approximate indication of the probable compressive strength of the PORTLAND CEMENT 159 same cement and even then the age, the mixture, the sand used in mor- tars, etc., are factors which must be carefully taken account of. 147. Tensile Strength and Soundness. The fallacy of the belief, often held, that a cement which shows up well in the early neat tests of tensile strength is not apt to fail to meet satisfactorily the conditions of structural use, through lack of constancy of volume, is well illustrated by Fig. 85, which is based upon tests made by Mr. W. P. Taylor.* These curves are derived from over two hundred nearly consecutive tests of a single brand of Portland cement, one hundred of them faihng in the soundness test, and one hundred passing. It will be noted that the early strength of the neat tests of those samples faihng to pass the test is much the greater, while the oppo- site is true in the case of the mor- tar samples. These data therefore afford an additional bit of evidence to show that mortar tests should be preferred to neat cement tests. 148. Relationbetween Neat and Mortar Strength. The relation between neat strength and stand- ard 1 : 3 mortar strength in both tension and compression is shown by Figs. 86 and 87, which are based upon the series of tests quoted above in discussing tens\le vs. com- pressive strength., It has been stated above that the mortar strength or sand-carrying capacity of a cement is dependent largely upon the proportion of extremely fine particles in a cement. It is therefore not surprising to find that in the series of tests of seven Portland cements from which the curves of Figs. 86 and 87 are plotted, no definite relation can be estabhshed between neat and 1 : 3 standard mortar strength. The seven cements probably varied greatly in the percentage of flour present, although the facts could not be discovered by the ordinary sieve tests which were made. Only two general conclusions may be based upon the curves of Figs. 86 and 87: first, that the ratio of neat to 1 : 3 mortar strength in com- pression is about 1.8 times as large as the ratio of neat to 1 : 3 mortar strength in tension, and second, that the value of this ratio in both * W. P. Taylor, Trans. Am. Soc. for Test. Matrls., Vol. 3, 1903. 1 1 1 1 1 1 1 1 1 1 1 RELATION BETWEEN TENSfLE STRENGTH AND SOUNDNESS (W.P.Taylor) 900 a soo p. ^ ~- ^ ^ " (re rrr id_ £1 cnti '- Ul 700 H - ^ ^ rei r— pf ' — L ns undo me nta ^.500 / i i 1 glOO 1 ' ^300 y.-i Mo tar' St enRth Soi Dd,Ce lentB ^ 1 — * nei ta r- §200 H 100 5= '^ -^ ■^3t «u -u nso iStfCe S 60 Age in Days Fig. 85. 160 MATERIALS OF CONSTRUCTION compression and tension rapidly decreases for about three months, after which no pronounced further decrease is observed. 1 8 12 16-2024^ 32 3640444852 Fig. 86. i 8 121620 24 2832 3640 4448 52 Age Id Weeiss Fig. 87. isa 28.90 90.180 Age Interval, Days^ Fig. 88. 180.360 "" ~" "~" ■~ — — RATE OF INCREASE IN TENSILE STRENOTh OF STANDARD 1:3 MORTAR (EacH Result the Aye. of 9-70 Teats. " ~ ~ 160 L Wo , 1 ■ ^ 1 ' trJ H 1 •v ^ ~ / ~l «— 'U t 7 <1fl v; ^ i! I --e ■^ •V clOO / / ^ ^ - / „ ■^ ^ M so y -f •^ ■M ."•, B tt ^ ^ 1 ■v ^ ' ^ Sfio / *A jO at rt. -V s _ / l ^ «s ^ s S 40 __ / / ^ V, ^ a V V ^ 20 / ~ ~~ ■^ V ~ ~ ~ — 7-, m u- 90 M fs 9 Id 0. te 80 rv! n. Da ys 1 ». S60 Fig. 89. 149. Rate of Increase in Tensile Strength and Loss of Strength Observed in Long time Tests. Figs. 88 and 89 represent the rate of increase in tensile strength, in neat and standard 1 : 3 mortar mixtures, POETLAND CEMENT 161 for the seven representative American Portland cements used in the Struc- tural Materials Laboratory series of tests above mentioned. It will be noted that for these typical cements the rate of gain in tensile strength at all ages is approximately inversely proportional to the strength at seven days, those cements which show the lowest tensile strength at seven days maintaining the best rate of gain in strength at all ages. Figs. 90 and 91 bring out the same facts in connection with the rate of gain in compressive strength of the same cements. Specifications which require given percentages of increase in strength between the seven-day and the twenty-eight day period seem therefore to be justified, 500 n " ■" - - RATE OF INCREASE JN COMPRESSIVE STRENGTH OF NEAT CEMENT. (Kach'Besuit the Ave. of 20-70 Tebte ) 400 o t. a Vw jC d& 1 rtir jJlO / i -^ ^ > "ta J CJ 1 a: to- -!!■ ^ |200 cl80 Sl60 / « a* / — / 1 M sq. n. i_ ^ - ■ f'*19f. / ^ --] *■ „ -da — 1 12 100 80 60 40 20 / y - 5« m *■ aq ^ _^ L-l -- s K /b( S & 21 ^ 'A — ■" . a ■nq Ifi at- -di JS- - ^ Ofe ■^ — — ^ ~ ^ L 7.2S 28.00 90-180 Age Interral, DayA. Fig. 90. 180.360 500 ^ ~ ~ — ~ n "~ - IBATE OF INCREASE IN COMPRESSIVE 6TBENaTH OF STANPARD 1 :3 MORTAR < (Each Result the'kve. ot 20-70 Testa y ^■\ 400 ; / IS i 900O /' 8 i 8000 S 7000 y I COMpReseiVE STRENGTH OF T-PORTLAND CEMENTS, NEAT. _ tEach Result the Ave. of SO Teats " (CurvB IB Ave. of 210 Testa) 1 £ 6000 Id 1 4000 2000 1000 1 1 1 ~ " -1 "■ - ~i ■~ ^ ^ n "■ 4000 3800 } ' .-- O3100 ;J3200 ■i?^3000 »2800 Hi 2600 S2400 O ■^ a ^ L--- 1 >- J ■ / ° / / " gaooo S 1800 giooo ■a 1400 / , ^ CpMPRESSiVE STRENGTH OF . ' 7 PORTLAND CEMENT M0RTAB6.1 1 13 STANDARD iJlORTAR.i ■ (Each Result: the Ave. of 30 Teats) ^ ( Curve is Ave. of 210 Te'sts ) / 1 J 8 1 giooo § 800 O 600 400 200 fl I L. J,. 1 180 Age In Days, Fig. 92. 90 ISO Age 1q Days. Fig. 93. that the modulus is not a constant for a given specimen, but decreases rapidly with increased load. In general the modulus is ' found to be lower as the mixture becomes leaner, and for a given mix- ture increases with age. Fig. 94 gives values of the modu- lus determined by tests made at the age of three months for a single brand of cement in several mixtures. The values here given are slight- ly higher than the average cements and mortars show at this age. 164. Shearing Strength of Cement and Mortars. The shearing strength of mortar is important not only because of the intimate relation ; '."^ --- JNW -Ce'mio u T >, -"'M^ "•i-' L _ _ _ .2 : If » Li _ -^ - - - - -' - - "■ ' -p- - ."!* •i* MODULUS OF ELASTICITY OF CEMENT, AND MORTARS Z I StBudord Saod and same Cement)!-! ~j s' Vs^° o " ^■■- 1 I ;i s X w X * ^ * M ^ ^ -■ -^ i i^ ■^ = 5 8 S i s i i i t t i i i oi ■•■ « e a 2 Load Interval in Pounds per Sq.lu. Tig. 94. 166 MATERIALS OF CONSTRUCTION between shearing strength and compressive strength, but because of the fact that mortars and concretes are often subjected to shearing stress when used structurally. Very few tests of shearing strength have been made, however, probably largely because of the difficulty found in sub- jecting a specimen to a purely shearing stress. The tests made by Bauschinger and published in the Proceedings of the Munich Technical Institute in 1879 still remain almost the only source of information in this important property of mortars. The following table has been prepared from Bauschinger's report. The specimens used were prismatic in shape, the cross-section being 2.4 inches by 4.8 inches. The specimens were stored in water. Each result in the table is the average of nine tests. For comparative purposes the tensile and compressive strengths of the same mixtures are given in the table. All stresses are given in pounds per square inch. SHEARING STRENGTH OF CEMENT AND MORTARS Mix. Tension. Compression. Shear. 7 day. 28 day. 2yr. 7 day. 28 day. 2yr. 7 day. 28 day. 2 yr. Neat 1 :3 1 :5 224 95 64 294 169 103 292 272 232 1910 880 537 2490 1040 977 4680 3340 2960 271 116 77 346 188 131 415 ~ 375 364 It will be noted that in these tests the shearing strength, expressed as a percentage of the compressive strength, varies in the case of the neat cement from about 14.2 per cent at seven days to about 8.4 per cent at two years; in the 1 : 3 mortar it varies from about 13.2 per cent at seven days to about 11.2 per cent at two years; and in the 1 : 5 mortar it varies from 14.3 per cent at seven days to 12.3 per cent at two years. The shearing strength will in general be dependent upon the same factors as the tensile and compressive strength. In particular, the fineness of grinding of the cement, and the character of the sand used in the mortar will be important factors. The Adhesive Strength of Cement and Mortars 156. Adhesion to Steel. The adhesion of cement and mortar to steel is important in the consideration of reinforced concrete and in the con- sideration of the holding power of anchor bolts, etc., which are imbedded in cement or mortar, or are grouted in place in drilled holes. Few tests are available to show the adhesive power of cement and mortars on steel, PORTLAND CEMENT 167 though many tests have been made to determine the same property for concretes. The following data are quoted by Sabin * from tests made with an ordinary river sand. Each figure is the average of from five to fifteen tests made when the mortar was 'one month old, the imbedded steel being plain round rods. ADHESION OF CEMENT AND MORTAR TO IRON RODS Mixture. Neat. 1 :2 1 :3 Average adhesion, lbs. per sq. in. of contact 313 264 111 156. Adhesion to Brick. The adhesion of cement and mortar to brick is of importance in all brick masonry construction. The following table has been constructed from a series of tests made by Sabin. f In each case two bricks were cemented together flatwise with a J-inch mortar joint and pulled apart in tension after a given interval. The specimens jvere stored in damp sand after the first forty-eight hours, and tensile tests of the same mortar were made for purposes of comparison on bri- quettes stored in the same manner. The consistency of the mortars was rather more moist than a normal consistency, but was not equal to that of mortar as usually used in masonry construction. ADHESION OF CEMENT AND MORTARS TO BUILDING BRICK Adhesion or Cohesion. Age. Months. Tensile Strength, Pounds per Square Inch, of Mortars Containing Parts Sand to One Cement. None. i 1 2 3 Cohesion 1 1 3 3 6 6 632 48 676 64 723 50 596 42 728 52 764 56 589 24 694 41 679 39 409 20 423 24 524 20 270 Adhesion 11 Cohesion 325 Adhesion 12 Cohesion 374 Adhesion 14 It will be noted that in general there is a fairly close relation between tensile strength and adhesion at all ages, the ratio of adhesion to cohesion being about 8.6 for neat cement, 7.1 for 2 : 1 mortar, 5.2 for 1 : 1 mortar, 4.8 for 1 : 2, and 3.8 for 1 : 3, showing that the addition of sand decreases the adhesion to brick more rapidly than it does the cohesive strength. Tests made with mortars to which lime paste had been added showed that the addition of 10 per cent lime increased the adhesive strength 120 to * Sabin, " Cement and Concrete," p. 300. t Loc. cit., p. 293. Fig. 95. 168 MATERIALS OF CONSTRUCTION 140 per cent; 16.7 per cent lime, 130 to 160 per cent; 25 per cent lime, 110 to 120 per cent; and 50 per cent lime, 75 to 80 per cent. The addition of lime increased the ratio of adhesion to cohesion in all percentages, since only small percentages of lime are even moderately beneficial to tensile strength and large percentages are detrimental. 157. Adhesion of Mortar to Various Materials. The adhesion of mortars to various building materials is a matter of much importance in construction work, but it has been little investigated. Fig. 95 gives the results of tests made by Gen. Wheeler and reported in the "Report of Chief of Engineers," 1895. Discs of the materi- al concerned were pre- pared, 1 inch by 1 inch square and J inch thick, and inserted in the center of the briquette molds, which were subsequently filled with mortar and tested in the usual manner in tension. 158. Abrasive Resistance of Cement and Mortars. This property of cement and mortars is primarily of importance in the determination of the best mortar for use in the top coat of concrete floors, walks, and pavements. Resistance to abrasion will always be dependent not only upon the cement, as regards the tenacity with which it clings to the sand grains (which will be largely dependent upon its fineness and its lime content), but also upon the hardness of the sand used. Abrasion either wears away the cement and the sand particles or it pulls the sand grains out of the cement matrix. With soft sand particles the resistance to abrasion with a given cement decreases constantly as the percentage of sand is increased. With hard sand grains the abrasive resistance increases as the proportion of sand increases, until the volume of cement becomes relatively too small to bind the sand grains together thoroughly. This limit is found to be reached when the mortar contains not more than two parts of sand to one of cement. 159. Permeability and Absorptive Properties of Cement Mortar. The permeability of mortar is a measure of the rate at which water under a given pressure will pass through a given thickness of the material. The absorptive properties of a mortar constitute a measure of the rate POETLAND CEMENT 169 at which moisture will be absorbed when, the mortar is exposed in damp situations or covered with water under negligibly small heads. Permeability is an important consideration where watertightness of walls, etc., is required and percolation of water is not admissible. Absorptive properties of a mortar determine its value as a damp- proofing coat, particularly in the event of its use as a mortar over metal lath, which must be protected to prevent corrosion. In view of the disintegrating effect of expansion and contractipn of mortars used as a plaster, etc., the moisture content (which largely affects this expansion and contraction) should not be greatly variable. Thus the' least absorp- tive mortar will be most durable, up to the limit reached when the cement content is relatively so high that the expansion and contraction is dis- proportionately increased. The determining of precise information concerning each of these properties is dependent upon a standardization of methods of conducting tests. Such standard methods have not yet been adopted, and it is therefore impossible to quote data as to the absolute permeability or absorptive power of mortars. Tests to determine the relative permeability and absorptive power of mortars were made at the Structural Materials Laboratory at St. Louis in 1909, and are reported in Technologic Paper No. 3 of the Bureau of standards. Owing to the small number of tests made and certain unsatisfactory features of the' testing method employed, only a few general conclusions will be drawn from the report of these tests. (1) PermeabiUty decreases rapidly for all mixtures with increase in age of the specimens when tested, (2) permeabiUty decreases consider- ably with the continuation of the flow, (3) permeability increases with the leanness of the mixture, the dryness of the mixture, and increased coarseness of the sand. Absorption was found to be dependent upon the same factors: it de- creased with the age of the mortar as a rule, but not as rapidly as did the permeabiHty (especially with the leaner mixtures) ; it decreased but slightly with increased richness of the mixtures; and the wetter mixtures were slightly less absorptive than the dryer mixtures. 160. The Expansion and Contraction of Cement Mortars. Changes of temperature are accompanied by changes in volume in cement mor- tars as in other structural materials. The temperature coefficient for mortars has been found with considerable accuracy to be 0.0000055, i.e., a specimen will suffer an increase or decrease in lineal dimensions of 0.00055 per cent for an increase or decrease of 1° F. (0.00099 per cent per degree centigrade). 170 MATEEIALS OF CONSTRUCTION 12315(>7891012 Other volume changes due to the chemical processes of setting and hardening, and still others caused by variation in the moisture content of cement, constitute important considerations in the use of mortars and concrete. Professor A. H. White of the University of Michigan has made an excellent experimental study of this question, and his paper, entitled " Destruction of Cement Mortars and Concrete through Expan- sion and Contraction," read before the American Society for Testing Materials,* will be -exclusively used in the present short discussion of the question. Fig. 96 shows the average percentage of linear shrinkage of bars of neat cement kept constantly in air for periods up to four years, and also the average linear expansion of bars kept continuously in water for three years. Four different brands of Port- land cement, all passing standard specifications for constancy of volume, were used. It is evident that in the case of neat cement fully one-half the total expansion or contraction comes in the first month, and there is very little change after one year. The expansion and con- traction of 1 : 3 mortar was found to be very much less than that of neat cement, and in general, at the end of one year amounts to from one-quarter to one-third that of neat cement. When bars of cement or mortar, after lying in water for three years, were allowed to dry in air at room temperatures, they gradually contracted till, at the end of two months, they had shrunk not only to their initial volume, but in most cases showed a considerable further contraction. When the bars so air-dried were again immersed in water they recovered in one day 90 per cent of the length they had lost in sixty-five days' air drying, and in most cases their length within one month exceeded that which they showed at three years, before having been air dried. These percentages of expansion and contraction seem small, but * Proceedings, Am. Soc. for Test. Matrls., Vol. XI, 1911. — 1 iTTTTTTi"" FT !>? .-"--i-- A J St )(;■■' ^4- -,-4-- - -- ^* ' ' c " BARS'OF CEMENT At&l u Rtbpiv iionfho rrn"-- 21 Fig. 96. PORTLAND CEMENT 171 when closely examined appear to be far from negligible. The shrinkage of 0.322 per cent amounts to about 4 inches in 100 feet, and indicates clearly why neat cement cannot be used for interior floors or wall plasters, since if it did not crack, it would, when dry, have to withstand a . ensile stress of 7500 pounds per square inch. In the case of such a plaster being constantly immersed and constrained from expanding, compressive stresses up to 5000 pounds per square inch would be encountered. With 1 : 3 mortar instead of neat cement, the observed volume changes would, if expansion or contraction were restrained, mean the introduc- tion of tensile or compressive stresses of 1250 pounds per square inch. Such stresses would in the case of the tensile stress caused by contraction necessarily mean failure by cracking. CHAPTER VIII CONCRETE GENERAL 161. Concrete as a Structural Material. All of the masonry mate- rials heretofore discussed — plasters, limes, and cements — have been strictly cementing materials. Alone, they are not used as masonry, but are used in combination with such non-cementing masonry materials as brick, stone, terra cotta, etc., or are used as a plaster coat in walls. Concrete, considered as a cementing material, is unique in that it may be and commonly is used alone to form bulk masonry. If we consider the constituent materials from which concrete is made, however, instead of considering concrete masonry as one material, we find that we have to deal with a class of masonry which is made up of a large bulk of non- cementing materials l?ound together by a comparatively small amount of cementing material, just as is the case with stone or brick masonry. The distinction lies, however, in the fact that in concrete the non-cement- ing material is in a comparatively finely divided state, and is incorporated in a mix with the cementing material prior to placing on the work. In stone or brick masonry, on the other hand, the non-cementing material is brought upon the work in the shape of conveniently sized units with which the structure is built up, cementing material being used only as beds and joints to bind the stone or brick together. One other distinction belongs to concrete alone among masonrj' materials — the fact that it may readily be strengthened or reinforced by the placing of steel therein. This makes possible its use in situations where tensile stresses are encountered, whereas stone and brick masonry may be depended upon to withstand compressive stresses alone. The subject of reinforced concrete is so complex, the problems en- countered in design and construction ones so peculiar to this class of construction alone, and the bibliography of the subject already so exten- sive, that the discussion of concrete in this chapter will not be made to include reinforced concrete. 172 CONCRETE 173 CONCRETE MATERIALS The Cement 162. Selection of Cement. Most engineers consider it a wise rule to use Portland cement for nearly all classes of concrete work. The con- ditions most favorable to the use of natural cement have been discussed above under the head of natural cement. Any cement used should be accepted under such standard specifications as those of the American Sociejiy for Testing Materials and, except on the least important work, samples from each shipment should be subjected to the standard tests recommended by the American Society of Civil Engineers. 163. Storage of Cement. After delivery on the work the cement should be carefully stored in weather-tight buildings the floor of which is raised from the ground. The storing should be done in such a manner as to permit of easy access for inspection and identification of each shipment. 164. Inspection and Testing. Each shipment of cement should be inspected by a competent inspector whose duty it is to select a sample for tests. The sample should be a fair average of a bag or barrel and, if conditions permit, about 1 barrel in every 10 should be sampled. Usually tests are made on a mixture of the individual samples, but in some cases on important work the individual samples are tested separately. Sand foe Concbete Aggregate 165. Granulometric Composition. In general, the discussion of sands for mortars included in Chapter VII (Arts. 141 to 144) applies with equal force to the consideration of sands for concrete. Investigations have abundantly shown that the sand should be for the most part coarse, rather than fine, that passing a 20-mesh sieve and retained on a 30-mesh sieve ' showing highest mortar strength. A sand showing a gradation in size from fairly coarse to fairly fine is preferable to either a uniformly coarse or a uniformly fine sand. 166. Shape of Sand Grains. It is usually specified that concrete sand shall be " sharp," by which is meant that the grains of sand shall be angular rather than rounded. For this reason, other things being equal, a bank sand is preferable to a river or beach sand. The advantage in favor of the former may be offset, however, in some instancesby the greater cleanUness of the river or beach sand. 167. Foreign Matter in Sand. It has been shown in the chapter above referred to that the injurious effect of foreign matter in sand is dependent upon many factors. If the silt contains more than 10 per cent 174 MATERIALS OF CONSTEUCTION of organic matter, the latter constituting as much as one-tenth of 1 per cent of the sand, an appreciable injury results. Clay, if finely divided and uniformly distributed throughout the sand, appears to have little effect unless present in large percentages, perhaps 10 per cent or more. Mica is injurious even in very small percentages. It is important, therefore, that the sand used on concrete work should be subjected to careful examination and tested no less systematically than is the cement. It is not sufficient to specify that the sand be " clean and sharp " and accept it upon the basis of a casual inspection made by taking a bit between the fingers to establish its grittiness, and determining its cleanliness by the amount of discoloration produced by rubbing it in the palm of the hand. 168. Voids in Sand. The proportion of voids in a sand determines the density of a mortar of a given cement-sand proportion and bears a similar relation to the density of a concrete made by the addition of larger aggregate to the mortar. The strength of mortar has been shown to be directly proportional to the density, which means, therefore, that a sand having the least void space will give a mortar and hence a concrete of greatest strength. This is simply corroboratory of the statement above made that a sand shovping a fair gradation in size is in general preferable to one of uniform size. Broken Stone or Gravel Aggregate 169. Gravel vs. Broken Stone. Either class of coarse aggregate may be perfectly satisfactory, and neither can be said to be wholly superior to the other. If the consistency of the concrete is such as to constitute a rather dry mix, more tamping is necessary to obtain a dense concrete with broken stone than is the case with gravel consisting of smoother and more rounded particles. Gravel usually has a smaller percentage of voids than has broken stone, and therefore a compact concrete may be secured with a somewhat smaller amount of mortar than would be required for broken stone. On the other hand, if properly tamped, the broken stone will to some extent interlock, forming a dense and strong concrete, the same effect being possible with a well-puddled wet mix. Also, the rough surface of the broken stone usually results in developing a greater adhesive strength or bond between the stone and the mortar. This latter consideration cannot be taken to be universally applicable, however, for the adhesion of cement to stone is not wholly a matter of roughness or smoothness. 170. Crushing and Screening Stone. Stone for concrete aggregate is crushed in exactly the same manner and with the same types of machines CONCRETE 175 described in connection with the preparation of the rock for the manu- facture of cement. Stone or gravel should always be screened to remove all crusher dust and fine material, unless the material is of such a character that it may be considered as sand and due allowance for same made in fixing the proportions of the mixture. 171. Mechanical Analysis of Stone. On important work mechanical analysis of the aggregate is desirable since it affords a basis for determining the best proportions to be used with given materials, or for determining what sized material should be added to the aggregate to make it more satisfactory for use in a concrete of given proportions. Mechanical analysis consists simply in passing the material through a succession of screens the mesh of which is of increasing fineness. The residue on each sieve is weighed and, beginning with - ""^i: the amount which has passed the finest sieve, the weights are successively added, so that each sum represents the weight of material which has passed a sieve of given mesh. These sums expressed as per- centages of the total weight of material used are plotted as the ordinates to a curve the abscissae of which are the diameters of the particles. Fig. 97 represents the mechanical analysis of a rather typical gravel. 172. Size and Shape of Fragments of Stone. All material under I inch or even f inch should be removed from either crushed stone or gravel, this fine part being in some cases suitable for use as sand. The maximum size of coarse aggregate allowable depends upon the character of the work on which the concrete is used. Usually this limit is such a size as will pass a 2§-inch ring. For massive concrete, however, the size may be such as will pass a 3-inch ring, and for reinforced concrete a size to pass a 1-inch, or even a f-inch ring, is required. From the standpoint of minimum void space rounded stone are more desirable than irregular rough- fragments; from the standpoint of ability to bond with the mortar they are deficient, however, so that in general the shape of the particles will be found to be much less important than the size and the hardness of the stone. 173. Voids in Stone or Gravel. As above noted, rounded particles of stone will have a smaller void space than will irregular broken frag- ,1i' 1.00 1.26 Diameter oi Stone in Inches Fig. 97. 176 MATEKIALS OF CONSTRUCTION ments. The percentage of voids in natural gravel usually runs from about 30 to about 35 per cent, while in crushed stone the percentage will in general vary all the way from 30 to 45 per cent, depending upon the shape and gradation in size of the particles. THE MAKING OF CONCRETE Proportioning Concrete * 174. Importance of Proper Proportioning. Upon important work, par- ticularly if of large extent, a thorough study of the materials and the proper relative proportions will often effect better results with a saving in cost. The cement is always the most expensive ingredient, and therefore if it is possible, as it often is, to reduce the proportion of cement used by adjusting the proportions of the aggregates in such a manner as to produce a leaner mixture of equal or greater density and strength, economy is thereby effected. 175. Theory of Proper Proportioning. The theory of correct pro- portioning of concrete materials is based on one fact — namely, the greatest strength and imperviousness as well, is always obtained with a mixture of greatest density. The mortar will have a maximum density when there is just sufficient cement to fill the voids and coat the particles of the sand, and the concrete will have a maximum density when there is just sufficient mortar to fill the voids and coat the particles of stone. Going a step farther, the concrete will be of maximum density with the least proportion of cement used, when each of the aggregates shows such a gradation in size as to make its void percentage least. That is, the proportion of particles of each successive size in each of the aggre- gates should be just sufficient to fill the voids in the next larger size. 176. Proportions in Practice. It is customary to stiate concrete proportions by volume, giving the number of parts of sand and stone to one of cement. It is hardly possible to state arbitrarily the proportions used in practice for concrete in any particular situation. The following division is given, however, as fairly representative of present conservative practice: (a) 1 : IJ : 3 — a rich mixture used for columns and other structural parts Subjected to high stresses or requiring especial water-tightness. * A very detailed discussion of concrete proportioning written by Mr. William B. Fuller will be found in Chapter XI of Taylor and Thompson's " Concrete, Plain and Reinforced." CONCRETE 177 (6) 1:2: 4 — a standard mixture used for reinforced floors, beams, columns, arches, engine and machine foundations where vibration occurs, sewers, conduits, etc. (c) 1 : 2| : 5 — a medium mixture used for floors on the ground, ordinary machine foundations, retaining walls, abutments, piers, thin foundation walls, building walls, sidewalks, etc. (d) 1:3: 6 — a lean mixture for massive concrete, heavy walls, large foundations under steady load, stone masonry backing or fiUing, etc. •(e) 1:4: 8 — a very lean mixture used only on unimportant work in very large masses. 177. Ingredients Required per Cubic Yard of Cement. Cement is usually bought and measured by the barrel, the weight of the barrel being 376 or 380 pounds net and the volume about 4 cubic feet. Sand and stone are bought and measured by the cubic yard, so that it will usually be convenient to determiiie the number of cubic ystrds of sand and stone or gravel for each barrel of cement in a concrete of a given mix. The following rule will be found to be sufficiently acciu-ate for the purpose of making preliminary estimates of quantities: Let c = the number of parts of cement; s = the number of parts of sand; g = the number of parts of stone or gravel. 10.3 i. ' Then '■ — = P = barrels of cement per (3ubic yard concrete in place; c+s+g and PXsX4/27 = cubic yards of sand per cubic yard concrete; PXsfX4/27 = cubic yards of stone per cubic yard concrete. This rule has been devised by experimental determination of the constant 10.3. It will give fairly accurate results with all classes of mate- rials except with rather fine gravel, or stone which shows a very excellent gradation in size of fragments. In this event the quantities of each of the ingredients other than stone, as determined by the rule, will be found to be about 10 per cent in excess of actual requirements. Mixing Concrete 178. Hand vs. Machine Mixing. Good concrete may be either hand or machine mixed, the choice depending largely upon the quantity required, and the consequent relative cost of the two methods. For all except the smallest work machine mixing will be less expensive than hand 178 MATEEIALS OF CONSTEUCTION mixing, and for this reason, as well as the greater likeli'hood of obtaining uniformly well-mixed concrete of uniform consistency, machine mixing is generally preferred. Hand mixing is apt to be slighted because of the heavy labor demanded. 179. Method of Mixing by Hand. Hand mixing should be done upon a water-tight platform about 10 feet wide by 15 feet long. The measured quantity of sand having been spread over the surface of the mixing plat- form, the cement is spread evenly over the sand and the two mixed thor- oughly dry, after which the required amount of water and stone may be added and the mass turned back and forth from one side of the board to the other until the mass is homogeneous in appearance and color. From three to five turnings are required to mix the concrete thoroughly. 1 Fig. 98a.— Cube Type Concrete Mixer. 180. Mixing Machines and Machine Mixing. Concrete-mixing ma- chmes are of two general classes— batch mixers and continuous mixers In usmg the batch mixer, the materials are measured separately and charged mto the machine in quantities sufficient to make a batch suited to the capacity of the machine. The required amount of water is added and the mass is mixed and theii completely discharged, after which the machine is recharged. In the continuous mixers the materials are deliv- ered gradually to the machine, either already combined in the correct proportions or the rate of feed of each material may be regulated to pro- CONCRETE 179 duce the required mix as nearly as possible. Water is added as the materials pass slowly through the length of the mixer and the mass is discharged continuously from the lower end. Continuous mixers do not generally produce a concrete so thoroughly mixed or of as uniform a consistency as is obtained with the use of batch mixers. Specifications for important work very often stipulate that batch mixers be used. Most concrete mixers consist of a rotating chamber into which the materials are charged and mixed with a complicated motion, due either Fig. 986. — Cube Type Concrete Mixer. to the shape of the mixer chamber or to the action of baffle plates placed on the inside walls of the mixing chamber. The cube mixer (Figs. 98a and 986), requires no baffle plates because of its peculiar shape and manner of mounting. The drum-shaped mixer (Figs. 99a and 996), and the duo cone-shaped mixer (Figs. 100 and 101), have baffle plates. The former is discharged by inclining the pivoted spout downward and outward while the machine continues to revolve. The latter is discharged by revolving the entire mixer on its trunnions. Deposition of Conckete 181. Timber Forms. The investment in materials and the cost of labor in placing and removing forms often represents from 15 to 30 per 180 MATEEIALS OF CONSTRUCTION cent of the total cost of concrete work. It is evident, therefore that the correct design and construction of forms is a very important feature of the work. Forms must be substantially built, so thoroughly braced and Fig. 99a,~Drum Type Concrete Mixer. Fig. 996.-Drum Type Concrete Mixer. wired that the finished concrete shall conform to the designed dimen- sions and contours, and made tight to prevent the leakage of cement- cha.rged water. Fig. 100. — Duo-cone Type Concrete Mixer. Fig. 101. — Mixing Chamber of Duo-cone Type Concrete IMixcr. The cheaper grades of lumber are usually used, such as spruce, fir, or even hemlock. Green timber is preferable to seasoned timber, since it is less apt to be affected by the water in the concrete. Better grades CONCRETE 181 of lumber are often used when a particularly smooth finish is desired or where form lumber or sectional forms are used repeatedly. Oiling of the surface of the lumber is beneficial when it is to be used repeatedly, and plank planed on one side is practically essential. Tongued-and-grooved lumber is often used for the sake of waten-tightness, but lumber beveled on one edge is considered preferable by many engineers because the edges crush as the wood swells, thus preventing buckling. Forms should always be wetted just prior to the deposition of the concrete. 182. Transportation and Deposition in Forms. The one essential in the transportation of concrete from the place of mixing to the forms is that no opportunity be afforded for a segregation to take place between the mortar and the coarse aggregate. Chutes down which the concrete flows are often considered objectionable, if of any .considerable length, for just this reason. Chutes may be so constructed, however, by making the slope conform to the degree of wetness of the mix, that little, difficulty is encountered by reason of segregation. Most commonly, concrete on small work is transported in wheel- barrows or, in case of very thin walls, in hand buckets. On larger work with machine mixers, derricks which swing large dump-buckets, or cars carrying dump-buckets, are commonly used. With modern fairly slow- setting cement there is little danger of the concrete setting before reach- ing the place of deposition except in case of interruption of the work, in which event especial care must be exercised to see that no concrete be left in wheel-barrows or other conveyences. Regauging of concrete which has set or partially set, by stirring up the mass with or without additional water, is never permissible. Concrete should be deposited in nearly horizontal layers only a few inches thick and should never be allowed to flow down a slope, since in that event segregation of the fine and coarse material will invariably take place. On work where absolute continuity of the concrete is required the deposition must be carried on continuously until the work is com- pleted. There will invariably be a joint or plane of weakness where one day's work is stopped and deposition of new concrete resumed after twelve to fifteen hours. It is therefore important that the work be so planned and prosecuted that the planes of weakness he in the direction of least stress in the finished structure. For instance, in an arch the planes should run longitudinally and not transversely. 183. Consistency, Ramming or Puddling. The materials should be mixed wet enough to produce a concrete of such a consistency as will flow into the forms and about the reinforcement if such be present. At the same time it must not be so wet as to cause difficulty through segre- gation of the coarse aggregate and the mortar before final deposition. 182 MATERIALS OF CONSTRUCTION Formerly, it was the practice to mix concrete so dry that an excessive amount of tamping was necessary to compact the mass and prevent bridging of the fragments of the larger aggregate, causing voids. Specifications now usually require a mix so wet that ramming or tamping is unnecessary, a dense concrete and smooth surfaces being ob- tained by simply puddling the mass with a straight shovel or sUcing tool until the ingredients have settled into their proper place by gravity and any surplus water has been forced to the surface. In order to have a good finish it is well to make use of a straight shovel along each side of the forms, forcing all the larger fragments of the aggregate a short distance back from the face. 184. Bonding to Old Work. Since joints cannot be avoided in work not carried on continuously to completion, every reasonable pre- caution should be taken to make the bond of new to old work as strong as possible. In massive work with horizontal joints the question is one of less importance than in thin walls or situations requiring water-tightness. In the former case it will probably suffice to simply clean and wet the old work before laying new concrete. Where walls are thin, or waterproof- ness is required, the concrete previously placed should be roughened, thoroughly cleaned of foreign material and " laitance," and slushed with a thin grout of either neat cement or rich mortar, the proportion of sand to cement in the latter case not exceeding two to one. Laitance is a whitish scum which is washed out of concrete when there is excess of water, as when concrete is deposited in water or when water collects in pools on the surface of freshly laid concrete. The laitance consists of the finest flocculent matter in the cement together with dirt from the aggregates, and its formation will be understood by reference to the discussion of the formation of amorphous hydrates during the setting of cement. The composition of laitance is practically identical with the composition of the cement itself. This flocculent material remains suspended in the water for a long time, giving it a milky appear- ance, and settles slowly on the surface of the concrete. The laitance hardens only very slowly and never acquires much strength, so that, if not removed, it seriously interferes with the bonding of successive layers of concrete. 185. Facing of Walls. The cheapest and most satisfactory method of obtaining a smooth mortar face on concrete walls has already been indicated, namely, by the use of a straight shovel or slice bar along the forms, forcing the coarser aggregate back from the surface. Plastering with mortar after removal of forms is useless, because it will almost invariably scale off, owing to poor bond and unequal expansion. CONCRETE 183 Imperfections in the face must, of course, be patched up, but this should be done with a mortar of the same mix as that used in the concrete to prevent different-colored patches showing. Washing with a thin grout immediately after removal of forms is beneficial to some extent, and if the layer is not of appreciable thickness so as to form a continuous film it will not scale off. A layer of special mortar is sometimes placed next the forms by means of a movable sheet-steel diaphragm which is inserted in the form and kept the required distance from the face by suitable spacing blocks. The concrete and mortar are now filled in simultaneously, and the diaphragm is raised as the work proceeds, so that it is always only a few inches below the surface. In this manner the two mixes come into contact with each other before setting begins and the bond will not be imperfect. A " rubbed finish " is sometimes obtained by removal of the forms while the concrete is still green and rubbing with a wooden float. A " tooled finish " is sometimes produced after the concrete has partially hardened, by use of the tools which are used in finishing stone. A " brushed finish " is produced by brushing the green concrete with a stiff wire brush, after which a dilute solution of hydrochloric or muriatic acid is applied with a brush. The acid thoroughly cleans the stone and brings out the natural colors, but must be immediately removed by slush- ing with water. Otherwise acid discoloration will occur. 186. Depositing under Water. In many classes of subaqueous con- crete construction it is possible to use cofferdams or caissons from which the water may be excluded. The placing of the concrete will then not differ materially from methods commonly used on land. When such methods are not feasible, the problem becomes one of some difficulty, owing to the formation of laitance. Cement, sand, and stone are of course heavy enough to sink in water, but the laitance and some cement which is not immediately hydrated will be floated away. This therefore represents a considerable loss of cement. The problem is entirely one of placing the concrete in its final position under water without allowing the excessive formation of laitance or wash- ing out of cement. Many methods have been used with a greater or lesser degree of success, among which the following may be mentioned: The " tremie," a device often used, consists of a large tube of wood or sheet metal, so constructed as to make its length adjustable, and pro- vided with a hopper at the top. In use the tremie is supported vertically in the water by barges or derricks, provision being made for horizontal movement of the tube over the area occupied by the work. The lower 134 MATERIALS OF CONSTRUCTION end is allowed to rest on the bottom or is closed by a valve arrangement, and the tube is filled with concrete. The tremie is now hfted a few inches or feet and the concrete allowed to escape as the device is moved over the required area. A layer of concrete of any desired thickness is thus deposited, the tube being kept continuously filled to a point above the water line. Of course this method does not entirely prevent the for- mation of laitance and loss of cement, but it has been found satisfactory on many large works. Buckets, so constructed as to allow the material to flow out from the bottom, the top being closed, are used in a manner similar to the use of the tremie. A derrick lowers the closed bucket into place, the bot- tom doors are opened, and the material escapes as the bucket is hoisted. Buckets so used are usually of large capacity, since if several yards of concrete escape at once the material compacts better with less loss of cement. Bags of all kinds from paper to burlap and heavy jute have been employed in depositing subaqueous concrete. Paper bags are usually of a brown paper which is destroyed shortly after immersion. Cloth bags are not removed or destroyed but, the cloth being very porous, enough cement escapes to bind the bags quite firmly together. Bags are never filled completely, as it is desirable to have them pack together closely. Concrete is sometimes mixed and deposited in water altogether dry, sacks or buckets being used. This method is entirely unsatisfactory, as the escape of cement is very great and it is impossible to obtain as uniform and dense a concrete as is obtained by any of the above methods, using concrete mixed with water in the usual manner. THE MAKING OF CONCRETE UNDER SPECIAL CONDITIONS Laying Concrete in Freezing Weather 187. Effect of Low Temperatures. It has teen shown above (Art. 122) that low temperatures have a marked effect in increasing the setting time of cement, often from four to eight times as long a period being required to obtain a final set at a temperature of 32° F. as is required at normal temperatures. If water in concrete or mortar freezes before the cement has set, it is not available for the chemical action of setting and hardening and hence the concrete or mortar will not set at all until the ice melts. The above facts must be borne in mind when removing forms from con- crete placed during cold weather. CONCRETE 185 ■ If the temperature hovers above the freezing-point for some time after concrete is deposited, there is a possibility of the water drying out before the greatly delayed setting has taken place. If, however, the concrete has begun to set before the temperature drops considerably below the freezing-point, the expansion of the water in solidifying pro- duces an expansive force in. excess of the cohesive strength of the green concrete. This action results in destruction of the bond and crumbling of the concrete when the ice melts. If the temperature does not fall more than a degree or two below freezing, the result may simply be the further delaying of the set without appreciable injury. This is possible because the water may not have frozen, owing to the chemical heat of combination afforded by the slowly setting cement. Experiments made by Mr. E. R. Mathews and Mr. James Watson * led to the following conclusions: (1) Light frost (not more than about 3° F. below freezing) has a permanently injurious effect on cement if it occurs immediately after gauging, a lesser detrimental effect which is not permanent when it occurs twenty-four hours after the cement is gauged, and no effect after forty-eight hours. (2) Heavy frost (about 17° F.) has a permanent most injurious effect upon cement and mortar freshly mixed. (3) The injurious effect of light frost on mortars occurs more immediately than on neat cement, but the mortar recovers from the ill effect more rapidly. (4) The gauging of cement and mortar with warm water (100° F.) has a permanently injurious effect upon cement and mortar. 188. Methods of Concreting in Freezing Weather. " Concrete should not be mixed or deposited at a freezing temperature, unless special precautions are taken to avoid the use of materials containing frost or covered with ice crystals, and to provide means to prevent the concrete from freezing after being placed in position and until it has thoroughly hardened." Work may be carried on during freezing weather by either of two methods — keeping the materials and the work at a temperature above the freezing-point until the concrete has had time to set, or, for temperatures only a few degrees below freezing, by the addition to the water used in mixing of a substance which lowers the freezing-point of water. The first method is more generally recommended and used. The * Trans. Am. Soc. C.E., Vol. 64, p. 320. Ig6 MATERIALS OF CONSTRUCTION sand and stone are heated by piling them over heated iron conduits 'or steam pipes, and, the water is heated in a large supply tank fitted with steam coils. The tests quoted in the above article would indicate that there is perhaps some danger of having the water and other materials too hot at the time of mixing. The work may be protected from frost by covering with earth, canvas, boards, etc., if the temperature falls but very little below freezing, but in case of heavy frost heat must be artificially supplied. One of the most common and efiicient methods consists in covering the top of the work to a depth of several inches or a foot with manure, which is in turn covered with boards or canvas. The chemical action of decomposition of the manure is a source of sufficient heat to prevent frost reaching the work. When the work is in the nature of a building or structure of hmited extent it is practicable to house the work with sheathing or canvas. Fires are then kept going continuously in salamanders within the enclosure, thus keeping the temperature above freezing. The second method, by reducing the freezing-point of water, is not generally considered as favorably as the above-described methods, but is cheaper and hence often used. Common salt or calcium chloride is most commonly used. Approximately 1 per cent of salt in. the mixing water lowers the freezing-point 1° F. Beyond 10 per cent salt becomes ineffective and decidedly injurious. It has not been conclusively shown, however, that small percentages are injurious. Alcohol, glycerine, and other chemicals have an effect similar to that of salt in reducing the freezing-point of a water solution. All are, however, less effective than salt, and the latter being cheaper is commonly preferred. Concrete in Sea Water. Effect of Alkali on Concrete 189. Action of Sea Water on Concrete. The behavior of concrete in sea water is a problem which has occupied much of the attention of engi- neers for many years. The question has often been discussed, and many attempts have been made to determine experimentally the exact action of sea water upon concrete, and the causes of that action. The amount of accurate information available is rather meager, however, and the results of experimental investigations are inconclusive and often contra- dictory. Many concrete structures in sea water have remained intact and uninjured for many years; a few, constituting a small minority of all marine structures built, have been injured or destroyed under the same CONCEETE 187 conditions. In view of the conflicting results obtained experimentally it is difficult or impossible to explain why certain marine structures remain sound indefinitely, while others disintegrate more or less rapidly. We know that the salts in the sea water (magnesium sulphate, mag- nesium chloride, sodium chloride, and calcium sulphate) react in some way with the constituents of cement. It appears further that cements high in free lime or especially high in alumina are especially subject to the destructive attack of the salts in the sea water. Beyond this general statement we are not willing to attempt any explanation of the chemical action involved. In view of the recently acquired more definite knowledge of the con- stitution of cement and the chemical processes of hydration, a reason- able explanation of the chemical action of sea water upon concrete may soon be forthcoming. All attempted explanations have hitherto been predicated upon an imperfect knowledge of constitution and are in con- sequence untrustworthy. The chemical action is accompanied by various physical phenomena: sometimes the mass swells, cracks, and gradually falls apart, sometimes the mortar softens and gradually disintegrates, and occasionally a crust forms on the surface which later cracks off. Often the disintegration is facilitated by freezing or by imperfect construction, especially when proper means have not been taken to prevent the inclusion of the lai- tance, which forms to a much greater extent in salt water than in fresh water. 190. Expedients Adopted to Prevent Injury by Sea Water. Fore- most among the precautionary measures to be taken in the construction of marine structures of concrete is the securing of as dense and imper- meable a concrete as possible. This end may be secured by any of the means discussed in Art. 192 and sequence. An outer shell of especially dense materials is sometimes used with good results on marine structures. In this case a few inches of rich mortar (1 : 2 or 1 : 2J) is made to enclose and protect the inner portion of the concrete. It is of course necessary that this outer layer be cast at the same time as the inner portion in order that there may be a perfect bond between the two mixes. Sometimes certain substances, such as barium chloride, are dissolved in the mixing water for the mortar used on the outer shell. These, upon contact with the salts of the sea water, form insoluble sulphates which tend to close the pores in the mortar. Sesquicarbonate of ammonia or magnesium fluosilicate are some- times used as a coating apphed to the face of the finished work by brush or spray. These tend to form an impervious film of car- 188 MATEEIALS Of CONSTRUCtroN bonate of lime in the one case, and insoluble calcium fluoride and lime silicate in the second case, thus stopping the pores. Of course the latter methods remain effective only just so long as the impervious coating remains intact. 191. Effect of Alkali on Concrete. The effect of alkali on concrete is a problem resembling in many respects that of the action of sea water on concrete. The problem is of especial interest in connection with con- crete construction in the arid regions of the West, where soluble salts are present in the soil to an extent not usually found elsewhere. The principal salts encountered in alkali waters usually include: magnesium sulphate, calcium sulphate and sodium sulphate, magnesium chloride, sodium chloride, and potassium chloride, together with carbon- ates of magnesium, sodium, and potassium. Of these the sulphates appear to be most active in causing disintegration of concrete ; the chlorides also are active, while the carbonates appear to be without effect. The attempts at an explanation of the manner of attack of these salts upon concrete have hitherto encountered the same difficulty found in the case of sea water — an unsatisfactory knowledge of the constitution of cement. From the physical point of view the action exactly resembles the action of frost, except that it is more rapid. There exists, apparently, a disruptive force which quickly destroys the bond and causes disin- tegration. This action appears to proceed most rapidly in the parts of a structure subjected to alternate wetting with alkali water and drying in the air. In porous concrete the action proceeds much more rapidly than in dense concrete, where, indeed, it may make no progress at all. As in the case of the injurious action of sea water on concrete, instances of failure caused by alkali waters are merely isolated ones, presenting an interesting field for study, but not constituting a very serious menace to the future of concrete construction in the arid regions of the west. The remedy in the present state of our knowledge is, as in the case of marine structures, a matter of the possible physical precautions only — the secur- ing of the densest possible concrete, thus preventing injury by the exclu- sion of the salt-bearing waters. Concrete where Water Tightness is Reoutred 192. Proportioning the Mixture. The permeability of concrete is closely related to the porosity or void content, but the relationship is not always direct and by no means constant, since the continuity and size of the pores determines permeabiUty more than does the actual per- centage of voids. CONCRETE 189 Dense concrete may, as above noted, be most readily obtained by a careful proportioning of the mixture based on careful selection and mechan- ical analysis of the aggregates. Mixing several of different granulometric compositions to obtain one having a minimum void space may sometimes be resorted to on important work. Usually only the outer layer having a thickness of a few inches need be thus carefully proportioned. It may even be advisable for this outer shell, to use a mortar of 1 : 2 mix, in which case with careful deposition practical imperviousness may be secured. 193. Thickness Required for Water Tightness. The thickness required for a water-tight wall or waterproofing layer on heavy walls, etc., will of course depend upon the mix used, the care exercised in the selection ,of the aggregates and in depositing the concrete, and the pres- sure or head of water encountered. A 1 : 2 mortar carefully made with selected materials need not be more than 2 to 4 inches thick to remain practically water tight under all heads up to 40 or 50 feet; A 1 : 2 : 4 concrete carefully made of selected materials will be prac- tically water tight under heads up to 50 feet when not more than 1 foot thick; and An average 1:2:4 concrete made under average conditions without more than common intelligent selection of aggregates will be practically water tight under heads up to 10 feet when the thickness is made from 1 foot to I5 feet. 194. Use of Waterproofing Compounds. Waterproofing compounds may be classed in two general divisions : Inert fillers — that is, those mate- rials such as clay, finely ground sand or feldspar, hydrated dolomitic lime, etc., which serve simply as void fillers and do not have any action upon the cement nor change in themselves — and active fillers, which react with certain of the constituents of the cement to form inert insol- uble compounds, or in the presence of the cement react with water and precipitate insoluble compounds. In this latter class are included many patented compounds all consisting essentially of stearic acid combined with soda and potash or lime. Inert fillers are added to the dry cement before mixing the mortar or concrete in percentages usually amounting to from 10 to 20 per cent of the weight of the cement. Active fillers are also mixed with the dry cement before mixing, but the percentage used is not often more than 2 per cent by weight of the cement. Upon the addition of water to a stear- ate of Hme, a lime-soap is formed which is not only insoluble in water, but is not wet by the water. Hence these compounds are often spoken of as " water-repeUing compounds." In case the stearic acid is combined 190 MATERIALS OF CONSTRUCTION with soda or potash, instead of lime, the soda-soap or potash-soap is readily soluble, and these must combine with the lime in the compound to form the insoluble lime-soap. This is readily accomphshed, since the stear- ates in the compounds never amount to more than a very small per- centage, the greater part of the material being hydrated lime and magnesia. All of the inert fillers are fairly effective in reducing permeabiUty, clays being slightly more effective than ground sand or feldspar. The active fillers are also usually more or less effective in reducing permeabil- ity, though often to a lesser degree than some of the inert materials. The inert fillers have little effect upon either tensile or compressive strength of mortars and concretes. In fact, the clay in particular sometimes appears to be beneficial to strength. The active fillers, on the other hand, usually reduce both the tensile and the compressive strength of rich mor- tars (i.e., not leaner than 1 : 4) and only in very lean mixtures is their injurious effect upon strength no longer noted. Hydrated lime used in amounts not exceeding 10 to 15 per cent of the cement is one of the best materials for waterproofing concrete avail- able. Its action, as above noted (Art. 145), appears to be chiefly mechan- ical, in that it produces a fat, viscous mortar in which separation of sand and cement is reduced to a minimum and a uniform dense concrete secured. 195. Layers of "Waterproof Material. Layers of waterproof paper or felt applied with a coating of coal tar or asphalt are sometimes used as an impervious course in underground concrete walls, floors, etc. As- phalt is much superior to coal tar, since the latter deteriorates when long exposed to moisture. The asphalt is spread hot on the concrete already placed, followed by alternate layers of paper or felt and hot asphalt. Usually the waterproof course is laid 3-ply, 4-ply, or even 5- or 6-ply. Such a course is finally coated with asphalt again, and the remainder of the concrete deposited in place at once. A distinct joint in the masonry is necessarily formed in the plane of the impervious course, and this fact must not be overlooked in designing walls and floors in which a waterproof layer is incorporated. 196. Surface Treatments for Waterproofing. The principal classes of coating compounds are the following: (1) Oil Paints and Varnishes. These are usually not specially made for use as cement paints, but are ordinary paints consisting of resins, pigments, driers, etc., mixed with linseed oil. They are superficial, inelastic, short lived, and of little value. CONCRETE 191 (2) Bitumens. These include asphalt, petroleum residuum, and coal-tar pitches. All are applied as a hot Uquid with or without waterproof paper or felt, and become solid at ordinary temperatures. Bitumens give fairly satisfactory results owing to their great elasticity and durability except when exposed to the weather. (3) Liquid Hydrocarbons. These include solutions of paraffin in benzine or benzol, and emulsions of petroleum oil or fat in water secured by the use of ammonia. They are superficial, but may prove effectual until their efficiency is destroyed by the opening up of small surface cracks on the face of the masonry. (4) Soaps. Soaps are used either solid or in solution, and also in connection with alum. They are soluble in water, so their efficiency is limited to the possibility of chemical action resulting in the formation of insoluble lime-soap. (5) Cements in which Water-repellent Material has been Incor- porated during Manufacture. These may be used as an exterior coating, or they may be incorporated in the entire concrete. They differ in no respect from ordinary cement with which one of this class of waterproofing compounds has been mixed just prior to concrete or mortar mixing, and hence need not be separately discussed here. PROPERTIES OF CONCRETE 197. Compressive Strength. The compressive strength of concrete is dependent upon many factors which vary widely. All of the factors which affect the strength of cement mortars naturally affect concrete in a similar manner. The density, the character and granulometric composition of the aggregate, the consistency of the mixture, the actual mechanical strength of the stOne of the aggregate, and the conditions of mixing, deposition, and aging all have a direct bearing upon the strength of concrete. Thus it is impracticable to attempt to state the average strength of given mixes of concrete at definite ages in more than a very general way. For any work. of large extent it is always desirable to make a series of tests of the actual concrete used, the test specimens being, whenever possible, molded from the concrete actually mixed for the work. If this is not possible, the materials employed on the work should be used, and the conditions of mixing, etc., made to approximate those which will obtain on the work as closely as possible. Fig. 102 is inserted to give an approximate idea of the strength of concretes. The curves express the average results of compressive tests 192 MATEEIALS OF CONSTRUCTION of five concretes made from different brands of Portland cement. The tests were made in 1899 at the Watertown Arsenal.* All specimens were 12-inch cubes, the mixtures used being 1:2:4, 1:3:6, and 1:6:12. Fig. 103, which is based upon a series of long-time tests of 8-inch concrete cubes of 1 : 2 : 4 mix made by the author, is inserted because a five-year period is covered by the tests. All of the specimens were made 38oa .36UU i-!3tO0 ^3200 L.30UU aagou S 201)0 *^,24UU 52200 o2000 £ll800 M1600 £l400 Si 200 lIiooo g 800 o 600 400 20O ^ — 1 1 1 1 1 1 1 1 1 1 1 1 1 COMPRESSIVE STRENGTH OF CONCRETE WATERTOWN ARSENAL TESTS (CURVES AVERAGE TESTS OF 5 BRANDS OF PORTLAND CEMENT COHCRETE; 12*CUBES - - ^ y ■^ ^ J . 7-.» A Y ^ y- •^ \-- F A / ^ / / f 1 fj It 1: S: r-1 11 — " / ^ — ~ \ ~ " " ^ t ""~ ~ 1 1 1 1 1 1 1 1 M 1 M COMPRESSIVE STRENGTH OF CONCRETE LONG TIME TESTS CURVE AVERAGES RESULTS OF 6 TESTS) COF S'OUBESI 1:2:4 MIXTURE - - c ^2800 — < —J ,^ ^ y\ - -J V / S 1800 glOOO cfi 1200 > 1000 1 6»» 5 ^ Age in Months Fig. 102. 2 3 Age In Years Fig. 103. from the same mix with a well-known brand of Portland cement. The sand was an average, fairly clean and well-graded natural sand, and the stone was crushed trap rock from which all material under |-inch and over 2 J inches had been screened. The curve given is representative of the results of similar tests made with several different brands of cement. The slight drop in strength after long intervals was found characteristic of practically all tests. Specimens after molding were stored in damp air for forty-eight hours, and were then allowed to harden in air in the laboratory. 198. Tensile Strength. The tensile strength of concrete is a prop- erty of limited importance because, being low in comparison with the compressive strength, concrete is practically never designed to with- stand tensile stresses. It will usually be found to be more economical * Tests of Metals, 1899. CONCRETE 193 to use steel reinforcement than to depend upon the tensile strength of concrete. The character of the workmanship and the materials used will greatly influence tensile strength, perhaps to an even greater extent than they affect compressive strength. The series of tests enumerated in the following table were made by the author upon specimens molded in the field under ordinary field condi- tions, and shipped to the laboratory for testing. The compressive tests were made upon 6-inch cubes and the tensile tests upon prisms, 6 by 6 inches in cross-section. The mix- tures used were of three kinds: (a) First-class Limestone Con- cretes. One part Portland cement, two. parts bank sand, and four parts of limestone crushed to pass a 1 5-inch ring. (6) First-class Sandstone Con- crete. Same mix and materials as (o) except that cobbles, mostly sandstone, crushed to pass a 25-inch ring, were used in place of crushed limestone. (c) Second-class Sandstone Concrete. Same as (6) except that mix was 1 : 2| : 5 instead of 1:2:4. Fig. 104.— Tensile Test of Concrete. The compression tests were made in the usual manner, the compres- sion surfaces being bedded in plaster of Paris. The tensile tests were made by the use of specially con- structed shackles which gripped the specimens by means of hardwood wedges on all four sides. The load was transmitted to the shackles through a spherical joint at either end, thus insuring failure by direct axial tension, and not through a complication of stresses caused by bending action. The majority of the specimens broke at a point well away from the grips, and in every case the failure occurred in a plane practically at right angles to the stress, the break being clean and showing no indication of any stress other than direct tension. Fig. 104* shows one of the tensile specimens after failure, still held by the apparatus used for the tests. 194 MATERIALS OF CONSTRUCTION (Note that the values of the ratio of tensile to compressive strength in the following table would have been somewhat lower had the speci- mens been tested in compression at the same age they were in tension).* TENSILE AND COMPRESSIVE STRENGTH OF CONCRETE Approx. Age. Compressive Tensile Ratio Quality and Strength, Lbs. per Sq.in. Strength. Lbs. per Sq.in. Tensile Str. Mix. Compress Str. Tensile Tests. Compres. Tests 1st Class 6 Mo. 1 Mo. 2206 278 Limestone 2708 308 1:2:4 Average . . 2500 253 306 264 257 2505 278 11.1% 1st Class 6 Mo. 1 Mo. 1069 149 Sandstone 1375 142 ' 1:2:4 1417 1722 2000 2139 133 178 158 128 153 150 161 1620 150 9.3% 2d Class 6 Mo. 2 Mo. 1028 ' 121 Sandstone 1639 114 1 : 2i : 5 972 889 1042 2083 1472 1889 106 158- 114 97 179 129 Average. . . 1639 139 1406 129 9.1% 199. Transverse Strength. The transverse strength of plain con- crete is almost wholly dependent upon the tensile strength of the concrete. Experiments show, however, that the modulus of rupture is considerably greater than the strength in tension. The following table represents the results of transverse strength tests made upon concrete beams 8 inches wide, by 10 inches deep, supported on spans varying from 3 to 8 * For detailed account of these tests see The Cornell Civil Engineer Vol 19 pp. 106-113. ' CONCRETE 195 feet. The materials used were a high-testing Portland cement, a fairly clean bank sand of excellent granulometric composition, and crushed trap rock screened to remove all fragments under J inch and over 1 inch in size. The tests were made in the laboratory of the College of Civil Engineering, Cornell University, under the direction of the author. Each result is the average of from six to eight separate tests made at an age of about three months. TRANSVERSE STRENGTH OF PLAIN CONCRETE ■ Mix. Modulus of Rupture. 1 :2 1 :3 1 :4 3 4 6 8 470 389 216 112 200. Shearing Strength. The shearing strength of concrete is a most important property of the material, since it is the real determining factor in the compressive strength of short columns. The strength of concrete beams is also under certain conditions dependent upon the shear- ing strength of the material. Since the angle of shear in concrete compression members must be slightly greater than 45°, we should expect the direct shearing strength to be sUghtly less than one-half the compressive strength. This theory has been well borne out by experiment as the data of the following table taken from tests made at the University of Illinois will show.* SHEARING STRENGTH OF CONCRETE (Each Result the Average of tbom 1 to 17 Tests) Mix. (S) Siiearing Strengtli, Lbs. per Sq.in. (C) Compressive Strength Lbs. per Sq.in. Eatio S/C. 1:2:4 1193 3210 -.37 1:2:4 1257 3210 .39 Average 1225 3210 .38 1:3:6 679 1230 .55 1:3:6 729 1230 .59 1:3:6 905 2428 .37 1:3:6 968 .1721 .56 1:3:6 796 1230 65 1:3:6 692 1230 .56 1:3:6 879 1230 .71 1:3:6 1141 2428 .47 1:3:6 910 1721 .63 Average 856 1605 .53 * BuU. No. 8 Univ. of 111. Exp. Sta., p. 24. 196 R "n — 1 TT 1 — r — r — — •" ~ *~ ^ ' — — ~~ " - ~ ~ ~~ ~ ~ " ~ "~ — 3000- ~ ~ ~ S800- ^ — — "~ "~ " ■. V — „ _ _ 2600 ^ s^ ■» ■" , _ 2100- ~ ~ i^ >* S 2200 - "" ~ / "^ "■ f^ --' _ > ^ ' ■^1800- ~ y' , g 1600 - ^ ^ COMPRESSION OF CONCRETE STRESS -STRAIN. DIAGRAM AGE-3 MONTHS (T6sts of Metals 1899 p. 75ir _ _ _ / _ f ^ i 100 ?- - — — ~ ~ - - 200 I Straiii,Iuches/luch iUG. 105. MATERIALS OF CONSTRUCTION 201. Elastic Properties. The elastic properties of concrete are of importance not only because of their bearing upon the deformation of concrete structures under load, but also because in the design of rein- forced concrete it is necessary to know the relative stresses in the steel and the concrete under like distortions. Fig. 105 presents typi- cal stress-strain diagrams for short prisms of concrete in compression. Curve (1) represents 1:2:4 concrete, and curve (2) represents 1:3:6 concrete. Both are taken from tests made at Water- town Arsenal in 1899.* Fig. 106 presents typi- cal load-deflection dia- grams for concrete beams under transverse loading. (1) represents 1:2:4 concrete and (2) repre- sents 1:3:6 concrete. The tests were made by the author. It will be seen that the elastic properties of concrete vary with the richness of the mixture and with the intensity of stress. They also vary with the age of the con- crete, although this is not shown by the diagrams. Concrete is not per- fectly elastic for any range of loading, an ap- preciable permanent set occurring for even the smallest loads, and the deformation is not pro- portional to the stress at any stage of the loading. s 202 . Modulus of Elas- ticity. Since, as just stated, the deformation of concrete is not proportional to the stress at any stage of the loading, the modulus of elasticity is not a constant for any appreciable range of stress, but varies from point to point, decreasing as the load increases. ■" ^ ■" ~ ~ A'^ . 280 aooo (11 i.a* ^\ __ _ — < »260 , ■" % 220 lT.6 » -' — y rif^ " E 100 5 750 y ^ f- y / / ^ TRANSVERSE TESTS OF CONCRETE BEAMS BEAMS-sVlOE," lO'^DEEP.-loVsPAN AGEjl. MONTH Smgie Conce'ntrated Lond- / - / i 60 f. 1 / / / J ^ i %\ 1 i ; i i V 5 3 a 3 =! 3 1 i Deflection in Inches Fig. 106. ■ Tests of Metals, 1899, p. 751. CONCRETE 197 The modulus is higher for richer mixtures and increases with the age of the concrete. The instantaneous value of the modulus may be computed on the basis of the slope of the chord drawn between two points on the stress- strain curve representing a change of stress not exceeding a few hundred pounds. For instance, the modulus at a stress of 300 pounds for the 1:2:4 concrete of Fig. 105 will be determined by dividing the stress increment from 100 to 500 pounds per square inch by the strain increment for that same range, (i.e., at six months i? = 400 ^0.00012 = 3,300,000 pounds per square inch). Similarly at a stress of 1000 pounds per square inch the modulus is about 2,200,000 pounds per square inch. For the 1:3:6 concrete of Fig. 105 the modulus at three months is about 2,200,000 pounds per square inch at a stress of 300 pounds per square inch, and about 1,250,000 pounds per square inch at a stress of 1000 pounds per square inch. The value of the modulus which is of importance in design and con- struction of concrete is that which corresponds to the working stress of the concrete. Assuming the concrete to be about one to two months old and the working stress not in excess of 500 pounds per square inch, the value of the modulus to be used will be about 2,000,000 to 2,500,000 pounds per square inch for 1:2:4 mix, and 1,500,000 to 2,000,000 pounds per square inch for the concrete of 1 : 3 : 6 mix. 203. Elastic Limit. As stated in Art. 201, concrete shows a per- manent set under the smallest loads. There can therefore be no elastic limit in the true sense of the term. There appears to be a stress, however, below which repetition of the same load does not cause appreciable increase in set, while beyond this stress repetition of load causes increased set indefinitely, finally resulting in rupture far below the normal ultimate strength. For practical purposes, therefore, it is convenient to consider this stress as the elastic limit. Experiments made by Bach, Van Ornum, and others seem to place this stress at about 50 to 60 per cent of the ulti- mate strength. 204. Stress-strain Curves. The curves of Fig. 105 are typical stress- strain curves for concrete in compression. These curves have ofteii been found to approximate closely to parabolas the axis of which is vertical, the origin being located at the point representing the ultimate strength. This fact has been made the basis of some methods of concrete beam design wherein the variation in stress in the concrete is assumed to be parabolic from the neutral axis to the extreme fiber. This means that the design is really based, upon the ultimate strength of the material. This method has been replaced to a great extent by methods which assume the stress-strain curve to be a straight line for stresses under the allow- 198 MATERIALS OF CONSTRUCTION able working stress. The computations required are thus simplified, and the design is based upon safe working stresses instead of the ultimate 'strength of the material. "V 205. Coefficient of Expansion. The coefficient of expansion of 1:2:4 concrete has been determined by several investigators with con- siderable uniformity to be about 0.0000055 per degree Fahrenheit. Other experiments have placed the coefficient of expansion of 1:3:6 concrete at about 0.0000065! These values differ so slightly from the coefficient of expansion of steel that there is little danger of failure of the bond of concrete and steel in reinforced concrete by reason of temperature changes and the resultant volumetric changes. 206. Contraction and Expansion of Concrete. In addition to the volumetric changes due to temperature variation, concrete is subject to other volume changes caused, as in the case of mortars (Art. 160), by the chemical processes of setting and hardening or by variation in the moist- ure content. Experiments made to determine the expansion and con- traction of concrete while hardening are not numerous, but they show conclusively that concrete hardened in air contracts, and concrete har- dened in water expands, the amount of change in volume being dependent upon the richness of the mixture. Experiments made by White (Art. 160) indicate that the expansion or contraction even in the case of old concrete, when alternately wet and dried, is far from being negfigible. Pieces of concrete, presumably not leaner than 1 : 3 : 6 or richer than 1:2:4, sawn from a sidewalk after twenty years in service, showed an expansion of 0.05 and 0.06 per cent when placed in water, and the same contraction when subsequently allowed to dry in air. If this concrete were restrained so that no volume change could take place, the resultant stresses introduced, considering the modulus of elasticity of the cement to be 2,000,000 pounds per square inch, would amount to from 1000 to 1200 pounds per square inch, a stress probably equal to at least half the ultimate strength if in compression, and far exceeding the ultimate strength if in tension. 207. Weight of Concrete. The weight of concrete is a factor in design, in that it must be included in the dead load on any structure. The weight is dependent almost entirely upon the denseness of the con- crete. If the aggregate, both fine and coarse, be of very well-graded composition and the concrete deposited in a manner to insure the mini- mum of void space, the weight may run as high as 160 pounds per cubic foot, and for less carefully chosen materials or less perfectly executed work the weight may not exceed 140 pounds per cubic foot. For practical purposes of design it is customary to assume the weight of concrete to be 150 pounds per cubic foot. CONCRETE 199 208. Adhesion of Steel. The adhesion of concrete to steel is chiefly important in its bearing on the design of reinforced concrete. The bond strength is dependent principally upon the richness of the mix and the character of the surface of the steel. The following table based upon tests made at the University of Illinois in 1906 gives representative values of the bond between concrete and steel rods.* ADHESION OF CONCRETE TO STEEL RODS Steel Rods. Adhesive Mix. Strength, Kind. Size, Inches. . Depth Embedded, Inches. Lbs. per Sq.in. 1:2:4 Plain round .... i and f 6 438 1:2:4 Plain round .... 4 and f 12 409 1 : .3 : 5^ Plain round .... 5 and 1 6 364 1 : 3 : 5i Plain round .... i and f 12 . 388 1 : 3 : 5i Cold rolled shafting 1 and i 6 146 1 : 3 : 5i Mild steel flat.. AXli 6 125 1:3:6 Tool steel round 3 4 6 147 The adhesive strength of 1 : 2 : 4 concrete to plain round rods appears to be about 400 pounds per square inch. In situations where a higher bond strength is required it is customary to secure a mechanical bond by the use of some form of deformed bar. ' 209. Ratio Ec/Ej. The relative moduli of elasticity of concrete and steel determine the relative stresses in the two materials when the com- bined concrete and steel member is deformed a given amount. So long as the bond is not destroyed the ratio Ec/Es fixes exactly the relative stresses in the concrete and steel. Ec for 1 : 2 : 4 concrete, the mix almost exclusively used for reinforced concrete, has beeji stated to be about 2,000,000 pounds per square inch (Art. 202), and for steel Es is about 30,000,000 pounds per square inch. The value of the ratio Ec/Es is, therefore about 1/15. 210. Fire Resistant Properties of Concrete. Concrete as a fire resistant has been subjected to various experimental trials, but the best proof of its value so used hes in the experience afforded by many very severe fires wherein concrete well demonstrated its superiority over most other materials, which are used for fire protection. The value of concrete as a protection for steel work in case of fire is due to several considerations. In the first place, concrete is in itself incombustible; second, its temperature coefficient is practically the same * Bull. No. 8 Univ. of lU. Eng. Exp. Sta. 200 MATERIALS OF CONSTRUCTION as that of steel, thus giving it an advantage over materials like terra cotta, which expands much more rapidly than does steel, and hence tends to fail by reason of the destruction of the bond caused by unequal expan- sion; third, the rate of heat conductivity of concrete is very low, due in part to its porosity and consequent air content, and in part to the dehy- dration of the water of chemical combination, the volatilization of which absorbs heat. This latter action increases the porosity, and hence the conductivity of the concrete which has suffered dehydration is still further lowered, and the penetration of the dehydrating action proceeds very slowly. The concrete which thus becomes dehydrated is seriously injured, but the effect is seldom appreciable to a depth of more than a small frac- tion of an inch, except in the hottest- and longeat-burning fires. Con- crete called " cinder concrete," in which the usual coarse stone aggregate has been replaced by cinders, has been found quite as effective a fire resistant as is stone concrete. In general it is considered that a covering of concrete over steel work, 2 inches in thickness, is sufficient to effectually protect the steel against temperatures sufficiently high to cause warping and twisting, with con- sequent failure of the structure. 211. Protection of Steel from Corrosion. Experience gained at the time of the demolition of reinforced concrete structures after years of exposure in damp situations, and carefully conducted experiments, as well, have shown that concrete forms a most effective preventive of the corrosion of steel imbedded therein. Particularly is this true if the concrete be mixed sufficiently wet so that the steel is completely covered by a wash of thin grout.* Experiments made by Professor Charles L. Norton for the Insurance Engineering Station in Boston led to the following conclusions: " (1) Neat Portland cement, even in thin layers, is an effective pre- ventive of rusting. " (2) Concretes, to be effective in preventing rust, must be dense and without voids or cracks. They should be mixed quite wet where applied to the metal. " (3) The corrosion found in cinder concrete is mainly due to the iron oxide, or rust, in the cinders and not to the sulphur. " (4) Cinder concrete, if free from voids and well rammed when wet, is about as effective as stone concrete in protecting steel." 212. Working Stresses and Factor of Safety. The following work- ing stresses are recommended by the Committee on Concrete and Rein- forced Concrete of the American Society of Civil Engineers.! * See Trans. Am. Soc. C.E., Vol. 71, p. 200. t Trans. Am. Soc. C.E., Vol. 66, p. 431. CONCRETE 201 The allowable compressive stress " on a plain concrete column or pier, the length of which does not exceed twelve diameters, is 22.5 per cent of the compressive strength at twenty-eight days, or 450 pounds per square inch on 2000-pound concrete." The factor on the basis of the twenty-eight day strength is therefore about 4.5. " The extreme fiber stress of a beam, calculated on the assumption of a constant modulus of elasticity for concrete under working stresses, may be allowed to reach 32.5 per cent of the compressive strength at twenty-eight days, or 650 pounds per square inch for 2000-pound con- crete." The apparent factor is about 3.1. (The actual factor is con- siderably larger.) " Where pure shearing stress occurs, that is, uncombined with com- pression normal to the shearing surface, and with all tension normal to the shearing plane provided for by reinforcement, a shearing stress of 6 per cent of the compressive strength at twenty-eight days, or 120 pounds per square inch on 2000-pound concrete may be allowed." The factor here is about 6 or 7; " Where the shear is combined ivith an, equal compression, as on a section of a column at 45 degrees with the axis, the stress may equal one- half the compressive stress allowed. For ratios of compressive stress to shear between and 1, proportionate shearing stresses shall be used." The factor is here again about 4.5. " The bonding stress between concrete and plain reinforcing bars may be assumed at 4 per cent of the compressive strength at twenty- eight days, or 80 pounds per square inch for 2000-pound concrete; in the case of drawn wire, 2 per cent, or 40 pounds per square inch on 2000- pound concrete." The factors are here about 4.5 and 2.25 respectively. " It is recommended that . . . the modulus of elasticity of con- crete ... be assumed as one-fifteenth that of steel, as, while not rigor- ously accurate, this assumption will give safe results." NON-CEMENTING MASONRY MATERIALS CHAPTER IX BUILDING STONES AND STONE MASONRY BUILDING STONES General 213. Stone as a Structural Material. The term " building stone " is applied to all those classes of natural rock which are employed in masonry construction. Stones form, with the exception of timber, the only important class of materials which may without alteration of their natural state be used directly in the construction of engineering works. Stone has been employed since prehistoric times in the construction of walls, dwellings, etc. Its use in masonry foundations, dams, piers, and even arches and bridges is very ancient; it has been used as an orna- mental material in types of ihasonry other than stone masonry since these types originated, and in the form of carved stone it has been one of the chief sources of architectural adornment of structures for the architects of all ages. Aside from purely structural uses, great quantities of stone are utihzed on other kinds of engineering construction. Of all the stone quarried in the United States about 40 per cent is used for building and monu- mental purposes as rough or cut stone, about 5 per cent is used for flag- ging and curbing, something over 6 per cent for paving blocks, over 21 per cent as crushed stone for road building, nearly 12 per cent as crushed stone railroad ballast, and over 16 per cent as crushed stone concrete aggregate. 214. Classification of Rocks. Geological Classification. In the usual geological classification rocks are divided into Igneous Rocks formed by consolidation from a fused or semi-fused condition; Sedimentary Rocks, formed by the sohdification of material transported and deposited by water; and Metamorphic Rocks, which are formed by the gradual meta- 202 BUILDING STONES AND STONE MASONRY 203 morphism of the structure and character of igneous or sedimentary rocks through the agency of heat, water, pressure, etc. Greenstone, basalt, and lava are common examples of igneous rocks; sandstone, hmestone, and shale, of sedimentary rocks; and marble and slate, of metamorphic rocks. The geological classification has only a Umited bearing upon the con- sideration of rocks as building stones. Igneous rocks are usually non- laminated and more or less crystalline in structure; sedimentary rocks are distinctly stratified, having, therefore, original cleavage planes; meta- morphic rocks may or may not be laminated, depending upon the pressure encountered during metamorphism. Most of the metamorphic rocks which have been changed largely through the agency of pressure, water, and heat are crystaUine in structure. Physical Classification.- The following classification of rocks on the basis of physical structure is made by Professor Baker: * With respect to the structural character of large masses, rocks are divided into stratified and unstratified. The structure of unstratified rocks is, for the most part, an aggregate of crystalline grains firmly adhering together. Granite, trap, basalt, and lava are examples of this class. Stratified rocks may be divided into the following classes according to physical structure: 1. Compact crystalline structure (quartz-rock, marble). 2. Slaty structure (clay and hornblende slate). 3. Granular crystalline structure (gneiss, sandstone). 4. Compact granular structure (blue limestone). 5. Porous granular structure (minute shells cemented together) . 6. Conglomerate (fragments of one stone embedded in mass of another). Chemical Classification. Stones are divided according to the chemical nature of their predominating constituents into the following three classes: 1. Siliceous stones, in which siUca is the predominating chemical constituent. (Granite, syenite, gneiss, mica-slate, greenstone, ba- salt, trap, porphyry, quartz-rock, hornblende-slate, and sandstone.) 2. Argillaceous stones, in which alumina is the important con- stituent. (Slate, and graywacke-slate.) 3. Calcareous stones, in which carbonate of lime is the predomi- nating constituent. (Marble and limestone.) *I. 0. Baker, "Masonry Construction.'' Baker uses the term "stratified" in a broad sense to include both bedded sedimentary and banded metamorphic roclcs. 204 MATEEIALS OF CONSTEUCTION STONE QUARRYING AND CUTTING 215. Methods of Quarrying. After the exposure of rock by the stripping of the surface soil the quarrying is done by hand tools, by machine tools, by the use of explosives, or by a combination of two or more of these methods. Hand Methods. Hand methods may be employed when the stone occurs in thin beds. Such stone, which is usually inferior in quality, may be taken out by use of the drill and hammer, wedges, plug and feathers, pick, crowbar, etc. Holes f to f inch in diameter are drilled a few inches apart in rows. The rock is thereupon split in the plane of the holes by the driving in of wedges or by use of plug and feathers. The plug is a short steel wedge with plane faces, and the feathers are wedges having one cylindrical and one plane side. Fig. 123. When a plug is inserted between two feathers the three will slip into a cylindrical hole, and a great spUtting force in one direction only may be exerted by driving the plug. The drills used are either jumper drills, the ordinary type of drill which is held by one man and turned between sledge blows struck by a second man, or churn drills, which are long heavy drills used without a hammer by one or two men. The drill is lifted, turned, and allowed to fall back in the hole, the force of the blow being due to the impact of the falling drill alone. Machine Methods. The use of machinery driven by steam, com- pressed air, or electric motors, is usually combined with hand methods whenever quarrying operations are conducted on any but a very small scale. The conamonest application of power-driven machines in the quarry lies in the use of machine drills which cut much more rapidly than hand drills and are usually arranged to work at any angle. Machine drills are of two general types: Percussion drills, the cutting tool of which resembles the hand drill and is operated by a piston and automatically turned a sm&,ll angle between strokes (steam or compressed-air power), and rotary drills, which are hollow tubes provided with an annular cutting edge. Rotary drills are sometimes provided with hardened steel teeth which constitute the cutting edge, but more often the annular end of the drill is set with a number of black diamonds which constitute the cutters. These cutters project slightly inside and outside the wall of the tube, so that clearance is provided and water may be forced down through the tube for the purpose of removing the debris which is washed up through the narrow space between the tube and the rock. The solid core of rock which remains within the tube may be broken off and removed from time to time by withdrawing the drill. Holes may be bored to great depths by joining successive lengths of drill-rod and the soUd core BUILDING STONES AND STONE MASONRY 205 affords an excellent indication of the exact nature, stratification, etc., of the rock passed through. When it is not desired to preserve the core intact the cutters may be so arranged as to cut a cyhndrical hole instead of an annular hole. When it is desirable that the stone be removed in rectangular blocks advantage is always taken of the natural cleavage planes of the rock. The quarry is worked in benches, the width of which correspond to the dimensions of stones that can be handled, and the height of which corre- sponds to the thickness of .the rock strata. Rows of holes are drilled parallel and perpendicular to the edge of the bench, wedges or plug and feathers are then inserted in the holes, and the blocks of rock thus split along the planes of the holes. " Under-cutting," the driUing of — and use of wedges in— a series of horizontal holes, is resorted to when not rendered needless by the stratification of the rock. When the rock need not be removed in large or rectangular blocks, particularly when the rock is to be subsequently crushed for road stone, ballast, or concrete aggregate, explosives may be used in the drilled holes. When stone is quarried for building and monumental purposes on a scale sufficient to justify the use of more elaborate mechanical equipment than the ordinary machine drills, or when very large rectangular blocks of stone are desired, it is the practice to use a special machine called a " channeler." The channeler is a machine operating on rails or guide bars, which operates a gang of cutters with which long and deep, but narrow channels are cut as the machine slowly moves along. The sides of large blocks are thus freed, and the blocks are subsequently freed from their beds by wedging, or undercutting and wedging, as may be necessary. Explosives. The explosives used include gunpowder, dynamite, and, rarely, nitroglycerine. Gunpowder must be the coarse, slow-acting, cheap powder rather than the high-power, quick-acting grade. Gun- powder is exploded by fuse or electric spark. The term dynamite is the name applied to any explosive consisting of some granular substance saturated with nitro-glycerine. Nitro-glycerine is a fluid produced by mixing glycerine with nitric and sulphuric acid. It is too quick acting for quarrying operations in general, since the rock is shattered rather than split. Its use is dangerous, owing to the ever-present possibility of escape of the fluid from drilled holes through seams in the rock to some distant place where it may lodge and constitute a menace to life during subsequent operations. Dynamite is called " true dynamite " when the granujar absorbent is an inert material; if the absorbent itself contains explosives the mix- ture is called " false dynamite." True dynamite may contain at least 50 per cent nitro-glycerine. False dynamites may contain not more 206 MATERIALS OF CONSTRUCTION than 15 per cent nitro-glycerine but the absorbent generally contains large amounts of oxygen which is liberated upon explosion and aids in effecting the complete combustion of the gases arising from the nitro- glycerine. The pressure of a little moist sand, clay, or, in the case of nitro-glycerine, a little water, provides sufficient tamping. Dynamite and nitro-glycerine are exploded by means of a percussion cap, which is a hollow copper cylinder about i inch in diameter and from 1 to 2 inches long, filled with a cement consisting of fulminate of mercury mixed with some inert substance. The percussion cap may be ignited by a fuse, but is more commonly ignited by an electric spark. » 216. Stone Cutting. Tools. The principal tools used in stone cutting are illustrated by Figs. 107 to 124.* Any description beyond that afforded by the figures is unnecessary and their uses are indicated in the discussion which follows. Surface Dressing.'^ All stones used in building come under one of three classes: I. Rough stones. II. Stones roughly squared and dressed. III. Stones accurately squared and finely dressed. The first class includes all stones used as they come from the quarry without any special preparation. The second class includes stones roughly dressed on beds and joints with the face hammer or axe. The distinction between this class and the third class lies in the closeness of the joints. When the dressing on the joints is such that the general thickness of mortar required is § inch or more the stones properly belong to this class. Three subdivisions of this class may be made, depending on the character of the face of the stone: (a) Quarry-faced stones are stones whose faces are left un- touched as they come from the quarry. (6) Pitch-faced stones are those the edges of whose face are made approximately true by use of the pitching chisel. (c) Drafted stones are those whose faces are surrounded by a chisel draft, the space inside being left rough. This method is not ordinarily used on this class of stones. The third class includes all stone dressed to smooth beds and joints so that the thickness of mortar joints is less than J inch. * A description of these tools and their use in dressing stone may be found in Vol. 6, Trans. Am. Soc. C.E., also in Baker's " Masonry Construction." t Trans. Am. Soc. C.E., Vol. 6. BUILDING STONES AND STONE MASONRY 207 As a rule all of the edges of cut stone are drafted. Inside the draft any of the following methods of dressing the face may be used. (a) Rough pointed; projections i or 1 inch. (Used on lime- stone and granite.) • (6) Fine pointed; projections less than ^ inch. (c) Crandalled; effect same as fine pointed except that the tool marks are more regular, |-inch projections. Cross-crandalled; if worked in both directions. (d) Axed; or pean-hammered; face covered with parallel chisel marks. • (e) Tooth-axed; same finish as fine pointed. (/) Bush-hammered; (usually used only on limestone), follows rough pointing and tooth-axing. ig) Rubbed; sawn surfaces smoothed by grit or sandstone, (used on sandstones and marble). {h) Diamond panels; face inside of draft cut to flat pyramidal form. PROPERTIES OF BUILDING STONES * General Description of Stones 217. Granite. Granite is the term applied to a plutonic,t igneous rock, whose structure varies from finely granular to coarsely crystalline. Its principal mineral constituents are quartz and feldspar, with varying amounts of mica, hornblende, etc. Its prevailing color is gray, though greenish, yellowish pink, and red shades are found more or less frequently. Granite is more extensively used as a building stone than any other class of igneous rock. It works with difficulty, due to its hardness and toughness, but its quarrying is usually facilitated by the existence of planes of weakness, the " rift " extending either in vertical or horizontal planes, and secondary planes, " the grain " along which the rock maybe less readily spht at right angles to the rift. As a rule the quarry rock shows " joints " or fissures in the direction of the rift, and often a second- ary series of joints exists in the direction of the grain. The removal of rectangular blocks of large or small dimensions is thus facilitated. Granite is used for foundations, base courses, columns, and steps * For detailed information concerning building stones consult " Stone for Build- ing and Decoration " by G. P. Merrill. Also see " Building Stones and Clay Products " by Dr. Heinrich Ries, and " Engineering Geology " by Dr. Ries and Dr. T. L. Watson. t Plutonic rocks are igneous rocks formed by the solidification of molten material prior to its emergence on the earth's surface, volcanic rocks have cooled on the earth's surface. 208 MATERIALS OF CONSTRUCTION Pigs. 107-113.— Tools Used in Stone Cutting. BUILDING STONES AND STONE MASONEY 209 I *+ -^" Fig. 114 CRANDALL Fig. U9 POINT PITCHING CHISEL Fig. 121 TOOTH CHISEL Fig. 120 CHISEL c^: Elg. 122 SPLITTING CHISEL Fig. 124 DRILL z> SCALE IN INCHES I I t t ■ ■ ' I I I I 1 1 12315 6789 10 11 12 Figs. 114^124.— Tools Used in Stone Cutting. 210 MATERIALS OF CONSTRUCTION in building construction, and is suitable for any situation where strength or hardness is required. To a Umited extent it is used as an ornamental stone, its suitability being dependent upon color and texture. Granite of excellent quahty may be obtained in any of the New England States, in most of the Southern States, in the Rocky Mountain Region, and in Cahfornia and Minnesota. 218. Gneiss. Gneiss has the same composition as granite and resembles granite in appearance but differs in physical structure, the various constituents being arranged in more or less parallel bands. The rock therefore sphts readily into flat slabs, which renders quarrying less expensive than in the case of granite and makes the stone valuable for foundation walls, street paving, curbing, and flagging. It is found in the same general localities as granite. 219. Limestones. The term hmestone is commonly applied to all stones which, though differing from one another in color, texture, structure, and origin, possess in common the property of containing carbonate of lime, calcite, or calcite and the double carbonate of lime and magnesia, dolomite, as the essential constituent. In addition they contain as impur- ities oxides of iron, silica, clay, bituminous matter, talc, etc. Different limestones may be hsted according to structure and composition and mode of origin under the following heads: 220. Crystalline Limestone or Marble. The term "marble" is commonly applied to any limestone which will take a good poUsh. It is properly applied only to those limestones which have been exposed to metamorphic action and rendered more crystalline in structure, the color being changed or even lost. The structure of marbles varies from finely crystalhne to coarsely crystalline. Marbles are found in almost every conceivable color, and are often richly streaked with several colors. All varieties of marble work well, the finer grained white marble being especially adapted to carving. Marble has been used in this country principally for interior decoration, but many varieties are entirely suitable for exterior construc- tion. Most of the colored and mottle|d and veined marbles are imported into this country. Quantities of white and black marbles are quarried, however, in Vermont, Georgia, New York, Pennsylvania, Maryland, and California, and some beautiful colored and mottled marbles are obtained from Tennessee and Vermont. g21 Compact Common Limestones. These are usually quite fine- grained limestones of varying textures and colors, giving rise to many varieties. The best known and most widely used American limestone is the Bedford limestone. This is a light-colored, fine-grained, oolitic limestone (made up of small rounded concretionary grains cemented BUILDING STONES AND STONE MASONEY. 211 together by carbonate of lime). It is a light gray in color, works with remarkable ease, and hardens on exposure. It has been extensively used for exterior construction of buildings, for bridge piers, and for heavy cut-stone masonry in general. It is quarried at Bedford, Indiana, and BowUng Green, Kentucky. Travertine is a compact fine-grained limestone deposited on the surface by running streams and springs. The term onyx or onyx marble is often applied both to travertine and to stalactite and stalagmite, which are deposits of limestone, often beautifully banded and streaked with colors, formed on the roofs, walls, and floors of caves. This use of the term " onyx," though incorrect, since it properly apphes to a banded variety of chalcedony, has come into such general use that it may as well be accepted as referring either to travertine or to stalactite and stalagmite. (Which, is immaterial, since the three are identical in composition and appearance, differing only in manner of chemical deposition). Onyx marbles (using the term in the sense above mentioned) differ from marbles of the common type only in that they are purely chemical deposits rather than products of metamorphism from pre-existing cal- careous sediments. The travertine varieties are probably products of deposition from hot springs carrying lime carbonate in solution, together with small quantities of iron and manganese carbonates and other more rarely encountered constituents. The stalactite and stalagmite varieties differ only in manner of formation, being cold-water depositions made on the roofs, walls, and floors of limestone caves. Both varieties owe their banded structure and variegated colors to the intermittent character of the deposition and the varying content of impurities hke the metallic oxides. The onyx marbles are considered the most beautiful of decorative stones, they cut readily and take a high finish, and are largely used for interior decorations. A large part of the onyx marble used in the United States is imported from Mexico. There are, however, important quarries in Arizona and California. The foreign onyx comes largely from Algeria and Italy. 222. Sandstones. " Sandstones are composed of rounded and angu- lar grains of sand so cemented and compacted as to form a solid rock. The cementing material may be either silica, carbonate of lime, an iron oxide, or clayey matter." * Sandstones vary greatly in color, hardness, and durabiUty, but include many of the most valuable varieties of building stone for exterior construction. The quahties of sandstones as a structural material depend * Merrill, " Stones for Building and Decoration," p. 299. 212 MATERIALS OF CONSTRUCTION largely upon the character of the cementing material, the character of the sand grains being very nearly the same for all, i.e., a pure quartz. If the cement is silicious the stone is light colored, hard, and sometimes difficult to work, but very durable. If iron oxides comprise the greater part of the cementing material, the color is a red or brownish tone and the stone usually is not too hard to work well, though it does not always prove very durable. If the cementing material is lime carbonate the stone is light colored, soft, and easy to work, but less durable than either of the above varieties. Clayey sandstones are the poorest class. They are soft and easily cut, but are particularly subject to the disintegration caused by weathering because of their high absorption. Some sand- stones contain very little cementing material, but owe their strength largely to the pressure, under which they have been solidified. Such stones are a light gray color, work easily, and if they possess sufiicient cohesive strength, are very durable. A few sandstones contain varying amounts of grains of feldspar or mica, in which case they are inferior to ones the grains of which are entirely quartz. The following are the prin- cipal well-known sandstones in this country : The brownstones of Connecticut, Massachusetts, Pennsylvania, New Jersey, North Carolina, and a few other locaUties, are handsome dark reddish-brown stones, fine grained, easy to work, and capable of taking a good " rubbed " finish. With the exception of the Massachusetts stone, they occur in distinctly laminated beds and must be used on their natural beds. These stones, having iron oxides for a great part of their cementing material, are usually subject to the disintegrating effect of atmospheric agencies and therefore do not usually rank especially high in durability, but have been largely used in all the large eastern cities. The Ohio stone, Berea sandstone, or Amherst stone is a fine-grained light buff, gray, or, blue-gray stone having silica for the most part as its cementing material, the amount of cement being low. These stones cut and work readily, are well adapted to carving, and, when those por- tions containing iron pyrites are excluded, are very durable. The prin- cipal quarries are located at Amherst, and at Berea, Ohio, and the stone has been largely used in the cities of the middle West. The Waverly stone is a fine-grained homogeneous stone of a light drab or dove color, quarried only in the vicinity of Cincinnati, Ohio. It is sometimes called the Euclid Uuestone. It resembles the Ohio stone except that it has a finer and more compact texture. It works easily and, except for portions containing iron sulphides, is a handsome and durable stone. The Potsdam red sandstone from Potsdam, New York, is composed wholly of quartz grains cemented by a small amount of silica with just BUILDING STONES AND STONE MASONRY 213 enough iron oxide to give it a reddish or brownish-red color. It is fine- grained, handsome in appearance, works well though rather hard, and is the strongest and most durable of sandstones after becoming hardened by seasoning. Its composition being practically quartzite, it is as strong as granite, and particularly non-absorptive, making it practically proof against all disintegrating influences. The Lake Superior stone is a Potsdam stone of medium fineness. It has a Ught red-brown color and is often spotted gray. It is quarried largely at Marquette, Michigan. This stone resembles the Potsdam stone of New York except that it has rather more cementing material and therefore is not quite as hard and strong. It works well and is very durable. Its use has been principally limited to Michigan cities. The Medina sandstone of Western New York is a hard moderately fine-grained stone, either red or gray in color. The red variety resembles the Potsdam stone, except that it is not so fine in texture, and is similarly used. The gray variety is rather too hard to work for general building purposes, but is largely used for street paving and curbing, where it has the advantage of some other hard stones, like granite or trap, in that it does not wear smooth. The Rocky Mountain sandstones include many varieties of excellent building stone. The best-known ones are very fine-grained, soft-textured stones of a dark-brown color. They work well, take a good " rubbed " finish, and are fairly durable. 223. Slates. Slates perhaps may not properly be considered as building stones, but their extensive use as roofing material and for various interior building uses classes them as a building material. Ordinary slate is a, siliceous clay, compacted and more or less meta- morphosed after deposition as fine silt on ancient sea bottoms. The pressure due to thousands of feet of overlying material is largely responsi- ble for the solidification of the clay into rock having very marked cleavage planes. The most valuable characteristic of slate is its pronounced tendency to split into thin sheets having smooth regular surfaces. The non-absorptiveness of slate, its great toughness and mechanical strength and its non-conductiveness for electric currents, are other valuable attri- butes. Slate must be spHt while fresh from the quarry, and the quarry loss amounts to more than 60 per cent of the rock. Physical and Mechanical Properties of Building Stones 224. Selection of Building Stone. The selection of a proper stone for construction purposes is dependent to a great extent upon the chmate where the stone is to be used. The range of changes of temperature, the average humidity of the atmosphere, the possibiUty of acid fumes in 214 MATERIALS OF CONSTRUCTION the atmosphere of many cities, and the possibility of the stone being subjected to high temperatures by fire, are among the considerations which must be carefully taken into account. Very often the only con- siderations given weight by an architect are the cost and the appearance. He is very apt to take great care to secure a color which harmonizes with the general scheme of the structure on which it is used, but wholly over- look the question of whether the stone chosen possesses satisfactory weathering qualities. The actual mechanical strength of stone is seldom of great importance because of the fact that stones in masonry structures can never be loaded to their full capacity because of the comparative weakness of the mortar joints. 225. Properties of Various Stones. Durability. The durability of stones depends upon ability to withstand weathering agencies, and the structure, texture, and mineral composition are the real determining factors. Joint planes, cracks, or other structural imperfections afford an opportunity for water to enter and for disintegration to begin through frost action. Stones of coarse-grained texture are more subject to the disintegrating influence of temperature changes than fine-grained ones, and dense stones, owing to their practical imperviousness, are less apt to be injured through frost action than are porous ones. Of the mineral compounds which make up our common rocks sulphides are among the least resistant to weathering agencies, iron compounds in general are un- desirable in large quantities, calcium and magnesium carbonates weather rather rapidly, aluminates weather less rapidly, and silicates or silica are most resistant to decay. The fact must not be overlooked in this connec- tion that the three factors, structure, texture and mineral composition, are simultaneously operative so that a very dense fine-grained stone made up principally of carbonate may weather well, while a porous or structurally imperfect stone made up principally of silica may weather poorly. The estimated life of various building stones is indicated by the fol- lowing table quoted by Ries and Watson * from the observations made by A. A. Julian in New York City. _, . Kind of stone. Life in Years. Coarse brownstone g j^o j^g Fine laminated brownstone 20 to 50 Coiiipact brownstone iqq ^^ 20O Bluestone (sandstone) untried, '.'..'.'.'.'.'.'.'.' ...Perhaps centuries Coarse fossihferous limestone 20 to 40 Fine oolitic (French) limestone 30 to 40 Marble, coarse, dolomitic 4q Marble, fine, dolomitic '.'.'.'.'.'.'.'.'.'.'.'.'.'.'..'.'. 60 to 80 ^^'^}^' fi°e 50 to 100 ^^^"'^^ 75 to 200 ^^'^^ 50 years to many centuries * " Engineering Geology," p. 432. BUILDING STONES AND STONE MASONRY 215 Absorption. The absorption or absorptive power of stones is repre- sented by the weight of water that can be absorbed expressed as a per- centage of the dry weight of the stone. Absorption is directly dependent upon the porosity of stones, though this relation is not necessarily any fixed ratio. The gain and loss of moisture when a stone is first exposed in a damp or wet situation and then dried, will be most rapid if the pores are large or straight, and least rapid if they be small or tortuous. The following figures constitute an abstract from the results of absorption tests made by Hirschwald.* The lower values are those found when submersion was rapid, the higher values when submersion was slow. No pressures above atmospheric pressure were used in these tests. Percentage Absorption. Percentage of Pore Volume. 4.89- 7.33 ■0.36- 0.49 7.51- 7.88 0.51- 0.55 22.11-23.41 0.51- 0.91 52 97-64 88 Marble 59.47-84.27 35 45-37 20 Slate 72 92-79 16 Tuff 65 51-69 71 Granite ... 41 20-57 71 It appears from these tests that the absorption of igneous and meta- morphic rocks rarely exceeds i of 1 per cent. (The tuff is a very porous volcanic- rock which is not used as a building stone and hence may be disregarded.) The sandstones absorb at least ten times as much as granites, marble, and slate, and the limestones absorb even more moist- ure than the sandstone. Expansion and Contraction. Stones, like most other materials, expand upon being heated and contract when cooled. Unlike most other mate- rials, however, they do not quite return to their original volume when cooled after heating, but show a swelling which is permanent. Experi- ments made at the Watertown Arsenal f by heating from 32° to 212° F. and cooling through the same range, showed the following amount of permanent increase in length for the various stones tested: PERCENTAGE PERMANENT INCREASE IN LENGTH Min. Max. Ave. Granites 0.0085 0.0015 0.0120 0.0145 0.0355 0.0870 0.0595 0.0980 0.0200 Sandstones Limestones Marbles 0.0235 0.0350 0.0450 * Hirschwald, " Handbuch der Bautechnischen Gesteinspriifung." t " Tests of Metals," 1895, p. 322. 216 MATERIALS OF CONSTRUCTION The coefficient of temperature expansion per degree Fahrenheit for various building stones was found in a series of tests at the Watertown Arsenal * to be very variable, the range of values found being as fol- lows: Granites, 0.00000311 to 0.00000408; limestones 0.00000375 to 0.00000376; marbles 0.00000361 to 0.00000562; sandstones 0.00000501 to 0.00000622. Frost Resistance. Stones can be disintegrated by frost action only when the pores are practically filled with water before exposure to freezing temperatures. As stones seldom are used under such conditions .that the maximum amount of water is absorbed, instances of injury to good build- ing stones by frost action are very rare. Experimental work on the resistance of. stones to disintegration by frost indicates that the pores can be filled with water, so that the subsequent expansion upon freezing will cause rupture, only by the use of high pressure or by first exhausting the air by a vacuum. It will be apparent therefore that only stone of the greatest absorptive power combined with low structural strength can ever be injured by frost action under the conditions encountered in practice. Fire Resistance. Practically all building stones are seriously injured if exposed to such high temperatures as may be encountered in case of fires, and particularly so if exposed to the combined action of fire and water. The cause of disintegration is usually explained to be the internal stresses caused by unequal expansion of unequally heated portions of the material. This explanation is rendered more forcible by the observed fact that if highly heated stones be suddenly cooled on the exterior by apphcation of water, the resultant disintegrating action is much more pronounced than when the cooling is slow. The texture of the stone, and the relative coefficients of expansion of its individual mineral con- stituents probably are also factors of importance. Experience has shown that granites are particularly poor fire resist- ants. Probably on account of the irregularity of the structure and the complexity of the mineral composition, granites crack irregularly and spall badly. The coarbe-grained granites are most susceptible to the action of fire and water, and the gneisses often suffer even more severely because of their banded structure. Limestones suffer little from heat until a temperature something over 600° C. is reached, at which point the decomposition of the stone begins, owing to the driving off of carbon dioxide. The stone then has a tendency to crumble due to the flaking of the quicklime formed. Curi- ously enough the limestones do not suffer so much by sudden cooling as by a slow cooling. * " Tests of Metals," 1890, p. 1108. BUILDING STONES AND STONE MASONRY 217 Marbles, owing to the coarseness of the texture and the purity of the material, suffer more than limestones at temperatures below the point where calcination begins. The cracking is irregular and the surface spalls off as in the case of granites. Sandstones, especially if of a dense, non-porous structure, suffer from high temperatures and sudden cooling less than most other building stones. (We except limestones which are good fire-resistants only below the temperature of calcination.) The cracking of sandstones which does occur, appears mostly in the planes of the laminations, which should be horizontal planes as the stone is set. These cracks are therefore not as serious: as irregular cracks. Sandstones whose cementing ingredient is silica orilime carbonate are better fire resistants than ones whose grains are bound by iron oxide or clay. •Mechanical Properties of Stones. The following tabulation of mechan- ical properties of the principal classes of building stone has been com- piled from the Watertown Arsenal Tests * of 1894 and 1895 to serve as an approximate guide in the selection of a building stone. It will be noted by reference to the table that the properties of the different varieties of building stones of the same general class vary greatly. It is, therefore, not advisable to use the average strength indi- cated except in a very general way, the safe working stress being deter- mined by the use of a large factor of safety. (Taken at from 15 to 35, dependingi upon the structural use made of the stone, and the amount of variation shown in tests of the particular variety of stone used.) For all situations involving high stresses, as in the case of monolithic columns in buildings, the stone should be chosen only after tests have been made of the particular stones under consideration. STONE MASONRY 226. Classification.f AH stone masonry is classed as rubble masonry, squared stone masonry, or ashlar or cut-stone masonry. " Rubble masonry includes all stone masonry composed of unsquared stones. It may be uncoursed ruhhle, Fig. 125a, laid without any attempt at regular courses, or coursed rubble, Fig. 1256, leveled off at specified heights to a horizontal plane. ' i Squared Stone Masonry: "According to the character of the face, this is classified as quarry-faced. Fig. i25c, or as pitch-faced, , Fig. 125d. If laid in 'regular courses of about the same rise throughout, it is range work, Fig. 12.5e. If laid in courses that are not continuous throughout * "Tests of Metals," 1894, 1895. t Trans. Am. Soc. C.E., Vol. 6. 218 MATERIALS OF CONSTRUCTION MECHANICAL PROPERTIES OF BUILDING STONES Name. Locality. Comp. Strength. Lbs. per Sq.in. Mod. of Rupture. Lbs. per Sq.in. Shearing Strength. Lbs. per Sq.in. Weight, Lbs. per Cu.ft. Mod. of Elas., Lbs. per Sq.in. Granite: Little Rock, Ark . . Millbridge, Me. . . , Chesterfield, Va. . , Korah, Va Exter, Cal Rockville, Minn. . . Sioux Falls, Minn . Troy, N. H JBranford, Conn . . . Milford, Mass Average Marble : (White) Rutland, Vt (Blue) Rutland, Vt (Dark) Rutland, Vt Sutherland Falls, Vt. . . . (Fossil) St. Joe, Ark (Brown), St. Joe, Ark. . . (White) DeKalb, N. Y. Marble Hill, Ga Tate, Ga Average Limestone: La Motte, Vt Dodge Co., Minn. . Junction Cy., Kan. Fort RUey, Kan. . . Beaver, Ark Mt. Vernon, Ky. . . Rockwood, Ala. . . . Bowling Gr., Ky . . Bedford, Ind Average Sandstone: Cromwell, Conn E. Long Mead., Mass. Jasper. Ala Pike Co., Minn Cabin Creek, Ark .... Fort Smith, Ark Redfield, Kan Oakland, Cal Coos Co., Ore Olympia, Wash Chuckanut, Wash .... Tenino, Wash Pittsburg, Wash 21,559 19,917 15,350 23,446 22,557 18,121 18,176 26,174 15,707 21,235 Average 12,531 20,224 11,892 13,864 12,833 16,156 10,312 12,278 12,497 11,505 12,425 12,640 14,622 4,522 3,173 20,581 7,647 5,957 6,043 9,918 9,058 16,894 10,004 15,630 12,647 18,468 12,765 8,027 11,041 7,444 12,655 11,533 6,688 19,208 1520 2048 1672 1608 1853 1327 1216 2169 1249 1545 1630 1247 2057 1759 2293 1615 1614 839 1038 1079 1505 1640 253 628 2707 1314 690 1058 1736 1241 1872 941 1889 868 1666 1473 2088 1063 '2i85' 1488 495 1457 2206 2820 2066 2662 2419 1949 2086 2214 1834 2311 156,2 164.7 161.6 162.6 2267 1023 1217 1453 1565 1332 1316 1318 2173 1135 1022 1998 1705 978 1211 1119 1418 1526 1199 2202 1555 2479 1766 1940 1626 1248 1643 1809 1226 1685 161.3 167.6 167.8 168.7 168.0 139.1 139.1 133.8 133.8 7,040,000 9,800,000 9,420,000 6,010,000 6,997,000 5,927,000 7,632,000 4,456,000 7,267,000 9,290,000 12,680,000 8,210,000 10,740,000 12,000,000 5,515,000 4,020,000 8,231,000 14,720,000 6,645,000 4,000,000 9,290,000 7,160,000 8,363,000 7,711,000 1,582,000 3,900,000 3,630,000 3,277,000 2,120,000 1,020,000 3,306,000 BUILDING STONES AND STONE MASONRY 219 the length of the wall, it is broken range work, Fig. 125/. If not laid in courses at all, it is random work, Fig. 125g." Fig. 126a UNCOURSEO RUBB1.E RUBBLE Kg. 125b COUTiSED RireST-E SQUARED STONE QUARRY PITCH FACED FACED RANGB Fig. 125c FlK.USd Fig. 1256 BROKEN RANG£ Fig. 125f II II 1 1 1 1 1 ^ 1 1 ASHLAR Flg.l25h ASHLAR RANDOM Fig. 123g 1 , \^ 1 \ — r' — 1— \ , 1 - 1 J 1 1 1 1- — - 1 _| "~— ' — — -^ h- " BROKEN ASHLAR Fig. 1251 1 1 1 1 1 _ — J |_ 1 1 - — F : 1 1 - - 1 1 — - _^ 1 1 ■ T- — 1 ]_ Fig. 125. — Stone Masonry. Ashlar Masonry. "This is equivalent to cut-stone masonry, or masonry composed of any of the various kinds of cut stone mentioned above." 220 MATERIALS 'OF CONSTRUCTION As a rule the courses are continuous, Fig. 125^, but sometimes are broken by the introduction of smaller stones of the same kind and then it is called broken ashlar, Fig. 125i. 227. Compressive Strength. The actual compressive strength of stone masonry has not been satisfactorily determined experimentally. The few tests which have been made do not form a sufficient basis for the determination of the relative strengths of the different classes of masonry, nor the relative strengths of masonry of the same class when different kinds of stone are used. The following data are abstracted from the report upon a limited series of tests made by the Austrian Society of Engineers and Architects in 1898-1901. COMPRESSIVE STRENGTH OF MASONRY PIERS Compressive Strength, Single Stone. Lbs. per Sq.in. Compressive Strength, Masonry Piers. Ratio of Strength of Masonry to Strength of Stone. 1 : 2 Mortar Lbs. per Sq.in. 1 : 3i Mortar, Lbs. per Sq.in. 1 : 2 Mortar, Lbs. per Sq.in. 1 : 3i Mortar, Lbs. per Sq.in. 10,380 8,780 3220 3870 2730 2820 33.6% 40.4 28.5% 29.4 Ave. 9,580 3545 2775 37.0 29.0 The piers tested with the above results were built of hard sandstone blocks about 6 inches thick, and either about 12 by 20 inches, or 7 by 9 inches, in transverse dimensions. The stones were laid with either 1 : 2 or 1 : 3i Portland cement mortar, joints ^ to f inch thick, the vertical joints being broken. The whole constituted a pier about 40 inches high, with a 20 by 20-inch base, having six courses and seven mortar beds. There is no opportunity afforded by these tests to compare the rela- tive strength of masonry with stones of different strength or with different kinds of mortar other than the two mixes of Portland cement mortar. It is evident that the strength of any masonry will be largely dependent upon the mortar used, and it is also a fact that the masonry will be strongest when the joints are thinnest. An analogy may perhaps be considered to exist between the relative strength of stone masonry and the stone employed therein, and the relative strength of brick masonry and the brick employed therein. Experimental data to fix the latter point are by no means lacking, since comprehensive series of tests have been conducted in several different laboratories. A series of tests conducted at the Watertown Arsenal in 1904 has been used to abstract the following data: Tests of six classes of brick BUILDING STONES AND STONE MASONEY 221 have been selected, including tests of dry-pressed face brick, repressed mud brick, wire-cut mud brick, hard sand-struck brick, and light hard sand-struck brick. These bricks were tested singly in compression, and in piers built with a 12 by 12-inch base, 8 feet high. The mortars used were neat Portland cement; 1 : 3 Portland cement mortar, and 1 : 3 lime mortar. In spite of the wide variation in the strength of the differ- ent classes of brick used, the variation in the ratio of the strength of the masonry to the strength of the brick was only within narrow limits for each class of mortar used. For neat cement the ratio varied from about 23 per cent to about 28 per cent, averaging about 26 per cent; for 1 : 3 cement mortar the ratio varied from 19 per cent to about 28 per cent, averaging about 24 per cent; while for the 1 : 3 lime mortar the ratio varied from about 10 per cent to about 18 per cent, averaging about 13 per cent. It would appear then that the strength of stone masonry, if the above- mentioned analogy holds, is dependent almost entirely upon the strength of the mortar joints, the actual strength of the stone not being a great factor except with the strongest mortars laid with the thinnest joints. The manner of failure of masonry under compression is almost invar- iably by the compressive failure of the mortar, followed by lateral flow of the latter, thus setting up tensile stresses in the stone which open up longitudinal cracks. Actual compressive failure of the stone in masonry is very rare indeed. 228. Allowable Loads on Stone Masonry. The building laws of the city of Chicago recommend the following values as safe pressures for the different classes of stone masonry: ALLOWABLE LOADS ON STONE MASONRY Kind of Masonry. Pounds Sq. in. Rubble, uncoursed, in lime mortar 60 Rubble, imcoursed, in Portland cement mortar 100 Rubble, coursed, in lime mortar 120 Rubble, coursed, in Portland cement mortar 200 Ashlar, limestone, in Portland cement mortar 400 Ashlar, granite, in Portland cement mortar 600 CHAPTER X BRICKS AND OTHER CLAY PRODUCTS GENERAL 229. Clay Products as Structural Materials. The principal clay products used structurally are building brick, paving brick, fire brick, terra cottas, and various forms of tiles. Brick has been used since the earliest times as a masonry material. Brick masonry possesses the great advantage of being exceedingly strong and durable, cheap as compared with stone masonry, and, owing to their artificial character, a wide lat- itude is possible in size, form, color, and structural character of the bricks themselves. Brick may be made from a large number of different classes of mate- rial, the one essential being that the material be of a mineral nature and possess a considerable degree of plasticity when dampened. Common building bricks are usually made of a mixture of clay and sand (to which coal and other foreign substances are sometimes added), which is mixed and molded in various ways, after which it is dried and burnt. Paving bricks are made primarily as a material for street pavements. Certain classes of paving bricks are largely used, however, as a substi- tute for building brick. Fire bricks are of such a nature that they will withstand high tem- peratures. Their structural uses are largely confined to linings of flues, stacks, etc. Terra cotta is made of selected clays in much the same way as ordi- nary brick. The architectural terra cotta is much used for decorative effect on building interiors and exteriors, and terra cotta lumber and hol- low blocks are much used for both interior and exterior walls of buildings and as fireproofing to protect steel-work. Tiles, made by burning various classes of clay are used in various forms as roofing tile, wall or floor tiles, drain tiles, sewer pipe, etc. 230. General Classification of Bricks and Clay Products. The principal classes of bricks and other clay products which are used as 222 BRICKS AND OTHER CLAY PRODUCTS 223 building materials, are included in the following classified list, in which certain subdivisions of the materials above mentioned have been made: (1) Building bricks. (a) Common building brick. (h) Face, or pressed brick. (c) Enameled or glazed brick. (d) Ornamental brick, tapestry brick, and Roman tile. (e) Hollow brick. if) Sand-hme brick (not a clay product). (2) Paving bricks. (3) Fire bricks. (4) Terra cotta. (a) Architectural. (6) Terra-cotta lumber. (c) Hollow building blocks and fireproofing. (5) Roofing, wall, and floor tiles. (6) Drain tile and sewer pipe. The manufacture, properties, and uses of each of these classes of building material will be briefly discussed in the paragraphs which follow. MANUFACTURE OF BUILDING BRICKS 231. Kinds of Clay. Their Use in Brick Making. The different classes of clay from the standpoint of geological manner of formation have been discussed above in Art. 83. Of the three general classes, resid- ual, sedimentary, and glacial clays, all are used for brick making when they possess a sufiicient plasticity for molding and burn to a body of the proper hardness, but sedimentary clays are most frequently found satisfactory. The following classification of sedimentary clays is made by Ries: * Classification of Sedimentary Clays. (a) Marine clays or shales (deposits often of great extent). White-burning clays. Ball clays and plastic kaoUns. Fire clays or shales, buff burning. Impure clays or shales. (1) Calcareous. (2) Non-calcareous. * Dr. Heinrich Ries, " Economic Geology, " p. 128. 224 MATERIALS OF CONSTEUCTION (b) Lacustrine clays (deposited in lakes or swamps). Fire clays or shales. Impure clays or shales, red-burning. Calcareous clays, usually of surface character. (c) Flood-plain clays (usually impure and sandy). . (d) Estuarine clays (deposited in estuaries, mostly impure and finely laminated). Marine deposits of clay often stretch for hundreds of miles with a depth of 30 feet or more. Their composition is remarkably uni- form, and, except in the case of those which are too high limed or are excessively plastic, include the best clays obtainable for brick manu- facture. Lacustrine and estuarine clays occur in beds of limited extent and usually of no great depth. They may be sandy and are less advantageous for use in brick-making than are the marine clays. Flood-plain clays are sometimes very sandy or even contain large pockets of pure sand. They are sometimes very calcareous, and cannot usually be advantageously used. 232. Influence of Kind of Clay upon Character of the Brick. Most clays used in brick making contain, in addition to sihcate of alumina, variable quantities of lime, magnesia, iron oxide, and alkalies. A very plastic clay is apt to shrink, crack, and warp in drying and be very hard after burning. The presence of coarse sand in suitable amounts tends to prevent shrinking and cracking in burning, but an excess of sihca in the shape of sand destroys cohesiveness. Iron oxide acts as a flux and adds greatly to the hardness and strength of brick. It causes the clay to burn buff or red in color according to the amount of iron oxide present. Lime, considerably in excess of iron, causes the brick to burn buff and shrink strongly as vitrification is approached. The lime must be in a very finely divided state so that it will be completely hydrated or fluxed in the process of manufacture of the brick. The presence of lumps of unhydrated lime which become hydrated after long periods is the cause of unsightly defects on brick walls called "lime-pops." Magnesia and alkalies also act as fluxes. Hand Phocesses of Manufacture 233. Preparation of the Clay. A small proportion of the common brick made and used, apart from the large cities, is still made by hand, and many specially molded brick made for various special purposes are made by hand methods. BEICKS AND OTHER CLAY PRODUCTS 225 The clay must first be freed from any pebbles, soil, excessive sand, etc., by washing. This is occasionally done in a wash mill similar to those described in connection with the wet process of cement manufacture. More frequently screens are employed. With the better class of clays this step is not required. Clays not put through a wash mill sometimes require crushing to reduce them to a state in which they readily mix with water. This is usually accompUshed in either a set of roUs or in an edge-runner mill (dry pan), such as has been described above. Clays which re- quire an excessive amount of crushing are not usually used in hand processes. After crushing the clay is tempered by mixing in a moderate amount crf^ater and then turning it by hand and allowing it to stand for a time /before the final pugging of the mix. 234. Pugging. The final reduction of the mix to a plastic mass is done in a pug-mill. At this stage it may be necessary to add sand to reduce shrinkage, and more water, the exact proportions being carefully controlled -and kept uniform. Th^^ug-mill, Fig. 57, is usually a horizontal cylinder provided with one/or two power-driven blades. The revolving blades slice up and the mass till it is ejected through an opening at one end. 235. Molding. There are two common methods of hand-molding. In the first, known as slop-molding, the mold is dipped in water just before filling to prevent the adhesion of the clay to the mold. The workman kneads the clay into approximate shape with his hands, forces it into the mold and tamps it hard, and then strikes it even with the mold by the use of a straight-edge. The mold is not removed till the brick reaches the dryer room. This method is slow and yields many imperfectly formed bricks. The second method is called sand-^molding, and differs from the first in that, instead of dipping the molds in water to prevent the clay from sticking, the mold is sprinkled with sand to effect the same purpose. The operation of filling the mold is practically the same as in the ^rst method, but, owing to differences in the details of manipulation, th&^second method is more rapid and the bricks are usually cleaner and sharper than those produced by slop-molding. The latter method is now seldom used. / 236. Dr3ring. After being shaped in the mold the bricks are allowed to dry for several weeks before being fired. Artificial dryers are seldom used in connection with hand processes. Very often the drying is accom- plished in the open air on racks which support the brick in such a manner as to allow a maximum access of air and sunlight to all sides of the tiers of brick. In order to prevent injury by rain it is desirable to cover the 226 MATERIALS OF CONSTBUCTION tiers with a light roofing of boards. For the better quaUties of bricks sheds containing racks must be used, or an artificial dryer installed. These sheds are permanent structures provided with a weather-tight roof, and with side walls which are either fitted with shutters or which are built of perforated bricks. Drjdng requires a length of time which is dependent upon the amount of water contained in the brick. Soft- mud brick and slop-molded brick often require from three to six weeks drying; stiff-mud brick and sand-molded brick are sometimes fired after less than one week on the drying racks. 237. Pressing. When hand-made brick are to be pressed the dry- ing must be carefully watched and they must be taken to the press before becoming too dry and hard. The press used is a portable one which is operated by hand. The bricks are placed one by one in the machine between dies and compressed by a piston operated by a lever. The bricks are pressed and replaced on the rack as fast as they can be handled, the press being wheeled along between the racks as the work proceeds. The kilns used in hand processes differ in no respect from those used for machine-made brick (Art. 242), except that the more elaborate types of kilns and especially the continuous kilns are seldom used in connection with hand molding. Machine Processes 238. General. Brick-making upon a large scale is now done almost entirely by machinery, from the mining of the clay by steam shovel to the molding in automatic machines and burning in semi-automatic kilns. The equipment used for the manufacture of machine-made brick varies greatly, on account of variations in the character of the raw material available, and variations in the class and quality of brick desired. There are three general processes employed in brick manufacture, viz. : The soft-mud process, the stiff-mud process, and the dry-press process. These processes must be separately considered. ~ 239. Soft-mud Process. Preparation of the Clay. The soft-mud process is practically the same as that employed in hand processes except that hand work is largely replaced by machines. The clay, unless it is found very free from pebbles and excess sand, is first cleaned in a wash- mill, after which it is tempered, coal dust or sand being added, if desired, in the tempering tanks, and then pugged in the machine which fills the molds. Molding and Drying. The upper part of the molding machine is virtually a pug-mill which delivers the pugged mass to molds below where It IS pressed into place by a plunger. The filled mold is now moved to BRICKS AND OTHER CLAY PRODTJCTS 227 one side and a second mold is brought into place and filled while the brick last made is being removed from its mold. The amount of pres- sure exerted on the clay is necessarily under constant control since vari- ations in the stiffness of the clay make variations in pressure necessary. The machine generally uses gang-molds and so makes perhaps four or six bricks at each stroke of the plunger. The molds are usually sanded by htod between reversals of the table which shifts the molds back and forth under the plunger. • The capacity of a machine is about 8000 to 10,000 bricks per day, which is about twice or three times the speed of hand-molding of sand-struck brick, and five or six times the speed pos- sible in the molding of slop-brick by hand. The drying of the soft-mud machine-made brick is often accomplished in the same manner as has been described above, referring to hand- made brick. In some instances artificial drying is resorted to, the equip- ment being that described below. ,-■•^40. Stiff-mud Process. Preparation of the Clay. The making of brick by the above-described methods means the incorporation in every brick of a poimd or more of water which must later be dried out and driven off in the kiln. A process whereby the brick clay is only sufficiently moist to possess the requisite coherency under moderate pressure results therefore in economy of time in drying, and economy of fuel in burning. This process requires in the first instance a clay which is not too wet. The lumps are sometimes broken up with the addition of water in a pug- mill or mixer, or the clay may even be finely disintegrated in an edge- runner mill if necessary (Fig. 56). Usually it is pugged with or without the addition of water, either in a separate pug-mill or in the brick-making machine itself, without previous grinding. Molding. The brick-making machine may be either of two general types, viz., the auger type (Fig. 126) or the plunger type (Fig. 127). The auger machine consists of a closed tube of cylindrical or conical shape, in which, on the line of the axis of the tube, revolves a shaft to which is attached the auger and auger knives. The knives are so arranged as to cut and pug the clay and force it forward into the auger. The function of the auger is to compress and shape the clay and force it through the die. When the clay passes through the die it is compressed to as great an extent as it can be in its semi-plastic condition. "The opening in the die is made to conform to the dimensions of either the end or the side of a brick, and a continuous bar of clay is forced through onto a long table where it is cut into sections the size of a brick. If the cross-section of the bar is the same as the end of the brick the brick are called end- cut, and when the section corresponds to the side of a brick they are 228 MATERIALS OF CONSTRUCTION side-cut. When end-cut brick are made the clay often issues from the machine in several separate streams. Fig, 126. — Auger Type Brick Machine. Fig. 127.-Plunger Type of Brick Machine with Wire-cutting Table. rru°r ""V^^ commonest types of cutting table is shown in Fig. 127. The bar of clay as it issues from the die travels along the table on an BRICKS AND OTHER CLAY PRODUCTS 229 endless belt which is supported on rollers. At intervals the operator throws a lever which swings downward a rigid frame across which a series of wires are tautly stretched. Thus the bar is cut into sections of either the length or the thickness of a brick. When the lever is reversed the wire frame is restored to a vertical position and the cut brick are carried away on a belt conveyor. Many types of machines are used for cutting the bar of clay into bricks, but all use wires to do the actual cutting. The great majority of bricks made by the stiff-mud process are made in the auger type of machine. The plunger type of machine is used to some extent, however. In the plunger type of machine (Fig. 127) the clay after pugging is forced into a closed chamber which acts as a feeder for the pressure \mm/'mM//MM////MMM/M///M////M/MMM [3- h//M///MMW/MMM/MMMWMMMM/M/^. WM//mMM///MMM//MMM/7ZS^ — s •m/mMMMMMMMMM/M/^A Fig. 128.— Tunnel Dryer. cylinder into which it discharges, the' amount discharged being subject to control. The forward motion of the plunger now compresses the clay and forces it through the die. The clay bar thus formed passes onto the cutting table and is cut by wires in the manner above described. A few plunger machines are so devised that the clay is forced into molds instead of through a die. With this exception, all stiff -mud brick are wire-cut brick. Drying and Pressing. Some manufacturers of stiff-mud brick burn the brick without any intermediate drying. The practice is not a good one, however, since it is apt to cause excessive warping and cracking of the brick unless the firing is done with great care and very slowly. The average stiff-mud brick is improved by a moderate amount of drying before burning. The drying racks or dry-houses above described are sometimes used, but, since the bricks contain much less moisture and are stronger, the drying may be accelerated and artificial heat used. 230 MATERIALS OF CONSTRUCTION The continuous tunnel dryer (Fig. 128) is most commonly used for rapid drying of green bricks. The bricks are piled on cars which move slowly through tunnels wherein they are heated either by hot air or by steam pipes. The tunnels are usually built of masonry, are 100 feet or more in length, and y^- - the cars traverse the length of the tunnel in about twenty-four hours. The usual product of the stiff-mud brick machine may be sorted into a small proportion of the face brick and a large proportion of com- mon brick. When, how- ever, a special face-brick is desired, the brick may be repressed in a mold under a plunger worked by a steam cylinder. Fig. 129 shows a com- mon type of repressing machine. Repressing re- shapes the brick, rounds the corners if desired, trues it in outline, and improves its appear- ance. Well made stiff- mud brick are apt to be structurally injured and weakened by re- pressing. 241. Dry-press Process. Preparation of the Clay. The dry-press proc- ess is especially fitted for the handhng of clays which contain not over 12 to 15 per cent moisture when they come from the bank. The clay is quarried by steam shovel and stored under cover to allow further drying, ground up in a dry pan (edge-runner), delivered to a mixer, which is simply an open pug-mill, in which it is thoroughly mixed to give a homogeneous product of about the consistency of flour, and discharged into the hopper of the brick machine. Molding. The molding of dry-press brick is a difficult operation, Fig. 129. — Brick Re-pressing Machine. BRICKS AND OTHEE CLAY PRODUCTS 231 the success of which is largely dependent upon the class of clay employed and the efficiency of the brick machine used. The type of machine illus- trated (Fig. 130) is one of the most satisfactory ones in general use. In this machine the clay is fed into the die or molds by a reciprocating charger located below the machine hopper. At each revolution of the machine the charger moves forward, and when it is directly over the molds the bottom plunger in the molds descends, allowing the molds to be filled with clay. The charger is withdrawn, the clay supply shut off, and the top and bottom plungers move toward each other in the molds, compressing the clay between them. The pressure is now relieved and then applied a second time, the compression of the clay being car- ried a bit further than in the first place. The upper plunger is now withdrawn and the bottom plimger raises the brick to the level of the top of the mold. The next stroke of the plunger pushes the finished brick upon the mold table, whence it is removed to the dryers or kilns. Many machines used with dry clay are arranged to mold two, four, or more bricks at each stroke. ■ Dry-pressed bricks are very compact, show high compressive strength, and are well formed, but they are not generally considered to be as durable as bricks the clay for which has been tempered. They are, however, largely used as face brick. The term " pressed brick " is properiy used only in referring to bricks made by the dry-clay process. The so-called pressed brick made by repressing of soft or stiff-mud brick should be called " repressed brick." Building bricks are occasionally made by a process called the semi- dry clay process. It differs Httle in methods or equipment used from the dry-clay process. 242. Kilns and Burning. Brick kilns may be divided into Wo gen- eral classes, viz., intermittent kUns, and continuous kilns. Intermittent Fig. 130. — Dry Clay Brick Machine. 232 MATERIALS OF CONSTRUCTION kilns may be further sub-divided into up-draught kilns and down-draught kilns. Continuous kilns are seldom used in connection with hand mold- ing or in small plants of any type. Up-draught Kilns. The up-draught kiln was at one time almost exclusively used in this country and is still largely used in small yards where the hand process is used. The old-fashioned up-draught kiln is nothing but the green bricks themselves built into a pile about 20 to 30 feet wide, 30 to 40 feet long and perhaps 12 to 15 feet high. The sides and ends are plastered with mud to keep in the heat, and the top is covered with earth and sometimes roughly roofed. The bricks are piled in such a way as to form a series of arched open- ings extending entirely across the kiln, and in these arches the fires are built. The brick nearest the fires are badly over-burned, sometimes to vitrification, and are called " arch brick." The brick at the top of the kiln are under-burned, and are called " salmon brick," and only the inter- mediate brick are first-class building brick. The modern up-draught kiln has permanent sides made of 12 or 16- inch brick masonry walls, and the heat is generated in ovens outside. The flames and hot gases enter the kiln through fire passages in the walls and arches made through the brick within the walls. This type of kihi is more economical of fuel than is the early type, and yields a much larger percentage of first-class brick. The burning of a kiln requires about a week, after which time the openings are tightly closed and the kiln allowed to cool very slowly. Down-draught kilns. Kilns of this type require permanent walls and a tight roof. The floor has openings connecting with flues leading to a stack. Down-draught kilns are usually built in a circular or bee-hive shape, but are some- times rectangular. Fig. 131 shows a circular kiln in vertical section, and Fig. 132 shows the same kiln in horizontal section, the right half Fig. 1 31. —Down-draught Kiln. Vertical Section. BRICKS AND OTHER CLAY PRODUCTS 233 Fig. 132.— Down-draught Kiln. Horizontal Sections. being a section above the floor, and the left half a section through the chimney flues below the floor. Heat is generated in outside ovens and the flames and gases enter the kiln through vertical flues {b, Fig. 131), carried to about half the height of the kiln. The heat therefore enters the brickwork at the top and is drawn downward by the chimney draught to the flues below the floor, and thence to the chimney or stack. The effi- ciency of the down-draught kiln is much higher than that of the up- draught kiln, and terra cotta and pottery, as weU as brick, are burned very evenly in this type of kiln. Continuous Kilns. Many patent- ed types of continuous kilns are on the market, but all depend on practi- cally the same principle. A number of chambers are connected in series, and also individually connected with a stack. The stack flues and the flues between chambers are pro- vided with dampers. While one chamber is burning, the waste products of combustion are forced to traverse the whole series of charged cham- bers before reaching one which is open to the stack. The material is thus preheated before being fired, and a considerable heat economy effected. The Hoffmann ring kiln (Fig. 13) is typical of this class of kiln. Fuel is supplied to the chambers through holes in the roof of each chamber, and in the more modem types is burned on grates or in troughs placed in the upper part of each chamber. The flue openings are in the floor so that the down-draught principle is utilized. This type of kiln is expen- sive to install, but is more economical of fuel than any other kiln. The percentage of first-class brick is also high, provided that the fuel is burned on grates or in troughs, instead of in contact with the brick, as was the practice with the original Hoffmann kiln. 243. Sorting and Classification. Uses of Various Grades. In empty- mg the kiln, the bricks are commonly separated into various grades or qualities according to the degree of burning and freedom from imperfections. All bricks which show cracks or excessive kiln- marking as well as those which are badly warped must be put into an inferior class. Classification and Uses of Building Brick. All of the classes of brick whose manufacture is discussed above are classed under the head of 234 MATERIALS OF CONSTRUCTION common brick or pressed brick. The three grades of common brick have been mentioned, viz.: Arch or hard brick, those which, owing to their position in the kiln have been over-burnt, are apt to be misshapen, and are used only in foot- ings and for the " fiUing " of brick masonry. Red or well-burned brick, which amount to about half the output of the up-draught kiln and constitute the best grade of brick for all general construction purposes, and Salmon or soft brick, those which have not been sufficiently burned, are too weak for use as a first-class construction material, but are used for masonry filling and unimportant work not calling for high strength or great durability. Pressed brick or face brick are made by the dry-clay process as above noted or, using the commonly accepted meaning of the term, by repress- ing soft or stiff-mud process bricks. They are smooth and hard, have true surfaces and sharp angles, and can therefore be laid with a minimum thickness of mortar joint. The advantages they possess in appearance favor their use in the facing of masonry, and the fact that they are much more expensive than common brick practically limits their use to that purpose. Glazed and enameled brick. Glazed brick are those made by coating one, side of unburned common brick with a thin layer of what is called " sHp," a composition of ball clay, kaohn, flint, and feldspar, and then applying a second coat of transparent glaze resembling glass. The shp gives the color to the brick, and the glaze melts upon firing the brick, and forms a smooth transparent coating over the white slip. Enameled brick are made from a clay of pecuHar character, generally containing fire-clay, and the enamel is applied either to the unburnt brick or to the finished brick. In burning, the enamel fuses and unites with the body of the brick. It does not become transparent, but gives its own color to the brick. Enameled brick are expensive and more difficult to make than glazed brick, but the former are generally considered the more durable. Both glazed and enameled brick are usually made pure white in color, they do not acquire odor, and are impervious to moisture. They are, therefore, particularly adapted for use in interior finishing, lavatories, hospitals, etc. Ornamental bnck are simply common brick which have been formed in a special mold which imparts some relief design to the face. Specially selected clays may be used to impart some desired uniform color. They are made in various sizes and shapes. Tapestry bnck are made by the stiff-mud process, and subsequently BRICKS AND OTHER CLAY PRODUCTS 235 have their surfaces roughened by cutting off a slice by a wire. Tapestry brick are much used as a face brick. Roman tile are pressed or repressed brick of unusual form, their face dimensions being 12 by I5 inches, and the depth 4 inches. Hollow brick are common brick made by the stiff-mud process, a special die which forms hollow spaces through the length of the brick being used. This produces a Ught brick which is much used for interior wall facing and for partitions. When their surface is grooved to afford a hold for plaster, hollow brick are called furring brick. Sand-lime brick are not made from clays, but since their uses are identical with those of ordinary building brick, their manufacture will be separately discussed at this point. MANUFACTURE OF SAND-LIME BRICK 244. General. There are several classes of brick which are manu- factured by combining sand and lime together in proper proportions. Only one of these is important as a commercial article, however, and entitled to the name " sand-lime brick." Sand brick, or the " kalksandstein " of the Germans, are made up of a mass of sand bound together by calcium carbonate (calcium silicate also, according to the claims of their advocates), hardened by exposure in the air, or in air particularly charged with carbon dioxide, such as the gases from a lime kiln. Mortar brick which have simply calcium carbonate as the binding material, are really simply blocks of lime mortar hardened in air. They are structurally weak and unfit to be used as a building material. Sand-lime brick are made from the same raw materials, but with much more care by a materially different process. The binding material is here not definitely established. The bricks are molded under pres- sure, and they are hardened in an atmosphere of steam while under pressure. 246. The Sand. Almost any sand may be used if the process used be properly varied to suit the properties of the sand. A comparatively pure and clean sand is, however, essential to cheap manufacture. One well-graded in size and not too coarse is preferable. Fineness and Granulometric Composition of Sand. Experiments made by Professor M. Glasenapp in the laboratory of the Polytechnic Institute at Riga * show clearly that the binding action of lime and sand proceeds much more rapidly when the sand is very fine. From the stand- point of denseness and strength, however, a sand of such composition * Thonindustrie Zeitung, Oct., 1900. 236 MATEEIALS OF CONSTRUCTION as to secure a minimum of voids is preferable. A composition of at least four parts of sand which is between the 20-mesh and the 100-mesh screen size to one part of sand finer than the 150-mesh has been found to give a brick of maximum compressive strength. From the standpoint of weathering qualities it is apparent that a well- graded mixture will make a brick of lowest absorption and hence least apt to disintegrate due to weathering. Impurities in Sand. The principal impurities in sand are clay, iron oxide, mica, and feldspar. Clay was found by Mr. Peppel * to have a marked influence in decreasing both the compressive and tensile strength of sand-lime bricks. The addition of 20 per cent clay decreased the strength one-third. Iron oxide is probably inert and the same is true of mica. More than very small percentages of mica are injurious, how- ever, in a purely mechanical way, just as they are in cement mortars. Feldspar was found by Mr. Peppel to be detrimental to crushing strength, but beneficial to tensile strength when 10 per cent was present. 246. The Lime. Quality and Quantity Required. Either high- calcium lime or dolomitic lime may be used, but the former, whenever obtainable, is preferable. Mr. Peppel found the high-calcium lime gave a materially stronger brick both in compression and in tension. The amount of lime used in practice varies from about 5 to 10 per cent. Mr. Peppel found that while there is an increase in strength with increase in lime, the increase is less marked as the hme percentage is increased, and he concluded that the strength gained by addition of lime beyond 10 per cent would not justify the additional cost. All of the earlier writings on the subject of sand-lime brick assert that the hardening process employing steam under heavy pressure results in the formation of calcium hydrosilicate by combination of the lime with the silica of the sand. It has never been proven that any chemical combination does take place between lime and silica at such low temperatures as are encountered in the hardening of sand-lime bricks, and all investigations of the forma- tion of calcium silicates tend to prove that such compounds certainly cannot be formed at temperatures below 900° C. The calcium hydro- silicate theory has therefore been generally abandoned, though a more reasonable hypothesis has not been advanced to account for the observed fact that the steam pressure treatment does produce a material whose hardness and mechanical strength cannot be accounted for if the binding agent is ordinary calcium carbonate alone. 247. Preparation of the Sand. The preliminary treatment of the sand is dependent first upon the source of the sand, and second, upon the * Geol. Sur. of Ohio, 1906. BRICKS AND OTHEB CLAY PRODUCTS 237 subsequent details of manufacture. If a soft sand-stone rock is used it must first be crushed and then screened to separate out the larger par- ticles. If the sand be obtained by dredging it must be dried, the amount of drying being dependent upon the subsequent manner of operation. If an excess of clay is present, or if it be a seashore sand contaminated with the salts of sea-water, washing and subsequent drying are required. If the sand does not contain a sufiicient proportion of very fine quartz sand it is necessary to pulverize a portion of it in a tube mill or other type of fine-grinding machine and add the pulverized sand to the natural sand. In some foreign works it is considered advantageous to roast the sand before mixing, the idea being that chemical action is rendered more active and complete thereby. The efficacy of this latter expedient is open to some doubt. 248. Preparation of the Lime. The preparation of the lime is simply a matter of hydrating or slaking it either before or after the addition of the sand. In the majority of cases the lime is slaked before mixing with the sand, the one exception being the use of a very wet sand. In this event the unslaked lime may be mixed with a portion of the wet sand and thus slaked, and the balance of the sand after being dried is added to the mix. With a moist sand the lime may be slaked to a putty before being mixed with the sand. The best method, however, consists in the use of dry sand and dry hydrated lime, the latter prepared by one of the methods described under the head of hydrated lime. 249. Mixing. The thoroughness of the mixing process is the most essential detail of the entire process. Probably the best method of thoroughly incorporating the 5 to 10 per cent of lime used with the sand and water consists in mixing the lime and sand in a tube mill and then adding water to the mix in a pug-mill. The latter delivers the mix to a bin where it is allowed to stand for some hours before being delivered to the press. 250. Pressing the Brick. Owing to the gritty character of the mix it is not possible to make wire-cut sand-lime brick. They are therefore made in a mold under pressure in exactly the same manner that dry-clay brick are made. Experiments made by Mr. Peppel indicate that brick of maximum strength are obtained when the pressure to which the brick are subjected in the mold is about 15,000 pounds per square inch. 251. Hardening. The brick are not allowed to harden in air, but are hardened in closed chambers subjected to steam under a pressure of from 100 to 150 pounds per square inch. The hardening kettle used. Fig. 133, is a horizontal cylinder of steel, provided with a removable steam-tight cylinder head and tracks upon which cars carrying the brick are run into the cylinder. Experiments made by Mr. Peppel led to the 2^8 MATERIALS OF CONSTRUCTION conclusion that first quality sand-lime brick may be hardened in four hours if 150 pounds steam pressure is used, in six to eight hours at 120 pounds and in eight to twelve hours at 100 pounds. The usual practice is to employ steam under about 120 pounds pressure, main- tained for from eight to ten hours. 252. Uses of Sand-lime Brick. Sand-lime brick, when properly made, have been found an eminently suitable material for general construction purposes in any situation where pressed brick may be used. Their use has as yet been confined largely to the vicinities of the plants where they are made, but the industry shows a healthy growth with an accompanying in- crease in structural use. In 1913 there were 68 plants making sand- lime brick in the United States, having a total yearly output of about 190,000 M, valued at about $1,238,000. Fig. 133. — Hardening Cylinder for Sand-lime Brick. MANXIFACTDRE OF PAVING BRICK 253. General. The requisites for a good paving brick are that it shall be hard enough to resist the abrasive action caused by street traffic; tough, so that it will not be broken by the impact of wheels, horses' hoofs, etc.; and non-absorptive, so that it will resist weathering well. Its manufacture differs from that of common brick, first, in that the selection of a suitable clay is more limited, and second, in that it must be burned at a much higher temperature, vitrification or at least incipient vitri- fication being required. 254. The Clay. Three classes of clays have been used in the manu- facture of paving brick, viz., surface clays, impure fire clays, and rock clays or shales. Surface clays are generally unfit for use as paving-brick clays bc-ause, on account of their highly silicious character, the range of temperatures between incipient and viscous vitrification is so short that only a small proportion of the kiln charge is properly burned. Impure fire clays are used to a slight extent, but are usually very BRICKS AND OTHER CLAY PRODUCTS 239 hard to burn properly. A pure fire clay cannot be vitrified; it is neces- sary, therefore, that there be present at least from 5 to 7 per cent of flux- ing impurities (iron, lime, magnesia and alkalies), and the higher this percentage is the more cheaply it can be burned. These clays possess one advantage in that it is usually impossible to over-bum them, and the burning need not be watched so carefully, as in the case of the sur- face clays and the shales. Shales or rock clays are used for practically all paving brick made at the present time. They occur in larger bodies than either of the other classes of clays and, although the expense of crushing is increased owing' to their rock nature, they are so impure that the range of vitrification is often as much as 400°, making them an especially valuable material for paving-brick manufacture. Preparation of the Clay. The shale banks are usually worked by steam shovel with or without blasting of the rock. Fireclays are mined in underground chambers like coal, and surface clays are usually handled by hand tools, scrapers, carts, etc. / Upon delivery at the plant, shales are crushed either in dry pans (edge runners), rolls, or centrifugal disintegrators. Fireclays and sur- 'face clays are handled in the same types of disintegrators but with less expense. The crushed clay is screened to remove all particles not passed by about a 20-mesh sieve, and delivered to the pug-mill where just suf- ficient water is added to make a stiff mud. I 255. Molding and Drying. Practically all paving brick are now made by the stiff-mud process above described, the machine being usually of the auger type. The size of the die is larger than in the case of build- ing brick, since the usual size of paving brick is about 3^ inches wide by 8^ inches long and 4 inches deep. Both side-cut and end-cut brick are made, the choice depending largely upon local conditions at each plant. Many paving bricks are repressed immediately after molding, the repressing being useful in making the brick more uniform in size and shape. The repressing dies are so formed that the edges of the brick are rounded off, lugs are made on the sides to separate the brick slightly when laid in the street, and the imprint of the manufacturer is placed on the brick. Paving bricks which are simply wire cut without being repressed are usually what are called " wire-cut lug brick." They differ from ordi- nary wire-cut brick only in that the wire-cutting mechanism is so con- trived that instead of the wii;es cutting plane surfaces, they cut planes which are interrupted by two high points which serve as lugs. Such bricks must be side cut, not end cut. 240 MATERIALS OF CONSTRUCTION " Hillside " brick are ones which have one of the upper edges beveled off to make a pavement less smooth. Drying of paving brick is accomplished by exactly the same methods as those used for ordinary stiff-mud brick. The use of a drying tunnel 'most commonly resorted to. 256. Burning, Annealing and Sorting. The burning of the brick is accomplished either in the down-draught kiln or the continuous kiln above described. The burning requires from seven to ten days as a rule, and the temperature is a bright cherry heat (1500 to 2000° F. for shales, and 2000 to 2800° F. for fireclays), whereas only a red heat is attained in burning hard building brick. The temperature required, varies according to the clay used, and the proper temperature of vitri- fication for a given clay must not be exceeded, since it results in the brick becoming softened. At the best the brick in the lower portion of the kiln are usually kiln-marked owing to the weight of the charge above. When the brick are thoroughly burned the kiln must be tightly closed and allowed to cool down slowly for several days. Thus the brick are annealed and acquire a great deal more toughness than when quickly cooled. The brick must be sorted in emptying the kiln. The upper courses will be very hard burned but possibly air-checked. They are excellent brick for foundations or sewers and are sometimes called sewer brick. From the zone of checked brick to within a few courses of the bottom, the brick should be No. 1 pavers; the lower courses have not been suf- ficiently heated to be vitrified and are classed as No. 2 pavers. These latter brick are much used as building brick, for which purpose they form a most excellent material. With fireclay as high as 80 to 90 per cent of the kiln charge are No. 1 pavers, and with shale about 60 to 80 per cent of No. 1 pavers are obtained. MANUFACTURE OF FIRE BRICKS 257. General. The manufacture of firebrick which are able to withstand the moderate heat encountered in ordinary situations such as chimneys, flues, etc., has been carried on for a great many years. It is only comparatively recently, however, that a demand has arisen for firebrick and fire blocks to be used (especially in the metallurgical in- dustries) where they must not only withstand extremely high tempera- tures, but must also have certain chemicaj proclivities which make them withstand an oxidizing or a reducing action, either of which may be encountered in one type or another of metallurgical furnace. BRICKS AND OTHER CLAY PRODUCTS 241 The principal materials from which firebrick are made are fireclay, silica rock, ganister rock (natural or artificial), magnesia, bauxite, and chromite. The process of manufacture varies according to the class of brick and the nature of the raw materials used; the various processes will therefore be separately considered. 258. Acid Bricks. Fireclay Brick. One of the most refractory materials for use in the manufacture of firebrick is ordinary fireclay, to which flint clay, burnt fireclay, sand, or other refractory material has been added to prevent undue shrinkage in drying and burning. Fire- clays owe their superiority over brick clays as a refractory material to their greater purity, i.e., the absence of the fluxing impurities iron, lime, magnesia, and the alkalies. In order to afford a contrast between the composition of fireclays and typical brick clays the following table has been selected from Ries' " Ecojiomic Geology." COMPOSITION OF FIRECLAYS AND BRICK CLAYS SiOz. AljOj-lFezO,, FeO. CaO. MgO Alk. HjO. CO2. SO.. Plastic fireclay. . . Flint fireclay Brick shale Calcar. brick clay Blue shale clay. . . 57.62 59.92 24.00 27.66 1.90 1.03 1.20 0.70 Tr, 0.30 Tr. 0.70 0.64 13.20 10.82 0.35 54.64 38.07 47.92 14.62 9.46 14.40 5.69 2.70 3.60 5.16 15.84 12.30 2.90 8.50 1.08 5.89 2.76 2.70 4.59 2.49 4.85 4.80 20.46 9.50 1.44 Manufacture. Fireclays are often mined in underground workings, and the plastic clay which is mixed with the fireclay is handled in the manner described for common brick clays. Each material is separately ground, usually in a dry pan, after which they are screened and mixed in the required proportions. The relative amounts of the two ingredients vary greatly according to the clays used, but usually approximate about equal proportions. The use of burnt fireclay or sand is more common in England and Germany than in this country. The mixing of the materials is accompUshed either in an open mixer, or in the pug-mill, where the required amount of water is added. The plastic mass delivered from the pug-mill is usually allowed to stand for some time in order to allow the moisture to become more perfectly dis- tributed before being sent to the molder or brick machine. A large pro- portion of the brick are molded by hand by a process that is practically that of slop molding above described. Either the stiff-mud process or the dry-clay process may be used, however. The drying of the brick is effected in the manner that would be used for ordinary clay brick 242 MATERIALS OF CONSTRUCTION made by a similar process, and the firing is done in down-draught or con- tinuous kilns. The temperature of burning must be higher than is used with ordinary clay brick in order that there may be no further shrink- age of the brick when subjected to high temperatures in use as a furnace fining, or in any similar situation. A temperature of from 2500 to 3500° F. is required for proper burning, the lower figure applying to the lower grade of ordinary firebrick and the upper figure for the highest grade. The cooling should be rapid till the temperature is below 2500° F., after which point slow cooling is required. Silica Brick. Silica brick are made of sihca sand or siUca sandstone rock, mixed with a very small percentage of lime, which acts as a binding material. Silica brick made of the purest grade of materials will with- stand a temperature of about 3900° F., while the usual commercial article will fail at temperatures above about 3700° F. They are therefore slightly less refractory than the highest grade fireclay bricks, but are superior to the ordinary run of fireclay bricks. Sihca bricks can be made only of very 'pure materials (98 per cent or more silica), since small percentages of the fluxing impurities materially increase fusibility. The brick are extremely hard and brittle and upon being heated expand, instead of shrinking, as clay brick do. This expansion, amounting to about 10 per cent, must be allowed for in laying the brick. M anufactwe. The silica rock is crushed between rolls or in a gyratory crusher ana then pulverized in an edge-runner mill. In this latter mill lime is added, 1| to 2 per cent being used, and water is mixed with the mass. Usually the lime and water are added together in the form of lime-milk. The final composition of the brick will be 96 to 97 per cent silica, IJ to 2 per cent lime, and IJ to 2 per cent impurities. Usually silica brick are molded by hand, dried in dry-rooms or dry- ing tunnels, and fired in down-draught kilns. No particular precautions are necessary in firing, the temperature required being from 2600 to 3200° F. The cooling down of the brick must be accomplished with the greatest care, however, since sihca bricks are very sensitive to sudden changes in temperature. A very high-grade sihca brick is now produced by the process used in the manufacture of sand-lime brick. The materials are mixed as a stiff mud, molded in a press, and hardened under steam pressure for about eight or ten hours. Ganister Brick. Ganister brick are intermediate in grade between fireclay brick and silica brick. They are made from ganister rock, which is a dense silicious sandstone containing about 10 per cent clay. Ganister brick are made in a manner similar to silica brick except that lime is seldom added, since the clay forms a sufficient binder. Fir- BRICKS AND OTHER CLAY PRODUCTS 243 ing is done at about the same temperature as that required for silica brick and the cooHng must be slowly and carefully accompUshed. 259. Basic Bricks. Magnesia Brick, Bauxite Brick. All of the firebricks above described are of an " acid " character, that is, being composed largely of free silica, they are capable of acting as an acid material at high temperatures, and of fluxing in contact with a basic material such as lime. By decreasing the silica content and increasing the alumina content the material may be rendered " basic " in character. Materials which are used for making basic bricks are principally "bauxite" and "magnesia." Bauxite brick are made by grinding, bauxite (85 per cent or more AI2O3) in an edge-runner, mixing with about 25 per cent clay in a pug-mill where water is added, and molding by hand or by the stiff-mud machine. Burning is done at a temperature of about 2800° F. The composition of bauxite brick is about 70 per cent alumina, 16 to 18 per cent silica, and 12 to 14 per cent ferric oxide. Shrinkage upon heating is excessive, and the bricks are structurally very weak. Magnesia brick are made from a mixture of caustic magnesia (obtained by burning magnesite in a lime kiln) and sintered magnesia which con- tains a small percentage of ferric oxide to act as a flux. Dolomite (double carbonate of lime and magnesia) is sometimes used to make magnesia brick, but is inferior to a mixture of sintered and lightly burnt magnesia. The materials are ground in an edge-runner mill, water is added in a mixer or pug-mill, and the bricks are molded under a heavy press. Dry- ing must be carefully carried out in drying houses or drying tunnels and firing is accomplished in down-draught kilns or in gas-fired kilns at a temperature of from 3300 to 3450° F. The shrinkage of magnesia brick is irregular and excessive. Magnesia and bauxite bricks have a remarkable power of resisting the corrosive action of slag and limestone at high temperatures, but are extremely sensitive to the action of silica and acids generally, and both are very weak structurally. They are used in the lining of basic steel furnaces, in lead smelters, in cement kilns, etc. 260. Neutral Firebricks. Chromite (chrome iron ore) is principally used for the manufacture of a refractory brick, which is practically " neutral." Chrome brick are used in metallurgical furnaces to separate basic and acid brick linings, when the former is used for the lining of the melting chamber in contact with a basic charge and the latter used for the top arches, where greater mechanical strength is needed. Chromite used alone is almost without binding power, and is therefore usually mixed with fireclay or bauxite in such a proportion that the bricks contain about 50 per cent chromium oxide, 30 per cent ferrous oxide, and 20 per cent alumina and silica. 244 MATERIALS OF CONSTRUCTION Chrome brick are made by crushing the materials in edge-runner mills and molding under heavy pressure, as in the case of siHca bricks. The firing temperature is about 3000° F. TERRA COTTA 261. General. Terra cotta is composed of practically the same material as bricks, but requires a carefully selected, finely divided homo- geneous clay which burns to a desirable color with a slight natural glaze. 262. Architectural Terra Cotta. Terra Cotta Lumber, Building Blocks, and Fireproofing. Usually no single clay is used in the production of terra cotta, but each shade and tint requires the minghng of different clays. Fire clays are universally used for the making of architectural terra cotta. The clays after delivery at the factory are separately ground in wash- mills or edge-runner mills, mixed with grit and water in pug-mills and separately deposited in layers or strata. As many as ten or twelve strata are thus piled up, and from this mass perpendicular cuts are taken and the whole mixed together in a pug-mill into a plastic mass. The subsequent manner of molding depends largely upon the character of the ware being made. Architectural terra cotta, used for architectural purposes, is usually made by hand mold- ing in plaster casts which have been made by skilled hand work. Intricate designs are modeled without a mold and green casts are often further carved by hand before being dried and burnt. After drying and be- fore burning a coat- ing of " slip " is applied to the ware. This slip is made up of clay, feldspar, flint, etc., is opaque, and imparts the color desired for the finished product. Either a dull or a bright glazed finish may be procured. Burning is done with extreme care to prevent either distortion or discoloration by flames or gases. Special kilns in which the ware doea £% JS ^MANes9uc TViczc Sft]3.' (^CtlSlf WflLTCR- TASLC -RiaHT CollW<* S«»*i"on GenticD , ^ i-^';^ -m Smsmi fimrinffiff jyooL rjOOOD 'Wr^HT -SffUl^KR FiQ. 134.— Architectural Terra Cotta. BRICKS AND OTHER CLAY PRODUCTS 245 not come in contact with the gases are commonly used. Fig. 134 shows a number of typical designs in decorative terra cotta. Very elaborate designs of high artistic merit are sometimes executed. It is not practic- able to bum terra cotta in very large units, so it is often necessary to make a complete design of many comparatively small sections. Terra cotta blocks or terra cotta lumber is an entirely different ma- terial, made for strictly structural purposes and not at all for decora- tive effect, being always covered by plaster or mortar. The raw materials used for terra cotta lumber are terra cotta clays and finely cut straw or sawdust. The materials are thorough- ly mixed (after grind- ing) in a mixer or pug- mill to form a stiff mud, which is thereupon forced through a die by a plunger-type ma- chine. Practically all terra cotta lumber is made of hollow con- struction, with walls about 1 inch thick, and with partition walls about f inch thick. The blocks are wire cut, as in the case of or- dinary stiff-mud bricks, and are carefully dried before firing. Firing is usually done in a down- draught kiln and especial care is required in stacking up the ware in such a manner as to prevent distortion and unequal heating. The tempera- ture of burning is suflScient to burn out all the straw or sawdust incorpor- FiG. 135.— Terra Cotta Lumber. 246 MATERIALS OF CONSTRUCTION ated in the mix, and leaves a very light and porous material which is soft enough to be cut with a saw and into which nails or screws may be driven with ease. Fig. 135 illustrates various common forms of terra cotta lumber. Hollow building blocks and fireproofing are absolutely the same thing as terra cotta lumber except that no straw or sawdust has been incor- porated with the clay and the burning is carried to a higher temperature, almost to vitrification in fact. They are therefore much harder than terra cotta lumber and are not porous, but resemble ordinary hard- burned brick. Hollow building blocks are made in exactly the same shapes as terra cotta lumber and 'have the same uses. Both are used extensively as a fireproof material for constructing exterior and partition walls, ceilings, floor arches, furring for outside walls, roof sheathings, and jackets around beams and columns of steel-framed buildings. Porous terra cotta is more conveniently handled, due to its softness and easy-cutting qualities. It is also claimed to be tougher than hard- burned blocks. The latter are stronger, however, under static load. ROOFING, WALL, AND FLOOR TILES 263. Roofing Tiles. Roofing tile must be hard, strong, and of low absorption. They are made by a process similar to that employed for Spanish D D G Mlssloa Curved Types (Intedocking Tilo) SauareEnd Fig. 136.— Roofing TUe. Washington Cut Hat Types (Shinglo Tile) Round End repressed stiff-mud brick, the blanks being formed by an augur machine shown by Fig 36. Shmgle tile are laid like slate, being perfectly flat; mission style tiles are, in a measure, interlocking, being segments S hollow cones; and Spanish tile are distinctly interlockfng. The Tottte BRICKS AND OTHER CLAY PRODUCTS 247 forms are made by repressing slabs of green clay. The clay is selected with greater care than in the case of ordinary brick, and burning is done at a sufficiently high temperature to insure hardness, strength and nbn- absorptiveness. 264. Wall Tile. Wall tile are of two general types according to the process employed in making them — "dust-pressed" tile and "plastic" tile. The clay used in making dust-pressed tile may be a fireclay, or even a red-burning shale clay, if artificial coloring is to be used, but if a white body is required, as is usual, a mixture of white clays, feldspar, flint, etc., is employed. These materials after grinding and mixing are made up into a thin cream which is strained through a silk screen. The water is now drained or squeezed out and the material is ready to be molded if the plastic process be used. For dust-pressed tile this material is dried, crushed to powder, moistened slightly by steam, and stored until used. Molding is done in a dry press and any desired relief pattern is imprinted by the moving plunger of the press. The ware is now placed in fireclay boxes in which it is burned out of contact with the flames. After burning, a transparent or opaque glaze is applied, coloring matter being introduced if desired. A second firiiig is now required to fuse the glaze. Either a bright or a dull glaze is obtained by proper selection and application of the glaze materials. Plastic tile are made in practically the same manner as " dust-pressed " tile, except that a mixture of soft clay and powdered burnt clay is used, and the molding is done immediately after mixing and tempering. Mold- ing is done by hand in plaster molds which may or may not impart a design to the face. In the latter case the tile may be removed from the mold when partly dry and a design modeled by hand. Sometimes plastic tile are not glazed, and in that event no second firing is required. Most wall tile are dust-pressed tile. 265. Floor Tile. Floor tile are dry-pressed from the same materials as wall tile and the process employed is practically the same. Fireclays and red-burning clays are most commonly used, however, and if white- burning materials are used coloring matter (metalhc oxides) is usually introduced. No glaze is used and only one firing is necessary. DRAIN TILE AND SEWER PIPE 266. Drain Tile. Drain tile are made from a red-burning clay or mixture of clays of the character of those used in making terra cotta lumber. The clay is handled by the stiff-mud process, issuing from a special die as a hollow cyUnder, which is cut to convenient lengths by wires. 248 MATERIALS OF CONSTRUCTION Burning is conducted at temperatures compatible with the pro- duction of a very porous product, which possesses a considerable degree of mechanical strength, but is not vitrified or glazed. 267. Sewer Pipe. Sewer pipe is not intended to be porous or absorp- tive like drain tile, but on the contrary is made from clays which will form a very non-porous product of low absorption. The sti£f-mud process of mixing and molding is used whether the pipes are fitted with socket ends or not. Special shapes, such as elbows, Ys, and Ts, are made by joining parts of green pipe with slip clay. After drying in steam chambers the ware is burned in down- draught kilns. A special glaze called " salt glaze " is imparted to all surfaces of the pipe by throwing common salt into the kiln fires after a temperature of about 1150° C. has been attained. The sodium vapors freed by heat pass through the kiln, and by combination with with the clay form a dense hard glaze, which renders the pipe practically non-absorptive. Sewer pipe are intended exclusively for use as conductors of water, sewage, etc., and are laid with tight cement joints. They therefore do not take up and carry off the water in wet soil as drain tiles do. PROPERTIES OF BRICKS OF ALL CLASSES 268. Crushing Strength. Since, as will be later shown, the strength of brick masonry is only a fraction of the strength of the brick, the com- pressive strength of individual bricks is of only relative value in that it aifords a basis of comparison between different kinds of brick. The following table is inserted to show the approximate strength of the several grades of brick. The table has been selected partly from the Watertown Arsenal tests of 1895, 1897, and 1906 and partly from tests made by the author. The color is noted in each case simply by way of general description. It is not an indication of the degree of burning except when brick from one class of materials only is being considered. From inspection of the table we may conclude by way of a summary that the compressive strength of average good building brick is about 4000 pounds per square inch ; pressed brick about 8000 pounds per square inch; sand-lime brick 3000 to 4000 pounds per square inch; paving brick 10,000 pounds per square inch; fireclay brick 3000 to 6000 pounds per square inch; terra cotta blocks, 4000 pounds per square inch; and archi- tectural terra cotta 3000 pounds per square inch; BRICKS AND OTHER CLAY PRODUCTS 249 COMPRESSIVE STRENGTH OF BRICK (Pounds per Sq. In.) Common Building Brick Light-burned buff colored 2,987 1st brand Hard-burned buff colored 3,570 1st brand Hard-burned buff colored 4,561 2d brand WeU-burned cream colored 4,066 1st brand Well-burned cream colored 4,845 2d brand Well-burned cream colored 7,244 3d brand WeU-burned red colored 6,834 1st brand Well-burned red colored 5,307 2d brand Hard-burned red colored 9,520 1st brand Pressed or Face Brick Red face brick 12,898 1st brand Red face brick ■ 5,278 2d brand Red face brick 8,327 3d brand Red face brick 9,951 4th brand Red face brick 7,067 5th brand Buff-colored face brick 9,823 1st brand Buff-colored face brick 9,264 2d brand Buff-colored face brick 6,975 3d brand Chocolate-colored face brick 6,582 1st brand Sand-limb Brick Red sand-Hme brick 5,285 1st brand Cream or brown sand-lime brick 3,517 1st brand Light gray sand-lime brick 4,774 1st brand Light gray sand-lime brick 5,098 2d brand Light gray sand-lime brick 4,010 3d brand Light gray sand-lime brick 2,535 4th brand Light gray sand-lime brick 6,810 5th brand Paving Brick Red vitrified paving block 11,340 1st brand Red vitrified paving block 13,316 2d brand Red vitrified paving block 10,347 3d brand Chocolate vitrified paving block 10,871 1st brand Chocolate vitrified paving block 10,593 2d brand •Chocolate vitrified paving block 11,180 3d brand Buff vitrified paving block 9,610 1st brand Buff vitrified paving block 9,580 2d brand Fire Brick Buff fireclay brick 2,668 1st brand Buff fireclay brick 6,299 2d brand Buff fireclay brick 5,715 3d brand Buff fireclay brick 4,525 4th brand Buff fireclay brick 3,253 5th brand Buff fireclay brick 6,278 6th brand Buff fireclay brick. . .'; 2,646 7th brand Buff fireclay brick 8,589 8th brand Terra Cotta Lumber Partition tile 1 core (net section) 4,892 Partition tile 2 cores (net section) 4,663 Floor Arch Blocks 2 cores (net section) 3,591 Floor Arch Blocks 4 cores (net section) 4,965 1st brand Floor arch blocks 4 cores (net section) 3,730 2d brand Architectural Terra Cotta White ornamental blocks 2,985 Buff ornamental blocks 3,520 Salmon ornamental blocks ^' on Red ornamental blocks 2,830 250 MATERIALS OF CONSTRUCTION 269. Absorbing Power. The absorption of water by brick is often taken to be an important criterion of its probable durability. The freez- ing of water which fills the pores of brick will of course constitute a great disintegrating agency, but the importance of this factor is probably over- estimated. In the first place, bricks which are structurally so weak that they are apt to suffer through frost action are not fit material for construction purposes, and in the second place, bricks, even though they are very porous, are seldom injured by frost for the reason that water does not fill the pores completely and hence is able to expand upon freezing without exerting a great disruptive force. Tests of absorptive power of bricks are usually made to extend over so short a period that the results are apt to be misleading. The rate of absorption of bricks varies greatly, and therefore the usual forty-eight- hour test will sometimes show practically the ultimate absorptive power of a brick, while in other cases only a fraction of the ultimate absorption is shown. In general the absorption shown by common building brick in a forty- eight-hour immersion test is from 12 to 18 per cent of the weight of the dry brick. For pressed or face brick the absorption is about 6 to 12 per cent; for sandJime brick it is 12 to 15 per cent; for paving brick, 1 to 3 per cent; for fireclay brick, 8 to 12 per cent; and for unglazed building blocks of terra cotta, 10 to 15 per cent. 270. Transverse Strength. The transverse strength of bricks is one of the best indications of quality, not because bricks are often subjected to severe tests of their transverse strength in masonry, but because tests are easily made and an indication of their toughness is afforded. An indication of their approximate tensile strength is also obtained since the modulus of rupture bears a fairly close relationship to tensile strength. Experiments and experience show that the failure of brick masonry under compressive stress is really by failure and subsequent lateral flow of the mortar, thereby introducing tensile stresses in the bricks and causing cracks to open up in the masonry in the direction of the pressure. The fractured surface of a brick also affords a valuable indication of the care with which the materials have been ground and mixed, and the degree of burning is made evident to an experieiiced observer who is familiar with the normal appearance of the bricks under examination. The modulus of rupture of common building bricks varies from about 500 pounds per square inch to about 1000 pounds per square inch; pressed bricks, 600 to 1200 pounds per square inch; sand-lime bricks, 300 to 600 pounds per square inch; paving bricks, 1500 to 2500 pounds per square inch; fireclay bricks, 300 to 600 pounds per square inch; and unglazed terra cotta building blocks, 500 to 1000 pounds per square inch. BEIGES AND OTHER CLAY PRODUCTS 251 271. Shearing Strength. The shearing strength of bricks as shown by tests is a property of Uttle practical importance, chiefly because it is almost impossible to make the best results really show actual shearing strength. All methods of testing which have been devised are more or less subject to the same objection, i.e., the shearing stress is not acting alone, but bending is introduced, thus bringing tension and compression into play as well as shear- Tests made at the Watertown Arsenal indicate that the shearing . strength of common bricks is about 1000 to 1500 pounds per square inch; pressed bricks, 800 to 1200 pounds per square inch; sand-lime bricks, 500 to 1000 pounds per square inch; paving bricks, 1200 to 1800 pounds per square inch; and fireclay bricks, 500 to 1000 pounds per square inch. 272. Modulus of Elasticity. The modulus of elasticity of bricks is not a constant for any considerable range of loading. The elastic prop- erties as shown by the stress-strain curve for a compressive test are quite similar to those of concrete and mortars. For ranges of loading not exceeding one-fourth of the compressive strength the modulus of elas- ticity of common bricks is about 1,500,000 to 2,500,000 pounds per square inch; pressed brick, 2,000,000 to 3,000,000 pounds per square inch; sand-lime bricks, 800,000 to 1,200,000 pounds per square inch; and pav- ing bricks, 4,000,000 to 8,000,000 pounds per square inch, BRICK MASONRY 273. General. Brick masonry was at one time considered_as_s.imply an inferior substitute for stone masonry, an estimation which may have been justified by facts at the time when almost all bricks were hand made from materials indifferently selected. At the present time, how- ever, brick masonry built with bricks made by modern processes from carefully chosen raw materials, and with the aid of elaborate plant equip- ment, will for many purposes and situations equal or excel most stone masonry. Brick masonry is usually cheaper than stone, it is more easily built with less skilled labor, it is a better fire resistant, and it compares well as regards durability with the best stone masonry. Only one class of stoiie masonry is stronger than the best brick masoniy — cut stone ma- sonry — and this is so much more expensive than brick that its use is largely barred expept where great strength is the first requisite. 274. The Mortar and the Joints. Reference should here be made to the discussion of lime and Hme mortar in Arts. 59 to 64. Lime mortar is used in building a great proportion of the brick masonry constructed 252 MATERIALS OF CONSTEUCTION in this country. As stated above the slower setting magnesian or dolo- mitic limes are generally preferred for bricklaying on account of the fact that the rapidity of laying is increased by their choice in preference to high-calcium Umes. Hydrated lime is now coming into quite general use in preference to lump quicklime. Natural cement and Portland cement are often added to lime mortar to secure greater strength. The former is generally beUeved to confer greater plasticity and better work- ing qualities than the latter. The proportions commonly used in lime mortars are either one part lime to two parts sand, or, more commonly, one part lime to three parts sand. Ordinary lime mortar hardens very slowly and the gain in strength is too slow when high walls are built rapidly. For such construction natural cement or Portland cement should constitute a portion of the mortar or better still, a Portland cement mortar may be used altogether. For all construction below grade, or in any situation where moisture is often encountered, also in heavily stressed brick masonry in piers, arches, etc., cement mortar should always be used instead of lime mortar. All brick should be laid with the minimum thickness of joints con- sistent with proper bedding. Common bricks are usually somewhat rough and uneven, but may be and should be laid with joints from ^ to f inch in thickness. It is commonly specified that the height of eight courses of brick masonry shall not exceed the height of eight bricks laid dry by more than 2 inches. Pressed brick, being usually smooth and true, are laid with joints not exceeding | or ^ inch in thickness. 275. Bond. Bond in brick- work is the arrangement of the bricks in courses resorted to for the purpose of tying together all parts of walls more than one brick in thickness by the action of the weight of the overl5ang masojiry. The commonly adopted bonds for lajang brick masonry are common bond, English bond, and Flemish bond. In common bond, Fig. 137, all of the outside brick are laid as stretchers for from four to six courses, and then a course of headers is placed. This type of bond is more gen- erally used than any other in this country. Fig. 137. — Common Bond of Brickwork. BEICKS AND OTHEE CLAY PRODUCTS 253 In English bond, Fig. 138, heading and stretching courses alternate. This is the strongest type of bond, but has not a pleasing appearance and is seldom used except in England. In Flemish bond. Fig. 139, headers and stretchers alternate in each course, each header being centrally placed with respect to a stretcher in Fig. 138.— English Bond of Brickwork. Fig. 139.— Flemish Bond of Brickwork. the course below. This is a strong bond, but requires cutting brick for each course at corners. None of the above types of bond is appUcable' in walls faced with face or pressed brick, the reason being that the face brick are smaller and laid with thinner joints than the common brick of the filling or backing, so that the level of the beds is not the same for the p~j facing and the backing. It is l-...-™-.^.....,.. ?^ , i- also not usually considered desirable to destroy the sym- metry of the 'face by intro- ducing headers. It is there- fore the practice to tie the facing of the backing by some sort of metal tie, one form of which is shown in Fig. 140. 276. Laying the Brick. Mortar, unless very wet, does not adhere to dry brick nor set properly, for the reason that the water in the mortar is absorbed by the bricks. All bricks should therefore be wet before being laid. This provision is one of great importance, but is a very difficult one to enforce since it makes a material difi'erence in the amount of labor involved. It is difficult to lay pressed bricks dry, so there is not usually Fig. 140.— Metal Face Brick Ties. 254 MATERIALS OF CONSTRUCTION so much difficulty encountered on this score in the case of pressed brick as in the case of common brick. The proper method of laying brickwork consists in spreadmg a layer of mortar along the outer edge of the last course of bricks with the trowel, pressing the brick into place with a sliding motion which forces the moriiar to fill the joint, scraping off the excess mortar squeezed out on the face of the ^all by the trowel, and apply- ing this mortar to the outer vertical . angle of the brick just laid to fill the next joint. After the two out- side courses (face and back) have been laid in this manner the space between should be bedded with soft mortar and the filling brick .placed with a similar shding mo- tion. After several courses of brick have been placed the joints are " struck " on the face side of the wall by means of the point of the trowel held obliquely. Fig. 141 shows the easiest and cheapest way Fig. 141. Pointing Brickworlc. (Incorrect.) Fig. 142. Pointing Brickwork. (Correct.) of striking the joints, but for the outside or weather face of walls the joints should be struck as shown in Fig. 142. Strength of Brick Masonry 277. General. The appearance and durability of brick masonry is usually of greater importance than the actual crushing strength. In piers, however, and occasionally in arches and in the lower portion of the walls of high buildings built without a steel or concrete frame, the stresses encountered may be sufficiently high to make the crushing strength of the brick work an important consideration. The strength of brick masonry is always much more a function of the mortar used, the bond, and the workmanship than a function of the strength of the individual bricks. Where high stresses are encountered cement mortar must invariably be used, and this is particularly true when a heavy load must be carried within a short time after the masonry is laid. 278. Strength of Brick Masonry Shown by Tests. The tests quoted below comprise a summary of tests of twenty-four brick piers about BRICKS AND OTHER CLAY PRODUCTS 255 12 inches square and 8 feet high made at the Watertown Arsenal.* The two varieties of pressed bricks tested gave lower column strength with all mortars than did the better class of bricks molded without pressure. This is probably accounted for by a lower bond or adhesion between the bricks and the mortar. These tests clearly show the importance of using cement mortar in preference to lime mortar where strength is an important consideration. The strength of the 1 : 3 cement mortar masonry is approximately twice the strength of the lime mortar masonry, while there is little gained by the use of neat cement instead of 1 : 3 cement mortar except in the cases of the piers built with the highest quality bricks where the mor- tar of neat cement develops a greater proportion of the strength of the bricks. CRUSHING STRENGTH OF BRICK PIERS Watektown Arsenal Tests, 1904 Age 6 months Compress. Strength. Lbs. per Sq.in. Per cent of the Ave. Crush. Strength of Brick. Description of Brick. Neat Port. 1 Port., 3 Sand. 1 Lime, 3 Sand. Neat Port. 1 Port., 1 Lime, 3 Sand. 3 Sand. Dry pressed face brick I 2880* Re-pressed mud brick | 1925 Face Brick 2400 16^0 1517 1260 26 28 21 25 13 19 Common Brick Wire-cut stiff-mud brick . . . 4021 2410* 1420 31 19 11 Hard sand-struck brick 4700* 1800* 994 42 16 9 Hard sand-struck brick 1969 1800 733 44 40 16 Hard sand-struck brick 1400 1411 718 24 24 12 Light-hard sand-struck brick 1510* 1519 732 23 23 11 Light-hard sand-struck brick 1061 1224 465* 20 23 9 * Tested at age of 1 month. \ 279. Pressures Allowed in Practice. The following table lists the safe working stresses for brick masonry recommended by a committee of Chicago engineers and architects, to be incorporated in the building laws of that city. The table represents very conservative practice: * Tests of Metals, 1904, pp. 419-449. 256 MATERIALS OF CONSTRUCTION SAFE LOADS ON BRICK MASONRY Description of Brick. Kind of Mortar. Safe Load, Lbs. per Sq.in. Paving brick 1 : 3 Portland 350 Pressed and sewer brick (strength 5000 lbs.) 1 : 3 Portland 250 Select hard common brick (strength 2500 lbs.) 1 : 3 Portland 200 Select hard commqn brick (strength 2500 lbs.) 1 Portland, 1 lime, 3 sand 1 : 3 Portland 175 Common brick (strength 1800 lbs.) 175 Common brick (strength 1800 lbs.) 1 : 3 Natural 150 Common brick (strength 1800 lbs.) Common brick (strength 1800 lbs.) 1 Portland, 1 lime, 3 sand 1 : 3 lime 125 100 PART II THE FERROUS METALS CHAPTER XI PIG IRON GENERAL 280. Historical. The ferrous metals comprise only three general classes of material, cast iron, wrought iron and steel. All of these are derivatives of iron, produced artificially by the reduction of iron ores and subsequent treatment by various metallurgical processes of the pig iron which is directly produced from ores. Cast iron, wrought iron, and steel are distinctly different materials, judged by their comparative physical properties, in spite of the fact that metallic iron is present in all to the extent of at least 92 per cent. The production of iron and steel from ores and the fabrication of use- ful implements therefrom comprises one of the oldest mechanic arts known to man. Ores of iron are all essentially oxides of iron, adulterated with large or small amounts of earthy material. Thousands of years ago it was dis- covered that metalUc iron could be produced by heating iron ore in char- coal fires, the oxygen being removed by the charcoal, leaving metallic iron in a pasty condition. The primitive method of burning ore in heaps over a charcoal fire gradually developed into the Catalan forge process, by which wrought iron and,, more rarely, steel was produced in all parts of the world for many centuries. The Catalan forge in its early form was nothing more than a hole in the ground lined with masonry. This was filled with charcoal and ore, either ndixed or deposited in vertical layers, and an air blast, deUvered through a downward inclined tuyere which projected into the mass, supplied the necessary oxygen for combustion of the fuel. The temperature acquired is not sufficient to melt the iron, but will fuse the slag, which may then be tapped off, and the iron is withdrawn 257 258 MATEEIALS OF CONSTRUCTION as a pasty mass. If the tuyere be given a lesser incline and the process be continued for a longer time, the temperature attained is sufficient to promote the combination of a certain amount of carbon of the fuel with iron, thereby producing steel. Cast iron was never produced in the Catalan forge or other early form of device for reducing iron from the ore, because the temperature attained was never sufficient to melt the iron and effect the combination of iron with the 3 or 4 per cent of carbon which is primarily responsible for the structural and physical characteristics which distinguish cast iron from wrought iron and steels. Cast iron, or more properly speaking, pig iron (pig iron is commonly called cast iron after being remelted and cast in molds) was not produced until the blast furnace process of ore reduction was developed in com- paratively modern times. 281. Iron and Steel in Construction. From the" standpoint of the economist the iron and steel industry is by far the most important and valuable one in existence. The world's production of iron and steel represents a greater proportion of our manufacturing wealth than that of any other industry, and probably no other factor has contributed so much to the development of all arts and industries which has character- ized the last fifty years as has the progress made in the manufacture of better and cheaper iron and steel. Cast iron is made so easily by a simple process of remelting pig iron and casting in molds, that it is produced in practically every city of industrial importance in the civilized world. Its strictly structural uses are somewhat hmited, owing to the greater adaptabihty of wrought iron and steel to structural needs. As a material of construction in general, however, and especially as a material for use in construction of machines, implements, etc., it finds countless applications which cannot be as cheaply and satisfactorily served by any other material. Wrought iron has, since the earliest times, been an important mate- rial in many arts and industries. Until the development of cheap methods of steel making in the latter half of the nineteenth century, it was the prin- cipal metallic structural material. Even now it has many uses in fields which steels have not been able to invade, in spite of the fact that the latter may be more cheaply produced. Steel assumed its present place as the most important metal for gen- eral construction purposes with the development of the Bessemer process of steel making following its invention in 1855. Prior to this time steels had been produced only by comparatively slow and expensive methods, which made impossible any real competition with the relatively cheap wrought iron. In recent years open-hearth process steel has been grad- PIG IRON 259 ually replacing Bessemer steel in many fields, but the quantity of both used structurally, and in construction generally, remains enormous. (In 1913 over 31,000,000 long tons of steel were produced in the United States alone.) Pig iron has no construction uses, but is the raw form of iron from which wrought iron and steels as well as cast iron are made. 282. General Classification of Iron and Steel. In a later chapter a technical classification of iron and steel will be introduced. For present purposes, however, iron products will be grouped under the following heads: (1) Pig iron, the product obtained by the reduction of iron ores in the blast furnace. Carbon is present in amounts not usually below 2.5 per cent nor above 4.5 per cent. It is cast as it flows directly from the blast furnace into rough bars called " pigs.". (2) Cast iron, remelted pig iron after being cast or about to be cast in final form. It does not necessarily differ from the pig in composition, and is regarded by the metallurgist as the same thing as pig iron. It is not malleable at any temperature. (3) Malleable cast iron, a form of cast iron which by a special - armealing treatment after casting in final form has been rendered malleable or semi-malleable. (4) Wrought iron, a form of iron which is aggregated from pasty particles without subsequent fusion. Wrought iron contains slag enclosures, and is initially malleable, but normally possesses so little carbon that it will not harden when rapidly cooled. (5) Stejel, iron which has been cast from a molten mass, whose composition is such that it is malleable at least in some one range of temperature, and which may or may not harden upon sudden coohng. Steel which owes its distinctive properties chiefly to car- bon is called carbon steel. Steels whose distinctive properties are due chiefly to the presence of elements other than carbon are called alloy steels. THE RAW MATERIALS OF THE IRON INDUSTRY Ores of Iron 283. General. Ores of iron consist essentially of compounds of iron, usually oxides, mixed with " gangue " (silica, clay, calcic phosphates, etc.), and those of commercial importance contain from 25 to 70 per cent metallic iron (usually 40 to 60 per cent). Iron is extracted from ores by a process known as " smelting," which consists primarily in the heating of the ore to a high temperature under 260 MATERIALS OF CONSTEUOTION strongly reducing conditions in the presence of a basic flux. The reducing agent serves to remove the oxygen from the oxides of iron, leaving metallic iron together with such elements as carbon, sihcon, manganese, phos- phorus, and sulphur, which are invariably present either in the ore or in the fuel used in melting. The flux, usually limestone, combines with the gangue of the ore and the ash of the fuel, producing a fusible slag which may be separated from the metallic iron. The forms of iron ore of greatest commercial importance are hematite, limonite, magnetite, and iron carbonate. 284. Hematite. Hematite, sometimes called " red hematite " or " red iron ore," is anhydrous ferric oxide, FeaOs, containing when pure 70 per cent iron. It occurs usually in granular or massive structure, although it sometimes has a columnar formation or a brilliant scaly structure, when it is called " specular hematite." It is usually very hard and heavy, varies from black to brick red in color and often occurs as an earthy ore which is cheaply handled. An oolitic form of hematite occurs in enormous quantities in France, Lothringen, and Luxemburg. This contains only 30 to 35 per cent iron, but, owing to its lime content, is partly or wholly " self-fluxing." This ore is the basis of the great German, French and Belgian iron industries. Hematite is almost entirely non- magnetic, and when pure is almost identical in appearance and com- position with ordinary deep-red iron rust. 285. Limonite. Limonite, also called " brown iron ore," or " bog iron ore," is hydrated ferric oxide, Fe20s-F[n]H20, containing about 60 per cent iron. It differs in composition from red hematite only in that it contains about 14.5 per cent chemically combined water. It is softer than red hematite, lighter, and occurs usually in massive form. A familiar form of limonite is the newly formed fresh yellow rust on iron. 286. Magnetite. Magnetite is the magnetic oxide of iron, Fe304, con- taining when pure 72.4 per cent iron. It is a hard black mineral occur- ring in granulated or massive structure. It is almost as magnetic as pure iron, very heavy, and often (in U. S.) contaminated with silica, titanium, and phosphorus. It is identical in composition with the black mill scale which forms on iron at temperatures above redness. 287. Iron Carbonate. Iron carbonate, commonly called " siderite " or " spathic iron ore," FeCOs, contains when pure 48.3 per cent iron. It changes to limonite and hematite on weathering. It is rarely used as an ore in its raw state, but is subjected to a preliminary calcination to remove the CO2. 288. Extent of Ore Production in the United States. The great bulk of the hematite ore used in the United States comes from the Lake Superior district (Michigan, Minnesota, and Wisconsin), or from Alabama. PIG mON 261 Of the total amount of hematite produced in 1913 the Lake Superior district produced over 52,000,000 long tons, while Alabama produced about 4,000,000 long tons, and no other district produced over about 500,000 tons. The total production of hematite in 1913 was 58,018,295 long tons. The production of limonite in 1913 was 1,577,019 long tons, of which Alabama produced about 845,000 tons, Virginia about 354,000 tons, Tennessee about 125,000 tons and Georgia about 96,000 tons. No other district produced over 30,000 tons. The production of magnetite in 1913 was 2,357,274 tons, the bulk of which was produced by New York, New Jersey, and Pennsylvania. The total production of carbonate was 27,849 tons, Ohio being the largest producer. It appears therefore that hematite constitutes about 93.6 per cent of the country's ore production, limonite about 2.5 per cent, magnetite about 3.8 per cent, and carbonate only 0.05 per cent. The Lake Superior dis- trict's output of hematite alone amounts to more than 84 per cent of the entire tonnage of iron ore mined in the United States. 289. Ore Mining and Transportation. The manner of mining adopted depends to a great extent upon the physical condition of the ores. Many of the ranges of hematite ore in the Lake Superior district, notably, the famous Mesabi range, which produces two-thirds of the Lake Superior ore, produce ores so soft and finely divided that they are very easily and cheaply worked by steam shovels. Others in the same district, however, are hard and dense, and must therefore be drilled and blasted. These rock ores are more expensively mined than the soft ores, but, as we shall later see, the use of a certain proportion of rock ore with the earthy ore is indispensable to the successful operation of the blast furnace. The limonites are always rather soft; the magnetites, on the other hand, are usually very hard and dense and are necessarily mined or quar- ried by the use of explosives and crushed before smelting. Ores commonly require long transportation in order that the smelt- ing may be done in a district of cheap fuel. Since, as above noted, the Lake Superior district suppUes over 84 per cent of all the ore smelted in the United States, the problem is largely a question of the delivery of Lake Superior ores to the coal regions of Pennsylvania, Ohio, Illinois, and New York. These four States produce, therefore, about 81 per cent of the country's output of pig iron. (Pennsylvania produces about 42 per cent, Ohio 23 per cent, Illinois about 9.5 per cent, and New York over 6.5 per cent.) . No local smelting of the Lake Superior ores has been done in the past, but the recent construction of furnaces in the ore district gives 262 MATERIALS OF CONSTRUCTION promise of a change in this regard in the near future. The ores have always been transported practically exclusively by the ore boats of the Great Lakes. These boats (10,000 to 15,000 tons capacity), are loaded at Lake Superior ports in a very few moments, elaborate loading devices being used. Passing through the locks of the Sault Ste. Marie canal, the boats proceed either to Chicago by way of Lake Michigan, or to the ports of Pennsylvania, Ohio and New York by way of Lakes Huron, St. Clair, and Erie. Discharging is accomplished in a period of from four to six hours by the use of mechanical unloading devices. The ores of Alabama, which constitute the only other important deposits, are mostly smelted in the neighborhood of the mines. 290. Special Preliminary Treatment of Ores. Practically all of the hematite, and therefore the greater proportion of all ores used, is charged into the furnace without any preliminary treatment. Some ores, how- ever, behave more satisfactorily in the furnace after having been sub- jected to one or the other of the following preliminary processes. Calcination is resorted to for the purpose of removing water from limonites or hydrous ores; removing CO2 from carbonates; oxidizing a portion of the gangue of dense ores, particularly magnetites, thereby rendering them more accessible to the furnace gases; or, lastly, rendering the ore magnetic to facihtate subsequent magnetic concentration. The calcination is usually accomplished in vertical furnaces resembhng a mixed feed type hme kiln. The fuel is charged with the ore at the top and the temperature is controlled by regulation of the air supply. Roasting is resorted to solely for the purpose of removing sulphur from ores. The sulphur is present as pyrite, FeS2, which is decomposed at a moderate heat, liberating S and FeS. The "FeS is oxidized by air to form ferrous sulphate, FeS04, and further heating decomposes the sulphate, forming ferric oxide, Fe203, with the liberation of sulphur dioxide and oxygen, thus: FeS2+Heat = FeS-f-S; FeS+202 = FeS04; 2FeS04+Heat = Fe203-l-2S02-|-0. Roasting is accomplished either in large cylindrical kilns (20 feet in diameter by 30 feet high), the fuel being mixed and charged with the ore at the top (Fig. 143) or in a gas-fired kiln which consists of an annular brick-lined chamber surrounding a central cylindrical flue (Fig. 144). Gas is burned in a separate combustion chamber and the products of com- bustion are drawn through the ore into the central flue. PIG lEON 263 Concentration of ores is employed occasionally for the purpose of freeing ',he ore of a part of the gangue, and for enriching the ore before smelting. Wet concentration is sometimes used to remove clay, loam, etc., by a simple process of washing. Another method of wet concentration is by the use of jigs which separate pebbles and sand from the ore by agitation of perforated trays set in tanks of water. Dry concentration is usually accomplished by some type of magnetic separator. If the ore is not already magnetic it is magnetized by pre- jif — k, — .if — lir — -^ — ,£ Fig. 143. Coal-fired Kiln for Ore Roasting. Fig. 144. Gas-fired Kiln for Ore Roasting. liminary calcination. It is then crushed and passed in a thin layer before strong magnets, the magnetic portion being thus attracted away from the non-magnetic. Fig. 145 shows one type of magnetic concentrator. The crushed ore is fed on to belt B and carried between the poles of a power- ful electromagnet where the magnetic portion is attracted to the cross belts B', which are provided with a coating of magnetic material. The cross belts carry the metal to one side out of the magnetic field, when it drops into the hoppers provided below. The tailings remain on belt B and are discharged into a bin when the belt passes over a sheave. 264 MATERIALS OF CONSTRUCTION irnr^Fi rr-i Fig. 145. Type of Magnetic Separator for Ore Concentration. 291. Grades of Ore. Ores of iron are divided into two main classes knownas Bessemer ores and non-Bessemer ores. This division is due to the fact that acid Bes- semer steel must contain less than 0.1 per cent phosphorus, and neither the blast furnace reduc- tion of the ore nor the acid Bessemer steel proc- ess is able to reduce the phosphorus content. All ores in which the phos- phorus content does not exceed one-thousandth part of the iron content are therefore classed as Bessemer ores, and all ores carrying a higher percentage of phosphorus as non-Bessemer ores. The Bessemer ores are worth about 15 per cent more in the market than non-Bessemer ores. The Fltjx 292. Necessity for Use of Fltxx. The office of the flux, as has been stated above, is to render fusible the more or less infusible gangue of the ore and provide a fusible slag in which the non-metallic portion of the ore may be carried off. The exact character and amount of the flux needed will depend upon the composition of the ore and fuel used, and the char- acter of pig iron required. In general it may be said that a basic flux is required for acid gangues (high in silica, alumina, etc.), while an acid flux might be required were the gangue basic (high in lime, magnesia, or alkaline matter). As a rule gangues are acid in character and fluxes are almost invariably basic in character. The flux serves another purpose, as will be later shown, besides taking care of the acid gangue and ash. The sulphur in the charge, whether in the ore or in the fuel, combines with the lime of the flux, forming calcic sulphide, which is removed in the slag. 293. Fluxes Used. The cheapest form of basic flux is limestone, and hence it is almost invariably used. The use of quicklime, CaO, would seem to be advantageous, but it is found that the advantage gained in the operation of the blast furnace is offset by the additional cost of cal- cination of the limestone. Occasionally CaCOs in the form of oyster shells is used as a flux in place of limestone. Limestone used as a flux PIG mON 265 should be very pure, since the presence of acid impurities greatly impairs its efficiency in fluxing the silica, alumina, etc., in the gangue. At least 95 per cent pure CaCOa is desirable.. Pure or high calcium limestones are not always available, and mag- hesian or dolomitic limestones are sometimes used. The replacement of a considerable part of the calcium by magnesium does not appear to appreciably impair the efficiency of the flux and, in fact, several author- ities claim an advantage in the use of magnesian limestone. The Fuel 294. General. The fuel used in a blast furnace must serve as a reducing agent as well as a source of heat. The fuel requirements from the standpoint of heat required always exceed the requirements for reduction, however, and therefore only its thermal value need be consid- ered in this connection. The rapidity of melting attained in the furnace is dependent upon the rapidity of heat production, which, in turn, is dependent upon the rapidity of oxidation of the carbon of the fuel by the oxygen of the air. It is therefore desirable that a fuel, in addition to having a high calorific value, have a porous rather than a dense structure, thereby affording additional surface to the action of the oxygen. It is further necessary that it possess sufficient firmness while being heated so that it will not fill up the interstices of the charge, thereby impeding the flow of the gases. AU solid fuels consist of a combustible portion, carbon and hydro- carbons, which combine with oxygen to form gases, and an incombustible portion which remains as a solid residue called ash, which is insoluble and must be fluxed from the furnace. Three classes of soUd fuel have been used in the blast furnace; raw coal, coke, and charcoal. From the standpoint of structure and access- ibility to oxidation charcoal surpasses coke and coke surpasses coal. From the standpoint of firmness coke stands first and charcoal is least desirjjble. Most of the hard, well-made cokes withstand the pressure of the charge very well, although coke is a rather friable material. Bitu- minous coals absolutely melt down during heating, and anthracites under similar circumstances are apt to splinter into fine particles. Charcoal is the purest of the solid fuels and has the least ash. Anthra- cite coal is less pure than charcoal and has much more ash; it is, however, much purer than coke and has very much less ash. Each type of fuel therefore has its decided advantages and all are being used, the choice in a particular locality being dependent mainly upon relative cost. 266 MATERIALS OF CONSTRUCTION 296. Coal. Anthracite coal is the only natural fuel which can be successfully used in the blast furnace without preparation. It contains very little volatile matter, the fixed carbon amounts to about 90 per cent, it. is ignited with difficulty, and burns slowly. It becomes broken up very finely upon heating. Bituminous coals have been used in blast furnaces abroad, but, because of the above indicated character of the American bituminous coals, they have never been successfully used in this country. 296. Coke. Coke is the soUd residue obtained by the distillation of certain grades of bituminous coals called " coking coals." A coking coal when heated out of contact with air to certain temperatures swells, becomes pasty, and emits bubbles of volatile gases, chiefly hydro-carbons, leaving a residue of fixed carbon, together with the ash, the phosphorus, the sulphur, etc. The weight of the coke is usually about two-thirds that of the coal used. 297. Charcoal. Charcoal is the fixed carbonaceous residue obtained by heating wood without contact of air, the volatile gases being driven off. Charcoal possesses almost the original bulk of the wood used, but is very light and very porous. It is fragile, but will withstand fairly heavy steady pressure. The ash content is very low (less than 3 per cent). Those charcoals made of soft woods at low temperatures ignite most readily, and those made of hard woods at higher temperatures ignite with difficulty. 298. Coke Manufacture. Coke is manufactured by either of two general processes. The older method, which is still largely used, pro- duces " beehive coke," so-called because of the shape of the oven used. The beehive oven is a circular dome-shaped structure built of brick (Fig. 146), about 12 feet in diameter and 7 feet high. Two openings are provided in the masonry, one a 15-inch opening in the crown used for charging, the other a larger opening left at the base for the purpose of withdrawing the coke. A charge of about 5 tons of bituminous coal is dropped through the crown opening and leveled off. The door at the base is then bricked up, with the exception of a sHt at the top about 1 inch wide left as an inlet for air. The mass gradually gathers heat from the masonry, which is still hot from the previous charge, until the ignition point of the slowly distill- ing gases is reached and they begin to burn in the open space above the coal. Sufficient air to support their combustion is admitted through the slit left in the door. The heat of oxidation of the gases raises the temperature of the coal rapidly, and distillation proceeds quickly from top to bottom of the coal. The coal increases in volume considerably during distillation, then fuses xnto a pasty mass, and finally, when all volatile gases have been driven PIG IRON 267 off, it again becomes solid and shrinks below its original volume. The air supply must not be in excess of that required for the oxidation of the gases, as an excess means oxidation of the carbon of the coke. The time required for making furnace coke is about forty-eight hours, while the better grade of foundry coke requires about seventy-two hours. When distillation is complete, the door is opened and the coke is sprayed with water. The consequent cooling causes contraction and splitting up of the solid mass into columnar fragments whose lengths represent the depth of the bed of coke. The entire mass is now with- drawn and further cooled with water, and the oven is immediately recharged. The " by-product " coke process is now being rapidly introduced in place of the older method above described. This process differs from the Fig. 146. — Bee-hive Coke Oven. (Forsythe.) beehive method in that the coking is accomplished in a closed retort, the gases not being burned, but, on the contrary, saved for their com- mercial value. " Retort " or by-product coke is in this country practically always made in one of two types of retort ovens — the Semet-Solvay retort, and the Otto-Hoffman retort. The Semet-Solvay retorts (Fig. 147) are long narrow chambers arranged in batteries of a score or more, the individual retorts being separated by heavy masonry walls. The retorts are 30 feet long, 5 feet 6 inches high, and I65 inches wide. Three openings are provided above for the charging of coal, and one foi^ the escape of gas. The retort walls are lined with hollow flue tiles and a part of the gas driven off from the retort charge is mixed with air and burned in these flues. The heavy brick walls 268 MATEEIALS OF CONSTRUCTION Fig. 147a. — Semet-Solvay Retort Coke Oven. Longitudinal Section. between the retorts are thereby heated to a high degree and the heat retained tends to equalize the temperature in the retort at the beginning and end of the proc- ess. A charge of 4^ tons is cdked in eighteen to twenty-six hours. The retort is charged by overhead mechani- cal devices, and the coke is discharged by a mechanical pusher which forces the entire "retortful of coke out bodily. The water cooling is done out- side the retort. The distillation begins at the retort walls and proceeds inward, the gas escaping at the median plane of the mass, thereby causing a cleavage plane which makes the length of the coke after splitting only one-half the width of the retort. The products of dis- tillation are conveyed by a collecting pipe to the condensing plant where they are relieved of the tar, ammonia, etc. The yield per ton of Pitts- burgh coal is about 1300 pounds of coke, 5000 cubic feet of gas, 75 pounds of tar, and 20 pounds of ammonic sulphate. The Otto-Hoffman retort oven (Figs. 148 149), aside from construction details, differs from the Semet-Solvay retort principally in that provision is made for using a part of the heat of the escaping products of combus- tion to heat the incoming gas and air which are mixed and burned in the SECTION C Fig. 147b.— Semet-Solvay Retort Coke Oven. Transverse Sections. PIG IRON 269 flues surrounding the retorts. The retorts are about 33 feet long, 6 feet high, and 16 to 24 inches wide, and are arranged in batteries of about fifty. The intervening walls are 12 inches thick and contain the flues in which the gases are burned. For this purpose about one-half the puri- fied gas from the retorts is used. The operation of the furnace involves the regenerative principle. The escaping hot products of combustion in the flues are led through chambers (Fig. 148) in which brick are piled to form a loose checkerwork. In their passage they heat up the brickwork of the regenerative chamber before Fig. 148. — Otto-Hoffman Retort Coke Oven. Longitudinal Section. passing on to the stack. After a definite interval of time the course of the escaping burned gases is diverted to a second set of regenerative chambers. The air and gas which are to be used in heating the retorts are now admitted separately through flues beneath their respective hot regenerative cham- bers and in passing up through the hot checkerwork become heated to about 1100° C. before reaching the flues where they are burned. The direction of circulation of the gases through the flues is reversed at fre- quent intervals so that one set of regenerators is always being heated by the burned gases while the second set is performing its function of pre- 270 MATERIALS OF CONSTRUCTION heating the incoming gas and air. An increased heat efficiency is thus obtained. A 7-ton charge is coked in from twenty-four to thirty-six hours, the charging and discharging, and also the condensing of the products of distillation being accomplished in a manner similar to that used in the Semet-Solvay process. 299. Charcoal Manufacture. Charcoal was at one time commonly made by burning in heaps without recovery of the by-products. The volatile constituents of wood are much higher than is the case with coal, how- ever, and the waste in charcoal-making is there- fore much greater than in coke-making. This fact largely accounts for the early adoption of kilns and closed retorts in charcoal- making. Most of the char- coal now made is burned in by-product retgrts with recovery of wood alcohol, acetic acid, tar, gas, etc. As a type of the mod- ern retort charcoal plant, that of the Algoma Steel Company at Sault Ste. Marie is here described. The retorts are horizontal shells, 46 feet long, 6J feet wide, and 8^ feet high, built of boiler plate and fitted with air-tight end doors. The shells are set in brickwork like boilers, and fireboxes are provided at either end. The flues lead along the walls so as to heat the retort evenly. The wood (usually pine, ash, spruce, birch, willow, fir, or alder) is split, seasoned, loaded on trucks, and run into the retort where it is subjected to heat at the temperature of carbonization for from eighteen to twenty hours. The cars are then pushed into an exactly similar retort, which is not fired, allowed to cool for twenty-four hours, and then to a second similar cooler for another twenty-four hours, after which the charcoal is ready for the blast furnace. The products of distillation are led by a collecting pipe '^i Fig. 149.— Otto-Hoffman Retort Coke Oven. Transverse Sections. PIG IRON 271 to the by-product plant where the gas, alcohol, acetic acid, and, later, the wood-creosotes, heavy oils, and tars are successively recovered. The gas is usually used under boilers to generate steam for the operation of the power plant. 300. Relative Use of Different Fuels. The relative extent to which the different fuels above discussed are used in the production of pig iron is indicated from the following data abstracted from " Mineral Resources of the United States," 1913. Bituminous fuel, chiefly coke, was used in the production of about 30,326,000 tons of pig iron, or 97.9 per cent of the total. Anthracite fuel and coke was used in the production of about 278,000 tons of pig iron, or 0.9 per cent of the total. Anthracite alone was used in the production of about 22,000 tons of pig iron, or 0.07 per cent of the total. Charcoal was used in the production of about 340,000 tons of pig iron, or 1.1 per cent of the total. THE REDUCTION OF IRON ORES Manufactuee of Pig Iron 301. General. Practically all of the iron used commercially in the world to-day, whether it be used as cast iron, wrought iron, or steel, is first reduced from the ores in a blast furnace to form pig iron. Many direct processes, by which wrought iron and steel may be made directly fromthe ore, have existed for ages, but their present importance is only an historical one. The first statement above made means, in effect, that all the iron which is used in the manufacture of wrought iron and steels has had an excessive amount of carbon and silicon added in the smelting process, necessitating their subsequent removal at great expense. Neverthe- less, the blast furnace process has been so perfected and cheapened that no direct process can compete with it in any degree. 302. The Blast Furnace Process in General. "The process of smelt- ing iron in the blast furnace consists essentially of charging a mixture of fuel, ore, and flux into the top of the furnace, and simultaneously blowing in a current of air at the bottom. The air burns the fuel, forming heat for the chemical reactions, and for melting the products; the gases formed by this combustion remove the oxygen from the ore, thereby reducing it to metallic form; and the flux renders fluid the earthy materials. The gaseous products of the operation pass out at the top of the furnace, while the liquid products, cast iron and slag, are tapped off at the bot- tom. The escaping gases are combustible, and therefore are conducted 272 MATERIALS OF CONSTRUCTION through pipes to boilers and stoves, where they perform the useful service of heating the blast and raising steam or operating internal combustion engines." * The essential equipment of a smeltery consists of the blast furnace itself, the equipment for handling the charges and other equipment for handhng the products, stoves which preheat the air used for the blast, and blowing engines which supply, air under pressure and deliver it to the fur- nace. 303. The Blast Fur- nace and its Mechanical Equipment. The blast furnace (Fig. 150) consists of a vertical shaft built of steel and lined with firebrick . The lower por- tion,'called the " hearth " or " crucible," is cyhn- drical, about 8 feet high, and 15 to 17 feet in diameter. It contains the " tuyeres," the " cinder notch," and the "iron notch," and serves as a crucible in which the molten products of the operation are collected, .Above the hearth the walls diverge, forming an inverted truncated cone called the " bosh," which „. „ . ,. is 12 to 13 feet high and 22 to 24 feet m diameter at the widest point. Above the bosh extends the " stack " convergmg to a diameter of about 16 to 18 feet at the " throat " at a height of 45 to 60 feet above the bosh. The shell of the shaft above the bosh is so constructed as to be independent of the parts be ow It is supported by a steel ring called the " mantle, " resting upon columns. The walls of the shaft portion of the furnace, called the " inwaUs," have a lining usually 27 inches thick, of hard, high-silica, fireclay brick designed to resist abrasion. The linmg of the bosh is also 27 inches thick, except when the bosh is surface-cooled in which case a 9- or 13i-inch hning is used. The lining of the hearth usually increases * Forsythe, " The Blast Furnace and the Manufacture of Pig Iron," p. 90. Fig. 150. — Blast Furnace and Charging Mechanism. . (Campbell.) PIG IRON 273 in thickness downward, being more than 30 inches thick at the bottom. A more refractory but softer fireclay brick is used in the lining of the bosh and hearth. The ring of " tuyeres " pierce the hearth lining just below the bosh. These are 8 to 16 pipes having an internal diameter of from 4 to 7 inches, through which the hot blast of air is driven. Both the tuyeres and the tuyere-blocks are protected from burning by being made of hollow metal con- struction and cooled by water circulating through them. The hole for tapping off the liquid slag, called the " cinder notch," is located on the side of the hearth about 3 feet below the tuyeres. This also is protected by a water-cooled casting. It is closed by stopping up the hole by an iron bar having an enlarged end, until the slag itself has solidified and plugged the hole. The " iron notch " or " tap-hole," used for tapping out the molten iron, is located at the very bottom of the hearth in the front or " breast " of the furnace. It is commonly stopped by ramming in several balls of clay. The hottest part of the furnace, the bosh, is cooled in one of two ways: The older method involves the use of thick walls in which wedge- shaped hollow castings are inserted. These " cooling plates " are provided with inlet and outlet pipes and are kept full of circulating water. They are so placed as to form a series of rings from 1 to 2 feet apart vertically. The second method consists of " surface-cooling " of the bosh walls, which in this case are not over 13 J inches thick. The cooling is accomplished either by sprays of water directed against the bosh jacket from all sides, or by a spiral trough wind- ing about the boshes and liept full of running water. A late modification of the blast furnace con- sists in the appUcation of surface cooling to the entire stack of the furnace. This necessitates the use of a thin Uning of the " inwalls," usually 12 inches. Both the spiral trough and the spray method of cooling are used. Fig. 151 shows a furnace modified from the old form to permit water cooling and thin lining. The increase in sectional area as the material sinks below the level of the stock line allows for the natural expansion at the higher temperature, and tends to prevent clogging. The reduction in the bosh holds up the material until the fuel is burnt out, and the liquid iron and slag gradually drop into the hearth, where their difference in specific gravity causes them to separate, permitting them to be tapped out sepa- rately. The combustible gases generated during smelting are taken from the furnace just below the bell through outlets, one to four in number, which converge into a single SECTION Of TKIM LINED BLAST FUilNACE Fig. 151.— New Thin-lined Water- cooled Blast Furnace. 274 MATERIALS OF CONSTRUCTION large pipe called the " down-take." An auxiliary pipe connected with the gas outlet pipes, and provided with a valve called the " bleeder," serves as an emergency relief at. times of unusual gas pressure. Near the lower end of the down-take the " dust-catcher " is placed. This is simply an enlargement of the pipe designed to remove solid particles carried over into the down-take by reason of the velocity of the current of gases. Since this velocity is reduced in proportion to the square of the diameter of the conveying pipe, the great reduction in velocity as the gases pass the dust-catcher allows by far the greater part of the suspended particles to settle by gravity. The gas can now be used under boilers or in hot-blast stoves, but cannot be used in internal combustion engines without having first been icleansed in some type of washing device wherein all suspended matter is removed by water. In American practice two long rows of storage bins behind the blast furnaces are kept filled with ore by bottom-dump cars or a conveying device. Between and under the bins runs a track upon which " ore larries " are switched back and forth containing in succession weighed amounts of ore, flux and fuel. These cars discharge into the loading skip of the blast furnace. The amount of material that must be charged into the top of a large blast furnace every twenty-fout hours exceeds 2000 tons, and charging goes on twenty-four hours in the day the year round with never more than a few hours' stop except in the case of serious accident. The charging is accomplished by means of a long double-track inclined skipway, the mechanism of which is controlled from the ground level. The skips are loaded by gravity from the 'larries, elevated to the top of the furnace, and discharged into the hopper automatically. The upper hopper of the furnace is closed at the bottom by an inverted iron cone called the " bell." The counter-weight which holds the bell in place is controlled by the motion of a piston in a steam cylinder. By lowering the bell the contents of the upper hopper are allowed to fall into the hopper proper of the furnace. This larger hopper is also closed by a bell controlled in the same manner from the ground level. At intervals the bell is lowered and the contents of the hopper are distributed in an even layer upon the material already in the furnace. At this time the upper hopper is closed so that there is never a direct opening from the interior to the open air. Some- times elaborate devices are installed to insure an even distribution of the charge as it falls into the furnace, avoiding the segregation of the fine and coarse material that is apt to be caused by the ordinary bell. 304. Hot-blast Stoves. Each furnace is connected with from three to five hot- blast stoves (most commonly four). The design of the stove varies, but the same principles govern the operation of all, so that only one of the type most commonly used, the JuUan Kennedy modification of the Cowper stove, will be here described. The stove consists of a vertical steel cylinder, 20 to 22 feet in diameter and 80 to 110 feet. high, containing two firebrick chambers (Fig. 152). The central chamber is open, while the outer annular chamber is divided into a large number of small flues. Gas from the blast furnace and a definite proportion of air are admitted at the bottom of the open chamber and burned. The products of combustion rise to the top of the furnace and pass downward through the small flues and thence to the stack. The greater part of their heat is taken up by the brickwork of the flues. After burning gas in a stove for about three hours the latter is hot enough to heat the blast. Air from the blowing engines is now admitted at the bottom of the small flues in the outer chamber and passes upward, taking up the heat stored in the brickwork. Thence it passes downward through the central flue to the furnace. The blast tem- perature (425 to 650° C), is kept fairly uniform by working about four stoves per fur- PIG IRON 275 nace, keeping three always " on gas " while the fourth is " on air," and changing stoves about once an hour. The waste gases from the blast furnace amount to about 90,000 cubic feet per minute at a temperature of about 235° C. One-third of this amount, or about 30,000 cubic feet per minute, is required to heat the three stoves " on gas," and the balance is available for power purposes. 305. The Blowing Engines. The air for smelting is delivered under pressure by immense blowing engines, the largest of which develop from 2000 to 2500 horse-power, and will deliver from 45,000 to 65,000 cubic feet of air per minute at a pressure of from 15 to 30 pounds per square inch. The above capacity just about suffices for one large blast furnace. The blowing engines have in the past been steam driven; an important change in the last few years, how- ever, has been the introduction of internal combustion engines which utilize the gas from the blast furnace after washing. 306. Drying the Blast. The pres- ence of a variable amount of moisture in the air of the blast, sometimes amounting to as much as 1 pound per 1000 cubic foot of air, means an expenditure of a great amount for fuel which serves no purpose other than dissociating the water vapor. (The fuel consumed in dissociating 1 pound of moisture per 1000 cubic feet of air will amount in one day to about 25 tons of coke.) The dis- sociation of the water to free hydrogen and oxygen in the smelting zone therefore results in materially cool- ing that zone. The hydrogen and oxygen will recombine to form water vapor again, with the evolution of a corresponding amount of heat, in the upper part of the furnace, but this does not compensate for the loss of heat in the smelting zone, where it is most needed. For these reasons most furnaces have now been equipped with refriger- ating devices whereby the air is cooled far below its dew-point, and its moisture largely reduced by condensation before being drawn into the blowing engines. The method of blast desiccation first introduced by James Gayley Fig. 152.— Hot Blast Stove. Vertical Section. 276 MATERIALS OF CONSTRUCTION in 1904 consists in drawing the air through chambers containing coils of pipe through which a cold solution of calcium chloride is circulated. The heat of the air is transferred to the brine, and as its temperature is lowered its capacity for carrying moisture is reduced and the precipi- tated moisture is deposited on the pipes as frost. The pipes are freed from the frost by sprays of water about once in four days. The calcium chloride solution or brine is cooled to the necessary degree by an ammonia refrigeration plant. The principles governing the operation of such a plant may be briefly summarized as follows: Ammonia, (NH3), is one of the compounds which, although a gas at ordinary atmospheric temperatures and pressures, may be liquefied by sufficiently reducing the temperature or by subjecting it to heavy pres- sure at ordinary temperatures. Having once been liquefied under pres- sure, it will therefore take up_ heat (latent heat) if the pressure is relieved, thus permitting it to reassume the gaseous state. Any material may therefore be cooled to a great degree if it be so placed that its heat may be absorbed by ammonia while the latter is passing through the transition from a liquid to a gas. In practice the ammonia is first subjected to heavy pressure in large compressors and then conveyed to condensers where the heat derived from the compressor cylin- ders is removed and the gas liquefied. The usual type of condensers used consists of coils of 2ih3h pipe with IHnch pipes passing through them. The gas under pres- sure passes through the outer pipes while cooling water is circulated through the inner pipes in the opposite direction. The gas is thus sufficiently cooled to become liquid at the existing pressure and is conveyed as a liquid under pressure to the coolers. The coolers again consist of large pipes within which smaller pipes are placed. The brine is circulated through the inner pipes while the liquid ammonia is admitted to the outer pipes. The sudden increase in the size of the container relieves the pressure on the ammonia, causing it to reassume the gaseous state with the absorption of heat of vaporization from the brine in the adjoining container. The ammonia gas is now reconveyed to the compressors, while the cold brine passes to the coils over which the air passes on its way to the blowing engines. The saving effected by the use of dry blast is two-fold: the saving of fuel used per ton of iron produced, and the saving caused by increased uniformity in furnace operation and consequent increased uniformity of the product. The saving in fuel has been estimated to be about 20 per cent with the same output of iron, or 12 J per cent saving in fuel with an increase of 10 per cent in the iron output. The saving due to increased uniformity in operating conditions is due to the fact that a cubic foot of dry air will always weigh the same and provide practically the same amount of oxygen to burn the fuel, while a cubic foot of atmospheric air will vary in weight and oxygen content if it varies in moisture content. PIG IRON 277 The Functions of the Blast Furnace 307. General. The blast furnace has five distinct duties to perform: (1) It must deoxidize the iron ore; (2) It must carburize the iron; (3) It must melt the iron ; (4) It must render fusible and melt the slag, and (5) It must separate the molten iron and the slag. 308. Deoxidation of the Iron Ore. The deoxidation of the iron ore is of course the primary object of smelting. The recovery of iron would be impossible without deoxidation, because of the operation of the general principle upon which most metallurgical operations are funda- mentally based, i.e., that oxidized bodies in a state of fusion will not unite with unoxidized ones. The application of this principle to the metallurgical processes in iron- and steel-making may be otherwise stated as follows: First, when an element such as carbon, silicon, or phosphorus, exist- ing in chemical union with a metal, combines chemically with oxygen, the resulting oxidized product must, when melted, separate itself from the remaining metallic portion; and Second, if oxidized metal parts with its oxygen, i.e., becomes deox- idized or reduced to the metallic state, the newly liberated portion joins the metal in the furnace. If, therefore, the iron were not reduced, the iron oxide would not be recovered, but would be lost with the slag. A second essential result of the deoxidation of the iron is the fact that the immunity of the firebrick lining of the furnace from rapid destruction by the corrosive action of the slag is due to the deposition of free carbon on the walls during the process of smelting, and, if the slag contained any large amount of iron oxide, this iron oxide would oxidize and remove the carbon coating, leaving the brickwork unprotected against the destructive attack of the slag. 309. Carburization of the Iron. Carburization of the iron is essen- tial because at the temperature attained in at least the greater part of the melting zone it would be impossible to render molten free iron, whereas iron saturated with carbon is sufficiently superheated beyond its melting- point to make it very fluid, so that it easily becomes separated from the slag in the hearth. 310. Melting the Iron. Since, as has been indicated in Art. 308, when fusion takes place all oxidized bodies unite to form the slag and expel therefrom all fused unoxidized bodies, it is essential that the iron 278 MATERIALS OF CONSTRUCTION be fused in order that it may be expelled from the slag just as oil is rejected by water. The molten iron will necessarily absorb all deoxidized sub- stances such as silicon, manganese and phosphorus, which exist as free metals or metalloids in the lower portion of the furnace. Carbon will also be absorbed until the saturation point is reached at about 3i to 5 per cent carbon. It is further essential that the iron be not only fused but, indeed, superheated, in order that it may remain fluid until drained from the furnace and cast into pigs or transported to steel furnaces. 311. Conversion of Gangue to Fusible Slag. The function of the slag formed in the blast furnace is primarily the elimination of all non- volatile matter in the gangue of the ore and in the fuel which does not properly belong in pig iron. This can be accomplished only by giving to the slag such a composition that it will offer a greater attraction to the impurities than does the metal. The slag-making materials consist of the gangue of the ore, the ash of the fuel, and the hme of the flux. The chemical nature of the slag and consequent metallurgical action is controlled by varying the rela- tion of lime (added as a flux) to the other slag-making constituents in the furnace charge. The three slag-making materials upon fusion form a molten silicate of lime, together with magnesia and alumina. The alumina and earthy and alkaline bases naturally enter into the slag, since they exist as oxides and are not reduced in the furnace. In addi- tion, the bulk of the silicon will enter the slag as silica (Si02), and most of the sulphur, by an entirely chemical action, enters the slag as sulphide of calcium, which, although an unoxidized body, does not unite with the molten iron, but appears to dissolve in the slag. (This is the only important exception to the fundamental metallurgical principle above stated.) 312. Separation of Iron and Slag. The separation of molten iron from molten slag follows as a necessary consequence of the operation of the principle that unoxidized bodies in a state of fusion will not unite with oxidized ones. Since the two substances are chemically mutually repellant, the fact that both are very fluid and of different specific gravities affords the only circumstance necessary to make their mechanical separa- tion easy. The slag floats upon the molten iron in the hearth of the fur- nace and may be readily tapped off through the cinder notch above the level of the iron. Operation of the Blast Furnace 313. Starting the Furnace. A new furnace must be first dried for several days by a wood fire built in the hearth. After drying, a scaffold is built just above the tuyeres and two or three courses of cord-wood PIG IRON 279 placed vertically are laid thereon. The wood is followed by a blank charge of coke (mixed with enough lime to flux its ash) extending to a point about midway in the height of the furnace. Upon this bed of coke the charges of fuel, ore and flux are begun and the furnace completely filled. Kindling is placed beneath the scaffold, kerosene is introduced in the tuyeres, the top of the furnace is opened, a light blast started, and the kindling ignited. The top is kept open until the wood smoke disappears, after which the gases are taken care of by the down-take. Charging continues as the stock hne settles, the proportion of ore and flux to fuel being grad- ually increased until the normal burden is reached in a week or ten days. From now on the charging is continuous, the stock hne being kept at a practically constant elevation just below the down-take outlets. 314. Mechanical Control of Furnace and Accessories. The com- puted proportions of ore, flux, and fuel are weighed in the ore larries before being discharged into the loading skips. The upper hopper usually wiU hold only one skipload of material, but the lower hopper is not dis- charged until it contains several skiploads, constituting a properly pro- portioned furnace charge. Various devices are used for accomplishing an even distribution of the stock as it falls into the furnace, the intention being the prevention of segregation of the coarse and fine material. The hot-blast stoves are controlled by a series of valves, one of which regulates the admission of gas, a second the admission of air from the blow- ing engines, a third the outlet of hot gases to the furnace, and a fourth the chimney draught while " on air." All of these valves except the air-inlet, which is a simple gate valve, are usually water-cooled mush- room valves. The temperature of the furnace is mechanically controlled (aside from the efifect secured by varying the burden of the furnace) by control of the temperature and pressure of the air blast, and this is accomplished by regulation of the hot-blast stoves and by regulation of the revolutions per minute of the blowing engines. The slag must be tapped off within ten to fifteen hours after starting the blast and thereafter about every two hours, the interval becoming shorter as the level of the molten iron rises toward the level of the slag notch. The iron is tapped about twenty to thirty hours after starting the blast and thereafter at intervals of about four to five hours. Peep holes are provided in the furnace walls so that the proper time for tapping slag or iron may be observed. The cinder notch is opened up by use of a pointed bar, and closed by holding in the hole a bar (having an enlarged end) until the cinder chills against it. The iron notch is opened by drilling through the clay which 280 MATERIALS OF CONSTRUCTION Fig. 153. Gun for Plugging Tap Hole. closes it, and finally driving in a pointed bar. It is closed either by cut- ting off the blast and driving in balls of clay by hand tools, or by a machine called the " gun " (Fig. 153), which by action of a piston drives balls of clay into the tap-hole without re- quiring complete cutting off of the blast pressure. The subsequent handling of the products will be considered in a later chapter. 315. Metallurgical Control of Fur- nace. Control of Deoxidizing Agencies. The strength of the deoxidizing agencies may be regulated by controlling the hearth temperature, by varying the furnace burden, and by varying the slag composition. The hotter the furnace temperature is, the more powerful will be the deoxidizing action of carbon upon iron, sihcon, calcium, manganese, phosphorus, etc. Increasing the fuel ratio has naturally a direct effect upon furnace temperature, and an indirect effect in increasing the deox- idizing action. Increasing the lime-magnesia content of the slag has the effect of raising its melting-point and as a result, as we shall later see, the hearth temperature is again increased with consequent increase in the strength of deoxidizing agencies. Control of Hearth Temperature. The hearth temperature may be raised in three ways: by increasing the proportion of fuel to ore and flux; by raising the blast temperature and pressure; and by making the slag more infusible. The first and second methods need no further explana- tion. Making the slag more infusible raises the hearth temperature because a very fusible slag becomes very fluid high up in the melting zone and rapidly traverses the hottest part of the furnace without oppor- tunity to absorb much heat. A less fusible slag, on the other hand, will descend slowly through the hottest part of the furnace, gradually melting and becoming superheated, so that when it finally trickles down into the hearth it will convey a great amount of heat thereto. Burdening the Furnace. The proportion of ore and flux to the fuel in the furnace charge is called the furnace " burden." Successful opera- tion depends more upon the proper burdening than upon any other single factor in the furnace management. The method of determining the fur- nace burden is somewhat complex, and is dependent in its details upon the experience gained in the use of a given ore or ores. No attempt will here be made to explain the process in more than a general way. PIG IRON 281 The composition of the slag to be produced, or at least the ratio of acids to bases, must be first assumed, the assumption being based upon the experience of the manager in the use of given materials and upon the character of pig iron required. If, for instance, an iron suitable for basic steel-making is required, the silicon must be below 1.0 per cent and the sulphur below 0.04 per cent. If a low grade Bessemer pig is required, the silicon must not exceed about 1.5 per cent, the sulphur must not exceed 0.05 per cent, and the phosphorus must be below 0.1 per cent. A foundry pig requires 2 to 3 per cent silicon, less than 0.06 sulphur, and 0.5 to 1.0 per cent phosphorus. In general a basic steel pig requires a high temperature and a basic slag (Si02+Al203 below 45 per cent); Bessemer pig requires a high temperature and a neutral slag (Si02+Al203 = 47 to 48 per cent); and foundry pig requires a high temperature and an acid slag (Si02+Al203 above 50 per cent). It is now essential that the complete analysis of ore, fuel, and flux be known. The " available base " of the flux is determined by deducting from the total percentage of bases in the flux the portion required to neutralize the acids in the flux. The quotient obtained by dividing 100 by the available base is termed the " efflciency of the flux." The amount of flux required to neutralize the excess of acids over bases in the fuel ash may now be determined by multiplying the total acids by the ratio of acids to bases in the slag (the slag ratio), thus fixing the total bases required, subtracting therefrom the total bases in the fuel, and multiplying the remainder by the flux efficiency. The flux required to neutrafize the excess acids in the ore is similarly computed (taking account of the fact that a portion of the silica becomes reduced in the furnace and cannot therefore be included in the total acids). The flux required for each of the above offices is now expressed in tons of stone per ton of pig iron produced and the total, when a slight addition has been made to flux the sulphur, will express the total tonnage of Hmestone to be charged per ton of pig iron made. The tonnage of ore required per ton of pig iron produced may be calculated directly from the percentage of iron in the ore; and, lastly, the amount of fuel required per ton of iron produced is determined. This last factor is fixed by two considerations: first, the amount of carbon required to form and melt the slag, and second, the amount of carbon required to reduce, carbonize, and melt the pig iron. These requirements are separately computed, the factors employed being largely determined by the experience of the furnace manager. In general the weight of carbon needed to satisfy the slag requirements will be in the neighborhood of one-fourth the total weight of slag produced per ton of iron, and the iron requirement (although variable because of variation in silicon content 282 MATERIALS OF CONSTRUCTION of the pig), will usually be in the neighborhood of 70 per cent of the weight of the pig. The quotient of the total carbon required by the per- centage carbon in the fuel is the coke required per ton of pig produced., To give an approximate idea of the relative amounts of ore, flux, and fuel in a charge, it may be added that for an average Bessemer pig having about 1.5 per cent silicon and made from ore containing about 55 per cent iron, the weight of fuel required will average in the neighbor- hood of one-half the weight of the ore, and the flux will average roughly one-fifth the weight of the ore. Very commonly in American practice two distinct kinds of ore must be charged, one a soft earthy ore such as is mined on the Messabi range, the other a rock ore which must be used to prevent clogging of the furnace by the earthy ore. The problem of determining the furnace burden is not materially affected except that the two ores must be considered sepa- rately, and the proportion of each as well as the fuel and flux require- ments of each are separately determined. Varying the Slag Composition. It has been noted above that the hearth temperature, the sulphur content, and the siUcon content of the pig iron are largely dependent upon the melting-point of the slag, and the melting-point of the slag is controlled by varying the ratio of bases to acids in the slag. We have seen that increasing the lime-magnesia content raises the melting-point of the slag with consequent increase in hearth temperature and strength of deoxidizing agencies. This has the effect of raising the silicon content and lowering the sulphur content of the iron. The removal of sulphur is especially facilitated by the presence of an excess of lime and magnesia. The melting-point of the slag is con- trolled in actual practice by varying the proportion of limestone charged. 316. Action within the Furnace. Solids and Gases. In consequence of the fact that all of the solid material used in the blast furnace is intro- duced at the top, while the fourth necessary material, the air, enters at the bottom, we may consider that we have in the furnace two moving currents; one a slow current of descending solids, the other a rapid cur- rent of ascending gases. The interactions of these two currents consti- tute the greater part of the changes which take place outside of the smelting zone. The Fuel and Heat Development. The primary source of heat in the furnace is naturally the carbon of the fuel, and if all of the carbon were completely burned the total heat developed would simply be 14,550 B.T.U. per pound of carbon. Since the furnace gases are combustible, however, it is evident that not all of the carbon is completely burned, and since pig iron contains carbon, it is further evident that a part of the carbon is not burned at all. Pia IRON 283- The carbon that is used to reduce the oxides of iron, silicon, mangan- ese, phosphorus, etc., is all completely burned to CO2 with the develop- ment of 14,550 B.T.U. per pound. The balance of the carbon in the fuel, after that absorbed by the iron and a further quantity required to reduce the CO2 of the flux to CO (because CO2 cannot exist in the portion of the furnace where CaCOs breaks up into CaO and CO2) has been deducted^ is burned to CO with the development of only 4450 B.T.U. per pound. In addition to the heat developed by the burning of carbon there is a secondary source of heat — the heat in the air blast which enters at a tem- perature of 425° to 650° C. The heat so introduced directly into the hearth may amount to as much as one-third of the heat developed in the hearth, and perhaps one-fifth of the total heat developed inr the furnace. Chemicol Reactions. The exact nature and sequence of chemical reactions in the blast furnace is not easily determined and, in view of the divergence! in opinions of various authorities, the discussion will here be only a general one. Fig. 154, taken from Campbell's " Metallurgy of Iron and Steel," represents I graphically the chemical phenomena of a blast furnace with probably a fair degree of fidelity. The conditions assumed are as follows: Temperature at tuyeres, 1500° C. Ore = 60 per cent Fe; no water. Coke = 87 per cent C; 1888 pounds per ton of iron. Stone = 100 per cent CaCOs; 1010 pounds per ton of iron. Pig iron =4 per cent C; 1 per cent Si. Volume composition of escaping gases =1 part CO2, IJ parts CO. Temperature of escaping gases =260° C. Height of furnace = 90 feet. An analysis of the diagram will establish the following sequence of chemical actions: The ore as it enters the furnace encounters an atmosphere of gases made up of 16 per cent CO2, 24 per cent CO, and 60 per cent N, at a tem- perature of 250° C. The Fe203 of the ore immediately begins to be reduced by CO, forming Fe304 with the evolution of CO2 and attended by the freeing of a certain amount of carbon. (This carbon becomes deposited in the form of lamp-black upon the solid material of the charge and on the walls of the furnace.) At a temperature of about 500° C, reached when the material has sunk about 13| feet below the stock line, the newly formed Fe304 begins to be further reduced by CO to form.FeO, attended by the further evolution of CO2. This action is represented as completed at a temperature of about 580° C, at a depth of 19 feet. 284 MATEEIALS OF CONSTRUCTION The FeO begins to be reduced by CO to metallic iron at a temperature of 700° C, and practically none exists at temperatures above 800° C. The iron at this time assumes a spongy form. (It is probable that the diagram is in error, since it has been quite authentically shown that by far the greater part of the reduction of the ores to metallic iron is com- pleted at a depth of not more than 12 feet. It is also probably in error in representing the solid carbon deposit as constant all the way down to the tuyeres, because owing to the presence of CO2 liberated by the break- ing down of the limestone, it is rapidly oxidized to CO, and probably only a small portion of it reaches the bosh.) C BEffiESEHTATION OF THE WEIGHT OF SUBSTANCES Fi^nires represent lbs. per 2240 Ibs IN THE BLAST FURNACE of Pij;-lron. HurUontel DlBtADceS FezOs Fe^O. PeO Fe Coke CaCo.GsO CO are proportional to'welgbtB """" '"^ 1010 234C COr. Nitrogen "" FiQ. 154. — Graphical Representation ot Chemical Action within Blast Furnace (CampbeU.) The decomposition of the limestone begins at a temperature of about 800° C. and is represented as completed immediately, though it is prob- able that some raw hmestone exists at a considerably lower level. The CO2 liberated is, as shown above, rapidly reduced by C to CO, while the lime descends to the zone of fusion to flux the acid portion of the charge. The travel of the materials through the region between the point of completion of the decomposition of limestone and the upper limit of the smelting zone is characterized by no chemical action. During this transit the materials absorb much heat, however, and the temperature rises steadily as the tuyeres are approached. The portion of the furnace constituting the lower part of the bosh is called the " smelting zone." It is characterized by chemical actions mainly impossible above this zone because of insufficiently high tempera- tures. The air of the blast is immediately separated at the tuyere level mto oxygen and nitrogen, and the latter, being practically inert, passps upward m substantially its original volume. (A small amount may PIG IRON 285 combine with carbon and potassium or sodium to form cyanides.) The oxygen immediately burns the carbon of the coke to CO2, which is in turn rapidly reduced to CO. The smelting zone is from a chemical point of view primarily the zone of reduction of the metalloids — manganese, silicon, and phosphorus — through the agency of solid carbon, and the removal of sulphur through the joint agency of the lime of the flux and the carbon of the fuel. Manganese is present as the dioxide, Mn02; silicon as silica, Si02; phosphorus as phosphoric acid, P2O5; and sulphur as ferrous sulphide, FeS. The reduction of the oxides of manganese, silicon, and phosphorus by carbon is in each case attended by the evolution of CO. Perhaps two-thirds of the manganese will under ordinary conditions be reduced and therefore be found in the iron. Silica is not so readily reduced, the extent of the reducing action being largely dependent, as above noted, upon the hearth temperature and the basicity of the slag. In general the amount reduced and found as silicon in the pig will not exceed about one-fifth of the amount present in the charge. Phosphoric acid is readily reduced at the temperature of the smelting zone and practically all of the phosphorus in the charge will therefore be found in the pig iron. The behavior of sulphur in the smelting zone is not analogous to that of the metalloids above considered. Whatever the original form of the sulphur in the charge it will probably reach the smelting zone in the form of ferrous sulphide, which is sohible in iron. Sulphur is an acid radical, however, and therefore readily combines with the bases of the slag, par- ticularly lime, and is thereby removed in the slag ifi the form of calcic sulphicie, CaS. Iron is thereby restored to the metal, and the oxygen liberated in the presence of carbon reacts to form CO again. The reduction of metalloids in the smelting zone is not represented in Fig. 154 because the scale of the diagram is such that the small quan- tities of matter involved could only be represented by a single line. 317. Handling the Products. The Iron. The iron when tapped from the furnace is handled in one of two general ways — by casting into pigs, or still molten in ladles. Formerly all the iron was cast in sand pig-beds, which consist of a series of parallel depressions molded in a bed of silica sand on the floor of the cast-house. Fig. 155 shows the layout of such a pig-bed. The individual depressions are connected to cross runners which in turn connect with the main runner leading from the tap hole. Sand casting is now seldom used except in small works, principally because of the fact that the silica sand which adheres to the pigs renders the iron unfit for steel manufacture. A modification of the sand pig-bed, the " chill pig-bed," is now some- 286 MATERIALS OF CONSTRUCTION times used. It is simply a pig-bed made of cast iron, molded in shape very similar to that of the sand pig-bed, but not requiring any prepara- tion beyond sprinkling with a clay wash to prevent the pig iron from sticking to the molds or melting them. The iron pigs are broken from the cross runners by hand sledges and bars or by a mechanical pig breaker, and the cross runners are similarly broken up into con- venient lengths. A large proportion of the iron now molded into pigs is cast in pig- molding machines, one type of which is illustrated by Fig. 156. The machine consists essentially of a continuous series of pressed steel molds carried on an endless chain. The iron runner of the furnace delivers the molten iron into a ladle which is discharged into a spout whence the metal is poured into the molds as they slowly travel past. The ;=aim Fig. 155. — Sand Casting Pig-bed. Fig. 156. — Pig-moulding Machine. iron quickly chills and is discharged into a car when the mold passes over a sheave at the end of the run. On the return the molds are immersed or sprayed with limewater to prevent the pigs adhering. The cooling of the pigs is usually facilitated by depressing the chains and running them through a tank of water. The blast furnace is now so often operated in direct conjunction with a steel plant that the iron is very commonly not cast into pigs at all, but is run directly into ladles which transport it to the steel furnaces. Fig. 157 shows one common type of ladle. The ladle is built of steel, mounted on trunnions on a car-truck, and lined with firebrick. Its capacity is usually 20 tons or more. PIG IRON .287 Slag Handling. The slag which accumulates above the level of the cinder notch is tapped off at intervals of about two hours, while the iron notch is closed. When the iron notch is opened iron free from slag flows at first, but later on in the cast a quantity of slag accompanies the iron, floating on top just as it does in the hearth. This slag is easily separated from the iron by a " skimmer " placed in the main iron runner. The skimmer (Fig. 158) now usually consists of a per- manent cast-iron trough, hav- ing a depression followed by a dam over which the iron must flow. The skimmer is suspended (in slots in the side of the trough) over the depression at such a height that it rests on top of the stream of iron and effectually prevents the slag from being carried over the dam. An open- FiG. 157. — Hot-metal Ladle. (a) Fig. 158. — Cast-iron Slag Skimmer. ing in the side of the trough allows the slag to overflow into a runner, whereby it is carried to the main slag runner which leads from the cinder notch to the point where the slag is discharged into ladles and carried to the slag dump. 288 MATERIALS OF CONSTRUCTION THE ELECTRIC REDUCTION OF IRON ORES 318. General Considerations. The electric furnace has been used in the metallurgical industries for a considerable period, but it is only within the last few years that the difficulties in the way of practicable operations have been removed to such a degree that pig iron has been successfully produced on a commercial scale. The conditions under which the electric reduction of ores can be suc- cessfully carried out are definite, and restricted for the present at least to only a few districts. The following considerations will make the matter clear: In the blast furnace fuel must be supplied to serve two purposes: (1) the introduction of carbon, the oxidation of which supplies the neces- sary heat; and (2), the introduction of carbon to act as a reducing agent. In the electric furnace, on the other hand, the requisite heat is supplied by electrical means and the only carbon required is that needed for strictly reducing purposes. It has been shown that the electric furnace needs from one-third to two-ninths the amount of carbon required by the blast furnace. This therefore means a great saving in the amount of fuel required. On the other hand, the cost of heat produced by electrical means will greatly exceed the cost of heat produced by the combustion of fuel except under unusual conditions, which obtain only in ore districts where the price of fuel is very high and electric power very low. It is only in such districts, therefore, that the commercial extraction of iron by electric means has been or for the present can be successful. 319. The Electric Furnace. Many types of electric furnaces have been used in the metallurgical industries, the principal classes of which may be classed as follows: o. Furnaces using electrodes with an open arc, the heating being done by radiation, called " arc furnaces." b. Furnaces using electrodes that project into the charge or the bath, called " resistance furnaces," and c. Furnaces without electrodes where the bath forms the secondary of a transformer, called " induction furnaces." Only furnaces of the second or resistance type have been employed in the smelting of Ores. Fig. 159 shows the general arrangement of the fur- nace which has been in operation at Domnarfvet, Sweden, for several years, and Figs. 160 and 161 show a modification of the Domnarfvet furnace built at Trollhatten, Sweden, in 1910. This latter furnace is practically a duplicate of furnaces independently developed in the United PIG lEON 289 States at Heroult, California,, at about the same time, the furnace may be described as follows: In its essentials 'Three tu;ererU&e fhiB blow oool tunnel -bfaai. gases' Into the free spaoo above the ore, for the pnnxMe of cooling tfae .roof. Thej also kIvb a better distribution of beat throughout, the charge In the fuoiaoe shaft. isi^^i^^''-^-amF Fig. 159. — Electric Ore Smelting Furnace. Domnarfvet, Sweden. Fig. 160.— Electric Ore Smelting Furnace. TroUhatten, Sweden, The upper portion of the furnace presents an appearance very similar to that of an ordinary blast furnace except that it is much smaller. This portion, the . shaft, is supported on columns i chargipg noo^ over a large crucible which is lined with magnesiteand provided with the usual tap-holes for slag and iron. The shaft is provided with the usual charging ar- rangement of hopper and bell, as well as gas outlets, bleeder, down- take and dust-catcher. The current is intro- duced through six elec- trodes (24 inches in diameter by 6 feet long) which are carried by adjustable mountings so that the depth to which they project into the charge may be kept constant Fig. 161.- -TroUhatten Furnace, Showing Gas-circulating System. 290 MATEEIALS OF CONSTRUCTION as they become burned away. The electrodes are built of carbon and are fitted with a screw and socket so that they can be screwed together, end to end. When an electrode has been lowered as far as it will go, a new electrode is screwed on to its top and thus no part is wasted. Three-phase current, capable of adjustment between 50 and 90 volts, is used. The material of the charge conducts the current, and its elec- trical resistance develops the necessary heat for smelting. The actual rejJuction of the ore is accomplished as in the blast furnace through the agency of the carbon supplied by the fuel, the amount required being only one-third that required in the blast furnace. Charcoal has thus far always been used instead of coke in the electric furnace. The reduction of the oxides through the agency of carbon evolves CO and CO2 gases which pass upward, giving up their heat to the charge in the shaft and performing some reduction in that region through the agency of the CO. The weight of gases evolved in the electric furnace is hardly one-tenth the weight of gases derived in smelting the same quan- tity! of iron in the blast furnace, because no air is driven into the hearth. In Colisequence of this fact the gases have little heat-carrying capacity, and the heat generated in the hearth is not carried to the upper part of the furnace in anything like the degree Attained in the blast furnace. This results in failure to heat the charge to the red heat required for reduc- tion by CO in that region. The solution of this problem has thus far been met by repeated circulation of the gases. After passing the dust- catcher the gases are led through a condenser, where they are cooled sufficiently to remove most of their moisture, after which from one-half to two -thirds of the gas is forced by a blower back to the hearth through tuyeres arranged so that the gas impinges against the crucible arch, thereby cooling it. This arrangement results in causing three or four times the natural amount of gas to pass up the shaft, carrying a proportionally greater amount of heat into the preheating zone, and thus facilitates the reduction of ore by CO. No electric furnaces have been built in sizes to compare at all with the blast furnaces in daily output. The largest furnaces thus far built have a capacity of only 15 to 18 tons of pig iron per day. The industry is just in its infancy, but shows great promise for the future. 320. QuaUty of the Product. The pig iron produced in the electric furnace far excels in quality the usual output of the coke-fired blast furnace, and equals or excels the famous Swedish charcoal pig iron. This may be attributed largely to the fact that it does not contain any of the oxides which are frequently present in pig iron, owing to oxidation by the blast in passing the tuyeres. The absence of nitrogen is also PIG IRON 291 probably a large factor. Still another metallurgical advantage lies in the fact that the temperatures are very high and easily controlled, and the hearth lining is basic, so that a large proportion of lime may be charged, making possible a very basic slag which facilitates the removal of sulphur and even some phosphorus. These considerations have great weight when it comes to making high-grade steel from pig iron. THE USES OF PIG IRON 321. Classification of Pig Irons. Pig irons are classified according to (a) method of manufacture, (6) the purpose for which they are intended, and (c) composition. \ a. Method of manufacture. 1. Coke pig: smelted with coke and hot blast. 2. Charcoal pig: smelted with charcoal, with either hot or cold blast. 3. Anthracite pig: Smelted! with anthracite coal and coke, with hot blast. b. Purpose for which intended. 1. Bessemer pig: for Bessemer or acid open-hearth process. 2. Basic pig: for basic open-hearth process. 3. Malleable pig: for malleable cast iron. 4. Foundry pig: for gray cast iron. 5. Forge pig: an inferior foundry pig used for manufacture of wrought iron. c. Chemical composition. 1. Silicon pig: high in silicon. 2. Low phosphorus pig. 3. Special low phosphorus pig. 4. Special cast irons (spiegeleisen, ferro-manganese, ferro- chrome, etc.). Any of the above irons may be called " sand-cast pig," " chill-cast pig," or " machine-cast pig " according to the method of molding. Of the above classifications the second is by far the most commonly used. The composition of the different grades is usually specified within the following limits: Silicon. Sulphur. Phosphorus. Bessemer pig. Basic pig. . . . Malleable pig Foundry pig . Forge pig 1-2.. 00% under 1 . 00 0.75-2.00 1.50-3.00 under 1 . 50 not over 0.05% under . 05 not over . 05 not over . 05 under . 10 not over 0.10% not specified not over 0.20 0.50-1.00 under 1 . 00 292 MATEEIALS OP CONSTRUCTION 322. The Uses of Pig Iron. Pig iron as such has no structural uses, but a considerable amount is used after remelting in the shape of cast iron. By far the greater part of all of the pig iron made is converted into steel either by the "Bessemer process" or the "open-hearth process" or into wrought iron by the "puddling process." All conversion processes have for their primary object the elimination of the greater part of the non-ferrous elements present in the pig. Any of these processes will reduce the carbon content to any desired point, while the silicon and manganese are necessarily eliminated during the carbon reduction. Phos- 60,889,731 Tons Pic Inn 27,303,567 TOBJ 39.0;! I0,6H,CN» Tom Piff Iron Use Made of , .Pig Iron 33.4^ 9. 081.608 Tom t,H2,TJ2 2.6 JC Bule Oixn B*vth SU«I n Ton I Wd-lraft ,axi,oa i.750.« TODI Fig. 162.— Uses of Pig Iron, 1910. phorus and sulphur are also reduced by the puddling process and by a special form of open-hearth process called the " basic open-hearth proc- ess." The difference between various steels and wrought iron are not so much a matter of chemical composition as they are a matter of physical characteristics which are dependent largely upon the conditions of con- version. The details of the different conversion processes will be taken up later. By reference to Fig. 162 (based upon the statistics for 1910) it will be seen that of the total production of pig iron in the United States about 39 per cent is used in the production of Bessemer steel, 35.2 per cent for open-hearth steel, 23.2 per cent remelted into cast iron, and 2.6 per cent converted into wrought iron. (Proportionate areas on this diagram represent proportionate tonnages.) These figures do not represent the relative amounts of cast iron, PIG lEON 293 wrought iron, and steel produced, because, as we shall later see, the ratio of pig iron used to finished material pro- duced is not fixed, but varies for the different processes, being high for Bes- semer steel and cast iron, and low for open-hearth steel and wrought iron where large amounts of scrap steel and iron are used in the furnace charges. 323. Production of Pig Iron. The pro- duction of pig iron in the United States in its relation to ore production and steel production is shown graphically in Fig. 162 for the year of 1910, and the production for the last forty years is shown by the diagram of Fig. 163, taken from " Mineral Resources of the United States," 1913. The production of ore and pig iron during the period from 1910 to 1913 is reported as follows: (Long tons.) Illliiiil ^fiO Jfl i -- ^- :: " r *^46-- - - I •s -- ^ -- _ §*": :: I- : " ■ 3^_. _' "~ J " ^30-- -- -' - t\ «26-- -- -- J d7 = - -- - t- -I 12C- -- -- -- . JjI I -- -- ..^tltA I >±zz:-.zz;i-M'iU- H 10 - - . : , ? * 5 ^ §"-- - 4--j' s^ : 1 R-- ,-' i'^^nZ ^ ^ :> - -^^ 1 Fig. 163. Production of Pig Iron. 1910. 1911. 1912. 1913. Iron Ore 57,014,906 27,303,567 43,876,552 23,649,547 55,150,147 29,726,937 61,980,437 Pig Iron 30,966,152 CHAPTER XII CAST IRON INTRODUCTORY. PIG-IRON PRODUCTS CLASSIFIED. 324. Pig-iron Products. The principal commercial forms of iron and steel made from pig iron have been indicated above in Art. 322 and in Fig. 162. It will now be necessary, before considering the various individual forms of iron, to make a systematic classification of pig-iron products for purposes of future reference. The scheme presented below is, with minor modifications, that proposed by Howe for the purpose of clarifying the nomenclature of iron and steel. Classification of Iron I. Carbon Class. (Properties chiefly dependent on carbon content.) A. Weld Metal. (Aggregated from pasty mass without later fusion.) 1. Wrought iron. (C = 0.20% or less.) 2. Blister steel. (C = 0.20-2.20%.) B. Cast or Ingot Metal. (Cast as distinguished from aggregated.) 1. Steel. (Malleable when cast.) 1. Bessemer steel. 2. Open-hearth steel. 3. Crucible steel. 4. Electric steel. a. Low-carbon steel. (C = 0.20% or less) b. Medium steel. (C = 0.20-0.30%) c. High-carbon steel. (C = 0.30-2.20%) 2. Cast Iron. (Not malleable, C = 2.20% or more.) a. Gray cast iron. (Carbon in graphitic state.) b. White cast iron. (Carbon in combined state.) c. Mottled cast iron. (Carbon partly in graphitic state.) 3. Malleable Cast Iron. (Cast and then rendered malleable.) 294 CAST IRON 295 II. Alloy Class. (Properties chiefly dependent on content of ele- ments other than carbon.) A. Special or Alloy Steels. (Useful immediately for special purposes.) 1. Nickel steel. 5. Molybdenum steels. 2. Manganese steels. 6. Vanadium steels. 3. Chrome steels. 7. Titanium steels. 4. Tungsten steels. 8. SiUcon steels. (High-speed steels.) '. Ferro Alloys. (Used only to introduce certain elei steels.) 1. Ferro nickel. 5. Ferro molybdenum. 2. Ferro manganese. 6. Ferro vanadium. 2a. Spiegeleisen. 7. Ferro titanium. 3. Ferro chrome. 8. Ferro silicon. 4. Ferro tungsten. 325. Cast Iron as a Material of Engineering Construction. Cast iron differs markedly from, the other general classes of iron products — wrought iron and steel — both in chemical constitution and in physical characteristics. It possesses a very complex constitution which is extremely subject to variation with slight variations in details of manu- facturing processes; it is comparatively coarsely crystalline in structure, possesses considerable hardness, but lacks toughness, melts readily and passes suddenly into a very fluid state, in which condition it will take a quite perfect impression of a mold, but it is nonductile at all tempera- tures and cannot be deformed (as in forging operations) without being broken. Cast iron is not used structurally to nearly the extent to which both wrought iron and steel are so used. It is used to a certain extent, how- ever, for columns and posts in buildings, also for column bases, bearing plates, and innumerable minor structural parts. In machine construc- tion it finds its widest field of application, for no other metallic material which can be cast in complex forms can be produced with such ease and at so low a cost. In this field its principal competitors are malleable cast iron, cast steel, and cast brasses, bronzes, etc. Each of these mate- rials possesses valuable properties not possessed by cast iron, such as greater strength, toughness, or non-corrodibility. When the conditions of service are ones that cast iron can meet, • however, it has no competi- tors. 296 MATERIALS OF CONSTRUCTION THE REMELTING OF PIG IRON 326. Iron Melting in General. A certain small amount of iron cast- ings are made by running the metal into molds just as it comes from the blast furnace. Ingot molds, for instance, are made in this way at steel works. The variability of the blast-furnace product, however, and the difficulty in judging the character and correcting the composition of the molten pig iron, limits the making of " direct castings " to a rather restricted class of products. The bulk of all the iron used as " cast iron " is remelted either in the " cupola furnace " or the " air furnace " before being cast in molds. The process of remelting the iron in the cupola resembles the process of melting ore in the blast furnace, but the reducing action of the blast furnace is not here present. The only office of the air blast used is the oxidation of the fuel of the charge. Remelting in the air furnace bears no resemblance to cupola melting, since the charge is melted in a hearth out of contact with the fuel. The heat is supplied by radiation and reflection from the flame of a soft coal fire maintained in a firebox adjoin- ing the hearth. The Materials Used 327. Foundry Pig Iron. Foundry pig irons were formerly graded, largely upon the basis of appearance of the fracture, as No. 1 Foundry, No. 2 Foundry, No. 3 Foundry, Foundry Forge, etc., the lower numbers representing pigs high in total carbon and silicon, and low in sulphur, phosphorus, and manganese, while the higher numbers and the forge grades are less choice irons, running higher in sulphur, phosphorus, and manganese, and low in carbon and silicon. At the present time the iron used for foundry purposes is practically all bought by analysis, the content of silicon and sulphur being specified, and sometimes also the total carbon, manganese, and phosphorus. In addition to the irons properly classed as foundry pig irons, Bessemer pig, ferro-silicon, and a few other special pig irons are used at times to bring the composition of the cast iron within the required limits. 328. Scrap Iron. The term " scrap iron " is used to designate that considerable portion of the iron charged into the furnace which has been remelted one or more times. It consists mainly of castings discarded after having been in service, but includes also defective castings, gates, sprues, etc., which have never left the foundry. Some classes of castings such as water pipes, for instance, are made without any scrap. For others, the percentage sometimes runs as high as 30 or 40 per cent, the average amount used for all purposes being 20 to 25 per cent of the iron. CAST IRON 297 Scrap being less expensive than new pig iron, it is of course desirable that as much be used as is possible without rendering the cast iron unfit for the purpose for which it is intended. Scrap iron is necessarily ex- tremely variable in composition and, owing to the practical impossibility of obtaining a representative sample for analysis, it is impracticable to attempt to grade it according to chemical composition. It is possible, however, to grade it with a fair degree of accuracy simply by inspection. The character of the fracture is, to an experienced observer, a much better indication of composition than it is in the case of pig iron, where the conditions of cooling and consequent opportunity for crystalline growth are much more variable. 329. The Flux. The ofRce of a flux in the melting of iron is pre- cisely the same as in the smelting of ores, i.e., to abscJrb and carry off in a slag the non-metallic residue of the iron and the ash of the fuel, and to assist in the removal of sulphur. Since the impurities in the charge of the melting furnace constitute only a very small proportion when com- pared with the amount present in the charge of the blast furnace, the percentage of flux required is correspondingly small. The requirements vary greatly, but in general will average in the neighborhood of from ^ to IJ per cent of the weight of the iron. The flux used is calcium carbonate, usually in the form of Umestone, but oyster shells, marble chippings, dolomite, etc., are sometimes used, and a portion of fluorspar (CaF2) is often added to obtain a more liquid slag. 330. The Fuel. The office of the fuel in iron melting is simply as a source of heat. The cupola is capable of developing a higher heat suf- ficiency than the air furnace because the fuel is in direct contact with the metal. Coke is most commonly used as fuel in the cupola, although a mixture of coke and anthracite coal is sometimes used. The air furnace requires the use of a long-flaming bituminous coal. (In some cases gas is used.) The fuel requirements depend upon the character of the castings being made, small and thin castings requiring a hotter metal than large castings. The cupola requires about 25 per cent of the weight of metal in fuel for hot iron, and 8 to 10 per cent fuel for the largest castings. The fuel requirements of the air furnace are about double those of the cupola. The Furnace 231. The Cupola Furnace and its Equipment. The cupola in its essential arrangement is really a very small blast furnace, operated under a blast pressure hardly more than one-twenty-fifth that employed in the 298 MATERIALS OF CONSTRUCTION blast furnace, and intended only to melt the charge without aiiy attempt being made to attain reducing conditions. The type of cupola shown in Fig. 164 is representative of a great proportion of the cupolas used in good-sized foundries and in steel works. It consists of a vertical cylin- drical shell of wrought iron or steel, lined with firebrick set in fireclay grout. The structure is supported on four columns about four feet from the floor. The size of the cupola is quite variable, ranging from about 22 inches inside diameter to about 100 mches, according to the amount of iron required. (55 to 60 inches has generally been found to be the most satisfactory size). The height is dependent upon the diameter, the usual practice being to locate the charging door at a height above the bottom plate Fig. 164 — Foundry Cupola. D r Fig. 166. — Two-impeller Blower. equal to from 3^ to 4 times the internal diameter. The height of the stack above the charging door is governed by considerations of draft, and is therefore variable. The air blast enters the crucible through tuyeres leading from the "wind belt" or "air chamber" which surrounds the lower portion of the furnace. The tuyeres them- selves are not elaborately constructed, but are usually simply iron castings set in the brickwork of the lining. The combined area of all the tuyeres amounts to from 15 to 25 per cent of the cross-sectional area of the cupola. (The larger figure applies when coke is burned and hot iron required, the smaller when coal is used and a less hot iron desired.) Fig. 165 shows two of the most commonly used forms of tuyeres. The tuyeres are usually arranged in two horizontal rings some 12 or 15 inches apart vertically, the area of the upper tuyeres being only a small fraction of that of the lower ones as a rule. CAST IRON 299 The slag hole is situated just below the tuyeres at a, height of iiom 2 to 10 inches above the bottom plate, the height being governed by the amount of metal required per heat, and the provision or non-provision of a "fore-hearth," or "receiver" into which the metal flows from the hearth. (In American practice the English "fore- hearth" is not used, and the tuyeres are often just above the bottom except in the case of cupolas melting metal in steel works. In the latter case the amount of metal required is so great that extra hearth room is called for, necessitating the location of the tuyeres sometimes several feet above the bottom.) The bottom of the furnace is what is called a "drop-bottom." That is, the bottom consists of two or more flap-hinged castings which are normally held in position by a prop beneath. At the conclusion of a "heat" or "cast" the prop is knocked out, allowing all the material remaining in the furnace to be "dumped." The "tap-hole" or "spout" is located just at the level of the bed of sand which covers the bottom doors. The air blast is derived by use of a positive-acting blower, usually of the two-impeller type shown in section by Fig. 166. This blower is very simple and free from mechanical weakness, and is capable of delivering large quantities of air under low pressures. Attempts have been made to increase the efficiency of the cupola by heating the blast and by refrigeration of the air. All such attempts have ended in failure, however. 332. The Reverberatory or Air Furnace. The rever].eratory or air furnace is so, called because of the manner in which the furnace operates. The term " air furnace " is appropriate, since it distinguishes this type from the cupola, which, instead of natural draught, uses forced draught or blast pressure, and the term " reverberatory " applies, since the fur- nace employs the principle of utilization of heat (from flames sweeping through the melting chamber) derived by reflection from the roof upon the bath of the metal. The design of air furnaces shows many variations, but forpresent purposes a de- scription of one typical form will suffice. Fig. 167 shows such a furnace. The main portion of the furnace, the hearth (/), is flanked on one end by a firebox (g); and on the other by a flue leading to a stack (o). The walls of the furnace (a) are of very heavy brick masonry, incased in iron plates (6), and reinforced both ways by tie rods (c) between buckstays (d). The hearth bottom is a mixture of sand and fireclay supported by brickwork built to slope downward from the "fire bridge" (h) to the flue bridge. The "crown" of the furnace is similarly inclined downward toward the stack in order to deflect the heat of the flames downward on the iron in the hearth. The firebox is provided with iron grate bars, and fuel is introduced through a fire- door in one side wall of the firebox. Pig iron and scrap are' charged in through the charging door (j), and holes (m) are provided to facilitate the skimming of the bath of metal. The spout is not shown in the figure, but is so placed as to drain the metal from the lowest part of the hearth. Furnaces in which large amounts of iron, or bulky scrap iron, are melted often have removable sections in the crown through which the charge is lowered by cranes. 333. Relative use of Cupola and Air Furnace. The cupola is used for most foundry purposes in the production of gray iron castings and 300 MATEEIALS OF CONSTEUCTION for melting iron for steel furnaces. The use of the air furnace is prin- cipally confined to the production of white cast iron for malleable cast iron, but it is also used to produce irons of special purity or particular composition for special purposes. Fig. 167. — ^Air-furnace. Operation of Cxtpola 334. Starting the Furnace. The usual method of starting consists in laying a bed of shavings and kindling-wood on the bottom, followed by heavier wood and then fine coke, and lastly a charge of the regular fuel is placed. The tuyere doors are now opened, the shavings lighted at the front, and the fire allowed to burn up until all the fuel is well lighted. Then the tuyere doors are mostly closed and charging is begun. A new method of lighting up is sometimes adopted, an oil torch being employed without the use of wood kindling. Openings or flues through the coke are made directly upon the sand bed by the use of thin boards, and the fuel is ignited by the torch flame entering at the front. CAST lEON ]01 335. Charging. The bed of fuel having been properly prepared and leveled off, the charge of broken pig and scrap is carefully placed, an effort being made to fill up the interstices as far as possible and keep the charge level. The next charge of fuel is now placed on the iron, and alternate charges of iron and fuel continue until the height of the charg- ing door is reached. When a flux is required, the proper proportion (usually i to 1§ per cent of the iron) is charged on top of each bed of iron, except at the start and at the end of the heat, when it may be omitted. The charging of the furnace is often done two or three hours before the blast is started but, even if the blast is put on im- mediately after charging the iron, iron should begin to melt and appear at the tap-hole within fifteen or twenty minutes after starting the blast. If it does not, it indicates that the bed of coals has not been properly prepared. 336. Action within the Fur- nace.* When the furnace is properly operated it shows the following distinct zones of action, beginning at the bottom and proceeding upwards: (1) The crucible zone or hearth, (2) the tuyere zone, (3) the melting zone, and (4) the stack. These four zones are indicated in Fig. 168. (1) The crucible zone extends from the bottom up to the level of the tuyeres. It serves the sole purpose of collecting and holding the molten metal and slag until tapped out. If the metal is allowed to run out of the spout continuously, this zone will be very short, the tuyeres being very close to the bottom. If the tuyeres are located high in the hearth a bath of considerable depth may be allowed to accumulate, but because of the cooling action of the blast from the tuyeres, it will not be as hot iron as that obtained * Bradley Stoughton, " The Metallurgy of Iron and Steel." Fig. 168. — Zones in Cupola. (Stoughton.) 302 MATERIALS OF CONSTEUCTION by collecting the metal in a ladle outside of the furnace. If slag is formed and allowed to collect on the bath of metal it will help to protect the latter from the effect of the blast. (2) The tuyere zone is the zone of combustion. The blast here comes in contact with the red-hot coke and rapidly oxidizes it. A col- umn of coke always extends from the melting zone to the bottom of the crucible, and combustion occurs from the level of the molten metal to a point above the tuyeres, the height of which is dependent upon the pressure of the blast. The blast pressure should not be sufficient to make the top of this zone more than 15 to 24 inches above the uppermost tuyeres. This pressure varies from about 1 pound per square inch for the largest cupolas, to about 5 pound per square inch for the smallest ones. (3) The melting zone is situated directly above the tuyere zone. During the melting the iron is supported on a column of coke, extending to the bottom of the cupola, which is the only solid material below the melting zone. The iron as it melts trickles down to the bottom over the column of coke. Each layer of iron requires about 5 to 10 minutes to melt, and the column of coke is constantly sinking, so that the last of the iron melts several inches lower than the first. If the charges of iron and coke and the pressure of the blast are properly proportioned, each charge of iron will enter the top of the melting zone just before the last charge is completely melted at the bottom. (4) The stack extends above the melting zone to the level of the charging door. Its function is to contain material that will absorb heat to bring it into good condition for action in the melting zone, and to keep the heat in the melting zone as much as possible. The amount of air required for satisfactory operation is largely in excess of the amount theoretically required to burn the coke, on account of imperfect combustion and leakage. Thus, instead of using about 60 cubic feet of air per pound of coke burned, 100 cubic feet or more are actually supplied. 337. Chemical Changes. The chemical changes which take place in the cupola are for the most part incidental rather than intentional. The iron absorbs sulphur from the fuel in trickfing down into the hearth to an amount varying from 0.02 to 0.035 per cent of the weight of the iron. On account of the excess of fuel burned before tapping, the sul- phur is high at the beginning of the run, and, on account of the loss of metal through oxidation, the sulphur is high again at the end of the run. Limestone is decomposed, as in the blast furnace, and fluxes the dirt in the charge and the ash of the fuel. The slag formed absorbs a part of the sulphur, just as it does in the blast furnace, and also carries off what- ever oxides (such as silica or iron oxide) may be formed. CAST IRON 303 The air of the blast oxidizes a small proportion of the iron, and sim- ilarly oxidizes an amount of silicon amounting to from 0.25 to 0.40 per cent of the iron. These latter changes of course mean a loss of metal. The cupola gases consist mainly of the nitrogen of the air together with CO2 and some CO. The presence of the latter indicates a failure to attain complete combustion of the fuel. All efforts to utilize this gas have resulted in failure, mainly, perhaps, because of the discontin- uous operation of the cupola. 338. Duration of the Cupola Run. A foundry cupola is never run continuously, but is started anew for each casting and is " dumped " at the conclusion of the run. As a general rule the duration of the run does not exceed three or four hours, and it cannot exceed this period if no provision is made for draining off the slag. Some large cupolas operated in connection with steel plants are run continuously for -six days in a week, at the end of which time, if not before, stopping is gen- erally required to effect extensive repairs. 339. Tapping Out and Stopping In. The tap-hole is usually left open until the iron begins to run after the blast is started, at which time it is closed if an accumulation of metal is desired. Often the first iron is used only to warm the ladles which are subsequently to be employed in pouring the castings. The " stopping in " is accomplished by the use of the " bod " and " bod-stick." The bod is a plug made of fireclay, sand, or molding sand, molded in the shape of a cone which adheres to the enlarged end of the " bod-stick," the latter being simply an iron bar with an upset end. The bod is thrust in quickly to stop the flow of metal and bakes hard enough to withstand the pressure. A soft bod of molding clay and sawdust is used when the cupola is tapped and stopped very frequently. Tapping consists simply in piercing the bod by use of a round iron bar provided with a pointed end. It is an operation requiring great care, since the danger of causing molten iron to spill and burn those near at hand is ever present. Operation of Am Furnace 340. The Charge. The small scrap, gates, etc., are first placed in the furnace hearth, followed by the flux and the pig iron or the large scrap, the latter being carefully stacked up to utilize all the space. The furnace is either still hot from the last melt, or it is heated for some hours before charging, so that melting begins soon after the charge is placed. 341. Control of Melting. The temperature of the furnace is con- trolled by the regulation of the draft by means of a damper in the stack 304 MATEEIALS OF CONSTRUCTION or stack flue, and one in the firebox. The bituminous coal bums with a long flame which sweeps through the melting chamber, the condi- ditions being strongly oxidizing rather than reducing. A slag soon forms and covers the metal as it accumulates, so protecting it in a measure from oxidation. The: slag is not drained off, but must be skimmed from time to time. The furnace man also rabbles the charge, gradually push- ing the pig down into the bath. The time required for melting is much greater than in the case of the cupola, the actual time being dependent upon the capacity of the furnace. A 10-ton furnace will melt down a charge in three or four hours, while a 30-ton furnace will require eight or nine hours. The fuel requirement is about one-fourth the weight of the charge or about 35 per cent of the iron produced. The metal is tapped as rapidly as pos- sible and, since the hottest metal is at the top of the bath, a series of tap- holes at different levels are sometimes used successively. The loss of metal due to oxidation amounts to from 2 to 5 per cent. of the charge. 342. Advantages and Disadvantages. The product of the air fur- nace is purer than cupola iron, since the metal does not come in contact with the fuel. This means less absorption of sulphur and less absorption of carbon, resulting in general in a higher grade and stronger iron. The process being much slower than the cupola process, it is under better control and any desired composition can be more closely approached. The cupola, on the other hand, is a cheaper installation, requires less skillful management and is therefore cheaper to operate, the heat is more uniform so that all the metal of a melt has more nearly the desired temperature, very hot iron is more easily obtained, the furnace can be started and stopped more readily, the fuel efficiency is greater, and there is less loss of metal through oxidation and consequent removal in the slag. IRON FOUNDmG 343. Iron Founding in General. The art of founding consists in pouring molten metal into a mold of any desired special form, which the metal assumes and retains when cold. The most important part of iron founding is making the molds, a process which demands a very high order of mechanical skill and no small amount of manual labor. No portion of the technology of iron is of more importance to the engineer and machine designer than a thorough understanding of the molder's art. Sand, the material almost universally used for molds, cannot be molded into any conceivable shape, but has certain practical limitations. For the most part, the impression in the sand of the mold CAST IRON 305 is made by a pattern which must be so designed that its removal from the mold is possible. It is not as easy as it may first appear to avoid shapes which would call for a pattern which could not possibly be removed, leaving the mold intact. Another consideration is that sand, when confined, is a comparatively unyielding material, and there- fore the shape of castings must be such that the shrinkage which invari- ably occurs as the metal cools will not induce dangerously high internal stresses. Without mentioning any of the other factors it will be clear that the definite limitations of the molder's art constitute certain limitations for the engineer, and without a practical knowledge of the former iron castings cannot be intelligently designed. 344. Molds and Molding. The various methods of making sand molds may be divided into the following three classes of greatest importance: "green-sand molding" involves the making of an impres- sion of the desired form, by means of a pattern, in a mold composed entirely of sand in a damp state; " dry-sand molding " involves, the making of molds in damp sand by means of a pattern as in green-sand molding, after which the mold is dried in an oven until the moisture is expelled and the sand baked hard; " loam molding " does not involve the use of patterns, but its application is largely confined to articles whose surface are surfaces of revolution, the molds being built up of brickwork covered by a layer of loamy sand in which the desired imprint is made by a " sweep." Machine molding may replace hand work by either of these methods. 345. Green-sand Molding. Green-sand molding is employed in all foundries for the making of the great majority of all classes of iron castings. The exception is very large work which can be handled with greater safety by the dry-sand or loam methods. The sand used for green-sand molding varies according to the partic- ular usage to be made of it. In general it should be a mixture of sili- cious sand and some binder such as clay. The silica gives it a refractory character, while the alumina, in addition to its refractory character, acts as a binder to the sand. The amount of binder should be only sufficient to give the sand cohesiveness without causing it to bake into a non- porous mass which will not permit the gases to escape. The sand should in general be sharp and should be of finer grain for small castings than for large castings. It is used over and over, and needs only gradual replacement. Enough water to give it a proper consistency must be well mixed in after each time the sand is used. With the exception of a limited class of castings such as plates, etc., molds are made in open frames of wood or iron called " flasks." The 306 MATERIALS OF CONSTRUCTION flasks usually consist of two parts (sometimes three or more), so pro- vided with projecting pins and sockets that they may be taken apart and then returned to exactly the same relative position as before. The lower part of the flask is called the " drag," while the upper portion is called the " cope." Fig. 169 shows the type of flask commonly used with good-sized castings. Smaller flasks are not usually provided with the extension handles nor with the crossbars which are required to prevent the sand dropping out of large flasks during handling. Fig. 170 shows a type of flask called a " snap-flask " commonly used for small work. With this type the flask is removed from the mold after =^ Fig. 169.— Flask for Green-sand Mold. BENCH LIFTER SQUARE TROWEL SLICK AND SPOON Fig. 170.— Snap Flask. Fig. 171.— Tools Used in Hand Molding. (Stoughton.) the latter is completed and is used over again instead of remaining until the metal has been poured. Fig. 171 shows most of the tools used in hand molding. 346. Patterns and Cores. Pattern-making is an art in itself involving a considerable amount of technical skill. It will therefore be here possible only to mention briefly a few of the considerations affecting the making and use of patterns. Patterns may in general be divided into two classes, the flrst of which are used to produce solid castings, the second to produce hollow cast- ings. In most cases, however, the pattern is solid, the hollow portion CAST lEON 307 being formed by a " core " which is placed in the mold after the removal of the pattern. The great majority of all patterns are made of wood. Brass or other metals are sometimes used, however, for the sake of greater durability when a great many molds are to be made from the same pattern. The simplest patterns are merely solid wood dupli- cates of the desired castings except that an allowance of about 5 inch per foot is made in dimensions to compensate for the shrinkage of the cast- ings in cooling. Fig. 172 (a) shows a simple turned pattern which is a duplicate of the casting except for the allow- ance for shrinkage and, in addition, a slight tapering of the vertical plane surfaces in order to facilitate the with- drawal of the pattern from the mold. The taper is indicated by the dotted lines; it is called the draft of the pattern and, on the average, amounts to about ^ inch per foot. Such a pattern is not easily molded because of the necessity of arranging a parting of the mold on a diametral plane of the pattern in order that the latter may be withdrawn from the sand. The same mold may be made much more easily by making a " parted pattern," A ^ ^ \ Fig. 172a.— Simple Turned Pattern. HS 4 TiTl 4 W Fig. 1726.— Simple Parted Pattern. Fig. 172c. — Green-sand Molding. 1st Stage. Fig. 172d. — Green-sand Molding. 2d Stage. Fig. 172e. — Green-sand Molding. 3d Stage. One ting SI Fig. 172/. — Green-sand Molding. Section of Finished Mold. 308 MATERIALS OF CONSTRUCTION turned from two separate pieces of wood clamped together as shown in Fig. 172 (6). The two halves of the pattern after completion are fitted with " dowels " which hold them in the proper relation the one to the other. In order to improve the appearance of castings as well as save the molder extra trouble and obtain- stronger castings it is always necessary to avoid all sharp angles in making patterns. This is done by rounding off all outside corners and placing a " fillet," Fig. 173, on the inner corner. Fillets are cut from wood or leather and glued in place, or else beeswax is used, the fillets being fashioned in place by a hot fillet iron. Fig. 173.— Use of Fillets in Pattern Making. Fig. 174a. — Cast-iron Flanged Tee. Fig. 1746.— Pattern for Flanged Tee. Tig. 174c.— Core Box for Flanged Tee. When a core is used to obtain a hollow casting the pattern must be provided with "core-prints." These are simply projections beyond the pattern proper, built integral with the pattern. The prints form impressions in the mold which will subsequently be filled by extensions on the core. The cores will thus be supported in place in the mold. Fig. 174 (6) is the form of the pattern required for molding the flanged tee shown in Fig. 174 {a). Fig. 174 (c) shows the " core-box " in which the required core is molded. The cores are usually made from a mixture of fine silicious sand with clay or loam, which is packed in the core-box while damp. Some binder such as flour and water, rosin, or glue-water, is usually required CAST IRON 309 to give the core strength enough to permit handling, and holes are usually left running lengthwise through the core for the purpose of venting (allowing escape of gases). When the cores are dried in an oven prior to use in a mold they are called " dry-sand cores." The great majority of all cores are of this class. If a core is long, particularly if it is sup- ported horizontally over a considerable span, it is necessary to stiffen the core by insertion of iron or steel wires or even skeleton frames ef metal. Cores which are adjacent to thick parts of a casting are apt to be fused, unless protected from the heat of the slowly cooling molten metal. They are therefore often daubed with an insulating coating of " blackening." For this purpose pulverized graphite or plumbago is either applied wet as a wash, dry with a brush, or shaken from a cloth bag. 347. Making a Mold in Green-sand. The general features of green-sand molding may perhaps be best indicated by a description of the process of molding such a pattern as is shown in Fig. 172 (6). ' One-half of the pattern is placed upon the molding board within the "drag " of the flask, the latter being placed upside down. Sand is now sifted into the flask until the pattern and molding board are well covered, after which unsifted sand is shoveled in and carefully packed and rammed until the flask is filled (Fig. 172 (c)). The top is struck off with a straight-edge and covered with a " bottom board." The molder now lifts the flask and turns the mold over, replacing it on his bench in the position shown in Fig. 172 (d), the bottom board and molding board being held in place by clamps during the operation of "rolling over.'' The molding board is now removed and the surface is "sleeked " over with a trowel to give it a fine finish. "Parting sand " (usually obtained by fine screening of the burnt sand cleaned from the castings) is now sprinkled over the surface to prevent adhesion between the two bodies of damp molding sand, the upper portion of the pattern is put in place, and the cope is fitted to the drag. At this stage it is necessary to make provision for the admission of the molten iron and usually also for the removal of the slag and dirt. "Sprue-pins " (Fig. 172 (e)), consisting simply of cone-shaped pieces of wood whose lengths correspond to the depth of the cope, are placed in the cope at a proper distance from the pattern, the pins being held in place by a projecting pointed wire which is thrust down into the sand of the drag. The sand is now packed into the cope in exactly the same manner as was done with the drag, the sprue-pins are withdrawn, the cope lifted off, and the pattern care- fully rapped and drawn. Any imperfections are now patched up and sleeked down, and a coat of blackening (usually graphite) is applied to the surface of the sand where imprinted by the pattern. The gates must now be cut connecting the mold with the riser and the pouring- and feeding-gates, after which the mold is "vented " by thrusting a vent wire into the sand at frequent intervals all around the mold, each vent being connected by a channel with the outside of the mold. These vents tend to permit the escape of gases generated when the castings are poured. The section of the finished mold will now appear as shown in Fig. 172 (/). (Often on small castings like this one the dirt riser may be omitted, the feeding gates being depended upon to serve a double purpose). After completion of the gate, cutting and venting, the cope is replaced and clamped to the drag and the mold is ready for the molten metal. 310 MATERIALS OF CONSTRUCTION 348. Dry-sand Molds. Dry-sand molding is applied principally to the making of rather large castings, especially when smooth surfaces are required. The lesser degree of skill required in making molds in dry sand permits the employment of inferior workmen. Less gas and vapor are generated when the mold is poured than is the case with green-sand, therefore necessitating less careful venting of the mold. Wooden flasks cannot be used in dry-sand molding on account of the fact that the molds are thoroughly dried in ovens at a temperature of from 150° to 200° C. The sand required for dried molds must be of a more loamy nature than that used in green-sand molding, and in ramming the mold less care need be exercised to attain just the required degree of hardness. The manner of making the dry-sand mold differs little from the above described method of molding in green-sand, except in the matter of ramming and venting above noted. Drying the mold has the effect of driving off the moisture and leaving a firm, hard, semi-baked mass. The ovens are heated by coke, coal, or gas, and the temperature is carefully controlled to prevent burning and consequent crumbliness, or under-baking and consequent softness. The time required for oven drying is dependent upon the size of the mold. Small molds may perhaps be dried in an hour, while some large molds may require a day or more. Dry-sand molds have the advantage over the green-sand in that less skill is called for in the making. They are stronger, and therefore the danger of sand-holes caused by the inclusion of loosened molding sand in the casting is less, also there is less danger of the mold yielding under the pressure of the molten metal. Furthermore, they tend to produce sounder castings, owing to the fact that far less gas is generated at the time of pouring. On the other hand, the dry-sand mold is apt to shrink and become distorted during the drying; there is more danger of excessive cooling stresses and possible checking of the castings, due to the lesser degree of yielding of the sand as the cooling metal contracts; minor repairs to the mold are less easily made; and considerable extra expense is involved in the extra time and labor of handling in drying. 349. Loam Molds. The use of loam molds is restricted largely to that class of large castings whose surfaces are surfaces of revolution, the expense of the method not being justified as a rule under other cir- cumstances. Exceptions to the rule are a few special classes of castings, such as gear wheels, where, by the use of special machines, the mold may be formed in loam with a considerable saving by reason of the fact that the making of a complicated pattern is unnecessary. CAST IRON 311 To explain the general features of loam molding the making of a mold for a hollow cylindrical casting will be described. The first step in the process consists in cutting boards in such a way that, when bolted to projecting arms mounted on a vertical spindle, their edges constitute elements of portions of the cylindrical surface of the finished casting. Those boards are called " sweeps " or " loam-boards " (D Fig. 175 (a), F Fig. 175 (6), H Fig. 175 (c)), and it will be evident that when the sweep is rotated about the spindle its edge will generate the desired cylindrical surface. The steel "bottom-plate" (F Fig, 175 (o)) is leveled up on blocking, the spindle seat Fig, 175, — Successive Operations in Making a Loam Mold. (Bale.) is embedded in the floor in a central position, and the spindle set up and its bearing adjusted. The molder now constructs the foundation and baSe of the mold by laying up a brickwork ring which is cleared by the sweep by about i-inch on all sides. Loam, and not mortar, is used in laying up the brickwork. The joints are sometimes filled with fine cinders to aid in venting the mold. A coarse loam mixture is now daubed on the brickwork and the entire surface is swept up. A finishing coat of fine loam is next applied, and the sweeping continued until the surface has a smooth finish. Finally the entire base is lifted and taken to an oven and dried. 312 MATEEIALS OF CONSTRUCTION The " lifting ring " (0 Fig. 175 (d)) is now laid on the bottom plate outside the base of the dried mold and the brickwork of the " cope " {H Fig. 175 (d)) is laid thereon, the sweep F having replaced the one used for the base. After each few courses of brick are laid the loam mixture is daubed on and swept up and finished as before. When the cope is complete, chains are hooked into the lifting plate, and the cope is lifted off the base by a hoist and moved to the oven while the " core " is being built. The sweep and spindle have been removed during this last operation. The spindle is now replaced and the cope-sweep is replaced by the core-sweep {H Fig. 175 (c)). The core is now built up in the same manner as was the cope, after which the spindle is finally removed, the core is dried and replaced on the base, and the cope Ufted back into place. The top plate (T Fig. 175 (d)) is now lowered into place. This plate is studded with projections which hold in place a layer of loam which has been applied and dried on beforehand. Holes in which sprue-pins are inserted ((? Fig. 175 (d) ) are provided at intervals in the plate to provide pouring gates and risers. The inside of the core is now filled with sand, and more sand is rammed to a depth of a foot or more on top. Finally a circular runner is cut in this sand to connect the several pouring gates, the sprue-pins are removed and the mold is ready for pouring. Such castings as the one described are sometimes so molded in a sectional loam mold that the mold need not necessarily be destroyed in removing the castings. In this event the mold may be used for several castings with only minor repairs. 350. Chilled Castings. A " chilled casting " is one made in a mold, some parts of which, at least, are made of iron, such portions of the mold being called " chills." The purpose of introducing chills into a mold is to convey away the heat of the molten metal rapidly, a treatment which, for reasons to be discussed later, has the effect of causing the car- bon in the iron to remain in chemical combination, instead of separating therefrom in the form of graphite, as it normally does in slowly cooled castings of gray iron. The physical effect of chilling on the character of the iron is to greatly harden it for a certain depth, giving to the exterior of the iron the characteristic appearance of white iron, whije the body of the casting remains a gray iron. ChiUing is principally used for wearing surfaces of such castings as iron rolls and treads of car wheels. The manner of molding chilled castings differs in no respect from the methods above described except in the in- terposition of the chills in the moid. The dry-sand Fig. 176.— Use of Chills on Treads of Car Wheels, method of molding is most commonly used. Fig. 176 shows the manner of introducing chills to harden the tread of a car wheel. It is often difficult to secure the depth of chill desired (often | inch or more) on account of the shrinkage of the casting and the expansion of the chill. On this account several special types of sectional chills have been devised, their arrangement being such that they may be made CAST IRON 313 to contract as the casting contracts, thus keeping in contact therewith for a longer period. Chilled castings are allowed to cool till near atmospheric temperature before being removed from the mold, or else they are removed as soon as they are solid, and placed in armealing pits where they are cooled still more slowly for a period of two or three days. This slow coohng has the effect of relieving internal stresses caused by the non- uniform rate of cooling in the mold. 351. Pouring the Iron. In small foundries it is usually the practice to stop molding in the middle of the afternoon and pour off all the molds that have been made. For small work the metal is caught in hand ladles at the cupola tap-hole and conveyed to the molds by one or two men. The ladles must be heated by allowing a part of the first iron run to stand in them for a few moments before the actual pouring is begun. Other- wise the ladle will chill the metal and the iron will not enter the mold at the required temperature. Practically all foundry ladles are " top- pouring " (i.e., the metal is poured by tipping the ladle), rather than " teeming " ladles, which are provided with a valve in the bottom. This necessitates the use of a bar to keep back the slag which floats on the metal and would otherwise enter the mold. Care is exercised by the molder to hold the ladle as near the pouring gate as possible to lessen the impact of the stream of metal upon the sand of the mold. The proper time to cease pouring is indicated by the appearance of the metal at the top of the riser. Each mold -must be filled in one operation, and there- fore when the ladle does not contain enough metal to completely fill a mold its contents are emptied into pig-beds molded on the foundry floor. Larger foundries employ traveling cranes and large ladles holding perhaps a ton or more of metal. The ladles are, however, of the top- pouring type. During the filling of the molds, it is usually the practice to ignite the gas which escapes from the vent holes, thereby preventing accumulation of the gas in the molding room. The flasks are removed soon after the completion of the pouring, and the molds are dumped off the bottom boards in piles, from which the hot castings are hooked out and allowed to cool. The gates and run- ners are now broken off by a few sharp blows with a hammer and the castings removed for cleaning. 352. Cleaning the Castings. The sand which adheres to the cast- ings is usually removed by one of three methods: rattling them in a " tumbling barrel," pickUng, or sand blasting. Rattling is most commonly practiced in case of small castings. The tumbling barrel is simply a short horizontal cylinder which is mounted on trunnions. Fig. 177. The castings are piled into the barrel, together with a quantity of abrasive material in the shape of small, irregularly 314 MATERIALS OF CONSTRUCTION shaped, hard iron " stars " or " picks." The barrel is rotated slowly and the falling about of the castings and the stars gradually knocks the burnt sand and scale off the surfaces of the casting. This method has the disadvantage of producing a hard skin upon the castings which causes difficulty if they are subsequently to be machined. Rattling will never completely clean any but very simple castings, and a better method consists in pickling the castings by immersion in a dilute sulphuric or muriatic acid solution. The acid attacks the iron somewhat, thereby loosening the sand and scale. Pickling in a 15 per cent solution of sulphuric or muriatic acid requires about twelve hours and must be followed by a careful washing in water. Hydrofluoric acid now sometimes replaces sulphuric or muriatic acid. The former Fig. 177. — Tumbling Barrel. attacks the sand itself, instead of the iron, and with only about a 5 per cent solution castings may be cleaned in an hour or less. The sand blast is the most convenient method of cleaning large cast- ings, especially those of very irregular form, such as gears, etc. Very often the sand blast is followed by pickling. The final operation in the preparation of castings for use consists in smoothing up the irregularities left by breaking off the gates, the " fin " formed where the metal has run between the two portions of a mold, etc. With small castings this is most readily done with an emery wheel; even a file may be used. With larger castings chipping with a cold chisel is often necessary, and a pneumatic chipping tool is most efficient. Portable emery wheels fitted with a flexible drive are now sometimes used for this purpose. CAST lEOK 315 properties of cast iron Constitution 353. Essential Constituents of Cast Iron. The composition of cast iron is that of a complex alloy containing usually six important elements, together with other elements of less frequent occurrence. The ele- ments invariably present are, in the approximate order of their impor- tance, iron, carbon, siUcon, phosphorus, sulphur, and manganese. In addition, copper, nickel, oxygen and nitrogen are often present, and aluminum, titanium, and vanadium are sometimes added. The constitution of cast iron is much more complex than the composi- tion, because of the variety of compounds which the eleinents present combine to form. The most important consideration affecting the character and properties of cast iron is the carbon content and, in par- ticular, the form assumed by the carbon, i.e., whether free as graphite, or in chemical combination with the iron as a carbide. The importance of the elements other than the iron and carbon is chiefly due to their influence upon the state assumed by the carbon. Before we can state even in a general way the essential constituents of cast iron we must recognize the existence of three principal classes of cast iron, the difference in character being due to the different states in which the carbon occurs. Gray cast iron is that in which the carbon occurs chiefly in the graphite state. White cast iron is that in which the carbon occurs chiefly as the car- bide of iron. Mottled cast iron is a mixture of particles of gray iron with particles of white iron. The essential constitution of gray iron is that of an aggregate of very impure " steel " mechanically mixed with graphite. The matrix which we speak of as " -steel " has two main constituents, " ferrite " and " cem- entite." Ferrite is that part of the steel which is practically pure iron, containing only a trace, or. even no carbon in solid solution. Cementite is a definite carbide of iron having the formula FesC. White cast iron has the constitution of very high carbon steel except for the great amount of impurities present. It consists of a large amount of cementite together with a small amount of " pearlite," the latter being a certain mixture, of a more or less definite composition, of ferrite and cementite. 354. Carbon in Cast Iron. When cast iron solidifies from the molten state the carbon probably remains in the combined condition as car- 316 MATERIALS OB' CONSTRUCTION bide of iron, FesC, which is partly free as cementite and partly in solid solution in the iron as " austenite." * The FeaC is an unstable com- pound, however, and when formed at a high temperature is readily decomposed into graphite and iron. The decomposition of the carbide with the consequent formation of graphite carbon is facilitated particularly by a slow rate of cooling and by the presence of silicon. It is retarded, on the other hand, by rapid cooling or by the presence of much sulphur or manganese. 355. Gray Cast Iron. Cast irons containing considerable amounts of graphite carbon are known as gray, cast irons, because of the grayish or blackish coarsely crys- talline appearance of their fractures. This appearance is caused by the presence of many irregular and generally elongated and curved plates of graphite im- bedded in the matrix of ferrite and 'cementite. These plates of graphite are made up of smaller plates, somewhat like sheets of mica, and may be split apart with ease. The individual sheets of graphite vary in size from microscopic pro- portions to one-eighth of a square inch or more in area. The character- istic structure of gray cast iron when highly magnified is shown in the photomicrograph of Fig. 178. The irregular dark bands are graphite plates, the intermediate area being the ferjite-cementite matrix. The actual percentage by weight of graphite in gray cast iron will be "between 2 and 4 per cent, the amount of combined carbon being under 1§ per cent. The volume content of graphite is much higher, however, since iron has a specific gravity about three and one-half times that of graphite. The volume content of graphite will, therefore, amount to from 7 to 14 per cent. * See page 422 for an explanation of the term " solid solution," and page 430 for a definition of "Austenite." Fig. its. — Gray Cast Iron. Magnified 100 Diameters. (Boylston.) CAST lEON 317 The great difference iil the character of gray iron and white iron is readily seen to be due primarily to the following considerations: The presence of much graphite means that httle iron carbide, which is very hard and possessed of great physical strength, will be present. At the same time the occurrence of weak and soft but tough ferrite is increased, and lastly, the graphite itself is weak and forms a more or less complete mesh separating the matrix into partially isolated particles to which it adheres very slightly. It is therefore to be expected that gray iron will be soft and weak, although comparatively tough, while white cast iron is hard and strong but brittle and difficult to work. 356. White Cast Iron. Cast iroii, the Inilk of whose carbon is present in chemical combination with iron as carbide of iron (FesC), or ccmentite, is called white cast iron be- cause of the white, highly metallic fracture which charac- terizes it. As noted above, the ferrite and a portion of the cementite together form pearlite, so that the ultimate constitution of an iron free from graphite will be a mixture of cementite and pearlite. PearUte is a mixture of ce- mentite and ferrite, the two components occurring most commonly as small wavy or parallel plates of alternately light and dark color. The characteristic appearance of laminated pearlite under high magnification is shown in Fig. 179. Ferrite and cementite combine in rather definite proportions to form pearlite the proportion of carbon to pure iron being 1 : 120. The appearance of white cast iron when highly magnified is shown in Fig. 180, wherein the Ught areas are free cementite, while the dark-banded areas are pearlite. While white cast iron is structurally not distinct from very high- carbon steel, if the occurrence of larger amounts of elements other than iron and carbon is disregarded, it lacks entirely the malleability of steel and is extremely hard and brittle because of the presence of a very large proportion of the hard and brittle free cementite. The dividing line between high-carbon steel and white cast iron lies at about 2.2 per cent carbon but, as a matter of fact, most steels do not approach 2 per cent Fig. 179.— Pearlite. Magnified 1090 Diameters. (Osmond.) 318 MATERIALS OF CONSTRUCTION carbon and few white cast irons have less than 2.25 per cent or even 2.50 per cent carbon. 357. Mottled Cast Iron. Irons which contain particles of gray cast iron mixed with particles of white cast iron, and which are therefore non- homogeneous in character, are sometimes produced. Such irons are called mottled cast irons. They have no special adaptation, and their production is largely unintentional. The microscopic structure of such an iron is shown in Fig. 181. Fig. 180.— White Cast Iron. Magnified 500 Diameters. (Wust.) Fig. 181.— Mottled Cast Iron. Magnified 500 Diameters. (Wust.; 358. Silicon in Cast Iron. After iron and carbon, silicon is, in its effects upon the character of the iron, the most important element pres- ent in cast iron. We have seen that the amounts of silicon and sulphur retained in pig iron are under the more or less complete control of the blast-furnace manager, and the amounts present in cast iron are similarly under the control of the foundry manager. The content of carbon and phosphorus, on the other hand, is not under control, nor is that of man- ganese, except to a slight extent. Silicon and sulphur, however, have a marked effect upon the condition of the carbon and, in consequence, they exert a powerful influence upon the properties of iron. Silicon combines with a part of the iron to form silicides (Fe2Si, FeSi, etc.), which dissolve in the ferrite portion of the iron. Silicoji's primary effect upon the carbon is as a precipitant, driving the carbon out of combination into the graphite form. The maximum precipi- tation of graphite seems to occur with about 2.5 to 3.5 per cent of siUcon, and when the siUcon exceeds this limit the effect is reversed, the propor- tion of combined carbon or cementite being increased. CAST IRON 319 Silicon in percentages below about 3.0 per cent acts, therefore, as a pronounced softener, producing soft gray iron, but larger percentages result in the formation of hard and brittle white iron. Small percent- ages also produce freedom from oxides and blowholes, promote fluidity, and decrease shrinkage and depth of chill. 369. Sulphur in Cast Iron. The influence of sulphur upon the form assumed by the carbon in cast iron is exactly the reverse of the influence of silicon. That is, the higher the sulphur content, the higher will be the proportion of combined carbon. This tendency upon the part of sulphur is much more potent than is the opposite tendency exhibited by silicon, however, a given amount of sulphur being able to neutralize about fifteen times as much silicon. Sulphur therefore tends to produce hard, brittle, white iron. Aside from the effect of sulphur upon the character and properties of iron consequent upon the fact that the carbon is driven into combina- tion as the carbide, cementite, sulphur inherently possesses the power to materially affect the behavior of iron in solidifying and cooling. Only a few tenths of 1 per cent of sulphur suffices to render iron very tender at a red-heat ("red-short"), and therefore apt to check or crack if in solidifying the shrinkage causes the casting to tend to crush the sand of the mold, thus resulting in the setting up of internal stresses in the iron. Sulphur also causes solidification to become very rapid, and often is responsible for the presence of blow-holes and sand-holes. Manganese, because of its great affinity for sulphur, will tend to rob the iron sulphide, FeS, of its sulphur, forming MnS, which latter compound is much less potent than FeS in affecting the proportion of combined carbon. A given percentage of sulphur may, in general, be neutrahzed by the presence of about twice as much manganese. Speci- fications usually limit the maximum sulphur content of gray cast iron to not over 0.10 per cent and often the maximum allowance does not exceed 0.05 per cent. 360. Phosphorus in Cast Iron. The effect of phosphorus upon the state assumed by the carbon is rather self-contradictory; chemically, it tends to increase the proportion of combined carbon, especially when the silicon is low and the phosphorus high. On the other hand, phos- phorus lengthens the time of sohdification, thereby affording additional opportunity for the precipitation of graphite. When the sihcon is high, therefore, the presence of moderate amounts of phosphorus actually increases the precipitation of graphite, but when the proportion of phosphorus is very large, the chemical effect is great enough to retain the carbon in the combined form in spite of the longer period of solid- ification. 820 MATERIALS OF CONSTEUCTION The presence of phosphorus in considerable amounts tends there- fore to produce a hard white iron, lacking in toughness and workability, and especially lacking in shock resistance when cold. Phosphorus reduces the melting-point of iron and makes it very fluid. It is therefore useful in making very thin castings where a less fluid iron will not take a per- fect impression of the mold. Not more than 0.05 per cent of phosphorus is allowed in best gray iron, while from 1.0 to 1.5 per cent is sometimes used when fluidity is more important than toughness. 361. Manganese in Cast Iron. Manganese increases the total carbon content of cast iron and also increases the proportion of combined carbon, though it is much less potent in this latter respect than is sulphur. As above noted, however, the effect of manganese cannot be con- sidered. apart from that of the sulphur. If no more manganese is present than is required to combine with the sulphur, forming MnS, its effect will not be to increase the proportion of combined carbon, but will be just the reverse, because the sulphur is taken from the sulphide, FeS, which is so powerful in causing the carbon to assume the combined form. Any additional manganese unites with carbon to form the carbide MnsC, and this carbide unites with the FesC, causing the cementite to be made up in part of the double carbide of iron and manganese (FeMn)3C. It appears, therefore, that manganese up to the amount which com- bines with sulphur to form MnS tends to lower the proportion of com- bined carbon and consequently decreases the hardness and brittleness of the iron. Any additional manganese, however, has a marked effect in causing the carbon to assume the combined form, and is therefore a hardener. Large percentages of manganese are sometimes added to cast iron designed for use as " spiegeleisen " or " ferromanganese " in steel making, but for ordinary castings the manganese seldom exceeds 2 per cent and may be as low as 0.10 per cent. Behavior of Iron in Cooling. 362. Shrinkage. The shrinkage of cast iron is an important con- sideration for the pattern maker, because due allowance for shrinkage must be made in the dunensions of the pattern if the casting is to con- form to the size called for by the drawings. It is also an important consideration for the designer and the founder, because the stresses set up in cooling and the consequent danger of checking are directly depend- ent upon the degree of shrinkage if the casting be of such a shape that its shrinkage tends to crush the sand in the mold. CAST IRON 321 All metals expand upon heating and contract when coohng, and the total expansion in melting a metal will correspond to its total shrinkage in soUdifying and cooling. Pure iron shrinks "about 0.3 of an inch per foot; a less pure iron usually shrinks less, because impurities, particularly carbon, usually lower the melting-point. In addition, the separation of carbon as graphite exerts a power- ful influence upon the total net shrinkage of iron because of the expan- sion which its separation causes. The factors which chiefly determine the amount of shrinkage are therefore the factors which chiefly control the separation of graphite, i.e., the silicon content and the rate of cooling. Moreover, since the lat- ter is largely dependent upon the size of the castings, the shrinkage becomes largely a func- tion of silicon content and size. This relation- ship is shown graphically by Fig. 182, which is based upon experiments made by W. J. Keep.* The shrinkage is shown to be inversely propor- tional to the per cent of siUcon, and for an iron of given composition the shrinkage decreases as the size of casting increases. Other elements whose presence affects the separation of graphite, either directly or by affecting the rate of cooling, naturally have an effect upon shrinkage. Sulphur, which drives carbon into combination with iron, therefore increases shrinkage unless neutralized by other elements. Manganese increases the total carbon and also the proportion of graphite if not present in amounts sufficient to provide an excess over that required for combination with the sulphur. It therefore decreases shrinkage except when excessive amounts are present. Phosphorus by lowering the rate of cooling tends to promote the separation of graphite and decrease the shrinkage. The separation of an iron phos- phide from the solid solution near the point of solidification also causes an expansion, as is the case when graphite separates. This circumstance, therefore, increases the effectiveness of phosphorus in decreasing shrinkage. * W. J, Ke6p, " Cast Iron," p. 48, .18 ,17 .16 — — — ^ ■ ~ -^ JUn '■e C.13 S.12 "^ ^ ■~ -.. ^ r.^ l[ a7-i ■ — ■ ^ -■ ■" L ^ *> -~ ^ "^ _, ^ ^ '" ■^ il.io £.09 B.OS ^ — _ h- L., ~ — ■-- iV, 1/ "^ ^ ■— i ^ H ■~. a .06 §.05 m.04 .03 .02 ^ ■* ~. ^ -~ -^ ^^ ■^ ^ ■- ^ — ~ ■ .91 1 1 .5 2 .0 2 .5 i! .0 3;i> PfirOont of Silicon Fig. 182.— Approximate Relationship Between Shrinkage, Silicon Content, and Size, 322 MATERIALS OF CONSTRUCTION 363. Checking. Liability upon the part of iron to check while cool- ing is dependent upon the magnitude of the stresses caused by the con- tracting of the metal upon'the sand, and upon the weakness of the metal at a temperature shghtly above a black heat (the temperature at which checking usually occurs). The factors which govern shrinkage, there- fore, determine the stresses to which the cooling metal will be subjected, if the casting be of such a shape as to compress the sand in shrinking. Sulphur is the most deleterious element affecting the strength of iron at a temperature just above a black heat. At this temperature both the iron sulphide and the magnesium sulphide are still in a pasty con- dition, thereby causing weakness of the metal. Phosphorus, by decreas- ing shrinkage, should decrease the liability of the metal to check, but this effect may be more than offset by the tendency of phosphorus to cause the metal to assume a coarsely crystalline structure. Manganese tends to counteract this last tendency upon the part of the phosphorus and therefore tends to prevent checking. 364. Segregation. Segregation in castings is the collecting together of impurities in spots. The primary cause of segregation is the effect • of impurities in lowering the freezing-point of iron. This results in forming a fluid solution which remains molten after the remainder of the metal has sohdified, and which runs to that, part of the casting which has the loosest texture. These spots, often called " hot spots," are apt to occur in the middle of the larger sections of the casting. They are often porous, and are usually extremely hard and brittle. The tendency to segregation is proportional to the amount of impuri- ties present. Phosphorus is especially apt to cause segregation, and manganese and sulphur have the same effect to a less marked extent. Segregation is not a commonly encountered diflSculty in iron founding, however, since other considerations will usually require a degree of freedom from excessive amounts of phosphorus, sulphur, or manganese which will minimize the danger of segregation. 365. Chilling. The intentional chilling of iron by the insertion of metal chills in a mold has been discussed above in Art. 350. The pro- duction of properly chilled iron is a very difficult problem, mainly on account of the effect of variations in composition of the iron upon the depth of chill obtained. The most important factors in this regard are the contents of silicon and of sulphur. If the siUcon is comparatively high and the sulphur very low chilling is practically impossible. If both silicon and sulphur are very low a considerable amount of chill is ob- tained, and the best results are obtained when a low percentage of silicon is combined with a rather high percentage of sulphur. Phosphorus CAST IRON 323 and manganese have little effect upon the depth of chill; the latter, how- ever, increases the hardness of the chilled iron. PHYSICAL AND MECHANICAL PROPERTIES 366. Hardness. The precise meaning of the term hardness as applied to metals is not altogether fixed. Properly, hardness should be con- sidered simply the measure of the resistance of the metal to being cut or scratched by a tool, or to being worn away by abrasion. Tenacity and brittleness are properties which are quite distinct from hardness, yet many methods employed for measuring hardness do not recognize any distinction. The principal factor in determining the hardness of cast iron is the amount of combined carbon. This is due, first, to the hardness of cementite itself, and second, to the fact that increase in combined car- bon usually means a decrease in graphite carbon, which is very soft. (Graphite has a further effect in increasing the ease with which cast iron may be worked, because it acts as a lubricant for the cutting tool.) The influence of elements other than carbon upon the hardness of iron is, with the exceptions of manganese, directly dependent upon their power to increase the amount of combined carbon. Silicon, therefore, acts as a pronounced softener, unless its percentage exceeds about 3 per cent, when the effect is reversed. Sulphur and phosphorus act as hardeners in all percentages, and manganese, in addition to its indirect effect due to the formation of combined carbon, has a direct harden- ing influence owing to the hardness of the compound (FeMn)3C. Tensile Strength 367. Tensile Strength in General. The tensile strength of cast iron is dependent upon so many variable factors that no general state- ment may be made concerning it. The founding methods, the design, and the size of castings are always important factors influencing strength. In addition, the composition and, more particularly, the constitution exert an enormous influence upon strength. The effect of details in founding methods, and the importance of design, etc., have been noted in a previous chapter. The effect of varia- tions in composition and constitution is really a question of the state assumed by the carbon. No other factor can be considered to be nearly as important as the relative proportion of combined and graphitic car- bon present, and the importance of elements other than carbon is largely in proportion to their power to either increase or decrease the proportion of combined carbon. 324 MATERIALS OF CONSTRUCTION 368. Influence of Form of Carbon. Gray cast iron has been repre- sented to consist really of plates of weak, soft, wholly non-metallic graph- ite, forming a more or less complete mesh-work which separates the ferrite and cementite matrix into partially isolated particles. The cementite and ferrite matrix of gray iron is really a low- or medium-car- bon steel, possessing great toughness combined with great softness, ductility, and low strength. White cast iron, upon the other hand, is a mixture of cementite and pearlite, not distinct from high-carbon steel except in the relatively greater impurity of the white cast iron, and the higher percentage of free cementite. The presence of graphite plates in any proportion must necessarily decrease the strength of the steel matrix and, therefore, the strength of the iron as a whole, because the graphite breaks up the continuity of the steel matrix. This injurious effect of graphite is not necessarily directly proportional to the percentage of graphite present but, rather, is proportional to the degree of continuity of the graphite mesh. Any diminution in the proportion of graphite will, in general, however, mean a diminution in some degree of the continuity of the graphite mesh and will, therefore, mean a lessening of its detrimental effect upon strength. At the same time, a diminution in the proportion of graphite means an increase in the proportion of combined carbon or cementite in the steel matrix. This increase in the proportion of cementite in the matrix means a proportionate increase in its strength until the percentage of cementife in the iron reaches a certain point, which is in the neighborhood of 1.2 per cent for an iron containing about 4.0 total carbon. Further increases in the percentage of cementite beyond this point increase hardness and brittleness at the expense of toughness, ductiUty, and strength. It is, therefore, evident that the strength of an iron will be greater when the percentage of combined carbon (cem- entite), does not exceed about 1.2 per cent than it will be if any higher percentage be present. Whether the highest strength is found with about 1.2 per cent cementite, or when the percentage of cementite is below this point will depend upon whether the loss of strength due to increase in graphite on the one hand, or the gain in strength due to the higher carburization of the steel matrix on the other hand, is the more influencial factor. This point cannot be settled on purely theoretical grounds, perhaps, but as a matter of common experience we know that the gain in strength of the steel matrix as the cementite content in- creases is really the controlling factor and, in consequence, the cast iron shows an increase in strength as the percentage of combined carbon increases up to about 1.2 per cent, after which further increases in com- bined carbon mean a loss in strength. CAST IRON 325 The whole substance of the above discussion is presented graphic- ally in the diagram of Fig. 183, which is abstracted from the original of Professor Henry M. Howe. In this diagram it has been assumed that the iron possesses a constant total carbon content (4 per cent), and an effort has been made to show graphically in an approximate degree, the effect of varying the proportion of combined carbon upon tensile strength. (Both the above discussion and Fig. 183 must be taken simply as indi- cating roughly the normal relationship of form of carbon to tensile strength of cast iron.) 369. Influence of Metalloids and Rate of Cooling upon Strength. The influence of the metalloids and the rate of cooling upon the strength of castings is, as above noted, largely an indirect one, dependent 100 U Graj tra jflron Gray Iron Close Sray Iron Mottled Iron White Iron^^^ "*■ '■« . "" ■ ^ " .. > C S |75 > ^ ^ Pi ^J f- !e It. S 3 s T't^ . i "/r. ■'i t^ rt . a s-" ''Vr ,, Sso V '^ 80, 1 V ((or fi>' tv .^ !? \ M "' 3 a / c tf ^ ^ ^ *• n'' f^ f =v ■k^ ^' 10 •- ^ J'e rl }c Q* •e i2 r >? , ^ ^ _J ,-- • rr^ - ' ^ ~ :; , 1^ 1^"°° — — ^ _ ■* .. '■' Cento ^ -' _ Combined CarboQ 0.5 Per Cent Graphite 3.5 1.0 3.0 1.5 2.5 2.0 2.0 2.5 1.5 3.0 1.0 3.5 0.5 4.0 0.0 Fig. 183. — ^Approximate Relationship Between Tensile Strength of Cast Iron and State of Carbon. (Howe.) upon the extent to which the separation of graphite is facilitated or retarded. Silicon in small amounts, by favoring the precipitation of graphite, exerts an influence which is beneficial to strength, provided that an excess- ive amount of combined carbon would otherwise be present. In this event the gain in strength of the matrix which accompanies the relief of brittleness more than compensates for the injurious effect of the increase of graphitic carbon. If, on the other hand, the additional graphite precipitation caused by silicon produces an iron whose matrix possesses too little combined carbon, the iron is weakened both because of the lowering of the strength of the matrix and because of the weaken- ing, and softening influence of the graphite. Large amounts of silicon (above about 3 per cent) have been observed to have exactly the reverse effect upon the form assumed by carbon, 326 MATERIALS OJ^ CONSTRUCTION the proportion of combined carbon instead of graphite carbon being in- creased. Therefore the effect of large amounts of siHcon upon strength is exactly the reverse of the effect of small amounts above noted. Aside from this indirect effect of silicon upon strength it possesses the power to increase strength directly, due to the formation of com- pounds which strengthen the matrix. When the silicon exceeds about 4 per cent this effect upon strength is reversed, owing to the increased brittleness of the iron, and with all percentages of silicon, this direct effect is of much less consequence than the indirect effect above stated. The proportion of silicon desirable for strong castings is closely depend- ent upon the size of the casting and the content of total carbon, sulphur, manganese, and phosphorus. In large or slowly cooled castings, a max- imum strength will be obtained when the silicon is low (only sufficient to offset the tendency of the sulphur and other metalloids to prevent the separation of graphite). In thin, quickly cooled castings, on the other hand, the silicon should be reasonably high, as should also be the case if the sulphur is high and not offset by manganese. The influence of sulphur is always as a weakener of cast iron, not only because it prevents the separation of graphite chemically and by hastening solidification, but also because it promotes the inclusion of flaws (blow- holes sand-holes, or shrinkage cracks), induces internal stresses, and causes coarse crystallization and brittleness. The baneful influence of sulphur may of course be more or less completely neutralized by much larger per- centages of silicon or by the presence of about twice as much manganese. Phosphorus usually tends to promote the formation of excess com- bined carbon and, therefore, to weaken .cast iron. When the sificon is high, however, a moderate amount of phosphorus may, by increasing the time of solidification, promote the separation of graphite, as above explained, thereby improving strength. The presence of more than about 0.05 per cent phosphorus will, however, always be detrunental to strength. The effect of manganese upon strength is always dependent upon the relative amounts of sulphur and manganese present. If the manganese content does not exceed twice the sulphur content, the manganese simply neutralizes the tendency of sulphur to decrease the proportion of graphite, and therefore the manganese increases strength. When the content of manganese exceeds the amount required to neutralize the sulphur, however, the excess manganese has a marked effect detrimental to strength because of the resultant excessive increase in the proportion of combined carbon. 370. Stress-strain Diagram for Cast Iron. It will be evident from the above discussion that cast irons must exhibit a great variability in elastic properties, since so many factors affect strength. CAST IRON 327 36,000 1* ^ — 32,000 fJVll ^ n »Sj c '3,28,000 , y >24,000 / / \v 3B , li / ■>i b^ "■ / N »> i> £1 / y- S 16,000 / u eO« ,^^ // k a /> '-' ■ ^ 8,000 f ^ -- i ^ ^ 4,000 f- /• ' / / f' .a 91 .0 02 .0 03 .004 In Fig. 184 typical stress-strain . diagrams for three radically dif- ferent cast irons are presented in order to illustrate their usual behavior under tensile stress. It will be observed that there is no well-defined elastic limit or yield point, but if we may consider the yield point to be the stress at which a marked increase in rate of deformation first appears, this point will be found to fall at about 60 per cent of the ultimate strength, except in the , case of the soft gray iron, when it falls at only 25 or 30 per cent of the ultimate strength. For the typical irons shown in Fig. 184 the ultimate strength falls at 35,500 pounds per square inch for the hard gray u-on, 22,500 pounds per square inch for the average gray iron, and 16,000 pounds per square inch for the soft gray iron. The yield points are found to fall at about 22,000 pounds per square inch, 14,000 pounds per square inch, and 5000 pounds per square inch, respectively. No constant proportionality of stress to strain exists for any consider- able load interval, and therefore we cannot properly say that cast iron has a modulus of elasticity. If we consider, however, that the term may be applied to the ratio of stress increment to corresponding strain increment for successive small load intervals, we find that the value of £?is initially about 30,000,000 pounds per square inch for hard cast iron, 24,000,000 pounds per square inch for average iron, and 14,000,000 pounds per square inch for soft iron. At 5000 pounds per square inch stress, the values of E have decreased to about 20,000,000 pounds per square inch, 15,000,000 pounds per square inch, and 7,000,000 pounds per square inch, respectively; and at 10,000 pounds per square inch stress, the values of E are about 15,000,000 pounds per square inch for hard iron and about 14,000,000 pounds per square inch for average iron. (The elastic limit of the soft iron has been exceeded at this stress.) The percentage elongation is small for all cast irons, rarely exceeding Straln.Inches per Inch Fig. 184. — Stress-strain Diagrams for Cast Irons. (Tension.) 328 MATERIALS OF CONSTRUCTION from 3 to 4 per cent for any grade, and the reduction of area is usually too slight to be appreciable. 371. Specification and Allowable Stress. The specifications of the American Society for Testing Materials recognize a distinction in elastic properties between light, medium, and heavy castings. Castings hav- ing any section less than § inch thick are classed as light castings; cast- ings having no section less than 2 inches thick are classed as heavy cast- ings; and castings not included in either of the above divisions are classed as medium castings. The minimum ultimate tensile strength of gray iron castings is not permitted to fall below: 18,000 pounds per square inch for light castings; 21,000 pounds per square inch for medium castings; 24,000 pounds per square inch for heavy castings. The factor of safety usually employed in the design of iron castings is about seven. The safe working stress of cast iron in tension is there- fore usually taken to be about 3000 pounds per square inch. Compressive Strength 372. Compressive Strength of Cast Iron. The compressive strength of cast iron, as is the case with all comparatively brittle materials, is largely a function of the shearing strength, since failure will inevitably occur along an obUque plane unless the specimen tested is sufficiently long to permit failure by lateral flexure. (In the latter case, stresses other than pure compression contribute to the failure.) The factors which control compressive strength are, however, exactly the same factors which control tensile strength. The most important consideration is, therefore, the state assumed by the carbon, and com- pressive strength will be benefited by all agencies which tend to increase the proportion of graphite carbon and decrease combined carbon, until the point of saturation of carbon in the steel matrix is reached (about 1.2 per cent combined carbon), beyond which point further increases in graphite mean a loss in strength of the matrix, and therefore a loss in strength upon the part of the iron as a whole. Compressive tests of cast iron show an enormously wide variation in strength values. If large specimens such as structural columns, etc., be tested, the ultimate strength will seldom be found to exceed 30,000 to 40,000 pounds per square inch. If, however, short specimens of small size be tested, the strength will be found to run to from 50,000 to 150,000 pounds per square inch. (This discrepancy is doubtless largely due to inevitable hidden defects in the large sections.) CAST lEON 329 Three typical stress-strain diagrams for compression of short blocks of cast iron are presented in Fig. 185. It will be observed that the yield point is much more clearly marked in compressive tests of cast iron than in tensile tests, although no absolutely constant ratio of stress to strain is maintained for any considerable load interval. The ultimate strengths of the three irons of Fig. 185 are 93,000 pounds per square inch, 63,000 pounds per square inch, and 44,000 pounds per square inch, for hard, average, and soft irons, respectively. The yield point falls at about 44,000 pounds per square inch for the hard iron, 30,000 pounds per square inch for the medium iron, and 20,000 pounds per square inch for the soft iron. The moduli of elasticity at 10,000 pounds per square inch are about 30,000,000 pounds per square inch, 20,000,- 000 pounds per square inch, and 12,000,000 pounds per square inch, for the three classes of iron, while at 20,000 pounds per square inch stress the value of E has dropped to about 25,000,000 pounds per square inch, 16,000,000 pounds per square inch, and 8,000,000 pounds per square inch, respectively. 373. Allowable Stress in Compression. The safe working stress usually assumed for cast iron in compression is about 16,000 pounds per square inch for short blocks. For columns or other structural mem- bers in compression this value may be reduced 50 per cent or more, depending upon the ratio of length to radius of gyration of the section. The factor of safety employed is, therefore, not more than about four in most cases. Cross-beeaking Strength 374. Cross-breaking Strength, Modulus of Rupture. The cross- breaking strength of cast iron is coming to be almost universally depended upon as the criterion by which the quality of the material going into gray iron castings is to be judged. Cross-breaking strength 90,000 - — ^ o 80,000 ^ s\ '' tij ■5, ?.*j S ^ -- ^ lO, t^ " S ^ t4 g 40,000 ^ / So Hj jro S. ^— — " " ' ^ '' 2 30,000 / "' / ^ §20,000 1 / J 10,000 ' .0 02 .0 04 Sf. )6 .0 )8 .0] .0 2 .0 4 .0 6 .0 8 .0 20 .022 .021 Straln.Inches per Inch Fig. 185. — Stress-strain Diagrams for Cast Irons. (Compression.) 330 MATERIALS OF CONSTRUCTION is closely allied to tensile strength, since bending stresses are a combina- tion of tensile stresses below the neutral axis of the beam with compres- sive stresses above the neutral axis. The extreme fiber stress on the ten- sion side of the beam is naturally the controlling factor, since cast iron is so much stronger in compression than in tension, and the load which produces failure is that load which causes an extreme fiber tensile stress equal to the ultimate tensile strength of the material. The cross-breaking strength is commonly expressed by use of the term " modulus of rupture," meaning the exti'eme fiber stress under the load which produces rupture as computed from the rule, wherein / is the extreme fiber stress, M is the bending moment, yi is the distance from the neutral axis to the extreme fiber, and / is the moment of inertia of the section. As a matter of mechanics, this procedure is a wholly imwarranted one, since it involves the assumption that the neutral axis remains a constant distance from the extreme fiber, and the further assumption that a constant proportionality of strain to stress obtains for all stresses up to the breaking stress, i.e., it is really assuming that the stress- strain curve for cast iron is a straight Une throughout. We have seen, however, that the proportion of strain to stress is not a constant for any range of stress, and it is a well-understood fact that the neutral axis shifts upward as the material deforms. Therefore the actual extreme fiber stress is far less than this value of / computed by the rule for the bending moment which exists under the load which produces rupture, and, indeed, the actual fiber stress is in the case of cast iron not more than 50 to 60 per cent of the modulus of rupture. For rectangular sections the above rule reduces to the form f=3Wl-^2bh^, the specimen being simply supported on a span I and broken by a central concentrated load W (6 is the width of the specimen and h is the height). The corresponding form of the rule for a circular section is - , f=Wl^nr^. The factors influencing cross-breaking strength are of course exactly the same factors as those discussed in their bearing upon tensile strength. CAST IRON 331 since failure under transverse loading is really failure by tension on the under side of the beam. In addition to these factors which have been discussed above, how- ever, it has been clearly shown that the transverse strength of cast iron is also dependent in some degree upon the size and shape of the speci- mens and the span upon which they are supported under load. Many series of tests, notably those of Mr. W. J. Keep,* have shown that specimens of small sections show relatively very much higher transverse strength than do specimens of large section, and are also much more sensitive to small variations in compositipn. Other tests have shown conclusively that rectangular specimens cast horizontally show" a much higher transverse strength than do round specimens cast vertically. In view of these circumstances efforts have been made by the Ameri- .can Society for Testing Materials to determine the size and shape of test specimen best fitted for adoption as a standard, and also the proper span upon which the beam should be supported. Their investigations have led to the widespread adoption of the " arbitration test bar " which is circular in section, Ij inches in diameter, and 15 inches long. The bar is broken under a central load upon a span of 12 inches. The propriety of maintaining the present standard 12- inch span in testing the arbitration bar has been seriously questioned. The point involved is well shown by the series of tests of arbitration bars made by Mr. C. D. Mathews f which are .represented by the diagrams of Fig. 186. In these diagrams breaking strength is expressed by the modulus of rupture. The average given represents 10 tests of the same iron except in the cases of the bars broken on a 12-inch span and those broken on a 24-inch span, in which cases nine tests were made. From the character of the results attained in these tests it would Beem desirable to increase the standard span for testing arbitration bars * W. J. Keep, loc. oit. p. 120. t Proc. Am. Soc. Test. Matrls., Vol. X, p. 304. S 58,000 g 51.000 S \ JS ".»»« \ Ru Pl ire s ^^/i k y^ \ (^ 4a!ooo £ 46,000 \ r ^> s ■-, J _■ \ ■^ ^ A lei f" eJIod ,s p lui tu 'e_ s __J gi 43^000 S ,^ ~v \ ^ — ' ■ K^ . tk up tu e \ "S 41,000 n 40,000 3 39,000 •O 38,000 «v ^ •v 'v -^ V ^ RELATION OF SPAN AND CROSS-BREAKING STRENGTH OF CAST IRON ^ 36,000 35.000 _ _ _i Mb W efl SL J _ IP 20 Span in Inches Fig. 186. • 332 MATERIALS OF CONSTRUCTION to 14 or 16 inches, either of which spans would seem to give more uniform results than are obtained with the present standard span or with longer spans than 16 inches. These tests are not sufficiently extensive to justify any change, however, unless their results are corroborated by further investigations. 375. Specificatioii and Allowable Cross-bending Stress. The Ameri- can Society for Testing Materials specifies that the minimum breaking strength of the arbitration bar under transverse load (the load being applied centrally on a span of 12 inches) shall not be under: 2500 pounds for light castings; 2900 pounds for medium castings; 3300 pounds for heavy castings. Also, " the deflection shall in no case be under 0.10 inch." These loads on a span of 12 inches correspond to the following values of the modulus of rupture: 39,000 pounds per square inch for light castings; 45,000 pounds per square inch for medium castings; 52,000 pounds per square inch for heavy castings. The specified minimum deflection of 0.10 inch is designed as a measure of the shock resistance of the material. It is a common practice to assume that the " modulus of shock-resistance " of cast iron is expressed by multiplying one-half the breaking load by the total deflection and dividing the product by the volume of the material of the specimen between the supports. The only error involved in this practice lies in the assumption that a constant ratio of load to deflection obtains up to the breaking load. This means that the load-deflection curve is assumed to be a straight line, so that the half-product of load and deflection will be the area of the load-deflection curve and therefore expresses the work done in deforming the beam. As a matter of fact, this curve is never "a straight Une, the ratio of load to deflection decreas- ing as the load increases. The modulus of shock resistance is, therefore, hke the modulus of rupture, simply an expression which will indicate approximately the relative behavior of different cast irons under trans- verse load, and is not a concrete expression of actual strength or resist- ance to shock. CHAPTER XIII MALLEABLE CAST IRON.* GENERAL 376. Definition of Malleable Cast Iron. Malleable cast iron is iron of special composition which, after having been cast in its final form, is rendered malleable by a process of annealing. It is essential that the iron used be a white iron before annealing, in order that the carbon may be almost wholly in the combined form. The armealing process will then result in the conversion of the combined carbon into free carbon in an amorphous condition, not resembling free carbon in the crystalline form as graphite. This amorphous carbon will exist as isolated particles in a continuous mesh of metal. The casting is through this circumstance rendered very much tougher than white or gray cast iron, and Its ductility and malleability are increased to such an extent that it may be bent or twisted to a considerable degree even when cold. 377. Malleable Cast Iron as a Material of Engineering Construc- tion. Malleable cast iron has no important applications as a purely structural material, but is largely used in the manufacture of machinery, iinplements, railway rolling stock, light hardware, etc. It combines the advantages of ordinary cast iron with respect to the ease with which complicated forms may be cast, with a degree of tough- ness, ductility, and strength approaching that of steel. Only cast steel or steel forgings can compete with malleable iron for the class of uses above noted, and where either material might be used the malleable iron will usually possess an advantage in cost. MANUFACTURE OF MALLEABLE CASTINGS 378. The Materials Used. The charge of the furnace of a malleable- iron foundry includes pig iron, sprues, annealed malleable-iron scrap, * Practically the only satisfactory modem treatments of this subject are the works of Richard Moldenke, " The Production of Malleable Castings," 1911, and " Malleable Castings," the latter published in three parts by the International Correspondence Schools. ^ 333 334 MATEEIALS OF CONSTRUCTION and steel scrap. Wrought-iron scrap, cast-iron scrap, and ferro-silicon are also used in some cases. The pig iron should contain not more than 0.60 per cent manganese, not more than 0.225 per cent phosphorus, and not more than 0.05 per cent sulphur; total carbon need not be specified. The silicon require- ment varies according to the castings made. Heavy castings require from 0.75 to 1.50 per cent silicon, while light castings require from 1.26 to 2.00 per cent. Several irons of different silicon contents are usually- kept on hand. The sprues or " hard scrap " include the gates and scrap castings that have not been annealed. Thorough cleaning of the sprues in tumbling barrels to remove the burnt sand is very necessary. The pro- portion of sprues in the charge may run from 25 per cent to as much as 60 per cent. Malleable scrap is difficult to melt because of the comparative infusi- bility of the skin of malleable castings. It contributes greatly to the strength of the castings made, however, and if the large scrap is broken up before charging, it may be handled without serious difficulty. The proportion of malleable scrap used depends upon the proportion of sprues, but should not generally exceed 20 per cent. Steel scrap of any sort, providing only that it be not too heavy, may be used as a part of the charge with beneficial results, if added after the balance of the bath is molten. (Otherwise, it would become oxidized.) The addition of steel scrap affords a method of reducing total carbon and produces stronger castings. Steel scrap is not used for cupola melt- ing, and is never allowed to amount to more than 10 per cent with no malleable scrap present. Wrought-iron scrap is sometimes substituted for steel scrap or malle- able scrap, a much lower proportion being used. About one-fifth as much wrought iron may be substituted for steel, or one-twentieth as much for malleable scrap. Cast-iron scrap is only used incidentally in place of so much pig iron. The maximum allowance does not exceed 5 per cent. Ferro-silicon is used only in emergencies to save a heat which has been improperly managed or burned, or, by additions in the ladle, to render iron intended for heavy castings suitable for light, castings. 379. The Furnace. Three types of furnaces are principally used m melting iron for malleable casting: the cupola, the air furnace, and the open-hearth furnace. The cupola furnace has been fully described in Art. 331 and the air furnace in Art. 332. The open-hearth fiXrnace used in the malleable-iron foundry differs in no essential respect from the open- hearth steel furnace, hereinafter described, except in size, the capacity CAST IRON 335 of the open-hearth furnace used in the malleable-iron industry being usually from 10 to 20 tons only. 380. Melting Malleable-iron Mixtures. The cupola process for melting iron for malleable-iron castings differs in no respect from ordi- nary gray-iron foundry practice, described in Arts. 334 and 339, except in the higher proportion of fuel charged. The fuel charge amounts to about one-fourth the weight of the metal charge when the iron is of the composition required for malleable castings. The advantages of the cupola process lie in the cheapness of installation and operation, the com- parative ease with which the furnace is controlled, and the small loss of silicon in melting. The disadvantages are the extreme liability of burn- ing, owing to the direct contact of metal and fuel, and the extremely close structure of the hard castings produced, which causes trouble in annealing. Very little malleable scrap, and no steel, can be used in the cupola, because of the great danger of burning. Cupola iron requires an annealing temperature 100 to 150° C. higher than does iron melted in the air furnace or open-hearth. The operation of the air furnace has been fully discussed in Arts. 340 to 342. The advantages of the air furnace as compared with the cupola are principally the better grade of castings produced, the wider range of scrap material used, the shorter time required for pouring, the less serious consequences of a breakdown, and the better control over process and product. The disadvantages of the air-furnace process are the greater expense of equipment, the greater skill required in operation, and the longer time required in melting. A few years ago the air furnace was rapidly replacing the cupola in malleable-iron foundries, and it still remains the most important type of furnace used in this country. At the present time, however, the general preference is for the open-hearth furnace. The open-hearth furnace is operated in the malleable-iron foundry in almost exactly the same manner as in the production of open-hearth steel, described in a later chapter. Its advantages over the air furnace are the saving of time (about one hour) required for melting, the very exact control of the process, and the resultant high efficiency and gain in the percentage of first-grade castings. The disadvantages are the high cost of installation, the heavy repair bill, the necessity of having gas fuel — meaning the installation of a gas producer if natural gas is not available — and the necessity of continuous operation. 381. Molding Methods for Malleable Castings. Molds for malleable castings are made in exactly the manner above described for gray-iron castings made in green sand. Particular care must be exercised to provide proper gating in handling white-iron mixtures, and risers or feeders must be provided where thin 336 MATERIALS OF CONSTRUCTION sections are encountered to prevent cooling of the metal at these points before the mold is completely filled, resulting, because of the excessive shrinkage, in the production of spongy spots. Chills are very commonly used in molds for malleable castings, par- ticularly for the sake of cooUng the larger parts of castings rapidly, thereby preventing the possibility of graphite separating out, as it tends to do with slow cooling. 382. Pouring the Castings. Molten white iron is a very different material from molten gray iron. The former must be poured very hot, and as rapidly as possible, to insure proper complete filling of the mold. White iron at a scintillating white heat is not as hot as gray iron at a bright red heat, the melting-point of white iron being more than 100° C. below that of gray iron. Iron which has chilled slightly, or which has been burned in the furnace, will be sluggish, and must not be poured except in pig-beds for use in subsequent heats. 383. Subsequent Treatment of the Castings. Hard castings for malleable iron are cleaned by any of the methods common to the gray- iron foundry. Tumbling is very commonly resorted to, and the sand blast, or pickling in acid solutions are not infrequently used. ' Very careful inspection of the cleaned castings is necessary, and all defective castings are rejected before being annealed. The gates are also chipped off and the castings are separated into a number of classes for annealing. I 384. The Annealing Process. The annealing process consists in heating the castings to a red heat and maintaining them at that tem- perature for a sufficient time to change practically all of the carbon from the combined form to the free amorphous form called " temper " car- bon. In order to prevent oxidation at this temperature, and also warping of the castings, the latter must be packed in annealing pots surrounded . by a proper packing material. This packing material might be sand, clay, or other inert material and the heat alone would effect the desired change in the state of the carbon and produce malleable castings. Higher grade and stronger ca,stings are produced, however, when the packing material is a decarbon- izing agent such as iron oxide. This results in the migration of carbon from the outer shell of the casting, producing a layer resembling steel about ^ inch thick, encased in a skin of almost carbonless iron on the surface. This skin may subsequently be enriched in carbon by a case- hardening process, and if the reduction of carbon has previously been carried to the maximum depth possible (about I inch), the resultant material will greatly resemble cast steel. It may even be hardened and tempered. MALLEABLE CAST IRON 337 The annealing pots are usually tl;ree or four cast-iron boxes, without bottoms, stacked one above another, and placed on a " stool " which supports them above the floor of the furnace. The usual size of the annealing pots does not exceed 18 by 24 inches, 15 inches high, but pots of specially large size or of special form are sometimes required. The packing material used is commonly the cinder or slag squeezed from the wrought-iron puddle balls. This slag is a very rich iron oxide high in silica. Hematite ore in a pulverized state is also used as a pack- ing material, especially for cupola iron whose annealing temperature is sufficient to cause puddle cinder to cake together and fuse onto the castings. Rolling-mill scale is also used, and magnetite has recently been found very satisfactory. If sand has not been perfectly removed from the castings, it causes trouble by combining with the oxide of the packing material, increasing its fusibility and causing the castings and packing to bake together. The annealing oven or furnace is built in many quite different forms, the principal factors being the type of fuel used and the mechanical equipment of the plant. In general, the oven usually utilizes the down- draft principle and, since the castings must cool slowly in the furnace itself, it cannot be continuous in opera- tion, and does not employ the regenera- tive principle. Fig. 187 illustrates one common type of an- nealing oven employ- ing natural-gas fuel. The flames are de- flected downward from a vaulted roof, pass between the pots, and escape through floor openings to flues beneath leading to the stack. A damper in the stack flue is depended upon to control the draft. Large ovens are pro- vided with gas burners at either end. Producer gas may be used in exactly the same way except that, owing to its lower calorific value, it must be burned in greater quantities. When coal fuel is used the gas burners are replaced by grates, a firing door, and an ash pit. Ovens are usually built in batteries of from six to twelve, adjacent ovens having common side walls. When overhead cranes are used for charging and discharging, the roof arches may be built in such a way that they may be removed in sections by the charging crane. Fig. 187. — Annealing Oven for Malleable Cast Iron. 338 MATERIALS OF CONSTRUCTION Two standard annealing processes are in use, depending upon the class of malleable produced. " Black-heart " malleable is usually pro- duced in American practice, while " white-heart " malleable is usually produced in European foundries. The latter process requires a higher annealing temperature, which is maintained for a much longer time. Most annealing ovens will attain the required temperature of 700° to 750° C. (800° to 900° C. for cupola iron) in from 24 to 36 hours; some ovens, however, require from 48 to 60 hours in heating up. The anneal- ing temperature is maintained for a period- of about 60 hours in the case of black-heart castings, and 100 or even 120 hours for white-heart cast- ings. In the black-heart process but little of the temper carbon is removed, as is shown by the fracture, wherein only a very narrow band of gray material encloses the velvety black interior. The white-heart process is seldom applied to metal more than J inch thick, and the removal of carbon is carried to the ultimate degree possible, the decarburization being effected to a depth of as much as yb or even j inch. The cooling of the castings after annealing should be as slow as pos- sible. Usually from fifty to sixty hours are permitted before the pots are withdrawn from the oven. The entire annealing process there- fore requires, as a rule, about six days. This period may be shortened to about four days, but not with best results. 385. Treatment of Annealed Castings. The aimealed castings must be cleaned to remove the scale which has formed. This is accomplished in tumbling barrels, the only abrasive used being small pieces of annealed scrap. For delicate castings, or when a polish is desired, a part of the scrap may be replaced by wooden blocks, leather, etc. In order to test the quality of the annealed iron, " test-plugs," which are simply small projections, about f by | by 1 inch long, are cast on the more important work. These are broken off and the fracture examined. If normal, the fracture should have a black velvety surface in the interior, surrounded by a band of dark gray about ^ inch thick, and this in turn is encased in a band of white not more than ^ inch thick. Too low silicon increases the width of this band. PROPERTIES AND USES OF MALLEABLE CASTINGS 386. Chemical Composition and Constitution. The proper com- position of good malleable iron is indicated by the following analysis: Silicon 0.45-1.00% Manganese about 0.30 Phosphorus not more than 0.225 Sulphur. not more than 0.06 Total carbon (before annealing) at least 2 . 75 MALLEABLE CAST' IRON 339 The constitution of malleable iron is extremely variable, even in a single casting, because the effect of the anneaUng process is largely- dependent upon the thickness of the casting. The outermost skin is practically carbonless iron, and will therefore consist largely of ferrite together with a very small proportion of impurities which were origi- nally present in the metal. The intermediate gray portion of black- heart malleable consists largely of ferrite, but contains scattered par- ticles of free carbon in the amorphous state called temper carbon. The black interior consists of ferrite in which many isolated particles of tem- per carbon are interspersed. White-heart malleable castings have the constitution of the intermediate portion of black-heart castings for the most part, but the outermost band of practically pure ferrite is much thicker in the case of the former than in the case of the latter. 387. Physical Properties. Tensile Strength and Ductility. The tensile strength of malleable castings is much more important than the tensile strength of ordinary gray-iron castings, and the tensile test is a better indication of the quality of malleable iron, because the duc- tility of the material allows it to be evenly gripped in the jaws of the testing machine without danger of introduction of cross-bending stresses. Tensile tests of malleable iron are commonly made on 1-inch square bars, 14 inches long, especially cast without the use of chills. Elonga- tion is measured over a 2-inch gauged length, the fracture being included in this length. The average strength of black-heart malleable castings is usually not less than 40,000 pounds per square inch and sometimes exceeds 50,000 pounds per square inch. The elongation in 2 inches is seldom less than 21 per cent, and may amount to as much as 7 per cent. 388. Transverse Strength. The transverse test of malleable iron is important because it is an indication of ductility and toughness as well as transverse strength. The transverse strength is determined by sup- porting a 1-inch square bar on a span of 12 inches and applying a cen- tral concentrated load. Observations of deflection under the breaking load are made. The usual breaking load for such specimens is not less than 3000 pounds, with a deflection of | inch and for specially well-made malleable castings the breaking load may amount to as much as 5000 pounds, with a deflection of IJ inthes. (The above loads correspond to moduh of rupture of 54,000 pounds per square inch, and 90,000 pounds per square inch, respectively.) 389. Toughness or Shock Resistance. Impact tests, unfortunately, have not been standardized for malleable cast iron, and no quantitative data concerning its shock resistance are available. An approximate measure of its toughness is obtained, however, from the observed deflec- tion in the transverse test, the so-called modulus of shock resistance 340 MATERIALS OF CONSTRUCTION or resilience being computed (as in the case of tests of gray-iron arbitra- tion bars) by dividing one-half the product of breaking load and total deflection by the volume of the specimen between the supports. The very excellent performance of malleable iron under shock is well understood, however, and it has long been considered to be espe- cially adapted to use where oft-repeated light shocks are encountered. For this class of service it is often considered superior to cast steel. 390. Uses of Malleable Castings. The uses for which malleable castings are especially adapted are very numerous. As above noted, this material is especially useful in the manufacture of that large class of articles whose form is too complicated for economical forging, but which must possess a strength and toughness not attainable in gray castings. Among the more common applications of malleable iron may be especially mentioned its use on railroad work. Couplers are commonly made of malleable iron, as are the journal boxes, brake fittings, and many other small fittings for rolling stock. Other uses include many agricul- tural implements and parts of agricultural machinery, all manner of pipe fittings, elbows, unions, valves, etc., and all kinds of household and harness hardware such as parts of locks, hinges, window and door fit- tings, buckles, swivels, etc. Another class of articles are made of malleable iron which has been case-hardened after prolonged annealing. The material then closely resembles cast steel, and is often sold as such. This class of articles includes many carpenter tools, such as hammers, hatchets, chisels, planer irons, etc., also pistol parts, skates, shears, etc. CHAPTER XIV WROUGHT IRON GENERAL 391. Historical. The history of the development of the art of the iron maker, and the use of iron in the making of implements, machines, and structures, is almost a chronicle of the advance of civilization. Since the eariiest ages iron has been the one metal of greatest value to man, and when or where the first discoveries of its virtues were made are facts which are shrouded in the mists of antiquity. Antiquarians have found evidence which seems to show that the Egyptians, whose civilization is the oldest of which we have any accurate knowledge, possessed " bars of wrought metal, and vessels of copper, and of bronze, and of iron " * seventeen centuries before the beginning of the Christian era. Iron was known to the Carthaginians probably one thousand years before Christ, and was also employed by the Chal- deans, the Babylonians, and the Assyrians, who were contemporaries of the early Egyptians. Old . Testament History is replete with references to the use of iron in implements of both peace and war. Herodotus, writing in the fifth century before Christ, speaks of " the Chalybians, a people of iron workers," and also mentions the use of both iron and steel by the Persians, the Medes, and the Parthians. Irons and steels are known to have been made in India probably a thousand years before Christ, and the famous swords of the city of Damascus, the " Damascus blades " of song and story, were made ages ago from Indian and Persian steel. The primitive methods of derivation of iron from ores were all direct methods, as has been above explained, and the irons produced in the char- coal fire and the Catalan forge and other types of ore refineries developed at a somewhat later date, were for the most part essentially what we call " wrought iron " to-day. Steel was similarly made at almost as early a period, but only under especially favorable conditions obtaining in certain districts,, and its production was never so common among * From inscription found at Karnak. 341 342 MATERIALS OF CONSTRUCTION ancient peoples as was the production of wrought iron. (Primitive methods of iron refinement roughly resembling the modern crucible steel process are believed to have been used in the production of the an- cient Indian and Persian steels.) Charcoal bloomeries and Catalan forges, not extensively modified or improved upon, were the only source of iron for many centuries. Furnaces which were able to render molten iron from the ores, by utili- zation of tall vertical shafts and stronger blast pressures (the first true blast furnaces), were developed in Belgium during the fourteenth century, and the first method of production of wrought iron from pig iron, in what was essentially a puddling furnace, was developed in England 400 years later, mainly by Henry Cort, whose patent for producing iron by the puddling process was obtained in 1784. From the time of Cort the development of the wrought-iron industry was very rapid until the invention of the Bessemer steel process in 1855. It was confidently predicted at that time that the availability of cheap Bessemer steel would speedily put an end to the production of wrought iron, by crowding it from every field which had theretofore been pecu- liarly its own. This prediction has not been fulfilled, however, although wrought iron has been crowded from many fields, and has fallen far behind steels in the importance of its industry and in rank among materials of engineering construction. 392. Definition of Wrought Iron. The term " wrought iron " is commonly applied to that commercial form of iron which is obtained by the refining of a mixture of pig iron and scrap iron at a temperature not sufiicient to maintain the metal in a molten state after the removal of its impurities, but only in pasty condition, the iron being intermixed with a considerable amount of the slag formed in the process. The Committee on Uniform Nomenclature of Iron and Steel of the International Association for Testing Materials defines wrought iron as "malleable iron which is aggregated from pasty particles without subsequent fusion, and containing so little carbon that it does not harden usefully when cooled suddenly." English writers often employ the term " malleable iron " meaning " wrought iron " as above defined. 393. Wrought Iron as a Material of Engineering Construction. Sixty years ago wrought iron was the most important metallic material for general structural and- construction purposes. It was rolled in all manner of shapes, and was used for frames of buildings, for bridges, ships, tanks, and structures of all kinds. Tools and implements were made of wrought iron whenever a hardened edge or surface was not required. Its uses in machine construction were legion, its strength, WROUGHT IRON 343 toughness, ductility, and forgeability making it available for many pur- poses which could not be served by cast iron, but which did not justify the use of the relatively expensive steel then available. The introduction and rapid decrease in cost of Bessemer and open- hearth steel in the latter part of the nineteenth century gradually forced the abandonment of the use of wrought iron as a structural material for the frames of buildings, bridges, etc., and various steels are now preferred for many special construction purposes, for most tools, imple- ments, machine parts, etc. Nevertheless, wrought iron still possesses an important place among materials of construction, and bids fair to continue so to do indefinitely. Its principal present uses are as a mate- rial for general forging operations, particularly where welding is involved, as rolled rods and bars, as wire, as welded pipe, and as a metal for roofs and sides of buildings and for tanks, etc. THE MANUFACTURE OF WROUGHT IRON -The Wet'Puddling Process 394. The Puddling Process in General. The usual process of man- ufacture of wrought iron consists in the melting of the pig iron in the hearth of a reverberatory furnace which is lined with iron oxides, result- ing in the elimination of most of the carbon, silicon, magnanese, phos- phorus, and sulphur present in the charge, by oxidation. The metal becomes pasty toward the end of the process, owing to the decreased fusibility of the purer iron, and is removed as a plastic ball from which the slag must be removed as completely as possible by squeezing or ham- mering. The resultant " puddled bloom " is rolled into large bars called " muck bars." The bars are cut into short lengths, piled up in bundles which are wired together, heated to a white heat, and rolled down to a smaller size called " merchant bars." This process results in the production of the purest of the common iron products, if we disregard the presence of from 1 to 3 per cent of slag, which the rolling process has caused to assume the form of greatly elongated particles in the direction of rolling. This circumstance accounts for the characteristic fibrous structure of wrought iron, and the purity of the metal accounts for its remarkable ductility and weldability. 395. The Iron Used. The pig iron commonly used for puddling is a rather inferior grade called " forge-pig." Its composition is from 1.00 to 1.50 per cent siHcon, 0.25 to 1.25 per cent manganese, not more than 1.00 per cent phosphorus, and not more than 0.10 per cent sulphur. Comparatively high silicon is desired in order to provide sufficient slag to cover the bath of metal and prevent excessive oxidation of the iron. 344 MATERIALS OF CONSTRUCTION in Manganese need not be carefully watched, because it is largely removed ■~ the process. Phosphorus, and particularly sulphur, are not com- pletely removed, and must therefore be kept fairly low, although their injurious effect upon the quality of the iron is by no means as marked as in the case of steel. The weight of the furnace charge varies according to the size and type of furnace used. Ordinary single furnaces handle from 200 to 600 pounds of pig, while large furnaces, or double furnaces built in pairs without a dividing wall, may handle as much as-1500 pounds per charge. Charging is usually done by hand, the pig iron being thrown Fia. 188.— 500-lb. Puddling Furnace. ^^ through the firing door. (Macfarlane.) 396. The Puddling Furnace. The puddling furnace is a rectangular masonry structure lined with firebrick and tied together by steel tie-rods and iron plates and buckstays. Fig. 188 illustrates a common type of puddling furnace of 500 pounds' capacity, worked from one side, and Fig. 189 Ulustrates a 1500-pound furnace with two work- doors. The furnace is of the reverberatory type, the heat depend- ed upon to bring the working chamber to the desired tempera^ ture being largely that which is reflected from the sloping firebrick roof of the melting chamber, the fuel being burned in a separate chamber out of contact with the metal. Iron castings support the working bed of the furnace at a conveni- hoIlow'faittoT' *J' ^uZ' *':'"' P""'""^ '^' ''"' '''''''^'^- °f ^^ beneath. A klo^ tHf .^■"'':''"'lse" separates the firebox from the working chamber. It 18 protected from the flames sweeping across into the working chamber by a covering Floor Line Fig. 189.— 1500-lb. Puddling Furnace. WROUGHT IRON 345 of refractory bricks or blocks, and is cooled by air driven through its hollow interior. The " flue-bridge " over which the waste gases pass from the working chatnber to the stack is of similar construction. It is customary to arrange two or four furnaces to be worked by one stack, the flues from the individual furnaces being provided with dampers. Furnaces are not infrequently built in pairs back to back. The area of the fire grate is large, usually about one-third of the area of the working part. The grate bars are usually of wrought iron and are easily replaceable. A steam injector is sometimes provided to supply extra air for the combustion of the fuel and to assist in the oxidation of impurities- in the metal. The fuel is a bituminous coal which burns with a long flame. It is introduced through a " firing hole " in the front wall. This hole, not being provided with a door, is usually partially stopped by lumps of coal. The large " working door " in the front wall of the working chamber has a heavy iron projecting sill and is normally closed by an iron-bound slab of brickwork which is suspended in place and balanced by a counter- weight. A small opening in the lower edge of this slab affords an opportunity for the puddler to insert his " rabbling iron " to stir or " rabble " the charge. The slag is tapped off when necessary through a tap-hole provided below the working door. 397. Preparation of Furnace for Charging. The Fettling. The furnace hearth is lined, or " fettled," with strong iron oxides of basic character. The fettling requires extensive renewals after practically every melt. The principal " fettling " materials used are: Basic slag from reheat- ing furnaces; slag from the puddling furnace itself, usually roasted, but sometimes allowed to remain in the furnace from the previous melt; hammer scale, or roll scale, from the finishing mill, and hematite ore. Sufficient fettling is usually employed to cover the iron plates to a depth of about 5 inches. The character of the fettling is variable, depending upon the character of the pig iron used. Easily fusible iron, i.e., very impure iron, requires a fusible fettling, and comparatively pure iron, which is therefore less fusible, requires a fettling which is less easily fused and less rich in oxides. 398. Furnace Operation. Chemical and Physical Changes. The furnace having been charged and a melting temperature attained, the subsequent process may be considered in four stages which merge into one another. (1) The " Melting-down Stage." The pig iron gradually becomes red hot and is turned about by the puddler to insure uniform heating. The more fusible slag begins to melt within twenty minutes, and in the course of thirty to thirty-five minutes the pig iron will have become completely melted down. During this stage the oxidation of the metal- loids will begin. The first to be oxidized is the silicon, followed by the manganese, and later by the phosphorus and sulphur. The oxides natur- ally leave the metal and join the slag in accordance with the fundamental metallurgical principles above stated. During this stage most of the 346 MATERIALS OF CONSTRUCTION silicon and manganese are tiius eliminated, together with a small propor- tion of the phosphorus and a very small proportion of sulphur. (2) The " Clearing Stage." During this stage, which occupies only from seven to ten minutes, it is usually necessary to add ore or mill scale, thus making the slag still more basic, and to close the dampers to cool the furnace sufficiently so that the carbon will not be oxidized before the phosphorus and sulphur have been disposed of. Very vigorous rabbluig is necessary to promote oxidation by intimate contact between the pig iron and the fettling. During this stage the removal of silicon and manganese is almost completed, so far as it ever will be, and a consider- able further amount of phosphorus and sulphur is eliminated. (3) The " Boiling Stage." This stage is principally characterized by the removal of carbon, through the agency of the fettling first, and later by the oxygen of the air. The ferric oxide is reduced by the carbon with the formation of carbon monoxide gas which bubbles to the sur- face, causing the boiling appearance characteristic of this period. It is especially essential that the slag be very strongly basic at this time, as it will otherwise be unable to retain the oxides of phosphorus and sul- phur which the carbon monoxide might so easily reduce. The carbon monoxide is burned by the oxygen of the air to form carbon dioxide, resulting in the appearance of the light yellow flames at the surface of the bath called " puddler's candles." The expansion of the bath during the boil causes its level in the hearth to rise greatly and a considerable amount of slag escapes through the open slag hole. During the boil the puddler continues to rabble the mass vigorously, and after about twenty minutes the metal begins to " come to nature," i.e., non-fluid iron begins to collect in patches in the bath and on the surface in a pasty condition, owing to its lesser fusibility in its now nearly carbonless condition. Great care is required upon the part of the puddler to prevent this pasty iron becoming chilled by sticking to the hearth, or oxidized by exposure above the slag. At the end of from twenty to thirty minutes all of the metal will have come to nature and the removal of carbon will have reached its limit. A considerable further quantity of phosphorus and sulphur will have been eliminated during this stage (20 to 30 per cent of the phosphorus and sulphur is never removed.) (4) The " Balling Stage." When all the metal has come to nature the balling stage begins. The puddler carefully gathers all of the pasty metal into one mass, which he subsequently subdivides into portions of such size as he is able to withdraw from the furnace. Each portion is worked into a ball and welded together as completely as possible. The balls are rolled up under the protection of the fire-bridge to prevent excessive oxidation of the iron before being withdrawn. The balls are WEOUGHT IRON 347 100' S50 540 Mn_ — ■ — ■ ^ ^ "sT / / / / / ^ ^ / 5^ / x^- ^ / / / /s , / y / / / ^^ ^ y f- ^ 10 20 30 40 80 60 70 80 90 100 Time (Minutes) Fig. 190. — Sequence of Removal of Metalloids in Puddling. (Stoughton.) finally gripped with tongs, one by one, and drawn out of the door over the fore-plate to a conveyance which takes them to the squeezers as quickly as possible. The usual weight of these " puddle balls " is about 100 pounds, but is sometimes as much as 200 pounds. The balling stage requires about twenty minutes altogether. It is characterized by no chemical changes, but the furnace tempera- ture must be main- tained as high as pos- sible in order that the puddle balls may be hot enough so that the slag is still very fluid when they reach the squeezer. Fig. 190 shows graphi- cally the approximate sequence of the removals of the metalloids during the operation of puddling. 399. Removal of Slag. Squeezing or Shingling. The puddle ball when removed from the furnace is a very loosely agglomerated mass of pasty iron, honeycombed with pockets of slag. This slag must be excluded, so far as is possible by mechanical means, and the iron compacted and welded together by one of the operations known as " squeezing," and " shingling," respectively. In American practice some form of squeezer is practically invariably used. A very common type of squeezer is that shown in Fig. 191. A wheel mounted on a vertical axis revolves within an encircling cylinder of cycloidal form. Fig, l91.-Rotary Squeezer, ^oth wheel and cylinder have corrugated surfaces. The puddle ball is introduced at the point where the space between the two surfaces is a maximum, and the rotation of the inner wheel causes the ball to roll as it is carried around the annular space which is constantly decreasing in width. When, the ball is finally ejected a large part of the slag will have been excluded 348 MATEEIALS OF CONSTRUCTION and it will have been compacted by the kneading and squeezing to less than half its original diameter. In European countries the puddle ball ifi usually " shingled," i.e., it i;= forged down by some type of power hammer. The steam hammer is now usually employed in shingling. Between blows the puddle ball is constantly turned by the operator until it has been thoroughly welded together and the slag largely excluded. The compression of the porous mass of metal either in squeezing or shingling results in a considerable rise in temperature of the mass, which circumstance favors the expulsion of the slag by retaining it in a very fluid condition. 400. Rolling Mill Operations. The puddled "blooms" from the squeezer or the shingling mill are immediately transferred to the rolling mill wherein the finished bar or shape is produced. The bloom is first passed through a bar mill which reduces it to rectangular bars called " muck bars " which are 2 to 4 inches thick. These muck bars are then cut into strips which are piled up (Fig. 192), tied with wire, reheated to a welding heat, and again rolled down to form " merchant bars." A further quantity of slag is squeezed out in rolling. Usually the bars in a pile are all laid the same way, but occasionally they are " cross-piled," i.e., alternate layers are laid crosswise. This practice results in the production of a cross-network of fibers, instead of having all the fibers running lengthwise of the bar. Bars which have been piled and rerolled are commonly called mer- chant bar or " single-refined iron "; when subjected to a second piling, heating, and rerolling, " double-refined bar " is produced. The effect of repeated rerolling is principally the further elongation of the strands of slag in the direction of rolling, thereby rendering the iron still more fibrous, in its structure. No advantage is gained by piling and rolling more than about three times. The final rolling is usually done in a mill called a " merchant bar mill," Fig. 193. A series of passes through the rolls serves to gradually reduce the bar to the desired size and shape. Roughing rolls are shown at B and D, Fig. 193, while finishing rolls for ovals and rounds are shown at F and H. 401. Mechanical Puddling. Efforts to avoid the very severe manual labor involved in hand puddling, by the use of devices which mechanically rabble the charge, have characterized the history of the Fig. 192. Method of Piling Muck Bars. WROUGHT IRON 349 industry for many years. None of these efforts have met with marked success, however. The best-known mechanical furnace is the Roe furnace, shown in Fig. 194. It is suspended upon trunnions and caused to oscillate 65 degrees Fig. 193.— Merchant Bar Mill. in each direction. Oil fuel is used, the oil and air for combustion being introduced through the hollow trunnions. A stack is provided at either end. The oscillations of the furnace keep the bath and slag well mixed, thus avoiding hand rabbling. The entire charge of the furnace, weigh- ing about 4000 pounds, is discharged in one ball from which the slag is squeezed in a large hydraulic squeezer of special design. 402. Wrought Iron from Scrap. A large proportion of the wrought iron made in this country is made, not by refining pig iron, but by heating and rolHng scrap wrought iron. Usually this scrap is simply bundled together and wired in a pile roughly resembUng a pile of muck bars, heated to a welding heat, and rolled. A second method of utilizing scrap is called " busheling scrap." Scrap iron of small size is gathered together, heated in a small furnace resembling a puddling furnace, and the product treated as an ordinary puddle ball. Fig. 194. — Roe Mechanical Puddling Furnace. 350 MATERIALS OF CONSTRUCTION A third method is called "fagoting," or "box piling." A rough box is made of muck bars which form the sides, bottom, and top, while the interior is filled with miscellaneous small scrap. This mass is wired together and handled just as a pile of muck bars would be handled. All of the methods of making wrought iron from scrap without re- melting result in the production of an inferior grade of material. PROPERTIES AND USES OF WROUGHT IRON 403. Composition and Constitution. The composition of wrought iron approaches pure iron more closely than any other commercial form of iron. The usual impurities — carbon, silicon, phosphorus, sulphur, and manganese — are always present in small amounts, however, in addi- tion to the slag which is invariably present. The following analyses are examples of the composition first, of an ordinary grade of wrought iron, and second, of a very pure Swedish charcoal iron: Element Total carbon Silicon Phosphorus Sulphur Manganese Slag and rare elements Iron (by difference) . . . Ordinary Wrought iron Swedish Iron % % 0.100 0.050 0.200 0.015 0.150 0.055 0.030 0.007 0.010 0.006 2.800 0.610 96.710 99.257 The constitution of wrought iron is quite simple as compared with that of cast iron, because of the very low percentages of carbon and other impurities in the iron. The great bulk of the material is nearly pure ferrite (contaminated with small amounts of silicon, phosphorus, etc.), which consists structurally of crystalline grains of iron. The appear- ance of a longitudinal section of wrought iron under high magnification is shown in the photo-micrograph of Fig. 195. The presence of slag, appearing as many irregular black Unes of varying thickness, is clearly evident. The crystalline nature of the ferrite is also plainly to be seen, thus refuting the belief which was long held that the peculiar properties of wrought iron were due to the fact that the minute structure of the iron was fibrous rather than crystalline as in steel. The micro-photograph of Fig. 196 shows the appearance of the transverse section of wrought iron. The structure is in every way similar to that revealed by the longitudinal section except that the slag here appears as irregular dark areas corresponding to the cross-section of the slag fibers. WROUGHT IRON 351 ^^^*^>1U^ Fig. 195. — Wrought Iron. Longitudinal Section. Macnified 100 Diameters. (Boynton.) A third constituent besides the ferrite and the slag is invariably recognized, owing to the fact that carbon is always present in some degree. This carbon will normally combine with the iron to form cement- ite, which latter unites with a definite portion of the ferrite to form pearl- ite. Since the carbon content is normally very low, the pearlite con- stituent is not conspic- uous, but occurs only in isolated small patches between the grains of ferrite. The constitution of the slag itself is rather complex, being made up of siUcates and phosphates of iron and manganese, formed by the com- bination of the acid and basic oxides which have joined the slag during the puddling process (the Fe203, FeO, MnO, Si02, P2O5, etc.). 4&t. Classes of Wrought Iron. Wrought irons may be classed according to method of manu- facture, or according to the class of uses for which they are intended. The first classi- fication comprises the follow- ing: (1) Charcoal iron is that made by refining either the ordinary blast furnace product or the product of blast fur- naces worked with charcoal fuel in a charcoal hearth. Knobbled charcoal iron is simply ' iron made in a special type of charcoal hearth after melting Fia. 196.-Wrought Iron. Transverse Section, in a coke refinery. Charcoal Magnification not Stated. (GuiUet.) irons are the purest grades of 352 MATERIALS OF CONSTRUCTION wrought irons and are much used for conversion into cementation steel and crucible steel by a process of recarburization. (2) Puddled iron is that made by the wet puddling process above described. Newly puddled iron is sometimes called muck-bar iron, or puddled-bloom iron, to distinguish it from scrap puddled-iron. Box- piled iron may be made entirely from puddled iron or may be made from muck-bar iron and scrap. (3) Busheled scrap is a heterogeneous product made by heating and rolUng busheled or fagoted scrap by the processes above described. The principal classes of wrought iron according to the uses for which their qualities fit them are staybolt iron, engine-bolt iron, refined-bar iron, and wrought-4ron plate. Staybolt iron is made from puddled or knobbled charcoal iron. It is the highest grade of wrought iron and, while not the strongest, it is the toughest and most ductile iron, the best for forging and welding. Engine-bolt iron is made from the same class of material as stay- bolt iron, which it slightly surpasses in strength. It is slightly less tough and ductile, however. Refined-bar iron is made from a mixture of muck-bar iron and iron scrap. It is inferior to stay-bolt and engine-bolt irons in strength, tough- ness, ductility and forgeability. Wrought-iron plate is made in two grades or classes: Class A is made wholly from puddled iron, and is a strong hard iron, but lacks ductility and toughness when compared with the best grades of iron. Class B is made from a mixture of puddled iron and scrap material, and is inferior to class A iron plate in every respect. Neither class of wrought-iron plate is intended for forging or welding. 405. Tensile Strength and Elongation. The tensile strength of a given wrought iron is dependent upon the direction of stress with respect to the " grain " of the iron. This is naturally to be expected, since, as we have seen, the continuity of the metal in a direction transverse to the direction of rolling is interrupted by numerous strands of slag which are structurally very weak. The tensile strength of wrought iron in a transverse direction has usually been found to be between 0.6 and 0.9 of the strength in a longitudinal direction. It will therefore be safe to assume, in the rare instances when we are concerned with the strength in a transverse direction, that this strength is about three-fourths the strength in the direction of rolling. When the muck bars have been "cross-piled" the strength in a transverse direction may practically equal the strength in the longitudinal direction. The tensile properties of wrought iron in general are quite variable; but, when we recognize the several grades above listed, the properties WEOUGHT IRON 353 of a given grade become quite definitely fixed. The specifications of the American Society for Testing Materials adopted in 1912 and 1913. pre- scribe the following tensile properties: Staybolt Iron. Engine- bolt Iron. Refined Bar Iron. Wrought-iron Plate. Property. 6-24 Ins. Wide. 24-90 ins. Wide. A B A B Tensile strength, pounds per squafl:e inc . . k Yield Doint 49-53,000 °«CsTr:) 30 48 50-54,000 25 40 48,000 25,000 22 49,000 26,000 16 48,000 26,000 14 48,000 26,000 12 47,000 26,000 Per cent elongation in 8 ins. Per cent reduction of area . 10 The " Modulus of Elasticity " of wrought iron varies little for all the different grades. It will usually be found to be between 26,500,000 pounds per square inch and 28,000,000 pounds per square inch. Atypical stress-strain curve for high-grade wrought iron is shown in Fig. 197. 406. Relationship be- tween Tensile Properties of Wrought Iron and Reduction in Rolling. A fairly definite relation- ship exists between the tensile properties of wrought iron and the amount of reduction in rolling. It is often as- serted that the strength of wrought iron is inversely proportional to the cross-sectional area. It is probable, however, that this is true only in so far as the smaller sizes represent a greater percentage reduction in roUing from the original pile of muck bars. Extensive tests made by the U. S. Board on Testing Iron and Steel in 1881 show that practically the same tensile properties are shown by all sizes of wrought- iron rods provided only that the ratio of finished size of bar to size of pile be kept constant. 15 000 - - - - "^ I. - = :3 = i. ^ ^ ^ /• / IT / =t — ■" ■ u /• & / C Y el il 'oi nt m" ^ '""' / } ^ EU St c LI mit / - h X5 000 / TjapicalStress-strala Curve for Wrought Iron (Curye 11 is Portion A-B of I Enlarged) - Oi / / 5 000 / > / 1 1 1 1 00 02 .0004 xm 00 08 .0010 iooja oo;i 00|6 0(^8 0020 0o!!2l00 ei. .02 .04 .06 .08 .10 .12 .14 .10 .18 Strain Inches per Inch Fig. 197. — Typical Stress-strain Curve for High Grade Wrought Iron. 354 MATERIALS OF CONSTEUCTION Fig. 198 has been plotted from the data obtained in the series of tests above mentioned. It will be noted that the relationship between ten- sile properties and reduction in rolling is by no means constant, but both ultimate tensile strength and elastic lunit are raised considerably by decreasing the ratio of section of finished bar to section of pile. The ■ effect is much more marked with respect to the elastic limit than with respect to the ultimate strength. 5900 5700 5500 5300 5100 4900 4700 d "4500 a g4300 P. ^4100 a 3000 ^3700 3500 3300 3100 2900 2700 2500 2300 9 \ V O "o^ s c o^ i,,,,,,^^ c •"•Qp o lUItl mate Ten silcJ trei ^h o \__^ . "^~~ — ° .1 \ o y I \^ O \ o ^Relationship Between Tensile Properties of Wrought Iron and Reduction In Roiling (Watertown Arsenal Testa, 1881) o \ \, x Elas tic Limit 1 o ■^ o — o 6 7 8 Per Cent^ 9 10 11 la 13 14 15 16^ Vrea of Bar Area of Elle Fig. 198. — Relation Between Tensile Properties of Wrought Iron and Reduction in Rolling. 407. Effect of Previous Straining or Cold Working upon Tensile Properties. The effect of previous straining of wrought iron upon the elastic limit and ultimate strength revealed by subsequent test is greatly to raise the elastic limit and considerably increase the ultimate strength. The magnitude of the change effected will be dependent upon the extent to which the previous straining was carried beyond the elastic limit of the material in its original condition, and the maximum effect WEOFGHT lEON 355 ! 30,000 20,000 10,000 will be noted when the specimen is actually broken in the initial test and the subsequent test made upon the portion of the specimen immediately- adjacent to the fracture. The effect of previous straining upon the tensile properties of wrought iron as revealed by a subsequent test is shown forcibly by Fig. 199, wherein I is the stress-strain curve obtained in the original test of a wrought-iron bar,* and II the stress-strain curve derived in a subsequent test of the portion of the original bar adjoining the fracture. The elastic Umit of the bar upon retest has been increased from 30,000 pounds per square inch to 59,000 pounds per square inch, while the ultimate strength has been raised from 53,700 pounds per square inch to 64,000 pounds per square inch. The elongation after fracture has been reduced from 16 per cent in a 100-inch gauge length to 5.3 per cent in a 50-inch gauge length. (These elongations appear small because the gauge length is unusuallylarge.) The modulus of elastic- ity is not changed in any degree by the previous straining. Curve III, Fig. 199 shows the effect of annealing upon uhe tensile properties of the previously strained bar. It will be noted that the effect of the former strain has not only been removed, but the elastic limit and ultimate strength have even been lowered below the values noted for the original bar. The explanation of this phenomenon will appear in the discussion which follows. Cold working of wrought iron, i.e., deforming it by rolling, hammering, or pressing, at temperatures below about 680° C, affects the structure and the mechanical properties of iron in much the same way as straining it beyond the elastic limit in a test does. The elastic limit is considerably raised, the ultimate strength is sUghtly raised, and the elongation or ductility is usually lowered. . The beneficial effect of cold working is largely due to the fact that it * Watertown Arsenal Tests. 1882. r TT ^ — ~' _ _ _ \ _ _ _ _ _ _. _ - - - — - =: "■ -= ^ :::; zA / II „ ^ / ^ "^ N ' ^ / Effect of Previous Straining of Wrauffht Iron _j 0.01 0.02 0.03 0.01 0.05 0.06 0.07 0.08 0.09 0.10 0.11 0.12 Steain Inches per Inch Fig. 199. — ^Effect of Previous Straining and Annealing upon Tensile Properties of Wrought Iron. (Wat. Ars. Tests, 1882.) 356 MATEEIALS OF CONSTRUCTION closes up the grain of the metal, and eliminates blow-holes. It also in- creases the cohesion and adhesion of the crystals, however, and if cold work follows hot work the size of the crystals will be decreased with consequent lessening of brittleness. Of the several methods of cold working above mentioned, rolling is the most rapid and the cheapest, although it does not work the metal as well as hammering does. Pressing works the metal to a greater depth than either hammering or rolling and is especially necessary when the work is large. Hammering is commonly accomplished by the steam hammer, while pressing is done by hydraulic presses of from 500 to 1000 tons capacity. 408. Heat Treatment and Crystalline Structure. The size of the crystalline grains of ferrite in wrought iron is dependent — first, upon the temperature from which it cools; second, upon the length of time it is maintained at that temperature; third, upon the rate of cooling and the amount of mechanical work to which it is subjected during cool- ing; and fourth, upon the temperature at which working is discontinued. In general, the higher the temperature, the larger the crystals will be, providing the time allowed at that temperature be not too short. Slow cooling also promotes crystalline growth, provided that the cool- ing be undisturbed. The application of pressure, however, either by rolls, hammer, or press, retards the formation and growth of crystals and breaks up or destroys pre-existing crystals. Thorough working to a temperature below the critical temperature above mentioned (about 690° C), is necessary to overcome the injurious effects of the coarse crystallization which occurs at such high temperatures as are required, for instance, in welding. 409. Compressive Strength of Wrought Iron. The properties shown by wrought iron in compression do not differ materially from its tensile properties, i.e., its elastic Umit, ultimate strength, and modulus of elas- ticity are about the same in compression as in tension, provided that the ratio of length to radius of gyration does not approach the point where lateral flexure occurs. In the latter event the flow of the metal when the yield point is reached immediately causes eccentricity of stress, and ulti- mate failure ensues under an average stress which exceeds the yield point by very little or none at all. The compressive strength of wrought iron may therefore be taken to be from 45,000 to 60,000 pounds per square inch if the length is very short in proportion to the radius of gyration. Usually, however, this proportion is too great to make it possible to disregard flexure, and the ultimate compressive strength must be taken to be only equal to the stress at the yield point, or from 25,000 pounds per square inch to 35,000 WROUGHT IRON 357 pounds per square inch, according to the character and condition of the iron. All of the considerations above discussed with reference to the modi- fication of tensile properties by variations in the amount of reduction in rolling, heat treatment, previous straining, etc., apply with equal force when we are concerned with compressive strength, except that we must not expect that previous straining in tension will improve the compressive properties revealed on subsequent tests. We find in fact that exactly the reverse is the case and that straining the material beyond its elastic limit in either tension or compression impairs its elastic properties under the opposite kind of stress. 410. Shearing Strength of Wrought Iron. Remembering what has been said above concerning the structure of wrought iron, it will at once be evident that the resistance of the material to shearing stresses will be far less on a plane parallel to the direction of the " grain " than on a plane which cuts the fiber of the iron transversely. It is a fact, indeed, that the shearing strength on a transverse plane is often twice that shown on a longitudinal plane. The actual shearing strength shown by tests is very variable, but in general will be from 20,000 to 35,000 pounds per square inch on a longi- tudinal plane and from 30,000 to 45,000 pounds per square inch on a transverse plane. 411. The Welding of Wrought Iron. One of the most valuable attributes of wrought iron is the comparative ease with which it may be welded, i.e., the joining together of two pieces by simply pressing or hammering them together while at a very high temperature. Wrought iron possesses the property of weldability to a greater degree than any other metal except an equally pure low-carbon steel. Its superiority is due largely to its comparative purity, since all impurities, especially carbon, silicon, and sulphur, reduce weldability in a marked degree. The exact temperature of welding is not known, but it is moderately close to the melting-point, probably between 1300° and 1400° C At this temperature (a white heat) the metal becomes plastic, almost pasty, and remains so for a considerable range of temperature. In welding operations, however, two factors are operative which tend to prevent a perfect union, and the production of a joint as strong as the original metal. The first of these is the formation of iron oxide by the oxygen of the air; the second is the coarse crystaUization which is always apt to occur at such high temperatures. (See Art. 408.) Iron oxide forms very rapidly on iron at a welding heat, forming a coating of liquid slag which effectually prevents a proper union of the surfaces which are to be joined. The remedy lies in the removal of the 358 MATERIALS OF CONSTRUCTION iron oxide from the joint. This may be effected in large measure by simply making the two surfaces convex to each other. They then come in contact first at their centers, and as the area of contact spreads the slag is squeezed out. The exclusion of the slag is facihtated by the use of a flux which is a solvent for the slag and therefore renders it more liquid. Borax is commonly used for this purpose, and of late years " welding plates " (thin plates of flux which are inserted in the joint just before closing it) are sometimes used. Ordinary sand is a satisfactory flux in welding steel. The effect of the rapid growth of the crystals at a welding heat is to produce a coarse-grained iron which lacks ductility and possesses great brittleness. The remedy lies in the refining of the metal by working it under the hammer or in a press until the critical range of temperatures has been passed. The danger due to coarse crystallization usually is not that the metal will be brittle just at the weld, for this portion will probably have been thoroughly refined by the working necessary to properly close the joint, but rather that the metal some few inches from the joint will have been overheated without subsequent heat refinement under the hammer. Such a joint may upon testing fail at a point some- what removed from the weld itself, but this does not prove that the weld is stronger than the original bar. The usual practice in welding consists in first " upsetting " the bars (enlarging the section, to provide extra metal for subsequent reduction in working, by endwise pressure or hammering of the two pieces) while at a forging heat, raising them to a white heat in a forge or otherwise, placing the pieces together on the anvil or block, adding the flux, and quickly closing the joint by hammering or by use of a press. Working is continued down to a red heat to effect a proper refinement of the crystals. It is desirable after cooling to reheat the parts to a point just above the critical temperature in order to restore the grain size. The commonest types of welds are the " lap " or " scarf " weld, Fig. 200a; the " butt " or " jump " weld. Fig. 2006; and the " V-weld," Fig. 200c. The proper shape of the individual pieces for these welds are shown by Figs. 200rf, 200e, and 200/, respectively. Electric welding is an art quite distinct from ordinary hand welding. Here less difficulty is experienced in avoiding the injurious effect of oxidation, since no strongly oxidizing air blast is used, as in the case of the ordinary forge, and the pieces may be heated in direct contact with each other. The temperature required may be attained electrically either by direct contact with— or by radiation from— an external electric arc; by making the joint itself the negative pole of an electric arc, the post- WROUGHT lEON 359 tive pole of which is a carbon rod; or by making the current traverse the pieces themselves, the increased resistance at the joint giving rise to the necessary welding temperature. Owing to the possibility of great variations in the extent to which the above-mentioned factors may be operative, the strength of welded Fig. 200a. FlQ. 200(1. Lap or Scarf Weld. 1 Fig. 200!>. Fig. 200e. Butt or Jump Weld. .'-' Fig. 200c. Fig. 200/. [ Tongue or V Weld. Fig. 200.— Various Types of Welds. joints is extremely variable. Hand forging is usually less efficient than power forging and electric heating is more efficient than forge heating. • In general the strength of hand-forged joints of average character is about 60 to 70 per cent of the strength of the bar. Power forging will raise the efficiency to from 70 to 90 per cent, and even higher efficiencies are sometimes attained by use of electric methods CHAPTER XV STEEL GENERAL 412. Definition. It is not possible to make a strict definition of steel which is concise and at the same time not in conflict with current usage. The term is applied to practically all the forms of iron pro- duced either in the cementation furnace, the crucible, the converter, the open-hearth furnace, the electric furnace, a combination of the con- verter and the open-hearth, or a combination of either the converter or the open-hearth with the electric furnace. Still there are excep- tions to the above statement which caimot be disregarded, among the most important of which is the use of the open-hearth in the production of malleable cast iron. In the early days of the iron industry it was easy to distinguish between steel and cast iron by declaring all forms of iron which were malleable to be either steel or wrought iron; steel if it would harden upon sudden cooling, otherwise wrought iron. These distinctions will not hold at the present time, however, since many grades of steel are now made which will not harden upon sudden cooUng; some are not malleable except through a certain range of very high temperatures; and one grade of iron, malleable cast iron, is malleable (after annealing) but cannot be classed as steel. Professor H. M. Howe has proposed the following definition, which, although cumbersome, covers the ground adequately: " Steel is that form of iron which is malleable at least in some one range of temperature, and in addition is either (a) cast into an initially malleable mass; or (6) is capable of hardening greatly by sudden cooling; or (c) is both so cast and so capable of hardening." The provision in the definition that steel " is malleable at least in some one range of temperature " distinguishes steel from cast iron and pig iron without excluding certain special steels like chrome and manganese steel, which are malleable only through a short range of high temper- atures; the provision that it is " cast into an initially malleable mass " 360 STEEL 361 excludes malleable cast iron, which is rendered malleable by special treatment after being cast; and either the provision that it is " cast," or the provision that it "is capable of hardening greatly by sudden cooling " serves to differentiate it from wrought iron (which is never cast and is never capable of hardening), without excluding " cementa- tion " or " blister " steel, which is not cast but will harden, or many low- or medium-carbon steels which are cast but will not harden. The Committee on Uniform Nomenclature of Iron and Steel of the International Association for Testing Materials * has proposed the following definition of steel: " Iron which is cast from the molten state into a mass which is usefully malleable initially at least in some one range of temperature." The definition excludes bUster steel, which is of plastic instead of fluid origin. The committee therefore proposed the following supplementary definition of steel of plastic origin: " Iron which is aggregated from pasty particles without subsequent fusion; is malleable at least in some one range of temperature; and contains enough carbon (say 0.30 per cent or more) to harden usefully on rapid cooling from above its critical range." 413. Classifications of Steels. Steels are classified according to method of manufacture as: BUster or cementation steel, made by the carburization of wrought iron at a bright red heat in the cementation furnace. Crucible steel, made by the carburization of wrought iron in a molten state in crucibles. Bessemer steel, made in the Bessemer converter by blowing finely divided air currents through molten pig iron, thereby removing the impurities in the latter. Open-hearth steel, made by subjecting pig iron and steel scrap to the oxidizing flame of gas and air burned in a reverberatory furnace. Electric steel, usually made by refining steel partially purified in the converter or open-hearth in the hearth of a furnace heated by electric induction, electric resistance, or the electric arc. (Some electric furnaces produce steel from pig iron and scrap, and some produce steel direct from the ore), and Duplex steel, made by completing in the open-hearth the refinement of steel partially purified in the Bessemer converter. Steels are classified according to carbon content as: Soft, mild, or low-carbon steel, containing from 0.08 to 0.10 per cent carbon in the case of Bessemer steel, and 0.08 to 0.22 per cent carbon in the case of open-hearth steel. * Proceedings Sixth Congress of International Association for Testing Materials, 1912. 362 MATERIALS OF CONSTRUCTION • Medium or medium-carbon steel, containing from 0.18 to 0.30 per cent carbon, and Hard or high-carbon steel, containing from 0.30 to about 1.5 per cent carbon. Steels are also classified according to the uses for which their proper- ties fit them as: Boiler-rivet steel, softest, most ductile, least strength. StructuralMvet steel, very soft, very ductile, low strength. Boiler-plate steel, soft, ductile, medium strength. Structural steel, medium hardness, ductility, and strength. Machinery steel, fairly hard, less ductile, high strength. Rail steel, hard, low ductility, high strength. Spring steel, rather soft, low ductiUty, very high strength, and Tool steel, extreme hardness, very low ductiUty and extremely high strength. 414. Steel as a Material of Engineering Construction. The place of steel among construction materials has been indicated above in the consideration of cast iron, malleable iron, and wrought iron. From the time of the development of cheap processes of steel production following Sir Henry Bessemer's invention of 1855, steel has held the foremost position among metallic materials of construction. In the form of rolled sections and shapes it is almost exclusively used as a metallic structural material for the construction of frames of buildings, bridges, and all manner of structures above the ground. No other material comes in competition with rolled steel for the construc- tion of railroad rails; as a material of machine construction, either as rolled or as cast steel, it has many applications which cannot be as advantageously served by either cast iron, malleable iron, or wrought iron; as rolled plate it is almost exclusively used for the shells of boilers, tanks, etc., for the hulls of vessels, and for the web portion of girders; as sheet steel it is used to a large extent as a cheap form of roof sheathing and both exterior and interior wall covering for buildings; special steels are exclusively used in the construction of ordnance and armor-plate; and in the form of either high-carbon steel or alloy steel it is almost abso- lutely without a competitor as a material for cutting tools or for springs. ^ The uses of steel in construction generally are so multitudinous that it is altogether impossible to indicate even in a general way its important applications. Its pre-eminence among iron products is indicated, how- ever, by -the fact that about 75 per cent of the total tonnage of iron prod- ucts of the United States is steel of one form or other. STEEL 363 the manufacture of steels General 415. Steel-making Processes. The making of steel was in the early days a process of conversion of iron ore into steel by direct methods whose general nature has been above indicated. Indirect methods, by the carburization of wrought iron or by the refining of pig iron, were developed at a much later date, but have almost wholly superseded the original direct methods. Two methods of steel-making by the carburization of wrought iron have been developed and are still used: (1) the cementation process, which produces blister or cementation steel by the carburization of wrought iron without fusion, and (2) the crucihle process, which produces crucible steel by the carburization of wrought iron in a fused condition. The two principal methods of steel-making by the refining of pig iron (with or without the admixture of iron and steel scrap) are (1) the Bessemer process, which produces Bessemer steel by blowing finely divided air currents through molten pig iron contained in a retort-shaped fur- nace called a " converter," the impurities being oxidized and thus removed in the slag, carbon being subsequently added; and (2) the open-hearth process, which produces open-hearth steel by subjecting pig iron and scrap to the oxidizing flame of gas and air burned in a rever- beratory regenerative furnace, carbon being restored after the removal of the oxides in the slag. Other processes of making steel from pig iron or pig iron and scrap are usually combinations of the Bessemer process and the open-hearth process, called duplex^ proceszes, or combination of electric furnace methods with either the Bessemer or the open-hearth process. The present relative importance of the various steel processes is indicated by the fact that about 33 per cent of the steel production of the United States in 1912 was Bessemer steel, over 66 per cent was open- hearth steel, less than 0.4 per cent was crucible steel, and about 0.06 per cent was electric steel. (Blister steel is not made to any extent outside of England, and duplex steel has been included above with open- hearth steel.) Carburization of Wrought Iron 416. The Cementation Process. The principle which underlies the operation of steel-making by the cementation process is that iron at a bright red heat will absorb carbon by an action which appears to be a 364 MATERIALS OF CONSTRUCTION traveling of solid carbon into solid iron, thereby forming a solid solu- tion of iron and carbon. The cementation furnace (Fig. 201), as used at Sheffield, England, consists of a brick or stone structure of rectangular plan, 15 to 20 feet long and 12 to 15 feet wide, provided with an arched roof and chimneys on either side. Manholes are provided in each end wall for charging and discharging. The entire furnace is enclosed within an outer masonry wall which con- tinues upward as a conical hood or stack to a height of 20 to 30 feet above the furnace. The in- termediate space between the furnace walls and the stack is filled with earth or sand, open passageways being provided oppo- site the furnace doors. Within the furnace are located two " convert- ing pots " which are 2 5 to 4 feet wide and high, 8 to 15 feet long, built of stone, and supported on short piers above the furnace floor to allow free access of the heat to all sides of the pots. A fire is maintained in a firebox which extends the length of the furnace between the converting pots and beneath the level of the floor. A space provided beneath the grates serves as an ash pit. The bars of wrought iron which are used for the cementation process in England (very little cementation steel is made outside of England) are usually very pure Swedish iron made by the charcoal-hearth process. They are usually 2| to 3 inches wide, ^ to | inch thick, and of a length corresponding to the length of the cementation pots. The operation of the process is begun by lining the bottom of each pot with small particles of charcoal upon which alternate layers of iron bars and charcoal are placed until the pots are filled. A space is left between the bars in each layer so that each bar is completely surrounded with charcoal. The top of the pots is now luted tight with a wet mixture called " wheel-swarf," composed largely of material derived from the wear of the grindstones in the cutlery mill. This material permits the escape of gases at first, but later becomes air-tight. The total charge of a furnace amounts to from 8 to 13 tons of iron. The fires are now started and the required bright red heat is attained in from two to four days. The progress of the operation is tested from Fig. 201. — Cementation Furnace. STEEL 365 time to time by withdrawing and examining the fracture of trial bars which have been placed so that they project through small apertures called " moles," specially provided for the purpose in the ends of the pots. The required temperature (650° to 700° C), having been attained, the carbon begins to soak into the iron at a rate of about |-inch per 24 hours. The time required for the completion of the process depends upon the grade of steel produced. Mild heats require the maintaining of the maximum temperature for a period of from 7 to 8 days, medium heats require about 9^ days, and high-carbon heats require about 11 days. Since the carburization proceeds from the exterior of the bars inward, the carbon content decreases progressively toward the center, and an unaltered core will be found in very mild bars. This is illus- trated by the following cases: (a) f-inch bar, per cent C at outside = 0.98, per cent C at center = 0.10, ave. per cent C = 0.45; (b) J-inch bar, per cent C at outside = 1.50, per cent C at center =1.15, ave. per cent C=1.33. The different grades of cementation steel made at Sheffield are classi- fied as follows:* No. 1. Spring heat 0.50% C. No.' 2. Country heat 0.63 No. 3. Single-shear heat 75 No. 4. Double-shear heat. . .' 1 .00 No. 5. Steel-through heat 1 • 25 No. 6. Melting heat 1 . 50 When the carburization has proceeded to the desired point the fire is withdrawn, and the furnace is allowed to cool for about a week before the bars are removed. The presence of some slag in the original wrought-iron bars is re- sponsible for the appearance of bUsters on the surface of the bars, which have been formed by the evolution of carbon-monoxide gas when the carbon combined with the ferrous oxide of the slag. The presence of these blisters accounts for the use of the term " blister-steel," which is often appUed to steel made by the cementation process. * Bradley Stoughton, " The Metallurgy of Iron and Steel," p. 70. 366 MATERIALS OF CONSTEUCTION The Cktjcible Process 417. General. Although the steel produced in very ancient times by the Chaldeans, Egyptians, or other early civilizations was probably made by methods similar to the crucible process, the actual invention of the modern crucible process was not made till 1740, when the efforts of Benjamin Huntsman of Sheffield, England, to devise a steel process superior to the cementation process, were finally crowned with success. For over one hundred years thereafter the crucible process remained the principal and almost the only method of steel-making, the cementa- tion process being relegated to the unimportant position it has always since held, principally as an intermediate stage in the production of the finest cutlery steel, its product being remelted in crucibles. The cost of crucible steel is very high, however, and steel was unable to supplant wrought iron as a structural material until after the inven- tion of the Bessemer process in 1855, and the open-hearth process in 1861. In spite of the fact that Bessemer steel and open-hearth steel may be produced at a much lower cost than crucible steel, the latter still occupies a distinctive field from which the cheaper processes seem unlikely to crowd it. Its use is now restricted mainly to the making of high-grade tools and cast-steel machine parts, where first cost is less important than intrinsic quality. For these purposes it is superior to the best open- hearth steel, even though the two steels may show exactly the same com- position. The crucible process consists essentially in the melting of wrought iron in closed crucibles of refractory material, the carburizer being placed in the crucible with the iron, together with any special alloying element desired. The details of the process vary according to the type of furnace and fuel used: for the most part the gas-fired regenerative furnace is used, but in England the earlier type of coke-fired furnace is employed to a considerable extent. The two types of furnace will be briefly considered separately. 418. The Coke-Furnace or Melting-hole. The English type of coke-furnace or melting-hole is shown by Fig. 202. In the steel-melting house the melting-holes are ranged along the sides of the building below the floor level. Each hole is built of masonry to hold two crucibles. The crucibles are set directly in the fuel on short stands of refractory material which are supported by the grates. The top of each chamber is covered by a slab of firebrick set in an iron frame. The melting temperature is maintained by controlling the chimney draught, a chimney opening being provided for the purpose below the grates. The fuel used is a hard-burned coke which completely surrounds both crucibles. The fuel must be replenished several times during the operation of melting. STEEL 367 The crucibles used in the melting-hole are of clay, having the general shape and size shown in Fig. 203. The making of clay crucibles must be carefully done, and, on account of the danger of cracking or excessive distortion due to shrinkage stresses, they are subjected to a long course of slow drying on shelves provided on the walls of the melting- house above the coke-holes. After drying, the crucibles are burned and carefully annealed, and are then heated up in the melting-hole before being charged. The maximum life of a clay crucible does not exceed 3 heats, after which the danger of breakage is very great, and fre- quently a crucible must be rejected after only one heat. 419. The Gas-fired Regenerative Furnace. The gas-fired regenerative crucible-steel furnace is simply an adaptation of the Siemens type of furnace which is used in open-hearth steel making. The general features of the furnace are shown by Fig. 204. The gas-fired furnace contains anywhere from two to twenty melting-holes, each large enough to hold either four or six crucibles. Six or eight inches of coke dust is placed in the bottom of each melting-hole, in the center of which a hole is provided so that, if a crucible breaks, the steel may be taken up by the coke, or will escape through the hole to the vault below, where it may be cleaned up. The melting-holes are con- nected on either side by means of " ports " and vertical flues or " up-takes " with chambers called ■' regenerators," placed on a level below the melting-holes. The regenerators are filled with a loose checker-work of brick and are connected below with flues which lead to reversible valves,, whereby the passage may be made to lead to the stack upon the one hand or to the gas main or the outside air upon the other hand. Four regenerators are required for each furnace, two upon each side. The two smaller regenerators nearest the center line of the furnace are alternately used to preheat the incoming gas, while the two larger outer regenerative chaTibers alternately preheat the air required for com- bustion. While the two regenerators on one side, of the furnace are preheating the incoming gas and air the gaseous products of combustion are passing out through the ports and vertical flues on the opposite side to the second pair of regenerators, to whose brickwork they give up a large part of their heat before passing on to the stack. When the gas and air valves are reversed, the current of in- coming air and gas is caused to pass through the chambers which have just been heated by the burned gases, and by timing the reversals of direction of the current of gases to come about once every fifteen or twenty minutes, Fig. 202.^English Crucible Steel Melting-hole. (Macfarlane.) A . Crucible being gently dried. B. Shelf and Support. C. Stack. D. Cover of melting hole. E. Handle. F. Furnace. G. Crucible lid. H. Crucible. /. Flue. J. Stand. K. Fire-grates. L. Grate bearer. M. Stack damper. Fig. 203.— Clay Crucible. 368 MATEEIALS OF CONSTRUCTION bJ transverse section emm mma mesea esmsi the temperature of the incoming gases may be maintained at a fairly uniform degree. This utilization of the heat of the burned gases to preheat the incoming gases explains the use of the term " regenerative " in speaking of the Siemens furnace. The gas used may be either natural gas or producer gas, the thermal value of the former greatly exceeding that of the latter. The arrangement of the ports must always be such that the gas enters below the air, mixing therewith as it rises in the ports. The type of crucible used in the gas-fired furnace is not made entirely of fireclay, but is a mixture of about 50 per cent graphite, 40 per cent or more clay, and the balance silica sand. The graphite crucible costs more than clay, but it lasts longer and is stronger, thus per- mitting larger crucibles to be. used. (The capacity is usually nearly twice that of a clay crucible). The graphite crucible is made in practically the same manner as a clay crucible and must be burned and annealed with equal care. It is tougher than the clay when cooled, although the latter is held to be tougher at a melting heat. The average life of a graph- ite crucible is much greater than that of a clay crucible, however, most crucibles serving for at least three heats, and sometimes for as many as eight heats, before rejection. 420. The Charge of the Crucible. The cru- cible charge is usually about 50 pounds for a new clay crucible, but must be reduced to about 44 pounds for the second charge, and 38 pounds for the third charge, in order that the slag-line, where the wall of the crucible is attacked and weakened, may be lowered each time. The usual charge of the graphite crucible is from 85 to 100 pounds. The English practice in charging clay crucibles is to heat the crucibles in the coke-hole and then introduce the charge through a funnel. The PLAN-SECTION "A-A" Fig. 204.— Gas-fired Regenerative Crucible Furnace. STEEL 369 American practice is to charge the crucible outside the furnace. The larger pieces of iron are first inserted, next the charcoal and ferro-man- ganese or oxide of manganese, and lastly the smaller pieces of iron. The iron of the crucible charge is supposed to be pure puddled iron, but in many cases, especially in American practice, wrought-iron scrap and even soft-steel scrap is substituted for a considerable part of the charge. This practice means cheapening the process and lowering its quality. Blister steel made by the cementation process is very fre- quently used in place of wrought iron in English practice, and rarely in American practice. The resultant product is claimed by the Sheffield steel-makers to be so superior to any steel made that the extremely high cost is justified. The carburizing agent used is almost invariably charcoal, which' is added in small lumps. Ferro-manganese or oxide of manganese is added to aid in forming a liquid slag and to add a little manganese to the metal. Other " physics," such as salt and potassium ferro-cyanide, are sometimes used to aid in slag formation and in the absorption of carbon. Special elements such as chromium, tungsten, manganese, vanadium, etc., are sometimes added when special steels of the class called " alloy steels " are to be produced. 421. Operatioii of Process. The crucibles having been charged either before or. after having been placed in the melting-hole, the covers are placed upon the crucibles and the temperature is gradually brought to a melting heat by building up the coke fire to the top of the crucibles in the case of the coke-furnace, or by turning on the air and gas in the case of the gas-furnace. The process is thereafter divided into two stages called " melting " and " killing." The melting requires from two to four hours, depending largely upon the composition of the charge. Low-carbon heats take much longer than high-carbon, since low-carbon stock melts at a much higher temperature than high-carbon. The melting time averages somewhat less with the gas-furnace than with the coke-furnace, rarely exceeding 2| to 3 hours for the former. The fuel of the coke-furnace requires replenishing two or three times during melting down. There is no sharp line of demarkation between the melting and the killing stage, as this is interpreted by the judgment of the melter, who removes, the cover and examines the charge to make sur6 that it is entirely molten. Killing or " dead melting " consists simply in holding the steel at a melting temperature until it becomes tranquil, i.e., does not evolve gases, and will pour '" dead," i.e., without ebullition, and produce sound ingots. The change which occurs during killing consists either, in.boiling the gases 370 MATERIALS OF CONSTRUCTION out of solution in the metal, or in acquiring soundness by combination of the gases with silicon, which is reduced from the walls of the crucible by carbon. If killing be prolonged too much the usual effects of excess silicon are noted, the steel being hard, brittle, and weak. The killing time depends upon the temperature, the carbon content of the metal, and the content of silicon, sulphur, phosphorus, etc. The hotter the furnace, the lower the carbon, and the purer the metal is, the shorter will be the time required for killing. The actual time required varies according to the above conditions from about i to If hours. The total time required for the completion of the process is 3^ to 5| hours in the English coke-furnace, and 3 to 3i (rarely 4) hours in the gas- fired furnape. (As much as 1 hour longer may be required in using a cold crucible for the first time.) The net results attained in the process are the addition of carbon, the elimination of slag (removed before pouring), and the addition of some silicon and manganese. Since the process is entirely an acid process no sulphur or phosphorus is removed. When the operation of killing is complete the crucibles are gripped by a suitable pair of tongs and hfted out of the melting-hole. Usually the operation of " drawing " the crucibles is performed by a workman who must straddle the melting-hole while doing so. The use of over- head cranes with special tongs has been occasionally resorted to. The cover is now removed and the slag floating on top of the steel is skimmed off by means of an iron rod with a ball of slag on one end, against which it chills. The steel is now poured or " teemed '' into an ingot mold, any remaining slag being kept back by holding a bar against it. The mold used is of cast iron and has commonly a cross-section of 3 to 4 inches square; it is split lengthwise and is held together by rings keyed on. The surface of the mold is coated with smoke from a gas flame, or from burning coal-tar or resin, in order to prevent sticking and to improve the surface of the ingot. The capacity of the mold is commonly that of one crucible. If more than one crucible is used for one ingot the teeming must be done in such a manner as to prevent even a momentary interruption of the stream of metal entering the mold. Otherwise the surface of the metal in the mold will freeze over, causing a serious defect called a " cold shut." In any case the teeming must be done very slowly and carefully, never permitting the stream of metal to strike the sides of the mold. 422. Grades of Crucible Steel. Very wide ranges in composition and properties of crucible steels are obtainable. The composition, in fact, is sometimes uncertain owing to the variable amount of carbon and silicon that may be absorbed from the crucible walls. Ingots are STEEL 371 therefore always graded by breaking off the worthless upper portion containing the " pipe " and examining the fracture. An experienced man is thus able to estimate the carbon content quite closely. Chem- ical analysis is now usually employed to supplement examination of the fracture, and the ingots are separated into several grades of similar analysis. No sharp subdivision of grades and uses of crucible steel can be made, but the following table shows in a general way the character of steel required for different classes of tools: APPROXIMATE GRADING OF CRUCIBLE STEELS (Carbon Steels Only) Uses Carbon % Manganese % Silicon % Sulphur % Phos. % Battering tools ] Hot-work tools . . . . f^ Dull-edge tools, etc. J .45-. 65 .20-. 50 .20-. 30 .02-. 060 .015-. 050 Dies, axes, large ■ ■ ■ \ drills, reamers, etc. ) .65-. 85 .20-. 40 .20-. 30 .015-. 030 .012-. 025 Chisels, knives 1 Drills, lathe tools. . . / .85-1,10 .15-. 30 .15-. 25 .010-, 020 .010-. 020 Razors, fine lathe ... tools and drills.. . . Gravers tools, etc. . . 1.10-1.50 .10-. 25 .12-. 25 .005-. 015 .005-, 015 423. Cost of Crucible Steels. The great difference in cost between crucible steels and Bessemer and open-hearth steels is due largely to the much greater ratio of labor expended to tonnage produced by the former processes compared with the two latter processes. The excessive fuel cost is also a contributing factor, the weight of coke fuel burned in the coke furnace being three or four times the weight of steel produced, and the weight of coal burned in gas producers for the gas-fired furnace being nearly equal to the weight of the steel produced. The high cost of crucibles and the necessity of using expensive raw materials are also important factors. The Bessemer Phocess 424. Historical. The Bessemer process for the manufacture of steel was invented by Sir Henry Bessemer in England in 1855. It is doubt- ful if any single invention or discovery has ever had such a wonderful 372 MATERIALS OF CONSTRUCTION effect upon industry and manufacturing in general. By producing large quantities of steel at a cost far below that of any previously known process, it rendered possible the great industrial development of the world which has characterized the last fifty years. For thirty-five years the Bessemer process led even the open-hearth process, both in tonnage produced and in perfection of methods, but the perfection of the basic open-hearth process has gradually resulted in the production of steel whose superior quality has enabled it slowly to replace Bessemer steel in one field after another, until finally in 1908 the tonnage of open-hearth steel produced in the United States exceeded the tonnage of Bessemer steel for the first time. 425. The Bessemer Process in General. The Bessemer process consists essentially in the removal of most of the impurities in pig iron by oxidation, through the agency of finely divided air currents Uown through a bath of molten iron contained in a vessel called a " converter." In American practice the addition of a " recarburizer " after " blowing " is necessary to give the " blown metal " the required carbon content for steel. The following operations constitute the essential features of the American Bessemer steel process: (1) Molten pig iron is brought from the blast furnace plant in hot- metal ladles and discharged into a large reservoir called the " mixer." (Many of the older plants are not operated in direct conjunction with a blast-furnace plant, in which case the pig iron is remelted in large cupolas within the steel plant.) (2) The mixer supplies molten iron as required to charging ladles, which in turn discharge into the converters, the latter being rotated into a horizontal position during charging. (3) The air blast of the converter is started and the vessel is elevated into a vertical position. The finely divided air currents pass up through the molten metal for a period of about ten minutes, by which time the impurities will have been practically eliminated by oxidation. (4) The converter is again turned into a horizontal position and the wind is cut off. A predetermined amount of recarburizer is now added to the bath in order to obtain a steel of any desired carbon content. (5) The molten steel is poured from the converter into a ladle which IS swung by a crane over a series of cast-iron ingot molds (mounted on cars), mto which the metal is teemed. (6) When the ingots have cooled sufficiently, the molds are stripped off and the mgots are placed in " soaking-pits " or reheating furnaces, where they remain until their still molten interiors have solidified and the temperature of the metal has become equalized throughout STEEL 373 (7) The hot ingots are transferred to the rolling mills where by a series of rolls they are reduced first to " blooms " and then to any desired shape for use in construction. (Presses sometimes replace rolls.) These operations will be further considered in the discussion which follows: 426. The Pig Iron Used. The Bessemer process universally used in American practice is an acid process, i.e., the slag formed is of an acid character. In consequence, it is impossible to remove either phos- phorus or sulphur from the iron, and a grade of pig iron specially low in these elements is required. The usual Hmits of composition of Bessemer pig iron are: Silicon. Manganese. Carbon. Phosphorua. Sulphur. % % % % % 1.0-2.0 0.4-0.8 3.5-4.0 0.07-0.10 0.02-0.07 At least 1.0 per cent of silicon is required in order to insure the pro- duction of a sufficient quantity of slag of satisfactory character, and also to provide heat. In fact, the oxidation of the silicon is the principal source of heat in the converter, the amount so derived being about twelve times that derived from the oxidation of the carbon and ten times that derived from the oxidation of the manganese. (Assuming the iron to have a composition corresponding to the minimum of each constit- uent in the above analysis.) 427. The Bessemer Converter and other Equipment of the Bessemer Plant. The Bessemer converter consists of a heavy steel sheet of cylindrical form supported upon two trunnions upon which it can be rotated. The upper portion of the shell is conical and may either be concentric (Fig. 205), or eccentric (Fig. 206). The former form of " nose " has now almost entirely replaced the latter. The clear opening at the converter mouth is usually from 2 feet to 2| feet in diameter, the inside diameter of the cylindrical portion is about 8 feet, and the height from inside of bottom to " mouth " is about 15 feet. (These dimensions apply to the average-sized converter, having a capacity of 15 tons. Converters are used, however, having capacities all the way from 1 ton to 20 tons.) The lining of the converter is usually from 12 to 13 inches thick, and is made of very refractory material of strongly acid character, silica being the principal constituent. In American practice, ganister blocks or bricks laid with thin fireclay joints are usually employed. In p]ngland, a natural ganister rock is commonly used. The lining is repaired between heats with a mixture of sihcious material and clay, and more extensive repairs are made during shut-downs. Under average conditions a lining may be made to last for several months — perhaps 10,000 to 15,000 heats— before it need be entirely replaced. A new lining must be dried, and a cold lining must be heated to a red heat before the converter is charged. The bottom of the converter is pierced with a great number of small holes, through which the air blast enters from the " wind-box " at the bottom of the vessel. The 374 MATERIALS OF CONSTRUCTION wind-box is connected by a pipe and slip ring with a hollow trunnion which communi- cates with the blower. This arrangement makes it possible to put on the blast regard- less of the position of the converter with re,spect to the normal vertical position. The trunnion opposite the hollow one carries a pinion which engages with a rack which is moved forward and backward by a hydrauUc piston to rotate the converter. The hning of the bottom is made up of damp silicious material bound together with clay in which the molded tuyere brick are set. The number of tuyeres varies from 15 to 30 and each contains from 10 to 18 holes of f to f inch diameter. The total thickness of the bottom Uning is commonly from 24 to 30 inches. On account of the fact that uncombined iron oxide has a strongly corrosive action on the Uning, the bottom is corroded very rapidly, especiaUy m the vicmity of the tuyeres, where the air encounters the molten iron. This limits the life of the bottom lining to about 20 or 25 heats, even though repairs are made between heats. Fig. 205. — 12-15 Ton Bessemer Converter. Concentric Type. Fig. 206. — 12-15 Ton Bessemer Converter. Eccentric Type. On this account the bottom of the converter is made easily detachable, the fasten- ings to the body being links secured by keys which can be quickly removed. The worn-out bottom is lowered by an hydraulic jack, located beneath the converter, on to a car which conveys it to the repair room. Meanwhile a new bottom on a second car is lifted into place and keyed on. The new bottom is carefully dried and heated to a high temperature before being put in place, and the joint with the body lining is daubed with mud before the bottom is forced into place by the hydraulic ram. In some cases the average time required to replace a bottom does not exceed 20 minutes. The blast is derived from blowing engines of either the vertical or horizontal type operating on steam or, in the later installations, on blast-furnace gas. A pressure of from 20 to 30 pounds per square inch is maintained. The turning on and off of the blast, as well as the movement of the converter, is all under ^the control of the " blower " who stands on a raised platform called the " pulpit " within full view of the entire operation. If the Bessemer plant is not operated in conjunction with a blast furnace the pig iron is melted in cupolas which differ in no respect from the ordinary foundry cupola except in size, the usual dimensions being from 8 to 12 feet in internal diameter and 40 STEEL 375 to 60 feet in height. Usually from 3 to 6 cupolas are required for a Bessemer plant. The cupolas are commonly worked continuously for several days or a week, after which time they must be closed down to remove the " scaffolds " and make necessary repairs. The metal is tapped at intervals into ladles which convey it directly to the converter. More commonly, in recent years, the pig iron is run into ladles at the blast furnace, and transferred (perhaps a considerable distance) to the steel plant, where the ladles are lifted by a traveling crane and discharged into the mixer, which serves as a source of supply for the converters. The mixer (Fig. 207) is simply a large steel reservoir Uned with refractory brick and mounted on rollers. Hydraulic cylinders located at the corners serve to tip the mixer to pour out the metal. The capacity of the mixer is from 150 to 500 tons. Be- cause of its large size the mixer wUl hold the product of several furnaces, and tends to Fig. 207.— Metal Mixer. Fig. 208. — Positioii of Converter while Receiving Charge. average the irregularities in the different irons, thus lessening the variability of the converter charge. The mixer also serves to keep the pig iron molten for an indefinite length of time, compensates for delays either at the blast furnace or at the steel plant, and affords an opportunity for the addition of special pig iron, if necessary to correct the composition. When a converter is to be charged, a ladle on a car is run under the pouring spout of the mixer and metal is poured into the ladle untU the latter contains the amount required for a converter charge, as indicated by the track scales upon which the ladle car rests. The ladle car is now transferred to a point opposite the converter, the latter is rotated to a nearly horizontal position, and the charge is poured in. 428. Operation of Process. The positioii of the converter while receiving its charge is shown by Fig. 208. The concentric converter may be charged from either side, but the eccentric converter is always inclined in the same direction. The bath of metal never reaches the height of the tuyeres before the vessel is righted, and the blast is turned on after charging and before righting in order to prevent the metal from entering the tuyeres. The bath of metal occupies only a small portion of the volume of the converter (perhaps 18 inches depth in a 15-ton con- 376 MATEEIALS OF CONSTRUCTION verter), on account of the great increase in the volume of the bath caused by the violent ebullition of the metal during the blow. After two or three minutes a reddish-yellow flame begins to pour from the mouth of the converter, indicating the beginning of the oxida- tion of carbon. This flame becomes rapidly augmented until a white- hot flame, 20 to 30 feet in height, pours out with a loud roaring sound, and a shower of sparks appears, owing to the ejection of slag and metal. Soon the flame begins to flicker and shorten, indicating that the carbon is practically burned out, whereupon the converter is immediately turned down and the blast shut off. 429. Chemistry of Process. Under the extremely active oxidizing influence of air driven through the bath of molten metal, all of the ele- ments are oxidized, almost without regard to their relative affinities for oxygen. Thereupon the slag formed by oxidized iron and other elements pos- sessing a lesser affinity for oxygen, attacks the more easily oxidized ele- ments and the latter become eliminated first. As soon as the blow is on the silicon and the manganese be^n to be burned to Si02 and MnO, the action being partly direct oxidation by the oxygen of the blast, and partly indirect through the agency of the FeO, and the CO which are easily robbed of their oxy- gen. Some iron oxide always survives, how- ever, and with the silica forms a siUcate of iron, FeSiOa, which is in large part re- tained in the slag. Additional iron oxide dissolved in the Cirlo n ■^ - -^^ ^ ~^ s V \ n^- \ ,*' ^H ^ * ^"^ •» V n* «o: -iH combined volume of the four regenerators of a 50-ton furnace amounts to about 10,000 cubic feet, the air regenerator being about one-third larger than the gas regenerator. The usual arrangement of flues and valves is similar to that shown in Fig. 204. The gas flues leading from the regenerators converge to- ward the gas valve, which is set over the stack flue on the center-line of the furn- ace installation. A common type of water-sealed valve is shown in Fig. 217. The base-plate of the valve holds several inches of water into which the shell of the valve and the movable diverting hood project. The opening to the stack flue is shown at 6 and the two gas flues at c c. By reversing the arms which carry the hood by means of the outside lever, the hood may be shifted in such a manner as to afford a sealed outlet from either gas flue to the stack flue, the incoming gas — which enters the valve through a pipe at the top — being at the same time diverted to the flue not covered by the hood. Similarly, the air flues converge to the air valve, which is also set over the stack flue, and which acts in exactly the same manner as the gas valve above de- scribed. A common type of air valve, called the "butterfly " valve, is shown in Fig. 218. From the valves the waste gases pass directly to the stack, which must have sufficient draught to exijaust the melting-cham- ber and draw the gases through the regenerators, flues, and valves. Often one stack is made to serve two furnaces, stack dampers being placed in the flues thereto. (t) TiUing or Rolling Furnaces. The tilting or rolling furnace of the Welhnan 390 MATERIALS OF CONSTEUCTION Fig. 217. — Water-sealed Gas Valve. type is shown in Fig. 219a (transverse section), and in Fig. 2196 (longitudinal section). The furnace consists of a heavy steel casing of rectangular form, lined with masonry like the stationary furnace, but mounted on two steel rockers which rest upon heavy bed-castings. Two large hydraulic cylinders on the pouring side of the furnace serve to rock the furnace forward or backward during the operation of pouring. The interior of the melting-chamber of the tilting furnace differs in scarcely any respect from that of the stationary furnace except that the spout is located above the level of the bath until the furnace is tilted on its rollers, and the design of the port opening is differ- ent. A cast-iron water- cooled ring is fitted around the port openings on the outside, and matches a similar ring on the port. The ports are built inside a heavy steel framework and are mounted on wheels or tracks carried by the up-takes from the regenerators. When it is desired to tilt the furnace, the ports are drawn back away from the furnace ends and, since the port openings are exposed while the furnace is in its tilted position, the gas and air supply must be stopped during the operation. A common adjunct of the rolling furnace is the " fore-hearth," a substitute for a teeming ladle, which is bolted on in place of a pouring spout on the back of the furnace. The fore-hearth resembles a teeming ladle in its design and is provided with two teeming valves through which the metal is discharged into ingot molds while the furnace is in its pouring position. The design and arrangement of the slag-pockets, regenerators, valves, and flues for the rolUng furnace differs in no essential respect from that of the stationary furnace above described. 446. Stationary vs. Tilting Furnaces. The stationary furnace is less complicated in its design and requires less elaborate equipment for its operation. It is less expensive to install, and the cost of its upkeep is much less. On the other hand, the tUting furnace has the advantage of never causing delay and consequent oxi- dation of the bath, owing to difficulty in clearing the tap-hole; the slag can be poured off at any time, which is a great advantage in the basic furnace especially; repairs to the hearth bottom may be made with much greater facility between or even during heats; metal may be poured off at any tune and transferred between furnaces (a necessary operation in the Talbot proc- ess, later described); and boiUng and violent action during pouring is lessened by the chilhng of the metal caused by the entrance of cold air through the open ports. 446. Life of Furnace and Repairs. The life of the open-hearth furnace is extremely variable, dependmg upon the quality of the materials used and the management of the furnace. The ports are usually the first portion to become excessively injured and o ten need replacement long before the furnace must be entirely rebuilt. The remova- Fig. 218. — Siemens Air Valve. ble ports of the tilting furnace have a special advantage on this account. STEEL 391 The bottom usually requires repairs in spots between heats, and is more extensively- repaired during temporary shutdowns. In this way this bottom lasts ahnost indefi- nitely. The root of the furnace fails sooner or later by burning through in spots or by falling in. When this happens the furnace must be practically rebuilt. Fig. 219o. — The Wellman Tilting Open-hearth Furnace. Transverse Section. Fig. 2196. — The Wellman Tilting Open-hearth Furnace. Longitudinal Section. The regenerators finally give out, either by becoming choked up with dust, or by cracks opening up in the walls. A shutdown is required in either case. In general the life of a furnace is from 200 to 600 heats, averaging about 350 heats, or from three to six or eight months' operation. The acid furnace lasts longer, from 800 to 1200 heats, or from ten to sixteen months. The Furnace Fuel 447. Natural-gas FueL Natural gas is the ideal fuel for the opera- tion of the open-hearth furnace, but its use is necessarily limited to the 392 MATERIALS OF CONSTRUCTION vicinities of the somewhat restricted districts of Pennsylvania, Ohio, West Virginia, etc., where it is available. Natural gas is essentially marsh gas or methane, CH4, with varying admixtures of other members of this series of hydrocarbon gases, together with hydrogen. It usually contains 60 to 80 per cent of methane and 20 to 30 per cent of hydrogen. It occurs usually associated with oil, and is probably produced by distillation from oil or coal within the earth. The wells are commonly several thousand feet deep. Natural gas is superior to artificial gas in that it has a higher calorific value (about 1000 B.T.U. per cubic foot), it is much purer, and it is much cheaper unless it is necessary to pipe it for very long distances. About 5000 to 6000 cubic feet are used in the open-hearth per ton of steel pro- duced. Preheating of natural gas for the open-hearth furnace is often omitted, the gas being introduced directly into the ports of the furnace. 448. Producer Gas. Only one form of artificial gas is of any im- portance in steel-making, this one being producer gas. The modern type of gas producer is a cylindrical shell of steel, lined with fire- brick. A common type of producer is shown by Fig. 220. Coal is charged into the hopper at the top and the cover is closed before the bell is lowered to dis- charge the contents into the producer. A supply of air, insufficient for complete com- bustion of the coal, is admitted through openings in the wind-box below the grates, and also through the pipe in the center. A steam injector is used to draw in the air. The ash falls through the grates into the ash pit, which is filled with water into which the shell of the producer projects, thus forming a water-seal. The air admitted through the grates of the producer converts a ^^ .... part of the carbon of the fuel to CO2, which IS reduced to CO in passing upward through the incandescent bed ot fuel. The carbon monoxide is diluted by the nitrogen of the air introduced and somewhat enriched by the hydrogen of the decom- posed steam and other moisture present. The actual ultimate compo- sition of producer gas is variable, but averages about 61 per cent N, 20 to 25 per cent CO, 6 to 10 per cent H, 3 to 8 per cent CO2, and 2 to 4 per cent methane. Fig. 220.— Water-sealed Gas Producer. STEEL 393 One ton of average good bituminous coal yields about 150,000 cubic feet of gas whose thermal value averages about 130 B.T.U. per cubic foot. Therefore to produce 1 ton of steel, which calls for the develop- ment of from 5,000,000 to 6,000,000 British thermal units of heat, it is necessary to bum an average of from 570 to 700 pounds of coal in the producer. The Basic Open-hearth Pkocess 449. General. The basic open-hearth process, like the basic Bessemer process, differs from the acid process mainly in that it utilizes stock higher in phosphorus and sulphur, and a basic slag must be produced by the addition of strong bases to the charge in order to effect the removal of this excess phosphorus and sulphur. In order to resist the corrosive action of the basic slag the hearth must be lined with basic material as above described, but the hearth lining, whether acid or basic, plays no part in the purification of the iron. 450. The Furnace Charge. Operating Practice. Basic open-hearth practice is characterized by a number of quite distinct modifications, depending upon the choice of materials which constitute the furnace charge. The resulting processes are characterized as: (a) The pig-and-ore process; (6) The pig-and-scrap process, and (c) The all-scrap process. The pig iron used must contain from 1 to 2J per cent P, less than 1 per cent of silicon, and at least 1 per cent of manganese. The car- bon will usually run from 2.5 to 3.5 per cent. The pig iron may be charged either solid or molten, one form being used about as often as the other in American practice. The substitution of scrap iron for pig iron results in shortening the time required for the operation, since there will be a lower percentage content of impurities to eliminate. When the pig iron is entirely re- placed by scrap in the " all-scrap process," there is insufficient reducing material present to prevent excessive oxidation of the iron in melting, and it becomes necessary to use some form of carbon to supply the need for a reducing agent. In the " pig-and-ore process " the charge is principally pig iron, to which ore is added in order to hasten the process. The limiting amount of ore is reached when the boiling of the charge becomes excessive. Unless pig iron is cheaper than scrap, the pig-and-ore process is not a com- mercially practicable one. Molten pig iron is sometimes used in this process, the iron being poured in on the ore, which has first been charged. 394 MATERIALS OF CONSTRUCTION A mixer is commonly interposed between blast furnace and steel furnace as in the Bessemer process. The " pig-and-scrap " process is. now the most usual process, the average practice being to use about 50 per cent of each material, the exact proportions used in a given case being largely a question of relative costs. Very commonly a small proportion of ore or mill scale is added to hasten the process. The order of charging varies, but the usual method consists in charg- ing first a small quantity of small scrap to protect the hearth, after which the lime flux and the ore or mill scale is charged upon the scrap. Most of the balance of the scrap and about one-third of the pig is now charged and the furnace started. The balance of the scrap and the pig are added after an hour or two, when the former charge has melted down to make room. The flux used is commonly lime, the amount required being depend- ent upon the contents of phosphorus, and especially sulphur, in the charge. High silicon also increases the amount of lime required in order to render the slag basic. In general the charge of lime amounts to from 10 to 30 per cent of the total charge, and the resulting slag contains from 35 to 45 per cent of CaO. 451. Chemistiy of Basic Process. Melting on the basic hearth is attended by oxidation of the metalloids, the most easily oxidized ones being eliminated first. The general sequence of removals of metalloids is shown in Fig. 221.* The silicon, manganese and carbon are all con- siderably reduced in amount during the period of melting (the first four or five hours). The phosphorus, however, remain* practically unaffected until the end of this period, when it begins to be rapidly oxidized. It is necessary, in order to prevent the bath from becoming too cool and to prevent oxidation of the iron toward the end of the opera- * Bradley Stoughton, loc. cit., p. 133. - — — — ~ ^ \ d t, _ -JS w - \ v ~^ 'v ■^ s . ~ X V 1.6 as _ \> .^ 1- s v s JS. rt N\ 1 ~^ ^fe ' 'ff s A i ~ 1.0 'A \ " ~ "^^ = -i K *r- _ _ — \ -— J ^ - - -- - - - 0,6 ^ \ \ ~ ~ ijorJ \ \ / -J ullih !■■ -- -,; _ ^ ^ i- _ _ _ « 1 a 3 i ■ 6 1 fou c ■ r" ) 1 — ' ' — ' ^ Fig. 221. — Removal of Impurities in Basic Open- hearth. (Bradley Stoughton.) Steel 395 tion, to see to it that the carbon is eliminated last. If the carbon is disappearing too early it is therefore the practice of the melter to add pig iron to provide additional carbon. Sometimes, when the phosphorus burns out rapidly, and the carbon too slowly, it is necessary to hasten the oxidation of carbon by adding ore. The progress of the operation is tested from time to time by ladhng out a small amount of metal, cast- ing a small test billet, breaking it, and examining the fracture. Billet tests are now commonly supplemented by chemical analysis of frequent samples. The slag performs several very important offices in the operation of the basic hearth. Its chief function is to take up and retain the oxides of the metalloids — sihcon, manganese, phosphorus, and sulphur. It must also act as a protection to the bath from excessive oxidation by the furnace gases, and, by virtue of its contained oxides, assist in the oxidation of the impurities. For efficient action as a deoxidizing agent it must be very fluid in order that it may mix intimately with the bath. The slag must also be rich in bases in order to retain the oxides of phosphorus and sulphur, and the removal of sulphur is greatly facilitated if the slag is rendered still more strongly basic by the addition of calcium fluoride just before the end of the operation. No oxides, such as ore or scale, can be added toward the end of the operation, lest the final slag be an oxidizing one, resulting in the metal being full of oxides and there- fore turbulent when poured and very unsound after solidification. 452. Recarburization. Recarburization of basic steel cannot be accomplished in the furnace because the carbon, silicon, and manganese of the recarburizer would reduce the phosphorus in the slag and restore it to the metal. On this account the recarburizer, in the form of ferro-manganese, together with coal, charcoal, or coke, is added to the stream of metal as it flows into the ladle. Provision is made for the removal of the greater part of the slag by overflowing at the top of the ladle. Spiegeleisen is not used as a recarburizer, because it must be used in a molten state and a cupola could not be operated to supply it in proper condition at the infrequent intervals at which a recarburizer is required in open- hearth operation. Low-carbon steel is always poured only after the carbon has been reduced to about 0.10 or 0.15 per cent, the additional carbon required being added by the recarburizer. High-carbon steel is preferably handled in the same way, but it is cheaper and takes less time simply to reduce the amount of carbon slightly below the desired point, and then recar- burize. The latter method is often used, but entails the danger of pour- 396 MATERIALS OF CONSTRUCTION ing steel while it is still too high in phosphorus and in oxides which cause unsoundness of the ingots. 453. Pouring the Ingots. The steel is discharged from the pouring spout or tap-hole of the furnace into a teeming ladle or fore-hearth, from which it is teemed into ingot molds mounted on cars as in the usual Bessemer practice above described. The ordinary open-hearth ingot may weigh as much as 10 tons and is therefore much larger than the usual Bessemer ingot. The Acid Open-heahth Process 454. General. The acid open-hearth process differs from the basic open-hearth process principally in the character of the iron used, the omission of the flux, and the time required for the operation. Since the slag formed is acid, it is unable to retain oxides of phosphorus and sulphur, and a pig low in these elements is required. An acid lining of the hearth is of course required to prevent rapid corrosion by the acid slag. The time required for the operation is shorter than that for the basic process because the iron contains less impurities to be removed (especially because a larger proportion of scrap is used), because the process need not be prolonged to remove phosphorus, and because no part of the heat is consumed in melting and accomplishing the function of the flux. 455. The Furnace Charge. The charge of the acid furnace usually consists of approximately one-third pig iron and two-thirds scrap. The pig iron is fairly low in sihcon, and low in manganese, phosphorus, and sulphur. The usual limits of composition are from 0.8 to 2.0 per cent sihcon, 0.3 to 0.5 per cent manganese, and less than 0.05 per cent of both phosphorus and sulphur. The carbon content will be from 3.0 to 4.0 per cent. The scrap will be of variable composition, but will average about 0.2 to 0.3 per cent carbon, 0.1 to 0.3 per cent silicon, 0.4 to 0.8 per cent manganese, and less than 0.05 per cent of both phosphorus and sulphur. Ore is not usually initially charged in the acid furnace, but may be added during the process for the sake of increasing the oxidizing agencies to hasten the removal of carbon. The pig iron is charged into the fur- nace before the scrap in order to prevent the scorification of the hearth, which would occur if scrap were charged first. 456. Chemistry of Acid Process. The melting operation is in the main an oxidizing action, though the flame may be a comparatively reducing one during charging in order to prevent excessive oxidation of the pig iron and, more particularly, the scrap during this stage. (The STEEL 397 (H) — — 70 r fid i£| °fl S- S N s "n \ S — k s S ' — ^ l^ ?s r ' S S V . ^ S^'n II H" .09 S &!, 1 - 1 t ( V4 ■v; p% "< s \ s o o "i s s ^ s g \ s Ufa .03 'ho !£. 21 us — s _ ^ =5 N ^ J- .<» Isi IP IIU - p - - - 3 k- te tt SI V „09 ' k Lh DUI 1.02 1.32 6.02 5.32S.52 i.n i.i7 B.U reducing flame is secured by cutting down the air supply below the quan- tity required for complete combustion.) The metalloids are largely eliminated during the melting-down stage, which requires some three or four hours, the silicon usually disappearing first, closely followed by the manganese. The amount of carbon oxi- dized during melting is dependent largely upon the amounts of the more easily oxidized elements (silicon and manganese) present. The lower the content of the latter elements, the greater will be the proportion of carbon oxidized. In any event two-thirds of the carbon will be re- moved very soon after the charge is altogether melted. The balance will be oxidized only very slowly, its disap- pearance usually being accelerated by the addi- tion of ore to the bath. The slag in the acid process never constitutes the important oxidizing agency that it does in the basic process. The sequence of removals of the metalloids in a typical acid heat is shown by Fig. 222. 457. Recarburizing. Recarburizing in the acid process is accomplished in the furnace, rather than in the ladle, because the considerations which prevent recarburization in the basic furnace are not here opera- tive. The practice as to the degree of recarburization varies, but in gen- eral the carbon is reduced to the practical minimum in the case of mild or medium steel, and the recarburizer added twenty to forty minutes before the heat is tapped. In the case of melting for high-carbon steel the carbon is usually reduced only slightly below the desired amount before the recarburizer is added. In this event, indeed, the addition to the bath is rather more a deoxidizer than a recarburizer, the primary object being the restoration to the bath of iron which has become oxi- dized. The recarburizer or deoxidizer is, in this process, ferro-man- ganese and ferro-silicon, as a rule. Coal is sometimes added in the ladle as in the basic process. The addition of the recarburizer necessarily largely increases the content of silicon and manganese in the steel, as well as the carbon content. Fig. 222. — Removal of Impurities in Acid Open- hearth. (Bradley Stoughton.) 398 MATERIALS OF CONSTRUCTION Special Open-hearth Processes 458. Duplex Processes. There are two methods by which the acid Bessemer and the basic open-hearth processes may be combined in a so-called " duplex process," whereby the silicon, manganese, and part of the carbon are eliminated in the converter, the phosphorus and the remainder of the carbon being removed in the open-hearth. In the most generally used duplex process the pig iron is blown in the acid converter until the silicon and manganese are practically eliminated and the carbon reduced to about 1 per cent, after which the converter charge is transferred to a mixer and thence to the basic open-hearth, wherein the phosphorus and the remainder of the carbon are removed. The second method differs from the one just mentioned in that the Bessemer blow is continued till the carbon is reduced to only a few tenths of 1 per cent, after which the product is transferred to the open-hearth, mixed with a large proportion of pig, generally molten, and the process completed exactly as in' the case of the latter part of the ordinary pig- and-scrap process. The advantage of the duplex process over the Bessemer process lies in the fact that lower grade, high-phosphorus pig iron may be used and yet produce an open-hearth steel which is admittedly superior in quality to Bessemer steel, and which commands a higher price in the market. The advantages of the combined process over the open-hearth proc- ess are the saving of about one-half of the time ordinarily required in the open-hearth, and the saving effected in cost of renewals of the hearth lining, because of the fact that the silica is removed before the metal enters the hearth. Forty to fifty heats per week are made in the duplex process, as compared with the eighteen heats usually made by the ordinary method. 459. The Talbot Process. The MoneU Process. The Talbot process is a modification of the basic open-hearth process, whereby the hearth is worked continuously. The tilting furnace must be used, and the depth of bath is twice that ordinarily employed. The capacity of the furnace may exceed 200 tons. The furnace is charged and conducted in the usual manner for the pig-and-ore process (no scrap being used), until the carbon is reduced to the desired point. The slag is then poured off and about one-third of the steel is poured into the ladle, recarburized, and teemed. Iron ore and limestone are now added to the bath to form a new slag, and molten pig iron is poured in to make the charge equal to its original amount. Oxidation proceeds very rapidly, owing to the large amounts of iron oxide in the slag. The interval between the addition of pig and the STEEL 399 pouring of steel is only from three to six hours, and three to four heats are poured per day. The principal advantages of the process are the large tonnage pro- duced arid the excellent temperature control possible. The furnace is expensive, however, both to build and to maintain. The Monell process is another modification of the basic open-hearth process, wherein a very strongly oxidizing slag is prepared before molten pig iron is charged. The slag is produced by melting limestone with ore or other iron oxide equal to about one-fourth of the pig-iron charge. Molten pig iron is poured into this slag and oxidation proceeds with ex- treme rapidity, most of the silicon, manganese, and phosphorus being eliminated within one hour. The balance o/ the process is exactly simi- lar to the ordinary pig-and-scrap process, and the total time required is about the same. Electric Repining of Steel 460. Electric Refining Processes in General. The part which elec- tricity plays in the various electric steel processes is, as in the case of the electric reduction of ores, simply that of a source of heat. Slags, which are strongly oxidizing in character, must be added to the bath in order to effect any refinement. In practice, the slag used is a strongly basic iron-oxide slag, because such a slag will oxidize phosphorus as well as retain the oxide formed. If much phosphorus is to be removed, or if it is to be reduced to a very low point, it is necessary to use two slags, skimming off the first after it becomes highly phosphorized. Sulphur can be reduced only by removing the iron-oxide slag, after the elimination of phosphorus, and producing a slag that is made up almost entirely of lime. The presence of manganese favors the removal of sulphur, because manganese sulphide is more readily taken up by the slag than iron sulphide, thus making possible the formation of cal- cium sulphide, in which form the sulphur is retained in the slag. Oiie of the greatest advantages of electric refining processes lies in the fact that- dissolved gases, occluded oxides, etc., are readily removed from the steel through the agency of heat alone. The lack of pertur- bation of the bath and the length of the process, combined with the very high temperature attained, provide ideal conditions for the slow migra- tion of these very small particles to the surface. No subsequent deoxi- dation of the metal is required after finishing. The general nature of the processes of electric refining, as carried out in the various types of furnaces, will be briefly indicated in the dis- cussion of furnaces which follows: 400 MATERIALS OF CONSTRUCTION 461. Types of Electric Refining Furnaces. Three quite distinct types of electric steel refining furnaces have been developed to the point of practical commercial application. These are: (1) Furnaces employing an open arc between electrodes above the bath, the latter being heated by radiation alone. The Stassano furnace is of this type. (2) Furnaces em- ploying an arc between electrodes and the bath, the latter forming a part of the electric circuit. The metal is heated largely by con- duction from the slag bath, which carries much of the current and is heated both by radiation and by reason of its electrical resist- ance. The H&oult, the Girod, and the Keller furnaces are of this type. (3) Electric induc- tion furnaces, wherein the bath forms the secondary of a trans- former consisting of a closed circuit with but a single turn. The Kjellin and Rochling- Rodenhauser furnaces are of this type. 462. Open-arc Fur- naces. The Stassano Furnace. The Stassano furnace is the best-known type of open electric-arc furnace. Its general arrangement is shown by Figs. 223a and 2236. The working chamber is a cyUndrical structure, hned with basic refractory brick, mounted upon a circular track at a slight inclination from the vertical, and slowly rotated in order to increase the mixing effect. SECTION C-D 2.0 Meters Fig. 223. — Stassano Open-arc Electric Furnace. STEEL 401 Electric energy is converted into heat through the medium of an arc between carbon electrodes just above the level of the bath, the bath being heated by radiation. Three-phase alternating current is used, three electrodes being necessary. The power consumption is about 1000 kilo- watt hours per ton of steel produced from cold scrap. Any grade of scrap or pig may be used, and the refining carried to any desired degree, the steel produced being comparable to best crucible tool steel in quality. The maximum capacity of the Stassano furnace is about 5 or 6 tons of metal per heat. 463. Arc-resistance Furnaces. The Heroult, Girod, and Keller Furnaces. The Heroult electric furnace resembles closely the open- hearth furnace except that heat is suppUed by electrical means instead of by the combustion of gas fuel. The form of the furnace is shown by Figs. 224a and 2246. Current enters through one of the electrodes which are suspended above the bath, arcs to the charge through which it passes, and thence arcs to the negative electrode. The H6roult furnace may be used to refine any grade of scrap, but its most important applications have been in conjunction with Bessemer converters and with the open-hearth in duplex units. The metal is, in this case, brought in a molten state from the converter or from the open-hearth, and practically any desired grade of steel may be produced. Superior grades of rail steel, axle steel, steel for wire, steel for castings, tool steels, and special alloy steels have been made with great success by refining either Bessemer or open-hearth metal in the Heroult furnace. The capacity of the largest Heroult furnaces is about 15 tons of metal, and the refining of Bessemer metal has been carried on for weeks at the rate of twelve to fifteen heats per day, meaning a daily output of about 180 tons of super-refined steel. The power consumption in refining molten Bessemer metal is from 100 to 200 kilowatt hours per ton of steel, while in melting and refining cold scrap the power consumed is 600 to 800 kilowatt hours per ton. Three-phase alternating current is used. The Girod electric furnace. Figs. 225a and 225b, like the Heroult furnace, employs an arc between the positive electrode and the bath, but the current is conveyed away by conductors built in the magnesite lining of the bottom of the hearth, instead of arcing to a negative elec- trode above the bath. The negative pole is formed by a number of pieces of soft steel which are prevented from fusing to too great a depth by water cooHng. The Girod furnace has been used abroad for practically the same pur- poses as the H6roult furnace has been used in this country. It works well as a duplex unit with the Bessemer or open-hearth and is also very 402 MATEEIALS Or CONSTEUCTION STEEL 403 efficient m melting and refining cold scrap. Its power consumption and capacity is about the same as the H^roult, except that the largest fur- naces have a capacity not exceeding 8 to 12 tons. The Keller electric furnace, Fig. 226, resembles the Girod furnace in having a conducting bottom, but instead of having only six negative pole pieces, iron bars 1 to 1| inches in diameter are placed vertically over the entire bottom, being spaced only about 1 inch apart. The intermediate spaces are rammed tightly with magnesia. After becoming heated up the entire bottom becomes a con- ductor, the magnesia itself becoming conductive. It is claimed for such a bottom that the more uniform distribution of energy flow means added efficiency as compared with the Girod furnace, and that the bottom is more durable. It is evident that with both the Girod and the Keller furnace the current density or amperage must be double that required for a furnace like the Heroult, where the electrodes are in series, the bottom not being one of the poles. 464. Induction Furnaces. The Kjellin and Rochling- Rodenhauser Furnaces. The electric inductiqn furnace dif- fers radically in theory and design from the furnaces employing electrodes. The induction furnace is in principle simply a stepdown transformer, wherein the metal under treatment forms a closed secondary circuit. One of the simplest forms of induction furnace is the Kjellin fur- nace shown in Figs. 227a and 2276. The primary coil, through which Fig, 225. -Girod Arc-resistance Electric Furnace. 404 MATERIALS OF CONSTRUCTION high-voltage alternating current is passed, is shown at AA ; the magnetic core of laminated iron, in which an alternating current is induced by the primary, is shown at BB; and the bath of metal under treatment shown at CC (contained in the circular crucible DD, built of calcined magnesia or dolomite) forms the secondary in which current is induced by the magnetic core. The resistance offered by the molten metal to the passage of the induced current is the source of heat whereby any desired degree of refinement may be ac- complished. This furnace ranks high as to thermal efficien- C3^, but its use is limited to rather small capacities because of the limited contact area between slag and metal, which means slow refining. Few fur- naces of this type have been built with capacities exceeding § ton, and its principal application has been in the refinement of metal for tool steel. If the metal charged is not al- ready molten, a portion of each charge must be left in the crucible to es- tablish a complete circuit until the solid metal has been melted down. The Rochling-Rodenhauser furnace, .shown in horizontal section in Fig. 228, does not differ in principle from the Kjellin furnace, being in fact simply a combination of three simple induction furnaces designed to use three-phase current. There are three primary coils, AAA, the cores of which, BBB, are conne.cted together above and below by a horseshoe-shaped member built of laminated iron. The' adjacent por- tions of the annular cavities, CCC, containing the bath, widen and join to form a central open chamber, D, which serves as a working chamber, upon which slags, ores, alloys, etc., may be charged and manipulated! This central chamber greatly increases the expedition with which refine- ment is accomplished. In order to prevent loss through magnetic leakage from the primary coils a few turns of heavy copper wire are placed out- FiG. 226. — Keller Arc-resistance Electric Furnace. STEEL 405 side these coils and connected with steel plates, EEE, embedded in the masonry of the lining. The refractory material covering these plates SECTION A-B 0.5 1:0 1.5' 2.0 Meters 1 .... I . I . , I . , , , I , I , , I Fig. 227. — ^Kjellin Electric Induction Furnace. is of such a nature that it acquires electrical conductivity when heated, thus allowing the current which flows to the terminals to pass directly to the bath. 406 MATERIALS OF CONSTRUCTION The usual capacity of the Rochling-Rodenhauser furnace does not exceed 2 or 3 tons; the power consumption is 200 to 250 kilowatt hours per ton (using molten charges), and the product may be super-refined mild steel, high-grade tool or alloy steel, steel castings, or even rail steel. 465. Applications and Limitations of Electric Furnaces and Electric Refining Processes. The electric furnace has already made itself a formidable rival of the crucible process, because it is able to make large:' tonnages of tool steel of crucible quality at a lower cost. In the field of steel castings it has also a considerable advantage over the highest grade ;-TT Fig. 228. — Rochling-Rodenhauser Electric Induction Furnace. Horizontal Section. crucible-made castings in the matter of cost, though it cannot compete with the open-hearth in the production of lower-grade castings. The electric furnace is not, and perhaps cannot under usual con- ditions be, a competitor of the Bessemer converter and the open-hearth in the production of mild and medium steel of ordinary quality. It has already become an important adjunct of both of these processes, how- ever, taking their product and super-refining it to produce steel for spe- cial purposes, a notable one being the production of steel rails which, without sacrificing hardness, are rhuch tougher than the ordinary Bes- semer product. In the production of special alloy steels, the electric furnace has a special advantage over other steel processes in that it need not be STEEL 407 operated under oxidizing conditions, but may be worked under either . neutral or actually reducing conditions. This is an important con- sideration in using certain valuable alloying elements which are very easily oxidized and lost iu other furnaces. Rolling Mill Opebations. The Finishing of Steel Reheating 466. Necessity for Reheating. An ingot cannot be sent to a rolling mill and rolled immediately after the ingot mold has been removed, because at that time the interior is still molten. If, on the other hand, the ingot were allowed to stand until the interior has solidified, the exterior would be too cold to be worked. It is therefore necessary to place the ingots, immediately after stripping, in a furnace where the interior may be solidified and the exterior kept at the required temper- ature for working. The process of rolling finished steel sections from ingots is a protracted operation, necessitating a great many passes of the metal through the rolls. It is therefore necessary, at one or more stages in the reduction of the section, to reheat the bloom or billet or slab which has been formed by the initial reduction of the ingot, and which has cooled below the proper working temperature. Two classes of reheating furnaces are therefore a necessary part of the equipment of a rolling mill: first, a furnace for heating ingots, or at least equalizing their temperature within and without, and second, a furnace in direct connection with the rolling mill, wherein billets or un- finished shapes may be reheated at any stage in the process of rolling. 467. Reheating Furnaces. Practice of Reheating. There are three principal classes of reheating furnaces: (1) the "soaking pit" and (2) the regenerative gas-fired pit-furnace for ingots, and (3) continuous furnaces for billets and other small sections. Special types of reheating furnaces are also required for reheating large blooms and slabs. The original soaking pit, which is still used abroad but has been replaced by the gas-fired piWurnace in this country, is simply a masonry chamber, built below the floor level, and charged through the top in order that the ingots may remain vertical while soHdifying. No fuel is employed, but the ingots are stripped and placed in the soaking pit as soon as possible after teeming, the heat of the stiU molten interior of the ingot being depended upon to bring the exterior to the proper temperature. The regenerative gas-fired pit-furnace is also a vertical furnace, built below the floor level, and charged through the top. Gas fuel is burned within the heating chamber) the arrangement of the furnace being exactly similar to that of the regenera- tive gas-fired crucible furnace. (Fig- 204.) 40S MATERIALS OF CONSTRUCTION BDlet-heating furnaces are now commonly of the reverberatory type, gas-fired, and recuperative in principle. Such a furnace is shown in Fig. 229. The billets are charged at the cool end of the furnace, and are pushed along through the length of the furnace by an hydraulic ram mounted at the charging end. Water-cooled pipes laid in the bed of the heating chamber and extending throughout its length, provide a sort of track along which the billets are pushed. The billets encounter hotter temperatures as they approach the end where the gas and air ports are located, and are there discharged and conveyed back to the rolls. The burned gases, upon leaving the heating chamber, are caused to pass through a series of pipes in a chamber below the working chamber, and the air which is to be used for combustion is caused to circulate through this chamber, Fig. 229. — Gas-fired Recuperative Billet Heating Furnace. thus becoming preheated. (This explains the application of the term " recuperative " to the furnace). The gas need not be preheated since it comes to the furnace directly from a producer worked in direct connection with the furnace. Rolling 468. General. The reduction of metal in rolls is, from the stand- point of the possible beneficial effect of the mechanical work involved upon the structure and mechanical properties of the steel, inferior to either hammering or pressing. It is, however, by far the most rapid method, and this fact coupled with its lower cost and lesser demand for labor is responsible for the almost universal use of rolls wherever their use is not practically prohibited by special considerations affecting only certain shapes and classes of products. The action of rolls not only compresses the metal in a direction along the radii of the rolls, but the traction force of the rolls in pulling the metal through causes longitudinal tension in the surface layers. In case the section is intricate, necessitating deep indentations in the rolls, the sur- STEEL 409 face velocity is much less where the rolls are deeply cut than where the diameter is greater, thus tending to drag different parts of the section through at different speeds. This renders the distribution of the amount of working still less uniform and in extreme cases causes tearing of the metal. Reduction can take place only in a vertical direction, but this reduc- tion is always accompanied by a certain amount of expansion sidewise and a large amount of extension lengthwise. The amount of reduction in each " pass " through the rolls is extremely variable, running all the way from 5 to 50 per cent, but usually not averaging over about 15 per cent in each pass for such shapes as rails, structural sections, etc. The amount of extension lengthwise may be judged from the fact that a 3-ton ingot will produce two 90-pound rails each about 80 feet long, even after a considerable discard has been made to get rid of the pipe, etc. The speed of rolling practiced is very great, some passes in rolling rails being made at as high a rate as 10 miles per hour, while rods are sometimes rolled at a rate of 30 miles per hour. (This speed is so great that rods will be heated by the distortion in rolling" more rapidly than the heat can be radiated, and they are finished at a higher temperature than at the beginning of rolling.) 469. Rolling Mills. The most essential parts of a rolling mill are of course the rolls. Cast-iron rolls, which have been chilled to produce a hard exterior, and turned in a lathe to produce a smooth surface of the desired form, are very commonly used, especially for finishing rolls. High-carbon steel is also used when the rolls must be very strong, and even nickel-steel rolls have been used. All of the rolls except the ones used' for finishing have their surface roughened in order to increase their grip on the metal. Rolls are turned in an infinite variety of shapes, varying from the plain cyhnders used for plates and some rectangular shapes to the rolls used for structural shapes, rails, corrugated bars, etc., which may be quite intricate in form. All rolls except plain cylindrical rolls make pro- vision for several passes of the metal, each pass approximating the final form of the section desired more closely than the last. Rolling mills may be in general classed under one of three heads: " two-high " mills, " three-high " mills, and " universal " mills. Two-high mills consist of a single pair of rolls mounted in the same vertical plane. One variation of the two-high mill is the " pull-over " mill, whose rolls always run in the same direction, so that the metal after each pass must be pulled back over the top of the rolls to be fed in for the next pass. This is the simplest form of mill and the cheapest, but its operation is slow, and it is adapted only for rolling small shapes which 410 MATERIALS OF CONSTRUCTION can be readily handled. A more important type of two-high mill is the " reversing " mill, the rolls ot which may be made to run in either direction by reversing the engines which drive them. Successive passes are therefore made in opposite directions through the rolls. The two-high reversing mill is often used in " cogging " ingots. Fig. 230.— Three-high Plate Mill. The three-high mill has three rolls geared together, so that the metal may make one pass between the lower and the middle roll, and the next pass in the opposite direc- tion between the middle and upper roll, without reversing the rolls. A very large proportion of all steel shapes are rolled or at least finished by a three-high mill. The universal mill is provided with two auxiliary rolls mounted vertically just in STEEL 411 front of the horizontal rolls. The distance between the axis of these rolls is adjustable horizont;illy and they are designed simply to keep the edges of the metal smooth with- out effecting any reduction. Universal mills are made with vertical rolls on only one side of the horizontal rolls or on both sides, and they may be either two-high or three-high mills. All rolling mills which handle anything except very light material must be provided 412 MATEEIALS OF CONSTRUCTION with a. series of rollers in front of and behind the rolls, known as the " roll tables." The roll tables for three-high mills must be capable of being raised or lowered at the end next the rolls, in order that the metal may be directed between either the upper or the lower set of rolls. Two views of a three-high plate mill are shown in Fig. 230, and two views of a two- high universal mill in Fig. 231. The sketches of Figs. 232, 233, and 234 illustrate the application of three-high mills in the cogging of ingots and the rolling of special Mi Fig. 232.— Cogging Rolls. 470. Examples of Rol- lingPractice. Steel Rails are rolled by a process whose general features are as follows: The ingot, after reheating in a pit furnace, is cogged down to a bloom whose cross- sectional dimensions are about one-half those of the origi- nal ingot. This operation is accomplished by a series of about 8 or 10 passes through a set of cogging rolls whose form is that shown by Fig. 232. The bloom is now sheared at each end to remove the pipe and the ragged end formed by the rolls, cut in two in the center, returned to the reheating furnace, and brought again to the proper temperature for rolling. The nexi r«= Fig. 233.— Roughing Rolls. Fig. 234.— Finishing Rolls. series of 4 to 6 passes are made in one or more trains of roughing rolls (Fig. 233), and the last 8 or 10 passes in a train of finishing rolls (Fig. 234). (The figures here given are merely representative of average practice. The number of passes in each roll train varies considerably in the different rail mills.) STEEL 413 Fig. 235.- — Method of Increasing Areas of Rolled Shapes. Structural Steel Sections are rolled in almost exactly the same manner as are rails. The different sectional areas for a given size of any structural shape such as angles, I-beams, channels, etc., may be produced with the same set of rolls by simply changing slightly the axial distance between the finish- ing rolls as indicated by Fig. 235. Plates are cogged from the ingot in a mill similar to that shown in Fig. 231 except that the vertical rolls are not used. The resultant " slab " is sheared up into smaller slabs, the pipe discarded, and the slabs reheated before completing the rolling operation in a mill of the type shown by Fig. 230. The successive reduction in the thickness of the plate is accomplished by bringing the roUs slightly nearer together between passes. The surface is cleared of the scale and a smoother finish secured if salt or sand is thrown upon the plate from time to time. Such procedure results in carrying between the rolls some of the cooling water which is kept running on the rolls, and the formation of steam when the water is pressed on the hot plate causes sharp explosions which tend to clear the surface of the mill scale. ■Bods are rolled in a manner similar to that described for steel rails except that the original bloom is usually cut up into a number of small sections before reheating and rolling. A mill called a " guide mill " is used, the material after each pass being bent around and guided into the next pass by a device specially attached to the mill for the purpose. Wire-making is only a rolling operation so far as the making of the wire-rod is concerned. A considerable portion of the reduction of the section is accomplished by a special operation known as " cold-drawing." The wire rod is rolled in the manner above noted, its final diameter being usually from M to K inch. The wire rod is then wound into coils and pickled in a dilute solution of sulphuric acid, which removes the scale. Water is next sprayed on to wash off the acid, and this is followed by immersion in a bath of lime- water, which removes the last traces of acid. The coils are now dried in an oven and sent to the wire-drawing mill. Cold drawing consists in successively reducing the section, and extending the length, by repeatedly pulling it cold through tapered holes in a die or "draw-plate" (Fig. 236). Each hole through which the metal is drawn is somewhat smaller than the preceding hole, the average amount of reduction being from 20 to 25 per cent. Some thick lubricant is appUed to the draw-plate to reduce friction and prevent too rapid wear on the hole. Drawing is performed on " draw-benches " which com- prise a frame upon which the draw-plate is mounted and a reel which is driven by power and serves to pull the wire through the plate and coil it up. Often several plates are mounted on one bench, the holes in each being smaller than those in the preceding one. In this case a power reel must be placed between A ' ■'■/ © © © © @ @ @ @ @ © @ © (o)(2)@(2) @(o)(o)(o) Fig. 236. Draw-plate for Wire. 414 MATERIALS Or CONSTRUCTION each pair of holes, around which the wire is given a few turns, since the strength of the wire after passing the last plate would not suffice to pull it through several holes. The wire must be annealed by heating to a low red heat in a closed receptacle after each 3 to 10 passes, because of the hardening of the metal caused by drawing. The finished wire is also annealed unless it is to be sold as hard-drawn wire. Wire is often galvanized by drawing it first through a weak pickling solution to remove the scale, through a rinsing bath, and then through a bath of molten zinc. The excess of zinc is removed by drawing the coated wire through asbestos plugs. Iron wire is made from wrought-iron billets in exactly the same manner as steel wire, except that the production may be accomplished in fewer passes and with less difficulty caused by hardening. Plow-steel wire is simply a high grade, high-carbon wire, made from crucible steel. It is so called because it was originally used for dragging steam plows. Lap-welded Pipes are made by the method indicated by Fig. 237. The metal is first rolled into flat strips called " skelp," of the desired thickness, then bent to a U- shaped section and, by another pass, to a circular section with the edges overlapping. (Small pipe may be bent to the circular section by drawing the skelp through a die.) The metal is now brought to a welding heat and is passed through a pair of welding rolls over a mandrel" which is supported be- tween the rolls on the end of a long rod. A second pass through " sizing " rolls is made to insure accuracy of size. Butt-welded Pipes are made by drawing the skelp at a welding heat through a die or " bell " (Fig. 238), which welds the edges together without lapping (1> Tbc Skelp. '^jj^/^/WJ^j/M.V'/^j.:.':nj^. (IJ Weldin? RoUo. Fig. 237. Making Lap-welded Pipe. Fig. 238. Bell for Making Butt-welded Pipe. Seamless Tubes (small sizes only) are made either by forcing a flat plate through a cylindrical die by means of a mandrel, or by piercing a biUet longitudinally, expanding the hole by forcing larger and larger tapered expanders through it, and finally roUing over mandrels until the section desired is attained. Cold-rolled Steel is steel the last pass of which through the rolls is made with com- paratively cold metal. Pickling is necessary before cold rolling to remove the scale, and the result is a great gain in accuracy in size and form, and in surface finish'. The strength and elastic properties are also greatly improved by cold rolling. STEEL 415 Finishing Steel by Steam Hammer and by Presses 471. Forging linder the Steam Hammer. In the early days of the steel industry the steam hammer or its predecessor, the helve, was the principal means whereby steel ingots were worked up into the desired final form; but as ingots became larger and heavier, the hammer was to a great extent replaced by rolls which, although they do not work the metal so well, are much more rapid in operation and involve less expense for labor and for reheating. In working metals under the hammer the pressure of the blow acts for only a very short interval and the metal recovers somewhat from the effect. This results in the deformation produced not being in propor- tion to the instantaneous force applied, and the process is therefore slow. For small sections, however, and for the exterior portion of larger sections, the metal is worked better than by any other method of reduc- tion, resulting in the refining of the grain of the metal by decreasing the size of the crystals, and improving the quality of the metal in every respect. For thick sections a very heavy blow would be required to properly work the metal all the way to the center, but this means a very heavy hammer and very expensive foundations for the anvil. The practical limit of size is now considered to be not heavier than 50 tons, and few hammers are now built with capacities beyond 30 tons. For all heavier forging work the hydraulic press is preferred. The steam hammer now finds little application in the steel industry except in the forging of high-grade crucible or cement steel (where the value of the product and the especial desirability of giving the metal the fine grain attained by hammering justifies the higher cost), and in the production of that large class of articles called " drop-forgings," whose form is too intricate for rolling but which are subjected to condi- tions in service too severe to be met by iron castings or even steel cast- ings. This class of articles includes a great variety of machine parts, small tools, automobile parts, etc. Drop-forgings are made by the use of dies, between which the metal is worked into the desired form by the blows of a steam hammer. The dies are made of hardened steel, the impressions formed in the faces of the dies corresponding to the impressions formed in the mold for a cast- ing. The metal is placed upon the lower die, which is made fast to the anvil, and the upper die is carried by the head of the hammer. Very often a series of dies are necessary to complete a forging, each set approx- imating more closely the final form required. The stages in making a simple drop-forging are indicated in Fig. 239. The metal, in the shape of a bar of steel, is first upset to gain metal 416 MATERIALS OF CONSTRUCTION for the larger end of the forging and is then placed in the die (a), which roughly approximates the desired form. The partly formed piece is shown at (a') after removal from die (a). It is now transferred to the second die (6) which completes the shaping of the piece but leaves a " fin " all around, shown at (&')• This fin is removed by shearing m the die or punch (c), the forging being withdrawn through the slot below the die. It now appears as shown at (c')- The tool is now cut off from the bar, and is finished by grinding upon an emery wheel. (a) (C) Fig. 239. — Stages in Making a Drop-forging. The majority of all drop-forgings for machine parts, etc., are not forged to the exact size required, but an allowance' is made for machin- ing. This results not only in securing a better finish, but the removal of the exterior coat of scale may reveal the presence of flaws which have been hidden. 472. Forging by Means of Presses. The effect of pressing steel, by the action of large hydraulic presses, differs from that produced by the action of a hammer, in that the force applied acts for an appreciable inter- val of time, and the distortion produced extends deeper into the metal. STEEL 417 In consequence, the press produces a better crystalline structure than does the hammer for all except very thin sections, and is therefore used in preference to the hammer for all heavy forging. The press is able to do more work with the same power expenditure than is the steam- hammer, but cannot compete with rolls except where the superior work- ing of the metal secured by pressing is more important than relative costs. The hydrauUc press consists essentially of a hydraulic cyUnder in which a plunger or ram moves vertically, and which is forced down upon the metal supported on an anvil block or bed as in the case of hammers. Presses vary in size from a few tons capacity up to 14,000-ton armor- plate presses which will handle ingots weighing 50 tons or more. Fig. 240 is a view of an hy- draulic press forging a 15,000-pound ingot. The appliances whereby the ingots are moved forward and turned under the ram are shown. In addition to their use in the reduction of in- gots in place of a cogging mill, and in the pressing of heavy plates, hydrau- lic presses are largely used in the production of forgings pressed between dies, as in the case of the steam hammer, and also in the production of a large class of articles made of thin plate steel which is pressed cold between dies. A press used ioT the latter class of uses is called a " flanging press." Fig. 240. — Hydraulic Press Forging Ingot. Defects in Ingots and their Cokrection 473. Blow-holes, Piping, Ingotism, and Segregation. One of the greatest sources of trouble for the steel-maker is the occurrence of blow- holes, which are almost unavoidably formed in some degree while the metal is solidifying in the ingot mold. Blow-holes are caused by the 418 MATERIALS OF CONSTRUCTION presence of gases such as hydrogen, nitrogen, and oxygen which are held in solution by the metal when molten but released as the metal solidifies, and they are also caused by the presence in the metal of iron oxide, which upon encountering carbon, forms CO gas. The gas in blow-holes is usually reduding in effect and therefore the surfaces of the holes do not become oxidized and will weld together when the ingot is subjected to pressure in the operation of rolling or pressing. Blow-holes near the surface of an ingot, however, are apt to break through to the exterior, allowing oxidation and preventing perfect welding in the rolls. Another defect in ingots, which cannot be corrected in rolUng, is the occurrence of the yiye or shrinkage cavity which forms during solidi- fication. Since the metal cools first in contact with the walls of the mold, the interior will remain molten after an outside solid shell has formed. The interior metal contracts as it solidifies progressively from the out- side inward, causing the formation of a cavity which becomes filled with gases evolved during solidification. Since the hottest metal is at the top of the ingot, the upper portion remains molten longer and acts as a feeder to fill the shrinkage cavity in the bottom portion. The pipe is thus localized in the upper third of the ingot. This portion must be cut off in the rolling mill and goes back to the steel furnace as scrap. Ingoiism, the formation of large crystals of steel, caused by too slow cooling or casting at too high a temperature, is a serious defect in ingots which causes the steel to be weak and low in ductility. The bad effects of ingotism may be largely or entirely corrected by careful roll- ing or forging. The compression of the metal crushes and reduces the size of the crystals, imparting to the steel a much superior degree of strength and ductility. Care must be taken in the initial rolling or forging to avoid the formation of cracks which cannot subsequently be welded up. Segregation of the impurities in steel ingots is caused by the fact that most impurities, notably carbon, phosphorus, and sulphur, are less soluble in iron when solidified than while molten. In conse- quence, a part of the impurities in the iron are progressively rejected by each layer of metal as it solidifies, being absorbed by the still molten portion, the net result being a tendency toward concentration of the impurities in that part or parts of the ingot which solidify last. Segre- gation cannot be altogether prevented, but it may be lessened by the addition of elements such as aluminum or titanium, which have the effect of quieting the steel. Casting in narrow ingots is also effective, but is not practicable in all cases because it would take so long to cast STEEL 419 many small ingots from one large ladle that the first metal would be too hot if the last metal were not too cold. The Heat Treatment of Steel Practice of Hardening, Tempering, Annealing, and Case Hardening * 474. General. If steel containing 0.6 per cent or more carbon be heated to a bright red heat and then rapidly cooled, as by plunging into water, it becomes very much harder, but at the same time much more brittle. If the hardened steel be heated to a temperature between 200° and 300° C. it will become softened or " tempered " and the brittleness largely removed, the degree of softness being directly dependent upon the temperature of tempering. If the steel is heated to the hardening temperature and then cooled very slowly it will be " annealed," i.e., rendered as soft as that particular steel can be made to be. 475. Hardening Steels. Every carbon steel possesses a certain critical range of temperatures within which important molecular changes occur in cooling. In general this range is from a low yellow down to a dull red heat. (From about 900° to about 690° C. for steels containing sufficient carbon to harden usefully.) These molecular changes do not take place instantly, and they will therefore be retarded by shortening the time of passing through the critical range of temperatures. The amount of retardation is directly dependent upon the rapidity of cooling. If the cooling could be made instantaneous a maximum degree of hardness would be obtained, and lesser degrees of hardness are obtained by various rates of cooling after having heated the steel to just the proper temperature. For instance, some degree of hardening is obtained by quenching in molten lead; a greater degree by quenching in heavy oil; water makes it harder still; and so on, extreme hardness being obtained by quenching in ice-water, ice-brine, mercury near its freezing-point, etc. In all cases the degree of hardening obtained by any treatment will be dependent upon the amount of carbon present, and other elements, such as manganesCj chromium, tungsten, etc., have not only a marked effect upon the hardness obtained, but they may greatly change the location of the critical range of temperatures, in some cases reducing it below atmos- pheric temperatures so that hardening may be accomplished by air- quenching. 476. Tempering Steels. Hardened steel is too brittle and too fragile for most uses without some degree of "tempering." The fact that *The effect of various heat-treatments upon the constitution and properties of steels will be discussed at a later point. 420 MATEEIALS OF CONSTRUCTION tempering may be accomplished at temperatures far below the critical range is due to the circumstance that hardened steel is in a state that is not natural to it at atmospheric temperatures, i.e., it is not in equihbrium. It does not change its state simply because of the rigidity and immobility of the material, which does not allow sufficient molecular freedom for the alteration to take place. Only a relatively low degree of heating is sufficient to materially decrease this rigidity and increase molecular activity, resulting in a degree of loss of brittleness and hardness in pro- portion to the temperature attained, and, to a lesser extent, in proportion to the time allowed at that temperature. Very often the steel is again quenched after the desired tempering heat is reached in order to pre- vent overheating by heat conducted to the hardened portion from the balance of the piece. The degree of heating in tempering carbon steels is plainly indicated by a film of oxide which forms on polished surfaces and shows a succes- sion of colors dependent upon the temperature and consequent thickness of the oxide film. These colors merge into one another gradually, but those usually .distinguished, with the temperatures and the class of tools for which they indicate a proper degree of hardness, are as follows: Color. Approx. Temperature. Tools so Tempered. Pale vellow . . 220° C. 230° C. 243° C. 255° C. 265° C. 277° C. 288° C. 297° C. 316° C. Engraving tools, fine drills, etc. Steel-cutting drills, milling cutters, etc. Straw Light brown Dies, taps, rook-drills, etc. Wood-cutting tools, etc. it ti It ti Brown-purple Pale blue Screwdrivers, needles, etc. Wood-saws, springs, etc. 477. Annealing Steels. Annealing has for its purpose (1) the reliev- ing of any internal strains originating during cooling or caused by work- ing, (2) the restoration of the grain of the steel to the minute size which is so desirable, and (3) the softening of the steel after hardening. The usual annealing temperatures are between 200° and 500° C. The. heat- ing must be done very carefully and uniformly out of contact with the fuel, and the pieces must be supported so that they will not become distorted while hot. For any refinement of the grain the heating must be carried above the critical temperature above discussed. "Usually the steel is brought to a point Just above the critical temperature and then cooled very slowly either by leaving it to cool down with the furnace or by removing it and letting it cool under a muffle. Small STEEL' 421 objects are enclosed in iron boxes or pots, and packed in charcoal or similar material to prevent decarburization. In principle, annealing means simply carrying the tempering process to the extreme by heating it to a high degree and cooling as slowly as possible. » 478. Case-hardening Steels. Case-hardening is a form of cementa- tion applied to low- or medium-carbon steels in order to , impregnate them with carbon to a depth of perhaps one-fourth of an inch or less, thus securing a high-carbon case which may subsequently be hardened by quenching. The advantage gained by this treatment is that a sur- face is produced which will withstand wear, abrasion, cutting, or inden- tation and at the same time the core is left soft and tough so that the shock resistance of the material is not impaired. Case-hardening is especially applicable to the construction of armor-plate, safes and vaults, the moving parts of machinery which are subjected to both shock and wear, such as crank shafts, pivots and axles, gears, etc., also for the bearings and knife-edges of weighing machinery, and for many other purposes in machine and implement construction. For many of the purposes above listed certain of the alloy steels which combine great hardness with great toughness are now often preferred. The steel used for case-hardening usually is one containing from 0.1 to 0.2 per cent carbon originally, and the operation is usually applied to the finished casting, forging, or otherwise fabricated object, so that no machine work need be done on the hardened surface. The general method of case-hardening consists in heating the steel in contact with carbonaceous matter such as potassium ferro-cyanide, charcoal, barium carbonate, bone dust, charred leather, etc. The usual temperature of carbonation is about 900° C. A more rapid penetration of the carbon may be secured by the use of higher temperatures, but this practice is attended by the danger of the growth of coarse crystals in the interior at the temperature of carbonation, resulting in loss of toughness and strength unless the steel is subsequently reheated to restore the grain size. A time of from two to twelve or more hours is required for the process, depending upon the temperature of the fur- nace and the class of carburizer used. Many special or alloy steels are often treated by a case-hardening process with very beneficial results. Nickel and chrome steels are especially valuable steels for case-hardening. Occasionally it is the practice to secure the hardening of steel which has thus been surface carburized by quenching the steel as it comes from the furnace. A far better practice, however, is to allow it to cool, and then reheat and harden in the usual manner. This operation not only tends to remove any lack of toughness caused by coarse crystalline growth, but also causes a diffusion of the carbon inward, leaving a less 422 MATERIALS OF CONSTRUCTION distinct plane of weakness between the carburized shell and the unal- tered core. THE PROPERTIES AND USES OF STEELS Structure and Constitution 479. The Constituents of Steels. The normal constituents of steels and irons have already been declared to be ferrite, which is .theoretically pure iron entirely free from carbon, cementite, which is the carbide of iron, FesC, and graphite, which is practically pure amorphous carbon. We must now recognize the existence of various modifications and solu- tions of these constituents whose existence is dependent upon the amount of carbon present and the rate of cooling from solution. These modifications of the normal constituents can only be considered in connection with the consideration of the behavior of iron carbon solu- tions in cooling. 480. Compounds and Solid Solutions. Eutectics. Steels in general, like wrought iron and cast irons, are not purely chemical compounds of iron with carbon, etc., but are in the nature of metallic alloys consisting of an intimate mixture or solution of metals with non-metals, forming when melted a homogeneous fluid. This molten steel is composed of liquid carbide of iron dissolved in liquid iron * and as the solution cools and freezes the carbide and iron remain a solution forming when solid- ified a " solid solution." A solid solution differs from a mixture in that there is no separation of the constituents in freezing, and even the microscope is unable to distinguish the different components. The solution appears to be a simple homogeneous body like a chemical compound. On the other hand, a solid solution differs from a chemical compound principally because it may contain widely varying amounts of each component while a com- pound must always hold certain definite relative amounts of each com- ponent, and in some multiple of their atomic weights. There is a limit, however, to the amount of carbon which can be car- ried with iron out of the liquid state in soUd solution. This limit is reached when the iron contains about 1.7 per cent of carbon (Sauveur). No solid solution of carbon and iron will form containing more than about 1.7 per cent of carbon, and any additional carbon must remain out of solution either as cementite or as graphite and ferrite. All irons, therefore, which contain more than the amount of carbon which can be held in solution (more than 1.7 per cent C.) are rightly classed as cast * According to the views of a number of authorities steels are solutions of carbon in iron rather than solutions of carbide of iron in iron. STEEL 423 irons, or products intermediate between cast iron and steel, and not as steels. Eutedics. When a solution of two dissimilar constituents is cooled, the freezing-point of one or the other of the constituents will normally be reached before the freezing-point of the solution as a whole is reached. There will always be, however, a certain definite solution of the con- stituents whose freezing-point is reached before either of the individual constituents has begun to crystallize out. This particular solution is called the " eutectic solution " and the substance formed when the eutec- tic solution crystallizes out is called the " eutectic " or, in the case of soUd solutions, " eutectoid." Phenomena of Slow Cooling of Ibon-caebon Alloys 481. Freezing of Iron-carbon Alloys. If a molten alloy or solution of carbide of iron (cementite) in iron contains an excess of the carbide over the eutectic ratio, carbide of iron will crystallize out in cooling until the remaining solution is of eutectic composition, when it will all crystallize out together. Such an alloy is called a " hyper-eutectic " alloy. On the other hand, if an excess of iron is present " austenite," the soUd solution of iron and carbide of iron which is normal at all tem- peratures above about 880° C, will crystallize out and continue to do so until the remaining solution is of eutectic composition. Such an alloy is called a " hypo-eutectic " alloy. The eutectic formed is a saturated solution of " austenite " * and cementite whose composition is defin- itely fixed at 95.7 per cent iron and 4.3 per cent carbon. The freezing of iron-carbon alloys is depicted graphically by the dia- gram of Fig. 241. The significance of the diagram may be explained as follows: With any hypo-eutectic alloy the temperature at which the austenite will begin to freeze is indicated by the line AB, and the temperature at which the freezing of austenite is completed is indicated by the line AD. Thus if an alloy contains 1 per cent carbon, its austenite will begin to freeze at about 1430° C. and it will have become wholly solid at about 1200° C. Any alloy of carbon and iron will behave similarly (passing through a transition stage and finally becoming a solid solution of cementite and 7-iron) so long as the carbon content does not exceed 1.7 per cent. When the carbon content lies between 1.7 per cent and 4.3 per cent a selective precipitation of austenite occurs until the liquid solution becomes so impoverished in iron that its composition approaches the eutectic ratio, and final solidification takes place when the eutectic * Austenite is defined in Art. 483. 424 MATERIALS OF CONSTRUCTION is reached. Thus an alloy containing 3 per cent carbon will begin to precipitate austenite at a temperature of about 1250° C and all of the austenite will have crystallized out before a temperature of about 1135° (line DBE, Fig. 241) is reached. As the remaining solution becomes impoverished in iron and continues to cool, its position moves along the line AB oi the diagram until the eutectic composition is reached at B, when further cooling causes it to crystallize out of solution. With any hyper-eutectic alloy (more than 4.3 per cent carbon) cool- ing causes cementite to begin to crystallize out as soon as the line CB, Fig. 241, is reached. The liquid solution thus becomes progressively usoo _, 1000 jic. jlrejO. ^Fe. 100 3 45 97 96 Percentage Composition Fig. 241. — The Freezing of Iron-carbon Alloys. improverished in cementite, and with continued cooling its position'moves along the line CB until the eutectic is formed upon cooling below point B. (A number of authorities consider that the saturation point of carbon in iron is 2.2 per cent carbon instead of 1.7 per cent carbon. According to this view the point D in Fig. 241 should represent 2.2 per cent carbon, and the line of demarcation between cast iron and steel is at 2.2 per cent carbon.) 482. Changes in Cooling below the Freezing-point. When the solid solutions of carbon and iron are further cooled they are found not to be stable, but decompose at various temperatures, dependent upon the car- STEEL 425 bon content. Austenite decomposes into cementite and ferrite, and as the cooling continues the ferrite successively assumes two allotropic forms at temperatures below that temperature at which the 7 form of ferrite exists. These allotropic forms of ferrite are |8-iron and a-iron. The molecular change in iron passing from the 7 form to the^^jForm, and subsequently from the P form to the a form, is attended by the evolution of a certain amount oTTieat which causes a " retardation " in the rate of cooling. Furthermore, a third retardation occurs in iron- carbon alloys when the temperature is reached at which the residual solution of the original solid solution (after precipitation of ferrite or cementite) finally decomposes into its constituents, ferrite and cemen- tite. These three temperatures at which retardations occur in cooling are called the " critical points " in coohng. They are usually designated as Ai, A2, and ^3, respectively, in order of increasing temperature. The range of temperature included between points Ai and A3 is known as the " critical range " of temperatures. Exactly the reverse phenomena occurs in heating steel through this same range, a sudden absorption of heat and a consequent falling off in the rate of heating being observed at each critical point. The critical points on heating are slightly higher than the corresponding critical points in cooling, owing to a lag or hysteresis effect, the physical change not occurring immediately when the temperature which should produce it is reached. The critical points in heating are usually distinguished from the corresponding points in cooling by designating the former Aci, Ac2, Ac3 and the latter An, Ar2, Ara* As the amount of carbon in the steel increases, the Ars and Ar2 points approach each other until they coincide, when the carbon content is about 0.5 per cent, and the critical point is then designated Ar3,2- Further increase in carbon to about 0.85 per cent causes the Ar3,2 point to coincide with An and the result- ant single critical point is designated Ar3,2,i- The approximate position of the critical points in cooling various steels is indicated by the time-temperature curves, or cooling curves, of Fig. 242. • By observing the position of the ciitical points in cooling for a con- siderable number of steels whose carbon content varies over a wide range, the complete decomposition diagrarn of iron-caubon solid solutions shown in Fig. 243 may be derived. The significance of this diagram may be explained as follows: Consider first a hypo-eutectoid alloy containing 0.2 per cent car- *r stands for the French refroidisaement, meaning cooling, and c for the French chauffage, heating. 426 MATERIALS OF CONSTRUCTION . i < A * / ,■!• ■3 1 / f / / f> S ^ ^ / ^ -^ / i /I <,-' / '* <, 7^ \ 't ) V -8- hH ^ iS g ^ d. -^ <* ro > ~ iC ^ -^ ^ — — - CO ^ ~ 1^ "1 i i \ : 1 S g % § 3 3 u g d " o o ,a o 5 s t1> 3 rrt a o ^. h h ^ ^ ■epBjSiijnao S39j3aid matHx «ctoid matrix t ctoid Ferrit9+ Cement te Irons) 3 4, 4.3 5 6 6.67 45 60 75 90 100 97 96 95 JM; 93.33 Percentage Compositioa Fig. 243. — Solid Solution Decomposition Diagram. about 900°. Cementite now begins to be precipitated from the solid solution and continues to crystallize out until the remaining austenite has reached the eutectoid composition, when it again becomes trans- formed into pearlite. Lastly, considering the changes which occur upon cooling the eutectic alloy of cementite and austenite below its freezing-point (about 1135° C), we find that pro-eutectoid cementite (cementite whose formation immediately precedes the formation of the eutectoid) first forms, followed by the decomposition of a part or all of the cementite into ferrite and graphite. This latter conversion process is greatly affected by the pres- 428 MATERIALS OF CONSTRUCTION ence of the various impurities such as silicon, manganese, phosphorus, and sulphur, which are always present in cast irons, and the precipita- tion of graphite in particular is especially affected by the rate of cooling. Cast Irons to. "0 i V H 15 ^ ■" ~* Fig. 251. — Distribution of Elongation. 442 MATERIALS OF CONSTRUCTION nomenon of necking down, the action being most pronounced in the case of medium-carbon steels. The fact that steel test specimens are reduced in cross-section as they are elongated under stress (in obedience with the law expressed by Poisson's ratio), and finally neck down, accounts for the fact that the breaking load, according to the stress-strain diagram, is below the ultimate strength. If the stress were at all loads computed upon the basis of the actual section then existing, instead of being computed (as it always is) upon the basis of the original section, the stress-strain curve would follow such a course as is indicated by the dotted line of Fig. 250. The modulus of elasticity can be determined from the slope of the first portion of the stress-strain diagram, but is usually computed from the ob- served values of stress and strain. The modulus of elasticity is strangely unaffected by the factors which influence other properties of steels and will, as above stated, be found in the neighborhood of 29,000,000 to 30,000,000 pounds per square inch for practically any class of carbon steel. The tensile properties of various steels called for in the standard specifications of the American Society for Testing Materials are sum- marized in the following table : TENSILE PROPERTIES OF VAUIOUS STEELS Character and Use of Steel. Rail Splice Bar Steel [Low carbon Medium carbon . High carbon I Extra high carbon . 55-65000 68000 85000 100000 I Structural . Structural steel for bridges j L Rivet . 55-65000 46-56000 f Structural . Structural steel for build- J '"63 l Rivet.... Structural steel for locomotives. Structural steel for cars Structural . Rivet and flange 1 plates j Structural steel for ships i f Structural . i Rivet . Tensile Strength Lbs. per Sq.in. 0.5 (u.t.s.) 0.5 (u.t.s.) 55-65000 46-56000 65-65000 50-65000 48-58000 58-68000 55-65000 Yield Poiut. Lbs. per Sq.in. 0.5 (u.t.s.) 0.5 (u.t.s.) 0.5 (u.t.s.) 0.5 (u.t.s.) 0.5 (u.t.s.) 0.5 (u.t.s.) 0.5 (u.t.s.) Elonga- tion. Min. in 8 ins. Per Cent. 1500000 u.t.s 1500000 u.t.s. 1400000 u.t.s. 1400000 u.t.s. 1.500000 u.t.s. 1500000 u.t.s 1500000 u.t.s 1500000 u.t.s. 1500000 Elonga- tion. Min. in 2 ins. Per Cent. 1600000 u.t.s.* 14 10 22 22 Reduc- tion of Area. Min. Per Cent. * u.t.s. stands for " ultimate tensile strength." STEEL TENSIEE PROPERTIES OF VARIOUS STEELS— (Cora.) 443 Character and Use of Steel. Tensile Strength. Lbs. per Sq.in. Steel Forgings (7 classes of f Lowest grade, un- carbon steel forgings are treated material recognized and different i Highest quality specifications are given I hardened and for each class).* ( tempered , Locomotive and car forg ings, hardened and tem- pered. Axles, shafts, etc. Max. thickness 4" Thickness 4"- 7' Thickness 7"-10' Thickness 10"-20" Locomotive forgings un- j treated f Max. thickness 8" Thickness 8"-12" f Max. thickness 8" Locomotive forgings an- Thickness 8"-12" nealed. i I Thickness 12"-20" Cold-rolled steel axles. -60000 90000 0.5 (u.t.s.) 55000 90000 85000 85000 82500 75000 75000 0.5 (u.t.s.) 0.5 (u.t.s.) 80000 80000 80000 70000 Yield Point. Lbs. per Sq.in. Elonga- tion. Min. in 8 ins. Per Cent. 55000 50000 50000 48000 0.5 (u.t.s.) 0.5 (u.t.s.) 0.5 (u.t.s.) eoooot Elonga- tion. Min. in 2 ins. Per Cent. 1500000 u.t.s. 2100000 2100000 u.t.s. 2000000 u.t.s. 1900000 u.t.s. 1800000 1600000 2500000 u.t.s. 4000000 4000000 u.t.s. 3800000 u.t.s. u.t.s 1500000 1800000 u.t.s. 1725000 u.t.s. 1 «50000 IS Reduc- tion of Area. Min. Per Cent. u.t.s. 3600000 u.t.s. 3400000 2200000 u.t.s. 2000000 u.t.s. 2640000 u.t.s. 2400000 35 Steel castings ( Hard. . . . \ Medium . ( Soft soooo 70000 60000 36000 31500 27000 15 18 22 20 25 30 f Flange . Boiler and firebox steel I Firebox. 65000 62000 0.5 (u.t.s.) 0.5 (u.t.s.) 1500000 u.t.s, 1500000 u.t.s. Boiler rivet steel . 45-55000 0.5 (u.t.s.) 1500000 Billet Steel Concrete Reinforcing Bars Plain Bars Deformed Bars Structural grade . . . , Intermediate grade. . Hard grade Structural grade ... Intermediate grade. Hard grade 55-70000 70-80C 80000 55-70000 70-85000 80000 Cold twisted bars.. Concrete reinforcing bars i from reroUed steel rails | f Plain Bars . Deformed bars . 80000 80000 33000 40000 50000 33000 40000 50000 55000 1400000 u.t.s. 1300000 u.t.s. 1200000 u.t.s. 1250000 u.t.s, 1125000 50000 50000 u.t.s. 1000000 u.t.s. 5 1200000 u.t.s, 1200000 * See Am. Soc. Test. Matrls. Year-book, 1914. t Elastic limit, not yield point. 444 MATERIALS OF CONSTRUCTION 90,000 80,000 , c a 70,000 M 2 "1 00,000 S. UU, Stren rtb 1=62,476 f/ Sq.Ii, 1 .60,000 4 10,000 T fint ioi _ ^ _ 1 ,^ ^ -L r ^ \ i^ r/ '' g 30,000 ; ^ .M.i trt Qgth = 3l.l .Si / 3q iJ %oiitt.\ =3 .6005 /s,. In TENSILE AND COMPRESSIVE STRESS-STRAIN DIAGRAMS FOR DEAD SOFT O.H. STEEL C— 0.0B* 20,000 ■s a 1 = i 1" 4- am jri BS on 'Si m *~ iij "ii.o'l! 1 1 1 .01. OSi. 03.04.06 .10 .16 Strain-Inches per Inch Fig. 252. 487. Effect of Carbon upon Physical Properties. The distinctive properties of tiie different grades of ordinary steels, (i.e., not special alloy steels) are due to variations in carbon content more than to any other single factor. Car- bon always acts as a hardener and strengthener, but at the same time reduces the ductility. The effect of car- bon upon the tensile and compressive prop- erties of steels is shown graphically by the series of stress- strain diagrams of Figs. 252 to 261 in- clusive. These curves have been plotted from the data of tests made at the Watertown Arsenal.* The steel is in all cases an open-hearth steel, the various test specimens having carbon percentages of 0.09, 0.20, 0.31, 0.37, 0.51, 0.57, 0.71, 0.81, 0.89, and 0.97 per cent respectively. The tensile specimens were 1 inch in diam- eter and the elon- gation was measured on a 30-inch gauged length: the compres- sive specimens were also 1 inch in diam- eter and were 12 inches long, the com- pression being meas- ured on a gauged length of 10 inches. It should be noted that the elongations of the tensile specimens appear low only because the long gauged length caused the great extension of the portion of the immediately ad,ioining the break to have lesser weight * " Tests of Metals," 1886 and 1887. ; ~ a 1-" " ~ 10,000 t It. St rei &" -6 S,3 51 */ ii , '1' en sic n -o- — — — — r ■S ^ ^ t. ^50,000- ^ — — ~ — — .2. t ^ l _ ~ ~ £ 40,000 1° 4 tJ Btr en ^ =1 "30.19:0 #/ Sq I a. ^ (_ " ■^ *i 30,000 d * field ,Pt!=3 1.000 V ^q. In ^0 TENSrLE AND COMPRESSIVE 6TRESS-STRAIN DIAGRAMS FOR SOFT O.H. STEEL c=o.aoi( ~ § 20,000 1 )n] pr es 31 80,000 _ t^ Ult. itrUgtll =\ SO G0( Tei^ sion , — ^ ■n Sq .In. S70,000 f 60.000 ^ f^ ^ f, / n y ^ 50,000 i J '•f / n Ul t. Strength' = ts WoifpS q- n 140,000 S 30,000 20,000 jO.OOO " .!^ ,^ ie di>ttli,5;00 (/Sq. In \ ^ s fl i c >1I] pi es sic a TEN SILE AND COMPRESSIVE ESS-STRAIN DIAGRAMS ± « l" ** C=a31* ^ t1 UJ tij .o|l.02 .10 .15 Straia-Inctaes per Inch Fig. 254. the total percentage elongation after fracture, than if this same local extension had been only averaged with the extension of the balance of the length of a standard 8-inch specimen. It should also be noted that the compressive specimens were of such a length, with respect to their transverse di- mensions, that failure invariably occurred by triple flexure at a load only slightly above the yield point. If these specimens had been very short, the com- pressive stress-strain curve would not differ greatly from the tensile curve. The direct relation- ship between carbon content and tensile strength, as shown by the series of tests above quoted, is summarized in Fig. 262. The relation between ultimate tensile strength and per cent of carbon is expressed with a fair degree of accuracy by the equation: Ultimate tensile strength = 45,000-|-115,000 (per cent carbon), and the relation between the yield point and the per cent of carbon is similarly expressed by the equation: Yield point = 30,000-1- 50,000 (per cent carbon). The relationship be- tween elongation and ultimate tensile strength, as shown by the same series of tests, is pre- sented by curve /, Fig. 263. This same relation- ship when only a length of 8 inches is considered (the extension of the four 1-inch intervals on either side of the fracture being added and divided go.ooo d i 1 en bio n Ult ,St re I Kt h= 85 11) U/ Sq In 80,000 — — " N y' \ / 1 / " 80,000 3 / / ft 180.000 Al V- t.E tre HB th =6 0,8 ■5 /S [1.1 1. pD ■* i' 40,000 '^ -1 Yi eld P i= 47, poo »/ pq In ? 1 30,000 (0 ^ C im^re 891 >ii s = y d TENSILE & COMPRESSIVE 6T.E1ES6-8TRAIN DIAGRAMS FOR MEDIUM HARD O.H. STEEL C = 0.37)S g* g t P> u n ■- 10,000 vs. i. " ,01.0- 2.0 3 .10 .IS Strain-Inches per Inch Fig. 255. 446 MATEEIALS OF CONSTEUCTION . 100,000 s m .Si re Ifft hO p.7 60 f/s J n. Te nsi Ln.., 1—1 N to 000 5 ^ \ a ' / ^ / ni / o •O / / m f I It. Str en Btt = SjOOOi'/Sq. [n. hi M ^ "' i 60,000 Yi Cld Pt ^ p7. OOO *l Sq In i " iO,000 m "f i" i 3o miJ reE sio n 30,000 1 f TENSILE & COMPRESSIVE STRESS-STRAIN DIAGRAMS FOR HARD O.H. STEEL 0=0.51^ - - 20.000 J- ^r 10,000 ^1 1 1 J c .^^0^..^3 — — — '~ ^ 1 1 1 1 J Ult. Str^ngth^ll7,«0jP/| -- ■ — ^ — ^ 1 3q. In. .^ Tension / J _-i / 1 y / f 1 d - ; 1 /jieid i^r 65 O00#/Sq. In .. Jtnt strength = ^65 at^sji. n. ,- • , f yiel4'Pt=f 5,300 #/ Sq In. n ' i ■^ C onipredslba «^ 40.000 B" ^ ^ 3 s rsT i ^ TENSILE AND C0MPRES6IVE STRESS-STRAIN DIAGRAMS FOR HARD O.H. STEEL C= 0.57 % i 8 10,000 II f^l ^ 1 ,.0i.0'2.03 _ _ _j .01.02.03.01.05 .10 .15 Strain- Inches per Inch Fig. 256. .01.02.03.01.06 .10 .15 Strain-Inches per Inch Fia. 257. TEN6ILE d COMPRESSIVE STRESS-STRAIN DIAGRAMS FOR HIGH CARBON O.H. STEEL -- =f 10 000 ^ ^ 1 J C=-0.8Bj6 " ~ ■»** M 1 1 1 1 .01.0^.03.04,05 .10 .15 Strain-Inches per Inch Fig. 260. 150,000 - ■■V mb. Strenglth= 152,660 '*/Sq l£ / ' Tensicln 130,000 - / ~ "~ ~ ~ ■~ ~ ~ 120,000 - ~ - 110,000 - 100.000 - d giOO 000 - -I J 4 - -jr ■^mt. Strenetli=91,5001#/Sq. li. - fe 80,000 , ^70,000 i 60,000 60,000 40,000 30,000 i 20,000 ; i 10,000 - t i r \ A fie d Ptii8i,000#/Sq. n. ~ fie d Pt.=79,0(Jo#>Sil, In. ~ ~ ■ J ~ — — — ~~ — ' Compression ' d ri , ~ ~ , CO t\ ^ - i- s ~ i ^ TENSILE AND COMPRESSIVE STRESS-STRAIN DIAGRAMS FOR HIGH CARBON O.H. STEEL C=0.97!t i~ II J lU ~ — _ .0 .0I2.OH " — . 01.02 M .04 ,06 .1( .1 i ■ Strain-Inches per £n£li Pig. 261. M 1 1 1 M 1 1 1 1 1 1 1 RELATION BETWEEN TENSrLE STRENGTH AND YIELD POINT "^IflOOOO \^ (V-'^ ^ S120 000 ftf ^ ^ 13^^110,000 -ff ^ r ^ ■S a 90,000 o '^ 80 000 i^ .»< ^ iv- ix ■^ ■Yt tl. ^W 7oioOO 1= 60,000 P a 60,000 ti 40,000 30,000 20,000 -r\\ [^ C iot -J^ — I ^ d-" 50 M ^ " ^ k " ti» ^» 2- ■i o Iv r- ^ *■' .40 .50 .60 .70 .80 .90 1.00 I.IO 1.2 Percentage o£ Carbon./ Fig. 262. by eight) is shown by curve II. The equation of this curve is expressed by the formula: Percentage elongation in 8 inches = 58— (ultimate tensile strength ) 2500 448 MATERIALS OF CONSTRUCTION For purposes of comparison curve III, which is expressed by the rule given first by a committee of the American Society of Civil Engineers, and used in many specifications, namely: 1,500,000 Percentage elongation in 8 inches = is plotted on the same figure. ultimate tensile strength' \' \ V, \ II ri. 180,000 "> 110,000 ^ o \ V S- N ' -V ^-^ \ >;1 120,000 5 110 000 \ ^ ^ ii ^ T 'fl S 100,000 J 90,000 Ifi 80,000 S 70,000 J V "S aa nt m '"( 'it *^^ k •>a '■»'«• "v. 1 "•i &- p--. ks «,?r fti '» nr- , 1 Vifi n RELATION BETWEEN ELONGATION AND ULTIMATE TENSILE STRENGTH K -Cj -"tr. ns. Str .^^ — d ^ tl g 50 000 ' ■> ^V^ Sm^ So 44^ ■ S 30,000 S i!«,000 p 10,000 Per Cent Elongation Fig. 263. ^f The marked effect of gauged length upon percentage elongation above referred to is shown by a comparison of curves I and //, the former being plotted from the elongation in the entire gauged length of 30 inches, while the latter is based upon the elongation of only a portion of the speci- men which was originally 8 inches in length, the portion taken being so chosen that it included the break at its center. This relationship is shown also by the series of curves plotted in Fig. 264, which have been based upon a . ^, ^ , , ^ , series of tests of open- hearth steel made by the author. These tests were made upon 1-inch round bars 8 feet long, gauge points having been marked at 1-inch intervals along 8 .10 IS U 10 18 ao 22 24 26 US so S2 34 38 38 40 42 14 M 48 50 Per Cent GlonKstlon Fig. 264. STEEL 449 the entire length of each specimen. Each of the gauged lengths over which measurements were taken after fracture, included the fracture and necked-down portion at the center, or within 5 inch of the center. 488. Effect of Heat Treatment and Mechanical Working. The effect of hardening, tempering, and annealing of steels upon structure, hardness, and brittleness, has been noted above in Arts. 474 to 478 and and in Art. 483. The effect of the same treatment upon the strength and ductility is best shown by the investigations of Brinell.* The curves shown in Figs. 265 and 266 constitute a sunmiary of a portion of BrinelFs experiments. The steel used was an acid open-hearth steel, hot rolled and then subjected to the various heat treatments indicated. The specimens were round bars, 18 millimeters in diameter (0.71 inch), and elongations were measured on a gauged length of 180 millimeters (7.08 inches). The steels were of average high quality and composition in all respects except that the silicon was unusually high. The limits of composition were as follows: Silicon 0.266 to 0.453 per cent (except for the lowest carbon steel, which contained only 0.005 per cent), sul- phur 0.01 to 0.02 per cent, phosphorus 0.025 to 0.030 per cent, and man- ganese 0.01 to 0.49 per cent. A study of the curves leads to the follow- ing conclusions: Annealing at 350° C. has little effect upon the strength of low- and medium-carbon steels, but increases the strength of high-carbon steels slightly; the ductility of all steels is increased slightly. Annealing at 750° C. has little effect upon the strength of low- and medium-carbon steels, but decreases the strength of high-carbon steels slightly; the ductility of all steels is increased considerably. Annealing at 850° C. has little effect upon the strength of any class of steel, but increases the ductility considerably, particularly the high- carbon steels. Annealing at 1000° C. has practically no effect upon the strength of low- or medium-carbon steels, but slightly increases the strength of high- carbon steels; the ductility of all steels is increased slightly, the low- carbon steels more particularly. Water quenching at 750° C. increases the strength of low-carbon steels greatly, medium-carbon steels slightly, and scarcely affects the strength of high-carbon steels; the ductility of low-carbon steels is reduced very greatly, medium-carbon steels slightly, and high-carbon steels are scarcely affected at all. Reheating to 550° C, after water quenching at 750° C. restores the steel to practically its original strength in all cases; the ductility is in every case very much greater than that obtained after quenching, but * Journal of the Iron and Steel Institute, 1901. 450 MATERIALS OF CONSTRUCTION is only slightly greater than that obtained with the steel in its original condition. Water quenching at 850° C. increases the strength of low-carbon and medium-carbon steels very greatly, but scarcely increases the strength Fig. 265.^Effect of Various Heat Treatments upon Tensile Strength of Various Steels. of high-carbon steels because of the increased brittleness; the ductility is reduced enormously for the low- and medium-carbon steels and is practically zero for high-carbon steels. Reheating to 550° C. after water-quenching at 850° C. largely nulli- STEEL 451 fies the gain in strength of low- and medium-carbon steels caused by- quenching, and enormously increases the strength of high-carbon steels* by removing brittleness; the ductility is restored to practically its original value before quenching. 36 30 24 CJ 20 16 1 %6 d 3 •p ■3? 1. 1 2 1 In 2 tp & 5- i 1 1 5 3 g 0.0!) ® ^ »« 1^ !1 ft I9 |.e cafe 0'*i a- ^9 / ^ jO.09 ^ ^ _c=o.o9ie "o.ou / =7 > U.IO 0.16^ I'o.OO ^1 V J.25 ^ / "■ ®\ / ^ y O.Mj 0.25 l\\ — io.25 / / tsr- 0.31 /A.! W \ / /0.34 , ,0.25 f / <5.« / /0.16 / - 'i.ii s (0.44 k4 / / // *(■*% c^o^ ^ -Ho.M/ / / // - 40.41 / / // V iA65 1 il.l7 0.09 0.05^ O.IS/' \ \ \ U 0.09 / \ fifiif—f 40.b5 cTro Ucij 0^ M \ W / /> ^^» ^ ^"^ Ofiif \ 0.05 0.91 J > M 1 W * M\ W / 0.S4 1/ Xo'^'^ - <»t^ y 1 1.17 \\ ,0.94 / \v- /J \///. \\ ^ \ \ ^.94 \\' Uko.ie/// \ ^ \i il/ // A \ 1.17 ''i.n \ j:-^ i.ir \ \l /#' Heat-Tr eatment \W/f0.34 0.91^31^.44 Fig. 266.— Effect of Various Heat Treatments upon Elongation of Various Steels. Oil quenching at 750° C. followed by reheating to 550° C, has little effect upon the strength of any class of steel, but increases the ductility of low- and medium-carbon steels slightly, and high-carbon steels con- siderably.^ 452 MATERIALS OF CONSTRUCTION Oil quenching at 850° C. followed by reheating to 550° C, increases the strength of all except the very low-carbon steels remarkably; the ductility of any grade of steel is not materially affected. The effect of hot mechanical working upon the structure and char- acteristics of steels has been mentioned in connection with the discussion of rolling, hammering, and pressing of steel. The first effect of working is the benefit derived from the elimination of flaws, blow-holes, etc., which become closed up. The coarse crystalline structure of steel slowly cooled from a high temperature is also improved by working, since the crystals become broken up, mixed intimately, thus destroying the continuity of their cleavage planes, and compacted together, thus increasing both their cohesive and their adhesive power. The amount of reduction necessary in rolling or forging the finished section from the ingot is dependent on many variable factors. It is a fact recognized in the ordinary practice of steel mills, however, that the finished section should never be more than 10 per cent of the ingot sec- tion, and it is commonly not more than 2 or 3 per cent, often being much less than 1 per cent. The temperature at which working is finished is a very important consideration, since, if this temperature is above a red heat, the crystals grow to a certain extent, thus diminishing strength and especially lower- ing the elastic Umit of the steel. In the ordinary practice of rolling structural steel the finishing heat is above the red heat and the elastic limit is therefore comparatively low. If the working be continued until the metal is not above a dull red heat the material is immobile, large crystals cannot reform, and strength and especially the elastic limit are greatly increased. Cold working of steels, i.e., the mechanical distortion of the metal below the critical range of temperatures, cannot be practiced except with low- or medium-carbon steels, because high-carbon steels are too deficient in ductility and too brittle to be worked below the critical range. The effect of cold working upon the existing structure (made up of pearl- ite with free ferrite) is simply to elongate the crystalline elements in the direction of working. Cold working does not. improve the crystalline structure as does working above the critical range, and the primary effect of cold worlcing upon physical properties is a marked decrease in ductility and increase in brittleness. One other effect which constitutes a great practical advantage in the case of steels used for certain pur- poses, is the very material extent to which the elastic limit of steel may be raised by cold working. This fact is taken advantage of in the man- ufacture of certain grades of hard wire (unannealed after completion of STEEL 453 the drawing process), and in the finishing of steel rods intended for par- ticular purposes such as concrete reinforcement. In the latter case the cold distortion of the metal by special rolls which deform the bars and produce alternate depressions and elevations, or by cold twisting of square bars, is commonly practiced. The purpose of such distortion is primarily to increase the hold of concrete on the steel by providing a mechanical bond, but this purpose could more cheaply be served by hot working were not the distinct advantage gained by cold working recognized. The extent of the effect of cold working is directly dependent upon how far below the critical range working is continued, and is most marked when working is done at atmospheric temperatures. 489. Effect of Silicon, Sulphur, Phosphorus, and Manganese. The direct effect of silicon, in the ordinary proportions commonly encountered in steels (usually not over 0.2 per cent) , upon strength and ductility is very slight. Whatever effect there may be is difficult to determine, because it is masked by the influence of other elements, like carbon and phosphorus, which cannot be made to be altogether constant factors, and which are much more influential in this respect. Increasing the silicon content intentionally to 0.3 or 0.4 per cent has the effect of raising the elastic limit and ultimate strength of the steel considerably, without reducing the ductility greatly. This is sometimes done in the produc- tion of steel castings. Special silicon steels containing over 1.0 per cent of silicon will be mentioned later. Sulphur, within the limits common to ordinary steels (0.02 to 0.10 per cent) has no appreciable effect upon the strength or ductility of steels. It has, however, a very injurious effect upon the properties' of the hot metal in lessening its malleability and weldability, thus causing difficulty in rolling called " red-shortness." If it were possible that the steel might contain an excess of sulphur over that which is neutralized by manganese, the effect would certainly be to reduce both strength and ductility. Specifications for structural steels commonly limit the sulphur content tq a maximum of from 0.04 to 0.05 per cent. Phosphorus is the most undesirable of all the elements commonly found in steels. Its effect upon the properties of steels is very capricious, but it is always detrimental to toughness or shock-resistance, and often detrimental to ductility under static load. Campbell states that the strength of steel under static load is increased by 1000 pounds per square inch for each 0.01 per cent of phosphorus so long as the total phosphorus does not exceed 0.12 per cent. Beyond this limit even static strength is diminished. Phosphoric steels are apt to break under very slight stress if this stress is suddenly applied or if vibration is .iSP:n 454 MATEEIALS OF CONSTRlHiPiSr encountered, and this fact alone is sufficient to bar phosphoric steels from most uses in construction. The loss of ductility, due to increased brittleness, is not always noted in tests of high-phosphorus steels, and it is therefore difficult to detect the presence of too high phosphorus by the ordinary methods of mechanical testing. Specifications commonly limit the phosphorus content of structural steels to a maximum of from 0.04 to 0.06 per cent. Manganese is an element which is commonly comparatively high in most steels because of the prevalent practice of using either spiegel- eisen or ferro-manganese as a recarburizer and deoxidizer in the Bessemer and open-hearth processes. The effect of manganese upon the proper- ties of steel is a rather involved question, but it has a tendency to in- crease the strength provided a certain limit is not exceeded. If this limit is exceeded its effect in the direction of increasing brittleness causes a reversal of its effect upon strength. With less than 0.3 or 0.4 per cent manganese the steel is apt to be impregnated with oxides whose harm- ful effect outweighs any beneficial effect due to the manganese. Between 0.3 or 0.4 per cent and about 1.5 per cent manganese, the beneficial effect is dependent upon the amount of carbon present. With 0.1 per cent of carbon (according to Bradley Stoughton), the strength is in- creased about 100 pounds per square inch for each 0.01 per cent of man- ganese over 0.3 or 0.4 per cent; with 0.2 per cent carbon, the gain is about 165 pounds per square inch for each 0.01 per cent of manganese; and with 0.4 per cent of carbon the gain is about 280 pounds per square inch for each 0.01 per cent of manganese. In all cases the beneficial effect is somewhat more marked with acid than with basic steel. As the content of manganese rises above 1.5 or 2.0 per cent the metal becomes so brittle as to be worthless, but as the content of manganese is further increased a curious reversal takes place with about 6 or 7 per cent of man- ganese. The properties of this special manganese steel will be discussed later under the head of special or alloy steels. 490. Shearing Strength. Direct Shear. Torsion. The shearing strength of steel is dependent to a very large extent upon the same factors as the tensile and compressive strength. It is therefore to be expected that a more or less constant relation may be found between' the shearing strength and the tensile strength. Experiments have amply demonstrated that such a relationship does exist, and have further shown that the value of the factor for most grades of steel is about 0.8, i.e., the shearing strength is about eight-tenths of the tensile strength. This fact is well shown by the appended table abstracted from the tests of Prof. A. B. W. Kennedy:* * Proc. Inst, of Mech. Engrs., 1885. STEEL 455 RELATION BETWEEN SHEARING AND TENSILE STRENGTH ^^^P Kind of steel Ult. Ten3. Strength Lbs. per Sq.in. Shearing Strength Lbs. per Sq.in. Ratio Onpn-hearth 57,000 63,500 64,000 69,000 71,000 78,000 82,000 118,000 47,500 51,000 52,000 56,000 51,000 64,000 59,000 79,000 0.83 a It (I " RpsRpmer fhard^ 0.80 0.81 0.81 0.72 11 « II « 0.82 0.72 0.67 Torsional stresses are nothing more than shearing stresses which, instead of being confined to a single section or a limited number of isolated sections of a specimen or part of a structure or machine, are common to every section between the planes of the external forces which produce the twisting moment called the torque. The intensity of the shearing stress on any section of a shaft is not uniform, however, but varies directly as the distance from the axis of the shaft (considering that the section is circular), and the maximum intensity of shearing stress is therefore found at the circumference. The mathematical expres- sion for the maximum shearing stress in the extreme fiber of circular shafts is: _2Pa ' (wTierein Pa is the torque, r is the radius of the section, and si the shear- ing stress in the extreme fiber of the shaft), and the shearing modulus of elasticity is expressed by the equation rd (wherein I is the distance between the planes of the external forces, and e is the angle of torque or twist). From the following table, which is a summary of a series of tests made by the author, it will be noted that the shearing strength as determined by torsion tests appears higher than it does in the case of direct shear. This is due to the fact that the expression for si above is true only as long as the material behaves elastically. When computed for the torque which produces rupture it is called the " torsional modulus of rupture " and corresponds to the true value of the shearing strength of the material only as the modulus of rupture corresponds to the actual extreme fiber stress. (See Art. 374.) 456 MATERIALS OF CONSTRUCTION Upton * has shown that a definite relation exists between the shear- ing modulus of rupture and the direct shearing strength, the relation being that expressed by the graphical construction shown by Fig. 267. On this diagram the curve OEA expresses the relation of torsional shear- ing stress to torsional strain and is the usual form of graphical representa- tion of a torsion test. The dotted line which represents the direct shear- ing stress (the true shearing stress in the extreme fiber of the torsion specimen also), is obtained by the following construction: At any point A on the curve draw a tangent and prolong same to. its intersection with Tangent of Helix Angle= -III I Fig. 267.— Relation between Torsional and Direct Shearing Stress. (Upton.) the axis OB. Lay off from A a distance A A' equal to i OB. The point A' will he on the curve of true or direct shearing stress, as Upton has shown by mathematical analysis. The curve EA' may thus be con- structed point by point. The most significant fact illustrated by this diagram is that as the latter portion of the curve becomes more and more nearly flat the ordi- nates to the curve of true shear become more and more nearly equal to three-fourths the corresponding ordinates of the torsional shear diagram, so that the true shearing strength of steel (or any ductile material) is practically equal to three-fourths the value of the, torsional modulus of * G. B. Upton. " The Sibley Journal of Engineermg," .June, 1913. STEEL 457 rupture. This fact is of value, because torsion tests are much more easily made than are direct-shear tests. STRENGTH OF STEEL IN TORSIONAL SHEAR Class of Steel Computed Extreme fiber stress Lbs. per Sq.in. Shearing Mod. ot Elasticity Lbs. per Sq.in. Number ot Complete Turns per Ft. of Length Number Testa Averaged of Mild Bessemer . . . Medium Bessemer Hard Bessemer . . Cold-rolled 64,200 68,300 74,000 79,900 11,320,000 11,570,000 11,700,000 11,950,000 2.28 2.26 2.01 1.81 15 10 10 12 491. Transverse Strength, Flexure and Deflection. The transverse strength of steel is directly dependent upon the tensile and compressive properties of the metal, for failure in cross-bending must necessarily occur either through failure of the loaded side of a beam in compression or by failure of the opposite side in tension. Whether the former or the latter is the controlling factor is dependent upon the form of the beam and the location of the neutral axis. If the section is symmetrical the failure will usually occur on the tension side if the steel is not a ductile grade, and on the compression side (when the maximum compressive stress has only just exceeded the yield point) if the steel is a ductile grade. Absolute rupture in cross-bending is not possible with any grade of mild or medium steel, since these steels may be bent 180 degrees flat without fracture. Some cast steels and most high-carbon steels may be actually ruptured. The modulus of elasticity of steel in bending is exactly the same as in tension and, since the flexure, and consequently the deflection, is inversely proportional to the value of E, the deflection of all steels within the limits of load which do not stress the most stressed fiber beyond its elastic Umit will be practically the same. (We assume that the section, the span, and the value of / are constant.) The maximum deflection attainable at the time of failure will be quite variable, however, even when I is constant, because it is dependent upon the yield point and the ductility of the steel. 492. Effect of Combined Stresses upon Elastic Properties. The situations in machines and in structures where steel members are subjected to a combination of stresses of different characters are so commonly encountered that the problem of the resultant elastic properties of the material becomes one of great practical importance. The following specific cases are merely illustrative of the great number of instances of a combination of stresses which might be cited : 458 MATERIALS OF CONSTRUCTION A steel rivet transmitting load between the two members of a riveted connection is subjected to shearing stress, which is combined with ten- sion caused by its longitudinal contraction after having been driven hot; a steel pin which holds together the various members of a pin-connected truss which meet at a joint, is subjected to a combination of shearing and bending stresses; a power transmission shaft which carries driving pulleys between the shaft hangers is subjected to combined torsion, flexure, and direct shear; a steel wire -rope power transmission is subjected to combined tension and flexure; a non-rectilinear strut is subjected to combined compression and flexure; shearing stresses practically never occur in members of structures or machines without being combined with flexure, etc., etc. Probably the most extensive study of the effect of combined stresses upon the resultant elastic properties of steels is that made by Mr. E. L. Hancock at Purdue University,* although important investigations have been made by others, notably those of Mr. J. J. Guest f and those of Mr. W. H. Scoble.t Hancock's results are summarized by Figs. 268, 269 and 270. The curves represent the following series of tests: (a) Tension tests of steel rods while under various fixed torsional stresses. (6) Torsion tests of steel rods while under various fixed tensile or compressive stresses. (c) Tension tests of steel tubes while under various fixed torsional stresses. (rf) Compression tests of steel tubes while under various fixed tor- sional stresses. (e) Flexure tests of steel shafts while under various fixed torsional stresses. (f) Shearing tests of steel rivets while under various fixed tensile stresses. From Fig. 268 it appears (1) that the elastic limit of steel in ten- sion, compression, or flexure is lowered by the coexistence of torsional stress, the amount of lowering of the elastic limit being proportional to the magnitude of the torsional stress relative to the torsional elastic limit. (2) The elastic limit of steel in torsion is similarly lowered by the coexistence of either tensile or compressive stress. Mr. Hancock found that a torsional shear in the extreme fiber of a * Proc. Am. Soc. Test Matrls., Vol. 5, p. 179; Vol. 6, p. 295: Vol. 7, p. 258- Vol; 8, p. 373; and Vol. 9, p. 427. > ^ > t Proc. Physical Soc. of London, Sept., 1900. J Philosophical Magazine, Dec, 1906. STEEL 459 £ S gg « 0.5 - W 0.4 - shaft or tube equaling the elastic limit in torsion lowered the elastic limit of the steel in flexure about 28 per cent, in compression about 42 per cent, and in tension about 50 per cent. A torsional stress of lesser magnitude lowered the elastic limit in each case proportionally. Conversely, he found that a tensile or compressive stress equal to the elastic limit in tension or compression, respectively, lowered the elastic hmit in torsion about 79 per cent. From Fig. 269 it appears (1) that the strain at the elastic limit in ten- sion or compression is lowered by the coexistence of torsional stress, but not as much proportionally as the corresponding stress, this meaning therefore that the modu- lus of elasticity is slightly i.^n lowered, (2) the deflection at the elastic limit in flexure is lessened by the coexist- ence of torsional stress, and the relative effect upon this deflection and the corresponding stress at the elastic limit is such that the modulus of elasticity in ilexure is also slightly lowered. From Fig. 270 it appears (1) that the shearing strength of rivet steel, is reduced about 12 per cent if it is at the same time stressed to its elastic limit in tension, and (2) that the shearing elastic limit is reduced about 33 per cent under the same conditions. The lowering of the strength and elastic limit in shear by a coexistent tensile stress below the tensile elastic limit appears again to bear a relation to that above noted nearly proportional to the ratio of the existent tensile stress to the tensile elastic limit. 493. Hardness of Steels. The precise meaning of the term hardness as applied to metals is, as above stated. Art. 366, not altogether fixed. Hardness referring to machine cutting tools means the ability to hold an edge while cutting metal; with respect to steel rails and the tires of car wheels it means the resistance to dry rolling friction; applied to the tires of locomotive drivers the term involves dry rolling friction combined with slippage; applied to axles and bearings it means resistance to wear a o S 0.2- V v^ \ \^ ^T ^ !"■. \ \ fe (1) H \^ \ \ \i 12) \ \ \ k \ (4) (1) Lowering of El, IJm. in Flexure doe to Torsion C2) « ** " " *» CompreBsion due to Torsion. tS) «' " " " "Teneion " " CIJ *• " " " " Torsion due to Tension and. Compression 0.1 V'e % 3/8 VS 6/8 6/5 Proportlonjjf j:iastl£Ximlt.A.ppUed Initially Fig. 268. — ^Effect of Combined Stresses upon Elastic Limit of Steels. MATERIALS OF CONSTRtJCTION y».j/« .'/» */« ^''« »/« AmouQt of Stress Applied In Fraction oLElastlc Limit _ Fig. 269.— Effect of Combined Stresses upon (1) Modu- meter lus of Elasticity; (2) Deformation at Elastic Limit: of inrlpn+ntC (3) Unit Fiber Stress at Elastic Limit, inaentation. between lubricated sur- faces.; respecting the per- formance of gears it means not only frictional wear between lubricated sur- faces, but also involves toughness; the hardness of the crushing faces of rock-grinding machinery is a measure of resistance to combined abrasion and shock; the hardness of ma- chinery steel is a measure of the difficulty experi- enced in cutting or ma- chining it, etc., etc. Properly, the tough- ness and resiliency of steel are properties distinct from hardness, and resistance to wear is quite a differ- ent thing from resistance to cutting or indentation, but as will be shown by the following considera- tion of methods of hard- ness measurement, the term hardness is used meaning any or all of these things. (1) The Brinell meth- od is based upon deter- mining the resistance offered to indentation by a hardened sphere, the latter being subjected to a given pressure. Brinell expressed the hardness by the pressure in kilo- grams per square milli- of spherical area Origi- STEEL 461 nally the spherical area of indentation was computed upon the assumption that the radius of curvature of the indentation is equal to the radius of the sphere used. This method was shown by several- investigators to be in error, owing to the fact that the sph ere flattens slightly under pressure, thus causing the radius of indentation to exceed that of the sphere. In consequence non-con- cordant values of the hardness numeral were obtained in using spheres of different diameters, or in em- ploying different pres- sures, in investigations of the hardness of dif- ferent materials. When the area of indentation is computed upon the basis of its actual ra- dius of curvature, how- ever, this disturbing factor • is eliminated and the results obtained by the Brinell test are truly comparable and of great practical value. P The fact that the hardness numerals by the rule A=— (A = hardness a numeral, P = pressure in kilograms, and a = area in square millimeters of the spherical impression) are referred to diameters of the impressions by the law of reciprocals, has been taken advantage of by M. Guillery,* who has expressed the relation by a straight line on a diagram having logarithmic coordinates. Fig. 271. The ordinates on this diagram are diameters of the impression produced by either a 10-millimeter ball under 3000 kilograms load (scale at left of diagram) or a 5-millimeter ball under 750 kilograms load (scale at right of diagram). The corre- *Proc. Sixth Congress of International Association for Testing Materials, 1912. Paper III5. g 0.8 H ■o a ° 0.7 E 3 0.6 n aei % ff -ii aei s 1= ^ $■ Qt ^ ^ __ ^ — — — "" ^ pT" An lea led ^ / '' , — p? iw r ool / / "^ .6^ ,/ y // / / — . — — — / S s *T ' " f — - -^ '& / ^ \^ / y i y n / 1 IJ 1/ / _ _ SO 100 ISO 200 Magnetic Force (CG.S. Units) Fig. 274.— Typical Induction or B-R Curves. The change of intensity of magnetism induced therefore lags behind a change of magnetizing force, and if the magnetizing furce is one which is periodically varying harmonically between a maximum positive value and a maximum negative value, as is the case when the magnetizing coil is excited by an alternating current, the loss of energy due to the constant necessity of overcoming molecular resistance to change of state becomes a very considerable portion of the total energy expended. This phenomenon of the lag of magnetization behind the magnetizing force is called " hysteresis," and the energy loss occasioned by it Is called " hysteresis loss." Consider, for instance, the magnetizing of the steel represented by the curves of Fig. 275, the magnetizing force being one produced in an STEEL 477 exciting coil by an alternating current. As the magnetizing force increases in a positive direction, the corresponding intensity of magnetism in the steel is represented by the dotted induction curve OA. At A the direction of the magnetizing force reverses and decreases to zero. The magnetism induced ceases to increase and immediately begins to decrease, but, owing to hysteresis, does not at first decrease in positive value at the rate corresponding to the magnetizing force which char- acterized its first increase. As a result when the magnetizing force has become zero the steel still possesses a positive magnetism, the intensity of which is indicated by OE. As the magnetizing force continues to decrease positively (increasing negatively), the induced magnetism decreases and becomes of zero intensity when the magnetizing force has the negative value B. Thereafter the ^+20,000^ magnetism becomes negative and increases negatively until the magnetizing force again reverses direc- tion at C. After the alternating current has passed through a few cycles the cyclic variation of the mag- netism induced comes to be represented by the closed loop 4BCZ), whose area represents the hysteresis loss. If the maximum magnetizing force had been that represented by A' and C", the hysteresis loss would be represented by the loop A'B'C'D'. The locus of the points A and C is evidently the induction curve. The loss due to hysteresis is for a given iron or steel approximately directly proportional to the rate of change of magnetism, or, as com- monly stated, the frequency. It also increases approximately propor- tionally to the 1.6 power of the maximum magnetic intensity which is attained during a cycle (Steinmetz's law), and since the latter is depend- ent upon the chemical character and past hjstory of the iron or steel the hysteresis loss varies with different steels. Eddy Current Loss. Besides the hysteresis loss, with cyclic variation of magnetizing force, a certain amount of energy is wasted in " eddy currents." These are currents of electricity produced in the iron or steel because of the varying magnetic fields. The eddy current loss -20,000 -250 -200-150-100 - 60 0+60 +100 +160 +200 +250 Magnetizing Torce (C.G.S. Units) Fig. 276. — Hysteresis Curves for Steel. 478 MATERIALS OF CONSTRUCTION in the cores of transformers, generators, motors, etc., is greatly reduced by building them up of thin sheets of metal, insulated from each other by a coating of varnish or paper, or by the natural scale on the surfaces. The sheets are so placed as to lie in a direction normal to the direction of flow, and the circuits of the eddy currents are thus broken up. Cwe Loss. The hysteresis loss and the eddy current loss cannot easily be separately determined. The total loss of energy due to both is therefore determined, and called the " core loss " or "iron loss." It is expressed by " the total power in watts consumed in each kilogram of material at a temperature of 25° C, when subjected to a harmonically varying induction having a maximum of 10,000 gausses and a frequency of 60 cycles per second ... It is represented by the symbol Wxo/eo-" * The magnetic properties of iron and steel which are of the greatest practical importance from the standpoint of the manufacture of electrical machinery are (1) the permeability, and (2) the core loss. Methods of testing these two properties have been standardized by the American Society for Testing Materials,! a-nd by the U. S. Bureau of Standards.! Magnetic tests are of fundamental importance to the manufacturer, because he thereby insures the uniformity of magnetic qualities in his materials, which is essential if the performance of electric machines is to be predetermined. Otherwise the manufacturing economy attained by the use of standard dimensions and windings for many machines of the same rating would not be possible. High permeability is desirable in an iron or steel used in the con- struction of electrical machinery not simply because weight and bulki- ness of machines is lessened, but particularly because the strength of the magnetizing force required is thereby reduced, which means that less copper in the exciting coils, and less current, is required to produce a given intensity of magnetic induction. By the substitution of cast steel for cast iron for the yokes, frames, and other parts of dynamos and motors which are not made of laminated sheet metal or copper wire, a considerable economy is effected because of the much greater permeability of cast steel, which is in fact about equal to that of good wrought iron except with low values of the magnet- izing force. (In spite of the higher efficiency which might be obtained, few electrical machines are built in this country with cast-steel frames, etc., except for export. Pressed steel is considerably used, however.) * " Standard Specifications for Magnetic Tests of Iron and Steel " YearhnoV Am. Soc. Test. Matrls., 1914. ' <='""""»■. t" Standard Specifications for Magnetic Tests of Iron and Steel" Am Sop Test. Matrls., Yearbook, 1914. " t Technical Paper No. 117, STEEL 479 For the laminated sheet metal used for the cores of armatures, the pole pieces of dynamos and motors, cores of transformers, etc., a high- permeability iron or low-carbon steel has until recent years always been used. Consideration of the possibility of reducing the core loss, by the use of a material whose hysteresis loss and eddy current loss are less than those of even a very pure iron, has led to the development of cer- tain alloy steels possessing an extremely low carbon content but con- taining 2 to 4 per cent of silicon. This silicon steel, or silicon iron, as it may perhaps be more properly called, exhibits an hysteresis loss very much below that of an ordinary iron, and, in addition, possesses a some- what higher degree of resistivity to the passage of current, and con- sequently reduces in a measure the eddy current loss. Relation between Magnetic Properties and Chemical Composition The following conclusions have been arrived at by de Nolly and Veyret as the result of an investigation of the hysteresis and eddy cur- rent losses of djmamo sheet metals of varying chemical composition.* Carbon. " The carbon percentage should be as low as possible and always remain below 0.1 per cent." A 0.15-per-cent-carbon steel is greatly inferior to one containing 0.10 per cent carbon. Silicon. " The presence of silicon diminishes the hysteresis losses considerably. The coefficient ri changes from 0.0016 for an iron free of silicon to 0.0009 for a metal containing 3.5 per cent siUcon." Silicon also increases the resistivity of the steel and therefore reduces eddy current losses by about 25 per cent. In a very weak field (below 5 to 10 gausses) sihcon increases permeability, but with stronger magnetizing force the effect is reversed and permeability is diminished very much. Manganese. Manganese appears. to be detrimental to magnetic properties if present in amounts exceeding about 0.3 per cent. The data secured bearing upon the effect of manganese are not very conclusive. Sulphur and Phosphorus. Both sulphur and phosphorus were found to be elements whose presence in amounts exceeding about 0.3 per cent (for both combined) constitute a distinct injury to magnetic proper- ties. A steel containing 0.15 per cent sulphur gave very poor results although it was a 0.1-per-cent-carbon steel. Relation between Magnetic Properties and Temperature Investigation of the changes in magnetic properties of steels as the temperature of the specimen under test is raised or lowered has been * Proc. Sixth Congress, International Assoc, for Test. Matrls., New York, 1912, paper IXb. 480 MATERIALS OF CONSTRUCTION made in only a few instances, and little detailed information is avail- able. The most extensive study yet made is probably that reported by the Chemical Laboratory of Messrs. Schneider's Works at Creusot,* and the following remarks are based wholly upon this report. The method of study followed consisted in subjecting steel speci- mens of various forms to uniformly increasing and decreasing tempera- tures in an electric furnace, and simultaneously exposing them to a chosen constant magnetizing force. The temperature indications of a thermo- couple mounted in the specimen, and the corresponding intensity of induced magnetism were continuously recorded autographically on a curve, the ordinates to which are values of 93, and the abscissae tem- peratures in degrees. One of the facts most emphatically shown by the test results was that the form of the curve obtained varies considerably for specimens of different forms. The maximum strength of magnetizing force employed also had a marked effect upon the character of the temperature-induct- ance curve. Because of these considerations the author has selected from the test results only those obtained with the form of test specimen which was least subject to disturbing influences (such as the demagnet- ization caused by the poles), and has further limited his selection to the results obtained with the maximum strength of field which seemed to bring out most clearly the characteristic differences of behavior of the various steels tested. The specimens used in the series of tests selected were 200 milli- meters long, 3 millimeters in diameter, and comprised a series of six steels varying in carbon content from 0.06 per cent to 1.20 per cent. The maximum strength of field employed was 20 gausses and the fre- quency 45 cycles per second. The test results are indicated by the curves of Fig. 276. The cooling 'curves are given negative ordinates in order to avoid confusion of the diagram. It will be noted at once that the most pronounced change of magnetic properties occurs when the critical range of temperatures is reached, at which point an almost total loss of magnetism takes place. The critical points Ai, A2, and ^3 may be located with a fair degree of accu- racy on the curves. Aside from the great change in magnetism which takes place in pass- ing the critical range of temperatures, other changes of lesser magnitude occur at lower temperatures. In the case of the three steels of lowest carbon content there is a progressive increase in induction up to a tem- perature of about 200 or 250° C. The explanation of this phenomenon * Proc. Sixth Congress, International Assoc, for Test. Matrls., New York 1912 paper IX(. ' ' STEEL 481 is not definitely understood, but seems to be attributable partly to the magnetic viscosity of mild steels at these temperatures, which may restrain magnetism. All of the medium- and high-carbon steels exhibit a certain falling off in magnetism in passing through the range just below 200° C, the loss becoming accentuated as the percentage of carbon increases. This behavior appears to be due to the fact that iron carbide loses its mag- B 16 000 i leatlpg 1 C=.0.0li4 14,000 12,000 10,000 <' O.Uft 17', F^ \ '^ c; Igj* \ \\ —- -^- c = O.iOJ ■^ \ \\\ ^ Ci .V ^ . ^ — -^ N \\\ 6000 40O0 2000 iV ii \l 1 5 n 1 1£ 2( 2c 8( 8 41 4! 5 5 6( 6; 7( V 1^ 2000 4000 6000 8000 10,000 12,000 Ccmi orat irc 1 7 T 1 / 1 1 1 1 / 1 / ^^ — £; 0.73S A /'/ / / n ^7 / ^ ■Vv - U.4U1. si— 'J '/ ^____ c= 1.20 )( 16,000 B \ c- 0.17 Ji y c 0.06^ )0Un r Fig. 276. — Relation of Magnetic Induction to Temperature. (Messrs. Schneiders' Works, Creusot.) netism to a very large degree at a temperature in the neighborhood of 200° C. The phenomenon is naturally most marked therefore when the total carbon content and therefore the carbide content is greatest. The fact that the magnetic properties of the 1.2-per-cent-carbon steel appear to excel those of the 0.73-per-cent and the 0.911-per-cent- carbon steels, appears to be due to the fact that the xnicro-structure of the latter two was largely that of lamellar pearlite, the series of inclu- sions of carbide in which might be considered to oppose the passage of magnetic lines of force much more effectively than when the carbide 482 MATERIALS OF CONSTRUCTION occurs in small globules disseminated throughout the ferrite, as was found to be the case in the 1.2-per-cent-carbon steel. Relation between Magnetic Properties and Mechanical and Thermal Treatment The effect of various mechanical treatments on a very low-carbon, medium-sihcon, dynamo sheet steel is indicated by the curves of Fig. 277, which have been selected from the series of tests made by de Nolly and Veyret and cited above. 18,000 id 50 60 70 Strength of Magnetizing Force Fig. 277. It appears that the cold working involved in stamping the specimens with a hollow punch is injurious to their magnetic properties, as is shown by a comparison of curves A and C, the latter representing specimen A after anneaUng. Curves D and E show that cold working by slight bending, followed by straightening with the mallet on wood, is injurious in proportion to the amount of distortion or cold working. Annealing at 300° C. removes the injurious effect of cold working in a measure (curve F), but is less effective than annealing at higher temperatures, as shown by curves G and J. The maximum benefit seems to be derived by annealing at 900° C, for when the annealing temperature is raised to 1000° (curve K) a less magnetic steel is derived. The effect of various hardening and annealing treatments upon the STEEL 483 magnetic properties of a 1.0-per-cent-carbon spring steel is shown by the curves of Fig. 278, which has been abstracted from a preliminary report Fig. l.Anaealed A Unhardened ?;9°.™f "S* '«'°'^. Critical Temperatura 3,Ab Beceiyed B Quenched in Oil H Quenched in Water 50 100 150 200 250 Strength ol Magnetizing Force 278. — Normal Induction Curves for a 1 per cent Carbon Spring Steel after Various Heat Treatments. (Burrows.) g 15,000 ^ fl y^ /B p / D^ — ^ ^ t / I / by Mr. C. W. Burrows upon the study of magnetic properties of irons and steels made at the U. S. Bureau of Standards.* It will be noticed that for this high-car- bon steel the magnetic properties are slightly improved by annealing the steel as received, scarcely affected by quenching below the critical temperature, very considerably im- paired by quenching in oil, and enormously impaired by water quenching. The effect of quench- ing, and quenching ^^^ (i,) followed by tempering, ^^^ 279.— Effect of Special Heat Treatment on Magnetic is shown again by the Properties of Steels. (Burrows.) curves of Fig. 279. It (a) l.O per cent Carbon (6) 0.8 per cent Carbon (1.0 should be particularly Steel. per cent Manganese) Steel, noted that these spring steels exhibit peculiar magnetic properties when tempering follows hardening. At low inductions the magnetic properties are rather * Proc. Am. Soc. Test. Matrls., Vol. 13, p. 570. so Too 50 100 ^1 strength oJ Magnetizing Force A. Annealed A. As Received B. Quenched in Oil 953° to 3i°C. B. Annealed Q j Quenched in Oil from 953" 0. C . Quenched in Oil from Sgo'c. * \ Drawn in Air at 665° C. ( Quenched in Oil from 820°C. "•\ Drawn in Air at 700°O. 484 MATERIALS OF CONSTRUCTION low, but at inductions exceeding about 13,000 gausses they excel the annealed material. Relation of Magnetic to Mechanical Properties The relation of magnetic properties of steels to their mechanical properties is under investigation at the U. S. Bureau of Standards, but the only information yet available is that relating to the study of mag- netic criteria of a 1.0-per-cent-carbon steel above alluded to. Fig. 280 exhibits the relation found between intensity of magnetic induction under a magnetizing force of 50 gausses, and the elastic limit in bending of the steel. Steels containing 1.0 per cent carbon were used 200,000 dl£0,000 -^^ o • ■=j "no III 'us S 1 t"'^"^ J 1 ■3 Quenched 10,000 11,000 12,000 lutCDSity of Magnetic Inductloo Fig. 280. — RelatioA between Maximum Fiber Stress at Elastic Limit and Magnetic Induction under a Magnetizing Force of 50 Gausses for a 1.0 per cent Carbon Steel. (Burrows.) in all tests, the elastic limit being varied by subjecting the individual, specimens to different hardening and tempering treatments. The curve shows that the highest induction corresponds to the lowest mechanical strength (that of the annealed steel), and the decrease in induction follows very closely with increase in quenching temperature, other things being equal. The mechanical strength increases up to a maximum value with decrease in induction, and thereupon decreases as the induc- tion continues to decrease. Fig. 281 exhibits the relation found between the induction secured at 50 gausses and the angle of cold bend at which 1.0-per-cent-carbon steels, variously heat treated, will rupture. The curve shows an increase of the angle of bend as the induction decreases; a maximum angle is passed, and further decreases i,n induction mean a decrease in the angle of bend. The maximum angle of bend occurs at a higher induction than STEEL 485 does the maximum fiber stress in bending. This would seem to indicate that the most serviceable steel would be one whose induction is a com- promise between that corresponding to the maximum fiber stress and SSI ,°^ 3 °S ol S ol cl ol s s 1 SS3 « o 1 g ol .X"^ •-^^^ / 1 • >. 1 I \ " "b o /• \ •» I \ „ ^ • Quenched a 100 E3 ■«1 10,000 11,000 12,000 13,000 6» Intensity of Magnetic Induction FiQ. 281. — Relation between Angle of Cold Bend and Magnetic Induction under Magnetizing Force of 50 Gausses for a 1.0 per cent Carbon Steel. (Burrows.) that corresponding to the maximum angle of bend. If the induction were lower, the steel might be expected to lack toughness; if higher, the mechanical strength will be low. i^vm a > 500 % ^^^ o tP p . ,. .. r.... .__ P oco O ">=» ( • Q \^ o >^ o -^-^ ■^-^ o 10,000 11,000 12,000 13,000 Intensity of Magnetic Induction 14,000 ^60 Fia. 282. — Relation between Brinnell Hardess and Magnetic Induction under Magnetizing Force of 50 Gausses for a 1.0 per cent Carbon Steel. (Burrows.) A comparison between magnetic induction and the hardness, as shown by the Brinell test, is afforded by Fig. 282. Both the hardness expressed by the ordinary Brinell number (based upon the area of spherical indentation) and the Devies hardness numeral (based upon the depth of indentation) are compared with induction at 50 gausses on this 486 MATEEIALS OF CONSTRUCTION diagram. The curves show a general increase in hardness accompanying a decrease in magnetic induction. 498. The Corrosion of Iron and Steel. The importance of the dura- bility of iron and steel when exposed to the various conditions of service has long been recognized, as is shown by the fact that the technical jour- nals and the transactions of the Societies devoted to engineering or its allied sciences have long teemed with contributions and discussions bearing upon the general problem of the corrosion of iron and steel. Unfortunately, much of the literature of the subject is of a distinctly controversial nature, owing to the fact that students of the subject are far from being agreed upon what the true explanation of the phenomenon is. Furthermore, various experimental studies of the relative corrodi- bility of different materials have, because of their contradictory results, led to opposing conclusions concerning the effect of various factors such as relative composition, constitution, structure, etc. The end product of the process of corrosion, i.e., rust, is well known to be simply ferric oxide plus a variable amount of combined water, and all authorities are agreed that corrosion of iron is possible only when both oxygen and water are present. Iron will not rust in perfectly dry air; on the other hand, it cannot rust in water which is entirely free from oxygen. It is further well known that certain conditions encoun- tered in practice, such as a very humid atmosphere, exposure to acid fumes, acid-bearing air or waters, or water containing dissolved salts, tend to accelerate corrosion. At about this point universal agreement between authorities ceases to exist, however, and it becomes necessary, in proceeding to consider further the vexatious problem of corrosion, to recognize the existence of conflicting theories, a number of which are strenuously supported by competent authorities. In the following very brief discussion of these theories the author has been guided principally by the works of Dr. A. S. Cushman, Director of the Bureau of Indus- trial Research in Washington, D. C, who is probably the foremost Amer- ican authority on the subject.* Of the many theories advanced to account for corrosion three have gained widest recognition. These are, in the order of their present importance: (1) the electrolytic theory, (2) the hydrogen peroxide theory, and (3) the carbonic acid theory. The electrolytic theory, which is now accepted by the greater number of authorities, is much more complex than are the two other theories above mentioned, and on this account the latter two will be briefly explained first. * See particularly the work of Cushman and Gardner, " The Corrosion and Pres- ervation of Iron and Steel," 1910, and the various contributions of Dr. Cushman in the Proceedings of the Am. Sec. Test. Matrls., 1906-1913. STEEL 487 The carbonic add theory asserts that the hydroxylation of iron cannot take place without the interaction of carbonic or some other acid. The acid changes the metal to a ferrous salt with evolution of hydrogen: 2Fe+2H2C03 = 2reC03+2H2. The water and oxygen now react with the ferrous salt, forming ferric hydroxide and setting free all the acid originally used: 2FeC03+5H20+0 = 2Fe(OH)3+2H2C03. A small amount of acid is sufficient to rust a large amount of iron, because it is set free again immediately after having combined with the iron to form the ferrous salt. This explanation of corrosion is entirely plausible, and the fact is well known that acids do often act a part in promoting the corrosion of iron, but the theory has been discredited by a number of investigators, including Cushman, who have shown, by methods so carefully planned and executed as to leave scarcely room for questioning the accuracy of their results, that iron will corrode in water containing oxygen, but not a trace of carbonic acid. Corrosion will even take place in dilute alkaline solutions in spite of the fact that the alkali would neutralize any acid present. The hydrogen peroxide theory explains corrosion by stating that water and oxygen react with iron to form ferrous oxide and hydrogen peroxide, whereupon one-half of the peroxide combines with the ferrous oxide to form hydrated ferric oxide, the excess of peroxide remaining to form a further amount of rust. Thus: 2Fe+202+2H20 = 2FeO+2H202=Fe203,H20+H202. This theory has failed to gain any substantial corroboration as the result of experimental investigation, and is largely discredited by the fact that the most delicate tests have always failed to detect even the transitory presence of the peroxide during the ordinary process of rusting. The electrolytic theory of corrosion cannot be explained without a preliminary statement of certain of the fundamental principles of the theory of solutions, electrolytic dissociation and electrolysis. Water is the one universal solvent for all forms of matter, and the common terms " soluble " and " insoluble " express simply relative solubility. A soluble substance readily passes into solution, but an insoluble one does so only with great reluctance. When a substance of the first class passes into solution " its molecules or atoms (depend- ing upon whether the substance in question is a compound or simple element) tend to distribute themselves equispacially among the mole- 488 MATEEIALS OF CONSTKUCTION cules of the solvent. The driving force which produced this tendency ... is known as solution tension or solution pressure " and is not a con- stant except for a given substance in a given physical and thermal state. As the number of particles in solution increases, there arises a back pressure, acting against the solution tension. This is what is known as osmotic pressure. "It is at once apparent that for any given sub- stance at a given temperature, its maximum solubility would be reached just as soon as the solution pressure and the osmotic pressure were equal." External causes may operate to reduce the osmotic pressure, however, and the extent of the solvent action be thereby increased. In analyzing certain studies of osmosis van 't HofI * made the momen- tous discovery that with a certain class of substances, chiefly of organic nature, the osmotic pressure of a solution of the substance is exactly equal to the gas pressure of a gas having the same number of mole- cules in a given volume. This meant that the osmotic pressure of these solutions follows the law of Boyle for gases, which states that the pres- sure of a gas varies directly with its concentration. Van 't Hoff further pointed out that another class of substances includ- ing the acids, bases, and salts of inorganic chemistry, did not obey the gas law, but in many cases showed osmotic pressure far out of proportion to the concentration of the solution. This observation led Arrhenius to question whether inorganic compounds entering solution might not simply dissociate molecularly, but actually break down into ultimate constituent particles or atoms which he called ions. This was the begin- ning of the modern theory of electrolytic dissociation. The theory of electrolytic dissociation asserts that when any com- pound of the class of the acids, bases, and salts of inorganic chemistry is dissolved in water to form a dilute solution, its molecules break down into atoms or groups of atoms called ions. The ions originating from the dissociation of a molecule all carry charges of static electricity. The charge of some ions is positive, and of others negative, but the posi- tively charged ions of a molecule are always opposed by ions whose total negative charge is exactly equal in amount. The positively charged ion is called a cation, and the negatively charged one an anion. Thus hydrochloric acid (HCl) dissociates into a cation, hydrogen (H), and an anion, chlorine (CI). Certain ions carry two or three times the ordi- nary charge, and are then termed bivalent or trivalent ions, while the ordinary ion is univalent. Thus calcic hydroxide (Ca(0H)2) forms the bivalent cation (Ca) and also two monovalent anions (OH) and (OH). * Zeit. Phys, Chem. I., 481 (1887). STEEL 489 Hydrogen is the cation of all acids, and the hydroxyl (OH) is the anion of all bases. No compound shows acid properties unless the characteristic "hydrogen ion is present, nor can a compound be basic without the char- acteristic hydroxyl ion. The anion of acids, the cation of bases, and both the anion and the cation of salts vary with different acids, bases or salts. A second method of formation of ions which characterizes many proc- esses of electrolysis, is the taking- of its charge from an ion by an atom, the ion thereby becoming an atom while the original atom becomes an ion. Thus copper may be precipitated from a solution of copper salt by the action of zinc atoms robbing the copper ions of their positive charge. Similarly a metal placed in an acid solution will form positively charged ions by taking the positive charge of the hydrogen ions. The third method of ion formation consists of the simultaneous for- mation of ions of opposite character by atoms of two different substances. Thus, when gold is placed in a chlorine solution the atoms of gold becomes positively charged, each gold atom becoming a cation, and at the same time a sufficient number of chlorine atoms assume a negative charge, becoming anions, so that a perfect balance between positive and nega- tive charges is maintained. The fourth and last known method of formation of ions depends upon the fact that an atom of a substance introduced into a solution of another substance may be able to assume a charge and become an ion by causing the valence of ah oppositely charged ion in the solution to be increased, thus maintaining equilibrium! For instance chlorine added to a ferrous chloride solution forms an anion and at the same time causes the ferrous (bivalent) ion to change to the ferric (trivalent) ion. This last example is illustrative of the true nature of the process known in chemistry as oxidation, and the reverse phenomenon is illustrative of the process of reduction. There is absolutely no conflict between the use of the term "valence" in the present instance as a measure of the electrical state of the ion, and the ordinary use of the term in chemistry meaning the combining power of a given atom for other atoms. Electrolysis is that phenomenon which takes place whenever a cur- rent of electricity passes through a solution which is capable of con- ducting the current. Only those substances which in solution form ions are capable of conducting current, and they do so not as a metaUic wire does, but by a migration of the ions toward the electrodes. Substances which form solutions which will carry current are known as electrolytes. If, for instance, a current is passed through a solution of hydrochloric acid the positive hydrogen ions proceed to the negative pole of the external circuit, give up their charge, and go out of the solution as atoms of gase- ous hydrogen. At the same time chlorine ions are proceeding to the pos- 490 MATERIALS OF CONSTRUCTION itive pole where an equal negative charge is given up and particles of chlorine are precipitated out. It is not necessary that there be an external source of electric cur- rent in order that electrolysis may take place, as is evidenced by the familiar voltaic cell or element. When two more or less dissimilar metals are placed in an electrolytic solution and connected by an external con- ductor, current will pass through the external circuit and electrolysis will take place in the solution. Whichever one of the metallic terminals or electrodes is most strongly electro-positive will discharge positive ions into the solution which, together with the positive ions of the solu- tion, proceed to the other terminal, give up their charge, and go out of the solution. At the same time the negatively charged ions of the solu- tion migrate to the more strongly electro-positive terminal and give up their charge and go out of the solution. Consider, for example, the voltaic cell, the electrodes of which are plates of zinc and copper, and the electrolyte sulphuric acid. Zinc is more strongly electro-positive than copper, and therefore discharges zinc ions, which join the hydrogen ions of the sulphuric acid solution, and proceed to the copper electrode, give up their positive charge, and leave the solution as hydrogen gas and precipitated zinc. At the same time the ions of SO4 in the electrolyte migrate to the zinc electrode, give up their negative charges and precipitate out. The copper thus becomes the positive electrode and the zinc the negative electrode, and the amount of current which passes through the external circuit from the copper to the zinc is equal to the total amount of the electrical charges carried by the ions of the electrolyte. The zinc gradually wastes away and the electrolyte becomes impoverished in acid, but the copper remains un- affected. So great a dissimilarity between the electrolyte as exists in the case of zinc and copper just cited is not at all necessary to make electrolytic action possible in some degree. Cushman and Gardner illustrate the point by stating that " even two steel needles from the same package are sufficiently dissimilar to show a slight difference of potential when coupled in such a way, and one will be protected while the other suffers accelerated corrosion." This fact is one of supreme importance in connection with the electrolytic theory of corrosion, as will be shown in the discussion which follows. The Electrolytic Theory of Corrosion Explained. Cushman and Gardner * have concisely stated the electrolytic theory of corrosion in the following terms: " Iron has a certain solution tension, even when the iron is chemically pure and the solvent pure water. The solution * Loc. cit., p. 44. STEEL 491 tension is modified by impurities or additional substances contained in the metal and in tiie solvent. The effect of the slightest segregation in the metal will throw the surface out of equilibrium, and the solution tension will be greater at some points than at others. The points or nodes of maximum solution pressure will be electro-positive to those of minimum pressure, and a current will flow, provided the surface points are in contact, through a conducting film. If the film is water, or in any way moist, the higher its conductivity the faster the iron will pass into solution in the electro-positive areas, and the faster the corrosion pro- ceeds. Positive hydrogen ions migrate to the negative areas, negative hydroxyls to the positives. . . . I " If the concentration of the hydrogen ions is sufficiently high, which . . . is only the same as saying, if the solution is sufficiently acid, the hydrogen ions will exchange their electrostatic charges with the iron atoms sweeping into solution, and gaseous hydrogen is seen escaping from the system. This takes place whenever iron is dissolved in an acid. If, however, as is usual in ordinary rusting, the acidity is not high enough to produce this result, the hydrogen ions will polarize to a great extent around the positive nodes without accomplishing a complete exchange, and the so-called electrical double layer of Helmholtz will be formed. This polarization effect resists and slows down action. Never- theless, although it cannot be seen, some exchange takes place and iron slowly pushes through, . . . (and) for every exchange of static charge between iron and hydrogen at the positive node, a corresponding negative hydroxyl ion appears at the negative node. . . . *In other words, as fast as the iron sweeps into solution the concentration of ferrous hydrox- ide grows, but the ferrous reaction appears in one place and the hydroxyl in another. It is now that the oxygen of the atmosphere dissolved in the solution takes up its work, the ferrous ions are oxidized to the insol- uble ferric condition, which results in the precipitation of rust. . . ." From the above discussion of corrosion it must be apparent (if the electrolytic theory is accepted) that the relative corrodibility of irons and steels under given conditions of exposure to air and moisture, with or without the accelerating effect of acid, will be in a large degree a func- tion of their relative chemical purity and structural homogeneity. It is very difficult to demonstrate these propositions experimentally, however, because of the complexity of the problem and the limitations imposed upon the investigator in his efforts to eliminate all factors except the one under special study. One of the favorite methods of investigating the relative corrodibility of various irons and steels has been by acceleration tests, the media employed being an acid solution which, compared with the ordinary con- 492 MATERIALS OF CONSTRUCTION ditions of rust formation, is a very potent electrolyte. It may well be doubted whether the comparative results obtained under such circum- stances can be relied upon as a certain indication of the behavior of the materials under the conditions of natural rust formation. Nevertheless it is probable that accelerated tests carefully conducted do indicate certain general tendencies which have an important bearing upon the practical problem of how irons and steels may best be made and handled to produce the least corrosive material. 100 BOO Tempering Heats 900 Fig. 283. — Effect of Various Tempering Heats on the Solubility of 0.95 per cent Carbon Tool Steel, Heated to 900° C. and Tempered by Reheating to the Tem- perature Indicated. (Heyn and Bauer.) The diagrams reproduced in Figs. 283 and 284* indicate the rela- tive corrodibility of a 0.95-per-cent-carbon tool steel and a 0.07-per-cent- carbon mild steel, respectively, when subjected to the action of a 1-per- cent solution of sulphuric acid, after the various heat treatments indicated. A study of Fig. 283 indicates that a tool steel is least corrodible when in a glass-hard condition, reheating not having been carried above * These tests are quoted by Cushman and Gardner from the studies of Heyn and Bauer, origmally published in the Jour. Iron and Steel Inst., May 1, 1909. STEEL 493 100° C. Further heating increases the solubility in acid until a maxi- mum solubility is reached when the tempering heat is 400° C. Higher temperature heats again reduce solubility until a minimum solubility, about twice that of the hardened steel, is attained when the metal has been completely annealed at 900° C. Fig. 284 indicates that the solubility of a very mild steel is increased as the temperature of tempering increases; a maximum solubility is reached 300 400 500 600 Tempering Heats Fig. 284. — Effects of Various Tempering Heats on the Solubility of 0.07 per cent Carbon Mild Steel, Quenched at 1000-1030° C. and Tempered by Reheating to the Temperature Indicated. (Heyn and Bauer.) when reheating is carried to about 300° C; and higher heats again reduce solubility, until, when the steel is completely annealed at 900° C, the sol- ubility is considerably below that of both the original rolled bar and the steel after hardening but without tempering. Attention is directed to the fact that a comparison of Figs. 283 and 284 will show that the tool steel is about 120 times as soluble as the mild steel when reheated to the temperature which produces maximum solubility, about 60 times as soluble in the original or the annealed condition, and about 25 times 494 MATERIALS OF CONSTRUCTION II s Relative Solubility Basis: Solubility of Annealed W o , o96 Hours At a- '■■K 'S^ s^ ^«^ ^i \ = 60 o 1 / \, Jkp >^^ ^ V fe^ K^ "I., V a ( s \ "uti s""^ V^ ^ >!>>, ■^ \ 1 gso \ s S, \ S^ o o \ s ^t V §40 1 q s ^ k ^^ h. K %> N ^ 1 ^^. tt" 1 h ^ "i"! ■^ 1?°'^ \\ ^ ^ ■v* ^ "S 'If \^i C"N \ 3 «20 ^ '\i ^ \ b ^ h \ ^ Ns ^ b 1 1 3.' 2 i .14 I 9 Perc enb 15 ige Jf N Icke I 25 26.24 30 Fig. 294.- -Relative Corrosion of Various Nickel Steels. (Friend, Bentley and West.) by the Bessemer or crucible process. metallic nickel, or ferro-nickel, charged with the rest of the stock Not counting armor-plate, where nickel is alloyed with chromium nickel steel is most used for structural work in bridges, railroad rails (on curves particularly), steel castings, ordnance, engine forgings shaft- ing (especially marine shafting), frames and engine parts of autoi^obiles wire cables, axles for cars and automobiles, etc. THE SPECIAL ALLOY STEELS 513 503. Manganese Steel. Manganese steel usually contains from 11 to 14 per cent manganese and from 0.8 to about 1.5 per cent carbon. Manganese steel was introduced by Sir Robert A. Hadfield in England in 1887-1888, and has ever since held a unique position among alloy steels. When cast in the ingot manganese steel is almost as brittle as glass, and is so hard that no carbon steel will cut it. Reheating to about 1000° C, followed by quenching in water, has the remarkable effect of rendering the material very much tougher and very much more duc- tile without materially altering its hardness. No other known metal or alloy combines in an equal degree extreme hardness and great ductil- ity. No treatment will materially soften manganese steel when cold, and it must therefore be usually cast to as nearly its final form as pos- sible and subsequently be finished by grinding. Manganese steel is very fluid when molten, and sound eastings are produced. The shrink- age is excessive, however, (often more than f inch per foot). The metal may be worked or forged with great diflaculty through a short range of temperatures above a red heat. It is practically non-magnetic under all circumstances unless it has long been maintained at about 450° to 500° C. Its habit of elongation after the yield-point is passed in tension differs from that of carbon steel in that it does not neck down, but its elongation and contraction of area are quite uniformly distributed over the entire length of the specimen. Structure and Constitution. The presence of manganese in the amount normally used in manganese steels has been shown to be responsible for the complete suppression of the allotropic changes which normally occur in the heating or cooling of carbon steels. Sir Robert Hadfield has recently shown * that the heating and the cooling curves of cast and forged manganese steels containing from about 10.9 to about 13.4 per cent manganese do not show the slightest retardations at any point between — 200° C, and -|- 1355° C. Manganese steel has no critical points, therefore, and the steel at atmospheric temperatures or any temperature above or below atmospheric temperatures must be a solid solution of 7-iron, manganese, and carbon. (One exception to the above statement must be . made in view of the fact that Sir Robert Hadfield has been able to develop critical points in the neighborhood of about 750° C, by prolonged heating at about 500° C. The steel so treated became slightly magnetic.) Professors J. 0. Arnold and A. A. Reed f have produced strong evidence tending to show that the manganese exists in combination with a portion of the iron and carbon as the double carbide, SFesC, MnsC. * " Heating and Cooling Curves of Manganese Steel," Jour. Iron and Steel Inst., Vol. 88, 1913, II, p. 191. t Jour. Iron and Steel Inst., Vol. 81, 1910, I, p. 169. 514 MATERIALS OF CONSTRUCTION In view of the fact that manganese steel possesses no critical points, it is impossible in the present state of our knowledge of the problem to make any explanation of what constitutes the nature of the extraor- dinary change of structure which accounts for the transformation upon quenching of an extremely brittle material into that condition in which possesses greater ductility than any other iron or steel. (In a very recent study of, this problem Mr. W. S. Potter * obtained heating and cooUng curves of manganese steels, which had been initially- very slowly frozen and slowly cooled, which he believed exhibited dis- tinct points in the neighborhood of 850° C, and less distinct points at a number of other temperatures. It appears not unreasonable to believe, however, that the slight retardations observed may have been due to instrumental errors.) Tensile Properties. The remarkable properties of manganese steel above alluded to are forcibly illustrated by the diagrams of Fig. 295, which exhibit the tensile properties of steels containing from 0.84 to 21.69 per cent manganese. Steels which have been tested as rolled, others which have been forged, but not otherwise treated, and still others which have been quenched in water from a white heat are included in the series of test results presented. These diagranis constitute a summary of the historically famous tests of manganese steels published by the discoverer of the valuable properties of this alloy steel. Sir Robert Hadfield himself. This particular set of data has not been selected on account of its historical value, however, but because it covers the ground more comprehensively than any more modern data found available. The following facts appear to be estabUshed by Hadfield's tests: (1) As the manganese content of steel is raised above 1 per cent the carbon content tends to increase, the proportionate increase of car- bon with respect to manganese gradually becoming greater with higher manganese steels. (2) The tensile strength of rolled manganese steel which has not been subjected to any heat treatment increases only very slightly with increase in manganese and, in fact, does not nearly equal the strength which might be expected to characterize a carbon steel of the same carbon content. Moreover, the increase of manganese rapidly reduces the ductility, and no rolled manganese steel containing more than about 2.5 per cent man- ganese shows more than about 5 per cent elongation in a gauged length of 8 inches. (3) The tensile strength of manganese steel which has been forged, * " Manganese Steel, with Especial Reference to the Relation of Physical Proper- ties to Micro-Structure and Critical Ranges," Bull. Inst. Min. Engrs., April 1914 p. 601. <^ J f 7 • J THE SPECIAL ALLOY STEELS 515 but not otherwise heat-treated, increases fairly rapidly with increase in manganese between 7 and 19 per cent, but is very low for the 7-per- cent-manganese steel and never equals that of an ordinary steel of the same carbon content. The ductility of these steels is of a very low order, although they may be forged with care. The elongation is never in excess of about 5 per cent. (4) The tensile strength of water-quenched steels containing in excess of 7 per cent manganese mounts rapidly with increase in manganese until a maximum of more than 140,000 pounds per square inch is reached with Sai5 ■1 ^ ^ ,^ 831.0 .^SoOJ H.UOOOO c Per lenta ge Cn R bon u L- -'- ■'' y- i2. — M)' °^i s rE= =-: — \\ 1 1 1 1 NoH roken It III ilmu Si 3 Ji d i? ^ load DdlOB «d I'i'iaoooo n S /l\ ,« 1 ? (Ju alleS rengt :fron lofte Whi Wat ^Hei 'V II 1 1 1^ \, / ' c A / s i' 1 1 ■ ^ k»9 ^ |100 000 1 \ y 7 ! , s '' :->'- >' I; \ t ! \ TenB UeSt) oillth as B lied J. .A enBllo Stran Ithof Test tars \ M 1 \ — -°- / \ ,-'" ,fyo^ rged eatmo vllbD it ofurt lerh at \ Tensile Strength and entase Elongation In 8"| I /. J V' / I lastlo Ltml of Q) enoht dBt« "k < /^^ — J\ ) / V K \ p ercet lueno S Ugll >ed-S longa ^on o \ Pero Dtage 1V..I IIIDIL ifBU lul loUed X J s s^. -->, S^ ^zr. »"• X. a«»r _ PorcenUi T gatlo ofF Jsi ;teelB _.otJ «r|fi. led S s < ( 1 1 2 u 1 S 1 8 2 9 2 i Percentage of Manganese Fig. 296. — Tensile Properties of Manganese Steels. (Hadfield.) about 13 or 14 per cent manganese. With still higher manganese content the strength rapidly falls off again. More important than the tensile strength of the quenched steels, however, is the wonderful ductility developed. With only 7 per cent manganese the elongation is only about 1.5 per cent, but it rapidly increases with increase in manganese until a maximum elongation exceed- ing 50 per cent is reached when the manganese content is about 13 or 14 per cent. Further increases in manganese rapidly decrease ductility. The ductility and the tensile strength remain practically directly pro- portional for all percentages of manganese. 516 MATEEIALS OF CONSTRUCTION A comparison of the elongation of the quenched steel and the forged steel containing about 13 or 14 per cent manganese reveals the fact that the elongation has been raised by quenching from about 1 or 2 per cent to more than 50 per cent. At the same time the strength has been increased about 100 per cent, and the hardness has not been materially- impaired. Some of these bars were bent cold, after testing, nearly 180°. In no case was there any sign of necking down, but the contraction of area and distribution of elongation were nearly uniform over the entire length of the reduced section of the test bars. (5) The elastic limit, as determined by the "first permanent set" is very low in proportion to the tensile strength. For the steels containing from 13 to 15 per cent manganese the stress at the elastic limit amounted to only 35 to 40 per cent of the tensile strength, being about 50,000 pounds per square inch for the steels whose tensile strength exceeded 140,000 pounds per square inch. (There are certain noticeable irregularities in these curves, certain test bars appearing to show abnormal properties. On account of the possibility that these anomalies might be explained by an abnormal carbon content, the amount of carbon in each steel has been indicated by the uppermost curve on the diagram, and what seems to be the normal carbon content is indicated by the dotted curve. A comparison between the carbon curve and the various strength and elongation curves now becomes very interesting. It would seem that practically every anomalous test result may be accounted for by an abnormal carbon content. Note particularly the steels containing respectively 10.11, 10.60, 12.60, 15.22, and 18.65 per cent manganese.) Manufacture and Uses. Manganese steel is made by the open-hearth process whenever large masses are required. It may be made in the crucible, however. Manganese is added to the steel in the form of ferro- manganese just before the completion of the process. Large quantities of manganese steel are used as steel castings, par- ticularly where great hardness and strength combined with great tough- ness are called for. It finds a special application in the construction of those parts of crushing and grinding machinery which are subjected to severe shock and abrasion. It is also used for curve rails, frogs, and crossings where hardness and freedom from brittleness constitute a great advantage, and to a Umited extent for axles and treads of wheels of railway rolling stock. Its principal limitation in machine construc- tion is the practical impossibility of machining it to final form by ordinary methods, on account of its excessive hardness. 504. Chrome Steel. Chrome steel usually contains from 1.5 to 2 or 2.5 per cent chromium together with from 0.7 to 1.5 or even 2.0 THE SPECIAL ALLOY STEELS 517 per cent carbon. Its value is due principally to its property of combin- ing intense hardness after quenching with very high strength and extremely high elastic limit. It is therefore especially well able to withstand abrasion, cutting, or shock. The quenching treatment does not improve its ductiUty, as is the case with manganese steels, but on the other hand it is more ductile than a similarly treated carbon steel of the same car- bon content. Structure and Constitution. Thermal Critical Points. The char- acteristic structure of chrome steels containing less than about 5 or 6 per cent chromium does not differ materially from that of carbon steels similarly treated except for the presence of emulsified or finely granu- lated chromium carbide. With more than about 6 per cent chromium the structure of annealed steel " consists of chrOmiferous ferrite con- taining particles of double or triple (chromium) carbides, the carbide masses varying in size, some being very minute specks, and others of considerable dimensions." * Arnold and Read found carbides in drastically annealed chrome steels corresponding to the following formulae: Per cent Carbon in Steel. Per cent Chromium in Steel. Formula which Approximately Agrees with Carbide Analysis. 0.64 0.84 0.835 0.85 0.85-0.88 0.65 0.99 4.97 10.15 15.02-23.7 20Fe3C,Cr3C2 12Fe3C,Cr3C2 4Fe3C,3Cr3C2,CriC FejC.CrsCeriC 2Fe3C,3Cr4C From 80 to 99.7 per cent of the total amount of carbon in the steel was accounted for by the carbon found in the double and triple car- bides. Chrome steels possess distinct critical temperatures corresponding to those of ordinary carbon steels. Professor McWilliams and Mr. Barnes of the University of Sheffield have established the position of the critical points in heating and cooling 2 per cent chrome steels (with regard to carbon content) as indicated in the following table: f * Arnold and Read, " The Chemical and Mechanical Relation of Iron, Chromium and Carbon," Jour. Iron and Steel Inst., Vol. 83, 1911, I, p. 258. t " Some Physical Properties of Two Per Cent Chromium Steels," Jour. Iron and Steel Inst., Vol. 81, 1910, I, p. 263. 518 MATERIALS OF CONSTRUCTION CRITICAL TEMPERATUE,ES OF 2 PER CENT CHROME STEEL Per cent Carbon. Deg. C. Ac. Deg. C. Aci. Deg. C. Ar,. Deg. C. An. Deg. C. An. Deg. C. 0,20 765 759 753 759 791 789 ?822 ?810 785 765 732 0,25 758 748 731 0.32 785 778 733 0.50 721 0.65 0.85 783 777 718 714 Tensile Properties. The tensile properties of a series of steels con- taining from 0.62 to 32.46 per cent chromium are exhibited by the diagrams of Fig. 296. In order to avoid the disturbing factors incidental to the past history (heat treatment and mechanical working) of the steel from which the test specimens were cut, all of these specimens have been heated three hours at 850° C, prior to testing. The tests were made by M. Albert M. Portevin and represent a portion of the investigation of the special ternary steels above referred to.* For the most part the iron-chromium carbon alloys represented by the diagram of Fig. 296 do not represent the particular alloys which are important as alloys steels. Only one or two facts illustrated by the diagram will therefore be noted, before considering certain special alloys which possess greater practical importance. First. It should be noted that all of the steels containing between 1 and about 20 per cent of chromium show remarkably high tensile strengths, only one steel in this series showing less than 170,000 pounds per square inch tensile strength. (The fact should be borne in mind that these steels have not been quenched nor thoroughly aimealed.) Second. The elastic limit of most of these steels is remarkably high in proportion to the tensile strength. Third. The ductility as indicated by the contraction', of area is rather low for all of the steels except those containing in excess of about 20 per cent chromium. Fourth. The tensile properties of all of the steels are closely depend- ent upon the carbon content, most of the anomalies in the various curves being directly attributable to an abnormal carbon content. The steels whose tensile properties are indicated by the curves of Fig. 297 more nearly represent the chrome steels which are commercially used. The tests upon which these curves are based constitute a portion of the very detailed study of the physical properties of 2 per cent chrome * See footnote, page 501. THE SPECIAL ALLOY STEELS 519 steels made by Professor Andrew McWilliam and Mr. Ernest J. Barnes and referred to just above. The relation between tensile properties, carbon content, and heat treatment of 2 per cent chrome steels is excep- tionally well shown by these tests. (It will be noted that this study of chrome steels parallels the study of nickel steels above quoted from another paper of the same investigators.) ■SO.9 50.8 •tf^O.7 .200 000 ^ ^ --* \ / \, f / ( V \ £90 000 \ r ',!' \ \ ^ f> 8 \^ ^ « 70 000 \. ,^-' ■'" H 50 000 AOOOO "ti 40 /^ - £ 30 § 20 a gio / NOI E; J / All steels 3 ho'urs . wer( t850 0. ed \ > y\ ont acti ono t Al- y before bt Dgt 35 ted \ ^ y \. / \ !»— ^ 9 8 U) 12 U le IS 20 22 21 28 Percentage of Obromlum Fig. 296. — ^Tensile Properties of Various Chrome Steels. (Portevin.) The following general statements cover the most important facts established by this series of tests of 2 per cent chrome steels: (1) The tensile properties of chrome steels, whether heat-treated in any manner or not, are dependent upon the carbon content to a very marked degree. (2) In general the tensile strength and yield-point increase rapidly, and the ductility decreases, as the carbon content increases. The max- 520 MATERIALS OF CONSTRUCTION imum beneficial effect of carbon would seem in most cases to be attained when the carbon content does not greatly exceed 0.7 or 0.8 per cent. (An exception to this statement must be made in the case of steels which have been neither hardened and tempered nor annealed.) (3) The yield-point of 2 per cent chrome steels is remarkably close to the tensile strength in case the steel has been quenched and tempered. r.3 O.i 0.S 0.6 Percentage ot Carbon Fia. 297. — Relation of Tensile Properties of 2-per-cent-Chromium Steels Content and Heat Treatment. (McWilliam and Barnes, to Carbon ) With moderate tempering the yield-point may exceed 95 per cent of the tensile strength. (4) The tensile properties of chrome steels which have not been heat- treated (hardened and tempered) do not excel those of ordinary steel of similar carbon content in any way except that they are slightly more ductile. THE gPEClAI. ALLOY STEELS 521 (5) The tensile strength is about doubled, the yield-point about tripled, and the ductility reduced about one-half, by quenching from 800° C, followed by tempering at 400° C. Tempering at higher heats reduces the strength and yield-point and increases the ductility in pro- portion to the tempering heat employed. Quenching followed by moderate tempering raises the tensile strength of 0.2 per cent carbon steel from 67,000 to 137,000 pounds per square inch, and the yield-point from 42,000 to 134,000 pounds per square inch. When the car- bon content is 0.5 per cent, this same treatment raises the tensile strength from 122,000 to 228,000 pounds per square inch, and the yield-point from 84,000 to 224,000 pounds per square inch. (6) By drastic annealing at high heats, long prolonged, the tensile strength is very much reduced and the yield- point is lowered even more in proportion, while the ductility is very notably inci-eased. When it is recalled that the possession of this tremen- dous strength, and a yield- point unapproached by any other steel (except nickel- chromium steel), is combined with a hardness almost equal to that of manganese steel, the commercial value of this class of steel may be appre- ciated. It is rather lack- ing in ductility, but this is often unimportant in view of its extremely high elastic hmit. Corrodibility. The relative corrodibiUty of three chrome steels and a 0.29 per cent carbon steel in various media is exhibited by the curves of Fig. 298. This diagram is based upon data secured by Messrs. J. N. Friend, J. L. Bentley, and W. West in the experimental study of the problem above referred to in discussing the corrosion of nickel 1.12 2 3 3.68 4 5 5.30 Percentage of Chromium Fig. 298. — Relative Corrosion of Various Chrome Steels. (Friend, Bentley and West.) 522 , MATERIALS OF CONSTRUCTION steels.* These corrosion tests constitute a portion of the same series of tests above quoted, and were made in the manner heretofore described. The comparative corrodibility of these several chrome steels under the conditions of the tests may be expressed as follows: (1) All chrome steels corrode less rapidly than ordinary carbon steel, but the advantage is far less marked in the case of the 1 per cent chrome steel than it is in the case of a 5.3 per cent chrome steel. (2) The superiority of chrome steels over carbon steel with respect to resistance to corrosion is generally less marked after six months' exposure in any medium than it is after only two months. (3) The superiority of chrome steels is most marked when the exposure consists in allowing the steel to be alternately wet and dry. The advantage conferred by the addition of the chromium is somewhat less marked on exposure to weak acids, sea water, or tap water, but is still considerable, and especially so in the case of 5.3 per cent chrome steel. (Note that the findings of Messrs. Friend, Bentley, and West in the two series of corrosion tests — nickel and chrome steels — are almost iden- tical in character.) Manufacture and Uses. Chrome steel is made in the crucible or in the open-hearth furnace. Chromiunl is added in the form of ferro- chrome, and since the latter oxidizes easily, the loss will be very heavy in the open hearth unless it is added just before the end of the process. Chrome steel is used where a hard surface and shock resistance are desired. It is commonly used in the manufacture of projectiles and (in a quaternary alloy with nickel) armor plate. It is also used for a limited class of tools and dies, for gears and other parts of automobiles and machines generally, for the wearing parts of rock-crushing machin- ery, and for safes and vaults. In the latter apphcation of chrome steel it is welded with alternate layers of wrought iron into a composite three- or five-ply plate. The chrome steel resists cutting by drills, while the wrought iron introduces an element of toughness so that it is better able to withstand concussion. 505. Tungsten Steel. Tungsten has long been recognized to be a most valuable alloy element for special steels. It is, however, most commonly used in conjunction with chromium or manganese in a qua- ternary alloy, instead of being used with carbon alone as a ternary alloy. The ternary alloys of iron, tungsten, and carbon possess a certain amount of commercial importance, however, and their properties will accordingly be briefly considered. Tungsten steel usually contains from 3 to 10 per cent of tungsten * See footnote, page 511. THE SPECIAL ALLOT STEELS 523 and from 0.2 to 1.0 or even 2.0 per cent carbon. The tensile proper- ties of tungsten steel resemble those of high-carbon steel, the strength and especially the elastic limit being high, but the ductility low. After moderately rapid cooling from high temperatures, however, the tungsten steels exhibits remarkable hardness which is still retained upon heating to temperatures considerably above the ordinary tempering heats of car- bon steels. It is this property of tungsten which makes it a valuable alloy for use (in conjunction with chromium or manganese), in making of the so-called " high-speed " tool steels. The tungsten steel which contains about 4 or 5 per cent of tungsten and 0.5 to 0.7 per cent carbon possesses remarkable magnetic reluctance. It is a valuable material therefore for use in constructing permanent magnets, since when once magnetized, it will retain magnetism much longer than ordinary iron or steels. Structure and Constitution. Thermal Critical Points. A general explanation of the structure and constitution of tungsten steels, includ- ing an explanation of the structural changes which take place when the steel is heated to a high degree and cooled with moderate rapidity, has never been satisfactorily made. Bohler,* H. LeChateher,t Osmond,} H. C. H. Carpenter,! Swinden,|| Edwards, 1[ and others have, however, made important contributions to the study of the structure and thermal critical points of tungsten-carbonriron alloys. All of these investiga- tors found retardations, or critical points, in the cooling curves of tung- sten steels, and, furthermore, it was shown by several of these gentlemen that a new critical point far below the ordinary critical range of tem- peratures appears when the steel has been cooled from temperatures in the neighborhood of 1200° C. This critical point has generally been understood to be the Ari point of carbon steels, i.e., the temperature of formation of pear lite. Bohler first proposed the theory that a sufficient amount of tungsten (or tungsten and chromium) is able to lower the lowest critical point below ordinary temperatures so that the change to the comparatively soft pearlite condition does not occur. Osmond went further and stated that this point of conversion to the pearUte state is lowered in propor- * " Wolfram und Rapid Stahl," 1903. t " Revue de M6tallurgie," 1904, pp. 334-347. t " Contribution k la des Aoiers rapides," 1904, § " The types of Structure and the Critical Ranges on Heating and Cooling of High-speed Tool Steels under Varying Thermal Treatments," Jour. Iron and Steel Inst., Vol. 67, 1906, I, p. 433. II " Carbon-Tungsten Steels," Jour. Iron and Steel Inst., Vol. 73, 1907, I, p. 291. If Function of Chromium and Tungsten in High-speed Tool Steels," Jour. Iron and Steel Inst., Vol. 77, 1908, II, p. 104. 524 MATEEIALS OF CONSTRUCTION tion as the quantity of tungsten, chromium, etc., is raised above the eutectoid composition. Carpenter found, however, that the low critical point occurs at about 400° C, in all cases when tungsten, molybdenum, tungsten-chromium, or molybdenum-chromium high-speed steels are cooled from 1200° or 1250° C. Later, Edwards showed that all the alloys containing more than 6.0 per cent of tungsten and 3.0 per cent of chromium show a low critical point at about 380° C, when cooled from 1200° C, but exhibit no critical point below 900° C, when cooled from 1320° C. (almost the melting temperature). Edwards also observed that if the effect of the tungsten is simply a depression of the An carbon change point it is implied " that the low point is still the carbide of iron change, and a steel slowly cooled below this temperature ought to revert to the annealed condition. If so, on again heating to much below the lowering temperature, say 900° C, and inmiediately cooling, the carbon change point (Ari) ought to appear at the normal temperature, since an initial temperature of 900° C. has no effect on this point . . . This is not the case. Thus after cooling below the low point several times, it is necessary to soak this sample for a half hour before the Ari point appears at the normal temperatures, and even after this treatment the l,ow point is still visible." Accordingly, Mr. Edwards expressed the opinion " that the low point is not Ari lowered by tungsten, but that a carbide of tungsten is slowly formed at about 1200° C, which has a critical temperature quite independent of Ari." The influence of the chromium in tungsten-chromium steels was shown by both Carpenter and Edwards not to be in the direction of lowering the Ari point, but actually the reverse. Edwards therefore concluded that, since the tungsten-chromium steels cooled from 1320° C, did not show any low point, that " the carbide of chromium first formed at 1200° C. combines with the chromium at higher temperatures to form the double carbide. This double carbide is held in solution on cooling, even slow cooling, but on heating it is deposited from solution at 670° to 730° C, and is slowly decomposed at slightly higher temperatures, first into carbide of tungsten, and on soaking, iron carbide is again formed. . . . Thus the double carbide is held in solution even when slowly cooled in air. To prevent the carbide of tungsten being deposited, the steels must be quenched in an air blast; whilst the carbide of iron is only kept in solution by quenching in water." Tensile Properties. The tensile properties of two series of steels con- taining from 0.4 to 27.05 per cent tungsten, together with about 6.2 per cent carbon, and about 0.8 per cent carbon, respectively, are exhibited by the diagrams of Fig. 299. This diagram is based upon a series of tests THE SPECIAL ALLOT STEELS 525 made by Portevin.* It will be noted that for both series of steels the tensile strength and elastic limit increase with increase in tungsten until a maximum strength is reached with 10 per cent tungsten, and 12 per cent tungsten, in the respective cases of the high- and low-carbon steels. The ductility is also reduced with increase in tungsten. (These steels have' been, heated three ^ hours at 850° C, prior to testing.) The marked differ- ence in the properties of a high- and a low- carbon tungsten steel is shown by a comparison of the curves of Fig. 299. The important relation of the carbon content to the properties of tungsten steel is shown to better advantage, however, by the tests of Mr. Thomas Swinden, f which are sum- marized by the curves of Fig. 300. The steels comprising this series all contained approximately 3 per cent tungsten, while the carbon content varied from 0.144 to 1.07 per cent. They were normal- ized by heating to 950° C . , maintaining this temperature for fifteen minutes, and cooling freely in air. It is very evident that increase in carbon is beneficial to the strength and elastic limit of this 3-per-cent-tungsten steel, until a limiting per- centage of about 0.9 per cent is reached. The ductility naturally de- creases with increase in strength. (The maximum beneficial percent- age of carbon may be higher than 0.9 when the tungsten content exceeds 3 per cent.) * " Contributions to the Study of the Special Ternary Steels," p. 26a t " Carbon-Tungsten Steels," p. 294. Si-o V U-o— U- 1 1 O0.5 f.C, rbou High Carbon itee bI_ j^ —o 200000 190 000 S. 180 000 ™170 000 •^160 000 S-Co fboii to\v"Car|bott-Steel r' ' ■^ rs / (H ;'ens le Strentr'th Oarbon Conter s t) s / \ / / \ i > 1 130 000 u 120 000 \ \ f' ; V .M' / ^j V ■giooooo S 90000 g 80 000 to 70 000 g 60 000 ^Si^ ^/ V ^/ \ '' ^ ~~ T5- ~S 'V/' ' \ &-- s \ /" •^ i/ > ^ "/a ~C7 .Si/ *-., -.. / / s ii X '^-. J ;> .^ # ^ ••i o 40 ^ To Oeu CLo\J tRe Car bon on :ontJ Are nt) \ \ 1^" \ ^ S 20 5,n V' /^ -<^ >^ Per Cen Red iSic nof Aral \ V -i^ Ueh parb mC mte: i>o ^ . Fig. 299.- 4 6 8 10 12 14 16 18 20 22 21 26 Percentage of Tungsten -Tensile Properties of Tungsten Steels. (Portevin.) 526 MATERIALS OF CONSTEUOTION The upper curves of Fig. 300 indicate that increase of carbon con- tent beyond about 0.3 per cent is detrimental to the ability of the steel to withstand cold bending or alternating stresses. With respect to the alternating stress tests it may be added that the showing of these steels is not equal to that of ordinary boiler-plate steels, which are expected to stand about 350 te- versals of bending stress in the Arnold machine. Manufacture and Uses. Tungsten steel is made almost exclusively by the crucible process,the tung- sten being added in the form of ferro-tungsten, as wolframite (WO4), or less commonly, as metal- lic tungsten. Tungsten steel is used for few purposes other than for machine tools. These must be forged roughly to form and fin- ished by grinding, fhe great advantage of this steel over ordinary high- carbon tool steel is its ability to hold its hard- ness at high tempera- tures, thus making it possible to run machines at high speed with heavy cuts. It is excelled in this respect, however, by the special steels classed as " high-speed " tool steels, which contain chromium or manganese in addition to tungsten. The uses of tungsten steel as a magnetic steel has been mentioned above. 506. Molybdenum Steel. The action of molybdenum in steels is generally accepted to be exactly similar in character to that of tungsten in so far as the influence of the alloy upon critical temperatures, hardening power, physical properties, etc., are concerned, but the effect of molybdenum differs in magnitude from that of tungsten, 1 per 0.1 0.2 0.3 0.4 0.6 0.6 0.7 0.8 0.9 1.0 1.1 1.3 1.3 Percentage ol Carboa Fig. 300. — Relation of Tensile Properties of 3-per-cent- Tungsten Steels to Carbon Content. (Swinden.) THE SPECIAL ALLOY STEELS 527 cent of molybdenum being apparently equivalent to 2 or 3 per cent of tungsten. Most molybdenum steels are really quaternary alloys, since chro- mium or manganese is usually present in notable amounts. The best molybdenum ternary alloy steels contain from 1 to 2 or 3 per cent molybdenum and not more than 2 per cent carbon. The general physical characteristics of molybdenum steels are identical with those of tungsten steel. Structure and Constitution. Thermal Critical Points. Mr. Thomas Swinden concluded as a result of an extensive study of the constitution of carbon molybdenum steels * that " molybdenum does not exist as double carbide, and is not in solid solution in the iron, but is probably dispersed in the ferrite in a manner suggesting the existence of a solid colloidal solution of an iron-molybdenum compound in iron." This condition is of course in opposition to Edward's theory respecting tung- sten steels above quoted. Swinden shows by a series of cooling curves that if the initial tem- perature has not exceeded a certain definite minimum the steel behaves in every way like plain carbon steel. If the iriitial temperature has exceeded a certain minimum " lowering temperature," however, a new low critical point appears far below the normal position of Ari, just as has been shown to be the case with tungsten steels. The minimum lowering temperature is somewhat lower in molybdenum steels than in tungsten steels, and when the molybdenum content is high (8 per cent) appears to be as low if not lower, than the Ac point. The temperature at which the maximiun lowering effect is derived increases with increase in carbon and with increase in molybdenum. Thus a steel containing about 1.2 per cent carbon and about 1.0 per cent molybdenum showed the lowest critical point obtainable upon cooling from 1000° C, while a steel of about the same carbon content, but having 4.0 per cent molyb- denum, was similarly affected only by a temperature of 1200° C. The position of the low point is absolutely independent of the car- bon content, but becomes lower as the molybdenum content increases. With 1.0 molybdenum its position was found to be about 560° C, but with 4 per cent molybdenum it is probably below 440° C. The actual explanation of the behavior of molybdenum steels upon cooling with moderate rapidity from the temperature above called the " lowering temperature " is still a subject of much controversy, just as is the case with tungsten steels and the quaternary steels which exhibit similar properties. The fact of the greatest practical importance, however, is that such heat treatment does cause some structural change * " Carnegie Scholarship Memoirs," Vol. 5, 1913, p. 100. 528 MATERIALS OF CONSTEUCTION the result of which is the acquirement of intense hardness, which is not lost on reheating until temperatures considerably above that temper- ature at which ordinary steel assumes the soft pearlite state is reached. Tensile Properties. The tensile properties of a series of molybdenum steels containing from 1.0 to 8.0 per cent molybdenum and 0.2 to 1.2 per cent carbon are exhibited by the curves of Fig. 301. These tests constitute a portion of Swinden's study of molybdenum steels above referred to. a 260 000 0.2 0.1 0.6 0.8 1,0 1.2 0.2 0,4 0,6 0.8 1.0 1.2 0.2 0.1 0.6 0.8 1,0 1,2 0,2 0,1 0.6 0.8 10 12 Feccentage of Carbon Fig. 301. — Tensile Properties of Molybdenum Steels. (Swinden.) Tensile Strength " Elastic Limit °- N. Normalized, 15 min. at 900° C, Air cooled. A. Annealed, 5 hrs., at 950° C, Cooled in Furnace. H. Hardened and Tempered, Quenched at 950-800° C. in Oil. The diagram is self explanatory and it will not be necessary to dis- cuss it beyond the point of simply calhng attention to the fact that with a given molybdenum content the strength and elastic limit are increased and the ductiUty decreased rapidly as the carbon content is raised. Quenching from a temperature of from 800° to 900° C. is very, bene- ficial to tensile properties with the exception of ductility. Ductility as Indicated hy Cold Bending. The results of cold-bend- ing tests of the steels of the series used for tensile tests are shown by the THE SPECIAL ALLOY STEELS 529 curves of Fig. 302. Contrary to what might have been expected increase in molybdenum content seems to improve the ductility of steels contain- ing a given amount of carbon (providing they have not been quenched). This fact is shown by both the elongation in tensile tests and the degree of bending cold. Alternating Stress Resistance. Fig. 302 shows that molybdenum steels do not rank exceptionally well when compared with carbon steels on the basis of their relative performance under alternating stress. The steels containing from 2 to 4 per cent molybdenum seem to rank highest, 0.2 0.4 0.6 0.8 1.0 l.li 0.2 0.4 0.6 0.8 1.0 1.2 0.2 0.4 0.6 0.8 1.0 1.2 0.2 0.4 0.6 0.8 1.0 1.2 Percentage of Carbon Fig. 302. — Cold Bending, Alternating Stress, and Brinell Hardness of Molybdenum Steels. (Swinden.) but the carbon content is a much more potent factor than the molyb- denum content unless the steels have been quenched. Brinell Hardness. The Brinell hardness, as is shown by Fig. 302, increases rapidly with the carbon content, and seems to be only slightly affected by the molybdenum content when the steels have not been quenched at temperatures above the minimum lowering temperature. (No data have been found available to show the hardness of these steels after cooling from the lowering temperature.) Manufacture and Uses. Molybdenum steels are made in the crucible in the same manner as tungsten steels, and have practically the same 530 MATERIALS OF CONSTRUCTION uses. The manufacture of these steels in the United States is controlled by patents and the production has been limited by this circumstance. 507. Silicon Steel. Silicon steels resemble nickel steels in their general properties, but have not been used as extensively as a material of engineering construction. When 1 or 2 per cent of silicon is combined with from 0.1 to 0.,4 per cent of carbon the resultant steel is one which may be classed as a high-strength structural steel resembling nickel -structural steel. It is chiefly valuable because its elastic limit is very high, compared with that of an ordinary steel of equal carbon content. It is rather hard, however, and gives some difficulty in rolling. The most valuable silicon steel is the one which is made particularly for use in electrical machinery and which was developed by Hadfield and patented in this country in 1907. Hadfield's silicon steel contains about 3 per cent of silicon (2.75 per cent recommended) and the smallest possible amounts of carbon, manganese, and other impurities. This steel acquires its remarkable magnetic properties (very high perme- ability and low core loss) only after a special heat treatment. It is heated to between 900° and 1100° C. (1070° C. recommended), cooled quickly to atmospheric temperatures, reheated to between 700° and 850° C. (750° C. recommended), and cooled very slowly. Sometimes it is again heated arid cooled very slowly from about 800° C. The magnetic properties of this class of steel have been discussed above. (See Art. 497 and Fig. 277.) Tensile Properties. The tensile properties of a series of very low- carbon silicon steels are exhibited by Fig. 303. These curves are' based upon tests made by Mr. Thomas Baker in 1913.* These steels do not excel ordinary medium high-carbon steels in tensile strength nor duc- tility, but do show a very high elastic Hmit. In the presence of so small an amount of carbon (0.04 per cent), the addition of silicon seems to be beneficial to strength only in amounts not exceeding about 5 per cent, and all of these steels acquire great brittleness if about 2.0 per cent silicon is exceeded. Similar tests made by Hadfield at an earlier date f with steels con- taining about the same range of silicon content, but having from 0.14 to 0.26 per cent carbon (Fig. 303), exhibit about the same properties as the very low-carbon steels used by Baker, except that the strength and elastic limit is slightly higher. Hadfield found forging of his steels impossible when the silicon content exceeded 5.5 per cent. * " The Influence of Silicon on Iron," Jour. Iron and Steel Inst., Vol. 64, 1913, II t Jour. Iron and Steel Inst., 1889, II, p. 222. THE SPECIAL ALLOY STEELS 531 Manufacture and Uses. Silicon steel may be made in the crucible, but is more often made in the open hearth. The alloy element is added in the form of ferro-siUcon. A small amount of silicon steel has been used as a hard, high-strength, structural steel for purposes similar to those above noted in discussing nickel steel. The principal conunercial application of this alloy, however, is as thin sheet steel for the construc- tion of the cores, pole pieces, etc., of electrical machinery. 508. Vanadium Steels. Vanadium is, with the single exception of car- bon, the most powerful element for alloying with iron yet discovered. Only 0.1 to 0.15 per cent of vanadium raises the tensile strength and elastic limit of low- or medium-carbon steel 50 per cent or more without any RoUsd- Tested OB rolled (No treatment) Normal -Heated to 1000 C. end oooled In air. Aimealed- Soaked 40 brB.atOGO C. oooleil slowly during 170 hour period. In the toBtB of.Bll BteelB contolnlDg 4.0 per cent Billcon or more, the eleBtlo limit coincided with the tensile Btrongth and the elongction wee nil. 1.0 2.0 3.0 4.0" 6.0 6.0 7.0 " 8.0 Percentage of Silicon Fig. 303. — Tensile Properties of Steels Containing 0.04 per cent Carbon. (Baker.) sacrifice of ductility. With high-carbon content (about 0.8 per cent carbon), and about 0.2 per cent vanadium, the tensile strength is about equal to that of an ordinary steel of corresponding carbon content, but the elastic limit is very much higher and the steel is much more ductile than a similar carbon steel. Little more than 0.2 per cent vanadium seems to be advantageous in any steel, and amounts exceeding about 0.3 per cent are very detri- mental to strength. Vanadium steels may be forged or rolled with only minor special precautions; they respond readily to heat treatments; are enormously strong when hardened by quenching and moderately tempered; and are very tough, and stand impact, vibration, or reversal of stress very well. 532 MATERIALS OF CONSTRUCTION Vanadium has an important quieting influence upon molten steel when cast in the ingot, and therefore promotes soundness by prevent- ing the occlusion of gases. The use of vanadium in steels designed for a great variety of pur- poses is becoming more common every year, and, in addition to the ternary alloys of vanadium carbon and iron, vanadium is used in a variety of quaternary alloy steels in which chromium, nickel,- etc, are also present. Structure and Constitution. The structure and constitution of vanadium steels, either normal or heat-treated, do not differ from those of corresponding carbon steels. The vanadium appears to exist for the most part as a carbide, but a small amount is usually also present in the free ferrite. The thermal critical points seem not to be markedly affected by the small amount of vanadium used. The beneficial efifect of vanadium upon the strength and ductility of steels cannot be definitely explained, but is probably 'due partly to its beneficial effect upon the behavior of the molten metal above noted. The formation of carbide and the presence of vanadium in the ferrite may also constitute advantageous factors. Tensile Properties. The ten- sile properties of two series of steels containing about 0.2 per cent, and about 0.8 per cent carbon, respectively, are shown by Fig. 304. The tests represented by this diagram constitute a portion of Portevin's study of ternary alloys referred to above. The beneficial effect of from 0.1 to 0.3 per cent vanadium upon tensile strength and, more particularly, the elastic limit, is plainly shown by these tests. The remarkable ductiUty of these steels, as indicated by the percentage of reduction of area, is also worthy of special note. 123466789 10 Percentage o£ Vauadium Fig. 304. — Tensile Properties of Various Vanadium Steels. (Portevin.) THE SPECIAL ALLOY STEELS 533 The important influence of carbon content and heat treatment upon the tensile properties of a series of 0.2-per-cent-vanadium steels is shown by Fig. 305. This diagram is based upon another of the wonderfully painstaking researches of Messrs. McWilliam and Barnes.* This diagram shows that the tensile strength and yield-point of 0.2-per-cent-vanadium steels increase rapidly with increase in carbon 200 000 190000 180 000 no 000 M 160 000 2 150 000 .^ 140 000 .£1 1-1 130 000 120 000 •g 110 000 S 100 000 ■a S 00 000 (« •a 8O00O o ej ^ 7O00O g 60000 m so 000 V S 40 000 P H 30 000 20 000 |2» HEAT TREATMENTS R. None.Tested as Received N. Norinalized.900°C.30 min.,Cooled in Air A. Annealed.950°C.18 hra., Cooled in Furnace Y. Quenched from 850°G., Tempered at lOO^O. X. Quenched from SljO°C., Tempered at SOO^C. W. Quenched from 850 C, Tempered at 700 C. — Tensile Stength — Yield Point 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 Percentage of Carbon Fig. 305. Relation of Tensile Properties of 0.2-per-cent-Vanadium Steels to Carbon Content and Heat Treatment. (McWilliam and Barnes.) until a steel containing about 1.0 per cent carbon is reached. Beyond this point increase in carbon content is not beneficial, and may be the reverse. The ductiUty decreases only very slowly with increase in carbon, and is still considerable when the 1 per cent carbon steel is reached. * " Influence of 0.2 Per Cent Vanadium on Steels of Varying Carbon Content." Jour. Iron and Steel Inst., Vol. 83, 1911, I, p. 294. 634 MATEEIALS OF CONSTRUCTION The diagram also shows that vanadium steels are very responsive to heat treatments, the effect of the hardening treatment followed by moderate tempering being especially remarkable because of the ex- tent to which the yield-point and, less notably, the tensile strength is raised. Manufacture and Uses. Vanadium steels were originally made in the crucible, but are now commonly made in large masses in the open hearth. The vanadium is added in the form of ferro-vanadium, intro- duced after the recarburizer has been added when the conditions of the process are reducing, rather than oxidizing, in order to avoid loss by oxidation. Vg,nadium steel is used to a considerable extent for castings and forgings for machines, automobiles, and railway rolling stock. It is also used for axles, springs, shafting, and gears, and a structural grade of vanadium steel has recently been introduced, the material being rolled to the ordinary structural shapes, and used in the construction of long- span bridges and other structures subjected to severe conditions of service. Quaternary Alloys 509. The Quaternary Alloys. Many of the quaternary alloy steels are no less important commercially than the ternary alloys discussed in some detail above. Unfortunately, however, the quaternary alloys have not been as systematically studied as have most of the ternary alloys, and no effort will therefore be made to consider separately the properties of each one in detail. The most comprehensive study of quaternary alloys available is that made by Dr. Leon Guillet in 1906,* and all of the data quoted hereinafter is derived from Dr. Guillet's valuable paper. On account of the fact that the science of metallography has developed to so great an extent since the date of this study, the section of the paper devoted to micrography will not be considered. The study of the mechanical properties of quaternary alloys will be extensively quoted, however. On account of the number of variable factors involved, the proper- ties of the quaternary alloys cannot conveniently be represented by curves. For this reason the following series of tables are inserted. The tables are intended to be self-explanatory, and it is therefore con- sidered unnecessary to discuss the results of the very large number of tests which are summarized. The author has omitted a certain portion of the tests which seemed * " Quaternary Steels," Carnegie Scholarship Memoirs, Jour. Iron and Steel Inst., Vol. 70, 1906, II, pp. 1-142. THE SPECIAL ALLOY STEELS 535 least significant, has rearranged the data for convenience in reference, and has converted the strength values from metric to English units. The tensile specimens used were round bars, 13.8 millimeters in diam- eter, and 100 millimeters between gauge points (J inch by 4 inches, nearly). All specimens marked " normal " had been normalized by heating at 900° C, followed by slow cooling. The " quenched " specimens were quenched in water from 850° C, except certain of the tungsten-chro- mium steels whose special treatment is noted. .PROPERTIES OF QUATERNARY ALLOYS NICKEL-TUNGSTEN STEELS -Approx. Coiiiposition. Tensile Strength, Lbs. per Sq. In. Elas. Limit, Lbs. per Sq. In. Elonga- tion, per cent in 4" Brinell Hardness. C. Ni. w. Normal. Quenched. Normal. Normal. Normal. Quenched 0.2 6.0 0.3 78,100 227,400 63,700 21.0 146 387 6.0 0.3 96,700 226,000 78,800 17.0 179 418 6.0 0.7 2.0 6.0 90,900 105,300 125,400 220,500 177,700 81,000 81,000 90,700 16.0 14.0 16.0 192 207 226 444 387 302 0.4 3.0 0.3 93,500 66,800 13.5 196 321 4.0 4.0 4.0 0.7 2.0 5.0 88,400 105,300 108,800 64,700 89,500 85,700 15.5 13.0 14.0 174 207 217 555 477 495 NICKEL-VANADIUM STEELS Approx. Composition. Tensile Strength, Lbs. per Sq. In. Elas. Limit, Lbs. per Sq. In. Elonga- tion, per cent in 4" Brinell Hardness. C. Ni. Va. Normal. Quenched. Normal. Normal. Normal. Quenched 0.2 6.0 0.1 0.3 0.5 0.7 86,700 108,100 103,100 119,400 218,900 223,200 204,700 226,700 69,700 81,800 82,700 93,100 24.5 18.0 1*9.0 15.5 166 192 235 235 302 293 321 321 0.4 3.5 0.1 96,400 . 69,700 21.0 179 402 3.0 0.3 98,300 73,900 20.0 196 460 3.5 0.5 103,800 183,500 81,000 19.0 183 460 3.0 0.7 109,500 79,900 16.0 166 375 536 MATERIALS OF CONSTRUCTION NICKEL-MANGANESE STEELS Approx. Composition. Tensile Strengtli, Lbs. per Sq. In. Elas. Limit, Lbs. per Sq. In; Elonga- tion, per cent in i" Brinell Hardness. C. Ni. Mn. Normal. Quenched. Normal. Normal. Normal. Quenched 0.15 2.0 5.0 7.0 15.0 201,300 99,800 218,700 140,800 41,700 6.0 "7'0 311 364 187 340 444 187 12.0 5.0 7.0 15.0 152,800 87,800 93,400 122,800 89,300 58,600 50,200 65,400 15.5 36.5 35.5 212 146 170 223 131 118 30.0 5.0 7.0 100,400 102,700 92,000 83,600 38,400 38,400 30.0 22.0 149 124 126 107 0.75 2.0 5,0 7.0 15.0 105,000 114,200 107,000 91,400 116,000 62,000 70,500 3.0 18.0 277 235 212 255 223 183 12.0 5.0 143,400 59,200 11.5 174 153 30.0 5.0 7.0 132,600 123,600 104,800 102,700 36,300 38,400 32.0 30.0 174 196 153 137 CHROMIUM-MANGANESE STEELS Approx. Composition. Tensile Strength, Lbs. per Sq. In. Elas. Limit, Lbs. per Sq. In. Elonga^ tion, per cent in 4" Brinell Hardness. C. Mn. Cr. Normal. Quenched. Normal. Normal. Normal. Quenched 0.2 3.0 3.0 5.0 136,300 174,200 168,600 154,300 89,800 174,200 5.0 0.0 293 444 364 418 10.0 3.0 5.0 126,000 108,300 155,400 117,200 68,900 33,400 4.0 29.0 248 196 444 234 15.0 3.0 125,700 122,800 40,100 19.5 114 105 0.8 2.0 3.0 5.0 136,000 118,700 153,300 138,700 119,800 92,500 8.0 10.0 364 302 477 418 12.0 3.0 5.0 101,600 122,600 97,400 105,700 50,100 58,700 25.0 14.0 183 217 159 179 THE SPECIAL ALLOY STEELS 537 NICKEL-CHROMIUM STEELS Approx. Composition. Tensile Strength, Lbs. per Sq. In. Elas. Limit, Lbs. per Sq. In. Elonga- tion, per cent in 4" Brinell Hardness. C. Ni, Cr. Normal. Quenclied. Normal. Normal. Normal. Quenched 0.20 5.0 3.0 10.0 143,800 162,100 203,400 241,800 117,800 98,100 10.0 8.0 248 402 444 418 12.0 3.0 10.0 20.0 236,000 174,800 130,800 276,700 129,300 236,000 93,800 110,600 6.0 14.0 20.0 430 277 225 475 332 217 30.0 3.0 10.0 98,100 128,700 91,700 125,600 69,700 97,900 26.0 10.0 121 143 137 196 0.80 5.0 3.0 10.0 21,750 204,700 21,750 20,480 1.5 2.0 255 555 402 512 12.0 3.0 10.0 174,100 118,000 209,000 130,800 113,100 56,800 29.5 33.5 311 286 302 196 0.30 2.5 0.5 1.0 2.0 3.0 6.0 77,800 78,200 98,300 197,600 230,400 'i76,bbb' 190,500 230,300 266,000 60,600 57,200 65,100 197,600 191,900 24.4 23.0 20.0 0.0 4.5 137 146 166 275 248 418 430 430 444 444 0.20 6.0 0.5 1.0 2.0 5.0 6.0 108,100 170,500 223,300 238,800 155,000 203,400 228,700 201,900 187,700 214,600 64,700 146,500 174,800 193,300 155,000 18.0 12.0 7.0 6.0 0.0 183 286 375 402 460 277 460 418 387 402 MANGANESE-SILICON STEELS Approx. Composition. Tensile Strength Lbs. per Sq. In, Elas. Limit. Lbs. per Sq. In. Elonsa- tion, per cent in 4' Brinell Hardness. C. Mn. Si. Normal. Quenched. Normal. Normal. Normal. Quenched 0.2 2.0 0.5 0.7 1.5 70,600 79,300 85,600 93,100 151,900 164,200 40,600 57,900 65,900 17.5 15.5 14.0 107 107 153 234 248 269 15.0 1.0 2.0 112,500 134,300 100,100 153,100 33,200 70,600 10.0 5.0 212 248 196 300 0.5 0.5 1.0 2.0 109,000 118,300 71,500 77,100 12.0 10.0 248 293 12.0 0.5 1.0 2.0 103,000 107,700 103,400 68,600 65,900 69,800 19.0 15.0 10.0 196 202 202 538 MATERIALS OF CONSTRUCTION NICKEL-SILICON STEELS Approx. Composition. Tensile Strength, Lba. per Sq. In. Elas. Limit, Lbs. per Sq. In. Elonga- tion, per cent in 4" Brinell Hardnes,?. C. Ni. Si. Normal. Quenched. Normal. Normal. Normal. Quenched 0.2 2.0 0.5 1.0 1.5 2.0 5.0 72,600 84,700 104,500 114,300 137,900 239,500 196,300 183,500 179,000 162,800 51,000 50,700 73,100 74,200 108,700 20.5 16.0 19.0 13.0 16.0 124 126 179 163 235 418 351 351 430 277 6.0 0.5 0.8 1.0 2.0 5.0 85,000 103,800 110,900 166,300 96,700 219,000 209,000 142,900 69,700 59,700 78,200 125,200 96,700 23.5 23.0 19.0 11.0 0.0 170 170 163 248 286 340 321 304 444 556 12.0 0.5 1.0 2.0 3.0 5.0 7.0 200,500 204,300 241,500 167,300 202,600 225,900 196,300 165,300 288,500 230,700 153,600 173,000 200,500 204,300 241,500 167,300 202,600 225,900 1.0 0.0 9.0 0.0 2.0 0.0 351 375 375 402 430 460 387 444 375 460 - 418 387 30.0 2.0 ■ 5.0 88,000 94,100 85,400 94,100 34,700 30,700 36.0 40.0 143 159 140 143 0.4 3.0 0.5 1.0 1.5 2.0 3.0 7.0 93,200 124,400 112,900 138,700 152,200 64,700 80,200 87,000 101,700 120,800 20.0 12.5 6.5 14.0 2.5 170 174 202 228 262 286 477 477 555 0.8 2.0 2.0 5.0 180,000 119,100 100,200 98,100 8.0 9.5 293 351 49'5 512 12.0 2.0 147,100 147,100 0,0 153 364 15.0 1.0 2.0 3.0 118,000 120,800 240,300 118,000 120,800 240,300 0.0 0.0 0.0 143 179 196 333 364 364 20,0 1.0 3.0 133,700 170,600 98,100 103,800 32,700 38,400 33.0 27.0 143 134 116 166 THE SPECIAL ALLOY STEELS 539 TUNGSTEN-CHHOMITTM STEELS (Quenched steels heated to 850° C.) Approx. Composition. Tensile Strenfi;th Lbs. per Sq. In. Elas. Limit. Lbs. per Sq. In. Elonga- tion, per cent in 4" Brinell Hardness. C. Cr. W. Normal. Quenched. Normal. Normal. Normal. Quenched 0.2 2.0 2.0 16.0 77,800 96,100 140,800 92,100 39,700 36,100 17.0 10.0 126 153 261 170 3.0 13.0 91,000 47,800 15.6 166 163 10.0 2.0 15.0 243,700 122,800 186,800 115,300 210,000 47,100 4.5 10.5 477 196 375 225 20.0 2.0 71,600 103,200 36,800 20.0 179 166 0.4 3.0 13.0 130,900 219,000 58,300 10.0 223 351 0.5 1.0 13.0 116,600 82,200 228 430 3.0 5.0 8.0 13.0 20.0 103,900 127,300 124,800 91,400 128,000 64,000 93,100 72,500 75,400 16.5 9.0 11.5 7.0 166 217 217 156 364 430 364 196 8.0 13.0 127,300 150,900 106,100 6.0 228 321 0.6 3.0 13.0 130,500 197,700 67,300 12.0 228 364 0.7 3.0 13.0 136,000 82,200 10.0 288 314 ,0.8 2.0 2.0 15.0 179,100 222,000 143,800 199,000 4.5 0.0 518 652 555 ? 3;o 13.0 122,200 183,600 90,100 15.0 217 430 10.0 2.0 204,800 179,100 3.0 253 ? 20.0 2.0 15.0 127,400 112,300 114,200 115,100 53,600 65,400 10.0 18.0 207 179- 192 179 540 MATEEIALS OF CONSTEUCTION TTJNSTEN-CHROMItTM STEELS Effect of various heat treatments on hardness Approx. Composition. Treatment. BrineU C. Cr. W. 2 3.0 13.0 Oupnchpd at 1200° O in water. , . . . - 228 4 3.0 13.0 Quenched at 1200° C in water. 496 0.5 1.6" 13.0 Quenched at 1200° C in water. 460 3.0 5.0 8.0 13.0 13.0 20,0 Quenched at 1200° C in water Cracked 418 Quenched at 1200° C in water. 532 Quenched at 1200° C, in water Quenched at 1200° C, in water. 302 8.0 13.0 Quenched at 1200° C, in water. 512 0.6 3.0 13.0 Quenched at 1200° C, in water. Crackpd 0.7 3.0 3.0 13.0 13.0 Quenched at 1200° C, in water Cracked 0.8 Quenched at 1200° C, in water 532 0.5 3.0 13.0 Quenched at 1200° C, after heating 1 minute. . . Quenched at 1200° C, after heating 5 minutes . . Quenched at 1200° C, after heating 10 minutes.. Quenched at 1200° C, after heating 15 minutes. . Quenched at 1200° C, after heating 20 minutes. . Quenched at 900° C, after heating 1 minute Quenched at 900° C, after heating 5 minutes Quenched at 900° C, after heating 10 minutes. . . Quenched at 900° C, after heating 15 minutes.. . Quenched at 900° C, after heating 30 minutes.. . Quenched at 1200° 3., in air after heating 5 min. Quenched at 1200° C, in current of air 5 minutes. Quenched at 1200° C, in oil, 5 minutes.. 532 555 600 600 600 364 477 655 600 600 600 600 578 Quenched at 1200° C, in water at 15° C, 5 min. Quenched at 1200° C, in large volume of mercury 565 555 PART III THE NON-FERROUS METAI.S AND ALLOYS AND TIMBER CHAPTER XVII THE NON-FERROUS METALS AND ALLOYS THE PURE METALS General 510, The Non-ferrous Metals of Industrial Importance. The non- ferrous metals of greatest industrial importance comprise: copper, zinc, lead, tin, aluminum,' and nickel. Those of secondary importance include bismuth, antimony, cadmium, mercury and the precious metals — silver, gold, and pla,tinum. A number of these latter metals are chiefly important as alloy elements, and many others such as magnesium, chro- mium, cobalt, vanadium, tungsten, molybdenum, titanium, etc., have almost no industrial applications except as alloy elements. The non-ferrous alloys of greatest importance are the alloys of copper with tin (the bronzes), and alloys of copper with zinc (the brasses). Many important special bronzes and brasses are made, how- ever, in which a third alloy element has been included. For this pur- pose tin or zinc, lead, phosphorus, manganese, aluminum, sihcon, iron, and vanadium are most commonly used. Aside from the bronzes and brasses, copper forms more or less valu- able alloys with practically all of the metals above listed; zinc is the principal metal of a number of other important alloys, and the same thing is true of lead, tin, aluminum, nickel and a few others. The principal non-ferrous metals and alloys are Hsted in the classi- fication which follows: 541 642 MATERIALS OF CONSTRUCTION CLASSIFICATION OF NON-FERROUS METALS AND ALLOYS NON-FERROUS METALS A. Metals of Primary Importance (1) Copper (2) Zinc (3) Lead B. Metals of Secondary Importance (7) Bismuth (8) Antimony (9) Cadmium (10) Mercury C. Metals Chiefly Important as Alloy Elements (14) Magnesium (15) Chromium (16) Cobalt (17) Vanadium (4) Tin (5) Aluminum (6) Nickel (11) Silver (12) Gold (13) Platinum (18) Tungsten (19) Molybdenum (20) Titanium NON-FERROUS ALLOTS A. Ordinary Bronzes. Copper-tin Alloys. B. Special Bronzes. Copper-tin . Zinc Lead Phosphorus Manganese Silicon Aluminum Vanadium , Nickel C. Ordinary Brasses. Copper-zinc AUoys. Lead Aluminum . Manganese D. bpecial Brasses. Copper-zinc I Iron Vanadium Phosphorus Silicon Aluminum Manganese Phosphorus Silicon E. Other Binary Alloys of Copper. Copper., "j Chrom^iim Tungsten Antimony Bismuth Lead Arsenic I Lead i Tin Antimony [ Bismuth ) F. Alloys of Zinc. Zinc . Alloys. Alloys. Alloys. Alloys. THE NON-FERROUS METALS AND ALLOYS 543 G. AUoys of Lead. Lead . H. AUoys of Tin. Tin. I. Alloys of Aluminum. Aluminum . Tin Antimony Bismuth \ Alloys. Cadmium Arsenic Cadmium Antimony Bismuth \ Alloys. Nickel Zinc-lead Zinc Copper Magnesium Copperrzinc Copper-manganese Nickel } Alloys. Tin Manganese Tungsten Chromium Titanium J. Alloys of Nickel. Nickel-Copper-Zinc Alloys. K. Special Bearing Metals. Bronzes (Ordinary and Special) Lead-antimony Tin-copper-antimony Lead-tin-antimony Lead-copper-antimony Zinc-tin-antimony Lead-tin-bismuth Alloys. It would not be justifiable to devote space in the present chapter to a detailed consideration of all of the metalS and alloys above listed. In fact, detailed information is not available in the cases of many alloys. A. brief treatment of the metallurgy, properties and uses of the more important metals and alloys will be presented, supplemented by a few general statements concerning the properties and uses of the less important ones. COPPEE 611. General. Classification of Commercial Forms of Copper. Cop- per is, with the exception of iron, the most useful and valuable metal found in nature. Its ores are found in every important country of the world, and native copper is found in enornious quantities in the region abutting upon the south shore of Lake Superior, and, to a lesser extent, in a few other localities. Copper ores exist in a great variety, of forms, usually as sulphide or oxide. The greater proportion of the world's sup- ply of copper is derived from copper pyrites. 544 MATERIALS OF CONSTRUCTION Owing to the fact that copper exists native and may be reduced from its ores with comparative ease, it has been known and used since pre- historic times. The early Egyptians and some of their Asiatic con- temporaries were conversant with the art of extracting copper from its ores and making bronzes therefrom, and the prehistoric nations of America used copper in large quantities. Many articles of copper and bronze made by the Aztecs have survived to the present day and exhibit excellent quahties and very fine workmanship. The classification of copper most commonly used "in the United States is not a pairticularly rational one, but is one which trade conditions have imposed. The American Society for Testing Materials * recog- nizes three general classes of American copper, which may be defined as follows: A. Electrolytic copper is copper derived by the electrolysis of a copper sulphate solution with anodes of crude copper and cathodes of pure copper. (Copper migrates from the anode, leaving its impurities behind, and is deposited on the cathode.) B. Lake copper is copper which has originated on the northern penin- sula of Michigan, U. S; A. (Southern shore of Lake Superior.) C. Casting copper is more or less impure copper which is either (a) fire-refined copper from virgin sources; (6) copper electrolytically pro- duced by deposition from impure liquors, or (c) copper reclaimed from secondary sources. Electrolytic copper has in recent years largely replaced all other classes of copper for electrical uses, for alloying purposes, and for all other uses demanding a very pure grade of copper. A large propor- tion of the electrolytic copper produced is derived from copper pyrites and has previously been smelted and fire-refined. The process of elec- trolysis results in the recovery of practically all the precious metals com- monly present in copper ores, and since this gain, as well as the added market value of electrolytically refined copper, will usually pay the extra cost of electrolytic refining, the amount of casting copper produced is in consequence relatively very small. A considerable portion of even the lake copper is electrolytically refined, either to free it from impurities or to recover the silver content. Nearly 24 per cent of the output of lake copper is thus refined, but the product is usually either mixed with furnace-refined copper or sold as some brand of lake copper. 512. Occurrence in Nature. Ores of Copper. The most important pres of copper are the following: Native copper occurs in large quantities on the south shore of Lake Superior in the upper peninsula of Michigan. It is also found in New * Proceedings, Vol. 13, 1913, p. 206. THE NON-FEREOUS METALS AND ALLOYS 545 Mexico, in Chili, and in South Australia. The lake copper which has always been an important factor in the copper industry of the world is from this source. Copper pyrites or chalcopyrite (CuFeS2), is the ore from which the greater proportion of the world's copper supply is derived. It con- tains when pure, 34.4 per cent of copper, often contains both silver and gold, and is generally associated with iron pyrites. It occurs in the United States in Vermont, Virginia, Georgia, Tennessee, Alabama, and Montana. It also occurs in Spain, Portugal, France, Germany, Austro- Hungary, England, Ireland, Italy, Sweden, Norway, Russia, New- foundland, Canada, Cuba, Peru, Australia, and Africa. Chalcocite or copper glance (CU2S) contains 79.89 per cent copper. It occurs in large quantities in Montana, Arizona, New Mexico, and Texas, and is also found in England, Italy, Russia, Chili, and Australia. Bornite or erubesdte (CusFeSs) contains on the average 55.6 per cent of copper. It generally occurs intermixed with chalcopyrite and chal- cocite, and is doubtless a decomposition product, as is shown by its vari- able content of copper, ranging between 42 and 70 per cent. It occurs in Montana, also in England, Italy, and Chili. Cuprite (CU2O) contains 88.8 per cent of copper. It is abundant in New Mexico and Arizona and in South Austraha. It occurs also in Chili, Colombia, and Russia. Tile ore is an intimate mixture of cuprite and limonite. Cuprite is a decomposition product of sulphur com- pounds of copper. Melaconite or black copper contains 79.8 per cent of copper, but usually occurs contaminated with oxides of iron and manganese. Like cuprite, it is a decomposition product of sulphides and was formerly found abundantly in Tennessee, North Carolina, and Virginia. The black copper was long ago exhausted, however, and the original sul- phides are now being mined in some of the old workings. Malachite (CuCOs, Cu(0H)2) contains 57.33 per cent of copper and is another decomposition product of sulphides. It is abundant in Eastern Russia, in Chili, and in South Australia, and has been encoun- tered in Arizona and New Mexico. Azurite or chessylite (2CuC03, Cu(0H)2) contains 55.16 per cent of copper and is another decomposition product of the sulphides. It generally occurs with malachite, and has been mined principally in France (at Chessy) and in Russia. It has also been found with malachite in Arizona and New Mexico. Chalcanthiie or copper vitriol (CUSO4-I-5H2O) is a crystalline decom- position product containing 25.4 per cent of copper, or is found in solu- 546 MATEEIALS OF CONSTRUCTION tion in the waters of mines that are working on mixtures of sulphide ores and iron pyrites, etc. Atacamite (CuCl2+3Cu(OH)2) contains 59.4 per cent of copper and occurs plentifully in Chili and Peru. Chrysocolla is a hydrat^d silicate of copper containing an average of 39.9 per cent of copper. It is a decomposition product of sulphides and is generally associated with malachite. It occurs plentifully in Eastern Russia and in Chili. Covellite or indigo copper (CuS) contains 66.0 per cent of copper and is found to a considerable extent in Chili. Bournonite (PbCuSbSs), enargite ([4CuS+Cu2S]As2S3), and fahlore or tetrahedrite (4[Cu2S, FeS, ZnS, Ag2S, Hg2S]Sb2S3, AS2S3) are ores of rare occurrence and little commercial importance. 513. The Extraction of Copper from its Ores. The metallurgy of copper is very complex on account of the great variety of ores used and the frequent necessity of providing for the recovery of not only copper, but also the precious metals which occur in copper ores. The niunber of metallurgical methods of extraction of metallic copper is therefore very large. The chief sources of copper have been shown to be ores which are essentially more or less complex sulphides, ores which, although origi- nally sulphides, have by atmospheric agencies been altered to oxides and carbonates, and native copper ores wherein the copper exists as free metal. (A considerable proportion of the copper of commerce is derived from secondary sources, i.e., it is remelted or resmelted scrap.) Sulphide ores are usually treated by one of three general processes which will be briefly considered under the heads: (A) Roasting, smelt- ing, and , converting, {B) pyrite smelting, and (C) alternate oxidation and reduction. When the ores are wholly oxidized the copper may be recovered by a process of direct reduction. With only minor exceptions, all of the products from the foregoing processes must be further subjected to a refining treatment before being marketed. Only two general methods of refining possess any present importance. These are (1) electrolytic refining, and (2) fire or furnace refining. The former is now the method most commonly applied. Aside from the above-mentioned methods of derivation of copper, which may all be characterized as dry methods, wet or chemical methods, depending upon the getting of the copper into aqueous solution by means of suitable solvents and subsequently recovering it from the solution by electrolysis or by the use of suitable precipitants, have been experi- mented with for many years and perhaps may now be considered to have attained a position of some commercial importance, THE NON-FERROUS METALS AND ALLOYS 547 In the following paragraphs the several processes of derivation of copper above mentioned will be briefly considered. Roasting, Smelting and Converting. Roasting has for its principal objects the burning of the sulphur contents of the ore to sulphur dioxide (SO2), which passes away as gas, and the changing of the metal with which the sulphur was combined into an oxide. No effort is made, how- ever, to ehminate entirely the sulphur, or completely oxidize the metal, since some sulphur and lower oxides are desirable in the subsequent smelting operation. Two methods of roasting which have been practiced for generations are hea-p roasting and stall roasting. Heaps are built up on a bed of fuel in such a manner that the coarser lumps are in the center of the pile, while the smaller ore, in continually diminishing sizes, forms the exterior. The heap is covered with a layer of fines or concentrates. The fuel is cordwood, brushwood, heather, or coal, and passages are provided in the bed for draft. The shape of the heap is that of a truncated square pyramid, and its dimensions may be anything up to about 85 to 100 feet long, 40 to -50 feet wide, and 8 feet 6 inches high. According to the amount of copper and sulphur in the ore, and the height of the heap, the time required for roasting may be anything from about forty days to more than one hundred days. Stall roasting resembles heap roasting except that the heaps are enclosed within low masonry walls on three sides. Stacks are usually provided to increase the draft, the top of the pile being made air-tight by a layer of fines, or a cover provided for the entire stall. In stall roasting the time is considerably shorter than with heap roasting, and a smaller proportion of the ore is insufficiently roasted. Roasting in reverberatory furnaces similar to the furnaces used for smelting hereinafter described (except that, being worked at lower temperatures, the firebox is much smaller in proportion' to the size of the hearth) has been practiced for many years. The ore must be com- paratively finely divided (a conisiderable proportion of it much finer than \ inch), and this may mean an additional expense for crushing and grind- ing if this has not been previously required for purposes of mechanical concentration. Nevertheless, a considerable saving in time and fuel is effected by the use of the reverberatory roasting furnace in preference to the older methods of roasting out of doors in heaps or stalls. On account of the severe labor involved in stirring or rabbling the ore, mechanical rabbhng devices are often employed. I he most modern type of roasting furnace, Fig. 30G, is a shelf fur- nace containing six circular hearths one above the other. The hearths are provided with openings alternately at the center and at the periphery 548 MATERIALS OF CONSTRUCTION to permit of the descent of the ores. A vertical hollow shaft traverses the center of the furnace carrying rabbles which are set to move the material toward the outlet opening of each hearth. The air required for combustion is forced in by a fan at the bottom, and the products of combustion (principally sulphuric acid gas), escape through a flue pro- vided at the top. The ore to be roasted is delivered by hoppers to a dry- ing hearth on the top of the furnace; rabbles gradually transfer it to the peripheral opening leading to the top roasting hearth, and by the motion of successive sets of rab- bles it is gradually dropped from hearth to hearth, falling alternate- ly at the center and the periphery, until it leaves the lowest hearth in a roasted condition. - A temporary firebox must be used to heat a cold furnace to the required temperature, after which raw ore is charged and firing kept up until cal- cination has cormnenced and the ores have reached a dull red heat. Cal- cination thereafter con- tinues by the combustion of the sulphur of the ore without any extraneous fuel. The Herreshoff furnace illustrated is pro- vided with an air-cooling arrangement, air under pressure being forced up the central tube and through the hollow arms which carry the rabbles. Any desired portion of this heated air may be taken from the, top of the furnace and returned to the lowest hearth, thus raising the working temperature and making possible the roasting of ores quite low in sulphur without extraneous fuel. Smelting of copper ores has for its object simply the concentration- of the ore by removal of the earthy portion or gangue, in order that only the metallic portion may have to be treated by the subsequent more expensive refining process. It is not possible with sulphide ores to Fig. 306. — Herreshoff 6-Hearth Ore Roasting Furnace. THE NON-FERROUS METALS AND ALLOYS 549 derive metallic copper as metallic iron is derived in the blast furnace; because, though the ' smelting operation is similarly conducted as a reducing process, the copper did not originally exist as an oxide. Iron pyrites and other sulphide minerals are not rejected but, upon melting, join the metallic portion of the material in the furnace. It is only because some portion at least of the undesirable metalUc minerals have by the preliminary process of roasting been changed into the oxidized condition that their removal in the slag is in any degree possible. The ultimate prodixct of the smelting operation is therefore a large amount of worthless slag, made up principally of silica combined with ferrous oxide, alumina, and lime from the flux, and a small metallic por- r — ^ 1 in i ^ E i '^ i yE ^ £ i M,~iS-^% ft a r°^ S mF r "»| ' t S LONGITUDIMAL SECTION TRANSVERSE SECTION Fig. 307. — ^Water-jacketed Copper Blast Furnace. (Peters.) tion called matte, which is essentially a mixture of metallic sulphides of copper, iron, and other metals originally present in the ore. The. smelting of copper ores is accomplished in one of two general types of furnace, viz., the blast furnace and the reverberatory furnace. The two methods of smelting will be briefly considered separately. A modem type of copper blast furnace is illustrated by Fig. 307. This is a 56-inch by 51-foot furnace (inside measurements at tuyere level), and has an average daily capacity on roasted ore of about 1600 tons of charge. Much larger furnaces have been built by increasing the length without materially increasing the width between tuyeres, the number of crucibles and spouts being increased. The furnace of Fig. 307 has two crucibles B and B', located near the eiids of the shaft A. Sets of tuyeres D, D', and, D^, penetrate the side walls of the shaft, the crucible 550 MATEEIALS OF CONSTRUCTION tuyeres D' and D^ being at a slightly lower level than the hearth tuyeres D. The bottom of the shaft slopes in each direction from the center, so that the molten metal flows into the crucibles. The crucibles are pro- vided at their front ends with slag spouts G, which are trapped so that the outlet is considerably higher than the opening through the crucible wall. Slag flows continuously, and the blowing of the blast out through the slag spout is prevented by the head of slag maintained above the slag outlet by the trap. At the opposite end of the crucibles tap holes H are provided which allow a continuous stream of metal to flow into large forehearths. The mechanical operation of the furnace involves charging of the ore and fuel, removal of the slag and the matte, the collection of the flue dust and the volatilized metals, and the control of the blast. Charging is done by means of side-dump cars through the charging doors I, ore, flux, and fuel being charged alternately. The slag is gran- ulated by a stream of water which also serves to carry the slag to the dump. The matte is received in a cylindrical forehearth which holds a great quantity of metal and keeps it molten. The forehearth is tapped at intervals into a ladle which conveys the matte to the converters. The flue dust is recovered by dust catchers in the gas downtake, just as in the case of the iron blast furnace, and is subsequently resmelted, usually after having been briquetted. The volatilized metals are recovered by straining the gases through cloth. The equipment used for this pur- pose, the bag-house, will be mentioned hereinafter in connection with lead smelting. The blast has until recently been derived most often from rotary impeller blowers, but with the increased pressure and air-volume require- ments of larger furnaces positive piston blowers or blowing engines are now preferred. The blast pressure utihzed by the furnace above illus- trated is about 2| pounds per square inch. In the reverberatory furnace the fuel and ore are separated, not mixed as in the blast furnace, the fuel (usually either a long-flaming bitumi- nous coal, petroleum, or dry pulverized coal), being burned in a separate firebox, or introduced by burners at one end, and the ore is heated by the flame passing over its surface and the heat reflected from the arch of the furnace. Fig. 308 illustrates a large modern reverberatory furnace in which coal fuel is used. It is provided with a firebox 9 by 12 feet inside dimen- sions, a cross-wall, b, of brick divides it into two compartments, each of which is provided with firing holes in the roof and air inlets for forced draught. The hearth is 17 feet wide, about 44 feet long, and is served by four working doors d, and five charging holes in the arched roof. A THE NON-FERROUS METALS AND ALLOYS 551 door e is provided in the flue wall for skimming the bath. The arch and side walls, as well as the fire bridge, are of firebrick, while the hearth bottom is of quartz sand 24 to 36 inches deep, sintered in place. This size of furnace will smelt something like 200 tons of raw ore per day, the fuel requirement being about 1 ton of coal per 5 tons of ore charged. The reverberatory process is not a reducing process as operated for matte production. The charge must be in a ' condition favorable to quick melting, and therefore fine ore, not lumps, composed of silicious sulphide concentrate, roasted concentrate, and flue dust are used. Sev- eral ore charges are allowed to melt down, and then the matte is tapped as the converters require it. The slag is tapped off at a higher level from time to time, but neither slag nor matte is ever entirely tapped A- i„,.„i_ji I . (i Fig. 308. — Large Reverberatory Copper Smelting Furnace. out. The introduction of oil fuel in certain reverberatory furnaces in this country has made possible the smelting of higher silica ores without requiring that the charge be in so finely divided a state.' The reverberatory furnace furnishes a richer matte than the blast furnace, and on this account the molten materials from the two types of furnaces are sometimes mixed before conversion in the Bessemer converter. The conversion of the matte into metallic copper is accomplished in a Bessemer converter, the principle of operation of which does not differ essentially from that of the ordinary Bessemer steel converter. The converter is always side blown, however, and at the present time the lining is invariably a basic one -made of magnesite brick. Fig. 309 illustrates a modern type of copper converter of the upright type. Another common form of converter has the form of a horizontal cylin- der. It is mounted on trunnions, however, and operated in much the same manner as the upright converter. 652 MATERIALS OF CONSTRUCTION There are two main stages in the operation of Bessemerizing copper mattes. The first is essentially the elimination of the iron sulphide; the second the final sulphur ehmination. During the first stage the o^jsgen of the air blown into the molten matte forms oxides of iron, Sulphur, and copper, and the latter immediately reacts with the remain- i^^J^oii sulphide, reforming copper sulphide with the production of more iron oxide. The iron oxide now reacts with the silica which has been introduced with the charge or during the blow, and produces a great quan- tity of ferrous silicate slag which must be poured off at the end of this stage. The sulphur oxidizes to SO2, which is driven off. 60 H.P. Motor-626 RP-M. Ratio 63 to 19 'SIDE ELEVATION Fig. 309. — Modern Upright Copper Converter. The product of the first stage of the operation is a " white metal," which may be considered to be practically pure copper sulphide. The white metal is now blown to produce blister copper, the sulphur being eliminated by the action of copper oxide, first produced, on the copper sulphide present. Blowing to white metal requires about fifty to sixty minutes, and metallic copper is obtained after about sixty minutes' more blowing. The progress of the operation is judged prin- cipally by the appearance of the converter flames. Pyrite Smelting. The process of pyrite smelting is one developed in the United States under circumstances permitting of the exclusive use of a highly pyritous ore. The term "pyrite smeltmg " is applied to the smelting of a pyritous ore mainly by the heat of combustion of its THE NON-FERROUS METALS AND ALLOYS 553 own sulphide constituents. The great advantage of pyrite smelting, when practicable, is that it is a process of smelting raw ore, and the expense of the usual roasting treatment is therefore saved. Pyrite smelt- ing is practiced only by a few American smelters, and little will here be said concerning the process. Ores suitable for pyrite smelting contain a large amount of iron sul- phide, and the iron pyrites (FeS2) loses part of its sulphur by dissocia- tion in the upper portion of the blastfurnace, leaving iron sulphide (FeS). The iron sulphide now reacts with oxygen and sihca, forming a ferrous silicate slag and hberating sulphur dioxide. A considerable amount of uncombined silica or quartz is a necessary part of the charge, and the degree of concentration accomplished is controlled by varying the quartz charge. Pyrite smelting requires more flux than ordinary smelting, on account of the slag formed. It also makes more slag, the metal loss is greater^ and a stronger blast pressure is called for, but these disadvantages are more than offset by the advantages gained. Alternate Oxidation and Reduction. This process is the old Welsh process, still used to some extent abroad but long ago abandoned in this country. It consists in a long series of alternate calcinations and fusions in the reverberatory furnace. The removal of impurities is imperfect, but the sulphur is gradually driven off and the other impurities become segregated in a portion of the product. This is the origin of the English " best selected " copper. Refining of Blister or Coarse Copper. Electrolytic refining, has come to be the method by which all the finer grades of copper, such as are used for electrical purposes, are refined. Nevertheless, the older method of refining in a reverberatory furnace still possesses some commercial value and will be briefly considered. Fire refining is based upon the scorifying effect of cuprous oxide upon base metals contained in a batch of molten copper. Air is blown upon a bath of molten coarse copper contained in the hearth of a rever- beratory furnace. Cuprous oxide is very rapidly formed, and becomes dissolved in the bath. A very active oxidizing agent is thus intimately mixed with all of the molten material; the impurities become oxidized and form a slag which may be skimmed off, while the cuprous oxide is reduced to metallic copper. An excess of cuprous oxide usually remains after the removal of impurities has been carried to the practicable limit (traces always remain), and this must subsequently be reduced by the use of -a layer of charcoal or coke spread over the molten metal. Green poles are then thrust beneath the surface and in this way the dissolved cuprous oxide is gradually reduced until the oxygen content does not exceed about 554 MATERIALS OF CONSTRUCTION 0.05 to 0.07 per cent. No recovery of precious metals is possible, and the copper bars formed by casting the product are still relatively impure com- pared with electrolytic copper. Electrolytic refining has displaced other methods of refining, primarily because the product is purer than is otherwise generally obtainable. The fact that the precious metals are separated by the same operation is an important consideration, however. The fundamental principles of the process of electrolysis have been explained above and need not be elaborated upon. The crude copper cast into anode plates is electrolyzed in a strongly acid copper sulphate solution. Copper is deposited on the cathode, while the impurities, including the precious metals, are insoluble in the electrolyte and fall to the bottom, forming sUmes which are subsequently refined for recovery of silver, gold, etc. The purity of the cathode copper is dependent only upon the character of the electrolyte, though there is a practical limit to the degree of impurity of the anode copper. Electrolytic cathodes commonly contain about 99.95 per cent copper, the adulterants being principally hydrogen, with perhaps 0.02 per^ cent of metallic impurities. The cathodes are not generally marketed with- out' remelting in a reverberatory furnace, because their form is not one which is suitable for shipment, small pieces being easily broken off. The remelting usually partakes of the nature of a fire refinement in some measure, but this is only because of the contamination of the copper inevitably incidental to remelting. The extraordinary conductivity of cathode copper is slightly impaired by fire refinement. Lake copper is derived from low-grade native-copper deposits which are mechanically concentrated after crushing. The resultant " min- eral " contains about 85 per cent of copper. This is melted in rever- beratory furnaces, the slag skinmied off, and the metal is fire refined by the method above described. In certain cases where the impurities, particularly the arsenic, run rather high, and where the content of pre- cious metals is sufficient to pay the cost, electrolytic refining of lake cop- per is practiced. The resultant product is marketed as lake copper — not as electrolytic copper, however. Unless the metal is electrolytically refined, lake copper is usually sufficiently alloyed with arsenic, or phos- phorus, aluminum, silica, etc., to be rendered unfit for electrical purposes because of low conductivity. The presence of these elements is not necessarily undesirable, however, because they are for the most part help- ful in developing desirable mechanical properties. Wet or chemical methods of derivation of copper (sometimes called . hydrometallurgical methods or copper leaching) from low-grade ores or tailings are now being experimented upon by most of the important cop- THE NON-FERROUS METALS AND ALLOYS 555 per-producing companies of this country. The process is still in an experimental stage and details of practice vary considerably. In general the practice usually consists in subjecting the ore to lixi- viation by sulphuric acid in tanks or vats, no great difficulty being encoun- tered in getting the copper into solution. The precipitation of the dissolved copper is an operation which has caused more difficulty than the ffi-st stage of the process. The recovery of the copper by electrolysis is being tried extensively with considerable success. Recently th,e use of insoluble anodes of fused magnetite, and the introduction of sulphur dioxide gas into the electrolyte, thus reducing the current requirements, have been found to offer distinct advantages. Aside from the use of the electrolytic method, copper is most fre- quently recovered from solution by the use of sponge iron as a precipitant. The latter is obtained by a roasting of pyrites, followed by reduction of the iron oxide formed. The sulphur driven off in roasting the pyrite is used to make acid solvent. 514. The Properties and Uses of Copper. The properties of copper which possess the greatest practical importance are its electrical con- ductivity (or conversely, resistivity), and its tensile properties. Since something over 50 per cent of all the copper consumed in this country is used for electrical purposes, high conductivity may properly be con- sidered to be the first requisite of nearly all high-grade copper and ten- sile properties are next most important. Electrical Resistivity The maximum resistivity of various classes of copper and copper products permitted under the standard specifications of the American Society for Testing Materials * is as follows, the resistivity being ex- pressed in international ohms per meter-gram at 20° C. : Low-resistance lake copper wire bars (annealed) . 15535 High-resistance lake copper (minimum) 0. 15694 Electrolytic copper wire and cakes, slabs, and billets for electrical purpose " ■ 15535 Electrolytic coppei ingots and ingot bars, cakes, slabs, and billets not intended for electrical Uc,es . 15694 . \ diameters 0.460 to 0.325 in 0. 15775 Hard-drawn copper wire ^ ^^^^^^^^ „ 324 to 0.040 iii 0.15940 . j diameters 0.460 to 0.325 in 0. 15694 Medium-hard drawn ropper wire -j diameters 0.324 to 0.040 in 0. 15857 Soft or annealed copper wire 0- 15614 The electrical conductivity is principally dependent upon the purity of the copper, and the specifications for lake and electrolytic copper therefore contain the following stipulations as to the mininum metal content: * Year-book.. 1914. 556 MATERIALS OF CONSTRUCTION Class of Copper Minimum Metal Content Low-resistance lake copper 99.88 (silver counted as copper) High-resistance lake copper. . . . 99.88 (silver and arsenic couni^ed as copper) Electrolytic copper (all shapes) . 99.88 (silver counted as copper) Tensile Properties The tensile properties of hard-drawn, mediuna hard-drawn, and soft or annealed copper wire of various sizes are required to satisfy the following specifications of the American Society for Testing Materials: Area. Hard-drawn Wire. Medium Hard-drawn Wire. Soft Wire. Diameter. Minimum Minimum Tensile Strength. Minimum Minimum Minimum Mils. Tensile Strength. Elonga- tion. Elonga- tion. Tensile Strength. Elonga^ tion. Lbs. per Per cent Minimum Maximum Per cent Lbs. per Per cent Sq.in. in 10 Ins. Lbs. per Sq.in. Lbs. per Sq.in. m 10 Ins. Sq.in. in 10 Ins. 0.460 211,600 49,000 3.75 42,000 49,000 3.75 36,000 35 0.410 168,100 51,000 3.25 43,000 60,000 3.60 36,000 35 0.365 133,226 62,800 2.80 44,000 61,000 3.26 36,000 35 0.325 105,625 54,500 2.40 45,000 52,000 3.00 36,000 35 0.289 83,520 66,100 2.17 46,000 53,000 2.75 37,000 30 0.258 66,565 57,600 1.98 47,000 64,000 2.60 37,000 30 0.229 52,440 59,000 1.79 Per cent in 60 Ins. 48,000 66,000 2.25 Per cent in 60 Ins. 37,000 30 0.204 41,615 60,100 1.24 48,330 55,330 1.25 37,000 30 0.182 33,125 61,200 1.18 48,600 65,660 1.20 37,000 30 0.165 27,225 62,000 1.14 37,000 30 0.162 26,245 62,100 1.14 49,000 66,000 1.16 37,000 30 0.144 20,735 63,000 1.09 49,330 56,330 1.11 37,000 30 0.134 17,956 63,400 1.07 37,000 30 0.128 16,385 63,700 1.06 49,660 66,660 1.08 37,000 30 0.114 12,995 64,300 1.02 50,000 57,000 1.06 37,000 30 0.104 10,815 64.800 1.00 37,000 30 0.102 10,404 64,900 1.00 50,330 67,330 1.04 38,500 25 0.092 8,464 65,400 0.97 38,500 25 0.091 8,281 65,400 0.97 60,660 57,660 1.02 38,600 25 0.081 6,561 66,700 0.96 61,000 58,000 1.00 38,500 25 O.OSO 6,400 66,700 0.94 38,500 25 0.072 6,184 65,900 0.92 51,330 58,330 0.98 38,600 26 0.065 4,226 66,200 0.91 38,500 25 0.064 4,096 66,200 0.90 51,660 58,660 0.96 38,500 26 0,057 3,249 66,400 0.89 52,000 59,000 0.94 38,600 25 0.051 2,601 66,600 0.87 62,330 59,330 0.92 38,500 25 0.045 2,025 66,800 0.86 62,660 69,660 0.90 38,500 25 0.040 0.021 020 1,600 67,000 0.86 63,000 60,000 0.88 38,500 38,500 25 25 0.003 40,000 40,000 20 20 THE NON-FEEEOUS METALS AND ALLOTS 557 The gradual increase in the tensile strength requirement and the decrease in ductility called for, as the size of wire becomes smaller, is in conformity with the well-known fact that the tensile strength increases rapidly, and ductility decreases, as the amount of cold working to which the copper is subjected in drawing increases. The effect of partial and complete annealing is also recognized by the slightly lower strength values and shghtly higher degree of ductility called for in the cape of the medium hard-drawn wire, and the much lower strength and very much greater ductility required in the case of the soft wire. The yield point of copper wire does not appear distinctly in tests, and the stress-strain curve exhibits no sudden break, as is the case with mild and medium steels. Tests made by members of the committee which framed the above specifications established the average value of the elastic limit of hard-drawn wires at 55 per. cent of the specified ulti- mate tensile strength in the case of the wires more than 0.324 inch in diameter, and 60 per cent in the case of wires less than 0.325 inch in diam- eter. For medium hard-drawn wire the average value of the elastic limit was found to be 50 per cent of the specified tensile strength. The modulus of elasticity of drawn copper is usuallj' found to be in the neighborhood of 16,000,000 pounds per square inch. The tensile properties of copper in forms other than wire are not well known, nor are they commonly specified at all. Hot-rolled plate shows an elastic limit usually not above 7000 or 8000 pounds per square inch, and an ultimate tensile strength of from 30,000 to 40,000 pounds per square inch. By finishing cold or by cold hammering its elastic limit may be doubled, or more than doubled, while the tensile strength is increased in a lesser degree. Uses of Copper The relative proportions of the copper output of the United States appfied to various classes of uses are indicated by the following Govern- ment stati^ics: Forms in which Coppek was Cast in 1913, Pekcentage of Total * Per cent Wire bars (used principally for electrical purposes) 58 Ingots and ingot bars (for castings and for brass and bronze industries). . . 2.3 Cakes (principally for rolled copper) 9 Cathodes (principally for brass and bronze industries) 8 Other forms •, 2 100 * " Mineral Resources," 1913. 558 MATERIALS OF CONSTEUCTION According to the figures of Aron Hirsh & Sons in their " Copper Statistics for 1913 " * the consumption of copper in the United States was divided as follows: Per cent Electrical industry (including copper wire) 52. 1 Brass mills 28.7 Copper sheets 16. i Miscellaneous (chiefly castings and alloys) 5.5 100.0 Production of Copper 515. Statistics of Copper Production. The total production of copper from primary sources in the United States, as compiled from the smelter returns for the last foUr years, is reported as follows: * Pounds 1910 1,080,159,509 1911. . 1,097,232,749 1912 1,243,268,720 1913 1,224,484,098 Of the primary copper produced in 1913 about 86.7 per cent was electrolytic copper, about 9.6 per cent lake copper, and about 3.7 per cent casting and pig copper. Of the total production of copper includ- ing (1) that from primary sources, and (2) that produced from secondary sources by either regular refiners or by plants that treat secondary mate- rials exclusively in the production of copper and copper alloy products, the proportion of secondary copper from all sources amounts to about 14.4 per cent. (The production of lake copper in 1913 was abnormally low because of unfavorable labor conditions in the lake district. In 1912 and 1911 the proportion of lake copper was about 15 per cent of the total production of primary copper.) Zinc 516. General. Commercial Forms of Zinc. After copper and lead, zinc is the next most important of the non-ferrous metals. It occurs in some measure in almost every important political division of the world, usually as a sulphide, a carbonate, or a silicate. It is used not only as metaUic zinc, in which form it is known to trade as " spelter," but also in the form of zinc dust, which is formed in the distillation of zinc, and as zinc pigments, such as zinc oxide, leaded zinc oxide, and * " Mineral Resources," 1913. THE NON-FEEROUS METALS AND ALLOYS 559 lithopone. A considerable proportion (about 25 per cent) of the zinc used commercially is derived from secondary sources, and not directly from ores. 517. Occurrence in Nature. Ores of Zinc. The most important ores of zinc are zinc blende, calamine, and hemimorphite. Zinc blende or sphalerite (ZnS) is at present the main source of zinc production. It contains 67.15 per cent of zinc when pure, but usually contains manganese, iron, cadmium, etc. Marmatite is a common variety containing 10 per cent or more of iron as sulphide. Zinc blende is found in practically every one of the European States, also in Asia, Africa, Australia, South and Central America, Canada, and — in the United States — in Arkansas, Utah, New Mexico, Tennessee, Virginia, Kentucky, Iowa, New Jersey, Missouri, Pennsylvania, Wisconsin, and Colorado. Calamine or zinc spar (ZnCOs) was formerly the chief source of zinc, but has been relegated to a secondary position, owing to the exhaustion of many deposits. Calamine is rarely pure, but contains as a rule car- bonates of cadmimn, iron, and manganese. When pure, it contains 52 per cent of zinc, while the presence of other carbonates may reduce the zinc content below 40 per cent. Calamine occurs almost as univer- sally as does zinc blende, and is found, or at one time was found, in most of the districts in which blende occurs. Hemimorphite or zinc silicate (Zn2Si04+H20) contains 53.7 per cent of zinc when pure, and is often intermixed with, or underlies, de- posits of calamine. It occurs in Altenburg, in Sardinia, Spain, Tunis, and — in the United States — in New Jersey, Pennsylvania, Missouri, Kansas, Virginia, and Wisconsin. Franklinite (Fe, Zn, Mn)0, (FeMn)203, willemite (Zn2Si04), zinci- tite or red zinc ore (ZnO), and hydrozindte or zinc bloom (ZnCOs -|-2ZnH202) are comparatively rare zinc ores of lesser commercial importance. 518. Extraction of Zinc from its Ores. The chemical properties of zinc are so different from those of other common metals that the metallurgical methods by which metallic zinc is derived from ores are unique ones, and are, furthermore, complex, and metallurgically imper- fect. Whatever the original state of the zinc in the ore, it must be in the form of an oxide before the metallic zinc is obtainable. In this form it may be reduced by carbon at high temperatures, but this temperature is above the volatiUzing point of the metal, so that it is always obtained as a vapor which must be condensed. Moreover, condensation of zinc vapor must be accomplished at a temperature above the point of fusion if the metalUc zinc is to be obtained in the more commonly useful com- mercial form as spelter. Otherwise, i.e., by condensation at a temperature 660 MATERIALS OF CONSTRUCTION which produces soHd instead of Uquid metal, a powder called zinc dust or zinc fume is derived. This dust has certain special uses, but oxidizes very readily and cannot be remelted and cast to form spelter. A further complication is introduced by the fact that zinc cannot be reduced in the presence of even minute quantities of carbon dioxide without becoming oxidized, and it is therefore essential that the reduction be accomplished in the presence of an excess of carbon and in a closed fetort without access of air. Preliminary Treatment of Ores, Concentration, Calcination, and Roasting. All ores must be crushed to a finely divided state before being otherwise treated. Some measure of mechanical concentration is usually accompUshed at this stage, and recently so-called flotation concentration has been widely adopted. In this process the ore is reduced to a six-mesh size and the tailings from the jigs are finely pul- verized, thickened to three parts water to one part ore, mixed with a small quantity of oil and sulphuric acid, and agitated in a tank. A float concentrate is then taken off, the gangue goes to a second-treat- ment tank, where the process is repeated, and finally to a third treat- ment tank^ where the final concentrate is taken off and the remaining gangue discharged as worthless tailings. Calamine (zinc carbonate) and zinc silicate ores usually must be calcined prior to reduction and distillation, for the purpose of driving off the carbonic acid and water. Some anhydrous calamines and sili- cates are not calcined, but are reduced in the raw state. All American ores of these classes are so handled. Calcination, when practiced, is accompUshed in shaft furnaces, resembling lime kilns, and the operation is quite analogous to lime burning. Blende must be roasted prior to reduction and distillation, in order to convert the sulphide into oxide as completely as possible. Any zinc remaining as a sulphide is lost in the retort residues after distillation. The roasting operation results in obtaining zinc oxide, and is attended by the derivation of quantities of sulphur dioxide gas, which latter is usually collected and neutrahzed, because of its destructive attack upon "vegetation if allowed to escape freely near the ground level. Not infre- quently the SO2 is collected and used in a by-product sulphuric acid plant. Many forms of furnaces have been employed in roasting blende, but for the most part modern roasting furnaces are reverberatory multiple- hearth furnaces fired by coal or gas. Fig. 310 is a transverse section of such a furnace, the three lower hearths of which are muffled, i.e., the charge is out of contact with the products Of combustion. Mechanical means for agitating the material and the distribution and removal of same is provided. THE NON-FEREOUS METALS AND ALLOYS 561 Distillation and Condensation. The retorts in which distillation is accomplished are usually of cylindrical form, are about 8 to 10 inches in internal diameter, about 4 feet 6 inches long, closed at one end and provided with a connection with the condenser at the other end. Retorts are made with great care from specially selected clays and are molded under heavy pressure. Before firing, they are carefully dried, and when placed in the furnace must be red hot. The condensers are short clay, cylinders or cones, which are attached to the end of the retort in such a manner that they may be easily and quickly detached for emptying, although luted tight to the retort while in use. The distillation furnace consists of a long masonry chamber which is usually built with two parallel compart- ments, the side walls of which are simply open frames of firebrick. The retorts are placed in these chambers, the closed end being supported by projections on the longitudinal division wall, while the open end is carried by the brick frame-work of the face. The retort is so supported as to incline slightly downward and outward, and the condenser projects beyond the face of the brickwork, the space around it being tightly luted with clay ^^^^^^^^^^^^^^^^^^^^^^^^^^mmmz^ or clay and coal. Fig. 311 jtig. 310.- shows a simple form of distil- lation furnace which is much used. The furnace may be heated by coal burned on grates beneath, but is now commonly heated by gas, which is admitted at one end. Air for combustion is admitted at different points along the length of the furnace in order that the gas may be burned gradually. The furnace illustrated has 864 retorts, six rows high. Sometimes the number of retorts may exceed 1000 to a furnace. The operation of the process of distillation and condensation is briefiy as follows: mm -Hegeler Muffle Zinc Roasting Furnace. 562 MATERIALS OF CONSTRUCTION The retorts are charged (after the furnace is hot) with a mixture of crushed ore and coal which is packed tightly in place and then vented by thrusting in a small rod near the top of the retort. The condensers are next fitted in place and luted tight, and their mouths are partly stopped with coal to exclude air and conserve heat. The gas generated is ignited as it escapes from the mouth of the condenser until a smoky greenish flame indicates that zinc is beginning to come off. After six or eight hours enough zinc has been distilled to make it necessary to draw it by scraping the liquid zinc from the condensers to a ladle held below. The end of the condenser is now stopped with coal again and the process proceeds, the condensers being usually drawn three times, after which they are removed, the re- - mm M Bfc torts scraped clean of all / „-T— rTT~r~i — rTl — n~-7-~~. residue, and recharged. About twenty-four hours are required to work off a charge, which, for a furnace of the capacity illustrated, will be about 20 tons of ore. The molten metal caught in the ladle is skimmed and immediately poured into molds. The ladle skimmings,condenser scrapings, etc., are collect- ed and added to subse- quent charges. Under conditions where fuel is expensive this simple type of reverberatory furnace is impracticable, and resort is had to the Siemens regenerative principle, regenerators for the air and gas, special ports, and equipment for the reversal of direction of the gases through the entire system being provided. Refining Crude Spelter. The crude spelter, cast just as it comes from the distillation furnace, is invariably marketed without any further treat- ment, so far as practice in the United States is concerned. Many spelters made in Europe are excessively high in lead, however, and must be refined. The usual practice consists in melting the bars of zinc in a reverberatory furnace which provides for the preheating of the solid bars on a charging plate at the firing end before they are pushed into the bath ot metal. The temperature is carefully controlled, and is just suf- FiG. 311.— Hegeler Gas-fired Zinc Distillation Furnace. THE NON-FERROUS METALS AND ALLOYS 563 ficient to melt the zinc. The air supply is restricted in order to maintain a reducing flame and prevent oxidation of the zinc. The molten metal collects in a sump at one end of the hearth and, after the slag has been skimmed off, the pure zinc which floats on top is ladled off, the lead being drawn from time to time by means of an outlet provided in the bottom of the sump. A stratum of " hard zinc " (a difiicultly fusible alloy of zinc, lead, and iron), floats upon the lead and must be removed from time to time by means of perforated ladles which allow the escape of the more Uquid lead. 519. Properties and Uses of Zinc. Commercial metallic zinc or spelter contains varying amounts of impurities up to a maximum per- missible amount of about 1.5 per cent for the lowest grade. The prin- cipal impurities are lead, iron, and cadmium, according to the content of which spelters are divided into the following four grades by the Ameri- can Society for Testing Materials: * Grade. Maximum Content of Lead. Per cent. Maximum Content of Iron. Per cent. Maximum Content of Cadmium. Per cent. Maximun) Content of Lead, Iron, and Cadmium. Per cent. A. High grade B. Intermediate. . . . C. Brass special. . . . D. Prime western. . . 0.07 0.20 0.75 1.50 0.03 0.03 0.04 0.08 0.05 0.50 0.75 0.10 0.50 1.20 Grades A, B, and C shall be free from aluminum. The following tabulation is a summary of mechanical tests of cast zinc made by Gilbert Rigg and G. M. Williams, f The specimens used in the tests were taken from specially prepared castings made by pour- ing remelted spelter in cast-iron molds. The castings produced were 17| inches long and Ij inches in diameter. The upper 2^ inches, con- taining the pipe, was rejected. For transverse test specimens the rough castings IJ inches in diameter by 15 inches long were used, and were supported on a 12-inch span and centrally loaded. The tensile speci- mens were of the standard form used for gray iron castings, the diameter being 0.8 inch, and the length of the reduced section 1 inch. The com- pressive specimens were cylinders 1 inch in diameter and 2.6 inches long. The average values given in the following table are based upon four tests of each sample in each character of test : * Year-book, 1914. t Proc. Am. See. for Test. Matrls., Vol. 13, 1913, p. 669, 564 MATERIALS OF CONSTRUCTION THE STRENGTH OF CAST ZINC Grade. Analysis per cent. Ultimate Tensile Strength Lbs. per Sq.in. Bending Modulus of Rupture Lbs. per Sq,;n. Pb Fe Cd Max. Min. Ave. Max. Min. Ave. 1. High grade 2. High grade 1. Intermediate. . . . 2. Intermediate. . . . 1. Brass special. . . . 2. Brass special .... 1. Prime western No. 1 0.041 0.040 0.194 0.190 0.474 0.484 1.190 1.420 0.680 1.150 2.090 0.014 0,016 0.016 0.017 0.013 0.031 0.032 0.087 0.010 0.011 3.510 0,000 0.000 0.000 0.000 0.000 0.000 0.250 0.079 0.274 0.046 0.043 8,326* 4,973 5,451 6,088 13,821* 5,351 10,444* 4,377 13,543* 5,331 8,136 4,098 4,109 3.740 4,487 10,125* 4,098t 4,198 2,984t 7,122 3,640 6,326 6,262 4,330 4,340 5,095 11,980* 4,330t 7,710 3,700 10,800 4^670 7,340 12,150 11,060 10,660 13,820 22,220* 16,260* 12,830 10,750 18,390* 11,350 16,100 11,110 10,040 9,640 9,930 11,620 8,580t 8,440t 9,400t 13,300 9,390t 13,750 11,630 10,570 10,160 12,360 16,550* 13,110 11,020 2. Prime Western No. 1 10,050t 1. Prime Western No. 2. 16,250* 2. Prime Western No. 2 10,370 Dross ... 15,300 Grade. Total Deflection in Bending on 12-inch Span Inches. Compressive Strength 10 Per cent Compression Lbs. per Sq.in. Compressive Strength 20 Per cent Compression Lbs. per Sq.in. Max. Min. Ave. Max. Min. Ave. Max. Min. Ave. 1. High grade 2. High grade 1. Intermediate. . . . 2. Intermediate. . , . 1. Brass special. . . . 2. Brass special .... 1. Prime Western No. 1 0.280 0.300 0,360 0.440 0.430 0.280 0.160 0.200 0.190 0.220 0.070 0.220 0.200 0.260 0.150 0.160 0.160 0.090 0,100 0.160 0.170 0,040 0,250 0,250 0.300 0.310 0,290 0,210 0.130 0.130 0.180 0.190 0.050 17,710 16,700 17,380 17,180 20,910 18,030 29,450 22,780 29,920 22,080 35,880 15,170 16,130 15,150 15,180 19,000 16,190 27,210 20,020 27,040 17,710 35,750 16,600 16,400 16,000 16,130 20,110 17,190 28,270 20,950 28,700 20,080 35,820 25,080 24,000 24,260 23,950 28,580 25,670 38,870 31,390 39,980 29,990 23,500 23,430 22,100 22,150 27,290 23,680 38,640 29,280 38,760 24,230 23,640 23,030 23,150 28,070 24,830 38,770 29,970 39,490 2. Prime Western No. 1 1. Prime Western. . No. 2 2. Prime Western No. 2 Max. Load, Lbs.per sq.in. Dross 38,530 30,360 34,380 * High strength value explained by finely crystalline structure, t Low strength value explained by coarsely crystalline structure. Fig. 312 gives the average stress-deformation curves for the com- pressive tests of the above series. The effect of the common impurities on the properties of spelters may be summarized as follows: Lead in moderate quantities tends to make spelter softer in rolling, but weakens the coating formed in galvanizing. In quantities above about 0.7 per cent, it causes castings to crack badly. Iron hardens spelter and renders it more brittle, An excessive amount of dross is formed in galvanizing. :THE NON-FERROUS METALS AND ALLOYS 565 Cadmium hardens spelter greatly and makes it very brittle. It is therefore particularly undesirable in galvanizing, because it is easily cracked off. Cadmium also tends to cause cracking of castings. Spelter is largely used as a galvanizing coating; a considerable quan- tity is rolled into sheet zinc; a further quantity is used in making cast- ings, one variety of which, called " slush castings," are hollow because the metal is poured back into the ladle as soon as a thin layer next the mold has solidified; and a considerable amount is used in combination with copper and other metals in making brass and other alloys. 40000 35 000 0.05 O.iO 0.15 0.20 0.25 0.30 0.35 0,40 0,45 0.50 Compiession in 2.6 Inches Fig. 312. — Strength of Cast Zinc in Compression. (Rigg and Williams.) A considerable amount of zinc is also used in the form of zinc dust. A paste of zinc dust is used in the cotton-dyeing industry under the name of " indigo auxiliary." Zinc dust is also used in the making of hydrogen, in galvanizing by various dry methods, in making the electro- lyte used for electro-zincing, and in the cyanide treatment of gold and sil- ver ores. The various zinc pigments made from zinc or zinc ore are much used. Zinc oxide is the most important of these, but leaded zinc oxide, zinc- lead oxide, and lethopone (a mixture of chemically precipitated zinc sulphide and barium sulphate) are used to a certain extent as white pig- ments for mineral paints. 566 MATERIALS OF CONSTRUCTION A large amount of zinc in the form of zinc chloride is also used as a wood preservative. 520. Statistics^ of Zinc Production. The total production of zinc from primary sources in the United States for the last four years is reported as follows,* the figures being based upon the reports of spelter smelted: Short tons 1910 269,184 1911. . . ' 286,526 1912 .' 339,806 1913 346,676 In addition to the spelter produced" from primary sources the follow- ing amounts of zinc were derived from secondary sources by the dis- tillation or remelting of old spelter, by the recovery of metallic zinc from brass and other remelted alloys, and by the recovery of zinc in pigment and zinc chloride: Short tons 1910 68,998 1911 74,747 1912 94,111 1913 89,528 Lead 521. General. Commercial Forms of Lead. Lead has been one of the most commonly used metals since the earliest times of which we have any record, and the consumption of the commercial forms of lead is at the present time second to that of only one other non-ferrous metal, namely, copper. Lead occurs in almost every part of the world, usually as a sulphide, but sometimes as oxidized decomposition products of the original sulphide. It usually occurs associated with a small amount of silver, and often its ores contain a notable amount of antimony. Owing to the common occurrence of the silver and antimony in lead ores, and the further fact that these metals are not removed by the ordinary process of smelting, the output of lead smelters is classed under three heads as follows: Soft lead is derived by smelting ores which are normally so low in silver that desilverization is not necessary or practical. (Ores derived principally from Missouri and other Mississippi Valley States.) Desilverized lead is that obtained by special desilverizing treatment of the ordinary argentiferous lead ores. ' Mineral Resources," 1913. H( tc THE NON-FEREOUS METALS AND ALLOYS 567 Antimonial lead is that which carries an average of about 17 per cent of antimony and about 2 per cent of other metals such as arsenic, cop- per, etc. It is, of course, an alloy and is sometimes called " type metal," since it is often used in type-founding. It is made from the ore, not from pure metals. In addition to the forms of metallic lead above hsted, lead is used commercially in the form of pigments, and as litharge. The principal forms of lead pigments are leaded zinc oxide, which contains from 4 to 20 per cent of lead sulphate, the remainder being zinc oxide with a small portion of zinc sulphate; zinc lead oxide, containing 46 to 50 per cent of lead sulphate, 52 to 46 per cent zinc oxide, and a small quantity of zinc sulphate; sublimed white lead, which consists of 50 to 53 per cent lead sulphate, 41 to 38 per cent lead oxide, and small proportions of lead sulphide, lead sulphite, and zinc oxide. All of these load pigments are much used as white pigments in the manufacture of mineral paints. Litharge is lead oxide (PbO), containing 92.83 per cent of lead. It is produced direct from ore, or may be prepared by heating metallic lead in a current of air. It is used in making flint glass, in glazing pottery, as an oxidizing agent in many processes, and has many other commercial applications. 522. Occurrence in Nature. Ores of Lead. The only important ore of lead is galena (PbS), containing 86.57 per cent of lead when pure. It usually contains silver either as silver sulphide or as silver ores, the amount varying from the merest trace up to about 1 per cent (usually less than 0.2 per cent). Galena is found in almost every important country of the world. In the United States the principal deposits are found in Missouri, Idaho, Utah, Colorado, Nevada, Arizona, California, Montana, and New Mexico. Cerussite (PbCOa), containing 77.52 per cent of lead, anglesite (PbS04), containing 68.3 per cent of lead, and pyromorphite (Pb3P208), con- taining 69.5 per cent of lead, are all decomposition products and occur only in the upper portions of deposits of galena. Only 76.8 per cent of the primary load produced in the United States in 1913 was derived from the lead ores above listed; 11.4 per cent of the total was derived from lead-zinc ores, 8.2 per cent was derived from zinc ores, and 3.6 per cent was derived from all other ores.* In addition, secondary lead, amounting to about 16 per cent of the total quantity of. primary lead produced, was derived from old metal, alloys, skimmings, etc. The average yield of metallic lead per 100 pounds of ore was 5.6 pounds from the lead ores, 4.0 pounds from the lead-zinc ores, and 0.4 pound from the zinc ores. * " Mineral Resources," 1913. 668 MATERIALS OF CONSTRUCTION 523. Extraction of Lead from its Ore. The metallurgical processes involved in the extraction of lead from its ores comprise two main opera- tions, namely, roasting or sintering of the ore, and smelting in the blast furnace, and a number of secondary operations, including treatment of the matte which forms a portion of the products of the blast furnace, the recovery of dust from roasting and smelting furnaces and from matte converters, and the desilverizing of the lead if the silver content is sufficient to justify it. The methods of roasting lead sulphide ore (galena) in almost univer- sal use at the present day are two forms of a process known as " blast roasting." The process consists essentially in forcing air through finely divided metallic sulphide with the object of partially removing the sulphur, oxidizing the metal, and agglomerating the material in a form suitable for use in the blast furnace. The first of these methods, known as the Huntington-Heberlein process, is an intermittent process carried out in pots which utilize the up-draft principle. Fig. 313 shows the arrange- ment of the blast roasting-pot. The charge, consisting of a mixture of finely divided galena concentrate and limestone, is placed in the cast- iron kettle A which is provided with a perforated cast-iron plate / which serves as a grate. A baffle plate g distributes the blast which is admitted through the inlet pipe below. The detachable hood C is now fitted in place, and the gases driven off, together with the dust which they carry in suspension, are conveyed away by the off-take d and the dust flue e. The lime- stone in the charge serves a two-fold purpose: mechanically, it accelerates the speed of roasting by keeping the particles of sulphide separate so that they may become thoroughly oxidized, and, chemically, it acts as a flux which serves to agglomerate the partly roasted ore by the formation of a siUcate sinter. The portion of the charge first placed in the kettle is usually hot roasted ore in order to facihtate the starting of the operation, and the process is continued until the sulphur content is reduced to from 4 to 8 per cent. The sinter derived as a result of the operation is broken up before being charged into the blast furnace. The Dwight-Lloyd sintering machine, Fig. 314, utilizes the down- FiG. 313.^ — Huntington-Heberlein Lead Blast Roasting-pot. THE NON-FERROUS METALS AND ALLOYS 569 draft principle (with only a very shallow charge of ore to be penetrated) and is continuous in operation. The charge, in a finely divided state, is fed by a hopper into boxes or trucks called " pallets," which, in com- bination, form a conveyer, the continuity of which is broken at only one point in the circuit. The pallets are carried on wheels which engage guides by which they are supported. A coal-fired ignition furnace is supported above the trains of pallets, and is so arranged that the fiames from the firebox are defiected down upon the ore in each pallet in succes- sion, igniting same just before the pallet comes within the influence of the suction box. While the pallets are passing over the suction box their grate-like bottoms are in direct contact with the planed top of the suction box and all joints between the pallets are closed air tight. A down- FiG. 314. — Dwight-Lloyd Straight-line Lead- Ore Sintering Machine. ward draft through the ore to the suction box is maintained by an exhaust fan, and the ore already ignited on top is rapidly sintered. The sinter cake formed is discharged from each pallet in succession as the latter drops into the discharge guides at the end of the circuit. Sprocket wheels now again engage the pallet and gradually return it to the charg- ing position again. The Dwight-Lloyd machine is essentially a sin- tering machine, and the sulphur removal is not usually carried to the ulti- mate extent possible in roasting. The machine is used to some extent in agglomerating fine ores, granulating slag, matte, flue dust, etc., for subsequent smelting treatment. The smelting of lead ores is accomplished in rectangular blast furnaces which closely resemble those used in copper smelting. The furnace charge consists of sinter, matte added for remelting or enriching, a small amount of limestone and coke, and a considerable amount of slag. The 570 MATERIALS OF CONSTRUCTION blast-furnace products are metallic lead, Avhicli is tapped from the lead well directly to a molding machine and which may or may not be sub- sequently refined; matte which contains varying amounts of lead and copper, and which is usually granulated at the furnace, reroasted, and sub- sequently treated in converters for recovery of lead and copper; and slag which is wasted if clean, but granulated and returned to the furnace if it contains more than 2 or 3 per cent of lead. The operation of treating the matte for recovery of lead and copper does not differ essentially from the treatment of ordinary copper mattes in the basic converter. Separate lead and copper converters are often necessary. The operation of the roasting or sintering furnaces, as well as the blast furnaces and converters, result in the carrying off of a great amount of metal in the shape of dust carried in suspension by the gases. The latter are therefore collected and conducted through long cooling and settling flues and finally brought to a " bag house," where the final separation of the solid material is accomplished. A great number of cotton or woolen bags, about 18 inches in diameter and 30 feet long, into which the gases are led and from which dust is collected, are verti- cally supported in long rows. A fan forces the gases in under light pressure, and mechanical agitation of the bags is usually provided. The material recovered in the bag house sometimes assays as high as 75 per cent lead. Lead derived by the general methods of operation above described commonly contains varying amounts of, arsenic, antimony, tin, zinc, nickel, cobalt, iron, sulphur, copper, and bismuth, in addition to silver, and their removal before marketing the lead is necessary in case they exceed limited proportions. With the exception of copper, bismuth, and silver, all of these impurities are more easily oxidized than is lead, and their removal may therefore be accomplished by an oxidizing fusion. Copper, however, forms with lead an alloy which is less fusible than lead alone, and this separates out when the lead is melted as a scum which may be removed. Silver is usually removed through the agency of zinc by a special process which also removes the last traces of copper, and bismuth can only be removed with great difficulty, but is usually not present in such amounts that it need be taken care of. Refining, for the sake of removing the impurities other than those specially mentioned above, may be accomplished to a certain degree by skimming the metal in a forehearth after it has cooled to the casting temperature, but if the impurities are present in more than a small amount the refining is commonly done in reverberatory furnaces or in melting pots. Air, sometimes assisted by the introduction of steam THE NON-FERROUS METALS AND ALLOYS 571 or lead oxide, is depended upon to oxidize the impurities. Desilveriza- ation is usually accomplished by throwing zinc onto the surface of the molten lead, melting the zinc, stirring the whole thoroughly, allowing the bath to cool, and removing the scum of lead-zinc silver alloys which forms on the surface. This scum not only contains all the silver origi- nally present, but also removes the last vestige of copper. From this mixture of alloys the zinc can be separated by distillation and the lead is subsequently separated from the silver by an oxidizing melting in a reverberatory furnace, the lead oxide formed being withdrawn in a molten condition while the silver remains behind. 524. Properties and Uses of Lead. The commercial uses of lead are of such a nature that the only physical properties commonly taken account of are its hardness, its malleability, and its resistance to cor- rosion after a thin oxide film has formed on the surface. Antimony, arsenic, copper, and zinc harden lead, when present to any considerable extent, but, with the exception of antimony, they are seldom present in refined lead to an extent which renders their effect appreciable. Anti- monial lead or " hard lead " is much used as type metal, bearing metal, and for shot and bullets. Lead used for rolling into sheet metal for pipes must, however, be as free as possible from antimony and arsenic as well as metallic sulphides or other hardening elements. Lead is much used in the shape of pipe (formed by forcing molten lead through a die, against which it chills, by means of an hydraulic press), and in the form of sheet lead, used in the linings of vats, tanks, chambers, etc., for chemical manufacturing processes, and for flushings, gutters, and other roofing purposes. (Sheet lead is rolled cold after casting in the form of a comparatively thin sheet.) Tea lead is a form of very thin sheet lead, so called because used primarily in lining tea-chests. The char- acter and uses of lead pigments and litharge have been indicated above. 525. Statistics of Lead Production. The total production of lead in the United States for the last four years is reported as follows * (quantities arc expressed in short tons) : Primary Lead. Secondary Lead. Year. Antimonial Lead. Desilverized Soft Lead. Other Soft Lead. Total Soft Lead. Pig Lead. Lead in Alloys. Total. 1910 1911 1912 1913 14,069 14,078 13,552 16,665 328,954 331,032 339,646 330,593 141,318 155,947 141,248 131,867 470,272 486,979 480,894 462,460 29,492 27,389 30,266 33,104 25,930 26,895 36,902 39,730 55,422 54,284 67,168 72,834 Mineral Resources," 1913. ,572 MATERIALS OF CONSTRUCTION Tin 526. General. Commercial Forms of Tin. Tin is used quite exten- sively in the form of sheet tin, and as tinfoil, but has almost no other commercial appUcation except as a constituent of many valuable aUoys. It is one of the few important metals which is found abundantly m other parts of the world which is not also abundant in the Umted States. The principal sources of tin are the Federated Malay States m the East Indies (referred to as the Straits or Straits Settlements), England (Corn- wall and Devon), Australia, and Bolivia. 527. Occurrence in Nature. Tin Ore. The only ore of tm which is used for extraction of the metal is tin stone or cassiterite (Sn02), con- taining 78.6 per cent of the metal. It occurs both in lodes or veins, when it is called lode tin, and in secondary deposits of water-worn particles, called alluvial deposits of stream tin. By far the greatest portion of the world's tin supply is derived from the alluvial deposits in the Federated Malay States. Bolivia ranks next, being the greatest producer of lode tin. Australia produces considerable amounts of both alluvial and lode tin, and, in England, where alluvial deposits had been worked from the time of the Phoenicians, only lode tin is now available in the old workings. In the United States only relatively small amounts of tin are mined, the principal source being Alaska. 528. Extraction of Tin from its Ores. The extraction of tin from its ore involves the mechanical, or combined mechanical, thermal, and chemical, concentration of the ore, reduction by smelting with char- coal or coke in shaft furnaces or reverberatory furnaces, and refining of the crude tin derived by smelting. When the ore is in the form of alluvial deposits a large proportion of the earthy material, quartz, siUcates, and metallic oxide impurities may be removed by mechanical separation methods. Lode ores usually require an oxidizing roasting in reverberatory furnaces. Sulphur and arsenic are thus expelled and the metals with which they were combined are converted into oxides. These metallic oxides are subsequently removed by washing, a second roasting and washing treatment being sometimes required when the arsenic content of the ore is very high. When copper and bismuth are present, their oxides formed by roasting are removed by leaching with dilute sulphuric or hydrochloric acid before further washing. Two-thirds of the tin produced is smelted in small shaft furnaces in which wood charcoal is used as fuel and reducing agent. Finely divided ores require the addition of a loosening agent (slags), in order to make smelting possible, and the increase in the amount of slag produced thereby THE NON-FEREOUS METALS AND ALLOYS 573 increases the amount of tin lost in the slag. The furnace must be worked with a low depth of charge and low blast pressure in order to pre- vent excessive loss by volatilization and by the carrying off of fine dust by the blast. Dust chambers are usually provided to collect this ore dust and volatihzed tin. In order that the tin may be subjected to the oxidizing blast as little as possible, the furnace is usually worked with an open tap hole, and the separation of the slag and tin is effected in a forehearth. The reverberatory furnace is able to smelt fine ore without the addition of a slag, but coal for reduction purposes is used in addition to that burned as fuel in the separate firebox. The tin produced in the reverberatory furnace may be purer than that from the shaft furnace, and little loss is occasioned by volatiUzation or as flue dust. More tin is lost in the slag, however. The slags, skiimnings, furnace accretions, etc., are treated by proc- esses of smelting with a lime flux or smelting with scrap iron, the lime or the iron serving to reduce the oxide of tin. The crude tin obtained by the above processes must usually be refined in order to remove its impurities^ron, copper, lead, antimony, and arsenic. The most common refining treatment is simply a liqua- tion process, the pure tin being melted out on a hearth and allowed to escape while the less fusible alloys of the impurities remain behind as liquation dross. Sometimes liquation is followed by boiling or tossing. The boiling process consists in maintaining the tin in a molten condi- tion in a pot, the metal being agitated by the gases and vapor given off when a bundle of green twigs is thrust into the bath. This treatment brings every portion of the metal in contact with the air and the more easily oxidized metals are oxidized and form a scum on top which is skimmed off. When there is no longer any scum on the metal it is allowed to cool, in order that the heavy metals (principally copper and iron) may settle to the bottom, after which the metal is ladled into molds. The upper portion of the metal in the pot is purest, the lower layers are of intermediate quality, and the portion at the bottom is usually liquated and boiled again. Tossing differs from boiling only in that a workman continually takes up a ladleful of metal and pours it back into the melt- ing pot from a considerable height, thus faciUtating oxidation. 529. Properties and Uses of Tin. The properties of tin of commer- cial importance are its extreme malleability at ordinary temperatures, and its high resistance to corrosion when pure. A very large portion of the world's production is cold rolled into sheet tin and used in the manufacture of cans, as roofing material, etc. A smaller proportion is used as a coating on sheet iron or steel, or as tinfoil. When heated 574 MATEEIALS OF CONSTRUCTION above atmospheric temperatures it becomes brittle, until at 200° C, it can be powdered by hammering. Iron in considerable amounts makes tin hard and brittle and less rust resistive. Arsenic, antimony, and bismuth, in amounts exceeding about 0.05 per cent, lower its strength considerably, and copper and lead (1 to 2 per cent) increase its hard- ness and strength, but render it less malleable. 530. Statistics of Tin Production and Consumption. The produc- tion of primary tin in the United States is a negligible amount when compared with our consumption of tin, which in 1913 amounted to 40.8 per cent of the world's production. From 5000 to 6000 tons of secondary tin is annually recovered from scrap tin, however. The consumption of tin, excluding that contained in imported ore, tin foil, etc., is reported as follows: * Short tons 1910 52,528 1911 ". 53,527 1912 58,016 1913 53,315 Aluminum 531. General. Commercial Forms of Aluminum. The history of aluminum as a metal of commercial importance is very brief, the first practicable processes for the production of metallic aluminmn having been developed since 1886. Prior to that time the metal was merely a chemical curiosity, but in the years that have followed the discovery of electrolytic extraction methods it has assumed a position of great importance among non-ferrous metals and now has a multitude of impor- tant every-day applications in the arts and industries. It is recommended particularly by its lightness (sp.gr. = 2.60-2.74) combined with a high degree of strength, great ductihty, and malleabihty, non-corrosiveness, and immunity from acid attack. Aside from its very common use as rolled, pressed, drawn, or cast metal, metallic aluminum is considerably used in the form of aluminum foil, like tinfoil, and as powdered aluminum which is used extensively as a paint pigment, in explosives, and in lithographing and in printing 532. Occurrence in Nature. With the exception of oxygen and sihcon, alummum is the most abundant element on the globe; yet there are few minerals which have been successfully used for its extraction The most important are banxiie and cryolite, while rocks containing alum- mum sulphate, kaolin, and clay are of much less importance. Whatever ' Mineral Industries," 1913. Nc tl THE NON-FERROUS METALS AND ALLOYS 575 the original source of the material, it is converted into alumina (AI2O3), before the metal is extracted. Bauxite is a mixture of alumic and ferric hydrates containing widely varying amounts of alumina, ferric oxide, silica, calcium and magnesium carbonates, water, etc. Important deposits occur in France, Germany, Ireland, Italy, Australia, and in the United States in Arkansas, Georgia, and Tennessee. Most bauxites carry from 40 to 55 per cent alumina. Cryolite is a double fluoride of sodium and aluminum, represented by the formula Al2F6+6NaF, and containing, when pure, 13.07 per cent aluminum. 533. Extraction of Aluminum. The only methods of extraction of aluminum of commercial importance consist in the electrolysis of compara- tively pure alumina dissolved in a bath of molten cryolite. Alumina for the purposes of electrolysis is at present principally made from baux- ite, but may be prepared by treating silicious bauxite, kaolin, or clay with sulphuric acid, and subsequently driving off the sulphuric acid from the aluminum sulphate produced, by ignition. Bauxite is treated with a soda solution, and the alumina is extracted as sodium aluminate. The alumina may be precipitated from the solution as hydroxide by carbon dioxide, and subsequently washed, filtered, and dehydrated by heating, or it may be precipitated as hydroxide by stirring the solution with pure aluminum hydroxide. The general features of the production of aluminum by the electrol- ysis of alumina dissolved in molten cryolite are as follows: A bath is formed by the melting of cryolite in a small pot which is built of plate iron and provided with a thick carbon bottom lining. The construction of one form of electrolytic alumina reducing pot is shown in a general way by Fig. 315. Carbon anodes are suspended at frequent intervals in such a manner that they project into the bath, but do not touch the carbon cathode which forms the bottom lining of the pot. The cathode carbon is in contact with the iron shell and the latter forms the negative pole. Provision is "made for raising and lowering the anode carbons. In operation, alumina in a finely divided state is spread over the bath of molten cryolite and stirred in. Thereupon it becomes melted by the heat developed by the resistance of the bath to the passage of current, and dissolves in the bath. Electrolysis now begins immediately, the alumina being decomposed into aluminum and oxygen. The alum- inum migrates to the cathode, where it is precipitated, and the oxygen at the anode forms carbon monoxide gas. As the bath becomes impover- ished, fresh alumina and cryolite are added from time to time. The aluminum collects on the bottom and is ladled or tapped out at fre- 576 MATEEIALS OF CONSTRUCTION quent intervals. In large works a number of baths are arranged in series, and operation is continuous. The details of modern methods of electrolytic reduction of alumma have not been permitted to become generally known, nor is information available concerning the detail design of aluminum extracting _ equip- ment. According to Borchers * the temperature of the bath is 750 C, the tension is 7.5 volts, and the current density 6000 amperes per square yard of hath area, when the horizontal section of molten material is 2 feet 6 inches by 5 feet, and the depth of bath 6 inches. The early Adjustment Copper Leads. L, Poaitivej C~j live (— e-T- 1o| |g ^1 ^ IKU Copper ii, Anode Carbons Molten SiiSS." Bath Copper Leada + Fig. 315. — Bath for Electrolysis of Solutions of Alumina. baths were completely lined with carbon, but it was later found that elec- trolyte which sohdifies on the side walls forms all the protection required by the iron plates, the latter being cooled by circulating air. The metal derived as the product of tlie above operation is commer- cial aluminum, and iS usually about 98 to 99.5 per cent pure. Special grades are obtainable, containing as much as 99.9 per cent aluminum, and some second grade metal is sold which contains not over 95 or 96 per cent aluminum. The principal impurities are siUcon and iron. 534. Properties and Uses of Aluminum. The general properties of commercial aluminum are indicated by the following table. The * " Electrometallurgie," 1902, p. 154. THE NON-FERROUS METALS AND ALLOYS 577 minimum values apply to metal in an annealed condition, and the max- imum values to metal that has been cold worked to an extreme degree by forging, cold rolling, or drawing, -without subsequent annealing. Tension. Compression. iVIodulua of Elastic Limit. Lbs. per Sq.in. Ultimate Strength. Lbs. per Sq.in. Reduction of Area. Per cent. Elastic Limit. Lbs. per Sq.in. Ultimate Strength. Lbs. per Sq.in. Elasticity. Lbs. per Sq.in. Castings Rolled sheets and Bars r i [ I 6,000 to 10,000 10,000 to 25,000 15,000 to 35,000 15,000 to 25,000 20,000 to 35,000 25,000 to 55,000 10 to 20 20 to 40 40 to 70 3000 to 5000 2000 to 6000 10,000 to 15,000 9,000 to 12,000 8,000,000 to 10,000,000 12,000,000 to 15,000,000 15,000,000 to 20,000,000 Cold drawn wire. . The very marked effect of cold working in increasing the strength and improving the elastic properties of the metal is exhibited by the above maximum values. The ductility is very considerable, and it works well, at temperatures below about 200° C, in rolling or forging. Cold working hardens aluminum excessively, and in wire drawing the metal must be frequently annealed to restore ductility. The low electrical resistance of aluminum is one of its most valuable properties, since a relatively high conductivity, combined with its light- ness and strength, makes it especially well adapted for use on long-span transmission lines. The conductivity of 98.5 per cent pure aluminum is about 55 per cent of that of copper, and the specially pure grades rank somwhat higher, the conductivity of 99.9 per cent pure aluminum being about 65 per cent of that of copper. Aside from the electrical uses of metaUic aluminum immense quanti- ties are used in the manufacture of many articles of every-day domestic use and in many industries where tanks, cooking vats, etc., which must be heat-conductive, non-corrosive, and non-poisonous, are used. For these purposes the metal is either cast or rolled, and many articles are finished in a press. A further quantity of aluminum is finished in the form of seamless tubing which has many important appUcations. Aluminum may be used for casting purposes where lightness and softness are required rather than hardness and strength. Most of the aluminum used for these purposes, however, is slightly alloyed with other metals which harden it and materially increase its strength. Much 578 MATERIALS OF CONSTRUCTION of the rolled and drawn aluminum is also slightly alloyed, with an improvement in strength and hardness. 535. Statistics of Production and Consumption of Aluminum. The production and consumption of aluminum in the United States are reported as follows; all figures are given in pounds: 1910 1911 1912 1913 Production * 32,990,000 65,607,000t 49,601,500 Consumption t 47,734,000 46,125,000t 72,379,0901: * As reported by the Metallgesellschatt, Frankfurt am Main, t "Mineral Resources," 1913. t Leaf aluminum, table, kitchen, and hospital utensils, and other miscellaneous manu- factures of aluminum are not included. Nickel 636. General. Commercial Forms. of Nickel. Nickel is by far the least important of the metals classified above as metals of primary impor- tance. The world's production of nickel amounts to less than 3 per cent by weight of that of either lead or copper, and amounts to only 3.5 per cent of that of zinc. Most of the metallic nickel produced is derived from pyritic or silicate ores, or as a by-product of the refining of blister copper. Metallic nickel has very few commercial applications. A limited amount is incorporated with iron as ferro-nickel and used in the making of nickel steel; a further quantity is used in alloys of non-ferrous metals; and a small quantity is used as anode plates for electrolytic nickel plating of various metals. In the form of the double sulphate with ammonium (Ni(NH4)2(S04)2), it is also used as the electrolyte in nickel plating. 537. Occurrence in Nature. Ores of Nickel. Nickel is a constit- uent of many minerals, sometimes as the principal metallic element, but more frequently as a secondary element only extracted because smelting is justified by the presence of other metals, such as copper, in the natu- ral mixture of minerals which includes nickel compounds. The only minerals from which nickel is obtained to any considerable extent are nickel pyrites and the siUcate garnierite. Nickel pyrites or millerite (NiS) occurs in large quantities in the Sudbury district of Ontario, Canada, and is found in a few other unim- portant deposits. It contains 64.5 per cent of nickel and is almost invariably associated with iron and copper pyrites. The Sudbury ores average about 3 per cent of nickel. THE NON-FEREOUS METALS AND ALLOYS 579 Garnierite is a nickel-magnesium silicate of variable composition usually containing from 9 to 17 per cent of nickel oxide, with varying amounts of silicon, magnesia, ferric oxide, etc. The principal source of garnierite is on the island of New Caledonia. These ores usually contain about 7 or 8 per cent of nickel. 538. Extraction of Nickel from its Ores. The methods of extrac- tion of nickel from sulphur compounds on the one hand, and the extrac- tion from silicates on the other hand, differ in certain respects, and will therefore be briefly considered separately. The smelting of sulphur compounds, after removal of gangue, involves principally the separation of nickel from sulphur and from iron, and in most cases from copper also. The first operation consists in roasting the pyrite ore, the methods and equipment used differing in no respect from those used in treating other classes of sulphide ores. The object of roasting is to remove sulphur until only enough remains to combine with nickel, copper, and a portion of the iron, during the smelting opera- tion which follows. Most of the iron becomes oxidized during roasting and the sulphides of the nickel, copper, and iron are in part converted into sulphates. The second step in the process is the smelting of the roasted ore in a blast furnace with carbon and silicious matter. All of the iron which has been previously oxidized passes into the slag, while the residue of undecomposed iron sulphide forms a matte containing all of the nickel and copper sulphides. Any oxides of nickel and copper react with iron sulphide in the blast furnace, decomposing and slagging the latter as iron silicate (with the assistance of the silica), and themselves becoming sulphides. The smelting furnace used resembles the copper blast furnace and the resultant coarse matte should contain 15 to 25 per cent of nickel unless the ore is very lean. In the latter case the coarse matte is enriched by a repetition of the roasting and smelting process. Complete separa- tion of the iron cannot be effected in the smelting operation because of the loss of nickel in the slag that would result, and the next step in the opera- tion is therefore the removal of the iron sulphide by an oxidizing fusion, the iron being converted into ferrous oxide, which combines with silica and is slagged off. This operation is accomplished in a converter, and the removal of iron is complete. Any nickel or copper which is oxidized reacts with ferrous sulphide, forming ferrous oxide (which is slagged), and sulphides of nickel or copper. It is not possible, however, to sepa- rate metallic nickel as copper is separated, by continuing the blow. The slag produced usually contains notable amounts of nickel and copper and is therefore returned to the smelter. The refined matte produced usually contains from 30 to 40 per cent of nickel, and any amount of copper up 580 MATERIALS OF CONSTRUCTION to, or slightly exceeding, the amount of nickel. Both are in the sulphide form. The derivation of metaUic nickel and copper from the refined matte is accomplished by a number of methods, the details of which are not generally known. The Orford process consists in removing the copper and the little residual iron by repeatedly smelting it in small blast fur- naces with sodium sulphate (Glauber salt) and carbon (coal). The copper forms a complex sulphide of sodium, copper, and iron, and nickel sulphide remains behind. This is roasted to nickel oxide, which is finally reduced by carbon (charcoal) in crucibles. A considerable proportion of the nickel oxide is marketed as such, since it naay be used in this form for alloying purposes. The Mond process consists in roasting the matte, extracting part of the copper with sulphuric acid and reducing by smelting to a nickel- copper alloy. By subjecting this to the action of carbon monoxide at the proper temperature, the nickel is converted into the gaseous nickel carbonyl, and the latter is finally decomposed by heat into metaUic nickel and carbon monoxide. Electrolytic methods of refinement have also been used successfully in treating refined matte. The latter is roasted, smelted for copper- nickel alloy, and the latter electrolyzed. The extraction of nickel jrom the silicate differs from the methods em- ployed in treating sulphur compounds principally in that the initial smelt- .ing of the ore is done in a blast furnace with the addition of sulphur compounds, in order to produce a matte containing the nickel in the foi-m of sulphide. The ore is mixed with calcium sulphide (derived from the Leblanc soda process), or with gypsum, and with coal or coke. The mixture is crushed to powder, pressed into briquettes, and charged into the blast furnace. If gypsum is used it is reduced to calcium sulphide by the carbon in the furnace, and the calcium sulphide is decomposed by the nickel silicate of the garnierite, nickel sulphide and calcium silicate being formed. The resultant matte is rather richer in nickel sulphide than that obtained in smelting sulphur compounds, and the subsequent sepa- ration of the iron sulphide, the refinement of the second matte, and the final derivation of metallic nickel is accomplished by practically the same methods that are used in treating matte from sulphur compounds. ! 539. The Properties and Uses of Nickel. The most important property of nickel, aside from the advantages which it may confer upon steel or non-ferrous metals with which it is alloyed, is its non-corrosive- ness. On this account and because of its silvery appearance, one of the commonest commercial applications of metallic nickel is in plating THE NON-FERROUS METALS AND ALLOYS 581 iron, steel or other metals. Its mechanical properties are excellent, sometimes equaling those of medium-carbon steel, but it is too rare and expensive for general use. It is quite ductile, and fairly malleable, but is rendered brittle and incapable of being rolled by not more than 0.1 per cent of arsenic or sulphur. Most of the other impurities common to commercial nickel are not injurious to its properties, and some are beneficial in limited amounts. The principal uses of nickel have been indicated above in Art. 536. 540. Statistics of Nickel Production and Consumption. No nickel is produced in the United States except as a by-product of the refin- ing of blister copper, and by refining matte, nickel oxide, and copper- nickel alloys which are imported. The United States production of nickel by the refining of blister copper amounted to about 480,000 pounds in 1913, and the importations of nickel in the form of nickel, matte, nickel oxide, and copper-nickel alloys (from Canada for the most part, a small amount coming from New Caledonia) in the last four years are reported as follows: * NICKEL CONTENT IN POUNDS OF IMPORTATIONS OF NICKEL IN VARIOUS FORMS 1910 1911 1912 1913 32,373,251 29,829,268 46,317,078 47,446,520 THE NON-FERROUS ALLOYS General 541. The Non-ferrous Alloys in General. The study of metallic alloys is a very complex one, not only on account of the great number of combinations of two or more metals which have been found to possess valuable attributes, but also because of the inherent complexity of the problem of the interaction of metals combined in a state of fusion. True alloys of metals are never mechanical mixtures of the constituents, but are either (a) soUd solutions of the metals, (b) soUd solutions of a chem- ical compound of the metals in the metal which is in excess, or (c) mix- tures of such solutions with definite substances which have crystallized out during cooling. The composition, constitution, and structure of many alloys may therefore vary widely, and the physical characteristics hke- wise vary, and can be predicated upon a knowledge of the properties of the constituent metals to only a sUght extent. * " Mineral Resources," 1913. 582 MATERIALS OF CONSTEUCTION Many important investigations of the more important alloys, or groups of alloys, have been made in recent years, and no small portion of our present knowledge of the subject has been gained through the application of modern methods of research in physical chemistry, by the use of microscopic methods, and by study of the phenomena of fusion and solidification. These studies have an important bearing upon the study of purely physical properties which alone possesses a direct interest for the engineer or artisan. In the present discussion of the subject, however, no effort will be made to cover anything more than the composition, physical properties, and uses of the alloys which possess a distinct commercial importance. Copper-tin Allots. Bronzes 542. Ordinary Bronzes. Alloys of copper and tin are among the most useful of all the non-ferrous alloys, and have been known and used since prehistoric times. The influence of tin upon the properties of copper is that of a pronounced hardener and strengthener, so long as a limiting percentage of 20 or 25 per cent is not exceeded. The range of composi- sion of ordinary commercial bronzes is not wide, all of the important ones containing 80 per cent of copper or more. An immense amount of information concerning the properties of bronzes has been collected. Fig. 316 presents a summary of the mechanical properties of a series of bronzes tested by the Committee on Alloys of the United States Board to Test Iron, Steel, and Other Metals.* The work was done under the direct supervision of Professor Robert H. Thurston. In spite of the exercise of the greatest care in conducting the investigation, the proper- ties of the individual specimens of the 29 alloys tested varied widely, and the comparatively smooth curves of Fig. 316 simply average up the results in a general way. All of the specimens used in this series of tests were of cast bronze, the transverse test specimens being the original cast bars 1 inch square -in section and centrally loaded on a span of 22 inches. The tensile specimens were turned from the ends of the trans- verse test specimens, the reduced portion being 6 inches long and 0.798 inch in diameter (0.5 sq.in. in area), and deformations were measured on a 5-inch gauged length. The compressive specimens were cylinders 2 inches long and 0.625 inch in diameter turned from the ends of the tensile specimens. The bronzes exhibiting the greatest tensile strength and bending strength, and the highest yield point, are those containing more than 80 * Report of U. S. Board to Test Iron, Steel, and other Metals, 1881. THE NON-FERROUS METALS AND ALLOYS 583 per cent of copper. The compressive strength appears to increase with decrease in copper content until an alloy containing about 75 per cent 20,000,000 19,000,000 f >^ 1 18,000,000 ? SI cot 17,000,000 .■.c> d^ flex irej o. ^ ^ \ 16,000,000 . O A / ^^ / (X) 16,000,000 1 / ' / / i4 11,000,000 g / 000 / 1 13,000,000 ^ / 000 r \ 12,000,000 » / \ \ o 11,000,000 1 10,000,000 ^ 1 V ,000 \ \ \, / \ % \ \ 9,000,000 // l». \ y / t9b \ 10 / I \ \\ 35 £ 000 4 ^ / \ \ ( \ 30 a yU ^-«>^ %■*' ^ 1 \ M 25 /^- V'' \ \ \ 1; r d 20 o f ^^ \ A \ . 15 9 \ \ r 1 .ttv»; ce "^ ^^ 10 a finny f — \ \ \ < Tisi ^as. toi c St ■eng h —7 ^ 5 ^ V -■^ :::i -— ■ieia Poin t ^ ^ -~- 1 DO i 1 6 i Per cent 6 S age Com 4 poai Hon 'd 1 ■I 8 U 1 9 K ^ Cu. KXSn, Fig. 316.— Properties of the Copper-Tin Alloys. (U. S. Test. Board.) of copper and 25 per cent of tin is passed. Beyond these limits the strength decreases rapidly in all cases with further additions of tin. The 584 MATERIALS OF CONSTRUCTION stiffness appears to increase until a 50 per cent tin alloy is reached, but the ductility reaches its maximum with only about 4 per cent of tin, and is entirely lacking with more than 25 per cent of the latter, until alloys are reached which begin to approach pure tin in properties, when the great ductility of the latter comes to be characteristic of the high-tin alloys. All of the alloys containing between 25 and 75 per cent of tin are extremely brittle and weak, and those containing more than 75 per cent of tin are weak and soft. The strength and ductility of bronzes are considerably affected by heat treatment. M. Guillet * found that, with bronzes containing over 92 per cent of copper, quenching between 400° and 600° C. slightly increases the strength and ductility. With less than 92 per cent of copper, both strength and ductility increase decidedly as soon as the quenching temperature exceeds 500° C. The maximum strength of all alloys was found to be reached by quenching at about 600° C, the beneficial effect becoming more marked as the copper content is reduced. The most commonly used types of ordinary bronzes are the following (it will be noted that a number of these alloys classed as ordinary bronzes sometimes contain small amounts of zinc, lead, etc.): Machinery bronzes are used principally as hard bearing metal, as metal for cut gears, and for valves, bushings, stuffing boxes, piston rings, steam whistles, plumbing fixtures, etc. The average machinery bronze contains from 81 to 87 per cent of copper, and 19 to 13 per cent of tin. Most gear bronzes contain phosphorus or manganese or both, and will therefore be considered under the head of " special bronzes." Fre- quently bearing bronzes contain from 2 to 4 per cent of zinc, as do also many of the bronzes used for the various machine parts above Usted. Locomotive bronzes and piston rings may contain as much as 8 or 9 per cent of zinc. (The properties of the ternary bronzes are considered in the following article.) Gun metal usually contains from 88 to 92 per cent of copper and 12 to 8 per cent of tin. It is one of the strongest of all the bronzes. It was at one time commonly used for casting guns, but has now been entirely replaced by steel in ordnance concentration and is used only as a material for strong castings. It is not infrequently alloyed with small percen- ages of zinc or even lead. Bell metal usually contains from 75 to 80 per cent of copper and 25 to 20 per cent of tin. It is largely used for bell founding, because of its resonance, but is hard and brittle. (A number of brasses and ternary alloys are also used in bell founding.) Small amounts of zinc or lead * " Etude Industrielle des Alliages Metalliques," 1906. THE NON-FEKROUS METALS AND ALLOYS 585 are occasionally used in bell metal, and silver is sometimes added with the idea that the tone is improved. Speculum metal usually contains 65 to 70 per cent of copper and 35 to 30 per cent of tin. Before the present state of the art of manufactur- ing silvered glass reflectors was attained it was always used for the specula of reflecting telescopes. Its value is due to the fact that it is extremely hard, and therefore takes a smooth poUsh, and possesses a silvery white color. Statuary bronzes usually contain from 90 to 78 per cent of copper, 2 to 10 per cent of tin, and any amount of zinc up to 10 or 15 per cent. They often contain more zinc than tin, and therefore ought to be classed as special bronzes or brasses. Coin or medal bronzes contain from 90 to 97 per cent of copper and 10 to 3 per cent of tin. Occasionally very small amounts of zinc or lead are added. 543. Special Bronzes. Copper-tin-zinc Bronzes. The copper-tin- zinc alloys are among the most valuable and commonly used of all the bronzes. The range of composition of the commercial bronzes of this class is from 50 to 95 per cent of copper, 1 to 15 per cent of tin, and 5 to 50 per cent of zinc. The tensile strength of copper-tin-zinc alloys is exhibited by the diagram of Fig. 317, which is based upon the report of the U. S. Board to Test Iron, Steel, and Other. Alloys above referred to. The triaxial diagram used in this instance is based upon the geometrical principle that the sum of the normals from any point in an equilateral triangle upon the sides is equal to the altitude of the triangle. The altitude is made to represent 100 per cent of any of the constituents. Thus the altitude measured to one vertex from the opposite side represents 100 per cent of copper, that to the second vertex 100 per cent of tin, and that to the third vertex 100 per cent of zinc. Each side of the tri- angle represents zero content of the constituent represented by the oppo- site vertex, and the distance of a point within the triangle to any side, expressed as a proportion of the altitude, expresses the percentage of the constituent represented by the vertex opposite the side chosen. Any point in the diagram therefore represents a certain definite alloy or mixture of three constituents. Thus point A on the diagram represents an alloy containing 13 per cent of copper, 57 per cent of tin, and 30 per cent of zinc. Each alloy tested is therefore represented by some one point on the diagram, .and when all such points have been plotted and marked with the figures representing the particular property under in- vestigation (tensile strength in this case), lines or contours may be drawn connecting points representing alloys possessing the property in question 686 MATERIALS OF CONSTRUCTION to an equal degree, and the result is such a diagram as is here presented. . a- • a. It is evident at a glance that all of the alloys possessing sufficient strength to be considered of any value are those containing more than 50 per cent of copper, less than 20 per cent of tin, and less than 50 per cent of zinc. The strongest alloys contain nearly 60 per cent of copper, 1 to 2 per cent of tin, and nearly 40 per cent of zinc. The diagram does :*»" Fig. 317. — Tensile Strength 6f Copper-Tin-Zinc Alloys. (Compiled by J. B. Johnson from Report of U. S. Test. Board.) not show the ductility of the various alloys, but the data of the original tests reveal the fact that the most ductile alloys are in the upper portion of the field comprised within the 30,000-pound-per-square-inch contour line. All of the alloys containing more than 10 to 20 per cent of tin (the higher, the lower the zinc content), are extremely brittle and fragile, and are altogether worthless. The alloys having a tensile strength of from 30,000 to 40,000 pounds per square inch are the most generally useful ones, since they are comparatively tough and are very ductile, showing from 10 to 30 per cent elongation at rupture in tension. THE NON-FERROUS METALS AND ALLOYS 587 The character and uses of the most important copper-tin-zinc bronzes have been indicated above under the head of machinery bronzes. Copper-tin-lead Bronzes. Lead is not infrequently added to bronzes used for bearings, for the purpose of increasing the plasticity just enough to allow the metal to adapt itself to the running surface so that the bearing is uniform over the whole surface. The addition of from 10 to 40 per cent of lead to a bronze containing 5 to 10 per cent of tin cuts down the wear on the bearing very materially without increasing the friction more than very slightly. The degree of softness desirable in a bearing bronze is dependent upon the conditions of service, and the lead content and tin content are varied accordingly. In general, the lower the content of tin, the higher the content of lead desirable. An alloy of this class commonly used for railway service and other pur- poses is known as " plastic bronze." It contains 65 per cent of copper, 30 per cent of lead, and 5 per cent of tin. Phosphor Bronze. The addition of very small percentages of phos- phorus to any bronze has a remarkable effect upoii its properties. The tensile strength is considerably increased, and the elastic limit and endurance under repetition of stress are very greatly increased. Phos- phorus added in small amounts acts principally as a deoxidizer, and the marked improvement in the properties of the bronze is principally due to the elimination of copper oxide. In slightly larger quantities the phosphorus combines with copper to form a compound which greatly hardens the bronze. Phosphor bronzes, intended for use as engine parts, valve metal, etc., usually show 6nly traces of phosphorus and contain about 7 per cent of tin. Phosphor bronze for gears is harder, owing to the pres- ence of 8 to 12 per cent of tin, and phosphorus varying from mere traces up to a maximum of about 0.2 per cent. Bearing bronzes most commonly contain 8 to 10 per cent of tin and from 0.2 to about 0.9 per cent of phosphorus. Certain bearing bronzes, particularly those for railway service, contain about 10 per cent of lead in addition to 10 p^r cent of tin and about 0.3 per cent of phosphorus. These bronzes ..excel any other ckss of bearing metal in resistance to wear under severe conditions, and in addition possess a very low coefficient of friction. Phosphor bronzes of proper composition may be rolled or drawn into wire, and when so fabricated exhibit about the same tensile prop- erties as does medium structural steel. Working cold has the same effect that it has in the case of steel, the strength, and especially the elastic, limit, being raised to a marked extent. All of the phosphor bronzes are remarkably resistant to corrosion, and are much used on subaqueous cojistruction and in other situations on this account. 588 MATERIALS OF CONSTRUCTION Manganese Bronze. The so-called manganese bronze, which is now commonly used for a great variety of purposes, is really a special brass, since it contains very large amounts of zinc and little or no tin. This class of alloys will therefore be considered under the head of special brasses. Silicon Bronze. Silicon added to ordinary bronzes has about the same effect as similar additions of phosphorus. Its principal action added in small quantities, is that of a deoxidizer, and it has about the same effect upon the properties as has phosphorus, the strength and especially the elastic limit being raised, although only a trace of silicon may appear in the finished bronze. In one important respect silicon differs in its effect from phosphorus. The latter, even in very small amounts, is very detrimental to electrical conductivity, but silicon is not. In consequence, silicon bronze is a much better conductor than phosphor bronze, and is considerably used for telephone wires, etc. Aside from its electrical conductivity, silicon bronze possesses most of the valuable properties of phosphor bronze, and is used for similar purposes where a strong, hard, and non-corrosive alloy is desired in the form of castings, rolled sections or sheets, or drawn wire. Aluminum Bronze. The alloy commonly known as aluminum bronze contains no tin, but is simply an alloy of copper and aluminum, and so is not really a bronze at all. Its character and properties will be con- sidered under the head of binary alloys of copper other than bronzes and brasses. Vanadium Bronze. Vanadium has occasionally been added to bronzes in very small amounts with remarkably beneficial effects upon proper- ties. These bronzes have not become commonly known or used, how- ever. Nickel bronzes are for the most part quaternary alloys containing about 30 to 40 per cent of lead, 5 per cent of tin, and 1 per cent of nickel, the balance being copper. The nickel serves particularly to prevent the separation out of lead during solidification, and produces a homo- geneous alloy containing more lead than can ordinarily be used without danger of segregation. This alloy is an important bearing bronze and is considerably used in machine construction and for railway service. CoppER-ziNC Alloys. Brasses 644. Ordinary Brasses. Alloys of copper and zinc, with or without addition of other elements (i.e., ordinary brasses or special brasses), share with the bronzes the most important position among non-ferrous THE NON-FEREOUS METALS AND ALLOYS 589 alloys. The influence of zinc upon the properties of copper is in the direction of increasing both strength and ductihty, so long as certain limiting percentages are not exceeded, but these limiting percentages are much higher than in the case of tin, zinc being a less potent element than tin in similar amounts. Fig. 318 presents a summary of the mechanical properties^ of the copper-zinc alloys. The curves are based upon the report of the Com- mittee on Alloys of the United States Toard to Test Iron, Steel, and Other Metals,* the fourth report of Professor W. C. Roberts-Austen to the Alloys Research Committee of the Institution of Mechanical Engineers, t and an independent study of the tensile strength of the copper-zinc alloys by Dr. J. M. Lohr.J The tests of the U. S. Test Board were made by Professor Robert H. Thurston, the specimens used being made from castings in the manner above described in connection with the tests of copper-zinc alloys quoted from the same source. The tests of the Alloys Research Committee were made upon specimens which had been mechanically worked, and those of Dr. Lohr were made upon cast specimens which were quenched in water immediately after sohdification, in order to prevent the loss in strength incidental to the acquirement of a coarse crystalline structure during slow cooling. The addition of zinc to copper gradually increases the strength in tension until about 30 or 35 per cent is present. Further additions rapidly increase the strength until a maximum beneficial effect is obtained with about 45 per cent of zinc. A rapid faUing off in strength occurs with additions of zinc beyond 45 per cent, and the brasses containing more than about 50 per cent of zinc are brittle and worthless. (In the series of brasses tested by Thurston the gain in strength is fairly uniform with increase in zinc content up to a maximum at about 40- per cent of zinc, the strength thereafter falling off rapidly.) The beneficial effect of preventing coarse crystallization by quenching the castings is shown by a comparison of the results of Lohr with those of Thurston, the strongest alloy of the former showing a tensile strength of about 72,000 pounds per square inch, with 45 per cent of zinc, as com- pared with one having a tensile strength of about 50,000 pounds per square inch with 40 per cent of zinc. The beneficial effect of mechanical work- ing is also shown by a comparison of the curve representing the results of Roberts-Austen with that representing the results of Lohr. The strongest cast brasses in compression appear to be those con- taining between 50 and 70 per cent of zinc; the maximum strength in * Report of U. S. Board to Test Iron, Steel, and Other Metals, 1881. t Proceedings of the Institution of Mechanical Engineers, 1897. i Journal of Physical Chemistry, Vol. 17, No. 1, Jan., 1913. 690 MATERIALS OF CONSTRUCTION flexure corresponds to the maximum tensile strength; and the max- imum ductility is shown by alloys containing from about 25 to about 35 per cent of zinc. The strongest brasses show very low ductility. 100 60 60 40 30" 40 50 60 70 Percentage Composition 20,000,000. 19,000,000 10 0)t Ou. 80 100;(; Zn. Fig. 318.— Properties of the Copper-Zinc Alloys. A comparison of the curves of Fig. 318 with those of Fig. 316 shows that the beet of the brasses excel the best of the bronzes in tensile THE NGN-FEEROUS METALS AND ALLOYS ' 591 strength and ductility, but the latter seem to excel in compressive strength. The various common commercial forms of brasses may be classi- fied, as follows according to composition and uses: Tombac and pinchbeck contain 85 to 90 per cent of copper and 15 to 10 per cent of zinc. They take a good polish and are principally used for ornaments, imitation of gold, etc. Hard brazing metal contains 80 to 90 per cent of copper and 20 to 10 per cent of zinc. That containing 20 per cent of zinc is known as " quarter metal " and is considerably used for the purpose indicated by the name. Red brass usually contains about 80 per cent of copper and 20 per cent of zinc. The name is used loosely, however, to cover any of the high-copper brasses which possess a reddish color, especially after pickling in acid. This class of brass is much used for ornamental work and soft castings. Standard brass contains about 66 to 70 per cent of copper and 34 to- 30 per cent of zinc. It is the most generally useful and most commonly used of all brasses. It is very ductile, works well hot or cold, and can readily be rolled into sheets and drawn into tubes or wire. It is less corrosive than any of the brasses of lower copper content, and is espe- cially adapted for use in locomotive and steamship boiler and condenser tubes. Practically all ordinary sheet brass and drawn brass is of this composition, and the best of the brass castings contain about 66 or 67 per cent of copper. Muntz metal is a brass containing 60 per cent of copper and 40 per cent of zinc. It can only be rolled hot, but was formerly much used as a sheathing for wooden vessels. Sea water attacks it and forms zinc salts, which prevent the fouling of the bottoms of ships by. livuig organ- isms such as barnacles, etc. Yellow brass contains from 48 to 56 per cent of copper and 52 to 44 per cent of zinc. A mixture of approximately equal proportions of the two constituents is commonly used in making brass castings which do not require great strength or toughness. At a red heat' yellow brass becomes so brittle that it may be easily pulverized, and in this form is much used for yellow solder for brazing purposes. White brass solder contains 34 to 44 per cent of copper and 66 to 56 per cent of zinc. It is extremely weak and brittle and is used only in a powdered condition for brazirig purposes. White brass contains less than 10 per cent of copper and more than 90 per cent of zinc. The metal possesses most of the characteristics of zinc, but is somewhat hardened and strengthened by the small amount 692 MATEEIALS OF CONSTRUCTION of copper used, the coarse crystalline structure of cast zinc being largely- destroyed. The material is principally used in making ornaments which are plated with bronze and sold under the name " French bronze." 545. Special Brasses. Copper-zinc-lead brasses. The addition of small percentages of lead softens brass and renders it more easily cut by machine tools. The presence of the lead lowers thf strength and decreas- es the ductility con- siderably. More than 5 per cent of lead cannot be profitably used because of the danger of segregation, and the usual addi- tion is not in excess of 3 per cent. Copper-zinc-alumi- num brasses. Alumi- ntmi is added to brass in amounts up to about 5 per cent with beneficial effects upon the tensile properties. The tensile strength and elastic limit are considerably raised, and the hardness in- creased and ductility decreased. The effect of aluminum upon the tensile strength and ductility of rolled and cast brass is exhibited by Fig. 319 which is based upon tests reported by M. Guillet.* The addition of aluminum is also beneficial in that it facilitates the making of good brass castings. Aluminum brass is principally used in making castings for machinery, marine work, etc., for forgings, and for rolled * " Etude Industrielle des AUiages Metalliques," 1906. 1.5 2.0 2.5 3.0 3.5 4.0 1.5 5.0 6.5 Percentage of Aluminum Fig. 319. — Effect of Aluminum on Tensile Properties of Brass. (Guillet.) THE NON-FERROUS METALS AND ALLOYS 593 bars, plates and shapes designed for any purpose requiring a strong brass, or a strong and non-corrosive metal. Copper-zinc-^manganese alloys. Manganese bronze. The most val- uable copper-zinc alloy in use at the present time is the so-called " Man- ganese Bronze." The presence of small amounts of manganese in the finished alloy is beneficial to strength in a measure, but the commercial alloy known by this name is simply a brass with which a small amount of manganese, in the form of ferro-manganese, or manganescrcopper produced in the electric furnace, has been incorporated while molten for the sake of the important deoxidizing effect which the manganese exerts. The resultant alloy usually contains no manganese, or only a trace, because it has been oxidized out and fluxed off. The specifications of the American Society for Testing Materials * for manganese bronze ingot metal for sand castings call for an alloy having the following composition and tensile properties: Chemical Composition. Per cent. Minimum Tensile Strength. Lbs. per Sq.in. Minimum Elongation in 2 Inches. Per cent. Copper, 53-62 Zinc, 36-45 Aluminum, 0.05-0.5 Lead, not over 0.15 I 70,000 20 The average composition of the commercial alloy, according to Mr. C. R. Spare, t is about 56 to 57 per cent of copper, 38 to 40 per cent of zinc, about 1 per cent of tin and 1§ per cent of iron. The iron is pres- ent inadvertently, owing to the use of ferro-manganese as the deoxidizer, and a little aluminum is introduced to facilitate the making of cast man- ganese bronze, but is considered undesirable in bronze intended for forg- ing or rolling. The range of composition of various brands of manganese bronze is quite wide and the properties of the metal vary accordingly. A soft grade is made especially for use in situations where it must withstand shock and vibration, as in naval ordnance construction. This metal has a tensile strength of about 60,000 pounds per square inch with an elongation of 40 to 50 per cent in 2 inches. A very hard grade is made which has a tensile strength of 90,000 to 100,000 pounds per square inch with elongation of 15 to 25 per cent. The average grade of manganese bronze shows about the following properties: * Yearbook, 1914. t Proc. Am. Soc. Test. Matrls., Vol. 8, p. 395. 594 MATERIALS OF CONSTRUCTION Tension. Compression. Elastic Limit. Lbs. per Sq.in. Tensile Strength. Lbs. per Sq.in. Elongation, 2 Ins. Per cent. Elastic Limit. Lbs. per Sq.in. Compressive Strength. Lbs. per Sq.in. Cast manganese Bronze Rolled or forged manganese bronze i 30,000 to 40,000 40,000 to 50,000 70,000 to 80,000 80,000 to 110,000 20 to 35 15 to 30 35,000 to 40,000 50,000 to 60,000 90,000 to 100,000 130,000 to 150,000 Probably no other metal or alloy possessing equal strength and toughness can be cast in intricate forms so successfully as can manganese bronze. In addition, it is particularly resistant to corrosion by sea water and alkali waters, and is proof against attack by dilute acids. Manganese bronze is very commonly used for steamship propellers, and is much used for other ship fittings, for piston rods, shafts and axles, and for all manner of castings, forgings, etc., used in general machine construction, and in locomotive and automobile construction. It has many appli- cations in subaqueous work of various types, and has also been found to be one of the best materials obtainable for the blades of high-speed steam turbines. For the latter application it not only possesses the high strength required, but also resists the corrosive action of high pressure steam very well. Copper-zinc-iron alloys. Two brass alloys containing iron have been commonly used: Sterro metal contains about 60 per cent of copper, 38 to 38.5 per cent of zinc, and 1.5 to 2 per cent of iron. Delta metal varies in composition, but usually contains about 55 per cent of copper, 41 per cent of zinc, 3 per cent of iron, and 1 per cent of manganese, phosphorus, and other elements. These metals, particularly the delta metal, possess a considerably higher strength and better working qualities than the brass would possess without the iron addition. They are also more resist- ant to corrosion. They possess, to a lesser extent, the characteristic properties of manganese bronze and aluminum brass, and have been adopted for the same class of uses, principally on marine con- struction. Other special brasses. Vanadium is a metal which is only very rarely used in brasses. Phosphorus is also occasionally used in small amounts for the purpose of deoxidizing the copper, but neither vanadium brass nor phosphorus brass are alloys possessing any commercial impor- tance. A number of quite important alloys of copper, zinc, and nickel THE NON-FERROUS METALS AND ALLOYS 595 are made and sold — usually under the name of " German silver." These latter alloys will be considered under the head of Nickel Alloys. Binary Alloys of Copper other than Bronzes and Brasses 546. Copper-Aluminvim Alloys. The alloys of copper and aluminum are the most valuable of the copper alloys other than the bronzes and brasses. The principal commercial alloy of copper and aluminum is the so-called aluminum bronze, which usually contains from 90 to 95 per cent of copper and 5 to 10 per cent of aluminum. Since no tin is present, the alloy is not really a bronze, but it is, nevertheless, commonly called " aluminum bronze," perhaps on account of its resemblance to some of the bronzes in its properties. The characteristics of the copper-aluminum alloys have been made the subject of an extremely painstaking investigation by the Alloys Research Committee of the Institution of Mechanical Engineers. The results of this investigation constitute the Eighth Report of this Com- mittee.* Fig. 320 constitutes a summary of the results obtained in a portion of the investigation of tensile properties of a series of copper- aluminum alloys containing up to 13 per cent of aluminum. The addition of aluminum to copper is shown to be responsible for a gradual raising of the tensile strength and yield-point, and a rapid increase in ductihty, until about 7.35 per cent of aluminum is present. Further additions cause a m.ore rapid raising of the strength and yield- point, accompanied by a very rapid decrease in ductility, until an alloy of maximum strength (€0,000 to 74,000 pounds per square inch for cast- ings) is reached with about 10 per cent of aluminum present. With additions of aluminum beyond 10 per cent, the strength rapidly falls off and the ductihty becomes practically nil, the yield-point coinciding with the ultimate strength with about 13 per cent of aluminum present. Comparing the properties of the castings made in sand with those cast in chills, it appears that chilhng scarcely affects the properties of alloys containing less than 6 per cent of aluminum, but raises the strength and yield-point of the higher-aluminum alloys in proportion to the aluminum content. The strength of the chilled 10 per cent aluminum alloy is about 20 per cent higher than that of the corresponding sand-cast alloy. The IJ-inch rolled bars were produced by hot rolling of an ingot cast 3 inches in diameter, but turned down to about 2^ inches, so that the reduction in rolling was from this diameter down to 1} inches. The rolled bars show practically the same ductihty as the cast bars, but excel the latter sUghtly in tensile strength and yield-point. The beneficial * Proceedings Institution of Mechanical Engineers, 1907, p. 57. 596 MATEEIALS OF CONSTRUCTION ^, 6 6 7 8 9 10 11 12 eroentage of Aluminum Fig. 320.— Tensile Properties of Aluminum Bronzes. (Alloys Research Committee.) THE NON-FERROUS METALS AND ALLOYS 597 effect of rolling was found to be only slightly increased when the reduc- tion in rolling was continued down to xf inch diameter. The cold-drawn bars were produced by rolling hot from IJ inches down to xf inch, cold drawing to | inch, annealing, and then cold- drawing down to xi inch. The effect of cold-drawing to this moderate extent appears to be a moderate increase in tensile strength, a great raising of the yield-point, and a great loss of ductility. The effect of slow cooling from 800° C, and the effect of quenching from the same temperature were also investigated and led to the con- clusion that slow cooling from this temperature has no effect upon alloys containing less than 7.35 per cent of aluminum, and injured the higher- aluminum alloys, owing to coarse crystallization and consequent brittle- ness. Quenching was found to have no effect upon the alloys below 7.35 per cent aluminum, but greatly raised the tensile strength and yield- point, and lowered the ductiUty of the higher-aluminum alloys. By way of summing up the discussion it may be stated that the aluminum bronzes containing less than 7.35 per cent of aluminum show moderate strength, a yield-point that is relatively low, and a ductility that is remarkably high. The high-aluminum alloys (8 to 11 per cent aluminum), show a remarkably high strength, a relatively low yield- point, and low ductility. Most of the aluminum bronzes work well in the foundry, and may easily be rolled below a bright red heat, forged at a low red heat, or drawn into bars, shapes, sheets, tubes, wires, etc. The metal is highly resistant to corrosion and has therefore been used on marine works and as ship's fittings, including propeller blades. In ordinary machine con- struction and in automobile construction it has found many special uses, and, because of the pecuhar smooth unctuous surface which it ac- quires, it has been found to be an excellent anti-friction metal. The high-copper aluminum bronze is much used as imitation gold. 547. Binary Alloys of Copper with Manganese, Phosphorus, Silicon, etc. Alloys of copper are occasionally made with manganese, phos- phorus, sihcon, vanadium, chromium, tungsten, antimony, bismuth, lead, arsenic, etc. None of these possess special properties which render them particularly useful commercially, and no space therefore will be herein devoted to a consideration of their characteristics. ALLOYS OF ZmC, LEAD, TIN, ALLUMINUM, AND NICKEL 548. Binary Alloys of Zinc. (Non-cuprous.) Zinc forms no binary alloys of commercial importance with the exception of the brasses. The character of the various other binary zinc alloys may be briefly indicated as follows: 598 MATERIALS OF CONSTRUCTION Lead cannot readily be alloyed with zinc in a binary alloy. Tin may be alloyed with zinc with care, forming white metals of almost any composition. These alloys have little practical importance, however, and are rarely made on account of the high cost of tin. Alloys of anti- mony- and zinc are difficult to make, and have little value because of their brittleness and the readiness with which they are oxidized. Bismuth forms alloys with zinc which are extremely brittle and worthless. 549. Binary Alloys of Lead. (Non-Cuprous.) Lead-tin alloys. Lead and tin alloy in all proportions, the most important series of alloys being those used as plumber's solder, and which contain from 33 to 50 per cent of tin. The best solder for wiping joints in lead pipe is one containing about 2 parts of tin to 1 part of lead, but, owing to the expensiveness of tin, the content of this constituent is often reduced. The especial value of the lead-tin solders lies in the fact that the metal passes through a pasty stage in solidifying, the lead solidifying gradually before the entire mass freezes. Another alloy of lead and tin formerly much used for mak- ing various domestic utensils, slush castings, etc., is that which is called pewter. As tin is added to lead the latter is hardened and strengthened gradually until an alloy of maximum strength is reached with about 70 to 75 per cent of tin present. Ordinary pewter contains from 50 to 80 per cent of tin, the quality being best when the. tin content exceeds 70 per cent. A little antimony or copper may be used to harden pewter. Lead-antimony alloys. Antimony is often present in lead accident- ally, as has been above noted, the resultant impure lead being known as " antimonial lead " or " hard lead " which may be used as type metal with or without the addition of tin, bismuth or copper. - Antimony hardens lead very rapidly and in large amounts makes it very brittle, but at the same time forms an alloy which casts well and takes a very sharp impression of the mold. The average type metal contains about 17 per cent of antimony and may also contain from 10 to 20 per cent of tin or small amounts of bismuth, copper, etc. Lead-antimony alloys used for shot, bullets, etc., contain about 12 to 16 per cent of antimony. Lead-bismuth alloys. Lead-bismuth alloys are easily made, so long as the proportion of bismuth is less than that of lead. The resultant alloys exceed pure lead in ductility and malleabihty, but have no impor- tant commercial applications. Lead-cadmium alloys are seldom produced, except accidentally, and possess no commercial importance. Lead-arsenic alloys are made intentionally for only one purpose— the making of shot. The addition of from 0.5 to 1 per cent of arsenic renders the lead more fusible, lengthens the time of solidification, thus facihtatmg the assumption of a spherical form by the lead in its' drop THE NON-FERROUS METALS AND ALLOTS 599 in the shot tower, and makes the metal, when solidified, somewhat harder than pure lead. 550. Binary Alloys of Tin. (Non-cuprous.) The principal classes of binary tin alloys have been considered above. Cadmium forms alloys of little practical value, and antimony forms alloys which are similarly without commercial importance with the exception of one alloy, known as Brittannia metal, which is used to a slight extent for ornamental castings and stamped or wrought forms which are usually polished or plated. The usual composition of this alloy is 80 to 90 per cent of tin and 5 to 15 per cent of antimony, the balance of the composition being made up of small amounts of copper, zinc, etc., added for the sake of their hardening effect. The binary alloys of tin vath bismuth, nickel, etc., possess no great industrial importance. In general, none of the alloys of tin, except- ing the brasses and the tin-lead alloys, possess valua;ble properties, and their high cost therefore bars their use where cheaper alloys will serve as well. 551. Aliuninum Alloys. A great number of metals, and some non- metallic elements, are added to aluminum for the purpose of strength- ening or hardening it, without materially increasing its weight. These alloys are generally classed as light aluminum alloys. Aluminum-zinc alloys are the most valuable and also the cheapest of the hght alloys. Proportions up to about 33 per cent of zinc are used, the most ductile and malleable alloys containing less than 15 per cent of zinc, while those containing higher zinc contents are still useful in cast- ings which permit of a certain degree of brittleness. These alloys are readily made, are harder and more fusible than aluminum, and are still very light. The properties of the aluminum-zinc alloys have been exhaustively studied by the Alloys Research Committee * and a summary of a por- tion of their report is presented by Fig. 321. It is shown by the diagram that all of the alloys containing more than about 15 per cent of zinc are very non-ductile and brittle, and that even when ductiUty is unimportant no gain in strength, or at least in yield-point, is obtained by increasing the zinc content beyond about 30 per cent. The beneficial effect of hot and cold working of these alloys is shown by the diagram of Fig. 322. Aluminum-coppa- alloys. Copper is one of the most commonly used hardening agents in aluminum alloys, the amount used in binary alloys rarely exceeding about 6 or possibly 8 per cent. The copper raises the strength and the yield-point considerably, but causes a rapid loss of ductility. * Tenth Report, Proc. Inst. Mech. Engrs., 1912, p. 319. 600 MATERIALS OF CONSTRUCTION The. properties of this series of alloys have also been studied by the Alloys Research Committee, and a portion of their report is summarized 60 30 10 30 ._ 40 50 60 rO Percentage Composition 10 90 O^Al. 100 ^Zn, Fig. 321.— Tensile Properties of Aluminum-Zinc Sand Castings. (Alloys Research Committee.) SOOOO 45000 o-lOOOO a 530000 Co a S £25000 S 20000 15000 I'OOOO — ■^^ ""■ y ^ — ^ ^ / y / ?> r^ / < & .? / / . V / / 4 / / ^^ / ^ y^ / / ^ ^ / .<>'■ ^ / / .v<^^ ^ ^ / c ^^ / /I i> y" / ^ ^ ^ /■ -<: s^ >- /- ^ ^ / 10 15 Per Cent of Ziao 20 S3 30 Fig. 322. — Effect of Mechanical Working on Strength of Aluminum-Zinc Alloys. (Alloys Research Committee.) by the diagrams of Fig. 323. A comparison of this diagram with the diagrams of Figs. 321 and 322 shows that the aluminum-copper alloys THE NON-FERROUS METALS AND ALLOYS 601 are considerably inferior to the aluminum-zinc alloys in strength, but greatly excel them in ductility. The yield-point also is not relatively nor actually as high in the copper alloys as in the zinc alloys. Aluminum-magriesium alloys. The alloy of aluminum and mag- nesium possessing the most valuable properties is one containing less than 2 per cent (usually about 1.6 per cent) of magnesium. This alloy is moo 4 5 6 Per Cent of Copper Fig. 323. — Tensile Properties of Aluminum-Copper Alloys. (Alloys Research Committee.) slightly lighter than pure aluminum, but shows a tensile strength of 25,000 to 40,000 pounds per sguare inch when cast and rapidly cooled or when rolled without annealing. The commercial alloy of aluminum and magnesium usually contains small percentages of copper, nickel, tin or lead. Mr. J. W. Richards * quotes tests of aluminum-magnesium alloys showing the following properties: * Proc. Am. Soc. Test Matrls., Vol. 3, p. 245. 602 MATERIALS OF CONSTRUCTION 2 Per cent Mg. 4 Per cent Mg. 6 Per cent Mg. Tens. Str. Lbs. per Sq.in. Per cent Elong. Tens. Str. Lbs. per Sq.in. Per cent Elong. Tens. Str. Lbs. per Sq.in. Per cent Elong. Cast in sand , 17,900 28,600 40,000 25,600 41,300 3.0 2.0 1.0 18.0 2.7 28,600 28,200 44,900 2.6 ' ' 8.0 2.1 57,600 28,100 44,100 Castings water chilled. . . Annealed sheet. 1.0 17.0 1.0 8 Per cent Mg. 10 Per cent Mg. Tens. Str. Lbs. per Sq.in. Per cent. Elong. Tens. Str. Lbs. per Sq.in. Per cent Elong. Cast in sand 21,400 33,600 61,100 2.4 Cast in chills. 3.4 Castings water chilled. . . Annealed sheet 54,900 1.6 4.2 Hard sheet Aluminum-copper-zinc alloys have been produced commercially to a slight extent as light casting alloys. Various alloys of this type contain from 9 to 27 per cent of zinc and 3 to 5 per cent of copper. The Alloys Research Committee * found the alloy containing about 25 per cent of zinc and 3 per cent of copper to possess quite valuable proper- ties. When cast in sand or in chills it showed a tensile strength of 36,500 and 40,440 pounds per square inch, respectively, but was almost absolutely non-ductile. When hot rolled, however, it developed a ten- sile strength of over 60,000 pounds per square inch with a yield point of about 44,000 pounds per square inch, and an elongation of about 16.5 per cent. Aluminum-copper-manganese alloys have not attained any con- siderable importance as commercial alloys, but have been made the sub- ject of special study by the Alloys Research Committee f principally because, from the analogy of aluminum bronzes to ordinary copper-tin bronzes, it was anticipated that the addition of the strongly reducing manganese might benefit the former, as it is known to benefit the latter. The results of the investigation show that manganese does have the effect anticipated in a measure, b^it does not alter the properties of aluminum bronze to nearly the same extent that it does ordinary bronze. * Appendix to Tenth Report. t Ninth Report, Proc. Inst. Mech. Engrs., 1910, p. 119. THE NON-FERROUS METALS AND ALLOYS 603 The strength of the corresponding copper-aluminum alloy is only slightly raised by the addition of manganese below 4 per cent, but the ductility is considerably decreased. (The composition of the ternary alloys above discussed is about 9 per cent of aluminum, less than 4 per cent of manganese, and the balance copper. It might therefore be more properly called a copper alloy than an aluminum alloy, and cannot be classed among. the light alloys.) Minor alloys of aluminum. Alloys of aluminum with nickel, tin, manganese, tungsten, chromium, titanium, silver and antimony have been made commercially, but none possess such valuable distinctive properties as to have won for them an important place among the light alloys. The commercial alloy known as " nickel aluminum alloy " usually contains only a trace of nickel, and is for the rest an alloy of 2 to 7 or 8 per cent of copper in aluminum. 552. Alloys of Nickel. German Silver. The principal nickel alloys not Considered above are a large number of alloys of copper, nickel, and zinc, to which other metals are sometimes added in small amounts, and which are known collectively as German silver. This alloy is chiefly valuable because of its silvery white color and its non-corrodibility. A typical composition of German silver is nickel 18 to 20 per cent, zinc 28 to 32 per cent, and copper 50 to 56 per cent. The whitest metal is obtained with about 25 to 30 per cent of nickel and 20 to 25 per cent of zinc, but, since the nickel is the most expensive constituent, its pro- portion is often cut down. The alloy containing 30 per cent of zinc and about 18 per cent of nickel is quite white, makes smooth sound cast- ings, and is best from the standpoint of malleability and toughness. Aluminum in amounts up to about 25 per cent makes the metal more fluid while molten and is therefore desirable in castings. It also toughens the cooled casting. Iron hardens the metal, and makes it whiter when pres- ent in amounts not exceeding 1 to 2 per cent. Tungsten up to 1 or 2 per cent is also occasionally used to form an alloy called " platinoid " which has very high electrical resistance, and is used for electrical purposes. German silver is principally used in the making of domestic utensils, table ware, decorative objects, physical and scientific instruments, coin- age, etc. 553. Special Bearing or Anti-friction Metals. The bearing bronzes and the lead-antimony bearing metals have been considered above. Aside from the bearing bronzes, the best-known bearing metals are those composed of tin, copper, and antimony which are known as Babbitt metal. The composition of this alloy is extremely variable, but the usual .limits are tin 80 to 90 per cent, copper 3 to 10 per cent, and antimony 8 to 12 604 MATERIALS OF COKSTRtJCTION per cent. The quantity of antimony should always exceed the amount of copper in order to prevent brittleness. The ultimate constitution of Babbitt metals appears to be that of a ground mass of soft tin with hard crystals of a copper-antimony compound and a tin-antimony compound scattered through it. The hard particles carry the load and resist wear, while the soft ground mass allows the metal to adjust itself to the surface of the shaft and equalize the bearing pressure Alloys of lead, tin, and antimony have been considerably used as bearing metals, the best compositions being those containing 10 to 15 per cent of antimony, 10 to 20 per cent of tin, and the balance lead. Alloys of lead, copper, and antimony have occasionally been used as bearing metals where heavy loads are encountered. A typical composi- tion is 65 per cent of lead, 10 per cent of copper, and 25 per cent of antimony. Alloys of zinc, tin, and antimony, and alloys of lead, tin, and bismuth, have been used as bearing metals, but their general application has been very limited on account of a tendency toward fragility on the part of the former, and the high cost of bismuth in the case of the latter. CHAPTER XVIIl TIMBER * GENERAL 554. Timber as a Material of Engineering Construction. Timber has been one of the primary materials of engineering construction since the earliest times, and, despite the fact that it has been largely super- seded by concrete and steel in the construction of certain classes of struc- tures, still the total consumption of timber for structural and other com- mercial purposes is steadily increasing year by year. In spite of the great number of species of trees (something like 500 grow in the United States alone), only a very limited number of kinds of timber are of great commercial importance, the larger part of all timber used structurally being derived from only twelve distinct species of trees — those commonly known as -pine, fir, oak, hickory, hemlock, ash, poplar, maple, cypress, spruce, cedar, and walnut. These common names of species usually include several varieties, which may show quite diverse characteristics and possess, therefore, very different values as timbers of construction. Much confusion exists as to the common nomenclature of woods, the same species or variety being known by many different local names, and different species are sometimes known in different localities by the same name. The recognized botanical Latin nomenclature affords a dependable guide to species, but is too cumbersome for general use. In consequence, various national societies like the American Society for Testing Materials and the American Railway Engineering and Main- tenance of Way Association have adopted standard classifications of those particular structural timbers most used commercially, listing the various common names of a given species under a single name whose meaning is thus defined. Any study of the characteristics and properties of woods must include * The writer is indebted to the pubUcations of the U. S. Forest Service for most of the data upon which this chapter is based. The admirable text-books of Mr. Samuel J. Record, " Economic Woods of the United States," 1912, and " The Mechan- ical Properties of Wood," 1914, have been frequently consulted, however. 605 606 MATERIALS OF CONSTRUCTION some information concerning trees. The point of view of the engineer need not be that of the botanist, and his study may be confined to the general features, conditions and manner of growth of a Hmited number of species of trees, but physical and mechanical properties of timbers are closely dependent upon structure, and structure is not only dependent upon variety, species, and genus, but within a given variety, upon con- ditions of growth — climatic and soil conditions. No detailed botanical consideration of trees will be included herein, but a general classifica- tion of trees will be made, together with some study of growth and structure. TIMBER WOODS. GROWTH AND STRUCTURAL CHARACTERISTICS 555. Classes of Tretss. All trees are primarily divided into two botanical groups according to their manner of growth : Exogenous trees, or exogens, increase in diameter by the formation between the old wood and the bark of consecutive rings or layers of new wood which envelops the entire living portion of the tree. They lengthen by a sort of telescopic extension at the tips, each consecutive layer increasing the . length because of its conical form. Practically all classes of commercially important timbers are derived from trees of this group. Endogenous trees, or endogens, grow both diametrically and longi- tudinally, principally the latter, by the addition of new wood fiber intermingling with the old. Most endogens are small plants like corn, sugar-cane, wheat, rye, etc.-, but others like the palm, the yuccas and the bamboo have some value as a source of structural material. Exogenous Trees 556. Conifers. Conifers or gymnosperms, the needle-leaved, naked- seeded trees, form an important portion of our timber trees, compris- ing principally the pines, the spruces, fir, hemlock, larch, tamarack, cedar, cypress, and redwood. The conifers are widespread throughout the northern hemisphere. They are usually hght and soft, hence often called " soft woods." The trees may invariably be recognized by their needle leaves, their resinous bark and the cones which they bear. They are for the most part " evergreens." (The larch and the bald cypress are not evergreens, as they shed their needles annually, and some of the pines, some spruces, and tamarack are not soft woods but are quite hard.) 557. Broadleaf Trees. Broadleaved trees or dicotyledons (two seed- leaves) provide a source of timber second in importance only to the con- TIMBER 607 ifers. They comprise many varieties of oak, ash, hickory, poplar (cot- tonwood), maple, walnut,_ elm, chestnut, birch, beech, cherry, locust, basswood (Hnden), whitewood (tulip), sycamore, catalpa, butternut, buckeye, alder, willow, eucalyptus, gum, horse-chestnut, holly, boxwood, laurel, lignum-vitse, mahogany, satinwood, and many other species of lesser commercial importance. The broadleaf trees are found in wide- spread areas scattered over most of the globe. They are usually heavy and hard, hence often called "hard woods," and, as a rule, they are decid- uous, although many broadleaf trees are evergreen in certain climates. The broadleaf woods are not used for structural purposes to anywhere near the same extent as the conifers, but are specially adapted to use for interior finishing, cabinet work, furniture, etc. (Many broad- leaved woods are neither heavy nor hard, as for instance poplar, chest- nut, basswood, whitewood, willow, etc.) Endogenous Trees 558. Endogenous Trees. Monocotyledons. Monocoiyledonous trees or monocotyledons (one-seed-leaf) are largely confined to tropical or semi- tropical regions. The palms, because of their long straight stems and comparative immunity from the destructive action of the teredo (a form of marine wood-borer which is very active in some waters), are some- times locally used as piles, but have practically no other commercial uses. Only a few varieties are native to the United States. The yuccas find little application to commercial uses except as paper-pulp and as a veneer adapted to certain special uses where its lightness and flexibility are advantageous. The bamboo is not native to the United States and, where found, has been transplanted from Asia. The bamboo grows with extreme rapidity, but requires years to harden after its growth is attained. Bamboo has many commercial uses in Asia, particularly in Japan and in China, where it is even used structurally to a considerable extent. Its use in the United States is largely confined to small house- hold articles, furniture, etc. Exogenous Growth of Wood * 559. Pith, Wood, and Bark. The section of any exogenous tree exhibits first a central portion composed of loosely aggregated thin- walled cells called the pith. It is circular, star-shaped, ovoid, or tri- * For a much more detailed treatment of the growth and structural elements of wood see Record's " Economic Woods of the United States." The present discussion is largely based on this work. 608 MATERIALS OF CONSTRUCTION angular in shape; black, red brown, or gray in color; usually of small diameter, and does not increase in size after the first year. It probably assists the life processes of the tree at first by alternately storing and giving up plant food like starch and tannin, but it becomes inactive after a very few years and sometimes the pith cells disappear, leaving a pith cavity. Often the pith becomes compressed and is sometimes scarcely evident in sections of mature trees. Outside the pith the wood appears in concentric zones or rings of annual growth, the demarkation between which is evident because of the different structure of the wood slowly formed toward the end of one season and that rapidly formed in the succeeding spring. The various elements entering into the wood structure will hereinafter be considered in some detail. The outermost portion or periphery of the section is formed by mate- rial of variable and very complex structure called the bark. The origin, growth, and structure of the bark will be briefly considered in the dis- cussion which follows. The bark of many trees possesses a distinct com- mercial value. Many trees furnish bark for medicinal purposes, others, like the hemlocks and the oaks, supply a great part of the tannin used in the leather industries, several serve as a source of coloring matter, others furnish fiber for cloth and cordage, and one particular species provides the cork of commerce. 560. Primary Wood, Cambium, and Secondary "Wood. The tissue which forms the apex of a growing shoot is composed of simple thin- walled similar cells called the primordial meristem. This tissue soon becomes differentiated into three portions known as the protoderm (out- ermost), the procambium strands, and the ground meristem (innermost), respectively. The protoderm soon changes into epidermis, the outer- most portion of newly formed bark; the ground meristem forms the pith, the primary rays, that portion of the bark called the pericycle, and between epidermis and pericycle, the primary cortex. The procambium strands become vascular bundles, which occupy the zone between pith and pericycle and are separated from each other only by the primary rays. The vascular bundles comprise three classes of tissue, the phloem, which constitutes the innermost portion of the bark; the cambium, that layer of generative cells between bark and wood; and the xylem or wood fiber, which constitutes all the woody portion between pith and bark. These tissues, being formed prior to the development of the cambium^ make up the primary wood, so called in contradistinction to the second- ary wood, which is generated by the cambium. The epidermis is destroyed at an early'period and replaced by cork, formed by a cork cambium originating in the epidermis or in the cells TIMBER 609 just beneath. The development of cork cuts off successive portions of the cortex which dry up and ultimately scale off as outer bark. The pericycle is to a large extent made up of fibrous tissue which imparts toughness to the bark and protects the deUcate tissues beneath. The phloem is made up of several elements which resemble corresponding elements in wood. It contains tubes similar to the vessels in wood, which assist the life processes of the tree, particularly by allowing a downward circulation of food materials. The portion of the original vascular bundles called the cambium, which is capable of generation and growth, is originally isolated in the several bundles, but ultimately becomes united in a continuous sheath separating the entire woody cylinder from the bark. The cambium is made up of thin-walled cells which are particularly delicate when satu- rated with sap during the period of most vigorous growth. The cambial cells, by division and development, generate new wood or xylem on the one side, and bark or phloem on the other. All wood formed from the cambium is called secondary wood, and this constitutes all but a negligibly small part of the wood of a tree. The secondary wood provides mechanical support for the tree, affords a medium for the ascent of sap from the roots, and alternately stores and gives up the starchy foods necessary to the life processes of the tree. The structure of secondary wood is quite complex, and is- subject to wide variation, but is invariably made up of some or all of the follow- ing four distinct elements, viz., (1) vessels, (2) tracheids, (3) wood fibers, and (4) parenchyma. 561. Structural Elements of Wood. Vessels are tubular elements formed by the union of original cambial cells, the end walls of which have become wholly or partly absorbed, thus giving rise to a contin- uous tube of indeterminate length. The point of union of the segments is always marked by a constriction in the walls, and its plane may be either square or oblique. In the latter case the perforation between the segments is often sclariform, that is, the opening is partially closed by a series of parallel cross-bars. In most cases the walls of the vessels are provided with many small gaps called pits. A pit is simply a small portion of the wall where the original cellulose membrane of the primary cambial cell has not become thickened by the addition of Ugnin. A canal is thus formed, closed only by the thin membrane of the primary cell wall. This canal often widens toward the primary cell wall, and, if the widening occurs suddenly, the pit is called a bordered pit, otherwise, it is a simple pit. The function of the pits is to facilitate the passage of water and food between adjacent cells. (Water only in the case of pits in the walls of vessels.) 610 MATERIALS OF CONSTRUCTION The vessels are always continuous for great lengths, often the entire length of the tree. In diameter they are sometimes very small (less than 0.005 inch) as in the poplar, in others, hke the chestnut they are large (0.01 to 0.03 inch) and visible to the naked eye as pores in the cross- section. The function of the vessels is to provide unobstructed passages through which water may ascend from the roots to the branches. They con- tain no protoplasmic matter after becoming fully developed, but those in the older part of the wood may become filled with gums, resin, lime- carboiiate, etc. Tracheids are elongated single cells of tubular form, closed at their ends and characterized by the presence of bordered pits in their side walls. They are polygonal in cross-section, arranged in radial rows, and become flattened and thicker-walled toward the end of each season's growth, the lumen or interior opening being therefore smaller toward the periph- ery of each growth ring. The tracheids form the bulk of the wood of conifers, wherein they attain a length of from 0.1 to 0.2 inch. Their diameter seldom exceeds 0.002 inch. In broad-leaved woods the tracheids are much less impor- tant elements than in the conifers, and may even be entirely lacking. If present, they are much smaller and less uniform in size, shape, and arrangement. The function of the tracheids, aside from affording mechanical sup- port to the tree, is to assist in the circulation of water. Wood fibers are narrow, elongated, sharp-pointed single cells, having very thick walls and a very small lumen. They usually have slit-like, oblique simple pits, but occasionally show small bordered pits. Wood fibers are not found in coniferous woods, but constitute the principal source of strength, hardness, and toughness of broadleaf woods. They occur most abundantly in the intermediate portion of a ring, and attain a size varying from 0.02 to 0.10 inch in length. The fibers are usually straight and have tapered ends, but sometimes they become distorted and interwoven, or the ends may be forked or saw-toothed. This produces a wood of irregular grain which is extremely tough and offers great resistance to splitting. Parenchyma are elements made up of rows of thin-walled cells joined end to end. They resemble wood fibers except for the presence of cross- walls (which are as thick as the side walls), and the shape of the pits, which are rounded simple pits instead of slit-like oblique, simple pits. Occasionally, as in most oaks, the cross-walls form small chambers, each of which contain a single crystal, usually of calcium-oxalate. Paren- chyma fibers are found in all classes of woods, both in the vertical di.rec- , TIMBER 611 tion and the horizontal (in the rays). In broad-leaved woods they may be scattered throughout the growth rings (comprising the periphery of the growth ring), ranged in tangential or radial bands, or ranged around the large vessels. In the wood of the conifers parenchyma are called resin cells. Sometimes their arrangement is scattering, sometimes in concentric zones, and sometimes they occur in groups, especially around resin ducts. The chief function of the parenchyma is the distribution and storage of elaborated food materials. In the conifers they are invariably asso- ciated with the formation and storage of resin. The typical forms of various wood elements are shown by Fig. 324. fee Tb. p. . n ' L Fig. 324.— Typical Wood Cells. A, Wood fiber with very narrow lumen; B, Wood fiber with large lumen showing oblique, sUtlike simple -pits (s.p.); C, End of wood fiber showing saw edge; C, End of wood fiber showing forked structure; D, Ends of two tracheids from Pinus showing numerous bordered ipits (6. p.); -E , Tracheids from Quercus; F, Wood parenchyma fiber showing individual cells and simple pits (s.p.) ; G, Chambered wood-parenchyma fibers from Juglann showing crystals of calcium oxalate: H, Conjugate parenchyma cells; K, Portion of vessel segment showing simple perforation (p) ; L, Portion of a Vessel segment showing scalariform perforation (Sc. p.). (Record.) 562. Rays, Resin Ducts, and Pith Flecks. Rays, often called medul- lary rays or pith rays, are radial, horizontal lines or bands of cells which cross the growth rings at right angles. Those which originate in the pith are called primary rays, while those which have originated in the cam- bium at any point are known as secondary rays. All rays are continuous from their origin into the bark. Rays consist of radial series of cells, usu- 612 MATERIALS OF CONSTRUCTION ally elongated horizontally, but sometimes elongated vertically m certam broad-leaved woods. In conifers the rays are for the most part only one cell wide and not more than twenty cells high, but in the resm-bearing conifers like the pines, spruces, larch, and hemlock, they may contam resin ducts and be several cells wide. Such rays are called fusiform rays (Fig 325.) In the broad-leaved woods the rays vary from one or two cells in width to very large rays, 25 to 75 cells wide and several hundred cells high (amounting to an inch or more sometimes, Fig. 326). In the coniferous woods the rays are composed largely of paren- chyma, but several species, par- ticularly the resin-bearing ones, show ray tracheids composing at least the upper and lower rows of cells, their presence being made evident by the bordered pits in their walls. In broad-leaved species the rays are composed wholly of parenchjrma. The ray parenchyma are provided with simple pits in their side walls and particularly in their end walls. The principal function of the rays is the lateral distribu- tion of plant food. Resin ducts are simply long, narrow, intercellular channels sur- rounded by parenchyma or resin cells. Unlike the vessels, they have no walls of their own. „ one T, ^icj 4.- cou ii en- Resin ducts are coHimon Only to Fig. 325. — Tangential Section of Shortleaf Pme . •' Showing Rays. (Magriification 125 Diam- resm-bearing trees and usually eters.) (Bull. 101, U. S. For. Ser.) occur particularly in that por- tion of the growth rings between the early and late wood. Fig. 327. The average diameter of the larger resin ducts is about 0.01 inch. Besides the large resin ducts which extend vertically, other smaller ducts are found running horizontally in the larger rays, and the two series are united at frequent intervals. Abnormal resin ducts may be developed as the result of injury, and in the case of the long-leaf pine the outer layers of the sapwood are intentionally so chipped as to form inclined ducts through which the resin may be collected and tapped. This is the source of most of the turpentine of commerce. TIMBER 613 Pith flecks, or medullary spots, Fig. 328, are small, brown, crescent- shaped patches appearing on the cross-sections of certain woods, espe- cially the birches, the maples, cherry, poplar, willow, etc. The origin of pith flecks is pathologic, they being caused by the -work of the larv£e of certain insects. These insects deposit their eggs in the bark of the smaller branches and the larva subsequently develop in the cambium layer. In the early part of the growing season the larva travels downward in search of food, leaving a channel or mine which extends through the cambium to the base of the tree. The larva then doubles back, and for the next few weeks mines back and forth, seldom going more than 10 feet above the ground. By the middle of the summer it travels into the roots and thence presses out into the soil. The passage of the larva destroys those cells in its im- mediate path and at first the mine is scarcely more than 0.05 inch wide. As the larva grows the mine is widened until it may finally be 0.1 or 0.2 inch in the circumferential direction. It rarely occupies the entire thick- ness of a growth ring, since the larvae begin their operations after the beginning and leave before the end of the growing season. The cambium becomes united again toward the periphery of the growth ring and the mine is left behind in the annual ring. The passage thus left independent of the cambium soon becomes filled with new cells similar in character to those in the rays, this action being due to the formation of wound tissue. The remains of dead cells and excrement of the leavse are responsible in most cases for the discoloration of the pith fleck. Occasionally pith flecks cause dis- integration of surrounding normal cells and thus give rise to intercellular spaces. Pith flecks often render lumber unfit for certain uses by marring the beauty of its grain and, in the case of the cherry, by seriously impair- FiG. 326. — Tangential Section of Circassian Walnut. (Cir. 212, U. S. For. Ser.) V, vessel: P-r., pith ray; w./., wood fibers; w.p.f., wood- parenchyma fibers; c.w., cross wall; b.p., bordered pits. Magnified 30 diameters. 614 MATERIALS OF CONSTRUCTION ing its quality owing to the dis- integration of the adjoining woody tissue. Where a pith ray encounters a pith fleck its course is abruptly terminated and the physiological activities of the tree are thereby interfered with, sometimes resulting in the sur- rounding wood becoming dark- ened prematurely as heart wood. 563. Annual Growth Rings. Spring and Sununer Wood. The growth of all exogenous trees has been above explained as a process of formation of new wood fiber between the old wood and the inner bark, through the agency of the cambium. Owing to the inability of trees to sus- tain their physiological activities indefinitely, and the effect of the alternation of seasons in all Fig. 327.-Transverse Section of Shortleaf Pine ^gjj^pgj.^^g zones, this growth is Showing Resin Duct. (Magnification 125 .,..,. i ,i /• diameters.) (Bull. 101, U. S. For. Ser.) intermittent, and the zones of growth correspond to the annual periods. The succeeding rings of growth may easily be distinguished from one another in most species because of the different structure of the wood formed rapidly in the spring and that more slowly added in the summer. (No wood is added during the winter months.) The distinc- tion between adjoining growth rings is sometimes augmented by the deposition of infiltrated pigments or resin in the late wood. Thus it happens that the age of most trees may be ac- curately determined by counting the annual rings on a section of the stem. The difference in appearance between the spring wood and the „ or.„ r„ • , ^ . 1 • . „ Fig- 328.— Tangential Section of Silver iSInple, summer wood IS occasionally SO showing Pith Flecks, i Natural Size. (Cir! marked as to make the two appear 215, U. S. For. Ser.) IMIfag^ o ^ ^ hI ^^paBB^aa^^t^y^aL^^^^'y^ ss; ;ru r2K^^^j{a^^^K^5?^5=? ® b; wMImP^K/J^^ \ife S S W^^Fifryff^ yftp ® ■P ^fe&JP^"^'" ^^^" jp & |^^ym\>^ iiVfe ,« ^^^ps^siiK)yj\ \ ^>^^^*^^y^^ fel « B^B^^fe i 1' i 3 = >» ®®*®®^® o o i ii 1 o o o Q 1 a c c c i s 1 o c o c C3 tie S , « tM 3®,^ ?•«?•. ^ 2® 1 ^ ss ^li!l h5^ a ^ •* J- i] M H J- tdtju ^is *> M^^^i \% r arf^^B-^ «^ 4Ui ■w" Ii TIMBER 615 as distinct bands within a single annual ring, as in the species commonly called, collectively, hard or yellow pine. More commonly, however, the spring wood merges gradually into the summer wood, and the only sharp line of demarkation is between the summer wood of one season and the spring wood of the following season. The structural difference between spring wood and summer wood may consist (1), in either an abrupt or gradual reduction in the number or the size of vessels in the later wood, e.g., oaks, chestnut, ash, locust, etc., show an abrupt change and are called " ring porous," Fig. 329, while other broad-leaved woods like maple. ^' -i . ! I I 1 i I . .rrn:-( M .'I A ■ * .1 I I if !;■•■; ■ill/' Fig. 329.— Cross-section of ' Ring-Poroug Pig. 330.— Cross-section of Diffuse-Porous Wood. (Red Oak). (BuU. 126, U. S. Wood. (Hard Maple.) (Bull. 126, U. For. Ser.) S. For. Ser.) walnut, beech, birch, etc., show only a gradual change and are called " diffuse porous," Fig. 330; (2) in a change in the kind of wood elements, e.g., where vessels are absent from the later wood, being replaced by wood parenchyma or tracheids; or, (3) where the cell wall^ of the various wood elements become thicker as the season progresses, the lumen becom- ing correspondingly smaller and the wood more dense. This structure is characteristic of white pine, spruce, etc., Figs. 327, 331, 332, 333. It is a, commonly observed fact that the rate of growth of trees is quite variable, not only in different species, but even for different speci- mens of the same species. This means that the growth rings are of van- 616 MATERIALS OF CONSTEUCTION able thickness. When the conditions of soil, light, heat, moisture, etc., are such as to produce normal thrifty growth, the width of the rings is greatest near the pith, decreasing outward, and it is also normally greater at the base of the stem, decreasing upward. Unfavorable soil or climatic conditions, or unfavorable seasons, will disturb the normal regularity of growth, however, and the cross-section of the stem there- fore presents a history of the growth in succeeding favorable and unfavor- 1 I -hi 1 \ ri. ^ i 1 1' f I^H ill 1 1 ^1 1 t1 1 1 11 ■ 1 1 i J|] ' ^ = ■ Fig. 331. — Transverse Section of Norway- Pine. Magnified 25 Diameters. (Bull. 139, U. S. For. Ser.) e.w., earlywood; Z.to,, late wood; (., tracheids; p.r., pith ray; r.c, resin cauaL. G. 332. — Radial Section of Norway Pine. Magnification 25 Diameters. (Bull. 139, U. S. For. Ser.) e.w., early wood; l.w., late wood; (., tracheids; ■p.r., pith ray; r.c, resin canal in pith ray; also a longitudinal canal. able seasons. The thickness of the ring is not even uniform circum- ferentially, because of unequal acceleration of the growth on different sides; thus the section often becomes oval and, even if circular the pith is eccentric (Fig. 334). ' The maximum thickness of growth rings attained during the period of thriftiest growth rarely exceeds 0.5 inch for either conifers or broad- leaved trees. For most trees a thickness of 0.10 inch to 0.15 inch indi- TIMBER 617 cates a good thrifty growth, and, for trees grown under unfavorable conditions, as well as for the outer wood of very old trees the ring thickness may not exceed 0.005 inch to 0.02 inch. Trees grown in dense forests always grow less rapidly than trees grown in the open, and the growth is apt to be less rapid (the rings thinner) at the base of the stem than farther up. (Note that this is a reversal of the normal habit of growth.) In most of the conifers the dis- tinction between spring wood and Fig. 333. — Tangential Section of Norway Pine. Magnification 25 Diameters. (Bull. 139, U. S. For. Ser.) t., tracheids; p.r., pith raya; r.c, resin canal in pith ray. Fig. 334. — Transverse Section of the Stem of a Young Balsam Fir Tree, Showing Annual Rings of Growth, a. r. i Natural Size. (Bull. 55, U. S. For. Ser.) summer wood is due to a thickening of the walls of the tracheid cells and a flattening of the cells radiallyj resulting in a greater density in the sum- mer wood. The proportion of summer wood in a growth ring is normally least in the wood of the sapling, greatest in the intermediate period of thrifty growth of the tree, and falls below the average again in old age. It should average 40 to 50 per cent of the wood near the base of the stem, and falls slightly below this average at the base of the limbs. The proportion of summer wood largely determines the heaviness, strength, and structural value of the wood. 618 MATERIALS OF CONSTRUCTION In most of the broad-leaved trees the distinction between spring and summer wood is largely due to a diminishing in the space occupied by vessels, and the summer wood is again the denser, heavier, and stronger portion of the wood. Unlike most coniferous woods, however, the ring- porous broad-leaved woods form their densest and strongest wood dur- ing the period of most rapid growth. This fact is due to a difference in character of the wood elements formed under different conditions of growth. When growth is most rapid, and the rings therefore widest, the middle portion of each ring contains a great abundance of the thick- walled, strong, and tough wood fibers. When the rings become narrower, it is at the expense of these wood fibers, whose strength-giving qualities cannot be equaled by the thin-walled vessels and parenchyma present. This accounts for the general preference for " second-growth " hickory, ash, etc., which has grown in less dense forest than the virgin timber, and whose growth has therefore been more rapid. 564. Sapwood and Heartwood. As the process of formation of annual rings of new wood adds layer after layer of vigorous healthy tissue over that previously formed, the latter gradually ceases to take an active part in the physiological activities of the tree, loses its proto- plastic contents, and dies. Decay does not usually follow immediately, however, and the dead wood continues sound and provides mechanical support for the tree. The Hving elements of the tree are called " sapwood " and the dead elements " heartwood." There is usually a sharp line of demarkation between the sapwood and heartwood, although the vigor of the living wood decreases progressively from the cambium inward. The pro- portion of sapwood varies considerably in different species and also between individuals of the same species. Certain woods such as maple, ash, beech, hickory, etc., usually form thick sapwood, while the juniper, the catalpa, the locust, the yew, and many others, normally form thin sapwood. Since the sapwood zone is the outer one, it forms a consider- able percentage of the volume of the wood. The normal percentage of sapwood in the hickory, for instance, is about 75 per cent; in the maple, ash, beech, etc., it exceeds 50 per cent; while in the juniper, catalpa, yew, etc., it does not usually exceed 25 per cent, and in the case of the locust may not amount to more than 15 per cent. Within the same tree the percentage of sapwood usually decreases from the base upward and is least in the branches. All young trees show a higher percentage of sapwood than do old trees of the same species. The distinction in color between sapwood and heartwood, which is characteristic of most woods, is due to the darkening of the dead wood by the presence of infiltrated pigment, gums, resins, etc., which permeate TIMBER 619 the cell walls and sometimes also the cellular and intercellular cavities. Certain woods, like spruce, fir, hemlock, poplar, willow, gum, etc., show little or no difference in appearance between the two portions. As a rule, the heartwood is more highly valued than the sapwood of the same variety. Important exceptions, however, are the hickory, the ash, the birch, all the paper-pulp woods, and timber to be impreg- nated with preservatives, where the sapwood is considered preferable. Endogenous Growth of Wood 565. Endogenous Growth. The general features of endogenous growth have been indicated above in Arts. 555 and 558. Any detailed consideration of the growth and structure of endogens (monocotyledons) is not justified, because of the relative economic unimportance of trees of this class as sources of structural timber. The elements of the wood of endogens are similar to those of exogens, but their disposition and arrangement are radically- different. The vascular bundles found in endogens are not grouped in concentric cir- cles, but are scattered throughout the volume of the wood, each one being isolated from neighboring ones by thin-walled tracheary cells which form a pith which offers little resistance to the growth of the bundles. The bundles are not even parallel to the stem in most cases, but each one curves inward in a vertical plane (from the point where it entered a leaf), and then outward again, thus crossing many other bundles and making the structure more complex. Usually the development of new tissue is more rapid at the periphery and in the outer zone of the stem than in the interior. This results in the production of much more dense tissue toward the exterior, the solidity of the wood decreasing from the circumference toward the center. Often, when the growth is very rapid, sluggishness on the part of the interior fiber causes it to become ruptured and the enlargement of the cavity finally produces the hollow stem character- istic of the bamboo and many grasses, but not of the palms, yuccas, etc., which have a pithy center. The occurrence of knots or joints at frequent intervals is common to many endogens, particularly those which form a hollow stem. The knots mark the places whence leaves have emerged. Structure op Wood of Exogens 566. Structure of Wood of Conifers. The characteristic structure of coniferous woods or gymnosperms, Figs. 327, 331, 332, 333, 334, is very simple and uniform, consisting almost wholly of tracheids, resin ducts, and small parenchymatous rays. The tracheid cells are usually 620 MATEEIALS QP CONSTRUCTION arranged in straight radial rows and form the great bulk of all the wood. The uniformity of this structure is disturbed only by the resin ducts, which are scattered through the summer wood portion of each annual ring, and by the rays which are small and composed usually of parenchy- matous cells. Conifers contain no vessels, and pores therefore do not appear on the section; wood fibers are wholly absent and parenchyma occur only around the resin ducts and in the rays. The tracheids have thin walls and large lumen in the spring wood, but become thick-walled and flat- tened radially in the summer wood, the lumen being therefore much reduced in size. 567. Structure of Wood of Broadleaf Trees. The structure of the wood of broadleaf trees or dicotyledons, Figs. 326, 329, 330, is very com- plex as compared with that of the conifers. This is due principally to the greater diversity of elements present, and to the presence of numer- ous and very large rays. Vessels are numerous in the wood of practically all dicotyledons and pores are usually prominent in the cross-section of the early wood, causing the spring wood to appear darker than the summer wood. Tracheids in broadleaf wood are always subordinate elements, and often are found only in the immediate vicinity of vessels or not at all. There are no tracheids in the rays of dicotyledons. Wood-fibers are common and constitute the principal source of strength and tough- ness in most broadleaf woods. They are invariably most abundant in the median portion of a growth ring and particularly in the rings of the greatest width. Parenchyma fibers are present in the wood of prac- tically all dicotyledons and are often very prominent. Various common arrangements of the parencyhma fibers have been noted above in Art. 561. The rays of dicotyledonous woods are composed wholly of paren- chyma and are, as above noted, often large and conspicuous. PHYSICAL CHARACTERISTICS OF WOOD 568. Grain and Texture of Wood. The physical appearance of wood, so far as that is dependent upon the character and arrangement of wood elements, the width of the growth rings, etc., is described by the common terms " fine grained " and " coarse grained," " even," " smooth," or "straight grained," and "uneven," "twisted," or "cross grained," and, in a few instances, by the less common terms " curly," " mottled/' " birds-eye," etc. Woods are fine grained if the growth has been slow, resulting in narrow rings; coarse grained if the rings are wide. If the main wood elements of a tree run parallel to its axis the log will be straight grained; often, however, the fibers follow a spiral course around the TIMBER 621 tree, producing a log of twisted grain. Sometimes the fibers may be oblique in one direction for several years' growth, and then become oblique in the opposite direction for a time, producing a log of cross-grained tim- ber. When a straight-grained log is sawn, either straight-grained or cross- grained timber may be obtained, depending upon the parallelism of the plane of cutting and the axis of the log. Straight-grained lumber may be obtained from a log of twisted or cross grain only by splitting, lumber cut parallel to the axis being inevitably cross grained. When the layer of wood newly formed beneath the bark becomes pitted or marked by prominences, due perhaps to the presence of dormant buds, these depressions or elevations may persist for several years, result- ing in the production of the beautiful effects known as " birds-eye " marking when boards or veneer are cut tangentially. The maple has a special tendency to preserve the contour of the growth rings for many years, and bird's-eye markings and curly grain are more often encountered in maple wood than in most other woods, wherein the tendency to com- pensate for inequalities is strong and accidental elevations or depressions in the growth rings do not persist more than two years. The term " texture " as applied to woods refers to the size, char- acter, and arrangement of the wood elements in so far as they affect the structural characteristics of the wood. Coarse texture and fine texture are terms applied to woods having many large elements (particularly if many large pores are present), and to woods wherein the opposite condition prevails, respectively. The texture is even, or uniform, if the wood elements show little variation in size, and uneven if the contrary is true. The characteristic wood of the redwood tree and coffee tree, for instance, is coarse textured, while that of the jumper, poplar, willow, etc., is fine textured. The bald cypress, the juniper, the redwood, etc., are even textured, while all ring-porous woods, hke oak, ash, elm, and all woods with very different early and late wood like Southern long-leaf pine, Douglas spruce, etc., are uneven textured. Knots originate in the timber cut from the stem or branches of a tree because of the encasement of a limb, either hving or dead, by the successive aimual layers of wood. Most limbs originate at the pith of the stem, and the knots found deep in a log are therefore small, increasing in size toward the bark. So long as the Hmb is growing, its layers of wood are a continuation of those of the stem. But a majority of the Hmbs die after a time, and, if a portion of a dead limb is subsequently encased by the growing stem, there will be no intimate connection between the new stem wood and the dead wood of the limb, and a board so cut as to intercept this portion of the log will contain a loose knot. A board cut from the log at such a depth that the limb is intercepted at a point where 622 MATEEIALS OF CONSTRUCTION it was encased while still living will contain a sound knot, unless the knot has rotted, become badly checked, or contains a large pith cavity. A sound knot is usually harder than the surrounding wood and in conifer- ous woods is apt to be very resinous. On this account it may consti- tute a defect because of its non-retentivity of varnish or paint. Other- wise it constitutes a defect only on account of the disturbance to the grain and difficulty caused in working, or, in the event of its occurrence on the under side of a timber used as a beam, a weak point exists owing to its small resistance to tensile stress. A knot constitutes an impedance to the splitting of timber, since the fibers of the stem wood above a hmb bend aside and pass around the limb while the fibers below run contin- uously into the limb. Thus it happens that a cleft started above a limb will never run into a knot, but one started below is very apt to do so. 569. Color and Odor. Color is a great aid to the identification of species and variety, and, within a given variety, often constitutes an important criterion of quahty. New wood is practically colorless, but becomes yellowed after a few years and, except in the cases of those species which do not form a distinctive heartwood, a decided deepening of color occurs when the wood ceases to take an active part in the fife processes of the tree. The sapwood of practically all species shows no great variation from the characteristic light yellow, but the heartwood shows great variation, distinctive colors being characteristic of a given species or variety. The following enumeration of a few common woods illustrates the wide variation in colors characteristic of the heart wood: Creamy white Holly, buckeye Yellow Papaw, sumac, osage orange Dark brown Black walnut, sweet gum Brownish-red Redwood, cedar Yellowish-white Tulipwood. poplar Light brown White oak' chestnut Red-brown Red oak, red ash ■"'^''k Persimmon In many instances the color is uniform, as in the wood of the walnut, the oaks, the chestnut, the elm, the redwood, the cedar, the holly, the buckeye, etc. In other cases the color is variable, as in the wood of the sweet gum, the rich brown of which is streaked and mottled with black the tulip wood, which varies from a deep greenish yellow to brown the blue ash, which is Hght yellow with brown streaks, and the osage orange whose yellow wood is richly streaked with orange and red. Deep color is, as above explained, always due to the infiltration of resms, pigments, tannin, etc., which, as a rule, possess an antiseptic nature and so contribute to the durability of the wood. In most in- stances, therefore, deep color in a specimen of a given variety is an indi- TIMBER 623 cation of greater durability than characterizes lighter colored specimens, particularly the sapwood, of the same variety. Dark colors are usually desirable in woods used for cabinet purposes and interior finish, and such woods are very commonly artificially darkened by the use of a filler containing Vandike brown or some other pigment. Many woods also are often stained to imitate more valuable cabinet woods, a familiar example of which is the use of birch or white mahogany stained to resemble red mahogany. All woods darken with age upon exposure, and some, like the cherry and mahogany, darken very appre- ciably. On this account the natural color of woods can only'^be observed on newly cut sections. All woods darken, some to an extreme degree, upon immersiori in water. The odor of wood, Uke the color, is due not to the wood itself, but to the foreign chemical compounds present. It is therefore usually more pronounced in heartwood than in sapwood, and is less pronounced after exposure than when wood is freshly cut. Sometimes the odor of green wood is entirely lost upon exposure, and, in some instances (e.g., the oaks), a new and often disagreeable odor is acquired, owing to the decom- position of organic compounds present. All woods possess a characteristic odor in some degree, although it is not readily perceptible. Pines have a marked resinous odor; cedar has a strong aromatic or spicy scent which persists for many years, and which is valued because it is offensive to certain insects which attack woolen goods, furs, etc.; bald cypress possesses a rancid odor; catalpa smells somewhat like kerosene; and many other species such as the juniper, arbor vitse, hemlock, sassafras, camphor-tree, etc., possess well- known distinctive odors which are not easily described. Decaying timber usually possesses a strong odor which may be very disagreeable, as in the case of the poplar, or fragrant, as in the case of decaying red oak. The odor of green wood, seasoned timber or decaying timber often provides an infallible indication of species. 570. Density and Weight. The density of woods of different species, different individuals of the same species, and even portions of the same individual, varies considerably, owing to differences in structure (par- ticularly, differences in the average thickness of cell walls), and to dif- ferences in the amount of water present in the cells and cell walls. The specific gravity of the ultimate wood fiber of all species, however, is about 1.6, so that it is apparent that no wood would float in water were it not for the buoyancy of the air present in the cells, walls, and intercellular spaces. In all varieties of trees the sapwood is heavier than the heartwood because of its greater water content, and the summer wood is heavier than the spring wood because of the smaller lumen and thicker walls 624 MATERIALS OF CONSTRUCTION of the cells of the late wood. A fiber from the dense summer wood of such a tree as the long-leaf pine will not float, regardless of its water con- tent, and many woods will not float while stiU very green. Even sea- soned timber of many species may be made to become " water-logged," i.e., a large part of the imprisoned air displaced by water. . Within a single tree of a given species specimens taken from differ- ent portions of the stem and branches exhibit wide variation in weight. All trees form their heaviest wood at the butt and at the base of the limb (knots), and the weight decreases toward the upper portion of the stem and the tips of the branches. If the weight of the several zones of wood formed during successive periods of growth be compared, it will be found again that the weight varies considerably, the manner of variation being fairly constant for all specimens of a given species grown under normal climatic and soil conditions, but differing between species of radically different habits of growth. Two factors chiefly influence the weight of wood formed at succeeding periods of growth: first, the proportion of summer wood formed, and second, the size and number of pores present in the early wood. The first is the controlling factor in the case of most conifers, and such trees as pine, spruce, etc., form light wood as saplings when the proportions of early wood is greatest, the heaviest wood dur- ing the period of thriftiest growth when summer wood preponderates, and lighter wood again in old age when the proportion of summer wood drops below the maximum. In most dicotyledons, on the other hand, especially those whose pores are conspicuous, hke the oak, chestnut, ash, etc., the heaviest wood is that of the sapling and the wood becomes slightly lighter with each succeeding year's growth. This is due to the fact that the pores in the wood near the pith are very minute, but larger and larger pores are formed during succeeding periods of growth. The weight of wood is in itself an important quality in many of its applications to structural uses. Weight is also closely related to strength, providing the disturbing factor of variation in moisture con- tent is eliminated, so that, at least within a given species the relative strength of different specimens are directly proportional to their weights. The weight of wood is experimentally determined by subjecting thin discs of the wood to an oven temperature of 100° C, till they cease to lose weight by evaporation of moisture. Taking into account the con- siderations above discussed, it is evident that the results of experimental determinations will be extremely variable, and the value usually assigned to a given species is simply the average of a large number of tests Such a value has only a general application, and especially so because the moisture content of lumber is variable and always amounts to at least 8 or 10 per cent unless the lumber is kept in a dry-kiln. TIMBER 625 The following table, abstracted from the Tenth Census of the United States, gives an approximate idea of the relative specific gravity and FIFTY TREES OF THE UNITED STATES, ARRANGED IN ORDER OF AVERAGE SPECIFIC GRAVITY (Abstract from 10th U. S. Census) Common Name. Lignum vitae Live oak WMte. hickory . . . Pignut hickory. '. . Osage orange . . . . Cuban pine White oak Locust Blue beech Mahogany Cedar elm Blue ash Pecan hickory. . . . Long-leaf pine . . . Slippery elm Sugar maple Beech Yellow birch. ... White ash Red oak White elm Red ash Tamarack Black walnut . . . Short-leaf pine . . Paper birch Sweet gum Western juniper. White birch Sycamore Loblolly pine . . . Douglas spruce. . Red juniper Bull pine Red fir Black spruce. . . . Bald cypress. . . . Basswood Chestnut Catalpa White basswood. Hemlock Redwood Butternut White spruce . . . Incense cedar . . . White pine Balsam fir. . .». . . White fir Cork wood Gen. Clas9. Broad leaf Conifer Broad leaf Conifer Broad leaf Approx. Sp.gr. Conifer Broad leaf Conifer Broad leaf Conifer Broad leaf Coni'.' Broad leaf Conifer t ( Broad leaf Conifer Broad leaf Approx. Wt. Lbs. per Cu.ft. .14 71 .95 59 .84 52 .82 51 .77 48 .75 47 .75 47 .73 46 .73 46 .73 46 .72 45 .72 45 .72 45 .70 44 .70 44 .69 43 .69 43 .66 41 .65 40 .65 40, .65 40 .63 39 .62 39 .61 38 .61 38 .60 37 .59 37 .58 36 .58 36 .57 36 .54 34 .52 32 .49 31 .47 29 .47 29 .46 29 .45 28 .45 28 .45 .. 28 .45 28 .43 27 .42 26 .42 26 .41 26 .41 26 .40 25 .38 24 .38 24 .36 22 .21 13 Wt. ClasB. Very heavy Heavy ' ti Medium Light Very light 626 MATERIALS OF CONSTRUCTION weight of fifty of our most important lumber woods. The limited application of these average values should be remembered, and their use confined largely to the comparison of particular specimens of equal moisture content, and to a general classification of woods as very heavy, heavy, medium, light, and very fight. The weight classification here given is that adopted by Roth.* 571. Moisture Content of Wood. Water occurs in the sapwood of living trees in three states: (1) It forms more than 90 per cent of the protoplasmic contents of the living cells, (2) it saturates the cell walls, (3) it partially or entirely fills the lumina of empty lifeless cells, fibers and vessels. The heartwood water occurs only in the second state, i.e., it exists in the cell walls. As a typical illustration of the distri- bution of water, Roth states that in the fresh sapwood of white pine water comprises about one-half of the total weight, distributed about as fol- lows: 5 per cent in the contents of living cells; 35 per cent in the cell walls; 60 per cent in the lumina of empty cells. The highest percentage of water is found in the wood near the bark, decreasing gradually toward the pith unless heartwood is formed, in which ^latter event, an abrupt decrease in the moisture percentage occurs at the heartwood limit. From the above considerations it will be apparent that, in the case of trees which do form heartwood, the moisture content at any section, will vary with the proportion of sapwood, and is there- fore greater in the upper than in the lower portion of the stem, still greater in the limbs, and greatest of all in the roots. It is impossible to determine the percentage of water in wood exper- imentally. For practical purposes, however, wood is considered thoroughly dry when a thin cross-sectional disc ceases to lose weight in a constant temperature of 100° C. At higher temperatures additional water will be given off, but chemical destruction sets in before all the water is driven off. 572. Seasoning of Timber. The seasoning or drying of timber neces- sarily precedes its application to structural purposes. The natural dry- ing of timber by long outdoor exposure to the action of the air is called " seasoning," while artificial drying by exposure for a limited period to high temperatures in a closed chamber is called " kiln-drying." In either event the loss of moisture is entirely by evaporation, and the treat- ment is appreciably beneficial to strength, stiffness, and durability. The rate of drying of timber depends upon the dimensions of the piece and the structure of the wood. Thin boards dry much more rapidly than thick planks or heavy timber, tod light porous woods much more rapidly than heavy dense woods. Moisture evaporation from the cross- * Bulletin 10, Forestry Division, U. S. Department of Agriculture. TIMBER 627 section of porous wood is, at first, very rapid, that from the radial section very much less rapid, and that from the tangential section still less rapid. The total moisture evaporation from these different sections in a given time will bear this relation one to the other, however, only when equal areas are exposed on each section. This is due to the fact that drying proceeds inward from any section at a rapidly diminishing rate. Numerous experiments made by Mr. H. S. Betts of the U. S. Forest Service upon large timbers of loblolly and long-leaf pine estab- lished the following conclusions: * " (1) The drying-out process takes place ahnost wholly through the faces of the beams and not longitudinally, except near the ends. " (2) The rate of evaporation through a surface is proportional to the rate of growth or density of the wood near the surface, being most rapid in case of the sapwood. " (3) If the whole stick is made up of heartwood or the proportion of sapwood is uniform throughout, the longitudinal distribution of moisture is very regular. If the proportion of sapwood is not uniform, on the other hand, the portion containing the most sap is the most sus- ceptible to moisture influences, i.e., it will dry or will absorb the moist- ure the most rapidly. " The average of two cross-sections of long-leaf pine sticks, 12 by 12 inches and 8 by 16 inches, and 16 feet long, which were air dried for two years, showed an average moisture content in the outer portion, cut half, way from surface to center, of 17.7. per cent, while the inner part contained 25.7 per cent. " From this it is quite evident that where timber of structural sizes is used, the strength ordinarily reckoned upon should not be greater than that of the green condition." In artificial drying temperatures of 158° to 180° F., are usually employed.f Pine, spruce, cypress, cedar, etc., are dried fresh from the saw, allowing four days for 1-inch boards. Hard woods, especially oak, ash, maple, birch, sycamore, etc., are air seasoned for three to six months to allow the first shrinkage to take place more gradually, and are then exposed to the above temperatures in the kiln for about six to ten days for 1-inch lumber. Freshly cut poplar and cottonwood are often dried directly in kilns. By employing lower temperatures, 100° to 120° F., green oak, ash, etc., can be seasoned in dry kilns without danger to the material. Steaming the lumbei; is commonly resorted to in order to prevent checking and case hardening, but not, as has frequently been asserted, to enable the board to dry. Yard-dried lumber is not dry, and its moisture is too * Bulletin 70, U. S. Forest Service, p. 123, t Roth, Bulletin 10. 628 MATERIALS OF CONSTRUCTION unevenly distributed to insure good behavior after manufacture. Care- ful piling of the lumber, both in the yard and kili^ is essential to good drying. Pihng boards on edge or standing them on end is believed to hasten drying. This is true only because in either case the air can circulate more freely around them than when they are piled in the ordi- nary way. Boards on end dry unequally — the upper half dries much faster than the lower half and horizontal piUng is, therefore, preferable. Since the proportion of sapwood and heartwood varies with size, age, species, and individual, the following figures must be regarded as mere approximations: POUNDS OF WATER LOST IN DRYING 100 POUNDS OF GREEN WOOD IN THE KILN (1) Pines, cedars, spruces, and firs (2) Cypress, extremely variable (3) Poplar, Cottonwood, basswood (4) Oak, beech, ash, elm, maple, birch, hickory, chestnut, walnut and sycamore Heartwood 16-25 18-60 40-60 30-40 The lighter kinds have the most water in the sapwood, thus sycamore has more than hiclcory. 573. Shrinkage, Warping and Checking in Drjring. The shrinkage of wood ill drying is due solely to the loss of moisture from the walls of the cells. Variation in the water content of the lumina of the lifeless cells, and in the proto- plasmic contents of the li-ving cells, does not affect the volume of the wood in any way. As moisture is evaporated, the cell walls be- come thinner; the lumina become larger, and the exterior cross-sectional dimensions become smaller. The contraction of a wood element in a longi- tudinal direction is scarcely appreciable, however. The total volumetric change of an element is roughly proportional to the original thickness of its walls. The aggregate volumetric change of a mass of thin-walled wood elements is therefore much less than that of a mass of thick-walled elements. Fig. 335. This explains the fact that summer wood almost invariably shrinks more than spring wood, its wood elements being thicker walled, and dense wood shrinks more than lighter wood of the same species for the same reason. Fig. 335. — Shrinlcage of thick-walled and thin- walled fibers. (Bull. 10, U. S. For. Div.) TIMBEE 629 Irregularities in the structure of all wood prevents shrinkage being uniform throughout. The rays, whose elements are for the most part arranged at right angles to the main wood elements, oppose shrink- age in a radial direction and tend to cause longitudinal shrinkage. The bands of dense summer wood are continuous in a tangential direction, but in a radial direction they alternate with less dense spring wood. On this account tangential shrinkage commonly amounts to at least twice the radial shrinkage. Longitudinal shrinkage will be greatest in woods having an abundance of large rays or those exhibiting wavy or spiral grain (the lack of parallelism of the wood elements with the pith in this case being accountable for the existence of a component of the trans- verse cell shrinkage in a direction parallel to the pith), but in any event the longitudinal shrinkage amounts to only a few tenths of 1 per cent. The following table gives the results of shrinkage tests made at the Yale Forest School by Mr. H. P. Baker and quoted by Record.* The values given represent the average percentage shrinkage (original dimen- sions taken as a basis) in reducing green wood to a kiln-dry condition. Wood. Length, Per cent. Radius, Per cent. Circumference, Per cent. Area of Cross-section, Per cent. Volume, Per cent. Red juniper Butternut 0.32 .36 .25 .24 .15 .10 .04 2.7 2.9 3.0 3.7 4.3 6.1 7.4 5.6 6.9 4.9 8.2 9.3 11.5 9.2 6.9 7.3 11.2 10.4 12,6 17.1 19.4 5.9 7.6 Red oak Tulip tree Black gum White hickory .... 13.7 18.0 19.8 The warping of lumber is due either to unequal drying of different portions, or to unequal shrinkage on account of irregularities in structure. Any straight-grained green board exposed on one side only to air and heat will become concave on the exposed side because of the more rapid dry- ing and consequent shrinkage of that side. Boards cut tangentially from the log tend to become convex on the side toward the pith when dried, because of the greater shrinkage of the wood in a direction parallel to the annual growth rings. (Each growth ring tends to shorten, thus causing the edges of the board to curl away from the pith as shown by Figs. 336 and 337. If lumber is cross-grained, the component of the shrinkage m a longitudinal direction causes a warping lengthwise as well as in the transverse direction, and where the grain is spiral boards may become badly twisted. Warping is always more apt to be encountered * " Economic Woods of the United States," p. 57. 630 MATERIALS OF CONSTRUCTION in the lumber from woods of great irregularity. Serious deformation of such lumber can be avoided only by most careful handling in drying. Checking of timber in drying is a consequence of the inability of the wood to accommodate strains consequent upon unequal shrinkage. A great many small checks occur, particularly in the ends of timbers, owing to the more rapid drying from the cross-section and the consequent excess of shrinkage of the end portion over that of the balance of the timber. This results in a tendency to bend the fibers. Fig. 338, and their stiffness may be sufficient to cause rupture between strands, thus reliev- ing the stress. Similar checks occur on the sides of logs and timbers because of the precedence of the shrinkage of the outer portion over that of the inner portion, which has scarcely begun to lose its moisture. Both of these classes of seasoning checks are considered temporary, because they close up and become imperceptible as the inner portion of the timber dries and shrinks. They are still there, whether visible or not, however, and always impair the structural quahties of the wood in some measure. Another class of checks, more important than the temporary checks, because they are apt to be larger and are permanent, are those caused principally by the greater shrinkage of timber Fig. 336.— Shrinkage of in a tangential direction along the rings than that Thick- and Thin- along the radius. The occurrence of the rays in walled Fibers. Warp- radial planes often contributes to the formation of f,^ TT^ J^T^' o ^^n""' these large radial checks, because they form a plane 10, U. S. For. Ser.) . , , . , , , , . of weakness at the very point where the strains are the greatest and most complex (two severe stresses existing at right angles to each other, owing to the shrinkage of the rays opposing the shrinkage of the main wood fibers). The danger of the occurrence of large checks of this nature constitutes a serious difficulty in seasoning large timbers, and especially round timbers such as poles, piles, and posts. Too rapid seasoning always increases the danger of injury by excessive checking. ■ Some woods, mostly hard woods, become " case-hardened " when rapidly dried in the kiln; that is, the outer part dries and shrinks, and commonly checks, while the interior is still practically in its original condition. The drying of the interior is thus retarded, but when it does occur great internal strains are set up, resulting in the formation of large or numerous radial checks which follow the rays. When these checks are comparatively small, but numerous, the wood is said to be TIMBER 631 "honeycombed." The case-hardening of timber may be avoided by air seasoning before placing it in the kiln or by occasionally admitting steam to the kiln. Wood, when dried, has the ability to reabsorb water from the atmos- phere. This property is termed " hygroscopicity." The amount of hygroscopic water thus acquired always exceeds the moisture content A / m 1 2 E I i 'TiII I Fig. „„.. . (BiiU. 10, .^^jts of Shrinkage. U. S. For. Div.) Fig. 338.— Formation of Checks. (BuU. 10, U. S. For. Div.) of the air, but varies with the humidity. The consequent shrinking and swelling of the wood is a serious hindrance to its use where exact fitting is desired. The hygroscopicity of wood may be reduced, but not eliminated, by prolonged exposure to temperatures in the neighbor- hood of 100° C, or by boiling, steaming, or prolonged soaking. The harmful effect of the shrinking or swelling of wood used in inte- rior finishing may be minimized on the one hand by avoiding large con- tinuous surfaces by paneling, etc., or, on the other hand, by rendering the wood more or less impervious to moisture by the use of coatings of oils, paints, varnishes, etc. 632 MATERIALS OF CONSTRUCTION MECHANICAL PROPERTIES OF WOODS 574. General. The intelligent use of wood for any structural pur- pose requires a general knowledge of the mechanical properties of dif- ferent woods, in order that the one selected may conform in its structural qualities to the requirements imposed, and in order that a given pur- pose may be served at a minimum expense. Wood is not like many other structural materials in that its mechan- ical properties are extremely variable, not only between different species and different trees of the same species, but also between specimens cut from different portions of the same tree. An assumption as to the mechanical properties of certain timber of a given variety can therefore be predicated upon the results obtained in tests of timber of that variety only in a most general way, unless detailed information concerning the many factors governing the mechanical properties of both test timber and commercial timber are known. Some of these important factors are: (1) correct identification of species and variety, (2) age and rate of growth of trees, (3) position of test specimens in the tree, (4) moist- ure content, and (5) relative freedom of test specimens and commercial timber from defects such as knots, checks, etc. Even if all this informa- tion is available the conclusion reached concerning the probable strength is only approximate, because the exact weight of each of the above fac- tors has not been definitely established. We have, however, fairly definite information concerning the allowance to be made for variation in the moisture content and rate of growth. The variation in test results for which these factors must be considered principally accountable is indicated by the following inspection of a series of compressive and bending tests of 32 species of American woods, made Proportion of Tests within 10 Per cent Proportion of Tests within 25 Per cent ot Average. of Average. Minimum Maximum Average Minimum Maximum Average Proportion Proportion Proportion Proportion Proportion Proportion (One Species) ^ S.2 c= £ S S £t5£ ill IP g.2 d ai > '^ £ii S.2 H m S 2 a ■g^s 3 e M 2 5i ^ IhS =* C M 3 ffl PM uptu Ben Per 2 C M o rt o rt o rt o rt o rt o Pi 28 20 79 64 56 38 65 58 100 95 93 76 In the above summary variability of results due to differences in moisture content has been eliminated by correction to a constant moisture content. TIMBER 633 by the U. S. Forestry Division.* The number of tests of each kind made upon each species varied from 10 to more than 1200, and averaged about 180. 575. Tensile Strength. The tensile strength of timber is not an important property except in so far as the tensile strength is involved in all cases of transverse loading. Timber used in construction is prac- tically never subjected to pure tensile stresses for the simple reason that the end connections cannot be so devised that they do not involve either shear along the grain or compression across the grain, and, since the resistance offered by any timber to compression across the grain, or shearing stress along the grain, never amounts to more than a small fractional part of the tensile strength, it is evident that considerations of economy in the design of structures will call for the use of iron or steel instead of timber for those members which must withstand tension. Tensile tests of woods are seldom made for the reason above indicated. When they are made, the specimens must have a reduced section, the area of which is not more than about one-tenth the area of the ends which are gripped by the jaws of the testing machine or pierced by the cross-bolt of a shackle through which the load is applied. Failure in tension across the grain involves principally the resistance offered by the thinner-walled wood elements to being torn apart longi- tudinally. The thin-walled parenchymatous cells and vessels are most often the determining factor. The thick-walled wood fibers are not torn, but are pulled apart. Wood sustains in tension across the grain only a small fraction (one-tenth to one-twentieth, perhaps) of the load carried in tension along the grain. The tensile strength of a number of varieties of timber woods in a direction perpendicular to the grain (tested in a green condition) are hsted in Table I at the end of this chapter. Failure in tension along the grain involves principally the resist- ance offered by the wood elements to being torn apart transversely or obhquely. The strands of wood elements are practically never pulled apart by failure of the union between adjacent strands or fibers. The nature of the prevailing wood elements is a factor in so far as it affects the actual area of cell walls encountered in a cross-section. A more important factor, however, is the structure of the wood as a whole, the tensile strength being especially dependent upon the arrangement of the wood- elements. Cross grain is prejudicial to tensile strength, and rays, owing to their transverse position with respect to a load apphed along the grain and their small resistance to tension in a direction normal to the direction of their fibers, greatly weaken the timber. Their disturb- * Circular 15. 634 MATERIALS OF CONSTRUCTION ance of the normal regularity of arrangement of the main wood fibers is also injurious. Knots similarly weaken wood subjected to longitudinal tension because of the likelihood of their being either loosely connected with the adjoining wood substance, badly checked, or structurally im- paired by a pith cavity. On account of the considerations mentioned above, little experimental data are available concerning the strength of wood in longitudinal ten- sion. The only importance of tensile strength in the direction of the grain arises from the circumstance that the lower side of a beam subjected to transverse load is obliged to withstand tensile stresses. This case will later be discussed under the head of cross-breaking strength and stiffness. 576. Compressive Strength. The compressive strength of wood in a direction normal to the grain is simply a matter of the resistance offered by the wood elements to being crushed or flattened. The cells with the thinnest walls collapse first and the action proceeds gradually, the so-called " elastic limit " being usually fairly well marked by the depart- ure of the load-deformation curve from a straight line. A load which cannot be exceeded will be finally reached (the specimen continuing to distort without increase of load), but there is no such thing as a breaking load in transverse compression of wood. The endwise compressive strength of wood, i.e., the strength of wood in compression along the grain, is dependent upon the anatomical struc- ture and the moisture content of the wood, and the manner of failure is fixed by these same factors. The individual fibers (or other elements) of wood act as so many hollow columns bound firmly together, and failure involves either buckhng or bending of the individual fibers or bundles of elements which finally come to act almost independently. Buckling is characteristic of any thin-walled cells when dry, as in the case of the early wood of many species, the vessels of others, and the wood of seasoned white pine, spruce, etc., as a whole. Bending is typical of all green or wet woods, and of those portions of any wood which are made up principally of thick-walled cells, whether wet or dry. The isolation of bundles of elements is very frequently caused by longitudinal splitting, which occurs by a tearing of the fibers of the wood, including those of the rays. The presence of moisture in wood decreases the stiffness of the cell walls, and lowers the compressive strength very materially in consequence. Moisture also facilitates the separation of the fibers from one another. From the above considerations it will be apparent that the principal factors affecting the compressive strength along the grain are: (1) the actual amount of wood substance encountered in a section, i.e., the density of the wood, (2) the strength of the union between the wood TIMBER 635 fibers or longitudinal strands of wood elements (particularly as affected by contained moisture), (3) the stiffness of the wood fibers or longitudinal strands of elements (again largely a matter of moisture content), and (4) the continuity of the course of the longitudinal strands in a direction parallel to the axis of the piece. Woods in which the separate elements are closely interlaced and bound together, without the existence of dis- tinct longitudinal planes of weakness, such as exist in ring-porous woods, for instance, will be stronger than woods of opposite character; well- seasoned wood will be stronger than green woods, the stiffness of which is decreased by a high moisture content; and straight-grained woods will be stronger than cross-grained or knotty woods, or woods in which rays are large and abundant. Thus we find that the strongest of our common woods in compres- sion with the grain are the dense and tough hickories, birches, hard maple, locust, etc.; oaks, long-leaf pine, elm, ash, etc., rank next; short- leaf pines, western hemlock, beech, cedar, cypress, etc., are placed in a third division; while basswood, butternut, chestnut, hemlock, spruce, soft pines, fir, etc., are ranked among our weakest woods. Tabulations of the mechanical properties of some of the principal lumber woods of the United States will be found at the end of this chapter. The resistance offered to crushing across the grain is dependent prac- tically entirely upon the density of the wood, since failure can occur only by flattening of the wood elements as their lumina are closed up. The crushing strength across the grain is therefore least for the lightest, most porous woods, and greatest for the heaviest and densest woods. An inspection of the data given in' Table II reveals the fact that the rela- tion of compressive strength across the grain to compressive strength along the grain is represented by the following factors: viz., 13 to 14 per cent for white pine, cedar, cypress, and spruce, 15 to 16 per cent for the various grades of hard pine, 18 to 26 per cent for elms, 21 to 26 per cent for ash, 22 to 26 per cent for oaks, and 23 to 31 per cent for hickories. 577. Cross-breaking Strength and Stiffness. The cross-breaking strength of any material is necessarily closely related to the tensile and compressive strength of that material, since the stresses encountered in a specimen loaded transversely are principally tensile stresses in the lower portion of a beam and compressive stresses in the upper portion. There also exist certain shearing stresses, consequent upon flexure, which are relatively so small that they may be neglected in considering materials whose shearing strength is a large fraction of the compressive strength. With timber, however, the strength in shearing along the grain is so very small that these stresses cannot be wholly neglected, and it is often ob- served in tests of timber that the initial failure is by longitudinal shear. 636 MATERIALS OF CONSTEUCTION It has been shown above that the tensile strength of all timbers is greatly in excess of its compressive strength (about three times as much on the average), and the latter factor will therefore usually be the deter- mining factor in limiting the cross-breaking strength. (Compressive strength will always be the determining factor, assuming there exist no defects such as knots, uneven grain, etc., on the under side of the beam, if the load is not a single concentrated load, but is applied at several points or distributed along the beam so that the shearing stressrs are reduced to a minimum.) It is apparent, therefore, that the considerations which fix the compressive strength similarly affect the cross-breaking strength, and, in fact, there is no reason why the compressive stress in the uppermost fiber of a beam at time of failure . should not be pre- cisely the same as the ultimate compressive strength of the wood. The computed value of the fiber stress under the breaking load always greatly exceeds the ultimate compressive strength, however, because this apparent stress exceeds the actual stress. (We compute the extreme fiber stress at rupture, i.e., the modulus of rupture, by the rule f=My/I, which involves the assumption that the material is still behaving elas- tically up to actual failure and, therefore, that the fiber stress is still directly proportional to the distance of the fiber from the neutral axis. As a matter of fact, the elastic hmit of the material has long since been passed when the breaking load is reached, the neutral axis has shifted, and the extreme fiber stress is no longer proportional to the bending moment.) Although the compressive. strength is always the determining factor which limits the transverse strength of wood, it is often only the initial failure which occurs in compression. The fibers on the compression side are caused to buckle or fold (slightly, at least), with the result that the neutral axis becomes shifted and the portion of the area of the sec- tion which is subjected to compression probably becomes relatively so much greater than the area subjected to tension that the intensity of tensile stress is disproportionately increased and ultimate failure in tension follows.* The stiffness of timber used structurally as beams is often quite as important as its cross-breaking strength. In plastered ceilings, for in- stance, the maximum deflection permitted is usually limited to 1/360 of the span, so that timber beams must be so designed as to not only carry the load imposed safely, but also to do so without excessive deflection. The stiffness of timber, as in the case of any material which behaves * This theory of transverse failure is attributed to Prof. Suenson of Copenhagen and has received some experimental verification. See Proc. Sixth Congress Internat Assoc, for Test. Matrls., 1912, XXIIIa, pp. 12-17. TIMBER 637 elastically, is most conveniently expressed by the modulus of elasticity, which quantity is computed from the expression: wherein P is a central concentrated load on a simply supported beam of span I; dm the deflection due to P; 6 is the width; and h is the height of the beam. Comparing the above expression with the similar expres- sion for the modulus of rupture, i.e., .^ (3w0 •^ (26/i2)' it will be noted that, whereas the strength is proportional to the span, the deflection is proportional to the cube of the span, and, whereas the strength is inversely proportional to the square of the depth of the beam, the deflection is proportional inversely to the cube of the depth. It is evident from this comparison that the stiffness rather than the strength will often be the governing factor in selecting a timber beam for a given situation. Stiffness of timber is largely dependent upon the same factors as strength. Dense woods are always stiffer than open, porous woods, and heavy woods are stiffer than light woods except in so far as the weight is attributable to moisture contained. Thus the wood of the long-leaf pine is stiffer than than of the white pine, and the wood of the hickory is stiffer than that of the chestnut. All woods are stiffer when well seasoned than when green, and a given timber of a variety showing distinct bands of summer wood will be stiffer if so placed that the general direction of the annual rings in its cross-section is vertical rather than horizontal. (See Tables I and II.) 578. Moisture and Strength. The important effect of moisture upon the strength and stiffness of woods has been often mentioned in the preceding discussion, but no effort has hitherto been made to show exactly what difference in strength and stiffness of a given timber is caused by a given change in moisture content. All woods gain both in strength and in stiffness when thoroughly air seasoned or kiln dried. The actual net gain which, from a compar- ison of strength and stiffness of small specimens in the green and oven- dry condition, would appear to be attainable, cannot be even approxi- mately reaUzed in practice, however, because of the operation of several factors which greatly modify the effect of lessened moisture. Checking, for instance, always occurs to some extent in drying lumber and will 638 MATEEIALS OF CONSTRUCTION partially or entirely counterbalance the gain due to drying the extent of this effect being dependent upon the size and the variety of the timber. The general character of the effect of various moisture contents upon the 45000 40000 .01 .02 .03 .04 .05 .06 .07 .08 .09 .10 .11,, .12 .13 .14 .15 .16 .17 .18 Compression-In the Piece 5?^ long -Inches Fig. 339. — Typical Stress-strain Diagrams for Three Species of Wood, Showing the Eilect of Different Degrees of Moisture upon Strength and Elastic Properties. Tested in Compres sion Parallel to the Grain. (Cir. 108, U. S. For. Ser.) strength and stiffness of three varieties of timber woods is shown by Fig. 339.* * U. S. Forest Service, Circular 108. TIMBER 639 Large timbers dried only by air-seasoning, even though the process is prolonged for several months or even years, seldom lose sufficient moisture to benefit their strength to more than a sUght degree. Such timbers, therefore, cannot be safely depended upon to show any greater strength than if they were in the original green condition. The explana- tion of this fact is that a great part of the moisture which is first evapo- i / y- / 18 16 ^ y / r- 14 12 ■r* ^ .^^ t 10 >f ^^ ^ B •V 1 ^ p^ ^^ 0^6 II i * 1 2 &« o |2 £ ^ ^ c^ ^ y i^ ^»V ^ V' / ° «>'> ^ X /I > 0- ^ -^ r " ^ ^ f, *«t:s r^ Ir- • ^ -^ ?,*1i i^ r ^ y ^ ^ ^ sy^ ^ ^ r -»^ ^ ^ c^ ^ o " 1 6 i 2 n ^ y ..^»' -5^ O ^ ^ • ^ !^ ^ ^ ,>. ^ 10 15 20 25 30 35 40 Molstare-Per cent of Dry Weight = M x 100 P^G. 340. — Relation between Swelling and Moisture Content. (Black Dots indicate Specimens that were Kiln-dried and then Allowed to Reabsorb Moisture.) (Cir. 108, U. S. For. Ser.) rated from wood is water which exists only as " free water " in the lumina of vessels and cells, whereas only variation in the moisture content of the walls of the wood elements affects strength in any way. " The degree of moisture at which maximum absorption by the cell walls is reached is called the ' fiber-saturation point ' of the wood. After this point is reached 640 MATERIALS OF CONSTRUCTION added moisture does not lessen ihe strength of wood. At this point also, wood ceases to swell," Fig. 340. The fiber-saturation point is determined experimentally by tests of the strength of very small specimens covering a large range of moisture content. When further moisture no longer lessens the strength the fiber- saturation point is reached. Tiemann found this point for various woods to be within the following limits: * Per cent. Long-leaf pine ' 24-26 Red spruce 29-35 Chestnut 24-28 Loblolly pine (heartwood) 23-24 Loblolly pine (sapwood) 24^26 White ash 19-23 Red gum 25 Douglas fir 23 Norway pine (heartwood) 27-34 Norway pine (sapwood) 26-30 Tamarack 27-33 (See Fig. 340.) Experimental determinations of the strength and stiffness of wood as influenced by moisture have been made in the laboratories of the U. S. Forest Service and are reported in Bulletin 70 and Circular 108. While these investigations have not covered all of our important timber woods;^ much important information has been derived from them. The woods studied were long-leaf pine, red spruce, chestnut, loblolly pine, Douglas fir, tamarack, and Norway pine. In the cases of the first three of the above varieties the study covered the effect of moisture upon compres- sive strength along the grain, strength (modulus of rupture) in bending, and strength in shearing along the grain, also stiffness (modulus of elas- ticity) in compression and in bending. For the remaining varieties only the effect of moisture upon compressive strength along the grain has been studied. From the facts observed in this series of tests the following conclu- sions may be drawn: (1) Compressive Strength Parallel to the Grain. (a) Loss of moisture does not affect strength. Fig. 341, or stiff- ness, Fig. 342, in any way until the total moisture content has been * U. S. Forest Service, Circular 108. TIMBER 641 reduced below the critical percentage which represents the fiber- saturation point. (b) The fiber-saturation point having been passed, the rate of gain in strength and stiffness varies considerably for different woods, always becoming greater as the total moisture decreases. 16 000 \^ ,14 000 d M t !> \1. |12 000 ;t-M li^iV- i g 10 000 ^ ^^ ' v^^ V \V =^. s ^ i.%V ^^^t A 1 UI 6000 a s \ &^ ^^t;^ >■ \ s ^ t s ^: s S^o t- ^ $<. ^ k> '. / Pibr 3-Sa iirat on] oint I O 4000 ^ k **^ ^ A ^l^ — v j^ "^ss ^Or> 's" ~~- __ "^ -- - 2000 ~ ' ~ 10 15 20 25 30 35 40 MoisturesPer cent fiaaed on Dry Weight Fig. 341.— Relation between Strength in Compression Parallel to the Grain and Moisture Content for Several Woods. (Cir. 108, U. S. For. Ser.) (c) The green strength of the woods tested amounted, as a rule, to from 50 to 60 per cent of the strength in the normal air-dry condition (12 per cent moisture); and the strength of the kiln- dry wood (3.5 per cent moisture), exceeded the strength of the air-dry woods by from 50 to 70 per cent. The actual numerical relations found in the tests were as follows: 642 MATERIALS OF CONSTRUCTION Wood. Ratio of Strength, Green, to Strength Air-dry. Ratio of Strength, Kiln-dry to Strength Air-dry. Long-leaf pine .58 .41 .55 .60 .59 .59 .51 .49 1.67 Red spruce 1.53 Chestnut 1.55 Loblolly pine (heartwood) . . Loblolly pine (sapwood).. . . Douglas fir 1.56 1.78 1.55 Tamarack 1.63 Norway pine 1.66 .§1600 1 \ \ \ ?; \ f? \l \ \ \ \ \ \ \ s, ^ V Pibi »-Sa urat on^ oint \ \ s s 4 ^ K \ S. X ^ \ --^ \ ^ "N, \ ^ 10 20 30 40 60 60 Moisture— Per cent of Dry Weight Fig. 342. — Variation of Stiffness in Com- pression Parallel to the Grain with Moist- ure Content. (Cir. 108, U. S. For. Ser.) 23000 1 I 18 000 \ M £16000 \ \ \ c* Cfi \ A -a a §12 000 \ \ ^ \ % <3 \ £ "aioooo '^X \ . \ Lo iKle« f Pi 'P iulus of i\ \ \ \ 's ^ 6000 \ CI eetn It \ S )mc 400O ■^ 10 20 30 40 50 Molsture-Per cent of Dry Weight FlG. 343. — Variation of Strength in Bend- ing with Moisture Content. (Bull. 70, U. S. For. Ser.) (2) Strength in Bending. Modulus of Rupture. (a) Same as (a) above, Fig. 343. (6) Same as (6) above. (c) The modulus of rupture of green woods of the varieties tested amounted to from 50 to 70 per cent of the value found for the same woods when air-dry, and the modulus of rupture of kiln-dry woods TIMBER 643 exceeded the value found for the same woods when air-dry by from 33 to 65 per cent. The numerical relations established were as follows: Wood. Ratio of Mod. of Rupt., Green, to M. of R., Air-dry. Ratio of Mod. of Rupt., Kiln-dry, to M. of R., Air-dry. Long-leaf pine .67 .52 .65 1.65 1.45 1.33 Red spruce Chestnut (3) Stiffness in Bending. Modulus of Elasticity. (a) Same as (a) above, Fig. 344. (6) Same as (6) above. Effect of moisture is not nearly so marked in case of stiffness as in case of either compressive or bend- ing strength. (c) The modulus of elasticity of green woods of the varieties tested amounted to from 80 to 84 per cent of the value found for the same wood when air-dry, and the modulus of elasticity of kiln- dry woods exceeded the value found for the same woods when air- dry by from 14 to 34 per cent. The numerical relations established were as follows: Wood. Ratio of E, Green, to E, Air-dry. Ratio of E, Kiln-dry, to E, Air-dry. Long-leaf pine .84 .80 .84 1 34 Bed spruce 1.14 Chestnut 1.21 579. Weight and Strength. It has long been known that the strength of a given species of wood is directly dependent upon the dry weight (specific gravity). Tests of the U. S. Forest Service made in 1896 * indicated further that this law of variation holds good not only for a given species, but irrespective of species for the four principal pines of our Southern States. In 1897, further tests f indicated that the appli- cation of the law might be extended further, and it was cautiously stated as a probability that " in woods of uniform structure strength increases with specific weight, independently of species and genus distinction, i.e., other things being equal, the heavier wood is the stronger." The relation indicated between weight and compressive strength parallel * Circular 12. ' t Circular 15. 644 MATERIALS OF CONSTRUCTION with the grain was expressed by a single rule which agreed approxi- mately with the observed results in tests of most classes of timber, except the oaks, which apparently follow a somewhat different law, the strength corresponding to a given weight being considerably lower than in case of other woods. 2800 \ \ \ ■g % a g 2100 \ r<3 a \ \ g 2000 P4 \^ V N \ \ \ ►> s N X ^ \, \ N**, ■"> "^ ^ m s \ ^ [^ f^'r 1400 s N ' N "~~~ --^ 1200 \ ~^ ^*'". ^ ■ — -. ^ ^ ^ ^ 1000 ■■ ■— -_, "~- 10 16 20 25 30 35 Molsture-Per cent of Dry Weight Fig. 344. — Relation between Stiffness in Bending (Modulus of Elasticity) and Moisture Content, for Three Species. (Cir. 108, U. S. For. Ser.) More extensive tests, carried out later in a more careful manner, have tended to qualify the original general conclusions, and it has been repeatedly brought out in the analysis of tests of the Forest Service that, -while the mechanical properties of wood vary directly with its dry weight for a given species, the relation is far from being constant in all species. Bulletin 108 of the Forest Service gives detailed information on the TIMBER 645 subject so far as most of the conifers are concerned. Reliable informa- tion is not yet obtainable concerning the broad-leaved woods. The curves of Fig. 345 express the relation between dry weight and bending strength (modulus of rupture) and stiffness (modulus of elas- ticity) of small clear beams. 11000 10000 9000 8000 7000 6000 5000 4000 3000 2000 1000 20 25 30 35 iO Pounds per Cubio.Foot (Dry) Fig. 345a.— Relation of Strength and Stiffness of Conifers in Bending to Dry Weight. (Bull. 108, U. S. For. Ser.) Since the position of a specimen in the log (height or position in the section), influences the density of the wood, the influence of this factor is covered in the above discussion. 580. Rate of Growth, Proportion of Summer Wood, and Strength. The average rate of growth of timber is readily computed by counting the annual rings along a radial line and dividing by the length of the line. The relation between rate of growth and bending strength for the tim- bers considered in the preceding article was also investigated and 646 MATERIALS OF CONSTRUCTION reported upon in Bulletin 108 of the Forest Service. The indications from these studies were that for most species there is a rate of growth 10 25 30 1 / Red FOOC 11000 in Lbs. per Sq. In. Fibre Stress at B.L. in Lbs. per Sq. In. / 1 / 10000 / / f Mnrli ilus of Elasticity / in in ^ We iteri HeJnloc i / / / / it-T ima: ack / / f y^ Red WOOi I / / / / /' / ^A y^ -No ■waj Pin c / /' V 50CO > / V /^i ^ost srn leni ocli / / J ''/ / <«-T UllU -jcU 4000 / /, / / / 4 / / A Noj (vay Plni 3000 / y ^ -^ 2000 ' ,-^ cste •nH smlcj clt I ^ Rod svoo 1 1000 p^ —-^ M '^•\ amj racb -— ■ "" - ^^ Noi ivay Pint Pounds per Cubic Foot (Dry) 35 40 Fig. 345b. — Relation of Strength and Stiffness of Conifers in Bending to Dry Weight. (Bull. 108, U. S. For. Ser.) which, in a very general way, is associated with the greatest strength. For the species tested this appeared to be as follows: Rings per inch. Douglas fir 24 Short-leaf pine 12 Loblolly pine g Western hemlock jg Tamarack. 20 Norwajr pine l^g 'Redwood 3q It was found, however, that the variation between different indi- viduals of a species was very great, so that, although an effort was made TIMBEE 647 to express the average relation between rate of growth and strength for each species, no great rehance could be placed upon the average rela- tion so determined, and the conclusion was stated " that rings per inch are not a reliable index to the mechanical properties of timber, especially structural timbers containing knots and other defects." The proportion of summer wood in the section of a timber of any species is closely related to the dry weight. It is therefore to be expected that the relation between the proportion of summer wood and strength should agree closely with the relation between dry weight and strength. The studies of the Forest Service show well the identity of these two relations as factors influencing strength, and the proportion of summer wood is therefore an important guide in judging the quality of timber. 581. The Time Factor in Tests of Timber. Timber differs from most other materials in that small variations in the rate of application of load have a more pronounced effect upon the strength and stiffness shown by a specimen under test. If a timber-compression block or beam is loaded rapidly, it will appear to have a higher elastic limit and ultimate strength, and will also appear to be stiffer, than it will if it is loaded less rapidly. This behavior is due to the fact that the deformation lags far behind the load, and if any load is permitted to remain upon a specimen for a sensible time interval the deformation increases, the amount of increase becoming greater for heavier loads. Actual failure appears to be consequent upon the attainment of a certain limiting ajnount of deformation or strain, rather than a limiting load or stress. This con- dition of affairs makes it necessary to standardize tests by adopting certain speeds of loading for each class of specimens, or, rather, to so proportion the speed of the moving head of the testing machine to the dimensions of the specimen that the resultant rate of fiber strain will be a certain specified amount. The usual practice is to adhere to the standards established by the Forest Service for the use of engineers of timber tests. These standards are as follows : * Character of Tests. Character of Specimens. Rate of Fiber Strain in Ins. per Minute. Bending tests ". Timber of structural sizes Small test specimens 0.0007 0.0015 Compression with grain Compression with grain Compression across grain Compression across grain. . . . Timber of structural sizes Small test specimens. 0.0015 0.0030 Timber of structural sizes Small test specimens 0.0070 0.0150 * " Instructions to Engineers of Timber Tests," Cir. 38 (revised), U. S. Forest Service. 648 MATERIALS OF CONSTRUCTION When constant loads amounting to a large fraction of the ultimate strength of timber are sustained for very long periods, the deformation may continue to increase until rupture occurs, even though the stress encountered is far below the ultimate strength of the timber as originally determined. A number of series of long-time tests have shown that for timbers loaded either transversely or in compression with loads amount- ing to from 50 to 60 per cent of the ultimate strength usually found, fail- ure will ultimately occur, the time required varying from a few weeks to several months. This important conclusion may therefore be drawn: "The strength of timber under any kind of permanent load is only about one-half its strength as found by actual (short-time) tests." * 582. Tabulation of Mechanical Properties of Structural Timbers. The following tabulation of the mechanical properties of American tim- ber has been compiled from the tests of the U. S. Forest Service. Table I is a summary of tests of forty-nine species of wood tested in a green condition (Circular 213), and Table II is a summary of tests of thirty- two species of wood, all values being obtained with — or reduced to — 12 per cent moisture (Circular 15). Small clear-test specimens were used in all of the tests. ' 583. Factor of Safety and Working Stresses. The factor of safety used in the design of timber structures is a very variable quantity. Merriman t gives 8 for steady stress, 10 for varjdng stress, and 15 for shocks; Rankin { gives 4 to 5 for dead load and 5 to 10 for live load; and the Cambria Steel Co. Handbook gives 10 for tension, 6 for extreme fiber stress in bending, 5 in compression along the grain, and 4 for com- pression across the grain and for shear. The variabiUty of the factors given by different authorities indicates that the factor of safety for tim- beriis in a large measure an expression of ignorance or lack of confidence in the reliability of values of strength upon which designing is based. As an example of the best current practice the following table of safe working stresses recommended in 1909 by the Committee on Wooden Bridges and Trestles of the American Railway Engineering Association is given. In comparing these figures with those given above, account must be taken of the fact that the present figures apply to green timbers (all large timbers must always be assumed to be no stronger than green timber) of structural sizes used for railway bridges and trestles where always subjected to vibration and shock. The conmiittee recommends that these values be increased 25 per cent for highway bridges, and 50 * Johnson, " Materials of Construction," p. 4686. t " Mechanics of Materials." t Civil Engineer's Handbook. ItKSlILTS OF TESTS ON 40 SI»KC1KS OF WOOD TESTED IN A GREEN CONDITION IN THE FORM OF SMALL CLEAR PIECE (U. S. Forrst Service Circuliir 213.) [TcHt Hpr'rirnrnn nro 2 by 2 inrhcH in siTtion. B«n(iinK ipnrimoni opo cut 30 inchoA long; others are ■hort«r, depending on kind of teat.) Species: Common Nam« ami Uutauical Namv. IIardwoodb. Ash: Black (Frftxinua nigra) Whitft (KraxiiiUH Anu^rirana). Uu . Basawood (Tilia Americana) Beech (KaguM atrupuiiirra) Birch, yellow (liptula lutoa) Elm: Rock *1hppcry (LMmuii puboHccna). White (Ulinua Anicricann). l.ocultty w)i<*rc gr^iwti Marathon Cnunty. Wis Stone County. Ark Hichliinil riinah, La Marntlion ("imnty. Wis HcnilrickH and Morgan (\>unti<'i), Intl Marathon (\iunty, Win Gum, red (l.ii4uidanibur atyraciHua) Hackberry (Oltia uccidcntalin) Hickory: Big abellbark (Hiooria larinioaa) . Do Bitteruut (Hicoria ntiniinn) Mockcruut (Hiooria alba) Uo Do Nutmvg (Hicoria myriaticvfurmia) Pignut (Hicoria glabra) Do Do Do Shagbark (Hiooria ovata) Do Do Do Water (Hicoria aquatica) Locuat. honpy (Oleditaia triacanthos) Maple: Hi'd (Acer rubrum) iSugar (Acer naccharuni) Do Poat itjuercus minor) H»?d (Uuercua rubra) Do Do Hwamp white (Quereua platanoidea). . I'anbtirk (Quercua densiflora) White (Querrua alba) Do Oak: Do. Yellow ((juerruH velutina) . Do Osage orange (Toxylhio .... do Sardia, Miaa Cheater County, Pa Webdter County, W. Va SardiN. Mma . . . tlo Napoleon, t >hio Cheater County, Pa Webnter County. W. Va Sardin, Minn Napoleon. Ohio Chester Cunty, Ind WilhtH. Cul Stone County. Ark Hentlrickn, Marion, ami Morgan CountieH, Ind Richland Parish, l.a .Stone Ciiunty. Ark Marathon County, Wis Morgan Ct»unty. Ind llenurickH County. Ind St. John the HaptiHt PariHh, La Shawano County, Wis Weed, Cal St. John the Baptist Pariah, La. Cjrand County, Colo Dee. Oregon Johnaon County. Wyo Madera County. Cal Marathon County. Wia (irand County. (. olo Jcdinaon County. Wyo Tangipahoa Pariah. La Shawano County, Wia Malvern. Ark Madero t^ounty. Cul Coconino C^ounty, Arii Madera County, t^al Douglaa County. Colo Shawano County, Wia llumbohlt County, Cal Mendocino County, Cal (irand County. Colo San Miguel County, Colo Cooa County, N.li. do. Marathon and Shawano Counties, Wia. 1 lu V II K II I r> 4 10 27 lU 4 U 10 2 Pro- tion of ■uni- mer wood 2.1 (1 14 N M 1) \i:t 2 III II lit 7 :iii 1) H 4 lU ■>-i :i Pit con I. .11 (Ml :to 2N OH n :i 70 2:i i» (JO l:i « 71 117 71 lu :i 4,1 lU U9 ;ii ,1 01 22 1 5N 17 H UO IN .1 •14 IN U 117 22 1 ii;) 20 2 (12 17 2 71 24 2 (13 l» It (14 1.1 4 60 3 U 84 20 lU U 22 U SU 4 01 111 02 10 7 04 10 4 70 1,1 5 71 22 1 05 1.1 (1 00 III 07 12 ,1 71 lu.o U ,5 N2 19 2 15 U 20 2.1 4 30 1.1 5 24 8 3N 1.1.0 14 7.4 34 17 3 22 U U 30 24 4 31 20 1 17 30 3 14 1« .1 37 22 1 41 13 4 11 « 34 21 4 2(1 13 31 31.9 20 1C.2 31 14 ,1 24 24 4 32 17 1 10 11 3 13 22 9 24 5.N 22 19 9 38 .Moii*- (uri. I»ir rt'iit, 77 3N 47 110 (II 72 4(1 fi7 Oil 71 ,10 04 .1.1 (1.1 04 .17 4N 70 ,19 .14 .1,1 .12 0,1 ,1H (14 00 74 .13 09 57 50 04 N3 NO 90 74 NN 5N 02 7K 77 NO 31 Nl 121 80 79 47 117 32 1,10 129 44 5N (13 54 52 123 9N 125 93 74 Nl (19 45 150 31 41 62 S|». gr., nv« .Iry, l.ii-i.l . ,SlirinkiiK<^ fri>rn Hrrt'ii lu ovrli-ilry .'(iiiilitiun. 0.400 . 5,10 .510 .315 5,1(1 .'.45 57N ,14 I 430 434 .K)4 •101 0(10 (124 (100 (102 (Kill 5,1N 027 (107 (.07 1107 Oils 040 017 053 030 . 095 512 .540 577 .190 5(19 50N .105 .037 ..1K5 . 5U4 .003 .0(HI .173 .5.10 701 .4.14 .475 293 303 .452 300 .3N3 .41N 3,K) .340 370 : .371 I .52N .440 I .477 300 . 353 .377 391 .303 .334 . 300 .325 .■299 .390 .318 .491 .S3H 520 .545 513 .321 X V 13 .-. 12 I Vol wIh'Ii III Uii- Taii- t.VI'll Viil- • liiii Ki,|l- dry. iliiif. tijil. I'cr JVr I'lT pi'nl ci'iit. ronl. 040 12 4 3 4 .19(1 11 7 ... .374 14 5 (12 N.4 009 1 5 1 4 III 5 .001 17.0 7.9 9 ' 039 16.5 5 1 9 9 570 14 II 4 2 N 9 17 Ii 7 4 11 2 2(1 9 7.9 14 2 hi 5 9 10 4 IN 9 N 4 11 4 15 5'0 9 N 15 3 3 9 5 10 9' ON 10 II 21 2 N 5 13 .S 111 5 HI 2 IH 4 7 9 1 1 4 15. 5 j 0.5 9 7 7.19 043 8.0 14 3 4^9 5.7 9 1 732 10 III .002 14 5 4 2 N 3 OOO 13.1 3 7 N 3 .792 17.7 1 5 5 10 \ 704 15 N 2 N 3 090 14 3' 4 9 9 70N 10 4 N 9 2 0011 14 2 4 5 9 7 Htaiie bfndinn. 7 3 7 9 4 9 II 7 I .4,18 10 9 3 7 .437 10 2 3 4 7.0 394 9.2 2 3 5 415 113 4 2 7 1 .407 III 1 3 5 9 5911 12 N 7 . ,107 115 4.5 7.2 '. 3N0 N .4 29 5 .395 9 2 4 1 4 433 115 4 3 7.3 .435 9 9 3.8 5.8 .391 7.8 2.2 5 9 .359 10.5 3.7 9 .335 10.3 3.0 0.2 .558 13.0 3.7 7.4 Kihor It I I'luMtir limit. l.hn. per Hll.ill. 2.1N0 51 NO 44.10 24NII 4490 4190 4290 55011 2N.1I) 3400 3320 0370 4800 .1470 .111011 05.10 0890 4800 0430 0820 0140 58(>0 0220 .1430 .1900 0120 5980 0020 44,10 40,30 41120 4720 3890 3490 4370 5380 0580 4320 4780 4410 ,10110 3720 7700 2820 4300 20(X) 311.10 4430 230(1 4000 3570 3880 3410 3080 27,1(1 .niyo 3740 4300 3330 2000 3 LSI) 3310 3410 4.1,30 .10211 2740 2180 3440 3100 4200 .MniliihiN 111 ru|»- liirc. l.bn. (MT Hiiin. O.IKKI 9.920 8.SN0 4,4.10 8.0111 8.390 9.430 9.510 11,1140 11.4.10 7,8IHI 11.110 II, SSI) 10,280 III.SIO 11,110 12,7211 W.llliO 11,780 12,300 I l,4.''.l) I 1,810 I 1,3.10 10,1190 10,170 ll,0t*r I- II in. 41 10 87 45 90 02 91) 32 44 .10 1 47 1 23 I 22 1 22 1 50 141 1 00 1 42 1 71 1 34 12 1 34 1.27 1 01 I 22 I 29 I 28 .78 .NN .85 1 39 .70 .00 .75 1 05 I 49 .95 1 IN 94 1 20 .71 2 53 51 1 IH) .90 .00 .59 .77 .73 .14 .45 .NN .19 .00 .47 .52 . ,19 .50 .30 .02 .,18 .84 To maii- mum load. In-lljH. piT rii in 13 1 20 III 5 8 14 1 14.2 19 4 117 II 8 Total. 19 24 3 30 2 20 18 31 7 24 1 22 8 24 7 27 7 34 9 30 10 7 .34 1 119 18 3 18 8 17 3 9 8 12 7 9 1 12 7 II 9 4 8 14.5 12 1 13 3 N 9 11.7 13 2 37 9 7 1 7.8 5.7 5.1 4 4 5.2 5 1 5 3 8 1 5 8 5 4 9 4 3 5.9 4.8 5.0 CO 7.2 Ill-lt.M. per ru in. 38 9 43 7 24 8.9 31 4 31 5 47 4 44 2 27 4 52 9' 78 O 99 75 5 58 2 Impact bending. Kiber Ntr«-f*« at t'laMtio limit. ModuluN of elati' ticity. l.liH. per Nij.in. 7,840 11,710 11,720 5.480 11,700 1 1 ,080 12,090 11,700 9,910 10^420 84 4 98 7 58 2 05 1 88.7 87 80 7 04 2 80 4 75 3 72.3 52 9 04 4 17 1 32 10.2 17 4 40 5 20 I) 20 4 37.0 13.400 11.070 11,080 13.780 11.200 10.800 10.580 10.250 13,280 30.7 35.5 19 3 30 7 30 3 101 7 13 20 9 9.5 15 4 7 4 13.0 14 9 12 9 7 4 119 34 8 28 110 12 4 15 12 13 3 0.1 0.5 30 9,800 10,980 11,7.10 10.840 1 1 ,590 15,. 120 8,180 7,0.10 .,290 8,290 5,280 8,870 7.220 0.330 0.870 0,410 9,080 7,4.80 0,740 0.100 7.070 0,190 0,490 0.300 5.3,10 7,750 1000 lbs. pfT Mij.in 9.15 1.104 1388 917 1.101 1812 1307 1 509 1138 i398 Work in bending tu elaatic limit. lu-lbR. per ru.in. 3 09 4.93 5 55 1.84 5.10 3.79 0.52 4 86 4 82 ' 4.48 2114 1411 1080 1581 1590 1.153 1500 1803 2048 Weigbt of ham- mer. Lb*. 50 50 60 60 50 50 60 50 100 4.76 5 45 Height of drop causing complete failure 50 60 50 50 50 60 60 50 60 1414 1507 1481 1479 1177 1498 1105 1310 778 1431 982 1579 1320 1025 1142 1108 1739 1438 1083 979 1115 1203 11.10 1070 900 1203 1 84 4 31 5.23 4 44 44 8 92 3 22 2.49 2.04 2 71 1.69 2.79 2 21 2 19 2 31 2 07 3.02 2.18 60 50 50 50 50 100 50 50 50 50 60 60 50 50 60 50 50 60 50 50 60 50 60 2 09 1.75 2.07 _ 50 50 50 Inchcfl. 30 33 32 10 43 30 48 40 17 '62" 66 28 37 28 38 43 30 40 50 40 45 35 43 35 60 24 26 23 9 20 18 17 16 10 35 28 17 17 21 20 18 13 >16 28 Compreuion | grail Fiber Btreaa elutio limit. l.bfl. per ■q.in. 3610 3210 780 2770 2670 3410 2730 3570 1820 4330 3600 3990 6230 302O 3460 4100 4270 3520 4160 2840 2780 3730 3240 4300 3040 I 2780 I 2640 2290 2670 3470 2400 3150 3030 2870 3980 2320 2-280 1420 2710 3560 1660 2763 2390 2610 j2110 2290 1910 3420 2470 2340 1900 2 too 2260 2370 3420 3560 1880 1690 3010 NoTi.— The blank apace* io the tabic arc due tu the furl that Bonio of the teata were made before the more comprebeuaive scries at present under way was instituted. RESULTS OF TESTS ON 49 SPECIES OF WOOD TESTED IN A GREEN CONDITION IN THE FORM OF SMALL CLEAR PIECES (11. S. Forest Sprvico Circular 213.) [Test apecimens are 2 by 2 inch(>a in section. Bending apecimena are rut .10 inrhea lonx; othera nn- nhorirr, depending on kind of teat] Num- ber of iraaa. Rings per inch. Pro- IKjr- tioo of sum- mer nood. Mois- ture ct)n- tunt Sp. gr., oven dry, bane*! on Shrinkage from green to oven-dry rondition. .Static bending. Impact bending. Compression pa grain rallel to Com- pression perpen- dicular to grain, liber stress at elastic limit. Hardness — Load re- quired to embed a 444-in. ball to i its diameter. Shearing strength paral- lel to grain. Cleavage strength per in. of width. Tension perpendicular to grain. • grown Vol. when green. Vol. when oven dry. In vol- ume. Utt- dial. Tan- gen- tial. Fiber HtreHH at elastic limit. Modulu.-* u( rup- ture. Mu,940 fl,4.5(J 7,800 11,110 9,880 10,280 10,840 11.110 12,720 9,000 11,780 12,300 11.4.50 ll.HlO 1 1.330 10 9911 10.170 1I.(H)0 10.740 12.300 8,310 8,800 8,820 7,380 8,101) 7.780 8.120 9,800 10,710 8,090 8,040 7,700 8,7.50 7,0.50 13,0(U) 0.300 7,380 4,250 0.040 7.110 4.450 0,570 0.340 5,970 5,770 5,130 5,170 8,030 0,430 7,710 5,270 4.700 5,180 5,400 5,310 0,560 7,400 4,550 3,860 6,820 5,200 7,170 KMM) l,b.i per mj.in. 9('>0 1410 1319 842 1353 1.597 1222 1314 1 052 1138 1170 1.502 1099 1399 1025 1,508 18X3 1289 11105 1.5.53 1 110.5 17119 lii:)S 1340 1392 17,52 1.503 1732 1445 1402 1437 913 1248 1208 1474 1.593 1078 1137 1311 1194 1219 1121 1329 904 1045 043 7.54 1378 801 1323 1242 1131 917 1015 972 1002 1384 l;t95 900 879 1111 10.53 1073 1024 1101 800 798 1143' 908 1230 In-lbn. per cu.in. 41 1 H) ,M7 45 .90 02 .90 1 . 32 44 In-lbi. per cu.in. 13.1 20. 10 6.8 14 1 14 2 19 4 11 7 11.8 In-lbM. per cu.in. 38 9 43 7 24 8 9 31 4 31 5 47 4 44 2 27 4 I,b«, p4T stj.in, 7.840 11.710 11.720 5.480 11.700 1 1 .080 12.090 1 1.7IH) 9.9 10 l(XX)lbii per Hij in 955 1.504 1388 917 1.501 1H12 1307 1.509 I13N lu-lbfl. per cu in. 3 09 4 93 5 55 1 84 5 10 3 79 52 4 80 4 82 I.tix ,50 50 ■50 ,50 50 .50 50 ,50 100 Inches. 30 33 32 10 43 30 48 40 17 Lbs. per Hij.in. 2200 4220 4040 1820 348(1 34(K) 3740 3990 2700 209O 3310 4520 3200 4570 4(100 4320 5240 3980 4870 4700 4820 4820 5000 4300 4370 4000 400O 4970 3680 3070 4(K)0 3330 344(1 3210 3400 430O 4840 3520 3530 3490 3700 3080 5810 2790 3550 1990 3030 3900 2000 3040 2920 2800 27,50 2530 24(X) 4280 3080 3570 20(X) 2220 2420 2000 2720 3820 4100 2170 1800 2920 1940 3480 lUOOIbs. sq.in. 1.531 1497 1010 1412 1915 1453 Lbs. per s(].in. 452 889 801 197 005 439 090 730 292 I.bs 505 1121 842 273 1012 827 954 919 530 I.bs. ,542 KMX) 729 220 897 708 883 757 4.50 I.bs. 540 1017 734 217 918 739 893 087 497 lbs. per sq.in. 870 1300 1239 50O 11.54 1103 1210 1197 778 bs. per sq.in. 832 1312 1100 017 1,375 1188 1330 1174 872 I.bs. 275 333 357 130 339 294 282 401 210 I.ba. 200 340 341 108 527 287 377 424 270 bs. per aq.in. ba.per aq.in. 5 5 3 5 S 0.640 .590 .374 .009 .661 12.0 117 14.5 10.5 17.0 4.3 0.4 645 010 226 633 446 671 611 fia D Countiea, Ind. ria 6 2 4.0 7.9 8.4 10.5 9 303 909 626 )d 1 .639 15.6 5.1 9.9 766 832 'ia .... Mo 2U 1 10 » 11 8 11 S 4 10 ai 19 4 9 1 10 2 1 ad 8.3 23 9 13.9 11.7 19.3 16.6 31.5 22 1 17.8 18.5 18.9 22.1 20.2 17.2 24.2 19.9 15.4 3.6 20.0 18.0 22.0 30.4 11.1 10.7 10.4 1ft. S 70 60 71 71 55 09 61 58 60 64 67 63 62 71 63 64 06 84 ei" 62 64 70 71 .676 14.0 4.2 17.0 7.4 20.9 7.9 8 9 11.2 14 2 50 1 47 1 23 1 22 1 22 1 50 141 1 00 1 42 1 71 1 34 1 12 1 34 1.27 1 01 1 22 1 29 1.28 .78 .88 .85 1 39 .70 .00 .75 1 05 1 49 95 1 18 94 1 20 .71 2 53 51 1 ()0 .00 19 24 3 :io 2 20 18 31 7 24 1 22 8 24.7 27 7 34 9 30 10 7 34 1 11 9 18 3 18 8 17 3 9 8 12 7 9 9 1 12 7 114 9 8 14 5 52 9 78 99 7J 5 .58 2 84 4 98 7 58 2 05 1 88 7 87 9 80 7 04 2 80 4 75 3 72 3 52 9 04 4 17 1 320 10 2 17 4 40 5 20 20 4 37 10.420 1398 4 48 .5(1 02 2730 3570 1820 4330 3000 3990 5230 3020 3450 41(N) 4270 3520 4100 2840 2780 3730 3240 4300 3040 2780 2.540 22'.lO 2070 3470 2400 31.50 3030 2870 3980 2320 2280 1420 2710 3500 1000 2703 2390 2010 , 2110 2290 1910 3420 2470 1008 1058 1257 1010 1081 1359 1904 1411 1474 1081 1911 1980 1809 1478 1029 1943 1920 1530 1403 "ioo2 1335 1295 1924 1489 940 1308 1181 1405 133i 1073 1280 754 808 1738 882 1579 1440 1332 1054 1219 991 1890 ; 1040 io29 , 1040 ! 1271 1174 1318 1175 1272 1021 971 1590 .575 994 1000 980 1005 9.58 829 795 773 1095 1134 1094 1134 1105 1251 1101 1191 1330 1348 1375 1313 422 436 661 786 . ■. . i I..:::: 10 5 (^ *^ 10 4 114 18.9 8.4 ..... V» i ^ . . . . . 938 1101 1224 1130 1114 1158 1080 .:...! 1 1 1010 12(N) 1235 1312 1.334 1200 1337 1053 1215 1020 1405 1457 1318 1505 1245 1297 1490 2090 15.0 15 3 10 9 21.2 16.0 18.4 5 6 3 6.8 8.5 6 6 9.8 9 5 10 9 13 8 10 ■) 1 1 1 1 ' 1 \iL'.'.'.'.'.'.'.'.'.'.'.. :;::::::i"" ; '.'.'.'.'.'.'.'. '.'.'.'.'.'.]'.'.'.'.'.'.'.'.'.'.'.'. ' 7 9 ^ :"::::: i .:::;;i;;::::i:":;; v« 15.5 6.5 9.7 972 1088 1084 606 053 870 1148 844 807 082 943 1355 829 727 1004 912 802 2200 433 451 288 518 548 307 334 427 445 420 304 332 491 3,58 400 3,53 342 320 410 314 539 018 302 279 322 262 480 1230 1390 1885 1 1 1 1 1 .759 8.6 13.400 11,070 11,080 13,780 11,200 10,800 10,. 580 10,2.50 13.280 9.800 10,980 11,7.50 10,840 11,. 590 15.520 8.180 7,0.50 5,290 2114 1411 1080 1581 1.590 1553 I.5O0 1803 2048 1414 \M7 1481 1479 1177 1498 1105 1310 778 4 70 5 45 4 55 6 75 4 41 4 21 4 10 3 30 4 79 ' '■( 84' 4 31 5 23 4 44 44 8 92 3 22 2 49 2 04 50 .50 50 .50 30 50 ,50 .5(1 50 ,50 ,50 ,50 50 .50 KM) 50 50 50 50 28 37 28 38 43 30 40 .50 40 45 35 43 35 00 24 25 15 1802 700 992 905 1139 11.39 1107 924 1205 1800 021 918 815 1008 1044 1020 884 1217 1832 620 901 815 1081 1039 1002 901 1099 552 297 376 357 354 307 380 305 428 382 342 371 379 337 010 330 513 451 487 446 470 410 536 457 500 474 470 305 1133 610 1446 1 1130 1330 n CounUea, Ind. 5 5 e s 1 .643 14 3 4 9 9 1 1193 1340 1 1190 1215 1132 1047 1198 1414 1090 1187 1138 1102 1270 1455 1530 1402 1220 1195 10.59 1394 1414 1292 1425 1308 1190 1320 864 .732 .002 .660 10 5.7 100 14 5 4.2 1 8 3 714 659 039 683 757 622 632 624 728 924 880 D Counties, Ind. 13.1 3.7 8.3 874 787 d .702 17.7 6.6 1 10.6 909 5 5 5 5 ... 5 1 5 9 4 & 14 9 5 5 5 5 5 5 70 5 S 5 5 5 22.1 is.e 16.0 12.5 19.0 6.5 19.2 IS. 9 23.4 15 5 24.8 15.0 7.4 17.3 9 9 24.4 ■20 1 - :«) 3 16.5 22.1 13.4 11.0 21 4 13 31.0 16.2- 14.6 24.4 17.1 11.3 22 9 ft. 8 1».0 66 60 67 71 82" "26 36 "38" 14 34 22 30 31 K 17 14 37 41 34" 26 31 20 - 31 24 32 16 13 24 22 38 .704 .696 .708 069 ^838 526 .545 .315 16. 8i 0.2 ! 8.3 14 3' 4 9 9 10.0 4 8 9.2 14.2 4 5 9.7 ' 8. '9 '.'.'.'.'.'.'.'.'.'. 13 5 50 7.3 12 1 13 3 8 9 11 7 13 2 37 9 7 1 7.8 5.7 30 7 35 . ,5 19 3 30 7 30 3 101 7 13 20 9 9 5 1113 1087 1183 1093 847 1838 004 814 321 1030 1049 1103 1083 800 2312 500 000 225 943 1047 1147 1031 790 1702 599 733 228 749 and Morgan 997 986 929 d 900 978 017 013 830 573 517 853 742 790 672 709 1000 812 740 702 044 080 094 049 1102 1084 014 062 800 654 039 858 723 813 747 719 953 741 070 714 080 700 717 039 205 277 148 425 380 139 540 472 241 781 12 4, 4 4 7.0 2.1 7.9 4.9 796 236 Pariah, La. .513 .321 11.5 3.8 0.0 2.5 6 7.1 90 .00 5 1 4 4 15 4 7 4 8,290 5,280 1431 982 2 71 1 .59 50 ,50 23 9 ■ .^j, 18 17 10 10 35 28 400 284 355 203 354 235 107 130 154 133 242 261 .458 .437 .394 .415 .407 .599 .607 10 9[ 3 7 10 2' 3 4 6.6 7.0 5.0 7 1 5 9 7.0 7.2 .59 .77 .73 54 .45 .88 59 6 6 6 2 6 6 1 6 3 8 1 6 8 13 14 9 12 9 7 4 11 9 34 8 28 8,870 7,220 (1,330 0,870 0.410 9.080 7,480 1579 1320 1025 1142 lio.s 1739 1438 2 79 2 21 2 19 2 31 2 07 3 02 2 18 .50 .5(1 ,50 50 50 50 .50 415 381 403 288 310 574 355 399 322 354 .307 318 502 345 416 334 334 310 319 521 340 139 145 108 142 253 187 101 127 187 151 140 200 180 154 213 271 240 179 304 a 9.2 M 3 10.1 12 8 11.5 2 3 4 2 3.0 0.0 4.5 323 ' 298 206 .386 .396 .433 .435 .391 8.4 9.2 11.6 9 9 7.8 2.9 4 1 4 3 3.8 2.2 6.6 0.4 7.3 5.8 6.9 00 47 .52 .59 .02 6 4 9 4 3 6.9 11 12 4 15 12 13 3 0,740 0,100 7.070 0.1 '.H) 0.490 1083 979 1115 1203 1150 2 34 2 13 2 51 2 20 2 0(1 50 50 50 .50 50 17 17 21 20 18 2340 19f)0 2100 2250 2370 3420 3500 1880 1590 334 310 310 315 304 307 311 300 330 294 342 318 323 332 299 168 152 1 102 ! 152 144 189 179 187 154 160 239 281 230 226 304 321 252 a 6 il 6 4 2 5 "360 .336 10.6 10.3 3.i' 3.0 6 9 6.2 .'.50' ' ' .30 .02 .58 .84 . . .^.^ . 5.0 6.0 6 6 7.2 1 5 5,3.50 1070 900 2 09 1.75 50 .50 13 > 15 272 231 "253 216 "274 220 "607 538 754 645 883 "624 600 783 615 843 110 1-29 136 143 lO Countiaa. Wia. 1 .568 13.6 's!?' ■ 7.4 30 7 J50 1203 2.07 _ 1 "56" "28" j 3010 401 380 370 167 169 236 274 ■ j^,„, Th* blank apacca io the Ubie are due to the fact that aome of the taata were made before the more comprebeasive series at preaent under way was instituted TIMBER 649 Table II MECHANICAL PROPERTIES OF 32 SPECIES OF AMERICAN WOODS (All Values Reduced to 12 per cent Moisture.) (Small Specimens Used.) (U.S. Forest Service Circular 15) Species Common Name. No. of Sp.gr. Tests. 30 .89 75 .85 25 .77 137 .81 37 .78 72 .78 410 .63 218 .80 14 .73 117 .73 44 .74 10 .62 1230 .61 31 .73 251 .73 256 .74 660 .53 216 .74 40 .72 170 .62 57 .72 153 .72 87 .62 49 .80 118 .59 100 .50 18 .54 330 .57 655 ,46 41 .51 130 .38 87 .37 Strength in Pounds per Square Inch. Cross-bending. Appar- ent Elastic Limit. Mod. of Hupt. Mod. of ■ Elas. -H 1000. Com- pres- sion along Grain. Com- pres- sion across Grain. Shearing along Grain. Pignut hickory Mockernut hickory . Bitternut hickory, . . Shagbark hickory. . . Pecan hickory Nutmeg hickory. . . . Cuban pine. ....... White oak Water hickory Texan oak Cedar elm Green ash Long-leaf pine Water oak Spanish oak Cow oak Loblolly pine Overcup oak Yellow oak Spruce pine Red oak Willow oak White ash Post oak Sweet gum Red pine White elm Short-leaf pine Bald cypress Douglas spruce White pine White cedar • • 12,600 11,700 11,100 11,200 11,600 9,300 11,100 9,600 9,800 9,400 8,000 8,900 10,000 8,800 8,600 7,600 9,200 7,500 8,100 8,400 9,200 7,400 7,900 8,400 7,800 7,700 7,300 7,800 6,600 6,400 6,400 5,800 18,700 15,200 15,000 16,000 15,300 12,500 13,600 13,100 12,500 13,100 13,500 11,600 12,600 12,400 12,000 11,500 11,300 11,300 10,800 10,000 11,400 10,400 10,800 12,300 9,500 9,100 10,300 10,100 7,900 7,900 7,900 6,300 2730 2320 2280 2390 2530 1940 2370 2090 2080 1860 1700 2050 2070 2000 1930 1610 2050 1620 1740 1640 1970 1750 1640 2030 1700 1620 1540 1680 1290 1680 1390 910 10,900 10,100 9,600 9,500 9,100 8,800 8,700 8,500 8,400 8,100 8,000 8,000 8,000 7,800 7,700 7,400 7,400 7,300 7,300 7,300 7,200 7,200 7,200 7,100 7,100 6,700 6,500 6,500 6,000 5,700 5,400 5,200 3200 3100 2200 2700 2800 2700 1200 2200 2400 2000 2100 1700 1260 2000 1800 1900 1150 1900 1800 1200 2300 2300 1900 3000 1400 1000 1200 1050 800 800 700 700 1200 1100 1000 1100 1200 1100 770 1000 1000 900 1300 1000 835 1100 900 900 800 1000 1100 800 1100 900 1100 1100 800 500 800 770 500 500 400 400 650 MATERIALS OF CONSTRUCTION per cent for buildings and similar structures wherein the timbers are pro- tected from the weather and are not subject to impact. The conamittee has used the following factors of safety: viz., 5 for extreme fiber stress in bending, 4 for shear along the grain, 3 for compression along the grain, and 2 for compression across the grain. For steady loads these factors become 34, 2|, 2, and 1| respectively. SAFE WORKING STRESSES FOR STRUCTURAL TIMBER (Adopted by Am. Ry. Eng. Assn., 1909) (All values expressed in pounds per square inch) Compression. Bending Extreme Fiber Stress. Shearing along Along Grain. Across Grain. Grain, Douglas fir Long-leaf pine Short-leaf pine White pine 1200 1300 1100 1000 1100 800 1000 1200 900 1100 900 1300 310 260 170 150 180 150 220 220 150 170 230 450 1200 1300 1100 900 1000 800 900 1100 900 900 800 1100 170 180 170 100 150 Norway pine 130 170 Western hemlock. . . Redwood. 160 80 Bald cypress Red cedar 120 White oak. 210 In using above values do not increase the live-load stresses to take account of impact. DURABILITY, DECAY, AND PRESERVATION OF TIMBER 584. Durability and Decay of Timber.* " The decay of wood is not an inorganic process like rusting of iron or the crumbling of stone, but is due to the activities of low forms of plant life called bacteria and fungi." Bacteria are among the lowest forms of life, often only a single cell which multiplies by division into two cells which in turn divide again. Several such cells united together form a thread or filament of micro- scopic size. Fungi consist merely of tiny threads, known as mycelium, which in most of the higher forms grow together to form compact masses as, for instance, the toadstools which grow on rotting logs and the " punks " or " brackets " on the trunks of forest trees. The action of fungi and bacteria in destroying wood is, so far as is known, about the same, and the two will therefore be considered together. * Bulletin 78, U. S. Forest Service. TIMBER 651 The mycelium starts from a single spore (a primitive substitute for seed, which are born in infinite numbers by the fruiting fungi, and are distributed by the wind). Under favorable conditions the spore sprouts or germinates, sending forth a single thread or hypha which branches and rebranches quickly. The hyphae creep through the tissues of the material in which they are growing and absorb the materials necessary for their growth. In living cells they attack the protoplastic contents ■ — the starches, sugars, and oils. In dead cells they attack the cell walls. After a time some hyphae form the fruiting body which bears a crop of spores. The action of the fungi results in complete or partial solution of the substance attacked. The sugars may be absorbed by the fungus directly or changed to a more easily digested sugar, the starches are changed to sugar and then absorbed. The nitrogenous substances undergo similar changes. Wood fibers may be dissolved entirely, or only the cellulose or lignin elements may be affected. The chemical changes bring about physical changes, the hard elastic fibers becoming a mushy mass or a dry brittle substance which falls to pieces at the least touch. At first the hyphae simply puncture the cell walls here and there, later they increase in number until the strength of the wood is lost entirely. The solvent action of the ferments of the fungi upon the cellulose or wood £ber often causes cracks, while other ferments dissolve the tissue in the intercellular spaces so that the cells ultimately fall apart. Some fungi require oxygen, and others grow best without; some require sugar and starches, while others do not; all require some moisture. Susceptibility to decay and comparative resistance to decay vary with different classes of timber. The determining factors are as yet almost unknown. Hardness, density, specific gravity, and strength seem to have no influence one way or the other. Some very hard tropical timbers decay very rapidly while others last very long. Hard strong oak decays much faster than fight porous cypress; tamarack and hem- lock decay rapidly, while cypress and cedar are lasting; elm and birch are short lived, the locust long lived, etc. 685. The Preservation of Timber. The simplest way to prolong the life of timber exposed to the attack of fungi' is to reduce the moisture content by seasoning. By piling timber so as to permit free access to air all around it, the moisture may be reduced to about 15 to 18 per cent. Of course the chmate and the size of the timbers have a great influence upon the rate of seasoning and the total amount of moisture lost. The moisture content of air-dry wood may be still further reduced by kiln drying, but the effectiveness of all methods of moisture reduction as a preservative measure must be hmited by the fact that timber so dried 652 MATERIALS OF CONSTEUCTION readily reabsorbs moisture upon being again exposed to dampness, and and benefits obtained are only moderate. " By far the best method of checking the growth of fungi, however, is by poisoning their food supply — by injecting poisonous substances into the timber, so changing the organic matter into powerful fungi- cides." It is a widespread idea that the germs of decay are inherent in the wood, needing only an opportunity for development to bring about its destruction. On the contrary, all wood-destroying agencies start from the outside and may even be excluded by certain paints which merely coat the surface of the timber but which are poisonous enough to prevent the germination of spores. "The ancients were in the habit of painting their statues with oily and bituminous preparations to pre- serve them from decay. The great wooden statue of Diana at Ephesus, which was supposed to have descended miraculously from heaven, was protected from earthly decay by oil of nard. When the preservative fluids were extracted from the heart of an Egyptian mummy that had resisted decay for over 3000 years, decomposition immediately set in. This showed that it was the presence of the antiseptic which prevented decay, and not a chemical change of the tissues." The first deliberate attempts to preserve timber from decay date back many centuries, when wood was charred to make it more resistant. Later came the period when wood was coated with preservative paints, then came attempts to inject preservatives into the wood. " Of the many antiseptics which have been proposed for the preser- vation of timber only four have been largely used with success in the United States. These are creosotes (dead oil of coal tar), zinc chloride, corrosive sublimate (bichloride of mercury), and copper sulphate. At the present time copper sulphate has fallen into almost total disuse, and corrosive sublimate is restricted to two plants in New England." There are many patented preservatives known by various names, practically all of which have either creosote or zinc chloride as their base. Creosote is a by-product of coal tar produced at illuminating gas plants and by-product coke oven plants. Creosote is not a simple sub- stance like zinc chloride, but is a complex compound of phenols (tar acids), naphthalene, anthracene, and residues. Wood creosote, distilled like coal tar from wood, also possesses strong antiseptic properties. Zinc chloride is obtained by dissolving metallic zinc in hydrochloric acid. This is diluted with water before being used as a wood preserva- tive. Creosote has the advantage of insolubility in water, so that it will not wash out of timber; zinc chloride has the advantage of much greater cheapness, and so is preferred for use in comparatively dry situations. The processes, by which preservatives are injected into timber may TIMBER 653 be diArided into two general classes, the " pressure processes " and the " non-pressure processes," the distinction being only in the fact that in the former force pumps, air compressors, etc., are utilized, while in the latter only atmospheric pressure is relied on. The Pressure Processes The Breant Process. One of the earliest practical pressure processes was that devised by Breant in France in 1831. The timber is inserted in a vertical cylinder and the liquid forced in almost to the top. A vac- uum is produced by opening a connecting valve to an auxiliary chamber which has been exhausted by an air pump, after which the valve is closed and the liquid forced in till a pressure of 10 atmospheres is reached. The time required for impregnation is about six hours. The Bethall process, a modification of the Breant process, using creo- sote, and the process of Burnettizing, using zinc chloride, are the most widely used pressure processes. The method of injection is practically the same for each. The timber to be treated is placed on iron trucks or " cylinder buggies " and run into huge horizontal cylinders 8 or 9 feet in diameter and often as much as 150 feet long, after which the doors are hermetically sealed. Live steam is admitted at a pressure of 20 pounds per square inch for several hours, then the steam is blown out, the vacuum pump started, and within a few hours as much air and moisture as pos- sible are exhausted. Finally, the preservative is run into the cyUnder at a temperature of about 160° F., the pressure pumps are started, and the desired amount of preservative is forced into the timber. The surplus preservative is now blown back into the tank, the timber allowed to drip for a few moments, and then withdrawn. The boiling process is used principally for Douglas fir on the Pacific coast. The timber is placed in the cylinder, which is filled with creo- sote slightly above the boiUng-point of water, and kept in this condi- tion for a period of from a few hours to two days, during which time water and volatile oils are driven off. Finally a pressure of 100 to 125 pounds per square inch is applied and the temperature is allowed to drop, thus forcing the preservative into the timber. In the A. C.W. process the procedure is like that of the Bethall except that air pressure (15 pounds per square inch) is apphed after the vacuum, and maintained while the creosote is admitted in order to prevent unequal absorption during the fiUing of the cylinder. Then 100 pounds per square inch pressure is applied till the penetration reaches the desired degree, after which the preservative is drawn off and air forced in at a pressure of 60 to 80 pounds per square inch to increase the penetration of the preservative. 654 MATERIALS OF CONSTRUCTION The Raping process is largely used with creosote, and is the best known empty-cell process. The timber is air-dried before being placed in a cylinder. Air is forced in at 75 pounds per square inch pressure till the wood is filled with the compressed air. Then, without abating the pressure, the oil is admitted at a higher pressure (80 to 85 pounds per square inch). The pressure is subsequently increased to 225 pounds per square inch. This forces the penetration of the oil into the timber. Finally the valves are opened, the excess oil is drawn off, so relieving the pressure around the timber, and the expansive force of the compressed air in the wood forces out much of the oil, leaving only a coating around the cell walls, thus resulting in a deep penetration with a light absorption. The Lowry process tries to accomplish the same effect without using compressed air. The air-seasoned timber is introduced into the cylinder, oil is admitted, and a great pressure applied, causing a compression of the air in the wood cells and intercellular space. Finally the oil is with- drawn, a strong vacuum is formed and the expanding of the compressed air forces out the oil as in the Rtiping process. The creo-resinaie process consists in impregnating the timber with a mixture of creosote and resin (50 to 75 per cent creosote). The method differs from the Bethall process only in. using dry heat instead of a steam bath before the vacuum. The Wellhouse process is an attempt to prevent the leaching out of zinc chloride from treated timber. It reUes for its efficiency upon the tendency of glue and tannin, when combined, to form a leathery, water- proof substance, or " leatheroid." A ^ per cent solution of glue is mixed with the zinc chloride solution and forced into the timber, and a tannin solution is afterward forced in. The formation of " leatheroid " in the cell openings tends to prevent the absorption of water and the leaching out of the salt. The Allardyce process is an attempt to devise a process cheaper than one using creosote and not involving the leaching out of zinc chloride. Zinc chloride in a 2 or 3 per cent solution is first forced into the timber (12 pounds per cubic foot) by a process similar to " Burnettizing " and then creosote is injected (3 pounds per cubic foot of timber). The creo- sote remains largely on the exterior and is designed to protect the sol- uble chloride in the interior. The card process is a similar one in its objects. The preserving fluid contains 15 to 20 per cent of creosote, and a 3 to 5 per cent solution of zinc chloride for the remainder. After a steam bath and vacuum the preservative is admitted under pressure. The two preservatives will not mix, and are of different specific gravity, so they are kept in a mechanical mixture by a centrifugal force pump which draws the solu- TIMBER 655 tion from the top and forces it back through perforated pipes running along the bottom of th^ cyhnder. The Novrpressure Processes Injection of preservatives by the non-pressure processes depends on different principles from those involved in the pressure processes. The wood, after being seasoned in air, is immersed in a bath of hot liquid in a suitable containing vessel for from one to six hours, during which time the air and moisture in the wood expand and partially pass off. Then, as quickly as possible, the wood is changed to a bath of cooler preserva- tive. This causes contraction of the imprisoned air and moisture, creat- ing a partial vacuum which is subsequently destroyed by the entrance of the preservative. Green-timber processes prolong the hot bath until much of the moisture in the green wood has been expelled. Air is not present to help in forming a vacuum, so the treatment is slow and unsatisfactorj', as well as expensive, through loss of time and evaporation of the pre- servatives. The oven process is founded upon the fact that the hot bath serves only to expand the air and moisture, and drive them out, with practically no penetration of the preservatives until the cold bath is reached. Much the same results may sometimes be secured by subjecting the timber to dry heat in an oven or kiln and then immersing quickly in a bath of cold proservative. Only in special cases is the process desirable, as when a metal-corroding preservative is used and metal tanks and steam coils cannot be employed, in which case the dry hot timber is plunged into a cold bath in a wooden tank. The Empty Cell Non-pressure Process In the non-pressure processes empty cells cannot be secured by vac- uum, force pumps, etc., so that the problem of securing a good pene- tration with a small quantity of preservative used presents itself. But here again the non-pressure processes have proven successful, and for some of the more porous woods the depth of penetration and the economy of preservative compares favorably" with empty-cell processes in pressure treatments. Practically the only difference in this process from the usual non- pressure process is that here the length of time of immersion in the cold bath is shortened. Suppose that the hot bath has partially expelled the air and moisture to a depth of 3 inches in a timber. When the cold bath is appUed the drop in temperature first takes place in the outer 656 MATERIALS OF CONSTRUCTION band of wood, causing the partial vacuum to first form there. This causes an inrush of the preservative which fills the intercellular and cel- lular space. If the timber is withdrawn before the preservative has penetrated more than 2 inches, instead of the 3 inches affected by the hot bath, the subsequent cooling of the third inch causes a further vacuum to form, and subsequently be destroyed by the admission of superfluous oil from the outer saturated portion, thus resulting in a depth of penetra- tion of 3 inches with the amount of preservative required to saturate 2 inches. A second hot bath is sometimes employed with similar results. The oil is a better heat conductor than the wood, causing a further heating of the interior, resulting in second expansion and expul- sion of air and moisture, with the formation of a partial vacuum which is destroyed by the increased penetration of the excess oil in the outer portion when the stick is exposed in air. The low-pressure process, developed by the Forest Service, consists in subjecting the seasoned timber first to a hot bath to cause expansion and expulsion of air and moisture followed by a cold bath. In this latter bath atmospheric pressure is not depended upon, but low pressure, not exceeding 70 pounds per square inch, is applied by means of the pumps which force in the preservative. The treatment is carried out in cylinders as in ordinary pressure processes. The non-pressure processes and the low-pressure processes can be used only in a restricted field. They are not adapted to use with non- porous woods or where large quantities of timber must be treated in a short time. Superficial Treatments- Less efficient but cheaper treatment can be secured by painting the face of the timber with at least two coats of hot creosote or some similar preservative . (the brush method). The liquid will not penetrate to any great extent, but as long as there remains an unbroken antiseptic zone around the surface the spores of the fungi cannot enter. Thorough air- seasoning before painting is necessary, since otherwise checks may form and provide a means of access to the interior of the timber for the spores of the fungi. A still less expensive treatment than the brush treatment is the method of dipping the timber in an open vat of the preservative. Usually the timbers are carried through the bath on chain conveyors and remain submerged only a few minutes at most. Dipping is not only more economical of time and labor, but gives better results than the brush method. TIMBER 657 Effect of Preservative Treatments upon Strength of Timber The effect of preservative treatments of timber upon strength has been investigated by the Forest Service under the charge of Professor W. K. Hatt,* Fig. 346. It appears in general that the strength of treated timber depends, first, on the percentage of moisture remaining in the wood, and second, on whether the wood has been subjected to injuriously high temperatures during the process of steaming and the 7500 6 r 8 9 10 # Number of Kings per Inch U 12 13 14 Fig 346. — Comparative Strength of Loblolly Pine Untreated and Creosoted without Steaming. (Cir. 39, U. S. For. Ser.) vacuum treatment. A high degree of steaming is injurious to wood, the limit of safety depending upon the quality of the wood, the degree of seasoning, and the temperature and duration of the steam bath. For loblolly pine this limit was found to be 30 pounds per square inch for four hours or 20 pounds per square inch for six hours. During steaming the amount of moisture in the wood must be increased with consequent weakening of the fibers if the vacuum following the steam is not able to remove the added moisture. There seems to be no ground for behev- ing that non-pressure processes can be injurious to the strength of the timber. The presence of creosote in itself will not weaken wood, since it appears not to enter the cell walls and fibers, but only to coat them, * Circular 39, U. S. Forest Service. 658 MATERIALS OF CONSTRUCTION and so, if the wood is not perfectly seasoned, the temperature of the bath may aid the seasoning and increase the strength of the wood. The presence of zinc chloride will not weaken wood except by the addition of water by reason of the fact that it is in itself a water solution. Subsequent seasoning will obviate this difficulty. A too-concentrated solution may cause chemical dissolution of some of the wood fibers, but this danger can easily be avoided. INDEX Abrasive Resistance. (See Material in question) Absorption. (See Material in question) Acid Bricks, 241-243 Adhesion, Concrete to Steel, 199 Adhesive Strength. (See Material in question) Alabaster, 3 Alca Lime, 56-58 Alkali Waste, 98 AUotropic Forms of Iron. (See Steel, Structure and Constitution) Allotropic Theory of Hardening. (See Steel, Heat Treatment) Alloy Steels, 500-540 Definition of, 500 Quaternary, 500, 534-540 Ternary, 500, 501-534 Alloys, Non-Ferrous. '(See Non-Ferrous Alloys) Alpha-Iron. (See Steel, Structure and Constitution) Aluminum, 574^578 Castings, 577 Commercial Forms of, 574 Electrical Resistance, 577 Extraction of, 575-576 Electrolytic Process, 575-576 Occurrence in Nature, 574-575 Production Statistics, 578 Properties, 576-578 Cold-working, Effect of, 577 Rolled-, 577 Uses, 574, 577-578 Wire-, Cold-drawn, 577 Aluminum Alloys, 599-603 Alimiinum Bronze, 588 Aluminum-Copper Alloys, 599-601 Aluminum - Copper - Manganese Alloys, 602-603 Aluminum-Copper-Zinc Alloys, 602 Aluminum-Magnesium Alloys, 601-602 Aluminum-Zinc Alloys, 599 Annealing. (See Steel, Heat Treatment; also Malleable Cast Iron) Anthracite Pig Iron, 291 Anti-Friction Metals, 603-604 Arbitrfition Bar, 331 Arc-Resistance Furnace, 400, 401-403 Austenite. (See Cast Iron, Constitution; also Steel, Structure and Constitu- tion) Babbitt Metal, 603-604 Basic Bricks, 243 Basic Pig Iron, 291 Basic Steel. (See Steel, Bessemer and Open-hearth) Bauxite Brick, 243 Bearing Metals, 603-604 Bell Metal, 584 Bessemer Pig Iron, 291 Bessemer Steel. (See Steel) Beta-Iron. (See Steel, Structure and Constitution) Beta-Iron Theory of Hardening. (See Steel, Heat Treatment) Black-heart Malleable, 338 Blast Furnace, 272-274 Blast Furnace Smelting. (See Pig Iron Manufacture) Blister Steel, 294, 361, 363-365. (See also Steel) Blowing Engines, 374 Blow-holes, 319, 326, 417-418 Boiler Steel, 362, 439, 443, 467 Box-piled Iron, 352 Brass, Red, 591 White, 591 White, Solder, 591 Yellow, 591 Brasses, 588-595 Copper-Zinc-Aluminum, 592 Copper-Zinc-Iron, 594 659 660 INDEX Brasses — Continued Copper-Zinc-Lead, 692 Copper-Zinc-Manganese, 593 Delta Metal, 594 German Silver, 595 Hard Brazing Metal, 591 Manganese Bronze, 588, 593-594 Muntz Metal, 591 Ordinary, 588-591 Composition, 588-591 Properties, 588-591 Phosphorus-, 594 Pinchbeck, 591 Red-, 591 Special-, 592-595 Standard-, 591 Sterro Metal, 594 Tombac, 591 Vanadium-, 594 Brazing Metal, Hard, 591 Brick Machines, 226-231 Brick Masonry, 251-256 Allowable Loads for, 255-256 Bond of, 252-253 Common, 252 English, 253 Flemish, 253 Joints, 251-252, 254 Mortars, 251-252, 256 Cement Mortar, 251-252, 256 Lime Mortar, 251-252, 256 Operation of Laying, 253-254 Strength of, 254-256 Ties, Wall-, for Face Brick, 253 vs. Stone Masonry, 251 Bricks, 222-244, 248-251 Building Bricks, 223-238, 248-251 Absorbing Power, 250 Classes of, 233-235 Common Brick, 234 Compressive Strength, 248-249 Glazed and Enameled Brick, 234 Hollow Brick, 235 Manufacture, 223-235 Clays, Classified, 223-224 Clays, Influence of Kind of, on Character of Bricks, 224 Dry Clay Process, 231 Hand Processes, 224-226 Kilns and Burning, 231-233 Machine Processes, 226-231 Semi-dry Clay Process, 231 Soft-mud Process, 226 Sorting and Classification, 233-235 Stiff-mud Process, 227 Modulus of Elasticity, 251 Bricks — Continued Building Bricks — Continued Ornamental Brick, 234 Pressed or Face Brick, 234 Properties, 248-251 Roman Tile, 235 Sand-lime brick, 235 Shearing Strength, 250 Tapestry Brick, 234 Transverse Strength, 250 Uses of Various Grades, 233-235 Clay Products Classified, 223 Firebricks, 240-244, 248-251 Absorbing Power, 250 Compressive Strength, 248-249 Manufacture, 240-244 Acid-, 241-243 Basic-, 243 Bauxite-, 243 Chrome-, 243 Fireclay-, 241-242 Canister-, 242-243 Magnesia-, 243 Neutral-, 243 Silica-, 242 Properties, 248-251 Shearing Strength, 251 Transverse Strength, 250 Paving Brick, 238-240, 248-251 Absorbing Power, 250 Compressive Strength, 248-249 Manufacture, 2'38-240 Burning, Annealing, and Sorting, 240 Clays for, 238 Molding and Drying, 239 Preparation of Clay, 239 Modulus of Elasticity, 251 Properties of, 248-251 Shearing Strength, 251 Transverse Strength, 250 Sarid-Ume Brick, 235-238, 248-251 Absorbing Power, 250 Classes of, 235 Compressive Strength, 248-249 Manufacture, 235-238 Hardening, 237-238 Lime for, 236 Mixing, 237 Preparation of Lime, 237 Preparation of Sand, 236 Pressing, 237 Sand for, 235 Structural Forms and Uses, 222 Bridge Steel. (See Steel) Broadleaf Trees. (See Trees) INDEX 661 Bronzes, 582-588 Aluminum-, 588 Bell Metal, 584 Coin-, 585 Copper-Tin-Lead-, 587 Copper-Tin-Zinc-, 585-587 Gun Metal, 584 Machinery-, 584 Manganese-, 588, 593-594 Medal-, 585 Nickel-, 588 Ordinary, 582-585 Composition of, 582-585 Properties of, 582-585 Phosphor-, 587 Silicon-, 588 Special-, 585-588 Speculum Metal, 585 Statuary-, 585 Vanadium-, 588 Building Blocks. (See Terra Cotta) Building Bricks, 223-238, 248-251 Building Stones. (See Stones, Building) Burden. (Blast Furnace). 280-282 Busheled Scrap, 352 Calcination. (See Material in Question) Calcium Limes. (See Quicklimes and Hydrated Lime) Cambium, 608-609 Carbon in Cast Iron, 315-316, 323, 324-325, 328-329 Carbonic Acid Theory of Corrosion, 487 Case-Hardening. (See Malleable Cast Iron, Annealing Process; also Steel, Heat Treatment; and Wood) Cast Iron, 294^332 Arbitration Test Bar, 331 Behavior in Cooling, 320-322 Checking, 322 Cause of, 322 Manganese, Effect of, upon, 322 Phosphorus, Effect of, upon, 322 Sulphur, Effect of, upon, 322- Temperature of, 322 Chilling, 322 Manganese, Effect of, upon, 322 Phosphorus, Effect of, upon, 322 Silicon, Effect of, upon, 322 Sulphur, Effect of, upon, 322 Segregation, 322 Cause of, 322 Definition, 322 Hot-Spots, 322 Manganese, Effect of, upon, 322 Phosphorus, Effect of, upon, 322 Cast Iron — Continued Behavior in Cooling — Continued Segregation — Continued Sulphur, Effect of, upon, 322 Shrinkage, 320-321 Graphite Separation and, 321 Impurities, Effect of, upon, 321 Manganese, Effect of, upon, 321 Phosphorus, Effect of, upon, 321 SiUcon, Effect of, upon, 321 Size of Casting, Effect of, upon, 321 Sulphur, Effect of, upon, 321 Blow-Holes, 319, 326 Brittleness, 319-320 Carbon in, 315-322, 323, 324-325 Checking. (See Behavior in Cooling) ChiU, Depth of, 312 Chilling. (See Behavior in Cooling) Classes of, 294, 315 Cleaning Castings, 313-314 Pickling, 314 Sand Blast, 314 Tumbling, 313-314 Composition, 315 Compressive Strength, 328-329 Carbon, Effect of Form of, 328, 329 Factor of Safety, 329 Modulus of Elasticity, 329 Plane of Rupture, 328 Safe Working Stress, 329 Shearing Strength, Relation to, 328 Stress-Strain Diagrams, 329 Ultimate Strength, 328, 329 Long Columns, 328 Short Columns, 329 Yield Point, 329 Constitution, 315-320 Austenite, 316 Carbon, 315-316 Cementite, 315, 316, 317 Classes of Cast Iron, 315 Constitution, Essential, 315 Ferrite, 315, 316, 317 Graphite, 315, 316, 317 Gray, Cast Irons, 315, 316, 317 Manganese in Cast Iron, 320 Brittleness, Effect upon, 320 Carbon, Effect on Form of, 320 Ferro-Manganese, 320 Hardness, Effect upon, 320 Limits, Usual, 320 Spiegel-Eisen, 320 Sulphur with. Effect of, 320 Toughness, Effect upon, 320 Mottled Cast Iron, 315, 318 Pearlite, 317 662 INDEX Cast Iron — Continued Constitution — Continued Phosphorus in Cast Iron, 319-320 Brittleness, Effect upon, 320 Carbon, Effect on Form of, 319 Fluidity, Effect upon, 320 Fusibility, Effect upon, 320 Hardness, Effect upon, 320 Maximum per cent Allowed, 320 Sihcon with. Effect of, 319 Solidification, Effect upon Rate of, 319 Thin Castings, Use in, 320 Toughness, Effect upon, 320 Silicon in Cast Iron, 318-319 Blow-Holes, Effect upon, 319 Carbon, Effect upon Form of, 318 Chill, Effect upon Depth of, 319 Fluidity, Effect upon, 319 Hardening Effect, 319 Reversal of Effect upon Carbon, 318 Shrinkage, Effect upon, 319 Softening Effect, 319 Toughness, Effect upon, 319 Sulphur in Cast Iron, 319 , Brittleness, Effect upon, 319 Carbon, Effect on Form of, 319 Checking and Cracking, Effect upon, 319 Hardness, Effect upon, 319 Manganese, Neutralized by, 319 Maximum per cent Allowed, 319 Red-Shortness, 319 Shrinkage, Effect upon, 319 Silicon vs., 319 Solidification, Effect upon Rate of, 319 Toughness, Effect upon, 319 White Cast Iron, 315, 317-318 Cross-Breaking Strength, 329-332 Arbitration Test Bar, 331 Dimensions, Effect upon, 331 Factors Influencing, 329-330 Importance, 329 Modulus of Rupture, 330 Modulus of Shock Resistance, 332 Rectangular vs. Round Bars, 331 Span, Relation to, 337 Specification (Arbitration Bar), 332 Elongation, 328 Factor of Safety, 328 Fluidity, 319, 320 Flux Used, 297 Founding of. (See Manufacture) Fuel Used, 297 Cast Iron — Continued Furnaces Used. (See Manufacture) Fusibility, 320 Gray, 315, 316-317 Carbon Content, 316 Characteristics, Physical, 317 Constitution, 316 Structure, 316 Hardness of, 318, 319, 320, 323 Carbon, Effect of Form of, upon, 323 Definition, 323 Factors Governing, 323 Manganese, Effect of, upon, 323 Phosphorus, Effect of, upon, 323 SiUcon, Effect of, upon, 323 Sulphiir, Effect of, upon, 323 Manganese in. (See Constitution) Manufacture of, 296-314 Action within Furnace, 301 Charging Furnace, 301, 303 Chemistry of Operation, 302 Control of Operation, 300, 303 Duration of Operation, 303 Flux, 297 Founding, 304-314 Cleaning Castings, 313-314 Molds and Molding, 305-313 Pouring the Iron, 313 Fuel, 297 Coal, 297 Coke, 297 Furnaces, 297-304 Cupola Furnace, 297-298, 300-330 Reverberatory or Air Furnace, 299, 303-304 Iron Melting in General, 296 Pig Irons, Foundry, 296 Scrap Iron, 296-297 Starting Operation, 300, 303 Stopping in, 303 Tapping out, 303 Mechanical Properties, 323-332 Melting in General, 296 Modulus of Elasticity, 327, 329 Modulus of Rupture, 330 Modulus of Shock Resistance, 332 Molds and Molding, 305-313 Chilled Castings, 312-313 Chills, 312 Effect of Chilling, 312 Uses of Chilled Castings, 312 Dry Sand, 310 Advantages and Disadvantaees 310 ' Applications, 310 Drying, 310 INDEX 663 Cast Iron—Continued Molds and Molding — Continued Dry Sand — Continued Molding Operations, 310 Sand, 310 Green Sand, 305-309 Applications, 305 Cores, 308-309 Flasks, 305-306 Molding Operations, 309 Parting Sand, 309 Pattern Making, 306-308 Sand, 305 Sprue-Pins, 309 Venting, 309 Loam, 310-312 Applications, 310 Brick-work, 311 Loam, 311 Loam-boards, 311 Molding Operations, 311-312 Mottled, 315, 318 Constitution, 318 Structure, 318 Patterns, 306-308 Pearlite, 317 Phosphorus. (See Constitution) Pig Iron Used, 296 Properties, 315-332 Red-Shortness, 319 Reduction of Area, 327 Safe Working Stresses, 328, 329 Scrap Iron Used, 296-297 Segregation. (See Behavior in Cooling) Shrinkage. (See Behavior in Cooling) Silicon. (See Constitution) Solidification, Rate of, 320 Specifications for, 328, 332 Strength of, 323-332 Tensile Strength, 323-328 Carbon, Effect of Form of, upoil, 324^325 Cementite and, 324-325 CooUng, Effect of Rate of, upon, 325-326 Elastic Limit, 327 Elongation, 328 Factor of Safety, 328 Factors Governing, 323 Graphite and, 324-325 Manganese, Effect of, upon, 326 Modulus of Elasticity, 327 Phosphorus, Effect of, upon, 326 Reduction of Area, 327 Safe Working Stresses, 328 SiUcon, Effect of, upon, 326 Cast Iron — Continued Tensile Strength — Continued Size of Casting, Effect of, upon, 326 Specification, 328 Stress-Strain Diagrams, 327 Sulphur, Effect of, upon, 326 Ultimate Strength, 327 Yield Point, 327 Thin Castings, 320 Toughness, 319, 320 White, 315, 317-318 Carbon Content, 317-318 Characteristics, Physical, 317 Constitution, 317 Yield Point, 327, 329 Casting Machines (Pig Iron), 285-286 Catalan Forge, 257 Cement Plaster, 1, 7, 8, 14, 18 Cement Rock, 96 Cementation Index, 60-62 Cementation Steel, 294, 361, 363-365. (See also Steel) Cementite. (See Cast Iron, Constitution) Cementite. (See Wrought Iron, Constitu- tion) Cementite. (See Steel, Structure and Constitution) Chalk. (See Portland Cement, Manufac- ture of) Charcoal Iron. (See Wrought Iron, Classes of) Charcoal Pig Iron. (See Pig Iron) Checking. (See Cast Iron, Behavior in Cooling) Checking. (See Wood) ChiUing of Cast Iron. (See Cast Iron, Behavior in Cooling) Chrome Brick. (See Brick, Fire-) Chrome Steel, 516-522 Composition, 516-517 CorrodibiUty, 521-522 Critical Points, 517-518 Hardness, 517, 522 Manufacture, 522 Structure and Constitution, 517-518 Tensile Properties, 518-521 Carbon Content and, 519-521 Chromium Content and, 518-519 Heat Treatment and, 519-521 Toughness, 517, 522 Uses, 522 Chromium-Manganese Steels, 536 Clay Products. (See Bricks, Terra Cotta, Tile, or Pipe) Clays, Brick-, 223-224, 238 664 INDEX Clays, Shales, and Slates. (See Portland Cement, manufacture of) Coin Bronze, 585 Coke Furnace, 366-367 Coke Pig Iron. (See Pig Iron) Cold-rolled Axles, 443 Cold-twisted Bars, 443, 447 Common Brick, 234 Composition, Chemical. (See Material in Question) Compressive Strength. (See Material in Question) Concrete, 172-201 Adhesion of, to Steel, 199 Alkali, Effect of, on, 188 Bonding to old work, 182 Coefficient of Expansion, 198 Compressive Strength, 191-192 Consistency, Proper, 182 Contraction, 198 Corrosion of Steel in, 200 Deformation, 196 Deposition of, 179-184 Deposition in Forms, 181 Deposition under water, 183-184 Bags, Buckets, or Tremie, Use of, 184 Ec/Es , 199 Elastic Limit, 197 Elastic Properties, 196 Expansion, 198 Facing of, 183 • Factor of Safety, 200-201 Fire-Resistant Properties, 199-200 Forms, 179 Laitance, Formation of, 182, 183, 184 Making, 176-191 Mixing, 177-179 Hand Mixing, 178 Hand vs. Machine Mixing, 177 Machine Mixers and Machine Mix- ing, 178, 179 Modulus of Elasticity, 196 Proportioning, 176, 177 Importance of proper, 176 Ingredients per cu. yd., 177 Practice in, 176, 177 Theory of, 176 Puddling or ramming, 182 Shearing Strength, 195 Stress-Strain Curves, 197-198 Subaqueous, 183-184 Tensile Strength, 192-194 Transportation to Forms, 181 Transverse Strength, 194-195 Water-proofing Compounds, 189-191 Water-proof Material, Layers of, 190 Concrete — Contimted Water-tight Construction, 188-191 Proportioning Mix, 188 Surface Treatments, 190-191 Thickness Required, 189 Weight of, 198 Working Stresses, 200-201 Concrete Beams, Deflection of, 196 Concrete in Sea Water, 186-188 Action of Sea Water, 186-188 Injury, Methods of Prevention of, 187-188 Barium Chloride, Use of, 187 Dense Outer Layer, 187 Magnesium Fluosilicate, Use of, 188 Sesquicarbonate of Ammonia, Use of, 188 Salts in Sea Water, 187 Concrete Making in Freezing Weather, 184-186 Cold, Effect of, upon, 184^185 Methods Employed, 185-186 Concrete Materials, 173-176 Cement, 173 Sand, 173-174 Foreign Matter in, 173-174 Granulometric Composition, 173-174 Shape of Grains, 173 Voids in, 174 Stone or Gravel, 174-176 Crushing and Screening, 174^175 Gravel vs. Broken Stone, 174 Mechanical Analysis, 175 Size and Shape of Fragments, 175 Voids in, 175-176 Conifers. (See Trees) Constitution. (See Material in Question) Contraction of Concrete, 198 Contraction of Mortars. (See Lime, Portland Cement, etc.) Converter. (See Steel, Bessemer) Copper, 543-558 Blister, 552 Casting, 544 Classification, Commercial, 544 Electrical Resistivity, 555-556 Electrolytic, 544 Extraction from Ores, 546-553 Conversion of Matte to Blister or Coarse Copper, 551-552 Bessemer Converter, 551-562 Pyrite Smelting, 552-553 Roasting, 546-548 Heap or Stall, 647 Object of, 647 INDEX 665 Copper — Continued Extraction from Ores — Continued Roasting — Continued Reverberatory Furnace, 547-548 Smelting, 548-551 Blast Furnace, 549-550 Matte, 549, 550, 551 Object of, 548 Reverberatory Furnace, 550-551 Welch Process, 553 Wet Methods, 554-555 Lake, 544, 554 Ores of, 544-546 Production Statistics, 558 Properties, 555-558 Refining Blister, 553-554 Electrolytic, 554 Fire, 553 Uses, 557-558 Copper-Aluminum Alloys, 595-597 Aluminumi Bronze, 595-597 Composition of, 595, 597 Properties of, 595-597 Copper-Tin-Lead Alloys, 587 Copper-Manganese Alloys, 597 Copper-Phosphorus Alloys, 597 Copper Sheet or Plate, 555, 557, 558 Copper-Silicon Alloys, 597 Copper-Tin Alloys. (See Bronzes) Copper-Tin-Zinc Alloys, 585-587 Copper Wire, 555, 565-557, 558 Copper-Zinc Alloys. (See Brasses) Copper-Zinc-Aluminum Alloys, 592 Copper-Zinc-Iron Alloys, 694 Copper-Zinc-Lead Alloys, 592 Copper-Zinc-Manganese Alloys, 593 Corrosion. (See Steel, Physical Proper- ties) Corrosion of Steel in Concrete, 200 Core Loss. (See Steel, Magnetic Prop- erties) Creosote. (See Timber, Preservation of) Critical Temperatures. (See Steel) Crucible Steel, 294, 361, 366-371. (See also Steel) Decay of Wood, 650-651 Delta Metal, 594 Deposition of Concrete, 179-184 Direct Processes. (See Pig Iron, Manu- facture) Dolomitic Lime. (See Quicklimes and Hydrated Lime) Drain Tile, 247-248 - Dry-Sand Molds. (See Cast Iron, Molds and Molding) Duplex Steel, 361, 398 Durability of Wood, 650-651 Dynamite, 205-206 Eckel's Rule, 99-101 Eddy Current Loss. (See Steel, Mag- netic Properties) Elastic Limit. (See Material in Question) Electric Furnace Smelting, 288-291 Electric Furnaces, 288-290 Electric Steel, 294, 261, 399-407 Electrolytic Theory of Corrosion, 489-491 Enameled Brick, 234 Endogens. (See Trees) Engine-bolt Iron, 352 Eutectic. (See Steel, Structure and Con- stitution) Eutectoid. (See Steel, Structure and Constitution) Exogens. (See Trees) Expansion of Concrete, 198 Expansion of Mortars. (See Lime, also Portland Cement) Explosives, 205-206 Eye-bar Steel, 467 Face Brick, 234 Facing of Concrete Walls, 183 Factor of Safety. (See Material in Ques- tion) Fatigue. (See Steel, Physical Properties) Ferrite. (See Steel, Structure and Con- stitution; also Cast Iron, Constitu- tion of; and Wrought Iron, Consti- tution) Fettling, 345 Fiber-Saturation Point. (See Woods, Mechanical Properties) Fineness. (See Material in Question) Fire-box Steel, 443, 467 Firebricks, 248-251 Fire-clay Brick, 241-242 Fireproofing, 246 Fire Resistance of Concrete, 199-200 Fire Resistance of Stones, 216-217 Floor Tile, 247 Flooring Plaster, 2 Flux (Blast Furnace), 264-265, 280, 281 Limestone, 264-265 Necessity for, 264 Oyster Shells, 264-265 Requirement, 280, 281 Forging Steel. (See Steel, Mechanical Treatment) Forgings, 443. (See Steel) Founding. (See Cast Iron, Manufacture) 666 INDEX Foundry Pig Iron, 291, 296 Freezing Weather, Concreting in, 184-186 Frost Resistance of Stones, 216 Fuel (Blast Furnace), 265-271, 280, 281, 282 Charcoal, 266, 270, 271 Manufacture, 270-271 By-products of, 271 Classes of, 265 Coal, 266, 271 Coke, 266-270, 271 Manufacture, 266-270 Beehive, 266-267 By-product or Retort, 267-270 Otto-Hoffmann, 268-270 Semet-Solvay, 267-268 By-products, 268 Ovens, 267, 268 Relative Use of Different Fuels, 271 Fungi. (See Timber, Durability of) Gamma-Iron. (See Steel, Structure and Constitution) Canister Brick. (See Bricks, Fire-) Gayley Dry Air Blast, 275-276 German Silver, 595, 603 Girod Furnace, 401-403 Glazed Brick, 234 Gneiss. (See Stones, Building), 210 Grain. (See Wood) Granite, 207 Graphite. (See Cast Iron, Constitution of) Graphite. (See Steel, Structure and Con- stitution) Grappier Cement, 63, 65, 66 Gray Cast Iron, 315, 316, 317 Gravel for Concrete, 174-176 Green-Sand Molds. (See Cast Iron, Molds and Molding) Gun Metal, 584 Gun Powder, 205 Gypsum Plasters, 1-18 Accelerators, Use of, 7, 8 Adhesive Strength, 14^15 Applications of (Non-Structural), 2, 18 Cement Plaster, 1 Classification of, 1 Compressive Strength, 13, 14 Definition, 1 Flooring Plaster, 2 Hard Finish Plaster, 2 Hard Wall Plaster, 1 Manufacture, 2, 8 Patent Plaster, 1 Plaster of Paris, 1 Plasticity, 16 Gypsum Plasters — Continued Production Statistics, 18 Properties and Uses, 8, 17 Relative Applicability of Various, to Structural Uses, 15, 17 Retarders, Use of, 7, 8, 14 Setting and Hardening, 8 Structural Uses, 2, 18 Tensile Strength, 9-13 vs. Lime Plasters, 16-17 Gypsum Rocks, 2 Hard Finish Plaster, 2 Hard Wall Plaster, 1 Hardening Steel. (See Steel, Heat Treat- ment) Hardness Tests. (See Steel, Physical Properties) Heartwood. (See Wood) Heat Treatments. (See Steel) H^roult Furnace. (See Steel, Electric) High-Calcium Lime. (See Quicklimes and Hydrated Lime) High-Speed Steel. (See Tungsten and Tungsten Chromium Steel) Hollow Brick. (See Bricks, Building) Hot Blast Stoves, 274 Hydrated Lime, 60-58 Alca Lime, 56-58 Chemical Composition, 31 Compressive Strength, 43, 54-55 Definition, 50 Dolomitic, 53-55 High Alumina, 56-58 High Calcium, 53-55 Hydration of, 50-51 Hydrators for, 51 Manufacture, 50-53 Plasticity, 53-56 Production Statistics, 58 Properties and Uses, 53-56 Sand-carrying Capacity, 53-56 Tensile Strength, 41, 53-64 Uses, 55-66, 68 vs. Quicklimes, 64, 55-66 Yield, 53, 56 Hydraulic Cementing Materials, 59-62 ' Classification of, 60, 61, 62 Hydraulic Index, 59-60 Hydraulic Limes, 59-69 Chemical Composition, 65-66 Classification, 63 Compressive Strength, 67-69 Definition, 62 Eminently Hydraulic Limes, 63 66 68- 69 INDEX 667 Hydraulic Limes — Continued Feebly hydraulic Limes, 63, 66, 67 Grappier Cement, 63, 65, 66 Limestones for, 63-64 Manufacture, 63-65 Non-staining Cements, 63, 65. Properties, 65-69 Selenitic Lime, 63, 65, 67 Tensile Strength, 67-69 Uses, 63, 69 Hydraulicity Defined, 59 Hydrogen Peroxide Theory of Corrosion, 487 Hysteresis. (See Steel, Magnetic Prop- erties). Impact Tests. (See Steel, Physical Prop- erties) Induction Furnace, 400, 403-406 Ingotism, 418 Ingots, 417-419 Iron, Historical Development of Use of, 257-258 Iron. (See also Pig Iron, Cast Iron, Malleable Cast Iron, and Wrought Iron) Iron Ores, 259-264 Bessemer, 264 Calcination of, 262 Classes of, 259-260 Hematite, 260 Iron Carbonate, 260 Limonite, 260 Magnetite, 260 Composition of, 259-260 Concentration of, 263 Dry, 263 Magnetic, 263 Wet, 263 Gangue, 259 Grades of, 264 Mining of, 261 Non-Bessemer, 264 Production Statistics, 260-261 Reduction of, 271-291 Roasting of, 262 Smelting, 271-291 Transportation of, 261-262 Keller Furnace, 403 Kiln. (See Lime; Bricks; Natural Cement; Portland Cement, Manufacture) Kjellin Furnace, 403-404 Knobbled Iron, 351 Knots. (See Wood) Laitance, 182, 183, 184 Lead, 566-571 Antimonial, 567 Binary Alloys of, 598 Commercial Forms of, 566-567 Desilverized, 566. Extraction from Ores, 568-570 ■ Conversion of Matte, 570 Dust Recovery, 570 Roasting Ores, 568-569 Sintering Ores, 568-569 Smelting, 569-570 Litharge, 567 Ores, 567 Oxide, 567 Production Statistics, 571 Properties, 571 Refining Crude, 570-571 Desilverization, 571 Soft, 566 Uses, 566, 571 White, 567 Lead- Antimony Alloys, (Type Metal), 598 Lead-Arsenic AUoys, 598 Lead-Bismuth Alloys, 598 Lead-Cadimum Alloys, 598 Lead-Tin Alloys, (Plumber's Solder and Pewter), 598 Lime. (See Quicklime, Hydrated Lime, and Hydraulic Lime) Lime Kilns, 25-27 Lime Mortars. (See Quicklimes and Hy- drated Lime) Lime Plaster. (See Quicklimes and Hy- drated Lime) Limestones. (See Portland Cement Manu- facture; also Stones, Building) Loam Molds. (See Cast Iron, Molds and Molding) Locomotive Steel, 439, 442, 466 Low-Carbon Steel, 294, 361 Machinery Bronze, 584 Machinery Steel, 362, 439 Magnesia Brick, 243 Magnesian Lime. (See Quicklimes and Hydrated Lime) Magnetic Properties. (See Steel, Physical Properties) Malleable Cast Iron, 294, 333-340 Annealing, Process, 336-338 Annealing Ovens, 337 Annealing Pots, 337 Annealing Temperature, 338 Black-heart Process, 338 Carbon, Change of Form of, 336 668 INDEX Malleable Cast Iron — Continued Annealing, Process — Continued Case-hardening, 336, 340 Cooling, Annealed Castings, 338 Decarburization, Depth of, 338 Packing Castings, 336 Packing Materials, 336, 337 Ternper Carbon, 336, 338 Time Required, 338 White Heart Process, 338 Cleaning Castings, 336, 338 Composition, 338 Constitution, 339 Definition, 333 Furnaces for Melting, 334 Inspection of Hard Castings, 336 Inspection and Testing of Castings, 338 Manufacture, 333-338 Materials Used in Manufacture, 333- 334 Cast-iron Scrap, 334 ' Ferro-Silicon, 334 Malleable Scrap, 334 Pig Iron, 334 Sprues or Hard Scrap, 334 Steel Scrap, 334 Wrought-iron Scrap, 334 Melting, 335 Cupola Process, 335 Fuel Requirements, 335 Open-hearth Process, 335 Reverberatory Furnace Process, 335 Molding Methods, 335, 336 Chills, 336 Gating, 335 Molds, 335 Risers and Feeders, 335 Shrinkage, 336 Pouring the Castings, 337 Burnt Iron, 337 Chilled Iron, 337 Temperature, 337 Properties and Uses, 338-340 Tensile Strength, 339 Elongation, 339 Transverse Strength, 339 Breaking Load, 339 Deflection, 339 Modulus of Rupture, 339 Toughness, Shock Resistance, 339-340 Modulus of Resilience, 339-340 Uses, 333, 340 Malleable Pig Iron, 291 Manganese Bronze, 588 Manganese in Cast Iron, 320 Manganese Steel, 513-516 Manganese Steel — Continued Composition, 513 Hardness, 513, 516 Manufacture, 516 Structure and Constitution, 513-514 Tensile Properties, 514r-516 Carbon Content and, 514-516 DuctiUty, 515-516 Manganese Content and, 514r-516 Toughness, 513, 516 Uses, 516 Manganese-Silicon Steels, 537 Manufacture. (See Material in Question) Marble, 210 Marl, 96, 115-117 Martensite. (See Steel, Heat Treatment) Mechanical Properties. (See Material in Question) Medium Steel, 294, 362 Medullary Rays, 611-612 Melting Hole, 366-367 Merchant Bars, 348 Modulus of Elasticity. (See Material in Question) Modulus of Rupture. (See Material in Question) Modulus of Shock Resistance. (See Cast Iron, also Malleable Iron, Cross- breaking Strength) Molds and Molding. (See Cast Iron) Molybdenum Steel, 526-530 Alternating Stress Resistance, 529 Composition, 527 Critical Points, 527-528 Ductility, Cold-bending, 528-529 Hardness, 526, 529 Manufacture, 529 Structure and Constitution, 627-528 Tensile Properties, 528-529 Carbon Content and, 528-529 Molybdenum Content and, 528-529 Tungsten vs., 526, 529 Uses, 530 Monell Process, 399 Mortar. (See Gypsum Plasters, Quick- limes, Natural Cement, Portland Cement, etc.) Mottled Cast Iron, 315, 318 Muck-bar, 348 Muck-bar Iron, 352 Muntz Metal, 591 Natural Cements, 79-89 Calcination, 81-82 Chemical Composition, 84 CUnker, 82 INDEX 669 Natural Cements— CorUimied, Compressive Strength, 86 Definition, 79 Distinction between, and Portland Cement, 79 Fineness of Grinding, 85 Free Lime in, 82 Grinding and Grinding Machinery, 82-83 KUns, 81-82 Manufacture, 80-84 Manufacturing Costs, 83 Modulus of Elasticity, 87 Natural Cement Rocks, 80 Production Statistics, 87 Properties, 84-87 Status of Industry, 80, 88-89 Tensile Strength, 85^86 Time of Settmg, 84-85 Uses, 80, 87, 89 Neutral Bricks, 243 Newberry's Rule, 99 Nickel, 578-581 Commercial Forms, 578 Consumption Statistics, 581 Extraction from Ores, 579-580 Electrolytic Methods, 580 Extraction of, from Silicate, 580 Smelting of. Sulphur Compounds, 579-580 Ores, 578-579 Production Statistics, 581 Properties, 580-581 Uses, 578, 580-681 Nickel Alloys, 603 Nickel Bronze, 588 Nickel-Chromium Steels, 537 Nickel-Manganese Steels, 536 Nickel-Silicon Steels, 538 Nickel Steel, 501-512 Alternating Stress Resistance, 501 Composition, 501 . Corrodibility, 511-512 Critical Points, 501 Impact Strength, 608-509 "Invar", 501 Irreversibility, 601 Magnetic Preperties, 509-511 Manufacture, 512 "Platinite,"501 Tensile Properties, 501-508 Carbon Content and, 501-508 Heat Treatment and, 505-508 High-Carbon Steels, 504-605 Low-Carbon Steels, 503 Medium-Carbon Steels, 604 Nickel Steel-— Continued Tensile Properties — Continued Nickel Content and, 601-508 Uses, 512 Nickel-Tungsten Steels, 535 Nickel- Vanadium Steels, 535 Nitro-Glycerine, 205 Non-ferrous Alloys, 641, 542-543, 581-604 Classification, 542-543 Most Important, 541 Nature of, 581-582 Non-ferrous Metals, 541-581 Classification, 542-543 Most Important, 641 Non-staining Cements, 63, 65 Onyx Marble, 210 Open-arc Furnaces, 400-401 Open-hearth Furnace, 386-391 Open-hearth Steel. (See Steel) Ornamental Brick, 234 Ores, 269-264, 644-646, 663, 667, 672. 574, 578 Patent Plaster, 1 Paving Brick. (See Bricks, Paving-) Pearhte. (See Cast Iron, Constitution; also Steel, Constitution, Heat Treat- ment; and Wrought Iron, Constitu- tion) Permeability. (See Portland Cement Mor- tars) PermeabiUty. (See Steel, Magnetic Prop- erties) Phosphor Bronze, 587 Phosphorus Brass, 594 Phosphorus in Cast Iron, 319-320 Pig Iron, 257-293 Anthracite, 291 Basic, 291 Bessemer, 291 Charcoal, 291 Classes of, 291 Coke, 291 Foundry, 291, 296 Malleable, 291 Manufacture, 269-291 Blast Furnace Smelting, 271-287 Action within Furnace, 282-285 Blast Furnace, 272-274 Blowing Engines, 275 Burdening the Furnace, 280-282 Carburization of the Iron, 277 Casting the Pigs, 285-286 Charging Mechanism, 274 Chemical Reactions, 283-285 670 INDEX Pig Iron — Continued Manufacture — Continued Blast Furnace Smelting — Continued Construction of Furnace, 272-274 Conversion of Gangue to Slag, 278 Deoxidation of Ore, 277 Deoxidizing Agencies, Strength of, 280 Function of Blast Furnace, 277 Hearth Temperature, Control of, 280 Heat Development, 282-283 Hot Blast Stoves, 274 Mechanical Control, 279 Melting the Iron, 277-278 Metallurgical Control, 280-282 Operation of Blast Furnace, 278- 285 Products Classified, 294^295 Products, Handhng the, 285-287 Refrigeration of Blast, 275-276 Separation of Iron and Slag, 278 Slag Composition, Control of, 282 Slag Handling, 287 Solids and Gases, 282 Starting the Operation, 278 Direct Methods of Smelting, 267-258 Catalan Forge, 257 Electric Furnace Smelting, 288-291 Arc Furnace, 288 Carbon Requirement, 288, 290 Cost of, 288 Economic Status, 288, 290 Electric Furnaces, 288-290 Heat Distribution, 290 Induction Furnace, 288 Product, Quality of, 290-291 Resistance Furnace, 288-290 Raw Materials, 259-271 Flux. (See Flux, Blast Furnace) Fuel. (See Fuel, Blast Furnace) Ores. (See Iron Ores) Production Statistics, 293 Remelting. (See Cast Iron) Uses, 292 Pinchbeck, 591 Pipe, Sewer, 248 Pipes, 414 Piping, 418 Pith, 607-608 Pith Flecks, 613-614 Pith Rays, 611-612 Plaster of Paris, 1 Plaster Calciners, 4-6 Plates, Rolled Steel, 413 Portland Cement, 90-171 Portland Cement — Continued Abrasive Resistance, 168 Adhesive Strength, 166-168 Brick, Adhesion to, 167 Steel, Adhesion to, 166 Various Materials, Adhesion to, 168 Composition, 126-129 Compressive Strength, 148, 156-158, 164-165 Increase in. Rate of, 161-162 Retrogression in, 161-163 Significance of, 148, 156-158, 164 Specifications, 164 Tensile vs., 148, 156-158 Tests, 165 Constitution, 129-132 Cost, 88 Definition, 92, 93 Distinction between Natural and, 79 Fineness of Grinding, 140-143 Sand-carryiag Capacity, Effect upon, 142 Setting Time, Effect upon, 141 Significance of, 140 Soundness, Effect upon, 141 Specifications, 141 Strength, Effect upon, 142-143 Tests, 141 Invention and Development, 90, 91 Limestones, 95 Manufacture, 94-125 Addition of Retarder, 114 Burning the Mixture, 109-114 Calcination, Theory of, 101-102 Calculating the Mixture, 99-101 Changes in Composition during Calcination, 113 Control during Operation, 101 Coohng the CUnker, 114 Cost, 125 Crushing the Rock, 104 Dry Process, 102-115 Drying the Rock, 104 Fineness of Grinding, 108-109 Fuel Used, 110-112 Grinding the Rock, 105-106 Kilns. (See Mill Equipment) Mill Equipment, 119-125 Mixing, 107 Packing, 115 Preliminary Treatment, 101 Proportioning Mixture, Machine for. 107 Pulverizing the Rock, 107-109 Raw Materials, 94-98 Alkali Waste, 98 INDEX 671 Portland Cement — Continued Manufacture — Continued Raw Materials — Continued Blast Furnace Slag, 98 Cement Rock, 96 Classification of, 94 Clays, 97 Combinations of, 95 Excavation, 102-103 Limestones, 95 Marl, 96 Proportioning, 98-101, 107 Shales and Slates, 97 Semi-dry Process, 118 Storing Cement, 115 Wet Process, 115-119 Raw Materials, 115 Use of Chalk and Clay, 117-118 Use of Marl and Clay, 115-117 Modulus of Elasticity, 165 Mortars, 148, 160-156, 157, 158, 159- 163, 164, 165-171 Absorptive Properties, 168-169 Contraction, 169-171 Expansion, 169-171 Hydrated Lime in, 155-156 Permeability, 168-169 Sands for, 151-155 Physical and Mechanical Properties, 126-171 Production Statistics, 88, 126 Proportioning of, 98-101 Eckel's Rule, 99-101 Newberry's Rule, 99 Setting and Hardening, 132-137 Shearing Strength, 165-166 Soimdness, 145-148 Fineness, Effect of, upon, 147 Seasoning, Effect of, upon, 147 Significance, 145-146 Specifications, 146-147 Sulphates, Effect of, upon, 147 Tensile Strength vs., 159 Specific Gravity, 137-140 Adulteration, Effect of, upon, 138- 139 Seasoning, Effect of, upon, 139-140 Significance, 137 Specifications, 138 Tests, 138 Thoroughness of Burning, Effect of, upon, 138 Tensile Strength, 148-164 Compressive vs., 148, 156-158 Increase in, Rate of, 160-163 Mortars, 150-156 Portland Cement — Continued Tensile Strength — Continued Mortars — Continued Density of. Relation of, to, 153 Lime, Hydrated, Effect of, Addi- tion of, upon, 155-156 Sand, Effect of Cleanliness of, upon, 154 Sand> Effect of Mica in, upon, 153 Sand, Standard, for Tests, 151 Significance, 148, 150 Specifications, 151 Neat, 148-150 Fineness of Grinding, Effect of, upon, 150 Lime Proportion, Effect of, upon, 149-150 Significance, 148 Specification, 149 Temperature of Burning, Effect of, upon, 150 Tests, 149 Neat vs. Mortar, 159-160 Retrogression in, 160-163 Soundness vs., 159 Time of Setting vs., 143-145 Fineness, Effect of, upon, 141 Seasoning, Effect of, upon, 144 Significance, 143 Sulphates, Influence of, upon, 144 Temperature, Influence of, upon, 143 Tests, 143 Water, Effect of per cent of, upon, 143 Preservation of Timber. (See Timber) Pressed Brick, 234 Primary Wood. (See Wood) Puddled Bloom Iron, 352 Puddled Iron, 352 Puddling Furnace, 344-345, 349 Puddling Process, 343-349 Puzzolan Cements, 70-74 Compressive Strength, 73-74 Definition, 70 Manufacture, 72-73 Natural Puzzolanic Materials, 71-72 Properties and Uses, 73-74 Puzzuolana, 71, 72 Santorin, 71, 72 Tensile Strength, 73-74 Trass, 71, 72 Tuff or Tufa, 71 Puzzuolana. (See Puzzolan Cements) Quarrying Methods, 204-206 672 INDEX Quicklimes, 19-49 Applications (Non-Structural), 20-21, 49 Calcium, 20 Classification of, 19-20, 30 Composition, 30-31 Compressive Strength (Mortars), 42-47 Definition, 19 Dolomitic, 20, 21, 23, 24, 31-49 Hardening, 35-36 Hardness, 38, 44 High Calcium, 19, 20, 21, 23, 24, 31-49 Hydration, 31-35, 45 Lump, 19 Magnesian, 20, 21, 23, 24, 31-39 Manufacture, 22-30 Plasticity, 36-37, 44, 45 Production Statistics, 49 Properties, 30-49 Impurities, Effect of, upon, 43-45 Sand, Effect of Character of, upon, 46 Size of Test Specimens, Effect of, upon, 47-48 ( Slaking, Effect of Method of, upon, 45-46 Temperature of Calcination, Effect of, upon, 45 Pulverized, 19 Run-of-kiln, 19 Sand-carrying Capacity, 37, 44 Selected, 19 Setting, 35-36 Shrinkage, 39 Slaking, 31-35, 45 Structural Uses, 21, 48, 49 Tensile Strength (Mortars), 39-42, 44, 45 Time of Setting, 39 vs. Gypsum Plasters, 16-17 Waste, 38, 45 Yield, 37-38, 44 Rail Steel, 362, 439 Rails, Rolling of, 412 Rays, Pith or Medullary, 611-612 Red Shortness, 319, 453 Refined Bar Iron, 362 Refrigeration (Air Blast), 275-276 Regenerative Furnace, 367-368, 386-391 Reheating Furnaces, 407-408 Reinforcing Bars, 443, 467 Resin Ducts, 612 Resistance Furnace, 288-290 Retarder (Portland Cement), 114 Rivet Steel, 439, 466 Rochling-Rodenhauser Furnace, 404-406 Rocks, Classification of, 202-203 Rods, Rolling of. 348, 409-414 Rolling Mill Operations, 348, 409-414 Roman Tile, 235 Roofing Tile, 246-247 Rust, Nature of, 486 Sand for Concrete and Mortars, 151-155, 173-174 Sand-Lime Brick. (See Bricks) Sandstones, 210-213 Santorin, 71-72 Sapwood. (See Wood) Seasoning Timber, 626-628, 651 Sea Water, Effect on Concrete, 186-187 Segregation, 322, 418-419 Selenite, 3 Selenitic Lime, 63, 65, 67 Sewer Pipe, 248 Shearing Strength. (See Material in Question) Shrinkage, Cast Iron, 320, 321 Shrinkage, Lime Mortars, 39 Siemens-Martin Process. (See Steel, Open- hearth) Silica Brick, 242 Silicon Bronze, 588 Silicon in Cast Iron, 318-319 Silicon Steel, 530-531 Composition, 530 Hadfield's, 530 Hardness, 530, 531 Heat Treatment, 530 Magnetic Properties, 530 Manufacture, 531 Tensile Properties, 530, 531 Silicon Content and, 530, 531 Heat Treatment and, 530, 531 Uses, 531 Slag (Blast Furnace), 74, 75, 98, 282, 287 Slag Cements, 71, 74-78 Chemical Composition, 77 Compressive Strength, 78 Definition, 74 Hydrated Lime for, 76 Manufacture, 75-77 Properties and Uses, 75, 77-78 Slag Composition, 74 Slag Granulation, 74, 75 Tensile Strength, 77 Slag in Wrought Iron, 350 Slates, 213 Smelting. (See Pig Iron, Manufacture) Solid Solutions, 422 Sorbite, 431, 432 INDEX 673 Soundness (Portland Cement), 145-148 Special Steels. (See Alloy Steels) Specific Gravity. (See Material in Ques- tion) Specification. (See Material in Question) Speculum Metal, 585 Spelter. (See Zinc) Spring Steel, 439 Springwood. (See Wood) Stassano Furnace, 400-401 Statuary Bronze, 585 Staybolt Iron, 352 Stead's Brittleness, 439 Steam Hammer Forging, 408, 415-416, 417 Steel, 360-499 Adhesion of Cement and Mortars to, 166-167 Alloy. (See Alloy Steels) Annealing. (See Heat Treatment) Effect on Strength. (See Physical Properties) Bessemer, Manufacture, 361, 371-383 Acid Process, 371-380 Blast Pressure and Volume, 374 Blowing Engines, 374 Casting the Ingots, 372, 378-379 Chemistry of Process, 376-377 Converter Construction, 372, 373- 374 Cupolas, 372, 374 Deoxidation of Blown Metal, 378 Essential Features of Process, 372-373 Heat Development and Utihzar tion, 377 Historical, 371-372 Invention of, 371 Ladles, 372, 375, 379 Mixer, 372, 375 Operation of Converter, 375-376 Pig Iron Used, 373 Recarburizers and Recarburization, 372, 377-378 Rolling Mills, 373, 407-414 Basic Process, 380-383 Basic Converter Lining, 380 Basic vs. Acid Process, 382-383 Converter Construction, 380 Heat Development, 382 Historical, 380 Limitations of, 380 Operation of Converter, 381-382 Pig Iron Used, 380-381 Recarburizers and Recarburiza* tion, 382 Slag Removal, 382 Steel — Continued Blister. (See Cementation) Boiler Plate, 362, 439, 443, 467 Boiler Rivet, 362, 439, 443, 467. Bridge Structural, 439, 442, 466 Building Structural, 439, 442, 466 Carbon Content and Strength. (See Physical Properties) Castings, 443, 467 Cementation, Manufacture, 294, 361, 363-365 Blisters, Cause of, 365 Carburization, Rate of, 365 Cementation Furnace, 364 Charge, 364 Grades of Steel Produced, 365 Operation of Process, 364-365 Temperature Required, 365 Time Required, 365 Wrought Iron Used, 364 Classification of Steels, 294, 361, 362,439 Compressive Strength. (See Physical Properties) Constitution. (See Structure and Con- stitution) Corrosion of Iron and. (See Physical Properties) Corrosion of, in Concrete, 200 Critical Temperatures, 425-428 Crucible, Manufacture, 294, 361, 366- 371 Alloys Used, 369 Carburizing Agent, 369 Charge, 368-369 Coke Furnace or Melting Hole, 366- 367 Composition of Product, 371 Crucibles, 367-368 , Fuel, 366, 368 Gas-fired Regenerative Furnace, 367- 368 Operation of Process, 369-370 Time Required, 369-370 Tools, Grade of. Required for Vari- ous, 371 Crystalline Structure, Refinement of, 415, 416, 419, 438, 439 Defects in Ingots. (See Steel Ingots) Definition, 360-361 Ductility. (See Physical Properties) Duplex, Manufacture, 361, 398 Advantages, 398 Description of Duplex Processes, 398 Elastic Limit. (See Physical Properties) Electric, Manufacture, 294, 361, 399- 407 674 INDEX Steel — Continued Electric, Manufacture — Contimted Applications and Limitations, 406- 407 Alloy Steels, Production of, 406- 407 Bessemer, Competition with, 406 Crucible, Competition with, 406 Duplex Processes, Use in, 401, 406 Open-hearth, Competition with, 406 Steel Castings, Production of, 401, 406 Tool Steel, Production of, 404, 406 Electric Processes in General, 399 Furnaces and Operation of Processes, 400-407 Arc-Resistance Furnaces, 400, 401- 403 Girod Furnace, 401-403 H6roult Furnace, 401 Induction Furnaces, 400, 403-406 Keller Furnace, 403 KjeUin Furnace, 403-404 Open-arc Furnaces, 400-401 Rochling-Rodenhauser Furnace, 404-406 Stassano Furnace, 400-401 Elongation. (See Physical Properties) Eye-bar, 467 Fatigue. (See Physical Properties) Fire-box, 443, 467 Forgings, 443 Gauged Length. (See Physical Prop- erties) Grades of, 439 Grain-refinement of, 415, 416, 419, 438-439 Hardening. (See Heat Treatment) Hardness. (See Physical Properties) Heat Treatment of, 419-422, 430-439, 449-453 General Considerations, 419, 430 Practice, 419-422 Annealing, 420-421 Grain Size, Refinement of, 419 Internal Strains, Relieving, 419 Purpose of, 419 Softening after Hardening, 419 Case Hardening, 421-422 ' Applications of, 421 Carbonation, Depth of, 421 Carbonizing Material, 421 Hardening Treatment, 421 Heating, 421 Steel Used, 421 Steel — Continued Heat Treatment of — Continued Practice — Continued Case Hardening — Continysd Temperatures Employed, 421 Time Required, 421 Hardening, 419 Heating, 419 Quenching, 419 Tempering, 419-420 Purpose of, 419 Temper Colors, 420 Tempering Heats, 420 Strength and Ductility, Effect upon, 449-452 Annealing, 449 Hardening, 449-452 Tempering, 449-452 Structure and Constitution, Effect upon, 430-439 Annealing, 437-439 Coarse Crystallization, Cause of, 438 Grain Size, Refinement of, 438- 439 Internal Strains, 437-438 Softening after Hardening, 439 Stead's Brittleness, 439 Hardening, 430-435 Allotropic Theory of, 434-435 Austenite, 430, 431, 432 Beta-Iron Theory of, 434-435 Carbon Content and, 434 Cementite, 431, 432, 433 Eutectoid Steel, 432, 433 Ferrite, 431, 432, 433 Hyper-eutectoid Steel, 432, 433 Hypo-eutectoid Steel, 432, 433- Martensite, 430, 431, 432 Pearlite, 431, 432 Sorbite, 431, 432 Transformations, 432, 433 Tempering, 435-437 Austenitic Steel, 435, 436 Martensitic Steel, 435, 436 Temperatures, 435, 436, 437 Time Factor in, 437 Troostitic Steel, 435, 436 Troosto-Martensitic Steel, 435, 436 Troosto-Sorbitic Steel, 435, 436 High Carbon, 294, 362 High-speed. (See Tungsten Steel) Impact Strength. (See Physical Prop- erties) Ingots. (See Steel Ingots) INDEX 675 Sieel-^CorUimted Invar. (See Nickel Steel) • Locomotive, 439, 442 466 Low Carbon, 294, 361 Machinery, 362, 439 Magnetic Properties. (See Physical Properties) Manganese. (See Manganese Steel) Manganese and Strength. (See Physical Properties) Mechanical and Thermal Treatment of, 407-417 Forging by Presses, 416-417 ■ Forging under Steam Hammer, 408, 415-416, 417 Reheating, 407-408 Billet, 408 Ingot, 407-408 Rolling, 408-414 Action of Rolls, 408-409 RoUing Mills, 409-412 Rolling Practice, 412-414 Cold RoUing, 414 Pipes, 414 Plates, 413 Rails, 412 Rods, 413 Structural Sections, 413 Tubes, Seamless, 414 Wire-drawing, 413-414 Medium, 294, 362 Modulus of Elasticity. (See Physical Properties) Molybdenum. (See Alloy Steels) Monell Process, 399 Nickel. (See Nickel Steel) Open-hearth, Manufacture, 294, 361, 383-399 Acid Process, 396-397 Charge of Furnace, 396 Chemistry of Process, 396-397 Deoxidation of Product, 397 Ferro-Manganese, 397 Ferro-SiUcon, 397 Ore Addition, 396, 397 Oxidation of Metalloids, 397 Pig Iron Used, 396 Proportions of Charge, 296 Recarburizers and Recarburiza- tion, 397 Scrap Used, 396 Slag, Functions of, 397 Basis Process, 384, 393-396 Acid vs. Basic Process, 384, 393, 396 Charge of Furnace, 393-394 All-Scrap Process, 393 Steel — Continued Open-hearth, Manufacture — Continued Basic Process — Continued Charge of Furnace — Continued Flux Requirement, 394 Mixer, Use of, 394 Order of Charging, 394 Ore Requirement, 393 Oxide Addition, 394 Pig and Ore Process, 393, 394 Pig and Scrap Process, 393, 394 Pig Iron Used, 394 Proportions of Charge, 394 Chemistry of Process, 394^395 Assimilation of Oxides by Slag, 395 Calcium Fluoride Addition, 395 Carbon Addition, 394-395 Deoxidizing Action of Slag, 395 Oxidation of Metalloids, 394 Oxide Addition, 395 Phosphorus Removal, 395 Slag, Functions of, 395 Sulphur Removal, 395 Ingot Casting, 396 Recarburization, 395-396 Coal, Charcoal, or Coke, 395 Extent of, 395 Ferro-Manganese, 395 High Carbon Steel, 395 Low Carbon Steel, 395 Essential Features of Processes, 384 Fuel, 391-393 Natural Gas, 391-392 Producer Gas, 392-393 Furnaces, 386-391 Life of, 390-391 Regenerative, in General, 386 Stationary, 387-389 Tilting or Rolling, 389-390 Tilting and Stationary, Compared, 390 Historical Development of Process, 383-384 Invention of Regenerative Furnace, 383 Invention of Siemens-Martin Process, 383 Monell Process, 399 Plant Equipment, 385-386 Air Valve, 386 Casting Equipment, 386 Charging Equipment, 385-386 Furnaces, 385, 386 Gas Valve, 386 Ingot Molds, 386 676 INDEX Steel — Continued Open-hearth, Manufacture — Continued Talbot Process, 398-399 Advantages, 399 Capacity, 398 Furnace and Operation, 398 Phosphorus and Strength. (See Physical Properties) Physical Properties, 439-498 Alternating Stresses. (See Repetition and Reversal of Stress) Boiler Steel, 439, 443, 467 Boiler Rivet Steel, 439, 443, 467 Carbon Content, Relation of, to, 444-449 Cold-bending. (See Ductility, Cold- bending) Cold-rolled Axles, 443 Cold-twisted Bars, 443, 467 Combined Stresses, 457-459 Compression and Torsion Combined, 458-459 Compressive Strength. (See Tensile and Compressive Strength) Corrosion of Iron and Steel, 486^98 Carbonic Acid Theory, 487 Cold-working and, 494 Corrodibility, Relative, of Various Irons and Steels, 494^498 Corrodibility Tests, 491-498 Electrolysis, 489-490 Electrolytic Dissociation, 488-489 Electrolytic Theory, 487-488, 490- 491 Heat Treatment and, 492-494 Hydrogen Peroxide Theory, 487 Importance of Corrosion Problem, 486 Rust, Nature of, 486 Theories of, 486 Ductility, Cold-bending, 466-467 Cold-bending Tests, Value of, 466 Specifications for, 466-467 Elastic Limit. (See Tensile and Com- pressive Properties) Elongation. (See Tensile and Com- pressive Properties) Eye-bar Steel, 467 Fatigue. (See Repetition and Rever- sal of Stress) Fire-box Steel, 443, 467. Flexure and Torsion Combined, 458- 459 Grades of Steel, 439 Hardness, 459-466 Applications of Tests, 466 Steel — Continued Physical Properties — Continued Hardness — Continued Bauer. Drill Test, 465 Brinell Test, 460-464 Cone Test, 464^65 Norris Dry Rolling Friction test, 465-466 Sceleroscope Test, 465 Tensile Strength and, 462-464 Heat Treatments, Effect of, upon, 449-452 Impact Strength, 467-470 Importance of, 467 Static vs., 469-470 Testing Methods, 468^69 Machinery Steel, 439 Magnetic Properties, 475-486 Chemical Composition and, 479 Cold-bending and, 484-485 Core Loss, 478 Eddy Current Loss, 477-478 Elastic Limit and, 484 Heat Treatment and, 482-484 Hysteresis, 476-477 Importance of, 475, 478 Intensity of Magnetism, 475 Magnetizing Force, 475 Mechanical Properties and, 484r-486 Mechanical Treatment and, 482 Permeability, 475-476 Temperature and, 478-482 Terms and Units, 475 Manganese and Strength. (See Ten- sile and Compressive Properties) Modulus of Elasticity (See Tensile and Compressive Properties) Phosphorus and Strength. (See Ten- sile and Compressive Properties), Rail SpUce Bar Steel, 443, 466 Rail Steel, 439 Reduction of Area. (See Tensile and Compressive Properties) Reinforcing Bars, Concrete, 443, 467 Repetition and Reversal of Stress, 470-475 Endurance of Steels, 473-475 Fatigue Failure, 470-472 Testing Methods, 472-473 Rivet Steel, 439, 466 Shearing Strength, 454-457 Direct Shear, 454-455 Tensile Strength and, 454-455 Torsion, 455-457 Torsion and Direct Shear, Relation. 456 INDEX 677 Steel — Continued Physical Properties— Corrfinwed Shearing Strength — Continued Shock Resistance. (See Impact Strength) SJlicoH and Strength. (See Tensile and Compressive Properties) Specifications, 442-443, 466-477 Spring Steel, 439 Steel Castings, 443, 467 Steel Forgings, 443 Stress-Strain Diagram. (See Tensile and Compressive Properties) Structural Steel for bridges, 439, 442, 466 Structural Steel for Buildings, 439, 442, 466 Structural Steel for Cars, 439, 442, 466 Structural Steel fo^ Locomotives, 439, 442, 466 Structural Steel for Ships, 442, 467 Sulphur and Strength. (See Tensile and Compressive Properties) Tensile and Compressive Properties, 439-454 Behavior under Stress, 440-442 Breaking Strength, 441 Carbon Content and, 444-449 Compressive Strength and Yield Point, 440 Elastic limit, 440 Elongation, 441, 442, 443 and Gauged Length, 448-449 Distribution of, 441 Factors Influencing, 440 General Considerations, 439-440 Heat Treatments, Effect -of, 449- 452 Manganese, Effect of, 454 Modulus of Elasticity, 442 Necking Down, 441-442 Phosphorus, Effect of, 453-454 Reduction of Area, 441, 442, 443 Silicon, Effect of, 453 Specifications for, 442-443 Stress-Strain Diagram, 440-441 Sulphur, Effect of, 453 Tensile Properties, Various Steels, 442-443 Ultimate Strength, 441, 442, 443 Yield Point, 441, 442, 443 Tension and Shear Combined, 458- 459 Tension and Torsion Combined, 458- 459 Steel — Continued Physical Properties — Continued Tool Steel, 439 Torsion. (See Shearing Strength) Transverse Strength, 457 Ultimate Strength. (See Tensile and Compressive Properties) Working Stresses, 498-499 Yield Point. (See Tensile and Com- pressive Properties) Production Statistics, 499 Rail, 362, 439 Red-shortness, 453 Reinforcing Bars, 493, 467 Repetition and Reversal of Stress. (See Physical Properties) Rivet Steel, 439, 466 Shearing Modulus of Elasticity. (See Physical Properties) Shearing Strength . (See Physical Prop- erties) Ship Structural, 443, 467 Shock Resistance. (See Physical Prop- erties, Impact Strength) Silicon. (See Silicon Steel) Silicon and Strength. (See Physical Properties) Softening. (See Heat Treatment) Special Steels. (See Alloy Steels) Spring, 362, 439 Strength. (See Physical Properties) Structural, 362, 439, 442, 466 Structural Rivet, 362, 439, 442, 466 Structure and Constitution, 422-439 a-Iron, 425 AUotropic Forms of Iron, 425 Austenite, 423 0-Iron, 425 7-Iron, 425 Cementite, 422 Changes Below the Freezing-point, 424^429 Constitution, Rapidly Cooled Steels. (See Heat Treatment) Constitution, Slowly Cooled Steels, 429-430 Eutectoid Steel, 429 Hyper-Eutectoid Steels, 429, 430 Hypo-Eutectoid Steels, 430 Cooling Curves, 425-426 Critical Points, Ai, As, A3, 425 Decomposition Diagram, 425-428 Eutectoid Alloy, 427-428 Hyper-Eutectoid Alloys, 427 Hypo-Eutectoid Alloys, 425-427 Equilibrium Diagram, 428-429 678 INDEX Steel — Continued Structure and Constitution — Continued Eutectic, 423 Eutectoid, 423 Ferrite, 422 Freezing of Iron-Carbon Alloys, 423- 424 Hyper-Eutectic Alloys, 424 Hypo-Eutectic Alloys, 423-424 Graphite, 422 Martensite. (See Heat Treatment) Pearlite, 426, 427, 429 Solid Solutions, 422 Sorbite. (See Heat Treatment) Structure, Rapidly Cooled Steels. (See Heat Treatment) Structure, Slowly Cooled Steels, 429- 430 Eutectoid Steel, 429 Hyper-Euteotoid Steels, 429, 430 Hypo-Eutectoid Steels, 429 Troostite. (See Heat Treatment) Sulphur and Strength. (See Physical Properties) Tempering. (See Heat Treatment) Tensile Strength. (See Physical Prop- erties) Tool, 362, 439 Electric, 404, 406 Torsional Strength. (See Physical Prop- erties) Transverse Strength. (See Physical Properties) Tungsten. (See Tungsten Steel) Uses in Construction, 362 Vanadium. (See Vanadium Steel) Working Stresses, 489^99 Yield Point. (See Physical Properties) Steel Castings, 443, 467 Steel Forgings, 443 Steel Ingots, Defects in, 417-419 Blow-holes, 417-418 Correction of, in Rolling, 418 Deep-seated, 418 Occurrence of, 417-418 Surface, 418 Ingotism, 418 Causes and Correction of, 418 Piping, 418 Causes of, 418 Removal of Pipe, 418 Segregation, 418-419 Causes of, 418 Remedies for, 418-419 Sterro Metal, 594 Stone for Concrete, 174-176 Stone Masonry, 217-221 Allowable Loads on, 221 Classes of, 217-220 Ashlar, 219-220 Rubble, 217 Squared Stone, 217 Compressive Strength, 220-221 Stones, Building, 202-217 Absorption, 215 Classifications of, 202-203 Compressive Strength, 218 Cutting, 206-207 Description of Varieties, 207-213 DurabiUty, 214 Expansion and Contraction of, 215-216 Fire Resistance, 216-217 Frost-Resistance, 216 Gneiss, 210 Granite, 207 Limestones, 210-211 Bedford, 210 Marbles, 210 Onyx Marbles, 211 Stalactite and Stalagmite, 211 Travertine, 211 Modulus of Elasticity, 218 Modulus of Rupture, 218 Quarrying, 204-206 Sandstones, 210-213 Amherst, 212 Berea, 212 Brownstone, 212 Euchd Bluestone, 212 Lake Superior, 213 Medina, 213 Ohio, 212 Potsdam, 212 Rocky Mountain, 213 . Waverly, 212 Selection of, 213-214 Shearing Strength, 218 Structural Uses, 202 Surface Dressing, 206-207 Varieties of, 207-213 Weight, 218 Strength. (See Material in Quqption) Stress-Strain Diagram. (See Material in Question) Structural Steel, 439, 442, 466-467 Rolling, 413 Subaqueous Concrete, 183-184 Sulphates in Cement, 144 Sulphur in Cast Iron, 144 Summer Wood. (See Wood) Talbot Process, 398-399 INDEX 679 Tapestry Brick, 234 Temper Carbon, 336, 338 Tempering. (See Steel, Heat Treatment) Tensile Strength. (See Material in Ques- tion) Ternary Alloys, 500, 501-534 Terra Cotta, 244r-246, 248-251 Absorbing Power, 250 Building Blocks, 245-246 Compressive Strength, 248-249 Decorative, 244 Fireproofing, 246 Lumber, 245-246 Properties, 248-251 Shearing Strength, 251 Transverse Strength, 250 Tile, 246-248 Dram, 247-248 Floor, 247 Roofing, 246-247 Wall, 247 Timber, 605-658. (See also Wood) Construction Uses, 605-606 Decay of. (See Durabihty) Durability of, 650-651 Causes of Decay, 650, 651 Process of Decay, 651 Relative Durability of Woods, 651 Moisture Content and Strength, 651- 658 Preservation of, 651-658 Antiseptics, Injection of, 652-656 Charring, 652 Creosote, 652 Kihi Drying, 651 Low-pressure Processes, 656 Non-pressure Processes, 655-656 Preservatives, 652 Pressure Processes, 653-655 A. C. W. Process, 653 AUardyce Process, 654 Bethall Process, 653 BoUing Process, 653 Breant Process, 653 Burnettizing, 653 Card Process, 654 Creo-resinate Process, 654 Lowry Process, 654 Rtiping Process, 654 WeHhouse Process, 654 Seasoning, 651 Strength of Treated Timber, 657-658 Creosote, Effect of, 657-658 Steam Treatment, Effect of, 657 Zinc Chloride, Effect of, 658 Superficial Treatments, 656 Timber — Continued Preservation of, — Continued Superficial Treatments — Continued Brush Method, 656 Dipping, 656 Zinc Chloride, 652 Properties. (See Woods, Mechanical Properties) Seasoning, Effect of, upon Durability, 651 Strength of Treated. (See Preservation of) Tests of. (See Woods, Mechanical Properties) Trees Used for, 605 Time of Setting. (See Material in Ques- tion) Tin, 572-574 Commercial Forms, 572 Extraction from Ores, 572-573 Concentration, 572 Refining Crude, 573 Smelting, 572, 573 Ores, 672 Production Statistics, 574 Properties, 573, 574 Uses, 572, 574 Tin Alloys, 599 Brittannia Metal, 599 Tombac, 591 Tool Steel, 591 Tracheids. (See Wood Elements) Transverse Strength. (See Material in Question) Trass, 71, 72 Travertine, 211 Trees. (See also Timber, and Wood) 605-607 Broadleaf, 606-607 Classes of, 606 Conifers, 606 Endogens, 606, 607 Exogens, 606, 607 Species Used for Timber, 605 Troostite. (See Steel, Heat Treatment) Tubes, Seamless, 414 Tufa-Portland Cement, 71 Tungsten Steel, 522-526 Composition, 522-523 Critical Points, 523-524 Hardness, 523, 526 High-speed Steels, 523, 526 Magnetic Reluctance, 523 Manufacture, 526 Structure and Constitution, 523-524 Tensile Properties, 524-526 680 INDEX Tungsten Steel — Continued Tensile Properties — Continued Carbon Content and, 525-526 Tungsten Content and, 525-526 tJses, 526 Tungsten-Chromium Steels, 539-540 Hardness, 540 Heat Treatment and, 540 Vanadium Bronze, 588 . Vanadium Steel, 531-534 Composition, 531 Manufacture, 534 Structure and Constitution, 532 Tensile Properties, 532-534 Carbon Content and, 533 Heat Treatment and, 533, 534 Vanadium Content and, 532 Uses, 534 Vessels. (See Wood Elements) Wall Tile, 247 Warping. (See Wood) Water-proofing of Concrete, 188-191 Water-proofing Compounds, 188-191 Welding (Wrought-iron), 357-359 White Cast Iron, 315, 317-318 White Heart Malleable, 338 Wire-Drawing, 413-414 Wood, 605-658 Bark, 608 Broadleaf. (See Trees, also Wood Structure) Cambium, 608-609 Case-hardening, 630 Checking in Drying, 630 Color, 622-623 Conifers. (See Trees, also Wood Struc- ture) Density, 623-626 Diffuse Porous, 615 Drying, 626-631 Elements. (See Wood Elements) Generation of, in Cambium, 608-609 Grain, 620-621 Growth, Endogenous, 619 Growth, Exogenous, 607-609 Growth Rings, 614-618 Heartwood, 618-619 Honeycombing, 630-631 Hydroscopieity, 631 Knots, 621, 622 Medullary Rays, 611-612 Medullary Spots, 613-614 Moisture Content, 626-628 Wood — Continued Odor, 623 Pith, 607-608 Pith Flecks, 613-614 Pith Rays, 611-612 Primary, 608 Properties. (See Woods, Mechanical Properties) Rays, 611-612 Resin Ducts, 612 Ring Porous, 615 Sapwood, 618 Seasoning, 626-628 Secondary, 609 Shrinkage in Drying, 628-629 Specific Gravity, 623-626 Spring Wood, 614-618 Strength. (See Woods, Mechanical Properties) Texture, 621-622 Warping in Drying, 629 Weight, 623-626 Wood Elements, 609-611 Parenchyma, 610-611 Tracheids, 610 Vessels, 609-610 Wood Fibers, 610 Wood Preservation. (See Timber, Pres- ervation of) Wood Structure, 606-620 Broadleaf Trees, 620 Conifers, 619-620 Diffuse Porous Woods, 615 Growth Rings, 014-618 Thickness of, 617-618 Heartwood, 618-619 Medullary Spots, 613-614 Pith Flecks, 613-614 Rays, 611-612 Fusiform, 612 Medullary, 611 Pith, 611 Primary, 611 Secondarj', 611 Resin Ducts, 612 Ring Porous Woods, 615 Sapwood, 618 Spring Wood, 614-618 Summer Wood, 614-618 Wound Tissue, 613 Woods, Mechanical Properties, 632-650 Compressive Strength, 634-635 Across vs. Along Grain, 635 Deformation, 634 Factors Governing, 634-635 Failure in Compression, 634, 635 INDEX 6^1 Woods, Mechanical Properties — Cont'd Cross-Breaking Strength and Stiffness, 635-637 Deflection, 636 Extreme Fiber Stress, 636 Factors Governing, 635-637 Modulus ot Elasticity, 637 Modulus of Rupture, 637 Relation to Tensile and Compressive Strength, 635-636 Factors of Safety, 648, 650 Growth, Relation of Rate of, to, 645- 647 Moisture Content, Relation of, to, 637- 643 Air Seasoning, Effect of, upon, 637 Compressive Strength and, 640-642 Cross-Breaking Strength and, 642- 643 Fiber-Saturation Point, 639-640 Kiln Drying, Effect of, upon, 637 Stiffness" and, 643 Stiffness. (See Cross-Breaking Strength, also Moisture Content) Summary of Tests of Timber, 648-649 Summer Wood, Relation of Proportion of, to, 647 Tensile Strength, 633-634 Across vs. Along Grain, 633 Factors Influencing, 633, 634 Failure in Tension, 633-634 Test Results, 648, 649 Time Factor in Tests, 647-648 Variability of, 632 Weight, Relation of, to, 643-645 Working Stresses, 650 Wrought Iron, 294, 341-359 Box-piled, 352 Busheled Scrap, 352 Charcoal, 351 Composition, 350 Compressive Strength, 356-357 Yield Point and, 356 Constitution, 350-351 Cementite, 351 Ferrite, 350 Pearlite, 351 Slag, 350 Definition, 342 ■ Engine-bolt, 352 History of Iron Making, 341 Knobbled Charcoal, 351 Manufacture, 343-350 Mechanical Puddling, 348-249 Roe Furnace and Operation, 349 Muck-bar or Puddled Bloom, 352 Wrought Iron — Continued Plate, 352 Properties, 350-359 Puddled, 352 Puddling Process, 343-347 Charge of Furnace, 343-344 Chemical and Physical Changes, 345- 347 Fettling, 345 Forge Pig, 343 Furnace, 344-345 Operation of Furnace, 345-347 Balhng Stage, 346-347 Boiling Stage, 346 Clearing Stage, 346 Melting-down Stage, 345 Refined Bar, 352 Rolling Mill Operations, 348 Bar Mill, 348 Blooms, 348 Double-refined Iron, 348 Merchant Bar Mill, 348 Merchant Bars, 348 Muck Bars, 348 Singled-refined Iron, 348 Scrap Processes, 349-350 Box-piled Scrap, 350 Bundled Scrap, 349 Busheled Scrap, 349 Fagoted Scrap, 350 Shearing Strength, 357 Slag Removal, 347-348 Staybolt, 352 Tensile Strength, 352-356 Cold Working, Effect of, upon, 354-356 Direction of Rolling and, 352 Elastic Limit, 353, 354, 355 Elongation, 353, 355 Heat Treatment and Crystalline Structure, 356 Modulus of Elasticity, 353 Previous Straining, Effect of, upon, 354^355 Reduction in RoUing, Relation of, to 353-354 Stress-Strain Curves, 353, 355 Ultimate Strength, 353, 355 Yield Point, 353 Uses, 342, 343 Welding, 357-359 Crystalline Growth, Danger of, 357, 358 Efficiency, 359 Electric, 358, 359 Fluxes, 358 682 INDEX Wrought Iron — Contimied Welding — Continued Oxide Formation, 357 Temperature of, 357 Types of Welds, 358 Zinc, 558-566 Binary Alloys of, 597-598 Commercial Forms of, 558 Dust, 558 Extraction from Ores, 559-563 Concentration, Roasting and Cal- cination, 560 Zinc — Contimied Extraction from Ores — Continued DistUlation and Condensation, 561- 562 Grades of, 563 Ores, 559 Pigments, 558 Production Statistics, 566 Properties, 563-565 Refining Crude, 562-563 Spelter, 558 Zinc Chloride. (See Timber, Preservation of)