BOUGHT WITH THE INCOME ©F THE SAGE ENDOWMENT FUND THE GIFT OF HENRY W. SAGE 1891 H^e date shows when this volume was taken. To renew this book copy tlie call No. and give to the Ubradan. .HOME USE RULES All Books BUbject to recall All borrowers misfit regis- ■' ter in the library to borrow books for home use. All books must be re- turjied at end of college year for inspection and repairs. Limited l;!ooks must be "• returned within the four W3ek limit and not renewed. Students must return all ^ books before leavinjg town. Officers should arrange for the return of books wanted during their absence from town. Volumes of periodicals and of pamphlets are held in ^ the library as much as possible. For special pur- poses they are given out for n limited time. Borrowers should not use thisir library privileges for the benefit of other persons. Books of special value .and gift books, when the ...(.«( , giver wishes it, are not allowed to circulate. Readers are asked tp re- port, all cases of books marked "or mutilated. i; Do not deface books by marks and writing. arVl9448*^'"™" ""'»"»«/ Library ^iiMiSSffiSi9n..Pf iron- Cornell University Library The original of tliis 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/cu31924031235652 WORKS MANAGEMENT LIBRARY THE CORROSION OF IRON A SUMMARY OF CAUSES AND PREVENTIVE MEASURES BV L. C. WILSON NEW YORK THE ENGINEERING MAGAZINE CO. 1915 7r Copyright, 1915 By THE ENGINEERING MAGAZINE CX}. PREFACE This little volume is the result of an attempt to collect and put in simple form for reading and ready reference some of the more interesting and important facts connected with the corrosion of iron and its protection therefrom. It requires no argument to prove the widespread nature and serious import of this process of decay, and the volume of literature to he found on this sub- ject is evidence of the increasing attention which is being paid to it. It is only comparatively recently, however, that corrosion has been studied in a scien- tific manner and something of its true character as- certained, so it naturally follows that the same thing is true in large degree of the measures employed for protection. This is, perhaps, especially the case with paint pigments and vehicles, where careful experi- ments have brought to light many fundamental truths of the greatest value regarding the nature and action of some of these materials, thus enabling us to use them with much greater intelligence and effective- ness. Much of the published work, however, is scat- tered throughout many different scientific and tech- nical magazines, consequently is largely inaccessible to one who does not have the time or facilities for such investigation. It is accordingly hoped that this book, by pre- senting some of this material in condensed form, may help to give the student or busy engineer a better understanding of the problems involved in the successful preservation of one of our most useful building materials, and point the way to their solu- tion. Brooklyn, New York. July 31, 1915. CONTENTS Chapteb I. The Eust PEOBiiEai. ...<<...... 1 Its Importance to Constmcting Engi- neers, Architects and Owners — Susceptibil- ity of Iron and Steel to Corrosion — Eco- nomic Bearings of the Matter — What Cor- rosion Is — The Contest of Iron versits Steel — Effect of Modem High-speed Meth- ods of Steel-making — Eemarkable Exam- ples of Preservation of Iron from Eust — ■ Explanation of this Good Behavior of Cer- tain Articles — The Chemical and Physical Facts Involved in Corrosion — Theory and Process of Solution — Ions and Ionization — Theory of Dissociation — Precipitation — Practical Bearing of these Things on the Busting and Protection of Iron and Steel. Chapter II. Theories of Corrosion 23 The Three Principal Theories— The Hy- drogen Peroxide Theory — Evidence for and against It — The Carbonic Acid Theory — Experiments Showing Its Partial Truth — The Electrolytic Theory — Data Supporting It — ^How It Agrees with Observed Facts — How It Influences the Choice of Protective Measures — Bad Influence of Stresses and Strains — Of Pickling — Of Indentations and Eoughness of Surface — Conclusions, iii iv contents Chaptke hi. Peotectitb Measttees 46 Practical Application of the Principles Discovered — Three Methods of Protection: by Coatings of Oxide, by Coatings of Other Metals, by Covering with Paint or Similar Materials — Pioneer Work in Oxide Coatings by Lavoisier and Faraday — The Bower- Barff Process — The Wells Process — The Gesner Process — The Dewees-Wood Process — The Bontempi Process — The Buffington Process — ^Heat Bluing — Coslettizing — ^Pro- tection by Coatings of Other Metals: Gal- vanizing by Dipping, Electrolysis, and Vapor — Hot Galvanizing — Methods of Testing — Electro-depositing of Zinc : Meth- ods and Defects — ^Vapor Processes — Sher- ardizing: Details and Advantages of the Process — The Molten Zinc Process — Preece and Walker Tests for Galvanized Work — Copper-coating — Nickel-coating — Lead- coating — ^Lohmannizing — The Schoop Proc- ess. Chapter IV. Paint Materials 74 Wide Applicability of Painting as a Pro- tective Measure — Nature of Paint — ^Two Methods of Making Paint — Comparative Merits of these Different Procedures — In- gredients of Paint — ^Linseed Oil — Drying and Driers and Their Effects on the Qual- ity of the Oil — Adulterants of Linseed Oil — Other Paint Oils — Tung or China Wood Oil — Soya Bean Oil — Menhaden Oil — Thin- ners — Turpentine — Petroleum Oils — Pig- CONTENTS V ments — Their Influence on Durability — Findings of the Committees of the Amer- ican Society for Testing Materials Tabu- lated and Discussed — Pigments Described — Zinc Chromate — ^Zinc and Barium Chrom- ate — Zinc Oxide — Zinc-lead White — Barium Chromate — ^Ultramarine Blue — Chrome Green — Prussian Blue — Litho- pone — Willow Charcoal — Litharge — ^Basic Carbonate-White Lead — Calcium Sulphate — Prince's Metallic Brown — Orange Min- eral — Calcium Carbonate — Sublimed Blue Lead — Lemon Chrome Yellow — 'Orange Chrome Yellow — Medium Chrome Yellow — Chrome Green — Venetian Eed — Bone Black — Asbestine — China Clay — Eed Lead — Indian Bed — American Vermilion — Sub- limed White Lead — Mineral Black — ^Barytes — Graphite — Ochre — Carbon Black — ■ Lampblack. Chapter V. Protective Paints 100 Quality of Paint for Ironwork Depends on Inhibitive, Insulating and Excluding Properties — Action of Various Pigments Discussed — Tests of Paint Films — Cush- man and Gardner's Eesults — Tests Made by the American Society for Testing Ma- terials — Formulae of the Paints Tested — Eesults — Conditions Controlling the Char- acter of a Paint Film — How to Work up a Good Paint Formula — Effect of Various Ingredients — Japans and Enamels — Lac- quers — Preparatory Treatment of Iron and Steel Surfaces before Painting. VI CONTENTS Chaptee VI. Influence oe Dipeeeent Ele- MENTS ON THE COKEOSION OF IeON , 137 Definite Work of Eeeent Date — Manufac- ture of Iron Summarized to Show Nature of the Problem — Natural Impurities and Their Effect on Electrolysis — Influence of Normal Constituents Individually Determined — Silicon — Sulphur — Manganese — Arsenic — Cobalt and Nickel — Silver, Tin, Tungsten and Selenium — Copper — Tests of Copper- Steels — Buck's Results Tabulated — Excel- lent Restdts Secured in Protection from Corrosion under Severe Conditions — Hope- ful Outlook for Copper-Bearing Steel as a Rust Resister. Chaptee VII. Coeeosion op Wrought-Ieon AND Steel Pipe 149 Controversy over their Relative Merits — Importance Due to the Enormous Use of Pipe — Puddling and Bessemer Processes Summarized — Their Tendencies Pointed Out — ^Wrought Iron Subject to Lamination — Bessemer Steel Free from this Tendency — Tests Made by the Navy Department — Results — ^Data Collected by the American Society for Testing Materials — Thomson's Experiments — Tests of the American So- ciety of Heating and Ventilating Engineers — ^Woolson's Tests in New York City Public Baths — Walker's Experiments on Pipe Lines — All Conclusions Negative to the Superior- ity of Iron over Steel as to Corrodibility, and Tendency to Pitting Somewhat Greater in the Case of Wrought Iron. THE CORROSION OF IRON Chapter I THE ETJST PEOBLEM A MONG the many questions wHcli the ar- chitect and builder are called upon to consider, none perhaps, is more important than that relating to the corrosion of iron and steel. This tendency to decay is peculiar to these materials, since none of the ordinary metals exhibits it to any comparable degree, and it is so strong that an unprotected piece is soon reduced to a shapeless mass of rust. It is one thing to design and erect a mighty sky-scraper, but an entirely different matter to protect it from those influences which, if allowed to do their work, would ruin the greatest structure in a short time. It seems strange, in a way, that unless iron is well protected it is far less resistant to natural agencies than wood or other building materials. It is being used more and more in the fabrication of large buildings and other 1 3 COEEOSIOW OF IRON structures of a permanent type, and the length of their life, to say nothing of the safety of the people employed in and around them, depends on the success with which cor- rosion can be prevented. Painters are constantly employed on such structures as the Brooklyn or Forth bridges, scraping away rust spots and repainting. When serious rusting is once started it is very hard to check, and as it would inevitably lead to the destruction of the framework, a collapse would be certain to occur sooner or later, probably with an appalling loss of life. The economic side of the question is also of serious import. It is being recognized more and more that the increasing use of materials is depleting our stock of natural resources, and it is evident that when these are irrevocably lost through some form of decay, we have sustained a real and serious loss. The production of pig iron in the United States alone has increased amazingly until now many millions of tons are produced annually. How much of this is wasted for lack of suitable protection, or otherwise, can- not be stated very accurately, but the total is considerable. How long our ore supply wiU hold out against the steadily increasing de- THE RUST PEOBLEM 3 mand can only be vaguely guessed, but if present conditions keep up there is certain to come a time when the scarcity and high price of iron will not permit it to be used so prod- igally as now. Again, the coal supply is by no means unlimited, and since it requires about four tons of coal, or its equivalent, to produc^ one ton of steel, it is apparent that when we allow iron to rust or otherwise go to waste we are losing not only iron, but coal, one of our most valuable assets. I do not mean to imply that the corrosion of iron is a new problem, or that technolo- gists generally do not realize its importance, since the articles and discussions which have been appearing in the technical press for the last few years are evidence of the work that is being done along these lines. This litera- ture, however, is distributed through a large number of publications and is, therefore, not readily accessible to the busy engineer, so in this small volume an attempt will be made to collect some of this information re- garding the study of corrosion and its prac- tical application in the protection of iron and steel, and to present it, together with the results of personal observations and tests, in a simple, compact form. COEEOSION OF lEON Before taking up any of the theories as to the nature and cause of corrosion, it may be interesting to note a few points about the process in general. It is a matter of common knowledge that iron soon becomes covered with a heavy coating of rust, if it is exposed to moisture without first being protected by paint or other good preservative, and that in time the piece will be entirely honeycombed and will assume a lace-like appearance. It is also known that different pieces of iron and steel exhibit widely varying tendencies in this respect ; that is, some are surprisingly resist- ant and keep in fairly good shape for a long time even under very adverse conditions, while others soon show the most serious rust- ing when exposed to the same conditions. The manufacturers of wrought and charcoal irons and the various kinds of steel have each contended strenuously that his particular product is most resistant to corrosion, but it is not the province of this chapter to state who is in the right, since it appears that each of these types of metal may vary greatly in rust resistance and that there are good and bad irons as well as good and bad steels. There is plenty of good evidence to support this view, and it seems worse than useless, THE EDST PEOBLEM 5 therefore, for mill men to engage in wordy controversy over the particular merits of their products. The recent researches of sev- eral investigators show that internal stresses or strains in a metal, as well as improper treatment resulting in occluded gas and a porous metal with blow holes, are conducive to rapid corrosion. Modem high-speed meth- ods, unless carefully performed, tend to pro- duce a metal in which various substances, which are present either as impurities or are added to produce certain effects, are more or less segregated instead of being thor- oughly mixed, and the iron is pounded and rolled into shape and cooled so quickly that the molecules do not have time to adjust themselves and heavy internal strains may be produced. Of course, the remedy for all this is more careful workmanship and ade- quate annealing. Metal which is strained beyond its elastic limit can be shown to be more easily corroded than metal which has not been strained, and it apparently makes little difference whether these strains are produced during the process of making the part or afterward. There seems to be no doubt that at least some of the metal produced previous to the 6 OORKOSION' OF IKON introduction of modem metliods (modem as far as speed is concerned) was superior in its resistance to corrosion to the present-day product, although the latter should not be condemned too hastily on this one point alone. Thus, I have seen various iron articles, espe- cially nails, which showed far less rust after an exposure of a hundred years or so, than the modern variety sometimes does in. a few weeks. Another piece, an old flint-lock pistol, was especially interesting. It was found by a friend in a patch of woods in Vermont and had evidently lain there for many years, since a piece of newspaper dated 1796 had been used as wadding in loading it. All the iron parts were rather rough and pitted and cov- ered with rust, but the arm was in surpris- ingly good shape, considering the conditions to which it had been subjected. The spring, hammer, and trigger were still capable of performing their functions and very little effort was required to put the old weapon in decidedly presentable condition after its ex- traordinary exposure to the weather. The old bridge at Newburyport, Mass., has also attracted much attention because after something like a hundred years' exposure to THE BUST problem: 7 the elements it shows remarkably little evi- dence of deterioration. Going farther back, a number of iron arti- cles have come down to us from ancient times, and perhaps the most noted of these is the famous pillar in the temple of Kutab Minar at Delhi, India. This old shaft, which pro- jects some 30 feet above the surface of the ground, was erected about 900 B. 0. Today- it shows little trace of rust, although it has had no protective coating other than that which the process of manufacture and the atmosphere itself have formed upon it. In this connection mention may also be made of what are known as the Ceylon link and the Singhalese chisel, nail and bUl hook, all specimens of great antiquity. To explain how or why these various arti- cles have so successfully withstood the at- tacks of air and moisture for so many years is not as easy as it might seem, for appar- ently some of our commonly accepted notions and theories as to what a good iron should and should not be do not hold, or at least are offset by other conditions. For instance, some of the links of the Newburyport bridge have been examined and found to be of very ordinary purity and of a most heterogeneous 8 CORROSION OF IRON structure, segregation of the impurities being decidedly noticeable. Tbe question of any inherent superiority in the iron itself was settled by heating and rolling one of the links into a sheet and exposing it to the weather, when it was found that it corroded quite as readily and extensively as the material of to- day. Again, tests made on the old relics referred to above show that only in manganese content do they approach modem irons in composi- tion; the phosphorus is from 22 to 60 times and the silicon from 9 to 50 times as high in these as in the present product. From this it would seem that on the score of composi- tion they are in no way superior to our own material, and in addition they contain several per cent of slag and cinder, which are admit- tedly very undesirable. The explanation of their good behavior in service is probably found in the fact that the iron from which they were produced was re- fined, and the objects themselves fashioned slowly and laboriously by hand in a sort of forge. The prolonged heating and slow forg- ing would tend to insure thorough annealing and freedom from strains, while the constant hammering produced a dense skin on the sur- THE RUST PEOBLEM 9 face, and this was doubtless covered with a more or less heavy coating of scale or oxide which also helped to protect the metal be- neath. Thus it seems that the ability of these irons to withstand atmospheric attack is ia- cidental to the method of working, rather than due to greater purity or freedom from segregation, as compared with modern irons. Of course, to apply such methods to the enor- mous output of iron and steel nowadays would be utterly impossible. Mention has been made of the impurities iu iron, and it may not be out of place to enum- erate some of them at this time. All except very highly purified iron contains silicon, sul- phur, phosphorus, manganese, and combined carbon, at least in small amounts. In making alloys, chromium, vanadium, tungsten, nickel, copper and several other elements may be added. Iron alloys or combines with prac- tically all the other elements, and is so easily influenced by some of them that oftentimes only a very small percentage is required to alter its character markedly; as a conse- quence, specification requirements are gener- ally drawn very rigidly. A good deal of importance has been at- tached to the influence which sulphur, phos- 10 COREOSION OP lEOlT phorus, and the other normal constituents of iron are supposed to exert on the process of corrosion, and manganese, in particular, has come in for an unusual amount of con- denmation. For example, it is known that it decreases the electrical conductivity of iron, to a certain point ; therefore, unless it is uni- formly distributed, the electrical conductivity wiU vary in different parts and on different surfaces, which has an important bearing on the problem in hand, as will be seen later on. Further, it is well known that manganese and sulphur tend to combine when both are present in iron and steel, and the resulting sulphide shows a difference of electrical po- tential against iron, which is a highly unde- sirable condition. In general, it may be said that, according to the electrolytic theory, any condition which tends to produce a difference of potential between surrounding parts is bad and will excite or accelerate rapid corro- sion. A detailed discussion of the influence of the various elements, as far as corrosion is concerned, will be taken up later on, but it may be said here that the results of a vast amount of experimental work show that even extremely pure iron corrodes about as readily THE RUST PBOBLEM 11 as a good, commercial grade of material; therefore the question is apparently not so much -what amount of impurities may be present as whether these are thoroughly dis- tributed throughout the metal or segregated here and there in patches. Before we can discuss some of the theories regarding the formation of rust and thus approach the subject in a scientific manner, however, it will be necessary to review briefly certain chemical and physical facts involved in the process, in order that there may be no misunderstanding of the terms and principles employed. Primarily we will be concerned with the phenomenon of solution. Water is the great solvent for aU forms of matter and there is probably nothing which it does not affect in some way. Most chemical changes take place in water, or in its pres- ence, so it is not only destructive in its action, as we often feel, but also constructive, as new compounds may be formed as well as old ones broken up. When we speak of the solu- bility of a substance we refer to the ease with which it goes into solution; therefore we say it is iasoluble if we find that no appre- ciable amount has been dissolved. Prac- tically, however, there is almost nothing 12 CORROSION OP IRON which is not dissolved to some extent, so the term is only relative, and there are all de- grees of solubility ranging from those oases in which the solvent dissolves but an in- finitesimal amount of the material, to those in which it takes more than its own weight of the substance into solution. This is not the place to discuss why a cer- tain solid does or does not dissolve in a cer- tain liquid, so we will confine ourselves to substances which are easily recognized as soluble in water, that is, dissolve readily in it. When such a body is placed in water it dis- appears from view more or less rapidly, and its molecules, if it is a compound body, dis- tribute themselves uniformly among the molecules of the solvent. This distribution is produced by a very definite force, known as the "solution pressure," which is exactly an- alogous to the pressure exerted by a gas con- fined in a closed vessel. Under such circum- stances the gas tends to occupy all the space at its disposal, and therefore produces a cer- tain uniform pressure on all parts of the vessel. If we had steam at a pressure of 100 pounds per square inch in a boiler and wished to introduce more steam into the lat- ter, it is evident that we would have to have THE BUST PKOBLKM" 13 the incoimiig vapor at a pressure slightly higher than 100 pounds per square inch, in order to overcome the back pressure, as it is called. Again, as a boiler is generating steam and the tension increases, it becomes ever harder for the water to assume the vapor state, against the pressure of the steam al- ready formed. The same thing happens ia the case of so- lutions. As more and more of the substance dissolves, there is produced a back pressure, known as osmotic pressure. It is evident that when the osmotic and solution pressures are equal, no more of the material will go into so- lution and a state of equilibrium will result. If any outside influences disturb this equilib- rium and lower the osmotic pressure, for example, it is easily seen that more of the substance will pass into solution until equilib- rium is again established. Osmotic pres- sure may be of surprising magnitude, in cer- tain cases, and can be shown by placing a glass tube closed at one end by animal parch- ment and filled with alcohol, in a glass of water. Alcohol and water have a strong ten- dency to diffuse or mix, and water can easily pass through the parchment into the tube, but the alcohol caimot; therefore the column 14 COBEOSION OF IRON of liquid ia the tube will rise, showing an increased pressure. It may be interesting to note that the solution pressure of a very- soluble substance, like sugar, is considerable. One investigator found that the pressure in the case of a 6 per cent sugar solution amounted to nearly 60 pounds per square inch. Coming now to the behavior of substances in solution, we know that if we dissolve salt in water it apparently disappears, but since we realize it must be present in some form we conclude that it has been resolved into ex- tremely fine particles. Experiment has shown that such a solution will conduct a current of electricity and two entirely new substances wiU make their appearance, one around each electrode, so it is plain that something in the solution is capable of carrying the electricity from one pole to the other, and that in so doing certain changes have taken place. The term "ion" was introduced, therefore, to denote the particle of matter which travels in the solution, along with the electricity, toward the poles. A further distinction is made in that the particle of matter whicb travels toward the anode or positive pole is called an an-ion, and that which migrates THE RUST PROBLEM 15 toward the cathode or negative pole, a cat- ion. The theory of dissociation holds, at the present time, that when acids, bases, and salts are dissolved in water they break down into ions, which are considered to be atoms or groups of atoms carrying large charges of static electricity. These charges are not apparent in the undissociated molecule, be- cause the positive charges exactly neutralize the negative ones. Thus, hydrochloric acid dissociates into positive hydrogen and nega- tive chlorine ions ; sodium hydroxide (caustic soda) forms positive sodium and negative hydroxyl ions; ferrous chloride forms posi- tive ferrous and negative chlorine ions, and soon. Putting it in the form of a principle, with- out taking the time to demonstrate it, we may say that the cation of all acids is hydrogen; the anion varies with the acid. Likewise, the anion of all bases is the hydroxyl group (OH) ; the cation varies with the base, the metal with which it is combined. Since hy- drogen is the characteristic ion of all acids, it can be seen that whenever its ions are pres- ent, acid properties are exhibited; in like manner, hydroxyl is the characteristic ion of 16 COEEOSIOIT OP IKON" all bases, and when hydroxyl ions are present basic properties are in evidence. With salts, both the anion and cation vary according to the elements composing the salt. It is not possible to discuss this interesting topic fur- ther, but anyone who wishes more informa- tion along these lines should consult some good text book on physical chemistry. Per- haps the application of the foregoing to the corrosion of iron may seem a little obscure, so we shall anticipate enough to say that this process is greatly retarded or accelerated by the presence of different ions, therefore some knowledge of their formation and action is necessary. Another form of breaking down or disso- ciation, known as hydrolysis, is also of im- portance in the study of corrosion. If a solu- tion of the carbonate, say, of sodium, potas- sium, or other strongly basic metal be tested with litmus paper, it will be found to have a strongly alkaline reaction. On the other hand, chromium, bismuth, aluminum and other weak bases tend to form salts which have an acid reaction; some are even too weak to exist. The principle may be deduced, therefore, that a compound formed from a weak acid and a strong base will be alkaluie THE EUST PEOBLEM 17 in reaction, while one formed from a strong acid and a weak base will be acid. Wben considering tbe theory of dissocia- tion it was said that hydrogen ions are the cause of an acid reaction and hydroxyl ions the cause of an alkaline reaction. Conversely, it can be said that an acid reaction denotes the presence of hydrogen ions, while hy- droxyl ions are shown by an alkaline reac- tion. It has just been noted that different ions may exert great influence on the process of corrosion ; therefore, when we go into this more deeply later on we should be able to ex- plain, in many cases at least, why certain pig- ments are stimulators of corrosion, or why they tend to act on the oil in paint films and cause early decay and failure. Basic car- bonate of lead — ^white lead — ^for example, is so alkaline in its reaction that it saponifies or makes lead soaps of the linseed oil, leading to what is called chalking and deterioration of the paint film. A further investigation of solutions shows that some will conduct a current of electricity while others will not. A substance which will conduct a current when in solution is called an electrolyte. For example, potassium ni- trate, saltpetre, in solution will conduct a 18 COEEOSION OP lEON current very readily, but syrup offers a high resistance to its passage. It can be shown that potassium nitrate ionizes highly in so- lution, which is not true of sugar, and it has been noted previously that when a current is passed through a conducting solution a me- chanical movement of the ions takes place toward the electrodes, where they give up their charges and "plate out" in metallic or gaseous form. This may be familiar to the reader as the process of electro-plating. If a solution of zinc sulphate is made to conduct a current, the positive zinc ions proceed to the negative pole, give up their charges, and assume the atomic condition. An equivalent amount of sulphuric acid is set free at the same time at the positive pole. If this is of zinc, as in electro-galvanizing, the acid at- tacks it and a further quantity of metal is taken into solution. Going a step farther, it is not necessary that an external source of elec- tricity be provided before electrolysis can take place. Strips of two different metals will generate a current when they are put into an electrolyte and the upper ends touched to- gether or joined by wire. This may be ex- plained by saying that the more electro-posi- tive metal shoots positive ions into the solu- THE BUST PEOBLEM 19 tion and thus leaves itself negatively charged, consequently appearing as the negative pole. It is more or less rapidly disintegrated, while the other is protected. As will be shown later on, this phenomenon plays a very important part in the preservation of iron, especially by zinc coatings. As everyone knows, all but a few of the elements are capable of combining and form- ing compounds with the other elements. The capacity of an atom to combine with other atoms is known as its "valence" and varies with the different elements. For instance, one atom of chlorine can combine with only one atom of another element; one atom of zinc can combine with two monovalent atoms, and so on. Sometimes the same element va- ries in its ability to combine with others ; for example, copper combines sometimes with one and sometimes with two atoms of chlor- ine. Manganese, gold, iron, and many others exhibit this peculiarity, which is thought to be due to the electrical state of the ion. By appropriate means such an element may be made to combine with a larger quantity of another element, if it is in the lower state of oxidation, or to give up a part of the element with which it is combined, if it is in the higher 80 COBEOSION OF lEON" state. Passing from a higher to a lower state of valence is known as "reduction" and from a lower to a higher state as "oxidation." Often one of these states is very unstable, tending to pass into the other under ordinary- circumstances. Thus iron forms two classes of compounds, ferrous and ferric, which have a valence of two and three, respectively, but the ferrous compounds oxidize very easily to the ferric state. If we immerse pure iron in pure water a certain amount of it wiU be dissolved, since iron has a small but definite solution pressure in water. It goes into solution as ferrous hydroxide; therefore, the iron is present as positive ferrous ions, the hydroxyls taking the negative charge. As soon as the ferrous ions come into contact with the oxygen of the air they are oxidized to the ferric state. Fer- ric hydroxide is insoluble in water, therefore it precipitates out; and since its influence on the solution is then at an end, more iron will dissolve and so the process will be repeated over and over until all the iron is used up. Meanwhile the precipitated iron compounds accumulate and the rust rapidly grows thicker. We referred a moment ago to the process THE RUST PROBLEM 81 of precipitation and it will be interesting to consider briefly some of the properties of tbe product of tbis action. Under certain condi- tions, precipitates may take the form of very fine crystals and are termed "crystalloids," or they may appear in a gummy, gelat- inous condition and are then called "col- loids," a word which is derived from the Greek word meaning glue. There are two important properties of colloids which should be carefully noted. The first is their peculiar propensity to absorb large amounts of liquids and impurities from the solutions from which they were precipitated. Since many pig- ments are first formed as colloidal precipi- tates, it will be seen that a serious source of danger is presented as the absorbed impuri- ties may more than neutralize any inhibitive qualities of the pigment itself. The second poiat of interest is the fact that colloidal precipitates suspended in an electrolyte be- have just as ions do, carrying electrical charges and migrating to the electrodes when a current is passed. Colloidal ferric hydrox- ide and all basic hydroxides collect around the cathode, therefore they may be consid- ered to have a positive charge. Hydroxides of an acid nature move toward the anode 22 COEEOSION OF lEON and thus have a negative charge. The most likely explanation of this is that a basic hy- droxide, for example, would split off negative hydroxyl ions and leave the residue with a positive charge, while an acid hydroxide would split off positive hydrogen ions, leav- ing the remainder with a negative charge. The practical significance of these facts is that they help to explain why the surface of a piece of badly rusted iron shows, as everyone has observed, many small but dis- tinct pits or depressions at some places and wart-like growths at others. Under the influ- ence of the electric currents formed during the process of corrosion, the ferric hydroxide is transported from the poiuts having a posi- tive charge, leaving tiny depressions, and de- posited at points of negative polarity, thus raising these above the rest of the surface. This feature will be taken up later. Chapter II THEORIES OP COREOSION TN the preceding chapter we considered ■*■ some of the fundamental principles of physics and chemistry on which the explana- tion of corrosion is based so that now we are in better condition to undertake the study of this subject itself. Speculation concerning this process has been rife for a great many years and various theories have been put forth from time to time which have endeav- ored to explain satisfactorily all the observed facts and, incidentally, point the way to a remedy. Some of these theories could not stand rigid investigation and consequently fell by the roadside, but three more or less distinct propositions have gradually been evolved, and practically all students of this subject are adherents of some one of them. The hydrogen-peroxide theory was based on a scheme of oxidation proposed by a Ger- man chemist, Traube. It assumes that when iron, water, and oxygen are in contact, a re- 23 34 COEEOSION OF IRON" action takes place in wHcli ferrous oxide (ox- ide of iron) and hydrogen peroxide are formed. The ferrous oxide then reacts with one-half of the hydrogen peroxide to form al basic or hydrated oxide; rust, an insoluble basic oxide, is the final product. Part of the excess of peroxide then acts on more iron, forming ferrous oxide and water, whereupon more of the peroxide reacts with the iron compound to produce a further quantity of basic oxide, and so the process is repeated in- definitely. Perhaps these relations may be made clearer by the following chemical equa- tions : Fe + Os + HjO FeO + HjO. Iron Oxyeen Water Ferrous Hydrogen oxide peroxide 2FeO + HjOa S9 Fe202(OH)2 = Fe203,HzO Ferrous Hydrogen Hydrated Bust oxide peroxide oxide Then, Fe + H2O! = FeO + HjO Iron Hydrogen peroxide Ferrous oxide Water 2FeO + H2O2 = FeiQjCGH). = FezOj.HzO Ferroua Hydrogen Hydrated Rust oxide peroxide oxide As a whole, this theory has seemed to de- rive some support from the fact that slight traces of hydrogen peroxide have been de- tected during the slow oxidation of certain metals. However, the same tests have not shown this substance to be present while iron is rusting. THEOKIES OF CORROSION 3S Another argument which, for a time, seemed to be strong evidence for the theory- was the observation that certain oxidizing agents, as chromic acid and its salts, prevent rusting when iron is immersed in water con- taining them; and hydrogen peroxide is known to be broken up by chromic acid. It has been pointed out, however, that because certain substances destroy hydrogen perox- ide, they do not necessarily prevent its for- mation, a contention which is borne out by analogous reactions with other compounds. A strong argument against the theory is the fact that other oxidizing agents which de- stroy the peroxide do not act the same as chromic acid and prevent rusting. Alto- gether, this theory does not seem to be suffi- ciently in accord with the known facts to serve as a working thesis. Of the two remaining theories, the carbonic acid and the electrolytic, the former has had the greater number of adherents and until recently has been considered to explain satis- factorily the phenomena of corrosion. As a matter of fact, it does express the truth to a certain degree, as will be shown later. It postulates that rusting is always started by an acid, even one as weak as carbonic suffic- 36 COEEOSION OP IKON ing to do this, the iron being first converted into a ferrous salt, with the escape of free hydrogen. Under the influence of oxygen and water the ferrous salt is oxidized to fer- ric hydroxide (which is insoluble and settles out) at the same time that the carbonic acid which was required to form the ferrous salt is set free. This again attacks the metallic iron with the formation of more ferrous com- pound which is again decomposed, ferric hy- droxide settling out and carbonic acid being set free to attack more iron and thus make the process a continuous one. The action may be shown by these equations : 2Fe + 2HJC0S 2FeCO. + 2H> Iron Carbonio acid Ferrous carbonate Hydrogen 2FeC0i + SHjO + O = 2Fe(0H). + 2H.C0i Ferrous Water Oxygen Ferric Carbonio carbonate hydroxide acid In order to prevent rusting, therefore, all we have to do is to neutralize the acid with an alkali, and those who held this to be the true explanation of the cause of corrosion pointed to the fact that iron does not rust in strongly alkaline solutions. It can be shown, however, that iron rusts very easily in dilute alkaline solutions, so that bit of evidence is thereby weakened. THEOEIES OF OOEEOSIOIT 37 Various investigators have lined np on both sides of the question and have defended their views with a series of ingenious and painstaking experiments, the object of which was to determine whether or not iron can rust in oxygen and water in the complete ab- sence of carbonic acid. The result of this work seems to show that iron will corrode rapidly in the absence of this substance if oxygen and liquid water are present, and the experiments were so carefully made that no doubt remains that carbonic acid is only a slight factor in the corrosion of iron and that this process can go on very readily without it. In this connection it may be interesting to cite an experiment which Cushman and Gard- ner describe in their book on this subject. Two Jena glass flasks were nearly filled with freshly distilled water and boiled vigorously one-half hour, at the end of which time bright, polished strips of charcoal iron and steel were slipped into the flasks and the boiling contiuued a few minutes longer. The flasks were then completely filled with the boiled water and every trace of air removed. No rust appeared on the strips, even when they were kept indefinitely in this boiled water. Pure oxygen was then freed from 38 COEEOSION OF IKON every possible trace of carbonic acid and passed into the flasks, whereupon rust formed on the bright pieces in about five minutes and was heavy in an hour. The action seemed to take place in patches having the appear- ance of a pattern, corresponding to the physi- cal structure of the metal. This experiment was also made using phenolphthalein, a sub- stance which shows the presence of exceed- ingly small amounts of acid or alkali, and in- variably a pink color developed, showing that no carbonic acid was present. This would seem to be good evidence, therefore, that water and oxygen can cause corrosion with- out the aid of other substances. The experiment was also tried of substi- tuting pure carbonic-acid gas, perfectly free from oxygen, in place of the pure oxygen used in the above tests. Under these circum- stances, no apparent action took place, even after several hours, although no doubt a slight amount of iron went into solution as the carbonate. When pure oxygen was al- lowed to enter and mingle with the carbonic- acid gas rusting began very shortly, a char- acteristic blue-green color, which always ac- companies the early formation of rust in the presence of carbonic acid, being noticed. In THEORIES OF COBEOSION' 39 normal rusting in the air this color is not to be seen. It can be further shown that, even if it is not possible to remove all traces of carbonic acid from the apparatus used in these experi- ments, the hydrogen ions which would be given by the dissociation of small amounts of this substance would not be much greater in number than those furnished by the dissocia- tion of pure water. On the whole, it may be said that the carbonic-acid theory ex- presses a partial truth, in that hydrogen ions must be present before iron can be attacked — a point which will be taken up later. The electrolytic theory is steadily gaining adherents because it seems to offer a satis- factory explanation for a greater number of observed facts than do any of the others. The modern conception of the reactions which take place between different substances in the presence of water holds that they are ac- companied by certain readjustments of the electrical state of the reacting ions, and ac- cording to the electrolytic theory iron must go into solution as a ferrous ion before it can be oxidized in the wet way. When metallic iron is put into a solution of copper sulphate (blue vitriol) the copper 30 COEEOSION" 0¥ lEON is precipitated in metallic form and the iron goes into solution, the ions of this taking the electrical charge of the copper ions. Now hydrogen is a metal, or at least it behaves like one, and has a lower solution tension than iron; therefore if we place a strip of iron in a solution containing hydrogen ions, its atoms will assume an electric charge and it will be dissolved, as it was in the copper solution, while hydrogen vnll give up its charge, changing from the ionic to the atomic condition and escaping as a gas. Ferrous salts, especially in solution or freshly precipi- tated, are rapidly oxidized to the ferric state by atmospheric oxygen, so that all the elec- trolytic theory must do is account for the solution of the iron in the first place. It has been shown by the careful experi- ments cited that iron is slightly soluble in pure water, but as long as no oxygen is per- mitted to act on the solution no sign of cor- rosion can be seen. When it is admitted, however, the red ferric hydroxide soon pre- cipitates out. So, from the results of these and other experiments, there can hardly be any doubt that iron dissolves in pure water without the aid of other substances. A little reflection wUl show that before THEOEIES OP COEEOSIOlir 31 iron can rust it must jfirst go into solution, at the same time that hydrogen is set free, and oxygen must be present to oxidize the ferrous salts formed. An exchange of elec- tricity between the reacting ions is therefore demanded and the case may be considered to be one of electrolysis, as the disappearance of a hydrogen ion, as the hydrogen goes off as gas, means the appearance of a ferrous ion somewhere else. Going a step further, it can be seen that corrosion is due to the iron being attacked by hydrogen ions and taken into solution, so we should naturally suppose that anything which increases the concentra- tion of these ions would increase the rapidity of corrosion and, conversely, anything which decreases their concentration would tend to retard or prevent it. And this is precisely the case. It was shown in the preceding chap- ter that hydrogen is the characteristic ion of acids and hydroxyl that of bases; there- fore we should expect acids to stimulate cor- rosion and alkaline substances to inhibit it, which is in accord with fact. Taking a bird's-eye view of the process of corrosion as explained by the electrolytic theory, we see that it lines up about as fol- lows. When iron and water are brought to- 33 COEEOSION OP lEON gether a certain amouiit of the metal goes into solution, since it has a certain solution pressure — or, to put it a little differently, water even when very pure is dissociated slightly into hydrogen and hydroxyl ions, the former attacking the iron as outlined before. It can be shown that the solution tension va- ries at different parts of the surface, this variation being due to strains and segrega- tion of impurities; impurities in the solvent also tend to increase the solution tension. Therefore, the points where this is greatest will be positive to those at which it is of les- ser degree, and a current will flow between them if there is a conducting medium present. Remembering that hydrogen acts like a metal, it is seen that its ions will tend to col- lect around the negative poles, while the hydroxyl ions will move toward the positive poles and try to exchange their static charges with the iron, so that the latter will go into solution and the hydrogen escape as a gas. If the hydrogen ions are in sufficient concen- tration, that is, if an acid is present, this ex- change of charges takes place very rapidly and the iron is quickly dissolved. Under or- dinary conditions, however, the action pro- ceeds much more slowly as the concentration THEOEIES QV COEEOSION" 33 of the ions is comparatively low. Iron goes into solution, nevertheless, as can be shown by means to be described shortly, and for every ferrous ion which is formed a hydroxyl ion must be formed also and appear else- where. The oxygen of the air now acts on the ferrous salts and changes them into the ferric state, in which form they precipitate out as the familiar red rust, and so the proc- ess is repeated indefinitely. Mention has already been made of so-called indicators, substances which show by a change of color the presence of very small quantities of acids or alkalis, and these are employed to show the positive and negative poles. Through the work of several investi- gators the reagent known as f erroxyl was de- veloped. It consists essentially of phe- nolphthalein, which shows the negative poles, and potassium ferricyanide, which indicates the positive poles, in a solution of gelatin or agar. It has proved to be a very valuable help in these lines of study and has widened our knowledge considerably. Nearly everyone has observed that not all pieces of iron rust in the same way. Some are covered more or less evenly with a coat- ing of rust which is of fairly uniform thick- 34 CORKOSION OP IKON ness ; others are deeply pitted in some places, whUe adjacent parts of tlie surface may show far less corrosion than might be expected. Without the aid of the ferroxyl test it is diffi- cult to explain just why these things should be. When a strip of iron is properly mounted in this reagent, various parts of the surface develop either a red or a blue color after a few hours, thus showing where the positive and negative poles are, which is equivalent to saying where the iron is being attacked by hydrogen ions and where it is being protected by the presence of the hydroxyl ions. Often these poles reverse and a positive pole be- comes negative, or vice versa; in fact, this reversal may be repeated several times. In this connection, perhaps it would be best to describe the results of such a test made by Cushman and Gardner and given in their book, to which reference has already been made. A strip of steel was placed in ferroxyl and carefully observed. When the colors first developed two dark blue nodes formed at the opposite ends of the test piece, with a large pink area at the centre, where for a time the metal remained quite bright. Very slowly, how- ever, the poles changed and the pink central area THEOEIES OP COEEOSION 35 disappeared and gave way to a large blue node which enveloped three-quarters of the test piece, with a small opposed pinkish spot. Again and again a re- versal and change of poles took place and at least five such changes could be observed. As a result of this action the metal strip was rapidly covered over its entire surface with the same superficial, loosely adherent coating of hydroxide which is obtained in many cases when certain samples of iron and steel are allowed to rust under a layer of water. It is pre- sumable that as the surface of the metal is eaten into by the solution of the iron at the positive poles, a new condition of equilibrium occurs, resulting in changes and even reversal of the positive and negative nodes. This would indicate that in the case of metals which suffer from local action or pitting, the segre- gation conditions are of a different nature from those which exist in the case of metals which rust more evenly. A rough analogy may be drawn by imagin- ing an imperfect mixture of black and white sand, the respective grains of which may lie in streaks, spots and layers, or may tend to arrange themselves in some more or less uniform relation to each other. The best demonstration that the rusting and corrosion of iron and steel in all its forms is essentially an electrolytic phenomenon is afforded by the fact that it has not as yet been possible to find a specimen of such purity that no trace of positive or negative nodes will be formed in the ferroxyl indicator. Having now outlined the principle of the electrolytic theory, it will be instructive to consider it in its practical relation to the problems at hand and see if it agrees with observed facts. It was stated in the preced- 36 COEEOSION OF lEON ing chapter that colloidal ferric hydroxide be- haves like an ion and travels to the negative pole under the influence of a current, and since a current must flow, according to the theory, when iron corrodes we should expect the ferric hydroxide to be affected. Now anyone who will examine, under a glass, a piece of badly rusted iron will often find a place where the rust is piled up like a cone and surrounded by a ring which has been eaten into the metal, or perhaps, the metal has been dissolved out of a certain spot and the ferric hydroxide piled up all around it, like a crater. It can be assumed, therefore, that the cone marks a point of negative po- larity, while in the other case the depression in the centre of the ring is positive. It will be evident from what has gone be- fore that anything which tends to prevent the formation or flow of current also lessens the tendency to corrosion, a point which will be touched upon later in connection with insulat- ing and conducting pigments. So we should expect that very pure iron would be more re- sistant than that which is impure, and this is found to be true, at least in large measure. As evidence of the truth of this general prop- osition, it has been shown that if two nails, THEOEIES OF COEEOSIOIT S? say, are placed in water or f erroxyl and one of them is connected to a piece of platinum wire while the other is left free, the former will corrode much more rapidly than the other. This may be explained by saying that the hydrogen which is set free can deposit on the wire and thus be removed from the scene of action much faster than m the other case. In a similar way, when two pieces of iron, one of which is attached to a piece of zinc, are immersed together in water or some elec- trolyte, the free strip of iron will rapidly de- cay, while the piece connected to the zinc will be protected, although the latter slowly dis- solves. Zinc is electro-positive to iron ; there- fore a current flows from it to the iron, caus- ing hydrogen to deposit on the latter and thus protecting it. (Gaseous or atomic hydrogen must not be confused with that which is in the ionic state. It has been shown that the latter attacks iron and takes it into solution, but gaseous hydrogen, settling out in the form of bubbles on the iron, will make a thin film of gas which is a good protective from further action. This is illustrated by the "polarizing" of an ordinary sal ammoniac battery.) 38 COEEOSIOIT OF IRON" These examples help to show why zinc, as applied in galvanized coatings, is such a good protector of iron. Through its impermeabil- ity it keeps the iron away from harmful influ- ences as long as possible; then when mois- ture and air do reach it, an electrolysis is set up which effectually protects it until most of the zinc is destroyed. If a piece of iron oxide is substituted for the zinc in the foregoing experiment, it is found that the free piece of iron corrodes much more slowly than the one in contact with the oxide, showing that this must be electro-negative to the iron, therefore hasten- ing its corrosion. The coupling together of two parts, one of which is electro-positive to the other, with the result that the corrosion of the former is much accelerated, brings to mind a possibility which is ever present in iron structures : that is, through some cause or other one member may be positive or nega- tive to the others and, unless proper precau- tions are taken, all the conditions for early decay of some of the members are present. At least one instance of such action has been recorded, when certain members of some structural iron-work were almost entirely de- stroyed, while others were in practically per- THEORIES OF OOEEOSION 39 feet shape. There seems to be no doubt that the parts which were destroyed were electro- positive to the others. The obvious remedy for such a condition is to assemble only parts having practically the same chemical compo- sition and treatment and, further, to insulate them from each other as far as possible by coating all contact surfaces with red lead in oil. Mention has been made of the fact that stresses and strains, whether produced dur- ing the process of manufacture or afterward, as well as improper heat treatment and segre- gation of impurities, are great factors in de- termining the extent and magnitude of the differences of electrical potential which will be present. To these may be added the effect of pickling in acids, a process which is widely used in preparing metal parts for finishiug. Investigation has shown that iron which has been pickled may contain considerable quan- tities of hydrogen, being made markedly harder thereby, and is easily oxidized. In the experience of the writer, pickling is dis- tinctly bad and should be avoided whenever possible, especially in cases where even slight traces of rust are objectionable. Of course, this applies only to painted or japanned 40 COREOSIOIT OF IRON work, as all electro-plating processes demand some such preparation of the parts and bad results do not seem to follow. It is interest- ing to note, however, that electrolytic iron which contains a large amount of hydrogen is corroded with difficulty. Other considera- tions lead to the belief that the hydrogen in pickled iron is not present in very stable combination, while that in electrolytic iron forms a true alloy, which may be much more resistant than iron alone. Coming now to iron which is actually in service in buildings and similar structures, it is found that this often behaves in a rather unexpected manner as regards its tendency toward corrosion. Those parts immersed in a few inches of water may corrode with great rapidity, while other parts placed a consider- able distance under the surface will remain in practically perfect shape almost indefi- nitely. Similar conditions are found in parts buried in the ground. The explanation is to be found in the fact that both water and oxygen are necessary to produce corrosion. If, therefore, iron parts are immersed in water or buried in the ground to a depth greater than that to which oxygen can pene- trate, they will be preserved because the con- THEOEIES OF COEEOSIOIT 41 ditions for corrosion have not been attained. The literature contains many examples of such action and it is impossible to do more than refer to it here. Somewhat along the same liae is the ob- servation that iron in swiftly moving water does not corrode as rapidly as that in more quiet situations, dissolved oxygen being pres- ent in both cases. It is a matter of record that pipe lines corrode from the outside rather than from the inside. It may be con- jectured, on the basis of the electrolytic theory, that under these constantly changing conditions the nodes or points having a different potential are changed too fre- quently for serious rusting in one spot to take place. A little observation will show that "busy" iron, as it has been termed, does not rust as readily as that which is not in service. Per- haps steel rails are the commonest example of this and it is well known that little or no corrosion is seen in lines on which the traffic is fast and heavy. Sang thinks that this may be explained by the "asstunption that vibra- tion causes a shedding of the rust as soon as it is formed on the spots not protected by mill scale and there is, therefore, no accelera- 42 COREOSIOK OF IRON tion of the action due to the accumulation of spongy and electro-negative rust." It is apparent that corrosion does not al- ways take place with the same rapidity or uniformity; therefore certain processes or treatments may render iron less susceptible to decay, and the word "inhibitive" has been used by Cushman, in the sense of limiting or checking it. If a piece of iron is dipped into concentrated nitric acid for a short time, it is rendered passive; that is, it will stand a much longer exposure to adverse conditions than ordinary iron, and it precipitates cop- per from its solutions only very slowly. Just what the action is may be open to question, but apparently there is a lowering of the electrical potential. This state is the abnor- mal or unstable one, however, so the piece gradually returns to the normal condition and increasing corrosion takes place. Chro- mic acid and its salts also act like nitric a^cid and can produce the passive state. Work has been done in the way of putting these facts to practical use, and it may be that some effi- cient process will be discovered whereby iron can be made much more resistant. Other ways of inducing passivity are by electrolyz- ing in distilled water or caustic potash, some THEORIES OE COEEOSION 43 investigators using weak and others heavy currents. We have seen that hydrogen ions are stimu- lators of corrosion and hydroxyl ions are in- hibitors, so strong alkalis are perfect rust preventives. The explanation is that neither hydrogen ions nor hydroxyl ions can exist in the same solution in excess, in relation to each other. Now, any condition which tends to further the escape of gaseous hydrogen and thus eliminate its polarizing and damp- ing effect on corrosion acts as an aid to this process, so it has been found that under cer- tain conditions a paint film may actually in- crease, rather than diminish it. "Walker men- tions a case where the application of a high- grade varnish to the inside of tinned-iron fruit cans led to highly stimulated corrosion. This was explained on the basis that the coat- ing, which was probably perfect when made, was ruptured and destroyed at certain points during the process of manufacture of the cans. When the fruit juices set up electro- lytic action the varnish acted as a depolarizer and combined with the hydrogen, thus stimu- lating corrosion to the point of destruction. Thus it is that a film of linseed oil alone may encourage rather than prevent decay, since 44 OOEEOSION OF lEON it is porous and unsaturated cliemically. It has often been the custom in the past to coat structural steel work with linseed oil before it left the shop, and many cases of rapid cor- rosion can be traced to this practice. Direct experirgents have shown that iron parts which have been coated with linseed oil and the film then destroyed at certain spots by abrasion, corrode much faster than ordina- rily. This is probably due to the fact that the film of oil absorbs the hydrogen as fast as it is formed and thus acts as a depolarizer. If pigments are mixed with the oil, however, it loses this power to a great extent. A paint, therefore, in order to accomplish the purpose for which it was designed must be free from voids and of such a character that it will not act as a depolarizer. Another source of corrosion which must not be lost sight of is in wounds, indentations, and general roughness of the surface. Such peculiarities are almost always electro-posi- tive to the adjacent parts and deep pitting is produced. For this reason razors and other articles which have a high polish will, with the most ordinary care, resist corrosion and remain in good condition for many years, long after a rough piece would be destroyed. THEOEIES OF COEEOSION 45 It seems to be a principle, therefore, that the possession of a uniform surface, from the standpoint of chemical composition and free- dom from strains as well as the removal of all irregularities and the production of a more or less high polish, tends to make the part much less susceptible to the attack of corrosive influences than is the case with less carefully made articles. Chapter III PEOTBCTIVE MEASUEES A FTER a working knowledge of the causes ■^ ^ and general nature of corrosion is ob- tained, the next step is to apply it toward the protection of iron and steel from those malignant influences so familiar to every en- giueer. Since these materials can not stand exposure very long, while unprotected, it is evident that they must be covered with some substance or preparation which will exclude moisture and air and itself be impervious to the action of these agents. We may say that this represents the ideal and that in practice it is almost, if not absolutely, impossible to attain these conditions completely, although improvements are constantly being made in consequence of the great amount of work which is being done along these lines. Roughly, the problem of the protection of iron and steel falls into three classes: 1. Protection by a coating of oxide produced on the surface by various processes. 2. Protect- ee PEOTECTIVB MBASUKES 47 ive coatings of other metals. 3. Protect- ive coatings formed by paints and similar materials. Many years ago Lavoisier, a famous French chemist, noted the formation of mill scale, which is mostly the magnetic oxide of iron, and commented on its stability and im- pervious character. Later on Faraday de- scribed the protective action of the oxide formed in the tubes of a steam superheater. Another observer discovered its formation when iron is acted on by highly heated air, and Prof. Barff, of London, became inter- ested in this oxide as a protective measure through his observations of a pipe carrying superheated steam. His process, which was first made known in 1876, consisted in heating the parts to 1,000 degrees C. and passing steam superheated to 538 degrees C. over them. A Mr. Bower and his son tried using air instead of steam, but were unable to obtain the desired results until they employed pro- ducer gas to reduce the red oxide formed by heating the parts in air. The air treatment lasted forty minutes and the gas treatment twenty minutes, these being repeated alter- nately four to eight times. Finally they pur- '48 COKEOSIOK OP IKON chased the BarfP patent and the Bower-Barff process was originated. In this process the parts to be treated are heated to nearly 900 degrees C. in a closed retort; superheated steam is then led in for twenty minutes and a coating consisting of a mixture of the black and red oxides of iron is formed. Producer gas is now substituted for the steam and passed in for about the same length of time. If the coating thus formed is not sufficiently thick, these operations are repeated as often as necessary. Paraffine or some other oil is afterward applied to the parts, whereby a fine black color is obtained and additional protection afforded. As regards cost, this process is expensive, the price ranging from $5 to $20 per ton, but it can be applied to almost anything which will not be injured by the heat and is small enough to be placed in the furnace, and it af- fords very efficient protection, especially against sea-water, acid fumes, and like in- fluences. Paints and enamels can be applied and adhere strongly. Parts given this treat- ment will stand almost any degree of heat without injury, but they may not be bent or machined. Of course, the process is not ap- plicable to tools or tempered pieces. Also, PROTECTIVE MEASURES 49 the parts are enlarged slightly so that those which must maintain close limits on dimen- sions should not be given this treatment. An improvement introduced later into Am erican practice, known as the Wells proc- ess, consists in finishing the work at one op- eration, by using steam and producer gas together instead of applying them alter- nately. The Gesner process, which is a modification of the preceding, came into use a few years later. The coating is said to be a compound of hydrogen, iron, and carbon, and has less tendency to scale. Analyses have shown that not less than 2 per cent of hydrogen is pres- ent. The method of application consists in keeping the parts in the coating retort at a temperature of about 600 degrees C. for twenty minutes, then allowing steam at low pressure to act at intervals for thirty-five minutes. It is passed through a red-hot pipe in the bottom of the retort and thus is par- tially decomposed into hydrogen and oxygen. After the steam treatment, a small quantity of naphtha or other hydrocarbon is poured in and allowed to act for fifteen minutes, where- by any red oxide is reduced and the surface carbonated. 50 CORROSION OF IRON A variety of work has been finished by this process, apparently at a somewhat lower cost than for Bower-BarfiSng. It is also claimed to have an advantage over the latter process in that the much lower heat does not burn or warp the parts and that little or no iu- crease in size takes place. Perhaps mention should also be made of the Dewees-Wood process which was almost identical with that of Gesner and preceded it. In this, sheet iron was subjected to the ac- tion of hydrocarbon vapors, or gas, and superheated steam in a closed, heated cham- ber. The Bradley process was patented in 1908 and is similar in principle to the last two. In carrying it out, the parts are prepared by tumbling, pickling, or sand-blasting, pref- erably the latter, and heated to a low red in a muffle, hydrogen gas being passed in. This is followed later by small quantities of gaso- lene which improves the color of the coating. The articles are left in the furnace for an hour, or until a sufficiently heavy coating of the oxide is produced, after which they are taken out, allowed to cool, and treated with linseed oil or paraffine. The Bontempi process is another recent PEOTECTIVB MEASURES 51 modification or improvement on tlie original Bower-Barfl method. It consists in heating the article as in the Bradley process and then passing in steam and the fumes of zinc or some heavy hydrocarbon as tar or pitch. A new patent has been recently granted cover- ing the use of certain substances which are claimed to further the formation of the black oxide. A uniform finish of a deep black color is obtained, which is said to resist corrosion indefinitely. Among the several methods which have been used for producing a black oxide coating by the use of moderate heat, that of BufSng- ton may be mentioned. Briefly, it consists in melting manganese dioxide and potassium nitrate (saltpetre) in an iron pot, immersing the cold parts in the mixture for a few mo- ments, then hanging them over the pot in the fumes and finally putting them in boiling water to cool. A fine blue to bronze color is produced. In about the same way, films of sulphide or phosphide of iron, produced by subjecting the metal to the vapors of these elements at high temperatures in the absence of air, have also been tried, but have not attained to any very extended use. 62 COKROSION OF IRON Heat-bluing, which is familiar to every one, is applied to watch and clock hands, buttons, buckles, and a large variety of steel articles, and gives a finish of a pleasing blue or black color, which is not, however, so very resistant to corrosion. It is produced in several ways, such as dipping the parts in a bath of molten saltpetre, heating them on an iron plate in the air, or tumbling them in a sheet-iron bar- rel heated by a gas flame. Revolvers and similar parts are very often given a fine blue finish by heating in charcoal. One of the best and simplest methods of producing a protective black coating is of rather recent origin, and is known as Coslett- izing, from the name of the discoverer, Cos- lett. In this process the parts are first cleaned by pickling or sand-blasting, then im- mersed in a boiling water solution of phos- phoric acid in which iron or zinc filings are placed. They are left in this for a period varying from one-half hour to three hours, depending on the nature of the work or the thickness of the coating desired. A very slight amount of the surface of the article treated is converted into certain phosphates of iron, but most of the coating comes from the solution itself. One advantage of this PEOTECTIVE MEASURES 53 method is that it can be applied to small and delicate parts, tempered pieces, edged tools, and so on. Typewriter manufacturers and others who are using it claim that it retards corrosion to a remarkable degree, especially on articles which are in use in the tropics and other places where unusually good protection .is required. In summing up all the above processes, it may be said that Bower-Barffing or any other procedure which gives a heavy coating con- sisting essentially of the magnetic oxide of iron wiU afford a durable, efficient finish. It is not absolutely rust-proof, as experience shows that tiny rust spots will generally de- velop in the course of time. Under extremely bad conditions these appear sooner and doubtless the finish would go to pieces event- ually; but although this coating is strongly electro-negative to iron, a circumstance which we should think would tend to accelerate cor- rosion rapidly when it is once injured, the breaking down is very slow. Perhaps the perfect adherence of the oxide allows the iron to rust only at the exposed places, so that, although it may pit deeply, it is not so no- ticeable as though it were more extensive. As noted, however, those methods which em- 54 OOEEOSION OP IRON ploy high temperatures are not applicable to all classes of parts. Coslettizitig appears to be the best of those processes employing low degrees of heat, and it can be applied to practically any iron or steel article, giving it a durable finish. It ia not very complicated or expensive, although the cost is increased through the fact that it is a patented process ; therefore a royalty must be paid. Lastly, those methods which produce merely a darkening of the surface ordinarily do not afford much protection from rust, es- pecially on parts exposed to high humidity. They have their place, however, and when used in conjunction with lacquers or like ma- terials may be very useful in producing a fairly good finish. Coming now to those processes which de- posit another metal or alloy on the surface of iron, for the purpose of protecting it, it is found that those involving the use of zinc in some manner are by far the most impor- tant. Several factors tend to make this true, but we may mention the comparative cheap- ness of ziuc and the fact that it is peculiarly suited to this work, since it is the most elec- tro-positive to iron, as has been mentioned, PEOTEOTIVE MEASURES 55 of all the metals which can be used practically for such purposes. Unlike most metals, ziuc may be applied to iron objects in. three widely different ways : by dipping them in the molten metal ; by de- positing it electrolytically from a plating bath ; and by the vapor or Sherardizing proc- ess. All these methods are widely used, and each has advantages which may especially commend it in certain cases. In the hot or dip method, the parts are cleaned by pickling, generally dipped in some acid flux, and then immersed in a bath of molten zinc. Sal-ammoniac or other fluxing agent is used in addition to insure the clean surface requisite for perfect adherence of the coating. After the parts have become thor- oughly heated and covered with the zinc, they are withdrawn and the excess of metal is shaken or wiped off, dependiug on their shape. At first sight it would seem that a very pure coating of zinc is thus produced, but a little consideration will show that such is not the case. In the first place, the zinc bath is not strictly pure, but generally contains a fairly considerable amount of impurities in the shape of other metals, with perhaps slight 56 COKEOSION OF IKOK quantities of dissolved oxides or oxy-chlo- rides. Iron, too, is more or less attacked by molten zinc, so the galvanizing kettle and the parts themselves contribute a good deal of iron. After the iron content reaches a cer- tain point a definite alloy of iron and ziuc settles out, and to prevent it from adhering and becoming burned onto the bottom of the kettle it is the practise of many foremen to add lead to the bath. This melts and, being heavier than either the zinc or the alloy, set- tles to the bottom and everything else rests on top of it. The removal of the alloy, there- fore, is not difficult and is done at regular in- tervals. In addition to the foreign sub- stances above mentioned, many galvanizers add a small amount of aluminum or tin to aid in the production of a smoother and better looking deposit. It is easily seen, therefore, that the coating actually produced is a mixture of zinc with a number of other metals, and perhaps some actively corrosive bodies derived from the fluxes. The resulting alloy is very easily at- tacked by acids, due, no doubt, to the electro- lytic action set up between the various met- als. According to some authorities the solu- tion tension of zinc alloy with iron and PROTECTIVE MEASUKE8 S? other impurities is greater than that of the pure metal. When a hot galvanized part is exposed to adverse conditions, therefore, it is plain that unexpectedly severe corrosion may take place. A good deal of work is be- ing done to improve the quality of hot gal- vanized coatings, in the way of employing harmless fluxes and keeping the zinc as pure as possible, so it is likely that much progress wiU be made, although the very nature of the process is such that a very pure coating can not be produced, at least commercially. Nevertheless, this process is used very widely and enormous amounts of iron are finished by it. In using galvanized or other material which has been given a protective finish, it is necessary to have some means of determin- ing the probable efficiency of the coating as a preventive of rust, and this is done indirectly by measuring its weight and thickness. For anyone who has access to a good balance and a few chemicals, perhaps the simplest way is to clean and weigh a small piece of the ma- terial, and then boil it in strong caustic soda until the zinc is all removed. The piece is then washed, dried, and weighed again. Sub- tracting this weight from the former one 68 COEEOSION OP IRON gives the weight of zinc present, and by divid- ing it by the area of the part ia square inches the result can be expressed as milligrams per square inches or ounces per square foot. It is not always practical or possible, however, to inspect material ia this way, so another method known as the Preece test is in wide use. This depends on the fact that ziac, when dipped into a strong solution of a copper salt, is dissolved and takes the place of the copper, which appears in metallic form. Iron does the same thing as the zinc, but much more slowly, so that the copper deposit formed on the zinc is very coarse and easily removed while that on the iron is more or less hard and bright. Putting these together, when a galvanized piece is dipped in copper-sulphate solution, the copper first deposited is easily wiped off, but as soon as the zinc is all gone the copper plates out on the iron, this point being determined by the difference in ap- pearance and behavior of the coating. Prac- tically, it is sometimes a little difficult to tell just when all of the zinc is removed; also there are many factors which may work to impair the value of the results obtained, so exact directions have been drawn up cover- PEOTECTIVE MEASURES 59 ing the application of the test and the means by which the critical point may be recognized. Thus there is, ordinarily, a good degree of uniformity in the manner of using it. For ex- ample, the American Steel and Wire Co. di- rects that the copper solution shall be neu- tralized by an excess of pure copper oxide, it shall be filtered before using, and shall have a specific gravity of 1.186 at 65 de- grees F. Samples must be thoroughly cleaned of all grease and dirt and the temperature of the solution must be not lower than 65 degrees F. nor more than 70 degrees F. In making the test the carefully prepared samples are immersed in the fresh solution for exactly one minute. They are then taken out, rinsed in water, and wiped with clean cotton waste, after which they are immersed again for ex- actly one minute, washed and dried. If the material is required to stand more than two dips, the process is repeated the requisite number of times; then the parts are care- fully examined to see if there are any traces of bright copper. If no deposit appears, the parts are considered to have passed the test. Although this procedure finds wide appli- cation, there are several disadvantages or 60 COREOSION OF IRON faults in it, to which attention has been called by investigators. These will be referred to later, after the other methods of producing zinc coatings have been considered, inasmuch as this test has been applied to all these indiscriminately. It is only within comparatively recent years that it has been possible to obtain a good deposit of zinc electrolytically, on a commercial scale. A great deal of work has been done along these lines, however, so that we are now able to produce very fair coat- ings, although there is still plenty of room for improvement. In brief, the process con- sists in immersing the carefully pickled and cleaned parts in a solution composed essen- tially of zinc sulphate, with a very slight amount of free sulphuric acid. Various or- ganic substances are also added to assist in the production of a smooth, dense deposit. There is a large number of formulae for these baths, many of which are patented or secret. The plating time varies from twenty minutes to an hour or more, depending on the current density, the weight of deposit required, and so on. Some of the solutions permit the use of current densities as high as fifty amperes per square foot, which is much higher than PEOTECTIVE MEASUEES 61 was possible wlien the process first came into use. After the plating period is over, the parts which have been plated are washed in water and dried. The coating which is produced is of great purity; therefore, it would seem to be su- perior, in this respect at least, to that given by hot galvanizing. "When properly carried out, a very smooth, uniform deposit is ap- plied, so this method is applicable in many cases, as with small parts, where the hot proc- ess would clog up small holes or threads and prove generally unsatisfactory. It also possesses the great advantage of very easy control — ^that is, a coating of definite weight can be produced at will. On the other hand, the deposit is more or less porous, depending on the conditions un- der which it is made, and certain impurities may be included which will become active stimulators of corrosion. As far as protec- tion is concerned, there does not seem to be much to choose between the hot dip and elec- tro methods, when testing parts of the same weight of deposit. As before, the actual weight of zinc per unit area may be determined by stripping in caustic soda, weighing the parts before 63 COEEOSION OF lEON and after stripping, and dividing the differ- ence in weight by the area of the part. In testing electro-galvanized work to see if the coating is of the desired weight, sam- ples are often suspended in distilled water through which air 'is bubbled, the time re- quired for the first sign of rust to appear be- ing carefully noted. Under such conditions the zinc is converted to the hydroxide or basic carbonate and settles to the bottom of the vessel, exposing the iron. In making the test it is important that certain definite con- ditions be always maintained, especially as regards the rate at which the air is passed, if comparative results are to be obtained. In passing, it may be said that the value of such an acceleration test, as a measure of the efficiency of a coating in protecting iron from corrosion, may be a doubtful quantity but it does roughly indicate the weight of the coat- ing. The Preece test is also applied to elec- tro-galvanized work, and when properly car- ried out should be of great value in determin- ing the quality of such a finish. Sherardizing was developed by Sherard Cowper-Coles, although some of the funda- mental facts regarding it have been known for a long time. In this process the thor- PEOTECTIVE MEASURES 63 oughly cleaned parts are placed in an iron drum with a quantity of finely powdered zinc, whicli generally contains a considerable amount of zinc oxide. This drum is then placed in a gas-fired or electrically heated furnace and brought to a temperature of about 800 degrees F. which is maintaiued for a varying length of time, depending on the thickness of coating desired. On the average a furnace will turn out two charges a day. "When the drums are sufficiently cool they are emptied over a grid or screen so that the zinc dust falls through and collects in bins, to be used again. One great advantage of this process is the fact that the zinc coating is applied very evenly and uniformly over all surfaces ac- cessible to the powder. In this respect it is superior to the methods previously consid- ered, especially the electrolytic, with which difficulty is experienced in getting the zinc to deposit iu inner corners and recesses. The size of the parts is increased very slightly by this treatment, but perhaps the main action is one of alloying with the iron rather than depositing zinc on the surface. Dr. Walker says that a polished and etched specimen of Sherardized iron presents a relatively 64 COEEOSION- OF lEON complex structure. The zinc penetrates the iron with the formation of deep layers of two alloys B and C, the latter being richer in iron than the other. Imposed on B there are a number of more or less unknown alloys containing varying amounts of zinc and iron. Upon the surface there is generally a layer of relatively pure zinc, although frequently the process is carried to the point where only a deep layer of alloys is formed. "When ex- amined under a microscope this alloy is seen to be covered with cracks or fissures, as though it had contracted in forming. The claim is made that even if the outside coat- mg, which is rather brittle, is removed by blows or rough usage, the part will still be protected by the deeper zinc-iron alloys and will not rust, and practical experience seems to bear this out. Further, if a deep notch is filed in a piece of Sherardized work and the latter is then exposed to the weather or placed in aerated water, rust will form in the notch and fill it up completely, then stop. It apparently can not crawl or loosen the surrounding coating. "When parts finished by this process are exposed to the weather or immersed in water, a yellowish-brown color which may easily be PEOTBCTIVE MEASUEES 65 mistaken for rust develops. TMs is called the curing color and is probably caused by the corrosion of part of the iron in some of the alloys. At any rate the surface rapidly turns darker and soon becomes almost black, from the formation of the magnetic oxide which is in itself a good protector. On con- tinued immersion a few weak spots will gen- erally show up and tiny rust spots form, but as remarked, these do not spread and for all practical purposes may be disregarded. Altogether, the work produced by this method seems to be as nearly rust-proof and durable as can be produced by commercial means at the present time. The fact that it is adapted to the finishing of both large and small parts, even those having threads, as screws or pipe, greatly increases its applicability and it is being used very widely. In the "molten zinc" process the parts are placed in a hollow drum which is placed in- side a larger one. The zinc is heated to the volatilizing temperature, while hydrogen or other reducing gas is forced in to prevent it from oxidizing, and it combines with the iron, producing about the same results as Sherardizing. Aside from its use as a means of affording protection, this method is also 66 coKEOsioiir of iron capable of application in decorative work. It is possible to get very beautiful effects by forming alloys with other metals — ^for ex- ample, the production of a brass design on a copper background. Another procedure, of which only the bar- est mention can be made, has been developed for finishing wire. It consists in passing a heavy current through the wire, thereby heating it to redness, and drawing it through a mixture of crushed coal and ziuc oxide in a furnace. The heat reduces the oxide to the metal, which then alloys with the iron of the wire, forming a coating similar to Sherardizing. Whether or not this process will come into very extended use remains to be seen. Some little reference has already been made to the Preece test for galvanized work. As a whole it has been objected to on the ground that it assumes that the zinc is re- moved ia direct proportion to the time of im- mersion and that it really determines the thickness of the coating at its thinnest part. "When applied to electro-galvanized or hot dipped work it is a fairly satisfactory means of indicating the thickness and uniformity of the coating, and meets commercial require- PKOTEOTIVB MEASURES 67 ments of speed and reliability. When ap- plied to Sherardized material, however, it is the opinion of many investigators that very misleading results may be obtained, in spite of the fact that some manufacturers are us- ing it. The trouble apparently is that some of the zinc-iron alloys are indicated as iron, while in reality they have very good pro- tective powers. Perhaps if carried out un- der very definite conditions and making cer- tain allowances it would prove of value as a ready test, but there are other means of de- termining the quality of the coating and these are much more accurate, though longer. A very good way is to strip a piece in caustic soda, as described before. Immersion in water is also a reliable means of testing all such coatings. One of the largest producers of Sherardized material in the country is said to test all of its product by subjecting samples to a strong salt-water spray for something like sixty hours. It should be remembered that all the fore- going methods of testing ia which the weight of the deposit is determined do not give much information regarding the freedom from pores. Walker therefore devised a scheme of testing electro- and hot-galvanized ware 68 CORROSION OP IRON for pinholes in the coating by immersing the part in strong, hot caustic soda. Any such imperfections will then be revealed by the formation of streams of small gas bubbles. Short and simple tests of quality are a necessity and are of real value when rightly used, but the best method of all is by actual exposure to the weather. It requires con- siderable time to obtain definite results, but when it is possible to do this it is generally worth the time and trouble, as it decides im- partially and unmistakably between various finishes or different samples of the same finish. Of the other metallic coatings produced by electro-plating, perhaps copper and nickel are the most familiar. These are applied both for decorative purposes and to afford protection from corrosion, and are used very widely for finishing an infinite variety of small and medium sized parts, although there are certain disadvantages connected with the use of such coatings. The processes for pro- ducing them are rather expensive and call for a high degree of skill ; to be really effective under bad conditions a heavy deposit is re- quired, and even then there is a decided ten- dency to wear away, especially on edges, if PROTECTIVE MEASURES 69 the parts are subjected to much handling. For many purposes, however, no better methods of protection have been devised. Copper is deposited from two kinds of baths, acid and cyanide. The former is es- sentially a solution of copper sulphate with a small amount of free sulphuric acid, and is used for plating on any metal which is not attacked by dilute sulphuric acid. Iron, for example, can not be plated in it because it reduces a copper solution, as we have seen, and is itself dissolved by the acid, so the cop- per cannot adhere to it and the cyanide solu- tion must be resorted to. This is made by dissolving basic copper carbonate in a slight excess of potassium cyanide. It does not at- tack iron or other metals to any extent and rapidly produces a fine deposit. Nickel solutions are generally made from nickel sulphate ; some platers use the double nickel-ammonium salt, a small amount of free acid being present in both cases. Boric acid is added to brighten the deposit. Among the metallic coatings produced in other ways, mention may be made of copper- clad steel, which seems to be finding a con- siderable field of usefulness, perhaps espe- cially in the form of wire. Space does not 70 COEEOSION OP lEON permit a description of the process beyond saying that the copper is applied in one of two ways, by alloying or by welding. In the former the copper is in the plastic state, while in the other it is molten. Copper is very resistant to atmospheric influences, and exposure tests of copper-clad wire have shown it to be practically unchanged after long periods. We should naturally expect that the con- tact of steel with copper would cause the former to corrode very rapidly, but this does not seem to be borne out by experiment and it has been conjectured that after a cer- tain amount of rust has formed a thin film of copper oxide mixed with copper is deposited between the iron oxide and the unattacked iron and that this acts as a preservative coat. Altogether it would seem that such copper and iron combinations should come into wide use, especially as the former metal is not destroyed by exposure to moisture and air as is zinc. Lead has been used to some extent, and the Lohmannizing process, which is perhaps the best kaown of these, is said to consist in the immersion of the parts in a bath of molten pure lead or lead alloys, the secret PBOTEOTIVB MEASURES 71 being in the use of the proper fluxes. While an alloy of lead and tin is used in the manu- facture of terne plate, it is somewhat doubt- ful if the use of pure lead in this way will come to be very extensive. Another procedure, which differs consider- ably from any of those so far discussed and presents some novel features, is that known as the Schoop process, wherein some metal is sprayed by means of an air blast onto the parts which it is desired to protect. The apparatus for doing this is more simple than might be supposed and consists of a sort of pistol, in appearance not unlike an automatic revolver, having three hose leads, through which are conducted oxygen, hydrogen, and compressed air. Inside the pistol the two gases are burned in a burner tube concentric to the air nozzle. Metal is supplied in the form of a wire, which is drawn in by a suitable feeding de- vice actuated by a small air turbine in the body of the pistol. When the wire reaches the flame zone the intensely hot oxygen- hydrogen flame instantly melts it, while at the same time a strong blast of air issues from the nozzle and blows the molten metal into a cloud or spray, so that it is discharged 72 COEEOSIOBT OP lEOIT from the pistol with a velocity said to be approximately 3,000 feet per second. It might appear at first sight that the metal would reach its mark in a molten state, but this does not seem to be the case since a few inches from the nozzle it is of fairly low temperature, low enough to be directed upon the hand for a moment, or upon silk or other cloth, or even matches. Perhaps the tiny particles are really solid when they strike the part under treatment, but are liquefied by the heat of the collision. Whatever the ac- tion may be, however, the fact remains that a dense, strongly coherent coating is pro- duced, which is said to be amorphous and vitreous, rather than crystalline. Any desired weight or thickness of deposit may be obtained at will and almost any metal or alloy can be used, so the process would seem to have a wide field of usefulness be- fore it, both from protective and decorative standpoints. While it is yet comparatively new and has not come into extended use, there does not seem to be any good reason why it could not be employed to coat bridges or other forms of structure work, either wholly or at any desired point, as the neces- sary apparatus is not very cumbersome or PEOrEOTIVE MEASURES 73 expensive. It is claimed that lead can be sprayed for less than two cents per square foot and one pound can be deposited in less than one minute. Various other alloys and processes are be- ing proposed from time to time and doubt- less some of them will prove to be of real value, but so far as the writer is aware none of them has displaced to any great extent the older protective methods. The protection of iron and steel by paints and similar materials will be considered in succeeding chapters. Chapteb IV PAINT MATBEIALS TN the preceding chapter the most impor- ■*■ tant protective coatings for iron and steel (other than those produced by paint and similar substances) were considered, and now we come to the study of these latter ma- terials. Galvanizing, plating, and the other methods which have been discussed are very widely used and are of the greatest value, but the use of paint is hardly secondary to these and is worthy of the deepest attention. Per- haps one reason why this is true is that the plating methods previously referred to are not always applicable, for some cause or other. For example, it would be impractica- ble, both from cost and manufacturing stand- points, to galvanize or copper-plate the mighty members of a railroad bridge or simi- lar structure. Also, in much of the piping, railing, machinery, and other iron work of everyday life, it would be altogether too ex- PAINT MATBIEXAIjS 75 pensive to apply a plated finish, and even if this was done it would be necessary to use a finishing coat of some other material, if only for the sake of appearance. In other words, the methods previously described are limited in their applicability by their cost and the size of the parts which can be treated. In all such cases, therefore, recourse must be had to some form of paint. Everyone knows, in a general way, that paint is composed of certain pigments sus- pended in linseed oil, with perhaps the addi- tion of a drier and some turpentine or ben- zine, and that it is put on with a brush. In order to study paint intelligently, however, it will be necessary to review some of the facts regiarding its manufacture and the in- gredients composing it. In general, there are two ways of making paint. In the first, the pigment is ground in a mill with only enough linseed oil to form a stiff paste, which is thinned when needed for use by mixing with more oil, turpentine, and drier. A little varnish may also be added to improve the quality of the film. In the other method, the pigment is thoroughly mixed with the amount of oil necessary to make a paint of the desired consistency, and 76 COEEOSION OF IRON this mixture is then ground by running it through a paint mill several times. Both of these procedures have certain ad- vantages, but it is doubtful if the pigment and oil can ever be as thoroughly incorpo- rated, when the first method is used, as when they are ground together; so when the best results are expected it is better to obtain the paint ready mixed from the factory. An- other point in favor of this is that such a paint is far more likely to have all the iagre- dients mixed in the right proportions. For everyone who uses paint, and espe- cially for engineers who may be called upon to draw up specifications or make recom- mendations regarding it, some knowledge of the materials employed in paint making and the processes by which they are produced is necessary, and the principal ingredients found in the ordinary paints used on metal work will be briefly considered. The one substance which is common to all paints is liuseed oil. This is obtained either by grinding and pressing the seeds of the flax plant or by extracting them with naphtha. The pressed oil contains a good deal of for- eign matter, "foots," which may be partly removed by filtration, the rest settling out on PATN-T MATEEIAIS 77 long standing. If the oil is painted upon a strip of glass or other surface and exposed to the air, it will slowly absorb oxygen and become converted into a tough film. With raw oil this forms too slowly, however, for practical purposes, so recourse is had to the process known as boiling, whereby the drying properties of the oil are greatly increased. Heating it to a temperature of 400 degrees F. to 500 degrees F. has the effect of making it considerably darker and more easily dried. The addition of salts of lead or manganese to oil heated to a high temperature also pro- duces this effect, and this was the old method of boiling oil. A very dark oil, however, is objectionable for many purposes and it is now the practice, at least at many places, to make a drier of the metallic oxides by heat- ing them at high temperatures with a small quantity of the oil, and then adding the residue, dissolved in benzine or turpentine, to the main body of oil kept somewhat above the boiling point of water. By this method fuel, time, and oil are saved and the result- ing product is lighter in color than with the other processes. Eegarding the quality of the oil thus pro- duced there seems to be a considerable differ- IIS COEEOSION OF lEON ence of opinion. Some people assert it is as good in every way as that which has been heated to a high degree with the salts or oxides; others maintain that an inferior grade of oil is produced. Amid these con- flicting views it is hard to decide, but it is probable that when it is made into a properly designed paint it will give satisfactory serv- ice, at least in many cases. In making up a paint a certain percentage of raw oil is generally added to the boiled oil, or raw oil may be used alone, a small quantity of a japan or oil drier being added in each case. Japan driers are made by melt- ing resins with salts and oxides of lead and manganese or other metals, afterward thin- ning the mass with turpentine or light min- eral oils. The action of such a drier is very rapid. Oil driers are made, as noted be- fore, by heating linseed oil with metallic salts or oxides to a high temperature and thinmng with more oil and volatile solvent. The oxides of lead and manganese are most widely used in the making of driers. Man- ganese salts start the drying action and cause the surface to dry quickly. Lead, on the other hand, causes the oxidation to proceed through the film and is generally used in the PAINT MATERIALS 79 larger proportion. Red lead makes paint films brittle; litharge tends to make them very elastic. A raw linseed-oil paint dries so slowly that it is impractical and not suited for ordinary use, although it would produce an excellent film in time. The addition of a certain amount of drier is justifiable and necessary, therefore, but it must be borne in mind that at least some action continues long after the oil is dry and eventually tends to bring about the destruction of the film. The kind and amount of drier used, accordingly, has a great influence on the life and durability of the paint and close attention should he paid to this point. It is poor economy to sacrifice a year or two of the usefulness of the paint film in order to make it dry a few hours sooner. On the whole, the oil driers do the least harm, but even they should be used in moderation. Since the tests for the purity of linseed oil are to be found in many text and refer- ence books they will not be given here. Ac- cording to price conditions and so on, min- eral oils, turpentine, rosin and rosin oil, and com, fish and cottonseed oils may be used to adulterate it. The detection and estima- 80 COEROSION OF IRON tion of these is the work of the analytical chemist and further consideration of them would be out of place here. Among the other oils which may be added to, or used as substitutes for, linseed, per- haps tung or China wood oil is the most im- portant. When properly boiled and treated it yields a film which is hard and elastic, with heavy body and high gloss. One great ad- vantage of this oil is that it forms paints which will dry in damp atmospheres. It has long been used by the Chinese and Japanese and is finding wide use for marine and water- proof paints, and there is no apparent reason why it should not be used more extensively for protective paints for iron and steel. Soya bean oil is obtained from the soya bean which grows in Manchuria, although it will also grow well in many parts of the United States. It is very similar to linseed oil, so much so that it is difficult to detect its presence in the latter. Efforts are being made, and with success, to utilize it in the manufacture of paint. Menhaden or fish oil will be mentioned, as it is used in special paints which are required to stand heat and light, as on smoke-stacks. Such a paint is considerably more expensive PAINT MATERIALS 81 than a linseed-oil paint, so it can be seen that menhaden oil is a valuable aid and not an adulterant, as often claimed. Proper treat- ment of the oil is very expensive, as the loss by evaporation is large and certain volatile products are formed which are very offensive to the workmen. Mixed with linseed oil, it is used to some extent for making water-proof paints for various purposes and seems to give good satisfaction. Coming to the materials used for thinning and diluting paints, turpentine is the most valuable. It is the product of the sap of several species of pine trees growing in the South. The sap is collected and distilled with steam, the turpentine passing over and the rosin being left behind. Foreign turpentines differ somewhat in specific gravity and odor, but have the same composition. When added to paint, turpentine causes the latter to work more easily and increases the spreading power. It evaporates slowly and completely without leaving a stain, and causes more rapid oxidation of the paint oils. The one objection to turpentine is its compara- tively high price, and many substitutes are on the market, as such or as adulterants. Some of the best of these are light mineral 83 CORROSION OP IKON oils derived from the refining of petroleum and other substances. In their flash and boil- ing points they resemble turpentine, possess the same flowing qualities, and apparently are just as good. At any rate they are cheaper and are being used widely. One of these, benzol, may be especially mentioned as it is used very largely in coal-tar and asphaltum paiats and in paint and varnish removers. Experience has shown that the value and durability of any paint depends in large measure on the pigments used, and this is especially tru6 of protective paints for iron and steel. It has been known for many years that certain pigments give better results than others, but it is only recently that efforts have been made to determine the action of some of these materials. Considerable work has been done on this subject and it has been possible to classify pigments as stimulators, inhibitors, and inerts or indeterminates, on the basis of extended researches carried out by several investigators. Finally the Ameri- can Society for Testing Materials decided to undertake a similar series of tests and the re- sults were so interesting and instructive that they are given on pages 84 to 86. Briefly, the method of conducting the test was to PAINT MATEUALS 83 place pieces of carefully weighed steel in a series of bottles, each of which contained a pigment suspended in distilled water through which air was bubbled. After about three weeks the steel pieces were cleaned, dried, and weighed, the loss in weight being care- fully noted. In performing this experiment several investigators worked independently and these results are taken from the Proceed- ings of The American Society for Testing Materials, Vol. 10, 1910. In considering these results it must be borne in mind that they are merely indica- tive of the general nature of the pigments when suspended in water, and it does not necessarily follow that oil paints made from these would show up ia service in exact agree- ment with the recorded test. A pigment sus- pended in a paint fihn is under vastly dif- ferent conditions from one in water, but the results of practical tests are in broad accord with the tabulated data, and it ap- pears that those pigments which, in water suspension, depend upon the inhibitive action of ohromate salts or the presence of hy- droxylions, through hydrolysis, will, in gen- eral, give a good account of themselves when made into an oil paint. 84 OOEEOSIOK OF lEON P a « o> 00 >o o» us w m P) t- CJ >o CO 09 c» -4 "S b g bO iHC ■ -I— . -- b KM '-ig'3 5 5«a ooooooi-it-ii-ir-ji-iwNwwe^NNNNeii nc5p>oroo5-*t~ •>-' -riit-t-t-NO • •m eoc05piO'^0'*eocD -i-i •t-(Nwo5"3N " •»h OOOOOi-tO(N(N -(M • (N N C« >-l(N iH • -G* m S ^^Seo^QOcoing; 3 2 ©"Ort-HOOOOO >nOOOlONt; COTj1C0 05CO(»»O^OSOCPOOCDt:-0 CONrHt^COlflf-tOOlO-^^COCOOSOi i-lrHf-tTHf-trHTHi-lt-li-ti-tC*Cfli-trH PAINT MATERIALS 85 § CO CO (X> t- 00 o: c eOT-<'^.^>Oi-iioMNto^eo^t^50eqtDNeO!Otoo»u3 -NsOTitt* g lO iS 3 -^ tJI CO -^ (N CO t- i-H CO CO r-< •* 00 j< m T-< N C<> M (N >0 ■* Tl< cc CO ^ CO iffl CO lO t» ■^ 3 -WiOlOiO tH i-H eq T-l N CO N MUM (N rH (M T-l »H (N CO C<1 N CO CO CO CO C^ ■* -COiflti t»OtDinu5iHcDi-lOOinc005t-iOIMQ0>00C0t»tDI>C0 •lOCDoO CO t* W3 TjH N O CO CO W T-l lO CO CO -^ »0 CO O lO 1-H 00 CO 05 CO O 'lOQOi t~ CO (»■ (N O CO CO CTl r~ 05 U5 ■^ t- i-H 00 m 00 >0 O 00 •* 00 CO ^ -ClOOO CO CO ■* CO Ttt Tji -^ in ■^ ■n >o i> CO oo 'coooeo OiOC5Coco>!i<-<*eo'n'>j.ooo50<-ieocot>.ooojO'-!geo^igSi2002 CORROSION OF IRON ,^ I l-ill *£) "!^ ^ rt n OS l-^ OnUOPQ I I? ^ ^ :§ I -2 .^ s I sS tS ;3 O P5 Hi >■ (5 III a a ao (^ a n *£] CD 3 :| 5 I i a ^ a) 03 1-^ PAINT MATBEIA18 87 Regarding the method of manufacture of some of these materials, zinc chromate is a yellow pigment made from zinc salts and potassium dichromate. It is fairly soluble in water and generally contains other chro- mates and ziuc oxide, with some impurities. It has a specific gravity of 3.5 and grinds to a paste in 25 per cent of oU. Considerable drier is required as it is a slow drying ma- terial, but it has proved to be one of the most inhibitive of all pigments in use and is valuable in even small amounts in protective paints. Its rather high price is its only dis- advantage. Zinc and barium chromate is made by pre- cipitating a solution of the chlorides of these metals by sodium chromate. It is less soluble than the zinc salt and is a very good inhibi- tive. Zinc oxide is one of the most valuable white pigments and is produced by roasting and subliming certaia ores of zinc. It is very opaque, has excellent spreading qualities, and is generally mixed with basic carbonate of lead — ^white lead — ^for use in white paints. By thus combining these, certain undesirable qualities of both are overcome. For instance, zinc oxide dries to a hard surface which does 88 COEEOSION OF IRON not stand changes of temperature as well as the softer lead pigment. On the other hand it tends to eliminate the tendency to- ward chalking which the lead possesses. It is widely used as a base for delicate colors and in enamels. As has been shown, it is a fairly good inhibitor. Zinc-lead white is made from zinc-lead ores containing sulphur. The metals are vola- tilized and oxidized, forming a white smoke or fume composed of about equal parts of zinc oxide and lead sulphate, which is col- lected in large bags. When placed on the market it is extremely fine, has a specific gravity of 4.4 and grinds in 12 per cent of oil. It is often mixed with white lead and zinc oxide and inert pigments, is very stable chemically, and generally possesses inhibitive qualities. Barium chromate is pale yellow in color and is made by treating barium chloride with sodium chromate. Its inhibitive value is not particularly good, probably because of im- purities. Ultramarine blue is made by calcining silica, China clay, soda ash, and sulphur in furnaces and grinding the product. It is of a bright blue color, has a specific gravity of PAINT MATERIALS 89 2.4, and grinds in 30 per cent of oil. It tends to turn darker on iron on account of tlie for- mation of sulphide of iron, from tlie sulphur in it, and is not considered a good inhibitive. Chrome green (blue tone) is made from lead nitrate, sodium chromate, and sulphuric acid, precipitated on white lead and Chinese blue. It contains a variety of impurities, whereby its protective powers are lowered, is a fairly heavy pigment, and grinds in 25 per cent of oil. Prussian blue is made by mixing solutions of iron sulphate and potassium ferricyanide and oxidizing the resulting precipitate. It is a pretty blue, has a specific gravity of 1.9, and grinds in 55 per cent of oil. It may or may not be inhibitive in character, depending on its freedom from impurities; therefore it should be tested before use. It exerts a marked preservative action on the oil with which it is combined so that a glossy surface is presented even after long exposure. Lithopone is produced by mixing solutions of ziao sulphate and barium sulphide The resulting precipitate consists of zinc sulphide and barium sulphate ; it is filtered off, heated to a high temperature, and dropped into water, whereby it is thoroughly disinte- 90 CORROSION OF IRON grated. It is washed and filtered again, dried and ground. All of these operations require a good deal of care and attention to produce a satisfactory pigment. Lithopone is very stable and is probably the whitest pigment known, being very widely used ia high-grade enamels. Exposed to light and dampness it darkens, but will often regain its whiteness again. It is, therefore, not suited to outside use, by itself, but needs to have a large amount of some stable pig- ment combined with it. It has a specific grav- ity of 4.25 and grinds in 13 per cent of oil. As an inhibitive it stands fairly well, but, as noted, is not adapted to outside use unless mixed with other pigments. Willow charcoal is made by charring cer- tain kinds of wood, and contauis a slight amount of alkali, which is probably the cause of its good inhibitive powers. It is a light pigment and grinds to a paste in 33 per cent of oil. Litharge, lead monoxide, is formed by heat- ing lead intensely for several hours. It is yellowish red, very heavy, and grinds in 9 per cent of oil. It finds extensive use in the manufacture of boiled oils and is an excellent inhibitor. PAINT MATERTATiS 91 Basic carbonate-white lead is made ia sev- eral ways. The old Dutch process consists in placing lead plates or grids in clay pots with dilute acetic acid, stacking the pots up, and covering with tanbark. Fermentation of the latter causes a rise in temperature and the production of carbon dioxide, which acts on the lead in the pots and gives a basic lead carbonate. At the end of two months the white lead is ground in water and dried, be- ing ground later in oil. The "quick process" requires but two weeks and is carried out by acting on finely divided lead with dilute acetic acid and car- bon dioxide, from burning coke, in revolving cylinders. In the "mild process" no acid is used, the finely divided lead being agitated in water through which air is blown. Hydrate of lead is formed and later carbonated. A very good product, free from all acid, results. The result of all of these methods is a very valuable pigment of specific gravity 6.8, grinding in 9 per cent of oil. It is very opaque and has much body, but is rather low in spreading power and is generally mixed with zinc oxide or pigments of similar na- ture. Sulphurous gases blacken it easily and 93 COEEOSION OP lEON it has a great tendency to chalk, due to the fact that it is naturally alkaline and thus acts on the linseed oil in the paint. It is generally inhibitive in its action and is widely used. Calcium sulphate, gypsum, is found in na- ture and is widely used in paint manufacture. It is very stable, but is somewhat soluble in water and has a tendency to wash out of paint films. Since it is easily ionized in the presence of water, with the formation of large numbers of hydrogen ions, it should not be used on iron or steel as corrosion is certain to occur. Prince's metallic brown is one of a number of iron ores which are used in paint manufac- ture. It is found as a natural hydrated iron oxide, or as a carbonate, and is prepared by roasting at a red heat and grinding. It contains a good deal of silica and alumina and is considered a standard pigment for protective paints. Orange mineral has the same composition as red lead, but a different tone, and is pro- duced by the oxidation of white lead. Owing to the variety of impurities which it contains its inhibitive power is variable. Calcium carbonate, whiting, is found exten- sively as chalk. "When properly prepared it PAINT MATEEIALS 93 has a specific gravity of 2.8 and grinds in 20 per cent of oil. It spreads well and is used to quite an extent, partly on account of its power of neutralizing any free acid in the linseed oil. The artificial form of this ma- terial is lighter and requires more oil for grinding. It generally contains impurities which make its use in protective, paints in- advisable. Sublimed blue lead is made by burning a mixture of galena and soft coal. The fumes are drawn through cooling pipes and into large bags. The pigment has a specific grav- ity of 6.4 and grinds in 10 per cent of oil. It is bluish black in color and has given very good results when mixed with iron oxide or lampblack. Lemon chrome yellow is a mixture of sul- phate and chromate of lead and is made by precipitating a lead salt by bichromate of sodium and sulphuric acid. Owing to its occluded impurities it is not a good inhibitive. Orange chrome yellow is essentially a mix- ture of the neutral chromate and basic chromate of lead, and is made by precipitat- ing a lead salt by chromate of sodium and an alkali. It has a specific gravity of 6.9 and grinds in 20 per cent of oil. It contaias a 94 OOEEOSION OH lEON variety of impurities, but its inhibitive power is good if these are not acid in character. Medium chrome yellow is the pure neutral chromate of lead and is made from a lead salt and sodium chromate. Its specific gravity is 5.8 and it grinds in 30 per cent of oil. The strength of its color depends on the method of manufacture, and it is used as a tiating material. It is liable to carry down impuri- ties during the precipitation and therefore does not possess very reliable inhibitive pow- ers. Chrome green, oxide of chromium, is made from chromium salts. It is much used in rail- road work as it is very permanent and stable. In addition to other desirable qualities it is inhibitive to a slight extent, but its high cost prohibits its use in many cases. Venetian red is essentially ferric oxide and calcium sulphate, produced by heating iron sulphate and lime. It is not considered a good material to use on iron work on account of its content of calcium sulphate, which, as noted before, is fairly soluble in water and ionizes easily. Some varieties of this pigment are made by mixing oxides of iron with calcium carbonate and sulphate and, as they are free from acid, are safer for use. PAINT MATEEIALS 95 Bone black is made by heating bones to a high temperature for several hours and grinding the residue. It contains a large amount of calcium phosphate and carbon, has a specific gravity of 2.68, and grinds in 50 per cent of oil. It is generally an excellent inhibitor and often replaces carbon black and lampblack in dark inhibitive paints. Asbestine is a natural silicate of magnesia, occurring also in the form known as talcose. Both varieties are very stable and are used in paints, largely to prevent the settling of other pigments and to strengthen the paint film. They are light pigments, grinding in 32 per cent of oil. China clay is really aluminum silicate and is found in nature. It is a very permanent, fine, white powder and is much used in paints. It is a light pigment and grinds in 28 per cent of oil. Red lead is a bright red pigment made either by oxidizing litharge in furnaces or by heating it with sodium nitrate in iron pots. According to conditions of manufacture and other details the color varies somewhat. It is widely used for the protection of iron and is considered to be one of the best pigments known. It is generally mixed with oil, when 96 CORROSION OP IRON required for use, in the proportion of thirty pounds of pigment to one gallon of oil. It exerts such a drying action on the oil that no other drier is necessary. Sulphurous gases tend to turn it brown and it is often mixed with certain inert materials. Its inhibitive power varies with the method of manufacture and the impurities which it contains. Indian red is the name applied to certain hematite ores. They vary in shade and phys- ical characteristics, but are much used in in- hibitive and other paints. American vermilion is essentially a basic chromate of lead, being formed by boiling white lead and sodium chromate, with the addition of small amounts of sulphuric acid. It has a specific gravity of 6.8 and grinds in 16 per cent of oil. Its very good inhibitive properties are probably due to the presence of free chromates and would probably be increased by omitting the acid treatment. Sublimed white lead is a basic sulphate white lead made by volatilizing galena, lead sulphide. The fumes are oxidized in the air to a basic sulphate of lead and are drawn through cooling pipes to bags. It is very fine pigment of specific gravity 6.2 and grinds in 10 per cent of oil. It is extremely stable in PAINT MATEEIALS 97 every way and is not blackened by sul- phurous gases, as are many pigments. Mineral black is made by grinding certain forms of slate and is mainly used as an inert filler for paints. What is called Keystone filler is made from a bituminous schist ore and contains a large amount of silica, with alumina and carbon. It is used widely on iron and steel, especially machinery. Barytes, barium sulphate, is found in large quantities in nature and is used extensively in paiuts. It is very stable and can be used as a base for the most delicate colors. It is very heavy and grinds in 10 per cent of oil. Blanc fixe is the artificial form and is made by precipitating a barium salt by a solu- ble sulphate. In some ways it is a more valuable pigment than the natural form. Both varieties may contain acids and should be tested before application to iron sur- faces. Graphite occurs in two forms, the natural and the artificial. The latter is purer than the other, but both have about the same spe- cific gravity and grind in 45 per cent of oil. Both have an excessive spreading rate and are generally mixed with some of the lead or other heavy pigments. On account of 98 CORROSION OF IRON their conducting power they are not con- sidered good inhibitors and should not be used as contact coats, although otherwise they form very valuable paints and are widely used. Ochre is similar to umber and sienna and' is essentially an iron-oxide pigment. These vary in their chemical composition and phys- ical characteristics and are not often used alone in paints for iron. Carbon black is made by burning natural gas and is very pure carbon. It is very light, grinds in 84 per cent of oil, and has been used in conjunction with white lead in protective paints, but it now appears to be a stimula- tor and as such is not safe to be used as a contact coat, although it can be employed very well as an excluder, on top of an in- hibitive paint. Lampblack is made by burning oils and is also a very pure form of carbon. It has a specific gravity of 1.82, grinds in 75 per cent of oil, and has unusual tinting power, wherefore it is used in large quantities for this purpose. Other characteristics are its great stability, its very slow rate of drying, and a preserving action on the oil with which it is combined. It resembles the graphites in PAINT MATERUXS 99 its ability to conduct electricity; therefore it should not be used as a contact coat, but possesses perhaps unusual merit when used as a top coating. Chapter V PEOTECTIVE PAINTS TT^EOM the facts wMch have been presented ■*• regarding the manufacture of paint mate- rials and the tests to which they have been subjected, it should be possible to predict, in a general way at least, the behavior of a pro- tective paint, provided we know its composi- tion. It must be noted, however, that there are many things to be considered in designing or passing on the probable value of a paint from its formula or analysis, as the question is not altogether whether a good quality of linseed oil and stimulative or inhibitive pig- ments have been used, although these points are of great importance. In the first place, it appears that even though the inhibitive powers of the pigment are the main factor ia protecting a painted surface from rust, this action is enhanced greatly by the use of pigments which are good insulators, as far as the ability to con- duct a current of electricity is concerned, for 100 PROTECTTVE PAINTS 101 the base. Again, a study of paint films has shown that some are much better excluders or water shedders than others. It has been shown that linseed-oil films are markedly porous, with the result that water soon pene- trates to the surface below and, if it is of iron, sets up corrosion. Most dried paint films are porous, although this porosity may be les- sened or prevented by the use of varnish, thus producing a film which is a good ex- cluder. By the use of finely ground pigments the size and number of pores in a film may be very materially reduced, but even then the film is not necessarily a water shedder. Cer- tain pigments seem to convey a peculiar prop- erty to a film in that it becomes wet with difficulty or not at all. Water rapidly evapo- rates or runs off from it and apparently none is retained upon the surface or in the pores. Perhaps the action of different pigments in excluding water from paint films may best be shown by citing an experiment performed and described by Cushman and Gardner. A number of pigments were ground ia oil and films made therefrom. It was not pos- sible to make all films of exactly the same thickness, but there is no doubt that these results are roughly indicative of the ability 102 COEBOSION 01" IRON I 0^«DI>t"0'S<«ONei|t"«0»-<>Oi-IOSO^O'!ll ^t050i-ib»o»ooo50i'niHooos'OOg«pi5'2>o oooooooooooooooooooo •s CQ Ui CO I 00 I •a I ;s s Ot^OO^t^OiHIQCQOOOOOlOt^a^COOX Soo6»Soso5i-ii-H^>ocnt>oot-'«*'*e9»g«^ i-(iHTHci.-ii-((Me«cOeiot-r-iQOOcoejg5 Oi-H»-lr-i»-irHt-lT-ti-HrHi-li^i-tNWNWC'>!ii'«j(t-eo'^ lOOdOOOOOOOOOOOOO OOOOOOOOOOOOOOOOOOOO PROTECTIVE PAINTS 103 ja ■*_Ot>.0>t>.0J«O001>M000S _ dddddddddr-i rti-i CO •S 00 S THoeoi-i COTH'nsDlOtO'^lOtDOON^ ddddddddddT-iiH OS U3 00 a> o 00 o ^ O) CO ^ C3 t* Cp ^ •i ddddddddddi-ii-i >0 00 to 1-1 OS 1-1 Tjt ..- - tOOCOtOTtlTjl'* e»e ,-1 oooooooooooo .r f^ iH»oomt^«3i-40500'-iN » ff «0000i(NO^'-IM-*O'-ll> .j -a "H ■-< iH N N u3coa30 fe fe oooJOsoo'-ii-ii-iN'^asi-i •^ ^ OOOrHi-li-trHi-iTHi-HiHN .B ^ dddddddddddd ^ *a § 60 g i 1^ g< g f H -Sb m o ead and Zinc Chromate. hinese Blue (Stimulative enetian Red 1— <- >> i I 2 3" 1 a i 9 -9 -I S a a .a J3 ^ " ^.. III I Sis ■i^tg iz; 1^ S n p ii; P3 ^ n 104 COREOSION OP IRON of these materials to act as excluders. The method of making the test was as follows : A series of small glass bottles with wide mouths, holding about two ounces, were half filled with con- centrated sulphuric acid, and paint films were tightly sealed over the mouths of the bottles, using Canada balsam. The bottles were then carefully labeled, numbered, accurately weighed on chemical balances, finally exposed to air saturated with moisture and kept at constant temperature under a large glass re- ceptacle. The bottles were removed from the re- ceptacle every week and weighed. The increase in weight, due to the amoimt of moisture which had penetrated the films, and which had been taken up by the sulphuric acid, owing to its hygroscopic na- ture, was thus determined. It is interesting to note the effect which a varnish gum has in closing the pores of a vehicle, as shown in this test. Thus iron oxide used alone stands about half way down the list, but when it contained 2 per cent of gum in addition to the oil it stood very high. It has been shown that the addition of a high- grade Kauri gum varnish to a paint increases its protective power considerably, this in- crease, up to a certain point, corresponding to the amount of varnish added. Putting together all of the above facts and coming to the practical design and testing of protective paints, The American Society for PROTECTIVE PAINTS 105 Testing Materials has for a long time been carrying on very thorough studies along these lines. The results of these tests, which have been made under service conditions as far as possible, are available in the reports of the Society, but for the sake of the readers who do not have access to them or who do not have time to look them up and go through the many papers, they wiU be given here ia part, as the writer feels they will be of more value than any theoretical discussions could be. One of the first tests was on the Pennsyl- vania Railroad bridge over the Susquehanna Eiver at Havre de Grace, Maryland. Nine- teen different paints were applied to panels placed on the bridge and to the bridge itself, brushing coats of definite weights. Very careful data were kept of all the conditions of the test, analyses were made of the paints, and everything, including the inspections, was carried out with a great deal of thor- oughness. The final results are now avail- able, but it is considered inadvisable to con- sider them further here, and mention is made only because it is thought that some reader of this volume might wish to make a thor- ough study of such a test. 106 COEEOSION OF lEON m *g'MOOOOOeO"*l>5000t-T|lOO «•! ,^§. O t« OOOU3 C4 >0 M'^'OICO ^ .-. ^ 0 "3 M N CO O <0 '^ 00 00 CO N M N N N i-l i-t N N N iH iH »-( i-l iH i-l T-l i-ti-liHi-i I ft O H TO ■* 00 1~ W 00 «0 1-1 O 00 O r-l O e<3 (N N t-T 0» CO N 00 OS N "g gj 'l|eOoJt-^«Di-(00'*rtOo6eOINod "3 l-i- il.alllli'S^iill^yo^i i-ldCO^U3cOtvO>OC<)^>OCOt~03 PEOTECTIVE PAINTS 107 o-*qt>ujr-;Nqqou3U3weo t^ooooeoousoo N N m 00 00 >n Tjl ■* M CO -!i< CO M M ■* ■* O O 5 CO (N w U5 lO eocot^-ioi-iriit-i* ooa5i-imo"j CO CO oo t- ^ ct q t-(D0000OI>OOC0lOlOl>-^lI5 OOOO'Ot-COt-lOUS O O ■* eD«000000 o ^g.9.go»THt»tHOooioo>rHi»'*N!0'*>n«ooooeoi>ooos o ^"§2 §'-ioodddo A S l^_g ■M.2t^N»oqooceoeor^oei5ooo»cooo(Niot>oooqi«. ^■^^ gM->jtTHdo6eodoot>iodT-iddTH 9) 0> ffi s a d a'b ID a> & f^Pn R S 2 H ^ rA -§■3 a?3 s g-i '111 SOOtg n _H ^ ^4 .■a a a a .aa-a-i 3 S S « IHC^eo^u30l^"050c^^u3«ot~05QTHT^l^-ooo50'H iHT-tr-llHiHr-ir-INNNNCTNeOeiS PEOTEOTIVE PAINTS 113 i fe'^S S "-I '-'' o» "5 00 1- 1> ffl ui d 00 1> tfS °i & N 00 ION OS 00 ON t> g'_g-S.^|>ososoaOU3«DOONO(MO § fe ta 2 S <^ '-' o"0 00 00 00 1> i> o oJ 00 oooqt-(N N5oq 00 303r-lOt-«5lOOOO"3OI>m S3-S'-i Somot-cs 05 0503X00303 iooqiot«. eqb-eiio QOOOTfd o6t>l>Q6 fl 00 USCDt^OO IQOt-OO 114 OOEROSION OF IROST J c3 ^ Q Ifi ^ccisg. 00 CO 03 CO 00 >0 CO lfi.2 'ji CO ri ^«S| rH CO t» PEOTECTIVB PAINTS 115 but a few comments may not be out of place. First, it will be noticed tbat not all of the pigments have shown up as might have been expected from the results of the water sus- pension test, quoted in the preceding chapter. In other words, some of the materials which were considered to be inhibitive in their ac- tion showed up considerably inferior to some which were thought to be stimulators. There are probably several reasons why such a re- sult was obtained. It has been previously noted that a pigment in water suspension is under vastly different conditions from one in a paint film, that is, in oil. It will be recalled that in the preparation of the paints the same amount by volume of each pigment was ground in the unit quantity of oil. Pigments vary greatly in the amount of oil which is required to grind them into paste or paint form, and there can be little doubt that dif- ' ferent results would be obtained if various quantities of oil had been mixed with each pigment. Again, the fineness of the pigment makes a great deal of difference to the paint, as it determines the extent to which the pores and voids in the oil film will be filled up, thus producing a more impervious film. Then, too, 116 OOEKOSION OF lEON the freedom of the pigment from impurities can easily be seen to be of prime importance. In this test the materials were of commercial purity. In the case of paints made from a combina- tion of pigments, the deviation of the results obtained from those which might have been expected was considerable. In such a case we have to regard the action of several pig- ments in a film, instead of one, and at present such action is not always easily foretold. There is no doubt that the proper combina- tion of two or more pigments makes a better and more durable paint than one used alone. In the main this effect is mechanical or phys- ical, each component contributing certain properties toward the production of a film having the desired qualities. On the whole, it seems that these tests show what each of the pigments tested has done, and probably would do regularly, when made up in a certain vehicle and in a cer- tain way, and that valuable data have been obtained for use both in the practical design- ing of protective paints and in the carrying out of future experimental work. In working up a paint formula attention must be paid to the particular requirements PEOTECTIVE PAINTS 117 of each case, since no one combination will give good results in all situations. For in- stance, a paint which must stand exposure to light and heat is of different composition from one which would be used in a cool, damp place. It is not possible to consider this ques- tion further, but it is hoped that the data which have been given will be of at least some assistance along these lines. The method of protecting iron, particularly building and structural material, employed by the manufacturers of "Asbestos Protected Metal, ' ' is worthy of notice. Briefly, the proc- ess consists in cleaning and heating the parts, then immersing them in some asphaltic com- pound, also kept at a high temperature. Upon removal they are passed through rolls, to se- cure more perfect union of the steel and the coating compound and remove air bubbles, then through another set of rolls between two sheets of asbestos felt, made from a special, long-fibred material. Of course, the use of asphaltum in one way or another as a protective agent is not new and when properly prepared it may give very good results. Heating the iron before the ap- plication of the coating is an excellent plan, as all moisture is thereby removed from the 118 CORROSION OF IRON" surface and pores; the advantage in doing this is easily seen. The asbestos acts as a mechanical protection to keep the underlying coating from injury by blows or scratches and shields it from the destructive action of the sun and air. The final stage of the proc- ess is a treatment designed to make the as- bestos water-proof, so it would appear that parts protected in this manner should last almost indefinitely. Coming now to protective materials which are slightly different from the foregoing, so- called baking japans are used widely in fin- ishing a large variety of metal work requir- ing a decorative finish which is cheap and durable. In general, black japans are made by combining linseed oil and asphalt, gilson- ite, stearine pitch, coal-tar pitch, and similar substances at high temperatures. The re- sulting product is thinned with turpentine, benzine, or light mineral oils, and when ap- plied by spraying or dipping and baked at a temperature ranging from 200 degrees F. to 450 degrees F. or so, will give a more or less thick coating which is quite hard, fairly im- pervious, and of a high lustre. The baking period, which varies from two to six hours or longer, and the temperature of baking, de- PROTECTIVE PAINTS 119 pend on the composition of the japan. Dull and medium gloss japans are made by grind- ing the bright japan with lampblack or car- bon black in a fine mill. The method of ap- plication and baking is the same as given. Colored japans or enamels are essentially pigments ground in a boiled linseed oil which may or may not contaia a percentage of var- nish gums. They are generally baked at con- siderably lower temperatures than the black japans. The film given by these baking japans is much harder than an ordinary paint film and will stand rougher handling and usage, a cir- cumstance which gives them a marked advan- tage in many cases. "While they are hardly intended for outdoor exposure and perhaps are not as durable and resistant under such conditions as the best oil paints, they give, nevertheless, very good protection even un- der severe service. Two or three coats may be applied in much less time than it takes for one coat of paint to become thoroughly dry, so this is an added advantage of some im- portance. Altogether, baking japans furnish a means of producing quickly and cheaply a coating which has good protective and decorative qualities. 120 COEROSION OF IRON An interesting application of this sort of process is in the coating of large pipe for water mains and so on. Many years ago Dr. Eobert Angus Smith, an English scientist, patented a method of treating such objects by dipping them in a mixture of coal tar, from which most of the volatile constituents had been distilled, and linseed oil, kept at a high temperature. The part was thor- oughly cleaned and coated with linseed oil, which was baked on, then it was coated with the coal-tar mixture and the residual heat of the pipe served to bake this on. In pass- ing, it is worthy of comment that not very much improvement has been made in the matter of coating cast-iron pipes since the time of Dr. Smith. The application of a coal-tar coating does not give very good re- sults on sheet-steel riveted pipe, however, and a more or less soft asphaltum is now used. A material known as maltha and obtained from California petroleum as a residue which remains in the retort after the more volatile portions have distilled off, is used to a large extent for this purpose. The dipping ma- terial is thinned with high boiling mineral oils and is heated to a high temperature for application, but the pipe is not baked after- PEOTECTIVE PAINTS 131 ward, the coating being allowed to cool in the air. Sabin has improved this process by the use of asphaltmn and linseed oil, whereby it is claimed to be possible to obtain a coating of any desired degree of elasticity by adding the requisite amount of the latter and baking properly. The use of volatile solvents is avoided by heating the compound to 300 de- grees F. for application and a thin uniform coating results. After dipping the pipe is allowed to drain for a time, until the excess has run off, which ordinarily takes one half- hour or less. It is then put in the oven, which is heated somewhat above 300 degrees F., and allowed to remain two hours or so, or until the volatile substances have been driven off and the oil oxidized. A tough, hard coat- ing is produced and the Sabin process, as it is called, is finding a Avide use. It is inter- esting to note that the steel underfloors of the roadways and sidewalks of the Williamsburg bridge over the East River in New York City were pickled and coated in this manner, with very satisfactory results. A coating for steel and wrought-iron pipe, thoroughly protected by some sort of heavy fabric, is much superior to one produced by 133 COEROSIOK OF IRON simple immersion in a bath of the hot bitum- inous compound, and the "National Coat- ing," produced by the manufacturers of Na- tional pipe, may be mentioned as a sample of this type. In applying this the pipe is cleaned and dried, then dipped at a temperature above the boiling point of water into a hot bath of a special refined bituminous compound, after which it is removed and allowed to drain in a vertical position and cool. The next operation consists in placing it in a winding machine where a strip of fabric, saturated in the hot compound, is wound spirally over the surface so as to overlap an inch or so on each turn; if necessary two or three layers of fabric may be applied. The final result is a hard, but tough and elastic, coating of considerable thickness which is strengthened and protected by the heavy fabric embedded in it. Naturally, gas and water pipes are liable to undergo rather strenuous treatment before they are finally laid to rest in the ground, and an ordinary coating may be badly scraped and damaged, thus greatly reducing its use- fulness. A coating such as described above, however, is very resistant to mechanical in- PROTECTIVE PAINTS 133 jury and should remain in good condition al- most indefinitely. There is still another class of protective materials, known as lacquers, which may be considered briefly here, although they afford relatively very slight protection and are not used to any great extent on iron. In general there are three kiads of lacquers — gum or resin, pyroxylin, and those made by combin- ing these two. The first kind is made by dissolving shellac, kauri, damar, sandarac, and so on, in alcohol, fusel oil, or other solvents. Eesin lacquers are applied by brushing, as a rule. A pyroxylin lacquer is essentially a solu- tion of cellulose nitrate in amyl acetate. The nitrate is generally made by treating cotton with a mixture of nitric and sulphuric acids, whereby a reaction takes place in which the cellulose unites with two or three molecules of a nitrogen compound, depending on the strength of the acid and the temperature and time of treatment. The cotton is then washed and all acid neutralized, and after drying is dissolved in a suitable solvent. Amyl acetate is perhaps the best, but it is very expensive, so is often adulterated with wood alcohol, acetone, benzine, or other substances. 134 OOEEOSION OF IRON The finest lacquers of this class are made from nitrated tissue paper, since this gives a whiter and better appearing product than cotton. This "water white" lacquer is used almost exclusively on silverware. The combination lacquers are made by mix- ing solutions of pyroxylin and resins, thereby producing a material which has a far wider range of usefulness than either of the pre- ceding. By the proper blending and mixing the most desirable qualities of both of the constituents may be brought out — ^hardness and covering power may be increased with- out increasing the flow, and so on. Probably the combinations of pyroxylin and shellac are the best and most widely usedj Any of the above lacquers may be colored by grinding pigments in them or by adding suitable dyes. They are mostly used, of course, for decorative work. All of these ma- terials are applied by dipping, brushing, or spraying, depending on the nature of the work and the qualities of the lacquer, and are allowed to dry in the air or subjected to gentle heat to hasten the process and assist in the production of a harder coating. As noted before, lacquers have a rather limited field of usefulness on iron work, but they are PKOTEOTIVE PAINTS 135 used very extensively on brass and copper articles, fixtures and tlie like. In conclusion, there is one important item wMch has been mentioned only incidentally before, but which is deserving of thorough treatment and consideration, and that is the proper cleaning of the surfaces of iron and steel parts before paint or other protective materials are applied. Paint which is put on a surface covered with grease or dirt does not come into good contact with the metal and consequently does not adhere to it. Sooner or later the dirt will fall or be knocked off and the paint will come with it. Beams and other parts which are coated with heavy oil or which have been allowed to lie on the ground until they are covered with mud and dirt, should be thoroughly cleaned before painting if it is desired that protection of any permanence shall be secured. Mill- scale is another substance which is almost always found on iron work, especially struc- tural material. It is hard and perhaps has some protective power when firmly attached, but all loose scale should be removed as com- pletely as possible by hammer and chisel or sand-blasting, as it tends to crack off slowly, taking the paint with it and exposing the bare 126 ooEEOsiosr of iron iron. It is a matter of common experience that when rust once gets started it is very- difficult to stop it, so progressive rusting may take place under a paint film which is itself intact and in good condition. It is important, therefore, to see that all surfaces are free from active rust, loose miU-scale, dirt, and other foreign substances, and that they are perfectly dry before painting. In addition, warming the parts slightly, whenever this is possible, just before the paint is applied, also tends to promote better adherence and more satisfactory results all aroimd. It must be remembered, too, that these remarks apply not only to the first, but to all subsequent paintings. Chapter VI INPLUBFCE OF DIFFEEENT ELEMENTS ON THE COEEOSION OF lEON T T has probably occurred to everyone who -■■ is forced to rely on protective coatings of one sort or another to keep iron and steel from rust and decay, that this method of com- bating the corrosion problem is of the same order as locking the bam door after the horse is stolen. In other words, since rusting is a consequence of the tendency of iron to go into solution when in contact with water, why not add something to it or treat it in some way so as to eliminate or decrease this ten- dency? An enormous amount of labor and material, to say nothing of time,' is used up every year in coating and protecting iron from adverse conditions, and with the rapid increase in the demand for this metal a cor- respondingly greater amount of labor, ma- terial, and time is required. Obviously, any procedure which would lessen the need for such thorough protection would be a very real saving. 127 138 COEEOSION OP IRON Definite work along this line is of com- paratively recent date, and the available in- formation is perhaps not as complete as could be wished; but enough has been done to throw some light on the possibilities of ultimate success and to enable us to use to good advantage some of the facts which have been obtained. In effect, of course, the problem resolves itself into the study of the influence of the various elements (meaning the different metals and metalloids) on the process of cor- rosion. Chemically pure iron is little more than a laboratory curiosity; certainly it is not available for commercial use in building bridges and sky-scrapers, and experiments have shown that, on the whole, it is very little or no more resistant to corrosion than well made wrought iron or steel ; so the answer to the problem does not lie in this direction, al- though this should not be taken to mean that any pains are to be spared La making iron and steel of high purity and as free from segrega- tion as possible. Perhaps a better idea of the difficulties presented in accomplishing these ends may be obtained by reviewing briefly the early stages of the manufacture of iron. ELEMENTS INPLUENOING COEEOSION 139 This metal is found ra nature principally in the form of the oxide, and by far the larg- est part of the ore is smelted in blast fur- naces, which are a sort of immense stove, 75 or 100 feet ia height. The ore, mixed with coke (more rarely with charcoal or coke and anthracite coal) and enough flux (generally limestone) to remove some of the impurities in the form of slag or ciuder, is charged in at the top. Air heated to a temperature of sev- eral hundred degrees Centigrade is forced at a pressure of about 15 pounds per square inch into the furnace near the bottom, through pipes known as tuyeres. This blast is further heated by the combustion of the fuel and passes up through the charge, melting and reducing the ore. The metallic iron formed trickles down and collects on the bottom of the hearth, with the cinder floating on top of it ; the furnace is tapped at intervals of sev- eral hours, and the iron cast into pigs or sent to the steel miU where it is purified as will be described in the next chapter. The cinder is allowed to escape at longer intervals. The iron produced in this manner is a com- paratively impure material, since it contains very considerable amounts of carbon and sili- con, especially, and less of manganese, sul- 130 COKEOSION OF IRON phur and phosphorus, with varying quantities of several other elements, depending on the character and quality of the ore. Altogether, these impurities will amount to 6 or 7 per cent. Pig iron is melted and cast into various forms, constituting the gray or cast iron of the foundry, but it can not be rolled or other- wise worked without further treatment. Omitting for the present the details of this subsequent treatment, the product is known as wrought iron or steel, which latter is a rather indefinite term. In these, however, the percentage of other elements will gener- ally be well under 1 per cent, except where one or more of them is added purposely, as in making alloys, or for conferring certain desirable characteristics. It is impossible to remove all of the sulphur, phosphorus, and so on, and if the attempt is made to do so, the iron itself undergoes considerable oxida- tion, the oxide present mingling with the mol- ten metal and introducing a new and perhaps greater danger. All the ordinary impurities in iron, with the exception of manganese, are electro-nega- tive to it, which means that a current will flow from one to the other in the presence of an electrolyte. Under such circumstances the EtBMBNTS INFLUENCING OOEEOSION 131 solution of the iron will be hastened, as ex- plaiued in a preceding chapter, and it will corrode much more quickly. It is in a way reasonable, therefore, for the idea to have arisen that the purer steel and iron can be made, the less subject to corrosion will they be, thus rather tacitly blaming the tendency to rust on the inherent impurities of the iron. Without stopping to discuss the soundness of this view, the fact remains that it is not prac- ticable to refine iron beyond a certaiu point ; hence experiments have been tried to ascer- taiu the effect of adding other elements be- sides those usually found in iron, in the hope that in some way the end desired might be accomplished by this means. Taking up first the normal constituents of iron and steel, it appears from the data at hand that silicon rather tends to increase cor- rosion, although this is contrary to what might be expected as it is weU known that high sUicon irons are markedly resistant to acid attack. It is generally agreed, however, that the resistance to acid is not always a re- liable measure of corrodibility, since many discrepancies are observed and perhaps this is one of them. The silicon content of ordi- nary steel is not very high, so this element. 133 COEEOSION OP IRON when present in normal amounts, probably exerts little effect as far as corrosion is con- cerned, but even if it did have an inhibitive action it could hardly be used in any quantity as it renders steel weak and brittle. Sulphur is harmful in its effects, materially increasing the tendency to corrode; even fairly small amounts exert an appreciable in- fluence in this direction. Manganese is popularly supposed to be the cause of much trouble, and the supposedly greater resistance of the older irons over modem steel has been largely ascribed to their low content of thi& metal. Recent work, however, shows that it does not play as im- portant a part as has been thought. In the tests conducted by Burgess and Aston the results were not altogether consistent, but in general it appeared that atmospheric cor- rosion increased with higher manganese con- tent, although in all cases this was less than that observed for pure electrolytic iron. On the whole, probably the influence of manga- nese, unless it is present in amounts greatly in excess of what would ordinarily be the case, may be considered as practically negli- gible; at any rate its action is not strongly marked one way or the other. ELEMENTS INFLUENCING COEEOSION 133 Phosphorus seems to exert a beneficial ef- fect, rather than otherwise, when present in reasonable quantities. Of course, the phos- phorus content of a steel affects the physical properties greatly and hence the permissible amount is low; specifications are usually very strict on this point. Carbon is another element which deter- mines to a large extent the characteristics of iron and steel, and the amount allowable is rather limited ; but from some investigations it appears that the corrodibility of annealed steel rises with carbon content up to about .89 per cent., then decreases. In quenched and tempered steels no decrease in corrodi- bility could be noted up to .96 per cent, car- bon. Aside from this the effect of varying amounts of some of the other normal impuri- ties of iron has apparently not been studied in detail, but it is rather doubtful if results of any particular value would be obtained. These metals and metalloids are always pres- ent to some extent in iron and steel; to re- move them entirely is a very difficult if not impracticable thing commercially, and in any case the weight of evidence goes to show that pure iron corrodes, in the long run, at a rate little or no different from an ordinary, com- 134 COEROSION OF lEON mercial grade of material, assuming that this is well made and reasonably free from segre- gation. The point is, that none of these ele- ments, with the possible exception of sulphur, appears to exert any particular influence, either good or bad, on the process of corro- sion when they are present in the amounts permissible from the standpoint of the physi- cal properties which they confer. There seems to be no reason to hope, therefore, that any benefits will accrue either from removing them completely or increasing the amount of any or all of them over that usually met with. Some time ago Burgess and Aston carried out experiments on the alloys produced by adding various quantities of different metals to electrolytic iron — ^that is, iron which had been deposited from a solution of a salt of this metal by means of the electric current and therefore was of great purity. Care was taken to exclude impurities as far as possible, in making up the samples, and after the al- loys had been brought into proper shape for testing, small pieces were exposed to the weather for something less than six months, then cleaned and the loss in weight noted, and an accelerated acid corrosion test applied. Before considering the results it may be ELEMENTS INELUENOIHTG COREOSION 135 noted that these tests were not made on a commercial grade of iron, the electrolytic ma- terial being practically free from carbon, sili- con, manganese, sulphur and phosphorus and other impurities so that the effect of these was lost, but while this may be objectionable in some ways there is little doubt that the results indicate, in a general way at least, the influence of the metals tested. In trying out various grades of commercial irons of different degrees of purity, it was found that some corroded more, some less, than the electrolytic, the samples of the latter representing an average product, as far as composition is concerned. In its ability to withstand acid attack or atmospheric corro- sion it did not stand out at all prominently, and these results are in general accord with those obtained by other investigators; that is, on very pure iron rust may be somewhat slower in starting than with the ordinary va- rieties, but the difference in behavior soon be- comes negligible and in the course of an ex- posure test covering several months it ap- pears that there is no decided superiority one way or the other. A sample of pure iron containing 1.33 per cent of aluminum showed a slightly decreased 136 COEKOSION 01" IRON tendency to corrode in the air, but was more easily attacked by acid. Neither it nor an- other sample containing a much smaller amount of aluminum gave results of any par- ticular interest. Arsenic confers no especial resisting power against atmospheric corrosion, even when used in amounts up to 3.56 per cent. It is well known that this metal is an objection- able impurity in sulphuric-acid pickling solu- tions as it slows down the action consider- ably, and it has been thought that soikething of this sort might be exhibited when it is al- loyed with iron, but apparently the reverse is true since the rate of solution in acid be- comes higher with larger amounts oi it pres- ent. In any case iron-arsenic alloys possess un- desirable qualities, such as weakness and brit- tleness, which would render them unsuited for structural purposes, aside from the fact that they are rather difficult to make. In like manner, the alloys with cobalt do not present any unusual characteristics. The resistance to acid attack increases wiilh higher cobalt content, and this is true of thb behav- ior of the specimens exposed to the weather. As in the case of arsenic, considerable per- ELEMENTS INFLUBNCIHTG CORROSION 137 eentages of cobalt would be required to pro- duce an alloy having sufficient resistance to corrosion to be of much value. The com- paratively high cost of the metal is one ob- jection to the use of such alloys and, further, the addition of such large quantities of it changes the properties of iron and steel more or less profoundly and takes us into the realm of special steels. This is also true of several other metals besides cobalt, The addition of nickel gives results some- what similar to those obtained with cobalt; perhaps this might be expected in view of the close chemical relationship between the two metals. Nickel, however, confers very fair resistance to both acid and atmospheric at- tack and this corresponds in a general way to the amount of this metal present. As in previous cases, however, a fairly large amount of nickel is needed (1 per cent or over) to produce very decided results from the standpoint of corrosion, and this in- creases the cost while altering the physical properties as well; up to about 3.5 per cent it gives greater strength, raises the elastic limit and promotes soundness. Iron-nickel alloys have properties which have been found of value for certain purposes, but it is rather 138 C0EK09I0N OP lEON doubtful if they come into wide use for gen- eral work solely on the score of decreased corrodibiKty. Experiments with silver, tin, tungsten and selenium indicate that these elements when alloyed with iron do not alter very much its tendency to corrode. In the case of silver and selenium the alloys contained only small amounts of these metals and it would be un- wise to attempt to draw very definite con- clusions as to their action. In any event it is not likely that either would be used very widely, on account of the cost. With both tin and tungsten there is a tend- ency toward increased resistance to atmos- pheric corrosion as the content of these met- als in their respective alloys becomes greater, but it is nothing remarkable at any time ; the alloys with low tin content display very good resistance to acid attack, but this decreases appreciably with the higher alloys. The cor- responding figures for tungsten show up well, but they are irregular and lack consistency. As in the other instances, these metals, to be of much effect, would have to be added in amounts which are sufficient to change the properties of the iron markedly; ia general the iron-tin alloys are brittle and lack ELEMENTS INFLUENCING COEEOSION 139 strength, while those with tungsten are noted for their great hardness and are used under the name of high-speed tool steels. Obviously they would be unfit for ordinary purposes, aside from cost considerations. Copper may be regarded as a normal con- stituent of iron and steel, although it is some- times present only in the barest traces, but the effects which even small additions of it produce are remarkable and a considerable amount of very careful work has been done in investigating the possibilities which it pre- sents along the line of reducing the tendency of iron to corrode at the least opportunity. The results are so interesting that they are worth considering in some detail. As long ago as 1901 two experimenters, Stead and Wigham, made some tests on cop- per- and non-copper-bearing steels and came to the conclusion that, under the conditions of their tests, at least, the steels with copper showed up to better advantage. Within the last few years, however, other experimenters have taken up this subject, and of these Buck has probably done more than anyone else and his work merits more than passing atten- tion. In a paper read before the annual meeting 140 COEEOSION OF IRON of the American Chemical Society at Mil- waukee, March 25, 1913, he gave the follow- ing information regarding the preparation of the samples and the method of conducting the tests. In order to avoid the possible uncertainty in com- paring different heats of steel with and without cop- per, and in order that the conditions, except the cop- per content, should be identical, it was decided for these comparisons to copperize portions of heats, leav- ing other portions of the same heats in their original conditions. Three heats were used. One was a regular basic open-hearth of the following analysis: Carbon Manganese Sulphur Phosphorus .10 .34 .034 .019 A second basic open-hearth heat was re-phosphor- ized, giving this analysis : Carbon Manganese Sulphur Phosphorus .13 .45 .036 .042 The third heat was regular Bessemer steel of the following analysis: Carbon Manganese Sulphur Phosphorus .08 .46 .070 .096 In pouring the open-hearth heats several ingots were first poured without the introduction of copper, then to four ingots sufficient copper was added to ob- tain in two of them about .15 per cent, and the ELEMENTS INELUENOING COEEOSION 141 other two about .25 per cent copper in the finished product. The Bessemer heat was treated in exactly the same way, except that, since the average Bessemer heat is too small to furnish six ingots of the size desired, only two ingots were copperized, aiming at the same contents as in the case of the open-hearth. The copper was added to the moulds a little at a time as they were filling, and that the resultant steel was uniform in its copper content, was demonstrated by many analyses of the bars and of the finished sheets. Indeed, that copper easily diffuses through the bath of molten steel, and does not segregate on cooling, is a well established fact'. Six ingots were then taken from each of the open- hearth heats, two plain, two with .15 per cent copper and two with .25 per cent copper, and three ingots from the Bessemer heat, one plain, and one with each content of copper. The fifteen ingots thus prepared were carried through the usual mill operations, each bar as cut and each sheet as rolled being chalk-marked so that no confusion could possibly occur, and in the end each lot was again carefully analyzed as a double check on the operations. One ingot of each grade of open-hearth was rolled into 16-gauge and the other into 27-gauge sheets, 30 by 96 inches ; while in the case of the Bessemer steel, one-half of each ingot was rolled into 16- and the other half into 27-gauge. All grades were subjected to exactly the same treatment, being rolled by the same crews, and annealed in the same furnace at the same time, and the finish was such as to conform with that of the competitive sheets used in this test. From 24 to 36 sheets of each of the 9 grades, both gauges, making 18 lots in all, were then sheared to 24 by 96 inches, thus obtaining a strip 6 inches wide 14a OOEEOSION OF IKON from each sheet. These strips were sheared into 2 by 4 inch test pieces, stenciled with distinguishing marks, and were used for corrosion tests which will be described later. The 34 by 96 inch sheets were corrugated in the usual way, and eight to twelve sheets of each grade shipped to each of three testing stations. One of these is located in the Pennsyl- vania coke regions, where the air contains notable amounts of sulphurous and sulphuric acids and other fumes from the coke ovens. In this district iron and steel, unless protected, corrode very fast. Another station is located on the sea coast, where the air car- ries sodium chloride. The third is in a rural com- mxmity, where the air is quite pure and free from added corrosive agents. A sloping wooden framework was erected at each place and the sheets fastened thereon, with the precautions to see that the drip from one sheet did not fall on the next one below. They were thus held several feet from the ground, so that there was a free cir- culation of air on aU sides, and exposed to the weather without a protective coating of any sort except the thin film of oxide nor- mally present on the surface of black sheets. There were likewise exposed at the same time and under the same conditions sheets of both 16- and 27-gauge, purchased on the open market and showing the following average analysis ; ELEMENTS INFLUENCING COEEOSION 143 Carbon Manganese Sulphur Phosphorus Copper .03 .03 .034 .003 .06-.07 All the sheets were placed in position dur- ing November, 1911, and were frequently in- spected; by means of the small test pieces cut from the large sheets and exposed at the same time, it was possible to determine the loss of weight of each grade. First, considering briefly the results of the examinations of the large sheets after vary- ing periods of exposure, at the end of seven months in the coke regions the Bessemer steel sample without copper had failed entirely, while those from the same heat but with cop- per additions were in very good condition. The same thing was true of the basic open- hearth steels, the copper-bearing samples showing up well, while those without it had failed. One from a re-phosphorized heat, but containing no copper, was in somewhat better condition than the corresponding panels from the other heats, although it was showing the effects of corrosion, while the copper-bearing panels were in good shape. It was noted that the color of the oxide on the non-copper steels was bright red and loosely adherent. The copper steels were 144 CORROSION OF IROKT CO S 03 O to V'Oi (NNOSO r^t-Nto os©'*'-l tOQOOOO CO CO ^ ^ ^ ^ ^ w* coTiteot- io«3cD>-i t~ooeoio t22S S2SS (NCOMi-l MNi-lN c<0iOiO SSto 6:S2S2 •*eooo nmcoos ooiooto OOOi-l OOOO T-irHi-1,-1 OOt-hO oo>-io t^t*«Dt* t^t^t^o C05050CO t* coooo"* T)iio>no oooo oooo OS »-* CO CO 8233 oo-*iooi »H 05 O 1-* oooo NOOCOO OOOO OOOSOOO f— t i-M O OS OO^O U5(N00>O too CO CO OOOO lOOOsCO cor- to CO OOOO IOCS 00 lO CO CO CO CO OOOO CO CO CO CO oooo CO-* top CO CO r^ r^ oooo ^S miot^o OOOrt 8 ogt-o OOi-H s CO OS CO OOi-l 8 o>noo i-iOO 1 MM 8 sKn i ^ CI a PQMOO fe ^ s s O O P<0< Km a a s s °° °^ oom« 146 CORROSION OF IRON m w [» O " S § IOOCOt~QO IOCOt-IW i-HCOrHN CONr-liH C^l C4 O O NOOO SSS8 ooosoi-i lOooMN eoosmeo i-lOSNO 0»>-l^^ TH-r^OiQ OOOW oooo oooo §SSS co>0"3N os«3"3io ooeoont Mr-ccoo tocoMco mcot?!? OOOO oooo oooo 00CS>Ol> CO CO CO CO oooo rococo CO t* cot^ oooo lO^TitTii cceoOTfi t^mioco co-*co-* TfcOTjc* ■*eo->*o ^S^^ Tjl«OlOU5 COtH CO ^ (oujouj footo t-oooeo 00*-*0 OOrHrH O »-t O O ^ C3 CO Oi i-lO.-(0 OOOCOIO iHOOO t^I^COt* 5Dfr*t^CD t^iDCDtD COt^CDt* NOi-IN rHOC^^i-l C4<-It-IiH i-IC^iHO 5 e % a ^ S i S i OPQOm EtEMBNTS INFLUENCING COREOSION 147 dark brown and the oxide adhered tena- ciously. An inspection several months later dis- closed that the steels without copper had failed completely and fallen to the ground, while the copper steels were still in fair shape. The low-carbon and manganese ma- terial, American ingot iron, was in rather bad shape and after about eighteen months of exposure had disappeared entirely al- though all of the copper-bearing sheets were still in place. While the samples at the sea-shore exhib- ited relatively less rapid corrosion than those in the coke regions, the same differences in the behavior of the copper and non-copper steels were shown here as in the other case. In something less than two years all the nor- mal steel samples had failed completely, which was also true of the American ingot iron, with the exception of two sheets, while the copper-bearing steel panels were still in- tact. At the test station located in the country district, too, the copper steels showed up to approximately the same advantage, over the normal materials, as at the other stations. Space does not, however, permit of a detailed 148 CORROSION OF IRON discussion of the results of the inspections and in any case a clearer idea of the showings of the various materials under test may per- haps be obtained from a consideration of the results obtained on the small test pieces. It may be noted that each result in the preced- ing table is the average of six pieces ; as men- tioned previously these were 2 by 4 inches in size. These results need no comment or explana- tion, but it is apparent that some progress in the right direction has been made. Whether it will be possible sometime so to alter the characteristics of iron, by the addition of some other metal or combination of metals, that it will not rust at all or only with diffi- culty, and at the same time not change its mechanical properties objectionably, seems to be doubtful, although further advances along this line will probably be made as time goes on. Just now, the best available material of this sort is copper-bearing steel, which is finding a wide and rapidly growing use ; any- one can make it and the present output is very large. Chapter VII THE COREOSION OF WEOUGHT lEON AND STEEL PIPE T T has been shown in the preceding chapter ^ that iron and steel may be made less sus- ceptible to corrosion by the addition of cer- tain elements, but it is a matter of common experience that even without these additions some parts are more resistant to such attack than others, and brief mention has already been made of the fact that the iron manu- factured today is not quite the same, in some respects, as that produced formerly. The substitution of modem methods and machine processes for the old, hand way of refining these materials has resulted in a different product, both chemically and physically, and coincident with these changes in method and product there has been a tremendous increase in the use of iron and steel, and perhaps an equally great alteration of the conditions of service ; so it is not strange that a good deal of difference of opinion, doubtless honest 149 150 OOEEOSION OF IRON" enough at bottom, has arisen regarding the relative merits of the old and new varieties, especially their resistance to corrosion. The controversy over wrought-iron and steel pipe, for instance, has been very animated. Perhaps one reason for this is that nearly everyone uses pipe, or comes in contact with it in some way; then the conditions to which a steam or water pipe are exposed are rather imusual and just the opposite of what gen- erally obtains. On the whole, ordinary iron work is subjected to a large excess of air containing varying amounts of moisture and gases; only at long intervals, as from rain and snow, does the surface become wet. The inside of a pipe line, on the other hand, is in contact with a large excess of water, which may contain much or little oxygen or air. In a line where the water is frequently drained out of the pipes, or where it holds a large amount of air, it will be seen that the conditions for rapid and severe corrosion are present and it is not surprising that a line, or section of it, may go to pieces in a comparatively short time. When such a case happens it is very easy, and perhaps natural, to ascribe it to inferior materials or some- thing of that sort, and unless trouble is taken IRON AND STEEL PIPE 151 to ascertain all the facts in the case very er- roneous notions may, and often do, become current. Such a one is the belief held by some that steel pipe is considerably inferior, in point of corrodibility, to that made from wrought iron. There is perhaps no doubt that the steel produced ten or fifteen years ago was not the equal in all respects of that turned out to- day, but by reason of the improvements which are being introduced from time to time the quality has risen until it may safely be said that there is no ground for comparing it unfavorably with wrought iron, at least without a fair trial. It may be remarked that this subject is by no means new, and there is available a good deal of literature concerning it, but it is con- sidered to be worth while to review some of this briefly and, without in the least desiring controversy with any whose opinions or in- terests lead them to disagree with the con- clusions stated therein, to present the results of various investigations in the hope that this may help to clear up any false impres- sions. Much money is lost every year through ignorance and prejudice, and at least some of this occurs in buying ferrous metals, 152 CORROSION OF IRON since certain varieties of these are claimed to be superior to others and a higher price is charged for them. If the facts show that this extra premimn is warranted, well and good; if not, it is a sheer waste of money. Before considering the data bearing on this point, however, it may be well to take up briefly the processes employed in the manufacture of these two materials so that a more definite idea of what each is may be obtained. The way in which iron ore is reduced in the blast furnace, with the production of pig iron containing several per cent of other met- als and metalloids, has been previously de- scribed. As noted, this does not permit of rolling or similar treatment; therefore, it must be refined before it is suitable for use other than in making castings. Iron may be refined in several ways, but for our purpose only the puddling furnace and Bessemer con- verter need be considered, as the terms wrought iron and steel, as used in the manu- facture of pipe, refer to the product of these. In puddling, the iron is melted on a bed of iron ore which gives up part of its oxygen to oxidize the carbon, silicon and manganese and, to a less extent, the sulphur and phos- IRON ANO STEEL PIPE 153 phorus. The greater the purity of iron, the higher the temperature at which it melts; therefore, the metal in the furnace becomes pasty after a time, as the impurities are burned out, and is worked into a ball and re- moved. This is rolled into bars of convenient size which are cut and piled up, then reheated to a welding temperature and rolled again into a bar of the desired dimensions. Wrought iron produced in this manner is very pure, the greatest and most objection- able part of the impurities being the slag which is scattered all through the metal and gives it the apparently fibrous structure so characteristic of it after rolling. These fine slag lines are easily detected and form a ready means of distinguishing between wrought iron and that produced by other methods ; it is only necessary to smooth down a small spot with a file when they will show up, best under a magnifying glass. It can be seen that the puddling process is rather slow and expensive, since only a lim- ited amount of metal, a few hundred pounds at most, can be treated at a time, while some- thing like an hour and a half is required for a charge and considerable skill and care are 154 COEEOSION OF lEON needed in order to obtain a good product. The presence of 2 per cent or so of slag dis- tributed irregularly throughout the mass can hardly tend to increase the homogeneity, and it is generally accepted that lack of this is of more importance, from the standpoint of cor- rosion, than the presence of a reasonable amount of impurities thoroughly and uni- formly mixed in. Further, it is very hard to weld perfectly sheets or bars several inches in width, so laminations are met with and may give serious trouble. In the Bessemer process the molten iron from the blast furnace is run into huge mix- ers, then into big converters, shaped some- thing like an egg, wherein cold air is blown through it, entering by means of tubes in the bottom. The oxygen rapidly bums out the impurities, thereby generating a large amount of heat, so no difficulty is experienced in keeping the metal melted. The length of time the air is blown through is determined by the color of the flame issuing from the mouth of the converter. As might be ex- pected, more or less of the iron is oxidized at the same time ; but the addition of a small amount of manganese, for example, will take care of these oxides, so the requisite quan- lEON AND STEEL PIPE 155 tity of this, calculated from an analysis of the iron, is added, together with such other elements as may be needed. It requires only ten or fifteen minutes to convert as many tons of iron into steel, so in point of speed and general economy the Bes- semer process is superior to the other; it is especially adapted to making low-carbon steel and, as a matter of fact, gives a prod- uct which is more nearly pure, considered on the score of total impurities, than that from the puddling furnace. As a general proposi- tion, the silicon, carbon, sulphur, manganese, and phosphorus may be lower in the wrought iron than in steel, but the former will contain about two per cent of oxides while in steel these will not run over one- or two-tenths of one per cent. From the nature of the proc- ess whatever impurities are present are well distributed, and as the metal is rolled directly into skelp, without cutting and re-welding as in the case of wrought iron, the danger from laminations is greatly reduced. Up to about twenty-eight years ago hand- puddled iron was used entirely for making pipe ; for various reasons steel had not been successfully welded, and as the quantities needed were comparatively small there was 156 COEROSION OF lEOK no trouble in supplying the demand. Witli an increasing use of pipe, however, the discov- ery that steel could be welded when treated properly, and the discovery of cheaper and better processes of making this material, caused it to be employed more and more and two classes of pipe appeared on the market, iron and steel. With many manufacturers in the field different grades of these appeared, and now there are a number of such; in fact, there is as much difference between some of the wrought irons as there is between wrought iron and the steels, and there is a corresponding disparity between the latter. All of this has tended to increase the confu- sion; and inasmuch as any new material is liable to be looked upon with suspicion, it is perhaps only natural that there should have arisen much lack of agreement regarding the length of life which iron pipe, as contrasted with that made from steel, might be expected to give. As a consequence a good deal of experi- mental work has been done and several in- vestigations of the conditions found in actual service carried out to find the answer to some of these questions. An experiment condupted somewhat over IRON AND STEEL PIPE 157 twelve years ago by the Bureau of Steam En- gineering of the Navy Department on lap- welded Bessemer steel, lap-welded iron, seam- less cold-drawn steel and seamless hot-drawn steel boiler tubes, may be of interest. Care was taken to secure good, representa- tive samples of these materials and analysis showed them to be of the following average composition : ANALYSIS Cold- drawn seamless Hot- drawn seamless Bessemer steel Iron Silicon .005 .043 .015 .50 .16 .23 .005 .039 .016 .49 .14 .30 .008 .081 .110 .31 .063 .28 .024 Sulphur .014 Phosphorus .038 Trace Carbon Trace Oxide 1.03 Without going too much into detail, the test consisted in weighing the parts, then iiri- mersing them in distilled water, through which air was bubbled, for a total of 64 weeks, divided into four periods of sixteen weeks each. At the end of each period the sam- ples were removed, thoroughly cleaned and weighed. The results may be averaged thus : 158 COEKOSION OF IRON ATEBAOB I.OS8 IN QBAMS PEB BQUABB INCH After 16 Wis. After 32 Wka. After 48 Wks. After 64 Wks. Total Av. Lose Compar- ing Iron as 100 Hot-d'n O. H. Steel .3034 .5333 .4950 .5564 .5896 .3510 .4043 .4502 .3966 .5070 .4945 .5017 .4893 1.6947 1.7085 1.8315 1.8081 93.7 .3147 94.6 Cold-d'n O. H. Steel .3232 101.3 .3326 100. It will be seen by comparing the figures of loss given above that there was no great dii¥erence in the behavior of these four classes of materials, as far as loss in weight is concerned; the charcoal-iron samples, which showed the greatest initial loss, ending up with a decrease about 6.3 per cent larger than that of the open-hearth hot-drawn tubes, which suffered the least. From data collected by the American So- ciety for Testing Materials, 1908, and experi- ments reported in the Iron Age, the table on page 159 has been drawn up summarizing the results of a number of other laboratory tests, largely on wrought iron and steel skelp. The specimens were exposed to the weather or im- mersed in different kinds of water and the loss in weight determined at the end of the experiment. These results show that even the steel man- ufactured fifteen years ago was slightly bet- IRON AND STEEL PIPE 159 ■HO Tj(0 a>o mo oo o "=. O tH tH OS s\ S^ ■3 1 S^ m^ ^ fl a o ■^ 1—4 iNrH .a- s -si 1 1.& IE u CM PMPL, PL, PL, J3 3.9* jg 3.& UPLlPW ■■ss 03 a g I i CO «*H a n o sr§ S-^ S*' '■*5 c5 -af=i ^ B ;y- _ ^N fe 5 ^ g a • fH " U 02 rt I &: 1> n < a oTSi 3 U -ga (U ^00 0)0 -§§ fern <; QE2 00 s 00 o o> o OS to CO o a o I ll w s § § w w w 1 1 1 1 a a 6 i lU a> o> o M ^ .£} .a ^ ^ ^ ^ ^ s 1 1 1 1 ri o 160 COEEOSION OF lEON ter than the wrought and charcoal irons of the same period. But while laboratory tests are useful, in giving indications of what may be expected in practice, the results obtained in actual service are naturally more reliable, so about the time the last of the tests were being car- ried out, a new line of investigation in the shape of trial installations was started. Prof. Thomson put in a hot-water line, using wrought-iron and steel pipe alter- nately, so that they would all be subjected to the same conditions. After a year of service he concluded that steel might be expected to stand up • about 7.5 per cent longer, under such conditions, than wrought iron. A short time later a test carried out under similar circumstances under the auspices of the American Society of Heating and Ven- tilating Engineers, using steel and iron pipes supplied by several makers, checked up closely enough with the previous one and the report of the trial concludes : We believe (this test) demonstrates that modern steel pipe of good quality is at least as durable as modern strictly wrought-iron pipe of good quality and is very much superior to a poor quality of wrought iron in this class of work. lEOlSr AND BTBBECi PIPE 161 Various other tests carried out by different observers under a variety of conditions liave led to the same general conclusion. Thus, several experiments conducted by the Pitts- burg Coal Co. and the H. C. Frick Coke Co., to mention only two, wherein the samples were immersed m running mine water, which generally contains considerable quantities of acid, show that steel corrodes at no greater rate than wrought iron. Again, the work of Howe and Stoughton gave results in broad agreement with this. Commenting on the tests of these two experimenters and the evi- dence which they have collected, Alfred Sang says in his book : Of ten different tests made by different observers in different places, seven resulted decisively in favor of steel; in the other three cases the results were very slightly in favor of the iron, but in only one of the latter was the material of modern manufac- ture. The tests which resulted in favor of steel were as follows, all except the two first being carried to de- struction: Seven months in hot, aerated salt water; sixteen months buried in dampened ashes; exposed to sulphuric acid coal-lime water; in railroad inter- locking and signal service; in locomotive boiler service. It was also found that steel tubes made in 1906 pitted much less than those of 1897 from the same makers, indicating the superiority of modem steel 162 COKEOSION OF lEON over that of some years back in this particular re- spect. Not all of the work on this subject has been confined, however, to either laboratory or short installation tests, and at least two rather extended investigations have been un- dertaken on pipe lines as they are found in service. The first was conducted by Prof. Woolson on some of the public bath houses of New York City, where much trouble was be- ing encountered from the rapid corrosion and consequent leaking of the hot-water pipes. A thorough examination of the situation showed that conditions in these bath houses were not materially different from those found in similar installations in other places. Croton water was used, being heated in vari- ous standard types of heaters to a tempera- ture of 160 degrees to 200 degrees F. The quantity of water passing could not be de- termined, but it was considerable, as many thousands of baths were, and are, given in some of these places during the hot weather. It developed, further, that all the pipe which had been in use four years or over had be- gun to give trouble, and in a few places pipe only two or three years old had perforated. Ill all, eleven houses were inspected, but for IKON AND STEEL PIPE 163 various reasons samples could be collected from only eight ; from these eighty-nine spe- cimens, perforated and worthless, were ob- tained for examination. They varied in size from % inch to 4 inches in diameter, the ma- jority being 1^^ to 2 inches. Small pieces were sawed from each of the specimens, crushed, and the fracture care- fully examined by two experts independently to determine whether the material was wrought iron or steel ; chemical analyses were then made of all samples which were at all doubtful in classification. The results of these tests were further checked to avoid all possibility of error. ANALYSES OF PIPE FROM NEW TOBK CITY BATH HOUSES Manganese, Per Cent Carbon, Per Cent Oxide, Per Cent Classified as .48.. .32.. .42.. .38.. .50.. Trace Trace Trace Trace Trace .09 .08 .08 .08 .08 Trace Trace Trace Trace Trace .20 .26 .21 .30 .38 2.85 2.40 2.20 2.50 1.75 Steel Steel Steel Steel Steel Wrought iron Wrought iron Wrought iron Wrought iron Wrought iron 164 CORROSION OF IRON As a matter of interest, in showing the dif- ference between the two kinds of metal, the table on page 163 presents typical results of the chemical analysis. From all this work it was decided that sev- enteen out of the eighty-nine samples were wrought iron; these had come from six of the eight houses yielding samples. Of the other two, no samples were examined from one and those collected from the last were proved to be steel. The above data may be arranged in. tabular form as follows : BATH HOUSES AND SAMPLES OBTAINED FROM EACH Material Supplied Bath House Classification Probably steel nth St. 76th St. 23d St. 41st St. Rivington St. 109th St. 2 Iron, 6 Steel Probably steel 2 Iron, 6 Steel Not certain 2 Iron, 4 Steel Steel 5 Iron, 7 Steel Steel 2 Iron, 23 Steel Steel 4 Iron, 11 Steel 17 56 Total of mixed samples 73 Steel, Allen Street , 16 Steel Total of all samples 89 The contract specifications for all except two of these houses called for wrought-iron pipe, galvanized, but plumbers and dealers IRON AND STEEL PIPE 165 often do not make any close distinction be- tween iron and steel pipe, so the latter may be supplied alone or mixed with the other. A few years ago many mills were making both kinds of pipe and it would be very easy to mix them in shipment. However that may be, it will be seen that of the mixed samples 23 per cent were wrought-iron and, bearing in mind that the two kinds of pipe were used indiscriminately, therefore all exposed to the same conditions, that all samples were tested to destruction and that probably not more than 10 per cent of the welded pipe on the market is wrought iron, it appears that steel showed up to very fair advantage. After his investigation Prof, Woolson de- cided : In my judgment, from the evidence collected, there was absolutely no difference in the corrosion of the two classes of pipe. They appeared to be equally sus- ceptible to the attack. A similar study was made by Dr. Walker in an attempt to determine how these ma- terials stood up in service when both were used in the same system, being separated 166 COEEOSION OF lEON from each other only by a coupling. Under these circumstances the wrought-iron and steel pipe would be subjected to as nearly identical conditions as could be obtained, so a fair basis of comparison should be af- forded. Briefly, the method pursued was to exam- ine a large number of steam and hot-water systems operating in different places and un- der different conditions and to collect samples of wrought iron and steel pipe. It chanced that the majority of the samples collected came from hot and cold-water feed systems, but enough were obtained from live and ex- haust-steam lines and other installations to make the results of general interest. The specimens were split lengthwise and carefully cleaned of rust and accumulated scale by the use of ammonium-citrate solu- tion, which does not attack iron, after which the extent of corrosion was calculated by measuring with a micrometer, the ten deepest pits occurring withia a length of about 12 inches. In all, it was possible to obtain sixty-four comparisons between iron and steel and the results may be summed up as shown in the table at the top of page 167. IRON AND STEEL PIPE IGl* COMPAEATIVB OOEBOSION OF WEOTTGHT-IEON AND STEEL PIPE Instances where iron corroded more than steel. . 20 Instances where steel corroded more than iron. . 18 Instances where iron and steel corroded equally . 9 Instances where corrosion was negligible 17 In the light of the facts brought out by this work, Dr. "Walker felt justified in concluding that on the average there is no difference in the corrosion of iron and steel pipe. Immersion in sulphuric acid of a certain strength has been used as a means of deter- mining the comparative resistance to corro- sion of different samples of iron and steel, and as an interesting sidelight to the above work, Dr. Walker conducted some experi- ments to find out whether this test has any real value. As he reports it: In order to show what relation may exist between the so-called acid corrosion test and the real corro- sion as found in service, eleven pairs of iron and steel were selected and subjected to 20 per cent sul- phuric acid for four hours at room temperature. Four pairs were selected in which the steel was de- cidedly better than the iron in service, four in which the iron had shown decidedly better than the steel, and three in which there was no difference between the two metals. In six instances the relative corrosion as shown by the sulphuric acid test corresponded with the cor- 168 C0EE08I0N OF IRON rosion as found in service. In five instances corro- sion as shovm by the acid test was exactly contrary to that found in service. Although the greatest care was taken to have the specimen of the same size, cleaned in the same way, and in the same physical condition, the results show that no reliance can be placed in this accelerated acid test, but that it may be entirely erroneous and very misleading. Not only did the acid test not agree with the service test when steel was compared with iron, but the steels failed to agree among themselves, and the irons showed no agreement when considered by themselves. It may be added that these conclusions have been reached by other investigators, working independently. After all, it appears that, no matter whether one is testing pipe or paint or almost anything else, an actual service test is the only really reliable crite- rion of merit, or the lack of it. Viewing impartially all of the data pre- sented so far, there seems to be little to choose between wrought-iron and steel pipe, on the whole, as regards their resistance to corrosive influences, but one point may be mentioned with reference to the manner in which these materials corrode. With steel the rusting takes place more or less uniformly over the surface, while wrought iron shows a decided inclination to form deep pits. That this is a dangerous tendency can hardly be IRON AND STEEL PIPE 169 doubted : to paraphrase an old saying, a pipe wall is no stronger than its thinnest spot; therefore, to the extent in which wrought iron exhibits this defect in greater measure than steel, it may be considered correspondingly inferior to the latter. In the course of the investigation men- tioned above. Dr. Walker found that in the most corroded samples (those being practi- cally worthless) the average depth of pitting was .130 of an inch for the iron and .118 for the steel. Likewise, it was found that of twenty-six samples of iron pipe removed from hot-water boiler-feed lines in the power plants of the H. 0. Frick Coke Co. the aver- age of the deepest pit in each was .112 of an inch, while the average of twenty-six steel samples taken from the same lines was .108. Of course, these differences are not at all startling, but they would seem to indicate a weak point in wrought pipe. THE EOT) INDEX PAOB Acid test, as measure of corrodibility 167 Alkaline solutions, effect on corrosion 36 Alloys, as protective coatings 73 " hydrogen-iron 40 Aluminum, effect on corrosion 135 " silicate, China clay 95 American vermilion 96 Ancient irons 7 Angus Smith method of protecting pipe 120 An-ions 14 Annealing 5, 8 Arsenic, effect on corrosion 136 Asbestine, magnesium silicate 95 Asbestos Protected Metal 117 Asphaltum, as protective coating 117 Assembling iron parts of different polarity 38 Atlantic City, paint tests at 109, 113, 113 japans 118 Barium chromate 88 Barytes, barium sulphate 97 Basic carbonate, white lead 91 Benzol 83 Bessemer process 154 Bone black 95 Bontempi process 50 Boiler tubes, tests on materials for 157 Bower-Barff process 48 171 172 INDEX FAQB Bradley process 50 BufBngton process 51 Calcium carbonate, ■whiting 93 " sulphate, gypsum 93 Carbon, effect on corrosion 133 Carbon black 98 Carbonic acid theory of corrosion 25 Cat-ion 15 China clay, aluminum silicate 95 China wood oil, tung oil 80 Chrome green 94 " " blue tone 89 Chromic acid, as rust inhibitor 35, 43 Cleaning of parts before painting 135 Cobalt, effect on corrosion 136 Colloids 31 Combinations of pigments 116 Copper, effect on corrosion 139 " electro-deposition of 69 Copper-clad steel 69 Copper sulphate solution, action on iron 30 " " " standard for Preece test 57 Corrosion due to attack of hydrogen ions 31 " effect of aluminum on 135 " " " arsenic on 136 « « " carbon on 133 " " " cobalt on 136 " " « copper on 139 " " " iron oxide on 38 " " " manganese on 133 " " " nickel on 137 " « " phosphorus on 133 " " " pickling on 39 " " " platinum wire on 37 t< INDEX 173 PAOB Corrosion, effect of segregation of impurities on 5 " " " selenium on 138 " " "silicon on 131 " " " silver on 138 " " " sulphur on 133 " " " stresses and strains on.... 5 " " " tin on 138 " " " tungsten on 138 " " " vibration on 41 " " " wounds on 44 " " " zinc on 37 of iron parts, in tlie ground 41 « « « « running vrater .... 41 " relative, of wrought iron and steel. . . 161 " resistance to, of pure iron 138 " theories of, carbonic acid 35 " " " electrolytic 39 " " " hydrogen peroxide 33 " varying tendency of irons and steels to 4 Coslettizing process 53 Crystalloids 31 Dewees-Wood process 50 Dissociation, theory of 15 Driers, oil and japan 78 " use of 79 Electro-deposition of copper 69 " " " nickel 69 « « « zinc 60 Electrolyte 17 Electrolytic iron 134 " theory of corrosion 39 Electro-plating 18 Enamels 119 174 INDEX PAQB Examination of pipe lines in New York City bath-houses 163 Examination of pipe lines in various installa- tions 165 Ferroxyl reagent < 33 Fish oil, menhaden oil 80 Galvanizing, electro 60 " hot dip method 55 Gesner process 49 Graphite 97 Gypsum, calcium sulphate 92 Havre de Grace bridge, paint tests on 105 Heat bluing 53 Hydrogen, alloy with iron 40 " as a metal 30 " peroxide theory of corrosion 23 Hydrolysis 16 Hydroxyl i. 15 Indeterminate pigments 86 Indian red 96 Indicators ,. 33 Inhibitive pigments 86 Ions, formation of 14 Iron, alloy with hydrogen 40 " electrolytic 134 " impurities in 9, 130 " ore, reduction in blast furnace 129 " oxide, effect on corrosion 38 " pure, resistance to corrosion 128,135 " refining of 152 " " " Bessemer process 154 « " " puddling « 152 INDEX 175 PAOII Iron, sulphide and phosphide films as protec- tive coatings 51 " wrought ,.., 153 Japan drier , , 78 Japans, baking 118 Keystone filler .. .,, ,. 97 Lacquers 123 Lampblack 98 Lead, basic carbonate white 91 " protective coatings of 70 " red 95 " sublimed blue 93 « " white 96 Lemon chrome yellow 93 Linseed oil 76 " « boiling of 77 " " film, as depolarizer 44 Litharge 90 Lithopone : S9 Magnesium silicate, asbestine 95 Manganese, effect on corrosion 132 " " " properties of iron 10 Medium chrome yellow 94 Menhaden oil, fish oil 80 Mineral black 97 Molten zinc process 65 National Coating, for pipe. 132 Newburyport bridge ,.,. 6 Nickel, effect on corrosion 137 " electro-deposition of 69 1761 INDEX FAOB Ochre - 98 Oil drier 78 Oils, China wood, timg 80 " linseed 76 " menhaden, fish 80 " soya bean 80 Old iron parts 6 Orange chrome yellow 93 " mineral 92 Osmotic pressure 13, 14 Oxidation 30 Oxides in steel 155 " " wrought iron 155 Paint films, excluding 101 " " porosity of 101 " " varnish in 104 " manufacture of 75 " tests, at Atlantic City 109, 113, 113 " " on Havre de Grace bridge 105 Passive condition 43 Phenolphthalein, indicator 38 Phosphorus, effect on corrosion 133 PicHing, « " " 39 Pigments, combinations of 116 " indeterminate 86 " inhibitive 86 " stimulative 86 " properties of 94 Pipe lines, examination of 163, 165 " " service tests of 160 Potassium ferricyanide, in ferroxyl reagent. . . 33 Preece test 57 Princes metallic brown 93 Protective coatings, classes of 46 INDEX 177 PAGB Prussian blue 89 Puddling process 152 Eed lead 95 Eeduction 20 Eelative corrosion of wrought iron and steel. . 161 Sabin process, for pipe 131 Schoop process 71 Segregation of impurities, effect on corrosion. . 5 Selenium, effect on corrosion 138 Sherardized iron, structure of 63 Sherardizing , 63 Silicon, effect on corrosion 131 Silver, " " " 138 Solution, phenomenon of 11 " pressure . ^ 13 " " eslct on, of strains and segregation of impurities. 33 Soya bean oil 80 Steel 155 Stimulative pigments 86 Stresses and strains, effect on corrosion 5 Sublimed blue lead 93 " white " 96 Sulphur, effect on corrosion 132 Theory of dissociation 15 " " corrosion, carbonic acid . • 35 « « " electrolytic 39 " " " hydrogen peroxide 33 Tin, effect on corrosion 138 Tung oil, China wood oil 80 Tungsten, effect on corrosion 138 Turpentine 81 178 INDEX PAOB TJltramarine blue 88 Valence . . ..-, 19 Varnish film as depolarizer 43 " in paint, effect of 104 Venetian red 94 Vibration, effect on corrosion 41 Walker's test for porosity of zinc coatings. ... 67 Wells process 49 Willow charcoal 90 White lead, basic carbonate 91 " sublimed 96 Whiting, calcium carbonate 92 Wounds, effect on corrosion 44 Wrought iron 153 " " and steel, relative corrosion of. . 161 « « « " skelp, tests on 159 " " " " pipe, service tests on. 160 Zinc and barium chromate 87 " chromate 87 " coatings, weight of 57, 62, 67 " effect of on corrosion . .>. .v.. 37 " electro-deposition of » 60 " lead white 88 " oxide , 87 li