5 / / (21179-1 OM-7-29) DEPARTMENT OF REGISTRATION AND EDUCATION DIVISION OF THS STATE GEOLOGICAL SURVEY URBANA, ILLINOIS f: \pt Invest 4-/ .C.&.ftj URBANA NOIS STATE GEOLOGICAL SURVEY 1 1IL 667 3 3051 00005 6667 STATE OF ILLINOIS HENRY HORNER, Governor DEPARTMENT OF REGISTRATION AND EDUCATION DIVISION OF THE STATE GEOLOGICAL SURVEY M. M. LEIGHTON, Chief REPORT OF INVESTIGATIONS— NO. 41 L SMOKELESS BRIQUETS: IMPACTED WITHOUT BINDER FROM PARTIALLY VOLATILIZED ILLINOIS COALS Progress Report of a Laboratory Investigation IL SMOKE INDEX: A QUANTITATIVE MEASUREMENT OF SMOKE BY R. J. PIERSOL PRINTED BY AUTHORITY OF THE STATE OF ILLINOIS URBANA, ILLINOIS 1936 Digitized by the Internet Archive in 2012 with funding from University of Illinois Urbana-Champaign http://archive.org/details/ismokelessbrique41pier C 2) STATE OF ILLINOIS HENRY HORNER, Governor DEPARTMENT OF REGISTRATION AND EDUCATION DIVISION OF THE STATE GEOLOGICAL SURVEY M. M. LEIGHTON, Chief REPORT OF INVESTIGATIONS — NO. 41 L SMOKELESS BRIQUETS: IMPACTED WITHOUT BINDER FROM PARTIALLY VOLATILIZED ILLINOIS COALS Progress Report of a Laboratory Investigation II. SMOKE INDEX: A QUANTITATIVE MEASUREMENT OF SMOKE BY R. J. PIERSOL ^SSSSSSKsfesS PRINTED BY AUTHORITY OF THE STATE OF ILLINOIS URBANA, ILLINOIS 1936 STATE OF ILLINOIS Hon. Henry Horner, Governor DEPARTMENT OF REGISTRATION AND EDUCATION Hon. John J. Hallihan, Director Spring-field BOARD OF NATURAL RESOURCES AND CONSERVATION Hon. John J. Hallihan, Chairman Edson S. Bastin, Ph.D., Geology William A. Notes, Ph.D., LL.D., Chem. D., D.Sc, Chemistry John W. Alvord, C.E., Engineering William Trelease, D.Sc, LL.D., Biology Henry C. Cowles, Ph.D., D.Sc, Forestry Arthur Cutts Willard, D. Engr., LL.D., President of the University of Illinois STATE GEOLOGICAL SURVEY DIVISION Urbana M. M. Leighton, Ph.D., Chief GEOLOGICAL RESOURCES Coal, G. H. Cady, Ph.D. Oil and Gas, A. H. Bell, Ph.D. Non-Fuels, J. E. Lamar, B.S. Areal and Engineering Geology, G. E. Ekulaw, Ph.D. Subsurface Geology, L. E. Workman, M.S. Stratigraphy and Paleontology, J. M. Weller, Ph.D. Petrography, R. E. Grim, Ph.D. Physics, R. J. Piersol, Ph.D. geochemistry F. H. Reed, Ph.D., Chief Chemist Fuels, Gilbert Thiessen, Ph.D. Non-Fuels, C. F. Fryling. Ph.D. Analyses, 0. W. Rees, Ph.D. mineral economics W. H. Voskuil, Ph.D., Mineral Economist TOPOGRAPHIC MAPPING (In cooperation with the United States Geological Survey) publications and records g§^7 (2358) PREFACE In 1931 in planning a research program to improve the utilization of Illinois coal, one problem which presented itself was the processing of slack coal into a product which would extend its marketability and value. Success in such a project would be of importance to the State, since it would provide a better domestic fuel from its own resources, and would promote the develop- ment of the coal industry of the State, since approximately one-half of the sized coal produced in Illinois is slack coal (less than 2 inches) which offers problems of marketing for many mines, particularly during certain seasons of the year. The experimental investigation was begun along the line of briquetting slack coal without a binder. A binder, such as tar, is not only ■ expensive, costing approximately 70 cents per ton of briquets, but it also adds to the smokiness of the resultant fuel. Preliminary attempts were made to briquet Illinois coals without binder by heating and applying steady pressure, but this method did not show promise of commercial success, as noted in the first part of Eeport of Investi- gations No. 31. However a systematic investigation of the combined effect of heating and impact blow, rather than steady pressure, yielded excellent briquets, without binder. This was reported in the second part of Eeport of Investigations No. 31 and in Eeport of Investigations No. 37. Still later it was discovered that this same process could be used in mak- ing smokeless briquets without binder, by first removing the smoke producing constituents from the coal. If common binder were used the resultant briquet would not be smokeless. The term smokeless, as used in this report, is used in the trade sense. The trade designates certain low volatile bituminous coals, like those occur- ring in certain beds in West Virginia and adjacent states, as smokeless. They are not truly smokeless, but by the ordinary processes of combustion they yield relatively little smoke as compared with high volatile bituminous coals. In view of this common usage of the term smokeless, the same term is used for products made from high volatile bituminous coals which yield the same amount of smoke as, or less smoke than, the so-called smokeless coals. In the progress of the work it became desirable to develop a laboratory method for a quantitative measurement of the amount of smoke liberated in combustion in order to permit comparison of the smokiness of impact briquets made from coal with partial volatilization with that of the corresponding raw coals. This method is herein referred to as the smoke index method. The smoke index method made it possible later to determine the degree of volatili- zation of the coal necessary to make smokeless briquets. Part I of this preliminary report deals with smokeless briquets made by impact without binder from partially volatilized Illinois coals. Part II deals with the smoke index method and its application to the measurement of smoke in naturally occurring and in processed coals. 14] CONTENTS PAGE I. SMOKELESS BRIQUETS: IMPACTED WITHOUT BINDER FROM PARTIALLY VOLATILIZED ILLINOIS COALS 7 II. SMOKE INDEX: A QUANTITATIVE MEASUREMENT OF SMOKE.. 31 [5] I. SMOKELESS BRIQUETS: IMPACTED WITHOUT BINDER FROM PARTIALLY VOLATILIZED ILLINOIS COALS Progress Report of a Laboratory Investigation R. J. PlERSOL Contents PAGE Chapter I — -Summary 11 Chapter II — Introduction 13 Need foe smokeless fuel 13 Comparison of present and formerly described briquets 14 Protection of process by patent 14 Acknowledgements . 14 Chapter III — Smokeless briquets 15 Introduction 15 Desirability of smokeless coal briquets 15 Methods of production of smokeless briquets 15 Briquetting without binder, by impact, of processed smokeless coal fines 16 Coals used in the investigation 16 Equipment used in the partial volatilization of Illinois coals 17 Rotary oven 17 Exhaust system 19 Equipment used for briquetting 19 Impact apparatus 19 Briquetting die 19 Equipment used in determining mechanical strength 19 Tumbling barrel 19 Smoke index apparatus 20 Experimental procedure in making smokeless briquets 20 Preparation of coal samples for partial volatilization 20 Removal of low-temperature volatile matter 20 Briquetting technique 20 Procedure in making tumbling tests 21 Procedure in determining smoke indices 21 Preparation of samples 21 Smoke index method 21 Experimental results 22 Tumbling tests 22 Effect of amount of volatile matter removed on mechanical strength of briquets 22 Effect of briquetting temperature on mechanical strength of Will County smokeless briquets 24 Time-temperature curve for 15 per cent volatile matter loss 25 Smoke index determinations 28 Estimated energy costs 29 Future investigations 30 m Tables PAGH 1. Proximate analyses of coals used for smoke index and briquetting tests 18 2. Mechanical strength of Will County briquets as affected by volatile matter 23 3. Mechanical strength of Franklin County briquets as affected by volatile matter 24 4. Mechanical strength of Will County smokeless briquets as affected by briquetting temperature 24 5. Time-temperature curve for optimum volatile matter loss of Will County coal. ........ 26 6. Summary of data showing the effect of amount of volatilization on smoke index of Will County briquets 27 7. Effect of amount of volatilization on smoke index of Franklin County briquets 29 Illustrations 1. Mechanical strength of Will County briquets as affected by volatile matter content. . . 22 2. Mechanical strength of Franklin County briquets as affected by volatile matter con- tent 23 3. Effect of briquetting temperature on mechanical strength of Will County smokeless briquets 25 4. Time-temperature curve for 15 per cent volatile matter loss for Will County coal. ... 26 5. Effect of the amount of volatilization on the smoke index of Will County briquets. ... 27 6. Effect of the amount of volatilization on the smoke index of Franklin County briquets 28 r io j L SMOKELESS BRIQUETS: IMPACTED WITHOUT BINDER FROM PARTIALLY VOLATILIZED ILLINOIS COALS Progress Report of a Laboratory Investigation CHAPTER I— SUMMARY ^MOKELESS BRIQUETS may be made from prepared Illinois smokeless ** coal fines from which 15 per cent of the volatile matter (dry basis) has been driven off, with the same briquetting equipment and the same magnitude of impact as briquets made from raw coal, using no artificial binder. The briquetting temperature of the prepared smokeless fines, however, must be between 300° and 400°C. as contrasted to a temperature of 250°C. for the natural coal fines. The mechanical strength of smokeless briquets, as determined by tumbling tests, is slightly greater than that of briquets made from raw coal by the impact method, using optimum conditions for both. Actual power consumption during commercial production can be de- termined only by commercial scale production. However, the impact energy necessary to make one ton of briquets could be supplied by 50 pounds of coal, if the coal is so burned as to produce one horsepower-hour for each two pounds of coal. In order to drive off the desired 15 per cent volatile matter (dry basis), the coal must be preheated up to a temperature of about 900°F. (483°C). Thus, with the above combustion efficiency, it may be calculated that in general about 150 pounds of coal are required for preheating one ten of smokeless briquets. These calculations are dealt with more fully on pp. 29-30. [11] CHAPTEE II— INTRODUCTION NEED FOR SMOKELESS FUEL There is a special need for a cheap smokeless domestic fuel, particularly in the Middle West. While modern stoker equipment has been rather widely employed in industry, with the result that the destructive and harmful effects of smoke from industrial plants have been considerably reduced, yet in the domestic field much less has been accomplished. Estimates based upon a recent survey 1 made in Chicago are that the smoke produced by the domestic group in Chicago since 1911 has increased ninefold for 302,000 homes and four-flat apartments, and sixteenfold for 16,000 six-flat or larger apartments. At present, according to this report, 63 per cent of the smoke in Chicago is produced by the domestic furnace, notwithstanding the fact that it consumes only 49 per cent of the coal. It is a difficult matter to compel domestic consumers to install better equipment. Possible solution lies in two lines of effort: (1) education to im- prove firing methods and to install mechanical equipment wherever it can be afforded; and (2) placing upon the market a prepared coal which will burn without smoke at a cost that will attract the domestic consumer to its use. There is no one universal choice of a smokeless fuel within the price range of the majority of domestic users. Coke is favored by some and it appears that its use is gaining headway. Up to the present, however, Illinois coals have not been extensively coked, but indications are that they will be in- creasingly made into coke. In view of the vast resources of coal in Illinois and the importance of knowing their coking possibilities, the State Geological Survey is carrying on investigations of their coke, gas, and by-product making properties, and plans to publish a report at an early date. The present report, however, concerns the preliminary results of another line of study, namely the preparation of smokeless briquets made by impact without binder from partially volatilized Illinois coals. The discoveries made appear to be promising for providing the domestic consumer with another type of smokeless fuel at a cost that will allow it to enter the competitive market with other fuels. 1 BlackweU, H. D., Chicago smoke survey shows need of education, better equipment Coa Heat 27, March, p. 5 (1935). I 13 J 14 SMOKELESS BRIQUETS COMPARISON OF PRESENT AND FORMERLY DESCRIBED BRIQUETS The partially volatilized coal product used for making impacted smoke- less briquets differs in composition, properties, and appearance from that of previously described products made from Illinois coals. 2 The present product retains a high volatile content (from about 25 to 30 per cent) as contrasted to other smokeless products of relatively low volatile content (usually less than 15 per cent) ; it retains its granular form as contrasted to other smoke- less products possessing a cellular coke-like texture; and it has a high specific gravity, sinking in water, as contrasted to other semi-coked products which have a specific gravity from 0.5 to 0.7. The specific gravity of the smokeless briquet is over 1.2. Because of their relatively greater density, smokeless briquets burn with less rapidity than porous coke. PROTECTION OF PROCESS BY PATENT The process of making smokeless briquets without binder has been pro- tected in the interest of the people of Illinois by U. S. Patent No. 2,021,020. Also, it is planned to apply for patent to protect the discovery that a non-coke smokeless fuel, retaining a relatively high volatile matter content, may be processed from Illinois coals by the preferential distillation of the low-temper- ature fractions of the volatile matter. ACKNOWLEDGMENTS Samples of Illinois coal for processing and for smoke index tests were furnished through the courtesy of various Illinois coal mining companies. The impact machine of the Department of Theoretical and Applied Mechanics of the University of Illinois was used through the courtesy of Pro- fessor M. L; Enger, Dean of the College of Engineering. J. M. Nash and H. C. Eoberts, assistants in the Physics Division, and for a brief time, Dr. F. W. Cooke, carried on the laboratory tests and assisted in construction of apparatus. Chemical analyses were made in the analytical laboratory of the Survey under the direction of Dr. O. W. Pees, Associate Chemist. Assistance in preparation of the report was furnished by Dr. G. H. Cady, Head of the Coal Division of the Geological Eesource Section, and by Dr. M. M. Leighton, Chief. 2 The reader is referred to the following- publications by S. W. Parr : Anthracizing of bituminous coal, Illinois State Geol. Survey Bull. No. 4, p. 196, 1906 ; The modification of Illinois coal by low temperature distillation, Univ. of Illinois Eng. Exp. Sta. Bull. No. 24, 1908 ; The coking of coal at low temperatures, Univ. of Illinois Eng. Exp. Sta. Bull. No. 60, 1912 ; The coking of coal at low temperatures with special reference to the properties and composition of the product, Univ. of Illinois Eng. Exp. Sta. Bull. No. 79, 1915 ; and Low temperature carbonization of coal, Second International Conference on Bituminous Coal, p. 54, Vol. I, 1928. CHAPTER III— SMOKELESS BRIQUETS INTRODUCTION The use of fine sizes of coal in the production of a briquet by impact and without use of an artificial binder has been shown to be possible. 1 By a par- tial volatilization of the coal before impacting, smokeless briquets can be made by the same general technique. DESIRABILITY OF SMOKELESS COAL BRIQUETS Smokeless coal, prepared by partial volatilization, would be in an un- marketable form for many uses unless subsequently briquetted. The smoke- less impact briquets possess several advantages over many other solid fuels. They are preferable to briquets made from carbonized coal with the aid of a binder because such briquets are usually smoky. Their ignition and main- tenance temperatures are believed to be lower than those of most natural smokeless ccals because of their remaining higher volatile matter content. In the tests made in an open grate they burned without swelling and dis- integration, presumably on account of the prior removal of low-temperature volatile fractions. Due to their dense structure, they burned from their sur- face inwardly, similar to a hard coal rather than a soft coal. They are clean to handle as compared with raw bituminous coal and possess the same advantages as any other briquetted coal in respect to uni- formity of size and structure. METHODS OF PRODUCTION OF SMOKELESS BRIQUETS There are at least four methods by which essentially smokeless briquets may be made, namely: (1) subsequent carbonization of coal briquets, formed with binder; (2) subsequent carbonization of coal briquets formed without binder; (3) briquetting of carbonized coal using a binder (which adds to the smokiness of the resultant briquet to an extent depending upon the smokiness of the binder) ; and (4) the process herein described which consists of briquet- ting without binder, by impact, processed smokeless coal fines. In regard to (1), several processes have been devised for subsequent car- bonization of coal briquets containing binder, but none of these has assumed commercial importance in the United States. 1 Piersol, R. J., Briquetting Illinois coals without a binder by impact. Second Report: Illinois State Geol. Survey Report of Investigations No. 37, 1935. ri5] 16 SMOKELESS BRIQUETS In regard to (2), a summary of the various methods of producing briquets without binder for subsequent carbonization was presented in a previous report. 2 At an early stage in the present investigation, an attempt was made to partially carbonize briquets made from natural coal by impact without binder. Preliminary results showed that ordinary briquets could be transformed into smokeless briquets with but slight swelling and cracking, but involving a con- siderable loss in mechanical strength. In the meantime, it was discovered that coal prevolatilized to a smokeless degree could be impacted without binder directly into a smokeless briquet possessing a mechanical strength greater than that of a briquet made direct from the same natural coal by impact without binder. Therefore, the former line of attack was dropped in favor of this method, In regard to (3), various carbonized fuels, such as coke breeze, petroleum coke, and charcoal, and also anthracite fines are briquetted with binder. An excellent summary of the literature on briquetting carbonized fuel with binder is given by Stillman, 3 and a review of the more recent literature is given by the author in the two reports previously cited. BRIQUETTING WITHOUT BINDER, BY IMPACT, OF PROCESSED SMOKELESS COAL FINES The fourth general method of producing smokeless briquets is the process herein described, which consists of briquetting by impact, without binder, processed bituminous coal fines from which the smoke-producing content has been removed. So far as is known, there is no previous literature on this line of investigation. COALS USED IN THE INVESTIGATION For determining the relationship between the volatile matter of raw coals and their smokiness a series of six banded bituminous coals were used of rank varying from high volatile bituminous C to low volatile coal. The high volatile bituminous C rank was represented by coal from Will County, Illi- nois, with a rank index of 120 4 and by coal from Washington County with a rank index of 126 5 ; the high volatile bituminous B rank was represented by coal from Franklin County with a rank index of 131 6 ; medium volatile 2 Piersol, R. J., Briquetting- Illinois coals without a binder by compression and by im- pact : Illinois State Geol. Survey Report of Investigations No. 31, 1933, pp. 14-15. 3 Stillman, A. L., Briquetting applied to carbonized coal, "Briquetting" pp. 336-357, The Chemical Publishing Company (1923). 4 State Geol. Survey Bulletin 62, p. 222, Mine Index No. 359. 5 Idem. p. 279. Mine Index 86. 6 Idem. p. 314, average for Franklin County. VOLATILIZATION EQUIPMENT 17 bituminous coal was represented by coal from the Jewell bed, Wyoming County, West Virginia, with dry mineral-free volatile matter content of 23.3 per cent ; and low volatile bituminous coal by two coals from the Beckley bed, one from Beckley County and the other from Ealeigh County, West Virginia, with dry mineral-matter-free and volatile matter content of 18.1 and 16.4 per cent respectively. Coals of intermediate rank — high volatile bituminous A were not used in the investigation. These coals were also used as a basis of comparison of the smokiness of partially devolatilized coals in the form of briquets and of natural smokeless coal. Table 1 shows the proximate analyses of the Illinois coals used for mak- ing smokeless briquets and of the Illinois and West Virginia coals used for smoke index tests of natural coals. Analyses herein reported for the Illinois coals were made in the Analytical Laboratory of the Geological Survey and those for the West Virginia coals were obtained from Black's Directory, Fourth Edition, 1935. The detailed results of the effect of the degree of volatilization on the smoke index and the influence of the percentage of naturally occurring volatile matter on the smoke index of the coal are given in Part II. EQUIPMENT USED IN THE PARTIAL VOLATILIZATION OF ILLINOIS COALS The laboratory equipment for processing coal by removal of low-temper- ature volatile fractions consists of a rotary oven in which the coal is partially volatilized and an exhaust hood for removing the escaping gases. The equip- ment for the quantitative measurement of the smoke content of both natural and processed coals consists of a smoke index apparatus, to be described. Rotary oven. — The rotary oven used for the partial volatilization of coal in this investigation is a modification of that 7 used previously for the pre- heating of coal to be briquetted. The present rotary oven consists of a heating cell, constructed from a 5%-inch length of 3%-inch pipe, so mounted as to rotate within a stationary 6-inch length of S^-inch pipe, around which is wound the heating element. For the insertion of a thermocouple, a l/4-inch copper tube, with its inner end closed, extends to the center of the cell through the rear end which is remov- able by means of a spanner wrench. The front end of the cell is closed by a permanent steel inset, through which there extends outwardly a 3-inch length of 14-inch steel tubing that serves both as an outlet for the escaping gas and as a means for rotating the heating cell. The rear end of the stationary pipe is closed by a transite inset with an opening through which the thermocouple passes; the front end is open. 7 Piersol, R. J., Briquetting Illinois coals without a binder by impact. Second Re- port : Illinois State Geol. Survey Report of Investigations No. 37, Fig-. 1, p. 21, 1935. 18 SMOKELESS BRIQUETS PQ t^ CO x-9. |l •33 0-2 o-53 CD j>> Q § C5 CM CO CO CM CO Tin CO t-H 00 ©' to o to Oi CO OS >o' CO CO to i>^ CM OC to Ttl tC 14750 14957 15670 OCM CM* CO CM CO O T-l cm cm' do co d coco do t^O CM t>! CO CO (m'co lO t-H t^CM TjH tO C5lO t^ CO* CO t-H T-H CM CO toco CC >i>! tO^JH r-H to coo COl^ CO CO CM CO CO cd 1>I >o CO CM Ol CM CM CO O t^CM rH*io" CM 00 O Oi do' T— 1 T-H CM CM to" to' o to' O T-H tJh'tIh 3 c t— I to CO co d ^1 o X > CD cu PQ. >> CD o CD PQ "a; CD >-3 CO 3 CO 6 « « O U c C c c -+- b e ) c c IS e 5 a 'e 'b > C C ,c b "5 5 Ph ) s .2 'c 'b > C O > CD O CD pa D C .5 'c 'b > C C hi C 1 c > ) CO CO Oi 1— 1 CO Oi T-l CO OS t-H tF CO O T-H CM as to co OS 6 2 >c CI ^ i-c; Jj CD a cd .So •2& la u rtfO 0) O O ° O .5 ^ cmW)bX) g o^-o-dggoJ W +j . rt CD CD fn tl tn o nj^^ a) P cp ■j 3 cc3t! "i^'C'O'd BRIQUETTING EQUIPMENT 19 Exhaust system. — The exhaust system consists of a 10 X 12-inch hood, supported over the front end of the rotary oven, and connected through a flexible tubing to a vacuum cleaner unit that discharges outside of the build- ing through 2-inch piping. EQUIPMENT USED FOR BRIQUETTING The laboratory equipment used in making smokeless briquets consists of a Turner impact machine and a briquetting die. A tumbling barrel is used to determine the mechanical strength of the briquets. Impact apparatus. — The Turner impact machine 8 consists of two ver- tical standards serving as guides for drop hammers of various weights, from 50 to 500 pounds, which are raised to the desired height by an electromagnet and dropped by breaking the electric circuit. Briquetting die. — The impact dies used in making smokeless briquets are all of the same design as the compaction die used previously. 9 The spool- shaped die is made of cold rolled steel, No. 2320 S. A. E., 3.5 per cent nickel, the wearing parts of which are case-hardened. The outer sleeve of the spool is wound with a heating coil, 20 feet of No. 19 resistance wire, which is cov- ered with an asbestos jacket. Eheostat control permits maintenance of the temperature of the die at any desired value up to 400° C. At the higher temperatures hardened clarite steel, quenched at 600 °C, is used for die parts. The temperature of the die is measured by a thermocouple inserted into an opening in the lower part of the outer sleeve. The cylindrical briquet, 1%- inch diameter, is impacted within the space confined by an inner sleeve, a fixed bottom plunger, and a movable top plunger. The impact blow from the ham- mer is transmitted to this movable plunger through an auxiliary plunger extending above the top of the die. EQUIPMENT USED IN DETERMINING MECHANICAL STRENGTH Tumbling barrel. — The same tumbling barrel is used as previously de- scribed in Eeport of Investigations No. 31. It consists of an 8-inch length of an 8-inch inside diameter pipe with 14-inch wall, the ends of which are closed by round steel plates, ^-inch thick, one being removable for the inser- tion and removal of briquets. Three equally spaced 1-inch angle irons that run the length of the barrel act as baffles. The barrel is half filled with flint pebbles, with a total weight of 5000 grams and an approximate weight of 25 grams each. 8 Illustrated in Fig. 8, p. 31, Report of Investigations No. 31. 'Illustrated in Fig. 2, p. 18, Report of Investigations No. 31. 20 SMOKELESS BRIQUETS SMOKE INDEX APPARATUS The smoke index equipment 10 used in this investigation consists of, (a) an electric muffle furnace, so equipped that a specified temperature and rate of air supply can be maintained; (b) a light-absorption tube through which all smoke is drawn; and (c) a smoke density system composed of a source of a beam of constant intensity which passes through the absorption tube, a photo-electric cell at the other end of the tube, and a galvanometer. EXPERIMENTAL PROCEDURE IN MAKING SMOKELESS BRIQUETS Preparation of coal samples for partial volatilization. — Since the size of material which gave most uniform volatilization results and was best adapted to briquetting was minus 4-mesh, all samples were reduced to this size upon receipt at the laboratory, being stored in air-tight receptacles to avoid excessive moisture loss. Immediately before use, in either volatilization or briquetting tests, the samples needed were obtained by quartering from the storage sample. In order to obtain approximately the same size briquets, either 45-gram or 50-gram samples were used, depending on the moisture content of the coal. Removal of low-temperature volatile matter. — In making volatiliza- tion tests the steps in the procedure were as follows: (1) the temperature of the rotary oven (measured by the thermocouple inserted in the copper tube) was raised to a predetermined value by use of an appropriate equilibrium heating current, which maintained a constant temperature throughout the test; (2) the heating cell was removed from the stationary pipe, its removable end opened, the weighed quantity of coal inserted, the end closed, and the loaded cell replaced, the entire operation requiring about 30 seconds; (3) the exhaust motor was started; (4) throughout the test, the heating cell was hand-rotated rapidly at 1 -minute intervals to prevent sticking of the coal and to insure uniform distribution of temperature and degree of volatilization; (5) at the end of the predetermined period of volatilization, the temperature of the coal was recorded; (6) the heating cell was removed; (7) the coal was cooled to a pre-determined temperature; and then (8) the coal was trans- ferred from the heating cell to the impact die, previously heated to the same pre-determined temperature. Briquetting technique. — In the formation of smokeless briquets, the partially volatilized coal at various specified temperatures was transferred to the impact die previously heated to various selected temperatures. The top surface of the coal in the die was leveled, and the movable plunger was lightly pressed down so that it entered the cylinder for a short distance. The loaded Piersol, R. J., Smoke index : a quantitative measurement of smoke. This report, pp. 49-51. DETERMINING SMOKE INDICES 21 die was clamped to the foundation directly beneath the impact hammer as described in the preliminary report already cited. The auxiliary plunger sup- porting a 1-inch steel plate 4X4 inches, was inserted on top of the movable plunger. The 500-pound hammer was dropped from various selected heights, care being taken to avoid a second impact on the rebound. The die was undamped, opened, and the briquet was pressed out of the inner cylinder by means of a hydraulic press at pressures between 500 and 1000 pounds. With downward taper (approximately 0.020-inch increase of diameter per inch) of the inside wall of the portion of the die surrounding the finished briquet, only a slight pressure is necessary to cause the briquet to fall out of the die. Each briquet was weighed immediately in order to determine its combined moisture and volatile loss. PROCEDURE IN MAKING TUMBLING TESTS The tumbling barrel was rotated at 40 r.p.m. for 2 minutes in the de- termination of the tumbling loss for smokeless briquets, all conditions being identical to those previously reported for the determination of the tumbling loss of ordinary briquets. Also, the weight and size of a smokeless briquet made from a 50-gram sample of coal is approximately the same as that of an ordinary briquet made from a 45-gram sample of coal. Therefore, the tumbling losses of the two kinds of briquets are directly comparable. PROCEDURE IN DETERMINING SMOKE INDICES Preparation of samples. — In preparing samples for smoke index de- termination seven or eight 1-cm. cubes were cut from each briquet and three or four cubes from the center of a lump of each coal tested. The latter were cut immediately before testing to avoid air-drying loss as much as possible. They were all approximately the same weight, as determined by actual weighing. Smoke index method. — A complete statement of the smoke index method is given in Part II. Briefly, the procedure for the determination was as follows : The cube of coal, on a nickel dish set on a movable tray, was placed in the center of the furnace. The furnace was maintained at a temper- ature of 600° C. and with an air supply of 4 cubic feet per minute. Galvano- meter readings were taken at 5-second intervals, starting at the instant the sample was placed in the furnace and continuing throughout the period of smoke liberation. The total smoke was calculated as the product of the average amount of smoke produced and the time required for its liberation. The smoke index (smoke per gram) was obtained by dividing this total smoke content by the initial weight of the sample. 22 SMOKELESS BRIQUETS EXPERIMENTAL RESULTS TUMBLING TESTS The results concern (1) the effect of the amount of volatile matter re- moved on the mechanical strength of the resultant briquets, and (2) the effect of the briquetting temperature on the mechanical strength of the briquets. Effect of amount of volatile matter removed on mechanical strength of briquets. — The influence of degree of volatilization on the mechanical strength of briquets was ascertained for Will County and Franklin County coals. Will County coal was volatilized at temperatures of 373 °, 448°, 460°, 466°, 475°, 485°, 494°, and 505°C. for ten minutes, then cooled to 300°C. and impacted by a 4%-foot drop of the 500-pound hammer. It is shown in Table 2 (Fig. 1) that the volatile matter in briquets made from volatilized Will County coal may be as low as 31.9 per cent without 25 20 Z LU U a. iu 15 CL- IO o! z; (CI 5" ol HI z[ MATTER CONTENT OF COAL LUi CO ° O O > ! O l 10 20 30 40 50 VOLATILE MATTER (PER CENT) Fig. 1. — Mechanical Strength of Will County Briquets as Affected by Volatile Matter Content. detrimentally affecting their strength. Such a briquet has a smoke index value less than one-third of that of West Virginia coals, as is shown on pages 28 and 106. Franklin County coal was volatilized at temperatures of 425°, 440°, 455°, 470°, and 482° C. for ten minutes, then cooled to 300° C. and impacted by a 4l/2-foot drop of the 500-pound hammer. EXPERIMENTAL RESULTS 23 Table 2. — Mechanical Strength of Will County Briquets as Affected by Volatile Matter (Data for Fig. 1) Volatilization coal temperature °C. Oven temperature °C. Volatile matter (per cent) Tumbling loss (a) (per cent) 373 425 490 500 510 520 525 530 540 550 43.7 40.5 39.8 38.5 36.8 36.4 35.7 33.0 31.9 4 8 448 1.8 460 2.7 466 1.3 475 2.9 475 1.8 485 3.4 494 2 2 505 2 8 (a) Percentage loss in weight after 2 minutes tumbling. 25 20 15 I 10 -J CO o\ Z| _l' CD 2J =>l MATTER CONTENT OF COAL Zi LLl! *-\ o ! o o ! 9 o ; 10 20 30 40 VOLATILE MATTER (PER CENT) 50 Fig. 2. — Mechanical Strength of Franklin County Briquets as Affected by Volatile Matter Content. The results of tumbling loss tests on a similar series of briquets made from Franklin County coal are shown in Table 3 (Fig. 2). The strength of these briquets is satisfactory for volatile matter content as low as 23.6 per cent. Such a briquet has a smoke index value less than one-seventh of that of West Virginia coals, as shown on pages 28 and 106. 24 SMOKELESS BRIQUETS Table 3. — Mechanical Strength of Franklin County Briquets as Affected by Volatile Matter (Data for Fig. 2) Volatilization coal temperature °C. Oven temperature Volatile matter (per cent) Tumbling loss (a) (per cent) 250 275 460 480 500 520 540 35.9 33.0 30.9 28.6 23.6 22.6 2.8 425 3.6 440 455 1.8 3.5 470 4.9 482 disintegrated (a) Percentage loss in weight after 2 minutes tumbling. Effect of briquetting temperature on mechanical strength of Will County smokeless briquets* — In this series of tests all conditions were main- tained constant except the briquetting temperature (that of the impact die and that to which the coal was cooled). Briquets were made from Will County coal volatilized to 33.1 per cent volatile matter content at a tempera- ture of 460° C. (coal temperature) maintained for five minutes, the shorter period and lower temperature being due to continuous rotation of the oven. Then the coal was cooled to the briquetting temperatures of 250 °, 300 °, 350°, and 400 °C, and then impacted by a 4 1 /2-foot drop of the 500-pound hammer. Table 4.— Mechanical Strength op Will County Smokeless Briquets Containing 33.1 per cent Volatile Matter, Impacted by a 43^-foot Drop of the 500-pound Hammer, from Coal Volatilized at 460 °C. (Coal Temperature) for 5 Minutes, as Affected by Briquetting Temperature (Data for Fig. 3) Briquetting temperature (°C) (a) Tumbling loss (6) (per cent) Individual Average 250 4.8 5.1 1.9 3.5 1.9 1.1 1.3 1.6 0.8 250 5.0 300 300 2.7 350 , 350 1.5 400 400 1.2 400 (a) The briquetting- temperature is that of the impact die and, also, that to which the coal is cooled subsequent to volatilization to 33.1 per cent volatile matter. (b) Percentage loss in weight after 2 minutes tumbling. EXPERIMENTAL RESULTS 25 Table 4 (Fig. 3) indicates that the mechanical strength of smokeless briquets increases with increasing briquetting temperature. Although a briquetting temperature of 400° C. results in a slightly stronger briquet than that obtained at 300 °C, nevertheless the latter tem- perature was selected for use in other tests since it did not require an impact die made of clarite steel. 25 •20 ^' 5 o O 10 o o 9 6 200 250 300 350 TEMPERATURE (DEGREES C ) 400 4 50 Fig. 3. — Effect of Briquetting Temperature on Mechanical Strength of Will County Smokeless Briquets. No tests were made to determine whether the die should be held at the same temperature to which the coal has been cooled, because in the rapid commercial production of briquets the die would tend to maintain the same temperature as the coal. TIME-TEMPERATURE CURVE FOR 15 PER CENT VOLATILE MATTER LOSS There is a wide range of time (8.5 minutes to 40 minutes) possible for volatilization to produce the required reduction in volatile matter, the length of time decreasing with increased temperature of volatilization as shown in Table 5 (Fig. 4), but in commercial practice, in order to maintain the optimum volatile matter content, it would be essential to adjust the tempera- ture to the period of volatilization, or vice versa. This is discussed more fully in Part II of this report, pages 107-110. 26 SMOKELESS BRIQUETS 50 40 ER CENT) 50 Fig. 5. — Effect of the Amount of Volatilization on the Smoke Index of Will County Briquets. Table 6. — Summary of Data Showing the Effect of Amount of Volatilization on Smoke Index of Will County Briquets (Data for Fig. 5) Volatilization coal temperature (°C.) 250° 477° 485° 505° 515° 535° Weight loss (per cent) 0.0 43.9 7.6 39.3 12.7 35.8 17.6 31.9 25.9 24.3 32.9 Volatile matter (a) (per cent) 16.4 Test No. Smoke indices 1 3230 3670 4090 3800 3890 3460 3360 3540 (6) 1880 3360 3400 2030 3040 1730 1910 1640 1590 1470 1880 1980 1820 1920 1820 263 777 550 273 871 526 324 475 245 98 127 64 183 66 104 238 159 2 96 3 140 4 188 5 6 7 8 Average 3640 2610 1770 507 141 146 (a) Percentage volatile matter calculated from determined weight loss. (&) Individual samples showed variation far beyond that to be expected from the heterogeneous structure of the coal, possibly due to greater volatilization of the smaller grains of coal. 28 SMOKELESS BRIQUETS SMOKE INDEX DETERMINATIONS Table 6 (Fig. 5) is a summary table which shows the influence of the remaining percentage of volatile matter on the smoke index of briquets im- pacted from processed Will County coal. A 10-minute volatilization, at a coal temperature of 505 °C, results in 17.6 per cent loss in volatile matter (also weight loss on a dry basis) ; thereby reducing the volatile matter content from 43.9 per cent to 31.9 per cent and reducing the smoke index from an average of about 3600 for the dry coal to an average of about 500 for the briquet im- pacted from the processed coal. This value of smoke index is less than one- third of that of a so-called West Virginia smokeless coal, the average smoke index value of such a coal, with a volatile content of 16.2 per cent, being 1770 (Table 25, West Virginia A coal, Part II of this report). 5000 4000 X u 3000 Q z o 5 2000 1000 A / !, ■VOLATILE MATTER CONTENT OF COAL / \~ • / ...... 1 10 20 30 VOLATILE MATTER (PER CENT) 40 50 Fig. 6. — Effect of the Amount of Volatilization on the Smoke Index of Franklin County Briquets. Table 7 (Fig. 6) is a similar summary table for the smoke index of briquets impacted from processed Franklin County coal. A 10-minute volati- lization, at a coal temperature of 495 °C, results in 16.1 per cent loss in volatile matter; thereby reducing the volatile matter content (dry basis) from 35.9 per cent to 23.6 per cent and reducing the smoke index from an average of about 2600 to an average of about 250, the latter value being less than one-seventh of that of the West Virginia smokeless coal. Eeference may be made to Part II of this report for the detailed results on the smoke index of smokeless briquets. EXPERIMENTAL RESULTS 29 Table 7. — Effect of Amount of Volatilization on Smoke Index of Franklin County Briquets (Data for Fig. 6) Volatilization temperature (°C.) Weight loss Volatile matter (a) . . . . Test No. a b e d e f g h Average 250 450 465 480 4.3 33.1 7.3 30.9 10.3 28.5 Smoke indices 495 16.1 23.6 2480 2120 1140 956 2400 1940 1830 1580 2480 2490 1800 934 2670 2090 1910 1310 2710 2300 1650 937 2710 2580 1290 1610 2820 1620 1590 1220 2440 2590 2160 1600 1220 175 86 176 414 47 449 170 479 250 (a) Percentage volatile matter calculated from experimental weight loss. ESTIMATED ENERGY COSTS Although actual power consumption can be determined only by com- mercial scale production, nevertheless, certain estimates may be made of the cost of the mechanical and heat energy required in making smokeless briquets. Approximately the same impact energy is required as that for making ordinary impacted briquets. It was estimated in Eeport of Investigations No. 37 that 50 pounds of coal (so burned as to produce one horsepower-hour for each two pounds of coal) would be sufficient to supply the mechanical energy to make one ton of briquets. On the above combustion efficiency, about 100 pounds of coal are neces- sary to preheat the coal for one ton of ordinary briquets, but for smokeless briquets approximately 50 per cent more heat is required, so that this method would require about 150 pounds of coal per ton of briquets. An approximate calculation of the heat necessary to raise dry coal to a given temperature may be based on the specific heat of coal. In order to drive off the desired 15 per cent volatile matter the coal must be preheated up to a temperature of about 900°F. (482°C). If the coal has an average temperature of 70 °F. before preheating, this requires an increase of about 830° F. Since coal possesses a specific heat of approximately 0.3. it thus requires 249 B.t.u. to preheat one pound of dry coal. If coal is used as a source of fuel to volatilize the coal and if it is so burned as to deliver 5000 30 SMOKELESS BKIQUETS B.t.u. per pound of coal, then 100 pounds of coal is needed to preheat one ton of dry coal or 120 pounds of coal per ton of smokeless briquets. If the coal is not dry, an additional quantity of fuel is necessary to drive off the moisture, the quantity depending upon the percentage of moisture present. Thus in general about 150 pounds of coal are required for preheating one ton of smokeless briquets. The total energy consumption is, therefore, equivalent to about 200 pounds of coal per ton of briquets. The 15 per cent volatile matter liberated in processing coal to a smoke- less product is readily combustible. Future tests will be made to determine the volume produced per ton of coal processed and its B.t.u. per cubic foot. If desired, this liberated gas may be burned to furnish a part, or perhaps all, of the fuel required to operate the briquetting unit. The shrinkage of coal in producing a ton of smokeless briquets is equal to the percentage moisture plus 15 per cent, which is the amount of volatile matter removed. Therefore, for a particular Illinois coal containing 10 per cent moisture, the total shrinkage in making smokeless briquets would be 25 per cent of the coal used. FUTURE INVESTIGATIONS The present preliminary report describes an exploratory laboratory in- vestigation in which it is shown beyond reasonable doubt that Illinois coal may be processed into a smokeless product by the removal of the very low- temperature fractions of its volatile matter and, second, that this product may be impacted without binder into strong smokeless briquets. This is being followed by further investigations which include : (1) Operating range, including temperature and time of volatilization, impact die temperature, and impact pressure. (2) Properties of smokeless briquets, including mechanical strength, smoke index, weathering and burning characteristics. (3) Effect of the rank of coal and the coal components (banded in- gredients), sulfur, and ash. (4) Systematic tests on a variety of Illinois coals. (5) Further detailed study of the processing of Illinois coal fines into a smokeless product for domestic consumption without subsequent briquetting. IL SMOKE INDEX: A QUANTITATIVE MEASUREMENT OF SMOKE R. J. PIERSOL Contents PAGE Chapter I — Introduction 37 Purpose of investigation 37 Scope of investigation 38 Summary of findings 39 Acknowledgments 40 Chapter II — Previous methods of smoke determination 41 Smoke density measurements 41 Analytical methods 42 Smoke recorders 43 Chapter III — Description of the smoke index method 47 Chapter IV— Equipment and procedure 49 Equipment 49 Furnace 49 Absorption tube 50 Procedure 52 Calibration of apparatus 52 Standardization of equipment with naphthalene 55 Procedure in making smoke index tests on coal 58 Graphical method of calculating smoke index 60 Algebraic method of calculating smoke index 63 Chapter V — Results 65 Effect of air supply 65 Effect of temperature 72 Reproducibility of smoke index 79 Chapter VI— Application of smoke index method. ....■ 81 Smoke index of the natural coals 82 Will County coal 82 Washington County coal 85 Franklin County coal 86 West Virginia coals 88 Relationship between smoke content and volatile matter of natural coals 90 Calculation of volatile matter in partially volatilized coals 91 Smoke index of briquets made by impact from partially volatilized coals 93 Will County briquets 93 Franklin County briquets 100 Relationship between smoke index and volatile matter content of briquets made by impact from partially volatilized coals 106 Relationship between smoke index and volatile matter content of natural bituminous coals 106 Smoke indices of briquets made by impact from processed Illinois coals compared with those made directly from natural coals 107 , : of eliminating the high-smoke-index fraction of the volatile matter 107 Relationship between temperature and time in effecting different amounts of volatili- zation 107 Effect of volatilization temperature on amount of volatile matter removed 108 Will County coal 108 Franklin County coal 108 Time-temperature curve for 15 per cent volatile matter loss 110 Discussion 110 Bibliography Ill [33] Tables PAGE 1. Calibration data for Weston photo-electric cell 52 2. Melting point of naphthalene 55 3. Standardization of equipment (naphthalene tests) 57 4. Standardization of equipment (naphthalene tests) 58 5. Analysis of coal sample used in smoke index tests 59 6. Graphical method of calculating smoke index 61 7. Effect of air supply (2 cubic feet per minute) on smoke index 66 8. Effect of air supply (3 cubic feet per minute) on smoke index 67 9. Effect of air supply (4 cubic feet per minute) on smoke index 68 10. Effect of air supply (5 cubic feet per minute) on smoke index 69 11. Effect of air supply (6 cubic feet per minute) on smoke index 70 12. Effect of air supply on smoke index (summary) 71 13. Effect of temperature (600°C.) on smoke index 73 14. Effect of temperature (700°C.) on smoke index 74 15. Effect of temperature (800°C.) on smoke index 75 16. Effect of temperature (900°C.) on smoke index 76 17. Effect of temperature (1000°C.) on smoke index 77 18. Effect of temperature on smoke index (summary) 78 19. Reproducibility of smoke indices 79 20. Smoke index data on Will County (B) coal (Series No. 1) 82 21. Smoke index data on Will County (B) coal (Series No. 2) 84 22. Smoke index data on Washington County coal 85 23. Smoke index data on Franklin County (B) coal 87 24. Smoke index of West Virginia coals 88 25. Effect of amount of naturally occurring volatile matter on smoke index of coal. ... 90 26. Analyses of briquets volatilized to various stages of volatilization 92 27. Smoke index of nonvolatilized Will County briquet containing 43.9 per cent volatile matter at temperature of 250°^ for 10 minutes 94 28. Smoke index of Will County briquets volatilized to 39.3 per cent volatile matter at temperature of 477°C. for 10 minutes 95 29. Smoke index of eight 1-cm. cubes cut from a Will County briquet volatilized to 35.8 per cent volatile matter at temperature of 485°C. for 10 minutes. 96 30. Smoke index of Will County briquets volatilized to 31.9 per cent volatile matter at temperature of 505°C. for 10 minutes 97 31. Smoke index of Will County briquets volatilized to 24.3 per cent volatile matter at temperature of 515°C. for 10 minutes 98 32. Smoke index of Will County briquets volatilized to 16.4 per cent volatile matter at temperature of 535°C. for 10 minutes 99 33. Smoke index of nonvolatilized Franklin County briquets containing 35.9 per cent volatile matter heated at temperature of 250°C. for 10 minutes 101 34. Smoke index of Franklin County briquets volatilized to 33.1 per cent volatile matter at a temperature of 450°C. for 10 minutes 102 35. Smoke index of Franklin County briquets volatilized to 30.9 per cent volatile matter at a temperature of 465°C. for 10 minutes 103 36. Smoke index of Franklin County briquets volatilized to 28.5 per cent volatile matter at a temperature of 480°C. for 10 minutes 104 37. Smoke index of Franklin County briquets volatilized to 23.6 per cent volatile matter at a temperature of 495°C. for 10 minutes 105 38. Volatile matter content of Will County coal as affected by various volatilization tem- peratures maintained for 10-minute periods 108 39. Volatile matter content of Franklin County coal as affected by various volatiliza- tion temperatures maintained for 10-minute periods 110 [34] Illustrations FIGURE PAGE 1. Ringelmann chart of smoke densities 42 2. Hamler-Eddy smoke recorder 43 3. Chart made by Hamler-Eddy smoke recorder 44 4. Diagram of combustion furnace 49 5. Photo-electric unit for determination of smoke index 50 6. Assembled smoke index apparatus 51 7. Calibration curve for Weston photo-electric cell 53 8. Calibration of manometer 54 9. Naphthalene tests 56 10. Smoke graph 60 11. Smoke graph 62 12. Smoke graph 62 13. Effect of air supply on smoke index 71 14. Effect of temperature on smoke index 78 15. Effect of the amount of naturally occurring volatile matter on the smoke index of coal. 90 16. Effect of amounts of volatile matter on the smoke index of Illinois and West Vir- ginia coals and on briquets made from Franklin and Will County coals 106 17. Volatile matter content of Will County coal as affected by various volatilization temperatures maintained for 10-minute periods 109 18. Volatile matter content of Franklin County coal as affected by various volatiliza- tion temperatures maintained for 10-minute periods 109 [35 j SMOKE INDEX: A QUANTITATIVE MEASUREMENT OF SMOKE CHAPTER I— INTRODUCTION PURPOSE OF INVESTIGATION IT HAS BEEN discovered that Illinois slack coal may be briquetted by impact without the use of an artificial binder/ and that excellent smokeless fuel briquets may be impacted without binder from Illinois slack coal which has been partially volatilized. 2 In the pursuit of the research on smokeless briquets, it became desirable to develop a quantitative measurement of smoke. Such a method of measurement permits the accurate determination of the smokiness of both naturally occurring coals of various volatile matter content and of briquets impacted from coal fines processed to various volatile matter content by the method herein described. The Encyclopedia Britannica describes smoke as follows : "Smoke is a general term applied to the visible exhalations from burning materials. "Nearly all fuels consist essentially of carbon, hydrogen, oxygen and nitrogen, in various proportions and variously combined. In addition, they usually contain a little sulfur, while in solid fuels varying amounts of incombustible mineral ash are also incorporated. If complete combustion were always attainable, no fuel would emit smoke, the final products in such an ideal case being limited to carbon dioxide, water vapor, and free nitrogen, all quite innocuous gases, and invisible unless the water vapor condenses to a cloud of steam. There would, however, if sulfur were present, also be produced small quantities of sulfur dioxide gas, which, also invisible, has a pungent smell, and in contact with air and moisture tends rapidly to be converted into a corrosive acid; while the mineral constituents would remain unburned in the form of ash. "To achieve such finality it is necessary only that a fuel should be brought into contact with enough air for full oxidation while maintained at a temperature sufficiently high for combustion to take place. These conditions, although appar- ently simple, are by no means easy to realize, and in practice some proportion of a fuel always eludes complete combustion. The unburned products vary widely in amount and in composition according to the nature of a fuel and the manner of its use, being in some circumstances inappreciable, in others very large. They are moreover not necessarily in the form of smoke, since with insufficient air 1 Piersol, R. J., Briquetting- Illinois coals without a binder by compression and by im- pact : Illinois State Geol. Survey Report of Investigations No. 31, 1933. 2 Part I of this report, pp. 7-30. T37] 38 SMOKE INDEX carbonaceous materials may emit gaseous intermediate products such as carbon monoxide and unsaturated hydrocarbons; but whether or not smoke is produced, incomplete combustion is always indicative of thermal loss. "Thorough admixture with air is relatively easy to secure in the case of gaseous fuels, which in properly constructed and properly adjusted burners pro- duce neither smoke nor other unburned products in appreciable quantity. An inadequate air supply, however, or the chilling or smothering of the flames, may result in the evolution of unburned gaseous products, including carbon monoxide and oxides of nitrogen, both highly poisonous; or in extreme cases may even cause the deposition of soot. "Owing to the relatively high density of solid fuels, the problem of bringing them into contact with sufficient air for complete oxidation is greatly intensified, and, even with an air supply far in excess of that theoretically required, perfect combustion cannot in practice be counted upon. "With bituminous coals, smoke production to a greater or lesser degree, accord- ing to the circumstances, is practically unavoidable; for such coals are subject to decomposition at temperatures below the ignition point, with the evolution of combustible gases and condensable tarry vapors. These are of so complex a character, and under the action of heat are subject to such complicated chemical changes, that although the more readily ignitible constituents may burst into spasmodic flames, others almost inevitably escape unburned. Coal smoke consists of such unconsumed distillation products, in association with carbon and tarry matter condensed by premature chilling of flame, together with dust and ash entrained by the upward rush of hot air and gases from the grate. Some of this settles on the walls of the flue as soot; the remainder is carried out through the chimney into the atmosphere with the excess air and gaseous products of com- bution, both burned and unburned." SCOPE OF INVESTIGATION The present investigation included, first, a survey of previous methods used for smoke determination; second, development of a laboratory method and equipment for determining the total amount of smoke emitted during the combustion of a small sample of fuel; third, standardization of equipment under most suitable conditions for testing; fourth, methods of calculating the smokiness of a sample of fuel in terms of a factor, referred to as the smoke index ; fifth, the determination of the smoke index of naturally occurring coals of various volatile matter contents ; and sixth, the determination of the smoke index of briquets impacted from Illinois coal fines processed to various volatile matter contents. The method described in this report measures the total quantity of smoke produced by the combustion under standard conditions of a small weighed sample of fuel; previous methods have determined only the gross effect of smoke produced in commercial furnaces. The present equipment is so designed that the total smoke liberated, under controlled conditions, in combustion of a small sample of fuel of known INTRODUCTION 39 weight, passes through a tube for light absorption where the smoke density is measured at regular time intervals by means of a photo-electric cell. The galvanometer deflections produced by the photo-electric cell are used to cal- culate the smoke index of the sample. In order for such a method to be of service it is necessary that results be reproducible within the desired degree of accuracy. Also, it is desirable that the equipment be so standardized that comparative results may be obtained in other laboratories. As is well known, coal does not possess a simple chemical composition, and it is therefore difficult, if not impossible, to use coal as a standard for calibration of smoke measurement equipment in various laboratories. Thus it was necessary to seek a material which might serve as a suitable standard for this purpose. Further, it is desirable to use an expression which signifies the total amount of smoke liberated in burning a unit quantity of fuel under standard conditions of testing. For this, the term smoke index is used. It is arrived at by multiplying the average percentage absorption of light by the time (in seconds) of smoke emission of one gram of fuel. As shown later, the value of the smoke index may be calculated algebraically or determined graphically. SUMMARY OF FINDINGS The equipment developed for the measurement of smoke index consists of an electric muffle furnace in which the sample of fuel may be burned at a certain temperature and with a specified supply of air. All the smoke pro- duced is drawn through an absorption tube at a constant rate of flow and the smoke density is measured by its absorption of a beam of light, of known intensity, passing lengthwise through the tube. The amount of light thus absorbed is determined by a Weston Photronic cell and a galvanometer, the readings being taken at regular intervals. Smoke equipment of this type is not described in the literature, so far as known, and therefore it is impossible to calibrate it against any previous standard. Thus it was desirable to devise some method by which it, or any similar equipment built by another laboratory, may be calibrated against a material having a constant smoke index for specified conditions of burning. It is found that naphthalene, when burned at a furnace temperature of 90°C. (10°C. above its melting point) in a container of suitable size and shape, possesses a constant smoke index. Naphthalene is obtainable in pure form as moth balls. Thus any duplicate equipment for measuring smoke index of coal or other fuel may be calibrated by using naphthalene as a standard material. The smoke index and the rate at which the smoke is emitted are influ- enced greatly by the conditions under which the fuel is burned. In order to 40 SMOKE INDEX find the most suitable temperature for measuring the smoke index of coal, the range from 600° C. to 1000° C. was investigated. Lower temperatures were not investigated since some bituminous coals have ignition temperatures approaching 600° C. The data obtained indicated that the most reproducible results are obtained at 600° C. Therefore this temperature was selected for use in all tests. For the same reason, the influence of air supply on the smoke index was investigated. With an air flow less than 4 cubic feet per minute an excessive quantity of soot was deposited on the inner surfaces of the equipment. Data showed that reproducible results are obtained with an air supply of 4 cubic feet per minute. Also a series of tests were made to determine the degree of reproduci- bility on as nearly identical samples of coal as were obtainable, these samples differing slightly, however, in weight. The time required for combustion indicated a high degree of uniformity of the samples. The results showed a maximum deviation from the average smoke index of 6.3 per cent and a mean deviation of 3.7 per cent. The results of the investigation show that there appears to be a direct proportionality between the smoke index and the percentage of volatile mat- ter in the Illinois and West Virginia coals investigated. For the Illinois coal fines processed by the method herein described, there is also a linear relationship between the smoke index and the percentage of volatile matter in the partially volatilized product, although the decrease in the smoke index is far more rapid than the decrease in volatile matter. Thus a smokeless briquet may be prepared with a much higher volatile matter content than that of a naturally occurring smokeless coal. ACKNOWLEDGMENTS Mr. J. M. Nash, Physics Assistant of the Survey staff, carried out the larger part of the experimental work, with assistance furnished by the Civil Works Administration as follows : Dr. F. W. Cooke, Physicist, Dr. J. J. Gibbons, Physicist, Dr. E. W. Tyler, Physicist, and Mr. P. G. Jones, Physics Assistant. Mr. H. C. Eoberts, Physics Assistant of the staff, designed and constructed the equipment. Dr. C. F. Fryling, Chemist, Non-fuels Division of the Survey, suggested the use of naphthalene for standardization. Dr. O. W. Eees, Associate Chemist, Analytical Division, of the Survey, supervised the chemical analyses. The Peabody Coal Company furnished the sample of coal. CHAPTER II— PREVIOUS METHODS OF SMOKE DETERMINATION Previous methods of smoke determination have included smoke density measurements, analytical measurements, and smoke recorders. Also furnace air drafts have been controlled by photo-electric relays located in the smoke stacks. SMOKE DENSITY MEASUREMENTS One of the most common methods for measuring smokiness has been the determination of the optical density, or opacity, of the smoke as it issues from the stack. This method and all its modifications have the limitation of mea- suring only the density of a continuous column of smoke, thus failing to de- termine the total amount of smoke. However, this method gives valuable information in that the density of smoke liberated from various fuels, or the same fuel at various stages of combustion, may be compared. By far the most widely used of these is the Ringelmann method 1 of de- termining the density of smoke. This method, which was devised by Professor Ringelmann, includes the use of a set of six comparison charts and a method of calculation. The six charts (Fig. 1) represent different degrees of gray, ranging from white to black. This method is described in the smoke abate- ment report of the Chicago Association of Commerce Committee of Investiga- tion on Smoke Abatement as follows : "Each chart is numbered for reference purposes, to-wit: No. — 100% white; No. 1—80% white and 20% black; No. 2—60% white and 40% black; No. 3—40% white and 60% black; No. 4—20% white and 80% black; No. 5—100% black. The white on the chart is the clear background of the white paper on which the chart is made, and the black is in the form of lines in cross-section, for charts Nos. 1, 2, 3, and 4, the width of the lines being such that the proportional amount of black is shown. The whole surface of chart No. 5 is black. "By placing the Ringelmann scale far enough from the eye, the cross-section lines on the four charts, Nos. 1, 2, 3, and 4, become diffused to the eye and appear as different shades of gray, while the white chart, No. 0, and the black chart, No. 5, appear unchanged in color. "Comparative observations of smoke by the Ringelmann scale are made by placing the scale of six charts at the proper distance between the observer and the 1 Smoke Report, Chicago Association of Commerce Committee of Investigation on Smoke Abatement and Electrification of Railway Terminals. [41] 42 SMOKE INDEX smoke to be observed, with a clear background for the smoke, and with no direct rays of the sun entering the eye of the observer. The color of the smoke emitted is then compared with the colors of the six charts — "The formula for computing the per cent of density of smoke is Smoke units x 20 = per cent of density Stack minutes "A 'stack minute' corresponds to the observation of one stack for one minute . . . A 'smoke unit' corresponds to the emission of No. 1 density of smoke for one minute, or its equivalent." Fig. 1, ■■■■■■*■■■•■■■ 2 3 4 RlNGELMANN CHART OF SMOKE DENSITIES. This method evidently is limited to data on the gross amount of smoke passing through a stack. Therefore, it gives pertinent information as to optimum conditions of combustion, but, of necessity, it cannot give a quantita- tive measurement of the amount of smoke liberated per unit weight. ANALYTICAL METHODS A number of tests involving physical and chemical analyses have been developed. For determining the amount of solids in smoke, and extracting these solids for testing purposes, one method is as follows. The stack is fitted with a sampling tube into which part of the smoke is drawn, by means of a vacuum pump, through a fine paper filter which retains the solid part of the smoke. This solid part is weighed, screened to determine particle size, and then analyzed chemically. Solid particles larger than 20-mesh are classed as "coarse cinders," those between 20-mesh and 200-mesh, as "fine cinders," and those passing a 200-mesh screen are called "fuel dust." The solids are then analyzed to determine amounts of tarry matter, combustible matter, min- eral matter or ash, and the sulfur compounds. The gaseous part of the smoke, after having had the solids filtered out, may be analyzed to determine the amount of carbon dioxide, carbon monoxide, oxygen, nitrogen, sulfur compounds, etc., that may be present. Standard methods of gas analysis may be used for this. For convenience, the Hayes portable gas-analysis outfit is often used. These chemical and physical tests give excellent and accurate information on the character, especially the destructive character and offensiveness, of the PREVIOUS METHODS 43 smoke produced. They require complicated and expensive apparatus with readings made by several operators at specified intervals and are consequently expensive. Also the results are similar to those obtained by smoke charts in that information is lacking as to the amount of smoke produced per pound of coal. SMOKE RECORDERS In many industrial districts adjacent to residential areas, ordinances are enforced which limit the density of smoke emitted by stacks. It is impossible to station an observer to take readings continuously at each stack, and to X IIP !!*,.., If IliiilliilW ^•wsiiSilt. . nil ittipt ; ^^^^^^^M sssww^ist^S^^^^^^^&i Fig. 2. — Hamler-Eddy Smoke Recorder. solve this problem, smoke recording devices have been constructed, such as the Hamler-Eddy smoke recorder (Fig. 2). In this instrument, a small amount of smoke is being constantly drawn from the stack, dried, and forced in a fine jet against the white paper surface of a revolving drum to which some 44 SMOKE INDEX 3 o ^ . so ■gS X ri o> c o ft O X5 3 s? -' a r/; v - 5Jj Q ~ s +j H o °.e « CD pj H S-. s ^ M o> ^ <+-, ,0 ?H "1 >M +J P w 1- o c3 a) Pi «W © U2 o o v, ^ ^ ^>- 03 23 ^ H <5i O cd -d _ M | GQ a o 88 CD +-> *h ^ c^ n 3 +J £> X ^ M r^ O tLi a o;^ o f-i A£ s§ £ C3 rS 03 +J ^ +J r£ - CU W)«E, 1—1 m Q3 o CD 5 4-1 rl -M £ W) 0*3 a q-l 2.8 as T3 Cfl 02 CD CD Oi ^ ^ PiHH PREVIOUS METHODS 45 of the solids in the smoke adhere, giving a line of variable density, its density being proportional to the opacity of the smoke. A more recent type of smoke recorder involves the use of the photo- electric cell set in the stack to determine continuously the density of the smoke. With the advent of the smoke abatement movement, a number of such devices have been installed in commercial plants for stack regulation. These instruments are all of the same general type although they differ some- what in detail. A lamp giving a constant intensity of illumination is set inside, in the wall, or outside the stack, and its beam is directed across the stack, through the path of the smoke, to a photo-electric cell near the opposite wall of the stack. All or a definite fractional part of the smoke passing through the stack is allowed to obscure the beam of light, the density fluctua- tions of the smoke thereby causing corresponding fluctuations in the current generated by the photo-electric cell. The output of the cell may be amplified and the fluctuations recorded on a moving drum thus giving a continuous record of the smoke density. Or the cell may be connected to a sensitive relay which will automatically sound a warning signal when the smoke density exceeds the specified value. All of the methods referred to above are intended for measurement of the density of smoke ejected into the atmosphere by a smoke stack. None of them gives a quantitative measurement of the amount of smoke given off by a given quantity of fuel while it is being burned to ash. Therefore, these methods are of little value in determining the inherent smokiness of a fuel, under a given set of conditions. CHAPTEK III— DESCRIPTION OF THE SMOKE INDEX METHOD The smoke index method is a quantitative test which may be performed on any sample of fuel of any convenient weight. However, the present equip- ment has been developed for a sample of approximately one gram. The pos- sibility of using a small sample has its advantages. For instance, tests may be made on one coal ingredient only, such as clarain; or tests may be made to determine the effect of size and shape of small particles (or to determine the effect of texture of a lump sample on the smoke produced, small lump samples may be used). In this study, the samples used were all cut to the same size and shape on a carborundum saw, their weights being slightly variable. In direct contrast to all known previously used methods of measuring smokiness, the smoke index method measures the total amount of smoke pro- duced. Every particle of smoke passes through the smoke absorption tube, re- quiring an appreciable time for its passage. The use of the smoke index method eliminates two defects of the previously used methods : i.e., the neces- sity for the by-passing of a small fraction of the total smoke through the apparatus; and the irregularities in smoke emission due to the periodic addi- tion of fresh coal under ordinary firing conditions. Since all the smoke passes through the absorption tube in the smoke index equipment, it is subject to measurement at every stage of burning, from before the time of ignition to complete ashing, if desired. The exces- sively large amounts of smoke liberated in the initial stages of burning are manifested, not being masked as they would be if other fuel in advanced stages of combustion were present at the same time. In fact, observed data are plotted so that they show clearly the relation of smoke produced to the stages of combustion. The sample to be treated is placed in a muffle furnace, where conditions of temperature and air supply may be accurately controlled. The heat capacity of the furnace is sufficiently large that the heat which the small sample gives off while burning does not raise the temperature of the furnace appreciably. The present furnace weighs about 30 pounds, being constructed of materials which have an average of about 0.2 specific heat. Thus the heat capacity of the furnace is equivalent to that of 6 pounds of water. The calorific value of bituminous coal is usually less than 15,000 B.t.u. Therefore the combustion of a one-gram sample of coal liberates not more than 33 B.t.u. [47] 48 SMOKE INDEX which would raise the temperature of the furnace about 5.5°F. or 3.0°C. The sample is thus exposed to almost constant furnace temperature (within 3.0° C.) throughout all stages of combustion. The smoke, as it issues from the burning sample, is drawn through the absorption tube. A beam of light is constantly passing axially along this absorption tube, striking the light-sensitive photo-electric cell which is con- nected to a galvanometer. When no smoke is present to obscure the light beam, the galvanometer shows a maximum deflection. As smoke enters, it partially intercepts the light falling on the photo-electric cell, and the amount of obscurity, or the smoke density, may be calculated from the change in the galvanometer reading. That is, the amount by which the galvanometer read- ing is decreased, in per cent, represents the proportion of light intercepted by the smoke in the absorption tube, in per cent. This in turn is in direct ratio to the amount of smoke present. If these individual smoke-density percentages, taken at regular intervals, are averaged, the result is the average per cent smoke density produced during the entire period of combustion. If this average is multiplied by the time required for combustion, in seconds, the total amount of smoke given off by that sample, in units of percentage smoke density and seconds time, is obtained. And if this total amount of smoke is divided by the weight of the sample used, the amount of smoke given off per gram of fuel, in terms of percentage smoke density and seconds time, is obtained. This value is called the smoke index. The actual value of the smoke index for any particular sample may be obtained either by a graphical method or by a numerical calculation. Both methods are described herein. CHAPTER IV— EQUIPMENT AND PROCEDURE A detailed description of equipment and procedure is here included in order to permit other investigators to duplicate smoke index determinations. EQUIPMENT The essential equipment consists of an electric muffle furnace, a tube for light absorption by the smoke, a source of air supply, means for drawing the smoke through the absorption tube, a source of constant illumination, a photo- electric cell, and a galvanometer. Furnace. — The electric muffle furnace (Fig. 4) constructed for use in the investigation consisted of a three-inch inside diameter alundum tube, A, Fig. 4. — Diagram of Combustion Furnace. (Illinois State Geol. Survey- Report of Investigations 37, Fig. 15, p. 69, 1935.) 18 inches long, wound with two heating elements, B, of No. 19 "Chromel" wire and each having a resistance of 15 ohms. Each of the two elements had a separate controlling rheostat C, and an ammeter D, in series with it, so that r 49 ] 50 SMOKE INDEX the temperatures of the front and rear parts of the furnace could be controlled separately, if desired. In the naphthalene standardization tests, the two heating elements were connected in series, in order to secure the low temperature of 90° C. In all the smoke index tests on coal, the two heating elements were connected in parallel, so as to make possible the higher furnace temperatures from 600° C. to 1000° C. In this case the same amount of current was passed through each coil, thereby insuring uniform temperature along the length of the furnace. The alundum tube, with the heating coils, was given a coating of alundum cement about i/^-inch thick, fired, and then placed in a steel and transite case E, 8 inches square and 20 inches long, packed with "Sil-o-Cel". A steel tube, F, 2y 2 inches inside diameter and 26 inches long, was fitted into the alundum muffle to protect it and to increase the heat capacity of the furnace and there- by minimize small fluctuations in temperature. A steel tray G, 1% inches wide, %-inch deep, and 30 inches long, was used to carry the container H, in which the sample I, was placed. The thermocouple J, was mounted in the tray with its junction directly under the sample container. The thermocouple leads extended to a potentiometer near the open end of the furnace. Air was introduced into the furnace through a small iron pipe K, leading to the back of the furnace, passing along the bottom of the muffle beneath the tray. The amount of air admitted to the furnace was measured by a calibrated differential manometer of orifice type. Absorption tube. — The smoke given off by the burning sample is drawn from the mouth of the furnace A, through the absorption tube B, by a com- r -TO EXHAUST FAN 15"- t TO COMPRESSED AIR 26" J €>- nS> Fig. 5. — Photo-Electric Unit For Determination of Smoke Index. pressed air aspirator C (Fig. 5). This tube is 34 inches long and 1% inches in- side diameter, and its inner surfaces are blackened. At the end of the absorp- tion tube nearest the furnace there is mounted a 15-watt, 110-volt A. C. inside- frosted incandescent bulb D, in the sleeve, in such a manner that the rounded surface of the bulb is 1 inch from a glass window F, closing the end of the absorption tube. On the other end of the absorption tube is a EQUIPMENT AND PROCEDURE 51 similar sleeve in which the photo-electric cell E is mounted. As the smoke passes through the absorption tube the beam of light from the incandescent bulb, D, is partially obscured, the intensity of the transmitted light being measured by the photo-electric cell E. In the direction parallel to the axis of the tube and with no smoke present, the intensity of illumination on the photo-electric cell is 0.75 foot-candle. The ends of the tube are closed by the thin glass plates F, which are placed 4 inches from the inlet and outlet of the Fig. 6. — Assembled Smoke Index Apparatus. absorption tube. These windows may be cleaned as required, but, due to their position, remain fairly clean throughout one test. The assembled smoke index apparatus is shown in figure 6. The photo-electric cell used in this investigation was a Weston Photronic cell, Model 594. This cell, while not as sensitive as some of the alkali cells, is very durable. It was connected in series with a 3000 ohm resistance and a D'Arsonval type galvanometer (sensitivity = .00082 milli-amperes per milli- meter at a distance of 1 meter). 52 SMOKE INDEX PROCEDURE Before using the smoke index equipment, it was necessary to calibrate the photo-electric cell, the incandescent lamp, and the differential manometer. Calibration of apparatus. — The incandescent lamp, used as a light source, was an ordinary commercial 15-watt General Electric bulb. Its candle- power was determined, using a Bunsen photometric bench and a standard lamp. The candlepower of the incandescent bulb, in the direction along its axis from its rounded end, was found to be 20.4. Table 1. — Calibration Data for Weston Photo-Electric Cell (Data for Fig. 7) Deflection (millimeters) Intensity (foot-candles) 250 .729 234 : .695 214 .632 193 .573 174 .528 157 .481 144 .445 130 .410 121 .381 114 .354 105 331 99 .307 93 .289 87 272 83 ?56 79 .241 75 .227 72 .214 69 204 67 .192 64 .183 A part of the surface of the incandescent lamp is obscured by the sleeve of the absorption tube when the lamp is in place in the apparatus. A deter- mination of the light intensity at the photo-electric cell end of the absorption tube was made entirely separate from other calibrations. The galvanometer deflection from the photo-electric cell was 257 mm. with the lamp in place and supplied with an exact potential of 110 volts. The intensity of illumination on the photo-electric cell necessary to produce this deflection was found to EQUIPMENT AND PROCEDURE 53 be that given by the standard lamp, 46.9? candlepower, at a distance of 7.91 feet. The intensity of illumination was thus calculated, by dividing 46.97 by 7.91 squared, to be 0.75 foot-candle. Also the photo-electric cell was calibrated, using the above standard lamp and varying the intensity of illumination on the light-sensitive surface of the photo-electric cell by placing the standard lamp at various distances. From the known candlepower of the standard lamp, it is possible to calculate the intensity of illumination (in foot-candles) which varies as the reciprocal of the square of the distance between the lamp and the surface of the photo- electric cell. The data for the calibration of the photo-electric cell are re- corded in Table 1 and the corresponding calibration curve is plotted ( Fig. 7 ) . 250 200 | 150 z o o o < 0/ 20 4 60 80 PRESSURE (MM. Hg.) Fig. 8. — Calibration of Manometer. 100 EQUIPMENT AND PROCEDURE 55 Standardization of equipment with naphthalene. — In order to provide a means for determining the reproducibility of smoke index results, it was necessary to standardize the equipment by the smoke index of some common substance which would be of the identical constituency wherever obtained. Coal is heterogeneous, and hence cannot be used as a standard material from laboratory to laboratory. The material finally selected was naphthalene, the reason for its choice being that it may be obtained anywhere in a high degree of purity in the form of moth balls. An index to the purity of naphthalene is its melting point. For the desired degree of purity necessary for reproduci- bility of smoke index, the melting point of naphthalene should be between 80.0 and 81.5 °C. Moth balls were purchased from three different sources and their melting points determined. Table 2. — Melting Point op Naphthalene (Moth Balls) Laboratory number Sample number Melting point °C. 0-968 N-l N-2 N-3 80.5 to 81.5 0-969 80.5 to 81.0 0-970 80.5 to 81.0 The results, given in Table 2, show practically no variation. When burned, naphthalene gives off a very dense, black smoke and the amount per unit weight can be easily and accurately determined by the smoke index method, using a certain controlled set of burning conditions. In determining the smoke index of naphthalene, samples exactly one gram in weight were used. The furnace was maintained at a temperature of 90° C, which is approximately 10° C. higher than the melting point of naphthalene. The air supply was held at a rate of flow of four cubic feet per minute. The weighed samples of naphthalene were placed in the container and put into the furnace and allowed to melt. As soon as the sample was entirely melted, the container was drawn to the mouth of the furnace and the naphthalene ignited by means of a Bunsen blue flame. Immediately upon ignition the container, holding the burning naphthalene, was pushed back to the middle of the furnace and galvanometer readings started. In making these tests the size and shape of the container was found to be very important as regards the rate of burning of the naphthalene. It was desirable to choose the type of a container which would give about the same rate of burning as a one gram sample of coal so that, during any stage of burning, the light passing through the absorption tube would not be completely obscured by the smoke. Since the smoke index method is based on the measurement of per- centage light absorption of smoke, it is evident that the density of the smoke 56 SMOKE INDEX to be measured must be less than that necessary to give complete absorption of light. The container which was eventually found to meet these conditions was made from nickel steel, cylindrical in shape, 1 inch inside diameter, 9/16 inch inside depth, with y 8 inch wall and 1/32 inch bottom. As stated previ- ously, the container was placed about half way back in the furnace for the tests. The bottom of the container was % i ncn f rom "the bottom of the furnace tube. 150 100 50 o v £ S 5 K & s £ ^— ,i b iz \ t ^V V 2ai IE>. 2|50 2 100 u w 50 * """■^ae ^m, ^z r~ t f ■w- ^t f- v ^ v^ ,- V r X IS ^ H ±£ \1 \ ^^A s*- ^ t- (&) \ i 5i t2t 50 100 150 200 250 \* V- ^-^^~ ~—H -,<=' ^-~^^^- ,_ *--^ v £ \- 1 \ r V it \ 4 L it X it YtV i X -/iv Vlt 3T ^ v i 5i \± 5i : x±: X£. .: : \j ■) 50 100 150 200 250 TIME (SECONDS) Fig. 9. — Naphthalene Tests. 50 100 150 200 250 A series of nine duplicate smoke index tests was made, using 1-gram samples of naphthalene, at a furnace temperature of 90° C, an air supply of 4 cubic feet per minute, following the procedure described above. The data for this series of tests are shown in Table 3 (Fig. 9). Table 4 indicates the degree of reproducibility which can be secured with the apparatus at the present time. The average deviation from the average value was 3.1 per cent, the maximum deviation being 7.1 per cent. EQUIPMENT AND PROCEDURE 57 Table 3. — -Standardization of Equipment (Naphthalene Tests) (Data for Fig. 9) Time (seconds) Galvanometer deflections (millimeters) 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175 180 185 190 195 200 205 210 215 220 225 230 235 240 245 141 140 130 118 127 127 118 129 122 126 123 125 126 116 114 114 120 119 113 113 113 111 104 99 101 99 91 88 78 72 68 47 28 32 21 7 4 1 1 3 2 30 90 106 114 114 118 120 117 123 137 136 116 111 114 110 113 114 110 113 116 111 112 110 110 115 110 115 109 107 99 101 87 93 90 87 83 81 74 56 56 48 32 19 8 5 2 1 1 6 70 96 109 111 115 117 118 119 128 124 106 106 106 104 101 99 102 103 107 102 101 100 100 98 97 98 96 92 94 92 85 86 84 76 67 66 48 46 40 31 19 15 15 6 3 2 36 77 94 98 100 102 102 102 103 162 162 150 143 141 139 138 135 134 134 130 131 129 134 130 130 129 125 121 114 114 113 105 108 99 101 85 74 80 61 55 42 28 27 14 12 2 2 2 46 100 122 126 124 127 129 128 141 137 124 116 114 122 115 115 114 117 116 112 113 117 111 119 115 112 110 111 103 96 95 92 86 78 72 65 55 40 25 21 11 i 4 3 3 4 24 82 99 109 115 114 114 115 117 117 118 138 141 120 113 122 119 116 114 112 114 112 115 110 113 113 109 103 111 101 98 89 88 89 83 72 70 62 57 42 33 30 24 16 7 4 2 1 1 1 16 74 95 103 106 109 108 109 111 113 112 138 141 125 117 120 116 118 118 116 116 113 115 115 111 108 112 103 101 102 95 86 86 78 74 64 60 46 52 34 28 16 8 3 3 2 1 5 55 91 105 108 110 111 113 113 114 114 114 136 138 120 110 115 113 110 114 114 114 112 113 111 105 109 111 108 110 100 104 97 98 99 91 82 84 75 61 53 47 44 30 99 1 1 7 60 93 96 108 111 112 110 111 112 112 147 148 137 123 129 126 126 124 124 120 125 124 124 119 117 123 121 114 116 102 108 97 99 92 74 67 58 39 39 31 23 14 5 2 2 1 1 7 65 110 117 117 119 122 123 123 124 124 58 SMOKE INDEX Table 3. — Concluded. Time (seconds) Galvanometer deflections (millimeters) 1 2 3 4 5 6 7 8 9 Total Number of readings A B X T S w I (smoke index) 4563 50 91.3 132.0 30.8 245 7546 1.0 7550 4173 49 85.2 128.0 33.4 245 8180 1.0 8180 3759 50 75.2 115.5 34.9 245 8550 1.0 8550 4752 50 95.0 145.0 34.5 245 8453 1.0 8450 4382 50 87.6 129.5 32.4 245 7938 1.0 7940 4121 50 82.4 125.0 34.1 245 8355 1.0 8350 4094 49 83.5 126.0 33.7 245 8257 1.0 8260 4202 50 84.0 124.0 32.3 245 7913 1.0 7910 4564 50 91.3 135.5 32.6 245 7987 1.0 7990 A = average deflection. B =: average of initial and final deflections. X = average smoke density (percentage) = B — A B X 100. T =s total time (seconds). S = total smoke = X X T. W = wt. of sample (grams). I = smoke index = S/W. Table 4. — Standardization of Equipment (Naphthalene Tests) Test number Smoke index Deviation from average Per cent deviation from average 1 7550 8180 8550 8450 7940 8350 8260 7910 7990 —580 +50 +420 +320 —190 +220 + 130 —220 —140 7.1 2 0.6 3 5.2 4 3.9 5 2.3 6 2.7 7 1.6 8 2.7 9 1.7 Average 8130 252 3.1 Procedure in making smoke index tests on coal. — In making smoke index tests to determine the total amount of smoke liberated in the burning of a given quantity of coal, either powdered or lump samples may be used. Powdered samples of coal will give more representative results if it is desired that the smoke index should be that of composite coal, but lump samples are essential if the effect of the texture of coal on its smokiness is to be con- sidered. Furthermore, the burning of lump samples more nearly approaches the household use of coal. Lump samples of coal were therefore used for smoke index tests made in this investigation. EQUIPMENT AND PKOCEDURE 59 In order to secure duplication of results, all samples were cut from one block of coal, selected on the basis of its apparent uniformity throughout. The most uniform available coal was a column sample of No. 5 seam from Mine No. 30 of the Black Mountain Corporation, Kenvir, Kentucky, with analyses as shown in Table 5. The banded constituent of this coal block was clarain throughout. Table 5. — Analysis of Coal Sample Used in Smoke Index Tests Laboratory number C-1047 "As received" Air dried Moisture- free Moisture and ash free Unit coal Moisture 2.3 35.6 59.6 2.5 0.8 14,251 1.5 35.9 60.1 2.5 0.8 14,362 Volatile matter 36.4 61.1 2.5 0.8 14,581 37.4 62.6 Fixed carbon Ash Total sulfur 0.8 14,960 B. t. u 15,013 The samples were cut to approximately one-centimeter cubes with a carborundum saw. Because of the breaking of the coal upon cutting, it was not possible to get the small cubes of equal weight although most of them were approximately so. Each sample was accurately weighed before being tested. The procedure in making a smoke index test was to have the furnace at the desired temperature, the rate of air flow being set at some given value, and then introduce into the furnace one of the small lump samples. The sample was placed on a small shallow nickel dish H (Fig. 4), which per- mitted a free circulation of air around the coal. The dish was placed on the furnace tray G, which was pushed into the furnace so that the sample, when burning, occupied a position about half way back. Since the lowest temperature (600°C.) used in making smoke index tests was well above the ignition temperature of Illinois coals (maximum recorded value of ignition temperature of Illinois coals being 558°C. 1 ), the sample started to smoke soon after being put in place. Galvanometer readings were taken at five-second intervals starting the instant the sample was placed in the furnace. They were continued until the sample stopped smoking as evi- denced by the change from a yellow flame to a blue flame and also by the return of the galvanometer deflection to an approximately constant value. 1 Arms, R. W., The ignition temperature of coal : Univ. of Illinois Eng. Exp. Sta., Bull. No. 128 (1922). 60 SMOKE INDEX Graphical method of calculating smoke index. — A convenient method for evaluating smoke index data is to plot the densities of smoke as ordinates and the intervals of combustion as abscissae, the enclosed area representing the total amount of smoke produced. Figure 10 shows a smoke graph where actual galvanometer deflections are plotted as ordinates (zero deflection repre- senting no light passing through the absorption tube, and initial deflection no 300 c 250 200 O 150 100 ■\r 50 100 150 20O TIME (SECONDS) Fig. 10. — Smoke Gkaph. light absorption by smoke) and periods of combustion as abscissae. A straight line is drawn between initial and final galvanometer deflections, this decrease in . deflection representing the slight increase in absorption of light by a small deposition of soot on the tube windows during combustion. Consecutive points on the graph are connected by straight lines, instead of by an average curve, since the fluctuations in the smoke produced are better measured in this way. The actual galvanometer deflections are then converted from millimeters to percentage values. The maximum light and total darkness readings were taken as 100 and per cent respective for each deflection. Table 6 lists the EQUIPMENT AND PROCEDURE 61 actual experimental values of galvanometer deflections in millimeters (Fig. 10), the same galvanometer deflections in per cent, and corresponding smoke densities in per cent. Eef erring to figure 10, the galvanometer deflec- tion in per cent for each point is determined by the ratio of the vertical distances between the zero-deflection line AB and this point, and between this zero deflection line and the datum line CD, through this point. Table 6. — Graphical Method op Calculating Smoke Index (Data for Figs. 10, 11, and 12) Time (seconds) 5 10 .15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175 Observed galvanometer deflection (mm)=A 262 261 259 257 252 241 196 134 127 90 36 33 38 29 22 13 16 13 25 27 28 42 39 59 64 89 103 122 130 145 172 219 233 234 240 238 Corrected initial galvanometer deflection (mm)=B 262.0 261.3 260.6 259.9 259.3 258.6 257.9 257.2 256.5 255.8 255.1 254.5 253.8 253.1 252.4 251.7 251.0 250.3 249.7 249.0 248.3 247.6 246.9 246.2 245.5 244.9 244.2 243.5 242.8 242.1 241.4 240.7 240.0 239.4 238.7 238 Percentage galvanometer deflection mm.=A/B x 100 100.0 99.9 99.4 98.9 97.2 93.2 76.0 52.1 49.5 35.2 14.1 13.0 15.0 11.5 8.7 5.2 6.4 5.2 10.0 10.8 11.3 17.0 15.8 24.0 26.1 36.3 42.2 50.1 53. 59. 71. 91.0 97.1 97.7 100.0 100.0 Decrement in galvanometer deflection (mm)=B-A 0.0 .3 1.6 2.9 7.3 17.6 61.9 123.2 129.5 165.8 219.1 221.5 215.8 224.1 230.4 238.7 235.0 237.3 224.7 222.0 220.3 205.6 207.9 187.2 181.5 155.9 141.2 121.5 112.8 97.1 69.4 21.7 7.0 5.4 0.0 0.0 Percentage decrement in galvanometer deflection =B-A x 100 B 0.0 .1 .6 1.1 2.8 6.8 24.0 47.9 50.5 64.8 85.9 87.0 85.0 88.5 91.3 94.8 93.6 94.8 90.0 89.2 88.7 83.0 84.2 76.0 73.9 63.7 57.8 49.9 46 40 2S 9 2 2 0.0 0.0 62 SMOKE INDEX r~s 75 \- 2 LU U a: Id Q. z 50 O H O u _l u. u Q 25 A ( 1 y / 1 •x> J 50 100 150 TIME (SECONDS) 200 Fig. 11. — Smoke Graph. 1 00 1 50 TIME (SECONDS) 200 Fig. 12. — Smoke Graph. EQUIPMENT AND PKOCEDTJRE 63 Since the galvanometer deflection is a measure of the light transmitted rather than the light absorbed by the smoke, the smoke density percentage is the complement of the per cent galvanometer deflection (100 per cent minus the per cent galvanometer deflection). Figures 11 and 12 show respectively graphs of galvanometer deflection in per cent and smoke density in per cent for the data recorded in Table 6. The smoke index of the sample in question can then be determined by measuring the area enclosed in its per cent smoke density-time graph by means of a planimeter, and dividing this area by the weight of the sample tested. Another approximate method might be to cut cardboard (made especially for such type of work) of the same shape as the per cent smoke density-time curves and weighing the cards. The ratio of the weight of such a card to the weight of a card consisting of the total area available in each instance, would be the average smoke density over the total time of combustion. This ratio (in terms of per cent) multiplied by the time of combustion and divided by the weight (in grams) of the sample tested, would give the smoke index. In determining the smoke density-time graphs, a more exact method would be to use an automatic recorder. The expense of such an apparatus prohibited its use in this investigation. The areas generated by such a recorder might be evaluated by either of the above described methods. Algebraic method of calculating smoke index. — Since the graphical method of calculating smoke index is subject to certain errors due to inaccura- cies in measuring areas and, in addition, requires considerable time, it was thought advisable to use an algebraic or numerical method of calculation. All of the values of smoke index given in this report have been calculated by this method which is as follows : The sum of all the galvanometer readings over the period of smoke production is obtained, and the average galvanometer deflection is then calculated by dividing this sum by the total number of readings taken. Let this average galvanometer deflection be called A. The next step is to determine the mean value of the initial and final deflections. Let this mean value be called B. Then the average smoke density in per cent, during the total time of combustion of the sample, will be this mean deflec- tion B, minus the average galvanometer deflection A, the difference being divided by the deflection B, and multiplied by 100; that is, average smoke density (X) = B — A X 100. This value of average smoke density when B multiplied by the total time of combustion T gives the value of the area S, enclosed by the smoke density-time curve, and the area S represents the total amount of smoke given off by the sample tested. Then in order to convert 64 SMOKE INDEX to a common basis for comparison, this area S is divided by the weight of the sample W, which gives the smoke index I of the sample in units of per cent smoke density times seconds, per gram, or I = S/W = % S. D. X T ~w~ This gives a simple and reliable method of calculating the smoke index of a fuel. CHAPTER V— RESULTS EFFECT OF AIR SUPPLY It is well known that the rate and the degree of completeness of the combustion of any fuel depends in part upon the supply of oxygen (air) admitted into the combustion chamber. Therefore it was deemed advisable to investigate the effect of the air supply on the amount of smoke produced by the samples burned in the smoke index tests. To this end a series of 25 tests were made using air supplies of 2, 3, 4, 5, and 6 cubic feet per minute, five tests being made for each rate of air supply. The samples used in these tests were all cut from the same block of coal so that individual sample differences might be minimized. Their weights varied from 0.69 to 1.13 grams, most of them weighing between 0.90 and 1.10 grams. They were all approximately cubical in shape. The furnace temperature was held at 600° C, and the tests were carried out in the usual manner except that the rate of air supply was changed every five tests. The data taken during these tests are tabulated in Tables 7, 8, 9, 10, and 11. At the bottom of each table are given the essentials of the calcula- tions involved in determining the smoke indices. The last row of each table gives the value of smoke index for each of the five tests. These results are summarized in Table 12 (Fig. 13) which gives the individual values of smoke index for each of the five tests, for each air supply, and also the average value of smoke index for each air supply. They indicate that increased air supply decreases the amount of smoke given off. For air supplies less than 4 cubic feet per minute, there was quite an appreciable amount' of soot deposited on the inner walls of the apparatus. For air supplies of 4 cubic feet per minute, or greater, this deposit was small. The results indicate only a small variation of smoke index with variation of the air supply from 3 to 5 cubic feet per minute. Therefore for all subse- quent tests the value of air supply was set at 4 cubic feet per minute, which is sufficiently high to avoid excessive soot. [65 J 66 SMOKE INDEX Table 7. — Effect of Air Supply (2.0 Cubic Feet per Minute) on Smoke Index Time (seconds) Galvanometer deflections (millimeters) 1 2 3 4 5 243 241 237 228 201 185 131 87 58 59 67 22 5 14 29 46 16 15 59 99 163 181 183 186 202 221 235 231 221 209 165 121 98 72 19 11 17 19 26 77 66 86 50 33 35 99 147 137 141 160 175 199 212 216 218 219 248 240 239 237 233 226 210 190 150 26 29 22 17 47 57 91 56 50 57 61 29 39 74 95 149 202 223 225 224 223 218 206 181 153 105 24 19 13 16 18 19 14 15 41 45 36 34 88 130 139 106 157 206 210 219 5 213 10 197 15 178 20 152 25 123 30 142 35 31 40 6 45 11 50 13 55 18 60 26 65 24 70 17 75 56 80 109 85 185 90 201 95 207 100 209 105 110 115 120 125 130 135 140 145 150 . Total 3413 27 126.4 239.0 47.1 130 6123 1.03 5940 3479 29 119.9 225.0 46.7 140 6538 1.04 6290 3522 28 125.8 236.5 46.8 135 6318 1.03 6130 2640 26 101.5 217.0 53.2 125 6650 1.06 6270 2337 21 A B. ./ 111.3 214.0 X 48.0 T S 100 4800 W I (smoke index) .77 6230 A = average deflection. B = average of initial and final deflections. X = average smoke density (percentage) = B — A X 100. T = total time (seconds). S = total smoke = X X T. W — wt. of sample (grams). I =s smoke index = S/W. RESULTS 67 Table 8. — Effect of Air Supply (3.0 Cubic Feet per Minute) on Smoke Index Galvanometer deflections (millimeters) Time (seconds) 1 2 3 4 5 243 240 238 220 200 65 19 16 15 33 46 53 39 44 105 141 121 122 151 183 207 225 227 219 218 217 208 195 171 141 123 89 69 27 23 24 33 40 49 47 77 53 88 97 52 38 70 110 109 154 184 185 251 253 260 240 190 126 103 106 54 59 48 61 89 61 32 49 39 50 76 110 126 143 220 235 253 252 246 229 186 61 13 17 14 19 41 78 82 97 87 104 141 74 42 104 114 155 211 219 213 5 210 10 209 15 208 20 196 25 178 30 138 35 99 40 81 45 73 50 48 55 38 60 15 65 38 70 39 75 27 80 43 85 81 90 96 95 76 100 137 105 178 110 187 115 120 125 130 135 140 145 150 Total 2953 23 128.3 235.0 45.4 110 4994 .99 5040 3110 29 107.2 202.0 46.9 140 6566 1.13 5810 2981 24 124.2 243.0 48.9 115 5624 .99 5680 2839 24 118.2 236.0 49.9 115 5739 1.03 5570 2608 23 A 113 3 B 200 X 43 3 T 110 S 4763 w 76 I (smoke index) 6270 A = average deflection. B = average of initial and final deflections. X = average smoke density (percentage) = B X 100. T === total time in seconds. S =s total smoke = X X T. W = wt. of sample (grams). I = smoke index = S/W. 68 SMOKE INDEX Table 9. — Effect of Air Supply (4.0 Cubic Feet per Minute) on Smoke Index Time (seconds) Galvanometer deflections (millimeters) 1 2 3 4 5 198 195 194 194 192 184 182 171 126 88 69 56 45 41 63 53 53 102 94 90 171 182 186 246 245 242 236 230 217 190 164 135 95 65 75 49 84 110 106 112 113 129 147 80 104 155 194 220 238 237 229 228 223 200 160 123 78 51 41 32 47 27 12 18 24 22 109 101 82 118 149 177 198 204 202 232 222 224 229 209 194 75 17 15 25 33 29 38 61 83 109 131 132 123 113 118 138 166 189 192 224 5 223 10 221 15 215 20 201 25 176 30 35 40 139 104 62 45 36 50 55 24 38 60 15 65 26 70 47 75 35 80 60 85 121 90 132 95 149 100 182 105 197 110 . 115. 120 125 130 135 Total 2929 23 127.3 192.0 33.7 110 3707 .69 5370 4218 27 156.2 241 . 5 35.3 130 4589 .86 5340 2855 25 114.2 215.5 47.0 120 5640 1.06 5320 3097 25 123.9 212.0 41.6 120 4992 1.00 4990 2627 Number readings 22 A 119.4 B 210.5 X 43.3 T . 105 S 4547 W .80 5680 A — average deflection. B = average of initial and final deflections. X = average smoke density (percentage) = B X 100. T = total time (seconds). S = total smoke = X X T. W = wt. of sample in grams. I =: smoke index = S/W. 69 Table 10. — Effect of Air Supply (5.0 Cubic Feet per Minute) on Smoke Index Galvanometer deflections (millimeters) Time (seconds) 1 2 3 4 5 236 235 233 233 231 222 197 167 138 128 107 94 94 86 72 77 54 51 54 66 90 105 134 134 126 167 211 225 213 211 209 201 175 140 105 77 59 42 32 31 36 39 38 40 111 106 104 103 107 126 149 177 194 191 209 205 194 167 137 93 76 72 70 73 69 61 64 64 58 61 90 107. 133 140 130 169 185 186 207 206 204 186 137 114 113 104 85 64 53 47 37 67 77 86 121 123 144 182 194 213 5 209 10 194 15 181 20 142 25 104 30 78 35 52 40 39 45 50 55 41 26 18 60 30 65 70 75 42 61 87 80 80 85 88 90 104 95 142 100 158 105 159 110 161 115 162 120 162 125 167 130 135 140 Total 3967 28 141.7 230.5 38.5 135 5197 .96 5410 3016 26 116.0 202.0 42.6 125 5325 .97 5490 2813 . 24 117.2 197.5 40.7 115 4681 .94 4980 2551 21 121.5 200.5 39.4 100 3940 .72 5470 2900 26 A B X T S W I (smoke index) 111.5 190.0 41.3 125 5163 1.02 5060 A ■= average deflection. B =3 average of initial and final deflections. X = average smoke density (percentage) = B — A X 100. T = total time (seconds). S =5 total smoke = X X T. W = wt. of sample (grams). I = smoke index = S/W. 70 SMOKE INDEX Table 11. — Effect of Air Supply (6.0 Cubic Feet per Minute) on Smoke Index Galvanometer deflections (millimeters) Time (seconds) 1 2 3 4 5 215 212 202 195 175 154 128 119 129 89 62 96 82 88 74 50 74 65 139 173 178 212 211 211 206 195 175 153 138 110 87 71 68 58 70 104 129 147 142 131 137 137 101 125 167 172 172 173 174 203 202 200 187 170 155 125 67 39 44 73 77 85 84 89 97 93 106 132 154 163 165 202 196 176 142 111 75 62 79 101 104 85 111 118 124 103 123 151 167 169 195 5 193 10 192 15 188 20 175 25 156 30 133 35 105 40 107 45 92 50 93 55 93 60 97 65 81 70 61 75 55 80 78 85 73 90 90 95 112 100 137 105 151 110 167 115 169 120 125 130 , 135 140 Total 2699 21 128.5 196.5 34.6 100 3460 .87 3980 3976 28 142.0 193.0 26.4 135 3564 .99 3600 2710 22 123.2 184.0 33.0 105 3465 .91 3810 2399 19 126.3 185.5 32.0 90 2880 .83 3470 2993 A 24 124.7 B 182.0 X : T 32.0 115 s 3680 w .91 I 4040 A = average deflection. B =3 average of initial and final deflections. X = average smoke density (percentage) = B X 100. T =s total time (seconds). S = total smoke = X X T. W =s wt. of sample (grams). I =s smoke index = S/W. 71 7000 6000 5000 4000 3000 2000 1000 o I 7 o I 2 3 4 5 6 AIR SUPPLY (CUBIC FEET PER MINUTE) Fig. 13. — Effect of Air Supply on Smoke Index. Table 12. — Effect of Air Supply on Smoke Index (Summary) (Data for Fig. 13) Smoke Indices Air supply — cu. ft. per minute Test number 2.0 3.0 4.0 5.0 6.0 1 5940 6290 6130 6270 6230 5040 5810 5680 5570 6270 5370 5340 5320 4990 5680 5410 5490 4980 5470 5060 3980 2 3600 3 3810 4 3470 5 4040 Average 6170 5670 5340 5280 3780 72 SMOKE INDEX EFFECT OF TEMPERATURE The amount and character of smoke given off during combustion depends also, in part, on the temperature of the furnace. After determining a suitable rate of air supply, a series of tests was made to determine the temperature at which most accurate and reproducible results were obtained. The highest temperature of the upper portion of the bed of coal in a domestic furnace is from 600°C. to 1000°C. ; accordingly, this range was investigated, in five steps of 100 °C. each. Ten tests were made at each temperature, using dupli- cate samples. In order to minimize sample differences, all samples were taken from the same block of coal. The ignition temperature of the sample was well below the lowest temperature investigated. All samples were approxi- mately cubical in shape, and their weights varied from 0.65 to 1.30 grams. For such small samples, the effect of such differences in size appears to be negligible. The tests were carried out in the manner described previously, the air supply being held at a value of 4 cubic feet per minute. The temperature was raised 100° C. after every ten tests, starting at 600° C. and continuing to 1000°C. The data from these tests are shown in Tables 13, 14, 15, 16, and 17. At the bottom of each table are shown the calculations for smoke indices and the last row in each table gives the ten values of smoke index obtained for that specified temperature. Table 18 summarizes these results, giving the ten individual values of smoke index for each temperature used and, also, the average value of the smoke index for each temperature (Fig. 14). The results of this series of tests show clearly a distinct decrease in smoke index with increasing furnace temperature, the average smoke index for 600° C. being 6390 and for 1000° C. being 2150, which is only one-third as great. This decrease takes place in a fairly uniform manner, although it is less pronounced in the middle than at the ends of the temperature range. As regards the individual values of smoke indices, Table 18 shows that by far the most reproducible results were obtained at a temperature of 600 °C. Since the values of smoke indices are much smaller at higher temperatures, any errors caused by irregularities in combustion will cause a greater per cent error. The temperature of 600° C. was selected for all tests since it gives the most reproducible results. RESULTS 73 Table 13. — Effect of Temperature (600°C.) on Smoke Index (Air Supply — 4 Cubic Feet per Minute) Time Galvanometer deflections (millimeters) (seconds) 1 2 3 4 5 6 7 8 9 10 11 5 262 261 259 257 252 241 196 134 127 90 36 33 38 29 22 13 16 13 25 27 28 42 39 59 64 89 103 122 130 145 172 219 233 234 240 238 236 234 231 224 209 182 145 106 76 51 31 22 15 12 10 17 33 35 50 57 77 99 114 144 184 210 216 257 253 251 247 245 235 211 171 122 95 62 49 31 21 14 14 16 45 26 33 54 76 58 95 155 217 235 223 222 221 217 160 90 72 112 94 70 59 49 52 39 26 29 19 28 45 39 39 60 55 82 104 123 154 143 162 161 204 233 232 231 230 229 226 213 186 159 117 94 81 54 42 41 31 35 29 27 21 46 40 24 24 41 63 62 89 104 135 210 218 216 211 209 206 194 156 105 59 31 20 18 12 12 37 38 42 56 61 81 110 146 185 204 247 242 240 222 192 150 105 65 47 30 28 19 11 13 17 17 30 42 61 90 131 153 167 182 187 204 236 236 232 195 130 129 140 120 84 51 35 31 24 13 12 24 23 47 69 81 90 114 126 147 154 164 210 219 231 229 221 212 204 180 161 110 80 57 47 36 45 37 18 29 74 94 107 98 99 115 128 135 195 218 239 235 226 213 171 130 90 62 47 42 38 32 32 32 40 55 68 72 89 107 113 147 190 209 204 203 10 202 15 196 20 25 183 153 30 85 35 46 40 21 45 15 50 21 55 14 60 28 65 59 70 69 7^ 55 80 80 85 110 90 125 95 186 100. . 197 105 193 110. 115 120 125 130 135 140 145 150 155 160 165 170 175 Total.... Number readings . . . A B X T S w I (smoke index) . 4488 36 124.7 250.0 50.1 175 8768 1.30 6740 3020 27 111.9 226.0 50.5 130 6565 1.06 6190 3288 27 121.8 246.0 50.5 130 6565 1.08 6080 3153 31 101.7 213.5 52.4 150 7860 1.27 6190 3783 33 114.6 224.5 49.0 160 7840 1.24 6320 2193 22 99.7 207.5 52.0 105 5460 .82 6660 3128 27 115.9 241.5 52.0 130 6760 1.03 6560 2900 27 107.4 227.5 52.8 130 6864 1.02 6730 3160 26 121.5 224.5 45.9 125 5738 .96 5980 2679 24 111.6 224.0 50.2 115 5773 .89 6490 2445 22 111.1 198.5 44.0 105 4620 .74 6240 A = average deflection. B =: average of initial and final deflections. X = average smoke density (percentage) = T ==! total time ( seconds ) . S = total smoke = X X T. W =s wt. of sample (grams). I = smoke index = S/W. X 100. 74 SMOKE INDEX Table 14. — Effect of Temperature (700°C.) on Smoke Index (Air Supply — 4 Cubic Feet per Minute) Galvanometer deflections (millimeters ) Time (seconds) 1 2 3 4 5 6 7 8 9 10 222 220 195 197 207 239 251 240 229 222 5 220 195 70 209 134 63 182 110 59 152 82 51 130 71 53 120 66 55 238 218 197 238 234 226 226 222 87 217 10 216 15 190 20 24 35 39 42 48 47 47 112 48 60 25 20 34 41 36 45 44 34 61 60 39 30 22 79 36 31 59 38 47 72 54 36 35 31 98 61 31 73 35 66 67 55 43 40 41 81 91 42 55 55 81 90 104 58 45 42 84 105 61 51 84 91 99 115 71 50 57 43 112 103 43 84 108 68 148 82 55 74 72 31 56 131 140 117 132 57 42 100 127 81 70 63 76 132 115 86 60 116 65 96 84 145 148 67 127 91 76 123 125 70 72 96 154 149 107 120 110 69 138 120 75 61 131 161 158 66 170 123 106 153 113 80 135 179 174 172 49 197 183 103 163 87 85 174 182 189 87 137 208 138 158 171 172 182 184 186 72 90.. 106 95 151 100 152 105 158 110.. 163 115 165 120 167 125 168 130. . 170 135 171 140 171 145 172 150. . Total 1810 19 1846 18 1936 17 1704 17 1447 19 1708 17 2244 18 2639 21 2724 20 3867 Number readings. . . . 30 A 95.3 202.0 52.8 90 102.6 204.5 49.8 85 113.9 184.5 38.3 80 100.2 184.5 45.7 80 76.2 172.0 55.7 90 100.5 218.0 53.9 80 124.7 229.5 45.7 85 125.7 206.0 39.0 100 136.2 207.5 34.4 95 128.9 B 197.0 X 34.6 T 145 S 4752 1.14 4170 4233 .87 4870 3064 .77 3980 3656 .70 5220 5013 1.01 4960 4312 .87 4960 3885 .83 4680 3900 .93 4190 3268 .75 4360 5017 W .88 I (smoke index) . 5700 A = average deflection. B = average of initial and final deflections. X = average smoke density (percentage) = B — A X 100. T = total time (seconds). S = total smoke = X X T. W=jwt. of sample (grams). I =s smoke index = S/W. RESULTS 75 Table 15. -Effect of Temperature (800°C.) on Smoke Index (Air Supply — 4 Cubic Feet per Minute) Time Galvanometer deflections (millimeters) (seconds) 1 2 3 4 5 6 7 8 9 10 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 195 190 160 92 50 29 46 47 56 51 64 95 88 103 97 126 149 150 214 210 120 63 26 27 48 61 42 50 59 77 88 74 94 110 125 143 149 148 147 145 142 144 145 146 149 200 197 121 56 43 38 67 76 86 71 99 110 109 108 145 168 174 175 176 177 201 194 165 71 40 24 51 68 47 45 75 85 100 119 132 133 135 146 161 194 191 170 103 81 54 49 57 43 40 20 38 77 62 75 94 110 137 197 190 174 115 75 68 60 62 71 104 87 87 116 107 132 155 157 159 192 185 120 33 31 45 74 75 64 97 69 64 120 130 121 153 166 166 168 180 140 75 32 27 40 41 55 67 108 119 124 135 141 144 150 152 153 184 180 145 85 49 46 59 60 71 30 47 63 48 51 67 91 135 142 142 142 143 146 187 184 133 82 26 39 34 63 69 117 90 126 147 168 95 100 105 110 115 120 125. . 130 135 Total. . Number readings A B X T S 1788 18 99.3 172.5 42.4 85 3604 1.04 3470 2946 27 109.1 181.5 39.9 130 5187 1.15 4510 2396 20 119.8 188.5 36.4 95 3458 .92 3760 2010 19 105.8 181.0 41.5 90 3735 1.03 3630 1595 18 88.6 165.5 46.5 85 3953 .96 4120 2116 18 117.6 178.0 33.9 85 2882 .74 3890 2073 19 109.1 180.0 39.4 90 3546 .87 4080 1883 18 104.6 166.5 37.2 85 3162 .93 3400 2126 22 96.6 165.0 41.5 105 4358 .80 5450 1465 14 104.6 177.5 41.1 65 2672 w.... I (smoke index) .75 3560 A =s average deflection. B = average of initial and final deflections. X = average smoke density (percentage) = B — A X 100. T = total time in seconds. S =: total smoke = X X T. "W =s wt. of sample (grams). I = smoke index = S/W. 76 SMOKE INDEX Table 16. — Effect of Temperature (900 ? C.) on Smoke Index (Air Supply — 4 Cubic Feet per Minute) Time Galvanometer deflections (millimeters) (seconds) 1 2 3 4 5 6 7 8 9 10 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 195 190 115 18 18 55 71 64 55 104 88 106 127 142 140 142 147 205 202 160 45 22 41 31 39 42 61 7 13 35 70 95 145 147 205 199 145 74 36 48 56 100 92 84 91 90 130 142 153 154 155 213 212 195 113 12 8 37 37 42 73 30 60 126 146 152 153 211 210 135 31 14 17 36 72 122 123 134 146 152 157 157 158 158 160 202 200 145 41 19 40 66 75 85 70 13 50 101 133 139 138 139 140 210 208 193 82 12 34 70 55 101 127 134 136 148 161 169 169 169 170 171 171 172 173 200 131 35 19 48 81 117 80 71 79 105 131 153 161 201 196 165 59 31 49 71 40 25 44 75 104 116 124 129 132 133 200 184 150 96 69 103 82 69 102 127 141 155 167 90 95 100 105 110 Total. . Number readings A B X T S w.... I (smoke index) 1777 17 104.5 171.0 38.9 80 3112 1.02 3050 1360 17 80.0 176.0 54.5 80 4360 1.17 3730 1954 17 114.9 180.0 36.2 80 2896 1.04 2780 1609 16 100.6 183.0 45.0 75 3375 .92 3670 2193 18 121.8 185.5 34.3 85 2916 .91 3200 1796 18 99.8 171.0 41.6 85 3536 .93 3800 3035 22 138.0 191.5 27.9 105 2930 .75 3910 1411 14 100.8 180.5 44.2 65 2873 .85 3380 1694 17 99.6 167.0 40.4 80 3232 1.06 3050 1645 13 126.5 183.5 31.1 60 1866 .75 2490 A = average deflection. B = average of initial and final deflections. X = average smoke density (percentage) = B X 100. T = total time in seconds. S = total smoke = X X T. W = wt. of sample (grams). I = smoke index = S/W. RESULTS 77 Table 17. -Effect of Temperature (1000°C.) on Smoke Index (Air Supply — 4 Cubic Feet per Minute) Time Galvanometer deflections (millimeters) (seconds) 1 2 3 4 5 6 7 8 9 10 5 10 15 20 25 30 35 40 45 50 55 60 186 182 130 133 125 87 96 136 163 179 179 181 187 183 160 25 18 52 64 1 20 71 129 153 154 156 159 161 162 202 197 100 49 76 142 132 95 162 161 180 189 190 181 178 116 33 55 91 116 141 165 168 171 188 189 187 156 21 24 106 157 167 176 177 179 182 175 60 10 39 69 118 114 80 61 125 155 162 189 180 115 36 21 35 79 111 120 139 156 163 166 200 188 157 61 58 25 52 88 115 156 169 195 190 120 30 57 70 66 91 136 157 174 178 180 179 171 130 13 4 20 68 108 141 144 147 151 65 70 75 80 85 Total. . Number readings A B X T S w.... I (smoke index) 1777 12 148.1 183.5 19.3 55 1062 .65 1630 1855 17 109.1 174.5 37.5 80 3000 1.00 3000 1875 13 144.2 196.0 26.4 60 1584 .78 2030 1603 12 133.6 184.5 27.6 55 1518 .78 1950 1539 11 139.9 184.0 24.0 50 1200 .84 1430 1350 13 103.8 172.0 39.7 60 2382 1.07 2230 1510 13 116.2 177.5 34.5 60 2070 .98 2110 1269 11 115.4 184.5 37.5 50 1875 .90 2080 1644 13 126.5 187.5 32.5 60 1950 .71 2750 1276 12 106.3 165.0 35.6 55 1958 .85 2300 A = average deflection. B = average of initial and final deflections. X = average smoke density (percentage) = X 100. T = total time in seconds. S = total smoke = X X T. W •= wt. of sample (grams). I = smoke index = S/W.* 78 SMOKE INDEX 7000 6000 5000 £4000 Q Z UJ | 3000 2000 1000 O o 6 o o 500 600 700 800 900 TEMPERATURE (DEGREES C ) 1000 Fig. 14. — Effect of Temperature on Smoke Index. Table 18. — Effect of Temperature on Smoke Index (Summary) (Air Supply — 4 Cubic Feet per Minute) (Data for Fig. 14) Smoke indices Temperature °C. C00°C. 700°C. 800°C. 900°C. 1000°C Test number— 1 6740 6190 6080 6190 6320 6660 6560 6730 5980 6490 4170 4870 3980 5220 4960 4960 4680 4190 4360 5700 3470 4510 3760 3630 4120 3890 4080 3400 5450 3560 3050 3730 2780 3670 3200 3800 3910 3380 3050 2490 1630 2 3000 3 2030 4 1950 5 1430 6 2230 7 2110 8 2080 9 2750 10 2300 Average 6390 4710 3990 3310 2150 79 REPRODUCIBILITY OF SMOKE INDEX A series of tests was made, using the preferred air supply of 4 cubic feet per minute and the preferred temperature of 600° C. in order to ascertain the degree of reproducibility of smoke index values. The individual results are shown in Table 19. It is seen that the per cent deviation from the average value varies from 0.9 to 6.3 per cent, with a mean deviation of 3.7 per cent. The temperature of the furnace was maintained at 600 ±3° C, by means of a thermocouple and potentiometer. As noted previously, the temperature increases only about 3.0° C. due to the heat liberated by the coal sample. The potential on the lamp fluctuates between 107 and 111 volts during the tests, but these fluctuations are rapid and therefore do not introduce an appreciable error. The air supply remains constant to within a pressure difference of 2 mm. of mercury, which is equivalent to approximately 0.1 cubic foot per minute (see calibration curve for manometer, Fig. 8) for the rate of air flow used, namely 4 cubic feet per minute. Also, there are additional errors due to temperature changes in the photo- electric cell, and of course no block of coal can provide two identical samples. But all these various errors are not cumulative and, therefore, the method has a maximum error of 6.3 per cent and an average error of 3.7 per cent for similar samples of coal. Table 19. — Reproducibility op Smoke Indices (Air Supply — 4 Cubic Feet per Minute) Test number Smoke index Deviation from average Per cent deviation from average 1 6740 6190 6080 6190 6320 6660 6560 6730 5980 6490 6240 +360 —190 —300 —190 -60 +280 +180 +350 —400 +110 —140 5.6 2 3.0 3 4.7 4 3.0 F> 0.9 6 4.4 7 2.8 8 5.5 9 6.3 10 1.7 11 2.2 Average 6380 233 3.7 CHAPTER VI— APPLICATION OF SMOKE INDEX METHOD The smoke index method was first developed as a means for comparing the amounts of smoke given off by impacted briquets (made without binder) and the natural coals from which the briquets were made. It has also been used in this laboratory to determine the relative smokiness of smokeless fuel briquets as compared with ordinary briquets and natural coals. Smokeless fuel briquets, as herein referred to, are briquets impacted without binder from partially volatilized bituminous coal. In this last connection, the smoke index method served admirably the purpose of indicating just when a smokeless fuel condition was reached, thus determining exactly how much volatile matter must be driven off before a smokeless fuel briquet is obtained. The smoke index method may be extended to compare the smokiness of two different types of briquets, of briquets and natural coals of all kinds, and of two different natural coals; in short, it can be used to compare the relative smokiness of any two kinds of fuel within limits. When the method is used merely for comparative purposes, the apparatus need only give comparative results since absolute values are not needed, the tests all being made in the same apparatus held under constant operating conditions, and the relative smokiness of the fuels thus being accurately determined. However, the method need not be limited to one of comparison only. Such a method can be made suitable for determining the inherent smokiness of any fuel under certain controlled conditions and it was with this idea in mind that the series of smoke index tests on naphthalene was made. A standard material should be used for calibrating the apparatus if results are to be reproduced in other laboratories. When this is done there is no reason why the same smoke index of a fuel cannot be obtained in various laboratories. With sets of apparatus giving reproducible results and with representative samples, the smoke index method affords the means of determining the smokiness of any fuel in terms of the amount of smoke given off per unit weight. The smoke index method will give the relative smokiness of two different types of coal without actually burning large quantities of the two coals. The use of the smoke index method is illustrated by the following detailed results on the smoke index of naturally occurring coals of various volatile matter content and that of Illinois coal fines processed to various volatile matter contents by the method herein described. r si i 82 SMOKE INDEX SMOKE INDEX OF THE NATURAL COALS Will County coal. — The experimental data and computations on the smoke index of eight portions of a sample of Will County coal in the first series of tests are given in Table 20. The average analysis of this coal (Table 1, Part I of report) shows a content of 43.5 per cent volatile matter and 9.1 per cent moisture (partly air dried). The smoke index values, computed as described above, ranged from 4330 to 6260, and averaged 5350. This variation in smoke index values appears to be due to the banded character of bituminous coals, banded ingredients varying in their respective smoke content. On account of the heterogeneous character of coal, therefore, the smoke index value is obtained by averaging the values from several determinations. Approximately three months later a second series of six smoke index tests (Table 21) were made on remaining portions of the same sample of coal to determine the effect of storage. As shown by the table, the values ran lower, ranging from 3630 to 4730,, and averaging 4220. These lower values may reflect a possible loss of moisture and volatile matter during storage. Table 20. — Smoke Index Data on Will County (B) Coal (Series No. 1) (Average Analysis of this Coal bed: Volatile Matter 43.5 Per Cent ;Moisture 9.1 Per Cent Partly Air Dried) Time (seconds) Test Samples 2 3 4 5 6 7 8 Galvanometer deflections (a) (mm.) 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 194 195 198 191 194 193 187 184 176 165 145 115 116 119 100 85 76 53 50 47 243 227 250 235 237 210 222 242 227 248 233 235 207 224 241 227 248 230 237 205 222 240 227 250 228 233 205 223 238 227 241 225 232 201 222 237 204 246 224 226 197 219 235 180 230 219 213 186 218 230 160 220 206 186 179 214 224 140 200 200 180 155 205 216 120 200 180 155 140 183 207 100 150 138 147 120 171 199 80 130 110 136 105 175 190 60 90 100 98 164 175 47 66 20 92 77 128 160 43 52 17 79 61 111 142 36 36 30 75 58 110 124 40 30 60 65 110 120 66 30 60 57 60 80 116 36 24 67 50 73 74 111 34 30 77 50 82 57 APPLICATION OF METHOD Table 20.— Concluded 83 Time (seconds) Test Samples 2 3 4 5 6 7 8 Galvanometer deflections (a) (mm.) 100 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175 180 Total Number readings.. A B X T S w I (smoke index) . . . 45 105 44 35 60 66 57 54 64 98 44 16 90 69 50 90 31 90 62 46 80 65 61 30 79 85 69 21 80 67 60 30 43 80 65 90 100 75 61 47 35 85 93 45 90 78 35 60 44 90 110 150 120 82 28 110 57 78 112 100 130 85 36 105 70 65 180 100 200 105 60 150 80 80 200 150 190 150 105 205 63 95 190 160 190 212 160 194 74 101 196 250 200 188 180 199 100 107 200 230 195 200 205 130 114 200 235 195 195 184 180 220 235 196 192 185 220 220 176 4239 5823 4246 4834 4679 4753 4158 4606 37 37 34 35 34 35 35 32 114.6 157.4 124.9 138.1 137.6 135.8 118.8 143.9 185.0 231.5 213.5 242.5 215.0 216.5 201.0 210.5 38.1 32.0 41.4 43.1 36.0 37.3 40.9 31.6 180 180 165 170 165 170 170 155 6858 5760 6831 7327 5940 6341 6953 4898 1.26 1.20 1.22 1.17 1.17 1.22 1.15 1.13 5440 4800 5600 6260 5080 5200 6050 4330 Average smoke index 5350 A = average deflection. B = average of initial and final deflections. X= average smoke density (percentage) = A B X 100. T = total time (sec.) S =3 total smoke = X X T. W = weight of sample (grams). I = smoke index = S/W. a — The highest value of the galvanometer deflection represents no smoke and the lowest value, maximum smoke. See Part II of the Report for complete description of the smoke index method. 84 SMOKE INDEX Table 21. — Smoke Index Data on Will County (B) Coal (Series No. 2) (Same Coal as Series No. 1 after 3 Months Storage) Test samples Time (seconds) 1 2 3 4 5 6 Galvanometer deflections 213 204 197 179 169 126 119 99 79 64 46 46 41 43 26 23 23 33 130 59 109 107 139 132 122 120 143 151 161 175 196 203 199 193 186 159 117 112 112 117 132 70 92 63 85 83 73 72 87 72 43 85 . 83 78 78 92 128 128 122 115 137 180 204 201 191 195 166 140 126 105 79 69 86 40 69 71 58 88 111 105 123 128 102 116 106 110 122 131 107 149 172 182 179 186 202 199 194 192 180 136 160 123 159 96 66 100 70 105 52 90 72 85 115 85 125 122 119 127 115 106 134 132 136 152 191 212 208 202 160 115 76 80 91 80 83 69 60 77 55 72 96 76 92 82 74 115 79 148 141 169 189 204 195 175 124 103 131 65 95 87 102 107 111 124 127 111 128 122 94 93 86 132 134 138 116 138 145 165 174 186 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145 150 . 155. . Total 3474 31 112.1 204.5 45.2 150 6780 1.47 4610 3496 31 112.8 191.5 41.1 150 6165 1.35 4570 4017 32 125.5 195.0 35.6 155 5518 1.38 4000 3940 31 127.1 196.5 35.3 150 5295 1.46 3630 2901 26 111.6 200.5 44.3 125 5538 1.17 4730 3712 29 128.0 195.0 34.4 140 4816 1.28 3760 A B X T Average S smoke W index I (smoke index) 4220 A = average deflection. B = average of initial and final deflections. X= average smoke density (percentage) = B — X 100. T = total time (sec.) S = total smoke = X X T. "W = weight of sample (grams). I = smoke index = S/W. APPLICATION OF METHOD 85 Washington County coaL — Table 22 gives the experimental data on Washington County coal. The smoke index values for nine tests ranged from 3890 to 5050 and averaged 4380. The character of this coal is approximately that indicated by the analysis of Washington County coal in Table 1 (Part I, p. 18). It contained 41.5 per cent volatile matter and 8.5 per cent moisture. Table 22. — Smoke Index Data on Washington County Coal (Analysis of Sample: Volatile Matter 41.5 Per Cent; Moisture 8.5 Per Cent Partly Air Dried) Time (seconds) 220 5 220 10 218 15 218 20 215 25 212 30 195 35 180 40 171 45 152 50 139 55 133 60 127 65 99 70 90 75 76 80 95 85 38 90 38 95 89 100 47 105 55 110 123 115 102 120 82 125 100 130 115 135 137 140 204 145 184 150 190 155 195 160 165 170 175 180 185 Test samples 2 3 4 5 6 7 8 9 Galvanometer deflections (mm.) 226 210 196 223 184 200 218 188 222 210 195 223 182 200 219 190 223 210 198 223 185 198 219 188 222 210 194 222 180 200 216 186 218 208 155 220 182 197 214 184 220 209 136 218 182 190 189 172 213 190 120 212 168 186 185 125 184 177 100 200 151 156 142 91 181 180 65 190 146 125 122 110 146 155 65 173 140 103 75 74 142 140 20 169 118 96 79 64 146 104 50 156 105 109 60 30 119 66 56 117 95 80 51 4 119 79 45 86 45 1 59 29 101 35 82 94 30 30 56 25 90 10 84 69 41 30 83 33 75 17 52 101 46 59 9 33 70 19 68 56 56 37 47 65 25 70 50 62 54 39 35 70 32 76 29 47 27 42 45 30 32 70 57 67 90 29 35 55 60 71 100 46 95 55 60 5] 54 81 41 79 81 56 50 71 63 70 68 90 96 67 50 96 85 54 70 99 90 70 80 109 74 101 60 107 80 77 54 120 99 119 95 124 101 77 87 119 144 172 101 162 110 113 117 131 150 155 105 144 115 124 159 139 150 169 190 153 130 184 140 154 151 173 184 150 145 165 146 170 156 177 148 173 173 146 207 158 184 152 159 172 150 192 163 184 150 163 174 161 199 194 190 194 190 150 158 160 161 163 160 175 153 157 86 SMOKE INDEX Table 22.— Concluded Test samples Time (seconds) 1 2 3 4 5 6 7 8 ; 9 Galvanometer deflctions (mm.) 190 196 160 160 173 195 200 Total . . Number readings A B X T S w.... I (smoke index) . 4459 32 139.3 207.5 32.9 155 5100 1.31 3890 5669 39 145.4 211.0 31.1 190 5909 1.38 4280 4025 34 118.4 186.5 36.5 165 6023 1.45 4150 3262 31 105.2 184.5 43.0 150 6450 1.55 4160 4781 35 136.6 206.5 33.8 170 5746 1.34 4290 4324 36 120.1 171.0 29.8 175 5215 1.31 3980 5062 41 123.5 186.5 33.8 200 6760 1.40 4810 4025 35 115.0 196.5 41.5 170 7055 1.47 4800 3598 36 99.9 172.5 42.1 175 7368 1.46 5050 Average smoke index 4380 A = average deflection. B = average of initial and final deflections. X= average smoke density (percentage) = T = total time (seconds). S = total smoke = X X T. W = weight of sample (grams). I = smoke index = S/W. X 100. Franklin County coal. — -Table 23 gives the experimental data on Frank- lin County coal. The smoke index values for seven tests ranged from 3200 to 3810 and averaged 3650. This coal is similar to that represented by the analysis of Franklin County coal in Table 1 (Part I of report). It contained approximately 33.8 per cent volatile matter and 8.7 per cent moisture. APPLICATION OF METHOD 87 Table 23. — Smoke Index Data on Franklin County (B) Coal (Analysis of Sample: Volatile Matter 33.8 Per cent; Moisture 8.7 Per Cent) Test samples Time (seconds) 1 2 3 4 5 6 7 Galvanometer deflections 5 10 15 216 213 210 208 204 195 178 160 137 125 107 90 74 70 55 50 50 55 90 60 80 120 90 150 125 120 170 177 195 205 202 195 194 190 175 160 142 142 115 102 90 89 100 17 47 42 55 55 85 110 130 105 130 140 190 175 180 190 188 187 188 185 180 174 160 145 135 123 119 115 10 55 40 80 65 80 89 90 92 103 108 136 128 130 130 179 178 176 176 174 163 145 130 110 104 91 81 84 35 40 40 50 45 52 49 83 104 99 114 104 117 133 142 162 157 159 180 178 172 171 158 152 129 119 112 101 76 73 59 90 50 35 38 54 57 59 47 57 78 86 79 106 107 127 135 135 141 137 230 228 225 215 210 197 193 179 169 156 174 50 36 58 64 76 58 64 77 115 63 74 123 128 143 165 194 195 197 197 199 260 250 251 253 240 225 210 205 180 165 142 120 110 30 70 55 80 105 125 120 105 140 135 165 20 105 130 142 170 230 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140. 145 170 175 150 155 Total 3774 29 130.1 205.5 36.7 140 5138 1.35 3810 3562 28 127.2 192.5 33.9 135 4577 1.25 3660 3770 30 125.7 182.5 31.1 150 4665 1.30 3590 3476 31 112.1 169.0 33.7 150 5055 1.37 3690 3298 32 103.1 158.5 35.0 155 5425 1.42 3820 4452 31 143.6 214.5 33.1 150 4965 1.55 3200 4538 30 151.3 245.0 38.2 145 5539 1.46 3790 A B X T Average S smoke W index I (smoke index) 3650 A = average deflection. B = average of initial and final deflections. X= average smoke density (percentage) = T = total time (seconds). S = total smoke = X X T. W = weight of sample (grams). I = smoke index = S/W. X 100. 88 SMOKE INDEX West Virginia coals. — Table 24 gives the experimental data on West Virginia (A) and West Virginia (B), Beckley bed, and West Virginia (C), Jewell bed. For West Virginia (A) the smoke index values for four tests ranged from 1540 to 2070 and averaged 1770; for West Virginia (B) the values for four tests ranged from 1580 to 2200 and averaged 1820; and for West Virginia (C) the values for three tests ranged from 2550 to 2970 and averaged 2720. According to Black's Directory, Fourth Edition, 1935, the coals contain 16.2, 17.7 and 22.5 per cent volatile matter, and 0.7, 0.0, and 1.4 per cent moisture respectively. Table 24. — Smoke Index or West Virginia Coals Time (seconds) 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 Sample A (Beckley Co.) (Volatile matter 16.2 per cent) "as received" 181 181 181 181 181 181 181 179 179 178 172 166 159 150 137 134 134 132 116 137 142 146 151 149 105 83 104 108 122 Sample B (Raleigh Co.) (Volatile matter 17.7 per cent) "as received" Sample C (Jewell Co.) (Volatile matter 22.5 per cent) "as received" Test samples 1 2 1 2 1 2 1 2 1 2 3 Galvanometer deflections (mm.) 184 173 188 187 184 193 186 185 187 184 173 188 187 184 192 186 184 187 184 173 188 188 184 192 186 184 185 184 173 187 188 183 193 186 183 186 184 173 188 187 183 193 185 179 185 183 172 188 187 184 193 186 172 185 181 171 187 186 183 192 186 162 182 179 167 187 186 182 190 184 148 176 175 162 185 184 182 189 183 138 166 169 153 182 180 181 186 180 133 159 162 154 176 173 179 183 172 119 153 161 151 166 171 176 178 163 113 141 152 142 156 165 172 171 138 106 141 149 134 143 153 168 161 132 99 124 156 123 139 141 161 156 138 93 109 131 116 134 129 156 148 164 91 67 123 132 128 116 146 141 159 91 70 139 130 123 107 139 132 172 103 70 136 130 116 108 145 116 177 76 69 133 132 115 139 103 103 154 79 76 114 142 106 126 98 97 134 82 78 98 147 111 118 106 120 121 77 84 76 147 141 115 115 112 132 81 87 101 155 143 98 102 116 161 86 93 109 163 156 105 98 129 181 92 99 117 143 168 117 130 148 184 102 103 114 96 175 123 120 160 108 113 126 110 177 129 125 175 118 119 132 118 180 133 123 182 126 122 188 188 187 186 187 184 177 173 166 166 103 148 149 92 83 83 84 84 87 98 109 123 131 136 143 147 154 161 166 APPLICATION OF METHOD Table 24. — Concluded 89 Sample A (Beckley Co.) (Volatile matter 16.2 per cent) "as received" Sample B (Raleigh Co.) (Volatile matter 17.7 per cent) "as received" Sample C (Jewell Co.) (Volatile matter 22.5 per cent) "as received" Time (seconds) Test samples 1 2 1 2 1 2 1 2 1 2 3 Galvanometer deflections (mm.) 145 131 134 136 139 141 143 153 167 168 147 172 174 132 113 131 152 161 169 170 171 181 183 142 151 162 181 184 130 137 92 105 130 151 174 186 132 138 143 145 150 153 164 166 167 169 131 135 142 148 159 165 170 174 171 150 177 155 178 160 179 165 170 175 180 185 190 Total. . . . Number read- ings A 5662 38 149.0 174.5 14.6 185 2701 1.48 1830 4729 32 147.8 179.0 17.4 155 2697 1.30 2070 5454 37 147.4 172.0 14.3 180 2574 1.67 1540 4985 31 160.8 185.5 13.3 150 1995 1.21 1650 5146 34 151.4 185.5 18.4 165 3036 1.38 2200 5311 36 147.5 179.0 17.6 175 3080 1.60 1930 4827 30 160.9 189.5 15.1 145 2190 1.39 1580 4330 26 166.5 185.0 10.0 125 1250 0.79 1580 5037 39 129.2 177.0 27.0 190 5130 1.73 2970 4940 37 133.5 180.5 26.0 180 4680 1.78 2630 4848 33 146.9 B X T S 183.5 19.9 160 3184 w I (smoke index) . . . 1.25 2550 Average=1770 Average=1820 Average=2720 A = average deflection. B = average of initial and final deflections. X= average smoke density (percentage) = B X 100. T = total time (seconds). S = total smoke = X X T. "W = weight of sample (grams). I = smoke index = S/W. 90 SMOKE INDEX RELATIONSHIP BETWEEN SMOKE CONTENT AND VOLATILE MATTER OF NATURAL COALS The foregoing results are summarized in Table 25 and also in a graph (Fig. 15) in which the average smoke index is plotted against the percentage of volatile matter, an inspection of which indicates -an approximation of a straight-line relationship for the seven coals tested. 6000 5000 3000 2000 1000 • ^ . O yS s s s s y O WEST VIRGINIA COALS • ILLINOIS COALS s s s s s s s s 20 30 40 VOLATILE MATTER (PERCENT) 50 Fig. 15. — Effect of the Amount of Naturally Occurring Volatile Matter on the Smoke Index of Coal. Table 25.— Effect of Amount of Naturally Occurring Volatile Matter on Smoke Index of Coal (Data for Fig. 15) Location Will County (Series No. 1) Will County (Series No. 2) . Washington County ...... Franklin County West Virginia (A) West Virginia (B) West Virginia (C) Bed 6 6 Beckley Beckley Jewell Moisture (a) (per cent) 9.1 (b) 8.5 8.7 0.7 0.0 1.4 Volatile matter (a) (per cent) 43.5 (b) 41.5 33.8 16.2 17.7 22.5 Average smoke index 5350 4220 4380 3650 1770 1820 2720 (a) As received basis. (&) Same coal as used in Will County Series No. 1 after three months storage. APPLICATION OF METHOD 91 CALCULATION OF VOLATILE MATTER IN PARTIALLY VOLATILIZED COALS Analyses of coal heated from 275 °C. to about 500° C. indicate that losses in weight are due, as would be expected, to loss in volatile matter (Table 26). In the present studies this loss in weight is therefore used directly as a means of determining the volatile matter content of the partially volatilized coal according to the following formula : VM — L D D VM x = 100 — L D where VM D is the per cent volatile matter in the raw coal (dry basis) and L D is the per cent loss in weight above 275 °C. or on the dry basis. This calculation may be illustrated by an example taken from values in Table 26. In order to calculate the per cent volatile matter in sample C-737 from the volatile matter in sample C-738, the above equation becomes as follows : 43.9 — 13.4 VM D = = 35.3 100 — 13.4 Eemoval of volatile matter will of course produce a corresponding increase in fixed carbon and ash according to the formula Fixed carbon (or ash) = Fixed carbon (or ash) 100 — L D T = H where T = L is the per cent at a certain temperature and T = H is that at a higher temperature. The volatile matter lost probably will include some sulfur, hence the amount of sulfur present in the coal at increasingly higher temperatures above 275 °C.., can be known only by analysis. The data in Table 26 indicate that for these samples more than 25 per cent of the sulfur is volatilized between 275°C. and 530°C. 92 SMOKE INDEX 53 O5NO00CD(N00 00rH CiCSi-OoO-rtiCOOOOlOO (MNiMNNHHOOl CO CO 00 CO CO M CO CO (N Total sulfur (per cent) d la O ^OiONHNNCOCO CO CO Th Th i-I t-I i-J i-J i-J Is TflHHCOHOOOOl COCOCOCOHHHHO Fixed carbon (per cent) d 13 u OOCOr^OOOTtHO^t^ ' O 00 © ^ © (M* iO !>.' iO'OCOCD>OcOcOCDCD la <1 OOcO^CDO^^TtfiO doOONNO (M* td 1>^ * id i> i>* oo" oo 13 H < COOOJCOHCO^OOH lOcOcOCOt^l^t^l>00 Volatile matter (per cent) d la o CiCOCO!>-OiT-il^iOOi CO »0 CO O >C Q (Mo'tDi* "*COCOCOCOCOCOC\|o © >o © >o o \o NOHCOMONON C-»tC»0>O(NrtiTtiL0iO PQ > 53 O O ■ '• '■ ".s.s'.d.s.s r _ r _ r _ r _cispitifl 773173^^0303030303 GO OONCDiOOTtnocDN COCOCOCOTtlOOOOOOOO l>-l>-t^t^t^00000000 666666666 a Ul 3 a o M Q ~ 0) £ 3 CD r/7 (0 rt C o S3 «S '£ 8£!J TO _£?r-H ojoo pro WW i3 re APPLICATION OF METHOD 93 SMOKE INDEX OF BRIQUETS MADE BY IMPACT FROM PARTIALLY VOLATILIZED COALS The samples of coal from Will and Franklin counties were volatilized to various degrees, and briquets without artificial binder were made by the impact process in order to determine the smoke indices of the partially vola- tilized coal. The minus 4-mesh coal was first heated for 10 minutes at a temperature of 275 °C, the loss of weight being assigned to moisture loss; subsequent losses in weight at higher temperatures were assigned to volatile matter, and from such loss the volatile matter remaining in the coal was calculated, using the formula given above. Will County briquets, — Separate portions of the sample of Will County coal were preheated for 10 minutes, each at a different temperature. These products were then briquetted and from four to eight duplicate smoke index tests were made on 1-cm. cubes cut from the briquets. The data are shown in tabular form as follows : Table 27 for briquetted coal preheated at 250°C. (coal temperature), retaining its original 43.9 per cent volatile matter; Table 28 for the product partially volatilized at 477° C. and containing 39.3 per cent volatile matter; Table 29 for the product par- tially volatilized at 485 °C. and containing 35.8 per cent volatile matter; Table 30, for the product partially volatilized at 505 °C. and containing 31.9 per cent volatile matter; Table 31, for the product partially volatilized at 515 °C. and containing 24.3 per cent volatile matter; and Table 32, for the product partially volatilized at 535° C. and containing 16.4 per cent volatile matter. The data are averaged and summarized in Table 6 (Part I of report), including the individual smoke indices with the exception of those of the sample prevolatilized at 477 °C. shown in Table 28. Apparently volatilization was not uniform throughout this sample, possibly due to the greater volatiliza- tion of the smaller grains of coal. In figure 5 (Part I of report) the smoke index is plotted against the volatile matter content of the partially prevolatilized briquets. This curve indicates a linear relationship for Will County coal for briquets volatilized at temperatures of 250°, 477°, 485°, and 505°C. Those volatilized at tem- peratures of 515° and 535°C. possess a smoke index of less than 150. 94 SMOKE INDEX Table 27. — Smoke Index of Nonvolatilized Will County Briquet Containing 43.9 Per Cent Volatile Matter at Temperature op 250°C. for 10 Minutes Test samples Time (seconds) 1 2 3 4 5 6 7 Galvanometer deflections (mm. ) 201 199 191 171 136 112 90 91 89 86 78 79 90 100 111 101 115 100 100 100 100 105 115 124 139 155 160 175 219 214 216 215 206 186 164 151 90 66 90 71 81 82 92 92 96 101 84 85 84 93 99 86 104 121 140 161 169 187 188 195 203 195 186 183 130 84 79 63 80 45 71 71 79 85 94 100 95 85 100 102 112 115 131 138 86 110 143 155 174 178 173 174 177 178 179 230 229 222 150 98 115 92 86 103 80 76 103 82 73 103 98 103 81 86 105 110 129 140 176 170 177 191 210 206 185 144 125 104 106 64 56 76 70 79 91 95 96 99 100 115 113 130 91 109 105 163 "182" 212 200 181 143 125 110 134 71 97 102 100 99 110 107 98 104 106 96 104 114 114 119 150 157 170 183 192 189 188 191 193 204 5 202 10 202 15 200 20 199 25 179 30 147 35 113 40 98 45 92 50 89 55 82 60 95 65 110 70 89 75 96 80 120 85 133 90 136 95 147 100 142 105 153 110 172 115 181 120 198 125 130 135 140. . 145 150 155 160 165 170 Total 3413 28 121.9 188.0 35.2 135 4752 1.47 3230 4228 32 132.1 207.0 36.2 155 5611 1.53 3670 4353 35 124.4 191.0 34.9 170 5933 1.45 4090 3408 27 126.2 210.5 40.0 130 5200 1.37 3800 2914 25 116.6 196.0 40.5 125 5063 1.30 3890 4259 31 137.4 202.5 32.1 150 4815 1.39 3460 3579 Number of readings A 25 143.2 B 201.0 X 28.8 T 120 S 3456 w 1.03 I (smoke index) .... 3360 A i= average deflection. B = average of initial and final deflections. X = average smoke density (percentage) = B X 100. T = total time (seconds). S = total smoke = X X T. W => wt. of sample (grams). I =a smoke index == S/W. APPLICATION OP METHOD 95 Table 28. — Smoke Index of Will County Briquets Volatilized to 39.3 Per Cent Volatile Matter at Temperature of 477°C. for 10 Minutes Test samples Time (seconds) 1 2 3 4 5 6 7 8 Galvanometer deflections ( mm.) 236 233 224 199 152 151 224 221 221 218 212 188 224 222 218 174 159 141 214 212 208 180 134 100 210 209 208 207 204 194 215 209 208 202 178 139 206 202 195 174 169 157 212 5 208 10 210 15 206 20 190 25 169 30 119 175 115 112 191 129 149 162 35 124 150 130 81 174 105 140 139 40 122 128 111 108 165 96 143 145 45 125 123 123 95 155 129 129 122 50 123 122 122 98 152 96 148 134 55 126 115 118 110 130 129 99 93 130 106 115 115 135 146 139 60 134 65 121 110 145 203 144 129 88 107 113 109 113 119 126 137 151 70 139 75 109 164 124 113 117 116 133 126 80 126 166 145 101 113 122 141 135 85 129 170 161 97 125 134 148 132 90 133 160 105 106 130 132 151 139 95 130 154 181 109 127 118 163 135 100 123 173 159 100 143 113 173 143 105 132 167 165 110 152 142 177 167 110 174 195 218 212 213 198 205 218 216 222 193 188 205 207 206 125 155 158 175 194 168 182 203 149 159 179 191 198 198 202 173 115 185 120 187 125 195 130 201 135 215 221 216 218 220 222 207 207 209 210 195 196 203 140 202 145 204 150 155 160 Total 5487 4771 5143 3863 3987 3921 3842 4987 Number readings 33 27 31 29 25 27 24 30 A 166.3 176.7 165.9 133.2 159.5 145.2 160.1 166.2 B 229.0 223.0 217.0 205.0 206.5 206.5 204.0 208.0 X 27.4 20.8 23.5 35.0 22.8 29.7 21.5 20.1 T 160 130 150 140 120 130 115 145 S 4384 1.24 3540 2704 1.44 1880 3525 1.05 3360 4900 1.44 3400 2736 1.35 2030 3861 1.27 3040 2473 1.43 1730 2915 W 1.53 I (smoke index) 1910 A =s average deflection. B = average of initial and final deflections. X = average smoke density (percentage) = B — A B X 100. T = total time (seconds). S — total smoke = X X T. W =s wt. of sample (grams). I =a smoke index = S/W. 96 SMOKE INDEX Table 29. — Smoke Index op Eight 1-cm. Cubes Cut from a Will County Briquet Volatilized to 35.8 Per Cent Volatile Matter at Temperature of 485°C. for 10 Minutes. Test samples Time (seconds) 1 2 3 4 5 6 7 8 Galvanometer deflections (mm.) 222 221 219 219 215 199 180 165 142 141 143 157 156 178 166 166 191 177 190 184 194 196 206 213 217 217 216 206 196 189 172 159 154 156 146 144 147 150 161 158 156 171 162 164 174 179 195 195 199 209 217 215 208 214 208 201 181 165 151 143 151 141 137 126 148 135 143 148 147 149 166 155 172 180 184 201 210 213 209 203 200 183 167 156 135 142 131 139 123 130 117 127 140 130 147 154 170 176 192 202 205 208 207 205'" 196 191 174 160 151 136 136 128 130 120 123 137 132 138 145 148 151 162 168 183 191 198 207 211 204 202 191 184 171 156 145 121 132 118 127 125 126 113 143 135 146 151 163 171 166 173 183 197 206 206 5 217 202 189 178 157 156 151 140 155 141 160 147 167 164 162 168 170 174 186 187 203 214 217 221 203 10 203 15 197 20 188 25 173 30 167 35 156 40 137 45 143 50 134 55 123 60 136 65 124 70 126 75 139 80 132 85 143 90 155 95 166 100 177 105 195 110 196 115 200 120 203 4448 25 177.9 221.5 19.7 120 2364 4435 24 184.8 219.0 15.6 115 1794 1.13 1590 4592 26 176.6 217.0 18.6 125 2325 1.58 1470 4379 26 168.4 212.5 20.8 125 2600 1.38 1880 4099 25 164.0 210.5 22.1 125 2652 1.34 1980 4017 25 160.7 207.0 22.4 125 2800 1.54 1820 4160 26 160.0 208.5 23.3 125 2913 1.52 1920 4122 Number readings 25 A 164.9 B 204.5 X 19.4 T 125 S 2425 w 1.44 1640 1.33 1820 A = average deflection. B =? average of initial and final deflections. X = average smoke density (percentage) = B — 'A X 100. T = total time (seconds). S = total smoke = X X T. W = wt. of sample (grams). I = smoke index = S/W. APPLICATION OF METHOD 97 Table 30.- -Smoke Index op Will County Briquets Volatilized to 31.1 Volatile Matter at Temperature op 505°C. for 10 Minutes Per Cent Test samples Time (seconds) 1 2 3 4 5 6 7 8 Galvanometer deflections (mm.) 231 229 223 224 220 217 211 210 207 215 206 224 216 217 218 221 222 225 228 229 231 226 222 213 205 194 186 187- 177 183 181 184 190 187 197 197 224 225 228 237 234 235 233 230 226 221 216 218 217 215 210 213 210 206 212 210 209 217 223 231 233 234 235 241 234 239 236 233 231 229 229 225 225 222 226 218 230 227 232 233 239" 242 241 240 233 234 226 229 221 222 219 213 220 209 210 204 208 205 211 202 179 184 189 194 212 232 243 255 249 250 246 249 237 242 234 231 226 240 233 226 232 224 223 219 223 231 233 249 252 240 239 236 233 233 233 226 227 225 222 227 227 218 217 215 218 213 226 232 235 239 237 5 233 10 15 20 25 30 231 236 232 232 224 35 223 40 223 45 221 50 219 55 60 214 216 65 212 70 216 75 213 80 218 85 213 90 213 95 217 100 225 105 233 110 233 115 234 120 237 Total 4393 20 219.7 230.0 4.48 95 426 1.62 263 3837 19 201.9 229.5 12.0 90 1080 1.39 777 5325 24 221.9 236.0 5.97 115 687 1.25 550 4391 19 231.1 241.5 4.31 95 409 1.50 273 5380 25 215.2 242.0 11.1 120 1332 1.53 871 5204 22 236.5 253.5 6.71 105 705 1.34 526 4781 21 227.7 239.5 4.93 100 493 1.52 324 5605 Number readings 25 A 224.2 B 237.0 X 5.40 T 125 S 675 W 1.42 I (smoke index) 475 A = average deflection. B = average of initial and final deflections. X^ average smoke density (percentage) = B X 100 T =2 total time (seconds). S = total smoke = X X T. W=!wt. of sample (grams), I =s smoke index = S/W. 98 SMOKE INDEX Table 31.- -Smoke Index of Will County Briquets Volatilized to 24.3 Per Cent Volatile Matter at Temperature of 515°C. for 10 Minutes Test samples Time (seconds) 1 2 3 4 5 6 7 8 Galvanometer deflections (mm.) 227 225 223 216 216 214 215 215 213 215 220 196 218 209 217 219 226 220 228 224 221 221 221 219 216 213 217 219 219 219 220 223 226 230 226 224 227 227 226 221 221 223 220 221 220 226 226 225 224 230 229 227 226 225 227 229 227 227 227 226 227 221 227 218 222 218 221 220 225 230 229 229 223 220 217 213 206 207 213 218 225 225 238 235 235 234 234 236 236 236 234 230 234 233 230 238 233 230 227 228 227 227 224 227 227 229 226 229 233 229 234 232 5 10 228 229 15 226 20 205 25 ao 208 213 35 40 45 50 55. . 215 221 224 234 229 60 65 70 229 233 75. . 228 80 . 85.. 90.. Total Number readings A 4132 19 217.5 227.5 4.39 90 395 1.61 245 3078 14 219.9 225.0 2.27 65 148 1.51 98 3817 17 224.5 230.0 2.39 80 191 1.50 127 4269 19 224.7 227.0 1.01 90 90.9 1.42 64 3083 14 220.2 229.0 3.84 75 288 1.57 183 3283 14 234.5 238.0 1.47 65 95.6 1.44 66 3430 15 228.7 233.5 2.06 70 144 1.38 104 3126 14 223.3 B 232.5 X 3.96 T S 65 257 W 1.08 238 A =: average deflection. B = average of initial and final deflections. X = average smoke density (percentage) = X 100 T = total time (seconds). S = total smoke = X X T. W = wt. of sample (grams). T =s smoke index = S/W. APPLICATION OF METHOD 99 Table 32. — Smoke Index or Will County Briquets Volatilized to 16.4 Per Cent Volatile Matter at Temperature of 535°C. for 10 Minutes Test samples Time (seconds) 1 2 3 4 Galvanometer deflections (mm.) 222 225 224 229 5 217 223 221 224 10 217 220 216 224 15 217 220 209 225 20 219 221 204 224 25 217 215 209 223 30 214 216 210 224 35 211 218 215 223 40 210 211 214 218 216 218 221 45 223 50 209 222 222 221 221 223 223 55 223 60 ..- 210 210 215 220 221 224 222 65 224 70 226 75 226 80 226 85 225 90 224 95 220 100 223 105 226 Total 3662 2857 2586 4928 17 215.4 13 219.8 12 215.5 22 A 224.0 B 221.5 224.5 223.5 227.5 X 2.75 2.09 3.58 1.54 T.... 80 60 55 105 S 220 125 197 162 W 1.38 1.30 1.41 0.86 I (smoke index) 159 96 140 188 A = average deflection. B = average of initial and final deflections. X = average smoke density (percentage) = B X 100 T =s total time ( seconds ) . S = total smoke = X X T. W=swt. of sample (grams), I =s smoke index = S/W. 100 SMOKE INDEX Franklin County briquets. — A similar series of tests was made for Franklin County coal. These smoke index results are shown in Tables 33, 34, 35, 36, and 37 for the products partially volatilized at temperatures of 250°, 450°, 465°, 480°, and 495°C., and containing 35.9, 33.1, 30.9, 28.5, and 23.6 per cent volatile matter, respectively. The same data are averaged and summarized in Table 7 (Part I of report). In figure 6 (Part I of report) the smoke index is plotted against the volatile matter content for Franklin County coal in a manner similar to that in figure 5 (Part I of report) for Will County coal. This curve likewise indicates a linear relationship between volatile matter content and smoke index for Franklin County coal. APPLICATION OF METHOD 101 Table 33. — Smoke Index or Nonvolatilized Franklin County Briquets Containing 35.9 Per Cent Volatile Matter Heated at Temperature of 250°C. for 10 Minutes Test samples Time (seconds) 1 2 3 4 5 6 7 8 Galvanometer deflections (mm.) 170 166 166 166 167 163 156 144 147 139 132 91 105 100 106 102 106 109 111 111 109 112 107 101 106 110 126 135 146 161 157 157 158 159 163 161 156 154 146 130 120 72 102 96 99 97 93 100 98 97 98 110 111 99 110 108 128 138 145 144 145 145 147 148 171 163 151 136 121 114 112 120 105 115 109 113 108 108 106 109 109 118 117 109 102 120 129 136 148 159 162 163 162 160 152 136 115 102 84 82 84 101 83 102 90 98 97 97 100 93 93 102 113 124 136 149 165 161 151 132 116 100 85 92 83 91 92 85 117 100 97 101 106 105 101 108 109 106 114 131 140 151 156 159 158 156 151 141 132 116 109 104 117 62 89 87 87 94 85 92 83 84 86 99 104 126 128 143 148 147 148 151 157 155 154 149 135 117 107 92 85 77 68 67 95 105 82 98 103 114 119 118 128 130 141 152 148 150 152 156 5 154 10. . 153 15 152 20 153 25 151 30 140 35 123 40 103 45 95 50 82 55 96 60 81 65 98 70. 91 75 98 80 94 85 102 90 102 95 107 100 105 105 107 110 a 96 115 105 120. . . . 134 125 137 130 144 135 140... 145 150 155 160 165 Total 4501 34 132.4 164.5 19.5 165 3218 3660 30 122.0 155.5 21.5 145 3118 1.30 2400 3370 27 124.8 166.5 25.0 130 3250 1.31 2480 2818 25 112.7 156.0 27.8 120 3336 1.25 2670 3095 27 114.6 160.5 28.6 130 3718 1.37 2710 3386 29 116.8 155.0 24.6 140 3444 1.27 2710 3198 27 118.4 154.5 23.4 130 3042 1.08 2820 3159 Number readings 27 A 117.0 B 150.0 X 22.0 T 130 S 2860 1.30 2480 1.17 I (smoke index) 2440 A = average deflection. B = average of initial and final deflections. X = average smoke density (percentage) = T^ total time (seconds). S = total smoke = X X T. W == wt. of sample (grams). I = smoke index = S/W. X 100 102 SMOKE INDEX Table 34.- -Smoke Index of Franklin County Briquets Volatilized to 33.1 Per Cent Volatile Matter at a Temperature of 450°C. for 10 Minutes Test samples Time (seconds) 1 2 3 4 5 6 7 Galvanometer deflections (mm ) 159 157 153 145 131 118 109 98 97 85 79 99 119 106 113 106 109 103 105 107 112 112 118 131 136 150 154 155 154 152 142 125 103 96 88 95 90 89 117 95 97 102 101 108 108 111 113 12/6 128 141 147 151 152 155 154 143 127 106 101 83 78 100 85 80 89 85 89 85 88 89 99 107 116 107 121 116 148 150 149 155 152 153 154 153 149 138 124 105 101 98 97 110 90 83 94 93 97 113 120 140 143 149 151 149 150 151 153 148 139 122 113 95 86 76 76 120 95 96 99 88 88 96 105 118 134 143 152 150 157 154 147 130 121 105 90 102 91 95 90 87 88 96 90 97 108 105 129 147 159 157 158 5 155 10 155 15 149 20 152 25 147 30 132 35 120 40 123 45. . 97 50. . 84 55. 127 60. . 105 65 119 70 119 75 123 80. . . 124 85 123 90 125 95. . 129 100. . 132 105. . 138 110.. 145 115. . 152 120 153 125 . 155 135 140 . Total Number readings A B X 3211 27 118.9 156.5 24.0 130 3120 1.47 2120 3086 26 118.7 153.5 22.7 125 2838 1.46 1940 3210 29 110.7 154.0 28.1 140 3934 1.58 2490 2952 24 123.0 152.5 19.3 115 2220 1.06 2090 2492 22 113.3 151.5 25.2 105 2646 1.15 2300 2545 22 115.7 157.0 26.3 105 2762 1.07 2580 3441 26 132.3 156.5 15.5 T 125 S 1938 w 1.20 1620 A = average deflection. B == average of initial and final deflections. X = average smoke density (percentage) = T =i total time (seconds). S = total smoke = X X T. W — wt. of sample (grams). I =3 smoke index = S/W. X 100 APPLICATION OF METHOD 103 Table 35. -Smoke Index op Franklin County Briquets Volatilized to 30.9 Per Cent Volatile Matter at a Temperature op 465°C. for 10 Minutes Test sampl 3S Time (seconds) 1 2 3 4 5 6 7 Galvanometer deflections (mm .) 160 158 156 154 144 141 120 111 98 95 149 123 134 133 132 138 141 146 152 159 160 156 151 143 137 128 124 113 115 103 92 97 87 96 93 95 104 90 120 140 132 143 147 151 159 161 159 156 158 154 152 142 129 116 100 104 96 96 126 120 116 116 116 126 134 134 157 159 156 150 140 126 121 106 101 103 95 90 91 86 93 95 104 108 112 125 136 146 151 155 158 159 160 159 159 158 155 142 136 125 98 95 91 95 121 109 116 122 129 128 136 142 152 158 159 161 160 159 159 156 149 142 130 103 123 113 117 113 119 109 121 117 119 133 133 146 157 158 5 153 10 146 15 139 20 127 25 115 30 101 35 93 40 95 45 82 50 100 55 135 60 111 65 121 70 118 75 116 80 125 85 125 90 122 95 133 100 140 105 145 110 154 115 156 120 Total. . 2744 20 137.2 159.5 14.0 95 1330 1.17 1140 3076 25 123.0 159.5 22.9 120 2748 1.50 1830 2868 22 130.4 159.0 18.0 105 1890 1.05 1800 3066 25 122.6 159.0 22.9 120 ' 2748 1.44 1910 3045 23 132.4 159.5 17.0 110 1870 1.13 1650 2939 22 133.6 159.0 16.0 105 1680 1.30 1290 3010 Number readings A 24 125.4 B 157.0 X 20.1 T 115 S 2312 W 1.45 I (smoke index) 1590 A = average deflection. B = average of initial and final deflections. X = average smoke density (percentage) = B — A X 100 T <=: total time (seconds). S = total smoke = X X T. W = wt. of sample (grams). I =: smoke index = S/W. 104 SMOKE INDEX Table 36. -Smoke Index of Franklin County Briquets Volatilized to 28.5 Per Cent Volatile Matter at a Temperature of 480°C. for 10 Minutes Test samples Time (seconds) 1 2 3 4 5 6 7 Galvanometer deflections (mm.) 156 153 154 156 154 153 149 143 136 129 126 127 117 122 111 116 140 149 140 147 151 153 157 157 154 155 154 152 155 152 149 144 137 129 126 119 115 107 111 98 108 113 106 110 113 120 128 137 148 156 154 165 164 164 161 155 147 147 126 134 124 122 121 125 140 128 141 144 151 160 163 162 161 159 154 146 139 140 121 127 122 119 116 110 115 112 118 118 122 132 140 151 157 158 159 160 163 161 159 160 161 157 150 137 129 124 125 135 127 133 131 131 134 138 142 144 144 149 153 158 160 162 161 159 152 140 126 113 110 103 99 113 109 110 119 128 123 129 138 140 152 161 158 160 5 157 10 158 15 154 20 146 25 137 30 126 35 119 40 115 45 115 50 105 55 107 60 124 65 123 70 116 75 125 80 128 85 130 90 133 95 140 100 147 105 151 110 156 115 125 135 Total 3239 23 140.8 156.5 10.0 110 1100 1.15 956 3707 28 132.4 155.5 14.9 135 2012 1.27 1580 2882 20 144.1 164.0 12.1 95 1150 1.23 934 3418 25 136.7 161.0 15.1 120 1812 1.38 1310 3605 25 144.2 161.5 10.7 120 1284 1.37 937 2905 22 132.0 160.0 17.5 105 1838 1.14 1610 3072 23 A 133.6 158.0 X 15.4 110 S 1694 1.39 I (smoke index) 1220 A = average deflection. B = average of initial and final deflections. X = average smoke density (percentage) = B X 100 T =3 total time (seconds). S = total smoke = X X T. W = wt. of sample (grams), I = smoke index = S/W. APPLICATION OF METHOD 105 Table 37. — Smoke Index of Franklin County Briquets Volatilized to 23.6 Per Cent Volatile Matter at a Temperature of 495°C. for 10 Minutes Test samples Time (seconds) 1 2 3 4 5 6 7 8 Galvanometer deflections (mm.) 160 159 157 156 155 156 155 153 153 155 153 153 152 154 156 158 158 160 160 156 156 157 157 157 152 156 158 158 159 160 161 161 159 156 159 159 156 157 153 153 155 155 155 154 157 156 155 159 161 161 159 155 154 148 147 145 143 138 141 145 144 144 160 157 158 158 158 160 159 159 160 159 159 158 161 159 157 158 157 156 154 149 146 143 142 142 138 138 14 1 139 142 138 148 147 162 159 169 165 166 168 166 166 164 162 159 159 168 161 158 161 165 168 168 5 167 10 166 15 165 20 163 25. , 156 30 155 35 150 40 145 45 141 50 137 55 146 60 147 65 141 70 157 159 161 146 75 167 80 161 85 165 90 168 95 100 . 2625 16 164.1 168.5 2.61 75 195.8 1.15 170 Total 2803 18 155.7 160.0 2.69 85 228.7 1.31 175 2047 13 157.5 160.5 1.87 60 112.2 1.30 86 2820 18 156.7 161.0 2.67 85 227.0 1.29 176 2401 16 150.1 161.0 6.77 80 541.6 1.31 414 2066 13 158.9 160.5 1.00 60 60.0 1.29 47 3118 21 148.5 159.0 6.60 100 660.0 1.47 449 2954 Number readings A 19 155.5 B 168.0 X 7.44 T 90 S 669.6 W 1.40 479 A = average deflection. B = average of initial and final deflections. X = average smoke density (percentage) = B X 100 T i= total time ( seconds ) . S — total smoke = X X T. "W = wt. of sample (grams) I =: smoke index = S/W. 106 SMOKE INDEX RELATIONSHIP BETWEEN SMOKE INDEX AND VOLATILE MATTER CONTENT OF BRIQUETS MADE BY IMPACT FROM PARTIALLY VOLATILIZED COALS For both naturally occurring and artificially reduced volatile matter con- tents of the bituminous coals investigated, an approximate linear relationship exists between the smoke index and the volatile matter content. The slope 6000 5000 4000 3000 2000 1000 • NATURALLY OCCURRING ILLINOIS COALS O NATURALLY OCCURRING W. VA . COALS © FRANKLIN COUNTY BRIQUETS <• WILL COUNTY BRIQUETS • ^. / o/ / y So y y y y y y / &/ y y y y y y 3 fo /O / f 20 30 VOLATILE MATTER ( PER CENT) 40 50 Fig. 16. — Effect of Amounts of Volatile Matter on the Smoke Index of Illinois and West Virginia Coals and Briquets Made From Franklin and Will County Coals. (Fig. 15, p. 90, and Figs. 5 and 6, Pt. I, pp. 27, 28.) of the straight line for natural coals differs radically, however, from that of coal processed by the method herein described, as shown by figure 16 which is a composite of figure 15 and figures 5 and 6 (Part I of report). RELATIONSHIP BETWEEN SMOKE INDEX AND VOLATILE MATTER CONTENT OF NATURAL BITUMINOUS COALS Eef erring again to figure 15, which shows a straight-line relationship between smoke index and volatile matter content for the natural coals investi- gated, it may be noted that the dotted extrapolation line intersects the axes at their zero value. In other words, there is an approximate direct propor- tionality between smoke content and volatile matter content for these coals. This seems to indicate that (with respect to its smoke content) the type of APPLICATION OF METHOD 107 volatile matter present in these various coals is practically identical, the amount of the smoke per gram of the volatile matter being practically the same for the bituminous coals investigated. However, it is well known that some coals possess a widely different smoke content from that of other coals containing the same percentage of volatile matter. It remains for future investigation to show whether or not a family of smoke index curves characterize coals having different botanical constitution. SMOKE INDICES OF BRIQUETS MADE BY IMPACT FROM PROCESSED ILLINOIS COALS COMPARED WITH THOSE MADE DIRECTLY FROM NATURAL COALS An examination of figure 16 shows a contrast in the rate of decrease of smoke index with volatile matter content for processed and natural coals. For example, in the instance of processed Will County coal (Fig. 5, Part I of report), the briquetted sample with a volatile matter content of 31.9 per cent (reduced from a natural volatile matter content of 43.9 per cent) has a smoke index of 250. Thus, processed coals from Will and Franklin coun- ties can be made which possess less than one-third and one-seventh, respectively, of the smoke index of that of a natural West Virginia coal, even though the latter has a lower percentage of volatile matter. IMPORTANCE OF ELIMINATING THE HIGH-SMOKE-INDEX FRACTION OF THE VOLATILE MATTER Eef erring again to figures 5 and 6 (Part I of report), which show a straight-line relationship between smoke index and volatile matter content for briquets of partially volatilized coals from Will and Franklin counties, it may be noted that the dotted extrapolated lines intersect the volatile matter axis at 29 per cent and 23 per cent, respectively. Thus the smoke index decreases far more rapidly than the volatile matter content. This seems to indicate that in the process herein described there is a fractionation of the volatile matter whereby the high-smoke-index fraction is liberated, whereas the low-smoke-index fraction is retained in the processed coal. Therefore for the purpose of obtaining a smokeless fuel for compaction into briquets from Illinois coals, it is essential only to apply heat sufficient to remove volatile matter which is driven off at comparatively low temperature. RELATIONSHIP BETWEEN TEMPERATURE AND TIME IN EFFECTING DIFFERENT AMOUNTS OF VOLATILIZATION This study consisted of: (a) the determination of the effect of volatiliza- tion temperature on the amount of volatile matter removed during a 10-minute volatilization period; and (b) the determination of time-temperature curve for 15 per cent volatile matter loss. 108 SMOKE INDEX Effect of volatilization temperature on amount of volatile matter removed. — The effect of the temperature of volatilization on the amount of volatile matter reduction was determined for both Will and Franklin County coals using a volatilization period of 10 minutes. Will County coal. — The effect of the temperature of volatilization, with a range from 350° to 530°C. (coal temperature), on the percentage of volatile matter in Will County coal, volatilized for a 10-minute period, is shown in figure 17 (Table 38). It may be noted from the figure that the volatile reduction starts at 420 °C, the amount of reduction increasing with tem- perature. Table 38. — Volatile Matter Content of Will County Coal as Affected by Various Volatilization Temperatures Maintained for 10 Minute Periods (Data for Fig. 17) Volatilization coal temperature (°C.) Oven temperature (°c.) Weight loss (per cent) Volatile matter (a) (per cent) 350 400 425 450 480 475 490 500 510 520 525 530 540 550 575 0.0 0.5 1.2 1.9 2.8 5.8 6.9 8.8 11.8 11.3 12.7 16.2 17.6 (h) 43.9 373 43.7 395 43.2 426 42.8 430 42.3 448 40.5 460 39.8 466 38.5 475 36.4 475 36.8 485 35.7 494 33.0 505 31.9 530 (&) (a) Percentage volatile matter calculated from experimental weight loss. (&) Weight loss could not be determined because no briquet was formed. Franklin County coal. — Similar results for briquets made from Franklin County coal, volatilized for a 10-minute period at a coal temperature ranging from 250° to 48£°C. are shown in figure 18 (Table 39). By extra- polation the initial temperature of volatile matter reduction appears to be about 410 °C. APPLICATION OF METHOD 109 50 h-40 Z ui O (£. UJ ^30 20 \ \ vo LATI LE K. IATTE R C ONTE NT F CO \L * X i "o^ O ' §10 350 360 370 380 390 400 410 420 430 440 450 460 470 480 490 500 510 520 TEMPERATURE (DEGREES C) Fig. 1 7. — Volatile Matter Content of Will County Coal as Affected by Various Volatilization Temperatures Maintained For 10-Minute Periods. 50 y-40 z Ul o a. UJ Q. -30 20 10 \ " V «T "o^. ^< > \ VOL UTILE MA rTER CON TENT OF COAL ^O C r^^ ^ % >. --.^ ^< 350 360 370 380 390 400 410 420 430 440 450 460 470 480 490 500 510 520 TEMPERATURE (DEGREES C) Fig. 18. — Volatile Matter Content of Franklin County Coal as Affected by Various Volatilization Temperatures Maintained For 10-Minute Periods. 110 SMOKE INDEX Table 39. — Volatile Matter Content of Franklin County Coal as Affected by Various Volatilization Temperatures Maintained for 10 Minute Periods (Data for Fig. 18) Volatilization coal temperature (°C.) Oven temperature (°c.) Weight loss (per cent) Volatile matter (a) (per cent) 250 275 460 480 500 520 540 0.0 4.3 7.3 10.3 16.1 17.2 35.9 425 33.0 440 30.9 455 28.6 470 23.6 482 22.6 (a) Percentage volatile matter calculated from experimental weight loss. TIME-TEMPERATURE CURVE FOR 1 5 PER CENT VOLATILE MATTER LOSS Table 38 shows the volatile matter content of Will County coal voktilized at various temperatures for various periods of time. As shown previously (p. 28), on "as received" basis, a 15 per cent reduction in volatile matter results in a smokeless coal, with a smoke index less than one-third that of a West Virginia coal. For a coal containing about 10 per cent moisture, 15 per cent reduction on "as received" basis is equivalent to between 16 and 17 per cent reduction on a dry basis. Thus an optimum volatile matter loss of 16 per cent, dry basis, reduces this Will County coal from 43.9 to about 34 per cent volatile matter, which was selected as the optimum volatile matter content for smokeless briquets from this coal. Figure 4 (Part I of report) is a graph of the time-temperature curve for such an optimum volatile matter loss. As expected, the period necessary for the desired degree of volatilization decreases rapidly with increasing temperature. DISCUSSION For Will County coal, prevolatilized for a 10 -minute period, volatilization starts at about 420° C, and the percentages of remaining volatile matter decrease linearly with temperature. For Franklin County Coal, volatilization starts at about 410 °C, or 10° less than that for Will County coal. The percentages of remaining volatile matter, likewise, decrease linearly with temperature for the same period, the rate of decrease being apparently the same as that for Will County coal. BIBLIOGRAPHY 1. Andreev, N. N., Determination of mean sizes of particles in dispersed systems with the aid of the photo-electric cell: Kolloid— Z, .57, 42-7 (1931). Determination of size of particles for which scattering stops and reflec- tion begins. 2. Arms, R. W., The ignition temperature of coal: University of Illinois. Eng. Exp. Sta. Bulletin No. 128 (1922). Defines ignition temperature as glow point; determined glow points for coals from six counties in Illinois. 3. Bean, R. D., Brown smoke density meter: Blast Furnace and Steel Plant, 21, 166-7 (March, 1933). Describes unit put out by Brown Instrument Company for measuring density of smoke in stacks. Density is recorded on Ringelmann scale. 4. Buchholz, M., Detection of gases, vapors or smoke. French Patent 656,160 (June 2, 1928). Presence and density of gases, vapors, or smoke is determined by their permeabilities to heat rays. 5. Bumgardner, H. E., Smoke density scales: Mech. Eng. 55, 200 (March, 1933). Gives revised Ringelmann scale taking into consideration the width of the smoke column. 6. Carter, W. A., Measurement of smoke density and sootfall: Power 74, 678-81 (1931). Discusses a number of visual and photo-electric methods of smoke density determinations, 7. Chicago Association of Commerce Committee of Investigation on Smoke Abatement and Electrification of Railway Terminals. Smoke Report (1914). Gives results of an extensive investigation of smoke throughout the Chicago area, covering (1) density determinations by use of Ringelmann scale, (2) physical and chemical analyses of solid and gaseous constituents of smoke passing through stacks, (3) density determinations by use of Hamler-Eddy Smoke Recorder. 8. Committee on Power Test Codes, Instruments and apparatus; smoke density determinations: Mech. Eng. 52, 999-1001 (Nov., 1930). Describes following smoke recording apparatus: Ringleman Chart, Bryan Donkins' Smoke Recorder, Eclipse Smoke Recorder, Roberts' Smoke Chart, Sawford's Smoke Density Meter, Umbrascope, "Westinghouse Smoke Indicator; all of which are for commercial use in stacks. 9. Dobson, G. M. B., British Patent 373,744 (April 28, 1931). Apparatus for measuring the intensity of smoke. 10. Freygang, W. H., British Patent 272,914 (June 17, 1926). Photo-electric cell and associated devices for giving warnings of sus- pended matter in gases. Ill 112 SMOKE INDEX 11. Fry, J. S., & Sons, Ltd., and Wrightson, F. B., British Patent 237,948 (April 7, 1924). Turbidimeter for recording the density of smoke, fog, etc. 12. Gray, R. W. W., The phenomena associated with finely divided particles in air: J. Soc. Chem. Ind. 48, (Trans.) 1071-2 (1929). Finds that large particles fall according to Stokes' law while smaller ones experience Brownian movements. 13. Owens, J. S., New instrument for measuring smoke emission: Engineer 156, 40 (1933). Compares smoke density with gray shades obtained from a rapidly revolving sector disc, in manner of Ringelmann scale. 14. Patterson, H. S., and Gray, R. W., On the densities of particles in smoke: J. Franklin Inst., 203, 605-6 (1927). Measurements of physical densities of metal particles in smoke show their densities to be different from those of the metal in its normal state. 15. Prince, C. E., Light sensitive cells; their use in the development of smoke prevention equipment: Electrician 110, 335 (Mar. 10, 1933). Describes use of photo-cell for measuring density of smoke in stack. Recorder is either a milli-ammeter graduated in Ringelmann units or a moving chart. 16. Sawford, F., Smoke-density meter: Mech. Eng. 49, 999-1004 (Sept. 1927). Describes smoke density measurement unit comprised of lamp, lens system, 3-inch pipe through stack (with orifice through which smoke passes), photo-cell, amplifying unit, recorder, etc. 17. Semikov, N. M., Apparatus for measuring the intensity of smoke discharged from boiler furnaces: Russian Patent 25,668 (Mar. 31, 1932). Light is passed through the smoke and onto a photo-electric cell which measures the fluctuations. 18. Shaw, J. F., and Hurley, T. F., Methods and standards of smoke measure- ment: Fuel Econ. Rev. 9, 97, 99-105 (1930). States determination of smoke in terms of suspended matter per unit weight of chimney gas is too complicated for plant use. Discusses Ringel- mann and similar methods for determining smoke densities. 19. Siemans and Halske, Apparatus for estimating smoke densities by absorption of light or heat radiation: British Patent 373,545 (Nov. 18, 1930). 20. Matson, P. D., and Kibler, A. S., The relation between obscuring power and particle number and size of screening smoke: J. Phys. Chem. 35, 1074-90 (1931). Tests are made to check as to whether obscuring power of smoke (chemical) was related to its particle number and size. 21. Wordley, W. A., Some experiments on the measurement of smoke under indus- trial conditions: Fuel Econ. Rev. 10, 89-92, 94-6 (1931). Instrument consists of a source of light passing through a tube across the smoke stack, a selenium cell at the other end of the tube, relay and signal bell. 22. Bailey, Compensated smoke recorder: Power Plant Eng. 38, 388 (Aug. 1934). Describes photo-electric smoke recorder which uses two cells^in parallel, one for detection, the other for compensation. Differential voltage is recorded, accurately representing the smoke density regardless of fluctuation in the system. BIBLIOGRAPHY 113 23. Electrical eye detects smoke: Power Plant Eng. 34, 447 (Apr. 1930). Describes unit developed by Zworykin at Westinghouse for measuring smoke density in stack. Lamp and cell are both mounted in tubes outside stack. Continuous recorder used. 24. Hannigan-McPhebson smoke indicator: Power Plant Eng. 35, 869 (Aug. 1931). Describes apparatus in which smoke is passed through a glass chamber with a light behind it. Visual observations are made. 25. Indicator for determining smoke density: Mech. Handling 19, 21 (Jan. 1932). Describes "Smoke meter" in which light is sent across the path of the smoke and the absorption measured visually by comparison with Ringel- mann scale. 26. Instrument for indicating and recording the density of smoke, liquids or dust: Power 69, 679-80 (1929). Describes photo-electric unit consisting of voltage regulator, lamp, photo-cell, amplifier, recording milliammeter, chart recorder. The detector is a length of pipe which admits a definite fraction of the total smoke. 27. Leads and Northrup smoke recorder: Power Plant Eng. 35, 868 (Aug. 1931). Describes use of a lamp and thermopile as a means of measuring the density of the smoke in a stack. 28. Photo-electric smoke recorder. Engineering 134, 165-6 (1932). Describes recorder made by Cambridge Instrument Company in which no light falls upon photo-cell unless smoke particles enter and act as secondary sources by diffraction. 29. Smoke detection. Engineer 155, 606-7 (1933). Describes photo-electric unit manufactured by Automatic Light Control, Ltd., for stack regulation. Uses selenium cell, washed windows in stack walls constant film of water. Uses temperature and voltage regulators. Unit makes a chart, sounds warning signals and lights lamp. 30. Encyclopedia Eritannica. 14 Ed., Vol. 20, p. 839. Discussion of smoke.