EQUILIBRIUM OF CARBON IN STEEL IN AN ATMOSPHERE OF HYDROGEN O'H'O'MWQM A METALLOGRAPHIC METHOD OF DETERMINING THE CARBON CONTENT IN STEEL TITC'SIS for the Degree. of M. S. MICHIGAN STATE COLLEGE Frank W. Littleford I939 1W. ‘5 III "a IIIIIIIIIIIIIIIII " 3‘; It a}! *1 ‘ I '~ 0' I. h? I "I‘d.“ - .5: ,. .g‘ x _ . * .3' L ‘1 ,‘Izta"§(‘:l‘t(‘ti Mull“. ” l I - ' . 1 ars- 5? to ”s MSU ’37" "I ‘ mounts it: 4.- RETURNING MATERIALS: PIace in book drop to remove this checkout from your record. FINES wiII be charged if book is returned after the date stamped below. EQUILIBRIUM OF CARBON IN STEEL IN AN ATMOSPHERE OF HYDROGEN * * * A METALLOGRAPHIC METHOD OF DETERMINING THE CARBON CONTENT IN STEEL Thesis submitted to the faculty of Iichigan State College In partial fulfillment of the requirements for the Degree of Master of Science Frank wl‘fltiisford Chemical Engineering Department 1959 THESIS TABLE OF CONTENTS Title Acknowledgement Abstract Survey of the Literature Apparatus and Materials Used Method of Procedure and Results Discussion A Netallographic Method of Determining the Carbon Content of Steel Index to Photomicrographs,charts, and graphs Bibliography 1214.83 page 10 14 29 51 59 4O ACKNOWLEDGEMENT The author wishes to dedicate this thesis to his Mother, who, through the years, has furnished a constant source of inspiration and sound advice. He also wishes to express his thanks to Professor H. E. Publow and the chemical engineering faculty for their invaluable assistance and suggestions. ABSTRACT One of the topics under discussion in the graduate seminar was Harrington's and Wood's article 'Critical points in iron-carbon allqys.‘ This was a rather exhaustive work and very carefully done, however, the thing that distinguishes it from many similar studies are the results obtained. These results differ widely from previous works along the same line so the question naturally arises as to why this should be so, and as previous investigators are in more general agree- ment the answer should lie in Mr. Harrington's and Ir. Wood's eXperiment. The major difference from this work and previous works lies in the preparation of the samples, Some of these samples were prepared by Professor E. D. Campbell by a method devised by him. This method is a rather novel method of preparation and not very widely used. It is based on the theory that steel samples of varying carbon content will reach an equilibrium in an atmosphere of hydrogen at a temperature of 1742QF. Therefore the object of this problem was to check Professor Campbell's theory and see if it could be duplicated. Also the ground- work for a new method of determining the carbon content by metallographic means was layed. This method is based on the fact that the percent of carbon in steel is proportional to the pearlitic area in a slowly cooled sample. Upon a photographic plate the pearlitic area appears white and the ferrite area black. Therefore if a light were placed on one side of the plate and a photoglectric cell on the other, the intensity of the reading from the cell would be prOportional to the area of pearlite, which is in turn prOportional to the carbon content. The results of the experiment on equilibrium of carbon in steel in an atmosphere of hydrogen confirmed the experiments run by Dr. Campbell at the University of lichigan. While the results of the photoelectric method of determining the carbon content in steel were not all that could be desired, they showed that the method has great possibilities pending further deve10p- lento SURVEY OF THE LITERATURE The only previous investigation on this subject, to the best of the authors knowledge, was done by Dr. E. D. Campbell at the University of’Michigan, the account of which is given in a paper published in “The Journal of the Iron and Steel Institute", No. II, 1925, vol. CVIII. To give clearly the background for this work it is better that I quote verbatim from this paper. 'It has been known for many years that the carburization of iron is readily affected by gaseous hydr-carbons. This reaction forms the basis of well-known commercial processes, such as case-hardening. Many authorities assume that, even where carburization is carried on by packing the iron or steel in solid carburizing compounds, the carbon actually entering the steel does so in the form of some gaseous com— pound of carbon. On the other hand, it has also been known for a good many years that hydrogen at temperature above 700°C reacts with steel, bringing about decarburization to a greater or less extent. This decarburization of steel by hydrogen forms the basis of a number of experiments carried on and reported to the institute some four years ago. There are here two reverse reactions involving both solid and gaseous phases. “In consideration of this fact, it should theoretically be possible to obtain certain conditions under which an equilibrium between these two reactions might be attained. The attainment of such an equilibrium would involve the existence of part of the carbon in a gaseous phase, and should not be confused with, those experiments which have been carried on during the past 40 years or more on the diffusion of carbon in steels where solid phases only are involved. The existence of such an equilibrium was explained on theoretical ground by M. Le Chatelier in his discussion of the author's paper to which reference has just been made. In his discussion M. Le Chatelier clearly indicates that on theoretical consideration both carburization and decarburization might take place simultaneously. Simultaneous carburization and decarburiza- tion by transfer of carbon in the form of gaseous compounds from hight low carbon metal has been experimentally shown in a qualitative way by J. H. Whitely.* “The object of the present investigation was to determine the amount of carbon transfered from a solid solution of carbides in iron to one of low concentration, when the materials were held in a still atmosphere of dry hydrogen at 950 C until enough time had elapsed to insure equilibrium.“ It can be readily seen from the above that the idea of equili- brium by means of hydrogen is not new and not original with Dr. Campbell. In a paper entitled 'A Laboratory Method for the Preparation of Small Steel Bars differing only in Carbon Content," in the American Society for Steel Treating of July - Dec. 1924, Dr. Campbell gives another summary of the literature which forms a basis of his investi- gation. From this I quote, 'More than a century ago Proust apparently recognized solid solution, for it was in 1806 he stated that white cast iron must be considered as a solidified solution of carbides of *Exp. on deox. of Steel withlig", Journ. of Jr. and Steel Inst., 1920 iron in iron. Again in 186$, A . Mathbssen considered steel and other alloys as solidified solutions and determined the effect of certain constituents on the specific resistance. It was about fifteen years after Mathiessen's work before metallurgical chemists began to use the solution theory as a means of interpreting the properties of steel and more than twenty-five years before physical chemists, led by Van't Hoff in 1890, adopted the idea of solid solution and began to develop it along chemical lines. During the last decade of the 19th century, the electrolytic dissociation theory of solution, first announced by Arrhenius in 1887, was made the subject of many researches all over the world and very much extended in its application. During the same period the so-called solution theory of steel was being used to interpret the properties of steel, as influenced by variations in composition and concentration of the solutes dissolved in the solvent iron, but no effort was made to develop a general theory of solution which would apply to steel, although it was obvious that the electrolytic dissociation theory could not apply to a solution in which the solvent was a good conductor of elec- tricity. The solution theory of steel, that is, the assumption that hardened steel is a solution in the strict sense of the word, has formed a back- ground for most of the researdiwork on steel carried on in this laboratory since 1891. During the first ten years of this period, experimental evidence was developed which indicated that the carbides of iron, as well as those of other metals, should, in the light of organic chemistry, be considered as metallo-carbons, that is, derivatives of hydrocarbons in which the hydrogen has been completely replaced by metals. Evidence was also given to show that the molecular constitution of the carbides of iron is much.more complex than is indicated by the simple formula F650, and that the molecular weight of the carbides in steel is influenced by the total carbon concentration and the heat treatment. It is now generally recognized that carbides are the most important solutes in steel and that the mechanical, electrical, thermal, and magnetic properties depend more on the composition and concentration of the carbides in solution than on any other factor? Dr. Campbell goes on to tell of the apparatus and materials, used in his experiments so, as that is not essential, I shall skip to the section which gives the theory of the equilibrium of carbon in hydrogen. "The principle involved in this method is that if two solutions differing in carbon concentration, held in a still atmosphere of pure dry hydrogen at a temperature of about 950°C, that is, well above the Ac; point of pure iron, hydrogen will tend to combine with carbon of the dissociated carbides, removing it in the form of hydrocarbons. These compounds coming into contact with the solution of low carbon concen- tration, will react, the hydrocarbons giving up their carbon with liberation of hydrogen. There are thus two reversable reactions and theoretically then should be an equilibrium established between the carbide concentrations of the two solid solutions and the hydrocarbon concentration in the gas. Since the reaction between gaseous hydrocarbons takes place at the surface of the low carbon solution, and since the solutions are solid, time will be required for the resulting solution to become homogeneous throughout the entire cross-section, as diffusion in such solid solutions must necessarily proceed much more slowly than in liquid solutions. APPARATUS AND MATERIALS USED The equilibrium experiment was carried on in a modified carbon train apparatus. Hydrogen was generated by electrolysis of water, the direct current for which was taken from the regular direct current source and was regulated by a light bank consisting of two fifty watt bulbs. After some experimentation it was found that this arrange- ment was the most satisfactory source of current supply. As the direct current was cut off during the night several (6) dry cells were substituted during these hours. These did not supply enough current to cause the hydrogen to bubble through the setup but were enough to keep up a steady pressure. This was done during the first three trials, but in the last two runs the rubber tubing on each side of the furnace was clamped and the hydrogen generator shut off during the night. This made no difference in the results as far as could be noticed. In fig. 1, a diagram of the apparatus, parts B and C are the hydrogen generator. The oxygen was allowed to bubble off into bottle A, of fig. 1. This bottle contained water, the height of which was equal to the pressure on the hydrogen side of the apparatus. On the hydrogen side the gas flowed first into bottle D, which was simply an overflow bottle placed there in case a sudden leak on that side should cause the water to flow into the apparatus. This happened once in the beginning of the trials so it was placed there because of eXper- ience rather than foresight. If it weren't there the water would be drawn into drying bottle E which contains concentrated H2804, which in turn would go into U-tube F and react with the CaClg. This not only stops the experiment but is rather difficult to clean up and much time 10 Figure 1- bottle containing water, oxygen outlet anode, oxygen evolved cathode, hydrogen evolved overflow bottle wash bottle, sulfuric acid U-tube, calcium chloride wash bottle, askerite constant pressure valve rheostat controlling furnace silicon tube furnace -».."\4 -a. 1-1'" ‘I-oJ f‘“ P ,4 « .,. 1.. _- .q I t, S . . .. \J . .I ti Iii: - J. .‘IY‘I ' 7 Id _' o", 3 -' ,- —4 FGf’T . “4'. Q . J —. I " n, .J F. "1 LI ‘- DILGRAN OF cannon DIFFUSION APPARATUS ll is wasted. The wash bottle G contains a commercial drying preparation known as askerite. At point H, in fig. 1, is a T-section of glass tubing. The vertical stem of the T runs down into a bottle containing just enough H2804 to force the hydrogen thru the heated furnace, if it had been allowed tg.gg_§2, As the exit to the silicon tube in the furnace was clamped shut the hydrogen bubbled off at the T-section and thus a con- stant hydrogen pressure was maintained in the furnace. In fig. 1, part I represents the rheostat which was used to control the furnace and keep it at a constant temperature. This rheostat was adjusted by the use of a thermocouple and potentiometer and was set at 1800°F (982°C), which is over 950°C used by Dr. Campbell. This was done to allow for any variation in current, which in using alternating current could easily happen. Part J, of fig. 1, is the silicon tube and part K the electric fUrnace. All connection, rubber tubing to glass, were bound with capper wire and all other sources of gas escape was painted with rubber cement. The materials used in this experiment were 5/8" sq. steel bars furnished by the Mechanical Engineering Department and the type used in their heat-treating courses. For these the author is indebted to Mr. Cockrell of that department. It was felt that using a commerical steel would possibly give results more useful to the average person than if specially prepared steels had beenused. The mill analysis of these steels are as follows: 12 Figure 2 Photograph of equilibrium apparatus S smug-:31"! sudsmsqqs mui1dIIIUps lo dqeagododq 13 LC (low carbon-inherently fine grained) - Carbon .18 Mn. .47 P .015 S .051 MC (medium carbon) — Carbon .57 Mn. .78 P .016 S .021 HO (high carbon-inherently coarse grained) - Carbon .88 Mn. .51 P .015 S .012 Si .22 The symbols LC, MC, and HC will be used hereafter in this paper to designate the low carbon, medium carbon, and the high carbon steels, respectively. These symbols plus a numeral will identify the steel and tell from which trial it came. For instance, HC 5 would indicate that it was the result of high carbon steel after trial three (5). Method of Procedure and Results of each Trial. The first thing to check was the uniformity of the steel bars to be used in the experiment. This was done by taking photomicrographs of sections taken from each end of the sample to be run in the experi- ment. The result of this showed that the steel bars to be used were homogenous throughout and are shown in figures 5 and 4. The next thing to be done was to grind each sample to take the paint off and to polish them as diffusion takes place more readily l4 -- - - .._--_,.H4 -A._ l - .. -~ 4.s~ “ -A'<- .- ~ — -<‘{~H‘l‘ ..‘ .1. “he--. i‘ ‘”_,..“l -<,—.--.. V-- Figure 5 H C as received .88 carbon 100x vFigure 4 L-C as received .18 carbon 100x 6 euuglq bevieseu as O H nodvso 88. xOCI a evugii bevisoeu as O J nodmso 81. xCCl . . . its": 9-“ up,» at , f1, . . 3.5 '. :37 z’.‘ ‘. '_ rail'iii‘sguifcfi ‘ 3:5 fl": r};3?"\% 3w. .. . ~t-I'..' 4' ‘vn- "-'.- a 'J ‘. ‘ H....<“:‘<1“§'Q,:P.yfinjga .r‘fz'gnv-F. ,rg: my.” .. v 1“" z ,, “who (.0. ,4 Eng“ J“: ~ . . ‘5‘“;va A‘éfl'fi”;""“ '4- -. .~ «v.1 Mn - r. -. .~-- *4 r 1“ :3.".“-".: .33“, a. (i ~L‘: .2 . u- . Tam fi'“"-’~.: , . me a H‘.‘ 15 through a smooth surface. After this was done the samples were care- fully wiped and accurately weighed to the fourth decimal place on a delicate balance. The samples were then placed in the furnace, approximately in the middle, the rubber st0pper put in place with rubber cement, and the hydrogen generator started. For several hours the hydrogen was allowed to pass through the furnace to sweep out all oxygen before the furnace was plugged in. After the furnace was started the exit to the Opposite end of the silicon tube was clamped and the hydrogen allowed to bubble off through the T-section, at part H of fig. 1. The above is the general procedure followed, but as each trial was slightly different they will have to be described separately. Trial 1. In this experiment the hydrogen was allowed to bubble slowly through the silicon tube and no T-section was used. The direction of flow of the hydrogen was from high to low carbon. The samples were placed and to end %' apart in an alundum boat. Seven days was allowed for the experiment. The data for the experiment is as follows: Sample Wt. before Wt. after Difference L0 16.4119 gm. 16.4057 gm. -.0092 H0 11.220 gm. 11.1100 gm. 33320 Totals 55.6479 gm. 55.5157 gm. -.1522 Figuring the carbon content by weight, assuming the mill analysis to be correct, the samples were decarburized to a great extent. Photo— micrographs show this to be so and can be seen in figs. 5 and 6. 16 - -.-n w... .-.,~.._—~ _ — o..-u-.... - . lFigure 5 H C 1 av. .13 carbon lOOx Figure 6 .Lc'l av. .14 carbon *3 _100x :1 >- a .r u’ ‘1 .' Q” , .9“ \Q ' 7 i \ -Vr-~——. 'vn‘ ~-n- ‘ ‘-'—-.--~ a etugli I O H nodtfio 51. .vs '- r I AC4 -.L B o @1031? I O J noddso £1. .ve KC(i l7 Upon microsc0pic examination of the entire cross-section of each sample it was seen that the pearlite seemed to be uniformly distributed, which shows that this method is distinct from other methods of decarburization where the outside portions of the sample are more decarburized than the center section. A comparison of the carbon content of these samples arrived at by different methods can be seen in fig. 17. Trial 2. Because the preceding trial had resulted in almost complete decarburization the constant pressure arrangement was used in this trial so that the hydrogen would not flow through the furnace and tube. A new alundum boat was used, but not weighed before the experiment. Six days was allowed for this trial. The data is as follows: Sample Wt. before Wt. after Difference LC 7.9860 gm. 7.9781 gm. -.OO79 H0 M219 gm. L220. gm. 2M Totals 15.9850 gm. 15.9151 gm. -.O699 By weight the carbon content of the two samples was even lower than in the preceding experiment, which is substantiated by the different methods of analysis used. It was noticed that the new alundum boat was very discolored, eSpecially where the samples rested, and in the light of the experiment to follow it is to be assumed that the carbon went into the boat. The photomicrographs of these samples are to be seen in figs. 7 and 8, and the analyses in fig. 17. 18 ‘he—ds-qu .o- 7. .43. Figure 7 H C 2 av. .11 carbon 130x Figure 8 . L C 2 at: .08 carbon 100x u. w_‘—.‘-—. ‘fi'fih.-ffl v 'x .. - .— Y equal? 3 O H morass II. .vs xfiii 8 Sinai? S O J nodwso 80. .vs XCOI Trial 5. Because of the suspicion that the alundum boat had absorbed the carbon in the second trial, this possibility of error was eliminated in this run as no boat was used. This created a new problem of how to get the samples into the silicon tube and be sure that they were in the middle and also neither too close together nor too far apart. In Dr. Campbell's experiment he bound the samples with iron wire, weighing it before and after the experiment. As the temperature at which the experiment is run, 989°C, is below that of the melting point of c0pper, 108500, by some 105°C, I thought it would be safe to bind the samples with copper wire. However, after they were taken out, in five days, it was found that the cOpper had fused to the samples no that they could not be accurately weighed. A photomicro- graph was taken at the point of fusion and shows the influx of the copper into the steel. The picture shows that small globules of copper penetrated quite some distance into the steel and also shows why a steel containing c0pper could be easier machining. The photos of the results of this experiment are to be seen in fig: 9, 10, and 16. The later shows the infiltration of copper. The data is as follows: Sample Wt. before LC 12.1724 gm. H0 12.5456 gm. The photomicrographs revel that the carbon content is approximately the same in the two samples and other methods prove it. The H0 5 photo shows a decided widmanstatten structure while the LC 5 has a more normal structure. 20 s . Vifl..... ‘ . . -vfi-I- «H..- ..... -.ro.o‘-.-u. ~mv-s m ‘\ C 6 ;‘ .‘\ l- .‘ 1’ ‘0. ‘ ‘ c 0 ,. .1. A-Cfi-‘Q‘ o-.—. figure 9 was : is}. .42 carbon 1100: ? Figure 10 ”L c 5 _av.‘.52 carbon "100: f i 1.- d\ .. ‘ Q " .- ., 1 a '.‘ ‘5 q - .1. 2-». ‘ 1.. ‘5‘" _ -... ...A-- r-- -.. e exugi? nodtec 9;. .vs x001 or emugza a o J median 35. .vs _ ... .. ., ... , .9 . .... ,r... . . - 4 ._ . w fill» ....» p ~ ‘I f 7.‘ F111 . d. v a; y \ p r .. , D 5 . .V I I f l ; /. ... . . . . . I. At 4 \ VI. 0 N a . . . . \c ..., "rJ—qpn . . ,a). . . K .. . I. a. . o l. l ..4 as a o ..In V n‘ A an .1 v . .... s e a . .1 . _ r .\ x t. l ,u L .. . u l \ . . . y 0 _ .9 .w r . . le . . . . . I ' {fir ”raga... .2... guns. .2 ..4» . ... “I 2] Trial 4. The foregoing run gave evidence that the set-up needed no alterations so again samples were cut and weighed out. This time no binding of any sort was used and the two samples merely passed to the center of the silicon tube, leaving approximately i of an inch between. The data is as follows: Sample Wt. before Wt. after Difference LC 9.7555 gm. 9.7570 gm. #.o255 HG 2:25.12 an. 2.2229. gm. 29.220 Totals 19.0945 gm. 19.0790 gm. -.0155 The photomicrographs of these samples, figs. 11 and 12, indicate that they approximately have the same carbon content as do all the methods of analysis used. Again the H0 sample showed a widmanstatten as compared to the more normal structure of the LC sample. The total loss of carbon was 15.5%. Trial 5. If a state of equilibrium really exists between the steel samples, then distance between the samples should make little difference in the result. Therefore in this trial the samples were placed a greater distance apart and three samples instead of the regular two were used. Because of the greater total mass a greater length of time should theoretically be necessary, so eight days was allowed for equilibrium to be attained. 22 Si emugifi b 0 J nodmso 9%. .ve xCCl The data for this experiment Sample LC HC Totals Wt. before 17.5754 sm- l6.6216 gm. 50.5146 gm. is as follows: Wt. after 17.4242 gm. 16.5522 5110 50.4957 gm. Difference /.0508 gm. '00] 52 $1. -.0189 gm. Again the HC sample showed the characteristic widmanstatten structure as in the preceding experiments and the medium and low carbon samples were more normal in this regard. The loss of .0189 gm. of carbon is 7.9% of the total carbon in the samples. The photomicrographs of this experiment are figs. 15, 14 and 15. 24 ' --'-- l “'c c- 0-.- n—n-u—nu— 81 equal? a O H nodvso Yb. .vs xCCl bi amuglq ash nodwso as. .vs xLCl 25 ... -.-—‘.v ,7 L..-§.7 - ~17» ... W‘AH‘ ..- Figure 15 L C 5 av. .45 carbon 100x 81 95'313 a D J nodmeo B>. .Vfi xCLl 26 aecqoo lo xulinl galvods A xttl 27 Figure 17 Chart showing comparison of carbon analysis by different methods Y1 amnglq slaviana medias lo noslieqmoo gnlwode JIBdO abodiem eueaeiilb vd Comparison of Results Of Carbon Analysis by Different‘lethods % of Carbon Chemically by weight ' by squares "Pearlimeter" H C as rec. spheroidal spheroidal L C as rec. .18* .15 .14 H C 1 .17 .12 .12 L C 1 .12 .12 .19 H C 2 . .11 .11 .12 L C 2 . .08 .08 .07 H C 5 - .44 .40 L C 5 .52 .55 H C 4 .47 .48 .41 L C 4 .42 .58 .51 H C 5 .45 .45 .57 u c 5 .57* $3.. .40 .40 .56 L C 5 .47 .45 .42 28 DISCUSSION There can be little doubt that the results check and support the conclusions made by Dr. Campbell in his previous experiments on this subject. However, the samples in the last two trials did not come to an exact equilibrium, by any method of analysis. Possibly the time allowed in trials 5 and 4 wasn't enough, but the real reason is more likely to be found elsewhere. One of the more likely reasons is that the temperature of the furnace wasn't precisely constant throughout the entire experiment, or even during one trial. It would take but small differences in temperature, at that relatively high temperature, to cause great differences in the solubility of carbon needed to bring about equili- brium. There is also the possibility that inhibitors used in the manufacture of the steel could cause differences in carbon content of the resulting samples because of their influence on the solubility of carbon in steel. The other constituents such as phosphorus, sulphur, and silicon undoubtedly affect the equilibrium in some such manner. The experiment also shows that grain size is a function of time and temperature, and that inherently fine or coarse—grained steels will both arrive at a constant maximum sigs, depending on the temperature and the time at that temperature. While 1800°F isn't really over- heating the long time allowed for the experiment caused very large grains to form. The fact that the HC sample usually resulted in a widmanstatten structure was pointed out. This is normal structure for a higher carbon steel which has been overheated. While the temperature wasn't too high the time allowed was considerable so the net result was the same as overheating. In micrographs of HC 1, LC 1, HC 2, LC 2, figs. 5, 6, 7, 8 will be seen numerous small black spots. These are definitely not polish- ing pits but are, in my estimation, small isolated spots of pearlite separating from larger pearlite areas, or are pearlite areas which have decreased in carbon content till they have become that size. The latter isn't very probable. The former suggests a movement of carbon inside the sample, probably in the form of some hydrocarbon. Upon cooling it would seem that the hydrocarbon gave up the carbon and hydrogen was freed. 50 A METALLOGRAPHIC METHOD OF DETERMINING THE CARBON CONTENT OF STEEL A METALLOGRAPHIC METHOD OF DETERMINING THE CARBON CONTENT OF STEEL This is an entirely new method, and like all new laboratory methods it will require a great deal of experimentation before it can be used with any great amount of accuracy or confidence. The principle involved is very simple, that the light intensely projected through a photographic plate is directly proportional to the area of pearlite of the plate. The percent of carbon is then computed by comparing the millivolt reading of the unknown with a reading from a 100% pearlite sample. Then, 85 percent of that is the percentage of carbon. The botany department have employed the same principle in the measurement of the areas of leaves in their work. However, this is comparatively simple as they work with either black or white areas entirely. A diagram of the apparatus can be seen in fig. 18, and a picture of it in fig. 19. The instrument was built and designed by the author. While it is rather crude in appearance the results seem satisfactory. Starting on the left, part A is a sliding rod which controls the panel upon which the light socket is located. By means of this the intensity of light may be regulated. This was necessary in the first instrument but may be omitted in the future as the intensity giving the best results will be known. B, in fig. 18, is the light source, which was a 50 watt bulb. An inclined light diffusing board, part C, was included to both diffuse the light and to turn it up towards the plate and the electronic cell. A ground glass was also used for further diffusion, 51 Figure 18 A- red to control light intensity 8— 50 watt light bulb C inclined light reflector D- ground glass E— cone F— photoelectric cell scale i": l” 81 Sisal? x+isnedni ddgii Iomduoo 03 Dow -A dIud Jfigll stw C8 -8 ioioellei Jdpii benllonl —D ssslg buncmg -C enoo -3 lies oliJoeIeoJodq -1 ”1 ="fi elnoa S\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\§8 \\‘ o. V////////A magma ogggomm ho Iguana 52 part D. A polished tin cone, part E, was used to collect the light coming through the plate. The photoelectric C311 was mounted on the top of the cone on the turned edge of the cone and held there by elastic bands stretching up over the cell. Velvet was used to make the instrument light tight around the base of the cone and on the turned edges supporting the cell. The plate holder fitted into a slide made for it under the cone in such a way that the plate always came to the middle of the cone. The cone itself was always placed on the base the same way each time so as to standardize any errors caused by the construction of the cone. Perhaps a good way of presenting the difficulties involved in this method is to list them. They are: A. In the preparation of the sample. (1) Polishing pits-cause light areas on a plate (2) The intensity of etch (a) Concentration of etching reagent (b) Time in etching solution. B. In taking the photomicrograph. (1) Time of exposure (a) Iris and diaphram openings on camera (2) Fluctuation in current causing fluctuations on intensity of light (5) Differing intensity of light caused by scattering of arc lamp, if it is used. (4) The correct exposure time will vary with the amount of pearlite. (5) Amount of magnification. 55 C. In developing the plats. (1) Possibility of fogging the plate (2) Concentration and age of developer (5) Type of developer - normal or contrast (4) Time in deve10per (5) Acidity of hypo (6) Thoroughness of the wash D. In using the QPearlimeter' (1) Variation of current affecting intensity of light. (2) Possibility of light leakage around base of cone or cell. All of these difficulties can be eliminated by careful control, but until that time the method will be still in the development stage. It can be readily seen that the greatest difficulties lie not in the apparatus itself but in the plate and its preparation. The most important thing to standardize is the etch. The author was not quite so aware of this until towards the end of the term when he tried to duplicate readings of .9 carbon steel. The steel sample was photographed and then repolished and re-etched. The two plates were handled identically with the exception of the etch, and the two readings were 11.75 and 18.8 millivolts. Thus it can be seen how vitally the etch affects the method. The exposure time is equally important, in the author's opinion. However this is more commonly controlled in everyday practice so it presents no new problems. The effect of the concentration of the developer is not known to the author but it seems reasonable that the time of development should 54 Figure19 4 Photograph of photoelectric apparatus Elaiuglq \ sudsthqs slidselsoiodq lo dqswgododq ‘1 55 be increased as the developer is used. In commercial deve10pers, such as D-76 and cepecially in fine grain developers, a 10% increase in time should be used after development of a certain number ofiilms. This should apply to the development of metallographic plates equally as well. A contrast developer, D-8, or D-9, was chosen for this work because it would darken the ferrite areas on the plates and decrease the possibility of transmission of light through the ferrite-area. The error caused by polishing pits and grain boundaries, while probably very slight, should be investigated and determined. The are lamp should not be used because it is too susceptable to fluctuations in current and because the carbons themselves do not burn steadily. The best source of light is the mazda lamp. This takes longer but any error in timing is proportionately less. It is hardly necessary to point out that the same magnification should be used each time. A magnification which will give a plate containing a representative area should be choosen and then used thereafter. A 100 x was chosen for all the work done by the author. Possibly the etch has more effect on the totally pearlite sample than on any other. The resulting photomicrographs may be seen in figs. 20, 21, 22 and 25. The developing of the plate, printing time, and grade of paper was the same for all. The exposure times were the same, therefore only the depth of etch was the determining factor. In fig. 29 is shown the limits of accuracy as determined by the author. This with further experimentation will b narrowed down and then the method will compare favorably with any other, in comparison of results. n- —- —-— —.---fi \Iigwnsjgg .90 carbon‘ _ 'I' 7 v .-_ " in '7... it. '. V .' ' ' O‘,’ 'i? ’5 ~ ' ‘ 16.8‘Itkaae1tl’ii ' fans {'4 \_ . . . ~‘.-' ‘ ft ."'.' . .- .1 ’- ski" \ . . . I' I 1001"vw-~-w,~ r... m . w a. . a. . ..' .. 'I O 3. ‘1 i a ,. ~1.‘ . '»\(.‘ ‘* 5' , ..- :.'.\‘.v w . -,. . I l _u\\ z . “a T - ~. ~ . V‘w‘u ‘ .. . ST , -. - ‘ . .\ , _ V ‘ ... " ‘. s ‘. “i. -. ... V ,. u . . ‘5 i '6 ...?” V, I. . 5. lo. ’. ’a’a- . " ' \ D O - - ‘ .. irignre 22' .90'carbon is.25 nillivolts‘ 100x,fl-- ' ' Figure 25 190 carbon _4.0 sillivolts , 100x 12 51331? 03 sungli nodaao C6. nodtso CG. aalovfliim 6?.11 sJIov-lim 8.8I xQLI thI 39 sangiH SR eaugiq; medias CE. . medias CG. silovlillm 0.9 sJIoleilm 3.8 xCCI . x001 3/. 1‘ up: ‘2, ‘ l e .'\- ' a‘ {h ' a . l‘w 12¢»- a "v‘t‘ - 9 J 57 Figure 24 graph showing limits of accuracy 43 equal? 5-..- {ostmcss lo aJImII galwods dqsmg . W m _ ._-..._ .— 58 INDEX TO PHOTOMICROGRAPHS, DIAGRAM82 GRAPHS Diagram of Carbon Equilibrium Apparatus Picture ” " ” " H C and L C as received — H C 1 and L C 1 — H C 2 and L C 2 — A —— ~ H C 5 and L C 5 — A H C 4 and L C 4 ~—- -— — H C 5 and I C 5 A L C 5 _ A __ L C 5 + Cu ------ --- Charta of Results — ~— Diagram of Photo-electric apparatus ~ Picture ” " " — 4 —— Comparison of .9 carbon steels —— Graph showing limits of accuracy ——4 53 ll 15 15 17 19 21 25 25 26 27 57 59 BIBLIOGRAPHY (1) American Society for Steel Treating - July - Dec. 1924 (2) The Journal of the Iron and Steel Institute 40 «Two. in. gistzlzfr. . . . 15...... _., aer......ifi.§é..s§ . . 4. . c . 1 v lass??? "TI" Ill 1'41 WW W11 1m WWI 93103