THE PROPERTIES OF MANGANESE STEELS CONTAINING ONE TO FIVE PER CENT MANGANESE HARRY THEODORE WARREN THESIS FOK THE I ) E G R K E O F B A G HELO R O F SCI E NGE IN CHEMISTRY COLLEGE OF LIBERAL ARTS AND SCIENCES UNIVERSITY OF ILLINOIS 1 922 ) 322 YY2 5 UNIVERSITY OF ILLINOIS M ay 31 j iq2 THIS IS TO CERTIFY THAT THE THESIS PREPARED UNDER MY SUPERVISION BY Harry Theodore Warren ENTITLED. Tiie_. Exopua r_t i Slq _ _Qf— Mangiar-e aa - -Sta.e la _ -Cent ai nlng. _ One. _ to — Five Per Cent Manganese" IS APPROVED BY ME AS FULFILLING THIS PART OF THE REQUIREMENTS FOR THE DEGREE OF Ba ch el o r _ of _ _S c_i_e n_c e _ in _ _C h em i_s try. 500286 Digitized by the Internet Archive in 2015 https://archive.org/details/propertiesofmangOOwarr HISTORICAL The first work on manganese steels was apparently con- fined to those steels containing less than two and one-half per cent manganese. This was carried on in France by the Terre Noire Company, which also made high-grade ferromanganese for the manufacture of these steels. Sir Robert Hadfield first produced what is now known as manganese steel. In 1838 he pub- lished a paper on the subject, summarizing his previous papers. He made steels containing from nine to fifteen per cent manga- nese, and devised a satisfactory heat treatment for them. He found that quenching in cold water from 900-1100° C. (lemon- yellow) produced steels of great toughness. One steel, contain- ing 13*75 per cent manganese and 0.85 per cent carbon, when quenched thus, had a tensile strength of 145,000 pounds per square inch, and an elongation of 50.7 per cent in eight inches. Though too hard for impression by a drill press, the steel could be dented with a hammer. Comparatively small deformations produced a permanent set. It is noteworthy that Hadfield* s steels were all rather high in carbon, some containing as much as 1.5 per cent. He could not make low-carbon steels with the ferromanganese available. The manganese 3teels containing less than three per cent, are, according to Sauveur, pearlitic, while further additions of manganese and carbon produce martensite, and finally austen- ite. Hoyt states that the structure of the intermediate steels - 2 - is troostite, instead of martensite, when the carbon content is greater than 0.30 per cent. The oearlitic manganese steels ( 1 to 3 per cent manganese) are generally ignored by manufac- turers, for two reasons: first, the belief that these steels are too brittle, and second, the difficulty of manufacture of low-carbon manganese steels. The belief that these steels are brittle is the result of Hadfield’s statement that they are hopelessly so. Since this time, however, some metallurgists, notably G-uillet, have found that this applies only to manganese steels of high carbon content cooled rapidly. Low-carbon pearl- itic manganese steels, slowly cooled, are not at all brittle. Until quite recently, it has been difficult to manufacture such steels, because the only f erromanganese available was that made in the blast furnace. This f erromanganese has a high car- bon content. Nov;, however, f erromanganese, low in carbon, is successfully manufactured in the electric furnace or by the thermit process. The manufacture of these steels should there- fore be given more attention. Another possibility worthy of investigation is that of casehardening articles made of these steels, to produce austenitic cases while the cores remain pearlitic. Manganese forms a continuous series of solid solutions with iron, and this tendency is not lost even in the presence of carbon, though manganese does form a carbide, Mn^C. At high temperatures, this carbide forms a continuous series of solid solutions with manganese. The exact chemical composition of the carbide present in manganese steels is undetermined. Arnold -3- and Read made an exhaustive study of the chemical relations between iron, manganese , and carbon, using a series of steels of the same carbon content, between 0.3 and 1.0 per cent, and manganese contents varying from 0.41 per cent to 19.59 per cent. They were unable to ascertain definitely whether the double carbides we re really compounds or merely mixtures of Fe-jC and Mn^C. However, they reported that there appeared to be only a mixture in steels containing 4.98 per cent, or less, of manga- nese, while higher percentages of manganese seemed to produce the double carbides ^Fe-^C. Mn^C and 2Fe3C. Mn-^C. The carbide, possibly one of these compounds, that appears in high-carbon, high -manganese steels, as cast, is apparently a source of weak- ness. It can be seen as membranes around the polyhedral crystals of austenite, and also segregated in chunks. When the steel has been quenched in water from a high temperature, and thus greatly toughened, the structure shows no carbide. The quench- ing prevents the carbide from separating out, and leaves it in solution. Manganese depresses the critical temperatures of steel, so that the forms that exist only at high temperatures in car- bon steels are present as normal (though not necessarily stable) constituents of manganese steels at ordinary temperatures. Manganese steels cool more rapidly than carbon steels, and have a greater contraction, which tends to cause piping and settling in the molds. The specific gravity of manganese steels is somewhat greater than that of carbon steels. -4- EXPERI MENTAL It was proposed to investigate the properties of manganese steels with manganese content of one to five per cent, with a view to finding whether the brittleness could be eliminated in these steels. If these steels can be made malleable, they will be valuable commercially. It was desired to study first the carbide of manganese, and the absorbtion of carbon by manganese. The absorbtion of carbon by manganese was to be brought about by casehardening some pure manganese. Attempts were made to determine the crit- ical point for manganese, in order to find the best temperature for the casehardening. Some pure manganese was ground up and put into a small fireclay crucible. A hole was bored in the side of the crucible and the thermocouple, enclosed in a silica tube, was put into the metal through this hole. The metal was then heated in an electric furnace. Two different attempts were made in this way, but no sign of a critical point appeared on either cooling curve, and though there was an indication of a critical point at 1009° G. on one heating curve, no such indi- cation appeared on the other heating curve. In the two attempts, the highest temperatures reached were 1025° C. and 1062° C. , respectively. In each case, the metal was badly oxidized, and fused into a solid mass. The experiment was repeated, and this time a steady current of CO2 was kept flowing through the fur- nace, to prevent oxidation of the manganese. The metal was again found to be completely fused, however, and no indication of a critical point was observed. The highest temperature was 1025° C. -5- On the fourth attempt, nitrogen was substituted for C0 2 , and the temperature reached was 1049° G. The metal was slightly fused. An indication of a critical point at 1021° G. was found on heating, but no point was obtained oh cooling. For the last attempt, the manganese was used in a different form. Two pieces of manganese (about 80 to 100 grams each), each with one flat surface, were used. A groove was ground in each of these two surfaces and the two pieces were wired together in such a way that the grooves were opposite each other and formed a hole into which the thermocouple could be inserted. The metal was heated to 1050° C. , and again no critical point was found on cooling, though on heating, a slight temperature pause was found at 1013° G. No more attempts were made to determine the critical point for mangahese, but 1050° G. was selected as the best temperature for casehardening. The first casehardening mixture used con- sisted of two parts, hy weight, of BaCO^, two parts of powdered charcoal, and three parts of bone meal. This was not quite satisfactory, and a mixture of sixty per cent charcoal and forty per cent BaCO^ was substituted, with good results. At first the casehardening was done in small fireclay crucibles, with a cover of alundum cement paste. The alundum cracked a little around the edges and when the crucibles were kept at 1050° C. for more than two or three hours, a great deal of the charcoal burned out. An iron pipe, eight inches long and two and one-half inches in diameter inside, was then substituted for the crucibles. One end of the pipe was closed with an iron top; alundum cement was used to stop the other. Not enough - 6 - carbon burned out to prevent good carbonization when the pipe was used. Pieces of Armco iron, pure manganese, and .20 carbon steel were put into the casehardening mixture together. The manganese showed the best carburization of the three. Casehard- ening was tried at 925° C. , but the carburization was not as complete at this temperature as at 1050° C. The duration of the process varied from two to six hours, the latter time giving the best results. Figure 2 shows a good example of casehardened manganese. Some manganese carbide was next made by melting some ground manganese in a fireclay crucible with powdered charcoal. At a temperature of 131C° C. , the molten metal ran through the crucible. Enough of it was saved to make a small button. This button could be ground on an emery wheel without cracking, but it broke into bits under a, pressure of 500 kilograms in a Brin- ell machine. The structure of the material was more fine-grain- ed than that of the case-hardened manganese. A second quantity was made up, and had similar properties. A large gas pot-furnace was used to melt the metal. Photographs of this carbide are shown in figures 3 and 4. Attempts were now begun to make manganese steels of from 3 to 4.5 per cent manganese. The first sample consisted of 112 grams of 0.30 carbon steel and 5 grams of pure manganese, which would give a steel with 0.29 per cent carbon and 4.27 per cent manganese. A graphite crucible was used, in a gas pot- furnace. The crucible and the sides of the furnace were melted, but the steel was not. A large oil spray pot-furnace was then tried. On the second attempt, a melt was obtained and the steel -7- was poured into a small steel mold. On cooling, however, the steel stuck to the mold, and could not even he hammered out. The mold was sawed in two, so as to get a surface of the man- ganese steel to polish and photograph. No tests could he made on this steel. % Another quantity of steel and manganese was melted up, in the proportions necessary to give a steel containing 3*3 per cent manganese and 0,28 per cent carhon. This steel wa.s poured into a sand mold. The sand was evidently wet, and the casting was so full of holes as to he quite worthless. The steel seem- ed to he extremely hard and brittle. A quantity of .30 carhon steel, .10 carhon steel and pure manganese was melted to obtain a steel of 3*4-9 per cent man- ganese and 0.236 per cent carhon. A lined graphite crucible and the oil-spray pot-furnace were used. Two small sound cast- ings were obtained,- one in a steel mold, and one in sand. A small surface was polished on the steel mold casting and the photograph shown in figure 5 was obtained. The steel apparently contains a great deal of carbides. This bar was then heated to orange heat and hammered. It was fairly malleable, and was hammered to about half Its original thickness. After cooling it could be sawed in two with a hack saw. Figure 6 shows a photograph of this forged steel, with the flattened grain struc- ture resulting from forging. One piece of the bar was then heated to 1200° G. for thirty minutes and quenched in oil. The structure v/as found to be mostly martensite, with some austen- ite. (Figures 7 and 8) The sand casting v/as too hard, as cast, to be sawed before - 3 - heat treatment. It was packed in calcium oxide in the pipe used in the casehardening work. The pipe was sealed with alun- dum cement. It was heated in a gas muffle furnace. The temper- ature was raised rather slowly to 7 00° C. , and then rather rap- idly to 1220° C. It had been intended to go only aboutto 1150° G. but the needle of the pyrometer stuck at 1100°, and this was discovered and remedied, but it was found that the temperature was 1220° G. The gas was shut off, and the furnace opened amid allowed to cool to 6 50° C. The furnace was then closed and re- heated to 1050° C. The temperature was kept at 1025-1050° for forty-five minutes, and then allowed to fall very slowly, six hours being required to cool to 450° C. The casting was found to have been softened considerably by the heat treatment, and could be sawed in two fairly easily. The microphotograph (Fig. 9) shows the coarsely crystalline structure of normalized low- carbon steel, with needles of manganese carbide. An attempt was next made to produce a manganese steel con- taining 2,0 per cent manganese and .20 per cent carbon. The oil-spray pot-furnace was tried first, but the attempt was un- successful. A gas furnace was then used. This furnace had a half -inch lining of alundum cement with a three-inch insulation of Sil-O-Cel between the lining and the jacket of sheet iron. The flame entered a hole just above the bottom of the furnace, from a special burner. The nozzle of this burner consisted of a three-eighths inch iron pipe, inside of which was a smaller pipe. Air came out through the smaller pipe, and gas through the larger one. The pipe which conducted the air to the nozzle was heated for a distance of about three feet, by a battery of -Q- gas burners (from a dismantled combustion furnace). The first attempt to melt the steel in this furnace failed, evidently because the gas pressure in the mains was too weak. The second attempt was successful, but the crucible slipped while the steel was being poured, and nearly all of the metal splattered out of the steel mold. An attempt to make a steel containing 4.18 per cent man- ganese and .214 per cent carbon failed because the metal was not hot enough when taken out of the furnace, and froze while it was being poured. At this time another furnace was built of firebrick. The cavity in the furnace was about ten inches square, and the depth was nine inches. The opening was made almost half-way up from the bottom of the furnace, and at one corner. The other three corners were built up v/ith alundum cement, to give a round- ed surface. This caused the flames to encircle the inside of the furnace completely, and made it possible for the crucible to get the full heat without being placed directly in the full blast of the flame. This arrangement made the heating of the crucible fairly even all around, and reduced the tendency of the crucible to soften and give way on one side. The furnace was unsuccessful at first, however, and this was found to be due, partly at least, to the fact that the opening for the flame was too high up. The flames rose after entering the fur- nace, and the bottom of a crucible remained comparatively cool. (The burner was the same as used for the circular furnace pre- viously described. ) Better results were obtained after the furnace was rebuilt with the opening for the flames close to - 10 - the bottom. A steel containing 3*54 per cent manganese and .23 per cent carbon was now made in the furnace described. It was cast in a sand mold, and appeared to be a sound casting. It was too hard to be sawed in two with a power hack saw before heat treatment. It was packed in CaO in the pipe used for previous heat treatments, and heated in a gas muffle furnace to 1080° G. It had been intend- ed to heat the steel to about 1200° C., but the gas pressure was too low to allow a temperature of more than 108C° G. to be reached. The steel was then cooled to 750° C. , and reheated to 1050° C. It was kept at 1050-1060° G. for one hour and five minutes, and then cooled very slowly in the furnace. The casting could then be sawed easily. It had a Brinell number of 286. When it was sawed, it was found, however, that it was too full of holes to be suit- able for making a test bar. Three giore attempts were made to obtain a casting large enough and sound enough to make a test bar, but all of them fail- ed. It was found that, as a rule, the gas pressure was sufficient to yield the necessary heat only late in the evening, and as the furnace was being used by another man also, the opportunities for obtaining a melt were few. Also, before enough experience was gained in making castings to give satisfactory techMque, the school year was ended. One reason why the castings made in the sand molds were unsound was that no "riser" or "feeder" was provided for in the molds. If a bell-shaped enlargement is made at the top of the mold, and enough metal is poured in to fill the mold to the top of this riser, the castings are far more likely to be sound. The holes are made by the contraction of the metal on . - - 11 - freezing. If there is a supply of molten metal at the too of the casting, it will he drawn down by the shrinking of the low- er part, and will fill up the holes there. Owing to the fact that no casting suitable for a test bar was made, no figures showing the tensile strength, elastic lim- it, elongation, and reduction of area could be obtained. The Brinell hardness was not taken on the small castings made early in the work, as they were really only preliminary efforts, and it was hoped at that time that it would be possible to make large castings from which all necessary significant data could be obtained. CON CLUSIONS Should this investigation be continued, the work already done shows that two essential features must be developed. These are, first, crucibles that will stand melting steel and slag up to about 1700° C., and second, a furnace of capacity to melt steel to cast an ingot, probably three inches in diameter, and long enough to permit a sound test bar to be cut from it. At the time this investigation was started, work was begun on an electric carbon resistance furnace designed to give a temperature of 1800° C. Receipt of the refractory parts of this furnace was delayed long past the expected date of delivery. The furnace was assembled and the test runs made on it during the last few days of the semester. No time was available for using it in this investigation, as was the intention when the work was started. Although the results of this work were so meagre, they - 12 - give noteworthy indications of the probability that steels containing one to five per cent of manganese can be freed from brittleness by proper heat treatment. This result was apparently attained in this investigation by heating the steel to a temper- ature above 1 100° G. , cooling to about 700° , reheating to 1025-1050°, and finally cooling very slowly in the furnace. It has at least been shown that the suoject merits further and extensive research. BIBLIOGRAPHY Journal of the Iron and Steel Institute, 1388, No. 2, pp. 49, ff. Manganese Steels, Sir Robert Hadfield. Sauveur- The Metallography and Heat Treatment of Iron and Steel. Hoyt- Metallography. Von Juptner- Siderology. Hiorns- Steel and Iron. Hoffman- General Metallurgy. Mars- Die Spezialstahle. x 100 Fig. 3* Manganese carbide, made by melting ground manganese with powdered charcoal. Etched with HNO*. 3 x735 Fig. 4. Same as fig. 3. Shows martensitic structure more plainly. x 1 00 Fig. 5. Steel as cast. Manganese. 3*5 per centJ carbon, ,24 per cent. Full of carbides. Etched with KNO3. x 100 Fig. 1. Manganese, etched with HNO-z . Shows usual polyhedral structure of pure metals. Black spots are inclusions of impurities. xlQO Fig. 2. Manganese, casehardened for six hours at 1050 G. Etched with HNOj.