THE EFFECT OF DEOXIDIZING AGENTS ON THE PHYSICAL PROPERTIES OF STEEL CASTINGS BY HUGO CHRISTIAN LARSON A.B. Augustana College, 1919 L THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN CHEMISTRY IN THE GRADUATE SCHOOL OF THE UNIVERSITY OF ILLINOIS, 1922 URBANA, ILLINOIS \ J (A. UNIVERSITY OF ILLINOIS THE GRADUATE SCHOOL MAY-3L, -IQ2-2 I HEREBY RECOMMEND THAI' THE THESIS PREPARED UNDER MY SUPERVISION BY HTTCf) CHRIS T IAN LARSON ENTITLED THE EFFECT O F DEOXIDIZING AGENTS OR — THE -PHYSICAL PROPERTIES _QF_STEEL CAST LEGS BE ACCEPTED AS FULFILLING THIS PART OF THE REQUIREMENTS FOR THE DEGREE OF IMS TER OF SCIENCE TN CHEMTSTRY 7±bb w. ( -JJ. In Charge of Thesis o<_a_a_ (JL-C.'tt Ci Head of Department Recommendation concurred in* Committee on Final Examination* •Required for doctor’s degree but not for master’s ; ' '• • • ACKNOWLEDGEMENT The writer wishes to express his most sincere appreciation to Dr. W. S. Putnam, whose personal assistance and encouragement were invaluable during this investigation. He wishes also to acknow- ledge his indebtedness to Professor C. W. Parmelee for his work in connection with the preparation of an electric furnace and zircon crucibles, and to Mr. E. Sunstrom for his kindness in preparing drawings and photographs of the apparatus used. • : f ; : u : 1 ' TABLE OP CONTENTS Page I. Introduction 1 II. Historical and Theoretical 3 III. Apparatus and Experimental 6 Purnaces 6 Crucibles 16 Slags 18 Molds 19 IV. Summary 26 V. Bibliography 27 VI. Figures I. Oil furnace 7 II. First type of furnace 9 III. Burner with air preheater 11 IV. Pinal type of furnace 13 V. Photographs of furnace 14 VI. Muffle furnace for slag melting point determinations 20 VII. Types of molds 21 VII. Tables I. Analysis of gas 12 II. Analysis of Flue 15 III. Calculation of slags 22,23 Digitized by the Internet Archive in 2016 https://archive.org/details/effectofdeoxidizOOIars THE EFFECT OF DEOXIDIZING AGENTS ON THE PHYSICAL PROPERTIES OF STEEL CASTINGS I INTRODUCTION Although deoxidizing agents have been employed successfully in the metallurgy of iron and steel, the effect which such agents have on the physical properties of steel castings is not found to any great extent in literature. This problem was suggested by the following statement of Giolitti^in regard to the special methods of melting, pouring, deoxidation and heat treatment; "An explanation in great detail is at present impossible, owing to the fact that they have been brought to their present perfection in very recent times, and are still held as industrial secrets only partly protected by patents." (p£42) The purpose of this investigation will be to note the variation in tensile strength, hardness, ductility and crystalline structure with varying amounts of the ferro-alloy of the less common deoxidizing agents, such as vanadium, titanium, and uranium. The steel used wiU be forty carbon (o.40$ C) and will be cast into bars of sufficient size to be machined down for use in the tensile strength testing machine. The hardness will be determined by the Brine 11 method and micro -photographs will demonstrate the structure. The length of time of heating, the percentages of deoxidizer added, the length of time of addition before pouring and the temperature of pouring are variables which will be controlled. The castings will be uniformally heat treated according to the methods suggested by Giolitti for the treatment of soft and . ' . ■■ - 2 - medium carbon steel. -3- II HISTORICAL AND THEORETICAL The use of deoxidizing agents, or "Scavengers”, in removing oxides, nitrides and occluded gases dates as far hack as the Bessemer process for making steel. In the year 1856, Sir Henry Bessemer pub- ( B ) lished his epoch-making paper in which he called attention to the necessity of manganese as an addition product in the final purifica- ( 2 ) tion of the steel. Shortly afterwards, Yalton , director of the Terre-Noire Steelworks, gave an explanation of the part manganese played as a deoxidizer in the Bessemer process, and in 1876, Gautier ( 3 ) communicated to the Iron and Steel Institute the value of ferro- manganese in the open-hearth process. Some years prior to Gautier, Siemens called attention to the action of manganese on steel. In 1875, Pourcel manufactured f erro -mangane se in the blast furnace. Robert s-Ansten concluded an address to the Iron and Steel Institute in 1900 ^) with the remark: "May we not hope that in the next century vanadium, molybdenum, titanium and glucinium will prove as faithful allies as manganese?" As early as 1894, Rossi ^ observed the beneficial effect of titanium in the manufacture of steel, and the ferro-alloy of that element was studied by Wohler and St. Clair Devilled) about the / c \ same time. Vanadium was discovered in 1830 by Sef stronr but it was not until 1896 that Chambley^) showed the advantages of vanadium in steel-making. Aluminum/ 8 ^ , cerium/ 9 ! uranium, molybdenum and glucinium have all been used as deoxidizers with various degrees of success. In practice, the steel is first partially deoxidized by f erro -manganese or spiegeleisen, and this is followed by addition of : • i : ■ ■ ' 3 l ■ i‘. % i jv.-liisl . 5 * - If ■ -4- the ferro-alloy of vanadium, titanium or uranium. In this way, all the impurities which have been acted upon by the manganese compounds are removed by the more active deoxidizers. Various theories have been advanced to explain the effect of ( 7 ) these agents on steel. Arnold and others believe that the bene- ficial effect is due to the deoxidizing or ’’scavenger" action of the substance. Norris^ 11 ^ contends that the improvement of physical prop- erties is due to the fact that they alloy themselves with the iron ( 12 ) and form troostitic and sorbitic pearlite. Anderson states that besides acting as deoxidizers, they tend to render slags more fluid, causing them to separate more completely from the metal. The pres- ence of the oxide of titanium or vanadium lowers the fusion point of the occluded slags in the metal, thereby imparting fluidity. It has been found ^ 13 ^ that 6-§"$ of Ti0 lowers the melting point to 1290°0. and 13$ addition lowered it to 1190°C. The great affinity of these agents for oxygen is a sure means of reducing the occluded oxides such as PeO, PegOg, and such gases as oxygen, nitrogen and carbon monoxide. The fact that titanium and vanadium unite readily with nitrogen make them valuable in removing that element from the metal. The harmful effect of nitrogen on iron and steel was brought out by LeGhatelieJ*^ in a paper before the Congress of Metallurgists in Belgium in 1905. He showed that nitrogen to the extent of from .02 to .045 percent is sufficient to make the steel so brittle that elongation and reduction of area is decreased to a marked extent. A great increase in nitro- gen will cause the steel to become so brittle that it can be crumpled in the fingers. Another beneficial effect of deoxidizers is their (15) tendency to retard the segregation of sulphur, phosphorus and carbon. Very little of work done by metallurgists in studying the -5- problem of deoxidation has been concerned with steel castings, but the theoretical considerations involved are practically the same. The metallurgists in charge of the steel-working in the munition factories of several European countries have succeeded in making large steel-castings which have proved to be just as resistant to the great strains they must bear as those which are made by tedious ma- f 16 ) chinery and forging. The Guldsmedshytte Company, Ltd. , in Sweden, make guns, anchors and other large castings of cast steel. The Gio. Ansaldo Company of Italy manufacture large steel castings for gun- carriage and gun-mount parts. The metallurgist at this plant, F. Giolattil^ 7 ) states that this was accomplished by "obtaining per- fect deoxidation of the metal, and the highest possible elimination o: emulsified non-metallic inclusions which possess an oxidizing power upon the mass of steel. This was due in part to the specific action of the titanium and vanadium used in their manufacture which caused a great frequency of the centers of alpha crystallization which form in austenite during allotropie transformation." The heat treatment given these castings by Giolitti is another important factor which helped to produce a strong, tough steel with uniform and homogeneous crystalline structure throughout the material. - 6 - III APPARATUS AND EXPERIMENTAL In the attempt to solve the problem of the effect of certain agents on steel castings, it was necessary to secure homogeneous castings of 40-carbon steel with certain percentages of the various agents added. This necessitated a temperature sufficiently high, not only to melt the steel but also to bring it about 100°C. above its melting point in order that the steel would not solidify during the ( 18 ) pour. The melting point of the steel used is about 1425°C, and it was thought that a temperature of 1600° would be high enough to secure a pour. It has been stated/\that the longer the metal is kept in the furnace in the liquid state and the higher the temperature of the pour, the greater will be the purification accomplished. Furnaces: The first furnace used was an oil-blast furnace (See FIG. 1) After several attempts, the hope of obtaining a good pour from this furnace was abandoned, because the highest temperature reached was 1450°C. So much heat was lost by radiation during the removal of the crucible from the furnace that the steel was left in a pasty condition which made it impossible to pour. To obtain a higher temperature, gas was used as fuel and various types of furnaces were developed. At first, a small gas-air blast furnace was tried, but the highest temperature reached in this type was only 1350°C. Preheating the gas and air in an iron tube over a series of Bunsen burners raised the temperature to 1375°, a tempera- ture at which the steel was slightly softened, altho still in the solid state. The walls of the furnace were constructed of a fire- clay refractory only two inches thick, and this caused such a great • m - 7 - /7f.-/CS - 8 - loss of heat by radiation that the temperature could not be brought up to the melting point of the steel. An additional heat insulator was built up by constructing a sheet-iron cylinder around the furnace and filling this with pbwdered silocel made into a paste with water. When this was dried slowly, all the water was expelled leaving a porous insulating wall about two inches thick which surrounded the smaller furnace. This prevented the loss of heat to a great extent, and on a trial run, the temperature registered about 1,400°C. The silocel insulator melted down, however, and fire-clay bricks held together with alundum cement were substituted. (FIG?. 2) It was thought that the preheating of the gas caased an expan- sion which decreased the actual quantity of gas, and so the air alone was pre-heated. It was found by running a blank test on the heating- value of the preheated air, that the temperature was raised 300°0. This was sufficient to raise the temperature the necessary amount above that due to the gas to produce a melt. A few trials were made by passing the gas thru a cylinder containing naphthalene, the idea being to enrich the gas with a hydrocarbon of high heat value. Out of five trials, only one was successful, which indicated the addition of naphthalene did not insure high temperatures, but rather that the variable pressure of the gas from the mains was responsible for the failure of the attempts. It was found that the greatest heat was obtained between 9 and 11 P.M. , on account of the increased pressure of the gas at that time. An oxygen tank was connected into the air-passage in such a way that the air could be enriched with varying quantities of oxygen, thus providing for the complete combustion of the gas. This was found to supply more oxygen than was necessary for the burning of the gas. - 9 - * jy^^'^fZ K N i \ 7 -/^r-^ee/sx' y^k^r a r ^r~ _ Lk F<' \A — > ijM — y* a “ L / 1 v /' 4^ fv 4\ ~~* = M — ^Jx r,\ y \«' — l'\ % ~J& TyiJ'.yi MhMb y/^-yy yyyyy yyyy yy yyyyyyy; ■ -/ V- t?y yvyy’yy&y*?' -15- By means of a tube and a gas-sampling apparatus, two liters of exhaust gas was withdrawn and analyzed. The low percentage of carbon monoxide and the slight excess of oxygen indicate that nearly all the heat value of the gas was being obtained. (See TABLE II) TABLE II ANALYSIS OP FLUE GAS COMPOSITION PERCENTAGE C0 2 8.0 CO 0.5 02 5.0 % 86.5 It was planned to use an electrical resistance furnace for melt- ing the steel, but owing to difficulties experienced by the Department of Ceramics, this was not completed until the latter part of the term. En the first trial run, it was found that a temperature of 900°C. was obtained in nine minutes. The carbon electrodes became heated to such a ,n extent that a system of cooling was necessary. A copper sleeve vas cast, and fastened around the electrode, and water circulated thru It. This proved quite successful and a temperature of 1,800°C. was obtained without causing much oxidation of the electrodes. As an addi- tional protection against oxidation, the electrodes were covered with alundum cement. Altho this prevented oxidation, it reacted to some sxtent with the electrode. A silmenite coating was tried and this proved very satisfactory. The furnace consisted of two rings of Acheson graphite placed about six inches apart and insulated from ' -16- tha air by two walls of refractory material. The first layer was one inch in thickness and made of a very high heat-resisting refrac- tory, so as to protect the outer layer which consists of a 4 <§■ inch wall. The crucible was imbedded in finely powdered graphite between the rings, and the temperature is raised by the heat developed from the resistance offered to the passage of the current. In practice, the current was started at 45 amperes and 30 volts, and after about 30 minutes this was stepped up to 95 amperes and 40 volts. (See FIG 6 Crucibles : Considerable difficulty was experienced in securing crucibles which would withstand the high temperatures to which they were subjected. Dixon graphite crucibles were used, but these did not prove very successful because the out ear surface was burned con- siderably However, when a paste of alundum cement was spread over the outer surface to a thickness of \ inches, this was prevented. The basic slags, used in the melt, reacted with the crucibles and caused their decomposition. When an alundum cement inner lining was tried, it was found that the basic slag reacted with the alundum, forming aluminates which went into the slag. Spinel (magnesium alum- inate) was mixed with a starch paste and s pread on the inside of the crucible, but on drying, it flaked off so that it could not be used as a protective casting within the crucible. By the use of an acidie slag, the decomposition of the crucible was avoided. The crucibles were softened by the heating to such an extent that, unless the greatest care was taken, they would fall to pfejes on being removed by the crucible tongs. Some ” zircon” crucibles were prepared under the direction of Professor Parmelee of the Ceramics Department. They were made of -18- zirconium silicate (ZrSiO^.), and contain, therefore, 49.5$ zirconium, 15.5 $ silicon and 35$ oxygen. The melting point of this substance is 2 , 550°0 . { ^ ^which is sufficiently high to warrent its use. Two zircon crucibles were filled with pieces of steel and slag and placed in the furnace together with Segar cones. After one and one -half hours of heating in which the highest temperature was between 1,650 and 1700°0 S they were found to have softened and practically dis- solved in the slag, so that only a few particles remained unaffected. The chemical action of the basic slag and the combination with the iron were probably the causes of their complete failure. Later determinations in an electric furnace (See FIG. 6) have shown that the crucibles do not melt under 1800°C., when free from impurities, but they are very much affected when any foreign substance is placed in them. Slags : The melting points and properties of various combina- tions of slags were determined in order that a suitable slag could be found for the purification of the steel, as well as to prevent atmospheric oxidation. The oxides used in forming the slags were thoroly mixed and ground to pass a 100-mesh screen. Pyramids of the size of Segar cones were formed by compressing the substance into a wood mold. Oil, molasses and water were tried as binders, the first- mentioned giving the best results. These cones were dried and care- fully bales d in a gas muffle at about 1000 °C. for one hour. They were then placed in the center of a gas muffle and an alundum cement rod was rested upon the point of the cone. The rod extended thru a hole in the top of the muffle in such a way that when the cone melted the rod would sink. A platinum-platinum-rhodium thermocouple, -19- insulated in a silica tube, was extended thru an opening in the front of the furnace. PIG. 7 shows the arrangement of the muffle for determining the melting points of slags. TABUE II (A and B) shows the composition of the slags used, and the calculations of their silicate degree, and melting points. It was found that Slag #2 had a higher melting point than Slagjfl, caused by the removal of KgO from Slag #1. Upon adding B2O3 to Slag #2, the melting point was considerably lowered. Substituting GaP for KgO in Slag § 1, increased the melting point about 20°. The fluidity was found to increase upon the addition of borax and calcium fluoride, but to decrease when EgO was present. Slag #1 was found to be too viscous and Slag #2 had too high a melting point to be used satisfactorily. Slag #4 proved to be quite suc- cessful!, having a melting point about 100° below that of the steel, and being very fluid at the temperature of the pour. Molds : After the solution of the problems of obtaining high temperatures, suitable slags and crucibles which would hold up, that of suitable molds was studied. The first molds tried were made of a mixture of powdered fire-clay and sand with water as a binder. A cylindrical form, 6 z | inches was used in a vertical position into which the steel was poured. Blow-holes and contraction-holes were formed, however, and it was believed that if a mold were made with the upper part terminating in the shape of an inverted bell and sufficient metal poured to fill this enlarged opening, the holes caused by the contraction of the metal would be eliminated. (See FIG. 8A) When this was tried, it was found that the blow-holes were formed as before. This was not due to a scarcity of metal, but to the fact that the gases did not have time to escape. - 20 - ry*/*' &/* -32- TABLE III (A) CALCULATION OP SLAGS COMPO- SITION MOL. WT. AMOUNT T MOL. RATIO # COMPO- SITION M.P. OP SUBSTANCE SLAG # 1 CaO 56.1 g. 67.32 g 1.20 39.4 1995°C A1 2 0 3 102.2 26.57 .26 15.5 3000 Si0 2 60.4 59.70 .99 35.0 1750 Fe 2 0 3 159.8 12.78 .08 7.8 1541 k 2 o 55.0 2.20 .04 1.2 890 MgO 40.3 2.02 .05 1.1 1900 SLAG # 2 CaO 56.1 70.68 1.26 45.8 1995 AI2O3 102.2 26.57 .26 17.2 2000 S10 2 60.4 42.28 .70 27.4 1750 PegO 3 159.8 12.78 .08 8.3 1541 MgO 40.3 2.02 .05 1.3 1900 SLAG # 3 CaO 56.1 70.68 1.26 44.8 1995 A120 3 102.2 26.57 .26 16.8 2000 SiO 2 60.4 42.28 .70 26.9 1750 Pe 2 0 3 159.8 12.78 .08 8.1 1541 MgO 40.3 2.02 .’05 1.2 1900 B 2°3 154.0 3.08 .02 2.2 577 SLAG # 4 CaO 56.1 67.32 1.20 39.1 1995 A1o0 3 102.2 26.57 .26 15.4 2000 Si0 2 60.4 59.79 .99 34.8 1750 PGgOg 159.8 12.78 .08 7.4 1541 MgO 40.3 2.02 .05 1.2 1900 CaP 2 78.0 3.12 .04 2.1 1378 V . -23- TABLE III - (B) SLAG no. M.P. OXYGEN IN BASE ” " ACID SILICATE DEGREE 1 1320-1330 230/198 = 1.1/1 Mono-silicate 2 1400-1425 233/140 = 1.6/1 Sub-silicate 3 1250-1275 234/143 = 1.63/1 n n 4 1350-1375 228/198 = l.l/l Mono-silicate To eliminate these gases, a mold was constructed as shown by FIG. 8-B. The metal is poured down a vertical passage into a hori- zontal cylindrical opening, at the other end of which, a vertical passage was constructed for the escape of gases. A very rough cast- ing was obtained with a hollow space or ”pipe” extending thruout its length. In an attempt to eliminate the roughness of the surface, an iron pipe 1-4 x 12 inches, closed at one end was embedded in a sand box. A crucible from which the bottom had been removed was cemented onto this pipe, the purpose being to provide a bell into which the excess metal could be poured. This gave a smooth, external appearance but synmetrically arranged blow-holes appeared thruout the casting, and a '’pipe” extended thru three inches of the length. A mold con- structed as the one shown in FIG. 8 - C is recommended. The steel is poured down one passage, passes thru a horizontal passage and rises Into the other arm, thus driving out the gases. Altho about twenty-five attempts were made to obtain a casting free from blow-holes and of such size that a test-bar could be ’ . -24- machined from it, the results were entirely disappointing. Every precaution was taken in regard to the control of temperature in pouring, the fluidity of the slag, the type of mold, and the peiiod of solidification, hut hlow-pipes or holes appeared in every case. The failure of the castings may he attributed to several causes: First, the small quantity of steel, which of necessity was used, caused too rapid solidification. This resulted in the inclusion of the gases in the pasty mass of the metal and their retention hy the metal on solidification. In the smaller furnaces, about one pound was melted, the purpose of the small eastings being to obtain a piece which could be photographed. The enlarged furnaces were capa- ble of holding a crucible of five-pound capacity, or enough metal to fill a mold 12 x li inches. It is suggested that an ingot three inches in diameter and 14 inches long be cast into a hot mold (See FIGr. 8-0) in an attempt to eliminate too rapid solidification. This would require 28.4 pounds of steel and a crucible and furnace of sufficient size would have to be constructed. An analysis of the blow-hole gase £22) has shown that they con- tain hydrogen, nitrogen and earbon monoxide. The first two are probably absorbed from the gases in the combustion chamber, while the last -mentioned gas is formed by the action of the carbon in the steel upon the ferrous oxide dissolved during the melt. This reac- tion may be represented by the equation: C + FeO — > CO + Fe. The higher the temperature, the more FeO will be dissolved and conse- quently the more CO will be liberated. The length of time of solidi- fication, however, will be greater, and thus the gases will have 022 ) more time to escape. It has been recommended that if the metal is stirred just before pouring, the escape of the gases will be ■ . ' ' ' ; • ; -25- hastened. It was thot that by adding a sufficient amount of deoxidizing material for about twenty minutes before pouring, the FeO would be reduced so that no CO would be formed. This was tried, but the small quantity of metal used was probably responsible for the blow- holes produced. The problem of blow-holes in steel castings is one which has been a source of trouble and annoyance to the steel industry. As one investigator expressed it:^ 23 ^ ,T There is no rapid ar royal road to the production of sound steel eastings; this is .: 2 01 (1910) 16. "Iron and Steel in Sweden", Aktiebolaget Svenska Teknologf’dren- ingens fOrlag, p.31 (1920) -28- 17. Giolitti, F. , ’’Heat Treatment of Soft and Medium Steel’ 1 , p. 243. 18. Sauveur, "Metallography and Heat Treatment of Iron and Steel", p. 438. 19. Sisco, F. T., "De-oxidation and Desulphurization in the Her Eurnace", Ghem. & Met., 26: 17 (1922) 21 Washburn and Libman, J. Am. Ger. Soc., Aug., 1920. 20. Parr, S. W. , "Fuels, Gas, Water and Lubricants", p. 186. 22. Von Malt it z, E., "Blow-Holes in Steel Ingots", Am. Inst. Min. Ihg 38:412 (1907) 23. Hadfield, "Aluminum Steel", J. Ind. and S. I., Vol. II, p. 174.