THE- RELM’WE EFFECTS OF THE {NOCULANTS CALCIUM-SEUCON AND FERRO-SlLiC-OH EN THE PRGDLifiTION 0F NGDULAR CA5? IRON Thesis fer the Degree of M. S. MICHIGAN STATE COLLEGE 30hr: E. Kmse 195G THESE? This is to certify that the ' thesis entitled "The Relative Effects of the Inoculents Calcium-Silicon and Ferro-Silicon in The Production of Nodular Cast Iron” presented by John E. Kruee has been accepted towards fulfillment of the requirements for Met. E. L degree in 11% y / Manor professor Date December 1, 1950 0-169 THE W EFFECTS OF THE INOCULANTS CALCIUM-SILICON AND FEPw-SILICON IN THE PRODUCTION OF NODUIAR CAST IRON John E. Kmse A THESIS Submitted to the School of Graduate Studies of Michigm State College of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of MASTER OF some! Department of Metalltu'gcal Engineering 1950 Acknowledgnents The author wishes to express his thanks for the cooperation given him by his msdor professor, Dr. A. J. Smith. He wishes to express his deepest gratitude to Mr. H. L. Womochel, under whose direction this investigation was conducted. The assistance in pouring the beats and making the tensile tests, given by Mr. D. J. Harvey, is greatly appreciated. 244561 II III Table of Contents Introduction Survey of Literature Original Research Sampling Procedure Chemical Determination and Preparation of Specimens Macro-examination Results Micro—examination Results Statistical Approach to Variance in Nodule Size Tensile Test Procedure and Results Discussion of Results Microstructure and Chemical Composition Chill Characteristics Tensile Properties 81mm and Conclusions Bibliography 54 No. l \oooqcrmhmto ESSE'S 15 16 List of Tables Title Base Iron Composition (Per Cent) Some Magnesium Alloys Used Alloying Efficiencies of Various I-Iagnesimn Alloys Magnesium Recovery Crystal Systems Base Iron Composition Furnace Charge Compositions (Per Cent) Per Cent Composition of Inoculants Analysis Using Ca-Si as the Inoculant Log of Heart J-ll Ftn'nace Charges (Pounds) Chemical Compositions (Per Cent) Statistical Results on Nodular Diameters Tensile Properties Microhardness (DPH Numbers) Metallographic Data Page 10 14 15 18 22 23 25 1‘3 42 1+2 1+3 43 List of Figures No. Title Page 1 Photomicrograph (Heat J-7) 1.4 2 Photomicrograph (Heat J-9) 45 3 Photomicrograph (Heat J-ll) 46 4 Photomicrograph Representative of Heats J—l through J —6 4'7 5 Photomicrograph.(Heat J-5) 48 6 Photomicrograph Showing Area of Segregation Around Intercellular Area 49 7 Photomicrograph Illustrating Area of what is Assumed to be Steadite 5O 8 Photomicrograph (Heat J-9) 51 .9 Photomicrograph (Heat J-ll) 51 10 Keel Block Illustrating Shrinkage 52 ll Keel Block Illustrating Sampling Locations 52 12 Chill Blocks 53 13 . Arrangement of Equipment for Experimental Work 53 Diagram I Keel Block Dimensions 1.0 I. Introduction The commercial grades of gray cast iron range in tensile strength from 20,000 psi minim to 60,000 psi minimum, with practically no ductility. Ductility as defined in the Metals Handbook is "the property that permits permanent defamation before fracture by stress in tension.’ The size, shape, and location of the paphite flakes is largely responsible for the lack of ductility and the limited strength of these cast irons. Mdleable cast iron is produced from specific metal mixtures. The metal as-cast is composed of a very hard structure known as cememtite. It is necessary to heat treat all as-cast malleable iron to obtain a structure consisting almost entirely of ferrite and paphite in the form of nodules. Malleable irons having this structure possess ductility, whereas the commercial cast irons have practically none. However, malleable cast iron is much more costly to produce. The structure ofthe iron as-castmaybe changedbythe addi- tion of materials to the molten metal as it is being poured into the ladle of Just before it is poured into the mold. The change in plwsical properties that is obtained is of such an extent that it cannot be explained by Just the change in composition resulting from the use of these materials. Those materials are known as ino- culants. Two inoculate that have been used for new years are calcium-silicon and farm-silicon. Morrcgh md Willimns in 1948 published their results on the utilization of cerium for the production of cast iron with nodular graphite in the as-cast contition. During the discussion of this paper, T. H. Wickenden of the International Nickel Compazw announced a method utilizing magnesium for the production of cast iron with nodular graphite in the as-cast condition. The magnesium or cerium aidition is followed by the addition of a graphitizing inoculent. This type of inoculant is one which contains elements that pro— mote the appearance of carbon in the form of graphite. Some of these are farm-silicon, calcium-silicon, ferromanganese-silicon, zirconium-silicon, silicon manganese-zirconimn, and nisiloy. The advantages of nodular iron—high strength, (100,000 psi), some ductility in the mat condition, increased shock resistance over commercial gray irons, and production cost equivalent to gay iron—have been widely acclaimed. As yet there is neither enough control nor basic knowledge of the process for its adaptation on a camercial basis. Some recent work done at Michigan State College(6) has indi- cated the superiority of calcium-silicon over fem-silicon as the inoculant in mercial gramr irons. As either of these graphitizing inoculants can be used in the production of modular cast iron, m investigation as to their relative effectiveness is the purpose of this research. II. Sm'vey of Literature Some of the terms designating this new type of cast iron are: spheroidal-graphite cast iron; spherulitic cast iron; ductile cast iron; and nodular cast iron. As reviewed in an article from the American Foundryman magazine (7) nodular irons may be ductile or relatively non-ductile , hard or relatively soft, magetic or non- magietic, heat resisting or not pm'ticularly heat resisting, etc. Therefore it is not possible to characterize them by any one pro- pertyandinthiswaytonmnethem. A nodular structure is "characterized by nodules“. A nodule is I'a rounded mass of irregular shape". These definitions from Webster' 3 Collegiate Dictionary could describe the graphite in malleable cast irons as well. The nodular yaphite in malleable cast irons is usually referred to as “temper carbon nodules", thus implying that these nodular graphite structures are produced by heat treatment. There are We forms of nodular graphite distinguishable in malleable cast irons. The form which occurs when manganese sulphide is present is described as 'graphite flake aggregate temper carbon nodules". This structure is characterized by an aggegate of small, apparently randomly orientated graphite particles, and is typical of hnerican or blackheart malleable iron. When the sulphur that is present is predominantly in the form of iron sulphide, the temper carbon nodule is described as spheru- 1itic. This form of nodule is found in the European whitebaart malleable and also in the as—cast nodular cast irons. "Spherulitic" does not mean the same as ”spheroidal". Spher- ulites may have many shapes. Usually the spherulites of gaphite in cast iron are approximately spheroidal in outline. "Each spherulite consists of a number of graphite crystallites radiat- ing from a common center, the basal planes of each hexagonal gaphite crystallite being oriented so that they are approximately at right angles to radii of the spheroid of which the spherulite can be considered to be composed."(8) Therefore when we speak of "nodular iron" or "nodular cast iron" in reference to the cast iron that is obtained with the proper magnesium or cerium treatment, it is implied that all or a substantial part of the graphitic carbon is in the form of spheru- litie nodules in the as-cast condition. The production of nodular cast iron is not so simple as Just adding mamsim to iron. The base iron must be within a particular range, and an optimmn mount of magnesium must be retained. Some of the recomnended base iron compositions are listed in .04 Table I. Table I - ase I_r_qg Mgitiogg (Pg; Cent) 1'0 Si Mn P 5 Reference 3.0-3.5 2.0-3.0 0.5-1.0 .15 .03 3 3.2-3.6 1.8-2.5 0.3 .05 - 13 3.2-3.6 1.8-2.5 0.1 .01. - 13 3.1-3.6 2.0-2.3 0.1-0.2 .02-.03 - 13 3.2-3.6 1.8-2.5 0.3-0.1. .05 .03-.01. 9 3.2-3.5 2.0-2.3 .42-.80 .06-.13 .01 14 3.6-4.2 1.6-2.0 0.1 .1 11 In the cerium process it is necessary that the iron be hypereutectic, that is, the carbon content should be higher C = 4,3 ——(/325H 3/”) than the value sly-WM. In the magnesium process the iron may be either hyper- or hypoeutectic. It is believed that there may be no fundamental difference between the mag- nesium and cerium treatments other than that more cerium is required to produce a given effect.(35) Ziegler, Meinhart, and Goldsmith”) comment that a carbon content under 3.0 per cent will result in a stronger metal, but it will be more difficult to machine. By increasing the amount of carbon while the rest of the composition remains essentially the same, Myskowski and Dunplw“) with a- series of photomicrographs show that the size of the nodules increases. Morral(m) recommends a carbon con- tent cf at least 3.5 Per cent to prevent excessive shrinkage. As a general rule the carbon content is wound 3.5 per cent. , For the cerium process it is recommended(36) that the silicon content after the double treatment should be in excess of 2.31 per cent, but not greater than 3.3 per cent. When nickel and copper are present the minimum silicon content may be slight- ly lower, but not lower than 1.8 per cent provided both or either of these elements is present to give an equivalent sili- con content of 2.3 per cent. This is assuming 1 per cent sili- con equals 3 per cent nickel or copper. Similarly for the mag- nesium process Ziegler, Meinhm't, and coldanthU) state that less than 2.0 per cent silicon will produce white iron, while over 3.0 per cent will cause ferrite to be too brittle. The general effect of increasing mnemts of manganese is 12 the inhibiting of the formation of ferrite, even with fairly high silicon contents. Stated another way, manganese acts as a carbide stabilizer. For the cerium process it is recommended<36) that the manganese content should be between 0.5 per cent and 0.9 per cent except when the element is used to obtain special proper- ties and structures, For the magnesium process Ziegler, Meinhau't, and Goldsmith”) state that a manganese content of less than 0.3 per cent has resulted in poorer physical qualities than when it was kept between 0.5 and 1.0 per cent. They offer no explana- tion for the effect on physical qualities with less than 0.3 per cent manganese. There is some disagreement as to the effect of manganese with respect to magnesium Holdeman and Stearns(9) cement that if the sulphur content is initially high, a higher manganese content such as 0.8 per cent will reduce the necessary amount of magnesium. Mm: Kmianslqm), on the other hand, with reference to the production of nodular cast iron on a commercial basis, states that by keeping the manganese low, a lesser percent- age of magnesium alloy was necessary. He recommends a 0.40 per cent maximum manganese content. The sulphur composition of the base iron he used is low, 0.01. per cent. The general purpose of manganese in commercial gay iron is to act as a scavenger with regard to the sulphur. is magnesium or cerium act as desulphur- izers in nodular iron, the manganese will act primarily as a car- bide stabilizer. It is believed<36) that phosphorus has a harmful effect on the properties of nodular cast iron. With phosphorus contents in success of 0.5 per cent the solubility of cerium is 13 reduced to a figure below the minimum required for the produc- tion of nodular irons. The recommended range is 0.1-0.5 per cent phosphorus. 0n the other hand, E. K. Smithu5) referring to some experiments using the magnesium treatment for nodular iron states that phosphorus is not a deterrent to nodule formation. However, the phosphorus contents in these heats ran from only 0.28 per cent to .045 per cent. Ziegler, Meinhart, and Gold- mithe) state that phosphorus is not affected by magnesium. Contrary to this, Rehder(23) states that magnesium additions produce a definite dephosphorizing action. Some addition agents are more effective than others with as much as '70 per cent de— phosphorization being possible. He also states that the sec- tion sensitivity, with regard to the per cent minimum molesting magnesium required, is less at lower phosphorus content. In order to produce nodulizing effects it is necessary to use a low sulphur base metal (0.03 per cent sulphur, maxim) or to desulphurize with soda ash, or to add excessively large amounts of magnesium. Table 2 lists some of the magnesium alloys that have been used, the per cent magnesium that has been added, and the per cent magiesium that is found in the iron on final analysis. A more complete picture of the efficiencies of various magnesium alloys is shoun in Table 3. It must be remembered that these alloying efficiencies are dependent on the temperatures of the melt at the time of additionué) (increasing temperature decreasing the recovery) and most certainly on the sulphur content of the iron. (15 ’ 17) Table 2' - me Magesium glgzs Used llloy Per Cent Mg Per Cent Mg Reference Added Retained 80Ni-20Mg o. 25 0.03-o:1 9 mil-mg 0035'00 55 0.05.0007 10 SOCu-b’OMg 0.35-0.55 0.05-0.07 10 80Ni-20Mg 0. 25 0.18 3 500u-50Mg 1. 5 0.12 15 Table 3' - gigging Efficiencies of Various Mflegium flozs 9 Alloy Approx. Alloying Efficiency, Per Cent Mg-Cu 50-50 5-8 30-70 8-15 20-80 9-15 10-90 20-25 Mg—Ni 50-50 5-8 20-80 25-40 10-90 30-40 Mg-Sb 20-80 5 Ms-n 55-45 15 20-80 8 14in 90-10 5 Mg-m 50-50 5 25-75 5 Mg-Cu-Zn 50-35-15 5 Mg-Bi 50-50 5 15-85 10 Mg—Bi 63—27 3-10 Cell Magmaium 100 5 15 Table I. shows the magnesium recovery from various addition agents as determined by three investigators. Data on original sulphur contents are lacking, but it is probable that the sulphur contents before additions are in the range 0.02—0.04 per cent.(23) Table 4 - Wm) Average Recovery, per cent Donoho Holdeman Rehder Addition Agent & Ste arns 011-112 50-50 12 5-8 10-14 70-30 13 8-15 - 80-20 15 9-15 - 90-10 - 20-25 30-35 Iii-Ms 50-50 ' a 5—8 - 82-18 27 25-40 30-35 90-10 - 30-40 - J. B. Rehder<23) cements that the data of table I. are out of date and of little commercial interest, but nothing has been published of other results. The actual mount of sulphm- remaining in the iron after the magiesium treatment has a meJor effect on the mount of residual magnesium in the iron necessary to the formation of nodular graphite. When the following assunptions are made relative to determining the magnesium that actually takes part in the formation of nodular graphite, it follows that the amount of magnesium as magnesium sulphide should be subtracted from the total magnesium present. 1. lame that Mgs exists as such. 2. Mgs does not take part in the nucleation process for nodules. 3. 111 the sulphur present is as MgS. The nucleating magnesium will be equal then to the total magnesium 16 minus per cent sulphur x 0.75. Rehder<23> believes the molesting magnesium exists as a carbide. Also a certain minimum nucleating magnesium content is necessary for a completely nodular structure to be formed. Below this amount some proportion of nodular graphite below 100 per cent is formed. Above this minimum amount free car- bide appears in light sections and then in heavier sections as the minimum amount is exceeded. The nucle ating magie sium is apparently independent of actual sulphur content and type or method of mag- nesium addition. It is probably dependent on other chemical com— position and method and type of post-inoculant. DeSy(24) gives a formula for estimating the amount of mag- nesium to be added. Mg(residual) '+ o.75(sl - 82) Mg(u1ded) = Y It is necessary to know the apparent recovery (Y) and also the sulphur content of the original iron. The residual magnesium is fixed. when speaking of the residual magnesium and sulphur it is necessary to place the location of the samples. A curve plotted for mechanical properties vs. Mg(residual) shows a marked mazimmn. The general ideal conditions necessary for an element to be capable of forcing the spherulitic crystallization of graphite, according to Desy(24), are as follows. 1. Should be a strong deoxidizer. 2. L stabilizer of carbides. 3. Should not form components which may precipitate in the 17 hexagonal system in the liquid. 1.. If not mluble in iron, should be capable of emulsifica— tion with iron as liquid or gas. 5. Element should be a graphitizer. Bogart(25) from his talk "Observations on Nodular Iron" states that mechanical properties reach a maximum at about 0.025—0.0[.5 per cent residual magnesium. Best properties are obtained at the lowest magnesium residual which produces a fully nodular structure. The Ford Motor Company operates with residual magnesium around 0.035 per cent with additions ganged to allow for interval between inoculation and pouring. With the holding of 50 pound melts at constant temperature it has been found that the magnesium nodulizing effect largely disappears in about ten minutes. There is a gradual deterioration of the nodulizing effect instead of a sharp line of demarcation as between treated and untreated metals.(5’9) It is possible that with larger ladles and dropping temperatures the effect would per- sist for a longer time. J. E. Rehder, Albert DeSy, and Harold Bogart have been in agreement on the existence of a nucleus for nodular graphite forma- tion. 11. Morrogh has been opposed to this belief. Rehder in sumnarizing his work<25) believes as follows. 1. Nodular graphite is the result of growth on specific nuclei. 2. There are radial and non-radial grqahite types. 3. A nucleus exists in every nodule. 1.. The nucleus is hard and relatively chemically inert. 5. 7. 8. 18 The nucleus is hexagonal in crystal habit. Size and size distribution of nodules is influenced by the nodulizing agent used. Nodules contain Si, Fe, and Ti, with the existence of T102 and Fe203 confirmed by x—ray diffraction. The nucleus may be a carbide or a nitride. Morrogh‘35) believes that extraneous nuclei (other than graphite) are not required, and that the spots in the center of the nodules are due to optical effects. These effects may possibly be a result of the plane of polishing coinciding exactly with the basal plane of a radial crystallite. Albert DeSy has advanced the hypothesis that the crystal system of the nuclei or solid particles suspended in molten gray iron de- termines whether nodular" graphite or flake graphite will precipi- tate. Table 5 shows the crystal systems of anstenite, cementite, and graphite, and some compounds which are generally solid at the temperature of the cast iron eutectic. These compounds mw be present in suspension in the cast iron as a result of the melting or as additions of deozidizing or inoculating materials. Table 5 - Wm anic lustenite M80 Cal) 310 Ed} MID F80 M83 Gas SrS 3&3 MnS Li Oe02 1‘10 320 21C TiN er Tetragonal Orthorhombic Hexagonal Rhombohedral Omentite Graphite 0302 (me) 23102 FeS 019312 1102 (Fe0)28102 3102 A1203 01-02 (FeO.MnO)SiOz Sic CBC 2 L102 216104 19 In order to obtain nodular graphite Fe3C must be precipitated, and this carbide decomposed toward the end of or after the solidi- fication of the eutectic. If the iron is deoxidized with elements having a mater affinity for oxygen than silicon, which do not crystallize in the hexagonal system, and which precipitate in the solid state above the eutectic temperature, the compounds will act as nuclei for austenite. As a result of the lack of graphite form- ing nuclei, the tendency will be towards solidification in the metastable system at the eutectic, i. e. austenite, cementite, and liquid. Such elements are Mg, Li, Ba, Ca, Sr, V, Ce,and Ti. Von Keil and his co-workers(26) have suggested that a sub- microscopic silicate slime nucleates the melt to give coarse graphite, while the absence of the slime allows the formation of 1mdercooled graphite. Norbury and Morgan(27) in order to explain the effect of Ti and 002 suggested that this process modified the slime in such a way as to make it soluble, thus allowing the forma- tion of fine undercooled graphite. About this time the process of inoculation was introduced; The silicate-slime theory could not explain the inoculating effects of ferrosilicon or calcium silicon, or graphite. Piwowarslcy'(28) developed a A graphite-nuclei theory to explain the tendency towards undercooling as a result of superheat- ing. is graphite dissolves very rapidly above the liquidus, this theory was weakened. Sasha”) goes along with the graphite-nuclei theory in that the inoculating effect of ferrosilicon is due to localized silicon concentrations reducing the carbon solubility. Morrogh and willians(26) believe that calcium-silicon acts simi- larly to fem-silicon in this manner. 20 The only reference giving the composition of the melt and per cents of addition agents utilizing calcium-silicon as the moment was by DeSy(31). This data is found in Table 8. Other references”: 1) mention that calcium-silicon has been used but is not so effective as other inoculants. Usually 75-90 per cent silicon in the ferro-silicon inoculant has been used to give a 0.4-0.6 per cent silicon addition. III Original Research The purpose of this research was to determine the relative effects of calcium-silicon and ferro-silicon inoculants in the production of nodular cast iron. It was decided to choose a base iron composition similar to those generally used, to make a mag- nesium addition as 80Ni-20Mg alloy, and to inoculate this molten metal with calcium-silicon or fem-Silicon in amounts such that equivalent silicon per cents are available for pick-up. Table 6 - Base Iron Composition 0 Si Mn P S 3.75 2.60 0.1.0 0.10 0.03 Mar The above base iron composition has been used in the casting of step blocks for dilatometer studies of the annealing of nodular iron. (3) In the one inch section a predominantly pearlitic back- ground was obtained. The per cent magnesium to be added was selected from the survey of literature to be 0.40 per cent magnesium in the form of an 80Ni-20Mg alloy. This magnesium alloy was added to the furnace crucible. It was believed that the temperature drop on pouring a 30 pound heat into a preheated ladle plus the loss in temperature obtained on the addition of mags sium would be too peat. The inoculant was also added to the furnace crucible. A 35 KVA Ad ax High Frequency Induction Furnace with a basic crucible was used. A 30 pound charge was melted for each pour. The magnesium alloy was added through a "stove pipe" fixture hung over the open crucible. This was a 3 foot length of pipe about 2 22 inches in diameter that was welded to a flanged bottom 3 inches long and 8 inches in diameter. The flanged bottom covered the open top of the crucible, and hung about an inch above it. Half way up the 3 foot length of pipe was a slide on which the magnesium alloy was placed. By pulling out the slide the alloy would drop into the molten metal. The purpose of the fixture was to afford a quick and safe method for adding the alloy plus protection from the resulting flash and any molten metal that might splash out. After the magnesium alloy was added, the fixture was swung clear of the crucible and the inoculant added. The power was turned on for mixing and also to raise the temperature of the melt. A standard chill block was cast for each heat. It was de- cided to use a keel block casting for the purpose of obtaining sound tensile bars. Diagram I shows the dimensions of the keel block pattern. It is an adaptation of the type suggested to the A.F.S. and A.S.T.M. committees for consideration as an alternate standard.(30) The molds for the keel blocks and the chill specimens were made of a core sand mix and baked at 35001? overnight (12 hours). The furnace charge was made up of the following constituents. Table 7 - Furnace Charge Compositions (Per Cent) Hanna Malleable Pig Farm-silicon Ingot Iron C 4. 2 C 0.44 Fe 100 Si 1. 52 Si 27.4 S 0.025 S 0.018 P 0.099 P 0.033 Mn 0.36 Mn 0.84 The inoculants were commercial grades of ferro-silicon and calcium-silicon. Their compositions are listed in Table 8. Table 8 — Per Cent Commsition of Inoculants Fem-Silicon Celsius—Silicon Si 92. 73 Si 63. 9 A1 10 68 Ca 31 o 4. Ca 0. 25 Fe 1. 7 Ti 0. 16 Al 1. 05 As the Hanna malleable pig was~ in 50 pound pigs it was nec- essary to cut it into slices about 3/4 inch thick and then break the corners off so that the slice would fit ihto the crucible. A slice weighed approximately 2 and 3/4 pounds. Three or four pigs were cut up at the same time and their identity maintained so that approximately an equal amount of each pig went into a charge. The 27 per cent form-silicon was selected at random from the bin and entered the charge in chunks about one inch square. The ingot iron was selected at random from the bin and was in the form of punchings one inch in diameter and from 1/8 - 1/4 inch thick. The ingot iron was placed in the crucible first. For heats J -1 through J-S all the 27 per cent ferro-silicon was charged next. On top of this was charged Hanna malleable pig in slices and checks to fill the crucible. The power was then turned on. For heats J-6 through J -11 the ingot iron was on the bottom and next was Hanna malleable pig to fill the crucible. The 27 per cent ferro-silicon was added later in the melt. In both procedures the rest of the pig was added as the melt down progressed. After the melt down the slag was removed and a temperature reading taken with an optical pyrometer. Up to heat J -I. no attempt was made to control ultimate melt down temperature as furnace and melt down characteristics were unknown. After heat J-A an attempt was made to hold the melt-down temperature to 2785°F and also to make the Ni-Mg addition at this temperature. The pipe fixture was swung over the open crucible, the power turned off, and the magnes- ium alloy addition made. The power was turned off in case any of the molten metal splashed out and shorted out the furnace. The alloy was in chunks about % inch square. After the addition the fiximre was swung away from the crucible, the slag removed, the power turned on, and the inoculation made. The stirring action of the induced currents in the molten metal was utilized to assure thorough mixing of the inoculant in the metal. From one to two minutes were allowed for mixing. The power was turned off, the slag removed, and the heat poured. In heats J-l through J-S the heat was poured in the following manner. 1. mall amount poured in a cast iron mold to get a com- pletely chilled sample for the carbon detemination. 2. Chill block mold poured. 3. Three keel block molds poured. 4;. Rest of the heat pigged. For heats J-6 through J -12 a small plug, one inch in diameter by 1%- inches high, was poured Just after the first slag was removed and before any magnesium or inoculating additions were made. The purpose was to determine the melt down composition of carbon, 25 silicon, and sulphur. The remainder of the casting procedure was the same as above. At the time this investigation was started, the only refer- ences that were found on the use of calcium-silicon as an inoculant stated that it had been tried but was not as effective as ferro- silicon.’ Due to the lack of information using a Ca-Si inoculant and the lack of experience of this investigator on producing nod- ular irons, it was necessary first of all to develop a procedure using the equipment and material that was available. For heats J -1 through J -6 the per cent magnesium added was varied. In heat J-5, 0.15 per cent magnesium was added. Micro examination of a specimen showed flake graphite and nodules. The flake graphite was located on the surface of the casting (see Figure 5). Morrogh and Williams(26) describe a similar occurance and attribute it to the inoculating effect of the mold wall on a low cerium content casting. Heats J-l, J-A, and 51-6 were treated with 0.40, 0.20, and 0.30 per cent magnesium respectively. The Brinell hardness of 51-1 was 330. The hardnessee of J—4 and J-6 were 300. '0330 per cent * Late in the investigation a notaol) was found giving the following composition. Table 9 - Analygis Using Ca-Si as the Inoculant C Si Mn S P 3.33 2.73 0.9 0.02-0.01. 0.10 0.5 per cent magnesium alloy was added. This we assumed to be a 50-50 alloy because of the large amount added. 0.2 per cent of a Ca—Si alloy was added as an inoculant. The letter went on to say that this gave a nodular iron. , No information was included on the type of Ca-Si alloy or if the 0.2 per cent referred to the available silicon in the inoculant. 26 magnesium addition was selected as this amount was common in the practice of adding 80Ni-ZOMg. The 0.20 per cent addition seemed too close to the 0.15 per cent that gave the flake graphite-nodular graphite structure. Through the heat J—6, Ca—Si had been the inoculant. The microstructure of all these heats was nodular graphite, pearlite, and a considerable amount of interdendritic cementite. We were attmpting to achieve a pearlitic matrix with the graphite in nodules. Heat J-7 was made using Fe-Si inoculant, again adding an mount to make available 0.1. per cent silicon. The microstruc- ture of this heat was the one desired. It was known then that the melt down and addition procedure would give the desired results and control could be maintained. The log of heat J-ll is given in Table 10 as an example of the procedure followed. The calculated composition of heat J-9 was changed so as to be more in line with the base composition of Table 6. This resulted in a lower carbon and a lower silicon content. The desired micro- structure was again obtained using Fe-Si as an inoculant. From the previous heats it was now apparent that the silicon loss for this melt down procedure was about 0.1 per cent. This loss was taken into consideration on calculating the charge for heat J -10. As a result the melt down silicon content differed by 0.06 per cent from that of heat J-9. It was suspected by now that the sil- icon pick-up from the Ca-Si inoculant was not as much as that from the Fe-Si inoculant. Therefore an mount of Ca-Si inoculant was added to make available 0.5 per cent silicon. The silicon pick-up for heat J-lO was the same as that for J-6, 0.16 per cent. The microstructure of J-6 and J-lO was similar; nodular graphite, pearl- ite, and interdendritic cmentite. The charge for J-ll was calculated to give the same final composition as J-9. This was accomplished by adding a greater ammount of ferro-silicon (2'7 per cent silicon grade) in the initial charge. The amount added was the difference between the final sili- con contents of J-9 and J-lO. This difference was 0.28 per cent and was rounded off to 0.30 per cent. Again 0.5 per cent silicon as Ca—Si was added as the inoculant. The silicon pick-up this time was 0.21. per cent. The final compositions of J-9 and 3-11 were similar. The microstructure was of the desired type and matched those of 21-7 and J-9. The charge for heat 51-12 was the same as J-ll. This boat was made with the idea in mind that the calcium was going into the melt and enforcing the carbide stabilizing ten— dency of the magnesium that was already there. _By adding Ca-Si in an amount that would make available the same mount of silicon for pick-up that was actually found in heat J-ll, the mount of calcium entering the melt would be less and the desired microstructure might be obtained. The Brinell hardness of this heat was 269 and the mi- crostructure was nodular graphite, pearlite, and interdendritic cementite. There was less interdendritic cementite present in this heat (J-l2) than in heat J-lO. M Procedure The 1:} x 1:;- x 6 inch test blanks were cut from the keel blocks and machined into standard .505 test bars. Drillings were made with a ~41 inch drill as shown in Figure 11 in the riser part of the keel 28 block. Drillings were also taken from the base iron plug. Samples for carbon determination were prepared by breaking up intentionally chilled treated iron with a hammer and mortar. (32) This chilled iron was obtained by pouring a mall mount into a cast iron pig in the course of casting the keel blocks. Chemical pgteminatiog _a_n_d_. Preparation 2:: §pecimen§ The carbon determinations were made by combustion and volumet- ric analysis. Duplicate smples were run for each heat. The pro- cess was checked continually with Bureau of Standards cast iron smples. The silicon in one gram smples was determined by evaporation with perchloric acid. (33) Silicon contents of heats J-l through J-12 were determined from drillings in the keel block. The melt down silicon contents of heats J-6 through J-12 were determined similarly from the one inch diameter plug samples. The manganese in 0.75 gram samples was determined by the ammonium persulfate oxidation method. (p. 196 Reference 33) The phosphorus in 2 gram samples was determined by the alkali- metric method. (p. 220 Reference 33) Sulphur was determined by combustion and volumetric analysis. Smples were run from the base iron plugs and the keel blocks. The metallographic specimens were mounted in bakelite and polished using the Schaeffler technique.(37) The etchanf. in all cases was 4 per cent picral. An attempt was made to polish samples using alcohol instead of water with the AB Micro Polish on AB Micro Cloth. This was done in 29 view of the apparently water soluble inclusions mentioned by J. E. Rehder.(34) In order to identify the intercellular-type materialzmicro- hardnesses were taken of the ferrite in the rings about the nodules, known cementite, and the intercellular-type material. An attempt was also made to differentiate any iron phosphide that.might have been present from cementite by etching a sample in 4 per cent picral, then in 8 per cent chromic acid, and finally heat tinting the etched sample. The chill blocks were broken in a vise using a 10 pound sledge hammer. No notch was necessary. Due to the characteristic light gray fracture of nodular cast irons it was difficult to see the amount of chill. The broken surface was removed with a cut-off wheel and the burned surface removed with a surface grinder. The surface was wet polished on 280 and 400 grit paper and then etched. All the smples for macro-exmination were etched with a mixture of 12 per cent sulfuric acid, 38 per cent hydrochloric acid, and 50 per cent water at 160°F. 1Macro-examination.Result; As all the keel blocks showed a shrinkage cavity as shown in Figure 10 and the sawed surface of the test blanks indicated poro- sity, one of the keel block castings was sectioned transversally and etched for macro-examination. The porosity showed up, but did not extend far enough into the test bar blank to influence the tensile results of the finished test bars. ~All the chill bars except those of heats J-7, J-9, and J-ll 30 were completely chilled. The chill bars of these heats are shown in Figure 12. A11 of the heats represented by these bars had the same melt down procedure, and addition temperature. Heat J-7 had a higher final silicon content than J—9, but the same mount of inoculant (0.4 Per cent silicon as Fe-Si) was added and the mount of silicon pick-up was essentially the same for both heats. Heat J-ll had a higher melt down silicon content than J-7 or J-9 to offset the low silicon pick-up when using Ca—Si inoculant. The final silicon content of this heat was the same as J-9. More chill is noticeable in the sample from heat J-ll inocu— lated with Ca-Si than for heat J-9 inoculated with Fe-Si. The least amount of chill is noticeable in the ample from heat J-7 which contained 0.16 more silicon than J-9 or J-ll. Inverse chills is also noticeably associated with this sample. Micro-examination .R_e_sults Figures 1, 2, 3, and l. are photomicrographs illustrating typical structures found in this series of heats. The structure of Figure I. is typical of all heats of Brinell hardness around 300. The structure of J -12 showed much less interdendritic cementite and had a Brinell hardness of 269. All of these heats but J-12. were too hard to cut with the power hack saw... Figures 2 and 3 are from beats with the same final composition,but heat J-9, Figure 2 was inocu- lated with Fe-Si, and heat J-ll, Figure 3 with Ca-Si. The size and spacing of the nodules are similar and there is about the same mount of intercellular material. There are a few more nodules without ferrite rings in Figure 1 than in Figure 2. More nodules 31 are visible in Figure 1 from heat J-7 than in Figures 2 or 3. Also there is more ferrite present. The mount of intercellular material seems to be about the same. Photomicrog'aphs were taken at x1500 magnification to inves- tigate the nature of the intercellular material. In all cases there was found adjacent to the white intercellular material flake graphite, non-metallic inclusions, and what is believed to be segre- gated iron phosphide. The non-metallic inclusion is presumed to be Fayalite, 2Fe0.Si02 as this type of mineral melts at 2201°F, is med— ium to dark gray in color by reflected white light, and is color- less, light green, yellowish green, or yellowish brown in color by transmitted white light. Seemingly attached to these inclusions are flakes of graphite. The etching followed by heat tinting procedure was to distin- guish between iron pho sphide md cementite in pho sphide eutectic of cast iron. The iron phosphide should be colored darker. It is usually difficult to detect ste adite when the phosphorus content is about 0.10 to 0.12 per cent. This amount of phosphorus is as- sumed to stay in solid solution. In cast iron carbon and phosphorus are both soluble in gamma iron to about the same extent 1.7 per cent. With the high carbons of this series of heats, probably what hap— pened is that during solidification . the phosphorus was segregated to the outside of the growing austenite areas, which habit is charac- teristic of normal gray iron. After solidification it would be found in the last portion to solidify. Figure 6 could very well il— lustrate such a series of events. Two possible ways of explaining this are (one), as the concentration of phosphorus is high in these 32 areas, and carbon and phosphorus are incompatible at high concen- trations, one has to precipitate out. The graphite flakes are visibly attached to non-metallic inclusions in and near to this area. Apparently the inclusions nucleate the graphite. (Two), it is believed that the graphite nodules have an area of influence. As these areas of segregation are removed from the areas of nod- ules, it might be that the influence of the nodule is not great enough to prevent flake graphite from forming. Close examination of Figure 1 will show nodules immediately adjacent to those areas showing segregation. A possible reason for the cementite persisting is the belief that phosphorus might retard the graphitizing reaction of silicon. There is also the possibility of magnesium segregating to this area and causing the stabilization of cementite; The ino- culants of this series of heats do not appear to be affecting these areas of segregation. After the samples were polished with alcohol as the carrier for the micropolish instead of water, inclusions similar to those mentioned by Rehderwlt) were visible. The fine graphite flakes in the intercellular-type areas were more clearly defined. Fig- ure 8 and Figure 9 are photomicrographs at 50X magnification of lmetched specimens from heats J-9 and J -ll respectively. The in— tercellular—type location of the flake graphite is noticeable in both. The size of the flake yaphite seems to be about the same, but there seems to be more in J-9 than J-ll. The flake graphite in J -9 seems to be more continuous. 33 Statistical Approach _t_q Variance _i._n_ 393313 _S_i_zg The number of the nodules in Figure 8 was 129 and in Figure 9 was 121.. The diameters of the nodules were measured and tabulated in a frequency distribution. From this tabulation the means and standard deviations for the respective heats were calculated. From these results an approximation was made to the standard error of the difference of the means. The ratio of the difference between the mean dimeters to the standard error was assumed to follow a normal distribution. From this the probability of occurance of a differ- ence as large or larger than the observed difference was calculated to be 0.3182. The observed difference in mean diameters may well have been due to chance variation. This data is tabulated in Table 13. Tensile _'I_'_e__s_t Procedure gig Results 0.505 inch dimeter shoulder-type tensile bars were tested on a 100,000 pound mechanical tensile testing machine. The ultimate strengths, per cents elongation, and hardnesses are found in Table 11.. Two tensile bars were made up from heats J-7, J-9, and J-ll. These heats were chosen as heats J -9 and J ~11 had like compositions and structures. J-9 was inoculated with Fe-Si and J-ll with Ca—Si. J—7 was also inoculated with Fe-Si, but had a higher silicon con- tent. One of the two tensile bars from both heats J-9 and J-ll broke outside the two inch test length. 34 IV Discussion of Results An effort was made to control as many of the variables as possible that might influence the characteristics of the final casting. All the constituents of the charges and the late addi— tions were from the same lots. The total melt down times were essentially the same, one hour and forty minutes. The final melt down temperature was 278501? for heats J —6 through J-12, with the exception of heat J-7. The final temperature of J—7 was 2950°F. A 0.3 per cent magnesium addition as 80Ni-20Mg was added at Z7850F. The range of equivalent carbons was 4.40-4. 57. Flake graphite, nodular graphite, cementite, steadite, and ferrite were the micro- constituents present. Microstructure pp__d Chemical Compition Having as many variables as possible under control, it was assumed that different mounts and types of inoculants would produce differences in the microstructure of two like irons. The following differences might be apparent. 1. The quantity and distribution of flake graphite. 2. The number and size of nodules. 3. The mount and location of intercellular material. I... The size of pearlite spacing. Fine graphite dispersion can be attributed to undercooling. Iagnesium md cerium promote this undercooling in nodular irons. Two other factors affecting graphite dispersion that are usuale considered are time (cooling rate) and composition. For the series of heats J -6 through J -12 the magnesium addition and melt down temperature were about the same (undercooling control). The 35 cooling rates were essentially the sme, i.e., same size casting poured at the same temperature. In mine heats the same motmt of silicon was available for pick-up from the inoculant. Where the melt down compositions were the same there was a noticeable differ- ence in the microstructures obtained using different inoculants (Figures 2 and 4). However, fine flake graphite was noticeable in both, located in the last areas of the austenite to solidify. This suggests a form of carbon segregation which has been mentioned in connection with gray irons(38). An ordinary gray iron that would solidify with a white or mottled structure could be forced to solidify with a gray structure by the addition of a graphitizing inoculant. The effect of this addition would be of such an extent that it could not be attributed to the change in composition alone. The charge of J-lO was calculated to give the same composition as J-6 but 0.10 per cent more silicon as Ca-Si was added. It was hoped that this added silicon in the inoculant would force the decomposition of the interdendritic cementite. However, the same mount of late-addition silicon was picked up and the structure was the sane as J-6 (Figure 4). The mount of silicon going into the melt averaged 1.0 per cent of that available from the Ca-Si against 100 per cent of that available from the Fe-Si for the addition procedure followed in these heats. It was thought that a higher melt down silicon content to- gather with the amount of silicon from the Ca-Si usually picked up, would produce an iron with the desired structure. This was 36 substantiated in heat J-ll (Finns 3). There were now available for comparison two nodular irons of the same final composition obtained through the use of the different inoculants. The only difference in flake graphite distribution was that it appeared to be more continuous in Figure 8 (J-9). This could be a result of the greater silicon pick-up from Fe-Si. The number of nodules was about the. same, 129 (J-9) and 123 (J-ll). There was a difference in the mean size of the diameters of the nodules. Those from heat J-ll were larger. However, this difference was not significant. The mount and location of the intercellular material was the same (Figure 8 and 9). The intercellular material in J-9 and J-ll was assumed to be steadite. In the etched microphotographs of Figures 2 and 3 the segregation and intercellular material seems water in Figure 3 (J-ll). The size of the pearlite spacing was the same. Chill Qharacteristicp The addition of the inoculant Fe-Si was more effective in reducing chill for the procedure used in this investigation. The lesser chill reduction obtained with Ca—Si could be a result of the lower silicon pick-up, a carbide stabilizing effect of the calcium, or a combination of both. Tmsile mpgrties The Brinell hardnesses of J-9 and J-ll were the same, 255. J -9 inoculated with Fe-aSi had one per cent more elongation. The ultimate tensile strength of J-9 was greater. The lower ultimate tensile strength and per cent elongation of the heat J ~11 inoculated with Ca—Si might be attributed to larger intercellular areas. Fig- ures 2 and 3 indicated such a difference, while the unetched photo-'- micrographs of Figures 8 and 9 did not. V Summary and Conclusions It is essential that a @aphitizing inoculant be used in the production of nodular cast iron. Two of the inoculants that have been used are Fe-Si and Ca-Si. The references to the use of these inoculants on nodular iron stated that Ca-Si was not so effective. However, no evidence was presented in support of these statements. The method of inoculation used in this investigation has sub- stantiated the reports that Ca-Si was not so effective as Fe-Si. The inoculants were added to the surface of the molten metal. The stirring action of induced currents was utilized for mixing. A 100 per cent pick-up of silicon from Fe-Si was possible while [.0 per cent was the average pick-up with Ca-Si. When the effects of Ca-Si and Fe-Si were compared with regard to chill reduction, tensile strength, and elongation in nodular irons of the same final composition, the inoculant Fe-Si showed a greater tendency (1) to reduce chill, (2) to increase tensile strength, and (3) to give a greater per cent elongation. The size and number of nodules appeared to be the same. The Brinell hardnesses were the same. 39 Table 10 - Log of Heat J-ll Charge 27.2 pounds Hanna malleable pig 405 grams ingot iron 600 grams 27 per cent ferro-silicon 113.6 grams Ca-Si (0.5 per cent Si addition; 204.3 grams Ni-Mg (0.3 per cent Mg addition on sition 13,2 alculation (Without Si of Inoculation) C Si S P Mn 3.87 2.61 0.02 0.09 0.36 2:45 Hydrogen on 405 grams ingot iron on bottom 10 slices pig to fill crucible 3:05 Power on 18 EN 4:05 Conical chunk floating on surface. Turned over with skimmer. 4:13 Mushy chunks floating on surface of bath. Broken up with skimmer. Some kish. Added half the 27 per cent ferro-silicon. Added 8 small chunks of pig. 4: 25 Mushy chunks floating on top of bath. Broken up with skimmer. Some kish. Added rest of 2'7 per cent fem-silicon. Added rest of small pieces of pig. 4:38 Some mushy chunks floating on bath. Broken up with skinner. No kish. 4:45 mall amount of slag floating on bath. No kish. hushed to side of crucible. 27850 F. Power off. Skimmed off slag. Poured base iron sample. 4:47 2785° F. Power off. 4:48 Added 80Ni-20Mg. About 30 seconds later the reaction was over. Skinned. 4:48 Power on 15 KW. Added 4OCa—60$i. 4:49 Power off. Skinned as much slag off as possible. 4:50 Poured. Total melt down time one hour and 45 minutes.. Table 11 - Furnace Charge (Pounds) Heat Hanna Malls able 27% Fe-Si Ingot Iron Inoculant Pig Amount Type J-7 27.9 1.08 0.892 .132 Fe-Si J-9 27.2 1.01 1.68 .132 Fe-Si 51-11 27.2 1. 32 .892 . 251 Ca—Si J-12 27.. 2 1.32 .892 .124 Ca—Si 0.45 pounds Ni-Mg added to heats 3.7 through 3.12 Table 12 4- Chemical Compgsitions (Per Cent) Heat Total Melt Down Final Inoculant Carbon Silicon Silicon Amount Si Type J-6 3.72 2. 20 2.36 .40 Ca—Si J-7 3.78 2. 36 2. 90 . 41 Fe-Si J-9 3.73 2. 24 2. 71. . 1.1 Fe-Si J-lO 3.65 2.30 2.1.6 . 50 Ca—Si J -ll 3.73 2. 53 2. 7'7 . 50 Ca-Si J-lz 3. '75 2. 46 2. 56 . 25 Ca-Si O. 3 per cent magnesium added as 8OI‘Ii-20Mg Average manganese . . . . . . . . . 0.32 Calculated phosphorus . . . . . . . 0.09 Average melt down sulphur . . . . . 0.03 Average final sulphur . . . . . . . 0.018 Table 13 — Statistical Results on Modular Diameters Figure 8 Cell Figure 9 Heat J -9 Boundaries Heat J -11 '7 0-1 11 42 1-2 29 35 2-3 30 26 3—4 40 .11 4-5 ll. 129 123 2. 56 Mean 2. 62 1.15 Standard Deviation 1.15 Standard Error 0.145 Probability of Occurance 0.318 Table 14 - Tensile Progrties He at Tensile Elongation BEN Note Strength (psi) (55) J -7-2 92,700 3. 0 241 J-7-4 95,650 3. O 255 J -9-2 95,150 - 255 A J-9-4 101, 200 3. 0 255 J-11-5 92, 600 2. 0 - 269 J -ll-6 69,000 - 269 B Note A This tensile bar broke 3/4 inches below the supporting shoul- der. There was a slag inclusion located at the fracture. Note B This bar broke just below the supporting shoulder. The cause of fracture was not determined. Table 15 - MicroharLdness (DPH Numbers) 43 He at Ferrite Cementite Intercellular Ring 'Iaterial J -4 247. 0 919. 5 J -5 193. 0 919. 5 J-7 223. 5 988.0 802. 5 919. 5 988. 0 Table 16 - I-ietallogaphic Data Magnification Objective Eyepiece 50X 5. 6x0.12 NJ. 53 1001 8.0x0.20 N.A. 7.5 H.P. 1500x 85.0x1.25 N.A. 10x Bellows Setting (cm.) 46.3 '45. 5 42.8 Heat J-7, As—cast structure, etched, 1001 Inoculant: 0.4 per cent Si as Fe-Si 45 Heat J-9, As-cast structure, etched, 1001 0.4 per cent 81 as Fe-si Inoculant: Heat J-ll, As-cast structure, etched, 1001 0.5 per cent Si as Ca—Si Inoculant: Figure 4 Representative of heats J-l through J-6, etched, lOOX Inoculant: 0.4 per cent Si as Ca-Si 47 Heat J-S, As-cast structure, etched, 1001 0.4 per cent Si as Car-Si Inoculant: 0.15 per cent Mg as 80Nii-20Mg «9‘ 6': g P :9, M o. ‘ Figure 6 As-cast structure showing area of segregation around intercellular area, etched, 15001 49 Figure 7 As-cast structure illustrating intercellular area of what is assumed to be steadite, etched, 1500K 1" '~ ' 5 i' .-:..’- ‘ _. . 08"}:5. . 1‘ . . T‘ el' ~{ ‘ ' 't. T. . ~ ‘ - I . a . , t ‘ "1’. .334 in“, . . I . f L e. - Figure 8 Heat J-9, As—cast structure, unetched, 50X Inoculant: 0.4 per cent Si as Fe—Si I ‘. .- 5 _. .'. ‘ .' _ ~ 0 .. - - c “O 9 :1 ‘ . a.“ f. - ‘. ~ .5 7 “5" . ' ‘ . .‘ c . , _. ‘ . ; : . - ,. l . C“ 1. I' J. I‘ 9.. UQ' "1‘ V-g..¢. o'- ‘. . 9. '. .i ' ' ' it“ 0. . - Q' "' ‘ ‘.r ”Dr I gr 'ijc. , .' or I, l "l. ’ ;‘ . .1.“ :g I ”if" ' ‘3 ‘ Y r. 9. .. O- ' ._" . . .. ::’\.’ ‘. - . . g. . _ : . . \" A a; _ .‘h ‘3 ' “a . o" ' h_ I . f". ‘. ',,., |' ' Figure 9 Heat J-ll, As-cast structure, unetched, 501 Inoculant: 0.5 per cent Si as 08’31 I‘l' 51 52 Figure 10 Keel block lying on its side to illus- trate shrinkage Figurell Riser of keel block, inverted to show sampling locations 53 Figure 12 Center..Chill block from heat J—ll Right...Chill block from heat J-7 uuuuuuuu , , {//////////I ' Figure 13 Arrangement of equipment for melting, making additions, and pouring 2. 3. 7. 9. 54 VI. Bibliography H. Morrogh, "Nodular Graphite Structures Produced in Gray Cast Irons," Reprinted from American Foundryman, April, 1948 Albert De Sy, ”Belgian Research Advances Nodular Graphite Theory," American Foundryman, 15, Jan., 1949, p. 55 N. A. Ziegler, W. L. Meinhart, and J. R. Goldsmith, "Dilate- meter Studies of Nodular Cast Iron," American Foundrymen's Society Preprint No. 50-62 E. T. Wskowski and R. P. Dunphy, "A Graphite Nodulizing Alloy," Iron Age, 164, Sept. 8, 1949, pp. 78-9 0. K. Donoho, "Producing Nodular Graphite with Magnesium," Amarican Foundryman, 15, Feb., 1949, p. 30 Norman C. McClure, "Ferro Silicon, Calcium Silicon, and -Ieta11ic Calcium as Ladle Inoculants for Gray Iron,n Thesis for the degree of Master of Science, Michigan State College, 1949 D. E. Krause, "What's in a Name," American Foundryman, 16, 13017., 1949, p. 45 H. Morrogh, "What's in a Name," American Foundryman, 16, IIOVO, 1949, P. 44 G. E. Holdeman and J. C. H. Sterne, "Variables in Producing Nodular Cast Iron by Magnesium Treatment," American Foundry- mm, 16, 11113., 1949, P. 36 10. 12. 13. 14. 15. 16. 17. 18. 19. 55 C. 0. Burgess, "Progress Report on Nodular Iron," The Mm Kuniansky, I”Problems in Producing Ductile Iron," The Foundry, Vol. 78, No. 1, Jan., 1950, p. 62 E. M. State and B. L. Stott, "Producing Ductile Iron," The Fomidry, V01. 78, No. 7, July, 1950, p.80 A. P. Gagnebin, K. D. Millie, and N. B. Filling, "Ductile Cast Iron, " Reprint from Iron Age, Feb. , 17, 1949 T. E,__Eagan and J. D. James, "A Practical Evaluation of Ductile Cast Iron," Reprinted from Iron Age, Dec.15, 191.9 E. K. Smith, "Experiments in Nodular Iron," American Fomldryman, 16, Bees, 1949, P. 45 "Nodulizing Gray Iron with Dilute Magnesimn Alloy," Iron Age. Sept. 29, 1949, p- 55 J. E. Rehder, flDesulpl'mrizing Graerast Iron with Magnesium," Amarican Formdryman, Sept., 1950, p. 40 "Nodular Graphitic Irons," Iron Age, April 28, 1949. p. 82 ”Magnesium Treatment for Nodular Graphite in Cast Iron,” Iron Age, Feb. 21p, 1949, P. 97"99 F. R. I-‘Iorral, "Nodular Iron," The Foxmdry, March, 1950, v01. 78, NO. 3, p. 135 22. 23. 25. 300 56 Elbert De Sy, "Further Results of Belgian Nodular Cast Iron Research," Amarican Foundryman, 17, May, 1950, p. 75 D. J. Reese, "Factors Affecting Development of Ductile Iron," The Foundry, Vol. 78, No. 1, Jan, 1950, p. 58 J. E. Rehder, "Magnesium Additions and Desulphurization of Cast Irons," American Foundryman, 16, Sept., 1949, p. 33 Albert De Sy, "Further Results of Belgian Nodular Cast Iron Research," American Foundryman, May, 1950, Vol 17, No. 5 H. F. Scobie, "Nodular Iron Wain Shows We Still Have Iong Way to Go," American Foundryman, 18, Oct. 1950, p. 46 H. Morrogh and W. J. Williams, flGraphite Formation in Cast Iron and in Nickel Carbon-Cobalt Carbon Alloys,” Journal, Iron and Steel Institute 155, 1947, pp. 322-371; 156, 1947, RP. 491-496; 157, 1947, PP. 193-197 A1 Norbury and E. Morgan, Journal of Iron and Steel Insti- tute, 1936, No. 11, pp. 327-346 E. Piwowarsky, Transactions, American Foundrymen's Association, 1926, Vol. 21., p. 914 _J. T. Eash, Transactions, American Foundrymen's Association, 1941, p. 887, Vol. 49 R. A. Flinn, R. W. Kraft, "Improved Test Bars for Standard and Ductile Grades of Cast Iron," American Foundrymen's Association Preprint No. 50-1 31. 32. 33. 34. 35. 36. 38. 39. 57 Albert Do Sy, Letters to the Editor, American Fotmdryman, 15, Feb., 1949, p.60 R. E. Deas and L. T. Conradi, "Determination of Total Carbon in Pig Iron, High Carbon Iron and Nodular Cast Iron," The Foundry, 77, July, 1949 , pp. 68-9 Lundell, Hoffman and Bright, "Chemical Analysis of Iron and Steel, " p. 261 J. E. Rehder, "An Introduction to the Annealing of Nodular Iron, " American Foundrymen' 8 Society Preprint N6. 50-17 "Nodular Iron; Theory and Practice,' Metal Progress, Vol. 58 NO. 5, N070, 1950, P. 729 ”The Production of Nodular Cast Irons with Cerium— Some Practical Details," Cerium Metals Corporation A. L. Schaeffler, I'Rapid Hand Polishing of Micro Samples," Metal Progress, Vol. 46, 1944, p.285 J. W. Bolton, "Graphitization and Inclusions in Gray Iron, " Transactions of the American Foundrymen's Association, 24,, 1937, p. 467 R. M. Allen, "The Microscope in Elementary Cast Iron Metal- lurg" p. 58 58 Albert De Sy, "Some Results of Belgian Research on Nodular Cast Iron," The Foundry, Nov. 1949, p. 128 Albert De Sy, "Producing As Cast—Ferritic Nodular Iron in Heavy Sections," Metal Progress, Jan., 1950, 57, p. 79 H. Morrogh and W. J. Williams, "The Production of Nodular Graphite Structures in Cast Iron," Journal, Iron and Steel Institute, 158, 1948, p. 306; discussion 160, 1948, p. 21 E. T. Myskowski and R. P. Dunphy, "New Graphite Nodulizing Alloy," The Foundry, Oct., 1949, p. 72 F. H. Allison, Jr., "Nodular Iron a New Engineering Material," United Effort, Vol. 30, No. 1, Jan.-Feb., 1950, p. 6 A. Boyles, "The Freezing of Cast Iron," American Institute of Mechanical Engineers Transactions, 1937, Vol. 125 A. Boyles, uThe Pearlite Interval in Gray Cast Iron," Transactions of the Alnerican Fomdrymen's Association, March, 1941, Vol. 48, p. 531 A. Boyles, "Influence of Atnosphere and Pressure on Structure of Iron-Carbon-Silicon Alloys," Transactions of the Anerican Institute of Mechanical Engineers, Vol. 135, 1939, p. 376 D. J. Reese, "Economic Considerations in Producing Magesium- bearing Iron,‘r The Fomidry, Vol. 78, No. 5, May, 3:950, p. 120 Journal, Iron and Steel Institute, No. 1, 1941, Vol. 143, pp. 195- 205 ROOM 'uo'E ONLY .. . . , . . . . 4:." . . . . .. _a . . . I. . . . .. .. . . . I. p. . . . . . ... . ‘ %. taqallvl.!;(.ib ”4‘; .fl..._1L..uk.¥.-y 4.... ...v .....WCY...... rap 0’. .... . . . \.P... . A\ . e . . . ..N’. . . . Mafia-May ....LLKLI 94k}?! .o..‘.’t’~.khk%\.fl I . . . . . . .h u .. .n . as r .v. x . u . . .741 ‘_. . . .IHU. Iv, ..lr... (f. Y 3 .1. .wlu . I..... ...ot.»4..fl... ’7 qil . ...“) v. - (AV. I ' . I. . . . . . . I u . .. .. . {slurp -.ro. N1. . .. .. . . . .. . . .. . Z . . '3'... :f. ....x‘.. .. . . l t.‘ In» 1.2, . . . . . . . . , . . .. ru...’..?..~.%.x. .. a . . . I . 4 1; . . . . . .. . . .. . . . .. . Q .J a . .. VI . . . f . . .. .. . _. . . . . . . . . . . . . . n . . .s r... _. qI-J’I’O X’s th‘...l I‘.N, . ... . . . . . ., h . . . \u . 7 , .. . . . . . ..u ...). ... . . . . . "’ ID . . I 3 M ROOM USE ONLY- 30 I I l I I f I. ’ ‘ l 1 D ‘ I r x ‘I ’ I l . 1 Q o I . \. u . . .. .. I I A ... . I n 5 b l: \ It! I IV I If n I I l . , . ,2. I. . a .... .I I I . I. I 1.. .. tn F a t V . . -\ . ‘ « . I I l A I. - 4 ‘ I . d c c II ' a I n H I i \ r I 1 a p I. a" n o | . ... . a I a s.. .. I . It r I I I. I I \ I‘ . ~a V _ n, . 44.. & . . . a . .w\\.. I l K\ . ... J _ . . I an 5 h r \ . Q Fl .. . . .8. cl. . A k V l QM. . I do . 4 c. . . . | A \ I I c I I .c P . I r c A\n\lw .\ o \f 5.1 I. ‘ ‘ e . QI \. 1 I. IV .1 l - b a -I ’ l r I t .l l 1 \I L c. (M It \. _ . I ll. .5 .n n \. o . I - .. .. . , 7 l N I .I s p. u I _Iov . u \ .. fl/ . '0 1.4. I . A b . . a .P bl .14. u . J n . — . I I \ I ‘ o n n .‘ u ' Ill _ I ‘ ' c on ,5. I.“ o . . x . I I 4 .I 4 I t 1 l y K r. (I .. .60 u w . div... ..4 a Ax... I. u a . .w r/ lht\.wd a . r .I . ... . A . _ - . y e , - ....t . . a .. .-...I . . _. . v .I.» , _ O o \ .1 .a. I . . n J . . on I ...I o e l u. 1 t I I 'J . '~ - I u /< 4 I \ . \- I! o I . \ l c L N ILrL \ l I I I . (Arl/ ‘ l . «H I I. . a . J - - . . ...... F.-. . . _ . . a ... _. - .. .. . . ll...‘ - .. .. . . .(~\\.w-.f. . s .. It. AI. . . . . . . o -.. . . wally». ... 2...... an... 3.11;... ... I... ...-.. .o . . ‘7 I . I. N _ . . .0 Q . \ °\ .1 ... m I «I: ..l\.« . 'ul ... . In. a.* 5 . 1‘ c .i D r J . . . c l ’ ll. \ . . I .4. :WA 0 u 5.. n u r .. . 1 L .. l .I. .14 I. (V Ir. I... r ’ 1 I.“ Y o I4 \D... I d J \1’ I I. l . ~ - v\.\4 I I .c ) to. 7.. vls’ a ‘6 ~ I. ) . .N u . J I o a x '1 u . 0 LI.‘ . L v.1. u .. . o nlw .IIIKJG 3.9““. 0"“ \V.' 8.. I t 5 I I c K th‘k.’ \. I. .1... 4. § . . Inf v It; 05 lo. 1 I an . .. \.. \‘V Ill“ r .2. A . I. v . p I v If \ o! .m. s .0“... c. a r 90%. ... . . I“! .~ 2‘. e ...“ Th... “F . kt}. ;r..I\ s . .vl o . I n t v . q u. I. t I A . . OI l‘ w v .. \ ’l 4 . .4 ., r o I . ~ ... I 3“. . dV\ IrluV. ..." a \t b. r 1 ..N c. 1‘ Q a. XI . . . n . s . . . I . . t . _ . o . I v . l . c P l ~> - \ '1 . v i. \- . \3 I. ‘- .o, ‘»~.,l. I a... . _ p n MTITIWHILEMIW [illilllMWlfijl/I‘ITI'Es