STEEL FOUNDERS’ SOCIETY OF AMERICA TECHNICAL SERVICE REPORT #104

ANALYSIS OF INCLUSION SOURCES IN BARS CAST FROM ACID MELTED CARBON STEEL

Published by the

STEEL FOUNDERS’ SOCIETY OF AMERICA

Mr. Malcolm Blair Technical and Research Director

August, 1990 IDENTIFICATION OF INCLUSION SOURCES IN BARS CAST FROM ACID MELTED CARBON STEEL C.E. Bates and John Griffin ABSTRACT

Defects from an SFSA member company were examined in the current techni- cal effort to determine the principal source of macro-inclusions. The majori-

ty of the inclusions in the samples examined in this study were composed of deoxidizers and deaxidation products. Very little iron or manganese oxide were present to suggest reoxidation during pouring. No mold erosion products were present since the were poured in graphite molds. Typical micro- graphs of foreign debris are presented.

1 IDENTIFICATION OF INCLUSIONS SOURCES IN BARS CAST FROM ACID MELTED CARBON STEEL

INTRODUCTION Oxide macro-inclusions have always been troublesome in steel castings. A general review of the steel production process suggests a number of possible sources of nonmetallic inclusions. Among these are: 1. furnace refractories melted or fluxed during meltdown,

2. metal oxides produced during oxygen blowing, 3. oxidation of the metal and deoxidizers including silicon and manganese during tapping, 4. formation of aluminum deoxidation products in the ladle, 5. fluxing and erosion of ladle refractories,

6. reoxidation of the metal during reladling, 7. fluxing and erosion of bottom pour nozzles and stoppers, 8. reoxidation of metal during pouring into the mold cavity,

9. reoxidation of metal by reaction with mold binder decomposition products)

10. fluxing of mold refractories by oxides carried into the mold, 11. loose sand on cores and in mold cavities, and 12. thermal expansion defects resulting in loose sand falling into the molten metal during pouring. These inclusions can be classified by their source of oxygen which might include air, water in refractories, molding sand, and oxidized slag. Mechani- cal entrapment of oxides such as refractories and sands is also a possibility. Mechanical entrapment is an obvious source and has often been cited as the major source of oxide macro-inclusions. However, many sources of inclusions exist, and a summary is presented in Table I. This table is a compilation of

2 information presented in References 1-20. The major inclusion sources include reoxidation of metal and entrapment of existing oxides in the metal, ladles and molds, Reoxidation

There is substantial evidence that reoxidation is a major cause of oxide macro-inclusions. Approximately 425 defects from carbon and low alloy steels were removed and examined by optical and scanning electron microscopic tech- niques to determine the nature and form of oxide nonmetallic materials pres- ent, Castings came from a wide variety of practices including acid and basic melting furnaces, large and small pouring ladles, high silica and high alumina refractory ladle linings and poured in green (clay-water bonded) sand molds and chemically bonded sand molds.(24)

The results are illustrated in Figure #1. Approximately 80% of the castings defects consisted principally of reoxidation products and 15% con- sisted of eroded sand. Slag, refractory and deoxidation products were rela- tively minor contributors. Reoxidation is considered to be the reaction of elements in steel with oxygen after the steel has been deoxidized. This oxygen may come from the air, reactions with slag or refractories, or the reduction of water from air, refractories, or the mold. The formation of deoxidation and reoxidation products is schematically illustrated in Figure 2.(16) Deoxidation products are small, typically 10 microns or less in diameter, and are composed almost exclusively of alumina

(Al2O3) in aluminum deoxidized carbon and low alloy steels. Reoxidation products generally consist of alumina particles in single or multiphase glob- ules formed by the oxidation of aluminum followed by silicon, manganese and iron (16,21). The matrix in which the alumina particles lie is typically rich

3 in manganese and silicon oxides, but in extreme cases may contain considerable amounts of iron oxide. The presence of iron oxide may cause gas holes as the iron oxide is reduced by carbon to produce carbon monoxide gas during solidi- fication. Typical oxide macro-inclusions visually appear to be green to white pow- der. The inclusions consist of a mixture of corundum crystals, silica, and manganese oxides. The presence of corundum indicates that the inclusion began as a reaction product between oxygen and dissolved aluminum. Most refracto- ries are composed of mullite or lower alumina content materials and are not expected to produce corundum.(1,5,6,8,20,19,20)

Ladle Refractory Attack Many investigators have concluded that ladle refractories are a major source of nonmetallic inclusions, Vingas and Zrimsek reported the main sources of inclusions in castings to be ladle refractories, ladle slags, oxidation of the deoxidizers, and eroded sand grains.(2,19) Lyman and Boulger similarly concluded that the major sources of ceroxide inclusions were reoxidation and refractory materials.(1) Castings poured from ladles lined with high alumina refractories contained more alumina inclu- sions than castings poured from ganister linings suggesting that alumina was being eroded by the molten metal. It is certainly possible for liquid metal and the contained oxides to attack refractories. Liquid steel tapped at 1590 °C (2900 °F) or higher can produce a molten silica-rich phase in alumina-silica refractories containing less than 70% alumina. Refractories containing less than 70% alumina have a simple eutectic that melts at 1590 °C (2900 °F). Progressively higher temper- atures produce progressively more liquid phase. The liquid lowers the refrac-

4 tory strength and can cause sagging or result in refractory erosion during filling or pouring from a ladle.

Mold Related Inclusions There are several sources of mold related inclusions as metal enters the mold cavity. Obvious sources include poor housekeeping and molding practices that produce sand from loosely compacted molds, brittle sand, loose sand from rough spots on patterns, debris from overhead conveyors, or mold crushes caused by rough handling, improperly sized core prints, and improper core placement. Correcting these problems lies more in improving the work prac- tices than in altering the molding or pouring practices. However, Sanders has shown that inclusion defects cannot be eliminated solely by "good housekeeping practices". Inclusions were present in steel poured in clean graphite molds.

(22) Silica sand is a common ingredient of nonmetallic inclusions, although rarely found by itself. Lyman and Boulger (1) and Caine (23) analyzed inclu- sions and found many to contain free sand. Excluding the obvious sand grains, the nonmetallic material was found to contain about 53% silica, 20% alumina,

20% manganese oxide, and 10% iron oxide.(1) Silica sand was a major compo- nent of the inclusions and could be picked up either by direct erosion or by having the reoxidation products stick to the sand and then be pulled off by the flowing metal. Whatever the mechanism, free silica sand was a major constituent of the inclusion and was found blended with mixed oxide phases.(1) Air contained in the mold cavity and water in clay bonded sand molds provide opportunity for oxidation as the metal fills the mold cavity. Since silica is wet by iron oxide, mold erosion and inclusion formation may be aggravated by oxidation which produces iron oxide and manganese oxide.(7) These materials wet and adhere to the sand, and the agglomerate can be eroded off by the

5 flowing molten metal. Iron and manganese oxide dissolve silica to form a fluid iron silicate melt which may aggravate erosion.

CURRENT STUDY During the course of several heats of steel, a sample of the steel was poured into a graphite mold to produce a bar approximately 4" diameter by 9" long. Each bar was poured at the approximate Middle of the heat. The purpose of pouring the bar in the graphite mold was to eliminate sand as a source of inclusions. The bars were then put in a lathe and turned to remove the inclu- sions. The inclusions were examined with an optical and scanning electron microscope to determine their nature and probable source. All of the heats were acid melted with a carbon content of about 0.5%. Oxygen was blown into the furnace to reduce the carbon to a value of about

0.20-0.24% and then the heat was blocked. After adjusting the composition, the metal was deoxidized with 6 lbs per ton Hypercal. when higher impact prop- erties were required, and the remaining heats were deoxidized with a combina- tion of aluminum, Hypercal, and Calsibar. The weight of the deoxidizing additions was calculated to add approximately 0.075% aluminum to the metal.

The castings are normally poured into phenolic urethane bonded sand molds although all of the test bars were poured in graphite molds. The deoxidizers were placed in the ladle during tapping when the ladle was approximately 1/3 full. The furnace was lined with 70% alumina brick and the bottom was prepared with silica mixed with sodium silicate. Ladles were lined with 70% alumina on both the side wall and bottom using 60% alumina mortar. The test bars were poured from the ladle when about half way through the heat. After cooling to room temperature, the bars were turned in a lathe to remove metal and the dust collected and placed in an envelope.

6 Samples of dust from each of the envelopes were placed in a tube and the tube was filled with epoxy and evacuated to remove all air bubbles. The epoxy was cured with an application of a small amount of heat under vacuum to firmly hold the oxides for metallographic polish and SEM specimen preparation. The debris was examined in a scanning electron microscope, and energy dispersive x-ray (EDX) analyses of the debris were obtained.

RESULTS AND DISCUSSION A typical micrograph of dust sample #765 is illustrated in Figure 3 at a magnification of 60X. Numbers beside the particles indicate the specific areas analyzed by EDX. The spectrum of area #1 in Figure 2 is illustrated in Figure 4. Particle #1 consisted almost completely of iron. Only traces of oxygen, silicon, and magnesia were present. The EDX spectrum from area #2 is illustrated in Figure 5. This particle consisted almost exclusively of alumi- na as seen by the EDX spectrum. A micrograph of dust sample #810 is illustrated in Figure 6 at a magnifi- cation of 60X. Again, the numbers and arrows indicate areas analyzed by EDX. Area #1 in Figure 6 was composed almost exclusively of alumina, although a small amount of calcium and manganese were present as shown by the spectrum in

Figure 7. Area #2 was similarly composed of alumina and oxygen with traces of silicon and calcium. (See Figure 8)

The appearance of dust from sample #747 is illustrated in Figure 9. The debris particles illustrated in Figure 9 were composed almost completely of calcium oxide, as seen in the EDX spectrum illustrated in Figure 10.

The appearance of a dust sample from #757 is illustrated in Figure 11. Three areas in Figure 10 were analyzed in the spectrum and are presented in

Figures 12, 13, and 14, respectively. Area #1 appeared to be a complex oxide

7 composed of aluminum, silicon, manganese and iron. The presence of the manga- nese suggests some reoxidation occurring during pouring. Area #2 was composed almost exclusively of alumina with small amounts of calcium and manganese oxide. The light area labeled #3 in Figure 10 was iron. The appearance of the dust sample from specimen #789 is illustrated in Figure 15. Three areas were chosen for EDX analysis and the results are shown by the spectra in Figures 16, 17 and 18. Area #1 again was composed almost exclusively of iron. Area #2 was composed of principally iron with some alumina silica and oxygen. Area #3 was composed principally of alumina with small amounts of calcium and iron.

SUMMARY AND CONCLUSIONS The majority of the inclusions in the samples examined in this study were composed of deoxidizers and deoxidation products. Very little iron or manga- nese oxide were present to suggest reoxidation during pouring. No mold ero- sion products were present since the castings were poured in graphite molds. It is recommended that additional studies of this type be conducted where sections of metal containing the inclusions are removed, mounted and polished to be sure the entire sample of foreign material is examined, rather than only the material collected as dust during .

SRI-MMER-90-792-7048-000.F

8 USE OF THIS REPORT AND INFORMATION CONTAINED THEREIN Publicity This report and the information contained therein is the property of the individual or organization named on the face hereof and may be freely dis- tributed in its present form. We do ask, however, that no advertising or publicity matter, having or containing any reference to Southern Research Institute, shall be made use of by anyone, unless and until such matter shall have first been submitted to and received the approval in writing of the Institute. (The Institute does not usually approve any type of endorsement advertising.)

Limitation of Liability Southern Research Institute has used its professional experience and best professional efforts in performing this work. However, the Institute does not represent, warrant or guarantee that its research results, or product produced therefrom, are merchantable or satisfactory for any particular purpose, and there are no warranties, express or implied, to such effect. Acceptance, reliance on, or use of such results shall be at the sole risk of Sponsor. In connection with this work, Southern shall in no event be responsible or liable in contract or in tort for any special, indirect, incidental or consequential damages, such as, but not limited to, loss of product, profits or revenues, damage or loss from operation or nonoperation of plant, or claims of customers of Sponsor. Report No.: SRI-MMER-90-792-7048-000.F

To: Steel Founders' Society of America Cast Metals Federation Building 455 State Street Des Plaines, IL 60016

Date: August, 1990

9 BIBLIOGRAPHY

1. W. Stuart Lyman and F.W. Boulger, "An Investigation of Factors Producing the Ceroxide Defect on Steel Castings: Part I--Composition and Sources,' SFSA Research Report NO. 48, February 1961. 2. George J. Vingas, Arthur H. Zrimek, and LeRoy Carney, "Ceroxide Origins, A Tracer Program," AFS Transactions, Vol. 72, pp 1-8, 1964. 3. M.C. Ashton and S.G. Sharmon, "Control of the As-Cast Quality of Steel Castings," AFS Transactions, Vol. 90 pp 719-728, 1982. 4. INTERNATIONAL ATLAS OF CASTING DEFECTS pub by Am. Foundrymen's Society, pp 269, 275, 280, 283, 288 and 290, 1974. 5. R.A. Flinn, L.H. Van Vlack and G.A. Colligan, "Macro-inclusions in Steel Castings, "AFS Transactions, Vol. 74, pp 485-512, 1966. 6. R.E. Preece, "Identification of Surface and Subcutaneous Marco-inclusion," AFS Cast Metals Research Journal, pp 12-20, September 1965. 7. G.A. Colligan, L.H. Van Vlack, and R.A. Flinn, "The Effect of Temperature and Atmosphere on Iron-Silica Interface Reactions," AFS Transactions, Vol. 66, pp 452-458, 1958. 8. L.H. Van Vlack, R.A. Flinn and G.A. Colligan, "Nonmetallic Macro-inclusion Causes in Steel Castings Deoxidized with Aluminum," AFS Transactions, Vol 74, pp 132-135, 1966. 9. L.H. Van Vlack and R.A. Flinn, "Reactions Between Refractories and Molten Steel Containing Aluminum," AFS transactions, Vol 68, pp 136-144, 1960.

10. R.A. Flinn and L.H. Van Vlack, "Deoxidation Defects in Steel Castings," AFS Transactions, Vol 67, pp 295-299, 1959. 11. G.R. Fitterer, J.W. Linhart, B.B. Rosenbaum, J.B. Kopect, S. Poch and W.G. Wilson, "Acid Open Hearth Slag Fluidity and Its Significance," Bulletin #1 of Acid Open Hearth Research Association, Inc., September 1945. 12. W.H. Sutton, J.C. Palmer and J.R. Morris, "Development and Evaluation of Improved High Temperature Ceramic Foams for Filtering Alloys," Sixth World Conference on Investment Casting, Washington, D.C., Paper No. 15, October 10-13, 1984.

10 13. W.H. Fettweis and R.K. Nangia, "An Engineering Blueprint for Upgrading and EAF Shop by Implementing EBT Technology and a. Clean Steel Practice," presented at the ISS 42nd Electric Furnace Conference held in Toronto in December, 1984, Iron & Steelmaker, pp 28- 35, May 1985. 14. R.A. Flinn and G.A. Colligan, "Macro-inclusions in Steel Castings, " Report of Research Project Sponsored by AFS Training & Research Institute Directed by the Research Committee of AFS Steel Division published by Am. Foundrymen's Society, Des Plaines, IL., 1966 15. J. Savage and J.M. Middleton, "Final Report on Mold Erosion," Journal B.S.C.R.A., pp 17-26, No. 68, August 1962 16. J.W. Farrell, P.J. Bilek and D.C. Hilty, "Inclusions Originating from Reoxidation of Liquid Steel," Electric Furnace Proceedings, pp 64-88, 1970. 17. Olga Repetylo, Michel Olette and Paul Kozakevitch, "Deoxidation of Liquid Steel with Aluminum and Elimination of the Resulting Alumina," Electric Furnace Proceedings, pp 7-11, 1966.

18. J. Tsubokuro, I.D. Summerville and A. McLean, "Factors Influencing the Effectiveness of Tundish Part III, " Iron & Steelmaker, July 1985, pp 48-50. 19. Arthur H. Zrimek and George J. Vingas, "Inhibitors for Elimination of Ceroxides in Steel Castings," Modern Castings, Vol #43, pp 203-208, 1963. 20. J. Brokloff, H. J. Chao, L.H. Van Vlack and R.A. Flinn, "Macro-inclusion Studies in Steel Castings," AFS Transactions, Vol 71, pp 783-790, 1963.

21. L.J. Heaslip, A. McLean, and I.D. Sommerville, "Chemical and Physical Interactions During Transfer Operations, ", Vol 1, pp 1-122, 22. C.A. Sanders, H.J. Heine and R.F. Marande "Shooter, Ceroxide, Cermet or Slag," AFS Transactions, Vol 69, pp 340-44, 1961. 23. G.J. Vingas and J.B. Caine, "Causes and Mechanics of the Scum Defect," reprinted from Foundry, June 1959. 24. John A. Griffin and Charles E. Bates, "Development of Casting Technology to Allow Direct Use of Steel Castings in High Speed Machining Lines, SFSA, Research Report No. 100, May, 1987.

11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30