Sic vs. 75% FeSi: comparing pre-inoculation effects.

The advancement of thin wall iron casting for weight reduction could play a significant role in the design engineer and purchasing community. By increasing the level of magnesium recovery through silicon carbon (SiC) pre-inoculation, thin wall ductile iron castings have improved flowabiity, reduced slag, less chill and reduced shrinkage, making ductile iron more appealing to design engineers.

This article investigates the pre-inoculation effect of SIC on thin wall ductile iron castings in comparison with 75% ferrosilicon (FeSi). The areas compared include fluidity/castability, the microstructures of thin wall ductile iron test plates and the overall magnesium recovery in the process of ductile iron production.

The base iron used in the experiment was produced in a 300-lb medium frequency coreless induction furnace from a charge consisting of nodular grade pig iron, ductile iron returns, AISI 1010 steel punchings and low sulfur carbon risers. The furnace additions were SiC containing 67.9% Si and 30% C or 75% FeSi containing 76% Si. The hypereutectic base iron was tapped from the furnace at approximately 2760F (1516C) into the flow-through magnesium-reaction chamber, containing treatment additions.

Experimental Series

The first series of experiments was done to set the baseline of iron quality, using 1.875 lb 75% FeSi as the primary furnace addition. The flow-through technique (Fig. 1) was used to treat the base iron with magnesium ferrosilicon (MgFeSi)-masteralloy, which contained 3.65% Mg and 45.49% Si. In these series, MgFeSi and the FeSi inoculant were placed in the flow-through reaction chamber. No late additions of any inoculant (stream or mold) were used.

The second series of experiments investigated the pre-inoculation effect of 2.15 lb SiC furnace additions, ensuring the same silicon content in the base iron as when FeSi was used. The standard additions of 4.5 lb MgFeSi and 1.875 lb 75% FeSi were placed in the flow-through reaction chamber. The recovery of magnesium was found to be considerably higher here than in the baseline heats, leading to a higher than anticipated residual magnesium level of approximately 0.039%.

In the third and the fourth series of experiments, the pre-inoculation capabilities of SiC were investigated using the same amount as in series two. However, the quantity of MgFeSi was reduced by 11% in the third series and by 22% in the fourth series in an attempt to produce a final iron with a residual magnesium level closer to that achieved in the first series with the FeSi pre-inoculant. For consistency, each series of experiments was run at least three times and the results are an average of at least three heats.

Test castings were poured in nobake, uncoated sand molds, assembled of seven 7.25-in diameter, 0.75-in. thick cores. A pouring cup was glued on the top of the assembly. The typical test casting consisted of three levels of plates, creating an "experimental tree." Each level was comprised of four 2-in, long and 1-in, wide plates with four varying thickness levels (0.039 in., 0.059 in., 0.079 in, and 0.118 in.).

The iron was tapped into the flow-through chamber containing MgFeSi and FeSi and then teemed into the ladle for an average treatment time of 31 sec. The treated iron was then poured into three experimental molds, Average pouring time of the first mold was 4 sec., while the average pouring time of all three molds was 21.5 sec.

Simultaneously with pouring and thermal analysis and right-after treatment, chill buttons were poured to evaluate the chemical composition of the base and treated ductile iron via spectrometer. The cooling curves of base iron that were obtained from the thermal analysis system were used to study solidification behavior of the base irons pre-inoculated with FeSi or SiC.

The average values of residual magnesium obtained are given in Fig. 2. The heats using SiC and standard additions of MgFeSi-master alloy averaged the highest level of residual magnesium, 0.039%. The experimental heats using FeSi pre-inoculant and standard additions of MgFeSi residual magnesium averaged 0.032%. Even when using 11% less MgFeSi, the iron pre-inoculated with SiC still contained higher residual magnesium (0.033%) than in heats with FeSi and standard additions of MgFeSi. In heats pre-inoculated with SiC and treated with 1.4% FeSiMg-masteralloy (reduced by 22%), the residual magnesium content was 0.029%.

Therefore, magnesium recovery was the highest (61.7%) in experimental heats using SiC pre-inoculant and standard FeSiMg-masteralloy additions, up from 55.3% when the iron was pre-inoculated with FeSi with the same amount of FeSiMg-masteralloy. When the SiC pre-inoculant and reduced additions of the FeSiMg-masteralloy were applied, the magnesium recovery observed was still higher (58% for both cases) than in those heats using FeSi and standard FeSiMg-masteralloy additions (55.3%).

Test Results

Castability--The castability of the iron plates was studied using a ruler. The distance from the ingate to furthest end of each plate was considered as the castability/fluidity criterion. The average results of castability in thin wall ductile iron using FeSi and SiC and treated with standard and reduced additions of FeSiMg-master alloy, are shown in Fig. 3. Because the experimental ductile iron plates with 0.059-in., 0.079-in, and 0.118-in. thickness were completely filled in all tests, the main emphasis was the castability evaluation of the 0.039-in, plates.

Better castability was found in those heats using SiC with both reduced FeSiMg-masteralloy additions and in heats using FeSi with standard FeSiMg-masteralloy additions. Lower castability was found in the heats using SiC pre-inoculant with the standard additions of FeSiMg-masteralloy, where the residual magnesium content was the highest at 0.039%.

Chill Depth--The thin wall plates at each level were separated from the riser and marked according to level and experimental tree identification. Plates that were taken from the middle level of each experimental tree were broken in half to study chill depth on the fractures. The chill area of each plate, generally located at the edges of the fracture was measured with a rule as a length of chill from the edge toward the center, and calculated in percentage as a ratio of the chill to the whole fracture length.

It was found that all 0.039-in. plates had white fracture--evidence of complete chill--regardless of materials used in the furnace, final chemistries, pouring temperatures or amount of FeSiMg-masteralloy additions used. Similar to the castability results, the chill depth in 0,059-in, plates, 0.079-in. plates and 0.118-in, plates was the lowest when SiC was added in the furnace with 1.6% or 1.4% FeSiMg additions, or when FeSi was used with the standard FeSiMg additions.

The heats where SiC is used in the furnace with the standard additions of FeSiMg-masteralloy, which resulted in higher residual magnesium, also tended to have a higher concentration of iron carbides. Chill depth was subsequently 76% in 0.059-in, plates, 52% in 0.079-in. plates and 32% in 0.118-in, plates, as expected. This may be explained as a result of excessive residual magnesium serving as an iron carbide stabilizer in ductile iron.

Microstructural Analysis--The same plates that were taken for chill tendency evaluation were used for microstructural analysis. Plates taken from the middle level of each experimental tree were stacked together, in sequence, according to thickness (thinnest to thickest). This set, representing each experimental tree, was then mounted in bakelite premolds, polished and marked for identification number.

The microstructure analysis was conducted utilizing both an optical microscope and an imaging system. The first stage was to determinate graphite morphology at magnification X200 on an unetched surface using the image analysis system. The periphery and center of each plate were studied, evaluating the area of graphite, nodule count and percent nodularity.

The second stage comprised of evaluation of metallic matrix at X200 using optical microscope. For this evaluation, all mounted sets were etched for 3 sec. The ferrite/pearlite ratio and iron carbide percentage were evaluated at the periphery and the center of each sample.

Area and Shape Factor of Graphite and Nodule Count--Area of graphite, nodule count and shape factor of graphite were evaluated during the first stage of microstructural analysis. The second stage comprised of evaluating metallic matrix--iron carbides and ferrite/pearlite ratio.

Obtained results showed that the area of graphite for all series of experiments was the lowest in 0.039-in, plates (7.39-10.1%) and highest in 0.118-in. plates (9.08-11.6%). At the same time, the nodule count in 1-mm plates was the highest (1134 to 865 nodules/mm2) and lowest in 0.118-in, plates (867 to 658 nodules/mm2). Thin wall ductile iron plates pre-inoculated with FeSi and SiC and treated with standard MgFeSi additions had the highest nodule count in comparison with the SiC heats with reduced MgFeSi additions. Nodularity in all samples, regardless of plate thickness, type of pre-inoculant used or amount of MgFeSi added, was more than 95%.

Percentage of iron carbides in 0.039in. plates in all series of experiments was much higher (57 to 42%) than the other plates. An inverse chill along the center of fracture was found in all 0.079- and 0.118-in, plates. The percentage of iron carbides was the highest in plates pre-inoculated with SiC and treated with standard MgFeSi additions, due to the high levels of residual Mg. The lowest iron carbides content was found in heats pre-inoculated with SIC and treated with 1.6% and 1.4% MgFeSi additions.

The percentage of ferrite in metallic matrix was the highest in 0.118-in. plates and the lowest in 0.039-in, plates, regardless of which pre-inoculant was used. In plates pre-inoculated with 75% FeSi, the percentage of ferrite was higher than the thin wall plates pre-inoculated with SiC and treated with the same amount of MgFeSi. All experimental heats where iron was pre-inoculated with SiC and treated with reduced FeSiMg-masteralloy additions had approximately the same ferrite content as observed in ductile iron plates preinoculated with 75% FeSi and treated with the standard FeSiMg additions.

Analyzing the Results

Based on the results of the study, pre-inoculation with SiC in production of thin wall ductile iron plates provides better magnesium recovery, which can lead to less chill tendency, high castability, less slag and reduced shrinkage, when the residual magnesium content is at the same level as in iron pre-inoculated with FeSi.

Residual magnesium level and calculated magnesium recovery are much higher in those heats utilizing SiC as furnace additions. With the same magnesium alloy additions, the SiC heats averaged 21% higher residual magnesium (0.032% vs. 0.039%). Those heats produced with SiC but using reduced MgFeSi additions (1.6%) still had higher residual magnesium levels than the base FeSi heats, therefore making it possible to reduce consumption of magnesium ferrosilicon.

Heats that were pre-inoculated with SiC and treated with lower alloy additions, showed high castability and a low percentage of iron carbides. Trial heats using SiC and higher MgFeSi additions exhibited the greatest amount of iron carbides due to the considerably higher residual magnesium levels. In all plates preinoculated with SiC, a higher percentage of pearlite was found in comparison with those using 75% FeSi.

For More Information

"A New Approach to Ductile Iron Inoculation," T. Skaland, 2001 AFS Transactions (01-078), AFS, Des Plaines, IL.

"Iron Inoculation: An Overview of Methods," Y. S. Lerner and M. V. Riabov, MODERN CASTING, June 1999.

[FIGURE 2 OMITTED]

About the Authors

Kenneth W. Copi is plant metallurgist at International Agile Manufacturing, Statesboro, Georgia.

Yury S. Lerner is a professor in the Industrial Technology department at the Univ. of Northern Iowa, where Nick S. Laukhin is a graduate student.