Ductile (spheroidal graphite) iron was invented in the 1940s and quickly grew in use, replacing malleable iron and forged and welded steel in industrial applications. Austempered ductile iron (ADI), a specially heat-treated ductile iron, commercialized in the 1980s, exhibited double the strength of as-cast ductile iron for a given level of ductility. A new family of high-silicon ductile irons emerged in the early 2000s having a silicon-strengthened ferritic matrix with excellent tensile strength and machinability, but lower-than-desired low temperature properties. Worldwide, ductile iron production now exceeds 27 million metric tons per year.

The ductile iron process blends pig iron (a byproduct of the titanium oxide reduction process), recycled steel scrap, internal returns, and small quantities of silicon, copper, and other alloys to produce a “designer” alloy with properties particular to the purpose.
Engineers continue to develop new, better ductile iron alloys to allow designers to create lighter, less expensive components. Ductile iron can compete technically and economically with aluminum, steel, and titanium castings, forgings, and weldments.

BACKGROUND
Ductile iron is a very low embodied energy structural material.
Although it is a relatively lower cost structural material and exhibits a high strength-to-weight ratio and excellent recyclability, barriers exist to its wider use.
For truly lightweight components, ductile iron must be produced carbide-free in very thin (3mm) sections. High-volume, green sand foundries generally will not quote as-cast ductile iron castings with sections under 5mm because of the risk of brittle carbides forming on edges in critical sections. Carbide formers in the alloy, like Mn, Ti, Cr, Mo, V, combined with a locally low silicon content can promote the formation of deleterious carbides. Increased silicon content promotes ferrite formation and can mitigate the formation of pearlite and carbides.
Producers aim to keep manganese below about 0.35% to produce ductile iron with good ductility and machinability. But low manganese steel scrap is becoming harder to acquire as most steels have 0.60% to 0.90% Mn, and even more in the new generation of high strength automotive steels. This requires ductile iron producers to use more expensive pig iron (a low silicon, low Mn, 4.3% carbon iron alloy) in their charge.

Alloys like copper, tin, and antimony can promote pearlite in ductile iron but also reduce the ductility of the iron. Nickel, although expensive, stabilizes austenite in ductile iron and is relatively benign, being soluble in iron to 100%. It is possible that nickel could increase the ductility of ductile irons with a variety of chemical compositions and microstructures.
In this research, various techniques were explored to achieve high-strength, high-ductility, carbide-free, 3mm wall ductile iron.

EXPERIMENTAL PROCEDURES
For each composition that was tested, 3 x 28 x 100 mm “fingers” were cast in order to look for the formation of carbides in thin sections (Fig. 1).
For certain compositions, keel blocks were cast to determine the mechanical properties of the alloys. To simulate the cooling rate of a thin section, the keel blocks were cast with chills. These keel blocks were machined into ASTM E8 standard tensile bars and ASTM E23 notched Charpy impact bars for testing at -40C.

Three different types of sands were also explored—a 60-mesh crystalline silica sand, a 70-mesh non-crystalline fused silica sand, and a 70-mesh engineered non-silica sand.
Joyworks typically melts around 175 lbs. of metal in an alumina crucible comprised of around 65 lbs. of sorel, 60 lbs. of ductile iron pig, 40 lbs. of 1010 steel punchings, and alloys. The other charge alloys used in the ductile iron melts include FeSi, graphite, FeMn, iron pyrite, copper, and nickel.

Chemical analysis was conducted per ASTM E1019-18 for carbon/sulfur and ASTM E1999-18. Once the heat is treated in the tundish ladle, it is transferred into a pouring ladle. During the transfer, the iron was inoculated with 0.25% of a Fe-Si based inoculant containing cerium, having sulfur and oxygen less than 1%. From there, the molten metal was quickly poured into the molds. The pouring temperature was between 2,500F (1,370C) and 2,600F (1,430C).
Nodularity and nodule count were both determined using commercially available software. According to ASTM standard E2567, the minimum size should be 10μm. However, this standard fails to count many of the smaller nodules that form due to the increased cooling rate in thin sections. Instead, a minimum diameter of 5μm was used, as suggested for thin section ADI by Pedersen and Tiedje. After measurements on the nodules, the specimens were etched using 10% ammonium persulfate to reveal the carbides in the microstructure (Fig. 2).
Two measurements—carbide depth and flash distance—were used to further catalogue the state of the carbides and the casting (Fig. 3).

CHEMISTRY EFFECT ON CARBIDES
Table 1 shows the main alloy compositions that were studied. The compositions are based around a baseline of 3.1 wt% carbon, 3.9 wt% silicon, which yield a carbon equivalent of 4.4. This was calculated as wt% carbon + ⅓ wt% silicon. In addition, the baseline has 0.25 wt% manganese, 0.3 wt% copper, and 2 wt% nickel. This composition was selected based on previous research and showed promising results in producing thin ductile iron sections that have no carbides. As such, alloy compositions are labeled with the purposeful change from this baseline, and it is how they will be referred to throughout the report. Note that the 0.5% Mn composition includes a higher copper fraction as well.

Table 2 contains the nodule characteristics of the alloy compositions when cast in non-crystalline fused silica sand. The first of these alloys showed signs of producing no carbides in non-crystalline fused silica sand, but when cast in crystalline silica sand or an engineered non-silica sand, they did show carbides. These sands have higher heat conductivity, which caused their castings to produce carbides for most of the alloys studied. While the alloys all have high nodularity, they have inconsistent nodule counts and, for these non-fused silica sands—even with the highest nodule counts—carbides were present.

This is at odds with findings by Almanza et al. in 2021 (Effect of Cobalt Additions on the Microstructure and Mechanical Properties of As-cast Thin-Wall Ductile Iron), whose study found that increasing nodule count was strongly correlated with decreased carbide phase fractions. Data from the Joyworks study shows a much stronger correlation for the thermal conductivity of the sand and less correlation for nodule count.

However, the highest recorded nodule count, in the 0% nickel alloy, had a finger thickness of only 1.99mm, much lower than the desired 3mm sections, which would result in a higher cooling rate and thus the high nodule count and carbide presence. The other alloy that produced carbides had high manganese content at 0.44 wt%. Overall, the main factor affecting carbide formation in all the specimens was cooling rate coupled with high silicon content; 3mm sections cast in fused silica have no carbides if carbide-inducing elements are not included at high concentrations.

When all the alloys are plotted in Figure 4, no apparent correlation appears between higher nodule counts and the formation of carbides. However, the highest carbide intrusion also corresponded to the thinnest finger at 1.99mm. This makes it more difficult to determine what combination of higher nodule count or thinner fingers was responsible for the increased carbide presence without taking more data.

MECHANICAL PROPERTIES
Table 3 shows the as-cast mechanical properties of alloy compositions studied in this report. In general, they exhibit good yield and tensile strengths with good ductility—with one exception. The 0.5wt% Mn composition, for which carbides formed in the thin sections, also resulted in significantly lower elongation, which is characteristic of carbide presence in the microstructure.
These high Si alloy mechanical samples were also austempered to ASTM grade 2 standards. The overall results in Tables 3 and 4 show that, in general, these high silicon alloys can have excellent mechanical properties, contrary to some findings that silicon content is deleterious to physical properties. Interesting to note: The lowest nickel content produced the highest mechanical properties, and as the microstructural results revealed, it seems that a high Si is important for producing no carbides in 3mm sections. It was thought that Ni additions might improve the low temperature impact properties, but this was not found to be the case. The need for some nickel in the alloy remains inconclusive, and there may be another root cause in the difference between these alloys and allows for further research. Finally, the higher manganese content in 0.5% Mn resulted in poor yield strength, lowered elongation, and bad impact properties due to the presence of carbides. There was also interest in -40F Charpy impact testing; the results are shown in Table 5. The 4% Ni showed some of the lowest impact properties on par with the 0.5% Mn samples.

While the aim was to produce 3mm sections for all specimens, there is a range due to the inherent difficulties with casting thin sections in a small scale. When this is plotted in relation to the carbide depth, the most obvious relationship is that non-crystalline fused silica sand produced the most consistent results in having carbide-free thin sections. Nearly all non-silica sand and crystalline silica sand thin sections have carbides.

Flash is produced due to mismatching between the cope and the drag of a mold and can be of considerable relative size when producing thin sections. Therefore, some measurements were taken of the flash at the cross section where microstructural properties were measured. These measurements can be seen in Figure 5 and are a relatively random distribution in relation to the carbide depth. The expected relationship was that more flash acted as cooling heat sinks for the edges of the finger molds and a deeper carbide depth, but these results show that is not the case. Instead, there is mild correlation between higher flash resulting in a lower carbide depth, which needs further investigation.

With this development of ductile iron alloys not typically used, it was of interest to determine susceptibility to environmentally assisted fracture, or EAF, in the presence of water. An EAF test was proposed whereby one of the alloy test specimens was wrapped in a wet towel and mechanically tested. The composition designated Low Si was tested.
The results in Table 6 show the susceptibility to EAF remains, with the ultimate tensile strength reduced by about 100 MPa. Yield strength was unaffected, but ductility is significantly reduced.

CONCLUSIONS
The goal of this investigation was to produce carbide-free, 3mm-thick ductile iron and measure the austempered properties of the various compositions studied. Within the limits of this investigation, the following conclusions can be made:
•    It was found that carbide-free sections with good tensile properties can be achieved using high-silicon content alloys.
•    The use of non-crystalline fused silica sand molding material was found to be necessary to produce 3mm-thick carbide free ductile iron.
•    These alloys have excellent austempered mechanical properties even with a high Si content.
•    These alloys also have modest low temperature impact properties even with the addition of up to 4% Ni, but they are not high enough for some industrial applications.
•    Thin section, carbide-free ductile compositions, but it is possible that a new frame of reference for what constitutes a proper nodule needs to be produced for these thin sections. Traditional nodule measurements may be missing significant numbers of nodules that should be measured.
•    There could be more avenues for the advancement of ductile iron to a further range of applications that include taking advantage of its high strength-to-weight ratio using thin sections. CS