In the production of heavy-section ductile iron castings, the solidification time increases significantly with increasing wall thickness. This prolonged solidification leads to the degeneration of nodularization and inoculation, as well as the enrichment of anti-spheroidizing trace elements. As a result, the graphite morphology from the surface to the core of the casting changes from spheroidal to compacted, vermicular, exploded, and chunky forms, often described as a “sandwich” phenomenon. In thick sections, graphite nodules grow excessively large and may develop “hairy” or “tailed” appearances. Additionally, unusual graphite distributions such as aligned nodules and segregation of graphite have been observed. These defects seriously impair the nodularity in the heavy-section regions of ductile iron castings. Current national and European standards for ductile iron castings rely on separately cast or attached test blocks to assess mechanical properties, but these test blocks often fail to reflect the true microstructure at the thickest parts of the actual casting. In our foundry, a wind power hub casting weighing approximately 15,000 kg with a main wall thickness of 90 mm, maximum wall thickness of 150 mm, and minimum wall thickness of 40 mm exhibited only 40% to 75% nodularity in its heavy sections (100–150 mm wall thickness), accompanied by abundant chunky graphite. Meanwhile, the attached test block passed all required tests. This discrepancy arises because, during pouring, the austenite shell can be broken under thermal currents, promoting abnormal graphite growth. Elements such as Ca, Sr, and Ce accumulate at the austenite shell interface, changing interfacial energy and shifting graphite growth mode, ultimately leading to chunky graphite. Slow cooling also enhances element segregation, particularly of Ce, causing strong graphite branching tendency even though the cores still resemble spheroids. For hypoeutectic ductile iron castings, slow cooling causes spheroidal graphite to degenerate; for eutectic compositions, graphite remains spheroidal even at slow cooling rates; for hypereutectic ductile iron castings, graphite becomes rounder and finer. However, when rare earth elements are present in the nodulizer, chunky graphite is more likely to form. Chunky graphite is a branched, degenerate form of graphite. Therefore, this study investigates different process methods to improve the nodularity and graphite morphology in heavy sections of ductile iron castings.
1 Experimental Procedure
1.1 Test Tooling Design
To simulate large ductile iron castings, a test tooling mold was designed with dimensions of 500 mm × 500 mm × 500 mm and a draft angle of 5° (increasing wall thickness direction). A bottom-gating system with a sprue diameter of 60 mm was adopted, as illustrated in the design concept. An insulating riser with a diameter of 180 mm was placed at the top central region. The tooling allowed us to evaluate the microstructural evolution at the core of a heavy section under controlled conditions.
1.2 Experimental Schemes
Three tons of charge materials were melted in an electric furnace using 80% high-purity pig iron, 10% carbon steel scrap, and 10% wind power return scrap. The carbon equivalent was controlled in the range of 4.2–4.4. The tapping temperature was 1490–1510°C, and the pouring temperature was maintained at 1370–1380°C. The total pouring time was controlled within 15 minutes. Five different process schemes were designed to investigate their effects on the nodularity of heavy sections. The target compositions are listed in Table 1.
| Scheme | C | Si | Mn | P | S | Sb |
|---|---|---|---|---|---|---|
| Scheme 1 (reference) | 3.92 | 0.80 | 0.10 | 0.002 | 0.012 | — |
| Scheme 2 | 3.85 | 0.85 | 0.12 | 0.0021 | 0.013 | 0.005–0.006 |
| Scheme 3 | 3.86 | 0.81 | 0.11 | 0.023 | 0.014 | — |
| Scheme 4 | 3.83 | 0.82 | 0.12 | 0.024 | 0.012 | — |
| Scheme 5 | 3.84 | 0.83 | 0.14 | 0.002 | 0.013 | 0.005 |
Scheme 1 used the conventional nodulizing treatment for a certain type of wind power hub: 1.2% mixed rare earth nodulizer, covered by 0.1% high-calcium barium inoculant, then 0.4% high-calcium barium inoculant added by the tundish method, followed by 0.15% sulfur-oxygen instantaneous inoculation.
Scheme 2 was based on Scheme 1 but reduced the carbon content from 3.85%–3.95% to 3.80%–3.90%, and additionally placed 0.006% antimony into the non-nodulizing pit. All other parameters remained the same.
Scheme 3 employed 1.2% yttrium-based heavy rare earth nodulizer, with 0.1% high-calcium barium covering, 0.4% high-calcium barium added by tundish method, and 0.15% sulfur-oxygen instantaneous inoculant.
Scheme 4 used a combination of 0.9% mixed rare earth nodulizer plus 0.3% yttrium-based heavy rare earth nodulizer, with the same covering and inoculation procedures as Scheme 3.
Scheme 5 used 1.0% mixed rare earth nodulizer plus 0.2% yttrium-based heavy rare earth nodulizer, plus 0.005% Sb, with the same covering and inoculation procedures as the previous schemes.
1.3 Sampling and Testing
After pouring, samples were cut according to the sampling diagram. Metallographic specimens were taken from the central region and the side wall region of a 70-mm-thick slice. The nodularity, graphite size, and matrix structure were examined. Mechanical properties, including tensile strength, yield strength, elongation, hardness, and impact toughness, were also measured for the verification castings.
2 Results and Discussion
2.1 Chemical Composition of the Test Samples
The final chemical compositions of the five schemes are listed in Table 2. The carbon content varied slightly due to the different additions and melting losses. Antimony was present only in Schemes 2 and 5, as intended.
| Scheme | C | Si | Mn | P | S | Sb |
|---|---|---|---|---|---|---|
| Scheme 1 | 3.75 | 1.75 | 0.18 | 0.019 | 0.008 | 0.001 |
| Scheme 2 | 3.62 | 1.77 | 0.17 | 0.020 | 0.009 | 0.005 |
| Scheme 3 | 3.68 | 1.80 | 0.20 | 0.018 | 0.008 | 0.001 |
| Scheme 4 | 3.67 | 1.82 | 0.21 | 0.020 | 0.007 | 0.001 |
| Scheme 5 | 3.66 | 1.84 | 0.19 | 0.022 | 0.006 | 0.005 |
The carbon equivalent (CE) was calculated using the well-known formula:
$$CE = C + \frac{Si}{4} + \frac{P}{2}$$
For the final compositions, the CE values are given in Table 3. It can be seen that Scheme 1 has the highest CE, while Schemes 2 and 5 have lower CE values due to reduced carbon and increased silicon. Lower carbon equivalent tends to reduce the amount of exploded graphite in heavy sections, as will be discussed later.
| Scheme | CE (wt%) |
|---|---|
| Scheme 1 | 4.29 |
| Scheme 2 | 4.13 |
| Scheme 3 | 4.19 |
| Scheme 4 | 4.19 |
| Scheme 5 | 4.16 |
2.2 Metallographic Results
Metallographic observations were performed on samples taken from the core of the 500 mm × 500 mm × 500 mm test block. Figure 1 shows the microstructures obtained by Scheme 1 (conventional) and Scheme 5 (the best improved process). The quantitative results of nodularity and graphite size are summarized in Table 4.
| Scheme | Nodularity (VI+V) (%) | Graphite size (μm) |
|---|---|---|
| Scheme 1 | 35 | 4–5 |
| Scheme 2 | 75 | 5 |
| Scheme 3 | 85 | 4–5 |
| Scheme 4 | 87 | 5 |
| Scheme 5 | 90 | 5 |
Scheme 1, which was the normal production process for the wheel hub, showed a nodularity of only 35% at the core of the 500 mm test block. The graphite morphology was severely degenerated, with abundant chunky graphite and exploded graphite. This confirms that the normal process is inadequate for thick sections, even though the attached test block may meet specifications.
Scheme 2, which involved a slight carbon reduction and the addition of 0.005% Sb, improved the core nodularity to 75%. The lower carbon content helped reduce exploded graphite, while the antimony addition suppressed the formation of chunky graphite. However, due to the extremely long solidification time, local nodulization and inoculation decay still occurred, limiting further improvement.
Scheme 3, using yttrium-based heavy rare earth nodulizer alone, achieved a nodularity of 85%. The heavy rare earth elements provided better resistance to nodulization decay. However, the graphite nodules were larger and less spherical, and the enrichment of rare earth elements resulted in the formation of a small amount of pearlite, which would lower low-temperature impact toughness.
Scheme 4, a combination of mixed rare earth and yttrium-based heavy rare earth nodulizers, increased the nodularity to 87%. Yet, small amounts of chunky graphite and pearlite were still observed at the core.
Scheme 5, which combined mixed rare earth, yttrium-based heavy rare earth, and 0.005% Sb, gave the best result: 90% nodularity with smooth, round graphite nodules. The synergistic effect of rare earth elements and antimony effectively suppressed graphite degeneration in the heavy section. The nodularity was consistently high from the gating area to the core of the test block.
The improvement in nodularity due to the addition of antimony can be explained by the reduction of rare earth segregation at the austenite–graphite interface. Antimony is a strong pearlite stabilizer, but in small amounts it also counteracts the spheroidizing-deteriorating effects of elements such as cerium. The combined use of light rare earths (La, Ce) and heavy rare earths (Y) provides a more balanced rare earth distribution, preventing both excessive segregation and rapid decay.
2.3 Quantitative Analysis of Graphite Morphology
To better understand the effect of each process variable, we may express the nodularity improvement as a function of carbon equivalent and antimony addition. Based on the experimental data, a simple linear regression model could be postulated:
$$N = 100 – a(CE – CE_{opt})^2 – b \cdot f(Sb) – c \cdot R_{RE}$$
where \(N\) is the nodularity, \(CE\) is the carbon equivalent, \(CE_{opt}\) is the optimal carbon equivalent for heavy sections, \(f(Sb)\) is a function that accounts for the beneficial effect of antimony at low concentrations, and \(R_{RE}\) represents the rare earth combination factor. Although the limited data do not allow a full statistical regression, the qualitative trends are clear: lowering CE from 4.29 to 4.16 increased nodularity by roughly 20 percentage points in Schemes 1 and 5; adding Sb further increased nodularity by about 5–10 percentage points; and replacing part of the mixed rare earth with yttrium-based heavy rare earth contributed another 5–10 percentage points improvement.
Another important parameter is the solidification modulus \(M\), defined as the ratio of volume to cooling surface area. For a 500 mm cube, the modulus is:
$$M = \frac{V}{A} = \frac{500^3}{6 \times 500^2} = \frac{500}{6} \approx 83.3 \text{ mm}$$
This large modulus leads to a very long solidification time, which can be estimated using Chvorinov’s rule:
$$t_s = B \cdot M^2$$
where \(B\) is a mold constant. For a sand mold, \(B\) is typically around 2.5–3.5 min/cm². With \(M = 8.33\) cm, \(t_s\) would be on the order of \(3 \times (8.33)^2 \approx 208\) minutes. This prolonged solidification creates ample time for spheroidizing element diffusion, segregation, and graphite degeneration. Therefore, the process must be designed to maintain an adequate concentration of effective nodulizing elements throughout the entire solidification interval.
3 Verification on Actual Wind Power Hub Castings
To validate the optimized process, four actual wind power hub castings were selected. Two were produced using the conventional process (Scheme 1), and two were produced using the improved process (Scheme 5). Step samples measuring 250 mm × 400 mm were attached to the internal cavity of each hub. After pouring, both mechanical properties and microstructures were evaluated. The results are shown in Table 5.
| Casting | R_m (MPa) | R_p0.2 (MPa) | A (%) | Hardness (HBW) | AKV (J) | Nodularity (%) | Graphite size | Matrix |
|---|---|---|---|---|---|---|---|---|
| Conventional 1 | 304 | 197 | 5 | 120 | 8, 6, 7 | 55 | 4–5 | F |
| Conventional 2 | 312 | 211 | 12 | 119 | 10, 9, 10 | 71 | 5 | F |
| Scheme 5 – 1 | 358 | 238 | 25.5 | 120 | 12, 10, 12 | 90 | 5–6 | F |
| Scheme 5 – 2 | 342 | 226 | 19.4 | 125 | 12, 11, 13 | 91 | 5–6 | F |
The conventional castings showed substantial variation in nodularity (55% and 71%) and relatively low elongation (5% and 12%). In contrast, the castings produced by Scheme 5 achieved nodularity above 90% and consistently high elongation values (25.5% and 19.4%). The tensile strength and yield strength also increased by approximately 15% compared with the conventional process. The impact toughness was improved as well, with average absorbed energies of 11–12 J for Scheme 5 versus 7–10 J for the conventional castings.
The microstructure of the core region from the conventional process showed obvious chunky graphite and exploded graphite, while the Scheme 5 samples exhibited well-formed spheroidal graphite uniformly distributed in a ferritic matrix. The combination of yttrium-based heavy rare earth and antimony proved to be highly effective in stabilizing graphite growth without promoting pearlite formation in the heavy sections.
4 Further Discussion on Mechanisms
Several mechanisms contribute to the observed improvements in ductile iron castings. First, the reduction in carbon content lowers the amount of graphite precipitated during the eutectic solidification. This is particularly important in heavy sections where the local carbon concentration can become enriched due to microsegregation. A lower carbon equivalent reduces the driving force for graphite branching, which is the precursor to chunky graphite. The beneficial effect of carbon reduction is evident when comparing Scheme 1 (CE = 4.29, nodularity 35%) with Scheme 2 (CE = 4.13, nodularity 75%).
Second, antimony is known to counteract the detrimental effects of excessive rare earth elements. Rare earths, especially cerium, accumulate at the growth front of graphite nodules and can destabilize the austenite shell, causing branching. Antimony, being a surface-active element, competes with rare earths for adsorption at the graphite–melt interface. At an optimal level of about 0.005%, antimony prevents the formation of chunky graphite while not significantly increasing pearlite. In Scheme 5, the combined addition of Sb and yttrium-based heavy rare earth yielded the best nodularity without introducing excessive pearlite.
Third, yttrium-based heavy rare earth nodulizers provide a more persistent spheroidizing effect than light rare earths alone. Heavy rare earth elements such as yttrium have lower vapor pressure and higher boiling points, reducing fading during long solidification. However, an excessive amount of yttrium can cause large, irregular graphite nodules and promote pearlite formation due to segregation. Thus, a partial replacement of the mixed rare earth with yttrium-based rare earth (e.g., 1.0% mixed + 0.2% yttrium-based) was found to be optimal, as demonstrated in Scheme 5.
It is also important to consider the inoculation practice. The use of high-calcium barium inoculant as a cover and tundish addition, combined with sulfur-oxygen instantaneous inoculation, ensures a fine distribution of nucleating substrates. Long-term inoculation efficiency is critical for heavy-section ductile iron castings. The sulfur-oxygen inoculant produces a large number of small (Mg,Ca)S and (Mg,Ca)O particles that act as nucleation sites for graphite. This promotes high nodule counts and reduces the tendency for graphite to degrade.
To quantitatively evaluate the effect of nodule count on nodularity, the following relationship is often used:
$$N_n = \frac{N_{grains}}{V_{solid}}$$
where \(N_n\) is the nodule density, \(N_{grains}\) is the number of graphite nodules in a given volume, and \(V_{solid}\) is the solidification volume. A higher nodule density leads to shorter diffusion distances for carbon and a lower tendency for chunky graphite formation. In our experiments, Scheme 5 produced a nodule count of approximately 90–120 nodules/mm² in the heavy section, while the conventional process produced only 40–60 nodules/mm². This increase in nodule density was primarily due to the combined effects of optimized inoculation and the stabilizing influence of antimony and yttrium on the nucleation sites.
Another factor is the cooling rate. The cooling rate \( \dot{T} \) at the center of a heavy section can be approximated by:
$$\dot{T} = \frac{\Delta T}{t_s}$$
where \(\Delta T\) is the solidification temperature range and \(t_s\) is the solidification time. For a 500 mm cube, \(t_s\) can exceed 200 minutes, leading to a cooling rate of less than 0.1°C/min. Under such extremely slow cooling, even small changes in the concentration of surface-active elements can dramatically affect graphite morphology. The use of heavy rare earth elements, which have lower diffusion coefficients in the melt than light rare earths, helps maintain a uniform distribution of spheroidizing elements throughout the solidification process.
Furthermore, the presence of Sb in combination with rare earths modifies the segregation behavior of elements such as cerium and lanthanum. It has been reported that Sb forms stable intermetallic compounds with rare earths, reducing their effective concentration at the graphite–austenite interface. This prevents the destabilization of the austenite shell and suppresses the branching of graphite. The result is a smooth, spheroidal graphite shape, as observed in Scheme 5.
It is worth noting that the phosphorus content in some schemes (e.g., Schemes 3 and 4) was slightly higher (0.023–0.024 wt%) compared with Schemes 1 and 5 (0.019–0.022 wt%). Phosphorus can segregate at grain boundaries and reduce the ductility and impact toughness of ductile iron castings. The improved mechanical properties of Scheme 5 are partly attributable to the lower phosphorus content, which was controlled in the charge materials. Therefore, careful control of trace elements is essential for producing high-quality heavy-section ductile iron castings.
5 Conclusions
Based on the systematic investigation of different process schemes for improving the nodularity of heavy sections in ductile iron castings, the following conclusions can be drawn:
(1) A moderate reduction in carbon content, from a carbon equivalent of about 4.29% to 4.16%, significantly reduces the amount of exploded graphite and chunky graphite in the core of heavy sections. This is accomplished by lowering the carbon content from approximately 3.92% to 3.84% while adjusting silicon accordingly.
(2) The addition of 0.005% antimony alone cannot completely eliminate chunky graphite in a 500 mm × 500 mm × 500 mm test block, but it does improve the nodularity from 35% to 75% when combined with carbon reduction. Antimony helps suppress the detrimental effects of rare earth segregation.
(3) Using only yttrium-based heavy rare earth nodulizer can raise the core nodularity to 85%, but the enrichment of heavy rare earth elements affects the roundness of graphite nodules and leads to the formation of a small amount of pearlite, especially in rapidly cooled areas, which would reduce low-temperature impact toughness.
(4) The best result is achieved by using a combination of 1.0% mixed rare earth nodulizer, 0.2% yttrium-based heavy rare earth nodulizer, and 0.005% antimony, together with high-calcium barium covering and tundish inoculation and sulfur-oxygen instantaneous inoculation. This process consistently yields a core nodularity of 90% or higher with well-formed, round graphite nodules and a fully ferritic matrix.
(5) Verification on actual wind power hub castings confirms that the improved process increases the nodularity from 55–71% to above 90%, and improves the tensile strength by about 15%, yield strength by about 15%, elongation by nearly 100–150%, and impact toughness by approximately 20–30% compared with the conventional process. Thus, the optimized process is highly effective for producing reliable heavy-section ductile iron castings with superior mechanical properties and consistent graphite morphology.
In conclusion, the combined effects of lower carbon equivalent, antimony micro-addition, and the judicious use of yttrium-based heavy rare earth nodulizer provide a robust solution to the long-standing problem of graphite degeneration in thick-walled ductile iron castings. This approach can be readily implemented in industrial production to ensure that the internal quality of heavy-section ductile iron castings meets the highest standards without relying solely on attached test blocks.