The microstructure evolution in ductile cast iron with magnesium addition was observed in situ by using X-ray radiography (two-dimensional observation) and time-resolved tomography (three-dimensional observation) in the BL20XU of a synchrotron radiation facility, SPring-8 (Hyogo, Japan). In the two-dimensional observation, graphite nodules nucleated in the melt and floated up immediately after nucleation. The floating was terminated by engulfment of graphite nodules into austenite dendrites. The radiography indicated that the average floating distance was shorter than the dendrite arm spacings in the 100-μm-thick specimen. Because the short distance could be influenced by the sample confinement, time-resolved tomography was performed by using a pink X-ray beam in the BL28B2 of SPring-8. Graphite nodules that nucleated in the melt (probably on magnesium–oxygen–sulfur inclusions in the melt) floated and were engulfed by austenite dendrites within several seconds, even in the bulk specimen. Although the average distance in the bulk specimen was approximately twice as large as that in the 100-μm-thick specimen, floating after nucleation and engulfment into austenite dendrites within a short duration were observed commonly from both techniques. The sequence of nucleation and engulfment had a critical effect on the number and size of the graphite nodules.
A small addition of magnesium (Mg) or cerium (Ce) into cast iron modifies the graphite morphology from flake-shaped to a compacted or spheroidal shape.1,2 Divorced eutectic growth, in which austenite and graphite grow independently as the primary phase, occurs in ductile cast iron that contains Mg or Ce. Inclusions3,4,5,6,7,–8 or bubbles9,10,–11 can act as nucleation sites for graphite and contribute to the formation of the spheroidal shape. Because the solidification sequence in the ductile cast iron is complex,12,13 the observation of solidification in ductile cast iron in situ is of interest to understand how graphite grows during solidification and how spheroidal graphite nodules remain in the matrix.
Recently, time-resolved and in situ observations using synchrotron radiation X-rays were performed to observe solidification in cast iron.14,15,16,17,18,–19 In early studies,15,16,–17 specimens (3.79C–3.1Si–0.14Mn–0.015S–0.04 Mg in mass%, carbon equivalent = 4.8) were melted and solidified repeatedly. Spherical graphite nodules grew in the melt and were surrounded by austenite dendrites or the coupled eutectic front of austenite and graphite.15,17 In the second or further runs, graphite particles transformed visibly from spheroidal to a flake-like shape before being engulfed by austenite dendrites or the coupled eutectic front. The temperature range in which the graphite nodules grew as the primary phase before austenite dendritic growth increased with a decrease in Mg concentration.16,17 X-ray imaging showed that Mg addition into the melt modified the phase equilibrium between the liquid, austenite and graphite. The influence of Mg on the carbon equivalent is expressed approximately by the following relationship:
where CE′ is a modified carbon equivalent. The contribution of Mg is 12 times larger than that of Si. Thus, graphite and austenite can start to grow simultaneously even at a carbon equivalent CE of more than 4.3.
In recent studies,18,19 the coupled eutectic solidification, which did not occur in ductile cast iron, was observed in an Al2O3 holder. In contrast, only the divorced eutectic solidification was observed until the end of solidification in a MgO holder. Thus, the microstructure evolution of ductile cast iron was reproduced in the thin specimen (100 μm thickness) using the MgO holder. The previous study19 showed that graphite nucleated on inclusions, such as Mg-containing oxysulfide, in the melt and most graphite nodules floated after nucleation because of the buoyancy force. The floating graphite nodules were engulfed by austenite dendrite within several seconds. The short floating distance of the graphite nodules compared with the dendrite arm spacing suggests that nucleation near austenite dendrites and solute exchange between the austenite dendrites and graphite nodules promoted the engulfment of graphite nodules into austenite dendrites. However, observations were performed by transmission imaging though a 100-μm-thick thin specimen. Floating of graphite nodules can be influenced by specimen confinement. Thus, the floating of graphite nodules in a bulk specimen should be observed to understand floating and engulfment better.
Recently, time-resolved and in situ tomography (four-dimensional computed tomography, 4D-CT) using a monochromatized X-ray has been developed to observe solidification in Fe-based alloys.20 In the 4D-CT, the specimen is rotated at 0.25 rps and reconstructed three-dimensional (3D) images are obtained every 4 s. Higher temporal resolutions are required to observe the floating of graphite nodules. A pink X-ray beam, which is reflected from an X-ray mirror, is an alternative to the monochromatized X-ray beam and is expected to improve the temporal resolution. This paper demonstrates a 3D observation of microstructure evolution during solidification in a hypereutectic ductile cast iron with Mg addition by 4D-CT using a pink beam. Nucleation events and the floating of graphite nodules in a bulk 1-mm-diameter specimen are compared with the previous results that were observed in the thin specimen. The primary and secondary nucleation events were disused based on in situ observations.
Previous results18,19 are presented briefly to compare the 4D-CT results. Conventional ductile cast iron, which was produced by the addition of Mg-containing spheroidizing agent before casting, was used as a specimen for the in situ observation (two-dimensional (2D) imaging). The chemical composition was 3.67C–2.63Si–0.45Mn–0.02 Mg–0.003S (mass %), and the carbon equivalent (CE) was 4.5. Time-resolved and in situ observations (2D) were performed at beamline BL20XU in SPring-8 (third-generation synchrotron radiation facility in Japan). The setup for the observation has been described in previous studies.14,17 A 100-μm thin specimen was melted in a MgO holder. Specimens with a typical microstructure of ductile cast irons (Figure 1) were used, and no treatment was conducted during the in situ observation procedures. The X-ray energy that was used for the observations was 19 keV. The pixel size was 0.5 μm × 0.5 μm, and the typical frame rate ranged from 1 to 2 fps. The specimen was melted and cooled at 0.5 K/s in a vacuum (less than several 100 Pa).
Time-resolved and in situ tomography (4D-CT) was performed to observe solidification at beamline BL28B2 in SPring-8. A pink X-ray beam was used, in which X-rays with an energy greater than 30 keV were cut by a mirror. The pink X-rays offer a higher brilliant beam to improve the temporal resolution compared with the monochromatized X-ray. Figure 2a shows a schematic illustration of the 4D-CT setup. A 1-mm-diameter specimen in a MgO tube was rotated at 2 rps (a 360° rotation took 0.5 s), and the transmission images (projection images) were acquired at 800 fps. The pixel size of the transmission images was 6.5 µm × 6.5 µm. The transmission images through the specimen in the MgO tube were recorded continuously during cooling. Figure 2b shows a typical procedure for 4D-CT. Two hundred projection images with a pixel matrix of 200 pixel × 200 pixel over a 180° rotation were used for the reconstruction. In the reconstructed images, ring artifacts and beam-hardening artifacts remained. To improve the image quality by removing the artifacts, the reconstructed images were normalized by the liquid-phase images immediately before solidification started.
Results and Discussion
Transmission Images (2D Observation)
The sequence of solidification in the hypereutectic ductile cast iron,18 which was obtained by the transmission imaging, is presented for comparison with the tomography results. The specimen was solidified in a MgO holder. As shown in Figure 3, austenite dendrites and graphite nodules grew independently. It should be noted that austenite and graphite grew simultaneously. In a previous study,17 Mg addition (500 ppm) shifted the eutectic composition, as expressed in Eqn. 1. The sequence in Figure 3 is consistent with that observed in the previous study.17 The graphite nodules tended to nucleate near the austenite dendrites and floated up immediately after nucleation. Graphite-nodule floating showed that nucleation events occurred in the melt. Graphite-nodule engulfment into austenite occurred when the floating graphite particles touched the austenite dendrite arms. Only the divorced eutectic was selected until the end of solidification. As a result, spheroidal graphite nodules that were surrounded by ferrite and perlite as a matrix were distributed in the observation specimen. Thus, the typical microstructure of the ductile cast iron, which is the so-called Bull’s eye, was reproduced in a thin specimen and the evolution was observed.
Previous studies3,4,5,6,7,–8 found that Mg–O–S inclusions act as nucleation sites because the inclusions exist in the core of graphite nodules. In an Al2O3 holder, graphite tended to transform from spheroidal to compacted/flake shapes.18 Holder material (MgO and Al2O3) that contacted the melt controls the oxygen potential in the melt. The oxygen potential in the Al2O3 holder is higher than that in the MgO holder. Mg–O–S and MgS inclusions, which existed initially in the melt, were unstable at a higher oxygen potential and could transform to spinel (MgAl2O4) or MgO. Because the specimen thickness was 100 μm, inclusion modification occurred over the entire specimen. The decrease in number of Mg–O–S inclusions in the melt reduced graphite nucleation events, if the Mg–O–S inclusions act as a graphite nucleation site. The influence of holder material is explained consistently by considering the modification of inclusions in the melt.
When inclusion modification occurs in the melt, sulfur atoms return to the melt. Sulfur atoms in the melt can also influence the graphite growth kinetics and carbon diffusion in the melt. Further studies on the influence of sulfur on the growth kinetics of graphite are required to discuss this aspect.
The shape evolution of graphite nodules18 is shown in Figure 4. Graphite nodules that were indicated by an arrow nucleated and grew in the melt as shown in Figure 4a. Because most nodules were moved upward by the buoyancy force, nucleation events occurred in the melt. Spheroidal graphite nodules immediately after engulfment are shown in Figure 4b. After engulfment, the graphite nodules, as shown in Figure 4c, became larger, but retained their spheroidal shape. The observation proved that the spheroidal graphite particles continued to grow even in the austenite. According to SEM/EDS observations of graphite nodules in previous works,18,19 the inclusions that contained Mg–O–S are a candidate for the graphite nucleation sites.
Time-Resolved Tomography (4D-CT)
Figure 5 shows the 3D evolution of graphite nodules during solidification in the spheroidal graphite cast iron at a cooling rate of 0.5 K/s. The 3D images were reconstructed from two hundred projections over 180°. Time in the figure was designed to be zero when cooling started. The interval between the reconstruction was 0.5 s, and the voxel size was 6.5 µm × 6.5 µm × 6.5 µm. Because the difference in X-ray absorption coefficient between the liquid phase and austenite was too small, γ-Fe dendrites (austenite dendrites) were not identified in the reconstructed images. The pink X-ray beam improves the temporal resolution, but decreases the contrast resolution. In the temporal resolution (2 rps), the floating of graphite nodules was observed as shown in Figure 5.
Small graphite nodules, probably immediately after nucleation, were identified at t = 18.5 s. Arrows at t = 33.50 s indicated the movement of some graphite nodules from the previous 3D image. The graphite nodules floated up because of the buoyancy force after nucleation, and the floating stopped. The floating events were stopped when graphite nodules were engulfed by austenite dendrite. Most graphite nodules floated up once after the nucleation. Thus, the floating behavior that was observed by 4D-CT is the same as the results from the transmission images in the previous section.18,19
The graphite nodules were distributed uniformly in the entire specimen at t = 55.0 s, and the number of graphite nodules increased even after 55.0 s. The increase means that nucleation events occur between the graphite nodules that nucleated previously and were engulfed by austenite dendrites after t = 55.0 s. In general, graphite-nodule nucleation hardly occurs between the graphite nodules that are growing in the melt, because carbon saturation in the melt is reduced around the growing graphite nodules. The higher diffusivity of carbon atoms (interstitial element in the solid) reduces the saturation. In this study, graphite-nodule nucleation between the growing graphite nodules was defined as secondary nucleation. Secondary nucleation events that were induced by the engulfment of carbon nodules into austenite will be a key event to determine the number and size of graphite nodules in the solidification structure.
Floating of Graphite Nodules
The floating of graphite nodules was evaluated from the time-resolved 3D datasets. The floating distance and nucleation time of the graphite nodules are plotted in Figure 6. The nucleation time was defined as the time at which the graphite nodule was detected. At the beginning of solidification, the floating distance of the graphite nodules ranged from 10 to 120 µm. The long distance of 120 μm, which was comparable to the dendrite arm spacing measured in the 2D observation, suggests that the volume fraction of austenite dendrites was low and sufficient space remained for floating. The floating distance of the graphite nodules became less than 10 µm after t = 55.0 s when the secondary nucleation events occur. The shorter distance shows that the graphite nodules nucleated in the γ dendrite network and were engulfed immediately by austenite dendrites. Although it was not always easy to distinguish secondary nucleation from primary nucleation, the graphite nodules tended to nucleate between the previously nucleated graphite nodules after t = 55.0 s. In addition, the number of secondary nucleation events was larger than that of primary nucleation.
A comparison of the floating distance in a thin specimen (2D observation) and a bulk specimen (3D observation) is of interest. Figure 7 shows histograms of the floating distance of graphite nodules. For the thin specimen, the floating distance was less than 10 μm for most graphite nodules. The floating distance in the bulk specimen was approximately twice as large as that in the thin specimen. Thus, sample confinement in the thin-specimen holder influenced the floating behavior. However, the floating distance of 20 μm in the bulk sample remains shorter than the primary dendrite arm spacing that was measured in the 2D observation.18,19 Thus, graphite nucleation events near austenite dendrites and the interaction between graphite nodules and austenite dendrites through carbon diffusion reduced the floating distance. Nucleation in the melt and the short floating distance promoted secondary nucleation and consequently influenced the number and size of graphite nodules in the solidification structure.
Influence of Engulfment on Secondary Nucleation
The engulfment of graphite nodules into austenite dendrites in a short duration promotes the secondary nucleation of graphite nodules. The sequence of nucleation and engulfment was explained by considering carbon transport around the graphite nodules.19 Figure 8 shows a schematic illustration of the secondary nucleation of a graphite nodule after austenite engulfment. During cooling, the austenite phase nucleates and grows into the melt in advance or simultaneously. Carbon atoms are rejected at the solidifying front of the austenite, and the carbon content near the solidifying front increases. If the melt temperature is below the eutectic temperature of the austenite and the graphite or the liquidus temperature of graphite, carbon-atom rejection at the liquid–austenite interface increases the graphite nucleation driving force (ΔCGr) near the austenite dendrite, as shown in Figure 8a. Once graphite nucleates, supersaturated carbon atoms are consumed by graphite growth. As a result, the driving force for graphite nucleation decreases rapidly as shown in Figure 8b. Thus, graphite nucleation in the melt suppresses further graphite nucleation. The diffusivity of carbon (interstitial element) is relatively high. Iron in the melt around the graphite nodule is supersaturated for austenite (ΔCγ), which leads to austenite growth toward the graphite nodule, as shown in Figure 8b. Austenite growth toward the graphite nodule reduces the floating distance. The engulfment of graphite nodules into the austenite dendrite eliminates the liquid–graphite interface, as shown in Figure 8c. After engulfment, carbon atoms are rejected to the liquid phase at the liquid–austenite interface again and the driving force of graphite nucleation in the melt is increased. Therefore, the sequence of graphite nucleation in the melt and engulfment into the austenite dendrite contributes to an increase in the number of graphite nodules during solidification.
Time-resolved and in situ X-ray imaging (2D transmission imaging) and tomography (4D-CT) were performed to observe solidification in the hypereutectic ductile cast iron with Mg addition.
The typical microstructure of ductile cast iron, in which graphite nodules that were surrounded with ferrite phase were distributed in the perlite matrix, was produced in the thin (2D observation) and bulk (3D observation) specimens. The X-ray imaging techniques demonstrated the microstructure evolution in ductile cast iron.
Graphite nodules tended to nucleate in the melt near the austenite dendrites and floated up until they were engulfed into the austenite dendrites. Although the austenite dendrites were not visible in the 4D-CT observation, the results that were observed by 4D-CT were consistent with the transmission imaging results.
The floating distance was shorter than the dendrite arm spacings. Nucleation of graphite near austenite dendrites and carbon exchange between graphite and austenite though diffusion in the melt promoted the engulfment.
The sequence of nucleation and engulfment of graphite nodules caused the secondary nucleation of graphite nodules and determined the number and size of graphite nodules in the solidification structure.
H. Morrogh, J.W.J. Grant, Iron Steel Inst. 155, 321 (1947)
A.P. Gagnebin, K.D. Millis, N.B. Pilling, The Iron Age. 77 (1949)
K.M. Muzumdar, J.F. Wallace, AFS Trans 80, 317–328 (1972)
K.M. Muzumdar, J.F. Wallace, AFS Trans. 81, 412–423 (1973)
G.X. Sun, C.R. Loper, AFS Trans. 91, 639–646 (1983)
T. Skaland, Metall. Trans. A 24A, 2321–2345 (1993)
M.H. Jacobs, T.J. Law, D.A. Melford, M.J. Stowell, Metals Technol. 1, 490–500 (1974)
N. Klorca-Isern, J. Tartera, M. Espanol, M. Marsal, G. Bertran, S. Castel, Micron 33, 357–364 (2002)
H. Itofuji, AFS Trans. 104, 79–87 (1996)
S. Yamamoto, Y. Kawano, Y. Murakami, B. Chang, R. Ozaki, Metal Sci. 12, 56–60 (1978)
M.J. Lalich, J.R. Hitchings, AFS Trans. 84, 652–658 (1976)
D.M. Stefanescu, Metall. Mater. Trans. A 38A, 1433–1447 (2007)
H. Zhao, M. Zhu, D.M. Stefanescu, Key Eng. Mater. 457, 324–329 (2011)
A. Sugiyama, H. Yasuda, T. Nagira, M. Yoshiya, K. Uesugi, K. Umetani, I. Ohnaka, J. Jpn Foundry Eng. Soc. 82, 131–136 (2011)
K. Yamane, H. Yasuda, A. Sugiyama, T. Nagira, M. Yoshiya, K. Uesugi, K. Umetani, C. Ushigome, A. Sato, J. Jpn Foundry Eng. Soc. 85, 760–770 (2013)
K. Yamane, A. Sugiyama, T. Nagira, M. Yoshiya, H. Yasuda, Y. Tanaka, A. Sato, K. Uesugi, A. Takeuchi, Y. Suzuki, H. Honda, K. Sato, J. Jpn Foundry Eng. Soc. 86, 461–470 (2014)
K. Yamane, H. Yasuda, A. Sugiyama, T. Nagira, M. Yoshiya, K. Morishita, K. Uesugi, A. Takeuchi, Y. Suzuki, Metall. Mater. Trans. A 46A, 4937–4946 (2015)
H. Yasuda, A. Sugiyama, C. Kiattisaksri, K. Morishita, T. Nagira, M. Yoshiya, K. Uesugi, A. Takeuchi, Mater. Sci. Forum 925, 104–109 (2018)
K. Chatcharit, A. Sugiyama, K. Morishita, H. Yasuda, J. Jpn Foundry Eng. Soc. 90, 602–612 (2018)
H. Yasuda, T. Hashimoto, N. Sei, K. Morishita, M. Yoshiya, IOP Conf Ser: Mater Sci Eng 529, 012013 (2019)
This paper was presented at the 2nd Carl Loper Cast Iron Symposium, which was held on September 30 to October 1, 2019, in Bilbao, Spain. This study was supported by a Grant-in-Aid for Scientific Research (S) from MEXT (17H06155). The synchrotron radiation experiments were performed as general projects at beamlines BL20XU, BL20B2 and BL28B2 at SPring-8 (JASRI). We thank Laura Kuhar, PhD, from Edanz Group (www.edanzediting.com/ac) for editing a draft of this manuscript.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This paper is an invited submission to IJMC selected from presentations at the 2nd Carl Loper 2019 Cast Iron Symposium held September 30 to October 1, 2019, in Bilbao, Spain.
About this article
Cite this article
Chatcharit, K., Sugiyama, A., Morishita, K. et al. Time Evolution of Solidification Structure in Ductile Cast Iron with Hypereutectic Compositions. Inter Metalcast 14, 794–801 (2020). https://doi.org/10.1007/s40962-020-00424-3
- spheroidal graphite
- ductile cast iron
- time-resolved tomography
- in situ observation
- synchrotron radiation X-ray