Materials Processing
Electron Microscopy on Cu Element Distribution in Spheroidal Graphite Cast Iron
2020 Volume 61 Issue 9 Pages 1853-1861
Details
2020 Volume 61 Issue 9 Pages 1853-1861
The solidification microstructure of spheroidal graphite cast iron was analyzed by electron probe micro analysis (EPMA) and scanning transmission electron microscopy (STEM), focusing on the distribution of Cu. EPMA and STEM results clarified the following tendency in the distribution of Cu in spheroidal graphite cast iron containing Cu within the pearlite matrix: (1) Cu–Sn–Mg enriched regions were distributed in the pearlite matrix, and (2) the distribution of metallic elements of Cu, Sn, and Mg in Cu–Sn–Mg enriched regions was not homogeneous, and a composite of the metallic and oxide phases was formed. EPMA with a field-emission electron gun (FE-EPMA) and FE-STEM apparatus with a silicon drift detector were highly effective in clarifying the distribution of Cu, compared with conventional EPMA employing a thermionic-emission electron gun (TE). The high-resolution observation using FE-STEM, and the combination with TE-EPMA, FE-EPMA and FE-STEM, are powerful tools for clarifying the distribution of Cu element in spheroidal graphite cast iron.
This Paper was Originally Published in Japanese in J. JFS 91 (2019) 512–520. Minor corrections in abstract, main text, figure and table captions, and references were performed with translation from Japanese to English and proofreading by native speakers. Reference 18) was replaced from “H. Ito, I. Narita, H. Miyahara: Reports of JFS meeting, 170 (2018) 4.” to “H. Miyahara, G. Ito, I. Narita: J. JFS 91 (2019) 703–709.”
Spheroidal graphite (SG) cast irons are Fe-based alloys that can be cast in the atmosphere; herein, the spheroidal graphite is dispersed in the iron matrix. SG cast irons exhibit good mechanical properties and high productivity, resulting in their wide use as industrial metallic materials.1–3) A number of studies have been performed on the mechanism of SG formation during the solidification of cast irons. Based on the viewpoint of materials engineering and processing, a process technology that facilitates the production of SG cast irons has been established. In contrast, the detailed mechanism for the formation of SG has never been elucidated. In recent years, there has been a renewal of interest with regard to the SG formation mechanism; this is because of the upgradation of equipment and the development of the technique employed in electron microscopy. Stefanescu et al.4,5) used the word “reexamination” in the title of their report. Based on investigations of SG cast irons by transmission electron microscopy (TEM) and scanning transmission electron microscopy (STEM) that have been carried out since 2014, the studies can be classified into the following two groups, namely: (1) defect structure analysis of SG using TEM and (2) fine-scale elemental analysis using STEM. Majority of the research associated with TEM on SG cast iron was performed focusing on the defect structure analysis in SG, which is based on the concept that the defect in SG is the key factor that governs the morphology of graphite. Qing et al. reported the dual structure in SG composed with an inner region with circumferential growth of the curved graphene and an outer region of graphite with rotation and stacking faults.6) Theuwissen et al. suggested the importance of atomic-scale defects in the SG7) and reported the crystal orientation as well as phase maps of SG by the use of automated crystal orientation mapping (ACOM) in TEM.8,9) Muhammad et al. reported the relationship between the trace elements (F, H, O, B, N, S, P, Se, etc.) and graphite growth morphologies in SG by employing high-resolution TEM (HRTEM).10) Hara and Lee et al. reported the heterogeneous structure in SG composed with an amorphous-like central region, annual rings of a layered intermediate region, and an outer region composed of large polygonal crystalline platelets in a mosaic-like structure.11–13) Maeda et al. reported the results of the analysis of the microstructure of SG by 3D TEM tomography using the Ultra-High Voltage Electron Microscope installed at Osaka University.14–17) Miyahara and Narita et al. clarified the misfit between SG and oxides by TEM.18) Igarashi et al. reported that the microstructure of SG varies with the growth of graphite in SG cast irons; this was achieved by the TEM and HRTEM focusing on the interface between the iron matrix and SG, radial textured graphite, and annual-ring textured graphite regions in SG.19) The above works reported similar results to those of previous studies,20–25) wherein SG exhibited a heterogeneous structure with various types of defects. The majority of the above-described TEM and HRTEM studies have focused on the nucleus of SG or have reported only the results of the microstructure analysis of the SG without a detailed discussion of the relationship between the defect structure and growth process.
STEM is the electron microscopy that employs beam scanning and signal mapping, which is similar to scanning electron microscopy (SEM), while TEM is the electron microscopy that utilizes lens imaging. STEM coupled with a high-angle annular dark-field (HAADF) detector is capable of high-spatial resolution imaging with an excellent elemental contrast (Z-contrast). STEM is recognized as a powerful technique for the analysis of the distribution of elements with high-spatial resolution; this is because of the HAADF image with Z contrast, scanning with a nanoscale electron beam, and suppression of the electron beam spreading. STEM analysis of element distribution in SG was reported in 1993 by Skaland et al.26) It was reported that the distribution of Mg, Ca, Al, and Si at the central region of SG with respect to the nuclei was inhomogeneous; however, detailed information was not reported because of the lack of spatial resolution. Compared with TEM, which is based on a parallel electron beam, STEM is based on scanning with a focused electron beam probe. As a result, STEM observation is extremely susceptible to the magnetic properties of the specimens. A minor attention has been given to the application of STEM for the microstructure analysis of SG cast irons; this is because of the difficulty caused by the magnetics of the iron matrix. The recent progress in the performance of STEM equipment has resulted from the improvement in the electrical and mechanical stability of STEM equipment. The development of the spherical aberration (Cs) corrector and the upgradation of the silicon drift detector (SDD) has been effective to improve the performance of STEM. The recent progress of the focused ion beam (FIB) system is particularly beneficial for the reduction of damage and artifacts during sample processing. The above-mentioned advancements have rendered it possible to evaluate the high-spatial elemental distribution in SG cast irons by STEM.27–32)
The tensile strength, hardness, and wear resistance of SG cast irons increase with the formation of a pearlite matrix in SG cast irons. The addition of Cu is known to be effective for the formation of a pearlite matrix in SG cast irons.1,2) Although the distribution of Cu in SG cast irons has been an object of study for a long time, there is an insignificant understanding on the position and morphology of the Cu-enriched region in SG cast irons. Ishiguro et al. reported Cu enrichment at the interface between SG and perlite matrix in Cu-added high-strength SG cast irons; this was observed by the energy-filtered backscattering electron images using field-emission SEM.33) Zou et al. observed Cu-film formation at the surface of SG by SEM using polished samples etched with degassed 7% HCl aqua solution for 2 h.34) In contrast, Tsujikawa et al. reported that the Cu film was not observed at the interface between SG and the perlite matrix in the bright field (BF) image of TEM.35) Igarashi et al. investigated the microstructure of boron (B)-added SG cast irons using SEM and TEM.36) They found triangular precipitates consisting of Fe–Si–Cu–B with a size of ∼150 nm at the surface layer of the graphite, and the Fe–Si–Cu–B region was composed of Fe-rich nanocrystalline grains and Cu–B-rich amorphous regions. The Fe–Cu alloy system shows a phase diagram with a flat liquidus and a metastable liquid miscibility gap.37) The liquid-phase separation phenomenon is generally observed in alloys wherein the mixing enthalpy (ΔHmix) between constituent elements shows a large positive value (ΔHmix ≫ 0). In contrast, the amorphous phase tends to be formed in alloys having large negative values (ΔHmix ≪ 0) among the constituent elements.38–41) Immiscible alloys with liquid-phase separation and amorphous alloys exhibited contrary ΔHmix of constituent elements. Amorphous-phase formation in Fe–Cu-based alloys with liquid-phase separation has been considered to be difficult for amorphous-phase formation because of the positive ΔHmix (+13 kJ/mol) in Fe–Cu atomic pair. Nevertheless, amorphous-phase formation in Fe–Cu-based alloys was found in 2005,42) and various Fe–Cu-based immiscible alloys with an amorphous phase have subsequently been developed. Based on the research on the Fe–Cu-based immiscible alloys with amorphous-phase formation, the Si and B were clarified to be the essential elements for this formation.43,44) B was the key element for enhancing the liquid-phase separation in Fe–Cu-based alloys,45) and it was the most essential element for the amorphous-phase formation in Fe–Cu-based immiscible alloys.42–45) These reports indicate that B has an enormous effect on the distribution of Cu in SG cast irons. Regarding the distribution of Cu in SG cast irons, a previous report36) related to the Fe–Si–Cu–B alloy-based amorphous phase containing B and other reports33–35) on SG cast irons without B should be studied separately. To the best of the authors’ knowledge about the reports focusing on the Cu distribution in SG cast irons by electron microscopy, there have been no reports about the detailed microstructure analysis of SG cast irons without the addition of B by using STEM. In this study, the distribution of Cu in Cu-added SG cast iron composed with a perlite matrix and SG was investigated; this was achieved by using an electron probe microanalyzer with a thermionic-emission electron gun (TE-EPMA), FE-EPMA, and SDD-equipped field-emission STEM (FE-STEM).
SG cast iron was prepared by using a green sand mold and high-frequency induction furnace with a total amount of 4.0 kg. Electrolytic iron (Fe: 99.95 mass% or more, hereafter all mass%), low-sulfur artificial graphite (C: 99.45%), ferrosilicon (Fe–75.08% Si–1.44% Al), electrolytic manganese (Mn: 99.9% or more), pure copper (Cu: 99.9% or more), pure tin (Sn: 99.9% or more), spheroidizing agent of graphite (Fe–44.83% Si–5.25% Mg–2.30% Ca–1.32% Ce–0.66% La) and inoculant (Fe–70.61% Si–1.39% Ca–1.61% Ba) were used as the raw materials. Graphite crucibles lined with alumina cement and colloidal silica were used. The melting and casting were performed with the following steps. (1) The mixture of ferrosilicon, low-sulfur artificial graphite, and a part of electrolytic iron was set in the crucible, and the heating of the crucible was started from room temperature. (2) After the melting of the mixture of raw materials in the crucible, the remaining electrolytic iron, electrolytic manganese, pure copper, and pure tin were put in the crucible. (3) After confirming the melting of all raw materials by stirring the molten metal with a graphite rod, the temperature of the crucible was raised to 1823 K. (4) The spheroidizing treatment was performed by a sandwich method. The graphite spheroidizing agent was used for the cover material, and the timing of inoculation was the same as the spheroidization treatment. The amount of spheroidizing agent was 76 g (the expected yield of magnesium was 36%), and that of inoculum was 0.3% of the total dissolved amount (4 kg). (5) The pouring of molten metal was performed after spheroidizing treatment and inoculation. The pouring temperature was approximately 1623 K. The time from the start of heating to pouring was approximately 3.6 × 103 s. The ingots with steplike morphology were obtained via sand mold casting using the model shown in Fig. 1(a). Figure 1(b) shows the outer appearance of an SG cast iron specimen. Table 1 shows the chemical composition of a SG cast iron (mass%), together with the value of carbon equivalent (C.E.). To minimize the effect of graphite precipitation at solid state caused by the decrease in the solid solubility limit of carbon in austenite along the Acm line and the diffusion of carbon during cooling, the pearlite-promoting element of Sn1,2) was added together with Cu in this study.
Schematic illustration of sand mold (a) and outer appearance (b) of spheroidal graphite cast iron specimen.
The specimens for TEM and STEM were obtained from the central region of the position P in Fig. 1(a) with a thickness of 6.00 mm (measured 6.35 mm) for reducing the size of the SGs embedded in the iron matrix. Optical microscopy (OM) observation was performed for the ingots after mechanical polishing. The OM microstructure of some specimens was observed using the polished specimens that had undergone 3% Nital corrosion. Microstructure analysis by EPMA was performed using TE-EPMA (JEOL JXA-8800R, Osaka University, W filament type TE, acceleration voltage 20 kV, probe current ∼1.0 × 10−8 A) and FE-EPMA (JEOL JXA-8530F, Kindai University, thermal emission type FE, acceleration voltage 15 kV, probe current ∼1.5 × 10−8 A). TEM observation was performed using a conventional TEM apparatus (Hitachi H-800, Osaka University, W filament type TE, acceleration voltage 200 kV). STEM observation was performed for microstructure observation and the analysis for elemental distribution (JEOL JEM-2100F, Kobe University, Schottky type FE, acceleration voltage 200 kV). The TEM-BF, STEM-BF, and STEM-HAADF images were obtained for the microstructure analysis. Wavelength dispersive X-ray spectroscopy (WDS) was applied for the elemental distribution analysis and chemical composition analysis in TE-EPMA and FE-EPMA, while energy dispersive X-ray spectrometry (EDS) with SDD was applied in STEM element mapping and chemical composition analysis. Table 2 summarizes the specifications of the apparatus and experimental conditions of the TE-EPMA, FE-EPMA, and FE-STEM. Stage scan mode was used in WDS using TE-EPMA, while beam scan mode was used in WDS using FE-EPMA and EDS using FE-STEM. In Table 2, the minimum electron beam probe size is the catalog value for TE-EPMA and FE-EPMA under the conditions of an acceleration voltage of 10 kV and a beam current of 1 × 10−8 A, and that of FE-STEM indicates the minimum setting value for the electron beam probe. The significantly high spatial resolution for elemental analysis using STEM was because of the suppression of the electron beam diffusion region by the thinning effect as well as the smaller electron beam probe size compared with EPMA. The thin-film approximation method was applied as a qualitative measurement method of elements in spectroscopic analysis of characteristic X-rays in FE-STEM EDS.
TEM and STEM specimens were prepared by flat-type polishing techniques via mechanical polishing and ion-milling processes,20,25) and the details are as follows. (1) Thin-sliced specimens with approximately 1-mm thickness in SG cast irons were cut from the ingots mechanically. (2) Thin specimens were wet-polished to a thickness of approximately 70 µm using water-resistant emery paper, where the final SiC grain size was approximately 10 µm (Grade 2000). (3) Thin specimens were mirror-polished using diamond lapping film, where the final diamond grain size was about 0.1 µm. (4) The thinnest part of the thin specimens was polished to approximately 20 µm in the thickness by a dimple machine, where diamond paste particle sizes of 6, 3, and 1 µm were used. (5) Thin films were thinned by Ar ions using Gatan PIPS with an acceleration voltage of 4.5 kV and incident angles of 4° at room temperature. Finally, the specimens were finish-polished at an acceleration voltage of 1.0 kV at room temperature. Instantaneous adhesives of the room-temperature curing type were used for bonding the glass substrate and the thin specimens of SG cast irons. Acetone was used for detaching the glass substrate from the thin specimens of SG cast irons and for removing the adhesive from the TEM and STEM specimens.
Figure 2 shows the OM microstructures of an SG cast iron specimen for electron microscope observation, where Fig. 2(a) is a no-etching image and Fig. 2(b) is a 3% Nital-etched image, respectively. The iron matrix showed a pearlite structure, and the existence of a chilled structure with a small area fraction was also seen in OM images. The existence of a ferrite matrix was not detected. The average size of the SG was approximately 15 µm.
OM microstructure of spheroidal graphite cast iron specimen for electron microscope observation. (a) No etching image, (b) 3% Nital etch.
Figure 3 shows TEM-BF images of an SG cast iron specimen. In Fig. 3(a1) and the magnified image in Fig. 3(a2), the iron matrix shows a pearlite structure. Figure 3(b) shows a TEM-BF image focusing on the interface between the pearlite matrix (M) and SG. TEM images shows that SG directly contacted M, indicating that graphite precipitation at solid state because of the decrease in the solid solubility limit of carbon in austenite during cooling was suppressed. The existence of thin Cu film at the interface between the iron matrix and SG33,34) was not detected in the TEM-BF images (Fig. 3(b)). This result was similar to that in the report by Tsujikawa et al.35) The TEM images in Fig. 3 are not the experimental evidence for the absence of Cu film but for the difficulty in the observation of the Cu film by conventional TEM observation, even if the Cu film exists. Further investigation using HRTEM and/or high-resolution STEM is necessary for reaching a conclusion about the absence or existence of the Cu film between the iron matrix and SG.
TEM-BF images of spheroidal graphite cast iron specimen. (a1), (a2) pearlite matrix (M), (b1), (b2) boundary region between spheroidal graphite (SG) and M.
Figure 4 shows the element mapping of Fe, C, Cu, Sn, Mg, and S in an SG cast iron specimen using TE-EPMA-WDS with W filament. In Fig. 4, the indices SG, M, X, Y, and Z denote the spheroidal graphite (SG), the iron matrix (M), Cu–Sn–Mg-enriched region (X), Mg–S enriched region (Y), and Mg-enriched region without the enrichment of Cu, Sn, and S (Z), respectively. In the Cu, Sn, Mg, and S element mapping images shown in Fig. 4(a), bright spots embedded in the pearlite matrix (M), which were difficult to identify as shot noise or not, were observed. The bright spots in the Cu, Sn, Mg, and S element mapping images were not completely randomly distributed as shown in Fig. 4(b), and the following three features were observed. (1) Bright spots of Cu, Sn, and Mg mapping images were observed in the region indicated by index X. (2) The bright spots of Mg and S mapping images were observed in the area indicated by index Y. (3) The bright spots of Mg mapping images were observed in the area indicated by index Z, while these of Sn and S were not observed in the area indicated by index Z. This indicates that the bright spots in the Cu, Sn, Mg, and S element mapping images were the regions where specific elements were concentrated rather than just shot noise. The region Z may be explained by the Mg-based oxide or nitride region;25) however, the conclusion was not reached in the present study because of the lack of experimental data about oxygen and nitrogen.
Element mapping of Fe, C, Cu, Sn, Mg and S in spheroidal graphite cast iron specimen using TE-EPMA-WDS with W filament. Indices SG, M, X, Y, and Z denote the spheroidal graphite, Fe-based matrix, Cu–Sn–Mg-enriched region, Mg–S-enriched region and Mg-enriched region, respectively.
Figure 5 shows secondary electron (SE) image, backscattering electron (BSE) image, and element mapping of Fe, C, Cu, Sn, Mg, and S in SG cast iron specimen using FE-EPMA focusing on the boundary of SG and pearlite matrix (M). In Fig. 5, the indices XX and Y denote the Cu–Sn- and Mg–S-enriched regions, respectively. XX existed at the interface between SG and M. The Mg–S-enriched region (Y) existed in a region apart from the interface between the SG and M; in the other words, Y was embedded in M. Figure 6 shows the SE image, BSE image, and element mapping of Fe, C, Cu, Sn, Mg, and S in SG cast iron specimen using FE-EPMA focusing on M. The indices SG, X, Y, and Z denote SG, the Cu–Sn–Mg-enriched region (X), the Mg–S-enriched region (Y), and the Mg-enriched region without the enrichment of Cu, Sn, and S (Z). In the Fe mapping image, regions with low Fe concentration were observed, and these regions corresponded to X and Y. Bright spots that were difficult to distinguish from shot noise in WDS element mapping using TE-EPMA (Fig. 4) were clearly observed not as shot noise, but as the particular-element-concentrated regions in WDS element mapping using FE-EPMA, shown in Figs. 5 and 6. The distribution of Cu, Mg, and Sn in XX in Fig. 5 and X in Fig. 6 was not homogeneous. The detailed information of the inhomogeneous distribution of Cu, Mg, and Sn in the regions X and XX was not obtained by FE-EPMA because of the limitation of spatial resolution.
Secondary Electron (SE) image, Back Scattering Electron (BSE) image, and element mapping of Fe, C, Cu, Sn, Mg and S in spheroidal graphite cast iron specimen using FE-EPMA focusing on the boundary of spheroidal graphite (SG) and pearlite matrix (M). The index XX and Y denote Cu–Sn-enriched region and Mg–S-enriched region, respectively.
Secondary Electron (SE) image, Back Scattering Electron (BSE) image, and element mapping of Fe, C, Cu, Sn, Mg and S in spheroidal graphite cast iron specimen using FE-EPMA focusing on the pearlite matrix (M). The index SG, X, Y and Z denote spheroidal graphite, Cu–Sn–Mg-enriched region, Mg–S-enriched region and Mg enriched region, respectively.
Figure 7 shows the element mapping of an SG cast iron specimen using STEM-EDS focusing on Cu–Sn–Mg enriched region embedded in the iron matrix, which corresponded to region X in Fig. 6. Metallic elements of Cu, Sn, and Mg elements were enriched in the brighter-contrast region in the STEM-HAADF image (Fig. 7(a)). The distribution of Cn, Sn, and Mg was inhomogeneous. Sn showed the tendency to be distributed as a small region embedded in the Cu–Mg-rich region. The enrichment of the carbon in the Cu–Sn–Mg-enriched region was not observed, indicating that the Cu–Sn–Mg-enriched region was not carbide.
Figure 8 shows the chemical composition analysis of a Cu–Sn-enriched region embedded in the iron matrix corresponding to region X in Fig. 6 using STEM-EDS. The Cu–Sn-enriched region in Fig. 8 corresponds to that in Fig. 7. Table 3 shows the chemical composition analysis results (mass%) in regions H, I, and J in Fig. 8 by STEM-EDS analysis. Not a single-phase image contrast but the contrast for a composite structure was observed in the STEM-BF image (Fig. 8(a)). In regions H and J in Fig. 8, sharp peaks caused by oxygen were observed in the EDS spectra (Figs. 8(d1) and 8(d3)). In contrast, the sharp peak corresponding to oxygen was not observed in region I (Fig. 8(d2)). The existence of the peaks corresponding to nitrogen was not observed in regions H, I, and J. Table 3 shows that the ratios of Cu:Sn:Mg in the regions H (Cu:Sn:Mg = 1:2.8:4.3), I (1:0.05:0.94), and J (1:0.04:2.9) were different. The STEM-BF image (Fig. 8(a)), HAADF image (Fig. 8(b)), element-mapping images (Figs. 8(c1)–8(c3)), and EDS spectrum (Figs. 8(d1)–8(d3)) obtained by FE-STEM with SDD show that the Cu–Sn–Mg-enriched regions dispersed in the pearlite matrix were not a single phase but a composite of metallic and oxide phases.
Composition analysis of a Cu–Sn-enriched region embedded in Fe-based matrix (corresponding to the region X in Fig. 6) using STEM-EDS. The Cu–Sn-enriched region in Fig. 8 corresponds to that in Fig. 7(a). (a) STEM-BF image, (b) STEM-HAADF image, (c1) Element mapping of Cu, (c2) element mapping of Sn, (d1) EDS spectrum of the region H, (d2) EDS spectrum of the region I, (d3) EDS spectrum of the region J.
Figure 9 shows element mapping of an SG cast iron specimen using STEM-EDS focusing on Mg–S-enriched region embedded in an iron matrix corresponding to the region Y in Figs. 4–6. The enrichment of Mg and S elements was observed at the darker region in the STEM-HAADF image (Fig. 9(a)). The enrichment of carbon, oxygen, and nitrogen was not observed in the Mg–S-enriched region, indicating that the Mg–S-enriched region was not carbide, oxide, and/or nitride. The Cu-enriched region may be observed in the Cu element-mapping image; however, the enrichment of Cu at the interface between the Mg–S-enriched region and iron matrix was not concluded, because the characteristic X-ray intensity caused by Cu was extremely weak.
The bright spots in the element-mapping image embedded in the iron matrix, which were difficult to distinguish from shot noise in TE-EPMA, were clearly observed as the particular-element-enriched regions in FE-EPMA and FE-STEM. Focusing on the distribution of Cu elements in the iron matrix, FE-EPMA and FE-STEM observation results indicated the following tendency in the Cu-enriched regions embedded in the pearlite matrix. (1) Sn and Mg were enriched together with Cu. (2) Cu–Mg–Sn-enriched regions were not a single phase. (3) Composites with metallic and oxide phases were formed in Cu–Mg–Sn-enriched regions.
The existence of Cu thin film at the interface between SG and an iron matrix has been an object of study for a long time, but there is little agreement,33–36) as described in Section 1. Even in a report excluding the research example in SG cast irons with intentionally added B,36) little is known about the Cu thin film in SG cast irons.33–35) In the results of FE-EPMA element mapping shown in Fig. 5, the Cu–Sn-enriched region (XX) at the interface between SG and the iron matrix was observed. The Cu–Sn alloy system is well known as a basic alloy system of bronze. The melting temperature of bronze is much lower than that of cast iron. In other words, Fig. 5 shows that there is a Cu–Sn-enriched region (XX) with a very low melting temperature compared with cast irons in the boundary region between SG and the pearlite matrix. Based on the results shown in Fig. 5, the following mechanisms can explain the Cu film formation reported in the literature.33,34) During the solidification and/or cooling process, Cu and Sn are concentrated in the boundary region between the spherical graphite and iron matrix. After solidification, Cu–Sn-enriched region had a size that can be observed with the spatial resolution of FE-EPMA element mapping between the spherical graphite and the iron matrix. If the Cu–Sn-enriched liquid with a low melting temperature spreads around the spherical graphite, a Cu film layer33,34) may form, as previously reported. When the Cu film layer33,34) had a thickness on the order of nanometers, it was difficult to observe in FE-EPMA and TEM-BF images, because of the lack of spatial resolution. High-resolution FE-STEM was thought to be effective in observing the Cu film layer reported to surround SG.33,34) However, the selection of the specific position for the observation was strictly limited in the present study because of the flat-type TEM and STEM specimen preparation techniques. Furthermore, it was considered difficult to increase the spatial resolution of STEM observation further because of the difficulty in the stigmatism correction caused by the magnetism of SG cast irons. The FE-STEM applied in the present study did not have enough spatial resolution for determining the existence or absence of the Cu film. Regarding the existence of Cu film layers, HRTEM is considered to be an effective method for detecting the existence; however, the image contrast of HRTEM is formed through a complicated process, and the interpretation of the image is not intuitive compared with STEM. In contrast, STEM makes the direct observation of element distribution by HAADF image and element mapping possible. Based on the reciprocal theory,46) elastic STEM-BF images are equivalent to TEM-BF images. STEM-BF images with a defect structure similar to TEM-BF images in SG in SG cast irons have been reported.27,47) HRTEM and high-spatial-resolution STEM have different characteristics based on different imaging principles. High-resolution STEM, as well as HRTEM, is considered to be an effective method for experimental verification of the existence of Cu film33,34) in SG cast irons, and this is planned for future work.
The Cu-enriched region embedded in the iron matrix was clarified as the Cu–Sn–Mg-enriched region (X in Figs. 4 and Figs. 6–8). The Cu-enriched region at the interface between the SG and iron matrix was the Cu–Sn-enriched region (XX in Fig. 5) without the concentration of Mg. The difference in the enrichment of Mg in the Cu-enriched region in the above-described regions may be explained by the performance limitation of FE-EPMA and/or the difference in the chemical composition and microstructure. Further investigation using HRTEM and/or high-resolution STEM should clarify the origin of the difference in the elements, especially for Mg in Cu-enriched regions.
TE-EPMA is an effective technique for elucidating the elemental distribution in SG cast irons. In fact, important experimental data and findings obtained by TE-EPMA have been reported recently.48–50) The present study demonstrates that FE-EPMA and FE-STEM had much higher spatial resolution in element mapping and chemical composition analysis than TE-EPMA. When the distribution of Cu elements in SG cast iron was only investigated by TE-EPMA, the lack of spatial resolution in TE-EPMA led to the difficulty of clarifying the difference of bright spots and shot noise in element-mapping images, as shown in Fig. 4. In contrast, FE-EPMA and FE-STEM can clarify the morphology and chemical composition of the Cu-enriched region in SG cast irons. Compared with EPMA, STEM can achieve higher spatial resolution in element mapping and chemical composition analysis by miniaturizing the electron probe size and suppressing the electron beam spreading region resulting from the thinning film effect in STEM specimens as shown in Figs. 7–9.
One of the weak points in STEM observation compared with EPMA was the difficulty of specimen preparation. FE-EPMA has superior features in terms of simple sample preparation compared with TEM and STEM, and simple measurement similar to TE-EPMA. WDS measurements are available for EPMA, while WDS measurement could not be performed in STEM. TE-EPMA, FE-EPMA, and FE-STEM have different characteristics and need to be used according to the specimens and the purpose for the observation. The FE-STEM element mapping used in this study corresponds to observation results with the lowest spatial resolution for FE-STEM. In the other words, the FE-STEM observation results shown in Figs. 7–9 had magnification close to the minimum setting values. This means that there is an area that cannot be covered in spatial resolution between TE-EPMA element mapping and FE-STEM element mapping. In contrast, FE-EPMA can acquire an element-mapping image whose spatial resolution can overlap with that in FE-STEM. The present study demonstrated not only that FE-STEM element mapping has a very high spatial resolution whose spatial resolution is much higher than that of FE-EPMA, but also that trans-scale observation combined with TE-EPMA, FE-EPMA, and FE-STEM is effective for the elemental distribution across the size range from micrometer order to nanoscale in SG cast irons. Finally, it should be emphasized that the high-spatial-resolution element mapping by FE-STEM and the trans-scale observation combined with TE-EPMA, FE-EPMA, and FE-STEM are effective for clarifying the elemental distribution not only of Cu, but also other elements in SG cast irons.
In this study, the distribution of Cu in SG cast iron containing Cu was elucidated using TE-EPMA, FE-EPMA, and FE-STEM. The conclusions obtained are summarized as the follows.
A part of this research was carried out with the support of a research grant from the Japan Foundry Engineering Society (JFS). The authors are grateful to Dr. T. Kaizu and Mr. K. Morita at Kobe University for their help with the STEM observation, Mr. K. Watanabe at Kansai University for help with the TE-EPMA analysis, and Mr. T. Uemura at Kindai University for help with the FE-EPMA analysis installed at the Joint Research Center of Kindai University.
