2025 年 65 巻 1 号 p. 133-141
This study identifies a method for shortening the duration of annealing in the grain boundary control process to achieve a high frequency of CSL boundaries in austenitic stainless steel by focusing on decreasing stacking fault energy. Si-added SUSXM15J1, which has significantly lower stacking fault energy, was used to examine the impact of a decreased stacking fault energy on the duration of annealing after cold-rolling, necessary to introduce a high frequency of CSL boundaries, by comparing it with SUS304 austenitic stainless steel. It was found that a decrease in stacking fault energy significantly contributed to shortening annealing duration. The frequency of the CSL boundaries in SUSXM15J1 increased from 55% to 75% through 5% cold rolling and subsequent annealing at 1323 K for only 60 s. Ex-situ and in-situ EBSD observations revealed that the strain-induced grain boundary migration, accompanied by the formation of twin boundaries, likely occurred in SUSXM15J1 compared to SUS304, as the recovery process was hindered by the lower stacking fault energy resulting from Si addition.
Austenitic stainless steels are widely utilized due to their desirable strength, ductility, corrosion resistance, and heat resistance. In practical applications, various alloying elements are incorporated to obtain their properties. For instance, Mo and Si are added to enhance corrosion and oxidation resistance, respectively.1) Because most materials are utilized in polycrystalline form, the grain boundaries are inevitably included and often act as preferential sites for corrosion and oxidation, for example. However, extensive studies have shown that the properties of grain boundaries, such as fracture and corrosion, depend on their character and structure. In particular, the coincidence site lattice (CSL) grain boundaries have been shown to possess a higher resistance against the intergranular degradation than random grain boundaries.2) This suggests that grain boundaries may not always have a deterimental influence on bulk properties. In 1983, Watanabe proposed a concept of the grain boundary design and control to a desired bulk properties,3,4) now commonly referred to as “grain boundary engineering”. He introduced a new parameter “grain boundary character distribution (GBCD) to quantify the fraction of grain boundaries with different character. Many researchers have been drawn to this concept, and experimental evidence has demonstrated that an increased fraction of CSL boundaries can enhance fracture toughness,5,6) fatigue strength,7) creep resistance,8) and reduce the susceptibility of materials to intergranular corrosion,9) stress corrosion cracking,10) and oxidation.11,12) Consequently, the grain boundary engineering is now widely accepted as a useful tool for achieving enhanced bulk properties of polycrystalline materials.
As for the grain boundary engineering type material processing, Palumbo et al.13) have first reported that the fraction of CSL boundaries can be increased by conducting cold rolling and annealing steps repeatedly for a Ni-base alloy with a low stacking fault energy (SFE), and showed that the grain boundary engineered materials have enhanced resistance to stress corrosion cracking. Consequently, a large number of studies on the grain boundary control via a similar processing route have been reported.14,15,16) In contrast, Shimada et al.2) have demonstrated the feasibility of increasing the fraction of CSL boundaries in the SUS304 (18%Cr-8%Ni) stainless steel through a single-step rolling and annealing processes, rather than the typically employed repeating process. They reported that 5% cold rolling followed by annealing at 1200 K for 72 h led to increase in the proportion of CSL boundaries from 65% to 86.5% and enhanced corrosion resistance. However, in the practical application of grain boundary engineering, a significant challenge is that a longer heat treatment is necessary to complete the processing. Consequently, there is a strong demand to reduce the duration of grain boundary engineering-type material processing from an industrial perspective to make grain boundary engineering feasible for practical manufacturing processes. Moreover, the majority of the previous research on grain boundary engineering for austenitic stainless steels utilized conventional SUS304, SUS316 or their low carbon series.2,7,8,9) There is a notable absence of reports17,18) on the effect of alloying elements on the grain boundary control. There are few papers on austenitic stainless steels other than the conventional steels mentioned above. We hypothesize that the grain boundary control may be more readily achieved in materials with low stacking fault energy. In this study, we attempted to control grain boundaries using SUSXM15J1 (19%Cr-13%Ni-3%Si), which has a lower stacking fault energy than typical heat-resistant austenitic stainless steels such as SUS304 steel. Our objective was to identify strategies for reducing the duration of grain boundary control process for achieving high frequency of CSL boundaries.
The materials used in this study were austenitic stainless steels SUSXM15J1 and SUS304, with their chemical compositions outlined in Table 1. The austenitic stainless steel SUSXM15J1 contains 3 mass% Si, aimed at enhancing oxidation resistance.19) The SFEs estimated using Pickering’s equation given by Eq. (1)20) are 6 mJ/m2 and 40 mJ/m2 for SUSXM15J1 and SUS304, respectively.
| (1) |
Therefore, particular attention was paid to the stacking fault energy to explore the possibility of reducing the duration of grain boundary control processing through a decrease in the stacking fault energy. SUSXM15J1 and SUS304 sheets, with dimensions of 300 mm length, 150 mm width, and 4.5 mm thickness, were used. These were cold-rolled either on a production mill or a laboratory mill, and subsequently annealed. These materials were referred to as “non-GBEM”. The average grain size at the surface and at the center position in the thickness direction of the sample was 36 μm for the non-GBEed SUSXM15J1 and 44 μm for SUS304.
| Steel | C | Si | Mn | Ni | Cr | N | SFE20) |
|---|---|---|---|---|---|---|---|
| SUSXM15J1 | 0.04 | 3.3 | 0.8 | 13 | 19 | 0.02 | 6 |
| SUS304 | 0.06 | 0.4 | 1.1 | 8 | 18 | 0.04 | 40 |
The non-GBEed samples were cold-rolled to 3%, 5%, 15%, 30%, and 50% reduction on a laboratory mill, and the cold-rolled sheets were annealed in a box furnace to control grain boundary character distribution. The annealing temperatures were 1223 K, 1323 K, and 1423 K, with an average heating rate of 3 K/s and a holding time at the annealing temperature of 60 s, followed by air-cooling. Additionally, some samples were annealed at 1323 K for up to 10 h to confirm the effect of the long-term annealing. Those materials subjected to the grain boundary control processing were referred to as the GBEM.
2.3. Microstructure ObservationsThe grain boundary microstructure such as the grain size and the grain boundary character distribution at the surface and the center of thickness for the Non-GBEM and the GBEM were evaluated, using scanning electron microscopy (SEM) and electron back scattered diffraction (EBSD). In this study, the grain boundaries with Σ1 to Σ29 coincidence relation were defined as the CSL boundaries.21) The frequency of CSL boundaries was determined by calculating the length ratio of the CSL boundaries to the total grain boundaries.
The SEM-EBSD analyses were performed on a JEOL Field Emission Scanning Electron Microscope (FE-SEM, JSM-7000F) equipped with TSL’s EBSD-OIM system, operating at an accelerating voltage of 25 kV, a current of 130 μA and a working distance of 14 mm. The electron beam was scanned with a step size of 1.0 μm.
To investigate the segregation of Si to the grain boundaries in non-GBEed SUSXM15J1, Cs-corrected Scanning Transmission Electron Microscope (Cs-TEM, JEOL’s NEOARM) with EDS (JED-2300T) was used. Specimens for Cs-STEM observation were obtained from the center of sample thickness, and mechanically polished to several 10 μm thickness. Subsequently, the specimens were trepanned into 3 mm disk, and then thinned using twin-jet electropolishing.
2.4. In-situ EBSD Measurements at High-temperatureThe in-situ EBSD observations at high-temperature were performed for SUSXM15J1 and SUS304 using a HSEA-1000 heating stage on JEOL’s FE-SEM equipped with TSL’s EBSD-OIM system. The samples were 1 mm thick sheets cold-rolled to 5% reduction, contacting to electric heater on a backside of the analyzed surface. The temperature of the sample was measured using a type-K thermocouple (Chromel/Alumel) attached to the surface of the sample near the analyzed area. Figure 1 shows the history of setting temperature and sample temperature during in-situ observations. The procedure of heating to 1253 K, holding for 30 s, and lowering the temperature of 673 K for EBSD observation at 673 K, where no microstructural change occurs,22) was repeated for eight sets. The EBSD analysis with 1.0 μm in step size was performed on the same area during heating.
To confirm the effect of grain boundary engineering, the intergranular corrosion resistance for Non-GBEM and GBEM in SUSXM15J1 was evaluated using 10% oxalic acid etching tests (JIS G 0571) after sensitization at 923 K for 60 min and air cooling. Microstructural observations of the samples after 10% oxalic acid etching tests were carried out by means of optical microscope.
Figure 2 shows the grain boundary (GB) maps obtained at the surface and at the center along the thickness direction for SUSXM15J1 and SUS304. Figures 2(a)–2(d) exhibit the GB map at the surface and Figs. 2(e)–2(h) at the center of the specimens. In these figures, low-angle grain boundary, CSL grain boundary and random grain boundary are represented by blue, red and black lines, respectively. These maps represent of a single phase austenite structure where annealing twins were found in the grains. Comparing Figs. 2(a) and 2(e), this material appears to have a homogeneous microstructure throughout the thickness. The average grain size at the surface and at the center was 36 μm when twins were excluded, and 22 μm when they were included.
The frequencies of CSL boundaries (FCSL) of the surface and center were 50% and 59%, respectively, with an average of 55%. Figures 2(b) and 2(f) show the GB maps at the surface and at the center, respectively, for GBEed SUSXM15J1, which was cold-rolled at 5% reduction and subsequently annealed at 1323 K for 60 s, producing a homogeneous microstructure throughout the thickness. The average grain size at the surface and at the center was 110 μm when twins were excluded, and 35 μm when twins were included. The frequencies of CSL boundaries of the surface and centers were 75% and 76%, respectively, with an average of 75%. The grain size of the GBEM was found to be coarser than that of the Non-GBEM. Surprisingly the frequency of CSL boundary considerably increased despite the short duration (60 s) of annealing after cold rolling. Figure 3 presents the grain boundary character distribution of non-GBEed and GBEed SUSXM15J1 subjected to a cold rolling at a 5% reduction and subsequent annealing at 1323 K for 60 s. The frequencies of the Σ3, Σ9, Σ27 and random boundaries in non-GBEM were 48%, 1%, 0.2%, and 37%, respectively. In contrast, those in GBEM were 64%, 5%, 3%, and 20%, respectively.
In SUS304, the average grain size at the surface and at the center was 44 μm when twins were excluded, and 31 μm when twins were included, as shown in Figs. 2(c) and 2(g). The frequency of CSL boundaries was 53% in both the surface and center. Prior to the GBE process, the frequency of the CSL boundaries exhibited a only 2% difference between SUSXM15J1 and SUS304. Figures 2(d) and 2(h) represent the GB maps of the surface and center of the GBEed SUS304 processed with the same conditions as the GBEed SUSXM15J1. The average grain size was 86 μm at the surface and at the center when twins were excluded, and 42 μm when twins were included. The grain size, excluding twins, was smaller in the GBEed SUS304 than in the GBEed SUSXM15J1, though the initial grain size in SUS304 was slightly larger than in SUSXM15J1. These findings suggest that the grain boundary migration was more likely to occur in SUSXM15J1 than in SUS304. In addition, the frequencies of CSL boundaries at the surface and at the center were 73% and 62%, respectively, with an average of 68% in GBEed SUS304, whereas in the Non-GBEed SUS304, it was 53%. The frequency of CSL boundary increased in SUS304 due to the GBE process. However, the rate of increase in the frequency of CSL boundaries is greater in SUSXM15J1 than in SUS304. This can be attributed to Si addition, which promotes increase in grain boundary migration.
3.2. Effects of Cold Rolling Reduction and Annealing Temperature on Frequency of CSL BoundariesFigure 4 shows the effects of the cold-rolling reduction and the annealing temperature on the frequency of CSL boundaries in SUSXM15J1, with an annealing time is 60 s. Data for SUS304 annealed at 1323 K are also included for comparison. At an annealing temperature of 1223 K, the frequency of CSL boundaries increased by a few percent for cold-rolling reduction of 3% and 5% compared to those in non-GBEM. However, it exhibited a maximum value of approximately 70% at a 15% reduction and decreased with further increase in the reduction ratio. Previous studies have suggested that the strain-induced grain boundary migration is accompanied by the formation of the twin boundaries.22,23,24) Thus, the strain introduced by cold-rolling at less than a 15% reduction may act as a driving force for grain boundary migration, leading to an increase in CSL boundaries frequency. However, beyond the 15% reduction, the strain promotes recrystallization rather than the grain boundary migration.
At annealing temperature of 1323 K, the frequency of CSL boundaries reached a maximum value of approximately 75% at a 5% cold-rolling reduction in SUSXM15J1, and conversely decreasing with further increases in the reduction ratio. Notably, the frequency of CSL boundaries in the SUSXM15J1 increased from 55% to 75% with short annealing duration of only 60 s. Importantly, the peak value in the SUSXM15J1 was considerably higher than that in SUS304, suggesting that the decrease in the stacking fault energy due to Si addition contributes to an increase in CSL boundary frequency. The decrease in CSL boundary frequency at a 15% reduction may result from the recrystallization due to high strain energy, as the SUSXM15J1 cold-rolled at a 15% reduction possessed a much smaller grain size of 34 μm compared to that cold-rolled at a 5% reduction with a grain size of 110 μm. Occurring recrystallization may have reverted the microstructure to its initial state with a low frequency of CSL boundaries. For annealing at 1423 K, changes in CSL boundary frequency with the cold rolling reduction showed a similar trend to those at 1323 K, but the maximum frequency of CSL boundaries in the SUSXM15J1 sample was smaller at 1423 K than at 1323 K.
3.3. Effect of Annealing Time on Frequency of CSL Boundaries and Grain GrowthTo assess more effective conditions for further increasing the frequency of CSL boundaries, the annealing of 5% cold-rolled SUSXM15J1 and SUS304 were extended to longer duration at 1323 K. Figure 5 represents changes in the frequency of CSL boundaries and the average grain size, excluding twin boundaries, as a function of annealing time. The frequencies of CSL boundaries were observed to increase up to 75% and 66% in SUSXM15J1 and SUS304, respectively, after annealing for 60 s. However, there was only a slight increase in the CSL boundary frequency with further annealing. Consequently, the CSL boundary frequency in SUSXM15J1 reached to approximately 80% after 15 min of annealing, whereas in SUS304 it reached 71% even with the annealing time extended to 10 h. These results indicate that prolonged annealing does not necessarily result in an increase in the frequency of CSL boundaries. Regarding the grain size, the grain size of both samples increased with increasing of annealing time at 1323 K in a parabolical manner, but grain growth of SUSXM15J1 during annealing was faster than that of SUS304. The difference in grain growth will be discussed in the next section, along with the results of in-situ EBSD observation.
The in-situ EBSD analysis was performed at high-temperatures on the specimen cold-rolled to a 5% reduction to reveal the microstructural change during grain boundary control process in detail. Figures 6(a)–6(c) show the IPF maps of SUSXM15J1 with the following observations: (a) as-cold rolled, (b) and (c) exposed at 1253 K for 1.3 min and 2.6 min, respectively. The arrows in these figures indicate the areas where observable microstructure changes occurred due to heating. It is observed that the migration of high-angle grain boundaries, accompanied by the formation of new annealing twins. On the other hand, Figs. 6(d)–6(f) show the IPF maps of SUS304 as-cold rolled, exposed at 1253 K for 1.3 min and 2.6 min, respectively. There was little change in the microstructure of SUS304 due to heating compared to SUSXM15J1.
Figure 7 shows the Kernel Average Misorientation (KAM) maps for as-cold rolled SUSXM15J1 and SUS304, as well as for the samples exposed at 1253 K for 1.3 min and 2.6 min. The KAM value corresponds to the dislocation density of geometrically necessary dislocations (GN dislocations). In these maps white-colored areas indicate high KAM values, which indicate a high dislocation density. It is observed that a considerable amount of strain was introduced near the grain boundaries in cold-rolled sheets. The KAM maps of SUSXM15J1, as shown in Figs. 7(a)–7(c), exhibited that the grains with low KAM values, indicated by arrows in the figure, underwent successively expansion with increasing holding time at 1253 K, which suggests that strain-induced grain migration occurred. A noteworthy finding is that the grain boundary migration resulted in an increase in CSL grain boundaries, particularly the Σ3 grain boundary. In contrast, as shown in Figs. 7(d)–7(f), the microstructure and the KAM values in SUS304 exhibits less changes even after holding for 2.6 min. The in-situ EBSD observations suggested that the movement of high-angle random boundaries occurs more easily in SUSXM15J1 than in SUS304 at 1253 K after cold rolling, and that the migration of random boundaries may contribute to an increase in the frequency of CSL grain boundaries.
The more quantitative analysis of the in-situ EBSD observations was conducted. Figure 8 shows the changes in the frequency of CSL boundaries and the grain sizes, excluding twin boundaries, in the SUSXM15J1 and SUS304 samples during in-situ EBSD observations at 1253 K. As observed in the ex-situ observations, the grain size and the frequency of CSL boundaries exhibited concurrent increase with increasing exposure time at 1253 K for both steels. However, these changes occurred at a faster rate in the SUSXM15J1 than in SUS304. In particular, the frequencies of CSL boundaries were found to increase by up to 62% and 48% in SUSXM15J1 and SUS304, respectively, with annealing at 1253 K for 2.6 min, and in SUS304 only increased to 55% even with an extended time to 9 min. The average grain sizes, excluding twin boundaries, were found to increase up to 124 μm and 59 μm in SUSXM15J1 and SUS304, respectively, with annealing for 2.6 min. In SUS304, the grain size increased to 106 μm with an extended annealing time of 9 min. Figure 9 shows the effect of exposure time at 1253 K on the number of microstructure change points of grain boundaries in the SUSXM15J1 and SUS304 samples cold rolled at a 5% reduction rate. Here, the number of microstructural changes was defined as the number where grain boundary migration occurred within the EBSD observation area. After an annealing time of 2.6 min, the number of microstructural changes in SUSXM15J1 was sixteen times higher than in SUS304, leading us to conclude that the grain boundary migration occurred more readily in SUSXM15J1 than in SUS304. Ex-situ and in-situ annealing observations confirmed that the strain-induced boundary migrations lead to significantly faster grain growth in SUSXM15J1 compared to SUS304. Consequently, the grain boundary control process achieved a CSL boundary frequency exceeding 75% in SUSXM15J1 within a fairly short time. Unlike recrystallization, which involves nucleation and growth, the strain-induced migration of high-angle grain boundaries in materials with low-stacking fault energy is accompanied by formation of twins, where the resulting twin boundaries are Σ3 grain boundaries in cubic structures, as confirmed by both previous study22) and this study. Specifically, in cases where the strain-induced boundary migration proceeds rapidly, as observed in SUSXM15J1, twins formation also occurs rapidly, leading to a significant increase in frequency of Σ3 grain boundaries within a short period.
To confirm whether the strain-induced grain boundary migration had already started prior to reaching the annealing temperature, the 5% cold-rolled samples were water-cooled immediately after reaching 1323 K, followed by SEM-EBSD observations, as exhibited in Fig. 10. The KAM maps indicate that some grains without strain are present, as indicated by arrows in SUSXM15J1, suggesting that the strain-induced grain boundary migration may have occurred to some extent during heating to 1323 K. In contrast, no such areas were observed in SUS304.
Figure 11 shows the optical micrographs of the surfaces and the cross-sections for non-GBEed and GBEed (5% cold rolling and annealing at 1323 K for 60 s) SUSXM15J1 after the sensitization heat-treatment at 923 K for 60 min and 10% oxalic acid etching test. In the non-GBEM, most grain boundaries were severely etched and formed ditches. Thus, the intergranular corrosion susceptibility was markedly high. In contrast, in the GBEM not all grain boundaries were found to be corroded. In addition, the intergranular corrosion depth (5-point average) in the GBEM was measured to be 7 ±1.9 μm, whereas the depth in the non-GBEM was found to be 10 ± 2.6 μm. Therefore, this study using 3% Si-added austenitic stainless steel, as in the previous study,9) confirmed the beneficial effect of grain boundary control on suppressing intergranular corrosion. The GBEM has a larger amount of CSL boundaries which are immune to sensitization due to precipitation of Cr carbides. The high intergranular corrosion resistance can be attributed to the severing of the random boundary network by CSL boundaries and suppression of the precipitation of Cr carbides during sensitization annealing.2)
This study revealed that in SUSXM15J1 (19%Cr-13%Ni-3%Si), a typical heat-resistant austenitic stainless steel comparable to SUS304 but including 3% Si, a high CSL boundary frequency of more than 75% can be attained in an annealing time of 60 s through a grain boundary control process. Furthermore, the strain-induced grain boundary migration, which resulted in the formation of the twin boundaries (Σ3 grain boundary), was found to occur readily in SUSXM15J1 compared to SUS304. Consequently, it is considered that the addition of Si will mainly contribute to a decrease in annealing time for achieving a significant increase in CSL boundary frequency. Therefore, in this section, we discuss the origin of the influence of Si addition on the observed phenomena.
4.1. Grain Boundary SegregationA solute element sometimes segregates to the grain boundaries and often cause a decrease in the mobility of grain boundary migration due to the dragging solute atmosphere, as observed in Fe-3%Si steel.25,26) However, grain boundary segregation of solute atoms does not necessarily result in a decrease in the mobility of grain boundary migration. Indeed, there is a report that demonstrate an acceleration of grain boundary motion in Al by addition of Ga.27) In any case, it is important to confirm whether Si segregates to grain boundaries in austenitic stainless steel with fcc structure. Thus, chemical analyses were performed in the vicinity of the grain boundary in non-GBEed SUSXM15J1 using EDS equipped with Cs-corrected STEM. Figure 12 shows an example of the chemical distribution profiles near a high-angle grain boundary in a non-sensitized SUSXM15J1. Although Cr was found to segregate to the grain boundary, this is commonly observed in austenitic stainless steels.28) However, there is little evidence of Si segregation to the grain boundary in SUSXM15J1, consistent with the several reports that Si does not segregates to the grain boundary in solution treated austenitic stainless steel.28,29) Therefore, it cannot be concluded that the addition of Si segregates to grain boundaries and accelerate the grain boundary motion.
From the experimental results, the addition of Si can influence the strain-induced grain boundary migration. According to Turnbull’s theory,30) the grain boundary migration rate can be expressed as follows:
| (2) |
where v the grain boundary migration rate, ΔF the free energy difference when an atom is incorporated in the next grain by grain boundary migration, λ the width of grain boundary, DG the diffusion coefficient of mass-transfer with grain boundary migration, R the gas constant, and T the absolute temperature. According to Eq. (2), when the free energy difference between adjacent grains is large, the grain boundary migration tends to occur in the direction of low energy, and the grain boundary migration occurs into the grain with higher strain to reduce the strain energy. As mentioned in section 2.1, the addition of Si will contribute to a decrease in the stacking fault energy, and thus SUSXM15J1 has a lower SFE than SUS304. A reduction in the SFE results in a greater difficulty in recovery due to a large extend of partial dislocations.31) Consequently, the strain introduced by cold rolling prior to annealing may have remained to a greater extent in SUSXM15J1 than in SUS304 during annealing, and it can act as a driving force for grain boundary migration. In order to confirm this consideration, Fig. 13 compares the distributions of KAM values in the areas where no grain growth occurred in Fig. 10 between SUSXM15J1 and SUS304. Figure 13 reveals that in the higher range of KAM values of approximately 0.4, the fraction of KAM values is lower for SUS 304 than for SUSXM1J1. Conversely, in the lower range of KAM values of approximately 0.3, the fraction of KAM values is higher for SUS304 than for SUSXM15J1. This finding suggests that strain recovery is less likely to occur in SUSXM15J1 than in SUS304. In other words, because SUSXM15J1 has a larger ΔF at high-temperatures than SUS304, the strain-induced grain boundary migration is more pronounced in SUSXM15J1 than in SUS304, resulting in a high frequency of CSL boundaries being achieved in a short annealing time. It can therefore be concluded that the effectiveness of Si addition in reducing the annealing time to introduce high frequency of CSL boundaries is attributed to prevention of recovery by the reduction of stacking fault energy.
In this study, we identify a method to reduce the duration of annealing in the grain boundary control process to achieve a high frequency of CSL boundaries in austenitic stainless steel by focusing on decreasing stacking fault energy. We employed Si-added austenitic stainless steel SUSXM15J1, which has significantly lower stacking fault energy compared to conventional austenitic stainless steels such as SUS304. We examined the impact of this reduction in stacking fault energy on the duration of annealing in grain boundary control process necessary to achieve a high frequency of the CSL boundaries. The main results obtained were as follows.
(1) The frequency of the CSL boundaries in SUSXM15J1 increased from 55% to 75% through a 5% cold rolling and subsequent annealing at 1323 K for only 60 s. This improvement was more pronounced than that observed in SUS 304, indicating that a decrease in stacking fault energy is more advantageous to shortening the annealing duration in the grain boundary control process.
(2) We compared the grain boundary migration between SUSXM15J1 and SUS304, considering why SUSXM15J1, with its lower stacking fault energy due to the addition of Si, can increase the frequency of CSL boundaries in a fairly short time. Ex-situ and in-situ EBSD observations revealed that SUSXM15J1exhibited faster grain growth compared to SUS304. It was observed that, along with grain boundary migration, the formation of twin boundaries (Σ3 grain boundaries) occurred, which contributed to the increase in the frequency of CSL boundaries. This phenomenon is likely due to strain-induced grain boundary migration in SUSXM15J1, because the recovery process hindered by lower stacking fault energy resulting from Si addition.
(3) The effect of intergranular corrosion suppression was confirmed in the short-time grain boundary control process for SUSXM15J1.
The authors declare no conflict of interest.
The authors express their gratitude to Mr. N. Ikeda, a graduate student at Kumamoto University, for his assistance with certain experiments. Furthermore, two authors (C. T. and J-I. H.) extend their sincere thanks to Ms. M. Ikado of NIPPON STEEL Stainless Steel Corporation, for her supporting EBSD measurements.
