Transformations and Microstructures
Effect of Cold Rolling on Stability of HCP and FCC Phases in Fe–Mn Alloys
2020 Volume 60 Issue 2 Pages 359-368
Details
2020 Volume 60 Issue 2 Pages 359-368
The phase transformation behavior of an Fe–20%Mn alloy during a heating process after various cold-rolling reductions was investigated, and the phase stabilities of the γ and ε phases were discussed. The initial hot-rolled material was composed of an ε martensite matrix and a small amount of the γ austenite phase at room temperature. The deformation of the martensite alloy in the cold rolling was not homogeneous, and the microstructure of some regions was clearly adopted from that in the hot-rolled sample. Moreover, a residual γ phase was still detected even after 35% cold-rolling reduction. In the heating stage, a remarkable reverse transformation to the γ phase started at 200°C or higher, and its finishing temperature clearly increased with the rolling reduction ratio. However, the in situ X-ray diffraction and electron back scatter diffraction (EBSD) observations revealed that the reverse transformation had already started from the residual γ phase particles even at temperatures below 200°C. In addition, from the EBSD–image-quality map, the distribution of the dislocations was considered to remain in the γ phase even after the reverse transformation.
Manganese (Mn) is an important alloying element that is widely used in high-strength steels together with carbon because it lowers the α’-BCC (BCT) martensitic transformation temperature (Ms) of iron and stabilizes the γ-FCC austenite phase. Furthermore, when the concentration of Mn increases, the alloy exhibits a stable non-magnetic property at extremely low temperatures in combination with excellent mechanical properties. Thus, high-Mn non-magnetic steel has attracted much attention, and various types of steels have been developed.1)
Recently, twinning-induced plasticity (TWIP) steel sheets have been studied and developed by adding 20% or more of Mn to obtain an austenite single phase, improving the work-hardenability by deformation twinning.2,3,4,5,6,7) Because TWIP steel sheets have unprecedented formability together with a high strength, their development for automotive use is underway because these automotive applications require excellent press formability. Therefore, extensive research has been conducted on the microstructural changes due to deformation. Although the formability of medium-Mn steel sheets with an Mn content of about 10% is not as high as that of TWIP steel sheets, they display a better balance of strength and ductility than the conventional dual-phase (DP) steel and TRIP transformation-induced plasticity (TRIP) steel when the amount of Mn is adjusted from the perspective of weldability, corrosion resistance, and castability. This type of alloy sheet is called a third-generation high-strength steel sheet.8)
Because Mn stabilizes austenite and lowers the stacking fault energy of iron, ε martensite with a hexagonal close packed (HCP) structure may be generated in the manufacturing process.9,10) When cooled from a high temperature in an austenite single-phase state, a γ/ε martensitic transformation may occur at temperatures in the range from 100 to 200°C, and the presence of ε martensite is thought to deteriorate the formability. Therefore, from the viewpoint of formability, a steel chemistry and/or production method that does not generate ε martensite has been adopted.
On the contrary, numerous studies have examined Fe–Mn–Si-based ferrous shape memory alloys, damping alloys, and other materials utilizing ε martensite, and some of the developed alloys have been used in practical use.1,11,12,13,14,15,16,17) For example, Watanabe et al.12,13) reported the improvement of the damping performance (damping capacity) by the training effect of repeated heat treatments of composite microstructures including the austenite and ε phases in systems such as Fe–20Mn. In addition, Sawaguchi et al. conducted research on shape memory alloys in an Fe–Si–Mn multicomponent system, and those alloys have been applied to seismic damper materials for high-rise buildings.14,15,16,17) As described above, in high-Mn austenitic steels and Fe–Mn–Si shape memory alloys, the martensitic transformation from γ austenite to ε martensite provides the materials various high functions and excellent mechanical properties. In particular, the γ/ε duplex microstructure developed under reversal cyclic loading is reported to be effective for improving low-cycle fatigue properties because it responds reversibly to repeated loadings.
Therefore, the stability of the ε (martensite) phase is important for iron-based high-Mn alloys. The stability of austenite and ε martensite was studied thermodynamically by Ishida et al.,18) and systematic studies were conducted on the influence of adding a third element on the stability. Recently, phase stability has been discussed by ab initio calculations,19) where an equilibrium state was assumed. Because the above-mentioned material is processed by a combination of cold working and heat treatment in the actual process, further study concerning the stabilities of austenite and ε martensite is necessary for application to a practical production process. For Fe–Si–Mn-based alloys, Tasaki et al. studied the effect of a tensile deformation of up to 10% on the phase transformation occurring in the subsequent heat treatment; however, it is necessary to investigate the influence on the phase transformation under other practical deformations for relatively heavier processing conditions.
In this study, to discuss the stabilities of austenite and ε martensite in a high-Mn ferrous alloy, the phase transformation and microstructural change were investigated by in situ X-ray diffraction and in situ electron back scatter diffraction (EBSD) measurements at high temperatures. The effect of cold rolling on the phase stability was discussed by varying the cold-rolling reduction ratio.
The chemical composition of the steel used was Fe–20 mass% Mn (0.001% C–0.048% Si–19.8% Mn–0.022% P–0.0034% S–0.023% Al–0.0028% N). The sample steel was prepared by vacuum melting, and the resulting ingot was hot forged and hot rolled to a thickness of 3.5 mm. The sheet was then reheated to 1000°C for 20 min and reduced to a finishing thickness of 2 mm by a second hot rolling. The finishing temperature was 900°C. The hot-rolled sheet was pickled, surface-polished, and then cold rolled at various rolling reduction ratios from 0 to 35%. When rolled at a rolling reduction ratio of 35% or more, a transverse crack expanded from the end face of the sheet, and fracture occurred on the delivery side of the rolling mill. Each cold-rolled sheet was subjected to the following evaluations.
Microstructure observation was conducted on samples heat treated in an electric furnace kept at 200, 300 or 400°C for 300 s. For the preparation of the specimens for the microstructure observation, the surface was finally polished by using an OPU colloidal silica suspension (Struers Corp.), and backscattered electron (BSE) images were captured with a LEO Gemini 1530 (Carl Zeiss) scanning electron microscope (SEM). Transmission electron microscopic observation was performed on both the hot-rolled and cold-rolled sheets. During the sample preparation, the thickness was reduced to about 60 to 70 μm by wet polishing, and the thin film was jet polished. Tenpole 5 was used for electrolytic polishing, and the liquid temperature was set as −35°C with an A-3 electrolytic solution (perchloric acid, 2-n-butoxyethanol, methanol). This thin film sample was observed with a Titan (FEI) 300 kV scanning transmission electron microscope (STEM) to observe the substructure from the sheet surface direction. The heat-treated materials were also subjected to Vickers hardness tests. The testing machine was a Vickers hardness tester (Akashi), and the test load was 500 g.
Thermoanalysis was conducted to observe the annealing process after cold rolling using a DSC 404 C (NETZSCH) differential scanning calorimeter (DSC). The sample was heated to 600°C at a heating rate of 10 K/min, held at that temperature for 30 min, and then cooled to room temperature at 10 K/min. Furthermore, for the material with a high rolling reduction ratio of 35%, the above-mentioned cycle was repeated thrice to investigate the variation in the transformation temperature. The DSC measurements were performed in an Ar atmosphere, and the Ar flow rate was 50 mL/min. The Vickers hardness test (load: 500 g) before and after the heat treatment was also performed in the same manner as described above.
An in situ X-ray diffractometer (SmartLab (3 kW; Rigaku)) was used to confirm the phase transformation corresponding to the exothermic/endothermic peaks obtained by the DSC measurements at high temperatures. The sample was cut to a size of about 9 × 12 mm from a hot-rolled sheet that had been ground to a thickness of 1 mm on one side and a cold-rolled sheet (0.85 mm) with 35% rolling reduction. The measurement surface was polished and mounted on the stage with Si grease. The heating and cooling rates were set as 10 K/min. After reaching the predetermined temperatures, a 5-min holding was conducted before the X-ray diffraction measurements. The measurement temperatures were 20, 100, 150, 200, 250, 300, 350, and 400°C for the heating process and 350, 300, 250, 200, 150, 100, and 30°C for the cooling process. The X-ray diffraction measurements were performed by the parallel-beam method; the sample rolling direction was set perpendicular to that of the incident X-ray.
Furthermore, in situ EBSD measurements were conducted to observe the crystal orientation and strain accumulation state of each grain in the heating process. A cold-rolled sheet (0.85 mm thick) with a 35% rolling reduction was used for the measurement. Surface preparation of the sample mold was performed by paper grinding up to #1000, diamond buffing of 0.25 μm, and vibration polishing using an OP-U colloidal silica suspension. The sample was then mounted on a heating stage (HSEA-1000; TSL Solutions Company). The sample was heated to 450°C, and the EBSD scan measurement was performed from 200°C with 50°C increment in a device equipped with a field emission scanning electron microscope (XL-30 FEG; Philips Co., Ltd.) with an orientation imaging microscopy (OIM) detector (Hikari Super, TSL). The degree of vacuum in the SEM sample chamber was about 3.0 × 10−6 Pa. To avoid changes in the microstructure during a scanning measurement, the measurement time was adjusted to be within 10 min. Specifically, a 20 × 20 μm region was measured for a measurement, with a step size of 0.3 μm. Because the image may drift when the temperature changes, the scanning position was adjusted before each measurement. The OIM Data Collection (OIM-DC) and OIM-Analysis 6.2 (TSL) software were used for the data collection and analysis, respectively.
Figure 1 shows the BSE micrographs of the material after applying different hot-rolling, cold-rolling, and heat-treatment conditions. As shown in Fig. 1(a), the hot-rolled material mainly exhibits a microstructure of band-shaped structures of about 1 to 2 μm that overlap each other. In the present Fe–20% Mn alloy, the austenite parent phase transforms into ε martensite during cooling after the hot rolling, as described subsequently. The band structures in Fig. 1(a) are the result of the interweaving of the {0001} ε (HCP) variants, which originate from the {111} planes of FCC.
BSE micrographs of the hot-rolled, cold-rolled, and annealed sheets: (a) as hot-rolled, (b) as 35% cold-rolled, and (c), (d), and (e) annealed at 200, 300, and 400°C, respectively, for 300 s after cold-rolling, respectively. Vickers hardness is also indicated in each figure.
When the hot-rolled material was cold rolled, the microstructure became refined, showing a fiber-like contrast, as shown in Fig. 1(b). However, regions with the band-structure feature can also be seen similar to the hot-rolled material. The undeformed regions are found surrounded by fiber-like heavily deformed regions. Even after the heat treatment at 200°C (Fig. 1(c)) or 300°C (Fig. 1(d)), this type of heterogeneous microstructure remains. Although some crystal orientation dispersion remains, coarse grains appear in the material after the heat treatment at 400°C, as shown in Fig. 1(e); this is apparently different from the hot-rolled microstructure. The Vickers hardness values are shown in the figures. The cold rolling at 35% reduction results in a remarkable increase of 100 points of the hardness. This is because of the fact that the slip system of HCP is restricted to the basal slip system; thus, the work hardening rate is considered to be higher than those of the conventional steels.20) Conventionally, for a cold-rolled material, softening is expected during the subsequent annealing processes; but, a slight hardening effect is found at 200°C (Fig. 1(c)). The hardness decreases as the heat-treatment temperature increases, although the final hardness (354 HV) is still high, which is a value between those of the hot-rolled and cold-rolled materials.
3.2. Thermal Analysis and Phase Transformation Behavior by High-temperature X-ray DiffractionThe results of the DSC measurements conducted on the hot-rolled and 5%, 10%, 21%, and 35% cold-rolled materials are shown in Fig. 2. As shown in the curves of the heating process (Fig. 2(a)), an endothermic peak appears in each sample above 200°C. This is considered to correspond to the reverse transformation from ε martensite to γ. The peaks are clear, particularly in the hot-rolled material; with increasing cold-rolling reduction ratio, the peaks become broader and their positions shift to the high-temperature side. For an Fe–28Mn–6Si–5Cr (mass%) alloy, it was reported by Tasaki et al. that14) the endothermic peak in the heating stage was unclear under a 10% tensile deformation. However, in this study of Fe–20% Mn, an obvious peak is observed even under a 35% cold rolling reduction ratio. The DSC curves of the cooling process are shown in Fig. 2(b). The hot-rolled specimen exhibits an exothermal peak at about 70°C; the peak becomes broad under cold-rolling reduction ratios of 10% or more. The transformation finishing temperature (Mf) is unclear and cannot be determined. Assuming the Ms temperature as the point rising slightly from the high temperature baseline, the influence of the cold reduction on Ms is small (122°C to 126°C).
(a) Heating and (b) cooling DSC curves for the hot-rolled and cold-rolled samples, where the reverse transformation starting and finishing (As and Af) temperatures and forward transformation starting (Ms) temperature are indicated by arrows.
The reverse transformation start temperature (As) is practically constant or rises slightly (from 214°C to 235°C) with increasing cold rolling ratio. In comparison, the reverse transformation finish point (Af) increases monotonically with increasing cold rolling reduction ratio (244°C to 325°C). Ms, As, and Af are plotted against the cold reduction ratio in Fig. 3.
As, Af, and Ms temperatures against the cold-rolling reduction. The Af points apparently increase with increasing reduction, whereas the As and Ms points are nearly constant. (Online version in color.)
To examine the influence of the training effect on the transformation behavior, the DSC measurements were conducted three times between room temperature and 600°C on the 35% rolled material. The results are shown in Fig. 4. As the measurement repeats, the characteristic peaks of heating become unclear and the transformation temperature shows a tendency to change toward the low temperature side.
Cyclic DSC curves between room temperature and 600°C for the 35% cold-rolled specimen. The As, Af, and Ms temperatures are indicated by arrows. The transformation peaks become weaker with the heat cycle.
The results of the Vickers hardness before and after the DSC measurements are shown in Fig. 5. As shown in Fig. 1, the hardness increases remarkably by cold rolling. Regarding the change in the hardness due to the cold reduction, work hardening tends to be high, particularly in the initial stage of the cold rolling. After heat treatment, the hardness of the hot-rolled sheet does not change or is slightly decreased by about 10 points, from 324 HV to 313 HV. However, in the cold-rolled specimen, the difference in the hardness before and after the heat treatment widens on increasing the rolling reduction ratio. Softening of the cold-rolled material occurs until 10% rolling reduction, but the hardness increases compared to the hardness of the hot-rolled sheet. (At the rolling reduction of 10%, the hardness decreases from 369 HV to 324 HV, or 45 points due to the heat treatment). The hardness of the 20% cold-rolled material decreases to less than 300 HV after the heat treatment, it being softer than the hot-rolled material. Considering that the hardness of the heat-treated material is softened to the level of the hot-rolled material under the conditions of cold-rolling reduction of more than 20% or of repeated heat treatment, a structural change, such as recrystallization of the γ phase, may have occurred during heating up to 600°C.
Change in Vickers hardness of the cold-rolled sheets before and after the DSC measurement. The decrease in the hardness after annealing is more pronounced at reduction rates higher than 20%. The changes in Vickers hardness of the cold-rolled sheets before and after the DSC measurements are plotted as closed and open circles, respectively. (Online version in color.)
Because the DSC peak was attributed to the phase transformation, the change in the crystal structure during heat treatment was investigated by in situ X-ray diffraction. Figure 6 shows the X-ray peak profiles at various measurement temperatures. The measurement surface normal corresponds to the sheet plane normal direction. Fig. 6(a) corresponds to the hot-rolled material and (b) to the cold-rolled material (35% rolling reduction). In both the specimens, the main phase is the HCP (ε phase) in the initial step before heating. The calculated lattice constants are: a = 2.524 Å, c = 4.071 Å; thus, the c/a ratio is 1.613, which is close to that of magnesium (1.625).21) In the HCP (ε) phase, peaks of
Change in the X-ray diffraction profiles during heat cycle: (a) hot-rolled and (b) 35% cold-rolled specimens, where the initial crystal texture is affected by the cold rolling. Although the peaks from the ε phase decrease with temperature in the both cases, the ε phase in the cold-rolled specimen remains even at high temperatures.
The change in the X-ray profile during heat treatment will be discussed. In Fig. 6(a) showing the case of the hot-rolled sheet, the peaks corresponding to the ε phase disappear when the temperature is above 200°C, whereas the (111) and (200) peaks of γ appear above 250°C. It is difficult to distinguish the γ (220) and (311) peaks from the peaks of the ε phase because they are located at practically the same positions. During cooling, the ε
For the 35% cold-rolled material (Fig. 6(b)), ε
For a more detailed study of the change in the X-ray profile during heat treatment, we focused on the ε
Temperature dependence of the
In situ heating observation via EBSD were performed on a cold-rolled sheet with 35% rolling reduction. The measurement consisted of heating up to 400°C and cooling down to 50°C. Although this was a low-temperature heat treatment, some effects such as the oxidation and evaporation of the Mn from the surface due to the vacuum condition may have occurred; a sharp Kikuchi pattern could not be obtained as the experiments proceeded. Because the transformation behavior and state of strain in the grains could be evaluated to some extent at temperatures up to 300°C, we focused only on the analysis of this heating stage.
Figure 8 shows the image quality (IQ) and inverse pole figure (IPF) maps of the ε phase and γ phase in the heating process. The reference direction of the IPF map is the normal direction of the measured surface. The in situ measurements are made in the area surrounded by the squares shown in Fig. 8(a).
Inverse pole-figure (IPF) maps of the ε and γ phases and image quality (IQ) maps at (a) RT, (b) 200, (c) 250, (d) 300, and (e) 350°C during heating, as obtained by the in situ EBSD measurement for the 35% cold-rolled sheet. The transformation from the ε to γ phase is obviously observed at 300°C. The transformed area at 300°C corresponds to the regions with a high value of brightness, which have a relatively high IQ at room temperature.
The IPF map shows that the fraction of the γ phase increases rapidly above 300°C, and practically the whole measured area transforms to the γ phase at 350°C. IQ value of the cold-rolled sheet is not uniform. This suggests the existence of a region with less deformation, similar to the band-like microstructure in the cold-rolled materials, as shown in Fig. 1. Such regions with high IQ values (white in the map) are surrounded by the region with a low IQ (black) value. Because the IQ value corresponds to the local deformation to some extent, There are coexisting regions with different dislocation densities in the sample. Comparing the IQ map of the cold-rolled material with that of the material heated to 300°C, at which the γ phase appears, the former clearly shows that ε crystal grains with a high IQ value, i.e., with a lower dislocation density, become the primary nucleation sites for the transformation to γ. This corresponds with the results of the DSC measurements, where the transformation from the ε phase to the γ phase is delayed by cold rolling. The contrast pattern of the IQ map does not change much even after the ε→γ transformation, suggesting that the γ phase may inherit some of the deformation state of ε. The material heat treated at 400°C shows a different morphology from that of the hot-rolled structure, as shown in Fig. 1. It corresponds to the fact that the DSC cooling curve with cold-rolled materials has a broader transformation peak (indicated ‘*’ mark in Fig. 2(b)) than that of the hot-rolled material.
Figure 9 shows the relationship between the crystal orientations of the γ precipitated grains and adjacent ε grains at 250°C, at which the FCC phase begins to appear. The surrounding ε (0001) plane coincides with one of the four poles of γ (111). The ε phase around the γ grain is inferred to be the variant originally transformed from the same γ grain. Although there is no clear peak of the γ phase in the cold-rolled material in the X-ray diffraction measurement, it is thought that the γ grain, remains partially. And in EBSD measurements, due to the scanning time under the in situ observation, the selected step size of approximately 0.3 μm was too large to observe the fine γ. In the heating stage, this grain is easily transformed (grown) at an early stage because it follows the Shouji–Nishiyama (SN) relation22,23) ((111)γ//(0001)ε, [
Orientation relationship between the evolving γ phase and surrounding ε phase at 250°C: (a) phase map and (b) IPF map of the HCP structure. The (111) poles of mall γ grain A are coincident with the (0001) poles of each surrounding ε phase.
The phase transformation was influenced by the atomic-scale dislocations caused by the deformation process. The substructures of the hot-rolled material and 10% and 35% cold-rolled materials were studied by STEM. Figure 10 shows the observation results of the hot-rolled material. Figure 10(a) shows the bright field (BF) image and (c) shows the annular dark field (ADF) image within the frame of (a). The microstructure of the hot-rolled material shows the band feature by the interweaving ε martensite variants. Dislocations are found inside the ε grains, and their distribution is found not uniform.
STEM micrographs of the hot-rolled specimen: (a) bright field image, (b) selected area diffraction pattern taken from the spotted area in (a), and (c) magnified annular dark field image corresponding to the square region in (a). A high density of dislocations is heterogeneously distributed in the ε martensite variants in the hot-rolled specimen. (Online version in color.)
The observed micrographs of the 10% cold-rolled material are shown in Fig. 11. The dislocation density is increased remarkably; there are some parts where the dislocations accumulate locally, particularly on the interfaces of the ε variants, and large defects are also formed to traverse the ε phase as indicated arrows in Fig. 11(a). The observed images of the 35% cold-rolled material are shown in Fig. 12. Although a region similar to the band-shaped structure of the hot-rolled material is observed near the center of the image, there is a heavily deformed region around the unreformed region. This tendency is the same as the results of the SEM observation. Figure 12(c) shows the BF image of another region showing the interweaving band-like microstructure. The diffraction patterns shown in Figs. 12(d) and 12(e) correspond to regions A and B in Fig. 12(c), respectively. The γ phase remains inside the ε grains. A slight distortion of the hot-rolled structure occurs as indicated arrows in Fig. 12(b), despite of a considerable accumulation of dislocations. This region can have been the origin (nucleus) of the γ phase grown in the in situ heating measurement by the EBSD.
STEM micrographs of the 10% cold-rolled specimen: (a) bright field image, (b) selected area diffraction pattern taken from the spotted area in (a), and (c) magnified annular dark field image corresponding to the square region in (a). Some variant interfaces are curved as a result of the cold rolling, whereas the dislocation structure is nearly the same as that in the as-hot rolled sample in Fig. 10. (Online version in color.)
STEM micrographs of the 35% cold-rolled specimen: (a) bright field and (b) annular dark field images, (c) bright field image of another region, and (d) and (e) selected area diffraction patterns taken from the areas labeled A and B in (c), respectively. The γ phase remains in some limited regions, as shown in (c) and (d), and a heterogeneous distribution of dislocations similar to that in the hot-rolled specimen is also observed.
In the deformation of HCP materials, the critical resolved shear stress (CRSS) of the basal slip is considered to be two orders of magnitude smaller than that of other slip systems at room temperature; only the basal slip is considered to be active.24,25,26,27) Recently, it has been reported that the workability was remarkably improved in fine-grained Mg alloys subjected to heavily strained deformations, and 40% of the total dislocations were generated by the non-basal slip systems.28,29) Yoo et al. have published a review of this phenomenon,30) and the <a + c> dislocation structure played an important role in the stability of the HCP lattice.31) A similar model has been reproduced by molecular dynamics (MD) calculation.20) According to Tomoda et al.,9) a thermally generated ε martensite plate was considered to be a major obstacle to the dislocation movement (against plastic deformation). Both the deformed twins of the FCC metal and γ→ε transformation were caused by the movement of the Shockley partial dislocations. The difference between the two was in the movement of the partial dislocations in each layer of the {111} plane toward or further away in every other layer to form the HCP structure. Compared with the twinning mechanism, ε had a smaller softening effect (shear strain was smaller) and the hardening effect was larger, due to a difference in the slip systems of the FCC and HCP phases.
It was expected from the X-ray diffraction measurement that the basal slip aligned to the (0002) plane to the plate surface direction was the main slip system in the cold-rolling process. However, there was a possibility that a non-basal slip had occurred microscopically due to some large local deformation. In such heavily deformed areas, the progress of the transformation to FCC was expected to be significantly changed.
3.5. Effect of Cold Rolling on Stability of ε and γ Phases in Heating StageThe transformation behavior during the heating process was investigated in the Fe–20%Mn alloy. The major phase of the initial hot-rolled material was ε martensite, having an HCP structure. As the cold-rolling reduction ratio increased, the dislocation density inside the ε phase also increased (Figs. 10, 11, 12); thus, a uniform deformation became difficult. Consequently, a heavily deformed structure could be generated locally by the non-basal slip, resulting in a complex and heterogeneous microstructure. There was also a region inherited by the microstructure that was similar to the band-shaped structure of the hot-rolled material (Figs. 1, 12). The TEM observation confirmed that the γ grains were retained before heating (Fig. 12), although those grains were difficult to distinguish by X-ray diffraction. The γ grains were not completely transformed to ε martensite in the cooling stage after hot rolling, and the γ phase remained in a condition surrounded by the ε variants.
When Watanabe et al.13) investigated the damping performance by changing the training conditions with a solution treatment using Fe–20 mass% Mn, the volume fractions of ε were 60 to 70% under the training treatment condition only with a thermal history and the fraction of ε varied in the training process combined with the mechanical working, whilst the initial hot-rolled material in the present work was almost the ε single phase. Nakatsu et al.1) pointed out that the amount of ε formation was suppressed, and the ε transformation onset temperature (Ms γ→ε point) decreased with the refinement of the γ grain size. These phenomena were particularly remarkable in the region where the γ grain size was 30 μm or less. When the γ grain size became finer, the existence of the grain boundaries was considered to suppress the chain reaction of ε formation, apparently stabilizing the γ phase. The difference in the resultant phase fraction in comparison with the results reported by Watanabe et al. was not clear but considered to be due to the effect of working, such as hot rolling, or the cooling condition.
When a cold-worked material is heated, the phase transformation is observed macroscopically at temperatures exceeding 200°C via DSC and X-ray diffraction (Figs. 2, 6). However, according to the results of the X-ray diffraction measurement (Fig. 7) and in situ observations by EBSD (Fig. 8), the γ transformation begins to progress locally even at relatively lower temperatures. In the previous experiment by Tasaki et al.14) with the Fe–28 Mn–6 Si–5 Cr (mass%) alloy, in which a slight deformation was applied by a 5% or 10% tensile test, the integral intensity of the ε phase in the X-ray diffraction profiles decreased slightly from 300 K under the 10% deformation, which is a tendency similar to the present work. The reason why the strength of the X-ray diffraction increases once at 300°C with the cold-rolled material is the texture change, as seen in the in situ observation by EBSD.
There is a high possibility that the γ phase remaining in the region with a relatively low degree of work becomes the origin or nuclei of the γ growth during heating. This is because the growth is easy owing to the small degree of working of this region and its good crystal coherency with the surrounding ε phases. However, the growth seems to have been limited to only some extent owing to the disorder associated with the presence of dislocations. The growth of the γ phase becomes prominent at around 250°C or higher, but as in the DSC measurement, the transformation completion temperature increases as the cold-rolling reduction increases. The deformed microstructure, such as the shear bands introduced in the hot-rolled structure by cold rolling, can be an obstacle in the growth of the γ phase. In addition, there is no clear change in the IQ map after the transformation in the EBSD measurement; the transformation peak in the cooling stage of the DSC measurement becomes broad with the cold-rolling reduction, and the hardness of the material heat-treated at 400°C is higher than that of the hot-rolled sheet. These tendencies suggest that the dislocations or lattice curvature remain even after the γ transformation, affected by the deformed substructures before the transformation.
There is a slight difference in the temperature necessary to quench the ε phase in the DSC measurement and X-ray diffraction measurement. The possible reasons are: (1) There was a possibility of overestimated temperature because the measurements were performed while varying the temperature; (2) the transformation point was evaluated from the intersection of the linear parts, and there was a possibility that the transformation onset temperature was over-evaluated, unlike the temperature where the shift from linearity started.
Thermal analysis, high-temperature X-ray diffraction analysis, and high-temperature EBSD measurement were conducted to study the effect of cold rolling on the stability of the HCP and FCC phases of the Fe–20Mn alloy. The stability of the ε and γ phases changed intricately, depending on the cold-rolling reduction and heating temperature. The following knowledge was obtained:
(1) When the hot-rolled material showing ε martensite as the major phase was cold rolled, uniform deformation became difficult because the degree of the cold-rolling reduction ratio increased. Because the work hardening rate was high, the basal slip in the HCP structure was the main slip system, but there was a possibility that the non-basal slip might also have occurred locally. Although the dislocation density increased after cold rolling, there were regions that inherited the band-like structure similar to that in the hot-rolled material. Retained γ grains were also found in the ε-deformed structure, but not detected by X-ray diffraction.
(2) During heating, a reverse γ transformation was remarkably observed macroscopically above 200°C, and the transformation finish temperature rose with increasing cold rolling reduction ratio. However, it was thought that the γ transformation progressed microscopically at a relatively lower temperature. It was considered that the γ phase was retained locally in the region with a relatively small degree of work before heating and served as the initiating sites of the growth.
(3) No clear change was seen in the IQ maps before and after the transformation by the EBSD measurements; the transformation temperature during cooling clearly broadened. Furthermore, even when the specimens were heat-treated at 400°C, their hardness did not return to that of the hot-rolled material. These facts suggested that the dislocations or lattice curvature remained under the influence of the deformed structure before the transformation, even if it underwent transformation to the γ phase during heating.
