Regular Article
Morphology Dependence on Mechanical Stability of Second-phase Austenite in Martensitic Steels
2024 Volume 64 Issue 2 Pages 421-429
Details
2024 Volume 64 Issue 2 Pages 421-429
To investigate the stand-alone effect of austenite morphology on transformation-induced plasticity (TRIP), martensitic steels with second-phase austenite, whose morphology was controlled by the dissolution of alloy cementite, were fabricated with 0.6%C–3%Mn steel and their mechanical properties were evaluated in terms of the mechanical stability of austenite. The morphology of alloy cementite in the initial microstructure was controlled to lamellar and spherical through Mn partitioning pearlitic transformation and subsequent spheroidization. Alloy cementite was replaced with retained austenite after austenitization due to local austenite stabilization by Mn enrichment. As a result, two types of martensitic steels with lamellar- and spherical-shaped austenite grains could be prepared, where composition, fraction, and number density of the second-phase austenite were kept the same, while only the morphology was changed significantly. Compression tests revealed that the steel with lamellar austenite maintains higher strain hardening rate than that with spherical austenite. The higher strain hardenability was attributed to the TRIP effect, indicating that the lamellar austenite has lower mechanical stability. Furthermore, nanoindentation tests was conducted to directory evaluate the elastic-plastic deformation behavior of second-phase austenite. The deformation behavior of the second-phase austenite showed a clear morphology dependence; the lamellar austenite shows larger strain hardening from the early stage of plastic deformation. Detailed analysis suggested a possibility that the interphase boundary between the martensite matrix and second-phase austenite acts as a preferential nucleation site for deformation-induced martensitic transformation.
In recent years, a demand for high-strength steels has been increasing in order to build a sustainable society. High-strength steels with fine austenite grains as second-phase are being developed as one of advanced high-strength steels, because they have excellent strength-ductility balance.1,2,3,4) The high strength of the steel is ensured by martensite or bainite matrix, while second-phase austenite undergoes deformation-induced martensitic transformation during plastic deformation, resulting in the improvement of ductility due to transformation-induced plasticity (TRIP) effect. Therefore, it is essential to take full advantage of the TRIP effect to optimize the strength-ductility balance of this type of steel, and it is important to understand and control the stability of the second-phase austenite in response to deformation-induced martensitic transformation, i.e. mechanical stability. The mechanical stability of austenite has been the subject of many studies. In addition to the chemical composition,5,6,7) crystallographic orientation,8) and size9,10,11) of austenite, various factors, such as the type of matrix phase12) and the strength difference between matrix and austenite,13) have been reported to affect austenite mechanical stability. In the austempering, which is a general heat treatment to obtain retained austenite in low-alloy carbon steel, carbon (C) is enriched in the untransformed austenite through bainitic transformation, low-alloy TRIP assisted steel. The untransformed austenite formed during the bainitic transformation is left with film-shape between individual bainitic ferrite, while blocky austenite grains remain near the prior austenite grain boundaries. In previous studies, it has been reported that the blocky austenite is more unstable than film one. This is most likely due to differences in C enrichment behavior during austempering, and thus the effect of morphology itself on the mechanical stability of austenite has not been evaluated. However, in order to change the austenite morphology during austempering, the temperature and time of bainitic transformation must be changed, which inevitably changes the properties of the bainite matrix and C-enrichment behavior.14,15) That is, it is very difficult to change only the austenite morphology independently in low-alloy TRIP assisted steels.
On the other hand, the retainment phenomenon of austenite caused by dissolution of alloy carbides has attracted attention.16,17,18,19,20,21) When the initial microstructure with alloy carbides containing high concentrations of Mn is heated to austenite single-phase region, the carbides are completely dissolved in austenite matrix. However, Mn does not diffuse fast in austenite, and Mn is locally held in the region where the carbides have dissolved. These localized Mn-enriched regions are called Mn ghost and remains as second-phase austenite even after quenching because the Mn-enrichment lowers the martensitic transformation start temperature Ms. As a result, a dual-phase microstructure with austenite dispersed in the martensite matrix can be obtained. This technique could be applied to arbitrarily control only the austenite morphology and to discuss the morphology dependence on the mechanical stability of second-phase austenite.
In this study, we tried to control the morphology of second-phase austenite dispersed in martensite matrix by using two types of initial microstructures with lamellar and spherical alloy cementite particles in 0.6%C–3%Mn steel (mass%). And compression tests were then performed on both steels to investigate the dependence of strain hardening behavior and the TRIP effect on austenite morphology. In addition, local mechanical tests using nanoindentation tests were also conducted to examine the effect of morphology on the mechanical stability of austenite.
0.6%C–3%Mn steel (mass%) with the chemical composition listed in Table 1 was used in this study. Specimens cut from a sufficiently homogenized hot-rolled plate were subjected to two types of heat treatment shown in Fig. 1 to obtain alloy cementite with different morphology. For lamellar pearlite material (LPM), after solution treatment at 1293 K for 1.8 ks, it was isothermally held at 923 K in a salt bath furnace for various times to complete pearlitic transformation with Mn partitioning between ferrite and cementite. And, for spheroidized pearlite material (SPM), lamellar pearlite isothermally transformed at 923 K for 180 ks was cold rolled at 67% in thickness reduction and annealed again at 923 K for 360 ks to spheroidize the cementite. Cylindrical specimens with a diameter of 3 mm × 10 mm were cut from LPM and SPM and heated to 1173 K at heating rate of 50 K/s for austenitization, followed by gas quenching using a heat treatment furnace (Formastor-FII developed by Fuji Electronic Industrial Co., Ltd). They were named as lamellar and spherical austenite materials (LAM and SAM), respectively. The austenite reversion start and finish temperatures, Ae1 and Ae3, estimated from bulk composition by Thermo-Calc. software (TCFE10), were 971 K and 983 K, respectively. Furthermore, as described below, the temperature at which the alloy cementite is completely dissolved by C diffusion under local equilibrium condition, i.e. Partitioning to Non-partitioning Transition Temperature: PNTT,20,21) is estimated to be 1127 K. The microstructures were observed using an optical microscope and a field emission scanning electron microscope (FE-SEM, JSM-7001F developed by JEOL Ltd.) at an acceleration voltage of 15.0 kV. Mn distribution was analyzed using an energy dispersive X-ray spectrometer (JED-2300 Analysis Station Plus, JEOL Ltd.) attached to the FE-SEM. Crystal orientation analysis was performed by Electron Back Scattering Diffraction (EBSD), which was then analyzed by OIM Data Collection ver. 7.3.1 developed by TSL Solutions, Inc. The working distance was 15.0 mm, and the step size was 50 nm. Uniaxial compression tests were performed using an autograph universal testing machine (AG-100kNX, Shimadzu Corporation). Cylindrical specimens with a diameter of 3 mm × 7 mm were used for compression tests at room temperature under an initial strain rate of 1.0 × 10−3 s−1. The true stress-strain curves were obtained under constant volume condition. Local mechanical tests were performed using a nanoindentation system (Hysitron TI Premier developed by BRUKER) with a Berkovich indenter. Load-displacement curves (P-h curves) were obtained with controlled loading and unloading rates of 200 μN/s, maximum load of 1000 μN, and loading time of 2 s. The composite elastic modulus E and nano-hardness Hn were evaluated from the obtained P-h curve.
| C | Mn | Ni | P | S | Fe | |
|---|---|---|---|---|---|---|
| mass% | 0.59 | 2.92 | <0.003 | <0.002 | 0.0010 | Bal. |
Figure 2 shows optical images of transformation microstructures of LPM isothermally held at 923 K for (a) 36 ks and (b) 180 ks. After 36 ks holding (a), although pearlite (P) formed partially, the large area was occupied by martensite (M) with high hardness, which was transformed from untransformed austenite. On the other hand, the pearlitic transformation was completed after 180 ks (b). Considering that the pearlitic transformation in carbon steel takes several tens of seconds to complete, this extremely slow pearlitic transformation suggests that the pearlitic transformation was controlled by Mn diffusion, which is consistent with the report by Ishigami et al.22) Figure 3 shows the SEM images of (a) LPM isothermally held for 180 ks and (b) SPM prepared by cold rolling and subsequent annealing at 923 K for 360 ks. LPM (a) showed a homogeneous lamellar structure with interlamellar spacing of 0.33 μm, while spherical cementite particles were uniformly dispersed in SPM (b). The area fractions of the cementite were 8.6% and 8.5%, which are almost the same as the calculated equilibrium cementite fraction of 8.9% at 923 K. The Mn concentration of cementite was analyzed at 19.2%Mn for LPM and 17.1%Mn for SPM, which also corresponds to the equilibrium value of 18.9%Mn calculated by Thermo-Calc with TCFE10 database. The above observations allowed us to prepare an initial microstructure in which only the alloy cementite morphology differs significantly.
Figures 4 and 5 show the microstructural evolution of LPM and SPM through austenitization at 1173 K, respectively. Each figure represents (a, b) phase maps, (c, d) IPF maps of the matrix phase (bcc), and IPF maps of the second phase (e) cementite or (f) fcc-austenite. In the phase maps, the bcc (α), fcc (γ), and cementite (θ) phases are colored blue, yellow, and red, respectively, and high angle grain boundaries with misorientaiton higher than 15° are indicated by solid white lines. In LPM before austenitization (Figs. 5(a), 5(c), 5(e)), cementite lamellae tended to be arranged on the same direction in each pearlite block with identical ferrite crystal orientation. The crystal orientation of the cementite is hardly indexed by EBSD because of its low crystallinity and the thinness of the lamellae, but it is expected that the orientation is identical at least for each block or colony.23) After austenitization (Figs. 5(b), 5(d), 5(f)), the matrix phase changed to martensite with a fine sub-structure and high hardness, and austenite was distributed as second phase. The austenite had the same lamellar morphology and distribution as the cementite before austenitization, indicating that the cementite in the initial microstructure was replaced with austenite through austenitization. When we look at the crystallographic orientation of the second-phase austenite, it is found that it changed discontinuously with a certain region as a unit. This suggests that austenite structure with these regions had existed as grains in the austenite single-phase state at high temperatures. Similarly, austenite originating from cementite were observed in SPM after austenitization (Fig. 6). In other words, after austenitization, SAM had a duplex microstructure with spheroidized second-phase austenite dispersed in the martensite matrix. The area fractions of second-phase austenite in LAM and SAM were 18.6% and 18.2%, respectively. Although the differences between the materials were small, both tended to increase from that of cementite in the initial microstructure, LPM and SPM. This is thought to be due to the diffusion of Mn ghosts in the austenite single-phase state, which will be discussed later. EDS analysis revealed that the lamellar and spherical austenite grains had almost the same Mn concentration at 16.2%Mn and 16.5%Mn, respectively. In addition, the Ms of the matrix was measured approximately 473 K, and there was no difference between both materials. These results strongly suggest that LAM and SAM were composed of a martensite matrix and second-phase austenite with almost the same chemical composition. In addition, there is also little difference in the number density of austenite grains; 0.402/μm2 for LAM and 0.597/μm2 for SAM. From these results, it can be said that two martensitic steels with significantly different morphology of austenite grains had been successfully fabricated.
In order to predict the formation of Mn enriched austenite caused by the dissolution of alloy cementite, a one-dimensional analysis using DICTRA was performed. The variation of phase fraction of ferrite, cementite, and austenite was simulated on heating from 923 K at constant heating rate of 50 K/s (Fig. 6). The total cell length was set to 0.75 μm, which corresponds to the average interval of dispersed cementite particles in SPM. It was also confirmed that the following simulation results remain almost unchanged, when the cell length is slightly changed to the size corresponding to LPM. And the equilibrium composition and fraction at 923 K were entered for ferrite and cementite as initial conditions, while austenite was treated as inactive phase which nucleates at the ferrite/cementite interphase. Austenite was formed at approximately 1000 K and then grew with increasing temperature, but the growth rate was not constant. After the nucleation, austenite formation proceeded slowly, and above 1035 K, the growth rate increased rapidly with the rapid shrinking of ferrite. This indicates that the rate controlling of the α/γ interface migration transited from Mn to C diffusion over 1035 K. Thereafter, ferrite disappeared completely at 1065 K, and the austenite transformation rate stagnated again. Yang et al.20) defined the lowest temperature at which alloy carbides completely dissolve by the C diffusion as PNTT-II (Partition to No-Partition Transition temperature) and reported a method to calculate this temperature thermodynamically. According to this method, the PNTT-II predicted from the initial microstructure of the alloy is about 1127 K. The DICTRA calculation result (1150 K) was slightly higher than the PNTT-II. This is considered to be due to an overheat phenomenon during the continuous heating process. However, since the experimental austenitization temperature was 1173 K, there would be no doubt that the dissolution of alloy cementite had completed and a single phase of austenite was obtained. Figure 7 shows the concentration profiles of C and Mn in the austenite single-phase state after reaching 1173 K. The concentration profiles in the initial state are also displayed as a dashed line for comparison. Mn was kept at high concentration in the regions where cementite existed even after austenitization, indicating the formation of Mn ghost. However, comparing to the Mn profile in the cementite in the initial microstructure, the Mn concentration in the Mn ghost decreased at the edges, which suggests that Mn diffused outward from the Mn ghost. Here, we consider the austenite stability of such Mn ghost. The effect of alloying elements on the Ms of carbon steel has been formulated by various researchers, and Andrews’ formula,24) which is expressed as follows, is well known for high-carbon steels.
| (1) |
If C = 0.6 mass% and the effects of other alloying elements are ignored, the Mn concentration at which the Ms reaches room temperature (293 K) is roughly estimated to be 8.7%Mn. We notice that the position where the Mn concentration reaches 8.7 mass% after austenitization was slightly larger than the initial α/θ interface. This indicates that the Mn ghosts formed by the dissolution of alloy cementite remain as austenite even after quenching, but their size tends to be coarser than that of the initial alloy cementite. This agrees with the experimental fact that the area fraction of austenite in LAM and SAM was higher than that of the initial cementite, LPM and SPM, as shown in Figs. 5 and 6.
3.2. Effect of Morphology on Mechanical Stability of Second-phase Austenite 3.2.1. Morphology Dependence of Second-phase Austenite on Strain-hardeningFigure 8 shows the results of compression tests on LAM and SAM with second-phase austenite. The true stress-strain curve (σt–εt) and the strain hardening rate-true strain curve (dσt/dεt–εt) are shown as solid and dashed lines, respectively. Both materials exhibited a very high yield stress of approximately 2000 MPa due to high-carbon martensite matrix, followed by stable strain hardening. The dσt/dεt decreased monotonically with plastic deformation and remained higher in LAM than in SAM over the entire strain range. Figure 9 shows the change in the second-phase austenite fraction fγ with compression true strain εt, according to the following equation proposed by Sugimoto et al.25,26)
| (2) |
Where fγ0 is the initial austenite fraction before deformation. On the other hand, k (≥ 0) is an index of austenite mechanical stability, and a larger value means that the austenite is unstable and prone to deformation-induced martensitic transformation. The austenite fraction was evaluated averagely after EBSD analysis was performed several times in the longitudinal and radial directions on compressed cylindrical test specimens. In both materials, the austenite fraction decreased with increasing compression strain, indicating deformation-induced transformation of the second-phase austenite. However, there was a large difference in the transformation behavior between the materials. Although the EBSD analysis has errors in the measurement data due to the relatively small observation area, LAM showed a larger k than SAM, clearly indicating that austenite in LAM was less mechanically stable. In LAM, more than half of the second-phase austenite underwent deformation-induced martensitic transformation after compression deformation at approximately εt = 0.10, resulting in greater strain hardening than in SAM due to the TRIP effect. In other words, the mechanical stability of second-phase austenite is also highly dependent on its morphology, and it is clear that lamellar austenite is more susceptible to deformation-induced martensitic transformation than spherical austenite in martensitic steels.
3.2.2. Phase Identification of Deformation-induced MartensiteGenerally, in the martensitic transformation of steel, fcc (face-centered cubic) austenite transforms to bcc (body-centered cubic) or bct (body-centered tetragonal) α’-martensite. However, it is known that hcp (hexagonal close-packed) ε-martensite prefer to form in high Mn austenitic steels such as Hadfield steels and TWIP steels, due to the reduction of stacking fault energy by the addition of Mn. In this study, steels with large amounts of C and Mn were intentionally added, and in particular, the composition of the central part of the Mn ghost was expected to be around 19%Mn-1.1%C (see Fig. 7), and the formation of ε-martensite is a concern. In order to clarify the martensite phase transformed from the second-phase austenite, EBSD analysis was performed on (a) LAM and (b) SAM after compression at εt = 0.10 (Fig. 10). In this figure, reliable measurement points with a Confidence Index (CI) value higher than 0.1 are displayed. As shown in Fig. 8, deformation-induced transformation was sufficiently advanced by the compression, and the austenite fraction of the second phase was sufficiently small compared to the initial value (fγ0 = 0.18) in this observation area, so it was judged that deformation-induced martensite was formed. The second-phase austenite was still uniformly distributed in both LAM and SAM, and the same trend was confirmed in several observation fields. This suggests that deformation-induced martensitic transformation occurred relatively uniformly in the material, although the deformation-induced martensite could not be distinguished from the pre-existing martensite matrix. As for the constituent phases, the complete absence of ε-martensite in both materials indicates that deformation-induced fcc-hcp martensitic transformation did not occur. In other words, the second-phase austenite in the materials transformed to α’-martensite. It is not clear why highly Mn-enriched second-phase austenite does not transform to ε-martensite, but the simultaneous enrichment of C with Mn increases the stacking fault energy of the austenite, resulting in a suppression of fcc-hcp martensitic transformation.
Compression tests showed that second-phase lamellar austenite is susceptible to deformation-induced martensitic transformation, which effectively increases the strain hardening rate through the TRIP effect in martensitic steels. However, when a metallic material composed of multiple phases is plastically deformed, the stress/strain partitioning between the phases must be taken into account. Park et al.27) investigated the effect of second-phase martensite distribution on the tensile deformation behavior in Dual Phase steel consisting of soft ferrite and hard α’-martensite. As a result of detailed strain distribution analysis using digital image correlation technique, they reported that the strain distribution behavior between ferrite and martensite changes depending on the distribution of martensite, and has a significant effect on strain hardening behavior. In addition, Koyama28) conducted a simulation of stress-strain curves for a two-phase microstructure using Secant method. As a result, he reported that the lamellar arrangement of the hard second phase carries high stress and increases the strain hardening of the entire material. Although there is no experimental evidence of strain or stress concentration in the lamellar austenite in LAM used in this study, it is undeniable that this might contribute to the deformation-induced transformation consequently. In other words, to clarify the intrinsic morphology dependence of austenite mechanical stability, it is necessary to deform only the second-phase austenite. In the following sections, we attempted to directly evaluate the mechanical stability of austenite using nanoindentation tests as a simplified test that is not affected by the martensite matrix as much as possible.
Figure 11 shows EBSD-Phase maps of LAM and SAM before and after the nanoindentation test in the same field of view. The points with CI value lower than 0.1 are shown in black. The nanoindentation was repeated at several locations, and the measurement positions were mechanically determined on a square grid. Comparing the phase maps before and after the test, it can be seen that the equilateral triangular indentations shown in black were regularly distributed. The indentations indicated by the black arrows are located in the center of individual austenite grains, and it is highly probable that the local mechanical properties of the second-phase austenite could be evaluated. Figure 12 shows P-h curve in nanoindentation tests on austenite grains in LAM and SAM. In addition to the results for the second-phase austenite, results for the martensite matrix are also shown for comparison. Here, when the object is elastically deformed in the nanoindentation test, the P-h curve can be expressed by the following equation using the Hertz contact equation.29)
| (3) |
R is the radius of curvature of the indenter tip, and E is the combined modulus of the sample and indenter, which can be estimated from the unloading behavior in the P-h curve. The P-h curve of austenite predicted by substituting R = 40 nm and E = 217 GPa into Eq. (3) (Hertz’s curve) is supplemented with a black dashed line. First, focusing on the P-h curve in the martensite matrix, there was no differences between the materials in the entire region. This means that the martensite matrix of them exhibited the same mechanical properties. On the other hand, the deformation behavior of the second-phase austenite was obviously different. In particular, as indicated by the black arrows, the P-h curves of the lamellar and spherical austenite grains differed significantly in the load range higher than the stress Pe, which deviates from the black dashed line predicted by the Hertz contact equation, with the lamellar austenite showing a smaller h at a given P. This indicates that the plastic deformation behavior of austenite was quite different between two materials, and that the lamellar austenite has a relatively greater strain-hardenability. Further analysis was conducted to understand the plastic deformation behavior of the lamellar and spherical austenite grains in more detail. It has been reported that the relationship between P and h in nanoindentation tests can be approximated by the following quadratic equation.30)
| (4) |
Where the coefficient a is a coefficient that depends on the mechanical properties of the material and b is another one corresponding to the indenter tip shape and the stiffness of the load frame. Dividing both sides of Eq. (4) by h gives
| (5) |
The result of arranging Fig. 12 into a P/h-h curve according to Eq. (5) is shown in Fig. 13. In this case, the slope of the P/h-h curve corresponds to a in Eqs. (4) and (5). In both materials, a changed discontinuously and significantly before and after he corresponding to Pe, confirming once again that the region below Pe is the elastic deformation state. Next, focusing on the plastic deformation region above he, the P/h-h curve was slightly convex downward in both materials, indicating that a is not constant but increased continuously. In particular, as indicated by the arrows, the lamella austenite deformed with a larger increase in a for h > 17 nm (hc), indicating that the deformation resistance of the second-phase austenite in LAM increased continuously from the relatively early stage of plastic deformation. Now considering the geometry of the Berkovich indenter34) and the elastic deformation of the sample, the length of a piece of equilateral triangular indentation formed on the sample surface was estimated to be 88.1 nm at hc = 17 nm. This size is well below the average width of 500 nm for lamellar austenite and the average grain size of 1000 nm for spherical austenite. In general, the plastic deformation area formed around the indenter is reported to be about three times larger than the indent size. That is, even considering the plastic zone around the indenter, the mechanical response at h < hc was less directly affected by the martensite matrix and the γ/α’ interface, which reflects the properties of the second phase austenite itself. On the other hand, at hc < h, the influence of the γ/α’ interface was expected to become apparent in lamellar austenite. The γ/α’ interface provides strain hardening of the second-phase austenite by pilling up dislocations as well as elastic restriction. In addition, such stress/strain concentration will induce martensitic transformation, resulting in further strain hardening. Considering that the TRIP effect was induced by the deformation-induced transformation of the second-phase austenite in Figs. 9 and 10, it can be inferred that the γ/α’ interface was a preferential nucleation site for the deformation-induced martensitic transformation, although the possibility that the interface itself is a deformation resistance cannot be dismissed. Ohmura et al.31) performed nanoindentation tests on polycrystalline ferrite and found that pop-in, which corresponds to the onset of plastic deformation, is low near grain boundaries, indicating that grain boundaries are sites for dislocation nucleation. The γ/α’ interface has a similar effect, and if dislocations promote martensite nucleation, they may act as a preferential nucleation site for strain-induced martensitic transformation. When the second-phase austenite is dispersed under constant volume fraction and number density conditions, the austenite shape becomes spherical, minimizing the total area of the γ/α’ interface. Considering the above, as shown schematically in Fig. 14, it is thought that the mechanical stability of lamellar austenite with high interface density is lower due to the higher frequency of deformation-induced martensite nucleation. The EBSD analysis of Figs. 10 and 11 did not confirm the fact that martensite formation at the γ/α’ interface. However, as shown in the simulation results using DICTRA (Fig. 8), the Mn concentration relatively decreases near the γ/α’ interface due to the diffusion of Mn ghosts during austenitization. Therefore, from a thermodynamic point of view, it is consistent that the γ/α’ interface was a preferential nucleation site for deformation-induced martensite.
As described above, the mechanical stability of second-phase austenite has morphology dependence, suggesting that the γ/α’ interface is the preferential nucleation site for the deformation-induced martensite. It should be noted that the decrease in Mn concentration at the γ/α’ interface is a unique phenomenon in this study. That is, since C is enriched in untransformed austenite in low-alloy TRIP assisted steels during austempering, the chemical driving force of martensitic transformation may be reduced at the γ/α’ interface with high C concentration. Therefore, further investigation is needed to determine whether the mechanical stability of the second-phase austenite in low-alloy TRIP steels is morphologically dependent.
To investigate the stand-alone effect of austenite morphology on transformation-induced plasticity (TRIP), the morphology of second-phase austenite dispersed in martensite matrix was controlled by using two types of initial microstructures with lamellar and spherical alloy cementite particles in 0.6mass%C–3mass%Mn steel. And compression tests were then performed on both steels to compare strain hardening behavior caused by TRIP effect. In addition, the morphology dependence of austenite mechanical stability was locally investigated using nanoindentation tests. The findings are summarized as follows:
(1) Alloy cementite was replaced with retained austenite after austenitization due to local austenite stabilization by Mn enrichment. Therefore, the morphology of the second-phase austenite can be controlled to lamellar and spherical by Mn-partitioning pearlitic transformation and subsequent spheroidization.
(2) The second-phase austenite remains with the alloy cementite morphology in the initial microstructure, but the second-phase austenite is slightly coarser than the alloy cementite due to diffusion of Mn in the Mn-enriched regions.
(3) The second-phase austenite transforms deformation-induced bcc-martensite by plastic deformation. In compression tests, lamellar austenite undergoes deformation-induced transformation more readily than spherical austenite, resulting in greater strain hardening due to TRIP effect.
(4) The elastic-plastic deformation behavior analysis by nanoindentation tests revealed that lamellar austenite has higher strain-hardenability than spherical austenite from early stage of plastic deformation. This suggests a possibility that the interface between martensite matrix and second-phase austenite acts as preferential nucleation site for deformation-induced martensite.
