Deformation-induced Martensite Transformation Behavior during Tensile and Compressive Deformation in Low-alloy TRIP Steel Sheets

Abstract

Transformation-induced plasticity (TRIP) is a phenomenon that improves the deformability of high-strength steel. TRIP depends on deformation-induced martensite transformation behavior. To clarify the mechanism of the transformation in low-alloy TRIP steel, we evaluated the transformation behavior via in-plane tension and compression experiments. During tensile and compressive deformation, the volume fraction of austenite (V_γ) decreased as strain and stress increased. The rate at which V_γ decreased during compressive deformation was slower than that during tensile deformation. However, after continuous deformation (i.e., tensile deformation under compression and vice versa), V_γ depended on stress, not strain. The transformation behavior was controlled by the applied stress, regardless of strain path and stored strain. It is appropriate to apply a stress-dominant strain-induced transformation model to explain this macroscopic transformation behavior.

1. Introduction

High-strength steels that contain multiple phase structures are widely used. However, the applications for high-strength materials are limited owing to their poor formability. The transformation induced plasticity (TRIP)¹⁾ is a phenomenon that counteracts this problem in steel.²⁾ Metastable austenite (γ) in steel transforms to hard martensite with deformation, and this transformation increases the macroscopic work-hardening rate of steel, which dominates its uniform elongation (UEL).³⁾ Zackey et al.⁴⁾ produced high-performance steel containing martensite with metastable γ, which they called “TRIP steel.” Although the balance between strength and elongation in TRIP steel is excellent (e.g., 1.3 GPa maximum tensile strength with 75% total elongation), using it is rarely economically feasible because it contains expensive alloying elements (Ni, Mo, and Cr) and the manufacturing process is difficult. Sakuma et al.^5,6) reported the fabrication of CSiMn simple composition steel sheets that contained a small volume fraction of metastable retained γ. This steel is called low-alloy TRIP steel,⁷⁾ and it has been primarily used to produce automotive steel sheets^8,9) because its performance and manufacturing cost are balanced. Recently, so-called “3rd generation steels”¹⁰⁾ are attracting attention as the next advanced high-strength steel. Representative 3rd generation steels, including quench and partitioning steel,^11,12) medium manganese steel,^13,14) carbide-free bainitic steel,^15,16) and TRIP-aided annealed martensitic steel,¹⁷⁾ include metastable γ and utilize the TRIP effect to improve performance.¹⁸⁾ Therefore, it is important to understand the transformation behavior of metastable γ during deformation in such steels.

Many studies on the effects of deformation on transformation have been conducted.²⁾ In 1932, Scheil¹⁹⁾ found that the amount of deformation-induced martensite generated at temperatures higher than that of the initiation of martensite transformation, M_s, decreases with increasing deformation temperature in high-Ni ferrous alloy. Therefore, no martensite forms above a specific deformation temperature, M_d. Scheil suggested that a critical resolved shear stress is required to initiate the transformation from γ to martensite. Olson and Cohen²⁰⁾ classified the deformation-induced martensitic nucleation from γ into stress-assisted and strain-induced nucleation. A schematic diagram of their model is shown in Fig. 1(a) for FeNiC alloys studied by Bolling and Richman.^21,22,23,24) This diagram shows the critical stress required to initiate a martensite transformation as a function of temperature. At temperatures slightly higher than M_s, nucleation occurs with applied stress that is less than the elastic limit. As the temperature increases, the critical stress increases until it equals the yield stress, σ_y, at temperature M_s^σ. Consequently, strain-induced martensitic nucleation, which occurs after yielding, depends on the creation of a nucleation site by plastic deformation. An intersection of two shear bands was proposed as the nucleation site from experimental observations of strain-induced nucleation in 304-type stainless steel.²⁵⁾ Strain-induced nucleation behavior depends on a strain which controls nucleation site, and the critical stress follows σ_y between M_s^σ and M_d.

Fig. 1.

Critical stress models for the initiation of martensite transformation as a function of temperature. (a) Original schematic.²⁰⁾ (b) Modified schematic.^26,27)

Onodera et al.^26,27) observed deformation-induced martensite in a Fe-25Ni-0.65C (wt%) alloy that nucleated around a deformation twin above M_s^σ and analyzed its crystallographic orientation relationship with the parent γ. Their model proposed that martensite transformation induced by plastic deformation depends on stress concentration due to the impingement of deformation twins and applied stress. They proposed a critical stress diagram based on their experimental results (Fig. 1(b)) that is a modification of the model shown in Fig. 1(a). The model depicted in Fig. 1(b) shows that the strain-induced martensite transformation is the same as the stress-assisted transformation, except for the stress concentration due to plastic strain.²⁸⁾ Consequently, the critical stress between M_s^σ and M_d is located just below the dotted line, which indicates an extension of the critical stress between M_s and M_s^σ, and the difference between these two values corresponds to the stress concentration.

Material designs for utilizing the TRIP phenomenon will differ depending on the deformation-induced transformation (DIT) model because the dominant factor for initiating the transformation is different. The model depicted in Fig. 1(a) depends primarily on the applied strain, and the model shown in Fig. 1(b) depends primarily on the applied stress. Moreover, as Olson and Cohen mentioned,²⁰⁾ these models explain only the nucleation phenomenon and do not necessarily depict macroscopic transformation behavior.

Hiwatashi et al.^6,29) presented the DIT behavior of low-alloy TRIP steel under four deformation conditions, as shown in Fig. 2. They found that amount of transformation from retained γ to martensite that occurred by a shrink flanging deformation mode was less than under the other conditions (i.e., uniaxial, plane strain, and equi-biaxial deformation). They suggested that compressive hydrostatic stress during shrink flanging deformation prevented the transformation progress. This result indicates that the DIT behavior in low-alloy TRIP steel depends on the magnitude of applied stress and its character. To identify the dominant factor of the DIT transformation in low-alloy TRIP steel, we measured the relationship between the transformation behavior and macroscopic strain and stress applied during in-plane tension and compression experiments.³⁰⁾

Fig. 2.

Changes in the volume fraction of austenite versus equivalent plastic strain under the four deformation conditions.²⁹⁾

2. Experimental Procedure

2.1. Materials

Two CSiMn low-alloy TRIP steel sheets were fabricated at laboratory scale, and the transformation behaviors during tensile deformation were compared with those of four steel sheets that had been evaluated in previous studies. Table 1 lists the details of the two steels used to fabricate the samples tested in this study (Steels A and B) and those of the steels evaluated in the referenced data (Steels W, X, Y, and Z). Steels A and B contain 0.1 and 0.2 mass% carbon, respectively, and their microstructures consist of polygonal ferrite, blocky retained austenite, and small amounts of bainite.^5,6) Therefore, the 13.8% initial volume fraction of austenite (V_γ,0) in Steel B is higher than that in Steel A (5.5%). Although the values of the 0.2% proof stress (σ_0.2%) and tensile strength of Steel B are 1.3 and 1.4 times those of Steel A, respectively, the UEL of Steel B is only 0.9 times that of Steel A. The balance between strength and elongation is improved in Steel B because of the increase in V_γ,0.

Table 1. Low-alloy TRIP steel sheets used in this study and referenced steel sheets from previous studies:^30,31,32) their chemical compositions, microstructures, and tensile properties. The microstructure in this table corresponds to the shape, the volume fraction, and the solute carbon content of retained γ.

Steel	Thickness (mm)	Chemical compositions (mass%)	Microstructure			Tensile properties			Ref.
			Shape	Vγ,0 (%)	Cγ (mass%)	σ0.2% (MPa)	TS (MPa)	UEL (%)
A	1.2	0.1C–1.2Si–1.5Mn	blocky	5.5	1.22	417	626	28
B	1.6	0.2C–1.2Si–2.0Mn	blocky	13.8	1.11	544	896	25
W	1.4	0.3C–1.5Si–2.0Mn	blocky	22.9	1.11	541	994	26	[31]
X	1.4		plate-like	24.6	1.19	700	1006	32
Y	1.6	0.2C–1.5Al–1.5Mn	blocky	15.1	1.30	433*	867	28	[32]
Z	3.0	0.05C–0.3Si–8.5Mn	plate-like	30	0.12	830	1034	18	[33]

The data for the CSiMn low-alloy TRIP steel sheets that correspond to Steels W and X were obtained from graphs in Tsuchida et al.³¹⁾ The carbon content of these steels is 0.3 mass% higher than that of Steel B; consequently, the V_γ,0 values are higher than that of Steel B. The microstructure of Steel W consisting of polygonal ferrite and blocky γ grains is similar to those of Steels A and B. In contrast, the ferrite and γ grains of Steel X are plate-like, and the values of σ_0.2% and UEL are superior to those of Steel W. Steel Y is a CAlMn low-alloy TRIP steel sheet studied by McDermid et al.³²⁾ The carbon content and V_γ,0 of Steel Y are as high as those of Steel B, and the shape of the microstructure of Steel Y is similar to Steels A, B, and W. Steel Z is a hot-rolled medium manganese steel sheet studied by Han et al.³³⁾ Steel Z contains plate-like ferrite and plate-like γ grains, like those of Steel X. Steel Y has an obvious yield point and yield plateau,³²⁾ and the other steel samples exhibit round-house type stress-strain curves.

2.2. In-plane Stress Reversal Test

We used the apparatus and method for conducting deformation tests proposed by Kuwabara et al.³⁰⁾ This procedure applies continuous in-plane stress reversals on a specimen while measuring the tension and compression forces applied by a load cell connected to the apparatus. The geometry of the test specimen was type JIS #5 along the transverse direction of the steel sheet. The strain was measured with a non-contact digital video extensometer, and the gage length was 10 mm. A blank holding pressure of 32 MPa was applied to each specimen to prevent it from buckling during in-plane compression.

The relationship between the nominal plastic strain measured by the extensometer without loading and the change in the cross-sectional area (S/S₀) at the center of the specimen is shown in Fig. 3. S₀ and S are cross-sectional areas of samples as-prepared and after deformation, respectively. We applied simple tensile deformation on Group T samples. Group C samples were deformed by simple compressive deformation without buckling. Group TC samples were subjected to compressive deformation immediately after tensile deformation, and Group CT samples received tensile deformation after compression. These deformations were interrupted with several strains to avoid inducing necking and buckling. The broken line corresponds to a constant volume condition, and each measured value of the cross-sectional area is on this line. We used the measured nominal plastic strain values to calculate true plastic strain values.

Fig. 3.

Relationship between the nominal plastic strain and the change in the cross-sectional area measured in Steel B samples deformed using an in-plane stress reversal test.³⁰⁾ T, C, CT, and TC correspond to strain path: simple tensile deformation, simple compressive deformation, tensile deformation after compressive deformation, and compressive deformation after tensile deformation, respectively.

The strain rate during tensile and compressive deformations was constant at 5.0 × 10⁻⁴ s⁻¹. This value is similar to those reported in the reference data, 3.3 × 10⁻⁴ s⁻¹,³¹⁾ 6.7 × 10⁻⁴ s⁻¹,³²⁾ and 1.0 × 10⁻⁴ s⁻¹.³³⁾

2.3. Observation

To avoid the region of Mn segregation, we observed the microstructure of the samples on a plane parallel to the sheet surface at one-quarter of the thickness. X-ray diffraction (XRD) with a Mo K_α source was used to evaluate the retained γ. Its solute carbon content, C_γ (mass%), was measured from the lattice parameter (a_γ) determined by the (2 0 0), (2 2 0), and (3 1 1) FCC diffraction peaks, using the following:³⁴⁾   

$$\begin{array}{c}{a}_{\gamma }\left(\mathring{\mathrm{A}}\right)=3.572+0.0012\times \mathrm{M}\mathrm{n}\left(\mathrm{m}\mathrm{a}\mathrm{s}\mathrm{s}\%\right)-0.00157\\ \times \mathrm{S}\mathrm{i}\left(\mathrm{m}\mathrm{a}\mathrm{s}\mathrm{s}\%\right)+0.033\times {\mathrm{C}}_{\gamma }\left(\mathrm{m}\mathrm{a}\mathrm{s}\mathrm{s}\%\right)\end{array}$$(1)

V_γ,0 was determined by comparing the integrated intensities between the FCC peaks referred to above and the (2 0 0) and (2 1 1) BCC diffraction peaks.

The microstructure after deformation was observed at one-quarter of the thickness and at the center of the width and length of each specimen. The volume fraction of γ in the deformed specimen, V_γ, was measured like V_γ,0. However, we did not evaluate the solute carbon content of γ in deformed specimens, because that value is difficult to obtain using XRD due to the lattice parameter changes that occur during deformation and DIT.

To evaluate phase strain, which we divided into ferrite and retained γ, we observed the microstructure of deformed samples using field-emission scanning electron microscopy (FE-SEM; JEOL-6500F) with an electron backscattered diffraction (EBSD) analysis system (OIM Data Collection ver. 7). The observation area was 80 μm × 100 μm with a step of 0.15 μm. We did not use pixel data with a confidential index (CI) less than 0.02 or an image quality (IQ) less than 15000 to remove error and eliminate data from martensite that contained multiple dislocations without deformation. We used Karnel average misorientation (KAM) parameter analysis to evaluate the phase strain in ferrite³⁵⁾ and γ.³⁶⁾ The KAM value was used to determine the geometrically necessary dislocation (GND) density.^35,37) We used the peak value of the KAM distribution over the entire area as the KAM value of a deformed specimen.

3. Results

3.1. Transformation during Tensile Deformation

The relationship between DIT behavior and true strain during simple tensile deformation (Group T) is shown in Fig. 4. We experimentally measured the data reported for Steels A and B and obtained the data for Steels W, X, Y, and Z from the literature.^31,32,33) The retained γ in each steel sheet decreases continuously as tensile strain increases. These behaviors indicate that DIT is occurring.

Fig. 4.

Relationship between the applied macroscopic strain and transformation behavior of retained austenite in steel sheets during tensile deformation.

V_γ appears to decrease as deformation begins in Steels B, C, D, E, and F. However, the retained γ in Steel A did not begin to transform at low strain (0.039), and the rate at which V_γ decreases as unit tensile strain increases is less than those in other steel sheets. The transformation behavior of Steel B was similar to those of Steels W, X, and Y. Notably, the values of V_γ,0 in Steels B and Y are almost equal, and their V_γ curves during increasing tensile strain overlap. In Steels W and X, the rates at which V_γ decreases are similar to those in Steels B and Y; however, V_γ remains constant during heavy deformation. V_γ in Steel Z decreases during deformation more readily than in other steel sheets.

A V_γ vs. strain graph is typically used to illustrate the DIT behavior in steel because it is suitable for comparing macroscopic work hardening behaviors^{5,7,29,31,32,33,38,39,43,44)} and considering transformation behaviors using the strain-dominant strain-induced transformation (SIT) model^{5,7,31,32,38,39,40,41,42,43,44)} shown in Fig. 1(a).

In contrast, it is preferable to compare V_γ with applied stress⁴⁵⁾ when considering behaviors using the stress-dominant SIT model, as shown in Fig. 1(b). The relationship between the DIT behavior shown in Fig. 4 and the macroscopic true stress during simple tensile deformation is shown in Fig. 5. V_γ decreases linearly in each steel sheet as the macroscopic true stress increases. Notably, V_γ in Steel Z³³⁾ begins decreasing from a V_γ,0 of approximately σ_0.2% (830 MPa) and decreases linearly during the entire range of uniform deformation. This result indicates that the SIT, (i.e., the DIT during plastic deformation) depends on applied stress in Steel Z.

Fig. 5.

Relationship between the applied macroscopic stress and transformation behavior of retained austenite in steel sheets during tensile deformation.

The thin lines in Fig. 5 represent the predicted transformation behaviors in each steel sheet if their retained γ transforms in the same manner as Steel Z. In Steels A, B, W, and X, which are CSiMn steel sheets that contain similar amounts of Si and Mn but different C content, the rates at which V_γ decreases with applied true stress are equal. Moreover, the predicted values of critical stress at which transformation begins are similar (~510 MPa). In contrast, the rates at which V_γ decreases with applied true stress in Steel Y, a CAlMn steel sheet, and Steel Z, a medium manganese steel sheet, are faster than those in CSiMn steel sheets. The predicted value of the stress at which transformation begins around yielding for Steel Y was less than those for CSiMn steel sheets.

A Kocks-Mecking plot^46,47) around yielding in Steels A and B is shown in Fig. 6. The values of the deformation hardening rate, dσ/dε, are constant until the applied stress of 300 MPa is achieved. This result means that the steels undergo elastic deformation below 300 MPa. Above 300 MPa, the hardening rates begin to decrease, which indicates that plastic deformation has started. Therefore, the DIT in Steels A and B, as shown in Fig. 5, started during macroscopic plastic deformation during tensile deformation.

Fig. 6.

Yield behaviors of Steels A and B during tensile deformation, using a Kocks-Mecking plot.

3.2. Transformation during Compressive Deformation

The relationship between DIT behavior and applied strain in Steels A and B during simple compressive deformation (Group C) are shown in Fig. 7. V_γ in Steel B decreases as compressive strain increases, as it does during tensile deformation. However, the rate at which V_γ decreases during compressive deformation is substantially less than that during tensile deformation. In Steel A, the value of V_γ under large strains is slightly higher than that during tensile deformation, although this tendency is not obvious.

Fig. 7.

Relationship between the applied macroscopic strain and transformation behavior of retained austenite in Steels A and B during compressive or tensile deformation.

The relationship between the transformation behavior and macroscopic true stress in Steels A and B during compressive deformation is shown in Fig. 8. V_γ decreases linearly as compressive stress increases, as it does during tensile deformation. In Steel B, the rate at which V_γ decreases with compressive stress is similar to that with tensile stress, and the critical stress at which transformation begins during compressive deformation is notably higher than that during tensile deformation.

Fig. 8.

Relationship between the applied macroscopic stress and transformation behavior of retained austenite in Steels A and B during compressive or tensile deformation.

Conversely, in Steel A, the at which V_γ decreases during compressive deformation is slower than that during tensile deformation, and the critical stress is similar despite the change in the mode of deformation. Although the retained γ is more stable during compressive deformation than during tensile deformation, the details of this phenomenon might differ between Steels A and B.

The relationship between yielding behavior and compression stress in Steels A and B is shown in Fig. 9. Although the elastic limits under compressive stress in Fig. 9 are less obvious than those under tensile stress, they occur compressive stress values below 400 MPa. This result means that the DIT during compressive deformation in Steels A and B begins after the plastic deformation has started; therefore, the DIT is the SIT.

Fig. 9.

Yield behaviors of Steels A and B during compressive deformation, using a Kocks-Mecking plot.

3.3. Transformation during Continuous Reverse Deformation

Group CT and TC samples were subjected to continuous reverse deformation: from compression to tension and from tension to compression, respectively. Table 2 lists the measured true strains of Group CT and TC samples. The column named “1st deformation” indicates the first half of the deformation applied to the samples (i.e., compressive deformation for CT samples and tensile deformation for TC samples). The samples listed in Table 2 received 1st deformation with in-plane nominal strain ε_n and were then unloaded briefly. The column named “2nd deformation” indicates the second half of the deformation applied to the samples immediately after unloading. The final apparent strain is the ratio of nominal strain to the initial gage length after the 2nd deformation. For example, the CT sample listed first in Table 2 received 4% tensile deformation after 4% compression to establish the initial shape of the specimen; consequently, the final apparent strain is ± 0%. The values of ε_p in the 1st and 2nd deformation columns are the plastic true strain without stress, and the label (C) or (T) indicates that the associated value is compressive true plastic strain (ε_c) or tensile strain (ε_t), respectively. The macroscopic elastic limits of Steels A and B shown in Figs. 6 and 9 are smaller than their respective 0.2% proof stress values; therefore, the samples listed in Table 2 undergo DIT after plastic deformation begins.

Table 2. Specimens subjected to continuous reverse deformation.

Steel	Group	1st deformation		2nd deformation		Final apparent strain	Vγ (%)
		εn	εp	εn	εp
A	CT	−4%	0.038 (C)	+4%	0.036 (T)	±0%	4.9
			0.038 (C)	+6%	0.056 (T)	+2%	4.5
			0.038 (C)	+8%	0.075 (T)	+4%	4.0
			0.038 (C)	+12%	0.113 (T)	+8%	3.3
		−8%	0.080 (C)	+12%	0.116 (T)	+4%	2.5
			0.080 (C)	+18%	0.172 (T)	+10%	1.8
	TC	+4%	0.037 (T)	−4%	0.035 (C)	±0%	4.6
			0.037 (T)	−6%	0.054 (C)	−2%	4.4
			0.037 (T)	−8%	0.075 (C)	−4%	4.4
			0.037 (T)	−12%	0.117 (C)	−8%	3.8
		+10%	0.093 (T)	−14%	0.130 (C)	−4%	2.9
			0.092 (T)	−18%	0.172 (C)	−8%	1.8
B	CT	−2%	0.018 (C)	+4%	0.035 (T)	+2%	10.6
			0.017 (C)	+8%	0.072 (T)	+6%	7.7
		−6%	0.058 (C)	+12%	0.112 (T)	+6%	6.0
	TC	+2%	0.018 (T)	−6%	0.055 (C)	−4%	10.3
			0.016 (T)	−8%	0.073 (C)	−6%	8.9
		+6%	0.055 (T)	−12%	0.112 (C)	−6%	6.7

Group CT and TC samples with their final apparent strains are shown in Fig. 10, which illustrates the differences in the DIT behavior between Steels A and B. The data for Steel A specimens are shown in Fig. 10(a). The values of V_γ for Group CT samples appear to be smaller than those for Group T samples. For Group TC samples, which received 4% tensile deformation during the 1st deformation, the values of V_γ appear to be similar to those of Group C samples. The data for Steel B specimens are shown in Fig. 10(b). The values of V_γ for Group CT samples appear to be similar to those for Group T samples. The values of V_γ for Group TC samples appear to be smaller than those for Group C samples. However, the sorting in Fig. 10 has little value because the plastic strain, which induces transformation according to the model shown in Fig. 1(a), does not correspond to the final apparent strain.

Fig. 10.

Relationship between transformation behavior and apparent nominal strain. Four groups (T, C, CT, and TC) of samples were subjected to simple tensile deformation, simple compressive deformation, tensile deformation after compressive deformation, and compressive deformation after tensile deformation, respectively. (a) Transformation behaviors of Steel A. (b) Transformation behaviors of Steel B.

The relationship between V_γ and the true plastic strain in Group CT and TC samples is shown in Fig. 11. In Figs. 11(a) and 11(b), the horizontal axis corresponds to the total true plastic strain, which equals ε_c + ε_t for each sample, regardless of its deformation path. The data for Steel A specimens are shown in Fig. 11(a). V_γ in Groups CT and TC samples are slightly larger than those in Groups T and C samples at the same amount of total true plastic strain. The data for Steel B specimens are shown in Fig. 11(b). V_γ in Group CT and TC samples are between those in Groups T and C samples at the same amount of total true plastic strain.

Fig. 11.

(a), (b) Relationships between the DIT behavior and applied total plastic strain (ε_t + ε_c). (c) (d) Relationships between transformation behavior and tensile plastic strain, ε_t. (e), (f) Relationships between transformation behavior and compressive plastic strain, ε_c. (a), (c), and (e) show data for Steel A, and (b), (d), and (f) show data for Steel B.

The relationship between V_γ with ε_t are shown in Figs. 11(c) and 11(d). In Steel A (Fig. 11(c)) and Steel B (Fig. 11(d)), V_γ in Group CT samples is similar to that in Group T samples at the same amount of ε_t, regardless of ε_c. The relationships between V_γ and ε_c are shown Figs. 11(e) and 11(f). In Steel A specimens (Fig. 11(e)), V_γ in Group TC samples is similar to that in Group C samples at the same amount of ε_c, regardless of ε_t. Therefore, in the CT and TC samples of Steel A and the CT samples of Steel B, the transformation behaviors under plastic strain during the 2nd deformation appear to be similar to those under simple one-way deformation, regardless of the 1st deformation mode. However, in Fig. 11(f), V_γ in Group TC samples of Steel B is smaller than that in Group C samples at the same amount of ε_c.

As shown in Figs. 10 and 11, it is difficult to find a trend that is common among the four deformation groups of Steel A and B specimens by comparing the DIT behavior and applied strain.

However, the relationships between the DIT behavior and applied stress are simple. We examined the relationships between V_γ and applied true stress at the end of the 2nd deformation for Groups CT and TC in Fig. 12. The thin lines in Fig. 12 correspond to the transformation behaviors of Group T and C, as shown in Fig. 8. For Steels A and B, the V_γ values of Groups CT and TC overlap those of Groups T and C, respectively. Therefore, the values of V_γ in Steel A and B specimens depend on the macroscopic true stress applied at the end of the deformation, regardless of the 1st deformation mode.

Fig. 12.

Relationship between the deformation-induced transformation behavior and applied stress at the end of deformation. (a) Steel A. (b) Steel B.

4. Discussion

4.1. DIT during Plastic Deformation

The initiation of DIT after macroscopic yield in Steels A and B is similar during tensile and compressive deformations. Typically, the DIT in low-alloy TRIP steel is called the SIT^{5,7,31,32,38)} because it occurs during macroscopic plastic deformation. However, the initiation of macroscopic plastic deformation in low-alloy TRIP steel is not necessarily the initiation of plastic deformation in retained γ because of the microstructural constituents of low-alloy TRIP (e.g., ferrite, bainite, and retained γ), and the elastic limits of these structures are different. Therefore, to apply the models for SIT shown in Fig. 1, it is essential to identify the initiation of plastic deformation start in retained γ.

Tomota et al.⁴⁸⁾ evaluated the DIT behavior in CSiMn low-alloy TRIP steel during tensile deformation using in-situ neutron diffraction. Although they showed that the initiation of plastic deformation in retained γ occurred after macroscopic plastic deformation, the interval between them was small. According to this and other^{49,50,51,52,53,54)} in-situ neutron diffraction data, the retained γ in low-alloy TRIP steel deformed plastically immediately after macroscopic yielding, and the start of transformation occurred with plastic deformation. These results mean that, macroscopically, the DIT in low-alloy TRIP steel is the SIT that occurs with plastic deformation in γ.

The microstructures, tensile properties, and transformation behaviors in Steels A and B are similar to those observed in materials used in previous studies. Therefore, the finding that the DIT is the SIT in Steels A and B was also expected.

4.2. Strain-dominant SIT

The SIT behaviors in Steels A and B during tensile deformation shown in Fig. 4 are similar to those reported for steels used in previous studies,^{38,39,40,41,42,43,44)} in which the transformation behaviors were described by empirical formulas⁵⁵⁾ based on the model shown in Fig. 1(a). For example, Sugimoto et al.^39,56) proposed the following equation:   

$$ln{V}_{\gamma }=ln{V}_{\gamma ,0}-k\cdot \epsilon ,$$(2)

where V_γ depends on the applied plastic strain ε, and k is parameter of the instability of γ. It is possible to consider that the transformation in Steels A and B during tensile deformation can be modeled as described in these previous studies because their behaviors are similar.

The transformation behavior during compressive deformation shown in Fig. 7 was slower than that during tensile behavior. However, these behaviors are alike except for the rate at which V_γ decreases; therefore, it is possible to consider that the transformation behavior during compressive transformation in Steels A and B can be modeled as described in these previous studies. For example, to express the transformation behavior during compression with Eq. (2), the instability parameter k is smaller than that for tensile deformation. In Steel B, the values of k for compression and tension are 3.21 and 8.26, respectively.

However, it is difficult to explain the transformation behavior during continuous reverse deformation shown in Fig. 11. The relationship between the KAM value and total plastic strain, ε_t, in Steel B is shown in Fig. 13. The KAM value corresponds to dislocation density;^35,37) therefore, V_γ is expected to decrease as the KAM value increases from the perspective of the strain-dominant SIT model in Fig. 1(a). As seen in Fig. 13, KAM increases as ε_t increases in all groups. However, in Fig. 11(b), the rate at which V_γ decreases depends on the strain path of the specimens. Moreover, from the perspective of strain-dominant SIT, it is difficult to understand the stabilization of γ by deformation.⁵⁷⁾

Fig. 13.

Relationship between the KAM value and total plastic strain (ε_t + ε_c) in Steel B. (a) KAM in austenite. (b) KAM in ferrite + bainite matrix.

4.3. Stress-dominant SIT

Values of V_γ in Steels A and B decrease linearly as applied stress increases. In previous studies,^50,51,53) this same behavior has been measured using in-situ neutron diffraction during tensile deformation. McDermid et al.³²⁾ compared V_γ during tensile deformation with the amount of work hardening normalized to the yield strength in low-alloy TRIP steel. The results suggested that the DIT behavior could be expressed with a simple equation containing flow stress and yield strength.

According to the stress-dominant SIT model^26,27,28) shown in Fig. 1(b), it is preferable to compare the SIT behavior and the applied stress, without considering the yield stress. The data shown in Figs. 5, 8, and 12 demonstrate that V_γ depends on the magnitude and character of applied stress, regardless of the strain path. From these results, the SIT behavior in low-alloy TRIP steel can be described by the following simple equation:   

$$V_{\gamma }=V_{\gamma ,0}-h\cdot \left(\sigma -{\sigma }_s\right),$$(3)

where σ_s and h correspond to the critical stress required for transformation to start and the rate of decrease, respectively. These are the parameters of the instability of retained γ. Therefore, they will depend on chemical,^32,33,58) morphological,^{31,33,48,53,59,60,61)} and crystallographic^50,61,62) factors under deformation conditions.^{29,31,41,54,63)}

In low-alloy TRIP steel, the stress-dominant SIT model shown in Fig. 1(b) does not directly depend on applied stress because of the microstructural constituents of TRIP steel. Therefore, an estimation of the stress in retained γ is required to understand the transformation behavior. Perlade et al.⁶⁴⁾ proposed a model to predict the strain and stress in the retained γ in low-alloy TRIP steel using the iso-work (Iso-W) hypothesis.^65,66) Our calculation of the relationship between the applied stress and flow stress in the ferrite + bainite matrix and retained γ in Steel B based on this model is shown in Fig. 14. The predicted flow stress in retained γ increases almost linearly as applied stress increases after the yielding of γ at 750 MPa. Therefore, this calculation suggests that V_γ has a linear dependence with the stress in retained γ, just as it does with the applied stress, as shown in Figs. 5, 8, and 12.

Fig. 14.

Relationship between applied the macroscopic stress and flow stress in austenite and ferrite + bainite matrix, as predicted by the Iso-W hypothesis.^64,65,66)

As mentioned above, it is appropriate to explain the SIT behavior in low-alloy TRIP steel using the stress-dominant SIT model shown in Fig. 1(b). However, it should be noted that this assertion applies to the macroscopic transformation behavior, and does not necessarily apply to the local nucleation behavior.

5. Summary

The aim of this study was to clarify the DIT behavior of retained γ in low-alloy TRIP steel. We evaluated the change in the volume fraction of γ (V_γ) during in-plane tension and compression experiments.

(1) V_γ decreased with tensile deformation as in previous studies of low-alloy TRIP steel, and the rate at which V_γ decreased was proportional to the macroscopic true stress.

(2) V_γ also decreased with compressive deformation. The rates at which V_γ decreased with macroscopic strain and stress were slower than those during tensile deformation.

(3) V_γ depended on the macroscopic true stress applied at the end of the entire deformation after continuous reverse deformation from tensile to compressive and from compressive to tensile. In contrast, there was no apparent relationship between V_γ and applied strain.

(4) In this study, the initiation of DIT occurred after macroscopic yield. The transformation observed in this study would be SIT because the retained γ in low-alloy TRIP steel deformed plastically immediately after macroscopic yield in previous studies with in-situ neutron diffraction.

(5) It is appropriate to apply the stress-dominant SIT model first proposed by Onodera et al. to explain the results of this study.

Acknowledgments

The authors would like to express their sincere thanks to Dr. Norimitsu Koga (Kanazawa University, Japan), Dr. Takayuki Yamashita, and Dr. Satoshi Morooka (Japan Atomic Energy Agency) for stimulating discussions and valuable comments.

References

• 1)   E. C.  Bain: Trans. Am. Soc. Steel Treat., 8 (1925), 14.
• 2)   V. F.  Zackay,  M. D.  Bhandarkar and  E. R.  Parker: Advances in Deformation Processing, eds. by J. J. Burke and V. Weiss, Plenum Press, New York, (1978), 351.
• 3)   C. K.  Divers: Met. Prog., 86 (1964), 115.
• 4)   V. F.  Zackay,  E. R.  Parker,  D.  Fahr and  R.  Busch: Am. Soc. Met. Trans. Q., 60 (1967), 252.
• 5)   Y.  Sakuma,  O.  Matsumura and  H.  Takechi: Metall. Trans. A, 22 (1991), 489.
• 6)   Y.  Sakuma,  N.  Kimura,  A.  Itami,  S.  Hiwatashi,  O.  Kawano and  K.  Sakata: Nippon Steel Tech. Rep., (1995), No. 64, 20.
• 7)   B. C.  De Cooman: Curr. Opin. Solid State Mater. Sci., 8 (2004), 285.
• 8)   M.  Takahashi: ISIJ Int., 55 (2015), 79.
• 9)   K.  Sugimoto and  M.  Mukherjee: Automotive Steels: Design, Metallurgy, Processing and Applications, eds. by R. Rana and S. B. Singh, Woodhead Publishing, Cambridge, UK, (2016), 217.
• 10)   D. K.  Matlock and  J. G.  Speer: Proc. 3rd Int. Conf. on Structural Steels, ed. by H. C. Lee, KIMM, Seoul, (2006), 774.
• 11)   J.  Speer,  D. K.  Matlock,  B. C.  De Cooman and  J. G.  Schroth: Acta Mater., 51 (2003), 2611.
• 12)   D. V.  Edmonds,  K.  He,  F. C.  Rizzo,  B. C.  De Cooman,  D. K.  Matlock and  J. G.  Speer: Mater. Sci. Eng. A, 438–440 (2006), 25.
• 13)   R.  Miller: Metall. Mater. Trans. B, 3 (1972), 905.
• 14)   T.  Furukawa: Mater. Sci. Technol., 5 (1989), 465.
• 15)   F. G.  Caballero and  H. K. D. H.  Bhadeshia: Curr. Opin. Solid State Mater. Sci., 8 (2004), 251.
• 16)   F. G.  Caballero,  M. J.  Santofimia,  C.  Capdevila,  C.  Garcia-Mateo and  C. G.  de Andrés: ISIJ Int., 46 (2006), 1479.
• 17)   K.  Sugimoto,  A.  Kanda,  R.  Kikuchi,  S.  Hashimoto,  T.  Kashima and  S.  Ikeda: ISIJ Int., 42 (2002), 910.
• 18)   L.  Liu,  B.  He and  M.  Huang: Adv. Eng. Mater., 20 (2018), 1701083.
• 19)   E.  Scheil: Z. Anorg. Allg. Chem., 207 (1932), 21.
• 20)   G. B.  Olson and  M.  Cohen: J. Less-Common Met., 28 (1972), 107.
• 21)   G. F.  Bolling and  R. H.  Richman: Philos. Mag., 19 (1969), 247.
• 22)   G. F.  Bolling and  R. H.  Richman: Scr. Metall., 4 (1970), 539.
• 23)   G. F.  Bolling and  R. H.  Richman: Acta Metall., 18 (1970), 673.
• 24)   R. H.  Richman and  G. F.  Bolling: Metall. Trans., 2 (1971), 2451.
• 25)   J. A.  Venables: Philos. Mag., 7 (1962), 35.
• 26)   H.  Onodera,  H.  Oka and  I.  Tamura: J. Jpn. Inst. Met., 42 (1978), 898 (in Japanese).
• 27)   I.  Tamura: Met. Sci., 16 (1982), 245.
• 28)   T.  Suzuki,  H.  Kojima,  K.  Suzuki,  T.  Hashimoto and  M.  Ichihara: Acta Metall., 25 (1977), 1151.
• 29)   S.  Hiwatashi,  M.  Takahashi,  T.  Katayama and  M.  Usuda: J. Jpn. Soc. Technol. Plast., 35 (1994), 1109 (in Japanese).
• 30)   T.  Kuwabara,  Y.  Kumano,  J.  Ziegelheim and  I.  Kurosaki: Int. J. Plast., 25 (2009), 1759.
• 31)   N.  Tsuchida,  S.  Okura,  T.  Tanaka and  Y.  Toji: ISIJ Int., 58 (2018), 978.
• 32)   J. R.  McDermid,  H. S.  Zurob and  Y.  Bian: Metall. Mater. Trans. A, 42 (2011), 3627.
• 33)   J.  Han,  S.-J.  Lee,  J.-G.  Jung and  Y.-K.  Lee: Acta Mater., 78 (2014), 369.
• 34)   R.  Ruhl and  M.  Cohen: Trans. Metall. Soc. AIME, 245 (1969), 241.
• 35)   N.  Allain-Bonasso,  F.  Wagner,  S.  Berbenni and  D. P.  Field: Mater. Sci. Eng. A, 548 (2012), 56.
• 36)   H.  Kimura,  Y.  Wang,  Y.  Akiniwa and  K.  Tanaka: Trans. Jpn. Soc. Mech. Eng. A, 71 (2005), 1722 (in Japanese).
• 37)   D. P.  Field,  P. B.  Trivedi,  S. I.  Wright and  M.  Kumar: Ultramicrosopy, 103 (2005), 33.
• 38)   O.  Matsumura,  Y.  Sakuma and  H.  Takechi: Scr. Metall., 21 (1987), 1301.
• 39)   K.  Sugimoto,  M.  Kobayashi and  S.  Hashimoto: Metall. Trans. A, 23 (1992), 3085.
• 40)   H. C.  Shin,  T. K.  Ha and  Y. W.  Chang: Scr. Mater., 45 (2001), 823.
• 41)   M. Y.  Sherif,  C.  Garcia-Mateo,  T.  Sourmail and  H. K. D. H.  Bhadeshia: Mater. Sci. Technol., 20 (2004), 319.
• 42)   M.  Mukherjee,  S. B.  Singh and  O. N.  Mohanty: Metall. Mater. Trans. A, 39 (2008), 2319.
• 43)   N.  Tsuchida,  T.  Araki,  Y.  Yamaguchi and  K.  Fukaura: Tetsu-to-Hagané, 98 (2012), 558 (in Japanese).
• 44)   N.  Tsuchida,  Y.  Yamaguchi,  Y.  Morimoto,  T.  Tonan,  Y.  Takagi and  R.  Ueji: ISIJ Int., 53 (2013), 1881.
• 45)   J. R.  Patel and  M.  Cohen: Acta Metall., 1 (1953), 531.
• 46)   H.  Mecking and  U. F.  Kocks: Acta Metall., 29 (1981), 1865.
• 47)   U. F.  Kocks and  H.  Mecking: Prog. Mater. Sci., 48 (2003), 171.
• 48)   Y.  Tomota,  H.  Tokuda,  Y.  Adachi,  M.  Wakita,  N.  Minakawa,  A.  Moriai and  Y.  Morii: Acta Mater., 52 (2004), 5737.
• 49)   P. J.  Jacques,  Q.  Furnemont,  S.  Godet,  T.  Pardoen,  K. T.  Conlon and  F.  Delannay: Philos. Mag., 86 (2006), 2371.
• 50)   O.  Muránsky,  P.  Šittner,  J.  Zrník and  E. C.  Oliver: Acta Mater., 56 (2008), 3367.
• 51)   A.  Narui,  S.-C.  Chen,  Y.  Tomota and  T.  Kamiyama: Trans. Jpn. Soc. Mech. Eng. A, 75 (2009), 501 (in Japanese).
• 52)   S.  Harjo,  N.  Tsuchida,  J.  Abe and  W.  Gong: Sci. Rep., 7 (2017), Article No. 15149.
• 53)   N.  Tsuchida,  T.  Tanaka and  Y.  Toji: Tetsu-to-Hagané, 105 (2019), 918 (in Japanese).
• 54)   T.  Yamashita,  S.  Morooka,  S.  Harjo,  T.  Kawasaki,  N.  Koga and  O.  Umezawa: Scr. Mater., 177 (2020), 6.
• 55)   G. B.  Olson and  M.  Cohen: Metall. Trans. A, 6 (1975), Article No. 791.
• 56)   K.  Sugimoto,  N.  Usui,  M.  Kobayashi and  S.  Hashimoto: ISIJ Int., 32 (1992), 1311.
• 57)   R. H.  Richman and  G. F.  Bolling: Metall. Trans., 2 (1971), 2451.
• 58)   J.  Wang and  S.  van der Zwaag: Metall. Mater. Trans. A, 32 (2001), 1527.
• 59)   Y.  Matsuoka,  T.  Iwasaki,  N.  Nakada,  T.  Tsuchiyama and  S.  Takaki: ISIJ Int., 53 (2013), 1224.
• 60)   A.  Kisko,  R. D. K.  Misra,  J.  Talonen and  L. P.  Karjalainen: Mater. Sci. Eng. A, 578 (2013), 408.
• 61)   D.  De Knijf,  C.  Föjer,  L. A. I.  Kestens and  R.  Petrov: Mater. Sci. Eng. A, 638 (2015), 219.
• 62)   T.  Yamashita,  N.  Koga and  O.  Umezawa: ISIJ Int., 58 (2018), 1155.
• 63)   T.  Iwamoto,  T.  Tsuta and  Y.  Tomita: Int. J. Mech. Sci., 40 (1998), 173.
• 64)   A.  Perlade,  O.  Bouaziz and  Q.  Furnémont: Mater. Sci. Eng. A, 356 (2003), 145.
• 65)   O.  Bouaziz and  P.  Buessler: Rev. Métall., 99 (2002), 71.
• 66)   O.  Bouaziz and  P.  Buessler: Adv. Eng. Mater., 6 (2004), 79.