1 Introduction

Transformation-induced plasticity (TRIP) steels, with an outstanding combination of strength and ductility, are excellent candidates for structural applications in the automotive industry (Ref 1,2,3,4). The excellent mechanical properties of TRIP steel are mainly due to the TRIP effect, i.e., strain-induced transformation of retained austenite (RA) to martensite, leaded to the simultaneous increase of strength and ductility (Ref 4,5,6). To meet the new structural performance requirements, without considerable increase in the automotive mass, TRIP steels are indispensable (Ref 7, 8). Using TRIP steels to make automotive parts would increase energy absorption ability and improve automotive safety performance (Ref 9,10,11,12).

Understanding the dynamic deformation behavior of high strength steels (TRIP steels, medium Mn steel, high Mn steel, etc) is essential for its potential applications (Ref 13,14,15,16,17,18,19). Choi et al. (Ref 15) investigated the deformation behavior of low carbon TRIP sheet steels at strain rates in the range of 10−3-2.5 × 102 s−1, and indicated that with the increase of strain rates, the yield strength (YS) and ultimate tensile strength (UTS) of the experimental steel was increased. Van Slycken et al. (Ref 16) showed that with the increase of strain rates from 5.6 × 10−4 to 2 × 103 s−1, the YS, UTS and total elongation (TEL) of low-alloyed TRIP steels was increased. Wei et al. (Ref 17) studied the dynamic tensile behavior of TRIP-aided steels over wide range of strain rate (400-1600 s−1) and demonstrated that with the increase of strain rates, both YS and UTS of TRIP-aided steels was increased, but the TEL was decreased. They pointed out that the adiabatic temperature rise at high deformation rate increased the stability of retained austenite (RA), resulting in the decrease in the amount of transformed RA. Gao et al. (Ref 18) indicated that when the strain rate was increased from 0.01 to 200 s−1, UTS of Si-Mn TRIP steel was increased, and the TEL first decreased and subsequently increased. Additionally, they demonstrated that the transformation of RA to martensite was not suppressed by adiabatic heating.

He et al. (Ref 19) studied the tensile deformation behavior of Si-Mn TRIP steel over a wide range of strain rate (0.001-2000 s−1), and indicated that the TEL during dynamic tensile straining was always less than that at quasi-static tensile test because the transformation of RA to martensite was inhibited by localized deformation during dynamic tensile test. In general, these studies mainly focused on the dynamic mechanical properties of TRIP steels. However, there are relatively few studies on the microstructure evolution of TRIP steels deformed at high strain rates. Furthermore, it may be noted from these studies that there is controversy if the adiabatic temperature rise can enhance the stability of RA and inhibit the transformation of RA to martensite. This aspect is unclear and is the focus of the study described here.

In addition, in view of high strain rate, it is difficult to realize the controlled strain experiment to study the transformation process of RA to martensite in TRIP steel during dynamic deformation, which limits our understanding of the behavior of TRIP steels. This also affects the development of stamping technology of TRIP steel and the improvement of anti-collision performance.

Therefore, the mechanical behavior of a high strength low alloy (HSLA) TRIP steel under different strain rates was systematically studied here, and the influence of adiabatic temperature rise on the stability of RA and martensitic transformation under dynamic deformation condition is clarified. To elucidate this aspect, an original tensile specimen with variable cross section was designed to realize the controllable strain experiment and study the dynamic deformation process of HSLA TRIP steel, and enhance the understanding of microstructure evolution and plasticity-enhancing mechanism of HSLA TRIP steel during dynamic deformation.

2 Experimental

The nominal chemical composition of the experimental steel was Fe-1.47Mn-1.40Si-0.21C-0.025Nb (in wt.%). The steel was melted in a 50 kg vacuum furnace and cast as ingot. The ingot was heated at 1200 °C for 2 h and hot forged into a rectangular slab of size of 100 × 30 mm, followed by air cooling to room temperature. Then, the forged ingot was soaked at 1200 °C for 2 h and hot rolled to 3 mm thickness by rolling temperature in the range of 1150-900 °C. Finally, the hot rolling sheet was pickled in 20 vol.% HCl solution and cold-rolled to 1 mm thickness.

Two-stage salt bath heat treatment was adopted. First, the experimental steel was annealed at 825 °C for 3 min, and then quenched to the isothermal bainite transformation (IBT) temperature at 410 °C for 3 min, followed by cooling in air to ambient temperature. Tensile specimens had flat dog-bone geometry with a gage length of 10 mm and width of 5 mm were cut from the heat-treated sheet along the rolling direction. Quasi-static tensile tests were performed on SANSCMT5000 machine at low strain rate of 0.01 s−1. Dynamic tensile tests were performed on Zwick HTM 5020 machined at different high strain rates (10, 100, 200 s−1). The dimension of variable cross-section tensile specimen is shown in Fig. 1. A circular opening with different radius (0.5, 0.75, 1.0 and 1.25 mm) was made in the middle of the gage length section of tensile specimen to control the cross section of specimen, with the aim to study the dynamic deformation process of TRIP steel under controllable strain. In order to verify the feasibility of variable cross-section specimen tensile method, ABAQUS finite element software was used to simulate the strain condition in different variable cross-section specimens after tensile deformation at strain rate of 100 s−1 before high-speed tensile test. As shown in Fig. 2, with the increase of opening depth from 0.5 to 1.25 mm, the strain in the gage section at both the ends of the opening was gradually decreased. This suggested that the dynamic deformation process of TRIP steel under controllable strain can be studied by means of variable cross-section specimen tensile method. For convenience, variable cross-section specimens with opening depths of 0.5, 0.75, 1.0 and 1.25 mm are henceforth referred as 0.5-sample, 0.75-sample, 1.0-sample and 1.25-sample, respectively.

Fig. 1
figure 1

The dimension of variable cross-section tensile specimen (unit: mm)

Full size image
Fig. 2
figure 2

The strain state of samples with variable cross section simulated by ABAQUS during necking (a) 0.5-sample; (b) 0.75-sample; (c) 1.0-sample; (d) 1.25-sample

Full size image

In order to minimize the impact of strain state, which was different from area to area, after fracture, the open area was cut, then the specimen for microstructural examination was cut from the middle of the gage length section of tensile specimen. Microstructural examination was carried out by SEM (Zeiss Ultra Plus) equipped with electron backscatter diffraction (EBSD) and TEM (Tecnai G2 20). The detailed preparations for samples were described in reference (Ref 20). The volume fraction of RA (Vγ) was determined by XRD (Rigaku, D/Max2250/PC) with CuKα using direct comparison method (Ref 20).

3 Results

3.1 Mechanical Properties

Figure 3 shows the engineering stress–strain curve of the investigated steel at different strain rates. Some oscillations were observed in the engineering stress–strain curve, which was mainly due to the high-speed tensile test noise. The statistical results of YS, UTS, TEL and the product of UTS and TEL (PSE) of investigated steel at different strain rates are shown in Table 1. It can be observed from Fig. 3 and Table 1 that with the strain rate increasing from 0.01 to 200 s−1, the YS and UTS increased monotonously from 552 to 614 MPa and 815 to 860 MPa, respectively. However, the TEL (PSE) first decreased from 29.7 to 21.7% (24.2-17.9 GPa%) with the strain rate increasing from 0.01 to 10 s−1, and then increased to 29.8% (25.6 GPa%) when the strain rate was increased to 200 s−1.

Fig. 3
figure 3

The engineering stress–strain curve of the experimental steel after tensile deformation at different strain rates

Full size image
Table 1 Mechanical properties of the experimental steel at different strain rates
Full size table

3.2 Microstructure

Figure 4 shows SEM and TEM micrographs of experimental steel after two-stage heat treatment. As shown in Fig. 4(a), the microstructure of the experimental steel after two-stage heat treatment consisted of dark gray polygonal ferrite, light gray island bainite, and white granular retained austenite (RA). Some of the blocky RA were located at the ferrite grain boundary, while others were located inside the ferrite grain (Fig. 4b).

Fig. 4
figure 4

SEM and TEM micrographs of the experimental steel after two-stage heat treatment: (a) SEM micrograph; (b) TEM micrograph (F: Ferrite, B: Bainite, RA: Retained Austenite)

Full size image

Figure 5 shows the TEM micrographs of the experimental steel after tensile deformation at different strain rates. As shown in Fig. 5(a, c, e, g), a number of deformation twins in martensite were present in the experimental steel after tensile deformation at different strain rates, which implied that the TRIP effect occurred during the tensile deformation at different strain rates. Simultaneously, with the strain rate increasing from 0.01 to 1000 s−1, the dislocation density in ferrite was increased and the density of dislocation tangles and interaction became stronger (Fig. 5b, d, f, h).

Fig. 5
figure 5

TEM micrographs of the experimental steel after tensile deformation at different strain rates: (a, b) 0.01 s−1; (c, d) 10 s−1; (e, f) 100 s−1; (g, h) 200 s−1

Full size image

Figure 6 shows XRD patterns and volume fraction of RA in the experimental steel before and after deformation at different strain rates. It can be seen from Fig. 6(b) that the RA content in the experimental steel was 15.4% before deformation. After tensile deformation at strain rate of 0.01, 10, 100 and 200 s−1, it decreased to 5.2, 5.4, 5.7 and 6.2%, respectively. In other words, with the increase of strain rate from 0.01 to 200 s−1, the degree of transformation of RA in steel decreased from 10.2 to 9.2%. It can be deduced that the stability of RA increased with the increase of strain rate.

Fig. 6
figure 6

(a) XRD patterns and (b) measured retained austenite fraction of the experimental steel before and after tensile deformation at different strain rates

Full size image

The variable cross-section specimen tensile method was used to further study the transformation process of retained austenite to martensite in TRIP steel during dynamic deformation. Figure 7 shows the macrographs of the variable cross-section specimen before and after tensile deformation at strain rate of 100s−1. It can be seen from Fig. 7(b) that the TEL of the samples decreased when the opening depth of the variable cross-section specimen increased from 0.5 to 1.25 mm. The TEL of 0.5-sample, 0.75-sample, 1.0-sample and 1.25-sample was 18.1, 10.3, 8.7 and 5.2%, respectively. Figure 8 shows the EBSD maps of the experimental steel after tensile deformation at strain rate of 100 s−1 at various strain levels. The distribution of RA and α-bcc phases are marked in red and blue, respectively. The low angle grain boundaries between 2 and 15 deg are drawn by white lines, the high angle grain boundaries with more than 15 deg misorientation are drawn by black lines. As shown in Fig. 8(a, b), majority of the large blocky RA at the ferrite grain boundary transformed to martensite during the initial stage of dynamic deformation (ε ≤ 5.2%). With the increase of strain (5.2 < ε ≤ 10.3%), the RA with smaller average grain size transformed to martensite (Fig. 8c, d). Near the end of deformation, only a few small granular RA that were too stable did not transform (Fig. 8e, f).

Fig. 7
figure 7

Macrographs of samples with variable cross section before and after tensile deformation with strain rate of 100 s−1: (a) before tensile deformation; (b) after tensile deformation

Full size image
Fig. 8
figure 8

EBSD maps of the experimental steel after tensile deformation with strain rate of 100 s−1 at various strain levels: (a) ε = 0; (b) ε = 5.2%; (c) ε = 8.7%; (d) ε = 10.3%; (e) ε = 18.1%; (f) ε = 26.6%. (The RA are marked in red, the α-bcc phases are marked blue, the low angle grain boundaries are drawn by white lines, the high angle grain boundaries are drawn by black lines.) (Color figure online)

Full size image

Figure 9 shows XRD patterns and the statistical results of RA content in the experimental steel under different strain after quasi-static (strain rate of 0.01 s−1) and dynamic (strain rate of 100 s−1) tensile deformation. It can be seen from Fig. 9 that during the initial stage of tensile deformation (ε ≤ 5.2%), the transformation amount of RA was ~ 3.9% under dynamic tensile deformation, which was slightly less than ~ 4.5% under quasi-static tensile deformation. With the increase of strain (5.2 < ε ≤ 10.3%), the transformation amount of RA increased from ~ 3.9 to ~ 6.4% under dynamic tensile deformation, while it increased from 4.5 to 7.2% during quasi-static tensile deformation. After dynamic and quasi-static tensile deformation, the transformation amount of RA was ~ 9.7 and ~ 10.2%, respectively.

Fig. 9
figure 9

XRD patterns of the experimental steel under different strain (a) after quasi-static tensile deformation (strain rate of 0.01 s−1), (b) after dynamic tensile deformation (strain rate of 100 s−1) and (c) the statistical results of RA contents

Full size image

4 Discussion

4.1 Effect of Strain Rate on Microstructural Evolution

As shown in Fig. 5(b, d, f, h), a large number of dislocations were present in the ferrite grains of experimental steel after tensile deformation at different strain rates. However, the dislocation density in ferrite was relatively low at strain rate of 0.01 s−1 (quasi-static deformation) (Fig. 5b). With the increase of strain rate to 10-200 s−1 (dynamic deformation), the deformation speed was extremely fast, the proliferation of dislocations in a short deformation time made the dislocation slide difficult, resulting in significant increase in dislocation density in ferrite and the density of dislocation tangles and interaction was stronger (Fig. 5d, f, h). Simultaneously, it can be observed from Fig. 5(a, c, e, g) that RA transformed to martensite after tensile deformation at different strain rates. According to Figs. 8 and 9, the transformation amount of RA was larger during the initial stage of deformation, irrespective of whether the process was quasi-static or dynamic tensile deformation. This is because the large blocky RA with lower stability is prone to transform to martensite (Ref 21). It should be also noted that with the strain rate increasing from 0.01 to 200 s−1, the transformation of RA into martensite was suppressed gradually (Fig. 8) and the transformation rate of RA in quasi-static tensile deformation was higher than that in dynamic tensile deformation (Fig. 9), which is because of different stabilities of RA during quasi-static and dynamic tensile deformation. This is discussed in section 4.2.

4.2 Effect of Strain Rate on the Stability of RA

The thermodynamic stability of austenite and martensite is presented in Fig. 10 (Ref 22). \(M_{s}\) is the temperature, at which martensite starts to form at pre-existing nucleation sites during cooling, driven by the difference in the chemical free energy \(\Delta G_{{M_{s} }}^{\gamma \to \alpha \prime }\) of austenite and martensite (Ref 23). Assuming that the critical driving force needed to trigger martensitic transformation remained constant over the temperature range of interest, the martensitic transformation can also occur at temperature \(T_{1}\) (below \(T_{0}\) but above \(M_{s}\)) under the required additional mechanical driving force \(U^{\prime}\), if this additional driving force added to the chemical driving force \(\Delta G_{{T_{1} }}^{\gamma \to \alpha \prime }\) equals \(\Delta G_{{M_{s} }}^{\gamma \to \alpha \prime }\), i.e., \(U\prime + \Delta G_{{T_{1} }}^{\gamma \to \alpha \prime } = \Delta G_{Ms}^{\gamma \to \alpha \prime }\).

Fig. 10
figure 10

Free energy of austenite and martensite as a function of temperature

Full size image

For high-speed tensile tests (strain rate ranging from 10 to 200 s−1), the temperature of experimental sample was significantly increased due to adiabatic heating. According to our previous study (Ref 24) and Curtze et al. (Ref 25), the adiabatic temperature rise (ΔT) can be estimated using the following equation:

$$ \Delta T = \frac{\Delta Q}{{\rho C_{P} }} = \frac{\beta }{{\rho C_{P} }}\int\limits_{{\varepsilon_{1} }}^{{\varepsilon_{2} }} {\sigma {\text{d}}\varepsilon } $$

where ΔQ is the fraction of mechanical energy that is converted to heat energy, ρ = 7.8 gcm−3 is the density of steel, Cp = 0.46 kJ(kgK)−1 is the typical heat capacity of steel, β = 0.9 is a coefficient related to the fraction of mechanical energy that is converted to heat energy, σ is the true stress, and ε is the true strain. The total mechanical energy generated during tensile test was obtained by integrating the area under the σ-ε curve. Figure 11 shows the ΔT-true strain (ε) curve of the experimental steel deformed at the strain rate ranging from 10 to 200 s−1. Corresponding to strain rate of 10, 100 and 200 s−1, ΔT reached 41, 53 and 61 °C, respectively. As depicted in Fig. 10, the \(\Delta G_{{T_{1} }}^{\gamma \to \alpha \prime }\) for martensitic transformation decreased with the increase of ΔT, and consequently the required additional \(U^{\prime}\) increased. This is the reason why the RA stability increased with the increase of strain rate.

Fig. 11
figure 11

The estimated \(\Delta T\) as a function of true strain of the experimental steel deformed at different strain rates

Full size image

4.3 Effect of Strain Rate on Mechanical Properties

Considering Fig. 3 and Table 1, it can be seen that YS and UTS increased with the increase of strain rate from 0.01 to 200 s−1, which is in agreement with Choi et al’s (Ref 15), Van Slycken et al’s (Ref 16) and Wei et al’s (Ref 17) results. This is mainly because the dislocation slide becomes difficult and ferrite is strengthened at higher strain rate (Fig. 5). It may be noted that with the strain rate increasing from 0.01 to 200 s−1, UTS increases by 45 MPa, which is lower than the increase of YS (62 MPa). The possible reason is that the temperature was increased because of adiabatic heating during high-rate deformation, resulting in softening of the matrix (Fig. 11). However, with increase of strain rate from 0.01 to 10 s−1, the TEL decreased from 29.7 to 21.7%. This can be ascribed to two aspects. First, with increase of strain rate from 0.01 to 10 s−1, the amount of transformation of RA decreased insignificantly from 10.2 to 10%, meaning a weak TRIP effect (Fig. 6b). Second, according to He et al. (Ref 19) the significant multiplication of dislocations during the short deformation time made the dislocation slide difficult during dynamic tensile deformation, resulting in deformation localization and causing loss in TEL. As the strain rate was increased from 10 to 200 s−1, the TEL increased from 21.7 to 29.8%. The main reason is that the adiabatic heating started to play an important role in high strain rate. As shown in Fig. 11, with the strain rate increasing from 10 to 200 s−1, the ΔT increased from 41 to 61 °C, which caused the thermal softening and resulted in the increase of TEL.

5 Conclusions

In this study, the effect of strain rate on microstructural evolution and mechanical properties of the Fe-1.47Mn-1.40Si-0.21C-0.025Nb TRIP steel was elucidated. Additionally, an original tensile specimen with variable cross-section was adopted to realize the controllable strain experiment and study the dynamic deformation process in Fe-1.47Mn-1.40Si-0.21C-0.025Nb TRIP steel. The main conclusions are as follows:

  1. (1)

    Compared with quasi-static tensile deformation, the dislocation density in the experimental steel during dynamic tensile deformation was significantly higher, and the density of dislocation tangles and interaction was stronger. However, the transformation amount and the transformation rate of RA at dynamic tensile deformation was less than the quasi-static tensile deformation.

  2. (2)

    With the strain rate increasing from 0.01 to 200 s−1, the transformation amount of RA in the steel decreased from 10.2 to 9.2%, owing to the increased adiabatic temperature during dynamic tensile deformation, which led to the increase of required mechanical driving force for the transformation of RA.

  3. (3)

    The microstructure of experimental steel was mainly composed of polygonal ferrite, bainite and RA, and majority of large blocky RA at the ferrite grain boundary transformed to martensite during the initial stage of dynamic deformation. With the increase of strain, RA with smaller average grain size transformed to martensite. At the end of deformation, only a few small granular RA that were too stable did not transform.

  4. (4)

    With the strain rate increasing from 0.01 to 200 s−1, the YS and UTS of experimental steel increased monotonously from 552 to 614 MPa and 815 to 860 MPa, respectively. However, the TEL (PSE) first decreased from 29.7 to 21.7% (24.2-17.9 GPa%) with the strain rate increasing from 0.01 to 10 s−1, and then increased to 29.8% (25.6 GPa%) when the strain rate was increased to 200 s−1.