1 Introduction

The transformation-induced plasticity (TRIP) steels with excellent mechanical properties have been used for automotive applications for more than two decades [1, 2]. Conventional TRIP steels were developed based on the Fe–C–Mn–Si alloy system. Recently, some derivatives of conventional TRIP-aided steels in which Si is partially replaced by Al have been the subject of attention as a means to improve galvanizability [3]. Like Si, Al also prevents the formation of carbides, leading to a higher enrichment of C in retained austenite (RA) [4, 5]. In order to obtain high strength TRIP steels, many investigations have focused on the role of Nb [6, 7]. NbC precipitates are formed during hot-rolling processes, which reduce the size of the recrystallized grains by pinning the grain boundaries and refine the final microstructure, resulting in an increase in yield strength and tensile strength. In order to improve both strength and ductility, Nb/V [8, 9] and Nb/Mo [10, 11] complex microalloying for hot- or cold-rolled TRIP steels were also investigated.

The alloying elements in TRIP steels play important roles in determining the stability of retained austenite and final microstructures. It is a challenging task to distinguish elemental compositions of different phases and the effect on the stability of retained austenite. Recently, Caballero et al. [12] investigated the partitioning of the redistribution of carbon by atom probe tomography (APT) after tempering of a novel nanocrystalline bainitic steel. These results elucidated that carbon partitioning from supersaturated ferrite to austenite occurs during tempering. Pereloma and Russell [13] observed the composition, sizes, and distribution of C-containing clusters and fine particles in thermomechanically processed CMnSi TRIP steels, alloyed with Nb, Mo and Al, with 3-Dimensional atom probe (3-DAP). Scott and Drillet [14] also investigated the carbon distribution in retained austenite. The values of carbon concentrations in retained austenite were directly obtained by electron energy loss spectroscopy (EELS) in transmission electron microscope. Lee et al. [15] studied the partitioning of Mn to austenite during intercritical annealing (IA) in a medium Mn TRIP steel by means of X-ray diffraction (XRD) and energy-dispersive X-ray spectroscopy (EDS) analysis in transmission electron microscope (TEM) foils. They reported that the partitioning of Mn and the size effect of ultrafine austenite grain during intercritical annealing are the two main contributions to the austenite stability.

It is well known that the C concentration in TRIP steels is the important factor that governs the retained austenite mechanical stability [16]. However, it is of difficulty in analyzing C quantitatively because of the limitation in detecting light elements. Wavelength dispersive electron probe microanalysis (EPMA) is preferred over other techniques to measure carbon concentration in steels such as atom probe tomography, energy loss spectroscopy in transmission electron microscope, energy dispersive spectroscopy, owing to its good sensitivity to detect low carbon concentration. Between mesoscopic and nanoscopic techniques, EPMA occupies an intermediate position where micrometer to nanometer structures can be characterized [17].

In the present work, the investigated steel was processed by annealing at predetermined intercritical annealing (IA) temperature with different times. The partitioning of C and the microstructure evolution during heat treatment were examined by EPMA, SEM, and XRD. The mechanical properties were measured by tensile test. The work hardening of the samples was also evaluated by calculating the instantaneous n value as a function of strain to compare retained austenite stability in different heat-treated samples. The objective of this study is to explore the IA time effects on the microstructure and mechanical properties of investigated steel. Special interest will be focused on the selection of the heat treatment parameters to produce high strength and high ductility steel.

2 Experimental

The steel used in this study has a nominal composition of Fe–0.19C–1.49Mn–0.7Si–0.94Al–0.19Mo–0.058Nb (wt%). In this study, the experimental steel was prepared as a vacuum-melted 40 kg ingot. The ingot was reheated to 1,200 °C and hot-rolled to a thickness of 4 mm, finishing temperature being 860 °C and then coiled at 600 °C. After cold rolling to a thickness of 1.2 mm, a two-stage heat treatment was adopted. In this experiment, the IA time was changed from 1 to 30 min to investigate its effect on the microstructures and the partitioning of alloying elements. The strip was first annealed at 830 °C for different times to form a mixture of ferrite and austenite and then austempered at 440 °C for 5 min to stabilize austenite in a salt bath. The microstructure of the TRIP steels was analyzed by scanning electron microscopy (SEM) and TEM. EPMA was used in order to build the chemical maps of the elements. Elemental beam scans and mappings were acquired on a field-emission gun electron microprobe, JEOL JXA-8530F, using an accelerating voltage of 20 kV and a beam current of 0.1 μA. Beam scans were performed at a fixed position and magnification (8,000×) by moving the electron beam. The detection time for each point was 2 s in the single-scan mode and 100 scans with step of 50 ms in the multiple-scan mode. Tensile specimens were prepared parallel to the rolling axis with a cross-section of 1 mm × 10 mm and a gauge length of 25 mm. Uniaxial tensile tests were performed on an INSTRON 5,967 30 kN machine at a constant strain rate of 1.0 × 10−3 s−1 at room temperature. The volume fraction of austenite was determined by XRD with CuK α radiation using the direct comparison method. This method uses the integrated intensities of the (200)α and (211)α peaks and the (220)γ and (311)γ peaks [18]. The average carbon concentrate of RA was calculated from the integrated intensity of austenite peaks using the following equation: a γ = 3.578 + 0.033C γ [19], where a γ is lattice parameter of the austenite and C γ is the carbon concentrate of RA.

3 Results and Discussion

The cold-rolled and heat-treated microstructures are presented in Fig. 1, respectively. The cold-rolled sample has a mixed structure of ferrite (F), bainite (B) and pearlite (P) elongated along the rolling direction, showing a band structure. After two-stage heat-treatment, the constituents of the samples are ferrite, granular bainite, retained austenite/martensite (M/A) islands, and retained austenite (RA), shown in Fig. 1. In the SEM micrographs, ferrite appears dark and M/A islands and RA appear light. Bainite appears as granular type. As observed in Fig. 1, the volume fraction, morphology, and distribution of each phase structure are different, depending on the time of IA treatment. With the increasing IA time, the fraction of ferrite and bainite of the final microstructure increased. The longer time also causes austenite grain coarsening during IA, which is the reason for the increase of the average ferrite and granular bainite grain size. Polygonal ferrite, granular bainite, and RA can be found in the microstructure of in Fig. 1b. When the annealing time increases to 30 min, the microstructure exhibits mostly ferrite, granular bainite, and M/A island, less granular RA exists (Fig. 1e). The M/A islands are bigger in Fig. 1e than in Fig. 1b. Figure 2 shows the XRD results of the samples intercritically annealed at 830 °C for different holding times and isothermally heat treated at 440 °C for 5 min. Diffraction peaks from the austenite phase are clearly visible in the sample annealed for 1 min. However, only minor austenite peaks were observed for the sample annealed for 30 min. The volume fractions and the average carbon contents of the RA as a function of intercritical annealing time are shown in Table 1. It can be seen the volume fraction of RA as estimated from the XRD decreased from 13.8% to less than 5%, with the annealing time increasing from 1 to 30 min. Carbon enrichment is maximized when the sample is annealed for 1 min, indicating that the retained austenite in this sample has the highest mechanical stability. However, increasing annealing time further results in a decrease of carbon content.

Fig. 1
figure 1

SEM images of the cold-rolled sample and the heat-treated sample with different annealing times (F ferrite, B bainite, P pearlite, RA retained austenite, M/A M/A island): a cold-rolled; b 830 °C/1 min + 440 °C/5 min; c 830 °C/3 min + 440 °C/5 min; d 830 °C/15 min + 440 °C/5 min; e 830 °C/30 min + 440 °C/5 min

Full size image
Fig. 2
figure 2

XRD patterns of the samples after intercritically annealed at 830 °C for different holding times and then isothermally heat treated at 440 °C for 5 min prior to tensile deformation

Full size image
Table 1 Volume fraction of RA and the average carbon content in RA for samples after annealing at 830 °C for different times and isothermally heat treated at 440 °C for 5 min
Full size table

Alloying element concentration mappings were performed with EPMA on samples taken from three different samples, i.e., after cold rolling, after intercritical annealing at 830 °C for 1 min and 30 min and then austempered at 440 °C for 5 min, as shown in Fig. 3. It can be seen that long-range carbon segregation is more accentuated in the cold-rolled sample in comparison with the other two samples. This is because the areas where carbon is segregated are not completely homogenized before hot rolling, due to the reduced diffusion rate of carbon by the presence of manganese. The microstructure of the cold-rolled sample inherited from the hot rolled one. In the EPMA C elemental mapping image, retained austenite is highlighted by its specific high C concentration because C solubility is relatively high in fcc austenite than in bcc ferrite matrix. It was clear that the C atoms in the investigated steel unevenly distributed in different phases in Fig. 3. When kept at 830 °C for 1 min, carbon mainly existed among the granular retained austenite and small M/A islands. The carbon concentration in the retained austenite grains was higher than the one in other phases. The amount of carbon atoms partitioned in ferrite grains can be ignored, as the carbon content is quickly reduced to a very small value in this phase during intercritical annealling [20]. For annealing time of 30 min, most of the carbon atoms enriched in the larger M/A islands. It also can be found in Fig. 3c that carbon is rich in granular bainite. This result implies that annealing for a relatively long time results in a decrease in the C content of the final retained austenite.

Fig. 3
figure 3

EPMA images of cold-rolled sample and heat-treated sample with different annealing times (F ferrite, B bainite, P pearlite, RA retained austenite, M/A M/A island): a cold-rolled sample, b 830 °C/1 min + 440 °C/5 min, c 830 °C/30 min + 440 °C/5 min

Full size image

From these results, it could be deduced that carbon partitioning at IA is the major factor in determining the volume fraction of RA in the investigated samples. According to Ref. [21], the austenite formation during IA can be divided into three stages: pearlite dissolution, austenite growth with carbon diffusion, and elemental partitioning between austenite and ferrite. Pearlite dissolution finishes fairly quickly and thus can not be the major reason for the observed trend in austenite volume change. Hence, the final microstructure of the investigated steel should originate from the way of austenite growth and the occurrence of carbon partitioning during IA. When the samples are isothermally treated at the two phase region, the pearlite is in an unstable state and the ferrite recrystallization, cementite decomposition and austenite formation would occur [22]. The non-uniform distribution of carbon and other elements results in the concentration and structure fluctuations which is favorable to the austenite nucleation. Therefore, a number of fine austenite grains with high carbon content form in carbon-rich areas. At 830 °C, carbon diffuses much faster than Si and Mn, both in ferrite and austenite, indicating that carbon diffusion is dominant in the early stage of the IA process [23]. Carbon seems to have partitioned to austenite during IA despite the short annealing time of 1 min. These austenite grains are formed during the initial stage of IA and kept after isothermal bainite treatment (IBT). There, the sample annealed at 830 °C for 1 min obtains the highest fraction of retained austenite, as shown in Fig. 2 and Table 1. From Table 1, it also can be seen that the volume fractions and average carbon content of RA decrease remarkably when the annealing time increased. With IA time increasing, the more pearlite decomposed and more austenite form. This results in the decrease of ferrite content and increase of austenite content before IBT process. The austenitic phase fraction rises when the holding time increases from 1 to 30 min during intercritical annealing, and the average carbon content of the austenite should decrease if no cementite is formed [24, 25]. It is well known that carbon is the alloying element which increases austenite stability most effectively. When the carbon content in the retained austenite is less, the stability of the retained austenite is also reduced. When the samples are heated during IBT process, the unstable austenite transforms into granular bainite and M/A islands, which results in a lower retained austenite volume fraction. Since carbon atoms still can diffuse as interstitial atoms in both ferrite and austenite at 440 °C, the partitioning of carbon from ferrite into M/A island should continue. More carbon atoms are ejected into the M/A island or film-shaped RA during IBT process. Thus, the large M/A islands which is rich in carbon can be seen in Fig. 3c.

The distribution of two other main alloying elements, Nb and Mo, are also shown in Fig. 3. In the polygonal ferrite and bainitic ferrite of the Nb/Mo steel, a significant number of Nb-rich, Mo-rich precipitates of varying compositions was observed both after cold rolling and after intercritical annealing (Fig. 3). A large amount of NbC and (Nb, Mo)C precipitates distribute evenly in ferrite grains in Fig. 4a. In precipitated form as carbides or carbonitrides, Nb has a grain-refining effect, increasing the strength. In Fig. 4a, the big particles are probably NbC. The NbC carbides mainly precipitate from austenite during hot rolling and from ferrite during IA which due to the lower solubility of carbon in ferrite than that in austenite [7]. In the microstructure of the sample austempered for 30 min in Fig. 4b, some coarse particles are clearly visible, whereas not are observed in the sample annealed for 1 min. The NbC was precipitated during long time IA [10]. The other precipitation in Fig. 4a is (Nb, Mo)C carbide which only precipitate from ferrite during cooling from the intercritical annealing temperature to room temperature due to their much lower precipitation temperature than that of NbC, and thus their particles are always fine [26]. For the Nb/Mo TRIP steel, Mo and Nb are expected to precipitate as (Nb, Mo)C carbides in ferrite grains for precipitation strengthening.

Fig. 4
figure 4

EDX analysis results of precipitates in the heat-treated sample with different annealing time: a 830 °C/1 min + 440 °C/5 min; b 830 °C/30 min + 440 °C/5 min

Full size image

Engineering stress–strain curves of all samples processed under different heat-treatment conditions are given in Fig. 5a. From Fig. 5a it can be seen that the stress–strain curves of the samples annealed at 830 °C vary with annealing time significantly, in which the strength and elongation significantly decreased with increased annealing time. The yield stress (YS), tensile strength (TS), total elongation (TEL), and product of tensile strength and total elongation (TS × TEL) are shown in Table 2. The mechanical properties vary significantly with the annealing time. In the Table 2, the YS and TEL drop from 591 MPa and 21.3% down to 494 MPa and 18.5%, respectively, with increasing the annealing time, while the TS decreases from 1,103 to 871 MPa. The longer the IA holding time, the more the TS decreases. If the results from Fig. 1, Tables 1 and 2 were compared, it would be observed that the strength and elongation assumes a different evolution with changing of annealing time. The TS and TEL of the samples is affected by the volume fraction of retained austenite and the grain size. It was found that the volume fraction and carbon content of RA decreases with the increasing of annealing time. Meanwhile, longer annealing time causes grain coarsened, which was another reason for the decrease of the strength. Therefore, the YS and TS of the steel linearly decrease with increasing of grain size and decreasing RA fraction. It can also be seen from Tables 1 and 2 that the TEL gradually decreases with increasing annealing time. This result stipulates the idea that the RA fraction is a key factor in order to get good ductility. The TS × TEL which might be used for evaluating formability and strength were measured as over 23,000 MPa % in the samples annealing time is 1 min, and below 21,000 MPa % in other samples. It was found that TS × TEL also linearly decreased with increasing annealing time. This relationship between TS × TEL and annealing time is not new, but it could be help us to understand why the excellent product of strength and ductility could be obtained in the annealed sample with a relatively short time. The higher the austenite fraction and smaller grain size, the higher the TS × TEL. This result indicates that the excellent combination of strength and ductility could be obtained in a short time annealing when it was first annealed in the intercritical region.

Fig. 5
figure 5

a Engineering stress–strain curves of the samples after intercritically annealed at 830 °C for different holding times and then isothermally heat treated at 440 °C for 5 min; b the instantaneous strain hardening exponent exponent (n) versus true strain

Full size image
Table 2 Mechanical properties of the samples after annealing at 830 °C for different times and isothermally heat treated at 440 °C for 5 min
Full size table

Work hardening behavior is known to be related to the strain-induced transformation of austenite into martensite (TRIP effect). Figure 5b shows the instantaneous strain hardening exponent as a function of true strain of different samples. The instantaneous strain hardening exponent (n) is calculated as n = dlnσ/dlnε [27]. The instantaneous n value has been widely used as an indication of formability of sheet materials [28]. Materials with good formability will exhibit higher n values that maintain to higher levels of strain. Figure 5b shows that the 1-min sample has higher n values than other samples, as well as a higher maximum n value. While other samples exhibit a highest work hardening rate at very low strain of 0.023, the 1-min sample demonstrates highest hardening rate at tensile strain above 0.035, and its work hardening behavior sustains at larger strains where work hardening is critical for suppressing localization (e.g., necking) and failure. The observed work hardening behavior can be explained by the good TRIP effect. The highest fraction retained austenite with 1.21 wt% carbon content was obtained in 1-min samples. The strain-induced transformation to martensite occurs gradually at a more uniform rate when 1-min sample shows high work hardening values at high levels of strain. As a result, excellent mechanical properties are obtained in the 1-min sample. Therefore, the stability of RA appears to play a significant role in final mechanical properties of the investigated steel. In this work, IA time which is less than 1 min wasn’t studied and that will be analyzed in future work.

4 Conclusions

  1. (1)

    The prolonged intercritical annealing increased in the fraction of the austenite phase and improved the carbon distribution uniformity in austenite, which result in significantly decreased stability of the RA. Therefore, the volume fractions and the carbon contents of RA decreased with IA time increasing from 1 to 30 min.

  2. (2)

    Significant inhomogeneous distribution of C was observed in the cold-rolled sample prior to annealing, and the enrichment of C in the retained austenite during the annealing was clearly observed. The RA of the 1-min sample showed enough stability to be retained at room temperature. C atoms existed among the granular retained austenite and small M/A islands. In contrast, the austenite stability of the sample annealed for 30 min became too low and the most of C atoms partitioned in the large M/A islands.

  3. (3)

    In the investigated steel, Mo and Nb precipitated as (Nb, Mo)C carbides in ferrite grains. The formation of fine and dispersive carbides in ferrite grains plays an important role in the increase of strength of samples.

  4. (4)

    According to the present experiment, the strength and elongation of the samples decrease, with increasing IA time. The mechanical properties of the sample intercritically annealed at 830 °C for 1 min and isothermally held at 440 °C for 5 min were better than other samples. This sample also maintains a relatively high rate of work hardening at larger strains due to the TRIP effect. The optimal mechanical properties are tensile strength of 1,103 MPa, total elongation of 21.3%, and strength-ductility product of 23,493.9 MPa %. The conclusion of the paper has limitations, and the sample which is annealed at 830 °C for less than 1 min will be studied in future work.