Microstructure and tensile behavior of friction-stir welded TRIP steel

Microstructure and tensile behavior of friction-stir welded TRIP steel

Materials Science & Engineering A 717 (2018) 26–33 Contents lists available at ScienceDirect Materials Science & Engineering A journal homepage: www...

3MB Sizes 2 Downloads 121 Views

Materials Science & Engineering A 717 (2018) 26–33

Contents lists available at ScienceDirect

Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea

Microstructure and tensile behavior of friction-stir welded TRIP steel a,⁎

a

b

S. Mironov , Y.S. Sato , S. Yoneyama , H. Kokawa

a,d

a

, H.T. Fujii , S. Hirano

T

c

a

Department of Materials Processing, Graduate School of Engineering, Tohoku University, 6-6-02 Aramaki-aza-Aoba, Aoba-Ku, Sendai 980-8579, Japan Department of Mechanical Engineering, Aoyama Gakuin University, 5-10-1 Fuchinobe, Sagamihara 252-5258, Japan Hitachi Research Laboratory, Hitachi Ltd, 7-1-1 Omika-cho, Hitachi 319-1291, Japan d School of Materials Science and Engineering, Shanghai Jiao Tong University, 800 Dongchuan Road, Minhang District, Shanghai 200240, P.R. China b c

A R T I C L E I N F O

A B S T R A C T

Keywords: Iron alloys Plasticity methods Phase transformations Characterization Electron microscopy Stress/strain measurements

In this work, electron backscatter diffraction and digital image correlation techniques were employed to study the microstructure and microstructure-property relationship in friction-stir welded TRIP steel. It was found that the thermal effect of the welding process led to material softening in the heat-affected zone and promoted martensite transformation in the stir zone. These microstructural changes provided rapid strain localization during subsequent transverse tensile tests and thus resulted in premature failure of the welds. Material softening in the heat-affected zone was deduced to be a combined result of dissolution and spheroidisation of retained austenite as well as recovery in bainitic ferrite. The stir zone martensite was characterized by significant orientation spread and therefore its orientation relationship with austenite essentially deviated from the ideal Kurdjumov-Sachs relation. Moreover, the martensite transformation was found to be influenced by variant selection.

1. Introduction Transformation-induced plasticity (TRIP) steels belong to a relatively new generation of materials developed for automotive applications with the aims of weight reduction and fuel savings without compromising passenger safety, at no increased cost. The principal characteristic of such steels is the significant fraction of retained austenite. It is well known that austenite may transform to martensite upon cold deformation, thus giving rise to significant work hardening effect [1–4]. This provides a good combination of strength and ductility and hence excellent energy absorption capacity, which is required during a sudden vehicle crash. Unfortunately, the unique microstructure of the TRIP steels is totally destroyed during conventional fusion welding and the superior properties of these materials are therefore lost. In this context, friction-stir welding (FSW), an innovative solid-state process [5], has recently attracted attention as a possible candidate for joining such materials. Considering the potential advantages of this technique, several works have been undertaken recently to evaluate the effects of FSW on microstructure and mechanical properties of TRIP steels [6–9]. It was found that FSW results in complex microstructural changes. The heat input associated with the welding process induces the austenite-tobainite phase transformation in the heat-affected zone [6,9], leading to concomitant material softening [8]. Approaching the stir zone,



however, the welding temperature exceeds the A1 point, thus initiating a sequence of phase transformations that eventually lead to the formation of martensite [9]. In the stir zone, the peak temperature is believed to surpass the A3 point [8,9]. Accordingly, martensite becomes the dominant phase in this region [7–9]. However, the stir zone may also contain minor fractions of ferrite [6,8,9], bainite, and retained austenite [6–8]. These observations are explained in terms of the relatively low cooling rate during the weld thermal cycle [8,9]. The above works have significantly contributed to our current understanding of the effects of FSW on the microstructure and properties of TRIP steels. Nevertheless, several important aspects of microstructure evolution, including microstructural changes in the retained austenite and bainite as well as mechanisms of the phase transformations, remains poorly understood. This obstructs an establishment of microstructure-strength relationship. Attempting to provide deeper insight into these issues, electron backscatter diffraction (EBSD) and digital-image correlation (DIC) techniques are employed in the current study. 2. Material and experimental procedures The material used in the present investigation was 1.2 GPa TRIP steel. The material had the nominal chemical composition listed in Table 1 and was supplied as 1.4 mm thick uncoated sheets. Thermo-

Corresponding author. E-mail address: [email protected] (S. Mironov).

https://doi.org/10.1016/j.msea.2018.01.053 Received 12 September 2017; Received in revised form 12 January 2018; Accepted 12 January 2018 Available online 17 January 2018 0921-5093/ © 2018 Elsevier B.V. All rights reserved.

Materials Science & Engineering A 717 (2018) 26–33

S. Mironov et al.

specimens were machined. The specimens were centered at the weld line, had a gauge section 35 mm in length and 7 mm in width, and included all of the characteristic microstructural zones developed during FSW. Other details of the specimens’ geometry are given in Supplementary Fig. S3. For comparative purposes, appropriate tensile specimens were also machined from the base material. Tensile tests

Table 1 The nominal chemical composition (wt%) of studied TRIP steel. C

Mn

Si

Al

P

S

N

Fe

0.40

2.00

1.50

0.04

0.01

0.002

0.004

Balance

Fig. 1. Low-magnification optical image of the weld cross-section. RS and AS are the retreating side and advancing side, respectively. TD, ND, and WD are the transverse direction, normal direction, and welding direction, respectively.

to failure were conducted at ambient temperature and constant crosshead velocity corresponding to a nominal strain rate of 10−3 s−1 using an Instron 5969 testing machine. Five tensile specimens were tested for each material condition. The DIC technique [10,11] was employed to evaluate strain distributions generated during the tensile tests. To this end, a random ink pattern was applied to the sample surface and a high-speed XiQMQ042MG-CM digital camera equipped with a Nikon AI AF MicroNikkor 200 mm f/4D IF-ED lens was used for image recording. The images were taken at a rate of 1 image per second (~ 0.1% of tensile strain). The samples for the strain distribution measurements were mechanically polished prior to the tensile tests to achieve a uniform thickness with the final polishing step comprising of 1 µm diamond.

Calc calculations showed that the phase transformation temperatures A1 and A3 in this material were 950 K and 1060 K, respectively. The asreceived material was friction-stir welded in a bead-on-plate configuration at a tool travel speed of 60 mm/min. In an attempt to provide the lowest possible welding temperature, a relatively low tool rotational speed of 150 rpm was applied. To minimize surface oxidation, argon shielding was employed around the tool during FSW. The welding tool was fabricated from a Co-based alloy and consisted of a convex shoulder 12 mm in diameter and a 1.2 mm long probe. The probe was tapered from 5.0 mm at the shoulder to 3.0 mm at the probe tip. Further details of the tool design are shown in supplementary Fig. S1. The principal directions of the welding geometry are denoted throughout as the welding direction (WD), transverse direction (TD), and normal direction (ND). Microstructure characterization was performed on the transverse cross-section of the weld (TD × ND plane) using optical microscopy, scanning electron microscopy (SEM), and EBSD. For optical, and SEM observations, the samples were ground with water abrasive papers, mechanically polished with 1 µm diamond finish, and chemical etched with 3% Nital. A suitable surface finish for EBSD was obtained by mechanical polishing in a similar fashion, followed by electro-polishing in 95% acetic acid + 5% perchloric acid solution at approximately 10°C–15 °C with applied potential of 40 V. To view the microstructure distribution in the weld more broadly, the microhardness profile was measured across characteristic microstructural zones. Vickers microhardness data were obtained by applying a load of 1 kg with a dwell time of 10 s. To assist with interpretation of the evolved microstructures, a fullymartensite structure was produced. To this end, the as-received (base) material was heated to 900 °C, soaked for 15 min, and then water quenched. Typical microstructure and microhardness distribution are shown in supplementary Fig. S2. SEM and EBSD analyses were conducted using a Hitachi S-4300SE field-emission gun scanning electron microscope (FEG-SEM) equipped with a TSL OIM™ EBSD system and operated at accelerating voltage of 25 kV. Orientation mapping was performed using a triangular scanning grid. Depending on the particular purpose, different scan step sizes were used for mapping. Low-resolution (overview) EBSD maps were acquired using a scan step size of either 1 or 0.5 µm, whereas higher resolution maps were obtained using a scan step size of 0.2 µm. On each diffraction pattern, seven Kikuchi bands were used for indexing to minimize the possibility of misindexing errors. To ensure reliability of the EBSD data, all small grains comprising three or fewer pixels were automatically “cleaned” from the EBSD maps using the grain-dilation option of the TSL software. Furthermore, to eliminate spurious boundaries caused by orientation noise, a lower-limit boundary misorientation cut-off of 2° was used. A 15° criterion was applied to differentiate low-angle boundaries (LABs) and high-angle boundaries (HABs). To examine mechanical behavior of the welds, transverse tensile

3. Results 3.1. Microstructural observations A low-magnification optical image of the weld cross-section is shown in Fig. 1. No volumetric defects were found but the joint exhibited evidence of bending, presumably reflecting substantial residual stresses. The optical contrast across the weld was obviously inhomogeneous, indicating complex microstructural changes induced by FSW. Five microstructural zones were defined, i.e., base material, zone 1, zone 2, stir zone, and a characteristic band region on the retreating side of the stir zone. Typical microstructures observed in these zones are summarized in Fig. 2. The brighter phase in these secondary-electron SEM images is austenite, martensite, or cementite, whereas the darker phase is ferrite (including bainitic ferrite).1 To assist with interpretation of the microstructures, the microhardness profile was measured across these zones and the results obtained are shown in Fig. 3a. The base material exhibited a typical ausformed microstructure with clear prior-austenite (or packet-) boundaries as well as sandwiched austenite and bainite laths arranged in blocks (Fig. 2a). The volume fraction of retained austenite was measured to be ~ 20% (Supplementary Fig. S4). In zone 1, the austenite fraction was found to be reduced (Fig. 2b). Along with precipitation of fine cementite particles (an example is circled in Fig. 2b), this observation probably implies transformation of retained austenite to lower bainite, thus being consistent with literature data [6,9]. Moreover, nearly-equiaxed austenite particles also appeared (an example is arrowed in Fig. 2b), thus presumably reflecting a 1 The secondary-electron contrast in SEM images is well accepted to depend on electronic structure of an irradiated material as well as from the so-called edge effect. The latter one implies that any changes in surface topography enhance emission of the secondary electrons and thus result in brighter contrast. Due to relatively high density of defects, the retained austenite and the martensite typically have more developed topography in an etched surface that the ferrite and thus they appear relatively bright.

27

Materials Science & Engineering A 717 (2018) 26–33

S. Mironov et al.

Fig. 2. Typical SEM micrographs taken from base material (a), zone 1 (b), zone 2 (c), central section of the stir zone (d), and band region (e). Note: The reference frame and scale bar for all micrographs are given in the bottom right corner. In (b), selected area exemplifies precipitation of cementite whereas arrow indicates an example of nearly-equiaxed austenite particle.

Fig. 3. Microhardness profile measured across the weld (a) and typical deformation diagrams of base material and friction-stir welds (b). In (a), different microstructural zones are indicated and the microhardness of the fully-martensite structure is shown (broken line).

As follows from Fig. 3a, the microhardness substantially varied across the stir zone, thus probably mirroring the variation of fractions of martensite and ferrite. This may indicate inhomogeneous character of temperature and/or cooling rate distributions across the stir zone, as is often found during FSW [5]. Of particular interest, however, was the band region which characterized by an abnormally high proportion of ferrite (Fig. 2e) and, accordingly, anomalously low strength (Fig. 3a).

spheroidization occurring in this phase. The above microstructural changes agree with material softening observed in this area (Fig. 3a). Remarkably, no clear evidences of strain were found and therefore this microstructural region was suggested to be the heat-affected zone. The possible mechanisms for material softening are discussed in more detail in Section 4.1. Martensite formation was found in zone 2 (Fig. 2c), which resulted in abrupt material hardening (Fig. 3a). Again, this is in good agreement with scientific literature [9] and most likely indicates that the welding temperature in this region surpassed the A1 point and the material experienced ferrite-to-austenite-to-martensite transformation sequence during the weld thermal cycle. In the stir zone, the microstructure was dominated by martensite but contained a minor fraction of ferrite (Figs. 2d and 3a), thereby also being in excellent agreement with previous works [6–9]. Remarkably, the prior-austenite grains in the stir zone appeared to be finer than those in zone 1 (compare Fig. 2d and c), thus implying possible grain refinement during FSW.

3.2. Tensile behavior Typical deformation diagrams of the base and welded materials are shown in Fig. 3b. Both material conditions demonstrated nearly identical yield strengths of ~ 950 MPa, i.e., the joint efficiency for the yield strength was close to 100%. Meanwhile, the welds showed significant reduction in ductility. Specifically, their elongation-to-failure was found to scatter from 3.7% to 5.3%, thus being substantially lower than that of the base material (Fig. 3b). Accordingly, the ultimate tensile strength of the welded material was also lower (1100 MPa) than that of 28

Materials Science & Engineering A 717 (2018) 26–33

S. Mironov et al.

Fig. 4. Composite optical image showing typical deformation relief observed in failed tensile specimens machined from frictionstir weld. RS and AS are the retreating side and advancing side, respectively. Note: Tension direction is horizontal.

the parent material (1200 MPa), as shown in Fig. 3b. As discussed in the previous section, the welded material was characterized by a considerable microstructural gradient from base material through the heat-affected zone to the stir zone. Hence, the ductility degradation may be a result of strain localization. To examine this idea, deformation relief developed during the tensile tests was analyzed. To this end, several specimens were mechanically polished to a mirror finish prior to the tensile tests. The typical deformation relief is shown in Fig. 4. It is clear from the figure that strain distribution across the weld was, indeed, fairly inhomogeneous. In all cases, the stir zone exhibited no distinct evidences of strain whereas clear strain localization (and subsequent failure) occurred in the heat-affected zone. Attempting to provide more fundamental insight into this phenomenon, the strain distribution across the weld was measured using the DIC technique.

This phenomenon was suggested to be associated with strain hardening effect which resulted in gradual increase of tensile stress in the heataffected zone and its final equilibration with yield stress in the basematerial zone. Hereafter, the tensile load was mainly sustained by the base material. Noteworthy, the strain distribution was still asymmetric, being preferentially concentrated on the retreating side of the weld (Fig. 5b). The evolved strain distribution did not change much during almost the entire range of the tensile test (compare Fig. 5b and c). Shortly before fracture, however, the strain was again observed to concentrate in the heat-affected zone (Fig. 5b) and this finally led to failure in this microstructural region (Fig. 5e). The possible reason for this behavior is discussed in Section 4.2. As expected, the stir zone experienced almost no plastic strain (Fig. 5).

3.3. Strain distribution measurements

4. Discussion

The distributions of extensional strains measured in the upper weld surface as a function of accumulated tensile strain are summarized in Fig. 5. The distributions of the width- and shear strains are provided in Supplementary Figs. S5 and S6. During elastic loading, distinct strain concentration was found in the heat-affected zone (arrows in Fig. 5a). This effect was presumably attributable to material softening in this microstructural region (Fig. 3a). Remarkably, the plastic strain was more pronounced on the retreating side of the weld (Fig. 5a). Nearly before yielding, however, the plastic strain also encompassed the base material region (Fig. 5b).

4.1. Material softening in heat-affected zone As shown above, the mechanical behavior of the welded material was essentially influenced by the heat-affected zone. Therefore, the material softening observed in this microstructural region deserves particular attention. To examine the possible mechanism(s) of this phenomenon, several EBSD maps were taken across this area, as shown in Fig. 6. The low-resolution (overview) EBSD map in Fig. 6a was used for examination of general trends of microstructure evolution whereas high-resolution maps in Fig. 6b-d were employed for quantitative

Fig. 5. The distribution of true extensional strain in the upper surface of weld after (a) elastic loading, and global plastic strain of (b) 0.2%, (c) 1.5%, (d) 3.2% (i.e., peak load), and (e) 3.6% (i.e., immediately before fracture). The scale bar and reference frame is given in the bottom right corner of (e). AS and RS are the advancing side and retreating side, respectively. Tension direction is vertical. In (a), arrows point to strain initiation in the heat-affected zone (see Section 3.3 for details). In (e), selected area shows the region used for microstructural analysis of the TRIP effect (see Section 4.2 for details).

29

Materials Science & Engineering A 717 (2018) 26–33

S. Mironov et al.

Fig. 6. Composite EBSD phase map showing evolution of austenite phase in the heat-affected zone (a) with high-resolution images given at higher magnification in (b) - (d). In (b)-(d), low- and high-angle boundaries are depicted as blue and black lines, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article).

Fig. 7. Microstructural changes in zone 1: dissolution (a) and spheroidization (b) of retained austenite, as well as recovery in bainitic ferrite (c). See Section 4.1 for details.

phase is characterized by relatively high dislocation density, and therefore it may experience restoration processes. Since Fig. 6b-d show no clear evidences of recrystallization in the bainitic ferrite, the possible restoration mechanism may be a recovery. To examine the reliability of this idea, the evolution of grain orientation spread in this phase was measured using the standard kernel-average-misorientation (KAM) option of the EBSD software. This option is often employed to evaluate small orientation fluctuations within grains and involves the measurement of average misorientation between a pixel at the center of the kernel and all neighboring pixels. The obtained results are summarized in Fig. 7c. The clear decrease of the grain orientation spread in the bainitic ferrite seen in this figure presumably reflects recovery occurring in this phase. It may be concluded therefore that material softening in the heataffected zone was a combined result of dissolution and spheroidization of the retained austenite, as well as recovery occurring in bainitic ferrite. It is interesting to note that austenite was found to re-precipitate in zone 2 (Fig. 6a and d). Together with martensite formation in this region (Fig. 2c), this evidenced that the peak temperature there exceeded the A1 point.

characterization of this process, as shown in Fig. 7. The statistics of the high-resolution EBSD measurements is summarized in supplementary Table S1. It is clear from Figs. 6a-c and 7a that the fraction of retained austenite reduced drastically in zone 1. Due to relatively high carbon content, austenite is believed to be harder than ferrite and/or bainite, and therefore its dissolution may explain material softening in this area (Fig. 3a). As suggested in Section 1, the morphology of the survived austenite particles has also changed. To quantify this effect, aspect ratios of the austenite grains were calculated2 using EBSD software and characteristic distributions are shown in Fig. 7b. It is evident that lamellar-type austenite particles at the outer edge of the heat-affected zone transformed into nearly equiaxed particles at its inner border thus presumably indicating a spheroidization process. This should additionally contribute to material softening. The bainite present in the base material may also undergone microstructural changes in the heat-affected zone. Per definition, this

2 The grain shape aspect ratio used in the current work was given by the length of the major axis divided by the length of the minor axis of an ellipse fit to a grain.

30

Materials Science & Engineering A 717 (2018) 26–33

S. Mironov et al.

Fig. 8. Microstructural evidences of the TRIP effect that possibly occurred in the base material section of the weld during tension tests: (a) EBSD phase map showing reduction of the retained austenite fraction, and (b) higher-resolution EBSD image-quality map with superimposed austenite phase showing possible austenite-to-martensite transformation. See Section 4.2 for details.

crystallographic orientations of retained austenite and martensite were compared with each other, as exemplified in Fig. 9b. To ensure the reliability of the experiment, orientation measurements were performed only in areas directly adjacent to a phase boundary. For the sake of simplicity, crystallographic directions in austenite that were close to those in martensite are circled. It is seen that the 〈111〉 directions of austenite were nearly parallel to the 〈110〉 directions of martensite. On the other hand, the 〈110〉 and 〈112〉 directions of austenite slightly deviated from the respective 〈111〉 and 〈110〉 directions in martensite, i.e., the measured orientation relationship satisfied neither the exact K-S nor the N-W models. Such a situation is often reported for bainite- or martensite transformations in steels [12,13]. To gain additional insight into this issue, orientation data were extracted from several prior-austenite grains and compared with ideal orientations of the martensite variants expected from the K-S and the NW orientation relationships. A typical example is shown in Fig. 10. It is clear from Fig. 10c that the measured martensite variants were characterized by significant orientation spread. This observation is thought to be associated with significant dilatation effect induced during martensite transformation and leading to considerable elastic strain of the martensite. Due to the orientation spread, the orientations expected for the K-S and NW variants essentially overlap (Fig. 10). This perhaps explains the revealed deviations from the ideal orientation relationships (Fig. 9b). Nevertheless, considering a characteristic circular appearance for the variants in the measured pole figure (Fig. 10c), the orientation relationship is believed to be closer to K-S (Fig. 10a) than to N-W (Fig. 10b). It is also seen in Fig. 10c that the prior-austenite grains contained a limited number of martensite variants. This may indicate that variant selection occurred during transformation. In order to check this possibility, the misorientation distribution in martensite was examined, as shown in Fig. 11a. The martensite variants inherited from the same prior-austenite grain are characterized by very specific misorientations linked to the K-S orientation relationship (Supplementary Table S2) and therefore the analysis of mutual proportions of such boundaries may shed light on the possible variant selection. The variant pairing frequency diagram derived from the misorientation distribution is given in Fig. 11b.4

4.2. TRIP effect in the base material zone The change of the phase fractions in the heat-affected zone discussed above may also explain the secondary strain localization in this area during the tensile tests (Fig. 5d and e). Due to the TRIP effect, the base material should exhibit relatively high strain hardening and this may promote strain concentration in the austenite-free heat-affected zone. To check this idea, the microstructure of the base material section of the failed weld specimen was studied3 and the results obtained are summarized in Fig. 8. EBSD measurements showed that the fraction of retained austenite in the deformed base material reduced from the initial 20% (Fig. 6b) to ~ 4% (Fig. 8a). Moreover, it was found that the survived austenite particles were typically adjacent to microstructural domains with relatively low image-quality contrast (Fig. 8b), which presumably corresponded to martensite. The above microstructural evidences are believed to reflect the austenite-to-martensite phase transformation induced by the tensile strain. If so, the TRIP effect might indeed be a possible reason for the secondary strain localization.

4.3. Martensite transformation in stir zone In addition to the material softening in the heat-affected zone (and the TRIP effect in the base material region), inhomogeneous strain distribution in the welded samples was also associated with martensite formation in the stir zone (Figs. 2d and 3a). Attempting to provide deeper insight into this process, several EBSD maps were taken from the central section of this microstructural region and a typical example is shown in Fig. 9a. In this map, individual martensite grains are colored according to their crystallographic orientations relative to the WD and the black clusters represent retained austenite; the total fraction of the retained austenite was measured to be ~ 2.6%. It is well accepted that the martensite transformation in steels follows either the Kurdjumov-Sachs (K-S) orientation relationship, viz. {111}γ//{110}α, 〈110〉γ//〈111〉α, or the Nishiyama-Wasserman (N-W) orientation relationship, viz. {111}γ//{110}α, 〈112〉γ//〈110〉α. To examine the possible orientation relationship in the current study, local 3

4 Assuming that accuracy of EBSD in determination of misorientation axis is ~ 5°, the variant pairing frequency was calculated within 5° tolerance. This led to overlapping of misorientations between some variants, as shown in Fig. 11b.

The examined region is indicated by the selected area in Fig. 5e.

31

Materials Science & Engineering A 717 (2018) 26–33

S. Mironov et al.

Fig. 9. EBSD orientation map (a) and typical orientation relationship between retained austenite and martensite (b) in the stir zone. In (a), grains are colored according to their crystallographic orientations relative to the WD, and black clusters are retained austenite; LABs and HABs are depicted as white and black lines, respectively. In (b), the nearly parallel directions in austenite and martensite are circled. See Section 4.3 for details.

Fig. 10. Ideal {001} pole figures showing orientations of martensite variants transformed from a single austenite grain according to (a) K-S orientation relationship (after Kitahara et al. [14]), or (b) N-W orientation relationship (after Kitahara et al. [15]), and (c) measured pole figure derived from the grain selected in Fig. 9a. Note: The measured pole figure was appropriately rotated to facilitate comparison with the ideal pole figures.

Fig. 11. Misorientation distribution (a) and variant pairing frequency diagram (b) of the stir zone martensite. See Section 4.3 for details.

32

Materials Science & Engineering A 717 (2018) 26–33

S. Mironov et al.

Organization (NEDO) which is co-research with Innovative Structural Materials Association (ISMA) (Project code: P14014).

An evident prevalence of the martensite variants V2, V3/V5, and V4 is seen in this figure. Generally, this effect was expected because the variant selection in steels is believed to depend on carbon content [13]. Specifically, crystallographic preference of the variants V2 and V4 is expected in the studied material (i.e., carbon content of ~ 0.4 wt%) [13]. From this viewpoint, however, the predominance of the V3/V5 variants seems to be abnormal. In principle, this “additional” variant selection may be associated with crystallographic texture and/or grain refinement of the prior-austenite structure. To examine this idea, however, reconstruction of this structure is necessary and therefore this issue requires additional study.

Appendix A. Supplementary Information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.msea.2018.01.053. References [1] O. Matsumura, Y. Sakuma, H. Takechi, Enhancement of elongation by retained austenite in intercritical annealed 0.4C-1.5Si-0.8Mn steel, Trans. ISIJ 27 (1987) 570–579. [2] Y. Sakuma, O. Matsumura, H. Takechi, Mechanical properties and retained austenite in intercritically heat-treated bainitic-transformed steel and their variation with Si and Mn additions, Metall. Trans. A 22 (1991) 489–498, http://dx.doi.org/ 10.1007/BF02656816. [3] K.-I. Sugimoto, M. Kobayashi, S.-I. Hashimoto, Ductility and train-induced transformation in a high-strength transformation-induced plasticity-aided dual-phase steel, Metall. Trans. A 23 (1992) 3085–3091, http://dx.doi.org/10.1007/ BF02656127. [4] B.C. De Cooman, Structure-properties relationship in TRIP steels containing carbide-free bainite, Curr. Opin. Solid State Mater. Sci. 8 (2004) 285–303, http://dx. doi.org/10.1016/j.cossms.2004.10.002. [5] R.S. Mishra, Z.Y. Ma, Friction stir welding and processing, Mater. Sci. Eng. R 50 (2005) 1–78, http://dx.doi.org/10.1016/j.mser.2005.07.001. [6] C.C.P. Mazzaferro, F.D. Ramos, J.A.E. Mazzaferro, T.S. Rosendo, M.A.D. Tier, A.M. Silva, J.F. dos Santos, A. Reguly, Microstructure evaluation and mechanical properties of a friction stir spot welded TRIP 800 steel, Weld. Int. 25 (2011) 683–690, http://dx.doi.org/10.1080/09507116.2010.527477. [7] G.P. Medina, H.L. Ferreira, P.Z. Robledo, A.M. Perez, F.A.R. Valdes, Microstructural development in a TRIP-780 steel joined by friction stir welding (FSW): quantitative evaluations and comparisons with EBSD predictions, Soldag. Insp. 21 (2016) 146–155, http://dx.doi.org/10.1590/0104-9224/SI2102.04. [8] A. Mostafapour, A. Ebrahimpour, T. Saeid, Finite element investigation on the effect of FSSW parameters on the size of welding subdivided zones in TRIP steels, Int. J. Adv. Manuf. Technol. 88 (2017) 277–289, http://dx.doi.org/10.1007/s00170-0168749-3. [9] A. Mostafapour, A. Ebrahimpour, T. Saeid, Numerical and experimental study on the effects of welding environment and input heat on properties of FSSWed TRIP steel, Int. J. Adv. Manuf. Technol. 90 (2017) 1131–1143, http://dx.doi.org/10. 1007/s00170-016-9430-6. [10] H.A. Bruck, S.R. McNeill, M.A. Sutton, W.H. Peters III, Digital image correlation using Newton-Raphson method of partial differential correction, Exp. Mech. 29–3 (1989) 261–267, http://dx.doi.org/10.1007/BF02321405. [11] S. Yoneyama, Basic principle of digital image correlation for in-plane displacement and strain measurement, Adv. Compos. Mater. 25-2 (2016) 105–123, http://dx.doi. org/10.1080/09243046.2015.1129681. [12] N. Takayama, G. Miyamoto, T. Furuhara, Effects of transformation temperature on variant pairing of bainitic ferrite in low carbon steel, Acta Mater. 60 (2012) 2387–2396, http://dx.doi.org/10.1016/j.actamat.2011.12.018. [13] A. Stormvinter, G. Miyamoto, T. Furuhara, P. Hedstrom, A. Borgenstam, Effect of carbon content on variant pairing of martensite in Fe-C alloys, Acta Mater. 60 (2012) 7265–7274, http://dx.doi.org/10.1016/j.actamat.2012.09.046. [14] H. Kitahara, R. Ueji, N. Tsuji, Y. Minamino, Crystallographic features of lath martensite in low-carbon steel, Acta Mater. 54 (2006) 1279–1288, http://dx.doi.org/ 10.1016/j.actamat.2005.11.001. [15] H. Kitahara, R. Ueji, M. Ueda, N. Tsuji, Y. Minamino, Crystallographic analysis of plate martensite in Fe-28.5 at% Ni by FE-SEM/EBSD, Mater. Character. 54 (2005) 378–386, http://dx.doi.org/10.1016/j.matchar.2004.12.015.

5. Conclusions In this work, microstructure and tensile behavior of friction-stir welded TRIP steel were systematically studied in order to clarify the microstructure-property relationship in this material. To this end, EBSD was employed for thorough microstructure characterization and DIC technique was used for precise analysis of the tensile behavior. The main conclusions derived from this study are as follows. FSW led to significant microstructural changes. Specifically, the thermal effect of the welding process gave rise to substantial material softening in the heat-affected zone and induced martensite transformation in the stir zone. The above microstructural changes promoted fairly inhomogeneous strain distribution during subsequent transverse tensile tests, thus resulting in premature failure. Therefore, they were considered the key factors governing the tensile behavior of the weld. Material softening in the heat-affected zone was shown to be a combined result of dissolution and spheroidisation of retained austenite as well as recovery of bainitic ferrite. The stir zone martensite was characterized by significant orientation spread, presumably due to the dilatation effect. As a result, the orientation relationship between the austenite and martensite could not be precisely described by either the ideal Kurdjumov-Sachs or Nishiyama-Wasserman relation. Moreover, martensite transformation was found to be influenced by variant selection. The latter effect was attributed to a specific carbon content of the steel studied and was suggested to be additionally influenced by crystallographic texture and grain refinement in the high-temperature austenite phase. Acknowledgments The authors are grateful to Dr. K. Kobayashi, Mr. H. Furuya, Mr. H. Honda, and Mr. S. Yamamoto for technical assistance. One of the authors (S. Mironov) would like to express his hearty thanks to Tohoku University for providing a scientific fellowship. This paper was partially based on results obtained from a future pioneering project commissioned by the New Energy and Industrial Technology Development

33