ISIJ International
Online ISSN : 1347-5460
Print ISSN : 0915-1559
ISSN-L : 0915-1559
Regular Article
Micro-tensile Behaviour of Low-alloy Steel with Bainite/martensite Microstructure
Kwangsik KwakTsuyoshi MayamaYoji Mine Kazuki Takashima
Author information
JOURNAL OPEN ACCESS FULL-TEXT HTML

2016 Volume 56 Issue 12 Pages 2313-2319

Details
Abstract

Micro-tensile testing and numerical analysis using a crystal plasticity finite element method (CPFEM) were employed to elucidate the deformation behaviour of bainite/martensite structures of a low-alloy steel. The bainite single-phase specimens exhibited habit-plane-orientation-dependent yielding, similar to the martensite single-phase specimens. In the bainite/martensite dual-phase specimen, deformation concentrated in the bainite region oriented favourably for in-habit-plane slip, leading to low-ductility fracture. With consideration of the habit-plane-orientation-dependent yielding, the present CPFEM analysis successfully reproduced the anisotropic plastic deformation behaviour of the single-phase steels observed in the experiments. The numerical results for the bainite/martensite specimen showed slip localization in the bainite region and stress concentration near the interphase boundary. This suggests that the interphase boundary can be a site for the fracture origin.

1. Introduction

The increasing size of ships, constructions, and civil engineering structures has extended the applications of welding steels. Therefore, it has been a major challenge to ensure the durability and reliability of welded joints in developing welding steels. When building ships and constructions, thick plates for large structures experiences large heat input welding because it improves the work efficiency and economic performance. However, a heat-affected zone (HAZ) mainly comprising so-called upper bainite is formed adjacent to the welded joint through the repetition of heating and cooling during the welding process.1,2,3,4,5,6) In addition, untransformed austenite with a high carbon content is retained between the bainite laths through holding at an elevated temperature range in which upper bainite forms. The retained austenite is partly transformed to martensite by cooling or mechanical processing, which causes the so-called martensite-austenite constituent.7) The brittleness of martensite has been considered to decrease the fracture toughness of welded structures.8,9,10) In contrast, a tensile test study by Chen et al. revealed that severe deformation was localised within the bainite region, leading to stress concentration at the interface between the bainite and martensite regions.11) There is controversy in the role of martensite in the fracture mechanism of the welded structures. The complexity of hierarchical microstructures with mixed bainitic and martensitic phases hinders the understanding of the mechanical response of each phase.

Mine et al. have analysed the mechanical characteristics of the microconstituents in the single-phase lath martensite structure of a low-alloy steel using micro-tensile testing.12) This study revealed that lath martensite structures with single packets exhibited moderate ductility and that the plastic deformation behaviour depended on the habit-plane orientation relative to the loading direction. Thus, micro-tensile testing allows for the analysis of deformation behaviour on the microstructural scale. In addition, a micro-tensile test study by Ogata et al. indicated that pre-strained ferrite/martensite dual-phase steel was embrittled by ultra grain refinement in the ferrite region, rather than by cracking in the martensite region.13) This finding suggested that martensite was not always the main cause of embrittlement in high-strength steels. In the present study, micro-tensile testing was employed to characterise the mechanical response of each phase in a bainite/martensite structure. Additionally, numerical analysis with crystal plasticity finite element method (CPFEM) was performed to elucidate the microscopic mechanism of the experimentally observed deformation behaviour.

The CPFEM analyses have helped in understanding plastic behaviours,14,15) and studies on dual-phase steels successfully reproduced the observed deformation and fracture behaviours.16,17) However, few studies have applied CPFEM to bainite/martensite dual-phase-structured steel. Therefore, in the present study, we confirm the applicability of a simple CPFEM with linear hardening laws to the deformation behaviour of the bainite and martensite single-phase structures. Furthermore, by using the material parameters identified in the analyses of single-phase structures, the analysis of the bainite/martensite dual-phase structure was also performed to discuss the deformation mechanism.

2. Material and Methods

The material used in this study was low-alloy steel of which the chemical composition is listed in Table 1. This material was solution-treated at a temperature of 1323 K and subsequently held for 0.5 h at a temperature of 773 K in a salt bath, followed by air-cooling. This heat treatment scheme formed an upper bainitic microstructure with dispersed martensite precipitates. A fully bainitic single-phase microstructure was also obtained through fan-cooling after the first solution treatment at 1323 K. Platelets with dimensions of 10 mm × 10 mm × 1 mm were cut from the bulk material and ground to a thickness of ~20 μm using emery paper. Both surfaces of each sample were polished using colloidal silica paste. A field-emission gun scanning electron microscope, combined with an electron backscatter diffraction (EBSD) analyser was used at a scanning step size of 0.3 μm under an accelerating voltage of 20 kV to determine the crystallographic orientations of the samples. EBSD analysis was performed using TSL OIM software (v. 7.1.0).

Table 1. Chemical composition of the low-alloy steel used in the present study (mass%).
CSiMnPSNiAlBNOFe
0.20.42.02<0.0020.0011.520.0280.0010.0011<0.001Bal.

A micro-tensile specimen with a gauge section of ~20 mm × 20 mm × 50 mm was fabricated using a focused ion beam. Tensile testing was performed at a displacement rate of 0.1 μm s−1 at room temperature in atmospheric air. The gauge section of each specimen was monitored using an optical microscope during tensile testing. The strain measurement18) was performed using images obtained continuously during tensile testing.

3. Experimental Results

3.1. Crystallographic Characteristics of Bainite/martensite Microstructures

A transmission electron microscopy study by Lan et al. revealed the morphology of the isolated martensitic phase formed by a heat treatment that simulated typical welding conditions.19) The isolated martensite contained a small amount of remained austenite between laths.19) Therefore, martensitic islands are frequently referred to as martensite-austenite constituents (MA). Detailed studies of upper bainite and martensite were conducted by EBSD analysis. The metallographic natures of both phases have been reported to differ based on their carbon contents and transformation temperatures.20,21)

Twenty-four variants of bainite can be formed from a single austenite grain with the Kurdjumov-Sachs (K-S) orientation relationship, as with martensite. Prior austenite grains, divided into clusters, consist of six variants that have a common habit plane, denoted as a CP (close-packed) group. The K-S variants are also divided into three distinct clusters consisting of eight variants from the Bain lattice correspondence ([001]γ//[001]α, [100]γ//[110]α, and [010]γ//[110]α), denoted as Bain groups.23) In other words, the variants formed from one prior austenite grain can be classified into four CP groups and three Bain groups. The region of lath martensite can be identified by the above-mentioned CP group classification because the martensite variants with common habit planes show a marked tendency to form adjacently. As for bainite, the paring of adjacent variants differs based on the transformation temperature. While the metallographic nature of bainite transformed within the temperature range between 623 and 723 K is similar to that of lath martensite,20) bainite transformed at higher temperatures (~853 K) forms adjacent laths of variants belonging to the same Bain group, with a region containing many low-angle boundaries.23)

Figure 1 shows a crystallographic orientation map determined by EBSD analysis, an (001) pole figure of the bainite/martensite microstructure, colour-coded maps of Bain and CP groups, and a (110) pole figure corresponding to the CP map. The bainite and martensite contained in one prior austenite grain are formed with the K-S orientation relationship, as shown in Fig. 1(b). The regions consisting of Bain group (Fig. 1(c)) and variants having the same habit plane (Figs. 1(d) and 1(e)) correspond to the bainitic and martensitic phases, respectively, by classification based on the metallographic natures of bainite and martensite. This indicates that both phases of the bainite/martensite structure can be identified based on the metallographic information obtained by EBSD analysis.

Fig. 1.

(a) EBSD crystallographic orientation map and (b) (001) pole figure of the bainite/martensite microstructure. (c) Bain map obtained in region C and (d and e) CP map and (110) pole figure obtained in region D in (a). White and black lines are low-angle (5°<θ<15°) and high-angle (θ>15°) boundaries, respectively. (Online version in color.)

3.2. Micro-tensile Behaviour in Bainite Single-phase Structures

Plastic deformation behaviour in martensite is greatly dependent on the habit plane orientation.12) We prepared micro-tensile specimens with different habit plane orientations with respect to the loading direction (LD) to investigate the anisotropy in the plastic deformation behaviour of the bainitic single-phase microstructure, which has a lath structure similar to that of martensite. Figure 2 shows colour-coded maps of the gauge sections of tensile specimens based on EBSD analysis and the corresponding (110) pole figures. The habit-plane orientation of the B-I specimen was parallel to the maximum shear stress plane (Fig. 2(b)), while the habit plane of the B-P specimen was oriented parallel to the LD (Fig. 2(d)). The broken line in each pole figure indicates the trace of the habit plane. The bainite used in the present study shows a packet structure because the cooling rate was chosen to prevent martensite from mixing into the bainite. However, it was considered that the structure within the block was essentially the same as that of bainite/martensite dual-phase steel.

Fig. 2.

EBSD colour-coded maps and the corresponding (110) pole figures for the B-I specimen (a and b) and the B-P specimen (c and d) before micro-tensile testing. LD and TD denote the loading and transverse directions, respectively, in micro-tensile testing. (Online version in color.)

Figure 3 shows the stress–strain curves obtained by micro-tensile testing. The yield and maximum tensile stresses of the B-I specimens are 594 and 675 MPa, respectively. The B-I specimen was fractured at a strain of approximately 25%. Meanwhile, the yield and maximum tensile stresses of the B-P specimen are 736 and 914 MPa, respectively, exceeding those of the B-I specimen. The strain of failure is approximately 11%. This value is lower than that of the B-I specimen. As mentioned previously, a micro-tensile test study by Mine et al. revealed that lath martensite, which has a hierarchical structure similar to the bainite, exhibited habit-plane-orientation-dependent yielding.12) From these results, it is considered that the mechanical response of the bainite greatly differs based on the habit plane orientation.

Fig. 3.

Stress-strain curves for the B-I and B-P specimens.

Figure 4 shows optical microscope images obtained during tensile testing in the B-I specimen. A slip trace appears at a strain of 0.3%, which has an angle of 48° relative to the LD (Fig. 4(b)). This slip trace corresponds to the (110) habit plane orientation (Fig. 2(b)). Slips parallel to the habit plane (i.e. in-habit-plane slip) are possibly activated. At the strain of 20%, deformation is localised in the site where the initial slip trace is observed. At greater strain, the specimen was fractured along to the habit plane.

Fig. 4.

Optical micrographs showing the deformation process of the B-I specimen.

Figure 5 shows the optical microscope images obtained during tensile processing and the (112) and (111) pole figures obtained by EBSD analysis in the initial state for the B-P specimen. The circle and triangle symbols in the pole figures indicate the normal direction of the slip plane and the direction, respectively, in the slip system that exhibits the highest Schmid factor among the {110} <111> and {112} <111> slip systems. A slip trace is observed on the surface of the specimen at a strain of 4% (Fig. 5(b)). This trace has an angle of 71° relative to the LD, which coincides with the {112} <111> slip system exhibiting the highest Schmid factor (Figs. 5(d) and 5(e)). Therefore, it is considered that slip across the habit plane (i.e. out-of-habit-plane slip) is preferentially activated in the B-P specimen. At a strain of 11%, deformation is localised in the site where the initial slip trace is observed (Fig. 5(c)). After that, the specimen was fractured across the habit plane at the strain of 11%.

Fig. 5.

(a–c) Optical micrographs showing the deformation process of the B-P specimen. (d and e) (112) and (111) pole figures obtained for the B-P specimen in the initial state. SF represents the highest Schmid factor. Colours in the pole figures correspond to those in Fig. 2(c). (Online version in color.)

The B-I specimen, in which the in-habit-plane slip was activated, shows a lower stress level compared to the B-P specimen in which the out-of-habit-plane slip was activated. The resolved shear stresses of the in-habit-plane and out-of-habit-plane slip systems were 285 and 360 MPa, respectively, and the ratio of the latter to the former was 1.3, similar to the 1.5 ratio of lath martensite. It is suspected that the difference in resolved shear stress between the in-habit-plane and out-of-habit-plane slip systems is caused by the existence of retained austenite between laths, as Maresca claims,24) carbides included in the matrix, or block boundaries. These results revealed that both bainite and martensite show habit-plane-orientation-dependent yielding.

3.3. Micro-tensile Behaviour in Bainite/martensite Dual-phase Structure

Figure 6 shows the initial microstructure of the B-MA specimen, which consists of bainite/martensite dual-phase structure; Figs. 6(a)–6(d) show the crystallographic orientation distribution, determined by EBSD analysis, in the gauge section of the specimen, the colour-coded map for CP group, (110) pole figure, and (111) pole figure, respectively.

Fig. 6.

(a) EBSD crystallographic orientation map, (b) CP map, and (c and d) the corresponding (110) and (111) pole figures for the B-MA specimen. (Online version in color.)

Figure 7 shows the micro-tensile testing result obtained using the B-MA specimen shown in Fig. 6. The yield stress of the B-MA specimen is 520 MPa, while the maximum tensile stress of 795 MPa occurs by strain hardening. The specimen was fractured at a strain of 7.2% from the stress–strain curve shown in Fig. 7(a). When compared with the stress–strain curves for the bainite single-phase specimens shown in Fig. 3, the B-MA specimen exhibits markedly decreased ductility, despite its yield stress being equivalent to that of the B-I specimen. Figures 7(b)–7(d) show the deformation behaviour observed by optical microscopy during tensile testing, which correspond to each point labelled in Fig. 7(a). The initial slip trace appears in the region indicated by the arrow in Fig. 7(b), which coincides with the site indicated by the arrow in the CP map shown in Fig. 6(b). This indicates that the slip trace appears in the bainite phase adjacent to the martensite phase. Moreover, the slip trace has an angle of 91° with respect to the LD, which nearly corresponds to the trace of the habit plane shown in Fig. 6(c). Therefore, the initial slip trace is attributed to the activation of the in-habit-plane slip system in the bainite phase. The habit plane has an angle of 54° relative to the through-thickness direction. The Schmid factor of the in-habit-plane slip system has the highest value of 0.44. With the deformation procedure, other slip traces are also observed in different areas (Fig. 7(c)) and the deformation is localised in these areas where slip traces are observed (Fig. 7(d)).

Fig. 7.

(a) Stress-strain curve and (b–d) optical micrographs showing the deformation process of the B-MA specimen.

Discussion of the active deformation mechanism based only on the above results is difficult because it requires the consideration of complex interactions depending on the sizes and configurations of the bainite and martensite phases. In this study, therefore, we performed a numerical analysis that considered the slip of each phase as an elementary step. The deformation mechanism is discussed based on the obtained stress and strain distributions.

4. Discussion

4.1. Crystal Plasticity Finite Element Analysis

In the present study, a large-strain FEM with a rate-dependent crystal plasticity model by Peirce et al.25) was used. The slip rate of slip system i is calculated by   

γ ˙ ( i ) = γ ˙ 0 sgn( γ ( i ) ) | τ ( i ) g ( i ) | 1/m (1)
where γ ˙ 0 and m are the reference slip rate and strain rate sensitivity parameter, respectively. τ(i) and g(i) in Eq. (1) are the resolved shear stress and reference stress for slip system i, respectively. The following evolution equation is used for g(i):   
g ˙ ( i ) = j=1 N h| γ ˙ ( j ) | (2)
where h is the hardening parameter. While the hardening behaviour should generally depends on the loading history, the active slip system and the interaction between the slip systems, a constant hardening parameter for all slip systems is used for simplicity in the present calculation because of the lack of understanding of the detailed hardening behaviour of bainite and martensite. For bainite and martensite, the {110} <111> and {112} <111> slip system families are considered as deformation mechanisms.

Figure 8(a) shows the finite element mesh used in the analyses of single-phase structures. The geometry of the model corresponds to the experimental specimen. The model is divided into 5600 finite elements by 20-node solid elements. The boundary conditions are applied as shown in Fig. 8(b). The bottom surface is fixed in the LD. For the top surface, uniaxial tensile loading is applied with a constant increase in displacement along the LD that corresponds to the experimental displacement rate.

Fig. 8.

(a) Finite element mesh for CPFEM analysis and (b) boundary conditions. Analysis models and calculated and experimental stress-strain curves for B-I (c and e), B-P (d and f), M-I (g and i), and M-P (h and j). (Online version in color.)

Figures 8(c) and 8(d) show the analysis models corresponding the bainite single-phase structures shown in section 3.2. The Euler angles obtained by EBSD analysis are allocated to each finite element as its initial crystallographic orientation. The material parameters are identified by fitting to the experimental stress–strain curves (Fig. 3), as shown in Table 2. The values of the experimentally obtained resolved shear stresses as shown in Table 3 are applied to the reference stresses. That is, different reference stresses were used for the in-habit-plane and out-of-habit-plane slip systems by considering the experimental results showing the yielding behaviour depends on the habit plane orientation.

Table 2. Material parameters other than initial reference stresses in the present CPFEM analysis.
Young’s modulus, E (GPa)200
Poisson’s ratio, ν0.3
Strain rate sensitivity, m0.02
Reference slip rate, γ ˙ 0 (s−1)0.001
Strain hardening constant, h (MPa)200
Table 3. Initial reference stresses (MPa) in the present CPFEM analysis.
{110}<111>slip systems{112}<111>
slip systems
In-habit-planeOut-of-habit-plane
B-I, B-P285360360
M-I, M-P93113331333

Figures 8(e) and 8(f) show the calculated stress–strain behaviour obtained by the crystal plasticity analysis of the bainite single-phase models. The experimental stress–strain behaviour (Fig. 3) is shown for comparison. The calculated stress–strain curve in the initial plastic deformation stage disagrees with the experimental stress–strain curve. This disagreement is attributed to the work-hardening used in this calculation being modelled as simple linear model. However, the overall stress–strain behaviours of both the B-I and B-P models show good agreement with the experimental results. Therefore, it is confirmed that the plastic anisotropy of the bainite can be quantitatively evaluated using the present analysis method.

Figures 8(g) and 8(h) show the analysis models of the martensite single-phase structures. To construct the analysis model of martensite single-phase structure, the result of carbon steel 55c that the authors’ research group studied previously were used, because the carbon contents of martensite in the bainite/martensite dual-phase structure specimen were between 0.37 and 0.52%. The initial crystallographic orientation and material parameters were identified in the same way as the analysis method for the bainite single-phase structure, as described above.

Figures 8(i) and 8(j) show the calculated stress–strain curves of the martensite single-phase models. The experimental stress–strain curves are also shown for comparison. The calculated stress–strain curves of the martensite single-phase models show good agreement with the experimental stress–strain curves. From these numerical analysis results for each single-phase model of bainite and martensite, it is confirmed that the crystal plasticity analysis used in this study can quantitatively describe the tensile deformation behaviour of the bainite/martensite single-phase structures in dual-phase low-alloy steel.

4.2. Inhomogeneous Plastic Deformation of Bainite/martensite Dual-phase Structure

In this section, the tensile loading analysis of the bainite/martensite dual-phase structure (B-MA specimen) is performed using the material parameters determined in the previous section to discuss the deformation mechanism.

It is difficult to obtain information on the crystallographic orientation and phase arrangement inside a specimen by only observing the specimen surface in such cases. In the present section, therefore, the analysis model of one element of thickness was constructed based on the observation results of the specimen surface. The boundary conditions in the analysis were also modified as shown in Fig. 9(a) where one side of the model surface on TD-LD plane was fixed in ND. Figure 9(b) shows the analysis model based on the B-MA specimen, which is colour-coded according to each block. For the initial crystallographic orientation, the Euler angles obtained by EBSD analysis are allocated to each finite element, as with the analyses of the single-phase structures in the previous section.

Fig. 9.

(a) Boundary conditions and (b) analysis model for B-MA specimen. White lines indicate the boundaries surrounding the martensite regions. (Online version in color.)

Figure 10 shows the calculated developments of the distributions of equivalent strain and equivalent stress with increased tensile strain. A high equivalent strain distribution is observed in the bainite region adjacent to martensite, as shown in Figs. 10(a)–10(c). When compared to the experimental results shown in Figs. 6 and 7, the region of high equivalent strain shown in Fig. 10 coincides with the region in which the initial slip trace is observed. On the other hand, the high stress is shown in the high-angle boundary, i.e. Bain boundary, in Figs. 10(d)–10(f). The stress is especially high in the interface between the bainite, which shows a large equivalent strain, and martensite.

Fig. 10.

Distributions of equivalent strain (a–c) and equivalent stress (d–f) analysed with consideration of the habit-plane-orientation-dependent yielding in the B-MA model. (Online version in color.)

According to the slip system activity obtained in present crystal plasticity analysis, the in-habit-plane slip system is preferentially activated after the onset of plastic deformation, which corresponds to the experimental results shown in Figs. 6 and 7. In contrast, slip in the martensite phase is not entirely activated, which also coincides with the experimental results in that a definite slip trace is not observed in the martensite phase.

For the deformation mechanism in the inhomogeneous deformation of the dual-phase structure, considering the distributions of equivalent strain and equivalent stress obtained in the calculated results, it is expected that cracks may nucleate inside the largely deformed bainite adjacent to the martensite. These cracks could propagate along the interface between bainite and martensite due to high stress concentrations at the interface.

4.3. Plastic Anisotropy of Bainite in Dual-phase Structure

In the previous section, the crystal plasticity analysis was performed with consideration of the habit-plane-orientation-dependent yielding in the bainite and martensite phases. In this section, we performed a crystal plasticity analysis in which we assumed that the reference stresses for all slip systems in the bainite phase are equivalent, to confirm the effect of the habit-plane-orientation-dependent yielding on the deformation behaviour.

Figure 11 shows the distributions of equivalent strain and equivalent stress in the case where a reference stress of 285 MPa is applied for all the slip systems. The distributions of equivalent strain and equivalent stress are significantly different based on the consideration of the habit-plane-orientation-dependent yielding, as shown in Figs. 10 and 11. According to the analysis of the active slip systems, the {112} <111> slip system exhibiting the highest Schmid factor is activated in the region where the equivalent strain is localised, as shown in Fig. 11. Moreover, stress concentrations appear in the interfaces between the bainite and martensite. However, not only is the value of the stress small, but also the regions of stress localisation are different when compared with the calculation results shown in Fig. 10 with consideration of the habit-plane-orientation-dependent yielding.

Fig. 11.

Distributions of equivalent strain (a–c) and equivalent stress (d–f) analysed without consideration of the habit-plane-orientation-dependent yielding in the B-MA model. (Online version in color.)

From these results, it is found that the habit-plane-orientation dependency is crucial to evaluate the inhomogeneous deformation behaviour of bainite/martensite dual-phase structures; otherwise, the stress concentration can be underestimated in analysis.

5. Conclusions

Micro-tensile testing, combined with analysis using CPFEM, was employed to elucidate the deformation behaviour and mechanical characteristics of bainite/martensite dual-phase-structured low-alloy steel. The conclusions are summarised as follows:

(1) The bainite single-phase specimens exhibited habit-plane-orientation-dependent yielding, similar to the martensite single-phase specimens.

(2) In the bainite/martensite dual-phase specimen, deformation concentrated in the bainite region was oriented favourably for in-habit-plane slip, leading to low-ductility fracture.

(3) When considering the habit-plane-orientation-dependent yielding, the present CPFEM analysis can reproduce the experimentally observed plastic anisotropy in the stress–strain behaviour of the bainite and martensite single-phase structures.

(4) In the CPFEM analysis for the bainite/martensite dual-phase structure, a large strain was accumulated in the bainite region, with the stress concentrated near the bainite/martensite boundary. This suggests that the interface between different phases can be a site for fracture origin.

Acknowledgement

The present work was supported in part by a Research Group ‘Fracture Toughness of High-Strength Steels’ in the Iron and Steel Institute of Japan (ISIJ) and in part by a Grant-in-Aid for Scientific Research 15H02302 from the Japan Society for the Promotion of Science (JSPS).

References
 
© 2016 by The Iron and Steel Institute of Japan
feedback
Top