1. Introduction
High strength with preferable ductility is one of the essential properties demanded for sheet steel.1) The low alloyed transformation induced plasticity (TRIP) aided steel has been designed for this demand with an application of bainitic transformation.2,3,4,5) The TRIP aided steel typically includes small amount of Si because the alloying of Si prevents from the precipitation of carbide. As a result, the TRIP aided steel consists of bainitic ferrite and the retained austenite as meta-stable phase at the room temperature.6,7) The retained austenite has a potential for the deformation-induced martensitic transformation (DIMT), realizing high work hardening that is necessary for the preferable ductility. In order to obtain the further higher strength, the TRIP aided steel with higher carbon content has been studied by several researchers. For example, Tsuchida et al.8) reported the tensile properties at various strain rates of the 0.3%C-0.15%Si-2%Mn (mass%) steels which have the tensile strength is more than 1 GPa. The TRIP aided steels with the tensile strength higher than 1.5 GPa have been reported as well.9) This trend in the research indicates the importance of the effect of carbon content in the TRIP aided steels.
By the way, the sheet forming, such as bending, are necessary for the engineering application of the steel sheets. Generally, both tensile and compressive conditions are distributed spatially at the sheet forming. Thus, the effect of deformation mode (compression, tension, etc.) on the DIMT should be clarified for the design and the modelling of the sheet forming. Although the DIMT behaviors at the various stress conditions have been studied at the basic perspectives by some scholars, such as Olson and Cohen,10) limited number of papers have reported for the TRIP-aided steels. Figure 1 shows the previously published data of the change in the volume fraction of austenite during uniaxial deformation of the various TRIP aided steels.8,11,12,13,14,15,16,17,18,19,20,21) These data were plotted as a function of absolute strain applied at room temperature. Most of the data were examined by tensile deformation and the available data for the compression are limited, except for some studies such as the paper by Kawata et al.,19) Ishimaru et al.20) and Yasutomi et al.21) Consequently, the aim of this paper is to clarify the kinetics of DIMT in the TRIP aided steels with various carbon contents both at tension and at compression.
Fig. 1. Fraction of austenite during deformation of various austenitic stainless steels and low-alloyed TRIP aided steels at ambient temperature. The bold lines indicates that these data contains both at tension and at compression.
2. Experimental Procedure
Si-bearing steels (2.0%Si - 0.2%Mn - 0.001%P - <0.001%S - 1.0%Cr - X%C) (in mass%) with different carbon contents (X = 0.15%, 0.25% and 0.62%) were studied. The steel with the largest content of carbon was studied by the authors previously,11) and the other two steels were prepared additionally for this study. All these samples were provided to austempering heat treatment in order to obtain the retained austenite with bainitic ferrite. The hot rolled bars with a dimension of 14 mm × 14 mm × 150 mm were austenitized at 950°C (the 0.15%C and 0.25%C steels) or 850°C (the 0.62%C steel) for 600 sec, and then soaked in a salt bath at 400°C for 1800 sec. The austempered samples were cut to the tensile and compressive test pieces. Conventional tensile test was conducted with a test piece having a gage length of 42 mm and a diameter of 6 mm and tested at room temperature at a constant cross head speed of 0.85 mm/min (an initial strain rate of 0.33 × 10−4/sec). Conventional compression test with a cylindrical sample with a height of 12 mm and a diameter of 8 mm was conducted at room temperature at a constant strain rate of 10−4/s.
Microstructure was examined by a scanning electron microscope (SEM) with an electron back scattering diffraction (EBSD) measurement system. The observation surface for SEM was prepared by mechanical gliding and electrochemical polishing with a solution of 10 vol% perchloric acid + 90 vol% acetic acid at an applied voltage of about 30 V for a polishing time of around 1 min at room temperature. SEM (JEOL 7000F) observation was conducted at an accelerated voltage of 15 kV and EBSD measurements (TSL OIM Data Collection) were done at a scanning pitch of 0.2 μm. BCC and FCC phases were set as the possible phases.
In order to confirm the volume fraction of austenite, in situ measurements was conducted with a high resolution and high intensity time-of-fright (TOF) neutron diffractometer at the Beam Line No.19 (TAKUMI) at the Japan Proton Accelerator Research Complex (J-PARC) center.22) The specimens examined at TAKUMI have dimensions with a parallel part length of 30 mm and a diameter of 6 mm for the tensile test, and with a height of 12 mm and a diameter of 5.5 mm for the compressive test. These test pieces were deformed at a quasi-static strain rate, and the diffraction patterns were repeatedly measured at the initial and the loaded states. Data were analyzed with using a Rietveld software called Z-Rietveld.23) The volume fraction of retained austenite was measured with the diffraction peaks indexed {110}α, {200}α, {211}α, {111}γ, {200}γ and {311}γ.
3. Results and Discussion
3.1. Microstructure of the As-austempered Bainitic Steels
Figure 2 shows the phase color maps of the 0.15%C (a), 0.25%C (b) and 0.62%C (c) bainitic steels. These data were obtained with the EBSD measurement. All of these show bainitic structure with equiaxed prior austenite grains. The diameter of the prior austenite grains in the 0.15%C and the 0.25%C bainitic steels were larger than that of the 0.62%C steels partially due to the higher austenitization temperature. Most of the retained austenite show blocky morphology while the remaining shows elongated shape. The volume fraction of austenite increases with increasing the carbon content. This change was related to the mechanism of bainitic transformation where the bainitic ferrite evolves with the carbon diffusion into retained austenite. The carbon content of retained austenite increases with the progress of the bainitic transformation until the carbon content reaches close to T0 content at which the Gibbs energy of the retained austenite equals to that of the bainitic ferrite.24) The isothermal holding temperature adopted in this work was constant (400°C) for all the three steels, thus the T0 content for these bainitic reaction are predicted to be the same. Supposed that the carbon content of the bainitic ferrite can be relatively negligible, the mass valance rule brings the following relation,
|
f
V
C
¯
γ
/
V
¯
m
=
C
all
| (1) |
where fv,
C
¯
γ
,
V
¯
m
and Call are the volume fraction, the atomic carbon content and the molar volume of retained austenite, and the atomic carbon content of alloy, respectively. This equation implies the linier relation between the volume fraction of austenite and the carbon content of alloys. Actually, the experimental data follows the liner relation as shown in Fig. 3. This means that the average carbon content of the retained austenite is the same in all the three steels.
Fig. 2. Phase color maps of the 0.15%C (a), 0.25%C (b) and 0.62%C (c) bainitic steels as austempered. the white line indicates the prior austenite grain boundary. (Online version in color.)
Fig. 3. Area fraction of retained austenite as a function of atomic fraction of carbon in the alloys.
Figure 4 shows the inverse pole figures indicating the austenite orientation parallel to the rolling direction at the sample preparation before the austempering process. This direction is parallel to the loading axis for the tensile and the compressive tests. The number denoted with the contour line indicates the density which is expressed in multiple of random density. All of these figures imply no significant texture component. This is because the texture was randomized by the diffusional transformation at the reheating stage before the austempering.
Fig. 4. Inverse pole figures showing the loading axis of the 0.15%C (a), 0.25%C (b) and 0.62%C (c) bainitic steels as austempered. The orientation showing maximum density indicates the cross mark. Area fraction of FCC phase was denoted at top-left of each figure. (Online version in color.)
Figure 5 shows the true stress - true strain curves at the tension or compression of the bainitic steels with various carbon contents. These curves were evaluated with the macroscopic change of loads and the dimensions (length or height) under the volume constant rule. Both the tensile and compressive tests clarified the tendency that the flow stress becomes higher with increasing of the carbon content. When compared between the tension and the compression at the same carbon content, the following characteristics were found:
(i) the yield stress at tension was smaller than that at compression,
(ii) the work hardening rate at tension was higher than that at compression.
Fig. 5. True stress - true strain curves of the 0.15%C (a), 0.25%C (b) and 0.62%C (c) bainitic steels. (Online version in color.)
One of the possible reasons for the former should be related to the plastic anisotropy some of which is typically discussed for the so-called Bauschinger effect with internal back stress.25) Since the bainite reaction is complex structural change, the resultant residual stress is difficult to be estimated. However, it seems to be possible that the compressive back stress can be evolved in the retained austenite due to the volume expansion accompanied with the transformation from the austenite matrix to the bainitic ferrite. The latter character (ii) should be related to the change of the DIMT behavior by the stress polarity. Because the work hardening is enhanced by the DIMT, the smaller work hardening rate at the compression indicates the suppression of the DIMT.
There are several studies which have pointed out the dependence of the matrix orientation on the occurrence of DIMT.13,26,27,28) In this work, the retained austenite in the samples strained up to several strains were examined by EBSD to measure the area fraction and the texture to clarify both the orientation dependence and kinetics of the DIMT. Figure 6 shows the inverse pole figures of retained austenite indicating the tensile orientation of the tensile samples. At all three samples, tensile deformation brought the concentration of tensile orientation around at <111>. This texture is one of the major components of deformation texture in FCC metals tensile-deformed uniaxially.29,30) The relatively low density was found in the 0.15%C bainitic steels than those of the other two samples. This is probably because the 0.15%C bainitic steel has the relatively small size of retained austenite grains. Since the plastic deformation can be constrained by the surrounding boundaries and other phases, the lattice rotation of the finer austenite grains in the lower carbon steels should be different from conventional rotation as found in the higher carbon steel. On the other hand, the compressive deformation brought the different texture in retained austenite which is shown in Fig. 7. The orientation concentration evolved around the <101> in all the bainitic steels. It should be noted that the change of stress polarity provides significant change in the developed texture. The relation between the texture evolution and the DIMT is discussed later.
Fig. 6. Inverse pole figures showing the tensile axis of the 0.15%C (a1, a2), 0.25%C (b1, b2) and 0.62%C (c1, c2) bainitic steels tensile strained to fracture. Area fraction of FCC phase measured with EBSD was denoted at top-left of each figure. (Online version in color.)
Fig. 7. Inverse pole figures showing the compressive axis of the 0.15%C (a1, a2), 0.25%C (b1, b2) and 0.62%C (c1, c2) bainitic steels. Area fraction of FCC phase measured with EBSD was denoted at top-left of each figure. (Online version in color.)
Figure 8 shows the volume fraction of the austenite in the various bainitic steels with different carbon contents, presenting by the linier - liner plot (a) or the logarithm - liner plot (b). The decreasing of the volume fraction indicates the occurrence of DIMT. The volume fraction of the 0.62%C bainitic steels was measured by in situ neutron diffraction at TAKUMI/J-PARC in order to examine the data with further accuracy than the EBSD measurement. The EBSD data were somehow scattered probably due to the sample surface condition. Especially, the 0.62%C bainitic steel tensile strained to a strain of 0.26 shows large difference between the data by EBSD and those by in situ neutron diffraction. This difference should be brought by the difficulty in the EBSD measurement of the largely strained condition. However, both the EBSD and the in situ neutron diffraction data revealed the similar tendency, thus the EBSD data in this work represent not only the surface condition but also the bulk condition.
Fig. 8. Fraction of austenite in 0.15%C (a), 0.25%C (b) and 0.62%C (c) bainitic steels as a function of applied true strain. (a) liner fraction of austenite - liner strain, (b) logarithmic fraction of austenite - line strain. The small circle points indicate the data of the 0.62%C bainitic steel by the measurements of in situ neutron diffraction at J-PARC. (Online version in color.)
Both these two plots (a,b) clearly show that the decreasing at the tension is faster than that at the compression in all the three samples, while the different phenomenological meanings can be drawn from these two plots. The liner - liner plot (a) indicates that the difference between the tension and the compression at the early stage of deformation (around 0.1 strain) becomes larger with the higher carbon content; while in the logarithm - liner plot (b), the data of the samples with different carbon content at the same deformation mode (tension or compression) can be represented by the regression lines with the similar slope. The liner relation at the logarithm - liner plot (b) indicates that the change in the fraction of retained austenite was found in several studies5) and this tendency can be ruled by the following differential equation;
Vγ is the volume fraction of retained austenite and ε is the absolute strain (positive value). This equation means that the volume fraction of the newly formed martensite par strain is proportional to the volume fraction of remaining austenite. The constant k has been evaluated occasionally by fitting to the experimental data with the logarithm - liner plot. Additionally, k has been considered to be proportional to the change in the chemical Gibbs energy in the transformation,
k1 is a constant and ΔGα′γ is the difference of Gibbs energy between the parent and the transformed phases.31) As indicated with Fig. 3, the average carbon contents of the retained austenite were the same among the as-austempered conditions in the three bainitic steels. This means that ΔGα′γ has no difference among all the three samples and it is consistent to the appearance in Fig. 8, where the slope k is constant at the same deformation mode. Consequently, the dependence of the DIMT kinetics on the stress polarity can be considered separately from the effect of the chemical composition.
The relationship between the deformation mode and the DIMT has been discussed by several researchers. Olson and Cohen10) calculated the threshold stress (transformation stress), which is necessary for the occurrence of the DIMT, either at tension or at compression based on the model proposed by Patel and Cohen.32) In their model, the DIMT at compression needs higher stress than at tension due to the volume expansion at the martensitic transformation. The result in this work appears consistent with their model. However, as shown in Fig. 8, it is difficult to find any significant difference of the start stress at which the fraction of retained austenite starts decreasing between at the tension and at the compression. This indicates that the absence of the clear threshold stress which is key parameter in the Orson and Cohen model. Thus, the different mechanism for the effect of stress polarity on the DIMT in the TRIP aided steel should be considered.
The recent works on the the crystallographic orientation dependence13,26,27,28) gives the key to consider one possible mechanism different from the mechanical energy calculation by Olson and Cohen. These studies have reported that the retained austenite grains with the tensile orientation nearly parallel to <111> remains preferably at the tension. For example, Ueji et al.,26) discussed this texture evolution can be explained at the perspective on the mechanism for martensitic transformation originally proposed by Bogers and Burgers.33) This mechanism includes the first primitive dislocation motion of the martensitic transformation of which Burgers vector is parallel to the leading partial dislocation in FCC crystal.34) As is well known, the imperfect slip of the leading partial can be regarded as the controlling process of the deformation twinning as well,34) and the preferable orientation for the deformation twinning by the uniaxial deformation has been regarded as that near <001> at tension and that near <111> for compression.34) This difference should be related to the dependence of the stress polarity on the DIMT, because the different deformation texture evolves at the different stress polarities. As shown in Fig. 7, the compressive deformation provides the texture in which the loading direction turns to <101> which is not preferable orientation for the DIMT. On the other hand, the tensile deformation in FCC crystal brings the two texture-components that are the staying the tensile orientation around <001> and the rotation to <111>. The condition that tensile orientation parallels to <001> is preferable for the DIMT at tension. Consequently, the effect of stress polarity on the DIMT should appears not only due to the mechanical interaction energy between the martensitic expansion and external stress, but also the different texture evolution due to the different deformation mode.
