1. Introduction
Ausforming processes are those thermomechanical treatments in which austenite is plastically deformed before either a martensitic or a bainitic transformation takes place.1,2) Although the deformation of an austenitized steel has many benefits, such as the reduction of the Ms temperature,3,4,5,6) the refinement of the subsequently formed bainitic ferrite (αB),7,8) the acceleration of phase transformations9) and the increase of the content of carbon-enriched retained austenite (γ+),10,11) the deformation of austenite at intermediate temperatures can also induce displacive stress or strain induced transformations, even at temperatures at which they usually do not appear.12,13,14) When ausforming treatments with deformation temperatures in the bainitic range are intended, especially at the industrial level, these stress and strain induced transformations have been shown to be unavoidable.3,15) Since they cannot be avoided, the possibility of using these types of phase transformations as a way to tune the microstructure and obtain better properties arises.
The effect of stress of the bainitic transformation has been studied during the last decades up to some extent. With respect to the effect of stress on transformation kinetics, it has shown to accelerate the transformation,16) regardless of whether stresses are below or above the austenite yield strength, because of the increase in driving force1,17) and the increase in nucleation sites, if the applied stress is higher than the austenite yield strength.1) Moreover, stresses have shown to give rise to transformation plasticity strain, i.e. there are different changes in length associated to the bainitic transformation, depending on the observation axis.17,18,19,20,21) This transformation plasticity is due to the preferential selection of variants promoted by the applied stress.19,20,21,22,23,24,25) However, one question that has not been answered yet, to the best knowledge of the authors, is what the effect of stress on the resultant microstructure, in terms of bainitic ferrite plate thickness and fraction of bainitic ferrite formed, is.
Bainitic ferrite plate thickness has been shown to be a key factor to improve the strength of the final microstructure,26,27,28,29) mainly affected by the austenite yield strength at the transformation temperature,30,31) and hard impingement events. Hard impingement events are dependent on the prior austenite grain size (PAGS)32,33) and the transformation rate, where the latter one is associated to the driving force for the transformation of austenite into ferrite.30,31) No information on the thickness of the αB plates formed while austenitized samples are strained, as compared to a stress-less structure, has been found up to date. With respect to the effect of a constant stress on αB plate thickness, constant stresses below the yield strength applied during isothermal holding in the bainitic range have shown to lead to way coarser bainitic structures than the structures formed in the absence of stress.34,35) While Su et al. explained this phenomenon based on the favorable formation of crystallographic variants, whose associated slip dislocation systems had preferential activation, promoting the accommodation of transformation strains,35) Pak et al.34) showed that coalescence occurred when applying a constant stress to the structure, during the isothermal holding.
Another very relevant parameter, when it comes to mechanical properties of bainitic microstructures, is the content of retained austenite, which can transform to martensite (α’) during service by Transformation Induced Plasticity (TRIP) effect.36,37) Thermal and mechanical stability of such austenite is strongly affected by the stiffness of the surrounding matrix (fraction and size of bainitic ferrite plates) and its chemical composition, dependent on the volume fraction of bainitic ferrite formed. Shipway and Bhadeshia conducted a study in which they assessed the effects of constant stresses below the austenite yield strength on bainitic transformation in a Fe-0.45C-2.08Si-2.69Mn (wt.%) steel.17) They presented results as a function of the dilatometric dilation, associating this variable to the volume fraction of formed bainitic ferrite. However, this assumption is not necessarily valid, as transformation anisotropy can alter the dilatometric signals, as reported in the same work. Although it was reported that there was a small increase in bainitic ferrite volume fraction as stresses increased, this trend was not quantified by techniques such as X-Ray Diffraction. The effect on the volume percentage of bainitic ferrite under stresses above the yield strength is not clear. Although previous results have shown that the introduction of dislocations prior to the transformation mechanically stabilizes the austenite against displacive transformations,10,11,38) those works studied the influence of previous deformation in a stress-free state. Further work is needed to elucidate how constant stress above the yield strength affects the volume percentage of bainitic ferrite and the stability of residual austenite.
This work aims to better understand the effect of stress on the bainitic transformation and its resultant microstructures, considering bainitic transformations happening while deformation happens, as well as those phase transformations occurring under constant stress, both below and above the yield strength, i.e. stress in the elastic regime (SER) and stress in the plastic regime (SPR). Experiments were carefully designed to address the mentioned unanswered questions and conducted in a dilatometer. Microstructures were characterized by Scanning Electron Microscopy (SEM) and X-Ray Diffraction (XRD).
2. Experimental
In this study, a commercial medium C (0.4 wt%) high Si (3 wt%) steel - Sidenor’s SCM40 steel - alloyed with other substitutional elements, such as Cr, Mo and Mn (5.2 wt.% in total), was used. This steel was selected for its high Si content, which ensures that the bainitic transformation is essentially carbide free.39) 10 mm long cylindrical specimens with 4- and 5-mm diameters were used depending on whether they were used for pure dilatometry or for deformation tests. The remaining substitutional elements were added for hardenability purposes, to ensure that the steel can reach the target temperature after austenitization before any phase transformation starts.
Thermal/thermomechanical treatments were carried out in a Bahr 805D high-resolution dilatometer. This piece of equipment enables to monitor the longitudinal changes in the sample length, to later associate them to phase transformations. An induction heating coil is used to heat up the specimens while, for cooling, Helium is blown directly into the specimen, where the temperature changes are monitored by a K type thermocouple welded to the central part of the sample surface. Two different specific modules were used: a dilatometer module with fused silica push rods for experiments without deformation, and a compression unit for tests involving deformation, where Si3N4 deformation tools separated from the sample by Mo films, to reduce friction and to increase the temperature homogeneity in the specimens.
Dilatometric data was used to determine the martensite start and finish temperatures, Ms and Mf respectively, on quenching experiments (Q) as well as on cooling to room temperature after isothermal treatments (ISO). For the ISO tests, dilatometric data was used to determine the start and end of the bainitic transformation, following the procedures described in Ref.40)
XRD analysis was used to estimate the volume percentage of retained austenite (FCC phase) and bainitic ferrite/martensite (BCC/BCT phases). Samples were step-scanned in a Bruker D8 Advance diffractometer with Bragg-Brentano geometry, equipped with a graphite monochromator and a Vantec position sensitive detector. The diffractometer has a rotating Co anode X-ray tube as a radiation source and its current and voltage were, respectively, 35 mA and 45 kV. The measurements were made by a coupled θ-2θ scan in the range 35–135°, with a step size of 0.026° and a counting time per step of 2 s. Retained austenite content was determined by comparing the areas under the BCC/BCT peaks {110}, {200}, {211} and {220} with the areas under the FCC peaks {111}, {200}, {220} and {311}, assuming that the microstructure consisted of only two phases.41) Using several peaks avoids non-accurate results related to crystallographic texture.42)
The central part of the samples, where the local plastic strain was maximum,43,44) was observed under a field emission gun scanning electron microscope (FEG-SEM). Both the transverse and the longitudinal sections were observed to identify possible anisotropy effects in the development of the microstructure.45)
Samples were metallographically prepared by standard metallography and the microstructure was revealed by etching with a 2% Nital solution for approximately 20 seconds. In the particular case of the XRD samples, surface preparation involved an additional set of etching and polishing cycles to remove the plastically deformed layer.
The bainitic ferrite plate thickness, t, was measured according to the procedure described in Ref.,46) and about 200 linear intercepts were measured on SEM micrographs to obtain reliable statistics. A stereographic correction was subsequently applied to the mean linear intercept (
L
α
¯
),
t=2
L
α
¯
/π
with a 95% confidence error
E=±1.96⋅
σ
L
α
/π
N
, where
σ
L
α
symbolizes the standard deviation of the intercepts and N is the number of measurements. Although this method may be inaccurate for anisotropic microstructures,45) if measurements on several sections are considered, it can be helpful to obtain a good estimate of the microstructure refinement level.
Vickers hardness measurements were made on the transverse section using a load of 10 kg, where at least three different measurements were performed.
3. Results and Discussion
3.1. Thermomechanical Treatments Design
This work aims to characterize bainitic microstructures obtained at fixed temperatures, under different stress states: stress-free – pure isothermal treatments (ISO), during the application of stress up to the elastic and plastic ranges – stress in the elastic/plastic range, i.e. SER/SPR - and under constant stresses in the elastic and plastic ranges. To compare these microstructures, different dilatometry experiments were designed. The first of them was an ISO treatment, in which samples were austenitized at 5°C/s up to 990°C and held for 4 minutes before cooling at 15°C/s to the selected isothermal temperature (TISO), which was maintained for 2 hours. Finally, all samples were cooled down to room temperature at 25°C/s. Refer to Fig. 1(a) for a sketch of the described ISO treatments. To study stress/strain induced transformations, i.e. transformations formed while deforming the prior austenite in the elastic/plastic regime, treatments SER/SPR+Q in Fig. 1(b) were conducted. Samples were austenitized at the same temperature as the ISO treatments and cooled down to a deformation temperature TDEF equal to TISO, temperature at which they were deformed up to ε at a strain rate
ε
˙
of 0.15 s−1 before finally quenched to room temperature. Finally, to study the bainitic transformations formed under constant stresses, in the elastic and plastic range, once a given deformation ε was applied, the stress was sustained during the whole isothermal treatment, after which it was removed and the sample cooled to room temperature, SER/SPR+ISO experiments in Fig. 1(c).
Fig. 1. Sketches showing the temperature/stress vs. time evolution during (a) quenching, Q, and isothermal, ISO, treatments; (b) single hit compression, SHC, tests, stress in the elastic/plastic regime and quenching, SER/SPR+Q, treatments and (c) stress in the elastic/plastic regime and isothermal, SER/SPR+ISO, treatments. TISO, TDEF, ε and
ε
˙
stand for isothermal temperature, deformation temperature, strain and strain rate, respectively.
The selection of the TISO=TDEF and ε parameters was made on the basis of results from a previous work.47) The bainitic temperature range, was determined by ISO type treatments, Fig. 1(a), with TISO in the range 300–600°C, in steps of 50°C, leading to the construction of the Temperature-Time-Transformation (TTT) diagram shown in Fig. 2(a). If needed, Ms and Mf temperatures were also determined on final cooling after these ISO experiments, see Figs. 2(a) and 2(c). These values were compared to the bulk Ms and Mf temperatures (280°C and 113°C, respectively), determined by quenching experiments (Q) of the type shown in Fig. 1(a). Illustrative examples of the different microstructures and their corresponding HV10 values can be found in Figs. 2(d)–2(g) and 2(b) respectively.
The as-quenched microstructure, see Fig. 2(d), was fully martensitic (α’), with a hardness of 744 HV10, Fig. 2(b). This microstructure was very similar to the ones obtained after the ISO treatments in the TISO range of 500–550°C, e.g. the one in Fig. 2(f), as no bainitic transformation occurred during the isothermal holding. This revealed that the Bs temperature was within this range. Any microstructure formed above 550°C consisted of ferrite (α) and cementite (θ), see Fig. 2(e), while any microstructure formed between 280°C and 450°C, e.g. Fig. 2(g), consisted of bainitic ferrite and carbon enriched austenite (αB+γ+) and, above 400°C, a small fraction of martensite formed during final cooling, as evidenced by dilatometry. The reason why microstructures formed about the Bs temperature do not show the expected pearlitic colonies may be that isothermal time was long enough to start a spherization of the structure,48,49) although further research would be required to clarify this point. The minimum value of hardness was achieved for the ISO treatment at 400°C, as can be seen in Fig. 2(b). This low hardness is most likely the product of the combination of a low fraction of martensite (note that the Ms temperature for this treatment is about 75°C, Fig. 2(c)) and a low fraction of coarse bainitic ferrite.1) Further information about these treatments and microstructures can be found in Ref.47) Based on this TTT diagram, it was decided to select three different TDEF = TISO of 550°C - in the hiatus between the pearlitic and the bainitic regions - 400 and 350°C, both of them in the bainitic range. Based on previous works in the same steel with similar conditions, the application of stress should only induce the formation of αB at these selected temperatures, and always in the plastic range.12)
The last critical parameter to be determined for this study was the strain ε values corresponding to the elastic and plastic regimes at the different selected TDEF (= TISO: 550, 400 and 350°C). To this end, single hit compression tests (SHC), schematized in Fig. 1(b), were performed, using the same process parameters already established for previous tests. The goal of these experiments was to observe the compression behavior of austenite to then select appropriate strain values for the subsequent experiments. Compression was applied until the limit of the dilatometer load cell was reached, at which point, for security reasons, the dilatometer stopped and the samples naturally cooled down. Figure 3 includes the true stress – true strain and the strain hardening rate (derivative of the true stress with respect to the true strain) – true strain curves corresponding to each TDEF. Strain hardening rate can be a good indicator of the presence of phases induced by straining an austenitic structure, as discussed in Ref.12,50,51) While the strain hardening rate – true strain curve keeps decreasing with true strain for TDEF = 550°C, this is not the case for the remaining two TDEF, at which the strain hardening rate reached a minimum point, after which its value stabilized. This trend suggests that a transformation was induced during the deformation step at those TDEF.12,50) Based on these results, a very low elastic deformation (needed to reach 150 MPa) and a plastic deformation of 20% (true strain of 0.223) were selected for the subsequent experiments.
Once treatment design was finished, ISO, SER/SPR+Q and SER/SPR+ISO treatments, as sketched in Fig. 1, were conducted. Figure 4 gathers the change in length vs. time curves corresponding to the isothermal holding steps during the ISO and SER/SPR+ISO treatments. The curves corresponding to the ISO treatments at temperatures of 400 and 350°C, present the typical expansion with a sigmoidal shape, characteristic of the isothermal bainitic transformation, where, after a certain period of time, the transformation starts and rapidly proceeds, slowly reducing its transformation rate as it advances, up to the point where the transformation stops and a plateau is reached. Under isotropic conditions, larger changes in length are related to higher fractions of bainitic ferrite,40) which is consistent with the higher signal at 350°C.1) Finally, as expected, isothermal holding at 550°C did not show any change in length variation during the whole treatment, as no phase transformation was expected to occur in the hiatus between the pearlite and the bainite regions, see the TTT diagram shown in Fig. 2(a).
Fig. 4. Change in length vs. time during the isothermal holding at (a) 550°C; (b) 400°C and (c) 300°C, including insight showing the first 500 s.
Regardless of the TDEF, the application of stress in the elastic and plastic regime at a constant temperature, during the SER+ISO and SPR+ISO treatments, respectively, led to a negative change in length. Such a behavior has been previously explained in terms of transformation anisotropy.52,53) At this point, it is important to note that, during the application of stress, anisotropic changes in length measured by the dilatometer are the summation of two contributors, i.e. anisotropic changes in length due to the Greenwood and Johnson mechanism - micromechanical plastic strain in the austenite from the expansion associated to the bainitic ferrite formation -53) and the anisotropic change in length associated to the often referred to as Magee effect,54) where the formation of specific crystallographic variants of bainitic ferrite, promoted by the applied stress, lead to anisotropic bulk transformation strains.17,20,21) While both mechanisms operate for displacive transformations,52,55,56,57) the latter one has been shown more relevant, according to several studies.57,58,59) On the other hand, reconstructive transformations, such as the pearlitic one, possible at 550°C in this study, have also shown transformation plasticity associated to the former mechanism.56,60,61) In addition, it is to be noted that the application of stresses above the yield strength led to lower change in length signals than those recorded during the application of elastic stresses, which could be related to: a) a higher degree of phase transformation (most likely lower bainite, although other transformations, such as pearlitic transformation, could be expected at 550°C) and/or b) a stronger transformation plasticity, due to either of the previously mentioned factors.
The fact that a change in the dilatometric signal was detected for the SER/SPR+ISO treatments at 550°C comes to confirm that stresses were able to shift the bainite start temperature to higher values.12) Moreover, the fact that the signal decreases constantly with time at 550°C indicates that 2 hours was not enough time to reach the end of the transformation. One more time, the magnitude of the signal during this treatment indicates that, either a significant fraction of bainitic ferrite formed during the isothermal step at 550°C, or the level of anisotropy at this temperature was remarkedly high. Given the proximity to the Bs temperature, the second option is the most probable. For the remaining temperatures, 400 and 350°C - Figs. 4(b), 4(c), results confirm that stresses also accelerated the bainitic transformation, given the shorter starting and finishing transformation times, as compared to results obtained for ISO treatments, in good agreement with the literature.16)
Figure 5 shows the transverse SEM micrographs corresponding to the conducted thermal and thermomechanical treatments as a function of TDEF=TISO. Each micrograph is accompanied by a text box, where the phases present in each microstructure are listed. For the conditions where bainitic ferrite is formed, values for bainitic ferrite plate thickness (measured on the transverse and longitudinal sections – tT and tL, respectively), retained austenite content, Ms and Mf temperatures and HV10 are included in Fig. 6.
Fig. 5. Transverse SEM micrographs corresponding to the following treatments (a–c) ISO; (d–f) SER+Q; (g–i) SPR+Q; (j–l) SER+ISO and (m–o) SPR+ISO with TDEF=TISO temperatures of (a, d, g, j, m) 550°C; (b, e, h, k, n) 400°C and (c, f, i, l, o) 350°C. α’ refers to martensite, α’COAL is coalesced martensite, αB stands for bainitic ferrite, αW is Widmanstätten ferrite, α+θ to pearlite and γ+ refers to retained austenite.
Fig. 6. (a, b) Bainitic ferrite plate thickness measured on the transverse (tT) and longitudinal (tL) sections; (c) retained austenite content; (d, e) Ms and Mf temperatures and (f) HV10 values as a function of TDEF=TISO for the performed heat treatments. The values corresponding to the quenching (Q) treatment are represented by a horizontal dashed line. (Online version in color.)
The microstructures resultant of the ISO treatments - Figs. 5(a)–5(c) - were bainitic microstructures, except at TISO = 550°C, temperature at which no bainitic transformation took place, and the microstructure was hence martensitic and similar to that obtained in the quenching experiment, as confirmed by its Ms and Mf temperatures, as well as by its HV10, see Figs. 6(d)–6(f). The microstructure formed at 400°C also contained a small fraction of martensite, as elucidated by its Ms of 125°C, as can be observed in Fig. 6(d). An increase in the TISO from 350 to 400°C led to an increase in bainitic ferrite plate thickness and retained austenite content, as shown in Figs. 6(a), 6(b), 6(c), as well stated in the literature.46,62)
The microstructures resultant of the SER/SPR+Q treatments are shown in Figs. 5(d)–5(i). The aim of these treatments was to characterize the αB formed during the applied deformation. It is evident that applying stress below the yield strength to the austenitic structure did not promote significant bainitic transformation, in good agreement with previous studies in this same steel.12) Only a few plates of bainitic ferrite formed during straining at 400 and 350°C, see Figs. 5(e), 5(f). Although the measured average thickness of these bainitic ferrite plates was smaller than the corresponding values obtained for the ISO microstructures, Figs. 6(a), 6(b), these values should be handled with caution as the number of measurements used is low and therefore the uncertainty in their value is high. Other than those few bainitic ferrite plates and a high fraction of martensite laths/plates, the microstructures resultant of these SER+Q treatments showed coalesced martensite (α’COAL), see example highlighted in Fig. 5(d). Coalesced martensite has been previously characterized by several authors.63,64) Consistently with the results obtained in this work, Pak65) showed that applying stress to a structure increases the propensity for martensite plates to combine. With respect to the retained austenite content, the Ms/Mf temperatures and the hardness values obtained for these SER+Q treatments, Figs. 6(c)–6(f) shows how they barely changed with respect to the as-quenched microstructure.
Stress above the yield strength applied at 550°C during the SPR+Q treatment did not induce any type of transformation, other than the coalescence of martensite, previously discussed. On the other hand, stresses in the plastic regime applied at 400 and 350°C induced a profuse bainitic transformation. It is worth mentioning that the bainitic ferrite plates nucleated at prior austenite grain boundaries, in good agreement with the literature,1) and subsequently grew in parallel to each other, as shown in Figs. 5(h), 5(i), most likely indicating a variant selection phenomena. The volume percentage of induced bainitic ferrite cannot be assessed by XRD analysis and, based on the micrographs, see Fig. 5(h) vs. (i), it is difficult to say if more or less bainitic ferrite is formed at a specific temperature. However, the fact that the microstructure stressed at 350°C has a lower retained austenite content, indicates that, most likely, a higher fraction of bainitic ferrite is formed at that temperature. Moreover, the Ms and Mf values determined during final cooling and included in Figs. 6(d), 6(e) can give further information about the microstructure. The first detail to point out is that Ms temperature barely changed with respect to the bulk Ms (Q treatment). This is most likely due to the previously mentioned heterogeneous distribution of bainitic ferrite plates, which predominantly formed on the prior austenite grain boundaries, leaving the central part of the prior austenite grains untransformed. These central areas were not transformed and were far from the prior austenite grain boundaries, where bainitic ferrite plates formed and partitioned carbon to their surroundings. Because of that, martensite is formed in these areas most likely transformed at a temperature very close to the bulk Ms. If one looks at the Mf temperatures for SPR+Q treatments in Fig. 6(e), one can appreciate that Mf temperatures considerably increased with respect to the bulk Mf temperature. Once again, this can be explained by the described heterogeneity, as the martensitic transformation initiated in the center of the prior austenite grains rapidly consumed the unstable austenite, only leaving the residual austenite in the proximity of the bainitic ferrite plates at the prior austenite grain boundaries. Residual austenite was most likely stable in those regions, due to its carbon enrichment, reason why it did not transform to martensite during final cooling. With respect to the size of the induced bainitic ferrite plates, one can observe in Figs. 6(a), 6(b) how plates grew finer than the ones formed under stress below the yield strength in the homologous treatment, and how a coarser microstructure was obtained at 400°C, as compared to the one corresponding to 350°C. Finally, it is to be noted that the microstructure subjected to stress above the yield strength at 350°C has a higher HV10 value than the microstructure subjected to SPR+Q at 400°C, most likely due to the finer bainitic ferrite and the lower retained austenite content. Therefore, these results suggest that bainitic ferrite plates formed at a given temperature while deforming an austenitized structure up to the elastic or plastic regime are finer than the ones that would form in the absence of stress, being the effect of temperature similar, in terms of bainitic ferrite plate thickness and volume percentage, than the effects that have been reported for stress-less transformations.62)
Lastly, SER/SPR+ISO microstructures were characterized. Note that these microstructures contained bainitic ferrite formed during the deformation step, if any, and previously characterized by the SER/SPR +Q treatments, and bainitic ferrite formed under constant stress. Only a few coarse bainitic ferrite plates were induced under stress in the elastic regime at 550°C, regardless of the stress level, as can be seen in Figs. 5(j), 5(m), barely affecting the retained austenite content, Ms/Mf temperatures and hardness levels, see Fig. 6. The microstructures formed under stress in the plastic regime at 550°C were microstructures consisting different phases: pearlite (α+θ), ultra-fine bainitic ferrite, granular bainite (αGB), Widmanstätten ferrite (αW), martensite and coalesced martensite. A complete identification of these phases can be found in the Supplementary Material. Note that the first four named structures (α+θ, αB, αGB, αW) typically form at temperatures around the hiatus, hence local deformation and/or chemical variations could promote the formation of all of them at the same time. A better understanding of the transformations at this temperature and stress level is required, although in-situ techniques may be required. The retained austenite content was not affected by these transformations, as compared to the as-quenched microstructure, whereas the Ms/Mf temperatures and the hardness levels were only slightly decreased, see Fig. 6. As for microstructures formed at lower temperatures, whereas the microstructures formed under stresses below the yield strength presented weak variant selection at first glance, as most of the plates were randomly oriented – see Figs. 5(k), 5(l) -, microstructures formed under stresses in the plastic regime and shown in Figs. 5(n), 5(o) showed a fair parallelism among plates that suggests that strong variant selection phenomena took place, in good agreement with the previously shown dilatometry results and with other works in the literature.17,18,19,20,21,22,23,24,25) Deeper crystallographic studies would be needed to confirm variant selection. XRD results included in Fig. 6(c) confirmed that these specimens present a lower fraction of austenite than the ISO microstructures. The effect of the temperature was kept, regardless of the treatment, as higher austenite content was detected at 400°C. No martensite was formed during final cooling for either of the treatments at 400 and 350°C, confirming that the bainitic ferrite volume percentages formed under constant stresses were higher than the ones obtained for ISO treatments and that the stability of the residual austenite was consequently higher. These results evidence that the introduced mechanical driving force shifted the T0 curve to higher carbon content values. This would imply that the transformation was still thermodynamically viable until the austenite reached a higher carbon content, thus leading to a higher percentage of bainitic ferrite. Regarding bainitic ferrite plate thickness, it is evident that those plates formed under stress in the plastic regime were coarser than the ones that were formed during the deformation step, and those formed under a stress-less state in the ISO treatments, see Figs. 6(a), 6(b). The effect of temperature was kept, i.e. the higher the temperature, the coarser the structure, for a homologous treatment. As previously mentioned, studies in the literature have already reported that constant stresses applied during isothermal holding in the bainitic range lead to coarser bainitic structures than the structures formed in the absence of stress.34,35) The explanations given for this observation were: a) the favorable formation of crystallographic variants with promoted accommodation of transformation strains35) and b) bainitic ferrite plates coalescence.34) No coalesced plates were detected in the present work, which indicates that the formation of specific crystallographic variants could be the plausible hypothesis. Moreover, as has already been noted, SPR+ISO treatments potentially led to stronger variant selection, which would be in good agreement with why plates are coarser for those specific conditions.
This hypothesis could also be supported by a deeper look at the factors affecting plate thickness, as defined by Singh and Bhadeshia30) and revised in Ref.46,66) According to the cited works, plate thickness mainly depends on austenite yield strength and austenite to ferrite driving force for bainite transformation, in descending importance. Previous calculations have shown how the mechanical driving force added by a constant stress is higher in absolute value than the driving force reduction by the introduction of dislocations during plastic deformation.12) Therefore, either a stress in the elastic or plastic regime increases the driving force for the transformation, in good agreement with the reported accelerated kinetics. A driving force increase would then be associated to a plate thickness reduction, in disagreement with the evidence. The reason why plate thickness is coarsened under stress must then rely on the other term, the austenite yield strength and the accommodation of plastic deformation, as suggested by Su et al.35) One could then wonder why plates formed during straining are refined (SER/SPR+Q treatments), if they also show variant selection. The answer could be related to: a) the characteristics of variant selection during the transformations happening in those two different scenarios did not follow the same rules, which led to different variants being formed, i.e. different plastic accommodation; b) the interaction of bainitic ferrite plates growth and dislocation formation during straining led to refinement, while this interaction was not so strong during isothermal holding, as not so many new dislocations formed during the transformation and as isothermal holding led to recovery.
With respect to the hardness of the resultant microstructures, both retained austenite content and plate thickness, as well as plastic deformation in the case of the SPR+ISO treatments, affect HV10 values, see Fig. 6(f). As can be observed, microstructures formed under stresses above the yield strength are harder than the isothermally treated ones, even if their average plate thickness was much coarser. Such hardening, as compared to the samples transformed under stress in the elastic regime, suggests that plastic deformation played a significant role in that sense.
Therefore, these results show that the application of a constant stress at a constant temperature and while the bainitic transformation occurs promotes coarsened microstructures, most likely associated to variant selection and plastic accommodation in the residual austenite, with larger percentages of bainitic ferrite, due to the driving force increase and the shift of the T0 curve to higher carbon contents. Temperature effects expected for stress-less transformations, i.e. higher temperature, coarser bainitic ferrite with larger austenite content, is kept regardless of the treatment.62)
