Abstract
An Fe–13Cr–3.4Mn–0.47C lean-alloy stainless steel was made austenitic by solution annealing at 1250°C. Tensile tests between 20 and 200°C indicated enhancement of ductility at higher temperatures. At 200°C where planar glide, manifested as deformation twinning, was the dominant deformation mechanism, a uniform tensile elongation of 102% was achieved. At 20°C where deformation-induced α′-martensitic transformation replaced deformation twinning as the dominant deformation mechanism, tensile elongation was significantly impaired. The tensile elongation contribution by the planar glide was estimated to be at least four times that of the α′-TRIP (transformation-induced plasticity) mechanism. The results indicate that inexpensive lean-alloy austenitic stainless steels exhibiting pronounced α′-formation at room temperature could become highly formable at higher temperatures.
Introduction
The stacking fault energy (SFE) of fcc metals and alloys, including austenitic steels, decreases at lower temperatures [1]. The associated changes in the glide mode have important consequences for the deformed microstructures; the wide separation of a/6<112>Shockley partial dislocations at low temperatures, where the SFE is low, limits the cross slip and promotes the planar glide of dislocations. As the deformation temperature increases, the reduced separation of partial dislocations and the possibility of constriction enable the cross slip and a transition to the wavy glide mode [2]. Examples of austenitic stainless steel microstructures formed at various deformation temperatures may be found in Ref. [3–9].
The high stacking fault probability at relatively low temperatures can lead to a number of deformation processes collectively termed the by-products of glide planarity. These include deformation twinning, ε-martensite formation, and α′-martensite formation which are commonly activated in that sequence as the temperature decreases [10]. The manner in which stacking faults interact or overlap determines which type of by-product is formed. The tendency of an austenitic steel to form ε-martensite [7], twinning [8] or both [11,12] depends on the alloy chemistry. In high Mn steels, for instance, alloying with Si increases the likelihood of ε-martensite formation at the expense of deformation twinning [13]. The latter is responsible for the twinning-induced plasticity (TWIP) effect. At sufficiently low deformation temperatures, the extremely high dissociation tendency of dislocations enables the formation of α′-martensite to which the transformation-induced plasticity (TRIP) effect is attributed. Nevertheless, a review of the temperature dependence of tensile elongation in austenitic stainless steels reveals that straining at temperatures below
, where the latter mechanism is operative, is associated with the impairment of ductility [5,7,14–23]. Furthermore, the theoretical calculation of the strain due to the α′-TRIP suggests a possible exaggeration of the role of the α′-TRIP effect [24]. Therefore, the plasticity contribution of the α′-TRIP effect needs a re-evaluation.
The present work aims to evaluate the contribution of the α′-TRIP effect to the ductility of austenitic steels by tensile testing of a lean-alloy austenitic stainless steel at a temperature where this mechanism is dominant. This is to be compared with the tensile elongation obtained at a higher temperature where TWIP becomes the dominant deformation mechanism. This experimental comparison is particularly justified in view of the difficulty of obtaining meaningful estimations of the contributions of these effects to the tensile elongation of polycrystalline materials based on the theoretical calculation of the shear contributions due to each mechanism [25]. In particular, any model aiming at the estimation of the elongation contribution due to the α′-TRIP effect must incorporate a realistic picture of the shears involved in the transformation and take into account the variant selection. Furthermore, continuous lattice rotations due to the tensile deformation must be taken into consideration.
Experimental procedure
Chemical composition of the studied steel in wt-%.
Results and discussion
Preliminary dilatometry experiments indicated a progressive decrease in the α′-martensite start (Ms) temperature of the steel upon solution annealing at higher temperatures. Ms temperature after annealing at 1180°C was in the vicinity of room temperature (RT). This is lower than that for a Mn-free variant of the experimental steel [26]. As shown in the optical micrograph of Figure 1, solution annealing at 1250°C resulted in an almost fully austenitic microstructure. Based on magnetic saturation measurements using specimens taken from the grip section, α′-martensite fraction before tensile tests was less than 5 vol.-%. The micrograph in Figure 1 also reveals the presence of a small fraction of undissolved Cr-rich carbides containing approximately 50 wt-%Cr according to energy-dispersive spectroscopy analysis in FESEM.
Optical micrograph of the steel before tensile tests. The sample was etched with the V2A reagent.
Figure 2(a) shows the stress–strain curves obtained at various tensile test temperatures. The corresponding true strain hardening curves are shown in Figure 2(b). Serrated flow occurred during tensile tests particularly at 20 and 80°C. Serrated flow during RT tensile deformation of an Fe–17Cr–7Mn–4Ni–0.05C–0.05N stainless steel has been attributed to the diffusion of interstitial atoms in the martensitic constituent [27]. Owing to the absence of serrated flow in the Mn-free variant of the present steel with an initial α′-martensite fraction of nearly 44 vol.-% [26], this mechanism does not hold for the present case. In high Mn steels with a low SFE (high glide planarity), serrated flow due to the dynamic strain aging is attributed to the destruction of C–Mn pairs due to the motion of Shockley partial dislocations. The latter converts octahedral sites into tetrahedral sites thereby necessitating the relocation of carbon to a neighbour octahedral site, a process which can occur locally within Portevin–LeChatelier (PLC) bands [28]. The presence of Mn coupled with the high planarity of glide in the present alloy, particularly at lower tensile test temperatures, makes the above mechanism also conceivable for the present alloy. As the tensile test temperature increased, a significant enhancement of elongation took place. The highest uniform tensile elongation of approximately 102% was obtained at 200°C. According to Figure 2(b), at stress levels above approximately 500 MPa, the strain hardening rates were almost constant for the tensile test temperatures of 160 and 200°C, where the highest elongations were achieved. This is in agreement with the observations made for high-alloy austenitic stainless steels where near-linear hardening resulted in the highest tensile elongations [8]. At both 160 and 200°C, the product of uniform elongation and ultimate tensile strength was in excess of 93 000 MPa% which is superior to that of conventional Fe–Cr–Ni austenitic stainless steels [14,29,30]. The excellent ductility of the steel at these temperatures makes it a potential candidate for warm forming operations.
(a) Engineering stress–strain curves at various temperatures; (b) true strain hardening curves corresponding to (a); (c) temperature dependence of uniform tensile elongation and deformation-induced α′-martensite fraction.
The temperature dependence of uniform tensile elongation is summarised in Figure 2(c). It resembles the behaviour exhibited by austenitic steels at temperatures below their respective
temperatures [5,7,14–23]. Accordingly, magnetic saturation measurements confirmed the occurrence of deformation-induced martensitic transformation during tensile tests. The deformation-induced α′-martensite fractions in Figure 2(c) represent the difference in the ferromagnetic phase content of 3 mm-thick cylindrical specimens taken from the gauge and grip sections of deformed tensile specimens. Owing to the lower stability of austenite at lower tensile test temperatures (lower SFEs), higher fractions of deformation-induced α′-martensite formed at lower temperatures. Furthermore, these higher fractions of α′-martensite were induced by the application of significantly lower tensile strains. The faster kinetics of α′-martensite formation at lower temperatures is also reflected in the higher rates of strain hardening at lower temperatures. Therefore, the low ductilities at 20 and 80°C can be attributed to the formation of α′-martensite at small strains. As the tensile test temperature increased, the formation of α′-martensite was postponed to higher strains which enhanced the ductility. The results clearly indicate that the deformation-induced formation of α′-martensite has a detrimental effect on the ductility of austenitic steels. The loss of ductility upon deformation-induced formation of α′-martensite is in line with numerous studies involving tensile testing of austenitic steels at various temperatures [5,8,14–23].
According to the preceding discussion, the high tensile elongations at 160 and 200°C must be justified by mechanisms other than the TRIP effect. To identify the deformation mechanisms operative at the studied tensile test temperatures, the microstructures of tensile specimens deformed at 20 and 200°C were examined by LOM and ECCI (Figure 3). The specimens were taken from the vicinity of the fracture surfaces and the planes of polish were normal to the tensile direction. Both microstructures exhibit a high density of nearly straight features which highlight the traces of {111}
γ
glide planes. The prevalence of nearly straight features indicates the high glide planarity and the low SFE of the alloy even at the highest tensile test temperature of 200°C. The higher curvature and the denser packing of such features at 200°C (Figure 3(a)–(c)) can be justified by the applied tensile strain of nearly 100%. The coarse planar features in the specimen tested at 200°C (marked by arrow in Figure 3(b)) were confirmed by EBSD (Figure 4(a)) to be deformation twins. Many of the finer planar glide features resolved by ECCI could not be resolved by EBSD. Nevertheless, due to the absence of ε-martensite peaks in the XRD pattern of the specimen tested at 200°C (Figure 4(b)), the finer features in the ECCI micrographs too are expected to be deformation twins. The XRD pattern of Figure 4(b) further indicates that a high density of grains are oriented with their <111>
γ
and <001>
γ
crystal axes parallel to the tensile direction (<111>
γ
and <001>
γ
fibres, respectively). This is in agreement with observations made for low SFE austenitic steels subjected to tensile straining or wire drawing [31–34]. According to viscoplastic self-consistent crystal plasticity calculations, <111>
γ
and <001>
γ
fibres are stable orientations under tensile loading conditions [34,35]. As the EBSD inverse pole figure map in Figure 4(a) shows, the high intensity of the <001>
γ
fibre is also due to the occurrence of deformation twins in grains oriented close to the <111>
γ
fibre. Twins formed by tensile loading along the <111>
γ
crystal direction will have their <115>
γ
axis oriented parallel to the tensile direction, the latter being only 16° misoriented from the <001>
γ
direction.
Light optical (a,d) and ECCI (b,c,e,f) micrographs of tensile specimens deformed at the indicated temperatures. Planes of polish are normal to the tensile direction. Samples for LOM were etched with the Beraha-I reagent. EBSD inverse pole figure maps (a,c) and XRD patterns (b,d) of tensile specimens deformed at the indicated temperatures. Planes of polish are normal to the tensile direction. Colours in the EBSD maps indicate crystal directions of austenite parallel to the tensile direction. To highlight deformation twins in (a), boundaries with misorientations within 60 ± 3° are shown in white. Grey in (a) and (c) denotes α′-martensite or data points with confidence indices below 0.1. In (c), ε-martensite is highlighted in black.

The phase contrast due to the α′-martensite which is mainly localised within the glide bands is readily noted in the micrographs of the specimen strained 7% at 20°C (Figure 3(d)–(f)). The tensile test temperature of 20°C is almost equal to the Ms temperature of the alloy below which the α′-martensite formation can occur spontaneously by the interaction of partial dislocations [36,37]. Therefore, an extremely low SFE is expected for the alloy at 20°C. Owing to discrepancies regarding the effect of alloying elements on the SFE [38–40], no calculated SFE value is reported here. The low SFE at 20°C implies that the glide bands mainly consist of stacking faults and the by-products of partial dislocations. This is especially likely for grains belonging to the <111> γ fibre where the Schmid factor for the primary Shockley partial glide system is higher than that for the primary perfect dislocation glide system (0.31 vs. 0.27). In contrast to the specimen deformed at 200°C, deformation twins could not be identified by EBSD (Figure 4(c)). Instead, in addition to α′-martensite, a small fraction of ε-martensite was detected by EBSD. The occurrence of ε-martensite was further confirmed by the XRD (Figure 4(d)). The appearance of ε-martensite peaks in XRD patterns is argued to be due to the high density of stacking faults in glide bands [41]. A similar transition in the type of deformation-induced products from twinning to ε-martensite by a reduction in temperature occurs in many austenitic steels [5,10,42].
The early formation of α′-martensite during tensile straining, originating from an extremely high glide planarity, is the main reason behind the low tensile ductility at 20°C. In particular, the relatively high carbon content of α′-martensite decreases its ductility. As a result, the present lean-alloy austenitic stainless steel shows a higher sensitivity to embrittlement caused by the deformation-induced α′-martensite formation than conventional austenitic steels with lower interstitial contents [43,44]. At higher temperatures where glide planarity is still high but not sufficient to cause the early formation of α′-martensite, on the other hand, promising tensile properties are obtained.
The tensile elongation of the present polycrystalline steel at 20°C enables to make an upper bound estimate of the α′-TRIP contribution to the tensile ductility of austenitic steels. This can be done by assuming that the tensile elongation of 7% is entirely accommodated by the TRIP effect due to the formation of 29 vol.-% α′, namely by neglecting dislocation activities not resulting in α′-martensite formation. This yields a tensile elongation of approximately 24% for the formation of 100 vol.-% α′. This is approximately a quarter of the tensile elongation at 200°C where plasticity is mainly accommodated by dislocation activities other than the α′-martensite formation, most notably by deformation twinning. The results emphasise that as long as ductility is the main requirement, a high glide planarity without α′-martensite formation is desirable. In other words, the role of the α′-TRIP effect as a plasticity mechanism is exaggerated [45]. Accordingly, the knowledge of the chemical composition dependence of the
temperature could be used as a guide in the design of ductile austenitic steels.
Conclusions
An Fe–13Cr–3.4Mn–0.47C lean-alloy stainless steel was made almost fully austenitic by solution annealing at 1250°C. Tensile tests in the temperature range of 20–200°C indicated enhanced ductility at higher temperatures. At 200°C where deformation twinning (TWIP effect) was the dominant deformation mechanism, a uniform tensile elongation of 102% was achieved. The product of uniform elongation and ultimate tensile strength was 93 000 MPa% which is superior to the RT values for conventional Fe–Cr–Ni and Fe–Cr–Ni–Mn austenitic stainless steels. At 20°C where the deformation-induced α′-martensitic transformation (α′-TRIP effect) replaced twinning as the dominant transformation mechanism, on the other hand, tensile elongation was significantly impaired. An optimistic estimate of the tensile elongation contribution by the α′-TRIP yields a value almost a quarter of that due to the occurrence of planar glide features other than the α′-martensite formation, most notably deformation twinning.
