X-ray Diffraction
The diffraction patterns of AR, WR 60%, and WR 80% are shown in Fig. 4, and the quantitative results of the fraction volume phases are summarized in Table 2. The results confirm the presence of austenite and ferrite in all the samples. There is a slight difference between the warm-rolled conditions when compared with the as-received. The volume fraction of α-ferrite in the WR 60% decreased by a small amount compared with that in the as-received condition. In WR 80%, the fraction of α-ferrite increased. It should be noted that the precision of the method does not allow for major conclusions regarding the phase variation. As expected, no deleterious phase was observed.
Table 2 Phase volume fraction calculated using X-ray diffraction data Scanning Electron Microscopy
Figure 5 shows SEM micrographs of the AR, WR 60%, and WR 80% samples. The gray color difference between these phases is a result of their different corrosion resistances during etching, where ferrite is the darkest and most corroded phase. An elongated morphology of alternative bands is noticeable for ferrite and austenite along the RD in the deformed structure. These lamellae appear as approximately parallel-lined grains from the early stages of deformation.
In Fig. 5b and c, it is also possible to see that the bands or lamellae of the phases (α + γ) became wavy because of the intense warm deformation imposed on the steel. The refinement of the microstructure because of the thickness reduction of ferrite and austenite lamellae is also noteworthy.
The microstructure in Fig. 5 observed after industrial hot rolling is slightly different from that after warm-rolling processing [7]. During plastic deformation, the behavior of ferrite and austenite phases is quite different. Austenite presents slip bands, while ferrite presents dislocation cells. Strain hardening also occurs in different ways in both phases, as the causes of shear bands, characterized by the wavy microstructure observed in Fig. 5b and c. Shear banding is a common phenomenon in metals deformed to high strains and is favored in low-SFE metals, as austenite, which develops layered microstructures, is often partially twinned, which inhibits dislocation glide and thereby promotes shear banding [22]. This is caused by local plastic instabilities when work hardening becomes exhausted [22]. The lower SFE of austenite also controls the deformation and recrystallization texture proportions and the stored energy during deformation before annealing [23,24,25].
Microtexture
Figures 6 and 7 illustrate the orientation distribution function (ODF) for φ2 = 45° for ferrite and φ2 = 0º, φ2 = 45°, and φ2 = 65° for austenite, respectively. Some important fibers and orientations in the Euler space (φ2 = 0°, φ2 = 45°, and φ2 = 65° section) are also plotted. The microtextures represented in the ODF sections were measured from the half-width of the ND–RD plane.
The ferrite texture is characterized as a typical rolling texture, primarily centered on the rotated cube orientation ({001} <110>) and α-fiber orientations under warm-rolled conditions, with the strongest component being {112} <110>. The austenite phase shows a cube component {001] <100> transforming to typical copper rolling texture features with a pronounced fraction of copper orientation {112} <111> as the deformation increases, along with S (123) <364> and brass {110} <112>.
The average grain diameter of the phases measured using EBSD are shown in Fig. 8. It is possible to see the effect of the warm-rolling strain on the ferrite and austenite grain size. As opposed to results reported by Ma and co-authors [26], where the austenite morphology remains almost undeformed after straining in a super DSS, in this study, austenite was the first phase to accommodate deformation, visible by its pronounced thickness reduction, Fig. 5b and c. This observation was reported in other study realized by Ahmed and Bhattacharjee of a warm rolled super DSS type with high Ni content (10.5%) [18].
The presence of brass components after deformation in austenitic stainless steels and fcc metals has been studied by several authors [23, 25, 27,28,29,30,31,32], and a possible mechanism is due to the low SFE, such as deformation twinning and shear banding. Some authors [33] have attributed the presence of an additional Goss ({110}) component in DSSs after hot rolling with an elevated deformation temperature, which tends to produce extra shear strain, especially near the sheet surface [34, 35]. A decrease in brass ({112}) and an increase in Goss and copper ({112}) components after annealing in cold- and warm-rolled samples has also been reported [8]. However, an increase in the S ({123}) component was only reported in cold working [8].
When the stored energy during cold deformation was not steep, a partial retention of these deformation texture components was observed throughout annealing. For the α-phase, the texture is typical of hot rolling, in which a strong rotated cube component is usually created. Additionally, the cube component incorporates a strong tendency to recover rather than recrystallize, and consequently, to be conserved and still reinforced during the annealing of various steels (pure iron, low carbon, or stainless steels) [23, 31, 34,35,36].
The microstructure after cross-warm rolling of a high Ni (10 wt.%) super DSS revealed a lamellar structure with alternating bands of the two phases (α, γ) [35]. Strong brass and rotated brass texture components were observed in austenite in this steel. Ferrite in the warm-rolled steel showed a remarkably strong α-fiber and {001} component. An increase in ND-fiber in warm-rolled 10.5 wt% Ni DSS has also been reported [19]. This trend was seen at this steel only at warm rolling with 80%, where strong α-fiber and the first components of γ-fiber could be seen. As reported by Ahmed et al. [20], ferrite in DSS warm-rolled showed much stronger α-fiber (RD//<110>) than γ-fiber (ND//<111>). This can be explained by the phenomenon of interaction of carbon atoms with dislocations [19]. The warm-rolling temperature (425 and 615 °C) is within the dynamic strain aging regime of DSS alloys resulting in the preferential retardation of slip on the {110} <111> system and in the strengthening of the α-fiber components at the expense of the γ fiber components as evidenced in the texture of the ferrite in Fig. 6b and c. Ahmed [20] noticed that the strengthening of the α-fiber happens beyond 70% deformation during warm-rolling being affected by carbon locking of dislocations. Therefore, while the strength of the α and γ texture fibers in carbon steels is known to be greatly affected by deformation [36, 37], in the present case the texture development is also greatly affected by the temperature of warm-rolling. The advancement of texture in the austenite and ferrite after cross-warm rolling can be explained by the stability of the texture components. Along with isothermal annealing of the 90% cross-warm-rolling DSS, the lamellar microstructure was retained before the breakdown of such morphology to the mutual interpenetration of the phase ensembles [35]. Ferrite showed recovery leading to an annealing texture quite like the deformation texture. Due to the low-temperature (600 °C) deformation applied here, there is no recrystallization, only recovery. Development of γ fiber from strong {111} <112> orientation is additionally evident during warm rolling in carbon steel [14, 36]. Verbeken et al. [36] observed a strong γ-fiber in an ultra-low carbon steel after cold rolling to 70 and 90% of reduction. In this direction, Samajdar et al. [37] found also the same behavior in an IF steel cold rolled to 90%. The point to enhance in the work of Samajdar et al. [37] is the initial texture of the steel, a hot band completely recrystallized, i.e., an extraordinarily strong γ-fiber. In the present work, the initial texture of the as-received material was a rotated cube, Fig. 6a. This component was maintained along the warm rolling processing, Fig. 6b and c.
In this study, it can be concluded that the warm-rolling textures of austenite in DSS undergo the cube and brass to copper texture transition during warm rolling, indicating some possible restoration phenomena [8]. Goss is also noticeable in the WR 80%. In ferrite, the texture can be adequately described by a strong cube component and RD fiber (RD//) and slight development of ND fiber (ND//). The evolution of deformation and annealing texture in two phases of the DSS developed, in principle, independently of one another, corroborating the results reported above [25, 38].
Transmission Electron Microscopy
TEM micrographs corresponding to observations made in the RD-TD plane are displayed in Figs. 9, 10, 11, 12, 13, 14 and 15. Figures 9 and 10 show the substructure formed in the DSS in the AR condition. Figure 9a shows the presence of deformation slip bands [23, 24, 27, 38], characteristic of the austenite phase (region in the upper right corner of the micrograph, indicated by an arrow). Figure 9b shows subgrains in ferrite, and Fig. 9c illustrates the planar arrangements of dislocation slip occurring in the austenite grain. Figure 9d illustrates discrete arrangements and dislocation loops in ferrite with defined high-angle grain boundaries. In Fig. 10a, it is possible to see the substructure of ferrite containing some subgrains, and Fig. 10b reveals the intragranular dislocation arrays, indicated by an arrow. It is also noticeable that the grain size, according to the EBSD results, is 3 µm for ferrite and 2 µm for austenite.
The 60% warm-rolled sample substructure in Fig. 11a–d contains a cellular structure (Fig. 11a and b) and dense dislocation walls (DDWs). Dislocation tangles can also be seen in Fig. 11c. Figure 11d displays an electron diffraction pattern with respect to austenite in the region indicated by an arrow in the micrograph of Fig. 11c. Figure 12 shows TEM bright-field images of the 60% warm-rolled specimen, where Fig. 12a and b illustrates parallel sub-boundaries formed by dislocation walls [39]. Austenite shows a substructure with tangles and dislocation walls in the austenite, and the mottled contrast in ferrite is associated with substructure formation. The grain size, according to the EBSD results, is 1.7 µm for ferrite and 1.5 µm for austenite.
Representative TEM micrographs after 80% reduction by warm rolling are shown in Figs. 13, 14 and 15. Long DDWs bounded by free dislocations in austenite grain interior result from the relatively low SFE of austenite, as seen in Fig. 13a and b. Conversely, in ferrite, randomly distributed tangled dislocations were dominant. A large portion of the phase boundaries were apparently random, as shown in Fig. 14a. The EBSD results in Fig. 8 show a grain size of 0.9 µm for ferrite and 0.8 µm for austenite. This indicates that the microstructure developed by the warm work was not fully balanced. The EDS chemical analysis (Fig. 15b and c) helps to identify the ferrite phase (point 1) with less Ni and more Mo contents. In contrast, the austenite phase had more Ni and less Mo contents (point 2). This means that an intense dislocation was formed at the α/γ interface that was concentrated in the γ phase (see Fig. 15a, upper corner).
During the warm deformation of steel at 600 °C, dislocations begin to cluster inside individual grains, and these clusters eventually join to form a cell structure as the amount of deformation is increased along with the warm rolling passes. After more than approximately 10% strain, most grains are subdivided into light-defined cells whose walls consist of a tangled network of dislocations. Within each cell, the dislocation density is low. The low temperature at which the specimen is deformed increases the number of long, straight screw dislocations by inhibiting cross-slip. The substructure of steel after large deformation, approximately 20% reduction or more, cannot be described solely based on dislocation glide, dislocation cell formation, and changes in cell dimensions. It has become apparent that in heavily deformed metals, deformation occurs by a series of different processes, each of which operates within a particular range of deformation. With increasing deformation of DSS, the sequence is: (1) slip by dislocation glide, (2) formation of microbands, and (3) macroscopic shearing.
Figures 9, 10, 11, 12, 13, 14 and 15 show TEM images of samples in the as-received and warm-rolled conditions. It is clear that warm deformation introduces a large amount of dislocations in the microstructure, resulting in a characteristic arrangement, depending on the phase, ferrite or austenite. The original grains of ferrite and austenite are subdivided into small arrangements of cell blocks, DDW subgrains, and irregular assembly of dislocations, called forests. The micrographs (Figs. 9, 10, 11, 12, 13, 14, and 15) confirm that grain sizes are reduced to the micrometer scale, or even to lower dimensions, that is, the nanometer scale. The dislocation structure in austenite changed from a planar slip and discrete loops to subgrains. In ferrite, there were cellular block structures. Formation of DDW occurs in both austenite and ferrite phases. This intense formation of DDW in both phases was observed after warm deformation at 600 °C. Hong and Lee [40] working on a low-cycle fatigue (LCF) 316L stainless steel emphasized that for temperatures below 250 °C, a cellular substructure changes to planar arrangements when the temperature reaches 250–600 °C in a dynamic strain aging regime (DSA). At a high strain rate, it changed back to equiaxial cells. In addition, there was a subgrain structure formation at a low strain rate above 600 °C, for instance, in Figs. 10b and 11a. This indicates that the plastic deformation mechanism goes from the wavy slip mode in the non-DSA regime to the planar slip mode in the DSA regime, thus restricting the cross-slip of screw dislocations [40,41,42]. This appears to be consistent with the gradual development of larger-angle planar dislocation walls on the background of misoriented tangled cells, observed by TEM in hot deformed austenite at low strains, as described by Cizek [32].
Dini and Ueji [30] have reported that the planar glide mode is the dominant mechanism before mechanical twinning commencement in TWIP steels (austenitic matrix). According to the authors, at low strains, the activation of primary slip systems from dislocation sources during plastic deformation leads to the formation of dislocation pile-ups against grain boundaries. With increasing strain, the number of dislocation sources increases, and with the activation of two or three slip systems and the lack of mobility, dislocations rearranged themselves in a complex geometric configuration, as can be seen in Fig. 9c, which is a Taylor lattice. In addition, the deformed ferrite microstructure consisted of undefined or undeveloped dislocation cells, as shown in Figs. 11 and 12. In previous cold-rolled DSSs, the formation of equiaxed and elongated cells was reported for a ferritic structure when it was deformed at room temperature and low deformation, ε = 0.18 [43].
Experiments were carried out in austenitic stainless specimens to elucidate the dislocation substructure under LCF loading conditions at 20 and 350 °C. Typical dislocation substructures were found to be cell block structures at 20 °C, as shown in Fig. 11a and b. Still, slip planarity seemed to prevail at 350 °C [42]. In agreement with a previous report [42], for the current research, this condition was found for the as-received sample. Figure 9c represents the typical well-developed planar slip band that characterizes the Taylor lattice. Ferrite, on the other hand, has a crystalline bcc system with high SFE. Because the probability of generating stacking faults is very small, the screw dislocations can cross-slip, allowing for the accommodation of large deformation with a small change in the orientation of their crystallographic substructure. Thus, grain refinement is achieved by successive subdivision of the original grains in the dislocation walls [23, 40, 43,44,45].
TEM images of the fine substructures that evolved at the final stage of warm deformation clearly show the development of very intense gliding of dislocations in the austenite, primarily at the interphase boundary (Fig. 15a and b). Conversely, the ferrite phase is characterized by a conventional cold-worked substructure containing a high amount of DDWs and dislocation cell blocks. In the austenite phase, deformation twin boundaries are usually distinguished by their characteristics, such as straight shapes, close parallel locations, and misorientations that quickly change with straining. However, no evidence was found in warm-rolled DSS specimens because of the high temperature employed for the latter. The increase in deformation temperature increases the SFE, reducing the probability of mechanical twinning formation [46]. The dislocation substructures in the austenite appear as very dense irregular parts with rather thick diffuse sub-boundaries, presenting a wide spectrum of low- to high-angle misorientations (Figs. 12b and 13a). The presence of the cell structure of dislocations is attributed to cross slipping in screw dislocations, observed in ferrite grains [47]. The dislocation density in ferrite is lower than that in austenite because of the recovery process through annihilation of dislocations; only this mechanism operates at a low temperature of 600 °C [29, 45, 47, 48]. This condition can be confirmed by examination of Figs. 9 and 10. The austenite phase has more non-resolved Kikuchi bands during the EBSD examination, that is, the austenite grain structure is not resolved as ferrite grains, as shown in Fig. 5. This means that a high deformation state is translated by intense dislocation gliding. This condition implies that it is more difficult for an EBSD system to identify and classify the Kikuchi bands [31, 35].
Tensile and Hardness Tests
Figure 16a corresponds to the engineering stress–strain curve for the as-received and warm-rolled samples, as well as the curves for all the samples after annealing at 1050 °C for 300 s in Fig. 16b. Figure 16a shows that warm rolling led to a significant increase in ultimate tensile strength and a reduction in ductility. In addition, the Vickers hardness (Table 3) of the warm-rolled samples converged to an average value of 408 HV, well above the steel in the as-received condition. Therefore, it is possible to conclude that the warm-rolling route of the DSS used here is suitable to produce a high-strength microstructure, as the warm-rolled condition can double the yield and tensile strength (Fig. 16a).
Table 3 Results of Vickers microhardness tests (4.9 N–10 s) The Vickers hardness of the annealed samples (Table 3) converged to an average value of 255 HV, slightly inferior to that of the as-received steel. In addition, when the warm-rolled and as-received samples were annealed at 1050 °C for 300 s, there was an increase in ductility and strength reduction, as shown in Fig. 16b. Comparing the as-received sample in Fig. 16a with the three conditions in Fig. 16b, there is a possible advantage in applying warm rolling and annealing to the steel, to reach industrial conditions, without the need for cold rolling and subsequent annealing. However, this strength level has already been reached in hot rolling conditions. However, to reach a thickness of 1–1.5 mm for commercial applications, warm rolling would be necessary.
As shown in Figs. 5 and 8 the grain size results, the austenite phase undergoes a more intense deformation, resulting in a fine substructure as consequence of strain hardening. In contrast, the ferrite was not able to undergo strain hardening at the same intensity, so it was prone to deformation by a cleavage mechanism [7].
There is a pronounced loss of ductility associated with warm rolling at 600 °C, as shown in Fig. 16a. However, it is worth observing that the 80% warm-rolled sample did not have its ductility reduced compared with the 60% sample. In fact, it showed a slightly larger ductility than the sample warm-rolled with 60% reduction (Fig. 16a).
After warm rolling, both the austenite and ferrite phases are self-distinguishable because of differences in the structure and intense heterogeneity (Fig. 5b and c). Conversely, the as-received sample has a more uniform microstructure, with a lamellar structure oriented according to the rolling direction (Fig. 5a). However, the dominant dislocation arrangements in both conditions (60 and 80% reduction) are still the same: intense formation of DDWs in austenite and subdivision in ferrite grains by dislocation, forming extended dislocation walls. In the specimen with the highest total strain, the appearance of deformation was very swift during the tensile tests. Necking occurred above the plastic deformation over the yield strength. The total elongation of the severely strained steel in a small volume was significantly reduced. Such behavior is typical for nanostructured materials [48, 49]. According to the literature [27], microcracks or shear bands (Fig. 5b) should terminate at the grain boundaries. However, it is difficult to determine the mechanisms responsible for the fracture because of the small volumes (reduced grain size, Fig. 8) in which the deformation takes place, for instance, grain diameter or α/γ interface [27].
The engineering stress–strain curves of the tensile specimens, corresponding to the microstructures shown in Fig. 5, are presented in Fig. 16a. For comparison, the tensile curves of the specimens annealed at 1050 °C for 300 s, after warm rolling, are also plotted in Fig. 16b. A distinct loss in strength when compared with its warm-rolled counterpart process can be easily observed. The difference stems from the grain size in both phases, which is greatly reduced to approximately the nanometer scale, for the warm work applied here. The structure after deformation at 600 °C, containing an ultrafine substructure, showed a stress peak soon after the yield point and failed at a percentage near the total elongation (Fig. 16a). These warm-worked samples displayed a small uniform strain-hardening characteristic, but showed a reasonable post-necking deformation, that is, total elongation. The lack of strain hardenability of the microstructure with nanostructured configuration has been anticipated by some authors [40, 47] and is considered to be the primary cause of low uniform tensile elongation. In addition, the ultrafine grained (UFG) size microstructure consisted of an intense deformed substructure with cellular regions, tangles of dislocation arrays, and high-density dislocation walls (HDDW), which strongly contribute to high strength and a lesser strain work hardenability [50, 51], in agreement with a previous report [51, 52].
A similar level of mechanical strength to that depicted here was reached in an experiment that involved a hydrostatic extrusion process with a true deformation of 1.4 in a DSS with 0.3%C, 22%Cr, 5%Ni, and 3.0%Mo. In this case, a tensile strength of 1370 MPa was reached, but with only 5% of the total elongation [27].
The high level of mechanical strength found in the present work is due to the substructure formation: dislocation tangles, cell block dislocations, incipient formation of subgrains, and DDWs. An additional increase in disorientation by dislocation accumulation; development of deformation heterogeneities, such as deformation and shear bands, which intensify to a large gradient of orientation in a small volume of the microstructure; and coalescence of the contours for large deformations to form cell blocks also contribute to the strength. The texture formation mechanisms involve relative rotation between different parts of a grain for different final orientations because of differences in the slip systems selection [28,29,30].
Ratuszek and Witkowska [53] have reported that the process of texture formation proceeds in different way depending on the initial orientation distribution and, also, that the formation and character of the final deformation textures is strongly influenced by the previous morphology of DSS phases. They show that ferrite–austenite band-like structure exerts considerable effect on slip behavior. Since Burgers vectors of slip dislocations in each DSS phase are different thus only slip transfer may occur across the α/γ interfaces. It seems that the interphase boundaries are strong barriers even for the co-ordinate dislocation motion throughout the band-like structure and the extension of shear banding is apparently strongly restrained by the α/γ interfaces, affecting the rolling texture.
Notwithstanding of processing conditions, the residual dislocation density in the all samples subjected to cold or warm rolling followed by annealing at different temperatures varies locally in a rather wide range [28]. The density is not the same in different recovered grains. As should be expected, the values of local dislocation density decrease with an increase in the warm rolled pass due to the interval between them, when the steel recovery the temperature for the next pass. As a result, the variation range of the residual dislocation density tightens with a progress in recovery because of reduction of local high dislocation densities first. Nevertheless, the dislocation density varies locally in a wide range even in apparently well recrystallized microstructures of an austenitic stainless steel after annealing, as suggest by Odnobokova et al. [28].
The grain refinement mechanisms are different between austenite and ferrite, for instance, during the ECAP process was described by Chen et al. At the early stage of ECAP, the ferrite should be refined by successive subdivision of dislocation walls due to the operation of multi-slip systems. Further deformation will increase the portion of high-angle grain boundaries owing to rotation of subgrains, leading to the formation of nanostructure in ferrite [29]. However, the deformation mechanisms of austenite are associated with dislocation slip and deformation twinning. The intersection of deformation twins may result in rapid subdivision of coarse grains into submicron-sized blocks, which could be further refined by the formation of micro twins and dislocation boundaries inside the interiors [29, 30]. During the warm rolling, as function of temperature, the deformation twin is not formed.
Deformation microstructures of polycrystalline single phase materials with medium to high SFE show a continuous evolution with strain in which stable glide dominates. An important characteristic of this evolution is the subdivision of the grains into cell blocks, which deform by scarcer slip systems than specified by the Taylor criterion for strain accommodation. However, the groups of neighboring cell block fulfill the Taylor criterion collectively, with increasing strain the size of the cell block shrinks faster than usual dislocation cells. Subgrains are formed as a last step in this evolution. New cells blocks can be formed principally through the formation of microbands [50].
Hence, the grain refinement mechanisms are different between austenite and ferrite in DSS, leading to heterogeneous microstructure after thermomechanical pressing [28]. Therefore, the work hardening can be a function of both the dislocation structure and the specific microstructural features such as cell blocks and DDW.
It is generally accepted that austenite with low SFE results in limited cross-slips, and hence, a low recovery rate. Consequently, the main softening mechanism in austenite for hot working would be static or discontinuous recrystallization. Ferrite, in turn, has high SFE, so its recovery rate is high, alongside a higher diffusion rate. Consequently, the recrystallization driving force remains far too low, making it unobserved, mostly for low temperatures employed in warm work at 600 °C.