Influence of quenching conditions on texture and mechanical properties of ultra-high-strength steels

The investigations on the influence of the processing route were performed on four different steels. These steels display industrial available UHS steel grades, which were provided from the coil as 6-mm-thick sheets. Their chemical composition is listed in Table 1. All steels experienced the same rolling conditions, and the FRT amounted 850 °C. The steels were labeled according to their strength classes and are numbered from 1 to 4. Steels 1 and 2 represent temper-resistant UHS steels and possess a carbon content of 0.09%. Additions of Mn, Si and B retard the γ to α transformation to ensure a martensitic microstructure [34‐38]. Mo and V facilitate a TM process route through an elevation of T_NR [22, 34, 39, 40]. Moreover, they compensate for the softening during tempering and enhance weldability [41]. Compared to the base alloy 1, steel 2 has higher additions of Cr, and additionally, Ni is added for further temper resistance [37, 42]. Alloy 3 represents a wear-resistant steel with a carbon content of 0.17%. Elevated additions of Si and Mn enable a through hardening [35]. Compared to steel 3, temper resistance is given in steel 4 with elevated amounts of Cr, Ni, Mo and Cu. Further, steel 4 is alloyed with Nb, which serves as strongest MAE in retarding recrystallization and providing additional strength after tempering [10, 43].

The steels were investigated in the as-rolled condition and compared to re-austenitization and quenching, for which three different austenitization procedures were performed. The different heat treatments are listed in Table 2. The conventional re-austenitization and quenching (RQ) consists of an austenitization at a temperature of 930 °C for 5 min followed by water quenching. In the second route, quenching is performed after normalization treatment and an extended austenitization time of 30 min (NRQ). In the third route, quenching was executed after a higher austenitization temperature of 1050 °C (HRQ).

Tensile tests were performed according to DIN EN ISO 6892-1 on flat tensile specimens in longitudinal and transverse directions of the 6-mm-thick sheet of each condition. The notch impact tests at 0°, − 20 °C and − 40 °C were performed according to DIN EN ISO 148-1 on 6-mm-thick specimens, and the obtained notch impact work was standardly converted to 10 × 10 mm full-sized samples by linear upscaling. Tensile tests were applied twice in each direction, and for impact tests, three samples of each condition were tested.

For the microstructural analysis, three different conditions (DQ, RQ and NRQ) of the four steels were investigated. Since the measuring principles of XRD and EBSD differ, the perspective on the sheet is also different. The relevant position to detect rolling textures was set to one-fourth of the sheet thickness [44]. For EBSD measurements, the transverse direction (TD) of the sheet is observed, as the electron beam can be focused to the significant position on the sheet.

The samples were hot embedded and then ground from 320 grit to 4.000 grit SiC paper for at least 30 s. Subsequently, the samples were polished with 3 μm diamond paste for at least 3 min and with 1 μm for 30 s. A silicate polishing was used (Struers OPS) for 10 min prior to a finishing electrolytic ablation of 5 s with 35 V. The samples were dipped in a diluted Nital etchant and plasma cleaned for 9 h before performing EBSD measurements. The measurements were taken using a EDAX Hikari XP EBSD system, installed in a FIB Versa FEI 3D DualBeam. An acceleration voltage of 30 kV was applied, and a step size of 100 nm was chosen. The data acquisition was performed with EDAX Team 32. The data were evaluated with EDAX OIM Analysis 7.3, in which a neighbor orientation correlation cleanup and grain dilatation cleanup with one single iteration were executed. High-angle grain boundaries with an angle of > 10° were highlighted. The misorientation distribution was analyzed in steps of 2°.

In contrast to the electron beam, the X-ray beam cannot be focused. Therefore, the surface of the sheet was milled down to one-fourth of the thickness prior to grounding and polishing in order to place the beam on the relevant position of the sheet. In the following, electrolytic polishing was performed for 1 min with 35 V. The utilized XRD, Bruker D8 advance, was operated with a voltage of 40 kV at 40 mA and a K_α Cu X-ray tube. In steps of 5°, {110}, {200} and {211} pole figures were measured by means of a LYNXEYE XE detector and evaluated with the program DiIFFRAC.SUITE.measurement™. The orientation distribution functions (ODF) of the individual sample conditions were calculated from the obtained pole figures in the software package MTEX at an applied half width of 5° [45]. Volume portions are then calculated to describe the relative volume of the relevant crystals, which obtain a certain orientation.

The yield strength of the 4 steels investigated is provided in Fig. 1. Steels 3 and 4 possess significantly higher YS compared to steels 1 and 2. The strength of the DQ condition is considerably above those after re-austenitization (RQ) for each of the steels investigated. The additional normalization (NRQ) and the higher austenitization temperature (HRQ) result in a minor decrease in strength for steels 1 and 4. Yet, this phenomenon is less pronounced as compared to DQ. Steels 1 and 2 show higher YS in T than L, whereby these differences are decreasing after re-austenitization, NRQ and HRQ. The L–T differences between steel 3 and steel 4 show opposite tendencies, and the YS in L is little higher compared to T. The heat treatments, however, do not reveal a clear trend concerning the L–T anisotropy of the YS. Clearer differences in the L–T anisotropy are evident in the Charpy impact toughness, which are displayed in Fig. 2 for the investigated steels at a testing temperature of 0 °C. The toughness behaves diametrical to the strength, and steels 1 and 2 have the highest Charpy impact toughness values, followed by steels 4 and 3. Steels 1–2 experience a deterioration of the impact toughness values through re-austenitization and quenching, steel 3 first after normalization. The toughness of steel 4, however, increases through re-austenitization. The L–T anisotropy of steels 1 and 2 benefits from the re-austenitization. The more intense the annealing in the γ region is (e.g., NRQ and HRQ), the smaller the L–T difference becomes. For steels 2, it disappears completely. The anisotropy for steels 3 and 4 is neither deteriorated nor improved through different treatments in the γ region. The temperature dependence of the Charpy impact toughness is displayed in Fig. 3 for temperatures between − 40 and 0 °C for steels 2–4. The decrease in the impact toughness of steel 2 is less pronounced at 0 °C; however, at − 20 °C, it deteriorates to values below 50 J. The impact toughness of steel 4 is clearly improved through re-austenitization and amounts even at − 40 °C almost 100 J for the normalized and quenched variant NRQ.

The inverse pole figure maps for steels 2–4 in the TD are given in Fig. 4. In the DQ condition, steels 2 (a) and 4 (c) show a strong deformation of the former γ-grains toward the rolling direction. Steel 4 exhibits a very fine martensitic microstructure, which is divided in blocks and laths, originates from the also very fine prior austenite grains. The martensitic microstructure of steel 3 originates from equiaxed PAGs, which are obviously larger than in steel 4. During re-austenitization, the PAGs form to a equiaxed shape, which is shown in Fig. 4d–f. Whereas steels 2 (d) and 3 (e) coarsen significantly through reheating in the γ-region, the grain growth of steel 4 is limited. Both the PAGs and the martensitic components are still extremely fine after re-austenitization.

Figure 5 shows the distribution of misorientation angles of steels 2–4. Four peaks are visible in the distribution of all investigated specimen. Augmented peaks are found at 6.5° and 59.5° and peaks with weaker appearance at 17° and 53°. The magnification of the relevant positions of the diagram in Fig. 5 delivers an influence of the heat treatment on the frequency of these angles. Steels 2 and 4 show an increase in the intensity of the 6.5°—angles after RQ, compared to DQ. Steel 3 does not exhibit a change of this fraction after re-austenitization. At the misorientation at 59.5°, in particular steels 3 and 4 show higher fractions after RQ, and steel 2 only possesses a weak intensification.

Figure 6 illustrates the ODF at a constant angle ϕ₂ = 45° of steels 1–4 dependent on the processing condition. A significant increase in the rotated cube component {001} <110>, evaluated from the ODF with an angular tolerance of 10° was observed for steels 1–3, especially in the DQ condition. The volume portions (Fig. 7) amount 10% for steels 1 and 2 and 11% for steels 3 and 4. During re-austenitization and quenching, all texture components decrease in favor of a slight increase in {112} <110>. All texture components of steel 3 fall below 5% even after a single re-austenitization, concluding that only little fragments from the rolling history remain. In contrast to the aforementioned, steel 4 exhibits increased intensities in {111} <112>, {112} <131> and {554} <225> after DQ. These components are marginally reduced on behalf of a growth of the {001} <110> constituent during RQ and NRQ. However, all texture components of steel 4 are still present to an increased extent.

The development of the YS revealed at first sight that an intense austenitization with a prior normalization can have a positive effect on the L–T anisotropy. Regarding the 0.09% C steels 1 and 2, the YS anisotropy behaves differently compared to steels 3 and 4. Steels 1 and 2 show higher YS in T than in L, whereas steels 3 and 4 show little higher YS in L. Due to the fact that the DQ strip is coiled after cooling and straightened prior to further processing, residual stress is induced. It is experienced that the more intense this straightening is, the lower the longitudinal YS becomes by the induced residual stresses. The transverse YS is not affected, as the strains are only applied along the rolling direction. The L–T anisotropy of the YS of steels 1 and 2 thus are disguised by these strains. The apparent decrease in the YS anisotropy after re-austenitization consequently can be associated with recovery. If steels 3 and 4 now are withdrawn from a rather outer part of the coil, the straightening is reduced due to the bigger coil radius, which subsequently leads to smaller residual stresses. This could explain the different behavior of steels 1 and 2 compared to steels 3 and 4. Nevertheless, the tensile tests have been performed on a class 1 tensile testing machine, which possesses an accuracy of 1%. Consequently, these deviations of the YS of steels 3 and 4 fall below the measuring tolerance so that this delivers no sufficient basis for interpretations of the L–T anisotropy. By re-austenitization, however, the absolute YS is significantly decreased. Regarding steel 1, the YS decreases by ~ 17%, steels 2 and 3 by 20% and steel 4 even by 24% after a single re-austenitization in the T direction. These observations can be attributed to the grain growth during re-austenitization, as depicted by the IPF map in Fig. 4, but cannot alone be the reason for this significant decrease. Although for all steels the YS is deteriorated in the same amount, the microstructure of steel 4 still maintains very fine after re-austenitization, whereas the grain size of steels 1–3 increase unambiguously. A detailed explanation can be given with a view to the different alloying concepts. All investigated steels were rolled with an FRT of 850 °C. The rolling below T_NR leads to a high deformed γ grain (pancaking), which results in a very fine transformed microstructure. Nb (in steel 4) has a by far stronger effect than V (steels 1 and 2) to retard recrystallization [10, 21, 22, 34, 43, 46]. The consequence is that the strength contribution through TMP for steel 4 is much more augmented than compared to steels 1 and 2. Although the grain coarsening of steel 4 through RQ is less pronounced, the lack of a high deformed γ grain accompanied by the recrystallized austenite contributes to the 24% decrease in YS. These effects as well are observed in the temperature dependence of the Charpy impact toughness, which is displayed in Fig. 3. The detected grain coarsening of steel 2 not only reduces the impact toughness, but shifts the transition temperature to higher values. Although steel 3 is dispensed with alloying elements which are known to optimize the impact toughness (Mo, V, Ni) [41, 47], the toughness still is above 50 J at 0 °C and does not deteriorate noticeably through re-austenitization. Steel 4 profits from the fine grain, which is still maintained after re-austenitization and results in higher impact toughness, which is even improved through RQ in elevating the entire level of the DBTT to higher toughness.

The investigations on the microstructure delivered details on the different behavior of the four investigated low-C steels regarding their microstructure–property relationship. Several peculiarities of the mechanical properties have been explained through the grain size and structure by means of EBSD measurements. A view on the distribution of the misorientation angles (Fig. 5) gave further insight into the microstructure. Augmented peaks were observed at angles of 6.5° and 59.5° and peaks with weaker appearance at 17° and 53° for all investigated specimens. These peaks correspond to variants of the Kurdjumov–Sachs orientation relationship of packet, sub-block and block boundaries [8, 48, 49]. It was observed that the fraction of the misorientation at 6.5° of steels 2 and 4 is increased during re-austenitization compared to the direct-quenched variants. Due to the fact that steel 3 does not exhibit this phenomenon but rather shows an unchanged distribution at 6.5° between DQ and RQ, this difference can be attributed to the pancaking of the PAGs due to the MAE of steels 2 and 4. After RQ, the equiaxed PAGs of steels 2 and 4 contribute to an increase in the low-angle grain boundaries. The PAGs of steel 3 are already equiaxed after DQ, so that a change of the frequency does not occur. Furthermore, also the high-angle grain boundaries at 59.5° show an increase after re-austenitization. Nevertheless, these differences are too little to contribute to an explication for a remaining anisotropy.

Consequently, there is still a missing link of the remaining anisotropy which in case of steels 3 and 4 cannot be eliminated completely, and therefore, measurements of the texture were taken. EBSD reaches its limit if the detected microstructure is very small as for the present case [50]. For a proper texture analysis, at least 10.000 grains are suggested to be measured [51]. The step size needs to be reduced immensely when detecting microstructures with high grain boundary densities, as it is the case for martensitic steels. This would result in an extremely high effort for textures analysis with EBSD. The XRD measurements, however, provided the orientation distributions over the whole sample for the relevant sheet position of 1.5 mm below the surface. The orientations of interest and their distributions are highlighted in Figs. 6 and 7. All steels investigated contain a decisive fraction of the {001} <110> component, especially steels 1–3 after DQ from the rolling heat. This orientation arises from the recrystallized austenite cube texture [52, 53] and thus is sparely present in steel 4 which is alloyed with strong recrystallization retarding elements (Nb) preventing a recrystallization prior to transformation. The {001} <110> texture is known to be unfavorable for impact toughness and provokes delamination and crack propagation [52, 54]. However, a clear link to the A_V/T diagram in Fig. 3 cannot be made. Despite a decrease in the intensity of {001} <110> (Fig. 7), the Charpy impact toughness of steels 1 and 2 decreases too and already falls after NRQ below 50 J at a temperature of 0 °C. These values are even lower than the values measured for steel 3, although a higher toughness is expected for steel 4 according the diametrical trend of strength and ductility. The prevailing components in steel 3 after DQ are {111} <112>, {112} <110> and {112} <131> which emerge from a deformed austenite [29]. The latter ranks as a very stable orientation for higher toughness toward the rolling direction [31, 33, 52‐54]. Furthermore, all steels contain significant fractions of {554} <225> arising as well from a strongly deformed γ grain [20, 52]. This component is inherited to RQ and even present after NRQ. The relatively high proportions of the {001} sheet plane component in the DQ condition of steels 1–3 might explain differences between L and T. This is confirmed by the fact that a reduction in the {001} <110> partitions reduces the impact toughness anisotropy after RQ. However, the question remains, why an almost levelled-out texture distribution as it exists in steel 3 after re-austenitization (Figs. 6 and 7) still delivers small but present differences between L and T. And in contrast to that, steel 4 which exhibits a significant amount of defined texture components even after HRQ exhibits excellent Charpy impact values in each condition (DQ and RQ) with a relatively small anisotropy. Consequently, a clear link between a remaining anisotropy and prevailing textures thus cannot be made with certainty. Furthermore, a strict distinction whether a mechanical anisotropy can be attributed to an elongated γ grain, texture fractions or nonmetallic inclusions [23] is not possible. In a previous publication [10], it was found that variances in the FRT can even affect the properties of Q + T steels, contrary to the assumption that the FRT has no effect on the properties of steels which are later re-austenitized and quenched. The present investigations confirm these findings, in specific, that the relation of FRT and T_NR not only plays a major role during TMP but rather influences the mechanical properties through the inheritance of microstructural characteristics. Although the usage of MAE provokes through a highly deformed γ grain a mechanical anisotropy, this does not mean simultaneously that the microstructural cancellations of these peculiarities are the key to a mechanical isotropy. Rather it can be observed that the higher the pancaking is, the more intense is the inheritance of the properties of the prior process steps.