Formability of a HSAS based on hot processing maps and texture analyses

Aiming to improve the formability of a HSAS Docol 1500 Bor, hot processing maps were obtained based on Prasad, Babu and Murty instability criteria. The hot processing maps based on the above instability criteria are similar and the explanation of its similarity is given. Recrystallization and misorientation in typical quenched specimens were observed by using SEM with a EBSD system. It was found that the fraction values of HAGBs in quenched martensite are all below 0.4 under experimental conditions. Flow location bands occurs at lower deformation temperatures and higher strain rates. The textures in martensite mainly include 〈110〉//ND and 〈110〉//RD components. Based on N-W OR, the textures in deformed austenite are mostly 〈111〉//ND and 〈112〉//RD〉 components. Prasad and Babu instability criteria are more conservative than Murty instability criterion in obtaining the processing maps of the tested steel. To reduce the anisotropy of quenched workpieces because of the textures at room temperature, the quenching temperature in the stamping process of the tested steel should be lower.


Introduction
The high strength automobile steels (HSAS) with high yield strength and high tensile strength have been widely used in production of automobile stamping parts. To improve the formability of HSAS, many studies have been carried out on the composition [1], mechanical properties [2], processing technologies etc [3,4]. The processing map is an effective and popular tool to study the deformation behaviors of a wide range of materials such as steels [5], aluminum alloys [6] and Ni-based superalloys [7]. By using processing map, the instability regimes can be predicted and the optimized process parameters can be obtained [8,9].
Many researchers have proposed some stability and instability criteria of materials [10,11]. Gegel and Malas constructed similar stability criteria by using Lyapunov functional stability criterion [12,13]. Prasad [14], Babu [15], Murty and Rao [16] had developed different instability criteria based on the maximum principle of irreversible thermodynamics applied to large plastic rheology. Some typical research works had been carried out to compare the prediction accuracy of these stability or instability criteria [17][18][19][20]. Li et al constructed the processing and instability maps of titanium alloy Ti-6.5Al-3.5Mo-1.5Zr-0.3Si based on the Prasad, Murty and Malas criteria. Li et al found that the predicted results based on the Murty and Prasad criteria were not much different, which were quite different from the results predicted by the Malas criterion [17]. Wang et al proved that the Gegel and Alexander-Malas stability criteria were mathematically equivalent and thus predicted the same unstable region in ultrahigh strength stainless steels [18]. Zhou et al compared the processing maps of superalloy GH4742 with different instability criteria of Prasad, Gegel, Malas, Murty and Semiatin. The results show that Murty instability criterion was most suitable for superalloy GH4742 [19].
For the high strength automobile steel (HSAS) products, the microstructure obtained is homogeneous quenched martensite after hot stamping and quenching processes. The microstructure characteristics of Original content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence.
Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. quenched martensite affect the performance of workpieces to a large extent, which inherit from austenite at high temperatures [21,22].
In this work, the hot deformation behaviours of a HSAS are studied by using the processing maps constructed based on different criteria at isothermal compression conditions. Furtherly, the influence of deformation conditions on textures in the quenched microstructure are discussed theoretically by using scanning electronic microscope (SEM) Canta650 with a EBSD system. The aim is to analyse the relationship between the textures of quenched martensite and the deformation conditions at high temperatures.

Material and hot compression test
The chemical compositions of the tested HSAS Docol 1500 Bor are 0.22%C-0.2%Cr-1.15%Mn-0.3%Si-0.003% B-0.03%P-0.01%S in weight percentage. To dissolve the precipitates such as carbides and obtain a homogenous microstructure, the tested steel was solution annealed. The solution temperature was 1120°C and the solution time is 3 h. The cylindrical specimen with dimensions of f8 mm×15 mm from the homogenized piece were isothermally compressed on a thermal-mechanical simulator Gleeble-3800. The specimens were heated to 1050°C and held for 5 min to ensure the uniformity of temperatures then cooled with cooling rate of 10°C/s to the test temperatures of 900, 950, 1000 and 1050°C and held for 3 min followed by compression tests at the engineering strain of 0.2, 0.4 and 0.6, respectively, while the strain rates ranged from 0.01to 10 s −1 . Tantalum sheets were stuck on the both end faces of cylindrical specimens to reduce friction. The specimens were immediately quenched in water after compression tests.

Microstructure observation
The room temperature microstructures after deformation at 900°C and strain rates of 0.01-10 s −1 were carefully examined on an optical microscope. The quenched specimens were sliced along the longitudinal direction, then prepared by mechanical polishing and revealed by using 4 wt% Nital etching for 9 s. Microstructure evolution in instability and stability domains was observed by using SEM with a EBSD system. Then some typical deformed specimens were selected and sliced along the longitudinal direction for EBSD examination, which were carefully prepared first by mechanical polishing and followed by electropolishing in a perchloric acid solution (8 vol% HClO 4 , 90 vol% C 2 H 5 OH and 2 vol% C 3 H 8 O 3 ) at 17 V for 20 s. The EBSD analysis was performed by S-4800 field emission scanning electron microscopy coupled with a fully automated with a TSL-OIM™ EBSD attachment, operated at 25 kV. For the compressed specimens in this work, scans were performed over an area of 250 μm×50 μm with a step size of 0.3-1.5 μm depending on the grain size. Figure 1 shows the flow stress curves of the tested steels during compression tests at different temperatures and strain rates. The critical stress σ c , critical strain ε c , peak stress σ p and peak strain ε p was shown in figure 1. The critical strain was the valley value of the dθ/dσ−σ curves, where θ was the work hardening rate, q s e = ¶ ¶ e  ( ) / [23,24]. It can be seen from figure 1 that the flow tresses decrease with increasing deformation temperatures or decreasing strain rates. At higher temperatures such as 1000 and 1050°C, the flow stress curves show obvious peaks at all experimental strain rates, which indicate that continue dynamic recrystallization (DRX) occurs during hot compression. In contract, at lower temperatures such as 900 and 950°C, the most of the flow stress curves are nearly flat, which shows the typical characteristic of dynamic recovery (DRV) during hot deformation. Figure 2 show the strain hardening rate-true stress curves at 950°C and strains of 0.01, 0.1, 1 and 10 s −1 . The strain hardening rate will taper off to zero when no DRX occurs. The presence of DRX causes a downward inflection in the strain hardening rate-true stress curves, leading to zero and negative strain hardening rates, which corresponds to peak and post peak softening, respectively. In this work, the stress peaks can be precisely measured as 58.2, 79.3 and103.3 MPa for HSAS Docol 1500 Bor steel at strain rates of 0.01,0.1,1 and 10 s −1 , respectively. Figure 3 shows the transformed martensite after at 950°C and strain rates of 0.01, 0.1, 1 and 10 s −1 when the strain is 0.6. It can be seen from figure 3 that a large amount of lath martensite exits in the deformed specimens. The morphology of quenched martensite grains reflects that the deformed grain size decreases with increasing strain rates. At the strain rates of 1 and 10 s −1 , the finer martensite grain results from the deformed austinite at high temperatures, where DRV is the major phenomena as shown in figures 3(c) and (d). Because martensitic transformation occurs after deformation, the deformation streamline at high temperatures could not be seen  clearly. However, the local deformation bands can also be observed in figures 3(c) and (d), which indicates the occurrence of flow localization.

Processing maps
At present, the processing map has become a very powerful approach to optimize the hot working processes and further control the microstructures and properties of the metal products. Basing on the dynamic materials model (DMM), The total power P absorbed by the work piece is the sum of two complementary functions: where G, J, σ and e  represent the power dissipations in the form of a temperature rising, the one caused by microstructure evolution such as DRX, DRV and phase transition, the flow stress and the strain rate, respectively. The strain rate sensitivity, m describes the power portioning between G and J, which can be expressed as follows: The efficiency of power dissipation, η presents the power dissipated through microstructure changes during hot deformation, which could be derived as a function of strain rate sensitivity: where J max is the theoretically maximum value of J. Then the power dissipation map plotting the efficiency of power dissipation η against temperatures and strain rates can be obtained, in which different domains with various η values present the diverse microstructural mechanism. It is generally accepted that the domains with high efficiency are considered as the optimum deformation conditions. However, for the instances that instable plastic deformation or flow localization occur, the efficiency of power dissipation alone is not sufficient to identify the unsafe domains. Fortunately, this problem has been overcame in virtue of instability criterion ξ P ( , e  T) for plastic flow proposed by Ziegler [25].
Equation (4) is deduced by the assumption that the flow stress-strain rate curve follows a power-law relation K , m s e =  in which m is constant. However, m varies with e  in reality. Murty and Rao proposed an instability criterion applicable for any type of stress-strain rate curves [16].
where M h is the efficiency of power dissipation in instability criterion ξ M ( , e  T), which should be deduced by its original definition.
Babu et al deduced another instability criterion based on the one proposed by Murty, which was also applicable for any type of stress-strain rate curves. Equation (5) can be modified as follows: Take partial differentiation of e  on both sides of equation (9): Both sides of the equation (12) are divided by σ, then the instability criterion ξ B ( , e  T)can be obtained as same as equation (8).  x e e = ¶ ¶ + It can be found that an extra m 3 exits in the instability criterion ξ p ( , e  T) by comparing equations (8) and (18). Because m 0 1, < < m 3 has little influence on the calculation results of equation (18). Then it can be known that the instability criterion ξ p ( , e  T), ξ B ( , e  T) and ξ M ( , e  T) are the same in theory, which can be used to explain the similarity reflected in figures 4, 5 and 6.

Recrystallization and texture analysis
Martensitic transformation is non-diffusive phase transformation. The size of austenite at high temperatures determine the size of quenched martensite. Then, the size of quenched martensite can be used to study the degree of recrystallization of deformed austenite. In this work, the grain tolerance angle of 1°was used to determine whether two neighbouring scan points belong to the same grain. Figure 7 shows the misorientation angle distributions of quenched martensite grains under different deformation conditions. It can be seen that the fraction of high angle grain boundaries (HAGBs)(15°-180°) of martensite increase with increasing the deformation temperatures and deceasing strain rates of parent austenite at the same strain as shown in figure 7. At the engineering strain of 0.2, the fraction values of HAGBs of martensite at the deformation temperature of 1050°C and strain rates of 1 and 10 s −1 are 0.286 and 0.268, respectively, while those at the strain rate of 1 s −1 and temperatures of 950 and 1050°C are 0.286 and 0.31, respectively. The same change law exists at the engineering strain of 0.6 as shown in figures 7(e)-(h). At the same time, the fraction of HAGBs of martensite increases with increasing strain at the same temperatures and strain rates as shown in figures 7(d) and (h), which increases from 0.299 to 0.366 at the deformation temperature of 1050°C and strain rates of 10 s −1 when engineering strain increases from 0.2 to 0.6. The increasing of the fraction of HAGBs of martensite means that more recrystallization has completed in deformed parent austenite at high temperatures. However, referring to figure 1, no complete dynamic recrystallization occurs at temperatures of 950, 1050°C and strain rates of 1, 10 s −1 , respectively, when the strain was 0.2, which leads to the low fraction of HAGBs as shown in figures 7(a)-(d).
The experimental orientation relationship (OR) between plate martensite and prior austenite is close to the Nishiyama-Wassermann (N-W) OR ( 111 110 , . These textures in lath martensite influence the plasticity at low temperatures, causing different deformation behavior along different directions of the sheet workpieces. For correlating anisotropic properties of the HSAS products with textures, it is necessary to describe the textures in a quantitative way. To achieve the statistical purpose, in several uniform deformation region was selected and the quenched martensite textures under specific deformation conditions was measured by using TSL OIM analysis software, of which the typical grain distribution was shown in figure 8.  The measured orientation has no obvious preferred orientation and no texture exit in quenched martensite, which should result from no preferred orientation appears in deformed austenite at high temperatures because of the low strain. Meanwhile, the fraction values of recrystallization at the strain rates of 1 and 10 s −1 at the temperature of 950°C have no significant difference, while the same situation exits at the temperature of 1050°C as shown in figures 7(c) and (d). Therefore, the deformation is easy to occur at the engineering strain of 0.2 and the presences of the Unstable zones I P and I B in figures 4(a) and 5(a) are not reasonable, which agrees with literature [17]. Then the processing maps based on the instability criterion ξ M ( , e  T) are more reasonable compared with the two ones based on the instability criteria ξ P ( , e  T) and ξ B ( , e  T) as shown in figures 4(a), 5(a) and 6(a). For figures 8(e), (f), (g) and (h) at the engineering strain of 0.6, the distributions of textures show obvious preferred orientation. The quenched martensite texture mainly includes 〈110〉//ND and 〈110〉//RD components, which correspond to the 〈111〉//ND and 〈112〉//RD texture components in recrystallization austenite after compression deformation at high temperatures. In generally, recrystallization of compressiondeformed steel tends to retain the deformation texture component [111]//ND [26]. Then the recrystallization austenite has the similar texture characteristics with the deformed austenite. For figures 8(f) and (g), the fraction of quenched martensite texture components 〈110〉//ND is less, from which we can deduce that less deformed austenite texture components 〈111〉//ND appears. Then the deformability of tested steels was deteriorated, which verify the correctness of the processing maps in this work. At room temperature, the fraction of martensite texture component 〈110〉//ND increased with increasing deformation temperatures and decreasing deformation rates, which has the same variation rule with the austenite dynamic recrystallization at high temperature. Then the 〈111〉//ND texture mainly results from recrystallized austenite. The content of the textures in quenched martensite is lower at the deformation temperature of 900°C, which is beneficial to reduce the anisotropy of workpieces. Therefore, the quenching temperature in the stamping process of the tested steel can be lower, for example, 900°C.

Conclusion
The following conclusions can be drawn from this work: (1) The processing maps adopting the instability criteria ξ P ( , e  T) and ξ B ( , e  T) are relatively more conservative compared with that using the instability criterion ξ M ( , e  T), which are more reliable. (2) HAGBs in quenched martensite increase with increasing deformation temperatures and deceasing strain rates of deformed austenite at the same strain, which also increase with increasing strain. In this work, the fraction values of HAGBs in quenched martensite are all less than 0.4 under all deformation conditions.
(3) For steel Docol 1500 Bor, no texture appears in quenched martensite at the engineering strain of 0.2. The deformed austenite texture mainly include 〈111〉//ND and 〈112〉//RD〉 components at the engineering strain of 0.6. The recrystallization texture component 〈111〉//ND in deformed austenite decreases with decreasing the deformation temperatures and increasing the strain rates. The flow location bands, the decrease of DRX and texture component 〈111〉//ND deteriorates the deformability of tested steels.
(4) Generally, the texture in quenched martensite mainly includes 〈110〉//ND and 〈110〉//RD components. The content of textures in quenched martensite is lower at the deformation temperature of 900°C. To reduce the anisotropy of quenched workpieces, the quenching temperature in the stamping process of the tested steel should be lower.