Mechanical behavior and microstructure evolution during high strain rate deformation of AA7075-T651

The current study presents the effects of strain and temperature on the mechanical response and microstructure evolution in AA7075-T651 at high strain rates. Compression tests have been performed at room temperature (RT), 200, 300 and 400 °C using a Split-Hopkinson pressure bar (SHPB) setup with strain rates ranging between 1400 and 5300 s−1. For deformation at RT, the flow stress increases with increase in strain rate. Whereas deformation at elevated temperatures show a non-monotonous behavior of the flow stress with respect to the strain rate. This trait is attributed to the pronounced effects from the adiabatic shear bands (ASBs); namely, distorted shear bands (DSBs) and transformed shear bands (TSBs); and cracks resulting from the plastic deformation instability during hot deformation. The sequence of microstructure evolution is: inhomogeneity in the initial microstructure – DSB – TSB – crack –fracture. The feasibility of formation and growth of ASBs and cracks increases with increase in strain and temperature, neglecting any significant effect from the strain rate. During the compression tests, temperature of the material rises due to adiabatic heating. Considering a certain strain developed in the material, this adiabatic temperature rise decreases as the deformation temperature is increased. Furthermore, during individual deformation processes, the temperature rise increases with increasing strain. The adiabatic temperature leading to the formation of TSB is approximated to be 0.7 times of the melting temperature of the alloy. These results from the current study are to be used in developing a physics-based material model for the alloy. Article Highlights At elevated temperatures, compression with Split-Hopkinson bars produce two types of shear bands and cracks. Evolution of shear bands and cracks is promoted by increase in strain and temperature irrespective of strain rate. Adiabatic temperature approximating to 70% of the melting point forms refined grain structure of transformed band. At elevated temperatures, compression with Split-Hopkinson bars produce two types of shear bands and cracks. Evolution of shear bands and cracks is promoted by increase in strain and temperature irrespective of strain rate. Adiabatic temperature approximating to 70% of the melting point forms refined grain structure of transformed band.


Introduction
Aluminum alloy 7075 (AA7075) in peak aged temper condition finds wide application in the structural parts in aerospace, defense and automotive industries due to the remarkable combination of properties, such as high mechanical strength, light weight, dimensional stability, machinability and specific stiffness, obtained through the T6 heat treatment [1][2][3]. T6 temper comprises the following two steps [4,5]: (i) solution heat treatment (SHT), which dissolves the second phase precipitates present in the Al matrix, and thus creates a super saturated solid solution (SSSS) upon quenching to room temperature (RT), and (ii) age-hardening, which leads to the formation of finely dispersed precipitates in the matrix according to the usual sequence in 7XXX series of Al-based alloys [6][7][8][9]: SSSS (α-phase) = > Guinier-Preston (GP) zones = > Metastable η′ phase = > Stable η (MgZn 2 ) phase. The η′ phase, which has a hexagonal crystal structure with lattice parameters of a = 0.489 -0.496 nm and c = 0.868 -1.403 nm [8,[10][11][12], is regarded as the main constituent strengthening factor owing to its abundance in the T6 state and its strong pinning effect on the dislocations [10,13,14]. The T651 temper of AA7075 alloy signifies a stress relief process by stretching the material by 0.5-2% after the T6 heat treatment [15].
The processing route of AA7075 includes hot forming, and therefore, many researchers have studied the deformation behavior of the alloy within a range of low strain rates between 1 × 10 -5 s −1 and 10 s −1 at ambient as well as elevated temperatures. However, the effects of higher strain rates on the mechanical properties of the alloy did not get much scientific or engineering attention until quite recently when several defense technologies as well as the increased usage of manufacturing techniques such as high-speed wire drawing, cold rolling, etc. required extended knowledge of the alloy behavior in such extreme conditions [16]. Thus, as compared to the low strain rate behavior of AA7075, fewer studies investigating the high strain rate behavior of the alloy within the order of 10 2 to 10 3 s −1 have been published so far, whose findings are summarized below.
The already existing literature data on mechanical behavior of AA7075 during its high strain rate deformation reveal that, for deformation at RT, the flow stress increases slightly when the magnitude of applied strain rate is increased within the order of 10 2 and 10 3 s −1 [17][18][19][20]. The ductility and strain rate sensitivity of the alloy at RT seem to increase with strain rates within these ranges [17,21]. Furthermore, Rahmaan et al. [22] have reported that the temperature rise due to adiabatic heating also increases with increase in the strain rate for deformation at RT. As the deformation temperature is increased at a constant strain rate, the flow stress of the alloy decreases significantly [21,23]. This decrease in the flow stress is amplified when the temperature increases from 200 to 300 °C [21].
From the failure mechanism and more specific microstructure evolution point of view, Lee et al. [23] reported that, at cryogenic temperatures with high strain rates, AA7075-T6 alloy fails due to loss of fracture resistance, whereas the failure at 0 °C takes place via localized melting due to adiabatic deformation. Sasso et al. [18] compared the high strain rate behavior of AA7075 in T6 and O (annealed) tempered conditions at RT. They found that the alloy in T6 state undergoes fracture mechanism characterized by cleavage, whereas O state experiences a microvoid coalescence fracture process. Lee et al. [21] inferred in their study that the high strain rate deformation behavior of AA7075-T6 between RT and 200 °C is governed by a dislocation drag mechanism. In the same study, they also reported possible dynamic recrystallization (DRX) in the alloy based on the grain size decrease observed at deformation temperatures between 25 and 300 °C.
In addition, adiabatic shear bands (ASBs) have also been observed to form in this alloy at these high strain rates [17,21,23]. During the deformation process, most of the plastic dissipated energy is converted into heat. In case of deformation with high strain rates, there is not sufficient time for the heat to get dissipated from the test specimen. As a result, the temperature of the material rises due to adiabatic heating, leading to a thermal softening effect. When the effect from thermal softening exceeds that from the work hardening, the plastic deformation becomes unstable and gets localized in narrow regions which are observed as ASBs. This is considered to be one of the most important failure mechanisms for many materials during high impact deformation, such as in explosive fragmentation, ballistic impact, high speed shaping and forming, dynamic compaction and welding, machining, grinding, etc. [24], and hence should be analyzed for AA7075 as well. Although Zhang et al. [25] and Khan et al. [26] have analyzed the formation of ASBs in Al-Zn-Mg-Cu alloys, their material composition was different than the standard composition recommended for industrial AA7075 [2]. ASBs in different heat treated states of AA7075 have previously been studied through the thick-walled cylinder (TWC) external explosive collapse technique which deliberately promotes the formation of shear bands in the material [27][28][29].
The above assessment reveals that the literature lacks data on the high strain rate deformation behavior of AA7075-T651, especially at elevated temperatures, thereby providing an open field for investigation into the effects of strain and strain rate at different deformation temperatures on the mechanical response and the microstructure evolution in the alloy at these high strain rates, which has been addressed to in the current study through compression tests using a Split-Hopkinson pressure bar (SHPB) [30][31][32] setup. Furthermore, the literature is deficient in studies on the analysis of ASBs and crack formation during the deformation of AA7075-T651 with SHPBs at elevated temperatures, and that is why it is the primary focus of our current work.
In the next section, we have detailed the methodology employed in the current study which can be used to reproduce our experimental results. Section 3 presents the observations made from the mechanical and microstructural characterizations. The explanations behind those observations are discussed in Sect. 4. Finally, Sect. 5 highlights the conclusions drawn from the current research.

Material
The material used in the current study was supplied by Alumeco [33]. According to the manufacturer, the initial extruded AA7075 cylindrical bars were drawn to a diameter of 10 mm, followed by a thermomechanical treatment corresponding to T651 temper conditions. The chemical composition provided by Alumeco is given in Table 1.

Mechanical characterization
In the current study, high strain rate compression tests were performed by using a SHPB setup which was previously implemented by Kajberg and Sundin [34]. Figure 1 shows a schematic representation of the setup. In our study, cylindrical samples having 10 mm diameter and 5 mm length were used for the tests with the compression direction parallel to the extrusion axis of the alloy. The tests were performed at RT, 200, 300 and 400 °C.
In order to carry out the compression tests, the sample was positioned in between and aligned with the centerline of the two pressure bars (incident and transmitter bars). For hot compression, the sample was heated with the help of an induction coil. The temperature of the sample was measured using a K-type thermocouple attached to the center (mid height) of its curved surface. Figure 2 is an example of the heating rate showing the temperature versus time during the heating of the sample for compression test at 400 °C. Once the sample was at the desired temperature (ambient or elevated), the striker bar was accelerated in an air gun tube and hit the incident bar, subsequently deforming the sample. The deformation in the sample was captured by the two strain gages attached to the pressure bars. Right after the impact, the sample immediately fell into the water filled container kept at RT (indicated by the dashed box in Fig. 1), ensuring rapid cooling to freeze the microstructural state obtained during deformation. More details about the placement of the sample, minimization of heat dissipation from sample to bar ends and movement of the bars can be found in the work by Kajberg and Sundin [34]. The resulting stress, strain and strain rate in the samples were calculated using the classical SHPB theory assuming uniform deformation under force equilibrium [16]. The strain rate produced during individual SHPB test was not constant throughout the compression, which is also evident in the literature [16,[35][36][37]. Therefore, each strain rate reported here in our study corresponds to the average of its distribution over the time during which it supposedly transmitted most of the energy to the sample in order to deform it. The strain rate for the tests was varied by changing the distance of the striker bar from the incident bar (known as projectile distance) and/or the pressure in the air gun tube. At each set of projectile distance and pressure, minimum two tests were performed in order to achieve a consistent strain rate. The strain rates obtained during deformation at different temperatures varied between 1400-5300 s −1 . The original stress-strain curves obtained from the data acquisition system were smoothened using a moving average algorithm.

Microstructural characterization
The alloy in the as-received (AR) state along with the samples after the high strain rate deformation were prepared using standard metallographic procedure for microstructure analysis. The surface perpendicular to the compression axis of the samples was mechanically ground using P240, P600, P1200 and P4000 grit SiC paper, and then polished with 1-µm diamond paste and 0.05-µm colloidal silica (OPS) solution. The samples were subsequently etched with Keller's reagent, having a chemical composition of 2 ml HF, 3 ml HCl, 5 ml HNO 3 and 190 ml H 2 O, for 20 s.
The macrographs and optical images of the etched samples were captured using a Nikon D7000 camera and a Nikon Eclipse MA100 inverted optical microscope (OM), respectively. The secondary electron (SE) images were acquired using Everhart-Thornley detector (ETD) and through-lens detector (TLD) in an FEI Magellan 400 field emission XHR scanning electron microscope (SEM). The current study did not employ advanced microscopic techniques such as electron back-scattered diffraction (EBSD) or transmission electron microscope (TEM), as the conclusions could be done using partially the data available in the literature.

Flow stress behavior
The true stress-strain curves obtained from the SHPB tests of AA7075-T651 comprised the deformation stage (during loading) as well as the unloading stage. Figure 3 represents such entire true stress-strain curves indicating the unloading stage for compression tests with 1400 s −1 at RT, 2000s −1 at 200 °C and 3200 s −1 at 300 °C. The maximum strain undergone by the material in a SHPB test corresponds to the strain at the end of the plastic deformation stage and the beginning of the unloading stage. Therefore, in the subsequent analysis of the flow stress behavior, the stress-strain parts pertaining to the unloading stages were excluded. Figure 4 shows the true stress-strain curves from the compression tests at all experimental conditions of temperature and strain rate, after excluding the unloading part. Thus, the true strain at the end of each curve corresponds to the maximum strain undergone by the material for the respective experimental condition. The curves corresponding to the lowest, intermediate and the highest strain rates obtained for each deformation temperature are presented by dotted, dashed and solid lines, respectively. It can be observed that the lowest flow stress level exhibited by the alloy at each deformation temperature is related to the largest strain for the corresponding temperature irrespective of the strain rate, with an exception at RT. The flow stress decreased with increase in deformation temperature from RT to 400 °C, as expected. Upon comparing the slope at the beginning of the plastic part of the stress-strain curves, the steeper positive slope in case of deformation at RT suggests that the thermal softening took place at a later stage as compared to in case of deformation at elevated temperatures. Furthermore, compared to other temperatures, deformation at 200 °C displayed a relatively wide spectrum of flow stress levels for the obtained strain rates. In order to study the effect of strain rate on the flow stress, the true stress values at a true strain of 0.1 are plotted against the strain rates for all deformation temperatures in Fig. 5. For deformation at RT, the flow stress increased with strain rate. This can be explained by the strain rate hardening effect. In this case, as the strain rate is increased, the time available for achieving a certain strain in the material is shorter; thus, higher stress needs to be exerted for moving the dislocations past the obstacles in order to produce the desired strain. On the other hand, in case of deformation at elevated temperatures, the flow stress exhibited non-monotonous behavior with respect to the strain rate.

Microstructure of AR material
To understand the microstructure evolution during deformation, the microstructure of AA7075-T651 in AR state was first analyzed. Figure 6a shows the macrograph of the surface of AR material perpendicular to the extrusion direction. Figure 6b1 and b2 are the OM images at the approximate center (area enclosed by black rectangle in Fig. 6a) and close to the edge (area enclosed by red rectangle in Fig. 6a) of the specimen, respectively. A near equiaxed grain structure is observed in the center of the sample, whereas, close to the edge, the grains are comparatively elongated having an arc-shaped morphology with their radius approximately facing towards the center  of the sample. As it is discussed later, this kind of initial grain morphology, which is typically imparted in the material during the processing, played a significant role in the subsequent microstructure evolution during the deformation of the material.

Macrostructure evolution in deformed specimens
The macrographs after deformation at different conditions are shown in Fig. 7. No noticeable damage was observed in case of deformation at RT and 200 °C ( Fig. 7a and b, respectively). However, more detailed investigation presented in the succeeding sections revealed the formation of ASBs at 200 °C. The appearance of ASBs became very conspicuous at macroscopic level for deformation at 300 °C, as can be seen in Fig. 7c-e. Deformation at 400 °C with 4100 s −1 strain rate produced a crack along with ASB ( Fig. 7f ). At an increased strain rate of 4900 s −1 for the same temperature, the crack propagation became very severe, leading to the rupture of part of the specimen (Fig. 7g). In addition, secondary micro-cracks were formed from the primary crack along a radial direction of the sample surface. The specimen compressed with 5300 s −1 at 400 °C underwent complete failure (Fig. 7h).
All the ASBs and primary cracks shown in Fig. 7 have an arc-shape, in some cases near-circular shape morphology, with their radius facing towards the center of the sample surface. Similar observation in case of high strain rate compression of AA7075 alloy has previously been reported by Lee et al. [21]. It can also be inferred that the ASBs and cracks have developed in the region close to the edge of the original sample. Furthermore, even at this macroscopic level, ASBs developed in these specimens seem to exhibit more than one morphology.

Types of ASBs
The OM images indicated the formation of two types of grain morphologies related to the ASBs: (i) distorted elongated grains, as shown in Fig. 8a for the specimen deformed with 4700 s −1 at 200 °C, and (ii) recrystallized ultra-fine grains (UFGs), as shown in Fig. 8b for the specimen deformed with 4300 s −1 at 300 °C. These grain structures are characteristic features of two types of ASB, i.e., distorted shear band (DSB) and transformed shear band (TSB), respectively [24,25,27,38,39]. The DSB itself has a concentric arc-shape, however, the distorted grains constituting the DSB are perpendicular to this arc (Fig. 8a). It is also observed that the DSBs form alone (Fig. 8a), whereas the TSBs are accompanied by the DSBs (Fig. 8b).

Formation of DSBs
The DSB shown in Fig. 8a is quite prominent and fully developed. In contrast, the DSB shown in Fig. 9a, which was formed in specimen deformed with 2000 s −1 at 200 °C, seems to be in its initiation stage and not quite developed or prominent yet. It also has been formed as an arc close to the edge of the specimen. Moreover, the grains around the DSB have an elongated arc-shaped morphology. As already indicated in Sect. 3.2, this kind of microstructure at the edge of the sample was imparted during the initial drawing process and seems to be retained even after the deformation. Therefore, the distorted elongated grains in DSB can be supposed to have been originated from the initial arc-shaped elongated grains close to the edge of the sample. Figure 9b, corresponding to the specimen deformed with 3200 s −1 at 300 °C, shows the presence of arc-shaped grains at the edge of the sample and their gradual transformation into fully distorted elongated grains. Figure 10 shows the initiation site for the TSBs in two specimens deformed at 300 °C with 4300 and 5000 s −1 . It is quite evident that the TSBs have started to originate in the area where the elongated grains in DSBs have become so distorted that they localize in a very narrow region, thereby resulting in the formation of UFGs by subsequent recrystallization. The continuation of localization of severely distorted grains leads to the propagation of TSB. This could explain the observation made in Sect. 3.4.1 that the TSBs are always accompanied by the DSBs. Figure 11a is a zoomed-in image of a portion of the primary crack shown in Fig. 7g, corresponding to the specimen deformed with 4900 s −1 at 400 °C. It can be seen that the crack propagation takes place at the boundary between the distorted grains, characteristic feature of DSB, and the UFGs, indicative of TSB. This crack was back traced to its initiation point, shown in Fig. 11b, which revealed that the crack was originated from the TSB and further propagated along the interface between the TSB and DSB.

Formation of cracks
In order to analyze the initiation of the crack, SEM analysis was performed on the primary crack shown in Fig. 7f corresponding to the specimen deformed with 4100 s −1 at 400 °C. Figure 12a shows the entire primary crack. Figure 12b is the magnified image of the crack initiation site as highlighted by the red box in Fig. 12a, showing the propagation of a micro-crack which supposedly engineered the initiation of the crack. The micro-crack is seen to be surrounded by UFGs that are characteristic of the TSB. The area highlighted by the yellow box in this image indicated the presence of a void constituting the micro-crack. Upon zooming-in onto this area, it manifested the micro-voids generated amid the UFGs (Fig. 12c) that had subsequently led to the formation of the micro-crack. From this image, the grain size inside the TSB that constituted the crack could be estimated to be of less than 100 nm. Previously researchers have reported similar grain size of less than 200 nm within the TSB for different Al alloys by employing sophisticated microscopic techniques such as EBSD and TEM [26,27,38]. Figure 12d is the magnified image of the crack termination site as highlighted by the black box in Fig. 12a, with Fig. 12e showing the exact termination point. In contrast to the initiation site, no noticeable micro-crack or micro-voids could be found at the termination site. It can be noted that this kind of crack formation could subsequently lead to the fracture of the alloy, which apparently did occur in our case, as illustrated in Fig. 7g and h.

Deformation at 200 °C
By referring to the stress-strain curves for deformation at 200 °C in Fig. 4, the true strain of ~ 0.18 and ~ 0.20, which were respectively induced by strain rates of 2500 and 2000s −1 , initiated DSB in the material. In both cases, the DSB was neither quite developed nor very prominent, one of which has been shown in Fig. 9a. The largest strain that could be obtained at this deformation temperature was ~ 0.27 with 4700 s −1 , and in this case, the DSB produced in the microstructure was quite fully developed (Fig. 8a). Figure 13 shows the OM images of the specimens deformed with 4300 and 5000 s −1 at 300 °C. Both strain rates produced TSB in the material. It can be observed that when the strain developed in the material increased from ~ 0.50 to ~ 0.57 (corresponding to 5000 s −1 in Fig. 13a and 4300 s −1 in Fig. 13b, respectively), the intensity of the TSB has been increased.

Deformation at 400 °C
For the deformation temperature of 400 °C, the least strain of ~ 0.36 from 4100 s −1 resulted in the formation of a crack (Fig. 7f ). Increase in the true strain to ~ 0.41, corresponding to 4900 s −1 , promoted severe crack propagation along with secondary micro-cracks formation leading to partial rupture of the material (Fig. 7g). The largest true strain of ~ 0.65, that could be attained with 5300 s −1 strain rate, led to the complete fracture of the specimen (Fig. 7h).

Estimation of adiabatic temperature rise
The adiabatic temperature from the high strain rate impact is typically estimated by two methods: (a) overall temperature of the sample based on the total mechanical work input [24,27,38,40], and (b) localized temperature inside the ASB according to von Mises criterion [26,41,42].
The estimation of overall sample temperature considers that most of the energy from the plastic deformation is converted into heat and thereby increases the temperature of the sample in general. This temperature rise due to adiabatic heating is calculated by Eq. (1): where T is the final temperature of the material including the adiabatic temperature rise, T 0 is the test temperature, is the fraction of plastic work converted into heat (referred to as Taylor-Quinney parameter), is the density of the material, C is the heat capacity of the material and ∫ 0 d is essentially the area under the stress-strain curve.
The calculation based on the von Mises criterion considers more localized strain from the microstructure evolution during the deformation and is given by Eq. (2): where is the angle between the streamline of ASB and the orientation of original grain structure. A better microscopic understanding of can be found in the work reported by He et al. [41] and Wu et al. [42]. The resultant strain is then substituted in Eq. (1) to calculate the temperature. This method provides a better estimation of the localized temperature in an ASB. However, it requires more careful observation into the microstructure to obtain information about the exact locations of the ASBs, whereas the method for overall temperature estimation is quite straightforward and easy to perform. Thus, the adiabatic temperature rise in our work is estimated using the overall temperature method. The specific heat capacity value, C , for AA7075 at RT, 200, 300 and 400 °C is taken as 960, 974, 1012 and 1128 J/ (kg-°C), respectively [2,43,44]. However, due to the lack of experimental data in the literature, the Taylor-Quinney parameter, , and the density, , for AA7075 are assumed to be independent of strain, strain rate and temperature. Here we have used = 0.9 (assuming 90% of the plastic work is converted into heat) and = 2810 kg/m 3 [2]. Substituting these values along with the mechanical work input (i.e., the area under each stress-strain curve shown in Fig. 4) in Eq. (1), the adiabatic temperature rise is estimated in case of each experimental condition. Figure 14a displays the temperature rise in the material up until true strain of 0.1, whereas Fig. 14b shows the temperature rise due to the maximum strain (indicated within parentheses inside the graph) undergone by the material for all testing temperatures and strain rates.
As can be observed from Fig. 14a, when the material is deformed up to a true strain of 0.1, the temperature rise due to adiabatic heating decreases with increasing deformation temperature. On the other hand, when the adiabatic temperature rise is estimated up to the maximum strain attained by the material during deformation (Fig. 14b), the abovementioned trend of temperature rise with respect to deformation temperature is not obvious anymore. In both cases shown in Fig. 14, the adiabatic temperature rise increases with strain rate for the tests conducted at RT, whereas it does not follow any pattern with respect to strain rate for the tests performed at elevated temperatures. Furthermore, Fig. 14b manifests that, at a constant deformation temperature, the highest adiabatic temperature rise is induced for the largest strain attained by the material irrespective of the strain rates.

Deformation path leading to the failure of material
Based on the formation and growth of ASBs and cracks demonstrated in Sects. 3.4 and 3.5, a sequence of microstructure evolution leading to the final fracture of AA7075-T651 during high strain rate compression is schematically represented in Fig. 15. The loading direction during the compression test is perpendicular to the plane of the schematic diagrams shown in the figure. The arc-shaped elongated grains in the initial microstructure close to the edge of the drawn rods (Fig. 6b2) are simplistically represented by straight lines in the schematic diagrams. As per previous reports, the inhomogeneities, such as small notches, precipitates, inclusions or other imperfections, that are inevitably present in the microstructure of a material, act as nucleation sites for the formation of ASBs [27,29,45]. In our case, the typical arc-shaped grain structure appears as a suitable site for the instability in plastic deformation to take place, thereby promoting the initiation of ASB. As there exists a critical strain for the development of ASB during dynamic deformation [24], the initial arc-shaped grains in our alloy get distorted and initiate a DSB once the strain reaches the critical value. Further localization of strain after this critical value leads to the growth of DSB both in width and length. When the intensity of strain localization attains a critical value, the distorted grains of DSB get restricted inside a confined region [38], thereby further increasing the local temperature and subsequently forming recrystallized UFGs which characterize a TSB. This is in agreement with the conclusion made by Timothy [46] that the formation of DSB corresponds to an earlier stage of adiabatic strain localization, whereas that of a TSB represents an advanced stage of the process. Once a TSB is generated, further strain localization leads to its propagation. Owing to a greater thermal softening because of the high temperature inside the TSB, the material within the shear band is softer than that in the surrounding matrix, which is comparatively colder, resulting in a lower flow stress inside the TSB [47,48]. This leads to the nucleation of micro-voids among the UFGs, which readily grow and coalesce under the effects of heat and shear stress within the TSB to form cracks [45,48]. The origination of crack from TSB has also been reported in case of AA7075-T73 alloy tested under TWC external explosive collapse technique [27]. The crack subsequently leads to the complete fracture of the material.

Effect of strain on microstructure evolution
The microstructure evolution observed in all deformed specimens have been summarized in Fig. 16 alongside their corresponding true stress-strain curves. It is quite logical that larger is the strain induced in the material, greater is the probable extent of strain localization into narrow regions, and hence higher is the feasibility of formation and subsequent development of ASBs and cracks. This can account for the microstructure evolution pertaining to the plastic strain induced in the material reported in Sect. 3.5, as explained below. For compression at 200 °C, the true strains of ~ 0.18 and ~ 0.20 accumulated just enough strain concentration for the initiation of a DSB (Fig. 9a shows the DSB developed at the strain of ~ 0.20). However, the amount of strain induced in both cases were not sufficient for the subsequent growth of the DSB. When the strain in the material  increased up to ~ 0.27 at the same deformation temperature, it provided enough strain localization for the growth of the DSB into a relatively developed and prominent structure (Fig. 8a). At 300 °C, the strain of ~ 0.33 could only produce a DSB in the material (Fig. 9b). Whereas, when the true strain was increased to ~ 0.50, it led to the formation of TSB from a DSB (Fig. 10b). Further increase in the strain to ~ 0.57 intensified the strain localization, thereby escalating the confinement of the distorted grains of the DSBs to form greater number of recrystallized UFGs, which emerged as increased intensity of the TSB (Fig. 13). Likewise, when the strain increased from ~ 0.36 to ~ 0.41 and then finally to ~ 0.65 at 400 °C, the strain localization became more and more severe, thereby progressively intensifying the constituent TSB and hence the crack propagation, ultimately leading to failure of the specimen at ~ 0.65 strain (Fig. 7f-h).
Furthermore, Fig. 13 shows that the intensity of TSB at a true strain of ~ 0.57 with 4300 s −1 at 300 °C is greater than that at a strain of ~ 0.50 produced by a "higher" strain rate of 5000 s −1 at the same deformation temperature. Therefore, it can be suggested that the evolution of ASBs and cracks is strongly influenced by the plastic strain rather than the strain rate.

Effect of temperature on microstructure evolution
As summarized in Fig. 16, deformation at RT did not produce any ASB in the material. As the deformation temperature was increased from RT to 400 °C, it resulted in the gradual formation of DSBs, TSBs and cracks in the material. Therefore, it is inferred that the increase in temperature promotes the plastic strain localization, which contradicts the observations reported by Zhang et al. [25]. Here it can be argued that this kind of severe strain localization in our case may have been stimulated by the true strain rather than the deformation temperature, as emphasized before in Sect. 4.2. However, Fig. 16 also displays that a strain of ~ 0.27 with 3600 s −1 at RT did not produce any ASB, whereas a smaller strain of ~ 0.20 with a "lower" strain rate of 2000s −1 at 200 °C generated a DSB in the material. Furthermore, the material deformed up to a true strain of ~ 0.50 with 5000 s −1 at 300 °C developed a TSB, whereas deformation up to a lower strain of ~ 0.36 with a "lower" strain rate of 4100 s −1 at 400 °C produced a crack in addition to a TSB. Thus, it can be deduced that the evolution of ASBs and cracks is influenced more by the deformation temperature along with the plastic strain rather than the strain rate. Moreover, an increase in the deformation temperature promotes the formation and growth of ASBs and cracks.

Effect of temperature, strain and strain rate on the adiabatic temperature rise
As per Eq. (1), the adiabatic temperature rise during a given compression test is directly proportional to the area under the corresponding stress-strain curve. At strain rates with magnitude of the same order and a constant strain, the flow stress decreases with deformation temperature, thereby reducing the area under the stress-strain curve. As a result, the adiabatic temperature rise also decreases with deformation temperature, which is depicted in Fig. 14a. Similar results for AA7075 can be found in case of deformation with both high (in the order of 10 3 s −1 ) [21] as well as low (between 0.1 and 10 s −1 ) [43,49] strain rates. Furthermore, as the maximum strain obtained in all our tests are different (Fig. 4), the estimated adiabatic temperature rise corresponding to the entire deformation process does not clearly display the abovementioned decreasing trend with respect to the deformation temperature (Fig. 14b), suggesting that the strain accumulated in the material also plays a significant role in the rise in temperature.
Increasing the strain in the material increases the amount of heat generated by the mechanical work, which leads to the rise in temperature of the material. Such increase in the temperature rise with increase in strain is quite common at a constant strain rate, which have been reported for different materials [50][51][52]. Similarly, in our case, the area under the stress-strain curve at a constant deformation temperature and strain rate increases with increase in the strain (Fig. 4), implicating an increase in the resultant adiabatic temperature rise. In addition, it is also quite conspicuous from Fig. 14b that, at a constant deformation temperature, the material undergoing the largest strain irrespective of the strain rate experiences the highest temperature rise. Here it can be argued that the adiabatic temperature rise estimated for 200 °C does not seem to increase with increase in the strain. This can be explained by the severity of the DSB formed (Fig. 8a) in case of the largest strain (~ 0.27 at 4700 s −1 ) induced in the material, which brings down the corresponding flow stress level, thus reducing the mechanical work input from the plastic deformation. This characteristic influence of formation and intensity of ASBs on the flow stress behavior is further illustrated in Sect. 4.7.
During the SHPB test, the impact energy from the incident bar is partially converted into adiabatic temperature rise. So, theoretically, at a constant deformation temperature, when the strain rate is increased by increasing the projectile distance and/or pressure in the air gun tube (as described in Sect. 2.2), more amount of impact energy is supposed to be transmitted to the material, leading to an increase in the adiabatic temperature rise. Our results corresponding to deformation at RT (Fig. 14) corroborates this theory. Similar results of increase in temperature rise with increasing strain rates in the order of 10 2 -10 4 s −1 for RT deformation have previously been reported for different materials [22,42,50,[53][54][55]. However, the abovementioned theory does not seem to be valid in case of our material for deformation at elevated temperatures (Fig. 14). In these cases, unlike at RT, increase in the strain rate does not always result in an increase in the temperature rise. The most plausible explanation for this is, at elevated temperatures, the effect of strain rate gets dominated by the formation of ASBs and cracks resulting from the combined effects of deformation temperature and the plastic strain induced in the material. However, the literature lacks experimental data for high strain rate deformation of materials performed at elevated temperatures to compare with this explanation. Figure 16 depicts that TSB, which typically comprises recrystallized grains owing to quite high temperature inside the shear band, is formed in case of deformation at 300 °C for two out of the three strain rates. The final temperatures for 4300 and 5000 s −1 strain rates that produced TSBs in the deformed samples are calculated (as described in Sect. 3.6) to be 348 and 343 °C, respectively. The corresponding values in SI units are 621 and 616 K, which are ~ 0.7T m , considering T m for AA7075 = 908 K, where T m is melting temperature of the material.

Adiabatic temperature for the formation of TSB
The recrystallization temperature for a metallic alloy is usually taken as 0.4-0.5T m [24,27,38]. The temperature estimated by previous researchers in case of different Al alloys, that supposedly led to the formation of recrystallized grains in TSBs, ranges between 0.42-0.8T m [26,27,40,41]. In our case, the temperatures calculated above using a rather generalized method of overall sample temperature, ~ 0.7T m , are already well within these ranges. So, it is safe to presume that the usage of more accurate method of localized strain based on von Mises criterion, which calculates a higher temperature in localized regions compared to the overall sample, would definitely estimate the temperature inside the TSB to be more than 0.7T m . Thus, it can be suggested that when the temperature of AA7075-T651 during its dynamic deformation approximates to 0.7T m , it can generate refined grain structure constituting a TSB via recrystallization.

Probable mechanism behind recrystallization within TSB
Recrystallization (RX) is the formation of a new grain structure in a deformed material by the formation and migration of high angle grain boundaries (HAGBs) driven by the stored energy introduced by plastic deformation [56]. The recrystallization taking place during deformation at elevated temperatures is known as dynamic recrystallization (DRX) [57]. AA7075 alloy has previously been reported to have undergone DRX via different mechanisms, such as continuous (CDRX) [49,[58][59][60][61], discontinuous (DDRX) [59] and geometrical (GDRX) [49,59,60], during its deformation with strain rates ranging between in the order of 10 -3 s −1 and 10 s −1 . At these low strain rates, the material could undergo DRX owing to the ample time available for the completion of the deformation process. However, in our current scenario, when the deformation is carried out at a very high strain rate, the actual deformation time is much shorter for the material to undergo DRX via conventional grain boundary migration mechanism. Moreover, the time taken for the formation of the TSB is lower by several orders of magnitude than the time required to form recrystallized grains by grain boundary migration [24]. So, the conventional mechanism of DRX cannot be employed to explain the formation of recrystallized UFGs within the TSBs.
On the other hand, Meyers et al. [62][63][64] had proposed a rotational recrystallization mechanism which may be able to explain the feasibility of the formation of refined grain structure of a TSB during a very short period of time. As per this mechanism [24], a random dislocation distribution with high energy configuration takes place at the beginning of the deformation. This random distribution leads to the formation of elongated cells. As the deformation continues and the misorientation increases, these cells become elongated sub-grains. These elongated sub-grains get further plastically deformed, subsequently generating approximately equiaxed refined grains. Lately this mechanism has been widely accepted by many researchers in order to account for the formation of TSBs during high strain rate deformation of different Al alloys [24,27,42].

Flow stress behavior
Conventionally, the flow stress increases with increase in strain rate owing to the strain rate hardening effect. Moreover, in case of SHPB tests at a certain temperature, larger is the combination of the projectile distance and the pressure in air gun tube, greater is the supposed energy being transmitted to the sample, and consequently higher is the extent of deformation undergone by the material. Thus, in case of compression at RT, the highest strain rate generated the highest flow stress level, in agreement with the reports available in the literature [17,18,21], and simultaneously induced the largest strain in the alloy (Fig. 4).
On the other hand, deformation at elevated temperature involves enhanced diffusion of atoms due to thermal activation and the formation of ASBs and cracks due to instability in the plastic deformation. The combined effects of these phenomena overturn the effect of the strain rate mentioned above. As discussed in Sects. 4.2 and 4.3, a larger strain as well as high temperature enhances the feasibility of the formation of ASBs and cracks irrespective of the strain rate. These ASBs and cracks are the resultant of pronounced strain localization which essentially facilitates the reduction in the flow stress level. The decreasing order of the flow stress exhibited by the alloy is in the sequence of specimen having DSB, then the one with TSB and finally the one that has crack in the deformed microstructure. However, to be more accurate, the flow stress level is determined by not only the formation of the crack and the specific type of the ASB but also by their intensities. For example, in case of deformation at 200 °C, all the three strain rates produced DSB in the alloy, but the deformation which produced the largest strain (~ 0.27), and hence the most intense DSB (Fig. 8a), exerted the lowest flow stress level among the three conditions (Fig. 16).
Since the mechanical behavior during hot compression depends strongly on the microstructure evolution, which is principally driven by strain and deformation temperature in our case, it explains the inconsistency in the trend of flow stress behavior with respect to strain rate displayed in Fig. 5. The said irregularity of the strain rate effect at elevated temperatures is in contradiction with the data presented by El-Magd et al. [17]. They reported an increase in the flow stress with an increase in the strain rate having magnitude in the order of 10 3 s −1 for deformation at 200 °C. The discrepancy in our observations and the one by El-Magd et al. [17] is attributed to the absence of shear bands in the study by the latter.
The plastic deformation part of individual stress-strain curve is the resultant of the competing effects from strain hardening and thermal softening. The deformations at RT lack diffusion of atoms as well as ASB/crack formation. Therefore, the strain hardening effect dominates over the softening effect, manifested by the steeper positive slope at the beginning of the plastic part in the stress-strain curves in Fig. 4, up to a larger true strain. Whereas in case of elevated temperatures, due to the enhanced diffusion of atoms and the instability in plastic deformation, the effect of thermal softening takes place at an earlier stage. As a result of which, the steepness of the slope at the beginning of the plastic part is reduced within a lower true strain range.
At these extreme strain rates in the order of 10 3 s −1 , which are not constant throughout the actual compression with SHPBs, it is difficult to determine the exact condition of strain, strain rate and temperature for the formation of different ASBs and cracks in the alloy during the deformation process. The reason being that if the loading condition (a combination of strain, strain rate or temperature) is close enough to the threshold of formation of ASB/crack, then a minor change in that condition whatsoever would significantly influence the microstructure as well as the resultant stress-strain curves, as compared to the deformation under more stable conditions. The results of which can actually be observed in Fig. 4 as the large scattering in the flow stress levels for deformation at 200 °C.

Conclusions
The mechanical and microstructural behavior of AA7075-T651 alloy under compression with high strain rates of magnitude in the order of 10 3 s −1 at temperatures between RT and 400 °C have been studied and analyzed with the help of a SHPB setup, and macro-and microscopic tools. The major takeaways from this study are as following: 1. The flow stress increases with increase in strain rate at RT, thereby following the conventional trend. However, it does not display any tendence with respect to strain rate at elevated temperatures. In these cases, the flow stress is rather determined by the resultant effect from the formation and growth (in intensity) of the ASBs and cracks. 2. ASBs are observed perpendicular to the compression axis for deformation at elevated temperatures. They are formed as an arc with radius facing towards the center of the specimen. Two types of ASB are identified, namely, (i) DSB: characterized by distorted elongated grains, and (ii) TSB: characterized by recrystallized UFGs. 3. The arc-shaped elongated grain structure close to the edge of the as-received circular bar, which is a typical microstructure induced from the initial drawing process, acts as the nucleation site for the formation of DSB. DSB is developed via distortion of those parent arc-shaped elongated grains. TSB is developed from the extreme localization of distorted grains of DSB into confined regions producing recrystallized equiaxed UFGs. Crack is produced via nucleation, growth and coalescence of micro-voids inside a TSB. The severity of crack can lead to the failure of the material. 4. At elevated temperatures, the microstructure evolution in the alloy is strongly influenced by the plastic strain and the deformation temperature rather than the strain rate. Localization of plastic strain, and hence formation and growth of ASBs and cracks are promoted with increase in strain and deformation temperature. 5. Adiabatic heating resulting from the high strain rates increases the temperature of the material. At a constant strain, this adiabatic temperature rise decreases with increase in deformation temperature. And during individual deformation process, it increases with increase in strain. The temperature rise also increases with strain rate but only at RT. At elevated temperatures, the strain rate effect loses its significance. 6. When the adiabatic temperature of the alloy approximates to 0.7T m , it can lead to refined grain structure constituting a TSB via recrystallization.
The results from the current study pertaining to the mechanical and microstructural characterizations for high strain rate compression of AA7075-T651 along with the corresponding ones for low strain rate deformation at 0.01 and 1 s −1 , which have already been completed previously [49], will be used to develop a physics-based material model for the alloy that can predict the final state of the material at the end of the manufacturing chain. Data availability The raw and processed data required to reproduce these findings cannot be shared at this time as the data also forms part of an ongoing study.

Declarations
Conflict of interest All the authors, namely, Biswajit Dalai, Marie Anna Moretti, Paul Åkerström, Vladimir A. Esin and Lars-Erik Lindgren, declare that they have no conflicts of interest.
Ethical approval This article does not contain any studies with human participants or animals performed by any of the authors.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.