A Review on the Adiabatic Shear Banding Mechanism in Metals and Alloys Considering Microstructural Characteristics, Morphology and Fracture

Abstract

The current review work studies the adiabatic shear banding (ASB) mechanism in metals and alloys, focusing on its microstructural characteristics, dominant evolution mechanisms and final fracture. An ASB reflects a thermomechanical deformation instability developed under high strain and strain rates, finally leading to dynamic fracture. An ASB initially occurs under severe shear localization, followed by a significant rise in temperature due to high strain rate adiabatic conditions. That temperature increase activates thermal softening and mechanical degradation mechanisms, reacting to strain instability and facilitating micro-voiding, which, through its coalescence, results in cracking failure. This work aims to summarize and review the critical characteristics of an ASB’s microstructure and morphology, evolution mechanisms, the propensity of materials against an ASB and fracture mechanisms in order to highlight their stage-by-stage evolution and attribute them a more consecutive behavior rather than an uncontrollable one. In that way, this study focuses on underlining some ASB aspects that remain fuzzy, allowing for further research, such as research on the interaction between thermal and damage softening regarding their contribution to ASB evolution, the conversion of strain energy to internal heat, which proved to be material-dependent instead of constant, and the strain rate sensitivity effect, which also concerns whether the temperature rise reflects a precursor or a result of ASB. Except for conventional metals and alloys like steels (low carbon, stainless, maraging, armox, ultra-high-strength steels, etc.), titanium alloys, aluminum alloys, magnesium alloys, nickel superalloys, uranium alloys, zirconium alloys and pure copper, the ASB propensity of nanocrystalline and ultrafine-grained materials, metallic-laminated composites, bulk metallic glasses and high-entropy alloys is also evaluated. Finally, the need to develop a micro-/macroscopic coupling during the thermomechanical approach to the ASB phenomenon is pointed out, highlighting the interaction between microstructural softening mechanisms and macroscopic mechanical behavior during ASB evolution and fracture.

1. Introduction

Plastic deformation reflects a fundamental mechanism that dominates several applications, from industrial and manufacturing processes to the structural integrity and operability of mechanical systems. In all cases, the plastic deformation mechanism can be described either as undesirable regarding structural integrity cases or as necessary in the case of forming manufacturing processes where homogenous and controllable deformation is required. However, plastic deformation can be dominated by strain instabilities reacting to uncontrollable and catastrophic fractures. From this aspect, a deep understanding of the dominant mechanisms that drive plastic deformation evolution, as well as the connection between unstable plasticity and fracture, allows for the prevention of undesirable deformation instabilities and the optimization of manufacturing processes by maximizing productivity and minimizing material defects.

Metals and alloys develop stable and homogenous plastic deformation under low strain and strain rates, with slipping and twinning being the dominant mechanisms for the expansion of dislocations and deformation progress. In contrast, at higher strain and strain rates, unstable deformation mechanisms such as adiabatic shear banding (ASB) may occur, leading to dynamic failure. Marchant and Duffy [1] reported that high-strain-rate plastic deformation develops at three stages as the strain increases: stable and homogenous deformation, unstable region, and dynamic failure. The genesis of ASB has coincided with plastic instability initiation, where shear bands are firstly manifested via severe shear localization along narrow zones, leading next to microstructural softening mechanisms and local mechanical degradation, bringing plastic instability and finally providing uncontrollable and dynamic damage. In particular, the intense shear localization reacts to significant deformation energy, which is converted to internal heat by a portion of about 90%, as expressed by the Taylor–Quinney coefficient [2]. In fact, high strain rates do not allow for enough time for thermal diffusion, representing adiabatic conditions, as described by Zener and Holloman [3], storing the internal heat inside the ASB and resulting in a significant temperature increase, which thermally softens the microstructure inside the ASB. Further, the temperature increase due to the produced heat can lead to additional degradation mechanisms except for thermal softening, such as dynamic recrystallization (DRX), as mentioned by Landau et al. [4], which both facilitate ASB expansion and uncontrollable fracture under rapid evolution [5]. For this reason, the ASB mechanism has been described as a thermoplastic shearing phenomenon [6], reflecting a thermomechanical instability that is generated when the thermal softening effect on plastic flow stress overcomes strain and strain rate hardening, thus leading to unstable deformation and final dynamic failure [7,8].

Due to their dynamic and uncontrollable behavior, a large amount of important research interest on ASBs has been concentrated on understanding their morphology and microstructural characteristics in order to predict instability and fracture. Bai et al. [5] identified a shear banding width of about 10–100 μm in steels under a 100 s⁻¹ shear strain rate ($\dot{\gamma }$), while Dodd and Bai [9] proposed an analytical expression for the ASB width. Also, several studies have focused on revealing a macroscopic criterion for ASB formation, with the majority agreeing on a critical shear strain and critical shear strain rate to be required for adiabatic shearing occurrence. Specifically, Timothy and Hutchings [10] proposed that an ASB is generated at the maximum load in the shear stress–strain (τ–γ) curve after a critical strain, while Rogers [6] and Recht [11] reported the critical strain rate as the key factor for ASB initiation. In addition, Zurek [12] revealed a 0.5 critical strain and a 1.8 × 10⁴ s⁻¹ critical strain rate for adiabatic shearing instability in 4340 steels, while Wang et al. [13] provided experimental $\dot{\gamma }$–γ maps for ASB formation and cracking progress for TB2 titanium alloy. Moreover, Bai [14] and Wright [15] introduced analytical expressions regarding the propensity of materials against an ASB, revealing that the strain and strain rate hardening parameters, thermal softening parameter, thermal diffusivity and critical strain/strain rate are the dominant material properties that affect its sensitivity in the ASB. Both expressions suggested that deformation instability occurs when the thermal softening effect on plastic flow stress prevails over that of the strain/strain rate hardening. Finally, Xu et al. [16] differentiated ASBs into “deformed bands” and “transformed bands”, with the latter ones mainly being reported in ferrous alloys. In fact, ASBs initially appeared as deformed bands at moderate strain levels via severe shear localization and the intense distortion of the deformed microstructure, which leads to elongation of grains along the ASB with possible fragmentation, while a transition to a transformed band occurs at a higher strain and strain rate, where a significant temperature rise activates the phase transformation and/or recrystallization mechanisms, further leading to dynamic failure via micro-voiding coalescence and cracking propagation [4,17]. The consecutive stages of ASB evolution from genesis to fracture are illustrated in Scheme 1 via their discrete events.

The formation of ASBs in manufacturing processes has also been investigated, aiming to minimize their presence and, consequently, material defects in the optimum process and tool designs with respect to material sensitivity against ASBs. Regarding ASB formation in the bulk and sheet forming of metals, forging and rolling processes have been studied regarding bulk deformation and punching, blanking and deep drawing processes for sheet forming. In forging processes, Tresca [18] first reported the appearance of shear bands through X-shaped single or multiple patterns along the workpiece cross-section, also describing them as heat zones due to the intense temperature rise caused by severe shear localization. On many occasions, ABSs follow the trajectories of slip lines during the deformation progress, affecting the forging force as supported by the Upper Bound Method (UBM) [19]. Semiatin and Lahoti [20] studied ASB formation in the hot forging of the Ti-6242 alloy, introducing a flow localization parameter for ASB initiation and suggesting that ASBs start to manifest as X-shaped bands. Also, Tang et al. [21] investigated the morphology of shear banding in the forging of the Inconel 718 nickel-superalloy. The results indicated that a low temperature and high strain rate loading conditions benefit the formation of S-shaped ASBs, in contrast to mild loading conditions, where X-shaped bands are formulated. Finally, shear bands can be separated into stronger and weaker ones depending on interfacial friction and non-uniform instantaneous heating. Moreover, Kang et al. [22] studied the cold forging of carbon steels, indicating diagonal shear bands with possible 45° deep cracking inside the transformed shear bands in the case of pearlitic steel, while steels of higher pearlite volume fractions proved to be more sensitive to ASB development. In the case of the rolling process, Wei et al. [23] reported two pairs of shear bands at the early stages of cold rolling of pure tungsten, while dynamic loading facilitates ASB formation due to the superiority of the thermal softening effect. Wang et al. [24] studied the hot rolling of the Mg-6Zn-0.5Zr magnesium alloy, reporting DRX at the grain boundaries for a low rolling reduction, while the increase in rolling reduction reacted to the ASB widening at 45° to the rolling direction [25]. Regarding sheet forming processes, ASBs may occur in the cup wall area during deep drawing via linear external stripes and V-shaped cracking fracture due to the coalescence of two shear bands. Finally, S-shaped ABSs have been reported to the adiabatic blanking process of steels [26,27] initiating around the punch and die corners, while greater punching velocity results in longer and wider ASBs.

Regarding metal machining processes, ASBs have been cited in high-speed machining in both continuous and discontinuous chips of steel, titanium and aluminum [28,29]; however, they are strongly connected to serrated or saw-tooth chip formation. In particular, Childs [30] revealed that increases in the cutting depth and cutting speed benefit the formation of serrated chips through ASB formation, which is further amplified by materials of low thermal diffusivity. Also, serrated chip formation has been attributed to micro-cracking inside the ASB, where micro-voiding initially occurs and expands through an ellipsoid shape, reacting to the coalescence mechanism and micro-cracking genesis [31]. In addition, orthogonal cutting tests revealed that high cutting speeds and depths, together with the low rake angle of the tool, facilitate adiabatic shearing, bringing serrated chip deformation in AISI 1045 steel [32] and transforming ASBs with a saw-tooth chip in the Ti-6Al-4V alloy [33] where a further increase in feed reacts to wider shear bands.

The ASB mechanism regarding penetrators and targets has also been studied in ballistic applications via several works. Specifically, ASB formation has been captured in both uranium [34] and tungsten [35,36] penetrators with 20 μm and 50$–$100 μm banding widths, respectively, while localized shear bands around the penetrator’s nose are manifested in the case of uranium [37], in contrast to the tungsten penetrator, whose ASBs result in mushrooming nose deformation [38]. Finally, the ASB mechanism has been conjugated into three deformation modes regarding target materials: plugging [39], discing [40] and spalling [41].

In the current work, an extended summary is developed on the microstructural characteristics and morphology of ASBs, the softening and damage mechanisms and ASB propagation through distinct stages and final fracture. Also, the material propensity to ASB formation is further evaluated via ASB susceptibility indexes, while the effects of mechanical and thermal properties on ASBs are discussed. Further, the effects of microstructural properties such as the type of initial microstructure, porosity, pre-existing twins and graining size on ASB susceptibility are analyzed. Further, aside from conventional metals and alloys, non-conventional metals like nanocrystalline and ultrafine-grained ones, bulk metallic glasses, metallic laminated composites and high-entropy alloys are also reviewed. Moreover, the propagation of ASB networks is discussed through their driving mechanisms and their connection to fracture. Also, the utilized testing methods and microstructural characterization techniques for ASB examination are reported, aiming to sum up their advantages and disadvantages and assess their suitability to specific challenges.

Finally, the current review aims to highlight critical characteristics and phenomena during ASB formation in metals and alloys in order to provide a clear understanding and fill possible gaps in the existing literature by underlining the ASB aspects that remain fuzzy. Specifically, the purpose of this work is to focus on attributing a more consecutive character to the ASB despite its rapid evolution, revealing its stage-by-stage propagation through discrete events and indicating its dominant microstructural mechanisms and their interactions. This work aims to highlight the need for a micro-/macroscopic coupled approach when analyzing ASBs, as microstructural softening mechanisms are strongly conjugated to macroscopic mechanical behavior. Also, several ASB aspects, such as the sequence of discrete stages during ASB evolution, the interaction between thermal and damage softening, the effect of temperature rise and its timing regarding ASB genesis, and the energy conversion between plastic work and internal heat are some issues which this review proposes to be reconsidered for more objective understanding.

2. Review Structure

The current work is organized into eight sections in total. First, a general introduction defines an ASB and the conditions under which it is developed, while the main dominant microstructural mechanisms are briefly discussed regarding their sequence during ASB evolution. Also, several applications in which an ASB has been recorded are quoted, and finally, the aim of the current work and the topic with which it deals are listed. The current section describes the structure of this review, briefly discussing the content of the following sections and the directions according to which this work was carried out.

The third section presents the microstructural characteristics of ASBs regarding their morphology and the dominant softening and evolution mechanisms. Specifically, the basic banding types are deeply discussed with respect to their characteristics, their differences and their contribution to ASB evolution, while finally, the ASB’s hardness and dominant softening mechanisms that lead to its propagation are analyzed.

The fourth section describes the material propensity to form an ASB, focusing on how different properties promote or prevent ASBs in order to indicate their effects and assess their responsibilities against ASBs. In particular, the effects of various mechanical and thermal material properties are discussed, as well as the effects of pre-existing twinning, initial microstructure, grain size and microstructural porosity on ASB promotion. In that way, this review aims to cover both the effects of macroscopic and microstructural properties and the characteristics of materials. Further, the conversion of plastic energy to internal heat energy is discussed, while finally, the occurrence of an ASB in various conventional metals and alloys, as well as in less conventional ones such as nanocrystalline and ultrafine-grained materials, bulk metallic glasses, metallic laminated composites and high-entropy alloys, is analyzed. The selection of the examined materials is based on the tendencies of the research literature. Last but not least, some basic metrics for ASB susceptibility are reported, aiming to quantify the material propensity against ASBs.

The fifth section discusses the connection between ASBs and fracture, highlighting their interaction and the transition from ASB to failure. Specifically, ASB propagation is analyzed at first by focusing on the widening mechanism and the developed network. Also, the transition to fracture is reported by discussing damage evolution and driving factors.

The sixth section consists of the microstructural observation and characterization techniques utilized, describing their applicability and the testing methods regarding experimental tests and manufacturing applications. However, the aim of the current work is not focused on reviewing the testing methods for ASB analysis. Instead, the sixth section aims to provide the reader with a better understanding of the tools utilized for ASB investigation and their suitability.

Finally, the seventh section contains the conclusions extracted from the current review work, and the eighth section describes some recommendations for future research regarding ASB aspects, which remain fuzzy according to the findings of this review.

3. Microstructural Characteristics of ASB Morphology and Mechanisms

3.1. Deformed and Transformed ASB

Plastic deformation under a high strain rate predisposes ASB formation due to thermomechanical strain instability, which occurs as the thermal softening effect on the plastic flow stress predominates against strain and strain rate hardening. The high magnitude of thermal softening required in order to reach deformation instability is attributed to a significant temperature increase inside the shear banding that is caused by the large amount of internal heat produced under intense adiabatic shearing deformation work along the ASB. A high strain rate loading is of significant importance for ASB genesis because it does not allow enough time for thermal diffusion storage. That way, the internal heat is produced inside the ASB. In fact, adiabatic shearing requires a critical strain rate, which provides a greater production rate than the diffusion one regarding internal heat [6,11,12]. Further, a large amount of deformation energy is required to provide high enough internal heat to react to a significant temperature rise by being trapped inside the ASB. Thus, large strain levels are also required for ASB formation in order to allow for high enough plastic work [42]. Therefore, ASB initiation presupposes a critical strain and critical strain rate for sufficiently large internal heat production under an adiabatic mode [13,43].

Figure 1 illustrates a strain–strain rate map for the TB2 titanium alloy, showing the distinct areas for stable deformation, ASB formation and cracking initiation regarding the strain/strain rate. In particular, stable plastic deformation dominates at lower strain and strain rates via slipping and twinning deformation mechanisms without forming any ASBs. As the strain and strain rates increase, unstable deformation occurs under the ASB, which is manifested via deformed banding at early stages, followed by transformed banding at higher strain and strain rates, finally leading to cracking failure. The stages through which an ASB is generated and evolves consist of severe shear localization along narrow shear-affected zones (SAZs) and deformed band formation, followed by transformed banding until the final cracking and fracture. During early deformation stages, homogenous and stable plastic deformation takes place under twinning and slipping, preceding ASB formation and providing heavily but homogeneously deformed large graining outside the ASB. Following this, as a deformed ASB forms under intense shear localization along the maximum shear direction, the grains become elongated along the ASB and even rotate, providing highly angled grain boundaries and a severely distorted microstructure inside the ASB [44]. Further, as deformation progresses at higher strains, transformed bands may develop, activating additional thermal softening and mechanical degradation mechanisms such as DRX [4,45,46,47], phase transformation [48,49,50], dynamic recovery [51,52] or even local melting [53].

In particular, deformed ASBs form at the early stages of adiabatic shearing at lower strain and strain rates compared to transformed ASBs. Deformed bands are manifested via severe shear localization along maximum shear directions, providing highly distorted and elongated graining along the ASB with possible fragmentation, without, however, reacting to phase change. Deformed ASBs have been observed in various non-ferrous metals and iron or carbon steels with coarse graining, as Figure 2 shows in the case of dynamically compressed AISI 4340 steel cylinders. On the other hand, higher strain and strain rates favor the formation of transformed ASBs, which are manifested via microstructural transformation through DRX or/and phase transformation mechanisms due to the significant temperature increase inside the ASBs. Transformed ASBs have been captured mainly in steels [3,12,54,55,56,57] and pure titanium and titanium alloys [10,58,59,60], but also in aluminum alloys [16], magnesium alloys [61,62,63] and nickel 718 superalloys [64], while both deformed and transformed bands have been noticed in U-Mo1.5 uranium alloys [34] and Al-Li alloys [16], among others.

Figure 3 depicts deformed and transformed ABSs for an Al-Li alloy in which critical strains of 0.14 and 0.17 are required for the formation of deformed and transformed banding, respectively. In fact, aluminum and nickel reveal high resistance against transformed ASB formation due to high thermal diffusivity and high strength, respectively. More specifically, the high thermal diffusion of aluminum does not allow for a significant temperature rise as the internal heat is quickly released, preventing adiabatic deformation, while in the case of the nickel 718 superalloy, its high enough strength makes it difficult to reach sufficient strain levels in order to reach instability.

On the contrary, high Co-Ni M54 steel reveals both banding types under a high strain rate compression, initially showing the formation of a deformed ASB, which then turns into a transformed one [65]. As Figure 4 depicts, the ASB is formulated at a 23% strain level along the diagonal direction of maximum shear, while significant shear failure occurs inside the white-etched ASB under cracking initiation. Further, as deformation progresses, secondary thinner ASBs form and cracks propagate alongside the primary ASBs. Also, a more detailed view at a 23% strain, depicted in Figure 5, shows that the central region of the ASB is subject to a transformed banding type compared to the ASB’s edges, which form as deformed bands, while the transformed band consists of fine equiaxed nanograins due to rotational dynamic recrystallization (RDR). A scanning electron microscopy (SEM) image reveals that the ASB core consists of transformed banding inside of which cracks propagate, while the ASB transition region shows a deformed type, indicating a greater width of 5.2 μm compared to a 2.1 μm wide transformed band. Finally, a microstructural composition analysis showed an increased carbon concentration within both deformed and transformed bands compared to the initial composition of M54 steel.

Moreover, during transformed ASB formation, possible DRX occurrence reacts to ultrafine equiaxed graining, which is severely elongated along the ASB, forming low-angled (2°–15°) misoriented grain boundaries [59,66], while possible phase transformation provides a martensitic microstructure inside the transformed ASB of steels. Both ultrafine elongated graining and possible martensite have been considered responsible for the white or slightly yellow color of the transformed bands, naming them white bands [67]. Transmission electron microscopy (TEM) and X-ray observations have also indicated martensite [57,66] and fine equiaxed delta-ferrite and martensite with carbide [68] microstructures inside the transformed ASBs in steels, while SEM and optical microscopy are not accurate methods for identifying if the white color of the transformed bands is attributed to the martensitic microstructure or highly elongated ultrafine grains. On the other hand, SEM and optical microscopy can sufficiently capture the presence of cracking propagation along the softened microstructure inside the transformed ASB [69], as depicted in Figure 6, while Figure 7a illustrates the sequence of deformed and transformed ASB formation with final cracking inside the transformed band at different stages of dynamically compressed armox 500T steel with inverse pole figure (IPF) maps and graining size distributions inside the ASB, and Figure 7b highlights the ultrafine graining generation in transformed banding due to DRX [70].

In summary, Scheme 2 contains the fundamental microstructural characteristics and dominant mechanisms of deformed and transformed bands as a comparative metric between the two main ASB types. Therefore, a deformed ASB is characterized by severe shear strain localization and dominates during ASB initiation, while the temperature rise is responsible for the transition to the transformed ASB, which is dominated by intense thermal softening, which activates dynamic recrystallization and phase transformation. Finally, a deformed band consists of highly distorted and elongated graining with possible fragmentation, while ultrafine equiaxed graining of low-angled misorientation boundaries dominates within the core of the transformed ASB.

3.2. ASB Hardness

Both martensite and ultrafine DRX graining have been considered responsible for the significantly high hardness of the microstructure inside transformed ABSs. More specifically, the presence of martensite inside the transformed bands is attributed to the rapid cooling of ASB compared to the surrounding microstructure, as the significant temperature increase via which transformed ASBs form is followed by a rapid quenching of a cooling rate of about 10⁷ K/s [28,71,72], resulting in significantly high hardness in the transformed band due to quenched martensite.

Also, in the case of DRX nanograin nucleation, highly hardened transformed banding is provided according to the Hall–Petch law for the grain sizing effect on yield stress. In fact, a commonly accepted aspect regarding the high hardness of transformed bands considers ultrafine graining generation to be the responsible mechanism, given that phase transformation does not always occur inside white bands [62,73]. Figure 8 illustrates the hardness distribution transversely to the ASB in the case of dynamically compressed 500T steel at a 28–30–33% strain magnitude, indicating that deformed bands form at earlier stages (28% strain) and transformed bands form at higher strains (30–33%), while deformed bands are revealed to be wider than transformed ones, with the heat-affected zone expanding transversely along the deformed bandwidth. Also, the hardness inside the ASB is significantly higher than the surrounding structure, while the further transformed bands show much higher hardness than the deformed ones due to ultrafine graining.

However, in contrast to conventional metals, ASBs in bulk metallic glasses (BMGs) have revealed softened interior microstructures showing lower hardness compared to the surrounding matrix [74]. The latter is attributed to both thermal softening [75] and free-volume coalescence softening [76], as the molecular dynamic simulations indicated by examining the dislocation activity, highlighting the presence of nonaffine displacements that precede amorphization [77,78].

Therefore, transformed ASBs have revealed significantly high hardness, which has been attributed to rapid cooling, which provides a quenched martensitic structure; phase transformation, which provides harder phases; and finally, ultrafine graining, which hardens the microstructure inside the ASB core.

3.3. Phase Transformation

A phase transformation mechanism during ASB formation has been indicated in steels [79], titanium alloys [80] and Fe-Cr-Ni monocrystals [81] and has been attributed to either the thermodynamics and kinetics of phase stability when a solvus temperature is reached or to a rapid cooling rate, which can reach even 10⁷ K/s [28,71,72]. Due to the rapid cooling of a massively high-temperature ASB compared to the surrounding microstructure, quenched martensite is generated inside the transformed band, while heavily deformed large grains exist outside the ASB, where twins dominate. Also, both DRX and phase transformation have been observed in the two-phase α-β Ti alloy (Ti-3Al-5Mo-4.5V) during transformed ASB formation, showing that the α-phase recrystallized into ultrafine graining at a temperature of about 1100 K, and the β-phase transformed to an α’’ structure [82]. Furthermore, another example of a parallel existence between grain refinement and phase transformation has been reported for a dual-phase α-β Ti alloy (Ti-5.5Mo-7.2Al-4.5Zr-2.6Sn-2.1Cr), whereas the ASB transition region revealed clearer grain boundaries in the α/β interface compared to the fuzzy microstructure of the ASB core [83].

Moreover, a metastable β Ti-6Mo-3.5Cr-1Zr alloy tends to develop a stress-induced ω-phase during high strain rate compression, while grain refinement, together with ω→α→β phase transformations at 500 °C and 700 °C, respectively, dominate during ASB evolution [84]. However, the stress-induced ω-phase causes a strain hardening increment, delaying ASB initiation. Also, the Ti1023 alloy revealed deformation-induced phase transformations during ASB, where β→α and martensitic transformations take place [85]. In addition, phase transformation was captured in local areas of sparse β-stabilizing elements (Mo and Cr), which were refined to smaller α-grains, while elsewhere, β-stabilizing elements recrystallized to finer β-grains under the DRX mechanism. Finally, an α→β phase transformation at 1000 °C has been captured during ASB evolution in the segmented chip of Ti-6Al-4V due to a temperature increase, while a softer β-phase accelerated the shearing instability [86].

Therefore, phase transformation reflects one fundamental mechanism of a transformed ASB, activated by either a temperature rise or rapid cooling. In the first case, a phase change takes place when the temperature overcomes the solvus temperature, while in the case of rapid cooling, a phase change happens, providing a quenched martensitic structure. Finally, a phase transformation has been reported together with DRX in many titanium alloys.

3.4. Dynamic Recrystallization

The DRX mechanism has been reported during transformed ASB formation in steels [79,87] and titanium [4,88], indicating that DRX nanograins are generated inside ASBs, while RDR has also been mentioned, resulting in the transformation of elongated sub-grains to ultrafine equiaxed grains via a 30° rotation of grain boundaries [79]. In addition, the ASB neighboring microstructure, in the case of a dynamically compressed AZ31 magnesium alloy, has been separated into three regions regarding grain crystal orientation [89]. Specifically, the ASB core has shown a random crystal orientation composed of ultrafine equiaxed DRX grains, the ASB transition zone has revealed a textured crystal orientation and, finally, the surrounding matrix also showed a random orientation, as Figure 9 illustrates through a grain orientation diagram. Further, DRX inside the ASB of an AZ31 alloy is driven by twin-induced DRX, continuous DRX and discontinuous DRX. The very dense twins control the first one, developed under severe shear deformation in an AZ31 alloy, while continuous DRX starts to nucleate finer grains within the initial grains, appearing mostly in alloys with high stacking fault energy or in Mg alloys during hot working. In contrast, discontinuous DRX nucleates the recrystallized grains along the initial grain boundaries and is responsible for local microstructural softening, providing a zigzag-like tendency in the stress–strain curve. Figure 10 depicts that the ASB core of the AZ31 alloy consists mainly of a recrystallized and deformed structure due to severe shear strain and a significant temperature rise, while the transition zone and surrounding matrix contain a substructure instead of a recrystallized one [89]. Also, the rising temperature during DRX is considered responsible for grain boundary migration and coalescence, revealing equiaxed nanograins of low-density imperfections [90]. DRX nanograins of 10–500 nm have been observed in 304 stainless steel and titanium alloys, while 50–100 μm graining occurs outside transformed ASBs.

DRX has been described as a mechanical degradation mechanism of a microstructure causing material softening and facilitating failure, while DRX has also been reported to precede thermal softening and, consequently, ASB failure [47]. Further, Landau et al. [4] reported regions of high density in DRX nanograins around crack tips, which facilitate cracking propagation along an ASB by softening the microstructure, while instead, sparse DRX regions were observed in the surrounding microstructure away from the ASB. DRX softening behavior has also been attributed to the decrease in dislocation density, which facilitates plastic deformation progress, resulting in strain instability [46]. Long et al. [91] reported that DRX results in strain softening, speeding up ASB evolution, while further TEM images of Ti-6Al-4V indicate local DRX as the responsible mechanism for micro-ASB formation, as depicted in Figure 11. Finally, a two-step DRX mechanism has been proposed during micro-ASB formation, showing an initial grain refinement in the ASB transition region and a following refinement within the ASB core [91]. In particular, the severe strain localization along the banding concentrates high-density dislocations (Figure 12a), which activate discontinuous DRX under nucleation at grain boundaries (Figure 12b). Thus, a DRX zone is progressively formulated, driven by the strain-softening of the new dislocation-free recrystallized grains (Figure 12c). Following this, during the second step of the DRX mechanism, a strain increase reacts to grain elongation inside the ASB core (Figure 12d), while some of these grains are divided into smaller equiaxed nanograins, subjecting them to a rotation mechanism (Figure 12e). Finally, several micro-ASBs are generated inside the DRX core and then expand and multiply as the strain increases (Figure 12f). Therefore, the proposed two-step DRX mechanism reveals an initial recrystallization level at the transition zone and a following second DRX step inside the core where micro-ASBs form and propagate.

To summarize, DRX provides ultrafine equiaxed graining inside the ASB core of random crystal orientation in contrast to the ASB transition region, which is dominated by textured crystal orientation. Also, three types of DRX have been reported: twin-induced, continuous and discontinuous DRX. Moreover, micro-ASBs have been considered to form under a two-step DRX mechanism, which causes grain refinement initially in the transition zone and next inside the ASB core. Finally, DRX has been found to precede thermal softening, while dense DRX regions around crack tips have indicated DRX as one of the main mechanical softening mechanisms that facilitate cracking propagation.

4. Materials

4.1. Material Properties

During plastic deformation, high strain and strain rate are required in order to initiate ASB formation by causing deformation instability. In fact, ASB genesis presupposes a large enough plastic deformation work, which is converted to internal heat, resulting in a significant rise in temperature inside the ASB. That high enough temperature increase requires adiabatic conditions, which are achieved through a high strain rate under which the heat production rate is greater than its diffusion rate. Following this, the significant temperature increase reacts to mechanical material degradation via thermal softening or other mechanisms such as DRX, RDR or phase transformation, facilitating ASB evolution and leading, finally, to fracture.

Table 1. ASB width and graining size inside ASB for different alloys.

Alloy	ASB Width (μm)	Grain Size Inside ASB (nm)	Shear Strain (–)	Shear Strain Rate (s−1)	Reference
Pure α-Ti	45	100$–$500	2.5	3.2 × 104	[92]
Ti-6Al-4V	1$–$8	200	0.13$–$0.92	104	[93]
Ti-5Al-5Mo-5V-1Cr-1Fe	20	100	2	105	[94]
AISI 4340 steel	9	20	0.5	1.8 × 104	[12]
AISI 304 stainless steel	5$–$50	100$–$200	1$–$100	104	[79]
Zirconium Z702	10$–$25	200	25$–$100	104	[95]
8090 Al-Li	10$–$30	200	1.2$–$1.4	103	[16]
CrMnFeCoNi	10	100$–$300	7	1.6 × 103	[96]

summarizes various observations of ASB width with interior recrystallized grain sizing for different alloys together with the shear strain/strain rate in which observations took place.

In this aspect, ASB formation reflects a thermomechanical instability phenomenon, and thus, the propensity of materials against ASB depends on its mechanical and thermal properties, while loading and geometry affect the ASB mechanism through their effects on the facilitation of shear localization. Considering the thermomechanical behavior of the ASB mechanism, thermo-viscoplastic flow rules are necessary for describing material behavior, suggesting that plastic flow stress is a function of strain hardening, strain rate hardening and thermal softening. Further, given that ASB instability is caused by the superiority of thermal softening against strain/strain rate hardening, material sensitivity against ASB formation depends strongly on its strain/strain rate hardening and thermal softening coefficients. Therefore, materials of lower strain/strain rate hardening and higher thermal softening tend to favor ASB formation.

Table 2. Critical strain for ASB and material hardening/softening parameters for different alloys (data from [97]).

Alloy	n	m	∂τ/∂θ (MPa/K)	γcr
1043 steel	0.12	0.01	0.75	0.61
1006 steel	0.24	0.01	0.625	1.45
AISI 4340 steel	0.043	0.01	3.1	0.053
AISI 304 stainless steel	0.5	0.02	1.975	0.8
6061-0 aluminum	0.24	0.002	0.5	1.33
Pure titanium	0.3	0.025	1.25	0.6
Ti-6Al-4V	0.02	0.015	2.3	0.022

indicates the effect of strain hardening exponent n, strain rate hardening exponent m and thermal softening parameter ∂τ/∂θ on critical shear strain γ_cr required for ASB formation for various alloys. In more detail, strain hardening parameter n expresses the strain exponent, and strain rate hardening parameter m expresses the exponent of strain rate on the constitutive equation of plastic flow stress, both reflecting the material strengthening due to strain and strain rate increments, respectively. On the other hand, the thermal softening parameter ∂τ/∂θ expresses the derivative of plastic flow shear stress τ to the temperature θ, reflecting the effect of temperature increment on the reduction of the plastic flow stress. Critical shear strain reflects a reliable index for assessing material susceptibility against ASB as lower γ_cr indicates a higher sensitivity, facilitating the occurrence of ASB by reacting to earlier shear instability. As Table 2 shows, weak strain hardening and stronger thermal softening reduce γ_cr, increasing material sensitivity against ASB.

Regarding the effect of material thermal behavior on ASB, thermal properties that enhance adiabatic conditions facilitate ASB. In particular, lower thermal diffusivity increases the material propensity to ASB genesis as it decreases the heat diffusion rate, providing a high temperature rise inside the ASB and facilitating strain instability. In addition, lower thermal conductivity reacts similarly to decreasing heat diffusion rate, while further lower specific heat and density result in lower critical strain according to the Culver instability criterion [98]. Also, lower specific heat reacts to higher temperature increases for certain plastic deformation work. However, lower thermal conductivity makes ASB detection more difficult by decreasing its width, while a higher material strength brings a similar effect [9]. Therefore, aluminum alloys do not often enhance ASB due to their high thermal diffusivity, and nickel-based alloys delay ASB occurrence due to their high strength.

Finally, the importance of strength in ASB initiation has not yet been clarified clearly, as in the case of extremely large strength, γ_cr increases enough to provide higher resistance against ASB, such as in the case of Inconel 718 [99], while in contrast, when high strength is attributed to rapid cooling, as in the case of quenched martensitic steel, it enhances severe shear localization through reduced ductility, which decreases γ_cr, favoring ASB initiation [100]. Also, material sensitivity to ASB can be reduced by implementing alloying elements into pure metals, which prevents dislocation slipping and increases necessary flow stress, which reacts to higher γ_cr [101]. Further, cold work hardening enhances strain hardening behavior, rising γ_cr and reducing the material propensity to ASB, while processes that enhance homogenous plastic deformation also prevent ASB. Thus, heat treatment techniques such as annealing, tempering or normalizing reduce the material's propensity to ASB by homogenizing its microstructure. In fact, a heat treatment of 350 °C–1 h in the case of an AE21 magnesium alloy provided homogenous plastic deformation during the rolling process without revealing ASB formation, which was detected before the heat treatment [25].

Therefore, the material propensity to form ASB is mainly affected by mechanical and thermal properties, as strong strain hardening prevents ASB, while high enough thermal softening facilitates strain instability and ASB. Also, materials of low thermal diffusivity have shown high sensitivity to ASB, as they promote adiabatic deformation conditions. Finally, heat treatment solution and work hardening both enhance material resistance to ASB.

4.2. Taylor–Quinney Coefficient

The Taylor–Quinney (TQ) coefficient also plays an important role in ASB formation by expressing the fraction of plastic strain energy converted into internal heat. Thus, considering the plastic strain energy ${\int }^{ }d{W}_p$, material density ρ, specific heat c_p and temperature increase ΔΤ, the TQ coefficient is defined as follows:

$$TQ=\frac{\rho ·c_p·\Delta {\rm T}}{{\int }^{ }d{W}_p}$$

(1)

Assuming adiabatic conditions, a high enough TQ parameter increases the material susceptibility against ASB. Despite most works assuming a constant TQ value of 0.9, Rittel et al. [102] showed that TQ depends mainly on the material and secondary on loading mode in some cases, revealing that a constant value of 0.9 is not widely accepted for all materials. In particular, various materials such as Ti-6Al-4V, commercial pure titanium (CP Ti), aluminum alloys like AA5086 and AA2024, 304L stainless steel, low carbon 1020 steel and C300 maraging steel were examined through compression (C), tension (T) and shear (S) tests under the strain rate and maximum strain which are listed in

Table 3. Maximum strain and strain rate of compression–tension–shear tests of various materials for TQ quantification [102].

Material		Compression	Tension	Shear
Ti6Al4V	Strain rate (s−1)	2000	1500–3400	2800
	Maximum strain	0.25	0.04	0.37
CP Ti	Strain rate (s−1)	2000$–$3000	2000–2500	3300–5000
	Maximum strain	0.22	0.07	0.36
Al 5086	Strain rate (s−1)	2200$–$2500	2000–3000	5000
	Maximum strain	0.24	0.1	0.63
Al 2024	Strain rate (s−1)	2500–3000	1250–1600	5000
	Maximum strain	0.2	0.08	0.31
304L	Strain rate (s−1)	2000–2800	1750–2000	7000
	Maximum strain	0.3	0.16	0.4
1020	Strain rate (s−1)	1300–1500	1400–2000	4000–6000
	Maximum strain	0.18	0.04	0.61
C300	Strain rate (s−1)	1200–1800	1000–2500	1500–2500
	Maximum strain	0.09	0.02	0.11

. Figure 13 depicts the results regarding the average TQ coefficient, showing that the TQ value of pure Ti reveals a strong dependence on loading mode as TQ lies in about the 0.7–0.9 range in compression and shear tests, in contrast to tension, where it varies in the range of 0.44–0.65. The reason for this can be attributed to a twinning mechanism, which is strongly related to heat production and reveals a stronger magnitude during shear and compression instead of tension [102]. The other examined materials do not seem to be affected by the loading mode, with Ti-6Al-4V reaching a TQ value of 0.3–0.5, aluminum alloys lying in the range of about 0.2–0.4 and 1020 steel within the range of 0.8–0.92, while the 304L and C300 steels revealed a TQ value in the range of 0.4–0.7. In fact, the lower TQ of aluminum alloys can be partially attributed to their high thermal diffusivity.

Therefore, TQ reveals a strong dependence mainly on material, but also secondary on loading mode like in the case of pure Ti, while some materials showed a slight dependence on the strain level, like 304L steel, whose TQ increased with strain, in contrast to Ti-6Al-4V, whose TQ decreased regarding the shear tests [102]. Thus, the assumption of a constant TQ value of 0.9 should be abandoned and considered separately for each material and possibly the loading case, while finally, when its strain dependence is important, it could provide a highly varied ASB propagating rate.

4.3. Twinning Effect

Also, twin-induced ASBs have been reported in a hot-rolled Mg-1.1Zn-0.76Y-0.56Zr alloy [103] and a Mg-Al-Mn alloy subjected to a ballistic impact [104], indicating twin boundaries as favorable positions for shear localization. More specifically, the interaction of closely neighboring twins under a high strain rate results in shear deformation localization along twin boundaries instead of widening of the twins, leading to ASB formation.

Moreover, the presence of pre-existing twins in the initial microstructure affects the material propensity against ASB formation. In particular, pre-existing twins favor ASB occurrence, but they tend to decrease the degree of DRX expansion [105]. Figure 14 depicts various IPFs of the ASB neighboring structure in the case of the AM30 magnesium alloy for different volume fractions of pre-existing twins, illustrating that the increase in pre-existing twins tends to facilitate ASB formation and cracking failure, while the ASB transition zone is shrunken. Furthermore, strain increase can lead to additional twinning development, which, however, is localized mainly in the surrounding matrix rather than in the ASB core [105]. In more detail, Figure 15 shows that the fraction of formulated twins decreases as the fraction of pre-existing twins becomes higher, while the surrounding matrix contains almost 17% twins in contrast to the ASB core, whose twinning composition is about 1%. Finally, the increase in pre-existing twins reduces the extent of the DRX region, which is mainly localized into the ASB core and secondary into the transition zone, while instead, the surrounding matrix consists of deformed and sub-structured grains, as Figure 16 depicts.

Therefore, twinning plays an important role in ASB susceptibility as twin boundaries promote ASB initiation, while pre-existing twins also facilitate ASB formation. Finally, deformed twins tend to form mainly on the surrounding matrix, while the ASB core remains poor in twinning content.

4.4. Microstructural Porosity Effect

In addition, increased microstructural porosity promotes ASB development by facilitating micro-voiding nucleation and speeding up ASB evolution [106]. Thus, processes that reduce microstructural porosity, like hot isostatic pressuring, decrease material propensity to ASB formation. The effect of microstructural anisotropy induced by a highly deviated porous distribution has also been studied regarding its influence on strain localization and shearing failure propensity [107].

Further, Vishnu et al. [108] examined various additive manufactured metals like AlSi10Mg aluminum alloy, 316L stainless steel, Ti-6Al-4V titanium alloy and Inconel 718 Ni-alloy in order to study the effect of microstructural porosity on ASB sensitivity. The examined specimens consisted of thick-walled cylinders subjected to dynamic torsion, while the porosity was approached by implementing internal voids until 2% volume fraction and 110 μm size were reached with multiple spatial and size distributions. The results indicate that both the size and spatial distribution of voids affect ASB position and propagation, as the larger pores trigger ASB initiation, while ASB expands through the directions of higher voiding density, as depicted in Figure 17 for Ti-6Al-4V. Thus, increased porosity enhances material sensitivity against ASB formation as large pores indicate ASB location, while increased voiding volume fraction reduces critical shear strain, driving ASB propagation.

Figure 18a depicts that for a certain voiding volume fraction, a high mean voiding size results in a porous microstructure of sparse voids where a primary ASB forms around the largest pore, while in contrast, a lower mean voiding size reacts to more homogenous porosity of a similar voiding size, favoring the formation of multiple secondary ASBs. Figure 18b ensures a similar effect regarding the standard deviation of voiding size distribution, revealing that higher deviation favors ASB initiation around larger pores, while lower deviation reacts to multiple shear localization zones.

Therefore, microstructural porosity significantly affects ASB formation, as large pores trigger ASB initiation and spatial porosity distribution affects the ASB propagating path. Also, low porosity deviation provides multiple but weaker secondary ASBs in contrast to a higher deviated porosity, which promotes the formation of one primary ASB.

4.5. Effect of Initial Microstructure

Moreover, the effect of the initial microstructure on ASB susceptibility has been investigated in the case of titanium alloys. Specifically, Qin et al. [109] studied the adiabatic shearing behavior of a Ti-5553 alloy subjected to dynamic compression, examining the effect of lamellar and bimodal microstructures. The lamellar microstructure allowed for ASB formation even at low strain levels, revealing DRX β-nanograins within the ASB core, while the chemical redistribution of alloying elements was captured inside the ASB. In contrast, the bimodal initial microstructure seemed to prevent ASB formation, showing increased resistance, while no chemical redistribution of alloying elements was observed. Also, Chen et al. [110] investigated the effect of the initial microstructure on the ASB propensity of a dynamically compressed Ti-6321 alloy, examining an equiaxed microstructure (EM), duplex microstructure (DM) and Widmanstätten microstructure (WM). DM showed the greatest resistance against ASB, revealing the highest critical shear strain rate of 3000 s⁻¹, and EM provided a 2000 s⁻¹ critical shear strain rate, while WM revealed the lowest one of 1000 s⁻¹, promoting more ASB formation. Further, Lei et al. [111] recently studied how rim and web locations in a steel railway wheel favor shear banding evolution, depending on their initial microstructure, which was produced via different heat treatments. For this reason, hat-shaped specimens were tested in a split Hopkinson pressure bar. The results indicated that pre-eutectoid ferrite was initially deformed while following ASB evolution and was driven by structural softening, which is strongly conjugated to the spacing between pearlitic lamellae. In particular, the narrower pearlitic lamellae spacing of the rim specimen was found to promote a greater strain gradient, facilitating ASB formation. Thus, the rim revealed higher ASB propensity under the same forced shear test compared to the web specimen.

Therefore, duplex and bimodal types of initial microstructure showed greater resistance against ASB, in contrast to the WM lamellar one, which showed higher sensitivity. Finally, narrower lamellae spacing seems to promote more ASB evolution in the case of a pearlitic structure in steels, providing a higher strain gradient and facilitating shear localization.

4.6. Grain Sizing Effect

Moreover, the effect of the initial grain sizing on the microstructure has also been studied for an AZ31 magnesium alloy, indicating that material susceptibility against ASB increases as the grain sizing reduces and the strain rate increases [112], revealing wider ASBs with grain refinement along the intersection between the ASB and surrounding matrix, and a larger graining with twins outside the ASB, as shown in Figure 19. More specifically, coarse initial grain sizing reveals lower resistance against plastic deformation according to the Hall–Petch relation and may be subject to early cracking failure before ASB occurrence. In contrast, fine graining increases resistance against plastic deformation, leading to higher required strain/strain rate levels that enhance ASB formation containing the precursor of thermal softening, whose great magnitude is responsible for adiabatic shear instability. In fact, adiabatic temperature increases due to internal heat from plastic work, which reacts initially to grain refinement inside, which ASB forms and widens. Similarly, friction stir processed (FSP) AA6061 aluminum alloy is subject to increased ASB sensitivity compared to commercial AA6061 because the greater stored energy into the fragmented grains of FSP AA6061 favors ASB initiation in contrast to commercial alloy, whose coarse graining provides higher resistance against ASB formation [113]. Further, multi-pass FSP AA6061 facilitates ASB expansion through continuous DRX, which softens the microstructure, and together with abnormal grain growth, reacts to reduced flow stress, enhancing strain instability.

4.7. Crystal Structure

Regarding materials containing hexagonal close packing (HCP) crystal structures such as titanium, magnesium and zirconium alloys, HCP structures reveal greater resistance against plastic deformation, leading to a higher required strain/strain rate for shear localization appearance [101]. More specifically, ASB observations in α-Ti [114] and Ti-6Al-4V [115] titanium alloys revealed ASB width of up to 20 μm, with thermal softening activating dynamic recovery and DRX in α-Ti alloy and twinning-induced RDR in Ti-6Al-4V alloy, while both presented equiaxed recrystallized nanograins. Further, AM60B magnesium alloy has been found to develop both deformed and transformed banding with a significant elongation of equiaxed DRX grains inside ASB without, however, any phase transformation [71], while ZK60 alloy also revealed fine equiaxed grains with low dislocation density inside a 9 μm wide ASB with a severely deformed and stretched surrounding matrix along the maximum shear direction [64]. Also, dynamic deformation reacts to 200 nm fine graining inside zirconium ASBs, while a 5 μm wide transition region of equiaxed graining is captured in the intersection between the ASB and matrix [95]. Finally, laminated metallic composites with soft HCP and harder neighboring constituents can benefit from moderate shear banding deformation, showing increased ductility due to shear banding-induced pyramidal $\langle \mathrm{c}+\mathsf{\alpha }\rangle$ slip systems, which are activated instead of basal ones [116].

Regarding materials with a face-centered cubic (FCC) structure such as aluminum, nickel-based alloys and copper, dislocation slipping has been reported as the dominant mechanism in the plastic deformation of aluminum alloys, while equiaxed recrystallized grains develop with ASB evolution under crystal rotation and randomly distributed grain misorientation inside ASBs [117]. However, high thermal diffusivity in many aluminum and copper alloys does not enhance adiabatic shearing, significantly increasing γ_cr, and so extremely high strain rates are required for ASB formation, as in the case of shocked copper [118]. Moreover, low nickel austenitic stainless steel has been found to develop RDR equiaxed elongated nanograins inside ASB with low dislocation density [119].

Regarding base-centered cubic (BCC) structure alloys, AISI 4340 high-strength steel has revealed an initially deformed ASB formation evolving to transform ASB with heavily deformed pre-existing martensite and without any phase transformation [120]. Further, RDR has been indicated as the responsible mechanism for graining refinement inside ASBs via 30° grain boundary rotation [121]. Also, Ti-5Al-5Mo-5V-1Cr-1Fe alloy (Ti-55511) develops RDR, dynamic recovery and phase transformation mechanisms during ASB evolution, providing α’’ martensite and orthorhombic β-Ti phases inside ASB [94], while β-phase transforms to martensite inside the ASB of Ti-5553 alloy [122]. BCC single β-phase Ti-alloy also shows phase transformation due to temperature rise during ASB evolution as β-phase is initially transformed to α’’ martensite and follows to HCP α-phase [123]. Finally, BCC pure tantalum hat-shaped specimens have been found to form ASBs under dynamic compression at γ_cr equal to 4, with ASB core consisting of DRX ultrafine graining whose slip and rotation provided a strong texture [124].

Therefore, pure aluminum and copper have shown a reduced tendency to ASB formation, mainly due to higher thermal diffusivity, in contrast to steel, titanium and magnesium alloys. In particular, steels seem to promote both deformed and transformed ASB types with grain refinement and phase transformation because of quenching having been captured. Also, magnesium alloys have been proven to suffer from ASB, with DRX representing the main softening mechanism, while titanium alloys also promote ASB dominated by DRX, phase transformation, twin-induced RDR and dynamic recovery phenomena.

4.8. Nanocrystalline and Ultrafine-Grained Materials

In addition, nanocrystalline (NC) metals and ultrafine-grained (UFG) materials have been investigated by various recent studies regarding their susceptibility to ASB formation. More specifically, NC materials have shown a greater propensity to ASB failure compared to conventional metals due to their lower stacking fault energy despite the fact that they revealed sufficient work hardening [125]. Further, multilayered metallic composites have shown vulnerability against shear instability, which, however, can be improved via a gradient layer thickness distribution in the case of Cu/Zr nanolayered composite [126]. In particular, more and thinner nanolayers revealed higher resistance against ASB by increasing dislocation density and preventing strain localization. UFG materials also reveal high susceptibility against ASB formation due to their decreased strain hardening ability, which reduces critical shear strain for ASB initiation [91]. Also, studies regarding UFG copper [118] and UFG titanium [127] have indicated their greater thermal softening magnitude as the key factor for their greater sensitivity to ASB formation as strain instability is enhanced. Moreover, ASBs have been reported to develop at NC and UGF materials even at low strain rates and quasi-static loading, while DRX can also occur at low strain and temperature due to their high density in grain boundaries [91].

4.9. Bulk Metallic Glasses

Regarding BMGs, their low strain hardening magnitude increases their susceptibility to shear banding (SB) even at lower strain rates under quasi-static loading [7]. Thermal softening [75] and free-volume coalescence softening [76] have been considered as the two key mechanisms during SB formation in BMGs. Jiang and Dai [128] have underlined the distinct contribution of free-volume softening and thermal softening to SB evolution separately. Specifically, free-volume softening was considered the triggering mechanism for SB generation by facilitating shear localization, the intense magnitude of which leads to SB genesis. Also, free-volume softening was found to precede the significant temperature increase, indicating that thermal softening must be considered more as a consequence of strain localization in contrast to free-volume, which reflects the real cause. Therefore, despite the exhibition of coupled free-volume and thermal softening, shear banding instability seems to be dominated by free-volume. Similarly, Dai and Bai [129] have confirmed that SB width in BMGs is controlled by free volume instead of thermal softening. Finally, V- and T-instability criteria, together with SB width formulas, were derived regarding free-volume and thermal softening effects, respectively, concluding that free-volume drives SB evolution, with thermal softening simply accelerating instability progress. In addition, Greer et al. [130] also reported that shear-induced structural disordering and free-volume generation drive SB evolution by facilitating shear localization, while internal heating follows later, reacting to catastrophic failure.

Shear bands are manifested via several shear transformation zones (STZs) where nonaffine displacements and local heterogeneities have been captured by implementing molecular dynamic simulations [77,131,132,133], which further indicates that dislocation activity precedes amorphization during SB evolution [131]. The increased shear strain rate within the SB and the reduced one in the surrounding matrix enhance strain softening and SB formation [128]. Moreover, strain gradient induced by structural heterogeneity also affects SB propensity. In fact, a higher strain gradient provides severe strain localization, activating STZs and leading to SB formation [134]. Instead, a lower strain gradient favors wider and more homogenous plastic deformation. Further, Yang et al. [135] decoupled displacement and strain gradient tensors into three distinct parts attributed to shear, dilatation and rotation, respectively. Thus, shear-dominated zones (SDZs), dilatation-dominated zones (DDZs) and rotation-dominated zones (RDZs) can be distinguished during the progress of deformation. At low strain, all three types remain strongly conjugated and act synchronously, while as deformation progresses, DDZs are found to predominate, reacting to SDZs and RDZs in softer regions, while secondary RDZs are activated in harder neighboring regions. Therefore, dilatation softening seems to be the main cause of shear localization, preceding the other two types at early stages, while as the strain increases, the effects of dilatation, shear and rotation on SB generation start to act separately.

Also, SB evolution in BMGs depends on loading type, as compressive or tensile stress states activate different propagation mechanisms [7]. In more detail, SB is initially expanded along a thermally softened region (TSR) of a₁ length value, as Figure 20 illustrates in the case of compressive loading. In the following, initial TSR cools down, revealing a “frozen band”, while SB propagates along shorter TSR of a₂ length value. That sequence of frozen bands and reduced TSR length continue until TSR reaches a critical length where SB expansion stops, highlighting that SB propagation velocity decreases during its evolution. In contrast, Figure 21 indicates that SB initially forms due to STZ expansion and thermal softening in the case of tensile loading. Next, cracking is generated and expands along SB following its propagating velocity. For this reason, the compressive stress–strain curve reveals a jagged behavior when SB forms, showing local peaks and lows in contrast to the tensile curve where SB propagation is more uniform [136], as Figure 22 depicts. Therefore, SBs can be detected either as hot or cold due to stop-and-go cycles caused by sticking and slipping, which freeze and reactivate the primary SB, respectively [130]. Sample dimensions and loading mode have shown to have important effects on the transition from cold to hot SBs in BMGs. In fact, the SB width and shear sliding velocity were found to affect the effective local viscosity of the sticking–slipping behavior of primary SB, while severe structural disordering and free volume were reported in cold SBs, leading to nano-crystallization [130]. The last one is attributed to atomic mobility instead of internal heating, reacting to more diffuse SBs and promoting their multiplication, which can enforce stable plasticity against failure [130]. Finally, a more heterogeneous SB propagating network has been reported in BMGs subjected to high-pressure torsion without, however, affecting the spacing between SBs [137], while the SB mechanism in BMGs under cyclic loading has also been studied regarding its connection to fatigue behavior [138].

Moreover, BMGs have concentrated significant interest regarding SB evolution by implementing novel modeling approaches. In particular, a new modeling technique based on distributed dislocations proposed by Li et al. [139] suggests that shear concentrators like notches or voids act as triggering mechanisms for SB initiation. Further, distributed dislocations reveal possible SB propagation paths, indicating that an SB will expand along the direction in which it will reach its maximum possible length, allowing for the sufficient prediction of the SB propagating network. In addition, the SB evolution path has been reported to be strongly conjugated to the primary shear stress profile, as the maximum shear stress reflects the driving force for SB expansion [140]. Last but not least, the vortex field around STZs in BMGs has also been investigated via molecular dynamics [141]. The results indicate that a rotation field is provided from the quadrupolar strain distribution around an STZ, revealing the propagating paths of SBs, their interaction and their multiplication.

In conclusion, BMGs have shown increased susceptibility to SB formation due to their low strain hardening. SB evolution is driven by free-volume softening, which precedes the rise in temperature. Instead, thermal softening accelerates catastrophic failure, reflecting a consequence of SB rather than a cause of it. Also, as deformation progresses, a transition from synchronous to separate actions between dilatation, shear and rotation takes place, with dilatation representing the predominant softening mechanism that triggers SB generation. The loading type significantly affects the driving force of SB evolution, while TSR controls the SB propagating velocity and gradually reduces it. Finally, SB evolution often undergoes stop-and-go cycles due to the sticking–slipping behavior of the primary SB, which alternates between cold and hot banding driven by structural disordering and free-volume softening.

4.10. High-Entropy Alloys

Recent research has been carried out on the ASB sensitivity of high-entropy alloys (HEAs). FCC single-phase CrMnFeCoNi HEA has shown remarkable resistance against ASB formation due to its increased strain rate hardening and moderate thermal softening magnitude, revealing a critical shear strain of 7 [96]. However, in the case of ASB occurrence, twins and RDR ultrafine grains compose the ASB core. FCC single-phase CrCoNi HEA has also disclosed the presence of twins and recrystallized grains inside shear localization zones even at cryogenic temperature, while a transformation from the FCC to HCP phase occurred within the refined grains [142]. Further, FCC single-phase FeCoNiCrMo_0.2 has been reported to reveal a 25 μm wide ASB during hot extrusion, inside of which nano twins and elongated equiaxed grains dominated, with DRX being the softening mechanism instead of phase transformation, which did not take place [143]. Also, FCC Al_0.1CoCrFeNi HEA has generally shown high resistance against ASB formation due to its increased strain and strain rate hardening and modest thermal softening. However, Jiang et al. [144] reported the occurrence of ASB under dynamic compression in cryogenic temperature, which led to cracking failure. In fact, Figure 23 depicts the sequence of the stages that lead to ASB formation, indicating that the interaction between dislocations and nanotwin boundaries (TBs) reacts to DRX regions, whose expansion and coalescence results in ASB formation. Finally, DRX and grain subdivision have been indicated as the dominant microstructural mechanisms that drive ASB evolution during serrated chip formation in high-speed machining of CoCrFeMnNi HEA [145].

Therefore, high-entropy alloys have generally shown increased resistance to ASB formation due to their high enough strain hardening and moderate thermal softening. However, DRX-induced ASB has been reported mainly in cryogenic temperatures, dominated by DRX and RDR mechanisms.

4.11. Metal Matrix Composites

ASB formation has also been investigated in particulate-reinforced metal matrix composites (MMCps), with particle size revealing a strong effect on ASB initiation. In particular, Dai et al. [146] studied 2024 Al composites reinforced with SiC particles (SiC_p) at dynamic torsion of 1000 s⁻¹ strain rate utilizing a modified split Hopkinson torsion bar. The results indicated that ASB initiation is strongly influenced by particle size, while smaller particles facilitate the detection of ASBs by making them more easily readable. Finally, strain gradient was reported as the main driving force of ASB evolution in MMCp, as a higher gradient promotes strain localization and leads to shear instability. Similarly, SiC_p/Al2024-matrix composites with a 15% particulate volume fraction have been studied for a variety of particle sizes, which lie between 3.5 and 20 μm [147]. Dynamic compression tests were conducted in a split Hopkinson pressure bar at a 2000 s⁻¹ strain rate, showing that the particle size affects the ASB onset, while ASB was captured more clearly in specimens with smaller particles. Also, an analytical expression for critical shear strain was derived by implementing a strain gradient law into the classic deformation stability analysis [148], indicating that a lower particle size enhances strain localization in the matrix through increased strain gradient. Finally, Jia et al. [149] studied ASB development in Cu-Ag and Cu-Nb metal matrix composites subjected to plane strain compression, utilizing 2D crystal plasticity finite element models. The results indicated that ASB evolution around a single crystal is strongly conjugated to the initial crystal orientation and the orientations of its abutting crystals. Moreover, ASBs can initiate at heterophase interfaces, while cross-phase ASBs can evolve through highly localized strain paths in the interfaces.

In conclusion, particle size in MMCp was reported to massively affect the ASB onset, while smaller particles react to easier detection of ASB. Also, increased strain gradient seems to drive ASB evolution, while cross-phase ASBs have been captured to propagate through highly localized strain paths in the interfaces of heterophase alloys.

4.12. Metrics for ASB Sensitivity

Finally, several metrics have been expressed for assessing material propensity against ASB formation. Macroscopic instability criteria have been developed, aiming to identify critical shear strain γ_cr and strain rate ${\dot{\gamma }}_{cr}$, that indicate ASB formation. Considering thermoplastic constitutive relations, analytical expressions for γ_cr and ${\dot{\gamma }}_{cr}$ have been derived from Culver [98] and Recht [11], respectively. However, thermo-viscoplastic flow rules reflect a more detailed approach to ASB instability, further considering the effect of strain rate as Bai [14] and Staker [42] proposed, providing analytical expressions for γ_cr, indicating, however, that thermoplastic approaches can be sufficiently accurate for sensitivity materials with low strain rates. In all cases, low strain and strain rate hardening magnitudes and a strong thermal softening parameter result in lower γ_cr and ${\dot{\gamma }}_{cr}$, facilitating ASB formation.

Although γ_cr and ${\dot{\gamma }}_{cr}$ represent the most reliable indexes for evaluating material susceptibility against ASB, additional parameters have been derived. In particular, Wright [15] developed the ASB susceptibility parameter “χ” accounting for strain/strain rate hardening and thermal softening, suggesting that increased χ facilitates ASB formation. Also, the flow localization parameter “α” was introduced by Semiatin and Lahoti [20], indicating that α values greater than −5 are sufficiently high in order to cause severe shear localization, while finally, Molinari and Clifton [45] expressed a shear instability criterion for considering strain/strain rate hardening and thermal softening exponents.

5. Fracture and Damage

5.1. ASB Evolution

The ASB mechanism is strongly connected to dynamic and unpredicted failures due to its softening and degradation mechanisms, which result in deformation instability and rapid evolution. Except for the adiabatic condition needed for ASB formation, high strain rates secure rapid evolution, reacting to uncontrollable failure progress. Fracture and deformation maps of strain rate–temperature have been derived for several metals, indicating that ASB failure occurs under strain rates greater than 100 s⁻¹ for pure aluminum and Ti-6Al-4V and under strain rates greater than 1000 s⁻¹ for pure titanium, zirconium and mild steel [7]. ASB initiates at a critical strain when deformation instability occurs, progressing following a transition stage of inhomogeneous plastic deformation as deformed ASB, and finally leading to severely unstable deformation through rapid failure [100,150].

During that transition stage, where the magnitude of the softening mechanisms effect starts to equalize the strain/strain rate hardening effect, ASBs propagate through widening mechanisms, providing slight peaks and lows in the τ–γ curve. Figure 24 illustrates the three main ASB widening mechanisms, which are intersecting, branching and intertwining ABSs [151]. In particular, in the case of intersecting ASBs, widening starts by enhancing their junction, resulting in an increased area whose greater width expands along the ASBs (Figure 24a). Branching ASBs can develop via the bending of shear bands caused by the twist of the trapped matrix between two parallel shear bands (Figure 24b,c). That bending results in a slight deviation of the ASB direction from the maximum shear direction, allowing for the ASB to expand obliquely (SB3 in Figure 24d) while inactivating primary ASB expansion along the maximum shear direction. However, as the strain increases, a new ASB forms along the initial shear direction, reactivating it (new SB5 in Figure 24d), and both ABSs tend to reconnect along the maximum shear direction, increasing the initial ASB width (Figure 24e). Finally, two parallel and independent ASBs can intertwine by bending towards each other and connecting due to the softened surrounding matrix trapped between them (Figure 24f), which also similarly affects the reconnection of branched ASBs (Figure 24d,e).

Furthermore, ultrafine graining has been reported to trigger micro-ASB nucleation, which expands, interacts with each other and finally forms a micro-ASB network that divides the matrix into smaller subregions [91]. As the strain increases, macro-ASBs are generated from the coalescence of micro-ASBs and afterward widen and propagate. Zhang et al. [152] developed a coupled spall–ASB model in order to investigate the ASB network, which expands within dynamically expanding and collapsed thick-walled cylinders. Their model proposes that the strain energy is composed of deviatoric, tensile and compressive parts, and it is capable of predicting the spiral patterns of ASBs, their initial positions and their propagating paths. In fact, the coupling between the spall and ASB failure modes can generate complex cracks that can initiate from the interaction between two ASBs, finally leading to the rapid fragmentation of the shell. Regarding the ASB network, spiral patterns of the ASB propagating paths were revealed for both expanding and collapsed shells, while double-direction patterns occurred in the case of the expanding shell in contrast to the collapsed one, which revealed single-direction ASB paths, as Figure 25 depicts.

Also, Wang et al. [153] identified that the number and spacing of ASBs within a collapsed shell depend on both material properties and strain rate. For 304L stainless steel [153,154] and Ti-6Al-4V [153] collapsed/expanding shells, damage softening was reported as the key factor that drives ASB propagation, while prior thermal softening clarifies the locations where the ASBs initiate. Figure 26 depicts the stress and temperature fields from which the ASB evolution paths are detected, while Figure 26c ensures that damage softening reflects the driving force of ASB propagation, leading to stress collapse. Except for the thermal softening, possible material defects also act as ASB triggering mechanisms, clarifying their initiation points and their propagating paths, as Figure 27 illustrates. Finally, directions of higher density in micro-defects have also been considered to promote ASB nucleation and propagation, as proposed by Guo et al. [155] in the case of explosively hardened pure titanium whose micro-defects facilitated ASB evolution.

In conclusion, DRX ultrafine graining has been responsible for promoting micro-ASBs, which expand and widen through intersecting, branching and merging behaviors. As the strain increases, the developed network propagates, formulating primary macro-ASBs whose evolutions are driven by damage softening, finally leading to fracture. However, ASB initiation is triggered by prior thermal softening and possible microstructural defects.

5.2. Damage Evolution

As ASB formation progresses, transformed bands occur under a significant temperature rise inside the banding, which leads to strong thermal softening, reacting further to micro-voiding and even local melting. Micro-voids are initially nucleated inside the ASB and elongate as the strain increases, forming an ellipsoid shape. Following this, voiding coalescence leads to discontinuous micro-cracks, which propagate according to mode II shearing, facilitating cracking propagation along the ASB and final fracture. Figure 28 indicates micro-voiding as the responsible mechanism for the sudden sharp decrease in the τ–γ curve, causing rapid and uncontrollable damage [1,17,93]. However, recent findings have indicated material properties as the key factors that determine if ASB drives plasticity without any micro-voiding or if it leads to catastrophic damage through nucleating micro-voids [156]. Moreover, recent studies have reported that an additional massive temperature increase develops inside the ASB after its formation, which is attributed to the intense friction between the sliding crack flanks at the final stages of fracture [157,158], suggesting that its magnitude is significantly higher than the one of the temperature rise obtained during the transformed banding generation. Also, Guo et al. [159] reported through experiments that a massive temperature increase appears some microseconds (~30 μs) after ASB formation, concluding that the significant temperature rise reflects a result rather than a precursor of the ASB mechanism. In addition, the results indicated that ASB initiation occurred several microseconds after the sudden stress drop, which is attributed more to microdamage via voiding and cracking or the DRX-softened microstructure and not only to thermal softening.

Rogers [6] indicated that ASB can lead to either ductile or brittle fracture. Ductile fracture is manifested by severe plastic deformation before failure via micro-voiding nucleation in hot spots inside the ASB when the last one is weakened from the temperature rise. In fact, temperature rise plays an important role in the transition from ASB to failure, as it reflects the key factor of hot-spot nucleation. In particular, the temperature increase field within the ASB is described as having a discontinuous and periodic distribution, revealing hot spots that lead to micro-voiding genesis [160,161]. In contrast, brittle fractures do not appear to have significant deformation before failure, while they are manifested with deformation termination when the quenching of the ASB has led to a hard and brittle transformed structure. Thus, brittle fracture occurs under either transgranular or intergranular cracking propagation via mode II shearing, although without showing any micro-voiding. However, the ultrafine graining that occurred inside the ASBs contributed to inter/trans-granular cracking propagation, which is difficult to capture with optical microscopy due to the graining size, as Figure 29 depicts where the crack expands along the ASB path [44].

On the other hand, micro-voiding represents the main mechanism of cracking genesis in ductile fracture, reflecting the sudden drop in the τ–γ curve, leading to dynamic and rapid failure. Due to its significantly high temperature, the ASB-softened core has been reported as the preferred region for micro-void nucleation, growth and coalescence, leading to micro-cracking when the critical length is reached [162], as shown in Figure 30 [162,163]. In addition, Figure 31 illustrates TEM observations in pure α-titanium, indicating that high-density DRX regions occurred around the crack tips, locally softening the microstructure and facilitating cracking propagation along the ASB [4]. On the contrary, heavily deformed large grains (HDLGs) and twins are captured laterally to the ASB with the slight presence of sparse DRX nanograins, preventing mode I cracking opening into the surrounding matrix. In particular, Figure 31b depicts the width of DRX nanograins with white dashed lines around the crack tip (indicated with a black arrow at the bottom center), while laterally and between crack tips, the HDLG microstructure occurs together with sparse DRX islands and twins (Figure 31c,d).

Regarding the cracking extent, the prior α-phase in Ti-6Al-4V has been found to reduce the cracking propagation rate or even change its direction due to its high hardness and stiffness, which suppress the damage extent, as cracking expansion through graining requires greater energy [164]. Further, the magnitude of the DRX effect on the fracture extent has been quantified by Sergey et al. [165] through a damage criterion that implements the DRX temperature and strain rate in order to assess a critical temperature above which a sudden and sharp decrease in the τ–γ curve occurs, reflecting rapid failure. Finally, the number of cracks per square millimeter also reflects an index of the damage extent inside the ASB [166], as a higher cracking density and, as a consequence, a greater damage extent, is captured at the center of the ASB compared to its boundary with the surrounding matrix, as Figure 32 verifies through SEM observations for a dual-phase low-carbon steel.

To conclude, ASB is strongly conjugated to dynamic failure, the transition to which is driven mainly by damage softening. Specifically, hot spots inside the transformed ASB generate micro-voids, which gradually elongate along the ASB, receiving an ellipsoid shape and forming cracks through their coalescence. In that way, micro-voiding reacts to stress collapse in the stress–strain curve, leading to rapid failure through mode II cracking propagation, which is facilitated by dense DRX regions around crack tips that soften the ASB microstructure.

6. Testing Methods

6.1. Microstructural Characterization Techniques

Several experimental observation techniques have been implemented for ASB microstructural characterization. SEM and electron probe microanalysis (EPMA) have been utilized as proper tools for fractographic studies in order to obtain the fracture surface, dominant mechanisms and failure traces [4,162,166,167,168]. Also, a microstructural analysis coupled with SEM/TEM and electron backscatter diffraction (EBSD) has been conducted for sub-graining structure identification [169,170], while IPF and phase maps have been implemented for experimental illustrations of ASB interior nanograins [70,171]. Further, EBSD benefits from optical microscopy due to its higher spatial resolution, data accuracy and more adequate microstructural characterization. On the other hand, TEM represents the main tool for nano- and micro-structural observations, offering a higher resolution capacity [4,57,66,168,169,170], allowing for DRX regions to be obtained, especially when combined with X-ray diffraction [172,173], while bright/dark fields and diffraction contrast allow for crystal defects and dislocations to be captured in nanoscale. In a similar direction, a high-resolution analysis through X-ray nanoscale tomography, together with TEM, was implemented for examining the ASB evolution in BMGs [174]. However, various difficulties regarding sample preparation in TEM and X-ray techniques limit their application range, which, in contrast, can be addressed by implementing the focused ion beam (FIB) method [175,176], allowing for the observation of DRX and RDR nanostructures by utilizing TEM and automated crystal orientation mapping (ACOM-TEM) [176].

Despite the fact that a low ASB width makes ASB observations difficult due to a reduced electropolishing capability inside the ASB area, TEM remains a widely utilized tool for high-resolution ASB micro- and nano-structural examination. Therefore, TEM benefits in obtaining interior graining, characteristics of deformation mechanisms, DRX nanograins and transformed ASBs, while SEM/EBSD tools and optical microscopy are preferred for bulk materials and lower resolution observations of fracture topography, the identification of shear localization areas, deformed ASBs and grain sizing. Moreover, high-speed thermography techniques have been utilized for recording and visualizing the temperature field inside the ASB [158], while the fusile multi-coating method has been recently implemented for estimating the local temperature rise even up to melting point [177]. In fact, both a low ASB width and high enough temperature remain the key challenges in obtaining accurate temperature records, and they often lead to misestimations as a temperature rise of some hundreds of Kelvin degrees, such as 600 K [1,178] or 1100 °C [73], can be developed along a 10–20 μm ASB width. More specifically, infrared radiation (IR) thermographic techniques have been utilized for temperature measurements in low-carbon steels [179,180], while a two-pyrometer technique has been implemented via indium antimonide IR detectors and intensified CCD camera for low and high temperature ranges, respectively [181]. Finally, a novel plane-array infrared imaging system accompanied by digital image correlation (DIC) techniques has been utilized for visualizing the two-dimensional ASB propagating network in space [182].

To summarize, the main tools for microstructural observation and characterization consist of optical microscopy, SEM, TEM, EPMA, X-ray and EBSD with IPF mapping. In particular, optical microscopy has been utilized to identify the location of ASBs and for a preliminary macroscopic observation. SEM is also utilized for microscale observations, while fractographic examinations have been conducted by combining SEM, EBSD and EPMA tools. Also, the ASB structure regarding graining and phases can be assessed via EBSD and IPF maps. Finally, TEM remains the most utilized tool for nanoscale observations together with X-ray techniques, while DIC techniques have been utilized for visualizing strain fields.

6.2. Testing Methods and Applications

Regarding testing methods for ASB examination, numerous experimental techniques have been implemented utilizing both the geometrical softening of specimens and loading characteristics, which facilitate shear localization and, consequently, ASB formation. Regarding the first case, hat-shaped [79,120,183] and single/double-edge notched specimens [184,185] have been widely tested against ASB failure where geometrical discontinuities enhanced severe shear localization and ASB evolution. On the other hand, regarding the applied loading technique, torsional [150,186,187], compression [16,188,189,190] and compression/shear tests [191,192] have been mainly conducted to derive compressive and/or shear stress fields which benefit ASB formation through severe shear localization. In fact, the shear stress state enhances the material propensity to ASB formation, as loading scenarios of high shear-to-compression stress ratio, like angled compression tests considering specimens with inclination, increase the susceptibility to form an ASB even at lower strain levels [193].

In addition, several manufacturing process tests have been implemented in order to investigate their effects on the propensity to form an ASB, such as forging [18,20,21,22,194], extrusion [195] and rolling [23,24,25,196] in bulk forming, blanking/punching and deep drawing tests for sheet forming [25,26,27], high-speed machining [28,29,30,31,32,33,197,198,199,200] regarding cutting processes and projectile penetration via ballistic impact tests [34,35,36,37,38,39,40,41]. Figure 33a illustrates X-type and S-type ASB observations in Inconel 718 hot forging [21], while deformed and transformed ASBs are captured inside the serrated chip structure as the cutting speed increases in the high-speed machining of AISI 1045 steel, as shown in Figure 33b [30]. Figure 34 provides a comparative view of the ASB network through a cross-section of an AA2099 Al-Li alloy produced via conventional rolling (CR) and asymmetrical rolling (ASR) processes. In ASR, the two rollers rotate at different speeds, providing higher hardness near the fast-rolled surface [196]. Moreover, ASR seems to promote more ASB formation compared to CR, providing multiple types of ASBs with a greater range of tilt angle and width. In more detail, CR reveals a certain ASB type (SBa’) of a 30° tilt angle, while ASR reveals four different ASB types (SBa-SBb-SBc-SBd) with higher tilt angles and widths near a slow rolled surface. Finally, Liu et al. [195] studied the ASB susceptibility of Ti-23Nb-0.7Ta-2Zr-O gum metal produced via extrusion and equal channel angular pressing (ECAP). In the case of extruded gum metal, ASB consisted of deformed ultrafine graining with increased dislocation density in contrast to ECAP gum metal, which, however, reached a higher maximum temperature, providing a recrystallized structure to the ASB core. Eventually, extruded gum metal revealed an earlier ASB occurrence, showing higher propensity due to a lower critical shear strain. Nevertheless, the challenges in preventing ASB during manufacturing processes can be summarized by material propensity as a significant vulnerability to ASB has been reported between different materials and processing conditions, as specific cases promote adiabatic shearing, geometrical softening regarding the workpiece design and, finally, the optimum productivity which requires a sufficiently high processing rate.

Especially in the case of high-speed machining, experiments in Ti-6Al-4V cutting until speeds of 210 m/s showed that ultra-high cutting speeds cause the transition from continuously serrated chip formation to discontinuously segmented formulation, accompanied by a transition from ductile to brittle fracture [197]. In fact, ASB was captured to develop in the chip during the transition stages, resulting in a discontinuous segmented chip through brittle failure. A critical condition for segmented chip formation was also derived by introducing the Φ-parameter, which reflects the ratio of input energy to total dissipation energy. The Φ-parameter was found to increase proportionally to the cutting speed, while Φ-values greater than 1 were considered to result in segmented chip formation. Finally, material thermal convection and the strain rate sensitivity were indicated as key parameters for limiting segmentation. Also, Ye et al. [198] studied the instability of chip flow during high-speed machining by developing a cutting model that accounts for the inertial effect, thermal convection and tool-chip compression interaction. In particular, a critical condition for the onset of serrated chip flow was derived by introducing the dimensionless number D, which reflects the interaction between thermal softening and strain hardening. A d-parameter was found to control the transition from continuous to serrated chip flow, while a power law between serration frequency and Reynolds thermal number Pe was revealed. Cutting in metallic glasses has also been studied by Zeng et al. [199] regarding the transition from ductile to brittle SB. Specifically, ductile SBs led to a stable serrated chip flow at a low cutting depth, while brittle SBs resulted in the instability of serrated flow at a higher cutting depth, showing a discontinuous chip. Also, dilatation was reported to be the key factor for the transition from ductile to brittle SB, revealing a critical local dilatation volume of 16.7%. Finally, a linear relation between the dilatancy inside the SB and the SB evolution degree was identified. Bejjani et al. [200] also investigated the connection between ASB and segmented chips in the high-speed machining of Ti-MMC. The outcomes showed that the temperature within the ASB almost reached DRX levels, but no phase transformation occurred. ASB formation was captured at a γ_cr of 7.5 and a ${\dot{\gamma }}_{cr}$ of 4.5 × 10⁵ s⁻¹, while the ASB core consisted of elongated and equiaxed nanograins. Finally, hard graining inside the surrounding matrix was not found to either promote or suspend ASB.

In general, a proper testing method for ASB observation requires a sufficient loading velocity that is high enough for adiabatic shearing and large deformation levels in order to produce significant work and temperature rise via internal heat reacting to strain instability; it does not have a significant Poisson ratio effect like necking or barreling, which affect ASB expansion, and finally, an obtainable force–strain recording system is needed in order to verify microscopical observations with macroscopic instability in the τ–γ curve. In particular, cylindrical hat-shaped specimens have been widely examined by implementing a Hopkinson compression loading bar [201], as Figure 35a shows. The specimen is located between the incident bar activated by a striker bar and the transmitted bar, while the geometry of the specimen allows for shear stress concentration, bringing severe shear strain localization. By using the Hopkinson compression bar, high strain rates up to 104 s⁻¹ can be reached, while various states of deformation can be examined via stopper rings, which adjust the maximum stroke of the input bar and, in consequence, the final deformation. Also, single- or double-edge notched specimens have been tested against ASB failure, introducing a compact forced simple shear, as shown in Figure 35b via the initial tension or compressor in the Hopkinson bars, which reacts to intense shear localization due to geometrical softening [202]. Cracks propagate according to mode II shearing loading, while the ASB evolution and propagation velocity can be measured. However, the fact that the ASB is introduced by forced shear localization remains a disadvantage of that method as it does not allow for general conclusions regarding the dominant mechanisms around ASB evolution.

Moreover, torsion tests have been carried out in torsional Hopkinson bar [203], as depicted in Figure 35c, or impact torsional loading, allowing for strain rates of up to 10³ and 10⁴ s⁻¹, respectively. Solid and hollow cylinders have been used as examined specimens, with the latter benefitting from a more homogeneous stress field due to their lower stress/strain gradients with the radius in contrast to solid specimens. However, slender specimens are avoided for torsion tests due to their propensity to buckle. Finally, torsion tests offer easier recording of the stress–strain field, while failure initiation depends on both material and geometry. In addition, compression tests have concentrated a significant number of experimental studies on ASBs due to their simplicity regarding specimen geometry and test system configuration. Specifically, cylindrical specimens have been examined in either the Hopkinson pressure bar [204] (Figure 35d) or drop-weight (up to 600 kg) configurations, allowing for strain rates of 10⁴ and 10³ s⁻¹, respectively, while further electromagnetic guns have been implemented for pressure sources on incident bars, allowing for higher loads and strain rates of even 10⁶ s⁻¹. Compression/shear tests have also been implemented in cylindrical specimens with a 6° inclination, which facilitates collapse initiation and shear localization [205].

The already existing testing methods allow the field to develop novel procedures and techniques that would address several issues regarding easier implementation, stage-by-stage recording of the ASB evolution in fracture surfaces and the examination of multi-axial loading, which introduces more complex stress states representing better real loading scenarios. Also, real-time measurements of ASB propagation velocity and temperature rise attract significant interest as there is a wide field for improvements regarding utilized methods.

In conclusion,

Table 4. Comparative characteristics of testing methods.

Method	Utilization Frequency	Strain	Strain Rate	Shear Localization Source		Complexity	τ–γ Recording Difficulty
				Geometry	Material
Hat-shaped	$\times \times \times \times \times$	$\times \times \times \times \times$	$\times \times \times \times \times$ (104–106 s−1)	Y	Y	$\times \times$	$\times \times \times$
Single-/double-edge notched	$\times$	$\times \times \times$	$\times \times \times \times$ (103–104 s−1)	Y	N	$\times \times \times \times$	$\times \times \times \times \times$
Torsion	$\times \times \times$	$\times \times \times$	$\times \times \times$ (102–103 s−1)	Y	Y	$\times \times \times$	$\times$
Compression	$\times \times \times \times \times$	$\times \times$	$\times \times \times \times$ (103–104 s−1)	N	Y	$\times$	$\times \times$
Compression/shear	$\times \times$	$\times \times$	$\times \times \times \times$ (103–104 s−1)	N	Y	$\times$	$\times \times \times \times$

$\times$: very low; $\times \times$: low; $\times \times \times$: medium; $\times \times \times \times$: high; $\times \times \times \times \times$: very high; Y: yes, N: no.

contains a comparative overview of the mentioned testing techniques regarding their utilization frequencies, shear localization sources, the achieved strain/strain rate levels, the τ–γ curve recording capability and the complexity. Hat-shaped specimens and compression tests are the most widely implemented methods for ASB examination, while in contrast, notched specimen tests are not as frequent due to their geometrical softening effect, which significantly influences ASB formation and can lead to a misunderstanding of the material propensity factors or the characteristics of dominant mechanisms. Although compression tests allow for high strain rates, maximum deformation is restricted from bulk material resistance, which suppresses ASB occurrence due to lower plastic work and, consequently, internal heat, which provides strain instability through thermal softening. However, compression tests are characterized by higher simplicity, like hat-shaped specimens, while torsion tests offer easier τ–γ curve recording due to their simpler stress/strain field, thus allowing for a macroscopic assessment of the ASB initiation via the strain instability point.

7. Conclusions

This study summarized the critical characteristics of ASB’s microstructure and morphology, evolution and connection to fracture. The main goal was to highlight their stage-by-stage evolution, aiming to attribute a consecutive behavior to them by highlighting discrete propagating stages and dominant mechanisms. In more detail, the ASB is initially manifested as a deformed band via severe shear localization, which following is converted to a transformed band. The key transition factor is attributed to a significant temperature rise, which further activates thermal softening and microstructural softening mechanisms such as DRX, RDR and phase transformation. DRX enhances strain softening, facilitating micro-ASB formation, which expands and widens gradually following a two-step evolution mechanism. The ASB core consists of DRX ultrafine equiaxed grains of low-angled misorientation, while the surrounding matrix contains heavily deformed coarser graining. In fact, grain refinement, together with rapid cooling, has been considered responsible for the ultra-high hardness of the ASB core.

In addition, thermal softening and microstructural defects have been considered to trigger ASB initiation, while in contrast, damage softening drives ASB evolution. However, shear banding generation in BMGs has been attributed to free-volume softening, which precedes thermal softening, with the last one simply accelerating the transition to fracture. Hot spots inside the transformed ASB generate micro-voids that elongate along softened ASB cores and form discontinuous microcracks through their coalescence. As a result, micro-voiding reacts to strain instability and rapid stress collapse, bringing the transition to dynamic failure, which expands via mode II cracking propagation. The direction of the cracking expansion is along the ASB centerline, as dense DRX regions concentrate around crack tips, facilitating cracking extension along the softened ASB microstructure.

Regarding material propensity to ASB formation, both mechanical and thermal material properties play important roles, as the ASB is favored by weak strain hardening, strong thermal softening and low thermal diffusivity. From a microstructural point of view, increased pre-existing twins and increased porosity facilitate ASB formation. In particular, large pores trigger ASB genesis, while a non-uniform spatial porosity distribution promotes the formation of a single primary ASB instead of secondary and weaker ASBs, which are favored by a more uniform porosity. Also, duplex and bimodal microstructures have shown increased resistance against ASBs. Fine initial graining also seems to promote ASB evolution by enhancing DRX softening. Further, BMGs, NC and UFG materials revealed an increased propensity for ASB failure due to their reduced strain hardening, even at low strain rates. On the contrary, HEAs have not revealed a clear tendency against ASBs, as some revealed increased resistances due to high strain hardening and moderate thermal softening, while others suffered from DRX-induced ASB.

Finally, SEM has been widely utilized for microscale and mesoscale observations regarding shear localization and fractured surface, while in contrast, TEM and X-ray have been established for nanoscale microstructural characterization inside ASBs. However, the TEM and X-ray observation range is limited by difficulties faced during sample preparation, which can be addressed by implementing the FIB method. Also, the ASB graining structure can be assessed via EBSD and IPF maps. Regarding testing methods, hat-shaped specimens have been widely tested, benefiting from geometrical softening, while dynamic compression/shear tests have concentrated the larger number of experimental studies as they promote severely localized shear stress fields.

8. Recommendations for Future Research

By reviewing some critical points of the ASB phenomenon, this work aims to identify and highlight some aspects that remain dark and unclarified. Through this study, the following concerns are reported to remain fuzzy and unclear, allowing for further investigation in order to provide more objective findings.

8.1. Micro-/Macroscopic Coupled Approach

A micro-/macroscopic structural-thermal coupled analysis is needed in order to reveal a more objective approach by reflecting the effects of microstructural mechanisms like DRX, RDR and phase transformation on the material softening behavior and, as a consequence, strain instability. In specific, a micro-/macroscopic structural–thermal coupled approach would allow for the implementation of ASB evolution characteristics such as widening mechanisms, interacting phenomena, critical softening temperature and micro-voiding growing mechanisms, whose effects on damage criteria have yet to be widely implemented.

8.2. Damage Assessment

More objective criteria regarding material susceptibility against ASB have to be derived, taking into account more microstructural and thermal material properties. For example, most instability criteria that propose γ_cr as proportional to strain hardening lack reliability in the case of some NC materials, which have almost zero or even negative strain hardening exponents, estimating very small γ_cr values and an increased ASB propensity, which, however, does not agree with the experimental results [125]. That highlights the importance of developing stronger instability criteria, which would consider the effects of porosity, grain sizing and pre-existing twins, and also softening mechanisms like DRX and damage softening, which could affect the strain failure or react to critical softening temperature considered in the damage criteria.

8.3. Temperature Rise

The effect of a temperature rise on ASB evolution allows for further investigation regarding its timing, as several studies have reported its appearance during final fracture due to friction between the crack flanks [157,158], indicating, in contrast, a moderate temperature increase during thermal softening. Also, although it is commonly accepted that strain instability is triggered by the superiority of thermal softening against strain/strain hardening, recent findings report that ASB evolution is affected more by damage softening rather than by thermal softening. Therefore, temperature rise and thermal softening need to be reconsidered in order to determine if they precede ASB formation or if they are a result of them.

8.4. Taylor–Quinney Coefficient

Similarly, the aspect of energy conversion between plastic work and internal heat needs to be re-quantified more carefully. In particular, the widely utilized assumption of a constant Taylor–Quinney factor of 90% needs to be reconsidered, as recent results have shown a strong influence mainly on material and secondary on loading mode and strain level, revealing significant differences regarding its value [102].

8.5. ASB Propagation

The ASB propagating paths and velocity can also be examined more extensively, as they are affected by the loading type and respective stress states, as in the case of BMGs, which can be analyzed further by implementing crystal plasticity models [206] and molecular dynamic simulations [77,131,132,133], whose investigation remains shallow. In this direction, the effect of dynamic strain aging on shear localization can be studied, focusing on the interaction between solute atoms and dislocations. However, the effect of the strain rate sensitivity on the propensity to ASB remains to be seen, as ASB formation, even under quasi-static loading, has been reported [7,91], implementing different softening mechanisms during ASB propagation and allowing for further research directions.