New Metastable Baro- and Deformation-Induced Phases in Ferromagnetic Shape Memory Ni2MnGa-Based Alloys

Structural and phase transformations in the microstructure and new metastable baro- and deformation-induced phases of the Ni50Mn28.5Ga21.5 alloy, typical of the unique class of ferromagnetic shape memory Heusler alloys, have been systematically studied for the first time. Phase X-ray diffraction analysis, transmission and scanning electron microscopy, and temperature measurements of electrical resistivity and magnetic characteristics in strong magnetic fields were used. It was found that in the course of increasing the pressure from 3 to 12 GPa, the metastable long-period structure of martensite modulated according to the 10M-type experienced transformation into a final non-modulated 2M structure. It is proved that severe shear deformation by high pressure torsion (HPT) entails grainsize refinement to a nanocrystalline and partially amorphized state in the polycrystalline structure of the martensitic alloy. In this case, an HPT shear of five revolutions under pressure of 3 GPa provided total atomic disordering and a stepwise structural-phase transformation (SPT) according to the scheme 10M → 2M → B2 + A2, whereas under pressure of 5 GPa the SPT took place according to the scheme 10M → 2M → B2 → A1. It is shown that low-temperature annealing at a temperature of 573 K caused the amorphous phase to undergo devitrification, and annealing at 623–773 K entailed recrystallization with the restoration of the L21 superstructure in the final ultrafine-grained state. The size effect of suppression of the martensitic transformation in an austenitic alloy with a critical grain size of less than 100 nm at cooling to 120 K was determined. It was established that after annealing at 773 K, a narrow-hysteresis thermoelastic martensitic transformation was restored in a plastic ultrafine-grained alloy with the formation of 10M and 14M martensite at temperatures close to those characteristic of the cast prototype of the alloy.

There exist several ways to improve the ductility of intermetallics, among which grain size reduction and atomic disordering should be mentioned [5,18]. It has been established that the combined technologies of ultra-rapid quenching (URQ) from the melt [21][22][23][24][25][26][27][28][29][30] and severe plastic deformation (SPD) [31][32][33][34][35][36] can provide production of SM alloys of individual chemical compositions based on titanium nickelide and copper, in a high-strength and ductile fine-grained (FG) [or ultra-fine-grained (UFG)] state. A URQ from the melt entails an increase in the strength and ductility of FG alloys with a mean grain size in the interval 0.2-1.0 µm, depending on the regime of subsequent thermal treatment. SPD provides for the formation of nanocrystalline UFG structure of the alloys (with the size of nanograins less than 100 nm). This leads to their strong hardening, but entails a decrease in their ductility. However, the ductility of these SPD UFG alloys can be significantly improved by subsequent low-temperature short-term annealing. An understanding of the combined treatment under discussion has been successful in (for example) the creation of wire demonstrating high strength and good plasticity, and of rods or sheets with SME for alloys based on titanium nickelide and copper. Recently, work on the creation of UFG structures using these methods of extreme external influences was performed on Ni 2 MnGa Heusler alloys by URQ from the melt [37][38][39][40][41][42][43], magnetron sputtering techniques [44,45] or the SPD methods [24,[46][47][48][49], including ball milling [50].
In the present article, a comprehensive systematic analysis of structural-phase transformations was carried out for the first time in the three-component SM Ni 2 MnGa-based alloy, subjected to combined SPD by high-pressure compression (HPC) and subsequent high-pressure torsion (HPT), as well as heat treatment. This work studies the influence of applied HPC and HPT on the transformation behavior and evolution of physical electrical and magnetic properties in the investigated low elastic-modulus Ni 50 Mn 28.5 Ga 21.5 alloy.

Materials and Methods
The Ni 50 Mn 28.5 Ga 21.5 alloy was melted out from nickel, manganese, and gallium of 99.99% purity in an electric arc furnace (Institute of Metal Physics, Ekaterinburg, Russia) in a helium atmosphere with triple remelting and subsequent long-term homogenizing annealing in the temperature range of 1073-1173 K. The chemical composition of the alloy according to integral spectral analysis was Ni-28.50 at.% Mn-21.50 at.% Ga. The grain sizes of the cast alloy reached several millimeters.
Samples for deformation in Bridgman anvils were made in the form of disks with a diameter (D) of up to 10 mm and a thickness (t) of 0.5 mm. The amplitude of high pressure was determined by the ratio of the value of applied load and the square of the anvil. True logarithmic deformation was determined as e = ln(t i /t f ) + ln(πnD/t f ), where t i and t f are the initial and the final thickness (of the sample), n is the number of revolutions. HPC deformation was performed at 3 and 12 GPa, and HPT deformation was performed at 3 GPa and five revolutions, and at 5 GPa and two and five revolutions HPC and HPT were carried out at room temperature (RT). The subsequent isothermal vacuum-sustained annealing was carried out for 10 min at temperatures in the interval of 373-973 K.
The temperature measurements of electrical resistivity and magnetic characteristics were carried out by the potentiometric method, on the MPMS-5XL SQUID magnetometer     Figure 2 shows the hierarchy of the packet micromorphology of 10M martensite in the cast alloy. It is seen that the packet micromorphology is formed by primary microcrystals (Figure 2a), twinned pairwise according to the data from electron back-scattered diffraction (EBSD) (Figure 2c,d), and contains larger magnetic microdomains in accordance with Lorentz microscopy (Figure 2b).    Figure 3 shows that, according to TEM data, secondary nanotwins were present inside individual martensite crystals. Figure 3a demonstrates the mechanism of intragrain formation of a packet of nano-twinned 10M martensite crystals in a TEM in situ experiment after heating by an electron beam. In the regions of the residual initial high-temperature L21 phase, a typical pre-martensitic tweed contrast is visible, accompanied by the appearance of diffuse streaks along <110> and satellites in positions of type 1/6 <220> L21 in the SAED patterns (see Figure 3a). Cooling in situ TEM from RT to 120 K led to the second TMPT with the formation of nano-twinned long-period 14M martensite packets ( Figure 3b) [52][53][54].   Figure 3 shows that, according to TEM data, secondary nanotwins were present inside individual martensite crystals. Figure 3a demonstrates the mechanism of intragrain formation of a packet of nano-twinned 10M martensite crystals in a TEM in situ experiment after heating by an electron beam. In the regions of the residual initial high-temperature L2 1 phase, a typical pre-martensitic tweed contrast is visible, accompanied by the appearance of diffuse streaks along <110> and satellites in positions of type 1/6 <220> L2 1 in the SAED patterns (see Figure 3a). Cooling in situ TEM from RT to 120 K led to the second TMPT with the formation of nano-twinned long-period 14M martensite packets ( Figure 3b) [52][53][54].    Figure 3 shows that, according to TEM data, secondary nanotwins were present inside individual martensite crystals. Figure 3a demonstrates the mechanism of intragrain formation of a packet of nano-twinned 10M martensite crystals in a TEM in situ experiment after heating by an electron beam. In the regions of the residual initial high-temperature L21 phase, a typical pre-martensitic tweed contrast is visible, accompanied by the appearance of diffuse streaks along <110> and satellites in positions of type 1/6 <220> L21 in the SAED patterns (see Figure 3a). Cooling in situ TEM from RT to 120 K led to the second TMPT with the formation of nano-twinned long-period 14M martensite packets ( Figure 3b) [52][53][54].  According to the XRD data, HPC at 3 and 12 GPa affected the phase composition of the alloy (Figure 1b,c), leading to partial (after 3 GPa) or complete TMPT (after 12 GPa) into the so-called non-modulated martensite with tetragonal structure of type 2M ( Table 1). The grain sizes, as a rule, did not change, despite the tenfold excess of pressure at HPC over the strength limit of this low-modulus alloy. An application of HPT under a pressure of 5 GPa and two and five revolutions caused further radical phase changes with formation in the alloy-according to the XRD analysis-of the γ-FCC structure (of type A1) (see Figure 1d-f; and Table 1). At the same time, after HPT for two revolutions, the preservation of traces of the L2 1 phase was noted (Figure 1d, Table 1).
According to the TEM data, HPT under a pressure of 3 GPa and five revolutions completely changed the microstructure of the alloy (Figure 4a). The dark-field TEM image of the alloy structure shows that the sizes of nano crystallites were 10-20 nm. When cooled to 120 K in the mode of TEM in situ, the dimensional and morphological features of the nanostructured B2 state in the alloy were preserved (insert in Figure 4a). The reflections in the SAED pattern (see the insert in Figure 4a) were distributed over the rings and had the following hkl B2-phase indices: 100, 110, 200, 211, etc. Moreover, judging by the halo-a weaker individual diffusive ring near the 100 position-the alloy was in a partially amorphous state at cooling up to 120 K. Methodically reliable identification of the β phases L2 1 , B2, and A2 was provided only by visualization of weak superstructural reflections in the SAED patterns of the types 111 and 200 from L2 1 , 100 from B2, or their absence for the A2 structure. In this case, we can only conclude unequivocally that the (B2 + A2) nanostructure is dominant in the alloy, since super-structure reflection of 111 L2 1 type was absent in the SAED patterns. According to the XRD data, HPC at 3 and 12 GPa affected the phase composition of the alloy (Figure 1b,c), leading to partial (after 3 GPa) or complete TMPT (after 12 GPa) into the so-called non-modulated martensite with tetragonal structure of type 2M ( Table  1). The grain sizes, as a rule, did not change, despite the tenfold excess of pressure at HPC over the strength limit of this low-modulus alloy. An application of HPT under a pressure of 5 GPa and two and five revolutions caused further radical phase changes with formation in the alloy-according to the XRD analysis-of the γ-FCC structure (of type A1) (see Figure 1d-f; and Table 1). At the same time, after HPT for two revolutions, the preservation of traces of the L21 phase was noted (Figure 1d, Table 1).
According to the TEM data, HPT under a pressure of 3 GPa and five revolutions completely changed the microstructure of the alloy (Figure 4a). The dark-field TEM image of the alloy structure shows that the sizes of nano crystallites were 10-20 nm. When cooled to 120 K in the mode of TEM in situ, the dimensional and morphological features of the nanostructured B2 state in the alloy were preserved (insert in Figure 4a). The reflections in the SAED pattern (see the insert in Figure 4a) were distributed over the rings and had the following hkl B2-phase indices: 100, 110, 200, 211, etc. Moreover, judging by the haloa weaker individual diffusive ring near the 100 position-the alloy was in a partially amorphous state at cooling up to 120 K. Methodically reliable identification of the β phases L21, B2, and A2 was provided only by visualization of weak superstructural reflections in the SAED patterns of the types 111 and 200 from L21, 100 from B2, or their absence for the A2 structure. In this case, we can only conclude unequivocally that the (B2 + A2) nanostructure is dominant in the alloy, since super-structure reflection of 111 L21 type was absent in the SAED patterns. After HPT at a pressure of 5 GPa and five revolutions, similar TEM images of the nanostructured state of the alloy were observed. However, SAED patterns had fundamentally changed (inserts in Figure 4b,c). According to the results of their indexing, in accordance with the XRD analysis (Figure 1e,f) the alloy has a structure of γ-FCC up to 120 K (of type A1: with hkl indices 111, 200, 220, 311, etc.). It can be assumed that the amorphous component detected due to continuous diffuse halos (inserts in Figure 4 a-c) was localized on the blurred, sinuous intercrystalline interfaces of γ nanocrystals clearly visualized in direct atomic resolution images (Figure 4d). It is along these that the bending contours of extinction were localized in TEM images.
As was already noted, XRD analysis showed that the Ni50Mn28.5Ga21.5 alloy subjected to HPT at 5 GPa and five revolutions, in accordance with TEM data, had a γ-FCC structure of type A1 (Figure 1e,f), which was preserved during cooling in situ to 130 K. However, After HPT at a pressure of 5 GPa and five revolutions, similar TEM images of the nanostructured state of the alloy were observed. However, SAED patterns had fundamentally changed (inserts in Figure 4b,c). According to the results of their indexing, in accordance with the XRD analysis (Figure 1e,f) the alloy has a structure of γ-FCC up to 120 K (of type A1: with hkl indices 111, 200, 220, 311, etc.). It can be assumed that the amorphous component detected due to continuous diffuse halos (inserts in Figure 4a-c) was localized on the blurred, sinuous intercrystalline interfaces of γ nanocrystals clearly visualized in direct atomic resolution images (Figure 4d). It is along these that the bending contours of extinction were localized in TEM images.
As was already noted, XRD analysis showed that the Ni 50 Mn 28.5 Ga 21.5 alloy subjected to HPT at 5 GPa and five revolutions, in accordance with TEM data, had a γ-FCC structure of type A1 (Figure 1e,f), which was preserved during cooling in situ to 130 K. However, all the Bragg reflections were noticeably broadened. Annealing at 373 and 473 K did not change the described nanostructural γ state of the HPT-ed alloy (Figure 5a,b). After annealing at 573 K, the diffuse halo on the SAED patterns had almost disappeared, indicating the realization of a thermally activated process of devitrification of the amorphous component (Figure 5c). all the Bragg reflections were noticeably broadened. Annealing at 373 and 473 K did not change the described nanostructural γ state of the HPT-ed alloy (Figure 5a,b). After annealing at 573 K, the diffuse halo on the SAED patterns had almost disappeared, indicating the realization of a thermally activated process of devitrification of the amorphous component ( Figure 5c). Annealing at 623 K led to more noticeable structural changes in the alloy ( Figure 6). As a result of the primary recrystallization, the size of the nanograins significantly increased in the wide range of 50-150 nm. However, in the SAED pattern of the insert in Figure 6b, superstructural reflections distributed by the rings (of types 100 from the B2, or 111 and 200 from the L21 phases) were not practically resolved, excluding individual marked reflections (200 L21, 125 10M, 127 14M). Cooling to 120 K revealed the presence of 10M martensite reflections (for example, see Figure 6d, insert). Figure 6c,d shows brightand dark-field TEM images of twinning inside martensitic grains larger than 100 nm. It can be assumed that TMPT occurred in L21 grains. Annealing at 623 K led to more noticeable structural changes in the alloy ( Figure 6). As a result of the primary recrystallization, the size of the nanograins significantly increased in the wide range of 50-150 nm. However, in the SAED pattern of the insert in Figure 6b Figure 6c,d shows bright-and dark-field TEM images of twinning inside martensitic grains larger than 100 nm. It can be assumed that TMPT occurred in L2 1 grains. all the Bragg reflections were noticeably broadened. Annealing at 373 and 473 K did not change the described nanostructural γ state of the HPT-ed alloy (Figure 5a,b). After annealing at 573 K, the diffuse halo on the SAED patterns had almost disappeared, indicating the realization of a thermally activated process of devitrification of the amorphous component ( Figure 5c). Annealing at 623 K led to more noticeable structural changes in the alloy ( Figure 6). As a result of the primary recrystallization, the size of the nanograins significantly increased in the wide range of 50-150 nm. However, in the SAED pattern of the insert in Figure 6b, superstructural reflections distributed by the rings (of types 100 from the B2, or 111 and 200 from the L21 phases) were not practically resolved, excluding individual marked reflections (200 L21, 125 10M, 127 14M). Cooling to 120 K revealed the presence of 10M martensite reflections (for example, see Figure 6d, insert). Figure 6c,d shows brightand dark-field TEM images of twinning inside martensitic grains larger than 100 nm. It can be assumed that TMPT occurred in L21 grains. Recrystallization annealing at 773 K caused more noticeable grain growth and, obviously, a complete restoration of the perfect metastable L2 1 superstructure. The grain sizes did not exceed 400 nm. As a result, 10M martensite in the UFG state was observed at RT (Figure 7a,b). When cooled by the in situ TEM method, the alloy experienced a second TMPT of the 10M → 14M type (Figure 7c,d). After annealing at 873 K, the grain sizes did not exceed 1 micron. The nanotwin morphology of 14M martensite inherited from 10M martensite also differed, as in the case of 10M, by the exceptionally single-packet character of individual grains. Recrystallization annealing at 773 K caused more noticeable grain growth and, obviously, a complete restoration of the perfect metastable L21 superstructure. The grain sizes did not exceed 400 nm. As a result, 10M martensite in the UFG state was observed at RT (Figure 7a,b). When cooled by the in situ TEM method, the alloy experienced a second TMPT of the 10M → 14M type (Figure 7c,d). After annealing at 873 K, the grain sizes did not exceed 1 micron. The nanotwin morphology of 14M martensite inherited from 10M martensite also differed, as in the case of 10M, by the exceptionally single-packet character of individual grains. Thus, in the Ni50Mn28.5Ga21.5 alloy, it is established that as the pressure value and the degree of shear SPD increased, a baro-(pressure-provided) and deformation-induced stepwise structural-phase transformation occured according to the scheme 10M → 2M → B2 → A1, and at the same time cascade-wise atomic disordering occured. The effect of stabilization of the nanostructured phases B2-BCC, A1-FCC, and L21 austenite with respect to TMPT was found at a critical grain size of up to 100 nm when the alloy was cooled to 120 K. The obvious reasons for the suppression of TMPT in Ni-Mn-Ga-based Heusler alloys and the like, as well as in titanium nickelide alloys, are (i) deformation-induced reduction in the size of nanograins below the critical level, (ii) atomic disordering, (iii) amorphization, and (iv) the detected sequential SPT 10M → 2M → B2→A1. The subsequent recrystallization annealing ensures (i) the growth of grains characteristic of the UFG state, (ii) restoration of the atomic order inherent to the L21 type, and, as a result, (iii) restoration of reversible TMPTs.
The baro-induced transition 10M → 2M occurred (i) under the influence of high pressure at 3 and 12 GPa with an increase in the value of the specific volume, as well as (ii) in the nanophases B2 and A1 formed in the course of HPT at high pressure of 3 and 5 GPa (at n = 2 or 5, respectively, see Table 1). Consequently, the appearance of the observed phases in the sequence 10M → 2M → B2→A1 was due not so much to the pressure as to the actual intense SPD (for example, e = 7 after five revolutions at the site of half of the disk radius). Both pressure-induced (under the influence of high pressure) and deformation-induced (under shear SPD) phase transitions were accompanied by atomic and Thus, in the Ni 50 Mn 28.5 Ga 21.5 alloy, it is established that as the pressure value and the degree of shear SPD increased, a baro-(pressure-provided) and deformation-induced stepwise structural-phase transformation occured according to the scheme 10M → 2M → B2 → A1, and at the same time cascade-wise atomic disordering occured. The effect of stabilization of the nanostructured phases B2-BCC, A1-FCC, and L2 1 austenite with respect to TMPT was found at a critical grain size of up to 100 nm when the alloy was cooled to 120 K. The obvious reasons for the suppression of TMPT in Ni-Mn-Ga-based Heusler alloys and the like, as well as in titanium nickelide alloys, are (i) deformationinduced reduction in the size of nanograins below the critical level, (ii) atomic disordering, (iii) amorphization, and (iv) the detected sequential SPT 10M → 2M → B2→A1. The subsequent recrystallization annealing ensures (i) the growth of grains characteristic of the UFG state, (ii) restoration of the atomic order inherent to the L2 1 type, and, as a result, (iii) restoration of reversible TMPTs.
The baro-induced transition 10M → 2M occurred (i) under the influence of high pressure at 3 and 12 GPa with an increase in the value of the specific volume, as well as (ii) in the nanophases B2 and A1 formed in the course of HPT at high pressure of 3 and 5 GPa (at n = 2 or 5, respectively, see Table 1). Consequently, the appearance of the observed phases in the sequence 10M → 2M → B2→A1 was due not so much to the pressure as to the actual intense SPD (for example, e = 7 after five revolutions at the site of half of the disk radius). Both pressure-induced (under the influence of high pressure) and deformation-induced (under shear SPD) phase transitions were accompanied by atomic and structural disordering. It is this phenomenon that actually determined the key physical nature of the detected phase transitions. It is obvious that in this case, annealing, which restored the atomic order and type of superstructure, as well as grain size above the critical level with a significant decrease in the number and density of defects in their internal structure, led to the implementation of reversible TMPTs inherent in the alloy, as found in the investigations conducted.
A more complete physical interpretation of data obtained for the first time on new deformation-induced phases and structural-phase transformations in the Ni 50 Mn 28.5 Ga 21.5 alloy was provided by using the results of temperature measurements of electrical resistivity ρ(T) in the initial cast state and after HPT (Figure 8a,b). In the figure, the arrows show the directions of the temperature change cycles during measurements of ρ(T), starting from RT to 4.2 K during cooling, then up to 800 K during heating, reverse cooling to 4.2 K and again reheating to RT. First, it can be seen that the dependence ρ(T) changed after the HPT. Namely, the alloy subjected to HPT had high values of ρ = 140 µΩ·cm below 400 K, and its low-temperature behavior was characterized by an anomalous negative temperature coefficient of resistivity (TCR) compared to the normal TCR for the original cast prototype alloy. The value of the residual electrical resistivity ρ 0 at a temperature of 4.2 K in the studied HPT samples showed an almost four-fold increase compared to the value ρ 0 of the cast alloy.
in the investigations conducted.
A more complete physical interpretation of data obtained for the first time on new deformation-induced phases and structural-phase transformations in the Ni50Mn28.5Ga21.5 alloy was provided by using the results of temperature measurements of electrical resistivity ρ(T) in the initial cast state and after HPT (Figure 8a,b). In the figure, the arrows show the directions of the temperature change cycles during measurements of ρ(T), starting from RT to 4.2 K during cooling, then up to 800 K during heating, reverse cooling to 4.2 K and again reheating to RT. First, it can be seen that the dependence ρ(T) changed after the HPT. Namely, the alloy subjected to HPT had high values of ρ = 140 μΩ·cm below 400 K, and its low-temperature behavior was characterized by an anomalous negative temperature coefficient of resistivity (TCR) compared to the normal TCR for the original cast prototype alloy. The value of the residual electrical resistivity ρ0 at a temperature of 4.2 K in the studied HPT samples showed an almost four-fold increase compared to the value ρ0 of the cast alloy.
Second, in the high-resistivity HPT alloy, the characteristic anomalies inherent to the TMPTs were not observed on the curves ρ(T). Finally, subsequent heating in the interval from 600 to 800 K led to a decrease in ρ(T) almost to the value of the initial cast alloy, and then the appearance of a narrow hysteresis loop due to the reversible TMPT (Figure 8b). The measured critical temperatures of TMPTs are given in Table 2. The behavior of ρ(T) during heating after cooling, shown in Figure 8b, clearly reveals two stages: relaxation of elastic-plastic distortions (recovery) in the temperature range from RT to 550 K and recrystallization in the range 550-800 K.  Figure 8c shows the results of measurements of the magnetic properties of the Ni50Mn28.5Ga21.5 alloy in the cast state and after HPT. When analyzing the magnetization of M(T), measurements were carried out starting from RT down to 4.2 K when cooling, then when heating to 400 K (curves 1 and 2) or up to RT (curve 3). From the data obtained, it can be seen that the application of a strong magnetic field H = 4 MA/m increased the temperatures of Ms and Mf by 10 and 5 K, respectively (Table 2). Also, attention is drawn to the sharp decrease in the magnetization of M(T) at RT in the HPT alloy compared to M(T) of the prototype alloy. However, when cooled to 4.2 K, the value of M(T) increased noticeably (Figure 8c, curve 3). The magnetization M(T) on cooling approached the value 60 A·m 2 /kg, whereas for the alloy in the initial state the value of its magnetization M(T) was greater than 80 A·m 2 /kg. Thus, analysis of the results of the structural studies and physical measurements shows that when the temperature drops below RT, despite the Second, in the high-resistivity HPT alloy, the characteristic anomalies inherent to the TMPTs were not observed on the curves ρ(T). Finally, subsequent heating in the interval from 600 to 800 K led to a decrease in ρ(T) almost to the value of the initial cast alloy, and then the appearance of a narrow hysteresis loop due to the reversible TMPT (Figure 8b). The measured critical temperatures of TMPTs are given in Table 2. The behavior of ρ(T) during heating after cooling, shown in Figure 8b, clearly reveals two stages: relaxation of elastic-plastic distortions (recovery) in the temperature range from RT to 550 K and recrystallization in the range 550-800 K.   Figure 8c shows the results of measurements of the magnetic properties of the Ni 50 Mn 28.5 Ga 21.5 alloy in the cast state and after HPT. When analyzing the magnetization of M(T), measurements were carried out starting from RT down to 4.2 K when cooling, then when heating to 400 K (curves 1 and 2) or up to RT (curve 3). From the data obtained, it can be seen that the application of a strong magnetic field H = 4 MA/m increased the temperatures of M s and M f by 10 and 5 K, respectively (Table 2). Also, attention is drawn to the sharp decrease in the magnetization of M(T) at RT in the HPT alloy compared to M(T) of the prototype alloy. However, when cooled to 4.2 K, the value of M(T) increased noticeably (Figure 8c, curve 3). The magnetization M(T) on cooling approached the value 60 A·m 2 /kg, whereas for the alloy in the initial state the value of its magnetization M(T) was greater than 80 A·m 2 /kg. Thus, analysis of the results of the structural studies and physical measurements shows that when the temperature drops below RT, despite the lower magnetization value, the HPT-ed γ alloy is in a partially magnetically ordered state, obviously genetically related to the original L2 1 alloy structure.
In conclusion, we note that the HPT-ed alloy, like its prototype alloy, experienced brittle fracture-destruction when bending. At the same time, after the formation of the UFG structure during annealing at 773 K and 873 K, the alloy became sufficiently plastic to bend and under subsequent heating experienced reversible deformation into a flat shape, showing SM effect.

Summary and Conclusions
In this work, we established the effect of SPD on Ni 50 Mn 28.5 Ga 21.5 alloy using highpressure uniaxial compression (HPC) and torsion (HPT). The following main conclusions are made:

1.
It was found that with an increase in pressure from 3 to 12 GPa, a metastable longperiod 10M martensitic structure underwent a baro-induced transformation into a non-modulated 2M structure. 2.
HPT radically refined the polycrystalline grain structure to a nanocrystalline, partially amorphized state at the grain junctions. As the pressure increased (up to 5 GPa) and the degree of deformation achieved the value of true deformation e = 7 (after five revolutions), a deformation-induced atomic disordering evolved and a stepwise structural-phase transformation occurred according to the scheme 10M→2M → B2→A1.

3.
HPT shear of five revolutions under a pressure of 3 GPa provided incomplete deformationinduced atomic disordering and a stepwise structural-phase transformation according to the scheme 10M → 2M → B2 + A2.

4.
Annealing at a temperature of 573 K caused the amorphous phase to decompose, and recrystallization annealing at 773-873 K entailed the formation of the L2 1 UFG structure characterized by a narrow-hysteresis TMPT with critical temperatures close to those characteristic of the cast prototype alloy.

5.
It was found that the dimensional effect of complete stabilization of a nanostructured alloy with respect to TMPT when cooled to 120 K was realized at a grain size of less than 100 nm. 6.
The electrical resistivity of the HPT-ed γ alloy increased significantly during cooling, demonstrating a negative value of temperature coefficient of resistivity. In this case, the alloy was in a partially magnetically ordered high-resistivity state, the magnetization of which increased significantly in strong magnetic fields at low temperatures. 7.
It can be assumed that the creation of a UFG structure in the alloy allowed its ductile properties to increase to those necessary for the implementation of an SM effect.
Author Contributions: Conceptualization, methodology, analysis, writing-review V.P.; methodology, investigation, analysis, writing-original draft preparation A.K., N.K., E.M. and Y.U. All authors have read and agreed to the published version of the manuscript.
Funding: This work was performed within the framework of state task "Structure", grant No. AAAA-A18-118020190116-6.
Institutional Review Board Statement: Not applicable.