Design and Development of High-Strength and Ductile Ternary and Multicomponent Eutectoid Cu-Based Shape Memory Alloys: Problems and Perspectives

Abstract

An overview is presented on the structural and phase transformations and physical and mechanical properties of those multicomponent copper-based shape memory alloys which demonstrate attractive commercial potential due to their low cost, good shape memory characteristics, ease of fabrication, and excellent heat and electrical conductivity. However, their applications are very limited due to brittleness, reduced thermal stability, and mechanical strength—properties which are closely related to the microstructural features of these alloys. The efforts of the authors of this article were aimed at obtaining a favorable microstructure of alloys using new alternative methods of thermal and thermomechanical treatments. For the first time, the cyclic martensitic transformations during repeated quenching, methods of uniaxial megaplastic compression, or torsion under high pressure were successfully applied for radical size refinement of the grain structure of polycrystalline Cu-Al-Ni-based alloys with shape memory. The design of the ultra- and fine-grained structure by different methods determined (i) an unusual combination of strength and plasticity of these initially brittle alloys, both under controlled heat or hot compression or stretching, and during subsequent tensile tests at room temperature, and, as a consequence, (ii) highly reversible shape memory effects.

1. Introduction

For many years, the most widespread multicomponent structural and multifunctional metal steels and alloys that undergo diffusion-free martensitic transformations (MTs) and other related phase transformations (decomposition of supersaturated solid solutions and atomic ordering) have been a powerful fundamental basis for the most diverse areas of the world economy. MTs are deformation-induced phase transitions of the first kind; specific volume changes (ΔV/V), high-strength, plastic, and other performance characteristics are among the strengths of these widely used structural materials based on iron [1,2]. The mechanical properties of steels due to MTs are mainly determined by the high number density of dislocations inside highly disperse crystals of α-martensite, morphologically and orientationally connected by Kurdjumov-Sachs orientation relationships and regularly distributed in the volume of grains [2]. On the contrary, highly reversible thermoelastic martensitic transformations (TMTs), realized in various predominantly non-ferrous alloys, have a packet pairwise-twinned dislocation-free substructure, which, due to the proximity of the crystal lattices of austenite and martensite to each other and much lower values of ΔV/V, is completely reversible [3,4,5,6,7,8]. TMTs are responsible for a number of unusual and extremely important physical phenomena and shape memory effects [9,10,11,12,13,14].

It should be noted when discussing some of the special possibilities for the unusual applications of MTs and TMTs that back in the 1950s, their significant effects of heat release at a forward γ→α MT and, conversely, heat absorption at a reverse α→γ MT were found in steels [1,12]. The asymmetric thermal effects in the alloy Fe–25 at %Ni were measured to be –2.4 kJ/mol and +1.5 kJ/mol, respectively, demonstrating the feasibility of a real embodiment of the principle of a martensitic heat pump. Moreover, the reversible baro-induced polymorphic transformation in RbCl crystals was accompanied by the release of heat, both when pressure was applied and when it was reversibly removed [12]. It has recently been discovered that such unique thermal effects as magnetocaloric [15,16,17], electrocaloric [18,19], elastocaloric [20,21,22,23,24,25], barocaloric [26], which are in great demand in effective innovative heating and refrigeration technologies, with direct or reverse TMTs, are also observed depending on varying external impacts (thermal, magnetic, electrical, force-assisted).

The unique TMT-induced cyclically reversible giant effects of shape memory (SM), giant superelasticity (GS), damping, and many others (of effects) help to classify smart steels and alloys into a separate class of new promising and practically useful structural multifunctional application-related materials [9,10,11,12,13,14,24]. However, only TiNi-based alloys (based on titanium nickelide) are distinguished by a unique combination of the strength- and plastic properties and SM effects and are increasingly being used in engineering and medicine. Previously, it was found that their fine- and ultrafine-grained (FG and UFG) structure leads to improvements in the strength, plastic, and fatigue characteristics of SM alloys [27,28,29,30,31,32]. The special UFG structure of TiNi-based alloys is provided by various thermal deformation technologies using severe megaplastic deformation (SMD) methods, including multi-pass equal-channel angular pressing (ECAP) and high-pressure torsion (HPT) [33,34,35,36,37,38,39,40], rolling, drawing into strips, rods or wire [41]. A similar method of refining the grain structure of titanium alloys is warm abc-pressing [42].

A critical disadvantage of other known polycrystalline smart materials is their low strength, plasticity, and, on the contrary, increased brittleness [4,43]. As a consequence, the implementation of their unique effects in multi-cycle and even single-cycle conditions is impossible, which excludes their practical use. Therefore, plasticization and hardening of various polycrystalline smart materials becomes a key task, the solution of which must be provided through the development of various methods of primary synthesis, the choice of optimal alloying or thermomechanical treatments. In recent years, it has also become increasingly important to reduce the cost of the mass production of these materials in order to involve them in wide and diverse breakthrough industrial development.

Copper β-alloys of Cu-Al-Ni, Cu-Zn-Al, Cu-Zn-Sn and other systems have a much lower cost and better thermal and electrical conductivity and processability, even compared to titanium nickelide alloys [3,4,5,6,7]. Alloys in the mono-crystalline state demonstrate attractive SM effects. At the same time, in an ordinary coarse-grained (CG) state typical of the polycrystalline alloys under consideration, they have an extremely low level of ductility, fracture resistance, and durability, excluding SM effects characteristic of their monocrystals [4,43]. Under conditions of high elastic anisotropy, the production of SM effects via the initiation of TMT is impossible due to the catastrophic brittleness generated as a result of TMT in the alloys. This phenomenon in polycrystalline CG alloys is due to the progressive accumulation of coherent elastic stresses arising in the course of TMT due to an increase in the volume effect value |ΔV/V|, when grain boundaries of the general type become the only significant localization of elastic stresses. The problem in CG copper β-alloys is aggravated not only by the strong softening of the elastic modulus C′ and the growth of the elastic anisotropy A in the pre-martensitic state, but also by grain boundary decomposition, increasing their grain-boundary component, brittleness [43,44,45]. These reasons prevent the commercial use of these aging or eutectoid SM alloys. The development of technologies that ensure the size refinement of grains to increase their strength, plasticity and prevent brittleness is the only acceptable trend taking into account all three critical circumstances (grain size, elastic anisotropy, grain boundary decomposition).

In our previous work [27,46,47,48,49,50,51,52,53], the significant weakening, the embrittlement, of copper SM alloys was found due to a radical decrease in the size of grains at SMD and an increase in the length of the grain boundaries. On the contrary, various other methods employing alloying additives [54,55,56] such as rapid quenching [57], heat treatment [58,59,60], powder metallurgy [61], and a number of other methods turned out to be, as a rule, unsuccessful, and did not lead to noticeable size refinement of the grain structure of these alloys, nor to an improvement in their technological plasticity.

The present work aims to present a comparative systematic study on the structural and phase transformations and physical and mechanical properties of the multicomponent alloys of D0₃–CuAlNi compound systems, depending on high-doping and the development of special thermal and mechanical treatments, including SMD. Using these methods, the microstructure, phase composition, and martensitic transformations in eutectoid alloys of the Cu-Al-Ni system with widely varying concentrations of Al and Ni were studied based on our original investigations. It was found that austenite deformation as well as heat treatment at temperatures above the solubility limit of eutectoid decomposition (840 K) were accompanied by simultaneous partial pro-eutectoid decomposition with the precipitation of γ₂ and α phases below 840 K by their eutectoid precipitation, along with the appearance of ultradisperse B2′ particles (based on Ni–Al–Cu). The cooling of the alloys led to a thermoelastic martensitic transformation with the formation of the β′-and γ′-types of martensite.

2. Materials and Methods

To obtain a more complete picture of the structural-phase transformations and formation of physico-mechanical properties, two groups of master alloys were made from high-purity Cu, Al, Ni (99.99%) systems: Cu–xAl–3Ni (10 ≤ x ≤ 14 wt.%) and Cu–xAl–yNi (7.5 ≤ x ≤ 14 wt.%, 3 ≤ y ≤ 6 wt.%), following the well-known three-component sections of phase equilibrium diagrams presented in Figure 1a,b [4]. The alloys were smelted in an electric furnace using an atmosphere of purified helium. The chemical composition of the alloys was estimated by spectral analysis on a Bruker Q4 Tasman Spectrometer and is shown in

Table 1. Mechanical properties, average grain size <d> and temperatures A_s and A_f of Cu-xAl-3 wt.%Ni alloys.

x, wt.% Al	σu, MPa	σy, MPa	δ, %	<d>, μm	As, K	Af, K	Ms, K	Mf, K
10.0	520	240	5.0	80	890	920	880	850
10.5	500	250	4.0	100	-	-	-	-
11.0	490	260	5.0	130	750	780	740	710
11.5	460	260	4.0	200	-	-	-	-
12.0	450	260	4.0	350	600	630	580	550
12.5	420	240	5.0	500	-	-	-	-
13.0	390	200	5.0	750	440	470	420	400
13.5	330	150	5.0	900	-	-	-	-
14.0	250	120	3.5	1000	260	280	250	230

and

Table 2. Mechanical properties of Cu-xAl -4.5 wt.%Ni alloys.

Treatment	x, wt.% Al	σu, MPa	σy, MPa	δ, %
Quenching from 1223 K, 10 min in water	7.5	760	260	14.0
	8.0	840	320	14.5
	9.0	450	200	3.0
	10.0	480	280	4.0
Quenching from 1223 K, 10 min in water + 473 K, 1 h	8.0	470	-	2.6
Quenching from 1223 K, 10 min in water + 573 K, 1 h	8.0	880	-	2.0
	10.0	180	130	1.5
Quenching from 1323 K, 10 min in water	8.0	300	230	4.5
	10.0	265	140	5.0

. Ingots after homogenizing at (1173 ± 25) K for 8 h were subjected to hot forging into bars with a cross section of 20 mm × 20 mm at (1173–1273) K and final quenching from 1223 K, 10 min in water at room temperature (RT). Some alloys after quenching to martensite were re-quenched from 1273 K for 30 min in water at RT. A number of alloys were subjected to multi-pass rolling or drawing at different temperatures. Torsion under high pressure of 6 GPa (HPT) in Bridgman anvils was performed in high-hard bikes (flat or with a cylindrical recess). The samples for HPT had the shape of disks with a diameter of 10 or 20 mm and a thickness of 0.5 or 1.2 mm, respectively. Uniaxial megaplastic compression (MPC) at various temperatures and strain rates was carried out in an Instron 8862 electromechanical measuring system equipped with an electric furnace for deformation under isothermal conditions at temperatures up to 1073 K, on cylindrical samples with a diameter (d₀) of 7.5 mm and a height (h₀) of 9.2 mm. The prevention of decomposition in alloys after MPC was provided by quenching in water at RT. For mechanical tensile tests, testing machines were used: Instron 5982 (on standard samples with d₀ = 3 mm) and Instron 3545 (on flat samples with a working part of 1 mm × 0.2 mm × 3 mm). The Vickers HV microhardness was measured on a NanoTest-600 device. Structure-phase studies were carried out using optical (OM), scanning electron microscopy (SEM) on Quanta-200 Pegasus and Tescan Mira microscopes (at 30 kV) and transmission electron microscopy (TEM) on a Tecnai G² 30 microscope (at 300 kV), as well as X-Ray diffractometry (XRD) on a diffractometer Brucker D8 Advance in monochromatized Cu Kα radiation. The critical temperatures of the start (M_s, A_s) and finish (M_f, A_f) of the forward and reverse TMTs were determined by the method of tangents on temperature hysteresis curves of the dependences of electrical resistance ρ(T) and magnetic susceptibility χ(T). The unit of reversible deformation ε_m was determined by the method of tangent on curves at tensile tests.

3. Results

3.1. The Structure and Properties of the Quenched Alloys

Previously, it was found that cast and forged copper β-alloys during subsequent cooling experience pro-eutectoid decomposition β→β₁ + γ₂ or β→β₁ + α at temperatures above T_ED close to 840 K, and eutectoid decomposition β₁→α + γ₂ at temperatures below T_ED (Figure 1) [4]. However, their quenching procedure can prevent eutectoid decomposition [46,49,52,62]. In Figure 2a–c, typical examples of (a, b) the grain eutectic microstructure of (a) the alloy in the cast state, (b) the β₁-austenite after quenching, as well as of (c) the intragrain morphology of martensite of quenched alloys are given. According to XRD data, two known martensitic phases were revealed in alloys after quenching, namely: the β’₁ phase (the parameters of its long-period lattice 18R with a small monoclinic distortion are close to a = 0.443 nm, b = 0.533 nm, c = 3.819 nm, β = 89.0–89.5°) and the γ’₁ phase–2H (the parameters of its orthorhombic lattice are close to a = 0.440 nm, b = 0.534 nm, c = 0.424 nm).

Microstructural studies of quenched alloys using the whole complex of structural methods have shown the following. A typical feature of their martensitic structure is mainly the packet morphology of alternating twinned crystals with lamellar or the zigzag wedge-shaped morphology of martensite (Figure 3 and Figure 4). The main crystal-structural characteristics of the packages are the flat boundaries of primary twin-oriented crystals with crystallographic habit planes close to {110}β₁ and orientation relations of the Bain type. According to the joint data on TEM and of selected area electron diffraction (SAED) patterns, the wedge-shaped morphology is more often characteristic of thin-twinned β’₁-martensite, and the lamellar morphology of γ’₁-martensite and all of these phases contain thin secondary nanotwins (Figure 4). The observed microstructure of martensite is generally typical of martensite in poly-and mono-crystalline alloys based on Cu-Al-Ni and Cu-Zn-Al systems and others [44].

Measurements of linear dimensions of grains were carried out in quenched alloys previously subjected to hot forging and subsequent recrystallization annealing for 10 min at 1223 K with quenching in water. Table 1 and Figure 5 show that depending on the chemical composition in alloys of the Cu-xAl-3Ni system, the average grain size <d> changed naturally, increasing from 80 µm to 1000 µm (i.e., by more than an order of magnitude) with an increase in the Al content from 10 to 14% [48]. A more uniform equiaxed FG structure has a positive effect on the strength characteristics of alloys (tensile strength σ_u, yield strength σ_y). However, their relative elongation (δ) remains at a low level, namely, of 3–5%. Figure 5 demonstrates the concentration dependence of the limits of σ_u, σ_y, δ and the critical temperatures of TMTs (according to the measurements of ρ(T) and χ(T) [27]) on the content of Al. It turned out that CG alloys with a high Al content experience brittle intergranular fracture, whereas in FG alloys with the lowest Al content, the fracture mechanism is replaced by the tough-ductile one (Figure 6). In the alloys with intermediate Al content and intermediate grain sizes, it was mixed tough brittle. Figure 7 shows the «σ vs. δ» curves obtained in the tensile tests of a number of alloys of another Cu-xAl-4.5Ni system under study after quenching or isochronous low-temperature annealing at 473 or 573 K. The mechanical properties of the alloys are presented in Table 2.

A detailed study of the relationship between the chemical and the phase composition of alloys by OM and SEM methods found that quenching could not eliminate the chemical liquation observed in Figure 8.

Table 3. Data of chemical X-ray microanalysis of the (A) Cu-13 wt.%Al-3 wt.%Ni after quenching from 1223 K, 10 min in water and (B) Cu-14 wt.%Al-3 wt.%Ni after quenching from 1223 K, 10 min in water and repeated quenching from 1273 K, 30 min in water at RT.

A
Element	Integral Composition, wt.%	Matrix, wt.%	Area 1, wt.%	Area 2, wt.%	Area 3, wt.%
Aluminum	13.0	13.2	10.3	10.1	9.6
Nickel	3.0	2.9	2.7	2.1	4.0
Cupper	84.0	83.9	87.0	87.8	86.4
B
Element	Integral Composition, wt.%	Area 1, wt.%	Area 2, wt.%	Integral Composition, wt.%
	Quenching from 1223 K, 10 min			Repeated quenching from 1273 K, 30 min
Aluminum	13.9	10.9	14.5	14.0
Nickel	3.1	3.9	5.4	3.0
Copper	83.0	85.2	80.1	83.0

, as an example, shows the data of the integral spectral analysis and X-ray energy dispersion microanalysis for Cu-13Al-3Ni alloy. It can be assumed that the formation of the FG structure of the austenite alloys during recrystallization after forging was stipulated by the barrier effect of grain growth inhibition due to the detected chemical heterogeneity of liquation origin. This had a weakening effect, as the change in the concentration of Al shifted the composition of the alloys from the boarder that separated the phase states (α + β) and β or β and (γ₂ + β) (see Figure 1). However, in this case, in alloys with a more homogeneous chemical composition, the growth of austenite grains was more noticeable during annealing (see Table 1) and, as a result, CG alloys experienced brittle fracture during tensile tests, whereas FG alloys containing alloying elements near the solubility boundary were subjected to partial pro-eutectoid decomposition during quenching, mainly at grain boundaries, which also led to their embrittlement (Figure 5).

To eliminate structural-phase and chemical heterogeneity and obtain a homogeneous solid solution in austenite, reheating of the martensitic alloy was carried out at a temperature of 1273 K for 30 min, followed by re-quenching in water. Figure 9 shows images of the microstructure of the alloy after this quenching procedure. In this case, the chemical analysis showed that the heat treatment effectively ensured the chemical and phase uniformity of the alloy (compare Table 3A,B). However, most importantly, noticeable size refinement of the grain structure was detected. The decrease in the average grain size <d> from 1000 to 250 µm is obviously related to the recrystallization effect in an initially inhomogeneous martensitic alloy due to the inverse TMT, which is well known for steels [2]. This FG alloy showed good deformability during cold multi-pass rolling or drawing and, after certain treatment, showed the ability to reverse the effect of SM (see the example of the implementation of SM in Figure 10).

As is known, β-austenite of alloys at temperatures higher than T_ED and M_s have the time management to experience two consecutive phase transitions «disorder–order» (β→β₂(B2)→β₁(D0₃)) [3,44]. As a result of the multi-type nucleation mechanism, a special substructure of anti-phase (AP) domains is formed, the boundaries of which (APB) are clearly visible on dark-field TEM images in superstructural reflections (Figure 11). The long-range atomic order of the initial austenitic phase is inherited by the structure of martensite, determining its important role in the realization of the effects of orientational crystal-structural reversibility and phase thermoelasticity at a TMT [3,4,5]. APBs had predominant crystallographic faceting along the planes of {100} and {110} of cuboid domains (Figure 11). During the TEM study, tweed striped diffraction contrast was also observed in the light- and dark-field austenite images (Figure 11). As the analysis has shown, the tweedy pattern is oriented along various crystallographic directions, which are mainly intersections of the planes {110} with the foil surface [52]. This tweedy pattern (TP) has a fine structure formed by equiaxial and cubic elements of homogeneous contrast alternating according to the type of quasi-periodic 3D-modulated FCC structure. At the TP background, one can see a contrast from the dislocations and APBs shown above. At the same time, the contrast and dimensions of the tweed elements on the APBs shown above indicate their preferred heterogeneous localization along these coherent superstructural boundaries of subgrains. A similar increase in tweed contrast was also noted in images of dislocations, both belonging to the inclined and horizontally disposed planes [52].

A complex pattern of diffuse scattering was detected on SAED patterns along with Bragg reflections (Figure 11 and Figure 12). In the reciprocal lattice (r.l.–*), this diffuse scattering forms the flat diffuse layers {111}* passing through all hkl nodes, with the exception of the central node 000 [52,63]. The intensity of this diffuse scattering is more pronounced near the reflections than between them. The most intense diffuse streaks are located along the non-radial directions <110>*. Additional evidence of the increased intensity of the streaks <110>* is their visualization in the form of diffuse spots, when the streaks are located in r.l. at an angle to the diffraction plane and “pierce” the Ewald sphere. These «puncture holes» are due to the presence of the inclined streaks along <110>*, both in the vicinity of Bragg reflections and in between them (e.g., 1/2<112>* in the r.l. sections {110}*, {311}* or {210}*). Such a scattering was already observed to be higher at the temperatures of the observation in situ: M_s by 100–150 K (Figure 11d and Figure 12a).

As we approach M_s, the intensity of the streaks along <110>* increases and, most importantly, according to qualitative and quantitative assessment, the intensity of the extra-reflections (or satellites) in the positions of r.l., close to 1/2<110>*, 1/3<110>* and 1/6<110>* (Figure 11 and Figure 13) should exhibit intensification. The same is valid for the streaks along <112>*, namely, the intensity of the extra-reflections (or satellites) in the positions of r.l., close to 1/2 <112>* and 1/3<112>* must also display an increase. At the satellites-forming stage of the long-period modulation of the crystal lattice, several lattice waves of atomic displacements are distinguished, described by wave vectors ± k, with their polarization vectors e_k. In accordance with the experimental data, the following modes of atomic displacements have been established: (i) transverse wave with k = 1/2<110>*, its e_k is parallel to <1$\overline{1}$0>; (ii) transverse wave with k₁ = 1/6<110>*, its e_k is parallel to <1$\overline{1}$0>; (iii) transverse wave with k₂ = 1/3<110>*, its e_k is parallel to <1$\overline{1}$0> [52]. A transverse wave with k = 1/3<112>* is a superposition of two waves: the first one with k equal to <1/3 1/3 0>* and the second one with k equal to <1/3 0 1/3>* (generating satellites of the type 1/3<211>* in the diffraction pattern).

Thus, the typical features of the observed deformation-induced tweed contrast and diffuse electron scattering make it possible to interpret them as localized transverse and longitudinal short- and long-wave atomic displacements that periodically distort the original crystal lattice. Figure 14 shows the wave spectra of atomic displacements in the k space (in r.l.). The crystallographic analysis of the obtained data makes it possible to construct a physical model of a real microstructure and its evolution, and also establishes its important role in the mechanism of TMT initiation in Cu-Al-Ni alloys [52]. So, in alloys with a BCC lattice, the scattering effects between reflections in the form of the flat diffuse layers {111}* is caused by short-wave acoustic, uncorrelated displacements of the tightly packed (along <111>) chains of atoms relative to each other, prevailing in the oscillation spectrum. With the pre-martensitic softening of elastic moduli, especially C’ [8,45], the amplitudes and correlations of such peculiar dynamic linear defects of atomic displacements gradually increase primarily in closely packed planes {110}. If the correlations of the displacements of atoms in these planes are higher than that of the planes, relative to each other, diffuse scattering has the form of continuous streaks along <110>*. It follows from the analysis of tweed contrast that atomic displacements are localized in nano areas or nanodomains. The analysis of diffuse scattering and its extinction makes it possible to determine short- and long-wave (in the limit homogeneous) displacements of atoms, describing the short-range order of atomic displacements (SOD) by the wave spectrum vectors k and e_k in these nanodomains with distorted structure and symmetry (see Figure 14) [8,52]. Here, attention is drawn to the presence of longitudinal displacement waves of the types <100>*, <100>*, correlating with the longitudinal Bain distortion.

When weakly incommensurable satellites appear on the diffraction pattern below a certain T_IS temperature during cooling of the alloys, a new stage of intra-phase transformation of the austenite structure begins, designated as a quasi-static or static stage of weakly incommensurable satellites. These are mainly satellites of types “1/6”, “1/3”, and “1/2”, which correspond to long-period modulated shear nanostructures: PSS-I (for satellites of types “1/6” and “1/3”) and PSS-II (for satellites of type “1/2”). The satellite-forming stage as an independent state replaces the SOD. Figure 15 schematically shows atomic displacements describing ordered intermediate substructures of shear. Since all crystallographically equivalent variants of distortion nanodomains (orientation- and anti-phase ones) are located statistically in the volume of the austenitic phase, the structure of such metastable alloys, on average, retains the original cubic symmetry, macroscopically being an intra-phase state according to XRD data.

According to SAED data, the internal distortion and local symmetry of the PSS-nanodomains differ from the initial symmetry of austenite and, obviously, approach the structure of future martensitic phases while maintaining coherent matching of crystal lattices under the conditions of the specifics of the progressive local instability of the austenite phase lattice and its anharmonicity [8,52]. Experimentally, this structural mechanism of TMT, firstly, is confirmed by the fact that the reflections of the formed martensitic crystals of phases β_1′ and γ_1′ on SAED patterns are located virtually in place of satellites of types 1/3 and 1/2 (see Figure 11 and Figure 12). Secondly, during their nucleation and growth, a large number of planar chaotically arranged stacking faults (SF) appear parallel to the basic plane of the type (001) for both β_1′ and γ_1′ martensitic phases. The appearance of SFs is due to the multi-nucleation mechanism of crystal formation from the PSS-I and PSS-II nanodomains (see Figure 16). In this case, characteristic features of the contrast effect and diffuse scattering are observed in the form of sharp solid strokes through martensitic Bragg reflections (Figure 12).

In the crystallostructural sense, PSS domains are special “non-classical” nano nuclei (with a structure that is not identical to the structure of both austenitic and future martensitic phases) and, obviously, play the role of real physical centers of the nucleation of martensite crystals. Then, at a certain synchronization of homogeneous distortion of Bain type and transverse waves of static displacements of atoms that describe the structure of nanodomains via PSS-I waves of the types 1/6 and 1/3<110>_k<1$\overline{1}$0>_e, including the possible superposition, restructuring can be implemented into the actual structure of the martensite β_1′ (18R) in the studied alloys (see Figure 16a). Additionally, when combining the homogeneous Bain distortion and the mode of periodic reshuffling displacements of the type of doubling 1/2<110>_k<1$\overline{1}$0>_e, which form the structure of the PSS-II nanodomains, we have a crystallographic preset of the trend for restructuring phase β₁→(into) phase γ_1′ (2H) (Figure 16b). Local violations of the ideal packing of atoms along the basic plane, obviously, give rise to its various “failures” in the form of SFs. In conclusion, it should be noted that measurements of the temperature dependences ρ (T) and χ(T) in the course of the thermal cycles “cooling from RT to 90 K→heating to RT” (

Table 4. Critical temperatures of the start (M_s, A_s) and finish (M_f, A_f) of the TMT in the alloys Cu–14Al–3Ni and Cu–14Al–4Ni after different treatments.

	Treatment	Ms, K	Mf, K	As, K	Af, K	ΔT, K
Cu–14Al–3Ni	Quenching from 1223 K (ρ(T))	250	230	265	280	33
	Quenching from 1223 K (χ(T))	255	240	265	280	25
	HPT of 10 revs (1)	320	300	400	440	110
	HPT of 10 revs (2)	-	-	380	470	-
Cu-14Al-4Ni	Quenching	270	250	275	290	22
	MPC 973 K	295	285	300	310	15
	MPC 773 K	305	290	300	315	10

) not only revealed TMT hysteresis loops, but also pre-martensitic deviations from the linearity of the mentioned dependences, in anticipation of TMTs, both forward and reverse in the range of 10–20 K (Figure 17a).

In addition, as was already noted, the copper alloys studied in this work, which are metastable with respect to TMTs, are distinguished by a low magnitude of the modulus C’ and, accordingly, a high anisotropy of elastic modules A = C₄₄/C’ (of 12–13 units) [45], whereas, for example, for elastoisotropic low-modulus plastic alloys of titanium nickelide, the value of A is only 1–2 units [8,27]. An abnormally large elastic anisotropy causes the dominance—in the pre-martensitic state—of the localized SOD and PSS nanodomains of the type <110>_k, <1$\overline{1}$0>_e and, accordingly, the specific quasi-periodic 3D-FCC morphology of nanodomains visualized by deformation tweed contrast typical of them, as well as the crystallography of shear transformation at the TMT by a single channel of atomic displacements of the re-shuffling type {110}<1$\overline{1}$0>. Possible variations in long-period shear modulation of the “tripling” and “doubling” types correspond to two variants of martensitic transformation into the phases β_1′ (18R) and γ_1′ (2H).

3.2. An Effect of Isothermal Mega-Plastic Compression

Systematic studies of the effect of isothermal MPC at various temperatures were performed on the CG (<d> close to 1 mm) Cu-14Al-4Ni alloy in the initial quenched state [47,53]. The results of MPC tests at RT showed that the alloy experiences a sufficiently large plastic deformation to fracture, close to ε = 22%, at high values of the stresses σ_y and σ_u, close to 400 and 1150 MPa, respectively (Figure 18a). The curve «σ vs. δ» obtained on this alloy with uniaxial tension at RT is shown in the insert to Figure 18a. A comparison of the obtained data showed that at close values of the coefficient of deformation hardening θ = dσ/dε (~3.5 and 4.5 GPa, respectively, under compression and tension), the relative elongation of the alloy before fracture under tension (δ = 4%) is more than 5 times less than the value characteristic of relative compression before fracture (ε ≈ 22%).

Figure 19a,b shows a SEM image of the martensitic microstructure of the alloy after tensile and compression tests at RT. As can be seen after MPC compared to the tense test, the morphology has changed and the crystals of martensite in the dominant packet morphology have been significantly refined in size. Virtually no large arc-shaped joints were detected, and the sizes of both individual martensite crystals and their packets significantly decreased. Figure 20a–f shows typical images of the fracture surface obtained by fractographic analysis of samples after tensile or compression tests. The brittle and the tough-brittle, both intergranular and intragranular, fracture of the alloy in the martensitic state after tensile tests at RT occurred mainly along the boundaries of packets of twin-oriented martensite crystals (Figure 20a,b). The fracture during MPC of the alloy with the formation of small-dimple fracture sites indicates the predominance of a tough-ductile fracture mechanism at RT (Figure 20c,d).

Mechanical compression tests at elevated temperatures of 673, 773 and 873 K showed that the Cu-14Al-4Ni alloy is capable of experiencing large plastic deformation without fracture just up to high values of σ_u, and reached 1600–2000 MPa (Figure 18b–d). On the curves «σ–ε», the stages of the elastic deformation are visible; easy, steady uniform deformation, which differ depending on the temperature and velocity in the magnitude of σ_y and hardening coefficients (θ₁ = dσ₁/dε); a transitional stage of rapidly growing strain hardening; and finally, the stage of strong hardening (θ₂ = dσ₁/dε) when ending deformation.

A comparison of the data shown in Figure 18 b–d obtained at different compression rates v (0.5; 1; 5 mm/min) showed that the values of the coefficient of deformation hardening θ₂ and the accumulated relative compression ε (or, respectively, its true logarithmic value e = ln(h ₀/h_τ)) are close enough, and MPC with a higher rate virtually does not lead to higher strength characteristics of σ_u (

Table 5. Mechanical characteristics of the Cu-14Al-4Ni alloy after MPC at various temperatures and rates (v).

Treatment	σy, MPa	σu, MPa	ε, %	θ1, GPa	θ2, GPa
RT, v = 1 mm/min	400	1150	22	3.5	-
673 K, v = 0.5 mm/min	360	1550	82	0.7	5.3
673 K, v = 1 mm/min	380	1550	84	0.7	5.7
673 K, v = 5 mm/min	530	1580	76	0.2	4.6
773 K, v = 1 mm/min	250	1550	84	0.3	6.3
773 K, v = 5 mm/min	310	1620	83	0.3	5.9
873 K, v = 1 mm/min	70	1550	95	0.2	8.0
873 K, v = 5 mm/min	120	1550	92	0.1	8.1
973 K, v = 1 mm/min	50	2000	70	0.1	10.5
1073 K, v = 1 mm/min	50	2000	55	0.1	11.0

). A stronger influence of the deformation rate and temperature on σ_y was found. Thus, an increase in v from 1 mm/min to 5 mm/min led to an increase in the value of σ_y from 380 to 530 MPa at a deformation temperature of 673 K, from 250 to 310 MPa at 773 K, and from 70 to 120 MPa at 873 K. This indicates the predominance of structural-deformation hardening processes with an increase in the rate and a decrease in the deformation temperature compared with compensating softening processes, on the contrary, progressing under compression with an increase in temperature, a decrease in the deformation rate, and, accordingly, an increase in the holding time τ. Together with an increase in the value of σ_u, the plasticity ε of the alloy decreased slightly with an increase in the rate and a decrease in the deformation temperature (Table 5). Higher strength characteristics of the alloy were found during tests of mechanical properties for compression in the austenitic state at 973 and 1073 K (Figure 18d, Table 5). It is obvious that eutectoid decomposition in the Cu-14Al-4Ni alloy will occur during mechanical compression at elevated temperatures (673–873) K in accordance with Figure 1. According to the XRD data, two martensitic phases (β’, type 18R and γ’, type 2H) were present in the MPC alloy, as well as precipitations of γ₂ phase after compression at 873–1073 K, and possibly also the α phase after MPC at 673 and 773 K (Figure 21).

According to SEM data, the most typical examples of the FG structure of an alloy with grains of 100–200 μm in size, an order of magnitude smaller than the grain sizes in the original CG alloy, convincingly prove the effect of radical size refinement of the grain structure as a result of their dynamic recrystallization at MPC (Figure 22). At the same time, after MPC at 673, 773, 873 K, the grains contain much smaller crystallites of 1–2 μm in size, identified as the α and γ₂ phases (Figure 23 and Figure 24). Attention is drawn to the fact that at MPC, the predominant pro-eutectoid heterogeneous decomposition of the phase γ₂ along the boundaries of austenitic grains preceded the homogeneous intragrain precipitation of the phases γ₂ and α.

Figure 25 shows TEM images of the fine structure of the alloy after MPC at 773 (a–d), 873 (e,g), 973 (f), and 673 K (h). As it follows from their analysis, the crystallite–grains of the α phase during precipitation experienced noticeable plastic deformation with the formation of mesh—cellular dislocation—and twin substructures (see Figure 25a,b), in contrast to the more brittle, solid, and often twinned γ₂ phase (see Figure 25c,d). In addition, B2′ phase precipitations based on the Ni-Al system were observed, which had much smaller dimensions, not exceeding 100 nm, rounded shape and had formed in the initial austenitic matrix, at the boundaries and inside the crystallites. According to energy dispersive microanalysis, α crystals were somewhat depleted of aluminum (after offsetting at 673 and 773 K to 10% by weight), and γ₂ precipitations along with copper just contained up to 5 at.% Ni, while B2′ precipitations, along with nickel and aluminum, contained up to 5 at.% Cu. Figure 25e–h demonstrate TEM images of the martensite structure after MPC at (e) 873 and (f) 973 K and (g,h) direct resolution of β’-martensite (g, after MPC at 873 K) and two B2′ particles (h, after MPC at 673 K).

Thus, firstly, it was found that a dynamic recrystallization process took place in the alloy at hot MPC, as a result of which an austenite FG structure homogeneous in grain size was formed. This FG structure was significantly (by an order of magnitude) more disperse than in the original prototype alloy. Secondly, during compression at 873 K and in the less extent at 973 and 1073 K, partial pro-eutectoid decomposition was induced in austenite with heterogeneous and homogeneous precipitation of the γ₂ phase, the size of which increased with increasing compression duration, whereas at 673 and 773 K (i.e., below T_ED), along with the γ₂ phase, the α and B2′ phase particles were precipitated. Their boundaries as the places of predominant heterogeneous nucleation subsequently turned out to be decorated with the phases that underwent precipitation, since the size refinement of austenitic grains occurred as a result of dynamic recrystallization, which preceded the decomposition. As is known, the barrier effect of grain boundary precipitations can restrain the growth of grains with continuing hot deformation. Austenite after MPC, when cooled to RT, experienced a TMT with the formation of the β’ and γ’ martensitic phases of packet morphology.

Temperature measurements of (T) showed that the critical temperatures (T_cr) of the TMT of the MPC Cu-14Al-4Ni alloy increase compared to the temperatures of the quenched prototype alloy (Figure 17c–e). They (T_cr) become higher than RT obviously, due to some depletion of the alloy matrix of aluminum due to the ongoing partial decomposition.

Figure 26a presents the results of subsequent tests of the mechanical properties of the alloy already in the metastable austenitic state during tensile tests at RT. The properties of the FG alloy samples were measured after MPC at 973 and 1073 K to compare them with the properties of the CG alloy in its initial state after forging at 1223 K and quenching (see insert in Figure 18a). A unique feature of the mechanical behavior of this FG alloy under tension after compression was the presence of a stage of martensitic inelastic pseudo-yield (ε_m 4%) at a low starting stress of the reorientation of the twin martensitic structure in the direction of the acting tensile force (σ_m < 100 MPa). Then, a strong deformation hardening occurred, culminating in the fracture of the samples at high values of σ_u after significant plastic deformation for these alloys (δ_f = 14–16%) (see Figure 26a). The curve in the insert in Figure 18a illustrates, on the contrary, the low strength and plastic properties of a conventional hardened CG alloy.

Figure 20c,d shows the fractography of fracture «surfaces» after testing this FG alloy for tensile strength at RT. Despite the sufficiently large plastic deformation of martensite and its uniform elongation (14–16%), it can be concluded from the appearance of the fracture surfaces that the fracture occurred by mixed tough-ductile and tough-brittle mechanisms. On the surface of the fractures of the tensile-test samples after compression, there were cup-shaped fine-dimpled zones of tough-ductile fracture, flatter chips with a «river»: internal pattern and individual cracks very likely along the boundaries of both grains and disperse martensite crystals. However, in this case, the linear dimensions of the flat elements of the fracture surfaces are (5–10) μm, which is an order of magnitude smaller than the grain sizes (100–200 μm). This characterizes the nature of the subgrain structural elements responsible for the development of the mechanism of tough-brittle intragrain fracture. It is obvious that the fracture could occur both by a quasi-cleavage along the boundaries of the interface of disperse martensite packets and via a tough-ductile mode across the crystals mainly oriented along the axis of tensile direction.

3.3. The Effect of the Intense Plastic HPT Deformation and Annealing

This section is devoted to the use of SMD alloys for the synthesis of UFG structure [46,52]. Figure 17b demonstrates the temperature dependences ρ(T) for the Cu-14Al-3Ni HPT alloy when measured in the thermal cycles «cooling–heating» (curve 1) or «heating-cooling» (2) of the TMT hysteresis loop. The critical temperatures, determined, as was already noted, by the method of two tangents, are given in Table 4. It is important to note that the hysteresis of TMT (ΔT) in the HPT alloy increases more than threefold with a noticeable increase in all critical temperatures.

It has already been noted that in the studied quenched β₁-alloys Cu–14Al–3Ni and Cu–13.5Al–3.5Ni, when cooled below the temperature M_s, two martensitic phases appear, designated β_1′ (of type 18R) and γ_1′ (of type 2H) (Figure 27, curve a). It was found that HPT at 10 revolutions at RT causes a mixture of three deformation-induced martensitic phases, α’, β_1′, and γ_1′ (Figure 27, curve b), to appear. The detected Bragg reflections are significantly broadened (with a half-width of 2θ to 2 deg) and coincide with the strongest lines of these martensitic phases. According to XRD data, the HPT alloy annealed at 373 and 473 K preserves the martensitic phases β_1′ and γ_1′ and the γ₂ phase appears (see Figure 27, curves c, d). Annealing at temperatures of 573–773 K (which are higher than A_f) leads to the eutectoid decomposition of already β₁-austenite into the phases (α + γ₂) (see Figure 27, curves e–g). Finally, annealing at 873 K causes the decomposition of austenite with the precipitation of the γ₂ phase (Figure 27, curve h). Subsequent cooling of the HPT alloy to RT is accompanied by the TMT in the residual β₁-matrix; therefore, its reflections, unlike martensitic ones, are not detected by XRD analysis.

Analyses of the light- and dark-field TEM images of the alloys after HPT at 10 revolutions showed that the sizes of the observed arbitrarily oriented nanograins vary from 10 to 80 nm and were an average of 30 nm (see Figure 28). It can be seen that lamellar nanotwins are located in the larger ones. Decoding of SAED patterns showed that the nanocrystalline structure formed in alloys mainly contains three (α’ + β_1′ + γ_1′) martensitic phases. The ring-wise distribution of reflections indicates the presence of both small- and large-angle misorientations of martensitic nanophases that make up the UFG structure. It is established that an increasingly homogeneous nano-grain martensitic structure in both of the Cu-Al-Ni alloys, characterized by a ring-wise distribution of reflections in SAED patterns, is formed as a result of HPT, with an increase in the number of revolutions from 1 to 10, and, accordingly, the degree of deformation.

Annealing martensite at 373 and 473 K according to SEM data did not lead to noticeable dimensional and morphological changes in the martensitic UFG structure formed as a result of HPT, but a network arrangement of larger (up to 100 nm) nanograins became more noticeable (Figure 29a). Annealing austenite at temperatures ranging from 573 to 873 K provided a noticeable growth of globular grains during the formation of UFG states in HPT alloys (Figure 29b–f). The EBSD analysis made it possible to establish a special mechanism for the decomposition of HPT alloys, as a result of which a UFG structure of alternating nanophases of the type «nanoduplex» (martensite + γ₂ phase) or «nanotriplex» (martensite + α + γ₂ phase) was formed (see, for example, Figure 29e,f).

According to TEM data (Figure 30), annealing just forms globular nanocrystallites in the UFG β₁-matrix, mainly of two phases α and γ₂ according to SAED patterns (<d> of grain crystallites are given in Table 3). In the reverted β₁-austenite, the annealing at 573 K triggers a complex reaction of the recrystallization and eutectoid (α + γ₂) decomposition [52]. At the same time, against the background of larger recrystallized grains, smaller crystallite grains are still distinguished by a strong deformation contrast, and they remain almost unchanged in size; their reflections still retain a continuous ring-wise distribution; <d> is close to 100 nm (see Figure 30a).

An increase in the annealing temperature to 673 K led to almost complete primary recrystallization of the reverted β₁-austenite and the formation of a (β₁ + α + γ₂) nanotriplex structure in the composition of homogeneous UFG austenite at an increased <d> close to 150 nm. The boundaries of the crystallites–grains remained rounded and distorted, and an increased number density of defects was still preserved in the structure (Figure 30b). Intensive enlargement in size of the globules, formed in β₁-austenite, of the α and γ₂ phases, with <d> close to 350 nm, took place at a higher isochronous annealing temperature of 773 K (Figure 30c). After annealing at 873 K, <d> for the phases β₁ and γ₂ was already close to 400 nm (Figure 30d). The images of the boundaries inherited by martensite after those of grains–crystallites of the nanoduplex structure (β₁ + γ₂) became even more pronounced and the contrast from the twins in martensite was visible. More “sharp” point reflections in SAED patterns indicate the significant relaxation of the internal stresses and interfacial distortions in the UFG structure formed by the precipitated phases and martensite. The «ring-wise» character of the already less uniform distribution of point reflections in the SAED patterns from the HPT alloys after annealing at 673–873 K indicates the preservation of the small-angle and large-angle random misorientations of the globular grains–crystallites formed in the alloy (see insert in Figure 30a–d). Figure 31 shows another type of bimodal single-phase structure—the UFG structure of eutectoid HPT alloys of the Cu-Al-Ni system—formed due to short-term thermal exposure during hot annealing of 1073 K for 10 s.

Thus, it has been established that annealing, starting from temperatures above A_f, leads to recrystallization of reverted austenite in the HPT alloys Cu–14Al–3Ni and Cu–13.5Al–3.5Ni. At the same time, sufficiently homogeneous UFG states are obtained. The average grain sizes increase after annealing at 573–873 K (30 min) in the range from 30 to 400 nm or reach 3.5 µm after short-term annealing at 1073 K, 10 s. Reflections from the γ₂ phase were present in the obtained XRD and SAED patterns in all the cases. At low-temperature (below M_s) annealing at 373–473 K, the HPT alloy experienced partial decomposition of martensite with the precipitation of disperse γ₂-particles, whereas austenite under high-temperature annealing experienced the eutectoid (α + γ₂) decomposition at 573–773 K or pro-eutectoid (β₁ + γ₂) decomposition at 873 K. If the matrix austenitic β₁ phase was preserved after annealing, then a TMT occurred during subsequent cooling to RT.

The data of HV microhardness measurements at half the radius of the disks of HPT alloys Cu-14 Al–3Ni and Cu–13.5Al–3.5 Ni, depending on quenching and annealing temperature, are shown in Figure 32. It was found that Vickers microhardness (VH) increases after HPT, as the annealing temperature increases, although after double quenching it remains somewhat less than after single quenching. After annealing at 573–673 K, the Cu-Al-Ni HPT alloys had maximum VH values: up to 6000 MPa after single quenching from 1223 K and 5550 MPa after repeated quenching from 1273 K. HPT increases VH by 1000 MPa compared to the VH of the initial hardened austenitic state. Repeated quenching provided size refinement of grains and better homogenization of the alloy solid solution than single quenching and, as a result, slightly lower VH values.

Table 6. Tensile test results of the Cu–14Al–3Ni alloy after different deformation and heat treatments.

Treatment	σm, MPa	σu, MPa	εm, %	δ, %
Quenching from 1223 K	160	620	2	7
Quenching from 1273 K	60	400	2	11
HPT, 10 revs, (293 K)	-	820	-	4
HPT, 10 revs, (423 K)	-	1450	2	12
HPT 10 revs + 573 K, 30 min	120	450	2	6
HPT 10 revs + 773 K, 30 min	50	320	3	8
HPT 10 revs + 1073 K, 10 s	250	900	5	13

and Figure 26b present the results of measurements of the mechanical tensile properties of the Cu-14Al-3Ni alloy at RT. It turned out that the quenched CG alloy had a σ_u of 620 MPa, a critical stress of martensitic shear (σ_m) of 160 MPa, and a value of δ = 7%. The repeated quenching of the alloy due to the creation of the FG state led to an increase in δ up to 11%. In the hardened UFG alloy subjected to HPT at RT, the value of δ decreased to 4%, and the fracture occurred by brittle mode without the formation of a neck. The phase ductility was not fixed in this case, unlike the hardened CG or FG alloy (where ε_m = 2%). An increase in the temperature of the HPT of 10 turns to 423 K (by 130 K higher than RT) led to unusually high deformation hardening of the Cu-14Al-3Ni alloy and to a significant increase in δ by 12%. So, in the HPT alloy, the highest mechanical characteristics were achieved in the combined complex: for example, σ_y = 1400 MPa and σ_u = 1450 MPa, at a sufficiently high value of δ (12%). Significant changes in the mechanical properties of the alloy were observed after HPT at 10 revolutions and at varying annealing modes in the temperature range from 573 to 1073 K with different holding times (from 10 s to 30 min), leading to the creation of the UFG structure (see Table 6).

As was already noted, the fracture of the initial hot-forged quenched CG alloy occurred in brittle mode, mainly by cleavage along the grain boundaries and large packets of martensitic crystals (see Figure 6a). The nano-scale grain–subgrain structure after an HPT of 10 turns changed the type of fracture and the nature of the fracture of the samples [46]. Fractography has shown many centers of localization of deformation with the formation of small flat dimples and, accordingly, low ridges of separation, which is characteristic of the tough-ductile intragrain fracture mechanism operating alloys. However, the average size of the dimples was 2–5 µm, which was two orders of magnitude larger than the size of the elements of the UFG structure of the HPT alloys. This circumstance indicates a special inter-crystalline mechanism of their fracture, obviously along the large-angle boundaries of the UFG structure. At the same time, the size of cellular fragments (dimples) of separation during tough-brittle fracture became two orders of magnitude smaller compared to the size of grains and areas of brittle cleavage in the prototype CG alloy. This ultimately determined, in a number of cases, the increased ductility of the UFG alloys previously subjected to HPT.

Tensile tests of the HPT alloy at two rates v, (a) 10⁻³ 1/s and (b) 10⁻⁴ 1/s, were performed at temperatures in a wide range of 300–673 K below T_ED (Figure 33,

Table 7. Tensile test results at elevated temperatures for the UFG Cu–14Al–3Ni alloy after HPT to 10 revolutions.

№	Ttest, K	v, 1/s	σu, MPa	σy, MPa	dσ/dε, GPa	σm, MPa	δs, %	δl, %	δΣ, %	εm, %
1	300	10−3	820	-	20	-	5	-	5	-
2	423	10−3	1400	900	12.8	-	6	-	6	-
3	573	10−3	1100	800	5.3	400	9	11	20	1.5
4	673	10−3	560	400	3.2	200	6	14	20	1.5
5	423	10−4	280	200	4.0	100	6	-	6	1.5
6	473	10−4	360	800	2.7	200	10	-	10	2
7	573	10−4	600	400	1.7	-	8	11	19	-
8	673	10−4	200	100	1.6	-	10	17	27	-

) [64]. The mechanical behavior of the UFG alloy in the martensitic state (curves: 1—at 300 K; 2, 5—at 423 K; 6—at 473 K) and in the austenitic state (curves: 3, 7—at 573 K and 4, 8—at 673 K) differ radically. The «σ–ε» curves for the austenitic UFG alloy at elevated temperatures of 573 and 673 K are characterized by the presence of several typical stages of deformation: elastic (I), pseudoelastic with a section or plateau of phase yield (II, on curves 3, 4), uniform (III, up to the limit of σ_u), and two localized (IV and V): initially with gradual and then with accelerated softening in the emerging wedge-shaped region of the “neck” of the tensile-test samples. The «σ–ε» curves of the UFG alloy in the martensitic state at lower test temperatures (300–473 K) were distinguished, firstly by the dσ/dε hardening that progresses as the degree of deformation increases, which determines the unusual appearance of their elastic and pseudoelastic stages (I and II). Then, above σ_y value, there follows stage III of uniform deformation. Another striking difference is the absence of stages (IV, V) of localized deformation, which is obviously due to a different mechanism of plastic deformation and the absence of a softening effect in the martensitic alloy.

As is known, the phase yield stage, observed on the «σ–ε» curves at a lowered rate v, the beginning of which is identified by the stress σ_m, is due to the shear reorientation of martensitic crystals in the direction of the acting force at temperatures below A_f. At slightly higher temperatures, the phase yield is mainly determined by the development of the deformation-induced TMT mechanism itself (above critical temperatures A_s up to temperature of deformation induced TMT, M_d). So, the UFG alloy Cu–14 Al–3Ni at RT and 423 K is deformed with a high coefficient of deformation hardening at moderate uniform deformation and experiences fracture before the start of the process of localization of plastic deformation. With an increase in the deformation temperature to 573 and 673 K, on the contrary, the «tensile» curves take on the appearance usual for steels and alloys (Figure 33). At elevated temperatures, the magnitude of the concentrated deformation δ_l increases and the magnitude of the uniform plastic deformation δ_s of the UFG alloy Cu–14Al–3Ni decreases. An effect of the phase yield observed during tensile tests at 423 and 473 K of Cu-14Al-3Ni alloy is absent at elevated deformation temperatures of 573 and 673 K in the case of tests with a deformation rate of 10⁻⁴ 1/s, since the deformation takes place in stable austenite (above the temperature of A_f and the M_d; Figure 33b, curves 7, 8) and is not capable of inducing any TMT (unlike curves 5 and 6 in Figure 33). However, the effect of the phase yield appears in the case of increasing v by an order of magnitude, i.e., up to 10⁻³ 1/s (curves 3, 4 in Figure 33a). At the same time, ε_m is 1–2%.

It is important to emphasize that, in the process of tensile tests at a temperature of 573 K, in the UFG alloy, a high plasticity is achieved at a sufficiently high σ_y and at a noticeable decrease in stress at the stage of the localization of deformation δ_l. Perhaps this is due to the implementation in the UFG alloy Cu–14Al–3Ni of the grain boundary slip mechanism. However, the alloy did not exhibit the usual superplastic behavior at the selected temperatures and strain rates, apparently due to the noticeable dynamic grain growth at elevated test temperatures. As is known, grain boundary mutual slip, which is a necessary condition for superplasticity, becomes more difficult with increasing grain size. A decrease in v from 10⁻³ 1/s to 10⁻⁴ 1/s leads to a decrease in σ_u and σ_y by more than half, while maintaining plasticity at the same level. This can be explained both by the effect of a decrease in the deformation rate and by a longer duration of tests at elevated temperatures, which, accordingly, leads to a greater grain growth rate.

Fractographic analysis has shown that at elevated deformation temperatures, the character of the fracture according to deformation and microstructural features is «tough-ductile», with a high dispersity of the dimples (or cups) of separation (Figure 34b,d,f,g). The fracture occurs with the formation of a small-dimpled edged-blade fracture surface in all structural states. Deep equiaxial and oblong dimples are visible on the fractures in the areas of separation. The lateral surface of the samples after tensile tests in the area of localization of deformation contains uniformly distributed slip (Luder’s) bands (Figure 34a,c,e,h). Thus, it can be concluded that the character (or, precisely, the type or mode) of the fracture (FC) in the studied alloys depends on their structural state: it (FC) is «brittle and intergranular» in the CG state, «tough-brittle» in the martensitic UFG state, and becomes «tough-ductile (fine-dimpled)» in the austenitic UFG state at elevated deformation temperatures. However, the size of the dimples (or cups) was several micrometers, an order of magnitude larger than the average size of ultrafine alloy grains. This circumstance, again, indicates a special inter-grain (and not intragrain) type of tough-ductile fracture occurrence, obviously, along the high-angle boundaries of the UFG structure.

In conclusion, it should be noted that the established regularities of controlling the structure, phase transformations, and properties of the studied ternary alloys can be extended onto multicomponent alloys of these systems; additional alloying of these ternary alloys by Mn, Hf, B, C, and other chemical elements was previously used to refine the size of their grain structure [43,56,61].

4. Conclusions

In this article, scientific approaches and patterns of construction for high-strength and plastic SM materials have been developed for multicomponent copper-based alloys. Two groups of the most representative eutectoid triple systems of Cu-xAl-3Ni alloys (10 ≤ x ≤ 14 wt.%) and Cu-xAl-yNi (7.5 ≤ x ≤ 14 wt.%, 3 ≤ y ≤ 6 wt.%) were selected for a comprehensive study—by modern methods—of their fine structure, phase transformations, and mechanical properties in a wide range of force, deformation, and temperature external impacts. The research was aimed at creating attractive FG and UFG structures that are the most favorable and promising for plasticizing and hardening of the alloys, using new (alternative to known being effective) methods of thermal and mechanical treatments (re-quenching of martensite, SMD, uniaxial compression or HPT, and subsequent annealing). It was shown that the principle of a multilevel hierarchy of the structural-phase state is a key basis for the design and development of high-strength and plastic multifunctional multicomponent SM alloys. From the analysis of the results obtained, taking into account the known data, the following conclusions are made:

• TMTs—with typical moderate volumetric effects |ΔV/V| < 1%, as well as the presence of pre-martensitic softening of elastic constants with high values of elastic anisotropy (10–13 units) and a heterogeneous pre-martensitic state with inherent effects of diffuse electron scattering of tweed TEM contrast—are characteristic of all the studied quenched SM alloys.
• It was found that a repeated quenching of the SM alloys in the martensitic state, providing, along with chemical and phase uniformity, the effect of recrystallization of the grain structure due to a reverse TMT, leads to a noticeable size refinement of the CG structure of prototype alloys and, as a consequence, to an increase in their plasticity.
• Cold uniaxial MPC of the quenched SM alloys, radically size-refining individual martensite crystals and packet-martensite morphology leads to a simultaneous increase in the strength- and plastic characteristics (σ_u = 1150 MPa, σ_y = 400 MPa, ε = 22%) due to both the low elastic modules of alloys and the effective redistribution and auto-adaptation of the elastic-plastic stresses and deformations (distortions) during deformation-induced reorientation of mechanoelastic martensitic crystals.
• Hot isothermal MPC (in the range of 873–1073 K) of the CG quenched alloys, inducing dynamic recrystallization of austenite, leads to (i) a strong size refinement of the grain (FG, up to 100–200 µm) structure with the participation of the pro-eutectoid decomposition, which blocks the grain growth, and, consequently, (ii) to high strength and ductility under the mechanical compression tests (σ_u in the vicinity of 2000 MPa, ε in the range of 55–70%) and tensile tests (with σ_u = 1100–1600 MPa, σ_m = 80 MPa, ε_m = 2 ÷ 3%, and δ = 14–16%).
• The warm isothermal MPC (in the range of 673–873 K) of quenched alloys, inducing the mechanism of the complex dynamic recrystallization and eutectoid decomposition of austenite, leads to radical size refinement of the grain–subgrain (UFG, up to 1–5 µm) structure and, as a consequence, to a combination of the high strength and plasticity of the wrought alloys (σ_u in the region of 1500–1600 MPa, ε in the region of 80–95%).
• It is shown that the HPT (at pressure 6 GPa and 10 revolution) of alloys with subsequent annealing provides the UFG state and, as a result, high-strength (σ_u up to 1400 MPa) and improved plastic (δ = 12–13%) tensile properties.
• Unlike the CG alloys with brittle grain-boundary fracture, the FG and UFG alloys in both austenitic and martensitic states under uniaxial compression or tension experience fracture (i) by the tough-ductile and tough-brittle intragrain mechanisms acting along the boundaries of disperse martensite crystals, and (ii) by inter-grain mechanisms of the same modes of fracture, causing the increased plasticity and deformability of the alloys during shaping.