The thermoelastic B2 L1 martensitic transformations in binary Ni Mn and Ni Mn alloys are known to occur at high temperatures; therefore, it is rather difficult to study the structure and properties of these alloys. In [1,2], we comprehensively investigated the fine structure and the TMTs of these alloys and determined critical temperatures: M = 970 K, M = 920 K, A = 970 K, and A = 1020 K; M = 940 K, M = 930 K, A = 990 K, and A = 1000 K, respectively. The high temperature B2 L1 phase transformations were detected in many binary and multicomponent intermetallic alloys based on nickel and titanium, such as Ni–Mn, Ni–Mn–Al, Ni–Al, Ni–Al–Co, Ni-Cu-Al, Ti–Rh, Ti–Rh–Ni, Ti–Ir, and Ti–Ir–Ni [3–14]. We assumed that these transformations in the alloys based on these intermetallic compounds also has the character of TMT, which should cause shape memory effects in them. Nevertheless, the boundaries of phase transformations and the crystal-structure types of austenite and martensite phases have not been exactly determined for ternary Ni–Mn–Al quasi-binary system. As a rule, alloying with a third chemical element decreases critical temperatures M , M , A , and A , which facilitates the designing of alloys with given martensitic transformation parameters. The purpose of this work is to study the influence of aluminum alloying on the structure and the thermoelastic martensitic transformations, i.e., the critical points of the forward (M , M ) and reverse (A , A ) transformations, in quasi-binary Ni Mn Al alloys. There is reason to believe that this transformations in alloys based on such intermetallic compounds and in other B2 nonferrous alloys (titanium nickelide, copper based alloys, ferromagnetic Heusler alloys, alloys based on alloyed manganese nickelide) also have characteristics of TMT, and this fact should cause the shape memory effects in them.
2. Materials and Methods
The alloys were melted by the electric-arc method from high-purity metals (99,99-99,999% purity) in an inert atmosphere of purified argon. The ingots (certified by the chemical composition) for homogenization were subjected to remelting (at least three times) followed by longterm vacuum annealing at 1173 K. The alloy ingots were spark-cut into plates, which were again subjected to homogenizing annealing during 6 h followed by water quenching or slow ( 100 K/h) cooling from 1073 or 1173 K. The X-ray diffraction analysis by the θ/2θ method was carried out using a DRON-3M diffractometer in the Cu Kα radiation monochromatized by a graphite single crystal. The electron-microscopic studies were performed at the Center of Collaborative Access, Institute of Metal Physics, Ural Branch, Russian Academy of Sciences using JEM-200 CX (maximum accelerating voltage of 200 kV) and CM-30 transmission electron microscopes and Quanta 200 scanning electron microscope (maximum accelerating voltage of up to 30 kV), which was equipped with EDS and EBSD systems. Electron-microscopic transmission studies were performed using bright- and dark-field regimes. To identify phases, we analyzed selected area electron diffraction (SAED) patterns. Electrical resistivity ρ(T) was measured by potentiometry using a double bridge in the temperature range 100–1170 K.
3. Results and Discussion
In accordance with the transition temperatures determined from the resistivity curves ρ (T), a diagram of martensitic transformations was constructed (shown in Fig. 1). This diagram shows the temperature and concentration dependences of martensitic transformations of alloys, as well as atomic ordering B2 L2 of austenitic phases and magnetic transitions from a paramagnetic state to antiferro- and ferromagnetic. The phase composition of the alloys was determined by the X-ray method and from selected electron diffraction patterns [1-14].
In accordance with the presented diagram, it follows that at room temperature the structure of martensitic phases changes with increasing aluminum content in the following sequence: a crystal lattice of type L1 (2M), 14M and 10M.
The results of the microhardness measurements of the alloys are given in Table. Doping with aluminum in the range from 0 to 18 at. % reduces the resistance of martensite to the hardness tester indenter. It can be concluded that, being in a martensitic state 2M and 14M with lower elastic moduli, these alloys have a lower hardness. Increase of microhardness in alloys with 20 and 22 at. % Al can be associated, first, with a further change in the type of the crystal structure of martensite from 14M to 10M. But, most likely, this fact correlates with the fact that the critical temperatures in these alloys are close to room temperature, at which microhardness measurements were carried out. Therefore, an increase of microhardness in alloys with 20 and 22 at. % Al may be due to the presence of two-phase (B2 + 10M), induced and due to deformation in the testing of hardness, providing additional hardening of the alloys. In austenitic alloys with 24 and 25 at. % Al superstructural atomic ordering occurs from B2 and L2 and the microhardness decreases again slightly.
A fractography of the alloys was carried out using SEM in the secondary electron mode after bending tests of samples in order to determine the nature of the fracture. In fig. 2, a, b shows the image of the fractures of the studied alloys. Fractographic studies have shown that their destruction occurs both by means of the transcrystalline (along the grain boundaries) and intercrystalline types (mainly along the joints of the packets of martensitic crystals inside the grain). Obviously, the brittle destruction occurs along the grain (Fig. 2, a). Intercrystalline destruction occurrs by brittle and brittle-ductile types, which probably depends on the location of the packages of martensitic plates relative to the direction of the fracture crack. If the crack expends along the flat boundary of the package, then brittle fracture occurs. This behavior can be explained by the concentration of the strains and stresses near the boundaries packages caused by the realization of the TMTs in certain regions of the polycrystalline alloy. If the crack develops at an angle to the martensitic plates, then in such areas the fracture has a ductile or brittle-ductile nature (Fig. 2, b).
The thin structure of these alloys was studied by TEM. Fig.3 shows TEM images of the martensite in the Ni Mn Al alloy (a, b) and the corresponding selected area electron diffraction (SAED) patterns (c). It allows to identify a seven-layer crystal lattice corresponding to the 14M crystal lattice. The substructure is represented by thin plates of the martensitic phase. The boundaries of the packets of martensitic plates are runoffs of grain-boundary misfit dislocations.
TEM studies were also carried out on alloys doped with 22 and 25 at% Al. They showed that Ni Mn Al alloy is in single-phase austenitic state B2 (or L2 ). This is evidenced not only by dark-field and bright-field images of the tweed structure (Fig. 4 a, b), but also the analysis of SAED patterns (Fig. 4 c, d, e) and X-ray data. The analysis showed that the diffuse scattering bands, which are more intense near the reflections, are traces of the intersection of the Ewald sphere by flat scattering layers along the 111 * of reciprocal lattice, passing through its nodes, with the exception of 000. The strongest scattering is concentrated along the 110 * and 112 *.
The regular nature of diffuse scattering and tweed contrast is due to the presence of localized cooperative, mainly shear, atomic displacements in crystals . Flat scattering layers over 111 * are interpreted in elastic-soft low-modulus crystals as a result of destabilization with respect to local displacements of close-packed atomic chains by 111 and their shear correlations over close-packed 110 type planes. These phenomena indicate that the alloy under study was in the pretransition B2 state [4,5].
The crystal lattice of the Ni Mn Al alloy, according to X-ray analysis, is ordered by type L2 . The SAED patterns presented in Fig. 5, c, d show diffraction patterns that are identical to B2 or L2 phase diffractions. They clearly show diffuse effects in the form of band and satellites. In addition, the grain structure of the alloy has tweed diffraction contrast (Fig. 5, a, b). These effects indicate the pretransition state of the alloy.
• It is shown that with increasing in the aluminum content in alloys of the NiMn-NiAl quasi-binary system, the critical temperatures of the thermoelastic martensitic transformations decrease and the martensite crystal lattice changes in the sequence 2M - 14M - 10M.
• The microhardness of the studied alloys naturally changes depending on the chemical composition and structural type of the resulting martensite with a specific thin substructure and morphology of martensitic crystals.
• Destruction in the studied alloys occurs by transcrystalline and intercrystalline types. In alloys, brittle transcrystalline fracture cracks expend both along grain boundaries and along flat boundaries of martensite plate packs; intercrystalline brittle-ductile fracture occurs when cracks expand along shear systems at an angle to the habit of martensitic plates.
• As shown by TEM and SEM, martensite has a predominant morphology in the form of a hierarchy of packages of lamellar and internally finely twinned coherent crystals with flat habit boundaries close to 110 . The systems of twinning shear for all types of martensite are close to the soft mode 101 10 of B2 austenite.
• It was found that in the premartensitic state, the austenite of the studied alloys can be described by short-range atomic displacements of the type of future martensitic phase mainly by correlated and lateral displacements along the 101 101 system with the formation of nano-localized structures of the intermediate type.