In atomistic simulations of B2 NiAl usirn Molecular Dynamics and Emxbedded Atom Method Ni-Al potentials, both thermally induoed and stress induced crent martensitic nucleation and growth processes of a tetragonal Llo phase have been observed, including the formation of self-accommodating martensite variants. Nucleation is found to occur only at special heterogeneous sites that can be well characterised. Prior to nucleation, clear evidence of the buildup of large amplitude localized soft modes can be seen. A discussion of the kinetics of these processes is presented.

We show that the interplay between electronic structure properties, crystalline effects and local chemical order is quite distinct for Ni-Al, Ni-Ti and Cu-Zn, although they all exhibit a first order displacive transformation upon cooling from a high temperature partially ordered B2 phase. For this purpose, electronic structure and phase stability properties of these three alloys are investigated with a first principles approach. The study is based upon the Generalized Perturbation Method applied to the Korringa-Kohn-Rostoker multiple scattering description of the Coherent Potential Approximation, and temperature effects are taken into account within a generalized mean field approach.

Bulk energy considerations can predict the gross features of microstructures observed in martensitic phase transitions, but not finer details such as the length scale on which twinning occurs. These are determined by surface energy. A standard approach is to minimize the sum of bulk and surface energy within the class of “twinned” configurations. This leads to the conclusion that the twin width w and the twin length L should satisfy w ∼L∼. While that seems to be the rule, a different structure is sometimes observed, for example in In-Tl and Cu-Al-Ni. It involves successive coarsening of the twins away from an austenite/twinned-martensite interface.

Vacancy behaviors during ageing of Cu-26Zn-4A1 and Cu-14Al-4Ni alloys have been investigated and compared by means of positron annihilation (PA) and electrical resistivity measurement. For ageing in martensitic state after direct quenching, it is observed that the S parameter values of Cu-Zn- Al specimens, measured in liquid nitrogen, increase at first and then decrease, while those of Cu-Al-Ni remain unchanged. The activation energies calculated from the S parameter for increasing and decreasing stages are o.4lev and o.63ev respectively, and the former can be corresponding to the formation energy of vacancy clustering, while the latter may be regarded as the migration energy of effective vacancies. A mechanism is put forward that the clustering of quenched-in vacancies results in a decreasing of the ordering degree and a reduction of the stored energy in martensite, which is responsible for the early stage of the stabilization of martensite in Cu-Zn-Al alloys. However, the fact that Cu-Al-Ni alloy is not subject to the stabilization is assumed to be owing to the immobility of supersaturated vacancies in its martensitic state which may be associated with the strong binding force between Ni and Al atoms.

Static configurations of stress-induced martensite and pretransition microstructures in Ni62.5A137.5 have been studied with conventional as well as high resolution electron microscopy using single crystal TEM specimens containing defect sites of extremely high stress intensity. From images of different regions around such defects it can be concluded that the austenite lattice develops transverse displacement modulations with increasing amplitude and correlation directly related to the {110}<110> shear-plus-shuffle displacements required to form the martensite structures. Different steps in this transition sequence are presented and discussed.

An interesting problem within the research on shape memory is provided by the investigation of the pseudo-elastic hysteresis loop that accompanies the strain-induced austenitic-martensitic phase transition. Miller [1] suggested by thermodynamic arguments that the size of that loop should be related to the coherency energy at interfaces between martensite and austenite. That proposition received a boost when Müller & Xu [2] showed that the same thermodynamic arguments predicted the onset of internal yield and recovery inside the hysteresis loop on a well-definqd line of unstable phase equilibrium. Recently Fu, Müller and Xu [3] have reported results with improved specimens and methods of observation that may help to understand the nature of metastability inside the hysteresis loop.

The aging effects of two kinds of Cu-Zn-Al shape memory alloys (Cu-ll.4 Zn-18.7A1 (A) and Cu-ll.2Zn-17.lAl (B) in at%) have been examined by differential scanning calorimetry (DSC), transmission electron microscopy (TEM) and atom location by channeling enhanced microanalysis (ALCHEMI). In the directly quenched (D.Q.) state, alloy A was the parent phase, Ms being 253 K, and alloy B was the martensite phase. The alloy B was subjected another quenching treatment as follows: It was once quenched into an oil bath at 423 K and held for 300 s, followed by quenching into iced water (step quench (S.Q.) ). The D.Q. alloy B did not exhibit the reverse transformation because of a stabilization of the martensbite phase, but the S.Q. alloy B did and its As temperature of the reverse transformation was raised with the progress of aging at the martensitic state. Fraction of Zn atoms at the Cu(2) site examined by the ALCHEMI measurements was almost the same in the parent phase of D.Q. alloy A and its aged one, indicating no change in Cu and Zn atom sites, while it was gradually decreased in S.Q. alloy B with the progress of aging. The fraction of Zn atoms in D.Q. alloy B was much lower than those in the S.Q. alloy B and its aged one. TEM observation of the S.Q. alloy B revealed that stacking faults as the lattice invariant shear in the M18R martensites decreased in the density with the progress of aging. The decrease in the fraction of Zn atoms and in the density of stacking faults well corresponds to the increase in As temperature, and thus the martensite stabilization was attributed to a disordering between Cu and Zn atoms and to an annihilation of stacking faults.

Third element additions to TiNi provide a wide range of modifications of its shape memory properties. The advantages of Cu additions are to provide a more narrow hysteresis, less sensitivity to the Ti::Ni(+Cu) ratio of the temperature at which martensite starts to form (Ms), a larger strength differential between the austenite and martensite phases, and superior fatigue resistance. The substitution of a few atomic percent Cu for Ni does not significantly alter the crystal structure of either the cubic B2 austenite nor the monoclinic B19ʹ martensite phases; however, the addition of greater than 10% Cu results in an orthorhombic B19 martensite phase. For the case of 10% Cu, a two-step martensitic transformation occurs upon cooling, with the cubic austenite transforming to the orthorhombic B19 martensite and subsequently to the monoclinic B19ʹ martensite. As a result of this two-step crystallographic transformation, material properties such as resistivity and shape change also exhibit a two-step transformation.

In situ transmission electron microscopy heating and cooling experiments are used to observe the two-step martensitic transformation and to establish an orientation relationship between the B19 orthorhombic and the B19ʹ monoclinic structures. Strain vs temperature Ms tests establish the relative shape changes associated with both the cubic-to-orthorhombic transformation and the orthorhombic-to-monoclinic transformation. Similar Ms tests, where an applied load is removed during the transformation, establishes a crystallographic dependence between the two shape changes. Whereas binary TiNi is "trained" to undergo a specific shape change, this ternary TiNiCu alloy has a "natural" direction associated with the second step of its shape change.

A survey by differential scanning calorimetry (DSC) and recovery during heating of indentations on a series of nickel-aluminum alloys showed that the Ni-36 at.% Al composition has the best potential for a recoverable shape memory effect at temperatures above 100°C. The phase transformations were studied by high temperature transmission electron microscopy (TEM) and by high temperature x-ray diffraction (HTXRD). Quenching from 1200°C resulted in a single phase, fully martensitic structure. The initial quenched-in martensites were found by both TEM and X-ray diffraction to consist of primarily a body centered tetragonal (bct) phase with some body centered orthorhombic (bco) phase present. On the first heating cycle, DSC showed an endothermic peak at 121°C and an exothermic peak at 289°C, and upon cooling a martensite exothermic peak at 115° C. Upon subsequent cycles the 289°C peak disappeared. High temperature X-ray diffraction, with a heating rate of 2°C/min, showed the expected transformation of bct phase to B2 between 100 and 200°C, however the bco phase remained intact. At 400 to 450°C the B2 phase transformed to Ni2Al and Ni5Al3. During TEM heating experiments a dislocation-free martensite transformed reversibly to B2 at temperatures less than 150°C. At higher temperatures (nearly 600°C) 1/3, 1/3, 1/3 reflections from an ω-like phase formed. Upon cooling, the 1/3, 1/3, 1/3 reflections disappeared and a more complicated martensite resulted. Boron additions suppressed intergranular fracture and, as expected, resulted in no ductility improvements. Boron additions and/or hot extrusion encouraged the formation of a superordered bct structure with 1/2, 1/2, 0 reflections.

Samples of Ni-36 at% Al which were hot rolled, then solutionized at 1230°C and quenched, exhibited reversible transformations between the low-temperature, martensitic phase and the high temperature, austenitic phase. The initial peak transformation temperatures on heating (AP), cooling (MP), and the heats of transformation (ΔH) determined by differential scanning calorimetry (DSC) were ∼121°C, 119°C and 4 J/g, respectively for the as-solutionized material.

Decreases in transformation temperatures and heats after aging at 350°C were also determined using differential scanning calorimetry (DSC), and microhardness testing (DPH). The DPH values plotted against age time showed two distinct peaks, a result similar to precipitation hardening in other Al alloys.

Crystal structure of the ζ2' martensite in a Au-49.5at%Cd ally has been analyzed by the single crystal x-ray diffraction method. The crystal lattice was trigonal and the lattice constants were a:0.8095(3) and c=o.57940(6) nm. There were 18 atoms in a unit cell. The space group was P3, which was different from that previously determined by Vatanayon and Hehemann. The structure was refined by the full matrix least squares method to a final R factor of 7.8% and a weighted R factor of 4.1%.

We present an overview of a theory of martensitic transformations developed by J. M. Ball and the author, and additional related results by Bhattacharya, Chu and Collins and Luskin and Kinderlehrer. This theory is in a form that is amenable to detailed numerical computations of microstructure. We describe the energy minimizing microstructures of a single crystal under applied displacements and under detailed dead loading. The relation of the theory to the crystallographic theory of martensite is presented.

A consequence of the theory is the sensitivity of the patterns of microstructure to the precise lattice parameters. In the case of the wedge-like microstructure studied by Bhattacharya, it is found that this microstructure is only possible as a coherent, energy minimizing microstructure if very restrictive conditions on the lattice parameters are satisfied. Materials that exhibit the wedge satisfy this relation closely. The analysis suggests that optimal shape-memory behavior may be related to relations of this type.

The theory has a relation to the work of Khachaturyan, Roitburd and Shatalov. We compare and contrast the two approaches. We present results that suggest that any approach based on the kinematics of linear elasticity will make serious quantitative errors in the prediction of microstructure.

The thermodynamic theory of the deformation of heterophase solids containing polydomain phases is presented. The polydomain phases consisting of twins are considered. Elastic interaction between twins leads to the formation of an equilibrium polydomain structure. Its basic element is a polytwin: a plane-parallel plate consisting of alternating plane-parallel twins. In a uniform single crystal the polytwin structure is unstable and should disappear at any uniform external stress. But it can be stable in a heterophase system. The evolution of the equilibrium structure with the variation of the external stress exhibits reversible superplasticity and pseudoelasticity. Two contributions into total strain are considered: due to displacements of twin boundaries inside a polydomain (polytwin) phase and due to movement of interfaces between polydomain and matrix phases.

Localized phase transitions, as well as shock waves, can be modeled by material discontinuities satisfying appropriate jump conditions. One can show that the classical system of Rankine-Hugoniot jump conditions is incomplete in the case of subsonic phase boundaries. The supplementary condition which generalizes the condition of phase equilibrium, can be obtained from the traveling wave solution of the truly dynamic system of equations describing the interface structure.

The nature of the premartensitic state is still under debate. It is becoming evident that alloys which transform into a long period polymorph display the equivalent precursor strain modulations in the parent phases. The situation in alloys which transform in the center of the BZ is less clear, however. On the basis of electron microscopic [1,2,3] and X-ray diffraction data [4], it has been proposed for Fe-Pd alloy that the parent state consists of a coherent mixture of tetragonally distorted variants which may form a transitional phase. The interpretation is substantially based on the observation of 110 diffuse scattering and the softening of the elastic constant ½(C11 -C12) [1,2,5,6,7]. We have performed TEM studies of In-Tl alloys and investigated the low frequency mechanical response of In-Tl, In-Cd and Fe-Pd alloys. The phase diagrams of the three alloys [8,9,10], schematically presented in Fig. 1, show the common FCC-FCT martensitic transformation. Our studies indicate isotropic "elastic" softening and show that 111 streaking is present as well. This paper presents the mechanical results and their preliminary interpretation.

Enhanced room temperature toughness of the Zr50Pd35Ru15B2 phase alloy was found to be a result of the activation of an additional deformation mode besides the b=[001] dislocation slip mode - {114}-type mechanical twinning. The twinning is a true one, i.e. there is no change in the ordered crystal structure. Another additional mode of plastic deformation, expected for more Pd rich alloys, is the formation of stress-induced martensite. The martensite was found to have a CrBtype structure.

Conventionally cast and hot-rolled Ni-Fe-AI-B alloys containing 4-20 at.% Fe, 23.9- 31.5 at.% Al, and 300 wppm B were investigated in this study. After oil quenching from 1300°C, all the alloys—except SMA-15 (27A1-14Fe)—have at least a two-phase microstructure, one phase of which is martensite with the characteristic plate morphology, and the other a globular second phase distributed throughout the microstructure. The amount of second phase generally increases with increasing Fe content. Alloys containing less than 14% Fe were found to be quite brittle at room temperature, indicating that a ductile second phase is at least partly responsible for the improved room-temperature ductility in the high-Fe alloys. The best tensile ductility (12%) was obtained in SMA-17 (23.9AI-20Fe) which was shown by X-ray diffraction to consist of 40% (mostly disordered) fcc [(Ni,Fe)3 (AI,Fe)] + 30% (partly ordered) bct martensite + 30% B2. Differential scanning calorimetry showed that the transformation temperatures for this alloy were MP = 65°C and AP = 95°C. Room-temperature tensile strains of 2-3% could be almost completely recovered in SMA-17 by heating for 3 min. at 600°C with the load removed. Upon subsequent cycling (i.e., strain-anneal cycling), the amount of strain recovery increased dramatically from 70% in the first cycle to nearly 100% after 4-5 cycles, indicating that cold work may help in improving the shape memory characteristics of this alloy. SMA-15 was found to have significantly higher transformation temperatures (Mp = 143°C and Ap = 170°C) than SMA-17; however, it is relatively brittle compared to SMA-17.

Nickel aluminide intermetallic compounds are of interest due to their good high temperature properties. However, they are associated with room temperature brittleness. A method for enhancing the room temperature ductility of polycrystalline NiAI in the composition range of 62-65 at.% by using the martensitic transformation is presented. Transformationinduced ductilities up to 4.5% in specimens with coarse grains have been obtained in our investigation. The effects of Ni concentration, grain size and quenching rate from the parent B2 phase are studied and optimized to induce maximum transformation ductility. The original excellent high temperatu re properties are restored from the ‘soft’ room temperatu re martensitic phase by heating to an appropriate temperature to revert to the equilibrium phases.

Potential high transition temperature shape memory alloys based on NiMn have been studied with emphasis on the shape recovery, transformation temperatures and mechanical properties. Binary NiMn, which has been reported to be brittle, has a low shape recovery, but this can be increased with Al or Ti additions. Also, the transformation temperature can be changed and the room temperature ductility improved by ternary element additions. The various substitutional solute characteristics affecting the shape recovery, the transformation temperatures, and the ductility have been examined.

Periodic multilayered titanium-rich Ni-Ti thin films were prepared by magnetron sputtering from alternating Ni45Ti50Cu5 alloy and pure titanium targets, with an alloy-layerfl'i-layer thickness ratio of 9:1. The microstructure, martensite transformation behavior, and precipitate and defect structures were studied in films which had been annealed at 923K for one hour and furnace cooled at <20 K/min. Energy dispersive X-ray fluorescence measurements showed that the resulting films had a hyperstoichiometric titanium content of approximately 51 atomic percent. Ti2Ni precipitates were found in the annealed structures which were oriented with [1 1 0 ]Ti2Ni parallel to [110]B2 and (001)Ti2Ni parallel within +/− 10 to (001)B2. Differential scanning calorimetry (DSC) revealed an unusually low transformation enthalpy for the martensite reaction in the film (9.1 J/g as opposed to 20.7 J/g for the alloy sputtering target), and a significant fraction of residual B2 austenite was found at temperatures well below the nominal Mf. The martensite transformation was found to occur in two steps involving, on cooling, the initial formation of an orthorhombic martensite prior to transformation to the monoclinic martensite phase at low temperature.