Search

The tribological performance and underlying deformation behavior of Ni55Ti45, Ni54Ti45Hf1 and Ni56Ti36Hf8 alloys were studied using rolling contact fatigue (RCF) testing and transmission electron microscopy (TEM). TEM results of the as-machined RCF rods, prepared using focus ion beam, revealed some damage very close to the surface. TEM results after initial RCF cycling showed that additional damage was mainly confined to deformation bands that propagated several microns into the sample. These bands formed via localized dislocation slip, possibly on multiple slip systems, within the B2 matrix and/or within transformed B19 prime martensite phase under repeated applied contact stress. Further cycling of Ni55Ti45 and Ni54Ti45Hf1 led to shearing and dissolution of the strengthening precipitates within the deformation bands, followed by formation of nanocrystalline grains and finally amorphization of the remaining matrix material within the bands. The Ni56Ti36Hf8 alloy exhibited a notable increase in RCF performance and a smaller damage zone (1.5 microns) compared to the Ni55Ti45 and Ni54Ti45Hf1 alloys (over 6 microns). This was attributed to the low fraction of B2 matrix phase (less than or equal to 13 %) in the Ni56Ti36Hf8 alloy, leading to formation of narrow deformation bands (less than 11 nm) that were incapable of dissolving the much larger precipitates. Instead, the deformation bands were restricted to narrow channels between the dense cubic NiTiHf and H-phase precipitates. In contrast, broad deformation bands accompanied by shearing and eventual dissolution of the Ni4Ti3 precipitates were observed in the Ni55Ti45 and Ni54Ti45Hf1 alloys due to the high area fractions of B2 matrix phase (~49 %).

Machine learning (ML) is shown to predict new alloys and their performances in a high dimensional, multiple-target-property design space that considers chemistry, multi-step processing routes, and characterization methodology variations. A physics-informed featured engineering approach is shown to enable otherwise poorly performing ML models to perform well with the same data. Specifically, previously engineered elemental features based on alloy chemistries are combined with newly engineered heat treatment process features. The new features result from first transforming the heat treatment parameter data as it was previously recorded using nonlinear mathematical relationships known to describe the thermodynamics and kinetics of phase transformations in alloys. The ability of the ML model to be used for predictive design is validated using blind predictions. Composition - process - property relationships for thermal hysteresis of shape memory alloys (SMAs) with complex microstructures created via multiple melting-homogenization-solutionization-precipitation processing stage variations are captured, in addition to the mean transformation temperatures of the SMAs. The quantitative models of hysteresis exhibited by such highly processed alloys demonstrate the ability for ML models to design for physical complexities that have challenged physics-based modeling approaches for decades., Comment: Submitted to Journal, 34 pages, 6 main figures/tables, and 9 supplementary figures/tables

Many models have been developed for simulation of shape memory alloy (SMA) constitutive responses. The historical development of these models has shifted as fundamental understanding of the physics of SMAs has been advanced: from a focus on transformation behaviors in isolation to the study of cyclic deformations and fatigue, i.e., coupled transformation and plasticity. The motivation for the model we present in this study was to arrive at a single, unified model capable of simulating rate-independent, thermo-mechanical shape memory behaviors, and their coupling to plastic deformation with high computational efficiency. In formulating an energy landscape at a macroscopic scale that encompasses elastic, initiation, and saturation of phase transformation, martensite reorientation, plasticity, and the coupling of plasticity with phase transformation and martensite reorientation terms, and then deriving an implicit, variational discretization of the equations, we achieved this goal. Furthermore, the use of invariant-based transformation surfaces allows us to simulate processing textures that range from random (i.e., as-cast ingot) to transversely anisotropic (i.e., rod, wire, tube) with no additional computational expense. In this presentation, we will review the model framework. Calibration and verification of the model will be illustrated through simulation of empirical data sets that include proportional and nonproportional axial-torsion loading and cyclic tensile deformation to moderate (³10%) strains. Furthermore, we will verify the evolution of the martensite volume fraction simulated by the model with in-situ diffraction experimental data.

Ultra-high strength and energy storage has been observed of a nickel-rich (54-at. %), hafnium-lean (1-at. %), balance titanium alloy. Specifically, full recovery of 4.5% strain upon release of 2.3 GPa uniaxial stress is realized during compressive loading. Furthermore, in cyclic loading, accumulation of permanent set is not observed after the first cycle. Robustness to permanent set is attributed to precipitation of a relatively high volume fraction of fine Ni4Ti3 precipitates, which is the expected result. It was also anticipated that the ability to recover large strains would arise from superelasticity via martensitic transformation. However, in-situ diffraction experimentation revealed a different result. Martensitic (or any other) phase transformation was not observed. Instead, line profile analysis of the diffraction patterns shows the formation of ~25 nm subdomains of B2 structure within the ~5 µm parent grains, accompanied by heterogeneous deformation among similarly orientated diffracting domains. The subdomains fully revert upon unload. In this presentation, the in-situ diffraction characterization of this phenomenal deformation mechanism will be reviewed and conclusions will be supported with in-situ electron microscopy and macroscopic characterization data.