This joint project between Ohio State University (OSU), the University of Pennsylvania (UPenn), and the National Institute for Materials Science (NIMS) in Japan combines critical experiments from the Japanese side with multi-scale modeling and simulation from the U.S. side to understand and predict the nanoscale mechanisms underlying the effects of point defects on the thermodynamics and initiation and propagation kinetics of martensitic transformations (MTs). In collaboration with the partner institution in Japan on processing, testing and characterization of shape memory alloys and strain glasses, the US participants address fundamental questions of long standing interest concerning MTs and the effects of point defects, such as what elementary defect and defect process constitute the smallest unit of MT, what the activation pathway and barrier energy of these elementary defect processes are, and how point defects modify them. In particular, the investigators systematically probe the effects of random point defects on the multi-plane generalized stacking fault (MGSF) energy landscape along the minimum energy pathway (MEP) connecting the parent phase lattice to the martensitic phase lattice and on the vibrational entropies, which together constitute the crystalline free energy that replaces the phenomenological Landau free energy. A new microscopic phase field model of MTs based on the ab initio energetics and a reaction-coordinate theory will be tested.
The methodology and approach to be developed are rather general and applicable to a larger set of shear-dominated processes such as the displacive-diffusional transformations found in many advanced alloy systems and the shearing-reordering process during plastic deformation of ordered alloys. Computational tools developed will be disseminated widely through free download from the project website. The educational effort involving the exchange of graduate students, postdoctoral researchers and faculty will inject cutting-edge development in materials science and technology and provide a broad international perspective to the student curricula.
The unique properties of ferroic materials including ferroelectric, ferromagnetic and ferroelastic systems originate from structural phase transformations that produce self-accommodating polydomain structures. Sensing and actuation can be realized simultaneously through domain switching under external fields. However, the conventional ferroic materials suffer from slow response, large hysteresis and narrow window of operation, etc., originating from their relatively coarse (micrometer scale) domain structures. In this project we have explored the effects of stress-carrying defects on the scale and characteristics of domain structures in ferroic materials and their responses to external applied fields, using a combination of multi-scale modeling carried out at OSU and MIT and experimentation carried out at our MWN foreign institution, National Institute for Materials Science, Japan, with a central focus on ferrelastic systems. The most significant outcomes of the project include: (a) Major scientific discovery: a new mechanism for Invar and Elinvar anomalies and superelasticity with nearly vanishing hysteresis in a wide temperature range The discovery of Invar and Elinvar anomalies (i.e., zero thermal expansion and temperature invariant elastic modulus, respectively) in Ni-Fe alloys in 1897 by Guillaume was awarded with a Nobel Prize in physics in 1920, but the origin of the these anomalies has remained one of the most challenging puzzles in physics. The state-of-the-art theory is based on ferromagnetic configuration transition suggested by electronic structure calculations [Nature 400, 46 (1999)]. However, such a transition has not been confirmed by experiments. Even if it does exist, the mechanism would operate only in magnetic materials and cannot account for the Invar anomaly found recently in CuZnAl shape memory alloys and the Invar and Elinvar anomalies found in Ti-based GUM metals that are non-magnetic. In this project we have discovered a new, general mechanism underlying Invar, Elinvar, and nearly-zero hysteretic strain response through rendering martensitic transformations (MTs) continuous with defect engineering. In particular, we show via computer simulations that randomly distributed point and extended defects are able to convert a sharp MT (that leads to long-range ordered coarse domain structures) into a broadly smeared, continuous transformation (that leads to nanodomain structures) [Acta Mater. 66, 349 (2014)]. Such a continuous transition can lead to nearly zero thermal expansion (i.e., the Invar anomaly), invariant elastic modulus (i.e., the Elinvar anomaly), and nearly zero hysteresis over a broad temperature range [Acta Mater. 66, 349 (2014); Scientific Report 3 (2013) 2156 1-7, ibid 4 (2014) 3995, 1-5]. This new mechanism operates in both magnetic and non-magnetic materials and may have a broad impact on the development and design of Invar and Elinvar alloys, as well as alloys with ultra-low or tunable (positive or negative) thermal expansions, elastic modulus and hysteresis. (b) New computational methods Mechanistic studies of coupled displacive-diffusional atomic processes are crucial for understanding time-dependent phenomena in materials science. They require modeling capabilities at atomistic length scales but diffusional time scales, which are beyond the reach of current molecular dynamics (MD) methods. The new method, diffusive MD [PRB 84 (2011) 054103; PRB 86, 014115 (2012)], developed in this project, provides the materials community a new computational tool that could address effectively and efficiently these issues. Metallic glasses (MGs) have attracted great attention for their superb properties such as high strength and fracture toughness and large elastic limit that surpass their crystalline counterpart. However, their poor tensile ductility has limited severely their potential applications. Exploring the structural origin of deformation of MGs excludes the continuum-level modeling that are based on phenomenological constitutive laws for plastic flow and MD methods that may not be able to achieve the time and length scales of shear band formation. The heterogeneously randomized shear-transformation-zone (STZ) model and the nanoscale kinetic Monte Carlo (KMC) algorithm developed in this project [Int. J. Plasticity 40 (2013) 1; Acta Mater 73 (2014) 149] can bridge the atomic structural information to the macroscopic mechanical behavior and, thus, offers the materials community a new computational tool to explore the deformation behavior of glassy materials. These codes have been and are still being distributed to the community of materials science, chemistry and physics. A new version of the DMD code will be released soon. (c) Personnel development In addition to training graduate students and postdocs directly involved in the project, the knowledge gained on strain engineering to design new smart materials of much enhance properties has been incorporated into our courses and presented at national and international meetings. Landau theory and computer simulation methods such as the phase field method and DMD have been introduced in our courses on materials modeling and diffusion for senior undergraduate and graduate students. Frequent exchanges of students and ideas have occurred between OSU (focusing on mesoscale modeling), MIT (focusing on atomistic modeling) and NIMS/XJTU (focusing on experiment). (d) Publications Total 35 journal articles have acknowledged this NSF grant.
