Session A50: Computational Discovery of High-Entropy Materials

Theory-guided design of high-strength, ductile multi-principal-element alloys within physics-based metrics for machine-learning

For accelerated design of multiple-principal-elements alloys (MPEAs) as promising materials for next-generation energy technologies, we present a rapid theory-guided down-selection for combinatorial synthesis of high-temperature MPEAs having high-strength and ductility. We showcase simple physics-based metrics to predict and to assess rapidly properties for arbitrary metals and solid-solution alloys, in particular strength and ductility. For example, the intrinsic strength of any solid-phase metal (single- and poly-crystal and amorphous) is obtained directly from an electronic metric available from any density-functional theory (DFT) code. For design, we showcase these predictions to inform bulk combinatorial synthesis and characterization to verify down-selection of superior mechanical properties, or other properties including catalysis. Examples for numerous systems will be presented.

    

Zentropy

Macroscopic properties of a system represent the responses of the system to external stimuli, i.e., its internal processes. Their prediction remains a key challenge in science.  Based on the second law of thermodynamics, entropy drives all internal processes in systems.  While the total entropy of a system can be accurately obtained by integration of experimentally measured heat capacity, its theoretical prediction remains elusive due to the difficulty in theoretically sampling all configurations in the system.  Entropy of a system is counted theoretically by either statistical mechanics in terms of Gibbs distribution or quantum mechanics in terms of Fermi-Dirac and Bose Einstein distributions.  Our newly termed zentropy theory integrates them into a nested formula to account for disorder and fluctuations from the electronic scale to the macroscopic scale of the system 1.  In this presentation, the zentropy theory is introduced through the combined law of thermodynamics containing entropy production due to internal processes, and its capability is demonstrated through prediction of emergent behaviors in magnetic and ferroelectric materials including singularity at critical points, effects previously thought to be explainable exclusively via strong correlated physics.  Discussions in fundamental understanding and parameter-free prediction of the magnetic and ferroelectric properties will be given.  Furthermore, the entropy production in the combined law enables us to derive flux equations and coefficients of cross-phenomena from fundamental thermodynamics 2.

1.    Liu, Z.-K., Wang, Y. & Shang, S.-L. Zentropy Theory for Positive and Negative Thermal Expansion. J. Phase Equilibria Diffus. (2022) doi:10.1007/s11669-022-00942-z.

2.    Liu, Z.-K. Theory of cross phenomena and their coefficients beyond Onsager theorem. Mater. Res. Lett. 10, 393–439 (2022).

    

Plasmonic high-entropy carbides

Discovering multifunctional materials with tunable plasmonic properties, capable of surviving harsh environments is critical for advanced optical and telecommunication applications. We chose high-entropy transition-metal carbides because of their exceptional thermal, chemical stability, and mechanical properties. By integrating computational thermodynamic disorder modeling and time-dependent density functional theory characterization, we discovered a crossover energy in the infrared and visible range, corresponding to a metal-to-dielectric transition, exploitable for plasmonics. It was also found that the optical response of high-entropy carbides can be largely tuned from the near-IR to visible when changing the transition metal components and their concentration. By monitoring the electronic structures, we suggest rules for optimizing optical properties and designing tailored high-entropy ceramics. Experiments performed on the archetype carbide HfTa₄C₅ yielded plasmonic properties from room temperature to 1500K. Here we propose plasmonic transition-metal high-entropy carbides as a class of multifunctional materials. Their combination of plasmonic activity, high-hardness, and extraordinary thermal stability will result in yet unexplored applications.   

Phase prediction in high entropy alloys

The term “High entropy” alloys (HEA) refers to a relatively new class of multicomponent—usually five or more—metallic alloys in equal or near equal atomic concentrations. The complex compositions of these alloys, and their derivatives, lead to unique properties. 

Phase prediction in multicomponent alloys remains one of the most fundamental challenges.  We propose a Multi-Cell Monte Carlo algorithm, or (MC)^2, for predicting stable phases in chemically complex crystalline systems. This algorithm takes advantage of multiple cells to represent possible phases while eliminating the size and concentration restrictions in their previous counterparts. Free atomic transfer among cells is achieved via the application of the lever rule, where an assigned molar ratio virtually controls the percentage of each cell in the overall simulation, making (MC)^2 the first successful algorithm for simulating phase coexistence in crystalline solids. We test the method by successfully predicting the stable phases of known binary systems. We then apply the method to several families of “high-entropy” alloys. This method is particularly powerful when applied to multicomponent systems for which phase diagrams do not exist. 

    

Data driven discovery and optimization of high entropy alloys and their hydrides

Data-driven modeling in recent years has ushered in a new paradigm for rapid discovery of useful materials across a plethora of domains in the physical and materials sciences. These methods become particularly invaluable when investigating applications of high-entropy materials, where the combinatorial growth of explorable chemical space makes brute-force experimentation or first-principles simulation intractable. To assess the primary figure(s) of merit for 10,000s of possible materials for a given application and down-select only top candidates for more rigorous examination, accurate and efficient machine learning and data-driven techniques are required. This talk will survey a variety of data-driven (high entropy) hydride discovery exemplars representing efforts between Sandia and a group of international collaborators.  These range from traditional machine learning approaches for direct hydride thermodynamic property prediction to modern graph neural networks, trained as DFT surrogate models, that feed sampling-intensive first principles simulations. These modeling strategies can rapidly screen hydride thermodynamic properties and therefore experimentally target new materials or rationally tune existing ones for various hydrogen storage or compression use cases.  Coupled with additional simple objectives (i.e., raw material costs), multi-dimensional Pareto optimal materials can therefore be identified, targeted, and synthesized.