Research

Overview: We are a computational materials science group whose research emphasizes modeling at the atomic scale. The primary objectives are to characterize materials properties that are not easily measured experimentally, and to discover new materials with superior performance. We employ several computational techniques to achieve these goals, including: electronic structure methods (Density Functional Theory), classical approaches such as Monte Carlo and molecular dynamics, machine learning, and high-throughput screening. These tools allow us to predict the thermodynamic, kinetic, mechanical, and transport properties of materials. We have a keen interest in materials of relevance for the transportation and energy sectors. Recent focus areas include materials for electrical, thermal, and chemical energy storage and the mechanical and microstructural properties of metallic alloys. We make extensive use of the high-performance computing resources at the Texas Advanced Computing Center.

For more information, see our full list of publications and the summaries below.

Li-ion transport mechanisms in solid electrolytes [1]

Battery Materials

We characterize limiting phenomena and predict the performance of battery materials. Chemistries currently being explored range from Li-ion to metal-air, Li-sulfur, multi-valent (Mg), and solid-state batteries. Examples of our recent work in this space includes: transport mechanisms in solid electrolytes [1,2,3] and in electrodes [4,5]; multi-valent batteries [6,7]; and interfacial phenomena [8,9,10].

Adsorption of natural gas in metal-organic frameworks [6]

Gas Capture and Storage

This effort aims to identify new storage materials and to understand their processing-structure-properties relationships. Metal-organic frameworks (MOFs) complex hydrids are being explored as the storage medium, with an emphasis on volumetric storage density and mass & thermal transport. Our recent work on hydrogen storage has identified several materials with exceptional storage densities: [1,2,3,4]. Beyond hydrogen, computation has been used to discover materials for CO2 capture [5] and storage of natural gas [6,7].

Discovery of new salt hydrates for thermal energy storage [2]

Thermal Energy Storage

This research direction leverages the group’s expertise in MOFs and computational discovery. The goal is to identify materials that can store thermal energy at high densities by exploiting the latent heat of absorption of water in solid (salt) hydrates and through adsorption in MOFs. In prior work we have characterized all known salt hydrates reported in the Inorganic Crystal Structure Database [1], and evaluated several thousand hypothetical hydrates based on halogen salts [2].

Calculated maximum shear modulus as a function of crystallographic direction for Li, Na, Al, and Ca at
300 K. [7]

Mechanical and Microstructural Properties

Although commonly considered to be meso-scale in nature, the mechanical and microstructural properties of materials can ultimately be traced back to atomic-scale phenomena. We have used atomic-scale simulations to predict a number of these phenomena, including: (i.) interfacial properties related to the nucleation and growth of 2nd phase precipitates (resulting in precipitate strengthening) in aluminum alloys [1-3]; (ii.) twinning and dislocation nucleation in Ni-alloys [4]; adhesion at heterophase interfaces [5-6]; and, the elastic properties of metal electrodes [7].

Prediction of mechanical and electronic properties of interfaces [6]

Surfaces and Interfaces

Surfaces and interfaces control many properties of materials, and thus comprise a focus area of our research. For example: the composition and magnetic state of surfaces can influence their catalytic properties [1]; charge transfer across interfaces is an essential process for determining the stability and functionality of battery materials [2,3]; the structure of interfaces impacts the mechanical properties of solids [4]; and, the wettability of electrode-solid electroltye interfaces is a prerequisite for achieving practical solid-state batteries [5,6].

Elementary features that strongly correlate with cation mobility in anti-perovskite solid electrolytes [3]

Machine Learning

Our group applies machine learning (ML) to accelerate the rate of materials discovery and to reveal the elementary materials features that correlate with (or control) a desired property. In the former case we have trained ML algorithms to make accurate predictions of the useable hydrogen capacity on a dataset containing nearly one million MOFs using limited and elementary structural data as input [1]. The ML model is available as a web application that can be used to quickly estimate H2 uptake in new MOFs [2]. In the latter case, we have used ML to develop quantitative structure-property relationships (QSPRs) for solid electrolytes [3] and thermal energy storage materials [4]. These QSPRs help us to understand why some materials perform well and others do not, and inform design rules to guide subsequent discovery.

Models of adiabatic and non-adiabatic charge transport in lithium-surful battery cathodes [3]

Charge Transport

The movement of charge, whether ionic or electronic, is a performance-limiting process in many applications. We have modeled many aspects of charge transport through solids using multiple techniques in many contexts, including: defect mediated transport in electrode materials [1,2]; adiabatic vs non-adiabatic processes [3]; multi-scale models [4]; contributions from the presence/absence of lattice crystallinity [5]; and, correlations with the rotational dynamics of cluster anions [6,7]