We study the physics and chemistry of materials using atomistic computational methods and high-performance computing technology.
The Materials Project is a multi-institution, multi-national effort to compute the properties of all inorganic materials and provide the data and associated analysis algorithms for every material researcher free of charge. The ultimate goal of the initiative is to drastically reduce the time needed to invent new materials by focusing costly and time-consuming experiments on compounds that show the most promise computationally.
The Materials Project has already performed more than 1 million calculations across 70,000 distinct inorganic compounds (requiring over 100 million CPU hours), and has registered over 40,000 users from around the world. Furthermore, we are building tools to facilitate user-contributed data (both computational and experimental) to help researchers disseminate their data and further enrich the entire materials science community.
Leveraging the Materials Project data and capabilities, users have published papers on Li battery materials, solid state electrolytes, piezoelectric materials, and magnetic materials, amongst other applications. Several Python libraries developed by Materials Project are also released open source and are used by thousands of researchers around the world. The current focus of the project is to expand the range of predicted properties to mechanical, thermal, and surface behavior, as well as to predict never-before-seen materials in the area of clean energy production, harvesting and storage.
The Materials Project now routinely performs electronic structure calculations of inorganic materials, with its database of over 70,000 inorganic materials and millions of calculated properties. However, most of these calculations are for materials with either non-magnetic or ferromagnetic order. We aim to develop computational workflows to calculate complex magnetic orderings and associated magnetic properties at scale, to enable discovery of new magnetic materials. One application of this research is discovery of new magnetocaloric materials, which are materials that undergo a temperature change under an applied magnetic field and can be used for room-temperature refrigeration.
Interfaces and surfaces add complexity and engineering avenues to the properties of materials, and can dominate the performance of nanostructured materials, as for example in photovoltaic (PV) nanowires, photocatalysts, or energy storage electrodes. Furthermore, surface energies and related properties are often used in higher length scale theories (e.g. nucleation and growth theories, coarsening, stability, device performance). The objective of our research is to develop a grand canonical approach whereby surfaces are equilibrated under the chemical potentials of select species in the environment and make the data and algorithms available through the Materials Project.
Our interests range from the atomic, molecular, and interfacial to the bulk behavior of different battery components. This involves the study of ionic transport, electrochemical stability and electrolyte solvation structure. Current case studies include alloy anodes, Si-anode compatible electrolyte design, low permittivity and superconcentrated electrolytes, and electrode passivation phenomena. We employ computational tools that span multiple length scales, including first-principles (e.g. density functional theory or Hartree Fock based methods), semi-empirical classical molecular dynamics, and coarse grained continuum approaches.
The Electrolyte Genome & Multivalent Cathode Screening projects aim to accelerate the next generation of energy storage systems, through discovery of novel solid and liquid materials. The Electrolyte Genome uses both density functional theory (DFT) and molecular dynamics (MD) simulations to uncover the properties of potential electrolyte molecules. DFT computations are performed on a large scale to reveal properties such as electrochemical stability windows of tens of thousands of molecules. MD calculations are being semi-automated to be run over dozens of the most promising candidates to determine the structural and dynamical properties such as solvation structure and diffusion coefficient. The Multivalent Cathode Screening project uses high-throughput DFT to search for new chemical compounds that can serve as electrodes by exchanging multivalent cations such as Mg2+, Ca2+, and Y3+. Such formulations can potentially offer large improvements in energy density over today’s Li+ based batteries, but suitable electrode compounds must be identified amongst tens of thousands of potential candidates.
In both applications, the role of computations is to guide JCESR experimental collaborators towards next-generation, blockbuster battery materials. Both projects are therefore in close contact with experimental teams across the country to feed back information between computational and experimental channels.
Liquid electrolytes for Li-ion Batteries (LIB) can exhibit a variety of speciated salt complexes which are in equilibrium with each other. For example, free ions can be present alongside associated neutral pairs. We are interested in exploring speciation and solvation structure at an atomic and molecular level, as well as the consequences of speciation on electrochemical stability and transport properties. As an example, LiPF6 salt in Dimethyl Carbonate (DMC) presents both solvent separated ion pairs as well as contact ion pairs at low salt concentrations, yet exhibits high ionic conductivity at higher concentrations. This behavior may seem contradictory; however, at increasing concentrations, the salt species affects the dielectric properties of the electrolyte which allows for a shift in the equilibrium away from associated complexes, towards conductive free ions. Different salt species present different characteristic peaks in the imaginary part of the permittivity, which allow validation of the models used. As a result, this electrolyte can circumvent decomposition issues tied to the standard polar solvent electrolytes while maintaining desirable transport properties.
The silicon anode has attracted much attention due to its promise for greatly improved capacity over graphite in LIB. However, many challenges still persist, such as large volumetric strains, poor Li-transport, low Coulombic efficiency and a non-passivating solid electrolyte interphase (SEI). Numerous studies have tackled these challenges by modifying nanostructure, size, depth of discharge, and electrolyte additives. However, despite decades of studies, the correlation between the formation of the SEI, its composition and the Li permeability through the interface remains a mystery. Our examination of the Si anode reactions includes modeling of the solvation structures within the bulk electrolyte as well as the electrochemical stability. We are also interested in investigating the diffusivity of Li+ inside the silicate layer via non-equilibrium simulations.
While multivalent cations show potential for energy storage devices with greater capacities than current lithium-based technologies, their practical implementation is inhibited by the lack of suitable electrolytes. Among the challenges posed by electrolytes for multivalent cation batteries is chemical stability during the multi-electron charge transfer processes involved in metallic plating and stripping. During plating, the solvated divalent ion has to undergo a transient, partially charged valence state; e.g. Mg+ or Ca+, as it approaches the metal surface. The process is fundamentally different as compared to monovalent systems (e.g. Li and Na), where no intermediate, reactive cation state exists, as a one-electron transfer is sufficient for plating. Most solvent molecules as well as ion-paired anion molecules were found to be highly susceptible to rapid decomposition when in close contact with the reactive partially charged cation, thereby precluding stable, reversible plating behavior. This effect was elucidated for the first time using theoretical calculations of the bis(trifluoromethane)sulfonimide (“TFSI”) anion, a popularly studied candidate electrolyte species for multivalent ion batteries. Stability in the face of reactive radical monocations is among the many properties of viable candidate electrolyte species evaluated by theoretical calculations in our group.
Over four decades, researchers have unfolded about sixteen photoanodes that can catalyze the oxidation of water. Over three years, using high-throughput computational and experimental materials screening, the Persson group, within the collaboration of the Materials Project and the Joint Center for Artificial Photosynthesis, has added more than a dozen new photoanodes to the existing list of materials. This work was enabled by the large computational database of the Materials Project, which allows an exploratory search for materials based on targeted intrinsic properties. These computationally prescreened materials are subsequently tested by our collaborators at high-throughput experimentation facilities at California Institute of Technology. Our group also develops theoretical methods to assess the electrochemical behavior of materials in order to improve the long-term durability of catalysts. These theoretical methods are being used to find water-stable catalysts as well as materials that can be used as passivating films to protect less water-stable catalysts.
Current efforts are directed toward materials discovery for CO2 reduction and in the identification of design principles for enhancing efficiency and product selectivity for the conversion of CO2 to chemical fuels.
A significant fraction of technologically critical materials are metastable under ambient conditions, yet our understanding of how to rationally design and guide synthesis towards these materials is limited. Polar materials in particular are a promising class of functionality for enhanced performance in energy materials. Our research focuses on identifying novel piezoelectrics, elucidating the fundamental mechanism by which strong polar properties arise in some structures and compositions, understanding how disorder affects this performance and how this property can be exploited for multifunctional materials. A significant portion of this research is focused on identifying synthesizeable materials: developing metrics for this concept and tools that aide experimentalists in identifying the ideal conditions for materials growth. The Materials Project has developed an epitaxial substrate selector application to identify the most promising substrates for materials growth that is available for every material on the website. This capability enables scientists to target environmental conditions for thorough testing by optimizing the substrate selection process. These two capabilities have identified a high energy metastable polymorph of SrHfO3 as a promising theoretical piezoelectric material and assisted our collaborators at the National Renewable Energy Laboratory (NREL) in growing it using the epitaxial strain created by a substrate to stabilize this phase.
Current manufacturing methods for functional thin-film semiconductors, which are used in most electronics, can be replaced with an environmentally responsible and cost-efficient alternative: aqueous solution synthesis with nanoscale polyoxometalate cluster precursors. A common form of these clusters is the Keggin structure, seen to occur for Mo, W, Fe, Al, Ga, and Sn. These clusters’ speciation and stability depends on intrinsic properties and environmental effects (pH, counterions). Aluminum, gallium, and tin Keggin clusters are being investigated by our group using quantum chemical computing approaches. This research is in collaboration with the Center for Sustainable Materials Chemistry (CSMC).
The Local Spectroscopy Data Infrastructure (LSDI) Project will develop a completely integrated platform for first-principles calculations of the so-called “local” environment at atomic positions within a material, which can be revealed through NMR spectroscopy and local-probe X-ray absorption spectra. The developed infrastructure will broadly address the needs of the growing community of chemists and materials scientists who rely on local-environment probe spectroscopy methods, which are relevant for defective, non-stoichiometric, and nano-crystalline materials and interfaces. These classes of materials are increasingly important across a range of applications, and no standardized spectral measurements are available or catalogued to accelerate characterization, understanding and design. Our project will (i) create robust, benchmarked workflows for calculating NMR, XAS, EELS, and other spectra and (ii) develop tools to use these massive data sets to better understand the local-environments that these techniques probe. Currently, we have calculated XAS K-edge spectra on over 37,000 materials and developed an application to match experimentally measured spectra with our simulated spectra.