Joseph Montoya

Accelerating the discovery of materials for clean energy in the era of smart automation

The discovery and development of novel materials in the field of energy are essential to accelerate the transition to a low-carbon economy. Bringing recent technological innovations in automation, robotics and computer science together with current approaches in chemistry, materials synthesis and characterization will act as a catalyst for revolutionizing traditional research and development in both industry and academia. This Perspective provides a vision for an integrated artificial intelligence approach towards autonomous materials discovery, which, in our opinion, will emerge within the next 5 to 10 years. The approach we discuss requires the integration of the following tools, which have already seen substantial development to date: high-throughput virtual screening, automated synthesis planning, automated laboratories and machine learning algorithms. In addition to reducing the time to deployment of new materials by an order of magnitude, this integrated approach is expected to lower the cost associated with the initial discovery. Thus, the price of the final products (for example, solar panels, batteries and electric vehicles) will also decrease. This in turn will enable industries and governments to meet more ambitious targets in terms of reducing greenhouse gas emissions at a faster pace.The discovery and development of advanced materials are imperative for the clean energy sector. We envision that a closed-loop approach, which combines high-throughput computation, artificial intelligence and advanced robotics, will sizeably reduce the time to deployment and the costs associated with materials development.

A high-throughput framework for determining adsorption energies on solid surfaces

In this work, we present a high-throughput workflow for calculation of adsorption energies on solid surfaces using density functional theory. Using open-source computational tools from the Materials Project infrastructure, we automate the procedure of constructing symmetrically distinct adsorbate configurations for arbitrary slabs. These algorithms are further used to construct and run workflows in a standard, automated way such that user intervention in the simulation procedure is minimal. To validate our approach, we compare results from our workflow to previous experimental and theoretical benchmarks from the CE27 database of chemisorption energies on solid surfaces. These benchmarks also illustrate how the task of performing and managing over 200 individual density functional theory calculations may be reduced to a single submission procedure and subsequent analysis. By enabling more efficient high-throughput computations of adsorption energies, these tools will accelerate theory-guided discovery of advanced materials for applications in catalysis and surface science.Surface chemistry: an automatic sense of attractionAn automated procedure for determining the energy required for a molecule to adhere to a surface is developed by researchers in the United States. Joseph Montoya from the Lawrence Berkeley National Laboratory and Kristin Persson from the University of California, Berkeley, introduce an algorithm for finding the adsorption sites on an arbitrary surface. Knowing the amount of energy required for molecular adsorption is crucial for identifying the best materials for use in electronics and catalysis. Density functional theory can predict adsorption energies but usually requires human intuition to tune the calculations. With so many combinations of surface and adsorbate, an automated method is required. Montoya and Persson use open-source computational tools from the Materials Project to present a workflow for performing high-throughput density functional theory calculations for arbitrary slabs and adsorbed species.

Computational prediction of new auxetic materials

Auxetics comprise a rare family of materials that manifest negative Poisson’s ratio, which causes an expansion instead of contraction under tension. Most known homogeneously auxetic materials are porous foams or artificial macrostructures and there are few examples of inorganic materials that exhibit this behavior as polycrystalline solids. It is now possible to accelerate the discovery of materials with target properties, such as auxetics, using high-throughput computations, open databases, and efficient search algorithms. Candidates exhibiting features correlating with auxetic behavior were chosen from the set of more than 67 000 materials in the Materials Project database. Poisson’s ratios were derived from the calculated elastic tensor of each material in this reduced set of compounds. We report that this strategy results in the prediction of three previously unidentified homogeneously auxetic materials as well as a number of compounds with a near-zero homogeneous Poisson’s ratio, which are here denoted “anepirretic materials”.There are very few inorganic materials with auxetic homogenous Poisson’s ratio in polycrystalline form. Here authors develop an approach to screening materials databases for target properties such as negative Poisson’s ratio by using stability and structural motifs to predict new instances of homogenous auxetic behavior as well as a number of materials with near-zero Poisson’s ratio.

The Materials Project: Accelerating Materials Design Through Theory-Driven Data and Tools

The Materials Project (MP) is a community resource for theory-based data, web-based materials analysis tools, and software for performing and analyzing calculations. The MP database includes a variety of computed properties such as crystal structure, energy, electronic band structure, and elastic tensors for tens of thousands of inorganic compounds. At the time of writing, over 40,000 users have registered for the MP database. These users interact with this data either through the MP web site (https://www.materialsproject.org) or through a REpresentational State Transfer (REST) application programming interface (API). MP also develops or contributes to several open-source software libraries to help set up, automate, analyze, and extract insight from calculation results. Furthermore, MP is developing tools to help researchers share their data (both computational and experimental) through its platform. The ultimate goal of these efforts is to accelerate materials design and education by providing new data and software tools to the research community. In this chapter, we review the history, theoretical methods, impact (including user-led research studies), and future goals for the Materials Project. 1 History and Overview of the Materials Project Materials scientists and engineers have always depended on materials property data to inform, guide, and explain research and development. Traditionally, such data originated almost solely from experimental studies. In the past 10–15 years, it has become possible to rapidly generate reliable materials data using scalable computer simulations of the fundamental equations of physics such as the Schrödinger equation. This paradigm shift was induced by a combination of theoretical advances, most notably the development of density functional theory (DFT), algorithmic improvements, and low-cost computing. The Materials Project (MP, or “The Project”) was founded in 2011 as a collaborative effort to leverage ongoing advances in theory and computing to accelerate the research and design of new materials. The Project rests on a comprehensive database of predicted properties of materials that is the result of executing millions of DFT simulations on supercomputing resources. At the time of writing, this database includes >69,000 inorganic materials with crystal structures and total energies, >57,000 materials with electronic band structures, >48,000 with electronic transport properties (Fig. 1) (Ricci et al. 2017), >30,000 with XANES The Materials Project: Accelerating Materials Design Through Theory-Driven. . . 3 Fig. 1 Example of a large electronic transport data set in MP generated through computations. Each point represents one compound, with Seebeck coefficient versus electron conductivity (divided by τ ) plotted. The color represents the thermoelectric power factor (S2σ ), and the point size is proportional to the bandgap (Ricci et al. 2017). This data set is available through the MPContribs platform (see Section 6.2) at: https://materialsproject.org/mpcontribs/boltztrap k-edge spectra (Dozier et al. 2017), >15,000 with conversion battery properties, >6000 with elastic tensors (de Jong et al. 2015a), >3,000 with intercalation battery properties, >1,000 with piezoelectric tensors (de Jong et al. 2015b), >1,000 with dielectric tensors (Petousis et al. 2017), and > 1000 elemental surface energies (Tran et al. 2016). This database is continually expanding with more materials and more properties (see Fig. 2 for an example of properties listed in the current iteration). The Project launched its publicly accessible web site in October 2011 and has since grown into a multi-institution collaboration as part of the US Department of Energy Office of Basic Energy Sciences (BES). The web site provides access to the database as well as applications (or “apps”) that combine and visually present the data for specific analyses such as phase diagram generation or battery electrode evaluation. The MP web site hosts more than 40,000 registered users worldwide consisting of a diverse set of researchers and students from academia, industry, and educational institutions (Figs. 3 and 4). The diversity of the audience base highlights the usefulness of a theory-based materials database across the spectrum of education, research, and development activities. Apart from the core data and web site, MP helps develop and maintain a set of open-source software libraries for setting up, executing, analyzing, and deriving