Kiran Mathew

Automated generation and ensemble-learned matching of X-ray absorption spectra

X-ray absorption spectroscopy (XAS) is a widely used materials characterization technique to determine oxidation states, coordination environment, and other local atomic structure information. Analysis of XAS relies on comparison of measured spectra to reliable reference spectra. However, existing databases of XAS spectra are highly limited both in terms of the number of reference spectra available as well as the breadth of chemistry coverage. In this work, we report the development of XASdb, a large database of computed reference XAS, and an Ensemble-Learned Spectra IdEntification (ELSIE) algorithm for the matching of spectra. XASdb currently hosts more than 800,000 K-edge X-ray absorption near-edge spectra (XANES) for over 40,000 materials from the open-science Materials Project database. We discuss a high-throughput automation framework for FEFF calculations, built on robust, rigorously benchmarked parameters. FEFF is a computer program uses a real-space Green’s function approach to calculate X-ray absorption spectra. We will demonstrate that the ELSIE algorithm, which combines 33 weak “learners” comprising a set of preprocessing steps and a similarity metric, can achieve up to 84.2% accuracy in identifying the correct oxidation state and coordination environment of a test set of 19 K-edge XANES spectra encompassing a diverse range of chemistries and crystal structures. The XASdb with the ELSIE algorithm has been integrated into a web application in the Materials Project, providing an important new public resource for the analysis of XAS to all materials researchers. Finally, the ELSIE algorithm itself has been made available as part of veidt, an open source machine-learning library for materials science.Machine learning: Web application for matching X-ray absorption spectraAnalyzing X-ray absorption spectra (XAS) just became easier thanks to a publically-available database and spectra matching web tool. Interpreting XAS data is currently hindered by a lack of reference spectra, which are laborious to obtain as they require finely tunable X-rays only accessible at synchrotron facilities. Here, a collaboration led by Kristin Persson at the University of California Berkeley, and Shyu Ping Ong at the University of California San Diego developed a ‘high-throughput’ approach generating a large database of computed XAS data, along with a machine-learning algorithm matching unknown spectra with ones in the database. The program correctly identified the oxidation states and coordination environments of a diverse test set of materials with high accuracy. The authors hope their web app will provide a valuable public resource for materials science researchers.

Creation of an XAS and EELS Spectroscopy Resource within the Materials Project using FEFF9

We describe the development of an X-ray Absorption Spectroscopy (XAS) and Electron Energy Loss Spectroscopy (EELS) resource in the Materials Project (MP) [1] database as part of the Local Spectroscopy Data Infrastructure Project [2]. This integration within the MP allows material properties and the associated spectroscopies to be calculated using NERSC (National Energy Research Scientific Computing Center) computing facilities and retrieved through either the user portal or the Representational State Transfer (REST) interface [1]. This ability to search the MP database through its REST interface and carry out a comparative analysis with experimental spectra introduces an efficiency in characterization that will aid in understanding material functionality in various technologically important areas such as semi-conductors, batteries, reactions, etc.

Evaluation of thermodynamic equations of state across chemistry and structure in the materials project

Thermodynamic equations of state (EOS) for crystalline solids describe material behaviors under changes in pressure, volume, entropy and temperature, making them fundamental to scientific research in a wide range of fields including geophysics, energy storage and development of novel materials. Despite over a century of theoretical development and experimental testing of energy–volume (E–V) EOS for solids, there is still a lack of consensus with regard to which equation is indeed optimal, as well as to what metric is most appropriate for making this judgment. In this study, several metrics were used to evaluate quality of fit for 8 different EOS across 87 elements and over 100 compounds which appear in the literature. Our findings do not indicate a clear “best” EOS, but we identify three which consistently perform well relative to the rest of the set. Furthermore, we find that for the aggregate data set, the RMSrD is not strongly correlated with the nature of the compound, e.g., whether it is a metal, insulator, or semiconductor, nor the bulk modulus for any of the EOS, indicating that a single equation can be used across a broad range of classes of materials.Equations of State: which are best?A systematic comparison between the performances of several thermodynamic equations of state revealed the superiority of three equations. Equations of state are widely used to describe materials properties based on variables like temperature, pressure, volume, etc. Now, a team from University of California Berkeley and the Lawrence Berkeley National Lab aim to determine the most suitable one for various conditions. The authors used DFT calculations to model the properties of hundreds of elemental, binary and ternary crystalline solids and subsequently fit them with the most commonly-used equations of state. The Birch, Tait and Vinet equations showed the lowest deviation from calculated points, while fitting reasonably well experimental data; this holistic approach underlines that there is not one equation of state to fit all cases.

Author Correction: Automated generation and ensemble-learned matching of X-ray absorption spectra

The following text has been added to the Acknowledgements section: “L. F. J. P. acknowledges support from the National Science Foundation (DMREF-1627583)”

High-throughput computational X-ray absorption spectroscopy

X-ray absorption spectroscopy (XAS) is a widely-used materials characterization technique. In this work we present a database of computed XAS spectra, using the Greens formulation of the multiple scattering theory implemented in the FEFF code. With more than 500,000 K-edge X-ray absorption near edge (XANES) spectra for more than 40,000 unique materials, this database constitutes the largest existing collection of computed XAS spectra to date. The data is openly distributed via the Materials Project, enabling researchers across the world to access it for free and use it for comparisons with experiments and further analysis.