High-throughput data generation methods and machine learning (ML) algorithms have given rise to a new era of computational materials science by learning relationships among composition, structure, and properties and by exploiting such relations for design. However, to build these connections, materials data must be translated into a numerical form, called a representation, that can be processed by a machine learning model. Datasets in materials science vary in format (ranging from images to spectra), size, and fidelity. Predictive models vary in scope and property of interests. Here, we review context-dependent strategies for constructing representations that enable the use of materials as inputs or outputs of machine learning models. Furthermore, we discuss how modern ML techniques can learn representations from data and transfer chemical and physical information between tasks. Finally, we outline high-impact questions that have not been fully resolved and thus, require further investigation.
Electrochemical energy conversion technologies based on the oxygen evolution reaction (OER) are at the heart of many efforts to achieve a sustainable future, carbon-free fuel, and a circular economy. The sluggish kinetics of oxygen electrocatalysis, as well as the high overpotential required to attain practical current densities, limit the efficiency of several promising electrochemical technologies, including water and carbon dioxide electrolyzers, metal–oxygen batteries, and fuel cells. The most efficient OER catalysts are precious metals such as iridium- and ruthenium-based materials (i.e., IrO2 and RuO2). This fact represents a challenge against the cost-effective implementation of these electrolysis technologies.1-3 As a result, there is a necessity for the development of cost effective PGM-free OER catalysts, with equivalent or superior activity and durability to the PGM catalysts. This presentation will describe the application of machine learning (ML)-guided materials discovery and high-throughput synthesis to address these concerns, taking advantage of the intriguing properties and rich chemistry of nanoporous materials, the demonstrated capability of machine learning (ML)-guided materials discovery, and the high OER electrocatalytic activity of perovskites especially in alkaline media.4-10 Simulation of over 8,000 perovskites across a variety of cell sizes, space groups, and compositions using density functional theory (DFT) has been performed. Mining of the simulation data indicated that experimentally-known perovskites are characterized by low energy above the thermodynamic convex hull. Efficient search algorithms, deep learning-based models, and DFT calculations have been used to explore the space of perovskite oxides to produce novel compositions with tailored electronic descriptors. Promising compositions designed for high activity and stability are then selected for high throughput automated synthesis using the High-Throughput Research Facility at Argonne National Laboratory. A correlation between the phase purity, annealing temperature and OER activity has been identified. Acknowledgements This work was supported by the U.S. Department of Energy, Advanced Research Projects Agency-Energy (ARPA-E) under the DIFFERENTIATE program. This work was authored in part by Argonne National Laboratory, a U.S. Department of Energy (DOE) Office of Science laboratory operated for DOE by UChicago Argonne, LLC under contract no. DE-AC02-06CH11357. References Katsounaros, Ioannis, Serhiy Cherevko, Aleksandar R. Zeradjanin, and Karl JJ Mayrhofer. "Oxygen electrochemistry as a cornerstone for sustainable energy conversion." Angewandte Chemie International Edition53, no. 1 (2014): 102-121. Lee, Youngmin, Jin Suntivich, Kevin J. May, Erin E. Perry, and Yang Shao-Horn. "Synthesis and activities of rutile IrO2 and RuO2 nanoparticles for oxygen evolution in acid and alkaline solutions." The journal of physical chemistry letters3, no. 3 (2012): 399-404. Cherevko, S. et al. Oxygen and hydrogen evolution reactions on Ru, RuO2, Ir, and IrO2 thin film electrodes in acidic and alkaline electrolytes: A comparative study on activity and stability. Today 262, 170–180 (2016). Nahar, Lamia, Ahmed A. Farghaly, Richard J. Alan Esteves, and Indika U. Arachchige. "Shape controlled synthesis of Au/Ag/Pd nanoalloys and their oxidation-induced self-assembly into electrocatalytically active aerogel monoliths." Chemistry of Materials29, no. 18 (2017): 7704-7715. Farghaly, Ahmed A., Rezaul K. Khan, and Maryanne M. Collinson. "Biofouling-resistant platinum bimetallic alloys." ACS applied materials & interfaces10, no. 25 (2018): 21103-21112. Khan, Rezaul K., Ahmed A. Farghaly, Tiago A. Silva, Dexian Ye, and Maryanne M. Collinson. "Gold-Nanoparticle-Decorated Titanium Nitride Electrodes Prepared by Glancing-Angle Deposition for Sensing Applications." ACS Applied Nano Materials2, no. 3 (2019): 1562-1569. Farghaly, Ahmed A., Mai Lam, Christopher J. Freeman, Badharinadh Uppalapati, and Maryanne M. Collinson. "Potentiometric measurements in biofouling solutions: comparison of nanoporous gold to planar gold." Journal of The Electrochemical Society163, no. 4 (2015): H3083. Suntivich, Jin, Kevin J. May, Hubert A. Gasteiger, John B. Goodenough, and Yang Shao-Horn. "A perovskite oxide optimized for oxygen evolution catalysis from molecular orbital principles." Science334, no. 6061 (2011): 1383-1385. Hwang, Jonathan, Zhenxing Feng, Nenian Charles, Xiao Renshaw Wang, Dongkyu Lee, Kelsey A. Stoerzinger, Sokseiha Muy et al. "Tuning perovskite oxides by strain: electronic structure, properties, and functions in (electro) catalysis and ferroelectricity." Materials Today31 (2019): 100-118. Gómez-Bombarelli, Rafael, Jennifer N. Wei, David Duvenaud, José Miguel Hernández-Lobato, Benjamín Sánchez-Lengeling, Dennis Sheberla, Jorge Aguilera-Iparraguirre, Timothy D. Hirzel, Ryan P. Adams, and Alán Aspuru-Guzik. "Automatic chemical design using a data-driven continuous representation of molecules." ACS central science 4, no. 2 (2018): 268-276.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.