|About this Abstract
||2020 TMS Annual Meeting & Exhibition
||Functional Nanomaterials 2020: Translating Innovation into Pioneering Technologies
||High Throughput Screening of Nano Catalysts for PEMFC/AEMFC and Machine Learning Prediction of Chemisorption
||Soonho Kwon, Jung Woo Choi, Hyuck Mo Lee
|On-Site Speaker (Planned)
||Hyuck Mo Lee
For rational design of new cathode nano catalysts for sluggish oxygen ORR, theoretical approaches using DFT have provided comprehensive explanations for enhanced performance of new catalysts. Based on solid concepts in heterogeneous catalysis, we conducted high-throughput screening of nano catalysts for PEMFC and AEMFC application to find a new ORR electrocatalyst. In this study, we introduce a highly durable and active catalyst candidate and describe its chemistry on the corresponding surface by higher-level of calculations for verification.
Given that one has a proper size of the database for specific properties, prediction through machine learning can be a good choice. Here, using a large database from the screening performed above, we trained ANN to predict surface – adsorbate interaction to estimate the catalytic performance within an error range of < 0.2 eV. In this way, one can reduce computational cost significantly and broaden the screening window for materials exploration.
||Planned: Supplemental Proceedings volume