Machine Learning Predictions
of Adsorption Energies of CH 4 -Related
Species
Takashi Toyao, Ichigaku Takigawa, and Ken-ichi Shimizu
Abstract Recent developments in data science could greatly impact catalysis
research in both industry and academia. Machine learning (ML) can play a central
role in this paradigm shift. Here, we describe a simple and efficient ML approach
for predicting the adsorption energies of CH 4 -related species, namely CH 3 , CH 2 ,
CH, C, and H, on Cu-based alloys. The developed ML model with 12 descriptors,
which are readily available for the selected elements, is shown to predict adsorption
energies obtained by using density functional theory (DFT)-based calculations. The
accuracy and simplicity of the developed system suggest that adsorption energies
can be readily predicted without time-consuming DFT calculations. This system
will eventually allow the prediction of the catalytic performance of solid catalysts.
Keywords Catalysis informatics · Machine learning · Density functional theory
calculations · Brønsted–Evans–Polanyi (BEP) relation
T. Toyao (B) · K. Shimizu
Institute for Catalysis, Hokkaido University, N-21, W-10, Sapporo 001-0021, Japan
e-mail: toyao@cat.hokudai.ac.jp
K. Shimizu
e-mail: kshimizu@cat.hokudai.ac.jp
Elements Strategy Initiative for Catalysts and Batteries, Kyoto University, Katsura 615-8520,
Kyoto, Japan
I. Takigawa
RIKEN Center for Advanced Intelligence Project, 1-4-1 Nihonbashi, Chuo-ku, Tokyo 103-0027,
Japan
e-mail: ichigaku.takigawa@riken.jp
Institute for Chemical Reaction Design and Discovery (WPI-ICReDD), Hokkaido University
Kita, 21 Nishi 10, Kita-ku, Sapporo 001-0021, Hokkaido, Japan
© Springer Nature Singapore Pte Ltd. 2020
K. Yoshizawa (ed.), Direct Hydroxylation of Methane,
https://doi.org/10.1007/978-981-15-6986-9_7
135