Prediction of Transition-State Energies of Hydrodeoxygenation Reactions on Transition-Metal Surfaces Based on Machine Learning
Abstract
Computational catalyst discovery involves identification of a meaningful model and suitable descriptors that determine the catalyst properties. We study the impact of combining various descriptors (e.g., reaction energies, metal descriptors, and bond counts) for modeling transition-state energies (TS) based on a database of adsorption and TS energies across transition-metal surfaces for the decarboxylation and decarbonylation of propionic acid, a chemistry characteristic for biomass conversion. Results of different machine learning models for more than 1572 descriptor combinations suggest that there is no statistically significant difference between linear and nonlinear models when using the right combination of reactant energies, metal descriptors, and bond counts. However, linear models are inferior when not including bond count and metal descriptors. In conclusion, when there are missing data for reaction steps on all metals, conventional linear scaling is inferior to linear and nonlinear models with proper choice of descriptors that are surprisingly robust.
- Authors:
-
- University of South Carolina, Columbia, SC (United States)
- University of North Carolina, Charlotte, NC (United States)
- Publication Date:
- Research Org.:
- Univ. of South Carolina, Columbia, SC (United States)
- Sponsoring Org.:
- USDOE Office of Science (SC), Basic Energy Sciences (BES); National Science Foundation (NSF)
- OSTI Identifier:
- 1656918
- Grant/Contract Number:
- SC0007167; EPSCoR-1632824; DMREF-1534260
- Resource Type:
- Accepted Manuscript
- Journal Name:
- Journal of Physical Chemistry. C
- Additional Journal Information:
- Journal Volume: 123; Journal Issue: 49; Journal ID: ISSN 1932-7447
- Publisher:
- American Chemical Society
- Country of Publication:
- United States
- Language:
- English
- Subject:
- 37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CHEMISTRY; Free energy; Metals; Adsorption; Chemical reactions; Energy
Citation Formats
Abdelfatah, Kareem, Yang, Wenqiang, Solomon, Rajadurai Vijay, Rajbanshi, Biplab, Chowdhury, Asif, Zare, Mehdi, Kundu, Subrata Kumar, Yonge, Adam, Heyden, Andreas, and Terejanu, Gabriel. Prediction of Transition-State Energies of Hydrodeoxygenation Reactions on Transition-Metal Surfaces Based on Machine Learning. United States: N. p., 2019.
Web. doi:10.1021/acs.jpcc.9b10507.
Abdelfatah, Kareem, Yang, Wenqiang, Solomon, Rajadurai Vijay, Rajbanshi, Biplab, Chowdhury, Asif, Zare, Mehdi, Kundu, Subrata Kumar, Yonge, Adam, Heyden, Andreas, & Terejanu, Gabriel. Prediction of Transition-State Energies of Hydrodeoxygenation Reactions on Transition-Metal Surfaces Based on Machine Learning. United States. https://doi.org/10.1021/acs.jpcc.9b10507
Abdelfatah, Kareem, Yang, Wenqiang, Solomon, Rajadurai Vijay, Rajbanshi, Biplab, Chowdhury, Asif, Zare, Mehdi, Kundu, Subrata Kumar, Yonge, Adam, Heyden, Andreas, and Terejanu, Gabriel. Thu .
"Prediction of Transition-State Energies of Hydrodeoxygenation Reactions on Transition-Metal Surfaces Based on Machine Learning". United States. https://doi.org/10.1021/acs.jpcc.9b10507. https://www.osti.gov/servlets/purl/1656918.
@article{osti_1656918,
title = {Prediction of Transition-State Energies of Hydrodeoxygenation Reactions on Transition-Metal Surfaces Based on Machine Learning},
author = {Abdelfatah, Kareem and Yang, Wenqiang and Solomon, Rajadurai Vijay and Rajbanshi, Biplab and Chowdhury, Asif and Zare, Mehdi and Kundu, Subrata Kumar and Yonge, Adam and Heyden, Andreas and Terejanu, Gabriel},
abstractNote = {Computational catalyst discovery involves identification of a meaningful model and suitable descriptors that determine the catalyst properties. We study the impact of combining various descriptors (e.g., reaction energies, metal descriptors, and bond counts) for modeling transition-state energies (TS) based on a database of adsorption and TS energies across transition-metal surfaces for the decarboxylation and decarbonylation of propionic acid, a chemistry characteristic for biomass conversion. Results of different machine learning models for more than 1572 descriptor combinations suggest that there is no statistically significant difference between linear and nonlinear models when using the right combination of reactant energies, metal descriptors, and bond counts. However, linear models are inferior when not including bond count and metal descriptors. In conclusion, when there are missing data for reaction steps on all metals, conventional linear scaling is inferior to linear and nonlinear models with proper choice of descriptors that are surprisingly robust.},
doi = {10.1021/acs.jpcc.9b10507},
journal = {Journal of Physical Chemistry. C},
number = 49,
volume = 123,
place = {United States},
year = {2019},
month = {11}
}
Web of Science