Screening of Metal Catalysts for CO2 Conversion via Machine Learning and Molecular Simulations

Aug 5, 2026

This study improves catalyst discovery for converting CO2 to methane by evaluating earth-abundant metal catalysts using machine learning (ML) and molecular dynamics simulations. CO2 binding energies were computed across 61 metals, identifying nickel (Ni) as the strongest binder, alongside copper (Cu) and carbon-related behavior. Multiple ML regressors used easily accessible metal properties and simulation-derived features to predict binding energies. Models such as RANSACRegressor and SVR achieved high accuracy (R-squared 0.99) with RMSE 0.18–0.40. Density emerged as a leading structural feature influencing binding energy, enabling a high-throughput approach for sustainable catalyst selection.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages10

Reading Time25 min

AccessFree to download

Introduction

Sustainability-driven CO2-to-methane catalysis

Earth-abundant metal catalysts and their advantages

Computational acceleration and the role of ML/HTS

Screening of Metal Catalysts for CO2 Conversion via Machine Learning and Molecular Simulations Felix Otieno Okello a, Manda Timothy a , Livingstone Ochilo a, Fredrick Okumu a , Solomon Omwoma a , Denis Magero b , Anthony Pembere a* a Department of Physical Science, Jaramogi Oginga Odinga University of Science and Technology, P.O. Box 210, Bondo, Kenya bAlupe University P.O. Box 845 Busia-Kenya 50400 | A R T I C L E I N F O | | A B ST RA C T | | --- | --- | --- | | Article history:
Received: 24 January 2025
Revised: 11 March 2025
Accepted: 20 March 2025
Published online: 15 April 2025 | | This study's primary objective is to improve catalyst discovery by assessing earth-abundant metal catalysts for the conversion of CO2 to methane through the use of machine learning (ML) and molecular dynamics (MD) simulations. The highest CO2 binding energy on 61 metals was determined to be-9.75 eV for nickel (Ni), -8.7 eV for copper (Cu), and-7.75 eV for carbon (C) . Various ML models were developed to predict binding energies on the metallic surfaces. Easily accessible properties of the metals and features obtained from molecular simulations were used as input features. RANSACRegressor, LinearSVR, HuberRegressor, OrthogonalMatchingPursuit CV, and LarsCV models exhibited high prediction accuracy with R-squared values of 0.99 and RMSE ranging from 0.18 to 0.40. Feature significance analysis revealed that density (D) is among the most significant structural features affecting binding energy. This work offers a dependable, high-throughput method for identifying efficient CO2 conversion catalysts, advancing sustainable technologies. | | Keywords:
Metal Catalysts;
CO2 Conversion;
Machine Learning;
Molecular Simulations. | | | 1. Introduction The pressing need for sustainable energy solutions to mitigate climate change has intensified research efforts towards efficient conversion of CO2 into valuable fuels and chemicals. Among various conversion pathways, the catalytic conversion of CO2 to CH4 holds immense promise due to the abundance of CO2 and the high energy density of CH4 as a clean fuel[1]. However, the development of efficient catalysts for this reaction remains a significant challenge, primarily due to the complex interplay of reaction kinetics, selectivity, and catalyst stability[2] . Earth-abundant metal catalysts present a compelling avenue for sustainable CO 2 conversion, offering cost-effectiveness and scalability compared to precious metal counterparts. Nonetheless, the identification of optimal catalyst candidates from the vast chemical space remains a formidable task. Traditional experimental screening methods are time-consuming and resource-intensive, motivating the integration of computational techniques to accelerate catalyst discovery. Metal catalysts hold a paramount position in catalysis research due to their diverse chemical properties, tunable reactivity, and widespread applicability in a plethora of industrial processes[3] . In the context of CO2 conversion to methane, metal catalysts offer several distinct advantages that make them indispensable for this catalytic transformation. Firstly, metals exhibit a wide range of oxidation states, allowing for facile redox reactions involved in CO2 activation and subsequent methane formation[4] . The ability of metals to readily switch between different oxidation states enables efficient catalytic cycles, facilitating the conversion of CO2 to methane under mild reaction conditions. Secondly, metal catalysts possess high surface area-to-volume ratios, providing ample active sites for CO2 adsorption and activation[5, 6] . Moreover, metal catalysts exhibit tunable electronic properties, enabling modulation of the energetics of key reaction intermediates involved in CO2 conversion[7] . By controlling the electronic structure of the catalyst surface, it is possible to enhance the binding affinity of CO 2 and facilitate its subsequent reductio

FAQs

What is the main goal of this study?

To improve discovery of efficient catalysts for converting CO2 to methane by using machine learning together with molecular dynamics simulations.

How were CO2 binding energies evaluated in the screening?

CO2 binding energies were determined for 61 metals, and the strongest binders were identified to guide catalyst selection.

Which ML models and features were used to predict binding energies?

Several regression models (including RANSACRegressor and LinearSVR) were trained using metal properties and features derived from molecular simulations; density was found to be among the most significant structural features.

Aug 5, 2026

Upload by Genevieve