Concerted, Facile CO2 Hydrogenation in Ni-MOF-74 Revealed by Machine Learning Interatomic Potential-Based Enhanced Sampling Simulations
Abstract
Metal-organic frameworks (MOFs) with open metal sites are promising heterogeneous catalysts, but the atomistic and dynamic details that govern their reactivity are difficult to access experimentally due to spatiotemporal resolution limits. Computational studies, predominantly gas-phase density functional theory (DFT) calculations, have been used to study MOF-catalysed reactions, but typically rely on reduced cluster models of the metal center that neglect confinement effects or treat the framework as rigid during the reaction, limiting realistic modeling of catalysis at finite temperature and pressure. Here, we investigate, as a prototypical reaction, CO2 hydrogenation inside Ni-MOF-74 using machine learning interatomic potentials-based molecular dynamics and on-the-fly probability-based enhanced sampling (OPES) simulations, which delineate dynamic adsorption of CO2 and H2 in a flexible MOF and provide the underlying free energy for the CO2 reduction reaction. OPES simulations further reveal that H2 first adsorbs and undergoes heterolytic dissociation at the Ni-O site with a free energy barrier of ∼16 kcal mol−1, forming a proton and hydride species. Subsequent reduction of CO2 to HCOOH proceeds via a concerted mechanism, involving simultaneous hydride and proton transfer through a single transition state with a relatively low activation barrier of ∼6 kcal mol−1. These results demonstrate that explicit treatment of framework flexibility and thermodynamic conditions is essential to accurately capture reaction mechanisms and energetics in porous catalysts. Our integrated MLIP-OPES approach provides a generalizable computational framework for studying complex chemical transformations under confinement in MOFs and related framework materials.
Related
- Roles of oxygen vacancies and the Ni–support interface in CO2 hydrogenation over Ni/Y-doped BaZrO3 revealed by experimental and computational studies
- Accuracy, Reliability, and Efficiency of a MACE Machine-Learning Interatomic Potential for Aspirin: A Reproducible Benchmark for Molecular Simulation
- Transfer learning on universal interatomic potential embeddings improves generalization in structure-property defect models
- Mechanistic Basis of Sulfo Transfer in a PAPS-Independent ArylSulfate SulfoTransferases Revealed by Crystallography and QM/MM Simulations
- Hessian-Informed Conformation Augmentation for Potential Energy Surface Curvature in Machine Learning Interatomic Potentials
BibTeX
@misc{dhananjay2026concerted,
title = {{Concerted, Facile CO2 Hydrogenation in Ni-MOF-74 Revealed by Machine Learning Interatomic Potential-Based Enhanced Sampling Simulations}},
author = {Dhananjay and Nishtha Sharma and Tarak Karmakar},
howpublished = {ChemRxiv},
year = {2026},
month = sep,
doi = {10.26434/chemrxiv.15008834/v1},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008834/v1},
note = {CC BY-NC},
}
RIS
TY - UNPB
TI - Concerted, Facile CO2 Hydrogenation in Ni-MOF-74 Revealed by Machine Learning Interatomic Potential-Based Enhanced Sampling Simulations
AU - Dhananjay
AU - Nishtha Sharma
AU - Tarak Karmakar
PY - 2026
DA - 2026/09/15
PB - ChemRxiv
DO - 10.26434/chemrxiv.15008834/v1
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008834/v1
AB - Metal-organic frameworks (MOFs) with open metal sites are promising heterogeneous catalysts, but the atomistic and dynamic details that govern their reactivity are difficult to access experimentally due to spatiotemporal resolution limits. Computational studies, predominantly gas-phase density functional theory (DFT) calculations, have been used to study MOF-catalysed reactions, but typically rely on reduced cluster models of the metal center that neglect confinement effects or treat the framework as rigid during the reaction, limiting realistic modeling of catalysis at finite temperature and pressure. Here, we investigate, as a prototypical reaction, CO2 hydrogenation inside Ni-MOF-74 using machine learning interatomic potentials-based molecular dynamics and on-the-fly probability-based enhanced sampling (OPES) simulations, which delineate dynamic adsorption of CO2 and H2 in a flexible MOF and provide the underlying free energy for the CO2 reduction reaction. OPES simulations further reveal that H2 first adsorbs and undergoes heterolytic dissociation at the Ni-O site with a free energy barrier of ∼16 kcal mol−1, forming a proton and hydride species. Subsequent reduction of CO2 to HCOOH proceeds via a concerted mechanism, involving simultaneous hydride and proton transfer through a single transition state with a relatively low activation barrier of ∼6 kcal mol−1. These results demonstrate that explicit treatment of framework flexibility and thermodynamic conditions is essential to accurately capture reaction mechanisms and energetics in porous catalysts. Our integrated MLIP-OPES approach provides a generalizable computational framework for studying complex chemical transformations under confinement in MOFs and related framework materials.
ER -