Free energies of adsorption on catalyst surfaces are the most elementary quantities in microkinetic modeling. Density functional theory (DFT) based approaches only provide access to the electronic components of these free energies, while finite-temperature contributions are usually approximated. Standard approximations include treating the adsorbate as a harmonic oscillator or a as two-dimensional ideal gas. In this work, we present an alternative approach to calculate the non-electronic contributions to the free energies using canonical partition functions computed through a sampling of the potential energy surface. This sampling is performed using neural network potentials trained on a set of DFT calculations of simple adsorbates on transition metal surfaces. Tradeoffs in computational cost and accuracy of the free energy estimates are discussed.

Adsorption free energies using neural network based potential energy sampling

254th National Meeting (2017)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.