Even after decades of electronic structure theory development, accurate modeling of reactions at metallic surfaces remains challenging. The most widely used method, i.e., density functional theory, can yield quantitatively and qualitatively inaccurate descriptions of electronic structures and reaction kinetics, motivating the use of higher-level correlated wavefunction (CW) theories, which better capture electron correlation effects. Quantum embedding theories allow advanced CW methods to be applied to a local region that interacts with its environment through, for example, an embedding potential optimized within density functional embedding theory (DFET). To ensure the accuracy of embedded electronic structure calculations, the local region of interest in the presence of the optimized embedding potential must reproduce the behavior of the original full system. Here, we introduce a polarizable embedding scheme that couples an external local potential to provide attractive and repulsive interactions, i.e., similar to the foundation of DFET, with a dielectric bath to reproduce polarization in response to the electric field of the cluster. Particularly, the time-consuming step within DFET, i.e., the optimized effective potential process, is replaced by a physics-informed approach to generate the embedding potential, significantly reducing the computational cost. We evaluate our polarizable embedding scheme using the Cu(111) surface and confirm that our approach outperforms the standard DFET in predicting the Fermi level, charge states, and binding strength of multiple adsorbates. We anticipate that this more efficient and robust embedding scheme could accelerate the mainstream use of embedded correlated wavefunction theory in the heterogeneous catalysis community.
Building similarity graph...
Analyzing shared references across papers
Loading...
Lee et al. (Mon,) studied this question.
www.synapsesocial.com/papers/69ba43cb4e9516ffd37a55e8 — DOI: https://doi.org/10.1063/5.0323151
Kwanpyung Lee
Connor Fawcett
Qing Zhao
The Journal of Chemical Physics
Northeastern University
Building similarity graph...
Analyzing shared references across papers
Loading...