Electrocatalysis plays a key role in sustainable energy conversion and storage. Although tremendous efforts from the experimental side have been devoted to elucidating the reaction mechanism, the detailed reaction pathways are still controversial due to intrinsic difficulty of in situ spectroscopy under electrochemical conditions. Therefore, computational studies based on density functional theory (DFT) energetics serve as an important tool to clarify the reaction mechanism. However, several aspects such as solvation effects and the electrochemical potential effects are important for the electrochemical systems while such effects are often absent in the simulations. Moreover, current DFT exchange correlation functionals present certain qualitative and quantitative errors, while the combination of solvation treatments and the more advanced computational methods are not established. To address these concerns, this thesis work on two different levels, stressing on incorporating the necessary effects to model the electrochemical processes. At the DFT level, we model the complicated sulfur reduction reaction process on heteroatom doped holey graphene framework. Specifically, we elucidate the electrocatalytic origin of the improved battery performance with these catalysts and decipher the complex 16-electron process. At the more advanced many-body perturbation theory (MBPT) level, we focus on the random phase approximation (RPA), as a promising approach to address certain DFT errors such as the carbon monoxide (CO) adsorption puzzle: the commonly used functionals give incorrect prediction of the CO adsorption site and energy on transition metal catalysts, which is key for several catalytic processes including the industrial catalysis for methanol synthesis from synthesis gas, the water-gas shift reaction, and the electrochemical carbon dioxide reduction reaction. Nevertheless, the cost of RPA for surface systems is often unaffordable, and the combination of RPA with implicit solvation and further the grand canonical treatment of electrons to describe the electrochemical potential, is generally not established. In this thesis, to pave the way to further electrochemical applications using RPA, we exploit a k-space extrapolation scheme to reduce the cost for surface calculations. Then we further combine the RPA framework for electrified interfaces, including implicit solvation described using the linearized Poisson-Boltzmann equation and the grand canonical treatment of electrons. We show that the RPA results are qualitatively and quantitatively different from commonly used functionals and match better with the experimental results.
展开▼
