Electrified solid-liquid interfaces

Electrochemical Energy Storage and Conversion

Shifting the energy supply towards renewables necessitates the development of energy storage systems, in particular based on catalytic and electrochemical transformations of materials at solid-liquid interfaces such as in batteries or fuel cells. An in-depth, atomic-level characterization of these highly relevant solid-liquid interfaces is crucial for the understanding and improvement of such energy conversion systems.

Predictive-quality quantum mechanical (QM) simulations can make most valuable contributions to this end, however atomistic modelling of electrified interfaces under realistic conditions is still a challenging task, considering the complex structure and behavior of liquids at interfaces. Key features that recent interface models strive to describe are (1) the structure of the solvent at the electrified interface, (2) the long-range double layer effects of the electrolyte and (3) the energetics within the thermodynamic ensemble relevant for the experimental situation. Evaluation of the kinetics of electrochemical reactions at such interfaces is still limited due to the underdeveloped methodology to study chemical and electrochemical reactions consistently (4).

In our group we are addressing these 4 key challenges by using explicit and continuum solvation models, both of which allow to include the electrode potential in a straight forward and consistent way in the simulations and in the energetics. 

 

Thermodynamics of electrified solid-liquid interfaces within continuum solvation models

The fully explicit quantum-mechanical treatment of the liquid environment is only affordable for some selected model systems - even when relying on computationally most efficient approaches like semi-local density-functional theory (DFT). Recently developed continuum solvation models coupled to an ab-initio cavity, however, have been shown to yield results in very good agreement with experiment [1,2].

Calculations with such implicit models are computationally significantly cheaper especially due to the reduced number of atoms treated quantum-mechanically, allowing ultimately to study a much bigger number of different systems with excellent predictive quality. In addition, computationally inexpensive, coarse-grained continuum models allow a straight-forward inclusion of the dielectric response of the solvent and space-charge-layer-related energy contributions of the electrolyte allowing the evaluation of grand-canonical interface energies. These calculations which include the electrode potential explicitly, enable an understanding of non-trivial potential and electrolyte effects e.g. potential-dependent adsorption geometries and non-trivial pH dependencies, previously inaccessible to simulation[3].

In general, the validation and testing of implicit models and other modelling approaches of electrochemical interfaces are still underdeveloped. We are addressing these issues at the moment by studying well defined model systems such as low index coinage metal electrodes and comparing results between different levels of theory and experiment. These efforts involve in particular also the combination of appropriate and efficient sampling methods for all interface degrees of freedom. Recently, we could show, that the absolute potential alignment of semiconductor levels in water, which is relevant for the correct gauge of electrode potentials in such simulations, can be obtained accurately with implicit-explicit hybrid models when the first solvent shell is treated explicitly, which gives us confidence in the accuracy of implicit water models.

 

[1] Hörmann, Andreussi, Marzari, J. Chem. Phys. 150, 041730 (2019)

[2] Hörmann, Guo, Ambrosio, Andreussi, Pasquarello, Marzari, npj Computational Materials 5:100 (2019)

[3] in preparation (2020)