Multiscale Sampling of a Heterogeneous Water/Metal Catalyst Interface using Density Functional Theory and Force-Field Molecular Dynamics

Abstract

A significant number of heterogeneously-catalyzed chemical processes occur under liquid conditions, but simulating catalyst function under such conditions is challenging when it is necessary to include the solvent molecules. The bond breaking and forming processes modeled in these systems necessitate the use of quantum chemical methods. Since molecules in the liquid phase are under constant thermal motion, simulations must also include configurational sampling. This means that multiple configurations of liquid molecules must be simulated for each catalytic species of interest. The goal of the protocol presented here is to generate and sample trajectories of configurations of liquid water molecules around catalytic species on flat transition metal surfaces in a way that balances chemical accuracy with computational expense. Specifically, force field molecular dynamics (FFMD) simulations are used to generate configurations of liquid molecules that can subsequently be used in quantum mechanics-based methods such as density functional theory or ab initio molecular dynamics. To illustrate this, in this manuscript, the protocol is used for catalytic intermediates that could be involved in the pathway for the decomposition of glycerol (C 3 H 8 O 3). The structures that are generated using FFMD are modeled in DFT in order to estimate the enthalpies of solvation of the catalytic species and to identify how H 2 O molecules participate in catalytic decompositions.

Influence of an electrified interface on the entropy and energy of solvation of methanol oxidation intermediates on platinum(111) under explicit solvation

“…61 All cMD simulations are carried out at 300 K. In all cMD simulations, H 2 O molecules are allowed to move, while the positions of Pt atoms and surface species atoms are held fixed. 54 All simulations are carried out similar to our prior work. 28,53,54 Briefly, H 2 O molecules are added to cells #1 and #2 using the MCPliQ code.…”

“…54 All simulations are carried out similar to our prior work. 28,53,54 Briefly, H 2 O molecules are added to cells #1 and #2 using the MCPliQ code. 54 H 2 O molecule positions are then refined in a 0.5 ns cMD simulation performed in the microcanonical (NVE) ensemble to check for energy conservation.…”

“…28,53,54 Briefly, H 2 O molecules are added to cells #1 and #2 using the MCPliQ code. 54 H 2 O molecule positions are then refined in a 0.5 ns cMD simulation performed in the microcanonical (NVE) ensemble to check for energy conservation. Next, a cMD simulation is performed in the isothermal-isobaric (NPT) ensemble for a total of 5 ns to achieve the appropriate water density.…”

“…Finally, simulations are performed in the NVT ensemble in order to generate configurations of H Configurations are sampled every 300 ps (resulting in 10 total configurations), as in prior work. 54 Vacuum space is added to the top of the water column (see Section 2.1), and then the ensemble average is calculated by evaluating eqn (2 O molecule -OH bonds make with the surface normal (y 1 and y 2 ; see Fig. 2 for an example).…”

“…35 The aqueous environment also influences the energies of surface species, and this phenomenon is more pronounced when the surface species form hydrogen bonds with liquid water. 28,[52][53][54] However, understanding the interplay between an applied electric field and solvent effects and how this interplay influences the free energies of surface species remains unknown, 55 in part because of the need to simultaneously apply explicit liquid solvent and an electric field.…”