New Implicit Solvation Scheme for Solid Surfaces Muhammad Faheem, Suwit Suthirakun, and Andreas Heyden* Department of Chemical Engineering, University of South Carolina, 301 South Main Street, Columbia, South Carolina 29208, United States *S Supporting Information ABSTRACT: It is shown that the effect of water on the bonding characteristics of transition metal surfaces with adsorbates is short-ranged. As a result, adsorption energies in water can be evaluated by a combination of plane-wave density functional theory calculations in vacuum and properly chosen cluster model calculations with and without an implicit solvation model. The scheme is demonstrated for a model C-C cleavage reaction on Pt (111) and for predicting CO frequency shifts on Pd and Pt due to water. We conclude that these shifts originate from water-metal interactions and can be explained by changes in π back-donation. Overall, the results demonstrate that the proposed methodology represents a highly efficient computational approach for approximating the effect of solvents on elementary reaction steps occurring at solid-liquid interfaces of heterogeneous catalysts. 1. INTRODUCTION Computational investigations of chemical reactions at solid- liquid interfaces pose a unique challenge of accurately yet efficiently accounting for the effect of the liquid-phase environment. Liquid molecules can affect the activity and selectivity of a catalyst by stabilizing or destabilizing adsorbed intermediates and transition states 1 and by providing low- energy pathways for reactions, e.g., for proton transfer between neighboring active sites. 2 Free energy differences and rates of elementary reaction steps occurring at solid-liquid interfaces are often very different from the same processes occurring at solid-gas interfaces. To correctly account for the effect of a liquid phase on reaction rates, the dynamic fluctuations in the complex liquid and the long-range electrostatic interactions of the liquid molecules must be considered, requiring the simulation of a large number of liquid molecules over a (computationally) long time period. As a result, the use of ab initio molecular dynamics (AIMD) approaches 3 for systematic investigations of such processes becomes for the foreseeable future computationally prohibitive. A common procedure for modeling, e.g., liquid water at solid-liquid interfaces, consists of optimizing a hexagonally closed-packed ice-like structure at the metal interface before replacing one of the water molecules with the reactant species. 4,5 Although significantly faster than AIMD, this approach includes no or very limited sampling of the water configuration space and is error-prone for relatively large adsorbates where it is difficult to identify a meaningful initial configuration of the water molecules. Alternatively, Jacob and Goddard 6 have pioneered the use of implicit continuum solvation models 7-9 on large metal clusters of (111) surface shape. While continuum solvation models cannot accurately describe site-specific interactions between the adsorbates and the surrounding solvent molecules, they are computationally fast and reasonably accurate for computing free energies in solution. Furthermore, solvents and reaction conditions such as temperature can easily be changed with modern implicit solvation models developed for molecular systems. 10 Unfortu- nately, relatively large metal clusters must be selected to describe the long-range metal interactions and to overcome unwanted boundary effects 6 due to the finite size of the cluster. 11 Similarly, implicit solvation models have been developed for periodic systems. However, most implementa- tions only consider electrostatic effects although nonelectro- static contributions are crucial for obtaining accurate solvation free energies. 12 Considering furthermore that the implementa- tion of smooth gradients of free energies has been progressing slowly in plane-wave density functional theory (DFT) codes, 13-15 it would be very beneficial if current nonperiodic implicit solvation models could be used to describe the effect of solvents on processes occurring at “periodic” solid-liquid interfaces. In this paper, we propose a simple but potentially very powerful new approach for modeling reactions at solid-liquid interfaces with implicit solvation models, which we call implicit solvation model for solid surfaces (iSMS). The objective of this theoretical study is to validate this procedure for a model C-C cleavage reaction in water and by comparing predicted CO frequency shifts in water to experimental data. This paper is organized as follows: After introducing the iSMS methodology and describing the computational details in section 2, we present in section 3 the convergence properties of iSMS methodology with respect to the size of the cluster model and the size of the basis set for the C-C cleavage reaction in double-dehydrogenated ethylene glycol on Pt (111) in water. Next, we use iSMS to calculate CO frequency shifts in water on Pd (111) and Pt (111) that can be directly compared to Received: August 17, 2012 Revised: September 18, 2012 Published: October 4, 2012 Article pubs.acs.org/JPCC © 2012 American Chemical Society 22458 dx.doi.org/10.1021/jp308212h | J. Phys. Chem. C 2012, 116, 22458-22462