(281d) Comparison of Solvation Models for Metal Catalysis

Heterogeneous catalysis involving liquid phase is an important field of research, for instance in biomass conversion, oxidation reactions, and electrocatalysis. Yet, typically density functional theory (DFT) studies ignore the liquid phase and focus on the solid surface/gas phase interface. Advances in computational power and theory currently allow consideration of surrounding solvent for complex systems such as metal surfaces typical in catalysis. Of the two general approaches, implicit and explicit solvation, the former implicit approach is generally much more computationally tractable since the solvent is treated as a continuum. The efficacy of such models, however, is unclear in modeling relevant surface chemistry. 

 We have undertaken a systematic study of various solvation models in order to identify the accuracy and efficacy of such models. We have examined implicit solvation models (PCM-based models and COSMO) and compare with explicit solvation models (e.g. water layer above metal surface). Specifically, various DFT codes that have implemented periodic boundary conditions and implicit solvation models (i.e. VASP[1], jDFTx[2]) were used, as well as cluster models (using NWChem[3] in a similar manner to the iSMS method[4]). We have modeled adsorption of a large number of species, both organic and inorganic, over typical metal surfaces (Pt and Pd) in order to obtain statistical data comparing these various solvation methods. Our results show that for many species solvation effects change adsorption energies very little (less than 0.2 eV). We have also modeled several reactions, including the oxygen reduction reaction (ORR), water-gas shift reaction, and oxidation of formic acid, methanol, and ethanol. Our results for instance predict that the ORR energetics can change by up to 0.5 eV due to solvation effects when using implicit solvation models.  Finally, the bulk of our results were obtained with water solvent (polar protic), but we have also considered other liquids, such acetonitrile (polar aprotic) and n-octane (non-polar).

 Our results provide a comparison and framework for the various solvation models and suggest which solvation models may be appropriate and the limits of such models for metal surface chemistry.

[1] K. Mathew, R. Sundararaman, K. Letchworth-Weaver, T. A. Arias, and R. G. Hennig, The Journal of Chemical Physics, 140, 084106 (2014)

[2] R. Sundararaman, D. Gunceler, K. Letchworth-Weaver and T.A. Arias, JDFTx, available from http://jdftx.sourceforge.net (2012)

[3] M. Valiev, E.J. Bylaska, N. Govind, K. Kowalski, T.P. Straatsma, H.J.J. van Dam, D. Wang, J. Nieplocha, E. Apra, T.L. Windus, and W.A. de Jong, Comput. Phys. Commun. 181, 1477 (2010)

[4] M. Faheem, S. Suthirakun, and A. Heyden, J. Phys. Chem. C. 116 (2012)