▎ 摘　　要

The catalytic decomposition of formic acid was studied on a series of candidate catalysts using density functional theory calculations. The candidate catalysts were modelled by a single Pd atom embedded in mono- and divacancy in graphene (Pd-1m-G vs. Pd-1d-G), as well as a Pd dimer embedded in di-, tri-, and quadrivacancy in graphene (Pd-2d-G vs. Pd-2t-G vs. Pd-2q-G). These catalysts can effectively and selectively catalyze formic acid dehydrogenation into hydrogen. Pd-2d-G is the most favorable catalyst among the five models with the rate-determining step energy barrier of only 0.68 eV, which is comparable to one of the most active catalysts i.e., Pd(111). Pd-1d-G is comparatively less active, with the rate-determining step energy barrier of 0.90 eV. The rest of the three models, i.e., Pd-1m-G, Pd-2t-G and Pd-2q-G, have energy barriers of 1.26, 1.12 and 1.06 eV, respectively. The model catalysts studied in this work are promising for reducing usage of the precious and rare metal Pd compared with Pd bulk catalysts. Additionally, unlike Pd nanoparticle catalysts, the model catalysts in this work clarify the catalytic mechanism. (C) 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.