American Chemical Society
jp2c07246_si_001.pdf (472.91 kB)

Strain in Catalysis: Rationalizing Material, Adsorbate, and Site Susceptibilities to Biaxial Lattice Strain

Download (472.91 kB)
journal contribution
posted on 2022-12-02, 14:38 authored by Cheng Zeng, Tuhina Adit Maark, Andrew A. Peterson
The binding strengths of reaction intermediates on a surface are often the principal descriptors of the effectiveness of heterogeneous catalysts. Although strain is a well-known theoretical strategy to modify binding strengths, and experimental methods have been introduced to directly induce strain, there is comparatively little systematic understanding of the binding energy susceptibilities of different adsorbates, materials, and surface sites to strain. In this work, we employ electronic structure calculations to develop such a systematic understanding. We utilize density functional theory calculations with 10 simple reaction intermediates adsorbed on four binding sites of five metal fcc(111) surfaces under an in-plane biaxial strain of ±2.0%. The responsiveness to strain is quantified using a single parameter named strain susceptibility, which we define as the slope of the adsorption energy versus strain. Typical values for this slope are in the tens of meV per unit percent strain. Based on these calculations, several general trends are identified. First, the material susceptibility order is found to be (Au, Pt) > Pd > (Ag, Cu), which we show can be correlated with the relative changes in d-band widths with strain. Second, binding sites with a higher degree of coordination to the adsorbate tend to exhibit a higher strain susceptibility. Third, adsorbates having higher valency tend to exhibit larger susceptibilities, and among adsorbates having the same valency, N- and O-containing adsorbates exhibit similar susceptibilities, but both show higher susceptibilities than that of C-containing adsorbates. The resulting changes in binding energy are compared to the linear scaling relations of adsorption and are found not to follow the published slopes, but rather to scale more closely with coordination number. This analysis can help to make predictions of which reactions are likely to respond favorably to strain and which catalysts may exhibit enhanced activity. Finally, an eigenforce model is used to rationalize the strain trends. The model-predicted susceptibilities show decent agreement with values by electronic structure calculations, differing by a mean absolute error of 0.013 eV/% for a variety of adsorption systems.