Development of a Transferable Density-Functional Tight-Binding Model for Organic Molecules at the Water/Platinum Interface

Overview & Synopsis

A computationally efficient and transferable approach for modeling reactions at metal/water interfaces could significantly accelerate our understanding and ultimately development of new catalytic transformations, particularly in the context of the emerging field of biomass conversion. Here, we present a parametrization of Pt-X (X = H, O, C) density functional tight-binding (DFTB) for addressing this need. We firstly constructed Pt-H, Pt-O, and Pt-C repulsive potential splines. These pair-wise parameters were then augmented to include many-body interactions using the Chebyshev Interaction Model for Efficient Simulation (ChIMES). We compare the geometrical and energetic performances of both DFTB and DFTB/ChIMES methods to DFT reference data across a variety of organic molecules at platinum surface from nanoparticles to single-crystal surfaces. DFTB shows limited transferability between extended crystal surfaces to small nanoparticles. This transferability is significantly improved through the introduction of three-body interactions with Pt in DFTB/ChIMES, which provides consistent results across the various systems, with reductions in the RMSD from around 30 kcal/mol in DFTB to around 10 kcal/mol. We demonstrate the stability and reliability of the obtained parameters by performing metadynamics simulations for the adsorption of phenol on Pt(111). We observe that DFTB itself is undersolvating the surface, leading to only one or two chemisorbed water molecule in a c(4×6) unit cell. In contrast, DFTB/ChIMES leads to a coverage of about 0.5 ML and successfully captures the chemisorbed mode of phenol both at the solid/liquid and the solid/gas interface.Furthermore, in agreement with experimental measurements, the adsorption energy at the solid/liquid interface is significantly smaller compared to the solid/gas interface. Furthermore, we highlight that even with DFTB, where we can accumulate dynamics for more than 1 ns for a given system, the simulations are not fully converged. compared to gas-phase conditions. It is difficult to elucidate the geometric arrangements and elementary energetics at these interfaces via experimental techniques alone. 8 Therefore, computational models are largely exploited to achieve complementary and detailed insights.Atomic-scale simulations with explicit solvent are generally desired in order to gain access to relevant phenomena. However, liquid/solid interfaces require large systems and long phase-space samplings. Hence, methods that only partially allow for the description of bondbreaking and formation, such as molecular mechanics (MM) or hybrid quantum-molecular mechanics (QM/MM) with static QM regions are generally applied. 9-11 MM simulations are especially popular for the description of non-reactive interfaces, 12-16 while approximate QM/MM couplings are adapted for studying reactivity. 9,17-19 Still, QM/MM approaches suffer from three limitations: (i) Despite significant efforts over the last years, force fields for water/metal interfaces are of limited accuracy and/or transferability, 15,20,21 (ii) the impact of the explicit solvent on the adsorption structure is difficult to capture accurately and (iii) due to the delocalized electronic structure of metals it is intrinsically difficult to embed a metallic QM cluster in a MM region without major impacts on on the electronic structure of the QM region. Moreover, many-body effects are largely neglected by classical force fields, but have been found to account for 30-50% of the interaction energy between water and noble metals. 22,23 Furthermore, accounting for surface reconstructions and the reactivity of the solvent is not easily achieved with these methods. To address these challenges, the density functional tight-binding (DFTB) 24 method emerges as a promising candidate. DFTB, as an approximate DFT method, is in principle capable of capturing many-body effects and offers high computational efficiency with reasonable accuracy when compared to standard DFT. 25,26 Similar semi-empirical methods have recently been benchmarked for reactions occurring at the solid electrolyte interphase in Li-batteries, demonstrating that low-cost electronic structure methods are appropriate for investigating complex interfaces, where adequately sampling phase-space is too expensive at the DFT level of theory. 27 DFTB relies on minimal basis sets, parametrized two-center integrals and, in its standard