Abstract: Accurate barriers for rate controlling elementary reactions on metal surfaces are key to understanding, controlling, and predicting the rate of heterogeneously catalyzed processes. While barrier heights for gas phase reactions have been extensively benchmarked, dissociative chemisorption barriers for the reactions of molecules on metal surfaces have received much less attention. The first database called SBH10 and containing 10 entries was recently constructed based on the specific reaction parameter approach to density functional theory (SRP-DFT) and experimental results. We have now constructed a new and improved database (SBH17) containing 17 entries based on SRP-DFT and experiments. For this new SBH17 benchmark study, we have tested three algorithms (high, medium, and light) for calculating barrier heights for dissociative chemisorption on metals, which we have named for the amount of computational effort involved in their use. We test the performance of 14 density functionals at the GGA, GGA+vdW-DF, and meta-GGA rungs. Our results show that, in contrast with the previous SBH10 study where the BEEF-vdW-DF2 functional seemed to be most accurate, the workhorse functional PBE and the MS2 density functional are the most accurate of the GGA and meta-GGA functionals tested. Of the GGA+vdW functionals tested, the SRP32-vdW-DF1 functional is the most accurate. Additionally, we found that the medium algorithm is accurate enough for assessing the performance of the density functionals tested, while it avoids geometry optimizations of minimum barrier geometries for each density functional tested. The medium algorithm does require metal lattice constants and interlayer distances that are optimized separately for each functional. While these are avoided in the light algorithm, this algorithm is found not to give a reliable description of functional performance. The combination of relative ease of use and demonstrated reliability of the medium algorithm will likely pave the way for incorporation of the SBH17 database in larger databases used for testing new density functionals and electronic structure methods.

Abstract: The hyperdynamics method is a powerful tool to simulate slow processes at the atomic level. However, the construction of an optimal hyperdynamics potential is a task that is far from trivial. Here, we propose a generally applicable implementation of the hyperdynamics algorithm, borrowing two concepts from metadynamics. First, the use of a collective variable (CV) to represent the accelerated dynamics gives the method a very large flexibility and simplicity. Second, a metadynamics procedure can be used to construct a suitable history-dependent bias potential on-the-fly, effectively turning the algorithm into a self-learning accelerated molecular dynamics method. This collective variable-driven hyperdynamics (CVHD) method has a modular design: both the local system properties on which the bias is based, as well as the characteristics of the biasing method itself, can be chosen to match the needs of the considered system. As a result, system-specific details are abstracted from the biasing algorithm itself, making it extremely versatile and transparent. The method is tested on three model systems: diffusion on the Cu(001) surface and nickel-catalyzed methane decomposition, as examples of reactive processes with a bond-length-based CV, and the folding of a long polymer-like chain, using a set of dihedral angles as a CV. Boost factors up to 109, corresponding to a time scale of seconds, could be obtained while still accurately reproducing correct dynamics.