Click the poster below for some information about our research group thrusts, namely: (i) The computational investigation of amorphous and disordered materials for their use as catalysts and catalyst supports for alternative fuel production and pollution reduction; (ii) Gaining a deeper understanding of catalytic nanoclusters and atomically dispersed metal-complexes supported by metal oxides for natural gas conversion; (iii) Homogeneous organometallic catalysis for specialty chemical production, especially for C-H activation of small molecules; and (iv) Developing and applying data analytics/machine learning/data mining tools to accelerate discovery of catalysts and materials.

In collaboration with Susannah L. Scott's Group, B. Goldsmith and co-workers conducted a joint computational-experimental study of the reaction of CH₃ReO₃ (MTO) with H₂O₂ to understand the origins of large discrepancies in previously reported experimental reaction kinetics and thermodynamics compared with computational results. We also explored MTO-catalyzed olefin epoxidation by H₂O₂, as it shows strong water acceleration effects, even though no step in the catalytic cycle explicitly consumes water. The main cause of the observed acceleration is the water-dependence of the rates at which the active species are regenerated. Read here and here.

Reactant-induced structural changes of supported metal nanoparticles (NPs) have been widely reported during heterogeneous catalysis. One common structural change is the reactant-induced disintegration of the supported NPs, which could lead to catalyst deactivation or be employed as an effective way to achieve catalyst redispersion. In collaboration with the Wei-Xue Li Group, we conducted an ab initio thermodynamic study to understand the effects of CO and NO reactants on the disintegration of metal-oxide supported Rh, Pd, and Pt NPs into adatom-reactant complexes under a variety of experimentally relevant conditions. Read article here.

Modeling isolated sites on amorphous catalyst supports remains a major challenge. Typical strategies use cluster models with arbitrarily chosen constraints to model the rigid solid which impart arbitrary properties to the site. Alternatively no constraints are used, which results in sites with unrealistic flexibility. During Dr. Goldsmith's PhD studies in the Peters Group, they developed a systematic ab initio method to model isolated active sites on insulating amorphous supports using small cluster models. The goal is to use the algorithm to facilitate the testing of catalytic mechanistic hypotheses. Read article here.

As part of the Novel Materials Discovery Laboratory, a major goal is to develop data analytics tools to uncover scientific insights from large materials repositories. While a postdoctoral fellow at the FHI Theory Department, Dr. Goldsmith applied subgroup discovery to find and describe interesting local patterns in materials-science data. In collaboration with Dr. Boley, two illustrative examples were considered to: (1) discover interpretable models that classify the octet binary materials as either zincblende or rocksalt, and (2) elucidate structure-property relationships of gold clusters in the gas phase. Read here and here.

With the L. M. Ghiringhelli Group, we are examining the (meta)stable structures of gold clusters present at finite temperature using van der Waals (vdW) corrected density-functional theory and replica-exchange ab initio molecular dynamics. Inclusion of many-body vdW interactions is needed for predicting accurate isomer energetics, and its importance grows as the cluster size increases. Temperature effects are observed to typically stabilize three-dimensional structures over planar structures at finite temperature. Gold cluster structures are assigned using far-IR spectroscopy obtained by the Fieleke Group and theoretical predictions.