“Catalysis by design” has been a dream for decades. To specify the composition and structure of matter to effect a desired catalytic transformation with desired and predicted rate and selectivity remains a monumental challenge, especially in heterogeneous catalysis. With the advent of surface science techniques in decades past, the promise was perceived of turning increased molecular level understanding of reaction mechanisms and surface sites into principles of catalyst design. Surface science alone has not proven to be sufficient for this purpose. Over the past decade the rise of powerful, computationally efficient theoretical methods has shown promise, not just for identifying catalytic intermediates and reaction pathways accessible to experiments, but of providing quantitative predictions of energetics for elementary reaction processes not easily accessed experimentally.

Much of our work is aimed at the rational design of catalysts for direct epoxidation of olefins. This chemistry remains one of the most challenging problems in heterogeneous catalysis. Although the epoxidation of ethylene by silver catalysts to form ethylene oxide (EO) has been practiced for decades, little progress has been made in expanding this technology to other products and processes. We have made significant progress through the combination of surface science experiments, Density Functional Theory (DFT) calculations, and catalytic reactor experiments, toward understanding the mechanism of this reaction on silver catalysts, and to the rational improvement of selectivity. The key has been the demonstration of surface oxametallacycle intermediates as the species that control reaction selectivity. This discovery permits the influence of catalyst promoters on selectivity to be probed, and new catalyst formulations to be developed. We have proposed a new mechanism of catalyst promotion for this system by alkalis based on surface ion-dipole interactions. We have carried out experimental studies of bimetallic catalysts, e.g., Cu/Ag, predicted by DFT to offer enhanced selectivity for ethylene epoxidation. The Cu/Ag bimetallic catalysts formulated to embody our design model achieve much higher selectivities to ethylene oxide than Ag catalysts for a wide range of conditions. This example shows how the molecular level understanding of catalytic reactions from fundamental studies can drive the design of new catalytic materials. It represents one of the first examples of catalyst design for improved selectivity based on first principles approaches.