No abstract

The recent explosion of capabilities to fabricate nanostructured materials to atomic precision has opened many avenues for technological advances but has also posed unique questions regarding the identification of structures that should serve as targets for fabrication. One material class for which identifying such targets is challenging are transition‐metal crystalline surfaces, which enjoy wide application in heterogeneous catalysis. The high combinatorial complexity with which patterns can form on such surfaces calls for a rigorous design approach. In this article, we formalize the identification of the optimal periodic pattern of a metallic surface as an optimization problem, which can be addressed via established algorithms. We conduct extensive computational studies involving an array of crystallographic lattices and structure‐function relationships, validating patterns that were previously known to be promising but also revealing a number of new, nonintuitive designs. © 2016 American Institute of Chemical Engineers AIChE J, 62: 3250–3263, 2016

A novel methodology combining mathematical optimization with DFT calculations is developed to determine highly cohesive transition metal nanoclusters.

Advances in heterogeneous catalysis are being enabled by ever-growing control over the structure of surfaces at the atomic scale. To propose new defected catalytic surfaces, there is a need to systematically search through a combinatorial number of possible atom arrangements. Mathematical optimization is well suited to search this design space with algorithms that provide guarantees of optimality that are not possible with sampling algorithms. In this work, we develop several formulations for catalytic activity and stability as part of mathematical optimization models for designing nanostructured surfaces. Using the oxygen reduction reaction as an example, we show how optimization-based design can be used to efficiently explore the trade-off of activity against stability of surfaces. We furthermore demonstrate how nanostructuring can be applied to increase the expected activity of surfaces under various constraints on the coverage and overbinding of adsorbates. Our approach can be generally applied to other chemistries of interest and is suitably parameterized such that a wide array of systems can be considered, enabling evaluation of the sensitivity of different systems to constraints on the design space.