Modeling van der Waals Interactions with Density Functional Theory

Abstract

Materials physics has been a driving force in human industry and technology for thousands of years. This is truer now than ever before, with the theory of quantum mechanics providing the first-ever accurate model of materials at the atomic scale. With accurate quantum mechanical modeling, a path is opened to address pressing needs in industry by developing novel “functional materials”, such as the porous metal--organic frameworks (MOFs). These materials show remarkable adsorption properties, with possible uses in gas storage, fuel refinement, catalysis, and many other fields. Many MOF species have been developed and synthesized, but their usefulness is limited without adequate theoretical description of their behavior. To this end, my work uses density functional theory (DFT) to characterize and predict the behavior of these functional materials. In particular, I use van der Waals density functional theory (vdW-DFT) to model the subtle, long-range forces that drive adsorption in MOFs. For example, we examine differences in adsorption energy between guest molecules, which can be used to separate gas mixtures. With careful modeling, we can identify whether this separability is a function of the MOF's unique structure—as we see with the ``gate opening'' behavior of RPM3–Zn—or whether it arises from a thermodynamic property of the adsorbate mixture. We observe the latter in the case of Ca(H2tcpb), solving a long-standing mystery of its temperature-dependent uptake of C6 isomers. Observing and correctly defining these mechanisms requires a high level of accuracy, which can not always be assumed for approximation methods such as DFT. For this reason, several of my projects have aimed to improve the accuracy of vdW-DFT methods. Two projects examine the effect of exchange—which derives from the Pauli exclusion principle—on binding energy in dispersion-dominated systems. Drawing on the lessons learned from these studies, we also design vdW-DF3-mc, a third-generation nonlocal functional optimized for molecular solids, which can be used to accurately describe thousands of industrially significant materials. These projects demonstrate how vdW-DFT can be used to accelerate progress in industry, and the many available avenues to improving the method's predictive power.