A Computational Toolkit to Understand Orbital Overlap and Chemical Reactivity

Abstract

Computational chemistry has become a standard tool for investigations in all branches of chemistry. Visualizing and interpreting electronic structure simulations, and predicting the chemical reactivity requires interpretative tools. However, in several cases, the widely used tools such as electrostatic potential (ESP) and partial atomic charges (Q) provide inconclusive information. Similarly, use of frontier molecular orbitals as interpretive tools is limited to visualize one orbital at a time. A comprehensive understanding of reactivity requires information about the nature of all occupied orbitals because the majority of chemical reactions are controlled by the synergy of electrostatics and orbital overlap effects. Orbital overlap distance, D(r), complements the ESP maps by quantifying the size of “test orbital” that maximizes its overlaps with a system’s occupied orbitals at a point r. Compact orbitals tend to have smaller values of D(r), as compared to the diffuse orbitals. We applied the combination of ESP and D(r) surface maps to rationalize the binding of ligands and metal ions to proteins and extended their applications to medicinal chemistry by quantifying the chemistry of promiscuous binders and predicting centromere-associated protein E inhibitors. We used this combination to distinguish the relative nature of carbon atoms at the defect sites of graphene sheet and to visualize s-holes on molecules of group IV to VII elements and transition metal nanoclusters. Our studies established that quantitative analysis of molecular ESP and D(r) surfaces can predict binding energies of s-holes interactions, acid-base binding affinities, stability constants, and interactions of metal ions to graphene defects. We used this quantitative analysis to predict values of empirical solvent softness scales, develop a solvent versatility scale, and model the Marcus µ-values of ionic liquids. Our second tool, atomic average overlap distance, D_A, distinguishes the compact, chemically stable atoms from chemically soft and unstable atoms. We used Q_A and D_A to capture trends in aromaticity, nucleophilicity, allotrope stability, and substituent effects. We used the mother tool, EDR(r;d), to analyze the stretched and compressed bonds. EDR(r;d) captures aspects of fractional occupancy and left-right correlation in stretched covalent bonds. This dissertation introduces the diverse applications of this toolkit to the major fields of chemical sciences.