We devised a global optimization (GO) strategy for optimizing molecular properties with respect to both geometry and chemical composition. A relative index of thermodynamic stability (RITS) is introduced to allow meaningful energy comparisons between different chemical species. We use the RITS by itself, or in combination with another calculated property, to create an objective function F to be minimized. Including the RITS in the definition of F ensures that the solutions have some degree of thermodynamic stability. We illustrate how the GO strategy works with three test applications, with F calculated in the framework of Kohn-Sham Density Functional Theory (KS-DFT) with the Perdew-Burke-Ernzerhof exchange-correlation. First, we searched the composition and configuration space of C m H n N p O q ( m = 0–4, n = 0–10, p = 0–2, q = 0–2, and 2 ≤ m + n + p + q ≤ 12) for stable molecules. The GO discovered familiar molecules like N 2 , CO 2 , acetic acid, acetonitrile, ethane, and many others, after a small number (5000) of KS-DFT energy evaluations. Second, we carried out a GO of the geometry of Cu m Sn n + ( m = 1, 2 and n = 9–12). A single GO run produced the same low-energy structures found in an earlier study where each Cu m S n n + species had been optimized separately. Finally, we searched bimetallic clusters A m B n (3 ≤ m + n ≤ 6, A,B= Li, Na, Al, Cu, Ag, In, Sn, Pb) for species and configurations having a low RITS and large highest occupied Molecular Orbital (MO) to lowest unoccupied MO energy gap ( E g ). We found seven bimetallic clusters with E g 〉 1.5 eV.
Chemistry and Pharmacology