ENHANCEMENTS TO THE MODIFIED GENETIC ALGORITHM FOR CRYSTAL STRUCTURE PREDICTION

G. I. PAGOLA; J. C. FACELLI; A. LUND; M. B. FERRARO; A. M. ORENDT

BUENOS AIRES

Workshop; Primer Workshop Latinoamericano de Modelado Molecular y Simulación Computacional; 2016

CELBI y FCEN, UBA

Crystal structure prediction is an important field of study, both for the development of new compounds and materials, and for the advancement of understanding crystallization processes. The Modified Genetic Algorithm for Crystal Structure Prediction, MGAC, is a software package for structure prediction that has had varying success in predicting the structures of many molecules. However, several advancements in the field of structure prediction have prompted a revision to the software, both from a scientific and technical standpoint. In this dissertation, the evaluation of a new method for energy calculation and structural optimization, dispersion corrected density functional theory, is presented, along with practical parameterizations for using density functional theory in crystal structure prediction. Next, a preliminary implementation of MGAC using density functional theory is outlined, including some key changes to the construction of unit cells, along with successful prediction results for the molecules glycine and histamine. Finally, a new implementation of MGAC is proposed to handle multiple space group prediction effectively, with accompanying preliminary prediction results for histamine using the new implementation of MGAC, called MGAC2.