Files
Abstract
Developments in quantum chemistry are closely related to the rise of modern computational technologies. Application ofquantum chemistry methods to ever larger molecular systems has become possible through constant enhancement on software and hardware resources. In this work, we start by employing robust ab initio methods in the study of the atmospheric relevant Criegee intermediate. In this investigation, we found that uncertainty with respect to electron correlation is the limiting factor affecting the quality of computed quantities, such as enthalpies and barrier heights. Motivated by this observation, we demonstrate how externally corrected coupled cluster (ecCC) can be used to assess and recover strong correlation. In the proposed ecCC method, adaptive configuration interaction is used as the source of external corrections. Finally, we present Fermi.jl, a quantum chemistry framework written in Julia, a novel programming language increasingly popular in scientific computing. Julia builds upon modern software technology offering native efficiency along with a dynamical user and developer experience. We explore how its features can be leveraged in the design of computational chemistry algorithms. The feasibility of Fermi.jl is verified by benchmarks of popular wave function methods, where a competitive performance against traditional approaches to quantum chemistry is observed.