Université de Sherbrooke

Published purpose

One of the great successes of computational chemistry is the Density Functional Theory (DFT). By expressing the energy of a system in terms of a functional of the density instead of a wave function, the problem of determining the correct electronic structure for a given nuclei configuration becomes classically tractable. However, the exact form of this energy functional is unknown, leaving us with approximations which break down when electrons are strongly correlated. The design of better functionals is an active research field. The Project proposes to get around this difficulty by performing Hamiltonian simulation on quantum computers for a wide variety of nuclei configuration from which an accurate electronic density is extracted, total energy and pair correlation. By performing a regression on a database generated in this manner, a better approximation to the density functional could be learned. In turn, it would allow to use DFT for problems beyond the reach of the usual approximate functionals. Even when fault-tolerant quantum computers come into existence, one cannot expect them to be widely available. Therefore, the Project aims at gaining the most out of every computation; The recipients believe that applications that extend the abilities of classical computers are one of the best uses for quantum computers. Because of the considerable classical computation infrastructure already available, this project will allow the discovery of novel materials and molecules faster than exclusive use of quantum computers.