Quantum Chemistry via General Tensor Networks

Garnet Chan of the California Institute of Technology is supported by an award from the Chemical Theory, Models and Computational Methods program in the Division of Chemistry to develop new computational techniques to solve the equations of quantum mechanics. Some difficult problems of this kind can be found in the chemistry of molecules with transition metal atoms, because the electrons there behave in a concerted, correlated manner across different atoms.

Dr. Chan's main focus is to develop techniques to solve these classes of problems, and especially to create computational approximations that are “robust” in that they can be applied to many different problems in chemistry, without fear of them breaking down or providing an incorrect answer. His work draws on ideas from several disciplines, thus one focus of his broader work is to produce new educational material for the field that tackles different theoretical topics from the perspective of multiple disciplines.

Garnet Chan is pursuing two specific complementary intellectual directions. The first is to develop quantum chemical techniques to treat dynamical correlation in very large multi-reference problems, in order to bring the most complex electronic structure problems within the reach of quantitative 1 kcal/mol accuracy. The second is to forge a path toward electronic structure methods that remain robust regardless of interaction strength, sign structure, or single- or multi-reference nature.

This is based on foundational work to carry out three-dimensional tensor network simulations in realistic chemical problems. The ideas in the proposal, which draw on the areas of quantum chemistry, quantum information, and quantum circuits, will be condensed into online resources that will be material that can be used to train the next generation of quantum chemists.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.