02:00 PM - 03:00 PM
Sophya Garashchuk, University of South Carolina, Columbia
Joint Institute for Computational Sciences Seminar
Research Office Building (5700), Room F-234
Email: Jacek JakowskiPhone:
The classical dynamics of nuclei is adequate in many situations (high-temperature, high-energy processes), providing insight into chemical processes. Yet it is well-known that quantum features of nuclear behavior -- the zero-point energy, tunneling, and nonadiabatic dynamics -- are sometimes important in low-energy reactions and in photochemistry. We are interested in the regime when quantum-mechanical (QM) behavior of nuclei of a few selected bonds modestly affects reactivity while the full QM treatment is unfeasible due to exponential scaling of numerical cost with the system size for the conventional methods. To make qualitative predictions and cheap estimates of the nuclear QM effects we are developing approximate dynamics based on the quantum trajectory (QT) formulation of the Schrödinger equation. The QM effects are incorporated through the quantum potential, computed in the “mean-field” approximation, acting on the trajectory ensemble in addition to the classical potential. Large molecular systems are described in a mixed quantum/classical QT framework with the QM correction incorporated into selected degrees of freedom. The approach is energy-conserving; full-dimensional wave function can be recovered from dynamics of multiple QT “sub-ensembles." The approximate QT dynamics is combined with the Density Functional Tight Binding to compute the electronic structure on-the-fly for systems of up to 200 atoms. The approach is applied to study adsorption of quantum hydrogen colliding with the graphene model, C37H15 (which is allowed to move). Localization of the proton wave function and details of its coupling to C37H15 have large effect on reactivity.