Automated Learning of a Dense Manifold of Electronic States and Electronic Energy Transfer and Reactions in Singlet O Collisions with N2

Calculations of collisions involving excited electronic states play an important role in many high-energy environments, for example in simulating thermal energy content and heat flux in flows around hypersonic re-entry vehicles, and useful data is usually not available from either experiment or theory. In this work, compatibilization by deep neural network (CDNN) – an automatic coupled potential energy surface (PESs) learning method – is used to discover and fit an underlying adiabatic-equivalent compatible potential energy matrix (CPEM) for singlet oxygen collisions with N2 in the singlet A′ manifold of N2O. The procedure yields not only a fit to the CPEM and its gradient but also analytic representations of the adiabatic surfaces, and their gradients. The problem is challenging because we consider high-energy collisions involving a 13-state dense manifold of electronic states. Using the resulting representation of the PESs and their analytic gradients, we calculated electronically nonadiabatic cross sections for N2(X) + O(singlet S) collisions for various initial conditions by using a new asymptotically extended formulation of the curvature-driven coherent switching with decay of mixing (CSDM) semiclassical dynamics method, which needs only the adiabatic potential energy surfaces to compute the coupling between electronic states and resolves the conflict between differing symmetries of the interacting atom-diatom system and the completely separated final states. This application also opens the way for treating other difficult problems involving electronic energy transfer and reactions of electronically excited species at high energy for various applications in chemistry, physics, materials, and engineering.