Full-Dimensional Neural Network Potential Energy Surface for the Photodissociation Dynamics of HNCS in the S1 band.

Abstract

The full-dimensional potential energy surface (PES) for the photodissociation of HNCS in the S₁(¹A″) electronic state has been built up by the neural network method based on more than 48,000 ab initio points, which were calculated at the multireference configuration interaction level with Davidson correction using the augmented correlation consistent polarized valence triple-ζ basis set. It was found that two minima, namely, trans and cis isomers of HNCS, and seven stationary points exist on the S₁ PES for the three dissociation pathways: HNCS(S₁) → H + NCS/HNC + S(¹D)/HN(¹Δ) + CS(¹Σ⁺). The dissociation energies of two lowest product channels H + NCS and HNC + S(¹D) calculated on the PES are in good agreement with experimental results, validating the high accuracy of the PES. Furthermore, the quasi-classical trajectory calculations were carried out to investigate the photodissociation dynamics of HNCS(S₁) at the total energy ranging from 5.0 to 6.0 eV based on the newly constructed S₁ PES. It was found that two products H + NCS/HNC + S(¹D) are dynamically comparable with the branching ratios of ∼1:1 at high energies, and the product HNC + S(¹D) is favored at low energies, resulting from different topographies of the PES along the two dissociation pathways. Specifically, the translational energy distributions of the products H + NCS and HNC + S(¹D) were found to be exceedingly different. The former behaves like a Gauss-type function with a broad width and a center of the peak at relatively high energy, while the latter is dominated by the low energies and decays heavily as the translational energy increases, shedding light on the photodissociation dynamics of HNCS in the S₁ band.