ABSTRACT The paper introduces a modular Python‐based program with a graphical user interface (GUI) to automate collisional dynamics of rigid‐rotors in cold and ultracold environments. The program streamlines the construction of ab initio potential energy surfaces (PES), with optional neural network (NN) augmentation, thereby reducing computational costs. The multipole expansion module generates radial terms for Molscat to compute inelastic cross‐sections and state‐to‐state rate coefficients. These rate coefficients, especially with hydrogen and helium, are crucial for determining column densities and modeling the abundances of species in the interstellar medium. In the paper, the rate coefficients of ‐He/ are calculated for the first 11 rotational states to evaluate the accuracy of the ab initio and NN methods. Finally, the rate propensities for , , , , and N linear chains () are calculated based on the benchmark results and delivered as an ensemble NN model to allow the direct prediction of rate coefficients based on the physical parameters of various carbonaceous species. The GUI‐automated PES generation/fitting accelerates collisional studies, ensures robustness and reproducibility, and significantly reduces data processing time from weeks to hours, thereby providing an efficient and general framework for accurate state‐to‐state rate predictions in astrochemical and molecular dynamics applications.
Kushwaha et al. (Thu,) studied this question.