Full-Dimensional Global Potential Energy Surface for the KRb + KRb → K2Rb2* → K2 + Rb2 Reaction with Accurate Long-Range Interactions and Quantum Statistical Calculation of the Product State Distribution under Ultracold Conditions

A full-dimensional global potential energy surface (PES) for the KRb + KRb → K2Rb2* → K2 + Rb2 reaction is reported based on high-level ab initio calculations. The short-range part of the PES is fit with the permutationally invariant polynomial-neural network method, while the long-range parts of the PES in both the reactant and product asymptotes are represented by an asymptotically correct form. The long- and short-range parts are connected with intermediate-range parts to make them smooth. Within a statistical quantum model, this PES reproduces both the measured loss rates of ultracold KRb molecules and the K2 and Rb2 product state distributions, underscoring the important role of tunneling in ultracold chemistry. The PES also correctly predicts the lifetime of the K2Rb2* intermediate complex within the Rice–Ramsperger–Kassel–Marcus limit. It thus provides a reliable platform for future dynamical studies of the prototypical reaction.

本工作报道了基于高精度从头算（ab initio）计算得到的、用于KRb + KRb → K₂Rb₂* → K₂ + Rb₂反应的全维全局势能面（potential energy surface, PES）。该势能面的短程部分采用置换不变多项式-神经网络（permutationally invariant polynomial-neural network）方法进行拟合，而反应物与产物渐近区的长程部分则通过渐近正确形式进行表征。长短程区域通过中间程部分进行衔接，以保证势能面的平滑性。在统计量子模型框架下，该势能面不仅重现了超冷KRb分子的实测损耗速率，同时也复现了K₂与Rb₂的产物态分布，凸显了隧穿效应在超冷化学中的关键作用。此外，该势能面还能在莱斯-拉姆珀-凯瑟-马库斯（Rice–Ramsperger–Kassel–Marcus）极限下准确预测K₂Rb₂*中间络合物的寿命。综上，该势能面为该典型反应的后续动力学研究提供了可靠的研究平台。