Accurate quantum chemistry with spatially varying correlation factors

量子化学中的高精度第一性原理计算能够通过定量预测来支持研究。这种计算的一个挑战是达到电子薛定谔方程的完整基组极限，其中收敛速度特别受短程电子相关描述的影响。F12理论通过明确依赖于电子间距离的相关因子来增强波函数，并显着提高了基础方法的准确性。在F12理论中建立的标准程序中的一个缺点是在相关因子中使用固定范围参数，这不能被选择为对于所有目的都是最佳的，特别是如果分子包含具有强烈不同价电子密度的原子，或者如果核心-价相关效应是重要的。在建议的项目中，我们将引入相关因子的修改，允许本地选择最佳范围参数。与此同时，该方法承诺保留标准F12方法的几乎所有优点。将实施和测试，首先使用二阶微扰理论的相关处理，随后也耦合集群理论和多参考方法的新的ananterior。改进后的方法将被整合到高精度的协议中，目标是具有较重主族元素和过渡金属的特定化合物。

High-accuracy first-principles computations in quantum chemistry are able to support research by quantitative predictions. One challenge in such computations is reaching the complete basis set limit of the electronic Schrödinger equation, of which the rate of convergence is particularly governed by the description of short-range electron correlation. F12 theory augments the wavefunction ansatz by correlation factors that explicitly depend on the interelectronic distance and significantly increases the accuracy of the underlying method. One drawback in the established standard procedure in F12 theory is the use of a fixed range parameter in the correlation factor, which cannot be chosen optimal for all purposes, in particular, if a molecule contains atoms with strongly differing valence electron density or if core-valence correlation effects are of importance. In the proposed project, we will introduce a modification of the correlation factor that allows to locally choose the optimal range parameter. At the same time the approach promises to keep nearly all of the advantages of the standard F12 approach. The new ansatz will be implemented and tested, first for correlation treatments using second-order perturbation theory, subsequently also for coupled-cluster theory and multireference approaches. The improved methods will be integrated into high-accuracy protocols targeting in particular compounds with heavier main group elements and transition metals.