Ottochian Alistar, Morgillo Carmela, Ciofini Ilaria, Frisch Michael J, Scalmani Giovanni, Adamo Carlo
Chimie ParisTech, PSL Research University, CNRS, Institute of Chemistry for Life and Health Sciences, Paris, France.
Gaussian, Inc., Wallingford, Connecticut.
J Comput Chem. 2020 May 15;41(13):1242-1251. doi: 10.1002/jcc.26170. Epub 2020 Feb 19.
In this paper we present the implementation and benchmarking of a Time Dependent Density Functional Theory approach in conjunction with Double Hybrid (DH) functionals. We focused on the analysis of their performance for through space charge-transfer (CT) excitations which are well known to be very problematic for commonly used functionals, such as global hybrids.Two different families of functionals were compared, each of them containing pure, hybrid and double-hybrid functionals.The results obtained show that, beside the robustness of the implementation, these functionals provide results with an accuracy comparable to that of adjusted range-separated functionals, with the relevant difference that for DHs no parameter is tuned on specific compounds thus making them more appealing for a general use. Furthermore, the algorithm described and implemented is characterized by the same computational cost scaling as that of the ground state algorithm employed for MP2 and double hybrids.
在本文中，我们展示了一种与双杂化（DH）泛函相结合的含时密度泛函理论方法的实现与基准测试。我们着重分析了它们对通过空间电荷转移（CT）激发的性能，众所周知，对于常用泛函（如全局杂化泛函）而言，这种激发是非常棘手的问题。我们比较了两个不同的泛函家族，每个家族都包含纯泛函、杂化泛函和双杂化泛函。所得结果表明，除了实现的稳健性之外，这些泛函提供的结果准确性与调整后的范围分离泛函相当，相关的区别在于对于双杂化泛函，无需针对特定化合物调整参数，因此使其更具普遍应用的吸引力。此外，所描述和实现的算法具有与用于MP2和双杂化的基态算法相同的计算成本缩放特性。
