Bao Jie J, Gagliardi Laura, Truhlar Donald G
Department of Chemistry, Chemical Theory Center, and Supercomputing Institute, University of Minnesota, Minneapolis, Minnesota 55455-043, USA.
Phys Chem Chem Phys. 2017 Nov 15;19(44):30089-30096. doi: 10.1039/c7cp05156e.
Multiconfiguration pair-density functional theory (MC-PDFT) is a post multiconfiguration self-consistent field (MCSCF) method with similar performance to complete active space second-order perturbation theory (CASPT2) but with greater computational efficiency. Cyano radical (CN) is a molecule whose spectrum is well established from experiments and whose excitation energies have been used as a testing ground for theoretical methods to treat excited states of open-shell systems, which are harder and much less studied than excitation energies of closed-shell singlets. In the present work, we studied the adiabatic excitation energies of CN with MC-PDFT. Then we compared this multireference (MR) method to some single-reference (SR) methods, including time-dependent density functional theory (TDDFT) and completely renormalized equation-of-motion coupled-cluster theory with singles, doubles and noniterative triples [CR-EOM-CCSD(T)]; we also compared to some other MR methods, including configuration interaction singles and doubles (MR-CISD) and multistate CASPT2 (MS-CASPT2). Through a comparison between SR and MR methods, we achieved a better appreciation of the need to use MR methods to accurately describe higher excited states, and we found that among the MR methods, MC-PDFT stands out for its accuracy for the first four states out of the five doublet states studied this paper; this shows its efficiency for calculating doublet excited states.
多组态对密度泛函理论(MC-PDFT)是一种后多组态自洽场(MCSCF)方法,其性能与完全活性空间二阶微扰理论(CASPT2)相似,但计算效率更高。氰基自由基(CN)是一种分子,其光谱已通过实验得到充分确立,其激发能已被用作处理开壳层体系激发态的理论方法的测试平台,与闭壳层单重态的激发能相比,开壳层体系的激发能更难研究且研究较少。在本工作中,我们用MC-PDFT研究了CN的绝热激发能。然后我们将这种多参考(MR)方法与一些单参考(SR)方法进行了比较,包括含时密度泛函理论(TDDFT)和带单、双激发以及非迭代三激发的完全重整化运动方程耦合簇理论[CR-EOM-CCSD(T)];我们还与一些其他的MR方法进行了比较,包括组态相互作用单双激发(MR-CISD)和多态CASPT2(MS-CASPT2)。通过SR方法和MR方法之间的比较,我们更好地认识到使用MR方法准确描述高激发态的必要性,并且我们发现,在MR方法中,MC-PDFT在所研究的五个双重态中的前四个态的准确性方面表现突出;这表明了它在计算双重态激发态方面的效率。