Soydaş Emine, Bozkaya Uğur
Department of Chemistry, Atatürk University , Erzurum 25240, Turkey.
J Chem Theory Comput. 2015 Apr 14;11(4):1564-73. doi: 10.1021/ct501184w. Epub 2015 Mar 4.
An assessment of orbital-optimized MP2.5 (OMP2.5) [ Bozkaya, U.; Sherrill, C. D. J. Chem. Phys. 2014, 141, 204105 ] for thermochemistry and kinetics is presented. The OMP2.5 method is applied to closed- and open-shell reaction energies, barrier heights, and aromatic bond dissociation energies. The performance of OMP2.5 is compared with that of the MP2, OMP2, MP2.5, MP3, OMP3, CCSD, and CCSD(T) methods. For most of the test sets, the OMP2.5 method performs better than MP2.5 and CCSD, and provides accurate results. For barrier heights of radical reactions and aromatic bond dissociation energies OMP2.5-MP2.5, OMP2-MP2, and OMP3-MP3 differences become obvious. Especially, for aromatic bond dissociation energies, standard perturbation theory (MP) approaches dramatically fail, providing mean absolute errors (MAEs) of 22.5 (MP2), 17.7 (MP2.5), and 12.8 (MP3) kcal mol(-1), while the MAE values of the orbital-optimized counterparts are 2.7, 2.4, and 2.4 kcal mol(-1), respectively. Hence, there are 5-8-folds reductions in errors when optimized orbitals are employed. Our results demonstrate that standard MP approaches dramatically fail when the reference wave function suffers from the spin-contamination problem. On the other hand, the OMP2.5 method can reduce spin-contamination in the unrestricted Hartree-Fock (UHF) initial guess orbitals. For overall evaluation, we conclude that the OMP2.5 method is very helpful not only for challenging open-shell systems and transition-states but also for closed-shell molecules. Hence, one may prefer OMP2.5 over MP2.5 and CCSD as an O(N(6)) method, where N is the number of basis functions, for thermochemistry and kinetics. The cost of the OMP2.5 method is comparable with that of CCSD for energy computations. However, for analytic gradient computations, the OMP2.5 method is only half as expensive as CCSD.
本文对用于热化学和动力学的轨道优化MP2.5（OMP2.5）方法[Bozkaya, U.; Sherrill, C. D. J. Chem. Phys. 2014, 141, 204105]进行了评估。OMP2.5方法应用于闭壳层和开壳层反应能量、势垒高度以及芳香族键解离能。将OMP2.5的性能与MP2、OMP2、MP2.5、MP3、OMP3、CCSD和CCSD(T)方法的性能进行了比较。对于大多数测试集，OMP2.5方法的表现优于MP2.5和CCSD，并能提供准确的结果。对于自由基反应的势垒高度和芳香族键解离能，OMP2.5与MP2.5、OMP2与MP2、OMP3与MP3之间的差异变得明显。特别是对于芳香族键解离能，标准微扰理论（MP）方法严重失效，其平均绝对误差（MAE）分别为22.5（MP2）、17.7（MP2.5）和12.8（MP3）kcal mol⁻¹，而轨道优化对应方法的MAE值分别为2.7、2.4和2.4 kcal mol⁻¹。因此，使用优化轨道时误差降低了5至8倍。我们的结果表明，当参考波函数存在自旋污染问题时，标准MP方法会严重失效。另一方面，OMP2.5方法可以减少无限制Hartree-Fock（UHF）初始猜测轨道中的自旋污染。总体评估得出，OMP2.5方法不仅对具有挑战性的开壳层系统和过渡态很有帮助，对闭壳层分子也很有帮助。因此，作为一种O(N⁶)方法（其中N是基函数的数量），在热化学和动力学方面，相对于MP2.5和CCSD，人们可能更倾向于使用OMP2.5。对于能量计算，OMP2.5方法的成本与CCSD相当。然而，对于解析梯度计算，OMP2.5方法的成本仅为CCSD的一半。
