Ganzenmüller Georg, Berkaïne Nabil, Fouqueau Antony, Casida Mark E, Reiher Markus
Institut de Chimie Moleculaire de Grenoble (ICMG, FR-2607), Laboratoire d'Etudes Dynamiques et Structurales de la Sélectivité, Equipe de Chimie Théorique, Université Joseph Fourier (Grenoble I), F38041 Grenoble, France.
J Chem Phys. 2005 Jun 15;122(23):234321. doi: 10.1063/1.1927081.
Previous work testing density functionals for use in calculating high-spin-low-spin energy differences, deltaE(HL), for iron(II) spin-crossover transitions has tended to conclude that only properly reparametrized hybrid functionals can predict deltaE(HL) since it seems to depend critically on a correct description of the electron pairing energy governed by the exchange term. Exceptions to this rule are the previous three papers (I, II, and III in the present series of papers) where it was found that modern generalized gradient approximations (GGAs) and meta-GGAs could do as well as hybrid functionals, if not better, for this type of problem. In the present paper, we extend these previous studies to five more molecules which are too large to treat with high-quality ab initio calculations, namely, the series [Fe(L)('NHS(4)')], where NHS(4)=2.2'-bis(2-mercaptophenylthio)diethylamine dianion, and L=NH(3), N(2)H(4), PMe(3), CO, and NO(+). Since we know of no reliable experimental estimate of deltaE(HL), we content ourselves with a comparison against the experimentally determined ground-state spin symmetry including, in so far as possible, finite-temperature effects. Together with the results of Papers I, II, and III, this paper provides a test of a large number of functionals against the high-spin/low-spin properties of a diverse set of Fe(II) compounds, making it possible to draw some particularly interesting conclusions. Trends among different classes of functionals are discussed and it is pointed out that there is at least one functional, namely, the OLYP generalized gradient approximation, which is able to give a reasonably good description of the delicate spin energetics of Fe(II) coordination compounds without resorting to hybrid functionals which require the relatively more expensive calculation of a Hartree-Fock-type exchange term.
此前用于计算铁（II）自旋交叉跃迁的高自旋 - 低自旋能量差ΔE(HL)的密度泛函测试工作往往得出结论，即只有经过适当重新参数化的杂化泛函才能预测ΔE(HL)，因为它似乎严重依赖于对由交换项控制的电子配对能的正确描述。该规则的例外情况是之前的三篇论文（本系列论文中的I、II和III），其中发现对于这类问题，现代广义梯度近似（GGA）和元GGA即便不比杂化泛函更好，也能达到同样的效果。在本文中，我们将之前的研究扩展到另外五个因分子太大而无法用高质量从头算进行处理的分子，即[Fe(L)(‘NHS(4)’)]系列，其中NHS(4)=2,2'-双（2 - 巯基苯硫基）二乙胺二价阴离子，且L = NH(3)、N(2)H(4)、PMe(3)、CO和NO(+)。由于我们没有可靠的ΔE(HL)实验估计值，我们满足于与实验确定的基态自旋对称性进行比较，尽可能包括有限温度效应。结合论文I、II和III的结果，本文针对多种铁（II）化合物的高自旋/低自旋性质对大量泛函进行了测试，从而有可能得出一些特别有趣的结论。讨论了不同类型泛函之间的趋势，并指出至少有一种泛函，即OLYP广义梯度近似，它能够在不借助需要相对更昂贵计算Hartree - Fock型交换项的杂化泛函的情况下，对铁（II）配位化合物微妙的自旋能量学给出合理良好的描述。
