Department of Inorganic Chemistry, Faculty of Chemistry , Jagiellonian University , Gronostajowa 2 , Krakow 30-387 , Poland.
J Chem Theory Comput. 2018 Aug 14;14(8):4010-4027. doi: 10.1021/acs.jctc.8b00200. Epub 2018 Jul 17.
High-level multireference (CASPT2, NEVPT2) calculations are reported for transition metal aqua complexes with electronic configurations from (3d) to (3d). We focus on the experimentally evidenced excitation energies to their various ligand-field states, including different spin states. By employing models accounting for both explicit and implicit solvation, we find that solvation effect may contribute up to 0.5 eV to the excitation energies depending on the charge of ion and character of the electronic transition. We further demonstrate that with an adequate choice of the active space and the energetics extrapolated to the complete basis set limit, the presently computed excitation energies are in a good agreement with the experimental data. This allows us to conclusively resolve significant discrepancies reported in earlier theory works [e.g., J. Phys. Chem. C 2014 , 118 , 29196 - 29208 ]. For the benchmark set of 19 spin-forbidden and 24 spin-allowed transitions (for which experimental data are unambiguous), we find the mean absolute error of 0.15 or 0.13 eV and the maximum error of 0.56 or 0.42 eV for CASPT2 or NEVPT2 calculations, respectively. For the particularly challenging sextet-quartet gap for [Fe(HO)], we support our interpretation by additional calculations with multireference configuration interaction (MRCI) and coupled cluster theory up to the CCSDT(Q) level. By underlining a rather subtle interplay between the solvation and correlation effects, the findings of this Article are relevant not only for modeling and interpretation of optical spectra of transition metal complexes but also in further benchmarking of theoretical methods for the challenging problem of spin-state energetics.
高水准多参考(CASPT2、NEVPT2)计算报告了电子构型为(3d)到(3d)的过渡金属水合配合物。我们关注的是实验证实的激发能到它们的各种配体场态,包括不同的自旋态。通过采用考虑显式和隐式溶剂化的模型,我们发现溶剂化效应可能对激发能的贡献高达 0.5 eV,这取决于离子的电荷和电子跃迁的性质。我们进一步证明,通过适当选择活性空间和能量外推到完全基组极限,可以使目前计算的激发能与实验数据很好地吻合。这使我们能够最终解决早期理论工作中报告的显著差异[例如,J. Phys. Chem. C 2014 ,118 ,29196 - 29208]。对于基准集的 19 个自旋禁阻和 24 个自旋允许跃迁(对于这些跃迁,实验数据是明确的),我们发现 CASPT2 或 NEVPT2 计算的平均绝对误差分别为 0.15 或 0.13 eV,最大误差分别为 0.56 或 0.42 eV。对于[Fe(HO)]的特别具有挑战性的 sextet-quartet 间隙,我们通过额外的多参考组态相互作用(MRCI)和耦合簇理论计算,直至 CCSDT(Q)水平,为我们的解释提供支持。通过强调溶剂化和相关效应之间相当微妙的相互作用,本文的发现不仅对过渡金属配合物光学光谱的建模和解释具有重要意义,而且对自旋态能的挑战性问题的理论方法的进一步基准测试也具有重要意义。