We benchmark the accuracy of quantum-chemical methods, including wave function theory methods [coupled cluster theory at the CCSD(T) level, multicongurational perturbation-theory (CASPT2, NEVPT2) and internally-contracted multireference conguration interaction (MRCI)] and...

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.

According to the current paradigm, the metal− hydroxo bond in a six-coordinate porphyrin complex is believed to be significantly less reactive in ligand substitution than the analogous metal−aqua bond, due to a much higher strength of the former bond. Here, we report kinetic studies for nitric oxide (NO) binding to a heme-protein model, acetylated microperoxidase-11 (AcMP-11), that challenge this paradigm. In the studied pH range 7.4−12.6, ferric AcMP-11 exists in three acid− base forms, assigned in the literature as 2), and [(AcMP-11)Fe III (OH)(His − )] (3). From the pH dependence of the second-order rate constant for NO binding (k on ), we determined individual rate constants characterizing forms 1−3, revealing only a ca. 10-fold decrease in the NO binding rate on going from 1 (k on(1) = 3.8 × 10 6 M −1 s −1 ) to 2 (k on (2) = 4.0 × 10 5 M −1 s −1 ) and the inertness of 3. These findings lead to the abandonment of the dissociatively activated mechanism, in which the reaction rate can be directly correlated with the Fe−OH bond energy, as the mechanistic explanation for the process with regard to 2. The reactivity of 2 is accounted for through the existence of a tautomeric equilibrium between the major [(AcMP-11)Fe III (OH)(HisH)] (2a) and minor [(AcMP-11)Fe III (H 2 O)(His − )] (2b) species, of which the second one is assigned as the NO binding target due to its labile Fe−OH 2 bond. The proposed mechanism is further substantiated by quantum-chemical calculations, which confirmed both the significant labilization of the Fe−OH 2 bond in the [(AcMP-11)Fe III (H 2 O)(His − )] tautomer and the feasibility of the tautomer formation, especially after introducing empirical corrections to the computed relative acidities of the H 2 O and HisH ligands based on the experimental pK a values. It is shown that the "effective lability" of the axial ligand (OH − /H 2 O) in 2 may be comparable to the lability of the H 2 O ligand in 1.

No abstract