The multiconfigurational CASSCF/CASPT2 approach, along with various functionals of density functional theory, is applied to selected iron(II)-nitrosyl ({FeNO}(7)) complexes, both with heme and nonheme groups. The energetics of the lowest doublet and quartet spin states at the correlated ab initio (CASPT2) level is presented for the first time. Comparison of the CASSCF and (unrestricted) DFT spin densities indicates that the nonhybrid functionals yield the spin densities most closely to the ab initio ones. The analysis of the multiconfigurational CASSCF wave function in terms of the localized active orbitals allows one to resolve the nature of Fe-NO bonding as a mixture of Fe(II)-NO(0) and Fe(III)-NO(-) resonance structures (in comparable contributions) for both spin states and various ligands.

The energetics of various electromeric states for two heme complexes with an iron-oxo (FeO(3+)) group, FeO(P)(+) and FeO(P)Cl (P = porphin), have been investigated, employing DFT and correlated ab initio methods (CASPT2, RASPT2). Our interest focused in particular on tri- and pentaradicaloid iron(IV)-oxo porphyrin radical states as well as iron(V)-oxo states. Surprisingly, the iron(V)-oxo ground state is predicted for both models in vacuo. However, the presence of a polarizable medium, such as a solvent or a protein environment, favors the iron(IV)-oxo porphyrin radical cation, which is predicted to be the actual ground state of FeO(P)Cl under such conditions. Nonetheless, the iron(V)-oxo electromer is still expected to lie only a few kcal/mol above the ground state-a conclusion coming from both CASPT2 and RASPT2 calculations with a very large active space and further supported by a calibration with respect to coupled cluster CCSD(T) calculations for a simplified small model. The DFT results turn out to be strongly functional-dependent and thereby inconclusive. The widely used B3LYP functional-although correctly predicting the iron(IV)-oxo porphyrin radical ground state for FeO(P)Cl-seems to place the iron(V)-oxo states much too high in energy, as compared to the present CASPT2, RASPT2, and CCSD(T) results.

DFT calculations of the molecular structure of the intrazeolite η 1 {CuNO} 11 adduct and the 14 N and 17 O hyperfine and 63 Cu superhyperfine coupling constants were performed and compared with previous EPR results. The calculations confirmed the choice of signs adopted in the previous analysis of the experimental data and the character of the SOMO. The influence of the basis set and the exchange-correlation functional on the HFCC and the spin-density distribution was investigated and briefly discussed. The global repartition of the spin density over Cu (F ) 0.11), N(F ) 0.58), and O (F ) 0.34) atoms determined from the Mulliken population analysis compared well with the experiment. The 14 N hyperfine tensor was successfully reproduced with the LanL2DZ basis and BPW91 functional, whereas in the case of the 63 Cu superhyperfine dipolar tensor T the agreement, except for that of the T zz component, was less satisfactory because of an overestimated polarization of the 3d yz orbital, regardless of the computation level. For the calculation of a iso (Cu), because LanL2DZ treats inner electrons with the effective core potential, a 6-311G(df) basis set appeared to be the most appropriate, leading to excellent agreement between the experimental and calculated values.