The anisotropic hyperfine coupling constants for a 19-electron organometallic complex, CpCo(CO)2 -, were calculated using the B3LYP density functional in conjunction with several all-electron basis sets. The calculated hfc constants are generally within 10% of the experimental values as determined by EPR. The calculated wave function indicates that the unpaired electron has a much lower occupancy of the cobalt dyz orbital (0.17−0.31) than the value of 0.56 obtained from the EPR data. It is concluded that the traditional method of calculating atomic orbital spin populations from EPR hyperfine coupling constants, which neglects spin polarization and covalency effects, should be used with caution when applied to organometallic systems. No significant structural changes occur in the ligands of CpCo(CO)2 - as a result of having to accommodate an extra electron.

The anisotropic hyperfine coupling (hfc) constants for the 19-electron Mn(CO)5Cl- complex, 1, and for the 17-electron Mn(CO)4Cl- complex, 2, were calculated using density functional theory (DFT). The calculated hfc values for 1 are in good agreement with the experimental ones reported in two EPR studies, which were not able to distinguish between 1 and 2. The Mn−CO(axial) bond dissociation energy in 1 was calculated to be 19 kcal/mol, which shows that 1 is stable to loss of axial CO. These data indicate that the species observed in the EPR experiments was the 19-electron complex, 1. The unpaired electron population of the manganese d z 2 orbital indicated by the SOMO of the DFT wave function (ca. 0.2−0.3) is significantly less than that obtained from the two EPR analyses (0.49, 0.63), which neglected spin polarization. The calculations show that spin polarization in 1 causes the spin density (as measured by EPR) to differ significantly from the unpaired electron density (SOMO). It is concluded that neglect of spin polarization in the EPR analysis of open-shell transition metal compounds may lead to an overestimate of the unpaired electron population on the metal. The standard method for estimating atomic orbital populations by ratioing the observed hfc in a molecule to the atomic hfc is not reliable for organometallic compounds.

Density functional theory (DFT) was used to calculate the equilibrium geometries and EPR hyperfine coupling constants for the CpCr(CO) 2 NOand CpW(NO) 2 P(OMe) 3 19-electron complexes. The calculated EPR hyperfine parameters for both molecules were found to be in good agreement with the experimental values. The EPR study on CpCr(CO) 2 NOindicated that the nitrosyl ligand bends significantly when the neutral 18-electron parent molecule is reduced. This was confirmed by the DFT calculations, although the degree of bending is not as large as that predicted from the EPR analysis. The calculated geometry for CpW(NO) 2 P-(OMe) 3 showed that the nitrosyl ligands are not significantly bent, which is consistent with the crystal structure of the related compound CpW(NO) 2 P(OPh) 3 . In both CpCr(CO) 2 NOand CpW(NO) 2 P(OMe) 3 , the unpaired electron is localized to about the same extent on the NO ligands. It is concluded that, although distortion of a ligand implies significant localization of the unpaired electron on that ligand, an undistorted ligand does not imply that the unpaired electron has no appreciable amplitude on that ligand. DFT was also applied to the 19-electron CpMo(CO) 3 P(OMe) 3 molecule, whose role as a key intermediate in photochemical reactions of Cp 2 Mo 2 (CO) 6 with P(OMe) 3 has been inferred from mechanistic studies. It was found that loss of P(OMe) 3 from CpMo(CO) 3 P(OMe) 3 is slightly thermodynamically favorable (∆E ) ∼1 kcal/mol). This is consistent with the necessity of using an excess of the phosphorus ligand in reactions in which CpMo(CO) 3 PR 3 is believed to be the electron-transfer agent. There is no indication of P(OMe) 3 ligand distortion in this molecule, just as there is no indication of Cp ring slippage in any of the Cp-containing molecules. Finally, a calculation of the hyperfine coupling constants in the 18 + δ Co(CO) 3 L 2 (L 2 ) 2,3-bis(diphenylphosphino)maleic anhydride) complex was also carried out. The calculated values are in reasonable agreement with experiment.

Density functional theory (DFT) calculations were carried out on three 19-electron metal sandwich complexes, CpFe(η 6 -C 6 H 6 ), CpFe(η 6 -C 6 Me 6 ), and (C 5 Me 5 )Fe(η 6 -C 6 H 6 ), to determine their electronic structure. The nuclear quadrupole splittings for these three molecules were also calculated using DFT and were found to be in reasonable agreement with the experimental values observed by Mo ¨ssbauer spectroscopy. For CpFe(η 6 -C 6 H 6 ) and (C 5 Me 5 )-Fe(η 6 -C 6 H 6 ), the metal contributions to the singly occupied molecular orbital (SOMO) were found to be about 45% and 30%, respectively, lower than estimates based on the magnitude of the quadrupole splitting. In the case of CpFe(η 6 -C 6 Me 6 ), however, the SOMO is dominated by the permethylated arene ligand, while the metal contribution is only about 10%, much less than the estimate of 75% based on the quadrupole splitting. In all three molecules the spin density is concentrated around the iron atom, and the partial charge on the iron atom is 0.75, as calculated using the AIM method. It is concluded that the determination of the composition of a molecular orbital based on comparing the observed quadrupole splitting for a molecule to the estimated splitting for the free atom is too approximate to be reliable and that consideration of a single molecular orbital does not allow a reliable estimate of the net charge or spin density around an atom in a molecule.