48
M. Stein
electron densities. The calculated g-tensor principal values are almost identical which
shows that the spin–orbit treatment is not the most critical issue here, at least for this
system with a rhombic g-tensor. The use of the X2C Hamiltonian, on the other
hand, gives calculated g-tensor principal values of 2.148, 2.046, 1.974 and is in best
agreement with experiment.
The challenges of multireference perturbation theory for calculating g-tensors of
first row transition metal complexes were also investigated recently [73]. The scalarrelativistic complete active space self-consistent field (CASSCF) and N-electron
valence perturbation theory (NEVPT2) methods were evaluated for a series of S
½ transition metal complexes of which Ni(mnt)
−
2 was part of. CASSCF calculations based on active spaces that contain only metal-based orbitals significantly
overestimate the g-values and are not able to give reliable results. The inclusion of
dynamic correlation by means of the NEVPT2 method improves the results and a
lowering in the g-shift is obtained. Still, CASSCF leads to an over description of
the metal-ligand bond ionicity and, hence, the spin–orbit coupling matrix elements
are too large. The results show that wavefunction theory has significantly progressed
in computing the properties of large, open-shell transition-metal complexes. Some
methodological work, however, is still required in order to be able to obtain accurate
results.
Molybdenum is part of the active sites of a large number of enzymes such
as sulfite oxidase (SO) and xanthine oxidase (XO) and usually shuttling between
Mo
IV /Mo
V /Mo
VI oxidation states. The Mo
V model compound (N 2 S 2 )MoOCl shows
similarities with certain catalytic intermediates and species of the sulfite and xanthine oxidases. The lack of a crystalline sample with sufficient size and purity has
obstructed the determination of the full g- and A-tensors in the protein. Single crystal
EPR spectra for the molybdenum(V) model compound cis,trans-(L-N 2 S 2 )Mo
V OCl
(L-N 2 S 2 =N,N
-dimethyl-N, N
-bis(mercaptophenyl)ethylenediamine) [74] revealed
the orientation of the g-tensor principal axes in the coordination complex. The
INDO/S-CI calculated principal values of the g-tensor of g 1 1.985, g 2 1.963,
and g 3 1.960 compare reasonably well with the experimental frozen solution values of g 1 2.004, g 2 1.963, and g 3 1.946 (see Table 2). SOMF BP86/SV(P),
TZP calculations gave principal g-values of 2.021, 1.964, 1.950, which are in good
agreement with the experimental data and also the calculated g-tensor orientations
were fully consistent with the experimental findings. The g 1 axis was found to be
oriented along the Mo–O bond, with an angle of 30.8° for INDO/S and 14° for BP86
compared to ~10° found experimentally. The g 2 and g 3 axes are essentially oriented
in the xy-plane of the complex. The g 2 axis forms an angle of 14° (INDO/S) and
18° (BP86/ZORA) with the Mo–S2 bond, compared to ~37° found experimentally.
The g 3 axis is rotated by 14° (INDO/S) and 3° (BP86/ZORA) from the Mo–Cl bond,
respectively, compared to ~38° in experiment (see Table 2).
Following an earlier benchmark study of small and medium Mo
V complexes [75],
Fritscher et al. [76] performed computational studies of g- and molybdenum HFC
tensors for series of larger Mo
V complexes. The best agreement with experimental
data for both hyperfine and g-tensors was obtained with hybrid functionals containing approximately 30–40% Hartree–Fock exchange. Computed orientations of
M. Stein
electron densities. The calculated g-tensor principal values are almost identical which
shows that the spin–orbit treatment is not the most critical issue here, at least for this
system with a rhombic g-tensor. The use of the X2C Hamiltonian, on the other
hand, gives calculated g-tensor principal values of 2.148, 2.046, 1.974 and is in best
agreement with experiment.
The challenges of multireference perturbation theory for calculating g-tensors of
first row transition metal complexes were also investigated recently [73]. The scalarrelativistic complete active space self-consistent field (CASSCF) and N-electron
valence perturbation theory (NEVPT2) methods were evaluated for a series of S
½ transition metal complexes of which Ni(mnt)
−
2 was part of. CASSCF calculations based on active spaces that contain only metal-based orbitals significantly
overestimate the g-values and are not able to give reliable results. The inclusion of
dynamic correlation by means of the NEVPT2 method improves the results and a
lowering in the g-shift is obtained. Still, CASSCF leads to an over description of
the metal-ligand bond ionicity and, hence, the spin–orbit coupling matrix elements
are too large. The results show that wavefunction theory has significantly progressed
in computing the properties of large, open-shell transition-metal complexes. Some
methodological work, however, is still required in order to be able to obtain accurate
results.
Molybdenum is part of the active sites of a large number of enzymes such
as sulfite oxidase (SO) and xanthine oxidase (XO) and usually shuttling between
Mo
IV /Mo
V /Mo
VI oxidation states. The Mo
V model compound (N 2 S 2 )MoOCl shows
similarities with certain catalytic intermediates and species of the sulfite and xanthine oxidases. The lack of a crystalline sample with sufficient size and purity has
obstructed the determination of the full g- and A-tensors in the protein. Single crystal
EPR spectra for the molybdenum(V) model compound cis,trans-(L-N 2 S 2 )Mo
V OCl
(L-N 2 S 2 =N,N
-dimethyl-N, N
-bis(mercaptophenyl)ethylenediamine) [74] revealed
the orientation of the g-tensor principal axes in the coordination complex. The
INDO/S-CI calculated principal values of the g-tensor of g 1 1.985, g 2 1.963,
and g 3 1.960 compare reasonably well with the experimental frozen solution values of g 1 2.004, g 2 1.963, and g 3 1.946 (see Table 2). SOMF BP86/SV(P),
TZP calculations gave principal g-values of 2.021, 1.964, 1.950, which are in good
agreement with the experimental data and also the calculated g-tensor orientations
were fully consistent with the experimental findings. The g 1 axis was found to be
oriented along the Mo–O bond, with an angle of 30.8° for INDO/S and 14° for BP86
compared to ~10° found experimentally. The g 2 and g 3 axes are essentially oriented
in the xy-plane of the complex. The g 2 axis forms an angle of 14° (INDO/S) and
18° (BP86/ZORA) with the Mo–S2 bond, compared to ~37° found experimentally.
The g 3 axis is rotated by 14° (INDO/S) and 3° (BP86/ZORA) from the Mo–Cl bond,
respectively, compared to ~38° in experiment (see Table 2).
Following an earlier benchmark study of small and medium Mo
V complexes [75],
Fritscher et al. [76] performed computational studies of g- and molybdenum HFC
tensors for series of larger Mo
V complexes. The best agreement with experimental
data for both hyperfine and g-tensors was obtained with hybrid functionals containing approximately 30–40% Hartree–Fock exchange. Computed orientations of