NMR of Metal-CO Systems
J. Am. Chem. Soc., Vol. 120, No. 19, 1998 4773
use of the Z-surface method.21,22 This approach uses Bayesian
probability to deduce the µ,F terms of the HB equations. Typically,
the shielding results of three spinning speeds were averaged, incorporat-
ing from 10 to 20 sidebands per spectrum, using a program kindly
provided by Dr. H. Le.
Quantum Chemical Calculations of Nuclear Shieldings and
Electric Field Gradients. Nuclear shieldings were calculated with
the sum-over-states density-functional perturbation theory (SOS-DFPT)
approach in its LOC1 approximation,23,24 using individual gauges for
localized orbitals (IGLO).25 All calculations of shieldings and of
nuclear quadrupole couplings were carried out by using the gradient-
corrected PW91 functional,26 employing a version of the deMon
program27 (including the deMon-NMR modules23,24) modified for the
use of semilocal effective-core potentials (ECPs).
corresponding to delocalized cluster bonding) are excluded from the
localization, a combined delocalized/localized analysis of the shielding
tensor becomes possible. We will call this in the following a “partial
IGLO”.35 Second, in a recent DFT-IGLO study of 17O shieldings in
transition-metal oxo complexes,36 it was found that the usual IGLO
method based on the Boys localization37 was very sensitive to the way
the semicore metal orbitals were localized (either separate from or
together with the valence shell). In contrast, a modified IGLO
procedure based on a Pipek-Mezey (PM) localization38 (in the following
denoted IGLO-PM) was more stable in this respect. As an additional
feature, the PM localization provides σ-π separation in multiple bonds,
even when no symmetry plane is present, whereas the Boys localization
normally gives “banana bonds”. The σ-π separation is very useful for
the analysis of shielding tensors in the present carbonyl ligands, as we
describe below.
Quasirelativistic ECPs were used for the metals, with (8s7p6d)/
[6s5p3d] Gaussian-type orbital valence basis sets.28,29 IGLO-II all
electron basis sets25 were used either for all ligand atoms (for Fe(CO)5,
Fe2(CO)9, and Ni2(η5-C5H5)2(CO)2) or in a locally dense basis ap-
proximation for Rh6(CO)16, as described in ref 17. Further computa-
tional details are also as described previously17 for Fe2(CO)9 and
Rh6(CO)16. For Fe(CO)5, comparative calculations with the larger
IGLO-III ligand basis sets25 were also carried out. Experimental
structures30-33 were used for all species. In the case of Ni2(η5-C5H5)2-
(CO)2, the unreasonable experimental C-H distances were replaced
by the more reasonable value of 1.094 Å. We implicity take
intermolecular contributions to shielding (and the electric field gradient)
in these systems to be small, a view supported by, for example, the
lack of any significant changes in isotropic shifts between solution and
the solid state, as well as the observation that shift, shift tensor, and
electric field gradient tensor results are well described theoretically by
using the isolated complex structures.
Computed absolute shieldings were converted to relative shifts via
the theoretical absolute shieldings of TMS for 13C (187.5, 184.0 ppm
with IGLO-II and IGLO-III bases, respectively) and H2Oliq for 17O
(271.0, 289.4 ppm, calculated from the computed absolute shieldings
of H2Ovap ) 307.1, 325.5 ppm with IGLO-II and IGLO-III bases, and
the experimental 36.1 ppm gas-liquid shift of water39).
Nuclear quadrupole coupling constants were computed from the same
Kohn-Sham orbitals employed for the nuclear shielding calculations,
using the approach implemented in the deMon-NMR code.40 The
oxygen nuclear quadrupole moment was taken to be -0.02558 barn.41
Results and Discussion
We describe below our experimental and theoretical results
on the following four compounds:
Previous work has shown that, due to the complicated nature of the
NMR chemical shift parameter, it is beneficial to employ different types
of analyses in terms of molecular orbital (MO) contributions. Within
an IGLO procedure, localized MO (LMO) contributions give useful
insights. Alternatively, within a gauge-including atomic orbital (GIAO)
framework,34 or with a common gauge origin, a delocalized (canonical)
MO analysis is feasible (with a common gauge, care has to be taken
with respect to basis set convergence). Beyond these standard
procedures, we have found two other variants to be useful. First, if
within an IGLO-type treatment a few key orbitals (typically those
(21) Le, H.; Pearson, J. G.; de Dios, A. C.; Oldfield. E. J. Am. Chem.
Soc. 1995, 117, 3800-3807.
(22) Havlin, R. H.; McMahon, M.; Srinivasan, R.; Le, H.; Oldfield, E.
J. Phys. Chem. 1997, 101, 8908-8913.
(23) Malkin, V. G.; Malkina, O. L.; Casida, M. E.; Salahub, D. R. J.
Am. Chem. Soc. 1994, 116, 5898-5908.
(24) Malkin, V. G.; Malkina, O. L.; Eriksson, L. A.; Salahub D. R.
Modern Density Functional Theory: A Tool for Chemistry; Theoretical
and Computational Chemistry; Seminario, J. M., Politzer, P., Eds.;
Elsevier: Amsterdam, 1995; Vol. 2.
(25) Kutzelnigg, W.; Fleischer, U.; Schindler, M. NMR-Basic Principles
and Progress; Springer: Heidelberg, 1990; Vol. 23, 165ff.
(26) (a) Perdew, J. P.; Wang, Y. Phys. ReV. B 1992, 45, 13244-13249.
(b) Perdew, J. P. Electronic Structure of Solids; Ziesche, P., Eischrig, H.,
Eds.; Akademie Verlag: Berlin 1991. (c) Perdew, J. P.; Chevary, J. A.;
Vosko, S. H.; Jackson, K. A.; Pederson, M. R.; Singh, D. J.; Fiolhais, C.
Phys. ReV. B 1992, 46, 6671-6687.
Fe(CO)5, discussed as our first example, is of interest since
it has both axial and equatorial CO ligands. Perhaps surpris-
ingly, neither the 13C nor the 17O shielding tensor principal
(27) deMon program: Salahub, D. R.; Fournier, R.; Mlynarski, P.; Papai,
I.; St-Amant, A.; Ushio, J. Density Functional Methods in Chemistry;
Labanowski, J., Andzelm, J.; Springer: New York, 1991. St-Amant, A.;
Salahub, D. R. Chem. Phys. Lett. 1990, 169, 387-392.
(28) Dolg, M.; Wedig, U.; Stoll, H.; Preuss, H. J. Chem. Phys. 1987,
86, 866-872.
(35) As with calculations using a common gauge origin, care has to be
taken with respect to basis set completeness. Moreover, this is not a black-
box procedure, as symmetry and other requirements have to be considered
(Kaupp, M.; Fleischer, U., to be published).
(36) Kaupp, M.; Malkina, O. L.; Malkin, V. G. J. Chem. Phys. 1997,
106, 9201-9212.
(37) Edmiston, C.; Ruedenberg, K. ReV. Mod. Phys. 1963, 35, 457-
465. Edmiston, C.; Ruedenberg, K. J. Chem. Phys. 1965, 43, 597-602.
Boys, S. F. Quantum Theory of Atoms, Molecules and the Solid State;
Lo¨wdin, P. O., Ed.; Academic Press: New York, 1966; pp 253ff. This
procedure is often erroneously attributed to: Foster, J. M.; Boys, S. F. ReV.
Mod. Phys. 1963, 35, 457-465.
(29) Andrae, D.; Ha¨ussermann, U.; Dolg, M.; Stoll, H.; Preuss, H. Theor.
Chim. Acta 1990, 77, 123-141.
(30) Braga, D.; Grepioni, F.; Orpen, A. G. Organometallics 1993, 12,
1481-1483.
(31) Cotton, F. A.; Troup, J. M. J. Chem. Soc., Dalton Trans. 1974,
800-802.
(32) Byers, L. R.; Dahl, L. F. Inorg. Chem. 1980, 19, 680-692.
(33) Corey, E. R.; Dahl, L. F.; Beck, W. J. Am. Chem. Soc. 1963, 85,
1202-1203.
(38) Pipek, J.; Mezey, P. G. J. Chem. Phys. 1989, 90, 4916-4926.
(39) See, e.g.: McFarlane, W.; McFarlane, H. C. E. in ref 12, pp 403ff.
(40) See, e.g.: Fedotov, M. A.; Malkina, O. L.; Malkin, V. G. Chem.
Phys. Lett. 1996, 258, 330-335.
(34) See, e.g.: (a) Ditchfield, R. Mol. Phys. 1974, 27, 789. (b) Wolinski,
K.; Hinton, J. F.; Pulay, P. J. Am. Chem. Soc. 1990, 112, 8251-8260.
(41) Pyykko¨, P. Z. Naturforsch. A 1989, 47, 189-196.