Nuclear Magnetic Resonance Volume 15

A Review of the Literature Published Between June 1984 and May 1985

By G. A. Webb

The Royal Society of Chemistry

Copyright © 1986 The Royal Society of Chemistry
All rights reserved.
ISBN: 978-0-85186-382-5

Contents

CHAPTER 1 Theoretical and Physical Aspects of Nuclear Shielding By Cynthia J. Jameson,
CHAPTER 2 Applications of Nuclear Shielding By G.E. Hawkes,
CHAPTER 3 Theoretical Aspects of Spin – Spin Couplings By A. Laaksonen,
CHAPTER 4 Applications of Spin – Spin Couplings By D.F. Ewing,
CHAPTER 5 Nuclear Spin Relaxation in Liquids By H. Weingärtner,
CHAPTER 6 Solid State N.M.R. By P.S. Belton, S.F, Tanner, and K.M. Wright,
CHAPTER 7 Natural Macromolecules By D.B. Davies,
CHAPTER 8 Synthetic Macromolecules By A.V. Cunliffe,
CHAPTER 9 Conformational Analysis By U. Berg and J. Sandström,
CHAPTER 10 N.M.R. of Living Systems By P.G. Morris,
CHAPTER 11 Oriented Molecules By C.L. Khetrapal and K.V. Ramanathan,
CHAPTER 12 Heterogeneous Systems T. Cosgrove,
AUTHOR INDEX, 400,

CHAPTER 1

Theoretical and Physical Aspects of Nuclear Shielding

BY CYNTHIA J. JAMESON

1 Theoretical Aspects of Nuclear Shielding

A. Ab Initio Calculations.- Nuclear shieldings of Cu, Ag, Zn, Cd, and Mn in metal complexes were calculated by a finite perturbation SCF method using the MIDI-1 basis set for metal atoms and ligands. This basis set is of double zeta quality for valence orbitals. The gauge origin is taken at the position of the metal atom. In these complexes the primitive metal atoms are characterized by the ground state electronic configuration d10s1-2p0 except for Mn which is d5s2. The complexes studied are [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], and Mn(CO)5L (L = H, CN, CH3, Cl). The shielding contributions are analyzed according to diamagnetic and paramagnetic contributions and in terms of individual MO contributions to each. As expected, the ligand AOs make only minor contributions to the metal shielding. For example, in Cd CCH3)2 only 75/4851 of the diamagnetic term and -29/ 1090 of the paramagnetic term come from the two CH3 ligands. The diamagnetic term is dominated by the inner core MOs on the metal, the valence MO contributions are only 5-18% of the total diamagnetic term. Although the absolute shielding is dominated by the diamagnetic term, the differences in shielding between molecules are primarily determined by the valence MO contributions to the paramagnetic term. These findings support the general assumptions which are usually invoked in approximate treatments of chemical shifts which include only the valence MO contributions to the paramagnetic term. The small changes in the diamagnetic term from one complex to another result from the transferability of the core MOs and the small relative contributions of ligand orbitals. The calculated metal d orbital versus p orbital contributions to shielding are described qualitatively in terms of the empirical electron donating and withdrawing properties of the ligand. For the d10s1-2p0 configuration, an electron-donating ligand deshields the metal nucleus primarily by donation of electrons from the ligands to the metal p orbitals whereas an electron-withdrawing ligand deshields the metal nucleus via back-donation of electrons from the metal d orbitals to the ligands. The relative contributions of the d and p mechanisms depend on the metal and the number and nature of the ligands.

The calculations on the Mn complexes as representative transition metal compounds show that the paramagnetic term and the chemical shifts between complexes arise primarily from terms, which in the perturbation scheme, are represented by transitions which are largely dπ -> dσ. This explains why approximate treatments which consider only the dπ -> dσ terms using a simple crystal field model work at all. The paramagnetic contribution to the shielding in Fin is an order of magnitude larger than those of the IB and IIB metal complexes. In the former there is a paramagnetic shielding which arises due to an intrinsic open d shell, whereas in the latter, mechanisms of electron donation to or from ligands are necessary in order to create holes in the d shell or electrons in the outer p shell of the metal atom.

For nuclear sites of sufficiently low symmetry there are non-vanishing off-diagonal shielding components σij(j≠i) in the principal axis system, reviewed in the previous volume of this series. The antisymmetric components of the 13C shielding tensor in several compounds (oxetane, cyclopentane, 1,3-cyclopentadiene, 1,4 cyclohexadiene, nitromethane, acetone, acetaldehyde, cyclopentane, ethylene sulfide, ethylene oxide, cyclopropene, and cyclobutene) have been calculated using the individual gauge for localized orbitals (IGO) method. The root mean square antisymmetric components [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] range from 1.0 ppm in CH313CHD to 139.4 ppm in cyclopropene (13C). The contribution of the antisymmetric components of the shielding to spin-lattice relaxation by shielding interactions can be large enough in some cases (0.20 to 0.88 times as large as the shielding anisotropy contribution, and 2.0 times as large in the cyclopropene case), to be a significant contributor to this relaxation mechanism. Nevertheless, the experimental determination of the sum of the squares of the antisymmetric shielding components is problematical unless the shielding interactions mechanism can be made dominant by observing favorable molecules in very high fields. Nuclei much heavier than 13C should be more favorable. Unfortunately, the antisymmetric components of the shielding are less accurately estimated in these cases and the results now available from this work, and those reviewed in the previous volume of this series, do not show any obvious correlations between electronic structure and magnitudes of 13C antisymmetric components.

Calculations of 1H and 13C shielding in a variety of small molecules [FORMULA NOT REPRODUCIBLE IN ASCII] by Ditchfield’s GIAO SCF FPT method using 4-31G basis sets give relative proton shifts to ±0.5 ppm and relative C shifts to ±6 ppm compared to gas-phase data. The calculated absolute shieldings tend to be too shielded compared to experiment. Of course, some of the discrepancy can be accounted for by rovibrational corrections, but for tetrahedral carbons these tend to be small. For CH4 for example, it is estimated to be -3.3 ppm. SDS-CI calculations of 1H and 17O shielding in H2O2 using a 6-31G** basis set give σ(1H) = 22.24 ppm, which is 7.8 ppm less shielded than 1H in H2O by a similar calculation, σ(17O)=149.5 ppm which is 146.7 ppm less shielded than the calculated 17O in H2O, which in turn is about 50 ppm too deshielded compared to experiment. The 1H shieldings in isolated H2O, H3O+, and OH-, respectively are 31.2, 22.2, and 39.1 ppm at the equilibrium configuration. The 17O shieldings are 326.5, 298.9, 318.0 ppm respectively. In this calculation the 17O shielding in H2O is also too deshielded compared to the experimental value of 357 ±17 ppm, which is obtained from the absolute shielding scale recently reported by Wasylishen based on the molecular beam experiment on C17O. The gas phase value for H2O is 344.0 ppm at 293 K, which when combined with the 13.6 ppm vibrational correction by Fowler and Raynes gives σe(17O in H2O) = 357 ppm. The available 17O shielding calculations for H2O (reviewed in Vol. 12 of this series) also give results which are less shielded than experiment. Apparently, the paramagnetic contribution to 17O shielding is overestimated by theoretical calculations.

B. Semi-empirical Calculations.- An empirical relation between two experimentally independent quantities: the chemical shift on the one hand and an optical parameter which can be calculated from the visible-UV spectrum of the complex on the other hand has been presented. For d6 complexes such as those of Co(III) the optical parameter incorporates the nephelauxetic ratio (β = Bcomplex/Bgaseous ion, the ratio of the Racah parameters for interelectronic repulsion) which is calculated together with the L matrix elements, within the framework of the parametrized db model on the basis of the two absorption bands o f cubic parentage for the complex. The energies of the first and second cubic transitions are predominantly determined by 8. Thus, a parameter C is defined:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)

where h1(a1T1g(α)) α = x, y, z is the energy of the first cubic transition [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] as obtained from the visible-UV spectrum. The a in front of t h e term symbol indicates that this is the lowest term of the symmetry type in question. h2(a1T2g(α)) corresponds to the [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] transition.

In the interpretation of chemical shifts of d6 complexes both the spectrochemical series (A series) of ligands and the nephelauxetic series (β series) have been used. The present observation indicates that the order of transition metal chemical shifts induced by ligands might be governed by an “internal field strength” parameter Δ/B. The internal field strength series for ligands runs as

[FORMULA NOT REPRODUCIBLE IN ASCII]

in which the heavy halides have been moved forward with respect to the spectrochemical series due to their high nephelauxetism (low B values). The dimensionless internal field strength parameter Δ/B is empirically determined from those values of Δ and B which reproduce the observed transition energies h1 and h2 as differences between the lowest eigenvalues of the A1g(5×5), T1g(4×4), and T2g(7×7) matrices for the energies in an intermediate field (in which the off-diagonal elements represent the interelectronic repulsion). In the interval 1.10

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)

The paramagnetic term is then written in terms of the parameters [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)

This gives an improved linear correlation for 59Co shifts in Co(III) complexes and the slope of the straight line reflects the magnitude of the unknown parameter (/β). This “effective” r-3 value includes the influence of the ligands on the radial term, if β is small, and nephelauxetism (electron cloud expansion) is large.

The conformation dependence of the 13C chemical shifts has been calculated for carbonyl, Cα, and Cβ in polyglycine I and II and poly(L-alanine) α-helix and β-sheet forms. The results are found to be qualitatively consistent with the observed conformational dependence as determined by the cross-polarization MAS technique, 19F chemical shifts in BF4-n(OH)n- and BF4- n(OOH)n- were calculated and discussed in terms of the electronic structure of these molecules.

2 Physical Aspects of Nuclear Shielding

A. Anisotropy of the Shielding Tensor.- The 13C shielding tensors in linear and pseudolinear (containing only H atoms off a C∞ axis) molecules have been determined by cross-polarization and intermolecular cross-polarization experiments on solid solutions of the molecule of interest in argon. The results are shown in Table 1 as absolute shielding; the reported chemical shifts relative to external liquid TMS were converted using an absolute shielding for TMS = 185.4 ppm. There are several interesting points worth noting in this study. One is that the assignments of three principal values to three principal axes in Table 2 required the correct relative ordering provided by theoretical calculations. The results of such calculations by a coupled Hartree-Fock method using individual gauges for localized molecular orbitals (IGLO) are also reported. The calculated values of the components relative to CH4 differ by less than 40 ppm from experiment, except in CO. Agreement with σav was much better (to within 17 ppm). The calculations agree with only one permutation of the experimentally observed shielding components, thus allowing assignments with confidence. In linear molecules the paramagnetic contribution to σ is zero by symmetry so long as the gauge origin is chosen anywhere along the line of centers. The similarity of the o values in the linear molecules shows that the diamagnetic contribution varies only slightly from one electronic environment to another. With the presence of symmetry-breaking protons off-axis, the paramagnetic term is no longer “quenched” and contributes to the component along the long axis of the molecule in dramatic contrast to the linear case.

13C shielding tensors have also been reported in quinol, HCN, halobenzenes, PhX (X = F, Cl, Br, I), in trans-polyacetylene, and in L-asparagine monohydrate. In quinol and the halobenzenes, the component perpendicular to the molecular plane is the most shielded component and becomes more shielded with separation of the C atom from the substituent (with the exception of PhI), e.g., increasing from 91.4 ppm for the carbon to 185.4 ppm for the para carbon in PhF. In contrast the isotropic average does not show regular changes with the position of C in the ring. On the other hand, the shielding component of C1 perpendicular to the ring changes drastically with X: 179.2 ppm in benzene (H), 173.4 ppm (I), 133.4 ppm (Br), 131.4 ppm (Cl), 118.4 ppm (OH), 91.4 ppm (F), in the direction one might expect by using electronegativity arguments. In HCN the 13C shielding anisotropy is reported as 334±20 ppm, and there is some indication that the molecular structure of HCN varies with liquid crystal solvents. The orientation of the 13C shielding tensor axes in pairwise 13C-enriched undoped trans-polyacetylene has been determined. σ11 = -31.6 ppm along an axis which makes an angle of 40±5° with respect to the C-C bond and 80±5° with the C=C bond. σ22 = 42.4 ppm, σ33 = 140.4 ppm along an axis perpendicular to the molecular plane. A review of the anisotropy of 13C shielding has been published.

The 15N shielding tensor for the peptide bond in glycylglycine hydrochloride monohydrate has been reported. The 17O shielding tensors in the hexacarbonyIs of Cr, Mo, and W are shown in Table 3. The 13C shielding tensors in the same molecules were measured as well, however the recent values do not differ significantly from the earlier report in Volume 11 of this series. 19F nmr of single crystal sym-C6Cl3F3 yields the following absolute shielding components: ½ (σ11 + σ22) = 262.8 ppm, σ33 = 390.8 ppm, which may be compared to C6F6 in which they are 302.8 and 460.8 ppm respectively. The 19F powder spectrum at 77 K. in CFBr3 gives an anisotropy δσ = 142±12 ppm, which is comparable to that in CFCl3. In KZnF3 the 19F nucleus is in a nearly ionic bond so the anisotropy is very small (18.6 ppm), σ[parallel] = 390 ppm and σ[parallel] 371.4 ppm, where the unique axis is along the Zn-F-Zn direction.

The 31P shielding anisotropies in isolated, end, middle, and branching phosphate groups are related to the bond orders of the PD bonds around the P nucleus The ideal PO43- tetrahedron, an isolated phosphate ion, has 4 equal P-0t bonds each with a bond order ~ 5/4 and an isotropic 31P shielding. An end PO43- and a branching PO43- are axially symmetric PO43- environments which can be considered as evolving from the isotropic PO43- by transfer of π bond character from one bond to the other 3 bonds making 3 terminal P-Ot. bonds and one P-Ob bridging bond (this is an end PO43- moiety] or by transfer of π bond character from 3 bonds in the isotropic case to one bond, making one terminal P-Ot. bond and 3 P-Ob bridging bonds (this is a branching PO43- moiety]. In the end phosphate groups the P-Ob along the symmetry axis has lost π bond character to the other 3 and this is accompanied by a decrease of 31P shielding parallel to the axis and an increase perpendicular to this axis. A branching PO43- formed by transfer of π bond character to one P-Ot bond on the symmetry axis (a change of electron density in an opposite direction from that in the end phosphate group), has the opposite shielding anisotropy. The 31P shielding parallel to the symmetry axis increases and shielding perpendicular to the axis decreases relative to the free PO43- ion. In both the branching and end phosphates the change in shielding component parallel to the bond is in the same direction as the change in the bond order, whereas the component perpendicular to it changes in an opposite way. These effects are illustrated in Table 4. These results are in agreement with earlier data reviewed in Vol. 14 of this series. The shielding tensors of all three 31P nuclei in chlorotris(triphenylphosphine) rhodium(I) are reported. The most shielded component is parallel to the P-Rh bond in each case.

(Continues…)Excerpted from Nuclear Magnetic Resonance Volume 15 by G. A. Webb. Copyright © 1986 The Royal Society of Chemistry. Excerpted by permission of The Royal Society of Chemistry.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.