Geometrical isomerism and (19)F NMR spectroscopy of octahedral perfluoroethyl- and perfluoropropyl-substituted β-diketonates and monothio-β-diketonates of rhodium(III)

Article abstract of DOI:10.1139/v99-133

In solution in acetone at room temperature, tris(perfluoropropanoyl and perfluorobutanoyl)benzoylmethanate complexes of rhodium(III) are shown by (19)F NMR spectroscopy to exist as nearly statistical 1:3 mixtures of their respective fac and mer octahedral isomers, whereas the analogous perfluoroalkanoylthiobenzoylmethanate complexes have a fac-octahedral structure. For all four complexes studied, vicinal F-F coupling constants are too small (0-3 Hz) to lead to resolved multiplet peaks, so that perfluoroethyl groups show singlet resonances for their CF₃ groups. Since four-bond coupling constants are significantly larger, ~9 Hz, perfluoropropyl groups show the expected multiplet resonances for their CF₃ groups. In acetone, AB patterns, with geminal coupling constants equal to ~270 Hz, are observed for the CF₂ groups next to the chelate ring (except for the perfluoropropanoylthiobenzoylmethanate complex). (These observations are modified by solvent. In toluene, one of the AB patterns of a mer-octahedral complex collapses to a single peak.) The AB spectra, which are subject to substituent and solvent effects, can be interpreted in terms of perfluoroalkyl groups rotating rapidly about the C(ring) C(alky1) bond, with an unsymmetric double potential well, giving unequal rotamer populations. Changes in the spectra were not noticeable over the temperature range accessible in the acetone and toluene solvents used.

Full text of DOI:10.1139/v99-133

1734
19
Geometrical isomerism and F NMR
spectroscopy of octahedral perfluoroethyl- and
perfluoropropyl-substituted ␤-diketonates and
monothio-␤-diketonates of rhodium(III)
Gary A. Stern and John B. Westmore
Abstract: In solution in acetone at room temperature, tris(perfluoropropanoyl and perfluorobutanoyl)benzoylmethanate
complexes of rhodium(III) are shown by 19F NMR spectroscopy to exist as nearly statistical 1:3 mixtures of their
respective fac and mer octahedral isomers, whereas the analogous perfluoroalkanoylthiobenzoylmethanate complexes
have a fac-octahedral structure. For all four complexes studied, vicinal F–F coupling constants are too small (0–3 Hz)
to lead to resolved multiplet peaks, so that perfluoroethyl groups show singlet resonances for their CF3 groups. Since
four-bond coupling constants are significantly larger, ~9 Hz, perfluoropropyl groups show the expected multiplet
resonances for their CF3 groups. In acetone, AB patterns, with geminal coupling constants equal to ~270 Hz, are
observed for the CF2 groups next to the chelate ring (except for the perfluoropropanoylthiobenzoylmethanate complex).
(These observations are modified by solvent. In toluene, one of the AB patterns of a mer-octahedral complex collapses
to a single peak.) The AB spectra, which are subject to substituent and solvent effects, can be interpreted in terms of
perfluoroalkyl groups rotating rapidly about the Cring—Calkyl bond, with an unsymmetric double potential well, giving
unequal rotamer populations. Changes in the spectra were not noticeable over the temperature range accessible in the
acetone and toluene solvents used.
Key words: rotation, perfluoroalkyl groups, fluorine NMR spectra, fluorine–fluorine coupling constants, fluorinated
β-diketonate complexes, monothio-β-diketonates, rhodium complexes.
Résumé : Faisant appel à la spectroscopie RMN du 19F, on montre que, en solution dans l’acétone, à la température
ambiante, les complexes tris(perfluoropropanoyl et perfluorobutanoyl)benzoylméthanate de rhodium(III) existent sous la
forme de mélanges 1:3 pratiquement statistiques de leurs isomères octaédriques fac et mer respectifs alors que les
complexes perfluoroalcanoylthiobenzoylméthanates analogues existent sous la forme d’une structure octaédrique-fac.
Pour chacun des quatre complexes étudiés, les constantes de couplage F–F vicinales sont trop faibles (0–3 Hz)
conduire à des pics multiplets résolus; les groupes CF3 des groupes perfluoroéthyles apparaissent donc comme
singulets. Les constantes de couplage à travers quatre liaisons étant beaucoup plus importantes, ≈9 Hz, les groupes CF3
des groupes perfluoropropyles se présentent sous la forme de multiplets attendue. Dans l’acétone, on observe des
diagrammes AB, avec des constantes de couplage geminales d’environ 270 Hz, pour les groupes CF2 adjacents du
cycle du chélate (à l’exception du complexe perfluoropropanoylthiobenzoylméthanate). (Ces observations varient avec le
solvant. Dans le toluène, l’un des diagrammes AB d’un complexe octaédrique-mer ne donne qu’un seul pic). Les
spectres AB, qui sont soumis à des effets de substituant et de solvants, peuvent être interprétés en fonction de groupes
perfluoroalkyles en rotation rapide autour de la liaison Ccycle—Calkyle avec un puits de potentiel double et asymétrique
conduisant à des populations inégales de rotamères. Dans le domaine de température accessible avec des solutions dans
l’acétone ou le toluène, il n’a pas été possible de noter de différences dans les spectres.
Mots clés : rotation, groupes perfluoroalkyles, spectres RMN du fluor, constantes de couplage fluor–fluor, complexes
β-dicétonates fluorés, monothio-β-dicétonates, complexes du rhodium.
[Traduit par la Rédaction] Stern and Westmore 1744
Received November 5, 1998.
This paper is dedicated to Professor Ted Schaefer in recognition of his outstanding contributions to Canadian chemistry.
G.A. Stern1 and J.B. Westmore.2 Department of Chemistry, University of Manitoba, Winnipeg, MB R3T 2N2, Canada.
1Present address: Department of Fisheries and Oceans, Freshwater Institute, 501 University Crescent, Winnipeg, MB R3T 2N6,
Canada.
2Author to whom correspondence may be addressed. Telephone: (204) 474-9761. Fax: (204) 474-7608.
e-mail: westmor@cc.umanitoba.ca
Can. J. Chem. 77: 1734–1744 (1999)
© 1999 NRC Canada

Stern and Westmore
1735
The 19F NMR spectra of many metal complexes contain-
ing perfluoroalkyl groups have been reported (31–37). The
compounds include metal-alkyl bonded complexes of Mn,
Re, Fe, and Co (31); Co and Rh (32); β-diketonate and (or)
monothio-β-diketonate complexes of Ni, Pd, and Co (33); Be
(34); Ru, Rh, Pd, and Pt (35); Ni, Pd, and Co (36); and Zn
and Pd (37). In most cases the spectra were simply used for
characterization purposes. In this paper, we report a more
detailed analysis of the spectra of rhodium(III) β-diketonate
and monothio-β-diketonate complexes containing perfluoro-
alkyl substituents. Not only do the spectra reveal the geome-
try of the complexes, but they also lead to some interesting
conclusions about the rotational behavior of the
perfluoroalkyl groups.
Introduction
Octahedral metal complexes having three unsymmetric
bidentate ligands can adopt facial (fac) or meridional (mer)
structures. While the mer isomer is statistically three times
more likely to form than the fac isomer, the ratio is strongly
influenced by thermodynamic parameters or by kinetics of
the formation reactions. Under appropriate conditions,
isomerism of such complexes can be studied by nuclear
magnetic resonance spectroscopy. Our initial interest in this
topic was prompted by our syntheses of fluorinated β-dike-
tonate complexes of rhodium(III), which are sufficiently vol-
atile for study by electron ionization mass spectrometry. We
found, by single crystal X-ray diffraction (XRD), that
tris(thioacetyltrifluoroacetonato)rhodium(III) adopts the fac-
octahedral geometry (1). However, for the rhodium(III) com-
plexes 1–4 to be described in this paper, XRD was not an
option because we were unable to prepare single crystals of
suitable quality. For these complexes, 19F NMR spectros-
copy shows that not only does the X substituent (O or S) in-
fluence geometrical isomerism but also, as a bonus, has led
to some interesting conclusions regarding the rotational be-
havior of the perfluoroalkyl substituents.
Experimental section
Syntheses
The four ligands needed to prepare complexes 1–4 were
synthesized by published methods (38).
To prepare β-diketonate complexes 3 and 4, RhCl3·H2O
(0.32 mmol), in 10 mL of ethanol, was added to a solution
of the appropriate ligand (0.96 mmol) in 10 mL of ethanol.
The solution was refluxed for 1 h, then cooled, and a small
amount of water was added. This solution was then shaken
with methylene chloride; the methylene chloride extract was
dried over sodium sulfate. After evaporation of methylene
chloride, the resulting oily red residue was added to a silica
gel column (60–200 mesh) through which methylene chlo-
ride was passed. For each complex, two bands separated,
while a portion of the sample still remained at the top of the
column. The first band (red) eluted quite readily but did not
contain the desired complex. The second band (yellow),
which contained the desired product, eluted slowly, so that
the silica from this part of the column was removed and the
product was extracted from it with acetone. After evapora-
tion of acetone, the residue was subjected to fractional subli-
mation, under roughing pump vacuum, to yield the desired
rhodium chelate as a yellow sublimate. The molecular
masses and fragmentation patterns determined by mass spec-
trometry (VG 7070E-HF mass spectrometer) supported as-
signment of the yellow products as monomeric tris-chelate
Because our description of the NMR spectra presented in
this paper requires an interpretation of F–F coupling con-
stants, we first present a brief summary of this topic, along
with some key references. The coupling constant depends
upon the number and nature of the bonds intervening be-
tween the coupled fluorine atoms, dihedral angles between
bonds, substituent effects, and “through space” interactions.
In particular, the near-zero values (<~1 Hz) observed for vic-
inal coupling constants (2–18) may be a consequence of the
averaging effects of internal rotation about the C—C bond,
and, in part, to cancellation of the orbital, spin-dipolar, and
Fermi contact terms of nonuniform sign (19, 20). All three
of these terms make significant contributions to geminal
coupling constants (19, 20), which, experimentally, have
wide variations (150–400 Hz), that have been attributed
partly to the sensitivity of the Fermi contact term to changes
in the FCF bond angles, and also partly to the sensitivity of
the orbital and Fermi contact terms to substituent effects
(21–24). In contrast to the near-zero coupling constants be-
tween vicinal fluorines, substantially larger values for cou-
pling over four and five bonds have been observed (25–30).
+
complexes. Complex 3 m/z: 898 (100%), RhL3 (898,
+
calcd.); 848 (36%), [RhL3 – CF2]+; 633 (57%), RhL2 ; 181
+
+
(81%), HRhC6H5 . Complex 4 m/z: 1048 (100%), RhL3
+
+
(1048, calcd.); 733 (51%), RhL2 ; 181 (65%), HRhC6H5 .
The monothio-β-diketonate complexes 1 and 2 were pre-
pared by the same procedure. The desired products were
again found in the second band (red) on the silica gel
column and were subjected to fractional sublimation, in
vacuo, to yield the desired complexes as red sublimates. The
assignments of the red products as monomeric tris-chelate
complexes were supported by mass spectrometry. Complex 1
+
m/z: 946 (100%), RhL3 (946, calcd.); 896 (17%), [RhL3
+
– CF2]+; 665 (32%), RhL2 ; 383 (49%), [Rh(L – H)]+. Com-
+
plex 2 m/z: 1096 (100%), RhL3 (1096, calcd.); 765 (49%),
+
RhL2 ; 433 (55%), [Rh(L – H)]+.
19F nuclear magnetic resonance spectroscopy
Dilute solutions (5–7 mmol%) of the complexes were pre-
pared in acetone or toluene to which a small amount of C6F6
© 1999 NRC Canada

1736
Can. J. Chem. Vol. 77, 1999
Table 1. Fluorine chemical shifts (ppm) and coupling constants (Hz) for perfluoroethyl-substituted complexes 1 and 3.a
AB part
X3 part
Sub-spectrum
Complex
Solvent
Sub-spectrum
δA(CF2)
δB(CF2)
∆υAB (Hz)
JAB
δX(CF3)
JAX = JBX
1 (X = S)
Acetone
Acetone
–119.41
–119.41
0
140
194
282
121
135
170
98
—
–81.38
–81.72
–81.86
–81.92
–82.04
–82.24
–82.37
–82.56
–82.45
≤1.0
~2.0
~2.7
≤1.0
~3.0
~1.7
~2.3
≤1.0
~2.3
3 (X = O)b
1
2
3
4
1
2
3
4
–118.201
–118.197
–118.342
–118.391
–119.064
–119.054
–119.374
–119.557
–118.697
–118.883
–118.837
–118.821
–119.543
–119.656
–119.722
–119.557
267.5
266.5
268.0
268.0
269.0
267.2
267.3
—
a
b
c (fac?)
d
a
b
c (fac?)
d
3 (X = O)b
Toluene
0
aValues of δX, JAX, and JAB are experimentally determined. The other NMR parameters emerge from the computer simulations for which a line width of
1.5 Hz was used. The uncertainties are estimated to be 0.002 ppm for δA and δB, 0.01 ppm for δX, and 0.3 Hz for JAB. ∆υAB (Hz) is the frequency
difference between the A and B resonances.
bFor complex 3, sub-spectra 1–4 and a–d, respectively, are presented from low to high field. No match between a particular resonance occurring in the
AB region with another in the CF3 (X3) region is implied.
was added to act as an internal chemical shift reference. In
these solvents, δF(C6F6) was taken as –162.90 ppm with re-
spect to δF(CFCl3). Thus, the variations in δF that are re-
ported for samples in the two solvents will be a combination
of solvent effects for both the reference and the sample. The
chlorinated solvents, chloroform and methylene chloride,
were unsuitable because the complexes appeared to decom-
pose slowly in them. The solutions were transferred to 5 mm
o.d. NMR tubes, then degassed by the freeze–pump–thaw
procedure; the sample tubes were then flame sealed. All 19F
NMR spectra were recorded with a Bruker AM300 spec-
trometer, at a frequency of 282.4 MHz, with fluorines de-
coupled from protons by broad-band irradiation. For detailed
examinations of selected spectral regions, the spectral win-
dow was reduced, and the spectrum was accumulated with
up to 500 free induction decays; Fourier transformation
yielded spectra with a digital resolution of 0.076 Hz/point.
The probe was at room temperature (~27°C) for the results
described (we noticed no significant changes in the spectra
over the high- or low-temperature range accessible with the
acetone and toluene solvents).
periments, we have observed that the 103Rh–F coupling con-
stant in the analogous complex when RF = CF3 is ~0.5 Hz,
about the same as JHF when not proton decoupled. However,
in the spectra discussed in this paper, peak structure result-
ing from 103Rh–F coupling was not resolved. The low-field
resonance, at –81.38 ± 0.01 ppm (relative intensity = 3), and
the high-field resonance, at –119.41 ± 0.01 ppm (relative in-
tensity = 2), must be assigned to the fluorine atoms of the
CF3 and CF2 groups, respectively. The full width at half
maximum (FWHM) of each signal was ~2 Hz, i.e., not mea-
surably greater than that of the C6F6 reference peak; com-
bined with the lack of peak structure, and assuming that the
CF3 peak is no more than 0.5 Hz broader than the reference
peak. This implies that the F–F vicinal coupling constant is
no greater than 1 Hz, as anticipated from the introduction.
The single-peak CF2 resonance, unlike the AB patterns
shown by complexes 2–4 (later), indicates that the chemical
shifts of these two fluorines are identical, or nearly so, be-
cause geminal coupling constants are expected to be high
(see Introduction). To permit easy comparison of NMR pa-
rameters of all complexes, the chemical shifts for complex 1
are included in Table 1.
The spectrum of complex 2 (RF = CF2CF2CF3), a potential
seven-spin ABMNX3 system, contains three major reso-
nances, namely a triplet (centred on δ ≈ –79.5 ppm), a
quintet (centred on δ ≈ –117.7 ppm), and a singlet (δ
≈ –125.5 ppm), with relative intensities of 3:2:2, respectively
(Fig. 1). However, this spectrum is much simpler than its full
analysis as an ABMNX3 spectrum would require. The triplet
must be assigned to the fluorine atoms of the CF3 group. As-
signment of the remaining resonances is consistent with both
coupling constant and chemical shift evidence. We assign
the singlet to the fluorines of the β-CF2 group, since vicinal
coupling constants between these fluorines and those of the
adjacent CF3 and α-CF2 groups are expected, on the basis of
the earlier discussion, to be very small. Furthermore, there is
no evidence of any MN structure for this peak thus indicat-
ing that these fluorines have identical, or nearly identical,
chemical shifts. Consequently, we can interpret the other
The NMR spectra were analyzed with the assistance of
the NUMARIT spectral simulation computer program (39)3,
which was extensively modified by R. Sebastian, formerly of
this department.
Results and discussion
The 19F NMR spectra of the monothio-β-diketonate com-
plexes are simpler than those of the β-diketonates, so we
shall discuss them first.
Monothio- -diketonate complexes
The spectrum of complex 1 (RF = CF2CF3), recorded in
acetone at room temperature, consists of only two reso-
nances, both appearing as singlets. The deceptive simplicity
of this spectrum must be explained, for the CF2 group could
be an AB system because the plane of the chelate ring is not
a molecular plane of symmetry. In addition, in separate ex-
3 Reworked by R. Sebastian and incorporated into the XSIM NMR computational package as NUMMRIT by K. Marat of this department, to
whom queries should be addressed.
© 1999 NRC Canada

Stern and Westmore
1737
Fig. 2. CF2 19F{1H} spectral region (AB part) of complex 2:
Fig. 1. 19F{1H} NMR spectrum of complex 2 in acetone.
(a) experimental spectrum in acetone, (b) simulated spectrum.
parts of the spectrum in terms of coupling between the α-
CF2 and CF3 groups only. The remaining assignment of the
quintet to the α-CF2 group is also consistent with its chemi-
cal shift, which is similar to that of the CF2 group of com-
plex 1. This is as expected because the electronic
environment of the α-CF2 group in complex 2 is similar to
that of the CF2 group in complex 1, i.e., both are located be-
tween the ring carbon atom and a perfluoroalkyl group (the
inductive effects of the CF3 and CF2CF3 groups bound to the
CF2 group next to the ring should be very similar). On the
other hand, the β-CF2 group in complex 2 has a substantially
different environment from the α-CF2 group, and therefore a
different chemical shift. Similar assignments have been
made for the fluorine atoms in a metal complex containing
the Mn-CO-CF2CF2CF3 group (31). The quintet, with peak
intensity ratios of 1.01:4.02:6.02:4.01:1.00, is associated
with two outer sets, each of four low intensity peaks. The
quintet and triplet splitting patterns exhibited by the α-CF2
and CF3 groups, respectively, are consistent with a special
case of an ABX3 spin system that we analyze as follows. In
the general case (40), for an ABX3 spin system, the AB part
of the spectrum consists of 16 peaks, comprised of four AB
sub-spectra of relative intensities 1:3:3:1, each with the same
value of JAB, but with effective chemical shift differences
given by ∆*δAB = ∆δAB + ms(JAX – JBX)/f, where ∆δAB is the
chemical shift difference between A and B; ms is the resul-
tant spin of the X3 nuclei, with the values ±1/2, ±3/2; and f
is the operating frequency of the spectrometer. The repeated
equal to JAX. Figure 2 shows a comparison between the AB
portion of the experimental spectrum and a simulated spec-
trum (39) for which the input parameters δA, δB, δX, JAB, JAX
,
JBX and line width were adjusted until the best match be-
tween the X3 and AB regions of the experimental and simu-
lated spectra was obtained. The resulting NMR parameters
4
4
are presented in Table 2. The JAX and JBX values are 8.9 ±
0.3 Hz, i.e., significantly larger than the vicinal coupling
constants. The JAB value, 270.0 ± 0.3 Hz, is within the wide
range of values (150–400 Hz), noted in the introduction, that
have been observed for geminal coupling constants.
For either monothio-β-diketonate complex only one type
of perfluoroalkyl group can be recognized from the NMR
spectrum, and consequently, the three groups are equivalent.
Therefore, both complexes have a fac-octahedral structure
for which the perfluoroalkyl groups are necessarily equiva-
lent, owing to the presence of a C3 symmetry axis. (In con-
trast, for a mer-octahedral structure, which lacks a C3 axis,
the three perfluoroalkyl groups are non-equivalent, i.e., one
of the three perfluoroacyl residues is trans to a benzoyl resi-
due, while the other two are trans to each other but not
equivalent by symmetry. Thus, each perfluoroalkyl group of
the mer isomer has a distinct spectrum.) A fac-octahedral
structure for the monothio-β-diketonate complexes agrees with
our XRD structure determination (1) of tris(thioacetyltrifluoro-
acetonato)rhodium(III), and is consistent with structures of
spacings between the AB sub-spectra are equal to 1/2(JAX
BX). The X3 part consists of 14 resolved peaks of which
two together, containing half the total intensity of the X3 re-
gion, are always separated by (JAX + JBX) (40). When JAX
BX, as expected here for a CF2CF2CF3 group with free rota-
+
J
=
J
tion about the C—C bonds, the X3 part collapses to a sym-
metrical 1:2:1 triplet with spacings equal to JAX. The
repeated spacings between the AB sub-spectra now equal
J
AX, and the effective chemical shift differences all equal
∆δAB. The eight strong central peaks of the four AB sub-
spectra can only collapse to the observed 1:4:6:4:1 quintet if
∆υ = JAX, where ∆υ is the common separation between the
central peaks of the AB sub-spectra. The remaining eight
weak peaks of the four AB sub-spectra are assigned to the
two sets of four low-intensity outer peaks, with spacings
© 1999 NRC Canada

1738
Can. J. Chem. Vol. 77, 1999
Table 2. Fluorine chemical shifts (ppm) and coupling constants (Hz) for perfluoropropyl-substituted complexes 2 and 4, in solution in
acetone.a
AB part
X3 part
β-CF2 part
Sub-
spectrum
Sub-
spectrum
Sub-
spectrum
4
δA(α-CF2) δB(α-CF2) ∆υAB (Hz) JAB
4JAX = JBX
δX(CF3)
δ(β-CF2)
Complex
J(vic)
2 (X = S)
4 (X = O)b
–116.781
–115.462
–115.518
–115.718
–115.743
–117.041
73
270.0 8.9
–79.55
–79.43
–79.52
–79.56
–79.45
–125.34 ~1
1
2
3
4
–115.942 136
–115.914 112
–116.078 102
–116.102 101
270.8 8.9
270.9 8.8
270.4 8.7
270.4 8.7
a
b
c
d
w
x
y
z
–125.12 ~0.9
–125.21 ~0.9
–125.40 ~1.9
–125.47 ~1.6
aValues of δX, δ(β-CF2), and J(vic) are experimentally measured. The other NMR parameters emerge from the computer simulations for which a line
4
4
width of 1.5 Hz was used. The uncertainties are estimated to be 0.002 ppm for δA and δB, 0.01 ppm for δX, and 0.3 Hz for JAB, JAX, and JBX. υAB (Hz)
is the frequency difference between the A and B chemical shifts.
bFor complex 4, sub-spectra 1–4, a–b, and w–z, respectively, are listed from low to high field. No matches among particular absorptions occurring in
the the three spectral regions are implied.
various other tris(monothio-β-diketonato)rhodium(III) com-
plexes (35, 41).
Fig. 3. 19F{1H} NMR spectrum of complex 3 in (a) acetone
(b) toluene.
-Diketonate complex 3 (RF = C2F5)
Spectra of this complex, shown in Fig. 3, were recorded in
both toluene and acetone, in order to reinforce the correct-
ness of the assignments given below. In both solvents, the
spectra contain two groupings of peaks. One grouping (rela-
tive intensity = 3) occurs between –81.65 and –82.10 ppm
(in acetone) or between –82.20 and –82.60 ppm (in toluene),
and the other grouping (relative intensity = 2) occurs be-
tween –117.20 and –119.80 ppm (in acetone) or between
–118.20 and –120.60 ppm (in toluene). These must be
assigned to the fluorines of the CF3 and CF2 groups, respec-
tively. In both solvents, the first set contains four major reso-
nance peaks and a number of minor peaks, while the second
set contains at least six major resonance peaks in acetone,
seven in toluene, and a number of minor peaks. The minor
peaks in the CF3 region of the spectra are assigned to impu-
rities that, despite several attempts, we failed to remove
completely by column chromatography followed by frac-
tional sublimation of the complex under roughing pump vac-
uum.
The observation of four major peaks in the CF3 region in-
dicates that both fac and mer isomers are present in solution.
Three of these peaks, which are necessarily of equal inten-
sity, can be assigned to the mer isomer and the fourth to the
fac isomer. The relative areas of the shaded parts of the
peaks (Fig. 4), starting from low field, are
1.05:1.01:0.85:1.00 (in acetone) and 1.05:1.00:0.85:1.01 (in
toluene). These results suggest that the third peak, starting
from low field, can be assigned to the fac isomer, even
though the intensities of the three remaining peaks are not
exactly equal (owing probably to contributions from impuri-
ties). The fac:mer ratio is close to the statistical 1:3 ratio, in
contrast to the monothio-β-diketonate complexes 1 and 3 for
which the fac isomer is formed exclusively.
Although no splitting of the four resonances in the CF3 re-
gion is apparent, the peak widths differ. Thus, for the spec-
trum recorded in acetone solution, the approximate FWHM
values are 4.0, 4.7, 2.0, and 5.0 Hz, from low to high field,
respectively. If we interpret these peaks as the X3 parts of
four ABX3 sub-spectra then, according to the analysis de-
scribed above for complex 2, each of these peaks corre-
sponds to a symmetrical 1:2:1 triplet with spacings equal to
1/2(JAX + JBX) = JAX, since JAX = JBX. By noting that the
FWHM value for the C6F6 resonance used for calibration is
~2 Hz, then we find upper limits for the respective values of
JAX to be ~2.0, 2.7, ≤1, and 3.0 Hz. Similarly, the FWHM
values in toluene, i.e., 2.7, 4.3, 1.3, and 4.3 Hz, lead to the
JAX upper limit values 1.7, 2.3, ≤1, and 2.3 Hz, respectively.
Based on the earlier discussion, these small coupling con-
stants are as expected for coupling between vicinal fluorines.
The NMR parameters for these resonance peaks (referred to
as sub-spectra a–d) are summarized in Table 1.
© 1999 NRC Canada

Stern and Westmore
1739
Fig. 4. CF3 19F{1H} spectral region (X3 part) of.complex 3 in
(a) acetone and (b) toluene. Not all impurities could be
successfully be removed. Relative areas of the peaks are
determined from the shaded portions.
Fig. 5. CF2 19F{1H} NMR spectral region (AB part) of complex
3: (a) experimental spectrum in acetone, (b) simulated spectrum.
Not all impurities could be successfully be removed.
lyzed, by simulation as four ABX3 spectra, with the Numarit
program (39) by using the NMR parameters and line width
as input parameters that could be adjusted until a good
match to the experimental spectrum was obtained. (In the
form used, this program can simulate four ABX3 spectra, but
the four intensities must be equal, and the line widths in all
spectra must be the same. This presents little difficulty in the
present case because the four experimental spectra for the
C2F5 groups are of approximately equal intensity, and the
line width variations will not significantly influence the ap-
pearance of the spectrum. Because we could not match the
peaks in the AB (CF2) region to specific peaks in the X3
(CF3) region, we used as input the average δX value of the
peaks in the CF3 region, and also JAX = JBX = 1.6 Hz. The
effect of this approximation on the accuracy of the simula-
tions is, of course, minimal.) The experimental spectra, re-
corded in acetone and toluene, are compared to the
simulated spectra in Figs. 5 and 6. The NMR parameters for
the sub-spectra, labeled 1–4, recorded in both solvents, are
given in Table 1.
Strictly speaking, the CF2 region of the spectrum should
be interpreted as the AB parts of four ABX3 spectra. How-
ever, in view of the small JAX and JBX coupling constants
observed, it is more practical to regard this region as four
overlapping AB sub-spectra, with the peaks broadened by
the small vicinal couplings. Thus, we expect to observe eight
strong peaks in this region, unless some of them are super-
imposed. The presence of small amounts of impurities as
well as partial overlap of some peaks made it impossible to
measure the relative intensities of peaks accurately and to
associate them with a particular isomer. (As noted above, the
intensities of the three resonances for the mer isomer should
always be equal, and therefore, if one of the four signals is
smaller or larger than the other three, it must be assigned to
the fac isomer. In the present case, the intensities of the
peaks in the X3 region show that the intensities in the AB re-
gion will not be sufficiently different to allow us to assign
signals in the AB region to a particular isomer.) These AB
sub-spectra are therefore simply referred to as sub-spectra
1–4 (counting from low to high field). We have not matched
these sub-spectra to any particular peak in the X3 (CF3) re-
gion of the spectrum. Figures 5 and 6 show expanded views
of the AB (CF2) spectral region, recorded in acetone and to-
luene, respectively. As before, these sub-spectra were ana-
The experimental spectra can now be explained. In ace-
tone (Fig. 5), the two incompletely resolved peaks close to
–118.40 ppm arise from the A parts of sub-spectra 1 and 2,
while the two incompletely resolved peaks between –118.50
and –118.54 ppm are due to the B part of sub-spectrum 1
© 1999 NRC Canada

1740
Can. J. Chem. Vol. 77, 1999
Fig. 7. 19F{1H} NMR spectrum of complex 4 in acetone.
Fig. 6. CF2 19F{1H} NMR spectral region (AB part) of complex
3: (a) experimental spectrum in toluene, (b) simulated spectrum.
Not all impurities could be successfully be removed.
Fig. 8. β-CF2 19F{1H} NMR spectral region of complex 4 in
acetone.
and the A part of sub-spectrum 3. The single peak at
–118.56 ppm is the result of the B part of sub-spectrum 2
lying almost directly over the A part of sub-spectrum 4,
while the remaining single peak at –118.65 ppm results from
the superimposition of the B parts of sub-spectra 3 and 4. In
toluene (Fig. 6), we observed only three small peaks on each
side of the central region, implying that one of the potential
AB sub-spectra has collapsed to a single peak. The two in-
completely resolved peaks near –119.23 ppm correspond to
the A parts of sub-spectra 1 and 2, while their B parts occur
at –119.36 and –119.44 ppm, respectively. The peaks at
–119.52 and –119.58 ppm correspond to the A and B parts,
respectively, of sub-spectrum 3. Finally, the peak at –119.56
ppm corresponds to a CF2 group, where geminal coupling is
not observed (i.e., for spectrum 4, δA = δB).
–125.30 ppm must therefore be assigned to the fluorines of
the β-CF2 groups and, because they are simplest, will be dis-
cussed first.
When the spectral region around δ ≈ –125.3 is expanded
(Fig. 8), four peaks, none of which show evidence of split-
ting, become apparent. These peaks are assigned to the β-
CF2 group, in a mixture of fac and mer isomers. As noted
above, such fluorine atoms are expected to have small, or
zero, coupling to vicinal fluorines of the CF3 and α-CF2
groups. From low to high field, the FWHM values are 4.6,
4.6, 7.7, and 6.3 Hz, respectively. If we take the FWHM
value of the C6F6 calibration peak as 2 Hz, then the upper
limits for the vicinal coupling constants are 0.9, 0.9, 1.9, and
1.6 Hz, respectively. (This is based on the assumption that
these peaks are unresolved 1:5:10:10:5:1 sextets formed by
coupling of the β-CF2 fluorines to the CF3 and α-CF2
fluorines, with both sets of vicinal coupling constants equal,
and that the FWHM of the sextet equals 3Jvic.) The relative
areas of the four peaks (0.96:1.02:1.01:1.00 from low to
high field) indicate that the fac and mer isomers are formed
as a nearly statistical 1:3 mixture. These results are summa-
rized as spectra w–z in Table 2.
-Diketonate complex 4 (RF = n-C3F7)
The spectrum of this complex, recorded in acetone
(Fig. 7), consists of three groups of resonances falling in the
spectral regions centered on δ ≈ –79.54 ppm (relative inten-
sity = 3), δ ≈ –115.80 ppm (relative intensity = 2), and δ ≈
–125.30 ppm (relative intensity = 2). Thus, the resonances
near δ ≈ –79.54 ppm, which show 12 resolved or nearly re-
solved peaks, must be assigned to the fluorine atoms of the
CF3 group. The resonances near δ ≈ –115.80 ppm show evi-
dence of 16 peaks in the central part and 16, or more, small
peaks in the outer parts. These are assigned to the α-CF2
group and will be discussed later. The resonances near δ ≈
© 1999 NRC Canada

Stern and Westmore
1741
Fig. 9. CF3 19F{1H} NMR spectral region (X3 part) of complex
4 in acetone.
Fig. 10. α-CF2 19F{1H} NMR spectral region (AB part) of
complex 4: (a) experimental spectrum in acetone, (b) simulated
spectrum.
Figure 9 shows an expansion of the CF3 region of the
spectrum. The 12 peaks are easily analyzed as four symmet-
rical 1:2:1 triplets of nearly equal intensity. No further split-
ting of any of the peaks is apparent, but there are slight
variations in peak width, which can be assigned to non-zero,
but small, vicinal couplings, in agreement with the magni-
tude of the coupling constants deduced above from the β-CF2
spectral region. To simplify the rest of the analysis, we have
assumed that the only effect of the vicinal couplings is to
broaden the peaks observed in the α-CF2 and CF3 spectral
regions. As for complex 2, a four-bond coupling between the
CF3 fluorines and α-CFA and α-CFB fluorines results in an
ABX3 spin system. When JAX = JBX, the X3 spectrum col-
lapses to the observed symmetrical 1:2:1 triplet. The NMR
parameters for the CF3 region of the spectrum, labeled as
spectra a–d, are given in Table 2. Because the JAX values are
the same within experimental error for each of spectra a–d
and w–z, it will not be possible to match the peaks in this re-
gion to those in the α-CF2 region.
Based on the foregoing, the α-CF2 region of the spectrum
(Fig. 10) corresponds to the AB portions of four ABX3 spec-
tra of approximately equal intensity. Recalling the analysis
given above for complex 2, we see that, in the general case,
each AB region would contain 16 peaks, corresponding to
four AB sub-spectra with spacings equal to 1/2(JAX + JBX),
and of relative intensities 1:3:3:1, of which 8 strong peaks
would be in the central region, and 4 small satellite peaks
would be on either side. Thus, if the chemical shifts of the
four AB sub-spectra are not very different, we expect 32
peaks in the central region, and 16 smaller peaks on each
side. The number of peaks actually observed is much lower
than the maximum. This can be attributed to accidental su-
perimposition of some of the peaks and (or) to the equality
of ∆υ with n{1/2(JAX + JBX)} (n = 1, 2, 3, etc), within an AB
sub-spectrum, where ∆υ is the common separation between
the central peaks of the AB sub-spectra, as before. This was
the case above for complex 2 (n = 1). Depending upon the
value of ∆υAB, the two central quartets may overlap with one
another. Thus, the complicated pattern of overlapping peaks
observed in the central part of the experimental spectrum
can be regarded as eight partially overlapping 1:3:3:1 quar-
tets. Similarly, the outer peaks on each side correspond to
four overlapping 1:3:3:1 quartets. From the recognizable
patterns produced, it was possible to simulate the spectrum,
and to obtain the various chemical shifts and coupling con-
stants given in Table 2. Figure 10 shows the close corre-
spondence between the experimental and the simulated (39)
AB region of the spectrum, and also shows how the peaks in
the experimental spectrum arise. The peaks between –115.61
and –115.74 ppm result from the partial overlap of the cen-
tral A parts of sub-spectra 1 and 2, while those lying be-
tween –115.74 and –115.84 ppm arise from the overlap of
the central B parts of sub-spectra 1 and 2, with coincidence
of several of the peaks. Finally, the peaks between –115.84
and –116.10 ppm result from the closely overlapping central
A and B parts of sub-spectra 3 and 4, again with coincidence
of many peaks.
The spectrum (single peak or AB) of the CF2 group
adjacent to the chelate ring
The foregoing discussion has answered the question of the
basic structures (i.e., fac or mer octahedral) of the com-
plexes. However, the part of the spectrum exhibited by the
CF2 group next to the quasi-aromatic planar chelate ring de-
© 1999 NRC Canada

1742
Can. J. Chem. Vol. 77, 1999
Fig. 11. Conformations considered as rotameric minima for the
CF2R group bound to the planar chelate ring.
molecular planes of symmetry, either for chemically equiva-
lent chelate rings of the fac-octahedral isomer or for any of
the non-equivalent chelate rings of the mer-octahedral iso-
mer. Thus, the two faces of the chelate ring, denoted as α
and β, are not symmetrically equivalent, and in none of the
12 conformations are FA and FB equivalent.
We are now able to predict the type of spectrum exhibited
by the α-CF2 group for different limiting conditions of
interconversion rates and probabilities of the conformations,
and of the relative electronic environment of the two faces of
the chelate ring. In the case of slow rotation (as set by the
NMR experiment), each conformation present in significant
amounts will contribute to the observed spectrum. If one
conformation dominates, then an AB spectrum appropriate
to that conformation will be observed; if several conforma-
tions have a significant presence, then a complex spectrum
containing several overlapping AB contributions will be ob-
served. A single-peak spectrum can result only if the chemi-
cal shifts of FA and FB are the same, as would be the case in
conformations IIIa and IVa, if the two faces of the chelate
ring should be insubstantially different. In fact, we shall
show later that these conformations are unlikely.
In the case of rapid interconversion among conformations,
the time-averaged chemical shifts of the two α-CF2 fluorines
will depend upon the chemical shifts in the conformations
and upon their relative probabilities. Consequently, a single
AB spectrum will normally result, although a single-peak
resonance occurs under certain conditions. It is with the
rapid interconversion scenario that the NMR spectra are con-
sistent.
Preferred conformations and rotation of the perfluoroalkyl
group
For simplicity, we confine the following discussion to the
conditions most likely to prevail. We have found (1) for
tris(thioacetyltrifluoroacetonato)rhodium(III) in the solid
state that the orientation of the CF3 group (i.e., R = F and X
= S) corresponds to identical conformations IV(a, b, or c)
(Fig. 11) in which a C—F bond eclipses the C–H part of the
ring plane. The observed H···F separation of 226 pm is con-
siderably less than the sum (270–305 pm) of the hydrogen
and fluorine van der Waals radii, but also much larger than
the sum (108 pm) of their covalent radii (42) and is within
experimental error of their estimated4 minimum separation
(221 pm) for a rotating CF3 group. If this result represents a
genuine preference for this conformation, possibly as a re-
sult of hydrogen bonding, rather than an artifact of crystal
packing, then it is reasonable to suppose that IVb and IVc
represent low-energy conformations of the CF2R group of
the present molecules.
To assess the energy changes developed on rotation of the
CF2R group, we can note that in benzyl fluoride the con-
former of greatest stability in the gas phase is an eclipsed
form analogous to one of the equivalent forms IVa–c (R = F,
Fig. 11) with a barrier to internal rotation of 1.17 kJ mol–1
(43); however, in acetonitrile and cyclohexane solvents a
perpendicular conformer, analogous to one of the equivalent
forms I/II (R = F), is more stable, with respective barriers to
internal rotation of 3.22 and 0.72 kJ mol–1 in the two sol-
serves further analysis. The CF2CF3-substituted monothio-β-
diketonate complex 1 yields a single-peak resonance for the
CF2 group, whereas the CF2CF2CF3-substituted complex 2
exhibits an AB pattern for the α-CF2 group. For the β-
diketonate complex 3 in acetone, all four symmetrically non-
equivalent CF2CF3 groups of the fac and mer isomers show
AB patterns for the CF2 group, but one of these AB patterns
collapses to a single peak in toluene. Again, for the β-
diketonate complex 4, four AB sub-spectra are exhibited by
the α-CF2 parts of the four non-equivalent CF2CF2CF3 sub-
stituents.
Without, for now, weighing their relative probabilities or
rejecting unlikely conformations, we consider orientations of
the CF2R group with respect to the plane of the chelate ring
that would reasonably correspond to energy minima; in
Fig. 11 the view is along the Cring—CF2R bond, and the
plane of the chelate ring is represented edgewise by the bro-
ken double line. The neighboring CH and X parts of the che-
late ring are also shown. As depicted, conformations I(a–c)
and II(a–c) have the projection of one of the bonds of the
CF2R group perpendicular to the plane of the chelate ring;
conformations III(a–c) have one bond of the CF2R group
eclipsing the C—X bond of the chelate ring, while confor-
mations IV(a–c) have one bond eclipsing the C—CH bond
of the chelate ring. The planes of the chelate rings are not
4 Molecular Modeling Pro program. WindowChem Software Inc., Fairfield, Calif.
© 1999 NRC Canada

Stern and Westmore
1743
where kT/h = 6.25 × 1012 at room temperature (45). Further-
more, the rate would still be very fast even at the lowest
temperatures possible with the solvents used, in agreement
with our variable temperature experiments.
Fig. 12. Energy changes on rotation of the CF2R group. 0°
corresponds to conformation IVa in Fig. 11. The principal barrier
is taken as 9 kJ mol–1 and the 180° barrier as 3 kJ mol–1 (see
text).
Substituent and solvent effects
Chemical shift differences are smaller for the monothio-β-
diketonate than for the β-diketonate complexes — compare 1
with 3, Table 1, and 2 with 4, Table 2 (the difference actu-
ally becomes zero for complex 1). The ring substituent X
could influence the results by producing intrinsic changes in
the FA and FB chemical shifts and (or) by changes in the ro-
tational relative energy minima (i.e., changes in relative
rotamer probabilities. The chemical shift differences are also
solvent dependent — compare the results for complex 2 in
acetone and toluene (Table 1). In toluene, one potential AB
spectrum for a perfluoropropyl group has collapsed to a sin-
gle-peak resonance, thus implying that ∆υAB for FA and FB is
near zero in this solvent. It is more reasonable to attribute
this result to a near equalization of conformation probabili-
ties in toluene than to intrinsic changes in chemical shifts
producing an accidental near-coincidence of FA and FB
chemical shifts. These proposals are reminiscent of results
obtained for substituted benzenes, as, for example, for
benzyl alkyl ethers (29) where the conformations adopted
are sensitive to substituent and solvent effects.
vents. In the case of 1,1,1-trifluoro-2-phenylethane, the sta-
ble conformer is a perpendicular form analogous to Ia/IIa (R
= CF3) with a solvent independent barrier to internal rotation
of 9.02 kJ mol–1 (44).
It is clear from the foregoing discussion that conformer
IVa represents the highest energy form (by ~9 kJ mol–1 with
respect to the potential energy minimum), while all the
remaining 11 conformers differ in energy by no more than
~3 kJ mol–1. Less certain is the choice of low-energy confor-
mations, though the crystal structure evidence favoring an
eclipsed C—F bond, coupled with the preference noted
above for a perpendicular orientation for the C—R bond,
suggests that IVb and IVc are indeed favored choices. If so,
the energy changes for rotation of the CF2R group are of the
form depicted in Fig. 12, where the minima (corresponding
to the R group being on the α or β sides of the chelate ring)
are shown with unequal depths. Interactions of a rotating
CF2R group, in most conformations of the R group with the
X substituent, or with another chelate ring or its substituents,
appear minimal, but interactions with solvent molecules
(themselves influenced by their proximity to different parts
of the complexes) could differ on either side of the chelate
ring (the results do show that the spectra are, indeed, influ-
enced by the solvent) and give the rationale for the unequal
minima in Fig. 12.
This model can be used to explain the spectra generated
by the CF2R groups. When the minima are of unequal depth,
the probability of the two low-energy conformers IVb and
IVc are unequal, and consequently, the time-averaged chem-
ical shifts of FA and FB will differ, and an AB spectrum will
result. As the energies of the minima equalize, the probabili-
ties of the conformers also equalize, the time-averaged
chemical shifts equalize, and a single-peak spectrum results.
The scenario of rapid interconversion among conforma-
tions is consistent with observed chemical shifts. The values
of ∆υAB shown in Tables 1 and 2 range up to nearly 300 Hz.
If these values represent the differences between time-
averaged values of the A and B chemical shifts, weighted for
different populations of conformations, it is reasonable to
propose that the extreme chemical shift difference between
A and B resonances in different conformations is not more
than ~1000 Hz. Thus, interconversion among various confor-
mations must occur faster than this, as is certainly the case.
Even considering an unreasonably high rotational energy bar-
rier of 15 kJ mol–1, a very fast interconversion rate of ~1.5
× 1010 s–1 is predicted from rate = (kT/h) exp (–Ea/RT)
Comparisons with reported structures
In agreement with the observations in this paper, the
structures of tris(monothio-β-diketonate) complexes of vana-
dium(III) (46), iron(III) (41, 47), cobalt(III) (46, 48), techne-
tium(III) (49), ruthenium(III) (41), rhodium(III) (35, 41),
and indium(III) (50) have all been reported as fac-octahe-
dral. Similarly, bis(monothio-β-diketonates) of nickel(II) (41,
51–53), copper(II) (41), palladium(II) (41, 54), and plati-
num(II) (41, 54) have all been reported as cis square planar.
On the other hand, tris complexes formed from unsymmetric
β-diketones and cobalt(III) (46) or ruthenium(II) (55) adopt
mer-octahedral, as well as fac-octahedral structures, with a
preference for the mer-octahedral form of the cobalt(III)
complex (46). Square planar complexes of copper(II) (56),
palladium(II) (57–59), and platinum(II) (59) can adopt the
trans as well as the cis form.
The preferred formation of fac-octahedral and cis square
planar metal complexes from monothio-β-diketones has
prompted the suggestion of weak nonbonding interactions
between the sulfur atoms (46, 51). This proposal is consis-
tent with the relatively short S···S distance (less than twice
the van der Waals radius of sulfur) in a tris(monothio-β-
diketonate) rhodium complex (1). In some complexes, the
adoption of the fac-octahedral structure can also be assisted
by better metal dπ(t2g) → ligand π* bonding between the
metal and sulfur atoms when the three metal–sulfur bonds
are mutually perpendicular. This is consistent with the ob-
served shortening of the Rh—S bonds in a monothio-β-dike-
tonate rhodium complex (1).
Conclusions
In acetone and toluene, tris(monothio-β-diketonates) of
rhodium(III) exist as fac-octahedral complexes whereas
© 1999 NRC Canada

1744
Can. J. Chem. Vol. 77, 1999
31. E. Pitcher, A.D. Buckingham, and F.G.A. Stone. J. Chem.
Phys. 36, 124 (1962).
32. J.A. McCleverty and G. Wilkinson. J. Chem. Soc. 4200
(1964).
33. D.T. Haworth, J.W. Beery, and M. Das. J. Fluorine Chem. 20,
599 (1982).
34. D.T. Haworth. J. Fluorine Chem. 26, 331 (1984).
35. D.T. Haworth and M. Das. J. Fluorine Chem. 20, 487 (1982).
36. M. Das and D.T. Haworth. J. Inorg. Nucl. Chem. 43, 2317
(1981).
tris(β-diketonates) exist as nearly statistical mixtures of fac-
and mer-octahedral complexes. The 19F NMR spectra are
consistent with rapid rotation of the pendant perfluoroalkyl
groups.
Acknow ledgements
We thank Dr. M. Das for gifts of the perfluorinated β-
diketones, T. Wolowiec and M. Jang for recording the NMR
spectra, C. Beaulieu for preparation of sample tubes, and the
Natural Sciences and Engineering Research Council of Can-
ada for financial support.
37. P.J. Toscano, C. Dettelbacher, J. Waechter, N.P. Pavri, D.H.
Hunt, E.T. Eisenbraun, B. Zheng, and A.E. Kaloyeros. J.
Coord. Chem. 38, 319 (1996).
38. M. Das. Inorg. Chim. Acta, 48, 33 (1981).
39. J.S. Martin, A.R. Quirt, and K.E. Worvill. The NMR program
library. Daresbury Laboratory, Daresbury, U.K.
40. R.J. Abraham. The analysis of high resolution NMR spectra.
Elsevier, New York. 1971.
41. M. Das, S.E. Livingstone, J.H. Mayfield, D.S. Moore, and N.
Saha. Aust. J. Chem. 29, 767 (1976).
42. J.E. Huheey, E.A. Keiter, and R.L. Keiter. Inorganic chemistry.
Principles of structure. 4th ed. Harper Collins College Pub-
lishers, New York. 1993. p. 292.
43. T. Schaefer, C. Beaulieu, R. Sebastian, and G.H. Penner. Can.
J. Chem. 68, 581 (1990).
44. T. Schaefer, P. Hazendonk, and D.M. McKinnon. Can. J.
Chem. 73, 1387 (1995).
45. J.W. Emsley, J. Feeney, and L.H. Sutcliffe. High resolution nu-
clear magnetic resonance spectroscopy. Vol. 1. Pergamon
Press, Oxford. 1965.
46. R.H. Holm, D.H. Gerlach, J.G. Gordon, and M.G. McNamee.
J. Am. Chem. Soc. 90, 4184 (1968).
References
1. L.W. Tari, G.A. Stern, J.B. Westmore, and A.S. Secco. Acta.
Crystallogr. Sect. C: Cryst. Struct. Commun. C46, 197 (1990).
2. L. Petrakis and C.H. Sederholm. J. Chem. Phys. 35, 1243
(1961).
3. S. Ng and C.H. Sederholm. J. Chem. Phys. 40, 2090 (1964).
4. R.J. Abraham and L. Cavalli. Mol. Phys. 9, 67 (1965).
5. L. Cavalli and R.J. Abraham. Mol. Phys. 19, 265 (1970).
6. L. Cavalli. J. Magn. Reson. 6, 298 (1972).
7. D.F. Evans. Discuss. Faraday Soc. 34, 139 (1962).
8. N. Boden, J. Feeney, and L.H. Sutcliffe. J. Chem. Soc. 3482
(1965).
9. K.L. Servis and J.D. Roberts. J. Am. Chem. Soc. 87, 1339
(1965).
10. K.L. Williamson and B.A. Braman. J. Am. Chem. Soc. 89,
6183 (1967).
11. M.L. Huggins. J. Am. Chem. Soc. 75, 4123 (1953).
12. R.K. Harris and V.J. Robinson. J. Magn. Reson. 1, 362 (1969).
13. N.F. Ramsey and E.M. Purcell. Phys. Rev. 85, 143 (1952).
14. N.F. Ramsey. Phys. Rev. 91, 303 (1953).
15. H.M. McConnell. J. Chem. Phys. 24, 460 (1956).
16. M.J. Stephen. Proc. R. Soc. London Ser. A, A243, 274 (1957).
17. M. Karplus. J. Chem. Phys. 30, 11 (1959).
18. H.S. Gutowski, M. Karplus, and D.M. Grant. J. Chem. Phys.
31, 1278 (1959).
19. K. Hirao, H. Nakatsuji, H. Kato, and T. Yonezawa. J. Am.
Chem. Soc. 94, 4078 (1972).
20. K. Hirao, H. Nakatsuji, and H. Kato. J. Am. Chem. Soc. 95, 31
(1973).
21. G.E. Maciel, J.W. McIver, N.S. Ostlund, and J.A. Pople. J.
Am. Chem. Soc. 92, 4151 (1970).
22. S.P.S. Porto. J. Mol. Spectrosc. 3, 248 (1959).
23. W.F. Edgell, P.A. Kinsey, and J.W. Amy. J. Am. Chem. Soc.
79, 2691 (1957).
24. R.D. Chambers, L.H. Sutcliffe, and G.J.T. Tiddy. Trans. Fara-
day Soc. 66, 1025 (1970).
25. J. Hilton. J. Magn. Reson. 14, 241 (1974).
26. C. Schumann and F. Cavagna. J. Magn. Reson. 18, 172 (1975).
27. C.W. Haigh, J. Hilton, L.H. Sutcliffe, and G.J.T. Tiddy. J.
Magn. Reson. 18, 241 (1975).
28. S.L. Manatt and M.A. Cooper. J. Am. Chem. Soc. 99, 4561
(1977).
29. T. Schaefer, W. Niemczura, C.-M. Wong, and K. Marat. Can.
J. Chem. 57, 807 (1979).
47. B.F. Hoskins and C.D. Pannan. Inorg. Nucl. Chem. Lett. 11,
409 (1975).
48. J. Ollis, M. Das, V.J. James, S.E. Livingstone, and K.
Nimgirawath. Cryst. Struct. Commun. 5, 679 (1976).
49. G. Bandoli, U. Mazzi, H. Spies, R. Münze, E. Ludwig, E.
Uhlemann, and D. Scheller. Inorg. Chim. Acta, 132, 177
(1987).
50. C. Sreelatha, V.D. Gupta, C.K. Narula, and H. Noth. J. Chem.
Soc. Dalton Trans. 2623 (1985).
51. O. Siiman, D.D. Titus, C.D. Cowman, J. Fresco, and H.B.
Gray. J. Am. Chem. Soc. 96, 2353 (1974).
52. J. Sieler, P. Thomas, E. Uhlemann, and E. Höhne. Z. Anorg.
Allg. Chem. 380, 160 (1971).
53. D.C. Craig, M. Das, S.E. Livingstone, and N.C. Stephenson.
Cryst. Struct. Commun. 3, 283 (1974).
54. E.A. Shugam, L.M. Shkol’nikova, and S.E. Livingstone. Zh.
Strukt. Khim. 8, 550 (1967).
55. S. Aynetchi, P.B. Hitchcock, E.A. Seddon, K.R. Seddon, Y.Z.
Yousif, J.A. Zora, and K. Stuckey. Inorg. Chim. Acta, 113, L7
(1986).
56. P.-K. Hon, C.E. Pfluger, and R.L. Belford. Inorg. Chem. 5, 516
(1966).
57. B. Wenclawiak, A.A. Pinkerton, and N.J. Terrill. Inorg. Chim.
Acta, 149, 213 (1988).
58. P.-K. Hon, C.E. Pfluger, and R.L. Belford. Inorg. Chem. 6, 730
(1967).
59. S. Okeya, S. Ooi, K. Matsumoto, Y. Nakamura, and S.
Kawaguchi. Bull. Chem. Soc. Jpn. 54, 1085 (1981).
30. J.M. Miller, T.R.B. Jones, and G.B. Deacon. Inorg. Chim.
Acta, 32, L75 (1979).
© 1999 NRC Canada