EPR spectra of [Cr(CO)2L(η-C6Me6)]+ (L = PEt3, PPh3, P(OEt)3, P(OPh)3): Analysis of line widths and determination of ground state configuration from interpretation of

Article abstract of DOI:10.1021/om970251p

The preparation and characterization of [Cr(CO)₂L(η-C₆Me₆)] (L = PEt₃, PPh₃, P(OEt)₃ and P(OPh)₃, are reported. One-electron oxidation affords unstable Cr(I) cations, [Cr(CO)

Full text of DOI:10.1021/om970251p

Organometallics 1997, 16, 4369-4376
4369
EP R Sp ectr a of [Cr (CO)2L(η-C6Me6)]+ (L ) P Et3, P P h 3,
P (OEt)3, P (OP h )3): An a lysis of Lin e Wid th s a n d
Deter m in a tion of Gr ou n d Sta te Con figu r a tion fr om
31
In ter p r eta tion of P Cou p lin gs
Michael P. Castellani,† Neil G. Connelly,‡ Robert D. Pike,§,| Anne L. Rieger,§ and
Philip H. Rieger*,§
Department of Chemistry, Marshall University, Huntington, West Virginia 25755, School of
Chemistry, University of Bristol, Bristol BS8 1TS, U.K., and Department of Chemistry,
Brown University, Providence, Rhode Island 02912
Received March 24, 1997X
The preparation and characterization of [Cr(CO)2L(η-C6Me6)] (L ) PEt3, PPh3, P(OEt)3
and P(OPh)3, are reported. One-electron oxidation affords unstable Cr(I) cations, [Cr(CO)2L-
(η-C6Me6)]+, EPR spectra of which are reported. Detailed analysis of the anisotropic 31P
hyperfine interaction indicates that, in frozen CH2Cl2/THF, the phosphine and phosphite
2
2
complexes have A′ and A′′ ground states, respectively. The hyperfine anisotropy can be
accounted for by dipolar interaction of the 31P nucleus with spin density on Cr and, in the
case of the phosphite complexes, with ∼0.01 P 3py spin density resulting from π-backbonding.
Line width anomalies observed in EPR spectra of these and other Cr(I) and Mn(II) “piano
stool” complexes can be understood in terms of molecular distortions resulting from solvation
in frozen solutions.
In tr od u ction
small environmental effects can lead to either ground
state for the tricarbonyl complexes.
The ground state electronic configuration of Cr(I)
“piano-stool” complexes such as [Cr(CO)2L(η-C5R5)], L
) CO, phosphine, or phosphite, has been the subject of
considerable experimental and theoretical interest. The
parent tricarbonyl complexes are orbitally degenerate
in idealized C3v geometry and thus undergo a J ahn-
EPR studies of the phosphine-substituted complexes,
[Cr(CO)2(PPh3)(η-C5H5)]5 and [Cr(CO)2(PMe3)(η-C5Me5)],1
doped into single crystals of the Mn(I) analogs, indicated
2
a A′′ ground state for both complexes, but the LCAO-
HFS-MO study of Fortier et al.1 showed that the 2A′
state lies only 15 kJ mol-1 higher in energy, correspond-
ing to opening of the OC-Cr-CO bond angle from 83
to 100°. Thus, environmental effects may well dictate
the ground state of a phosphine- or phosphite-substi-
tuted complex. The reason for the bond angle depen-
dence of ground state is easily understood: π-overlap
of the metal d-orbitals with CO π-acceptor orbitals
changes significantly with the OC-Cr-CO angle. The
effect on the singly occupied MO (SOMO) and the two
highest energy doubly occupied MOs, computed using
extended Hu¨ckel MO theory,6 is shown in Figure 1 for
2
2
Teller distortion to Cs symmetry and either a A′ or A′′
ground state.1 EPR spectroscopic studies of [Cr(CO)3-
(η-C5H5)]2 and [Cr(CO)3(η-C5Me5)],1 doped into single
crystals of the Mn(I) analogs, indicated 2A′ and 2A′′
ground states, respectively. An LCAO-HFS-MO study1
of [Cr(CO)3(η-C5H5)] showed that the two ground states
differ primarily in the OC-Cr-CO bond angles: the
2
optimum conformation for A′ has angles of 85, 100, and
2
100° whereas A′′ has angles of 92, 84, and 84°. An EPR
study of [Cr(CO)3(η-C5Ph5)] in frozen toluene suggested
the possibility of two conformations in temperature-
dependent equilibrium.3 The single-crystal X-ray struc-
ture of this complex3 showed two somewhat differently
distorted molecules per unit cell; the OC-Cr-CO bond
angles of the two molecules match approximately the
predictions of Fortier et al.1 for the ground state
conformations corresponding to 2A′ and 2A′′.4 Thus,
2
2
the optimized A′ and A′′ structures of Fortier et al.1
The interchange of the SOMO and HOMO is clearly a
result of better π-overlap of the a′′ MO and poorer
overlap of the a′ MO with increasing bond angle.
The 31P hyperfine couplings observed in EPR spectra
of transition metal complexes with phosphorus ligands
hold potentially important information relating to mo-
lecular and electronic structure. In a recent paper,7 we
were able to make some progress in the interpretation
of the isotropic 31P couplings observed in spectra of
octahedral Cr(I) carbonylphosphine and -phosphonite
complexes. Unfortunately, the spectra of these com-
plexes were not sufficiently well resolved to enable the
† Marshall University.
‡ University of Bristol.
§ Brown University.
| Present address: Department of Chemistry, College of William and
Mary, Williamsburg, VA 23185.
X Abstract published in Advance ACS Abstracts, September 15, 1997.
(1) Fortier, S.; Baird, M. C.; Preston, K. F.; Morton, J . R.; Ziegler,
T.; J aeger, T. J .; Watkins, W. C.; MacNeil, J . H.; Watton, K. A.; Hensel,
K.; Le Page, Y.; Charland, J .-P.; Williams, A. J . J . Am. Chem. Soc.
1991, 113, 542.
(4) Extended Hu¨ckel MO calculations based on the coordinates of
the two structures confirm that one is 2A′ and the other 2A′′.
(5) Cooley, N. A.; Baird, M. C.; Morton, J . R.; Preston, K. F.; Le Page,
Y. J . Magn. Reson. 1988, 76, 325.
(6) EHMO calculations used the Alvarez collected parameters
supplied with the CACHE software. CACHE Scientific, Beaverton, OR.
(7) Cummings, D. A.; McMaster, J .; Rieger, A. L.; Rieger, P. H.
Organometallics 1997, 16, 4362.
(2) Krusic, P. J .; McLain, S. J .; Morton, J . R.; Preston, K. F.; Le
Page, Y. J . Magn. Reson. 1987, 74, 72. Morton, J . R.; Preston, K. F.
Cooley, N. A.; Baird, M. C.; Krusic, P. J .; McLain, S. J . J . Chem. Soc.,
Faraday Trans. 1 1987, 83, 3535.
(3) Hoobler, R. J .; Hutton, M. A.; Dillard, M. M.; Castellani, M. P.;
Rheingold, A. L.; Rieger, A. L.; Rieger, P. H.; Richards, T. C.; Geiger,
W. E. Organometallics 1993, 12, 116.
S0276-7333(97)00251-3 CCC: $14.00 © 1997 American Chemical Society

4370 Organometallics, Vol. 16, No. 20, 1997
Castellani et al.
Ta ble 1. An a lytica l, IR Sp ectr oscop ic, a n d
Electr och em ica l Da ta for [Cr (CO)2L(η-C6Me6)]
analysisa (%)
L
color
yellow 61.8 (61.9) 8.4 (8.5)
orange
C
H
IRb/cm-1 ν(CO) E°′/Vc
PEt3
PPh3
1849, 1788
1858, 1797
1868, 1806
1887, 1826
-0.09
0.03
0.10
d
d
P(OEt)3 yellow 55.1 (55.0) 7.8 (7.6)
P(OPh)3 yellow 66.2 (66.2) 5.7 (5.7)
0.32
a
b
Calculated values in parentheses. In CH2Cl2. c In CH2Cl2 at
a platinum disk electrode. Under the conditions used, the potential
d
for the couple [Fe(η-C5H5)2]+-[Fe(η-C5H5)2] is 0.47 V. Prepared
as in ref 12.
Exp er im en ta l Section
The preparation, purification, and reactions of the complexes
described were carried out under an atmosphere of dry
nitrogen or argon using dried, distilled, and deoxygenated
solvents. The compounds [Cr(CO)2L(η-C6Me6)] were prepared
by the UV photolysis of [Cr(CO)3(η-C6Me6)] in the presence of
L (L ) PPh3,10 PhCtCPh11) or by the thermal displacement
of PhCtCPh from [Cr(CO)2(PhCtCPh)(η-C6Me6)] by L.12 IR
spectra were recorded on a Nicolet 5ZDX FT spectrometer.
Electrochemical studies were carried out in CH2Cl2 using an
EG&G Model 273 potentiostat in conjunction with a three-
electrode cell. For cyclic voltammetry, the auxiliary electrode
was a platinum wire and the working electrode a platinum
disk. The reference was an aqueous saturated calomel elec-
trode (SCE) separated from the test solution by a fine-porosity
frit and an agar bridge saturated with KCl. Solutions were
0.1 × 10-3 mol dm-3 in the test compound and 0.1 mol dm-3
in [NBu4][PF6] as the supporting electrolyte. Under these
conditions, E°′ for the one-electron oxidation of [Fe(η-C5H5)2],
added to the test solutions as an internal calibrant, is 0.47 V.
Microanalyses were carried out by the staff of the Microana-
lytical Service of the School of Chemistry, University of Bristol.
Syn th esis of [Cr (CO)2{P (OP h )3}(η-C6Me6)]. A mixture
of [Cr(CO)2(PhC≡CPh)(η-C6Me6)] (134 mg, 0.30 mmol) and
P(OPh)3 (93 mg, 0.30 mmol) in n-hexane (50 cm3) was heated
under reflux for 3 h. The resulting yellow solution was filtered
hot and then allowed to cool to 0 °C. The yellow precipitate
was separated from the mother liquors and then dried in vacuo
to give [Cr(CO)2{P(OPh)3}(η-C6Me6)], yield 102 mg (59%). The
compound is air-stable in the solid state and dissolves in
solvents such as CH2Cl2, thf, and toluene to give moderately
air-sensitive solutions.
The complexes [Cr(CO)2L(η-C6Me6)] (L ) P(OEt)3, PEt3)
were prepared similarly in 30-50% yields after a reaction time
of 24 h and partial removal of solvent from the filtered reaction
mixture before cooling to 0 °C. They may be further purified,
if necessary, by recrystallization from n-hexane. Analytical,
IR spectroscopic, and electrochemical data for the three new
compounds, as well as [Cr(CO)2(PPh3)(η-C6Me6)], are given in
Table 1.
The cations [Cr(CO)2L(η-C6Me6)]+ (L ) PEt3, PPh3, P(OEt)3,
P(OPh)3) were prepared by addition of solid [Fe(η-C5H5)2][BF4]
to a solution of the neutral complex in 1:2 CH2Cl2/THF under
nitrogen; the solution was then transferred by syringe to a
nitrogen-filled EPR tube and frozen. The formation of the
cation was confirmed by IR spectroscopy on a separate aliquot.
X-Band EPR spectra of CH2Cl2/THF glasses were recorded at
90 K using a Bruker ESP-300E spectrometer equipped with a
Bruker variable-temperature accessory and a Hewlett Packard
5350B microwave frequency counter. The field calibration was
checked by measuring the resonance of the diphenylpicrylhy-
drazyl (dpph) radical before each series of spectra. Occasion-
F igu r e 1. The three highest energy occupied MOs, com-
puted using extended Hu¨ckel MO theory, for the conforma-
2
tions of [Cr(CO)2(PH3)(η-C5H5)] found to give A′′ and 2A′
ground states.
accurate measurement of the 31P hyperfine anisotropies.
Since each g-feature in spectra of [Cr(CO)2L(η-C5R5)]
and [Cr(CO)2L(η-C6R6)]+ is a simple 31P doublet, spectra
of these complexes offer a better opportunity to obtain
accurate measures of 31P hyperfine anisotropies. Un-
fortunately, except for L ) PPhMe2, spectra of [Cr(CO)2L-
(η-C5Ph5)]8 are marred by anomalously broad low-field
features which limited the accuracy of measurement of
the 31P splitting. The dilute single-crystal studies of
[Cr(CO)2(PPh3)(η-C5H5)]5 and [Cr(CO)2(PMe3)(η-C5Me5)]1
did not employ enough orientations to determine ac-
curately the 31P hyperfine matrices. The cationic ana-
logs reported here, [Cr(CO)2L(η-C6Me6)]+ (L ) PEt3,
PPh3, P(OEt)3, and P(OPh)3) proved to exhibit spectra
with sufficiently sharp features that complete accurate
measurement of the 31P coupling matrices was possible.
As we will show below, detailed interpretation remains
nontrivial, largely because of the uncertainty in the
orientation of the hyperfine matrix principal axes, but
substantial progress is possible, both in understanding
the origins of the 31P hyperfine couplings and in assign-
ing electronic ground states to the observed radical
species.
The electronic structure insights discussed above lead
to a semiquantitative understanding of the anomalous
variation of component widths observed in frozen solu-
tion EPR spectra of [Cr(CO)2L(η-C6Me6]+ and other Cr-
(I)8 and Mn(II)9 piano stool complexes.
(8) Hammack, D. J .; Dillard, M. M.; Castellani, M. P.; Rheingold,
A. L.; Rieger, A. L.; Rieger, P. H. Organometallics 1996, 15, 4791.
(9) Pike, R. D.; Rieger, A. L.; Rieger, P. H. J . Chem. Soc., Faraday
Trans. 1 1989, 85, 3913.
(10) Strohmeier, W.; Hellmann, H. Chem. Ber. 1964, 97, 1877.
(11) Strohmeier, W.; Hellmann, H. Chem. Ber. 1965, 98, 1598.
(12) Connelly, N. G.; Demidowicz, Z.; Kelly, R. L. J . Chem. Soc.,
Dalton Trans. 1975, 2335.

EPR Spectra of [Cr(CO)2L(η-C6Me6)]+
Organometallics, Vol. 16, No. 20, 1997 4371
ú
gxx ) g +
×
∑
e
k*0∆E0k
2
x
(c0,x2-y2ck,yz
+
3c0,z2
c
k,yz + c0,xzck,xy
)
(2a)
ú
x
gyy ) g +
(c0,x2-y2ck,xz
-
3c0,z2ck,xz -
∑
e
k*0∆E0k
2
x
c0,xzck,x2-y2
+
3c0,xzck,z2
)
(2b)
(2c)
ú
2
gzz ) g +
(2c0,x2-y2
c
k,xy + c0,xzck,yz
)
∑
e
k*0∆E0k
F igu r e 2. X-Band EPR spectrum of [Cr(CO)2(PEt3)(η-C6-
ú
Me6)]+ in CH2Cl2/THF (1:2) at 90 K.
x
gxz
)
(c0,x2-y2ck,yz
+
3c0,z2
c
k,yz + c0,xzck,xy) ×
∑
k*0∆E0k
ally spectra contained extra features attributed to radical
decomposition products; in general, when the oxidation step
was performed quickly and the solution immediately frozen,
clean spectra free of extra features were obtained.
(2c0,x2-y2
ck,xy + c0,xzck,yz) (2d)
contains only one important term, the off-diagonal term
is approximately related to gxx and gzz, gxz ≈ ((gxx - ge)-
(gzz - ge))1/2, so that the principal values are gX ≈ ge, gZ
≈ gzz + gxx - ge, consistent with the observed g-matrices
with gX < gY < gZ.
Resu lts
EPR spectra of [Cr(CO)2L(η-C6Me6)]+ (L ) PEt3, PPh3,
P(OEt)3, P(OPh)3) at 90 K in frozen CH2Cl2/THF show
three well-separated features corresponding to the three
components of the g-matrix. Each feature is split into
a doublet by hyperfine coupling to the 31P nucleus. The
spectrum of the PEt3 derivative is shown in Figure 2,
and EPR parameters are given in Table 2. The low-
field features broadened somewhat as the temperature
was increased, but the increase was less dramatic than
in the case of [Cr(CO)2L(η-C5Ph5)].8 Unlike the C5Ph5
complexes, which exhibited axial powder spectra at
temperatures just above the melting point of the solvent,
the spectra of [Cr(CO)2L(η-C6Me6)]+ collapsed into single
broad lines when the solutions were melted. Room-
temperature spectra of the L ) PPh3 and PEt3 deriva-
tives showed a variety of resonances, none of which
could be associated with the frozen solution spectrum
and thus apparently correspond to decomposition prod-
ucts. No attempt was made to record isotropic spectra
of the L ) P(OEt)3 or P(OPh)3 derivatives.
On the other hand, for a SOMO of a′′ symmetry, eqs
3 for gxx, gzz, and gxz each have two major terms and
ú
gxx ) g +
×
∑
e
k*0∆E0k
(c0,yzck,x2-y2
2
x
+
3c0,yzck,z2 + c0,xyck,xz
)
(3a)
(3b)
ú
2
gyy ) g +
(c0,xyck,yz - c0,yzck,xy)
∑
e
k*0∆E0k
ú
2
gzz ) g +
(2c0,xyck,x2-y2 + c0,yzck,xz
)
(3c)
∑
e
k*0∆E0k
ú
gxz
)
(2c0,xyck,x2-y2 + c0,yzck,xz) ×
∑
k*0∆E0k
x
(c0,xyck,xz
+
3c0,yzck,z2 + c0,yzck,x2-y2) (3d)
Discu ssion
that for gyy, which involves coupling to more distant
MOs of a′′ symmetry, has no major terms. Thus gyy
should be close to ge and may be slightly smaller than
ge as a result of coupling to distant empty MO’s. In
general, gxz < ((gxx - gc)(gzz - ge))1/2, so that gX > ge but
gZ will still be large. Again, the predicted g-matrix is
consistent with those observed.
In either case, the X and Z principal axes of the
g-matrix are displaced from the molecular axes by the
angle âg, given by eq 4.
g-Ma t r ices. The g-matrices for [Cr(CO)2L(η-C6-
Me6)]+ are very similar to those of [Cr(CO)2L(η-C5R5)]1,5,8
and [Mn(CO)2(PPh3)(η-C5H5)]+ 9 with one component
much larger than the free-electron value, a second
component slightly larger than ge, and a third compo-
nent slightly smaller than ge.
The frontier orbitals in d5 piano stool complexes are
essentially the octahedral t2g set, two of a′ symmetry,
eq 1a, and one of a′′ symmetry, eq 1b. For a SOMO of
|k,a′ ) ck,x2-y2|x2 - y2 + ck,z2|z2 + ck,xz|xz + ... (1a)
2gxz
tan 2âg )
(4)
gzz - gxx
|k,a′′ ) ck,yz|yz + ck,xy|xy ...
(1b)
The smallest g-component was found to correspond
to the molecular y-axis (normal to the plane of sym-
metry) in dilute single-crystal spectra of [Cr(CO)2(PPh3)-
(η-C5H5)]5 and [Cr(CO)2(PMe3)(η-C5Me5)],1 so that these
complexes almost certainly have an 2A′′ ground state.
However, because the relative energies of the highest
occupied MO’s of a′ and a′′ symmetry depend strongly
a′ symmetry, gxx and gzz differ from ge through spin-
orbit coupling of the SOMO with the nearby a′′ MO.
Similarly gyy is expected to be somewhat larger than ge
through coupling to the nearby a′ MO. The off-diagonal
component, gxz, also involves a′/a′′ coupling. The result-
ing expressions are given by eqs 2. Since each sum

4372 Organometallics, Vol. 16, No. 20, 1997
Castellani et al.
Ta ble 2. EP R P a r a m eter sa
(a) Cr(I) Complexes
P
P
P
complex
g1
g2
g3
A1
A2
A3
[Cr(CO)2(PEt3)(η-C6Me6)]+
[Cr(CO)2(PPh3)(η-C6Me6)]+
[Cr(CO)2{P(OEt)3}(η-C6Me6)]+
[Cr(CO)2{P(OPh)3}(η-C6Me6)]+
[Cr(CO)2(PPhMe2)(η-C5Ph5)]b
1.993
1.993
1.994
1.993
1.994
2.025
2.026
2.025
2.024
2.013
2.110
2.102
2.109
2.112
2.106
30.4
29.6
39.7
40.0
32.6
32.1
31.5
42.4
42.4
34.8
31.1
30.7
43.8
43.1
34.2
(b) Mn(II) Complexesc
complex
g1
g2
g3
A1
A2
A3
â
[Mn(CO)(dmpe)(η-C5H5)]+
2.000
2.009
2.021
2.014
2.187
2.090
35.6
12.1
≈0
115.3
105.1
45.5°
13.5°
[Mn(CO)2(PBu3)(η5-6-exo-PhC6H6)+
≈0
Estimated accuracy: g-matrix components, (0.001; A-matrix components, (0.2 × 10-4 cm-1
.
Data from ref 8. c Data from ref 9.
a
b
2
on the OC-Cr-CO bond angle, we cannot assume a A′′
ground state for similar complexes in different environ-
ments.
e.g., [Mn(PMe3)(η-C4H6)2], AP ) 25 × 10-4 cm-1 19 and
[Rh2(CO)2(PPh3)2{µ-PhNC(Me)NPh}2]+, where no 31P
coupling is detectable.20 Preston and co-workers18 have
attempted to rationalize some of these effects, but many
of the details are not understood.
In our earlier work on [Cr(CO)2L(η-C5Ph5)],8 we found
axial powderlike spectra from low-temperature liquid
solutions as a result of averaging of the molecular x and
y axes through rotation of the “piano stool legs” with
the very bulky C5Ph5 “seat” essentially stationary.
From the shift of the gZ features, we determined that
the gZ principal axis is displaced from the molecular z
axis (normal to the C5 ring) by âg ≈ 15°, but it was not
possible to determine the direction of tilt, i.e., toward
the phosphorus ligand or away from it. The C6Me6
ligand is not sufficiently bulky to give axial spectra in
low-temperature liquid solutions, but a similar axis
displacement is expected for [Cr(CO)2L(η-C6Me6)]+.
31P Hyp er fin e Cou p lin gs. Examples of 31P hyper-
fine coupling for phosphine or phosphite complexes can
be divided into two classes: (i) Isotropic couplings on
the order of (20-40) × 10-4 cm-1 and small anisotropies
are observed in cases where the P atom lies on or near
a SOMO nodal plane, e.g., in the octahedral V(0)13 and
Cr(I)7 carbonyl phosphine complexes, analogous Mn(II)
complexes,14 and the species described in this paper; the
coupling is probably negative and arises primarily from
spin polarization of P 3s and inner-shell s-orbitals by
spin density in a metal 3d orbital. (ii) Much larger
isotropic couplings and relatively large anisotropies are
often observed when the P atom is located on the z-axis
Here we focus our attention on the 31P hyperfine
anisotropy in spectra of [Cr(CO)2L(η-C6Me6)]+ as well
as [Cr(CO)2(PPhMe2)(η-C5Ph5)].8 There are two poten-
tial sources of hyperfine anisotropy: (1) dipolar coupling
of the 31P nucleus with spin density on Cr (major axis
Z′ along the Cr-P bond); (2) dipolar coupling with spin
density in a P 3p orbital. We assume that the P atom
lies on the plane of symmetry (xz). If the SOMO has
a′′ symmetry, P 3py character is possible and may be
significant for the phosphite ligands (major axis Y′
normal to symmetry plane). Assuming that both aniso-
tropic contributions are axial (i.e., that spin-orbit
coupling corrections are negligible), the 31P hyperfine
components are given by eqs 5 where a and b are given
AX′ ) A - a - b
AY′ ) A - a + 2b
AZ′ ) A + 2a - b
(5a)
(5b)
(5c)
by eqs 6 and 7. The point dipole approximation gives
PCr
r3
Cr
a )
F3d
(6)
(7)
2
and the SOMO is primarily a metal dz orbital, e.g., [Co-
(tpp)(PMe3)], AP ) 222 × 10-4 cm-1 15
,
[Rh2(O2CEt)4-
and [Co(C2Ph2)-
(PPh3)2]+, AP ) 165 × 10-4 cm-1 16
,
2
P
b ) PPF3p
(CO){P(OEt)3}2], AP ) 168 × 10-4 cm-1 17
;
in these
5
cases, the isotropic and anisotropic couplings arise pri-
marily from P 3s and 3pz character, respectively. There
are, however, several exceptions to these generaliza-
tions: thus [W(CO)4{P(OMe)3}]- exhibits a small iso-
tropic coupling and relatively large anisotropy,18 and
some complexes with a phosphine on or near the z axis
PCr ) 10.7 × 10-4 cm-1Å3, and, according to Morton and
Preston,21 PP ) 306 × 10-4 cm-1. For unit spin density
on Cr and a Cr-P distance of 2.38 Å, we expect a ) 0.8
× 10-4 cm-1
.
Since the g-matrix anisotropy is much greater than
that of the hyperfine matrix, the observed features
correspond to orientations along the g-matrix principal
axes. In order to relate the 31P hyperfine components
of eqs 5 to the experimental results, we need the
orientation of the hyperfine matrix principal axes rela-
tive to those of the g-matrix. We assume that the
g-matrix Y-axis and hyperfine matrix Y′ axis are normal
2
and metal dz character show much smaller couplings,
(13) McCall, J . M.; Morton, J . R.; Preston, K. F. Organometallics
1985, 4, 1272.
(14) Carriedo, G. A.; Connelly, N. G.; Perez-Carreno, E.; Orpen, A.
G.; Rieger, A. L.; Rieger, P. H.; Riera, V.; Rosair, G. M. J . Chem. Soc.,
Dalton Trans. 1993, 3103.
(15) Wayland, B. B.; Abd-Elmageed, M. E. J . Am. Chem. Soc. 1974,
96, 4809.
(16) Kawamura, T.; Fukamachi, K.; Hayashida, S. J . Chem. Soc.,
Chem. Commun. 1979, 945.
(17) DeGray, J . A.; Meng, Qingjin; Rieger, P. H. J . Chem. Soc.,
Faraday Trans. 1 1987, 83, 3565.
(18) Hynes, R. C.; Preston, K. F.; Springs, J . J .; Williams, A. J . J .
Chem. Soc., Dalton Trans. 1990, 3655.
(19) McCall, J . M.; Morton, J . R.; Preston, K. F. Organometallics
1985, 4, 1272.
(20) Boyd, D. C.; Connelly, N. G.; Herbosa, G. G.; Hill, M. G.; Mann,
K. R.; Mealli, C.; Orpen, A. G.; Richardson, K. E.; Rieger, P. H. Inorg.
Chem., 1994, 33, 960.

EPR Spectra of [Cr(CO)2L(η-C6Me6)]+
Organometallics, Vol. 16, No. 20, 1997 4373
Ta ble 3. F itted Va lu es of th e 31P Hyp er fin e Ma tr icesa
(a) gmin ) gY
complex
a
b
âA
AX′
AY′
AZ′
[Cr(CO)2(PEt3)(η-C6Me6)]+
[Cr(CO)2(PPh3)(η-C6Me6)]+
[Cr(CO)2{P(OEt)3}(η-C6Me6)]+
[Cr(CO)2{P(OPh)3}(η-C6Me6)]+
[Cr(CO)2(PPhMe2)(η-C5Ph5)]
0.8
0.8
0.8
0.8
0.8
0.8
0.9
1.5
1.3
1.0
33
35
63
53
38
-32.8
-32.3
-44.3
-44.0
-35.7
-30.4
-29.6
-39.7
-40.0
-32.6
-30.7
-29.9
-41.9
-41.3
-33.3
(b) gmin ) gX
complex
a
âA
AX′
AY′
AZ′
[Cr(CO)2(PEt3)(η-C6Me6)]+
[Cr(CO)2(PPh3)(η-C6Me6)]+
[Cr(CO)2{P(OEt)3}(η-C6Me6)]+
[Cr(CO)2{P(OPh)3}(η-C6Me6)]+
[Cr(CO)2(PPhMe2)(η-C5Ph5)]
0.9
0.9
0.4
0.6
0.9
52
57
-32.1
-31.5
-43.0
-42.7
-34.7
-32.1
-31.5
-42.4
-42.4
-34.8
-29.4
-28.8
-40.6
-40.3
-32.0
63
a
Hyperfine couplings in units of 10-4 cm-1
.
to the reflection plane and thus collinear. The Cr-P
bond axis (the hyperfine Z′ axis) is oriented 134° from
the Cr-C5 axis in [Cr(CO)2(PMe3)(η-C5Ph5)], and the
g-matrix Z axis is displaced âg ) (15° from this axis.8
Assuming similar angles for [Cr(CO)2L(η-C6Me6)]+, the
hyperfine Z′ axis is displaced from the g-matrix Z-axis
by âA ≈ 46 ( 15°. The observed splittings, AX and AZ,
then are related to the principal values, AX′ and AZ′ by
eqs 8.
2
2
2
A )
A
x
A
x
cos2 âA + AZ′ sin2 âA
(8a)
(8b)
X
X′
X′
2
A )
sin2 âA + AZ′ cos2 âA
Z
F igu r e 3. Average 31P coupling for [Cr(CO)2L(η-C6Me6)]+,
L ) P(OEt)3, P(OPh)3, PEt3, and PPh3, as a function of
Cr-P bond length in [Cr(CO)5L].
If we assume an a′′ SOMO, so that gmin ) gY, and a ≈
0.8 × 10-4 cm-1, eqs 5 give the values of b, AX′, and AZ′
listed in Table 3a; the values of âA were obtained from
AX′ and AZ′ using eqs 8. The parameter b suggests P
3py spin densities on the order of 0.011 for the phos-
phites and 0.007 for the phosphines. Extended Hu¨ckel
MO calculations6 on the model compounds [Cr(CO)2-
(PH3)(η-C6H6)]+, [Cr(CO)2{P(OH)3}(η-C6H6)]+, and [Cr-
(CO)2(PH3)(η-C5H5)] suggest a 3py spin density of ∼0.005
for P(OH)3 but essentially no P 3py character for PH3.
Thus, a spin density of 0.011 for the phosphites is
plausible, but 0.007 for the phosphines is much larger
than anticipated. The values of the noncoincidence
angle âA are in good agreement with those expected, i.e.,
either ∼31 or ∼61°, but it is puzzling to find one angle
for the phosphines and the other for the phosphites.
Such a result would suggest different signs for gxz for
the phosphine and phosphite complexes, a conclusion
difficult to rationalize in terms of a MO description of
these complexes.
pointed out above, very similar g-matrices are expected
2
2
for A′ and A′′.
In our earlier paper,7 we rationalized the isotropic 31
P
couplings in octahedral Cr(I) carbonylphosphine com-
plexes in terms of eq 9 , where the first term represents
3s
3d
3p
AP ) AP FP + QPCrFCr + QPPFP
(9)
s
the contribution of P 3s spin density, APs ) 4438 × 10-4
cm-1 21
the second term represents polarization of P ns
,
orbitals by spin density on Cr, and the third term
represents polarization of P ns orbitals by spin density
in a P 3p orbital. From an analysis of the 31P couplings,
we obtained an estimate of the metal polarization term
for phosphine complexes, QP ≈ -38 × 10-4 cm-1
.
Cr
Assuming negligible P 3s spin density, this term can
be used to estimate the metal spin densities in the
present phosphine complexes, ∼0.8-0.9, consistent with
estimates from extended Hu¨ckel MO calculations. Since
Alternatively, if we assume that the SOMO has a′
symmetry so that gmin ) gX and b ) 0 (P 3py character
is forbidden for an a′ SOMO), the phosphite data cannot
be fitted to eqs 8; however, the phosphine data give an
entirely sensible fit with a ) 0.9 × 10-4 cm-1 and âA ≈
60° (Table 3b).
the phosphites have isotropic couplings ∼11 × 10-4 cm-1
3p
larger in magnitude than the phosphines, and FP
≈
0.011, assuming the same value of QP and comparable
Cr
metal spin densities for the phosphite complexes leads
to QP ≈ -1000 × 10-4 cm-1
. This value is almost
certainly much too large in magnitude. A more likely
explanation for the larger phosphite couplings is that
P
Considering the parameters a, b, and â, the most
satisfactory results are obtained by assuming gmin ) gY
for the phosphites and gmin ) gX for the phosphines. As
we have seen, this implies a 2A′′ ground state for the
QP increases in magnitude with decreasing Cr-P bond
Cr
length. Figure 3 shows a plot of AP for [Cr(CO)2L(η-
2
C6Me6)]+ as a function of Cr-P bond lengths for
phosphites but A′ for the phosphines. A difference in
ground state symmetry is surprising given the very
similar g-matrices for the five complexes, but as we have
(21) Morton, J . R; Preston, K. F. J . Magn. Reson. 1978, 30, 577.

4374 Organometallics, Vol. 16, No. 20, 1997
Castellani et al.
F igu r e 4. Plots of total energy and the SOMO/HOMO energy difference from EHMO calculations as functions of the
2
2
OC-Cr-CO bond angle for [Cr(CO)2(PH3)(η-C5H5)] in the A′′ (a) and A′ (b) conformations found by Fortier, et al.1
Ta ble 4. Wid th of gz Distr ibu tion s for Cr (I) Com p lexes
a
a
complex
gZ
T/K
whigh
wlow
wg
wg/δgZ
[Cr(CO)2(PEt3}(η-C6Me6)]+
[Cr(CO)2(PPh3)(η-C6Me6)]+
[Cr(CO)2{P(OEt)3}(η-C6Me6)]+
[Cr(CO)2{P(OPh)3}(η-C6Me6)]+
[Cr(CO)2(PMe3)(η-C5Ph5)]b
[Cr(CO)2(PPhMe2)(η-C5Ph5)]b
[Cr(CO)2P(OMe)3}(η-C5Ph5)]b
2.109
2.102
2.111
2.112
2.104
2.106
2.130
90
90
90
8.8
9.9
14.0
14.4
11.0
8.9
3.9
3.2
3.7
3.4
2.2
2.4
2.5
0.0053
0.0062
0.0090
0.0093
0.0071
0.0057
0.017
0.050
0.062
0.082
0.084
0.070
0.055
0.13
90
120
120
125
25
a
b
Gaussian widths in units of gauss. Data from ref 8.
[Cr(CO)5L].22-24 While the correlation is monotonic,
suggesting that the bond length dependence is real, the
phosphine and phosphite points, taken separately, sug-
gest a smaller bond length dependence than that
observed when phosphines and phosphites are com-
pared. Thus either Cr-P(OR)3 bonds are intrinsically
There is literature precedent for the kind of orienta-
tion-dependent line width effect seen in the present
work. Thus, for example, Imagawa26 found orientation-
and quantum number-dependent line widths in spectra
of Cu2+ in various oxide glasses which he attributed to
a distribution of ligand field strengths in the glasses.
Similarly, Griscom27 found that the g features were
anomalously broad in spectra of Cl and Br atoms in
irradiated alkali halides and attributed the broadening
to a distribution in the spacing of the pz and px,y energy
levels, which is responsible for the departure of g from
ge. In general, powder or frozen solution ESR spectra
are subject to environmental variations which may
affect one or more of the spin Hamiltonian parameters.28
Gaussian width parameters29 are given in Table 4 for
the low-field and high-field features in spectra of [Cr-
(CO)2L(η-C6Me6)]+ and [Cr(CO)2L(η-C5Ph5)]. We as-
sume that all spectral features are subject to the same
inhomogeneous broadening mechanism (of unspecified
origin) and that the excess broadening of the low-field
features may be computed using eq 10, where we note
more easily spin polarized than Cr-PR3 bonds or QP
P
is relatively large and negative; quite possibly, both
explanations are correct. A negative sign is at first
surprising since the comparable parameter for carbon
is positive,25 but we note that the analogous Qs are well-
known to be negative for transition metals.
Lin e Wid t h E ffect s in E P R Sp ect r a of Cr (I)
Com p lexes. Given our understanding of the origins
of g-matrix anisotropy, discussed above, and the conclu-
sion that environmental effects can lead to either of two
ground states for [Cr(CO)2L(η-C6Me6)]+ and [Cr(CO)2L(η-
C5Ph5)], it is not difficult to understand the reason for
the anomalous widths of the low-field spectral features
corresponding to gZ. Small deviations of the OC-Cr-
CO bond angle from the average value lead to substan-
tial changes in the SOMO/HOMO energy difference.
EHMO calculations6 were used to model the effect of
wexcess
)
w2 - w2high
(10)
2
2
x
low
small deviations from the optimum A′′ and A′ confor-
mations of Fortier et al;1 the results are shown in Figure
4. The SOMO/HOMO energy difference changes about
5.0 and 4.7% per degree change in bond angle for the
that Gaussian line width contributions combine as the
(22) Cr-P bond lengths for [Cr(CO)5(PPh3)] and [Cr(CO)5{P(OPh)3}]
are from X-ray structures.23 Bond lengths for the PEt3 and P(OEt)3
analogs were estimated from the experimental values for PPh3 and
P(OPh)3 using the molecular mechanics results of Lee and Brown.24
(23) Plastus, H. J .; Stewart, J . M.; Grim, S. O. Inorg. Chem. 1973,
12, 265.
(24) Lee, K. J .; Brown, T. L. Inorg. Chem. 1992, 31, 289.
(25) Karplus, M.; Fraenkel, G. K. J . Chem. Phys. 1961, 35, 1312.
(26) Imagawa, H. Phys. Status Solidi 1968, 30, 469.
(27) Griscom, D. L. Solid State Commun. 1972, 11, 899.
(28) Taylor, P. C.; Baugher, J . F.; Kriz, H. M. Chem. Rev. 1975, 75,
203.
2
2A′′ and A′ conformations, respectively. (The difference
in energies of the SOMO and the next filled MO also
varies with angle, but by a much smaller relative
amount.) Since the SOMO/HOMO energy difference
appears in the denominator of the first terms of eqs 2
and 3 for gxx - ge, gzz - ge, and gxz, these parameters
are expected to vary with bond angle by about the same
amount. After matrix diagonalization, the effect is
concentrated in the maximum g-component, gZ; gZ - ge
then is expected to vary by about 9-10% per degree.
(29) The measured half-width at half-height is related to the
Gaussian width parameter w by ∆B1/2 ) (2 ln 2)1/2w.

EPR Spectra of [Cr(CO)2L(η-C6Me6)]+
Organometallics, Vol. 16, No. 20, 1997 4375
F igu r e 6. X-band EPR spectrum of [Mn(CO)2(PBu3)(η5-
6-exo-PhC6H6)]+ in CH2Cl2/C2H4Cl2 (1:1). (a) Experimental
spectrum at 120 K;9 (b) simulation with parameters of
Table 2 and constant 5.9 G Gaussian line widths; (c)
simulation with constant 5.9 G line widths, wA ) 4.0 ×
F igu r e 5. X-band EPR spectrum of [Mn(CO)(dmpe)(η-
C5H5)]+ in CH2Cl2/C2H4Cl2 (1:1): (a) experimental spectrum
at 120 K;9 (b) simulation with parameters of Table 2 and
constant 4.1 G Gaussian line widths; (c) simulation with
constant 4.1 G line widths, wg ) 0.0049.
10-4 cm-1
.
square root of the sum of squares. The excess width is
attributed to a distribution of gZ values. Consider the
resonant field for a molecule oriented with the Z axis
along the field, B ) hν/gZµB. The contribution to the
observed width of the feature then is given by eq 11,
creases with decreasing |mI|, the 55Mn nuclear spin
quantum number, as shown in Figure 5a. Spectra of
[Mn(CO)2(PBu3)(η5-6-exo-PhC6H6)]+ and other 6-exo-
PhC6H6 complexes also have broad parallel features, but
the widths increase with increasing |mI| at both ends
of the spectrum, as shown in Figure 6a. In either case,
simulations based on constant line widths, shown in
Figures 5b and 6b, give poor accounts of the spectra.
The interpretation of the line width effects in the
spectrum of [Mn(CO)(dmpe)(η5-C5H5)]+ is complicated
by the presence of 55Mn hyperfine coupling and the fact
that the principal axes of the g and 55Mn hyperfine
coupling matrices are displaced in the xz plane by 45°.
Because of the noncoincidence of the g- and A-matrix
principal axes, the various parallel features correspond
to different orientations of the magnetic field in the
g-matrix principal axis coordinate system; these orien-
tations are given in Table 5a. Thus if the excess width
is associated with orientation along the g-matrix Z axis,
we expect the widths to increase in the order mI ) -1/2
< -3/2 < -5/2 , +5/2 < +3/2 < +1/2, and this is indeed
observed. Including the 55Mn hyperfine interaction, the
field position of a spectral feature can be written as in
eq 12, where k is the (angle-dependent) hyperfine
wgB
wexcess ) wg|∂B/∂gz| )
(11)
gz
where wg is the width of the gZ distribution, presumed
to be Gaussian. Values of wg, computed using eq 11,
are given in Table 4. The quantity wg/(gZ - ge), a
measure of the relative width of the gZ distribution, is
on the order of 0.05-0.07 for the phosphine complexes
and 0.08-0.13 for the phosphites. Comparison with the
EHMO results suggests a solvation-induced bond angle
distribution with a width on the order of 1-2°. Note
that the phosphine and phosphite complexes show
somewhat different line width effects, consistent with
the difference in ground state deduced from analysis of
the 31P hyperfine matrices.
If the above explanation is correct, the line width
effect should be solvent- and temperature-dependent.
For [Cr(CO)2L(η-C5Ph5)],8 the low-field line widths
increase with increasing temperature for L ) PMe3 and
P(OMe)3 but decrease for L ) PPhMe2; the line width
effect for L ) P(OMe)3 is much greater in CH2Cl2/C2H4-
Cl2 than in toluene.8 These effects are consistent with
the steric requirements of the very crowded C5Ph5
complexes. Thus, smaller solvent molecules are more
likely to affect the OC-Cr-CO bond angle, and the
PPhMe2 ligand is more likely to exclude solvent mol-
ecules from the crucial region, particularly when its
vibrational amplitude increases. A more extensive
study might well lead to further insights into the nature
of solvation in frozen solutions.
kmI
hυ
B )
-
(12)
geffµB geffµB
coupling and geff is given by eq 13. Differentiating with
2
X
2
2
geff
)
(g cos2 φ + gY sin2 φ) sin2 θ + gZ cos2 θ
x
(13)
BgZ
∂B
) -
cos2 θ
(14)
2
∂gZ
geff
Lin e Wid t h E ffect s in E P R Sp ect r a of Mn (II)
Com p lexes. Spectra of the Mn(II) complexes fall into
two categories. Spectra of [Mn(CO)(dmpe)(η5-C5H5)]+
and other cyclopentadienyl complexes have anomalously
broad low-field “parallel” features, but the width in-
respect to gZ, eq 14, we see that the line width sensitiv-
ity of a feature depends not only on the orientation in
the field but also on the magnitude of the field. Values
of ∂B/∂gz are shown in Table 5a. Taking the width of

4376 Organometallics, Vol. 16, No. 20, 1997
Castellani et al.
distributed with a Gaussian width wA ) (4 ( 1) × 10-4
cm-1. A computer simulation with wA ) 4 × 10-4 cm-1
is shown in Figure 5c.
Ta ble 5. An a lysis of Wid th s of “P a r a llel” F ea tu r es
in Sp ectr a of Mn (II) Com p lexes
(a) [Mn(CO)(dmpe)(η-C5H5)]+ (115 K)
Extended Hu¨ckel MO calculations suggest that the
SOMO in [Mn(CO)2(PBu3)(η5-6-exo-PhC6H6)]+ is of a′
symmetry and is significantly delocalized over the
π-orbitals of carbons 1, 5, and 6 of the 6-exo-PhC6H6
ligand. This is consistent with the EPR parameters, in
particular the average 55Mn coupling, which is notice-
ably smaller than for the cyclopentadienyl analogs.
EHMO calculations further suggest a marked depen-
dence of the delocalization on the dihedral angle de-
scribing the tilt of C-6 out of the plane of the other five
carbon atoms, increasing the Mn 3d spin density by
∼1% per degree increase in the dihedral angle. The
observed component widths suggest a distribution in AZ′
of ∼4%, suggesting a distribution of dihedral angle on
the order of 4°. Variable solvation in the frozen solution
thus appears to tilt the 6-exo-Ph group (and thus C-6)
over a range of angles.
mI
5
3
1
/
/
/
2
-1/2
-3/2
-5/2
2
2
B|/G
θ|/
2865
14.5°
1240
7.2
2962
9.4°
1324
7.5
3053
3.3°
1392
8.2
3432
3532
3636
86.4° 81.1° 76.6°
8
3.1
|∂B/∂gz|/G
46
4.1
0
105
5.4
w/Ga
w
wg
excess/Ga,b 5.9
6.2
7.2
0.0048 0.0047 0.0051
(b) [Mn(CO)2(PBu3)(η5-6-exo-PhC6H6)]+ (115 K)
mI
5
3
1
/
/
/
2
-1/2
-3/2
-5/2
2
2
B|//G
θ|/
2973 3080 3186 3388 3450 3537
7.2°
5.1°
1.52
7.6
4.8
3.2
3.6°
0.49
6.6
3.0
(6.1)
78.8° 59.5° 43.4°
0.19
|∂B/∂Az′|/104 G cm 2.54
0.97
6.3
2.2
2.15
10.8
9.0
w/Ga
11.6
10.0
3.9
w
excess/Ga,c
wA/10-4 cm-1
(2.3)
4.2
a
Gaussian width parameter. b Assuming wx,y ) 4.1 G. c Assum-
ing wx,y ) 5.9 G.
Con clu sion
(1) [Cr(CO)2{P(OEt)3}(η-C6Me6)]+ and [Cr(CO)2-
{P(OPh3}(η-C6Me6)]+ in frozen CH2Cl2/THF have 2A′′
the -3/2 feature as whigh (the width of the mI ) -1/2
feature is anomalously narrow as a result of the
destructive interference of a divergence feature), eq 10
gives the corrected widths, wexcess, listed in Table 5a.
Equation 11 then leads to the values of wg listed in the
table. A simulation incorporating a Gaussian distribu-
tion of gZ with wg ) 0.049 is shown in Figure 4c. The
value of wg obtained for [Mn(CO)(dmpe)(η-C5H5)]+ is
somewhat smaller than those obtained for the Cr(I)
dicarbonyl complexes. This is expected since, for mono-
carbonyl complexes, both the SOMO and HOMO are
expected to be of a′ symmetry with the a′′ MO signifi-
cantly lower in energy; thus solvent-induced changes
in ligand bond angles are expected to have a smaller
relative effect on the a′/ a′′ energy difference and thus
on gZ. Line width effects are considerably more pro-
nounced in the spectrum of [Mn(CO)2(PPh3)(η-C5H5)]+,
consistent with a larger value of wg comparable to those
obtained for the Cr(I) phosphine complexes.
The spectrum of [Mn(CO)2(PBu3)(η5-6-exo-PhC6H6)]+,
shown in Figure 5, is typical of the 6-exo-PhC6H6
complexes in showing an increase in the width of the
parallel features with increasing |mI|. (The monocar-
bonyl complexes show a similar, but usually consider-
ably smaller, effect.) Although there may be a contri-
bution from the gZ distribution, the major effect here
appears to be a distribution in AZ′. Because of the
noncoincidence of the g- and A-matrix principal axes,
an analytical expression for ∂B/∂Az is unwieldy, but this
quantity is readily computed numerically using our
powder pattern analysis program.30 These derivatives
are listed in Table 5b for each of the parallel features,
together with the resonant fields, orientations of the
magnetic field in the g-matrix axis system, and the
measured Gaussian line widths. Assuming an underly-
ing Gaussian line width contribution of 5.9 G (the value
that gives the most consistent fit), we find that AZ′ is
ground state configurations; for these complexes, the 31
P
hyperfine anisotropy is adequately explained in terms
of dipolar coupling of the 31P nucleus to spin density on
Cr and to a small spin density (∼0.011) in the P 3py
orbital resulting from π-backbonding.
(2) [Cr(CO)2(PEt3)(η-C6Me6)]+, [Cr(CO)2(PPh3}(η-C6-
Me6)]+, and [Cr(CO)2(PPhMe2)(η-C5Ph5)] in frozen solu-
tion have 2A′ ground state configurations; the 31P
hyperfine anisotropy is accounted for by dipolar coupling
of the 31P nucleus to spin density on Cr only.
(3) For both the phosphine and phosphite complexes,
the major axis of the 31P hyperfine interaction is along
the Cr-P bond which differs from the g-matrix principal
z-axis by 52-63°, i.e., the C5-Cr-P or C6-Cr-P angle
complement plus âg.
(4) The anomalous broadening of the low-field fea-
tures of EPR spectra of [Cr(CO)2L(η-C6Me6)]+, [Cr(CO)2L-
(η-C5Ph5)], [Mn(CO)2L(η-C5H5)]+, and [Mn(CO)L2(η-
C5H5)]+ can be accounted for by variable distortion of
the OC-M-CO bond angle by solvation in frozen
solutions while the qualitatively different line width
effects observed in spectra of [Mn(CO)2L(η5-6-exo-
PhC6H6)]+ and [Mn(CO)L2(η5-6-exo-PhC6H6)]+ suggest
variable solvent displacement of C-6 relative to the
mean plane of the coordinated C atoms, presumably
through solvent interaction with the 6-exo-Ph group.
Ack n ow led gm en t. We thank the EPSRC for funds
to purchase an EPR spectrometer at the University of
Bristol, the National Science Foundation (Grant CHE-
9123178), and the donors of the Petroleum Research
Fund, administered by the American Chemical Society,
for support of the research at Marshall University. We
are indebted to a reviewer for suggesting the bond
length dependence of the hyperfine parameter, QP
.
Cr
(30) DeGray, J . A.; Rieger, P. H. Bull. Magn. Reson. 1987, 8, 95.
OM970251P