Synthesis and properties of cyclopentadienylniobium(III) complexes. A magneto-structural correlation for 16-electron four-legged piano stool complexes

Article abstract of DOI:10.1021/om960384v

Sodium reduction of (ring)NbCl₄ in the presence of the appropriate phosphine ligand affords the 16-electron Nb(III) complexes (ring)NbCl₂L₂ (ring = C₅H₅, L₂ = dppe or ring = C₅H₄Me, L = PEt₃). While the previously reported Cp*NbCl₂(PMe₃)₂ complex behaves as a Curie - Weiss spin triplet paramagnet, compound (C₅H₄Me)NbCl₂(PEt₃)₂ has a spin singlet ground state and a thermally populated excited triplet state (2.3 kcal/mol higher in energy) and (C₅H₅)NbCl₂(dppe) is essentially diamagnetic. The X-ray structure of (C₅H₄Me)NbCl₂(PEt₃)₂ shows a four-legged piano stool geometry, with Cl-Nb-Cl and P-Nb-P angles correlating with the experimentally determined singlet ground state. Theoretical calculations with geometry optimization at the MP2 level on both ¹A′ and ³A″ states for the (C₅H₅)NbCl2-(PH₃)₂ model system confirm the strong dependence of the angular parameters on the spin state.

Full text of DOI:10.1021/om960384v

Organometallics 1996, 15, 5489-5494
5489
Syn th esis a n d P r op er ties of
Cyclop en ta d ien yln iobiu m (III) Com p lexes. A
Ma gn eto-Str u ctu r a l Cor r ela tion for 16-Electr on
F ou r -Legged P ia n o Stool Com p lexes†
J . C. Fettinger, D. Webster Keogh, Heinz-Bernhard Kraatz, and Rinaldo Poli*
Department of Chemistry and Biochemistry, University of Maryland,
College Park, Maryland 20742
Received May 20, 1996X
Sodium reduction of (ring)NbCl4 in the presence of the appropriate phosphine ligand affords
the 16-electron Nb(III) complexes (ring)NbCl2L2 (ring ) C5H5, L2 ) dppe or ring ) C5H4Me,
L ) PEt3). While the previously reported Cp*NbCl2(PMe3)2 complex behaves as a Curie-
Weiss spin triplet paramagnet, compound (C5H4Me)NbCl2(PEt3)2 has a spin singlet ground
state and a thermally populated excited triplet state (2.3 kcal/mol higher in energy) and
(C5H5)NbCl2(dppe) is essentially diamagnetic. The X-ray structure of (C5H4Me)NbCl2(PEt3)2
shows a four-legged piano stool geometry, with Cl-Nb-Cl and P-Nb-P angles correlating
with the experimentally determined singlet ground state. Theoretical calculations with
3
geometry optimization at the MP2 level on both 1A′ and A′′ states for the (C5H5)NbCl2-
(PH3)2 model system confirm the strong dependence of the angular parameters on the spin
state.
In tr od u ction
molecules, especially the angle between the M-L bonds
and the vector joining the metal with the center (CNT)
of the ring, suggests a dependence on the spin state of
the molecule. Although all molecules are of the four-
legged piano stool type, independent of the spin state,
the CNT-M-X angles (X ) π donor ligand) are larger
for spin singlet molecules (in the range 125-130°)
relative to spin triplet molecules (in the range 111-118);
see Table 1. This difference has been ascribed to the
better overlap of the proper symmetry X lone pair with
Half-sandwich compounds with four monodentate
ligands typically show the four-legged piano stool struc-
tural motif.1-3 Given the 14 electrons located in the
metal-ligand bonds, this structure is allowed for metal
centers with electronic configurations d0 through d4, to
afford complexes with a total electron count of 14
through 18. For the 16-electron (d2) configuration, the
four-legged piano stool structure has been verified
crystallographically for compounds of group 4 metals
in the oxidation state II, e.g. (η6-arene)M[(µ-X)2AlX2]2
(M ) Ti, Zr, Hf; X ) Cl, Br, I)4,5 and (η6-arene)MX2L2
(arene ) toluene; M ) Zr, Hf; X ) Cl, I; L ) PMe3),6-9
for the V(III) compound CpVCl2(PMe3)2,10 and for a few
Mo(IV) systems, e.g. [CpMoCl2(PMe3)2]+ and Cp*MoCl3-
(PR3) (PR3 ) PMe3, PMePh2).11,12 The group 4 M(II)
complexes are all diamagnetic, whereas the isoelectronic
V(III) and Mo(IV) complexes are all spin triplet para-
magnets. An analysis of the structural details of these
the empty dz orbital in the spin singlet compounds.2
2
Several half-sandwich compounds of group 5 metals
in the oxidation state III are known, all those of
vanadium being paramagnetic, e.g. CpVX2(PR3)2 (X )
Cl, alkyl).10,13,14 For systems of Nb(III), on the other
hand, there are no reported structures and no detailed
magnetic studies. Compound Cp*NbCl2(PMe3)2 exhibits
contact shifted 1H-NMR resonances, which led to the
hypothesis of a spin triplet ground state.15 Similar
shifts were subsequently observed for Cp*NbCl2(PMe2-
Ph)2 with a magnetic susceptibility of 2.54 µB. However,
no variable-temperature magnetic or NMR studies
(which could probe the possible thermal population of
excited states with a different spin) were reported.16 In
addition, the NMR properties of the latter compounds
were interpreted as consistent with a pseudo-trigonal
bipyramidal structure, which, although established for
the d0 complex [Cp*WMe4]+, would be unprecedented
for a d2 system.
† Dedicated to the memory of Sir Geoffrey Wilkinson.
X Abstract published in Advance ACS Abstracts, December 1, 1996.
(1) Kuba´cek, P.; Hoffmann, R.; Havlas, Z. Organometallics 1982, 1,
180-188.
(2) Poli, R. Organometallics 1990, 9, 1892-1900.
(3) Lin, Z.; Hall, M. B. Organometallics 1993, 12, 19-23.
(4) Calderazzo, F.; Pampaloni, G. J . Organomet. Chem. 1992, 423,
307-328.
(5) Calderazzo, F.; Pampaloni, G. J . Organomet. Chem. 1995, 500,
47-60.
(6) Green, M. L. H.; Mountford, P.; Walker, N. M. J . Chem. Soc.,
Chem. Commun. 1989, 908-909.
(7) Green, M. L. H.; Walker, N. M. J . Chem. Soc., Chem. Commun.
1989, 1865-1867.
(8) Diamond, G. M.; Green, M. L. H.; Mountford, P.; Walker, N. M.;
Howard, J . A. K. J . Chem. Soc., Dalton Trans. 1992, 417-422.
(9) Diamond, G. M.; Green, M. L. H.; Walker, N. M.; Howard, J . A.
K.; Mason, S. A. J . Chem. Soc., Dalton Trans. 1992, 2641-2647.
(10) Nieman, J .; Teuben, J . H.; Huffman, J . C.; Caulton, K. G. J .
Organomet. Chem. 1983, 255, 193-204.
(13) Hessen, B.; Lemmen, T. H.; Luttikhedde, H. J . G.; Teuben, J .
H.; Petersen, J . L.; Huffman, J . C.; J agner, S.; Caulton, K. G.
Organometallics 1987, 6, 2354-2362.
(14) Hessen, B.; Buijink, J .-K. F.; Meetsma, A.; Teuben, J . H.;
Helgesson, G.; Kakansson, M.; J agner, S.; Spek, A. L. Organometallics
1993, 12, 2268-2276.
(11) Krueger, S. T.; Poli, R.; Rheingold, A. L.; Staley, D. L. Inorg.
Chem. 1989, 28, 4599-4607.
(12) Abugideiri, F.; Keogh, D. W.; Poli, R. J . Chem. Soc., Chem.
Commun. 1994, 2317-2318.
(15) Siemeling, U.; Gibson, V. C. J . Organomet. Chem 1992, 424,
159-161.
(16) de la Mata, J .; Galakhov, M. V.; Gomez, M.; Royo, P. Organo-
metallics 1993, 12, 1189-1192.
S0276-7333(96)00384-6 CCC: $12.00 © 1996 American Chemical Society

5490 Organometallics, Vol. 15, No. 26, 1996
Fettinger et al.
Ta ble 3. Cr ysta llogr a p h ic Da ta for 1
Ta ble 1. An gu la r P a r a m eter s (d eg) in Sp in Sin glet
a n d Tr ip let 16-Electr on F ou r -Legged P ia n o Stool
Com p ou n d s of Typ e tr a n s-(r in g)MX2L2
formula
fw
C18H37Cl2NbP2
479.23
cryst size (mm)
cryst system
space group
a, Å
b, Å
c, Å
0.20 × 0.20 × 0.20
monoclinic
P21/n
8.0735(5)
13.9677(9)
20.458(2)
90.3421(7)
2306.9(3)
4
1.380
0.710 73
0.420
293(2)
compd
S
CNT-M-X CNT-M-L ref
(η6-toluene)ZrCl2(PMe3)2
0
0
0
1
1
1
125.3(1)
130.1(1)
127.9(1)
117.0(0)
118.5(1)
110.9(1)
109.4(1)
109.8(1)
110.3(1)
113.7(0)
113.2(1)
121.6(1)
9
8
6
10
11
43
(η6-C7H8)ZrCl2(PMe3)2
a
(η6-C7H7SiMe3)ZrI2(PMe3)2
CpVCl2(PMe3)2
b
[CpMoCl2(PMe3)2]+
[Cp*MoCl2(PMe2Ph)2]+
â, deg
V, Å3
a
b
Z
C7H8 ) cycloheptatriene. C7H7SiMe3 ) (trimethylsilyl)cy-
cloheptatriene.
D
calc, g/cm
λ(Mo KR), Å
µ(Mo KR), mm-1
temp, K
Ta ble 2. NMR Da ta a
NMRa
θrange, deg
2.47-22.47
(8,-15,+22
3091
data collcd (hkl)
no. reflcns
complex
1H (δ, ppm)
31P (δ, ppm)
T ) 290 K
refinement method
data/restr/param
GOF
full-matrix least-squares on F2
2995/0/215
1.038
1
54.7 (b, w1/2 ) 893 Hz, Cp′)
2.20 (b, w1/2 ) 56 Hz, PEt3)
-21 (b, w1/2 ) 7100 Hz, PEt3)
R1, %
4.98
wR2, %
8.08
T ) 283 K
8.4-7.3 (m, 10 H, Ph2P)
6.6-5.5 (m, 10 H, Ph2P)
5.87 (s, 5 H, Cp)
2
41.4 (s, dppe)
Ta ble 4. Selected Bon d Len gth s (Å) a n d An gles
(d eg) for 1
3.20 (bm, 2 H, PCH2CH2)
2.95 (bm, 2 H, PCH2CH2)
Nb-Cl(1)
Nb-Cl(2)
Nb-P(1)
2.464(2)
2.458(2)
2.625(2)
Nb-P(2)
2.638(2)
2.018
Nb-CNTa
aThe solvent for all NMR determinations was C6D5CD3.
Cl(2)-Nb-Cl(1)
Cl(2)-Nb-P(1)
Cl(1)-Nb-P(1)
Cl(2)-Nb-P(2)
Cl(1)-Nb-P(2)
114.38(7)
81.66(7)
82.45(6)
80.95(7)
81.61(7)
P(1)-Nb-P(2)
Cl(1)-Nb-CNTa
Cl(2)-Nb-CNTa
P(1)-Nb-CNTa
P(2)-Nb-CNTa
148.96(7)
116.9
128.7
104.6
106.3
For these reasons, we have prepared and structurally
and magnetically characterized more examples of half-
sandwich Nb(III) compounds. The results of our studies
reveal the ubiquitous four-legged piano stool structure
for Cp′NbCl2(PEt3)2 (Cp′ ) η5-C5H4Me) and show that
the nature of the ground state for this class of deriva-
tives depends very delicately on the nature of the
ligands.
a
CNT is the center of gravity for C(1)-C(5).
Syn th esis of Cp NbCl2(d p p e) (2). CpNbCl4 (0.637 g, 2.12
mmol) was added to a THF solution (40 mL) of amalgamated
Na (0.104 g, 4.52 mmol in 10 g of Hg) and dppe (0.845 mL,
2.12 mmol). The solution was initially red and turned purple
upon stirring for 0.5 h at room temperature. Stirring was
continued for 2 days, during which time the solution turned
dark red. The solution was evaporated to dryness, the residue
was extracted with CH2Cl2 (3 × 5 mL), and the extracts were
Exp er im en ta l Section
Gen er a l Da ta . All operations were carried out under an
atmosphere of argon. Solvents were dehydrated by conven-
tional methods and distilled directly from the dehydrating
agent prior to use (THF and Et2O from Na/benzophenone,
heptane and toluene from Na, and CH2Cl2 from P2O5). NMR
spectra were recorded on Bruker WP200 and AF200 spectrom-
eters; the peak positions are reported with positive shifts
downfield of TMS as calculated from the residual solvent peaks
(1H) or downfield of external 85% H3PO4 (31P). For each 31P-
NMR spectrum a sealed capillary containing H3PO4 was
immersed in the same NMR solvent used for the measurement,
and this was used as the reference. The elemental analyses
were by M-H-W Laboratories, Phoenix, AZ, or Galbraith
Laboratories, Inc., Knoxville, TN. Cp′NbCl4,17 CpNbCl4,17 and
Cp*NbCl2(PMe3)215,16 were prepared by following the literature
procedures. PMe3 (Aldrich), PMe2Ph (Aldrich), and dppe
(Strem) were used as received. The NMR characterization
data for all new compounds are given in Table 2.
Syn th esis of Cp ′NbCl2(P Et3)2 (1). Cp′NbCl4 (1.031 g, 3.29
mmol) was added to a toluene solution (70 mL) of amalgamated
Na (0.157 g, 6.83 mmol in 17 g of Hg) and PEt3 (0.970 mL,
6.57 mmol). After being stirred overnight, the resulting dark
red-purple solution was evaporated to an oil; the residue was
extracted with heptane (200 mL) and filtered through Celite.
The resulting red-purple solution was placed at -80 °C for 2
h, and a red solid precipitated. The solid was filtered off and
dried under reduced pressure. Yield: 0.936 g, 62%. Anal.
Calcd for C18H37Cl2NbP2: C, 45.1; H, 7.8. Found: C, 44.9; H,
7.9. µeff ) 1.09 µB (Evans’ method in C6D6 at 277 K).
filtered through Celite until the washings were colorless. The
1
CH2Cl2 solution was then concentrated to approximately
/
3
the initial volume and heptane (15 mL) was added, precipitat-
ing a red-brown solid. The solid was filtered off, washed with
heptane (2 × 5 mL), and dried under reduced pressure.
Yield: 0.534 g, 45%. Anal. Calcd for C31H29Cl2NbP2: C, 59.3;
H, 4.7. Found: C, 58.7; H, 4.6.
X-r a y An a lysis of Com p ou n d 1. Single crystals were
obtained by cooling a concentrated heptane solution of 1 to
-80 °C. A single crystal was glued to the inside of a thin-
walled glass capillary, which was then flame sealed under
dinitrogen and mounted on the diffractometer. The cell
parameters and crystal orientation matrix were determined
from 25 reflections in the range 13.1 < θ < 18.5°; these
constants were confirmed with axial photographs. Data were
collected with ω/2θ scans over the range 2.0 < θ < 22.5°. The
position of the Nb atom was obtained from the analysis of the
Patterson map, and the positions of all the other heavy atoms
(Cl and P) were revealed by a subsequent DIRDIF run. The
structure was then refined by alternating least-squares cycles
and difference Fourier maps, revealing the positions of all
other non-hydrogen atoms, to convergence with all non-
hydrogen atoms anisotropic. Hydrogen atoms were finally
included in calculated position and used for structure factor
calculations, but not refined. Crystal data are reported in
Table 3, and selected bond distances and angles are collected
in Table 4.
(17) Cardoso, A. M.; Clark, R. J . H.; Moorhouse, S. J . Chem. Soc.,
Dalton Trans. 1980, 1156.

Cyclopentadienylniobium(III) Complexes
Organometallics, Vol. 15, No. 26, 1996 5491
Th eor etica l Ca lcu la tion s. Calculations were carried out
with the Gaussian 94 package.18 The geometry optimizations
were carried out with the LANL2DZ basis set without polar-
ization basis functions, which include both Dunning and Hay’s
D95 sets for H and C19 and the relativistic ECP sets of Hay
and Wadt for the heavier atoms.20-22 Electrons outside of the
core were all those of H and C atoms, the 4s, 4p, 4d, and 5s
electrons in the Nb atom, and the 3s and 3p electrons in Cl
and P atoms. The input coordinates for the CpNbCl2(PH3)2
model compound were adapted from the X-ray structure of
compound 1, which was idealized to Cs symmetry. The mean
value of the spin of the first-order electron wave function,
which is not an exact eigenstate of S2 for unrestricted Har-
tree-Fock calculations on open-shell systems, was considered
to identify unambiguously the spin state. The value of <S2>
for the UHF calculation on the triplet state was 2.0274 at
convergence, indicating minor spin contamination. A single-
point energy calculation on the geometry optimized as de-
scribed above was also carried out for both spin states by using
the same functions for H, C, and Nb atoms, and the 6-311G
basis set,23,24 supplemented with diffuse functions and 3 sets
of d and 1 set of f polarization functions for the Cl and P atoms.
The value of <S2> for the UHF calculation on the triplet state
in this case was 2.0285.
Resu lts
Syn th etic Stu d ies. Compounds with the general
formula (ring)NbCl2L2 [ring ) C5H4Me (Cp′), L ) PEt3
(1); ring ) Cp, L2 ) dppe (2)], which are formally 16-
electron, d2, Nb(III) complexes, have been synthesized
by reducing (ring)NbCl4 in the presence of 2 equiv of
phosphine; see eq 1. The products were isolated as
crystalline materials and characterized by 1H- and 31P-
NMR spectroscopy and by elemental (C,H) analysis.
Compound 1 was also characterized by X-ray crystal-
lography.
F igu r e 1. Side view (a) and top view (b) of 1. Thermal
ellipsoids are drawn at 40% probability level.
synthesis of other systems with intermediate steric
crowding in the coordination sphere, e.g. (ring)-
NbCl2(PMe2Ph)2 (ring ) Cp′ and Cp), was attempted
but only impure mixtures (presumably containing the
16-electron and 18-electron Nb(III) species and also Nb-
(IV) species of incomplete reduction, as shown by a
combination of NMR and EPR spectroscopies) were
obtained under a variety of different conditions, and the
preparation of these systems was therefore abandoned.
As mentioned in the Introduction, 16-electron four-
legged piano stool complexes can have one of two
possible ground states of different spin, a singlet and a
triplet. Preliminary NMR investigations for the isolated
complexes indicate that compound 2 is diamagnetic,
while compound 1 is paramagnetic with contact shifts
comparable to those reported previously for Cp*NbCl2L2
(ring)NbCl4 + 2Na + 2L THF or toluene8
(ring)NbCl2L2 + 2NaCl (1)
ring ) Cp′, L ) PEt3; ring ) Cp, L2 ) dppe
Half-sandwich compounds of Nb(III) with phosphine
ligands that are already reported in the literature are
either of the 18-electron type when the ligands are
relatively unencumbering, e.g. CpNbCl2(PMe3)3,25 or of
the 16-electron type for bulkier systems, e.g. Cp*NbCl2L2
(L ) PMe3,15 PMe2Ph2,16 and L2 ) dmpe16). Compound
1 does not appear to lead to an 18-electron derivative
when excess PEt3 is added (see later); thus, the combi-
nation of the Cp′ and PEt3 ligands is already sufficient
to prevent the achievement of a saturated structure. The
16
(L ) PMe3,15 PMe2Ph2 and L2 ) dmpe16); see Table 2.
Detailed variable-temperature NMR studies, however,
reveal a more complex and interesting situation (vide
infra).
(18) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
J ohnson, B. G.; Robb, M. A.; Cheeseman, J . R.; Keith, T. A.; Petersson,
G. A.; Montgomery, J . A.; Raghavachari, K.; Al-Laham, M. A.;
Zakrzewski, V. G.; Ortiz, J . V.; Foresman, J . B.; Cioslowski, J .;
Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala,
P. Y.; Chen, W.; Wong, M. W.; Andres, J . L.; Replogle, E. S.; Gomperts,
R.; Martin, R. L.; Fox, D. J .; Binkley, J . S.; Defrees, D. J .; Baker, J .;
Stewart, J . P.; Head-Gordon, M.; Gonzales, C.; Pople, J . A. Gaussian
94 (Revision A1); Gaussian Inc.: Pittsburgh, PA, 1995.
(19) Dunning, T. H., J r.; Hay, P. J . In Modern Theoretical Chemistry;
Schaefer, H. F., III, Ed.; Plenum Press: New York, 1976; pp 1-28.
(20) Hay, P. J .; Wadt, W. R. J . Chem. Phys. 1985, 82, 299-310.
(21) Wadt, W. R.; Hay, P. J . J . Chem. Phys. 1985, 82, 284-298.
(22) Hay, P. J .; Wadt, W. R. J . Chem. Phys. 1985, 82, 270-283.
(23) McLean, A. D.; Chandler, G. S. J . Chem. Phys. 1980, 72, 5639.
(24) Krishnan, R.; Binkley, J . S.; Seeger, R.; Pople, J . A. J . Chem.
Phys. 1980, 72, 650.
X-r a y str u ctu r e of Cp ′NbCl2(P Et3)2. Two different
ORTEP views of 1, the first 16-electron CpNbIII complex
to be crystallographically characterized, are shown in
Figure 1. Selected bond lengths and angles are given
in Table 4. The molecule exhibits a typical four-legged
piano stool geometry, with an η5-bound Cp′ ring and a
relative trans configuration for the Cl2(PEt3)2 set of
ligands. The molecule is structurally identical to a well-
studied and documented class of compounds of the form
CpMoX2L2 (L ) tertiary phosphine).26 A comparison of
the bond lengths and angles of this complex with the
(25) Alt, H. G.; Engelhardt, H. E. Z. Naturforsch. 1989, 44b, 367-
368.
(26) Poli, R. J . Coord. Chem. B 1993, 29, 121-173.

5492 Organometallics, Vol. 15, No. 26, 1996
Fettinger et al.
27
recently described CpMoCl2(PEt3)2 results in the
observation of a lengthening of the M-P bonds (2.632-
(2) Å for 1 versus 2.532(1) Å for the Mo analogue), which
can be partially attributed to the larger covalent radius
of the Nb(III) versus the Mo(III). A smaller effect is
observed for the M-CNT bond length with 2.018(2) Å
for Nb-CNT in 1 and 1.940(2) Å for the Mo-CNT in
CpMoCl2(PEt3)2. An 18-electron CpNb(III) isocyanide
complex, Cp*NbCl2(CN-2,6-Me2C6H3)3, has also been
crystallographically characterized, and a Nb-CNT bond
length was found to be 2.099(5),28 which is slightly
longer than that found for 1. In contrast to these
observations, the average M-Cl bond distance is slightly
shorter for the Nb complex (2.461(2) Å versus 2.476(1)
Å for the Mo analogue). The 18-electron Nb isocyanide
complex, Cp*NbCl2(CN-2,6-Me2C6H3)3, shows a signifi-
cantly longer Nb-Cl bond, 2.561(2) Å.28 All these trends
may be explained on the basis of changes in M-Cl and
M-PEt3 π-bonding. The M-PEt3 π-back-bonding in-
teraction is favored by a greater number of electrons in
the metal-based orbitals (3 for Mo(III) and 2 for Nb-
(III)), whereas the M-Cl π-interaction is favored by a
smaller number of electrons in the metal-based orbitals.
For the 18-electron Cp*NbCl2(CN-2,6-Me2C6H3)3, there
are no empty metal orbitals available for the establish-
ment of a Nb-Cl π-interaction. Another interesting
difference between the CpMCl2(PEt3)2 (M ) Nb, Mo)
structures concerns the CNT-M-L angles. The two
CNT-Nb-Cl angles (116.9(1) and 128.7(1)°) are quite
different from each other, presumably because of a steric
repulsion of the Cp′ methyl group on the chlorine atom
which is eclipsed with it (see Figure 1b). The Mo
system, having a Cp ring, does not experience the same
steric repulsion, and therefore the angles are similar
(118.0(1) and 121.6(1)°). The average CNT-M-Cl
angle for both systems is comparable [122.8(1)° for the
Nb and 119.8(1)° for the Mo system]. This angle,
however, is significantly greater with respect to the
corresponding angle of the isoelectronic vanadium sys-
tem, CpVCl2(PMe3)2.10 As stated in the Introduction,
this angle appears to be diagnostic of the spin state (see
Table 1); the angle observed for compound 1 falls in
between those observed for diamagnetic compounds and
for compounds that have two unpaired electrons, war-
ranting further magnetic investigations by variable-
temperature NMR spectroscopy.
F igu r e 2. Plot of δ versus 1/T for Cp*NbCl2(PMe3)2.
the phosphorus atom due to the vicinity of the quadru-
polar Nb nucleus (I ) 9/2, 100% abundance, Q ) -0.366-
(18) × 10-24 cm2, vs Q ) -0.00282(19) × 10-24 cm2 for
2H).29 A possible slight population of a paramagnetic
excited state is ruled out, since the position of the 31P-
NMR resonance does not shift upon further cooling to
213 K. Few diamagnetic Nb phosphine compounds
appear to have been investigated by 31P-NMR spectros-
copy, and these commonly exhibit broad resonances at
room temperature.25,30,31
As mentioned above, preliminary room-temperature
1H-NMR studies indicate paramagnetism for compound
1. The resonance at 59.7 ppm is tentatively assigned
to the Cp′ methyl protons, while the Cp′ ring protons
1
do not appear to be visible in the H-NMR spectrum.
This is the same phenomenon previously observed for
(ring)MoCl3L compounds: the Cp protons (e.g. for
CpMoCl3(PMe2Ph)) are not observed in the 1H-NMR
spectrum, while the Cp* protons for Cp*MoCl3(PMe2-
Ph) are observed at δ 0.6.32 The two additional reso-
nances at δ -21 and 2.20 for 1 are tentatively assigned
to the methylene and methyl protons of the PEt3 ligand,
respectively. Lowering the temperature, however, re-
sulted in a shift of all resonances toward the diamag-
netic region, contrary to the case of Cp*NbCl2(PMe3)2.
This phenomenon is consistent with a singlet ground
state (as suggested by the X-ray structure) and a
thermally populated triplet excited state. According to
the literature, the temperature-dependent chemical
shifts for a spin singlet/triplet equilibrium system are
related to fundamental molecular properties as shown
in eq 2.33-35
NMR Stu d ies. The magnetic properties of com-
pounds 1, 2, and the previously reported15,16 Cp*NbCl2-
1
(PMe3)2 were probed by variable-temperature H-NMR
6e-E/kT
gâH0A
studies. For compound Cp*NbCl2(PMe3)2, the contact-
shifted resonances for both the Cp* and the PMe3
protons further shifted away from the diamagnetic
region upon cooling. The observed linearity of the plots
of δ versus 1/T (see Figure 2) is consistent with Curie-
Weiss behavior for a spin triplet paramagnet. For
compound 2, on the other hand, a temperature-inde-
pendent 1H-NMR spectrum in the normal region (see
Table 2) indicates a pure diamagnetic system. No
resonances were observed in the 31P-NMR spectrum at
room temperature, but a singlet resonance at δ 41
becomes observable upon cooling to 283 K. The latter
phenomenon could be attributed to a fast relaxation of
δobs ) δdia
+
(2)
-E/kT
3kT(γ /2π)
1 + 3e
H
The constants in eq 2 have their usual meanings,
while δdia is the non-contact-shifted chemical shift, A is
the hyperfine electronic coupling constant, and E is the
(29) Raghavan, P. At. Data Nucl. Data Tables 1989, 42, 189-292.
(30) Schrock, R. R. J . Organomet. Chem. 1976, 21, 373-379.
(31) J amieson, G.; Lindsell, W. E. Inorg. Chim. Acta 1978, 28, 113-
118.
(32) Abugideiri, F.; Gordon, J . C.; Poli, R.; Owens-Waltermire, B.
E.; Rheingold, A. L. Organometallics 1993, 12, 1575-1582.
(33) Kriley, C. E.; Fanwick, P. E.; Rothwell, I. P. J . Am. Chem. Soc.
1994, 116, 5225-5232.
(34) Cotton, F. A.; Chen, H. C.; Daniels, L. M.; Feng, X. J . Am. Chem.
Soc. 1992, 114, 8980-8982.
(35) Cotton, F. A.; Eglin, J . L.; Hong, B.; J ames, C. A. J . Am. Chem.
Soc. 1992, 114, 4915-4917.
(27) Cole, A. A.; Fettinger, J . C.; Keogh, D. W.; Poli, R. Inorg. Chim.
Acta 1995, 240, 355-366.
(28) Alcade, M. I.; de la Mata, J .; Gomez, M.; Royo, P. Organome-
tallics 1994, 13, 462-467.

Cyclopentadienylniobium(III) Complexes
Organometallics, Vol. 15, No. 26, 1996 5493
Ta ble 5. Geom etr ica l P a r a m eter s for MP 2
Geom etr y-Op tim ized , 16-Electr on Cp NbCl2(P H3)2
calculated
CpNbCl2(PH3 2
)
exptl
Cp′NbCl2(PEt3)2
3A′′
2.483
2.575
2.706
1A′
2.374
2.537
2.656
Nb-C(Cp)(av)/Å
Nb-Cl (av)/Å
Nb-P(av)/Å
2.343(7)
2.461(2)
2.631(2)
Cl-Nb-P(av)/deg
Cl-Nb-Cl/deg
P-Nb-P/deg
81.7(7)
114.38(7)
148.96(7)
78.98
80.87
135.73
119.02
112.13
120.49
111.14
146.97
124.28
106.50
CNT-Nb-Cl(av)/deg 122.8(1)
CNT-Nb-P(av)/deg
105.4(1)
E(LANL2DZ)a
-294.3316 -294.3234
-294.9550 -294.9563
E(LANL2DZ/6-311G)a
a
Energies are in hartree units.
18-electron adduct in equilibrium with 1 and a rapid
PEt3 exchange (on the NMR time scale) for compound
1.
1
Th eor et ica l Ca lcu la t ion s. The model system,
CpNbCl2(PH3)2, was used for the theoretical calculations
in both possible spin states, singlet or triplet. The
calculations consisted of unrestricted open-shell SCF
followed by a second-order Møller-Plesset (MP2) ge-
ometry optimization (see Experimental Section).
The optimized geometrical parameters are collected
F igu r e 3. Plots of δ versus T for the H-NMR spectrum
of 1. Upper and lower plots correspond to the resonances
assigned to the Cp′ methyl protons and PEt3 methyl
protons, respectively.
singlet-triplet energy gap.36-40 Using the Curve Fit
program,41 a nonlinear least-squares fitting (quasi-
Newtonian method) of the data to eq 2 allows the
determination of the parameters δdia, A, and E. Two of
the resonances were sufficiently sharp to allow accurate
measurements and independently yielded an identical
energy gap within the experimental error [2.24(11) and
2.3(7) kcal/mol], as expected. Figure 3 shows plots of
the chemical shift versus temperature for these two
resonances, superimposed to the curve fits. The elec-
tronic coupling constant depends on the interaction of
the nucleus with the unpaired electron; thus a more
contact-shifted peak is expected to have a larger cou-
pling constant. This trend was observed for 1, with the
Cp′ methyl resonance yielding A ) -14(2) MHz and the
phosphine methyl resonance giving A ) -0.4(2) MHz.
The unperturbed diamagnetic shifts, δdiag, are 1(2) and
0.8(4) ppm, respectively, for the Cp′ methyl and PEt3
methyl protons. These values are in agreement with
expectations for the chemical shifts of such protons in
a diamagnetic environment, supporting the resonance
assignment.
1
in Table 5. All of the geometric parameters for the A′
system compare extremely well with those experimen-
tally found for 1, whereas the parameters optimized for
3
the A′′ system are quite different. This indicates that
the experimentally determined singlet ground state for
3
1 in solution is also retained in the solid state. The A′′
state shows, as expected, slightly longer Nb-ligand
distances relative to the 1A′ state. The most interesting
difference, however, is the variation of the CNT-Nb-
Cl and CNT-Nb-P angles as a function of the spin
state. The CNT-Nb-Cl angle closes from 124.28 to
112.13° upon going from the singlet to the triplet state,
while the CNT-Nb-P angle correspondingly opens
from 106.50 to 120.49°. This trend follows the experi-
mentally determined angles in spin singlet and spin
triplet systems (see Table 1) and can be attributed to
the change of Cl-Nb π-interactions, as previously
discussed in the literature.2,3
The 3A′′ state is calculated as the ground state at this
level of theory, while the 1A′ state is 5.15 kcal/mol higher
in energy or 6.02 kcal/mol when using the spin-projected
wave function. The small difference between these two
numbers is additional indication of the small spin
contamination of the wave function from the unre-
stricted open-shell calculation (see Experimental Sec-
tion). A single point calculation at the optimized
geometry by including diffuse and polarization functions
on the Cl and P atoms gave a smaller difference: the
singlet state is now more stable than the triplet by 0.82
kcal, whereas the triplet is more stable by 0.25 kcal if
the spin-projected wave function is used.
The addition of excess free PEt3 to a solution of
compound 1 did not show any evidence for coordination
with formation of a (diamagnetic) 18-electron com-
pound: The 1H-NMR spectrum of a solution of com-
pound 1 to which was added a large excess of free PEt3
showed only the unperturbed resonances of 1 together
with those of free PEt3 and no new resonances in the
diamagnetic region that could be attributable to the
hypothetical 18-electron Cp′NbCl2(PEt3)3. This result
rules out both the presence of a significant amount of
(36) Holm, R. H.; Hawkins, C. J . NMR of Paramagnetic Molecules;
Academic Press: New York, 1973.
(37) Boersma, A. D.; Phillipi, M. A.; Goff, H. M. J . Magn. Reson.
1984, 57, 197.
(38) Campbell, G. C.; Haw, J . F. Inorg. Chem. 1988, 27, 3706.
(39) Campbell, G. C.; Reibenspies, J . H.; Haw, J . F. Inorg. Chem.
1991, 30, 171.
(40) Hopkins, M. D.; Gray, H. B.; Miskowski, V. M. Polyhedron 1987,
6, 705.
Discu ssion
1
The variable-temperature H-NMR experiments pre-
sented here establish a Curie-Weiss behavior for the
previously reported15,16 spin triplet Cp*NbCl2(PMe3)2
compound. Conversely, compound 2 has a spin singlet
ground state without appreciable population of a triplet
excited state. Finally, compound 1 has an intermediate
(41) Raner, K. Curve Fit, version 0.7e; Clayton, Victoria, Australia,
1992.

5494 Organometallics, Vol. 15, No. 26, 1996
Fettinger et al.
Sch em e 1
predict that compounds such as CpNbCl2L2 with small
phosphine ligands should be diamagnetic.
The results of the theoretical calculations (e.g. a more
stable triplet state for the unencumbered CpNbCl2-
(PH3)2 model complex) presented in this work must not
be overinterpreted, since the two spin states experience
a different degree of electronic correlation and this is
only estimated as a second-order perturbation on the
monoelectronic Hamiltonian by the MP2 calculation.
However, the small calculated triplet-singlet gap of
5.15 kcal/mol (or -0.82 kcal/mol with an expanded basis
set on Cl and P) qualitatively agrees with the delicate
balance of this system. The more positive result of the
computational work is that the optimized geometries for
both spin states follow the experimentally determined
correlation of the CNT-M-X angles and the spin state
for a 16-electron four-legged piano stool geometry.
As a final note, it is interesting to compare the
magnetic properties of isoelectronic and isostructural
four-legged piano stool Zr(II), Nb(III), and Mo(IV)
systems. On one side, the Zr(II) systems6-9 are all
diamagnetic and their structures show large CNT-
Zr-X (X ) halogen) angles. On the other side, all the
Mo(IV) systems11,12 have a triplet ground state and their
structures show small CNT-Mo-Cl angles (see Table
1). In between these two extremes, the Nb(III) systems
examined in this work exhibit an intermediate behavior.
This trend is consistent with the expected change of
pairing energy (Zr2+ < Nb3+ < Mo4+), correlating with
variations of the effective nuclear charge along this
series of isoelectronic 4d2 ions.
behavior, showing a thermal equilibrium between a
ground state singlet and a thermally populated triplet.
The variation of magnetic properties along the examined
series of Nb(III) compounds deserved a detailed analy-
sis. The choice of the ground state depends, in simple
terms, on the competing effects of the pairing energy,
PE, which favors a high spin arrangement, and the
orbital splitting, which favors a spin-paired situation.
The nature of the ligands could have an effect on either
or both parameters.
The pairing energy is probably affected only by the
electron-donating power of the phosphine and ring
ligands, the better donor ligands expanding the metal
orbitals and lowering PE. However, the electron-richest
system, e.g. Cp*NbCl2(PMe3)2, is a pure paramagnet,
whereas the opposite would be predicted by the pairing
energy variation argument. A better correlation can be
found on the basis of a variation of orbital splitting,
which can be indirectly brought about by structural
changes enforced by steric effects. As discussed in the
literature,1-3 the M-Cl π-interaction diagram is as
shown in Scheme 1. The major π interaction between
the two Cl atoms and the metal is established between
Con clu sion s
The synthesis of open-shell intermediate oxidation
state organometallic niobium complexes has been ac-
complished by reduction of higher valent CpNb chlorides
in the presence of phosphines. Within the general class
of complexes, (ring)NbCl2L2 (L ) tertiary phosphine),
remarkably different magnetic properties can be found,
ranging from pure spin triplet paramagnets to spin
singlet/triplet equilibrium systems and finally to pure
diamagnetic systems. The examples provided in this
work, combined with isoelectronic systems of Zr(II) and
Mo(IV), are consistent with the expected increase of
pairing energies for higher oxidation states and reveal
a correlation between the CNT-M-X angle (X )
π-donor ligand) and the observed magnetic properties,
with those complexes having large angles being dia-
magnetic. This is attributed to a better overlap of the
2
the metal dz orbital and a symmetry-adapted linear
combination of two Cl lone pairs, which has the effect
of increasing the gap between the two frontier orbitals.
The interaction is maximized at a CNT-M-Cl angle
of 135°. A large steric interaction between the ring and
the phosphine ligands forces the CNT-M-L angle to
be large. This, in turn, induces a small CNT-M-Cl
angle because of the σ-rehybridization as detailed by Lin
and Hall,3 ultimately reducing the ∆ gap and favoring
higher spin configurations. This effect fully rationalizes
the observed trend, the bulkier systems having a
preference to be paramagnetic. This correlation is
applicable only for systems where the X2L2 set of ligands
adopts a relative trans configuration, for which the
angular trans influence operates.2,3 Compound 2 pre-
sumably has a cis configuration [as established crys-
tallographically for the similar CpMoBr2(dppe)].42 On
the basis of this correlation, however, it is possible to
2
donor lone pairs with the metal dz orbital .
Ack n ow led gm en t. We are grateful to the National
Science Foundation (Grant CHE-9508521) for support
of this work.
Su p p or tin g In for m a tion Ava ila ble: For compound 1
tables of crystal data and refinement parameters, fractional
atomic coordinates, bond distances and angles, anisotropic
thermal parameters, and H-atom coordinates (9 pages). Or-
dering information is given on any current masthead page.
(42) Krueger, S. T.; Owens, B. E.; Poli, R. Inorg. Chem. 1990, 29,
2001-2006.
(43) Abugideiri, F.; Keogh, D. W.; Kraatz, H.-B.; Poli, R.; Pearson,
W. J . Organomet. Chem. 1995, 488, 29-38.
OM960384V