Niobium-mercury heterometallic compounds as sources of niobium(II): Radical paths to organoniobium species

Article abstract of DOI:10.1021/om970779u

The niobium-mercury compounds [Cp′₂Nb(L)]₂Hg (Cp′ = η⁵-C₅H₄SiMe₃, L = CO (4), PMe₃ (5), or CN^tBu (6)) serve as stable precursors to the short-lived Nb(II) radicals of general formula [Cp′₂Nb-L]. The homolysis of the Nb-Hg bond may result from a slow thermal reaction (generating low concentrations of radicals) or from a photochemical process in which mercury extrusion is more rapid. Since the Nb(II) species show no evidence for Nb-Nb bond formation, they are useful in the synthesis of a variety of Nb(III) and Nb(IV) species. Reactions of 4 with dimeric species such as [CpFe(CO)₂]₂, [CpNi(CO)]₂, Co₂(CO)₈, or RSe-SeR give rise to new Nb-M or Nb-Se compounds, while reactions with potential π donors such as formaldehyde or azobenzene lead to displacement of the ligand L and formation of the Nb(IV) complexes. Crystallographic and variable-temperature NMR studies on the Nb-Fe compound Cp′₂Nb(μ-CO)₂Fe(CO)Cp (10) are consistent with a low-energy fluxional process involving exchange of bridging and terminal carbonyl ligands.

Full text of DOI:10.1021/om970779u

5884
Organometallics 1997, 16, 5884-5892
Niobiu m -Mer cu r y Heter om eta llic Com p ou n d s a s
Sou r ces of Niobiu m (II): Ra d ica l P a th s to
Or ga n on iobiu m Sp ecies
B. Thiyagarajan,1a Lucyna Michalczyk,1a Victor G. Young, J r.,1b and
J oseph W. Bruno*,1a
Department of Chemistry, Wesleyan University, Middletown, Connecticut 06459, and X-ray
Crystallographic Laboratory, University of Minnesota, Minneapolis, Minnesota 55455
Received September 3, 1997X
The niobium-mercury compounds [Cp′2Nb(L)]2Hg (Cp′ ) η5-C5H4SiMe3, L ) CO (4), PMe3
(5), or CNtBu (6)) serve as stable precursors to the short-lived Nb(II) radicals of general
formula [Cp′2Nb-L]. The homolysis of the Nb-Hg bond may result from a slow thermal
reaction (generating low concentrations of radicals) or from a photochemical process in which
mercury extrusion is more rapid. Since the Nb(II) species show no evidence for Nb-Nb
bond formation, they are useful in the synthesis of a variety of Nb(III) and Nb(IV) species.
Reactions of 4 with dimeric species such as [CpFe(CO)2]2, [CpNi(CO)]2, Co2(CO)8, or RSe-
SeR give rise to new Nb-M or Nb-Se compounds, while reactions with potential π donors
such as formaldehyde or azobenzene lead to displacement of the ligand L and formation of
the Nb(IV) complexes. Crystallographic and variable-temperature NMR studies on the Nb-
Fe compound Cp′2Nb(µ-CO)2Fe(CO)Cp (10) are consistent with a low-energy fluxional process
involving exchange of bridging and terminal carbonyl ligands.
In tr od u ction
a radical center to function as chain carrier. Many 17-
electron species are also activated toward ligand sub-
stitution reactions, and the disproportionation mecha-
nism (eq 1) illustrates the fact that associative pathways
are available; the kinetic enhancement (relative to
similar processes for diamagnetic analogues) is due to
the availability of the 19-electron intermediates and
associative reaction pathways that are not accessible to
saturated diamagnetic compounds.3c-f For these rea-
sons, the production of radical species is of considerable
interest and a common synthetic approach involves
photolytic homolysis of metal-metal bonds.5 Clearly,
however, the recombination of the resulting radicals
often represents an undesirable outcome, one which
may preclude the use of the radicals in synthetic
applications. Indeed, a wide range of radical combina-
tion reactions have been the subject of quantitative
study, and they exhibit second-order rate constants in
It is widely recognized that transition-metal-centered
radicals afford ample opportunity for the development
of new chemistry of interest in synthesis and catalysis.2
The majority of these radicals exhibit a 17-electron
configuration and are known to be susceptible to one-
electron processes, two-electron processes, or both.
These reactions include substitution or atom transfer
reactions yielding new metal-ligand bonds,3 as well as
reduction or oxidation reactions; oxidation usually
involves prior addition of a ligand to form a 19-electron
radical.4 As expected, the susceptibility of the odd-
electron species to both oxidation and reduction also
makes them useful in disproportionation reactions;
these may proceed as in eq 1 (LnM ) CpMo(CO)3), with
the range 107-109 M-1 s-1 5,6
One approach used to
.
circumvent simple coupling reactions involves the in-
corporation of large ligands into the coordination sphere;7
this serves to slow dimerization but can also diminish
the reactivity toward other substrates too. Hence, the
goal is to develop radical-based systems that are easily
accessible, exhibit substantial reactivity toward a va-
riety of substrates, and are resistant to dimerization
reactions. Clearly, these latter two criteria are often
addition of L′ to give the 19-electron reductant. The net
result is the production of two 18-electron species and
X Abstract published in Advance ACS Abstracts, December 1, 1997.
(1) (a) Wesleyan University. (b) University of Minnesota.
(2) (a) Organometallic Radical Processes; Trogler, W. C., Ed.;
Elsevier: Amsterdam, 1990. (b) Baird, M. C. Chem. Rev. 1988, 88,
1217-1227. (c) Sun, S.; Sweigart, D. A. Adv. Organomet. Chem. 1996,
40, 171-214.
(3) (a) Turaki, N. N.; Huggins, J . M. Organometallics 1986, 5, 1703-
1706. (b) Herrington, T. R.; Brown, T. L. J . Am. Chem. Soc. 1985, 107,
5700-5703. (c) Therien, M. J .; Ni, C.-L.; Anson, F. C.; Osteryoung, J .
G.; Trogler, W. C. J . Am. Chem. Soc. 1986, 108, 4037-4042. (d) Bagchi,
R. N.; Bond, A. M.; Heggie, C. L.; Henderson, T. L.; Mocelin, E.; Seikel,
R. A. Inorg. Chem. 1983, 22, 3007-3012. (e) Shi, Q.-Z.; Richmond, T.
G.; Trogler, W. C.; Basolo, F. J . Am. Chem. Soc. 1982, 104, 4032-
4034. (f) Kowaleski, R. M.; Basolo, F.; Trogler, W. C.; Ernst, R. D. J .
Am. Chem. Soc. 1986, 108, 6046-6048.
(4) (a) Tyler, D. R. Prog. Inorg. Chem. 1988, 36, 125-194. (b) Tyler,
D. R. Acc. Chem. Res. 1991, 24, 325-331. (c) Stiegman, A. E.; Stieglitz,
M.; Tyler, D. R. J . Am. Chem. Soc. 1983, 105, 6032-6037. (d) Goldman,
A. S.; Tyler, D. R. J . Am. Chem. Soc. 1984, 106, 4066-4067.
(5) (a) Scott, S. L.; Espenson, J . H.; Zhu, Z. J . Am. Chem. Soc. 1993,
115, 1789-1797. (b) Yao, Q.; Bakac, A.; Espenson, J . H. Organo-
metallics 1993, 12, 2010-2012. (c) Zhang, S.; Brown, T. L. J . Am.
Chem. Soc. 1993, 115, 1779-1789. (d) Meyer, T. J .; Caspar, J . V. Chem.
Rev. 1985, 85, 187-218.
(6) (a) Hughey, J . L., IV; Anderson, C. P.; Meyer, T. J . J . Organomet.
Chem. 1977, 125, C49-C52. (b) Hughey, J . L., IV; Bock, C. R.; Meyer,
T. J . J . Am. Chem. Soc. 1975, 97, 4440-4441. (c) Meckstroth, W. K.;
Walters, R. T.; Waltz, W. L.; Wojcicki, A.; Dorfman, L. M. J . Am. Chem.
Soc. 1982, 104, 1842-1846. (d) Wegman, R. W.; Olsen, R. J .; Gard, D.
R.; Faulkner, L. R.; Brown, T. L. J . Am. Chem. Soc. 1981, 103, 6089-
6092. (e) Walker, H. W.; Herrick, R. S.; Olsen, R. J .; Brown, T. L. Inorg.
Chem. 1984, 23, 3748-3752. (f) Laine, R. M.; Ford, P. C. Inorg. Chem.
1977, 16, 388-391.
S0276-7333(97)00779-6 CCC: $14.00 © 1997 American Chemical Society

Niobium-Mercury Heterometallic Compounds
Organometallics, Vol. 16, No. 26, 1997 5885
(CO)4], Co2(CO)8, and [CpNi(CO)]2 (Aldrich) were commercial
materials and used as supplied. Compounds 4-6 were
prepared using literature methods.8
at odds, and useful systems must exhibit a fine balance
between high reactivity and stability toward dimeriza-
tion.
NMR spectra were obtained on a Varian Gemini 300 FT-
NMR instrument, infrared spectra on a Perkin-Elmer model
1600 FT-IR spectrophotometer, ESR spectra on a Bruker
ESP300 spectrometer, and voltammetric data with an EG&G
VersaStat potentiostat interfaced to a 486 PC. The electrolytic
cell was evacuated on a Schlenk line and then loaded with a
solution of 0.5 M NBu4PF6 in THF. The reference electrode
was prepared by inserting a short (ca. 1/4 in.) length of 5 mm
porous vycor into a Teflon tube. The latter was filled with
0.1 M AgNO3/CH3CN solution, into which a 16 gauge silver
wire was immersed. The electrolytic solution was maintained
We have recently communicated our discovery of a
series of heterometallic niobium-mercury compounds
with a LnNb-Hg-NbLn structure pattern.8 These
compounds (4-6) were isolated from the reduction of
Nb(III)-halide precursors with sodium amalgam, and
were proposed to arise from the attack of intermediate
niobium(II) radicals on elemental mercury (eq 2, Cp′ )
η5-C5H4SiMe3). The method involves synthesis of d3
under
a nitrogen atmosphere by passing THF-saturated
nitrogen through the cell. Photochemical experiments were
carried out with a medium-pressure mercury lamp using Pyrex
glassware; under these conditions, the principal photolysis
wavelengths used are 313 and 365 nm.
Syn th esis of Cp ′2Nb(CO)(SeP h ) (9). [Cp′2Nb(CO)]2Hg (4,
0.100 g, 0.100 mmol) and Ph2Se2 (31 mg, 0.1 mmol) were
weighed into a gas inlet flask. Benzene (20 mL) was added,
and the resulting solution was stirred at ambient temperature
for ca. 15 min. NMR indicated that the reaction was quantita-
tive, and the solvent was removed in vacuo to yield green 9.
1H NMR (C6D6): 5.39 (br s, 2H, Cp′), 4.92 (br s, 2H, Cp′), 4.71
(br s, 4H, Cp′), 0.19 (s, 18H, SiMe). 77Se NMR (C6D6): -44.5.
IR(benzene): 1926 (s), 1251 (m), 1096 (w), 1022 (w), 839 (m),
812 cm-1 (s).
Syn th esis of [Cp ′2Nb(µ-CO)2F e(CO)Cp ] (10). A solution
of [Cp′2Nb(CO)]2Hg (4) (0.23 g, 0.255 mmol) and [FeCp(CO)2]2
(0.09 g, 0.255 mmol) in 25 mL of toluene was irradiated under
nitrogen for 45 min using a 450 W Hanovia medium-pressure
mercury lamp. The resulting purple solution was filtered to
remove mercury; concentration of this solution followed by
addition of excess hexanes led to the precipitation of a purple
solid, which was collected by filtration and dried in vacuo (0.18
g, 67%). 1H NMR (C6D6): 4.87, 4.21 (4H each, br s, Cp′-H),
4.38 (5H, s, Cp), 0.28 (18H, s, SiMe). 13C NMR (C6D6): 100.5,
99.1, 98.9 (Cp′), 90.98 (Cp), 0.00 (SiMe). IR (Nujol): 1909 (vs,
CO), 1698 (vs,CO), 1243 (s), 1168 (s), 902.7, 839.7 cm-1. Anal.
Calcd for C24H31FeNbO3Si2: C, 50.35; H, 5.45. Found: C,
50.50; H, 5.52. Compound 11 was prepared similarly from 4
and equimolar Co2(CO)8. 1H NMR (C6D6): 5.07 (br s, 4H, Cp′),
4.59 (br s, 4H, Cp′), 0.02 (s, 18H, SiMe). IR (benzene): 2029
(s), 1966 (s), 1949 (s), 1906 (m), 1251 (m), 838 cm-1 (m). Green
12 was prepared similarly from 4 and 1.5 equiv of [CpNi(µ-
CO)]2; the reaction was photolyzed until IR indicated con-
sumption of 4 (ca. 30-40 min), and 12 was isolated as a brown
solid by precipitation from cold hexane. 1H NMR (C6D6): 5.33
(s, 5H, Cp), 4.90 (br s, 4H, Cp′), 4.23 (br s, 4H, Cp′), 0.18 (s,
18H, SiMe). IR (benzene): 1745 (s), 1260 (m), 1164 (w), 1013
(m), and 838 cm-1 (s).
radicals, and if Nb-Hg bond formation is reversible, the
heterometallic Nb-Hg product has the potential to
serve as a convenient precursor for radical intermedi-
ates. Herein, we show that this is indeed the case and
that the radicals may be generated using either thermal
or photochemical stimulation; as such, the stable Nb-
Hg compounds 4-6 constitute a convenient means of
storing the Nb(II) radicals. These are susceptible to
both one-electron and two-electron processes; the former
generate niobium products with either (a) new nio-
bium-ligand or (b) new niobium-metal bonds, while
the latter provide access to ligand substitution reactions
and products that are best formulated as niobium(IV)
radicals. While the one-electron processes yield het-
erobimetallic compounds containing niobium and either
middle or late transition metals, there is no evidence
for dimerization of the niobium radicals.
Exp er im en ta l Section
Gen er a l Con sid er a tion s. All manipulations involving
metal complexes were carried out under an atmosphere of
nitrogen which was first passed through activated BTS
catalyst and molecular sieves. Standard Schlenk techniques
were used to handle solutions,9 and solids were transferred
in a Vacuum Atmospheres Corp. glovebox under purified
nitrogen. Elemental analysis (C, H) were performed by the
Oneida Research Services, Inc., New York. Solvents toluene,
hexane, and tetrahydrofuran (J . T. Baker) were distilled from
sodium benzophenone ketyl under nitrogen. NbCl5, [Cp2Fe2-
Syn th esis of Cp′2Nb(η2-P h NNP h ) (13). [Cp′2Nb(CO)]2Hg
(4, 0.100 g, 0.100 mmol) and Ph2N2 (36 mg, 0.2 mmol) were
weighed into a Pyrex tube with a gas inlet side arm. Benzene
(ca. 20 mL) was added, and the resulting solution was
irradiated (450 W mercury lamp) for ca. 45 min. NMR data
were used to verify the loss of diamagnetic starting materials,
and the solvent was removed in vacuo to yield green 13. ESR
(7) (a) McCullen, S. B.; Brown, T. L. J . Am. Chem. Soc. 1982, 107,
7496-7500. (b) Walker, H. W.; Rattinger, G. B.; Belford, R. L.; Brown,
T. L. Organometallics 1983, 2, 775-776. (c) Crocker, L. S.; Heinekey,
D. M.; Schulte, G. K. J . Am. Chem. Soc. 1989, 111, 405-406. (d) Cooley,
N. A.; MacConnachie, P. T. F.; Baird, M. C. Polyhedron 1993, 7, 1965-
1972. (e) Watkins, W. C.; Hensel, K.; Fortier, S.; Macartney, D. H.;
Baird, M. C.; McLain, S. J . Organometallics 1992, 11, 2418-2424. (f)
Fortier, S.; Baird, M. C.; Preston, K. F.; Morton, J . R.; Ziegler, T.;
J aeger, T. J .; Watkins, W. C.; MacNeil, J . H.; Watson, K. A.; Hensel,
K.; LePage, Y.; Charland, J .-P.; Williams, A. J . J . Am. Chem. Soc. 1991,
113, 542-551. (g) Kuksis, I.; Baird, M. C. Organometallics 1996, 15,
4755-4762. h) Fettinger, J . C.; Keogh, D. W.; Poli, R. J . Am. Chem.
Soc. 1996, 118, 3617-3625.
(C6H6): g ) 1.9888, a
) 10 G. IR (C6H6): 1588 (m), 1474
Nb
(m), 1248 (m), 902 (m), 839 (s), 803 (m), and 752 (s) cm-1
.
Cr ysta l Str u ctu r e Deter m in a tion of 10. Compound 10
was crystallized by slowly cooling a hexanes solution to 0 °C,
resulting in dark violet needles. A small, well-formed crystal
measuring 0.60 × 0.09 × 0.09 mm was attached to a glass
fiber and mounted on the Siemens SMART system for a data
collection at 173(2) K. An initial set of cell constants was
calculated from reflections harvested from three sets of 50
frames. These initial sets of frames were oriented such that
orthogonal wedges of reciprocal space were surveyed. This
produced orientation matrices determined from 510 reflections.
(8) Thiyagarajan, B.; Michalczyk, L. M.; Bollinger, J . C.; Bruno, J .
W. Organometallics 1996, 15, 2588-2590.
(9) Shriver, D. F.; Drezdzon, M. A. The Manipulation of Air-Sensitive
Compounds, 2nd ed.; Wiley-Interscience: New York, 1986; Chapter
1.

5886 Organometallics, Vol. 16, No. 26, 1997
Thiyagarajan et al.
Ta ble 1. Cr ysta l Da ta for 10
empirical formula
fw
C24H31FeNbO3Si2
572.43
cryst habit
cryst size
space group
temperature
a 12.0587(1) Å
b 12.0587 (1) Å
c 15.3204(3) Å
volume
violet-brown needles
0 .60 × 0.09 × 0.09 mm
P32
173(2) K
1929.31(4) Å3
Z (molecules/cell)
density (calcd)
abs coeff
3
1.478 mg/m3
1.125 mm-1
R indices
R
R1 ) 0.0292, wR2 ) 0.0709
90°
â
90°
120°
γ
F igu r e 1. Cyclic voltammograms for 1, 1 mM in THF with
1.0 M NBu4PF6. The working electrode is a platinum disk,
the counter electrode a platinum wire, and the reference
electrode is Ag/AgNO3.
Ta ble 2. Bon d Len gth s (Å) a n d An gles (d eg) for 10
Nb(1)-Fe(1)
Nb(1)-C(1)
Nb(1)-C(2)
Fe(1)-C(1)
Fe(1)-C(2)
C(1)-O(1)
C(2)-O(2)
Fe(1)-C(3)
C(3)-O(3)
2.8156(6)
2.183(4)
2.184(4)
1.969(4)
1.997(4)
1.190(5)
1.188(5)
1.748(4)
1.159(5)
to a description of the voltammetric experiments, the
results of which are depicted in Figure 1. This contains
the voltammogram obtained with 1 mM Cp′2Nb(Cl)(CO)
(1) in THF solution; use was made of a platinum disk
electrode and Ag/AgNO3 reference electrode, and the
potentials are related to that couple. It is clear that
there are two one-electron processes operating here. The
first of these is wholly irreversible at scan rates up to
500 mV s-1, and this remains true even if the scan is
reversed immediately after this first wave. This is
commonly observed for this class of compounds,11 and
we attribute the irreversible nature of the reduction to
the rapid loss of chloride. The second wave is quasi-
reversible and is presumably due to conversion of the
Nb(II) radical to the Nb(I) anion. The reduction of the
17-electron Nb(II) radical would give rise to an 18-
electron anion and should not lead to loss of any ligand.
Moreover, we see no evidence for production of this
anion in the synthetic studies using sodium amalgam,
which does not have sufficient reducing power to effect
this second redox process. As a result, we have sug-
gested that the production of the heterometallic com-
pounds 4-6 is due to attack of the Nb(II) radicals on
elemental mercury present in the amalgam.8 As a
means of understanding the chemistry involved in the
second reduction process (Figure 1), we carried out
reductions using sodium-naphthalene. This is a more
powerful reductant than is sodium amalgam, and the
use of sodium-naphthalene with compounds 1-3 con-
verts them to the corresponding hydrides Cp′2Nb(H)-
(L). The source of the hydride ligand is not clear, but
it appears that reagents powerful enough to carry out
two successive reductions with 1-3 convert them to a
reactive (highly basic) anion that leads to hydride. This
is presumably the reason the second reduction is not
fully reversible.
Nb(1)-C(1)-O(1)
Nb(1) C(2)-O(2)
Fe(1)-C(1)-O(1)
Fe(1)-C(2)-O(2)
C(1)-Nb(1)-Fe(1)
C(1)-Fe(1)-Nb(1)
C(2)-Nb(1)-Fe(1)
C(2)-Fe(1)-Nb(1)
Nb(1)-C(1)-Fe(1)
Nb(1) -C(2)-Fe(1)
144.1(3)
144.4(3)
130.6(3)
131.0(3)
44.17(9)
50.59(11)
44.92(11)
50.53(11)
85.24(14)
84.5(2)
Final cell constants were calculated from a set of 8192 strong
reflections from the actual data collection. The space group
P32 was determined based on systematic absences and inten-
sity statistics. A successful direct-methods solution was
calculated, which provided most non-hydrogen atoms from the
E-map. All non-hydrogen atoms were refined with anisotropic
displacement parameters unless stated otherwise. All hydro-
gen atoms were placed in ideal positions and refined as riding
atoms with individual isotropic displacement parameters. Data
are collected in Tables 1 and 2.
Resu lts
Volta m m etr ic Stu d ies of Nb(III) P r ecu r sor s. For
niobocene compounds, Nb(III) is the lowest oxidation
state commonly observed. A number of compounds of
general formula Cp′2Nb(Cl)(L) (L ) two-electron donor
ligand) are available from the known [Cp′2Nb(µ-Cl)]2,
and these clearly contain an electron-rich Nb(III) center;
many are prone to facile aerobic oxidation.10 Nonethe-
less, we have used reduction reactions in a variety of
synthetic processes which may proceed through Nb(II)
intermediates and chose to undertake a synthetic and
voltammetric study of this process. This led to the
production of the niobium-mercury compounds shown
in eq 2, and the details of the syntheses have already
been published.8 As such, we will devote this section
Niobiu m -Mer cu r y Com p ou n d s As Sou r ces of
Nb(II) Ra d ica ls. The production of heterometallic
species 4-6 was postulated to result from attack of
niobium-centered radicals on elemental mercury.8 Of
greater utility would be the reverse process,12 for this
(10) (a) Martinez de Ilarduya, J . M.; Otero, A.; Royo, P. J . Organo-
met. Chem. 1988, 340, 187-193. (b) Antin˜olo, A.; Espinosa, P.; Fajardo,
M.; Gomez-Sal, P.; Lopez-Mardomingo, C.; Martin-Alonso, A.; Otero,
A. J . Chem. Soc., Dalton Trans. 1995, 1007-1013. (c) Martinez de
Llarduya, J . M.; Otero, A.; Royo, P. J . Organomet. Chem. 1988, 340,
187-193. (d) Fermin, M. C.; Hneihen, A. S.; Maas, J . J .; Bruno, J . W.
Organometallics 1993, 12, 1845-1856.
(11) (a) Chivers, T.; Ibrahim, E. D. Can. J . Chem. 1973, 51, 815-
820. (b) Gubin, S. P.; Smirnova, S. A. J . Organomet. Chem. 1969, 20,
229-240. (c) Mugnier, Y.; Moise, C.; Laviron, E. J . Organomet. Chem.
1981, 204, 61-66. (d) Kotz, J . C. In Topics in Organic Electrochemistry;
Fry, A. J ., Britton, W. E., Eds.; Plenum: New York, 1986; Chapter 3.
(12) Carlton, L.; Lindsell, W. E.; Preston, P. N. J . Chem. Soc., Dalton
Trans. 1982, 1483-1488.

Niobium-Mercury Heterometallic Compounds
Organometallics, Vol. 16, No. 26, 1997 5887
would indicate that these compounds can serve as
sources of the radical intermediates. Therefore, we
sought evidence for exchange reactions indicating the
intermediacy of the radicals. In the first of these, an
equimolar mixture of 4 and 5 was dissolved in benzene-
d6 and allowed to equilibrate at room temperature.
After ca. 10 h, NMR spectroscopy showed evidence for
a new species 7 with inequivalent niobium centers;8 this
was attributed to the exhange reaction depicted in eq
3, a process thought to proceed via the niobium(II)
species. To eliminate the (unlikely) possibility that this
photolysis (medium-pressure mercury lamp, ca. 30 min)
in benzene-d6 also led to the production of hydride with
no deuterium incorporation. None of these reactions is
of synthetic utility, since the hydrides are readily
available via other routes. However, they help establish
a useful base line for the studies to be noted below. In
the absence of a hydrogen atmosphere, compounds 4
and 5 are stable for 1-2 days in solution at ambient
temperature. Eventually, they do convert to the hy-
drides; hence, in the reactions with added reagents
discussed below, a negative result (i.e., failure to react
with the added reagent) would be indicated by the
production of hydride. Only in rare circumstances do
4 and 5 exhibit gross decomposition to intractable
products. It is important to note that we have seen no
evidence for niobium-niobium bond formation under
any circumstances.
1
Other species with metal-metal bonds have been
treated with alkyl halides in an effort to trap radical
intermediates.5a,13 Indeed, reaction of 4-6 with CDCl3
leads first to the precursors 1-3 and then to the known
compound Cp′2NbCl2.14 Although we have not proved
that this is a radical-induced reaction, it nonetheless
suggested studies on the behavior of the mercury
adducts with other potential radical precursors. Thus,
we treated compound 4 with Ph2Se2 to look for Nb-Se
bond formation; Ph2Se2 is an excellent trapping reagent
for organic radicals and reacts with primary alkyl
process involved a double ligand substitution (CO for
PMe3 exchange without Nb-Hg homolysis), we also
carried out the exchange reaction depicted in eq 4; note
that 4 contains Cp′ ligands while the product (8)
contains both Cp′ and Cp ligands. A ligand substitution
8
radicals with rate constants of ca. 107 M-1 s-1 15
Indeed,
.
here would involve pairwise exchange of cyclopentadi-
enyl ligands, which is far less likely than homolysis of
the metal-metal bond. Thus, the exchange processes
shown in eqs 3 and 4 provide convincing evidence
indicating that 4-6 serve as sources of the Cp′2NbII(L)
radical. Indeed, we observed a 10-line ESR signal
during the production of 4, indicating the presence of a
niobium-centered radical in solution;8 this signal dis-
appeared as diamagnetic 4 was produced. Also note-
worthy is the observation (above) that the radicals do
not undergo reduction with sodium amalgam; hence
either precursors 1-3 or products 4-6 may be stirred
over excess sodium amalgam for days without exhibiting
reduction past the Nb(II) oxidation state.
reaction of 4 with diphenyldiselenide (eq 5) gives rise
to the niobium-selenium compound 9 within 15 min
at room temperature; this reaction is quite clean and
gives no other niobium-containing products, but an
excess of the diselenide converts mercury to Hg-
1
(SePh)2.16 Compound 9 exhibits the expected H NMR
resonances for Cp′ and phenyl groups, a 77Se NMR
signal at -44.6 ppm, and an infrared stretch for the CO
ligand at 1926 cm-1
.
Compound 9 could also be
Ra d ica l Rea ction s a n d th e F or m a tion of Nb-X
Bon d s. With evidence for the accessibility of radical
intermediates, we sought to characterize their synthetic
potential. In the first such reaction, 4 or 5 was dissolved
in benzene and treated with 1 atm of H2. Over a period
of 1 day at ambient temperature, the compounds were
converted smoothly to the known hydrides Cp′2Nb(L)-
(H). This constitutes very mild conditions for the
homolytic activation of hydrogen by a metal center
which is undergoing only a one-electron oxidation; e.g.,
there is no evidence for oxidation to a niobium(IV)
species here. Similarly, if benzene solutions of these
same compounds (4 or 5) are heated to 100 °C in an
evacuated, sealed NMR tube, they also convert cleanly
into the hydrides in a few hours; it should be noted that
the use of benzene-d6 also leads to the niobium hydrides,
with little or no incorporation of deuterium (as evi-
denced by 2H NMR). The C-D bond in benzene is very
strong, so we also carried out these thermolyses in
toluene-d8 (or THF-d8). These reactions also led only
the niobium hydride, and deuterium NMR showed no
evidence for Nb-D; we have been unable to identify the
source of the hydride in these reactions. Finally,
prepared by treatment of 1 with Na+PhSe- in THF, but
the low solubility of this salt meant that this metathesis
reaction required days to go to completion. A related
reaction of 4 with dibenzyldiselenide gave a mixture of
at least two products, one of which was the known Cp′2-
Nb(CO)(CH2Ph).17 Diselenides are known to exhibit
weak Se-Se bonds with BDEs of ca. 44 kcal/mol, and
(13) (a) Huber, T.; Macartney, D. H.; Baird, M. C. Organometallics
1995, 14, 592-602. (b) MacConnachie, C. A.; Nelson, J . M.; Baird, M.
C. Organometallics 1992, 11, 2521-2528. (c) Yesaka, H.; Kobayashi,
T.; Yasufuku, K.; Nagakura, S. J . Am. Chem. Soc. 1983, 105, 6249-
6252. (d) Herrick, R. S.; Herrington, T. R.; Walker, H. W.; Brown, T.
L. Organometallics 1985, 4, 42-45. (e) Lee, K.-W.; Hanckel, J . M.;
Brown, T. L. J . Am. Chem. Soc. 1986, 108, 2266-2273. (f) Lee, K.-W.;
Brown, T. L. J . Am. Chem. Soc. 1987, 109, 3269-3275. (g) Hanckel, J .
M.; Lee, K.-W.; Rushman, P.; Brown, T. L. Inorg. Chem. 1986, 25,
1852-1856.
(14) Hitchcock, P. B.; Lappert, M. F.; Milne, C. R. C. J . Chem. Soc.,
Dalton Trans. 1981, 180-186.
(15) (a) Paulmier, C. Selenium Reagents and Intermediates in
Organic Synthesis; Pergamon: Oxford, 1986; Chapter 2. (b) Iwasawa,
N.; Funahashi, M.; Hayakawa, S.; Narasaka, K. Chem. Lett. 1993,
545-548. (c) Esker, J . L.; Newcomb, M. J . Org. Chem. 1993, 58, 4933-
4940.
(16) (a) Bochmann, M.; Webb, K. J . In Inorganic Syntheses; Cowley,
A. H., Ed.; J . Wiley & Sons: New York, 1997; Vol. 31, pp 24-28. (b)
Okamoto, Y.; Yano, T. J . Organomet. Chem. 1971, 29, 99-104.

5888 Organometallics, Vol. 16, No. 26, 1997
Thiyagarajan et al.
this makes them susceptible to radical reactions;15 this
is undoubtedly the reason for the formation of 9 in eq
5, since homolysis of the Se-Ph bond would not be
likely. However, the benzyl radical is considerably more
stable than is the phenyl radical,18 and homolysis of the
Se-benzyl bond could, and presumably does, result in
the production of the niobium-benzyl species.
P r ep a r a tion of New Heter om eta llics. In a recent
review, it was noted that there exist relatively few
niobium-containing heterobimetallic species.19 Clearly,
radical cross-coupling reactions represent a potentially
useful synthetic route to such species, albeit one that
is often limited by competition with self-coupling reac-
tions leading to homonuclear metal-metal dimers.
Since the niobium species under study are not prone to
this latter process, we sought evidence for their use in
the formation of heterobimetallic species. In one such
attempt, equimolar quantities of [CpFe(CO)2]2 (Fp2, a
potential source of Fp•) and 4 were reacted in benzene
at 25 °C. After a period of of ca. 20 h, 1H NMR
suggested the formation of the three products shown in
eq 6; the ratios vary with subtle details of the experi-
ment, but 10 is usually the major product. The mixture
As a probe of the generality of the chemistry in eq 6,
4 was allowed to react with other homonuclear dimers.
Equimolar quantities of 4 and Co2(CO)8 reacted readily
under similar photochemical conditions, within 25 min
giving rise to a new compound exhibiting infrared bands
at 2029, 1966, 1949, and 1906 cm-1 and NMR signals
at 5.07 and 4.59 ppm. There is no evidence for bridging
carbonyls, and we assign the 1906 cm-1 band to the Nb-
CO ligand. The structure proposed for the major
product is shown in eq 7. Unfortunately, the production
of this compound was always accompanied by the
formation of an additional product which exhibited
NMR signals 5.12 and 5.60 ppm; this was usually a
minor product but was always present. The infrared
signature of this minor product was largely masked by
the strong bands of the major product; since we were
unable to separate these products, we are unable to
identify the minor product; it is worth noting that an
earlier study of some similar Nb-Co heterometallics
show evidence of unusual spectral data and facile
isomerization.20 A related reaction with [CpNi(µ-CO)]2
proceeded as indicated in eq 8;21 here, photolysis for 45
min gave the green 12 as the only metal-containing
product. Compound 12 exhibited a single infrared band
contains the known mercury adduct Fp2Hg,12 a product
tentatively identified as a Nb-Hg-Fe species (this
assignment is made difficult by the fact that the mixture
is not easily separable), and new compound 10. While
this reaction indicated the reactivity of the Nb-Hg
compounds, it was of little synthetic value. However,
when the same reaction was run under photochemical
conditions using a medium-pressure mercury lamp,
several changes were noted; the reaction went to
completion much more quickly (45 min of photolysis),
elemental mercury was deposited in the bottom of the
vessel, and 10 was the sole product. Compound 10 was
obtained as violet needles, which were deposited from
a saturated hexane solution cooled to -20 °C. It
1
at 1745 cm-1 and H NMR resonances for the Ni-Cp
(5.33 ppm) and Nb-Cp′ ligands (4.90, 4.23 ppm); this
compound is expected to contain two mirror planes, so
the spectral data are fully consistent with the formula-
tion given in eq 8. The success of these photochemical
routes suggests that products 10-12 result because they
are stable to photolysis; this was confirmed for 10, which
only exhibited slow photochemical conversion to an
unknown compound over a period of several hours.
Cr ysta llogr a p h ic Stu d y of 10. The violet needles
of 10 were found to diffract satisfactorily, and the results
of an X-ray study are shown in Figure 2. Crystal data
and key bond lengths and angles are presented in
Tables 1 and 2. Compound 10 crystallized in the
trigonal space group P32 with three molecules per unit
cell. As indicated in eq 6, the molecule consists of
niobium and iron fragments connected with a Nb-Fe
bond and two bridging carbonyl ligands. The Nb-Fe
separation is 2.8156(6) Å; this is less than the sum of
the covalent radii (3.00 Å), and the presence of the
metal-metal bond presumably accounts for the dia-
magnetism seen for 10. The four-membered ring is
approximately planar, with torsion angles of 2.40(14)°
(C(1)-Nb(1)-C(2)-Fe(1)), 176.6(2)° (C(2)-Nb(1)-Fe(1)-
C(1)), and 2.7(2)° (Nb(1)-C(2)-Fe(1)-C(1)). The bridg-
exhibited infrared bands at 1907 and 1698 cm-1
,
indicating the presence of both bridging and terminal
1
carbonyl ligands. The H NMR spectrum contained a
resonance for the iron-bound Cp ligand and two reso-
nances for the aromatic hydrogens of the niobium-bound
Cp′ ligands; since each of these Cp′ ligands contains two
pairs of diastereotopic protons (one set R to the SiMe3
substituent and one set â), two sets of Cp′ signals would
seemingly require the presence of two mirror planes in
10. This is not the case, and the spectral data were
rationalized with X-ray crystallography and VT-NMR
studies; the results of these will be presented below.
(17) Antin˜olo, A.; de Llarduya, J . M.; Otero, A.; Royo, P.; Lanfredi,
A. M. M.; Tiripicchio, A. J . Chem. Soc., Dalton Trans. 1988, 2685-
2693.
(18) (a) McMillen, D. F.; Golden, D. M. Ann. Rev. Phys. Chem. 1982,
33, 493-532. (b) Bordwell, F. G.; Zhang, X.-M. Acc. Chem. Res. 1993,
26, 510-517.
(19) Chetcuti, M. J . In Comprehensive Organometallic Chemistry
II; Abel, E. W., Stone, F. G. A., Wilkinson, G., Eds.; Pergammon:
Oxford, 1995; Vol. 10, Chapter 2.
(20) Wong, K. S.; Scheidt, W. R.; Labinger, J . A. Inorg. Chem. 1979,
18, 136-140.
(21) For related studies, see: (a) Chetcuti, M. J .; Eigenbrot, C.;
Green, K. A. Organometallics 1987, 6, 2298-2306. (b) Davidson, J .
L.; Sharp, D. W. A. J . Chem. Soc., Dalton Trans. 1973, 1957-1960.

Niobium-Mercury Heterometallic Compounds
Organometallics, Vol. 16, No. 26, 1997 5889
F igu r e 3. Erying plot for the fluxionality in 10 in toluene-
d8.
Ta ble 3. Kin etic Da ta for F lu xion a lity in 10 (See
Text)
F igu r e 2. ORTEP drawing of the structure of 10 showing
the atom-numbering scheme.
temp (°C)
k (s-1
)
9
2
-18
-30
-40
-50
474
164
14.0
4.80
0.80
0.29
ing carbonyls exhibit Nb-C distances of 2.183(4) and
2.184(4) Å, Fe-C distances of 1.969(4) and 1.997(4) Å,
Nb-C-Fe angles of 85.23(14)° and 84.5(2)°, Nb-C-O
angles of 144.1(3) and 144.4(3)°, Fe-Nb-C angles of
44.17(9)° and 44.92(11)°, Nb-Fe-C angles of 50.59(11)°
and 50.53(1), and Fe-C-O angles of 130.6(3)° and
131.0(3)°. Thus, with the exception of the Fe-C dis-
tances and Fe-Nb-C angles, the two carbonyl ligands
exhibit identical (to within 3σ) metrical parameters; it
is also clear that the carbonyl ligands lean toward the
iron center and away from the niobium center. Crabtree
has reviewed structural data for bridging carbonyls22
and has suggested that bridging and semibridging
carbonyls may be identified by comparison of the angles
θ (M-C-O) and ψ (Μ-Μ-Ì). Within this framework,
symmetrically-bridging ligands show θ and ψ param-
eters near 138° and 49° (with ranges of ca. 136-144°
and 44-52°) and bent semibridging ligands show θ and
ψ ranges of 136-180° and 50-75°. As such, it is clear
that 10 (θ ) 144.2(2)°, ψ ) 44.5(4)°) does not exhibit
metrical parameters within the ranges normally associ-
ated with semibridging carbonyls, even though there is
a slight distortion in that direction; the two carbonyls
in 10 are best described as bridging, not semibridging.
gave rate constants at two temperatures, and 2D-EXSY
(exchange spectroscopy)23 was utilized to determine rate
constants at temperatures down to -50 °C; these rate
constants are listed in Table 3, and the corresponding
Eyring plot is given as Figure 3. It is clear that rate
constants obtained in these two ways (EXSY and
coalescence) fit the same line, and the resulting activa-
tion parameters for the fluxional process are ∆Hq
)
15.1(7) kcal/mol and ∆Sq ) 6.8(8) cal mol-1 deg-1. This
serves to reconcile the solution and solid-state data, and
it is clear that the infrared spectrum consists of a Fe-
CO (terminal) stretch and the antisymmetric stretch for
the bridging carbonyls.
Rea ction s w ith Tw o-Electr on Liga n d s. In the
processes described above, the niobium(II) radicals
reacted with other radical sources to produce species
most reasonably formulated as Nb(III) compounds.
However, niobocene derivatives containing Nb(IV) are
also well-known,24 so we attempted a few reactions with
potential two-electron ligands. In the first of these, a
benzene solution of 4 was treated with excess paraform-
aldehyde and the solution was irradiated for 30 min.
At the end of this time, the 1H NMR indicated the
absence of diamagnetic species and infrared showed no
evidence for the presence of the niobium-bound carbonyl
ligand. Indeed, an ESR study of this solution indicated
that the reaction had proceeded with formation of the
known niobium-formaldehyde adduct Cp′2Nb(η2-CH2O);
we have previously prepared this compound by reduc-
tion of the complex Cp′2Nb(Cl)(η2-CH2O), which loses
chloride upon reduction.25 Thus, the reaction proceeded
as shown in eq 9, and complexation of the formaldehyde
ligand resulted in loss of the carbonyl ligand. In a
Va r ia ble-Tem p er a tu r e NMR Stu d ies of 10. The
NMR spectral data for 10 were not consistent with the
solid-state structure, and comparison with the cobalt or
nickel compounds indicates that bridged and nonbridged
structures are accessible. We undertook a variable-
temperature NMR study of 10 with the aim of under-
standing the 25° NMR data and to uncover any fluxional
processes. At 25 °C, the niobium-Cp′ ligands exhibited
two slightly broadened resonances at ca. 4.82 and 4.21
ppm, and these became sharper at 35 °C. However, at
temperatures below ca. 0 °C these decoalesced into four
signals; this is consistent with the solid-state structure
of 10, since the static structure has one mirror plane
(not two) and would have inequivalent Cp′ ligands.
Fortunately, the two pairs of signals exhibited unequal
chemical shift differences and decoalesced at different
temperatures; the signal at 4.21 decoalesced at ca. 9 °C,
while the signal at 4.82 decoalesced at ca. 2 °C. This
(23) (a) Perrin, C. L.; Dwyer, T. J . Chem. Rev. 1990, 90, 935-967.
(b) Abel, E. W.; Coston, T. P. J .; Orrell, K. G.; Sia, V.; Stephenson, D.
J . Magn. Reson. 1986, 70, 34-53.
(24) Wigley, D. E.; Gray, S. D. in Comprehensive Organometallic
Chemistry II, Abel, E. W., Stone, F. G. A., Wilkinson, G., Eds.;
Pergamon Press: Oxford, 1995; Vol. 5.
(25) Thiyagarajan, B.; Michalczyk, L.; Bollinger, J . C.; Huffman, J .
C.; Bruno, J . W. Organometallics 1996, 15, 1989-1999.
(22) Crabtree, R. H.; Lavin, M. Inorg. Chem. 1986, 25, 805-812.

5890 Organometallics, Vol. 16, No. 26, 1997
Thiyagarajan et al.
metal bonding framework.27 Irradiation into these
absorption bands gives extrusion of elemental mercury
in both cases, leading to production of Co2(CO)8 in the
first case and [Os18Hg2C2(CO)42]2- in the second. By
comparison, the UV-vis spectrum of Nb-Hg compound
5 exhibits a shoulder at 514 nm (ꢀ 4390 L mol-1 cm-1).
This band is partially obscured by a strong LMCT band
centered at 350 nm, and we do not observe the MMCT
band for 4; we believe it is blue-shifted (relative to 5)
and is obscured by the LMCT band. The MMCT band
of 5 tails into the visible region of the spectrum and is
responsible for the orange color of the compound; the
strength of the transition is indicative of charge-transfer
character. It is interesting to note that the energy of
the transition for 5 is between those of [Os18Hg3C2-
(CO)42]2- and Hg[Co(CO)4]2. Thus, the transition energy
for 5 is more comparable to that of the anionic osmium
cluster, and it is clear that the niobium moieties are
considerably more electron-rich than are the Co(CO)4
moieties.
Mech a n istic Con sid er a tion s. A detailed, quanti-
tative understanding of the mechanism for the photo-
processes described herein is not yet available because
the Nb-Hg compounds exhibit broad UV-vis bands
that strongly overlap those exhibited by the other
dimeric species used. Hence, we are unable to eliminate
the possibility of simultaneous photoprocesses generat-
ing multiple reactive intermediates. We are seeking
other substrates which are suitable for detailed photo-
physical and mechanistic studies, and these will be
reported later. However, for the present, we may
compare the photochemical and thermal results to offer
general mechanistic hypotheses. In so doing, we will
focus on the Nb-Fe system for which the supporting
information (e.g., on the chemistry and photochemistry
of Fp2) is most readily available.
similar vein, compound 4 was also treated with azoben-
zene and irradiated for 45 min. This reaction gave a
green paramagnetic product 13 which exhibited the 10-
line ESR spectrum typical for niobium complexes (93Nb,
9
100%, I ) /2), with a g value of 1.9888 and a
) 10
Nb
G. The infrared spectrum again indicated that the
carbonyl ligand had been lost, and this reaction is
proposed to yield the azobenzene complex 13 shown in
eq 10. We were unable to crystallize this compound
(which oils out of all common solvents), and the Nb(V)
precursor Cp′2Nb(Cl)(η2-PhNNPh) is unknown. How-
ever, reaction of the dimeric Nb(III) reagent [Cp′2Nb-
(µ-Cl)]2 with azobenzene yielded a precipitate of
Cp′2NbCl2; we filtered off the insoluble Cp′2NbCl2 and
determined that 13 remained in solution. Hence, while
the azobenzene is unable to give symmetrical cleavage
of the niobium(III) dimer [Cp′2Nb(µ-Cl)]2, it effects
asymmetric cleavage to yield 13. This latter reaction
constitutes the preferred route to 13, but the reactions
shown in eqs 9 and 10 indicate that the Nb(II) radical
species are capable of releasing a carbonyl ligand to
accommodate an incoming π-bound ligand and formal
two-electron oxidation.
Discu ssion
Th er m a l a n d P h otoch em ica l Rou tes to Ra d ica ls.
The exchange reactions depicted in eqs 3 and 4 indicate
that the Nb-Hg compounds serve as sources of Nb(II)
radicals. The reactions with dimeric species such as Fp2
(eq 6) suggest that thermal and photochemical routes
can exhibit differing synthetic outcomes, but we believe
they are similar in proceeding via Nb(II) radicals. They
contrast, however, in two important ways. First, the
photochemical reaction with Fp2 requires less time to
go to completion than does the thermal one. Second,
the photochemistry yields only 10, while the thermal
reaction gives three products. This suggests that the
photochemistry serves to generate a higher concentra-
tion of Nb(II) radicals while also resulting in the loss of
elemental mercury. Indeed, others have noted the
utility of photochemical approaches in the elimination
of mercury from heterobimetallic species. In a study of
Hg[Co(CO)4]2, it was noted that this compound exhibited
an absorption band at 328 nm (ꢀ ca. 25 000 L mol-1
cm-1), and this was assigned to a transition involving
the linear Co-Hg-Co framework.26 The Hg-Co bonds
in this species are polarized and exhibit Hg+-Co-
character; thus, either a σ-σ* or dCo-σ* transition
would have a significant component of Co f Hg charge
transfer (MMCT) character. Similarly, the cluster anion
[Os18Hg3C2(CO)42]2- consists of two Os9C carbide clus-
ters separated by a triangular Hg3 moiety; this exhibited
The reaction of 4 and Fp2 involves two photosensitive
reagents, and irradiation could generate both Cp′2NbL
and Fp radicals.28 Hence, eqs 11-13 constitute a simple
mechanism for the production of 10, involving simul-
taneous generation of these radical species. It is
entirely possible that this sequence contributes to the
production of 10, but it cannot be the only mechanism
operating. First, we noted above that photolysis of 4
alone yields the hydride Cp′2Nb(H)(CO); thus, the low
steady-state concentration of Fp radical (which recom-
(27) Gade, L. H.; J ohnson, B. F. G.; Lewis, J .; McPartlin, M.; Kotch,
T.; Lees, A. J . J . Am. Chem. Soc. 1991, 113, 8698-8704.
(28) (a) Abrahamson, H. B.; Palazzotto, M. C.; Reichel, C. L.;
Wrighton, M. S. J . Am. Chem. Soc. 1979, 101, 4123-4127. (b) Vitale,
M.; Lee, K. K.; Hemann, C. F.; Hille, R.; Gustafson, T. L.; Bursten, B.
E. J . Am. Chem. Soc. 1995, 117, 2286-2296. (c) Dixon, A. J .; George,
M. W.; Hughes, C.; Poliakoff, M.; Turner, J . J . J . Am. Chem. Soc. 1992,
114, 1719-1722. (d) Moore, J . N.; Hansen, P. A.; Hochstrasser, R. M.
J . Am. Chem. Soc. 1989, 111,4563-4566. (e) Caspar, J . V.; Meyer, T.
J . J . Am. Chem. Soc. 1980, 102, 7794-7795. (f) DiMartino, S.; Sostero,
S.; Traverso, O.; Rehorek, D.; Kemp, T. J . Inorg. Chim. Acta 1990,
176, 107-112.
an absorption band at 540 nm (ꢀ 43 800 L mol-1 cm-1
)
which was also thought to be localized in the metal-
(26) Vogler, A.; Kunkely, H. J . Organomet. Chem. 1988, 355, 1-6.

Niobium-Mercury Heterometallic Compounds
Organometallics, Vol. 16, No. 26, 1997 5891
bines with k ) 4.5(2) × 109 M-1 s-1 27c
)
should lead to
mercury is retained in some of the products. Indeed,
the exchange processes depicted in eqs 3 and 4 never
show visible evidence for loss of elemental mercury; for
this reason, we depict the thermal Nb-Hg homolysis
as a stepwise process which leaves one such linkage
intact (eq 16). The Cp′2Nb-CO radical is proposed to
react with Fp2 as before to form 10 and Fp• via adduct
14 (eq 17). Subsequently, the Fp• radical couples with
the mercury-centered radical to generate the Nb-Hg-
Fp product (eq 18). Although omitted to save space,
inefficient coupling of the Nb and Fp radicals and the
production of substantial amounts of the niobium hy-
dride; it does not. In addition, the thermal route to 10
must proceed without the prior generation of Fp radical,
since this homolysis reaction has been shown to have a
half-life of ca. 30 days.29 Hence, we propose that the
Nb(II) radical serves as the reactive intermediate and
that it reacts with intact Fp2 as well as (perhaps) Fp
radical.
There are detailed studies on the reactions of radicals
with metal-metal dimers, and these have exhibited two
common characteristics; the reactions are electron-
transfer in nature, and the reductant is a 19-electron
species formed via ligand addition to the 17-electron
radical (eq 1). A similar process is likely in the
production of 10, but no ligand has been added to
convert Cp′2Nb-CO into a 19-electron reductant. The
reactions are typically carried out in benzene, which is
a poor ligand for the bent metallocene; related Cp′2NbL
fragments bind alkynes readily,30 but styrene31 and
formaldehyde25 adducts are the only known examples
of complexes of ligands containing sp2-hybridized carbon
donor atoms. Hence, the only potential ligand for the
niobium center is Fp2 itself, which is known to bind
Lewis acids at the oxygen of the µ-carbonyl ligand;
indeed, other heterometallic complexes of metal car-
bonyl clusters have been characterized crystallographi-
cally and show binding of the electrophilic metal cen-
ter(s) in this way.32 Rather than invoke an electron-
transfer reaction that generates a 16-electron niobium
center, we therefore propose the formation of the Fp2-
niobium adduct 14 as a precursor to reduction of the
Fp2 (eqs 14 and 15). In this way, an intramolecular
subsequent Nb-Hg homolyses involving this species
would lead to Fp-Hg-Fp. Thus, the thermal reaction
to 10 is proposed to involve the same mechanism as does
the photochemical process; the overall reaction differs
only in that it retains mercury in the Nb-Hg• radical,
and this is the source of the mercury-containing final
products.
The proposed mechanisms are consistent with all of
the synthetic and control reactions carried out to date,
they relate the thermal and photochemical results in a
simple way, and they are consistent with the large body
of evidence favoring 19-electron reductants in related
reactions.3,4 While we believe they constitute the major
routes to the observed products, we note again the
possiblity of secondary photochemical routes that con-
stitute alternate pathways to the observed products;
identification of these will require detailed kinetic and
photophysical studies and will be the subject of future
work.
Str u ctu r a l P r op er ties. We noted earlier that in
comparison to other transition metals, there are rela-
tively few niobium-containing heterometallics.18 None-
theless, these have exhibited a wide-ranging structural
diversity, and A,33 B,34 C,20,35 and D36 indicate the
structure types known for niobium-metal compounds.
Although only four bonding modes are indicated, there
charge-transfer process generates 10 and releases 1
equiv of Fp radical. This can then couple to make a
second equivalent of 10 or Fp2. This postulate is
preferable to direct attack of the niobium radical on an
iron center, as we are unaware of any precedent for a
radical attack (at iron) leading to homolysis of the Fp2
dimer.
(33) (a) Darensbourg, M. Y.; Silva, R.; Riebenspies, J .; Prout, C. K.
Organometallics 1989, 8, 315-323. (b) Woodward, S.; Curtis, M. D.;
Rheingold, A.; Haggerty, B. S. Organometallics 1992, 11, 2140-2146.
(c) Oudet, P.; Kubicki, M. M.; Moise, C. Organometallics 1994, 13,
4278-4284. (d) Oudet, P.; Perrey, D.; Bonnet, G.; Moise, C.; Kubicki,
M. M. Inorg. Chim. Acta 1995, 237, 79-86. (e) Oudet, P.; Moise, C.;
Kubicki, M. M. Inorg. Chim. Acta 1996, 247, 263-268. (f) Brunner,
H.; Kubicki, M. M.; Gehart, G.; Lehrl, E.; Lucas, D.; Meier, W.;
Mugnier, Y.; Nuber, B.; Stubenhofer, B.; Wachter, J . J . Organomet.
Chem. 1996, 510, 291-295. (g) Sato, M.; Yoshida, T. J . Organomet.
Chem. 1975, 94, 403-408.
A very similar mechanism can be invoked to explain
the thermal reaction, which differs in the sense that the
(29) Castellani, M. P.; Tyler, D. R. Organometallics 1989, 8, 2113-
2120.
(34) (a) Oltmanns, P.; Rehder, D. J . Organomet. Chem. 1988, 345,
87-96. (b) Herrmann, W. A.; Bierack, H.; Balbach, B.; Ziegler, M. L.
Chem. Ber. 1984, 117, 95-106. (c) Balbach, B.; Baral, S.; Biersack,
H.; Herrmann, W. A.; Labinger, J . A.; Scheidt, W. R.; Timmers, F. J .;
Ziegler, M. L. Organometallics 1988, 7, 325-331. (d) Wong, K. S.;
Scheidt, W. R.; Labinger, J . A. Inorg. Chem. 1979, 18, 136-140.
(35) (a) Skripkin, Yu.V.; Pasynskii, A. A.; Kalinnikov, V. T.; Porai-
Koshits, M. A.; Minacheva, L. Kh.; Antsyshkina, A. S.; Ostrikova, V.
N. J . Organomet. Chem. 1982, 231, 205-217. (b) Pasynskii, A. A.;
Anthyshkina, A. S.; Skripkin, Yu.V.; Kalinnikov, V. T.; Porai-Koshits,
M. A.; Ostrikova, V. N.; Sadikov, G. G. Russ. J . Inorg. Chem. 1981,
26, 1310-1314.
(30) (a) An˜tinolo, A.; Gomez-Sal, P.; de Llarduya, J . M.; Otero, A.;
Royo, P.; Carrerra, S. M.; Blanco, S. G. J . Chem. Soc., Dalton Trans.
1987, 975-980. (b) Antinolo, A.; Martinez-Ripoll, M.; Mugnier, Y.;
Otero, A.; Prashar, S.; Rodriguez, A. M. Organometallics 1996, 15,
3241-3243.
(31) Antinolo, A.; Carillo, F.; Garcia-Yuste, S.; Otero, A. Organo-
metallics 1994, 13, 2761-2766.
(32) (a) Hu, N.; Nie, G.; J in, Z.; Chen, W. J . Organomet. Chem. 1989,
377, 137-143. (b) de Boer, E. J . M.; de With, J .; Orpen, A. G. J . Chem.
Soc., Chem. Commun. 1985, 1666-1667. (c) Nelson, N. J .; Kime, N.;
Shriver, D. F. J . Am. Chem. Soc. 1969, 91, 5173-5174. (d) Alich, A.;
Nelson, N. J .; Strope, D.; Shriver, D. F. Inorg. Chem. 1972, 11, 2976-
2983. (e) Kristoff, J . S.; Shriver, D. F. Inorg. Chem. 1974, 13, 499-
506.
(36) Pasynskii, A. A.; Skripkin, Yu.V.; Eremenko, I. L.; Kalinnikov,
V. T.; Aleksandrov, G. G.; Andrianov, V. G.; Struchkov, Yu. T. J .
Organomet. Chem. 1979, 165, 49-56.

5892 Organometallics, Vol. 16, No. 26, 1997
Thiyagarajan et al.
is considerable variation within these groups (beyond
simply the identity of the ancillary ligands on the two
metal centers). For example, compounds of type A
contain two bridging ligands L, but these may be neutral
(CO),33e monoanions (SR, SeR, or PR2),33a-e,g or dianions
(S),33f and there may or may not be a niobium-metal
bond. Compounds C and D are characterized by the
presence of semibridging carbonyl ligands, and these
may be linear or bent. The existence of types A, C, and
D suggests that there may be relatively small barriers
to isomerizations involving carbonyl ligands. Although
this hypothesis has not been tested previously, com-
pound 10 proved to be amenable to such a study. The
resulting ∆Gq (13.1 kcal/mol) compares closely with that
observed for cis-trans isomerization in [CpFe(CO)2]2
(∆Gq 10.4 kcal/mol),37 a process believed to proceed by
way of bridge-terminal interconversion and twisting
about the metal-metal bond. In addition, our small
value for ∆Sq (6.8(8) cal mol-1 deg-1) is not consistent
with a heterolysis process yielding the ions [Cp′2Nb-
(CO)]+ and [CpFe(CO)2]-. Nonetheless, we tested for
this process by dissolving 10 in acetonitrile and record-
ing the IR spectrum; this polar solvent is known to
facilitate the formation of free ions,38 but we saw no
evidence for loss of the bridging carbonyl ligands in 10.
Therefore, we propose that 10 behaves as does [CpFe-
(CO)2]2 and equilibrates by way of the process depicted
in eq 19.
istry constitutes a facile route for the elimination of the
mercury, thus rendering the Nb-Hg compounds a
useful means of storing the radicals. Further, the
inaccessibility of Nb-Nb bond formation makes the
radicals highly reactive toward a variety of substrates.
Their reaction with metal-metal dimers is related to
the radical-induced disproportionation of these dimers,3
with one notable exception. In cases in which a dimer
LnM-MLn is reduced by the action of radical LnM•, the
latter first adds a ligand L′; the net result is production
of LnL′M+ and LnM-, an ion pair containing two 18-
electron species. In the chemistry reported herein, the
17-electron niobium-containing radical leads to hetero-
metallic species 10-12 in the absence of added ligand
and we have postulated the intermediacy of an adduct
(14) to facilitate the process. Inasmuch as the hetero-
metallic species 10-12 appear to have low barriers to
interconversion of bridging and terminal carbonyls, we
expect them to have a rich chemistry. Future work will
be devoted to a study of potential associative processes
involving the niobium(II) radicals and added ligands,
of reactions of the radicals with other metal centers
which do not contain carbonyls to support bridging
interactions, and on the chemistry of adducts such as
10-12 toward small molecules.
Con clu sion s. The niobium-mercury compound 4
shows clear evidence for the generation of electron-rich
niobium(II) radicals, and this has proved useful in the
preparation of a variety of new organoniobium com-
plexes. The radical species are available via thermal
and photochemical paths, and the subsequent chemistry
differs mainly in the fate of the mercury. Photochem-
Ack n ow led gm en t. We thank the National Science
Foundation for support of this work and Professor T.
David Westmoreland and Mr. Tim Shea for assistance
with the EXSY experiments.
Su p p or tin g In for m a tion Ava ila ble: Text giving a sum-
mary of the crystallographic analysis, tables of crystal data,
bond lengths and angles, atomic coordinates, and thermal
parameters, and an ORTEP diagram of 10 (16 pages). Order-
ing information is given on any current masthead page.
(37) Bullitt, J . G.; Cotton, F. A.; Marks, T. J . Inorg. Chem. 1972,
11, 671-676.
(38) (a) Horowitz, C. P.; Shriver, D. F. Adv. Organomet. Chem. 1984,
23, 219-305. (b) Darensbourg, M. Y. Prog. Inorg. Chem. 1985, 33, 221-
274.
OM970779U