Comparison of H-H versus Si-H σ-Bond Coordination and Activation on 16e Metal Fragments. Organosilane, N2, and Ethylene Addition to the Agostic Complex W(CO)3(PR3)2 and Dynamic NMR Behavior of the Latter

Article abstract of DOI:10.1021/ic960870a

Variable-temperature ³¹P{¹H} NMR spectroscopy of the agostic complexes M(CO)₃(PCy₃)₂ (M = Mo, W) indicates dynamic behavior as evidenced by collapse below -20°C of a singlet to an AB signal plus a shifted singlet. The inequivalency of the phosphines is possibly due to the presence of conformational isomers resulting from hindered rotation of the M-P bond or, less likely, a geometric isomer with pseudo-as PCy₃ ligands. Further studies on the coordination chemistry of W(CO)₃(PR₃)₂ (R = iPr, Cy) were performed. The bridging dinitrogen complex [W(CO)₃(PiPr₃)₂]₂(μ-N ₂) (1) was cleanly formed in the reaction of W(CO)₃(PiPr₃)₂ with N₂. Complex 1 was structurally characterized and compared with other bridging dinitrogen compounds of tungsten. The ethylene complex W(CO)₃(PCy₃)₂ (η²-C₂H₄ (2) was synthesized and characterized by X-ray crystallography in order to compare the binding mode of ethylene with that of H₂. Phenylsilane reacted with W(CO)₃(PR₃)₂ (R = iPr, Cy) to form the thermally unstable oxidative addition (OA) products WH(SiH₂Ph)(CO)₃(PR₃)₂ (3, R = Cy; 4, R = iPr). Diphenylsilane reacted with W(CO)₃(PiPr₃)₂ at 60°C to form the bridging silyl species [W(CO)₃(PiPr₃)(μ-SiHPh₂)]₂ (5), which was confirmed by spectroscopic techniques and X-ray crystallography to have two 3-center 2-electron W·H·Si interactions. Detailed comparisons of the binding and activation of silanes versus H₂ on various 16e metal centers suggest a high degree of similarity, but relative ease of OA depends on the electrophilicity of the metal-ligand fragment and other factors such as bond energetics. Increasing the electrophilicity of the metal center (e.g., adding positive charge) may aid in stabilizing alkane coordination.

Full text of DOI:10.1021/ic960870a

Inorg. Chem. 1997, 36, 3341-3353
3341
Comparison of H-H versus Si-H σ-Bond Coordination and Activation on 16e Metal
Fragments. Organosilane, N2, and Ethylene Addition to the Agostic Complex
W(CO)3(PR3)2 and Dynamic NMR Behavior of the Latter
Matthew D. Butts, Jeffrey C. Bryan, Xiao-Liang Luo, and Gregory J. Kubas*
Chemical Science and Technology Division, MS J514, Los Alamos National Laboratory,
Los Alamos, New Mexico 87545
ReceiVed July 23, 1996X
Variable-temperature 31P{1H} NMR spectroscopy of the agostic complexes M(CO)3(PCy3)2 (M ) Mo, W) indicates
dynamic behavior as evidenced by collapse below -20 °C of a singlet to an AB signal plus a shifted singlet. The
inequivalency of the phosphines is possibly due to the presence of conformational isomers resulting from hindered
rotation of the M-P bond or, less likely, a geometric isomer with pseudo-cis PCy3 ligands. Further studies on
the coordination chemistry of W(CO)3(PR3)2 (R ) iPr, Cy) were performed. The bridging dinitrogen complex
[W(CO)3(PiPr3)2]2(µ-N2) (1) was cleanly formed in the reaction of W(CO)3(PiPr3)2 with N2. Complex 1 was
structurally characterized and compared with other bridging dinitrogen compounds of tungsten. The ethylene
complex W(CO)3(PCy3)2(η2-C2H4) (2) was synthesized and characterized by X-ray crystallography in order to
compare the binding mode of ethylene with that of H2. Phenylsilane reacted with W(CO)3(PR3)2 (R ) iPr, Cy)
to form the thermally unstable oxidative addition (OA) products WH(SiH2Ph)(CO)3(PR3)2 (3, R ) Cy; 4, R )
iPr). Diphenylsilane reacted with W(CO)3(PiPr3)2 at 60 °C to form the bridging silyl species [W(CO)3(PiPr3)-
(µ-SiHPh2)]2 (5), which was confirmed by spectroscopic techniques and X-ray crystallography to have two 3-center
2-electron W‚‚‚H‚‚‚Si interactions. Detailed comparisons of the binding and activation of silanes versus H2 on
various 16e metal centers suggest a high degree of similarity, but relative ease of OA depends on the electrophilicity
of the metal-ligand fragment and other factors such as bond energetics. Increasing the electrophilicity of the
metal center (e.g., adding positive charge) may aid in stabilizing alkane coordination.
Introduction
The role of metal to ligand back-bonding in the coordination
chemistry of strong π-acceptor ligands such as CO, ethylene,
and dinitrogen has long been established but is only now being
defined in the activation of σ-bonded ligands such as H2 and
silanes. A massive amount of experimental and theoretical data
for the binding and oxidative addition of these and related
ligands on a wide array of transition metal fragments has been
accumulated within the past 10 years. Key to the development
of this field has been the creation of new unsaturated 5-coor-
dinate precursors (including Cp and Tp systems) which often
can be isolated and bind sixth ligands reversibly. In early studies
of group 6 complexes, we found that the addition of trialkyl-
phosphines, PR3, with large cone angles (R ) iPr, Cy) to
(C7H8)M(CO)3 (M ) Mo, W; C7H8 ) cycloheptatriene) under
an atmosphere of H2 resulted in the formation of M(CO)3(PR3)2-
(H2), the first characterized examples of η2-H2 coordination to
a metal center (eq 1).1 These reactions were found to proceed
The same is believed to be true in the molybdenum analogues,
which like W(CO)3(PR3)2 displayed reduced ν(CH) due to the
agostic C-H in the range 2540-2710 cm-1 3
.
These species can be likened to low-temperature-stable alkane
adducts of often studied photochemically-generated M(CO)5
fragments,5 but they are stabilized entropically because the
agostic interaction is intramolecular and also presumably
electronically by the electron-donating phosphines. In the latter
regard, σ-bond interactions with metals are anchored by three-
center bonding but are greatly reinforced by metal dπ to X-H
σ* back-bonding. Back-bonding is favored by increasing the
electron richness of the metal center by electron-donating
ligands. Most
(1) (a) Kubas, G. J. J. Chem. Soc., Chem. Commun. 1980, 61. (b) Kubas,
G. J.; Ryan, R. R.; Swanson, B. I.; Vergamini, P. J.; Wasserman, H.
J. J. Am. Chem. Soc. 1984, 106, 451. (c) Kubas, G. J.; Unkefer, C. J.;
Swanson, B. I.; Fukushima, E. J. Am. Chem. Soc. 1986, 108, 7000.
(d) Kubas, G. J. Acc. Chem. Res. 1988, 21, 120. Recent reviews: (e)
Jessup, P. G.; Morris, R. H. Coord. Chem. ReV. 1992, 121, 155. (f)
Heinekey, D. M.; Oldham, W. J., Jr. Chem. ReV. 1993, 93, 913. (g)
Crabtree, R. H. Angew. Chem., Int. Ed. Engl. 1993, 32, 789. (h) Morris,
R. H. Can. J. Chem. 1996, 74, 1907.
(2) Zhang, K.; Gonzalez, A. A.; Mukerjee, S. L.; Chou, S.-J.; Hoff, C.
D.; Kubat-Martin, K. A.; Barnhart, D.; Kubas, G. J. J. Am. Chem.
Soc. 1991, 113, 9170. The Cr complex was prepared from Cr(CO)3-
(naphthalene): Gonzalez, A. A.; Mukerjee, S. L.; Chou, S.-L.; Zhang,
K.; Hoff, C. D. J. Am. Chem. Soc. 1988, 110, 4419.
(3) Wasserman, H. J.; Kubas, G. J.; Ryan, R. R. J. Am. Chem. Soc. 1986,
108, 2294. For improved syntheses of M(CO)3(PCy3)2 and precursors,
see: Kubas, G. J. Organomet. Synth. 1986, 3, 254; Inorg. Synth. 1990,
27, 1.
through the intermediate M(CO)3(PR3)2 complexes.1a,2,3 X-ray
crystallographic studies performed on the Cr2 and W3 complexes
M(CO)3(PR3)2 (M ) Cr, R ) Cy; M ) W, R ) iPr, Cy)
indicated that these species were in essence 6-coordinate
complexes with one metal site occupied by an agostic1g,4 C-H
bond of a phosphine ligand.
(4) Brookhart, M.; Green, M. L. H.; Wong, L.-L. Prog. Inorg. Chem.
1988, 36, 1.
X Abstract published in AdVance ACS Abstracts, June 15, 1997.
S0020-1669(96)00870-1 CCC: $14.00 © 1997 American Chemical Society

3342 Inorganic Chemistry, Vol. 36, No. 15, 1997
Butts et al.
a
Table 1. Types of Coordination and Reactivity with W(CO)3(PR3)2
reactn with
no reacn with
SO2, CO
CO2, Xe
NH3, RNH2, py
CH3CN, DMF
PR3, P(OR)3
R3N, R2NH
alkanes, arenes
large PR3
irreversible binding
reversible binding
{
H2, N2, C2H4
propylene, C2F4
H2O,b ROH, R2CdO
tetrahydrofuran
R2S, thiophene
furan
R2O
{
{
H2 (equil), SiH3Ph
HX (X ) Cl, BF4)c
RSH,d H2Sd
oxidative addition
free radical formation
I2,d RSSRd
a This work and refs 2a and 3 unless noted. b Reference 18a.
c Reference 31. d Reference 55.
importantly, increased back-bonding weakens and eventually
cleaves the σ-bond, so there is a crucial but enigmatic fine
balance of σ donation and π back-donation that needs to be
defined in order to fully understand σ-bond activation. An
effective course to follow is to characterize a large series of
complexes, varying the electronic properties of both the metal-
ligand set and the substrate (sixth ligand). This approach will
elucidate the electronic structure requirements for the optimal
design of metal fragments for σ-bond activation, possibly even
leading to the elusive stabilization of metal-alkane complexes.
The agostic C-H‚‚‚M interactions in M(CO)3(PR3)2 (M )
Cr, Mo, W; R ) Cy, iPr) have been shown to be easily displaced
by a large variety of small molecules.1a,2,3,6 Four different types
of binding and/or reactivity have now been observed as
summarized in Table 1, including oxidative addition and in some
cases unanticipated formation of stable 17e free radicals,
W(CO)3(PR3)2(L) (L ) I, SR). In certain instances, the large
phosphines sterically inhibit ligand binding to a fine degree (e.g.,
NH2Me binds but NHMe2 does not), but for the most part
electronics determine type of reactivity. This paper reports
further investigation of the coordination chemistry of the
tungsten complexes with strong π-acceptor ligands such as
ethylene, dinitrogen, and silanes where back-bonding is a key
component of the interaction as for η2-H2. In contrast to H2,
organosilanes have been found to undergo complete oxidative
addition to W(CO)3(PR3)2. Comparison of the binding and ease
of oxidative addition of silanes versus H2 can now be made for
a large series of d6 group 6-8 metal complexes and is reported
here also.
Figure 1. Variable-temperature 31P{1H} NMR spectra (202.46 MHz)
of Mo(CO)3(PCy3)2 in toluene-d8 under helium. The weak singlet near
54 ppm was unidentified.
solution whereby only a single resonance was observed in the
respective 31P{1H} NMR spectra, and the agostic hydrogens
could not be resolved by 1H NMR spectroscopy at any accessible
temperature.3 We have now further investigated the solution
behavior of the W and Mo complexes M(CO)3(PCy3)2 by low-
1
temperature 31P{1H} and H NMR spectroscopy with the goal
of better understanding the nature of the C-H‚‚‚M interaction.
Results and Discussion
Temperature-Dependent Behavior of the Agostic Com-
plexes M(CO)3(PR3)2. Because the energy of the agostic
interaction in M(CO)3(PR3)2 is a crucial parameter in determi-
nation of binding energies of H2 and other ligands (eq 1),6
attempts have been made to obtain this value using NMR
methods. However, the interaction is so dynamic that a static
agostic structure could not be frozen out by NMR, although
unusual temperature-dependent behavior was nonetheless ob-
served. The low-temperature (-40 °C) 31P{1H} NMR spectrum
of Mo(CO)3(PCy3)2 in toluene-d8 contained two doublets of
equal intensity (AB pattern) at 61.0 and 56.8 ppm (JPP ) 77
Hz) and a singlet at 55.6 ppm (Figure 1). On gradual warming,
these resonances coalesced (reversibly) to a single broad peak
which sharpened on further warming to room temperature (57.2
1
ppm). The H NMR spectrum at low temperature contained
As an additional feature for study, the agostic complexes
M(CO)3(PR3)2 were early found to be highly fluxional in
only broad resonances between 1 and 2 ppm attributed to the
cyclohexyl groups. Similar behavior was observed for W(CO)3-
(PCy3)2 except that at low temperature (-60 °C) only one
doublet of the AB pattern was observed (59.9 ppm, J ) 81 Hz)
together with a broad singlet at 53.0 ppm (Figure 2). It is likely
that the other doublet is buried under the broad singlet. Once
again, the 1H NMR spectrum at -60 °C contained only
resonances for the cyclohexyl groups. The spectra of both
M(CO)3(PCy3)2 compounds as a function of temperature were
indistinguishable from those of M(CO)3[P(Cy-d11)3]2 (M ) Mo,
W) in toluene-d8.
Because of the similarities in the variable-temperature NMR
spectroscopic characteristics, it seems likely that the phenom-
enon that gives rise to the new 31P signals is common to both
the Mo and W complexes, M(CO)3(PCy3)2 and also to [Re-
(CO)3(PCy3)2]BAr4, reported by Heinekey and co-workers to
undergo analogous behavior.7 Proving the origin of the AB
(5) (a) Graham, M. A.; Perutz, R. N.; Poliakoff, M.; Turner, J. J. J.
Organomet. Chem. 1972, 34, C34. (b) Perutz, R. N.; Turner, J. J. J.
Am. Chem. Soc. 1975, 97, 4791. (c) Brown, C. E.; Ishikawa, Y.;
Hackett, P. A.; Rayner, D. M. J. Am. Chem. Soc. 1990, 112, 2530
and references therein. (d) Andrea, R. R.; Vuurman, M. A.; Stufkens,
D. J.; Oskam, A. Recl. TraV. Chim. Pays-Bas 1986, 105, 372. (e)
Church, S. P.; Grevels, F.-W.; Hermann, H.; Shaffner, K. J. Chem.
Soc., Chem. Commun. 1985, 30. (f) Upmacis, R. K.; Gadd, G. E.;
Poliakoff, M.; Simpson, M. B.; Turner, J. J.; Whyman, R.; Simpson,
A. F. J. Chem. Soc., Chem. Commun. 1985, 27. (g) Sweany, R. L. J.
Am. Chem. Soc. 1985, 107, 2374. (h) Upmacis, R. K.; Poliakoff, M.;
Turner, J. J. J. Am. Chem. Soc. 1986, 108, 3645. (i) Hall, C.; Perutz,
R. N. Chem. ReV. 1996, 96, 3125. (j) Walsh, E. F.; Popov, V. K.;
George, M. W.; Poliakoff, M. J. Phys. Chem. 1995, 99, 12016. (k)
Ishikawa, Y.; Hackett, P. A.; Rayner, D. M. J. Phys. Chem. 1989, 93,
652.
(6) (a) Zhang, K.; Gonzalez, A. A.; Hoff, C. D. J. Am. Chem. Soc. 1989,
111, 3627. (b) Gonzalez, A. A.; Zhang, K.; Nolan, S. P.; de la Vega,
R. L.; Mukerjee, S. L.; Hoff, C. D.; Kubas, G. J. Organometallics
1988, 7, 2429. (c) Gonzalez, A. A.; Zhang, K.; Mukerjee, S. L.; Hoff,
C. D.; Khalsa, G. R. K.; Kubas, G. J. ACS Symp. Ser. 1990, 428, 133.
(d) Kubas, G. J.; Nelson, J. E.; Bryan, J. C.; Eckert, J.; Wisniewski,
L.; Zilm, K. Inorg. Chem. 1994, 33, 2954.
(7) (a) Heinekey, D. M.; Schomber, B. M.; Radzewich, C. E. J. Am. Chem.
Soc. 1994, 116, 4515. (b) Heinekey, D. M.; Radzewich, C. E.; Voges,
M. H.; Schomber, B. M. J. Am. Chem. Soc. 1997, 119, 4172.

H-H versus Si-H σ-Bond Coordination
Inorganic Chemistry, Vol. 36, No. 15, 1997 3343
this is sterically allowed, one would then expect o-xylene to
form a similar complex, but with equivalent phosphine environ-
ments (eq 3). However, the same spectrum was observed for
Mo(CO)3(PCy3)2 in an o-xylene solvent mixture (hexane and
Et2O added to prevent freezing). Solvent binding (CH2Cl2) was
not seen for [Re(CO)3(PCy3)2]+ either.7 Importantly, none of
the three scenarios considered thus far explain the observation
of the singlet resonance found in the low-temperature 31P NMR
spectra of each complex (Mo ) 55.6 ppm, s, Figure 1; W )
53.0 ppm, br s, Figure 2; Re+ ) 24.0 ppm7, br).
A fourth possibility is that an equilibrium exists between the
agostic complex, in which the C-H‚‚‚M interaction is too
dynamic to be observed in the slow exchange regime (producing
the singlet peak), and an isomer which has inequivalent
phosphines coupled to each other (giving the AB pattern).
Square pyramidal (spyr) or trigonal bipyramidal (tbp) species
with cis-phosphines and no agostic interaction are candidates:
Figure 2. Variable-temperature 31P{1H} NMR spectra (202.46 MHz)
of W(CO)3(PCy3)2 in toluene-d8 under helium.
spectrum is difficult, but several possibilities can be eliminated,
including that it represents the static agostic complex, which
does contain inequivalent phosphines. This assignment is very
1
unlikely because the agostic C-H was not observed in the H
NMR spectrum for any of the complexes at any temperature.
Agostic hydrogens are typically shifted significantly upfield
(relative to the geminal nonagostic hydrogens) in cases where
static agostic complexes have been characterized by low-
temperature NMR.8 Most importantly, the coalescence tem-
perature, which is near -10 °C for both the W and Mo
complexes, is far too high for freezing out such a highly dynamic
weak interaction (binding enthalpy is estimated6b,c to be ca 10-
15 kcal/mol). Typically temperatures below -80 °C are
required to observe static agostic structures,4 and decoalescence
occurs at -90 °C for the aryl C-H‚‚‚Mo interaction in Mo-
(CO)(Bz2PC2H4PBz2)2.8j The lack of resonances in the hydride
region of the H NMR also argues against the presence of
phosphine cyclometalation structures, i.e. scission of the C-H
bond to form M-C and M-H bonds.
A third explanation for the AB pattern could be an equilibrium
solvent-bound species (eq 2). The toluene-d8 solvent would
Closely-related Re(CO)3(PCy3)2 was in fact found by crystal-
lography to have a distorted spyr structure with no agostic
interaction,9 although this is a 17e radical with chemically
equivalent trans phosphines. The 16e species TcCl(dppe)2 and
[MCl(R2PCH4PR2)2]+ (M ) Ru, Os) have distorted tbp struc-
tures,10 although here the coordinative unsaturation is stabilized
by π-donation from chloride11 and chelating phosphines and
no carbonyls are present. Low-spin d6 metal systems normally
display the most pronounced preference for the spyr structure,
with preference for π-accepting ligands in the basal position.12
Despite the large bulk of the PCy3 ligand and common belief,
cis-(PCy3)2MLn compounds have been characterized,13 including
cis-Mo(CO)4(PCy3)2.13e Thus cis structures for M(CO)3(PCy3)2
are possible, and the spyr geometry should be electronically
and sterically favored over the tbp form.
1
However, the above rationale is problematic also. The
coupling constants, JPP ) 77 and 81 Hz, were more consistent
with trans- than with cis-PR3, near those measured for
have to be bound asymmetrically with respect to the phosphines,
e.g. η2 along the P-M-P axis (η6- or η4-arene binding is
prohibited due to steric crowding). In the unlikely event that
(9) Crocker, L. S.; Heinekey, D. M.; Schultz, G. K. J. Am. Chem. Soc.
1989, 111, 405.
(10) (a) Burrell, A. K.; Bryan, J. C.; Kubas, G. J. J. Am. Chem. Soc. 1994,
116, 1575. (b) Mezzetti, A.; Del Zotto, A.; Rigo, P.; Pahor, N. B. J.
Chem. Soc. Dalton Trans. 1989, 1045. (c) Chin, B.; Lough, A. J.;
Morris, R. H.; Schweitzer, C.; D’Agostino, C. Inorg. Chem. 1994,
33, 6278. (d) Lough, A. J.; Morris, R. H.; Schlaf, M. Acta Crystallogr.,
Sect. C 1996, C52, 2193.
(11) Caulton, K. G. New J. Chem. 1994, 18, 25 and references therein.
(12) (a) Pearson, R. G. J. Am. Chem. Soc. 1969, 91, 4947. (b) Rossi, A.
R.; Hoffmann, R. Inorg. Chem. 1975, 14, 365. (c) Hoffman, P. R.;
Caulton, K. G. J. Am. Chem. Soc. 1975, 97, 4221. (d) Albright, T. A.;
Burdett, J. K.; Whangbo, M. H. Orbital Interactions in Chemistry;
John Wiley & Sons: New York, 1985; pp 313-326.
(13) (a) Grumbine, S. D.; Tilley, T. D.; Arnold, F. P.; Rheingold, A. L. J.
Am. Chem. Soc. 1993, 115, 7884. (b) Hampden-Smith, M. J.; Ruegger,
H. Magn. Reson. Chem. 1989, 27, 1107. (c) Azizian, H.; Dixon, K.
R.; Eaborn, C.; Pidcock, A.; Shuaib, N.; Vinaixia, J. J. Chem. Soc.,
Chem. Commun. 1981, 1020. (d) Butler, G.; Eaborn, C.; Pidcock, A.
J. Organomet. Chem. 1979, 181, 47. (e) Watson, M.; Woodward, S.;
Conole, G.; Kessler, M.; Sykara, G. Polyhedron 1994, 13, 2455.
(8) (a) Carr, N.; Mole, L.; Orpen, A. G.; Spencer, J. L. J. Chem. Soc.,
Dalton Trans. 1992, 2653. (b) Conroy-Lewis, F. M.; Mole, L.;
Redhouse, A. D.; Litster, S. A.; Spencer, J. L. J. Chem. Soc., Chem.
Commun. 1991, 1601. (c) Brookhart, M.; Lincoln, D. M.; Bennett,
M. A.; Pelling, S. J. Am. Chem. Soc. 1990, 112, 2691. (d) Ozawa, F.;
Park, J. W.; Mackenzie, P. B.; Schaefer, W. P.; Henling, L. M.; Grubbs,
R. H. J. Am. Chem. Soc. 1989, 111, 1319. (e) Park, J. W.; Mackenzie,
P. B.; Schaefer, W. P.; Grubbs, R. H. J. Am. Chem. Soc. 1986, 108,
6402. (f) Benn, R.; Holle, S.; Jolly, P. W.; Mynott, R.; Romao, C. C.
Angew. Chem., Int. Ed. Engl. 1986, 25, 555. (g) Cracknell, R. B.;
Orpen, A. G.; Spencer, J. L. J. Chem. Soc., Chem. Commun. 1986,
1005. (h) Cracknell, R. B.; Orpen, A. G.; Spencer, J. L. J. Chem.
Soc., Chem. Commun. 1984, 326. (i) Brookhart, M.; Green, M. L. H.;
Pardy, R. B. A. J. Chem. Soc., Chem. Commun. 1983, 691. (j) Luo,
X.-L.; Kubas, G. J.; Burns, C. J.; Eckert, J. Inorg. Chem. 1994, 33,
5219.

3344 Inorganic Chemistry, Vol. 36, No. 15, 1997
Butts et al.
trans-W(CO)4(PR3)(PR′3),14 although this may not be definitive
if a distorted cis geometry is present. The 31P{1H} NMR spectra
of the isostructural3 PiPr3 complex W(CO)3(PiPr3)2 showed
some sign of temperature-dependent behavior, but it could not
be resolved at any achievable low temperature, while [Re(CO)3-
(PiPr3)2]+ showed no spectral changes at all.7b This would not
be expected sterics should favor a spyr isomer for smaller cis-
PiPr3. Thus the increased bulkiness of the electronically-similar
PCy3 appears to be a critical factor (see below). Although the
CO resonances could not be resolved in the 13C{1H} NMR
spectrum of Mo(CO)3(PCy3)2 at -45 °C, the spectrum for [Re-
(CO)3(PCy3)2]+ showed no evidence for geometric isomers.7b
Freezing out 5-coordinate isomeric structures by NMR is
generally difficult in any event. Finally, the AB pattern has
also been seen in the 31P NMR of 6-coordinate octahedral
complexes by Caulton15 and Heinekey,7b which in this case
cannot be explained by a spyr isomer.
Heinekey and co-workers7b offer a rationale for the AB signal
in [Re(CO)3(PCy3)2]+, where sterics (bulky PCy3) are important.
It directly relates to the report of Caulton15 concerning hindered
rotation about the M-P bonds in 5- and 6-coordinate Ru and
Ir systems containing bulky asymmetric trans-P-t-Bu2Me phos-
phines. Steric interactions with the meridonal ligands are
proposed to lead to the existence of conformers with either no
symmetry or mirror symmetry that result from freezing out the
rotation about the M-P bond.15 Heinekey took this one step
further and attributed the AB pattern in his Re system to a
rotational conformer in which the bulky PCy3 ligands are still
trans yet inequivalent because of a staggered orientation of the
cyclohexyl groups. The singlet signal is then proposed to be
due to a eclipsed conformer in which the phosphines are related
by a mirror plane of symmetry.
Figure 3. ORTEP diagram of [W(CO)3(PiPr3)2]2(µ-N2) (1; 50%
probability ellipsoids).
two scenarios are in principle consistent with most of the
experimental results. An equilibrium between the highly
fluxional agostic complex with trans phosphines and a minor
spyr isomer containing inequivalent pseudo-cis phosphines is
conceivable, but the existence of conformers resulting from
hindered M-P rotation appears to be a better explanation.
Synthesis, Crystal Structure, and Reactivity of a Tungsten
Dimer with a Bridging Dinitrogen Ligand. As mentioned
above, the agostic C-H‚‚‚M interaction in the M(CO)3(PR3)2
complexes was easily displaced by a number of small molecules
(Table 1). Early in the study of H2 complexes, it was found1a,3
that the deep purple agostic complexes readily reacted with 1
atm of N2 to form yellow Mo and W dinitrogen19 complexes.
For M(CO)3(PCy3)2, the products of these reactions were
determined to be the terminal dinitrogen complexes M(CO)3-
(PCy3)2(N2), largely based on the observation of NN stretching
frequencies (Mo ) 2159 cm-1, W ) 2120 cm-1) in the infrared
spectra of the solids.1a,3 In contrast, the IR spectra of the PiPr3
congeners did not contain peaks that could be assigned to an
NN stretch, and the complexes were orange rather than yellow.
The Raman spectrum of the product of W(CO)3(PiPr3)2 and N2
was reported to contain bands at 1996 and 1939 cm-1 that
shifted to 1978 and 1896 cm-1 on 15N2 substitution. It was
concluded that this is consistent with a dinuclear species with
a bridging dinitrogen ligand wherein strong vibronic coupling
exists between the NN stretch and a carbonyl stretch of the same
symmetry.3 It is likely that the 1939 cm-1 mode is assignable
to ν(NN) because the calculated isotopic shift is 41 cm-1 (for
diatomic N2), and the shift to 1896 cm-1 would then be 43 cm-1
(metal coordination and coupling to CO would affect the shift).
The 1996 cm-1 band would thus be due to a CO stretch mixed
with ν(NN), shifting to 1978 cm-1 for the 15N2 isotopomer. The
proposed structure, shown in eq 4, was recently confirmed in a
Both of these conformers (and no intermediate rotomers) have
been seen in the X-ray crystal structures of these types of
complexes. The structures of the agostic complexes for Cr, W,
and Re are all staggered,2,3,7a and those for their 6-coordinate
adducts are generally eclipsed.1b,16-18 The W-ethylene com-
plex (see below) has a staggered conformation, possibly because
of steric pressure from the ethylene. The size of the meridonal
ligands is important in freezing out the conformers,15 and it was
also suggested that the agostic interaction is a factor in increasing
the barrier to rotation of the M-P bond in [Re(CO)3(PCy3)2]+
(and presumably in M(CO)3(PCy3)2 also).7b
In summary, in the rationalization of the temperature-
dependent behavior of the agostic complexes M(CO)3(PCy3)2,
(14) Schenk, W. A.; Buchner, W. Inorg. Chim. Acta 1983, 70, 189.
(15) Notheis, J. U.; Heyn, R. H.; Caulton, K. G. Inorg. Chim. Acta 1995,
229, 187.
(16) Heinekey, D. M.; Voges, M. H.; Barnhart, D. M. J. Am. Chem. Soc.
1996, 118, 10792.
(17) Unpublished X-ray data for M(CO)3(PCy3)2(H2) for M ) Mo and W
( G. J. Kubas, P. J. Vergamini, H. J. Wasserman, and R. R. Ryan)
and trans-W(CO)4(PCy3)2 (L. S. Van Der Sluys, J. C. Huffman, and
G. J. Kubas).
(18) (a) Kubas, G. J.; Burns, C. J.; Khalsa, G. R. K.; Van Der Sluys, L. S.;
Kiss, G.; Hoff, C. D. Organometallics 1992, 11, 3390. (b) Khalsa, G.
R. K.; Kubas, G. J.; Unkefer, C. J.; Van Der Sluys, L. S.; Kubat-
Martin, K. A. J. Am. Chem. Soc. 1990, 112, 3855.
(19) Reviews on N2 complexes: (a) Hidai, M.; Mizobe, Y. Chem. ReV.
1995, 95, 1115. (b) Pelika´n, P.; Boca, R. Coord. Chem. ReV. 1984,
55, 55. (c) Chatt, J.; Dilworth, J. R.; Richards, R. L. Chem. ReV. 1978,
78, 589. (d) Chatt, J.; Leigh, G. J. Chem. Soc. ReV. 1972, 1, 121.

H-H versus Si-H σ-Bond Coordination
Table 2. Summary of Crystallographic Data
Inorganic Chemistry, Vol. 36, No. 15, 1997 3345
1
2
5
empirical formula C49H92N2O6P4W2 C41H70O3P2W C48H64O6P2Si2W2
a, Å
b, Å
c, Å
R, deg
â, deg
γ, deg
V, Å3
Z
13.676(3)
13.833(2)
16.447(4)
70.80(2)
84.90(2)
71.64(2)
2788.4(10)
2
11.250(5)
12.716(6)
14.458(6)
87.01(4)
81.03(3)
74.04(3)
1964(2)
2
15.760(2)
17.427(2)
17.957(2)
90.390(8)
However, it would seem that the phosphine iPr groups would
be further away from each other in the W and Rh complexes if
the angle was in fact 90° as pictured in eq 4. The potential
surface for twisting about the M-NtN-M axis is likely to be
very soft however, and the torsional angle could be established
by optimal inter- as well as intramolecular packing of the iPr
groups, as noted previously.20 Steric repulsion between phos-
phines on neighboring metal centers is significantly greater in
the PCy3-substituted complex, and only the terminal complex
W(CO)3(PCy3)2(N2) was observed. The structure of 1 is similar
to that proposed for [W(CO)3(PiPr3)2]2(µ-pyrazine), the synthesis
and electrochemistry of which were recently reported.22
The N-N distance in [W(CO)3(PiPr3)2]2(µ-N2) (1) of 1.136(6)
Å represents only minimal lengthening of the N-N bond
distance found in gaseous N2 (1.0976 Å).23 Interestingly, the
N-N bond length is quite a bit shorter than that for other
structurally-characterized W complexes with a bridging N2
ligand, indicating less N2 activation. For example, the N-N
bond lengths in [Cl2W(C2Ph2)(dme)]2(µ -N2),24a [Cp*WMe3]2-
(µ-N2),24b [Cp*WMe2(C6F5O)]2(µ-N2),25 and [Cp*WMe2(Smes)]2-
(µ-N2)25 have been reported to be 1.292(16), 1.334(26), 1.26(2),
and 1.27(2) Å, respectively.26 Only in [Cl2W(NPh)(PMe3)2]2-
(µ-N2) was the N-N bond length (1.19(2) Å) comparable to
that found in 1.27 Interestingly, each of these compounds
contained tungsten centers in high oxidation states compared
to 1, in which the metal center is formally W(0). A more
electron rich tungsten center should result in a more activated
dinitrogen ligand due to a greater degree of WfN back-
bonding,19b,d although this is clearly not the case with 1. The
structures of the above high-oxidation-state compounds suggest
that they are most accurately described as tungsten hydrazido-
(4-) complexes with a significant contribution from the WtN-
NtW resonance structure.25 In contrast, our structural studies
would suggest that the most important resonance structure of 1
is W-NtN-W. This situation is also present in [Mo(CO)-
(depe)2]2(µ-N2), where the N-N bond length is 1.127(5) Å.21
The congeners with phenyl28 (dppe) or benzyl8j instead of ethyl
substituents and also the cationic Mn analogue29a [Mn(CO)-
(N2)(dppe)2]+ bind N2 terminally.
4931.8(10)
4
fw
1296.8
P1h (No. 2)
-100
0.710 73
1.54
42.8
0.030
0.079
856.8
P1h (No. 2)
-70
0.710 73
1.45
30.6
0.029
0.061
1222.8
C2/c (No. 15)
-80
0.710 73
1.65
48.2
0.061
0.166
space group
T, °C
λ, Å
F
calcd, g cm-3
µ, cm-1
Ra
Rw
b
a R ) ∑||Fo| - |Fc||/∑|Fo|, based on Fo2 > 2σ(Fo ). b Rw ) [∑[w(Fo
2
2
- Fc2)2]/∑w(Fo )2)]1/2
.
2
Table 3. Selected Bond Distances (Å) and Bond Angles (deg) for
[W(CO)3(PiPr3)2]2(µ-N2) (1)
W(1)-P(1)
W(1)-P(2)
W(1)-C(1)
W(1)-C(2)
W(1)-C(3)
W(1)-N(1)
N(1)-N(2)
2.5039(14)
2.5152(14)
2.019(5)
1.965(5)
2.027(5)
2.117(4)
1.136(6)
W(2)-N(2)
W(2)-P(3)
W(2)-P(4)
W(2)-C(4)
W(2)-C(5)
W(2)-C(6)
2.121(4)
2.5255(14)
2.5111(14)
2.018(5)
1.968(5)
2.007(5)
P(1)-W(1)-N(1)
P(2)-W(1)-N(1)
P(3)-W(2)-N(2)
P(4)-W(2)-N(2)
N(1)-W(1)-C(1)
N(1)-W(1)-C(2) 175.7(2)
N(1)-W(1)-C(3)
N(2)-W(2)-C(4)
93.70(11) N(2)-W(2)-C(5) 178.1(2)
88.0(2)
92.46(11) N(2)-W(2)-C(6)
92.39(10) W(1)-N(1)-N(2) 177.1(4)
93.84(10) W(2)-N(2)-N(1) 178.3(4)
92.1(2)
P(1)-W(1)-P(2)
P(3)-W(2)-P(4)
C(1)-W(1)-C(3)
C(4)-W(2)-C(6)
173.56(4)
173.39(4)
176.1(2)
179.4(2)
87.8(2)
92.0(2)
single-crystal X-ray crystallographic study performed on [W(CO)3-
(PiPr3)2]2(µ-N2) (1). An ORTEP diagram is shown in Figure 3,
(4)
and crystal and data collection parameters as well as selected
bond lengths and bond angles can be found in Tables 2 and 3,
respectively. Both tungsten centers of 1 exist in an octahedral
ligand environment with trans phosphines. When viewed down
the W-NtN-W axis, the corresponding ligands on each metal
are staggered with respect to each other (L-M-M-L torsional
angles are 54° on average). This staggering angle is similar to
those in [RhH(PiPr3)2]2(µ-N2)20 (55°) and [Mo(CO)(depe)2]2-
(µ-N2)21 (∼45° on average; depe ) Et2PC2H4PEt2). From
electronic arguments, this angle should be around 90° so that
orthogonal π systems can be used by each metal for MfN back-
bonding.19b,d Offsetting steric effects usually come into play
in bulky phosphine systems, and 45° staggering clearly reduces
repulsion between the eight sets of phosphine organo groups in
the Mo complex.
It is difficult to sort out whether formation of terminal versus
bridged solid-state structures is a consequence of more elec-
(22) Bruns, W.; Kaim, W.; Waldho¨r, E.; Krejcik, M. Inorg. Chem. 1995,
34, 663.
(23) Wilkinson, P. G.; Houk, N. B. J. Chem. Phys. 1956, 24, 528.
(24) (a) Churchill, M. R.; Li, Y.-J.; Theopold, K. H.; Schrock, R. R. Inorg.
Chem. 1984, 23, 4472. (b) Churchill, M. R.; Li, Y.-J. J. Organomet.
Chem. 1986, 301, 49.
(20) Yoshida, T.; Okano, T.; Thorn, D. L.; Tulip, T. H.; Otsuka, S.; Ibers,
J. A. J. Organomet. Chem. 1979, 181, 183.
(21) Luo, X. L.; Kubas, G. J.; Burns, C. J.; Butcher, R. J.; Bryan, J. C.
Inorg. Chem. 1995, 34, 6538.
(25) Regan, M. B.; Liu, A. H.; Finch, W. C.; Schrock, R. R.; Davis, W.
M. J. Am. Chem. Soc. 1990, 112, 4331.

3346 Inorganic Chemistry, Vol. 36, No. 15, 1997
Butts et al.
trophilic metal centers, steric congestion, stoichiometry, or
differential solubilities. Indeed the terminal N2 form30 of Mo-
(CO)(N2)(depe)2 actually can be isolated in the presence of
excess N2 whereas the µ-N2-bridged complex was generated in
a deficiency of nitrogen. Complex 1 precipitated even with
excess N2 present and was not very soluble. In solution, 15N
NMR on 15N-enriched complexes is necessary to distinguish
between terminal and bridging forms for these species. The
Mn complex was found to contain only terminal N2 in
solution.29a
Consistent with the observation of a short N-N distance, the
µ-N2 ligand in 1 as well as that in the above Mo complex was
somewhat labile. Although toluene solutions required heating
to 100 °C under partial vacuum (or reflux in xylene under argon
for the Mo system) to remove N2 and regenerate the agostic
complex, reactions in general proceeded through N2 ligand loss.
For example, the N2 ligand of 1 was displaced at room
temperature by a small excess of acetonitrile or under 1 atm of
ethylene to form the 6-coordinate complexes W(CO)3(PiPr3)2-
(L) (L ) MeCN, C2H4). Complex 1 was stable, however, in
the presence of 12 equiv of THF in benzene. As a measure of
ligand strengths in these systems, the order of enthalpy of
binding toward W(CO)3(PCy3)2 has been determined to be THF
< H2 < N2 < MeCN (∼2-3 kcal/mol differentials).6b The
terminal N2 ligand of W(CO)3(PCy3)2(N2) was much more labile
than the bridging N2 in 1.3,6a-c A yellow toluene solution of
the former quickly became purple when subjected to a vacuum
at room temperature, indicating loss of N2 and formation of the
agostic complex. The terminal N2 ligand was also labile in the
solid state under vacuum. Similar observations were made with
the ethylene complex W(CO)3(PCy3)2(η2-C2H4) (see below) and
the previously reported dihydrogen compound W(CO)3(PCy3)2-
(η2-H2).1
(PCy3)2(Ν2), including color (yellow), solubility (low), and
reversible binding of L. Toluene solutions of each complex
rapidly became purple under reduced pressure at room temper-
ature, indicating loss of L and re-formation of the agostic
complex, and discoloration occurred in the solid state. The
yellow color was immediately restored upon placing the
complexes back under ethylene, dihydrogen, or dinitrogen. The
properties of 2 in solution had not been previously studied,
however, and NMR characterization (1H, 31P, 13C) of 2 is
reported here (see Experimental Section). The NMR is
consistent with complete ethylene ligation on addition of 1 atm
of ethylene to W(CO)3(PCy3)2.
The metal-H2 σ-bond interaction in dihydrogen complexes
has been considered to be directly analogous to metal-alkene
π-bonding on the basis of theoretical calculations.32,33 The
structure of W(CO)3(PiPr3)2(η2-H2) showed that the side-on-
bonded H2 ligand is oriented parallel to the P-M-P axis rather
than the OC-M-CO axis presumably to maximize metal dπ
to H2 σ* back-bonding.1b-d An alkene such as ethylene would
be expected to maximize back-bonding to its π* orbital similarly
by aligning along the metal d orbitals of higher energy involved
with the electron-donating phosphines. We were thus specif-
ically interested in whether or not the ethylene ligand of 2 would
be oriented in the same fashion as H2. The structure was also
of interest because the steric constraints imposed by the bulky
tricyclohexylphosphine groups could prevent this electronically
favored orientation for the larger ethylene ligand. Propylene
in fact did not form a stable complex (Table 1), presumably
because steric factors would disallow this orientation (although
it could conceivably fit along the OC-M-CO axis). Similarly,
CO2 did not bind to W(CO)3(PCy3)2, although it is not clear
whether the reasons for this are electronic, steric, or both. In
the X-ray structure of Mo(CO)3(PiPr3)2(SO2), the SO2 ligand
is η1-S-bound with a planar MSO2 unit perpendicular to the
P-Mo-P plane in order to align the 2b1 acceptor orbital of
SO2 for π back-bonding.34 It was therefore thought that the
structure determination of an η2 π-ligand complex such as 2
would be revealing.
Many examples of NH3 or NH2NH2 formation from the
reaction of mono- or multinuclear N2 complexes with acids have
been reported.19a,c However, the reaction of 1 with 3.5 equiv
of HBF4‚OEt2 led to N2 loss and formation of W(CO)3(PiPr3)2-
(H)(BF4) as shown in eq 5. This 7-coordinate complex was
isolated as a solid by reacting W(CO)3(PiPr3)2 with HBF4‚OEt2
in toluene and was identified by comparison with the known
W(CO)3(PCy3)2(H)(BF4).31
X-ray Crystal Structure of W(CO)3(PCy3)2(η2-C2H4). An
ORTEP diagram of 2 is shown in Figure 4. Table 2 contains
crystal and data collection parameters, and selected bond lengths
and bond angles can be found in Table 4. The most significant
aspect of the structure of 2 is that it shows unambiguously that
ethylene is bound to W in the same orientation as H2 is bound
in W(CO)3(PiPr3)2(η2-H2). This is expected from electronic
considerations (more favorable back-bonding to π* and σ*
Synthesis and NMR of W(CO)3(PCy3)2(η2-C2H4), 2. It has
been known for some time that ethylene displaces the agostic
interaction in W(CO)3(PCy3)2 (eq 6, Table 1).1a,3 The properties
of solid 2 that precipitates from toluene solution in eq 6 were
very similar to those of W(CO)3(PCy3)2(η2-H2) and W(CO)3-
(26) The X-ray structure of [W(N2)2(PEt2Ph)3]2(µ-N2) has been reported;
however, the state of the structure refinement did not allow discussion
of bond distances: Anderson, S. N.; Richards, R. L.; Hughes, D. L.
J. Chem. Soc., Chem. Commun. 1982, 1291.
(27) Harlan, C. J.; Jones, R. A.; Koschmieder, S. U.; Nunn, C. M.
Polyhedron 1990, 9, 669.
(28) Sato, M.; Tatsumi, T.; Kodama, T.; Hidai, M.; Uchida, T.; Uchida,
Y. J. Am. Chem. Soc. 1978, 100, 4447.
(29) (a) King, W. A.; Luo, X.-L.; Scott, B. L.; Kubas, G. J.; Zilm, K. J.
Am. Chem. Soc. 1996, 118, 6782. (b) King, W. A.; Kubas, G. J. To
be published.
(30) Kubas, G. J.; Burns, C. J.; Eckert, J.; Johnson, S.; Larson, A. C.;
Vergamini, P. J.; Unkefer, C. J.; Khalsa, G. R. K.; Jackson, S. A.;
Eisenstein, O. J. Am. Chem. Soc. 1993, 115, 569.
(31) Van Der Sluys, L. S.; Kubat-Martin, K. A.; Kubas, G. J.; Caulton, K.
G. Inorg. Chem. 1991, 30, 306.
(32) (a) Saillard, J.-Y.; Hoffmann, R. J. Am. Chem. Soc. 1984, 106, 2006.
(b) Burdett, K. J.; Eisenstein, O.; Jackson, S. A. In Transition Metal
Hydrides; Dedieu, A., Ed.; VCH Publishers, Inc.: New York, 1992;
pp 149-184 and references therein. (c) Lin, Z.; Hall, M. B. Coord.
Chem. ReV. 1994, 135/136, 845.
(33) (a) Jean, Y.; Eisenstein, O.; Volatron, F.; Maouche, B.; Sefta, F. J.
Am. Chem. Soc. 1986, 108, 6587. (b) Hay, P. J. J. Am. Chem. Soc.
1987, 109, 705. (c) Eckert, J.; Kubas, G. J.; Hall, J. H.; Hay, P. J.;
Boyle, C. M. J. Am. Chem. Soc. 1990, 112, 2324. (d) Craw, J. S.;
Bacskay, G. B.; Hush, N. S. J. Am. Chem. Soc. 1994, 116, 5937. (e)
Dapprich, S.; Frenking, G. Organometallics 1996, 15, 4547.
(34) Kubas, G. J.; Jarvinen, G. D.; Ryan, R. R. J. Am. Chem. Soc. 1983,
105, 1883.

H-H versus Si-H σ-Bond Coordination
Inorganic Chemistry, Vol. 36, No. 15, 1997 3347
of structurally characterized W(CO)3(PCy3)2(L) compounds. For
example, an average W-P bond length of 2.51 Å was found
where L ) H2 or CO.17 In the agostic complex W(CO)3(PCy3)2,
the W-P distance of the phosphine without the C-H‚‚‚W
interaction was 2.494(1) Å.3 Thus, overall it appears that the
coordination sphere around the metal center in 2 is expanded
slightly to accommodate the sterically less favored ethylene
orientation. This was presumably counterbalanced by the gain
in back-bonding here as opposed to the scenario where the
ethylene lies off-axis or along the OC-W-CO axis.
Reactions of the Tungsten Agostic Complexes with Silanes.
Silanes, like H2, can bind to transition metal centers as σ-bond
ligands in an η2-fashion.1g,37 Because of the close relationship
between these ligands (η2-H2 and η2-HSiR3), we recently
became interested in the reactivity toward silanes38 of a range
of Mo and W agostic complexes that form stable η2-H2
compounds. While W(CO)3(PR3)2 (R ) iPr, Cy) reacted in
solution with H2 to form an equilibrium mixture of the classical
dihydride complex and the η2-H2 species, only the oxidative
addition products were observed when these W complexes were
treated with 1 equiv of PhSiH3 (eq 7). Whereas the resonances
Figure 4. ORTEP diagram of W(CO)3(PCy3)2(C2H4) (2; 50% prob-
ability ellipsoids).
Table 4. Selected Bond Distances (Å) and Bond Angles (deg) for
W(CO)3(PiPr3)2(η2-C2H4) (2)
W-P(1)
W-P(2)
W-C(1)
W-C(2)
W-C(3)
W-C(1e)
2.5524(13)
2.5579(14)
2.007(4)
2.012(4)
1.977(4)
2.338(4)
W-C(2e)
2.339(4)
1.378(6)
1.158(5)
1.157(5)
1.159(5)
C(1e)-C(2e)
C(1)-O(1)
C(2)-O(2)
C(3)-O(3)
P(1)-W-P(2)
P(1)-W-C(1)
P(1)-W-C(2)
P(1)-W-C(3)
C(1)-W-C(2)
C(1)-W-C(3)
165.26(3)
87.88(12)
93.35(11)
81.42(12)
174.2(2)
96.6(2)
C(1e)-W-C(2e)
C(1)-W-C(1e)
C(1)-W-C(2e)
C(3)-W-C(1e)
C(3)-W-C(2e)
34.3(2)
86.9(2)
86.3(2)
163.36(15)
161.83(15)
orbitals is available along the P-W-P axis) and reflects the
similarity of metal-dihydrogen bonding to metal-olefin bond-
ing. From a steric standpoint, it seemed somewhat surprising
that this orientation is adopted. The hydrogens on ethylene were
located and, despite their close proximity to the bulky phosphine
ligands, appear to be in typical positions on the carbons (average
CdC-H bond angle is 118°). Thus, electronic effects are
clearly dominant here, and on going to propylene or C2F4, steric
effects (e.g., fluorine lone-pair repulsion) apparently prevent
binding by disallowing orientation along P-W-P. It is
interesting that the larger olefins do not give stable complexes
(Table 1) when forced to orient along the OC-W-CO axis or
off-axis (both of which should be sterically allowed but very
unfavorable for back-bonding). This graphically illustrates just
how critical back-bonding is in stabilizing olefin coordination,
as is the case for H2, which can always orient to receive
maximum back-donation because of its small size.
The ethylene CdC distance of 1.378(6) Å is at the lower
limit of what is normally observed in ethylene complexes of
tungsten, which typically range from about 1.38 to 1.45 Å.35
This CdC distance in 2 represents only moderate elongation
compared to that in free ethylene (1.337(2) Å).36 Accordingly,
the W-C distances of 2.338(4) and 2.339(4) Å are substantially
longer than normally observed in W(η2-C2H4) complexes35
although distances in this range have been observed previously.35f
The average W-P distance of 2.555 Å is the longest in a series
1
of the Si-bound hydrogens of free PhSiH3 appeared in the H
NMR spectrum in C6D6 at 4.23 ppm (25 °C, JSiH ) 200.0 Hz),
the Si-H resonance of W(H)(CO)3(PCy3)2(SiH2Ph) (3) at -45
°C in toluene-d8 was found shifted downfield at 5.79 ppm, with
smaller SiH coupling (176.8 Hz) as is typically observed in
metal-bound silyl.1g The W-H signal appeared at -4.51 ppm
(t, JHP ) 14.9 Hz). The strongest evidence for the assignment
of 3 as a 7-coordinate hydrido-silyl species rather than a
6-coordinate η2-H-SiH2Ph species is the fact that the W-H
resonance in the 1H{31P} NMR spectrum contained no Si
satellites but did reveal JWH ) 28.2 Hz. The H and 31P{1H}
1
NMR resonances for 3 broadened on warming to 0 °C, which
is attributed to exchange across the equilibrium shown in eq 7.
Broadening on cooling below -45 °C is most likely due to the
slow exchange regime being approached for the intramolecular
ligand rearrangement in the 7-coordinate compound (still
unresolved at -80 °C). Similar observations were made for
the analogous PiPr3 compound 4 with the added feature that
small coupling was observed between the W-H and the silicon-
bound hydrogens (JHH ) 1.8 Hz). The spectroscopic changes
observed in these reaction mixtures on warming were ac-
companied by a reversible color change from bright yellow at
-45 °C to brown at room temperature. Both 3 and 4 decomposed
to several products on standing at room temperature in solution.
Unlike the tungsten complex, Cr(CO)3(PCy3)2 did not react with
silanes.
(35) (a) Radius, U.; Sundermeyer, J.; Pritzkow, H. Chem. Ber. 1994, 127,
1827. (b) Clark, G. R.; Nielson, A. J.; Rickard, C. E. F. Acta
Crystallogr. 1993, C49, 2074. (c) Aoshima, T.; Tamura, T.; Mizobe,
Y.; Hidai, M. J. Organomet. Chem. 1992, 435, 85. (d) Chacon, S. T.;
Chisholm, M. H.; Eisenstein, O.; Huffman, J. C. J. Am. Chem. Soc.
1992, 114, 8497. (e) Grevels, F.-W.; Jacke, J.; Betz, P.; Kru¨ger, C.;
Tsay, Y.-H. Organometallics 1989, 8, 293. (f) Alvarez, C.; Pacreau,
A.; Parlier, A.; Rudler, H. Organometallics 1987, 6, 1057. (g)
Carmona, E.; Galindo, A.; Poveda, M. L.; Rogers, R. D. Inorg. Chem.
1985, 24, 4033. (h) Sharp, P. R. Organometallics 1984, 3, 1217.
(36) Bartell, L. S.; Roth, E. A.; Hollowell, C. D.; Kuchitzu, K.; Young, J.
E. J. Chem. Phys. 1965, 42, 2683.
(37) (a) Schubert, U. AdV. Organomet. Chem. 1990, 30, 151. (b) Schubert,
U. In AdVances in Organosilicon Chemistry; Marciniec, B., Chojnows-
ki, J., Eds.; Gordon and Breach: Yverdon-lesBains, Switzerland, 1994.
(c) Lichtenberger, D. L.; Rai-Chaudhuri, A. J. Am. Chem. Soc. 1990,
112, 2492. (d) Corey, J. Y.; Braddock-Wilking, J. MGCN, Main Group
Chem. News 1996, 4, 6. (e) Zhang, S.; Dobson, G. R.; Brown, T. L.
J. Am. Chem. Soc. 1991, 113, 6908.
(38) (a) Luo, X.-L.; Kubas, G. J.; Burns, C. J.; Bryan, J. C.; Unkefer, C.
J. J. Am. Chem. Soc. 1995, 117, 1159. (b) Luo, X.-L.; Kubas, G. J.;
Bryan, J. C.; Burns, C. J.; Unkefer, C. J. J. Am. Chem. Soc. 1994,
116, 10312. (c) Luo, X.-L.; Kubas, G. J.; Burns, C. J. Unpublished
results.

3348 Inorganic Chemistry, Vol. 36, No. 15, 1997
Butts et al.
The low-temperature 1H, 1H{31P}, and 31P{1H} NMR spectra
of W(CO)3(PiPr3)2 in the presence of Ph2SiH2 in toluene-d8
indicated the presence of a mixture of compounds which could
not be resolved (inconsistent with the presence of either the
mononuclear oxidative-addition product or the η2-H-SiHPh2
product). The complex W(CO)3(PiPr3)2 did, however, react
cleanly with an excess of Ph2SiH2 at 60 °C in toluene to form
the bridging silyl complex 5 shown in eq 8, which was isolated
Figure 5. ORTEP diagram of [W(CO)3(PiPr3)(µ-SiHPh2)]2 (5; 50%
probability ellipsoids).
Table 5. Selected Bond Distances (Å) and Bond Angles (deg) for
[W(CO)3(PiPr3)(µ-SiHPh2)]2 (5)
W-Si
2.523(3)
2.708(3)
2.528(3)
2.008(9)
1.984(11)
2.029(10)
W-W′
3.2256(8)
1.887(11)
1.883(10)
1.158(12)
1.160(14)
1.119(13)
as an orange crystalline solid in 34% yield. This reaction most
likely began with the oxidative addition of Ph2SiH2 such as
observed with PhSiH3, although no intermediates were observed
and it is not clear at what point dimerization and H2 loss
occurred. The dimeric silane-bridged structure, confirmed by
X-ray crystallography, is part of a growing class of complexes
of this type containing bridging agostic Si-H interactions.39
The 1H NMR spectrum of 5 in THF-d8 contained no resonances
in the expected range for a terminal Si-H group but did contain
a multiplet at -7.26 ppm which collapsed to a singlet in the
1H{31P} NMR spectrum. Interestingly, this resonance in the
metal-hydride region had two sets of satellites of equal intensity
(JWH ) 42.4 and 36.7 Hz). A three-center W‚‚‚Si‚‚‚H interac-
tion is suggested on the basis of the 29Si NMR spectrum of 5 in
THF-d8, which contained only a broad doublet at 146.3 ppm
(ν1/2 ) 23 Hz) with JSiH ) 52 Hz. The latter coupling constant
is much lower than the Si-H coupling constant of 199.0 Hz
for free Ph2SiH2. The fact that the resonance collapsed to a
singlet in the 29Si{1H} NMR spectrum confirmed that the
coupling observed was indeed that of Si-H. This JSiH is typical
of J values for η2-HSiR3 compounds, which usually fall in the
range 35-70 Hz.37,38 The large upfield chemical shift of the
W-H resonance suggests that the Si‚‚‚H bond may have
significant tungsten-hydride character and possibly a “stretched”
Si‚‚‚H bond distance analogous to stretched H‚‚‚H distances
over 1 Å in H2 complexes.1g A well-defined species with such
a M-H‚‚‚Si bonding description was only recently reported for
a complex quite similar to 5, [Fe(CO)3(µ-SiHPh2)]2, which has
a very long Si‚‚‚H distance of 2.10 Å (50% longer than that
(1.43 Å) in the nonbonded Si-H in the Ph2SiH ligand).39f The
value of JSiH, 24.3 Hz, is the lowest reported for a M‚‚‚H‚‚‚Si
interaction and is much lower than that for 5. In light of this
and the absence of structural location of hydrogen (see below),
an unambiguous distinction between the M-H‚‚‚Si interpreta-
tion and the conventional M‚‚‚H-Si formalism cannot be made
for 5.
W-Si′
W-P
Si-C(30)
Si-C(40)
C(1)-O(1)
C(2)-O(2)
C(3)-O(3)
W-C(1)
W-C(2)
W-C(3)
P-W-Si
P-W-Si′
P-W-C(1)
P-W-C(2)
P-W-W′
Si-W-C(1)
Si-W-C(2)
146.47(9)
108.92(8)
89.0(3)
80.3(3)
157.63(6)
81.5(3)
Si-W-C(3)
Si-W-Si′
W-Si-W′
C(30)-Si-C(40)
C(1)-W-C(3)
C(2)-W-W′
95.4(3)
103.96(7)
76.04(7)
101.6(5)
175.6(4)
122.1(3)
69.2(3)
X-ray Crystal Structure of [W(CO)3(PiPr3)(µ-SiHPh2)]2,
5. Structural studies on related dimeric compounds containing
bridging SiHnR3-n (n ) 1, 2) ligands have suggested that, in a
bonding arrangement such as shown in eq 8, one should expect
a longer M-Si distance in the 3-center, 2-electron bond than
in the normal M-Si bond.39 For this reason and also in an
attempt to find an elongated Si-H bond distance, we performed
an X-ray crystallographic study on 5. An ORTEP diagram of
5 is shown in Figure 5. Table 2 contains crystal and data
collection parameters, and Table 5 gives selected bond lengths
and bond angles.
The crystal structure of 5 confirmed that it is a dimeric species
with bridging SiPh2 groups. From the view provided in Figure
6 it can be seen that the four-membered ring is roughly coplanar
with the P atom and the C(2)-O(2) carbonyl group. The W-W
distance was found to be 3.2256(8) Å, consistent with the
existence of a M-M bond as in the closely related complex
[W(CO)4(µ-SiHEt2)]2, where CO is present in place of PiPr3
and W-W ) 3.183(1) Å.39e Both tungsten centers in these
dinuclear complexes have highly distorted octahedral coordina-
tion about them. Unfortunately, the hydrogens were not located,
but as illustrated in Figure 7, the P atoms are pushed away from
the Si atoms “cis” to them (P-W-Si ) 108.9°), indicating
that the hydrogens presumably attached to the Si and W atoms
are situated between P and Si in the plane of the four-membered
ring. The strongest structural evidence that 5 contains two
3-center, 2-electron W‚‚‚H‚‚‚Si interactions is the observation
of two distinctly different W-Si bond distances (2.523(3) and
2.708(3) Å), with the longer being cis to the position in which
the hydride ligand is proposed to exist. Such bond length
differences are always observed in bridging silane compounds
in which there is a proposed W‚‚‚H‚‚‚Si interaction as depicted
in Figure 7. For example, [W(CO)4(µ-SiHEt2)]2 had W-Si
(39) (a) Suzuki, H.; Takao, T.; Tanaka, M.; Moro-oka, Y. J. Chem. Soc.,
Chem. Commun. 1992, 476. (b) Fryzuk, M. D.; Rosenberg, L.; Rettig,
S. J. Organometallics 1991, 10, 2537. (c) Aitken, C. T.; Harrod, J.
F.; Samuel, E. J. Am. Chem. Soc. 1986, 108, 4059. (d) Auburn, M.;
Ciriano, M.; Howard, J. A. K.; Murray, M.; Pugh, N. J.; Spencer, J.
L.; Stone, F. G. A.; Woodward, P. J. Chem. Soc., Dalton Trans. 1980,
659. (e) Bennett, M. J.; Simpson, K. A. J. Am. Chem. Soc. 1971, 93,
7156. (f) Simons, R. S.; Tessier, C. A. Organometallics 1996, 15,
2604. (g) Takao, T.; Yoshida, S.; Suzuki, H.; Tanaka, M. Organo-
metallics 1995, 14, 3855.

H-H versus Si-H σ-Bond Coordination
Inorganic Chemistry, Vol. 36, No. 15, 1997 3349
Figure 6. Stereoview of 5 as viewed down the W-W axis. The isopropyl groups and phenyl carbons other than the ipso carbons are omitted for
the sake of clarity.
in-molecules formalism.41 Complexes with H‚‚‚H separations
>1.3 Å were best viewed as dihydride complexes with weak
H‚‚‚H interactions, while a separation of 0.82 Å corresponds
to a true dihydrogen complex; i.e., a clear bond path connects
the hydrogens. The value in W(CO)3(PCy3)2(H2) is 0.82 Å by
neutron diffraction but is probably closer to 0.89 Å, the solid
state NMR value.42 A distance near 1.0 Å represented an
intermediate state of activation with some residual H-H
bonding, and a similarly stretched Si-H bond (2.10 Å, which
roughly corresponds to a 1.05 Å H-H distance) has now been
identified.39f
Back-bonding from M to the X-H σ* orbital has widely been
accepted to control the X-H distance (X ) H, C, Si, etc.) along
Figure 7. Drawing of 5 with the crystallographically determined bond
the reaction coordinate toward OA and confers stability to H2
complexes; e.g., poor back-bonding fragments such as Cr(CO)5
give thermally unstable H2 complexes.5 However, highly
electrophilic cationic centers such as [Re(CO)3(PCy3)2]+ and
[Mn(CO)(dppe)2]+ can bind H2 by increased σ donation to the
metal, offsetting decreased back-bonding.7,29 σ donation alone
cannot cause H-H bond rupture; back-bonding must complete
this task. Sorting out these two components is important in
understanding σ-bond activation. Recent theoretical studies by
Ziegler on our group 6 complexes43 and also group 8 systems44
indicate that the back-donation component, EBD, of the total
M-H2 orbital interaction energy can be quantified and separated
from σ-donation ED and is comparable to, or exceeds, the latter
eVen when H-H distances are short (0.85-0.9 Å).
lengths and angles in the plane of the four-membered ring. The lower
figure shows probable positions of hydrogens.
distances of 2.586(5) and 2.703(4) Å with the same overall
geometry as found for 5.39e Unfortunately, no spectroscopic
evidence for W‚‚‚H‚‚‚Si interaction was offered.
Comparisons of H-H versus Si-H σ-Bond Activation and
Oxidative-Addition (OA) Processes. Much current effort is
being focused on σ-bond activation,1g and it would be quite
informative to correlate well-studied H2 bonding and activation
with that for silanes, which could model alkane activation.
Agostic complexes such as M(CO)3(PR3)2 (M ) Mo, W) and
M(CO)(R2PC2H4PR2)2 (M ) Mo, Mn+)8j,21,28-30 generally react
with silanes similarly to H2, i.e. to form both OA products as
well as M‚‚‚H-Si σ-complexes. As yet, few metal fragments
bind both silanes and H2, especially near the point of OA. W-
(CO)3(PR3)2 cleaves silanes yet binds H2, which unexpectedly
is opposite to the reactivity on Mo(CO)(Et2PC2H4PEt2)2. The
possible reasons for this will be discussed here.
For H2 binding on a wide variety of metal-ligand fragments,
a near continuum of H-H distances (0.85-1.5 Å) exists (cf.
0.74 Å in H2).1e-h This blurs the dihydrogen-dihydride
formalism, i.e. the point at which the H-H bond can be
considered to be fully broken (1.6 Å is often quoted). Maseras
recently studied this theoretically by applying Bader’s40 atoms-
EBD values as high as 21 kcal/mol were calculated, and
(41) Maseras, F.; Lledos, A.; Costas, M.; Poblet, J. M. Organometallics
1996, 15, 2947.
(42) (a) Zilm, K. W.; Merrill, R. A.; Kummer, M. W.; Kubas, G. J. J. Am.
Chem. Soc. 1986, 108, 7837. (b) Zilm, K. W.; Millar, J. M. AdV. Magn.
Opt. Reson. 1990, 15, 163.
(43) Li, J.; Ziegler, T. Organometallics 1996, 15, 3844.
(44) Li, J.; Dickson, R. M.; Ziegler, T. J. Am. Chem. Soc. 1995, 117, 11482.
(40) (a) Bader, R. F. W. Atoms in Molecules: a Quantum Theory;
Clarendon: New York, 1990. (b) Bader, R. F. W. Acc. Chem. Res.
1985, 18, 9.

3350 Inorganic Chemistry, Vol. 36, No. 15, 1997
Butts et al.
Table 6. Axial CO Stretching Frequencies (cm-1) for Complexes
Containing H2 versus Other Ligands
At least a dozen such equilibria have been observed for H2
activation but only two for Si-H bond activation.38a,39g Re-
markably, ∆Gq for eq 9 (16.0 kcal/mol at 298 K)18b is very
similar to that for eq 10 (15.3 kcal/mol at 333 K),38c indicating
that H-H and Si-H cleavages have reasonably similar reaction
coordinates. However, there are subtle differences in H-H and
Si-H activations, as seen by the work here that shows silanes
undergo complete OA on W(CO)3(PR3)2 whereas H2 does so
only in equilibrium fashion. It is instructive to examine Table
7 which compares activation of these ligands on various d6
centers, along with other π-acceptor, “electronic gauge” ligands.
The values of JHD, JSiH, and H-H distance correlate well as
indicators of the activation of the σ bond, and νNN and νSO
measure the metal’s electron richness (back-bonding ability).
Morris proposed that M-H2 binding is stable if νNN is 2060-
L
W(CO)5(L)a W(CO)3(PCy3)2(L)b Mo(CO)(dppe)2(L)c
SO2
H2
2002
1971
1873
1843
1901
1815
C2H4
N2
argon
1973
1961
1932
1926
1921
1919e
1834
1835
1813
1809
CH4/agosticd
pyridine
NR3
1797
1757
1788f
1723
1718
1723e
a References 5d, 56 (SO2), 57 (N2). b References 1a, 3. c References
30 (SO2, H2, agostic), 58. d L ) CH4 (matrix) for W(CO)5(L); agostic
e
C-H for others. NR3 ) NEt2H. f NR3 ) NH2Bu.
experimentally, H2 rotational barriers as high as 11 kcal/mol
have now been seen in H2 complexes,45 implying at least this
much back-donation. Surprisingly, EBD in Mo(CO)5(H2) is
about half ED despite the seemingly poor back-bonding ability
of this fragment. These unexpected findings suggest that H2
complexes might best be pictured as triangulo species with
partial bonds joining the three atoms rather than the usual
3-center representation.
2150 cm-1 48
and this has held up fairly well (highly electro-
,
philic centers may be an exception). Thus H2 (and silane)
binding to Cr(CO)5 is unstable (νNN ) ∼2230 cm-1) but H2
cleaves on the much more electron-rich Mo(CO)(depe)2 (νNN
) 2050 cm-1).30 However the H-H bond is never >0.9 Å for
any of the group 6 systems and breaks suddenly, whereas it
usually elongates for later transition metals. The presence of a
strong π-acceptor, CO, trans to H2 is a major factor because it
disfavors OA of H2.33
In comparison to H2, PhSiH3 did not oxidatively add to Mo-
(CO)(depe)2 eVen though silanes appear to be the better
π-acceptors because, unlike H2, silanes rearrange their coordina-
38
tion position to be cis to CO, ostensibly to aVoid competing
with CO for back-donation.
In line with this, our normal-coordinate vibrational analysis of
W(CO)3(PCy3)2(H2) shows that the H-H and W-H stretches
are highly coupled and that the force constants for these
coordinates are nearly equal, with the H-H constant reduced
4-fold from its value in free H2.46
Dihydrogen is an excellent π-acceptor, perhaps even superior
to CO47 and at least comparable to olefins and N2 judging by
similarity of νCO for CO trans to L ) H2, N2, C2H4 in M(CO)5L
in liquid Xe,5d heptane,5j and the gas phase5k and also in
W(CO)3(PR3)2(L) and Mo(CO)(dppe)2(L) (Table 6). Strong
acceptors (SO2) give high νCO while pure σ-donors give much
lower values; the values all drop proportionately as the electron
richness of the metal center increases from left to right.
Importantly, these data also show that CH4 and agostic C-H
are poor π-acceptors (similar to amines), which had been
rationalized theoretically32a by orbital mismatch (overlap is large
between metal d and H2 σ*). This lack of back-bonding helps
explain the instability of alkane complexes.5i
(11)
In contrast to “stretching” the X-H bond, some systems give
equilibria between σ-complex and OA products (eqs 9 and 10).
Theoretical calculations49 on Mo(CO)(PH3)4(H‚‚‚SiH3) also
favored coordination of SiH4 cis to CO for this electronic reason.
The reversal of results for H2 and PhSiH3 addition to W(CO)3-
(PR3)2 compared to that for Mo(CO)(depe)2 demonstrates that
a fine balance of forces exists here that may go beyond
electronics, e.g. steric factors or structural rearrangement barriers
for 6- versus 7-coordination. Theoretical and experimental
studies indicate overall bond energetics can be crucial.50 The
relative M-X energies can offset the X-H energy difference
(H-H is 104 kcal/mol versus only 72-90 kcal/mol for Si-
(48) Morris, R. H.; Earl, K. A.; Luck, R. L.; Lazarowych, N. J.; Sella, A.
Inorg. Chem. 1987, 26, 2674.
(49) Fan, M.-F.; Jia, G.; Lin, Z. J. Am. Chem. Soc. 1996, 118, 9915.
(50) (a) Musaev, D.; Morokuma, K. J. Am. Chem. Soc. 1995, 117, 799.
(b) Maseras, F.; Lledos, A. Organometallics 1996, 15, 1218. (c)
Esteruelas, M. A.; Oro, L. A.; Valero, C. Organometallics 1991, 10,
462. (d) Luo, X.-L.; Crabtree, R. H. J. Am. Chem. Soc. 1989, 111,
2527. (e) Poulton, J. T.; Sigala, M. P.; Eisenstein, O.; Caulton, K. G.
(45) Jalon, F. A.; Otero, A.; Manzano, B. R.; Villasenor, E.; Chaudret, B.
J. Am. Chem. Soc. 1995, 117, 10123.
(46) Bender, B. R.; Kubas, G. J.; Jones, L. H.; Swanson, B. I.; Eckert, J.;
Capps, K. B.; Hoff, C. D. J. Am. Chem. Soc., in press.
(47) Morris, R. H.; Schlaf, M. Inorg. Chem. 1994, 33, 1725.

H-H versus Si-H σ-Bond Coordination
Inorganic Chemistry, Vol. 36, No. 15, 1997 3351
Table 7. Comparison of Small-Molecule Binding and Activation on Various d6 Metal Fragments (OA ) Oxidative Addition)
H2
PhSiH3
JSiH
Ph2SiH2
JSiH
N2
νNN
SO2
νSO
a
c
c
d
e
metal fragment
Cr(CO)3(PR3)2
W(CO)3(PR3)2
JHD
H-Hb
35f
0.85g
0.89f
0.88
0.87
N.R.g
OA
57, weak
41
2128g
2120g
2090
2062
34, equil OAf
eq 8
weak
1237g
1209
1180
2050
Mo(CO)(Ph2PC2H4PPh2)2
Mo(CO)(Bz2PC2H4PBz2)2
Mo(CO)(Et2PC2H4PEt2)2
[Mn(CO)(Ph2PC2H4PPh2)2]+
CpMn(CO)2
34
30
OA
33
33
39
50
1164
0.89
weakh
2167
2173
∼63.5i
1270
[CpFe(CO)(PEt3)]+
31.6
58
∼1282j
a Hz. References (from top down): 6d, 1c, 30, 8j, 29a, 59, 52. b Å. From solid state 1H NMR (refs 6d, 41, 29a). Hz. For coordinated Si-H
c
bond (refs 38b, 37a, 52; values for uncoordinated Si-H ) 164-205 Hz). d cm-1. References 2, 3, 30, 8j, 29, 51. e cm-1; asymmetric stretch (refs
8j (Mo), 60-62). f R ) iPr. g R ) Cy. JSiH could not be measured because of broadening due to the quadrupolar Mn nucleus. i Value for MeCp
h
derivative. j Value for [CpFe(dppe)(SO2)]+.
H50f). Thus calculations by Maseras and Lledos50b showed little
energy difference between OsCl(CO)(PH3)2(H)(η2-H-SiH3) and
OsCl(CO)(PH3)2(SiH3)(η2-H2), although experimentally there is
preference for the latter.50c On the other hand, the [IrH2(η2-
HSiEt3)2(PPh3)2]+ structure was found on this fragment.50d
Addition of H2 to RuH2(CO)(PtBu2Me)2 gave the η2-H2 adduct,
but HSiMe3 addition led to OA.50e Thus the relative stability
of the two forms may depend on a subtle balance of thermo-
dynamics by themselves or in combination with back-bonding
arguments.
not interact with excess PhSiH3 (steric congestion around a
smaller first-row metal2 may inhibit silane binding). As
mentioned above, the highly electrophilic cationic [Mn(CO)-
(dppe)2]+ containing a double agostic interaction29a binds H2
similarly to Mo(CO)(dppe)2 because increased H2 to Mn+
σ
donation offsets decreased back-bonding. However, H2 binds
significantly more strongly than either PhSiH3, SiH4, or N2 to
Mn+ as compared to Mo.29b
Theoretical comparison between Si-H and H-H σ-bond
activation was recently made by Fan et al.,49 by Bader-type
analysis of the electron density around the Mo(H‚‚‚Si) triangle
in Mo(CO)(PH3)4(H‚‚‚SiH3). A generic conclusion was made
that coordinated Si-H lies closer to the classical hydrido-silyl
extreme but that “the H-H bond remains essentially unperturbed
by MfH2 σ* back-bonding, and it lies closer to the nonclassical
dihydrogen extreme”. Regarding Si-H binding, this could well
be the situation in many cases, particularly in [Fe(CO)3(µ-
SiHPh2)]2, where Si‚‚‚H is 2.10 Å (50% elongation),39f and
perhaps even in Mo(CO)(depe)2(PhSiH3), where Si‚‚‚H is
elongated 25%. However, for H2 activation, the conclusions
are too general since many complexes with elongated H‚‚‚H
exist that clearly are perturbed by back-bonding. Also, all
indications point to strong MfH2 back-bonding eVen in true
unstretched H2 complexes and similarity in the reaction coor-
dinate for OA of H-H and Si-H bonds. Inexplicably however,
in some cases, particularly for group 6 complexes, the H-H
bond does seem less perturbed by back-bonding than it should
be. Why are H-H distances not longer than 0.9 Å in H2
complexes perilously close to OA, such as Mo(CO)(dppe)2(H2)
where EBD is 21 kcal/mol, nearly twice that for ED? Also, why
do tautomeric equilibria occur between a σ complex with a short
X-H and the OA product (eqs 9 and 10) instead of bond
elongation?
A possible answer is that complexes with H-H distances of
0.85-0.95 Å and high JHD values (>30 Hz) may actually lie
further along the reaction coordinate than originally believed,
which is supported by the H-H force constant46 and EBD data.
Complexes with much longer H-H distances (“stretched H2
complexes” or “compressed dihydrides”), would then lie even
further along the reaction coordinate. H atom abstraction
enthalpies calculated by Morris indeed suggest that complexes
with H-H distance >1.1 Å contain little H-H bonding and
are more consistent with M-H bonding alone.1h All of this is
very subjective because of the continuum nature of the bond
activation. Even classical dihydrides have been shown calcu-
lationally to have weak residual H‚‚‚H interactions,54 and trying
It might be possible to conclude from Table 7 that silanes
undergo OA more easily than H2 as the electrophilicity of the
metal center increases, perhaps because of the superior π-ac-
ceptor strength of Si-H. As the metal becomes more electron
rich, the differences in π-acceptor strength of the σ-ligands
should not matter as much in bond cleavage, and H2 could even
undergo OA more easily than silanes for steric or other reasons.
The W(CO)3P2 center is more electrophilic than Mo(CO)P4 on
the basis of IR evidence in Tables 6 and 7 and calculations of
EBD43 and indeed cleaves silanes more easily than H2. However,
unlike H-H activation, Si-H (and C-H) actiVation can be
tuned two ways: varying the substituents at both M and Si.
Making the Si more electropositiVe using substituents such as
Cl favors back-bonding that increases binding energy and
ultimately leads to Si-H cleavage.37a-d Significantly, the
opposite trend is seen37e on highly electrophilic centers such as
Cr(CO)5, where ED is more important to the binding energy.
SiH4 undergoes equilibrium OA on Mo(CO)(depe)2 (eq 10), but
organosilanes bind only as σ-ligands As seen in Table 7, JSiH
is 11 Hz lower for the PhSiH3 complex than the Ph2SiH2
complex, while the Si-H distance (1.77(6) Å38b) is correspond-
ingly longer than in the latter (1.60(6) Å38c). Therefore
comparisons of silane and H2 activation are valid only for silanes
with similar substituents. The most electrophilic fragment in
Table 7, CpMn(CO)2, readily coordinates Ph2SiH2 and oxida-
37
tively adds HSiCl3 but was only recently found to give an
isolable H2 complex51 from supercritical CO2 (JHD ) 33 Hz;
52
i.e. relatively unactivated). Even more electrophilic CpV(CO)3
and [CpFe(CO)(PEt3)]+ 53 give unstable H2 complexes but easily
bind silanes and cleave52 HSiEtCl2.
Unlike the above systems, Cr(CO)3(PCy3)2 coordinates H2
under several atmospheres of pressure in solution2,6d but does
Inorg. Chem. 1993, 32, 5490. (f) The value of 72 kcal/mol for the
Si-H bond energy of SiH4 in ref 50b is probably too low because it
is an average value. The energy of the HSi3-H bond is 90 kcal/mol
according to: Walsh, R. Acc. Chem. Res. 1981, 14, 246.
(51) Banister, J. A.; Lee, P. D.; Poliakoff, M. Organometallics 1995, 14,
3876.
(52) George, M. W.; Haward, M. T.; Hamley, P. A.; Hughes, C.; Johnson,
F. P. A.; Popov, V. K.; Poliakoff, M. J. Am. Chem. Soc. 1993, 115,
2286.
(54) Jackson, S. A.; Eisenstein, O. J. Am. Chem. Soc. 1990, 112, 7203.
(55) (a) Lang, R. F.; Ju, T. D.; Kiss, G.; Hoff, C. D.; Bryan, J. C.; Kubas,
G. J. J. Am. Chem. Soc. 1994, 116, 7917. (b) Lang, R. F.; Ju, T. D.;
Kiss, G.; Hoff, C. D.; Bryan, J. C.; Kubas, G. J. Submitted to Inorg.
Chim. Acta.
(53) Scharrer, E.; Chang, S.; Brookhart, M. Organometallics 1995, 14, 5686.

3352 Inorganic Chemistry, Vol. 36, No. 15, 1997
Butts et al.
All J values are reported in hertz. IR spectra were obtained on a Bio-
Rad FTS-40 instrument. Elemental analyses were obtained from Oneida
Research Services, Inc.
Toluene, hexane, Et2O, and THF were distilled from sodium/
benzophenone. The M(CO)3(PR3)2 (M ) Mo or W, R ) Cy; M ) W,
R ) iPr) compounds were prepared according to the literature.3 All
other reagents were obtained from commercial suppliers and used
without further purification.
Sealed NMR tubes were prepared by connecting an NMR tube to a
Kontes vacuum adapter via a Cajon joint, freezing the NMR tube in
liquid nitrogen, and flame-sealing the tube with an oxygen/propane
torch. “Glass bomb” refers to a cylindrical, medium-walled Pyrex
vessel joined to a Kontes K-826510 high-vacuum Teflon stopcock.
Variable-Temperature NMR Spectroscopy of M(CO)3(PCy3)2 (M
) Mo, W). Inside the drybox, Mo(CO)3(PCy3)2 (12 mg, 1.62 × 10-5
mol) was added to an NMR tube followed by 0.5 mL of toluene-d8.
The tube was then flame-sealed under He, and the contents were
analyzed by NMR spectroscopy on a Bruker AMX 500 MHz instru-
ment. Samples of W(CO)3(PCy3)2, M(CO)3[P(Cy-d11)3]2 (M ) Mo,
W), and Mo(CO)3(PCy3)2 in a 1:1:1 o-xylene:Et2O:hexane mixture were
each prepared and studied in the same manner.
to place boundaries on categorizations such as “stretched
dihydrogen complexes” or pinpoint the exact moment of H-H
cleavage will always be debatable.
Summary and Conclusions
NMR investigation of the temperature-dependent behavior
of the agostic species M(CO)3(PCy3)2 indicates the presence of
conformational isomers resulting from hindered rotation of the
M-P bond or, less likely, a geometric isomer with pseudo-cis
PCy3 ligands. The structure of W(CO)3(PCy3)2(C2H4) demon-
strates that, despite crowding by the large phosphines, ethylene
has a strong preference for aligning along the P-W-P axis as
in the case of H2 coordination. Unlike the case of H2
σ
coordination, PhSiH3 oxidatively adds to W(CO)3(PR3)2. On
the other hand, Ph2SiH2 reacts at 60 °C to eliminate H2 and
PiPr3 and form [W(CO)3(PiPr3)(µ-SiHPh2)]2 with 3-center
W‚‚‚H‚‚‚Si interactions.
Evidence indicates that both H2 and silane ligands are
powerful π-acceptors like ethylene. Activation of these σ
ligands on metal fragments follows a relatively simple model
yet offers a wealth of opportunity for study. There are at least
five primary variables in M(η2-X-H) systems influencing the
electronics and hence activation toward OA: the nature of M
and X, the substituents at both M and X, and overall charge.
Additionally, steric factors and overall energetics of OA
processes could play important roles in determining the point
at which a σ bond breaks. Thus, despite the similarity of the
reaction coordinate for H-H and Si-H bond cleavage, on some
metal fragments, e.g. Mo(CO)(depe)2, H2 undergoes OA more
easily than silanes, but on others such as W(CO)3(PR3)2, the
reverse is true.
Correlations such as those in Table 7 should be valuable in
defining electronic structure requirements for the optimal design
of metal fragments for σ-bond activation. Indications are that
metal-alkane complexes could best be stabilized toward
isolation on extremely electrophilic, positively-charged metal
centers, which were recently found to give quite stable silane
and H2 complexes. Since back-bonding is generally weak in
alkane complexes, it might be more important to increase σ
donation as much as possible (this also avoids C-H bond
cleavage). Theoretical studies predict stronger stabilization of
charged bare-metal M+-CH4 complexes compared to their
neutral analogues,63 and experimental studies show that CH4
actually binds more strongly than H2 to Co+.64
[W(CO)3(PiPr3)2](µ-N2) (1). A 50 mL glass bomb was charged
with W(CO)3(PiPr3)2 (204 mg, 3.47 × 10-4 mol) followed by 15 mL
of toluene in the drybox. The solution was flushed with N2 on a vacuum
line by evacuating and then backfilling with N2 two times. The yellow-
brown solution was held at room temperature for 4 h, being backfilled
with N2 once more after the first 2 h. The solution was concentrated
to 8 mL under vacuum, during the course of which a small amount of
orange powder precipitated, which was subsequently redissolved by
warming the solution. Inside the drybox, the yellow-brown solution
was layered with 6 mL of hexane. Orange block crystals of 1 were
isolated in 66% yield (139 mg) upon cooling this solution to -35 °C.
1H NMR (C6D6): δ 2.30 (m, 12H, CH), 1.20 (m, 72H, CH3). 13C{1H}
NMR (C6D6): δ 211.42 (t, JCP ) 5.3, CO), 210.12 (t, JCP ) 6.5, CO),
28.17 (vt, JCP ) 9.4, CH), 20.25 (s, CH3). 31P{1H} NMR (C6D6): δ
40.15 (s, JPW ) 276.6). IR (KBr): νCO ) 1946 (s), 1857 (sh), 1836
(br) cm-1
. Anal. Calcd for C42H84N2O6P4W2: C, 41.87; H, 7.03.
Found: C, 42.22; H, 7.27.
W(H)(CO)3(PiPr3)2(BF4). The complex W(CO)3(PiPr3)2 (151 mg,
2.57 × 10-4 mol) was dissolved in 8 mL of toluene in a 50 mL flask
inside the drybox. A diluted solution of HBF4‚OEt2 (40%, 259 mg,
6.40 × 10-4 mol, 2.5 equiv) in 5 mL of Et2O was then added dropwise
with stirring. The dark purple solution turned dark yellow over the
course of the addition. The solution was allowed to stir at 25 °C for
1 h, during which a small amount of yellow precipitate formed. The
volatile materials were removed under vacuum, leaving a dark yellow
powder. This was washed with hexanes (2 × 5 mL) and then extracted
with 10 mL of toluene. The yellow-brown solution was filtered, the
filtrate was reduced to 5 mL under vacuum, and the concentrate was
layered with 5 mL of hexane. Cooling this solution to -30 °C afforded
yellow crystals of W(H)(CO)3(PiPr3)2(BF4) (108 mg, 62% yield). This
compound can also be synthesized by following this procedure using
1 as the starting material rather than the agostic complex. 1H NMR
(C6D6): δ 2.51 (br m, 6H, CH), 1.15 (br m, 36H, CH3), -6.06 (m,
1H, W-H). 13C{1H} NMR (C6D6): δ 208.56 (br, CO), 207.15 (t, JCP
) 6.8, CO), 25.43 (d, JCP ) 21.4, CH), 19.73 (s, CH3). 31P{1H} NMR
Experimental Section
General Materials and Procedures. Unless otherwise noted, all
reactions and manipulations were performed in dry glassware under a
helium atmosphere in a Vacuum Atmospheres drybox or by using
1
1
standard Schlenk techniques. All H, H{31P}, 13C{1H}, and 31P{1H}
NMR spectra were recorded on either a commercial 500 MHz Bruker
AMX series spectrometer or a Bruker WM 300 MHz spectrometer.
(C6D6): δ 45.5 (br AB pattern). IR (KBr): νWH ) 2023 (m), νCO
)
1920 (s), 1913 (sh), 1883 (s) cm-1. IR (toluene): νWH ) 2002 (s),
νCO ) 1906 (br), 1889 (br), 1872 (br) cm-1. This complex was not
stable in the solid state at room temperature, and thus a satisfactory
elemental analysis could not be obtained.
(56) Schenk, W. A.; Baumann, F.-E. Chem. Ber. 1982, 115, 2615.
(57) Burdett, J. K.; Downs, A. J.; Gaskill, G. P.; Graham, M. A.; Turner,
J. J.; Turner, R. F. Inorg. Chem. 1978, 17, 523.
NMR and Crystal Growth of W(CO)3(PCy3)2(η2-C2H4) (2). The
complex W(CO)3(PCy3)2 (19 mg, 2.29 × 10-5 mol) was added to a 7
in. J. Young NMR tube followed by 0.5 mL of C6D6, forming a dark
purple solution. On a vacuum line, the solution (only) was frozen in
liquid N2 and evacuated after which it was backfilled with 1 atm of
ethylene. The solution became yellow immediately upon thawing and
was analyzed by NMR spectroscopy. In a separate experiment, pale
yellow crystals were grown by slowly cooling in a freezer a nearly
saturated toluene solution of 2 under N2 in a Schlenk flask placed inside
a Dewar flask. 1H NMR (C6D6): δ 2.44 (br, 6H, Cy), 2.36 (br s, 4H,
C2H4), 2.17 (br m, 12H, Cy), 1.68 (m, 24H, Cy), 1.26 (m, 24H, Cy).
(58) (a) Sato, M.; Tatsumi, T.; Kodama, T.; Hidai, M.; Uchida, T.; Uchida,
Y. J. Am. Chem. Soc. 1978, 100, 4447. (b) Tatsumi, T.; Tominaga,
H.; Hidai, M.; Uchida, Y. J. Organomet. Chem. 1980, 199, 63.
(59) Lawless, G. A. Private communication (see also footnote 31 in ref
51).
(60) Kubas, G. J.; Jarvinen, G. D.; Ryan, R. R. J. Am. Chem. Soc. 1983,
105, 1883.
(61) Barbeau, C.; Dubey, R. J. Can. J. Chem. 1973, 51, 3684.
(62) Schenk, W. A.; Karl, U.; Horn, M. R. Z. Naturforsch., B 1989, 44B,
1513.
(63) Blomberg, M. R. A.; Siegbahn, P. E. M.; Svensson, M. J. Phys. Chem.
1994, 98, 2062.
(64) Haynes, C. L.; Armentrout, P. B. Chem. Phys. Lett. 1996, 249, 64.
13C{1H} NMR (C6D6): δ 212.88 (t, JCP ) 6.4, CO), 209.49 (t, JCP
)

H-H versus Si-H σ-Bond Coordination
Inorganic Chemistry, Vol. 36, No. 15, 1997 3353
6.4, CO), 28.60 (br t, J ) 4.2, CH2CH2), 27-39 (Cy). 31P{1H} NMR
(C6D6): δ 19.60 (s, JPW ) 227).
Hz, JSiH ) 52). IR (KBr): νCO ) 1977, 1909, 1874 cm-1. Anal. Calcd
for C48H64O6P4Si2W2‚1/2C4H8O: C, 48.23; H, 5.60. Found: C, 47.85;
H, 5.29.
Reactions of W(CO)3(PR3)2 with PhSiH3. (a) W(CO)3(PCy3)2-
(H)(SiH2Ph) (3). W(CO)3(PCy3)2 (12 mg, 1.45 × 10-5 mol) was
dissolved in 0.5 mL of toluene-d8 in the drybox. Phenylsilane (1.9
µL, 1.54 × 10-5 mol) was then added to the purple solution, which
immediately turned brown. The solution was transferred to an NMR
tube which was flame-sealed under He. The solution was bright yellow
on cooling (-30 °C) and converted back to brown on rewarming. The
Single-Crystal X-ray Diffraction Studies. A crystal of either 1
(orange cube), 2 (yellow cube), or 5 (red block) was mounted on a
glass fiber using silicone grease from a pool of mineral oil, bathed in
Ar. The crystal was immediately mounted on the goniostat and cooled
under a N2 stream. A summary of crystallographic data is given in
Table 2. Data were obtained on an Enraf-Nonius CAD4 (1), a Siemens
R3m/V (2), or a Siemens P4 (5) diffractometer using graphite
monochromated Mo KR (λ ) 0.710 73 Å) radiation. Intensities were
corrected for Lorentz and polarization effects, and empirical absorption
corrections based on a set of ψ scans were applied. The structures
were solved by direct methods and expanded in subsequent Fourier
syntheses. Calculations were carried out using Siemens’ SHELXTL.
The ethylene H atoms in 2 were located and refined with all their related
1,2 and 1,3 distances restrained to be equal. Each H atom was also
given an isotropic displacement parameter equal to 1.2 times the
equivalent isotropic displacement parameter of the atom to which it
was attached. All other H atoms in all three structures were placed in
calculated positions, refined using a riding model, and given isotropic
displacement parameters equal to 1.2 (CH, CH2) or 1.5 (CH3) times
the equivalent isotropic displacement parameter of the atoms to which
they were attached. Additionally, methyl H atomic positions were
allowed to rotate about the adjacent C-C bond. The hydride H atom
in 5 was not located. Full-matrix least-squares refinement of all data
1
reaction mixture was investigated by H and 31P{1H} NMR spectro-
scopy. The resonances for the oxidative-addition product W(CO)3-
(PCy3)2(H)(SiH2Ph) were broad at 0 °C and sharpened on cooling to
-30 °C. Further broadening at lower temperatures was attributed to
the slow-exchange regime being approached for the intramolecular
ligand exchange in the 7-coordinate complex. 1H NMR (toluene-d8,
-45 °C): δ 8.14 (d, J ) 7.1, 2H, o-Ph), 7.33 (t, J ) 7.4, 2H, m-Ph),
7.19 (t, J ) 7.7, 1H, p-Ph), 5.79 (s, JSiH ) 176.8, 2H, SiH), 2.35 (br,
6H, Cy), 1.77 (br, 12H, Cy), 1.61 (br, 24H, Cy), 1.23 (br, 24H, Cy),
1
-4.51 (t, JHP ) 14.9, JHW ) 28.2 (determined in an H{31P} NMR
experiment), 1H, WH). 31P{1H} NMR (toluene-d8, -20 °C): δ 31.7
(br s). IR (toluene, 25 °C): νCO ) 1973, 1898, 1854 cm-1
.
(b) W(CO)3(PiPr3)2(H)(SiH2Ph) (4). This compound was prepared
by the same method used for the analogous PCy3 complex. 1H NMR
(toluene-d8, -30 °C): δ 8.02 (d, J ) 7.1, 2H, o-Ph), 7.32 (t, J ) 7.4,
2H, m-Ph), 7.14 (t, J ) 7.7, 1H, p-Ph), 5.66 (d, JHH ) 1.8, JSiH
)
180.2, 2H, SiH), 2.32 (m, 6H, CH), 1.10 (vq, J ) 6.8, 36H, CH3),
-4.90 (dt, JHP ) 14.9, JHH ) 1.8, JHW ) 29.6, 1H, WH). 31P{1H}
NMR (toluene-d8, -30 °C): δ 32.9 (s, JPW ) 207). IR (toluene, 25
2
against F2 of the quantity ∑w(Fo - Fc2)2 was used to adjust the
positions and anisotropic thermal parameters of all non-hydrogen atoms.
All peaks in the final difference maps >1 e/Å3 were located near a
tungsten atom.
°C): νCO ) 1977, 1962, 1902 cm-1
.
(c) [W(CO)3(PiPr3)(µ-SiHPh2)]2 (5). In the drybox, W(CO)3(PiPr3)2
(402 mg, 6.83 × 10-4 mol) was added to a 50 mL bomb, followed by
10 mL of toluene. Diphenylsilane (380 µL, 2.05 mmol, 3 equiv) was
then added by syringe to the purple solution, which immediately turned
dark brown. The solution was heated in a 60 °C oil bath for 6 h, during
which the solution turned dark red. Inside the drybox, the volatile
materials were removed under vacuum, leaving a dark red oil, which
was triturated with hexane three times and then washed with hexane
(3 × 10 mL, dark red to yellow washings). Drying under vacuum
afforded a tan powder. This crude product was dissolved in 16 mL of
toluene, the solution was layered with 16 mL of hexane, and the mixture
was placed in a -30 °C freezer. Orange crystals were isolated in 34%
yield (142 mg). The sample for elemental analysis was recrystallized
a second time from THF/hexane at -30 °C. 1H NMR (THF-d8): δ
7.58 (m, 8H, Ph), 7.21 (m, 12H, Ph), 2.68 (m, 6H, CH), 1.22 (dd, J )
13.9, 7.2, 36H, CH3), -7.26 (m, 2H, WH). 13C{1H} NMR (THF-d8,
5 °C): δ 218.18 (d, JCP ) 16.5, CO), 202.65 (d, JCP ) 5.7, CO), 144.90
Acknowledgment. This work was supported by the Depart-
ment of Energy, Office of Basic Energy Sciences, Division of
Chemical Sciences, and the Laboratory Directed Research and
Development Office at Los Alamos. We are most grateful to
David M. Barnhart for performing the initial X-ray structure
determination of complex 2 (data were subsequently re-refined
against F2 by J.C.B.) and to Lori Van Der Sluys and G. R. K.
Khalsa for NMR investigations. We also thank Mike Heinekey
for helpful discussions concerning the temperature-dependent
behavior of the agostic complexes.
Supporting Information Available: X-ray crystallography files,
in CIF format, for the structure determinations of complexes 1, 2, and
5 are available on the Internet only. Access information is given on
any current masthead page.
(s, Ph), 136.98 (s, Ph), 129.51 (s, Ph), 127.64 (s, Ph), 31.56 (d, JCP
)
22.9, CH), 20.20 (s, CH3). 31P{1H} NMR (THF-d8, 0 °C): δ 36.7 (s,
JPW ) 234). 29Si NMR (THF-d8, -20 °C): δ 146.3 (br d, ν1/2 ) 23
IC960870A