Indenyl-nickel complexes bearing a pendant, hemilabile olefin ligand: Preparation, characterization, and catalytic activities

Article abstract of DOI:10.1021/om050285u

The reaction of (PPh₃)₂NiCl₂ with Li[Ind∧CH=CH₂] gave the neutral complexes (η: η⁰-Ind∧CH=CH₂)Ni(PPh₃)Cl (Ind = indenyl; ∧ = (CH₂)₂, 1a; Si(Me)₂CH ₂, 1b), which were subjected to Cl^- abstraction to give the corresponding cationic complexes [(η:η²-Ind∧CH= CH₂)Ni(PPh₃)]⁺ (∧ = (CH₂) ₂, 2a; Si(Me)₂CH₂, 2b). The bis(phosphine) derivatives [(η:η⁰-Ind-CH₂CH₂CH=CH ₂)Ni(PPh₃)₂]⁺ (3a) and [(η³-allyl)Ni(PPh₃)₂]⁺ (4) formed gradually from room-temperature solutions of 2a and 2b, respectively, even in the absence of added PPh₃. On the other hand, [(η:η⁰-Ind-SiMe₂CH₂CH=CH ₂)Ni(PPh₃)₂]⁺ (3b) was detected only when PPh₃ was added to a CD₂Cl₂ solution of 2b. The lability of the vinyl moiety in 2 allows these complexes to act as single-component precatalysts for the polymerization and hydrosilylation of styrene; the latter reaction requires little or no induction period with the hydrosilanes PhRSiH₂ (R= Ph, Me, H) and proceeds with up to 1000 catalytic turnovers. Compounds 1a, 1b, 2a, 3a, and 4 have been characterized by NMR and single-crystal X-ray diffraction studies, whereas 2b and 3b were identified by NMR spectroscopy. Structural information gleaned from both solid-state and solution data provide important information on the Niolefin bonding in 2a and 2b and indicate that the Ni-Ind interactions in these complexes are affected by the significant trans influence of the chelating olefin moiety.

Full text of DOI:10.1021/om050285u

Organometallics 2005, 24, 4003-4013
4003
Indenyl-Nickel Complexes Bearing a Pendant,
Hemilabile Olefin Ligand: Preparation,
Characterization, and Catalytic Activities
Daniel Gareau, Christine Sui-Seng, Laurent F. Groux, Franc¸ois Brisse, and
Davit Zargarian*
De´partement de Chimie, Universite´ de Montre´al, Que´bec, Canada H3C 3J7
Received April 13, 2005
The reaction of (PPh3)2NiCl2 with Li[Ind∧CHdCH2] gave the neutral complexes
(η:η0-Ind∧CHdCH2)Ni(PPh3)Cl (Ind ) indenyl; ∧ ) (CH2)2, 1a; Si(Me)2CH2, 1b), which were
subjected to Cl- abstraction to give the corresponding cationic complexes [(η:η2-Ind∧CHd
CH2)Ni(PPh3)]+ (∧ ) (CH2)2, 2a; Si(Me)2CH2, 2b). The bis(phosphine) derivatives [(η:η0-Ind-
CH2CH2CHdCH2)Ni(PPh3)2]+ (3a) and [(η3-allyl)Ni(PPh3)2]+ (4) formed gradually from room-
temperature solutions of 2a and 2b, respectively, even in the absence of added PPh3. On
the other hand, [(η:η0-Ind-SiMe2CH2CHdCH2)Ni(PPh3)2]+ (3b) was detected only when PPh3
was added to a CD2Cl2 solution of 2b. The lability of the vinyl moiety in 2 allows these
complexes to act as single-component precatalysts for the polymerization and hydrosilylation
of styrene; the latter reaction requires little or no induction period with the hydrosilanes
PhRSiH2 (R) Ph, Me, H) and proceeds with up to 1000 catalytic turnovers. Compounds 1a,
1b, 2a, 3a, and 4 have been characterized by NMR and single-crystal X-ray diffraction
studies, whereas 2b and 3b were identified by NMR spectroscopy. Structural information
gleaned from both solid-state and solution data provide important information on the Ni-
olefin bonding in 2a and 2b and indicate that the Ni-Ind interactions in these complexes
are affected by the significant trans influence of the chelating olefin moiety.
Scheme 1
Introduction
Recent reports have shown that the in-situ-generated
cationic indenyl-nickel complexes [IndNi(PPh3)]+ (Ind
) indenyl or its substituted derivatives) can catalyze
the dimerization of ethylene,1 oligo- and polymerization
of styrene and norbornene,2 and hydrosilylation of some
olefins and ketones.3 The catalytic activities of these
coordinatively unsaturated, highly electrophilic species
depend on the substrate concentration, such that the
catalysis proceeds efficiently in the presence of a large
excess of substrate but drops off precipitously with
decreasing substrate concentration. This loss of activity
stems from the tendency of the active species [IndNi-
(PPh3)]+ to decompose into an unknown number of side-
products, including the much less active bis(phosphine)
complexes [IndNi(PPh3)2]+ (Scheme 1).
with a tethered, hemilabile moiety L that can occupy
the open site around the Ni center to prevent the
formation of the bis(phosphine) derivatives. The ef-
fectiveness of this strategy is predicated on two fac-
tors: (a) the intramolecular LfNi interaction enables
the relatively weakly nucleophilic moieties L to compete
effectively with the much more nucleophilic phosphines
because of the chelation effect; (b) the reversible nature
of the LfNi interaction allows the substrates, which
are usually present in large excess, to approach the Ni
center, thereby facilitating the catalysis. Thus, we have
shown that the complexes [(η:η1-Ind∧NR′2)Ni(PR3)]+
(∧ ) various linkers tethering NR′2 moieties to Ind) are
efficient, single-component catalysts for the polymeri-
zation and hydrosilylation of styrene and dimerization
of ethylene; as expected, the catalytic activities of these
systems are highly dependent on the nucleophilicity of
the hemilabile amine moiety (Scheme 2).4
One strategy for circumventing this inherent deacti-
vation pathway consists of functionalizing the Ind ligand
* Corresponding author. Tel: 514-343-2247. Fax: 514-343-2468.
E-mail: zargarian.davit@umontreal.ca.
(1) (a) Vollmerhaus, R.; Be´langer-Garie´py, F.; Zargarian, D. Organo-
metallics 1997, 16, 4762. (b) Dubois, M.-A.; Wang, R.; Zargarian, D.;
Tian, J.; Vollmerhaus, R.; Li, Z.; Collins, S. Organometallics 2001, 20,
663.
(2) (a) Dubois, M.-A. M.Sc. Thesis, Universite´ de Montre´al, 2000.
(b) Sun, H.-M.; Li, W.-F.; Han, X.; Shen, Q.; Zhang, Y. J. Organomet.
Chem. 2003, 688, 132. (c) Li, W.-F.; Sun, H.-M.; Shen, Q.; Zhang, Y.;
Yu, K.-B. Polyhedron 2004, 23, 1473. (d) Jime´nez-Tenorio, M.; Puerta,
M. C.; Salcedo, I.; Valerga, P.; Costa, S. I.; Silva, L. C.; Gomes, P. T.
Organometallics 2004, 23, 3139.
(3) Fontaine, F.-G.; Nguyen, R.-V.; Zargarian, D. Can. J. Chem.
2003, 81, 1299.
10.1021/om050285u CCC: $30.25 © 2005 American Chemical Society
Publication on Web 06/28/2005

4004 Organometallics, Vol. 24, No. 16, 2005
Gareau et al.
compound (indenyl)Ni(PPh3)Cl.7 We have proposed
that the side-reaction leading to (PPh3)3NiCl is initiated
by the Ni-promoted coupling of the Ind ligands,
which results in the formation of (PPh3)nNi(0); the latter
then undergoes a comproportionation reaction with the
Ni(II) precursor (PPh3)2NiCl2 to give (PPh3)3NiCl. It is
not clear why this side-reaction occurs with Li[Ind-
(CH2)2CHdCH2] but not Li[IndSiMe2CH2CHdCH2]. The
different steric, electronic, and/or chelating properties
of these ligands are presumably responsible for the
observed differences in their reactivities, but no simple
correlation is discernible among the ligands Li[R-Ind]
that lead to the formation of (PPh3)3NiCl (e.g., R) H,7
CH2CH2CHdCH2) and those that do not (e.g., R ) Me,8
CH2CH2NMe2,4a-c SiMe2CH2CHdCH2). It should be
emphasized, however, that once isolated, both 1a and
1b are stable in solution over a few days.
Scheme 2
The success of the above approach prompted us to
investigate the reactivities of analogous systems bearing
hemilabile functional groups other than amines. We
elected to study the complexes [(η:η2-Ind∧olefin)Ni-
(PR3)]+ because they resemble a key intermediate in the
proposed mechanism for the reactions of olefins with
the in-situ-generated [IndNi(PR3)]+; these target com-
plexes would, therefore, provide a rare opportunity to
investigate the structural properties and reactivities of
species that are closely related to a fleeting intermediate
in the above-discussed catalytic reactions.
1
The H, 13C{1H}, and 31P{1H} NMR spectra of com-
This report presents the synthesis and characteriza-
tion of the complexes (η:η0-Ind∧CHdCH2)Ni(PPh3)Cl
(∧ ) (CH2)2, 1a; Si(Me)2CH2, 1b) and the corresponding
cationic derivatives [(η:η2-Ind∧CHdCH2)Ni(PPh3)]+
(∧ ) (CH2)2, 2a; Si(Me)2CH2, 2b) and [(η:η0-Ind∧CHd
CH2)Ni(PPh3)2]+ (∧ ) (CH2)2, 3a; Si(Me)2CH2, 3b). The
spontaneous formation of the known complex [(η3-allyl)-
Ni(PPh3)2]+ (4) from 2b is also described. NMR char-
acterization of compounds 1-4, solid-state structural
studies of complexes 1a, 1b, 2a, 3a, and 4, and reactivity
studies in the polymerization and hydrosilylation of
styrene will be described. Complexes 2 represent the
first indenyl-nickel complexes containing a π-bound
olefin, but the related complexes (η5:η2-(C5R4)(CH2)3-
CHdCH2)NiX (R ) H, Me; X ) Br, I, Me, Et, i-Pr) and
[(η5:η2-(C5Me4)(CH2)nCHdCH2)Ni(PPh3)]+ (n ) 2, 3)
have been reported previously.5
plexes 1a and 1b are quite similar to the corresponding
spectra of the previously studied derivatives (1-R-Ind)-
Ni(PPh3)Cl.9 For instance, each 31P{1H} NMR spectrum
displays a sharp singlet at ca. 30 ppm, while the 1H and
13C{1H} NMR spectra show the normally observed
resonances for this family of complexes, including the
characteristically upfield signals for H3 and C3 (see
Scheme 3 for the numbering scheme).10 Furthermore,
the sharp 13C {1H} signals observed for the side-chain
carbons imply that the vinyl moiety is not involved in a
dynamic process of coordination-dissociation; this con-
trasts with the observation of a dynamic, temperature-
dependent coordination-dissociation process for some
of the analogous amino-indenyl derivatives4a-d and
confirms that the vinyl moiety has a lower nucleo-
philicity toward nickel.
Comparison of the chemical shifts for some of the
13C signals in these complexes also revealed interesting
observations on the electronic effects of the silyl sub-
stituent.11 For instance, the 13C nuclei directly bonded
to Si in 1b (C1 and C9) resonate upfield of the corre-
sponding nuclei in 1a: 93.6 ppm vs 107.2 ppm, and
24.4 ppm vs 26.9 ppm, respectively. For the three
13C nuclei once removed from Si, the signals for two
(C2 and C7a) appear more downfield in 1b than in 1a
(109.7 ppm vs 103.4 ppm, and 134.4 ppm vs 130.4 ppm,
respectively), while the third (C10) shows the opposite
trend (135.4 ppm in 1b vs 139.0 ppm in 1a).
Result and Discussion
Synthesis and Characterization of Complexes 1.
The reaction of Li[Ind] with Cl∧CHdCH2 gave the
ligands H[Ind∧CHdCH2],6 which were deprotonated
and reacted with Ni(PPh3)2Cl2 to give the complexes
(η:η0-Ind∧CHdCH2)Ni(PPh3)Cl (∧ ) (CH2)2, 1a; Si-
(Me)2CH2, 1b) in 21% (1a) and 65% (1b) yields (Scheme
3).
Isolation of pure samples of 1a and 1b was compli-
cated by the high solubility of these complexes in all
common solvents, including hexane. The particularly
low yield of 1a was also caused by the formation, in
fairly large quantities, of (PPh3)3NiCl. This Ni(I) byprod-
uct has been encountered previously during the syn-
theses of other Ind-Ni derivatives, including the parent
Single crystals of 1a and 1b suitable for X-ray
analysis were obtained by slow diffusion of hexane into
a saturated Et2O solution. The crystal data and details
of data collection are listed in Table 1, a selection of
structural parameters is given in Table 2, and ORTEP
drawings are given in Figures 1 and 2. The solid-state
structures for 1a and 1b are very similar to those of
(4) (a) Groux, L. F.; Be´langer-Garie´py, F.; Zargarian, D.; Vollmer-
haus, R. Organometallics 2000, 19, 1507. (b) Groux, L. F.; Zargarian
D. Organometallics 2001, 20, 3811. (c) Groux, L. F.; Zargarian D.
Organometallics 2003, 22, 3124. (d) Groux, L. F.; Zargarian D.
Organometallics 2003, 22, 4759. (e) Groux, L. F.; Zargarian, D.; Simon,
L. C.; Soares, J. B. P. J. Mol. Catal. A: Chem. 2003, 193, 51.
(5) Lehmkuhl, H.; Na¨ser, J.; Mehler, G.; Keil, T.; Danowski, F.;
Benn, R.; Mynott, R.; Schroth, G.; Gabor, B.; Kru¨ger, C.; Betz, P. Chem.
Ber. 1991, 124, 441.
(6) The ligands H[Ind(CH2)2CHdCH2] and H[IndSi(Me)2CH2CHd
CH2] have also been prepared by slightly different procedures in the
following reports: (a) de Armas, J.; Kolis, S. P.; Hoveyda, A. H. J. Am.
Chem. Soc. 2000, 122, 5977. (b) Ready, T. E.; Chien, J. C. W.; Rausch,
M. D. J. Organomet. Chem. 1999, 583, 11. (c) Cano, J.; Gomez-Sal, P.;
Heinz, G.; Royo, P. Inorg. Chim. Acta 2003, 345, 15.
(7) Huber, T. A.; Be´langer-Garie´py, F.; Zargarian, D. Organo-
metallics 1995, 14, 4997.
(8) Huber, T. A.; Bayrakdarian, M.; Dion, S.; Dubuc, I.; Be´langer-
Garie´py, F.; Zargarian, D. Organometallics 1997, 16, 5811.
(9) Zargarian, D. Coord. Chem. Rev. 2002, 233-234, 157.
(10) We have shown previously (ref 9) that the upfield shifts of the
C3 and H3 nuclei result from the nearly sp3-hybridization of C3 and
the ring currents of PPh3 phenyl groups.
(11) For an in-depth discussion of the electronic effects of silyl
substituents on Cp-type ligands see the following reports: (a) Zach-
manoglou, C. E.; Docrat, A.; Bridgewater, B. M.; Parkin, G.; Brandow,
C. G.; Bercaw, J. E.; Jardine, C. N.; Lyall, M.; Green, J. C.; Keister, J.
B. J. Am. Chem. Soc. 2002, 124, 9525. (b) Langmaier, J.; Samec, Z.;
Varga, V.; Horacek, M.; Choukroun, R.; Mach, K. J. Organomet. Chem.
1999, 584, 323.

Indenyl-Nickel Complexes
Organometallics, Vol. 24, No. 16, 2005 4005
Scheme 3
Table 1. Crystal Data, Data Collection, and Structure Refinement Parameters of 1a, 1b, 2a, 3a, and 4
1a
1b
2a
3a
4
formula
C
31H28ClPNi
C32H32ClPSiNi
569.80
red, block
0.06 × 0.21 × 0.56
monoclinic
C2/c
28.763(9)
10.667(10)
19.103(11)
90
C55H48BPNi
809.42
red, block
0.16 × 0.16 × 0.19
monoclinic
P21/n
15.9152(2)
32.7827(3)
16.5352(2)
90
C73H63BNiP2
1071.69
yellow, block
0.50 × 0.43 × 0.08
monoclinic
P21/n
10.3056(7)
16.2965(11)
33.568(2)
90
C
39H35BF4NiP2
mol wt
525.66
711.1407
yellow, block
0.55 × 0.08 × 0.05
monoclinic
C2/c
cryst color, habit
cryst dimens, mm
symmetry
space group
a, Å
red, block
0.16 × 0.30 × 0.30
triclinic
P1h
9.15590(10)
10.4614(2)
15.5498(2)
90.5120(10)
106.2650(10)
113.8330(10)
2
25.0077(3)
17.1191(2)
17.6639(2)
90
b, Å
c, Å
R, deg
â, deg
97.44(3)
90
104.5180(10)
90
96.119(5)
90
118.0880(10)
90
γ, deg
Z
8
8
4
8
D(calcd), g cm-1
diffractometer
1.348
1.302
1.287
1.270
1.416
Bruker AXS
SMART 2K
223(2)
Nonius CAD-4
Bruker AXS
SMART 2K
223(2)
Bruker AXS
SMART 2K
100(2)
Bruker AXS
SMART 2K
100(2)
temp, K
223(2)
λ (KR), Å
µ, mm-1
scan type
F(000)
θmax (deg)
h,k,l range
1.54178 (Cu)
2.731
0.71073 (Mo)
0.875
1.54178 (Cu)
1.308
1.54178 (Cu)
1.368
1.54178 (Cu)
2.168
ω scan
ω scan
ω scan
ω scan
ω scan
548
72.91
2384
25.98
3408
72.98
2256
73.20
2944
72.90
-11 e h e 11
-12 e k e 12
-19 e l e 19
15 611
multiscan
SADABS
0.4100, 0.7300
0.0388, 0.1072
1.12
0 e h e 35
0 e k e 13
-23 e l e 23
5684
integration
Absorption
0.8400, 0.9500
0.0460, 0.1252
0.802
-19 e h e 19
-40 e k e 40
-20 e l e 20
16 463
multiscan
SADABS
0.6900, 0.8600
0.0434, 0.1126
0.958
-12 e h e 12
-20 e k e 20
-36 e l e 40
10 703
multiscan
SADABS
0.0400, 0.9300
0.0787, 0.1966
0.999
-30 e h e 30
-20 e k e 21
-21 e l e 20
6557
multiscan
SADABS
0.6000, 0.8800
0.0511, 0.1465
1.074
reflns used (I > 2σ(I))
absorption
correction
T (min, max)
R [F2 > 2σ(F2)], wR(F2)
GOF
the previously studied, closely analogous (1-R-Ind)Ni-
(PPh3)Cl complexes.9 The geometry of the Ni center in
these complexes can be described as distorted square
planar, the largest distortion arising from the small bite
angle of the Ind ligand (ca. 65°). The significantly longer
Ni-C3a and Ni-C7a distances relative to the Ni-

4006 Organometallics, Vol. 24, No. 16, 2005
Gareau et al.
Table 2. Selected Bond, Distances (Å) and Angles
(deg) for 1a, 1b, 2a, and 3a
1a
2.1727(5) 2.172(2) 2.2276(6) 2.1956(10)
2.1835(6) 2.192(3) 2.2136(11)
2.1302(19) 2.113(7) 2.0726(19) 2.143(3)
1b
2a
3a
Ni-P1
Ni-Cl or P2
Ni-C1
Ni-C2
Ni-C3
2.075(2)
2.034(2)
2.292(2)
2.332(2)
2.044(7) 2.069(2)
2.041(7) 2.073(2)
2.332(7) 2.297(2)
2.349(7) 2.330(2)
2.130(2)
2.089(3)
2.052(4)
2.299(3)
2.364(3)
Ni-C3a
Ni-C7a
Ni-C10
Ni-C11
C1-C2
2.072(2)
1.417(3)
1.421(3)
1.452(3)
1.461(3)
1.294(6)
1.407(10) 1.427(3)
1.393(9) 1.392(3)
1.457(10) 1.465(3)
1.469(10) 1.448(3)
1.304(10) 1.356(3)
96.31(9)
1.413(5)
1.409(5)
C2-C3
C3-C3a
C7a-C1
C10-C11
1.462(5)
1.482(5)
1.298(7)
P1-Ni-Cl or P2 97.94(2)
102.42(4)
C3-Ni-Cl or P2 162.42(7) 163.9(2)
162.24(10)
101.96(6) 94.12(11)
97.55(10)
C3-Ni-P1
99.51(7)
99.6(2)
97.4(2)
C1-Ni-Cl or P2 95.59(6)
Figure 2. ORTEP view of complex 1b. Thermal ellipsoids
are shown at 30% probability.
C1-Ni-P1
C1-Ni-C3
∆M-Ca (Å)
HAb (deg)
166.10(6) 166.2(2) 165.26(6) 159.96(11)
66.87(9)
0.23
10.35
9.94
66.7(3)
0.26
10.68
8.74
66.80(8)
0.24
11.15
11.99
66.27(14)
0.234
10.76
FAc (deg)
12.14
2 allowed us to recover ca. 45-50% crude yields, but
isolation of pure samples by recrystallization of the
crude material was complicated because these com-
pounds undergo a slow decomposition process, as fol-
lows.
a ∆(M-C) ) 0.5{(M-C3a + M-C7a)} - 0.5{(M-C1 + M-C3)}.
b HA is the angle between the planes formed by the atoms C1,
C2, C3 and C1, C3, C3a, C7a. c FA is the angle between the planes
formed by the atoms C1, C2, C3 and C3a, C4, C5, C6, C7 and C7a.
Monitoring freshly prepared samples of 2a by 31P{1H}
NMR spectroscopy showed only one singlet resonance
at ca. 34 ppm, but a minor set of AB doublets (2JP-P
)
27 Hz) emerged over a few hours at ca. 35 and 32 ppm.
Previous studies had shown that such AB signals
correspond to bisphosphine complexes of the type
[(R-Ind)Ni(PPh3)2]+;1a,12 we suspected, therefore, that
the new species arising from 2a was the bis(phosphine)
side-product [η,η0-IndCH2CH2CHdCH2)Ni(PPh3)2]+, 3a.
This was confirmed when we observed that adding
excess PPh3 to the sample of 2a accelerated the growth
of the aforementioned AB signals. Given the facile
conversion of 2a to 3a, isolation of pure samples of 2a
required that the abstraction of Cl- from 1a be carried
out rapidly (over minutes) and in highly dilute mixtures;
this approach furnished single crystals of both 2a and
3a suitable for X-ray analyses of these complexes (vide
infra).
Complex 2b was less susceptible to decomposition, but
solutions aged over 1 day or longer showed a number
of new 31P{1H} resonances; interestingly, however, the
AB doublets expected for the analogous bis(phosphine)
side-product [η,η0-(IndSiMe2CH2CHdCH2)-Ni(PPh3)2]+,
3b, were not detected in these spectra. The absence of
3b among the species arising from the decomposition
of 2b was intriguing and prompted us to test whether
excess PPh3 would facilitate the formation of 3b. These
tests showed that adding 1 equiv of PPh3 to a CD2Cl2
solution of 2b resulted in the replacement of the original
31P{1H} signal at ca. 33.7 ppm by a broad signal at ca.
31 ppm; this signal collapsed at 188 K into AB doublets
at 33.1 and 29.0 ppm (2JP-P ≈ 32 Hz). Comparison of
these signals to those obtained for 3a convinced us that
they were due to 3b, which appears to be in equilibrium
with 2b.
Figure 1. ORTEP view of complex 1a. Thermal ellipsoids
are shown at 30% probability.
C(allyl) bond lengths are in accord with the anticipated
lower hapticity of the Ind ligand, the so-called η5fη3
slippage, when it is coordinated to relatively electron-
rich metal centers. The calculated values of ∆M-C, the
slip factor, are in the usual range for the neutral Ni-
Cl derivatives (ca. 0.23-0.26 Å); similarly, the signifi-
cant asymmetry in the Ni-C(allyl) bond lengths
(Ni-C1 > Ni-C2 > Ni-C3) reflects the unequal trans
influences of the PPh3 and Cl ligands.9 The pendant
olefin moiety is oriented away from the Ni center in both
structures.
Synthesis and Characterization of Complexes 2
and 3. Abstraction of Cl- from 1a proceeded in the
presence of either NaBPh4 or AgBF4 in solvents such
as CH2Cl2 or C6H6 to give the cationic complex 2a in
ca. 50% yield, whereas the analogous conversion of 1b
into 2b required AgBF4 and proceeded in CH2Cl2 only
(Scheme 3). The lower solubility of cationic complexes
1
The H NMR spectrum was also consistent with the
above scenario. For instance, the addition of PPh3
(12) Fontaine, F.-G. Ph.D. Thesis, Universite´ de Montre´al, 2002.

Indenyl-Nickel Complexes
Organometallics, Vol. 24, No. 16, 2005 4007
diminished the intensity of the signals due to the
chelating vinyl moiety and resulted in the emergence
of signals at lower fields (ca. 5-6 ppm), assigned to the
nonchelating vinyl moiety. Similarly, the intensity of
the signals assigned to the chemically inequivalent
Si(CH3)2 groups in 2b diminished while a pair of new
signals appeared ca. 0.8 ppm upfield of the original
signals, presumably because of the interaction of the
SiMe2 protons with the PPh3 ligand; integration of the
two sets of SiMe2 signals indicated a 7:1 ratio (at room
temperature) in favor of the new species. These obser-
vations imply that 2b can be converted to 3b in the
presence of PPh3, but this process is reversible because
steric repulsion between the SiMe2 group and the
adjacent PPh3 ligand in 3b allows the olefin moiety to
compete effectively for coordination to Ni (Scheme 3).
The Cs symmetry of complexes 2a and 2b renders all
their nuclei chemically inequivalent, which complicated
1
the H and 13C{1H} NMR spectra of these complexes,
Figure 3. ORTEP view of one of the two independent
molecules found in the unit cell of complex 2a. Thermal
ellipsoids are shown at 30% probability.
but signal assignment was facilitated by the COSY and
HMQC spectra. Comparing the spectral features of 2
to those of their precursors 1 revealed a number of
interesting shifts in many of the signals, indicating
important structural differences between these com-
plexes. For instance, significant downfield shifts were
noted for H2/H3 (ca. 0.7-2 ppm) and for C1/C2/C3 (ca.
10-20 ppm). These shifts can be attributed to the
enhanced hapticity of Ind, which results from the
increased electrophilicity of the Ni center in cationic 2.
The downfield shifts noted for H3 also serve as an
indirect measure of the trans influence of the vinyl
moiety, and comparison to previously studied systems
allows us to attribute a very strong trans influence to
olefins, similar to that of CO.13
The chelation of the vinyl moiety in 2 was also
corroborated on the basis of spectral comparisons, as
follows. The primary indicators of the chelation were
the upfield shifts of the signals for the vinyl protons H10
(by ca. 0.5 ppm) and H11A/H11B (by ca. 1-2 ppm), as
well as the signals for C10 (by ca. 24 ppm) and C11 (by
ca. 40 ppm). These shifts can be justified by noting that
coordination of the vinyl moiety to Ni increases the sp3
character of C10 and C11 and places H11A and H11B
within the anisotropy cone of the PPh3 phenyl groups.14
On the other hand, chelation reduces the flexibility of
the side-chain, thereby magnifying the differences in the
chemical environments of the CH2 and Si(CH3)2 nuclei
and increasing the separation of their NMR signals. For
instance, the chemical shifts of the Si(CH3)2 signals are
less different in 1b (0.81 and 0.83 ppm) than in 2b (0.70
and 0.60 ppm).
Figure 4. ORTEP view of complex 3a. Thermal ellipsoids
are shown at 30% probability.
As discussed above, the coordination of two PPh3
ligands in 3 is most clearly evident from the AB doublets
in the 31P NMR spectra. The similar chemical shifts for
1
the H signals due to the vinylic protons H11 in 1 and
3 (ca. 5 ppm vs ca. 3-4 ppm for the corresponding
signals in 2) indicate that the olefin moiety in 3 is
dangling. On the other hand, the 1H and 13C signals for
Ind nuclei are fairly similar for the cationic complexes
2 and 3.
Single crystals suitable for X-ray diffraction studies
were obtained by repeated recrystallization in CH2Cl2/
Et2O for 2a and CH2Cl2/hexane for 3a. All attempts at
growing crystalline samples of 2b failed, giving instead
small quantities of a crystalline species identified as the
well-known complex [(η3-allyl)Ni(PPh3)2]+, 4 (vide infra).
The crystal data and details of data collection for 2a
and 3a are listed in Table 1, a selection of structural
parameters are listed in Table 2, and ORTEP drawings
are given in Figures 3 and 4. The solid-state structures
of 2a and 3a are described below.
(13) We have shown (ref 9) that the chemical shift of the H3 signal
in the analogous complexes (1-R-Ind)Ni(PPh3)X and [(1-R-Ind)Ni-
(PPh3)L]+ reflects the Ni-C3 bond strength and hence the trans
influence of the ligands X and L. The following trans influence order
has emerged for the ligands studied: Cl (3.30 ppm) < CC-Ph (3.9 ppm)
< phthalimidate (4.10 ppm) < Me (4.20 ppm) ∼ pyridine (4.21 ppm) <
PR3 (ca. 4.3-5.1 ppm) < ButNC: (4.80 ppm) < olefin (ca. 5.5 ppm) ∼
CO (5.53 ppm).
(14) The distances measured from the center of the phenyl ring
nearest the chelating vinyl moiety (i.e., ring C121-C126 in 2a) to the
vinyl hydrogen atoms are as follows: 5.356 Å for H10; 4.219 Å for
H11A.; 3.427 Å for H11B. In their report on the solid-state structure
of the related complex (η5:η2-(C5Me4)(CH2)3CHdCH2)NiBr (ref 5),
Lehmkuhl et al. have proposed an agostic interaction between Ni and
one of the vinylic protons (2.33(8) Å). Although comparable values were
noted for the corresponding Ni-H distances in 2a (2.44 Å for H111A,
2.54 Å for H111B, and 2.52 Å for H110), we are not convinced these
distances reflect significant Ni-H interactions.
The main structural features of the cationic complexes
2a and 3a are quite similar to those of the neutral
complexes 1, implying that the displacement of Cl by

4008 Organometallics, Vol. 24, No. 16, 2005
Gareau et al.
Chart 1
Chart 2
the pendant olefin moiety or a second PPh3 does not
result in major changes in the overall geometry of these
complexes. Interestingly, the ∆M-C values for 1a, 2a,
and 3a are virtually unchanged, implying no major
changes in the Ind-Ni interaction on going from the
neutral precursors to cationic derivatives. This contrasts
with the stronger Ind-Ni interaction inferred from the
solution NMR spectra of 2a and 3a (vide supra) and
expected on the basis of the presumed increase in the
electrophilicity of the Ni center in these cationic species.
Similarly, the observation that the Ni-P bonds are
longer by about 15-25 esd in cationic 2a and 3a relative
to neutral 1a and 1b might appear counterintuitive at
first since the more electrophilic Ni centers appear to
be forming weaker Ni-P bonds; this phenomenon can
be explained, however, in terms of the lower compat-
ibility between the positively charged, harder Lewis acid
Ni center and the soft Lewis base PPh3.15
2a. The C-C distance in this moiety is much longer
than that in the nonchelating complexes 1 (C10-C11)
1.294(6) Å in 1a and 1.304(10) Å in 1b vs 1.356(3) Å in
2a) but shorter than the corresponding C-C distance
in (η5,η2-C5Me4(CH2)3CHdCH2)Ni-Br (1.39(1) Å).5 Simi-
larly elongated C-C bonds (ca. 1.38-1.39 Å) have been
observed in the related, charge-neutral Cp-Ni(II) com-
plexes Cp*Ni(Ph)(CH2dCH2)16 and CpNi(R)(CH2dCH2)17
(Chart 1), whereas much greater elongation is observed
in an analogous Ni(II) complex bearing electron-
withdrawing CO2Me groups on the olefin moiety
(1.43 Å)18 and in one of the two ethylene moieties in
the Ni(0) complex (CH2dCH2)2Ni(PCy3) (1.42 Å).19 The
elongation of the C-C bond in the Ni-olefin moiety can
be attributed to the σ-donation from the olefinic π-elec-
tron density into empty Ni orbitals as well as π-back-
bonding from the filled Ni dπ orbitals into the empty π*
orbitals of the olefin moiety; the greater elongation
found in the charge-neutral Cp complexes is presumably
a reflection of more extensive π-back-donation from the
more electron-rich, neutral Ni center.
Finally, it is noteworthy that the Ni-C11 distance
in 2a is significantly shorter than Ni-C10 (by ca.
0.06 Å, about 30 times esd), which indicates that the
vinyl moiety coordinates to the Ni center in an unsym-
metrical fashion; in comparison, the difference in the
corresponding Ni-C bonds in (η5,η2-C5Me4(CH2)3CHd
CH2)Ni-Br5 is less significant (ca. 0.025 Å, about 3
times esd). It is tempting to propose that this nonsym-
metrical interaction implies that a charge-separated
canonical form makes significant contributions to the
bonding scheme in 2 (Chart 2); this proposal would help
explain why olefins susceptible to cationic polymeriza-
tion (e.g., styrene) are polymerized by this class of Ni
cations (vide infra).
Despite the overall similarities in the structures of
the neutral Ni-Cl precursors 1 and their cationic
derivatives 2a and 3a, there are important differences
in the coordination symmetry of the Ind ligand and the
disposition of the tethered olefin moiety in these com-
plexes. Thus, we observe highly symmetrical Ni-C
interactions for the allylic carbons in 2a (Ni-C1 ∼ Ni-
C3), implying that the vinyl moiety and PPh3 have
comparable trans influences. In contrast, the Ni-
C(allyl) interaction is as unsymmetrical in 3a as it is
in 1a (Ni-C3 < Ni-C1 by ca. 30 esd), even though both
of the allylic carbons in 3a are trans from a PPh3. We
attribute the unequal Ni-C1/C3 distances in 3a to the
unequal Ni-P bond lengths (Ni-P2 > Ni-P1 by ca. 18
esd), which are caused by the steric repulsion between
the Ind substituent and the PPh3 adjacent to it; in effect,
the two PPh3 ligands lead to unequal trans influences.
It is instructive to note that the Ni-C(Cp) interactions
in (η5,η2-C5Me4(CH2)3CHdCH2)Ni-Br5 are also unsym-
metrical because of the unequal trans influences of Br
and the chelating vinyl moiety. Moreover, the average
Ni-C(Cp) distance in the latter complex is intermediate
between the shorter Ni-C1/C2/C3 and the longer Ni-
C3a/C7a distances in 2a; as a result, the allyl-ene
distortion is much more evident in 2a.
Complex 4. As mentioned above, small amounts of
4 were obtained during the attempted recrystallization
The η2-bound pendant olefin moiety is perhaps the
most important feature of the solid-state structure in
(16) Lehmkuhl, H.; Keil, T.; Benn, R.; Rufinska, A.; Kru¨ger, C.;
Poplawska, J.; Bellenbaum, M. Chem. Ber. 1988, 121, 1931.
(17) Lehmkuhl, H.; Naydowski, C.; Benn, R.; Rufinska, A.; Schroth,
G.; Mynott, R.; Kru¨ger, C. Chem. Ber. 1983, 116, 2447.
(18) Battiste, M. A.; Griggs, B. G., Jr.; Sackett, D.; Coxon, J. M.;
Steel, P. J. J. Organomet. Chem. 1987, 330, 437.
(19) Kru¨ger, C.; Tsang, Y.-H. J. Organomet. Chem. 1972, 34, 387.
(15) For a discussion of Ni-phosphine and Ni-Cp bonding in the
complexes Cp*Ni(PR3)X see: Holland, P. L.; Andersen, R. A.; Bergman,
R. G.; Huang, J.; Nolan, S. J. Am. Chem. Soc. 1997, 119, 12800.

Indenyl-Nickel Complexes
Organometallics, Vol. 24, No. 16, 2005 4009
be polymerized by complex 2b alone (eq 1, Table 3).23
Optimization experiments demonstrated that the
choice of solvent was crucial for the efficacy of the
polymerization reactions: using chlorinated solvents
such as dichloroethane resulted in small catalytic
turnover numbers (ca. 40-60), whereas reactions car-
ried out in nonchlorinated solvents such as toluene or
no solvent at all gave ca. 500 turnovers. The GPC
analyses of the polymers obtained from the latter
experiments showed a bimodal distribution of low
molecular weight oligomers (Mw ∼ 1200-1600) and
much higher molecular weight polymers (Mw ≈ (5-7)
× 105); the polydispersities of both fractions were fairly
narrow (Mw/Mn ≈ 1.2-1.4). These results are qualita-
tively similar to those obtained from the polymerizations
catalyzed by [(η,η1-Ind-CH2CH2NMe2)Ni(PPh3)]+ or the
IndNi(PPh3)Cl/NaBPh4 combination. On the other hand,
the polymerizations catalyzed by 2b appear to be
somewhat more facile, as they proceed at milder tem-
peratures (60 °C vs 80 °C); indeed, the reactions
catalyzed by 2b can proceed even at room temperature,
albeit with lower catalytic turnover numbers (runs 4
and 5 vs runs 2 and 3).
Figure 5. ORTEP view of the major model (60%) for
complex 4. Thermal ellipsoids are shown at 30% prob-
ability. Selected bond lengths (Å) and angles (deg): Ni-P1
) 2.2049(8), Ni-P2 ) 2.1920(8), Ni-C1 ) 2.042(18), Ni-
C2 ) 2.015(4), Ni-C3 ) 2.03(2), C1-C2 ) 1.405(13), C2-
C3 ) 1.320(15), C1-Ni-C3 ) 71.4(5), P1-Ni-P2 )
103.12(3).
of 2b. The NMR spectra of 4 matched those reported
previously.20 The way in which the unexpected but
interesting transformation of 2b to 4 comes about is not
known. To be sure, desilylation of SiR3-substituted
indenyl ligands is precedented, but in most previous
cases it is the silyl substituent that is eliminated,
leaving behind an unsubstituted Ind ligand,21 whereas
in the present case it is the IndSiMe2 fragment that is
eliminated, leaving behind the allyl moiety. Since the
solid-state structure of this complex had not been
reported previously, we undertook an X-ray diffraction
study of the crystals obtained, as described below.
The solid-state structure of complex 4 is similar to
that of 3a in that two PPh3 ligands and the allyl moiety
are arranged around the Ni center. The central carbon
atom of the allyl moiety (C2) is distorted over two
positions, but the disorder was modeled in a satisfactory
fashion with two solutions (60:40). The ORTEP diagram
and a selection of bond distances and angles for the
major model are shown in Figure 5. The geometry
around the central Ni(II) is distorted square planar,
with a C1-Ni-C3 angle of ca. 71° and a P1-Ni-P2
angle of ca. 103°. The square plane defined by P1, P2,
C1, and C3 atoms makes an angle of ca. 115° with the
allyl moiety. The Ni-C1/C3 bond lengths are virtually
symmetrical, whereas the differences between Ni-P and
allylic C-C distances (C1-C2 vs C2-C3) are small but
significant. Fairly similar structural parameters are
found in the only other structurally characterized
Ni-(simple allyl) complex [(1,1,2-trimethylallyl)Ni-
(PPh3)2]+22 (Chart 1).
The polymerization reactions described above can, in
principle, proceed via either a Ni-initiated electrophilic
route or a Ni-based insertion mechanism. We used NMR
spectroscopy in an attempt to shed some light on the
reaction pathway; unfortunately, however, the spectra
recorded for C6D6 solutions (ca. 2 M in styrene and
1
0.02 M in 2b, at 60 °C) were uninformative: the H
NMR spectra showed only very broad, featureless
signals (at ca. 1-4 ppm), whereas the 31P{1H} NMR
spectra showed no signals at all. Thus, we have been
unable to determine which pathway is operative. It
seems reasonable, however, to suppose that the first
step in the polymerization process is the associative
displacement of the chelating vinyl moiety by styrene,
giving [(η,η0-Ind-SiMe2CH2CH2dCH)Ni(PPh3)(styrene)]+
(Scheme 4). The nature of the Ni-styrene interaction
in this intermediate would have a determining effect
on the polymerization mechanism. If we suppose a
(partial) positive charge on its R-carbon atom, it follows
that the coordinated styrene can undergo a nucleophilic
attack by a second molecule of monomer, as shown in
Scheme 4; subsequent attacks by styrene would then
lead to chain growth.
The above-proposed propagation sequence would then
continue until the Ni-CH2CH(Ph)-P moiety (P ) the
growing polymer chain) undergoes a â-H elimination to
give a Ni-H species and a carbocation-terminated
polymer chain. The latter would continue growing via
the electrophilic mechanism, with the carbocation acting
as the electrophile and the monomer as the nucelophile,
to give the high Mw poly(styrene) obtained in our
experiments.24 On the other hand, the Ni-H intermedi-
Catalytic Polymerization of Styrene. Previous
reports have shown that combining IndNi(PPh3)Cl and
NaBPh4 produces an active catalyst that polymerizes
styrene at 80 °C over 24-48 h to a mixture of oligo- and
poly(styrene).2 As anticipated, we found that styrene can
(23) Complex 2a was not included in the reactivity studies, because
its solutions were often contaminated by small amounts of [(1-CH2-
CH2CHdCH2-Ind)Ni(PPh3)2]+, 3a, which is inactive in the polymer-
ization reactions but active in the hydrosilylations.
(24) The proposed sequence of electrophilic attacks might also be
intramolecular, leading to the formation of cyclic oligomers.
(20) Castro, B.; Neibecker, D. J. Organomet. Chem. 1975, 85, C-39.
(21) Wang, B.; Zhu, B.; Xu, S.; Zhou, X. Organometallics 2003, 22,
4842.
(22) Zocchi, M.; Albinati, A. J. Organomet. Chem. 1974, 77, C40.

4010 Organometallics, Vol. 24, No. 16, 2005
Table 3. Oligo- and Polymerization of Styrene Catalyzed by 2ba
Gareau et al.
catalytic
turnovers
oligomer Mw (×103),
polymer Mw (×104),
oligomer:polymer
ratio
run
T (°C)
solvent (mL)
Mw/Mn
Mw/Mn
1b
2
60
60
60
25
25
(ClCH2)2 (8)
toluene (1)
none
toluene (1)
none
50
560
560
50
7.9, 2.55
61.7, 1.41
58.5, 1.40
64.1, 1.35
62.2, 1.42
1.4, 1.25
1.6, 1.39
1.4, 1.20
1.2, 1.27
90:10
85:15
80:20
65:35
3
4
5
120
a Unless otherwise noted, the polymerizations were run over 4 h using a catalyst:monomer ratio of 1:1000. b A 1:2000 ratio of 2b:
styrene was used in this run; reaction time ) 48 h.
Table 4. Catalytic Formation of
Scheme 4
a
PhCH(PhSiH2)CH3
run
catalyst
time (h)
yield (%)
1
2a
b
3a
b
1a/NaBPh4
1b/NaBPh4
24
6
24
6
24
72
6
24
24
24
24
24
24
24
70
70
74
<5
63
75
85
85
10
80
5
1b
c
4a
b
2b
5
Ni(PPh3)3Cl
6
3a
7
[(1-Me-Ind)-Ni(PPh3)2][BPh4]b
8
2bc
2bd
2be
50
93
52
9
10
a Unless otherwise specified, the reactions were run in C6D6 at
room temperature using a Ni:NaBPh4:styrene:PhSiH3 ratio of
1:10:100:100. b Taken from ref 3. c 1000 equiv each of styrene and
PhSiH3 were used in this experiment. d A 2b:styrene:PhSiH3 ratio
of 1:200:80 was used for this run. e A 2b:styrene:PhSiH3 ratio of
1:1500:1000 was used for this run.
Scheme 5
In the present study, we selected styrene hydrosilyl-
ation as a test reaction for evaluating the catalytic
activities of the complexes 1a, 1b, and 2b.23 Initial
experiments were conducted under the same conditions
as used in most of our previous studies, namely, room-
temperature reactions employing a 1:10:100:100 ratio
of Ni:NaBPh4 (when needed):PhSiH3:styrene and run
over 6-24 h. The catalytic activity of unactivated 1b
was also examined briefly, because a recent study has
shown that some Ni-Cl precursors can promote the
hydrosilylation of styrene in the absence of NaBPh4.26
It should be noted that all of these reactions gave the
same product, namely, PhCH(SiPhH2)Me, with virtually
none of the other regioisomer, PhCH2CH2(SiPhH2),
being detected by NMR. The quantitative results of the
catalytic runs are listed in Table 4 and described below.
Runs 1-4 showed that hydrosilylation reactions
catalyzed by 2b give higher yields than those by 1a or
1b. The superior activity of 2b was especially evident
when shorter reaction times were employed or when the
neutral precursors were used in the absence of the
cationic initiator NaBPh4. This point is better illustrated
from the plots of yield vs time obtained by monitoring
the progress of the catalysis promoted by 1b and 2b
(Figure 6). As can be seen, the catalytic activity of 2b
requires little or no induction period and reaches a
plateau 2-3 h after the start.
ate could then initiate a different propagation sequence
via an insertion mechanism and terminating via a
â-hydride elimination pathway.25
Catalytic Hydrosilylation Reactions. Previous
studies in our group have shown that the IndNi(PPh3)-
Cl/NaBPh4 combination can also catalyze the addition
of PhRSiH2 (R ) Ph, Me, H) to olefins. For instance, a
system consisting of (1-Me-Ind)Ni(PPh3)Cl/NaBPh4 can
catalyze the addition of PhSiH3 to styrene to give
PhCH(Me)(SiPhH2) in up to 80 catalytic turnovers.3 We
have proposed that the mechanism of this reaction
involves the insertion of styrene into the Ni-H bond of
an intermediate to give a new Ni-alkyl species; the
latter then reacts with a second PhSiH3 to liberate the
hydrosilylation product and regenerate the Ni-H in-
termediate. The initial Ni-H species arises from a
reaction between the in-situ-formed Ni-based cations
and PhSiH3 (hydride abstraction); alternatively, it can
be generated by reacting the Ni-X precursors with
LiAlH4 (X ) Cl) or directly with PhSiH3 (X ) Me;
Scheme 5). It should be noted, however, that the
IndNi(PPh3)Cl/NaBPh4 combinations give faster rates
and higher yields.3
The reactivities of the Ni(I) species (PPh3)3NiCl and
the bis(phosphine) species 3a were also probed briefly.
The Ni(I) species was found to be a poor precatalyst for
the hydrosilylation reaction (Table 4, run 5), which
implies that even if this side-product were present in
solutions of 1a, its contribution to the overall catalytic
(25) Previous observations on the analogous reactions of ethylene
or 1-hexene with Ind-Ni-based systems have shown that these
eliminations can be facile: Wang, R.; Groux, L. F.; Zargarian, D.
Organometallics 2002, 21, 5531.

Indenyl-Nickel Complexes
Organometallics, Vol. 24, No. 16, 2005 4011
Figure 6. Kinetic profiles for the addition of PhSiH3 or
PhMeSiH2 to styrene catalyzed by complex 1b or 2b. The
experiments were carried out in C6D6 using a 1:(10:)50:50
ratio of precatalyst:(NaBPh4:)styrene:silane.
Figure 7. Kinetic plot of the addition of PhSiH3 to styrene
catalyzed by 2b. New batches of substrates (50 equiv each)
were added to the reaction mixture at the time intervals
indicated by the dashed vertical lines.
activities would be insignificant. On the other hand,
complex 3a was found to be nearly as active as 2b for
promoting the addition of PhSiH3 to styrene (compare
runs 6 and 4b). Significantly, this level of activity is
much higher than that of [(1-Me-Ind)Ni(PPh3)2]+ (run
7) and implies that the catalysis proceeds more ef-
ficiently in the presence of the pendant olefin moiety.
Since the presence of a second PPh3 ligand in the
coordination sphere of Ni would be expected to hinder
the approach of the much more weakly nucleophilic
substrates, we propose that the higher reactivity of 3a
is due to the intramolecular displacement of the second
PPh3 ligand by the pendant olefin moiety. In other
words, under the conditions of the catalysis, 3a is
effectively converted to a catalytically active species,
probably via the intermediary of 2a (eq 2).
The sharp decline in the catalytic activity of 2b is
presumably caused by a deactivation process that
depletes the active species available for the hydrosilyl-
ation catalysis. We suspected that the most likely
deactivation pathway involves possible side-reactions
between PhSiH3 and 2b (or derivatives arising from it
during the catalysis). To confirm this, we monitored the
31P{1H} NMR spectrum of a mixture of 2b (ca. 0.02 M
in C6D6) and PhSiH3 (10 equiv) in the absence of
styrene, which showed that the signal due to 2b
disappeared completely within 5 min of the mixing time
and was replaced by several new 31P resonances in the
40-43 ppm range. We have not isolated any of these
new species, but their chemical shifts are in the range
associated with complexes of the type (R-Ind)Ni(PPh3)X
(X ) H and SiPhnH3-n moieties).28 On the other hand,
monitoring the 1H NMR spectra of the reaction mixture
indicated that some or all of these new species continue
reacting with PhSiH3. These observations are consistent
with the possibility that PhSiH3 undergoes competitive
side-reactions with 2b and derivatives arising from it
during the catalysis, thereby reducing the number of
catalytically active species available for hydrosilyl-
ation.29 It is reasonable to propose that such side-
reactions would be more competitive when the concen-
tration of styrene is depleted (toward the later stages
Optimizing the Catalytic Efficiency of 2b. To
determine the highest level of hydrosilylation activity
that can be obtained from 2b, we conducted a number
of experiments using smaller catalyst loadings, which
showed that increasing the catalyst:substrate ratio leads
to lower relative yields. For example, a 0.1% Ni loading
gave only ca. 50% yield (run 8); although this corre-
sponds to a higher number of catalytic turnovers in
comparison to the runs using 1% loadings (ca. 500 vs
75-85), the diminishing relative yields imply that a
deactivation takes place beyond the initial stages of the
catalysis. To confirm this, we conducted an experiment
in which additional batches of substrates were added
to the reaction mixture after an interval of ca. 24 h and
the reaction progress was monitored continuously. The
results are plotted in Figure 7, which shows that the
catalysis becomes increasingly sluggish after the initial
phase. The extent of decline in the catalytic efficacy of
2b is more clearly evident when we compare turnover
frequency numbers at various stages during the three-
day reaction period: 1.4 mmol/h during t ) 0-5 h;
0.034 mmol/h during t ) 5-25 h; 0.005 mmol/h during
t ) 25-52 h; 0.008 mmol/h during t ) 52-80 h.27
(26) Chen, Y.; Sui-Seng, C.; Boucher, S.; Zargarian, D. Organo-
metallics 2005, 24, 149.
(27) Careful inspection of Figure 7 seems to indicate an upswing in
the pace of catalysis toward the end of the 3-day period, presumably
because of the accumulation of unreacted substrates in the reaction
medium. This result should be treated with caution, however, given
the relatively large uncertainties associated with the NMR integration
values.
(28) The Ni-H derivatives are not stable enough to be isolated, but
signals detected in the range 41-44 ppm during the decomposition of
some bulky Ni-alkyl derivatives (alkyl ) t-Bu, i-Bu, sec-Bu) have been
attributed to them (refs 9 and 12). On the other hand, the 31P{1H}
NMR (C6D6) spectrum of IndNi(PPh3)SiCl3 shows a singlet resonance
at 49 ppm: Boucher, S.; Zargarian, D. Unpublished results.
(29) On the basis of previous studies, we believe that the most likely
side-reaction is the dehydrogenative coupling of PhSiH3 to give
(PhSiH)n oligomers, but the 1H NMR spectra recorded for the 2b/
PhSiH3 mixtures do not show the characteristically broad signals
associated with these materials. Other possibilities include insertion
of the pendant olefin into a Ni-H bond or reductive elimination of
R-Ind-H to from Ni(0) species. Or see: (a) Fontaine, F.-G.; Zargarian,
D. Organometallics 2002, 21, 401. (b) Fontaine, F.-G.; Kadkhodazadeh,
T.; Zargarian, D. Chem. Commun. 1998, 1253.

4012 Organometallics, Vol. 24, No. 16, 2005
Gareau et al.
of the hydrosilylatoin catalysis) or when a greater [2b]:
[PhSiH3] ratio is used. To examine the validity of this
proposal, we carried out the following experiments.
Experimental Section
General Comments. All experiments were conducted
under an atmosphere of nitrogen using standard Schlenk
techniques and/or a glovebox. Dry, oxygen-free solvents were
used throughout. The elemental analyses were performed by
Laboratoire D’analyse EÄ le´mentaire (Universite´ de Montre´al).
The NMR spectra were recorded at ambient temperature on
an AMX400 spectrometer. The GC/MS analyses were made
on an Agilent Technologies 6890 Network GC system equipped
with an HP-5MS capillary column and a 5973 MS selective
detector. Synthetic procedures for the two indenyl ligands,
1-{CH2dCH(CH2)2}-indene and 1-{CH2dCH(CH2)SiMe2}-in-
dene, have been reported previously.6
First, we conducted catalytic reactions using an excess
of styrene to determine if this would improve the
reaction yields. Thus, a reaction using a 1:200:80 ratio
of 2b:styrene:PhSiH3 gave an improved hydrosilylation
yield of 93% (run 9), implying that the extent of
unproductive side-reactions can be minimized in the
presence of excess styrene. On the other hand, a similar
reaction using a 1:1500:1000 ratio gave only 52% yield
(run 10), which is not a significant improvement over
the 1:1000:1000 reaction (run 8), implying that the
greater silane concentration in large-scale runs coun-
teracts the beneficial effects of excess styrene.
Next, we set out to determine if using secondary
hydrosilanes instead of PhSiH3 would affect the extent
of unproductive side-reactions; since these silanes are
less reactive than PhSiH3 in dehydrocoupling reactions,
they were expected to favor the hydrosilylation cataly-
sis.30 Indeed, the addition of PhMeSiH2 or Ph2SiH2 to
styrene was found to be nearly quantitative in less than
3 h, both in small-scale (1:100:100 ratio) and large-scale
(1:1000:1000 ratio) reactions (Figure 6).31 That the
large-scale hydrosilylation of styrene proceeds more
efficiently with a secondary silane such as PhMeSiH2
(up to 1000 catalytic turnovers) than PhSiH3 (up to 500
catalytic turnovers) is consistent with the proposal that
the hydrosilylation is inhibited by unproductive side-
reactions involving PhSiH3.
(η3:η0-Ind(CH2)2CHdCH2)Ni(PPh3)Cl (1a). An Et2O solu-
tion (50 mL) containing Ind(CH2)2CHdCH2 (0.270 g,
1.53 mmol) and n-BuLi (0.67 mL of a 2.5 M solution in hexane,
1.68 mmol) was stirred overnight and then transferred (drop-
wise over 3 h) to a stirred solution of (PPh3)2NiCl2 (1.20 g,
1.83 mmol) in Et2O (25 mL). After filtration and concentration,
the product was precipitated with hexane. Washing the filtrate
with cold hexane (2 × 20 mL, 0 °C) gave 0.168 g of red solid
(21%). Single crystals suitable for X-ray analysis were obtained
by slow diffusion of hexane into a saturated Et2O solution.
1H NMR (C6D6): δ 7.81 (m, PPh3), 7.55 (m, H7), 7.27 (t, 3JH-H
) 7.5, H6), 7.01 (t, 3JH-H ) 7.4, H5), 6.57 (d, 3JH-H ) 1.9, H2),
6.29 (d, 3JH-H ) 7.4, H4), 6.11 (m, H10), 5.30 (d, 3JH-H ) 16.9,
3
H11A), 5.17 (d, JH-H ) 10.0, H11B), 3.56 (s, H3), 2.80, 2.70,
2.47, and 2.20 (m, H8A, H8B, and H9A, H9B). 13C{1H} NMR
2
(C6D6): δ 139.3 (C10), 135.1 (d, JP-C ) 11.8, o-C), 133.2
1
(d, JP-C ) 43.7, i-C), 130.9 (p-C), 130.4 (C7a), 129.1 (m-C),
127.4 (C3a), 127.1 and 126.8 (C5 and C6), 119.2 and 117.5
2
(C4 and C7), 115.7 (C11), 107.2 (d, JP-C ) 13.2, C1), 103.4
(C2), 67.4 (C3), 31.7 and 26.9 (C8 and C9). 31P{1H} NMR
(C6D6): δ 30.7. Anal. Calcd for C31H28ClNiP: C, 70.83; H, 5.37.
Found: C, 70.61; H, 5.57.
Conclusion
(η3:η0-IndSi(Me)2CH2CHdCH2)Ni(PPh3)Cl (1b). An Et2O
solution (200 mL) containing IndSi(Me)2CH2CHdCH2 (1.05 g,
4.9 mmol) and n-BuLi (2.00 mL of a 2.5 M solution in hexane,
5.0 mmol) was stirred overnight and then transferred (drop-
wise over 3 h) to a stirred solution of (PPh3)2NiCl2 (3.5 g,
5.4 mmol) in Et2O (25 mL). After filtration and concentration,
the product was precipitated with hexane and washed with
cold hexane (0 °C, 2 × 20 mL) to give 1.87 g of red solid (67%).
Single crystals suitable for X-ray analysis were obtained by
slow diffusion of hexane into a saturated Et2O solution.
The present study has shown that incorporating a
pendant olefin moiety in Ind-Ni complexes renders the
otherwise electronically and coordinatively unsaturated
cations [IndNi(PPh3)]+ stable enough to be isolated and
characterized, while the hemilabile nature of the che-
lation preserves the reactivities of these species. The
spectroscopic and solid-state characterization of 2 pro-
vides a better understanding of the Ni-Ind and Ni-
olefin interactions in such species. We have learned, for
instance, that the unsymmetrical binding of the olefin
might lead to the accumulation of positive charge on a
vinylic carbon, thereby rendering the olefin susceptible
to nucleophilic attacks; the facile polymerization of
styrene in the presence of 2b is consistent with this
hypothesis. Thus, we believe that these complexes
are good models for the proposed intermediates in
Ni-catalyzed transformations of olefins. On the other
hand, the labile binding of the pendant olefin in 2b and
3a facilitates the efficient hydrosilylation of styrene. On
the basis of our experimental observations, we have
proposed a few working hypotheses on how these
catalytic reactions proceed and what type of deactivation
pathways might be operating, but a detailed mechanis-
tic picture will require further investigations.
3
1H NMR (C6D6): δ 7.94-7.81 (m, PPh3), 7.68 (d, JH-H ) 8.1,
3
3
H7), 7.30 (t, JH-H ) 7.3, H6), 7.04 (t, JH-H ) 7.5, H5), 6.78
3
3
(d, JH-H ) 2.5, H2), 6.52 (d, JH-H ) 7.6, H4), 6.13 (m, H10),
5.20 (d, 3JH-H ) 17.0, H11A), 5.13 (d, 3JH-H ) 9.1, H11B), 3.65
3
(d, JH-H ) 3.1, H3), 2.28-2.15 (m, H9A & H9B), 0.83 (s,
Si-Me), 0.81 (s, Si-Me). 13C{1H} NMR (C6D6): δ 135.4 (C10),
134.5 (d, 2JP-C ) 11.1, o-C), 134.3 (C7a), 132.3 (d, 1JP-C ) 43.7,
3
i-C), 130.5 (p-C), 129.5 (C3a), 128.5 (d, JP-C ) 10.3, m-C),
127.1 and 126.0 (C5 and C6), 121.5 and 117.6 (C4 and C7),
2
113.6 (C11), 109.7 (C2), 93.6 (d, JP-C ) 13.9, C1), 72.9 (C3),
24.4 (C9), -2.70 and -2.93 (C8A and C8B). 31P{1H} NMR
(C6D6): δ 28.58. Anal. Calcd for C32H32ClNiPSi: C, 67.45; H,
5.66. Found: C, 67.79; H, 5.56.
[(η3:η2-Ind(CH2)2CHdCH2)Ni(PPh3)][BPh4] (2a).
A
CH2Cl2 solution of 1a (70 mg, 0.13 mmol) and NaBPh4
(320 mg, 0.93 mmol) was stirred for 2 h at room temperature
and then filtered. The product was precipitated with Et2O to
give a brown-red solid (41 mg, 45%). Microcrystals were
obtained by repeated recrystallization from CH2Cl2/Et2O.
(30) Tertiray hydrosilanes R3SiH (R ) Et, Ph, Cl) proved completely
unreactive in our systems.
3
1H NMR (CDCl3): δ 7.69 (d, JH-H ) 7.1, H7), 7.60-7.01 (m,
(31) It should be mentioned, however, that hydrosilylation of styrene
with PhMeSiH2 gives a 90:10 mixture of the regioisomers PhCH-
(SiPhMeH)CH3 and PhCH2CH2(SiPhMeH), whereas the corresponding
reaction with Ph2SiH2 gives only PhCH2CH2(Ph2SiH). In other words,
the regioselectivity of the hydrosilylation depends on the nature of R
in PhRSiH2.
PPh3, H2, H5, and H6), 6.52 (d, 3JH-H ) 7.1, H4), 5.63 (s, H3),
3
3
5.48 (m, H10), 4.05 (d, JH-H ) 16.1, H11A), 3.09 (t, JH-H
)
7.1, 11B), 2.54 and 2.05 (m, H8A, H8B and H9A, H9B).
13C{1H} NMR (CDCl3): δ 133.3 (d, JP-C ) 11.1, o-C), 131.8
2
(p-C), 129.5 (d, 3JP-C ) 13.2, m-C), 128.3 (d, 1JP-C ) 46.5, i-C),

Indenyl-Nickel Complexes
Organometallics, Vol. 24, No. 16, 2005 4013
120.3 (C4), 119.5 (C7), 114.7 (d, 2JP-C ) 10.4, C1), 112.3 (C10),
105.5 (C2), 92.2 (C3), 71.8 (C11), 36.5 and 21.9 (C8 and C9).
31P{1H} NMR (CDCl3): δ 33.66. Anal. Calcd for C31H30BF4-
NiP‚H2O: C, 62.57; H, 5.08. Found: C, 62.41; H, 5.10.
General Procedure for Carrying Out Catalytic Po-
lymerization of Styrene. Polymerization reactions were
carried out by stirring a mixture of 2b (ca. 10 mg, 0.0175 mmol)
and a large excess of styrene for a given length of time, as
follows. Run 1 was carried out in dichloroethane (8 mL) using
ca. 2000 equiv of styrene; the mixture was stirred at 60 °C for
48 h. Runs 2-5 were carried out in toluene (1 mL) or neat
styrene (ca. 1000 equiv); the mixtures were stirred for 4 h at
60 or 25 °C (see Table 3). The final reaction mixtures obtained
from runs 1, 4, and 5 were evaporated to remove the solvent
and/or unreacted styrene, followed by dissolving the solid
residues in a minimum of toluene and transferring into ethanol
to precipitate an off-white solid. The reaction mixtures in runs
2 and 3 solidified in 3-4 h; workup of these mixtures consisted
of dissolving the solid residues in a minimum of toluene and
transferring into ethanol to precipitate an off-white solid. The
solid products were subjected to NMR and GPC analyses; the
latter were done using THF as solvent and were calibrated
on the basis of poly(styrene) standards.
General Procedure for Conducting Catalytic Hydro-
silylation of Styrene. Most runs were carried out in NMR
tubes containing C6D6 (ca. 0.8 mL), 2b (ca. 10 mg, 0.0175
mmol), styrene (ca. 200 µL, 1.76 mmol), and PhSiH3 (ca.
210 µL, 1.76 mmol); NaBPh4 (ca. 55 mg, 0.17 mmol) was also
added to the reaction mixture in some runs (see Table 4). The
tubes were placed in an ultrasonic bath at room temperature
for the desired period of time. NMR spectroscopy allowed us
to identify the reaction product by comparison to authentic
samples, which were obtained from large-scale runs, distilled,
and fully characterized by NMR and GC/MS. The yields were
determined from the relative integration values for the
substrate and product protons, on the basis of a calibration
curve prepared as described previously.26 The kinetic runs
were conducted similarly, but using 50 equiv of substrates.
[(η3:η2-IndSi(Me)2allyl)Ni(PPh3)][BF4] (2b). A CH2Cl2
solution of 1 (300 mg, 0.53 mmol) and AgBF4 (113 mg,
0.58 mmol) was stirred for 1 h at room temperature in an
ultrasonic bath and then filtered. The product was precipitated
and washed with hexane to give a brown solid (165 mg, 46%).
1H NMR (CDCl3): δ 7.52-7.20 (m, PPh3, H7, H6, H5, and H2),
3
6.56 (d, JH-H ) 7.6, H4), 5.48 (s, H3), 5.40 (m, H10), 3.90
(d, 3JH-H ) 16.2, H11A), 3.18 (t, 3JH-H ) 7.1, H11B), 2.52 (dd,
3
2
2JH-H ) 6.9, JH-H ) 12.1, H9A or H9B), 1.86 (dd, JH-H
)
3
4.5, JH-H ) 12.5 H9A or H9B), 0.70 (s, Si-Me), 0.60 (s,
Si-Me). 13C{1H} NMR (CD2Cl2): δ 133.3 (d, 2JP-C ) 10.7, o-C),
2
132.1 (p-C), 131.0 and 129.8 (C5 and C6), 129.5 (d, JP-C
)
10.7, m-C), 127.8 (d, 1JP-C ) 47.0, i-C), 126.9 (C7a), 122.1 (C3a),
121.4 and 121.3 (C4 and C7), 110.5 and 110.0 (C10 and C2),
2
104.5 (d, JP-C ) 8.8, C1), 91.6 (C3), 75.0 (C11), 26.3 (C9),
-0.18 and -2.68 (C8A and C8B). 31P{1H} NMR (CDCl3):
δ 33.70. Anal. Calcd for C32H32ClNiPSiBF4: C, 61.88; H, 5.19.
Found: C, 60.84; H, 5.04.
[(η3:η0-Ind(CH2)2CHdCH2)Ni(PPh3)2][BPh4] (3a). Mul-
tiple recrystallization of 2a in hot CH2Cl2/hexane gave 3a as
a yellow solid. 1H NMR (CD2Cl2): δ 7.50-6.91 (m, PPh3, BPh4,
H5 and H6), 6.56 (d, 3JP-H ) 2.0, H2), 6.30 (d, 3JH-H ) 6.9, H4
3
or H7), 6.05 (d, JH-H ) 7.0, H4 or H7), 5.55 (m, H10), 4.98
(d, 3JP-H ) 2.4, H3), 4.86 (d, 3JH-H ) 10.1, H11B), 4.75 (d, 3JH-H
) 17.1, H11A), 2.15-2.00 (m, H9A and H9B), 1.48 and 1.26
(m, H8A and H8B). 13C{1H} NMR (CD2Cl2): δ 163.9 (4-line
multiplet, 1JB-C ) 49, i-C of BPh4), 136.5 (C10), 136.0 (m-C of
2
2
BPh4), 133.7 (d, JP-C ) 10.9, o-C of Ph3P(1)), 133.5 (d, JP-C
) 11.2, o-C of Ph3P(2)), 131.3 and 130.7 (p-C of both PPh3),
1
130.3 (d, JP-C ) 46.6, i-PPh3), 129.1 to 128.3 (m-C of both
3
PPh3, C5 and C6), 125.6 (d, JP-C ) 2.3, o-C of BPh4), 124.2
and 122.8 (C3a and C7a), 121.7 (p-C of BPh4), 119.6 and 119.4
Acknowledgment. The authors gratefully acknowl-
edge the financial support provided by NSERC of
Canada and useful discussions with Sylvain Boucher
and Michel Simard.
2
(C4 and C7), 115.8 (C11), 105.8 (d, JP-C ) 10.1, C1), 104.3
(C2), 83.1 (d, 2JP-C ) 9.2, C3), 32.5 (d, 3JP-C ) 4.5, C8), 25.83
(C9). 31P{1H} NMR (CD2Cl2): δ 35.4 and 32.2 (d, 2JP-P ) 26.7).
Anal. Calcd for C73H63BNiP2: C, 81.81; H, 5.93. Found: C,
80.56; H, 5.78.
Supporting Information Available: Complete details on
the X-ray analyses of 1a, 1b, 2a, 3a, and 4, including tables
of crystal data, collection, and refinement parameters, bond
distances and angles, anisotropic thermal parameters, and
hydrogen atom. This material is available free of charge via
the Internet at http://pubs.acs.org.
[(η3-allyl)Ni(PPh3)2][BF4] (4). Recrystallization of 2b in
CH2Cl2/Et2O gave brown crystals, which were identified as the
known complex 420 on the basis of the NMR spectra and
microanalysis. 1H NMR (C6D6): δ 7.30 (m, PPh3), 5.75 (m, H2),
3.83 (m, H syn or anti), 3.17 (m, H syn or anti). 31P{1H} NMR
(C6D6): δ 26.66. Anal. Calcd for C39H35BF4NiP2: C, 65.87; H,
4.96. Found: C, 65.64; H, 5.02.
OM050285U