Solid state structures and phosphine exchange reactions of (1-Me-Indenyl)(PR3)Ni-Cl

Article abstract of DOI:10.1021/om010608w

The complexes (1-Me-Ind)(PR₃)Ni-Cl (Ind = indenyl; R = Me (2), Cy (3), Bu (4), and (CH₂)₂- (CF₂)₅CF₃ (5)) have been prepared by the direct reaction of the corresponding (PR₃)₂NiCl₂ with Li(1-Me-Ind) or via phosphine exchange reactions with (1-Me-Ind)(PR′₃)Ni-Cl (R′ = Ph (1) and Me). Solution NMR and single-crystal X-ray diffraction studies of the structures of 2 and 3 have allowed an analysis of the mode of coordination of the 1-Me-Ind ligand both in the solid state and in solution. Kinetic studies have shown that the substitution of PPh₃ in (1-Me-Ind)(PPh₃)Ni-Cl by PCy₃ follows a second-order rate (associative mechanism) with the following kinetic parameters: K₂ (× 10^-2 M^-1 s^-1) = 1.1 ± 0.1 at 233 K, 1.9 ± 0.4 at 245 K, 15 ± 2 at 284 K, 21 ± 2 at 293 K, and 34 ± 7 at 303 K; ΔH? = 6.40 ± 0.07 kcal/mol; ΔS? = -40 ± 4 eu. The implications of these results for the mechanisms of the polymerization reactions catalyzed by these complexes have been discussed.

Full text of DOI:10.1021/om010608w

5156
Organometallics 2001, 20, 5156-5161
Solid Sta te Str u ctu r es a n d P h osp h in e Exch a n ge
Rea ction s of (1-Me-In d en yl)(P R3)Ni-Cl
Fre´de´ric-Georges Fontaine, Marc-Andre´ Dubois, and Davit Zargarian*
De´partement de Chimie, Universite´ de Montre´al, Montre´al (Que´bec), Canada H3C 3J 7
Received J uly 9, 2001
The complexes (1-Me-Ind)(PR3)Ni-Cl (Ind ) indenyl; R ) Me (2), Cy (3), Bu (4), and (CH2)2-
(CF2)5CF3 (5)) have been prepared by the direct reaction of the corresponding (PR3)2NiCl2
with Li(1-Me-Ind) or via phosphine exchange reactions with (1-Me-Ind)(PR′3)Ni-Cl (R′ )
Ph (1) and Me). Solution NMR and single-crystal X-ray diffraction studies of the structures
of 2 and 3 have allowed an analysis of the mode of coordination of the 1-Me-Ind ligand both
in the solid state and in solution. Kinetic studies have shown that the substitution of PPh3
in (1-Me-Ind)(PPh3)Ni-Cl by PCy3 follows a second-order rate (associative mechanism) with
the following kinetic parameters: K2 (× 10-2 M-1 s-1) ) 1.1 ( 0.1 at 233 K, 1.9 ( 0.4 at 245
K, 15 ( 2 at 284 K, 21 ( 2 at 293 K, and 34 ( 7 at 303 K; ∆Hq ) 6.40 ( 0.07 kcal/mol; ∆Sq
) -40 ( 4 eu. The implications of these results for the mechanisms of the polymerization
reactions catalyzed by these complexes have been discussed.
In tr od u ction
however, the PR3 ligand in our complexes remains
coordinated to the Ni center during the catalysis.
Indeed, higher catalytic activities were displayed by
complexes bearing more strongly coordinating phos-
phines.7
Recent reports have documented the emergence of a
number of Ni(II) compounds that act as effective pre-
catalysts for the polymerization of ethylene and other
R-olefins.1 Notable among these are the complexes (η2-
(N,O)-salicylaldimine)Ni(PPh3)(Ar) and (R-diimine)NiX2
(X ) Br, Me). Brookhart et al. have shown2 that the
polymerization catalysis in the latter system proceeds
via a cationic mechanism, whereas Grubbs et al. have
shown3 that the salicylaldimine system follows a non-
cationic pathway involving the predissociation of PPh3.
In this context, we have reported that the complexes
(Ind′)(PR3)Ni-X (Ind′ ) indenyl and its substituted
derivatives; R ) Ph, Me, Cy; X ) Cl, Me, CC-Ph)4 act
as precatalysts in a number of oligomerization5 and
polymerization6 reactions, including the polymerization
of ethylene.6b The results of mechanistic experiments
have led us to conclude that the polymerization of
ethylene in this system involves noncationic intermedi-
ates as in the Grubbs system; unlike the Grubbs system,
As a follow-up to the polymerization studies men-
tioned above, we have studied the structural features
of some of these precatalysts with a view to finding
structure-activity relationships; in addition, we have
examined the phosphine exchange reactions in these
complexes in order to probe the likelihood of phosphine
dissociation reactions similar to those reported for the
Grubbs systems. The present paper reports the prepa-
ration and characterization of the complexes (1-Me-Ind)-
(PR3)Ni-Cl (R ) Me (2), Cy (3), Bu (4), and (CH2)2-
(CF2)5CF3 (5)), including the solid state structures of the
PMe3 and PCy3 derivatives, and describes the results
of a kinetic study on the ligand exchange reaction
between (1-Me-Ind)(PPh3)Ni-Cl, 1, and PCy3.
Exp er im en ta l Section
(1) Bauers, F. M.; Mecking, S. Macromolecules 2001, 34, 1165, and
references therein.
Gen er a l Com m en ts. All manipulations were performed
under an inert atmosphere of N2 or argon using standard
Schlenk techniques and a drybox. Dry, oxygen-free solvents
were employed throughout. The synthesis of (1-Me-Ind)(PPh3)-
Ni-Cl, 1, has been reported previously;4b (PMe3)2NiCl2 was
prepared from NiCl2 and PMe3. The reagent P{(CH2)2(CF2)5-
CF3}3 was donated by Dr. I. Horvath; all other reagents used
in the experiments were obtained from commercial sources and
used as received. The elemental analyses were performed by
the Laboratoire d’Analyse EÄ le´mentaire (Universite´ de Mont-
re´al). The NMR spectra were recorded using the following
spectrometers: Bruker AMX400 (1H at 400 MHz, 13C{1H} at
100.56 MHz, and 31P{1H} at 161.92 MHz) and Bruker AV300
(1H at 300 MHz).
(2) (a) J ohnson, L. K.; Killian, C. M.; Brookhart, M. J . Am. Chem.
Soc. 1995, 117, 6414. (b) J ohnson, L. K.; Killian, C. M.; Arthur, S. D.;
Feldman, J .; McCord, E.; McLain, S. J .; Kreutzer, K. A.; Bennett, M.
A.; Coughlin, E. B.; Ittel, S. D.; Parthasarathy, A.; Tempel, D.;
Brookhart, M. (UNC-Chapel Hill/DuPont) PCT Int. Patent WO 96/
23010, 1996; Chem. Abstr. 1996, 125, 222773t. (c) Svejda, S. A.;
J ohnson, L. K.; Brookhart, M. J . Am. Chem. Soc. 1999, 121, 10634,
and references therein.
(3) Younkin, T. R.; Connor, E. F.; Henderson, J . I.; Friedrich, S. K.;
Grubbs, R. H.; Bansleben, D. A. Science 2000, 287, 460, and references
therein.
(4) (a) Dubuc, I.; Dubois, M.-A.; Be´langer-Garie´py, F.; Zargarian,
D. Organometallics 1999, 18, 30. (b) Huber, T. A.; Bayrakdarian, M.;
Dion, S.; Dubuc, I.; Be´langer-Garie´py, F.; Zargarian, D. Organometal-
lics 1997, 16, 5811. (c) Bayrakdarian, M.; Davis, M. J .; Reber, C.;
Zargarian, D. Can. J . Chem. 1996, 74, 2194. (d) Huber, T. A.; Be´langer-
Garie´py, F.; Zargarian, D. Organometallics 1995, 14, 4997.
(5) Fontaine, F.-G.; Kadkhodazadeh, T.; Zargarian, D. J . Chem. Soc.,
Chem. Commun. 1998, 1253.
(1-Me-In d )(P Me3)Ni-Cl (2). A THF solution (130 mL) of
Li(1-Me-Ind) (477 mg, 3.5 mmol) was added dropwise to a
(6) (a) Wang, R.; Be´langer-Garie´py, F.; Zargarian, D. Organometal-
lics 1999, 18, 5548. (b) Dubois, M.-A.; Wang, R.; Zargarian, D.; Tian,
J .; Vollmerhaus, R.; Li, Z.; Collins, S. Organometallics 2001, 20, 663.
(7) For example, the PMe3 and PCy3 derivatives are almost an order
of magnitude more active in the polymerization of ethylene in
comparison to the PPh3 analogue (ref 6b).
10.1021/om010608w CCC: $20.00 © 2001 American Chemical Society
Publication on Web 10/30/2001

Reactions of (1-Me-Indenyl)(PR3)Ni-Cl
Organometallics, Vol. 20, No. 24, 2001 5157
stirring THF solution (40 mL) of Ni(PMe3)2Cl2 (1.0 g, 3.5 mmol)
which was kept at 60 °C. During the addition, the reaction
vessel was placed under dynamic vacuum for a few seconds
every 10 min in order to remove the free PMe3 liberated by
the reaction. The final mixture was stirred for 30 min and then
evaporated to give a red powder (940 mg, crude yield 90%).
Recrystallization from Et2O/hexanes gave 420 mg of dark red
-123.43 (br, 2F), -124.46 (br, 4F), -127.72 (br, 2F). 31P{1H}
NMR (CDCl3): δ 22.08 (s). Anal. Calcd for C34H21F39Ni1P1Cl1:
C, 31.53; H, 1.64. Found: C, 31.89; H, 1.66.
Cr ysta l Str u ctu r e Deter m in a tion s. A dark red crystal
of 2 was attached to a glass fiber and transferred rapidly and
under a cold stream of nitrogen to an Enraf-Nonius CAD-4
diffractometer equipped with a low-temperature gas stream
cryostat for data collection at 220(2) K. The data were collected
with graphite-monochromated Cu KR radiation; the refine-
ment of the cell parameters was done using CAD-4 software8
on 25 reflections, while NRC-2 and NRC-2A were used for the
data reduction.9 The crystal data for 3 was collected on a
Bruker AXS SMART 2K diffractometer using graphite-mono-
chromated Cu KR radiation at 293(2) K (SMART10 software).
Cell refinement and data reduction were carried out using
SAINT.11 Both structures were solved by direct methods using
SHELXS9612 for 2 and SHELXS9713 for 3, and the refinements
were done on F2 by full-matrix least squares. All non-hydrogen
atoms were refined anisotropically, while the hydrogens
(isotropic) were constrained to the parent atom using a riding
model. The procedure used to treat the twinning detected in
the crystal of 2 is described in the Supporting Information. In
the case of the noncentrosymmetric structure of 2, the overall
absolute structure was determined by the Flack parameter of
0.0(8).14 Crystal data and experimental details for both com-
pounds are listed in Table 1, and selected bond distances and
angles are listed in Table 2.
Kin etic Stu d ies. The rate of the ligand exchange reaction
between (1-Me-Ind)(PPh3)Ni-Cl, 1, and PCy3 was studied
under pseudo-first-order conditions using a large excess of
PCy3. The low-temperature data were obtained using NMR
spectroscopy. Toluene-d8 solutions containing accurately mea-
sured quantities of the reactants and Ph3PdO (as internal
reference) were placed in high-precision NMR tubes, and the
samples were cooled to the required temperature (245 and 233
K) inside the spectrometer probe. The gradually decreasing
intensity of the 31P{1H} signal for 1 was monitored vs time,
and the values of [1] were determined relative to [Ph3PdO].
Plots of ln([1]/[1]0) vs time gave straight lines passing through
the origin with slopes of kobs, while plots of the kobs values vs
different [PCy3]0 (10-, 15-, and 20-fold excess) gave straight
lines with slopes of K2. The rates of the exchange reaction at
high temperatures were determined by UV-vis spectroscopy.
(The electronic spectra of this family of complexes have been
reported previously.4c) Thus, the time course of the absorbance
was monitored at 412 nm, and the data were subjected to a
least-squares analysis in order to obtain the pseudo-first-order
rate constants, kobs. Plotting the kobs values vs different [PCy3]0
(10-, 15-, 20-, and 25-fold excess; [1]0 ) 1.1 × 10-3 M) gave
straight lines with slopes of K2. The activation parameters ∆Hq
and ∆Sq were obtained from a plot of ln(K2/T) vs 1/T (Figure
3) for the combined NMR and UV-vis data. The raw data for
the kinetic experiments and the kobs plots are given in the
Supporting Information.
1
3
crystals (40% yield). H NMR (C6D6): δ 7.05 (d, J H-H ) 7.6,
3
3
H7), 6.98 (t, J H-H ) 7.2, H5 or H6), 6.90 (t, J H-H ) 7.3, H5
3
3
or H6), 6.55 (d, J H-H ) 7.3, H4), 6.11 (d, J H-H ) 2.6, H2),
4
2
3.58 (br, H3), 1.61 (d, J H-P ) 2.6, Ind-CH3), 0.71 (d, J H-P
)
2.6, P-CH3). 13C{1H} NMR (C6D6): δ 130.21 (d, J C-P ) 1.4,
2
C7a), 126.29 (C3a), 125.75, 125.47, 118.15, 116.05 (C4, C5, C6,
2
C7), 103.77 (s, C2), 101.29 (d, J C-P ) 16.6, C1), 59.53 (C3),
14.68 (d, 1J C-P ) 28.4, P-CH3), 11.97 (d, 3J C-P ) 2.8, Ind-CH3).
31P{1H} NMR (C6D6): δ -10.61 (s). Anal. Calcd for C13H18Ni1P1-
Cl1: C, 52.15; H, 6.06. Found: C, 51.70; H, 6.24.
(1-Me-In d )Ni(P Cy3)Cl (3). A solution of (1-Me-Ind)(PMe3)-
Ni-Cl (400 mg, 1.34 mmol) and PCy3 (376 mg, 1.34 mmol) in
Et2O (40 mL) was stirred for 2 h. Evaporation of the solvent
followed by an EtOH washing gave a red powder (400 mg,
59%); single crystals suitable for X-ray studies and elemental
analysis were obtained by recrystallization from cold hexanes.
1H NMR (C6D6): δ 7.01, 6.86 (br s, H4/H7, H5/H6), 6.33 (s,
H2), 4.11 (s, H3), 1.94-1.10 (m, PCy3); the signal for Ind-CH3
was obscured by PCy3 signals. 13C{1H} NMR (C6D6): δ 129.72
2
(d, J C-P ) 2.8, C7a), 127.34 (C3a), 123.10, 121.67, 115.45,
2
115.33 (C4, C5, C6, C7), 101.30 (C2), 97.10 (d, J C-P ) 12.5,
C1), 55.42 (C3), 32.07 (d, 1J C-P ) 19.4, Cipso), 27.13 (d, 2J C-P
)
)
3
3
10.4, 2 Cortho), 24.82 (d, J C-P ) 4.9, Cmeta), 24.71 (d, J C-P
3
4.2, Cmeta), 23.65 (s, Cpara) 9.34 (d, J C-P ) 2.8, Ind-CH3). 31P-
{1H} NMR (C6D6): δ 37.17 (s). Anal. Calcd for C28H42Ni1P1-
Cl1: C, 66.76; H, 8.40. Found: C, 66.35; H, 8.89.
(1-Me-In d )Ni(P Bu 3)Cl (4). Neat PBu3 (165 µL, 0.66 mmol)
was added via a syringe to the red solution of (1-Me-Ind)-
(PMe3)Ni-Cl (200 mg, 0.66 mmol) in Et2O (40 mL). The
resulting mixture was stirred for 2 h with periodic venting to
remove the PMe3 generated in the reaction. Evaporation of
the solvent and extraction of the residue with hexanes (10 mL)
afforded, after removal of the volatiles, a black oil (236 mg,
84% crude yield) which contained some PBu3 and other
unidentified impurities. Further attempts to purify this oil by
recrystallization were unsuccessful. 1H NMR (C6D6): δ 7.05
3
3
(d, J H-H ) 7.6, H7), 7.00 (t, J H-H ) 7.2, H5 or H6), 6.88 (d,
3J H-H ) 7.2, H5 or H6), 6.24 (d, J H-H ) 2.5, H2), 3.75 (dd,
3
3J H-P ) 4.0, J H-H ) 2.8, H3), 1.60 (d, J H-P ) 5.4, Ind-CH3),
1.45-1.21 (m, P(CH2)3), 0.84 (t, 3J H-H ) 7.1, P(CH2)3CH3). 13C-
{1H} NMR (C6D6): δ 130.60 (C7a), 126.95 (C3a), 125.53,
125.36, 118.25, 116.43 (C4, C5, C6, C7), 103.93 (C2), 101.11
3
4
2
(d, J C-P ) 13.2, C1), 58.72 (C3), 26.42 (CH2CH3), 24.54 (d,
2J C-P ) 13.2, P-CHCH2), 24.03 (1J C-P ) 24.9, P-CH2), 14.01
(CH2CH3). 31P{1H} NMR (C6D6): δ 15.09 (s).
(1-Me-In d )Ni{P [(CH2)2(CF 2)5CF 3]3}Cl (5). A solution of
(1-Me-Ind)(PPh3)Ni-Cl (485 mg, 1.0 mmol) and P{(CH2)2-
(CF2)5CF3}3 (1.13 g, 1.05 mmol) in Et2O (30 mL) was stirred
for 2 h. The solvent was evaporated and the residue crystal-
lized from a cold Et2O/hexanes solution. A dark orange solid
was obtained (620 mg, 48%). 1H NMR (CDCl3): δ 7.18 (d, 3J H-H
Resu lts a n d Discu ssion
The complex (1-Me-Ind)(PMe3)Ni-Cl, 2, was prepared
in ca. 40% yield by the direct reaction of (PMe3)2NiCl2
3
3
) 7.8, H7), 7.14 (t, J H-H ) 7.5, H5 or H6), 7.04 (t, J H-H
)
3
3
7.4, H5 or H6), 6.83 (d, J H-H ) 7.6, H4), 6.34 (d, J H-H ) 2.6,
H2), 4.25 (pt, J H-H
)
3J H-P ) 3.4, H3), 2.19, and 1.89 (m,
4
3
(8) CAD-4 Software, version 5.0; Enraf-Nonius: Delft, The Neth-
erlands, 1989.
(9) Gabe, E. J .; Le Page, Y.; Charlant, J .-P.; Lee, F. L.; White, P. S.
J . Appl. Crystallogr. 1989, 22, 384.
(10) SMART, Release 5.059; Bruker Molecular Analysis Research
Tool; Bruker AXS Inc.: Madison, WI 53719-1173, 1999.
(11) SAINT, Release 6.06; Integration Software for Single Crystal
Data; Bruker AXS Inc.: Madison, WI 53719-1173, 1999.
(12) Sheldrick, G. M. SHELXL96, Program for the Solution of
Crystal Structures; University of Gottingen: Germany, 1990.
(13) Sheldrick, G. M. SHELXS, Program for the Solution of Crystal
Structures; University of Goettingen: Germany, 1997.
(14) Flack, H. D. Acta Crystallogr. 1983, A39, 876.
P-CH2CH2), 1.53 (d, J H-P ) 6.0, Ind-CH3). 13C{1H} NMR
(CDCl3): δ 129.11 (d, 2J C-P ) 2.1, C7a), 127.10, 127.04, 118.73,
115.98 (C4, C5, C6, C7), 125.59 (C3a), 105.54 (d, 2J C-P ) 13.2,
C1), 103.93 (C2), 60.91 (C3), 26.31 (t, 2J C-F ) 22.4, P-CH2CH2),
15.53 (d, 1J C-P ) 24.2, P-CH2), 11.98 (d, 3J C-P ) 3.5, Ind-CH3);
the CF2 were not detected presumably because of (a) the
inherently weaker intensities of signals due to quaternary C
1
2
nuclei and (b) the extended J C-F and J C-F couplings arising
from the diastereotopic fluorines. 19F{1H} NMR (CDCl3): δ
-82.33 (t,3J F-F ) 8.9, CF3), -116.05 (q, 3J F-F ) 15.2, CF2CF3),

5158 Organometallics, Vol. 20, No. 24, 2001
Fontaine et al.
Ta ble 1. Cr ysta l Da ta , Da ta Collection , a n d
Str u ctu r e Refin em en t of 2 a n d 3
2
3
formula
mol wt
cryst color
cryst habit
C
13H18ClNiP
C28H42ClPNi
503.746
dark red
block
299.40
dark red
block
cryst dimens, mm 0.53 × 0.27 × 0.19 0.14 × 0.12 × 0.06
symmetry
space group
a, Å
b, Å
c, Å
orthorhombic
P212121
7.783(2)
11.904(7)
15.041(4)
90
monoclinic
P21/c
13.4028(4)
39.7337(12)
10.3414(3)
90
101.8960(10)
90
5389.0(3)
8
R, deg
â, deg
90
90
γ, deg
volume, Å3
Z
1393.5(10)
4
1.427
D(calcd), g cm-1
diffractometer
1.2418
Nonius CAD-4
Bruker AXS
SMART 2K
298(2)
temp, K
220(2)
F igu r e 1. ORTEP diagram for complex 2 showing 30%
probability thermal ellipsoids and the atom-numbering
scheme.
λ(Cu KR), Å
µ, mm-1
scan type
θmax
1.54056
4.610
ω/2θ scan
69.82°
-9 e h e 9
-14 e k e 14
-18 e l e 18
2650
1.54178
2.587
Ω scan
73.19°
-16 e h e 16
-49 e k e 48
-12 e l e 10
10 606
h,k,l range
no. of reflns used
(I > 2σ(I))
abs corr
T (max, min)
R[F2 > 2σ(F2)],
wR(F2)
integration
0.48 and 0.24
0.0194, 0.0493
multiscan SADABS12
0.8900 and 0.6983
0.0630, 0.1474
GOF
1.030
1.007
Ta ble 2. Selected Bon d Dista n ces (Å) a n d An gles
(d eg) for 1-3
1
2
3a
3b
Ni-P
2.1782(11) 2.1586(7)
2.1865(10) 2.1898(7)
2.1875(10) 2.1897(10)
2.2114(10) 2.1884(10)
2.1316(18) 2.142(4) 2.146(3)
Ni-Cl
Ni-C(1)
Ni-C(2)
Ni-C(3)
Ni-C(3A)
Ni-C(7A)
C(1)-C(2)
C(2)-C(3)
C(3)-C(3A)
2.137(2)
2.072(2)
2.026(3)
2.308(2)
2.351(2)
1.403(4)
1.421(4)
1.451(4)
2.0511(19) 2.042(4)
2.042(3)
2.037(3)
2.392(3)
2.411(3)
1.403(5)
1.409(5)
1.459(5)
1.417(5)
1.456(5)
0.30
2.0246(18) 2.039(4)
2.3363(19) 2.388(4)
2.3608(19) 2.403(4)
1.398(3)
1.423(3)
1.461(3)
1.418(3)
1.459(3)
0.27
1.395(5)
1.424(5)
1.458(5)
1.424(5)
1.443(5)
0.30
C(3A)-C(7A) 1.417(3)
C(1)-C(7A)
1.458(3)
0.25
F igu r e 2. ORTEP diagram for complex 3 showing 30%
probability thermal ellipsoids and the atom-numbering
scheme. Only one of the two independent molecules present
in the unit cell is shown.
∆M-Ca
P-Ni-Cl
98.82(4)
161.51(8)
99.33(8)
95.48(8)
165.62(7)
95.89(2)
164.87(6)
99.18(6)
98.21(6)
164.81(6)
66.93(8)
94.22(4)
96.43(4)
C(3)-Ni- Cl
C(3)-Ni-P
C(1)-Ni-Cl
C(1)-Ni-P
161.29(11) 160.92(11)
104.39(11) 102.57(10)
95.56(10)
168.83(11) 168.26(10)
66.10(14) 66.11(14)
94.83(10)
and H5/H6, respectively; a doublet at ca. 6.1 ppm due
to H2; a broad singlet at 3.58 ppm due to H3; and two
doublets at 1.61 and 0.7 ppm attributed to Me-Ind and
C(1)-Ni-C(3) 66.50(10)
a
See ref 18 for the definition of ∆M-C and literature values
for other indenyl complexes.
(15) (a) The free PMe3 that is liberated in this reaction can
complicate the purification process and reduce the yield if measures
are not taken to remove it promptly from the reaction medium. Tests
have shown that isolated samples of 2 react with PMe3 to give (PMe3)2-
NiCl2, Ni(PMe3)4,15b and (1-Me-Ind)2.15c We suspect that the coordina-
tion of one or more molecules of PMe3 to the Ni center in 2 leads to
the formation of unstable η1-Ιnd′ species, which decompose to (1-Me-
Ind)2 and (PMe3)3NiCl; the latter then disproportionates to Ni(PMe3)4
and (PMe3)2NiCl2. (b) The compounds Ni(PMe3)4 and (PMe3)2NiCl2 were
identified on the basis of their singlet resonances in the 31P{1H} NMR
(C6D6) spectrum (at ca. -21.3 and -22. 3 ppm for Ni(PMe3)4 and
(PMe3)2NiCl2, respectively); in the 1H NMR spectrum, the complexes
can be distinguished on the basis of the doublet resonances for the
PMe3 ligands (at 1.15 and 0.93 ppm for Ni(PMe3)4 and (PMe3)2NiCl2,
respectively): Tolman, C. A.; Seidel, W. C.; Gosser, L. W. J . Am. Chem.
Soc. 1974, 96, 53. (c) (1-Me-Ind)2 was identified by comparing its 1H
NMR spectrum to the literature reports: Nicolet, P.; Sanchez, J .-Y.;
Benaboura, A.; Abadie, M. J . M. Synthesis 1987, 202.
and Li(1-Me-Ind),15 while compounds 3-5 were pre-
pared by the ligand exchange reactions between the
appropriate phosphine ligand and 1 or 2 (Scheme 1).
The characterization of complexes 2-5 by NMR spec-
troscopy was quite straightforward because all com-
plexes showed a singlet resonance in their 31P{1H} NMR
spectra, while their 1H NMR spectra matched those
obtained for other members of this family of complexes.4
For example, the 31P{1H} NMR spectrum of 2 showed
1
a singlet at ca. -10 ppm, while its H NMR spectrum
contained the following signals: two doublets and two
triplets between 6.5 and 7.0 ppm attributed to H4/H7

Reactions of (1-Me-Indenyl)(PR3)Ni-Cl
Organometallics, Vol. 20, No. 24, 2001 5159
Ch a r t 1
7a) - M-Cav (for C1,3)} is often used to measure the
extent to which Ind ligands have slipped away from the
ideal η5 hapticity wherein ∆M-C should be nearly
zero.18 The ∆Ni-C values for 3 (0.30 Å) and 2 (0.27 Å)
are somewhat larger than those observed for complex
1 (0.25 Å), reflecting the stronger nucleophilicity of the
tri(alkyl) phosphines.
Another commonly observed feature in these com-
plexes is the unequal Ni-C1 and Ni-C3 bond distances
(Ni-C1 > Ni-C3 in 2 and 3, ∆Ni-C is ca. 25-50 σ);
this type of “sideways slippage” can be attributed to the
unequal trans influences of the PR3 and Cl ligands.
Therefore, we propose that the 1-Me-Ind ligand in
complexes 1-3 adopts solid state hapticities that can
best be denoted as η5 T η3 T η1,η2 (Chart 1). Clark et
al. have put forth a similar mode of bonding, namely,
η3 T η1,η2, for the allyl ligand in the complexes (η-C3H5)-
Pt(PR3)X on the basis of solution NMR data.19
F igu r e 3. Eyring plot for the reaction of 1 with PCy3. The
K2 values (× 10-2 M-1 s-1): 1.1 ( 0.1 at 233 K; 1.9 ( 0.4
at 245 K; 15 ( 2 at 284 K; 21 ( 2 at 293 K; and 34 ( 7 at
303 K. The activation parameters: ∆Hq ) 6.40 ( 0.07 kcal/
mol; ∆Sq ) -40 ( 4 eu.
Sch em e 1
PMe3, respectively. In addition, complex 5 displayed 19F-
{1H} signals corresponding to the (CF2)5CF3 moieties.
Solid Sta te a n d Solu tion Ha p ticities of th e 1-Me-
In d Liga n d . The solid state structures of complexes 2
and 3 have been studied and compared to that of the
previously reported PPh3 analogue 1 with a view to
evaluating the influence of the phosphine ligand on the
structures of these complexes. The ORTEP diagrams for
2 and 3 are shown in Figures 1 and 2, while the crystal
data and selected bond distances and angles are pre-
sented in Tables 1 and 2, respectively. The unit cell of
complex 3 contains two independent molecules related
to each other by the orientation of the Cy groups in the
PCy3 ligand.16 The main features of these structures are
described below.
If the 1-Me-Ind ligand in complexes 2 and 3 can be
considered to be occupying two coordination sites (vide
infra), the geometry around the Ni center is approxi-
mately square planar, the largest distortion arising from
the small C1-Ni-C3 angle of ca. 66°. The Ni-P
distance is longer in complex 3 (av 2.188 Å) relative to
2 (2.159 Å; ∆(Ni-P) ≈ 30 σ) presumably because of the
much larger cone angle of PCy3. The 1-Me-Ind ligand
in both 2 and 3 is bound to the Ni center primarily
through the C1, C2, and C3 atoms (Ni-C(1-3) ) 2.02-
2.15 Å), whereas the Ni-C3a and Ni-C7a distances are
significantly longer (2.33-2.41 Å). This “slippage” away
from an ideal η5 hapticity is commonly observed in this
family of complexes4 and can be attributed to the
tendency of a Ni(II) center to avoid forming 18-electron
complexes.17 The parameter ∆M-C ) {M-Cav (for C3a,-
The solid state data described above suggest that
complexes 1-3 have very similar structures in the solid
state, and it is unclear whether the subtle structural
differences observed (e.g., different ∆Ni-C values)
might be responsible for the differences in the catalytic
activities of complexes 2 and 3 compared to 1.7 Given
that the solution structures of complexes 1-3 are more
likely to affect their reactivities, we have compared
below some of the structural features of these complexes
as gleaned from their solution 13C{1H} NMR spectra.
The solution hapticity of an indenyl ligand in a given
complex may be estimated by using the equation ∆δavC
) δav(C3a, 7a of η-Ind) - δav(C3a, 7a of Na+ Ind-) to
compare the 13C{1H} NMR chemical shifts of the C3a
and C7a nuclei in that complex to the corresponding
shifts in Na+Ind-, for which the Ind hapticity is
considered to be intermediate between η5 and η3.18a
According to this protocol, the solution hapticity of the
Ind ligand is thought to be closer to η5 when ∆δavC , 0
(e.g., ∆δavC for Ind2Fe is ca. -42 ppm),18a closer to η3
when ∆δavC . 0 (e.g., ∆δavC for (Ind)Ir(PMe2Ph)3 is ca.
+28 ppm),18a or intermediate when ∆δavC ≈ 0 (e.g.,
∆δavC for (Ind)2Ni is ca. +5 ppm).18b The ∆δavC values
for complexes 1-5 are ca. -2 ppm for 1, 2, and 4 and
ca. -1 ppm for 3 and 5, implying an intermediate (η5 T
η3) hapticity for the 1-Me-Ind ligand in all of these
complexes. The similarity of the ∆δavC values indicates
that the solution hapticity of the 1-Me-Ind ligands in
these complexes is also very similar; therefore, struc-
tural factors alone cannot explain the differences in the
(18) For a discussion of the correlations between the solid state and
solution hapticities for indenyl complexes, on one hand, and the ∆M-C
values and the 13C NMR data, on the other, see: (a) Baker, R. T.; Tulip,
T. H. Organometallics 1986, 5, 839. (b) Westcott, S. A.; Kakkar, A. K.;
Stringer, G.; Taylor, N. J .; Marder, T. B J . Organomet. Chem. 1990,
394, 777.
(16) We thank a reviewer for pointing out that a similar situation
exists in the solid state structure of complex (1-Me-Ind)Rh(η2-C2H4)-
(PCy3): Westcott, S. A.; Kakkar, A. K.; Taylor, N. J .; Roe, D. C.; Marder,
T. B. Can. J . Chem. 1999, 77, 205.
(17) For a detailed discussion of hapticity issues in the corresponding
CpNi complexes see: Holland, P. L.; Smith, M. E.; Andersen, R. A.;
Bergman, R. G. J . Am. Chem. Soc. 1997, 119, 12815.
(19) Clark, H. C.; Hampden-Smith, M. J .; Ruegger, H. Organome-
tallics 1988, 7, 2085.

5160 Organometallics, Vol. 20, No. 24, 2001
Fontaine et al.
Sch em e 2
observed reactivities between complexes 2 and 3 on one
hand and 1 on the other.
Liga n d Su bstitu tion Rea ction s. The results of the
ethylene polymerization reactions catalyzed by com-
plexes 1-36b raised the question of whether the phos-
phine ligands in these complexes dissociate prior to the
coordination of the incoming substrate molecule, as is
thought to be the case with the Grubbs system.3 Since
most square planar d8 complexes follow an associative
mechanism in ligand exchange reactions, the predisso-
ciation of the phosphine ligand in complexes such as
1-5 would seem unlikely; it should be noted, however,
that these compounds might, in principle, undergo a
dissociative substitution owing to the flexible hapticity
of the Ind′ ligand, which can adjust its electronic
contribution to the Ni center by slipping back and forth
between the η3 and η5 modes. This possibility prompted
us to study the mechanism of the phosphine substitution
reaction in the title complexes in order to determine
whether phosphine dissociation and/or hapticity changes
are involved in these reactions.
was carried out with the less bulky Cp analogue CpNi-
(PPh3)Cl, we observed that the substitution of PPh3 by
PCy3 was essentially over in the time of mixing even at
233 K, i.e., the exchange rate is 10-100 times more
rapid for the Cp analogue. The latter result not only
confirms the importance of the steric factors but also
implies that hapticity changes cannot play an important
role in these ligand exchange reactions since Cp com-
plexes would be expected to be less apt to undergo such
slippage than their Ind′ analogues. We conclude, there-
fore, that the ligand exchange reaction with complex 1
follows an associative mechanism and likely involves
the formally five-coordinate, 18-electron21 intermediate
or activated complex I shown in Scheme 2.
It is noteworthy that five-coordinate intermediates of
the type (π-allyl)NiL2R have also been postulated in a
number of other catalytic reactions,22 while many five-
coordinate d8 compounds have been shown to be rela-
tively stable species.23 Although the isolation of formally
five-coordinate, 18-electron complexes such as I has
eluded us so far, we have obtained spectroscopic evi-
dence for the existence (in solution) of a closely related
analogue wherein the incoming L ligand is tethered
The choice of a specific substitution reaction for the
kinetic studies was made on the basis of the following
considerations. The reaction of 1 with PRF was unsuit-
3
able due to the limited solubilities of both PRF and
3
complex 5. The rates of the substitution reactions
between complex 2 and the other phosphines were too
slow for studying, while those for the reactions of 1 with
PMe3 or PBu3 were too rapid for convenient monitoring.
The reaction of 1 with PMe3 was doubly unsuitable for
our purposes because the product of the reaction,
complex 2, reacts further with PMe3, which is present
in excess, to give a number of decomposition products.15
In contrast, the reaction of 1 with PCy3 showed a rate
profile that was convenient to monitor and did not lead
to secondary products; therefore, this reaction was
chosen for our kinetic studies.
(21) We propose that the 1-Me-Ind ligand occupies two coordination
sites in the intermediate I as opposed to slipping into an η1 mode,
because the formation of η1-Ind′ derivatives in this family of compounds
normally leads to decomposition involving the coupling of the Ind′
ligands (refs 4d and 15). Moreover, if slippage into an η1 mode were
required, the ligand exchange reaction would be expected to proceed
more slowly for the Cp analogue, whereas the opposite was observed.
However, a reviewer has pointed out that this conclusion may not be
warranted in view of the fact that the CpNiLn systems do not display
the same type of hapticity changes as the analogous Ind complexes.
For example, the paramagnetic complex Cp*Ni(acac)(PMe3), which
would be formally considered a 20-electron complex, does not show
any “slippage” but rather an overall lengthening of the Ni-C(Cp*)
bonds (Smith, M. E.; Andersen, R. A. J . Am. Chem. Soc. 1996, 118,
11119). A similar behavior is seen when comparing the structural
features of Cp2M vs Ind2M complexes (M ) Fe, Co, Ni): the Cp
complexes retain the 5-fold axis as the electron count goes from 18 to
19 to 20, but the M-C distances increase; on the other hand, the
corresponding Ind analogues undergo slip-fold distortions whose
degrees increase with the increasing electron count (ref 18b). Therefore,
in terms of their coordination to Ni, the Cp and Ind ligands seem to
respond differently to changes in the electronic environment around
Ni, so that it may be possible for the Cp and Ind complexes to both
undergo associative ligand exchange. In the Cp case, the association
of the incoming PCy3 would result in longer Ni-Cp distances without
causing hapticity changes, whereas in the Ind case there would be
subtle hapticity changes. We thank this reviewer for bringing this
different interpretation to our attention.
The pseudo-first-order rate of the reaction of 1 with
PCy3 (10-20-fold excess) at low temperatures was
studied by the 31P{1H} NMR monitoring of the disap-
pearance of 1 relative to OdPPh3.20 Plots of ln([1]/[1]0)
vs time gave straight lines passing through the origin
with slopes of kobs, while plots of the kobs values vs
different [PCy3]0 values (10-, 15-, and 20-fold excess)
gave straight lines with slopes of K2. The rate of the
ligand exchange reactions at high termperatures was
studied in a similar manner using UV-vis spectroscopy.
The activation parameters ∆Hq (6.40 ( 0.07 kcal/mol)
and ∆Sq (-40 ( 4 eu) were obtained from a plot of ln-
(K2/T) vs 1/T (Figure 3).
The results of the kinetic studies, namely, the clear
dependence of the rate on the [PCy3] and the negative
value of ∆Sq, imply that the phosphine substitution
reaction for complex 1 follows an associative mecha-
nism. Consistent with this supposition, we found that
steric factors exert great influence on the rate of this
ligand exchange reaction; thus, the substitution of PPh3
by PCy3 proceeds 100-1000 times more slowly for the
more sterically encumbered complex (1-Me,2-Ph-Ind)-
Ni(PPh3)Cl (exchange takes hours at room tempera-
ture). On the other hand, when a similar experiment
(22) (a) Bosnich, B.; Mackenzie, P. B. Pure Appl. Chem. 1982, 54,
189. (b) Cherest, M.; Felkin, H.; Umpleby, J . D.; Davies, S. G. J . Chem.
Soc., Chem. Commun. 1981, 681. (c) Consiglio, G.; Piccollo, O.; Roncetti,
L.; Morandini, F. Tetrahedron 1986, 42, 2043. (d) Kurosawa, H. J .
Organomet. Chem. 1987, 334, 243.
(23) (a) For a few trigonal bipyramidal compounds of d8 metals with
tripodal tetradentate ligands see: Aizawa, S.-i.; Ilida, T.; Funahashi,
S. Inorg. Chem. 1996, 35, 5163, and the references therein. (b) For a
few five-coordinate complexes of Rh(I) with multidentate imine ligands
see: Haarman, H. F.; Bregman, F. R.; Ernsting, J .-M.; Veldman, N.;
Spek, A. L.; Vrieze, K. Organometallics 1997, 16, 54, and the references
therein. (c) For square pyramidal Pd(II) complexes of the type (PR3)3-
PdX2 see: Louw, W. J .; de Waal, D. J . A.; Kruger, G. J . J . Chem. Soc.,
Dalton Trans. 1976, 2364. Collins, J . W.; Mann, F. G.; Watson, D. G.;
Watson, H. R. J . Chem. Soc. 1964, 1803. (d) For a trigonal bipyramidal
ethylene complex of Pt(II) see: Maresca, L.; Natile, G.; Calligaris, M.;
Delise, P.; Randaccio, L. J . Chem. Soc., Dalton Trans. 1976, 2386.
(20) Tests have shown that OdPPh3 is completely unreactive toward
both the starting material and the product in these reactions.

Reactions of (1-Me-Indenyl)(PR3)Ni-Cl
Organometallics, Vol. 20, No. 24, 2001 5161
to the Ind, namely, (η3,η1-IndCH2CH2-NMe2)Ni-
associatively activated intermediates play a major role
in the polymerization reactions catalyzed by our com-
plexes. We recognize, however, that the presence of a
large excess of Lewis acids such as MAO in the catalytic
settings might render the normally unfavorable dis-
sociation of PR3 more plausible. To be sure, NMR
monitoring of the reaction of 1 with 50-100 equiv of
MAO failed to detect the formation of the MAO‚PPh3
adduct, although this cannot rule out the possibility that
under the catalytic conditions (i.e., 500 equiv of MAO)
small amounts of a highly active species are generated
via phosphine dissociation.
(PPh3)Cl:24
(η3,η0-IndCH2CH2-NMe2)Ni(PPh3)Cl h
(η3,η1-IndCH2CH2-NMe2)Ni(PPh3)Cl
Con clu sion . The structural studies outlined in this
report have not revealed significant differences in the
hapticity of the 1-Me-Ind ligand to account for the
different reactivities of complexes 1-3. Thus, the solid
state hapticities are somewhat different in these com-
plexes, but the solution hapticities are virtually identi-
cal. The Ni-P bond length in 1 is intermediate between
the shorter distance in 2 and the longer one in 3, while
the order of ethylene polymerization activities is 2 ∼ 3
> 1.
The kinetic studies indicate that five-coordinate in-
termediates such as I (Scheme 2) are likely involved in
the phosphine substitution reactions. The species I is
analogous to the proposed6 intermediates for the po-
lymerization of Ph-CC-H and ethylene, namely, (Ind′)-
(PR3)Ni(X)(L), with L being a substrate molecule such
as ethylene and X being the growing polymer chain.25
Although substrates such as ethylene are less nucleo-
philic than PCy3, they are also significantly less bulky
and present in very large excess compared to PCy3. It
is, therefore, tempting to suppose that this type of
Ack n ow led gm en t. The Natural Sciences and En-
gineering Research Council of Canada and FCAR of
Que´bec are gratefully acknowledged for financial sup-
port. We are also indebted to Dr. I. Horvath for the
generous gift of a sample of P{(CH2)2(CF2)5CF3)}3, Dr.
R. Vollmerhaus for carrying out some of the initial
experiments on the synthesis of 2, Prof. D. Macartney
for help with the UV-vis studies, and Prof. S. Collins
for many stimulating discussions on the mechanism of
the polymerization reactions in our system.
Su p p or tin g In for m a tion Ava ila ble: Complete details of
the X-ray analysis of 2 and 3, including tables of crystal data,
collection and refinement parameters, bond distances and
angles, anisotropic thermal parameters, and hydrogen atom
coordinates; tables of the raw data for the kinetic study. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(24) Groux, L. F.; Be´langer-Garie´py, F.; Zargarian, D.; Vollmerhaus,
R. Organometallics 2000, 19, 1507.
(25) We have proposed that the reaction of this intermediate with
PMAO-type cocatalysts would serve to weaken the Ni-X bond (without
ionizing it), thus facilitating the insertion of ethylene (ref 6b).
OM010608W