Counterion effects on propylene polymerization using two-state ansa-metallocene complexes

Article abstract of DOI:10.1021/ja0207706

Propylene polymerization using unsymmetrical, ansa-metallocene complexes Me₂Y(Ind)CpMMe₂ (Y = Si, C, M = Zr, Y = C, M = Hf) and the co-initiators methyl aluminoxane (PMAO), B(C₆F₅)₃, and [Ph₃C][B(C₆F₅)₄] was studied at a variety of propylene concentrations. Modeling of the polymer microstructure reveals that the catalysts derived from Me₂Si(Ind)CpZrMe₂ and each of these co-initiators function under conditions where chain inversion is much faster than propagation (Curtin-Hammett conditions). Surprisingly, the microstructure of the PP formed was essentially unaffected by the nature of the counterion, suggesting similar values for the fundamental parameters inherent to two-state catalysts. The tacticity of PP was sensitive to changes in [C₃H₆] in the case of catalysts derived from Me₂C(Ind)CpHfMe₂ and PMAO, or [Ph₃C][B(C₆F₅)₄], but the average tacticity of the polymer produced at a given [C₃H₆] decreased in the order [Ph₃C][B(C₆F₅)₄] > PMAO. With B(C₆F₅)₃, the polymer formed was more stereoregular, and its microstructure was invariant to changes in monomer concentration. The PP pentad distributions in this case could be modeled by assuming that all three catalyst/cocatalyst combinations function with different values for the relative rates of insertion to inversion (Δ) but otherwise feature essentially invariant, intrinsic stereoselectivity for monomer insertion (α, β), while the relative reactivity/stability (g/K) of the isomeric ion-pairs present seems to be only modestly affected, if at all. Similar conclusions can also be made about the published propylene polymerization behavior of the C_s-symmetric Me₂C(Flu)CpZrMe₂ complex with different counterions. For every counterion investigated, the principle difference appears to be the operating regime (Δ) rather than intrinsic differences in insertion stereoselectivity (α). Surprisingly, the ordering of the various counterions with respect to Δ does not agree with commonly accepted ideas about their coordinating ability. In particular, catalysts when activated with B(C₆F₅)₃ appear to function at low values of Δ as compared to those featuring B(C₆F₅)₄ (less coordinating) and FAl[(o-C₆F₅)C₆F₄]₃ (more coordinating) or PMAO (more coordinating) counterions where the ordering in Δ is MeB(C₆F₅)₃ < B(C₆F₅)₄ < B(C₆F₅)₄ < FAl[(o-C₆F₅)C₆F₄]₃ ≈ PMAO. Possible reasons for this behavior are discussed.

Full text of DOI:10.1021/ja0207706

Published on Web 06/04/2003
Counterion Effects on Propylene Polymerization Using
Two-State ansa-Metallocene Complexes
Muqtar Mohammed,† Marcio Nele,‡ Abdulaziz Al-Humydi,§ Shixuan Xin,†
Russell A. Stapleton,§ and Scott Collins*,§
Contribution from the Department of Chemistry, UniVersity of Waterloo,
Waterloo, Ontario Canada, N2L 3G1, Progamma de Engenharia Quimica da COPPE,
UniVersidade Federal do Rio de Janeiro, Rio de Janeiro - RJ, Brazil, and
Department of Polymer Science, UniVersity of Akron, Akron, Ohio 44325-3909
Received June 3, 2002; E-mail: collins@polymer.uakron.edu
Abstract: Propylene polymerization using unsymmetrical, ansa-metallocene complexes Me2Y(Ind)CpMMe2
(Y ) Si, C, M ) Zr, Y ) C, M ) Hf) and the co-initiators methyl aluminoxane (PMAO), B(C6F5)3, and
[Ph3C][B(C6F5)4] was studied at a variety of propylene concentrations. Modeling of the polymer microstructure
reveals that the catalysts derived from Me2Si(Ind)CpZrMe2 and each of these co-initiators function under
conditions where chain inversion is much faster than propagation (Curtin-Hammett conditions). Surprisingly,
the microstructure of the PP formed was essentially unaffected by the nature of the counterion, suggesting
similar values for the fundamental parameters inherent to two-state catalysts. The tacticity of PP was
sensitive to changes in [C3H6] in the case of catalysts derived from Me2C(Ind)CpHfMe2 and PMAO, or
[Ph3C][B(C6F5)4], but the average tacticity of the polymer produced at a given [C3H6] decreased in the
order [Ph3C][B(C6F5)4] > PMAO. With B(C6F5)3, the polymer formed was more stereoregular, and its
microstructure was invariant to changes in monomer concentration. The PP pentad distributions in this
case could be modeled by assuming that all three catalyst/cocatalyst combinations function with different
values for the relative rates of insertion to inversion (∆) but otherwise feature essentially invariant, intrinsic
stereoselectivity for monomer insertion (R, â), while the relative reactivity/stability (g/K) of the isomeric
ion-pairs present seems to be only modestly affected, if at all. Similar conclusions can also be made about
the published propylene polymerization behavior of the Cs-symmetric Me2C(Flu)CpZrMe2 complex with
different counterions. For every counterion investigated, the principle difference appears to be the operating
regime (∆) rather than intrinsic differences in insertion stereoselectivity (R). Surprisingly, the ordering of
the various counterions with respect to ∆ does not agree with commonly accepted ideas about their
coordinating ability. In particular, catalysts when activated with B(C6F5)3 appear to function at low values
of ∆ as compared to those featuring B(C6F5)4 (less coordinating) and FAl[(o-C6F5)C6F4]3 (more coordinating)
or PMAO (more coordinating) counterions where the ordering in ∆ is MeB(C6F5)3 < B(C6F5)4 < FAl[(o-
C6F5)C6F4]3 ≈ PMAO. Possible reasons for this behavior are discussed.
evidence2 that suggested counterion effects did not influence
Introduction
PP tacticity using C2-symmetric metallocene catalysts, it would
The study of counteranion effects in metallocenium ion-
catalyzed, olefin polymerization has revealed profound influ-
ences of the counteranion on catalyst stability, activity, and
polymer molecular weight.1 Such effects have been properly
attributed to the degree of association of the counteranion with
the metallocenium ion, including specific bonding interactions
within contact ion-pairs in solution or the solid state, as well as
the dynamics of these ion-pairs in solution.
Surprisingly, the role(s) of the counteranion in influencing
other features of olefin polymerization using metallocenium ions
such as comonomer incorporation and/or poly(propylene) (PP)
tacticity are less well understood.1 While there was some early
appear that strongly coordinating counteranions (e.g., FAl[o-
C6F5-C6F4]3) markedly influence both the polymerization rate
(large decrease) and the tacticity (significant increase) under
controlled conditions.1,3
With Cs-symmetric complexes which produce s-PP, it would
appear that strongly coordinating counteranions can increase
polymer syndiotacticity again, at the expense of catalyst
activity,3-4 although here too there seems to be somewhat
inconsistent data on this point in the literature.5 Interestingly,
solvent effects6 can be important with this class of catalysts; a
(2) (a) Chien, J. C. W.; Song, W.; Rausch, M. D. Macromolecules 1993, 26,
3239-3240. (b) Chien, J. C. W.; Rausch, M. D. J. Polym. Sci., Part A:
Polym. Chem. 1994, 32, 2387-2393.
† University of Waterloo.
(3) Chen, X.-Y.; Metz, M. V.; Li, L.; Stern, C. L.; Marks, T. J. J. Am. Chem.
Soc. 1998, 120, 6287-6305.
‡ Universidade Federal do Rio de Janeiro.
§ University of Akron.
(4) (a) Chen, M.-C.; Marks, T. J. Abstr. Pap. 221st ACS National Mtg. 2001,
San Diego, CA, Inorganic Division, Paper 65. (b) Chen, M.-C.; Marks, T.
J. J. Am. Chem. Soc. 2001, 123, 11803-11804.
(1) For a review, see: Chen, X.-Y.; Marks, T. J. Chem. ReV. 2000, 100, 1391-
1434 and references therein.
9
7930
J. AM. CHEM. SOC. 2003, 125, 7930-7941
10.1021/ja0207706 CCC: $25.00 © 2003 American Chemical Society

Counterion Effects on Propylene Polymerization
A R T I C L E S
Scheme 1
more polar solvent led to an increase in catalyst activity but a
decrease in syndiotacticity.6b Recent work has clarified this issue
in terms of a “leveling” effect on the relative rates of insertion
versus chain back-skip, regardless of the counterion.4
systems would be affected (g/K and δ) and how large an effect
would be shown in, for example, the pentad distribution or other
observable.
In this paper, we summarize the main features of our kinetic
model, report the results of some propylene polymerization
studies involving some simple unsymmetrical metallocene
catalysts (1-4, Scheme 1) using different activators, and model
the stereosequence distributions and their response to changes
in polymerization conditions. In addition, we model some
published data4b relating to the Cs-symmetric ansa-metallocene
complex Me2C(Cp)FluZrMe2 and different counterions. We
believe the conclusions made about the effect of counterions
on PP microstructure could prove to be general for this class of
catalyst.
Last, although counteranion effects on PP microstructure have
been observed with C1-symmetric complexes,7 differing ex-
perimental conditions (e.g., T and [C3H6]) were involved which
also can affect PP microstructure with catalysts of this type.8
We have recently described a kinetic approach for modeling
stereosequence distributions in PP produced using two-state,
bridged, or fluxional metallocene catalysts.9 In favorable cases,
analysis of the pentad distribution and its response to changes
in [C3H6] at constant T for both C1- and Cs-symmetric ansa-
metallocene complexes can provide information on the intrinsic
stereoselectivity of monomer insertion (R and â, Scheme 1),
the mechanism of stereocontrol at each of the two states (e.g.,
site vs chain-end control), as well as estimates for the relative
reactivity/stability of two states (denoted by g/K, Scheme 1)
and the ratio of the rate constants for monomer insertion with
respect to chain back-skip or inversion for one of the states
(denoted by δ, Scheme 1).
It had occurred to us that counteranion effects might manifest
themselves through changes to all of these fundamental quanti-
ties under a given set of conditions, resulting in PP with different
microstructures and thus physical properties, simply by changing
the counterion. On the basis of existing work,1-7 it might be
anticipated that, for example, significant changes to intrinsic
stereoselectivity of the two states might only be accomplished
through use of strongly coordinating counteranions - yet it was
unclear how the other attributes of these types of catalyst
Results and Discussion
Summary of the Model. Although we and others have
written extensively about the basic features inherent to two-
state polymerization catalysts8 and how one can use a kinetic
model to describe stereosequence distributions,9 it is appropriate
to review some of this material here.
In the case where the two states are enantiomeric (Scheme
1, left-hand side), g ) K ) 1 by symmetry, while â ) 1 - R,
and thus the polymerization behavior of Cs-symmetrical catalysts
can be described by two fundamental parameters, R and δ. Both
of these can be reliably estimated by studying the response of
the pentad distribution to changes in [C3H6] at constant T, while
the temperature-dependent behavior of δ and R can be elucidated
by study of such catalysts at different T while varying [C3H6]
at each T.9a
Note that we do not consider chain-end epimerization in the
model. This is certainly expected at sufficiently low [M] with
all ansa-metallocene complexes.10 For the catalysts under study
here, degradation in tacticity was not seen at the lowest [M]
investigated, while it has been reported that the rmmr pentad
does not uniformly increase with decreasing [M] (as would be
expected for chain-end epimerization) with the Me2C(Cp)-
FluZrMe2 catalyst activated by different cocatalysts.4
It is important to note that all Cs-symmetric (and indeed all)
ansa-metallocene catalysts can be directly compared by defining
a variable ∆ which represents the rate of insertion to inversion
(i.e., ∆ ) δ[C3H6], assuming first-order kinetics in monomer,
Scheme 1). For example, if two Cs-symmetric catalysts produce
polymer whose microstructure responds to changes in [C3H6]
in an identical manner but one produces more stereoregular
(5) Ewen, J. A. In Catalyst Design for Tailor-Made Polyolefins; Soga, K.,
Terano, M., Eds.; Elsevier: Tokyo, 1994; pp 405-410.
(6) (a) Vizzini, J. C.; Chien, J. C. W.; Babu, G. N.; Newmark, R. A. J. Polym.
Sci., Part A: Polym. Chem. 1994, 32, 2049-2056. (b) Herfert, N.; Fink,
G. Makromol. Chem. 1992, 193, 773-8.
(7) Giardello, M. A.; Eisen, M. S.; Stern, C. L.; Marks, T. J. J. Am. Chem.
Soc. 1995, 117, 12114-12129 and references therein.
(8) (a) Miller, S. A.; Bercaw, J. E. Organometallics 2002, 21, 934-945. (b)
Veghini, D.; Henling, L. M.; Burkhardt, T. J.; Bercaw, J. E. J. Am. Chem.
Soc. 1999, 121, 564-573. (c) Mohammed, M.; Xin, S.; Collins, S. Am.
Chem. Soc. PMSE Prepr. 1999, 80, 441-442. (d) Kleinschmidt, R.; Reffke,
M.; Fink, G. Macromol. Rapid Commun. 1999, 20, 284-288. (e) Dietrich,
U.; Hackmann, M.; Rieger, B.; Klinga, M.; Leskela, M. J. Am. Chem. Soc.
1999, 121, 4348-4355. (f) Bravakis, A. M.; Bailey, M. P.; Pigeon, M.;
Collins, S. Macromolecules 1998, 31, 1000-1009. (g) Herzog, T. A.;
Zubris, D. L.; Bercaw, J. E. J. Am. Chem. Soc. 1996, 118, 11988-11989.
(h) Gauthier, W. J.; Corrigan, J. F.; Taylor, N. J.; Collins, S. Macromol-
ecules 1995, 28, 3771-3778. (i) Gauthier, W. J.; Collins, S. Macromol-
ecules 1995, 28, 3779-3786. (j) Rieger, B.; Jany, C.; Fawzi, R.; Steimann,
M. Organometallics 1994, 13, 647-653.
(9) (a) Nele, M.; Mohammed, M.; Xin, S.; Collins, S.; Pinto, J. C.; Dias, M.
Macromolecules 2001, 34, 3830-3841. (b) Nele, M.; Collins, S.; Pinto, J.
C.; Dias, M.; Lin, S.; Waymouth, R. M. Macromolecules 2000, 33, 7249-
7260.
(10) (a) Busico, V.; Caporaso, L.; Cipullo, R.; Landriani, L.; Angelini, G.;
Margonelli, A.; Segre, A. L. J. Am. Chem. Soc. 1996, 118, 2105-2106.
(b) Busico, V.; Cipullo, R. J. Am. Chem. Soc. 1994, 116, 9329-30.
9
J. AM. CHEM. SOC. VOL. 125, NO. 26, 2003 7931

A R T I C L E S
Mohammed et al.
polymer than the other, one may safely conclude that R differs
significantly between the two. If the two catalysts have identical
values of R but differ in their response to changes in [C3H6],
the pentad distributions that result will all fall on a single family
of curves when plotted versus ∆, and thus the differential effects
on δ for the two catalysts can be elucidated from their individual
responses.
C1-symmetrical catalysts are more complicated in that g *
K * 1 and â * 1- R as the two states are diastereomeric
(Scheme 1, right-hand side). A wider range of behavior is
expected and observed depending on the magnitudes of the
various parameters, while g/K, R, â, and δ can only be reliably
estimated for catalysts functioning under conditions where ∆
≈ 1.
first-order kinetics and the steady-state hypothesis being valid
for intermediates involved in stereosequence formation:9
rmrr )
{2[∆4(R2 - 2R3 + R4) + ∆3(4R4 - R3 - 3R + 1) + ∆2
(4R4 - 8R3 + 11R2 - 7R + 3) + ∆(4R2 - 4R + 3) + 1]}/
[(∆ + 2)4] (1)
There are higher order terms in R and ∆ ) δ[M] including
the product of the two, corresponding to the different ways of
forming this pentad through a combination of skipped insertion
and/or misinsertion. In the limit of R ) 1, eq 1 simplifies to
2[∆3 + 3∆2 + 3∆ + 1] 2(∆ + 1)3
rmrr )
)
(2)
(∆ + 2)4
(∆ + 2)4
For either type of catalyst, one can define two extremes of
behavior based on the magnitude of ∆. If ∆ , 1 under all
conditions investigated, the two-state catalyst behaves like a
single-state catalyst as the two states are fully equilibrated by
the inversion process. During steady-state chain growth under
these Curtin-Hammett (C-H) conditions, it can be shown that
the average stereoselectivity of a given insertion is given by
the following expression: jꢀ ) (R(g/K) + â)/((g/K) + 1),8f,9
where R, â, and g/K are as defined in Scheme 1.
For a Cs-symmetric catalyst, jꢀ ) 0.5, and thus atactic PP
will be produced, while for a C1-symmetric catalyst, isotactic
or atactic PP (or intermediate forms) may be formed depending
on the magnitude of R, â, and g/K. In the usual situation with
one state isospecific and the other stereorandom (R ≈ 1, while
â ≈ 0.5), if g/K . 1, isotactic polymer will be produced.
When ∆ . 1, the two states are trapped by the insertion
process - hence the term kinetic quenching (KQ); one forms
polymer by an alternating insertion mechanism, and syndiotactic
PP will be produced if both states are isospecific but with equal
and opposite facial selectivity. Again, with C1-symmetrical
catalysts, one may form hemi-isotactic PP if R ≈ 1 and â ≈
0.5, but other behavior is also observed.8a
Under C-H and KQ conditions, although PP microstructure
will be largely invariant to changes in [C3H6], one can
distinguish between these two possibilities (except when R )
â) as the stereosequence distributions are Bernoullian under the
former conditions (single-state behavior). To go from one regime
to another requires varying ∆ (i.e., monomer concentration) by
about 103, an impractical requirement. Thus, only catalysts
which function under intermediate conditions (0.1 < ∆ < 10)
yield reliable estimates of all intrinsic parameters (2 for Cs- and
4 for C1-symmetric catalysts). Further, as some of these
parameters may be correlated with one another (e.g., R and g/K
are frequently correlated as is evident in the equation for jꢀ),9a
it is generally inappropriate to estimate these parameters on the
basis of modeling the pentad (or higher n-ad) distribution
obtained at a single [C3H6].
and, in the limit ∆ . 2, this expression further simplifies to
2k1
2k2
2
∆
rmrr )
)
)
(3)
kAp [M] kBp [M]
In essence, only for Cs-symmetric catalysts functioning under
KQ conditions and which are close to perfectly stereoregulating
will the intensity of this pentad be directly equal to the ratio of
the rates of inversion to insertion. One can easily show that, if
rmrr ) 0.08 under the conditions investigated, this corresponds
to ∆ ) 19.6 (assuming R ) 1), whereas use of the simple
equation rmrr ) 2/∆ gives ∆ ) 25, an error of 28%. On the
other hand, if rmrr ) 0.02, the error in ∆ calculated using the
approximate equation is only about 5%. Larger errors can be
expected if R is significantly less than 1.
We only point this out because one might be tempted to use
the approximate relationship under all conditions, even where
it is clearly invalid. A corollary to this is that if one is not sure
what the magnitude of ∆ is at any particular [M], estimation of
both R and δ from the pentad distribution under these conditions
will be less reliable than using estimates obtained from
experiments at different [M].
Reaction of Dimethyl Complexes 1-4 with B(C6F5)3 and
[Ph3C][B(C6F5)4]. The simple, dimethylmetallocene complexes
1-4 (Scheme 1) were prepared as described in the literature11or
in the Experimental Section. Detailed studies of the reaction of
the Zr dimethyl complexes (1 and 2) with either B(C6F5)3 or
[Ph3C][B(C6F5)4] in toluene or bromobenzene solution, respec-
tively, have been conducted.12Some of the main observations
to date will be summarized here to facilitate discussion.
In essence, isomeric ion-pairs are formed from 1 or 2 and
B(C6F5)3 in toluene solution which are in slow exchange on
the NMR time scale over the entire T range that they are stable
(-80 to ca. +50 °C).13 The major ion-pair present is 5 (Scheme
2) as shown by NOESY spectra (see Supporting Information),
and the ratio of 5:6 varies from about 4:1 (X ) C) to 3:1 (X )
Si) and shows minor variation with T (increasing 5 at lower T).
Insight into the nature of exchange processes involving these
two isomeric ion-pairs was provided by EXSY spectroscopy.
In the case of Cs-symmetric catalysts, it is possible to estimate
both R and δ using a pentad distribution obtained from a single
experiment. However, we wish to emphasize that a complete
pentad analysis using an appropriate model is important to
extract meaningful estimates.
For example, the intensity of the “skipped defect” rmrr pentad
in s-PP produced using Me2C(Cp)FluZrX2 catalyst is given by
the following expression using our kinetic model, which is valid
under all conditions, subject to the underlying assumptions of
(11) (a) Compound 1: Cameron, P. A.; Gibson, V. C.; Graham, A. J.
Macromolecules 2000, 33, 4329-4335. (b) Compounds 2-4: Mohammed,
M. Ph.D. Thesis, University of Waterloo, 2001.
(12) Mohammed, M.; Xin, S.; Al-Humydi, A.; Collins, S.; Monwar, M.; Rinaldi,
P. L., manuscript in preparation.
(13) The ion-pairs formed from the Zr complexes 1 and 2 have been studied in
detail; their Hf analogues show qualitatively similar behavior.
9
7932 J. AM. CHEM. SOC. VOL. 125, NO. 26, 2003

Counterion Effects on Propylene Polymerization
A R T I C L E S
1
Figure 1. High field region (δ -1.6 to 2.6) of the H-1H EXSY NMR spectra of ion-pairs 5 and 6 (X ) Si) with peak assignments as shown. (a) τ ) 50
ms. (b) τ ) 500 ms. Exchange correlation peaks between B-Me and Zr-Me resonances are indicated with solid and dashed boxes, respectively.
Scheme 2
ion-pair reorganization involves two distinct steps (Scheme 2)
or a separate process leads to faster exchange of the B-Me
signals.
Similar behavior has been observed elsewhere in a degenerate
exchange process involving, for example, the ion-pair [rac-Me2-
Si(2-Me-4-tBu-Cp)2ZrMe][MeB(C6F5)3]14 in benzene solution,
where ion-pair reorganization leads to, for example, exchange
correlation of the SiMe2 signals. Rapid anion exchange between
ion-pairs in a bimolecular fashion was invoked - that is,
exchange of anions involving an ion-quadrupole as an inter-
mediate. If this process were involved here, it would have to
occur in a fashion where rapid exchange of anions occurs
without change to the configuration of the metal in the two
isomers; this seems somewhat unlikely but cannot be excluded
on the basis of present evidence.
More recent work using [(1,2-Me2Cp)2ZrMe][MeB(C6F5)3]
in bromobenzene at constant ionic strength and using [nBu4N]-
[MeB(C6F5)3] as a noninteracting source of the MeB(C6F5)3
anion has revealed that ion-pair reorganization is a two-step
process involving unimolecular, reversible, rate-determining
dissociation of the anion followed by inversion at the metal.15
We have seen similar behavior using ion-pairs 5 and 6 (X )
C) in bromobenzene solution where again the B-Me signals
exchange at a rate (k ) 680 s-1) that is 34× faster than the
overall rate of ion-pair reorganization (k′ ≈ 20 s-1 for the Zr-
Me signals, see Supporting Information). This behavior is
consistent with the mechanism proposed by Bercaw and Wendt
As shown in Figure 1, which depicts the high field region of
the EXSY spectrum of ion-pairs 5 and 6 (X ) Si), prominent
exchange correlation between the B-Me signals of the two
isomers is observed at short mixing times (50 ms, Figure 1a),
while, in addition, weaker correlation is seen between the two
Zr-Me signals at longer mixing times (500 ms, Figure 1b).
As indicated in Scheme 2, these two features are consistent
with the overall process of ion-pair reorganization which
permutes B-Me with B-Me and Zr-Me with Zr-Me signals
on the different isomers rather than borane dissociation where
B-Me signals are correlated with Zr-Me signals on different
isomers.1 However, because the B-Me signals exchange at very
different rates than the Zr-Me signals, either the process of
(14) Beck, S.; Lieber, S.; Schaper, F.; Geyer, A.; Brintzinger, H.-H. J. Am. Chem.
Soc. 2001, 123, 1483-1489 and references therein.
(15) Wendt, O. F.; Bercaw, J. E., manuscript in preparation. Personal com-
munication.
9
J. AM. CHEM. SOC. VOL. 125, NO. 26, 2003 7933

A R T I C L E S
Mohammed et al.
Table 1. Polymerization of Propylene Using Complexes 1-4 and
provided that the rate of (degenerate) anion reassociation is much
faster than inversion from the intermediate, solvent-separated
ion-pair in the case of 5 and 6. Future work will focus on
delineating these features.12
Various Cocatalystsa
[M]
M
w
entry complex cocatb (µM) [C H ]
R c
(K)
M /M
%mmmm
3
6
p
w
n
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
1
1
1
1
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
3
3
4
4
4
A
A
A
B
A
A
A
B
B
B
C
C
C
A
A
A
A
B
B
B
B
C
C
C
C
A
A
A
20
20
20
5
1.30
2.21
3.26
1.76
10
17
26
27
25
68
1.73
1.92
2.01
2.24
11.6
13.3
14.9
19.0
26.9
26.2
16.3
15.3
16.2
26.2
38.9
48.1
43.2
32.3
39.6
39.4
42.2
22.2
35.3
34.2
38.2
99.1
2.7
2.7
3.0
2.3
1.7
1.8
1.8
2.4
1.7
1.8
1.8
2.1
2.2
2.3
1.9
2.0
2.0
1.8
1.8
1.6
1.7
2.1
1.9
1.7
1.7
2.2
2.0
2.8
12
14
12
12
31
31
34
34
34
34
34
35
35
38
30
28
23
44
44
39
38
48
47
47
48
51
52
52
It should be mentioned that in the presence of excess borane
(even trace quantities formed as a result of adventitious
deactivation of the dialkyl used to form the ion-pairs) EXSY
spectra reveal additional correlation between Zr-Me and B-Me
resonances on different isomers which is more prominent at
short mixing time than either of the two processes discussed
above (see Supporting Information). As has been reported
elsewhere,14 the two isomeric ion-pairs may also interconvert
by an independent mechanism which involves electrophilic
attack of excess borane on the remaining Zr-Me group of these
ion-pairs. Because, in the current case, this process is evidently
more facile than either of the two unimolecular pathways
(Scheme 2), it should be borne in mind in experiments of this
2.5 1.15
2.5 2.21
2.5 3.26 104
10
10
10
15
15
15
20
20
20
20
1.76
2.48
3.26
1.76
2.48
3.26
0.96
2.21
4.18
8.82
49
79
91
11
15
20
1.5
3.4
5.5
d
1
kind. In particular, the exchange correlation peaks in the H-
5.0 0.96
5.0 1.76
5.0 3.26
2.5 4.41
12
1H EXSY spectra are identical to those expected for borane
dissociation!
43
64
81
Analysis of the intensity data16 from the EXSY spectrum
shown in Figure 1b gives an estimate of the exchange rate
constants of 6 ( 1 and 0.15 ( 0.025 s-1 for the two processes
for the Si-bridged complex. In the case of the C-bridged
analogue, the B-Me exchange process was detected at short
mixing time and occurs at essentially the same rate (kobs ≈ 6 (
1 s-1), while exchange correlation between the two Zr-Me
signals was barely evident at the longest mixing time employed
(5 s). An admittedly crude estimate of the overall rate of ion-
pair reorganization of 0.005 ( 0.0025 s-1 was obtained from
the analysis of this EXSY spectrum. In comparing the Si- with
C-bridged complex, it is therefore evident that ion-pair re-
organization is about 30× faster in the former case at equivalent
concentrations at 25 °C. This finding assumes prominence in
view of the different polymerization behavior of these two
catalysts (vide infra).
20
1.30
1.76
2.21
3.26
1.15
2.21
3.26
2.0
15
20
20
20
20
20
2.8
3.8
6.9
7.5
14
19
152
188
a Conditions: toluene solution (300-500 mL), 30 °C, 1000 rpm, 30-
120 min reaction time depending on catalyst/activator. b A ) solid PMAO
(1000:1 M ) Zr; 2000:1 M ) Hf); B ) [Ph3C][B(C6F5)4] (1.2:1 B:M)
added to reactor (presaturated with monomer and containing ca. 1-2 mM
of MAD) 1-2 min prior to complex 1-4; C ) B(C6F5)3 and metallocene
complex (1.2:1 B:M) combined in 20 mL of toluene at 25 °C and then
added to reactor presaturated with monomer and containing ca. 1-2 mM
of MAD. c Steady-state Rp [mol C3H6/[M] × s)] as measured by calibrated
mass flow meters (for representative flow profiles, see Figure 2). d Poly-
merization in liquid propylene, activity is 1.1 × 105 g of PP/mol Hf × h.
Polymerization of Propylene Using Complexes 1-4 and
Various Co-initiators. Summarized in Table 1 are some
polymerization rate and polymer property data for catalyst
precursors 1-4 in combination with PMAO (1000-2000:1 Al:
M), B(C6F5)3 (1.2:1 B:M), or [Ph3C][B(C6F5)4] (1.2:1 B:M) at
30 °C and various [C3H6] in toluene solution. All combinations
investigated gave rise to catalysts that were reasonably stable
to the reaction conditions as revealed by stable mass flow
profiles with time following attainment of steady-state conditions
(see Figure 2, for example). As would be expected, polymer-
ization activity (Rp) tracks with the nature of the counterion
with B(C6F5)4 ≈ MeMAO > MeB(C6F5)3, while polymer MW
varies to a lesser degree [B(C6F5)4 > MeB(C6F5)3 ≈ MeMAO]
and with the usual dependencies on Zr versus Hf.
In the case of the boron-based co-initiators, an excess of
MeAl(BHT)2 (MAD) was used as an “noninteracting” scrubbing
agent (ca. 0.5-2 mM MAD).17 Earlier we reported that this
compound does not react with B(C6F5)3 or the metallocenium
ion derived from reaction of this activator with, for example,
Cp2ZrMe2 or act as an inhibitor of catalyst activity even when
present in large excess. Similar comments also apply to the ion-
Figure 2. Propylene flow (10-4 mol C3H6 s-1) vs time (s) for catalyst 2
activated by B(C6F5)3 and [Ph3C][B(C6F5)4]. The time period between
catalyst injection and uptake of monomer corresponds to consumption of a
fixed quantity (∼0.4 g at 30 °C and 30 psi) of dissolved monomer so as to
produce a measurable pressure drop leading to initiation of flow. A more
active catalyst features a shorter lag time between injection and detection
of flow.
pairs (5 and 6, X ) Si, [Zr] ) 0.025 M) under study here; both
1H and 19F NMR spectra were unchanged in the presence of
ca. 5 equiv of this additive ([Al] ≈ 0.125 M) after 1 h at 25 °C
(16) Observed rate constants were calculated from kobs ) (1/τm) ln[(r + 1)/(r -
1)], where τm is the mixing time, and r ) ∑Id/∑Ix, where Id and Ix are the
intensities of the diagonal and cross-peaks, respectively. See: Perrin, C.
L.; Dwyer, T. J. Chem. ReV. 1990, 90, 935.
9
7934 J. AM. CHEM. SOC. VOL. 125, NO. 26, 2003

Counterion Effects on Propylene Polymerization
A R T I C L E S
[B(C6F5)4] required about 15 min to reach steady-state, despite
lower concentrations and in situ catalyst formation. The slower
buildup to steady state is consistent with slow initiation relative
to propagation in the case of the borane-activated complexes.
This behavior has been previously observed and quantified in
the polymerization of 1-hexene using [en(Ind)2ZrMe][MeB-
(C6F5)3] where kp/ki ≈ 70.18
It is worth pointing out that the values of Rp reported in Table
1 were those measured at steady-state following completion of
initiation and are thus characteristic of intrinsic catalyst activity.
As the results in the Figure 2 indicate, the difference in steady-
state rates of monomer consumption is only about a factor of 5
for these two counterions with B(C6F5)4 > MeB(C6F5)3.
As far as PP microstructure is concerned, of the various
complexes and co-initiators investigated, only the Hf complex
3 produced PP whose microstructure was sensitive to the nature
of the counterion or to changes in [C3H6] (mmmm ) 23-48%,
Table 1, entries 14-25). Zirconium complexes 1 and 2 furnished
oligomeric, “hemi-isotactic” PP (mmmm e 15%, entries 1-4)
and higher MW, marginally isotactic PP (mmmm ≈ 32-35%,
entries 5-13), respectively, while the Hf analogue of 2 (4)
furnished more stereoregular PP (mmmm ≈ 52%, entries 26-
28) whose microstructure was also invariant to changes in
monomer concentration.
In the case of complex 1 activated by PMAO, where oligo-
PP microstructure is insensitive to changes in [C3H6] (Table 1,
entries 1-3), we were unable to precisely fit the pentad
distributions to either limiting form (C-H or KQ) of our model,
even after correction of the pentad distribution for interference
from end groups.9 We suspect this result simply reflects the
inadequacy of our model (in particular, our model is only valid
where the kinetic chain length ) Rp/Rtr ) Xhn is long compared
to a pentad) rather than anything peculiar about this complex.
In view of this inadequacy, we did not systematically investigate
different activators, etc. with this catalyst particularly when use
of [Ph3C][B(C6F5)4] as an activator led to no change in
microstructure and insignificant changes in the degree of
polymerization (entry 4 vs 1-3).
In the case of Si-bridged complexes 2 and 4 where again no
significant variation in microstructure was found on varying
[C3H6] (Table 1, entries 5-13 and 26-28), we had earlier
shown that both of these complexes behave as single-state
catalysts on activation with PMAO.9a We therefore felt it was
instructive to only study the effect of counterion with one of
these complexes (2) as limited information is available from
analysis of the pentad distribution under C-H conditions (i.e.,
one parameter).
Figure 3. 1H and 19F NMR spectra of ion-pairs 5 and 6 (X ) Si) in the
absence and presence of 5 equiv of MeAl(BHT)2 after 1 h at 25 °C with
[Zr] ) 0.025 M and [Al] ≈ 0.125 M.
as shown in Figure 3. Because the absolute concentrations of
Zr and Al were >103 and 125× greater in this experiment than
those under polymerization conditions (i.e., [Zr] ) 2-20 µM
and [Al] < 2 mM), we can conclude that, on the time scale of
these polymerization experiments (i.e., 1-10 h) and at the
concentrations employed, MeAl(BHT)2 is inert toward these ion-
pairs and does not modify their chemistry.
MeAl(BHT)2 is not inert toward [Ph3C][B(C6F5)4] and
undergoes slow hydride abstraction to form Ph3CH and un-
identified Al byproducts in bromobenzene at room temperature
(t1/2 ≈ 1.5 h at [Ph3C][B(C6F5)4] ) [MAD] ) 0.0.086 M). For
this reason, in polymerization reactions using this activator, the
dimethylmetallocene was added separately, but immediately
after adding [Ph3C][B(C6F5)4] to the reactor presaturated with
monomer and containing excess MAD. We have previously
shown that this procedure minimizes initial degradation of active
cocatalyst in ethylene polymerization involving Cp2ZrMe2 and
this activator.17c
Selected pentad distributions and those calculated using our
kinetic model are summarized in Table 2; complete results are
provided as Supporting Information. As can be gleaned from
the data in Table 2, the pentad distributions of PP formed using
complex 2 (entries 6, 9, and 12) and the various activators are
essentially the same, invariant to changes in [C3H6] and
completely consistent with single-state behavior. As indicated
in the Introduction, only a single parameter jꢀ can be extracted
through analysis of the (Bernoullian) pentad distributions under
C-H conditions, and the fact that this parameter is the same
for all three cocatalysts (Table 3) implies that the fundamental
In comparing the behavior of the different activators em-
ployed, we noted that catalysts activated by B(C6F5)3 exhibited
a slow buildup to steady-state conditions, corresponding to a
constant concentration of growing chains, as compared to the
[Ph3C][B(C6F5)4] activator. As indicated in Figure 2, attainment
of steady-state conditions required nearly 40 min at 30 °C using
preformed ion-pairs at [Zr] ) 15 µM, while use of [Ph3C]-
(17) (a) Williams, V. C.; Dai, C.; Li, Z.; Collins, S.; Piers, W. E.; Clegg, W.;
Elsegood, M. R. J.; Marder, T. B. Angew. Chem., Int. Ed. 1999, 38, 3695-
3698. (b) Vollmerhaus, R.; Rahim, M.; Tomaszewski, R.; Xin, S.; Taylor,
N. J.; Collins, S. Orgaonometallics 2000, 19, 2161-2169. (c) Williams,
V. C.; Irvine, G. J.; Piers, W. E.; Li, Z.; Collins, S.; Clegg, W.; Elsegood,
M. R. J.; Marder, T. B. Organometallics 2000, 19, 1619-1621. (d)
Metcalfe, R. A.; Kreller, D. I.; Tian, J.; Kim, H.; Taylor, N. J.; Corrigan,
J. F.; Collins, S. Organometallics 2002, 21, 1719-1726.
(18) Liu, Z.; Somsook, E.; White, C. B.; Rosaaen, K. A.; Landis, C. R. J. Am.
Chem. Soc. 2001, 123, 11193-11207.
9
J. AM. CHEM. SOC. VOL. 125, NO. 26, 2003 7935

A R T I C L E S
Mohammed et al.
mrrm
Table 2. Selected Pentad Distributions of PP Prepared Using Two-State ansa-Metallocene Complexesa
Complexes 1-3 and Various Cocatalysts (this work)
entryb
mmmm
mmmr
rmmr
mmrr
xmrx
mrmr
rrrr
rrrm
2
0.137
(0.137)
0.118
(0.131)
0.314
0.148
(0.138)
0.138
(0.138)
0.176
0.053
(0.055)
0.054
(0.055)
0.045
0.182
(0.170)
0.163
(0.165)
0.153
0.152
(0.158)
0.155
(0.167)
0.104
0.059
(0.079)
0.075
(0.084)
0.057
0.069
(0.083)
0.060
(0.079)
0.025
0.128
(0.111)
0.141
(0.111)
0.055
0.108
(0.069)
0.126
(0.069)
0.085
4
6
(0.332)
0.341
(0.166)
0.194
(0.025)
0.024
(0.166)
0.171
(0.101)
0.102
(0.051)
0.054
(0.025)
0.013
(0.051)
0.032
(0.083)
0.074
9
(0.337)
0.346
(0.342)
0.303
(0.313)
0.443
(0.431)
0.468
(0.166)
0.194
(0.166)
0.173
(0.166)
0.170
(0.160)
0.167
(0.025)
0.024
(0.024)
0.033
(0.025)
0.022
(0.017)
0.021
(0.166)
0.162
(0.166)
0.195
(0.189)
0.171
(0.165)
0.165
(0.099)
0.104
(0.097)
0.072
(0.080)
0.051
(0.063)
0.049
(0.050)
0.054
(0.049)
0.027
(0.031)
0.014
(0.029)
0.014
(0.025)
0.013
(0.024)
0.040
(0.043)
0.019
(0.019)
0.012
(0.050)
0.035
(0.049)
0.072
(0.065)
0.024
(0.035)
0.024
(0.083)
0.075
(0.083)
0.097
(0.088)
0.088
(0.082)
0.081
12
15
19
23
(0.483)
(0.152)
(0.014)
(0.153)
(0.054)
(0.027)
(0.014)
(0.028)
(0.076)
Me2C(Cp)FluZrMe2 and Various Cocatalysts (refs 4b and 9a)c
cocatd
mmmm
mmmr
rmmr
mmrr
xmrx
mrmr
rrrr
rrrm
mrrm
Ae
B
(0.021)
0.020
(0.043)
0.044
(0.037)
0.107
(0.815)
0.649
(0.022)
0.027
(0.030)
0.021
(0.046)
0.056
(0.065)
0.040
(0.116)
0.171
(0.180)
0.049
(0.645)
0.455
(0.458)
0.791
C
D
(0.021)
(0.044)
(0.049)
(0.790)
a Observed values with calculated values in parentheses - full details are included as Supporting Information. b Entries correspond to those in Table 1.
c Conditions: toluene solution, 60 °C, 3 atm C3H6, 1000 rpm with [Zr] ) 1-10 µM or see ref 4b. d A ) PMAO (1000:1 Al:Zr); B ) [Ph3C][B(C6F5)4]; C
) B(C6F5)3; D ) [Ph3C][FAl(o-C6F5-C6F4)3]. e Calculated values extrapolated from data obtained at 30-50 °C using the known T dependence of δ - see
ref 9a.
Table 3. Modeling of PP Microstructure Produced Using
at constant [C3H6] and [2], the three ion-pairs appear to function
under C-H conditions where the rate of ion-pair reorganization
is faster than insertion (from both states). Clearly, the lack of
information content inherent in the polymerization behavior of
complex 2 is frustrating.
Two-State ansa-Metallocene Complexesa
cocatalyst
R
â
g/K
δ
(σ2)b
Me2Si(Cp)IndZrMe2 (2)
jꢀ ) 0.802
PMAO
[Ph3C][B(C6F5)4]
B(C6F5)3
c
c
c
5.9 × 10-4
1.5 × 10-3
1.6 × 10-3
jꢀ ) 0.804
jꢀ ) 0.807
Fortunately, much less ambiguity exists with respect to the
interpretation of the behavior of Hf complex 3 in the presence
of various co-activators. On average, the PP tacticity decreases
in the order B(C6F5)3 > [Ph3C][B(C6F5)4] > PMAO, and,
interestingly enough, only the latter two activators yield a
catalyst that produces PP whose microstructure is sensitive to
[C3H6] at 30 °C (Table 1). From the complete pentad distribu-
tions (see Table 2, entries 15, 19, and 23 and the Supporting
Information), it is possible to model the observed behavior using
two simple assumptions which will be discussed in order.
Me2C(Cp)IndHfMe2 (3) (Scenario I - R, â, and g/K the same)
PMAO
0.953
0.953
0.953
0.578
0.578
0.578
3.48 0.524 1.1 × 10-3
3.48 0.0942 7.3 × 10-4
3.48 0.0218 2.0 × 10-3
[Ph3C][B(C6F5)4]
B(C6F5)3
Complex 3 (Scenario II - R and â the same, g/K different)
PMAO
0.92
0.92
0.92
0.61
0.61
0.61
14
5.3
4.8
0.822 1.0 × 10-3
0.0953 7.4 × 10-4
0.0117 2.0 × 10-3
[Ph3C][B(C6F5)4]
B(C6F5)3
Me2C(Cp)FluZrMe2 (Scenario I - same R)
d
PMAO
0.974
0.026
0.026
0.026
0.026
1
1
1
1
39.4
31.8
9.99
4.02
e
[Ph3C][FAl(o-C6F5-C6F4)3] 0.974
1.0 × 10-4
3.4 × 10-4
4.5 × 10-4
[Ph3C][B(C6F5)4]
B(C6F5)3
0.974
0.974
Modeling of PP Microstructure Produced by Complex 3
- Scenario I - r, â, and g/K Are the Same. The simplest
explanation for the observed behavior is that, as with complex
2, R, â, and g/K are essentially unaffected by the nature of the
counterion and that only δ varies (Scenario I, parameters in
Table 3). Figure 4a shows the fit of this model to the mmmm
pentad of PP obtained under the different conditions in a plot
of pentad intensity versus the variable ∆; given the experimental
error inherent in the measurement of these stereosequence
distributions as well as the reproducibility of polymerization
experiments, it is clear that all of the data obtained using the
different counterions and monomer concentrations can be
adequately fit using this simple model.
Me2C(Cp)FluZrMe2 (Scenario II - different R)
PMAOb
[Ph3C][FAl(o-C6F5-C6F4)3] 0.975
0.975
0.025
0.025
0.025
0.034
1
1
1
1
39.4
31.1
9.78
4.33
e
d
1.0 × 10-4
3.3 × 10-4
3.4 × 10-4
[Ph3C][B(C6F5)4]
B(C6F5)3
0.975
0.966
a For definition of parameters and model description, see introduction
and ref 9a. b Standard variance of the fit, normalized to the number of
experiments. c Curtin-Hammett conditions apply (Bernoullian distribution),
and δ cannot be unambiguously determined. d â is equal to 1 - R. e The
data used for modeling purposes were extrapolated from earlier work at
30-50 °C - an error estimate is not applicable.
parameters R, â, and g/K are also similar for the different
counterions investigated.
Also, despite significant differences in Rp for the different
activators investigated, spanning about an order of magnitude
Scenario II - r, â Are the Same with g/K Different. If
one relaxes the restriction on g/K, allowing this parameter to
“float” for the different cocatalysts (Scenario II), the overall fit
9
7936 J. AM. CHEM. SOC. VOL. 125, NO. 26, 2003

Counterion Effects on Propylene Polymerization
A R T I C L E S
Figure 4. M4 pentad intensity vs ∆ for complex 3 activated by different cocatalysts: (a) R, â, and g/K are the same for all activators; (b) R and â are the
same but g/K is different for the three cocatalysts. Parameters and fits are summarized in Table 3, while representative experimental and calculated pentad
distributions appear in Table 2.
of the model to all of the data is not noticeably improved despite
the inclusion of an additional adjustable parameter (Table 3,
Figure 4b). In particular, the modeling suggests that g/K is
essentially identical for the B(C6F5)3- and [Ph3C][B(C6F5)4]-
activated catalyst while the PMAO-activated complex may have
a higher value of g/K.
It should be mentioned that in earlier work9a significant
correlation (-r2 ) 0.95) existed between the estimated R and
g/K parameters for the PMAO-activated complex and so the
current assumption of having all catalysts operating with the
same values of R and â (0.91 and 0.61, respectively) leads to a
different estimate for g/K (i.e., 14 vs 9 in earlier work). In fact,
this correlation is caused by the lack of asymptotic behavior
observed for PP microstructure versus ∆ and thus ambiguous
estimation of R and â versus g/K.
In comparing the values of the parameters obtained for these
two scenarios, as well as the goodness of the fit as reflected in
the variance (Table 3), it is obvious that the principal effect on
Figure 5. R4 pentad intensity vs ∆ for Me2C(Cp)FluZrMe2 activated by
B(C6F5)3, [Ph3C][B(C6F5)4], PMAO, and [Ph3C][FAl(o-C6F5-C6F4)3] at 60
°C. Data are taken from ref 4b or are extrapolated from polymerization
reactions conducted at 30-50 °C (ref 9a).
changing the counterion is one of changing the value of δ, the
ratio of the rate constants for insertion to inversion for the
aspecific state (similar comments apply to the isospecific state
examined by Chen and Marks, and in particular the dependence
of the pentad distribution on [C3H6] at 60 °C has been studied
by these authors. Selected results, along with the calculated
distributions obtained using our model, are summarized in Table
2, while a graphical representation of the rrrr pentad intensity
versus ∆ appears in Figure 5.19
Here, there are only two parameters involved (R and δ), and,
at least at 60 °C, all of the data can be adequately fit by assuming
that R is the same for all counterions investigated (i.e., R ≈
0.974, Scenario I, Table 3) and that the only thing changing is
the ratio of the rate constants for insertion to inversion (which
in this case corresponds to δ as both states are equivalent). In
comparing B(C6F5)3 with [Ph3C][B(C6F5)4], it is evident that a
similar change in δ is involved for this catalyst as compared to
complex 3 and the change is in the same direction with the
borane-activated complex functioning at lower values of ∆ )
δ[M].
as δ × g/K ) kA/k1). In particular, δ varies by more than an
p
order of magnitude in going from B(C6F5)3 to PMAO with the
[Ph3C][B(C6F5)4] intermediate between these two activators. In
contrast, if g/K does vary (and this is by no means proven from
the data), it does so by at most a factor of 3 for the different
activators investigated.
As should be evident from the foregoing discussion, even
with all of the available data, it was still necessary to make an
assumption about this two-state catalyst, that R and â (and likely
g/K) are the same for the different counterions, etc. While we
have justified this assumption on the basis of literature precedent,
intrinsically it is of interest to be able to address this issue -
one needs a simpler system (i.e., one with fewer parameters)
as in the following section.
Modeling of PP Microstructure Produced by Me2C(Cp)-
FluZrMe2 and Various Activators. As mentioned in the
Introduction, the activation of Ewen’s prototypical Cs-symmetric
metallocene complex Me2C(Cp)FluZrMe2 using B(C6F5)3, [Ph3C]-
[B(C6F5)4], and [Ph3C][FAl(o-C6F5-C6F4)3] has been recently
While we do not have access to strictly comparable data using
PMAO as activator, it seems, on the basis of work performed
(19) We thank Prof. Marks for discussing these results prior to their publication.
9
J. AM. CHEM. SOC. VOL. 125, NO. 26, 2003 7937

A R T I C L E S
Mohammed et al.
Table 4. Temperature Dependence of R and δ Estimated from Modeling of Pentad Distributionsa
MeB(C F )
B(C F )
FAl(o-C F −C F )
6 5 6 4 3
6
5 3
6
5 4
T (°C)
[C H ]
R
δ
σ (δ)b
−r2
σc
R
δ
R
δ
σ (δ)b
−r2
σc
3
6
-10
0
10
25
40
60
3.80
2.56
1.83
1.19
0.85
0.54
0.988
0.988
0.987
0.984
0.983
0.973
22.6
23.5
17.9
10.4
6.02
4.11
0.31
0.28
0.25
0.49
0.57
0.41
0.48
0.48
0.49
0.71
0.79
0.77
0.64
0.67
0.92
3.0
6.2
5.6
0.990
0.989
0.988
0.983
0.979
0.975
91.5
80.3
50.0
35.0
19.6
0.993
0.994
0.989
0.986
0.983
0.976
2.1
3.0
0.73
0.97
1.3
189
134
79.6
30.5
2.0
1.8
1.2
0.94
0.48
0.62
0.63
0.47
9.78
2.7
a Data from ref 4b. For calculated and observed pentad distributions, see Supporting Information. b Estimated standard deviation in the value of δ. c Estimated
standard deviation over all pentads (×102).
earlier at 30-50 °C and the T-dependent behavior of PMAO-
activated Me2C(Cp)IndZrCl2,9a that this catalyst resembles that
activated by [Ph3C][FAl(o-C6F5-C6F4)3] - both function close
to the KQ limit. It is perhaps appropriate to point out that the
values of δ obtained from modeling with these two activators
(39.4 and 31.8, respectively, Table 3) correspond to ∆∆Gq
between inversion and insertion of 2.43 and 2.29 kcal mol-1 at
60 °C. On this basis, it is perhaps correct to state that the
differential effects of the counterion cannot be reliably inter-
preted in the case of these two activators, while, for example,
the borane activated complex is significantly different!
In this simple, two parameter system, an examination of the
error structure reveals no strong cross-correlation of δ with R
when these are allowed to vary independently for the different
counterions (data in Table 3, Scenario II). The same basic
conclusion prevails, and only modest (and probably not
significant given the errors involved) differences in R between
the different activators are noted. In essence, counterion effects
manifest themselves mainly by altering the relative rates of
insertion to inversion in C1- or Cs-symmetric catalysts. Similarly,
the modeling results with the Cs-symmetric complex indicate
that R is largely unaffected by the nature of the counterion, and
thus our earlier assumption seems quite reasonable in the context
of an unsymmetrical system (3).
borane-activated complex at 60 °C (Table 4) with those obtained
from the experiments summarized in Table 3. Not only are the
parameters somewhat different, but the fit of the model to the
data is about 2.5× worse for the single point experiment. Finally,
as one approaches the KQ limit, it becomes impossible to
reliably estimate δ from a single point calculation; at the KQ
limit, it is also impossible to estimate δ even from experiments
at different [M]. This is certainly true for the catalyst partnered
with the fluoroaluminate counterion at low temperature where
we were unable to provide meaningful estimates of δ (Table
4).
This feature was also evident from plots of ln(δ) versus 1/T
where deviation from Arrhenius behavior was seen for some
of the counterions at low T. This is a reflection of our inability
to reliably estimate large values of δ from single pentad
distributions. From the slope of these plots, ∆∆Hq for insertion
versus inversion was estimated as -5.3, -6.2, and -8.2 kcal
mol-1 for the methylborate, borate, and aluminate counterions,
while ∆∆Sq ) -13, -14, and -18 eu was similar for all three
counterions. The values obtained for the fluoroaluminate coun-
terion were based on the three highest T’s reported as significant
downward curvature was observed in the Arrhenius plot at lower
T.
Conclusions
With these Cs-symmetric complexes, as mentioned in the
Introduction, it is possible to estimate both R and δ from single
experiments. In the work of Chen and Marks, polymerization
experiments were also performed at different T and at constant
pressure (1 atm) with the different activators. The pentad
distributions that result were analyzed using our model (see
Supporting Information), and the parameters obtained are
summarized in Table 4 along with estimates of errors in δ,
(negative) correlation coefficients between R and δ, and the
overall fit of the model to the data.
The polymerization experiments and modeling work presented
here clearly indicate the origin of counterion effects on PP
microstructure using two-state ansa-metallocene catalysts. In
essence, of the fundamental parameters that are characteristic
of these systems, it seems that δ is the only one strongly affected
by the counterion and that the ordering of the various counte-
rions appears to be independent of catalyst structure/polymer-
ization conditions at least for those systems functioning under
intermediate conditions.
As in earlier work,9a the T dependence of R was flat over a
small range, but in this case it is clear that, over a 70 °C range,
these catalysts become more stereoselective at lower T. A more
pronounced change in δ is observed for all counterions over
the same T range, and it is this feature which largely determines
the appearance of the pentad distribution at any particular T
with these catalysts.
It can be seen that the errors in δ and the overall discrepancy
between the model and the data are generally more significant
under conditions that deviate most from KQ. This is easy to
understand as the pentad distribution is a sensitive function of
both R and δ under such conditions (i.e., eq 1), and more reliable
parameter estimates are only available through study of this
dependence through, for example, changes to [M]. This is
evident in comparing the single point results obtained for the
The only outstanding issue is whether g/K is affected by the
nature of the counterion. Certainly our modeling of the data
does not support a scenario where g/K must be different. It is
plausible to argue that the most stable state in Scheme 1 is B
(at least if X- is weakly coordinating) and thus K, as originally
defined is >1 and 1/K < 1. Conversely, it is well known that,
for many metallocene complexes, isospecific insertion is ap-
parently faster than aspecific insertion and thus g is also expected
to be >1. The net effect then is that g/K may not be a very
sensitive indicator of counterion interactions as those factors
which affect relative stability may have an opposite effect on
reactivity.
In comparing complex 3 and Me2C(Cp)FluZrMe2, a more or
less uniform ordering of the counterions can be discerned as
far as their effect on δ is concerned. However, it is evident that
9
7938 J. AM. CHEM. SOC. VOL. 125, NO. 26, 2003

Counterion Effects on Propylene Polymerization
A R T I C L E S
the ordering of the counterion effect on δ [i.e., MeB(C6F5)3 <
B(C6F5)4 < FAl(o-C6F5-C6F4)3 ≈ PMAO] bears no strong
resemblance to their coordinating ability [FAl(o-C6F5-C6F4)3
. MeB(C6F5)3 > B(C6F5)4] nor to trends in polymerization
activity [FAl(o-C6F5-C6F4)3 < MeB(C6F5)3 , B(C6F5)4].
The observed behavior is counterintuitive; both inversion and
insertion are either known1,20 or thought21 to involve the
dissociation and/or displacement of the counterion. While there
could easily be a differential effect on the rates of these two
processes, thus leading to differences in δ, one might expect
the magnitude of this parameter to track with the coordinating
ability of the counterion. However, it is possible that while
intrinsic insertion rates are more sensitive to the nature of the
counterion, inversion rates are less so. Thus, for example, the
anomalous position of the MeB(C6F5)3 anion on the ∆ scale
could simply reflect much lower insertion rates as compared to
a lesser change in inversion rates.
At 25 °C where catalyst degradation is presumably less
problematic, the activity data reported in ref 4b for the B(C6F5)3,
[Ph3C][B(C6F5)4], and [Ph3C][FAl(o-C6F5-C6F4)3]-activated
Me2C(Cp)FluZrMe2 catalyst fall in the order A ) 0.44, 8.9, and
0.2 × 106 g PP mol Zr-1 atm-1 h-1, respectively. Using our
values of δ determined for these counterions at 25 °C (Table
4), one can estimate that the relative rates of inversion are 28,
170, and 1, respectively, which is in qualitative agreement with
the coordinating ability of these discrete counterions.22
with recent work from the Landis group on low T detection of
propagating ion-pairs where the MeB(C6F5)3 counterion appears
to be less coordinating during 1-hexene polymerization.20
Finally, why do some catalysts when activated with the same
cocatalyst appear to operate under C-H conditions (e.g., 2 or
4), while others have intermediate behavior (e.g., 3) or exhibit
highly alternating insertion [e.g., Me2C(Cp)FluZrMe2], despite
similar polymerization rates ? For example, in the case of the
PMAO-activated hafnocenes 3 and 4, the latter is about 4× more
active than the former under the same conditions (Table 1), and
yet 3 functions under intermediate conditions (∆ ≈ 1), while 4
behaves like a single-state catalyst (∆ < 0.1).
This result implies that the rate of ion-pair reorganization
differs greatly with the Si-bridged complex undergoing this
process 1-2 orders of magnitude faster than the C-bridged
catalyst! As alluded to earlier, the rates of ion-pair reorganization
are different for ion-pairs derived from Zr complexes 2 or 1
and B(C6F5)3 with the Si-bridged complex showing a signifi-
cantly higher rate (ca. 30× higher) based on EXSY spectra at
different mixing times. This result is consistent with the observed
behavior of their Hf analogues in propylene polymerization.
Experiments underway will clarify whether the observed dif-
ferences in rate are large enough to counterbalance the modest
activity difference between these two systems.
Experimental Section
We are, however, concerned whether the values obtained for
δ for the different counterions will quantitatively agree with
measured inversion versus insertion rates. In particular, the rates
of ion-pair reorganization of model ion-pairs 5 and 6 are some
300× faster for, for example, B(C6F5)4 versus MeB(C6F5)3 in
C6D5-Br solution at the same [Zr] and T (see Supporting
Information), yet δ differs only by a factor of about 2 and in
the wrong direction! Even more dramatic differences in inver-
General. All solvents and chemicals were reagent grade and purified
as required. All synthetic reactions were conducted under an atmosphere
of dry nitrogen in dry glassware unless otherwise noted. Tetrahydro-
furan, diethyl ether, hexane, toluene, and dichloromethane were dried
and deoxygenated by passage through columns of A2 alumina and Q5
deoxo catalyst as described in the literature.24
The ligands Me2C(IndH)CpH and Me2Si(IndH)CpH were prepared
according to reported procedures.8i,25 The compounds Zr(NMe2)4 and
Hf(NMe2)4 were prepared as described in the literature.26 Amine
elimination reactions26 were used to prepare the Me2X(Ind)CpM(NMe))2
complexes following published procedures (X ) C, M ) Zr, Hf, X )
Si, M ) Zr)8i,27 or as described elsewhere (X ) Si, M ) Hf).28 These
compounds could be converted to the known dichloride complexes8i,25,27
by reaction with excess Me3SiCl.26 The zirconium dimethyl complexes
1 and 2 could be prepared by alkylation of the latter compounds with
MeLi as described below. For the hafnium dimethyl complexes 3 and
4, the procedure of Kim and Jordan26e was employed using the bis-
(dimethylamido) complexes and AlMe3 as described below. MeAl-
(BHT)2 was prepared using the method reported by Ittel and co-
workers29 and was added to toluene solvent (ca. 50-100 µM) used to
dilute stock solutions of compounds 1-4 or cocatalysts prior to delivery
to the reactor. The activators B(C6F5)3 and [Ph3C][B(C6F5)4] were
generously donated by Nova Chemicals Ltd.
-
sion rates are reported for MeB(C6F5)3 versus FAl(o-C6F5-
C6F4)3- partnered with Me2C(Cp)FluZr+Me in toluene solution,
and while the change is in the expected direction, δ again only
changes by about an order of magnitude as compared to a ca.
103 difference in inversion rates.4
We suspect the discrepancy here may reflect the choice of
model ion-pair (or medium) used to determine inversion rates.
Beswick and Marks have recently noted pronounced steric
effects on the rate of ion-pair reorganization in [(1,2-Me2-
Cp)2ZrR][MeB(C6F5)3] ion-pairs with sterically hindered Zr-R
groups exhibiting lower barriers to reorganization.23 Differences
in free energies of activation as large as ∼4 kcal mol-1 for Zr-
t
Me versus Zr-CH2 Bu have been reported. The latter value
corresponds to a rate difference of 850 at 25 °C and provides
a reasonable explanation of why the B(C6F5)3-activated com-
plexes appear to operate at significantly lower values of ∆; in
essence, these catalysts have both (moderately) slower insertion
rates coupled with a dramatic increase in ion-pair reorganization
rates following initiation. The latter hypothesis is also consistent
(24) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.; Timmers,
F. J. Organometallics 1996, 15, 1518-1520.
(25) Green, M. L. H.; Ishihara, N. J. Chem. Soc., Dalton Trans. 1994, 657.
(26) (a) Hughes, A. K.; Meetsma, A.; Teuben, J. H. Organometallics 1993, 12,
1936. (b) Herrmann, W. A.; Morawietz, M. J. A. J. Organomet. Chem.
1994, 482, 169. (c) Diamond, G. M.; Rodewald, S.; Jordan, R. F.
Organometallics 1995, 14, 5. (d) Carpenetti, D. W.; Kloppenberg, L.;
Kupec, J. T.; Peterson, J. L. Organometallics 1996, 15, 1572. (e) Kim, I.;
Jordan, R. F. Macromolecules 1996, 29, 489. (f) Diamond, G. M.; Jordan,
R. F.; Petersen, J. L. J. Am. Chem. Soc. 1996, 118, 8024. (g) Diamond, G.
M.; Jordan, R. F.; Petersen, J. L. Organometallics 1996, 15, 4530. (h)
Diamond, G. M.; Jordan, R. F.; Petersen, J. L. Organometallics 1996, 15,
4038.
(20) For relevant experimental work on borane-activated catalysts, see: Landis,
C. R.; Rosaaen, K. A.; Sillars, D. R. J. Am. Chem. Soc. 2003, 125, 1710-
1711.
(21) For theoretical work on the involvement of the counterion on insertion,
see: (a) Lanza, G.; Fragala, I. L.; Marks, T. J. J. Am. Chem. Soc. 2000,
122, 12764-12777. (b) Chan, M. S. W.; Ziegler, T. Organometallics 2000,
19, 5182-5189. (c) Vanka, K.; Ziegler, T. Organometallics 2001, 20, 905-
913.
(27) Hermann, W. A.; Morawietz, M. J. A.; Priemeier, T. J. Organomet. Chem.
1996, 506, 351.
(28) Mohammed, M. Ph.D. Thesis, University of Waterloo, 2001.
(29) Shreve, A. P.; Mulhaupt, R.; Fultz, W.; Calabrese, J.; Robbins, W.; Ittel,
S. D. Organometallics 1988, 7, 409-416.
(22) We thank a reviewer for suggesting this type of analysis.
(23) Beswick, C. L.; Marks, T. J. J. Am. Chem. Soc. 2000, 122, 10358-10370.
9
J. AM. CHEM. SOC. VOL. 125, NO. 26, 2003 7939

A R T I C L E S
Mohammed et al.
Table 5. 1H NMR Shift Data for Ion-Pairs Formed from
1
Routine H and 13C NMR spectra were recorded in benzene-d6 or
a
Complexes 1,2 and B(C6F5)3
CDCl3 solution on either a Bruker AM-250, AC-200, or AC-300
spectrometer. 1H and 19F NMR spectra of ion-pairs were recorded using
an AC-200 or Varian Mercury 300 or Innova 400 spectrometer in
toluene-d8 or bromobenzene-d5 solution with 2,3,5,6-tetrafluoro-p-
xylene as an internal standard. Elemental analyses were determined
by Oneida Research Services, Inc., New York.
Polymerization Procedures and Polymer Characterization. Pro-
pylene polymerization in toluene solvent and characterization of poly-
(propylene) involved techniques and methods that have been presented
in detail elsewhere.8f Propylene solubility in toluene as a function of
P,T was estimated from literature data,30 and these concentrations appear
in Table 1. Mass transfer limitations can significantly affect PP MWD
and microstructure even using single-site catalysts if the polymer formed
is soluble in the reaction medium;9 diffusion-limited, polymerization
experiments in toluene can be detected by variation in monomer flow
with, for example, stirring speed. All polymerization experiments
reported here were conducted at catalyst concentrations and at stirring
speeds (1000 rpm) such that intrinsic behavior was studied. Temperature
control was generally better than (1 °C following catalyst introduction,
and catalyst concentrations were chosen so as to avoid exothermic
conditions.
Polymerization reactions were typically conducted for periods of
time corresponding to attainment of steady state as revealed by a stable
mass flow profile (see Figure 2) and then for an additional period equal
to the length of time required to reach steady state in most cases.
Reactions were halted by venting the monomer and emptying the reactor
contents into a round-bottomed flask containing methanol. Solvent was
partially evaporated to dryness in vacuo, and then the syrup was
transferred to an aluminum pie plate. Solvent was allowed to further
evaporate at room temperature in a hood and then at 60 °C in a vacuum
oven (∼1 mmHg).
X ) C (4:1)
X ) Si (3:1)
assignment
major (5)
minor (6)
major (5)
minor (6)
H4
H5
H6
H7
H3
H3′′
H3′
H2
H2′
H2′′
XMe2
XMe2
BMe
ZrMe
7.13 (d)
6.88 (t)
7.23 (d)
7.00 (t)
7.13
6.41
6.52 (t)
6.65 (t)
6.67 (d)
6.25 (d)
5.95 (dd)
5.86 (dd)
5.49 (d)
5.06 (dd)
4.40 (dd)
1.21 (s)
1.10 (s)
6.59 (d)
6.54 (d)
6.25 (dd)
6.07 (dd)
5.65 (d)
5.32 (dd)
4.85 (dd)
0.17 (s)
0.11 (s)
6.84
6.35 (d)
5.66 (dd)
6.06 (dd)
4.72 (d)
5.25 (dd)
4.39 (dd)
1.27 (s)
6.70 (d)
5.87 (dd)
6.38 (dd)
5.07 (d)
4.85 (dd)
5.37 (dd)
0.28 (s)
0.88 (s)
-0.01 (s)
0.81 (br s) -0.26 (br s)
-0.85 (s) 0.44 (s)
0.53 (br s) -0.54 (br s)
-0.84 (s)
0.54 (s)
a Toluene-d8 solution, 298 K, 300 MHz. Assignments based on NOESY
and COSY spectra.
warm slowly to room temperature, and stirring was continued for an
additional 1 h after reaching room temperature. The reaction mixture
was filtered via a cannula, and the ether was removed in vacuo. The
off-white solid was dissolved in a minimum amount of pentane and
was cooled to -35 °C overnight. The supernatant was decanted, and
the crystals were washed with cold pentane. The crystalline material
was dried under vacuum in a glovebox to give 113 mg (66% yield) of
Samples for GPC analysis were obtained by Soxhlet extraction of a
portion of the polymer with refluxing toluene containing 0.1 wt %
Irganox 1010 as an antioxidant. The soluble extracts were concentrated
to dryness under high vacuum, and the purified polymer was dissolved
in TCB (ca. 0.1-0.5 wt %) containing 0.1 wt % Irganox. The solutions
were then analyzed after dissolution periods of 1-3 h at 150 °C using
a Waters 150C+ instrument as described elsewhere.8f It is important
to use antioxidant during Soxhlet extraction and to minimize dissolution
time at 150 °C to prevent thermal degradation of the polymer.
Propylene Polymerization Using Borane or Borate Activators.
An autoclave reactor containing 480 mL of toluene and 1.0 g (ca. 2
mmol) of MAD was saturated with propylene at 30 °C at the desired
pressure for several hours until monomer flow had effectively ceased.
After another hour, solutions of the ion-pairs or dialkyl and cocatalyst
were added in the same manner as described below.
In the case of B(C6F5)3, a solution of the ion-pairs [7.5 mM in
complexes 1-3 and 9 mM in B(C6F5)3 B:M ) 1.2:1] was prepared by
mixing stock solutions of each component at room temperature. One
milliliter of this solution was transferred into a 50 mL sample bomb
and diluted with 19.0 mL of toluene prepurified using MAD. After a
period of 5-10 min, the contents were delivered to the reactor using
a 10 psi overpressure of dry N2.
In the case of [Ph3C][B(C6F5)4], a similar procedure was employed
except 10.0 mL of a solution of [Ph3C][B(C6F5)4] (0.15-0.30 mM
depending on metal and [C3H6]) was added to the reactor followed by
10.0 mL of a solution of 1-3 (0.125-0.25 mM B:M ) 1.2:1] within
a minute or less.
Preparation of Me2C(Ind)CpZrMe2 (1). A suspension of Me2C-
(Ind)CpZrCl2 (191 mg, 0.5 mmol) in 25 mL of diethyl ether was cooled
to -78 °C, and MeLi (1.0 mL of 1.0 M solution in ether, 1.0 mmol)
was added via a syringe while stirring. The solution was allowed to
1
spectroscopically pure compound 1. H NMR (250 MHz, C6D6): δ
7.46 (d, J ) 8.6 Hz, 1H, H4), 7.23 (d, J ) 8.9 Hz, 1H, H7), 7.10 (m,
1H, H5), 6.78 (m, 1H, H6), 6.64 (d, J ) 3.4 Hz, 1H, H3), 6.30 (m,
1H, H3′), 6.25 (m, 1H, H3′′), 5.52 (d, J ) 3.5 Hz, 1H, H2), 5.30 (dd,
J ) 2.5 Hz, 1H, H2′), 5.14 (dd, J ) 2.5 Hz, 1H, H2′′), 1.51 (s, 3H,
Me2C), 1.32 (s, 3H, Me2C), 0.06 (s, 3H, Me-Zr), -1.01 (s, 3H, Me-
Zr). 13C NMR (50.32 MHz, C6D6): δ 125.9, 124.1, 123.8, 123.7, 119.3,
114.7, 113.8, 113.4, 113.0, 104.6, 103.9, 103.3, 38.4, 34.5, 30.7, 26.3,
25.4. Anal. Calcd for C19H22Zr: C, 66.81; H, 6.49. Found: C, 66.59;
H, 6.39.
Preparation of Me2Si(Ind)CpZrMe2 (2). A suspension of Me2Si-
(Ind)CpZrCl2 (199 mg, 0.5 mmol) in 25 mL of diethyl ether was cooled
to -78 °C, and MeLi (1.0 mL of 1.0 M solution in ether, 1.0 mmol)
was added via a syringe into the suspension while stirring. The solution
was allowed to warm to 0 °C and was stirred at this temperature for an
additional 1.5 h. The reaction mixture was filtered to remove LiCl,
and the ether was removed in vacuo. The off-white residue was
dissolved in a minimum amount of pentane and was cooled to -35 °C
overnight. The supernatant was decanted, and the crystals were washed
with cold pentane. The crystalline material was dried under vacuum in
a glovebox to give 122 mg (68% yield) of spectroscopically pure
1
compound 2. H NMR (250 MHz, C6D6): δ 7.63 (br d, J ) 8.6 Hz,
1H, H7), 7.20 (m, 2H, H4-H5), 6.89 (d, J ) 3.3 Hz, 1H, H6), 6.63
(m, 1H, H3), 6.60 (dd, J ) 2.9 Hz, 1H, H3′), 6.53 (dd, J ) 2.9 Hz,
1H, H3′′), 5.70 (d, J ) 3.3 Hz, 1H, H2), 5.48 (dd, J ) 2.9, 1H Hz,
H2′), 5.37 (dd, J ) 2.9 Hz, 1H, H2′′), 0.44 (s, 3H, Me2Si), 0.31 (s,
3H, Me2Si), 0.12 (s, 3H, Me-Zr), -1.90 (s, 3H, Me-Zr). 13C NMR
(50 MHz, C6D6): δ 130, 126.9, 125.9, 125.0, 124.5, 120.6, 119.3, 118.5,
112.8, 112.2, 114.4, 35.2, 31.1, -3.2, -4.2. Anal. Calcd for C18H22-
SiZr: C, 60.44; H, 6.20. Found: C, 60.50; H, 6.35.
Preparation of Me2C(Ind)CpHfMe2 (3). A suspension of Me2C-
(Ind)CpHf(NMe2)2 (2.44 g, 5.0 mmol) in 25 mL of hexane was cooled
to -35 °C, and neat trimethylaluminum (1.44 g, 20 mmol) at -35 °C
(30) (a) Atiqullah, M.; Hammawa, H.; Hamid, H. Eur. Polym. J. 1998, 34, 1511-
1520. (b) Reid, R. R.; Prausnitz, J. M.; Poling, B. E. The Properties of
Gases & Liquids; McGraw-Hill: Singapore, 1987.
9
7940 J. AM. CHEM. SOC. VOL. 125, NO. 26, 2003

Counterion Effects on Propylene Polymerization
A R T I C L E S
was added into the suspension while stirring. The solution was warmed
to ambient temperature and stirred overnight. The reaction mixture was
vacuum-dried to provide crude product, which was recrystallized in
hexane at -35 °C to provide spectroscopically pure 3 as an off-white,
0.29 (s, 3H, Me2Si), -0.16 (s, 3H, Me-Hf), -1.28 (s, 3H, Me-Hf).
13C NMR (75 MHz, C6D6): δ 130.0, 126.6, 125.9, 125.2, 125.0, 124.7,
120.1, 118.7, 117.6, 111.7, 111.5, 110.5, 40.1, 38.0, -3.3, -4.1.
Formation of Ion-Pairs 5 and 6 from Complexes 1,2 and B(C6F5)3
in Toluene-d8. Stock solutions of Me2X(Ind)CpZrMe2 (1) (0.05 M)
and B(C6F5)3 (0.050 M) in toluene-d8 were mixed in a 1:1 ratio at -35
°C in a glovebox. The immediate color change from colorless to orange
red, upon mixing the two stock solutions, indicated a rapid reaction at
this temperature. The 1H and 19F NMR spectra of the resulting solution
show two sets of signals in a roughly 1:3 (X ) Si) or 1:4 (X ) C)
ratio at ambient temperature. The 1H NMR chemical shifts of the ion-
pairs formed are listed in Table 5.
1
crystalline powder in 96% yield. H NMR (300 MHz, C6D6): δ 7.42
(d, J ) 8.6 Hz, 1H, H7), 7.23 (dd, J ) 8.7, 1.0 Hz, 1H, H4), 7.05 (dd,
J ) 8.6, 7.0 Hz, 1H, H5), 6.74 (dd, J ) 8.6, 7.0 Hz, 1H, H6), 6.52 (d,
J ) 3.3 Hz, 1H, H3), 6.18 (dd, J ) 3.2, 2.4 Hz, 1H, H3′), 6.11 (dd, J
) 3.2, 2.4 Hz, 1H, H3′′), 5.40 (d, J ) 3.3 Hz, 1H, H2), 5.16 (dd, J )
3.2, 2.4 Hz, 1H, H2′), 5.03 (dd, J ) 3.2, 2.4 Hz, 1H, H2′′), 1.55 (s,
3H, Me2C), 1.33 (s, 3H, Me2C), -0.12 (s, 3H, Me-Hf), -1.20 (s, 3H,
Me-Hf). 13C NMR (75 MHz, C6D6): δ 126.5, 126.0, 124.3, 123.9,
123.7, 118.8, 116.6, 114.1, 113.0, 112.3, 103.3, 103.2, 102.7, 39.3,
38.4, 37.9, 26.4, 25.4.
Acknowledgment. The authors would like to thank the
Natural Sciences and Engineering Research Council and Nova
Chemicals Ltd. of Canada, and the University of Akron for
financial support of this work. M.N. wishes to thank CNPq
(Conseil Nacional de Pesquisa - Brazil) under the program
PROFIX for providing a scholarship.
Preparation of Me2Si(Ind)CpHfMe2 (4). A suspension of Me2Si-
(Ind)CpHf(NMe2)2 (1.206 g, 2.4 mmol) in 20 mL of hexane was cooled
to -35 °C, and neat AlMe3 (0.69 g, 9.6 mmol) at -35 °C was added
to the suspension while stirring. The solution was warmed to ambient
temperature and stirred overnight. The reaction mixture was dried to
provide crude product, which was recrystallized in hexane at -35 °C
to provide spectroscopically pure 4 as an off-white, crystalline powder
Supporting Information Available: Observed and calculated
pentad distributions and 1D and 2D NMR spectra of ion-pairs
5 and 6 (PDF). This material is available free of charge via the
Internet at http://pubs.acs.org.
1
in >95% yield. H NMR (300 MHz, C6D6): δ 7.55 (d, J ) 8.6 Hz,
1H, H7), 7.16 (m, 1H, H4), 7.12 (m, 1H, H5), 6.86 (m, 1H, H6), 6.82
(dd, J ) 3.0 Hz, 1H, H3), 6.47 (dd, J ) 1.7 Hz, 1H, H3′), 6.39 (dd, J
) 1.9 Hz, 1H, H3′′), 5.59 (dd, J ) 3.0 Hz, 1H, H2), 5.37 (dd, J ) 2.4
Hz, 1H, H2′), 5.30 (dd, J ) 2.4 Hz, 1H, H2′′), 0.41 (s, 3H, Me2Si),
JA0207706
9
J. AM. CHEM. SOC. VOL. 125, NO. 26, 2003 7941