Ethylene and propylene polymerization by cationic monocyclopentadienyl titanium catalysts containing the weakly coordinating anion [B(C6F5)4]-

Article abstract of DOI:10.1016/S0022-328X(98)01209-1

The compounds Cp*TiMe₂E (Cp=η⁵-C₅Me₅; E=Me, C₆F₅, OC₆F₅, Cl) react with trityl tetrakis(perfluorophenyl)borate, [Ph₃C][B(C₆F₅)₄], to form the thermally unstable dititanium complexes [(Cp*TiMeE)₂(μ-Me)][B(C₆F₅) ₄], all of which behave as sources of the highly electrophilic species [Cp*TiMeE]⁺. An investigation of the activities of these [B(C₆F₅)₄]^- salts as ethylene and propylene polymerization catalysts shows that they are more active than the analogous compounds Cp*TiMeE(μ-Me)B(C₆F₅)₃, as anticipated since [B(C₆F₅)₄]^- is a poorer ligand than is [BMe(C₆F₅)₃]^-. However, contrary to current perceived wisdom, substitution of a methyl ligand of [Cp*TiMe₂]⁺ by the more electron withdrawing C₆F₅, OC₆F₅ and Cl ligands in these monocyclopentadienyl systems does not generally result in catalysts exhibiting lower activities and producing lower molecular weight polymers. An EPR study of the Cp*TiMe₃/[Ph₃C][B(C₆F₅) ₄] system in chlorobenzene at room temperature indicates that <0.01% of the titanium is present occasionally during polymerization as a complex of titanium(III), suggesting that a contribution to the catalytic processes by titanium(III) species is unlikely.

Full text of DOI:10.1016/S0022-328X(98)01209-1

Journal of Organometallic Chemistry 579 (1999) 106–113
Ethylene and propylene polymerization by cationic
monocyclopentadienyl titanium catalysts containing the weakly
coordinating anion [B(C6F5)4]−
Sean W. Ewart, Mark J. Sarsfield, Edan F. Williams, Michael C. Baird *
Department of Chemistry, Queen’s Uni6ersity, Kingston, Ontario K7L 3N6, Canada
Received 11 September 1998
Abstract
The compounds Cp*TiMe2E (Cp*=h5-C5Me5; E=Me, C6F5, OC6F5, Cl) react with trityl tetrakis(perfluorophenyl)borate,
[Ph3C][B(C6F5)4], to form the thermally unstable dititanium complexes [(Cp*TiMeE)2(m-Me)][B(C6F5)4], all of which behave as
sources of the highly electrophilic species [Cp*TiMeE]+. An investigation of the activities of these [B(C6F5)4]− salts as ethylene
and propylene polymerization catalysts shows that they are more active than the analogous compounds Cp*TiMeE(m-
Me)B(C6F5)3, as anticipated since [B(C6F5)4]− is a poorer ligand than is [BMe(C6F5)3]−. However, contrary to current perceived
wisdom, substitution of a methyl ligand of [Cp*TiMe2]+ by the more electron withdrawing C6F5, OC6F5 and Cl ligands in these
monocyclopentadienyl systems does not generally result in catalysts exhibiting lower activities and producing lower molecular
weight polymers. An EPR study of the Cp*TiMe3/[Ph3C][B(C6F5)4] system in chlorobenzene at room temperature indicates that
B0.01% of the titanium is present occasionally during polymerization as a complex of titanium(III), suggesting that a
contribution to the catalytic processes by titanium(III) species is unlikely. © 1999 Elsevier Science S.A. All rights reserved.
Keywords: Ziegler–Natta catalysis; Polyolefins; Titanium
1. Introduction
Similar species are apparently formed on treating for
example metallocene dihalides with the oligomeric
methylalumoxane (MAO), and commercially viable
processes are based on use of this co-catalyst, although
research with B(C6F5)3 has perhaps provided greater
insight into polymerization mechanisms.
Relatively recent work has also focused on the re-
lated monocyclopentadienyl precursors Cp%TiR3 [3],
which may be readily converted to the active, 10-elec-
tron species [Cp%TiR2]+. These monocyclopentadienyl
catalysts are electronically less saturated and sterically
less hindered than their metallocene counterparts, and
have been used to polymerize a wide variety of olefins
with substantial success. We have previously shown
that the precursor Cp*TiMe3, when reacted in a 1:1
ratio with the strong Lewis acid B(C6F5)3, forms exclu-
sively the methyl bridged species Cp*TiMe2(m-
Me)B(C6F5)3 (1; Eq. (2)) [4].
There have in recent years been numerous investiga-
tions into the use of Group 4 organometallic complexes
as homogenous catalysts for the polymerization of
olefins [1]. The most studied and successful of the
precursor compounds have been metallocenes of the
type Cp%MR2 (M=Ti, Zr, Hf; R=alkyl; Cp%=substi-
2
tuted cyclopentadienyl), which undergo alkyl abstrac-
tion on reaction with Lewis acid co-catalysts, such as
the borane B(C6F5)3 (Eq. (1)) to give the formally
14-electron, cationic species [Cp%MR]+ [2], which are
2
believed to be the actual catalysts.
Cp2%MR2+B(C6F5)3“[Cp%MR]+ +[BR(C6F5)3]−
2
(1)
* Corresponding author.
0022-328X/99/$ - see front matter © 1999 Elsevier Science S.A. All rights reserved.
PII: S0022-328X(98)01209-1

S.W. Ewart et al. / Journal of Organometallic Chemistry 581 (1999) 106–113
107
type [Cp*TiMeE]+ (E=Cl, C6F5, OC6F5), similar to
the above-mentioned [Cp*TiMe2]+. However, although
the compounds Cp*TiMeE(m-Me)B(C6F5)3 containing
the more electron-withdrawing ligands E were expected
to be the better catalysts, the anticipated results were
only partially achieved. Although higher molecular
weight polypropylenes of low polydispersities (Mw/Mn
1.3–1.9) were formed at −78°C when E=Cl, C6F5,
OC6F5 than when E=Me, the amounts of polymers
formed were actually comparable or lower, apparently
because the greater Lewis acidity of the cationic species
[Cp*TiMeE]+ resulted in stronger coordination of the
[BMe(C6F5)3]− anion. It was in fact possible to show
that the compounds containing the more electron with-
drawing ligands coordinate the methyl borate anion
much more strongly, and thus the anion competes with
the monomer for the vacant site, reducing the catalytic
activity of the resulting system [6]. Rather similar ion
pairing effects have been found for metallocene systems
[2a], and much new research currently focuses on the
design of new co-catalysts that can provide a more
poorly coordinating counterion for more active cata-
lytic systems [2].
Cp*TiMe3+B(C6F5)3“Cp*TiMe2(m-Me)B(C6F5)3
(1)
(2)
The latter behaves as a source of the active species
[Cp%TiMe2]+ [4], which has been utilized as an initiator
for both Ziegler–Natta and carbocationic polymeriza-
tions of a variety of olefins [5].
Since cationic complexes of both the mono- and the
dicyclopentadienyl series generally form more active
polymerization catalysts than do neutral species [1,5d],
we have earlier hypothesized that substitution of one of
the methyl groups of [Cp*TiMe2]+ by more electron
withdrawing ligands could result in the formation of
more active initiators [6]. This expectation is apparently
contrary to at least some findings for metallocene sys-
tems [7], since, for instance, incorporation of electron-
With these developments in mind, we have endeav-
ored to synthesize
a series of new compounds
[Cp*TiMeE]X (E=Cl, C6F5, OC6F5; X=B(C6F5)4), in
which the counteranion B(C6F5)4−, for steric or elec-
tronic reasons, cannot coordinate strongly and thus
induce deactivation. In this publication we describe the
use of the trityl borate co-catalyst [Ph3C][B(C6F5)4],
which normally readily abstracts methyl anions to form
cationic complexes containing the relatively poorly co-
ordinating counteranion [B(C6F5)4]− [2a]. The four sys-
tems Cp*TiMe2E/[Ph3C][B(C6F5)4] all behave as good
olefin polymerization catalysts with ethylene and
propylene, the polymerization activities correlating with
the nature of E as indicated in part by complementary
solution dynamics study of these systems. We also
describe unsuccessful attempts to synthesize similar cat-
alysts utilizing the very bulky borane tris(2,2%,2%%-pe-
rfluorobiphenyl)borane, B(C12F9)3, which is expected to
react with Cp*TiMe2E to form the poorly coordinating
anion [BMe(C12F9)3]− [2i].
withdrawing groups
X on the indenyl rings of
metallocenes of the type (h5-5,6-X2C9H5)2ZrCl2 (2)
have been found to result in both decreased activity and
reduced polymer molecular weights [7c]. While these
results seem to imply that electron-withdrawing ring
substituents increase the rates of termination and/or
chain transfer relative to the rates of initiation and
chain propagation, in fact the reactions are expected to
be affected unpredictably by ligand substitution, and
the reasons for the observed deactivation were not
assessed in detail [7c]. Indeed, it has been shown that
enhanced Lewis acidity of the metal can result in
stronger binding of the counter anion [2], which could
result in both inhibition of monomer coordination and
more facile termination of polymer growth via olefin
ligand displacement. Thus, the effects of increased
metal ion Lewis acid strength on catalytic activity ar-
guably remain unclear.
In order to test the hypothesis concerning the effects
of replacing
Me)B(C6F5)3 by more electronegative ligands, we have
previously synthesized the chiral compounds
a
methyl group of Cp*TiMe2(m-
2. Results and discussion
The reaction of Cp*TiMe3 with one molar equivalent
of [Ph3C][B(C6F5)4] occurred cleanly at −50°C in
CD2Cl2 on an NMR scale to form only the thermally
unstable, dititanium compound [(Cp*TiMe2)2(m-
Me)][B(C6F5)4] (4a). The reaction involves abstraction
of a methyl ligand from half of the available Cp*TiMe3
by trityl cation, followed by preferential coordination
of the remaining Cp*TiMe3 at the vacant site produced
(Eq. (3)).
Cp*TiMeE(m-Me)B(C6F5)3 (3; E=C6F5, OC6F5, Cl),
for which the electron-withdrawing powers of the lig-
ands E in the precursor compounds appear to decrease
in the order C6F5, Cl\OC6F5\Me [6b,c]. All of these
compounds polymerize ethylene to high molecular
weight polyethylene and propylene to atactic, elas-
tomeric polypropylene, the major catalytically active
species in all cases being titanium(IV) complexes of the

108
S.W. Ewart et al. / Journal of Organometallic Chemistry 581 (1999) 106–113
2Cp*TiMe3+2[Ph3C][B(C6F5)4]
“[(Cp*TiMe2)2(m-Me)][B(C6F5)4] (4a)
+[Ph3C][B(C6F5)4]+Ph3CMe
plex [Cp*TiMe2]+ is believed to be the active species
that coordinates monomer and thus initiates the poly-
merization process. The equilibrium of Eq. (5) must lie
well to the left, however, as the neutral Cp*TiMe3 is
not available in the absence of monomer to react with
the remaining [Ph3C][B(C6F5)4], even upon warming to
0°C. Compound 4a can also be synthesized by the
reaction of Cp*TiMe3 with a half equivalent of
[Ph3C][B(C6F5)4], yielding 4a alone with no
[Ph3C][B(C6F5)4] remaining in solution. Interestingly, as
will be shown below, the full equivalent of co-catalyst is
needed to obtain optimal polymerization results.
Our observation that the trityl cation of
[Ph3C][B(C6F5)4] is apparently less effective at removing
a methyl carbanion from Cp*TiMe3 than is the neutral
borane B(C6F5)3 seems counterintuitive [2j]. However, a
factor increasing the spontaneity of the chemistry of
Eq. (1) is the ability of the methylborate counteranion,
[BMe(C6F5)3]−, to coordinate to the titanium in 1.
Although the type of non-classical bonding involved
might be expected to be weak, in fact the
[BMe(C6F5)3]− has also been shown to be a better
ligand in metallocene systems [2] than is [B(C6F5)4]−
and as is shown below, the same is true for the type of
compounds under consideration here. Thus, the driving
force for the methyl transfer reaction of Cp*TiMe3 with
[Ph3C][B(C6F5)4] is lowered in large measure by the
inability of the [B(C6F5)4]− to compete with Cp*TiMe3
for the inner coordination sphere of the [Cp*TiMe2]+
ion.
Our observation that the reaction of Cp*TiMe3 with
[Ph3C][B(C6F5)4] yields solely the dititanium species 4a
is of some interest in that it shows that previous
assumptions [3k] concerning the chemistry of Eq. (3)
may not apply. While attempting to rationalize the
appearance of complex EPR resonances in reaction
mixtures containing equimolar amounts of Cp*TiMe3
and [Ph3C][B(C6F5)4] in chlorobenzene, Grassi et al.
[3k] presumed that [Cp*TiMe2]+ was the dominant
species formed and that this cationic complex subse-
quently undergoes reduction to an EPR-active but oth-
erwise uncharacterized titanium(III) species. In fact, as
we have now demonstrated, the major species present
(3)
As indicated by Eq. (3), only one half of the
[Ph3C][B(C6F5)4] reacts, leaving a half equivalent re-
maining in solution and thus resonances of both the
trityl cation and 1,1,1-triphenylethane were observed in
the 1H-NMR spectrum of the reaction mixture. No
further reaction ensues, although our polymerization
results imply that [(Cp*TiMe2)2(m-Me)]+ reacts further
with [Ph3C][B(C6F5)4] in the presence of olefin (see
below).
Although the dititanium product 4a is thermally too
labile to isolate, a compound containing the same
cationic complex has been previously synthesized via
the reaction of Cp*TiMe3 and the borane B(C6F5)3 in a
2:1 ratio (Eq. (4)) [4c].
2Cp*TiMe3+B(C6F5)3
“[(Cp*TiMe2)2(m-Me)][BMe(C6F5)3]
(4)
As before, the unstable cation is readily identified on
1
the basis of its H-NMR spectrum which exhibits reso-
nances attributable to the three methyl proton environ-
ments, at l 2.06 (Cp*), 1.46 (TiꢀMe) and 0.15 (m-Me).
Interestingly, as with [(Cp*TiMe2)2(m-Me)][BMe(C6F5)3]
[4c], spin saturation transfer experiments showed that
irradiation of the terminal methyl resonance of 4a at l
1.46 results in the appearance of an otherwise vanish-
ingly weak resonance at l 0.61, assigned to the TiꢀMe
group of free Cp*TiMe3, and suggesting the equi-
librium shown in Eq. (5). A complementary experiment
showed that the Cp* resonances of 4a (l 2.06) and
Cp*TiMe3 (l 1.88) also undergo exchange.
[(Cp*TiMe2)2(m-Me)]+ vCp*TiMe3+[Cp*TiMe2]+
(5)
Metallocene complexes containing methyl groups bridg-
ing two Group 4 metals have been reported previously
[2d,8] and provide precedents for the structure (4)
proposed.
in
methylene
chloride
is
[(Cp*TiMe2)2(m-
Me)][B(C6F5)4], which is quite stable with respect to
reduction. Indeed, on monitoring the reaction of
Cp*TiMe3 with [Ph3C][B(C6F5)4] under the same condi-
tions as previously [3k], i.e. at room temperature in
chlorobenzene by EPR spectroscopy, we observe only a
weak doublet at g=1.992, exhibiting a hyperfine cou-
pling constant of 8.2 G. This finding is similar to both
the type of titanium(III) hydride reported from the
reaction of CpTi(OBu)3 with MAO [3c] and the very
weak doublet we have previously observed in reaction
mixtures containing equimolar amounts of Cp*TiMe3
and B(C6F5)3 in chlorobenzene (g=1.994, A=8.4 G)
[6c].
The equilibrium shown in Eq. (5) is vital to the
process of olefin polymerization since the cationic com-

S.W. Ewart et al. / Journal of Organometallic Chemistry 581 (1999) 106–113
109
Integration using standard TEMPO solutions showed
that the titanium(III) species being observed amounted
to only ca. 0.01% of the total titanium in solution, and
thus we cannot confirm the previous report that this
same system in chlorobenzene contains relatively high
proportions of Ti(III)-containing species [3k]. Working
under scrupulously anhydrous and anaerobic condi-
tions, we observe neither the stronger (by ca. 102) trityl
and titanium-centered resonances nor the resonances
attributed to putative alkoxytitanium(III) byproduct
species that have been previously reported [3k]. In view
of the highly reproducible and definitive NMR and
EPR spectroscopic evidence presented herein, however,
we feel that we have defined well the reaction between
Cp*TiMe3 and [Ph3C][B(C6F5)4].
Reactions of Cp*TiMe2C6F5 and Cp*TiMe2Cl with
[Ph3C][B(C6F5)4] do occur on warming to ca. 10°C,
however, and appear to form the dititanium species
[(Cp*TiMeC6F5)2(m-Me)][B(C6F5)4]
(4c)
and
[(Cp*TiMeCl)2(m-Me)][B(C6F5)4] (4d), respectively. Un-
fortunately, unambiguous identification of the products
is difficult as the release of methane (l 0.2) and the
appearance of new resonances in the spectra show that
thermal decomposition occurs at these elevated temper-
atures. Evidence that species 4c and 4d are formed,
however, is that only
a half equivalent of the
[Ph3C][B(C6F5)4] reacts, analogous to the formation of
1
4a and 4b, and that the H- and 13C{1H}-NMR spectra
of 4d exhibit pairs of TiꢀMe resonances, consistent with
the formation of chiral, diasteromeric products al-
though in this case they appear to formed in a ca. 3:2
ratio. Double irradiation experiments performed on 4d
gave no evidence of resonances attributable to
Cp*TiMe2C6F5 or [Cp*TiMeC6F5]+, suggesting there
was no observable equilibration between 4d and its
precursors, as is the case for 4a and 4b. This is perhaps
not a surprising result as we have previously shown that
Cp*TiMeC6F5(m-Me)B(C6F5)3 does not dissociate to
Cp*TiMe2OC6F5 reacts with [Ph3C][B(C6F5)4] in a
manner similar to Cp*TiMe3 at −50°C, forming the
interesting chiral complex [(Cp*TiMeOC6F5)2(m-
Me)][B(C6F5)4] (4b), which has also previously been
synthesized via the reaction of Cp*TiMe2OC6F5 with a
half equivalent of B(C6F5)3 [6b]. Like 4a, complex 4b is
thermally unstable and has been characterized by NMR
spectroscopy (1H, 13C{1H}, 19F). Unlike 4a, however,
1
4b contains two chiral titanium centers and the H- and
detectable
amounts
of
Cp*TiMeC6F5+
and
13C{1H}-NMR spectra are characterized by doubling of
many of the resonances because of the presence of
diastereomers in approximately equal amounts. Thus,
MeB(C6F5)3−, contrary to the behavior of the
analogous Cp*TiMe2(m-Me)B(C6F5)3 and Cp*TiMe-
OC6F5(m-Me)B(C6F5)3 [6b]. The more electron deficient
titanium center in Cp*TiMeC6F5+ acts to hold its neu-
tral counterpart much more strongly, not allowing dis-
sociation to occur. The variable temperature 19F spectra
of 4d is also of interest as the broadness of the ortho
and meta resonances changes with temperature, sharp-
ening as the temperature increases and suggesting hin-
dered rotation of the C6F5 ring, as was reported for
Cp*TiMeC6F5(m-Me)B(C6F5)3 [6b].
1
the H-NMR spectrum of 4b exhibits two TiMe reso-
nances (l 1.50, 1.47), while the 13C{1H}-NMR spec-
trum exhibits doubling of each of the C5Me5 (l 132.0,
131.9), C5Me5 (l 12.3, 12.2) and TiMe (79.7, 79.3)
resonances. Spin saturation experiments have, interest-
ingly, shown 4b to be in equilibrium with Cp*Ti-
Me2OC6F5 and [Cp*TiMeOC6F5][B(C6F5)4] in much
the same manner as described for 4a in Eq. (5). Irradi-
ation of the Cp* peak in 4b at l 2.01 gives rise to a new
peak at l 1.86, attributed to a very low concentration
of Cp*TiMe2OC6F5.
2.1. Polymerization results
The products of the reactions of Cp*TiMe2C6F5 and
Cp*TiMe2Cl with [Ph3C][B(C6F5)4] have proven to be
much more difficult to identify because they form only
at temperatures at which decomposition also occurs.
The Cl and C6F5 groups are more electron withdrawing
than the Me or OC6F5 groups [6b], apparently resulting
in more electron deficient titanium centers that hold the
anionic methyl ligands more tightly and thus inhibit
removal of a methyl carbanion. In any case, no reac-
tions occur between Cp*TiMe2C6F5 or Cp*TiMe2Cl
and [Ph3C][B(C6F5)4] at −50°C, a result in contrast
with the facile methyl abstraction reactions of
Cp*TiMe2C6F5 and Cp*TiMe2Cl with B(C6F5)3, which
occur readily at −50°C [6b]. Again, the neutral
B(C6F5)3 acts as a much more effective methyl abstrac-
tor than does the trityl cation. To our knowledge, these
precursors represent the only titanium methyl com-
pounds that exhibit such reluctance to react with the
trityl cation.
All four precursors Cp*TiMe3, Cp*TiMe2OC6F5,
Cp*TiMe2Cl and Cp*TiMe2C6F5 have previously been
shown to be highly effective catalysts for the polymer-
ization of ethylene and propylene when activated by
B(C6F5)3 co-catalyst (as in Eq. (1)) [6b]. As outlined
above, it was anticipated that the more electron defi-
cient catalysts derived from Cp*TiMe2C6F5 and
Cp*TiMe2Cl would be more active because of their
greater propensities to coordinate an olefin, the first
step in the polymerization mechanism. However, exper-
imental results showed that this was not the case as
these catalysts produced comparable amounts or less of
polymer, depending on the nature of E and the poly-
merization temperature.
Interestingly, however, the polypropylene formed at
−78°C in toluene by Cp*TiMeC6F5(m-Me)B(C6F5)3 ex-
hibited a significantly higher value of Mw (2.3×106;
Mw/Mn 1.7) than that formed under the same condi-

110
S.W. Ewart et al. / Journal of Organometallic Chemistry 581 (1999) 106–113
Table 1
those obtained when using the same four precursors
Data for ethylene polymerization
with the borane cocatalyst in Tables 1 and 2, where it
should be noted that the activities for the ethylene
polymerizations with the [Ph3C][B(C6F5)4] activated sys-
tems are in all cases lower limits. The polymerizations
were hindered by extremely viscous solutions that
formed, preventing effective monomer diffusion
through the solution. In addition, the insolubility of the
polyethylene in 1,2,4-trichlorobenzene at 140°C implies
that this material was both linear and of high molecular
weight.
Precursor
Activitiesa
Using B(C6F5)3
23
Using [Ph3C][B(C6F5)4]
Cp*TiMe3
Cp*TiMe2OC6F5 22
Cp*TiMe2C6F5
Cp*TiMe2Cl
30
23
24
24
8
16
a kg of polymer per (mol Ti atm h).
Interestingly, the molecular weights and polydispersi-
ties of the polypropylene obtained were in all cases
essentially identical to those prepared utilizing the bo-
rane as co-catalyst [6c], indicating that ion pairing has
little effect on the polypropylene molecular weights. As
with the polypropylene formed utilizing the borane as
co-catalyst [6c], the materials produced here were pri-
marily atactic (52–58% r dyads), and contained as
much as 18% 2,1 inserted regioerrors, as indicated by a
resonance at l 28.2 in the 13C{1H}-NMR spectra [6c].
These regioerrors are comparable with those observed
with soluble vanadium catalysts [9a] and heterogenized
monocyclopentadienyl titanium complexes [9b]. How-
ever, by using [Ph3C][B(C6F5)4] as a cocatalyst in lieu of
B(C6F5)3, we appear to have increased the polymeriza-
tion activities most notably for the two precursors,
Cp*TiMe2C6F5 and Cp*TiMe2Cl, with the strongly
electron withdrawing ligands. Although strict compari-
sons arguably should not be made in view of the
differing exotherms for the different reactions, it is
probably significant that polymerizations by the
Cp*TiMe2C6F5/[Ph3C][B(C6F5)4] system quickly be-
came much hotter than the others, consistent with
lower activities for the latter.
We are still left with the problem of trying to under-
stand how the neutral species remaining in solution
affect the polymerization activities. Upon initiation of
polymerization, the dititanium complexes 4a–d must
dissociate to allow monomer to coordinate. The neutral
precursors remaining in solution can then either react
with the excess [Ph3C][B(C6F5)4], also in solution, or
compete with the incoming olefins for the active site. In
order to study this further, we carried out some poly-
merization runs with only a half molar equivalent of
tions by Cp*TiMe2(m-Me)B(C6F5)3 (3.4×105; Mw/Mn
1.3) [6c]. Clearly the rate of chain propagation was not
decreased relative to those of chain termination/transfer
by the presence of the electron-withdrawing C6F5
group. Since the decreased activity was suspected to be
a result of stronger coordination of the [MeB(C6F5)3]−
anion to the titanium center competing with the incom-
ing monomer for the active coordination site, we have
now investigated the use of [Ph3C][B(C6F5)4] rather
than the borane because the trityl salt would incorpo-
rate
a
more poorly coordinating counteranion,
[B(C6F5)4]−, into the catalyst. Unfortunately, as is
shown above, the free active species [Cp*TiMeE]+
(E=C6F5, Cl) are not formed in solution with
[Ph3C][B(C6F5)4] at −78 °C, at which temperature high
molecular weight polypropylenes of narrow molecular
weight distributions are formed by clearly single site
catalysts, such as Cp*TiMeC6F5(m-Me)B(C6F5)3. The
neutral precursors Cp*TiMe2E compete extremely ef-
fectively for the active site of the cation of
[Cp*TiMeE]+[B(C6F5)4]−, no polymer was formed,
and the desired comparisons are impossible.
Polymerizations of ethylene and propylene by the
Cp*TiMe2E/[Ph3C][B(C6F5)4] systems were therefore of
necessity carried out at 25°C, at which temperature
some decomposition is apparent in both the B(C6F5)3
and the [Ph3C][B(C6F5)4] systems. Indeed, at this tem-
perature the B(C6F5)3 initiated catalysts give relatively
low molecular weight polypropylene (ca. 2×105) of
relatively high polydispersities (Mw/Mn ca. 5), consis-
tent with the presence of more than one catalytically
active species in solution [6c]. The activities obtained
with the [Ph3C][B(C6F5)4] co-catalyst are compared with
Table 2
Data for propylene polymerization
Precursor
Activitiesa using Mw (Mw/Mn)
Activitiesa using
[Ph3C][B(C6F5)4]
Mw (Mw/Mn) using
[Ph3C][B(C6F5)4]
Activitiesa using
12[Ph3C][B(C6F5)4]
B(C6F5)3
using B(C6F5)3
Cp*TiMe3
Cp*TiMe2OC6F5 160
Cp*TiMe2C6F5
Cp*TiMe2Cl
77
2.6×105 (5.2)
1.7×105 (4.9)
2.7×105 (3.9)
Bimodal
225
174
222
246
2.2×105 (4.5)
1.4×105 (3.7)
2.6×105 (3.9)
Bimodal
133
0
0
78
29
0
a kg of polymer per (mol Ti atm h).

S.W. Ewart et al. / Journal of Organometallic Chemistry 581 (1999) 106–113
111
[Ph3C][B(C6F5)4] for each precursor, conditions under
which NMR experiments demonstrate the formation of
the dititanium complexes 4a–d in the absence of unre-
acted [Ph3C][B(C6F5)4]. In all cases but Cp*TiMe3,
propylene polymerizations were quenched completely.
Some activity was observed with Cp*TiMe3, but less
than with a full equivalent of co-catalyst; this is proba-
bly due to the equilibrium between the bridged ditita-
nium species and the active species [Cp*TiMe2]+ for
this system. These results show clearly that the com-
pounds Cp*TiMe2E in solution compete very effectively
with propylene for the vacant sites of [Cp*TiMeE]+.
Furthermore, the decreased activities observed when
only a half equivalent of [Ph3C][B(C6F5)4] is reacted
with Cp*TiMe2E demonstrate clearly that the excess
[Ph3C][B(C6F5)4], which was originally remaining in
solution at the beginning of a polymerization run, does
eventually react with the neutral species obtained from
initiator dissociation, preventing it from recoordinating
to the active site. Therefore, we can conclude that after
initial induction periods, all of the precursors are con-
verted into the active species Cp*TiMeE+. It is also
interesting to note that for polymerization runs involv-
ing Cp*TiMe2C6F5 and Cp*TiMe2Cl, the induction
periods appear to be much longer, ca. 1 min, than for
polymerization runs involving Cp*TiMe3 and Cp*Ti-
Me2OC6F5 (ca. 10 s), measured by the time it takes for
the reaction to begin to evolve heat. This longer induc-
tion period is possibly a result of the lower dissociation
rate of compounds 4c and 4d, as indicated by the
double resonance experiments as well as the lower rates
of reaction between their neutral precursors and the
[Ph3C][B(C6F5)4] co-catalyst, as seen by variable tem-
perature NMR experiments.
spectra with respect to internal C6F6. GPC data were
obtained on Waters Model 150-C in 1,2,4-
a
trichlorobenzene solvent running at 145°C. EPR experi-
ments involving equimolar amounts of Cp*TiMe3 and
[Ph3C][B(C6F5)4] (0.01–0.06 mmol of each in 10 ml
solution) were run at 22°C on a Bruker R-B70 EPR
spectrometer at ca. 9.7 GHz with the field centered at
ca. 3400 G; the magnetic field was calibrated with
DPPH (g=2.0037), and standard solutions of TEMPO
(1×10−3–1×10−5 M) were used for quantitative esti-
mates of the concentrations of titanium(III) species
present. The compounds Cp*TiMe3 [10a], Cp*TiMe2Cl
[10b], Cp*TiMe2OC6F5 [6], Cp*TiMe2C6F5 [6],
B(C6F5)3 [11a,b], B(C12F9)3 [2i] and [Ph3C][B(C6F5)4]
[11c] were prepared by literature methods.
3.1. Polymerization experiments
In a typical polymerization run, a solution of 14 mg
(0.06 mmol) of titanium precursor in 25 ml toluene
saturated with monomer at 25°C was treated with a
solution of 56 mg (0.06 mmol) of [Ph3C][B(C6F5)4] in 10
ml toluene. Monomer was constantly bubbled through
the solution, which warmed somewhat for 15 min be-
fore cooling to nearly room temperature after 20 min.
For ethylene polymerizations, solutions became thick
and viscous as polyethylene precipitated. Polymeriza-
tion was continued for 20 min, and then 10 ml of
methanol were added to terminate polymerization. The
resulting polymers were collected by filtration, washed
with methanol to remove residual catalyst and dried in
vacuo.
Attempts to react Cp*TiMe3 with the sterically hin-
dered B(C12F9)3 failed at low temperatures, presumably
because of steric hindrance, and gave decomposition
products that were poor propylene polymerization cata-
lysts at room temperature. The compound
Cp*TiMe2C6F5 failed to react with B(C12F9)3 even at
30°C, and work with this borane, which does activate
metallocenes and some other monocyclopentadienyl
compounds [2i], was terminated.
3.2. In situ NMR studies of the reactions of
Cp*TiMe2E with [Ph3C][B(C6F5)4] to form
[(Cp*TiMeE)2(v-Me)][B(C6F5)4]
As the same basic procedure was followed in all
cases, we describe here only our investigation of
[(Cp*TiMe2)2(m-Me)][B(C6F5)4] (4a). In an NMR exper-
iment, 14 mg (0.06 mmol) of Cp*TiMe3 and 56 mg
(0.06 mmol) of [Ph3C][B(C6F5)4] were placed in an
NMR tube and dissolved in 0.5 ml CD2Cl2 at −78°C.
The resulting deep-red solution was then placed in the
NMR spectrometer probe at −50°C. 1H-NMR
(CD2Cl2) of 4a: l 2.06 (30H, C5Me5), 1.46 (12H, TiMe),
0.15 (3H, m-Me). 13C{1H}-NMR of 4a: l 129.1
(C5Me5), 12.7 (C5Me5), 79.1 (TiMe). 19F-NMR of 4a: l
−133.3 (m, 6F, ortho), −163.9 (m, 3F, para), −167.7
3. Experimental
All experiments were carried out using standard
Schlenk line techniques or a Vacuum Atmospheres
glovebox, dry, prepurified nitrogen atmosphere and
1
1
dried, thoroughly deoxygenated solvents. H-, 13C{1H}-
(m, 6F, meta). The H-NMR spectrum also exhibited
and 19F-NMR spectra were run using a Bruker AM 400
spectrometer operating at 400.14, 100.6 and 376.5
resonances attributable to the trityl cation at l 8.23 (br
m, 3H, para), 7.84 (br m, 6H, meta) and 7.65 (br m,
6H, ortho), and to 1,1,1-triphenylethane at l 7.27 (m,
9H, para, meta), 7.04 (m, 6H, ortho) and 2.14 (s, 3H,
Me).
1
MHz, respectively; H- and 13C{1H}-NMR spectra are
referenced with respect to internal TMS using residual
proton or carbon resonances from the solvents, 19F

112
S.W. Ewart et al. / Journal of Organometallic Chemistry 581 (1999) 106–113
1H-NMR
(CD2Cl2)
of
[(Cp*TiMeOC6F5)2(m-
latter function as sources of the highly electrophilic
species [Cp*TiMeE]+, and behave as more active olefin
polymerization catalysts than do the analogous com-
pounds Cp*TiMeE(m-Me)B(C6F5)3 since [B(C6F5)4]− is
a poorer ligand than is [BMe(C6F5)3]−. Interestingly,
substitution of a methyl ligand of [Cp*TiMe2]+ by the
more electron withdrawing C6F5, OC6F5 and Cl ligands
does not generally result in catalysts exhibiting lower
activities and producing lower molecular weight poly-
mers. Indeed, it seems that more electron withdrawing
ligands on the titanium result in better catalysis.
Me)][B(C6F5)4] (4b): l 2.01 (30H, C5Me5), 1.50, 1.47
(6H, TiMe), −0.31 (3H, m-Me). 13C{1H}-NMR of 4b:
l 132.0, 131.9 (C5Me5), 12.3, 12.2 (C5Me5), 79.7, 79.3
(TiMe). 19F-NMR of 4b: l −133.3 (m, 6F, o-F of
BC6F5), −159.4, −159.6 (m, 4F, o-F of TiOC6F5),
−163.9 (m, 3F, p-F of BC6F5), −163.6, −164.1 (m,
4F, m-F of TiOC6F5), −165.0 (m, 2F, p-F of
1
TiOC6F5), −167.7 (m, 6F, m-F of BꢀC6F5). The H-
NMR spectrum also exhibited resonances attributable
to the trityl cation at l 8.23 (br m, 3H, para), 7.84 (br
m, 6H, meta) and 7.65 (br m, 6H, ortho), and to
1,1,1-triphenylethane at l 7.27 (m, 9H, para, meta),
7.04 (m, 6H, ortho) and 2.14 (s, 3H, Me).
Acknowledgements
In similar attempts to synthesize [(Cp*TiMeC6F5)2(m-
Me)][B(C6F5)4] (4c) and [(Cp*TiꢀMeCl)2(m-Me)][B-
(C6F5)4] (4d), only resonances attributable to the start-
ing materials were observed at −78°C. The solutions
were therefore warmed to 10°C to give deep red colors,
The authors thank the Natural Sciences and Engi-
neering Research Council of Canada (Research and
Strategic Grants to M.C.B. and Graduate Scholarship
to S.W.E.) and Alcan (Graduate Scholarship to
S.W.E.) for financial support. They also thanks Profes-
sor J.K.S. Wan for allowing them time on the Bruker
RB-70 EPR spectrometer.
1
but the H- and 13C{1H}-NMR spectra were very com-
plicated, indicating extensive decomposition. Only par-
tial assignments of the NMR spectra of 4c are possible.
1H-NMR (CD2Cl2) of [(Cp*TiMeC6F5)2(m-Me)][B
(C6F5)4] (4c): l 2.16 (30H, C5Me5), 2.34, 2.35 (6H,
TiMe). 13C{1H}-NMR of 4c: l 139.2, 139.7 (C5Me5),
14.6 (C5Me5), 79.7, 79.3 (TiMe). Resonances attributed
the trityl cation and 1,1,1-triphenylethane (see above)
References
[1] For recent reviews see (a) H.H. Brintzinger, D. Fischer, R.
Mu¨lhaupt, B. Rieger, R. Waymouth, Angew. Chem. Int. Ed.
Engl. 34 (1995) 1143. (b) M. Bochmann, J. Chem. Soc. Dalton
Trans. (1996) 255. (c) W. Kaminsky, M. Arndt, Adv. Polym. Sci.
127 (1997) 143. (d) P.C. Mo¨hring, N.J. Coville, J. Organomet.
Chem. 479 (1994) 1.
1
were observed in a 1:1 ratio in the H-NMR spectra of
4c and 4d, as was a resonance attributed to methane (l
0.2).
[2] (a) L. Jia, X. Yang, C.L. Stern, T.J. Marks, Organometallics 16
(1997) 842. (b) Y. Kim, B.H. Koo, Y. Do, J. Organomet. Chem.
527 (1997) 155. (c) C. Sishta, R.M. Hathorn, T.J. Marks, J. Am.
Chem. Soc. 114 (1992) 1112. (d) M.A. Giardello, M.S. Eisen,
C.L. Stern, T.J. Marks, J. Am. Chem. Soc. 117 (1995) 12 114. (e)
X. Yang, C.L. Stern, T.J. Marks, J. Am. Chem. Soc. 116 (1994)
10 015. (f) P.A. Deck, T.J. Marks, J. Am. Chem. Soc. 117 (1995)
6128. (g) A.R. Siedle, R.A. Newmark, J. Organomet. Chem. 497
(1995) 119. (h) Y.-X. Chen, C.L. Stern, T.J. Marks, J. Am.
Chem. Soc. 119 (1997) 2582. (i) Y.-X. Chen, C.L. Stern, S. Yang,
T.J. Marks, J. Am. Chem. Soc. 119 (1997) 12 451. (j) P.A. Deck,
C.L. Beswick, T.J. Marks, J. Am. Chem. Soc. 120 (1998) 1772.
(k) Y.-X. Chen, M.V. Metz, L. Liting, C.L. Stern, T.J. Marks, J.
Am. Chem. Soc. 120 (1998) 6287.
[3] For recent studies, see (a) C. Pellecchia, P. Longo, A. Proto, A.
Zambelli, Makromol. Chem. Rapid Commun. 13 (1992) 265. (b)
C. Pellecchia, P. Longo, A. Proto, A. Zambelli, Makromol.
Chem. Rapid Commun. 13 (1992) 277. (c) J.C.W. Chien, Z.
Salajka, S. Dong, Macromolecules 25 (1992) 3199. (d) A. Kucht,
H. Kucht, S. Barry, J.C.W. Chien, M.D. Rausch, Organometal-
lics 12 (1993) 3075. (e) T.E. Ready, R.O. Day, J.C.W. Chien,
M.D. Rausch, Macromolecules 26 (1993) 5822. (f) A. Zambelli,
C. Pellecchia, L. Oliva, P. Longo, A. Grassi, A. Makromol.
Chem. 192 (1991) 223. (g) H. Kucht, A. Kucht, J.C.W. Chien,
M.D. Rausch, Appl. Organomet. Chem. 8 (1994) 393. (h) C.
Pellecchia, A. Immirzi, A. Grassi, A. Zambelli, Organometallics
12 (1993) 4473. (i) J.C. Flores, J.C.W. Chien, M.D. Rausch,
Organometallics 13 (1994) 4140. (j) J.C. Flores, J.S. Wood,
J.C.W. Chien, M.D. Rausch, Organometallics 15 (1996) 4944.
(k) A. Grassi, A. Zambelli, F. Laschi, Organometallics 15 (1996)
3.3. In situ NMR studies of the reactions of
Cp*TiMe2E with B(C12F9)3
In NMR experiments similar to the above, addition
of an equimolar amount of B(C12F9)3 to solutions of
the compounds Cp*TiMe3 and Cp*TiMe2C6F5 in
CD2Cl2 at −50°C was shown to result in no reaction.
In the case of Cp*TiMe3, warming to 10°C did result in
apparent methyl abstraction, but extensive decomposi-
tion of the product also occurred as the resulting NMR
spectrum was very complicated and indicated the evolu-
tion of methane. An attempted propylene polymeriza-
tion experiment using this system in toluene at 30°C
yielded only a very small amount of polypropylene.
There was no reaction between Cp*TiMe2C6F5 and
B(C12F9)3 even at 30°C, and this system failed to poly-
merize propylene at 50°C in toluene.
4. Summary
The compounds Cp*TiMe2E (E=Me, C6F5, OC6F5,
Cl) react with trityl tetrakis(perfluorophenyl)borate,
[Ph3C][B(C6F5)4], to form the thermally unstable ditita-
nium complexes [(Cp*TiMeE)2(m-Me)][B(C6F5)4]. The

S.W. Ewart et al. / Journal of Organometallic Chemistry 581 (1999) 106–113
113
480. (l) J. Sassmannshausen, M. Bochmann, J. Organomet.
Chem. 548 (1997) 23. (m) C. Pellecchia, D. Pappalardo, L. Oliva,
A. Zambelli, J. Am. Chem. Soc. 117 (1995) 6593. (n) Y.-X.
Chen, P.-F. Fu, C.L. Stern, T.J. Marks, Organometallics 16
(1997) 5958.
with B(C6F5)3 gives a mixture of compounds since two strong
and several weak Cp* methyl resonances are observed. Although
both methyl and chloro abstraction may be occurring, the poly-
merization activity of the mixture requires that there be present
some Cp*TiMeCl(m-Me)B(C6F5)3, the TiꢀMe resonance of which
is obscured by the Cp* resonances.
[4] (a) D.J. Gillis, M.-J. Tudoret, M.C. Baird, J. Am. Chem. Soc.
115 (1993) 2543. (b) D.J. Gillis, R. Quyoum, M.-J. Tudoret, Q.
Wang, D. Jeremic, A.W. Roszak, M.C. Baird, Organometallics
15 (1996) 3600. (c) Q. Wang, R. Quyoum, D.J. Gillis, D.
Jeremic, M.C. Baird, J. Organomet. Chem. 527 (1997) 7.
[5] (a) F. Barsan, M.C. Baird, J. Chem. Soc. Chem. Commun.
(1995) 1065. (b) D. Jeremic, Q. Wang, R. Quyoum, M.C. Baird,
J. Organomet. Chem. 497 (1995) 143. (c) Q. Wang, M.C. Baird,
Macromolecules 28 (1995) 8021. (d) Q. Wang, R. Quyoum, D.J.
Gillis, M.-J. Tudoret, D. Jeremic, B.K. Hunter, M.C. Baird,
Organometallics 15 (1996) 693. (e) M.C. Murray, M.C. Baird, J.
Mol. Cat. 128 (1998) 1.
[6] (a) T.L. Tremblay, S.W. Ewart, M.J. Sarsfield, M.C. Baird, J.
Chem. Soc. Chem. Comm. (1997) 831. (b) M.J. Sarsfield, S.W.
Ewart, T.L. Tremblay, A.W. Roszak, M.C. Baird, J. Chem. Soc.
Dalton Trans. (1997) 3097. (c) S.W. Ewart, M.J. Sarsfield, D.
Jeremic, T.L. Tremblay, E.F. Williams, M.C. Baird,
Organometallics 17 (1998) 1502. (d) S.W. Ewart, M.J. Sarsfield,
M.C. Baird, unpublished results. The reaction of Cp*TiMe2Cl
[7] (a) A.-L. Mogstad, R.M. Waymouth, Macromolecules 25 (1992)
2282. (b) N. Piccolrovazzi, P. Pino, G. Consiglio, A. Sironi, M.
Moret, Organometallics 9 (1990) 3098. (c) I.-M. Lee, W.J. Gau-
thier, J.M. Ball, B. Iyengar, S. Collins, Organometallics 11
(1992) 2115.
[8] M. Bochmann, S.J. Lancaster, Angew. Chem. Int. Ed. Eng. 33
(1994) 1634.
[9] (a) Y. Doi, Macromolecules 12 (1979) 248. (b) K. Soga, D.H.
Lee, Makromol. Chem. 193 (1992) 1687.
[10] (a) M. Mena, P. Royo, R. Serrano, M.A. Pellinghelli, A. Tiripic-
chio, Organometallics 8 (1989) 476. (b) A. Mart´ın, M. Mena,
M.A. Pellinghelli, P. Royo, R. Serrano, A. Tiripicchio, J. Chem.
Soc. Dalton Trans. (1993) 2117.
[11] (a) A.G. Massey, A.J. Park, J. Organomet. Chem. 2 (1964) 245.
(b) R. Uson, A. Laguna, Inorg. Synth. 21 (1982) 72. (c) J.C.W.
Chien, W.-M. Tsai, M.D. Rausch, J. Am. Chem. Soc. 113 (1991)
8570.
.