Ethylene polymerization with dimeric zirconium and hafnium silsesquioxane complexes

Article abstract of DOI:10.1021/om980687k

Treatment of the silanol (c-C₅H₉)₇Si₈O₁₂(OH) with Cp″Ti(CH₂Ph)₃ (Cp″ = 1,3-C₅H₃(SiMe₃)₂) or TiCl₄ selectively affords the mono(silsesquioxane) complexes Cp″[(c-C₅H₉)₇Si₈O ₁₃]Ti(CH₂Ph)₂ (1) and [(c-C₅H₉)₇Si₈O ₁₃]TiCl₃ (2), respectively, while with M(CH₂Ph)₄ (M = Ti, Zr, Hf) mixtures of products were obtained. When the disilanol (c-C₅H₉)₇Si₇O₉(OSiMe ₃)(OH)₂ is reacted with M(CH₂Ph)₄ (M = Ti, Zr), the bis(silsesquioxane) complexes [(c-C₅H₉)₇Si₇O ₁₁(OSiMe₃)]₂M (M = Ti (3), Zr (4), Zr-2THF (5)) are formed exclusively. With (PhCH₂)₂ZrCl₂OEt₂ as precursor, the mono(silsesquioxane) complex [(c-C₅H₉)₇Si₇On(OSiMe ₃)]ZrCl₂·2THF (6) can be isolated. M(CH₂Ph)₄ (M = Ti, Zr, Hf) reacts smoothly with the tris(silanol) (c-C₅H₉)₇Si₇O₉(OH) ₃ (III), giving the metallasilsesquioxane benzyl species, {[(c-C₅H₉)₇Si₇O ₁₂]MCH₂Ph}n (M = Ti, n = l (7); M = Zr, n = 2 (8); M = Hf, n = 2 (9)). Compounds 5 and 8 have been characterized by X-ray analysis. Dimer 8 consists of a zwitterionic-like structure with two electronically different metal sites. M - C bond hydrogenolysis of 8 and 9 affords the corresponding hydrides, which are active α-olefin hydrogenation catalysts. Without cocatalyst, the neutral dimers 8 and 9 are poor, though active ethylene polymerization catalysts (activity: (5 - 10) × 10³ g PE/(mol·h)). Addition of B(C₆F₅)₃ affords the cationic, mono(benzyl) complexes {[(c-C₅H₉)₇Si₇O₁₂] ₂M₂(CH₂Ph)}⁽⁺⁾ (M = Zr, Hf): single-site catalysts (activity: (2 8) × 10⁶ g PE/(mol · h)) that are considerably more active than the neutral 8 and 9. Whereas titanasilsesquioxanes 3 and 7 do not react with THF, the corresponding zirconasilsesquioxanes 4 and 8 form bis(THF) adducts, [(c-C₅H₉)₇Si₇O ₁₁(OSiMe₃)]₂Zr · 2THF (5) and [(c-C₅H₉)₇Si₇O ₁₂]ZrCH₂Ph · 2THF (10), which suggests that the titanium complexes are less electrophilic than the zirconium ones. Accordingly, the titanium complex 7 does not react with dihydrogen and is inactive in ethylene polymerization.

Full text of DOI:10.1021/om980687k

Organometallics 1998, 17, 5663-5673
5663
Eth ylen e P olym er iza tion w ith Dim er ic Zir con iu m a n d
Ha fn iu m Silsesqu ioxa n e Com p lexes
Robbert Duchateau,*,† Hendrikus C. L. Abbenhuis,† Rutger A. van Santen,†
Auke Meetsma,‡ Sven K.-H. Thiele,§ and Maurits F. H. van Tol§
Schuit Institute of Catalysis, Eindhoven University of Technology, P.O. Box 513, 5600 MB
Eindhoven, The Netherlands, University of Groningen, Nijenborgh 4, 9747 AG Groningen, The
Netherlands, and DSM Research B.V., P.O. Box 18, 6160 MD Geleen, The Netherlands
Received August 12, 1998
Treatment of the silanol (c-C5H9)7Si8O12(OH) with Cp′′Ti(CH2Ph)3 (Cp′′ ) 1,3-C5H3(SiMe3)2)
or TiCl4 selectively affords the mono(silsesquioxane) complexes Cp′′[(c-C5H9)7Si8O13]Ti(CH2-
Ph)2 (1) and [(c-C5H9)7Si8O13]TiCl3 (2), respectively, while with M(CH2Ph)4 (M ) Ti, Zr, Hf)
mixtures of products were obtained. When the disilanol (c-C5H9)7Si7O9(OSiMe3)(OH)2 is
reacted with M(CH2Ph)4 (M ) Ti, Zr), the bis(silsesquioxane) complexes [(c-C5H9)7Si7O11-
(OSiMe3)]2M (M ) Ti (3), Zr (4), Zr‚2THF (5)) are formed exclusively. With (PhCH2)2ZrCl2‚
OEt2 as precursor, the mono(silsesquioxane) complex [(c-C5H9)7Si7O11(OSiMe3)]ZrCl2‚2THF
(6) can be isolated. M(CH2Ph)4 (M ) Ti, Zr, Hf) reacts smoothly with the tris(silanol) (c-
C5H9)7Si7O9(OH)3 (III), giving the metallasilsesquioxane benzyl species, {[(c-C5H9)7Si7O12]MCH2-
Ph}n (M ) Ti, n ) 1 (7); M ) Zr, n ) 2 (8); M ) Hf, n ) 2 (9)). Compounds 5 and 8 have been
characterized by X-ray analysis. Dimer 8 consists of a zwitterionic-like structure with two
electronically different metal sites. M-C bond hydrogenolysis of 8 and 9 affords the
corresponding hydrides, which are active R-olefin hydrogenation catalysts. Without cocatalyst,
the neutral dimers 8 and 9 are poor, though active ethylene polymerization catalysts
(activity: (5-10) × 103 g PE/(mol‚h)). Addition of B(C6F5)3 affords the cationic, mono(benzyl)
complexes {[(c-C5H9)7Si7O12]2M2(CH2Ph)}(+) (M ) Zr, Hf): single-site catalysts (activity: (2-
8) × 106 g PE/(mol‚h)) that are considerably more active than the neutral 8 and 9. Whereas
titanasilsesquioxanes 3 and 7 do not react with THF, the corresponding zirconasilsesqui-
oxanes 4 and 8 form bis(THF) adducts, [(c-C5H9)7Si7O11(OSiMe3)]2Zr‚2THF (5) and [(c-C5H9)7-
Si7O12]ZrCH2Ph‚2THF (10), which suggests that the titanium complexes are less electrophilic
than the zirconium ones. Accordingly, the titanium complex 7 does not react with dihydrogen
and is inactive in ethylene polymerization.
In tr od u ction
processes as a result of which the necessary amount of
MAO can be reduced.4 Metallocenes physisorbed onto
MAO-pretreated supports yield systems that resemble
their unsupported analogues.5a,b However, a drawback
of these systems is that leaching remains a problem and
the activities are rather low. On the other hand, silica-
or alumina-grafted early transition metal complexes are
much more robust with respect to leaching and do not
need to be activated by a cocatalyst.2b,g,5c Although
recent studies show that oxide-grafted organometallic
complexes can be quite uniform in structure, reactivity,
Although metallocenes are among the most active
olefin polymerization catalysts known to date,1 their
commercial use remains limited. The industrial ap-
plicability of homogeneous metallocene based olefin
polymerization catalysts is not so much limited by the
activity or selectivity of the catalyst but more due to
restrictions such as poor polymer morphology and the
necessity of large quantities of the cocatalyst methyla-
lumoxane (MAO), which makes processing of the poly-
mer difficult and expensive. These problems can partly
be overcome by using immobilized catalysts.2 For ex-
ample, morphology control by replication can lead to
uniform polymer particles with a high bulk density,3
whereas site isolation reduces bimolecular deactivation
(2) For example see: (a) Zakharov, V. A.; Yermakov, Y. I. Catal.
Rev.-Sci. Eng. 1979, 19, 67. (b) Collette, J . W.; Tullock, C. W. U.S.
Patent 4.298.722 (Du Pont), 1981. (c) Keii, T., Soga, K., Eds. Catalytic
Polymerization of Olefins, Studies in Surface Science and Catalysis;
Elsevier: New York, 1986. (d) Kaminsky, W., Sinn, H., Eds. Transition
Metals and Organometallics as Catalysts for Olefin Polymerization;
Springer-Verlag: New York, 1988. (e) Marks, T. J . Acc. Chem. Res.
1992, 25, 57. (f) Marsden, C. E. Plast., Rubber Compos. Process. Appl.
1994, 21, 193. (g) Proceedings of SPO ′96 1996, 283. (h) Olabisi, O.;
Atiqulla, M.; Kaminsky, W. Makromol. Chem. Phys. 1997, C37 (3), 519,
and references therein.
(3) For example see: Hungenberg, K. H.; Kerth, J .; Langhauser, F.;
Marczinke, B.; Schlund, R. In Ziegler Catalysts; Fink, G., Mu¨lhaupt,
R., Brintzinger, H. H., Eds.; Springer-Verlag: New York, 1995; Chapter
20.
(4) Chien, J . C. W.; He, D. J . Polym. Sci.: Part A: Polym. Chem.
1991, 1603.
* Corresponding author. E-mail: R.Duchateau@tue.nl. Fax: +31 40
245 5054.
† Eindhoven University of Technology.
‡ University of Groningen.
§ DSM Research B.V.
(1) (a) Kaminsky, W.; Miri, M.; Sinn, H.; Woldt, R. Makromol. Chem.,
Rapid. Commun. 1983, 4, 417. (b) Spaleck, W.; Ku¨ber, F.; Winter, A.;
Rohrmann, J .; Bachmann, B.; Antberg, M.; Dolle, V.; Paulus, E. F.
Organometallics 1994, 13, 954. (c) Sinclair, K. B.; Wilson, R. B. Chem.
Ind. 1994, 857. (d) Hackmann, M.; Rieger, B. CATTECH 1997, 2, 79.
10.1021/om980687k CCC: $15.00 © 1998 American Chemical Society
Publication on Web 12/01/1998

5664 Organometallics, Vol. 17, No. 26, 1998
Duchateau et al.
Sch em e 1
and distribution, their exact structure often remains a
subject of debate.6 Despite advances in spectroscopic,
physical, and chemical techniques, investigation of the
processes taking place at these catalyst surfaces re-
mains inherently difficult.7 As a consequence, efforts
have been undertaken to develop homogeneous model
systems to mimic silica surface metal complexes. One
sophisticated example in which homogeneous model
systems have proven to be very useful for the under-
standing of reactions taking place at surfaces involves
silsesquioxanes. Elegant work of Feher et al.8 has
demonstrated that partly condensed silsesquioxanes are
suitable structural models for a variety of surface sites
ranging from isolated silanols to silanol nests, as found
in silicas and zeolitic materials. The power of these
model systems arises from the fact that standard
characterization techniques such as solution NMR
spectroscopy provide crucial information that cannot be
obtained from the corresponding heterogeneous sys-
tems. In addition, the high uniform character of the
silsesquioxanes provides an excellent possibility to fine-
tune synthetic procedures to anchor the metal complex
to the support. Recent results have shown that (metal-
la)silsesquioxanes are also very suitable to perform
quantum mechanical calculations which can lead to
reliable predictions for the corresponding silica-sup-
ported systems.9b
cently synthesized in our group,9 and Feher’s disilanol
(c-C5H9)7Si7O9(OSiMe3)(OH)2 (II) and trisilanol (c-C5H9)7-
Si7O9(OH)3 (III), respectively.8 With these ligands avail-
able, we aimed to prepare homogeneous complexes that
mimic the corresponding silica-supported group 4 metal
alkyl and hydride complexes as reported by, for ex-
ample, Schwartz and Basset.6b-f Since protolysis of
metal precursors containing a hydrolizable ligand is the
preferred route to synthesize silica-grafted systems,6 it
was also the selected method for the preparation of
silsesquioxane complexes.
Sila n ol I. In accordance with the grafting of group 4
metal alkyls onto partly dehydroxylated silica6 and the
alcoholysis with ROH (e.g., ROH ) (Me3C)3COH, 2,6-
(Me3C)2-C6H3OH, 2,2′-CH2(C6H2-4-Me-6-(CMe3)OH)2),10
introduction of silsesquioxane I to group 4 metal half-
sandwich complexes is conveniently achieved by pro-
tolysis.9 For example, as described in detail elsewhere,9
Cp′′Ti(CH2Ph)3 (Cp′′ ) 1,3-(Me3Si)2C5H3) reacts selec-
tively with I to produce the corresponding silsesquiox-
ane complex Cp′′[(c-C5H9)7Si8O13]Ti(CH2Ph)2 (1, eq 1).
Similarly, TiCl4 reacts with 1 equiv of I, liberating HCl
and affording the (silsesquioxane)titaniumtrichloride,
[(c-C5H9)7Si8O13]TiCl3 (2), in good yield (eq 2).11
Our interest in silsesquioxane chemistry is aimed at
developing homogeneous group 4 metal silsesquioxane
complexes and to compare the catalytic activity of these
metallasilsesquioxanes with that of corresponding het-
erogeneous catalysts. We report here the synthesis of
group 4 metal silsesquioxane complexes and their use
as olefin hydrogenation and polymerization catalysts.
These results provide insight into some very interesting
mechanistic aspects of silsesquioxane-based Ziegler-
Natta catalysts.
Resu lts a n d Discu ssion
Syn th esis of Silsesqu ioxa n e Gr ou p 4 Meta l Com -
p lexes. Three different silsesquioxanes, each represent-
ing a different type of surface silanol site, were selected
for our investigation (Scheme 1): the monodentate cube-
octameric spherosilicate (c-C5H9)7Si8O12(OH) (I), re-
Surprisingly, the protolysis of I by homoleptic group
4 metal benzyl complexes, M(CH2Ph)4 (M ) Ti, Zr, Hf),
is completely aselective. Instead of the expected product,
[(c-C5H9)7Si8O13]M(CH2Ph)3, only complicated product
mixtures were observed containing significant amounts
of the precursor M(CH2Ph)4. Even the reaction of
(5) (a) Soga, K.; Kaminsky, W. Makromol. Chem., Rapid Commun.
1992, 13, 221. (b) Soga, K.; Kaminsky, W. Makromol. Chem. 1993, 194,
1745. (c) Kaminsky, W.; Renner, F. Makromol. Chem., Rapid Commun.
1993, 14, 239.
(6) (a) Ballard, D. G. H. Adv. Catal. 1973, 23, 263. (b) Schwartz, J .;
Ward, M. D. J . Mol. Catal. 1980, 8, 465. (c) King, S. A.; Schwartz, J .
Inorg. Chem. 1991, 30, 3771. (d) Scott, S. L.; Basset, J .-M.; Niccolai,
G. P.; Santini, C. C.; Candy, J .-P.; Lecuyer, C.; Quignard, F.; Choplin,
A. New. J . Chem. 1994, 18, 115. (e) Vidal, V.; The´olier, A.; Thivolle-
Cazat, J .; Basset, J .-M.; Corker, J . J . Am. Chem. Soc. 1996, 118, 4595.
(f) Quignard, F.; Le´cuyer, C.; Choplin, A.; Basset, J .-M. J . Chem. Soc.,
Dalton Trans. 1994, 1153. (g) Ajjou, J . A.; Scott, S. Organometallics
1997, 16, 86. (h) Bade, O. M.; Blom, R.; Ystenes, M. Organometallics
1998, 17, 2524. (i) Theopold, K. H. Eur. J . Inorg. Chem. 1998, 15, and
references therein.
(7) (a) Tamura, K. Dynamic Heterogeneous Catalysis; Academic
Press: New York, 1978. (b) Delgass, W. N.; Haller, G. L.; Kellerman,
R.; Lunford, J . H. Spectroscopy in Heterogeneous Catalysis; Academic
Press: New York, 1979. (c) Delannay, F. Characterization of Hetero-
geneous Catalysts; Marcel Dekker: New York, 1984. (d) Fu, S. L.;
Rosyneck, M. P.; Lunsford, J . Langmuir 1991, 7, 1179. (e) Schnellback,
M.; Ko¨hler, F. H.; Blu¨mel, J . J . Organomet. Chem. 1996, 520, 227.
(8) (a) Feher, F. J .; Newman, D. A.; Walzer, J . F. J . Am. Chem. Soc.
1989, 111, 1741. (b) Feher, F. J .; Newman, D. A. J . Am. Chem. Soc.
1990, 112 (2), 1931. (c) Feher, F. J .; Budzichowski, T. A.; Blanski, R.
L.; Weller, K. J .; Ziller, J . W. Organometallics 1991, 10, 2526.
(9) (a) Duchateau, R.; Abbenhuis, H. C. L.; van Santen, R. A.; Thiele,
S. K.-H.; van Tol. M. F. H. Organometallics, in press. (b) Duchateau,
R.; Cremer, U.; Harmsen, R.; Mohamud, S.; Abbenhuis, H. C. L.; van
Santen, R. A.; Meetsma, A.; Thiele, S. K.-H.; van Tol. M. F. H.
Manuscript in preparation.
(10) (a) Lubben, T. V.; Wolczanski, P. T.; Van Duyne, G. D.
Organometallics 1984, 3, 977. (b) Latesky, S. L.; McMullen, A. K.;
Niccolai, G. P.; Rothwell, I. P. Organometallics 1985, 4, 902. (c) Okuda,
J .; Fokken, S.; Kang, H.-C.; Massa, W. Chem. Ber. 1995, 128, 221.
(11) Similar dehydrochlorination reactions have been reported for
both homogeneous and heterogeneous systems. For example: (a) Park,
J . R.; Shiono, T.; Soga, K. Macromolecules 1992, 25, 521. (b) Nomura,
K.; Naga, N.; Miki, M.; Ynagi, K.; Imai, A. Organometallics 1998, 17,
2152.

Ethylene Polymerization
Organometallics, Vol. 17, No. 26, 1998 5665
of M(CH2Ph)4 and [(c-C5H9)7Si7O11(OSiMe3)]2M (M ) Ti
(3), Zr (4), eq 3). As observed for silanol I, protolysis
reactions of II are nonselective in the sense that
incorporation of the second silsesquioxane ligand is
faster than the first one. Although undesired in view of
our goal to prepare olefin polymerization catalysts, this
phenomenon is in agreement with the reported activat-
ing effect of alumina and aluminosilicate carriers on
group 4 metal complexes.2,6 When soluble silsesquiox-
anes are used instead of these solid supports, bimolecu-
lar reactions exclude isolation of reactive intermediates
and inherently result in the formation of thermody-
namically stable complexes such as 3 and 4 or product
mixtures.
M(CH2Ph)4 with 2 or 3 equiv of I proved unsuccessful,
resulting in a mixture of products including M(CH2Ph)4.
The formation of these product mixtures indicates that
the rate of protolysis for the intermediate silsesquioxane
metal benzyl complexes is at least of the same order as
for M(CH2Ph)4. This contrasts with the mono(silsesqui-
oxane) complexes Cp′′[(c-C5H9)7Si8O13]Ti(CH2Ph)2 and
[(c-C5H9)7Si8O13]TiCl3 (2), which refuse to react with
another equivalent of I even under forced conditions
(toluene, reflux).
The reaction of M(CH2Ph)4 with silanol I is consider-
ably faster than with alcohols. Whereas the reaction of
Zr(CH2Ph)4 with I is nearly instantaneous at room
temperature, protolysis of tritoxH ((Me3C)3COH)10a
requires refluxing in benzene to proceed. With a cone
angle of ∼125°,10a tritoxH is sterically less hindered
than silanol I (cone angle ) 148°).9b Hence, the observed
difference in reactivity between tritox and I is probably
electronic rather than steric in origin. This assumption
is supported by results of Feher et al.,12 who reported
the [Si8O12] cage to be as electron-withdrawing as a CF3
group. Hence, it is justifiable to assume that silanol I
is a stronger Brønsted acid than tritoxH, giving faster,
and less selective, protolysis reactions.
Disila n ol II. Following the reactions with silanol I,
protolysis reactions were carried out with the disilanol
(c-C5H9)7Si7O9(OSiMe3)(OH)2 (II), a model for bipodal
silica silanol sites. Since they are suitable model sys-
tems for Schwartz’s silica-supported [tSiO]2ZrR2 (R )
H, alkyl)6b,c and potential precursors for catalytic olefin
polymerization, attempts were undertaken to synthesize
mono(silsesquioxane) metal bis(alkyl) complexes of the
type [(c-C5H9)7Si7O11(OSiMe3)]MR2. The reaction be-
tween equimolar amounts of II and M(CH2Ph)4 (M )
Ti, Zr) in toluene resulted selectively in a 1:1 mixture
Both metallasilsesquioxane complexes 3 and 4 could
be synthesized in excellent yield by treatment of toluene
solutions of M(CH2Ph)4 with 2 equiv of silsesquioxane
II (eq 3). During the course of this investigation, Crocker
et al. reported the synthesis of the cyclohexyl-substi-
tuted analogue of 3, [(c-C6H11)7Si7O11(OSiMe3)]2Ti, which
was used as an olefin epoxidation catalyst.13 Noteworthy
to mention is that in the presence of THF 4 readily
forms the bis(THF) adduct [(c-C5H9)7Si7O11(OSiMe3)]2-
Zr‚2THF (5, eq 4), whereas 3 refuses to react with Lewis
bases (THF, MeCN). This low tendency of 3 to form
Lewis base adducts cannot be explained by the differ-
ence in steric hindrance (ionic radii: Ti4+ ) 0.74 Å, Zr4+
) 0.86 Å)14 for 3 and 4 alone. Hence it appears that 3 is
significantly less Lewis acidic than 4 (vide infra).
Alternative routes to prepare mono(silsesquioxane)
metal bis(alkyl) complexes of the type [(c-C5H9)7Si7O11-
(OSiMe3)]MR2 were also investigated. By treatment of
(12) (a) Feher, F. J .; Budzichowski, T. A. J . Organomet. Chem. 1989,
379, 33. (b) Feher, F. J .; Tajima, T. L. J . Am. Chem. Soc. 1994, 116,
2145.

5666 Organometallics, Vol. 17, No. 26, 1998
Duchateau et al.
(PhCH2)2ZrCl2‚OEt2 with disilanol (c-C5H9)7Si7O9-
(OSiMe3)(OH)2 (II) incorporation of a second equivalent
of II was successfully avoided, yielding the correspond-
ing mono(silsesquioxane) zirconium dichloride, [(c-C5H9)7-
Si7O11(OSiMe3)]ZrCl2‚2THF (6, eq 5) in excellent yield.
Alkylation of 6 is currently under investigation, al-
though separation problems have until now prevented
the isolation of well-defined products. Alternatively, the
reaction of (PhCH2)2ZrCl2‚OEt2 with 1 equiv of the
thallium salt of II, (c-C5H9)7Si7O9(OSiMe3)(OTl)2,15 was
also attempted but did not afford a uniform product
either.
F igu r e 1. Ortep view of [(c-C5H9)7Si7(OSiMe3)O11]2Zr‚
2THF (5). Ellipsoids are scaled to enclose 30% of the
electron density. Hydrogens and cyclopentyl groups are
omitted for clarity. Selected bond distances (Å): Zr-O(1),
2.289(5); Zr-O(2), 2.287(5); Zr-O(3), 2.037(4); Zr-O(13),
1.969(4); Zr-O(15), 2.038(5); Zr-O(25), 1.957(4); Si(9)-
O(15), 1.590(5); Si(8)-O(14), 1.649(6). Bond angles (deg):
O(1)-Zr-O(15), 84.33(17); O(1)-Zr-O(13), 169.60(17);
O(2)-Zr-O(13), 89.88(17); O(2)-Zr-O(25), 167.85(18);
O(3)-Zr-O(25), 95.37(18); O(3)-Zr-O(15), 161.57(19);
Si(7)-O(14)-Si(8), 138.9(4); Si(1)-O(4)-Si(2), 156.3(3).
1
The single SiMe3 resonance in the H, 13C, and 29Si
substituents hampered the refinement. Nevertheless, all
relevant atoms could reliably be localized, providing
sufficient information concerning the structure.
NMR spectra of 3 and 4 is consistent with a tetrahedral
metal center containing C2 symmetry. In analogy with
its cyclohexyl-substituted analogue, [(c-C6H11)7Si7O11-
(OSiMe3)]2Ti,13 the 13C and 29Si NMR spectra of 3 show
seven equi-intense resonances (assigned to methine-CH
and silsesquioxane framework-Si, respectively), which
indicates that the local mirror symmetry of the silses-
quioxane ligands is lost.16 The larger ionic radius of
zirconium, compared to titanium, finds expression in a
more fluxional complex 4, as is demonstrated by the
coalescence of two pairs of methine-CH and framework-
Si resonances in the room-temperature 13C and 29Si
NMR spectra of 4, respectively. With five methine-CH
resonances (1:2:2:1:1 ratio) and a single Si(CH3)3 signal,
the 13C NMR spectra of 5 and 6 are very similar and in
agreement with a C2 symmetric structure in which the
mirror symmetry of the silsesquioxane ligands is re-
tained. The room-temperature 29Si NMR spectrum of 5
displays six framework-Si resonances (1:1:1:2:1:1 ratio),
The six-coordinated zirconium atom in 5 displays a
distorted octahedral geometry with pseudo-C2v sym-
metry formed by the coordination of two silsesquioxane
ligands and two THF molecules. The THF molecules are
positioned cis to each other. With an average distance
of 2.288(5) Å the Zrr:O dative bonds in 5 are between
the Zrr:O bond lengths observed for cationic zir-
conocene THF adducts and neutral zirconium complexes
(ranging from 2.122(14) Å to 2.662(3) Å).19 An indication
that the metal center in this zirconasilsesquioxane
complex is quite electrophilic is given by the fact that
the THF oxygen atoms in 5 are virtually planar (sum
of angles: O(1), 358.0(5)°; O(2), 358.6(5)°), which reflects
a rehybridization of the two oxygen donor orbitals to
produce an sp2 type σ-donor orbital and a π-donor orbital
of predominantly p character.19a,b To optimize the orbital
overlap associated with this additional π-interaction,
both THF molecules in 5 are orientated nearly perpen-
dicular (O(15)-Zr-O(1)-C(1) ) -8.1(5)°; O(3)-Zr-
O(2)-C(5) ) -9.6(5)°) to the plane formed by O(1), O(2),
O(13), and O(25). The Zr-O(3) (2.037(4) Å) and Zr-
O(15) (2.038(5) Å) bond lengths are considerably longer
than the Zr-O(13) (1.969(4) Å) and Zr-O(25) (1.957(4)
Å) bond distances opposite the coordinated THF mol-
ecules and the Zr-O bonds in [(c-C6H11)7Si7O12]ZrCp*
(1.958(32) Å).18a The Si-O distances (ranging from
of which two coalesce at 37 °C (∆Gq ) 63.5 kJ /mol)17
Tc
to give the expected set of five signals in a 1:1:1:2:2 ratio.
X-r ay Str u ctu r e Deter m in ation of [(c-C5H9)7Si7O11
-
(OSiMe3)]2Zr ‚2THF (5). Crystallization from hexane
at -30 °C afforded colorless block-shaped crystals of 5
suitable for an X-ray diffraction study. A perspective
view of the molecular structure is shown in Figure 1.
Crystal data are collected in Table 2. As for many
crystallographically characterized silsesquioxane com-
pounds,8,15,18 conformational disorder of the cyclopentyl
(13) Crocker, M.; Herold, R. H. M.; Orpen, A. G. Chem. Commun.
1997, 2411.
(14) Shannon, R. D. Acta Crystallogr. 1976, A32, 751.
(15) Feher, F. J .; Rahimian, K.; Budzichowski, T. A.; Ziller, J .
Organometallics 1995, 14, 3920.
(18) For example see: (a) Feher, F. J . J . Am. Chem. Soc. 1986, 108,
3850. (b) Feher, F. J .; Budzichowski, T. A.; Ziller, J . W. Inorg. Chem.
1992, 31, 5100. (c) Buys, I. E.; Hambley, T. W.; Houlton, D. J .;
Maschmeyer, T.; Masters, A. F.; Smith, A. K. J . Mol. Catal. 1994, 86,
309.
(16) This is in agreement with the solid-state structure of 3, which
will be described elsewhere. Vorstenbosch, M. L. W.; Abbenhuis, H.
C. L.; van Santen, R. A.; Spek, A. L. To be published.
(19) (a) J ordan, R. F.; Bajgur, C. S.; Willett, R.; Scott, B. J . Am.
Chem. Soc. 1986, 108, 7410. (b) Amorose, D. M.; Lee, R. A.; Petersen,
J . L. Organometallics 1991, 10, 2191. (c) Giannini, L.; Caselli, A.;
Solari, E.; Floriani, C.; Chiesi-Villa, A.; Rizzolli, C.; Re, N.; Sgamellotti,
A. J . Am. Chem. Soc. 1997, 119, 9198.
(17) T ) 310 K, ∆ν ) 56.2 Hz, ∆Gq ) -RTc ln(π∆νh/x2kbTc):
Tc
Kessler, H. Angew. Chem. 1970, 82, 237.

Ethylene Polymerization
Organometallics, Vol. 17, No. 26, 1998 5667
Ta ble 1. Su m m a r y of Eth ylen e P olym er iza tion Exp er im en ts
catalyst
(µmol)
cocatalyst
(mmol)
conditions
[min]/[°C]/[atm]
PE yield
(g)
activity
(g/(mmol‚h))
Mw
(×103)
Mw/Mn
7 (30)
7 (30)
8 (30)
8 (20)
8 (20)
8 (10)
8 (40)
9 (30)
9 (20)
30/25f50/5
30/25f50/5
7/25/3
0
0
a
BF15 (1)
0.6
5.5
trace
9.7
12.0
0.5
11.1
10
2400
∼0
BF15 (1)
7/80/5
7/80/5
7/29/5
6.6
2.3
b
BF20 (5.6)
Silicac/MAO
MCMe/MAO
8300d
2600d
<10
65
32
11.7
6.8
7/22-82/5f
15/25/3
7/80/5
BF15 (1)
4800
82
3.2
a
b
d
BF15 ) B(C6F5)3. BF20 ) [Ph3C][B(C6F5)4]. c Silica ) PQ3040. Assuming that all the catalyst is active. e MCM ) all-silica MCM-
41. f Strong temperature increase observed.
Ta ble 2. Deta ils of th e X-r a y Str u ctu r e Deter m in a tion of [(c-C5H9)7Si7O11(OSiMe3)]2Zr ‚2THF (5) a n d
{[(c-C5H9)7Si7O12]Zr CH2P h }2 (8)
5
8
formula
fw
C
84H160O26Si16Zr‚(C4H8O)0.72
(C42H70O12Si7Zr)2
2109.67
2178.69
cryst system
space group, no.
a, Å
b, Å
c, Å
monoclinic
P21/c, 14
20.197(3)
15.937(1)
35.006(3)
triclinic
P1h
13.971(1)
16.172(1)
27.004(2)
R, deg
94.439(6)
99.491(7)
99.810(7)
5894.7(7)
1.189
2
â, deg
97.02(1)
γ, deg
V, Å3
11 183(2)
1.294
4
D
Z
calc, g cm-3
F(000), electrons
µ(Mo, KR), cm-1
cryst size, mm
4660
3.4
2224
3.78
0.30 × 0.45 × 0.50
0.35 × 0.38 × 0.50
T, K
130
130
θ range, deg: min, max
1.01, 24.0
0.710 73
graphite
∆ω ) 0.85 + 0.34 tan θ
h: 0f23; k: 0f18; l: -39f39
1.24, 25.0
0.710 73
graphite
∆ω ) 0.90 + 0.34 tan θ
h: -16f16; k: 0f19; l: -32f31
λ (Mo, KR), Å
monochromator
ω/2θ scan, deg
index ranges
total no. of data
no. of unique data
no. of data with criterion: (Fo g 4.0σ(Fo))
18 841
17 499
11 440
21 882
20 604
16 611
0.038
0.066
20 604
1117
2
2
Rint ) ∑[|(Fo - Fo2(mean)|)/∑[Fo
]
2
Rsig ) ∑σ(Fo2)/∑[Fo
]
0.058
17 481
1150
2
no. of reflections (Fo g 0)
no. of refined params
final agreement factors: wR(F2) ) [∑[w(Fo - Fc2)2]/∑[w(Fo2)2]]1/2 0.2090
0.2568
2
2
for Fo > 0
weighting scheme: a, b
0.0633, 87.60
0.1269, 28.00
w ) 1/[σ2(Fo2) + (aP)2 + bP] and P ) [max(Fo2,0) + 2Fc2]/3
R(F) ) ∑(||Fo| - |Fc||)/∑|Fo| for Fo > 4.0σ(Fo)
0.0818
1.036
0.0866
1.125
2
GoF ) S ) [∑[w(Fo - Fc2)2]/(n/p)]1/2
n ) no. of reflns
p ) no. of params refined residual electron density in
final difference Fourier map, e/Å3
max. (shift/σ) final cycle
-1.0, 1.2(1)
<0.001
-0.88, 2.17(15)
<0.001
1.590(5) Å to 1.649(6) Å) and Si-O-Si angles (ranging
from 138.9(4)° to 156.3(3)°) in 5 are similar to those
observed in other crystallographically characterized
(metalla)silsesquioxane compounds.8,18a
deficiency, the (formally 8-electron) zirconium and
hafnium benzyl complexes (8, 9) form dimeric structures
1
which display much more complicated H, 13C, and 29Si
spectra (Figure 2, vide infra).
Tr isila n ol III. Compared to the reactions with I and
II, protolysis of the trisilanol III with equimolar amounts
of the homoleptic group 4 metal benzyl complexes,
M(CH2Ph)4 (M ) Ti, Zr, Hf), proceeds surprisingly
selectively, affording metallasilsesquioxanes with the
general formula [(c-C5H9)7Si7O12]MCH2Ph (M ) Ti (7),13
Zr (8), Hf (9), eq 6). In analogy with its cyclohexyl-
substituted analogue,13 both the 13C and 29Si NMR
spectra of 7 show three resonances in a 3:1:3 ratio,
consistent with a monomeric C3v corner-capped metal-
lasilsesquioxane complex.13,18a To reduce their electron
To investigate the fluxional behavior of these dimers
in more detail, a variable-temperature 29Si NMR experi-
ment on toluene solutions of 8 was carried out. In the
low-temperature limit (-60 °C), seven equi-intense
resonances indicate that all silicon atoms of each
metallasilsesquioxane moiety are different. This is in
accordance with the solid-state structure of 8 (Figure
2). Upon raising the temperature (from -60 °C to -15
°C), the fluxionality of the metallasilsesquioxane moi-
eties around the siloxy bridges increases, giving five
resonances (2:1:1:1:2 ratio) that are in agreement with

5668 Organometallics, Vol. 17, No. 26, 1998
Duchateau et al.
) 131 Hz)20 of ηn-bonding of the benzyl group in 7 could
be detected either. Furthermore, like [(c-C5H9)7Si7O11-
(OSiMe3)]2Ti (3), the titanasilsesquioxane benzyl com-
plex [(c-C5H9)7Si7O12]TiCH2Ph (7) does react with Lewis
bases and is recovered unchanged after treatment with
THF or acetonitrile. In contrast, 4 and 8, the zirconium
analogues of 3 and 7, readily form the bis(THF) adducts
[(c-C5H9)7Si7O11(OSiMe3)]2Zr‚2THF (5) and [(c-C5H9)7-
Si7O12]ZrCH2Ph‚2THF (10) when treated with or syn-
thesized in THF. The difference in ionic radii14 of
titanium and zirconium alone cannot explain that
titanium complexes (3, 7) refuse to react with THF
whereas the corresponding zirconium compounds (4, 8)
react with even 2 equiv of THF. Therefore, the low
affinity of 3 and 7 toward Lewis bases is probably
electronic in origin. However, why titanasilsesquioxanes
are electronically less unsaturated than the correspond-
ing zirconium complexes is not clear. Titanasilsesqui-
oxanes and zirconasilsesquioxane complexes also behave
differently with respect to hydrolysis. The tSiO-Ti
bonds in 3 and 7 are stable toward hydrolysis, whereas
the corresponding zirconium compounds (4, 8) hydrolyze
rapidly to form the corresponding silanols II and III,
suggesting that the titanium complexes are less ionic
than their zirconium analogues.21
F igu r e 2. ORTEP drawing of {[(c-C5H9)7Si7O12]ZrCH2Ph}2
(8). Ellipsoids are scaled to enclose 30% of the electron
density. Hydrogens and cyclopentyl groups are omitted for
clarity. Selected bond distances (Å): Zr(1)-O(1), 2.202(4);
Zr(1)-O(5), 2.458(5); Zr(1)-O(11), 1.981(4); Zr(1)-O(12),
1.952(5); Zr(1)-O(24), 2.164(4); Zr(2)-O(1), 2.166(5); Zr(2)-
O(13), 1.943(5); Zr(2)-O(23), 1.954(5); Zr(2)-O(24), 2.263-
(4); Zr(2)-C(8), 2.281(10); Zr(1)-C(1), 2.374(7); Zr(1)-C(2),
2.584(7); Si(1)-O(5), 1.682(5); Si(4)-O(5), 1.674(5); Si(1)-
O(2), 1.608(5); Si(5)-O(6), 1.633(5). Bond angles (deg):
C(8)-Zr(2)-O(24), 153.6(3); O(1)-Zr(2)-O(23), 159.32(17);
O(13)-Zr(2)-C(8), 98.1(3); O(13)-Zr(2)-O(24), 102.65(19);
O(13)-Zr(2)-O(23), 99.5(2); O(13)-Zr(2)-O(1), 98.69(18);
O(5)-Zr(1)-C(1), 147.6(2); O(5)-Zr(1)-C(2), 175.11(19);
O(12)-Zr(1)-O(24), 164.05(17); O(1)-Zr(1)-O(11), 145.02-
(18); Zr(2)-C(8)-C(9), 124.5(8); Zr(1)-C(1)-C(2), 80.9(4);
Si(1)-O(5)-Si(4), 134.0; Si(12)-O(18)-Si(13), 167.3(3).
X-ray Structure Determination of{[(c-C5H9)7Si7O12]-
Zr CH2P h }2 (8). Slow crystallization from a cooled (-30
°C) saturated hexane solution resulted in pale yellow
single crystals of 8 suitable for an X-ray diffraction
study. A perspective view of the molecular structure of
8 is shown in Figure 2. Crystal data are collected in
Table 2. As observed for 5, the structure of 8 showed
considerable conformational disorder of the silsesqui-
oxane cyclopentyl substituents. Yet, this did not affect
the localization of the most relevant atoms. From Figure
2 it is clear that 8 consists of an asymmetric dimeric
structure. The dimer is formed by two corner-capped
zirconasilsesquioxane moieties connected by two bridg-
ing siloxy functionalities. Typical for electron-deficient
dimeric metallasilsesquioxane complexes, O(1) and O(24)
in 8 are µ2-bridged to increase the metal’s coordination
number. Examples of similar metallasilsesquioxane
complexes containing µ2-bridged siloxy groups are
{[R7Si7O12]M}2 (M ) Ti, V, Al, Y; R ) c-C5H9, c-C6H11).22
Even for less electron-deficient metals dimeric struc-
tures are sometimes preferred. In these systems, {[(c-
C6H11)7Si7O12]M}2 (M ) B, TiCp, VdO, Mo),18b,c,23 the
dimeric structure is maintained by siloxy functionalities
that are σ-bonded to the adjacent metal center. Al-
though entropically unfavorable, these dimers are formed
to optimize O(pπ)-M(dπ) bonding, which is less effective
in monomeric structures for which the Si-O-M angles
local mirror symmetry of the silsesquioxane ligands.
Further increase of the temperature (from -15 °C to
60 °C) eventually yields three resonances with relative
intensities of 3:1:3, indicating that at elevated temper-
ature the dimeric structure becomes very fluxional or
may even dissociate.
(20) ηn-bonding of a benzyl group is accompanied with a highfield
1
shift of the ipso-carbon of the benzyl group, and a large J C-H (>140
Hz) for the benzylic CH2 group. For example see: (a) J ordan, R. F.;
LaPointe, R. E.; Bajgur, C. S.; Echols, S. F.; Willet, R. J . Am. Chem.
Soc. 1987, 109, 4111. (b) J ordan, R. F.; LaPointe, R. E.; Baenziger, N.
C.; Hinch, G. D. Organometallics 1990, 9, 1539. (c) Bochmann, M.;
Lancaster, S. J . Organometallics 1993, 12, 633. (d) Bochmann, M.;
Lancaster, S. J .; Hursthouse, M. B.; Malik, K. M. A. Organometallics
1994, 13, 2235.
(21) The stability toward hydrolysis of the titanium-siloxy bonds
in 3 and 7 is in agreement with that of TS-1 and Ti-Beta, which are
effective oxidation catalysts in an aqueous environment. For example
see: Corma, A.; Esteve, P.; Mat´ınez, A.; Valencia, S. J . Catal. 1995,
152, 18, and references therein.
The observation that, unlike 8 and 9, the titanasil-
sesquioxane 7 (formally an 8-electron species) is a
monomer in solution is rather surprising. No strong
evidence (13C NMR, ipso-C6H5 ) 128 ppm; CH2Ph, 1J C-H

Ethylene Polymerization
Organometallics, Vol. 17, No. 26, 1998 5669
do not strongly support such a structure, it is clear that
Zr(1) shows all the characteristics of a strong electron-
deficient metal center, similar to that in cationic zirco-
nium systems. In contrast, Zr(2) does not show any signs
of being very electron-deficient. Due to the coordination
of O(5) to Zr(1), the Si(1)-O(5) (1.682(5) Å) and Si(4)-
O(5) (1.674(5) Å) bonds are significantly longer than the
other Si-O bonds (1.608(5)-1.633(5) Å) that form the
silsesquioxane framework. Although the silsesquioxane
Si-O-Si span a wide range of angles (from 134.0(3)°
to 167.3(3)°), they are not exceptional.8,18,22,23
F igu r e 3. Proposed resonance structures for {[(c-C5H9)7-
Si7O12]ZrCH2Ph}2 (8).
are more acute.22c Accordingly, O(1) is more strongly
bonded to Zr(2) (2.166(5) Å) than to Zr(1) (2.202(4) Å),
whereas O(24) is more strongly bonded to Zr(1) (2.164(4)
Å) than to Zr(2) (2.263(4) Å). As expected, the µ2-bridged
oxygens in 8 form weaker bonds with the zirconium
than the nonbridging oxygens (1.943(5)-1.981(4) Å), for
which the distances are comparable to the Zr-O σ-bonds
in, for example, 5, [(c-C6H11O7Si7O12]ZrCp*,18a and [2,6-
(Me3C)2C6H3]OZr(CH2Ph)3.10b The coordination environ-
ment of Zr(2) can best be described as distorted square-
pyramidal with O(13) occupying the axial position
(C(8)-Zr(2)-O(24) ) 153.6(3)°, O(1)-Zr(2)-O(23) )
159.32(17)°, O(13)-Zr(2)-C(8) ) 98.1(3)°, O(13)-Zr(2)-
O(24) ) 102.65(19)°, O(13)-Zr(2)-O(23) ) 99.5(2)°,
O(13)-Zr(2)-O(1) ) 98.69(18)°). The coordination en-
vironment of Zr(1) is more complex due to the additional
coordination of O(5) and C(2) to the metal center.
Considering the benzyl C(1) and C(2) atoms to occupy
one coordination vertex, the zirconium environment can
best be described as distorted octahedral (O(5)-Zr(1)-
C(1) ) 147.6(2)°, O(5)-Zr(1)-C(2) ) 175.11(19)°, O(12)-
Zr(1)-O(24) ) 164.05(17)°, O(1)-Zr(1)-O(11) ) 145.02-
(18)°). Whereas for Zr(2) the benzyl group is bonded in
a normal η1 fashion (Zr(2)-C(8) ) 2.281(10) Å, Zr(2)-
C(8)-C(9) ) 124.5(8)°), the acute Zr(1)-C(1)-C(2) angle
(80.9(4)°) and the short Zr(1)-C(2) distance (2.584(7)
Å) are characteristic for an η2-bonded benzyl group.20,24
Besides the η2-bonding mode of the benzyl group, Zr(1)
gains additional electron density by coordination of O(5).
With a distance of 2.458(5) Å, the Zr(1)r:O(5) dative
bond is normal within the range found for other
zirconium complexes.19 A similar intramolecular inter-
action of a silyl ether was observed in a dimeric yttrium
silsesquioxane complex.22d Coordination of silyl ether
functionalities to Lewis acidic metal centers has also
been proposed to occur in corresponding oxide-supported
systems.6c,25 The additional coordination of O(5) and the
η2-bonding of the benzyl group suggest that Zr(1) is more
electron-deficient than Zr(2). A possible explanation for
this phenomenon is given by the zwitterionic resonance
structure as proposed in Figure 3. Although the com-
parable Zr(1)-O and Zr(2)-O distances in the system
M-C σ-Bon d Hyd r ogen olysis. Both the monomeric
(7) and dimeric (8, 9) silsesquioxane benzyl complexes
are interesting models for Basset’s silica-grafted group
4 metal carbyl and hydride systems.6d,e,26 Grafting of
MR4 onto partly dehydroxylated silica affords species
of the type [tSiO]MR3. Subsequent hydrogenolysis of
these immobilized tris(carbyl) complexes has been found
to yield mono(hydride) systems [tSiO]3MH (M ) Ti, Zr,
Hf), whose general structure is similar to that of
7-9.6d,e,26 As a first test to determine the suitability of
metallasilsesquioxane complexes as models for silica-
grafted species, M-C σ-bond hydrogenolysis of 7-9 to
form the corresponding metallasilsesquioxane hydrides,
“[(c-C5H9)7Si7O12]MH”, was attempted. Exposure of the
zirconium and hafnium silsesquioxane benzyl complexes
8 and 9 to dihydrogen (toluene, 50 °C, 3 atm H2) resulted
in the slow (10-20 h) hydrogenolysis of the M-C bond
with concomitant quantitative formation of toluene.
Poor crystallinity, complex NMR spectra, and insuf-
ficient thermal stability frustrated the isolation and
characterization of the zirconium and hafnium hydrides.
Nevertheless, the existence of these hydrides could be
proven by the catalytic hydrogenation of 1-hexene to
n-hexane (eq 7). The monomeric titanasilsesquioxane
benzyl complex 7 is much less susceptible to Ti-C bond
hydrogenolysis: after 48 h at 75 °C under 4 atm of H2
and in the presence of 1-hexene, both the benzyl complex
and the 1-hexene were recovered unchanged. Likewise,
the coordinatively saturated bis(THF) adduct 10 did not
react with dihydrogen under the conditions that proved
to be sufficient to hydrogenolyze 8 and 9.
Olefin P olym er iza tion . Oxide-supported zirconium
hydride and alkyl species are known to effectively insert
CdC double bonds. Recently, Basset et al. reported the
synthesis of a aluminosilicate-supported zirconium
hydride.26b The authors argued that introduction of
electrophilic aluminum centers on the support enhances
the olefin polymerization activity of the catalyst com-
pared to the all-silica system. That Lewis acidic alumi-
num sites on oxide supports can positively influence the
catalyst’s activity was also demonstrated by, for ex-
ample, Colette2b,g and Moroz,27 who found that alumina-
(22) (a) Feher, F. J .; Gonzales, S. L.; Ziller, J . W. Inorg. Chem. 1988,
27, 3440. (b) Feher, F. J .; Walzer, J . F. Inorg. Chem. 1990, 29, 1604.
(c) Feher, F. J .; Budzichowski, T. A.; Weller, K. J . J . Am. Chem. Soc.
1989, 111, 7288. (d) Herrmann, W. A.; Anwander, R.; Dufaud, V.;
Scherer, W. Angew. Chem. 1994, 106, 1338.
(23) (a) Feher, F. J .; Walzer, J . F. Inorg. Chem. 1991, 30, 1689. (b)
Budzichowski, T. A.; Chacon, S. T.; Chisholm, M. H.; Feher, F. J .;
Streib, W. J . Am. Chem. Soc. 1991, 113, 689.
(24) (a) Hughes, A. K.; Meetsma, A.; Teuben, J . H. Organometallics
1993, 12, 1936. (b) Davies, G. R.; J arvis, J . A. J . J . Chem. Soc., Chem.
Commun. 1971, 1511. (c) Latsky, S. L.; McMullen, A. K.; Niccolai, G.
P.; Rothwell, I. P.; Huffman, J . C. Organometallics 1985, 4, 902. (d)
Latesky, S. L.; McMullen, A. K.; Niccolai, G. P.; Rothwell, I. P.;
Huffman, J . C. Organometallics 1985, 4, 902. (e) Mintz, E. A.; Moloy,
K. G.; Marks, T. J . J . Am. Chem. Soc. 1982, 104, 4692.
(26) (a) Rosier, C.; Niccolai, G. P.; Basset, J .-M. J . Am Chem. Soc.
1997, 119, 12408. (b) Dufaud, V.; Basset, J .-M. Angew. Chem. 1998,
110, 848. (c) d’Ornelas, L.; Reyes, S.; Quignard, F.; Choplin, A.; Basset,
J .-M. Chem. Lett. 1993, 1931.
(25) Liebau, F. Structural Chemistry of Silicates; Springer: Berlin,
1985.

5670 Organometallics, Vol. 17, No. 26, 1998
Duchateau et al.
or aluminosilicate-supported zirconium alkyl complexes
are active ethylene polymerization catalysts (without
cocatalysts), whereas the unsupported zirconium alkyls
are inactive.
(10 equiv) of cocatalyst present, these systems react
exclusively with 1 equiv of B(C6F5)3, yielding the
expected cationic mono(benzyl) complexes, {[(c-C5H9)7-
Si7O12]2M2CH2Ph}+ (M ) Zr, Hf). In the 19F NMR
spectra, the small ∆δ(Fm - Fp) values for the [(C6F5)3-
BCH2Ph]- anion (8 + B(C6F5)3, ∆δ(Fm - Fp) ) 2.59 ppm;
9 + B(C6F5)3, ∆δ(Fm - Fp) ) 2.46 ppm) indicate that no
significant anion-cation interaction is present in these
systems.28 The cationic complexes have a limited ther-
mal stability and gradually decompose at room temper-
ature. When activated with [Ph3C]+[B(C6F5)4]-, the
polymerization activity of both 8 and 9 dropped dra-
matically compared to the B(C6F5)3-activated systems.
NMR studies showed that whereas B(C6F5)3 reversibly
abstracts only one benzyl group per dimer, [Ph3C]+-
[B(C6F5)4]- is electrophilic enough to irreversibly remove
both benzyl groups, yielding inactive complexes of the
type {[(c-C5H9)7Si7O12]M}+{B(C6F5)4}-.
The silsesquioxane-stabilized benzyl complexes 7-9
are probably the most realistic model systems for silica-
supported group 4 metal carbyl or hydride species
known to date. In the dimeric structure of 8, the electron
density appears to be unevenly divided between the
metal centers (Figures 2, 3). One could argue that Zr-
(2) acts as an internal cocatalyst (comparable to acidic
aluminum sites on oxide supports), withdrawing elec-
tron density from Zr(1).2a,b,26b,27 Such an intramolecular
activation would render the use of a cocatalyst superflu-
ous.
Preliminary ethylene polymerization experiments in
which 7-9 were used without the addition of a cocata-
lyst revealed that the dimeric metallasilsesquioxane
benzyl complexes 8 and 9 are indeed active ethylene
polymerization catalysts (eq 8), whereas the monomeric
titanasilsesquioxane 7 is inactive. Although the activity
of 8 and 9 in ethylene polymerization is low (Table 1),
the polarization in the dimeric structures (8, 9) is
sufficiently strong to activate the system at least to some
extent. This activating synergism between the two
metal centers is expected to intensify when one of the
two benzyl groups is removed by a cocatalyst, forming
a dimeric mono(benzyl) cation. In the presence of
B(C6F5)3 the ethylene polymerization activity of both 8
and 9 does indeed increase considerably (Table 1, eq 8).
Hence, introducing a positive charge on one of the metal
centers results in activation of the adjacent metal alkyl
bond.
The molecular weight of the polyethylene formed with
B(C6F5)3-activated 8 (Mw ) 6600) is considerably lower
than for the corresponding hafnium system (9: Mw
)
82000). With a Mw/Mn of 2.3 and 3.2, respectively, both
the zirconium and hafnium systems can be considered
as single-site catalysts. Opposite of what is usually
observed, the hafnium complex 9 forms a more active
catalyst than the zirconium complex 8 when activated
with B(C6F5)3 (Table 1). A plausible explanation for this
phenomenon is the fact that the zirconium complex is
thermally less stable than its hafnium analogue (quali-
tatively determined with 1H NMR), which will result
in faster catalyst deactivation of the zirconium complex
at higher polymerization temperatures (80 °C, Table 1).
Immobilization of 8 on MAO-pretreated silica (PQ3040)
and all-silica MCM-41 also gave active ethylene polym-
erization catalysts. The resulting heterogeneous cata-
lysts showed an activity comparable to or even higher
than observed for the unsupported 8 + B(C6F5)3 system
(Table 1). This is quite surprising since heterogenization
of homogeneous metallocene catalysts often goes to-
gether with loss of activity.2 As expected for a hetero-
geneous catalyst, the polyethylene prepared with im-
mobilized 8 showed a much better morphology (spherical
particles) and gave considerably less reactor fouling
than the homogeneously prepared polymer, which was
isolated as a gel. Immobilization of 8 resulted in increase
in molecular weight by a factor of 10; however, the
rather high polydispersity of the polyethylene indicates
that it is probably not formed by a single-site catalyst
(Table 1). This is not surprising keeping in mind that
without cocatalyst 8 is also an active ethylene polym-
erization catalyst. Furthermore, there is a realistic
chance that (partial) substitution of the silsesquioxane
ligands takes place since aluminumalkyls (including
MAO) are known to be able to split metal siloxy,
M-OSit, bonds.29 This could subsequently lead to new
catalytically active complexes which contribute to the
high polydispersity.
Although being reasonably active single-site ethylene
polymerization catalysts (Table 1), the B(C6F5)3-acti-
(28) (a) Pellecchia, C.; Immirzi, A.; Grassi, A.; Zambelli, A. Orga-
nometallics 1993, 12, 4473. (b) Horton, A. D.; de With, J .; van der
Linden, A. J .; van de Weg, H. Organometallics 1996, 15, 2672.
(29) Reactivity studies of metallasilsesquioxane complexes have
shown that aluminum alkyls are able to split M-OSit bonds, resulting
in leaching of the silsesquioxane ligand: (a) Feher, F. J .; Blanski, R.
L. J . Am. Chem. Soc. 1992, 114, 5886. (b) Reference 9.
1H, 11B, and 19F NMR studies on B(C6F5)3 containing
solutions of 8 and 9 showed that even with an excess
(27) Moroz, B. L.; Semikolenova, N. V.; Nosov, A. V.; Zakharov, V.
A.; Nagy, S.; O’Reilly, N. J . J . Mol. Catal. 1998, 130, 121.

Ethylene Polymerization
Organometallics, Vol. 17, No. 26, 1998 5671
[(c-C5H9)7Si8O13]TiCl3 (2). At room temperature, a solution
of (c-C5H9)7Si8O12(OH) (I: 2.92 g, 3.18 mmol) in hexane (30
mL) was added to a hexane (50 mL) solution of TiCl4 (1.5 mL,
13.7 mmol). Upon addition, the reaction mixture turned
reddish and colorless microcrystalline material precipitated.
The volatiles were removed in a vacuum, and the crude
product was dissolved in hot toluene (20 mL). Crystallization
at room temperature yielded 2 as block-shaped off-white
crystals (1.2 g, 1.1 mmol, 35%). Cooling of the mother liquor
to -30 °C yielded a second crop of 2 as white microcrystalline
material (0.7 g, 0.7 mmol, 21%). 1H NMR (CDCl3, δ): 1.80 (m,
14H, CH2-C5H9), 1.55 (m, 42H, CH2-C5H9), 1.07 (m, 7H, CH-
vated metallasilsesquioxanes 8 and 9 show no polym-
erization activity for R-olefins (propene, 1-hexene) at all.
As expected, the monomeric titanasilsesquioxane com-
plex 7 remains inactive upon addition of B(C6F5)3 or
[Ph3C]+[B(C6F5)4]-, which suggests that the dimeric
structure of metallasilsesquioxanes 8 and 9 plays an
important role.
Con clu d in g Rem a r k s
The results presented here emphasize the earlier
proposed suitability of silsesquioxanes to mimic surface
silanol sites. Besides being useful soluble model systems
that can provide significant information about reactions
taking place at silica surfaces, the metallasilsesquioxane
complexes 8 and 9 are very interesting catalysts them-
selves.
1
C5H9). 13C NMR (CDCl3, δ): 27.21 (t, CH2-C5H9, J C-H ) 129
Hz), 27.16 (t, CH2-C5H9, 1J C-H ) 129 Hz), 26.95 (t, CH2-C5H9,
1
1J C-H ) 129 Hz), 22.09 (d, CH-C5H9, J C-H ) 119 Hz), 22.00
1
1
(d, CH-C5H9, J C-H ) 119 Hz), 21.80 (d, CH-C5H9, J C-H
)
121 Hz). 29Si NMR (CH2Cl2, δ): -65.82, -67.04, -113.48 (3:
4:1). Anal. Calcd for C35H63Cl3O13Si8Ti: C, 39.26; H, 5.93; Ti,
4.47. Found: C, 39.72; H, 6.04; Ti, 4.25.
[(c-C5H9)7Si7O11(OSiMe3)]2Ti (3). At -80 °C, Ti(CH2Ph)4
(0.37 g 0.90 mmol) was added to a solution of (c-C5H9)7Si7O9-
(OSiMe3)(OH)2 (II: 1.6 g, 1.7 mmol) in toluene (30 mL). Upon
warming to room temperature, the solution decolorized. After
stirring for an additional hour at room temperature, the
solvent was evaporated and the product was extracted with
hexane (30 mL). Evaporation to dryness and subsequent
recrystallization from CH2Cl2 gave 3 as a white microcrystal-
line material which lost lattice solvent upon drying. Yield: 1.1
One of the shortcomings of homogeneous model
systems is clearly demonstrated by the reactions with
I and II: When soluble silsesquioxanes (I or II) are used
instead of solid supports, bimolecular reactions result
in the formation of thermodynamically stable complexes
such as 3 and 4 or product mixtures.
A peculiar observation is the significantly higher
electrophilic character of zirconium compared to tita-
nium. Although formally 8-electron species, titanasil-
sesquioxanes 3 and 7 refuse to react with strong Lewis
bases such as THF or even the sterically unhindered
MeCN. In contrast, silsesquioxane-stabilized zirconium
and hafnium complexes satisfy their electron deficiency
by forming dimeric structures (8, 9) or bis(THF) adducts
(5, 6, 10). The reason for this difference in electrophi-
licity is not yet understood.
Most intriguing of all is the apparent intramolecular
activating ability of the dimeric benzyl complexes 8 and
9, which, like their heterogeneous congeners,2b,g,26b are
active ethylene polymerization catalysts even without
a cocatalyst. This activating synergism between the two
metal centers in 8 and 9 is intensified by the addition
of B(C6F5)3, which increases the catalyst’s activity by
more than 2 orders of magnitude.
1
g (0.6 mmol, 67%). H NMR (CDCl3, δ): 1.80 (m, 14H, CH2-
C5H9), 1.62 (m, 42H, CH2-C5H9), 1.01 (m, 7H, CH-C5H9), 0.21
(s, 9H, Si(CH3)3). 13C{1H} NMR (CDCl3, δ): 27.79, 27.65, 27.53,
27.39, 27.09, 26.99 (s, CH2-C5H9); 24.36, 23.50, 23.21, 23.12,
23.02, 22.47, 22.35 (s, CH-C5H9); 1.93 (s, Si(CH3)3). 29Si NMR
(toluene, δ): 9.14 (SiMe3), -65.16, -65.70, -66.33, -67.09,
-68.02, -68.22, -68.33. Anal. Calcd for C76H144O24Si16Ti: C,
47.07; H, 7.48; Ti, 2.47. Found: C, 46.99; H, 7.59; Ti, 2.42.
[(c-C5H9)7Si7O11(OSiMe3)]2Zr (4). To a solution of Zr(CH2-
Ph)4 (0.40 g, 0.88 mmol) in toluene (20 mL) was added
(c-C5H9)7Si7O9(OSiMe3)(OH)2 (II: 1.66 g, 1.75 mmol) at -50
°C. Immediately the solution decolorized. The mixture was
allowed to warm to room temperature and was stirred for 1
h. The volatiles were removed in a vacuum, and the crude
product was extracted with and recrystallized from hot hexane
1
(10 mL). Yield: 1.05 g (0.53 mmol, 60%). H NMR (CDCl3, δ):
1.79 (m, 14H, CH2-C5H9), 1.63 (m, 14H, CH2-C5H9), 1.52 (m,
28H, CH2-C5H9), 0.98 (m, 7H, CH-C5H9), 0.23 (s, 9H,
Si(CH3)3). 13C{1H} NMR (CDCl3, δ): 27. 59, 27.49, 27.40, 27.34,
27.10, 27.04, 26.97 (s, CH2-C5H9); 24.26, 23.52, 23.29, 22.51,
22.42 (s, CH-C5H9). 13C NMR (CDCl3, δ): 1.88 (q, Si(CH3)3,
1J C-H ) 118 Hz). 29Si NMR (toluene, 323 K, δ): 9.44 (SiMe3),
-65.14, -65.55, -65.72, -66.63, -67.77 (1:1:1:1:2:2). Anal.
Calcd for C76H144O24Si16Zr: C, 46.04; H, 7.32; Zr, 4.60. Found:
C, 46.16; H, 7.45; Zr, 4.54.
[(c-C5H9)7Si7O11(OSiMe3)]2Zr ‚2THF (5). A THF (20 mL)
solution of Zr(CH2Ph)4 (0.61 g, 1.34 mmol) was treated with
(c-C5H9)7Si7O9(OSiMe3)(OH)2 (II: 2.52 g, 2.66 mmol) at -80
°C. The mixture was allowed to warm to room temperature,
upon which the color of the solution turned from bright to pale
yellow. The solvent was evaporated, and the crude product was
dissolved in CH2Cl2 (10 mL) and filtered to remove minor
amounts of insoluble impurities. Slow evaporation of the CH2-
Cl2 at room temperature yielded 5 as large block-shaped
crystals (1.64 g, 0.77 mmol, 58%). Since the crystals slowly
lost THF incorporated in the lattice, the crystals were pow-
dered and dried thoroughly. 1H NMR (CDCl3, δ): 4.20 (m, 4H,
THF-R-CH2), 1.96 (m, 4H, THF-â-CH2), 1.77 (m, 14H, CH2-
C5H9), 1.61 (m, 14H, CH2-C5H9), 1.51 (m, 28H, CH2-C5H9),
0.97 (m, 7H, CH-C5H9), 0.17 (s, 9H, Si(CH3)3). 13C{1H} NMR
(CDCl3, δ): 71.19 (s, THF-R-CH2), 28.16, 27.85, 27.74, 27.67,
27.60, 27.55, 27.45, 27.13, 26.97, 26.93 (s, CH-C5H9), 25.30
(s, THF-â-CH2), 25.00, 24.10, 23.86, 22.67, 22.49 (s, CH-C5H9,
Exp er im en ta l Section
Gen er a l Com m en ts. All manipulations were performed
under an argon atmosphere using box (Braun MB-150 GI) and
Schlenk techniques. Solvents were distilled from Na (toluene),
K (THF), Na/K alloy (ether, hexanes), or CaH2 (CH2Cl2) and
stored under argon. NMR solvents were dried over Na/K alloy
(benzene-d6, toluene-d8) or 4 Å molecular sieves (CDCl3). NMR
spectra were recorded on a Varian GEMINI 300 and Bruker
AC400 spectrometers. Chemical shifts are reported in ppm and
referenced to residual solvent resonances (1H, 13C NMR) or
external standards (11B, BF3‚OEt2 ) 0 ppm; 19F, CF3CO2H )
0 ppm; 29Si, SiMe4 ) 0 ppm). Elemental analyses were carried
out at the Analytical Department of the University of Gronin-
gen (The Netherlands). To reduce the often observed silicon
carbide formation, V2O5 was added to improve the combustion.
Silsesquioxanes (c-C5H9)7Si8O12(OH) (I),9 (c-C5H9)7Si7O9(OSiMe3)-
(OH)2 (II),8a (c-C5H9)7Si7O9(OH)3 (III),8a and group 4 metal
benzyl complexes M(CH2Ph)4 (M ) Ti, Zr, Hf)30a and (PhCH2)2-
30b
ZrCl2‚OEt2 were prepared following literature procedures.
ZrCl4 (Merck) was sublimed twice before use. HfCl4 (ACROS)
and TiCl4 (Aldrich) were used without further purifications.
Experimental details of Cp′′[(c-C5H9)7Si8O13]Ti(CH2Ph)2 (1) are
described in detail elsewhere.9

5672 Organometallics, Vol. 17, No. 26, 1998
Duchateau et al.
1:2:2:1:1), 2.00 (s, Si(CH3)3). 29Si NMR (THF, 283 K, δ): 8.18
(SiMe3), -64.10, -64.79, -65.88, -67.28, -68.40, -69.10 (1:
1:1:1:2:1:1). Anal. Calcd for C84H160O26Si16Zr: C, 47.44; H, 7.58;
Zr, 4.29. Found: C, 46.85; H, 7.45; Zr, 4.36.
the stirred mixture was allowed to warm to room temperature,
yielding a colorless solution. After stirring overnight, the
solvent was evaporated, leaving a white solid. To remove traces
of toluene, the sticky solid was dissolved in ether (10 mL) and
subsequently dried in a vacuum. This procedure was repeated
(2×) until a nonsticky product was obtained. The colorless foam
obtained was dissolved in hexane (30 mL). Within a few
minutes crystals appeared, and after subsequent cooling to
-30 °C, dimeric 9 was isolated as a white microcrystalline
material (1.70 g, 0.74 mmol, 43%). 1H NMR (benzene-d6, δ):
[(c-C5H9)7Si7O11(OSiMe3)]Zr Cl2‚2THF (6). At room tem-
perature, solid (c-C5H9)7Si7O9(OSiMe3)(OH)2 (II, 4.62 g 4.88
mmol) was added to a stirred THF (50 mL) solution of
(PhCH2)2ZrCl2‚OEt2 (2.02 g, 4.83 mmol). Immediately, the
yellow solution decolorized. After 5 min, the solvent was
evaporated and the product was dried thoroughly. The product
was dissolved in dichloromethane (20 mL) and filtered.
Concentration and cooling to -30 °C yielded 6 (2.50 g, 2.00
mmol, 41%) as colorless crystals. The mother liquor was
evaporated to dryness, and the remaining solid was dissolved
in hexane (10 mL). After filtration the clear solution was
concentration and cooled to -30 °C, yielding a second crop of
6 as microcrystalline material (1.3 g, 1.04 mmol, 22%). 1H
NMR (CDCl3, δ): 4.34 (m, 8H, THF-R-CH2), 2.03 (m, 8H,
THF-â-CH2), 1.77 (m, 14H, CH2-C5H9), 1.61 (m, 14H, CH2-
C5H9), 1.51 (m, 28H, CH2-C5H9), 0.97 (m, 7H, CH-C5H9), 0.10
(s, 9H, Si(CH3)3). 13C{1H} NMR (CDCl3, δ): 73.2 (s-br, THF-
R-CH2), 27.75, 27.58, 27.35, 27.19, 27.00, 26.96, 26.90, 26.87
(s, CH2-C5H9), 25.37 (s, THF-â-CH2), 24.85, 23.54, 23.21,
22.52, 22.33 (s, CH-C5H9, 1:2:2:1:1), 1.99 (s, Si(CH3)3). Anal.
Calcd for C46H88Cl2O14Si8Zr: C, 44.13, H, 7.08. Found: C,
43.84; H, 6.99.
3
3
7.43 (d, 2H, o-C6H5, J H-H ) 7 Hz), 7.28 (t, 2H, m-C6H5, J H-H
3
) 7 Hz), 6,87 (t, 1H, p-C6H5, J H-H ) 7 Hz), 2.63 (s, 2H, Hf-
CH2), 1.7 (m, 56H, CH2-C5H9), 1.2 (m, 6H, CH-C5H9), 0.9 (m,
1H, CH-C5H9). 13C{1H} NMR (benzene-d6, δ): 148.24 (s, ipso-
C6H5), 128.5 (s, C6H5), 128.07 (s, C6H5), 122.26 (s, C6H5), 65.39
(t, Hf-CH2, 1J C-H ) 112 Hz), 28.26, 28.05, 27.85, 27.73, 27.47
(s, CH2-C5H9), 23.67, 22.90, 22.51 (s, CH-C5H9, 3:3:1). Anal.
Calcd for {C42H70HfO12Si7}2: C, 44.17; H, 6.18. Found: C,
43.77; H, 6.30.
[(c-C5H9)7Si7O12]Zr CH2P h ‚2THF (10). Zr(CH2Ph)4 (0.95 g,
2.08 mmol) was dissolved in THF (20 mL) and cooled to -90
°C. Then, solid (c-C5H9)7Si7(OH)3 (1.82 g, 2.08 mmol) was
added, and the mixture was slowly warmed to room temper-
ature. The solvent of the pale yellow solution was evaporated,
and the crude product was dried thoroughly and extracted (30
mL) with hexane. Concentration and cooling to 4 °C afforded
10 as colorless crystals (0.90 g, 0.75 mmol, 36%). 1H NMR
(benzene-d6, δ): 7.25 (m, 4H, C6H5) 6.86 (m, 1H, C6H5), 3.94
(s (br), 8H, THF-R-CH2), 2.17 (s, 2H, CH2Ph), 1.95, 1.81, 1.72,
1.58, 1.51 (m, 56H, CH2-C5H9), 1.44 (s (br), 8H, THF-â-CH2),
1.21, 1.06, 0.89 (m, 7H, CH-C5H9). 13C NMR (benzene-d6, δ):
152.74 (s, ipso-C6H5), 128.16 (d, C6H5, 1J C-H ) 158 Hz), 126.26
[(c-C5H9)7Si7O12]TiCH2P h (7). A red solution of Ti(CH2-
Ph)4 (0.88 g, 2.13 mmol) in toluene (25 mL) was cooled to -80
°C and subsequently (c-C5H9)7Si7O9(OH)3 (III: 1.87 g, 2.14
mmol) was added. The resulting solution was allowed to warm
to room temperature and stirred for an additional hour. To
remove traces of toluene, the sticky solid was dissolved in
hexane (10 mL) and subsequently dried in a vacuum. This
procedure was repeated (3×) until a nonsticky product was
obtained. Recrystallization from hexane afforded 7 as a yellow
powder (1.91 g, 1.89 mmol, 89%). 1H NMR (benzene-d6, δ): 7.15
(m, 4H, C6H5), 6.86 (m, 1H, C6H5), 2.99 (s, 2H, CH2Ph), 1.88
(m, 14H, CH2-C5H9), 1.67 (m, 28H, CH2-C5H9), 1.48 (m, 14H,
CH2-C5H9), 1.15 (m, 7H, CH-C5H9). 13C{1H} NMR (benzene-
d6, δ): 142.92 (ipso-C6H5), 128.65 (C6H5), 128.59 (C6H5), 124.32
1
1
(d, C6H5, J C-H ) 159 Hz), 119.31 (d, C6H5, J C-H ) 151 Hz),
71.02 (t, THF-R-CH2, 1J C-H ) 146 Hz), 54.21 (t, CH2Ph, 1J C-H
) 117 Hz), 28.37, 27.99, 27.86, 27.55, 27.52, 27.42 (t, CH2-
C5H9, 1J C-H ) 129 Hz), 25.30 (t, THF-â-CH2, 1J C-H ) 133 Hz),
1
23.80, 23.17, 22.89 (d, CH-C5H9, J C-H ) 117 Hz). 29Si NMR
(THF, δ): -66.62, -66.74, -68.40 (3:1:3). The crystalline
product gradually lost THF. Therefore no satisfactory elemen-
tal analysis could be obtained.
1
(C6H5), 81.53 (t, CH2, J C-H ) 131 Hz), 27.90, 27.87, 27.78,
Hyd r ogen a tion of 1-Hexen e. Hydrogenation of 1-hexene
was carried out in pressure NMR tubes containing toluene
solutions of 40-60 µmol of the precursor (7-10) and 1 mmol
of 1-hexene under 4 atm of dihydrogen. For 8 and 9, after 10-
20 h at 75 °C toluene was quantitatively formed and all the
1-hexene was selectively hydrogenated to n-hexane. In the case
of 7 and 10, no observable reaction had occurred under the
same conditions.
27.47 (CH2-C5H9), 22.73, 22.65, 22.41 (CH-C5H9, 3:1:3). 29Si
NMR (toluene, δ): -64.59, -66.55, -67.68 (3:1:3). Anal. Calcd
for C42H70O12Si7Ti: C, 49.87; H, 6.98. Found: C, 48.66; H, 6.45.
{[(c-C5H9)7Si7O12]Zr CH2P h }2 (8). At -80 °C, a solution of
Zr(CH2Ph)4 (3.05 g, 6.69 mmol) in toluene (30 mL) was added
to a suspension of (c-C5H9)7Si7O9(OH)3 (5.85 g, 6.68 mmol) in
toluene (50 mL). The mixture was allowed to warm to room
temperature and subsequently stirred for 1 h. Evaporation of
the volatiles yielded crude 8 as a oily yellow solid. To remove
traces of toluene, the sticky solid was dissolved in hexane (10
mL) and subsequently dried in a vacuum. Then the product
was dissolved in hexane (50 mL) and filtered to remove minor
amounts of insoluble impurities. Concentration and cooling to
-30 °C gave dimeric 8 (5.2 g, 2.47 mmol, 74%) as pale yellow
block-shaped crystals. Recrystallization from hexane at -30
°C afforded crystals suitable for an X-ray structure determi-
nation. 1H NMR (benzene-d6, δ): 7.47 (d, 2H, C6H5, 3J H-H ) 7
Hz), 7.28 (dd, 2H, C6H5, 3J H-H ) 7 Hz), 6.99 (d, 1H, C6H5, 3J H-H
) 7 Hz), 3.12 (s, 2H, CH2C6H5, 1.7 (m, 50H, C5H9), 1.2 (m, 6H,
C5H9), 1.1 (m, 4H, C5H9), 0.8 (m, 3H, C5H9). 13C NMR (benzene-
NMR Tu be Rea ction of 8 w ith B(C6F 5)3. NMR tubes were
charged with toluene-d8 solutions of 8 (75-95 mg, 70-90 µmol)
and variable quantities (1, 2, and 10 equiv) of B(C6F5)3. The
reactions were followed by 1H, 11B, and 19F NMR. The NMR
(1H, 11B, 19F) spectra recorded within 5 min after the addition
of 1 equiv of B(C6F5)3 to a toluene-d8 solution of 8 showed the
instantaneous formation of a new zirconium benzyl complex
(ZrCH2 ) 2.73 ppm) and [PhCH2B(C6F5)3]- as the only boron
product (1H, BCH2 ) 3.52 ppm, 11B, PhCH2B(C6F5)3 ) -16.7
ppm). 19F NMR indicates that this borate anion is not
coordinated to the cationic zirconium complex: ∆δ(Fm - Fp)
) 2.59 ppm. Within 30 min at 50 °C, the cationic complex has
decomposed to unidentified product(s). When treated with 2
or 10 equiv of B(C6F5)3, complex 8 still reacted with only 1
equiv of B(C6F5)3, leaving the excess borane untouched.
Eth ylen e P olym er iza tion Exp er im en ts. Ethylene po-
lymerization experiments were carried out in a 1.3 L DSM
stainless steel reactor equipped with a stationary steel stirrer.
In a typical experiment, under nitrogen PMH (pentamethyl-
heptane, 600 mL) was brought into the reactor. The reactor
was heated under stirring to the required temperature.
Additionally, a constant ethylene pressure was maintained.
A toluene (10 mL) solution of the cocatalyst was syringed into
1
d6, δ): 142.3 (s, ipso-C6H5), 129.9 (d, C6H5, J C-H ) 160 Hz),
1
1
124.7 (d, C6H5, J C-H ) 157 Hz), 57.6 (t, CH2C6H5, J C-H
)
124 Hz), 13C{1H}: 28.4, 28.1, 27.8, 27.6, 27.5, 27.2, 25.6, 23.6,
23.0, 22.8, 22.6 (s, C5H9). 29Si NMR (toluene, 213 K, δ): -57.23,
-61.48, -62.05, -62.98, -65.14, -65.57, -67.07. Anal. Calcd
for {C42H70O12Si7Zr}2: C, 47.82; H, 6.69. Found: C, 48.31; H,
7.05.
{[(c-C5H9)7Si7O12]HfCH2P h }2 (9). A solution of Hf(CH2Ph)4
(1.88 g, 3.46 mmol) in toluene (25 mL) was cooled to -50 °C.
Solid (c-C5H9)7Si7O9(OH)3 (3.00 g, 3.43 mmol) was added, and

Ethylene Polymerization
Organometallics, Vol. 17, No. 26, 1998 5673
a 100 mL catalyst dosing vessel containing PMH (25 mL).
Subsequently, the solution was transferred into the reactor.
After 15 min, the desired amount of catalyst precursor was
introduced into the reactor in the same way, after which the
catalyst dosing vessel was washed automatically with PMH
(2 × 50 mL). The polymerization reactions were carried out
under isothermal conditions and constant ethene pressure. At
the end of the reaction, the reaction mixture was collected and
quenched with 20 mL of methanol. Antioxidant (Irganox 1076)
was added to stabilize the polymer. The polymer was dried
under vacuum at 70 °C for 24 h. If possible, the obtained
polymer was analyzed by SEC-DV (weight-averaged molecular
weight (Mw) and molecular weight distribution (MWD).
The positional and anisotropic thermal displacement param-
eters for the non-hydrogen atoms were refined on F2 with full-
matrix least-squares procedures minimizing the function Q )
∑h[w(|Fo - kFc |)2], where w ) 1/[σ2(Fo2) + (aP)2 + bP], P )
2
2
2
[max(Fo2,0) + 2Fc ]/3, Fo and Fc are the observed and calculated
structure factor amplitudes, respectively; a and b were refined.
Reflections were stated observed if satisfying F2 > 0 criterion
of observability. For 5, a subsequent Fourier synthesis showed
some electron density (not exceeding 2.9 e/Å3) in the holes of
the lattice which could be correlated to partly occupied and
disordered solvent molecules of THF. Using the BYPASS
procedure, the amount of THF was found to be 0.72 per
molecule of 5. For both 5 and 8 refinement was complicated
by conformational disorder of the cyclopentyl substituents.
Final refinement on F2 was carried out by full-matrix least-
squares techniques. Neutral atom scattering factors and
anomalous dispersion corrections were taken from Interna-
tional Tables of Crystallography.36 Al calculations were carried
out at a HP9000/735 computer at the University of Groningen
with the program packages SHELXL,37 PLATON,38 and
ORTEP.39
X-r a y Str u ctu r e Deter m in a tion of [(c-C5H9)7Si7O11
-
(OSiMe3)]2Zr ‚2THF (5) a n d {[(c-C5H9)7Si7O12]Zr CH2C6H5}2
(8). Suitable crystals were measured at 130 K with graphite-
monochromated Mo KR radiation on an Enraf-Nonius CAD-
4F diffractometer equipped with a low-temperature unit.31
Precise lattice parameters and their standard deviation and
orientation matrix were derived from the angular settings of
22 reflections (5, 18.21° < θ < 20.41°; 8, 18.07° < θ < 20.57°)
in four alternative settings.32 The unit cell was identified as
monoclinic (space group P21/c) for 5 and triclinic (space group
P1h) for 8.33 Intensity data were corrected for Lorentz and
Ack n ow led gm en t. We wish to thank DSM Re-
search B.V. for funding this research. We also thank
Dr. A. J . R. L. Hulst, University of Twente, The
Netherlands, for measuring the 19F NMR spectra.
2
34
polarization effects and scale variation and reduced to Fo
.
The structures were solved by Patterson methods, and exten-
sion of the model was accomplished by direct methods applied
to difference structure factors using the program DIRDIF.35
Su p p or tin g In for m a tion Ava ila ble: Tables of atomic
coordinates, thermal displacement parameters, bond lengths,
bond angles, and torsion angles for 5 and 8 (34 pages).
Ordering information is given on any current masthead page.
(30) (a) Zuccchini, U.; Albizzati, E.; Giannini, U. J . Organomet.
Chem. 1971, 26, 357. (b) Wngrovius, J . H.; Schrock, R. R.; J .
Organomet. Chem. 1981, 305, 319.
(31) Van Bolhuis, F. J . Appl. Crystallogr. 1971, 4, 263.
(32) De Boer, J . L.; Duisenberg, A. J . M. Acta Crystallogr. 1984,
A40, C410.
(33) No higher metric lattice symmetry or extra metric symmetry
elements were found: (a) Spek, A. L. J . Appl. Crystallogr. 1988, 21,
578. (b) Le Page, Y. J . Appl. Crystallogr. 1987, 20, 264. (c) Le Page, Y.
J . Appl. Crystallogr. 1988, 21, 983.
(34) Spek, A. L. HELENA Program for Data Reduction; University
of Utrecht: Utrecht, The Netherlands, 1993.
(35) Beurskens, P. T.; Beurskens, G.; Bosman, W. P.; de Gelder, R.;
Garc´ıa-Granda, S.; Gould, R. O.; Israe¨l, R.; Smits, J . M. M. The
DIRDIF-97 program system; Crystallography Laboratory, University
of Nijmegen: Nijmegen, The Netherlands, 1992.
OM980687K
(36) International Tables of Crystallography, Vol. C.; Wilson, A. J .
C., Ed.; Kluwer Academic Publishers: Dordrecht, The Netherlands,
1992.
(37) Scheldrick, G. M. SHELX-97. Program for the Solution and
Refinement of Crystal Structures; University of Go¨ttingen: Go¨ttingen,
Germany, 1997.
(38) Spek, A. L. PLATON. Program for the Automated Analysis of
Molecular Geometry; University of Utrecht: Utrecht, The Netherlands,
1996.
(39) J ohnson, C. K. ORTEPII. Report ORNL-5138; Oak Ridge
National Laboratory: Oak Ridge, TN, 1976.