Catalytic trimerization of ethene with highly active cyclopentadienyl-arene titanium catalysts

Article abstract of DOI:10.1021/om020765a

The mono(cyclopentadienyl-arene)titanium complexes [η⁵-C₅H₃R-(bridge)-Ar]TiCl₃, activated by methylalumoxane (MAO) cocatalyst, form a family of highly active catalysts for the trimerization of ethene, giving 1-hexene as the main product. Concomitant cotrimerization of ethene and 1-hexene, to give mainly 5-methylnon-1-ene, is also observed. The selectivity for trimerization depends on the presence of a pendant arene group on the cyclopentadienyl ligand and the nature of the bridge between these two ligand moieties. In the absence of a pendant arene, polyethene is the main product. The highest activity and selectivity for trimerization was observed for catalysts with a disubstituted C₁ bridge between the cyclopentadienyl and arene ligand moieties. A SiMe₃ substituent on the cyclopentadienyl ligand improves catalyst activity and selectivity, whereas methyl substitution of the arene decreases activity. Nevertheless, combining cyclopentadienyl SiMe₃ substitution with arene Me substitution gives rise to a catalyst with the highest activity and selectivity, evidence of the strongly nonlinear additivity of ligand substituent effects in this system. The cyclopentadienyl-arene ligand is likely to exhibit hemilabile behavior during catalysis, stabilizing intermediates by η⁶ coordination and dissociating or slipping to make room for the incoming substrate. The presence of two pendant arene groups on the cyclopentadienyl ligand diminishes the activity of the catalyst but greatly enhances its stability.

Full text of DOI:10.1021/om020765a

5122
Organometallics 2002, 21, 5122-5135
Ca ta lytic Tr im er iza tion of Eth en e w ith High ly Active
Cyclop en ta d ien yl-Ar en e Tita n iu m Ca ta lysts
Patrick J . W. Deckers, Bart Hessen,* and J an H. Teuben
Dutch Polymer Institute/ Center for Catalytic Olefin Polymerization,
Stratingh Institute for Chemistry and Chemical Engineering,
University of Groningen, Nijenborgh 4, NL-9747AG Groningen, The Netherlands
Received September 13, 2002
The mono(cyclopentadienyl-arene)titanium complexes [η5-C5H3R-(bridge)-Ar]TiCl3, ac-
tivated by methylalumoxane (MAO) cocatalyst, form a family of highly active catalysts for
the trimerization of ethene, giving 1-hexene as the main product. Concomitant cotrimerization
of ethene and 1-hexene, to give mainly 5-methylnon-1-ene, is also observed. The selectivity
for trimerization depends on the presence of a pendant arene group on the cyclopentadienyl
ligand and the nature of the bridge between these two ligand moieties. In the absence of a
pendant arene, polyethene is the main product. The highest activity and selectivity for
trimerization was observed for catalysts with a disubstituted C1 bridge between the
cyclopentadienyl and arene ligand moieties. A SiMe3 substituent on the cyclopentadienyl
ligand improves catalyst activity and selectivity, whereas methyl substitution of the arene
decreases activity. Nevertheless, combining cyclopentadienyl SiMe3 substitution with arene
Me substitution gives rise to a catalyst with the highest activity and selectivity, evidence of
the strongly nonlinear additivity of ligand substituent effects in this system. The cyclopen-
tadienyl-arene ligand is likely to exhibit hemilabile behavior during catalysis, stabilizing
intermediates by η6 coordination and dissociating or slipping to make room for the incoming
substrate. The presence of two pendant arene groups on the cyclopentadienyl ligand
diminishes the activity of the catalyst but greatly enhances its stability.
In tr od u ction
these catalysts is as yet unknown, and it is difficult to
control catalyst performance by well-directed catalyst
modifications, although a recent report on diphosphine/
CrCl3/MAO catalysts showed strong ligand effects on
catalyst activity and selectivity.5 Two catalyst systems
for the trimerization of ethene based on group 5 metals
(V,6 Ta)7 have been reported, but the activities of these
catalysts are orders of magnitude lower than those of
the Cr-based Ziegler-type catalysts.
Linear 1-alkenes are important chemical intermedi-
ates that are used in the production of detergent
alcohols (C6-C16) and synthetic lubricants, and as
comonomers in the production of linear low-density
polyethene (LLDPE, C6 and C8).1 They are largely
synthesized by catalytic oligomerization of ethene,
yielding a Flory-Schultz distribution of chain lengths.2
The selective synthesis of one specific linear alkene from
ethene would be very attractive but (apart from the
trivial dimerization of ethene to 1-butene)3 is difficult
to perform catalytically. The selective trimerization of
ethene to give 1-hexene as the main product can be
achieved with chromium Ziegler-type catalysts, consist-
ing of a combination of Cr(III) salts (usually carboxy-
lates) with aluminum alkyls in conjunction with some
Lewis basic donor (especially pyrroles or 1,2-diethoxy-
ethane).4 The precise nature of the active species in
(3) For examples of ethene (and R-olefin) dimerization, see: (a)
Muthukumaru Pillai, S.; Ravindranathan, M.; Sivaram, S. Chem. Rev.
1986, 86, 353. (b) Parshall, G. W.; Ittel, S. D. Homogeneous Catalysis;
Wiley: New York, 1992; p 72. (c) Chauvin, Y.; Olivier, H. In Applied
Homogeneous Catalysis with Organometallic Compounds; Cornils, B.,
Herrmann, W. A., Eds.; VCH: Weinheim, Germany, 1996; p 258. (d)
Britovsek, G. J . P.; Keim, W.; Mecking, S.; Sainz, D.; Wagner, T. J .
Chem. Soc., Chem. Commun. 1993, 1632. (e) Christoffers, J .; Bergman,
R. G. Inorg. Chim. Acta 1998, 270, 20. (f) Piers, W. E.; Shapiro, P. J .;
Bunel, E. E.; Bercaw, J . E. Synlett 1990, 74. (g) Hauptman, E.; Sabo-
Etienne, S.; White, P. S.; Brookhart, M.; Garner, J . M.; Fagan, P. J .;
Calabrese, J . C. J . Am. Chem. Soc. 1994, 116, 8038.
(4) (a) Briggs, J . R. U.S. Pat. Appl. 4668838, 1987. (b) Karol, F. J .;
Levine, I. J . U.S. Pat. Appl. 4777315, 1988. (c) Reagan, W. K.; Conroy,
B. K. Eur. Pat. Appl. EP-416304, 1990. (d) Reagan, W. K.; Conroy, B.
K. U.S. Pat. Appl. 5198563, 1993. (e) Tamura, M.; Uchida, K.; Iwanaga,
K.; Ito, Y. Eur. Pat. Appl. EP-699648, 1995. (f) Reagan, W. K.;
Pettijohn, T. M.; Freeman, J . W. U.S. Pat. Appl. 5523507, 1996. (g)
Araki, Y.; Nakamura, H.; Nanba, Y.; Okano, T. U.S. Pat. Appl.
5856612, 1997. (h) Wu, F.-J . U.S. Pat. Appl. 5744677, 1998. (i) Wu,
F.-J . U.S. Pat. Appl. 5968866, 1999. (j) Wass, D. F. PCT Int. Appl.
WO 0204119, 2002.
(5) Carter, A.; Cohen, S. A.; Cooley, N. A.; Murphy, A.; Scutt, J .;
Wass, D. F. Chem. Commun. 2002, 858.
(6) Romano, A. M.; Proto, A.; Santi, R.; Sommazzi, A.; Grande, M.;
Masi, F. PCT Int. Appl. WO 0168572, 2001.
* To whom correspondence should be addressed. E-mail: hessen@
chem.rug.nl.
(1) Vogt, D. In Applied Homogeneous Catalysis with Organometallic
Compounds; Cornils, B., Herrmann, W. A., Eds.; VCH: Weinheim,
Germany, 1996; p 245. (b) Parshall, G. W.; Ittel, S. D. Homogeneous
Catalysis; Wiley: New York, 1992; p 56.
(2) For examples, see: (a) Skupinska, J . Chem. Rev. 1991, 91, 613.
(b) Mecking, S. Coord. Chem. Rev. 2000, 203, 325. (c) Keim, W.;
Kowaldt, F. H.; Goddard, R.; Kru¨ger, C. Angew. Chem., Int. Ed. Engl.
1978, 17, 466. (d) Keim, W.; Schulz, R. P. J . Mol. Catal. 1994, 92, 21.
(e) Svejda, S. A.; Brookhart, M. Organometallics 1999, 18, 65. (f)
Killian, C. M.; J ohnson, L. K.; Brookhart, M. Organometallics 1997,
16, 2005. (g) Britovsek, G. J . P.; Mastroianni, S.; Solan, G. A.; Baugh,
S. P. D.; Redshaw, C.; Gibson, V. C.; White, A. J . P.; Williams, D. J .;
Elsegood, M. R. J . Chem. Eur. J . 2000, 6, 2221.
(7) Andes, C.; Harkins, S. B.; Murtuza, S.; Oyler, K.; Sen, A. J . Am.
Chem. Soc. 2001, 123, 7423.
10.1021/om020765a CCC: $22.00 © 2002 American Chemical Society
Publication on Web 10/19/2002

Catalytic Trimerization of Ethene
Organometallics, Vol. 21, No. 23, 2002 5123
Sch em e 1
in Scheme 2. Cyclopentadienyl ligands with -CR2Ar (R
) alkyl, Ar ) aryl) substituents are readily accessible
from the reaction of 6,6-dialkylfulvenes with the ap-
propriate aryllithium salts.12 The resulting lithium
cyclopentadienides, [C5H4CR2Ar]Li, can be optionally
quenched with trimethylsilyl chloride to afford the
corresponding (C5H4CR2Ar)SiMe3 reagents.9a Introduc-
tion of a -SiMe2Ph or -CH2Ph moiety on the cyclopen-
tadienyl ligand can be achieved by reaction of CpLi with
dimethylphenylsilyl chloride (PhMe2SiCl)9a or benzyl
chloride,13 respectively, to give C5H5SiMe2Ph or C5H5-
CH2Ph. Subsequent lithiation with n-BuLi affords the
corresponding lithium salts. Reaction of 6-phenyl-6-
methylfulvene with LDA (lithium diisopropylamide)
gives [C5H4C(dCH2)Ph]Li.14 All lithium salts and (tri-
methylsilyl)cyclopentadienyl reagents were character-
ized by NMR spectroscopy.
Recently, we communicated that a mono(cyclopenta-
dienyl)titanium catalyst system activated by methyla-
lumoxane (MAO) cocatalyst, (η5-C5H4CMe2Ar)TiCl3/
MAO (Ar ) Ph, 3,5-Me2C6H3), is able to trimerize ethene
with high activity.8 The active species appear to be
cationic titanium complexes in which the pendant arene
group coordinates to the metal center9 and in which the
cyclopentadienyl-arene ancillary ligand displays hemi-
labile behavior. We observed earlier that reaction of the
neutral (η5-C5H4CMe2-3,5-Me2C6H3)TiMe3 complex with
the Lewis acid B(C6F5)3 affords the cationic species [(η5:
η6-C5H4CMe2-3,5-Me2C6H3)TiMe2]+, in which the arene
moiety is coordinated to the titanium center (Scheme
1). In the presence of the Lewis base THF, the arene
moiety is displaced to give a [(η5-C5H4CMe2-3,5-Me2C6H3)-
TiMe2(THF)x]+ cationic species, thus behaving as a
substitutionally labile group.10
In this paper we describe the synthesis of a range of
(cyclopentadienyl-arene)titanium catalysts and evalu-
ate the effect of variations in the ligand system on the
performance in ethene trimerization catalysis. It is
shown that ligand variations can have a substantial
effect on the catalyst performance (activity, selectivity,
and stability). As a result of these studies, a proposal
for the pathways of generation and action of the species
responsible for the trimerization could be formulated,
and catalysts have been obtained that can produce >100
kg of ethene trimerization product/(g of Ti) under mild
reaction conditions (30 °C, 5 bar of ethene).
The various (cyclopentadienyl)titanium trichlorides
are readily prepared either via salt metathesis of the
appropriate (cyclopentadienyl)lithium reagent with TiCl4
in methylene chloride10 or via Me3SiCl elimination upon
reaction of the appropriate (trimethylsilyl)cyclopenta-
dienyl reagent with TiCl4.9a All titanium compounds
1
were characterized by H and 13C NMR spectroscopy
and microanalysis.
Eth en e Tr im er iza tion w ith (η5-C5H4CMe2P h )-
TiCl3 (1)/MAO. The results of catalytic ethylene con-
version experiments with the catalyst system (η5-
C5H4CMe2Ph)TiCl3 (1)/MAO (toluene solvent, Ti:Al ratio
of 1:1000) under various conditions are listed in Table
1. Analysis of the liquid fraction by GC (using cyclooc-
tane as internal standard), GC/MS, and NMR tech-
niques revealed that, under these conditions, the cata-
lyst produces olefin trimerization products with high
selectivity (>95 wt % overall). These trimerization
products consist of two fractions: C6 (trimers of ethene)
and C10 (cotrimers of ethene and 1-hexene, vide infra).
In addition to the trimerization products, smaller
amounts of C8 (1 wt %) and polyethene (PE, 1-3 wt %)
are produced.
Resu lts a n d Discu ssion
The rate of production of 1-hexene increases with
increasing ethene pressure. The C6 productivity is
around 550-600 kg/((mol of Ti) bar h) over a range of
2-10 bar of ethene pressure. Although no true kinetic
studies were undertaken, the roughly linear dependence
between the amount of C6 product formed over the fixed
run time of 30 min and the ethene pressure suggests
that the rate of the trimerization process has a first-
order dependence on the ethene concentration.
The C6 product fraction consists predominantly of
1-hexene (99.5%), with the remaining 0.5% being a
mixture of 2- and 3-hexenes. For entry 2 in Table 1, the
C10 fraction was isolated by evaporation of the low-
boiling (bp <80 °C at 180 Torr) volatile components of
the reaction mixture and analyzed separately by NMR
and GC/MS.15 1H NMR analysis of the olefinic residues
in this mixture indicates the presence of 90% RCH2CHd
CH2 end groups, 5% RCHdCHR′, and 5% RR′CdCH2.
Descr ip tion a n d Syn th esis of th e Ca ta lysts. The
titanium complexes used in this study can all be
described by the general formula [η5-C5H3R-(bridge)-
Ar]TiCl3. These mono(cyclopentadienyl)titanium trichlo-
rides can be converted into the cationic active catalyst
species by reaction with methylalumoxane (MAO) co-
catalyst. Alternatively, they can first be transformed
into their trimethyl derivatives, which can then be
reacted with various activators (Lewis acidic boranes,
Brønsted acid, or trityl salts of weakly coordinating
anions) that are commonly used with single-site olefin
polymerization catalysts.11
The routes employed for the synthesis of the [η5-
C5H3R-(bridge)-Ar]TiCl3 compounds are summarized
(8) Deckers, P. J . W.; Hessen, B.; Teuben, J . H. Angew. Chem., Int.
Ed. 2001, 40, 2516.
(9) (a) Sassmannshausen, J .; Powell, A. K.; Anson, C. E.; Wocadlo,
S.; Bochmann, M. J . Organomet. Chem. 1999, 592, 84. (b) Longo, P.;
Amendola, A. G.; Fortunato, E.; Boccia, A. C.; Zambelli, A. Macromol.
Rapid Commun. 2001, 22, 339. (c) Schwecke, C.; Kaminsky, W. J .
Polym. Sci. A: Polym. Chem. 2001, 39, 2805. (d) Flores, J . C.; Wood,
J . S.; Chien, J . C. W.; Rausch, M. D. Organometallics 1996, 15, 4944.
(10) Deckers, P. J . W.; Van der Linden, A. J .; Meetsma, A.; Hessen,
B. Eur. J . Inorg. Chem. 2000, 929.
(12) Renaut, P.; Tainturier, G.; Gautheron, B. J . Organomet. Chem.
1978, 148, 35.
(13) Erker, G.; Aul, R. Chem. Ber. 1991, 124, 1301.
(14) Erker, G.; Nolte, R.; Aul, R.; Wilker, S.; Kru¨ger, C.; Noe, R. J .
Am. Chem. Soc. 1991, 113, 7594.
(15) The yield of the C10 fraction was 3.8 g, including some residual
toluene solvent and cyclooctane standard, as seen by NMR.
(11) Bochmann, M. J . Chem. Soc., Dalton Trans. 1996, 255.

5124 Organometallics, Vol. 21, No. 23, 2002
Deckers et al.
Sch em e 2
GC/MS indicates that the C10 fraction mainly consists
of 5-methylnon-1-ene (83%). The only detectable product
in the C8 fraction (by GC) is 1-octene. DSC analyses of
the polyethene samples gave melting points higher than
128 °C, indicative of HDPE (high-density polyethene),16
suggesting that very little of the 1-hexene formed is
incorporated into the polymer.
The thermal stability of the catalyst system 1/MAO
is modest (Table 2): increasing the reaction temperature
decreases the overall catalyst productivity over the 30
min run period and increases the relative amount of PE
produced. This is likely to be associated with catalyst
(16) Karol, F. J . CHEMTECH 1983, 222.

Catalytic Trimerization of Ethene
Organometallics, Vol. 21, No. 23, 2002 5125
Ta ble 1. Ca ta lytic Eth en e Con ver sion w ith th e 1/MAO Ca ta lyst System a s a F u n ction of Eth en e P r essu r ea
P(ethene),
C6 products,
g (wt %)
C10 products,
g (wt %)
PE,
g (wt %)
productivity C6
productsb
trimerizn
products, wt %
trimerizn
bar
productivityc
2
5
10
8.0 (87)
20.9 (83)
47.2 (86)
1.0 (11)
3.5 (14)
5.1 (9)
0.2 (1.6)
0.5 (1.8)
1.4 (2.6)
535
555
630
98
97
95
41
110
239
a
b
Reaction conditions: toluene solvent, 30 °C, 15 µmol of Ti, Al:Ti ) 1000, 30 min reaction time. In g of C6 product/((mmol of Ti)
bar h). c In mol of olefinic bonds trimerized/((mmol of Ti) h).
Ta ble 2. Ca ta lytic Eth en e Con ver sion w ith th e 1/MAO Ca ta lyst System a s a F u n ction of Tem p er a tu r ea
C6 products,
g (wt %)
C10 products,
g (wt %)
PE, g
(wt %)
productivity
C6 productsb
trimerizn
products, wt %
trimerizn
T, °C
productivityc
30
50
80
20.9 (83)
12.4 (83)
3.3 (76)
3.5 (14)
1.6 (11)
0.2 (4)
0.5 (1.8)
0.7 (4.6)
0.8 (19)
555
330
90
97
94
80
110
64
17
a
b
Reaction conditions: toluene solvent, 5 bar of ethene pressure, 15 µmol of Ti, Al:Ti ) 1000, 30 min reaction time. In g of C6 product/
((mmol of Ti) bar h). c In mol of olefinic bonds trimerized/((mmol of Ti) h).
Ta ble 3. Ca ta lytic Eth en e Con ver sion w ith th e 1/MAO Ca ta lyst System in th e P r esen ce of Ad d ed 1-Octen ea
P(ethene),
1-octene,
g
C6 products,
g (wt %)
C10 products,
g (wt %)
C12 products,
g (wt %)
PE,
g (wt %)
productivity
C6 productsb
trimerizn
products, wt %
trimerizn
bar
productivityc
2
2
5
5
0
15
0
8.0 (87)
6.3 (69)
20.9 (83)
16.6 (75)
1.0 (11)
0.7 (8)
3.5 (14)
2.2 (10)
0.03 (0.4)
2.0 (22)
0.1 (0.4)
2.7 (12)
0.2 (1.6)
0.1 (1.1)
0.5 (1.8)
0.5 (2.3)
535
425
555
445
98
99
97
97
41
37
110
92
15
a
b
Reaction conditions: toluene solvent, 30 °C, 15 µmol of Ti, Al:Ti ) 1000, 30 min reaction time. In g of C6 product/((mmol of Ti)
bar h). c In mol of olefinic bonds trimerized/((mmol of Ti) h).
Ta ble 4. Ca ta lytic Eth en e Con ver sion w ith th e (η5-C5H4CMe2R)TiCl3/MAO Ca ta lyst System sa
C6 products,
g (wt %)
C10 products,
g (wt %)
PE, g
(wt %)
productivity
C6 productsb
trimerizn
products, wt %
trimerizn
catalyst (R)
1 (Ph)
2 (4-MeC6H4)
3 (3,5-Me2C6H3)
4 (Me)
productivityc
20.9 (83)
14.7 (88)
7.9 (93)
0.5 (17)
3.5 (14)
1.6 (10)
0.1 (4)
0.1 (4)
0.5 (1.8)
0.3 (1.6)
0.1 (1.3)
2.4 (76)
555
395
210
13
97
98
97
21
110
75
38
3
a
b
Reaction conditions: toluene solvent, 5 bar of ethene, 30 °C, 15 µmol of Ti, Al:Ti ) 1000, 30 min reaction time. In g of C6 product/
((mmol of Ti) bar h). c In mol of olefinic bonds trimerized/((mmol of Ti) h).
degradation, as suggested by the more rapid decrease
in ethene uptake rate during the runs at elevated
temperature. At 80 °C, ethene uptake stops completely
after about 10 min of reaction time, whereas at 30 °C
ethene is consumed over the whole run period, although
the uptake rate does slow gradually over the course of
the run.
The C10 product fraction is likely to be produced by
cotrimerization of ethene and 1-hexene rather than by
direct pentamerization of ethene. In separate experi-
ments, ethene was converted by the catalyst 1/MAO at
30 °C in the presence of 15 g of 1-octene added to the
reaction mixture (Table 3). In addition to the C6 and
C10 product fractions mentioned above, a considerable
amount of C12 products (cotrimers of ethene and 1-octene,
obtained in 2-3 g quantities, depending on the ethene
pressure) is now observed as well. This is likely to stem
from ethene/1-octene cotrimerization, and these results
indicate that the C10 product fraction in the catalytic
ethene oligomerization by 1/MAO is indeed formed via
cotrimerization of ethene and 1-hexene.
It can be seen that making the pendant phenyl group
more electron-rich, by adding one or two methyl sub-
stituents, respectively, significantly diminishes the
productivity of the catalyst with each methyl group
added, whereas the selectivity for trimerization is
retained. This reduction of catalyst productivity upon
methyl substitution of the pendant group reflects a
slowing down of the catalytic cycle, as no evidence was
found for a more rapid catalyst deactivation in these
systems. The absence of the pendant arene group leads
to the predominant formation of polyethene, indicating
that the pendant arene group is essential to obtaining
selective trimerization. These observations suggest that
the cyclopentadienyl-arene ligand is likely to display
hemilabile character in the course of the catalysis: it
has to be present for coordination, apparently helping
to generate the species responsible for the selective
trimerization, but if it binds too strongly it can slow the
catalytic reaction.
It may be noted that even for the tert-butylcyclopen-
tadienyl system 4/MAO a certain amount of ethene
trimerization product is observed in addition to the main
product, polyethene. A related observation was made
recently by Pellecchia and co-workers in ethene polym-
erization with the [(η5-C5Me5)TiMe2][MeB(C6F5)3] cata-
lyst in toluene solvent, where the PE obtained contains
a noticeable amount of n-butyl side groups17 It was
suggested that the catalyst is partly converted to a
species that trimerizes ethene to 1-hexene, which is then
Effect of th e P en d a n t Ar en e Gr ou p on Eth en e
Tr im er iza tion Ca ta lysis. The effect of the pendant
arene group on ethene conversion by the (η5-C5H4CMe2-
Ar)TiCl3/MAO catalysts was probed by studying the
catalysts with Ar ) Ph (1), 4-MeC6H4 (2), 3,5-Me2C6H3
(3) and comparing them with an analogous system
without a pendant arene, (η5-C5H4CMe3)TiCl3 (4). The
results are listed in Table 4.

5126 Organometallics, Vol. 21, No. 23, 2002
Deckers et al.
Ta ble 5. Ca ta lytic Eth en e Con ver sion w ith th e 1/MAO Ca ta lyst System in Tw o Differ en t Solven tsa
toluene,
vol %
n-octane,
vol %
C6 products,
g (wt %)
C10 products,
g (wt %)
PE,
g (wt %)
productivity
C6 productsb
trimerizn
products, wt %
trimerizn
productivityc
100
20
0
80
20.9 (83)
8.1 (87)
3.5 (14)
0.8 (9)
0.5 (1.8)
0.4 (4.3)
555
215
97
96
110
41
a
b
Reaction conditions: 5 bar of ethene pressure, 30 °C, 15 µmol of Ti, Al:Ti ) 1000, 30 min reaction time. In g of C6 product/((mmol
of Ti) bar h). c In mol of olefinic bonds trimerized/((mmol of Ti) h).
incorporated into the polymer. Our observations indicate
that transient coordination of the toluene solvent to the
metal center may be instrumental in this,18 albeit less
efficient than the pendant arene group in the (η5-C5H4-
CMe2Ar)TiCl3/MAO systems, as the interaction is in-
termolecular rather than intramolecular.
involve metallacyclic intermediates and a Ti(II)/Ti(IV)
couple.21 It is less obvious how the required (cationic)
low-valent Ti(II) species would be generated from a
mono(cyclopentadienyl)titanium(IV) trichloride complex
and MAO.
One possibility is that the cationic [(Cp-arene)Ti-
(CH2CH2R)2]+ species, likely to be generated initially
by alkylation of the (Cp-arene)TiCl3 by MAO followed
by alkyl anion abstraction and sequential ethene inser-
tions into the Ti-Me bonds, is in equilibrium with its
hydride-olefin isomer [(Cp-arene)Ti(H)(η2-CH2dCHR)-
(CH2CH2R)]+. This may be followed either by dissocia-
tion of the alkene (to yield a cationic hydrido alkyl Ti-
(IV) intermediate that is likely to undergo subsequent
reductive elimination of the alkane to give a cationic
Ti(II) species) or by reductive elimination of the alkane
(to yield a cationic Ti(II) olefin complex). In both
pathways, the coordination of the arene moiety can
provide additional stabilization to the Ti(II) state.
Presently it is impossible for us to distinguish between
the two pathways (although the latter appears to be the
more likely, as it avoids the formation of a “naked”
cationic (Cp-arene)Ti(II) species). A theoretical study
of the process using DFT calculations is presently in
progress. The catalytic cycle as proposed on the basis
of the information available at this moment is sum-
marized in Scheme 3.
To estimate the possible role of solvent interactions
on the selective ethene trimerization by the (cyclopen-
tadienyl-arene)titanium catalysts, we investigated the
catalyst system 1/MAO in an 80:20 (v/v) mixture of
n-octane and toluene (Table 5). The ethene conversion
results showed that the selectivity for trimerization is
largely retained but that the catalyst productivity over
the 30 min run is lower than in neat toluene. The initial
ethene uptake rate is identical with that of the corre-
sponding run in neat toluene, but the catalyst appears
to deactivate more rapidly in the n-octane/toluene
mixture. These findings indicate that the intrinsic
trimerization process and its selectivity do not depend
on the aromatic solvent but that having an aromatic
solvent may be beneficial for the catalyst stability.
P r op osed Ca ta lytic Cycle for Eth en e Tr im er iza -
tion . It has been proposed that the only family of
catalysts known thus far to perform efficient and
selective ethene trimerization (Ziegler-type catalysts
based on the combination of chromium salts with
aluminum alkyls and added Lewis bases)4 acts through
a mechanism involving metallacyclic intermediates.
This may proceed through initial oxidative coupling of
two ethene molecules by a low-valent chromium species,
generated in situ, to produce a chromacyclopentane
compound. Insertion of an additional ethene molecule
into one of the Cr-C bonds then affords a chromacy-
cloheptane species. Subsequent â-H abstraction and
reductive elimination can then lead to the formation of
1-hexene and the regeneration of the low-valent chro-
mium species.19
This sequence explains the lack of 1-butene formation,
as the chromacyclopentane is expected to be much more
stable toward â-H abstraction than the more flexible
chromacycloheptane.20 For the very recently reported
selective ethene trimerization by TaMe2Cl3, a similar
catalytic cycle via metallacycles has been proposed.7
It is likely that a similar mechanism is operative in
the selective trimerization performed by the (cyclopen-
tadienyl-arene)titanium catalysts presented here, in-
volving cationic Ti(IV) metallacyclic intermediates and
the Ti(II)/Ti(IV) couple for oxidative coupling/reductive
elimination. Several catalytic C-C coupling reactions
have been reported in neutral titanium systems that
If the rate of 1-hexene formation is indeed first order
in ethene (vide supra), it is very likely that the insertion
of the third ethene molecule into one of the metal-
carbon bonds of the proposed 16-electron cationic ti-
tanacyclopentane intermediate is the rate-determining
step. This reaction will require the displacement or
slippage of the η6-coordinated arene ligand to yield a
less electron-rich metal center that can capture and
insert the third ethene molecule. It is thus possible that
the process of arene dissociation or ring slippage from
a cationic Ti(IV) species is involved in the rate-
determining step, which would be consistent with the
observed decrease in catalyst activity upon increasing
the donor ability of the pendant arene group.
Kinetic studies of chromium-based ethene trimeriza-
tion catalysts have revealed a second-order rate depen-
dence on ethene concentration, indicating that in these
systems the formation of the chromacyclopentane (or a
bis(ethene) adduct leading to this species) is likely to
be rate-determining.22 However, the selective trimer-
ization of 1-hexene with triazacyclohexane-chromium
catalysts was recently reported to have as a rate-
(17) Pellecchia, C.; Pappalardo, D.; Gruter, G. J . Macromolecules
1999, 32, 4491.
(18) Gillis, D. J .; Tudoret, M. J .; Baird, M. C. J . Am. Chem. Soc.
1993, 115, 2543.
(21) (a) Thorn, M. G.; Hill, J . E.; Waratuke, S. A.; J ohnson, E. S.;
Fanwick, P. E.; Rothwell, I. P. J . Am. Chem. Soc. 1997, 119, 8630. (b)
Waratuke, S. A.; Thorn, M. G.; Fanwick, P. E.; Rothwell, A. P.;
Rothwell, I. P. J . Am. Chem. Soc. 1999, 121, 9111.
(19) Briggs, J . R. J . Chem. Soc., Chem. Commun. 1989, 674.
(20) Emrich, R.; Heinemann, O.; J olly, P. W.; Kru¨ger, C.; Verhovnik,
G. P. J . Organometallics 1997, 16, 1511.
(22) (a) Manyik, R. M.; Walker, W. E.; Wilson, T. P. J . Catal. 1977,
47, 197. (b) Yang, Y.; Kim, H.; Lee, J .; Paik, H.; J ang, H. G. Appl. Catal.
A: Gen. 2000, 193, 29.

Catalytic Trimerization of Ethene
Organometallics, Vol. 21, No. 23, 2002 5127
Sch em e 3
Ta ble 6. Ca ta lytic Eth en e Con ver sion w ith 5/Coca ta lyst System s Com p a r ed w ith 1/MAOa
C6 products,
g (wt %)
C10 products,
g (wt %)
PE,
g (wt %)
productivity
C6 productsb
trimerizn
products, wt %
trimerizn
cocatalyst
productivityc
[PhNMe2H][B(C6F5)4]d
14.6 (90)
5.8 (88)
13.8 (95)
20.9 (83)
1.2 (7)
0.3 (5)
0.6 (4)
3.5 (14)
0.3 (2.0)
0.4 (6)
n.d.
390
155
365
555
97
93
73
29
68
d
B(C6F5)3
MAO/SiO2
99f
97
e
1/MAOg
0.5 (1.8)
110
a
b
Reaction conditions: toluene solvent, 5 bar of ethene, 30 °C, 15 µmol of Ti, 30 min reaction time. In g of C6 product/((mmol of Ti) bar
h). c In mol of olefinic bonds trimerized/((mmol of Ti) h). B:Ti ratio ) 1.1. e 5 wt % MAO on silica, Al:Ti ratio ) 250. f Product distribution
is based on liquid fraction only (remaining 1 wt % is C8 and C12+ fractions). Al:Ti ratio ) 1000.
d
g
determining step the insertion of the third monomer
molecule.23
dominant formation of trimerization products, indicat-
ing that free AlMe3 is not instrumental in the generation
of the species that produces 1-hexene. MAO/SiO2 and
[PhNMe2H][B(C6F5)4] proved to be efficient cocatalysts,
affording active and highly selective (>97 wt %) trim-
erization catalysts. The selectivity of the 5/B(C6F5)3
catalyst system is slightly worse (about 93 wt % of
trimerization products), and its activity is rather low.
This may be associated with the relatively strong
coordinative ability of the [MeB(C6F5)3]- anion.25 It may
be taken into account that the experiments with the
B(C6F5)3 and [PhNMe2H][B(C6F5)4] activated systems
are conducted in the absence of impurity scavenger,
which might partially explain the lower activities,
compared to 1/MAO.
Effect of th e Br id ge betw een th e Cyclop en ta d i-
en yl a n d Ar en e Gr ou p s. To probe structure-perfor-
mance relationships in the (Cp-arene)Ti-based ethene
trimerization catalyst system, the effect of the bridge
between the cyclopentadienyl ligand and the arene
group on catalytic ethene conversion was investigated.
The precatalysts (η5-C5H4CH2Ph)TiCl3 (6), (η5-C5H4-
SiMe2Ph)TiCl3 (7), and (η5-C5H4CMe2CH2Ph)TiCl3 (8)
were tested in ethene conversions with MAO activator
and compared with the reference catalyst (η5-C5H4CMe2-
Ph)TiCl3 (1)/MAO (Table 7).
Zhu and co-workers reported recently that the cata-
lyst system (η5-C5Me5)Ti(OCH2Ph)3/MAO in heptane
produces short chain (ethyl and butyl) branched poly-
ethene in the homopolymerization of ethene, and they
proposed a reduced titanium species as being respon-
sible for in situ dimerization and trimerization of
ethene.24 They suggest that free AlMe3, present in MAO,
is essential in the reduction of titanium(IV) and the
subsequent oligomerization of ethene. Decreasing the
AlMe3 amount in the cocatalyst indeed diminishes the
degree of branching in the PE obtained with this
catalyst system. The presence of ethyl branches (in
addition to butyl branches) in the polyethene generated
by the (η5-C5Me5)Ti(OCH2Ph)3/MAO catalyst suggests
that a different oligomerization mechanism is at work
in this particular system, as neither with the (η5-C5Me5)-
TiMe3/B(C6F5)3/toluene catalyst17 nor with the (Cp-
arene)TiCl3/MAO systems described here have indica-
tions for the formation of significant amounts of 1-butene
been found.
To evaluate the effect of the cocatalyst on the trim-
erization catalysis, we performed catalytic ethene con-
version studies using (η5-C5H4CMe2Ph)TiMe3 (5), with
the cocatalysts [PhNMe2H][B(C6F5)4], B(C6F5)3, and
MAO/SiO2 (Table 6). All catalyst systems show pre-
The catalysts with the CH2 and SiMe2 bridges (6 and
7, respectively) both produce 1-hexene and polyethene
in comparable amounts, together with a Flory-Schulz
(23) (a) Ko¨hn, R. D.; Haufe, M.; Kociok-Ko¨hn, G.; Grimm, S.;
Wasserscheid, P.; Keim, W. Angew. Chem., Int. Ed. 2000, 39, 4337.
(b) Mihan, S.; Lilge, D.; Maas, H.; Molnar, F.; Ko¨hn, R.; Seifert, G.;
Kociok-Ko¨hn, G. Abstracts of Papers, 221st National Meeting of the
American Chemical Society, San Diego, CA, April 2001; American
Chemical Society: Washington, DC, 2001; INOR 114.
(25) For examples, see: (a) Yang, X.; Stern, C. L.; Marks, T. J . J .
Am. Chem. Soc. 1994, 116, 10015. (b) Bochmann, M.; Lancaster, S. J .
Organometallics 1994, 13, 2235. (c) LaPointe, R. E.; Stevens, J . C.;
Nickias, P. N.; Mark, M. H. Eur. Pat. Appl. EP-520732-A, 1992. For
recent reviews, see: (d) Piers, W. E.; Chivers, T. Chem. Soc. Rev. 1997,
26, 345. (e) Chen, E. Y. X.; Marks, T. J . Chem. Rev. 2000, 100, 1391.
(24) Zhu, F.; Huang, Y.; Yang, Y.; Lin, S. J . Polym. Sci. A: Polym.
Chem. 2000, 38, 4258.

5128 Organometallics, Vol. 21, No. 23, 2002
Deckers et al.
Ta ble 7. Ca ta lytic Eth en e Con ver sion w ith th e [η5-C5H4-(B)-P h ]TiCl3/MAO Ca ta lyst System sa
C6 products,
g (wt %)
C8 products,
g (wt %)
C10 products,
g (wt %)
C12-24 products,
g (wt %)
PE,
g (wt %)
trimerizn
products, wt %
catalyst (B)
1 (CMe2)
6 (CH2)
7 (SiMe2)
8 (CMe2CH2)
20.9 (83)
2.7 (42)
2.1 (36)
1.2 (83)
0.3 (1)
0.4 (6)
0.3 (5)
0.1 (7)
3.5 (14)
0.6 (9)
0.4 (7)
0.05 (3)
0.1 (0.5)
0.6 (9)
0.5 (8)
0.5 (2)
2.2 (34)
2.6 (44)
0.1 (7)
97
<51
<43
86
0.01 (0.7)
a
Reaction conditions: toluene solvent, 5 bar of ethene, 30 °C, 15 µmol of Ti, Al:Ti ) 1000, 30 min reaction time.
Ta ble 8. Ca ta lytic Eth en e Con ver sion w ith th e (η5-C5H4CR2P h )TiCl3/MAO Ca ta lyst System sa
C6 products,
g (wt %)
C10 products,
g (wt %)
PE, g
(wt %)
productivity
C6 productsb
trimerizn
products, wt %
trimerizn
catalyst (R2)
productivityc
1 (Me2)
9 (Et2)
10 (-(CH2)5-)
11 (dCH2)
20.9 (83)
18.5 (88)
24.4 (87)
17.3 (88)
3.5 (14)
1.4 (7)
2.9 (10)
1.4 (7)
0.5 (1.8)
1.0 (4.6)
0.6 (2.0)
0.9 (4.7)
555
495
650
460
97
95
97
95
110
92
125
87
a
b
Reaction conditions: toluene solvent, 5 bar of ethene, 30 °C, 15 µmol of Ti, Al:Ti ) 1000, 30 min reaction time. In g of C6 product/
((mmol of Ti) bar h). c In mol of olefinic bonds trimerized/((mmol of Ti) h).
type distribution of higher 1-alkenes. These systems
thus have a very poor selectivity for trimerization. The
catalyst 8 (with a bridge with a C2 backbone) produces
mainly 1-hexene, but at a very slow rate.
involves the coordination of the arene moiety to the
electron-deficient titanium center to generate the ansa-
Cp-arene Ti(II) species (Scheme 3). Apparently, both
the CH2 (6) and SiMe2 (7) bridged compounds are less
effective in this respect than the CMe2-bridged parent
compound (1). It is as yet unclear whether this is the
result of a slow switch from the polymerization catalyst
(Ti(IV)-dialkyl) to the trimerization catalyst (Ti(II)/Ti-
(IV)-metallacycle), promoted by (intramolecular) arene
coordination, or if the trimerization catalyst, once
formed, can readily be transformed back into a species
active in polymerization.
Catalyst precursors 1 and 6 have recently been
compared in the polymerization of styrene.9c The 1/MAO
(CMe2-bridged) system displayed significantly lower
polymerization activity and afforded syndiotactic poly-
styrene with a lower molecular weight and melting
temperature than the CH2-bridged catalyst 6/MAO,
presumably due to stronger arene coordination in the
cationic catalytic species derived from 1. Different
coordinating properties of the phenyl group in cationic
metallocene species with C5H4CR2Ph ligands (R ) H,
Me) were also observed for cationic species generated
from the reaction of (η5-C5H4CR2Ph)2ZrMe2 with 1 equiv
of [Ph3C][B(C6F5)4].26 Only for the cation with R ) Me
did NMR spectroscopy suggest some form of intramo-
lecular coordination of the pendant phenyl group. These
findings indicate that the formation of ansa-cyclopen-
From the above experiments it is clear that a disub-
stituted C1 bridge between the cyclopentadienyl and
arene moieties gives the best selectivity and activity in
catalytic ethene trimerization. To refine this point
further, we prepared three more precatalysts of the type
[η5-C5H4-(B)-Ph]TiCl3 that satisfy this requirement:
9 (B ) CEt2), 10 (B ) C[(CH2)5]), and 11 (B ) CdCH2).
The data in Table 8 show that, upon activation with
MAO, all these catalysts are indeed active and selective
in ethene trimerization. The performance of 10/MAO,
with the bridging carbon forming part of a 1,1-disub-
stituted cyclohexane moiety, even exceeds that of the
reference catalyst 1/MAO, whereas the catalyst with the
CEt2 bridge is slightly inferior to its CMe2 analogue in
both activity and selectivity. Interestingly, the catalyst
11/MAO, with an sp2 bridging carbon, also performs
reasonably well, with trimerization behavior comparable
to that of the catalyst with the CEt2 bridge.
From the data in Table 7, it can be seen that the
bridging unit plays a crucial role in the selective ethene
trimerization performance of these catalyst systems.
The transformation of the catalyst from a species active
in polymerization into one that is active in trimerization,
as put forward in the in the proposed mechanism,

Catalytic Trimerization of Ethene
Organometallics, Vol. 21, No. 23, 2002 5129
Ta ble 9. Ca ta lytic Eth en e Con ver sion w ith th e [η5-(3-R)C5H3CMe2P h ]TiCl3/MAO Ca ta lyst System sa
time,
min
C6 products,
g (wt %)
C10 products,
g (wt %)
PE,
g (wt %)
productivity
C6 productsb
trimerizn
products, wt %
trimerizn
catalyst (R)
productivityc
1 (H)
1 (H)
12 (SiMe3)
13 (CMe3)
14 (CMe2Ph)
14 (CMe2Ph)
30
120
30
30
30
20.9 (83)
27.9 (78)
25.2 (85)
14.7 (89)
11.9 (91)
46.6 (89)
3.5 (14)
6.6 (18)
3.3 (11)
0.9 (5)
0.6 (5)
4.1 (8)
0.5 (1.8)
1.0 (2.8)
0.4 (1.2)
0.7 (4.3)
0.3 (2.3)
0.8 (1.5)
555
185
670
390
315
310
97
96
96
94
96
97
110
38
130
73
59
59
120
a
b
Reaction conditions: toluene solvent, 5 bar of ethene, 30 °C, 15 µmol of Ti, Al:Ti ) 1000. In g of C6 product/((mmol of Ti) bar h). c In
mol of olefinic bonds trimerized/((mmol of Ti) h).
tadienyl-arene Ti(II) cations, the prerequisite for selec-
tive ethene trimerization activity, is likely to be facile
for the CMe2-bridged species, which could explain the
differences in ethene conversion for 1/MAO and 6/MAO.27
The catalyst performance of the SiMe2-bridged species
7 is most likely related to the poor accessibility of the
ansa-Cp-arene coordination mode of the ancillary
ligand. Due to the larger ionic radius of silicon (0.26 Å)
versus carbon (0.15 Å)28 the ligand will have to adopt a
much more acute Cp-Si-arene bend angle to accom-
modate ansa coordination.29 Bochmann and co-workers
reported that the reaction of (η5-C5H4SiMe2Ph)TiMe3
with either B(C6F5)3 or [Ph3C][B(C6F5)4] affords highly
thermally labile species (even at -60 °C), in contrast
to the much more stable complex [(η5:η6-C5H4CMe2Ph)-
TiMe2][B(C6F5)4], tentatively suggesting the absence of
the crucial arene coordination in the former cationic
species.9a These observations may indicate that in
7/MAO the conversion from an ethene polymerization
catalyst to an ethene trimerization catalyst is slow
compared to that for 1/MAO.
for a stronger η6-arene coordination in 8 than for 1 with
the CMe2 bridge.30 The “constrained geometry” of the
C1-bridged ansa-(η5: η6-Cp-arene)Ti species might fa-
cilitate the slippage or dissociation of the arene ring,
proposed for the rate-determining insertion of the third
ethene molecule into the titanacyclopentane, resulting
in a higher trimerization activity of 1/MAO relative to
8/MAO.
The differences in trimerization activity between the
1/MAO, 9/MAO, and 10/MAO systems are relatively
small (maximum 25%) and can be the result of the
different steric properties of the backbone substituents31
or other effects that influence the propensity for arene
“ring slippage” in the proposed rate-determining step.
Since ansa-cyclopentadienyl-arene coordination of the
ancillary ligand is crucial to ethene trimerization reac-
tivity (vide supra), the behavior of the [C5H4C(dCH2)-
Ph] ligand (11) is somewhat puzzling. To adopt the
required conformation, the Cp-C-arene bend angle in
this ligand has to be reduced even more from its
unstrained value (around 120°) than for the nonselective
SiMe2-bridged species 7. One possibility is that the
unsaturated moiety in the bridge of 11 gets alkylated
in the process, which would effectively turn it into a
CRR′-bridged species. Further study of the cationic
dialkyl derivative of 11 would be required to investigate
this issue.
The clean formation of ansa-cyclopentadienyl-arene
cationic titanium species for the reaction of the trim-
ethyl derivative of 8, (η5-C5H4CMe2CH2Ph)TiMe3, with
9a
B(C6F5)3 (and for the reaction of (η5-C5Me4CH2CH2-
Ph)TiMe3 with [Ph3C][B(C6F5)4])9d suggests that the
ansa coordination, crucial for trimerization selectivity,
is readily accessible for bridges with a C2 backbone.
Accordingly, a relatively good trimerization selectivity
of 86% is obtained for 8/MAO, albeit at a slow rate. The
longer, more flexible CMe2CH2 bridge probably allows
Effect of Su bstitu en ts on th e Cyclop en ta d ien yl
Gr ou p . The effect of the attachment of substituents on
the cyclopentadienyl moiety of the cyclopentadienyl-
arene catalysts was probed by the synthesis of three
precatalysts of the type [η5-(3-R)C5H3CMe2Ph]TiCl3: 12
(R ) SiMe3), 13 (R ) CMe3), and 14 (R ) CMe2Ph). The
results of catalytic ethene conversion with these cata-
(26) (a) Bochmann, M.; Green, M. L. H.; Powell, A. K.; Sassmann-
shausen, J .; Triller, M. U.; Wocadlo, S. J . Chem. Soc., Dalton Trans.
1999, 43. (b) Doerrer, L. H.; Green, M. L. H.; Hau¨ssinger, D.;
Sassmannshausen, J . J . Chem. Soc., Dalton Trans. 1999, 2111.
(27) Metal-arene interactions were proposed to explain the obser-
vation that in ethene polymerization with [η5-C5H4-(B)-Ph]2ZrCl2/
MAO systems (B ) CH2, CMe2, SiMe2), the CH2- and SiMe2-bridged
species are about 20 times as active as the CMe2-bridged species. See:
Alt, H. G.; Ko¨ppl, A. Chem. Rev. 2000, 100, 1205.
(30) The X-ray structure of {[(η5:η6-C5H4CMe2-3,5-Me2C6H3)Ti(µ-
Br)]2}[B(C6F5)4]2 indicates that the arene moiety does not coordinate
in a fully symmetrical η6 fashion but that the three carbons closest to
the bridge show shorter Ti-C distances than the other three (ap-
proaching η3 coordination).10
(31) For examples of backbone substituent effects in ansa-metal-
locenes, see: (a) Razavi, A.; Atwood, J . L. Macromol. Symp. 1995, 89,
345. (b) Alt, H. G.; Zenk, R. J . Organomet. Chem. 1996, 526, 295. (c)
Ewen, J . A.; Hapseslagh, L.; Atwood, J . L.; Zhang, H. J . Am. Chem.
Soc. 1987, 109, 6544.
(28) CRC Handbook of Chemistry and Physics, 80th ed.; Lide, D.
R., Ed.; CRC Press: Boca Raton, FL, 1999; pp 12-14.
(29) For example, in the X-ray structures of Me2X(η5-C5H4)ZrCl2 (X
) C, Si) a difference of 6.3° is observed in the respective bend angles.
See: (a) Koch, T.; Blaurock, S.; Somoza, F. B., J r.; Voigt, A.; Kirmse,
R.; Hey-Hawkins, E. Organometallics 2000, 19, 2556. (b) Bajgur, C.
S.; Tikkanen, W. R.; Petersen, J . L. Inorg. Chem. 1985, 24, 2539.

5130 Organometallics, Vol. 21, No. 23, 2002
Deckers et al.
Sch em e 4
lysts using MAO activator, and comparative data for the
reference catalyst 1/MAO, are listed in Table 9.
As expected, all three systems show good selectivity
for trimerization. The SiMe3-substituted catalyst 12/
MAO shows a somewhat improved activity over 1/MAO,
but the CMe3-substituted catalyst 13/MAO exhibits a
lower productivity and selectivity over the 30 min run
period. This appears to be due to a catalyst degradation
process in the latter system. Its initial ethene uptake
rate is comparable to that of 12/MAO but decreases
much more rapidly over time, the catalyst being fully
deactivated after 15 min. The catalyst 14/MAO, which
has two CMe2Ph substituents on the cyclopentadienyl
ring that, in principle, can both (alternately) coordinate
to the metal center, is selective but rather slow com-
pared to 1/MAO. Nevertheless, the catalyst 14/MAO
does have a very interesting feature. As mentioned
before, the thermal stability of 1/MAO is only modest,
and even at 30 °C catalyst degradation is taking place
noticeably. The top two entries of Table 9 describe the
behavior of 1/MAO at run times of 30 and 120 min,
respectively. It is clear that catalyst deactivation limits
the total production attainable by this catalyst to around
48 kg of trimerization product/(g of Ti). In contrast, 14/
MAO displays a comparable productivity (around 310
kg of C6/(mol h bar)) over both the 30 and 120 min runs,
affording a production of 71 kg of trimerization product/
(g of Ti) after 120 min, at which stage the catalyst is
still active. Thus, the presence of two CMe2Ph substit-
uents slows down the catalysis but greatly improves
catalyst stability. The increased stability may be related
to our earlier observation that 1/MAO degrades less
rapidly in neat toluene solvent than in an n-octane/
toluene (80/20 v/v) mixture, possibly indicating catalyst
stabilization by transient toluene coordination. The
lower activity may be caused by competition between
the loose pendant arene substituent and incoming
ethene on the titanacyclopentane intermediate.
the tert-butyl-substituted catalyst 13/MAO these trends
are similar (99.8% and 91%, respectively).
Considering all the possibilities to form C10 products
from ethene and 1-hexene via the proposed trimeriza-
tion mechanism, it appears that one pathway is by far
the most favorable. This is illustrated in Scheme 4 and
proceeds through selective oxidative coupling of ethene
with 1-hexene to give a titanacylopentane with the
n-butyl substituent on the â-carbon. This is followed by
an insertion of ethene into the Ti-C bond at the
unsubstituted side of the metallacycle, giving a â-n-
butyl-substituted titanacycloheptane, which then af-
fords 5-methylnon-1-ene via â-H transfer from the
unsubstituted side of the metallacycle and reductive
elimination.32 The similar product distribution within
the C10 fractions observed for the SiMe3 (12) and CMe3
(13) substituted catalysts suggests that the selectivity
of the cotrimerization is mainly controlled by the steric
properties of the substituted cyclopentadienyl ligand
rather than by its electronic properties.
Com bin in g Liga n d Effects. The investigation of the
effect of single ligand variations on the catalytic ethene
conversion with (cyclopentadienyl-arene)titanium spe-
cies (variation in bridging group, substituents on the
cyclopentadienyl moiety, and substituents on the aryl
moiety) has resulted in the identification of specific
features that appear to be advantageous to catalyst
efficiency. In comparison under standard conditions to
the reference catalyst 1/MAO, the catalyst with the
C[(CH2)5] bridging group (10) shows a 17% increase in
C6 productivity, and the catalyst with the SiMe3-
substituted cyclopentadienyl group (12) shows a 20%
increase. Combining these features in the catalyst {η5-
(3-SiMe3)C5H3C[(CH2)5]Ph}TiCl3 (15)/MAO leads to a
57% increase in C6 productivity relative to 1/MAO
(Table 10). This suggests that the effects of variations
in bridging group and cyclopentadienyl substituent are
roughly additive, with a slightly positive nonlinearity.
In contrast to the positive effect on C6 productivity of
substitution on the cyclopentadienyl group, the attach-
The stabilities of the catalysts 12 (R ) SiMe3)/MAO
and 13 (R ) CMe3)/MAO differ significantly. The actual
deactivation mechanism is not known presently, but
various possible pathways can be suggested. For in-
stance, it is possible that electronic effects play a role
in the stabilization of the catalytic species. As suggested
earlier, (transient) arene coordination improves the
catalyst stability. The electron-donating tert-butyl group
of 13 increases the electron density on the titanium
center, making it less prone to arene coordination and
thus possibly more amenable to catalyst deactivation.
A second possible deactivation pathway involves cyclo-
metalation of the EMe3 substituent (E ) Si, 12; E ) C,
13). Marks and co-workers reported that for cationic
zirconocene species [(η5-C5H4R)(η5-C5H4CMe3)ZrR′]+ this
process occurs more easily for E ) C than for E ) Si.25a
A similar trend in our systems could explain the greater
stability of 12/MAO relative to 13/MAO in selective
ethene trimerization.
The addition of a substituent on the cyclopentadienyl
ring also affects the composition of the trimerization
products formed. For the SiMe3-substituted catalyst 12/
MAO the 1-hexene content of the C6 fraction rises to
99.9% (99.7% for 1/MAO) and the 5-methyl-1-nonene
content of the C10 fraction to 92% (85% for 1/MAO). For
(32) A similar pathway has been proposed for the cotrimerization
of ethene with styrene with [CpRTiMe2]+ cations. See: Pellecchia, C.;
Pappalardo, D.; Oliva, L.; Mazzeo, M.; Gruter, G. J . Macromolecules
2000, 33, 2807.

Catalytic Trimerization of Ethene
Organometallics, Vol. 21, No. 23, 2002 5131
Ta ble 10. Ca ta lytic Eth en e Con ver sion w ith th e [η5-(3-R)C5H4CX2Ar ]TiCl3/MAO Ca ta lyst System sa
C6 products,
g (wt %)
C10 products,
g (wt %)
PE,
g (wt %)
productivity
C6 productsb
trimerizn
products, wt %
trimerizn
catalyst (R, X, Ar)
1 (H, Me, Ph)
productivityc
20.9 (83)
24.4 (87)
33.9 (84)
7.9 (93)
40.1 (84)
24.4 (91)
3.5 (14)
2.9 (10)
5.2 (13)
0.1 (4)
7.0 (15)
1.6 (6)
0.5 (1.8)
0.6 (2.0)
1.2 (3.0)
0.1 (1.3)
0.3 (0.6)
0.7 (2.8)
555
650
905
210
1070
650
97
97
97
97
99
97
110
125
177
38
211
121
10 (H, [CH2]5, Ph)
15 (SiMe3, [CH2]5, Ph)
3 (H, Me, 3,5-Me2C6H3)
16 (SiMe3, Me, 3,5-Me2C6H3)
17 (CMe3, Me, 3,5-Me2C6H3)
a
b
Reaction conditions: toluene solvent, 5 bar of ethene, 30 °C, 15 µmol of Ti, Al:Ti ) 1000, 30 min reaction time. In g of C6 product/
((mmol of Ti) bar h). c In mol of olefinic bonds trimerized/((mmol of Ti) h).
Ta ble 11. Over view of Tr im er iza tion Activity a n d
ment of methyl subtituents on the aryl group was seen
to lead to a decrease in C6 productivity relative to
Selectivity w ith [η5-(3-R)C5H3-(B)-Ar ]TiCl3/MAO
Ca ta lyst System sa
1/MAO (for the catalyst with the 3,5-Me2C6H3 group (3)
a reduction by 62%). Remarkably, the catalyst [η5-(3-
trimerizn
precat-
alyst
products,
wt %
trimerizn
SiMe3)C5H3CMe2-3,5-Me2C6H3]TiCl3 (16)/MAO, com-
bining Cp substitution with aryl substitution, turned
out to be the most active catalyst of all (Table 10), with
an increase in the rate of C6 production by 92% relative
to 1/MAO and a very high selectivity for the combined
trimerization products (99%). Aryl substitution in the
tert-butylcyclopentadienyl derivatives [η5-(3-CMe3)C5H3-
CMe2Ar]TiCl3/MAO (Ar ) Ph (13), 3,5-Me2C6H3 (17))
has no effect on initial ethene uptake rate but signifi-
cantly improves catalyst stability, as can be seen from
the ethene uptake profile, affording a catalyst with
comparable productivity to 1/MAO over a 30 min run
period. A judicious combination of substituents on the
key positions (backbone, cyclopentadienyl, and arene)
can thus improve the three major parameters in the
catalytic trimerization process: activity, selectivity, and
stability. The precise interplay of steric and electronic
effects is as yet unclear and warrants further investiga-
tion.
R
B
Ar
productivityb
1
2
H
H
CMe2
CMe2
Ph
4-Me-
C6H4
97
98
110
75
3
H
CMe2
3,5-Me2-
C6H3
Ph
97
38
6
7
8
H
H
H
H
H
H
CH2
SiMe2
<51
<43
86
95
97
95
96
94
96
Ph
CMe2CH2 Ph
CEt2 Ph
C[(CH2)5] Ph
CdCH2
CMe2
6
92
125
87
130
73
59
9
10
11
12
13
14
15
16
Ph
Ph
Ph
Ph
SiMe3
CMe3
CMe2Ph CMe2
SiMe3
SiMe3
CMe2
C[(CH2)5] Ph
97
99
177
211
CMe2
CMe2
3,5-Me2-
C6H3
3,5-Me2-
C6H3
17
CMe3
97
121
a
Reaction conditions: toluene solvent, 5 bar of ethene, 30 °C,
b
15 µmol of Ti, Al:Ti ) 1000, 30 min reaction time. In mol of
olefinic bonds trimerized/((mmol of Ti) h).
Con clu sion s
trimerization catalysts. Apparently, the conformational
constraints imposed by the bridging unit are of major
importance to catalyst performance. The introduction
of additional substituents on the cyclopentadienyl moi-
ety can influence both catalyst activity and stability. The
introduction of a trimethylsilyl substituent is highly
favorable for catalyst activity, whereas the addition of
a second CMe2Ph substituent greatly increases catalyst
stability (albeit at the cost of a lower activity). Combin-
ing two ligand variations at a time shows in some cases
a significant positive nonlinear additivity of the sub-
stituent effects on catalyst activity. In addition, steric
effects can be employed to improve the selectivity for
formation of a single monomethyl-branched R-olefin
product in the cotrimerization of ethene with 1-alkenes.
Although some ethene oligomerization processes that
produce 1-hexene in excess of the amount expected from
Schulz-Flory product distributions and ethene polym-
erization catalysts that produce branched polyethenes
are known (e.g., for certain Ziegler-type catalyst sys-
We have identified a catalyst system, (η5-C5H4CMe2R)-
TiCl3/MAO, that by attachment of a pendant arene
group to the cyclopentadienyl ancillary ligand (R ) aryl)
can be transformed from an ethene polymerization
catalyst into a selective ethene trimerization catalyst.
The intrinsic selectivity for ethene trimerization of the
catalytic species is essentially independent of solvent
and cocatalyst, and hemilabile behavior of the cyclo-
pentadienyl-arene ancillary ligand appears to be es-
sential for obtaining this high selectivity for trimeriza-
tion.
We have shown that a wide range of titanium pre-
catalysts of the type [η5-C5H3R-(B)-Ar]TiCl3 can be
activated with MAO to effect catalytic ethene trimer-
ization (summarized in Table 11). The nature of the
bridging moiety (B) between the cyclopentadienyl and
the arene group is crucial for obtaining a good selectivity
in 1-hexene production. Thus far, only disubstituted C1
bridges have afforded highly selective and active ethene

5132 Organometallics, Vol. 21, No. 23, 2002
Deckers et al.
tems),33 to our knowledge the present system is the first
highly active ethene trimerization catalyst that is not
based on chromium. Very recently catalyst systems for
ethene trimerization based on TaMe2Cl3 and (arene)2-
VX (X ) Br, Cl, BPh4) were reported that show good
selectivity for 1-hexene, but their activities are at least
2 orders of magnitude smaller than those of the tita-
nium catalyst systems described here. The catalyst
precursors, substituted mono(cyclopentadienyl)titanium
trichloride complexes, are readily synthesized, and
ligand variations can be easily introduced. They may
provide a well-defined, highly active alternative to the
currently available chromium-based ethene trimeriza-
tion catalysts.
instrument equipped with an HP-1 dimethylpolysiloxane
column (19095 Z-123). GC/MS analyses were conducted using
an HP 5973 mass-selective detector attached to an HP 6890
GC instrument. DSC analyses were performed on a Perkin-
Elmer DSC-7 differential scanning calorimeter. Elemental
analyses were performed at the Microanalytical Department
of the University of Groningen. Given values are the average
of at least two independent determinations.
P r ep a r a tion of (η5-C5H4CMe2-4-MeC6H4)TiCl3 (2). To a
solution of 2.60 g (9.6 mmol) of C5H4(SiMe3)CMe2-4-MeC6H4
in 40 mL of methylene chloride at -30 °C was added dropwise
1.06 mL (1.8 g, 10 mmol) of titanium tetrachloride. The brown-
red suspension was warmed to room temperature and was
stirred overnight. The solvent was removed in vacuo, and the
residue was stirred with 20 mL of pentane, which was
subsequently pumped off. Extraction with methylene chloride
and concentration and cooling of the extract to -40 °C afforded
1
Exp er im en ta l Section
2.32 g (6.6 mmol, 69%) of 2 as brown crystals. H NMR (300
3
3
MHz, C6D6): δ 6.88 (d, J HH ) 8.1, 2H, Ar H), 6.80 (d, J HH
)
Ma ter ia ls a n d Meth od s. All experiments were carried out
under a purified nitrogen atmosphere using standard Schlenk
and glovebox techniques. Deuterated solvents (Aldrich, Acros)
were either used as received (CDCl3) or dried over Na/K alloy
and vacuum-transferred before use (C6D6, THF-d8). Cyclooc-
tane (Aldrich, used as internal standard in the catalytic ethene
conversion experiments) and n-octane (Aldrich) were distilled
from sodium, and 1-octene (Aldrich) was distilled from CaH2
prior to use. Toluene (Aldrich, anhydrous, 99.8%) was passed
over columns of Al2O3 (Fluka), BASF R3-11 supported Cu
oxygen scavenger, and molecular sieves (Aldrich, 4 Å) under
a nitrogen atmosphere prior to use. Diethyl ether and THF
(Aldrich, anhydrous) were dried over Al2O3 (Fluka), and
pentane, hexane, and methylene chloride (Aldrich, anhydrous)
were dried over molecular sieves (Aldrich, 4 Å) under a
nitrogen atmosphere before use. Ethene (AGA polymer grade)
was passed over BASF R3-11 supported Cu oxygen scavenger
and molecular sieves (Aldrich, 4 Å).
3
8.4, 2H, Ar H), 6.31 (ps t, J HH ) 2.4, 2H, Cp H), 6.00 (ps t,
3J HH ) 2.4, 2H, Cp H), 2.07 (s, 3H, ArCH3), 1.56 (s, 6H,
C(CH3)2). 13C NMR (75.4 MHz, C6D6): δ 154.6 (Ar Cipso), 145.2
(Cp Cipso), 136.1 (Ar p-Cipso), 129.3, 126.1 (Ar o- and m-CH),
123.3, 121.7 (Cp CH), 40.6 (C(CH3)2), 28.8 (C(CH3)2), 20.7
(ArCH3). Anal. Calcd for C15H17TiCl3: C, 51.25; H, 4.87; Ti,
13.62. Found: C, 51.49; H, 4.89; Ti, 13.54.
P r ep a r a tion of C5H4(SiMe3)CEt2P h . To a solution of 4.85
g (58 mmol) of PhLi in 200 mL of diethyl ether, cooled to -50
°C, was added dropwise 8.0 g (60 mmol) of 6,6-diethylfulvene.
The reaction mixture was warmed to room temperature and
was stirred for 3 h. The yellow solution was then cooled with
an ice bath, and 7.6 mL (6.5 g, 60 mmol) of trimethylsilyl
chloride was added dropwise. The mixture was warmed to
room temperature and was stirred overnight. The reaction
mixture was poured into 250 mL of ice-water. The water layer
was extracted with 2 × 100 mL of light petroleum, after which
the combined organic layers were rinsed with 200 mL of brine.
The organic phase was dried on MgSO4. After the low-boiling
volatiles were evaporated in vacuo, the residue was distilled
using a Kugelrohr apparatus. The product was distilled at 110
°C at 0.5 Torr as a mixture of isomers. Yield: 9.21 g (32 mmol,
55%). 1H NMR (300 MHz, CDCl3, main isomer): δ 7.28 (m,
4H, Ph o- and m-H), 7.18 (m, 1H, Ph p-H), 6.40 (m, 1H, Cp
H), 6.31 (s, 1H, Cp H), 6.22 (m, 1H, Cp H), 3.27 (s, 1H, Cp H),
2.02 (m, 4H, CH2CH3), 0.72 (m, 6H, CH2CH3), 0.06 (s, 9H, Si-
(CH3)3).
The compounds C5H4(SiMe3)CMe2-4-MeC6H4,34 [C5H4C(d
CH2)Ph]Li,14 3-(R,R-dimethylbenzyl)-6,6-dimethylfulvene,35 and
36
B(C6F5)3 were prepared according to published procedures.
6,6-Diethylfulvene was prepared from cyclopentadiene and
3-pentanone analogously to the procedure described for 6,6-
pentamethylenefulvene,37 and 3-tert-butyl-6,6-dimethylfulvene
was prepared from tert-butylcyclopentadiene and acetone
analogously to 3-(R,R-dimethylbenzyl)-6,6-dimethylfulvene.35
The titanium complexes (η5-C5H4CMe2Ph)TiCl3 (1),9a (η5-C5H4-
CMe2-3,5-Me2C6H3)TiCl3 (3),10 (η5-C5H4CMe3)TiCl3 (4),38 (η5-
C5H4CMe2Ph)TiMe3 (5),9a (η5-C5H4CH2Ph)TiCl3 (6),9c (η5-C5H4-
SiMe2Ph)TiCl3 (7),9a and (η5-C5H4CMe2CH2Ph)TiCl3 (8)9a were
prepared as reported previously. A toluene solution of MAO
(9.8 wt % Al, Akzo Nobel Chemicals), MAO supported on silica
(5 wt %, Witco), and [PhNMe2H][B(C6F5)4] (Akzo Nobel Chemi-
cals) were used as received.
P r ep a r a tion of (η5-C5H4CEt2P h )TiCl3 (9). To a solution
of 6.30 g (22 mmol) of C5H4(SiMe3)CEt2Ph in 40 mL of
methylene chloride, cooled to -40 °C, was added 2.45 mL (4.2
g, 22 mmol) of TiCl4. The mixture was warmed to room
temperature and was stirred overnight. The methylene chlo-
ride was removed in vacuo, and the residue was stirred with
50 mL of pentane, which was subsequently pumped off.
Extracting with methylene chloride and cooling to -60 °C
afforded red-brown crystals of the title compound. Yield: 5.63
g (15.3 mmol, 70%). 1H NMR (300 MHz, C6D6): δ 7.24 (d, 3J HH
NMR spectra were recorded on Varian Gemini 200/300
spectrometers in NMR tubes equipped with a Teflon (Young)
1
valve. The H NMR spectra were referenced to resonances of
residual protons in the deuterated solvents (δ 7.15 ppm for
C6D6, δ 7.24 ppm for CDCl3). The 13C NMR spectra were
referenced to the carbon resonances of the deuterated solvent
(δ 128 ppm for C6D6). Chemical shifts (δ) are given relative to
tetramethylsilane (downfield shifts are positive); J values are
given in hertz. GC analyses were performed on an HP 6890
3
) 7.3, 2H, Ph o-H), 7.17 (t, J HH ) 7.3, 2H, Ph m-H), 7.06 (t,
3
3J HH ) 7.3, 1H, Ph p-H), 6.26 (ps t, J HH ) 2.8, 2H, Cp H),
3
6.04 (ps t, J HH ) 2.8, 2H, Cp H), 2.06 (m, 2H, CH2CH3), 1.86
3
(m, 2H, CH2CH3), 0.51 (t, J HH ) 7.3, 6H, CH2CH3). 13C NMR
(75.4 MHz, C6D6): δ 154.8, 142.1 (Ph and Cp Cipso), 128.8 (Ph
o-CH), 128.3 (Ph m-CH, overlap with solvent), 127.2 (Ph p-CH),
123.1, 121.8 (Cp CH), 48.6 (C(CH2CH3)2), 29.3 (C(CH2CH3)2),
8.5 (C(CH2CH3)2). Anal. Calcd for C16H19TiCl3: C, 52.57; H,
5.24; Ti, 13.10. Found: C, 52.75; H, 5.27; Ti, 12.99.
P r ep a r a tion of C5H4(SiMe3)C[(CH2)5]P h . To a solution
of 4.00 g (48 mmol) of PhLi in 200 mL of diethyl ether, cooled
to -50 °C, was added dropwise 6.95 g (48 mmol) of 6,6-
pentamethylenefulvene. The reaction mixture was warmed to
room temperature and was stirred for 3 h. Subsequently the
yellow solution was cooled with an ice bath and 6.4 mL (5.5 g,
(33) For example, see: Murtuza, S.; Harkins, S. B.; Long, G. S.; Sen,
A. J . Am. Chem. Soc. 2000, 122, 1867 and references therein.
(34) Sassmannshausen, J . Organometallics 2000, 19, 482.
(35) Gutmann, S.; Burger, P.; Hund, H. U.; Hofmann, J .; Brintz-
inger, H. H. J . Organomet. Chem. 1989, 369, 343.
(36) Massey, A. G.; Park, A. J . J . Organomet. Chem. 1964, 2, 245.
(b) Massey, A. G.; Park, A. J . J . Organomet. Chem. 1966, 5, 218.
(37) Stone, K. J .; Little, R. D. J . Org. Chem. 1984, 49, 1849.
(38) Hart, S. L.; Duncalf, D. J .; Hastings, J . J .; McCamley, A.; Taylor,
P. C. J . Chem. Soc. Dalton Trans. 1996, 2843.

Catalytic Trimerization of Ethene
Organometallics, Vol. 21, No. 23, 2002 5133
51 mmol) of trimethylsilyl chloride was added dropwise. The
mixture was warmed to room temperature and was stirred
overnight. The reaction mixture was poured into 250 mL of
ice-water. The water layer was extracted with 2 × 100 mL of
light petroleum, after which the combined organic layers were
rinsed with 200 mL of brine. The organic phase was dried on
MgSO4. After the low-boiling volatiles were evaporated in
vacuo, the residue was distilled using a Kugelrohr apparatus.
The product was distilled at 165 °C at 0.4 Torr as a mixture
P r ep a r a tion of [η5-(3-SiMe3)C5H3CMe2P h ]TiCl3 (12). To
a solution of 0.92 mL (1.6 g, 8.4 mmol) of TiCl4 in 50 mL of
methylene chloride, cooled to -50 °C, was added 2.75 g (8.4
mmol) of C5H3(SiMe3)2CMe2Ph. The reaction mixture was
warmed to room temperature and was stirred overnight. The
volatiles were removed in vacuo, and the residue was stirred
with 15 mL of pentane, which was subsequently pumped off.
Extracting with methylene chloride and cooling to -60 °C
afforded 2.76 g (6.7 mmol, 80%) of the title compound. 1H NMR
(300 MHz, C6D6): δ 7.1-6.85 (m, 5H + 1H, Ph H and Cp H),
6.57 (m, 1H, Cp H), 6.53 (m, 1H, Cp H), 1.63 (s, 6H, C(CH3)2),
0.12 (s, 9H, Si(CH3)3). 13C NMR (75.4 MHz, C6D6): δ 158.5,
148.5, 144.1 (Ph and Cp Cipso), 128.7, 128.6, 126.7, 126.1, 124.6
(Ph CH and Cp CH), 41.2 (C(CH3)2), 29.3, 29.0 (C(CH3)2), -0.8
(Si(CH3)3). Anal. Calcd for C17H23SiTiCl3: C, 49.84; H, 5.66;
Ti, 11.69. Found: C, 49.70; H, 5.68; Ti, 11.59.
1
of isomers. Yield: 8.96 g (30 mmol, 63%). H NMR (300 MHz,
CDCl3, main isomer): δ 7.40 (m, 2H, Ph o-H), 7.33 (m, 2H, Ph
m-H), 7.15 (m, 1H, Ph p-H), 6.43 (m, 2H, Cp H), 6.15 (s, 1H,
Cp H), 3.27 (s, 1H, Cp H), 2.17 (m, 4H, R-CH2), 1.65-1.40 (m,
6H, â- and γ-CH2), -0.03 (s, 9H, Si(CH3)3).
P r ep a r a tion of {η5-C5H4C[(CH2)5]P h }TiCl3 (10). Tita-
nium tetrachloride (1.4 mL, 2.4 g, 12.7 mmol) was added to a
solution of 3.70 g (12.5 mmol) of C5H4(SiMe3)C[(CH2)5]Ph in
40 mL of methylene chloride, cooled to -40 °C. The reaction
mixture was warmed to room temperature and was stirred
overnight. The methylene chloride was removed in vacuo, and
the residue was stirred with 30 mL of pentane, which was
subsequently pumped off. The residue was extracted with
methylene chloride. Crystallization from a 1:1 (v/v) mixture
of CH2Cl2 and pentane afforded red-brown crystals of the title
compound in 78% yield (3.68 g, 9.7 mmol). 1H NMR (300 MHz,
C6D6): δ 7.16-7.06 (m, 4H, Ph o- and m-H), 7.01 (m, 1H, Ph
P r ep a r a tion of [C5H3(CMe3)CMe2P h ]Li. To a solution of
3.80 g (45 mmol) of phenyllithium in 60 mL of diethyl ether,
cooled to -40 °C, was added 7.4 g (46 mmol) of 3-tert-butyl-
6,6-dimethylfulvene. The reaction mixture was warmed to
room temperature and was stirred overnight. The volatiles
were removed in vacuo. The yellow oil was suspended in 40
mL of n-hexane and was refluxed for 4 h. The resulting off-
white solid was filtered off and repeatedly rinsed with pentane
to afford 5.50 g (22 mmol, 49%) of product. 1H NMR (300 MHz,
C6D6/THF-d8): δ 7.53 (d, 3J HH ) 7.0, 2H, Ph o-H), 7.13 (t, 3J HH
) 7.3, 2H, Ph m-H), 6.98 (m, 1H, Ph p-H), 5.85 (m, 2H, Cp H),
5.79 (m, 1H, Cp H), 1.81 (s, 6H, C(CH3)2), 1.45 (s, 9H, C(CH3)3).
P r ep a r a tion of [η5-(3-CMe3)C5H3CMe2P h ]TiCl3 (13). To
a solution of 1.47 g (6.0 mmol) of [C5H4(CMe3)CMe2Ph]Li in
30 mL of methylene chloride, cooled to -20 °C, was added
dropwise 0.70 mL (1.2 g, 6.3 mmol) of TiCl4. The red-brown
solution was warmed to room temperature and was stirred
overnight. The solvent was removed in vacuo, and the residue
was stirred with 20 mL of pentane, which was subsequently
pumped off. Extraction with toluene afforded a brown oil that
could not be crystallized from pentane, hexane, toluene, or
methylene chloride. The oil was repeatedly rinsed with cold
pentane to give 1.98 g (5.0 mmol, 83%) of product (about 95%
purity as indicated by NMR spectroscopy). 1H NMR (300 MHz,
3
3
p-H), 6.31 (ps t, J HH ) 2.8, 2H, Cp H), 5.97 (ps t, J HH ) 2.8,
2
2H, Cp H), 2.45 (d, J HH ) 13.2, 2H, R-CHeq), 1.88 (m, 2H,
R-CHax), 1.37 (br, 3H, â- and γ-CH2), 1.25-1.05 (m, 3H, â- and
γ-CH2). 13C NMR (75.4 MHz, C6D6): δ 156.0, 142.1 (Ph and
Cp Cipso), 129.2 (Ph o-CH), 127.9 (Ph m-CH), 126.8 (Ph p-CH),
123.2, 120.9 (Cp CH), 45.1 (C[(CH2)5]), 35.8 (R-CH2), 26.1 (γ-
CH2), 22.4 (â-CH2). Anal. Calcd for C17H19TiCl3: C, 54.08; H,
5.07; Ti, 12.69. Found: C, 53.93; H, 4.90; Ti, 12.62.
P r ep a r a tion of [η5-C5H4C(dCH2)P h ]TiCl3 (11). To a
solution of 0.61 mL (1.06 g, 5.6 mmol) of titanium tetrachloride
in 40 mL of methylene chloride, cooled to -50 °C, was added
1.80 g (5.6 mmol) of [C5H4C(dCH2)Ph]Li. The reaction mixture
was warmed to room temperature and was stirred overnight.
The volatiles were removed in vacuo, and the green-black
residue was stirred with 30 mL of pentane, which was
subsequently pumped off. Extraction with pentane afforded
small amounts of the analytically pure title compound as dark
3
C6D6): δ 7.1-7.0 (m, 3H, Ph m- and p-H), 6.87 (d, J HH ) 7.0,
3
3
2H, Ph o-H), 6.60 (ps t, J HH ) 2.4, 1H, Cp H), 6.40 (ps t, J HH
3
) 3.3, 1H, Cp H), 6.29 (ps t, J HH ) 2.9, 1H, Cp H), 1.64 (s,
3H, C(CH3)2), 1.63 (s, 3H, C(CH3)2), 1.04 (s, 9H, C(CH3)3). 13C
NMR (75.4 MHz, C6D6): δ 157.5, 156.1, 148.8 (Ph and Cp Cipso),
128.6, 126.6, 126.1 (Ph CH), 120.5, 120.0, 119.6 (Cp CH), 41.6
(C(CH3)2), 34.7 (C(CH3)3), 30.4 (C(CH3)3), 28.9, 28.7 (C(CH3)2).
P r ep a r a tion of [C5H3-1,3-(CMe2P h )2]Li. To a suspension
of 2.28 g (27.1 mmol) of PhLi in 50 mL of n-hexane was added
6.14 g (27.4 mmol) of 3-(R,R-dimethylbenzyl)-6,6-dimethylful-
vene. The mixture was refluxed for 5 h. The precipitate was
poured onto a glass frit and rinsed with 2 × 20 mL of pentane.
Drying in vacuo yielded 4.18 g (13.6 mmol, 50%) of an off-
white solid. 1H NMR (300 MHz, C6D6/THF-d8): δ 7.55 (d, 3J HH
) 8.2, 4H, Ph o-H), 7.16 (m, 4H, Ph m-H), 7.01 (m, 2H, Ph
p-H), 5.87 (m, 1H, Cp H), 5.83 (m, 2H, Cp H), 1.79 (s, 12H,
C(CH3)2). 13C NMR (75.4 MHz, C6D6/THF-d8): δ 154.9, 129.0
(Ph and Cp Cipso), 127.8, 126.7, 124.7 (Ph CH), 100.8, 99.8 (Cp
CH), 39.8 (C(CH3)2), 32.5 (C(CH3)2).
1
red crystals. Isolated yield: 0.25 g (0.8 mmol, 14%). H NMR
3
(300 MHz, C6D6): δ 7.2-7.05 (m, 5H, Ph H), 6.35 (ps t, J HH
3
) 2.7, 2H, Cp H), 6.01 (ps t, J HH ) 2.7, 2H, Cp H), 5.58 (s,
1H, dCH2), 5.20 (s, 1H, dCH2). 13C NMR (75.4 MHz, C6D6): δ
142.5, 139.7, 139.6 (Ph, Cp Cipso and CdCH2), 128.8, 128.7,
128.5 (Ph CH), 123.4, 121.1, 120.5 (Cp CH and dCH2). Anal.
Calcd for C13H11TiCl3: C, 48.57; H, 3.45; Ti, 14.90. Found: C,
48.71; H, 3.55; Ti, 14.78.
P r ep a r a t ion of C5H3(SiMe3)2CMe2P h . To a solution of
2.25 g (11.8 mmol) of [C5H4CMe2Ph]Li in 50 mL of diethyl
ether and 20 mL of THF, cooled in ice-water, was added
dropwise 1.5 mL (1.3 g, 11.9 mmol) of Me3SiCl. The mixture
was warmed to room temperature and was stirred overnight.
The yellow solution was cooled in ice-water, and 4.8 mL (12
mmol) of a 2.5 M n-BuLi solution in hexanes was added. After
it was warmed to room temperature, the mixture was stirred
for 4 h. The white suspension was cooled in ice-water, and
1.6 mL (1.4 g, 12.7 mmol) of Me3SiCl was added dropwise. The
mixture was warmed to room temperature and stirred over-
night. The yellow suspension was poured into 125 mL of ice-
water. The water layer was extracted with 50 mL of light
petroleum, and the combined organic layers were dried on
MgSO4. After evaporation of low-boiling volatiles, the residue
was distilled using a Kugelrohr apparatus. The product was
distilled at 115 °C at 0.8 Torr. Yield: 2.87 g (8.7 mmol, 74%).
1H NMR (200 MHz, CDCl3): δ 7.4-7.1 (m, 5H, Ph H), 6.40 (d,
P r ep a r a tion of [η5-C5H3-1,3-(CMe2P h )2]TiCl3 (14). To a
solution of 1.31 g (4.2 mmol) of [C5H3-1,3-(CMe2Ph)2]Li in 30
mL of methylene chloride, cooled to -40 °C, was added
dropwise 0.47 mL (0.8 g, 4.2 mmol) of TiCl4. The dark brown
solution was warmed to room temperature and was stirred
overnight. The solvent was removed in vacuo, and the residue
was stirred with 40 mL of pentane, which was subsequently
pumped off. The residue was extracted with 50 mL of toluene,
which was replaced by a 1:1 (v/v) mixture of methylene chloride
and pentane (30 mL in total). Cooling to -40 °C afforded 0.22
3
1
3J HH ) 2.2, 2H, Cp H), 6.20 (t, J HH ) 2.1, 1H, Cp H), 1.53 (s,
g (0.5 mmol, 12%) of the title compound. H NMR (300 MHz,
6H, C(CH3)2), -0.03 (s, 18H, Si(CH3)3).
C6D6): δ 6.98 (m, 2H, Ph p-H), 6.96 (m, 4H, Ph m- or o-H),

5134 Organometallics, Vol. 21, No. 23, 2002
Deckers et al.
3
6.70 (m, 4H, Ph m- or o-H), 6.50 (m, 1H, Cp H), 6.40 (d, J HH
) 2.6, 2H, Cp H), 1.60 (s, 6H, C(CH3)2), 1.54 (s, 6H, C(CH3)2).
13C NMR (75.4 MHz, C6D6): δ 156.2, 148.8 (Ph and Cp Cipso),
128.4, 126.5, 126.0 (Ph CH), 121.5, 120.5 (Cp CH), 41.7
(C(CH3)2), 28.5, 28.4 (C(CH3)2). Anal. Calcd for C23H25TiCl3:
C, 60.62; H, 5.53. Found: C, 60.16; H, 5.56.
mmol, 72%) of light brown crystals. 1H NMR (300 MHz,
C6D6): δ 6.96 (m, 1H, Cp H), 6.69 (s, 2H, Ar o-H), 6.64 (m,
2H, Cp H and Ar p-H), 6.55 (m, 1H, Cp H), 2.08 (s, 6H, ArCH3),
1.70 (s, 6H, C(CH3)2), 0.13 (s, 9H, Si(CH3)3). 13C NMR (75.4
MHz, C6D6): δ 159.1, 148.5, 144.1 (Ar and Cp Cipso), 137.9
(Ar m-Cipso), 128.8, 128.4, 127.8, 124.7, 124.1 (Ar CH and Cp
CH), 41.2 (C(CH3)2), 29.3, 29.2 (C(CH3)2), 21.5 (ArCH3), -0.9
(Si(CH3)3). Anal. Calcd for C19H27SiTiCl3: C, 52.13; H, 6.22;
Ti, 10.94. Found: C, 51.72; H, 6.24; Ti, 10.76.
P r ep a r a tion of [C5H4(CMe3)CMe2-3,5-Me2C6H3]Li. A
solution of 1.88 g (17 mmol) of (3,5-dimethylphenyl)lithium
in 40 mL of diethyl ether was cooled to -40 °C. An equimolar
amount of 3-tert-butyl-6,6-dimethylfulvene (2.7 g) was added.
The reaction mixture was warmed to room temperature and
was stirred overnight. Removing the solvent in vacuo gave a
yellow oil which solidified in refluxing hexane. The solid was
repeatedly rinsed with pentane to yield 2.35 g (8.6 mmol, 51%)
of the title compound. 1H NMR (300 MHz, C6D6/THF-d8): δ
7.17 (s, 2H, Ar o-H), 6.63 (s, 1H, Ar p-H), 5.81, 5.77, 5.71 (m,
1H each, Cp H), 2.13 (s, 6H, ArCH3), 1.77 (s, 6H, C(CH3)2),
1.40 (s, 9H, C(CH3)3).
P r ep a r a tion of C5H3(SiMe3)2C[(CH2)5]P h . To a solution
of 1.67 g (7.3 mmol) of {C5H4C[(CH2)5]Ph}Li in 70 mL of diethyl
ether, cooled with ice-water, was added dropwise 0.8 mL (0.7
g, 6.4 mmol) of trimethylsilyl chloride. The reaction mixture
was warmed to room temperature and was stirred overnight.
The white suspension was cooled to -30 °C, and 7.3 mmol of
a 2.5 M solution of n-BuLi in hexanes was added dropwise.
After the mixture was stirred for 3 h at ambient temperature,
the reaction vessel was placed in ice-water and 0.9 mL (0.8
g, 7.4 mmol) of trimethylsilyl chloride was added. The reaction
mixture was warmed to room temperature and was stirred
overnight. The mixture was poured into 100 mL of ice-water.
The water layer was extracted twice with 50 mL portions of
light petroleum, and the combined organic layers were dried
over MgSO4. Kugelrohr distillation at 160 °C and 0.4 Torr
yielded 1.28 g (3.5 mmol, 55%) of the title compound. 1H NMR
(200 MHz, CDCl3): δ 7.45-7.1 (m, 5H, Ph H), 6.50 (m, 1H,
Cp H), 6.39 (m, 1H, Cp H), 6.18 (m, 1H, Cp H), 2.2 (m, 4H,
R-CH2), 1.55 (m, 6H, â- and γ-CH2), -0.07 (s, 18H, Si(CH3)3).
P r epar ation of {η5-(3-SiMe3)C5H3C[(CH2)5]P h }TiCl3 (15).
To a solution of 0.34 mL (0.6 g, 3.2 mmol) of titanium
tetrachloride in 40 mL of methylene chloride, cooled to
-40 °C, was added dropwise 1.20 g (3.3 mmol) of C5H3-
(SiMe3)2C[(CH2)5]Ph. The reaction mixture was warmed to
room temperature and was stirred overnight. The volatiles
were removed in vacuo, and the residue was stripped with
pentane. Extraction with methylene chloride yielded red-brown
crystals. Yield: 0.68 g (1.5 mmol, 47%). 1H NMR (300 MHz,
C6D6): δ 7.13 (m, 4H, Ph o- and m-H), 7.01 (m, 1H, Cp H),
6.93 (m, 1H, Ph p-H), 6.45 (m, 2H, Cp H), 2.54 (m, 2H, R-CHeq),
2.07, 1.86 (m, 1H each, R-CHax), 1.4 (br, 3H, â- and γ-CH2),
1.15 (br, 3H, â- and γ-CH2), 0.13 (s, 9H, Si(CH3)3). 13C NMR
(75.4 MHz, C6D6): δ 160.3, 143.6, 142.2 (Ph and Cp Cipso),
129.1, 128.7, 126.8, 126.7, 123.7 (Ph CH and Cp CH), 45.4
(C[(CH2)5]), 36.6, 35.7 (R-CH2), 26.1 (γ-CH2), 22.4 (â-CH2), -0.8
(Si(CH3)3). Anal. Calcd for C20H27SiTiCl3: C, 53.41; H, 6.05;
Ti, 10.65. Found: C, 52.83; H, 6.08; Ti, 10.52.
P r ep a r a tion of C5H3(SiMe3)2CMe2-3,5-Me2C6H3. To a
solution of 1.15 g (5.3 mmol) of [C5H4CMe2-3,5-Me2C6H3]Li in
50 mL of diethyl ether, cooled with ice-water, was added
dropwise 0.7 mL (0.6 g, 5.5 mmol) of trimethylsilyl chloride.
The reaction mixture was warmed to room temperature and
was stirred overnight. The white suspension was cooled to -30
°C, and 5.4 mmol of a 2.5 M solution of n-BuLi in hexanes
was added dropwise. After the mixture was stirred for 3 h at
ambient temperature, the reaction vessel was placed in ice-
water and 0.8 mL (0.7 g, 6.4 mmol) of trimethylsilyl chloride
was added. The reaction mixture was warmed to room tem-
perature and was stirred overnight. The mixture was poured
into 100 mL of ice-water. The water layer was extracted twice
with 50 mL portions of light petroleum, and the combined
organic layers were dried over MgSO4. Kugelrohr distillation
at 160 °C and 0.4 Torr yielded 1.26 g (3.5 mmol, 66%) of the
title compound. 1H NMR (200 MHz, CDCl3): δ 6.90 (s, 2H, Ar
o-H), 6.78 (s, 1H, Ar p-H), 6.37 (m, 2H, Cp H), 6.19 (m, 1H, Cp
H), 2.24 (s, 6H, ArCH3), 1.51 (s, 6H, C(CH3)2), -0.05 (s, 18H,
Si(CH3)3).
P r ep a r a t ion of [η5-(3-CMe3)C5H 3CMe2-3,5-Me2C6H 3]-
TiCl3 (17). To a solution of 1.54 g (5.6 mmol) of [C5H4(CMe3)-
CMe2-3,5-Me2C6H3]Li in 30 mL of methylene chloride, cooled
to -20 °C, was added dropwise 0.65 mL (1.1 g, 5.8 mmol) of
titanium tetrachloride. The mixture was warmed to room
temperature and was stirred overnight. Removing the vola-
tiles, stripping with pentane, and extracting with toluene
afforded a brown oil. Repeated rinsing with cold pentane
afforded 2.06 g (4.9 mmol, 88%) of the oil, which was 95% pure,
as seen by NMR spectroscopy. 1H NMR (300 MHz, C6D6): δ
6.7-6.5 (m, 2H, Ar p-H and Cp H), 6.63 (s, 2H, Ar o-H), 6.43
3
3
(t, J HH ) 2.9, 1H, Cp H), 6.29 (t, J HH ) 2.9, 1H, Cp H), 2.09
(s, 6H, ArCH3), 1.70 (s, 3H, C(CH3)2), 1.69 (s, 3H, C(CH3)2),
1.05 (s, 9H, C(CH3)3). 13C NMR (75.4 MHz, C6D6): δ 157.3,
156.5, 149.0 (Ar and Cp Cipso), 137.8 (Ar m-Cipso), 124.6 (Ar
p-CH), 124.1 (Ar o-CH), 120.6, 120.1, 119.5 (Cp CH), 41.7
(C(CH3)2), 34.7 (C(CH3)3), 30.4 (C(CH3)3), 29.0, 28.9 (C(CH3)2),
21.5 (ArCH3).
Gen er a l P r oced u r e for th e Ca ta lytic Eth en e Con ver -
sion s w ith MAO a s Coca ta lyst. A stainless steel 1 L
autoclave (Medimex), fully temperature- and pressure-con-
trolled and equipped with solvent and catalyst injection
systems, was preheated in vacuo for 45 min at 100 °C prior to
use. The reactor was cooled to the desired temperature,
charged with 200 mL of toluene, and pressurized with ethene.
After equilibration for 15 min, the appropriate amount of
MAO/toluene was injected, together with 25 mL of toluene.
Subsequently, a mixture of 2.50 g of cyclooctane (internal
standard) and 1.0 mL of a 15 mM stock solution of the titanium
halide complex in toluene was injected, together with 25 mL
of toluene, to start the reaction. During the reaction the ethene
pressure was kept constant to within 0.1 bar of the initial
pressure by replenishing flow. The run was ended by adding
an aliquot of ethanol, and the reactor was vented. The
remaining residual MAO was destroyed by adding further
ethanol, and samples of the reaction mixture were taken to
analyze and quantify the soluble components by GC and GC/
MS. The polymer was stirred for 1 h in acidified ethanol,
repeatedly rinsed with ethanol on a glass frit, and dried in
vacuo at 70 °C overnight. For entry 2 in Table 1, the volatiles
of the liquid fraction of the reaction mixture were removed on
a rotary evaporator operating at 80 °C and 180 Torr to give
3.8 g of a fraction, predominantly containing the C10 isomers
(1H NMR spectroscopy showed that the sample contained
about 10 mol % toluene and 10 mol % cyclooctane).
P r ep a r a tion of [η5-(3-SiMe3)C5H3CMe2-3,5-Me2C6H3]-
TiCl3 (16). To a solution of 0.35 mL (0.6 g, 3.2 mmol) of
titanium tetrachloride in 40 mL of methylene chloride,
cooled to -40 °C, was added dropwise 1.18 g (3.3 mmol) of
C5H3(SiMe3)2CMe2-3,5-Me2C6H3. The reaction mixture was
warmed to room temperature and was stirred overnight. The
volatiles were removed in vacuo, and the residue was stripped
with pentane. Extraction with pentane yielded 1.02 g (2.3
P r oced u r e for t h e Ca t a lyt ic E t h en e Con ver sion s
u sin g (η5-C5H 4CMe2P h )TiMe3 (5) w it h B(C6F 5)3 or
[P h NMe2H][B(C6F 5)4] a s Coca ta lyst. A stainless steel 500
mL autoclave (Medimex), fully temperature- and pressure-
controlled and equipped with solvent and catalyst injection

Catalytic Trimerization of Ethene
Organometallics, Vol. 21, No. 23, 2002 5135
systems, was preheated in vacuo for 45 min at 100 °C. The
reactor was cooled to the desired temperature, charged with
150 mL of toluene, and pressurized with ethene. After equili-
bration for 15 min, the appropriate amount of boron-based
cocatalyst in 5 mL of toluene was injected, together with 25
mL of toluene. Subsequently, a mixture of 2.50 g of cyclooctane
(internal standard) and 1.0 mL of a 15 mM stock solution of
the titanium trimethyl complex in toluene was injected,
together with 25 mL of toluene, to start the reaction. During
the reaction the ethene pressure was kept constant to within
0.1 bar of the initial pressure by replenishing the flow. The
run was ended by venting the reactor, and samples of the
reaction mixture were taken to analyze and quantify the
soluble components by GC and GC/MS. The polymer was
repeatedly rinsed with ethanol on a glass frit and was dried
in vacuo at 70 °C overnight.
After equilibration for 15 min, a slurry of 2.05 g of silica with
5 wt % MAO in 10 mL of toluene was injected together with
30 mL of toluene. Subsequently, a mixture of 2.50 g of
cyclooctane (internal standard) and 1.0 mL of a 15 mM stock
solution of the titanium trimethyl complex in toluene was
injected, together with 25 mL of toluene, to start the reaction.
During reaction the ethene pressure was kept constant to
within 0.1 bar of the initial pressure by replenishing flow. The
run was ended by adding an aliquot of ethanol, and the reactor
was vented. Remaining residual MAO was destroyed by adding
further ethanol, and samples of the reaction mixture were
taken to analyze and quantify the soluble components by GC
and GC/MS. The amount of polymer, precipitated on the silica
support, could not be quantified, since the suspended fine
polymer/support particles could not be separated quantita-
tively from the liquid fraction.
P r oced u r e for th e Ca ta lytic Eth en e Con ver sion s u s-
in g (η5-C5H4CMe2P h )TiMe3 (5) w ith MAO/SiO2 a s Coca ta -
lyst. A stainless steel 1 L autoclave (Medimex), fully temper-
ature- and pressure-controlled and equipped with solvent and
catalyst injection systems, was preheated in vacuo for 45 min
at 100 °C. The reactor was cooled to the desired temperature,
charged with 200 mL of toluene, and pressurized with ethene.
Ack n ow led gm en t. We thank Mr. G. Alberda for the
DSC analyses and Mr. A. J ekel for GC/MS and GPC
analyses.
OM020765A