Hafnocene catalysts for selective propylene oligomerization: Efficient synthesis of 4-methyl-1-pentene by β-methyl transfer

Article abstract of DOI:10.1021/ja063717g

A series of hafnocene complexes (η⁵-C₅Me ₄R¹)(η⁵-C₅Me₄R ²)HfCl₂ with [R¹, R²] = [H, H] (1), [Me, H] (2), [Me, Me] (3), [Et, Me] (4), [ⁱPr, Me] (5), [SiMe ₃, Me] (6), [^tBu, Me] (7), [ⁿBu, Me] (8), [^tBu, Me] (9), [Et, Et] (10), [ⁿBu, ⁿBu] (11), [ⁱBu, ⁱBu] (12) was tested as catalyst precursors for propylene oligomerization. Upon activation with methylaluminoxane or [Ph ₃C][B(C⁶F₅)₄]AlⁱBu ₃, complexes 2-4 and 8-12 catalyzed the dimerization of propylene to produce 4-methyl-1-pentene with selectivities ranging from 23.9 to 61.6 wt % in the product mixture. The selectivity was dependent on the nature of the substituents R¹ and R², with the highest value found for (η⁵-C₅Me₄ⁱBu) ₂HfCl₂ (12). Rapid deactivation was observed for 5-7, whereas (η⁵-C₅Me₄H)₂HfCl ₂ (1) polymerized propylene. 4-Methyl-1-pentene is proposed to form by repeated 1,2-insertion of propylene into the hafnocene methyl cation, followed by selective β-methyl elimination. Detailed analysis of the byproduct distribution (isobutene, 1-pentene, 2-methyl-1-pentene, 2,4-dimethyl-1-pentene, 4-methyl-1-heptene, 4,6-dimethyl-1-heptene), determined by gas chromatography, was performed with the aid of a stochastic simulation involving rate constants for the propagation by insertion, β-hydride elimination, and β-methyl elimination. The rate of termination is dependent on the structure of the growing chain of the active species as well as on the bulkiness of the cyclopentadienyl ligands. The selectivity highly depends on the reaction conditions (pressure, temperature, concentration of methylaluminoxane). The rates of β-methyl elimination leading to 4-methyl-1-pentene were proportional to propylene pressure for 2-4 and 8-10 but practically independent from propylene pressure for the sterically bulkier derivatives 11-12.

Full text of DOI:10.1021/ja063717g

Published on Web 09/12/2006
Hafnocene Catalysts for Selective Propylene Oligomerization:
Efficient Synthesis of 4-Methyl-1-pentene by â-Methyl Transfer
Yasuhiko Suzuki,† Takahiro Yasumoto,‡ Kazushi Mashima,‡ and Jun Okuda*
Contribution from the Institute of Inorganic Chemistry, RWTH Aachen UniVersity, Landoltweg 1,
D-52074 Aachen, Germany
Received June 6, 2006; E-mail: jun.okuda@ac.rwth-aachen.de
Abstract: A series of hafnocene complexes (η5-C5Me4R1)(η5-C5Me4R2)HfCl2 with [R1, R2] ) [H, H] (1), [Me,
H] (2), [Me, Me] (3), [Et, Me] (4), [iPr, Me] (5), [SiMe3, Me] (6), [tBu, Me] (7), [nBu, Me] (8), [iBu, Me] (9), [Et,
n
i
Et] (10), [nBu, Bu] (11), [iBu, Bu] (12) was tested as catalyst precursors for propylene oligomerization.
Upon activation with methylaluminoxane or [Ph3C][B(C6F5)4]/AliBu3, complexes 2-4 and 8-12 catalyzed
the dimerization of propylene to produce 4-methyl-1-pentene with selectivities ranging from 23.9 to 61.6 wt
% in the product mixture. The selectivity was dependent on the nature of the substituents R1 and R2, with
i
the highest value found for (η5-C5Me4 Bu)2HfCl2 (12). Rapid deactivation was observed for 5-7, whereas
(η5-C5Me4H)2HfCl2 (1) polymerized propylene. 4-Methyl-1-pentene is proposed to form by repeated 1,2-
insertion of propylene into the hafnocene methyl cation, followed by selective â-methyl elimination. Detailed
analysis of the byproduct distribution (isobutene, 1-pentene, 2-methyl-1-pentene, 2,4-dimethyl-1-pentene,
4-methyl-1-heptene, 4,6-dimethyl-1-heptene), determined by gas chromatography, was performed with the
aid of a stochastic simulation involving rate constants for the propagation by insertion, â-hydride elimination,
and â-methyl elimination. The rate of termination is dependent on the structure of the growing chain of the
active species as well as on the bulkiness of the cyclopentadienyl ligands. The selectivity highly depends
on the reaction conditions (pressure, temperature, concentration of methylaluminoxane). The rates of
â-methyl elimination leading to 4-methyl-1-pentene were proportional to propylene pressure for 2-4 and
8-10 but practically independent from propylene pressure for the sterically bulkier derivatives 11-12.
Introduction
trast, reportson the catalyzed synthesis of branched R-olefins
have been limited. 4-Methyl-1-pentene is one of the branched
Production of higher olefins by catalytic oligomerization of
lower olefins such as ethylene or propylene as raw materials
is an important process in the petrochemical industry.1 Recent
developments of homogeneous transition metal catalysts,
such as the use of carefully designed ancillary ligands, have
made the selective formation of linear R-olefins by dimer-
ization,2 trimerization,3 or tetramerization4 possible; in con-
R-olefins used for the production of (co)polymers with excel-
lent optical, thermal, and electrical properties.5 Currently,
4-methyl-1-pentene is manufactured by propylene dimerization
promoted by heterogeneous catalysts based on alkali metals.1d,6
There have been, to the best of our knowledge, only a small
number of publications describing attempts at developing
homogeneous catalysts based on transition metal metallocenes
with Cp* ligands (Cp*: η5-C5Me5) to produce 4-methyl-1-
pentene.7-10
† Permanent address: R&D Center, Mitsui Chemicals, Inc. 580-32
Nagaura, Sodegaura-City, Chiba 299-0265, Japan.
‡ Permanent address: Department of Chemistry, Graduate School of
Engineering, Osaka University, Toyonaka, Osaka 560-8531, Japan.
(1) (a) Parshall, G. W.; Ittel, S. D. Homogeneous Catalysis: The Applications
and Chemistry of Catalysis by Soluble Transition Metal Complexes; John
Wiley & Sons: New York, 1992. (b) Vogt, D. In Applied Homogeneous
Catalysis with Organometallic Compounds; Cornils, B., Herrmann, W.,
Eds.; VCH: New York, 1996; Vol. 1, pp 245-257. (c) Chauvin, Y.; Olivier,
H. In Applied Homogeneous Catalysis with Organometallic Compounds;
Cornils, B., Herrmann, W., Eds.; VCH: New York, 1996; Vol. 1, pp 258-
268. (d) Weissermel, K.; Arpe, H.-J. Industrial Organic Chemistry, 4th
ed.; Wiley-VCH: New York, 2003.
In 1990, Teuben et al. reported that cationic group 4
-
decamethylmetallocene complexes ([Cp*2MMe(C4H8S)]+BPh4 ;
M ) Zr, Hf; C4H8S: tetrahydrothiophene) catalyze the dimer-
ization or trimerization of propylene to give mainly 4-methyl-
1-pentene and 4,6-dimethyl-1-heptene.9a,b A mechanism con-
sisting of primary (1,2-) insertions of propylene into the metal-
(2) (a) Small, B. L. Organometallics 2003, 22, 3178-3183. (b) Tellmann, K.
P.; Gibson, V. C.; White, A. F.; Schmidt, G. F. Organometallics 2005, 24,
280-286. (c) Broene, R. D.; Brookhart, M.; Lamanna, W. M.; Volpe, A.
F., Jr. J. Am. Chem. Soc. 2005, 127, 17194-17195.
(5) (a) TPX, Mitsui Chemicals, Inc. http://www.mitsuichemicals.com/tpx.htm.
(b) Xu, G.; Cheng, D. Macromolecules 2001, 34, 2040-2047. (c) Irwin,
L. J.; Reibenspies, J. H.; Miller, S. A. J. Am. Chem. Soc. 2004, 126, 16716-
16717. (d) Gendler, S.; Groysman, S.; Goldschmidt, Z.; Shuster, M.; Kol,
M. J. Polym. Sci., Part A: Polym. Chem. 2005, 44, 1136-1146.
(6) Shaw, A. W.; Bittner, C. W.; Bush, W. V.; Holzman, G. J. Org. Chem.
1965, 30, 3286-3289.
(7) (a) Stevens, J. C.; Fordyce, W. A. (The Dow Chemical Company) U.S.
Patent 4,695,669, 1987. (b) Stevens, J. C.; Fordyce, W. A. (The Dow
Chemical Company) U.S. Patent 4,855,523, 1989.
(8) Watanabe, M.; Kuramoto, M. (Idemitsu Kosan Company) U.S. Patent 4,-
814,540, 1989.
(3) (a) For a review on trimerization, see: Dixon, J. T.; Green, M. J.; Hess, F.
M.; Morgan, D. H. J. Organomet. Chem. 2004, 689, 3641-3668. (b)
Deckers, P. J. W.; Hessen, B.; Teuben, J. H. Organometallics 2002, 21,
5122-5135.
(4) Bollmann, A.; Blann, K.; Dixon, J. T.; Hess, F. M.; Killian, E.; Maumela,
H.; McGuinness, D. S.; Morgan, D. H.; Neveling, A.; Otto, S.; Overett,
M.; Slawin, A. M. Z.; Wasserscheid, P.; Kuhlmann, S. J. Am. Chem. Soc.
2004, 126, 14712-14713.
9
10.1021/ja063717g CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 13017-13025
13017

A R T I C L E S
Suzuki et al.
Scheme 1. Proposed Mechanism for the Propylene Oligomerization by Metallocene Catalysts (M ) Zr, Hf)
Chart 1
alkyl bond of the active species, followed by â-methyl
elimination, was proposed to be crucial for the selective
formation of the vinyl compounds.9c
analysis. Complex 6 was found to be relatively air-sensitive,
while the others were air-stable.
At the same time, Yamazaki et al. reported that a combination
of Cp*2MCl2 (M ) Zr, Hf) and methylaluminoxane (MAO)
catalyzed the oligomerization of propylene with high activity
to produce mainly 4-methyl-1-pentene, as well as a considerable
amount of byproducts.10 The byproducts that contain up to 10
carbon atoms were identified, and a mechanism was proposed,
consisting of insertion and â-methyl elimination (major) and
â-hydride elimination (minor) pathways (Scheme 1).
Although it has been pointed out that sterically hindered
ligands in the cationic metallocenes are essential for this
catalysis, which is based on â-methyl elimination, there has been
no further attempt at improving the catalysts for selective
4-methyl-1-pentene production. At the same time, â-methyl
elimination remains a remarkable elementary step and is
attracting considerable interest.11-13 Although this reaction step
has been studied in detail, much has remained unclear, especially
with regard to the transition state.12,13
With the aim of controlling the selectivity of 4-methyl-1-
pentene formation, we comprehensively studied the steric effects
of hafnocene catalysts. Stochastic simulations to analyze kineti-
cally controlled byproduct distributions according to the mech-
anism proposed in Scheme 1 were performed. We report here
not only on our development of a significantly improved
peralkylhafnocene catalyst but also on new insights into the
process of homogeneously catalyzed propylene dimerization.
Li(C5Me4R1) + Cp*HfCl3 f
Cp*(η5-C5Me4R1)HfCl2 + LiCl (1)
(9) (a) Eshuis, J. J. W.; Tan, Y. Y.; Teuben, J. H. J. Mol. Catal. 1990, 62,
277-287. (b) Eshuis, J. J. W.; Tan, Y. Y.; Meetsma, A.; Teuben, J. H.;
Renkema, J.; Evans, G. G. Organometallics 1992, 11, 362-369. (c) Two
alternative mechanisms for the formation of 4-methyl-1-pentene are
conceivable. One involves double 2,1-insertion of propylene into a metal-
hydride bond followed by â-hydride elimination (ref 1c). This mechanism
is ruled out on the basis of the assumption that double 2,1-insertion within
a constrained peralkylmetallocene ligand sphere is sterically prohibitive.
The other alternative involves the oxidative coupling of two propylene units
at a divalent metallocene fragment to give 2,4-dimethylmetallacyclopentane
followed by â-hydride elimination/reductive elimination (McLain, S.;
Sancho, J.; Schrock, R. R. J. Am. Chem. Soc. 1980, 102, 5610-5618;
Takahashi, T.; Fujimori, T.; Seki, T.; Saburi, M.; Uchida, Y.; Rousset, C.
J.; Negishi, E. J. Chem. Soc., Chem. Commun. 1990, 182-183). This second
alternative can be excluded because it requires the formation of divalent
hafnocene species under the present conditions. Moreover, characteristic
byproducts of such an oxidative coupling/â-H elimination/reductive
elimination process are absent (2,3-dimethyl-1-butene, 4-methyl-2-pentene,
1-hexene, 2-hexene), and the formation of higher oligomers cannot be
explained by this process.
(10) Mise, T.; Kageyama, A.; Miya, S.; Yamazaki, H. Chem. Lett. 1991, 1525-
1528.
(11) (a) Watson, P. L.; Roe, D. C. J. Am. Chem. Soc. 1982, 104, 6471-6473.
(b) Moscardi, G.; Resconi, L.; Cavallo, L. Organometallics 2001, 20, 1918-
1931. (c) Yang, X.; Stern, C. L.; Marks, T. J. Angew. Chem., Int. Ed. Engl.
1992, 31, 1375-1377. (d) Resconi, L.; Piemontesi, F.; Franciscono, G.;
Abis, L.; Fiorani, T. J. Am. Chem. Soc. 1992, 114, 1025-32. (e) Yang,
X.; Stern, C. L.; Marks, T. J. J. Am. Chem. Soc. 1994, 116, 10015-10031.
(f) Repo, T.; Jany, G.; Hakala, K.; Klinga, M.; Polamo, M.; Leskela¨, M.;
Rieger, B. J. Organomet. Chem. 1997, 549, 177-186. (g) Weng, W.;
Markel, E. J.; Dekmezian, A. H. Macromol. Rapid Commun. 2000, 21,
1103-1107. (h) Guo, Z.; Swenson, D. C.; Jordan, R. F. Organometallics
1994, 13, 1424-1432. (i) Hajela, S.; Bercaw, J. E. Organometallics 1994,
13, 1147-1154. (j) Horton, A. D. Organometallics 1996, 15, 2675-2677
(k) Beswick, C. L.; Marks, T. J. J. Am. Chem. Soc. 2000, 122, 10358-
10370. (l) Lin, M.; Spivak, G. J.; Baird, M. C. Organometallics 2002, 21,
2350-2352. (m) Volkis, V.; Nelkenbaum, E.; Lisovskii, A.; Hasson, G.;
Semiat, R.; Kapon, M.; Botoshansky, M.; Eishen, Y.; Eisen, M. S. J. Am.
Chem. Soc. 2003, 125, 2179-2194.
Results
Synthesis. Metallocene complexes 2, 4-9, 12, and 13 (Chart
1) were prepared by the method used to synthesize 3.14 Lithium
salts of the appropriate cyclopentadiene derivatives were reacted
with Cp*HfCl3 (for 2, 4-9) (eq 1) or with MCl4 (M ) Hf, Zr;
for 12 and 13) (eq 2) under reflux in xylene. The complexes
were identified by 1H and 13C NMR spectroscopy and elemental
(12) Chirik, P. J.; Dalleska, N. F.; Henling, L. M.; Bercaw, J. E. Organometallics
2005, 24, 2789-2794.
(13) (a) Yang, P.; Baird, M. C. Organometallics 2005, 24, 6005-6012. (b) Yang,
P.; Baird, M. C. Organometallics 2005, 24, 6013-6018.
9
13018 J. AM. CHEM. SOC. VOL. 128, NO. 39, 2006

Hafnocene Catalysts for Propylene Oligomerization
A R T I C L E S
Table 1. Productivity of Production of 4-Methyl-1-pentenea
2 Li(C5Me4R1) + MCl4 f (η5-C5Me4R1)2MCl2 + 2 LiCl
(2)
complex
(R1, R2)
TOF
1
activator
productivityb
(min-
)
Preliminary Screening. For the catalyzed reaction with
propylene, hafnocene precatalysts (η5-C5Me4R1)(η5-C5Me4R2)-
HfCl2 were activated with MAO or [Ph3C][B(C6F5)]/AliBu3 and
the products were analyzed by gas chromatography (GC) after
aqueous workup and removal of polypropylene. Initial screening
showed that complexes 1 and 5-7 had low productivity for
the production of 4-methyl-1-pentene and were therefore unsuit-
able for further study (Table 1). The sterically less encumbered
complex (η5-C5Me4H)2HfCl2 (1) gave mainly atactic polypro-
pylene, with low selectivity for 4-methyl-1-pentene, although
the activity toward propylene was high (1027 g of polymer/
(mmol of Hf ‚ h)). Complexes 5-7 showed low productivities
and, moreover, were found to lose their activities at the early
stages of the reaction (Figure 1). It is interesting that irreversible
deactivation occurred more rapidly for complexes 6 and 7
bearing the tertiary alkyl substituents SiMe3 and tBu than for 5
1
MAO
TrBAr4
20
52
7.9
20.6
c
(H, H)
3
MAO
TrBAr4
261
318
103.4
126.0
(Me, Me)
4
MAO
TrBAr4
315
363
124.8
143.8
(Et, Me)
5
MAO
TrBAr4
303
100
120.0
39.6
(iPr, Me)
6
MAO
TrBAr4
12
14
4.8
5.5
(SiMe3, Me)
7
MAO
TrBAr4
4
2
1.6
0.8
(tBu, Me)
a Conditions: Hf, 0.008 mmol; MAO, 2 mmol; [Ph3C][B(C6F5)4], 0.009
mmol; AliBu3, 0.25 mmol; toluene, 20 mL; propylene, 0.2 MPa; 50 °C, 30
min. b (g of 4-methyl-1-pentene)/(mmol of Hf ‚ h). c [Ph3C][B(C6F5)4]/
AliBu3.
i
with the secondary alkyl substituent Pr, while 4 with only
primary alkyl substituents (Me and Et) did not undergo this
deactivation step. This suggests that R-methyl groups of the
substituents are involved in the deactivation. We propose that
C-H activation occurs, followed by the formation of a Hf-C
bond, to give a so-called “tucked-in” complex (Scheme 2).15
The other complexes 2-4 and 8-13 were all stable enough to
maintain their activities for at least an hour. The product
distributions were not affected by the reaction time over this
period under a constant propylene pressure. This observation
indicates that the detected products are not consumed further
during the catalysis.
Influence of Steric Effects on Selectivity. The product
distributions and activities of complexes 2-4 and 8-13 were
determined (Table 2). A run with decamethylhafnocene complex
3/MAO gave, with slight deviations, results similar to those
reported by Yamazaki et al.,10 and in addition to 4-methyl-1-
pentene (C6-1, 31.1 wt %), a considerable amount of other
byproducts also formed. Less than 0.3 wt % 2,4-dimethylpen-
tane, which may be formed by alkyl transfer to aluminum,
formed. The amount of higher oligomers obtained with Cp*-
(η5-C5Me4H)HfCl2 (2) was significantly more than that obtained
with Cp*2HfCl2 (3), which has one methyl group more in the
ligand sphere (3/MAO: 24.2 wt %, 2/MAO: 48.8 wt %).16
Conversely, a bulkier ligand prevented not only propylene
polymerization but also the formation of isobutene (C4),
2-methyl-1-pentene (C6-2), 2,5-dimethyl-1-pentene (C7), 4-methyl-
1-heptene (C8), and 4,6-dimethyl-1-heptene (C9-1). In summary,
the selectivity for 4-methyl-1-pentene (C6-1) increases in the
order of the bulkiness of the substituents [R1, R2]: [Me, Me]
Figure 1. Production of 4-methyl-1-pentene catalyzed by complexes 4, 5,
6, and 7. Conditions: Hf, 0.008 mmol; [Ph3C][B(C6F5)4], 0.009 mmol; Al-
iBu3, 0.25 mmol; toluene, 20 mL; propylene, 0.2 MPa; 50 °C; 30 min. *
represents productivity in g of 4-methyl-1-pentene/(mmol of Hf ‚ h).
Scheme 2. Proposed Mechanism of Catalyst Deactivation by
C-H Activation of an R-Methyl Group on a Cyclopentadienyl
Substituent (R ) H or alkyl)
n
< [Et, Me] < [nBu, Me] ≈ [iBu, Me] ≈ [Et, Et] < [nBu, Bu]
i
≈ [iBu, Bu]; under the standard conditions up to 61.6 wt %
C6-1 forms. The selectivity for 4-methyl-1-pentene (C6-1
)
formation by zirconocene complex 13 was not as high as that
of hafnocene congener 12, in agreement with the results for
the decamethylmetallocene complexes previously reported by
Teuben et al. and Yamazaki et al.9,10
(14) Roddick, D. M.; Fryzuk, M. D.; Seidler, P. F.; Hillhouse, G. L.; Bercaw,
J. E. Organometallics 1985, 4, 97-104.
Rate Constants. To obtain more information from the
selectivity of the oligomerization, ratios of rate constants for
each of the elementary steps in the proposed mechanism
(Scheme 3) were estimated. Values of k4/k5 and k8/k9 for the
competitive â-methyl and â-hydride elimination reactions were
(15) (a) Fischer, J. M.; Piers, W. E.; Young, V. G., Jr. Organometallics 1996,
15, 2410-2412. (b) Bercaw, J. E.; Brintzinger, H. H. J. Am. Chem. Soc.
1971, 93, 2048-2050. (c) Bercaw, J. E. J. Am. Chem. Soc. 1974, 96, 5087-
5095. (d) Luinstra, G. A.; Teuben, J. H. J. Am. Chem. Soc. 1992, 114,
3361-3367.
(16) A similar observation made with Cp*(η5-C5Me4H)ZrCl2/MAO and (η5-
C5Me4H)2ZrCl2/MAO was mentioned in ref 10.
9
J. AM. CHEM. SOC. VOL. 128, NO. 39, 2006 13019

A R T I C L E S
Suzuki et al.
Table 2. Catalytic Oligomerization with Activated Metallocene
and propagation, leading to higher selectivity for 4-methyl-1-
pentene. Another important observation is that k4/k5 is always
significantly larger than k8/k9, and k4 is always considerably
larger than k8 for a given catalyst. As a result, C5 formation is
suppressed and more C8 is formed.
Complexesa
selectivityc (wt %)
complex
(R1, R2, M)
activator activityb C4
C5 C6
C6
-2
C7 C8 C9
-1
others
-1
2
MAO
TrBAr4 935
977
0.9 0.5 23.9 0.2 2.9 3.4 19.4 48.8
2.9 1.3 24.0 1.2 6.9 8.3 14.5 41.5
d
(H, Me, Hf)
It is also noteworthy that a critical point of polymerization/
oligomerization change is found for complex 2, where the ratio
of elimination to propagation (k4+k5)/k1 changes from 0.94
(MAO) to 1.36 ([Ph3C][B(C6F5)4]/AliBu3).
3
MAO
923 10.2 3.2 31.1 3.2 4.4 8.5 15.2 24.2
TrBAr4 895
7.7 3.3 39.6 1.9 3.3 7.4 17.5 19.3
(Me, Me, Hf)
4
MAO
TrBAr4 891
921
8.9 3.5 34.3 3.0 4.1 7.9 16.0 22.3
7.1 3.5 40.9 1.6 2.6 7.6 18.2 18.5
Pressure Dependence of Rate Constants. The oligomer-
ization was investigated at various pressures (0.12-0.3 MPa)
to determine whether the rate constants depend on propylene
concentration (i.e., kobs ) k[propylene]), provided Henry’s law
is adhered to. Interestingly, only the bulkiest complexes 11-
13 gave more of the higher oligomers at higher pressure. The
other complexes 2-4 and 8-10 gave less C4, C5, C6-2, C7,
and C8 byproducts formed at higher pressure. Further analysis
of the rate constants showed that k1/k4 (propagation and â-methyl
elimination) was proportional to pressure for complexes 11-
13, while k1/k4 remained constant for the other complexes 2-4
and 8-10. In contrast, values of k4/k5 (â-methyl elimination
and â-hydride elimination) were constant for complexes 11-
13 but proportional to pressure for the other complexes 2-4
and 8-10. Typical examples are shown for complexes Cp*2-
(Et, Me, Hf)
8
MAO
TrBAr4 862
883
5.2 2.8 40.2 1.3 3.1 6.2 18.2 23.0
4.2 3.3 50.7 0.7 1.8 4.7 20.7 13.9
(nBu, Me, Hf)
9
MAO
TrBAr4 764
860
5.0 2.8 41.5 1.0 2.5 6.2 19.7 21.3
4.2 3.5 52.1 0.7 1.7 4.7 20.3 12.8
(iBu, Me, Hf)
10
MAO
TrBAr4 794
831
4.4 3.3 41.9 1.1 2.1 7.2 18.6 21.4
5.2 3.9 48.0 1.1 2.0 6.1 19.3 14.4
(Et, Et, Hf)
11
MAO
TrBAr4 432
450
4.3 4.6 54.2 0.9 2.2 4.6 18.1 11.1
4.4 5.1 58.3 0.7 1.5 3.9 18.3 7.8
(nBu, nBu, Hf)
12
MAO
TrBAr4 305
311
3.8 4.9 58.5 0.8 1.7 4.4 16.6 9.3
4.6 6.2 61.6 0.8 1.6 3.9 15.8 5.5
(iBu, iBu, Hf)
13
MAO
TrBAr4 237
265
2.0 1.9 28.8 0.8 2.9 3.8 20.1 39.7
3.9 3.3 34.3 1.2 2.5 4.1 21.4 29.3
(iBu, iBu, Zr)
HfCl2 (3) and (η5-C5Me4 Bu)2HfCl2 (12) with [Ph3C][B(C6F5)4]/
i
AliBu3 in Figure 2. The activity, considered to reflect the
propagation rate constant k1, was proportional to pressure for
all complexes (Supporting Information). Therefore, the â-methyl
elimination rate constant k4 does not depend on propylene
concentration for complexes 11, 12, and 13. However k4 values
depend on propylene concentration for complexes 2-4 and
8-10, while the rate constant for â-hydride elimination k5 does
not depend on propylene concentration for all complexes. The
results suggest that the transition state for â-methyl elimination
as the rate-determining step is associated with a propylene
molecule, provided the catalyst’s ligand sphere is sterically not
too encumbered.
a Conditions: catalyst, 0.008 mmol; MAO, 2 mmol, [Ph3C][B(C6F5)4],
0.009 mmol; AliBu3, 0.25 mmol; benzene, 19.5 mL; n-hexane, 0.5 mL;
propylene, 0.2 MPa; 50 °C; 30 min. b (Sum of all the observed products
in g)/(mmol of catalyst ‚ h). c C4: isobutene, C5: 1-pentene, C6-1: 4-methyl-
1-pentene, C6-2: 2-methyl-1-pentene, C7: 2,5-dimethyl-1-pentene, C8:
4-methyl-1-heptene, C9-1: 2,4-dimethyl-1-heptene, others: higher oligo-
mers; see Experimental Section. d [Ph3C][B(C6F5)4]/AliBu3.
calculated from the experimentally (GC) determined ratios (mol
of C6-1)/(mol of C7) and (mol of C5)/(mol of C6-2), respectively.
Estimation of ratios of rate constants for the elimination steps
(k3, k4, k6, k8) and propagation (k1) required stochastic simula-
tion17 using an algorithm developed for this purpose, because
the mechanism is far too complicated to be solved deterministi-
cally.18 Despite the approximations adopted for the mechanism
(viz, the rate constants k1, k4, and k5 are kept constant regardless
of the lengths of the growing chains), the simulation showed
excellent agreement with the experimental results (Supporting
Information).
During the simulation trials, we noticed that k6/k1 must
be small for appropriate amounts of C5 and C8 to be obtained
whenever a complex is sterically encumbered (e.g., k6/k1 ) 0
for complex 12, Table 3). The small k6/k1 value corresponds
to the assumption that the extent of â-methyl elimination for
the active species with an n-propyl chain is limited. This
limitation impedes the regeneration of the methyl or isobutyl
metal cation when the metal hydride cation is the initial active
species. Ratios of the rate constants are summarized in Table
4.
Temperature Dependence of Rate Constants. The oligo-
merization was also investigated in the temperature range 25-
75 °C. The trend was similar for all complexes; i.e., more of
the C4, C5, C6-2, C7, C8 byproducts formed at higher temper-
atures at the expense of higher oligomers. A decrease in higher
oligomers was previously reported for the decamethylmetal-
locene complexes.8,9b The increase of the byproducts (C4, C5,
C6-2, C7, C8) stemming from the increase of â-hydride elimina-
tion was unexpected, however, and the dependency differed
from one complex to another. Analysis with regard to the rate
constants clearly showed these dependencies (Figure 3 and
Supporting Information). The activities were lower at higher
temperature, e.g., 488 and 243 g/(mmol of Hf ‚ h) at 25 and 75
°C, respectively, for complex 12/[Ph3C][B(C6F5)4]/AliBu3), but
other factors (e.g., increase in solubility or increased formation
of propylene) precluded precise interpretation. The k8/k9 value
was always lower than the k4/k5 value [(k8/k9)/(k4/k5) was in the
range of 0.10-0.35].
As the bulkiness of the ligand increases, all ratios k4/k5, k8/
k9, k4/k1, and k8/k1 increase; i.e., for the bulkier catalysts,
â-methyl elimination is preferred to both â-hydride elimination
Influence of MAO Concentration. As shown in Table 2,
[Ph3C][B(C6F5)4]/AliBu3 is a slightly better cocatalyst than MAO
with regard to the selectivity of 4-methyl-1-pentene.19 As often
discussed in the field of homogeneous olefin polymerization,
MAO’s performance is variable and simply changes with
(17) (a) Schaad, L. J. J. Am. Chem. Soc. 1963, 85, 3588-3592. (b) Connors,
K. A. Chemical Kinetics: The Study of Reaction Rates in Solution; VCH:
New York, 1990. (c) Flisak, Z.; Ziegler, T. Macromolecules 2005, 38,
9865-9872.
(18) When â-H elimination becomes negligible, the Flory-Schulz theory is
applicable; see ref 9b.
9
13020 J. AM. CHEM. SOC. VOL. 128, NO. 39, 2006

Hafnocene Catalysts for Propylene Oligomerization
A R T I C L E S
Scheme 3. Definition of Rate Constants kia
a M ) Zr, Hf; Cp′ ) η5-C5Me4R1,2. R ) Me (transfer to the metal) or R ) iBu (transfer to propylene). k1: Propagation (insertion into alkyl-metal bond),
k2: â-methyl elimination from isobutyl-metal species, k3: â-hydride elimination from isobutyl-metal species, k4: â-methyl elimination from 2,4-dimethylpentyl
(2,4,6-trimethylheptyl, etc.)-metal species, k5: â-hydride elimination from 2,4-dimethylpentyl (2,4,6-trimethylheptyl etc.)-metal species, k6: â-methyl
elimination from n-propyl-metal species, k7: â-hydride elimination from n-propyl-metal species, k8: â-methyl elimination from 2-methylpentyl-metal
species, k9: â-hydride elimination from 2-methylpentyl-metal species.
Table 3. Results of Simulation for 12/[Ph3C][B(C6F5)4]/AliBu3
Table 4. Ratios of Rate Constants
selectivitya (wt %)
complex
(R1, R2, M)
activator
k4/k5
k8/k9
k4/k8/k1
k6/k1
C4
C5
C6
-1
C6
C7
C8
C9
-1
others
-2
2
MAO
TrBAr4
9.9
4.0
3.3
1.3
0.85/0.10/1
1.09/0.12/1
1.63 (k8/k1)
0
4.8
4.7
1.7
4.4
67.7
62.7
0.2
0.6
1.7
1.6
1.3
3.7
16.5
16.2
6.1
6.1
a
(H, Me, Hf)
3
MAO
TrBAr4
8.3
13.9
1.2
2.0
1.86/0.37/1
2.27/0.48/1
a Formation of ethylene and propylene is excluded. C4: isobutene, C5:
1-pentene, C6-1: 4-methyl-1-pentene, C6-2: 2-methyl-1-pentene, C7: 2,5-
dimethyl-1-pentene, C8: 4-methyl-1-heptene, C9-1: 2,4-dimethyl-1-heptene.
(Me, Me, Hf)
4
MAO
TrBAr4
9.9
18.2
1.4
2.6
2.02/0.45/1
2.34/0.49/1
(Et, Me, Hf)
concentration.20 The oligomerization was also strongly affected
by the concentration of MAO in both activity and selectivity
(Figure 4). The activities were improved as the concentration
was increased and surpassed those with [Ph3C][B(C6F5)4]/Al-
iBu3. Surprisingly, k4/k5, which represents the dominance of
â-methyl elimination over â-hydride elimination, concomitantly
decreased from values comparable to those obtained with [Ph3C]-
[B(C6F5)4]/AliBu3, regardless of their dependence on propylene
pressure (vide supra). Although the reason for this remains
unclear, the tradeoff makes it difficult to apply MAO for the
optimal conditions. At higher MAO concentration, k1/k4, which
determines the average molecular weight of oligomers, increased
only slightly (Supporting Information).
8
MAO
TrBAr4
14.7
32.4
2.5
5.5
2.16/0.48/1
2.60/0.75/1
(nBu, Me, Hf)
9
MAO
TrBAr4
19.7
35.2
3.3
5.9
2.20/0.50/1
2.75/0.80/1
(iBu, Me, Hf)
10
MAO
TrBAr4
23.3
27.7
3.5
4.2
2.27/0.59/1
2.65/0.72/1
(Et, Et, Hf)
11
MAO
TrBAr4
30.0
45.2
6.2
9.0
3.32/1.10/1
3.79/1.39/1
(nBu, nBu, Hf)
12
MAO
TrBAr4
39.3
46.2
7.8
9.2
4.27/1.42/1
4.84/1.63/1
(iBu, iBu, Hf)
13
MAO
TrBAr4
11.6
16.3
3.1
3.3
1.05/0.35/1
1.37/0.47/1
(iBu, iBu, Zr)
Discussion
a [Ph3C][B(C6F5)4]/AliBu3.
The influence of steric effects on â-methyl elimination
basically agrees with a model originally proposed by Teuben
et al. (Figure 5).9b Steric repulsions between the substituents of
the cyclopentadienyl ligands and the substituents in the â-posi-
tion of the growing chain in the active species force the â-methyl
group into an energetically more favorable position within the
equatorial plane where the LUMO is located. However, this
simple model requires modifications when the structure of the
(19) For complexes 1 and 2, MAO suppresses â-hydride elimination more
efficiently than [Ph3C][B(C6F5)4]/AliBu3.
(20) Ju¨ngling, S.; Mu¨lhaupt, R. J. Organomet. Chem. 1995, 497, 27-32.
9
J. AM. CHEM. SOC. VOL. 128, NO. 39, 2006 13021

A R T I C L E S
Suzuki et al.
i
Figure 2. Pressure dependence of k1/k4 (left) and k4/k5 (right) for complexes Cp*2HfCl2 (3) and (η5-C5Me4 Bu)2HfCl2 (12). Conditions: Hf, 0.008 mmol;
i
[Ph3C][B(C6F5)4], 0.009 mmol; Bu3Al, 0.25 mmol; benzene, 19.5 mL; n-hexane, 0.5 mL; 50 °C; 30 min.
n
i
Figure 3. Temperature dependence of k1/k4 (left) and k4/k5 (right) for complexes (η5-C5Me4 Bu)2HfCl2 (11) and (η5-C5Me4 Bu)2HfCl2 (12). Conditions: Hf,
0.008 mmol; [Ph3C][B(C6F5)4], 0.009 mmol; AliBu3, 0.25 mmol; benzene, 19.5 mL; n-hexane, 0.5 mL; propylene, 0.2 MPa; 30 min.
Figure 4. Influence of MAO concentration on activity (left) and k4/k5 (right) for complexes Cp*2HfCl2 (3) and (η5-C5Me4 Bu)2HfCl2 (12). Conditions: Hf,
i
0.4 mM; benzene, 19.5 mL; n-hexane, 0.5 mL; propylene, 0.2 MPa; 50 °C; 30 min. * represents (sum of observed products in g)/(mmol of Hf ‚ h).
9
13022 J. AM. CHEM. SOC. VOL. 128, NO. 39, 2006

Hafnocene Catalysts for Propylene Oligomerization
A R T I C L E S
Scheme 4. Formation of C8 Byproducta
i
a R ) Me (transfer to the metal) or R ) Bu (transfer to propylene).
Scheme 5. Active Species with n-Propyl Growing Chain
Scheme 6. Resting State and Subsequent Associative Exchange of Olefins
Scheme 7. Idealized Catalytic Cycle for 4-Methyl-1-pentene
Synthesis
This leads to facile coordination of propylene and subsequent
insertion (Scheme 5).
Recently, Baird et al. proposed â-methyl transfer to the
propylene monomer by a concerted mechanism for the zir-
conocene catalyzed propylene polymerization.13 The pressure
dependency of k4 for complexes 2-4 and 8-10 is consistent
with this proposal. The conformation model in Figure 5 can
also be applied here, as long as the π-orbital of the incoming
monomer is positioned in the equatorial plane of the metal-
locene. The requirement for a crowded transition state is in
agreement with the results observed for the bulkier complexes
11-13; direct â-methyl transfer to the metal center can be
assumed when k4 values are independent from pressure.
Another interpretation is that the â-methyl transfer to the
metal is fast and the subsequent associative exchange of olefins
constitutes the rate-determining step (Scheme 6). In this case,
it is reasonable to assume that the steric bulk of substituents
both on the cyclopentadienyl ligands and at the δ-position of
the growing chain serve as the driving force for expelling the
olefin by steric repulsion in the complexes 2-4 and 8-10.
Finally, the data shown in Table 3 could be interpreted as
that the selectivity is improved at the expense of activity. On
one hand, it is possible that the bulkiness of the catalyst hinders
propylene coordination and/or insertion to make the propagation
rate small, resulting in a low activity. On the other hand, it is
also possible that the catalytic activity is not strongly affected
by the bulkiness. In the observed activity the rate for propylene
formation is not included. The amount of propylene should be
larger when the rate ratio of elimination to propagation is higher.
growing chain is included. The relationship between the values
of k1, large k4, and small k8 to produce a very large amount of
C8 byproduct suggests that the growing chain with a secondary
carbon in the δ-position influences the selectivity for â-methyl
elimination effectively (Scheme 4). This steric effect of δ-sub-
stituents R3 and R4 is complimentary to that of the ring
substituents R1 and R2 and weakened (i.e., k4/k8 becoming
smaller) as the bulkiness of R1 and R2 increases.
In the case of the active species with an n-propyl growing
chain, it is probable that the terminal methyl group is directed
toward the outside of the wedge in the absence of another alkyl
group in the geminal position. This accounts for the small k6
values, in particular for the sterically encumbered complexes.
Conclusion
The introduction of bulky substituents in dhafnocene catalysts
greatly improved the selectivity for the formation of 4-methyl-
1-pentene, according to Scheme 7, and revealed some mecha-
nistic details of the catalysis. The analysis of relative rates
Figure 5. Steric effect of the ligands on â-methyl elimination.
9
J. AM. CHEM. SOC. VOL. 128, NO. 39, 2006 13023

A R T I C L E S
Suzuki et al.
showed that the bulkiness of the cyclopentadienyl ligands22
as well as the structure of the growing chain significantly
influence the selectivity. The dependence of the catalysis on
the propylene pressure gave further information on the â-methyl
elimination mechanism. We are continuing investigations aimed
at catalyst development and at clarifying the mechanism in more
detail.
CH3), 2.04 (s, 6H, C5(CH3)4Et), 2.03 (s, 6H, C5(CH3)4Et), 2.02 (s,15H,
C5(CH3)5), 0.94 (t, 3H, 3JHH ) 7.6 Hz, CH2CH3). 13C{1H} NMR (CD-
Cl3): δ 127.2, 122.1, 121.9, 120.9, 19.9, 14.4, 12.0, 11.9, 11.7. Anal.
Calcd for C21H32Cl2Hf: C, 47.24; H, 6.04. Found: C, 47.25; H, 6.10.
Synthesis of Cp*(η5-C5Me4 Pr)HfCl2 (5). The synthesis was carried
i
out according to the procedure to prepare 2 using 5-isopropyl-1,2,3,4-
tetramethylcyclopentadiene instead of 1,2,3,4-tetramethylcyclopenta-
diene; yield: 39%. 1H NMR (CDCl3): δ 3.06 (septet, 1H, 3JHH ) 7.1
i
Hz, CH(CH3)2), 2.05 (s, 15H, C5(CH3)5), 1.98 (s, 6H, C5(CH3)4 Pr),
3
1.10 (d, 6H, JHH ) 7.1 Hz, CH(CH3)2). 13C{1H} NMR (CDCl3): δ
Experimental Section
130.4, 122.8, 122.0, 121.2, 27.4, 21.9, 12.6, 12.0, 11.8. Anal. Calcd
for C22H34Cl2Hf: C, 48.23; H, 6.25. Found: C, 48.26; H, 6.13.
Synthesis of Cp*(η5-C5Me4SiMe3)HfCl2 (6). The synthesis was
carried out according to the procedure to prepare 2 using trimethyl-
(2,3,4,5-tetramethyl-2,4-cyclopentadien-1-yl)silane instead of 1,2,3,4-
tetramethylcyclopentadiene; yield: 35%. 1H NMR (CDCl3): δ 2.21(s,
15H, C5(CH3)5), 2.03 (s, 6H, C5(CH3)4SiMe3), 2.02 (s, 6H, C5(CH3)4-
SiMe3), 0.26 (s, 9H, Si(CH3)3). 13C{1H} NMR (CDCl3): δ 130.6, 124.7,
122.3, 121.9, 15.3, 12.0, 11.8, 2.1. Anal. Calcd for C22H36Cl2HfSi: C,
45.72; H, 6.28. Found: C, 45.71; H, 6.20.
General. All manipulations involving air- and moisture-sensitive
organometallic compounds were carried out under argon using the
standard Schlenk technique. Cp*HfCl3,21 5-isopropyl-1,2,3,4-tetra-
methylcyclopentadiene,23a and5-tert-butyl-1,2,3,4-tetramethylcyclopenta-
diene23b were prepared according to the literature. 5-n-Butyl-1,2,3,4-
tetramethylcyclopentadiene and 5-isobutyl-1,2,3,4-tetramethylcyclopenta-
diene were prepared according to the literature23a with slight modifica-
tions (i.e., use of n-butyllithium/hexane and isobutylmagnesium
chloride/Et2O, respectively). Benzene, n-hexane, n-pentane, toluene, and
xylene were dried and deoxygenated by distillation over sodium
benzophenone ketyl under nitrogen. MAO (Aldrich Chemical Co.) was
purchased as a 1.0 M toluene solution, and the remaining trimethyl-
aluminum was evaporated under a vacuum to obtain a white powder.
2,4-Dimethyl-1-pentene (C7) was purchased from TCI Europe N. V.
4-Methyl-1-heptene (C8) and 4,6-dimethyl-1-heptene (C9-1) were
obtained according to the literature10 and the references therein.
Propylene (Aldrich Chemical Co.), [Ph3C][B(C6F5)4] (Asahi Glass Co.),
Synthesis of Cp*(η5-C5Me4 Bu)HfCl2 (7). The synthesis was carried
out according to the procedure to prepare 2 using 5-tert-butyl-1,2,3,4-
t
tetramethylcyclopentadiene instead of 1,2,3,4-tetramethylcyclopenta-
1
t
diene; yield: 16%. H NMR (CDCl3): δ 2.26 (s, 6H, C5(CH3)4 Bu),
t
2.03 (s, 15H, C5(CH3)5), 2.01 (s, 6H, C5(CH3)4 Bu), 1.33 (s, 9H,
C(CH3)3). 13C{1H} NMR (CDCl3): δ 135.7, 122.3, 121.2, 36.5, 31.8,
15.8, 12.1 (one resonance for methyl group and one resonance for Cp
ring missing). Anal. Calcd for C23H36Cl2Hf: C, 49.16; H, 6.46.
Found: C, 49.17; H, 6.45.
1
and other commercially available reagents were used as supplied. H
NMR (200 MHz) and 13C NMR (50 MHz) spectra were measured on
a VARIAN-UNITY spectrometer at ambient temperature. Chemical
shifts were referenced to the residual solvent resonances and reported
relative to tetramethylsilane. Elemental analyses were performed by
the Microanalytical Laboratory of the Johannes Gutenberg-University,
Mainz, Germany.
Synthesis of Cp*(η5-C5Me4nBu)HfCl2 (8). The synthesis was carried
out according to the procedure to prepare 2 using 5-n-butyl-1,2,3,4-
tetramethylcyclopentadiene instead of 1,2,3,4-tetramethylcyclopenta-
1
diene; yield: 43%. H NMR (CDCl3): δ 2.43 (m, 2H, CH2CH2CH2-
n
CH3), 2.02 (overlapped, s, 27H, C5(CH3)5 and C5(CH3)4 Bu), 1.22-
3
1.32 (m, 4H, CH2CH2CH3), 0.90 (t, 3H, JHH ) 6.7 Hz, CH2CH3).
Synthesis of Cp*(η5-C5Me4H)HfCl2 (2). In a Schlenk flask, a 2.5
M solution of LinBu in hexane (0.44 mL, 1.1 mmol) was added to
1,2,3,4-tetramethylcyclopentadiene (122 mg, 1.0 mmol) in pentane (10
mL) at 0 °C. The mixture was stirred overnight at room temperature.
The resulting white suspension was decanted, washed with pentane (2
× 10 mL), and dried under a vacuum to yield a white powder (125
mg, 0.975 mmol). Solid Cp*HfCl3 (369 mg, 0.864 mmol) was mixed
with Li(C5Me4H), xylene (15 mL) was added, and the suspension was
refluxed for 2 days. All volatiles were removed under a vacuum. CH2-
Cl2 (50 mL) and 1 M hydrochloric acid (50 mL) were added. The
organic layer was separated, and the aqueous layer was extracted with
CH2Cl2 (2 × 25 mL). The combined CH2Cl2 fractions were dried over
Na2SO4 and filtered. The solvent was removed under a vacuum, washed
with pentane (3 mL), and dried under a vacuum to give a pale yellow
powder; yield: 269 mg (0.532 mmol, 62%). Recrystallization from hot
13C{1H} NMR (CDCl3): δ 126.3, 121.9, 121.8, 121.0, 32.4, 26.7, 23.1,
14.1, 12.0, 11.9, 11.8. Anal. Calcd for C23H36Cl2Hf: C, 49.16; H, 6.46.
Found: C, 49.20; H, 6.46.
Synthesis of Cp*(η5-C5Me4 Bu)HfCl2 (9). The synthesis was carried
i
out according to the procedure to prepare 2 using 5-isobutyl-1,2,3,4-
tetramethylcyclopentadiene instead of 1,2,3,4-tetramethylcyclopenta-
1
3
diene; yield: 51%. H NMR (CDCl3): δ 2.35 (d, 2H, JHH ) 7.5 Hz,
i
CH2CH(CH3)2), 2.04 (s, 6H, C5(CH3)4 Bu), 2.03 (s, 15H, C5(CH3)5),
i
2.02 (s, 6H, C5(CH3)4 Bu), 1.60-1.79 (m, 1H, CH(CH3)2), 0.82 (d, 6H,
3JHH ) 6.6 Hz, CH(CH3)2). 13C{1H} NMR (CDCl3): δ 125.6, 121.9,
121.4, 35.8, 29.7, 22.8, 12.6, 12.0 (one Cp ring resonance missing).
Anal. Calcd for C23H36Cl2Hf: C, 49.16; H, 6.46. Found: C, 49.14; H,
6.47.
Synthesis of (η5-C5Me4H)2HfCl2 (1). The complex was mentioned
in the literature without characterization.24 In a Schlenk flask, a 2.5 M
solution of LinBu in hexane (1.8 mL, 4.5 mmol) was added to 1,2,3,4-
tetramethylcyclopentadiene (549 mg, 4.50 mmol) in pentane (15 mL)
at 0 °C. The mixture was stirred overnight at room temperature. The
resulting colorless powder was washed with pentane (10 mL) and dried
under a vacuum to give a white powder; yield: 461 mg (3.60 mmol,
80%). Solid HfCl4 (576 mg, 1.80 mmol) was mixed with Li(C5Me4H),
xylene (15 mL) was added, and the suspension was refluxed for 2 days.
All volatiles were removed under a vacuum. CH2Cl2 (50 mL) and 1 M
hydrochloric acid (50 mL) were added. The organic layer was separated,
and the aqueous layer was extracted with CH2Cl2 (2 × 25 mL). The
combined CH2Cl2 fractions were dried over Na2SO4 and filtered. The
solvent was removed under a vacuum to give a pale yellow powder;
yield: 249 mg (0.50 mmol, 28%). Recrystallization from hot hexane
1
hexane afforded colorless crystals. H NMR (CDCl3): δ 5.57 (s, 1H,
C5(CH3)4H), 2.07(s, 15H, C5(CH3)5), 2.03 (s, 6H, C5(CH3)4H), 1.85 (s,
6H, C5(CH3)4H). 13C{1H} NMR (CDCl3): δ 130.4, 121.4, 116.5, 109.1,
12.4, 12.3, 12.1. Anal. Calcd for C19H28Cl2Hf: C, 45.12; H, 5.58.
Found: C, 44.98; H, 5.58.
Synthesis of Cp*(η5-C5Me4Et)HfCl2 (4). The synthesis was carried
out according to the procedure to prepare 2 using 5-ethyl-1,2,3,4-tetra-
methylcyclopentadiene instead of 1,2,3,4-tetramethylcyclopentadiene;
1
3
yield: 45%. H NMR (CDCl3): δ 2.46 (q, 2H, JHH ) 7.6 Hz, CH2-
(21) Llina´s, G. H.; Mena, M.; Palacios, F.; Royo, P.; Serrano, R. J. Organomet.
Chem. 1988, 340, 37-40.
(22) For a recent example of a dramatic effect on metallocene reactivity by a
ring substituent pattern, see: Pool, J. A.; Lobkovsky, E.; Chirik, P. J. Nature
2004, 427, 527-530. For a review on bulky cyclopentadienyl ligands,
see: Okuda, J. Top. Curr. Chem. 1992, 160, 97-145.
(23) (a) Evans, W. J.; Forrestal, K. J.; Ziller, J. W. Polyhedron 1998, 17, 4015-
4021. (b) du Plooy, K. E.; du Toit, J.; Demetrius, D. C.; Coville, N. J. J.
Organomet. Chem. 1996, 520, 265-268.
(24) Fries, A.; Mise, T.; Matsumoto, A.; Ohmori, H.; Wakatsuki, Y. Chem.
Commun. 1996, 783-784.
9
13024 J. AM. CHEM. SOC. VOL. 128, NO. 39, 2006

Hafnocene Catalysts for Propylene Oligomerization
A R T I C L E S
afforded colorless crystals. 1H NMR (CDCl3): δ 1.79 (s, 12H,
C5(CH3)4H), 2.04 (s, 12H, C5(CH3)4H), 5.23 (s, 2H, C5(CH3)4H). 13C-
{1H} NMR (CDCl3): δ 127.8, 118.4, 109.2, 13.7, 11.9. Anal. Calcd
for C18H26Cl2Hf: C, 43.96; H, 5.33. Found: C, 43.89; H, 5.21.
Synthesis of (η5-C5Me4Et)2HfCl2 (10). This compound was men-
tioned in the literature without characterization data.8 The synthesis
was carried out in analogy to the procedure to prepare 1 using 5-ethyl-
1,2,3,4-tetramethylcyclopentadiene instead of 1,2,3,4-tetramethylcy-
mL), a 0.50 M solution of AliBu3 in n-hexane (0.5 mL), and a 4 mM
solution of a metallocene complex in benzene (2 mL) were added via
syringe into an autoclave. After the atmosphere was replaced by
propylene, the mixture was heated to 50 °C and stirred for 10 min at
0.50 kgf/cm2 (gauge pressure). The reaction was started with an addition
of a solution of [Ph3C][B(C6F5)4] in benzene (3 mM, 3 mL) via syringe.
Alteration of Reaction Conditions in Propylene Oligomerization.
Propylene pressure, temperature, MAO concentration, and reaction time
in the above procedures were varied to the designated value in order
to investigate the influence of the reaction conditions. When the pressure
was 2.00 kgf/cm2, an inlet with a septum was exchanged for a screwed
plug with a Viton O-ring after the reaction was started. The reaction
was quenched after cooling at 10 °C for 2 min without a propylene
supply. In the case of preliminary screening, toluene was used instead
of benzene.
clopentadiene; yield: 43%. 1H NMR (CDCl3): δ 2.45 (q, 4H, 3JHH
)
7.6 Hz, CH2CH3), 2.02 (s, 12H, C5(CH3)4Et) 2.01 (s, 12 H, C5(CH3)4-
Et), 0.92 (t, 6H, 3JHH ) 7.6 Hz, CH2CH3). 13C{1H} NMR (CDCl3): δ
127.1, 122.2, 120.9, 20.0, 14.4, 12.0, 11.7. Anal. Calcd for C22H34Cl2-
Hf: C, 48.23; H, 6.25. Found: C, 48.24; H, 6.27.
Synthesis of (η5-C5Me4 Bu)2HfCl2 (11). The complex was men-
n
tioned in the literature without characterization.8 The synthesis was
carried out according to the procedure to prepare 1 using 5-n-butyl-
1,2,3,4-tetramethylcyclopentadiene instead of 1,2,3,4-tetramethylcy-
Stochastic Simulation. The program (SuzuKin.nb) was written in
Mathematica25 code and run on a Macintosh G4 (Apple Computer).
Active species which appear in the plausible mechanism (Scheme 1)
were indexed with the carbon number of an alkyl group attached to
the metal (e.g., cat[0] for hydride species, cat[1] for methyl species,
and so on). Because propylene insertion occurs only in a primary mode
without resulting in regioisomeric structures, each intermediate species
was uniquely defined by the index. The simulation started from cat[1]
when activated with MAO or cat[4] with [Ph3C][B(C6F5)4]/iBu3Al. After
a generated random number was referred to, probabilities of possible
competitive pathways were calculated on the basis of relative scales
of rate constants, and an action was selected.17c If a propagation path
was selected, cat[n] was changed to cat[n + 3]. If a â-methyl elimination
path was selected, cat[n] was changed to cat[1] (or cat[4] in the case
of â-methyl transfer to propylene). If a â-hydrogen elimination path
was selected, cat[n] was changed to cat[0]. If no action was selected,
cat[n] was not changed. To integrate first-order differential equations
with good precision,17a,b these trials were applied to a large number of
active species simultaneously (10 000) and for a large number of cycles
(1000). Formation of olefins was estimated with the use of a history
log of each active species. The algorithms in detail are described in
the Supporting Information. Values of k4/k5 and k8/k9 were estimated
from (mol of C6-1)/(mol of C7) and (mol of C5)/(mol of C6-2), and k1
was set to a constant (tentatively to 1). As for other rate constants,
major pathways in the mechanism were considered first, and then minor
pathways were adjusted subsequently. Values of k4 and k8 were
tentatively set to (mol of C9-1)/(mol of C6-1) and (mol of C5)/(mol of
C8)*(mol of C9-1)/(mol of C6-1)*k4. First k4 was adjusted using the
values of (wt % of C9-1)/(wt % of C6-1) and a total of other higher
oligomers with constraining {k2, k3, k6, k7, k8} to {2k4, k5, k4, 2k5, k4}
(the constant “2” stems from intuitive understanding that probability
of â-elimination is proportional to the number of methyl or hydride
substituents in the â-position), then k8 was adjusted using the value of
(wt % of C8)/(wt % of C5) with constraining {k2, k3, k6, k7} to {2k8, k9,
k8, 2k9}, then k3 was adjusted using the amount of C4 with constraining
{k2, k6, k7} to {2k8, k8, 2k3}, and finally k6 was adjusted using the amount
of C5 and C8. The remaining k2 and k7 were left as 2k8 and 2k3,
respectively; in fact they do not influence the distribution of the
observed products strongly. This procedure was repeated until the
experimental result was well simulated.
1
3
clopentadiene; yield: 44%. H NMR (CDCl3): δ 2.41 (t, 4H, JHH
)
n
7.1 Hz, CH2CH2CH2CH3), 2.01 (overlapped, s, 24H, C5(CH3)4 Bu),
1.07-1.39 (m, 8H, CH2CH2CH3), 0.89 (t, 6H, 3JHH ) 6.7 Hz, CH2CH3).
13C{1H} NMR (CDCl3): δ 126.3, 122.0, 121.0, 32.4, 26.8, 23.1, 14.1,
12.0 (one methyl group resonance missing). Anal. Calcd for C26H42-
Cl2Hf: C, 51.70; H, 7.01. Found: C, 51.74; H, 7.09.
Synthesis of (η5-C5Me4 Bu)2HfCl2 (12). The synthesis was carried
i
out according to the procedure to prepare 1 using 5-isobutyl-1,2,3,4-
tetramethylcyclopentadiene instead of 1,2,3,4-tetramethylcyclopenta-
1
3
diene; yield: 43%. H NMR (CDCl3): δ 2.36 (d, 4H, JHH ) 6.6 Hz,
i
i
CH2CH(CH3)2), 2.05 (s, 12H, C5(CH3)4 Bu), 2.02 (s, 12H, C5(CH3)4 -
Bu), 1.58-1.79 (m, 2H, CH(CH3)2), 0.60 (d, 12H, 3JHH ) 7.3 Hz, CH-
(CH3)2). 13C{1H} NMR (CDCl3): δ 125.6, 121.8, 121.4, 35.9, 29.7,
22.8, 12.6, 12.0. Anal. Calcd for C26H42Cl2Hf: C, 51.70; H, 7.01.
Found: C, 51.70; H, 6.98.
Synthesis of (η5-C5Me4 Bu)2ZrCl2 (13). The synthesis was carried
i
out according to the procedure to prepare 12 using ZrCl4 instead of
HfCl4; yield: 45%. 1H NMR (CDCl3): δ 2.31 (d, 4H, 3JHH ) 7.3 Hz,
i
i
CH2CH(CH3)2), 1.96 (s, 12H, C5(CH3)4 Bu), 1.94 (s, 12H, C5(CH3)4 -
Bu), 1.59-1.78 (m, 2H, CH(CH3)2), 0.81 (d, 12H, 3JHH ) 6.6 Hz, CH-
(CH3)2). 13C{1H} NMR (CDCl3): δ 127.3, 123.6, 123.1, 36.0, 29.6,
22.8, 12.7, 12.1. Anal. Calcd for C26H42Cl2Zr: C, 60.43; H, 8.19.
Found: C, 60.45; H, 8.22.
Propylene Oligomerization Activated with MAO. Benzene (15.5
mL), n-hexane (0.50 mL, internal standard), and a 1.00 M solution of
MAO in benzene (2.0 mL) were added via syringe into an argon filled
autoclave (100 mL Bu¨chi miniclave) equipped with an inlet/outlet for
propylene/vacuum, an inlet with a septum, a thermocouple, a magnetic
stirrer, and an oil bath. After the atmosphere was replaced by propylene,
the mixture was heated to 50 °C and stirred for 10 min at 0.50 kgf/cm2
(gauge pressure). The reaction was started with an addition of a solution
of a metallocene complex in benzene (4 mM, 2 mL) via syringe, the
septum was exchanged for a new one, and propylene was supplied at
1.00 kgf/cm2 (gauge pressure) for 30 min. After stopping the propylene
supply, water (0.1 mL) was added via syringe and the reactor was
cooled to 10 °C and settled for 30 min. The reaction mixture was
washed with 1.0 M HCl (20 mL) and then with water (20 mL). All
volatiles are trapped in a flask cooled with liquid nitrogen under a
vacuum (10-5 bar). The residue was weighed, and the trapped volatiles
were thawed and analyzed by GC. GC analyses were carried out on a
Shimadzu GC-2010A (FID) with a DB-1 column (60 m, J&W
Acknowledgment. We thank Ms. M. Kirch and Mr. Y.
Kuzuba for the synthesis of the starting materials.
Scientific) and hydrogen carrier gas. Amounts of C4, C5, C6-1, C6-2
,
Supporting Information Available: Experimental details of
simulation, oligomerization, algorithms (PDF), and a Math-
ematica notebook for simulation. This material is available free
of charge via the Internet at http://pubs.acs.org.
C7, C8, and C9-1 were estimated using the relative peak area of n-hexane
as an internal standard and scaling factors referenced to the authentic
compounds. Amounts of other higher oligomers were estimated on the
assumption that the peak area is approximately proportional to the
number of carbon atoms.
JA063717G
Propylene Oligomerization Activated with [Ph3C][B(C6F5)4]/
AliBu3. The above procedure was modified as follows: Benzene (14.5
(25) Mathematica 5.0; Wolfram Research, Inc. http://www.wolfram.com.
9
J. AM. CHEM. SOC. VOL. 128, NO. 39, 2006 13025