Highly selective ethylene trimerization catalyzed by half-sandwich indenyl titanium complexes with pendant arene groups and MAO

Article abstract of DOI:10.1016/j.molcata.2013.03.005

A series of half-sandwich indenyl titanium complexes [Ind-(bridge)-Ar] TiCl₃ (C1-C9) bearing pendant arene group on indenyl ring have been synthesized and used for the catalytic ethylene trimerization to 1-hexene in the presence of MAO. The mol

Full text of DOI:10.1016/j.molcata.2013.03.005

Journal of Molecular Catalysis A: Chemical 373 (2013) 85–95
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Highly selective ethylene trimerization catalyzed by half-sandwich indenyl
titanium complexes with pendant arene groups and MAO
Yanlu Zhang, Haiyan Ma, Jiling Huang∗
Laboratory of Organometallic Chemistry, East China University of Science and Technology, 130 Meilong Road, P.O. Box 310, Shanghai 200237, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of half-sandwich indenyl titanium complexes [Ind-(bridge)-Ar]TiCl3 (C1–C9) bearing pendant
arene group on indenyl ring have been synthesized and used for the catalytic ethylene trimerization to
1-hexene in the presence of MAO. The molecular structures of complexes C3 [Ind-C(cyclo-C5H10)-Ph]TiCl3
and C5 [Ind-C(cyclo-C5H10)-(p-MePh)]TiCl3 have been established by single-crystal X-ray diffraction
study. No intramolecular coordination interaction between the arene moiety and the titanium center
could be observed in the solid state of these complexes. At 0 ◦C and 0.8 MPa of ethylene pressure, upon
activation with MAO, C3 possesses the highest activity of 1968 kg of 1-hexene/(mol-Ti h) and the overall
selectivity of 95.9% by mass for 1-hexene, and is also more active than the corresponding cyclopentadi-
enyl analog [Cp-C(cyclo-C5H10)-Ph]TiCl3 (C10) under the identical conditions. The substituents of various
steric and electronic effects on the pendant arene group and the bridge unit between the indenyl and this
arene group exert great influence on the activity and selectivity of these indenyl titanium complexes for
ethylene trimerization to 1-hexene. Similar to cyclopentadienyl analogs, upon activation with MAO the
resultant indenyl catalytic systems also show great sensitivity to the temperature. With the increase of
the reaction temperature, both the activity and selectivity of 1-hexene declined.
Received 21 December 2012
Received in revised form 18 February 2013
Accepted 10 March 2013
Available online xxx
Keywords:
Titanium
Indenyl
Pendant arene
Ethylene
Trimerization
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
production [11]. Suttil and co-workers used the same ligand skele-
ton as Hessen’s to complex with Zr and Hf centers. The resultant
Linear ␣-olefins (LAOs), mainly produced via ethylene oligomer-
ization processes, have a variety of applications including as
co-monomers for the copolymerization with ethylene to gener-
ate linear low-density polyethylenes (LLDPEs) and as versatile
intermediates for many industrially products like plasticizers,
lubricants, detergents, and surfactants [1,2]. The catalytically selec-
tive trimerization of ethylene to 1-hexene has been extensively
studied in recent years [3–7]. Although the vast majorities are
based on chromium catalysts [8], complexes of other metals like
titanium can also effectively catalyze ethylene trimerization [9].
Since Hessen’s group found that [Cp-bridge-Ar]TiCl3/MAO systems
(Chart 1, 1) are able to effectively catalyze ethylene trimerization
to 1-hexene [10], a lot of efforts have been devoted in this area
[11–17]. The intrinsic high selectivity for ethylene trimerization of
these systems is suggested to be essentially depending on the hemi-
labile behavior of the cyclopentadienyl-arene ancillary ligand. The
nature of the bridging group between the Cp ring and the pendant
arene is also crucial for obtaining a good selectivity in 1-hexene
zirconium or hafnium complexes (Chart 1, 2) mainly exhibit ethy-
lene polymerization activity, although there is an evidence for
1-hexene formation in the liquid fraction. Obviously, such dis-
tinct catalytic behavior derives from the nature of different metal
centers [18]. Park’s group recently reported that titanatrate com-
plex 3 bearing a Cp-arene ligand (Chart 1, 3) could obtain a
selectivity of 92.3 wt.% and a catalytic activity of 1.03 × 104 kg of
1-hexene/(molTi-h) for ethylene trimerization to 1-hexene [19].
Huang and co-workers successfully developed a system involv-
ing half-sandwich titanium complexes with a pendant thienyl
group (Chart 1, 4), which could serve as efficient catalyst systems
for ethylene trimerization, indicating that ␩1-S coordination to
the titanium center plays an important role [20]. Suzuki and co-
workers recently reported a new system derived from FI catalyst,
containing Ti complexes bearing phenoxy-imine ligand with pen-
dant aryl-OR donor (Chart 1, 5). In the presence of MAO, these
complexes can selectively transform ethylene to 1-hexene with
unprecedented activity [3.15 × 105 kg of 1-hexene/(molTi-h)] [21].
Among the catalytic systems based on titanium species, Hes-
sen’s [Cp-bridge-Ar]TiCl3/MAO system is the most effective for the
formation of 1-hexene and thus has been thoroughly investigated,
with more work mainly focusing on theoretical studies recently
[22,23]. In comparison with the cyclopentadienyl analogs, titanium
∗
Corresponding author. Tel.: +86 21 64253519; fax: +86 21 64253519.
E-mail addresses: yingyub2006@126.com (Y. Zhang), haiyanma@ecust.edu.cn
(H. Ma), Jlhuang@ecust.edu.cn (J. Huang).
1381-1169/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcata.2013.03.005

86
Y. Zhang et al. / Journal of Molecular Catalysis A: Chemical 373 (2013) 85–95
R1
R2
R2
O
O Ti
O
Zr Cl
Cl
N
Ti Cl
Cl
Cl
R1 = SiMe3, H;
R2, R3 = alkyl
Cl
R3
3
2
1
Z
N
TiCl3
O
Ti Cl
S
Z = OPh, OMe
R = cumyl, adamantyl
Cl
Cl
4
R
5
Chart 1. Typical ethylene trimerization catalysts based on group IV metals.
complexes with pendant arene-substituted indenyl group have
never been reported and adopted for the selective ethylene trimer-
ization to date. Regarding the diversity and flexibility of indenyl
group, we would like to investigate the effect of different indenyl
groups on such trimerization process. Herein we report the synthe-
sis of a series of half-sandwich indenyl titanium complexes with
pendant arene groups and their catalytic behavior for selective
ethylene trimerization to 1-hexene.
organic layers were combined and dried with anhydrous MgSO4.
All the volatiles were evaporated in vacuo, and the residue was
purified via chromatography (petroleum ether as eluent, silica gel
column) to give L2 as light yellow oil (2.5 g, 28%). 1H NMR (400 MHz,
CDCl3): ı 7.41 (d, J = 7.2 Hz, 1H, Ind-H), 7.23–7.15 (m, 2H, Ph-H),
7.09–7.02 (m, 3H, Ind-H, 2Ph-H), 6.98 (t, J = 7.6 Hz, 1H, Ind-H), 6.73
(d, J = 7.6 Hz, 1H, Ind-H), 6.44 (s, 1H, Ind-H), 3.38 (s, 2H, Ind-H), 2.28
(s, 3H, PhCH3), 1.65 (s, 6H, CH3). 13C{ H} NMR (125 MHz, CDCl3):
1
149.4 (Ind-C), 145.4 (Ind-C), 140.5 (Ph-C), 135.1 (Ind-C), 129.3 (Ind-
C), 128.9 (Ind-C), 126.0 (Ind-C), 125.2 (Ind-C), 124.0 (Ph-C), 123.2
(Ph-C), 122.5 (Ph-C), 122.1 (Ind-C), 40.0 (Ind-C), 29.8 (CMe2-C), 24.9
(CMe2-C), 20.9 (PhMe-C).
2. Experimental
2.1. General considerations
Synthesis of 3-[(1-phenyl)cyclohexyl]indene (L3). The procedure
was similar to that of L2. Phenyl lithium (3.4 g, 40 mmol) and 6-
cyclohexylidenebenzofulvene (7.8 g, 40 mmol) were used to give
L3 as white solids (5.4 g, 46%). mp: 77.6–79.1 ◦C. 1H NMR (400 MHz,
CDCl3): ı 7.45–7.38 (m, 3H, Ind-H), 7.29–7.23 (m, 2H, Ph-H),
7.16–7.10 (m, 1H, Ph-H), 7.09–6.97 (m, 3H, Ind-H, 2Ph-H), 6.52 (t,
J = 2.0 Hz, 1H, Ind-H), 3.41 (d, J = 2.0 Hz, 2H, Ind-H), 2.40–2.32 (m,
2H, C5H10), 2.26–2.13 (m, 2H, C5H10), 1.71–1.56 (m, 5H, C5H10),
Unless otherwise stated, all manipulations were carried out
under a dry argon atmosphere using standard Schlenk techniques.
Solvents were purified by distillation over sodium benzophe-
none (diethyl ether, tetrahydrofuran, toluene, light petroleum and
n-hexane) and CaH2 (dichloromethane). MAO (toluene solution,
0.9 M) was purchased from Witco GmbH. Polymerization grade
ethylene was used without further purification. The benzofulvenes
[24,25], substituted phenyl lithium [26–28], ligand L1 [29] were
synthesized according to the literature procedures.
NMR spectra were recorded on a Bruker AVANCE-400 Hz spec-
trometer using tetramethylsilane (TMS) as an internal standard and
employing CDCl3 as solvent (dried over CaH2 for a minimum of 24 h,
distilled and stored under argon). Elemental analyses were per-
formed on an EA-1106 spectrometer. GC-MS measurements were
performed on a Agilent 6890N/5975 instrument, and GC analyses
were performed on a Fuli 9790 instrument equipped with Agi-
lent J&W HP-5 GC column. IR data were tested on Bruker TENSOR
27. MS spectra were recorded on Shimadzu Biotech Axima Perfor-
mance. Melting point was tested on XGW x-4 micro melting point
instrument.
1
1.53–1.39 (m, 1H, C5H10). 13C{ H} NMR (125 MHz, CDCl3): ı 149.8
(Ind-C), 147.3 (Ind-C), 145.1 (Ph-C), 143.8 (Ind-C), 129.5 (Ind-C),
128.0 (Ind-C), 126.9 (Ind-C), 125.5 (Ind-C), 125.2 (Ph-C), 123.8 (Ph-
C), 123.5 (Ph-C), 122.2 (Ind-C), 44.4 (Ind-C), 37.4 (C5H10-C), 36.3
(C5H10-C), 26.5 (C5H10-C), 22.9 (C5H10-C).
Synthesis of 3-{[1-(2-methylphenyl)]cyclohexyl}indene (L4). The
procedure was similar to that of L2. o-Methylphenyl lithium (3.1 g,
32 mmol) and 6-cyclohexylidenebenzofulvene (6.3 g, 32 mmol)
were used to give L4 as white solids (2.3 g, 25%). mp: 123.1–125.2 ◦C
1H NMR (400 MHz, CDCl3): ı 7.81–7.76 (m, 1H, Ind-H), 7.42–7.32
(m, 1H, Ind-H), 7.30–7.22 (m, 1H, Ph-H), 7.10 (td, J = 7.2 Hz, J = 0.8 Hz,
1H, Ph-H), 7.04 (td, J = 7.2 Hz, J = 0.8 Hz, 1H, Ph-H), 6.98–6.88 (m, 2H,
Ph-H), 6.84–6.72 (m, 1H, Ph-H), 6.54 (t, J = 2.0 Hz, 1H, Ind-H), 3.40
(d, J = 2.0, 2H, Ind-H), 2.44–2.25 (m, 4H, C5H10), 2.21–2.12 (m, 3H,
1
PhCH3), 1.75–1.49 (m, 6H, C5H10). 13C{ H} NMR (125 MHz, CDCl3):
2.2. Synthesis
ı 149.9 (Ind-C), 144.9 (Ind-C), 144.0 (Ph-C), 143.8 (Ind-C), 137.7
(Ind-C), 133.0 (Ind-C), 129.2 (Ind-C), 127.9 (Ind-C), 126.05 (Ph-C),
125.4 (Ph-C), 125.3(Ph-C), 123.8 (Ph-C), 123.5 (Ph-C), 121.9 (Ind-C),
44.8 (Ind-C), 37.2 (C5H10-C), 35.6 (C5H10-C), 26.6 (C5H10-C), 22.9
(C5H10-C), 21.8 (PhMe-C).
2.2.1. Synthesis of ligands
Synthesis of 3-{2-[2-(4-methylphenyl)]propyl}indene (L2). To a
solution of p-methylphenyl lithium (3.5 g, 36 mmol) in 50 mL of
diethyl ether at −20 ◦C was added dropwise a solution of 6,6-
dimethylbenzofulvene (5.6 g, 36 mmol) in 15 mL of diethyl ether.
The reaction mixture was warmed to room temperature, stirred
for 24 h, and then hydrolyzed with 100 mL of ice-water. The water
layer was extracted with light petroleum ether (50 mL), and the
Synthesis of 3-{[1-(4-methylphenyl)]cyclohexyl}indene (L5). The
procedure was similar to that of L2. p-Methylphenyl lithium (3.5 g,
36 mmol) and 6-cyclohexylidenebenzofulvene (7.1 g, 36 mmol)
were used to give L5 as white solids (7.5 g, 72%). mp: 84.2–85.7 ◦C.

Y. Zhang et al. / Journal of Molecular Catalysis A: Chemical 373 (2013) 85–95
87
1H NMR (400 MHz, CDCl3): ı 7.42 (d, J = 7.2 Hz, 1H, Ind-H), 7.34–7.28
(m, 2H, Ind-H), 7.12–7.00(m, 5H, Ind-H, 4Ph-H), 6.53–6.49 (m,
1H, Ind-H), 3.41 (s, 2H, Ind-H), 2.41–2.32 (m, 2H, C5H10), 2.29 (s,
3H, PhCH3), 2.24–2.14 (m, 2H, C5H10), 1.71–1.54 (m, 5H, C5H10),
(Ind-C), 147.8 (Ind-C), 131.6 (Ph-C), 130.1 (Ind-C), 129.4 (Ind-C),
128.8 (Ind-C), 128.5 (Ind-C), 127.6 (Ind-C), 126.9 (Ph-C), 126.6 (Ph-
C), 125.8 (Ph-C), 121.0 (Ind-C), 114.5 (Ind-C), 41.8 (CMe2-C), 31.4
(CMe2-C), 27.4 (CMe2-C). MALDI-TOF-MS (m/z): 351.0 (M+−Cl). IR
(cm−1): 2963, 2876, 1599, 1464, 1077, 1053, 752, 695. Anal. Calcd.
for C18H17Cl3Ti·(0.15C5H12): C, 56.53; H, 4.76; found: C, 56.72; H,
5.21%.
1
1.53–1.41 (m, 1H, C5H10). 13C{ H} NMR (125 MHz, CDCl3): ı 150.0
(Ind-C), 145.1 (Ind-C), 144.2 (Ph-C), 143.9 (Ind-C), 134.9 (Ind-C),
129.3 (Ind-C), 128.8 (Ind-C), 126.8 (Ind-C), 125.2 (Ph-C), 123.7 (Ph-
C), 123.5 (Ph-C), 122.3 (Ind-C), 44.1 (Ind-C), 37.3 (C5H10-C), 36.4
(C5H10-C), 26.5 (C5H10-C), 22.9 (C5H10-C), 20.8 (PhMe-C).
[Ind-CMe2-(p-MePh)]TiCl3 (C2). The procedure was similar to
that of C1. L2 (1.5 g, 6 mmol), n-BuLi (2.5 mL, 6 mmol) and TiCl4
(0.66 mL, 6 mmol) were used to give deep red crystals (1.5 g, 62%).
mp: 123.3–125.1 ◦C. 1H NMR (400 MHz, CDCl3): ı = 7.70–7.66 (m,
1H, Ind-H), 7.36–7.34 (m, 1H, Ind-H), 7.34–7.32 (m, 1H, Ph-H),
7.32–7.28(m, 1H, Ind-H), 7.28–7.25 (m, 1H, Ph-H), 7.25–7.22 (m, 1H,
Ph-H), 7.03–7.96 (m, 4H, 3Ind-H, Ph-H), 2.24 (s, 3H, PhCH3), 1.98
Synthesis of 3-{[1-(4-tert-butylphenyl)]cyclohexyl}indene (L6).
The procedure was similar to that of L2. p-tert-Butylphenyl
lithium (6.3 g, 45 mmol) and 6-cyclohexylidenebenzofulvene (8.8 g,
45 mmol) were used to give L6 as white solids (4.5 g, 30%).
mp: 93.1–95.6 ◦C. 1H NMR (400 MHz, CDCl3): ı 7.42 (d, J = 6.4 Hz,
1H, Ind-H), 7.38–7.30 (m, 2H, Ind-H), 7.29–7.23 (m, 2H, Ph-H),
7.21–6.97 (m, 3H, Ind-H, 2Ph-H), 6.50 (s, 1H, Ind-H), 3.41 (s, 2H, Ind-
H), 2.44–2.33 (m, 2H, C5H10), 2.25–2.11 (m, 2H, C5H10), 1.72–1.35
1
(s, 3H, CH3), 1.90 (s, 3H, CH3). 13C{ H} NMR (125 MHz, CDCl3): ı
148.9 (Ind-C), 144.8 (Ind-C), 136.1 (Ph-C), 131.6 (Ind-C), 130.2 (Ind-
C), 129.4 (Ind-C), 129.2 (Ind-C), 128.8 (Ind-C), 127.6 (Ph-C), 127.1
(Ph-C), 125.7 (Ph-C), 121.0 (Ind-C), 114.63 (Ind-C), 41,6 (CMe2-C),
31.4 (CMe2-C), 27.5 (CMe2-C), 20.8 (PhMe-C). MALDI-TOF-MS (m/z):
367.0 (M+−Cl) IR (cm−1): 2964, 2877, 1599, 1445, 1096, 1028, 901,
832, 751, 696. Anal. Calcd. for C19H19Cl3Ti: C, 56.83; H, 4.77; found:
C, 56.78; H, 4.71%.
1
(m, 6H, C5H10), 1.28 [s, 9H, C(CH3)3]. 13C{ H} NMR (125 MHz,
CDCl3): 149.9 (Ind-C), 148.1 (Ind-C), 145.2 (Ph-C), 144.1 (Ind-C),
144.0 (Ind-C), 129.5 (Ind-C), 126.5 (Ind-C), 125.3 (Ind-C), 124.9
(Ph-C), 123.7 (Ph-C), 123.6 (Ph-C), 122.5 (Ind-C), 44.1 (Ind-C),
37.4 (C5H10-C), 36.4 (CMe3-C), 34.2 (C5H10-C), 31.4 (CMe3-C), 26.6
(C5H10-C), 23.0 (C5H10-C).
[Ind-C(cyclo-C5H10)-Ph]TiCl3 (C3). The procedure was similar to
that of C1. L3 (1.6 g, 6 mmol), n-BuLi (2.5 mL, 6 mmol) and TiCl4
(0.66 mL, 6 mmol) were used to give deep red crystals of C3 (620 mg,
15%). mp: 143.4–145.1 ◦C. 1H NMR (400 MHz, CDCl3): ı 7.79 (d,
J = 8.8 Hz, 1H, Ind-H), 7.66 (d, J = 8.8 Hz, 1H, Ind-H), 7.44 (d, J = 7.2 Hz,
2H, Ph-H), 7.42–7.37 (m, 1H, Ph-H), 7.35–7.27 (m, 3H, Ind-H),
7.21–7.14 (m, 2H, Ph-H), 7.09 (d, J = 4.4 Hz, 1H, Ind-H), 3.07 (d,
J = 14.4 Hz, 1H, C5H10), 2.75–2.65 (m, 2H, C5H10), 2.23–2.11 (m,
1H, C5H10), 1.85–1.65 (m, 4H, C5H10), 1.50–1.30 (m, 2H, C5H10).
Synthesis of 3-{[1-(4-flurophenyl)]cyclohexyl}indene (L7). The
procedure was similar to that of L2. p-Fluorophenyl lithium (5.1 g,
50 mmol) and 6-cyclohexylidenebenzofulvene (9.8 g, 50 mmol)
were used to give L7 as white solids (3.9 g, 27%). mp: 78.2–79.7 ◦C
1H NMR (400 MHz, CDCl3): ı 7.42 (d, J = 7.2 Hz, 1H, Ind-H), 7.40–7.33
(m, 2H, Ind-H), 7.11–6.89(m, 5H, Ind-H, 4Ph-H), 6.53 (m, 1H, Ind-
H), 3.41 (s, 2H, Ind-H), 2.39–2.30 (m, 2H, C5H10), 2.22–2.08 (m,
2H, C5H10), 1.70–1.56 (m, 5H, C5H10), 1.51–1.39 (m, 1H, C5H10).
1
13C{ H} NMR (125 MHz, CDCl3): ı 150.7 (Ind-C), 142.3 (Ind-C),
1
13C{ H} NMR (125 MHz, CDCl3): ı 145.2 (Ind-C), 143.6 (Ind-C),
131.9 (Ph-C), 129.2 (Ind-C), 129.1 (Ind-C), 128.6 (Ind-C), 128.5 (Ind-
C), 128.3 (Ind-C), 127.8 (Ph-C), 127.2 (Ph-C), 126.5 (Ph-C), 119.9
143.0 (Ph-C), 143.0 (Ind-C), 129.6 (Ph-C), 128.5 (Ind-C), 128.4 (Ind-
C), 125.3 (Ind-C), 123.9 (Ind-C), 123.6 (Ind-C), 122.2(Ph-C), 114.8
(Ph-C), 114.6 (Ph-C), 44.0 (Ind-C), 37.4 (C5H10-C), 36.5 (C5H10-C),
26.4 (C5H10-C), 22.9 (C5H10-C).
(Ind-C), 115.0 (Ind-C), 45.0 (C5H10-C), 36.4 (C5H10-C), 34.9 (C5H10
-
C), 25.9 (C5H10-C), 22.5 (C5H10-C), 22.2 (C5H10-C). MALDI-TOF-MS
(m/z): 393.1 (M+−Cl). IR (cm−1): 2925, 2858, 1599, 1466, 1057,
1030, 750, 700. Anal. Calcd. for C21H21Cl3Ti: C, 58.98; H, 4.95; found:
C, 59.21; H, 5.23%.
Synthesis of 3-[2-(2-phenyladmantyl)]indene (L8). The procedure
was similar to that of L2. Phenyllithium (3.4 g, 40 mmol) and 6-(2-
adamantylidene)benzofulvene (9.9 g, 40 mmol) were used to give
pure L8 as white solids (8.8 g, 67%). mp: 121.1–123.3 ◦C. 1H NMR
(400 MHz, CDCl3): ı = 7.80 (d, J = 8.0 Hz, 1H, Ind-H), 7.60 (d, J = 8.0 Hz,
2H, Ind-H), 7.38 (d, J = 8.0 Hz, 1H, Ind-H), 7.26 (t, J = 8.0 Hz, 2H, Ph-
H), 7.20 (t, J = 8.0 Hz, 1H, Ph-H), 7.15–7.03 (m, 2H, Ph-H), 6.50 (t,
J = 2.0 Hz, 1H, Ind-H). 3.43–3.20 (m, 4H, 2 C9H14, 2Ind-H), 2.40–1.65
[Ind-C(cyclo-C5H10)-(o-MePh)]TiCl3 (C4). The procedure was
similar to that of C1. L4 (1.7 g, 6 mmol), n-BuLi(2.5 mL, 6 mmol) and
TiCl4 (0.66 mL, 6 mmol) were used to give deep red crystals of C4
(900 mg, 33%). mp: 159.7–162.1 ◦C. 1H NMR (400 MHz, CDCl3): ı
7.89 (d, J = 8.0 Hz, 1H, Ind-H), 7.67 (d, J = 8.0 Hz, 1H, Ind-H), 7.52
(s, 1H, Ph-H), 7.64–7.27 (m, 4H, 3Ph-H, Ind-H), 7.22–7.17 (m, 1H,
Ind-H), 7.13 (t, J = 7.2 Hz, 1H, Ind-H), 6.99–6.90 (m, 1H, Ind-H), 3.10
(d, J = 13.6 Hz, 1H, C5H10), 2.98 (d, J = 13.2 Hz, 1H, C5H10), 2.59 (t,
J = 12.8 Hz, 1H, C5H10), 2.2–1.8 (m, 3H, C5H10), 1.88 (s, 3H, PhCH3),
1.72 (d, J = 10.8 Hz, 2H, C5H10), 1.55–1.35 (m, 1H, C5H10), 1.20–1.05
1
(m, 12H, C9H14). 13C{ H} NMR (125 MHz, CDCl3): ı 150.6 (Ind-C),
145.9 (Ind-C), 145.3 (Ph-C), 143.8 (Ind-C), 129.1 (Ind-C), 127.9 (Ind-
C), 126.9 (Ind-C), 125.5 (Ind-C), 124.9 (Ph-C), 123.7 (Ph-C), 123.7
(Ph-C), 122.0 (Ind-C), 50.6 (Ind-C), 38.1 (C10H14-C), 37.3 (C10H14-C),
33.9 (C10H14-C), 33.4 (C10H14-C), 27.7 (C10H14-C), 27.3 (C10H14-C).
1
(m, 1H, C5H10). 13C{ H} NMR (125 MHz, CDCl3): ı 150.7 (Ind-C),
142.1 (Ind-C), 137.1 (Ph-C), 130.5 (Ind-C), 129.1 (Ind-C), 128.7 (Ind-
C), 127.9 (Ind-C), 127.3 (Ind-C), 127.0 (Ph-C), 125.7 (Ph-C), 125.3
(Ph-C), 123.8 (Ph-C), 123.5 (Ph-C), 120.6 (Ind-C), 114.3 (Ind-C),
44.9 (C5H10-C), 37.6 (C5H10-C), 35.7 (C5H10-C), 26.3 (C5H10-C), 23.3
(C5H10-C), 23.0 (C5H10-C), 22.4 (PhMe-C). MALDI-TOF-MS (m/z):
407.0 (M+−Cl). IR (cm−1): 2922, 2859, 1509, 1447, 1187, 1076, 736.
Anal. Calcd. for C22H23Cl3Ti·(0.11 C6H14): C, 60.33; H, 5.48; found:
C, 60.77; H, 5.94%.
[Ind-C(cyclo-C5H10)-(p-MePh)]TiCl3 (C5). The procedure was
similar to that of C1. L5 (1.7 g, 6 mmol), n-BuLi (2.5 mL, 6 mmol) and
TiCl4 (0.66 mL, 6 mmol) were used to give deep red crystals of C5
(1.6 g, 60%). mp: 146.7–148.1 ◦C. 1H NMR (400 MHz, CDCl3): ı 7.77
(d, J = 8.8 Hz, 1H, Ind-H), 7.61 (d, J = 8.8 Hz, 1H, Ind-H), 7.39–7.32 (m,
1H, Ind-H), 7.29–7.23 (m, 2H, Ph-H), 7.21 (d, J = 2.0 Hz, 1H, Ind-H),
7.12–6.99 (m, 4H, 2Ind-H, 2Ph-H), 2.99 (d, J = 13.6 Hz, 1H, C5H10),
2.64 (d, J = 7.2 Hz, 2H, C5H10), 2.24 (s, 3H, PhCH3), 2.15–2.05 (m,
2.2.2. Synthesis of titanium complexes C1–C8
Synthesis of [Ind-CMe2-Ph]TiCl3 (C1). To a solution of 0.84 g
(3.6 mmol) of L1 in 40 mL light petroleum ether, cooled to −78 ◦C
was added dropwise 1.5 mL (3.6 mmol) of n-BuLi. The reaction
mixture was warmed to r.t. and stirred overnight. Then 0.39 mL
(3.6 mmol) of TiCl4 was added to the mixture at −78 ◦C, the resul-
tant dark red solution was warmed to r.t. and stirred overnight. The
solvent and the unreacted TiCl4 were removed in vacuo. The residue
was extracted with light petroleum ether. After cooling the filtrate
to −20 ◦C, deep red crystals of C1 were obtained (423 mg, 30%). mp:
118.3–120.1 ◦C. 1H NMR (400 MHz, CDCl3): ı 7.74–7.69 (m, 1H, Ind-
H), 7.37–7.32 (m, 3H, Ph-H), 7.32–7.30 (m, 1H, Ind-H), 7.30–7.27
(m, 1H, Ph-H), 7.25–7.21 (m, 2H, Ind-H), 7.19–7.16 (m, 1H, Ph-H),
7.15–7.13 (m, 1H, Ind-H), 7.13–7.11 (m, 1H, Ind-H), 2.03 (s, 3H,
1
CH3), 1.95 (s, 3H, CH3). 13C{ H} NMR (125 MHz, CDCl3): ı 148.5

88
Y. Zhang et al. / Journal of Molecular Catalysis A: Chemical 373 (2013) 85–95
Table 1
1H, C5H10), 1.79–1.59 (m, 4H, C5H10), 1.43–1.29 (m, 2H, C5H10).
Crystallographic data for complexes C3 and C5.
1
13C{ H} NMR (125 MHz, CDCl3): 151.2 (Ind-C), 139.1 (Ind-C), 136.1
(Ph-C), 132.0 (Ind-C), 129.2 (Ind-C), 129.2 (Ind-C), 129.0 (Ind-C),
128.6 (Ind-C), 128.1 (Ph-C), 127.8 (Ph-C), 127.3 (Ph-C), 119.9 (Ind-
C), 115.0 (Ind-C), 45.7 (C5H10-C), 36.2 (C5H10-C), 34.9 (C5H10-C),
25.9 (C5H10-C), 22.6 (C5H10-C), 22.2 (C5H10-C), 20.8 (PhMe-C).
MALDI-TOF-MS (m/z): 407.0 (M+−Cl). IR (cm−1): 2922, 2860, 1529,
C3
C5
Formula
Formula wt
T (K)
Crystal system
Space group
C21H21Cl3Ti
427.63
293(2)
Monoclinic
P2(1)
7.843(4)
11.840(5)
11.463(5)
90
109.329(7)
90
1004.5(8)
2
1.414
C22H23Cl3Ti
441.65
293(2)
Triclinic
P-1
7.966(2)
10.293(3)
12.763(4)
85.394(4)
83.005(4)
85.127(4)
1032.4(5)
2
1447, 1259, 1018, 800. Anal. Calcd. for C22H23Cl Ti 0.3C H14: C,
·
3
6
˚
a (A)
61.15; H, 5.86; found: C, 61.53; H, 5.63.
˚
b (A)
[Ind-C(cyclo-C5H10)-(p-tBuPh)]TiCl3 (C6). The procedure was
similar to that of C1. L6 (1.2 g, 3.6 mmol), n-BuLi (1.5 mL, 3.6 mmol)
and TiCl4 (0.39 mL, 3.6 mmol) were used to give deep red crys-
tals of C6 (850 mg, 49%). mp: 144.6–146.7 ◦C. 1H NMR (400 MHz,
CDCl3): ı 7.82 (d, J = 8.8 Hz, 1H, Ind-H), 7.67 (d, J = 8.8 Hz, 1H, Ind-
H), 7.44–7.37 (m, 1H, Ind-H), 7.37–7.27 (m, 5H, 4Ph-H, Ind-H),
7.15 (d, J = 3.6 Hz, 1H, Ind-H), 7.07 (d, J = 3.6 Hz, 1H, Ind-H), 3.06
(d, J = 13.6 Hz, 1H, C5H10), 2.76–2.64 (m, 2H, C5H10), 2.25–2.08 (m,
˚
c (A)
˛ (◦)
ˇ (◦)
ꢀ (◦)
3
˚
V (A )
Z
ꢁcalcad (mg cm−3
)
1.421
0.806
456
ꢂ (mm−1
F (0 0 0)
)
0.826
440
ꢃrange (◦)
1.88–26.01
4555/3719
0.0555
0.9220, 0.8656
3719/1/227
0.0555
0.0412, 0.0541
0.0977, 0.0609
0.382, −0.246
1.61–26.01
4710/3950
0.0395
0.9608, 0.9310
3950/0/235
0.893
0.0396, 0.0945
0.0622, 0.1013
0.318, −0.279
1H, C5H10), 1.86–1.62 (m, 4H, C5H10), 1.35–1.19 [m, 11H, 2 C5H10
,
Reflections collected/unique
Rint
Max, min transmission
Data/restraints/parameters
Goodness of fit on F 2
R1, wR2 (I > 2ꢄ)
9C(CH3)3]. 13C{ H} NMR (125 MHz, CDCl3): ı = 151.3 (Ind-C), 149.2
(Ind-C), 138.9 (Ph-C), 132.0 (Ind-C), 129.2 (Ind-C), 128.9 (Ind-C),
128.5 (Ind-C), 127.8 (Ind-C), 127.6 (Ph-C), 127.4 (Ph-C), 125.3(Ph-
C), 119.9 (Ind-C), 115.0 (Ind-C), 45.6 (C5H10-C), 36.2 (C5H10-C),
34.8 (CMe3-C), 34.3 (C5H10-C), 31.2 (C5H10-C), 25.9 (C5H10-C), 22.6
(C5H10-C), 22.1 (CMe3-C). MALDI-TOF-MS (m/z): 449.2 (M+−Cl). IR
(cm−1): 2924, 2859, 1508, 1447, 1259, 1057, 875, 821, 800. 749.
Anal. Calcd. for C25H29Cl3Ti: C, 62.07; H, 6.04; found: C, 62.25; H,
6.41%.
1
R1, wR2 (all data)
Largest diff. peak and hole (e A
−3
˚
)
◦
˚
[Ind-C(cyclo-C5H10)-(p-FPh)]TiCl3 (C7). The procedure was sim-
ilar to that of C1. n-BuLi (2.5 mL, 6 mmol), L7 (1.7 g, 6 mmol) and
TiCl4 (0.66 mL, 6 mmol) were used to give deep red crystals of
C7 (983 mg, 37%). mp: 157.8–159.7 ◦C. 1H NMR (400 MHz, CDCl3):
ı 7.76 (d, J = 8.8 Hz, 1H, Ind-H), 7.68 (d, J = 8.8 Hz, 1H, Ind-H),
7.45–7.38 (m, 3H, 2Ind-H, Ph-H), 7.36–7.29 (m, 1H, Ph-H), 7.17 (d,
J = 3.6 Hz, 1H, Ph-H), 7.05 (d, J = 3.6 Hz, 1H, Ph-H), 7.04–6.98 (m, 2H,
Ind-H), 3.00 (d, J = 13.6 Hz, 1H, C5H10), 2.74–2.64 (m, 2H, C5H10),
2.18 (td, J = 13.6 Hz, J = 3.8 Hz, 1H, C5H10), 1.85–1.60 (m, 4H, C5H10),
(ꢅ = 0.71073 A). All data were collected at 20 C using the ␻-scan
techniques. The structures of C3 and C5 were solved by direct
method and refined using Fourier techniques. The absorption cor-
rection was applied based on SADABS [30]. Hydrogen atoms were
located and refined by the geometry method. All non-hydrogen
atoms were refined by full-matrix least-squares on F2 using the
SHELXTL program package [31]. The cell refinement, data collec-
tion, and reduction were done by Bruker SAINT [32]. The structure
determination and refinement were performed by SHELXS-97 [33]
and SHELXL-97 [34], respectively. The crystal and refinement data
are collected in Tables 1 and 2.
1
1.50–1.35 (m, 1H, C5H10), 1.17–1.30 (m, 1H, C5H10). 13C{ H} NMR
(125 MHz, CDCl3): ı 150.2 (Ind-C), 137.9 (Ind-C), 131.9 (Ph-C), 130.1
(Ind-C), 130.0 (Ind-C), 129.2 (Ind-C), 129.0 (Ind-C), 128.7 (Ph-C),
127.8 (Ph-C), 127.0 (Ph-C), 119.9 (Ph-C), 115.5 (Ph-C), 115.3 (Ind-C),
114.9 (Ind-C), 45.6 (C5H10-C), 36.5 (C5H10-C), 35.2 (C5H10-C), 25.9
(C5H10-C), 22.5 (C5H10-C), 22.1 (C5H10-C). MALDI-TOF-MS (m/z):
411.1 (M+−Cl). IR (cm−1): 2924, 2859, 1529, 1447, 1259, 1155, 1057,
876, 838, 800. 749, 736, 664. Anal. Calcd. for C21H20Cl3TiF: C, 56.60;
H, 4.52; found: C, 57.16; H, 4.73%.
Table 2
◦
˚
Selected bond lengths (A) and angles ( ) of complexes C3 and C5.
C3
C5
Ti(1) Cl(1)
Ti(1) Cl(2)
Ti(1) Cl(3)
Ti(1) C(1)
Ti(1) C(2)
Ti(1) C(3)
Ti(1) C(8)
Ti(1) C(9)
2.2394(17)
2.2210(17)
2.2145(15)
2.335(4)
2.275(5)
2.390(4)
2.2338(11)
2.2330(10)
2.2079(11)
2.427(2)
2.349(3)
2.288(3)
[Ind-C(cyclo-C9H14)-Ph]TiCl3 (C8). L8 (2.0 g, 6 mmol), n-BuLi
(2.5 mL, 6 mmol) and TiCl4 (0.66 mL, 6 mmol) were used to give
deep red crystals of C8 (150 mg, 5%). mp: 177.7–179.3 ◦C. 1H NMR
(400 MHz, CDCl3): ı = 8.35–8.25 (m, 1H, Ind-H), 7.88–6.91 (m, 10H,
5Ind-H, 5Ph-H), 3.65–3.15 (m, 2H, C9H14), 2.35–1.05 (m, 12H,
2.465(4)
2.416(4)
1
C9H14). 13C{ H} NMR (125 MHz, CDCl3): ı 150.6 (Ind-C), 146.0
Ti(1) C(4)
Ti(1) C(5)
2.370(2)
2.453(2)
(Ind-C), 129.1 (Ph-C), 127.9 (Ind-C), 127.0 (Ind-C), 126.7 (Ind-C),
126.2 (Ind-C), 124.8 (Ind-C), 123.9 (Ph-C), 122.5 (Ph-C), 122.0 (Ph-
C), 120.0 (Ind-C), 115.8 (Ind-C), 50.6 (C10H14-C), 49.1 (C10H14-C),
38.0 (C10H14-C), 33.9 (C10H14-C), 33.4 (C10H14-C), 28.8 (C10H14-C),
27.1 (C10H14-C), 25.2 (C10H14-C), 22.6 (C10H14-C), 19.4 (C10H14-C).
MALDI-TOF-MS (m/z): 445.2 (M+−Cl). IR (cm−1): 2923, 2858, 1508,
1447, 1259, 1155, 1057, 875, 837, 800. 749, 736, 670, 638. Anal.
Calcd. for C25H25Cl3Ti: C, 62.60; H, 5.25; found: C, 63.18; H, 5.79%.
Cl(2) Ti(1) Cl(3)
Cl(2) Ti(1) Cl(1)
Cl(3) Ti(1) Cl(1)
Cl(3) Ti(1) C(8)
Cl(1) Ti(1) C(8)
Cl(2) Ti(1) C(5)
C(2) Ti(1) C(8)
C(1) Ti(1) C(8)
C(3) Ti(1) C(8)
C(9) Ti(1) C(8)
C(3) Ti(1) C(5)
C(2) Ti(1) C(5)
C(4) Ti(1) C(5)
C(1) Ti(1) C(5)
C(9) C(10) C(16)
C(1) C(10) C(16)
102.25(6)
103.79(4)
102.14(4)
101.68(4)
101.36(6)
103.54(5)
85.22(12)
128.16(11)
85.60(7)
57.19(17)
56.32(16)
34.23(12)
34.23(13)
2.3. X-ray diffraction measurements
57.78(9)
56.71(9)
34.41(8)
34.33(8)
Single crystals of titanium complexes C3 and C5 were obtained
from saturated solutions of light petroleum ether at room temper-
ature. Diffraction data were collected on a Bruker SMART 1000 CCD
diffractometer using graphite monochromated Mo K␣ radiation
106.2(4)
105.4(2)

Y. Zhang et al. / Journal of Molecular Catalysis A: Chemical 373 (2013) 85–95
89
2.4. General procedure for ethylene trimerization
A 100 mL autoclave equipped with a magnetic stirrer was heated
under vacuum for 1 h over 100 ◦C, then allowed to cool to the
required reaction temperature under ethylene atmosphere, and
charged with 20 mL of toluene. The autoclave was pressurized to
required ethylene pressure. Stirring for about 30 min ensured that
the solution was equilibrated and the required reaction temper-
ature was established. An appropriate amount of MAO and the
catalyst in toluene were added. After that the reactor was pres-
surized with ethylene to reach the required value to start the
polymerization. The ethylene pressure was kept constant during
the polymerization. After the reaction was carried out for 0.5 h, the
reactor was quitted, and when the temperature was cooled to room
temperature, the pressure was released and 1 mL of ethanol was
injected immediately to quench the reaction. A certain amount of
heptane was added as the internal standard. An aliquot of upper-
layer clear solution was withdrawn from the reaction mixture to
analyze the content of oligomers by GC and GC/MS. The precipi-
tated polymer was filtered, washed with acidified ethanol (3% HCl),
and dried under vacuum at 60 ◦C to a constant weight.
T i
C l
C l
T i
C l
C l
C l
C l
C 9
C 1 0
Chart 2. The structures of C9 and C10.
just similar to the corresponding cyclopentadienyl complexes [36].
The arene group is above the indenyl plane and far away from the
metal center. The dihedral angle between the arene group and the
indenyl plane depends on the bridge unit between the arene group
and indenyl.
3. Results and discussions
As listed in Table 2, the Ti C bond lengths of five-member ring
in these complexes are uneven. The common character is that the
Ti C distance to the substituted carbon is longer than the other
Ti C(H) distances, but the longest Ti C distance is the one where
the carbon atom is adjacent to the substituted one and belongs to
the six-member ring. For the corresponding cyclopentadienyl com-
plexes, such Ti C bond lengths are almost equal [36]. It suggests
that the bulky substituent on the 3-position of indenyl ring leads
to the unsymmetrical bonding of Ti atom to the carbons of the five-
member ring in these complexes significantly. In addition, because
of the electrophilic property of the phenyl in the indenyl ring, the
average Ti C bond lengths are longer than those in the correspond-
ing cyclopentadienyl complexes [36], indicating that the indenyl-Ti
bonding is weaker than cyclopentadienyl-Ti bonding.
3.1. Synthesis of the complexes C1–C8
The half-sandwich titanium complexes (C1–C8) bearing various
pendant arene groups have been synthesized. The synthetic route
is summarized in Scheme 1.
According to the literature [29], the substituted indenes were
prepared from the addition reaction of benzofulvenes with appro-
priate aryl lithiums. The resultant lithium salts of indenes were
optionally quenched with trimethylsilyl chloride to obtain the cor-
responding trimethylsilyl reagents, or hydrolyzed with H2O to
get the substituted indenes. Generally speaking, half-sandwich
cyclopentadienyl or indenyl titanium complexes are synthesized
via the reaction of the corresponding trimethylsilyl reagents
with TiCl4 [11,35]; while we found that it was hard to obtain
the target titanium complexes by reacting the corresponding
trimethylsilyl reagents with TiCl4 (see supporting information). The
half-sandwich titanium complexes C1–C8 were successfully syn-
thesized through the reaction of the more active lithium salts of
indene ligands with TiCl4 in 1:1 molar ratio. Due to the presence
of unreacted benzofulvenes or other impurities, the lithium salt
of indenes formed in the first step were not used in the reaction
with TiCl4 directly, instead further hydrolysis and chromatographic
purification were carried out to afford the indenes L1–L8 in pure
form, which were deprotonated with n-butyl lithium to give the
corresponding lithium salts.
To make a comparison with complexes C1–C8, titanium com-
plexes [Ind-CH2-Ph]TiCl3 (C9) and [Cp-C(cyclo-C5H10)-Ph]TiCl3
(C10) were also synthesized according to the literature procedures
(Chart 2) [35,11].
3.2. Crystal structures of C3 and C5
Single crystals of [Ind-C(cyclo-C5H10)-Ph]TiCl3 (C3) and [Ind-
C(cyclo-C5H10)-(p-MePh]TiCl3 (C5) were obtained by slow evapo-
ration of the saturated petroleum ether solutions. Crystallographic
data and results of the refinements for complexes C3 and C5 are
summarized in Table 1, selected bond lengths and angles are listed
in Table 2. The CCDC Nos. are 896661 and 896659, respectively.
As shown in Figs. 1 and 2, both complexes show the familiar
piano-stool geometry. There are no indication of either intra- or
intermolecular interaction between the arene and the metal center,
Fig. 1. X-ray crystal structure of C3 (C21H21Cl3Ti). Thermal ellipsoids are shown at
the 30% probability level.

90
Y. Zhang et al. / Journal of Molecular Catalysis A: Chemical 373 (2013) 85–95
Scheme 1. The synthetic route of C1–C8.
3.3. Ethylene trimerization studies
Indenyl titanium complex C3 shows almost the same trimer-
ization selectivity as the corresponding cyclopentadienyl complex
C10, but is more active than C10. Comparing with the half-sandwich
Cp complex C10, the corresponding indenyl complex C3 has a
different property of steric and electronic effects. The electron
density of ␲-system is delocalized in the whole indenyl ring, and
the electron-withdrawing effect of phenyl moiety makes the five-
member ring of indenyl lack of electron density relatively, which
weakens the bonding between indenyl and the metal center. As
also indicated by the X-ray diffraction studies, the average Ti-C
bond lengths in indenyl complexes are longer than those in the
corresponding cyclopentadienyl complexes. Therefore, upon acti-
vation with MAO, the resultant metal cation bearing indenyl ring is
more electrophilic, which is more favorable for the coordination of
the pendant arene and ethylene to the metal center, leading to an
enhanced activity of indenyl complex C3 than C10 toward ethylene
trimerization.
3.3.1. Effect of indenyl ring versus Cp ring on ethylene
trimerization
In order to check the activity and selectivity of these indenyl
titanium complexes for ethylene trimerization, the ethylene con-
version experiments at 0 ◦C, 0.5 MPa of ethylene pressure, using
indenyl titanium complex C3 and the corresponding cyclopenta-
dienyl complex C10 activated by 1000 equivalents of MAO were
tested firstly. The results are listed in Table 3.
As expected, the C3/MAO system was found to be highly active
and selective for ethylene trimerization under the tested condi-
tions. The products of ethylene conversion consist of two fractions:
the liquid oligomers as the major fraction and a few amount of
polyethylene. The liquid fraction as analyzed by GC–MS and GC
mainly contains ethylene trimers with the content of 1-hexene in
C6 > 99.9%, and a trace amount of tetramer or higher oligomers
(denoted as C8+ oligomers). No 1-butene could be detected by
GC, which agrees well with the result of extensive DFT calcula-
tions by Blok and co-workers based on [Cp-CH2-Ph]TiCl3 system
[13] that in a trimerization process forming 1-butene needs more
free energy barriers than forming 1-hexene and other intermedi-
ates.
3.3.2. Effect of the structure features of indenyl titanium
complexes on ethylene trimerization
Based on the above results, ethylene trimerization experiments
catalyzed by indenyl titanium complexes C1–C9 with MAO as
cocatalyst were studied in detail. It is found that the activity and
Table 3
Ethylene trimerization catalyzed by titanium complexes with different ␲-system/MAO.a
Ti
Ti
Cl
Cl
Cl
Cl
Cl
Cl
C3
C10
.
Cat.
1-Hexene (mg)
C8+ (mg)
PE (mg)
Productivity of 1-hexeneb
Selectivity of 1-hexene (wt.%)c
C3
C10
1042
765
17
15
39
20
1389
1020
94.9
95.6
a
Reaction conditions: 20 mL toluene as solvent; 30 min run time; 1.5 ␮mol Cat.; 0.5 MPa of ethylene; 0 ◦C; Al/Ti = 1000.
In kg of 1-hexene/(mol-Ti h).
Percentage of 1-hexene in overall by mass.
b
c

Y. Zhang et al. / Journal of Molecular Catalysis A: Chemical 373 (2013) 85–95
91
unit might tend to push the arene group toward the metal center,
therefore leading to a stronger coordination interaction.
Complex C9 has the smallest CH2 bridge unit and supposed
to have the biggest ␣ angle in the resultant species. The pen-
dant arene group is very flexible, which would lead to a greatly
weak coordination with the cationic metal center. In line with
this, C9 displays very low activity and poor selectivity for ethy-
lene trimerization to 1-hexene, which also agrees with the results
of [Cp-CH2-Ph]TiCl3/MAO system that a CH2-bridge is believed to
be inefficient for an intramolecular metal-arene interaction [15].
When the bridge group is CMe2, which effectively pushes the
pendant phenyl to coordinate with the cationic metal center, the
related complex C1 displays high trimerization activity of 1053 kg
of 1-hexene/(mol-Ti h) and a selectivity of 95.3% for 1-hexene. As
expected, C3 with a cyclohexyl as the bridging unit, performs better
than C2 in ethylene trimerization to 1-hexene, obtaining an activity
of 1389 kg of 1-hexene/(mol-Ti h).
The only exception is that complex C8, possessing the biggest
adamantyl as the bridging unit and supposed to have the small-
est ␣ angle in the resultant cationic species upon activation with
MAO, shows very poor activity for the production of 1-hexene,
although with a relatively high 1-hexene selectivity. Hessen found
that in the rate-determining step of a trimerization process, accom-
panying with one ethylene molecular being captured, the loosing
of metal-arene interaction is needed [15], therefore a very strong
metal-arene interaction might not be favorable for the rate-
determining step and would lead to a decline of the activity. The
low activity of C8 is probably because of the big adamantyl as the
bridge making the metal-arene interaction too strong.
Fig. 2. X-ray crystal structure of C5 (C22H23Cl3Ti). Thermal ellipsoids are shown at
the 30% probability level.
Simiar phenomenon was also observed in our previous het-
eroatom systems (Chart 4), where the pendant arene group was
replaced with thienyl group. It was found that the pendant thienyl
system could also be highly selective for ethylene trimerization
[20]. The selectivity for 1-hexene is more than 95%, but the activity
is much lower than that in Hessen’s arene system, most likely due
to a stronger coordination interaction of sulfur atom to the cationic
metal center than the phenyl ring. When a more coordinating oxy-
gen atom is introduced [37], the corresponding catalyst is even not
active at low temperature, but does show a very low activity at high
temperatures; the selectivity of 1-hexene is still more than 95%, just
as thienyl and arene system.
Thus it can be concluded that, for the half-sandwich [Cp’-(B)-
(pendant group)]TiCl3 system: if the pendant group is too weak to
coordinate with the metal cation, both the activity and selectiv-
ity for ethylene trimerization would be poor just as C9; and if the
pendant group has too strong coordination with the cationic metal
center, the activity of the catalyst will decline greatly, but it will
still show high selectivity for 1-hexene just as C8.
selectivity for ethylene trimerization are greatly influenced by the
structure of titanium complexes and the reaction conditions (tem-
perature, ethylene pressure, Al/Ti ratio, etc.). Among the factors
influencing the activity and selectivity, the structure features of
titanium complexes, such as the bridging unit between indenyl and
pendant arene group, and the substituents of the arene group, play
a crucial and profound role.
3.3.2.1. Effect of the bridge between indenyl and pendant arene group.
The results of ethylene trimerization by complexes C1, C3, C8 and
C9 with different bridges between indenyl and phenyl are shown in
Table 4. Similar to the reported cyclopentadienyl titanium systems
[10,11], the bridging unit also exerts great influence on the catalytic
performance of these complexes.
When activated with MAO, C9 with the smallest CH2 bridge pro-
duces 1-hexene and polyethylene in comparable amount, together
with some higher oligomers (C8+) at a very low reaction rate. Com-
plexes C1 (CMe2), C3 [C(cyclo-C5H10)] and C8 [C(cyclo-C9H14)] all
display great tendency for ethylene trimerization to produce 1-
hexene, with the activities of C1 and C3 significantly higher than
that of C8. Thus, the order according to the size of the bridge is
C9 (CH2) < C1 (CMe2) < C3 [C(cyclo-C5H10)] < C8 [C(cyclo-C9H14)];
while, the order of the productivity of 1-hexene is C9 (CH2) < C8
[C(cyclo-C9H14)] ꢀ C1 (CMe2) < C3 [C(cyclo-C5H10)], which illus-
trates the effect of different bridges on ethylene trimerization.
In a trimerization process, when these indenyl titanium com-
plexes are activated with MAO, the hemilabile coordination of the
pendant arene group to cationic metal center is usually presumed
to be responsible [10]. The steric bulkiness of the bridging unit
between indenyl and pendant arene plays a crucial role in deter-
mining the orientation of the arene group, as indicated by the angle
of C(Ind)-C(bridge)-C(arene), named ␣ in Chart 3. Although no exact
␣ data in the cation of titanium complexes, it is deducible that
the greater the steric bulkiness of the bridge, the smaller the ␣
angle will be. It is considerable that the more steric bulky bridging
3.3.2.2. Effect of the substituent of the pendant arene group. The sub-
stituents on the pendant arene group can influence the catalytic
process with both steric and electronic effect.
As shown in Table 5, by comparing the reaction runs with
complexes C3–C7 that all have a cyclohexyl bridging unit as the
catalysts, the introduction of a substituent of either electron donat-
ing or electron withdrawing leads to a significant decrease of
the catalytic activity for ethylene trimerization without excep-
tion. Especially, complex C4 bearing a ortho-methyl substituted
phenyl, displays the lowest activity of 34 kg of 1-hexene/(mol-Ti h)
and a dramatic loss of the selectivity for ethylene trimerization. In
Hessen’s Cp system, when 4-methylphenyl or 3,5-dimethylphenyl
were introduced as the pendant arene group, the trimerization
activity also declined [11]. They suggested that increasing the ␲-
basicity of the arene by introducing methyl group would make the
metal-arene interaction too strong and this is not favored in the
rate-determining step of a trimerization process [15]. In our case, if

92
Y. Zhang et al. / Journal of Molecular Catalysis A: Chemical 373 (2013) 85–95
Table 4
Ethylene trimerization catalyzed by indenyl titanium complexes with different bridge groups/MAO systems. a
Ti
Cl
Ti
Cl
Ti
Cl
Ti
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
C1
C3
C8
C9
.
Cat.
1-Hexene (mg)
C8+ (mg)
PE (mg)
Productivity of 1-hexeneb
Selectivity of 1-hexene (wt.%)c
C1
C3
C8
C9
790
1042
91
15
17
5
24
39
11
22
1053
1389
121
31
95.3
94.9
85.0
24.5
23
49
a
Reaction conditions: 20 mL toluene as solvent; 30 min run time; 1.5 ␮mol Cat.; 0.5 MPa of ethylene; 0 ◦C; Al/Ti = 1000.
In kg of 1-hexene/(mol-Ti h).
Percentage of 1-hexene in overall by mass.
b
c
a
a
a
a
H3C Ti
C9*
31
H3C Ti
H3C Ti
C3*
1389
H3C Ti
C8*
121
C1*
1053
Productivity:
Kg/((mol-Ti·h))
* : cooresponding cation species after activated by MAO
Chart 3. Different angles of C(Ind)-C(bridge)-C(arene) in titanium complexes versus the productivity.
only taking the electronic effect into account, the activities of C4 (o-
Me) and C5 (p-Me) should be almost the same. Actually, C5 is much
more active than C4. It is believed that the steric hindrance of the
pendant arene moiety should be the main reason in this case. The
steric effect of the methyl on the ortho-position of the phenyl blocks
the metal-arene interaction, resulting in a significantly decline of
both activity and selectivity of C4.
The steric effect of alkyl substituent on the para-position of
phenyl ring is not so significant as that of ortho-substituent.
Although a decrease of trimerization activity was observed for
C5 (p-Me), C6 (p-tBu), with the activity tendency of C3 (H) < C5
(p-Me) < C6 (p-tBu), the selectivity of these complexes for ethylene
trimerization is dropped only slightly, indicating that a different
situation should be involved when compared to that of C4.
Complex C6 bearing a strong electron-donating “tBu” sub-
stituent on the para-position of phenyl, shows lower activity than
C3 (H), which is however consistent with Hessen’s observation
that the increased ␲-basicity of the arene group will prompt a
stronger coordination interaction with the metal cation and thus
make the activity declined. It is clear that, although tert-butyl
has a big size, it is on the para-position of the arene and far
away from the bridging unit; in this case, the electron-donating
S
Ti
Cl
Cl
O
Ti
Cl
Ti
Cl
Cl
Cl
Cl
Cl
Cl
.
.
.
Trace
57 Kg/(mol-Ti h)
>95%
Productivity of C6:
1020 Kg/(mol-Ti h)
>95%
120 Kg/(mol-Ti h)
>95%
Selectivity of 1-hexene:
Al/Ti = 1000, 0.5 MPa ethylene, 0 oC
80 oC
Chart 4. The influence on ethylene trimerization by titanium complexes with different pendant coordination groups.

Y. Zhang et al. / Journal of Molecular Catalysis A: Chemical 373 (2013) 85–95
93
Table 5
Ethylene trimerization catalyzed by indenyl titanium complexes with different substituent on the arene moiety/MAO systems. a
Ti
Ti
Cl
Cl
Cl
Cl
Cl
Cl
C1
C2
F
Ti
Ti
Ti
Cl
Ti
Cl
Ti
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
C3
C4
C5
C6
C7
.
Entry
Cat.
P (MPa)
1-Hexene (mg)
C8+ (mg)
PE (mg)
Productivity of 1-hexeneb
Selectivity of 1-hexene (wt.%)c
1
2
3
4
5
6
7
8
9
C1
C1
C2
C2
C3
C3
C4
C5
C5
C6
C7
0.5
0.8
0.5
0.8
0.5
0.8
0.5
0.5
0.8
0.5
0.5
790
914
798
976
1042
1476
26
591
655
161
389
15
14
25
23
17
20
7
46
19
7
24
60
20
38
39
42
80
17
18
11
36
1053
1219
1064
1301
1389
1968
34
788
874
214
518
95.3
92.5
94.6
94.1
94.9
95.9
23.5
90.4
94.9
89.9
82.6
10
11
46
a
Reaction conditions: 20 mL toluene as solvent; 30 min run time; 1.5 ␮mol Cat.; 0.5 MPa of ethylene; 0 ◦C; Al/Ti = 1000.
In kg of 1-hexene/(mol-Ti h).
Percentage of 1-hexene in overall by mass.
b
c
effect plays the dominant role. The dominant electronic-donating
Increasing the reaction temperature, both the productivity and
selectivity for 1-hexene decreased. On the one hand, it might be
the result of gradual degradation of the active trimerization species
at higher temperature. Most important, it is believed that the cru-
cial hemilable coordination interaction of the pendant arene group
with the cationic metal center for ethylene trimerization is less
favored at higher temperature [35,38], and relatively more active
effect of para-alkyl substituent could also be deduced by compar-
ing the ethylene trimerization results of complexes C5 and C2,
where C2 with smaller “CMe2” bridge is more active than C5 with
“C[cyclo-(C5H10)]” as bridge. The positive electron donating effect
of para-methyl in C2 offsets the weakening of the “arene-metal”
coordination caused by reducing the bridge size.
When a strong electron-withdrawing fluoro substituent is intro-
duced to the para-position of the pendant phenyl in C7, the activity
of C7 is not only lower than that of C3, but also lower than that of C5.
Comparing with complexes C5, C6 having para-alkyl substituent,
C7 shows evidently declined selectivity for ethylene trimerization.
It is therefore considerable that the electron-withdrawing fluoro
group might make the arene lack of electron density and weaken
the interaction between the arene group and the cationic metal
center significantly, and thus leading to a decrease of both activity
and selectivity for trimerization.
3.3.3. The influence of different reaction conditions
The ethylene trimerization processes catalyzed by indenyl com-
plexes/MAO systems were also tested under different reaction
conditions (temperature, ethylene pressure, concentrations of cat-
alyst, and the amount of co-catalyst MAO). It was found that the
activity and selectivity for ethylene trimerization to 1-hexene were
greatly influenced by these conditions especially by temperature.
Figs. 3 and 4 show the influence of the reaction temperature
on the activity and selectivity of C1–C10/MAO for ethylene trimer-
ization to 1-hexene. It can be seen that the reaction temperature
exerts significant effect on the behavior of these catalyst systems.
Fig. 3. The influence of temperature on productivity of 1-hexene catalyzed by C1-
C10. Reaction conditions: 20 mL toluene as solvent; 30 min run time; Al/Ti = 1000;
1.5 ␮mol Cat.; 0.5 MPa ethylene.

94
Y. Zhang et al. / Journal of Molecular Catalysis A: Chemical 373 (2013) 85–95
Table 6
The influence of different amount of catalyst on ethylene trimerization by C3/MAO system.a
Cat.
MCat. (␮mol)
1-Hexene (mg)
C8+ (mg)
PE (mg)
Productivity of 1-hexeneb
Selectivity of 1-hexene (wt.%)c
C3
C3
C3
3.0
1.5
0.75
1566
1042
196
9
17
14
18
39
3
1044
1389
523
94.7
94.9
92.0
a
Reaction conditions: 20 mL toluene as solvent; 30 min run time; Al/Ti = 1000; 0.5 MPa of ethylene; 0 ◦C.
In kg of 1-hexene/(mol-Ti h).
Percentage of 1-hexene in overall by mass.
b
c
Table 7
The influence of different ratio of Al/Ti on ethylene trimerization by C3/MAO system.a
Cat.
Al/Ti
1-Hexene (mg)
C8+ (mg)
PE (mg)
Productivity of 1-hexeneb
Selectivity of 1-hexene (wt.%)c
C3
C3
C3
C3
2000
1000
500
904
1042
651
9
17
14
9
18
39
15
14
1205
1389
868
97.1
94.9
95.7
95.9
200
554
739
a
Reaction conditions: 20 mL toluene as solvent; 30 min run time; 1.5 ␮mol of catalyst; 0.5 MPa of ethylene; 0 ◦C.
In kg of 1-hexene/(mol-Ti h).
Percentage of 1-hexene in overall by mass.
b
c
species accounting for polymerization are formed which lead to
While halving the amount of catalyst to 0.75 ␮mol, the mass of
1-hexene decreases dramatically, which makes the activity of 1-
hexene decrease to 523 kg/(mol-Ti h). The overall selectivities for
1-hexene fraction under these conditions however are hardly influ-
enced by changing the concentration of titanium catalyst.
The activity is not only influenced by the concentrations of
ethylene and catalyst but also depends upon the amount of the
co-catalyst. Table 7 shows the effect of different Al:Ti ratio on the
trimerization process. The trimerization activity increases with the
increase of the Al:Ti ratio up to 1000, due to an easily reduction
of the metal site at higher concentration of MAO. When the ratio
of Al:Ti is 2000, the activity slightly declines, probably due to the
absorption of co-catalyst on the catalyst center in competition with
the monomer [39]. The selectivity of 1-hexene, however, is not
influenced evidently with changing the ratio of Al:Ti.
the decrease of both productivity and selectivity for 1-hexene. Due
to the very weak nature of such arene-metal interaction, similar
to cyclopentadienyl analog, most of these indenyl titanium com-
plexes show higher activity and selectivity at low temperature of
0 ◦C. The only exception is C6, which gives the best activity at 30 ◦C,
probably the introduction of tert-butyl as electro-donating group
slightly enhances the arene-metal coordination, favorable for the
stabilization of the active trimerization species at relatively higher
temperature.
Generally, ethylene pressure plays a positive role on trimeriza-
tion activity. As shown in Table 5, comparing with the catalytic
behavior at 0.5 MPa (Entries 1, 3, 5 and 8), higher productivities
could be obtained with increasing the ethylene pressure to 0.8 MPa
(Entries 2, 4, 6 and 9). Although no practical kinetic studies are
undertaken, this increasing tread is just agree with that of corre-
sponding Cp complexes [11], probably due to the higher ethylene
solubility under higher ethylene pressure.
4. Conclusions
Table 6 shows the effect of catalyst amount on the activity and
selectivity for ethylene trimerization with C3 as the catalyst. It
is obvious that the productivity of 1-hexene increases with dou-
bling the amount of catalyst to 3.0 ␮mol accompanied by a slight
decrease of other fractions. The activity however declines slightly.
We successfully synthesized a series of half-sandwich indenyl
titanium complexes relevant to selective ethylene trimerization.
The crystal structures of complexes C3 and C5 show no intramolec-
ular coordination of the pendant arene group and the metal center
in the solid state. Upon activation with MAO indenyl complex C3
is more active than the corresponding cyclopentadienyl complex
C10 under otherwise the identical experiment conditions. For these
indenyl titanium complexes, the activity and selectivity for ethy-
lene trimerization are greatly influenced by the substituent of the
pendant arene group and the bridge between the indenyl and the
arene moiety. Among them, C3 is the most proper structure favor-
able for the trimerization of ethylene, and the combination of a
cyclohexyl bridge and a pendant phenyl group in C3 may afford the
most suitable hemilabile property accounting for a trimerization
process. The other complexes, assumed to have an “arene-metal
coordination” of either too strong or too weak, show inferior
catalytic performance than C3. It is suggested that a too weak
“arene-metal coordination” in the cationic species may lead to a
decrease of both the activity and selectivity for ethylene trimeriza-
tion; and a too strong coordination may only lead to a reasonable
decrease of the activity, and hardly influencing the selectivity for
1-hexene. Moreover, the catalytic system has a great sensitivity to
reaction temperature. Both the activity and selectivity for ethylene
trimerization to 1-hexene decrease with the increase of the reaction
temperature.
Fig. 4. The influence of temperature on selectivity for ethylene trimerization to 1-
hexene catalyzed by C1–C10. Reaction conditions: 20 mL toluene as solvent; 30 min
run time; Al/Ti = 1000; 1.5 ␮mol Cat.; 0.5 MPa ethylene.

Y. Zhang et al. / Journal of Molecular Catalysis A: Chemical 373 (2013) 85–95
95
Acknowledgments
[12] H. Hagen, W.P. Kretschmer, F.R.V. Buren, B. Hessen, D.A.V. Oeffelen, J. Mol. Catal.
A: Chem. 248 (2006) 237–247.
[13] A.N.J. Blok, P.H.M. Budzelaar, A.W. Gal, Organometallics 22 (2003) 2564–2570.
[14] S. Tobisch, T. Ziegler, Organometallics 22 (2003) 5392–5405.
[15] B. Hessen, J. Mol. Catal. A: Chem. 213 (2004) 129–135.
[16] H. Hagen, Ind. Eng. Chem. Res. 45 (2006) 3544–3551.
[17] M. Horacek, R. Gyepes, J. Merna, J. Kubista, K. Mach, J. Pinkas, J. Organomet.
Chem. 695 (2010) 1425–1433.
[18] J.A. Suttil, D.S. McGuinness, S.J. Evans, Dalton Trans. 39 (2010) 5278–5285.
[19] Y. Kim, J. Park, S. Yoon, S.B. Park, Transit. Met. Chem. 37 (2012) 439–444.
[20] J. Huang, T. Wu, Y. Qian, Chem. Commun. 39 (2003) 2816–2817.
[21] Y. Suzuki, S. Kinoshita, A. Shibahara, S. Ishii, K. Kawamura, Y. Inoue, T. Fujita,
Organometallics 29 (2010) 2394–2396.
This work is subsidized by the National Basic Research Program
of China (2005CB623801), National Natural Science Foundation
of China (20774027, 21274041), and Key Project of Chinese Min-
istry of Education (109064). All the financial supports are gratefully
acknowledged.
Appendix A. Supplementary data
[22] E. Otten, A.A. Batinas, A. Meetsma, B. Hessen, J. Am. Chem. Soc. 131 (2009)
5298–5312.
[23] H. Toulhoat, M.L. Fomena, T. Bruin, J. Am. Chem. Soc. 133 (2011) 2481–2491.
[24] H.G. Alt, M. Jung, G. Kehr, J. Organomet. Chem. 562 (1998) 153–181.
[25] M. Diekmann, G. Bockstiegel, A. Lützen, M. Friedemann, W. Saak, D. Haase, R.
Beckhaus, Organometallics 25 (2006) 339–348.
[26] P. Beak, G.W. Selling, J. Org. Chem. 54 (1989) 5574–5580.
[27] S. Schweizer, J.M. Becht, C.L. Drian, Tetrahydron 66 (2010) 765–772.
[28] M.D.P. Creapo, T.D. Avery, E. Hanssen, E. Fox, T.V. Robinson, P. Valente, D.K.
Taylor, L. Tilley, Antimicrob. Agents Chemother. 52 (2008) 98–109.
[29] H.G. Alt, A. Reb, K. Kundu, J. Organomet. Chem. 628 (2001) 211–221.
[30] SADABS, Bruker Nonius Area Detector Scaling and Absorption Correction-
V2.05, Bruker AXS Inc., Madison, WI, 1996.
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/
j.molcata.2013.03.005.
References
[1] I.A. Hussein, T. Hameed, B.F.A. Sharkh, K. Mezghani, Polymer 44 (2003)
4665–4672.
[2] I.A. Hussein, T. Hameed, J. Appl. Polym. Sci. 97 (2005) 2488–2498.
[3] D.S. McGuinness, P. Wasserscheid, W. Keim, C. Hu, U. Englert, J.T. Dixon, C.
Grove, Chem. Commun. 39 (2003) 334–335.
[31] G.M. Sheldrick, SHELXTL 5.10 for Windows NT, Structure Determination Soft-
ware Programs, Bruker Analytical X-ray Systems, Inc., Madison, WI, 1997.
[32] SAINT, Version 6.02, Bruker AXS Inc., Madison, WI, 1999.
[33] G.M. Sheldrick, SHELXS-97, Program for the Solution of Crystal Structures, Uni-
versity of Gottingen, Germany, 1990.
[34] G.M. Sheldrick, SHELXL-97, Program for the Refinement of Crystal Structures,
University of Gottingen, Germany, 1997.
[35] P. Foster, J.C.W. Chien, M.D. Rausch, Organometallics 15 (1996) 2404–2409.
[36] J. Sa␤mannshausen, A.K. Powell, C.E. Anson, S. Wocadlo, M. Bochmann, J.
Organomet. Chem. 592 (1999) 84–94.
[4] D.S. McGuinness, P. Wasserscheid, W. Keim, D. Morgan, J.T. Dixon, A. Bollmann,
H. Maumela, F. Hess, U. Englert, J. Am. Chem. Soc. 125 (2003) 5272–5273.
[5] W. Müller, A. Wc¸ hl, S. Peitz, N. Peulecke, B.R. Aluri, B.H. Müller, D. Heller, U.
Rosentha, M.H. Al-Hazmi, F.M. Mosa, ChemCatChem 2 (2010) 1130–1142.
[6] S. Licciulli, K. Albahily, V. Fomitcheva, I. Korobkov, S. Gambarotta, R. Duchateau,
Angew. Chem. Int. Ed. 50 (2011) 2346–2349.
[7] R.A. Muller, H. Tsurugi, T. Saito, M. Yanagawa, S. Oda, K. Mashima, J. Am. Chem.
Soc. 131 (2009) 5370–5371.
[8] T. Agapie, Coord. Chem. Rev. 255 (2011) 861–880.
[9] D.S. McGuinness, Chem. Rev. 111 (2011) 2321–2341.
[10] P.J.W. Deckers, B. Hessen, J.H. Teuben, Angew. Chem. Int. Ed. 40 (2001)
2516–2519.
[11] P.J.W. Deckers, B. Hessen, J.H. Teuben, Organometallics 21 (2002)
5122–5135.
[37] T. Wu, Y. Qian, J. Huang, J. Mol. Catal. A: Chem. 214 (2004) 227–229.
[38] J.C. Flores, J.S. Wood, J.C.W. Chien, M.D. Rausch, Organometallics 15 (1996)
4944–4950.
[39] E. Ahmadi, Z. Mohamadnia, Catal. Lett. 141 (2011) 1191–1198.