Group IV complexes of an amine bis(phenolate) ligand featuring a THF sidearm donor: From highly active to living polymerization catalysts of 1-hexene

Article abstract of DOI:10.1016/S0020-1693(02)01306-3

A new amine bis(phenolate) ligand bearing a THF donor on a sidearm and several dialkyl complexes of the group IV triad are introduced. The crystal structures of a zirconium dibenzyl complex and a titanium dibenzyl complex were solved, and revealed a strong binding of the oxygen donor of the THF group to the metal center. Upon activation with tris(pentafluorophenyl)borane the zirconium and the hafnium dibenzyl complexes lead to highly active 1-hexene polymerization catalysts. In comparison, titanium complexes (a dibenzyl complex and a dimethyl complex) lead to slower but living polymerization of 1-hexene at room temperature. Polymerization of neat 1-hexene under high dilution conditions was found to be living for an exceptionally long time of 6 days, leading to poly(1-hexene) of M_w=816 000 and PDI=1.09. A block copolymerization of 1-hexene and 1-octene at room temperature could be obtained using the titanium catalysts.

Full text of DOI:10.1016/S0020-1693(02)01306-3

Inorganica Chimica Acta 345 (2003) 137ꢁ144
/
www.elsevier.com/locate/ica
Group IV complexes of an amine bis(phenolate) ligand
featuring a THF sidearm donor:
from highly active to living polymerization catalysts of 1-hexene
Stanislav Groysman a, Israel Goldberg a, Moshe Kol a,*, Elisheva Genizi b,
Zeev Goldschmidt b,*
a School of Chemistry, Raymond and Beverly Sackler Faculty of Exact Sciences, Tel Aviv University, Tel Aviv 69978, Israel
b Department of Chemistry, Bar-Ilan University, Ramat-Gan 52900, Israel
Received 31 May 2002
Dedicated in honor of Professor Richard R. Schrock
Abstract
A new amine bis(phenolate) ligand bearing a THF donor on a sidearm and several dialkyl complexes of the group IV triad are
introduced. The crystal structures of a zirconium dibenzyl complex and a titanium dibenzyl complex were solved, and revealed a
strong binding of the oxygen donor of the THF group to the metal center. Upon activation with tris(pentafluorophenyl)borane the
zirconium and the hafnium dibenzyl complexes lead to highly active 1-hexene polymerization catalysts. In comparison, titanium
complexes (a dibenzyl complex and a dimethyl complex) lead to slower but living polymerization of 1-hexene at room temperature.
Polymerization of neat 1-hexene under high dilution conditions was found to be living for an exceptionally long time of 6 days,
leading to poly(1-hexene) of Mwꢀ
/
816 000 and PDIꢀ1.09. A block copolymerization of 1-hexene and 1-octene at room
/
temperature could be obtained using the titanium catalysts.
# 2002 Elsevier Science B.V. All rights reserved.
Keywords: Titanium; Zirconium; Living polymerization; 1-Hexene; Copolymerization
1. Introduction
and syndiotactic polymerization of propylene at room
temperature [7], and the extremely active catalysts for
polymerization of ethylene [8ꢁ11] employing imine
phenolate ligands. These studies show that by careful
design of the ligand system, a close control of the
activity of the catalyst and the properties of the polymer
may be achieved.
In recent years there has been a broad interest in the
developing of cyclopentadienyl-free catalysts based on
/
group IV metals for polymerization of alpha olefins [1ꢁ
/
3]. This interest was inspired, to a large extent, by the
pioneering work of McConville [4] and Schrock [5] that
reported for the first time a living polymerization of
high-olefins at ambient temperatures employing dia-
mido-type ligands. Several other important develop-
ments that followed include the living and isotactic
polymerization of high-olefins at room temperature
employing diamine diphenolate ligands [6], the living
Recently we introduced the amine bis(phenolate)
ligand family for group IV metals [12ꢁ17]. We found
/
that for zirconium and hafnium, the presence of an extra
donor on a sidearm was crucial for obtaining active
polymerization catalysts, as complexes lacking it led to
rapid catalyst deactivation [13,14]. The reactivity of
these precatalysts could be tuned by varying the nature
of this sidearm donor, and the most active zirconium
catalyst was found to have a methoxy donor on the
sidearm [15]. The corresponding titanium complexes
exhibited a different mode of reactivity. Whereas the
* Corresponding authors. Tel.: ꢂ
9293 (M.K.); tel.: ꢂ972-3-531 8307; fax: ꢂ
E-mail addresses: moshekol@post.tau.ac.il
goldz@mail.biu.ac.il (Z. Goldschmidt).
/
972-3-640 7392; fax: ꢂ
972-3-535 1250 (Z.G.)
(M. Kol),
/
972-3-640
/
/
0020-1693/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 0 2 0 - 1 6 9 3 ( 0 2 ) 0 1 3 0 6 - 3

138
S. Groysman et al. / Inorganica Chimica Acta 345 (2003) 137ꢁ144
/
sidearm donor was not a prerequisite for an active
catalyst, its presence led to living polymerization at
room temperature [16]. Especially, a titanium complex
of an amine bis(phenolate) ligand having a methoxy
donor on the sidearm led to living polymerization of 1-
hexene for an exceptionally long time of 31 h yielding a
polymer of Mw of 450 000 and PDI of 1.12 [17]. These
results led us to design a new amine bis(phenolate)
ligand having a different oxygen donor: a THF group.
In this work we describe the synthesis of dialkyl
complexes of zirconium, hafnium, and titanium of this
new ligand and their reactivity in polymerization of 1-
hexene.
ware [20], and were refined by full-matrix least-squares
with SHELXL-97 [21]. The crystals contain partly dis-
ordered solvent and partly disordered tert-butyl groups.
Some of the benzyl groups also exhibit large-amplitude
wagging motion. These give rise to a relatively large
fraction of weak reflections below the intensity thresh-
old, and affect to some extent the precision of the
structure determination.
2.2. Synthesis of the ligand precursor (LigH2)
A solution of 2,4-di-tert-butylphenol (6.0 g, 29
mmol), 2-(aminomethyl)tetrahydrofuran (1.0 ml, 9.7
mmol) and 36% aqueous formaldehyde (1.62 ml, 19.4
mmol) in methanol (10 ml) was stirred and refluxed for
40 h. The mixture was cooled and the supernatant
solution decanted. The solid residue was triturated with
ice cold methanol, filtered and washed thoroughly with
cold methanol, to give the ligand precursor as a colorless
powder (3.48 g, 67% yield), which could be further
2. Experimental
2.1. General
All experiments were performed under an atmosphere
of dry nitrogen in a nitrogen-filled glovebox. Ether and
tetrahydrofuran were purified by reflux and distillation
under dry argon atmosphere from Na/benzophenone.
Pentane was washed with HNO3/H2SO4 prior to dis-
tillation from Na/benzophenone/tetraglyme. Toluene
was refluxed over Na and distilled. Benzyl magnesium
chloride, 1-hexene, 1-octene, and chlorobenzene were
purchased from Aldrich Inc. Tris(pentafluorophe-
nyl)borane was obtained from Strem Chemicals Inc. 1-
Hexene and chlorobenzene were passed through alu-
mina prior to use. Zirconium tetrabenzyl, hafnium
tetrabenzyl, and titanium tetrabenzyl were synthesized
according to published procedures [18,19]. Titanium
tetrabenzyl was used shortly after its synthesis. NMR
data were recorded on Bruker AC-200 and Bruker
Avance AC-400 spectrometers and referenced to protio
impurities in benzene-d6 (d 7.15) or TMS in CDCl3 (d
0.00). Elemental analyses were performed in the micro-
analytical laboratory in the Hebrew University of
Jerusalem. The metal complexes were analyzed within
a few hours of being taken out of the freezer of the glove
box. Poly(1-hexene) molecular weights were determined
by gel permeation chromatography (GPC) using TSKgel
GMHHR-M, TSKgel G 4000 HHR, and TSKgel G
5000 HHR columns set on a Jasco instrument equipped
with a refractive index detector. Molecular weight
determination was carried out relative to polystyrene
standards using tetrahydrofuran (HPLC grade, distilled
and filtered under vacuum prior to use) as the eluting
solvent. The X-ray diffraction measurements were
purified by recrystallization from etherꢁ
/
methanol. M.p.
160ꢁ161 8C. To the crude oil which remained after
/
removal of the solvent, was added a small amount of
methanol. The mixture was left standing at room
temperature for 2 months, during which more of the
product precipitated, increasing the yield to a total of
90%. C35H55NO3 (537.7): Calc. for C, 78.16; H, 10.31;
N, 2.60. Found: C, 78.16; H, 10.29; N, 2.84%. 1H NMR
(CDCl3): d 8.86 (bs, 2H), 7.20 (d, Jꢀ/2 Hz, 2H), 6.88 (d,
Jꢀ2 Hz, 2H), 4.27 (m, 1H), 4.01 (m, 1H), 3.89 (m, 1H),
/
3.77 (m, 4H), 2.61 (m, 1H), 2.51 (m, 1H), 1.90 (m, 3H),
1.57 (m, 1H), 1.40 (s, 18H), 1.27 (s, 18H). 13C NMR
(CDCl3): d 153.02 (C), 140.63 (C), 136.05 (C), 124.95
(CH), 123.38 (CH), 121.43 (C), 77.54 (CH), 68.28
(CH2), 57.52 (CH2), 55.93 (CH2), 34.99 (CMe3) 34.10
(CMe3), 31.67 (CH3), 29.60 (CH3), 29.60 (CH2), 25.21
(CH2). HRMS (DCIꢁCH4): 537.418854 (C35H55 NO3)
/
(Mꢂ).
2.3. Synthesis of 1a
LigH2 (60 mg, 0.11 mmol) was dissolved in approxi-
mately 1 ml of toluene and added dropwise to a bright
yellow solution of ZrBn4 (51 mg, 0.11 mmol) in toluene
at room temperature. The solution was stirred at room
temperature for 2 h. The solvent was removed in vacuo,
and the resulting yellow solid was washed with pentane
(ca. 2 ml). The final yield was 70 mg (78%). Single
crystals of 1a were grown from toluene (see Table 1 for
crystallographic data). C49H67NO3Zr (809.3): Calc. C,
performed on a Nonius Kappa CCD diffractometer
˚
72.72; H, 8.35; N, 1.73. Found: C, 71.99; H, 8.12; N,
1
system, using Mo Ka (lꢀ
/
0.7107 A radiation). The
1.82%. H NMR (C6D6, 200 MHz) d 7.79 (d, Jꢀ
/
7.2
7.6 Hz,
7.4 Hz, 2H), 6.58 (t, Jꢀ
13.8 Hz, 2H), 3.27 (m, 3H), 2.89
(s, 2H), 2.65 (m, 2H), 2.61 (s, 2H), 2.08 (t, 2H), 1.79 (s,
analyzed crystals were embedded within a drop of
viscous oil and freeze-cooled to approximately 110 K.
The structures were solved by a combination of direct
methods and Fourier techniques using the SIR-92 soft-
Hz, 2H), 7.59 (t, Jꢀ
2H), 7.02 (m, 5H), 6.77 (t, Jꢀ
7.0 Hz, 1H), 3.5 (t, Jꢀ
/
2.4 Hz, 2H), 7.40 (t, Jꢀ
/
/
/
/

S. Groysman et al. / Inorganica Chimica Acta 345 (2003) 137ꢁ
/144
139
Table 1
Crystallographic data for 1a and 3a
(C), 140.8 (C), 137.4 (C), 137.1 (C), 128.8 (CH), 128.4
(CH), 128.0 (CH), 127.9 (CH), 125.4 (CH), 125.3 (CH),
124.9 (CH), 124.5 (CH), 124.6 (CH), 121.9 (CH), 120.8
(CH), 80.9 (CH), 75.0 (CH2), 71.6 (CH2), 70.8 (CH2),
64.7 (CH2), 64.3 (CH2), 56.7 (CH2), 35.7 (C), 35.6 (C),
34.5 (C), 32.0 (CH3), 30.6 (CH3), 30.3 (CH3), 28.5
(CH2), 25.2 (CH2).
3a×2C7H8
1a×2C7H8
Empirical formula
Formula weight
Temperature (K)
Crystal system
Space group
C63H83NO3Ti
950.20
C63H83NO3Zr
993.52
110
110
monoclinic
P21/n
monoclinic
P21/n
˚
2.5. Synthesis of 3a
a (A)
14.8410(2)
15.3280(2)
25.3630(6)
105.3950(6)
5562.62(17)
4
14.6110(2)
15.3350(2)
25.8920(5)
103.9060(6)
5631.32(15)
4
˚
b (A)
˚
c (A)
LigH2 (92 mg, 0.17 mmol) was dissolved in approxi-
mately 1 ml of toluene at room temperature and added
dropwise to a stirring dark red solution of TiBn4 in
approximately 1 ml of toluene. The color of the solution
remained dark red and it was stirred for 2 h at room
temperature. The solvent was removed under reduced
b (8)
3
V (A )
˚
Z
Dcalc (g cmꢃ1
)
1.135
0.7107
1.172
0.7107
˚
l (A)
m (mmꢃ1
Reflections measured
Independent reflections 7074
)
0.198
12 237
0.239
13 451
8302
pressure yielding a redꢁbrown solid, which was washed
/
with approximately 2 ml of pentane and dried in vacuo.
The final yield was 98 mg (75%). Crystals suitable for X-
2u Range (8)
Dataset
3.3ꢁ
0ꢁ/19/0ꢁ
/
55.8
16/ꢃ33ꢁ
5.5ꢁ
/
56.5
/
/
32 0ꢁ19/0ꢁ
/
/
20/ꢃ33ꢁ33
/
R1 (I ꢀ2s(I))
wR2 (I ꢀ2s(I))
R1 (all data)
0.0900
0.2112
0.1604
0.2522
1.015
0.0713
0.1684
0.1322
0.2044
1.049
ray analysis were grown from tolueneꢁ
/
pentane mixture
at ꢃ35 8C (see Table 1 for crystallographic data).
/
C49H67NO3Ti (757.9): Calc. C, 76.83; H, 8.82; N, 1.83.
1
Found: C, 76.55; H, 8.64; N, 1.93%. H NMR (C6D6,
wR2 (all data)
Goodness-of-fit (F2)
200 MHz) d 7.94 (d, Jꢀ
(d, Jꢀ7.8 Hz, 2H), 7.00 (m, 5H), 6.66 (t, Jꢀ
2H), 6.48 (t, Jꢀ7.2 Hz, 2H), 3.56 (m, 2H, CH2), 3.28
/
7.6 Hz, 2H), 7.72 (bs, 2H), 7.42
/
/7.1 Hz,
18H), 1.34 (s, 18H), 0.60 (m, 3H), ꢃ
0.01 (m, 1H). 13C
/
/
(m, 3H), 3.01 (m, 4H, PhCH2-Ti), 2.51 (dd, 2H, CH2),
2.20 (s, 18H, t-Bu), 1.86 (m, 2H), 1.37 (s, 18H, t-Bu),
NMR (C6D6, 50.29 MHz) d 158.9 (C-O), 158.2 (C-O),
148.5 (C), 147.5 (C), 141.1 (C), 149.8 (C), 136.8 (C),
136.5 (C), 129.2 (CH), 128.6 (CH), 128.5 (CH), 128.2
(CH), 128.0 (CH), 125.6 (CH), 124.7 (CH), 124.6 (CH),
124.3 (CH), 122.4 (CH), 120.7 (CH), 80.5 (CH), 70.4
(CH2), 68.3 (CH2), 65.2 (CH2), 65.0 (CH2), 64.6 (CH2),
56.7 (CH3), 35.8 (C), 35.6 (C), 34.5 (C), 32.0 (CH3),
30.6 (CH3), 30.3 (CH3), 28.7 (CH2), 25.1 (CH2).
0.55 (m, 3H), ꢃ
0.05 (m, 1H). 13C NMR (C6D6, 50.29
/
MHz) d 161.43 (C-O), 160.80 (C-O), 154.94 (C), 141.45
(C), 141.31 (C), 136.34 (C), 129.73 (CH), 129.27 (CH),
128.51 (CH), 128.40 (CH), 128.12 (CH), 128.01 (CH),
126.90 (C), 126.84 (CH), 126.05 (C), 125.64 (CH),
125.19 (CH), 124.39 (CH), 124.22 (CH), 124.09 (CH),
122.18 (CH), 121.93 (CH), 99.26 (CH2), 84.54 (CH2),
78.54 (CH), 68.29 (CH2), 65.50 (CH2), 65.00 (CH2),
56.82 (CH2), 36.08 (C), 35.97 (C), 34.41 (C), 31.91
(CH3), 31.13 (CH3), 30.84 (CH3), 28.52 (CH2), 24.55
(CH2).
2.4. Synthesis of 2a
LigH2 (149 mg, 0.28 mmol) was dissolved in toluene
and added dropwise to a bright yellow solution of
HfBn4 (150 mg, 0.28 mmol) in 6 ml of toluene. The
yellow color disappeared immediately and the solution
became colorless. The solution was stirred at room
temperature for two hours, and the solvent was removed
under reduced pressure. The product was washed once
with approximately 2 ml of pentane and dried in vacuo
to give 2a as a white solid. The final yield was 209 mg
(83%). C49H67NO3Hf (896.5): Calc. C, 65.64; H, 7.53;
2.6. Synthesis of 3b
50 mg (90 mmol) of LigH2 and 19 mg (0.19 mmol) of
Et3N were dissolved in 2 ml of THF and added dropwise
at room temperature to 31 mg (90 mmol) of TiCl4(THF)2
in 3 ml of THF. The reaction solution turned red slowly
and was stirred at room temperature for 2 h. The
solution was filtered through Celite (in order to remove
1
N, 1.56. Found C, 65.42; H, 7.61; N, 1.74%. H NMR
Et3N×
/
HCl) and cooled to ꢃ33 8C. MeMgBr (60 ml, 1.0
/
(C6D6, 200 MHz) d 7.73 (d, Jꢀ
2.2 Hz, 2H), 7.42 (t, Jꢀ7.7 Hz, 2H), 6.95 (m, 5H), 6.82
(t, Jꢀ7.7 Hz, 2H), 6.59 (m, 1H), 3.54 (m, 2H), 3.29 (m,
/
7.2 Hz, 2H), 7.63 (t, Jꢀ
/
M in diethyl ether) was added to the reaction vial, and
the solution turned brown immediately. The reaction
was allowed to stir for 20 min, after which it was
evaporated to dryness. 3b was extracted from the crude
product by pentane (ca. 1 ml). The final yield was 28 mg
(50%). C37H59NO3Ti (613.7): 3b gave low carbon
percentage due to its extreme thermal and hydrolytic
/
/
3H), 2.66 (m, 2H), 2.64 (s, 2H, PhCH2), 2.50 (s, 2H,
PhCH2), 2.10 (m, 2H), 1.88 (s, 18H), 1.35 (s, 18H), 0.57
(m, 3H), ꢃ
/
0.02 (m, 1H). 13C NMR (C6D6, 50.29 MHz)
d 158.9 (C-O), 158.2 (C-O), 149.2 (C), 148.9 (C), 140.9

140
S. Groysman et al. / Inorganica Chimica Acta 345 (2003) 137ꢁ144
/
sensitivity (Calc. C, 72.40; H, 9.69; N, 2.28. Found: C,
70.02; H, 9.42; N, 2.27%). 1H NMR (C6D6, 200 MHz) d
7.65 (bs, 2H), 6.99 (bs, 2H), 3.59 (m, 2H), 3.13 (m, 3H),
2.61 (m, 2H), 1.99 (s, 9H), 1.97 (s, 9H), 1.75 (s, 3H), 1.48
added to the polymerization vial after the complete
consumption of 1-hexene. The duration of the whole
experiment was approximately 30 h, after which a
complete consumption of 1-octene was observed.
(s, 3H), 1.42 (s, 9H), 1.41(s, 9H), 0.62 (m, 3H), ꢃ0.05
/
(m, 1H). 13C NMR (C6D6, 50.29 MHz) d 161.61 (C-O),
160.61 (C-O), 140.76 (C), 136.35 (C), 135.93 (C), 128.73
(C-H), 79.08 (CH), 69.12 (CH2), 64.65 (CH2), 64.09
(CH2), 62.49 (CH2), 61.39 (CH3-Ti), 56.91 (CH3-Ti),
35.93 (C), 35.78 (C), 34.59 (C), 32.16 (CH3), 30.49
(CH3), 30.11 (CH3), 28.89 (CH2), 25.15 (CH2).
3. Results and discussion
3.1. Catalyst design through ligand design
The ligand precursor (Scheme 1) was prepared by a
single step Mannich condensation between 2,4-di-tert-
butylphenol, formaldehyde, and the corresponding
amine (see Section 2 for details). The tetrahydrofuran
ring is connected through the 2? position to the central
nitrogen atom by a methylene bridge. The increased
2.7. Polymerization of neat 1-hexene by 1aꢁ2a
/
B(C6F5)3 (15 mol) was dissolved in approximately 1
ml of 1-hexene and added to a stirred solution of 1a (2a)
(12 mmol) in 1-hexene. The total volume of the monomer
polarity of THF relative to ether (1.75 vs. Â1.2 D)
/
was 20 ml. The resulting mixture was stirred for 20ꢁ40
/
should promote a stronger binding of the THF oxygen
to the metal atom. The rigidity of the THF group
relative to the O-Me group should also contribute to its
stronger coordination to the metal. The bulky tert-butyl
groups ortho to the phenolic oxygen are required for the
clean chemistry in the case of the large metal ions (Zr,
Hf) [14]. The resulting dibenzyl octahedral complexes
should possess C1 symmetry in contrast to all previously
reported amine bis(phenolate) complexes (featuring Cs
symmetry on the NMR timescale) due to an asymmetric
carbon atom. The low symmetry of this ligand may
direct the approaching monomer thus leading to a
defined stereochemistry of the resulting polymer if it is
expressed in the catalytic site (‘enanthiomorphic site
control’).
min until 1-hexene had started to boil. The remaining
olefin was evaporated in vacuo yielding poly(1-hexene)
as a colorless sticky oil. Samples of poly(1-hexene) were
analyzed by GPC. 13C NMR (CDCl3) d 40.39 (br,
CH2), 34.40 (br, CH2), 32.41 (CH), 29.87 (br, CH2),
23.41 (CH2), 14.35 (CH3).
2.8. Polymerization of 1-hexene diluted in chlorobenzene
by 1aꢁ2a
/
A solution of B(C6F5)3 (15 mmol) in 1-hexene (1 ml)
and chlorobenzene (4 ml) was added dropwise to a
solution of 1a (2a) (12 mmol) in 4 ml of 1-hexene and 16
ml of chlorobenzene. In a typical experiment the
polymerization process was monitored by taking mea-
sured aliquots over a period of 20 min. The samples
were evaporated to dryness, weighed and analyzed by
GPC.
2.9. Typical polymerization of neat 1-hexene by 3aꢁ3b
/
Scheme 1. The amine bis(phenolate) THF ligand.
B(C6F5)3 (ca. 15 mmol) was dissolved in 1 ml of 1-
hexene and added to a stirred olefin solution of 3a (3b)
(ca. 12 mmol). A typical volume of 1-hexene was in range
3.2. Synthesis and structure
of 40ꢁ500 ml. The polymerization process was followed
/
The desired dibenzyl complexes of the group IV triad
were prepared by protonolysis of a readily available
precursor (M(CH2Ph)4) with the ligand precursor
(Scheme 2). A different synthetic strategy had to be
developed for preparing the dimethyl complex 3b.
Reacting TiCl4(THF)2 with the ligand precursor in the
presence of NEt3 led to a clean mixture of two products
that were not fully characterized [22,23] but were further
reacted with 2 equiv. of MeMgBr (Scheme 3). The
desired product was extracted with pentane giving a
by GPC.
2.10. Block copolymerization of 1-hexene/1-octene by 3a
16 mg (31 mmol) of B(C6F5)3 were dissolved in 0.43 ml
of chlorobenzene and added to 20 mg (26 mmol) of 3a in
1.00 ml of chlorobenzene. The reaction solution turned
slowly yellow and was allowed to stir at room tempera-
ture for 1 h. 426 mg (5.2 mmol, 200 equiv.) of 1-hexene
in 1.43 ml of chlorobenzene were added to the reaction
vial, and the polymerization process was monitored by
GPC. The second monomer (1-octene, 396 mg, 136
equiv.) was dissolved in 2.50 ml of chlorobenzene and
yellowꢁorange solid, in ca. 50% yield.
/
As previously mentioned, the low symmetry of this
ligand must induce a low symmetry of the complex (C1).

S. Groysman et al. / Inorganica Chimica Acta 345 (2003) 137ꢁ
/144
141
Scheme 2. Synthesis of the dibenzyl complexes.
Scheme 3. Synthesis of the titanium dimethyl complex.
This could be confirmed by four different benzyl
protons and four different aromatic protons (observed
as multiplets), as well as by two AB systems for the
ligand methylenes (adjacent to the phenolate rings) for
both of the titanium complexes (3a and 3b). Interest-
ingly, 3a and 3b differ from each other by the number of
dissimilar tert-butyl groups: two for 3a and four for 3b.
The zirconium and the hafnium complexes exhibit
apparent ‘higher’ symmetry than the titanium com-
plexes: the benzylic protons are observed as sharp
singlets, and only two different tert-butyl groups could
be identified. Hence, it may be concluded, that the
influence of the formally low symmetry on the catalytic
site is rather small, especially for the case of the larger
metal atoms.
Fig. 2. Structure of 3a (50% probability ellipsoids). H-atoms and
solvent molecules are omitted for clarity. Key atoms are labeled.
Table 2
˚
Selected bond distances (A) and angles (8) for 1a
Bond distances
ZrÃO2
ZrÃO3
ZrÃO4
ZrÃN5
ZrÃC6
ZrÃC7
1.992(2)
1.998(2)
2.371(3)
2.444(3)
2.296(4)
2.297(4)
Bond angles
O2ÃZrÃO3
O4ÃZrÃN5
C6ÃZrÃC7
ZrÃC6ÃC8
ZrÃC7ÃC14
159.09(10)
69.36(9)
1a and 3a were recrystallized from cold toluene and
their crystal structures were solved. ORTEP views of 1a
and 3a are shown in Figs. 1 and 2, respectively. The
crystallographic data, key bond distances and angles are
101.42(14)
114.63(2)
105.80(3)
listed in Tables 1ꢁ3.
/
Table 3
˚
The two X-ray structures indicate that the ‘ligand
core’ remains symmetrical despite the formal absence of
a mirror plane in the molecule. The molecules are of a
distorted octahedral geometry, in which the two phe-
nolic oxygens are in trans configuration, and the two
benzyl groups are in cis configuration. According to our
Selected bond distances (A) and angles (8) for 3a
Bond distances
TiÃO2
TiÃO3
TiÃO4
TiÃN5
TiÃC6
TiÃC7
1.870(2)
1.880(2)
2.264(2)
2.359(3)
2.149(4)
2.197(3)
expectations, the TiÃ
/
O (THF donor) bond distance of
Bond angles
O2ÃTiÃO3
O4ÃTiÃN5
C6ÃTiÃC7
TiÃC6ÃC8
TiÃC7ÃC14
163.15(10)
72.94(9)
90.40(14)
109.67(2)
117.00(3)
˚
2.26 A is considerably shorter than the TiÃ
/
O bond
length of 2.42 A found for a methoxyalkyl sidearm
˚
donor [17]. The same trend was observed for the ZrÃ
/
O
˚
˚
donor bond of 2.37 A in 1a versus 2.45 A found for a
related zirconium compound with the same methox-
yalkyl sidearm [15]. The structural similarity between
the cores of these complexes and those of previously
reported complexes of that ligand family enables us to
attribute different reactivities to minor changes in the
ligand framework.
Fig. 1. Structure of 1a (50% probability ellipsoids). H-atoms and
solvent molecules are omitted for clarity. Key atoms are labeled. The
elongated ellipsoids indicate partly disordered atoms. The disorder of
the lower left tert-butyl group could not be resolved.

142
S. Groysman et al. / Inorganica Chimica Acta 345 (2003) 137ꢁ144
/
Table 4
Polymerization data for 1a/B(C6F5)3 and 2a/B(C6F5)3
Activitya
Mw
PDI
b
Precatalyst/solvent
1a/neat
1a/ClPh
2a/neat
2a/ClPh
21 000
2000
4000
1000
80
118
59
3.0
2.1
1.9
1.5
60
a
ꢃ1
cat
Measured in g mmol
103 g molꢃ1
hꢃ1
.
b
Fig. 4. Living polymerization of 1-hexene by 3a/B(C6F5)3: dependence
of Mw on the polymer weight.
.
3.3. Polymerization studies
hꢃ1. At the early stages of the polymer-
3 h) the PDI indices appear to be rather high
ꢃ1
12 g mmol
cat
ization (2ꢁ
/
For metallocenes, zirconium and hafnium catalysts
are generally more active than the respective titanium
systems [24]. For non-metallocene systems this tendency
does not appear to be general. Nevertheless, the amine
bis(phenolate) systems follow the same trend. Upon
activation with approximately 1.2 equiv. of B(C6F5)3, 1a
was found to be highly active toward polymerization of
(1.13ꢁ1.15) for a living process, but as the polymeriza-
/
tion proceeds (Fig. 3), there is a substantial lowering of
the PDI values (1.05). The lowering of the polydisper-
sity, the linear dependence of molecular weight on time
(Fig. 3), and the linear increase in molecular weight
versus polymer produced (Fig. 4) all suggest that the
polymerization is living and that its rate is constant
under these conditions. When the initial amount of 1-
hexene is very high (more than 50 000 equiv.), the
polymerization remains living for an exceptionally
long time: the 3a/B(C6F5)3 system catalyzes 1-hexene
polymerization at room temperature in a living fashion
for up to 48 h: after that time the resulting polymer has a
molecular weight of 316 000 and a PDI value as low as
1.05 (Figs. 3 and 4).
The relatively high polydispersitiy at the early stages
of the polymerization may result from slow activation of
the precatalyst (abstraction of the benzyl group by the
borane reagent). To try to improve the catalyst’s
performance at the early stage of the reaction we
prepared 3b. Adding B(C6F5)3 to 3b in 1-hexene resulted
in an immediate color change from dark red to bright
yellow, apparently signifying a rapid activation of the
precatalyst. As for 3a, the polymerization of 1-hexene
proceeded in a living manner, however, the initial PDI
values were not reduced substantially. When high
dilution conditions were employed (200 000 equiv.), the
polymerization continued for 6 days giving poly(1-
hexene) of extremely high molecular weight. The final
polymer had a molecular weight of 816 000 and a PDI
value of 1.09 (see Fig. 5 for details). We therefore
1-hexene. The highest activity was obtained in neat
hꢃ1) and was accompanied by
ꢃ1
cat
olefin (21 000 g mmol
boiling of the monomer. Diluting of 1-hexene in
chlorobenzene reduced the excess heat release, thereby
slowing various termination processes. Consequently, 1-
hexene was consumed totally, and a higher molecular
weight and lower PDI value were obtained (see Table 4
for details). The PDI values are around 2.0 in this case,
signifying that 1a/B(C6F5)3 is a single site catalyst.
The hafnium-based catalysts are generally slightly less
active than their zirconium analogues [25]. Activation of
2a with approximately 1.2 equiv. of B(C6F5)3 resulted in
the boiling of the monomer. As expected, the reaction
was slower than in the case of 1a, and the calculated
hꢃ1. When diluted in
ꢃ1
cat
activity was 4000 g mmol
chlorobenzene, complex 2a led to poly(1-hexene) having
significantly lower PDI values, without affecting the
molecular weight appreciably (Table 4).
Complex 3a was also found to be an active precatalyst
for 1-hexene polymerization. Addition of B(C6F5)3 to a
dark red solution of 3a in neat 1-hexene resulted in a
slow change of color to bright yellow. We propose that
the color change results from the precatalyst’s activa-
tion. The titanium-based catalyst is a moderate poly-
merization catalyst, and its activity was evaluated to be
Fig. 3. Living polymerization of 1-hexene by 3a/B(C6F5)3: dependence
of Mw on time.
Fig. 5. ‘Immortal’ polymerization by 3b/B(C6F5)3: dependence of
molecular weight (Mw) on time.

S. Groysman et al. / Inorganica Chimica Acta 345 (2003) 137ꢁ
/144
143
816 000 (PDIꢀ1.09). To the best of our knowledge, this
/
is the highest molecular weight ever obtained for
polymerization of 1-hexene in the living fashion.
4. Supplementary material
Crystal structures of 1a and 2a in a cif format are
available from the authors upon request.
Fig. 6. Block copolymerization of 1-hexene and 1-octene by 3a/
B(C6F5)3.
propose that the behavior of the catalyst is almost
unaffected by the nature of original alkyl groups in a
precatalyst, and that the polymerization in this system is
practically ‘immortal’. We attribute the eventual termi-
nation of the living polymerization process to technical
sources, such as increased solution’s viscosity or water/
oxygen traces in the glove box atmosphere.
Acknowledgements
This research was supported by the Ministry of
Science, Culture and Sports and was also supported in
part by the Israel Science Foundation founded by the
Israel Academy of Sciences and Humanities. We thank
Dr. Edit Y. Tshuva for helpful discussions.
The main application of living polymerization is the
synthesis of well-defined block copolymers. A necessary
condition for successful block copolymerization is the
total consumption of the first monomer. The system
should also be stable enough so as not to decompose
prior to addition of the second monomer. In order to
achieve a maximum activation of the precatalyst, the
catalyst was preconditioned, i.e. 3a and B(C6F5)3 were
mixed in chlorobenzene for 1 h after which 200 equiv. of
1-hexene were added (Fig. 6). After the complete
consumption of the first monomer, the resulting poly-
mer had a molecular weight (Mw) of 22 000 and a PDI
value of 1.16. The higher than expected molecular
weight (17 000) could arise from a partial decomposition
of the catalyst prior to 1-hexene addition or from
incomplete catalyst activation. 1-Octene was added to
the reaction mixture after the complete consumption of
the first olefin, and immediate increase in molecular
weight could be noticed (Fig. 6). The rise of Mw was
accompanied by retention of the polydispersity index
References
[1] G.J.P. Britovsek, V.C. Gibson, D.F. Wass, Angew. Chem., Int.
Ed. Engl. 38 (1999) 428.
[2] Y. Imanishi, N. Naga, Prog. Polym. Sci. 26 (2001) 1147.
[3] G.W. Coates, J. Chem. Soc., Dalton Trans. (2002) 467.
[4] J.D. Scollard, D.H. McConville, J. Am. Chem. Soc. 118 (1996)
10008.
[5] R. Baumann, W.M. Davis, R.R. Schrock, J. Am. Chem. Soc. 119
(1997) 3830.
[6] E.Y. Tshuva, I. Goldberg, M. Kol, J. Am. Chem. Soc. 122 (2000)
10706.
[7] J. Tian, P.D. Hustad, G.W. Coates, J. Am. Chem. Soc. 123 (2001)
5134.
[8] S.-i. Ishii, J. Saito, M. Mitani, J.-i. Mohri, N. Matsukawa, Y.
Tohi, S. Matsui, N. Kashiwa, T. Fujita, J. Mol. Catal., A Chem.
179 (2002) 11.
[9] S. Matsui, M. Mitani, J. Saito, Y. Tohi, H. Makio, N.
Matsukawa, Y. Takagi, K. Tsuru, M. Nitabaru, T. Nakano, H.
Tanaka, N. Kashiwa, T. Fujita, J. Am. Chem. Soc. 123 (2001)
6847.
(PDIꢀ1.16). After full consumption of the second
/
[10] M. Mitani, J.-i. Mohri, Y. Yoshida, J. Saito, S. Ishii, K. Tsuru, S.
Matsui, R. Furuyama, T. Nakano, H. Tanaka, S.-i. Kojoh, T.
Matsugi, N. Kashiwa, T. Fujita, J. Am. Chem. Soc. 124 (2002)
3327.
monomer, the resulting block poly((1-hexene)-(1-octene)
had an Mw of 34 000 (all actions were performed at
room temperature).
[11] J. Saito, M. Mitani, J.-i. Mohri, Y. Yoshida, S. Matsui, S.-i. Ishii,
S.-i. Kojoh, N. Kashiwa, T. Fujita, Angew. Chem., Int. Ed. Engl.
40 (2001) 291.
Despite the formal C1 symmetry of the complexes, the
polymers obtained by all of the catalysts were atactic,
according to 13C NMR [6].
[12] E.Y. Tshuva, M. Versano, I. Goldberg, M. Kol, H. Weitman, Z.
Goldschmidt, Inorg. Chem. Commun. 2 (1999) 371.
[13] E.Y. Tshuva, I. Goldberg, M. Kol, H. Weitman, Z. Goldschmidt,
Chem. Commun. (2000) 379.
In summary, we introduced a new amine bis(pheno-
late) ligand, featuring a THF donor on a sidearm.
Dibenzyl complexes of the group IV triad with the
ligand were obtained in a one-step reaction, and a
dimethyl titanium complex was prepared in a simple
‘one-pot’ two-step procedure. The zirconium and the
hafnium complexes showed an outstanding activity in
high-olefin polymerization proving to be single-site
catalysts. The titanium complexes promote an ‘immor-
tal’ polymerization of neat 1-hexene at room tempera-
ture that led to a polymer having a molecular weight of
[14] E.Y. Tshuva, I. Goldberg, M. Kol, Z. Goldschmidt, Organome-
tallics 20 (2001) 3017.
[15] E.Y. Tshuva, S. Groysman, I. Goldberg, M. Kol, Z. Gold-
schmidt, Organometallics 21 (2002) 662.
[16] E.Y. Tshuva, I. Goldberg, M. Kol, Z. Goldschmidt, Inorg. Chem.
Commun. 3 (2000) 611.
[17] E.Y. Tshuva, I. Goldberg, M. Kol, Z. Goldschmidt, Chem.
Commun. (2001) 2120.
[18] U. Zucchini, E. Alizzati, U. Giannini, J. Organomet. Chem. 26
(1971) 357.

144
S. Groysman et al. / Inorganica Chimica Acta 345 (2003) 137ꢁ144
/
[19] A.van der Linden, C.J. Schaverien, N. Neijboom, C. Ganter, A.G.
Orpen, J. Am. Chem. Soc. 117 (1995) 3008.
[22] We suggest that the ‘‘dichloro’’ mixture consists of two C1
symmetry products resulting from different wrapping of the
ligand around the metal. E.Y. Tshuva, unpublished results.
[23] T. Toupance, S.R. Dubberly, N.H. Rees, B.R. Tyrrell, P.
Mountford, Organometallics 21 (2002) 1367.
[20] A. Altomare, G. Cascarano, C. Giacovazzo, A. Guagliardi, M.C.
Burla, G. Polidori, M. Camali, J. Appl. Crystallogr. 27 (1994)
435.
[21] G.M. Sheldrick, SHELXL-97 Program, University of Gottingen,
¨
Gottingen, Germany, 1996.
¨
[24] W. Kaminsky, M. Arndt, Adv. Pol. Sci. 127 (1997) 144.
[25] W. Kaminsky, A. Laban, Appl. Catal., A Gen. 222 (2001) 51.