Highly active self-assembled group-IV-metal multinuclear catalysts for ethylene polymerization

Article abstract of DOI:10.1016/j.jorganchem.2015.05.036

Group-IV metals play an important role in the production of practical catalysts for polyethylene industry. To develop new highly active catalysts, we designed a new self-assembly strategy that leads to a new family of group-IV-metal multinuclear catalysts

Full text of DOI:10.1016/j.jorganchem.2015.05.036

Journal of Organometallic Chemistry xxx (2015) 1e13
Contents lists available at ScienceDirect
Journal of Organometallic Chemistry
journal homepage: www.elsevier.com/locate/jorganchem
Highly active self-assembled group-IV-metal multinuclear catalysts
for ethylene polymerization
He-Kuan Luo*, Cun Wang, Wendy Rusli, Chuan-Zhao Li, Effendi Widjaja, Pui-Kwan Wong,
Ludger Paul Stubbs, Martin Van Meurs
Institute of Chemical and Engineering Sciences, Agency for Science, Technology and Research, 1 Pesek Road, Jurong Island, 627833, Singapore
a r t i c l e i n f o
a b s t r a c t
Article history:
Group-IV metals play an important role in the production of practical catalysts for polyethylene industry.
To develop new highly active catalysts, we designed a new self-assembly strategy that leads to a new
family of group-IV-metal multinuclear catalysts with bis-phenoxy-imine ligands. These multinuclear
catalysts are formed via three bridging strategies, by linking the imine-N, the 4-position of phenolate, or
the mixed linking strategy. The coordination environment of the catalysts was investigated with far-IR
and Raman. The multiple imine-H shift found in 1H NMR revealed that the multinuclear catalysts are
typical multi-site catalysts comprising of different coordination geometries. All the multinuclear catalysts
were more active and stable, and produced polyethylene of higher MW than the corresponding mono-
nuclear catalysts. The bridging spacer has an evident effect on catalytic activity, stability, polymer MW
and incorporation rate of propylene into the polyethylene backbone.
Received 17 January 2015
Received in revised form
17 May 2015
Accepted 23 May 2015
Available online xxx
Keywords:
Titanium
Zirconium
Bis-phenoxy-imine
Catalysts
© 2015 Elsevier B.V. All rights reserved.
Ethylene
Polymerization
1. Introduction
As bidentate phenoxy-imine catalysts (see FI-Ti/Zr in Scheme 1)
have limited lifetime, much effort has been made to improve their
The development of high performance catalysts for olefin
polymerization is of great importance to the chemical industry.
Among the various catalysts studied so far [1e7], phenoxy-imine-
based Group 4 metal catalysts [8e17] (see Scheme 1 and model-1
in Scheme 2) have received much attention because they are a
class of non-metallocene catalysts with high initial activity that
may potentially compete with the commercial metallocene or half-
sandwich Group 4 metal catalysts. However, these phenoxy-imine
catalysts have been reported to have poor stability resulting in
limited lifetime which was attributed to the transfer of the sup-
porting ligand to the aluminum in co-catalyst [18e21], especially at
elevated temperatures used in commercial processes. As a conse-
quence these catalysts were usually studied at low temperature
and/or short reaction time [8e17]. Therefore there is substantial
incentive to develop robust phenoxy-imine catalysts with high
activity and minimal catalyst decay.
stability by using tetradentate ligands, which were expected to
form more stable mono-nuclear catalysts of coordination model-2
(see Scheme 2). Ishii et al. [17] reported that tetradentate ligands
of Cn-chain-bridged phenoxy-imine units (see II, n ¼ 2e6, in
Scheme 3) formed mono-nuclear catalysts with zirconium. The
results showed that the complexes with ligands of longer bridge
(n ¼ 5 or 6) displayed high activity for five minutes run, while the
complexes with ligands of shorter bridge (n ¼ 2e4) displayed low
activity, and the issue of catalyst rapid deactivation was not
addressed. Scott and co-workers [18e21] also believed that the
tetradentate ligands incorporating titanium and zirconium may
form more stable catalysts bearing the coordination model-2 with
two imine-N linked. However, experimental results demonstrated
that the tetradentate ligands III and XII (see Scheme 3) did not
afford olefin polymerization catalysts, principally because of a
destructive 1,2-migratory insertion of a metal-bound alkyl/poly-
meryl chain into the imine C]N unit [18e20]. Subsequently, Knight
et al. [19,20] found that introducing an alkyl group at the position
R4 (see ligand XI in Scheme 3) of a zirconium salicylaldiminato
complex leads to a long-lived catalyst (1 h test) for ethylene poly-
merization because of steric blocking of an intramolecular 1,2-
migratory insertion. However, this steric blocking promotes a
* Corresponding author. Current address: Institute of Materials Research and
Engineering, Agency for Science, Technology and Research, 3 Research Link, 117602,
Singapore.
E-mail address: luoh@imre.a-star.edu.sg (H.-K. Luo).
http://dx.doi.org/10.1016/j.jorganchem.2015.05.036
0022-328X/© 2015 Elsevier B.V. All rights reserved.
Please cite this article in press as: H.-K. Luo, et al., Journal of Organometallic Chemistry (2015), http://dx.doi.org/10.1016/
j.jorganchem.2015.05.036

2
H.-K. Luo et al. / Journal of Organometallic Chemistry xxx (2015) 1e13
attention in developing new catalysts. So far, some of the materials
have been demonstrated to be efficient catalysts with better
selectivity and/or stability relative to corresponding mono-nuclear
catalysts. In order to develop highly active Ti(IV) and Zr(IV) cata-
lysts of phenoxy-imine with improved stability, here we use self-
assembly strategy between metals and bisphenoxy-imine ligands
to form multinuclear catalysts. The key of this strategy is the use of
bis-phenoxy-imine ligand that has a bridging spacer to separate the
two phenoxy-imine coordination units so that they have to coor-
dinate with different metals, thus leading to the formation of
multinuclear catalysts over mono-nuclear catalysts. One-step self-
assembly (see Fig. 1) using one ligand and two-step programmed
self-assembly (see Fig. 2) using two different ligands were inves-
tigated. The ethylene polymerization studies reveal that these multi-
nuclear catalysts are much more active, decay at a significantly
slower rate, and produce higher MW polymers compared to the
corresponding mono-nuclear catalysts. The multinuclear Ti-catalyst
with short bridging spacer (-C6H5-) displayed improved propylene
incorporation compared to the mono-nuclear FI catalyst. Here we
report that, besides using polymeric ligand [24,25], self-assembly is
an alternative approach to synthesize highly active multinuclear
catalysts for ethylene polymerization.
Scheme 1. Synthesis of mono-nuclear Ti/Zr catalysts and five possible isomers.
2. Experimental
2.1. General considerations
All manipulations involving air-sensitive materials were carried
out by using standard Schlenk line techniques or in a glove box
under an atmosphere of argon. 4,40-methylenedianiline, benzidine,
1,4-diaminobenzene, 3-tert-butyl-2-hydroxy-benzaldehyde and
anhydrous hexane were purchased from SigmaeAldrich and used
without pre-treatment. 5,5-Methylene-di-3-tert-butyl salicylalde-
hyde was prepared according to literature methods [26e28].
Methanol was dried over 4 Å molecular sieves. DCM and THF were
purified using an MBRAUN-SPS solvent purification system. 1H
NMR and 13C NMR were recorded in CDCl3 on a BRUKER 400 MHz
spectrometer. HRMS(EI) was performed on a Thermo Finnigan MAT
95. HRMS(ESI) was performed on an Agilent LC-MS TOF. Elemental
Scheme 2. Comparison of catalyst models: multinuclear vs mononuclear.
new radical catalyst decomposition mechanism in certain in-
stances, thus resulting in far lower activities compared to the cor-
responding catalyst based on the FI ligand. In addition, all the
zirconium complexes of ligands IV ~ X have no activity probably
due to the lack of steric bulk in the phenolate 2-position [20]. In
their further studies, Clarkson et al. [21] investigated more tetra-
dentate ligands (see XIII-XVII in Scheme 3). For titanium com-
plexes, [(XIII)TiCl2] had no activity for ethylene polymerization
when treated with MAO because the two chloride ligands are in
trans-arrangement. The cis-complex [(XIV)TiCl2] however was also
unproductive, perhaps due to enhanced imine reactivity brought on
by ring-strain in the diamine backbone. Complex [(XV)TiCl2] pro-
duced only a trace of polymer. Although complexes [(XVI)TiCl2] and
[(XVII)TiCl2] demonstrated significantly improved activity at 25 ꢀC
(in excess of 2 ꢁ 103 KgPE molꢂca1t hꢂ1 barꢂ1 for a one hour test), the
overall productivities are much lower at 50 ꢀC resulting from more
rapid catalyst decomposition. For zirconium complexes, complex
[(XV)ZrCl2] produced only a trace of polymer. The complexes [(XVI)
ZrCl2] and [(XVII)ZrCl2] demonstrated only low activities. Similarly,
introducing a methyl group at the phenolate 5-position of the
simple phenoxy-imine ligand may block the intramolecular 1,2-
migratory insertion, however this steric blocking failed to
improve the catalytic lifetime [20]. Therefore, after many in-
vestigations of various tetradentate ligands, the development of
long-lived and highly efficient phenoxy-imine-based Group 4 metal
catalysts is still a challenge.
analysis was performed on
a EuroEA3000 Series Elemental
Analyzer. Metal content was tested with ICP. Methyl aluminoxane
solution (Al%: ~5.2%) in toluene was purchased from Chemtura
Organometallics GmbH and used directly without any pre-
treatment. The known ligand (FI) and known FI-Ti/Zr catalysts
were prepared following the reported methods [8e17]. The
multinuclear catalysts (MNTi-1~5 and MNZr-1~5) were synthesized
via the same procedure. High temperature GPC analyses of poly-
ethylene were performed on a Polymer Labs GPC-220 with a
refractive index detector. Typical operating conditions for analyzing
polyethylene are: two PLgel 10
7.5 mm) and one PLGel 10
m
m Mixed B columns (300 ꢁ
m
m guard column (50 ꢁ 7.5 mm) at
160 ꢀC using 1,2,4-trichlorobenzene stabilized with 0.0125 wt. %
BHT as the eluent. Polymer samples were prepared at a concen-
tration of 1 mg/mL using a Polymer Labs SP260 sample preparation
system at 150 ꢀC until dissolved (typically about 4e6 h), followed
by filtration where necessary. X-ray photoelectron spectroscopy
(XPS) spectra were recorded on an ESCALAB 250 spectrometer and
Al Ka radiation was used as the X-ray source. The C1 peak at
The coordination of metal with multi-dentate ligands form a
material known as MOF (Metal-Organic Frameworks) or coordi-
nation polymer where the ligand acts as a bridging spacer among
the metals [22,23]. Although the published catalytic applications
using MOF or coordination polymer generally suffer from a lack of
characterization with respect to sample homogeneity and purity
[22], this kind of materials have been receiving more and more
284.5 eV was used as a reference for the calibration of the binding
energy (BE) scale. FT-IR spectra were recorded on a Bruker Vertex
70 spectrometer. Laser Raman spectra were recorded on an InVia
Reflex instrument (Renishaw) equipped with a near infrared
enhanced deep-depleted thermoelectrically Peltier cooled CCD
array detector (576 ꢁ 384 pixels) and a high grade Leica micro-
scope. The methyl branching of the copolymers was measured with
Please cite this article in press as: H.-K. Luo, et al., Journal of Organometallic Chemistry (2015), http://dx.doi.org/10.1016/
j.jorganchem.2015.05.036

H.-K. Luo et al. / Journal of Organometallic Chemistry xxx (2015) 1e13
3
Scheme 3. Tetradentate ligands forming mononuclear catalysts [17e21].
FT-IR by comparison against standard samples.
in a 1 L stainless steel autoclave, which was heated by recycling hot
oil. Before reaction, the autoclave was dried under vacuum at
100 ꢀC for 2 h during which period the autoclave was swept with
dry argon at least three times. After the temperature was lowered
to the reaction temperature (60 ꢀC), hexane (600 mL) and MAO
(12.0 mmol) were introduced, followed by the addition of catalyst
2.2. General procedure for ethylene homo-polymerization in
300 mL reactor
The polymerization was carried out in a 300 mL stainless steel
autoclave equipped with a mechanical stirrer with adjustable stir-
ring rate. The autoclave was heated by a heating mantle. Before
reaction, the autoclave was dried under vacuum at 100 ꢀC for 2 h
during which period the autoclave was swept with dry argon at
least three times. Then the temperature was lowered to the reac-
tion temperature (60 ꢀC) and the reactor was evacuated and refilled
with ethylene. Hexane (100 mL), MAO (2.0 mmol) and catalyst
solution in DCM were added consecutively via syringe under
ethylene atmosphere (~10 PSI) at a stirring rate of 300 RPM. Then
the autoclave was quickly pressurized to 80 PSI (5.5 bar) with
ethylene and the stirring rate was adjusted to 500 RPM. After the
polymerization was run for the required time, the ethylene pres-
sure was vented quickly and the reaction was quenched with 2 mL
ethanol. The produced polyethylene was collected by filtration,
washed with ethanol and hexane and dried in vacuo at 50 ꢀC. The
obtained white polymer was weighed and analyzed with GPC. The
(6 mmol) solution in DCM (3 mL) at a stirring rate of 300 RPM. Then
the autoclave was quickly pressurized to 3.0 bar with propylene,
followed by ethylene being added to 7.0 bar total pressure. During
the polymerization process, when the pressure dropped to 5.0 bar,
1 bar propylene and 1 bar ethylene were pressurized in up to
7.0 bar. After the copolymerization was run for 0.5 h, the pressure
was vented quickly and the reaction was quenched with 12 mL
ethanol. The produced copolymer was collected by filtration,
washed with ethanol and hexane and dried in vacuo at 50 ꢀC. The
obtained white polymer was weighed and analyzed with GPC. The
activity was calculated against the total pressure in unit of KgPE
molꢂM1 hꢂ1 barꢂ1
.
2.4. Bis-phenoxy-imine ligand (BFI-1, see Scheme 4)
4,40-methylenedianiline (1.34 g, 6.76 mmol) was dissolved into
anhydrous methanol (25 mL). After being stirred for several mi-
nutes, 3-tert-butyl-2-hydroxy-benzaldehyde (2.65 g, 14.87 mmol)
was added, followed by several drops of formic acid. The resulting
mixture was stirred for one hour at room temperature and then
refluxed for one day under argon atmosphere. After cooling to
room temperature, the product was isolated by filtration, washed
activity was calculated in unit of KgPE molꢂM1 ꢂ1 barꢂ1
h .
2.3. General procedure for copolymerization of Ethylene and
propylene in 1 L reactor
The copolymerization of ethylene and propylene was carried out
Please cite this article in press as: H.-K. Luo, et al., Journal of Organometallic Chemistry (2015), http://dx.doi.org/10.1016/
j.jorganchem.2015.05.036

4
H.-K. Luo et al. / Journal of Organometallic Chemistry xxx (2015) 1e13
2.5. Bis-phenoxy-imine ligand (BFI-2, see Scheme 4)
The ligand BFI-2 was synthesized via the same procedure as
ligand BFI-1 using benzidine (1.06 g, 5.74 mmol) and 3-tert-butyl-2-
hydroxy-benzaldehyde (2.09 g, 11.49 mmol) in anhydrous meth-
anol (30 mL). The product was obtained as a yellow powder in 99%
yield (2.80 g). 1H NMR (CDCl3, 400 MHz,
d): 1.51 (s, 18H, C(CH3)3),
6.92 (t, 2H, Ph-H, J ¼ 7.60 Hz), 7.30 (dd, 2H, Ph-H, J ¼ 7.60 Hz,
J ¼ 1.6 Hz), 7.40e7.45 (m, 6H, Ph-H), 7.70 (d, 4H, Ph-H, J ¼ 8.40 Hz),
8.71 (s, 2H, CH]N), 13.96 (s, 2H, OH). 13C NMR (CDCl3, 100 MHz,
d):
29.36, 34.94, 118.41, 119.12, 121.74, 127.87, 130.48, 130.71, 137.72,
138.80, 147.64, 160.61, 163.11. Elemental analysis 34H36N2O2
C
(504.68): Calc. C 80.92%, H 7.19%, N 5.55%; Found C 80.98%, H 7.12%,
N 5.62%. HRMS(EI, m/z): Calculated 504.2777; Found 504.2823
(Mþ). Single crystal was crystallized from toluene solution (for the
crystal details, see the supporting information).
2.6. Bis-phenoxy-imine ligand (BFI-3, see Scheme 4)
The ligand BFI-3 was synthesized via the same procedure as
ligand BFI-1 using 1,4-diaminobenzene (0.61 g, 5.61 mmol) and 3-
tert-butyl-2-hydroxy-benzaldehyde (2.00 g, 11.22 mmol) in anhy-
drous methanol (45 mL). The product was obtained as a yellow
powder in 94% yield (2.25 g). 1H NMR (CDCl3, 400 MHz,
d): 1.50 (s,
18H, C(CH3)3), 6.91 (t, 2H, Ph-H, J ¼ 7.80 Hz), 7.29 (dd, 2H, Ph-H,
J ¼ 7.60 Hz, J ¼ 1.4 Hz), 7.39 (s, 4H, Ph-H), 7.43 (dd, 2H, Ph-H,
J ¼ 7.80 Hz, J ¼ 1.4 Hz), 8.69 (s, 2H, CH]N), 13.90 (s, 2H, OH). 13
C
Fig. 1. Self-assembly to design multinuclear catalysts.
NMR (CDCl3, 100 MHz,
d): 29.35, 34.93, 118.44, 119.08, 122.25,
130.51, 130.71, 137.72, 146.93, 160.58, 162.95. Elemental analysis
28H32N2O2 (428.58): Calcd. C 78.47%, H 7.53%, N 6.54%; Found C
78.31%, H 7.50%, N 6.55%. HRMS(EI, m/z): Calculated 428.2464;
Found 428.2485 (Mþ). Single crystal was crystallized from toluene
solution (for the crystal details, see the supporting information).
C
with methanol (12 mL) and dried in vacuo affording 3.45 g of a
yellow powder, yield 98%. 1H NMR (CDCl3, 400 MHz,
): 1.50 (s,
d
18H, C(CH3)3), 4.04 (s, 2H, CH2), 6.89 (t, 2H, Ph-H, J ¼ 7.60 Hz), 7.26
(s, 10H, Ph-H), 7.41 (d, 2H, Ph-H, J ¼ 7.20 Hz), 8.63 (s, 2H, CH]N),
13.96 (s, 2H, OH). 13C NMR (CDCl3, 100 MHz,
d): 29.38, 34.93, 41.04,
118.34, 119.13, 121.37, 129.89, 130.30, 130.62, 137.67, 139.64, 146.66,
160.55, 162.90. Elemental analysis C35H38N2O2 (518.71): Calc.: C
81.05%, H 7.38%, N 5.40%; Found: C 80.89%, H 7.41%, N 5.46%. HRMS
(EI, m/z): Calculated 518.2933; Found 518.2903 (Mþ).
2.7. Bis-phenoxy-imine ligand (BFI-4, see Scheme 5)
The ligand BFI-4 was synthesized via the same procedure as
ligand BFI-1 using aniline (1.01 g, 10.86 mmol) and 5,5-methylene-
di-3-tert-butyl salicylaldehyde (2.00 g, 5.43 mmol) in anhydrous
Fig. 2. Two-step programmed self-assembly to design multinuclear catalysts.
Please cite this article in press as: H.-K. Luo, et al., Journal of Organometallic Chemistry (2015), http://dx.doi.org/10.1016/
j.jorganchem.2015.05.036

H.-K. Luo et al. / Journal of Organometallic Chemistry xxx (2015) 1e13
5
Scheme 4. Synthesis of multinuclear catalysts by bridging the imine-N.
methanol (80 mL). After reaction, the yellow slurry was concen-
trated to about 10 mL. The product was filtered, washed with
methanol (2 ꢁ 5 mL) and dried in vacuo affording 2.56 g of a yellow
was added dropwise over a period of 10 min at ꢂ78 ꢀC. Then the
mixture was allowed to warm to room temperature and stirred for
two hours. The resulting solution was added dropwise to a stirred
solution of TiCl4 (0.366 g, 1.93 mmol) in THF (15 mL) at ꢂ78 ꢀC via a
cannula over a period of 20 min. The resulting mixture was again
warmed to room temperature and stirred overnight for 18 h. After
removal of THF, the residual solid was extracted with 30 mL DCM
and filtered to give a clear solution. Removal of DCM gave a deep
reddish-brown solid which was ground to a fine powder with a
spatula. The obtained catalyst was dried in vacuo at room tem-
perature for several hours until its weight was equal to the theo-
retical yield calculated with C35H36Cl2N2O2TiꢃTHF. The catalyst has
a general repeating unit of C35H36Cl2N2O2TiꢃxTHF. Elemental
analysis and 1H NMR showed that x was close to 1. The catalyst
yield was 1.35 g (99%). FT-IR (cmꢂ1): 1608 (M), 1595 (S), 1587 (SH),
1554 (S), 1503 (M), 1481 (W), 1425 (M), 1393 (SH), 1384 (M), 1362
powder (91% yield). 1H NMR (CDCl3, 400 MHz,
d): 1.46 (s, 18H,
C(CH3)3), 3.92 (s, 2H, CH2), 7.01 (d, 2H, Ph-H, J ¼ 2.04 Hz), 7.23e7.27
(m, 8H, Ph-H), 7.39 (t, 4H, Ph-H, J ¼ 7.80 Hz), 8.57 (s, 2H, CH]N),
13.79 (s, 2H, OH). 13C NMR (CDCl3, 100 MHz,
d): 29.41, 34.93, 40.41,
118.93, 121.18, 126.68, 129.36, 130.28, 130.70, 131.32, 137.76, 148.53,
158.94,163.38. Elemental analysis C35H38N2O2 (518.71): Calculated:
C 81.05%, H 7.38%, N 5.40%; Found: C 81.09%, H 7.46%, N 5.47%.
HRMS (ESI): Calculated for [M þ H]þ: 519.3006; Found 519.3004.
2.8. Multinuclear catalyst MNTi-1 (see Scheme 4)
To a stirred solution of BFI-1 (1.00 g, 1.93 mmol) in THF (20 mL),
a solution of n-butyllithium (1.6 M, 2.4 mL, 3.86 mmol) in hexane
Scheme 5. Synthesis of multinuclear catalysts by bridging the phenolate-4 position.
Please cite this article in press as: H.-K. Luo, et al., Journal of Organometallic Chemistry (2015), http://dx.doi.org/10.1016/
j.jorganchem.2015.05.036

6
H.-K. Luo et al. / Journal of Organometallic Chemistry xxx (2015) 1e13
(W), 658 (W), 600 (W), 564 (W), 498 (W), 460 (W), 360 (W). Raman
(cmꢂ1): 1590 (M), 1553 (W), 1424 (S), 1296 (S), 1175 (W), 889 (W),
552 (W), 439 (W), 362 (W), 325 (W). 1H NMR (400 MHz, CDCl3):
2.12. Multinuclear catalyst MNTi-5 (see Scheme 6)
2.12.1. Synthesis of BFI-1-(TiCl3)2 solution in THF
d
¼ 1.25e1.61 (m, 18H, t-Bu), 3.55e3.59 (m-broad, 2H, CH2),
To a stirred solution of BFI-1 (0.50 g, 0.96 mmol) in THF (15 mL),
a solution of n-butyllithium (1.6 M, 1.2 mL, 1.92 mmol) in hexane
was added dropwise over a period of 10 min at ꢂ78 ꢀC. Then the
mixture was allowed to warm to room temperature and stirred for
two hours. The resulting solution was added dropwise to a stirred
solution of TiCl4 (0.364 g, 1.92 mmol) in THF (15 mL) at ꢂ78 ꢀC via a
cannula over a period of 20 min. The resulting mixture was again
warmed to room temperature and stirred for 18 h affording a so-
lution of BFI-1-(TiCl3)2 in THF.
6.44e7.66 (m-broad, 14H, AreH), 7.83(s), 7.87e8.07 (m-broad), 8.13
(s), 8.14 (s), 8.18 (s) (2H, CH]N). Calculated for C35H36Cl2N2O2
-
TiꢃTHF (FW 707.55): C 66.20%, H 6.27%, N 3.96%, Ti 6.77%; Found: C
66.05%, H 6.38%, N 3.81%, Ti 6.52%.
2.9. Multinuclear catalyst MNTi-2 (see Scheme 4)
The title catalyst MNTi-2 was synthesized via the same proce-
dure as MNTi-1 using ligand BFI-2 (1.00 g, 1.98 mmol) in 30 mL THF
and equimolar TiCl4 (0.376 g, 1.98 mmol) in 30 mL THF. The
multinuclear catalyst MNTi-2 was obtained as a deep reddish-
2.12.2. Synthesis of catalyst MNTi-5
To a stirred solution of BFI-4 (0.50 g, 0.96 mmol) in THF (15 mL),
a solution of n-butyllithium (1.6 M, 1.2 mL, 1.92 mmol) in hexane
was added dropwise over a period of 10 min at ꢂ78 ꢀC. Then the
mixture was allowed to warm to room temperature and stirred for
two hours. The resulting solution was added dropwise to the so-
lution of BFI-1-(TiCl3)2 via a cannula over a period of 20 min
at ꢂ78 ꢀC. The resulting mixture was warmed to room temperature
and stirred for 18 h. After removal of THF, the residue was extracted
with DCM (40 mL) and filtered to give a clear solution. Removal of
DCM gave a deep reddish-brown solid which was ground to a fine
powder. The multinuclear catalyst was dried in vacuo at room
temperature for several hours until its weight was equal to the
theoretical yield calculated with C35H36Cl2N2O2TiꢃTHF. The catalyst
has a general repeating unit of C35H36Cl2N2O2TiꢃxTHF. Elemental
analysis and 1H NMR showed that x was close to 1. The catalyst
yield was 1.36 g (100%). FT-IR (cmꢂ1): 1606 (M,SH), 1590 (M), 1554
(S), 1502 (W), 1486 (W), 1430 (M,BR), 1392 (M,SH), 1362 (W), 650
(W,BR), 600 (W), 592 (W,SH), 564 (W), 498 (W,BR), 458 (W), 362
(W). Raman (cmꢂ1): 1584 (S), 1541 (S), 1427 (S), 1378 (M), 1302 (S),
1244 (W), 884 (M, SH), 866(M), 549 (M), 425 (W), 369 (W), 327 (W).
brown solid with a general repeating unit of C34H34Cl2N2O2
-
TiꢃxTHF. Elemental analysis and 1H NMR showed that x was close to
1. The catalyst yield was 1.37 g (100%). FT-IR (cmꢂ1): 1606 (SH),1595
(S), 1586 (M), 1554 (S), 1490 (S), 1466 (W), 1425 (M), 1393 (SH), 1384
(M), 1362 (W), 659 (W), 612 (W), 601 (W), 565 (W), 552 (SH), 460
(W), 360 (W,BR). Raman (cmꢂ1): 1593 (S), 1553 (M), 1424 (S), 1297
(S), 1177 (S), 889 (W), 560 (W), 447 (W), 364 (W), 334 (W). 1H NMR
(400 MHz, CDCl3):
d
¼ 1.24e1.66 (m, 18H, t-Bu), 6.65e7.66 (m-
broad, 14H, AreH), 7.90e8.21 (m-broad) (2H, CH]N). Calculated
for C34H34Cl2N2O2TiꢃTHF (FW 693.52): C 65.81%, H 6.10%, N 4.04%,
Ti 6.90%; Found: C 65.67%, H 6.01%, N 4.09%, Ti 6.75%.
2.10. Multinuclear catalyst MNTi-3 (see Scheme 4)
The title catalyst MNTi-3 was synthesized via the same proce-
dure as MNTi-1 using ligand BFI-3 (1.00 g, 2.33 mmol) in 30 mL THF
and equimolar TiCl4 (0.442 g, 2.33 mmol) in 30 mL THF. The
multinuclear catalyst MNTi-3 was obtained as a deep reddish-
1H NMR (400 MHz, CDCl3):
(m-broad), 3.83e3.91 (m-broad) (2H, CH2), 6.50e7.66 (m-broad,
14H, AreH), 7.81(s), 7.83 (s), 7.92e8.09 (m-broad), 8.13 (s), 8.15 (s),
8.17 (s), 8.18 (s), 8.22 (s) (2H, CH]N). Calculated for
d
¼ 1.20e1.65 (m, 18H, t-Bu), 3.55e3.59
brown solid with a general repeating unit of C28H30Cl2N2O2
-
Ti$xTHF. Elemental analysis and 1H NMR showed that x was close to
1. The catalyst yield was 1.43 g (99%). FT-IR (cmꢂ1): 1608 (M,SH),
1601 (S), 1586 (M), 1554 (S), 1506 (M,SH), 1497 (M), 1482 (W), 1427
(M), 1386 (M,BR), 1362 (W), 670 (W), 625 (W), 612 (W), 595 (W),
565 (W), 542 (W), 497 (W), 462 (W), 455 (W,SH), 402 (W), 363
(W,BR). Raman (cmꢂ1): 1587 (S), 1554 (W), 1427 (S), 1294 (S), 1171
(W) 893 (W), 564 (W), 458 (W), 367 (W), 323 (W). 1H NMR
C
35H36Cl2N2O2TiꢃTHF (FW 707.55): C 66.20%, H 6.27%, N 3.96%, Ti
6.77%; Found: C 66.34%, H 6.35%, N 3.96%, Ti 6.55%.
2.13. Multinuclear catalyst MNZr-1 (see Scheme 4)
(400 MHz, CDCl3):
broad,10H, AreH), 7.88e8.18 (m-broad) (2H, CH]N). Calculated for
28H30Cl2N2O2Ti$THF (FW 617.43): C 62.25%, H 6.20%, N 4.54%, Ti
d
¼ 1.24e1.68 (m, 18H, t-Bu), 6.46e7.62 (m-
The title catalyst MNZr-1 was synthesized via the same proce-
dure as MNTi-1 using ligand BFI-1 (1.00 g, 1.93 mmol) and equi-
molar ZrCl4 (0.450 g, 1.93 mmol). The multinuclear Zr catalyst was
obtained as pale yellow solid with a general repeating unit of
C
7.75%; Found: C 62.41%, H 6.19%, N 4.70%, Ti 7.73%.
C
35H36Cl2N2O2Zr$xTHF. Elemental analysis and 1H NMR showed
that x was close to 1. The catalyst yield was 1.42 g (98%). FT-IR
(cmꢂ1): 1610 (M), 1590 (S), 1553 (S), 1513 (M,SH), 1503 (M), 1482
(W), 1428 (M), 1389 (M,BR), 1360 (W), 650 (W,BR), 635 (W), 601
(W), 574 (W), 548 (W), 526 (W), 500 (W,BR), 480 (W), 445 (W), 426
2.11. Multinuclear catalyst MNTi-4 (see Scheme 5)
The title catalyst MNTi-4 was synthesized via the same proce-
dure as MNTi-1 using ligand BFI-4 (0.60 g, 1.16 mmol) in 20 mL THF
and equimolar TiCl4 (0.220 g, 1.16 mmol) in 15 mL THF. The
multinuclear catalyst MNTi-4 was obtained as a deep reddish-
(W), 370 (W), 330 (W). 1H NMR (400 MHz, CDCl3):
d
¼ 1.26e1.58
(m, 18H, t-Bu), 3.65 (broad, 2H, CH2), 6.68e7.65 (m-broad, 14H,
AreH), 7.97e8.10 (m-broad), 8.15 (s), 8.16 (s), 8.21 (s) (2H, CH]N).
Calculated for C35H36Cl2N2O2ZrꢃTHF (FW 750.91): C 62.38%, H
5.91%, N 3.73%, Zr 12.15%; Found: C 62.50%, H 6.10%, N 3.71%, Zr
12.15%.
brown solid with a general repeating unit of C35H36Cl2N2O2
-
Ti$xTHF. Elemental analysis and 1H NMR showed that x was close to
1. The catalyst yield was 0.81 g (99%). FT-IR (cmꢂ1): 1606 (M), 1591
(M), 1551 (S), 1487 (M), 1451 (W), 1433 (M), 1422 (M), 1384 (M,SH),
1362 (W), 596 (W), 564 (W), 498 (W), 471 (W), 443 (W), 362
2.14. Multinuclear catalyst MNZr-2 (see Scheme 4)
(W,BR). 1H NMR (400 MHz, CDCl3):
3.78e3.94 (m-broad, 2H, CH2), 6.71e7.43 (m-broad, 14H, AreH),
d
¼ 1.21e1.60 (m, 18H, t-Bu),
The title catalyst MNZr-2 was synthesized via the same proce-
dure as MNTi-1 using ligand BFI-2 (1.00 g, 1.98 mmol) in 30 mL THF
and equimolar ZrCl4 (0.461 g, 1.98 mmol) in 30 mL THF. The
multinuclear catalyst MNZr-2 was obtained as a pale yellow solid
7.83e8.10 (m-broad, 2H, CH]N). Calculated for C35H36Cl2N2O2
-
TiꢃTHF (FW 707.55): C 66.20%, H 6.27%, N 3.96%, Ti 6.77%; Found: C
66.34%, H 6.37%, N 4.02%, Ti 6.58%.
Please cite this article in press as: H.-K. Luo, et al., Journal of Organometallic Chemistry (2015), http://dx.doi.org/10.1016/
j.jorganchem.2015.05.036

H.-K. Luo et al. / Journal of Organometallic Chemistry xxx (2015) 1e13
7
Scheme 6. Synthesis of multinuclear catalysts with two different bis-phenoxy-imine ligands.
with a general repeating unit of C34H34Cl2N2O2Zr$xTHF. Elemental
analysis and 1H NMR showed that x was close to 1. The catalyst
yield was 1.43 g (98%). FT-IR (cmꢂ1): 1607 (M,SH),1590 (S),1553 (S),
1488 (M), 1428 (M), 1391 (M), 1361 (W), 654 (W), 602 (W), 558 (W),
445 (W), 370 (W), 330 (W). 1H NMR (400 MHz, CDCl3):
yield was 0.87 g (100%). FT-IR (cmꢂ1): 1607 (M), 1589 (S), 1551 (S),
1482 (M), 1451 (W), 1432 (M), 1390 (M), 1361 (W), 665 (W), 616
(W), 584 (W), 556 (W), 500 (W), 460 (W), 330 (W). 1H NMR
(400 MHz, CDCl3):
broad, 2H, CH2), 6.71e7.43 (m-broad, 14H, AreH), 8.04 (s-broad, 2H,
CH]N). Calculated for
d
¼ 1.21e1.46 (m, 18H, t-Bu), 3.81e3.92 (m-
d
¼ 1.24e1.63 (m, 18H, t-Bu), 6.79e7.70 (m-broad, 14H, AreH),
C
35H36Cl2N2O2ZrꢃTHF (FW 750.91):
C
7.88e8.30 (m-broad) (2H, CH]N). Calculated for
62.38%, H 5.91%, N 3.73%, Zr 12.15%; Found: C 62.56%, H 6.27%, N
3.90%, Zr 12.13%.
C
34H34Cl2N2O2ZrꢃTHF (FW 736.88): C 61.94%, H 5.74%, N 3.80%, Zr
12.38%; Found: C 61.83%, H 5.71%, N 3.94%, Zr 11.30%.
2.17. Multinuclear catalyst MNZr-5 (see Scheme 6)
2.15. Multinuclear catalyst MNZr-3 (see Scheme 4)
The title catalyst MNZr-5 was synthesized via the same proce-
dure as MNTi-5 using BFI-1 (0.50 g, 0.96 mmol), BFI-4 (0.50 g,
0.96 mmol) and ZrCl4 (0.447 g, 1.92 mmol). The catalyst MNZr-5
was obtained as a pale yellow solid with a general repeating unit
of C35H36Cl2N2O2Zr$xTHF. Elemental analysis and 1H NMR showed
that x was close to 1. The catalyst yield was 1.41 g (98%). FT-IR
(cmꢂ1): 1608 (M), 1589 (S), 1553 (S), 1485 (M), 1432 (M,BR), 1390
(M), 1361 (W), 665 (W), 650 (W), 601 (W), 585 (W), 558 (W), 500
(W), 445 (W,BR), 370 (W, BR), 330 (W,BR). 1H NMR (400 MHz,
The title catalyst MNZr-3 was synthesized via the same proce-
dure as MNTi-1 using ligand BFI-3 (0.50 g, 1.17 mmol) in 20 mL THF
and equimolar ZrCl4 (0.273 g, 1.17 mmol) in 15 mL THF. The
multinuclear catalyst MNZr-3 was obtained as a pale yellow solid
with a repeating unit of C28H30Cl2N2O2Zr$xTHF. Elemental analysis
and 1H NMR showed that x was close to 1. The catalyst yield was
0.77 g (100%). FT-IR (cmꢂ1): 1611 (M,SH),1597 (S),1589 (S),1553 (S),
1506 (M), 1495 (M), 1482 (W), 1460 (W), 1426 (M,BR), 1390 (M),
1361 (W), 662 (W), 618 (W), 445 (W,BR), 370 (W,BR), 330 (W,BR).
CDCl3):
d
¼ 1.19e1.61 (m, 18H, t-Bu), 3.65 (broad), 3.82e3.98 (m-
1H NMR (400 MHz, CDCl3):
d
¼ 1.26e1.62 (m, 18H, t-Bu), 6.44e7.58
broad) (2H, CH2), 6.70e7.40 (m-broad, 14H, AreH), 7.75(s), 7.78 (s),
8.01e8.12 (m-broad), 8.15 (s), 8.17 (s), 8.20 (s), 8.21 (s), 8.25 (s) (2H,
(m-broad, 10H, AreH), 7.92e8.22 (m-broad) (2H, CH]N). Calcu-
lated for C28H30Cl2N2O2Zr$THF (FW 660.79): C 58.16%, H 5.80%, N
4.24%, Zr 13.81%; Found: C 58.31%, H 5.92%, N 4.28%, 13.78%.
CH]N). Calculated for
62.38%, H 5.91%, N 3.73%, Zr 12.15%; Found: C 62.51%, H 6.12%, N
3.81%, Zr 12.09%.
C
35H36Cl2N2O2ZrꢃTHF (FW 750.91):
C
2.16. Multinuclear catalyst MNZr-4 (see Scheme 5)
3. Results and discussions
The title catalyst MNZr-4 was synthesized via the same proce-
dure as MNTi-1 using ligand BFI-4 (0.60 g, 1.16 mmol) in 20 mL THF
and equimolar ZrCl4 (0.270 g, 1.16 mmol) in 15 mL THF. The
multinuclear catalyst MNZr-4 was obtained as a pale yellow solid
with a general repeating unit of C35H36Cl2N2O2Zr$xTHF. Elemental
analysis and 1H NMR showed that x was close to 1. The catalyst
3.1. Three strategies to form Methylene-linked multinuclear
catalysts
Containing two phenoxy-imine ligands is one of the key features
of the mono-nuclear FI catalyst (see Scheme 1). Hence, here we use
Please cite this article in press as: H.-K. Luo, et al., Journal of Organometallic Chemistry (2015), http://dx.doi.org/10.1016/
j.jorganchem.2015.05.036

8
H.-K. Luo et al. / Journal of Organometallic Chemistry xxx (2015) 1e13
Table 2
three strategies to bridge the mono-nuclear catalyst to form
methylene-linked multinuclear catalysts: (I) Link the aniline frag-
ment to make the catalysts MNTi-1 and MNZr-1 (see Scheme 4), (II)
Link the phenolate fragment to make the catalysts MNTi-4 and
MNZr-4 (see Scheme 5) and (III) Use a combination of the above
two linking strategies to make the catalysts MNTi-5 and MNZr-5
(see Scheme 6).
Ethylene polymerization with MNZr and FI-Zra.
Entry Cat.
Time (min) PE (g) Activityb (ꢁ103) Mn (ꢁ103) Mw/Mn
1
2
3
4
5
6
7
8
MNZr-1
5
15
30
60
2.74 66.4
4.27 34.5
4.94 20.0
6.20 12.5
6.8
10.3
12.0
14.2
23.4
38.0
15.5
16.8
34.0
40.2
9.4
11.5
11.3
10.8
18.8
17.2
41.4
31.6
32.3
29.1
11.0
22.6
16.7
35.6
35.3
35.2
41.0
47.3
56.0
38.6
6.8
11.9
14.7
42.2
7.6
46.5
75.9
118
135
111
56.4
45.3
54.0
70.5
83.9
43.0
52.5
69.7
44.4
59.1
6.1
MNZr-1
MNZr-1
MNZr-1
MNZr-1 120
8.46
8.55
MNZr-2
MNZr-2
MNZr-2
MNZr-2
5
15
30
60
2.78 67.4
4.11 33.2
6.68 27.0
8.91 18.0
3.2. Catalytic activity and stability of methylene-linked
multinuclear catalysts
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
MNZr-2 120
9.73
9.83
The multinuclear catalysts were evaluated for ethylene poly-
merization for different reaction times using MAO as co-catalyst
(see Tables 1 and 2). The results showed that all the methylene-
linked multinuclear catalysts displayed higher activity and better
stability than the corresponding mono-nuclear catalysts (see Figs. 3
and 4, also see Fig. S1eS9 in supporting information). For multi-
nuclear Ti-based catalysts, the catalyst MNTi-1 showed much
higher activity and much slower catalyst decay compared to its
mono-nuclear counterpart FI-Ti at 60 ꢀC in hexane. Its cumulative
average activities after 30, 60 and 120 min were 1.4, 2.0 and 2.4
times higher than those of FI-Ti (see entries 1e3 and 16e18 in
Table 1). The average activities over the three reaction periods of
0e30 min, 30e60 min and 60e120 min, can be obtained by
comparing the productivity of three reaction times (30, 60 and
120 min). The average activities of the multinuclear MNTi-1 over
the above three reaction periods were 1.4, 4.5, and 5.8 times higher
than those of FI-Ti. The other two multinuclear catalysts MNTi-4
and MNTi-5 with different linking strategies also showed im-
provements in catalytic activity and stability (see entries 10e15).
Both MNTi-4 and MNTi-5 were 1.7 times more active than FI-Ti over
a 120 min period (see entries 12, 15 and 18 in Table 1). When
comparing the linking strategies, MNTi-1 with the anilines linked is
more active than MNT-4 with the phenolates linked.
MNZr-3
MNZr-3
MNZr-3
MNZr-3
5
15
30
60
1.93 46.8
3.97 32.1
6.29 25.4
8.16 16.5
MNZr-3 120
9.73
9.83
MNZr-4
MNZr-4
MNZr-4
MNZr-4
5
15
30
60
2.11 51.2
3.69 29.8
4.74 19.2
6.33 12.8
MNZr-4 120
7.38
7.45
MNZr-5
MNZr-5
MNZr-5
MNZr-5
5
15
30
60
3.11 75.4
5.78 46.7
8.00 32.3
10.23 20.7
12.45 12.6
1.44 34.9
1.76 14.2
MNZr-5 120
FI-Zr
FI-Zr
FI-Zr
FI-Zr
FI-Zr
5
15
30
60
120
3.43
3.63
4.29
4.56
5.10
10.2
19.1
67.8
36.0
1.81
1.89
1.96
7.31
3.82
1.98
a
300 mL stainless steel autoclave, 100 mL of hexane, 5.5 bar of ethylene pressure,
60 ꢀC, 2.0 mmol of MAO, catalyst loading: 0.09
m
mol metal.
b
kgPE molꢂM1 hꢂ1 barꢂ1
.
4.57 ꢁ 103 kgPE molMꢂ1 hꢂ1 barꢂ1, but FI-Zr dropped to a very low
activity after 1 h. The other two new catalysts MNZr-4 and MNZr-5
showed similar improvements in activity and stability. Compared
to the activities of FI-Zr after 5, 15, 30, 60 and 120 min, the cumu-
lative average activities of MNZr-4 were 1.5, 2.1, 2.6, 3.3 and 3.8
times higher (see entries 16e20, Table 2), while the cumulative
average activities of MNZr-5 were 2.2, 3.3, 4.4, 5.4 and 6.4 times
higher (see entries 21e25, Table 2).
The three methylene-linked multinuclear Zr-based catalysts
also displayed much improved activity and stability compared to
the mono-nuclear FI-Zr catalyst (see Fig. 4). The cumulative average
activities of MNZr-1 after 5, 15, 30, 60 and 120 min were 1.9, 2.4, 2.7,
3.3 and 4.3 times higher, respectively, than those of mono-nuclear
FI-Zr (see entries 1e5 and 26e30 in Table 2). During the second
hour of polymerization, MNZr-1 still attains a high activity of
Table 1
Ethylene polymerization with MNTi and FI-Tia.
Entry Cat.
Time (min) PE (g) Activityb (ꢁ103) Mn (ꢁ103) Mw/Mn
1
2
3
4
5
6
7
8
MNTi-1
MNTi-1
MNTi-1 120
MNTi-2
MNTi-2
MNTi-2 120
MNTi-3
MNTi-3
MNTi-3 120
MNTi-4
MNTi-4
MNTi-4 120
MNTi-5
MNTi-5
MNTi-5 120
FI-Ti
FI-Ti
FI-Ti
30
60
6.55 2.65
11.19 2.26
15.61 1.58
7.14 2.88
12.92 2.61
20.02 2.02
7.34 2.97
11.93 2.41
16.10 1.63
6.01 2.43
8.42 1.70
10.82 1.09
4.85 1.96
7.92 1.60
10.87 1.10
4.69 1.89
5.73 1.16
6.49 0.66
651
662
740
642
716
828
445
357
693
285
664
887
492
903
984
329
385
493
2.5
2.9
3.2
4.5
4.1
3.7
7.2
9.6
5.7
8.5
3.6
6.0
4.7
3.0
3.9
2.0
2.0
3.2
30
60
30
60
9
10
11
12
13
14
15
16
17
18
30
60
30
60
30
60
120
a
300 mL stainless steel autoclave, 100 mL of hexane, 5.5 bar of ethylene pressure,
60 ꢀC, 2.0 mmol of MAO, catalyst loading: 0.9
m
mol metal.
b
kgPE molꢂM1 hꢂ1 barꢂ1
.
Fig. 3. Productivity comparison of MNTi-1, 4 and 5 with FI-Ti.
Please cite this article in press as: H.-K. Luo, et al., Journal of Organometallic Chemistry (2015), http://dx.doi.org/10.1016/
j.jorganchem.2015.05.036

H.-K. Luo et al. / Journal of Organometallic Chemistry xxx (2015) 1e13
9
ligands BFI-2 and 3 and multinuclear catalysts MNTi-2 and 3. The
crystal structures of BFI-2 and BFI-3 clearly showed that the two
phenoxy-imine unites cannot coordinate to a single metal atom,
therefore it is impossible to form mononuclear catalyst (see Fig. S12
and S13 in supporting information). The ethylene polymerization
results showed that MNTi-2 with eC6H5C6H5ꢂ as the bridging
spacer displayed the highest activity and slowest decay with an
activity of 2.88 ꢁ 103 KgPE molMꢂ1 hꢂ1 barꢂ1 for a 30 min run (see
entry 4, Table 1). Its cumulative average activities after 30, 60 and
120 min run times were 1.5, 2.3, and 3.1 times higher (see entries
4e6, Table 1) compared to those of FI-Ti (see entries 16e18,
Table 1). When measured over the three reaction periods of 0e30,
30e60, and 60e120 min, its average activities were 1.5, 5.6, and 9.3
times higher relative to FI-Ti. When the bridging spacer was further
shortened to one phenyl (eC6H5e), the corresponding catalyst
MNTi-3 was still more active and stable than FI-Ti. However, MNTi-
3 produced polyethylene with the broadest MWD among the cat-
alysts MNTi-1~3 because the shortest bridging spacer results in the
most crowded metal centres. For example, after 60 min, the Mw/Mn
was 9.6 for MNTi-3, which is broader than in the case of MNTi-1
and MNTi-2 (Mw/Mn ¼ 2.9 and 4.1). Compared to the flexible
bridging spacer eC6H5CH2C6H5ꢂ, the two spacers eC6H5C6H5ꢂ
and eC6H5ꢂ are more rigid and hence the distance between the
metal centres may be better controlled. This could be the reason
why both MNTi-2 and MNTi-3 afforded higher activity and better
stability than MNTi-1 with a flexible spacer (eC6H5CH2C6H5e).
On the other hand, the multinuclear catalysts MNZr-2 and 3
containing two different bridging spacers (eC6H5C6H5e and
eC6H5ꢂ) were also prepared and investigated together with MNZr-
1. The results showed that all the catalysts displayed improvements
in catalytic activity and stability relative to their mono-nuclear
analogue (see entries 1e15, Table 2). After 120 min of run time,
the activities of MNZr-1~3 were 4.3, 5.0 and 5.0 times higher than
that of FI-Zr. The MNZr-3 with shortest bridging spacer (eC6H5e)
afforded the broadest MWD. For example, its Mw/Mn after 60 min of
polymerisation was 135, which is broader than those obtained with
MNZr-1~2 (Mw/Mn ¼ 56.0 and 42.2, respectively).
Fig. 4. Productivity comparison of MNZr-1, 4 and 5 with FI-Zr.
3.3. Molecular weight (MW) and polydispersity (Mw/Mn)
Besides being more active and stable, the methylene-linked
multinuclear catalysts also produced polyethylene of higher MW
compared to the corresponding mono-nuclear catalysts (see
Tables 1 and 2). For example, the Mn of the polyethylene produced
during 2 h of polymerisation was much higher for MNTi-1, MNTi-4
and MNTi-5 (Mn ¼ 740 ꢁ 103, 887 ꢁ 103 and 984 ꢁ 103) than that of
FI-Ti (Mn ¼ 493 ꢁ 103). Similarly, the Mn of the polyethylene pro-
duced by MNZr-1, MNZr-4 and MNZr-5 (Mn ¼ 23.4 ꢁ 103; 29.1 ꢁ 103
and 35.3
ꢁ
103) was much higher than that of FI-Zr
(Mn ¼ 5.10 ꢁ 103).
On the other hand, the multinuclear catalysts afforded broader
molecular weight distributions (MWD) compared to the corre-
sponding mono-nuclear catalysts. For example, after 60 min, the
Mw/Mn was 2.9, 3.6 and 3.0 for MNTi-1, MNTi-4 and MNTi-5,
respectively, while for FI-Ti the Mw/Mn was only 2.0. For the Zr-
based catalysts, both the MNZr catalysts and FI-Zr afforded very
broad MWDs after long reaction times (60 min and 120 min).
Contrastingly, after only 5 min reaction time, the MWD was much
broader for MNZr-1, MNZr-4 and MNZr-5 (Mw/Mn ¼ 35.2, 56.4 and
43.0) than that of FI-Zr (Mw/Mn ¼ 6.1). When comparing the
multinuclear Ti vs Zr catalysts, the MNTi were found to produce
polyethylene of higher MW and narrower MWD than MNZr (for 2-h
run time, MNTi-1: Mn ¼ 740 ꢁ 103, Mw/Mn ¼ 3.2 vs MNZr-1:
Mn ¼ 23.4 ꢁ 103, Mw/Mn ¼ 38.6, see entry 3 in Table 1, entry 5 in
Table 2). The broader MWD of polyethylene produced by the
multinuclear catalysts is consistent with them behaving more like
multi-site ZieglereNatta catalysts than single-site homogeneous
catalysts.
The bridging spacer can also affect the reactor fouling. MNTi-1
and MNTi-3 showed reduced levels of fouling compared to the
mono-nuclear catalyst FI-Ti which caused severe reactor fouling
(see Fig. S11 in supporting information). Interestingly, MNTi-2 can
significantly reduce reactor fouling besides displaying the highest
activity and the best stability. The reactor is still quite clean after 2 h
polymerization (see Fig. S10 in supporting information). In
conclusion, multinuclear catalysts with suitable bridging spacers
(such as eC6H5C6H5ꢂ) resemble heterogeneous catalysts with
respect to reactor fouling. This is important for achieving contin-
uous production in industrial applications.
3.5. The impact of bridging spacer in copolymerization of ethylene
and propylene
The multinuclear MNTi-1~3 catalysts were investigated for the
copolymerization of ethylene and propylene (see Table 3). The
levels of methyl branching obtained with MNTi-1~3 were 12.4%,
14.0% and 18.4% for the three different bridging spacers
(eC6H5CH2C6H5e, eC6H5C6H5ꢂ and eC6H5ꢂ), indicating that the
shortest bridging spacer afforded the highest propylene incorpo-
ration. Compared to the mono-nuclear FI-Ti, the MNTi-3 with short
bridging spacer (eC6H5e) demonstrated improved propylene
incorporation (methyl branching: 18.4% vs 13.5%) while also dis-
playing a higher activity (729 vs 419 KgPE molꢂM1 hꢂ1 barꢂ1). This
result is consistent with that of bi-metallic catalysts which was
reported to display a cooperative effect to improve their copoly-
merization capability [29e31].
3.4. The impact of bridging spacer in ethylene homo-
polymerization
For methylene-linked multinuclear catalysts MNTi-1 and MNZr-
1, the bridging spacer between the two imine-N is eC6H5CH2C6H5ꢂ
(see Scheme 4). Because the bridging spacer controls the distance
between the metal centres, it is expected to have an impact on the
catalytic capabilities. Hence it is essential to study the effect of
bridging spacer.
Here we selected two more different bridging spacers
(eC6H5C6H5e and eC6H5ꢂ) and prepared the corresponding
Please cite this article in press as: H.-K. Luo, et al., Journal of Organometallic Chemistry (2015), http://dx.doi.org/10.1016/
j.jorganchem.2015.05.036

10
H.-K. Luo et al. / Journal of Organometallic Chemistry xxx (2015) 1e13
Table 3
Copolymerization of ethylene and propylene with MNTi-1~3 and FI-Tia.
Entry
Cat.
Bridging spacer
PE (g)
Activityb
Branch (wt%)
Mn (ꢁ103)
Mw/Mn
1
2
3
4
MNTi-1
MNTi-2
MNTi-3
FI-Ti
-C6H5CH2C6H5-
-C6H5C6H5-
-C6H5-
12.0
16.7
15.3
8.8
571
795
729
419
12.4
14.0
18.4
13.5
191
190
94
3.0
3.6
5.5
2.5
N.A.
206
a
1L stainless steel autoclave, 600 mL of hexane, 4.0 bar of ethylene pressure, 3.0 bar of propylene pressure, 12.0 mmol of MAO, catalyst loading: 6.0
m
mol metal, 60 ꢀC.
b
kgPE molꢂM1 hꢂ1 barꢂ1
.
3.6. Catalyst analysis with FT-IR, Raman, 1H NMR and XPS
shift relative to corresponding ligands. The C]N vibrations
C]N) of MNZr-1e3, MNZr-5 and FI-Zr were all observed at the
(y
For practical applications, we want to use all the obtained cat-
alysts without further purification after they were isolated from
LiCl because washing with solvents or re-crystallization or re-
precipitation from solvents usually loses much catalyst resulting
in much lower yield. The solvent residue THF can be removed from
the catalyst by Al(III) in excess MAO since Al (III) is a stronger Lewis
acid [32]. Hence the catalyst can be completely activated by MAO
and display its catalytic capability.
same position of 1553 cm-1, while the C]N vibration of MNZr-4
was observed at 1551 cm-1 due to a different linking strategy as
well. In the far-IR spectra of MNZr-1e3, MNZr-5 and FI-Zr (see right
spectra in Fig. 6), the ZreO, ZreCl and ZreN bonds were observed at
445, 370 and 330 cm-1, respectively. The ZreN bond of MNZr-4 was
also observed at 330 cm-1, but the ZreO and ZreCl bonds could not
be identified.
The C]N vibrations (yC]N) and the titanium coordination
To understand the formation of multinuclear catalysts, their
metal coordination environments were studied with FT-IR and
Laser Raman spectroscopy. In the mid-IR spectra of Ti-based cata-
environment comprising of TieO, TieCl and TieN bonds were also
observed in the Laser Raman spectra (see Fig. 7). In general, the
MNTi-1~3, MNTi-5 and FI-Ti gave similar spectra. Their C]N vi-
brations were observed at 1590/1553 cm-1, 1593/1553 cm-1, 1587/
1554 cm-1, 1584/1541 cm-1 and 1586/1555 cm-1, respectively. Their
TieO bond vibrations were observed at 889/552/439 cm-1, 889/
560/447 cm-1, 893/564/458 cm-1, 884/866/549/425 cm-1 and 890/
558/446 cm-1, respectively. Their TieCl bond vibrations were
observed at 362, 364, 367, 369 and 363 cm-1, respectively. Their
TieN bond vibrations were observed at 325, 334, 323, 327 and
336 cm-1, respectively.
In general, the coordination bonds and coordination geometry
are the key factors of an active catalyst. The FT-IR and Raman
studies revealed that the multinuclear catalysts have the same
coordination bonds relative to the corresponding mono-nuclear
catalysts. However, all the multinuclear catalysts were improved
in catalytic activity and stability. This may indicate that the multi-
nuclear catalysts have different coordination geometries. Because
the imine-H is close to N that is coordinated to metal, the change in
lysts (see left spectra in Fig. 5), the C]N vibrations (yC]N) of li-
gands BFI-1, BFI-2, BFI-3, BFI-4 and FI were observed at 1616, 1611,
1605, 1618 and 1615 cm-1. After coordination to titanium, all the C]
N vibrations displayed a similar blue-shift. In MNTi-1~3, MNTi-5
and FI-Ti, all the C]N vibrations shifted to the same position at
1554 cm-1, while the C]N vibration of MNTi-4 shifted to 1551 cm-1.
This may be due to MNTi-4's linker (methylene) which make the
conjugated system (C6H5eC]N) more electro-rich. In the far-IR
spectra of Ti-based catalysts (see right spectra in Fig. 5), the TieO
bond vibrations (yTi-O) of MNTi-1e5 and FI-Ti were observed at 498,
500, 497, 498, 498 and 500 cm-1. The TieCl bond vibrations (yTi-Cl
)
of MNTi-1e5 and FI-Ti were observed at 460, 460, 462, 472, 458 and
458 cm-1. The TieN bond vibrations (yTi-N) of MNTi-1e5 and FI-Ti
were observed at 360, 360, 363, 362, 362 and 365 cm-1. In the
mid-IR spectra of Zr-based catalysts (see left spectra in Fig. 6), it can
be found that MNZr-1~5 and FI-Zr also displayed a similar blue-
Fig. 5. FT-IR of Ti-based multinuclear catalysts and FI-Ti. (Left: mid-IR. Right: far-IR).
Please cite this article in press as: H.-K. Luo, et al., Journal of Organometallic Chemistry (2015), http://dx.doi.org/10.1016/
j.jorganchem.2015.05.036

H.-K. Luo et al. / Journal of Organometallic Chemistry xxx (2015) 1e13
11
Fig. 6. FT-IR of Zr-based multinuclear catalyst and FI-Zr. (Left: mid-IR. Right: far-IR).
Fig. 7. Raman spectra of MNTi-1~3, 5 and FI-Ti.
coordination geometry should result in imine-H shift in 1H NMR. It
is reported that the mono-nuclear FI catalyst have five possible
isomers and it is believed that the most possible active isomer is cis-
I (see Scheme 1) [14,16]. In CDCl3 at RT, we observed one main
isomer and two small isomers of FI-Ti (see Fig. 8), the main isomer
is most possible to be cis-I. For FI-Zr, we observed only one isomer
cis-I in CDCl3 (see Fig. 8). After the mono-nuclear catalyst was
bridged with different strategies or spacers to form multinuclear
catalysts, the bridging spacer between the two NeO coordination
units may change the NeO coordination behavior with metals in
different ways affording a large number of coordination geome-
tries. Accordingly, the imine-H shift in 1H NMR displayed multiple
Please cite this article in press as: H.-K. Luo, et al., Journal of Organometallic Chemistry (2015), http://dx.doi.org/10.1016/
j.jorganchem.2015.05.036

12
H.-K. Luo et al. / Journal of Organometallic Chemistry xxx (2015) 1e13
Fig. 8. Comparison of 1H NMR shifts of imine-H of multinuclear catalysts and FI catalyst.
nature (see Fig. 8). When bridging the imin-N with three different
spacers, the imine-H shift became more and more broad with the
increasing of the rigidity of the spacer (rigidity order:
eC6H5CH2C6H5ꢂ < ꢂC6H5C6H5ꢂ < eC6H5ꢂ) (see Fig. 8). The mul-
tiple imine-H shift of MNTi is similar to that of MNZr with the same
ligand, indicating that the bis-phenoxy-imine ligand is crucial in
determining the coordination geometries. Especially, MNTi-5 and
MNZr-5 using combined linking strategy afforded very complicated
imin-H shifts, hence there are more coordination geometries
formed in this two catalysts.
MNZr-1 have a slightly higher binding energy than the mono-
nuclear FI-Ti/Zr catalysts (Ti2p of MNTi-1: 465.0 eV and 458.8 eV
vs FI-Ti: 464.7 eV and 458.5 eV,
MNZr-1: 185.2 eV and 182.7 eV vs FI-Zr: 184.6 eV and 182.2 eV,
D
¼ þ0.3 ev and þ0.3 ev; Zr3d of
D
¼ þ0.6 eV and þ0.5 eV). In principle, an atom in higher positive
oxidation state displays a higher binding energy due to the extra
coulombic interaction between the ion core and the photoemitted
electron. The higher binding energy of Ti or Zr in multinuclear
catalysts may imply that the metal contributes a little more
d electron cloud to form a stronger (d-p)
p bond with O and N,
Mono-nuclear FI catalyst is a typical fluxional catalyst, isomer
cis-I may transfer to other isomers that are less active. In insertion-
type polymerization systems, the active catalyst forms a closeeion
pair with the MAO-derived counter-anion, which is typically used
in large excess (e.g.1000 equiv). Hence transfer of the supporting
ligand to the aluminum in MAO is easy and results in fast catalyst
deactivation for the mono-nuclear catalysts. Bridging the mono-
nuclear catalyst may have two effects on catalyst structure. One is
restricting in the mobility of coordination unit NeO on metal,
hence reducing the transformation of cis-I to the other isomeric
structures. Another effect is limiting in the catalyst mobility to
decrease the transfer of ligand from Ti/Zr to Al contained in MAO,
this may result in a different interaction of the cationic catalyst and
the anionic MAO counteranion. Both of these two effects can sta-
bilize the catalyst to achieve higher activity and longer lifetime. In
addition, XPS analysis showed that the metal centers of MNTi-1 and
hence the multinuclear catalyst is more stable. On the other hand,
the higher oxidation state means the metal is more electronphilic,
hence it is easier to accept electrons from olefins to form a weak
bond for chain initiation and subsequent chain growth in catalytic
olefin polymerizations.
The multiple imine-H shifts in 1H NMR (see Fig. 8) may also
indicate that a multinuclear catalyst is a mixture of different struc-
tures. Hence the multinuclear catalysts are typical multi-site cata-
lysts behaving like the classical multi-site ZieglereNatta catalyst
used in industry. This is consistent with the GPC results that poly-
ethylene produced with multinuclear catalysts have broader MWD
compared to corresponding mono-nuclear catalyst. In addition the
multinuclear catalysts with different bis-phenoxy-imine ligands
have different structures. At this stage the bridging effect has been
clearly demonstrated that it can improve the catalytic activity and
stability to produce polymers of higher molecular weight, albeit the
Please cite this article in press as: H.-K. Luo, et al., Journal of Organometallic Chemistry (2015), http://dx.doi.org/10.1016/
j.jorganchem.2015.05.036

H.-K. Luo et al. / Journal of Organometallic Chemistry xxx (2015) 1e13
13
detailed catalyst structures are still difficult to be made clearer.
References
[1] T.J. Pullukat, R.E. Hoff, Catal. Rev. Sci. Eng. 41 (1999) 389.
4. Conclusion
[2] G.J.P. Britovsek, V.C. Gibson, D.F. Wass, Angew. Chem. Int. Ed. 38 (1999) 428.
[3] S.D. Ittel, L.K. Johnson, M. Brookhart, Chem. Rev. 100 (2000) 1169.
[4] V.C. Gibson, S.K. Spitzmesser, Chem. Rev. 103 (2003) 283.
[5] W. Kaminsky, J. Polym. Sci. Part A Polym. Chem. 42 (2004) 3911.
[6] J. Zhang, X. Wang, G.-X. Jin, Coord. Chem. Rev. 250 (2006) 95.
[7] C. Bianchini, G. Giambastiani, I.G. Rios, G. Mantovani, A. Meli, A.M. Segarra,
Coord. Chem. Rev. 250 (2006) 1391.
[8] S. Matsui, K. Tsuru, M. Nitabaru, T. Fujita, M. Mitani, Y. Takagi, T. Nakano, EP
0950667 (1999).
[9] N. Matsukawa, EP 1 013 674 (1999).
[10] T. Fujita, Y. Tohi, M. Mitani, S. Matsui, J. Saito, M. Nitabaru, K. Sugi, H. Makio, T.
Tsutsui, US 6 875 718 (2005).
[11] Y. Nakayama, H. Kaneko, H. Bandoh, J. Saito, S. Kojoh, M. Mitani, Y. Suzuki, S.
Matsui, N. Kashiwa, T. Fujita, US 6 897 176 (2005).
[12] Y. Inoue, S. Matsui, Y. Suzuki, Y. Yoshida, Y. Takagi, T. Fujita, T. Matsugi, T.
Nakano, US 7 053 159 (2006).
[13] S. Ishii, M. Mitani, J. Saito, S. Matsuura, N. Matsukawa, K. Tsuru, T. Fujita, US 7
109 284 (2006).
[14] 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.
[15] S. Ishii, M. Mitani, J. Saito, S. Matsuura, S. Kojoh, N. Kashiwa, T. Fujita, Chem.
Lett. (2002) 740.
[16] S. Matsui, Y. Tohi, M. Mitani, J. Saito, H. Makio, H. Tanaka, M. Nitabaru,
T. Nakano, T. Fujita, Chem. Lett. (1999) 1065.
[17] S. Ishii, M. Mitani, J. Saito, S. Matsuura, R. Furuyama, T. Fujita, Sci. Tech. Catal.
(2002) 49.
[18] P. Woodman, P.B. Hitchcock, P. Scott, Chem. Commun. (1996) 2735.
[19] P.D. Knight, A.J. Clarke, B.S. Kimberley, R.A. Jackson, P. Scott, Chem. Commun.
(2002) 352.
[20] P.D. Knight, G. Clarkson, M.L. Hammond, B.S. Kimberley, P. Scott, J. Organomet.
Chem. 690 (2005) 5125.
[21] G.J. Clarkson, V.C. Gibson, P.K.Y. Goh, M.L. Hammond, P.D. Knight, P. Scott,
T.M. Smit, A.J.P. White, D.J. Williams, Dalton Trans. (2006) 5484.
[22] D. Farrusseng, S. Aguado, C. Pinel, Angew. Chem. Int. Ed. 48 (2009) 7502.
[23] J.Y. Lee, O.K. Farha, J. Roberts, K.A. Scheidt, S.T. Nguyen, J.T. Hupp, Chem. Soc.
Rev. 38 (2009) 1450.
[24] M. Seitz, H.G. Alt, J. Mol. Catal. A Chem. 257 (2006) 73.
[25] H.-K. Luo, H. Schumann, J. Mol. Catal. A Chem. 227 (2005) 153.
[26] B.M.L. Dioos, P.A. Jacobs, Appl. Catal. A General 282 (2005) 181.
[27] R.I. Kureshy, N.H. Khan, S.H.R. Abdi, S.T. Patel, R.V. Jasra, Tetrahedron:
Asymmetry 12 (2001) 433.
We developed a new family of multinuclear Group-IV-metal
catalysts with self-assembly between metals and bis-phenoxy-
imine ligands. The bridging spacer between the two phenoxy-
imine units was selected such that tetradentate coordination of
the ligand to a single metal is impossible, and instead the ligand has
to coordinate to two metal atoms to form multinuclear catalysts.
These multinuclear catalysts can be formed via three bridging
strategies, by linking the imine-N, the 4-position of phenolate, or
the mixed linking strategy. FT-IR and Laser Raman studies show
that the multinuclear catalysts have similar metal coordination
environment relative to the corresponding mono-nuclear catalysts.
The multiple 1H NMR shifts of the imine-H are evidence for the
formation of multi-site catalysts comprising of different coordina-
tion geometries. All of these new catalysts displayed much higher
activity, better stability and produced polyethylene of higher MW
than the corresponding mono-nuclear phenoxy-imine catalysts.
The bridging spacer has an evident effect on the catalytic activity,
stability, polyethylene molecular weight and incorporation rate of
propylene into the polyethylene backbone. The multinuclear cata-
lyst with shorter bridging spacer, such as phenyl, afforded higher
propylene incorporation. In addition, multinuclear catalysts with a
biphenyl spacer can reduce reactor fouling in ethylene polymeri-
zations. We hope this study could inspire the development of
highly active multinuclear catalysts with self-assembly coordina-
tion strategy for olefin polymerizations.
Acknowledgments
We acknowledge the financial support from the Institute of
Chemical and Engineering Sciences (Project ICES-05-111004). Many
thanks to Dr. Zi-Yi Zhong of ICES for his kind assistance in XPS
analysis.
[28] K.B.M. Janssen, I. Laquiere, W. Dehaen, R.F. Parton, I.F.J. Vankelecom,
P.A. Jacobs, Tetrahedron: Asymmetry 8 (1997) 3481.
[29] M.R. Salata, T.J. Marks, J. Am. Chem. Soc. 130 (2008) 12.
[30] N. Guo, C.L. Stern, T.J. Marks, J. Am. Chem. Soc. 130 (2008) 2246.
[31] A. Motta, I.L. Fragala, T.J. Marks, J. Am. Chem. Soc. 131 (2009) 3974.
Appendix A. Supplementary data
ꢀ
[32] H.-K. Luo, R.-G. Tang, K.-J. Gao, J. Catal. 210 (2002) 328.
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.jorganchem.2015.05.036.
Please cite this article in press as: H.-K. Luo, et al., Journal of Organometallic Chemistry (2015), http://dx.doi.org/10.1016/
j.jorganchem.2015.05.036