Immobilization and activation of 2,6-bis(imino)pyridyl Fe, Cr and V precatalysts using a MgCl2/AlRn(OEt)3-n support: Effects on polyethylene molecular weight and molecular weight distribution

Article abstract of DOI:10.1016/j.molcata.2006.07.030

Ethylene polymerizations carried out with various bis(imino)pyridyl iron, chromium and vanadium complexes immobilized on a MgCl₂/AlR_n(OEt)_3-n support gave relatively broad polyethylene molecular weight distributions in the case of iron, but high molecular weight and a very narrow molecular weight distribution with vanadium, indicative of a single active species. The narrow MWD was confirmed by melt rheometry. Similar results were obtained after reaction of the bis(imino)pyridyl complex LVCl₃ (6) with MeLi or AlEt₃, where alkylation of the pyridine ring gives a complex L′VCl₂ (7). In the case of chromium, a bimodal distribution was obtained, with evidence of incomplete catalyst immobilization. The polyethylene molecular weights obtained with the iron complexes were strongly dependent on the substituents in the bis(imino)pyridyl ligand, and were somewhat higher than have been obtained in homogeneous polymerization. In contrast, the molecular weights obtained with the bis(imino)pyridyl chromium and vanadium complexes were much higher that those previously obtained under homogeneous conditions. In all cases, the activities of the immobilized catalysts were higher than those found in homogeneous polymerization.

Full text of DOI:10.1016/j.molcata.2006.07.030

Journal of Molecular Catalysis A: Chemical 260 (2006) 135–143
Immobilization and activation of 2,6-bis(imino)pyridyl Fe, Cr and V
precatalysts using a MgCl2/AlRn(OEt)3−n support: Effects on
polyethylene molecular weight and molecular weight distribution
Rubin Huanga,b, Nileshkumar Kukalyekara,b, Cor E. Koninga,b, John C. Chadwicka,b,∗
a Laboratory of Polymer Chemistry, Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven, The Netherlands
b Dutch Polymer Institute (DPI), P.O. Box 902, 5600 AX Eindhoven, The Netherlands
Received 24 April 2006; received in revised form 13 June 2006; accepted 12 July 2006
Available online 22 August 2006
Abstract
Ethylene polymerizations carried out with various bis(imino)pyridyl iron, chromium and vanadium complexes immobilized on a
MgCl2/AlRn(OEt)3−n support gave relatively broad polyethylene molecular weight distributions in the case of iron, but high molecular weight
and a very narrow molecular weight distribution with vanadium, indicative of a single active species. The narrow MWD was confirmed by melt
rheometry. Similar results were obtained after reaction of the bis(imino)pyridyl complex LVCl3 (6) with MeLi or AlEt3, where alkylation of
the pyridine ring gives a complex LꢀVCl2 (7). In the case of chromium, a bimodal distribution was obtained, with evidence of incomplete cat-
alyst immobilization. The polyethylene molecular weights obtained with the iron complexes were strongly dependent on the substituents in the
bis(imino)pyridyl ligand, and were somewhat higher than have been obtained in homogeneous polymerization. In contrast, the molecular weights
obtained with the bis(imino)pyridyl chromium and vanadium complexes were much higher that those previously obtained under homogeneous
conditions. In all cases, the activities of the immobilized catalysts were higher than those found in homogeneous polymerization.
© 2006 Elsevier B.V. All rights reserved.
Keywords: Catalysis; Ethylene; Immobilization; Magnesium chloride; Polymerization
1. Introduction
of active species is operative [4]. Strong evidence for the
presence of different active species has recently been provided
by Barabanov et al. [5], who used 14CO radiotagging to
determine the numbers of active centres and propagation
rate constants in homogeneous polymerization. The results
obtained indicated the presence of highly reactive but unstable
active centres producing a low molecular weight polymer
fraction, as well as less active but more stable species producing
higher molecular weight polymer. Iron-based precatalysts can
be activated by both MAO and aluminium trialkyls and it
has been reported that narrow molecular weight distribution
can be obtained using AliBu3 [6,7] or iBu2AlOAliBu2 [8].
Following these studies, carried out under homogeneous
polymerization conditions, several groups investigated the
immobilization and activation of bis(imino)pyridyl iron pre-
catalysts on various supports, including silica [9–11] and
magnesium chloride [12,13]. The MgCl2-supported systems
consistently gave broad polyethylene molecular weight distri-
bution, irrespective of the type of aluminium trialkyl used as
cocatalyst.
Bis(imino)pyridyl iron complexes of type {2,6-
[ArN C(Me)]2C5H3N}FeCl2 have been shown to be
very active precatalysts for ethylene polymerization [1,2].
However, in contrast to the great majority of homogeneous
olefin precatalysts, these systems typically give polyethylene
with relatively broad molecular weight distribution. In many
cases a bimodal molecular weight distribution is obtained and
evidence has been presented that, for systems activated with
methylaluminoxane (MAO), this is caused by the formation of
a low molecular weight fraction resulting from chain transfer
to aluminium, particularly in the early stages of polymerization
[3]. The question remains as to whether chain transfer to
aluminium is the only reason for the broad polydispersities
obtained with these systems, or whether more than one type
∗
Corresponding author. Tel.: +31 40 2475378; fax: +31 40 2463966.
E-mail address: j.c.chadwick@polymers.nl (J.C. Chadwick).
1381-1169/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2006.07.030

136
R. Huang et al. / Journal of Molecular Catalysis A: Chemical 260 (2006) 135–143
Recently, we and others have investigated the immobilization
ers [21]. {2,6-[ArN C(Me)]2(2-MeC5H3N)}VCl2 (7; Ar = 2,6-
diisopropylphenyl) was prepared by the reaction of 6 with MeLi
in toluene as described by Gambarotta and coworkers [21]; after
solvent removal under vacuum, the residue was redissolved in
ether and after filtration the ether was removed under vacuum
to yield a dark green solid.
AlEt3 (25 wt.% solution in toluene) and ZnEt2 (1.0 M solu-
tion in hexane) were purchased from Aldrich. AliBu3 (1 M
solution in hexane) and MAO (25 wt.% solution in toluene) were
purchased from Fluka and Akzo Nobel, respectively.
of a range of early- and late-transition metal precatalysts using
supports of type MgCl2/AlRn(OR)3−n, prepared by reaction of
AlR3 with either solid, spherical adducts of MgCl2 and ethanol
[12,14] or adducts of MgCl2 and 2-ethylhexanol in hydrocar-
bon solution [15]. Widespread implementation of homogeneous
and single-site catalysts in polyolefin production, especially in
gas-phase and slurry processes, is dependent on the develop-
ment of effective techniques for catalyst immobilization and
considerable research is being carried out in this field [16].
Our own studies have concentrated on the use of supports typi-
cally prepared by the reaction of AlEt3 with spherical, partially
dealcoholated adducts of MgCl2 and EtOH [17]. The spherical
particlemorphologyofthesupportisretainedandreplicateddur-
ing catalyst immobilization and polymerization, leading to the
formation of spherical polymer particles without reactor fouling.
In the present work, we have investigated the immobilization
and activation of different bis(imino)pyridyl iron(II) precata-
lysts, including the effect of cocatalyst type and concentration.
In order to assess whether with these systems the polyethylene
polydispersity is dependent on the transition metal, the ligand or
the metal/ligand combination, bis(imino)pyridyl chromium(II)
and vanadium(III) precatalysts have also been immobilized and
activated using the same MgCl2/AlRn(OEt)3−n support.
Ethylene (3.5 grade supplied by Air Liquide) was purified by
˚
passing over columns of 4 A molecular sieves and BTS copper
catalyst.
2.2. Support preparation and catalyst immobilization
Support preparation was performed by the addition of AlEt3
to a slurry of an adduct MgCl2·1.1EtOH (average particle size
d50 82 ␮m) in light petroleum (AlEt3/EtOH = 2) at 0 ◦C, after
which the mixture was kept at room temperature for 2 days
withoccasionalagitation. Theresultantsupportwaswashedwith
light petroleum three times and dried under argon flow and sub-
sequently under vacuum until free flowing. The Al contents of
the support were determined by the H. Kolbe Microanalytisches
¨
Laboratorium, Mulheim an der Ruhr, Germany. The ethoxide
content in MgCl2/AlEtn(OEt)3−n support was determined by
gas chromatography (GC) analysis of the ethanol content of a
solution obtained by dissolving 100 mg of support in 5 mL of
BuOH containing a known quantity of PrOH as an internal stan-
dard. The Al and OEt contents of the support were 3.89 and
4.89 wt.%, respectively, indicating an overall support composi-
2. Experimental
2.1. Materials
All manipulations were performed under an argon atmo-
sphere using glove box (Braun MB-150 G1 or LM-130)
and Schlenk techniques. Light petroleum (b.p. 40–60 ◦C) and
dichloromethane were passed over a column containing Al2O3
and stored over 4 A molecular sieves. All the solvents were
freeze-thaw degassed at least twice prior to use.
Precatalysts 1–4 (structures shown in Scheme 1) were pre-
pared according to various literature procedures [3,18,19]. {2,6-
[ArN C(Me)]2C5H3N}CrCl2 (5; Ar = 2,6-diisopropylphenyl)
and {2,6-[ArN C(Me)]2C5H3N}VCl3 (6; Ar = 2,6-diisopro-
pylphenyl) were prepared following procedures similar to those
described by Devore et al. [20] and by Gambarotta and cowork-
tion MgCl2·0.17AlEt2.25(OEt)0.75
.
Catalyst immobilization was effected by mixing the sup-
port (50–100 mg) with a precatalyst solution in dichloromethane
(2 mL, containing 0.5–1.0 ␮mol of precatalyst) and keeping at
room temperature overnight. The slurry of the immobilized cat-
alyst in dichloromethane was diluted with light petroleum and
used directly in ethylene polymerization.
˚
2.3. Polymerization procedure
Polymerization was carried out in a l L Premex autoclave
by charging the immobilized catalyst (50–100 mg, containing
0.5–1.0 ␮mol precatalyst), slurried in approximately 100 mL
light petroleum, to 400 mL light petroleum containing the
desired amount of cocatalyst, at 50 ◦C and an ethylene pressure
of 5 bar. After catalyst injection, polymerization was continued
at constant pressure for 1 h and with a stirring rate of around
1000 rpm. After venting the reactor, 20 mL of acidified ethanol
were added and stirring was continued for 30 min. The polymer
was recovered by filtration, washed with water and ethanol and
dried in vacuo overnight at 60 ◦C.
2.4. Polymer characterization
Molecular weights and molecular weight distributions of the
resulting polymers were determined by means of gel perme-
Scheme 1. Structure of bis(imino)pyridyl metal (Fe, Cr and V) precatalysts.

R. Huang et al. / Journal of Molecular Catalysis A: Chemical 260 (2006) 135–143
137
ation chromatography on a PL-GPC210 at 135 ◦C using 1,2,4-
trichlorobenzene as solvent. Melting points of the polymers
were obtained on a Q100 (TA Instruments) DSC in the stan-
dard DSC run mode. The particle morphologies of the polymers
were examined using a Philips S-250MK3 SEM-EDX.
Oscillatory shear measurements were performed on a Rheo-
metrics ARES melt rheometer. Strain controlled frequency
sweep tests were carried out at 180 ◦C over an angular frequency
range from 0.01 to 100 rad s−1. The constant dynamic strain
applied on a compressed PE plate of diameter 8 mm was 0.5%.
The measurements were performed using parallel plate geome-
try, maintaining a distance between the plates of about 0.5 mm.
All measurements were carried out under nitrogen atmosphere
on polymers stabilized with 0.1 wt.% Irganox 1010 (added as
a solution in acetone, after which the polymer was dried under
reduced pressure).
in light petroleum. Because of the limited solubility of the pre-
catalysts in hydrocarbon solvents, immobilization was carried
out at ambient temperature by contact of the support with a
dichloromethane solution of the precatalyst. A relatively low
precatalyst loading of 10 ␮mol/g support was used, our previ-
ous studies with other catalysts having shown that particularly
high catalyst activities can be obtained at low loadings [14b].
Quantitative catalyst immobilization was indicated by complete
discolouration of the liquid phase after contacting overnight, to
give sand-coloured supported catalysts. We found that it was not
necessary to remove the dichloromethane from the immobilized
catalyst slurry prior to polymerization. The slurry was simply
diluted with light petroleum and injected as such into a polymer-
ization reactor containing light petroleum and different amounts
of cocatalyst.
In order to determine the effect of cocatalyst type and concen-
tration on catalyst performance and in particular on the molec-
ular weight and molecular weight distribution of the polymers
obtained, polymerizations were carried out in the presence of
AlEt3, AliBu3 or ZnEt2. It has been reported by Gibson and
coworkers that, under homogeneous polymerization conditions,
low molecular weight polymers having very narrow polydis-
persity could be obtained in the presence of diethyl zinc [23].
Under such conditions, the chain growth process is characterized
by an exceptionally rapid and reversible exchange of the grow-
ing polymer chains between iron and zinc. We were interested
to see whether this effect could take place in a heterogeneous
system, or whether limited molecular mobility would restrict the
reversibility of chain transfer between iron and zinc.
3. Results and discussion
3.1. Bis(imino)pyridyl iron precatalysts
Our first studies were directed at the immobilization and acti-
vation of the bis(imino)pyridyl iron complexes 1–3. Previous
results have shown that high polymerization activity and good
polymer particle morphology could be obtained with precat-
alysts 1 and 2, using MgCl2-based supports similar to those
used in the present work [12]. Precatalyst 3 was selected in
order to investigate the effect on polyethylene molecular weight
of chloro substitution in the aryl rings of the ligand, it having
been reported that the electron-withdrawing effect of halogen-
containing ligands leads to a lowering of product molecular
weightinethylenepolymerizationandoligomerization[22]. The
immobilization of these precatalysts was carried out using a sup-
port of composition MgCl2·0.17AlEt2.25(OEt)0.75, prepared by
reaction of AlEt3 with a slurry of an adduct MgCl2·1.1EtOH
The results of the ethylene polymerizations carried out are
given in Tables 1–3. In order to ensure effective scavenging
of impurities, a cocatalyst concentration of at least 1 mmol/L
was used. These polymerizations were carried out at low cat-
alyst concentration (2 ␮mol/L), resulting in Al/Fe (or Zn/Fe)
molar ratios ≥500. It is apparent that very high catalyst activities
Table 1
Effect of cocatalyst in ethylene polymerization with {2,6-[ArN C(Me)]2C5H3N}FeCl2 (1; Ar = 2,4,6-trimethylphenyl) immobilized on a MgCl2/AlRn(OEt)3−n
support
Activity (kg mol−1 Fe bar−1 h−1
)
¯
Mw (g/mol)
¯
Mn (g/mol)
¯ ¯
Mw/Mn
Experiment
Cocatalyst
Al/Fe (mol/mol)
Zn/Fe (mol/mol)
1
2
3
4
5
6
AlEt3
AlEt3
AliBu3
AliBu3
ZnEt2
ZnEt2
500
1000
500
1000
0
0
0
0
17400
15090
4840
10000
1470
435800
425600
425800
418900
163300
144900
44400
38500
38500
36100
30100
24700
9.8
11.1
11.1
11.6
5.4
0
500
1000
0
1800
5.9
Table 2
Effect of cocatalyst in ethylene polymerization with {2,6-[ArN C(Me)]2C5H3N}FeCl2 (2; Ar = 2,6-diisopropylphenyl) immobilized on a MgCl2/AlRn(OEt)3−n
support
Activity (kg mol−1 Fe bar−1 h−1
)
¯
Mw (g/mol)
¯
Mn (g/mol)
¯
Mw/Mn
¯
Experiment
Cocatalyst
Al/Fe (mol/mol)
Zn/Fe (mol/mol)
1
2
3
4
5
6
AlEt3
AlEt3
AliBu3
AliBu3
ZnEt2
ZnEt2
500
1000
500
1000
0
0
0
0
4620
5200
1040
2600
960
748600
771400
403600
458400
275200
320900
133100
146000
103700
123400
41500
5.6
5.3
3.9
3.7
6.6
6.9
0
500
1000
0
2200
46400

138
R. Huang et al. / Journal of Molecular Catalysis A: Chemical 260 (2006) 135–143
Table 3
Effect of cocatalyst in ethylene polymerization with {2,6-[ArN C(Me)]2C5H3N}FeCl2 (3; Ar = 2-chloro-4,6-dimethylphenyl) immobilized on
a
MgCl2/AlRn(OEt)3−n support
Activity (kg mol−1 Fe bar−1 h−1
)
¯
Mw
¯
Mn
¯
Mw/Mn
¯
Experiment
Cocatalyst
Al/Fe (mol/mol)
Zn/Fe (mol/mol)
1
2
3
4
5
6
AlEt3
AlEt3
AliBu3
AliBu3
ZnEt2
ZnEt2
500
1000
500
1000
0
0
0
0
8200
7480
2320
4320
1120
1920
102100
103600
108800
103200
84300
17900
13400
13100
17100
13100
12300
5.7
7.7
8.3
6.0
6.4
6.5
0
500
1000
0
80500
were obtained, most notably with precatalyst 1 in combination
with AlEt3. This precatalyst also gave relatively broad molecu-
lar weight distribution, with a polydispersity of around 10–12,
but no narrowing in molecular weight distribution was observed
when AliBu3 rather than AlEt3 was used as cocatalyst. It is also
seen that a doubling in AlR3 concentration had relatively lit-
tle effect on polyethylene molecular weight, an indication that
chain transfer to aluminium was not the reason for the broad
polydispersity in this system. There was also no indication of the
bimodality typically observed in homogeneous, MAO-activated
polymerization, in which the formation of the low molecular
weight peak is ascribed to chain transfer to aluminium [1,3].
Ethylene polymerization with precatalyst 1 in combination with
diethyl zinc gave lower but still very appreciable activity and
also significantly lower molecular weights, confirming the effec-
tiveness of ZnEt2 as chain transfer agent with the immobilized
catalyst. However, polydispersitiesintherange5–6indicatethat,
in contrast to homogeneous polymerization [23], the reversibil-
ity of chain transfer between iron and zinc is limited, no doubt as
a result of restricted chain mobility in this heterogeneous system.
The results with precatalyst 2 show similar trends, but in
this case somewhat narrower molecular weight distributions
were obtained, with AlEt3 and AliBu3 giving polydispersities
of 5.3–5.6 and 3.7–3.9, respectively. Precatalyst 3 gave rela-
tively low molecular weights in comparison to precatalysts 1
and 2, comparison of catalysts 1 and 3 indeed showing that the
replacement of a 2-methyl by a 2-chloro substituent in each
aryl ring of the ligand leads to a significant molecular weight
reduction. Again, no significant effect of cocatalyst concentra-
tion on molecular weight or molecular weight distribution is
apparent. However, the results in Tables 1–3 indicate that, for
most of the systems investigated and particularly when using
AliBu3 or ZnEt2, an increase in metal alkyl concentration led to
increased catalyst activity. This could indicate that with these
alkyls the generation of the catalytically active species was
less efficient than was the case with AlEt3. AliBu3 has been
reported to be a less efficient activator than AlEt3 for precat-
alyst 2 in homogeneous polymerization [7]. It is likely that,
in the MgCl2-immobilized system, the different activities are
due to differences in the alkylating capacity of the cocatalyst.
The AlEtn(OEt)3−n present in the support is not likely to play
a significant role, although in previous work we have observed
that Cp2TiCl2 could be activated using MgCl2/AlEtn(OEt)3−n in
the absence of any additional cocatalyst [14a]. The precise role
of magnesium chloride in the immobilization and activation of
early- and late-transition metal catalysts is as yet unresolved but
will depend on the nature of the catalyst. In the case of metal-
locenes which are not easily activated using AlR3, Lewis acidic
sites on the support surface may participate in the formation of
cationic active species, as has been described by Marks [24].
This may be less important for the bis(imino)pyridyl iron cata-
lysts, which can be activated with simple aluminium alkyls, but a
general advantage of the use of a MgCl2 support for both early-
and late-transition metal systems is stabilization of the active
species (vide infra), preventing the rapid decay often observed
in homogeneous polymerization.
Comparison of the results obtained with precatalysts 1 and
2 shows that the presence of isopropyl as opposed to methyl
substituents in the 2,6-positions of the aryl groups in the ligand
leads, as has been noted previously [3,12b], to somewhat lower
activities. However, at the same time a significant narrowing
in polyethylene molecular weight distribution is obtained, using
either AlEt3 or AliBu3 as cocatalyst. Comparison of precatalysts
1 and 3 shows that the replacement of a methyl by a chloro sub-
stituent leads to somewhat lower activity as well as to the large
decrease in polymer molecular weight. Lower activity following
the substitution of a methyl group by a chlorine atom on the ortho
position of the aryl groups has also been noted in homogeneous
polymerization [25]. The steric bulk of a chloro substituent is
very similar to that of a methyl group, so the different behaviours
of catalyst 3 must arise from the electron-withdrawing nature of
the halogen substituent, which should diminish the electron den-
sity on the metal. The DSC melting temperatures (Tm2) of the
polymers prepared as in Tables 1–3 were all above 133 ◦C, indi-
cating the formation of linear, high-density polyethylene. The
melting enthalpies of the polymers prepared with precatalysts 1,
2 and 3 were in the ranges 196–212, 176–188 and 206–231 J/g,
respectively, indicating an increase in crystallinity with decreas-
ing molecular weight. Relatively stable activity was observed in
all of the above polymerizations. Typical plots are shown in
Fig. 1, from which only a moderate decay in activity is appar-
ent from changes in monomer mass flow and the temperature
difference between the reactor contents and the cooling mantle.
This result is a further illustration of the stabilizing effect of a
MgCl2 support [12,13] and is in striking contrast to the rapid
decay typically observed with these catalysts under homoge-
neous polymerization conditions [3,5,6].
The polymer molecular weights obtained with the MgCl2-
immobilized systems listed in Table 1 are higher than the
molecular weights obtained with these precatalysts under homo-
geneous conditions [3], but the difference is relatively small in
comparison to the very large increases in molecular weight fre-

R. Huang et al. / Journal of Molecular Catalysis A: Chemical 260 (2006) 135–143
139
3.2. Bis(imino)pyridyl chromium and vanadium
precatalysts
Tridentate bis(imino)pyridyl ligands have been widely used
in the preparation of highly active iron- and cobalt-based
catalysts for olefin polymerization, but there are fewer exam-
ples of their use in the preparation of early-transition metal
precatalysts. However, bis(imino)pyridyl chromium [20,27,28]
and vanadium [21,29] complexes have been synthesized and
have been shown to be active ethylene oligomerization and
polymerization catalysts when activated by methylaluminox-
anes under homogeneous conditions. Esteruelas et al. [27]
reported the formation of waxes and low molecular weight
polyethylene using {2,6-[ArN C(Me)]2C5H3N}CrCl2 or {2,6-
[ArN C(Me)]2C5H3N}CrCl3
(Ar = 2,6-diisopropylphenyl).
Small et al. [28] reported the formation of mixtures of
oligomers and polymers using a range of bis(imino)pyridyl
chromium complexes and provided evidence that the active
species derived from Cr(II) and Cr(III) complexes were
identical. Gambarotta and coworkers [21] investigated the
preparation of {2,6-[ArN C(Me)]2C5H3N}VCl3 (Ar = 2,6-
diisopropylphenyl), which after activation with MAO produced
polyethylenes having high polydispersities. GPC analysis
revealed bimodal distributions, indicating the formation of
different active species. Schmidt et al. [29] prepared a series of
bis(imino)pyridyl V(III) complexes, most of which gave low
molecular weight oligomers.
In order to determine whether bis(imino)pyridyl chromium
and vanadium precatalysts could be effectively immobilized and
activated using a MgCl2/AlRn(OEt)3−n support, and to deter-
mine whether or not single-site catalyst characteristics and nar-
row polymer molecular weight distribution could be obtained,
theCr(II)andV(III)complexes5and6weresynthesized. Precat-
alyst immobilization was carried out in dichloromethane under
conditions similar to those used for the bis(imino)pyridyl iron
precatalysts. The ethylene polymerization results obtained using
the Cr(II) precatalyst 5 are summarized in Table 4. The catalyst
activities are much lower than those achieved with the corre-
sponding iron catalyst 2, and reveal little effect of the cocatalyst.
GPC analysis of the polymers obtained revealed broad, bimodal
molecular weight distributions, irrespective of the cocatalyst
Fig. 1. Mass flow and temperature profiles in ethylene polymerization using
complex 1 immobilized on a MgCl2/AlEtn(OEt)3−n support and activated with
AlEt3 (ꢀ: internal temperature; ᭹: cooling mantle temperature).
quently observed when immobilizing various early-transition
metal precatalysts on similar supports [14a,b]. In order to fur-
ther assess the effect of precatalyst immobilization on product
molecular weight, an additional experiment was carried out with
precatalyst 4, which represents a new generation of ethylene
oligomerization catalysts, illustrating the very important role
that ligand steric effects play in transition metal-catalyzed olefin
polymerization and oligomerization [19]. In this experiment
AlEt3 (Al/Fe = 2000) was used as cocatalyst, resulting in the
formation of an oligomeric (light petroleum-soluble) fraction
and a solid, polymeric fraction, in the proportions 8.7:1, with
¯
used. The Mw values are, however, one to two orders of magni-
tudehigherthanthosereportedforhomogeneouspolymerization
using this complex [27], indicating a significant effect of immo-
bilization on MgCl2. The formation of a low molecular weight
fraction, inadditiontothemain(highmolecularweight)polymer
fraction, could arise from the presence of more than one active
species, possibly as a result of incomplete immobilization of the
catalyst on the support. In order to test this possibility, we carried
out a polymerization in toluene rather than light petroleum, tak-
ing into account that leaching of the catalyst into solution should
be more prevalent in toluene. The GPC traces of polymers pre-
pared in light petroleum and in toluene are given in Fig. 2, which
indeed shows that polymerization in toluene increased the pro-
portion of the low molecular weight component.
an overall catalyst activity of 36 720 kg mol−1 Fe bar−1 h−1
.
1H NMR analysis of the oligomeric fraction indicated that
the product comprised mainly (91%) vinyl-terminated chains,
¯
with Mn approximately 250. This result shows that precata-
lyst 4 remains predominantly an oligomerization catalyst when
immobilized on magnesium chloride, a further indication that
the effect on product molecular weight of immobilizing a
bis(imino)pyridyl iron catalyst on a MgCl2-based support is
relatively small. The formation of oligomeric and polymeric
fractions in MAO-activated homogeneous polymerization with
the closely related catalyst {2,6-[ArN C(Me)]2C5H3N}FeCl2
(Ar = 2-ethylphenyl) has been reported by Quijada et al. [26].
The polymerization results obtained using the vanadium(III)
precatalyst 6 are given in Table 5. It is immediately obvious

140
R. Huang et al. / Journal of Molecular Catalysis A: Chemical 260 (2006) 135–143
Table 4
Effect of cocatalyst in ethylene polymerization with {2,6-[ArN C(Me)]2C5H3N}CrCl2 (5; Ar = 2,6-diisopropylphenyl) immobilized on a MgCl2/AlRn(OEt)3−n
support
Activity (kg mol−1 Cr bar−1 h−1
)
¯
Mw (g/mol)
¯
Mn (g/mol)
¯ ¯
Mw/Mn
Experiment
Cocatalyst
Al/Cr (mol/mol)
1
2
3
4
5
6a
AlEt3
AlEt3
AlEt3
AliBu3
MAO
AlEt3
500
1000
2000
1000
1000
500
220
220
240
160
140
320
690000
1111000
635000
967000
1070000
637000
48000
38700
56300
43200
30700
30000
14.4
28.7
11.3
22.4
34.9
21.2
a
Polymerization in toluene.
Fig. 2. Effect of polymerization solvent on the MWD of PE produced with
complex (5) immobilized on a MgCl2/AlEtn(OEt)3−n support and activated with
AlEt3 (the curve numbers correspond to the experiment numbers in Table 4).
Fig. 3. Shear frequency dependence of storage modulus (Gꢀ) at constant strain
¯
¯
of (ꢁ and ꢀ) polyethylene (Mw/Mn = 1.8–2.1) prepared using an immobilized
¯
¯
V catalyst, compared to (ꢂ) a reference polyethylene (Mw/Mn = 5.6) prepared
using an immobilized iron catalyst.
that immobilization of this catalyst on the MgCl2 support gives,
in contrast to the iron- and chromium-based systems, not only
high polyethylene molecular weight but also very narrow molec-
ular weight distribution. The presence of a single active species
is indicated by the narrow (Schulz–Flory) distributions, with
try is a very effective tool for the qualitative comparison of the
molecular weight distributions in high molecular weight poly-
mers[14d]. TheresultsinFig. 3showaconstantstoragemodulus
over the entire frequency range for the V-derived polymers,
characteristic of polyethylene having narrow molecular weight
distribution [30]. In contrast, the decrease in storage modulus
with decreasing frequency for the Fe-derived polymer illustrates
the much broader molecular weight distribution of this polymer.
The successful immobilization and activation of precat-
alyst 6 contrasts with previous, unsuccessful attempts to
heterogenize this and related V(III) precatalysts on a silica
support, which resulted in little or no activity [29]. The activ-
ities of 500–560 kg mol−1 V bar−1 h−1 obtained with MgCl2-
¯
¯
Mw/Mn close to 2. In order to obtain further confirmation of the
narrow molecular weight distributions of the polyethylenes syn-
thesized with the MgCl2-immobilized bis(imino)pyridyl vana-
dium precatalyst, the melt rheological properties of polymers
prepared using either AlEt3 or AliBu3 as cocatalyst were com-
¯
¯
pared with those of a polymer having Mw/Mn = 5.6, prepared
(Table 2, experiment 1) using the related bis(imino)pyridyl iron
catalyst 2. Oscillatory shear measurements were carried out at
180 ◦C to determine the dependence of the storage modulus, Gꢀ,
on shear. As we have shown in previous studies, melt rheome-
Table 5
Effectofcocatalystinethylenepolymerizationwith{2,6-[ArN C(Me)]2C5H3N}VCl3 (6;Ar = 2,6-diisopropylphenyl)and{2,6-[ArN C(Me)]2(2-MeC5H3N)}VCl2
(7; Ar = 2,6-diisopropylphenyl) immobilized on a MgCl2/AlRn(OEt)3−n support
Cocatalysta
Temperature (◦C)
Activity (kg mol−1 V bar−1 h−1
)
¯
Mw (g/mol)
¯
Mn (g/mol)
¯ ¯
Mw/Mn
Experiment
Catalyst
1
2
3
4
5
6
6
6
6
6
7
7
AlEt3
AlEt3
AliBu3
MAO
AlEt3
AlEt3
50
70
50
50
50
70
560
2140
500
560
320
1770000
1961000
1090000
1050000
1531000
2040000
848000
1205000
590000
534000
1053000
1363000
2.1
1.6
1.8
2.0
1.5
1.5
907
a
Al/V molar ratio = 1000.

R. Huang et al. / Journal of Molecular Catalysis A: Chemical 260 (2006) 135–143
141
immobilized precatalyst 6 at 50 ◦C are higher than those reported
[21] for homogeneous polymerizations with MAO activation,
where it was also noted that rapid deactivation took place when
the temperature was increased to 140 ◦C. Rapid deactivation
during polymerization is a characteristic feature of many homo-
geneous vanadium catalysts [31], but very effective stabilization
can be achieved by immobilization on magnesium chloride.
This has been demonstrated by Nakayama et al. [15b], who
obtained very stable polymerization kinetics even at 75 ◦C using
a MgCl2-based support/activator in combination with an FI–V
complex, and recently we have observed stable polymerization
activity with an immobilized vanadium(III) amidinate catalyst
[14e]. In the present work, we carried out a polymerization with
MgCl2-immobilized precatalyst 6 at 70 ◦C. No deactivation was
apparent during the course of polymerization and the activity
(Table 5) was significantly higher than that obtained at 50 ◦C.
The above polymerization studies consistently gave spheri-
cal polymer particle morphology, with no evidence of reactor
fouling. Scanning electron micrographs of polymers produced
with immobilized bis(imino)pyridyl Fe(II), Cr(II) and V(III)
precatalysts are shown in Fig. 4, demonstrating that the spher-
ical morphology of the original support had been retained and
replicated throughout catalyst immobilization and polymeriza-
tion. It is, however, apparent that the particle morphology of the
Cr-PE is inferior to that of the polymers obtained with the iron
and vanadium catalysts, which may be a further indication of
incomplete catalyst immobilization in the case of the chromium
catalyst.
maximum amount of precatalyst that can be immobilized on
MgCl2 is similar to the amount of TiCl4 that can be adsorbed,
suggesting that these components adsorb on the same surface
acidic sites [13]. However, in this system good polymer par-
ticle morphology and absence of fouling do not necessarily
provide proof of strong anchoring of the precatalyst to the
support during polymerization, as the limited hydrocarbon sol-
ubility of the catalyst would tend to prevent leaching from the
support.
The possibility of side reactions involving the bis(imino)-
pyridine ligand when attached to different metals has been inves-
tigated by several groups and is a complex issue [32], although
Gibson and coworkers have reported that the ligand could
be recovered intact after MAO-activated ethylene polymeriza-
tions with the bis(imino)pyridyl iron complex 2 [3]. Reactions
between the free ligand and various lithium, magnesium and
zinc alkyls have been shown to give N-alkylated products aris-
ing from attack of the metal alkyl at the pyridine nitrogen atom
[33]. Very recent work by Budzelaar and coworkers [34] has
revealed that reactions of the free ligand with aluminium alkyls
are unexpectedly complex, alkyl additions to the imine carbon
and at the 2- and 4-positions of the pyridine ring being observed.
It has also been shown that the iron atom in a bis(imino)pyridyl
iron complex can be displaced by reaction with an aluminium
alkyl, but the resulting Al complex was catalytically inactive,
indicating that this reaction represents a deactivation pathway
[35]. Gambarotta and coworkers have investigated reactions of
various other bis(imino)pyridyl transition metal complexes with
alkylating agents. Ligand coupling via deprotonation of a methyl
group attached to the imino function was observed for a man-
ganese complex, while reaction of LCrCl3 with AlMe3 resulted
in reduction to LCr(␮-ClAlMe3)2 [36]. An unusual result was
obtained when the corresponding vanadium complex LVCl3 (6)
was reacted with excess MAO or MeLi. In this case, alkylation of
the pyridine ring took place, leading to the formation of species
7, as illustrated in Scheme 2 [21]. This reaction, leading to a
decrease in the metal coordination number, was suggested to be
a key factor contributing to the polymerization activity obtained
using complex 6 in combination with MAO and it was found
that complex 7 indeed gave activities and polymer characteris-
tics that were very similar to those obtained using complex 6
[21].
It is evident from the above studies that MgCl2-immobilized
bis(imino)pyridyl iron, chromium and vanadium precatalysts
exhibit very different characteristics in ethylene polymerization.
The iron systems give polyethylenes with molecular weight dis-
¯
¯
tributions (Mw/Mn) in the range 3–12, dependent on the steric
bulk of the bis(imino)pyridyl ligand. However, there was no evi-
dence of the low molecular weight peak typically observed in
polymers prepared under homogeneous conditions with MAO,
which has been ascribed to chain transfer to aluminium [3],
or when AlMe3 is used as cocatalyst with a MgCl2-supported
iron catalyst [13]. Changes in cocatalyst concentration had lit-
tle effect on molecular weight or molecular weight distribu-
tion and it therefore appears that these systems contain more
than one active species. The mode of coordination of the iron
species LFeCl2 to the support is likely to involve bridged
chloride species, Mikenas et al. having determined that the
Inordertocomparetheactivespeciesformedfromcomplexes
6 and 7 on a MgCl2-based support, we synthesized and immo-
Fig. 4. Scanning electron micrographs of polyethylene prepared using bis(imino)pyridyl iron, chromium and vanadium catalysts immobilized on a
MgCl2/AlEtn(OEt)3−n support: (a) Fe complex 3; (b) Cr complex 5; (c) V complex 6.

142
R. Huang et al. / Journal of Molecular Catalysis A: Chemical 260 (2006) 135–143
Scheme 2. Reaction of precatalyst 6 with MeLi [21].
Fig. 5. UV–vis spectra of complexes 6 and 7 in dichloromethane (a) before and (b) after contact with AlEt3.
bilized complex 7. During the immobilization of complexes 6
and 7 it was observed that the colour changed from light red
(6) or light green (7) in the starting solutions to give the same
sand colour in the final supported catalysts. Comparison of the
results of ethylene polymerizations carried out with each com-
plex (Table 5) reveals somewhat lower activities for complex
7, but the polyethylene molecular weights were very similar to
those obtained with complex 6, an indication that very similar
types of active species were present in each system. Further evi-
dence for the formation of similar active species was obtained
by comparing the UV spectra of complexes 6 and 7 in solu-
tion in dichloromethane, before and after contact with AlEt3
(Al/V = 100). Fig. 5a reveals a significant difference between the
spectra of complexes 6 and 7, but the similarity of the spectra
in Fig. 5b indicates that similar species are generated on contact
with triethylaluminium. It is therefore likely that, in the poly-
merizations carried out with MgCl2-immobilized complex 6 in
the presence of AlEt3, alkylation takes place to give a complex
very similar to 7, but with an ethyl rather than a methyl group
at the 2-position of the pyridine ring.
molecular weights obtained with the iron catalysts are somewhat
higher than those reported for homogeneous polymerization.
In contrast, immobilization of bis(imino)pyridyl chromium and
vanadium complexes on magnesium chloride gives much higher
molecular weights than have been obtained under homogeneous
conditions and furthermore, in the case of vanadium, a very nar-
row molecular weight distribution which can be ascribed to the
presence of a single active species formed via alkylation of the
pyridine ring.
Acknowledgements
We thank Drs. A.P. Jekel of the Center for Catalytic Olefin
Polymerization at the Rijksuniversiteit Groningen for GPC anal-
ysis and Dr. X. Zheng of Eindhoven University of Technology
for SEM characterization of polymer particle morphology. This
work is part of the Research Programme of the Dutch Polymer
Institute (DPI), under project number #495.
References
4. Conclusions
[1] B.L. Small, M. Brookhart, A.M.A. Bennett, J. Am. Chem. Soc. 120 (1998)
4049.
The molecular weight and polydispersity of polyethy-
lene prepared with bis(imino)pyridyl complexes of iron,
chromium and vanadium immobilized on supports of type
MgCl2/AlRn(OEt)3−n is strongly dependent on the nature of
the transition metal. Iron-based catalysts give polymers with
molecular weight distributions ranging from 3 to 12, depend-
ing on the substituents in the bis(imino)pyridyl ligand. The
[2] G.J.P. Britovsek, V.C. Gibson, B.S. Kimberley, P.J. Maddox, S.J. McTavish,
G.A. Solan, A.J.P. White, D.J. Williams, Chem. Commun. (1998)
849.
[3] G.J.P. Britovsek, M. Bruce, V.C. Gibson, B.S. Kimberley, P.J. Maddox, S.
Mastroianni, S.J. McTavish, C. Redshaw, G.A. Solan, S. Stro¨mberg, A.J.P.
White, D.J. Williams, J. Am. Chem. Soc. 121 (1999) 8728.
[4] Q. Wang, H. Yang, Z. Fan, H. Xu, J. Polym. Sci. A: Polym. Chem. 42
(2004) 1093.

R. Huang et al. / Journal of Molecular Catalysis A: Chemical 260 (2006) 135–143
143
[5] A.A. Barabanov, G.D. Bukatov, V.A. Zakharov, N.V. Semikolenova, L.G.
Echevskaja, M.A. Matsko, Macromol. Chem. Phys. 206 (2005) 2292.
[6] K.R. Kumar, S. Sivaram, Macromol. Chem. Phys. 201 (2000) 1513.
[7] K. Radhakrishnan, H. Cramail, A. Deffieux, P. Franc¸ois, A. Momtaz,
Macromol. Rapid Commun. 24 (2003) 251.
[8] (a) Q. Wang, L. Li, Z. Fan, J. Polym. Sci. A: Polym. Chem. 43 (2005) 1599;
(b) Q. Wang, L. Li, Z. Fan, Eur. Polym. J. 41 (2005) 1170.
[9] F.A.R. Kaul, G.T. Puchta, H. Schneider, F. Bielert, D. Mihalios, W.A. Her-
rmann, Organometallics 21 (2002) 74.
[10] R. Schmidt, M.B. Welch, S.J. Palackal, H.G. Alt, J. Mol. Catal. A: Chem.
179 (2002) 155.
[11] I. Kim, B.H. Han, C.-S. Ha, J.-K. Kim, H. Suh, Macromolecules 36 (2003)
6689.
(b) J.R. Severn, J.C. Chadwick, R. Duchateau, N. Friederichs, Chem. Rev.
105 (2005) 4073.
[17] M. Sacchetti, S. Pasquali, G. Govoni, US Patent 5698487 (1995), to Montell
Technology Company.
[18] R. Souane, F. Isel, F. Peruch, P.J. Lutz, C. R. Chim. 5 (2002) 43.
[19] B.L. Small, M. Brookhart, J. Am. Chem. Soc. 120 (1998) 7143.
[20] D.D. Devore, S.S. Feng, K.A. Frazier, J.T. Patton, WO 00/69923 (2000),
to Dow Chemical Company.
[21] D. Reardon, F. Conan, S. Gambarotta, G. Yap, Q. Wang, J. Am. Chem. Soc.
121 (1999) 9318.
[22] Y. Chen, R. Chen, C. Qian, X. Dong, J. Sun, Organometallics 22 (2003)
4312.
[23] (a) G.J.P. Britovsek, S.A. Cohen, V.C. Gibson, P.J. Maddox, M. van Meurs,
Angew. Chem. Int. Ed. 41 (2002) 489;
(b) G.J.P. Britovsek, S.T. Cohen, V.C. Gibson, M. van Meurs, J. Am. Chem.
Soc. 126 (2004) 10701.
[24] T.J. Marks, Acc. Chem. Res. 25 (1992) 57.
[12] (a) R. Huang, D. Liu, S. Wang, B. Mao, Macromol. Chem. Phys. 205 (2004)
966;
(b) R. Huang, D. Liu, S. Wang, B. Mao, J. Mol. Catal. A: Chem. 233 (2005)
91;
(c) R. Xu, D. Liu, S. Wang, B. Mao, Macromol. Chem. Phys. 207 (2006)
779.
[25] F. Pelascini, F. Peruch, P.J. Lutz, M. Wesolek, J. Kress, Eur. Polym. J. 41
(2005) 1288.
[13] T.B. Mikenas, V.A. Zakharov, L.G. Echevskaya, M.A. Matsko, J. Polym.
Sci. A: Polym. Chem. 43 (2005) 2128.
[26] R. Quijada, R. Rojas, G. Bazan, Z.J.A. Komon, R.S. Mauler, G.B. Galland,
Macromolecules 34 (2001) 2411.
[14] (a) J.R. Severn, J.C. Chadwick, Macromol. Rapid Commun. 25 (2004)
1024;
[27] M.A. Esteruelas, A.M. Lo´pez, L. Me´ndez, M. Oliva´n, E. On˜ate,
Organometallics 22 (2003) 395.
(b) J.R. Severn, J.C. Chadwick, Macromol. Chem. Phys. 205 (2004) 1987;
(c) J.R. Severn, J.C. Chadwick, V. Van Axel Castelli, Macromolecules 37
(2004) 6258;
(d) J.R. Severn, N. Kukalyekar, S. Rastogi, J.C. Chadwick, Macromol.
Rapid Commun. 26 (2005) 150;
(e) J.R. Severn, R. Duchateau, J.C. Chadwick, Polym. Int. 54 (2005) 837;
(f) J.R. Severn, J.C. Chadwick, WO 2005/077989 and 2005/092935 (2005),
to Stichting Dutch Polymer Institute.
[28] B.L. Small, M.J. Carney, D.M. Holman, C.E. O’Rourke, J.A. Halfen,
Macromolecules 37 (2004) 4375.
[29] R. Schmidt, M.B. Welch, R.D. Knudsen, S. Gottfried, H.G. Alt, J. Mol.
Catal. A: Chem. 222 (2004) 17.
[30] J.F. Vega, S. Rastogi, G.W.M. Peters, H.E.H. Meijer, J. Rheol. 48 (2004)
663.
[31] (a) H. Hagen, J. Boersma, G. van Koten, Chem. Soc. Rev. 31 (2002) 357;
(b) S. Gambarotta, Coord. Chem. Rev. 237 (2003) 229.
[32] J. Scott, S. Gambarotta, I. Korobkov, P.H.M. Budzelaar, J. Am. Chem. Soc.
127 (2005) 13019.
[15] (a) Y. Nakayama, H. Bando, Y. Sonobe, H. Kaneko, N. Kashiwa, T. Fujita,
J. Catal. 215 (2003) 171;
(b) Y. Nakayama, H. Bando, Y. Sonobe, Y. Suzuki, T. Fujita, Chem. Lett.
(Jpn.) 32 (2003) 766;
(c) Y. Nakayama, H. Bando, Y. Sonobe, T. Fujita, Bull. Chem. Soc. Jpn.
77 (2004) 617;
[33] I.J. Blackmore, V.C. Gibson, P.B. Hitchcock, C.W. Rees, D.W. Williams,
A.J.P. White, J. Am. Chem. Soc. 127 (2005) 6012.
[34] Q. Knijnenburg, J.M.M. Smits, P.H.M. Budzelaar, Organometallics 25
(2006) 1036.
(d) Y. Nakayama, H. Bando, Y. Sonobe, T. Fujita, J. Mol. Catal. A: Chem.
213 (2004) 141.
[16] For reviews see:
[35] J. Scott, S. Gambarotta, I. Korobkov, Q. Knijnenburg, B. de Bruin, P.H.M.
Budzelaar, J. Am. Chem. Soc. 127 (2005) 17204.
[36] H. Sugiyama, G. Aharonian, S. Gambarotta, G.P.A. Yap, P.H.M. Budzelaar,
J. Am. Chem. Soc. 124 (2002) 12268.
(a) G.G. Hlatky, Chem. Rev. 100 (2000) 1347;