Heterogenized iron(II) complexes as highly active ethene polymerization catalysts

Article abstract of DOI:10.1016/S1381-1169(01)00333-8

The synthesis of new iron(II) diiminopyridine complexes and their heterogenization to give highly active ethene polymerization catalysts is described. The ligands are characterized by ¹H NMR, ¹³C NMR spectroscopy and GC/MS. The compl

Full text of DOI:10.1016/S1381-1169(01)00333-8

Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
Heterogenized iron(II) complexes as highly active
ethene polymerization catalysts
Roland Schmidta, M. Bruce Welchb, Syriac J. Palackalb, Helmut G. Alta,∗
a
Laboratorium für Anorganische Chemie, Universität Bayreuth, Postfach 10 12 51, D-95440 Bayreuth, Germany
b
Phillips Petroleum Company, Research and Development Center, 74004 Bartlesville, OK, USA
Received 5 April 2001; accepted 16 August 2001
Abstract
The synthesis of new iron(II) diiminopyridine complexes and their heterogenization to give highly active ethene polymer-
ization catalysts is described. The ligands are characterized by 1H NMR, 13C NMR spectroscopy and GC/MS. The complexes
are paramagnetic and were characterized by elemental analyses and mass spectrometry. The complexes were activated and
heterogenized with a cocatalyst consisting of partially hydrolyzed trimethylaluminum (PHT) on silica gel and were used for
ethene polymerization.
The polymerization results and the polymer properties are presented. The influence of the catalyst structure, hydrogen and
1-hexene on the polymerization behavior is discussed. © 2002 Elsevier Science B.V. All rights reserved.
Keywords: Iron(II) complexes; Catalysis; Heterogenization; Ethene polymerization; Polymer properties
1. Introduction
neous. Metallocene catalysts can be made heteroge-
neous by either immobilizing the catalyst precursor
or the cocatalyst. An efficient way of immobilizing
the cocatalyst is to precipitate the MAO on a car-
rier material [12–15]. The catalyst precursor is then
added to the immobilized MAO to form the activated
catalyst.
A common way of immobilizing MAO consists of
depositing it on the surface of silica. Some of the re-
sulting ‘solid MAO’ are commercially available. A
significant disadvantage in using these materials is an
associated decrease in activity.
To overcome the disadvantages of homogeneous
and ‘solid MAO’ polymerizations, we prepared par-
tially hydrolyzed trimethylaluminum (PHT), a solid
formed by reacting a carrier and trimethylaluminum
(TMA) with water for the activation of new diimino-
pyridine complexes of iron(II). The resulting catalysts
were used for ethene polymerization.
In 1998, Small et al. [1], and Britovsek et al.
[2–5] independently described olefin polymerization
and oligomerization catalysts prepared by reacting a
well-known class of compounds [6], diiminopyridine
complexes of iron and cobalt, with an activator such
as methylaluminoxane (MAO) [7–11]. Unfortunately,
these homogeneous catalysts require a large excess
of expensive MAO when used to oligomerize and/or
polymerize ethene and result in reactor fouling in
slurry processes. For these reasons, these homoge-
neous catalysts are unattractive for industrial appli-
cations. In general, reactor fouling can be avoided in
the slurry loop process by making catalysts heteroge-
∗ Corresponding author. Tel.: +49-921-552555;
fax: +49-921-552157.
E-mail address: helmut.alt@uni-bayreuth.de (H.G. Alt).
1381-1169/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved.
PII: S1381-1169(01)00333-8

156
R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
Fig. 1. Synthesis of diketones.
2. Results and discussion
NMR and 13C NMR spectroscopy and mass spectro-
metry (Table 1).
2.1. Preparation of the ligand precursor
2.4. Preparation of the catalyst precursors 17–32
The common starting compound for the syntheses of
the desired tridentate ligands is 2,6-dicyanopyridine.
This pyridine is reacted with the appropriate Grignard
reagent in diethylether and subsequently treated with
dilute sulfuric acid to yield the corresponding diketone
(Fig. 1).
Complexes 17–32 were synthesized via a 1:1 ad-
dition reaction. Compounds 1–16 were added to
an iron(II) chloride solution in diethylether or di-
ethylether/tetrahydrofuran (Fig. 3). The resulting
complexes are paramagnetic and were characterized
by elemental analyses and mass spectrometry (see
Section 4).
2.2. Synthesis of the compounds 1–16
The diketone is reacted with a suitable aniline in
a condensation reaction (Schiff’s base reaction [6])
to give the corresponding diiminopyridine compounds
1–16 (Fig. 2).
2.5. Heterogeneous catalyst synthesis and ethene
polymerization results
Heterogeneous metallocene catalysts are frequently
formed by immobilizing the cocatalyst on the sur-
face of a carrier material followed by addition of
a homogeneous solution of the metallocene com-
plex. The metallocene complex is simultaneously
2.3. Characterization of compounds 1–16
Compounds 1–16 were prepared in accordance with
the procedure shown in Fig. 2 and analyzed via H
1
Fig. 2. Synthesis of the diiminopyridine compounds 1–16.

R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
157
Table 1
NMR spectroscopic and mass spectrometric data of compounds 1–16
No.
Compound
1H NMRa
13C NMRb
MS, M+
(m/e)
1
8.14 (d, 1H), 7.59 (t, 2H),
7.25–6.66 (m, 10H), 2.15
(s, 6H)
(Cq): 166.6, 160.8, 152.4 (CH):
138.7, 128.9, 124.1, 123.3,
119.7 (CH3): 16.5
n.dc
2
8.13 (d, 1H), 7.86 (t, 2H),
7.22–6.69 (m, 8H), 2.31
(s, 6H), 2.13 (s, 6H)
(Cq): 164.6, 154.7, 148.7, 133.1
(CH): 138.5, 130.1, 125.5, 124.3,
123.2, 118.4 (CH3): 18.0, 16.4
341
3
4
8.38 (d, 1H), 7.86 (t, 2H),
7.04–6.56 (m, 6H), 2.32
(s, 12H), 2.09 (s, 6H)
(Cq): 166.9, 155.1, 147.4, 132.9,
127.1 (CH): 136.7, 131.1, 126.9,
122.1, 118.1 (CH3): 20.8, 17.7,
16.2
369
369
8.23 (d, 1H), 7.86 (t, 2H),
7.04–6.52 (m, 6H), 2.35
(s, 12H), 2.13 (s, 6H)
(Cq): 167.2, 155.1, 148.7, 136.4,
125.4, 122.2 (CH): 127.9, 123.0
(CH3): 18.0, 16.5
5
6
8.36 (d, 1H), 7.88 (t, 2H),
6.95–6.75 (m, 6H), 2.32
(d, 18H), 1.90 (s, 6H)
(Cq): 167.4, 155.2, 146.2, 132.2,
125.3 (CH): 136.8, 128.5, 122.1
(CH3): 20.7, 17.9, 16.4
397
425
8.36 (d, 1H), 7.90 (t, 2H),
7.12–6.70 (m, 4H), 2.63–2.25
(m, 8H), 2.22 (s, 6H), 1.22–1.05
(m, 12H)
(Cq): 166.9, 155.1, 136.9 (CH):
147.8, 131.2, 127.7, 126.0,
125.9, 123.3, 122.2, (CH2): 24.6,
24.3 (CH3): 16.8, 13.8, 13.0
7
8
7.82 (d, 1H), 7.18 (d, 2H),
6.98–6.76 (m, 6H), 2.85 (sept,
2H), 2.14 (s, 6H), 1.96 (s, 6H),
1.21 (dd, 12H)
(Cq): 167.l, 155.2, 147.6, 136.2,
125.0 (CH): 138.4, 127.8, 123.2,
118.3, 28.3 (CH3): 23.1, 22.8,
18.1, 6.8
453
481
7.85 (d, 1H), 7.20 (d, 2H),
6.88–6.72 (m, 6H), 2.88 (sept,
4H), 2.12 (s, 6H), 1.20 (dd, 24H)
(Cq): 167.1, 158.2, 145.6, 138.2
(CH): 138.8, 125.8, 123.9, 122.4
(CH3): 26.1, 23.3, 16.6

158
R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
Table 1 (Continued )
No.
Compound
1H NMRa
13C NMRb
MS, M+
(m/e)
9
8.45 (d, 1H), 7.90 (t, 2H), 6.98
(d, 2H), 2.39 (s, 6H), 2.22 (s,
6H), 2.15 (s, 6H), 1.97 (s, 6H)
(Cq): 167.9, 154.8, 147.1, 132.2,
125.2, 124.0 (CH): 136.7, 129.5,
122.2 (CH3): 23.3, 18.3, 17.5, 16.4
556
10
11
8.38 (d, 1H), 7.85 (t, 2H),
7.03–6.80 (m, 6H), 2.75 (q,
4H), 1.98 (s, 12H), 1.44 (t, 6H)
(Cq): 171.4, 154.5, 148.4, 125.3
(CH): 127.9, 122.9 (CH2): 23.2
(CH3): 18.2, 11.3
397
n.d.
8.32–8.27 (m, 6H), 7.60 (d,
2H), 7.50 (s, 4H), 7.01–6.88
(m, 6H), 1.98 (s, 12H)
(Cq): 165.9, 152.9, 149.1, 132.0,
126.9 (CH): 139.6, 130.9, 130.0,
128.6, 126.0, 124.8, 124.0,
122.8 (CH3): 17.5
12
8.53 (d, 1H), 8.18–7.96
(Cq): 168.9, 154.9, 152.2, 141.0,
126.1 (CH): 137.6, 128.6, 128.5,
126.1, 125.0, 123.3, 122.8
549
(m, 4H), 7.26–6.81 (m, 14H),
3.70–354 (m, 4H), 3.13–2.71
(m, 4H), 2.00 (s, 12H)
(CH2): 29.7 (CH3): 18.1
13
14
15
8.14 (d, 1H), 7.59 (d, 2H),
7.26–6.67 (m, 8H), 3.00 (sept,
2H), 2.13 (s, 6H), 1.21 (d, 12H)
(Cq): 166.6, 155.5, 148.7, 138.1
(CH): 136.8, 126.1, 125.7,
124.0, 122.2, 118.4, 28.5 (CH3):
22.8, 16.4
397
425
397
8.14 (d, 1H), 7.59 (d, 2H),
7.19–6.80 (m, 8H), 2.11
(s, 6H), 1.18 (s, 18H)
(Cq): 170.2, 158.5, 148.2, 139.0,
35.5 (CH): 138.0, 128.4, 125.3,
124.0, 123.8, 123.1 (CH3): 31.2,
16.2
8.16 (d, 1H), 7.64 (d, 2H),
6.97–6.80 (m, 4H), 6.40 (m,
2H), 2.70 (q, 4H), 2.10 (s, 6H),
2.08 (s, 6H), 1.40 (t, 6H)
(Cq): 166.9, 159.1, 152.5, 139.5,
128.9 (CH): 138.0, 130.9, 130.0,
124.8 (CH2): 28.0 (CH3): 18.8,
16.6, 15.7
16
8.14 (d, 1H), 7.60 (d, 2H),
7.16–6.61 (m, 6H), 2.85 (sept,
2H), 2.28 (s, 6H), 2.10 (s, 6H),
1.25 (d, 12H)
(Cq): 167.1, 155.2, 147.6, 136.1,
125.0 (CH): 136.9, 127.8, 123.3,
123.3, 122.2, 28.4 (CH3): 23.1,
22.9, 18.2, 16.8
425
a 25◦C, in chloroform-d1, δ (ppm) rel. chloroform (7.24).
b 25◦C, in chloroform-d1, δ (ppm) rel. chloroform-d1 (77.0).
c n.d. = not determined.

R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
159
Fig. 3. Synthesis of the catalyst precursors 17–32.
activated and attached to the surface of the carrier.
Several patents and articles have been issued in re-
cent years that describe the heterogenization of MAO
and various catalyst precursors on a carrier material
[16–19]. Supporting catalyst precursor 20 on Witco
prepared HL/04 silica-supported MAO (Al/Fe molar
ratio = 253/1) produced a catalyst with an activity of
less than 500 g PE/gcat h.
Due to the very low activity of the heterogeneous
Fe/MAO catalyst, other methods of forming analo-
gous heterogeneous catalysts were tried. The method
producing the best results was the PHT method de-
scribed previously [7]. The addition of water to TMA
in the presence of dried silica was shown to produce
a significantly different heterogeneous support than
if TMA was added to wet silica. At the equivalent
Al/Fe molar ratio and reactor conditions to those used
for the Fe(II)/MAO system described above, the PHT
method produced a catalyst that was 30 times more
active (10,200 g PE/gcat h).
Then the desired amount of water was added to form
the PHT. Addition of the selected catalyst precursor
17–32 yielded the active heterogeneous ethene poly-
merization catalyst (Fig. 4). Using this method, the
water to aluminum content can be varied over a broad
range. The location and nature of substituents on the
diiminopyridine ligand were highly influential in de-
termining the catalytic activity of the activated com-
plex as well as observed properties.
2.5.1. Different aluminum/iron molar ratios
The activities of heterogeneous PHT/metallocene
catalysts are dependent on the aluminum/transition
metal molar ratios. In an earlier study [7] with a metal-
locene catalyst, the maximum activity was observed at
an Al/Zr molar ratio of about 260/1. Even slight devi-
ations in the Al/Zr molar ratio were found to dramati-
cally alter the activity. At molar ratios under 50/1, the
polymerization system exhibited almost no activity.
In contrast, the heterogeneous iron containing
catalysts exhibit high activities even at low Al/Fe
molar ratios, 8/1, as shown in Table 2 and Fig. 5.
Molar ratios lower than 8/1 were not tested as the
For the PHT method, the carrier material used was
calcined silica with a water content of <1.2 wt.%. The
silica was suspended in toluene and treated with TMA.

160
R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
Fig. 4. Synthesis of a PHT-supported catalyst.
Table 2
Activities, based on g PE/gFe h, go through a maxi-
mum at an Al/Fe molar ratio of about 220/1 (Fig. 5).
The activities are comparable to those observed for
many metallocene catalysts ([20–23] and references
therein).
Activity as a function of Al/Fe molar ratios
Al/Fe molar ratio, catalyst
precursor 20 + PHT
Activity
Activity
(g PE/gFe h)
(g PE/gcat h)a
8
21
43
55740
27860
17860
25600
11500
10200
2460
920000
1030000
1100000
2930000
2890000
4380000
1290000
155000
2.5.2. Varying water/aluminum molar ratios
84
One of the most surprising observations regarding
iron/PHT olefin polymerization catalysts is that water
to aluminum molar ratios over 1.0 result in the high-
est activities. In contrast, metallocene/PHT catalysts
frequently show their highest activities at a water/Al
molar ratio of 0.69 [7,24] and half-sandwich imino
complexes/PHT at a molar ratio of 0.83 [25].
Theoretical and mechanistic studies suggest that
MAO, of which 20–30% of the aluminum is present
as ‘free’ TMA, methylates the metallocene dichloride
170
254
339
509
1570
a The cat = catalyst precursor 20 on PHT.
catalyst precursor was not entirely immobilized under
these conditions. As expected, no catalytic activ-
ity was observed for any of the iron(II) complexes
studied in the absence of a cocatalyst.

R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
161
Fig. 5. Molar ratios Al/Fe vs. activity (g PE/gFe h). Catalyst precursor 20 on PHT. Polymerization conditions: H2O/Al molar ratio = 0.93;
31 bar ethene pressure at 90 ◦C in isobutane; 8.5 mmol hydrogen added at the start of the polymerization runs; 60 min polymerization time.
complex prior to any catalytic activity. After methyla-
tion, one methyl anion is abstracted. The metallocene
mono methyl cation formed is thought to be the active
species ([26–34] and references therein).
should be few remaining aluminum–methyl bonds and
very little if any ‘free TMA’.
Therefore, it is possible that any aluminum–methyl
bond is capable of activating selected iron(II) chlo-
ride complexes and thus initiating the polymerization
reaction. Each type catalyst precursor, activated with
PHT, may require a different water/Al molar ratio to
obtain its best catalytic activity. The reason for such
behavior might result from the fact that late transition
metals, e.g. iron, cobalt or nickel are less oxophilic
than titanium, zirconium or hafnium and more ro-
bust against typical Lewis base poisons. Thus, higher
water/transition metal molar ratios in the catalyst
may have less effect on the late than early transition
In contrast, the PHT activated iron containing cata-
lysts show little activity at water/Al molar ratios simi-
lar to those found in conventional MAO. Instead. these
catalysts are most active at a water/aluminum molar
ratio of 1.1 (Fig. 6). At this molar ratio it is unlikely
that free TMA still exists. This suggests that free TMA
is not needed to activate these complexes and that free
TMA may reduce activity for the iron catalysts stud-
ied. Activities drop to zero when the water/aluminum
molar ratio exceeds 1.5 as expected. At this point there
Fig. 6. H2O/Al molar ratio vs. catalyst activity (g PE/gcat h). Catalyst precursor 20 on PHT. Polymerization conditions: Al/Fe molar
ratio = 253/1; Al/Zr molar ratio 260/1; 31 bar ethene pressure at 90 ◦C in isobutane; 8.5 mmol hydrogen added at the start of the
polymerization runs; 60 min polymerization time.

162
R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
Fig. 7. Activity vs. silica/trimethylaluminum weight ratio. Catalyst precursor 20 on PHT. Polymerization conditions: H2O/Al molar
ratio = 1.1/1; Al/Fe molar ratio = 253/1; 31 bar ethene pressure at 90 ◦C in isobutane; 8.5 mmol hydrogen added at the start of the
polymerization runs; 60 min polymerization time.
metal/PHT catalysts. Likewise, it may also be that
an alternative activation mechanism or anion/cation
equilibrium exists. Further investigations are nec-
essary to discern the reasons for the observed
behavior.
solid could be used as a heterogeneous catalyst. Its
polymerization behavior was comparable to other in-
vestigated catalyst systems.
2.5.4. Influence of hydrogen and comonomers
during polymerization
2.5.3. Varying aluminum/silica molar ratios
Most ethene polymerizations were carried out in the
presence of hydrogen, and some were carried out in
the presence of both hydrogen and 1-hexene. Under
the chosen polymerization conditions, almost no ac-
tivity was observed in the absence of hydrogen. Traces
of hydrogen greatly improved activity (Fig. 8), but
further increases in hydrogen concentration decreased
activity. Addition of 1-hexene at constant ethene con-
centration results in a decrease in activity (Fig. 9).
However, when the polymers were analyzed by 13C
HT-NMR spectroscopy, no incorporation of 1-hexene
was found.
In order to prepare a PHT/metallocene catalyst sys-
tem, a carrier, such as silica, is generally used to pre-
vent reactor fouling. For the studies reported here,
Davison XPO 2410, a calcined silica containing less
than 1.2 wt.% total volatiles at 955 ◦C, was selected.
The activity of the catalyst was influenced by the silica
to TMA weight ratio (see Fig. 7). The degree of fouling
is dependent on the silica content of the final catalyst
[7,24]. The iron(II) dichloride/PHT system showed no
reactor fouling tendency even if no support material
was incorporated. Under these conditions, the obtained
Fig. 8. Increasing activity with decreasing hydrogen concentration at constant ethene concentration. Catalyst precursor 20 on PHT.
Polymerization conditions: Al/Fe molar ratio = 253/1; H2O/Al molar ratio = 1.15/1; 31 bar ethene pressure at 90 ◦C in isobutane; 60 min
polymerization time.

R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
163
Fig. 9. Influence of 1-hexene on the activity of catalyst precursor 20 on PHT. Polymerization conditions: Al/Fe molar ratio = 253/1;
H2O/Al molar ratio = 1.15/1; 31 bar ethene pressure at 80 ◦C in isobutane; 8.5 mmol hydrogen and 1-hexene fed continuously during
polymerization; 60 min polymerization time.
2.5.5. Influence of reactor residence time on
polymerization behavior
The activity decreases with the increasing bulk
of substituents in the 2- and 6-positions of the ani-
lines: dimethyl > diethyl > methyl-isopropyl >
diisopropyl (Fig. 11). Activity can be increased as
much as 42% by substituting the 2,6-dimethyl aniline
moiety with a 3-bromo-2,4,6-trimethyl aniline moiety
(Fig. 12).
Activity was found to be directly related to the
reactor residence time (Fig. 10). No decline in ac-
tivity was noted with run length. This kinetic profile
suggests that the catalytically active polymerization
centers do not decay over the time frame of the ex-
periments. In contrast, most metallocene complexes
lose activity rapidly when the run length increases. It
is likely that traces of oxygen and water fed with the
ethene into the reactor affect those catalytic centers
negatively. Unlike metallocenes, the Fe/PHT system
may benefit from those impurities.
The steric and electronic effects of substituents on
the transition metal polymerization site also influence
the nature of the polymers produced. Low molecular
¯
weight, waxy polymers (Mw = 2700 g/mol) are pro-
duced with catalysts that are only substituted in the
2-position of the aniline moieties. In contrast, substi-
tution at both the 2- and 6-positions produces high
¯
molecular weight polymers MI = 0.1 dg/min; Mw =
2.5.6. Steric and electronic effects on catalyst activity
Polymerization reaction kinetics are influenced by
both steric and electronic changes at the transition
metal catalytic center. Such changes occur when var-
ious functional groups are substituted onto either the
aniline moiety or onto the carbon atom at the imino
moiety.
455, 000 g/mol. These results are in agreement with
the work first reported by Brookhart and coworkers
[1].
It is important to note that both the melt index (MI)
and high load melt index (HLMI) data are dependent
on the polymerization run length. Additional or other
Fig. 10. Polymerization run-time vs. activity. Catalyst precursor 20 on PHT. Polymerization conditions: Al/Fe molar ratio = 253/1; H2O/Al
molar ratio = 1.15/1; 31 bar ethene pressure at 90 ◦C in isobutane; 8.5 mmol hydrogen added at the start of the polymerization runs.

164
R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
Fig. 11. Influence of increasing substituent bulkiness on activities. Catalyst precursors 20–23 on PHT. Polymerization conditions: Al/Fe
molar ratio = 253/1; H2O/Al molar ratio = 1.15/1; 31 bar ethene pressure at 90 ◦C in isobutane; 8.5 mmol hydrogen added at the start of
the polymerization runs; 60 min polymerization time.
Table 3
functional groups on the aniline rings have almost no
Influence of different imino substituents on polymer properties
influence on the molecular weights of the obtained
polymers (Fig. 13).
Imino
substituent
MI (dg/min)
HLMI
(dg/min)
HLMI/MI
The impact of imino substituents at the imino moi-
ety on resin properties are much less pronounced. Only
a small decrease in molecular weight is observed as
the bulk of the substituent is increased (Table 3).
Methyl
Ethyl
Phenyl
0.1
2.31
27.00
3.72
n.a.
23.1
112.5
372.0
n.a.
0.24
0.01
0.02
Phenylethyl
2.5.7. Other support materials (‘double shell’ PHT)
Using the PHT technique in the absence of a
carrier, a TMA solution in toluene was partly hy-
drolyzed in a water to aluminum molar ratio of
1.53. The fact that the solid was still pyrophoric
following the water treatment indicates that not all
aluminum–methyl groups were deactivated. Subse-
quently, the solid product obtained was then used as
the carrier material in another PHT preparation (‘dou-
ble shell’). In the second step, a water to aluminum
molar ratio of 1.15 was used. The catalyst precur-
sor was only added after the second preparation step
Fig. 12. Influence of different substitution patterns on the aniline moiety of the iron containing catalyst precursors on activity. Catalyst
precursors 20, 19, 25, 22 on PHT. Polymerization conditions: Al/Fe molar ratio = 253/1; H2O/Al molar ratio = 1.15/1; 31 bar ethene
pressure at 90 ◦C in isobutane; 8.5 mmol hydrogen added at the start of the polymerization runs; 60 min polymerization time.

R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
165
Fig. 13. Influence of various bulky substituents at the imino moiety on activity. Catalyst precursor 20, 26, 27, 28 on PHT. Polymerization
conditions: Al/Fe molar ratio = 253/1; H2O/Al molar ratio = 1.15/1; 31 bar ethene pressure at 90 ◦C in isobutane; 8.5 mmol hydrogen
added at the start of the polymerization runs; 60 min polymerization time.
(Fig. 14). The PHT ‘double shell’ catalyst was 35%
more active than the normal PHT catalyst (20,000 ver-
sus 13,000 g PE/gcat h, respectively).
mined. In addition, other test methods, e.g. high tem-
perature gel permeation chromatography, density and
high temperature NMR, were conducted.
3.1. HT–GPC results
3. Polymer properties
The molecular weights of polyethenes made with
homogeneous (MAO) and heterogeneous (PHT) di-
iminopyridine iron(II) dichloride catalysts vary signifi-
cantly. The polymer produced with the heterogeneous
The physical, mechanical and optical properties of
polyolefins determine their utility and quality. The MI
and HLMI of all obtained polyethenes were deter-
Fig. 14. Preparation of a ‘double shell’ PHT catalyst system.

166
R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
Table 4
Comparison of molecular weights and molecular weight distributions of homogeneously and heterogeneously polymerized polyethenes
made with catalyst 20
¯
¯
Mn (g/mol)
Mw (g/mol)
Polydispersity
Homogeneously polymerized PE
Heterogeneously polymerized PE
6700
41000
91000
412000
13.6
10.2
catalyst has a number average molecular weight six
times higher than the polymer made with the homo-
geneous catalyst (see Table 4). Similarly, the weight
average molecular weight is about four to five times
higher for the hetero- versus the homopolymerized
resin.
The molecular weight distribution of the heteroge-
neously polymerized resin is unexpectedly narrower
than for the homogeneously polymerized resin. The
HT–GPC curve for the homogeneously polymerized
polyethene indicates that this resin is bimodal while
the similar curve for the heterogeneously polymerized
polyethene does not show any bimodal character (see
Fig. 15). However, the heterogeneously polymerized
polyethylene has a much broader molecular weight
distribution than would be expected for a single site
catalyst.
Unexpectedly, the weight average molecular
weights of polyethenes made with a heterogeneous
catalyst were observed to be run-time dependent. The
resin obtained from a 2 min run had an Mw of about
60 kg/mol while the resin from a 60 min run had an
¯
¯
Mw of 260 kg/mol. The plot in Fig. 16 illustrates the
relationship between polymerization time and Mw.
¯
The molecular weight distribution of the resins,
made at different residence times, is also dependent on
the length of the polymerization. The polydispersity
increases with time as seen in Fig. 16. The polymer
from a 2 min run has a polydispersity of 5.1 while
the polymer from a 60 min run broadens to a polydis-
¯
persity of 12.4. The observed changes in the Mw and
polydispersity with polymerization time are difficult
to explain on the basis of ethene diffusion limitations
resulting from the growing fluff particles [28]. Rather
Fig. 15. Comparison of molecular weight distribution of homogeneously and heterogeneously obtained polyethenes. Catalyst: 20/MAO.
Polymerization conditions: Al/Fe molar ratio = 2000/1 (homogeneous) and 20/PHT/silica; Al/Fe molar ratio = 253/1; H2O/Al = molar
ratio 1.15/1 (heterogeneous); 31 bar ethene pressure at 90 ◦C in isobutane; 8.5 mmol hydrogen added at the start of the polymerization
runs; 60 min polymerization time.

R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
167
Fig. 16. Shifts of the molecular weight maxima with the polymerization time. Catalyst precursor 20 on PHT. Polymerization conditions:
Al/Fe molar ratio = 253/1; H2O/Al molar ratio = 1.15/1; 31 bar ethene pressure at 90 ◦C in isobutane; 8.5 mmol hydrogen added at the
start of the polymerization runs; 60 min polymerization time.
it appears from the HT–GPC curves that there is a
structure change of some of the active centers. As the
polymerization proceeds, some of the active polymer-
ization centers are changed such that they produce
higher molecular weight polymers. At 20 min into a
run one can see a distinct peak in the HT–GPC at
¯
log Mw of about 5.8 that grows as the time increases
and is not observed in resins from shorter run times.
The nature of the polymerization site transformation
is not currently understood, but such transformations
are known to occur for metallocene catalysts as well
[24].
Not only does the water/Al molar ratio influence
Fig. 17. Effect of water/Al molar ratio on the molecular weight of
the obtained resin. Catalyst precursor 20 on PHT. Polymerization
conditions: Al/Fe molar ratio = 253/1; 31 bar ethene pressure at
90 ◦C in isobutane; 8.5 mmol hydrogen added at the start of the
polymerization runs; 60 min polymerization time.
the activity of the iron(II) complex/PHT system, but it
also alters the molecular weight of the resulting resins
as shown in Fig. 17.
As the water/Al molar ratio increased the molecular
weight decreased linearly. No correspondence of ac-
tivity with molecular weight was observed. The reason
for the observed decrease in molecular weight with
increasing water levels is clearly a reflection of the
changing nature of the polymerization sites, but with-
out knowing the structure of these sites we cannot an-
ticipate what these changes might be.
As described earlier, substituents can influence
the molecular weight of the resulting polymer. Sub-
stituents in positions 2 and 6 or only in position 2
of the aniline moiety determine whether the result-
ing polymer is high or low in molecular weight.

168
R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
Table 5
Influence of different substituents on the aniline part of the catalyst precursor on MI, HLMI, HLMI/MI data
Catalyst precursor
Substituent
MI (dg/min)
HLMI (dg/min)
HLMI/MI
20
21
25
2,6-Dimethyl
2,4,6-Trimethyl
3-Bromo-2,4,6-trimethyl
0.01
0.03
0.009
2.31
3.9
2.28
231
130
253
Table 6
Influence of different substituents on the imino part of the catalyst precursor on MI, HLMI, HLMI/MI data
Catalyst precursor
Substituent
MI (dg/min)
HLMI (dg/min)
HLMI/MI
20
26
27
28
Methyl
Ethyl
Phenyl
0.01
0.24
0.01
0.02
2.31
27
3.72
n.a.
231
113
372
n.a.
Phenylethyl
Substituents in other positions or different substituents
on the aniline moiety do not influence the molecular
weight significantly (Table 5).
Substituents at the imino moiety have little effect on
resin molecular weight. Only the ethyl substituent sig-
nificantly altered the resin molecular weight (Table 6).
3.2. Branching
Fig. 19. Density of the polyethenes obtained at different run times
As polymerization reaction time increases, the den-
sity of the resulting polyethene decreases (see Fig. 18).
This decrease in density is not the result of short
chain branch formation. Rather the density decrease is
a reflection of the changes in the polymer molecular
weight (see Fig. 19).
¯
vs. log Mw. Catalyst precursor 20 on PHT. Polymerization condi-
tions: Al/Fe molar ratio = 253/1; H2O/Al molar ratio = 1.15/1;
31 bar ethene pressure at 90 ◦C in isobutane; 8.5 mmol hydrogen
added at the start of the polymerization runs.
Even when 1-hexene was present during the poly-
merization run, the resulting polymer contained no
detectable butyl branches by 13C HT-NMR spectro-
scopy.
4. Experimental
All procedures were performed under an inert gas
atmosphere using standard Schlenk techniques or a
glove box in order to eliminate traces of air or mois-
ture. Purified and dried argon (BTS catalyst, molecular
sieve) was used as the inert gas. All other preparations
were carried out in air.
Fig. 18. Dependence of polyethene density on the polymerization
time. Catalyst precursor 20 on PHT. Polymerization conditions:
Al/Fe molar ratio = 253/1; H2O/Al molar ratio = 1.15/1; 31 bar
ethene pressure at 90 ◦C in isobutane; 8.5 mmol hydrogen added
at the start of the polymerization runs.

R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
169
All solvents were purchased as technical grade pu-
rity and purified by distillation over Na/K alloy under
argon atmosphere. All other chemicals were commer-
cially available or were synthesized according to lit-
erature procedures.
The reaction products were monitored and con-
trolled via GC/MS. After completion of the reaction,
the mixture was hydrolyzed with diluted sulfuric acid
(ca. 1 M) and the mixture was stirred for an additional
30 min.
Sodium carbonate was added carefully to the mix-
ture until gas evolution stopped. The organic layer was
separated, and the water layer was extracted twice with
diethylether. The organic layers were combined, dried
over Na2SO4 and evaporated to dryness. The residue
was used without further purification.
NMR spectra were recorded in CDCl3 at 25 ◦C on
a Bruker ARX 250 instrument. The chemical shifts
1
of H NMR spectra were referenced to the residual
proton signal of the solvent (δ = 7.24 ppm for CHCl3);
the carbon resonances in 13C NMR spectra were also
referenced to the solvent signal (δ = 77.0 ppm for
CDCl3). Mass spectra were recorded on a Varian MAT
CH7 instrument (direct inlet system, electron impact
ionization 70 eV).
4.2. General procedure for the preparation of the
free ligands (condensation reaction)
Polymer molecular weights were recorded on a Mil-
lipore Waters GPC 150C instrument. For the separa-
tion four successively connected polystyrene columns
were used. The pore diameters of the single columns
were 500, 1000, 10,000 and 100,000 Å. For the re-
fractometric detection, a refractometer RI Waters 401
was used. The solvent was degassed 1,2,4-trichloro-
benzene (flow rate of 1 ml/min). Polymer samples
were dissolved in boiling 1,2,4-trichlorobenzene and
filtered. The HT–GPC was calibrated with polystyrene
standards. Polymerizations were conducted in isobu-
tane at 90 ◦C and 31 bar ethene pressure for 1 h.
The MI is a measure of the amount of molten poly-
mer that extrudes from an orifice under an applied
pressure in a specific length of time. The HLMI is
the same measurement but is 10 times the applied
pressure. The measurements were made according to
ASTM D1238 (conditions 190/2.16 and 190/21.6).
To a stirred solution of the corresponding diketopy-
ridine in toluene, 0.05 g of p-toluenesulfonic acid were
added. The aniline compound was added in excess.
The mixture was refluxed and the water formed was
separated via a Dean–Stark trap. Refluxing was contin-
ued for 3–25 h (depending on the aniline used and/or
the substituent at the diketopyridine). Completion of
the reaction was determined via GC/MS.
After cooling to room temperature, the mixture was
washed twice with a dilute solution of Na2CO3 in
water and twice with water. The organic layer was
separated and the water phase was extracted twice with
diethylether. The combined organic layers were dried
over Na2SO4 and evaporated to dryness. Ethanol was
added to the dried residue. If crystallization did not
start immediately, the solution was stored at −20 ◦C
overnight.
The crystals were filtered off, washed with ethyl-
alcohol and dried in air. The liquid residue was reduced
in volume to about 30–40%. A second crop of crystals
could be obtained after storing at −20 ◦C overnight.
Overall yield: 90–95%.
4.1. General procedure for the preparation
of substituted diketopyridines
A Grignard reagent of bromoalkyl, bromoaryl or
bromoalkylaryl compound was prepared. For this pur-
pose a solution of 40 mmol of the bromo compound,
dissolved in diethylether, was added dropwise to
0.97 g of magnesium (40 mmol) in diethylether. Next
a solution of 2,6-dicyanopyridine (1 g = 7.74 mmol)
in diethylether was prepared. Either the Grignard
reagent was added dropwise to the stirred solution
of 2,6-dicyanopyridine at 0 ◦C (inverse Grignard
reaction) or the 2,6-dicyanopyridine was added drop-
wise to the stirred Grignard reagent at 0 ◦C (straight
Grignard reaction).
4.3. General procedure for the preparation of the
iron complexes 17–32
All complexes were prepared in an analogous man-
ner. The corresponding diiminopyridine was stirred
in a mixture of diethylether and tetrahydrofuran (1:1)
at room temperature. Then one equivalent of finely
powdered iron dichloride was added. A rapid color
change indicated the reaction. Stirring was continued
for at least 10 h. Then the colored solids were filtered,

170
R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
Table 7
Mass spectrometric analysis (only highest mass peaks listed) and colors of the synthesized catalyst precursors
No.
Catalyst precursor
MS, M+ (m/e)
Color
Black
17
438
18
19
467
495
Light blue
Blue
20
21
495
523
Blue
Blue
22
551
Blue
23
551
Light blue

R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
171
Table 7 (Continued )
No.
Catalyst precursor
MS, M+ (m/e)
Color
Blue
24
25
26
607
682
523
Dark blue
Blue
27
28
29
30
619
676
524
552
Blue
Light blue
Light blue
Dark green

172
R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
Table 7 (Continued )
No.
Catalyst precursor
MS, M+ (m/e)
Color
Blue
31
32
552
551
Blue
washed with diethylether and dried. The yields varied
from 50 to 80% depending to what extent the solutions
were used for crystallization.
and 28.7 ml of a 1.764 M solution of TMA in toluene.
The mixture was vigorously stirred for 30 min.
Another flask containing 0.70 ml up to 1.5 ml of
water was connected via a Teflon tube to the mixture
and heated to 180 ◦C. The Teflon tube ended near the
bottom of the reaction flask. Dry nitrogen was purged
through the water containing flasks and a wet nitrogen
stream was formed. This wet nitrogen stream trans-
ferred the vaporized water to the TMA/silica/toluene
slurry. Purging continued until all the water was
vaporized and transferred to the reaction flask. The
slurry became highly viscous and the temperature in-
creased to about 60 ◦C. Some toluene (ca. 50 ml) was
added, and stirring was continued until the reaction
flask had cooled down to room temperature. To the
resulting slurry was added the iron(II) catalyst pre-
cursor (0.2 mmol), and stirring was continued for at
least 1 h.
4.4. Characterization of the catalyst precursors
17–32
The products 17–32 are paramagnetic and therefore
NMR spectroscopy was not applied for characteriza-
tion. Instead, elemental analyses (C, H, N, Cl) were
performed for most derivatives in order to confirm the
correct composition of the compounds. In addition, the
mass spectra showed molecular ions for all derivatives
(Tables 7 and 8).
4.5. General procedure for the preparation of a
supported catalyst
To 1 g of dried Davison silica (XPO 2410) in a
500 ml Schlenk flask were added 100 ml of toluene
The slurry was filtered over a frit. After removing
the toluene, the residue was washed with hexane and
Table 8
Elemental analyses
Complex
CCalc
CFound
HCalc
HFound
NCalc (ClCalc
)
Nfound (ClFound)
17
20
21
24
25
27
29
30
54.70
60.49
60.80
65.26
49.05
64.97
60.61
59.21
54.13
60.87
61.85
65.14
47.40
67.76
61.85
63.06
4.35
5.48
5.96
7.12
4.57
5.04
5.96
6.39
4.37
5.95
6.09
7.22
4.84
5.47
6.30
5.90
(16.11)
8.47
8.01
6.91
6.14
6.77
8.01
7.61
(16.10)
7.77
7.81
7.13
5.66
6.06
7.53
7.18

R. Schmidt et al. / Journal of Molecular Catalysis A: Chemical 179 (2002) 155–173
173
dried in vacuo to give a white solid. Overall yield was
>95%.
[13] J.A.M. Canich, G.F. Licciardi, Exxon Chemical Co., US
Patent 5,057,475 (1991).
[14] T. Tsutsui, T. Ueda, Mitsui Petrochemical Ind., US Patent
5,234,878 (1993).
[15] M. Chang, Exxon Chemical Co., US Patent 5,529,965 (1996).
[16] O. Uchida, R. Sugimoto, Mitzu Toatsu Chemicals, Jpn. Kokai
Tokyo Koko JP 04,234,405 [92,234,405] (1992); Chem. Abstr.
117 (1992) 2520034.
[17] R. Jackson, J. Ruddelsden, D.J. Thompson, R. Whelan, J.
Organomet. Chem. 125 (1977) 57.
Acknowledgements
We thank Witco company (Germany) for the dona-
tion of methylalumoxane.
[18] W. Kaminsky, F. Renner, Makromol. Chem. Rapid Commun.
14 (1993) 239.
[19] H.G. Alt, K. Patsidis, M.B. Welch, R.L. Geerts, B. Peifer,
S.J. Palackal, D.R. Fahey, H.R. Deck, Phillips Petroleum Co.,
Eur. Pat. Appl. Ep. 628566 (1994); Chem. Abstr. 123 (1995)
33867.
References
[1] B.L. Small, M. Brookhart, A.M.A. Bennett, J. Am. Chem.
Soc. 120 (1998) 4049.
[20] K. Soga, T. Shiono, Prog. Polym. Sci. 22 (1997) 1503.
[21] M. Bochmann, J. Chem. Soc., Dalton Trans. (1996) 255.
[22] V.K. Gupta, S. Satish, I.S. Bhardwaj, Rev. Macromol. Chem.
Phys. C 34 (1994) 439.
[23] W. Kaminsky, J. Chem. Soc., Dalton Trans. (1998) 1413.
[24] A. Köppl, H.G. Alt, M.D. Phillips, J. Appl. Polym. Sci. 80
(2001) 454.
[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, V.C. Gibson, D.F. Wass, Angew. Chem. 111
(1999) 448.
[4] G.J.P. Britovsek, V.C. Gibson, D.F. Wass, Angew. Chem. Int.
Ed. Engl. 38 (1999) 428.
[25] A. Reb, Dissertation, Universität Bayreuth, 2000.
[26] H.H. Brintzinger, D. Fischer, R. Mülhaupt, B. Rieger, R.
Waymouth, Angew. Chem. 107 (1995) 1255.
[27] H.H. Brintzinger, D. Fischer, R. Mülhaupt, B. Rieger, R.
Waymouth, Angew. Chem. Int. Ed. Engl. 34 (1995) 1143.
[28] G.G. Hlatky, H.W. Turner, R.R. Eckmann, J. Am. Chem. Soc.
111 (1989) 2728.
[29] R. Taube, L. Krukowka, J. Organomet. Chem. 347 (1988) C9.
[30] H. Schumann, D.F. Karasiak, S.H. Muhle, R.L. Haltermann,
W. Kaminsky, U. Weingarten, J. Organomet. Chem. 579 (1/2)
(1999) 356.
[5] 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. Strömberg, A.J.P. White, D.J. Williams, J. Am.
Chem. Soc. 121 (1999) 8728.
[6] E.C. Alyea, P.H. Merrell, Syn. React. Inorg. Chem. 4 (1974)
535.
[7] A. Köppl, H.G. Alt, S.J. Palackal, M.B. Welch, Phillips
Petroleum Co., US Patent 5,900,035 (1999).
[8] A. Andresen, H.G. Cordes, J. Herwiq, W. Kaminsky, A.
Merck, R. Mottweiler, J. Pein, H. Sinn, H.J. Vollmer, Angew.
Chem. 88 (1976) 889.
[31] R.F. Jordan, C.S. Bajgur, W.E. Dasher, A.L. Rheingold,
Organometallics 6 (1987) 1041.
[32] J.J.W. Esthuis, Y.Y. Tan, J.H. Teuben, J. Mol. Catal. 62 (1990)
277.
[33] L.A. Nekkaeva, G.N. Bondarenke, S.V. Ruykov, A.L.
Nekhaev, B.A. Krentsel, V.P. Marin, L.I. Vyshinskaya, I.M.
Khrapova, A.V. Polonskii, N.N. Korneev, J. Organomet.
Chem. 406 (1991) 139.
[9] A. Andresen, H.-G. Cordes, J. Herwiq, W. Kaminsky, A.
Merck, R. Mottweiler, J. Pein, H. Sinn, H.-J. Voilmer, Angew.
Chem. Int. Ed. Engl. 15 (1976) 630.
[10] H. Sinn, W. Kaminsky, Adv. Organomet. Chem. 18 (1980)
99.
[11] H. Sinn, W. Kaminsky, H.-J. Vollmer, R. Woldt, BASF AG,
US Patent 4,404,344 (1983).
[12] G.F. Schmidt, D.A. Hucul, R.E. Campbell Jr., Dow Chemical
Co., US Patent 5,015,749 (1991).
[34] H. Sinn, Makromol Chem., Makromol. Symp. 97 (1995) 27.