Bis(arylimino)pyridine iron(III) complexes as catalyst precursors for the oligomerization and polymerization of ethylene

Article abstract of DOI:10.1016/j.apcata.2011.06.010

A series of 26 bis(arylimino)pyridine iron(III) complexes containing either electron withdrawing or electron donating substituents in their ligand frameworks was synthesized and characterized. After activation with methylaluminoxane (MAO), these catalysts

Full text of DOI:10.1016/j.apcata.2011.06.010

Applied Catalysis A: General 403 (2011) 25–35
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Bis(arylimino)pyridine iron(III) complexes as catalyst precursors for the
oligomerization and polymerization of ethylene
Christian Görl∗, Tanja Englmann, Helmut G. Alt
Laboratorium für Makromolekulare Chemie, Universität Bayreuth, Universitätsstraße 30, Bayreuth D-95447, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 15 May 2011
Received in revised form 7 June 2011
Accepted 9 June 2011
Available online 16 June 2011
A series of 26 bis(arylimino)pyridine iron(III) complexes containing either electron withdrawing or elec-
tron donating substituents in their ligand frameworks was synthesized and characterized. After activation
with methylaluminoxane (MAO), these catalysts oligomerize or polymerize ethylene to give both linear
and branched products. In contrast to iron(II) complexes, the presence of at least one ortho-substituent
at the iminophenyl rings is not obligative for catalytic activities of the iron(III) complexes. A couple
of such iron(III) complexes containing meta- and para-substituted bis(arylimino)pyridine compounds
were accessible and their oligomerization behaviour revealed interesting differences to the well known
iron(II) analogues since both internal and branched olefins were found in the product mixtures beside
the expected linear ␣-olefins. The widths of the resulting molecular weight distributions and the degrees
of isomerization of the resulting oligomers strongly depend on the substitution pattern at the ligand
frameworks.
Keywords:
Bis(arylimino)pyridine iron complexes
Ethylene oligomerization
Isomerization of olefins
Polymerization
Polyethylene
© 2011 Elsevier B.V. All rights reserved.
1. Introduction
substituents at the iminophenyl rings can be exchanged with
aryl groups applying Suzuki coupling reactions. As an alternative
Gibson [1–4] and Brookhart [5–7] independently reported
the application of 2,6-bis(arylimino) pyridine iron complexes as
effective catalysts for the polymerization and oligomerization of
ethylene leading to highly linear products. Especially the sub-
stitution pattern of the iminophenyl rings has a great influence
on the polymerization activities and the product compositions
[1–17]. Iron complexes bearing small substituents (alkyl or halo-
gene) at the 2-positions of the iminophenyl rings proved to
be excellent catalyst precursors for the oligomerization of ethy-
lene to give low molecular weight ␣-olefins. These ␣-olefins
are industrially highly desired compounds which are useful,
e.g., for copolymerization reactions with ethylene to give linear
low density polyethylene (LLDPE). In this context, halogenated
2,6-bis(arylimino)pyridine compounds play an important role as
valuable ligand precursors and some of their transition metal
complexes are known in the literature [1,5,18–25]. In the past
five years, 2,6-bis(arylimino)pyridine iron complexes with halo-
gen substituted iminophenyl rings were reported by Qian et al.
and Ionkin et al. [8,9,14]. While fluoro, chloro, and bromo sub-
stituted compounds are known for a longer time [8–10,26–28],
the analogous iodo substituted compounds were described by our
group in 2007 [29]. As demonstrated by Ionkin et al. [14], bromo
method, Sonogashira coupling reactions of terminal alkynes with
2,6-bis(arylimino)pyridine compounds containing para-bromo and
para-iodo substituted iminophenyl rings proved to be a success-
ful pathway to give 2,6-bis(arylimino)pyridine compounds with
extended ligand frameworks [29]. The vast majority of these lit-
erature known complexes rely on iron(II) compounds. One of the
characteristics of 2,6-bis(arylimino)pyridine iron(II) complexes is
the fact that the iminophenyl rings of the ligand frameworks
must contain at least one substituent at the ortho-position to the
iminophenyl nitrogen atoms to be stable against ligand transfer
reactions. Although the synthesis of the iron(II) complex with the
tridentate ligand precursor 2,6-bis(1-(phenylimino)ethyl)pyridine,
a ligand without any substituents at the iminophenyl rings, was
described by Abu-Surrah et al. [30], this complex does not exist
in the common form (L)FeCl2 (L = bis(arylimino)pyridine ligand)
but can be isolated as an air stable ionic compound of the
composition [(L)2Fe]2+[FeCl4]2−. Analogous ion-pair complexes
were described for 2,6-bis(1-(2-fluorophenyl)ethyl)pyridine [28],
2,6-bis(1-(3,5-dibromo-4-methylphenyl)ethyl)pyridine [14], 2,6-
bis(1-(2,6-dibromophenyl)ethyl)pyridine [31], and 2,6-bis(1-(4-
nitrophenyl)ethyl)pyridine [13]. According to these results, methyl
groups seem to be the smallest ortho-substituents that prevent
this ligand transfer reaction. While a great deal of work has
been spent on bis(arylimino)pyridine iron(II) complexes, only a
few publications report about the corresponding iron(III) com-
plexes despite their higher catalytic activities towards olefin
∗
Corresponding author. Tel.: +49 921 55 2993; fax: +49 921 55 3206.
E-mail address: christian.goerl@uni-bayreuth.de (C. Görl).
0926-860X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2011.06.010

26
C. Görl et al. / Applied Catalysis A: General 403 (2011) 25–35
oligomerization and polymerization [14,24,29,32–39]. Although
both iron(II) and iron(III) complexes can be undoubtedly char-
acterized, the oxidation state of the iron centers (Fe(II) or
Fe(III)) after activation with aluminoxane cocatalysts and the
nature of the active species still remains unclear [4,36,40].
Actual DFT calculations by Cruz et al. [41,42] and Raucoules
et al. [43] give strong evidence for the enhanced oligomeriza-
tion/polymerization ability of active iron(III) species compared
with iron(II) species. However, iron centers in both oxidation
states may be present in the same catalyst system (“multi-
centered catalysts”) and lead, e.g., to polyethylenes with broad
or even bimodal molecular weight distributions [41,44]. The
increased activities of the bis(arylimino)pyridine iron(III) com-
plexes can be explained with the stronger Lewis acidic character
of iron(III) centers compared with iron(II) centers and, there-
fore, an enhanced affinity to coordinate electron rich olefin
molecules. Lopez Reyes [24], Devore [35], Bryliakov [36], and
Prades [37] describe some bis(arylimino)pyridine iron(III) com-
plexes with ortho-disubstituted iminophenyl rings useful for
ethylene polymerization reactions. As mentioned above, Ionkin
et al. [14,19] explored the reactivity of halogen substituted
bis(arylimino)pyridine compounds towards Suzuki coupling. Dur-
ing these studies, three iron(III) complexes were described along
with a huge number of iron(II) complexes which were applied
for ethylene oligomerization. The synthesis of ␻-alkenyl substi-
tuted bis(arylimino)pyridine compounds and their corresponding
iron(II) and iron(III) complexes was described by our group
[33]. While their ethylene oligomerization and polymerization
behaviour was superior compared with analogous alkyl substi-
tuted bis(arylimino)pyridine iron complexes, self-immobilization
[45] as described for metallocene complexes was not observed dur-
ing the oligomerization/polymerization runs. Due to the extreme
preference of the iron centers to ethylene molecules compared with
higher ␣-olefins, the incorporation rates of such higher olefins are
very low [46–48]. While the early literature mainly emphasizes the
influence of substituents at the ortho-positions of the iminophenyl
rings, more actual publications laid the focus on the meta- and
para-substituents [13,14,24,29]. The introduction of electron with-
drawing substituents into the ligand backbones resulted in an
enhanced temperature stability and, therefore, lead to higher activ-
ities. According to Gong [38], bis(arylimino)pyridine transition
metal complexes without any substituents at the iminophenyl
rings can be applied for the polymerization of 1,3-butadiene. These
results engaged us to synthesize bis(arylimino)pyridine iron(III)
complexes containing halogen or other electron withdrawing
substituents on the iminophenyl rings and to investigate their abil-
ities towards selective ethylene oligomerization/polymerization
including the question if catalytic activities are also observed when
the iminophenyl rings of the ligand frameworks do not contain
substituents in ortho-positions to the former amino groups.
30% in toluene) and Albemarle (Baton Rouge, USA/Louvain – La
Neuve, Belgium; 10% in toluene). Ethylene (3.0) and argon (4.8/5.0)
were supplied by Rießner Company (Lichtenfels). All other start-
ing materials were commercially available and were used without
further purification.
NMR spectra were recorded at 25 ◦C on a Varian inova 400 spec-
trometer. The chemical shifts in the 1H NMR spectra are referred to
the residual proton signal of the solvent (ı = 7.24 ppm for CDCl3) and
in 13C NMR spectra to the solvent signal (ı = 77.0 ppm for CDCl3).
EI mass spectra were routinely recorded at the Zentrale Analytik
of the University of Bayreuth with a VARIAN MAT CH-7 instru-
ment (direct inlet, E = 70 eV) and a VARIAN MAT 8500 spectrometer.
MALDI-TOF MS measurements were performed on a Bruker Dal-
tonic Reflex TOF using graphite as matrix. The laser intensity was set
to 60–70%. GC/MS spectra were recorded with a Thermo Focus gas
chromatograph in combination with a Thermo DSQ mass detector
(EI, 70 eV) using a HP-5MS GC column (length: 30 m, film thick-
ness: 0.25 ␮m, flow: 1.5 ml/min, split ratio: 1:50) and helium as
the carrier gas. The routinely used temperature program contained
a starting phase (2 min at 50 ◦C), a heating period (10 K/min for
24 min) and a plateau phase (15 min at 290 ◦C) resulting in a run
length of 41 min. At the Zentrale Analytik of the University of
Bayreuth, GC/MS spectra were routinely recorded with a HP5890
gas chromatograph in combination with a MAT 95 mass detector.
For the analysis of oligomer mixtures, GC spectra were obtained
with an Agilent 6890N gas chromatograph equipped with a HP-5
column (length: 30 m, film thickness: 1.5 ␮m, flow: 150 ml/min,
split ratio: 1:50). The temperature program included a starting
phase (6 min at 35 ◦C), two heating ramps (1 K/min up to 55 ◦C,
then 20 K/min up to 250 ◦C) and a plateau phase (20 min at 250 ◦C)
resulting in a run length of 55.75 min. This temperature program
allowed the separation of most of the hexene isomers. GPC mea-
surements were routinely performed by SABIC Company (Riyadh,
Saudi Arabia). Elemental analyses were performed with a VarioEl
III CHN instrument using acetanilide for calibration.
2.2. Synthesis of 2,6-bis(arylimino)pyridine compounds applying
molecular sieves and a heterogeneous SiO2/Al2O3 catalyst
(Method A)
To a solution of 2.6-diacetylpyridine (0.49 g; 3 mmol) in toluene
˚
˚
(20 ml) were added molecular sieves (4 A or 3 A; 15 g), the
corresponding amine or aniline compound (7 mmol), and the sil-
ica/alumina catalyst (0.5 g). The reaction mixture was heated to
45–50 ◦C for 24 h. If the reactions were not completed (according
to GC/MS analyses), the heating period was prolonged till comple-
tion. After cooling to room temperature, the mixture was filtered
over sodium sulfate, and the residue was washed several times
with toluene. After removal of the solvent, methanol was added
for precipitation. After storage at −20 ◦C for 24 h, the precipitated
bis(imino)pyridine compounds were isolated and dried in vacuo
(yields: 40–90%). Their spectroscopical data are given in the Sup-
porting Information (Tables A1 and A2).
2. Experimental
2.1. General considerations
2.3. Synthesis of 2,6-bis(arylimino)pyridine compounds applying
a Dean-Stark trap (Method B)
All experimental work was routinely carried out using Schlenk
technique. Dried and purified argon was used as inert gas. The
solvents n-pentane, diethyl ether, toluene und tetrahydrofuran
were purified by distillation over Na/K alloy. Diethyl ether was
additionally distilled over lithium aluminum hydride, toluene was
additionally dried over phosphorus pentoxide. Methylene chlo-
ride was dried over phosphorus pentoxide and calcium hydride.
Methanol and ethanol were dried over magnesium. 1-Butanol (p.a.)
was purchased from Merck and used without prior distillation.
Methylalumoxane was purchased from Crompton (Bergkamen;
To a solution of 2,6-diacetylpyridine (0.82 g; 5 mmol) in 150 ml
of toluene were added a substituted aniline (12.5 mmol; 2.5 equivs.)
and a few milligrams of para-toluenesulfonic acid. The reaction
mixture was heated under reflux for 8–48 h applying a Dean-
Stark-trap. After cooling to room temperature, a saturated sodium
hydrogencarbonate solution (200 ml) was added. The organic phase
was separated and filtered over sodium sulfate and silica. The sol-
vent was removed and methanol (20 ml) was added. The imino
compounds precipitated when stored at −20 ◦C for some days

C. Görl et al. / Applied Catalysis A: General 403 (2011) 25–35
27
(yields: 35–50%). Their spectroscopical data are given in the Sup-
porting Information (Table A1).
some ligand precursors were prepared using alkyl substi-
tuted anilines or cycloalkylamines. Condensation reactions using
2,4,6-triphenylaniline [11], 3-ethynylaniline, 4-ethynylaniline, 4-
ethynyl-2,6-dimethylaniline, and 1-aminoadamantane as amine
compounds did not lead to imino compounds but gave
condensation products of the corresponding anilines. How-
2.4. Sonogashira coupling reaction of 13 with phenylacetylene
In 15 ml of triethylamine, the iodo substituted 2,6-bis
(arylimino)pyridine compound 13 (0.35 mmol), phenylacety-
lene (0.70 mmol), bis(triphenylphosphino)palladium dichloride
(7 ␮mol) and copper(I) iodide (14 ␮mol) were dissolved. The mix-
ture was stirred for 20 h at room temperature. After removal of
the solvent, water (50 ml), n-pentane (50 ml), and methylene chlo-
ride (15 ml) were added. The organic phase was separated, and the
aqueous phase was extracted several times with a 3:1 (v/v) mixture
of n-pentane/methylene chloride. The combined organic phases
were dried over sodium sulfate. Removal of the solvent in vacuo
and recrystallization from methanol furnished the coupling prod-
uct 21 as a bright yellow powder (yield: 63%). The spectroscopical
data of 21 are given in the Supporting Information (Table A1).
ever,
a Sonogashira cross-coupling reaction of 2,6-bis[1-(4-
iodophenylimino)ethyl]pyridine [29] (13) with phenylacetylene
provided the phenylethynyl substituted compound 21 in good yield
(see Scheme 2). 2,6-Bis(arylimino)pyridine compounds contain-
ing electron withdrawing substituents on their iminophenyl rings
were found to be much more sensitive towards air and moisture
(easy hydrolysis of the imine groups!) compared with alkyl or
aryl substituted analogues. Additionally, all iodo substituted lig-
and precursors appeared to be light sensitive. Former literature
usually reported the use of para-toluenesulfonic acid as a cata-
lyst for the condensation reactions of 2,6-diacetylpyridine with
aniline derivatives (Method B) [2,5,49]. The great disadvantage of
this method is the comparably high reaction temperature of about
130 ◦C due to azeotropic water removal using a Dean-Stark trap and
the poor solubility of para-toluenesulfonic acid in toluene at lower
temperatures. As a result, some temperature sensitive aniline and
amine compounds decomposed under these quite harsh reaction
conditions and did not provide the desired bis(arylimino)pyridine
compounds. To overcome these problems, an elegant method intro-
duced by Qian et al. [8,50] was applied using a heterogeneous
silica/alumina catalyst (13% Al2O3/87% SiO2) and molecular sieves
2.5. General synthesis of the 2,6-bis(arylimino)pyridine iron(III)
complexes 28–54
An amount of 1 mmol of the 2,6-bis(arylimino)pyridine com-
pound was dissolved in 1-butanol (20 ml; alternatively, THF or
a 1:1 mixture of THF and diethylether were used) and reacted
with anhydrous iron(III)chloride (1 mmol) resulting in an imme-
diate colour change to brown, orange, or red). The mixture was
stirred for 3 h at room temperature (for 54 derived from benzy-
lamine: t = 5 min), whereby the complexes precipitated. n-Pentane
(20 ml) was added for complete precipitation, and the mixture was
stirred for another 15 min. The iron complexes were filtered over
a glass frit, washed three times with 15 ml n-pentane, and dried
in vacuo (yields: 50–95%). The data of MS and elemental analyses
of complexes 28–54 can be found in the Supporting Information
(Table A3).
˚
3 A as the water absorbing agent (Method A). The heterogeneous
catalyst acts as a Lewis acid and, additionally, as a scavenger for
impurities in the aniline compounds. Due to the lower reaction
temperatures (40–45 ◦C), no decomposition products were found
in the reaction mixtures and the bis(imine) compounds could be
readily isolated. However, nitro substituted aniline derivatives did
not give the corresponding bis(imine) compounds (see compounds
19 and 20 in Table 1) under these mild reaction conditions. Apply-
ing Method B, compounds 19 and 20 could be isolated in moderate
yields and were found to be extremely sensitive towards hydrolysis
(= back reaction). Table 1 gives an overview of the prepared 2,6-
bis(arylimino)pyridine compounds, while the 1H NMR, 13C NMR,
19F NMR, and MS data for compounds 1–27 can be found in the
Supporting Information (Tables A1 and A2).
2.6. Polymerization of ethylene in the 1 l Büchi autoclave
The desired iron complex (0.2–2 mg) was suspended in toluene
(5 ml). Methylaluminoxane (30% in toluene or 10% in toluene) was
added to maintain a ratio Fe:Al = 1:2500 resulting in an immediate
colour change. The mixture was added to a 1 l Schlenk flask filled
with 250 ml n-pentane. This mixture was transferred to a 1 l Büchi
laboratory autoclave under inert atmosphere and thermostated at
60 ◦C. An ethylene pressure of 10 bar was applied for 1 h. After cool-
ing down to 15 ◦C, the ethylene pressure was carefully released.
To the oligomer/polymer mixtures was added diluted hydrochloric
acid resulting in biphasic systems. The polymers were filtered off
using a glass frit, washed with water and acetone, and finally dried
in vacuo.
3.2. Synthesis of the iron(III) complexes
The bis(arylimino)pyridine compounds 1–27 were reacted
with iron(III)chloride either in n-butanol or a 1:1 (v/v) mix-
ture of diethylether and THF to give the corresponding
bis(arylimino)pyridine iron(III) complexes 28–54 (Scheme 3 and
Table 1) which could be isolated as orange or brown solids. When
2,6-bis[1-(2-iodophenylimino)ethyl]pyridine (5) was reacted with
iron(III)chloride, the complexation reaction failed and no product
could be obtained. This result may be explained with the increased
steric bulk caused by the iodo substituents at the ortho-positions
to the imino nitrogen atoms compared with other halogen sub-
stituents. Except complex 28 which was described by Gong et al.
[38] and used for the polymerization of 1,3-butadiene, complexes
29–54 are not known in the literature. Among these complexes,
there are the first examples for bis(arylimino)pyridine iron(III)
complexes whose ligands do not contain ortho-substituents at the
aryl rings (33–40, 46, 48–50).
The oligomer solutions were separated and dried over sodium
sulfate. Most of the solvent n-pentane was distilled off using a
Vigreux column. The resulting oligomer mixtures were character-
ized by gas chromatography.
3. Results and discussion
3.1. Preparation of substituted 2,6-bis(arylimino)pyridine
compounds
As described in the introduction chapter, bis(arylimino)pyridine
compounds without or with small substituents at the ortho-
positions of the iminophenyl rings immediately form ion-pair
complexes in reactions with iron(II)chloride. However, some exam-
ples for such iron(II) complexes were described as (L)FeCl2 type
complexes (e.g., see Ref. [30]), but after thoroughly reviewing
Condensation reactions of 2,6-diacetylpyridine with ani-
lines containing electron-withdrawing substituents yielded the
2,6-bis(arylimino)pyridine compounds (Scheme 1). Besides halo-
gen substituted anilines, aromatic amines containing phenoxy,
phenylthio, or nitro groups were applied. For comparison purposes,

28
C. Görl et al. / Applied Catalysis A: General 403 (2011) 25–35
Scheme 1. General synthesis of 2,6-bis(arylimino)pyridine compounds.
Scheme 2. Synthesis of bis(arylimino)pyridine compound 21 via palladium catalyzed Sonogashira coupling reaction.
Table 1
Synthesized 2,6-bis(arylimino)pyridine compounds and their corresponding iron(III) chloride complexes.
bis(Imino) compound
R1
R2
R3
R4
R5
Method of synthesis
Iron(III) complex
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
H
F
Cl
Br
I
H
H
H
H
H
H
H
H
F
Cl
Br
H
H
H
H
H
F
Cl
Br
I
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
F
Cl
Br
I
F
Cl
Br
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
t-Bu
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
28
29
30
31
[32]
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
O-phenyl
S-phenyl
H
Me
H
H
H
NO2
NO2
Phenylethynyl
H
H
A
A
B
B
H
H
H
H
t-Bu
t-Bu
Scheme 2
H
H
A
A
A
A
A
A
Amine compound: cyclopentylamine
Amine compound: cyclohexylamine
Amine compound: 2-methylcyclohexylamine
Amine compound: benzylamine
Scheme 3. Synthesis of the bis(arylimino)pyridine iron complexes 28–54.

C. Görl et al. / Applied Catalysis A: General 403 (2011) 25–35
29
Scheme 4. Literature known bis(arylimino)pyridine iron(II) complexes of the type (L)FeCl2 lacking ortho-substituents at the iminophenyl groups.
these articles only three complexes seem to exist in the state
(L)FeCl2 [29,51,52] (see Scheme 4). All other compounds should
be expressed as ion pairs [L2Fe]2+ [FeCl4]2− according to the given
experimental data (a good indicator is the deep purple or violet
colour of these ion pair complexes).
In contrast, reactions of such bis(arylimino)pyridine compounds
with iron(III)chloride cleanly provided the desired mono-ligated
complexes of the general formula (L)FeCl3. However, the resulting
complexes are not stable towards ligand transfer reactions when
they are kept in polar solvents (like THF or alcohols, see Refs. [14]
and [36]) for some hours. Additionally, at longer reaction times
these iron(III) complexes can act as oxidation reagents towards the
solvent THF resulting in the abstraction of hydrogen atoms, the loss
of HCl, and the formation of the corresponding iron(II) complexes
[53–55].
Therefore, the time limit for the complexation reactions and
the subsequent isolation procedures was determined to about 3 h.
The iron complex 54 (ligand derived from 2,6-diacetylpyridine
and benzylamine) exhibited an extreme instability in the cho-
sen reaction media (1-butanol or a 1:1 mixture of Et2O and THF)
and decomposed after about 15 min to give the ion-pair complex
[L2Fe]3+ [FeCl6]3− (as indicated by MS analyses). Since the desired
mono-ligated complex immediately started to precipitate after the
addition of iron trichloride to a solution of the bis(imine) compound
27, the reaction was worked up after 5 min and complex 54 could
be obtained in good yield (76%). All complexes were stored in the
dark to prevent the iron(III) centers from reduction to iron(II), as
light also seemed to have a detrimental influence on the complex
stabilities.
The iron(III) complexes were characterized by mass spectrome-
try (EI-MS and MALDI-TOF MS) and elemental analyses. Compared
with analogous iron(II) complexes, the iron(III) complexes are less
stable to fragmentation in EI-MS, so in most cases the molecular
ion was not detectable. In contrast, MALDI-TOF MS analyses gave
more significant results and the molecular ions could be found for
all complexes. As expected, mass distributions due to the presence
of chloro atoms (and in some cases additionally bromo atoms) were
obtained. Due to their paramagnetism, 1H NMR analyses of these
iron complexes appeared to be less significant, since no defined
signals could be found in the range between −2000 and 2000 ppm
(relative to Me4Si). The MS and elemental analyses data of com-
plexes 28–54 are given in the Supporting Information (Table A3).
Scheme 5 shows the MALDI-TOF mass spectrum of complex
36. Around m/z = 728, the molecular ion mass distribution was
Table 2
Ethylene oligomerization and polymerization results for the iron complexes 28–54 (solvent: 250 ml n-pentane, activator: MAO, Fe:Al = 1:2500, 10 bar ethylene, 1 h). The
oligomer share is defined as the content of liquid ␣-olefins in the product mixture.
Com-plex
Reaction
Activity [kg
prod./mol Fe·h]
Oligomer share
[wt.%]
Schulz–Flory
coefficient, ˛
C6 in the product
mixture [%]
1-Hexene in
the C6 fraction [%]
temperature [◦C]
28
29
29
29
30
30
30
31
31
31
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
60
40
60
80
40
60
80
40
60
80
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
60
4150
21530
4260
1850
87420
13770
13380
63540
19065
4410
7320
52050
22000
19070
4485
3250
4680
22825
25470
7760
9000
240
100
99
100
100
91
95
98
61
84
93
100
100
100
100
100
100
100
100
96
90
73
100
100
100
60
100
100
100
–
–
0.27
0.71
0.39
0.36
0.91
0.71
0.69
0.97
0.87
0.72
0.36
0.34
0.34
0.36
0.43
0.29
0.28
0.38
0.64
0.83
0.85
0.37
0.35
0.47
0.87
0.55
0.25
0.20
–
52.1
32.4
37.0
43.1
16.2
23.7
26.5
9.3
49.3
46.8
65.7
70.3
>99
97.9
97.0
> 99
15.2
28.9
40.2
26.4
26.5
42.9
41.7
54.0
51.5
46.0
31.4
14.2
14.6
38.5
57.4
39.1
13.7
34.6
44.5
47.1
–
98.6
98.1
56.3
68.6
80.8
55.4
41.5
49.3
49.3
45.9
60.1
95.7
98.6
87.6
92.0
57.6
93.5
57.5
70.9
73.0
–
940
14510
14750
8370
7900
20400
0
0
–
–
–
76.2
48.1
Traces of hexene isomers
100
100a
39.6
285
0.65
a
Only traces of hexenes were found in the product mixture.

30
C. Görl et al. / Applied Catalysis A: General 403 (2011) 25–35
Scheme 5. MALDI-TOF mass spectrum of complex 36 (ions with m/z < 500 were omitted for clarity).
observed. The base peak at m/z = 691/693 resulted from the loss
of one chloro ligand, while the ion at m/z = 656/658 had its origin
in the loss of two chloro ligands. At m/z = 566, an ion correspond-
ing to the protonated ligand [Mligand + 1] was detected. Similarly
to complex 36, MALDI-TOF mass spectra of most of the complexes
revealed that the base peaks mainly result from the loss of either
one or two chloro ligands (see Supporting Information, Table A3).
Slow diffusion of toluene into dichloromethane solutions of
complexes 36 and 37 afforded thin orange brown colored nee-
dles. However, due to their geometry these crystals were found
to be unsuitable for X-ray analyses. Similarly to the three literature
known crystal structures of bis(arylimino)pyridine iron(III) com-
plexes [14,32,38], the R values were too high (corresponding to an
increased uncertainty for the geometry of the molecule), so, finally,
the refinement procedures could not be finished successfully.
in Table 2) consisted mainly of linear low molecular weight ␣-
olefins and, in some cases, branched olefins. These mixtures can, to
some extent, dissolve higher molecular weight olefins (up to ∼C50),
however, their detected amounts decreased dramatically with
increasing carbon numbers. The obtained oligomer mixtures were
characterized by gas chromatography (olefins with chain lengths
of C4–C40 could be detected via GC or GC/MS analyses) while the
polymers were characterized using GPC analyses (Table 3).
For nearly all oligomerization/polymerization runs, very high
initial ethylene uptake rates could be observed (∼10 l of ethy-
lene/min). Due to the gradual catalyst degradation, the ethylene
flows drop distinctively during the experiments, and for the sys-
tems 28/MAO, 33/MAO and 54/MAO the ethylene flow had ceased
completely at the end of the regular run time (1 h) corresponding
to a complete deactivation of the active catalyst species. The “life
time” of a catalyst could be determined by the period in which
an ethylene flow was detectable. All 2,6-bis(arylimino)pyridine
iron complexes bearing at least one non-hydrogen substituent
at the ortho-positions of the iminophenyl rings in their ligand
frameworks were found to be still catalytically active after 1 h.
In contrast, the bis(cycloalkylimino)pyridine iron complexes 51
and 52 were completely inactive in ethylene oligomerization reac-
tions, and complex 53 only afforded traces of hexenes. Since the
sizes of cyclohexyl and phenyl rings are quite similar, this dra-
matic effect most likely originates from different electron densities
around the iron centers. Cycloalkyl groups stabilize the cationic
iron centers due to their electron donating nature (+I effect), so
the coordination of ethylene molecules is not favored. Complex 54
bearing benzyl groups in the ligand framework exhibits a slightly
higher electron deficit compared with the cycloalkyl derived com-
plexes 51–53. The coordination of an ethylene molecule therefore
becomes more facile, and a moderate oligomerization activity was
observed (285 kg prod./mol Fe·h). Complexes 28–50 whose lig-
ands are derived from substituted anilines showed distinctively
higher activities due to their greater lack of electron density at
the positively charged iron centers compared with the catalyst
precursors 51–54. Depending on the size, the nature, and the
position of a substituent at these aryl rings, a wide range of
olefinic products could be obtained from the activated complexes
28–50 and 54. At 60 ◦C (the routinely used reaction tempera-
ture for oligomerization/polymerization experiments), complexes
30–31 and 41–43 containing 2-halo- or 2,4-dihalo substituted
iminophenyl rings gave mixtures of liquid olefins and low molecu-
lar weight polyethylenes (see Table 2). With increasing the size of
the halogen substituent, the amount of polyethylene also increased
while correspondingly the oligomer shares, i.e., the liquid mass
3.3. Results of the catalytic oligomerization and polymerization of
ethylene
3.3.1. Oligomerization/polymerization activities and product
molecular weight distributions
The iron complexes 28–54 were applied as catalyst precur-
sors for the homogeneous polymerization and oligomerization
of ethylene. The complexes were activated with methylalu-
minoxane (MAO) applying a ratio Fe:Al = 1:2500. Oligomeriza-
tion/polymerization runs were routinely performed in a 1 l Büchi
steel reactor at a temperature of 60 ◦C over 1 h employing an ethy-
lene pressure of 10 bar. As a solvent n-pentane was used. Due to
the big differences in their activities, the complexes were used in
appropriate molar amounts to keep the exothermic reaction under
control. The oligomerization/polymerization results are given in
Tables 2 and 3. The liquid fractions (noted as “oligomer share”
Table 3
GPC analyses of the polyethylenes obtained at a polymerization temperature of
40 ◦C.
Complex
Polymerization
GPC data of the PE fractions
temperature [◦C]
Mn [g/mol]
Mw [g/mol]
PD
6.62
76.93
4.41
3.28
6.22
29
30
31
41
42
43
47
40
40
40
60
60
60
60
410
535
380
750
640
655
720
2690
41,150a
1680
2460
4010
6950
33,070
10.61
45.80
a
Bimodal MWD.

C. Görl et al. / Applied Catalysis A: General 403 (2011) 25–35
31
Fig. 1. GC spectrum of the oligomer mixture produced with 29/MAO at 40 ◦C (C4–C10 region; signals for higher oligomers were omitted for clarity). Offset: enlarged view of
the C6 region (hexenes).
fractions of the product mixtures, became lower. The catalytic sys-
tems 30/MAO and 31/MAO gave mixtures with oligomer contents
of 95%, and 84%, respectively, while the dihalo substituted com-
plexes 41–43 lead to mixtures containing 96%, 90%, and 73% of
liquid products. At 40 ◦C, also 29/MAO produced a small amount
of low molecular weight polyethylene. For both the 2-halo and the
2,4-dihalo substituted complexes, this result can be referred to the
increased steric bulk of the bromo substituents compared with the
much smaller fluoro and chloro atoms. The initial publications of
Brookhart and Gibson [1–7] depicted the same effect for differently
alkyl substituted 2,6-bis(arylimino)pyridine iron(II) complexes.
While the trends of the product compositions of complexes 29–31
and 41–43 are quite similar, their oligomerization/polymerization
activities exhibited dramatic differences. Within the series 29–31
(containing 2-halo substituted compounds), the 2-chloro deriva-
tive 30 showed an extremely high activity of 87,420 [kg prod./mol
Fe·h] which is distinctively higher compared with the 2-bromo
and the 2-fluoro derivatives 31 and 29. The “medium”-sized chloro
atoms apparently leave enough space for the coordination of ethy-
lene molecules and the growing chain, whereas the ortho-bromo
atoms may partially interfere with incoming ethylene molecules
leading to a somewhat lower activity (63,540 [kg prod./mol Fe·h]).
The comparatively low activity of the 2-fluoro substituted complex
29 could be explained with a faster deactivation of the catalytically
active centers due to the above described ligand transfer reac-
tion. A visible proof of that assumption was the purple colour of
the resulting oligomer mixture. The small size of the ortho-fluoro
atoms and the polar structure of the activated catalyst molecules
favors this ligand transfer reaction, so its rate constant becomes
dominant over the rate constant of ethylene insertion. A com-
pletely different result was found for the 2,4-dihalo substituted
complexes 41–43, where the 2,4-difluoro substituted compound
41 showed a much higher activity compared with the heavier
analogues 42 and 43. The introduction of halogen substituents
in the para-positions of the iminophenyl rings proved to be ben-
eficial in some cases [8,9,23,49,56,57], and the highest activities
for this complex type were found for 2-methyl-4-halogen substi-
tuted bis(arylimino)pyridine iron complexes [29,58]. Only a few
substitution patterns at the iminophenyl rings allowed such high
activities, and the presence of more than one halogen substituent
per iminophenyl ring makes explanations of reactivities more dif-
ficult, however, it can be stated that combinations of an electron
donating group (at ortho or para position) and an electron with-
drawing substituent (at para or ortho position) within one aryl
ring gave the best activities. As described by Ionkin [32] and
Zhang [59], ligand systems of oligomerization/polymerization cat-
alyst precursors must be electronically flexible to accept a partial
positive charge after activation with an aluminoxane co-catalyst.
Since the heavier halogen atoms can be polarized easier com-
pared with the small fluorine atoms, the complexes 42 and 43
containing four chloro, respectively, bromo substituents in their
ligand backbones may be better stabilized and are, therefore, less
prone to coordinate olefin molecules. Reduced catalytic activities of
iron catalysts containing para-chloro and para-bromo substituted
bis(arylimino)pyridine ligands compared with their fluoro or iodo
analogues were also observed for the para-substituted complexes
37–40 (see below) and were already described in the literature
[29,58].
The Schulz–Flory coefficient ˛ was found to be a useful param-
eter for the description of the product compositions, i.e., the
molecular weight distributions of the oligomer mixtures [60–65]. It
can be calculated from both the reaction rate constants and the GC
integrals of the oligomer fractions. Since the GC integrals are easily
accessible and are related to the molar amounts of the oligomers, ˛
can be defined as the quotient of the molar amounts of two subse-
quent oligomer fractions whose carbon numbers differ by a number
of 2:
kpropagation
mol(Cn+2
mol(Cn)
)
˛ =
=
(1)
kpropagation + ktermination
A higher coefficient ˛ directly corresponds to an increased propa-
gation probability resulting in higher molecular weight products.
The upper limit ˛ = 1 is only reached when molecular weight dis-
tributions are obtained in which the maximum peak appears at

32
C. Görl et al. / Applied Catalysis A: General 403 (2011) 25–35
Fig. 2. GC spectrum (taken from GC/MS analyses) of the oligomer mixture produced with 30/MAO at 40 ◦C (C6–C20 region). Offset: enlarged view of the C6 region (hexenes).
The peak for the solvent n-pentane was omitted.
higher carbon numbers (usually butenes or hexenes are the main
products) but in these cases the mathematical requirements for a
Schulz–Flory distribution are not fulfilled (an ideal Schulz–Flory
distribution is obtained for ˛ ∼ 0.7). Equation 1 shows that higher
values for ˛ are equivalent to increased amounts of higher olefins in
the product mixtures. Since such mixtures can be hardly separated,
it is desirable to design more selective catalysts. Many efforts have
been made especially to prepare catalysts which produce, e.g., only
1-hexene or 1-octene with high selectivities, since these ␣-olefins
are consumed in large amounts as comonomers in the synthesis of
linear low density polyethylenes (LLDPE). At present, chromium
catalysts based on [PNPN] ligands represent the upper limit for
the selective trimerization of ethylene to give 1-hexene ([66] and
references therein).
During our investigations of bis(arylimino)pyridine iron(II) and
iron(III) catalysts, we found that iron(III) catalysts usually pro-
duced mixtures of shorter chain length olefins compared to their
iron(II) analogues bearing the same bis(arylimino)pyridine
ligand [33,67]. The surprising result that 2,6-bis(1-(4-
iodophenylimino)ethyl)pyridine iron(II) chloride was active
in ethylene oligomerization [29] encouraged us to evaluate the
oligomerization behaviour of analogous iron(III) complexes in
more detail. As evident from Table 2, complexes 29–31 and 41–43
produced broad molecular weight distributions, and in all cases
low molecular weight polyethylenes were obtained besides the
desired ␣-olefins. The Schulz–Flory coefficients ˛ increased with
enlarging the sizes of the ortho-halogen substituents. Additional
para-halogen substituents in complexes 41–43 lead to substan-
tially lower ˛ values indicating a reduced steric bulk around the
catalytically active center. Probably, the para-substituents lead to
a better separation of the cationic iron center and the anionic MAO
cages.
␣-olefins (29 and 41), but additionally gave internal olefins (see
Fig. 1).
Similar product mixtures containing both ␣- and internal olefins
were also obtained with bis(arylimino)pyridine vanadium(III) cat-
alysts [23,67]. According to early publications of Brookhart and
Gibson [5,56], iron complexes bearing one ortho-methyl group at
the iminophenyl rings exclusively produced highly linear ␣-olefins
indicating that these methyl groups are voluminous enough to sup-
press isomerization reactions of terminal into internal olefins. For
the bulkier ortho-chloro and ortho-bromo substituted analogues
(30, 31, 42, and 43), mainly linear ␣-olefins (>95%) were detected
by GC analyses (see Fig. 2 for example) confirming this assumption.
The temperature dependences of the oligomer shares and,
correspondingly, the Schulz–Flory values were investigated for
complexes 29–31. Therefore, complexes 29–31 were additionally
tested for ethylene oligomerization at 40, 60, and 80 ◦C. For all
three catalysts, dramatically reduced activities were observed at
elevated temperatures corresponding to a faster deactivation pro-
cess. As could be expected, complex 29 was least stable among
these three catalysts due to its small ortho-fluoro substituents.
The oligomer shares increased at elevated temperatures, however,
complexes 30 and 31 still produced small amounts of low molecular
weight polyethylenes at 80 ◦C. For complexes 30 and 31, the result-
ing ˛ values decreased at higher temperatures, and were found to
fit a more ideal Schulz–Flory distribution (˛ = 0.69 for 30 at 80 ◦C,
˛ = 0.72 for 31 at 80 ◦C), while a dramatic decrease of the ˛ value was
observed for complex 29 (see Table 2) which predominantly pro-
duced 1-butene, 1-hexene, and 1-octene at 80 ◦C along with some
isomerization products (see Section 3.3.2).
Since the activated complex 29 appeared to be a good approach
to a catalyst system that produces low molecular weight ␣-olefins
more selectively, complexes with a further reduced steric bulk at
the ortho-positions of the iminophenyl rings looked very promis-
ing to give narrower Schulz–Flory distributions. This was realized
with complexes 28, 33–40, 46, and 48–50 which only contain
hydrogen atoms at the ortho-positions of the iminophenyl rings.
The size of the ortho-substituent at the iminophenyl rings also
effected the isomerization behaviour of the bis(arylimino)pyridine
iron(III) catalysts. If these ortho-substituents were either hydrogen
or fluorine atoms, the catalysts produced not only the expected

C. Görl et al. / Applied Catalysis A: General 403 (2011) 25–35
33
Fig. 3. C4–C8 region (content of oligomers with C > 8 is lower than 2%) of the oligomer mixture produced with 50/MAO.
As 28 was described to be catalytically active in the polymer-
ization of 1,3-butadiene [38], it seemed reasonable to test this
complex also for the oligomerization of ethylene. Indeed, after
activation with MAO, complex 28 exhibited a moderate activity
in ethylene oligomerization (4150 kg prod./mol Fe·h), however,
due to its lack of any substituents at the iminophenyl rings, com-
plete deactivation had to be noted after a run time of only 10 min
when no more ethylene was consumed. The Schulz–Flory coeffi-
cient was determined to ˛ = 0.27 indicating a very narrow product
distribution.
described by our group [29] which may be explained with the size
and the good polarizability of the voluminous iodo atoms due to
partial compensation of the positive charges at the iron centers
during the oligomerization process. The high activity (22,825 kg
prod./mol Fe·h) of 40/MAO is therefore in accordance with our
former results. In the presence of the smaller para-halogen sub-
stituents in complexes 37–39, distinctively lower activities were
noticed corresponding to less shielded active iron centers.
Differences in the Schulz–Flory coefficients were observed
when considering ortho-, meta- and para-halogen substituted
bis(arylimino)pyridine iron complexes. As expected, ˛ increases
in the 2-halogen substituted series 29–31 (29: ˛ = 0.71 < 30:
˛ = 0.91 < 31: ˛ = 0.97) with increasing size of the halogen sub-
stituent. The meta- and para-substituted derivatives 33–40 yielded
substantially lower ˛ values (0.28 < ˛ < 0.43). For the meta-halogen
series 33–36, ˛ remains nearly constant (˛ ∼ 0.35) with increas-
ing size of the halogen substituent, but no general trend could
be observed for the para-series 37–40. The narrowest molec-
ular weight distribution among the mono-halogen substituted
bis(arylimino)pyridine iron(III) complexes was found for the
para-bromo substituted complex 39 (˛ = 0.28). The para-halogen
substituted catalysts showed the highest tendencies towards
isomerisation products, and remarkably low 1-hexene contents
(<50%) among the C6 fractions were found (see Table 2).
Besides halogen substituted bis(arylimino)pyridine com-
pounds, complexes containing other electron withdrawing
substituents at their bis(arylimino)pyridine ligand frameworks
(44–47) also proved to be active in ethylene oligomerization reac-
tions. Literature known nitro substituted bis(arylimino)pyridine
iron(II) catalysts proved to be thermally more stable compared
with alkyl functionalized analogues [13]. Complex 46 containing
para-nitro groups at the iminophenyl rings showed a lower activ-
ity compared with the 4-iodo derivative 40 and, additionally, a
broader molecular weight distribution (˛ = 0.47 for the oligomer
mixture obtained with 46/MAO). Similarly to the ortho-halogen
substituted complexes 29–31, complex 47 produced a quite
high amount of low molecular weight polyethylene, however,
its productivity was much lower (14,750 kg prod./mol Fe·h). The
strong electron withdrawing effect of the nitro groups leads to an
Regarding the fast deactivation of 28, complexes 33–40 bear-
ing meta and para mono-halogen substituted iminophenyl rings
were considered to exhibit an improved shield of the iron center.
Similarly to 28, these complexes produced low molecular weight
olefins, and the determined Schulz–Flory coefficients range from
0.28 (for 39/MAO) to 0.43 (for 37/MAO). Compared with the ortho-
halogen functionalized complexes 29–31, the activities of the meta-
and para-halogen substituted complexes 33–40 were lower, as
partial deactivation by ligand transfer reactions occurred. While
among the meta-halogen substituted complexes 33–36 the chloro
compound 34 gave the highest activity, complex 40 bearing iodo
substituents at the ligand backbones proved to be the most active
candidate in the para-halogen substituted series 37–40. With an
activity of 52,050 [kg prod./mol Fe·h], 34/MAO actually is the most
active bis(arylimino)pyridine iron catalyst which does not con-
tain any substituents at the ortho-positions to the former amino
nitrogen atoms. Similarly to the ortho-substituted series 29–31, the
chloro compound revealed to be the most active candidate in ethy-
lene oligomerization reactions among the meta-series 33–36 what
may be explained with the medium size of the chloro atoms and
their high electronegativity. The latter leads to a strong decrease of
the electron density in the iminophenyl rings favouring the coordi-
nation of ethylene molecules. However, if chloro substituents are
introduced at the para-positions (complexes 38 and 42), the activ-
ities dropped. These results are in accordance to Pelascini et al.
[11] who also described a negative effect of chloro substituents
on the polymerization activities of a series of iron(II) and cobalt(II)
catalysts. The beneficial effect of introducing iodo substituents at
the para-positions of bis(arylimino)pyridine ligands was already

34
C. Görl et al. / Applied Catalysis A: General 403 (2011) 25–35
via GC analyses, since branched and internal alkenes are obtained
along with the expected 1-alkenes. Isomerization of terminal into
internal olefins probably proceeds through the coordination of an
␣-olefin and its insertion in a 2,1-mode into the postulated hydrido
iron species with subsequent ␤-hydrogen elimination to yield an
internal olefin. These results are in accordance with the calculations
of Ramos [46] regarding the inability of bis(arylimino)pyridine iron
catalysts bearing ortho-methyl substituents on the iminophenyl
rings to copolymerize ethylene with higher ␣-olefins.
The degrees of isomerization of the produced ␣-olefins, espe-
cially 1-hexene, which were observed for 29 at three different
temperatures, became interestingly lower at elevated temperature
(40 ◦C: 46.8%; 60 ◦C: 65.7%; 80 ◦C: 70.3%), and at 80 ◦C the high-
est relative content of 1-hexene (among the C6 fraction) could be
detected (see Table 2). This result could again be explained with an
accelerated deactivation process (ligand transfer) which does not
allow the “recoordination” and subsequent isomerization of a pro-
duced ␣-olefin molecule to the active center to a greater extent.
Less steric bulk around the iron centers most likely favors the
coordination of 1-butene or higher ␣-olefins which may then fur-
ther undergo 1,2- or 2,1-insertion either into a postulated cationic
iron hydride species [7,33,68] or into an iron-alkyl species (= the
growing chain). The different ways of coordination and subsequent
chain termination reactions lead to a broad spectrum of linear and
branched isomers. Fig. 1 shows the C6 fraction of the oligomer
mixture produced with 29/MAO at 40 ◦C along with an assign-
ment of the detected isomers which was proven by GC analyses
of the pure hydrocarbons. Obviously, both branched and internal
olefins were produced, however, the overall selectivity towards lin-
ear C6 alkenes (1-hexene, cis/trans-2-hexene) still exceeded 90%.
Compared to the results for complex 29, the degree of isomer-
ization of the hexene fraction for the catalytic system 28/MAO
is very similar, and the relative amount of 1-hexene in the C6
fraction was calculated to 49.3%. Besides the complexes 28 and
29, the para-halogen substituted complexes 37–40 exhibited the
greatest tendencies towards isomerization reactions, since the con-
tents of 1-hexene in the C6 fractions of the product mixtures was
always lower than 50% (see Table 2). Among this series, 37/MAO
only gave 41% of 1-hexene. The C6 region in the GC spectrum of
the oligomer mixture obtained with 39/MAO is shown in Fig. 4.
Beside the expected 1-hexene (retention time: 6.43 min), signals
for cis- and trans-4-methyl-2-pentene, 2-methyl-1-pentene, cis-
and trans-2-hexene, 2-methyl-2-pentene, 2-ethyl-1-butene, and
2,3-dimethyl-2-butene could be assigned by comparison of the GC
retention times of the pure compounds and further GC/MS corre-
lation using the NIST mass spectral database.
According to the assignments in Fig. 4, the para-halogen
substituted complexes 37–40 also afforded branched oligomers.
While the occurrence of linear internal olefins can be explained
with recoordination of the ␣-olefins to the iron center in a 2,1-
insertion mode and subsequent ␤-hydrogen elimination, branched
products additionally require the incorporation of at least one
ethylene or 1-butene molecule into the growing chain (see “chain
running” mechanism known for bis(imine)nickel catalysts [69])
before ␤-hydrogen elimination takes place. Especially the exclu-
sively para-substituted complexes produced remarkable amounts
of branched oligomers indicating that the lack of substituents in
both ortho and meta positions provided sufficient space for ade-
quate coordination of “non-ethylene” olefins. In conclusion, the
iron(III) complexes with ortho-unsubstituted iminophenyl rings
display the first examples of this complex type which are able to
isomerize and copolymerize small olefins (except an ␻-alkenyl sub-
stituted complex described by our group [33] for which another
chain propagation mechanism was proposed).
Fig. 4. Hexene fraction in the gas chromatogram of the oligomer mixture produced
with 39/MAO.
increased affinity of the cationic iron center towards electron rich
ethylene molecules resulting in longer oligomer/polymer chains.
Very low activities were recorded for the phenoxy- and
phenylthio substituted complexes 44 and 45 (see Table 2). The pres-
ence of oxygen and sulfur atoms in the ligand frameworks causes a
special deactivation reaction, since these heteroatoms are assumed
to coordinate irreversibly to the cationic iron centers. Despite their
low activities, the resulting product molecular weight distributions
were very narrow (˛ = 0.35–0.37) and only ␣-olefins were obtained.
For comparison purposes, complexes 48–50 containing meta-
alkyl (49 and 50) or para-phenylethynyl (48) substituted
bis(arylimino)pyridine ligands were tested for the oligomerization
of ethylene. While the catalytic system 48/MAO showed a moderate
activity of 8370 kg prod./mol Fe·h which is similar to the 4-halogen
substituted complexes 37–40, the resulting product distribution
(˛ = 0.55) is much broader compared with the distributions of the
4-halogen derivatives.
Since halogen substituents at the meta-positions of the
iminophenyl rings lead to narrow product molecular weight dis-
tributions (33–36) and, therefore, low values for the Schulz–Flory
coefficient ˛, complexes 49 and 50 revealed to be interesting
candidates for ethylene oligomerization reactions. Indeed, both
complexes produced mixtures with very narrow molecular weight
distributions (˛ = 0.25 for 49/MAO and ˛ = 0.20 for 50/MAO) indi-
cating high selectivities towards low molecular weight olefins
(Fig. 3). Additionally, their selectivities towards ␣-olefins are
decisively higher compared with the above described halogen sub-
stituted catalysts. The presence of two bulky tert-butyl groups per
iminophenyl ring in complex 50 protects the iron center moder-
ately from deactivation by ligand transfer reactions leading to a
good activity of 20,400 kg prod./mol Fe·h. On the other hand, its
lack of substituents at the ortho positions provides enough space
for the coordination of ethylene molecules.
3.3.2. Isomerization and in situ copolymerization of ˛-olefins
Complexes 28, 29, 33–41, 46, and 48–50 are the first examples of
bis(arylimino)pyridine iron complexes which, after activation with
MAO, isomerize and most likely copolymerize small ␣-olefins to a
greater extent indicating that the lack of ortho substituents at the
iminophenyl rings is an essential requirement for this behaviour
since all other complexes produced highly linear products with
selectivities to ␣-olefins greater than 90% (see Table 2). Both the iso-
merization of ␣-olefins and the copolymerization of ethylene with
small ␣-olefins (especially the gaseous 1-butene) can be verified
Compared with complexes 37–40, complexes 33–36 bearing
halogen substituents at the meta positions of the iminophenyl rings

C. Görl et al. / Applied Catalysis A: General 403 (2011) 25–35
35
afforded increased amounts of 1-hexene (55–80%) in the C6 frac-
tions (see Table 2) and, therefore, lower amounts of isomerized and
branched hexenes were detected.
[14] A.S. Ionkin, W.J. Marshall, D.J. Adelman, B. Bobik Fones, B.M. Fish, M.F.
Schiffhauer, Organometallics 25 (2006) 2978–2992.
[15] V.C. Gibson, N.J. Long, P.J. Oxford, A.J.P. White, D.J. Williams, Organometallics
25 (2006) 1932–1939.
As indicated above, the meta-tert-butyl substituted complexes
49 and 50 featured lower tendencies towards isomerization
compared with the complexes bearing electron withdrawing
substituents (33–36). Due to the bulky tert-butyl groups, the coor-
dination of higher ␣-olefins to the iron centers is hindered resulting
in lower amounts of isomerized and branched olefins. The selectiv-
ities towards 1-hexene (among the C6 fraction) were determined
to 70.9% (49) and 73.0% (50), respectively, representing the highest
selectivity values among the ortho-unsubstituted complexes.
[16] Y.V. Kissin, C. Qian, G. Xie, Y. Chen, J. Polym. Sci. A: Chem. 44 (2006) 6159–6170.
[17] F.A.R. Kaul, G.T. Puchta, G.D. Frey, E. Herdtweck, W.A. Herrmann,
Organometallics 26 (2007) 988–999.
[18] D. Takeuchi, R. Matsuura, Y. Fukuda, K. Osakada, Dalton Trans. 41 (2009)
8955–8962.
[19] A.S. Ionkin, W.J. Marshall, D.J. Adelman, B. Bobik Fones, B.M. Fish, M.F.
Schiffhauer, R.E. Spence, T. Xie, Organometallics 27 (2008) 1147–1156.
[20] D. Takeuchi, R. Matsuura, S. Park, K. Osakada, J. Am. Chem. Soc. 129 (2007)
7002–7003.
[21] K.P. Tellmann, V.C. Gibson, A.J.P. White, D.J. Williams, Organometallics 24
(2005) 280–286.
[22] M. Seitz, H.G. Alt, A. Aburaqabah, S.J. Palackal, EP 1716924 (2006).
[23] R. Schmidt, B.M. Welch, R.D. Knudsen, S. Gottfried, H.G. Alt, J. Mol. Catal. A:
Chem. 222 (2004) 17–25.
[24] M. Lopez Reyes, C. Martin Marcos, O. Prieto Acedo, J. Sancho Royo, J. Campora
Perez, P. Palma Ramirez, A.M. Naz Lucena, C.M. Perez Rodriguez, EP 2003166
(2008).
4. Conclusion
Bis(arylimino)pyridine iron(III) complexes containing electron
withdrawing substituents in their ligand frameworks were synthe-
sized and characterized. After activation with methylaluminoxane
(MAO), the resulting catalysts proved to be highly active in ethylene
oligomerization reactions. Both the size and the electronegativity of
the substituents as well as their positions at the iminophenyl rings
strongly influence the product compositions. The major difference
of the iron(III) complexes compared with analogous iron(II) com-
plexes is the fact, that catalytic activities are also observed when
the iminophenyl rings of the ligand frameworks do not contain
substituents in ortho-positions to the former amino groups. These
iron(III) complexes did not only produce ␣-olefins, but also internal
and branched olefins. An explanation for this behaviour consid-
ers the coordination and insertion of small ␣-olefins like 1-butene
into the growing chains. In other words, these complexes are the
first examples for bis(arylimino)pyridine iron catalysts which seem
to be able to copolymerize ethylene with higher ␣-olefins and to
isomerize terminal into internal olefins.
[25] A. Rodriguez-Delgado, J. Campora, A.M. Naz, P. Palma, M.L. Reyes, Chem. Com-
mun. (2008) 5230–5232.
[26] R. Schmidt, U. Hammon, M.B. Welch, H.G. Alt, B.E. Hauger, R.D. Knudsen, US
6,458,905 (2002).
[27] I.S. Paulino, U. Schuchardt, J. Mol. Catal. A: Chem. 211 (2004) 55–58.
[28] Z. Zhang, J. Zou, N. Cui, Y. Ke, Y. Hu, J. Mol. Catal. A: Chem. 219 (2004) 249–254.
[29] C. Goerl, H.G. Alt, J. Organomet. Chem. 692 (2007) 4580–4592.
[30] A.S. Abu-Surrah, K. Lappalainen, U. Piironen, P. Lehmus, T. Repo, M. Leskelä, J.
Organomet. Chem. 648 (2002) 55–61.
[31] R.-F. Chen, C.-T. Qian, J. Sun, Chin. J. Chem. 19 (2001) 866–869.
[32] A.S. Ionkin, W.J. Marshall, D.J. Adelman, B. Bobik Fones, B.M. Fish, M.F.
Schiffhauer, P.D. Soper, R.L. Waterland, R.E. Spence, T. Xie, J. Polym. Sci. A:
Polym. Chem. 46 (2008) 585–611.
[33] C. Goerl, H.G. Alt, J. Mol. Catal. A: Chem. 273 (2007) 118–132.
[34] A.S. Ionkin, WO 2007059015 (2007).
[35] D.D. Devore, S.S. Feng, K.A. Frazier, J.T. Patton, WO 2000069923 (2000).
[36] K.P. Bryliakov, N.V. Semikolenova, V.N. Zudin, V.A. Zakharov, E.P. Talsi, Catal.
Commun. 5 (2004) 45–48.
[37] F. Prades, R. Spitz, C. Boisson, S. Sirol, A. Razavi, WO 2007014889 (2007).
[38] D. Gong, B. Wang, C. Bai, J. Bi, F. Wang, W. Dong, X. Zhang, L. Jiang, Polymer 50
(2009) 6259–6264.
[39] P. Hao, Y. Chen, T. Xiao, W.-H. Sun, J. Organomet. Chem. 695 (2010) 90–95.
[40] E.P. Talsi, D.E. Babushkin, N.V. Semikolenova, V.N. Zudin, V.N. Panchenko, V.A.
Zakharov, Macromol. Chem. Phys. 202 (2001) 2046–2051.
[41] J. Martinez, V. Cruz, J. Ramos, S. Gutierrez-Oliva, J. Martinez-Salazar, A. Toro-
Labbe, J. Phys. Chem. C 112 (2008) 5023–5028.
Acknowledgements
[42] V. Cruz, J. Ramos, J. Martinez-Salazar, S. Gutierrez-Oliva, A. Toro-Labbe,
Organometallics 28 (2009) 5889–5895.
[43] R. Raucoules, T. de Bruin, P. Raybaud, C. Adamo, Organometallics 27 (2008)
3368–3377.
[44] I.E. Soshnikov, N.V. Semikolenova, A.N. Bushmelev, K.P. Bryliakov, O.Y. Lyakin,
C. Redshaw, V.A. Zakharov, E. Talsi, Organometallics 28 (2009) 6003–6013.
[45] B. Peifer, W. Milius, H.G. Alt, J. Organomet. Chem. 553 (1998) 205–220.
[46] J. Ramos, V. Cruz, A. Munoz-Escalona, J. Martinez-Salazar, Polymer 43 (2002)
3635–3645.
[47] N.I. Ivancheva, V.K. Badaev, I.I. Oleinik, S.S. Ivanchev, G.A. Tolstikov, Doklady
Phys. Chem. 374 (2000) 203–205.
[48] R. Souane, F. Isel, F. Peruch, P.J. Lutz, C. R. Chim. 5 (2002) 43–48.
[49] R. Schmidt, M.B. Welch, S.J. Palackal, H.G. Alt, J. Mol. Catal. A: Chem. 179 (2002)
155–173.
We thank Saudi Basic Industries Corporation (SABIC), Saudi Ara-
bia, for the financial support and Susanne Edinger (MC II, University
of Bayreuth) and Ruth Lohwasser (MC I, University of Bayreuth) for
recording the MALDI-TOF spectra.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.apcata.2011.06.010.
[50] C. Qian, F. Gao, Y. Chen, L. Gao, Synlett (2003) 1419–1422.
[51] G.-X. Jin, C. Liu, CN 1352204 (2002).
[52] C. Liu, G. Zhou, G.X. Jin, CN 1306012 (2001).
References
[53] R.F. Moreira, E.Y. Tshuva, S.J. Lippard, Inorg. Chem. 43 (2004) 4427–4434.
[54] R. Li, R.L. Smith, H.I. Kenttamaa, J. Am. Chem. Soc. 116 (1996) 5056–5061.
[55] V. Malatesta, K.U. Ingold, J. Am. Chem. Soc. 103 (1981) 609–614.
[56] G.J.P. Britovsek, S. Mastroianni, G.A. Solan, S.P.D. Baugh, C. Redshaw, V.C. Gibson,
A.J.P. White, D.J. Williams, M.R.J. Elsegood, Chem. Eur. J. (2000) 2221–2231.
[57] J.-Y. Liu, Y. Zheng, N. Hu, Y.-S. Li, Chin. J. Chem. 24 (2006) 1447–1451.
[58] M. Watanabe, H. Sato, M. Kuramoto, H. Tashiro, S. Tanaka, T. Tamura, WO
2001017967 (2001).
[1] 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) 849–850.
[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, J. Am. Chem. Soc. 121 (1999) 8728–8740.
[3] G.J.P. Britovsek, V.C. Gibson, D.F. Wass, Angew. Chem. Int. Ed. 38 (1999)
428–447.
[4] G.J.P. Britovsek, G.K.B. Clentsmith, V.C. Gibson, D.M.L. Goodgame, S.J. McTavish,
Q.A. Pankhurst, Catal. Commun. 3 (2002) 207–211.
[5] B.L. Small, M. Brookhart, J. Am. Chem. Soc. 120 (1998) 7143–7144.
[6] B.L. Small, M. Brookhart, A.M.A. Bennett, J. Am. Chem. Soc. 120 (1998)
4049–4050.
[7] B.L. Small, M. Brookhart, Macromolecules 12 (1999) 2120–2130.
[8] Y. Chen, C. Qian, J. Sun, Organometallics 22 (2003) 1231–1236.
[9] Y. Chen, R. Chen, C. Qian, X. Dong, J. Sun, Organometallics 22 (2003) 4312–4321.
[10] Z. Zhang, S. Chen, X. Zhang, H. Li, Y. Ke, Y. Lu, Y. Hu, J. Mol. Catal. A: Chem. 230
(2005) 1–8.
[11] F. Pelascini, F. Peruch, P.J. Lutz, M. Wesolek, J. Kress, Eur. Polym. Sci. 41 (2005)
1288–1295.
[12] K. Lappalainen, K. Yliheikkilä, A.S. Abu-Surrah, M. Polamo, M. Leskelä, T. Repo,
Z. Anorg. Allg. Chem. 631 (2005) 763–768.
[59] T. Zhang, W.H. Sun, T. Li, X.J. Yang, J. Mol. Catal A: Chem. 218 (2004) 119–124.
[60] G.V. Schulz, Z. Phys. Chem. B 30 (1935) 379–398.
[61] G.V. Schulz, Z. Phys. Chem. B 43 (1939) 25–46.
[62] P.J. Flory, J. Am. Chem. Soc. 62 (1940) 1561–1565.
[63] H. Kehlen, M.T. Rätzsch, Z. Phys. Chem. 265 (1984) 1049–1060.
[64] M. Seitz, Diploma thesis, University of Bayreuth, Germany (2001).
[65] S.A. Svejda, M. Brookhart, Organometallics 18 (1999) 65–74.
[66] S. Peitz, N. Peulecke, B.R. Aluri, S. Hansen, B.H. Mueller, A. Spannenberg, U.
Rosenthal, M.H. Al-Hazmi, F.M. Mosa, A. Woehl, W. Mueller, Eur. J. Inorg. Chem.
(2010) 1167–1171.
[67] M. Seitz, C. Goerl, W. Milius, H.G. Alt, Jordan J. Chem. 3 (2008) 109–145.
[68] B.L. Small, A.J. Marcucci, Organometallics 20 (2001) 5738–5744.
[69] C.M. Killian, L.K. Johnson, M. Brookhart, Organometallics 16 (1997)
2005–2007.
[13] A.S. Ionkin, W.J. Marshall, D.J. Adelman, A.L. Shoe, R.E. Spence, T. Xie, J. Polym.
Sci. A: Polym. Chem. 44 (2006) 2615–2635.