Iron(III) complexes with meta-substituted bis(arylimino)pyridine ligands: Catalyst precursors for the selective oligomerization of ethylene

Article abstract of DOI:10.1016/j.molcata.2011.10.011

Bis(arylimino)pyridine iron(III) complexes containing meta-halogen substituents at the iminophenyl rings were synthesized and characterized. 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 these iron(III) complexes. After activation with methylaluminoxane (MAO), these catalysts oligomerize ethylene to give also internal and branched olefins besides the expected linear α-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.

Full text of DOI:10.1016/j.molcata.2011.10.011

Journal of Molecular Catalysis A: Chemical 352 (2012) 110–127
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Iron(III) complexes with meta-substituted bis(arylimino)pyridine ligands:
Catalyst precursors for the selective oligomerization of ethylene
Christian Görl∗, Nadine Beck, Katharina Kleiber, 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:
Bis(arylimino)pyridine iron(III) complexes containing meta-halogen substituents at the iminophenyl
rings were synthesized and characterized. 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 these iron(III) com-
plexes. After activation with methylaluminoxane (MAO), these catalysts oligomerize ethylene to give also
internal and branched olefins besides 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.
Received 2 September 2011
Received in revised form 8 October 2011
Accepted 11 October 2011
Available online 18 October 2011
Keywords:
Bis(arylimino)pyridine iron complexes
Ethylene oligomerization
Oligomers
© 2011 Elsevier B.V. All rights reserved.
Isomerization of olefins
Branched alkenes
1. Introduction
and polymerization [14,24,29–38]. Although both iron(II) and
iron(III) complexes can be undoubtedly characterized, the oxi-
In 1998, 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. The
polymerization activities and the product compositions strongly
depended on the substitution pattern of the iminophenyl rings
[1–17]. Iron complexes bearing small substituents (alkyl or halo-
gen) 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 [1,5,8–10,14,18–29]. The
introduction of such electron withdrawing substituents into the
ligand backbones resulted in an enhanced temperature stability of
the catalysts and consequently lead to higher activities.
dation 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 [4,34,39]. Actual DFT calculations by Cruz
et al. [40,41] and Raucoules et al. [42] give strong evidence for the
enhanced oligomerization/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 [40,43].
The increased activities of the bis(arylimino)pyridine iron(III)
complexes can be explained with the stronger Lewis acidic
character of iron(III) centers compared with iron(II) centers and,
therefore, an enhanced affinity to coordinate electron rich olefin
molecules. During our investigations of bis(arylimino)pyridine
iron(II) and iron(III) oligomerization catalysts, we found that
iron(III) catalysts usually produced mixtures of shorter chain
length olefins compared to their iron(II) analogues bearing the
same bis(arylimino)pyridine ligand [31,44].
One of the characteristics of the longer known 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 triden-
While
a
great deal of work has been invested in
bis(arylimino)pyridine iron(II) complexes, only a few publica-
tions report about the corresponding iron(III) complexes despite
their higher catalytic activities towards olefin oligomerization
tate 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. [45], this complex probably does
∗
Corresponding author. Tel.: +49 921 55 2993; fax: +49 921 55 3206.
E-mail address: christian.goerl@uni-bayreuth.de (C. Görl).
1381-1169/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2011.10.011

C. Görl et al. / Journal of Molecular Catalysis A: Chemical 352 (2012) 110–127
111
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 also reported 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 [46], and 2,6-bis(1-
(4-nitrophenyl)ethyl)pyridine [13]. These ion–pair complexes
were catalytically completely inactive. Methyl groups seem to be
the smallest ortho-substituents that prevent this ligand transfer
reaction. In contrast to the data presented by Abu-Surrah et al. [45],
Bluhm et al. [47] disclosed a different synthetic route to the neu-
tral complex (2,6-bis(1-(phenylimino)ethyl)pyridine)FeCl2 using
FeCl2(THF)2 instead of the pure iron salt. The resulting complex
(as well as some other described complexes) may be stabilized by
the donor THF during the preparation but it showed a quite low
ethylene oligomerization activity which can be explained with a
fast ligand transfer reaction after activation with the cocatalyst
MAO.
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 [13,14,24,29,38]. However, due to the described ligand
exchange reaction, examples for 2,6-bis(arylimino)pyridine com-
plexes bearing only substituents at the meta- or para-positions
of the iminophenyl rings are still rare [38,47–52]. According to
Gong et al. [36], bis(arylimino)pyridine transition metal(III) com-
plexes without any substituents at the iminophenyl rings can be
applied for the polymerization of 1,3-butadiene. The lack of sub-
stituents therefore allows the incorporation of bigger molecules
than ethylene. The results of Gong et al. prompted us to investigate
the stabilities of 2,6-bis(arylimino)pyridine iron(III) complexes
without ortho-substituents at the iminophenyl rings [38]. We suc-
cessfully prepared a series of such iron(III) complexes which were
stable against ligand transfer reactions in contrast to the analo-
gous iron(II) complexes [38]. Employing such iron(III) catalysts in
ethylene oligomerization reactions, product mixtures with narrow
molecular weight distributions were obtained [38]. However, the
catalysts’ thermal stabilities were lower compared with catalysts
bearing ortho-substituents at the iminophenyl rings. Since halogen
substituents at meta- and para-positions were found to increase
both the catalytic activities and the thermal stabilities [13,14,24,29]
we focused on the synthesis of bis(arylimino)pyridine iron(III)
complexes containing meta-halogen substituted iminophenyl rings
and the investigation of their abilities towards selective ethylene
oligomerization.
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. For the analysis of oligomer mixtures, GC spectra
were obtained with an Agilent 6890 N 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. Ele-
mental analyses were performed with a VarioEl III CHN instrument
using acetanilide for calibration.
2.2. Synthesis of 3,5-dihalogen-nitrobenzenes
2,6-Dibromo-4-nitroaniline
or
2,6-diiodo-4-nitroaniline
(15 mmol) were dissolved in ethanol (100 ml) and cooled to
0 ◦C in an ice bath. Concentrated sulfuric acid (8 ml; 0.15 mol)
was added dropwise over 30–45 min with constant stirring. The
reaction mixture was heated to 60 ◦C and sodium nitrite (3.11 g;
45 mmol) was added to the reaction mixture in small portions.
The resulting yellow colored reaction mixture was heated slowly
to 90 ◦C and refluxed for 3 h whereby the color changed to brown.
After cooling to room temperature, the mixture was poured into
ice water. The resulting reddish brown solid was filtered off,
washed with water and dried yielding the desired products as
brown solid (1: 86%) and yellow solid (2: 91%), respectively.
2.3. Reduction of 3,5-dihalogen-nitrobenzenes to
3,5-dihalogenanilines
Compounds 1 or 2 (13 mmol) were dissolved in ethanol (30 ml)
and SnCl2·2H2O (11.3 g; 50 mmol) was added portion wise at room
temperature. The reaction mixtures were heated under reflux at
80 ◦C for 1.5 h (in case of 1), respectively, 8 h (in case of 2). After
cooling to room temperature, the solvents were evaporated under
reduced pressure and the crude solids were basified with 4N NaOH
to pH 12. The mixtures were extracted with ethyl acetate (3× 30 ml)
and the combined organic layers were washed with brine and dried
over sodium sulfate. The solvents were removed in vacuo providing
the desired compounds as a brown solid (3; 86%) and a yellow solid
(4; 89%).
2. Experimental
2.1. General considerations
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 and 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 was dried over magnesium turnings. Drying of ethanol
was accomplished with sodium. 1-Butanol (p.a.) was purchased
from Merck and used without prior distillation. Methylalumi-
noxane (MAO) was purchased from Crompton (Bergkamen; 10%
in toluene). Ethylene (3.0) and argon (4.8/5.0) were supplied
by Rießner Company (Lichtenfels). All other starting materi-
als were commercially available and were used without further
purification.
2.4. Synthesis of 2,6-bis(arylimino)pyridine compounds
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 cor-
responding aniline compound (7 mmol), and the silica/alumina
catalyst (0.5 g). The reaction mixture was heated to 40–45 ◦C for
24 h. If the reactions were not completed (according to GC/MS
analyses), the heating period was prolonged till completion. After
cooling to room temperature, the mixture was filtered over sodium

112
C. Görl et al. / Journal of Molecular Catalysis A: Chemical 352 (2012) 110–127
Table 1
1H NMR, 13C NMR, and MS data of compounds 1–16.
Compound 1H NMR ı [ppm]
13C NMR ı [ppm]
MS [m/z]
•
+
1
2
8.29 s (2H, Ar-H), 7.96 s (1H, Ar-H)
8.49 s (2H, Ar-H), 7.23 s (1H, Ar-H)
149.0 (Cq) 140.0 (Ar-CH) 125.6 (2C, Cq), 123.4 (2C, Ar-CH)
279/281/283 M (27/50/24)
235 M-NO2 (51)200 M-Br (3)
156 M-NO2-Br (31)
•
+
152.0 (Cq) 135.3 (Ar-CH), 131.8 (2C, Ar-CH), 94.2 (2C, Cq, C-I)
375 M (80)
329 M-NO2 (24)
202 M-NO2-I (22)
•
+
3
4
5
6.98 s (1H, Ar-H), 6.71 s (2H, Ar-H), 3.74 s (br, 2H, NH2) 148.6 (Cq), 123.6 (2C, Ar-CH), 123.3 (2C, Cq), 116.5 (Ar-CH)
7.16 s (1H, Ar-H), 6.91 s (2H, Ar-H), 4.71 s (br; 2H, NH2) 149.3 (Cq), 132.4 (Ar-CH), 121.8 (2C, Ar-CH), 94.4 (2C, Cq, C-I)
249/251/253 M (59/100/49)
170 M-Br (33)
•
+
345 M (100)
218 M-I (47)
•
+
8.34 d (2H, Ar-H), 7.86 t (1H, Ar-H), 7.40–7.34 m (4H,
167.3 (Cq, C N), 155.4, 151.3 (Cq), 136.8 (Py-CH), 129.0, 123.6
313 M (8)
Ar-H), 7.15–7.08 m (2H, Ar-H), 6.87–6.83 m (4H, Ar-H), (Ar-CH), 122.3 (Py-CH), 119.3 (Ar-CH), 16.2 (CH3)
2.38 s (6H, CH3)
298 M-Me (2)
118 Ph-N C-CH3 (54)
77 Phenyl (100)
•
+
6
8.33 d (2H, Ar-H), 7.85 t (1H, Ar-H), 7.31 dd (2H, Ar-H), 167.9 (Cq, C N), 163.3 d (Cq, 1JCF = 246.1 Hz), 155.6 (Cq), 153.0 d 349 M (100)
6.83–6.77 m (2H, Ar-H), 6.65–6.54 m (4H, Ar-H), 2.41 s (Cq, 3
(6H, CH3)
123.0 (Py-CH), 115.3 d (Ar-CH, 4
2JCF = 21.3 Hz), 107.0 d (Ar-CH, 2JCF = 23.1 Hz), 16.2 (CH3)
8.31 d (2H, Ar-H), 7.85 t (1H, Ar-H), 7.28 td (2H, Ar-H), 168.1 (Cq, C N), 155.1, 152.4, 134.7 (Cq), 137.0 (Py-CH), 130.1,
J
CF = 9.2 Hz), 137.4 (Py-CH), 130.2 d (Ar-CH, 3JCF = 9.3 Hz),
239 M-[FC6H4N] (17)
136 FC6H4N C-CH3 (100)
J
CF = 2.6 Hz), 110.7 d (Ar-CH,
•
+
7
8
381 M (46)
7.08 dd (2H, Ar-H), 6.84 d (2H, Ar-H), 6.71 dd (2H,
Ar-H), 2.39 s (6H, CH3)
123.7 (Ar-CH), 122.6 (Py-CH), 119.4, 117.5 (Ar-CH), 16.4 (CH3)
254 M-[ClC6H4NH2] (32)
152 ClC6H4N C-CH3 (100)
•
+
8.29 d (2H, Ar-H), 7.86 t (1H, Ar-H), 7.19–7.22 m (2H,
168.1 (Cq, C N), 155.1, 152.6, 122.7 (Cq), 136.9 (Py-CH), 130.4,
471 M (100)
Ar-H), 7.00 d (2H, Ar-H), 6.74–6.76 m (2H, Ar-H), 2.39 s 126.5 (Ar-CH)
(6H, CH3) 122.6 (Py-CH), 122.2, 117.9 (Ar-CH), 16.4 (CH3)
390 M-Br (11)
298 M-[BrC6H4NH2] (55)
196 BrC6H4N C-CH3 (96)
•
+
9
8.30 d (2H, Ar-H), 7.85 t (1H, Ar-H), 7.45 dd (2H, Ar-H), 168.0 (Cq, C N), 155.0, 152.5 (Cq), 94.4 (Cq, C-I), 136.9 (Py-CH), 565 M (100)
7.01–7.22 m (4H, Ar-H), 6.81 dd (2H, Ar-H), 2.39 s (6H, 132.5, 130.5, 127.9 (Ar-CH), 122.6 (Py-CH), 118.6 (Ar-CH), 16.4
438 M-I (9)
CH3)
(CH3)
346 M-[IC6H4N] (44)
244 IC6H4N C-CH3 (61)
•
+
10
8.34 d (2H), 7.86 t (1H), 7.33–7.26 m (2H, Ar-H),
7.16–7.12 m (2H, Ar-H), 6.88–6.86 m (2H, Ar-H),
6.67–6.64 m (2H, Ar-H), 2.41 s (6H, CH3), 1.33 s (18H,
t-Bu)
8.34 d (2H, Ar-H), 7.92 t (1H; Ar-H), 6.57–6.62 m (2H,
Ar-H), 6.39–6.42 m (4H, Ar-H), 2.44 s (6H, CH3)
167.2 (Cq, C N), 155.6, 152.2, 150.9, 34.8 (Cq), 136.8 (Py-CH),
128.6 (Ar-CH), 122.5 (Py-CH), 120.8, 116.6, 116.3 (Ar-CH), 31.3
(tBu-CH3), 16.2 (CH3)
425 M (100)
410 M-Me (17)
368 M-tBu (14)
•
+
11
12
168.5 (Cq, C N), 163.5 dd (Cq, 1JCF = 246.5 Hz), 154.8 (Cq), 153.8 385 M (13)
t (Cq, 3JCF = 9.2 Hz), 137.1, 122.9 (Py-CH), 102.4 dd (Cq, 2JCF = 22.8 366 M-F (7)
Hz), 98.7 t (Ar-CH, 2JCF = 25.5 Hz), 16.4 (CH3)
154 (F2C6H3)N C-CH3 (100)
•
+
8.27 d (2H, Ar-H), 7.86 t (1H; Ar-H), 7.09 t (2H, Ar-H),
6.72 d (4H, Ar-H), 2.39 s (6H, CH3)
168.7 (Cq, C N), 154.7, 153.1 (Cq), 137.1 (Py-CH), 135.3 (Cq),
123.5, 122.9 (Ar-CH + Py-CH), 117.8 (Ar-CH), 16.5 (CH3)
451 M (7)
413 M-HCl (22)
265 M-CH3-C N(C6H3Cl2) (30)
186 (Cl2C6H3)N C-CH3 (100)
13
14
8.23 d (2H, Ar-H), 7.82 t (1H; Ar-H), 7.35 s (2H, Ar-H),
6.88 s (4H, Ar-H), 2.35 s (6H, CH3)
8.25 d (2H, Ar-H), 7.84 t (1H; Ar-H), 7.77 s (2H, Ar-H),
7.15 s (4H, Ar-H), 2.38 s (6H, CH3)
168.8 (Cq, C N), 154.7, 153.4 (Cq), 137.1 (Py-CH), 128.9 (Ar-CH), 629 M·+ (100)a
123.1 (Cq), 122.9 (Py-CH), 121.1 (Ar-CH), 16.6 (CH3)
•
+
168.8 (Cq, C N), 154.7, 153.3 (Cq), 139.9 (Ar-CH), 137.1 (Py-CH), 817 M (100)a
127.4 (Ar-CH), 122.8 (Py-CH), 94.8 (Cq; C-I), 16.6 (CH3)
370 H3C-C N-[C6H3I2] (44)
•
+
15
8.31 d (2H, Ar-H), 7.84 t (1H; Ar-H), 6.76 dd (2H, Ar-H), 167.0 (Cq, C N), 155.5, 151.3, 138.6 (Cq), 136.7 (Py-CH), 125.2
369 M (62)
6.48 s (4H, Ar-H), 2.41 s (6H, CH3), 2.33 s (12H;
(Ar-CH), 122.1 (Py-CH), 116.8 (Ar-CH), 21.4 (Ar-CH3), 16.2 (CH3) 354 M-CH3 (42)
tBu-CH3)
146 H3C-C N-[C6H3(CH3)2] (100)
•
+
16
8.36 d (2H), 7.88 t (1H), 7.17 s (br, 2H, Ar-H4), 6.71 s
(br, 4H, Ar-H2/6), 2.44 s (6H, CH3), 1.33 s (36H, t-Bu)
167.0, 155.7, 151.5, 150.5, 34.9 (Cq), 136.8, 122.2, 117.5, 113.8
(CH), 31.5 (tBu-CH3), 16.2 (CH3)
537 M (100)
522 M-Me (24)
480 M-tBu (57)
a
Molecular weight too high for GC/MS analysis, thus EI-MS spectra were recorded.
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: 45–85%.
The analytical data for compounds 1–16 are given in
Table 1.
three hours at room temperature, whereby the complexes precipi-
tated. 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 analytical data for
complexes 17–28 are given in Table 2.
2.5. General synthesis of the 2,6-bis(arylimino)pyridine iron(III)
2.6. Oligomerization of ethylene in a 1 l Büchi autoclave
complexes 17–28
The desired iron complex (1.5–9.2 ␮mol) was suspended in
toluene (5 ml). Methylaluminoxane (10% in toluene) was added
to maintain a ratio Fe:Al = 1:2500 resulting in an immediate color
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 labo-
ratory autoclave under inert atmosphere and thermostated at 60 ◦C.
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 immediate
color change to brown, orange, or red. The mixture was stirred for

C. Görl et al. / Journal of Molecular Catalysis A: Chemical 352 (2012) 110–127
113
Table 2
MS and elemental analyses data of the iron complexes 17–28.
Complex
EI-MS [m/z] (rel. intensity
MALDI-TOF [m/z] (rel.
Elemental analyses
to base peak: [%])
intensity to base peak: [%])
Cexp [%]
52.88
Ctheor [%]
Hexp [%]
4.25
Htheor [%]
4.03
Nexp [%]
8.76
Ntheor [%]
•
•
+
+
17
474/476/478M (−)
476/478 M (27)
439/441 M-Cl (95)
404 M-2Cl (76)
53.03
49.30
8.84
404 M-2Cl (3)
348 M-Fe-2Cl (10)
313 M-Fe-3Cl (100)
314 [M + H]-Fe-3Cl (36)
•
•
+
+
18
19
510/512 M (−)
512/514 M (20)
475/477 M-Cl (80)
440 M-2Cl (100)
49.54
46.01
3.53
3.32
3.35
3.15
8.02
7.54
8.21
475 M-Cl (6)
440 M-2Cl (14)
383 M-Fe-2Cl (8)
349 M-Fe-3Cl (100)
542/544/546 M (−)
405 M-3Cl (44)
350 [M + H]-Fe-3Cl (72)
542/544/546 M (11)
•
•
+
+
46.32
7.72
509/511 M-Cl (5)
474/476 M-2Cl (3)
381 M-Fe-3Cl (47)
509/511 M-Cl (100)
474/476 M-2Cl (99)
381 [M + H]-Fe-3Cl (83)
•
+
20
21
632/634M (−)
632/634/636 [M+ (2)
595/597M-Cl (100)
561 M-2Cl (44)
40.07
34.71
39.82
34.68
2.83
2.42
2.71
2.36
6.46
5.60
6.63
5.78
505 M-Fe-2Cl (4)
471 M-Fe-3Cl (100)
•
+
726/728M (−)
727/729/731
692 M-Cl (3)
[M• + + 1](16)
600 M-I (5)
598 M-Fe-2Cl (6)
565 M-Fe-3Cl (28)
691/693 M-Cl (100)
656/658 M-2Cl (89)
566 [M + 1]-Fe-3Cl (29)
•
•
+
+
22
586/588 M (1)
549 M-HCl (4)
586/588/590 M (3)
571/573 M-Me (7)
570/572 M-CH4 (24)
551/553 M-Cl (48)
516 M-2Cl (100)
59.53
59.26
5.89
6.00
7.14
7.15
537 M-Cl-Me (23)
481 M-3Cl (6)
425 M-Fe-3Cl (20)
•
+
23
24
546/548 M (−)
385 ligand (78)
154 F2(C6H3)N C-CH3
(100)
547/549 [M + 1] (1)
511 M-Cl (100)
476 M-2Cl (62)
384 M-Fe-3Cl (2)
45.73
40.88
46.06
41.12
2.87
2.71
2.76
2.46
7.49
6.63
7.67
6.85
•
+
610/612 M (−)
613/611 [M + 1] (8/13)
576 M-Cl (100)
541 M-2Cl (1)
451 ligand (50)
416 ligand-Cl (33)
541 M-2Cl (61)
506 M-3Cl (30)
452 [M + 1]-Fe-3Cl (33)
451 M-Fe-3Cl (14)
•
•
+
+
25
26
27
790/792 M (−)
629 ligand (10)
547 ligand-Br (6)
794/792 M (6)
31.57
25.72
51.95
31.88
25.76
52.19
2.02
1.72
5.18
1.91
1.54
5.12
5.18
4.28
6.79
5.31
4.29
6.90
757/755 M-Cl (95/100)
722 M-2Cl (22)
628 M-Fe-3Cl-H (6)
•
•
+
+
978/980 M (−)
978/980 M (2)
817 ligand (2)
945/943 M-Cl (78/100)
908/910 M-2Cl (9/35)
816 M-Fe-3Cl (3)
343 [I2C6H3N] (82)
•
•
+
+
530/532 M (−)
530/532 M (2)
495 M-Cl (33)
495 M-Cl (3)
369 ligand (100)
461 M-Cl-HCl (100)
426 M-3Cl (34)
368 M-Fe-3Cl-H (31)
•
•
+
+
28
698/700 M (1)
665 M-Cl (19)
698/700/702 M (30)
665 M-Cl (80)
63.78
63.48
7.37
7.34
5.89
6.00
662 M-HCl (38)
628 M-2Cl (3)
629 M-Cl-HCl (100)
595 M-3Cl (14)
595 M-3Cl (2)
537 M-Fe-3Cl (100)
480 M-t-Bu-FeCl3 (77)
An ethylene pressure of 10 bar was applied and the mixture was
stirred for one hour, then the ethylene flow was stopped. After cool-
ing down to −15 ◦C, the ethylene pressure was carefully released
(with loss of traces of butenes and no loss of oligomers ≥ C6). A
sample (1 ml) was taken from the cooled mixture by syringe and fil-
tered over a small silica column which was externally cooled with
an isopropanol/dry-ice bath (−20 ◦C). From the filtered solution,
a volume of 2 ␮l was applied for the GC analysis. The relative
ratios of the oligomer fractions (C4, C6, C8) could be taken from the
gas chromatogram. The GC temperature program was optimized
especially to display the separation of the C6 isomers. For a quan-
titative analysis of the C6 fraction, diluted hydrochloric acid (20 ml
of a 1 M HCl) was slowly added to the whole oligomer mixture.
The resulting biphasic system was then allowed to come to room

114
C. Görl et al. / Journal of Molecular Catalysis A: Chemical 352 (2012) 110–127
temperature whereby most of the butenes were evolved. The
oligomer solution was separated and dried over sodium sulfate. The
residual butenes and the solvent n-pentane were carefully distilled
off using a Vigreux column (Texternal bath = 40 ◦C). The residue (con-
taining the C6 and higher oligomers and a small amount of toluene
from the MAO) was heated slowly, and the C6 fraction was collected
(at Texternal bath = 65–70 ◦C) and weighed after cooling down to room
temperature. From the corresponding GC integral of the C6 fraction
(thus serving as an internal standard), the amounts of the C4 and
C8 fractions were then calculated. Due to the similar boiling points
of toluene (111 ◦C) and the octene isomers (∼110–125 ◦C), a sepa-
ration of the C8 fraction by distillation could not be accomplished.
was applied using a heterogeneous silica/alumina catalyst (13%
Al2O3/87% SiO2) and molecular sieves 3 A as the water absorbing
˚
agent. Theheterogeneouscatalystactsasa Lewisacidand, addition-
ally, 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. Table 3 gives an overview of
the prepared 2,6-bis(arylimino)pyridine compounds, while the 1
H
NMR, 13C NMR, and MS data for compounds 5–16 can be found in
Table 1 (see Section 2).
3.3. Synthesis of the iron(III) complexes
The bis(arylimino)pyridine compounds 5–16 were reacted
with anhydrous ferric chloride either in n-butanol or a 1:1
(v:v) mixture of diethylether and THF to give the corresponding
bis(arylimino)pyridine iron(III) complexes 17–28 (Scheme 3 and
Table 3) which could be isolated as orange or brown solids.
3. Results and discussion
3.1. Synthesis of meta-dihalogen substituted anilines
While many meta-halogen substituted aniline derivatives
are readily commercially available, 3,5-dibromoaniline and 3,5-
As
described
in
the
introduction
chapter,
many
bis(arylimino)pyridine compounds without or with small
substituents at the ortho-positions of the iminophenyl rings
immediately form deep purple or violet ion–pair complexes
[L2Fe]2+ [FeCl4]2− in reactions with iron(II) chloride. However, in
some cases, the iron(II) complexes seem to exist as (L)FeCl2 type
complexes when bearing 4-iodo [29], 4-allyl [56], 3-vinyl [52],
3- or 4-trifluoromethyl [47], or 3- or 4-methoxy [47] substituted
iminophenyl rings.
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 [38]. However, the
resulting complexes are not stable towards ligand transfer reac-
tions when they are kept in polar solvents (like THF or alcohols, see
Refs. [14] and [34]) for some hours. Additionally, at longer reac-
tion 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 [57–59]. Therefore, the time limit for the com-
plexation reactions and the subsequent isolation procedures was
determined to about three hours. 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 the 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 were 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 are given
in Table 2.
Scheme 4 shows the MALDI-TOF mass spectrum of complex
21. Around m/z = 728, the molecular ion mass distribution was
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 corresponding
to the protonated ligand [Mligand + 1] was detected.
diiodoaniline had to be prepared in
a two-step synthesis.
Diazotation of the amino groups of 2,6-dibromo-4-nitroaniline
and 2,6-diiodo-4-nitroaniline was accomplished using an excess of
sodium nitrite and sulfuric acid (96 wt.%). Subsequent removal of
the diazo function by heating the reaction mixture for three hours
at 90 ◦C lead to the 3,5-dihalogen-nitrobenzenes 1 and 2 (Scheme 1)
[53].
Reduction of compounds 1 and 2 was achieved applying
SnCl2·2H2O as the reducing agent obtaining the desired 3,5-
dihalogenanilines 3 and 4 in high yields (Scheme 1) [53].
3.2. Preparation of substituted 2,6-bis(arylimino)pyridine
compounds
Condensation reactions of 2,6-diacetylpyridine with meta-
halogen substituted anilines yielded the 2,6-bis(arylimino)pyridine
compounds (Scheme 2). For comparison purposes, some ligand
precursors were prepared using meta-alkyl substituted anilines.
Condensation reactions using 3-methylaniline (m-toluidine) as
the amine compound surprisingly did not lead to the desired
bis(arylimino)pyridine compound but gave condensation products
of the aniline. 2,6-Bis(arylimino)pyridine compounds containing
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, the iodo substituted ligand
precursors appeared to be light sensitive. Former literature usu-
ally reported the use of para-toluenesulfonic acid as a catalyst
for the condensation reactions of 2,6-diacetylpyridine with ani-
line derivatives [2,5,54]. 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 tempera-
tures. As a result, some temperature sensitive anilines decomposed
under these quite harsh reaction conditions and did not provide
the desired bis(arylimino)pyridine compounds. To overcome these
problems, an elegant method introduced by Qian et al. [8,55]
While for complex 21 the molecular ion in the MALDI-TOF mass
spectrum was clearly visible, the intensities of the molecular ions of
2,6-bis(arylimino)pyridine iron complexes bearing dihalogenated
iminophenyl rings were distinctively lower but the ways of decom-
position are identical. Scheme 5 shows the upper mass region of
Scheme 1. Synthesis of 3,5-dibromo- and 3,5-diiodoaniline.

C. Görl et al. / Journal of Molecular Catalysis A: Chemical 352 (2012) 110–127
115
Scheme 2. General synthesis of 2,6-bis(arylimino)pyridine compounds 5–16.
Scheme 3. Synthesis of the bis(arylimino)pyridine iron complexes 17–28.
Table 3
hour employing an ethylene pressure of 10 bar. As a solvent
n-pentane was used. Due to the big differences in their activ-
ities, the complexes were used in appropriate molar amounts
to keep the exothermic reaction under control. Depending on
the size and the nature of the substituents at the iminophenyl
rings, a wide range of olefinic products were obtained from
the activated complexes 17–28. Their oligomerization results are
given in Table 4 while Table 5 provides a detailed comparison
of the most relevant oligomerization properties of ortho-, meta-,
di-meta-, and para-halogen substituted bis(arylimino)pyridine
iron(III) catalysts including some results of our previous study
[38].
In all experiments, the product mixtures mainly consisted
of low molecular weight ␣-olefins and, to a lower extent, of
internal and branched olefins. No solid products (=polyethylenes)
were obtained. The oligomer mixtures were characterized by gas
chromatography and GC/MS correlations using the NIST mass
spectral database. The C6 fractions were isolated and weighed
(thus serving as an internal standard), and the amounts of
butenes and octenes were then calculated from the corresponding
Synthesized 2,6-bis(arylimino)pyridine compounds and their corresponding
iron(III) chloride complexes (yields given in brackets).
Bis(imino) compound
R1
R2
Iron(III) complex
5 (77%)
6 (56%)
7 (74%)
8 (73%)
9 (54%)
10 (69%)
11 (55%)
12 (54%)
13 (55%)
14 (47%)
15 (68%)
16 (85%)
H
F
H
H
H
H
H
H
F
Cl
Br
I
Me
t-Bu
17 (94%)
18 (95%)
19 (83%)
20 (74%)
21 (65%)
22 (49%)
23 (69%)
24 (59%)
25 (79%)
26 (65%)
27 (82%)
28 (68%)
Cl
Br
I
t-Bu
F
Cl
Br
I
Me
t-Bu
the MALDI-TOF mass spectrum of complex 26 (diiodo substituted
iminophenyl rings) including a closer look at the region m/z = 900
up to m/z = 1000 while Scheme 6 shows the isotope distributions
of the molecular ion around m/z = 979 and of the base peak (loss of
HCl or Cl) around m/z = 943.
Similarly to complexes 21 and 26, MALDI-TOF mass spectra for
all other complexes revealed that the base peaks result from the
loss of either one or two chloro ligands (see Table 2).
Slow diffusion of toluene into a dichloromethane solution of
complex 21 afforded thin orange brown colored needles. 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) complexes [14,30,36],
the R value was too high (corresponding to an increased uncertainty
for the geometry of the molecule). Finally, the refinement proce-
dure could not be finished successfully.
3.4. Results of the catalytic oligomerization of ethylene
3.4.1. Oligomerization activities and product molecular weight
distributions
The iron complexes 17–28 were applied as catalyst precursors
for the homogeneous oligomerization of ethylene. The complexes
were activated with methylaluminoxane (MAO) applying a ratio
Fe:Al = 1:2500. Oligomerization runs were routinely performed
in a 1 l Büchi steel reactor at a temperature of 60 ◦C over one
Scheme 4. MALDI-TOF mass spectrum of complex 21 with clearly visible molecular
ion around m/z = 729 (ions with m/z < 500 were omitted for clarity).

116
C. Görl et al. / Journal of Molecular Catalysis A: Chemical 352 (2012) 110–127
Scheme 5. MALDI-TOF mass spectrum of complex 26 with less intensive molecular ion around m/z = 979 (ions with m/z < 700 were omitted for clarity). Offset: enlarged view
of the region m/z = 900 up to m/z = 1000.
Scheme 6. Enlarged sections of the MALDI-TOF mass spectrum of complex 26: (left) isotope distribution of the molecular ion M+ around m/z = 979 and (right) ion distribution
around m/z = 943 for the loss of HCl or Cl.
GC integrals. The catalysts’ activities were calculated both from
the consumed amount of ethylene (named as “theoretical yield”
in Table 4) and from the obtained amount of C6 oligomers
(given as “oligomer yield” in Table 4). Remarkably, the activities
determined from weighing the C6 fractions are only marginally
lower (1–5%) compared with the values calculated from the
ethylene consumptions which indicates that the workup pro-
cedure (see Section 2) sufficiently retains the butenes in the
mixtures).
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
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 [60–66]. It
can be calculated from both the reaction rate constants and the GC
kpropagation and ktermination describe the rate constants for chain
growth and chain termination. Cn and Cn+2 rely to olefins with n
or (n + 2) carbon atoms giving ˛ as the average value of at least two
subsequent propagation steps. In the present study, the C6 → C8

Table 4
Ethylene oligomerization results for the iron complexes 17–28 (solvent: 250 ml n-pentane, activator: MAO, Fe:Al = 1:2500, T = 60 ◦C, 10 bar ethylene, 1 h).
Complex
nCat
[␮mol]
Ethylene
consumption
[l]
Theoretical
yield [g]
Oligomer
yield [g]a
Activity [kg
prod./mol
Fe h]
Cb
M c
[g/mol]
Oligomer
yield [mol]
Schulz–Flory
coefficient
␣
C4 olefins in
the product
mixture
1-butene in the
[%]
[g]a
[mol]a
product
mixture [%]
C4 fraction
[%]
17
18
19
20
21
22
23
24
25
26
27
28
9.2
6.1
2.0
1.6
4.1
4.7
3.7
2.3
1.5
2.3
2.4
3.1
32.2
44.4
87.7
29.0
65.4
31.1
18.8
39.4
56.0
20.4
8.8
38.7
53.3
105.3
34.8
78.5
38.7
22.9
47.3
67.6
24.6
10.5
64.3
38.5
52.9
104.2
34.4
78.0
37.1
22.5
46.5
67.0
24.4
10.2
64.0
4150
7320
52050
22000
19070
7900
6.03
6.14
5.59
4.99
6.15
5.57
5.09
5.69
5.25
5.96
5.33
5.50
84.5
87.7
78.4
70.0
86.3
78.2
71.4
79.8
73.7
83.6
74.8
77.2
0.46
0.61
1.33
0.49
0.76
0.48
0.32
0.59
0.91
0.29
0.14
0.83
0.28
0.36
0.29
0.27
0.36
0.23
0.28
0.35
0.33
0.35
0.23
0.20
29.3
35.7
41.0
60.9
34.5
41.5
57.9
42.9
53.1
48.4
46.6
41.1
7.5
12.1
30.6
16.9
14.7
11.0
10.2
14.3
27.1
7.9
0.13
0.22
0.55
0.30
0.26
0.20
0.18
0.25
0.48
0.14
0.06
0.34
29.0
35.1
40.5
60.6
34.1
41.2
57.5
42.6
52.7
48.2
46.4
41.1
98.9
98.3
98.8
99.6
98.9
99.3
99.3
99.5
99.3
99.7
99.6
>99.9
6025
20650
44480
10420
4270
3.5
19.1
53.6
20400
Complex
C6 olefins in the
product mixture
1-hexene in the
C8 olefins in the
product mixture
1-octene in the
C10 olefins in the product mixture [%]
Higher oligomers [%]
[%]
[g]d
[mol]d
product mixture [%]
C6 fraction [%]
[%]
[g]a
[mol]a
product mixture [%]
C8 fraction [%]
17
18
19
20
21
22
23
24
25
26
27
28
49.5
38.9
39.0
29.8
41.1
45.1
30.8
35.6
29.4
33.2
41.2
45.9
19.0
19.8
43.6
12.4
26.2
18.0
8.2
17.8
22.5
8.1
0.23
0.24
0.52
0.15
0.31
0.22
0.10
0.21
0.27
0.09
0.05
0.38
24.4
21.9
26.4
24.3
22.8
29.4
25.5
27.3
22.4
27.4
32.1
32.1
49.3
51.9
68.6
80.8
55.4
64.8
82.8
76.7
76.4
82.6
77.8
70
14.7
13.4
10
7.5
9.1
14.9
3.2
11.1
5.6
2.7
8.5
8.9
3.5
1.3
0.07
0.08
0.13
0.03
0.1
0.05
0.02
0.07
0.08
0.03
0.01
0.09
9.4
9.5
8.2
5.5
10
9.9
7.1
11.7
7.9
10.4
8.7
55.2
71
82
96.2
76.6
82.9
93.9
91.8
90.9
97.9
4.4
5.7
3.2
2
5.5
2.3
2.4
4.4
3.3
4
2.3
6.3
6.8
1.5
5.8
0.6
1.3
7.3
5.5
3.7
1.2
–
5.8
13.1
10.5
7.6
12.8
8.7
10.7
8.7
11.4
4.6
32.0
100
91.2
2.3
2.6
10.6
10.4
a
Yields for the C4 and C8 fractions calculated using the C6 fraction as “internal standard”.
b
c
Average carbon number of the oligomer mixture (in analogy to the average polymerization grade defined by Henrici-Olive and Olive in Ref. [66]) calculated from the GC integrals.
Average molecular weight of the oligomer mixture (M = C × 14.027 since the olefins can be expressed as CnH2n).
Yields for the C6 fractions were weighed.
d

118
C. Görl et al. / Journal of Molecular Catalysis A: Chemical 352 (2012) 110–127
Table 5
Comparison of Schulz–Flory values, the overall contents of C6 olefins, and the amount of 1-hexene within the C6 fractions of ortho-, meta, di-meta-, and para-halogen or -alkyl
substituted 2,6-bis(arylimino)pyridine iron(III) complexes (compounds noted in brackets could not be obtained, therefore no values can be given).
Substituent at the iminophenyl ring in position
ortho
meta
di-meta
para
2-F
2-Cl
2-Br
[2-I]
17 (2-H)
2-Me
0.71
0.91
0.97
–
0.27
0.78
18 (3-F)
19 (3-Cl)
20 (3-Br)
21 (3-I)
0.36
0.29
0.27
0.36
0.23
–
23 (3,5-di-F)
24 (3,5-di-Cl)
25 (3,5-di-Br)
26 (3,5-di-I)
0.28
0.35
0.33
0.35
0.20
0.23
4-F
0.43
0.29
0.28
0.38
4-Cl
4-Br
4-I
Schulz–Flory
coefficient ˛
22 (3-tBu)
[3-Me]
28 (3,5-di-tBu)
27 (3,5-di-Me)
2-F
2-Cl
2-Br
[2-I]
17 (2-H)
2-Me
32.4
18 (3-F)
19 (3-Cl)
20 (3-Br)
21 (3-I)
38.9
23 (3,5-di-F)
24 (3,5-di-Cl)
25 (3,5-di-Br)
26 (3,5-di-I)
30.8
4-F
41.7
16.2
9.3
–
49.5
4.4
39.0
29.8
41.1
45.1
–
35.6
29.4
33.2
41.2
45.9
4-Cl
4-Br
4-I
54.0
51.5
46.0
C6 olefins in the
product mixture [%]
22 (3-tBu)
[3-Me]
28 (3,5-di-tBu)
27 (3,5-di-Me)
2-F
2-Cl
2-Br
[2-I]
17 (2-H)
2-Me
46.8
>99
>99
–
49.3
93.7
18 (3-F)
19 (3-Cl)
20 (3-Br)
21 (3-I)
51.9
68.6
80.8
55.4
64.8
–
23 (3,5-di-F)
24 (3,5-di-Cl)
25 (3,5-di-Br)
26 (3,5-di-I)
82.8
76.7
76.4
82.6
77.8
70.0
4-F
41.5
49.3
49.3
45.9
4-Cl
4-Br
4-I
1-hexene in the C6
fraction [%]
22 (3-tBu)
[3-Me]
28 (3,5-di-tBu)
27 (3,5-di-Me)
and the C8 → C10 propagation steps were used for the calculation
of ˛.
The oligomerization activities of complexes 18–21 exhibited
distinct differences. Analogously to the ortho-substituted catalysts
[38], the 3-chloro derivative 19 showed an extremely high activity
of 52050 [kg prod./mol Fe h] which is distinctively higher com-
pared with the 3-fluoro, 3-bromo, and 3-iodo derivatives 18, 20,
and 21. Actually, 19 is the most active bis(arylimino)pyridine iron
catalyst which does not contain any substituents at the ortho-
positions to the former amino nitrogen atoms. While the high
electronegativity of the chloro substituents leads to a decrease of
the electron density in the iminophenyl rings, the “medium”-sized
chloro atoms apparently leave enough space for the coordination
of ethylene molecules to the iron centers and for the growing
chain, whereas the meta-bromo atoms may partially interfere with
incoming ethylene molecules leading to a lower activity (22000
[kg prod./mol Fe h]). The comparatively low activity of the 3-fluoro
substituted complex 18 could be explained with a faster deactiva-
tion of the catalytically active centers due to the above described
ligand transfer reaction. A visible proof of that assumption was the
purple color of the resulting oligomer mixture due to the octahe-
drally coordinated Fe3+ ions which are bonded to the six nitrogen
atoms of the two bis(arylimino)pyridine ligands (see Scheme 7).
The small size of the 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. The neutral bis(arylimino)pyridine
iron(III) complexes are stable against this ligand transfer
reaction; it only occurs after activation of the complexes
with MAO in competition to the chain propagation reaction
(see Scheme 7). Due to the abstraction of an iron-methyl
group, the activated cationic iron(III) complexes are struc-
turally similar to neutral bis(arylimino)pyridine iron(II) complexes
which could only be isolated as ion pairs when the ligand
framework consisted of exclusively meta- or para-substituted
iminophenyl rings (with the above mentioned rare exceptions
[29,47,51,52]).
A higher value of the Schulz–Flory coefficient ˛ directly corre-
sponds to an increased propagation probability resulting in higher
molecular weight products. The upper limit ˛ = 1 is only reached
when molecular weight distributions are obtained in which the
maximum peak appears at higher carbon numbers (usually butenes
or hexenes are the main products) but in these cases the mathemat-
ical requirements for a Schulz–Flory distribution are not fulfilled
(an ideal Schulz–Flory distribution is obtained for ˛ ∼ 0.7). Eq. (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 cata-
lysts. Many efforts have been made especially to prepare catalysts
which selectively produce, e.g., only 1-hexene or 1-octene, 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 [67] and references therein]. For the catalysts 17–28, nar-
row molecular weight distributions were observed with ˛ < 0.40
(see Table 4).
For almost all oligomerization runs, very high initial ethy-
lene uptake rates could be observed (up to 10 l of ethylene/min).
Due to the gradual catalyst degradation, the ethylene flows
drop distinctively during the experiments, and for the sys-
tems 17/MAO and 18/MAO the ethylene flow had ceased before
the end of the regular run time (1 h). Complex 17 which was
described to be catalytically active in the polymerization of
1,3-butadiene [36], 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, complete
deactivation had to be noted after a run time of only 10 min
when no more ethylene was consumed. The Schulz–Flory coef-
ficient ˛ was determined to ˛ = 0.28 indicating a very narrow
product distribution. Regarding the fast deactivation of 17, com-
plexes bearing meta- or para-halogen substituted iminophenyl
rings were considered to exhibit an improved shield of the iron
center [38]. The oligomerization results of the iron(III) com-
plexes 18–21 with meta-substituted bis(arylimino)pyridine ligands
revealed that mainly butenes, hexenes, and octenes were pro-
duced, and the contents of higher alkenes did not exceed 12 wt.%
(Table 4).
Among the para-halogen substituted complexes, the iodo
derivative exhibited the highest activity (22825 [kg prod./mol Fe h],
see [38]) while for the fluoro, chloro, and bromo derivatives activi-
ties lower than 5000 [kg prod./mol Fe h] were observed. Therefore,
an improved shield of the iron center can be considered for
the meta-halogen substituted catalysts compared with the para-
substituted analogues. The beneficial effect of introducing iodo
substituents at the para-positions of bis(arylimino)pyridine ligands

C. Görl et al. / Journal of Molecular Catalysis A: Chemical 352 (2012) 110–127
119
Scheme 7. Activation of complex 18 with MAO and termination by ligand transfer reaction. [MAO-Me]− describes the MAO cages which act as counter anions.
[29] can be explained with the size and the good polarizability of
the voluminous iodo atoms due to partial compensation of the pos-
itive charges at the iron centers during the oligomerization process
(“electronical flexibility”, see [30] and [68]).
23, 24, and 26. In analogy to 18, the low activity of the 3,5-difluoro
substituted complex 23 could be explained with a fast deactivation
of the catalytically active centers due to ligand transfer reactions.
For 24 bearing two chloro substituents at the iminophenyl rings
a dramatically reduced activity was observed compared with the
mono-chloro derivative 19. This result can be explained with a
destabilization of the catalytically active species due to the pres-
ence of four strong electron withdrawing chloro substituents in the
ligand framework. The same tendency was found for the diiodo sub-
stituted compound 26 for which the activity reached only half of the
activity of the mono-iodo derivative 21. In contrast to the dichloro
derivative 24, the increased steric bulk provided by four iodo sub-
stituents may be responsible for the lower oligomerization activity
of the diiodo compound 26. According to the literature, the high-
est activities for bis(arylimino)pyridine iron complexes were found
for 2-methyl-4-halogen substitution patterns [29,69] and the pres-
ence of more than one halogen substituent per iminophenyl ring
made explanations of reactivities more difficult, however, it could
be stated that combinations of an electron donating group (at ortho
or para position) and an electron withdrawing substituent (at para
or ortho position) within one aryl ring gave the best activities.
For comparison purposes, complexes 22, 27, and 28 containing
meta-alkyl substituted bis(arylimino)pyridine ligands were tested
for the oligomerization of ethylene. Since halogen substituents
at the meta-positions of the iminophenyl rings lead to narrow
product molecular weight distributions and, therefore, low values
for the Schulz–Flory coefficient ˛, complexes 22, 27, and 28 also
revealed to be interesting candidates for ethylene oligomerization
reactions. Indeed, these three complexes produced mixtures with
very narrow molecular weight distributions (˛ = 0.23 for 22/MAO
and 27/MAO and ˛ = 0.20 for 28/MAO) indicating high selectivities
towards low molecular weight olefins.
The product mixtures obtained with catalysts 18–21/MAO bear-
ing meta-halogen substituted ligands exhibited ˛ values in the
range 0.27–0.36 indicating narrow molecular weight distributions.
These values are similar to those obtained for the corresponding
para-substituted catalysts and much lower than those of the ortho-
substituted iron(III) catalysts (see Table 5 and [38]) for which the
Schulz–Flory coefficients ˛ increased with increasing the size of
the ortho-halogen substituent (2-F: ˛ = 0.71 < 2-Cl: ˛ = 0.91 < 2-Br:
˛ = 0.97; the complexation reaction to give the 2-iodo substituted
iron(III) complex failed). While the overall contents of C4, C6, and
C8 olefins for the fluoro, chloro, and iodo derivatives 18, 19, and 21
were very similar (∼35%, 39%, 10%), the bromo substituted com-
pound 20 mainly produced 1-butene and less hexenes and octenes.
For the meta-dihalogen substituted catalysts 23–26, the overall
contents of C6 olefins were somewhat lower compared with cat-
alysts 18–21 corresponding to an enhanced steric bulk around the
iron centers. According to Table 5, the contents of C6 olefins roughly
increase in the order ortho- < di-meta- < meta- < para-halogen sub-
stitution.
Among the meta-dihalogen substituted bis(arylimino)pyridine
iron catalysts 23–26 the bromo compound 25 revealed to be
the most active candidate (44480 [kg prod./mol Fe h]). The
Schulz–Flory coefficients ˛ for the dihalogen substituted catalysts
23–26 were found in the same range as for the mono-halogen com-
pounds 18–21. Similarly to the mono-bromo derivative 20, 25/MAO
produced a distinctively higher amount of butenes (Schulz–Flory
coefficient ˛ = 0.33) and less amounts of the higher olefins com-
pared with the difluoro, dichloro, and diiodo substituted catalysts

120
C. Görl et al. / Journal of Molecular Catalysis A: Chemical 352 (2012) 110–127
Scheme 8. Dependence of the product compositions on the substitution pattern of the iron(III) catalyst (meta- and para-substituents were omitted in the lower picture due
to their marginal influence when at least one ortho-substituent is present).
3.4.2. Investigation of the isomerization and cooligomerization
behavior of complexes 17–28
dramatically for the 3-bromo compound 20. Scheme 9 shows the
C6 fraction of the GC spectrum obtained from the oligomer mix-
ture produced with the 3-fluoro compound 18/MAO along with an
assignment of the detected isomers. The assignment was proven
by GC analyses of the pure hydrocarbons.
Besides the expected 1-hexene (retention time tret = 6.42 min),
cis- and trans-2-hexene (tret ∼ 6.60 min and 6.75 min) were iden-
tified by comparison of the GC retention times of the pure
compounds and further GC/MS correlation using the NIST mass
spectral database while 3-hexene (cis or trans) was not observed.
Isomerization of the terminal 1-hexene into internal C6 olefins
probably proceeds through the coordination of 1-hexene to
the iron center and subsequent insertion in a 2,1-mode into
the postulated hydrido iron species [7,31,71] followed by ␤-
hydrogen elimination with the two possible pathways cis- and
trans-2-hexene (“chain-running mechanism”, see Ref. [72] and
Scheme 10).
Branched products additionally require the incorporation
of at least one molecule of ethylene or 1-butene into the
growing chain before ␤-hydrogen elimination takes place. Cor-
respondingly, less steric bulk around the iron centers favors
this coordination and incorporation (1,2- or 2,1-insertion) of 1-
butene. 3-Methyl-1-pentene (tret ∼ 6.10 min) and 2-ethyl-1-butene
(tret ∼ 6.60 min) result from cooligomerization reactions of one
molecule of ethylene and 1-butene. Starting with a 2,1-insertion
of 1-butene followed by incorporation of ethylene and termination
through ␤-hydrogen elimination, 3-methyl-1-pentene is obtained
(Scheme 11) which can further be isomerized to give 3-methyl-2-
pentene (not detected).
Alternatively, 3-methyl-1-pentene would be accessible from
ethylene and 2-butene (cis or trans). However, the contents of these
butene isomers in the mixtures are negligible. 2-Ethyl-1-butene
is produced when 1-butene is inserted into an Fe-ethyl bond in a
1,2-mode with subsequent ␤-hydrogen elimination (Scheme 12).
The presence of the residual C6 olefins (cis- and trans-
4-methyl-2-pentene, 2-methyl-1-pentene, 2-methyl-2-pentene,
2,3-dimethyl-2-butene) cannot be explained by simple isomeriza-
tion or cooligomerization reactions of ethylene and butenes. One
The size of the halogen or alkyl substituents at the
iminophenyl rings also effected the isomerization behavior of
the bis(arylimino)pyridine iron(III) catalysts. Complexes 17–28
as well as the 2-fluoro and all para-halogen substituted deriva-
tives produced not only the expected ␣-olefins, but additionally
gave internal and branched olefins (see Tables 4 and 5) due to
isomerization and cooligomerization reactions. The lack of ortho
substituents at the iminophenyl rings can be considered as an
essential requirement for this behavior since all complexes contain-
ing ortho-substituted 2,6-bis(arylimino)pyridine ligands (except
the 2-fluoro derivative) produced highly linear products with
selectivities for ␣-olefins above 95% indicating that the ortho-
substituents are voluminous enough to suppress isomerization
reactions of terminal into internal olefins.
Scheme 8 gives a general overview of the product composi-
tions obtained with differently substituted bis(arylimino)pyridine
iron(III) complexes. According to theoretical calculations of Ramos
[70], bis(arylimino)pyridine iron catalysts bearing ortho-methyl
substituents on the iminophenyl rings are unable to copolymerize
ethylene with higher ␣-olefins. Additional substituents in meta- or
para-positions (omitted in the lower part of Scheme 8 for clarity)
only have a minor influence on the product compositions of such
ortho-methyl substituted complexes.
For complexes 17–28, the isomerization of ␣-olefins and the
cooligomerization of ethylene with small ␣-olefins like 1-butene
can be verified via GC analyses, since branched and internal alkenes
were detected along with the expected 1-alkenes. While the selec-
tivities of catalysts 17–28 towards 1-butene in the C-4 fractions
always exceeded 98% (Table 4), a closer look on the C6 fractions
gave interesting results. The degree of isomerization of the hex-
ene fraction for the catalytic system 17/MAO was determined to
50.7% and, correspondingly, the relative amount of 1-hexene in
the C6 fraction was calculated to be only 49.3%. Among the meta-
monohalogen substituted complexes 18–22, the 3-fluoro and the
3-iododerivatives(18and21) exhibitedsimilartendenciestowards
the production of C6 isomers while the 1-hexene content increased

C. Görl et al. / Journal of Molecular Catalysis A: Chemical 352 (2012) 110–127
121
Scheme 9. C6 region of the GC spectrum of the oligomer mixture produced with 18/MAO.
Scheme 10. “Chain-running” of 1-hexene providing internal C6 isomers [72].
possible reaction pathway to these olefins could be the cleavage of
C–C bonds of internal C6 olefins within the coordination sphere of
the iron center yielding C3 “building blocks” which recombine fast.
These assumptions are based on mechanisms presented by Döt-
terl and Alt [73], Jacobson [74–76], Peake et al. [77], and Grubbs
and Miyashita [78] for rearrangement reactions of metallacyclic or
metal-bis(olefin) complexes. Dötterl proposed the initial coordina-
tion of two ethylene molecules to a formally dikationic nickel center
forming a metallacyclopentane. Partial ␤-hydrogen elimination of
that metallacyclic compound provided an alkenyl-hydride species
Scheme 11. Cooligomerization of 1-butene and ethylene yielding 3-methyl-1-pentene.

122
C. Görl et al. / Journal of Molecular Catalysis A: Chemical 352 (2012) 110–127
Scheme 12. Cooligomerization of ethylene and 1-butene yielding 2-ethyl-1-butene.
oligomerization reaction. Very probably, the third chloro ligand is
also exchanged against a methyl group due to the excess of MAO.
If that methyl group is then also abstracted by MAO, a dika-
tionic iron(III) species stabilized by two [MAO-Me]− cages would
result which can coordinate two olefin ligands at once. Since two
MAO cages are supposed to act as counter anions, there must be
enough space for a second cage. In fact, catalysts 17 and 18 featur-
ing the lowest steric bulk around the iron centers apparently fulfil
that requirement. The dikationic species may be formed upon the
initial activation with MAO (showing a Fe–Me unit) or after one
or more catalytic cycles of the commonly accepted monokationic
hydrido iron species (Fe-H moiety, see Scheme 14).
The formally dikationic catalyst species is much more destabi-
lized compared with the monokationic species leading to a faster
deactivation and, therefore, a lower activity. Both this destabi-
lization and the tendency towards ligand transfer reactions are
supposed to be responsible for the comparatively low catalytic
activities of complexes 17 and 18. Despite of its assumed dikationic
nature, the catalyst still produces 1-alkenes initially. According to
Scheme 13, a metallacycle could be formed easily from a dikationic
nickel center and two ethylene molecules. This metallacycle either
releases 1-butene or coordinates another ethylene molecule form-
ing 1-hexene or C6 isomers (depending on the insertion mode of
the 1-butene molecule). However, just the presence of an addi-
tional ethylene (or 1-butene) molecule in the coordination sphere
does not lead to the observed branched isomers. As indicated above,
cis- and trans-2-hexene were detected in the product mixtures.
The coordination of either cis- or trans-2-hexene to the dikationic
iron catalyst initially provides an iron-hex-2-yl species. Due to
the increased positive charge of the iron center (=strong Lewis
acid), stabilization may occur due to activation of an alkyl C–H
bond leading to a metallacyclic “ferracyclopentane” intermediate
(Scheme 15).
Scheme 13. Mechanism proposed by Dötterl for the MAO-free oligomerization of
ethylene in buffered ionic liquids with a formally dikationic nickel complex [73].
which either released 1-butene or coordinated another ethylene
molecule forming higher oligomers (see Scheme 13).
In analogy to these “naked” dikationic nickel species, dikationic
iron species can also be proposed for bis(arylimino)pyridine
iron(III) catalysts. While the pseudo-octahedral coordina-
tion sphere of the neutral iron(III) complexes (tridentate
bis(arylimino)pyridine ligand and three chloro ligands) has
been established (e.g., by Ionkin using crystallography [14,30]),
the nature of the third monodentate ligand at the iron atoms after
activation with MAO is still unknown. As described in Scheme 7,
methylation of the iron(III) catalysts followed by the abstraction
of a methyl anion generates a monokationic species with a [MAO-
Me]− cage as the counter anion. At least two of the chloro ligands
must have therefore been replaced by methyl groups to start the
The “dimethylferracyclopentane” species now can convert into
a hydrido-bis(propene)iron complex. Depending on the insertion
mode of the “propylene intermediates”, 2,3-dimethyl-1-butene
(twice 2,1-insertion) or 4-methyl-2-pentene and 2-methyl-1-
pentene (2,1- and 1,2-insertion) are obtained. 2,3-Dimethyl-
1-butene and 2-methyl-1-pentene can be further isomerized
to 2,3-dimethyl-2-butene and 2-methyl-2-pentene, respectively
(Scheme 16).
One aspect confirming the hypothesis that two alkyl/alkenyl
species are involved in the formation of these C6 olefins
could be the presence of 2,3-dimethyl-2-butene in the mixtures

C. Görl et al. / Journal of Molecular Catalysis A: Chemical 352 (2012) 110–127
123
Scheme 14. Proposed formation of a dikationic species from the bis(arylimino)pyridine iron(III) catalyst 18: initially bearing a Fe–Me group due to the MAO activation (upper
picture), after one or more catalytic circles bearing a Fe-H moiety due to ␤-hydrogen elimination of an olefin (lower picture).
produced with the unsubstituted catalyst 17, the 3-fluoro deriva-
tive 18, and the para-halogen substituted catalysts (see Table 5).
Among complexes 17–28, only catalysts 17 and 18 provide
enough space for the formation of the bulky 2,3-dimethylbutene
while the steric bulk of all other catalysts prevents its forma-
tion. Furthermore, similar product distributions containing cis-
and trans-4-methyl-2-pentene, 2-methyl-1-pentene, 2-methyl-2-
pentene, and 2,3-dimethylbutene are obtained with other iron or
nickel catalysts in propylene dimerization reactions [73,79–81]
while 2-ethyl-1-butene usually cannot be detected in propylene
dimerization experiments.
since cooligomerization of ethylene and propylene or 1-butene and
propylene cannot be excluded. However, according to the GC spec-
tra, C5 or C7 alkenes were not found in the product mixtures. In
Table 4, there are some examples which show that internal and
branched isomers are present in the product mixtures with con-
tents up to 50% of the C6 fractions (for catalysts 17 and 18). These
contents correspond to molar amounts of more than 100 mmol.
Even if all methyl groups of the applied MAO (max. ∼ 10 mmol Al)
would be used to methylate the iron centers (which is defi-
nitely not the case), the isolated amounts of isomers are much
higher.
As suggested by a referee, C3 building blocks may alterna-
tively be formed by an insertion of ethylene into Fe–Me bonds
(formed by initial methylation of the trichloride complex with
MAO or by regeneration of the catalyst during the oligomer-
ization) and subsequent ␤-hydrogen elimination of propylene
molecules which can then be dimerized. This reaction pathway
would also lead to oligomers with odd numbers of carbon atoms
The selectivities of the meta-disubstituted complexes 23–28
towards ␣-olefins, especially to 1-hexene, were decisively higher
compared with the meta-monosubstituted catalysts 18–22. Due
to the presence of two meta substituents in complexes 23–28,
the coordination of higher ␣-olefins to the iron centers is hin-
dered resulting in lower amounts of isomerized and branched
octenes. The exclusive production of 1-octene for 27/MAO could
Scheme 15. Proposed formation of a “ferracyclopentane” intermediate.

124
C. Görl et al. / Journal of Molecular Catalysis A: Chemical 352 (2012) 110–127
Scheme 16. Proposed formation of methylpentenes and dimethylbutenes over metallacyclic intermediates.
Scheme 17. GC spectrum of the oligomer mixture produced with 25/MAO (C4 to C10 region; small signals for higher oligomers were omitted for clarity). Offset: enlarged
view of the C6 region (hexenes).

C. Görl et al. / Journal of Molecular Catalysis A: Chemical 352 (2012) 110–127
125
Scheme 18. C6 regions of the GC spectra of the oligomer mixtures produced with meta-monosubstituted (left column) and meta-disubstituted catalysts (right column). For
the mixture obtained with the 3-fluoro derivative 18/MAO, the peak for 2,3-dimethyl-2-butene appears at tret = 6.95 min.

126
C. Görl et al. / Journal of Molecular Catalysis A: Chemical 352 (2012) 110–127
be explained with the quite low activity of the catalyst which is
caused by fast deactivation due to ligand transfer reactions. The
presence of two bulky tert-butyl groups per iminophenyl ring in
complex 28 protects the iron centers moderately from deactiva-
tion by ligand transfer reactions leading to a good activity of 20400
[kg prod./mol Fe h].
complex type, the meta-functionalized 2,6-bis(arylimino)pyridine
iron(III) complexes are rare examples which are able to copoly-
merize ethylene with higher ␣-olefins and to isomerize terminal
into internal olefins.
Acknowledgements
Scheme 17 shows the GC spectrum of the oligomer mixture
produced with 25/MAO with special attention to the C6 region.
For 25/MAO, the internal isomers cis-2-hexene (tret = 6.75 min)
and trans-2-hexene (tret ∼ 6.60 min) as well as the branched
isomers 4-methyl-2-pentene (tret = 6.12 min) and 2-methyl-2-
pentene (tret ∼ 6.60 min) were observed. Due to the increased steric
bulk of the bromo substituents compared to fluoro or chloro
atoms, the content of the “cooligomerization” product 2-ethyl-1-
butene significantly decreased (peak around tret ∼ 6.58 missing, see
Scheme 9 for comparison) indicating that the 1,2-insertion of 1-
butene into the growing chain is hindered (for the subsequent
formation of 2-ethyl-1-butene the growing chain is equivalent to an
ethyl group). trans-4-Methyl-1-pentene, 2-methyl-1-pentene, and
2,3-dimethylbutene were also not found in the mixture produced
with 25/MAO.
Among the iron(III) catalysts shown in Table 5, the para-
halogen substituted complexes and complex 17 with unsubstituted
iminophenyl rings exhibited the greatest tendencies towards iso-
merization reactions, and remarkably low contents of 1-hexene
were found in the C6 fractions of the product mixtures (always
lower than 50%, see Table 5). Compared with the para-halogen
substituted catalysts, the complexes 18–21 afforded increased
amounts of 1-hexene (52–81%) in the C6 fractions (see Table 5).
For meta-dihalogen substituted catalysts, a further increase of the
1-hexene content could be noted (except for the bromo substituted
compounds where the selectivity of the mono-bromo derivative 20
was marginally higher compared with the value for the dibromo
derivative 25). In analogy to the first publications of Brookhart and
Gibson [1–7], the ortho-substituted compounds showed the highest
selectivities towards ␣-olefins (except the 2-fluoro derivative). For
comparison, Scheme 18 shows the enlarged C6 fractions in the GC
spectra obtained for the meta-monosubstituted catalysts (18–22;
left) and corresponding meta-disubstituted catalysts (23–26 and
28; right) indicating the increased selectivities towards 1-hexene
of the latter ones.
We thank Saudi Basic Industries Corporation (SABIC), Saudi Ara-
bia, for the financial support and Christopher Synatschke (MC II,
University of Bayreuth) for recording the MALDI-TOF MS spectra.
References
[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.
[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.
[3] G.J.P. Britovsek, V.C. Gibson, D.F. Wass, Angew. Chem. Int. Ed. 37 (1999) 428.
[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.
[5] B.L. Small, M. Brookhart, J. Am. Chem. Soc. 120 (1998) 7143.
[6] B.L. Small, M. Brookhart, A.M.A. Bennett, J. Am. Chem. Soc. 120 (1998) 4049.
[7] B.L. Small, M. Brookhart, Macromolecules 12 (1999) 2120.
[8] Y. Chen, C. Qian, J. Sun, Organometallics 22 (2003) 1231.
[9] Y. Chen, R. Chen, C. Qian, X. Dong, J. Sun, Organometallics 22 (2003) 4312.
[10] Z. Zhang, S. Chen, X. Zhang, H. Li, Y. Ke, Y. Lu, Y. Hu, J. Mol. Catal. A: Chem. 230
(2005) 1.
[11] F. Pelascini, F. Peruch, P.J. Lutz, M. Wesolek, J. Kress, Eur. Polym. Sci. 41 (2005)
1288.
[12] K. Lappalainen, K. Yliheikkilä, A.S. Abu-Surrah, M. Polamo, M. Leskelä, T. Repo,
Z. Anorg, Allg. Chem. 631 (2005) 763.
[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.
[14] A.S. Ionkin, W.J. Marshall, D.J. Adelman, B. Bobik Fones, B.M. Fish, M.F.
Schiffhauer, Organometallics 25 (2006) 2978.
[15] V.C. Gibson, N.J. Long, P.J. Oxford, A.J.P. White, D.J. Williams, Organometallics
25 (2006) 1932.
[16] Y.V. Kissin, C. Qian, G. Xie, Y. Chen, J. Polym. Sci. A: Chem. 44 (2006) 6159.
[17] F.A.R. Kaul, G.T. Puchta, G.D. Frey, E. Herdtweck, W.A. Herrmann,
Organometallics 26 (2007) 988.
[18] D. Takeuchi, R. Matsuura, Y. Fukuda, K. Osakada, Dalton Trans. 41 (2009) 8955.
[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.
[20] D. Takeuchi, R. Matsuura, S. Park, K. Osakada, J. Am. Chem. Soc. 129 (2007) 7002.
[21] K.P. Tellmann, V.C. Gibson, A.J.P. White, D.J. Williams, Organometallics 24
(2005) 280.
[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).
[25] A. Rodriguez-Delgado, J. Campora, A.M. Naz, P. Palma, M.L. Reyes, Chem. Com-
mun. (2008) 5230.
In conclusion, iron(III) complexes with meta-substituted
iminophenyl rings proved to be highly active catalysts for
ethylene oligomerization reactions leading to product mixtures
with narrow molecular weight distributions and displayed the abil-
ity to isomerize and copolymerize small olefins like 1-butene.
[26] R. Schmidt, U. Hammon, M.B. Welch, H.G. Alt, B.E. Hauger, R.D. Knudsen, US
6458905 (2002).
[27] I.S. Paulino, U. Schuchardt, J. Mol. Catal. A: Chem. 211 (2004) 55.
[28] Z. Zhang, J. Zou, N. Cui, Y. Ke, Y. Hu, J. Mol. Catal. A: Chem. 219 (2004) 249.
[29] C. Goerl, H.G. Alt, J. Organomet. Chem. 692 (2007) 4580.
[30] 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.
[31] C. Goerl, H.G. Alt, J. Mol. Catal. A: Chem. 273 (2007) 118.
[32] A.S. Ionkin, WO 2007059015 (2007).
[33] D.D. Devore, S.S. Feng, K.A. Frazier, J.T. Patton, WO 2000069923 (2000).
[34] K.P. Bryliakov, N.V. Semikolenova, V.N. Zudin, V.A. Zakharov, E.P. Talsi, Catal.
Commun. 5 (2004) 45.
[35] F. Prades, R. Spitz, C. Boisson, S. Sirol, A. Razavi, WO 2007014889 (2007).
[36] D. Gong, B. Wang, C. Bai, J. Bi, F. Wang, W. Dong, X. Zhang, L. Jiang, Polymer 50
(2009) 6259.
[37] P. Hao, Y. Chen, T. Xiao, W.-H. Sun, J. Organomet. Chem. 695 (2010) 90.
[38] C. Goerl, T. Englmann, H.G. Alt, J. Appl. Catal. A: Gen. 403 (2011) 25.
[39] E.P. Talsi, D.E. Babushkin, N.V. Semikolenova, V.N. Zudin, V.N. Panchenko, V.A.
Zakharov, Macromol. Chem. Phys. 202 (2001) 2046.
[40] J. Martinez, V. Cruz, J. Ramos, S. Gutierrez-Oliva, J. Martinez-Salazar, A. Toro-
Labbe, J. Phys. Chem. C 112 (2008) 5023.
[41] V. Cruz, J. Ramos, J. Martinez-Salazar, S. Gutierrez-Oliva, A. Toro-Labbe,
Organometallics 28 (2009) 5889.
4. Conclusion
Iron(III)
complexes
containing
meta-substituted
2,6-
bis(arylimino)pyridine ligand frameworks were synthesized 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 strongly influence the product compositions.
Although the iminophenyl rings of the ligand frameworks do
not contain substituents in ortho-positions to the former amino
groups, good catalytic activities were observed. Besides ␣-olefins,
also internal and branched olefins were detected by GC analyses.
Explanations for this unusual behavior consider the coordination
and insertion of small ␣-olefins like 1-butene into the growing
chains and intermediate C–C bond cleavage reactions. Some of the
resulting methylpentenes require the presence of C3 units in the
reaction mixtures which are supposed to form upon coordination
of internal C6 olefins over metallacyclic intermediates. Among this
[42] R. Raucoules, T. de Bruin, P. Raybaud, C. Adamo, Organometallics 27 (2008)
3368.
[43] 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.

C. Görl et al. / Journal of Molecular Catalysis A: Chemical 352 (2012) 110–127
127
[44] M. Seitz, C. Goerl, W. Milius, H.G. Alt, Jordan J. Chem. 3 (2008) 109.
[45] A.S. Abu-Surrah, K. Lappalainen, U. Piironen, P. Lehmus, T. Repo, M. Leskelä, J.
Organomet. Chem. 648 (2002) 55.
[46] R.-F. Chen, C.-T. Qian, J. Sun, Chin. J. Chem. 19 (2001) 866.
[47] M.E. Bluhm, C. Folli, M. Döring, J. Mol. Catal. A: Chem. 212 (2004) 13.
[48] B. Cetinkaya, E. Cetinkaya, M. Brookhart, P.S. White, J. Mol. Catal. A: Chem. 142
(1999) 101.
[63] H. Kehlen, M.T. Rätsch, Z. Phys. Chem. 256 (1984) 1049.
[64] M. Seitz, Diploma thesis, University of Bayreuth, Germany, 2001.
[65] S.A. Svejda, M. Brookhart, Organometallics 18 (1999) 65.
[66] G. Henrici-Olive, S. Olive, Adv. Polym. Sci. 15 (1974) 1.
[67] 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.
[49] O. Dayan, F. Dogan, I. Kaya, B. Cetinkaya, Syn. React. Inorg. Metal-Org. Nano-
Metal Org. Chem. 40 (2010) 337.
[50] A.S. Ionkin, L.K. Johnson, US 2006/0009597 (2006).
[51] J. Jia, CN 2003-105297 20030228 (2004).
[68] T. Zhang, W.H. Sun, T. Li, X.J. Yang, J. Mol. Catal. A: Chem. 218 (2004) 119.
[69] M. Watanabe, H. Sato, M. Kuramoto, H. Tashiro, S. Tanaka, T. Tamura, WO
2001017967 (2001).
[70] J. Ramos, V. Cruz, A. Munoz-Escalona, J. Martinez-Salazar, Polymer 43 (2002)
3635.
[52] C. Liu, G. Zhou, G.X. Jin, CN 1306012 (2001).
[53] D.R. Haydon, L.G. Czaplewski, N.J. Palmer, D.R. Mitchell, J.F. Atherall, C.R. Steele,
T. Ladduwahetty, WO 2007148093 (2007).
[54] R. Schmidt, M.B. Welch, S.J. Palackal, H.G. Alt, J. Mol. Catal. A: Chem. 179 (2002)
155.
[55] C. Qian, F. Gao, Y. Chen, L. Gao, Synlett (2003) 1419.
[56] G.-X. Jin, C. Liu, CN 1352204 (2002).
[57] R.F. Moreira, E.Y. Tshuva, S.J. Lippard, Inorg. Chem. 43 (2004) 4427.
[58] R. Li, R.L. Smith, H.I. Kenttamaa, J. Am. Chem. Soc. 116 (1996) 5056.
[59] V. Malatesta, K.U. Ingold, J. Am. Chem. Soc. 103 (1981) 609.
[60] G.V. Schulz, Z. Phys. Chem. B 30 (1935) 379.
[71] B.L. Small, A.J. Marcucci, Organometallics 20 (2001) 5738.
[72] C.M. Killian, L.K. Johnson, M. Brookhart, Organometallics 16 (1997)
2005.
[73] M. Dötterl, H.G. Alt, ChemCatChem (2011), doi:10.1002/cctc.201100182.
[74] D.B. Jacobson, B.S. Freiser, Organometallics 3 (1984) 513.
[75] D.B. Jacobson, B.S. Freiser, J. Am. Chem. Soc. 105 (1983) 736.
[76] D.B. Jacobson, B.S. Freiser, J. Am. Chem. Soc. 105 (1983) 5197.
[77] D.A. Peake, M.L. Gross, D.P. Ridge, J. Am. Chem. Soc. 106 (1984) 4307.
[78] R.H. Grubbs, A. Miyashita, J. Am. Chem. Soc. 100 (1978) 7416.
[79] H.G. Alt, K.J. Schneider, E. Funk, Jordan J. Chem. 3 (2008) 367.
[80] H.G. Alt, K.J. Schneider, C. Goerl, Jordan J. Chem. 4 (2009) 33.
[81] M. Dötterl, H.G. Alt, Adv. Synth. Catal., doi:10.1002/adsc.201100557.
[61] G.V. Schulz, Z. Phys. Chem. B 43 (1939) 25.
[62] P.J. Flory, J. Am. Chem. Soc. 62 (1940) 1561.