Biphasic ethylene oligomerization using bis(imino)pyridine cobalt complexes in methyl-butylimidazolium organochloroaluminate ionic liquids

Article abstract of DOI:10.1016/j.molcata.2011.03.015

Several bis(imino)pyridine cobalt (II) complexes have been synthesized and used in ethylene oligomerization in 1-methyl-3-butylimidazolium tetrachloroaluminate (BMI·AlCl₄) ionic liquid activated by methylaluminoxane (MAO). The ligand substituti

Full text of DOI:10.1016/j.molcata.2011.03.015

Journal of Molecular Catalysis A: Chemical 340 (2011) 83–88
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Biphasic ethylene oligomerization using bis(imino)pyridine cobalt complexes in
methyl-butylimidazolium organochloroaluminate ionic liquids
Daniel Thiele, Roberto Fernando de Souza∗
Institute of Chemistry, UFRGS, Av. Bento Gonc¸ alves, 9500, P.O. Box 15003, CEP 91501-970, Porto Alegre, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history:
Several bis(imino)pyridine cobalt (II) complexes have been synthesized and used in ethylene oligomer-
ization in 1-methyl-3-butylimidazolium tetrachloroaluminate (BMI·AlCl4) ionic liquid activated by
methylaluminoxane (MAO). The ligand substitution pattern, with halogen, methyl or trifluoromethyl
units in the ortho- and para-positions of the aromatic group, has a marked effect on activity and selec-
tivity. The cobalt catalysts evaluated in this work exhibited activities in the range 4000–15,300 h−1 and
selectivities towards ethylene dimerization to butenes above 90%. These catalysts also exhibited high
selectivity to 1-butene formation, up to 87%.
Received 25 January 2011
Received in revised form 11 March 2011
Accepted 16 March 2011
Available online 31 March 2011
Keywords:
Ethylene
Oligomerization
Cobalt
© 2011 Elsevier B.V. All rights reserved.
Organoaluminate
Ionic liquid
1. Introduction
Despite efforts to develop more efficient oligomerization cat-
alysts, limited success has been achieved in terms of systems
The use of “green” solvents that allow catalyst recovery, low
environmental impact and reduced production costs is a great chal-
lenge in industrial processes. A strategic approach is the use of
biphasic catalysis using solvents that are immiscible with the reac-
tion products and are efficient at immobilizing transition metal
catalysts. The use of ionic liquids (ILs) allows for recycling of the
catalyst by simple decantation of the catalyst solution from the
products.
Since the pioneer reports by Chauvin et al. [1], concerning the
utilization of ionic liquids for the immobilization of oligomerization
catalysts, many efforts have been devoted to achieve a catalyst for
the selective synthesis of linear ␣-olefins by ethylene oligomer-
ization [2–8]. A great amount of the literature concerning the
oligomerization reaction is devoted to nickel complexes [2–7], but
examples of several other transition metal complexes that have
been shown to promote carbon–carbon bond formation are known.
Examples of a tungsten catalyst [8] and more recently, cobalt and
iron complexes [9] have been reported.
active for the selective synthesis of linear ␣-olefins in ionic liquids
[2]. As a matter of fact, nickel oligomerization catalysts immo-
bilized in chloroaluminate ionic liquids present, in parallel with
the oligomerization reaction, a high carbon–carbon double bond
isomerization rate [3–7,9–13], generating internal and branched
olefins. Another drawback of these ionic liquids observed in some
studies was diminished selectivity during the oligomerization of
olefin as a result of the abstraction ligand coordinated to the nickel
by the aluminum species [5,14]. This aspect has been minimized by
the addition of bases into the ionic liquid with an AlCl3/1-methyl-
3-butylimidazolium chloride (BMICl) ratio higher than 1 [14,15].
Chloroaluminate ionic liquids show good abilities in immobiliz-
ing coordination compounds and incorporating organoaluminum
co-catalyst due to their intrinsic characteristics. These character-
istics of the dialkylimidazolium organoaluminates allows for a
drastic reduction in catalyst and co-catalyst consumption. Some
authors have evaluated reductions in catalyst consumption by a
factor of ten and of the co-catalyst by a factor of at least two [16].
Complexes containing bis(imino)pyridine ligands have
attracted much attention for their application in oligo/
polymerization reactions [17–19]. Some of these complexes
are able to convert ethylene with high oligomerization rates,
producing a majority of ␣-olefins. The performance of these
complexes in homogeneous catalysis, the strong coordination of
this kind of ligand to transition metals [20] and the arrangement
of its pyridine and N-aryl ring almost orthogonally to each to other
make them suitable candidates for tests in oligomerization of low
molecular weight olefins when immobilized in chloroaluminate
A plethora of ligand structures are associated with transition
metals and were studied in these reactions. Some relevant exam-
ples of nickel complexes are shown in Scheme 1, like bidentate
[P,O] [2], structure A, ␣-diimine, structure B, bis(imino)pyridine [3],
structure C, 1,2-diiminophosphorane [5], structure D, phosphines
[6], bis-(salicylaldimine) [7], structure E.
∗
Corresponding author. Tel.: +55 51 3308 6303; fax: +55 51 3308 7304.
E-mail addresses: rfds@ufrgs.br, rfds@iq.ufrgs.br (R.F. de Souza).
1381-1169/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2011.03.015

84
D. Thiele, R.F. de Souza / Journal of Molecular Catalysis A: Chemical 340 (2011) 83–88
N
P
N
Ni
N
[SbF6]-
Ni
Cl
Cl
N
N
Ar
Ar
O
P
Ni
Cl
Cl
R
C
B
Br-
N
A
N
R
Br-
O
N
N
O
Ni
N
N
R
Ph3P
N
N
PPh3
Ni
Cl
Cl
E
R
D
Scheme 1. Nickel complexes for ethylene oligomerization in ionic liquids.
Ligand 1: Yield 51%. 1H NMR (300 MHz, CDCl3): ı ppm 8.35 (2H,
m-H Py), 7.87 (1H, p-H Py), 7.38 (4H, m-ArH), 7.12 (2H, p-ArH), 6.85
(4H, o-ArH), 2.41 (6H, CH3–C N). 13C NMR (75 MHz, CDCl3): ı ppm
167.3 (C N), 155.4 (o-C Py), 151.3 (i-C Ar), 136.8 (p-C Py), 129.0 (m-
C Ar), 123.6 (p-C Ar), 122.3 (m-C Py), 119.2 (o-C Ar), 16.2 (CH3–C N).
ionic liquids. Herein, we describe the performance of a series of
bis(imino)pyridine cobalt(II) complexes in an ethylene oligomer-
ization reaction using as the ionic liquid BMI·AlCl4 activated by
MAO.
IR: C N 1635 cm−1
.
2. Experimental
Ligand 2: Yield 42%. 1H NMR (300 MHz, CDCl3): ı ppm 8.41 (2H,
m-H Py), 7.89 (1H, p-H Py), 7.30–7.15 (4H, 3-ArH+5-ArH), 7.04 (2H,
4-ArH), 6.70 (2H, 6-ArH), 2.35 (6H, CH3–C N), 2.13 (6H, o-CH3Ar).
13C NMR (75 MHz, CDCl3): ı ppm 166.8 (C N), 155.4 (o-C Py), 149.9
(i-C Ar), 136.8 (p-C Py), 130.4 (3-C Ar), 127.1 (2-C Ar), 126.4 (5-C Ar),
123.6 (4-C Ar), 122.2 (m-C Py), 118.1 (6-C Ar), 17.8 (o-CH3 Ar), 16.3
All the manipulations were conducted under an argon
atmosphere using Schlenk tube techniques. Anilines, 2,6-
diacetylpiridine and silica–alumina grade-135 were purchased
form Aldrich and used without further purification. Solvents were
dried over Na and then distilled under an argon atmosphere.
CoCl2·6H2O was dehydrated as described [21]. The CoCl2·(THF)1.5
adduct was prepared as previously described [22]. The ionic liquid
BMI·AlCl4 was obtained by reacting 1-butyl-3-methylimidazolium
chloride (BMICl) with aluminum chloride as described in the
literature [23].
Elemental analyses were carried out in a Perkin Elmer M CHN/O
Analyser, Model 2400, with a tolerance for carbon, hydrogen and
nitrogen data of 0.4. The acquisition of 1H NMR and 13C NMR spec-
tra was performed on an Inova 300 MHz (Varian), with the samples
dissolved in CDCl3. The IR analyses of the ligands and complexes
were performed on a Varian 640-IR spectrometer. Data was col-
lected with 32 scans, resolution 4 cm−1 in a 400–4000 cm−1 range.
Chromatographic analyses were performed on a Varian 3400 CX
gas chromatograph, with a Petrocol DH capillary column: methyl
silicone, 100 m long, 0.25 mm i.d., and 0.5 ␮m film thickness. Ana-
lytical conditions: 30 min at 36 ◦C, heating rate of 5 ◦C/min until
reach 250 ◦C.
(CH3–C N). IR: C N 1637 cm−1
.
Ligand 3 was prepared as described by Alt and co-workers
[24]. To a stirred solution of 2,6-diacetylpiridine in 70 mL of
benzene, 1 equivalent of p-toluenesulfonic acid was added. The
2,6-dimethylaniline compound was added in excess (2.2 equiva-
lents to 2,6-diacetylpiridine). The mixture was refluxed and the
water formed in the reaction was separated using Dean–Stark distil-
lation. Refluxing was continued for 24 h and benzene was removed
under reduced pressure. Dichloromethane (30 mL) was added and
the ligand was neutralized in 30 mL of a solution saturated with
Na2CO3. The organic layer was separated and the aqueous phase
was extracted twice with dichloromethane. The combined organic
layers were evaporated to dryness. Ethanol was added to the dried
residue and the ligand was filtered and dried under reduced pres-
sure.
Ligand 3: Yield 62%. 1H NMR (300 MHz, CDCl3): ı ppm 8.49 (2H,
m-H Py), 7.92 (1H, p-H Py), 7.09 (4H, m-ArH), 6.95 (2H, p-ArH), 2.25
(6H, CH3–C N), 2.06 (12H, o-CH3Ar). 13C NMR (75 MHz, CDCl3): ı
ppm 167.2 (C N), 155.1 (o-C Py), 148.7 (i-C Ar), 136.8 (p-C Py), 127.9
(m-C Ar), 125.4 (o-C Ar), 123.0 (p-C Ar), 122.2 (m-C Py), 17.9 (o-CH3
2.1. Ligands and complexes synthesis
Ar), 16.4 (CH3–C N). IR: C N 1645 cm−1
.
Ligands 1 and 2 were prepared as described by Brookhart and
co-workers [17]. The synthesis of these ligands was done by mixing
2,6-diacetylpiridine and the corresponding aniline in a 1:3 molar
ratio in 20 mL of dichloromethane. Formic acid (0.4 mL) was added
and the solution was stirred for four days. Dichloromethane was
removed under reduced pressure and the ligand was precipitated
by the addition of ethanol. The solution was filtered and the solid
was washed with ethanol. The ligand was dissolved in hot ethanol
and precipitated a second time, filtered and dried under reduced
pressure.
The ligands 4–7 were prepared as described by Qian et al. [25].
2,6-Diacetylpiridine and 2.5 equivalents of the corresponding ani-
line were added to 40 mL of toluene. Silica–alumina grade-135
˚
(80 g/mol of 2,6-diacetylpiridine) and 1.5 g of molecular sieves 4 A
were added. The mixture was stirred for four days. The suspen-
sion was filtered and the solid washed with four aliquots of 5 mL
of toluene. The solvent was removed under reduced pressure and
the ligand washed with ethanol and dried under reduced pres-
sure.

D. Thiele, R.F. de Souza / Journal of Molecular Catalysis A: Chemical 340 (2011) 83–88
85
Ligand 4: Yield 52%. 1H NMR (300 MHz, CDCl3): ı ppm 8.37
(2H, m-H Py), 7.92 (1H, p-H Py), 7.69 (2H, 3-ArH), 7.53 (2H, 5-ArH)
7.20 (2H, 4-ArH), 6.82 (2H, 6-ArH), 2.38 (6H, CH3–C N). 13C NMR
(75 MHz, CDCl3): ı ppm 168.5 (C N), 154.9 (o-C Py), 149.6 (i-C Ar),
137.2 (p-C Py), 132.6 (5-C Ar), 126.5 (3-C Ar), 123.9 (CF3), 123.2 (4-C
Ar), 123.1 (m-C Py), 119.7 (6-C Ar), 119.5 (2-C Ar), 16.8 (CH3–C N).
Table 1
Ethylene oligomerization performance of cobalt-bis(imino)pyridine catalysts in
BMI·AlCl4.
Entry Catalyst Temperature
(◦C)
Al/Co TOF (h−1
)
C4 (%) 1-C4 (%) C6 (%)
1
2
3
4
5
6
7
8
9
10
11
12
13
14a
1a
2a
3a
4a
5a
5a
5a
5a
5a
5a
5a
6a
7a
1a
30
30
30
30
30
30
10
30
50
30
30
30
30
50
600
600
600
600
200
400
600
4300
6000
5300
9800
4800
8600
6200
98.5
96.0
94.8
96.8
98.5
94.7
97.2
95.6
89.4
96.3
98.3
96.0
96.5
63.7
84.1
87.0
79.2
80.1
95.8
83.8
89.3
80.1
67.1
85.7
95.2
83.6
83.2
96.9
1.5
4.0
5.2
3.2
1.5
5.3
2.8
4.4
10.6
3.7
IR: C N 1650 cm−1
.
Ligand 5: Yield 46%. 1H NMR (300 MHz, CDCl3): ı ppm 8.36 (2H,
m-H Py), 7.92 (1H, p-H Py), 7.42 (2H, 3-ArH), 7.26 (2H, 5-ArH) 6.79
(2H, 6-ArH), 2.38 (6H, CH3–C N). 13C NMR (75 MHz, CDCl3): ı ppm
169.5 (C N), 158.5 (4-C Ar), 154.8 (o-C Py), 145.6 (i-C Ar), 137.2 (p-C
Py), 123.2 (m-C Py), 123.0 (CF3), 121.4 (6-C Ar), 120.7 (2-C Ar), 119.5
600 15,300
600 17,000
(5-C Ar), 113.8 (3-C Ar), 16.8 (CH3–C N). IR: C N 1644 cm−1
.
800
1000
5100
3000
Ligand 6: Yield 33%. 1H NMR (300 MHz, CDCl3): ı ppm 8.35 (2H,
m-H Py), 7.92 (1H, p-H Py), 7.68 (2H, 3-ArH), 7.50 (2H, 5-ArH) 6.77
(2H, 6-ArH), 2.38 (6H, CH3–C N). 13C NMR (75 MHz, CDCl3): ı ppm
169.2 (C N), 154.6 (o-C Py), 148.0 (i-C Ar), 137.3 (p-C Py), 132.6 (5-C
Ar), 128.7 (4-C Ar), 126.7 (3-C Ar), 123.3 (m-C Py), 123.0 (CF3), 121.3
1.7
4.0
600 13,700
600 10,400
3.5
600
1200
36.3b
Reaction conditions: complex: 10 ␮mol; ionic liquid: 3 mL of BMI·AlCl4; solvent:
20 mL of toluene; ethylene: 10 bar; co-catalyst: MAO; time: 30 min.
(6-C Ar), 120.9 (2-C Ar), 16.9 (CH3–C N). IR: C N 1639 cm−1
.
a
Without ionic liquid.
C10 products were detected.
Ligand 7: Yield 41%. 1H NMR (300 MHz, CDCl3): ı ppm 8.35 (2H,
m-H Py), 7.92 (1H, p-H Py), 7.81 (2H, 3-ArH), 7.64 (2H, 5-ArH) 6.86
(2H, 6-ArH), 2.38 (6H, CH3–C N). 13C NMR (75 MHz, CDCl3): ı ppm
169.1 (C N), 154.6 (o-C Py), 148.5 (i-C Ar), 137.3 (p-C Py), 135.6 (5-C
Ar), 129.5 (3-C Ar), 124.8 (4-C Ar), 123.3 (m-C Py), 122.9 (CF3), 121.6
b
maintained at this temperature for six more hours. The solution
was filtered and used as prepared.
(6-C Ar), 121.2 (2-C Ar), 16.9 (CH3–C N). IR: C N 1643 cm−1
.
2.3. Oligomerization of ethylene
The bis(imino)pyridine cobalt (II) complexes were synthesized
by reacting the corresponding ligand with CoCl2·(THF)1.5 in THF.
The ligand and the cobalt adduct, 1.05 equivalents of the ligand
to adduct, were dissolved in 10 mL of anhydrous THF. The solution
was stirred for 24 h and then was filtered off, the solid washed with
four aliquots of 5 mL of cold THF. The complexes were dried under
reduced pressure. Scheme 2 shows the synthesis of the ligands and
cobalt complexes.
The catalytic runs were performed using a double-walled glass
reactor equipped with magnetic stirring, a thermocouple, and with
a continuous feed of ethylene at 10 bar. The reactor was fed with
10 ␮mol of the cobalt complex and 20 mL of toluene. The tempera-
ture was controlled by a thermostatic circulation bath. The system
was initially purged with ethylene, then a solution of the cobalt
catalyst dissolved in 3 mL of the ionic liquid BMI·AlCl4 was added
to the reactor. The reaction was started with the addition of the
desired amount of MAO. The reaction was stopped after 30 min.
The phases were allowed to separate and the organic phase was
withdrawn by a cannula and analyzed by GC.
Chromatographic analyses were performed on a Varian 3400
CX, equipped with a Petrocol DH capillary column: methyl sili-
cone, 100 m long, i.d. 0.25 mm, and film thickness 0.5 ␮m. Analysis
conditions: 36 ◦C for 15 min and a heating rate of 5 ◦C/min until
250 ◦C. The productivity was expressed as turnover frequencies,
TOF, defined as mol converted ethylene per mol pre-catalyst and
by reaction time (in hours).
Complex 1a: Yield 77%.
C = 56.91%; H = 4.32%; N = 9.48%. CHN found: C = 56.50%; H = 4.04%;
N = 9.20%. IR: C N 1590 cm−1
C21H19Cl2CoN3. CHN calculated:
.
Complex 2a: Yield 97% C23H23Cl2CoN3·0.5 THF. CHN calculated:
C = 59.19% H = 5.36% N = 8.28%. CHN found: C = 59.00%; H = 5.75%;
N = 8.08%. IR: C N 1589 cm−1
.
Complex 3a: Yield 96%. C25H27Cl2CoN3·0.5 H2O. CHN calculated:
C = 59.07%; H = 5.55%; N = 8.27%. CHN found: C = 59.28%; H = 5.49%;
N = 7.42%. IR: C N 1587 cm−1
.
Complex 4a: Yield 87%. C23H17Cl2CoF6N3·0.5 THF. CHN cal-
culated: C = 48.80%; H = 3.44%; N = 6.83%. CHN found: C = 48.82%;
H = 3.20%; N = 6.56%. IR: C N 1589 cm−1
.
Complex 5a: Yield 86%. C23H15Cl2CoF8N3·1 THF. CHN calcu-
3. Results and discussion
lated: C = 47.18%; H = 3.37%; N = 6.11%. CHN found: C = 47.17%;
H = 3.13%; N = 6.11%. IR: C N 1590 cm−1
.
Bis(imino)pyridine transition metal complexes have been
showed as efficient catalysts for the oligomerization and/or poly-
merization of olefins, and this has become a subject of intense
investigation [17–19]. More specifically, cobalt bis(imino)pyridine
catalysts allow for the selective synthesis of linear ␣-olefins, but
there are no available studies concerning their utilization under
biphasic conditions in ionic liquids.
Complex 6a: Yield 70%. C23H15Cl4CoF6N3·0.5 THF. CHN cal-
culated: C = 43.89%; H = 2.80%; N = 6.14%. CHN found: C = 43.80%;
H = 2.66%; N = 5.84%. IR: C N 1588 cm−1
.
Complex 7a: Yield 54%. C23H15Br2Cl2CoF6N3·0.5 THF. CHN cal-
culated: C = 38.84%; H = 2.48%; N = 5.44%. CHN found: C = 38.95%;
H = 2.25%; N = 5.29%. IR: C N 1591 cm−1
.
In the present study, several cobalt coordination compounds
were synthesized and their catalytic performance in biphasic
imidazolium–organoaluminate ionic liquids was evaluated. Table 1
shows the performance of selected complexes in the biphasic
oligomerization of ethylene under different reaction conditions.
2.2. MAO synthesis
The oligomerization tests of the cobalt catalyst were performed
using methylaluminoxane (MAO) as the co-catalyst. The MAO was
prepared as described by Reddy et al. [26]. A solution of 20.9 mL
(0.218 mol) of TMA in toluene was added dropwise to a suspension
12.107 g (18.2 mmol) of Al2(SO4)3·18H2O in toluene under vigor-
ous stirring at −10 ◦C. After the addition of trimethylaluminum, the
solution was slowly heated to 30 ◦C and maintained at this temper-
ature for 4 h. The temperature was increased gradually to 50 ◦C and
3.1. Effect of the Al/Co ratio
Compound 5a was used for optimization of the effect of the alu-
minum/cobalt molar ratio Al/Co on the ethylene oligomerization
ratio was evaluated. Values of Al/Co from 200 to 1000 were stud-

86
D. Thiele, R.F. de Souza / Journal of Molecular Catalysis A: Chemical 340 (2011) 83–88
NH2
R1
R3
+
R1
N
R1
2
N
Co
N
O
O
Cl
Cl
THF
CoCl2.(THF)1.5
R2
R2
R3
R3
R2
R1 R2 R3
i)
1
2
3
4
5
6
7
H
H
H
H
H
F
H
H
Me
H
H
Co 1a 2a 3a
4a
Me Me CF3 CF3 CF3 CF3
Cl Br
5a
6a
7a
- H2O
N
Me
Me
CF3
CF3
R1
R2
R3
H
H
H
H
H
H
Me
H
H
F
R1
R1
H
H
H
N
N
R3
CF3 Cl
CF3 Br
H
H
R2
R3
R2
Scheme 2. Synthesis of the ligands and cobalt complexes.
ied [entries 5, 6, 8, 10 and 11]. The increase of the Al/Co molar
ratio leads to an increase in the activity in the oligomerization rate,
reaching a maximum for Al/Co = 600. This behavior indicates that
an Al/Co ratio below 600 promotes the maximization of the num-
ber of active species, but a further increase in the Al/Co ratio results
in the deactivation of the catalyst.
with the cationic cobalt complex [9]. However, despite the same
downward trend, the comparison of these data showed that the
selectivities observed for bis(imino)pyridine cobalt catalysts were
much higher to those observed for cationic cobalt complex, tak-
ing into account the temperature and especially when considering
similar activities. For illustrative purposes, we compared the sys-
tem Co-AlEtCl2 activity 14,100 h−1 and 48.8% of butene-1 [9], with
5a-MAO activity of 15,300 h−1 and 80.1% of butene-1, demonstrat-
ing the potential of this class of catalysts for selective dimerization
of ethylene to butene-1.
The selectivity in the formation of butenes is slightly dependent
on the Al/Co molar ratio used in the catalytic runs. In all the Al/Co
molar ratios tested, the selectivity towards butenes remained in the
range of 94.7–98.5%. The selectivity for the formation of butene-1
changed with the Al/Co ratio in the range of 80.1–95.8%. As the co-
catalyst should act only to activate the complex into the catalyst, no
change on the selectivity from the catalyst should be expected. As
expected, no change was observed into dimer selectivity. The same
trend was not observed for the selectivity for butene-1. Increas-
ing the amount of the co-catalyst should only increase the number
of the active species until reach a maximum, without change the
structure of the catalyst with the amount of the co-catalyst. In face
of this observation, the change in the 1-butene selectivity is prob-
ably due the variation of the activity of the cobalt catalyst and not
from Al/Co ratio, activity and 1-butene selectivity have the same
trend with Al/Co ratio.
3.3. Influence of the substitution pattern on catalyst activity
All the cobalt catalysts were active for the oligomerization of
ethylene in BMI·AlCl4. The activity for catalyst 1a was low, 4300 h−1
.
The presence of substitutions on the bis(imino)pyridine ligands
proved to be beneficial for the activity of the cobalt catalysts,
regardless of the electronic nature of the substituent group, as in
entries 2–4, 8, 12 and 13. The presence of one methyl group at the
ortho position on catalyst 2a increased the activity to 6000 h−1
.
When the CF3 group was used as a substituent in the ortho posi-
tion of the ligand, the complex of the cobalt catalyst was more active
when compared to the 1a complex and the 2a complex. The activ-
ity for complex 4a for ethylene oligomerization was 9800 h−1 when
immobilized in BMI·AlCl4.
3.2. Effect of the reaction temperature
The reaction temperature showed a more pronounced influ-
ence in the activity of the oligomerization reaction of ethylene in
BMI·AlCl4. At 10 ◦C, the activity of the cobalt catalyst was restricted
to 6200 h−1 for the conversion of ethylene. Raising the reaction
temperature to 30 ◦C, the activity increased to 15,300 h−1. The tem-
perature reaction of 50 ◦C showed no further substantial increase
on the oligomerization activity. The activity for the cobalt complex
5a at 50 ◦C was 17,000 h−1. The selectivity for the dimerization of
ethylene was high at 10 ◦C, 97.2%. An increase in temperature to
30 ◦C for oligomerization resulted in a low change in selectivity,
95.6% for the formation of C4. A further increase of the oligomer-
ization reaction to 50 ◦C showed a sharper decrease in selectivity
in the formation of dimers to 89.4%.
The substitution at the ortho position in bis(imino)pyridine
ligands is beneficial for activity when this cobalt catalyst is
immobilized in BMI·AlCl4, regardless of the electronic nature of
the substituent. The requirement of a substitution at the ortho
position to improve the activity for oligomerization follows the
same trend observed for homogeneous catalysis with cobalt
bis(imino)pyridine catalysts [17,27,28]. Bazan and co-workers
observed for nickel catalyst a relationship of activity with respect
to the presence of a bulky substituent on the catalyst structure
[29]. These authors observed that, as the bulk of the substituents
increased, this was followed by a progressive decrease in the rate
of ring rotation at the more crowded coordination sphere around
nickel, increased monomer consumption activity [29].
The selectivity in the formation of butene-1 showed a strong
dependence on the temperature at which the oligomerization
reaction of ethylene was performed. At 10 ◦C, the catalyst had a
selectivity of 89.3% for butene-1 of total butenes. The test per-
formed at 30 ◦C showed selectivity for butene-1 of 80.1%, lower
than that obtained at 10 ◦C. The reduction in selectivity was even
more pronounced in the tests performed at 50 ◦C, with a selectiv-
ity of 67.1%. This profile is very similar to the decrease observed
Despite the increased oligomerization activity by the presence
of a substituent in the ortho position, the insertion of a second
methyl substituent proved useless for the activity of the catalyst.
For the case of the 3a complex, a decrease in oligomerization activ-
ity was observed when compared to the 2a complex, with only one
methyl substituent. This result suggests that although the presence
of substituent in this position is important, the effect on activity is
not only dependent on the spatial orientation of the ligand, but the

D. Thiele, R.F. de Souza / Journal of Molecular Catalysis A: Chemical 340 (2011) 83–88
87
electronic component is probably the predominant aspect in the
activity of a catalyst.
the electron density on the metal sites and generate a more elec-
trophilic Co center. Such increase in Co electrophilic character could
favoring the isomerization as consequence of the catalyst activity
increment, as observed with nickel catalysts [34]. For the cobalt cat-
alyst studied, only a slight increase in isomerization rate of olefins
was observed, suggesting that the product once formed remain on
organic phase preventing their isomerization in larger scale.
It is worth noting that the bis(imino)pyridine cobalt (II) cata-
lysts herein described, as far as we are concerned, show the best
balance of activity and selectivity in the formation of ␣-olefin for
a Ziegler–Natta oligomerization catalyst described in the literature
[3–7,9].
Gibson and co-workers showed that a cobalt catalyst, when sub-
stituted by a CF3 group in the ortho position of aromatic ring of
the ligand, could have its activity increased by the insertion of
a fluorine atom in the ligand structure in the R2 or R3 position
[30]. A similar trend in the activity of the oligomerization catalyst
bis(imino)pyridine cobalt was observed by Lutz and co-workers
[31]. These authors observed that the presence of a chlorine atom
and a CF3 group in the structure of the complex resulted in a more
active catalyst compared to the one that just had CF3 as a sub-
stituent. The cobalt complexes 5a, 6a and 7a activated by MAO
were shown to be the most active catalysts for the oligomerization
of ethylene when immobilized in ionic liquid BMI·AlCl4 and, among
these catalysts, the activity for oligomerization was dependent on
the electronegativity of the halogen present at R2 position.
4. Conclusions
Bis(imino)pyridine cobalt (II) catalysts activated by MAO were
active for ethylene oligomerization under biphasic (BMI·AlCl4)
conditions. The cobalt catalysts evaluated in this work showed
3.4. Influence of the substitution pattern on catalyst selectivity
activities for ethylene conversion in a range of 4000–15,300 h−1
.
The complexes of bis(imino)pyridine cobalt were selective for
the dimerization of ethylene to butenes. Is remarkable that the
selectivity for dimerization proved to be high regardless of the con-
ditions under which the oligomerization tests were performed or
the structure of the catalyst. These results are distinct from those
observed with these catalysts in homogeneous medium in which
they behave as catalysts for oligomerization with a Schulz–Flory
distribution of 0.63 for catalyst 2a [30] or oligo/polymerization
catalysts [17,30,31]. There are other cases where the selectivity of
the catalysts for oligomerization immobilized in ionic liquids was
significantly directed to a shorter product distribution [3,32].
It has been demonstrated that the oligomerization reac-
tion catalyzed by CoII bis(imino)pyridine complexes activated by
MAO shows a first-order dependence on monomer concentration
[18,27], which is consistent with chain growth via a classic inser-
tion mechanism. The chain end process for these cobalt catalysts
has been studied by NMR spectroscopy and revealed a first order
dependence of the reaction rate on the cobalt-alkyl species and a
zero order dependence on the monomer concentration [33], sug-
gesting an ␤-H elimination and formation of an intermediate metal
hydride species.
The most active catalysts were those containing electron with-
drawing groups CF3 and F, Cl or Br in their structure. All the
catalysts exhibited high selectivity for the dimerization of ethy-
lene into butenes in an ionic liquid, above 90%. The major product
was butene-1 for all the catalysts.
Acknowledgment
The authors thank CNPq for their financial support of this work.
References
[1] Y. Chauvin, B. Gilbert, I. Guibard, J. Chem. Soc. Chem. Commun. (1990) 1715.
[2] P. Wasserscheid, C.M. Gordon, C. Hilgers, M.J. Muldoon, I.R. Dunkin, Chem.
Commun. (2001) 1186–1187.
[3] V. Lecocq, H. Olivier-Bourbigou, Oil Gas Sci. Technol., Rev. IFP 62 (2007)
761–773.
[4] S. Einloft, F.K. Dietrich, R.F. de Souza, J. Dupont, Polyhedron 15 (1996)
3257–3259.
[5] K. Bernardo-Gusmão, L.F.T. Queiroz, R.F. de Souza, V. Leca, C. Loup, R. Réau, J.
Catal. 219 (2003) 59–63.
[6] L. Pei, X. Liu, H. Gao, Q. Wu, Appl. Organomet. Chem. 23 (2009) 455–459.
[7] K.-M. Song, H.-Y. Gao, F.-S. Liu, J. Pan, L.-H. Guo, S.-B. Zai, Q. Wu, Catal. Lett. 131
(2009) 566–573.
It is known that the solubility of ethylene is lower in ionic liq-
uids compared to organic solvents [32]. Based on experimental
observations of the independence of the monomer concentration
on the chain end process [33] and the low concentration of ethy-
lene in ionic liquid, the chain end is favored compared to chain
growth shifting to the formation of a shorter oligomer distribution.
This observation is also supported by the results with the catalyst
1a, comparing the selectivity of the reaction in ionic liquid with
the selectivity of the same catalyst in toluene, entries 1 and 14.
The reaction carried out in toluene showed and selectivity for the
ethylene dimerization of 63.7, with C10 formation. The reaction per-
formed in ionic liquid was very selective for ethylene dimerization
98.5%.
All catalysts studied showed a high selectivity for the forma-
tion of butene-1. Complex 1a showed selectivity for butene-1 at
84.1% of butenes. The presence of a methyl group at the R1 posi-
tion increased the selectivity for butene-1 to 87.0%. The catalyst 4a,
substituted with CF3 at R1 position and the catalysts 5a, 6a and 7a
substituted with CF3 at R1 position and F, Cl or Br at R2 position
showed a decrease in selectivity for butene-1 compared to catalyst
2a. Possibly, the decrease in the selectivity was consequence of the
low solubility of ethylene in the ionic liquid that might favor the iso-
merization of olefins under these conditions. Despite the decrease
in selectivity for these catalysts, it still remained high (above 80%)
for the most of the catalysts. Is worth to noting that the presence
of electron-withdrawing groups such as CF3, F, Cl, Br would reduce
[8] H. Olivier, P. Laurent-Gérot, J. Mol. Catal. A 148 (1999) 43–48.
[9] D. Thiele, R.F. de Souza, Catal. Lett. 138 (2010) 50–55.
[10] R.F. de Souza, B.C. Leal, M.O. de Souza, D. Thiele, J. Mol. Catal. A 272 (2007) 6–10.
[11] Y. Chauvin, H. Olivier, C.N. Wyrvalski, L.C. Simon, R.F. de Souza, J. Catal. 165
(1997) 275–278.
[12] R.F. de Souza, D. Thiele, A.L. Monteiro, J. Catal. 241 (2006) 232–234.
[13] D. Thiele, D. Thiele, R.F. de Souza, J. Mol. Catal. A 264 (2007) 293–298.
[14] Y. Chauvin, S. Einloft, H. Olivier, Ind. Eng. Chem. Res. 34 (1995) 1149–1155.
[15] B. Ellis, W. Keim, P. Wasserscheid, Chem. Commun. (1999) 337–338.
[16] F. Favre, A. Forestiere, F. Hugues, H. Olivier-Bourbigou, J.A. Chodorge, Oil Gas –
Eur. Mag. 31 (2005) 83–87.
[17] B.L. Small, M. Brookhart, A.M.A. Bennett, J. Am. Chem. Soc. 120 (1998)
4049–4050.
[18] B.L. Small, M. Brookhart, J. Am. Chem. Soc. 120 (1998) 7143–7144.
[19] 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.
[20] D. Zhu, P.H.M. Budzelaar, Organometallics 27 (2008) 2699–2705.
[21] A.R. Pray, in: T. Moeller (Ed.), Inorganic Syntheses, McGraw-Hill, New York,
1957, pp. 153–156.
[22] J.R. Kern, J. Inorg. Nucl. Chem. 24 (1962) 1105–1109.
[23] J.S. Wilkes, J.A. Levisky, R.A. Wilson, C.L. Hussey, Inorg. Chem. 21 (1982)
1263–1264.
[24] R. Schmidt, M.B. Welch, S.J. Palackal, H.G. Alt, J. Mol. Catal. A 179 (2002)
155–173.
[25] C. Qian, F. Gao, Y. Chen, L. Gao, Synlett 10 (2003) 1419–1422.
[26] S.S. Reddy, K. Radhakrishnan, S. Siravam, Polym. Bull. 36 (1996) 165–171.
[27] G.J.P. Britovsek, M. Bruce, V.C. Gibson, B.S. Kimberley, P.J. Maddox, S. Mas-
troianni, S.J. McTavish, C. Redshaw, G.A. Solan, S. Stro1mberg, A.J.P. White, D.J.
Williams, J. Am. Chem. Soc. 121 (1999) 8728–8740.
[28] 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. 6 (2000) 2221–2231.
[29] J.D. Azoulay, K. Itigaki, G. Wu, G.C. Bazan, Organometallics 27 (2008)
2273–2380.

88
D. Thiele, R.F. de Souza / Journal of Molecular Catalysis A: Chemical 340 (2011) 83–88
[30] K.P. Tellmann, V.C. Gibson, A.J.P. White, D.J. Williams, Organometallics 24
[32] P. Wasserscheid, C. Hilgers, W. Keim, J. Mol. Catal. A 214 (2004) 83–90.
[33] V.C. Gibson, K.P. Tellmann, M.J. Humphries, D.F. Wass, Chem. Commun. (2002)
2316–2317.
(2005) 280–286.
[31] F. Pelascini, F. Peruch, P.J. Lutz, M. Wesolek, J. Kress, Eur. Polym. J. 41 (2005)
1288–1295.
[34] H. Gao, Z. Ke, L. Pei, K. Song, Q. Wu, Polymer 48 (2007) 7249–7254.