Propylene bulk phase oligomerization with bisiminepyridine iron complexes in a calorimeter: Mechanistic investigation of 1,2 versus 2,1 propylene insertion

Article abstract of DOI:10.1016/S0022-328X(03)00670-3

Iron(II) complexes were synthesized with bisiminepyridine ligands of low steric demand. Activation with modified-methylaluminoxane (25 mol.% isobutyl groups) generated very active catalysts for propylene oligomerization. The oligomerizations were carried out in liquid propylene in a heat flow calorimeter. The oligomers were separated by preparative gas chromatography and the dimers and trimers analyzed using analytical gas chromatography, ¹ H-NMR- and ¹³C-NMR-spectroscopy. With knowledge of the dimer and trimer structure, we were able to establish a mechanistic pathway for propylene insertion and obtained knowledge about the iron alkyl species involved. Analysis of the various dimers formed allowed us to determine the percentage of 1,2 versus 2,1 propylene insertions. Considering the same iron alkyl species with ligands of different steric demand, a change in the probabilities for 1,2 versus 2,1 propylene insertions can be observed. With this knowledge, the catalyst behavior for ligands of varying steric demand can be predicted.

Full text of DOI:10.1016/S0022-328X(03)00670-3

Journal of Organometallic Chemistry 683 (2003) 209Á219
/
www.elsevier.com/locate/jorganchem
Propylene bulk phase oligomerization with bisiminepyridine iron
complexes in a calorimeter: mechanistic investigation of 1,2 versus 2,1
propylene insertion
Sebastian Thomas Babik, Gerhard Fink *
Max-Planck-Institut fur Kohlenforschung, Kaiser-Wilhelm-Platz 1, 45470 Mulheim an der Ruhr, Germany
¨
Received 27 March 2003; accepted 28 June 2003
Dedicated to Prof. M.T. Reetz on the occasion of his 60th birthday
Abstract
Iron(II) complexes were synthesized with bisiminepyridine ligands of low steric demand. Activation with modified-
methylaluminoxane (25 mol.% isobutyl groups) generated very active catalysts for propylene oligomerization. The oligomerizations
were carried out in liquid propylene in a heat flow calorimeter. The oligomers were separated by preparative gas chromatography
and the dimers and trimers analyzed using analytical gas chromatography, 1H-NMR- and 13C-NMR-spectroscopy. With knowledge
of the dimer and trimer structure, we were able to establish a mechanistic pathway for propylene insertion and obtained knowledge
about the iron alkyl species involved. Analysis of the various dimers formed allowed us to determine the percentage of 1,2 versus 2,1
propylene insertions. Considering the same iron alkyl species with ligands of different steric demand, a change in the probabilities
for 1,2 versus 2,1 propylene insertions can be observed. With this knowledge, the catalyst behavior for ligands of varying steric
demand can be predicted.
# 2003 Elsevier B.V. All rights reserved.
Keywords: Bisiminepyridine iron complex; Sterical demand of the ligand system; Propylene oligomerization; Propylene dimers; 1,2 and 2,1 propylene
insertion
1. Introduction
propagation and termination steps of the polymeriza-
tion reaction. If we have more information about these
steps we know in which direction we have to adjust for
instance the bulkiness of a substituent, the strength of an
electronic withdrawing group or the nature of the co-
catalyst. Our group has made mechanistic investigations
in metallocene catalyzed propylene polymerizations and
One goal of our research group is to learn more about
the mechanisms of polymerization reactions, which can
lead us to control these reactions by tuning the steric
and electronic properties of the reactive center. There
persists great industrial interest in producing polymeric
materials that satisfy the consumer needs exactly. The
aim of our research is to produce these ‘tailor-made-
polymers’. In general it is possible to adjust the polymer
properties by changing the ligand structure or varying
the reaction conditions, temperature, monomer pressure
norborneneÁethylene copolymerizations [5,6].
/
Recently there has been growing academic and
industrial interest in polymerization catalysts based on
bisiminepyridine complexes with iron. This type of
catalysts was developed and investigated by Brookhart
and co-workers and Gibson and co-workers in the late
1990s [7,8]. It is known that they are very active in
ethylene and propylene polymerization with activation
by methylaluminoxane, modified-methylaluminoxane
and co-catalyst ratio [1Á4]. It is an indispensable
precondition to know as much as possible about the
/
(MMAO) or Ph3C[B(C6F5)4] and AlR3 [9Á16,21]. We
recently published a paper describing a catalytic cycle
that involves an iron hydride species, which is the result
/
* Corresponding author. Tel.: ꢀ
306-2980.
E-mail address: fink@mpi-muelheim.mpg.de (G. Fink).
/
49-208-306-2240; fax: ꢀ49-208-
/
0022-328X/03/$ - see front matter # 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0022-328X(03)00670-3

210
S.T. Babik, G. Fink / Journal of Organometallic Chemistry 683 (2003) 209Á219
/
of a b-H-elimination [17]. This hydride is in our opinion
the result of a b-H-elimination. There is also the
possibility of a transfer reaction to the monomer as
discussed in the theoretical works by Ziegler and
Morokuma [22,23]. More experiments are necessary to
determine whether the termination step is a transfer
reaction or a b-H-elimination.
2.3. Synthesis of 2,6-bis[1-(phenylimino)ethyl]-pyridine
(L1) [18]
2,6-Diacetylpyridine (1.14 g, 7.0 mmol) was dissolved
in 15 ml of methanol in a 50 ml round-bottom flask.
Aniline (1.86 g, 20.0 mmol) was added. Four drops of
97% formic acid were added and the clear, orange
solution was allowed to stir in the sealed flask at
ambient temperature for 5 h. After stirring the resultant
pale yellow solid precipitate was collected by filtration,
washed with cold methanol and dried. The yield was
1.86 g (85%) of pure ligand.
To further analyze the mechanism of propylene
polymerization with bisiminepyridine iron complexes,
it would be ideal to isolate an iron alkyl species to
further understand the nature of the inserting and
propagating iron alkyl species. The problem being that
it has not been possible to isolate such a species. Thus,
we decided to regard this problem from another
perspective and isolate dimers and trimers of propylene
produced by less sterically demanding iron bisiminepyr-
idine complexes. After identifying the dimer or trimer it
is possible to extrapolate the nature of the previous iron
alkyl from which the dimer or trimer was formed.
After producing polymer when using bulky catalysts
it was surprising to find eight different dimers and at
least seven different trimers, when using a less sterically
demanding catalyst. This indicates a very complicated
system of insertion and subsequent eliminations. For
example, some iron alkyls prefer to insert propylene in
1,2 fashion, some in a 2,1 fashion, some do not insert
propylene at all. We also found a high fraction of 2-
olefins which show E/Z isomerization. In this manu-
script we report the product distribution of three
different iron bisiminepyridine catalysts and present
our conclusions concerning the nature of the iron alkyl
species involved and the influence of the steric properties
of the ligands on the reactivity of those alkyl species.
1H-NMR (CDCl3): dꢁ
7.78Á7.83 (t, 1, py-Hp); 7.28Á
7.08 (t, 2, Haryl); 6.76Á6.79 (d, 4, Haryl); 2.34 (s, 6, NÄ
CCH3).
/
8.23Á/8.26 (d, 2, py-Hm);
/
/
7.34 (t, 4, Haryl); 7.02Á
/
/
/
2.4. Synthesis of 2,6-bis[1-(2,4-
dimethylphenylimino)ethyl]-pyridine (L2) [8,13]
2,6-Diacetylpyridine (1.25 g, 7.7 mmol) was dissolved
in 20 ml of methanol in a 50 ml round-bottom flask. 2,4-
Dimethylaniline (2.88 g, 22.0 mmol) was added. Five
drops of 97% formic acid were added and the clear,
orange solution was allowed to stir in the sealed flask at
ambient temperature for 15 h. After stirring overnight
the resultant yellow solid precipitate was collected by
filtration, washed with cold methanol and dried. The
yield was 2.04 g (72%) of pure ligand.
1H-NMR (CDCl3): dꢁ
7.81Á7.86 (t, 1, py-Hp); 6.92Á
6.79 (m, 2, Haryl); 6.49Á
/
8.30Á
/
8.33 (d, 2, py-Hm);
/
/
6.98 (m, 2, Haryl); 6.78Á
/
/
6.53 (d, 2, Haryl); 2.26 (s, 6, NÄ
/
CCH3); 2.14 (s, 6, arylCH3); 2.02 (s, 6, arylCH3).
2.5. Synthesis of 2,6-bis[1-(2-ethylphenylimino)ethyl]-
pyridine (L3)
2. Experimental
2,6-Diacetylpyridine (1.19 g, 7.3 mmol) was dissolved
in 15 ml of methanol in a 50 ml round-bottom flask. 2-
Ethylaniline (2.49 g, 19.0 mmol) was added. Four drops
of 97% formic acid were added and the clear, orange
solution was allowed to stir in the sealed flask at room
temperature (r.t.) for 15 h. After stirring overnight the
resultant pale yellow solid precipitate was collected by
filtration, washed with cold methanol and dried. The
yield was 1.86 g (69%) of pure ligand.
2.1. General considerations
The handling of water- and air-sensitive compounds
was performed under an argon atmosphere using
Schlenk techniques.
2.2. Materials
1H-NMR (CDCl3): dꢁ
7.86Á7.93 (t, 1, py-Hp); 7.18Á
7.11 (m, 2, Haryl); 6.66Á6.69 (d, 2, Haryl); 2.48Á
(quart., 4, arylCH2Me); 2.37 (s, 6, NÄCCH3); 1.12Á
(t, 6, arylCH2CH3).
/
8.39Á8.42 (d, 2, py-Hm);
/
Methanol was dried over CaH2/Mg and distilled.
Toluene was distilled from sodium. THF was distilled
from MgH2. 2,6-Diacetylpyridine, aniline, 2,4-dimethy-
laniline, 2-ethylaniline, 97% formic acid and FeCl2 were
purchased from Aldrich and used without further
purification. MMAO (7 wt.% solution in toluene, 25%
isobutyl groups) was purchased from Texas Alkyls Inc.
Propylene (99.5%) was purchased from Messer-Grie-
sheim and used without further purification.
/
/
7.27 (m, 4, Haryl); 7.06Á
/
2.56
1.18
/
/
/
/
2.6. Synthesis of the iron complexes (C1Ã
/
C3) [7Á16]
/
Dry FeCl2 (one equivalent) in 20 ml dry THF was
stirred under an argon atmosphere in a 100 ml flame-

S.T. Babik, G. Fink / Journal of Organometallic Chemistry 683 (2003) 209Á
/219
211
dried two-neck-flask. A solution of Ligand L (1.05
equivalent) in 25 ml dry THF was added slowly at r.t.
via a dropping funnel. The brown suspension of FeCl2
turned immediately into a blue color. The mixture was
then stirred under argon for one additional hour. Then
pentane was added to the blue suspension and the solid
was filtered and dried under argon. The light-blue
complex was isolated in near quantitative yield. The
complexes were analyzed using ESI mass spectroscopy,
where the molecular peak of the iron complex and the
molecular peak of the corresponding ligand were the
only peaks detected.
use of multidimensional separation in coupled columns
with techniques such as ‘peak cutting’ and ‘back
flushing’ enables this instruments to separate larger
amounts of purified substances.
1H-NMRs (300 MHz) and 13C-NMRs (75.5 MHz)
were carried out using a Bruker DPX-300 spectrometer
and CDCl3 as solvent.
3. Results and discussion
Fig. 1 shows the catalysts synthesized in this work.
The main focus of this work will be on the product
distribution of the oligomers in order to draw conclu-
sions about the iron alkyls that are taking part in the
reaction. Fig. 2 is a GC plot of a typical oligomeric
mixture obtained with catalyst 1. In this figure, the
region of the dimers and trimers has been enlarged and
will be the focus henceforth.
2.7. Oligomerization process
All polymerizations were carried out in a 1.8 l high-
pressure steel reactor using the Mettler RC1 reaction
calorimeter [19,20]. The reactions were performed in the
isothermal mode using an anchor stirrer (250 rpm)
under the vapor pressure of liquid propylene at the
corresponding temperature. First 1 ml of the MMAO
solution was added via syringe to the reactor as a
scavenger. One liter of propylene was added to the
reactor by condensation, measured by the loss in the
propylene supply gas bottle. The reactor was then
heated to the reaction temperature. A modification to
the standard Mettler RCI is a separate thermostatically
controlled bath used to circulate maintain the reactor
head 2 8C above the reaction temperature to avoid
condensation and to minimize heat losses. The catalyst
in a suspension of 5 ml anhydrous toluene was injected
into the reactor immediately followed by another
solution of 0.5 ml of MMAO in 4.5 ml of toluene using
argon pressure. The injection system has been thermo-
stated to the temperature of the reactor by means of the
thermostatically controlled bath in order to minimize an
initial temperature disturbance. The polymerizations
were stopped by injection of water. Throughout the
entire reaction, temperatures, pressure and heat was
recorded. With this data, it is possible to determine a
kinetic profile for the reaction. After each experiment
the remaining propylene is vented. The organic phase,
which consists usually of more than 95% oligomers and
5% of toluene is separated from the aqueous phase
containing catalyst and the decomposed MMAO. The
organic phase is diluted with toluene and analyzed
directly using GC and preparative GC.
The identification of di- and trimers was performed
1
with H- and 13C-NMR after separation using prepara-
tive gas chromatographically. In some cases it was not
possible to separate the substances using this method. In
the enlarged analytical plot (Fig. 2) two or even three
substances are assigned to a single peak such as peak
26Á28 in the C9-fraction. However, it was possible to
/
identify the different olefinic groups because the 1-olefin
and the 2-olefin protons have distinct signals (Fig. 3).
It was also possible to identify E- and Z-isomers. It
was possible to compare synthesized olefins using
reference olefins with known configuration. As shown
in Fig. 4, it is possible to differentiate between E- and Z-
1
isomers using H-NMR. The Z-isomers show slightly
broader signals in the olefinic region and the shape and
chemical shift of the ‘D-group’ differs. The E-isomer is
slightly below 2 ppm and the Z-isomer above 2 ppm,
making it possible to distinguish between the two
isomers easily.
Once the dimers and trimers have been correctly
identified, it is possible to proceed with the mechanistic
considerations. Fig. 5 shows the isomer distribution of
the C6-fraction at three different temperatures produced
with catalyst 1.
With increasing temperature, there is an increase in
the percentage of dimer 1 (4-methyl-1-pentene) while the
amount of dimer 4 (E-4-methyl-2-pentene) is decreasing.
On the other hand, the percentage of dimer 3 (Z-4-
methyl-2-pentene) is increasing only slightly. The in-
creasing temperature favors the formation of the 1-
olefin and the Z-isomer of the 2-olefin. Furthermore,
the temperature has almost no effect on the distribution
2.8. Oligomer separation and characterization
The oligomers were analyzed using a GC with a Carlo
Erba 4100 equipped with a 60 m RTX-1 column (0.25
mm) and a FID. Identification of the dimers and trimers
necessitated isolation of each isomer. This was per-
formed using multidimensional preparative gas chroma-
tography, based on a Shimadzu Tandem-GC 14A. The
of the other dimers. The percentage of the dimers 1Á
/8 is
constant over the studied temperature range.
Scheme 1 illustrates all the olefins and corresponding
iron alkyls that occur in this reaction. Percentage near
the arrows indicates whether a 1,2 or 2,1 propylene

212
S.T. Babik, G. Fink / Journal of Organometallic Chemistry 683 (2003) 209Á
/219
Fig. 1. Complexes in study.
insertion occurs. The given numbers only relate to the
dimer distribution and neglect that some species are still
involved in the growth to trimers and so on.
incoming propylene: the 1,2- or the 2,1-insertion.
Analyzing the formed dimers, it is possible to conclude
that there is a 50 to 50 probability for a propylene
insertion via a 1,2 or a 2,1 step.
The three latter dimers 1, 3 and 4, result from the
same iron alkyl, shown in Scheme 1 in the bottom right.
From this position there are two possible directions for
the chain termination step. One towards the short
branch which forms the 1-olefin (1) and the other
direction towards the longer branch which form the
two 2-olefins (2 and 3). In the case of the dimers 6 (1-
hexene), 7 (E-2-hexene) and 8 (Z-2-hexene), one iron
alkyl leads to all three dimers. Concerning dimers 2 (2,3-
dimethyl-1-butene) and 5 (2-methyl-1-pentene) the first
species is a primary iron alkyl where only one direction
is possible for the termination, resulting in a vinylidene
double bond.
In the middle of Scheme 1 is the proposed starting
hydride species [17]. From this species there are two
possibilities for a propylene insertion. Moving left in
Scheme 1 (white arrow), the propylene inserts in a 1,2-
arrangement and forms the iron-n-propyl species. From
this species there are again two possible routes for an
A different situation is found when following the
black arrow to the right of the hydride and starting with
a 2,1 propylene insertion. The formed iron isopropyl
species favors the subsequent 2,1-insertion by more than
90% about the 1,2-insertion. The iron alkyl formed after
two 2,1 propylene insertions is the dominant species for
the further growth to tri- and higher oligomers. This is
confirmed by the distribution of the C9 fraction seen in
Fig. 2. The amount of substituted trimers is higher than
the more linear trimers. We can say it is more likely that
the propylene inserts in a 2,1- than in a 1,2-arrangement
both in the iron hydride and the iron alkyl when we use
the least sterically demanding catalyst 1.
What happens when the sterical demand of the system
is increased by using catalyst 2? Here there is a methyl
group in the ortho position of the phenylimin. In Fig. 6
is the distribution of the C6 fraction produced with
catalyst 2 at different temperatures.
Fig. 2. Propylene oligomer distribution and identified dimers and trimers; catalyst 1.

S.T. Babik, G. Fink / Journal of Organometallic Chemistry 683 (2003) 209Á
/219
213
Fig. 3. Comparison of olefinic end groups, 1H-NMR of 1- and 2-olefinic groups.
Catalyst 2 gives a completely different dimer distribu-
tion compared to catalyst 1. In this case, hexenes 6, 7
and 8 are the most populated dimers and no longer the
methyl pentenes. Catalyst 2 also shows different tem-
perature behavior, increasing temperature results in an
increase in the amount of the dimers 1, 3 and 4 formed.
When catalyst 1 was used, there was also an increase in
the dimers 1 and 3, but a decrease in dimer 4. This
difference in the termination behavior must relate to the
slightly higher sterical demand in the ligand sphere of
catalyst 2. The distribution of the linear hexenes 6, 7 and
8 shows that the amount of 1-hexene (6) remains
constant while the amount of dimer 7 (E-2-hexene)
and 8 (Z-2-hexene) is decreasing. Dimer 2 and 5
distributions are not effected changing the temperature
from 20 to 40 8C.
combined sterical demand of the ligand system and the
growing chain at the iron center, causing the increase in
the percentage of 2,1 insertions. In the case of the less
sterically demanding catalyst 1, the n-propyl group after
the first 1,2 propylene insertion can arrange in such a
way that there is no preference for either a 1,2- or a 2,1-
insertion. The two methyl groups of the iron isobutyl
species have a higher sterical demand than the linear n-
propyl group, causing a more hindered iron center. The
higher sterically demanding catalyst 2 prevents the
rearrangement of the linear n-propyl group and the
sterical demand nears this of the isopropyl group. With
catalyst 2, there seems to be no difference for the
incoming propylene between an iron-n-propyl and an
iron-isopropyl group.
With a higher sterical demand in the ligand system
there is an increasing preference for the first 1,2
propylene insertion and an increasing tendency for the
second 2,1-insertion is observed. The next system
investigated was the sterically most demanding, a
ortho-ethyl substituted catalyst 3. Fig. 7 shows the
distribution of the C6 fraction obtained with this
catalyst.
Catalyst 3 follows the same general trend in dimer
distribution observed with catalyst 1 and 2. With higher
sterical demand of the ligand system, more hexenes (6, 7
and 8) and fewer substituted pentenes (1, 3 and 4) are
formed. Dimers 2 and 5 were not detected at all. The
trend in oligomer distribution with changing tempera-
ture is similar to catalyst 2.
Scheme 3 is the mechanistic pathway for catalyst 3.
Beginning with the iron hydride species, almost the
same probabilities for a 1,2- and a 2,1-arrangement for
the initial propylene insertion are observed as with
catalyst 2. The biggest difference between the two
catalysts is found in the second step. In the case of
In Scheme 2, the mechanistic pathway for catalyst 2 is
shown. Comparing the behavior of catalyst 2 and
catalyst 1 there is an interesting shift in the mechanism.
The cycle begins again with the FeÃ
/
H species in the
middle of Scheme 2. The higher sterical demand of the
ligand system leads to a 69:31 favored 1,2 propylene
insertion (white arrow) in the iron hydride instead of a
88:12 favored 2,1 insertion in the case of catalyst 1.
Concerning the first insertion of propylene into the FeÃ
/
H bond the reason for this change must be the sterical
effect of the ligand system, because this is the only
difference between the two catalysts. Considering the
second propylene insertion after a previous 2,1 insertion
(black arrow) the same ratio is observed as with catalyst
1. However, upon the second insertion after a 1,2-
insertion (white arrow) there is a great change in the 2,1
insertion in the iron-n-propyl species from a 68:32 ratio
in the case of catalyst 1 to a 91:9 ratio for catalyst 2. It is
quite interesting that the ratio is almost the same as 93:7
ratio after the first 2,1 insertion. There must be a

214
S.T. Babik, G. Fink / Journal of Organometallic Chemistry 683 (2003) 209Á219
/
Fig. 4. Comparison of olefinic end groups, 1H-NMR of E- and Z-isomer.
catalyst 2, there is a 91:9 favored 2,1-arrangement for
the second insertion compared with a 100:0 ratio using
catalyst 3.
favored with an increasing sterical demand shown by
comparing the ratios in Fig. 5 for catalyst 1, Fig. 6 for
catalyst 2 and Fig. 7 for catalyst 3.
With increasing temperature, there is a great increase
in the production of dimer 1 and a slight increase for
dimers 3 and 4. Considering the hexene product
distribution, we observe a dramatic decrease in the
amount of the 2-hexenes (7 and 8) and an increase in the
amount of 1-hexene (dimer 6). Catalyst 1 formed only
2% 1-hexene, catalyst 2 had a constant ratio of about
10% and with catalyst 3 the amount of 1-hexene
increased up to 20%. In the case of the pentenes, the
formation of the terminal vinyl group is more and more
When comparing the three catalysts there is clearly a
different behavior as shown in Fig. 8. With a slightly
higher sterical demand in ortho position from a hydro-
gen to methyl, the dimer distribution dramatically
changes. The change from methyl to ethyl on the ligand
system does not change the distribution dramatically,
but increases the ratio of the 1-olefins.
Considering an even more sterically demanding
ligand, one would expect the first 1,2 propylene insertion
into the iron hydride to be followed by several 2,1

S.T. Babik, G. Fink / Journal of Organometallic Chemistry 683 (2003) 209Á
/219
215
Fig. 5. Propylene dimer distribution of catalyst 1, propylene bulk phase oligomerization in the Mettler RC1 calorimeter, conditions: [Fe]ꢁ
/
4ꢂ
/
10ꢃ5
mol lꢃ1; 1.5 ml MMAO: [Al]ꢁ 10ꢃ3 mol lꢃ1; liquid propylene 99.5%; 120 min.
/
3.5ꢂ
/
Scheme 1. Mechanism of propylene dimerization with catalyst 1; propylene bulk phase oligomerization in the Mettler RC1 calorimeter, conditions:
[Fe]ꢁ4ꢂ 3.5ꢂ
10ꢃ5 mol lꢃ1; 1.5 ml MMAO: [Al]ꢁ 10ꢃ3 mol lꢃ1; liquid propylene 99.5%; 120 min.
/
/
/
/

216
S.T. Babik, G. Fink / Journal of Organometallic Chemistry 683 (2003) 209Á219
/
Fig. 6. Propylene dimer distribution of catalyst 2, propylene bulk phase oligomerization in the Mettler RC1 calorimeter, conditions: [Fe]ꢁ
/
4ꢂ
10ꢃ5
/
mol lꢃ1; 1.5 ml MMAO: [Al]ꢁ 10ꢃ3 mol lꢃ1; liquid propylene 99.5%; 120 min.
/
3.5ꢂ
/
Scheme 2. Mechanism of propylene dimerization with catalyst 2; propylene bulk phase oligomerization in the Mettler RC1 calorimeter, conditions:
[Fe]ꢁ4ꢂ 3.5ꢂ
10ꢃ5 mol lꢃ1; 1.5 ml MMAO: [Al]ꢁ 10ꢃ3 mol lꢃ1; liquid propylene 99.5%; 120 min.
/
/
/
/

S.T. Babik, G. Fink / Journal of Organometallic Chemistry 683 (2003) 209Á
/219
217
Fig. 7. Propylene dimer distribution of catalyst 3, propylene bulk phase oligomerization in the Mettler RC1 calorimeter, conditions: [Fe]ꢁ
/
4ꢂ
/
10ꢃ5
mol lꢃ1; 1.5 ml MMAO: [Al]ꢁ 10ꢃ3 mol lꢃ1; liquid propylene 99.5%; 120 min.
/
3.5ꢂ
/
Scheme 3. Mechanism of propylene dimerization with catalyst 3; propylene bulk phase oligomerization in the Mettler RC1 calorimeter, conditions:
[Fe]ꢁ4ꢂ 3.5ꢂ
10ꢃ5 mol lꢃ1; 1.5 ml MMAO: [Al]ꢁ 10ꢃ3 mol lꢃ1; liquid propylene 99.5%; 120 min.
/
/
/
/

218
S.T. Babik, G. Fink / Journal of Organometallic Chemistry 683 (2003) 209Á219
/
Fig. 8. Propylene dimer distribution with different catalysts 1Á
/
3; propylene bulk phase oligomerization in the Mettler RC1 calorimeter, conditions:
30 8C, [Fe]ꢁ4ꢂ 3.5ꢂ
/
/
10ꢃ5 mol lꢃ1; 1.5 ml MMAO: [Al]ꢁ 10ꢃ3 mol lꢃ1; liquid propylene 99.5%; 120 min.
/
/
insertions. The terminating step should result in a
terminal double bond. This is exactly what is observed
when a 2-methyl-6-isopropyl substituted catalyst is used
[9,17].
arrangement leads to the same but larger alkyl species
and the chain is growing. Insertion in a 1,2-arrangement
leads to the b-methyl alkyl species which is a dead end
and stops the chain growth.
Thus far we have concentrated exclusively on the C6
fraction. The percentages for the 1,2 insertion will be
much lower, when considering trimers and tetramers.
The 2,1 insertions are still involved in the propagation
process. Separation of pure trimers by preparative GC
was not possible; however, it was possible to identify
several trimers with H-NMR. Scheme 4 is a mechan-
istic scheme that includes the trimers and explains which
iron alkyls are necessarily to produce the trimers.
4. Conclusion
Our approach to the clarification of mechanism was
the isolation of oligomers and the identification of the
obtained products. Knowing the exact product struc-
ture, it is possible to make reasonable approximations
on the structure of involved iron alkyl species in the
catalytic cycle without identifying the exact alkyl
species. With this knowledge we were able to establish
a mechanistic scheme for the propylene oligomerization
with the bisiminepyridine iron system.
The question of how we produce oligomers versus
polymers is a question of knowing how to control the
ratio of the 1,2 and 2,1 insertion. One method is to alter
the sterical demand in the ortho position of the ligand.
The more bulky the ligand the more often there happens
a 2,1 propylene insertion and therefore higher molecular
mass of the oligomer, i.e. polymer. Another important
observation is that with a higher sterical demand of the
catalysts the formation of the a-olefin is favored.
Mechanistic studies contribute to our understanding of
polymer reactions and allow greater control of polymer
properties, i.e. tailor-made-polymers.
1
In principle we are dealing here with two different
iron alkyl species. One is formed after a 1,2 propylene
insertion (white arrow) and forms an iron alkyl species
with a b-methyl group on the alkyl chain, seen in the
upper half in Scheme 4. Because we did not find any of
the possible oligomers that might be produced after a
propylene insertion in such an iron alkyl, we propose
that the elimination is the only possible pathway for
these species. The result is the formation of the
vinylidene olefin which was isolated with preparative
1
gas chromatography and identified using H-NMR.
The other possible iron alkyl species is formed after a
2,1 propylene insertion (black arrow) and results in an
iron a-methyl alkyl species. These alkyl species are
predominantly responsible for the oligomer or polymer
growth. Inserting another propylene molecule in a 2,1-

S.T. Babik, G. Fink / Journal of Organometallic Chemistry 683 (2003) 209Á
/219
219
Scheme 4. Mechanism of propylene oligomerization with bisiminepyridine iron complexes in liquid propylene.
[7] B.L. Small, M. Brookhart, A.M.A. Bennett, J. Am. Chem. Soc.
120 (1998) 4049.
5. Outlook
[8] 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.
Work that has started but is not completed so far is a
combined molecular mechanic/quantum mechanic cal-
culation for several iron alkyls and their energy for the
1,2 or 2,1 propylene insertion. We intend to calculate
and visualize the incoming propylene at the iron center.
This shall give us an idea how the sterical demand of the
ligand effects the processes near the iron center. A
visualization of the ligand movement will also give us an
idea of how the trajectory of the incoming propylene
molecule is hindered by the ortho substituents. Another
important investigation deals with the nature of the
coordination sphere around the iron. We want to know
which isomerization steps happen and what the differ-
ence in the energy is to switch for example from a
growing to a terminating configuration.
[9] B.L. Small, M. Brookhart, Macromolecules 32 (1999) 2120.
[10] A.M.A. Bennett, Chemtech (1999) 24.
[11] G.J.P. Britovsek, M. Bruce, V.C. Gibson, B.S. Kimberley, P.J.
Maddox, S. Mastroianni, S.J. McTavish, C. Redshaw, G.A.
Solan, S. Stromberg, A.J.P. White, D.J. Williams, J. Am. Chem.
Soc. 121 (1999) 8728.
[12] B.L. Small, M. Brookhart, J. Am. Chem. Soc. 120 (1998) 7143.
[13] 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.
[14] E.J.M. deBoer, H.H. Deuling, H. van der Heijden, N. Meijboom,
A.B. van Oort, A. van Zon, Shell Internationale Research
Maatschappij B.V., The Netherlands, PCT Int. Appl. WO
0158874 (2001); Chem. Abstr. 135 (2001) 181084.
[15] E.P. Talzi, D.E. Babushkin, N.V. Semikolenova, V.N. Zudin,
V.A. ZakharovÁ, Kinet. Catal. 42 (2001) 147.
/
[16] G.J.P. Britovsek, V.C. Gibson, S.K. Spitzmesser, K.P. Tellmann,
A.J.P. White, D.J. Williams, J. Chem. Soc. Dalton Trans. (2002)
1159.
[17] S.T. Babik, G. Fink, J. Mol. Catal. 188 (2002) 245.
[18] Alyea, Merrell, Synth. Inorg. Met. Org. Chem. 4 (1974) 535.
[19] F. Korber, K. Hauschild, G. Fink, Macromol. Chem. Phys. 202
(2001) 3329.
References
[1] R.F. Jordan, Adv. Organomet. Chem. 32 (1991) 325.
[2] H.-H. Brintzinger, D. Fischer, R. Mulhaupt, B. Rieger, R.
¨
[20] F. Korber, K. Hauschild, M. Winter, G. Fink, Macromol. Chem.
Phys. 202 (2001) 3323.
Waymouth, Angew. Chem. 107 (1995) 1255.
[3] K. Soga, T. Shiono, Prog. Polym. Sci. 22 (1997) 1503.
[4] G. Fink, Y. van der Leek, K. Angermund, M. Reffke, R.
Kleinschmidt, R. Goretzki, Chem. Eur. J. 3 (1997) 585.
[5] R. Kleinschmidt, Yu. Griebenow, G. Fink, J. Mol. Catal. A:
Chem. 157 (2000) 83.
[21] Y. Chen, C. Qian, J. Sun, Organometallics 22 (2002) 1231.
[22] L.Q. Deng, P. Margl, T. Ziegler, J. Am. Chem. Soc. 121 (1999)
6479.
[23] D.V. Khoroshen, D.G. Musaev, T. Vreven, K. Morokuma,
Organometallics 20 (2001) 2007.
[6] A.u. Wendt, G. Fink, Macromol. Chem. Phys. 202 (2001) 3490.