Modified pyridine-bis(imine) iron and cobalt complexes: Synthesis, structure, and ethylene polymerization study

Article abstract of DOI:10.1002/ejic.200600239

In this paper, we describe the synthesis of two new pyridine-bis(imine)s {4-chloro-2,6-bis[1-(2,6-dusopropylphenylimino)-ethyl]pyridine and 2,6-bis[1-(2,6-dimethylcyclohexylimino)-ethyl]pyridine) and their complexation with iron and cobalt dichloride. The solid-state structure of the iron complexes was solved and found to be very close to catalysts already described by Brookhart and Gibson. Their ability to polymerize ethylene was investigated after activation with MAO. It was thus shown that the substitution of the pyridine ring of the ligand is unfavorable: their catalytic activity is rather low. The replacement of the aromatic rings on the imine functions by dimethylcyclohexyl rings resulted in a complete loss of activity of the iron complex for oligomerization and polymerization of ethylene. With the cobalt analog, polymerization of ethylene could be achieved under the same conditions. Wiley-VCH Verlag GmbH & Co. KGaA, 2006.

Full text of DOI:10.1002/ejic.200600239

FULL PAPER
DOI: 10.1002/ejic.200600239
Modified Pyridine-Bis(imine) Iron and Cobalt Complexes: Synthesis, Structure,
and Ethylene Polymerization Study
Frédéric Pelascini,[a,b] Marcel Wesolek,*[b] Frédéric Peruch,*[a] and Pierre J. Lutz[a]
Keywords: N ligands / Iron / Cobalt / Ethylene / Polymerization
In this paper, we describe the synthesis of two new pyridine-
bis(imine)s {4-chloro-2,6-bis[1-(2,6-diisopropylphenylimino)-
ethyl]pyridine and 2,6-bis[1-(2,6-dimethylcyclohexylimino)-
ethyl]pyridine} and their complexation with iron and cobalt
dichloride. The solid-state structure of the iron complexes
was solved and found to be very close to catalysts already
described by Brookhart and Gibson. Their ability to polymer-
ize ethylene was investigated after activation with MAO. It
was thus shown that the substitution of the pyridine ring of
the ligand is unfavorable: their catalytic activity is rather low.
The replacement of the aromatic rings on the imine functions
by dimethylcyclohexyl rings resulted in a complete loss of
activity of the iron complex for oligomerization and polymeri-
zation of ethylene. With the cobalt analog, polymerization of
ethylene could be achieved under the same conditions.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2006)
methyl), was performed. The iron and cobalt complexes,
once synthesized, were activated with methylaluminoxane
(MAO) and tested for the polymerization of ethylene.
Introduction
In 1998, Brookhart and Gibson simultaneously described
the synthesis of very active pyridine-bis(imine) iron- or co-
balt-based catalysts for the oligomerization or the polymeri-
zation of ethylene.[1–3] When the substituents on the ligand
were bulky enough, they were found to be extremely ef-
ficient for the synthesis of linear polyethylenes. Many modi-
fications of these ligands have already been described, but
it was almost always the R1–R5 (Figure 1) groups that were
modified; some reviews are listed with the most recent of
these references.[4–17] So far, less work has been done on
other types of modifications (imine functions replaced by
amine functions, aryl rings replaced by amines or pyrrolyl
rings, pyridine ring replaced by thiophene or furan ring,
and so forth).[18–23] In this paper, we report new kinds of
modifications. We focus more specifically on the electronic
factors through the synthesis of two new types of pyridine-
bis(imine) ligands. With ligand 1, the presence of the chlo-
rine group in the para position of the pyridine ring in
the well-known 2,6-bis[1-(2,6-diisopropylphenylimino)ethyl]-
pyridine ligand (R1 = R5 = isopropyl) led us to study the
effect of the electronic-withdrawing chlorine group. For
ligand 2, the replacement of the aryl groups by aliphatic
Figure 1. Pyridine-bis(imine) iron or cobalt complexes.
Results and Discussion
Synthesis and Characterization of Ligands
4-Chloro-2,6-bis[1-(2,6-diisopropylphenylimino)ethyl]pyr-
idine (1) was synthesized according to already published
procedures for the synthesis of pyridine-bis(imine).[1,2] The
reaction is a condensation between 2,6-diacetyl-4-chloro-
pyridine (synthesis described by Constable[24]) and 2,6-di-
isopropylaniline (Scheme 1, i). The yellow product was iso-
rings
of
another
well-known
ligand,
2,6-bis-
1
lated and characterized by elemental analysis, and H and
[1-(2,6-dimethylphenylimino)ethyl]pyridine (R1 = R5
=
13C NMR spectroscopy.
Concerning the synthesis of 2,6-bis[1-(2,6-dimethylcyclo-
hexylimino)ethyl]pyridine (2), the same procedure was fol-
lowed but the substituted aniline was replaced by the corre-
sponding cyclohexylamine to yield aliphatic-substituted im-
ines (Scheme 1, ii). Besides, substituted cycloalkanes exhibit
several stereoisomers. In the case of 2,6-dimethylcyclohex-
ylamine, three different stereoisomers can be identified: the
[a] Institut Charles Sadron (UPR CNRS 22),
6 rue Boussingault, B. P. 40016, 67083 Strasbourg Cedex,
France
[b] Laboratoire de Chimie des Métaux de Transition et de Catalyse
(LC3-UMR CNRS 7177), Institut Le Bel,
4 rue Blaise Pascal, 67070 Strasbourg Cedex, France
E-mail: peruch@enscpb.fr
wesolek@chimie.u-strasbg.fr
Eur. J. Inorg. Chem. 2006, 4309–4316
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4309

F. Pelascini, M. Wesolek, F. Peruch, P. J. Lutz
FULL PAPER
Scheme 1. Synthesis of ligands 1, 2all-cis, 2all-trans and complexes 1a, 1b, 2a, 2b.
Figure 2. Most stable conformers for each stereoisomer of 2,6-dimethylcyclohexylamine.
cis–cis (all-cis), the trans–trans (all-trans), and the cis–trans the conformational exchange is slow; we observe a differ-
one (Figure 2). It was decided to use only the pure all-trans ence of 0.3 ppm between the axial and equatorial protons,
and all-cis amines. After double condensation with diacetyl the equatorial protons being the most strongly shielded.
pyridine, these two products will give few conformers com-
The attribution of the multiplet between 1.0 and 1.9 ppm
pared to the cis–trans one. Indeed, in the case of the cis– was possible thanks to a 2D experiment (1H–13C). Three
trans amine, the final pyridine-bis(imine) would probably broad multiplets that integrate respectively for 10, 2, and 4
be a mixture of isomers. Moreover, thanks to a thorough protons can be distinguished. One of the main differences
conformational analysis, it was shown that each stereoiso- between the two NMR spectra is the peak corresponding
mer had several conformers of different stabilization energy. to the proton of the cycle carried by the carbon linked to
For the all-cis and all-trans amines the most stable con- the nitrogen atom of the imine function, whose chemical
former is the one where the two methyl groups are in the shift changes from 2.9 ppm for the 2all-trans system to
equatorial position, thus minimizing steric repulsion. For 3.7 ppm for the 2all-cis system. Moreover, the coupling con-
the all-cis amine the amino group takes the axial position stant for the triplet goes from 9.3 Hz to 2.2 Hz. This differ-
and for the all-trans amine the same group is in an equato- ence may be easily explained with the Karplus and Conroy
rial position. Thus, normally only one conformer will be
obtained for each selected amine. Condensation was carried
out under ethanol reflux, with a slight excess of amine and
some drops of acetic acid.
Both ligands based on the amine, hereafter named 2all-
cis and 2all-trans, were characterized by elemental analysis,
and 1H and 13C NMR spectroscopy. In order to better allot
the resonance peaks, a COSY spectrum was recorded. It
was thus established that the molecule is in a blocked con-
figuration at ambient temperature. Indeed, in the case of
Figure 3. An ORTEP view of the crystal structure of 2all-cis. Se-
lected bond lengths [Å] and angles [°]: C(6)–C(9) 1.500(2), C(9)–
N(2) 1.270(2), N(2)–C(16) 1.465(2), C(18)–C(28) 1.530(3), C(12)–
C(24) 1.523(3), C(9)–N(2)–C(16) 121.85(15), N(2)–C(9)–C(25)
126.92(16), C(16)–C(18)–C(28) 111.60(16), C(18)–C(16)–C(12)
unsubstituted cyclohexane, it is known that there is a differ-
ence of 0.5 ppm between the chemical shifts of the axial
and equatorial protons only at low temperature, when the
exchange of cycle is slow compared to the timescale of the
NMR spectroscopy. In our case, at ambient temperature, 110.92(14).
4310
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 4309–4316

Modified Pyridine-Bis(imine) Iron and Cobalt Complexes
FULL PAPER
3
equation. Indeed, J180° coupling constants are always
3
strictly higher than J60° ones.
The X-ray structure of the ligand 2all-cis has been deter-
mined to complete the characterization and is represented
in Figure 3. Only one conformer is observed, as expected.
Synthesis and Characterization of Complexes
Synthesis and Structure of Iron and Cobalt Complexes with
4-Chloro-2,6-bis[1-(2,6-diisopropylphenylimino)ethyl]-
pyridine (1a and 1b)
The complexes were obtained from a stoichiometric reac-
tion between the ligand and a suspension of anhydrous
FeCl2 or CoCl2 in THF at room temperature. In order to
recover the complexes, the solution was concentrated and
hexane was added to the reaction medium. The iron com-
plex was isolated with a yield of 83% as a blue-green pow-
der. The cobalt complex was isolated with a yield of 81% as
a green powder. The isolated complexes were paramagnetic
Figure 4. An ORTEP view of the crystal structure of 1a. Selected
bond lengths [Å] and angles [°] for the subunit Fe1: Fe1–N(1)
2.213(4), Fe1–N(2) 2.080(3), Fe1–N(3) 2.243(4), Fe1–Cl(1)
2.247(1), Fe1–Cl(2) 2.305(2), Cl(1)–Fe1–N(1) 100.8(1), Cl(1)–Fe1–
N(2) 151.1(1), Cl(1)–Fe1–N(3) 97.2(1), Cl(2)–Fe1–N(1) 100.0(1),
Cl(2)–Fe1–N(2) 89.3(1), Cl(2)–Fe1–N(3) 102.5(1), Cl(1)–Fe1–Cl(2)
119.55(7), N(1)–Fe1–N(2) 73.7(1), N(1)–Fe1–N(3) 139.2(1), N(2)–
Fe1–N(3) 72.9(1).
1
(according to H NMR indicating chemical shifts from –40
to 170 ppm) and were characterized by elemental analysis,
infrared and X-ray analysis (iron complex only). The ele-
mental analysis results were in agreement with the formula
LMCl2.
difference is the reduction of the length of the Fe–Cl(1)
bond trans to the pyridine group (2.247 instead of 2.263 Å),
while the angle N(2)–Fe–Cl(1) where the chlorine and the
nitrogen are trans to the basal plane varies from 147.9 to
151.1° (for the chlorinated ligand). The apical iron chlorine
bond varies from 2.305 (for the chlorinated ligand) to
2.317 Å, in agreement with the expected electronic effect of
the chlorine on the para position on the pyridine, that is a
decrease of the electronic density on the iron atom. Never-
theless, the solid-state structures of the chlorinated and the
nonchlorinated complexes are very close. Besides, as we will
see in the structure of {2,6-bis[1-(2,6-dimethylcyclohex-
ylimino)ethyl]pyridine}iron dichloride (2a), the crystal
packing causes little difference in the structure.
Recrystallization of the iron complex redissolved in di-
ethyl ether yields crystals suitable for X-ray analysis. In the
solid state (Figure 4), the geometry of the metal center can
be considered as distorted square-pyramidal. One of the
chlorine atoms occupies the apical position, with an almost
right angle with the iron and nitrogen of the pyridine
[N(1)(2)(3)–Fe–Cl(2) ranging from 89.3° to 102.5°], whereas
the second occupies one of the bases of the formed square.
The apical bond seems to be the weaker, as the length of
the Fe–Cl bond goes from 2.305 Å for the apical position
to 2.247 Å for the longitudinal one. This is normally ex-
pected for that geometry. The iron atom is slightly above
the base of the pyramid [N(3)–Fe–N(1) = 139.2°]. Predict-
ably, the phenyl groups linked to the imine functions are
almost perpendicular to the plane formed by the nitrogen
atoms, although there is a slight deformation of the mole-
Synthesis and Structure of Iron and Cobalt Complexes with
2,6-Bis[1-(2,6-dimethylcyclohexylimino)ethyl]pyridine (2a
and 2b)
With the ligand 2all-cis, no iron or cobalt complexes
cule, one of the phenyl groups having a dihedral angle of could be isolated. Indeed, during the reaction between iron
94.7° and the other 85.1°. chloride and the ligand in THF at ambient temperature,
We can examine the effect of the chlorine atom in the the solution became slightly pink, but no precipitation was
para position of the pyridine ring by comparing the lengths observed, even after 24 h. The addition of hexane or diethyl
of the various bonds with X-ray data of the equivalent ether did not lead to the precipitation of the complex. Evap-
nonchlorinated iron complex.[2] The molecule is asymmetric oration of solvents to dryness yielded a pink powder. IR
in the solid state, one of the Fe–Nimine bonds measuring results prompted us to think that no complexation oc-
2.244 Å and the other 2.212 Å. This is probably due to the curred, as free FeCl2 and free ligand were detected. Similar
presence of a molecule of Et2O in the lattice, but the average observations were made when the reaction was run at 50 °C.
value is 2.228(9) Å, very close to the average value Thus, the stoichiometric reaction between the ligand 2all-
2.224(9) Å for the unsubstituted complex where the lengths cis and the metal chloride did not lead to the formation of
of these same bonds are almost identical (in this case half the expected complex of formula LMCl2. We first made the
a water molecule is present in the lattice). The decrease in assumption that a steric factor could be the cause of this
length of the Fe–Npy bond for the chlorinated ligand (2.080 nonreactivity, but an X-ray of the free ligand shown in Fig-
instead of 2.091 Å) can be explained by the fact that the ure 3 does not confirm this hypothesis.
chlorine on the pyridine group enhances the π back do-
On the contrary, the addition of 1 equiv. of the ligand
nation of the iron to the pyridine group. Another expected 2all-trans to 1 equiv. of metal chloride in THF at room
Eur. J. Inorg. Chem. 2006, 4309–4316
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
4311

F. Pelascini, M. Wesolek, F. Peruch, P. J. Lutz
FULL PAPER
temperature led to the formation of the desired complex in average equatorial one (2.28 Å), as expected from this type
good yield (75% for LFeCl2 and 65% for LCoCl2). Crystals of geometry. It is noticeable that these lengths are higher
suitable for X-ray analysis were obtained by slow crystalli- than for complexes bearing fully aromatic ligands. This
zation in THF. The X-ray crystal structure contains two seems to indicate that the electronic density on the iron be-
independent sets of iron complexes, Fe1 and Fe2, with comes higher. The iron atom is slightly above the base of
slightly different angles and distances (Figure 5, Table 1).
the pyramid: N(3)–Fe–N(1) 144.5°, which is close to the
value found for the aromatic trimethyl analog described by
Gibson et al.,[4] where the value is 145.5°. For comparison,
in complex 1a the angle is 139.2°.
The distances between the iron and both imines are al-
most identical (average 2.29 Å for Fe1 and 2.31 Å for Fe2).
These bonds are a bit longer than for the complex with the
2,4,6-trimethyl aryl substituent[4] (average 2.27 Å), revealing
the electronic effect generated by the replacement of an aryl
by a cyclohexyl group. Besides, the Fe–Npy bond length
with an average of 2.09 Å in complex 2a is shorter than for
the aryl analog, where the value is 2.11 Å. The angle N–
Fe–Cl where the nitrogen and the chlorine atom are trans
to the basal plane is somewhat greater, with an average
value of 154.4° in complex 2a compared to 131.3° in the
aryl one.[4]
Although the cyclohexyl groups are not planar and
adopt a chair conformation with the two methyl substitu-
ents and the imine in the equatorial position, they are al-
most perpendicular to the plane of the square base, like aryl
groups (dihedral angles: 117.1° and 111.8°). The average
distance between the iron atom and the methyl substituents
is about 4.3 Å, as already observed for the equivalent aro-
matic systems.
Figure 5. An ORTEP view of the crystal structure of 2a all-trans.
Only one of the two distinct units is represented. The differences
in lengths and angles are shown in Table 1.
Table 1. Selected bond lengths [Å] and angles [°] of the two subunits
Fe1 and Fe2 for the complex 2a all-trans.
Bond lengths [Å]
Fe1–N(1)
Fe1–N(2)
Fe1–N(3)
Fe1–Cl(1)
Fe1–Cl(2)
2.286(4)
2.087(4)
2.295(4)
2.349(1)
2.279(1)
Fe2–N(6)
Fe2–N(5)
Fe2–N(4)
Fe2–Cl(3)
Fe2–Cl(4)
2.318(4)
2.091(4)
2.308(4)
2.329(2)
2.286(1)
Bond angles [°]
Cl(1)–Fe1–N(1) 97.9(1)
Cl(1)–Fe1–N(2) 91.8(1)
Cl(1)–Fe1–N(3) 95.1(1)
Cl(2)–Fe1–N(1) 103.2(1)
Cl(2)–Fe1–N(2) 153.4(1)
Cl(2)–Fe1–N(3) 101.0(1)
Cl(1)–Fe1–Cl(2) 114.7(1)
N(1)–Fe1–N(2) 73.7(2)
N(1)–Fe1–N(3) 144.5(1)
N(2)–Fe1–N(3) 73.0(2)
Cl(3)–Fe2–N(6)
Cl(3)–Fe2–N(5)
Cl(3)–Fe2–N(4)
Cl(4)–Fe2–N(6)
Cl(4)–Fe2–N(5)
Cl(4)–Fe2–N(6)
Cl(3)–Fe2–Cl(4)
N(5)–Fe2–N(6)
N(4)–Fe2–N(6)
N(4)–Fe2–N(5)
95.9(1)
95.1(1)
102.0(1)
102.2(1)
155.4(1)
102.2(1)
109.4(1)
72.8(2)
Ethylene Polymerization with Complexes 1a, 1b, 2a, and 2b
Activated with MAO
All the complexes were activated with methylalumin-
oxane and tested for the polymerization of ethylene. The
results are summarized in Table 2 together with the experi-
mental conditions.
142.7(2)
73.2(2)
With complexes 1a and 1b, polyethylene was produced
As in the complexes with an aryl group on the imine with the same activity for the two metals, whereas, iron
functions, the geometry of the metal center can be consid- complexes generally present a higher catalytic activity than
ered as distorted square-pyramidal, with one of the chlorine their cobalt analogs. This result might be due to faster deac-
atoms occupying the apical position, with an almost 90- tivation processes of 1a versus 1b. The activities were com-
degree angle with the iron and the nitrogen atoms: pared with those observed for the nonchlorinated pyridine
N(1)(2)(3)–Fe1–Cl(1) ranging from 91.8° to 97.9° and complexes (diiPrFe and diiPrCo). As indicated in Table 2,
N(4)(5)(6)–Fe2–Cl(3) ranging from 95.1° to 102.0°. The the presence of the chlorine atom in the para position of
average apical Fe–Cl bond (2.35 Å) is longer than the the pyridine decreases the catalytic activity, from 2220 to
Table 2. Ethylene polymerization with iron and cobalt complexes/MAO.
Catalyst[a]
Yield
[g]
Activity
[g·mmol–1·h–1·bar–1] (gPE/gCat.
Productivity
Mn
Mw
MMD[c]
Tm
χ[d]
[b]
[b]
[d]
)
[g·mol–1]
[g·mol–1]
[°C]
1a
1.9
5.5
1.9
3.4
810
2220
810
295
904
294
556
800
35000
48000
29000
30000
44
22
4
97; 136
119; 136
135
0.82
0.82
0.82
0.84
diiPr-Fe
1b
2200
6400
7000
diiPr-Co
1410
4
137
[a] Catalyst: 10 µmol; [Al]/[Met] = 400; PEt. = 1 bar; T = 35 °C; t = 15 min; toluene (40 mL). [b] Average molar masses measured by
High Temperature Size Exclusion Chromatography in 1,2,4-trichlorobenzene at 150 °C with RI and viscosimetry detection. [c] Molar
mass distribution. [d] Melting temperature and crystallinity rate measured by DSC.
4312
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 4309–4316

Modified Pyridine-Bis(imine) Iron and Cobalt Complexes
FULL PAPER
810 g·mmol–1·h–1·bar–1 for the iron complexes and from peaks attributed to the protons on the pyridine are strongly
1410 to 810 g·mmol–1·h–1·bar–1 for the cobalt one. It was shifted. Although the spectrum was obtained with a large
thought that slightly decreasing the electronic density on window, from –55 to 250 ppm, some of the protons of the
the metal would render the interaction with the incoming complex cannot be observed, probably because they are too
ethylene more favorable and then would enhance the cata- large. Indeed the integration gives 35 protons instead of the
lytic activity.
39 expected. An NMR study at low temperature, down to
Actually, this is not what happens because lowering the –90 °C, does not reveal new peaks. On the contrary, the
electronic density on the metal simultaneously disfavors the broad peak at –11 ppm broadens even further at –70 °C and
π back donation of the metal to the incoming ethylene. is invisible at –90 °C.
Bennett did not observe significant falls of activity with a
cobalt complex substituted by a trifluoromethyl group in
the para position of the pyridine ring, but experimental con-
ditions were not comparable.[25] As the decrease in the cata-
lytic activity cannot be explained by the small differences
in the structures of the chlorinated and the nonchlorinated
complex, the possible interaction of the chlorine atom pres-
ent on the pyridine with MAO, used in a large excess, was
studied. To do this we have observed by 1H NMR the
chemical shifts of chlorobenzene (to mimic the ligand, this
one would coordinate first with the nitrogen groups) in deu-
terated toluene after addition of a quantity of MAO similar
to that used in the polymerization experiment. No signifi-
cant chemical shifts have been observed, which means that
there is probably no interaction between the chlorine atom
and the TMA present in the MAO, or the MAO itself.
Eventually, the loss of activity observed for the chlorinated
complexes could be attributed to the lack of stability of the
1
complex. Indeed, as can be seen on the H NMR spectrum
of the complex (Figure 6), a great amount of decomposition
product appeared within 20 min; this quantity increases
with time. The decomposition process has not been studied
in detail.
Figure 7. (a) 1H NMR spectrum of the iron complex 2a containing
THF marked with *; all the other nonmarked peaks are due to the
protons on the cyclohexyl group. (b) 1H NMR spectrum of the
cobalt complex 2b containing THF marked with * (300 MHz,
293 K, CDCl3).
All the spectra were recorded with pure crystals that con-
tained THF of crystallization; this is the reason two rela-
tively intense peaks of free THF can always be seen in the
spectra. In the case of the cobalt complex the proton NMR
shows us more important paramagnetic shift: the peaks are
1
Figure 6. H NMR spectrum of complex 1a. The peaks of the de-
composition product are marked with * (300 MHz, 293 K, CDCl3).
in the range –40 to 166 ppm (Figure 7, b). The integration
With the iron complex 2a no polymers or oligomers have of the peaks corresponds to what is expected.
been detected. However, with the cobalt complex 2b poly- These observations and the fact that there is a marked
mers are obtained (no oligomers could be detected by GC). difference in the reactivity between the iron and the cobalt
Bianchini obtained oligomers with dissymmetric pyridine- complexes bearing the same ligands prompted us to think
bis(imine) ligands [the imine substituents being an aryl that the ligand is in some way hemilabile in the case of iron.
group and a (cyclo)alkyl group].[21] The proton NMR spec-
The complete loss of activity can then be explained by
tra of the iron complex (Figure 7, a) reveals that almost all the assumption that the addition of MAO can hinder the
the peaks are in the region from –8 to 11 ppm; only the return of the imine group to the iron. On the contrary, for
Eur. J. Inorg. Chem. 2006, 4309–4316
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
4313

F. Pelascini, M. Wesolek, F. Peruch, P. J. Lutz
FULL PAPER
the cobalt complex 2b, the spectrum of which is shown in
Figure 7 (b), polymerization of ethylene is observed in the
presence of MAO. Under 4 bar of ethylene and after 3 h,
150 mg of polyethylene was isolated. We will not go into
more detail, as the subject is currently under investigation.
The polymers obtained with 1a and 1b were analyzed by
1H NMR, DSC, and HT SEC. 1H NMR spectra of polyeth-
ylenes (C2D2Cl4, 130 °C) present a main peak at 1.3 ppm
due to CH2 groups, and a small one at 0.9 ppm due to CH3
groups coming from chain ends or branching. Nevertheless,
the intensity of the peak at 0.9 ppm is too small to allow
reliable integration.
The results of SEC measurements are presented in
Table 2. The chlorine atom seems to have no influence on
the molar masses or the molar mass distributions when co-
balt is employed. For the iron complexes, broad (even bi-
modal) molar mass distributions were obtained. It is now
well established that this is due to transfer to aluminum
(TMA present in MAO). It was nevertheless a bit surprising
to get such broad distributions, as to avoid this transfer,
the MAO we employed had most of its TMA removed by
evaporation (5 mol-% residual TMA instead of 30 mol-%
Experimental Section
General Considerations
Anhydrous FeCl2 and CoCl2, diacetylpyridine, chelidamic acid,
thionyl chloride, Meldrum’s acid, and 2,6-diisopropylaniline (all
purchased from Aldrich) were used as received. 2,6-Dimethylcyclo-
hexylamine was kindly provided by BASF AG. Methylaluminoxane
(MAO) was purchased from Aldrich as a 10 wt.-% solution in tolu-
ene. Toluene and most of the trimethylaluminum (TMA) were re-
moved under vacuum to yield a white powder, still containing
5 mol-% of TMA. Tetrahydrofuran, hexane, and toluene were dis-
tilled from sodium benzophenone ketyl. All complex syntheses
were performed in a Jacomex glovebox or under thoroughly puri-
fied argon using a standard Schlenk technique. 2,6-Diacetyl-4-chlo-
ropyridine was prepared according to an already published pro-
cedure.[24]
Characterization
1H and 13C NMR spectra were recorded with a Bruker Avance-
300 apparatus (300 MHz) at room temperature. IR spectra were
recorded with a Perkin–Elmer 1600 FTIR spectrometer. Molar
masses were measured by Size Exclusion Chromatography using an
Alliance GPCV 2000 Permeation Chromatograph equipped with a
differential refractive index detector and a viscosimeter in 1,2,4-
in commercial MAO). For cobalt complexes, molar mass trichlorobenzene (150 °C) using two Styragel HT 6E and one HT 2
columns. DSC measurements were performed with a Perkin–Elmer
distributions were much lower, because these catalytic sys-
tems are less sensitive towards transfer to aluminum.
The results of DSC measurements are also presented in
Table 2. Like for SEC measurements, no huge differences
were observed for the chlorinated and nonchlorinated li-
gands. The chlorine atom seems to have almost no influence
on the melting temperatures and on the crystallinity rate.
For all polyethylenes, the crystallinity rate (Ͼ0.8) is consis-
tent with highly linear polymers. Polyethylenes obtained
DSC4 or DSC7, with a heating rate of 10 °C·min–1.
Synthesis of Ligands
4-Chloro-2,6-bis[1-(2,6-diisopropylphenylimino)ethyl]pyridine
(1):
2,6-Diacetyl-4-chloropyridine (0.5 g, 2.5 mmol), 2,6-diisopro-
pylaniline (0.98 g, 5.5 mmol), and a small amount of p-toluenesul-
fonic acid were poured into a three-necked round-bottomed flask
containing toluene (50 mL). The solution was refluxed for 24 h
with a Dean–Stark trap. After lowering the temperature, the tolu-
with the iron complexes exhibit two melting temperatures ene was removed under vacuum to yield a brownish oil. The ad-
dition of hexane (20 mL) led to a yellow solution and a brown
solid. After filtration and drying, a yellow solid was recovered
(0.94 g, yield 72%). H NMR (300 MHz, CDCl3, 293 K): δ = 8.48
corresponding probably to two populations of polymers of
different molar masses observed by HT SEC. With the co-
balt complexes, only one melting temperature, very close to
the one measured for highly linear polyethylene, was ob-
served.
1
3
(s, 2 H, py-Hm), 7.22–7.09 (m, 6 H, Ar-H), 2.74 (sept, JH,H
=
=
3
6.9 Hz, 4 H, CHMe2), 2.25 (s, 6 H, N=C-Me), 1.17 (d, JH,H
3
6.9 Hz, 12 H, CHMeMe), 1.16 (d, JH,H = 6.9 Hz, 12 H,
CHMeMe) ppm. 13C NMR (75 MHz, CDCl3, 293 K): δ = 166.0
(N=C), 156.5 (Py-Co), 146.0 (Ar-Cip), 145.4 (Py-C-Cl), 135.7 (Ar-
Co), 123.8 (Ar-Cp), 123.0 (Py-Cm), 122.2 (Ar-Cm), 28.3 (N=C-Me),
Conclusion
23.3 (CHMeMe), 22.9 (CHMeMe), 17.2 (CHMe ) ppm. IR: ν =
˜
2
1640, 1557, 1317, 1260, 1228, 1189, 1104, 1016, 935, 885, 828, 762,
722, 692, 535, 438 cm–1. C33H42ClN3 (516.16): calcd. C 76.79, N
8.14, H 8.20; found C 73.23, N 7.21, H 8.35.
In summary, two new pyridine-bis(imine)s were synthe-
sized, by substituting the pyridine ring, or by replacing the
aryl groups linked to the imine functions by an aliphatic
ring. These ligands were complexed with iron and cobalt
2,6-Bis[1-(2,6-dimethylcyclohexylimino)ethyl]pyridine (2all-cis): 2,6-
Diacetylpyridine (2 g, 12.3 mmol), 2,6-dimethylcyclohexylamine
dichloride and two of these four new complexes were char- all-cis (3.2 g, 25 mmol), and a few drops of acetic acid were poured
into a three-necked round-bottomed flask containing ethanol
(50 mL). The solution was refluxed for 18 h. After lowering the
temperature, a white precipitate appeared. After filtration, washing
acterized by X-ray analysis. All the complexes were tested
in the presence of MAO for the polymerization of ethylene.
The introduction of a chlorine atom onto the pyridine ring
was unfavorable; indeed the catalytic activity decreases
compared to the unsubstituted ligand. For the other ligand
two different results were observed. In the case of the iron
complex, the loss of aromaticity led to a complete loss of
catalytic activity towards ethylene polymerization and
oligomerization. On the other hand, for the cobalt system,
polyethylene could be obtained.
with pentane and drying, a white solid was recovered (3.7 g, yield
3
82%). 1H NMR (300 MHz, CDCl3, 293 K): δ = 8.18 (d, JH,H
=
3
7.6 Hz, 2 H, py-Hm), 7.68 (t, JH,H = 7.6 Hz, 1 H, py-Hp), 3.71 (t,
3JH,H = 2.2 Hz, 2 H, CHcHex-N=), 2.38 (s, 6 H, N=C-CH3), 1.88–
1.32 (br. m, 16 H, cHex-CH2-), 0.70 (d, 3JH,H = 6.6 Hz, 12 H, cHex-
CH3) ppm. 13C NMR (75 MHz, CDCl3, 293 K): δ = 164.0 (-C=N),
156.5 (py-Co), 135.9 (py-Cp), 120.8 (py-Cm), 65.2 (CHcHex-N=),
38.6 (cHex-Co), 28.9 (cHex-Cm), 27.1 (cHex-Cp), 19.3 (cHex-CH3),
4314
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 4309–4316

Modified Pyridine-Bis(imine) Iron and Cobalt Complexes
FULL PAPER
13.3 (CH -C=N) ppm. IR: ν = 1703, 1638, 1567, 1303, 1237, 1156,
–45.57 (s, 6 H, NCMe) ppm. C33H42Cl3FeN3 (642.91): calcd. C
˜
3
1116, 1086, 995, 980, 950, 857, 838, 818, 626, 558, 530, 491, 440,
338, 277 cm–1. C25H39N3 (381.60): calcd. C 78.69, N 11.01, H
10.30; found C 78.70, N 11.11, H 10.34.
61.65, N 6.54, H 6.58; found C 61.88, N 6.01, H 7.35. IR: ν =
˜
1689, 1574, 1500, 1320, 1282, 1237, 1115, 1015, 806, 722, 525, 365,
313 cm–1.
2,6-Bis[1-(2,6-dimethylcyclohexylimino)ethyl]pyridine (2all-trans):
2,6-Diacetylpyridine (2 g, 12.3 mmol), 2,6-dimethylcyclohex-
ylamine all-trans (3.2 g, 25 mmol), and a few drops of acetic acid
were poured into a three-necked round-bottomed flask containing
ethanol (50 mL). The solution was refluxed for 18 h. After lowering
the temperature, a white precipitate appeared. After filtration,
washing with pentane and drying, a white solid was recovered
{4-Chloro-2,6-bis[1-(2,6-diisopropylphenylimino)ethyl]pyridine}-
cobalt Dichloride (1b): CoCl2 (38 mg, 0.29 mmol) and 4-chloro-2,6-
bis[1-(2,6-diisopropylphenylimino)ethyl]pyridine (150 mg,
0.29 mmol) were dissolved in THF (25 mL). The reaction was let
run for 16 h. After precipitation with hexane, filtration, and drying,
a g r e e n p owd e r wa s r e c ov e r e d ( 1 5 0 m g , y i e l d 8 1 % ) .
C33H42Cl3CoN3 (646.00): calcd. C 61.36, N 6.50, H 6.55; found C
1
(3.6 g, yield 76%). H NMR (300 MHz, CDCl3, 293 K): δ = 8.09
59.52, N 5.42, H 6.70. IR: ν = 1582, 1325, 1264, 1204, 1103, 1057,
˜
3
3
(d, JH,H = 7.7 Hz, 2 H, py-Hm), 7.69 (t, JH,H = 7.7 Hz, 1 H, py- 1028, 938, 858, 854, 838, 798, 768, 722, 590, 487, 440, 347, 309,
Hp), 2.92 (t, JH,H = 9.3 Hz, 2 H, CHcHex-N=), 2.39 (s, 6 H, N=C 280, 213 cm–1.
3
3
Me), 1.82–1.10 (br. m, 16 H, cyclohexyl -CH2-), 0.74 (d, JH,H
=
{2,6-Bis[1-(2,6-dimethylcyclohexylimino)ethyl]pyridine}iron Dichlo-
ride (2a): FeCl2 (132 mg, 1.04 mmol) and 2,6-bis[1-(2,6-dimethylcy-
clohexylimino)ethyl]pyridine all-trans (400 mg, 1.04 mmol) were
dissolved in THF (25 mL). The reaction was let run for 16 h. After
precipitation with hexane, filtration, and drying, a blue powder was
recovered (400 mg, yield 75 %). 1H NMR (300 MHz, CD2Cl2,
298 K, δsolvent = 5.32): δ = 78.34 (s, 2 H, py-Hm), 10.14 (v br. s, 2
H, CcHex-H), 8.84 (s, 2 H, CcHex-H), 8.58 (s, 1 H, py-Hp), 3.46 (s,
2 H, CcHex-H), 1.73 (s, 4 H, CcHex-H), 1.02 (s, 4 H, CcHex-H), –1.29
(s, 6 H, N=C-CH3), –6.99 (br. s, 12 H, CcHex-CH3) ppm.
C25H39Cl2FeN3 (508.35): calcd. C 59.07, N 8.27, H 7.73; found C
6.6 Hz, 12 H, cHex-CH3) ppm. 13C NMR (75 MHz, CDCl3,
293 K): δ = 164.8 (-C=N), 156.8 (py-Co), 136.5 (py-Cp), 121.0 (py-
Cm), 72.8 (CHhex-N=), 37.9 (cHex-Co), 33.9 (cHex-Cm), 25.8 (cHex-
C ), 19.5 (cHex-CH ), 14.7 (CH -C=N) ppm. IR: ν = 1643, 1567,
˜
p
3
3
1305, 1237, 1172, 1153, 1116, 1088, 993, 948, 858, 839, 817, 779,
624, 555, 440, 340, 274, 228 cm–1. C25H39N3 (381.60): calcd. C
78.69, N 11.01, H 10.30; found C 78.45, N 11.41, H 10.38.
Synthesis of the Complexes
{4-Chloro-2,6-bis[1-(2,6-diisopropylphenylimino)ethyl]pyridine}iron
Dichloride (1a): FeCl2 (36 mg, 0.28 mmol) and 4-chloro-2,6-bis[1-
(2,6-diisopropylphenylimino)ethyl]pyridine (150 mg, 0.29 mmol)
were dissolved in THF (25 mL). The reaction was let run for 16 h.
After precipitation with hexane, filtration, and drying, a blue pow-
der was recovered (150 mg, yield 83 %). 1H NMR (300 MHz,
CDCl3, 293 K, δsolvent = 7.26): δ = 76.46 (s, 2 H, Hm, Py), 14.40 (s,
58.52, N 8.40, H 7.71. IR: ν = 1582, 1268, 1198, 1020, 948, 842,
˜
816, 718, 639, 568, 314, 285 cm–1.
{2,6-Bis[1-(2,6-dimethylcyclohexylimino)ethyl]pyridine}cobalt Di-
chloride (2b): CoCl2 (136 mg, 1.04 mmol) and 2,6-bis[1-(2,6-di-
methylcyclohexylimino)ethyl] all-trans (400 mg, 1.04 mmol) were
4 H, Hm Ar), –5.86 (br. s, 12 H, CHMeMe), –6.62 (s, 12 H, dissolved in THF (25 mL). The reaction was let run for 16 h. After
CHMeMe), –10.87 (s, 2 H, Hp Ar), –23.70 (br. s, 4 H, CHMe), precipitation with hexane, filtration, and drying, a green powder
Table 3. Crystal data and structure refinement.
1a
2a all-trans
2all-cis
Empirical formula
C37H52Cl3FeN3O
C33H42Cl3FeN3·C4H10O
717.05
C112H182Cl8Fe4N12O4
C100H156Cl8Fe4N12·3C4H8O·H2O
2267.8(1)
173
C25H39N3
Formula mass
Temperature [K]
381.59
173
173
Z
4
4
8
Crystal system
Space group
monoclinic
P21/c
monoclinic
C2/c
orthorhombic
Pbcn
a [Å]
b [Å]
c [Å]
10.3742(2)
20.2225(5)
18.6578(4)
97.982(5)
3876.3(1)
blue
0.20×0.06×0.02
1.23
1520
43.5402(5)
12.5601(2)
26.2792(3)
117.157(5)
12787.0(6)
violet
0.16×0.12×0.06
1.17
4800
23.0750(6)
14.6860(6)
13.7930(9)
90.00
4674.2(4)
colorless
0.20×0.20×0.15
1.085
1680
0.063
–29 Յ h Յ 29
–19 Յ k Յ 19
–17 Յ l Յ 17
1.77–27.48
5354
β [°]
V [Å3]
Color
Crystal size [mm]
Dcalcd. [g·cm–3]
F(000)
Absorption coefficient [mm–1]
Index ranges
0.627
0.661
–14 Յ h Յ 14
–26 Յ k Յ 28
–26 Յ l Յ 26
2.5–30.55
12118
3120
406
0 Յ h Յ 61
0 Յ k Յ 17
–36 Յ l Յ 32
2.5–30.01
19449
10571
632
θ range [°]
Reflections collected
Reflections observed [I Ͼ 3σ(I)]
Number of parameters
R
2862
253
0.0555
0.046
0.076
Rw
0.061
1.028
0.350, –0.122
0.094
1.119
0.868, –0.873
0.1278
0.954
0.183, –0.186
Goodness-of-fit on F2
Largest difference peak, hole [e·Å–3]
Eur. J. Inorg. Chem. 2006, 4309–4316
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
4315

F. Pelascini, M. Wesolek, F. Peruch, P. J. Lutz
FULL PAPER
was recovered (350 mg, yield 65%). 1H NMR (300 MHz, CDCl3,
293 K, δH = 7.26): δ = 164.79 (s, 2 H, CcHex-H), 103.50 (s, 2 H, py- [1] G. J. P. Britovsek, V. C. Gibson, B. S. Kimberley, P. J. Maddox,
S. J. McTavish, G. A. Solan, A. J. P. White, D. J. Williams,
Chem. Commun. 1998, 849–850.
Hm), 27.73 (s, 1 H, py-Hp), 18.21 (s, 2 H, CcHex-H), 3.70 (s, 2 H,
cHex-H), 0.79 (s, 2 H, CcHex-H), –4.72 (s, 4 H, CcHex-H), –11.59
(s, 6 H, N=C-CH3), –15.58 (s, 4 H, CcHex-H), –16.02 [s (v br. sh),
2 H, CcHex-H], –39.19 (s, 12 H, CcHex-CH3) ppm. C25H39Cl2CoN3
(511.44): calcd. C 58.71, N 8.20, H 7.69; found C 58.41, N 8.49, H
C
[2] B. L. Small, M. Brookhart, A. M. A. Bennett, J. Am. Chem.
Soc. 1998, 120, 4049–4050.
[3] B. L. Small, M. Brookhart, J. Am. Chem. Soc. 1998, 120, 7143–
7144.
7.69. IR: ν = 1620, 1582, 1266, 1248, 1198, 1115, 1023, 948, 864,
˜
[4] a) G. J. P. Britovsek, M. Bruce, V. C. Gibson, B. S. Kimberley,
P. J. Maddox, S. Mastroianni, S. J. McTavish, C. Redshaw,
G. A. Solan, S. Strömberg, A. J. P. White, D. J. Williams, J. Am.
Chem. Soc. 1999, 121, 8728–8740; b) V. C. Gibson, S. K. Spitz-
messer, Chem. Rev. 2003, 103, 283–315.
[5] J. Qiu, Y. Li, Y. Hu, Polym. Int. 2000, 49, 5–7.
[6] G. J. P. Britovsek, S. Mastroianni, G. A. Solan, S. P. D. Baugh,
C. Redshaw, V. C. Gibson, A. J. P. White, D. J. Williams,
M. R. J. Elsegood, Chem. Eur. J. 2000, 6, 2221–2231.
[7] Z. Ma, H. Wang, J. Qiu, D. Xu, Y. Hu, Macromol. Rapid Com-
mun. 2001, 22, 1280–1283.
[8] R. Chen, C. Qian, J. Sun, Chin. J. Chem. 2001, 19, 866–869.
[9] A. S. Abu-Surrah, K. Lappalainen, U. Piironen, P. Lehmus, T.
Repo, M. Leskelä, J. Organomet. Chem. 2002, 648, 55–61.
[10] Z. Ma, W.-H. Sun, Z.-L. Li, C.-X. Shao, Y.-L. Hu, Chin. J.
Polym. Sci. 2002, 20, 205–211.
[11] Z. Ma, W.-H. Sun, N. Zhu, Z. Li, C. Shao, Y. Hu, Polym. Int.
2002, 51, 349–352.
823, 816, 746, 716, 638, 568, 317, 287 cm–1.
Polymerization Procedure: Polymerizations were conducted in a
250-mL glass Büchi reactor equipped with a magnetic stirrer and
a temperature controller. Ethylene pressure was maintained con-
stant in the reactor during the reaction. A typical procedure was:
catalyst (10 µmol) in solution in toluene (30 mL) was added to the
reactor under nitrogen. The reactor was heated to the reaction tem-
perature and charged with ethylene. A solution of the desired
amount of MAO dissolved in toluene (10 mL) was then added to
the system. The reactor was charged with the desired differential
pressure of ethylene. The reaction was let run for a period of
15 min, after which acidified ethanol (20 mL) was added to the
miniclave to stop the polymerization. The content of the reactor
was then precipitated in ethanol. The precipitated polymer was
washed with ethanol and dried overnight in a vacuum oven at
60 °C.
[12] R. Schmidt, U. Hammon, S. Gottfried, M. B. Welch, H. G. Alt,
J. Appl. Polym. Sci. 2003, 88, 476–482.
[13] Y. Chen, R. Chen, C. Qian, X. Dong, J. Sun, Organometallics
2003, 22, 4312–4321.
X-ray Crystallography: Crystals of 1a, 2a, and 2all-cis coated with
vaseline were mounted onto the goniometer and placed on a Non-
ius Kappa CCD diffractometer with graphite-monochromated Mo-
Kα radiation (λ = 0.71073 Å). The structures were solved using the
Nonius OpenMoleN package and refined against F2 using the
SHELXL-97 software[25] with anisotropical thermal parameters for
all non-hydrogen atoms (except for one THF and the H2O in 2a)
and hydrogen atoms were introduced as fixed contribution
(SHELXL-97[26]) (except for one THF and the H2O in 2a) (see
Table 3).
[14] Y. Chen, C. Qian, J. Sun, Organometallics 2003, 22, 1231–1236.
[15] H. Wang, W. D. Yan, T. Jiang, B. B. Liu, W. Q. Xu, J. J. Ma,
Y. L. Hu, Chin. Chem. Lett. 2003, 14, 257–258.
[16] M. E. Bluhm, C. Folli, M. Döring, J. Mol. Catal. A 2004, 212,
13–18.
[17] a) F. Pelascini, F. Peruch, P. J. Lutz, M. Wesolek, J. Kress, Eur.
Polym. J. 2005, 41, 1288–1295; b) V. C. Gibson, S. K. Spitz-
messer, Chem. Rev. 2003, 103, 283–315; c) S. D. Ittel, L. K.
Johnson, M. Brookhart, Chem. Rev. 2000, 100, 1169–1203; d)
R. Mulhaupt, Macromol. Chem. Phys. 2003, 204, 289–327.
[18] G. J. P. Britovsek, V. C. Gibson, S. Mastroianni, D. C. H.
Oakes, C. Redshaw, G. A. Solan, A. J. P. White, D. J. Williams,
Eur. J. Inorg. Chem. 2001, 431–437.
[19] G. J. P. Britovsek, V. C. Gibson, B. S. Kimberley, S. Mastro-
ianni, C. Redshaw, G. A. Solan, A. J. P. White, D. J. Williams,
J. Chem. Soc., Dalton Trans. 2001, 1639–1644.
[20] C. Amort, M. Malaun, A. Krajete, H. Kopacka, K. Wurst, M.
Christ, D. Lilge, M. O. Kristen, B. Bildstein, Appl. Organomet.
Chem. 2002, 16, 506–516.
CCDC-255729 (for C33H42Cl3FeN3·C4H10O, 1a), -255730 (for
C100H156Cl8Fe4N12·3OC4H8·H2O, 2a), and -613556 (for C25H39N3,
2all-cis) contain the supplementary crystallographic data for this
paper. These data can be obtained free of charge from The Cam-
bridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif.
[21] C. Bianchini, G. Mantovani, A. Meli, F. Migliacci, F. Zanobini,
F. Laschi, A. Sommazzi, Eur. J. Inorg. Chem. 2003, 1620–1631.
[22] G. J. P. Britovsek, V. C. Gibson, O. Hoarau, S. K. Spitzmesser,
A. J. P. White, D. J. Williams, Inorg. Chem. 2003, 42, 3454–
3465.
[23] V. C. Gibson, S. K. Spitzmesser, A. J. P. White, D. J. Williams,
Dalton Trans. 2003, 2718–2727.
[24] R. Ali Fallahpour, E. C. Constable, J. Chem. Soc., Perkin
Trans. 1 1997, 2263–2264.
[25] A. M. A. Bennett, World Patent, WO9827124, 1998.
[26] G. M. Sheldrick, SHELXL-97, Program for the Refinement of
Crystal Structures, University of Göttingen, Germany, 1997.
Received: March 16, 2006
Acknowledgments
The authors want to thank the CNRS for financial support, Prof.
Richard Blackborow, BASF, and Dr. Rath for providing the 2,6-
dimethylcyclohexylamines, Suzanne Zehnacker for DSC measure-
ments, Emmanuel Ibarboure and Prof. Henri Cramail (LCPO, Bor-
deaux, France) for HT SEC measurements, Jean-Daniel Sauer for
the low-temperature NMR measurements, Nathalie Kyritsakas-
Gruber for X-ray refinements of the complexes, Dr. André De Cian
for all the X-ray measurements and the X-ray refinement of the
free ligand, and Dr. Marie-Thérèse Youinou for her helpful advice.
Published Online: August 24, 2006
4316
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 4309–4316