Ferrocene-substituted bis(imino)pyridine iron and cobalt complexes: Toward redox-active catalysts for the polymerization of ethylene

Article abstract of DOI:10.1021/om0509589

The synthesis, characterization, and ethylene polymerization behavior of a series of iron and cobalt halide complexes LMCl₂ (M = Fe, Co) bearing chelating ferrocenyl-modified bis(imino)pyridyl ligands L [L = 2,6-(FcArNCMe)₂C₅H₃N] are reported. X-ray diffraction studies show the geometry at the metal centers to be distorted square pyramidal. Electrochemical studies focus on an Fe (7) and a Co (10) complex and in each case show several quasi-reversible oxidation and reduction waves, the ferrocenyl iron atoms being easier to oxidize than the central iron (and cobalt) atoms. Chemical oxidation of these complexes is possible via addition of FcPF₆, and good yields of the doubly charged iron and cobalt complexes were isolated. Treatment of the complexes 7 and 10 with methylaluminoxane (MAO) leads to highly active ethylene polymerization catalysts, converting ethylene to highly linear polyethylene (PE), with polymer molecular weights in the region of 900 000. As expected, the Fe catalyst is approximately an order of magnitude more active than the Co species, with activities of 6900 and 440 g mmol^-1 bar^-1 h^-1, respectively. The effect of the polymerization activities by oxidation of the precatalysts to their ferrocenium counterparts was seemingly negated by the reducing properties of the alkyl aluminum activator (MAO).

Full text of DOI:10.1021/om0509589

1932
Organometallics 2006, 25, 1932-1939
Ferrocene-Substituted Bis(imino)pyridine Iron and Cobalt
Complexes: Toward Redox-Active Catalysts for the Polymerization
of Ethylene
Vernon C. Gibson,* Nicholas J. Long,* Philip J. Oxford, Andrew J. P. White, and
David J. Williams
Department of Chemistry, Imperial College London, Exhibition Road,
South Kensington, London, U.K., SW7 2AZ
ReceiVed NoVember 8, 2005
The synthesis, characterization, and ethylene polymerization behavior of a series of iron and cobalt
halide complexes LMCl2 (M ) Fe, Co) bearing chelating ferrocenyl-modified bis(imino)pyridyl ligands
L [L ) 2,6-(FcArNCMe)2C5H3N] are reported. X-ray diffraction studies show the geometry at the metal
centers to be distorted square pyramidal. Electrochemical studies focus on an Fe (7) and a Co (10) complex
and in each case show several quasi-reversible oxidation and reduction waves, the ferrocenyl iron atoms
being easier to oxidize than the central iron (and cobalt) atoms. Chemical oxidation of these complexes
is possible via addition of FcPF6, and good yields of the doubly charged iron and cobalt complexes were
isolated. Treatment of the complexes 7 and 10 with methylaluminoxane (MAO) leads to highly active
ethylene polymerization catalysts, converting ethylene to highly linear polyethylene (PE), with polymer
molecular weights in the region of 900 000. As expected, the Fe catalyst is approximately an order of
magnitude more active than the Co species, with activities of 6900 and 440 g mmol-1 bar-1 h-1,
respectively. The effect of the polymerization activities by oxidation of the precatalysts to their ferrocenium
counterparts was seemingly negated by the reducing properties of the alkyl aluminum activator (MAO).
We have an ongoing research program that features the
synthesis and characterization of redox-tunable polymerization
catalyst systems probing the influence of redox-tunable ligand
substituents on catalyst activities and selectivities. Previously,
we10 and others11-14 have shown that the ferrocenyl unit can
be successfully attached to the N-donor atoms of amide and
imide ligands. However, we found that when positioned this
close to the polymerization center, oxidation of the ferrocene
units becomes problematic due to the strong electron-withdraw-
ing effect of the pendant electrophilic metal center(s). Thus,
we decided to re-focus our efforts toward ligands that are known
to support polymerization active centers, but that have been
modified to contain a ferrocene substituent in the ligand
backbone quite remote from the active site but still within range
to influence the activity and selectivity of the active center.
These redox-active species offer exciting potential within
homogeneous transition metal catalysts, since the oxidation of
Introduction
The introduction of molecular polymerization catalysts has
played a central role in the development of new polymeric
materials. A key advantage of these new catalysts systems is
the added control they afford over macromolecular properties,
such as molecular weight, molecular weight distribution,
comonomer incorporation, and tacticity. In recent years, com-
plexes of the late transition metals have gained increasing
prominence as olefin polymerization precatalysts.1,2 In particular,
the families of iron and cobalt catalysts bearing terdentate bis-
(imino)pyridine ligands have attracted great interest.1,3-9 Here,
bulky substituents on the imino nitrogen donors were found to
play a pivotal role in preventing the formation of bis-chelate
complexes and controlling the rates of propagation and chain
termination, thus influencing both activity and molecular weight
of the resultant polymer.
* To whom correspondence should be addressed. E-mail: n.long@
imperial.ac.uk; v.gibson@imperial.ac.uk. Tel: 020 75945781. Fax: 020
75945804.
(9) (a) Britovsek, G. J. P.; Gibson, V. C.; Kimberley, B. S.; Solan, G.
A.; White, A. J. P.; Williams, D. J. Chem. Commun. 1998, 849. (b) Small,
B. L.; Brookhart, M.; Bennett, A. M. A. J. Am. Chem. Soc. 1998, 120,
4049. (c) Britovsek, G. J. P.; Bruce, M.; Gibson, V. C.; Kimberley, B. S.;
Maddox, P. J.; Mastroiani, S.; McTavish, S. J.; Redshaw, C.; Solan, G. A.;
Stromberg, S.; White, A. J. P.; Williams, D. J. J. Am. Chem. Soc. 1999,
121, 8728.
(10) (a) Gibson, V. C.; Long, N. J.; Marshall, E. L.; Oxford, P. J.; White,
A. J. P.; Williams, D. J. J. Chem. Soc., Dalton Trans. 2001, 1162. (b)
Gibson, V. C.; Halliwell, C. M.; Long, N. J.; Oxford, P. J.; Smith, A. M.;
White, A. J. P.; Williams, D. J. J. Chem. Soc., Dalton Trans. 2003, 918.
(11) Shafir, A.; Power, M. P.; Whitener, G. D.; Arnold, J. Organome-
tallics 2001, 20, 1365.
(12) Siemeling, U.; Kuhnert, O.; Neumann, B.; Stammler, A.; Stammler,
H. G.; Bildstein, B.; Malaun, M.; Zanello, P. Eur. J. Inorg. Chem. 2001,
913.
(13) Herberhold, M. Angew. Chem., Int. Ed. 2002, 41, 956.
(14) Boer, E. J. M.; Deuling, H. H.; van der Heijden, H.; Meijboom, N.;
van Oork, A. B. (Shell) WO0158874, 2001.
(1) Gibson, V. C.; Spitzmesser, S. K. Chem. ReV. 2003, 103, 283, and
references therein.
(2) Ittel, S. D.; Johnson, L. K.; Brookhart, M. Chem. ReV. 2000, 100,
1169.
(3) Gibson, V. C.; Wass, D. F. Chem. Br. 1999, 7, 20.
(4) Bennett, A. M. A. Chemtech. 1999, 29, 24.
(5) (a) Small, B. L.; Brookhart, M. J. Am. Chem. Soc. 1998, 120, 7143.
(b) Britovsek, G. J. P.; Mastroianni, S.; Solan, G. A.; Baugh, S. P. D.;
Redshaw, C.; Gibson, V. C.; White, A. J. P.; Williams, D. J.; Elsegood, M.
R. J. Chem. Eur. J. 2000, 6, 2221.
(6) Esteruelas, M. A.; Lopez, A. M.; Mendez, L.; Olivan, M.; Onate, E.
Organometallics 2003, 22, 395. Small, B. L.; Carney, M. J.; Holman, D.
M.; O’Rourke, C. E.; Halfen, J. A. Macromolecules 2004, 37, 4375.
(7) Reardon, D.; Conan, F.; Gambarotta, S.; Yap, G.; Wang. Q. J. Am.
Chem. Soc. 1999, 121, 9318.
(8) Tellmann, K. P.; Gibson, V. C.; White, A. J. P.; Williams, D. J.
Organometallics 2005, 24, 280.
10.1021/om0509589 CCC: $33.50 © 2006 American Chemical Society
Publication on Web 03/07/2006

Redox-ActiVe Catalysts for the Polymerization of Ethylene
Scheme 1. Formation of Compounds 1-12
Organometallics, Vol. 25, No. 8, 2006 1933
1
a redox-active substituent induces a higher electrophilicity and,
therefore, a change in reactivity of a catalytically active transition
metal center. This paper features the synthesis of new ferrocenyl-
modified bis(imino)pyridine ligands, their iron and cobalt
complexes, and their role as precatalysts in olefin polymeriza-
tion.
ligands were air stable and were characterized by H and 13C-
{1H} NMR spectroscopy, microanalysis, and FAB mass spec-
trometry. For example, the 1H NMR spectrum of 4 shows, due
to the mirror plane and the freedom of rotation of the isopropyl
groups, only nine sets of resonances. Of note are the ferrocenyl
resonances (a, b, c) characteristic of a monosubstituted ferro-
cenyl compound and the septet (e) and doublet (f) due to the
isopropyl groups.
Results and Discussion
Iron dichloride complexes of the bis(imino)pyridine ligands,
7-9, were formed by the reaction of a THF solution of the
respective ligand with anhydrous iron(II) chloride at 60 °C for
4 h (Scheme 1). After several hours a dark green precipitate
formed, which could be recrystallized from a dichloromethane/
pentane (30:70) solution to give air-sensitive green crystals of
7-9. The complexes formed in excellent yields, and detailed
analysis was obtained for 7, along with comparative data for 8
The ferrocenyl anilines 1-3 were synthesized by the Negishi
coupling reaction of ferrocenyl zinc chloride with the appropriate
iodoaniline in THF solvent, using Pd(PPh3) as a catalyst
15
(Scheme 1).
The reaction mixtures were stirred at room
temperature overnight to yield brown-purple solutions. The
solvent was then removed under vacuum to leave dark purple
oils, which were purified via column chromatography (silica
gel) to give a red-colored oil (1) or a red waxy solid (2 and 3).
1H NMR spectra include, in each case, a pattern of a singlet
and two pseudotriplet resonances in the ferrocenyl region, which
is indicative of monosubstitution of one of the Cp rings. From
these precursors, the bis(imino)pyridine ligands 4-6 were
conveniently synthesized by the condensation of the respective
ferrocenyl aniline with 2,6-diacetylpyridine (Scheme 1). The
reactions were carried out in refluxing ethanol, using HCl as a
catalyst, and after 3 days an orange precipitate could be isolated
by filtration and then recrystallized from ethanol. The new
1
and 9. Although the complexes are paramagnetic, H NMR
spectroscopy can be informative. A paramagnetic contact-shifted
1H NMR spectrum was obtained for 7, and the resonances may
be assigned on the basis of integration and proximity to the
paramagnetic center and by reference to literature values.9c The
spectrum shows the ketimine protons (at -37.9 ppm), isopropyl
protons (-20.6, -6.05, and -4.66 ppm), ferrocenyl protons
(4.97, 5.08, and 5.32 ppm), aromatic protons (15.34 ppm), and
pyridyl protons (80.99 and 81.79 ppm) (Figure 1).
X-ray diffraction quality crystals of 7 (solvated with dichlo-
romethane) were obtained by the slow diffusion of a saturated
dichloromethane solution of the complex into pentane and
(15) Yamamoto, A.; Morifuji, K.; Ikeda, S.; Saito, T.; Uchida, Y.;
Misono, A. J. Am. Chem. Soc. 1968, 90, 1878.

1934 Organometallics, Vol. 25, No. 8, 2006
Gibson et al.
Figure 1. 1H NMR spectrum and labeling of signals for 7 and 10.
isolated as green hexagonal needles. The structure of 7 is similar
to that of the closely related iron complex dichloro-(2,6-bis(N-
(2,6-diisopropylphenyl)acetylimino)pyridyl)iron(II), FECAT.9c
The molecule has, with the exception of the two ferrocenyl
substituents, approximate Cs symmetry about a plane containing
the metal center, the pyridine nitrogen, and the two chlorine
atoms. The orientations of the ferrocenyl units break this
symmetry, both being rotated in the same sense relative to the
planes of their adjacent 2,6-diisopropylphenyl rings; the torsional
twists about the C(14)-C(35) and C(26)-C(45) bonds are ca.
15° and 24°, respectively. The geometry at iron is severely
distorted square pyramidal, with trans-basal angles of 140.84-
(10)° [N(9),N(7)] and 153.06(8)° [N(1),Cl(2)] and axial/equato-
rial angles lying in the range 90.58(7)° to 116.36(4)°; the ligand
N,N bite angles are 72.95(10)° and 73.34(10)°. The iron atom
lies ca. 0.51 Å out of the ligand N3 plane in the direction of
Cl(1). This out-of-plane geometry is very similar to that
observed in the 2,6-diisopropylphenyl structure (0.56 Å) but in
striking contrast to the near-planar arrangement observed in,
Table 1. Selected Bond Lengths (Å) and Angles (deg) for 7
Fe(1)-Cl(1)
Fe(1)-N(1)
Fe(1)-N(9)
N(9)-C(9)
2.3200(10)
2.085(3)
2.207(3)
1.280(4)
Fe(1)-Cl(2)
Fe(1)-N(7)
N(7)-C(7)
2.2594(10)
2.198(3)
1.279(4)
Cl(1)-Fe(1)-Cl(2)
Cl(1)-Fe(1)-N(7)
Cl(2)-Fe(1)-N(1)
Cl(2)-Fe(1)-N(9)
N(1)-Fe(1)-N(9)
116.36(4)
98.33(8)
153.06(8)
99.37(8)
72.95(10)
Cl(1)-Fe(1)-N(1)
Cl(1)-Fe(1)-N(9)
Cl(2)-Fe(1)-N(7)
N(1)-Fe(1)-N(7)
N(7)-Fe(1)-N(9)
90.57(8)
101.36(8)
101.66(7)
73.34(10)
140.84(10)
for example, the mesityl derivative.9c The Fe-N(pyridine) and
Fe-Cl bond lengths in 7 (Table 1) do not differ significantly
from those observed in the 2,6-diisopropylphenyl structure,
though the Fe-N(imino) bonds are both ca. 0.03 Å shorter;
the reason for this difference is not immediately apparent. The
two ferrocenyl units have different staggered geometries with
stagger angles of ca. 11° for Fe(2) and ca. 25° for Fe(3); their
cyclopentadienyl rings are inclined by ca. 1° and 2°, respec-
tively. The crystal contained two included CH2Cl2 solvent

Redox-ActiVe Catalysts for the Polymerization of Ethylene
Organometallics, Vol. 25, No. 8, 2006 1935
Figure 2. Molecular structure of 7.
Figure 3. Comparison of the cyclic voltammograms of the iron complexes FECAT and 7. Conditions: 0.01 M dichloromethane solution
n
of compound; supporting electrolyte 0.1 M Bu4NPF6. Standard calomel electrode as reference electrode with palladium working and
auxiliary electrodes. Scan rate 100 mV s-1
.
molecules, one ordered, the other disordered. There are no
noteworthy packing interactions.
ferrocenyl protons (3.5, 4.2, and 4.9 ppm), and pyridyl protons
(53.4 and 118.2 ppm). X-ray diffraction quality crystals of
complex 10 were obtained by the slow diffusion of a saturated
dichloromethane solution of 10 into pentane and isolated as red
platy prisms, again with solvation by dichloromethane. The unit
cell of the cobalt complex 10 was determined and found to be
very similar to that of its iron analogue 7, indicating that both
the molecular and extended packing structures in the crystals
of 7 and 10 are essentially the same.
Electrochemistry of Iron and Cobalt Precatalysts: Com-
pounds 7 and 10. Due to its easily accessible and reversible
one-electron redox couple (Fe(II)-Fe(III)), ferrocene is an ideal
compound to study electrochemically. Substituents on the
cyclopentadienyl groups affect the electron density and redox
properties of the ferrocenyl iron center. Herein, the effect on
the central iron(II) or cobalt(II) centers by the ferrocene-sub-
stituted bis(imino)pyridine ligand is examined for 7 and 10 and
Cobalt(II) chloride complexes of ligands 4-6 could be
synthesized in a fashion similar to the corresponding iron(II)
complexes (Scheme 1) by stirring a mixture of ligand with
anhydrous cobalt(II) chloride in THF at 60 °C for 4 h. Following
filtration and washing with cold diethyl ether, the complexes
10-12 were isolated as red-brown solids in good yields and
could be recrystallized from a dichloromethane/pentane (30:
70) solution. As expected, the complexes are paramagnetic and,
1
as for the iron analogues, a contact-shifted H NMR spectrum
was obtained for one example, compound 10. Again the
resonances were assigned on the basis of integration and
proximity to the paramagnetic center and by reference to
1
literature values.9c The H NMR spectrum of 10 (Figure 1)
shows isopropyl protons (at -82.8 and -17.2 ppm), the
ketimine protons (1.6 ppm), aromatic protons (10.8 ppm),

1936 Organometallics, Vol. 25, No. 8, 2006
Gibson et al.
Figure 4. Cyclic voltammetry of the cobalt complexes COCAT and 10. Conditions: 0.01 M dichloromethane solution of compound;
supporting electrolyte 0.1 M nBu4NPF6. Standard calomel electrode as reference electrode with palladium working and auxiliary electrodes.
Scan rate 100 mV s-1
.
Table 2. Cyclic Voltammetry of Bis(imino)pyridine Compounds
a
c
a
c
c
complex
Ep (1) (V)
Ep (1) (V)
E1/2(1) (V)
Ep (2) (V)
Ep (2) (V)
E1/2(2) (V)
Ep (3) (V)
ferrocene
FECAT
7
COCAT
10
0.470
0.550
0.488
0.360
0.473
0.378
0.874
0.373
0.405
0.512
0.433
0.644
0.574
0.200
0.609
-0.624
0.484
0.429
their well-known and catalytically active non-ferrocenyl coun-
terparts FECAT and COCAT. The cyclic voltammograms are
shown as Figures 3 and 4, and results are tabulated in Table 2.
For compound 7, two oxidation and reduction waves are
observed. The initial oxidation is attributed to the ferrocenyl
groups and has a half-potential of 0.433 V. The second wave is
due to the quasi-reversible, one-electron transfer for the Fe-
(II)-Fe(III) redox couple of the central iron atom, with a half-
potential of 0.609 V. In comparison to FECAT the oxidation
potential is shifted by 0.907 V. The ferrocenyl iron atoms are
therefore easier to oxidize than the central iron atom, and the
resulting positive charge means that the central iron atom is
harder to oxidize. Oxidation of the ferrocenyl moieties therefore
changes the electron potential at the central iron atom, and its
effect on polymerization may be observed (see later). In
comparison, the cobalt complexes show a different pattern
(Figure 3). The main feature of the cyclic voltammogram of 10
is the single quasi-reversible Fe(II)-Fe(III) wave of the ferro-
cenyl substituents with a half-potential of 0.429 V. There is
possibly a further oxidation at 1.2 V, but this is at the limit of
the solvent window, and full electrochemical information could
not be obtained. In contrast the unsubstituted cobalt complex
COCAT shows three reduction peaks at 0.874, 0.200, and
-0.624 V. The reduction peaks have been tentatively assigned
to Co(III), Co(III)-Co(II), and Co(II)-Co(I) (Table 2).
Although the cyclic voltammetry studies show that the
oxidation potentials of the ferrocenyl units of the iron and cobalt
complexes are higher than that of ferrocene, the chemical
oxidation of complexes 7 and 10 was easily carried out by the
addition of 2 equiv of ferrocenium hexafluorophosphate in
dichloromethane solution (Scheme 2). The reaction was stirred
at 40 °C for 2 h. A color change of green to khaki was noted
for the iron complex 13 and from orange-brown to khaki-brown
for the cobalt complex 14. The solvent was removed in vacuo,
and the resulting crude solid was further heated under vacuum,
during which time orange crystals of ferrocene were seen to
sublime in the Schlenk tube. The solids were washed with
pentane, dried, and isolated in yields of 46% for 13 and 64%
for 14. The ferrocenium salts were insoluble in most organic
solvents and were metastable in air. They were stored under
dry nitrogen. Characterization (FAB MS and microanalysis)

Redox-ActiVe Catalysts for the Polymerization of Ethylene
Organometallics, Vol. 25, No. 8, 2006 1937
Figure 5. Reactivity profiles for ethylene polymerization in stainless steel reactor vessel.
Scheme 2
solution of activated catalyst at a constant pressure of ethylene
(4 bar). The results of the tests are given in Table 3, and the
reaction profiles are illustrated in Figure 5. Precatalysts FECAT
and COCAT were also tested in order to compare the effects
of the ferrocenyl groups. The nature of the metal center has a
large influence on catalyst productivity. It can be seen that the
iron catalysts show a larger activity (by about an order of
magnitude) than their cobalt equivalents, and this is the same
trend seen for the non-ferrocenyl analogues.9c A modest increase
in activity is seen for 7 relative to the non-ferrocenyl FECAT
precatalyst. which may reflect the greater electron-withdrawing
capacity of the ferrocenyl unit and in turn would lead to an
increase in electrophilicity of the catalytically active iron center.
However, this is not seen for the cobalt system, where a decrease
in activity is found for 10 versus COCAT.
Molecular weight data for 7 and 10 reveal Mw values of
865 000 and 19 000, respectively, with a molecular weight
distribution for the iron catalyst of 15.3 and 3.8 for the cobalt
system. These values are typical of bis(imino)pyridine iron and
cobalt catalysts.9 Interestingly, the oxidation of the ferrocene
moieties appears to have little or no effect on the catalyst activity
since the reactivity profile and activity of the ferrocene (7) and
ferrocenium (13) species are almost identical. Further, the Mw
of the resultant polyethylene using 13, 942 000 (PDI 11.1), is
virtually identical to that obtained using 7. This suggests that
the dicationic ferrocenium species is likely reduced by the
alkylaluminum cocatalyst. In this context, reduction of Fe(III)
to Fe(II) species by alkyl aluminum compounds has been
documented.15 To test this hypothesis, a sample of ferrocenium
hexafluorophosphate was treated with MAO. The blue-colored
solution immediately turned green and, on stirring, turned
orange; it was possible to isolate ferrocene from the reaction
mixture. Thus, to see any effect of ferrocene acting as a redox
switch for ethylene polymerization, a system free of MAO will
be required. Systems amenable to the stabilization of cationic
alkyl species using weakly coordinating anions are expected to
be more suited to this requirement.
Table 3. Summary of Reactor Trials for Ethylene
Polymerization
activity
(g mmol-1
polymer (g) bar-1 h-1
catalyst MAO effective ratio
mass of
catalyst (µmol) (µmol)
MAO:cat.
)
FECAT
1
1
1
4
4
100
100
100
400
400
4100:1
4100:1
4100:1
4100:1
4100:1
23.5
27.6
27.4
7.1
5875
6900
6850
440
7
13
10
COCAT
10.0
630
confirmed that 13 and 14 were the dicationic bis-hexafluoro-
phosphate salts.
Ethylene Polymerization Studies. The ethylene polymeri-
zation activity of the iron and cobalt precatalysts 7 and 10 was
been evaluated in a double-jacketed, 1 L stainless steel auto-
clave. Polymerization was commenced by injection of a toluene

1938 Organometallics, Vol. 25, No. 8, 2006
Gibson et al.
(multiplet, 1H, C6H3), 7.20 (multiplet, 2H, C6H3). FAB MS (+ve):
m/e 290 [M]+.
Conclusions
The synthesis and characterization of a family of iron- and
cobalt-based ethylene polymerization precatalysts featuring
ferrocenyl substituents have been described. As with the non-
ferrocenyl analogues, activation with MAO produces highly
active catalysts with the iron species being an order of magnitude
more active than the Co species.
As expected, the ferrocene-substituted complexes show a rich
electrochemistry, featuring quasi-reversible processes, though
it has not proved possible to establish the effect on the catalytic
activities of the oxidized precatalysts due to the reducing
properties of the activator. Thus, further work is underway to
understand more fully the role and influence of the ferrocenyl
substituents, particularly their use in catalytic systems that do
not require alkyl aluminum activators.
2,6-Diacetylpyridinebis(4-ferrocenyl-2,6-diisopropylaniline)
(4). To a solution of 2,6-diacetylpyridine (0.74 g, 4.5 mmol) in
absolute ethanol (50 mL) was added 4-Fc-2,6-diisopropylaniline
(3.30 g, 9.1 mmol). The mixture was purged with nitrogen for 10
min. After the addition of 2 drops of glacial acetic acid the mixture
was refluxed for 5 days, during which time an orange precipitate
had formed. The reaction mixture was cooled to room temperature
and filtered. The filtrate was washed with cold ethanol and dried
1
to give 4 (1.70 g, 44%). H NMR (CDCl3, ppm): 1.20 (d, 24H,
3JH-H ) 3.3 Hz, CHMe2), 2.31 (s, 6H, CH3), 2.79 (sept, 4H, 3JH-H
) 3.3 Hz, CHMe2), 4.07 (s, 10H, C5H5), 4.29 (t, 4H, C5H4), 4.63
3
(t, 4H, C5H4), 7.30 (s, 4H, C6H2), 7.94 (t, 1H, JH-H ) 3.8 Hz,
C5H3-py), 8.5 (d, 2H, C5H3-py). 13C NMR (CDCl3, ppm): 17.2
(CHMe2), 23.4 (CHMe2), 28.2 (NdCMe), 66.6 (C5H5-Cp), 68.3
(C5H4-Cp), 69.5 (C5H4-Cp), 87.5 (ipso-C C5H4-Cp), 121.4 (C6H2),
122.1 (C5H3-py), 133.8 (C5H3-py), 135.6 (C6H2), 136.9 (C6H2),
155.2 (C6H2), 161.8 (C5H3-py), 167.1 (NdC). FAB MS (+ve): m/e
849 [M]+. Anal. Calcd for C53H59Fe2N3: C, 74.91; H, 6.95; N,
4.94. Found: C, 74.78; H, 7.15; N, 4.82.
Experimental Section
General Details. All preparations were carried out under an
atmosphere of nitrogen using standard Schlenk techniques or an
inert atmosphere (nitrogen) glovebox, unless otherwise stated. All
solvents were distilled over standard drying agents under nitrogen
directly before use and deoxygenated. Silica gel (Kieselgel grade
60) was used for chromatographic separations. All NMR spectra
were recorded using a Delta upgrade on a JEOL EX270 MHz
spectrometer. Chemical shifts are reported in δ (ppm) using residual
proton impurities in CDCl3 (1H δ 7.25 ppm, 13C δ 77.0 ppm) or
CD2Cl2 (1H δ 5.32 ppm, 13C δ 53.8 ppm) as the reference solvent.
Mass spectra were recorded using positive FAB methods for
organometallic compounds, and EI for organic compounds, on a
micromass Autospec spectrometer (primary ion beam 35 keV Cs+
and 3-nitrobenzyl alcohol matrix) by Mr. J. Barton, Imperial
College. Microanalyses were carried out by Mr. S. Boyer of the
London Metropolitan University.
2,6-Diacetylpyridinebis(4-ferrocenyl-2,6-dimethylaniline) (5).
A procedure analogous to that described for 4 was followed but
using 2,6-diacetylpyridine (0.25 g, 1.6 mmol) and 4-ferrocenyl-
2,6-dimethylaniline (1.0 g, 3.2 mmol) to yield 5 (0.41 g, 34%). 1H
NMR (CDCl3, ppm): 2.06 (s, 12H, CH3), 2.28 (s, 6H, CH3), 4.08
(s, 10H, C5H5), 4.21 (t, 4H, C5H4), 4.62 (t, 4H, C5H4), 7.20 (s, 4H,
C6H2), 7.93 (t, 1H, 3JH-H ) 3.8 Hz, C5H3-py), 8.50 (d, 2H, C5H3-
py). 13C NMR (CDCl3, ppm): 16.6 (Ar-Me), 18.1 (NdCMe), 66.1
(C5H5-Cp), 68.4 (C5H4-Cp), 69.5 (C5H4-Cp), 86.0 (ipso-C C5H4-
Cp), 122.2 (C6H2), 122.4 (C5H3-py), 125.8 (C6H2), 133.5 (C6H2),
136.8 (C6H2), 147.0 (C5H3-py), 155.2 (C5H3-py), 167.3 (NdC).
FAB MS (+ve): m/e 737 [M]+. Anal. Calcd for C45H43Fe2N3: C,
73.27; H, 5.83; N, 5.69. Found: C, 73.16; H, 6.04; N, 5.48.
2,6-Diacetylpyridinebis(4-ferrocenyl-2-methylaniline) (6). A
procedure analogous to that described for 4 was followed but using
2,6-diacetylpyridine (0.55 g, 3.4 mmol) and 4-ferrocenyl-2-methy-
Polyethylene analysis was performed at BP Chemicals Limited
by J. Boyle (NMR) and G. Audley (GPC). GPC traces were
recorded using PL gel 2 x mixed bed-D, 30 cm, 5 µm columns,
trichlorobenzene eluent, and a flow rate of 1.0 mL/min at 150 °C
using a refractive index detector. Starting materials were prepared
according to adapted literature procedures (referenced where
1
laniline (2.0 g, 6.85 mmol) to yield 6 (1.51 g, 62%). H NMR
(CDCl3, ppm): 2.17 (s, 6H, CH3), 2.41 (s, 6H, CH3), 4.07 (s, 10H,
C5H5), 4.30 (t, 4H, C5H4), 4.64 (t, 4H, C5H4), 6.6 (m, 2H, C6H3),
7.33 (m, 4H, C6H3), 7.92 (t, 1H, 3JH-H 3.8, C5H3-py), 8.40 (d, 2H,
C5H3-py). 13C NMR (CDCl3, ppm): 16.4 (Ar-Me), 17.9 (NdCMe),
66.1 (C5H5-Cp), 68.6 (C5H4-Cp), 69.5 (C5H4-Cp), 85.8 (ipso-C
C5H4-Cp), 118.3 (C6H3), 122.2 (C5H3-py), 123.6 (C6H3), 124.2
(C6H3), 127.3 (C6H3), 128.0 (C6H3), 134.2 (C6H3), 147.9 (C5H3-
py), 155.4 (C5H3-py), 166.8 (NdC). FAB MS (+ve): m/e 709
[M]+. Anal. Calcd for C43H39Fe2N3‚1.25CH2Cl2: C, 65.15; H, 5.09;
N, 5.15. Found: C, 65.00; H, 5.04; N, 4.92.
appropriate) and were characterized by H NMR and 13C NMR
1
and mass spectrometry where appropriate. All other chemicals were
purchased from commercial sources and used without further
purification unless stated. Research grade ethylene (BOC) was used
for all ethylene polymerization experiments. Methyl aluminoxane
(MAO) was obtained as a 1.6 M solution in toluene. 4-Fc-2,6-
diisopropylaniline (1) was prepared using the method of Siemel-
(2,6-Diacetylpyridinebis(4-ferrocenyl-2,6-diisopropylaniline))-
FeCl2 (7). A solution of 4 (0.8 g, 0.94 mmol) in THF (30 mL) was
added to a solution of FeCl2 (0.12 g, 0.94 mmol) also in THF (30
mL) to yield a dark green solution. The mixture was stirred for 4
h, during which time a dark green solid had precipitated. The
mixture was filtered and the residual green solid washed with cold
diethyl ether (2 × 10 mL). The solid was dried to give a dark green
ing,16 and NaBAF,
FcBA,18 FECAT,9 and COCAT9 were
17
prepared using literature procedures.
Synthesis of 4-Ferrocenyl-2,6-dimethylaniline (2). A procedure
analogous to that described by Siemeling16 for 1 was followed but
using iodo-2,6-dimethylaniline to yield 2 (72%) as an waxy red
solid. H NMR (CDCl3, ppm): 2.21(s, 6H, CH3), 3.54 (br s, 2H,
NH2), 4.04 (s, 5H, C5H5), 4.22 (t, 2H, C5H4), 4.52 (t, 2H, C5H4),
7.08 (s, 2H, C6H2). FAB MS (+ve): m/e 304 [M]+.
Synthesis of 4-Ferrocenyl-2-methylaniline (3). A procedure
analogous to that described by Siemeling16 for 1 was followed but
using 1-iodo-2-methylaniline to yield 3 (54%) as a waxy red solid.
1H NMR (CDCl3, ppm): 2.21(s, 3H, CH3), 3.57 (br s, 2H, NH2),
4.04 (s, 5H, C5H5), 4.23 (t, 2H, C5H4), 4.54 (t, 2H, C5H4), 6.63
1
1
powder (0.74 g, 80%). H NMR (CD2Cl2 broad signals observed,
ppm): -37.9 (6H, NdC(Me)), -20.6 (4H, CH(CH3)2), -6.05 (12H,
CH(CH3)2), -4.66 (12H, CH(CH3)2), 4.97 (4H, C5H4-Cp), 5.08
(10H, C5H5-Cp), 5.32 (4H, C5H4-Cp), 15.34 (4H, Ar-Hm), 80.98
(1H, py-Hp), 81.79 (2H, py-Hm). FAB MS (+ve): m/e 975 [M]+.
Anal. Calcd for C53H59Cl2Fe3N3: C, 65.16; H, 6.04; N, 4.30.
Found: C, 64.92; H, 5.93; N, 4.20.
(2,6-Diacetylpyridinebis(4-ferrocenyl-2,6-dimethylaniline))-
FeCl2 (8). A procedure analogous to that described for 7 was
followed but using 5 (0.08 g, 0.1 mmol) and FeCl2 (0.012 g, 0.01
(16) Siemeling, U.; Neumann, B.; Stammler, H.-G.; Kuhnert, O.
Polyhedron 1999, 18, 1815.
(17) Brookhart, M.; Grant, B.; Volpe, A. F. Organometallics 1992, 11,
3920.
1
mmol) to yield 8 (0.07 g, 76%). H NMR (CD2Cl2 broad signals
(18) Heinekey, D. M.; Radzewich, C. E. Organometallics 1998, 17, 51.
observed, ppm): -26.27 (6H, NdC(Me)), 4.61 (10H, C5H5-Cp),

Redox-ActiVe Catalysts for the Polymerization of Ethylene
Organometallics, Vol. 25, No. 8, 2006 1939
4.81 (4H, C5H4-Cp), 5.32 (4H, C5H4-Cp), 12.84 (12H, Ar-CH3),
16.88 (4H, Ar-Hm), 35.27 (1H, py-Hp), 82.38 (2H, py-Hm). FAB
MS (+ve): m/e 863 [M]+. Anal. Calcd for C45H43Cl2Fe3N3: C,
62.51; H, 4.97; N, 4.86. Found: C, 62.49; H, 5.03; N, 4.85.
(2,6-Diacetylpyridinebis(4-ferrocenyl-2-methylaniline))-
FeCl2 (9). A procedure analogous to that described for 7 was
followed but using 6 (0.23 g, 0.32 mmol) and FeCl2 (0.04 g, 0.31
mmol) to yield 9 (0.19 g, 70%). FAB MS (+ve): m/e 835 [M]+.
Anal. Calcd for C43H39Cl2Fe3N3: C, 61.70; H, 4.66; N, 5.02.
Found: C, 61.69; H, 4.72; N, 4.85.
Cl2CoFe2N3P2F12: C, 50.10; H, 4.60; N, 3.31. Found: C, 49.60;
H, 5.00; N, 2.85.
Crystal data for 7: C53H59N3Cl2Fe3‚2CH2Cl2, M ) 1146.33,
monoclinic, P21/c (no. 14), a ) 18.501(2) Å, b ) 13.7329(12) Å,
c ) 21.937(2) Å, â ) 91.093(9)°, V ) 5572.5(9) Å3, Z ) 4, Dc )
1.366 g cm-3, µ(Mo KR) ) 1.096 mm-1, T ) 203 K, green
hexagonal needles; 9771 independent measured reflections, F2
refinement, R1 ) 0.046, wR2 ) 0.107, 7405 independent observed
absorption-corrected reflections [|Fo| > 4σ(|Fo|), 2θmax ) 50°], 616
parameters. CCDC 600245.
(2,6-Diacetylpyridinebis(4-ferrocenyl-2,6-diisopropylaniline))-
CoCl2 (10). A solution of 4 (0.40 g, 0.47 mmol) in 30 mL of THF
was added to a solution of CoCl2 (0.06 g, 0.47 mmol) also in THF
(30 mL) to yield a red-brown solution. The mixture was stirred for
4 h, during which time a dark brown solid had precipitated. The
mixture was filtered and the residual brown solid washed with cold
diethyl ether (2 × 10 mL). The solid was dried to give a red-brown
powder (0.35 g, 78%). 1H NMR (CD2Cl2, ppm): -82.8 (6H,
CH(CH3)2), -17.2 (12H, CH(CH3)2), 1.6 (6H, NdC(Me)), 3.5 (4H,
C5H4-Cp), 4.2 (5H, C5H5-Cp), 4.9 (4H, C5H4-Cp), 10.8 (4H Ar-
Hm), 53.4 (1H, py-Hp), 118.2 (2H, py-Hm). FAB MS (+ve): m/e
978 [M]+. Anal. Calcd for C53H59Cl2CoFe2N3: C, 64.96; H, 6.02;
N, 4.29. Found: C, 64.85; H, 5.81; N, 4.12.
(2,6-Diacetylpyridinebis(4-ferrocenyl-2,6-dimethylaniline))-
CoCl2 (11). A procedure analogous to that described for 10 was
followed but using 5 (0.55 g, 0.77 mmol) and CoCl2 (0.10 g, 0.71
mmol) to yield 11 (0.40 g, 62%). FAB MS (+ve): m/e 865 [M]+.
Anal. Calcd for C45H43Cl2CoFe2N3: C, 62.35; H, 4.96; N, 4.85.
Found: C, 62.13; H, 4.99; N, 4.62.
(2,6-Diacetylpyridinebis(4-ferrocenyl-2-methylaniline))-
CoCl2 (12). A procedure analogous to that described for 10 was
followed using 6 (0.23 g, 0.32 mmol) and CoCl2 (0.04 g, 0.31
mmol) to yield 12 (0.17 g, 62%). FAB MS (+ve): m/e 837 [M]+.
Anal. Calcd for C43H39Cl2CoFe2N3: C, 61.59; H, 4.59; N, 4.97.
Found: C, 61.52; H, 4.65; N, 5.00.
(2,6-Diacetylpyridinebis(4-ferrocenium-2,6-diisopropylaniline))-
FeCl2 (PF6)2 (13). To a suspension of ferrocenium hexafluoro-
phosphate (0.12 g, 0.36 mmol) in dichloromethane (25 mL) was
added a solution of 7 (0.17 g, 0.18 mmol) in dichloromethane. The
mixture was stirred for 4 h, and the volatiles were removed in vacuo.
The resulting green solid was washed with pentane (25 mL) and
dried to yield 13 (0.21 g 46%). FAB MS (+ve): m/e 1120 [M -
PF6]+. Anal. Calcd for C53H59Cl2Fe3N3P2F12: C, 50.20; H, 4.66;
N, 3.32. Found: C, 50.17; H, 4.54; N, 3.19.
Crystal data for 10: C53H59N3Cl2Fe2Co‚2CH2Cl2, M ) 1149.41,
monoclinic, P21/c (no. 14), a ) 18.527(2) Å, b ) 13.7570(14) Å,
c ) 21.911(2) Å, â ) 91.063(9)°, V ) 5583.7(12) Å3, Z ) 4, Dc
) 1.370 g cm-3, µ(Mo KR) ) 1.133 mm-1, T ) 203 K, red platy
prisms (only unit cell determined).
Cyclic Voltammetry. In all cases measurements were performed
using an Autolab PGSTAT12 potentiometer with Gpes software,
with 0.1 M tetrabutylammonium-hexafluorophophate as the elec-
trolyte in dichloromethane. The sample solution was approximately
1 mM concentration in this electrolyte. The working electrode was
platinium, circular and 1 mm in diameter, the reference was an
Ag/AgCl electrode, and the counter electrode was a platinum wire.
The cyclic voltammograms were measured at rates of 100 mV s-1
and the potentials calculated using ferrocene as a comparison.
Ethylene Polymerization Tests. A solution of the catalyst was
introduced to the reactor though the injector unit by an overpressure
of nitrogen. Constant ethylene pressure was maintained throughout
the reaction via a pressure regulator connected to a mass flow
controller, and constant temperature was maintained by the circula-
tion of thermostatically controlled oil. The ethylene flow rate,
polymerization temperature, and autoclave pressure were all logged
by computer. For a polymerization test the reactor was first heated
to 85 °C for 1 h under a continuous flow of nitrogen, to remove
atmospheric oxygen and moisture. The reactor was cooled to
reaction temperature (50 °C) and the isobutane solvent (0.5 l) added.
An oxygen/moisture scavenger (MAO) was added to the reactor
followed by the desired pressure of ethylene and the solvent/
scavenger mix stirred for 30 min. The test was terminated by venting
of the volatile components. The reactor contents were isolated,
washed with aqueous hydrochloric acid in methanol, and dried under
vacuum.
Acknowledgment. The Department of Chemistry, Imperial
College London, is thanked for a Teaching Assistantship (for
P.J.O.), and BP Amoco Chemicals Ltd. is thanked for GPC
measurements and analysis.
(2,6-Diacetylpyridinebis(4-ferrocenium-2,6-diisopropylanil))-
CoCl2(PF6)2 (14). A procedure analogous to that described for 13
was followed but using ferrocenium hexafluorophosphate (0.07 g,
0.21 mmol) and 10 (0.1 g, 0.11 mmol) to yield 14 (0.09 g, 64%).
FAB MS (+ve): m/e 1123 [M - PF6]+. Anal. Calcd for C53H59-
OM0509589