Cationic 2,6-bis(imino)pyridine iron and cobalt complexes: Synthesis, structures, ethylene polymerisation and ethylene/polar monomer co-polymerisation studies

Article abstract of DOI:10.1039/b106614p

The synthesis and characterisation of a series of cationic bis(imino)pyridine iron and cobalt complexes of the type [LMCl(D)]SbF₆ (D = CH₃CN or thf) and [LM(R₂acac)]SbF₆ (R = CH₃, CF₃, Ph) are described {L = 2,6-bis[1-(2,6diisopropylphenylimino)ethyl]pyridine, M = Fe or Co}. The solid state structures of these five-coordinate complexes vary between square-based pyramidal and trigonal bipyramidal geometries, depending on the ligands used. Attempts to synthesise cationic metal alkyl complexes of the type [LMR]⁺ were unsuccessful. However, these complexes serve as highly active ethylene polymerisation catalysts when activated with MAO. Polymerisation activities are comparable to the activities obtained with neutral dichloride precursors [LMCl₂] and the resulting polymer properties are nearly identical, suggesting that in all cases the same active species is being generated. The polymerisation activity is not inhibited by the presence of donors such as thf or CH₃CN and these cationic precursors can be activated with less co-catalyst than is normally used for neutral dichloride pre-catalysts. As little as 10 equiv. TMA, in combination with 1 equiv. B(C₆F₅)₃, afford a highly active polymerisation system. Co-polymerisation studies of ethylene with polar monomers such as methyl methacrylate (MMA) or styrene resulted in polymer production with high activities. However, in both cases no co-polymer is obtained. The activity of the catalyst is significantly reduced in the presence of methyl acrylate (MA) or 2-vinyl-1,3-dioxolane (VDO) and again no co-polymer is produced. Polar monomers such as vinyl acetate, acrolein and acrylonitrile deactivate the catalyst.

Full text of DOI:10.1039/b106614p

View Article Online / Journal Homepage / Table of Contents for this issue
Cationic 2,6-bis(imino)pyridine iron and cobalt complexes:
synthesis, structures, ethylene polymerisation and ethylene/polar
monomer co-polymerisation studies†
George J. P. Britovsek, Vernon C. Gibson,* Stefan K. Spitzmesser, Kilian P. Tellmann,
Andrew J. P. White and David J. Williams
Department of Chemistry, Imperial College, Exhibition Road, London, UK SW7 2AY.
E-mail: v.gibson@ic.ac.uk
Received 23rd July 2001, Accepted 13th December 2001
First published as an Advance Article on the web 20th February 2002
The synthesis and characterisation of a series of cationic bis(imino)pyridine iron and cobalt complexes of the type
[LMCl(D)]SbF6 (D = CH3CN or thf ) and [LM(R2acac)]SbF6 (R = CH3, CF3, Ph) are described {L = 2,6-bis[1-(2,6-
diisopropylphenylimino)ethyl]pyridine, M = Fe or Co}. The solid state structures of these five-coordinate complexes
vary between square-based pyramidal and trigonal bipyramidal geometries, depending on the ligands used. Attempts
to synthesise cationic metal alkyl complexes of the type [LMR]ϩ were unsuccessful. However, these complexes serve
as highly active ethylene polymerisation catalysts when activated with MAO. Polymerisation activities are comparable
to the activities obtained with neutral dichloride precursors [LMCl2] and the resulting polymer properties are nearly
identical, suggesting that in all cases the same active species is being generated. The polymerisation activity is not
inhibited by the presence of donors such as thf or CH3CN and these cationic precursors can be activated with less
co-catalyst than is normally used for neutral dichloride pre-catalysts. As little as 10 equiv. TMA, in combination
with 1 equiv. B(C6F5)3, afford a highly active polymerisation system. Co-polymerisation studies of ethylene with
polar monomers such as methyl methacrylate (MMA) or styrene resulted in polymer production with high activities.
However, in both cases no co-polymer is obtained. The activity of the catalyst is significantly reduced in the presence
of methyl acrylate (MA) or 2-vinyl-1,3-dioxolane (VDO) and again no co-polymer is produced. Polar monomers
such as vinyl acetate, acrolein and acrylonitrile deactivate the catalyst.
Introduction
During the past five years, there have been a number of
significant advances in the development of olefin polymeris-
ation catalysts. Early transition-metal Ziegler–Natta systems
and chromium based catalysts, both exploited commercially
over many years, have been joined recently by several families
of highly active late-transition metal catalysts based, for
example, on iron, cobalt and nickel.1–3 Amongst these new
systems, the bis(imino)pyridine iron and cobalt systems, dis-
covered independently by Brookhart, Bennett and ourselves,4–7
Fig. 1 Schematic representation of the dichloro (A), monochloro (B)
have attracted interest, both in academia and particularly in
industry,8,9 because of 1) their exceptionally high activities for
ethylene polymerisation,10 2) their remarkable selectivities for
ethylene oligomerisation to linear α-olefins11,12 and 3) their
suitability for adaptation to existing commercial processes.
Various modifications to the original bis(imino)pyridine ligand
framework have been reported, but have generally afforded less
active systems.13–21
The increased poison resistance of late metal systems has
fuelled the idea that these catalysts may be suited to the co-
polymerisation of hydrocarbon olefins with polar monomers.22
Noteworthy in this context is the work by Yamamoto from
the early 1970s, who showed that related late-transition metal
complexes such as (bipy)2FeEt2 are active catalysts for the
homo-polymerisation of polar monomers. A range of different
monomers, such as acrylonitrile (AN), methyl vinyl ketone
(MVK) and methyl methacrylate (MMA) were polymerised,
though generally with poor conversions.23–25
and acac complexes (C) (M = Fe or Co).
Bis(imino)pyridine iron() and cobalt() dihalide complexes
(A, Fig. 1) are commonly activated by methyl aluminoxane
(MAO) or other alkyl aluminium co-catalysts.26,27 As for other
late and early transition metal systems, the exact role of MAO
as a co-catalyst is not fully understood, but in general in situ
generated cationic metal alkyl species are proposed to be the
active sites in these catalyst systems. Theoretical investigations
have been based on the involvement of such species in the
polymerisation mechanism,28,29 but no experimental evidence
for the presence of cationic alkyl species has been obtained to
date. A particular difficulty in the synthesis of well-defined
cationic alkyl complexes of iron appears to be the instability of
potential iron() dialkyl precursors.30,31 Stable iron() dimethyl
or diethyl compounds are rare in organo-iron chemistry, the
32
two six-coordinate iron() complexes (dmpm)2FeMe2 and
(bipy)2FeEt2 33 being the only examples that have been structur-
ally characterised. In addition, the five-coordinate complex
(PPh3)3FeMe2 was reported by Yamamoto and co-workers to
decompose above Ϫ10 ЊC.34 Attempts to obtain LFeR2 species
via treatment of LFeCl2 with a range of main group alkyl
† Electronic supplementary data (ESI) available: the molecular struc-
ture of 4. See: http://www.rsc.org/suppdata/dt/b1/b106614p/
DOI: 10.1039/b106614p
J. Chem. Soc., Dalton Trans., 2002, 1159–1171
This journal is © The Royal Society of Chemistry 2002
1159

View Article Online
reagents usually afforded the coupled R–R product, most likely
via reductive elimination from an unstable LFeR2 species. Such
reactivity is well-precedented for iron() dialkyls.35 Cationic
alkyl derivatives based on the heavier second row metals,
ruthenium and rhodium, have been synthesised recently, but
their lack of polymerisation activity severely limits their useful-
ness in the study of the polymerisation mechanism.36
In order to address the co-polymerisation of hydrocarbon
olefins with polar monomers it is desirable to reduce, or prefer-
ably eliminate, the requirement for an aluminium activator
which makes the cationic alkyl species an important objec-
tive. We decided, therefore, to target cationic pre-catalysts of
iron and cobalt containing bis(imino)pyridine ligands with
readily displaceable co-ligands which, on the one hand, would
allow problematic LFeR2 to be avoided, and on the other hand,
could allow access to the desired cationic mono-alkyl species
via treatment of LFeXϩ with a small amount of a suitable
alkylating agent or activator.
Here we describe our results on the synthesis and character-
isation of cationic bis(imino)pyridine iron and cobalt com-
plexes containing either one chloro substituent and a neutral
donor D (B, Fig. 1) or a bidentate monoanionic acetylacetonate
(acac) ligand (C).37 The catalytic activity of these complexes as
catalysts for the polymerisation of ethylene with MAO and
other co-catalysts has been examined and compared with the
neutral dichloride pre-catalysts. In addition, co-polymerisation
experiments of ethylene with polar monomers have been
carried out.
Scheme 1
can also be prepared by sequential addition of AgSbF6 and
Ag(acac) in acetonitrile solution in one pot, without isolating
the intermediate cationic monochloride complexes. Introducing
differently substituted acetylacetonates requires a slight change
in the reaction procedure: silver salts of 1,3-diphenyl- and
1,3-bis(trifluoromethyl)-acetylacetonate are not commercially
available and therefore sodium salts of the desired substituted
acetylacetone were employed in the reaction with the mono-
chloro complex 2a to obtain the substituted acac complexes 6
and 7 in good yields. Other counter ions can be introduced by
the addition of Ag(acac), followed by the sodium salt of the
tetraphenylborate anion BPh4Ϫ in acetonitrile to afford complex
8. Initial reaction of the dichloride complex 1a with NaBPh4
did not yield a cationic complex and starting material was
recovered. This observation suggests the necessity for a silver
reagent in at least one of the halide abstractions.
Results and discussion
Synthesis and characterisation of cationic complexes
Initially, dialkyl iron complexes containing bis(imino)pyridine
ligands were targeted. Attempts to alkylate bis(imino)pyridine
iron dichloride 1a with Grignard reagents or alkyl lithium com-
pounds generally yielded decomposition products, probably as
a result of reductive processes.35 The reaction between bis-
(imino)pyridine cobalt dichloride 1b and 2 equiv. of MeMgCl
also led to reduction, in this case to give a relatively stable
bis(imino)pyridine cobalt() methyl complex, which is the
subject of a separate report.38
In order to avoid potential reductive decomposition path-
ways available to dialkyl iron() species, cationic monochloro
iron() complexes of type B (Fig. 1) were targeted, with the aim
of generating monoalkyl iron() complexes therefrom. The
reaction of bis(imino)pyridine iron and cobalt dichloride 1a
and 1b with one equiv. of AgSbF6 in acetonitrile or tetra-
hydrofuran results in the clean formation of the cationic mono-
chloro iron and cobalt complexes, containing one acetonitrile
(2a and 2b) or thf ligand (3) coordinated to the metal centre
(Scheme 1). It is interesting to note that in these complexes
the acetonitrile or thf ligand occupies the apical position of
the square based pyramid, in contrast to a closely related
bis(2,6-iminophosphoranyl)pyridine cobalt complex reported
recently.21
Several attempts have been made to generate a donor-free
four-coordinate monochloro iron() complex. For example,
when the reaction with AgSbF6 was carried out in dichloro-
methane, an orange–red solid was obtained. X-Ray analysis of
the crystallised product (CH2Cl2/pentane) revealed a five-
coordinated complex 4 containing one molecule of H2O bound
in the apical position (see ESI for a Figure of this complex).
The scavenging of water, which probably originated from
adventitious water in the solvent, is in keeping with the rare
occurrence and high reactivity of four-coordinate square planar
iron() complexes.39
All compounds were characterised by 1H NMR, UV/VIS and
IR spectroscopy, FAB mass spectrometry, magnetic suscepti-
bility and microanalysis. In addition, the crystal structures of
complexes 2a, 2b, 3, 4, 5a, 6 and 7 were determined by single-
crystal X-ray diffraction studies.
X-Ray crystallographic analysis. Complexes 2a and 2b are
isomorphous but with small differences in the included solvent
molecules. Compounds 4 and 6 were both found to crystallise
as different polymorphic forms, with monoclinic and triclinic
polymorphs being observed in each case: these different poly-
morphs have been denoted as 4i/4ii and 6i/6ii respectively. The
complexes fall into two distinct groups, those bearing, in
addition to the tridentate 2,6-bis(imino)pyridine ligand, two
monodentate substituents on the metal centre (2a, 2b, 3 and 4),
and those with a bidentate ligand occupying these two sites
(5a, 6 and 7). Representative illustrations of the structures of
compounds 2a, 3 and 5a are given in Figs. 2, 3 and 4. Fig. 5
provides a schematic diagram for the purpose of identifying the
principal geometric and conformational features of the series
of structures; comparative data relating to this figure are given
in Table 1.
Compounds 2a, 2b, 3 and both polymorphs of 4 all have
approximate molecular Cs symmetry about the X–M–Y plane
that also contains the pyridyl nitrogen atom N(1) (Fig. 5). The
geometry at the metal centre is, in all cases, distorted square
pyramidal with the metal atom lying ca. 0.35 Å out of the basal
The second chloride ligand can be abstracted with silver
acetylacetonate [Ag(acac)] to afford the donor-free iron and
cobalt acac complexes of type C (Fig. 1). Complexes 5a and 5b
1160
J. Chem. Soc., Dalton Trans., 2002, 1159–1171

View Article Online
Fig. 5 Schematic diagram identifying the principal geometric and
conformational parameters in structures 2a, 2b, 3, 4i, 4ii, 5a, 6i, 6ii
and 7.
than those to the imine nitrogens N(7) and N(9). The geometry
and conformation of the tridentate ligands are, in each case,
very similar with a retention of the double bond character for
the C(7)–N(7) and C(9)–N(9) bonds, and an approximately
orthogonal orientation of the two 2,6-diisopropylphenyl rings
to the basal coordination plane (δ and γ in Fig. 5). This con-
formation is stabilised, as has been observed previously,13,40,41
by C–H ؒ ؒ ؒ N(pπ) interactions (f in Fig. 5). The separations of
the “upper” and “lower” isopropyl carbon atoms, C(20) and
C(29) [d], and C(17) and C(32) [e] respectively, expand and con-
tract depending on the steric bulk of the apical ligand X; the
most pronounced difference between the two separations is in 3
where X = thf.
Fig. 2 The molecular structure of 2a.
The introduction of a bidentate ligand in the X and Y
positions (5a, 6 and 7) results in a change in the coordination
geometry at the metal to one that is intermediate between
square pyramidal and trigonal bipyramidal with N(7)–M–N(9)
as the “trigonal” axis. The nature of the substituents on the
β-diketonate moiety also has a dramatic effect on the resulting
geometry of the complex. In complexes 5a and 7 approximate
molecular Cs symmetry is retained, though in 5a there is a
ca. 11Њ out of plane fold about the O(1) ؒ ؒ ؒ O(2) vector of the
β-diketonate ligand. In both these structures the oxygen atoms
X [O(1)] and Y [O(2)] occupy, in the context of a severely dis-
torted square pyramidal geometry, pseudo-axial and basal sites
respectively with Y positioned approximately centrally between
the planes of the two 2,6-diisopropylphenyl rings (Figs. 6 and 7).
Fig. 3 The molecular structure of 3.
Fig. 4 The molecular structure of 5a.
coordination plane (N3Y) in the direction of the apical ligand
X. The principal distortion from square pyramidal geometry is
due to the coordination constraints of the tridentate 2,6-
bis(imino)pyridine ligand which results in trans-basal N(7)–M–
N(9) angles that are in the range 142.7(3)–145.2(3)Њ. The
equivalent metal–nitrogen bond lengths in these five structures
are, within statistical significance, the same (Table 2), with in
each case that to the pyridyl nitrogen atom N(1) being shorter
Fig. 6 View down the N(7) ؒ ؒ ؒ N(9) vector of the structure of 5a
showing the distorted coordination geometry and the positioning of
O(2) between the faces of the two 2,6-diisopropylphenyl rings.
J. Chem. Soc., Dalton Trans., 2002, 1159–1171
1161

View Article Online
Table 1 Comparative conformational and geometric parameters for compounds 2a, 2b, 3, 4i, 4ii, 5a, 6i, 6ii and 7
aa
ba
cb
θc
φc
ζd
de
ee
αf
βf
γf
δf
fg
ψh
2a
2b
3
4i
4ii
5a
6i
6ii
7
0.35
0.32
0.34
0.38
0.34
0.46
0.34
0.36
0.49
0.01
0.05
0.06
0.03
0.06
0.12
0.29
0.22
0.11
0.46
0.39
0.31
0.19
0.36
0.42
Ϫ0.17
Ϫ0.09
0.28
87
86
90
86
89
120
127
127
122
161
165
165
161
164
147
146
146
147
90
89
89
86
88
84
76
77
89
7.08
6.96
7.60
7.00
7.28
7.31
6.65
6.70
6.91
5.45
5.31
4.79
5.43
5.13
5.35
6.56
6.50
5.49
7
4
5
3
3
12
2
2
7
4
2
3
4
9
1
2
6
89
88
83
84
89
86
82
83
82
89
89
89
89
85
78
72
74
84
2.33–2.41
2.34–2.45
2.40–2.51
2.39–2.43
2.38–2.43
2.40–2.56
2.46–2.54
2.46–2.54
2.42–2.50
109–114
110–114
110–111
110–112
111–112
102–111
107–111
107–110
106–112
5
a a is the deviation (Å) of the metal atom from the N3Y plane, which is planar to within b Å. b c is the displacement of the metal atom out of the plane
of the pyridyl ring A towards X, a negative value indicates that the metal lies below this plane. c θ and φ are the N(1)–M–X and N(1)–M–Y angles (Њ),
respectively. d ζ is the dihedral angle (Њ) between the MN3 and MXY planes. e d and e are the separations (Å) of the isopropyl carbon atoms C(20)
and C(29), and C(17) and C(32), respectively. f α, β, γ and δ are the mean torsional twists (Њ) about the C(2)–C(7), C(6)–C(9), N(7)–C(11) and
N(9)–C(23) bonds, respectively. g f is the range (Å) for the isopropyl methine hydrogen H ؒ ؒ ؒ N distances. h ψ is the range (Њ) of C–H ؒ ؒ ؒ N angles for
the contact f.
Table 2 Comparative selected bond lengths (Å) and angles (Њ) for compounds 2a, 2b, 3, 4i, 4ii, 5a, 6i, 6ii and 7
2aa
2bb
3c
4id
4iid
5ae
6ie
6iie
7e
M–X
M–Y
M–N(1)
M–N(7)
M–N(9)
C(7)–N(7)
C(9)–N(9)
2.109(12)
2.231(3)
2.110(7)
2.221(8)
2.223(8)
1.275(12)
1.283(12)
2.049(11)
2.218(3)
2.052(8)
2.188(8)
2.174(9)
1.259(13)
1.281(14)
2.079(4)
2.228(1)
2.056(4)
2.205(4)
2.171(3)
1.285(6)
1.289(6)
2.121(6)
2.230(3)
2.072(8)
2.217(7)
2.227(7)
1.274(12)
1.236(11)
2.070(8)
2.237(3)
2.092(7)
2.222(7)
2.210(7)
1.284(12)
1.286(12)
1.960(5)
1.940(4)
2.101(5)
2.247(5)
2.211(6)
1.274(9)
1.261(9)
2.026(5)
2.006(5)
2.102(6)
2.222(6)
2.230(6)
1.287(10)
1.285(10)
2.027(7)
2.015(6)
2.109(7)
2.220(7)
2.223(7)
1.297(11)
1.280(11)
1.959(3)
1.960(3)
2.109(4)
2.205(4)
2.201(4)
1.279(6)
1.275(6)
X–M–Y
112.3(3)
87.0(3)
94.9(3)
95.3(4)
160.7(3)
103.6(2)
104.1(2)
73.2(3)
73.0(3)
144.1(3)
109.4(3)
85.9(4)
95.8(3)
97.2(4)
164.6(3)
102.6(2)
103.2(2)
75.1(3)
73.9(3)
145.2(3)
105.3(2)
89.6(2)
98.9(2)
97.8(2)
165.1(1)
103.2(1)
101.9(1)
74.4(1)
75.0(1)
144.8(1)
112.5(2)
85.7(3)
99.8(2)
93.5(2)
161.2(2)
106.4(2)
100.2(2)
73.3(3)
73.1(3)
142.7(3)
107.1(3)
88.6(3)
98.7(3)
96.4(3)
164.2(2)
105.4(2)
101.8(2)
73.3(3)
73.8(3)
143.2(3)
92.0(2)
120.3(2)
104.5(2)
101.1(2)
146.8(2)
93.2(2)
109.8(2)
72.8(2)
73.5(2)
87.3(2)
126.5(2)
106.1(2)
90.3(2)
145.7(2)
93.4(2)
116.5(2)
73.3(2)
73.8(2)
86.7(3)
127.0(3)
106.7(3)
91.7(3)
145.9(3)
93.9(3)
114.9(3)
73.6(3)
73.3(3)
90.9(1)
122.3(1)
105.4(1)
102.7(1)
146.8(1)
101.0(1)
100.6(1)
73.0(1)
X–M–N(1)
X–M–N(7)
X–M–N(9)
Y–M–N(1)
Y–M–N(7)
Y–M–N(9)
N(1)–M–N(7)
N(1)–M–N(9)
N(7)–M–N(9)
73.1(1)
144.0(1)
144.8(2)
146.8(2)
146.8(3)
a M = Fe, X = N(40), Y = Cl. b M = Co, X = N(40), Y = Cl. c M = Co, X = O, Y = Cl. d M = Fe, X = O(1), Y = Cl. e M = Fe, X = O(1), Y = O(2).
plane [O(1)–Fe–O(2)] being twisted about this vector by ca. 76Њ
to the “basal” N3M plane cf. 84 and 89Њ in 5a and 7 respectively.
Furthermore, the metal atom in 6i and 6ii lies slightly “below”
the plane of the pyridyl ring, whereas in each of the other seven
complexes it is positioned significantly “above” (Table 1). This
geometry is also highlighted by the increase in the X–M–N(1)
angle θ, and is particularly well illustrated in Fig. 8 when com-
pared with those in Fig. 6 and 7. The Fe–N bond distances are
unchanged from those observed for 2a, 4i and 4ii, as are the
C᎐N double bond lengths. In 5a and 7 the Fe–O distances are
᎐
all effectively the same (av. 1.955 Å) but those in 6i and 6ii are
noticeably longer (av. 2.019 Å), there being no differentiation
between pseudo-axial and basal Fe–O bonds. Both polymorphs
of 6 also exhibit larger differences in the torsional twists δ and γ
of the 2,6-diisopropylphenyl rings out of the N3M coordination
plane than those observed in the other complexes, and have
very much closer to equal isopropyl ؒ ؒ ؒ isopropyl separations
d and e (Fig. 5 and Table 1).
The packing of the molecules in 2a, 2b and 3 is similar with,
in each case, a pseudo-layer structure being formed based on a
co-alignment of the basal coordination planes, but with no
intermolecular interactions of note. In 4i there are inter-
molecular O–H ؒ ؒ ؒ F hydrogen bonds [2.74 and 2.87 Å]
between the apical aqua ligand and the SbF6 anion, a pseudo-
dimer being formed involving a centrosymmetrically related
pair of complex cations and SbF6 anions. In 4ii there is only a
single cation ؒ ؒ ؒ anion O–H ؒ ؒ ؒ F hydrogen bond [2.81 Å].
There are no intermolecular packing interactions of note in
complex 5a, the SbF6 anion being disordered. Both polymorphs
Fig. 7 View down the N(7) ؒ ؒ ؒ N(9) vector of the structure of 7
showing the distorted coordination geometry and the positioning of
O(2) between the faces of the two 2,6-diisopropylphenyl rings, compare
with Fig. 6.
By contrast, the two polymorphs of compound 6, with CF3
substituents on the β-diketonate ligand, both adopt a geometry
with near C2 symmetry about the Fe–N(1) axis, the X–M–Y
1162
J. Chem. Soc., Dalton Trans., 2002, 1159–1171

View Article Online
and 6ii also favours the formation of two weak F ؒ ؒ ؒ π contacts
to the pyridyl ring; in 6i these separations are both 3.14 Å,
whilst in 6ii they are 3.12 and 3.38 Å. In 7, however, there are
no such approaches, and the shortest Fe ؒ ؒ ؒ F separation is
3.78 Å.
Magnetic susceptibility. All of the described complexes are
paramagnetic. Their magnetic moments were determined by the
Evans’ NMR method (see Experimental section).44–47 The Fe()
complexes afford magnetic moments between 5.0 and 5.4 µB,
consistent with four unpaired electrons and a quintet ground
state for the iron() centres. The magnetic moments for the
cobalt() complexes are around 4.6 µB, suggesting a quartet
spin state with three unpaired electrons for the Co() centres.
Evidently, replacing a chloride ligand by ligands that exert a
stronger ligand field, such as CH3CN, thf or acac, does not
affect the high-spin configuration of the metal centres.
1H NMR spectroscopy. Despite the paramagnetic nature of
the cationic complexes, H NMR spectra can be obtained. All
Fig. 8 View down the N(7) ؒ ؒ ؒ N(9) vector of the structure of 6
showing the pseudo-trigonal bipyramidal geometry at iron, the
positioning of the iron atom “below” the plane of the pyridyl ring and
the sandwiching of the β-diketonate chelate ring between the faces of
the two 2,6-diisopropylphenyl rings.
1
signals appear as broad singlets and assignments of the reson-
ances can be made on the basis of integration, proximity to the
paramagnetic metal centre and by comparison with similar
paramagnetic complexes.10 For some proton signals, dramatic
differences in chemical shifts are observed between the neutral
dihalide precursors and the cationic complexes. In Fig. 10 the 1H
of 6 exhibit an unusual juxtaposition of the cations and anions
with, in each case, one of the fluorine atoms of the ordered
SbF6 anion being positioned proximal to the iron centre in the
vacant “octahedral” coordination site (Fig. 9). It should be
Fig. 9 The approach of one of the SbF6 fluorine atoms to the vacant
“octahedral” site of the iron centre in 6.
Fig. 10 1H NMR spectra of complexes 1a and 2a.
NMR spectrum of complex 2a is shown, together with the par-
ent dichloride precursor 1a. From these spectra it can be seen
that for both complexes, the meta-pyridyl protons (F) of the
ligand are found downfield at 90–100 ppm, which is indeed
common for most iron bis(iminopyridine) complexes. In sharp
contrast, the para-pyridyl proton (G), which is found at 83 ppm
in 1a, appears considerably further upfield in the case of the
cationic complexes, e.g. at about Ϫ14 ppm for 2a and Ϫ33 ppm
for 5a. The backbone methyls (A) are also strongly affected,
but their shifts are not easily rationalised, e.g. Ϫ37 ppm (1a),
Ϫ8 ppm (2a) and Ϫ61 ppm (5a). The aromatic protons (D and
E) and the isopropyl methyl groups (C, CЈ), which are further
removed from the iron centre, are largely unaffected. In some
cases not all of the expected signals are observed, for example
the methyl groups of the CH3CN ligand in 1a and 1b and the
noted that previous attempts to synthesis six-coordinate
bis(imino)pyridine iron() complexes of the type LFeCl2(D)
(where D = CO, P(OMe)3, PMe3) systematically failed and that
this is the first indication of a tendency towards octahedral
coordination in these five-coordinate complexes. In 6i and 6ii
the F ؒ ؒ ؒ Fe separations are 2.70 and 2.77 Å respectively, with
O(1)–Fe ؒ ؒ ؒ F angles of 159 and 160Њ. This distance is signifi-
cantly longer than the Fe–F bond between the iron atom and a
fluorine atom of the SbF6 anion in, for example, (hexafluoro-
antimonate)(meso-tetraphenylporphinato)iron() [ca. 2.11
Å].42 However, the Cambridge Crystallographic Database
(version 5.20) reveals only one example of a non-bonded
Fe ؒ ؒ ؒ F contact of less than 3.3 Å occurring between a five-
coordinate iron centre and a fluorine atom (3.24 Å).43 The
aforementioned close positioning of the anion and cation in 6i
J. Chem. Soc., Dalton Trans., 2002, 1159–1171
1163

View Article Online
thf signals in 3 are not observed. This may be due to exchange
processes or fluxional behaviour, resulting in an increased
broadening of these signals.
280–300 nm (ε = 7000–14000 dm3 molϪ1 cmϪ1) which are due to
ligand based transitions (π–π*). In the visible region, the iron
dichloride complex 1a displays a broad absorption around
700 nm (ε = 2500 dm3 molϪ1 cmϪ1) which is probably due to
metal to ligand charge transfer (MLCT). This absorption is
observed at 500–600 nm (ε = 400–600 dm3 molϪ1 cmϪ1) for the
iron monochloro (2a) and acac (5a) complexes, and is con-
sidered responsible for the blue (1a) and red (2a and 5a) colours
respectively.51,52 The cobalt dichloro complex (1b) also shows an
absorption at 700 nm, though much less intense compared to
the iron analogue, and which is shifted only slightly to 650 nm
for the monochloro complex (2b) and to 670 nm for the acac
complex (5b). Because of its low intensity (ε = 100–200
dm3molϪ1cmϪ1), we believe that this absorption more likely
results from a d–d transition. In addition, absorptions around
430 nm in the case of 1b and 2b and 480 nm (5b) are observed
(ε = 700–900 dm3 molϪ1 cmϪ1), most probably due to MLCT.
Addition of MAO activator to toluene solutions of the iron
pre-catalysts 1a, 2a or 5a results in all cases in a clear orange–
yellow solution. These solutions display a similar, rather
featureless UV/VIS spectrum (Fig. 11c), which shows that,
irrespective of the pre-catalyst employed (1a, 2a or 5a) the same
active species appears to be generated upon activation with
MAO. This suggests that the role of MAO is possibly twofold:
1) abstraction of one chloride from a dichloride precursor and
2) an exchange reaction of the remaining chloride (or acac)
ligand by an aluminium alkyl species, to generate the active
metal alkyl species (see Scheme 2). As mentioned previously,
alkylation of cobalt() precursors leads to a reduction of the
metal centre to Co().38 There is strong spectroscopic evidence
that reduction also occurs upon activation with MAO; these
studies will be reported separately in due course.
UV/VIS Spectroscopy. All the iron and cobalt complexes pre-
sented here are intensely coloured and, upon activation of
the pre-catalysts with the co-catalyst MAO, dramatic colour
changes are observed. For example, solutions of the bis-
(imino)pyridine iron dichloride complex 1a are blue, whereas
the monochloride 2a and acac iron complexes (5a, 6, 7) are red,
all of which turn orange after addition of MAO. The cobalt
analogues are green and change to purple upon activation.
UV/VIS spectroscopic studies of MAO-activated olefin
polymerisation systems have been carried out previously for
metallocenes and α-diimine nickel catalysts.48–50
The UV/VIS spectra of the dichloride (1a, 1b), monochloride
(2a, 2b) and acac (5a, 5b) iron and cobalt complexes as well
as the activated solutions have been recorded and are presented
in Fig. 11. All iron and cobalt complexes, as well as the free
ligand, show two characteristic absorptions in the UV region
between 224–234 nm (ε = 27000–45000 dm3 molϪ1 cmϪ1) and
Infrared spectroscopy. All of the complexes afford character-
istic absorptions for the bis(imino)pyridine ligands. The C᎐N
᎐
stretching mode ν(C᎐N) is observed at 1610–1630 cmϪ1 for the
᎐
coordinated bis(imino)pyridine ligand, which compares with
ν(C᎐N) in the range of 1640 to 1645 cmϪ1 for the free ligand.53
᎐
This shift to lower wavenumber upon complexation to a
metal is consistent with observations for other 2,6-bis(imino)-
pyridine metal complexes.54–56 Two strong absorption bands
between 1520 and 1580 cmϪ1 can be attributed to the combin-
ational stretching modes (ν(C᎐C) ϩ ν(C᎐O)) of the C᎐C and
᎐
᎐
᎐
the C᎐O stretches of the coordinated acetylacetonate
᎐
ligand.54,57–59 Complex 6 shows two strong absorptions at 1206
and 1147 cmϪ1 which can be attributed to the C–F stretching
mode, indicating the presence of the trifluoromethyl acetyl-
acetonate moiety.53,60 The characteristic absorption for the
Sb–F stretching mode ν(Sb–F) is observed at ca. 660 cmϪ1 61
.
Polymerisation results
Polymerisation tests with MAO as co-catalyst. Treatment of
the cationic Fe complexes 2a and 5a with 100 equiv. MAO
results in the formation of an orange–yellow solution (purple
for the Co complexes 2b and 5b), the same colour that is
observed after addition of MAO to a toluene solution of the
dichloride precursors 1a (1b for Co). The results in Table 3 show
that both in the case of iron as well as for cobalt, similar activ-
ities for the polymerisation of ethylene are obtained for the
dichloride (1a,b), monochloride (2a,b) and acac (5a,b) pre-
catalysts. The polymer properties are also comparable. These
polymerisation results, combined with the UV/VIS studies
presented above, suggest that the same active species is being
generated, irrespective of the pre-catalyst used (see Scheme 2)
and that this species does not have any halide or acac ligands
attached.
Fig. 11 UV/VIS spectra of (a) ligand and iron complexes 1a, 2a and
5a (CH2Cl2, c = 0.5 × 10Ϫ4 M for 200–400 nm, c = 5.0 × 10Ϫ4 M for 400–
900 nm), (b) ligand and cobalt complexes 1b, 2b and 5b (CH2Cl2, c = 0.5
× 10Ϫ4 M for 200–400 nm, c = 5.0 × 10Ϫ4 M for 400–900 nm) and (c) iron
complexes 1a, 2a and 5a ϩ MAO (100 equiv.) (toluene, c = 10Ϫ3 M).
A significant observation is that the presence of one equiv-
alent of a relatively strong donor such as acetonitrile does not
inhibit the polymerisation reaction. We believe that both the
1164
J. Chem. Soc., Dalton Trans., 2002, 1159–1171

View Article Online
Table 3 Results of ethylene polymerisation with precatalysts 1a,b, 2a,b and 5a,b in a 1 L autoclavea
Precatalyst/
µmol
Saturated
Unsaturated
chain endsd
Activityb
Mw
Mn
Mw/Mn
Mpk
chain endsd
c
c
c
c
Run
Yield/g
1
1a (0.5)
1b (0.6)
2a (0.5)
2b (0.6)
5a (0.5)
5b (0.6)
9.6
3.7
12.2
1.3
14.9
1.2
4800
450
6100
220
7500
200
500000
14000
278000
12000
317000
12000
53000
4200
64000
4500
78000
4600
9.5
3.3
4.3
2.6
4.1
2.7
285000
12000
254000
8900
179000
9100
Polymer insoluble
2e,f
3
2.4
0.4
3.8
0.3
2.8
1.9
0.1
2.8
0.1
2.6
4f
5
6f
a General conditions: isobutane solvent, triisobutylaluminium scavenger, 100 equiv. MAO, 4 bar ethylene pressure, 50 ЊC, reaction time 1 hour. b g
1
mmolϪ1 hϪ1 barϪ1
pressure.
.
c Determined by GPC at 135 ЊC. d Results from H NMR analysis, given per 1000 carbon atoms. e See ref. 10. f 10 bar ethylene
Table 4 Results of ethylene polymerisation with precatalysts in 500 ml Fisher–Porter bottlesa
Precatalyst/
µmol
Saturated
Unsaturated
chain endsd
Activityb
Mw
Mn
Mw/Mn
Mpk
chain endsd
c
c
c
c
Run
Yield/g
7
8
9
10
11
12
13
14
15
1a (2.0)
1b (10.0)
2a (2.0)
2b (10.0)
5a (2.0)
5b (10.0)
6 (2.0)
28.6
18.0
12.1
10.1
13.4
20.9
20.2
18.4
14.2
11400
1400
4840
810
5360
1670
8080
7360
5680
169000
15000
112000
19000
136000
16000
146000
121000
198000
46000
6300
33000
7200
52000
5900
40000
40000
49000
3.7
2.3
3.4
2.6
2.6
2.7
3.6
3.0
4.1
87000
13000
82000
13000
102000
13000
84000
79000
111000
0.3
2.8
0.5
2.0
0.5
2.2
0.5
0.5
0.5
0.3
2.2
0.4
2.1
0.4
2.2
0.4
0.4
0.4
7 (2.0)
8 (2.0)
a General conditions: toluene solvent (250 ml), 100 equiv. MAO, 5 bar ethylene pressure, r.t., reaction time 15 min. b g mmolϪ1 hϪ1 barϪ1
.
c Determined
by GPC at 135 ЊC. d Results from 1H NMR analysis, given per 1000 carbon atoms.
Scheme 2
acetonitrile ligand and the acac ligand are sequestered by the
co-catalyst upon generation of the active species. This is further
supported by the polymerisation results from Fisher–Porter
glass reaction vessels at 5 bar ethylene pressure. These experi-
ments were carried out to evaluate the influence of different
acac substituents and a different counter ion (BPh4Ϫ in complex
8). The results in Table 4 again show very similar activities and
polymer properties for all cationic iron acac complexes 5a–b
and 6–8. This observation suggests little interaction of the acac
moiety with the active species after activation with MAO. Fur-
thermore, no deleterious effect of the counterion is observed.
The higher activity of the neutral dichloride iron pre-catalysts
1a and 1b compared to their cationic counterparts under these
conditions is probably due to a higher initial activity within the
first few minutes after activation. This leads to a higher aver-
aged activity when shorter run times of 15 min are employed
(compare runs 1–4 in Table 3 vs. runs 7–10 in Table 4).
The cationic cobalt complexes 2b and 5b also produce poly-
ethylene with very similar activities and nearly identical molecu-
lar weight parameters compared to the dichloride pre-catalyst
1b when tested both at 10 bar and at 5 bar ethylene pressure.
This confirms that for both iron and cobalt bis(imino)pyridine
complexes, the nature of the pre-catalyst—neutral or cationic—
has no influence on the activity and molecular weight param-
eters in combination with MAO activator. From this observ-
ation we may conclude that the same active species is formed
upon activation with MAO.
Polymerisation tests with co-catalysts other than MAO.
Although MAO is the most frequently used activator for olefin
polymerisation catalysts, the reaction steps that lead to the
active species are still not adequately understood. Despite the
excellent properties of MAO as a co-catalyst, this lack of
understanding, combined with high cost and the fact that often
J. Chem. Soc., Dalton Trans., 2002, 1159–1171
1165

View Article Online
Table 5 Results of ethylene polymerisation with precatalysts 1–8 in Schlenk-line testsa
c
Mpk
Precatalyst
µmol
Saturated
Unsaturated
chain endsd
c
c
c
Run
Yield/g
Activityb
Mw
Mn
Mw/Mn
Peak 1
Peak 2
chain endsd
16e
17f
18e
19f
20e
21f
22e
23f
24g
25e
26f
27g
28e
29f
30e
31f
32e
33f
1a (9.4)
1a (9.2)
1b (9.8)
1b (14.2)
2a (9.8)
2a (9.5)
5a (9.6)
5a (9.4)
5a (10.2)
5b (9.8)
5b (10.7)
5b (9.1)
6 (10.2)
6 (9.4)
5.4h
1.6
4.8
1.2
4.9
1.2
4.9
2.6
0.2
4.3
1.1
1.0
5.1h
3.1
5.6h
2.8
5.6h
2.7
2300
700
1960
340
2000
510
2040
1110
80
1760
410
440
2000
1310
2290
1140
2240
1130
72000
101000
17000
26000
55000
477000
52000
117000
267000
19000
28000
23000
65000
136000
107000
114000
43000
82000
6000
9500
6700
21000
6500
155000
5500
40000
109000
7300
12000
10000
5300
44000
8100
39000
4300
12.1
10.7
2.5
2.5
8.5
3.1
9.5
2.9
2.5
2.6
2.3
2.3
12.2
3.1
13.3
3.0
10.1
4.3
1100
2200
43000
82000
12000
23000
27000
511000
20000
88000
249000
13000
21000
19000
26000
97000
58000
84000
14000
59000
3.7
2.5
2.4
2.4
4.0
0.5
4.8
0.5
1.5
2.2
1.5
2.0
4.0
0.9
2.7
0.6
5.9
1.5
0.4
0.2
2.1
1.9
0.7
0
0.7
0.4
0
1.9
1.2
1.5
0.6
0.3
0.3
0.3
1.4
0.4
1200
1000
900
900
800
7 (9.8)
7 (9.8)
8 (10.0)
8 (9.6)
19000
a General conditions: toluene solvent (100 ml), 1 bar ethylene pressure, r.t. b g mmolϪ1 hϪ1 barϪ1 c Determined by GPC at 135 ЊC. d Results from 1H
.
NMR analysis, given per 1000 carbon atoms. e 100 equiv. MAO, reaction time 15 min. f 1 equiv. B(C6F5)3/10 equiv. TMA, reaction time 15 min.
g 1 equiv. B(C6F5)3/10 equiv. TiBAl, reaction time 15 min. h Mass transport problems may have occurred.
a large excess is needed, render it less attractive. In addition, a
prime objective of this study was to reduce the co-catalyst con-
centration in order to make the catalyst more amenable to polar
co-monomers.
B(C6F5)3 prior to the addition of TMA results in a more stable
active species. This is illustrated by the fact that both MAO and
B(C6F5)3/TMA activated solutions remain active after ageing
for 10–15 minutes. Complexes activated with TMA alone are
only active immediately after activation. Addition of MeAlCl2
to the active catalyst solution, formed using 1 equiv. B(C6F5)3/
10 equiv. TMA, resulted in no polymerisation activity. This
suggests that chloro aluminium alkyl reagents poison iron and
cobalt catalysts, which is in line with the very low activities
observed when chloro aluminium alkyls are used as co-catalysts
for the neutral iron and cobalt complexes 1a and 1b.6,26
In Table 5 the ethylene polymerisation results for neutral
and cationic iron and cobalt complexes activated with different
co-catalysts are summarised. Generally, under the given condi-
tions, the neutral and cationic iron complexes bearing the 2,6-
diisopropylphenyl-bis(imino)pyridine ligand show an activity
between 2000 and 2300 g mmolϪ1 hϪ1 barϪ1 when activated with
100 equiv. MAO. With 1 equiv. B(C6F5)3 and 10 equiv. trimethyl-
aluminium (TMA) as co-catalyst, a decrease in activity to 900–
1300 g mmolϪ1 hϪ1 barϪ1 is observed. This may be due to a less
efficient scavenging of catalyst poisons in the case of a lower
aluminium alkyl concentration.
Very striking is the effect of the different co-catalysts on the
polymer properties. Under Schlenk-line conditions with 1 bar
ethylene pressure and 100 equiv. MAO as co-catalyst a bimodal
molecular weight distribution of the polyethylene is observed.
This is the case for neutral and cationic iron complexes. As
discussed previously, at low ethylene pressures and relatively
high aluminium alkyl concentrations, the rate of chain trans-
fer to aluminium and the rate of β-H transfer can become
competitive, leading to the formation of a fully saturated low
molecular weight fraction in addition to the high molecular
weight vinyl terminated polymer.10 Upon consumption of the
aluminium alkyl, only high molecular weight polyethylene is
produced with β-hydride transfer as the main transfer mechan-
ism. As a consequence of the lower aluminium alkyl concen-
tration, with 1 equiv. B(C6F5)3/10 equiv. TMA as co-catalyst
only high molecular weight polyethylene is produced with a
unimodal molecular weight distribution and relatively narrow
PDI. Analysis of the chain ends shows an approximate 1 : 1
ratio between saturated and vinyl end-groups indicating that
β-hydride transfer is the sole chain termination mechanism.
As seen before with MAO as co-catalyst, no effect of different
substituents on the acac ligand or of the different counter ions
is observed. Entry 24 in Table 5 shows that triisobutyl-
aluminium (TiBAl) in combination with B(C6F5)3 is a less effi-
cient co-catalyst for the iron system.
Co-polymerisation of ethylene with polar monomers. The
observation that, for the iron and cobalt bis(imino)pyridine
catalysts, the presence of one equivalent of acetonitrile or
tetrahydrofuran does not inhibit the polymerisation reaction
indicates that these catalysts possess good functional group
tolerance. This, along with the finding that these cationic
pre-catalysts can be activated with very small amounts of
alkyl aluminium co-catalyst, are important requirements for
hydrocarbon olefin/polar monomer co-polymerisations. Co-
polymerisation tests using ethylene and a variety of polar
monomers were carried out under Schlenk-line conditions with
1000 equiv. of the polar momoner and the results are summar-
ised in Table 6. Surprisingly, both the neutral pre-catalyst 1a
(activated with 100 equiv. MAO) and the cationic pre-catalyst
5a (activated with 1 equiv. B(C6F5)3/10 TMA) show high poly-
merisation activities in the presence of 1000 equiv. methyl
methacrylate (MMA) or styrene. In the presence of methyl
acrylate (MA) or 2-vinyl-1,3-dioxolane (VDO) the activity is
significantly reduced. The polymer properties are also affected.
The polydispersity of the polymer generated by the neutral
dichloride pre-catalyst 1a, in combination with MAO, is
lowered in the presence of polar monomer (see runs 34–36).
This could possibly be due to an interaction between the
polar group of the co-monomer and trimethyl aluminium
(present in MAO), thereby reducing the rate of chain transfer to
aluminium.
1
Analysis by IR and H NMR spectroscopy of the polymer
products obtained in runs 35, 36 and 38–41 revealed that in all
cases, in addition to the signals for polyethylene, minor signals
corresponding to the functional group of the co-monomer
were present. However, it should be noted that, at low levels of
co-monomer incorporation, the detection by IR and NMR
All cationic iron and cobalt acac complexes can be activated
with as little as 10 equiv. TMA. Different colour changes upon
activation and after stirring for a few minutes indicate a lower
stability of the active species in those cases. Addition of 1 equiv.
1166
J. Chem. Soc., Dalton Trans., 2002, 1159–1171

View Article Online
Table 6 Co-polymerisation results of pre-catalysts 1a and 5a with various polar olefins
Precatalyst/
µmol
Saturated
endsd
Unsaturated
endsd
c
c
c
c
Run
Polar olefin
Yield/g
Activityb
Mw
Mn
Mw/Mn
Mpk
34e
35e
36e
37f
38f
39f
40f
41f
1a (9.4)
1a (9.5)
1a (10.0)
5a (9.4)
5a (10.0)
5a (10.2)
5a (10.0)
5a (9.4)
5.4g
2.7
2.2
2.6
4.1g
0.7
3.1
0.4
2300
1220
440
1100
1600
140
72000
200000
137000
117000
92000
132000
214000
218000
6000
59000
34000
40000
1300
53000
70000
70000
12.1
3.4
4.1
2.9
6.8
2.5
3.1
3.7
43000
136000
84000
88000
43000
101000
111000
158000
3.7
0.6
0.7
0.5
0.7
0.2
0.8
0.6
0.4
0.3
0
0.4
0.4
0
MMA
MA
MMA
MA
Styrene
VDO
1220
180
0.2
0.1
a General conditions: toluene solvent (100 ml), 1000 equiv. polar olefin, 1 bar ethylene pressure, r.t., reaction time 15 min (except runs 36 and 39; 30
1
min). b g mmolϪ1 hϪ1 barϪ1
.
c Determined by GPC at 135 ЊC. d Results from H NMR analysis, given per 1000 carbon atoms. e 100 equiv. MAO
f 1 equiv. B(C6F5)3/10 equiv. TMA. g Mass transport problems may have occurred.
spectroscopy of true co-polymers, as compared to a blend of
two homo-polymers, is not straightforward. We therefore
employed a procedure62 of dissolving the crude polymer
products obtained in runs 35–36 and 38–41 (Table 6) in hot
1,2,4-trichlorobenzene, followed by precipitation in thf (see
experimental). Blank experiments were carried out using homo-
polymer mixtures of polyethylene and PMMA, PMA or poly-
styrene, respectively. IR spectroscopic analysis clearly showed
that these mixtures are easily separated via this procedure. In
the case of PMA the procedure had to be repeated several
times. IR spectroscopic analysis of the purified polymer no
longer showed the characteristic absorptions for the polar co-
monomers, clearly establishing that in all co-polymerisation
experiments a mixture of homo-polymers had been produced,
rather than a co-polymer. By comparison of the IR spectra with
those of the materials used in the blank experiments, it can
be estimated that the amount of polar homo-polymer was
generally very low. It is interesting to note that, under similar
reaction conditions, but in the absence of ethylene, homo-
polymerisation of these polar monomers does not occur.
In the case of other polar monomers such as vinyl acetate,
acrolein or acrylonitrile, no polymer is obtained when these
polar monomers are added to the activated catalyst solution
prior to the addition of ethylene (conditions as in run 38, Table
6). In another series of experiments where these polar mono-
mers were added 90 s after exposure of the activated catalyst to
ethylene, i.e. after ethylene polymerisation has commenced,
only polyethylene was produced; there was no evidence by
NMR and IR spectroscopy for polar monomer incorporation.
Co-polymerisation tests with methyl methacrylate (MMA),
methyl acrylate (MA), styrene and 2-vinyl-1,3-dioxolane
(VDO) were carried out. Both the neutral and the cationic iron
complexes are active polymerisation catalysts in the presence of
these polar olefins, whereas monomers such as vinyl acetate,
acrolein or acrylonitrile deactivate the catalyst. In all co-
polymerisation tests no incorporation of the polar monomer
into the ethylene chain can be substantiated.
Experimental
General considerations
All manipulations of water and/or moisture sensitive com-
pounds were performed by means of standard high vacuum
Schlenk and cannula techniques. Complexes were transferred in
a nitrogen filled glove-box and, unless stated otherwise, stored
at room temperature. Crystal data were collected on a Siemens
1
P4/PC diffractometer with Cu-Kα radiation. H NMR spectra
were recorded with a Bruker AC-250 spectrometer. Chemical
shifts were referenced to the residual proton signal of the
deuterated solvents. Infrared spectra were obtained on a Perkin
Elmer 1760X FT-IR spectrometer using thin film samples on
NaCl plates. Relative intensities of the IR-bands have been
described as w (weak), m (medium), s (strong). UV/VIS spectra
were recorded in 10 mm quarz glass cells on a Perkin Elmer
Lambda 20 spectrometer. Mass spectra were recorded on either
a VG Autospec or a VG Platform II spectrometer. Elemental
analysis were performed by the microanalytical services of the
Chemistry Departments of Imperial College and the University
of North London. GPC and NMR spectroscopic analyses of
polyethylene were performed by BP Chemicals at Sunbury.
Magnetic moments were determined by the Evans NMR
method.44–47
Conclusion
The synthesis and characterisation of a series of cationic iron
and cobalt complexes have been described, all containing the
same 2,6-diisopropylphenyl substituted bis(imino)pyridine
ligand. Activation with MAO produces highly active ethylene
polymerisation catalysts. As seen for the neutral dichloride
complexes, the cationic iron complexes show an order of
magnitude higher activity than the corresponding cobalt com-
plexes. The formation of cationic pre-catalysts with different
acac substituents and different counter ions has no adverse
affect on the catalyst activity. In fact, little difference in activity
and polymer properties is observed between neutral dichloro, or
cationic monochloro or acac pre-catalysts, which leads us to
conclude that the active species generated upon activation with
MAO, is likely to be the same in all cases and does not con-
tain any of the co-ligands (chloride, acac or acetonitrile).
Co-catalysts other than MAO have been tested and co-catalyst
concentrations as low as 10 equiv. TMA per metal centre
afford efficient ethylene polymerisation catalysts. The resultant
materials possess narrower molecular weight distributions
than those obtained from neutral dichloride precatalysts
where much larger activator concentrations are employed.
Solvents and reagents
Pentane and toluene were dried by passing through a column,
filled with commercially available Q-5 reactant (13 wt.% CuO
on alumina) and activated alumina (pellets, 3 mm). The follow-
ing solvents were dried by prolonged reflux over a suitable
drying agent under an atmosphere of nitrogen, being freshly
distilled prior to use. For diethyl ether and tetrahydrofuran
(thf ), sodium benzophenone ketyl was used as drying agent,
whereas acetonitrile (MeCN) and dichloromethane were
dried over calcium hydride. All solvents were used without
degassing unless otherwise stated. CD2Cl2 was dried over 4 Å
molecular sieves. Ethylene was purchased from BOC Gases and
used without further purification. Methylacrylate (MA) and
methylmethacrylate (MMA) were dried over calcium hydride
and vacuum-transferred and degassed prior to use. Styrene (St)
and 2-vinyl-1,3-dioxolane (VDO) were dried over 4 Å molecu-
lar sieves and distilled and degassed prior to use. The 2,6-bis-
[1-(2,6-diisopropylphenylimino)ethyl]pyridine ligand (UV/VIS
J. Chem. Soc., Dalton Trans., 2002, 1159–1171
1167

View Article Online
(CH2Cl2): λmax/nm (εmax/dm3 molϪ1 cmϪ1) = 343 (1400), 279
(11500), 234 (42000)) and the complexes 1a (UV/VIS (CH2Cl2):
λmax/nm (εmax/dm3 molϪ1 cmϪ1) = 695 (2500), 289 (7000), 224
(44500)) and 1b (UV/VIS (CH2Cl2): λmax/nm (εmax/dm3 molϪ1
cmϪ1) = 703 (400), 291 (7000), 228 (32000)) were prepared
according to published procedures.10 All other reagents are
commercially available and were used without further
purification.
solution was filtered off and the solvent was removed in vacuo.
The red residue was redissolved in 10 ml dichloromethane and
precipitated with 60 ml pentane. The supernatant solution was
filtered off and the residue was dried in vacuo. Yield: 390 mg
(91%). Crystals suitable for X-ray analysis were grown from a
concentrated dichloromethane solution layered with pentane.
1H NMR (250 MHz, CD2Cl2, r.t., all peaks appear as broad
singlets): δ Ϫ60.5 (6H, N᎐C–Me), Ϫ33.0 (1H, PyH ), Ϫ22.1
᎐
p
(6H, O᎐C–Me), Ϫ12.67 (2H, PhH ), Ϫ6.0 (12H, iPrMe), Ϫ2.8
᎐
p
(12H, iPrMe), 2.4 (4H, iPrCH or PhHm), 13.1 (4H, iPrCH or
Syntheses
PhH ), 30.5 (1H, O᎐C–CH ), 117.3 (2H, PyH ). IR (neat com-
᎐
m
m
2,6-Bis[1-(2,6-diisopropylphenylimino)ethyl]pyridine iron(II)
chloride hexafluoroantimonate acetonitrile adduct (2a). Silver
hexafluoroantimonate (1.22 g, 2.0 mmol) and 1a (687 mg, 2.00
mmol) were dissolved in 50 ml MeCN. After stirring at r.t. for
16 h the solution was filtered off and the solvent removed in
vacuo. The residue was washed with diethyl ether (2 × 30 ml).
After drying in vacuo a red powder was obtained. Yield: 1.59 g
(94%). Crystals suitable for X-ray analysis were grown from a
concentrated dichloromethane solution layered with pentane.
1H NMR (250 MHz, CD2Cl2, r.t., all peaks appear as broad
singlets): δ Ϫ13.8 (1H, PyHp), Ϫ13.1 (2H, PhHp), Ϫ8.7 (6H,
pound): 1625 (w, ν(C᎐N)), 1567 (s, ν(C᎐C) ϩ ν(C᎐O)), 1522 (s,
᎐
᎐
᎐
ν(C᎐C) ϩ ν(C᎐O)), 659 (s, ν(Sb–F)) cmϪ1. UV/VIS (CH2Cl2):
᎐
᎐
λmax/nm (εmax/dm3molϪ1cmϪ1) = 560 (500), 296 (11000), 228
(27500). FAB mass spectrum (m/z): 636 [(M Ϫ SbF6)ϩ], 620
[(M Ϫ SbF6)ϩ Ϫ CH3]. Anal. Calcd. for C38H50N3F6FeO2Sbؒ
1 CH2Cl2: C, 48.93; H, 5.47; N, 4.39. Found: C, 49.11; H, 5.33;
N, 4.33%. µeff (Evans’ NMR method) = 5.0 µB.
2,6-Bis[1-(2,6-diisopropylphenylimino)ethyl]pyridine cobalt(II)
acetylacetonate hexafluoroantimonate (5b). Silver hexafluoro-
antimonate (640 mg, 1.9 mmol) and 1b (1.14 g, 1.9 mmol) were
dissolved in 50 ml MeCN and stirred at r.t. for one hour. Then a
suspension of silver acetylacetonate (384 mg, 1.9 mmol) in
20 ml MeCN was added. After stirring at r.t. for 16 h the solu-
tion was filtered off and the solvent was removed in vacuo. The
residue was washed with diethyl ether and after drying in vacuo
the product was obtained as a pale green solid. Yield: 1.30 g
(78%). By using the procedure described above and 2b as the
starting complex, a green powder was obtained in 53% yield. 1H
NMR (250 MHz, CD2Cl2, r.t., all peaks appear as broad sing-
N᎐C–Me), Ϫ2.2 (24H, iPrMe), 10.6 (4H, iPrCH or PhH ), 12.4
᎐
m
(4H, PhHm or iPrCH ), 100.6 (2H, PyHm). IR (neat compound):
1617 (m, ν(C᎐N)), 670 (s, ν(Sb–F)) cmϪ1. UV/VIS (CH2Cl2):
᎐
λmax/nm (εmax/dm3 molϪ1 cmϪ1) = 540 (400), 298 (6000), 224
(39000). FAB mass spectrum (m/z): 572 [(M Ϫ SbF6)ϩ
Ϫ
MeCN]. Anal. Calcd. for C35H46N4ClF6FeSbؒ0.5CH2Cl2: C,
47.79; H, 5.31; N, 6.28. Found: C, 48.06; H, 5.09; N, 6.39%. µeff
(Evans’ NMR method) = 5.4 µB.
2,6-Bis[1-(2,6-diisopropylphenylimino)ethyl]pyridine cobalt(II)
chloride hexafluoroantimonate acetonitrile adduct (2b). By using
the procedure described above and 1b as the starting complex, a
green powder was obtained in 56% yield. Crystals suitable for
X-ray analysis were grown from a concentrated dichloro-
methane solution layered with pentane. H NMR (250 MHz,
CD2Cl2, r.t., all peaks appear as broad singlets): δ Ϫ65.9 (2H,
PhHp), Ϫ24.9 (12H, iPrMe), Ϫ11.6 (12H, iPrMe), Ϫ6.3 (4H,
lets): δ Ϫ73.7 (4H, iPrCH or PhH ), Ϫ35.6 (1H, O᎐C–CH ),
Ϫ12.3 (12H, iPrMe), Ϫ11.6 (12H, iPrMe), Ϫ9.9 (2H, PhHp),
᎐
m
Ϫ9.5 (6H, O᎐C–Me), 10.3 (6H, N᎐C–Me), 12.04 (4H, PhH or
᎐
᎐
m
iPrCH ), 35.2 (1H, PyHp), 116.2 (2H, PyHm). IR (neat com-
pound): 1624 (w, ν(C᎐N)), 1578 (s, ν(C᎐C) ϩ ν(C᎐O)), 1522 (s,
᎐
᎐
᎐
1
ν(C᎐C) ϩ ν(C᎐O)), 659 (s, ν(Sb–F)) cmϪ1. UV/VIS (CH2Cl2):
᎐
᎐
λmax/nm (εmax/dm3molϪ1cmϪ1) = 670 (100), 483 (900), 293
(14000), 228 (32000). FAB mass spectrum (m/z): 639 [(M Ϫ
SbF6)ϩ]. Anal. Calcd. for C38H50N3CoF6O2Sbؒ2CH2Cl2: C,
45.96; H, 5.21; N, 4.02. Found: C, 45.41; H, 4.93; N, 4.21%. µeff
(Evans’ NMR method) = 4.6 µB.
iPrCH or PhH ), 15.7 (6H, N᎐C–Me), 16.9 (4H, PhH or
᎐
m
m
iPrCH ), 50.6 (1H, PyHp), 125.4 (2H, PynHm). IR (neat com-
pound): 1618 (w, ν(C᎐N)), 662 (s, ν(Sb–F)) cmϪ1. UV/VIS
᎐
(CH2Cl2): λmax/nm (εmax/dm3 molϪ1 cmϪ1) = 647 (200), 293
(8000), 227 (28000). FAB mass spectrum (m/z): 575 [(M Ϫ
SbF6)ϩ Ϫ MeCN]. Anal. Calcd. for C35H46N4ClCoF6Sbؒ1.5
CH2Cl2: C, 44.72; H, 5.04; N, 5.71. Found: C, 44.47; H, 5.18; N,
5.36%. µeff (Evans’ NMR method) = 4.6 µB.
2,6-Bis[1-(2,6-diisopropylphenylimino)ethyl]pyridine iron(II)
1,3-bis(trifluoromethyl)-1,3-propanedionate
hexafluoroanti-
monate (6). Sodium 1,3-trifluoromethyl-1,3-propanedionate
(46 mg, 0.20 mmol) and 2a (170 mg, 0.20 mmol) were dissolved
in a mixture of 30 ml diethyl ether and 10 ml dichloromethane.
After stirring at r.t. for 16 h the solvent was removed in vacuo
and the residue was extracted with dichloromethane. After
removal of the solvent in vacuo a purple powder was obtained.
2,6-Bis[1-(2,6-diisopropylphenylimino)ethyl]pyridine cobalt(II)
chloride hexafluoroantimonate tetrahydrofuran adduct (3). Silver
hexafluoroantimonate (225 mg, 0.655 mmol) and 1b (400 mg,
0.654 mmol) were dissolved in 75 ml THF. After stirring for 4 h
at r.t. 15 ml dichloromethane were added to solubilise the
product. The solution was filtered off and the solvent was
removed in vacuo to obtain a green powder. Yield: 460 mg
(80%). Crystals suitable for X-ray analysis were grown from a
concentrated dichloromethane solution layered with pentane.
1H NMR (250 MHz, CD2Cl2, r.t., all peaks appear as broad
singlets): δ Ϫ24.9 (12H, iPrMe), Ϫ11.9 (12H, iPrMe), Ϫ6.5
1
Yield: 100 mg (51%). H NMR (250 MHz, CD2Cl2, r.t., all
peaks appear as broad singlets): δ Ϫ27.5 (1H, PyHp), Ϫ19.21
(1H, O᎐C–CH ), Ϫ14.6 (2H, PhH ), Ϫ6.4 (12H, iPrMe),
᎐
p
Ϫ2.2 (12H, iPrMe), 6.7 (4H, iPrCH or PhHm), 14.4 (4H, iPrCH
or PhHm), 116.8 (2H, PyHim). IR (neat compound): 1625
(w, ν(C᎐N)), 1552 (w, ν(C᎐C) ϩ ν(C᎐O)), 1522 (w, ν(C᎐C) ϩ
᎐
᎐
᎐
᎐
ν(C᎐O), 1206 (s, ν(CF ), 1147 (s, ν(CF )), 660 (s, ν(Sb–F)) cmϪ1
.
᎐
3
3
FAB mass spectrum (m/z): 744 [(M Ϫ SbF6)ϩ]. Anal. Calcd. for
C38H44N3F12FeO2Sb: C, 46.56; H, 4.52; N, 4.29. Found: C,
46.72; H, 4.48; N, 4.28%. µeff (Evans’ NMR method) = 5.3 µB.
(4H, iPrCH or PhH ), 15.7 (4H, iPrCH or PhH ), 18.0 (6H, N᎐
᎐
m
m
C–Me), 53.2 (1H, PyHp), 123.7 (2H, PyHm). IR (neat com-
pound): 1619 (m, ν(C᎐N)), 660 (s, ν(Sb–F)) cmϪ1. FAB mass
᎐
spectrum (m/z): 575 [(M Ϫ SbF6)ϩ Ϫ THF]. Anal. Calcd. for
C37H51N3ClOCoF6Sb: C 50.27, H 5.82, N 4.75. Found C 50.26,
H 5.81, N 4.84%. µeff (Evans’ NMR method) = 4.9 µB.
2,6-Bis[1-(2,6-diisopropylphenylimino)ethyl]pyridine iron(II)
1,3-diphenyl-1,3-propanedionate hexafluoroantimonate (7).
Sodium 1,3-diphenyl-1,3-propanedionate63 (74 mg, 0.30 mmol)
and 2a (250 mg, 0.30 mmol) were dissolved in 40 ml MeCN.
After stirring at r.t. for 16 h the solvent was removed in vacuo
and the red residue was extracted with dichloromethane. After
removing the solvent in vacuo a purple powder was obtained.
2,6-Bis[1-(2,6-diisopropylphenylimino)ethyl]pyridine iron(II)
acetylacetonate hexafluoroantimonate (5a). Silver acetylaceto-
nate (420 mg, 0.49 mmol) and 2a (102 mg, 0.49 mmol) were
dissolved in 30 ml MeCN. After stirring at r.t. for 16 h, the
1
Yield: 170 mg (57%). H NMR (250 MHz, CD2Cl2, r.t., all
1168
J. Chem. Soc., Dalton Trans., 2002, 1159–1171

View Article Online
peaks appear as broad singlets): δ Ϫ33.77 (1H, PyHp), Ϫ13.7
(2H, PhHp), Ϫ5.5 (12H, iPrMe), Ϫ1.9 (12H, iPrMe), 6.6
was prepared by dissolving the pre-catalyst in toluene and
adding MAO (10 wt% in toluene) in a Schlenk-tube. An aliquot
of this solution was then injected into the reactor under nitro-
gen pressure. The reactor pressure was maintained constant
throughout the polymerisation run by computer-controlled
addition of ethylene. The polymerisation time was 60 minutes.
Venting off all volatiles terminated the runs, the reactor con-
tents were isolated, washed with aqueous HCl and methanol,
and dried in a vacuum oven at 50 ЊC.
(10H, O᎐C-Ph), 12.8 (6H, N᎐C–Me), 22.5 (4H, iPrCH or
᎐
᎐
PhHm), 25.5 (4H, iPrCH or PhHm), 114.0 (2H, PyHm). IR (neat,
cmϪ1): 1624 (w, ν(C᎐N)), 1522 (s, ν(C᎐C) ϩ ν(C᎐O)), 658 (s,
᎐
᎐
᎐
ν(Sb–F)). FAB mass spectrum m/z): 760 [(M Ϫ SbF6)ϩ]. Anal.
Calcd. for C48H54N3F6FeO2Sb: C, 57.85; H, 5.46; N, 4.22.
Found: C, 57.95; H, 5.33; N, 4.14%. µeff (Evans’ NMR method)
= 5.4 µB.
2,6-Bis[1-(2,6-diisopropylphenylimino)ethyl]pyridine iron(II)
acetylacetonate tetraphenylborate (8). Silver acetylacetonate
(108 mg, 0.52 mmol) and 1a (316 mg, 0.52 mmol) were dis-
solved in 30 ml MeCN. After stirring at r.t. for 1 h, the solution
was filtered into a solution of 188 mg (0.55 mmol) sodium
tetraphenylborate in 20 ml MeCN. Further extractions of the
dark residue with MeCN were also added. After stirring at
r.t. for 16 h, the solvent was removed in vacuo and the red
residue was extracted with dichloromethane. The solution was
concentrated in vacuo to 5 ml. The product was precipitated by
addition of pentane. After removing the supernatant solution
and drying in vacuo a red powder was obtained. Yield: 351 mg
High pressure tests in 500 ml Fisher–Porter glass reaction
vessels. A 500 ml glass bottle with a stainless steel headset con-
taining a mechanical stirrer and an injection valve was evacu-
ated and flushed with nitrogen 3 times before it was filled with
toluene (250 ml). The catalyst solution was prepared by dissolv-
ing the pre-catalyst in toluene and adding MAO (10 wt% in
toluene) in a Schlenk-tube. An aliquot of this solution was then
injected under nitrogen pressure. Ethylene was introduced into
the reactor, and the reactor pressure was maintained at 5 bar
throughout the polymerisation run by addition of ethylene. The
polymerisation time was 15 minutes. Releasing the ethylene
pressure and adding aqueous HCl and methanol terminated the
runs. The solid PE was recovered by filtration, washed with
methanol, and dried in a vacuum oven at 50 ЊC.
1
(71%). H NMR (250 MHz, CD2Cl2, r.t., all peaks appear as
broad singlets): δ Ϫ61.5 (6H, N᎐C–Me), Ϫ34.5 (1H, PyH ),
᎐
p
Ϫ22.7 (6H, O᎐C–Me), Ϫ12.9 (2H, PhH ), Ϫ6.0 (12H, iPrMe),
᎐
p
Ϫ2.9 (12H, iPrMe), 2.3 (4H, PhHm or iPrCH ), 7.0 (4H,
BPhH ), 7.4 (8H, BPhH ), 8.5 (8H, BPhH ), 13.1 (4H, iPrCH or
Schlenk-line 1 bar ethylene tests. The pre-catalyst was dis-
solved in toluene (120 ml) and a solution of B(C6F5)3 in toluene
and trimethylaluminium (2.0 M in toluene) was added. The
Schlenk tube was purged with ethylene, and the contents were
magnetically stirred and maintained under ethylene (1 bar)
throughout the polymerisation run. The polymerisation time
was 15 minutes. Releasing the ethylene pressure and adding
aqueous HCl and methanol terminated the runs. The solid PE
was recovered by filtration, washed with methanol, and dried in
a vacuum oven at 60 ЊC.
PhH ), 30.4 (1H, O᎐C–CH ), 117.8 (2H, PyH ). IR (neat com-
᎐
m
m
pound): 1621 (w, ν(C᎐N)) cmϪ1. FAB mass spectrum (m/z): 636
᎐
[(M Ϫ B(C6H5)4)ϩ]. Anal. Calcd. for C62H70N3BFe: C, 77.90; H,
7.38; N, 4.40. Found: C, 77.90; H, 7.40; N, 4.63. µeff (Evans’
NMR method) = 5.1 µB.
Magnetic susceptibility
According to the NMR method described by Evans,44–47 a solu-
tion of the compound (≈ 3 mg) in a mixture of dichloro-
methane-d2/cyclohexane (95/5 v/v) was prepared in a 1.0 ml
volumetric flask. A portion of this solution was transferred into
a melting point capillary tube that was sealed with PTFE-tape
and dropped into a NMR tube containing the dichloro-
methane-d2/cyclohexane (95/5 v/v) mixture. The chemical shift
difference of the cyclohexane signal between the inner and the
outer tubes was measured at room temperature. From this dif-
ference in the chemical shift the molar susceptibility χM and the
magnetic moment µeff can be calculated with equations (1) and
(2) (given in S.I. units).
General co-polymerisation procedures
Schlenk-line 1 bar ethylene tests. The pre-catalyst was dis-
solved in toluene (120 ml) and a solution of B(C6F5)3 in toluene
and trimethylaluminium (2.0 M in toluene) was added. The
polar monomer (typically 1000 equivalents) was added via
syringe. The Schlenk tube was purged with ethylene, and the
contents were magnetically stirred and maintained under ethyl-
ene (1 bar) throughout the polymerisation run. The polymeris-
ation time was 15 minutes. Releasing the ethylene pressure and
adding aqueous HCl and methanol terminated the runs. The
solid polymer was recovered by filtration, washed thoroughly
with methanol, and dried in a vacuum oven at 60 ЊC. The crude
polymer product was subsequently dissolved in hot 1,2,4-
trichlorobenzene and the hot solution poured into a tenfold
excess of tetrahydrofuran at room temperature. The precipi-
tated polymer was collected by filtration, washed with thf and
dried in a vacuum oven at 60 ЊC.
(1)
(2)
X-Ray crystallography
where: χM is the molar susceptibility of the sample in m3 mol,
∆: is the difference in the chemical shift of a set of protons in
Hz, : is the frequency of operation of the spectrometer in Hz,
c is the concentration of the sample in mol dmϪ3, T is the
temperature in K.
Table 7 provides a summary of the crystal data, data collection
and refinement parameters for compounds 2a, 2b, 3, 4i, 4ii, 5a,
6i, 6ii and 7.
CCDC reference numbers 168887–168895.
See http://www.rsc.org/suppdata/dt/b1/b106614p/ for crystal-
lographic data in CIF or other electronic format.
General polymerisation procedures
High pressure tests in a 1 l steel autoclave. A 1 l stainless steel
reactor was baked under nitrogen flow for at least 1 h at >85 ЊC
and subsequently cooled to the temperature of polymerisation.
Isobutane (0.5 l) and triisobutylaluminium (2.0 ml, 1.0 M in
hexane) were introduced into the reactor and stirred at the reac-
tion temperature for at least 1 h. Ethylene was introduced into
the reactor by backpressure of nitrogen. The catalyst solution
Acknowledgements
The authors are grateful to BP Chemicals Ltd. for financial
support. Dr J. Boyle and Dr. G. Audley are thanked for
GPC and NMR measurements. Dr P. Maddox and Prof. D.
Goodgame are thanked for helpful discussions.
J. Chem. Soc., Dalton Trans., 2002, 1159–1171
1169

View Article Online
1170
J. Chem. Soc., Dalton Trans., 2002, 1159–1171

View Article Online
29 L. Deng, P. Margl and T. Ziegler, J. Am. Chem. Soc., 1999, 121,
References
6479.
1 G. J. P. Britovsek, V. C. Gibson and D. F. Wass, Angew. Chem.,
Int. Ed., 1999, 38, 428.
2 S. D. Ittel, L. K. Johnson and M. Brookhart, Chem. Rev., 2000, 100,
1169.
3 S. Mecking, Angew. Chem., Int. Ed., 2001, 40, 534.
4 G. J. P. Britovsek, V. C. Gibson, B. S. Kimberley, P. J. Maddox,
S. J. McTavish, G. A. Solan, A. J. P. White and D. J. Williams,
Chem. Commun., 1998, 849.
5 B. L. Small, M. Brookhart and A. M. A. Bennett, J. Am. Chem. Soc.,
1998, 120, 4049.
6 G. J. P. Britovsek, B. A. Dorer, V. C. Gibson, B. S. Kimberley,
G. A. Solan, World Pat., WO 99/12981 (BP Chemicals Ltd.), 1999;
G. J. P. Britovsek, B. A. Dorer, V. C. Gibson, B. S. Kimberley and
G. A. Solan, Chem. Abstr., 1999, 130, 252793.
7 A. M. A. Bennett, World Pat., WO 98/27124 (DuPont), 1998;
A. M. A. Bennett, Chem. Abstr., 1998, 129, 122973.
8 V. C. Gibson and D. F. Wass, Chem. Br., 1999, 35, 20.
9 A. M. A. Bennett, CHEMTECH, 1999, July, 24.
10 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 and D. J. Williams, J. Am.
Chem. Soc., 1999, 121, 8728.
11 B. L. Small and M. Brookhart, J. Am. Chem. Soc., 1998, 120, 7143.
12 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 and M. R. J.
Elsegood, Chem. Eur. J., 2000, 6, 2221.
30 T. Kauffmann, B. Laarmann, D. Menges and G. Neiteler, Chem.
Ber., 1992, 125, 163.
31 T. Kauffmann, Angew. Chem., Int. Ed. Engl., 1996, 35, 386.
32 W. K. Wong, K. W. Chiu, G. Wilkinson, A. J. Howes, M. Motevalli
and M. B. Hursthouse, Polyhedron, 1985, 4, 603.
33 W. Lau, J. C. Huffman and J. K. Kochi, Organometallics, 1982, 1,
155.
34 Y. Kubo, L. Sun Pu, A. Yamamoto and S. Ikeda, J. Organomet.
Chem., 1975, 84, 369.
35 T. Ikariya and A. Yamamoto, J. Chem. Soc., Chem. Commun., 1974,
720.
36 E. L. Dias, M. Brookhart and P. S. White, Organometallics, 2000, 19,
4995.
37 G. J. P. Britovsek, V. C. Gibson, and S. K. Spitzmesser, World Pat.,
WO 00/69869 (BP Chemicals), 2000; G. J. P. Britovsek, V. C. Gibson
and S. K. Spitzmesser, Chem. Abstr., 2001, 134, 5260.
38 V. C. Gibson, M. Humphries, K. P. Tellmann, D. F. Wass, A. J. P.
White and D. J. Williams, Chem. Commun., 2001, 2252.
39 D. P. Riley, J. A. Stone and D. H. Busch, J. Am. Chem. Soc., 1977,
99, 767.
40 P. N. Jagg, P. F. Kelly, H. S. Rzepa, D. J. Williams, J. D. Woollins and
W. Wylie, J. Chem. Soc., Chem. Commun., 1991, 942.
41 P. A. Cameron, V. C. Gibson, C. Redshaw, J. A. Segal, G. A. Solan,
A. J. P. White and D. J. Williams, J. Chem. Soc., Dalton Trans., 2001,
1472.
42 K. Shelly, T. Bartczak, W. R. Scheidt and C. A. Reed, Inorg. Chem.,
1985, 24, 4325.
43 J. -T. Ahlemann, H. W. Roesky, R. Murugavel, E. Parisini,
M. Noltemeyer, H.-G. Schmidt, O. Müller, R. Herbst-Irmer,
L. N. Markovskii and Y. G. Shermolovich, Chem. Ber., 1997, 130,
1113.
13 G. J. P. Britovsek, V. C. Gibson, S. Mastroianni, D. C. H. Oakes,
C. Redshaw, G. A. Solan, A. J. P. White and D. J. Williams,
Eur. J. Inorg. Chem., 2001, 431.
14 P. T. Matsunaga, World Pat., WO 99/57159 (Exxon Chemical Patents
Inc.), 1999; P. T. Matsunaga, Chem. Abstr., 1999, 131, 351802.
15 S. Al-Benna, M. J. Sarsfield, M. Thornton-Pett, D. L. Ormsby,
P. J. Maddox, P. Brès and M. Bochmann, J. Chem. Soc., Dalton
Trans., 2000, 4247.
44 D. F. Evans, J. Chem. Soc., 1959, 2003.
45 D. F. Evans, G. V. Fazakerley and P. R. F. Phillips, J. Chem. Soc. A,
1971, 1931.
16 D. M. Dawson, D. A. Walker, M. Thornton-Pett and M. Bochmann,
J. Chem. Soc., Dalton Trans., 2000, 459.
46 H. D. Crawford and J. Swanson, J. Chem. Educ., 1971, 48, 382.
47 D. H. Grant, J. Chem. Educ., 1995, 72, 39.
17 L. S. Moody, P. B. Mackenzie, C. M. Killian, G. G. Lavoie,
J. A. Ponasik Jr., A. G. M. Barrett, T. W. Smith, J. C. Pearson,
World Pat., WO 00/50470 (Eastman Chemical Company), 2000;
L. S. Moody, P. B. Mackenzie, C. M. Killian, G. G. Lavoie,
J. A. Ponasik Jr., A. G. M. Barrett, T. W. Smith and J. C. Pearson,
Chem. Abstr., 2000, 133, 208316.
18 R. Engehausen, W. Nentwig, P. Schertl, M. Arndt-Rosenau,
O. Pyrlik, M. Guillemot, Eur. Pat., EP1044992 (Bayer Aktienge-
sellschaft), 2000; R. Engehausen, W. Nentwig, P. Schertl,
M. Arndt-Rosenau, O. Pyrlik and M. Guillemot, Chem. Abstr.,
2000, 133, 310282.
48 J. -N. Pédeutour, D. Coevoet, H. Cramail and A. Deffieux,
Macromol. Chem. Phys., 1999, 200, 1215.
49 U. Wieser and H.-H. Brintzinger, in Organometallic Catalysts and
Olefin Polymerization (Catalysts for a New Millenium), ed. R. Blom,
A. Follestad, E. Rytter, M. Tilset and M. Ystenes, Springer Verlag,
Berlin, 2001, p. 3.
50 F. Peruch, H. Cramail and A. Deffieux, Macromolecules, 1999, 32,
7977.
51 P. Krumholz, Inorg. Chem., 1965, 4, 612.
52 B. de Bruin, E. Bill, E. Bothe, T. Weyhermüller and K. Wieghardt,
Inorg. Chem., 2000, 39, 2936.
19 K. Nomura, W. Sidokmai and Y. Imanishi, Bull. Chem. Soc. Jpn.,
2000, 73, 599.
20 D. D. Devore, S. S. Feng, K. A. Frazier, J. T. Patton, World Pat., WO
00/69923 (Dow Chemical Company), 2000; D. D. Devore, S. S. Feng,
K. A. Frazier and J. T. Patton, Chem. Abstr., 2000, 134, 5263.
21 K. Kreischer, J. Kipke, M. Bauerfeind and J. Sundermeyer, Z. Anorg.
Allg. Chem., 2001, 627, 1023.
53 D. Lin-Vien, N. Colthup, W. Fateley and J. G. Graselli, Infrared and
Raman Characteristic Frequencies of Organic Molecules, Academic
Press, San Diego, CA, 1991.
54 K. Nakamoto, Infrared and Raman Spectra of Inorganic
Coordination Compounds, Wiley, New York, 1989.
55 J. D. Curry, M. A. Robinson and D. H. Busch, Inorg. Chem., 1967, 6,
1570.
22 L. S. Boffa and B. M. Novak, Chem. Rev., 2000, 100, 1479.
23 A. Yamamoto, J. Chem. Soc., Dalton Trans., 1999, 1027.
24 A. Yamamoto, J. Polym. Sci. A, 1975, 9, 931.
25 A. Yamamoto, J. Organomet. Chem., 1986, 300, 347.
26 K. R. Kumar and S. Sivaram, Macromol. Chem. Phys., 2000, 201,
1513.
27 J. F. van Baar, P. A. Schut, A. D. Horton, T. Dall’occo, G. M. M. van
Kessel, World Pat., WO 00/35974, (Montell Technology Company
bv), 2000; J. F. van Baar, P. A. Schut, A. D. Horton, T. Dall’occo and
G. M. M. van Kessel, Chem. Abstr., 1999, 133, 59218.
28 E. A. H. Griffiths, G. J. P. Britovsek, V. C. Gibson and I. R. Gould,
Chem. Commun., 1999, 1333.
56 P. E. Figgins and D. H. Busch, J. Phys. Chem. A., 1961, 65, 2236.
57 D. A. Thornton, Coord. Chem. Rev., 1990, 104, 173.
58 S. Kawaguchi, Coord. Chem. Rev., 1986, 70, 51.
59 S. Komiya, M. Katoh, T. Ikariya, R. H. Grubbs, T. Yamamoto and
A. Yamamoto, J. Organomet. Chem., 1984, 260, 115.
60 C. J. Pouchert, The Aldrich Library of Infrared Spectra, Aldrich
Chemical Company Inc., Milwaukee, WI, 1970.
61 A. I. Popov, A. V. Scharabarin, V. F. Sukhoverkhov and
N. A. Tchumaevsky, Z. Anorg. Allg. Chem., 1989, 576, 242.
62 G. Desurmont, M. Tanaka, Y. Li, H. Yasuda, T. Tokimitsu, S. Tone
and A. Yanagase, J. Polym. Sci. A, 2000, 38, 4095.
63 H. D. Murdoch and D. C. Nonhebel, J. Chem. Soc., 1962, 2153.
J. Chem. Soc., Dalton Trans., 2002, 1159–1171
1171