Investigations into the mechanism of activation and initiation of ethylene polymerization by bis(imino)pyridine cobalt catalysts: Synthesis, structures, and deuterium labeling studies

Article abstract of DOI:10.1021/om049032b

The activation of bis(imino)pyridine cobalt(II) precatalysts by MAO leads initially to a bis(imino)pyridine cobalt(I) cationic species with no cobalt-C(alkyl) bond into which insertion can occur. Mechanistic studies have shown that the initiation of polym

Full text of DOI:10.1021/om049032b

Organometallics 2005, 24, 2039-2050
2039
Investigations into the Mechanism of Activation and
Initiation of Ethylene Polymerization by
Bis(imino)pyridine Cobalt Catalysts: Synthesis,
Structures, and Deuterium Labeling Studies
Martin J. Humphries, Kilian P. Tellmann, Vernon C. Gibson,*
Andrew J. P. White, and David J. Williams
Department of Chemistry, Imperial College London, Exhibition Road, London, SW7 2AZ, U.K.
Received December 8, 2004
The activation of bis(imino)pyridine cobalt(II) precatalysts by MAO leads initially to a
bis(imino)pyridine cobalt(I) cationic species with no cobalt-C(alkyl) bond into which insertion
can occur. Mechanistic studies have shown that the initiation of polymerization from this
species involves incorporation of alkyl groups from the cocatalyst, most likely involving attack
of methide anion (from the counteranion) on a cobalt-ethylene species.
Introduction
more commonly, a diene.6 In the iron-based polymeri-
zation system, there is good evidence for the binding of
trialkylaluminum reagents at the active center,7 which
explains the tendency for these systems to engage in
“chain transfer to metal” processes, although the oxida-
tion state of the metal center in the active species is
still open to debate.8
For cobalt, it would be tempting to expect that the
dihalide precursor would follow a “traditional” activation
pathway, leading to a cationic alkyl propagating species.
In recent years there has been much interest in the
discovery and development of olefin polymerization
catalysts based on the late transition metals.1 Following
the recognition of the potential of nickel complexes in
olefin polymerization,2 the subsequent discovery of
highly active catalysts based on iron and cobalt3 sign-
posted the way forward for further advances in the
design and applications of late transition metal systems.
While much is now known about the influence of ligand
substituents on the performance of iron- and cobalt-
based systems,3,4 there remains much to be understood
about the mechanism of activation and initiation of
polymerization.
In general, there are a number of common links
between late and early transition metal systems, both
in their activation pathways and the nature of the
propagating species. For example, catalyst activation
can usually be accomplished by treatment of a dihalide
metal precursor complex with an activator such as
MAO, leading initially to dialkylation of the metal
center, followed by methide group abstraction to afford
a cationic alkyl which, in the presence of a suitably
noncoordinating counteranion, functions as the propa-
gating species.5 It has also been shown for both late and
early transition metal species that a low-valent precur-
sor route can be used to generate the active site. This
involves attack by a strong Lewis acid, for example,
B(C6F5)3, MAO, or Al2Me6, on a coordinated olefin or,
(4) (a) Small, B. L.; Brookart, M. J. Am. Chem. Soc. 1998, 120, 7143.
(b) Britovsek, G. J. P.; Bruce, M.; Gibson, V. C.; Kimberley, B. S.;
Maddox. P. J.; Mastroianni, S.; McTavish, S. J.; Redshaw, C.; Solan,
G. A.; Stro¨mberg, S.; White, A. J. P.; Williams, D. J. J. Am. Chem.
Soc. 1999, 121, 8728. (c) 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. (d)
Ma, Z.; Wang, H.; Qiu, J.; Xu, D.; Hu, Y. Macromol. Rapid. Commun.
2001, 22, 1280. (e) Ma, Z.; Sun, W.-H.; Li, Z.-L.; Shao, C.-X.; Hu, Y.-
L., Li, X.-H. Polym. Int. 2002, 51, 994. (f) Britovsek, G. J. P.; Gibson,
V. C.; Hoarau, O. D.; Spitzmesser, S. K.; White, A. J. P.; Williams, D.
J. Inorg. Chem. 2003, 42, 3454. (g) Chen, Y.; Qian, C.; Sun, J.
Organometallics 2003, 22, 1231. (h) Schmidt, R.; Hammon, U.; Got-
tfried, S.; Welch, M. B.; Alt, H. G. J. Appl. Polym. Sci. 2003, 88, 476.
(i) Abu-Surrah, A. S.; Lappalainen, K.; Piironen, U.; Lehmus, P.; Repo,
T.; Leskela¨, M. J. Organomet. Chem. 2002, 55. (j) Chen, Y.; Chen, R.;
Qian, C.; Dong, X.; Sun, J. Organometallics 2003, 22, 4312. (k)
Ivanchev, S. S.; Tolstikov, G. A.; Badaev, V. K.; Oleinik, I. I.; Ivancheva,
N. I.; Rogozin, D. G.; Oleinki, I. V.; Myakin, S. V. Kinet. Catal. 2004,
45, 176. (l) Paulino, I. S.; Schuchardt, I. J. Mol. Catal. A 2004, 211,
55. (m) Bluhm, M. E.; Folli, C.; Do¨ring, M. J. Mol. Catal. A 2004, 212,
13. (n) Smit, T. M.; Tomov, A. K.; Gibson, V. C.; White, A. J. P.;
Williams, D. J. Inorg. Chem. 2004, 43, 6511. (o) Tellmann, K. P.;
Gibson, V. C.; White, A. J. P.; Williams, D. J. Organometallics 2005,
24, 280.
(5) (a) Siedle, A. R.; Lamanna, W. M.; Newark, R. A.; Schroepfer, J.
N. J. Mol. Catal. A 1998, 128, 257. (b) Svejda, S. A.; Johnson, L. K.;
Brookhart, M. J. Am. Chem. Soc. 1999, 121, 10634.
(6) (a) For an example on nickel see: Strauch, J. W.; Kehr, G.; Erker,
G. J. Organomet. Chem. 2003, 683, 249. (b) For an overview see: Erker,
G. Acc. Chem. Res. 2001, 34, 47.
(7) (a) Talsi, E. P.; Babushkin, D. E.; Semikolenova, N. V.; Zudin,
V. N.; Panchenko, V. N.; Zakharov, V. A. Macromol. Chem. Phys. 2001,
202, 2046. (b) Semikolenova, N. V.; Zakharov, V. A.; Talsi, E. P.;
Babushkin, D. E.; Sobolev, A. P.; Echevskaya, L. G.; Khysniyarov, M.
M. J. Mol. Catal. A 2002, 182-183, 283. (c) Bryliakov, K. P.;
Semikolenova, N. V.; Zakharov, V. A.; Talsi, E. P. Organometallics
2004, 23, 5375.
* To whom correspondence should be addressed. E-mail: v.gibson@
imperial.ac.uk.
(1) (a) Britovsek, G. J. P.; Gibson, V. C.; Wass, D. F. Angew. Chem.,
Int. Ed. 1999, 38, 428. (b) Ittel, S. D.; Johnson, L. K.; Brookhart M.
Chem. Rev. 2000, 100, 1169. (c) Gibson, V. C.; Spitzmesser, S. K. Chem.
Rev. 2003, 103, 283.
(2) (a) Starzewski, K. A. O.; Witte, J. Angew. Chem., Int. Ed. Engl.
1985, 24, 599. (b) Johnson, L. K.; Killian, C. M.; Brookhart, M. J. Am.
Chem. Soc. 1995, 117, 6414. (c) Johnson, L. K.; Killian C. M.; Bennett,
A. M. A.; Coughlin, E. B.; Ittel, S. D.; Parthasarathy, A.; Tempel, D.
J.; Brookhart, M. S. WO 9603010, 1996.
(3) (a) Small, B. L.; Brookhart, M.; Bennett, A. M. A. J. Am. Chem.
Soc. 1998, 120 , 4049. (b) Britovsek, G. J. P.; Gibson, V. C.; Kimberley,
B. S.; Solan, G. A.; White, A. J. P.; Williams, D. J. Chem. Commun.
1998, 849.
(8) (a) Britovsek, G. J. P.; Clentsmith, G. K. B.; Gibson, V. C.;
Goodgame, D. M. L.; McTavish, S. J.; Pankhurst, Q. A. Catal. Commun.
2002, 3, 207. (b) Bryliakov, K. P.; Semikolenova, N. V.; Zudin, V. N.;
Zakharov, V. A.; Talsi, E. P. Catal. Commun. 2004, 5, 45.
10.1021/om049032b CCC: $30.25 © 2005 American Chemical Society
Publication on Web 03/24/2005

2040 Organometallics, Vol. 24, No. 9, 2005
Humphries et al.
Scheme 1. Summary of the Syntheses of Cobalt(I)
Complexesa
However, it has been shown in preliminary studies that
the LCoCl2 precursors, rather than undergoing dialky-
lation to give LCoR2, are first reduced by the alkylating
agent to LCoCl and then alkylated by the same reagent
to give cobalt(I) alkyl species.9 These, however, do not
insert olefins. Rather, in the presence of a strong Lewis
acid, the alkyl group is abstracted to afford cationic
cobalt(I) species that are capable of binding ethylene9a
and that are the direct precursors to the propagating
species. In this paper we describe the synthesis and
characterization of the key intermediates in the activa-
tion process and describe studies directed toward ob-
taining an improved understanding of the mechanism
by which cationic cobalt(I) complexes convert to an
active propagating species.
Results
Synthesis and Characterization. Our initial in-
vestigations into the cobalt system commenced with
attempts to synthesize a cobalt(II) dimethyl complex,
with the anticipation of studying reactions of this
complex with well-defined activating agents. The syn-
thesis of the cobalt(II) species was attempted by the
reaction of [2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N]CoCl2
(1) with 2 equiv of alkylating agents such as MeLi,
MeMgX, and Me2Mg. In no case was the anticipated
dimethyl complex isolated. Rather, a diamagnetic
Co(I) complex was isolated and characterized as [2,6-
{MeCdN(2,6-iPr2C6H3)}2C5H3N]CoMe (2). The synthe-
ses of this and other cobalt species in this work are
summarized in Scheme 1. If the reaction with Grignard
reagents (X ) Cl, Br) did not go to completion, NMR
spectra of the crude products showed the presence of
the cobalt halides [2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N]-
CoCl (3) or [2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N]CoBr
(4) as intermediates, implying that the first step in this
reaction is reduction to the cobalt(I) halide, followed by
alkylation. The cobalt(I) chloride 3 can be accessed
directly via the reduction of 1 by excess zinc in toluene
at 50 °C. Upon treatment of 3 with trimethylsilylhalides,
halogen exchange occurs to afford the bromo (4) and iodo
(5) derivatives; these were not isolated but characterized
in situ by NMR spectroscopy.
a (i) Zn, toluene, 50 °C (Ar ) 2,6-iPr2C6H3 3; 2,4,6-Me3C6H2
10) or C8K, toluene, - 78 °C (Ar ) 2-EtC6H4 12); (ii) SiMe3X,
25 °C (Ar ) 2,6-iPr2C6H3, X ) Br 4, I 5); (iii) RMgX, Et2O, -
78 °C (Ar ) 2,6-iPr2C6H3, R ) Me 2, Et 6, nPr 7, CH2Ph 8; Ar
) 2,4,6-Me3C6H2, R ) Me 11); (iv) B(C6F5)3, 25 °C (Ar ) 2,6-
iPr2C6H3); (v) H2, 25 °C (Ar ) 2,6-iPr2C6H3); (vi) C2H4 or C3H6,
25 °C (Ar ) 2,6-iPr2C6H3) to afford 6 or 7, respectively; (vii)
C2H4, 25 °C (Ar ) 2,6-iPr2C6H3).
ally, no change in chemical shift was observed upon
cooling a sample in toluene-d8 to 193 K, nor upon the
addition of a donor solvent such as THF or diethyl ether.
It can reasonably be concluded that these unusual
chemical shifts are due to ring current effects arising
from the ligand aryl substituents, rather than due to
the presence of â-agostic interactions.
Other cobalt alkyl species can also be accessed
using Grignard reagents, including [2,6-{MeCdN(2,6-
iPr2C6H3)}2C5H3N]CoEt10
(6),
[2,6-{MeCdN(2,6-
iPr2C6H3)}2C5H3N]ConPr10 (7), and [2,6-{MeCdN(2,6-
iPr2C6H3)}2C5H3N]CoCH2Ph (8); the latter complex has
also been described by Gal and co-workers.9b The
resonances of the protons in the â-positions of the alkyl
chains show unusually high-field chemical shifts (δ
The cobalt(I) alkyl complexes containing hydrogen
atoms in the â-position of the chain are not stable over
long periods, decomposing to unidentified diamagnetic
products. The only volatile species isolable from these
decomposition reactions are the corresponding alkanes,
presumably arising via ligand cyclometalation reactions.
The decomposition is slowed significantly in the pres-
ence of ethers and alkenes. A full description of the
chemistry of these longer chain cobalt alkyl species has
been reported elsewhere.10
1
-1.18 to -0.73) with JC-H coupling constants in the
range expected for nonagostic alkyl groups (123-125
Hz). That the 1JC-H coupling constants are virtually the
same for 6 and 7 also indicates that there is no fluxional
agostic interaction; a shift to a lower value would be
expected on exchanging ethyl for propyl due to averag-
ing over two rather than three â-hydrogens.11 Addition-
(9) (a) Gibson, V. C.; Humphries, M. J.; Tellmann, K. P.; Wass, D.
F.; White, A. J. P.; Williams, D. J. Chem. Commun. 2001, 2252. (b)
Kooistra, T. M.; Knijenburg, Q.; Smits, J. M. M.; D. Horton, A.;
Budzelaar, P. H. M.; Gal, A. W. Angew. Chem., Int. Ed. 2001, 40, 4719.
(10) (a) Gibson, V. C.; Tellmann, K. P.; Humphries, M. J.; Wass, D.
F. Chem. Commun. 2002, 2316. (b) Tellmann, K. P.; Humphries, M.
J.; Rzepa, H. S.; Gibson, V. C. Organometallics 2004, 23, 5503.
The cobalt(I) hydride [2,6-{MeCdN(2,6-iPr2C6H3)}2-
C5H3N]CoH (9) can be prepared via the reaction of 2
(11) Brookhart, M.; Green, M. L. H.; Wong, L.-L. Progr. Inorg. Chem.
1988, 36, 1.

Bis(imino)pyridine Cobalt Catalysts
Organometallics, Vol. 24, No. 9, 2005 2041
Table 1. Variation in the Chemical Shifts of the Protons of the Ketimine Methyl Group and the
para-Position of the Pyridine Ring with Substitution at the Cobalt Center
complex
δ (ketimine)
δ (para-H, py)
[2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N]CoH,a 9
-1.66
-1.33
-1.17
-1.14
-0.37
-0.08
0.05
10.82
10.25
10.26
10.19
10.15
9.79
[2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N]CoEt,a 6
[2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N]CoCH2Ph,a 8
[2,6-{MeCdN(2,6-iPr2C6H3) 2C5H3N]CoMe,a 2
[2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N]CoI,a 5
[2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N]CoBr,a 4
[2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N]CoCl,a 3
9.54
[{{2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N}Co(C2H4)}{MeB(C6F5)3}],b 14
[{{2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N}Co(N2)}{MeB(C6F5)3}],b 13
0.75
8.00
7.55
1.11
a In benzene-d6. b In toluene-d8.
Table 2. Comparative Selected Bond Lengths (Å)
and Angles (deg) for 2, 3, 10, and 13
with dihydrogen, producing methane as a byproduct.10,12
9 is insufficiently stable to allow its isolation, and so
characterizing data have been obtained on samples
prepared in situ. Attempts to synthesize the deuteride
2
3
10
[X ) Cl]
13
[X ) N2]
[X ) Me]
[X ) Cl]
analogue via the reaction of 2 with 2H2 were unsuccess-
ful due to scrambling of the deuterium into the isopropyl
substituents of the ligand, providing a further indication
that reversible ligand cyclometalation processes occur
in these cobalt(I) bis(imino)pyridine complexes.
Co-X
1.960(4)
1.833(3)
1.907(3)
1.903(3)
1.335(4)
1.329(5)
1.443(5)
1.434(5)
1.445(4)
1.441(5)
1.366(4)
1.365(4)
2.1807(10) 2.1672(13) 1.841(3)
Co-N(1)
Co-N(7)
Co-N(9)
C(7)-N(7)
C(9)-N(9)
C(2)-C(7)
C(6)-C(9)
N(7)-Ar
1.797(3)
1.916(3)
1.912(3)
1.317(5)
1.322(5)
1.444(5)
1.435(5)
1.449(5)
1.440(5)
1.373(5)
1.370(5)
1.793(3)
1.914(3)
1.914(3)
1.328(5)
1.325(4)
1.432(5)
1.442(5)
1.437(4)
1.443(4)
1.365(4)
1.375(4)
1.812(3)
1.915(3)
1.908(3)
1.303(5)
1.303(5)
1.465(5)
1.459(5)
1.452(4)
1.444(4)
1.356(4)
1.354(4)
The chemistry outlined above is not unique to the di-
(isopropyl)phenyl-substituted ligand. Analogous com-
plexes containing the N-mesityl ligand, [2,6-{MeCdN-
(2,4,6-Me3C6H2)}2C5H3N]CoCl (10) and [2,6-{MeCdN-
(2,4,6-Me3C6H2)}2C5H3N]CoMe (11), and, for reasons
that will become clear shortly, the N-2-EtC6H4 deriva-
tive, [2,6-{MeCdN(2-EtC6H4)}2C5H3N]CoCl (12), have
also been prepared and characterized.
N(9)-Ar
N(1)-C(2)
N(1)-C(6)
N(1)-Co-X
N(7)-Co-X
N(9)-Co-X
178.94(17) 179.15(10) 176.65(11) 178.3(2)
98.95(14) 98.89(9)
98.70(15) 98.17(9)
99.40(9)
97.25(9)
98.23(14)
98.44(14)
N(1)-Co-N(7) 81.22(12) 81.35(14) 81.46(12) 81.67(12)
N(1)-Co-N(9) 81.11(12) 81.58(14) 81.96(12) 81.68(12)
N(7)-Co-N(9) 162.32(12) 162.92(12) 163.35(12) 163.31(12)
Certain features of the 1H NMR spectra of the cobalt-
(I) complexes are worthy of note. The chemical shifts of
certain ligand resonances are highly dependent upon
the nature of the substitution at the cobalt center, with
the singlet resonance of the ketimine methyl group and
the triplet resonance of the proton in the para-position
of the pyridine ring being particularly sensitive. The
shift of the ketimine resonance correlates with the
electrophilicity of the cobalt center (Table 1). Thus, the
ketimine resonance of the cation 13 (vide infra), at δ
+1.11, compares with shifts of δ 0.05 to -0.37 for the
halide derivatives 3-5, δ -1.14 to -1.33 for alkyl
derivatives 2, 6, and 8, and δ -1.66 for complex 9. The
resonance of the proton in the para-position of the
pyridine ring shows the opposite trend, being δ 7.55 and
10.19 for 13 and 2, respectively. The variation in the
chemical shifts has been explained by the presence of a
low-lying triplet state that can be thermally populated
from the singlet ground state.12,13
C(7)-N(7)-Ar 118.4(3)
C(9)-N(9)-Ar 119.1(3)
118.6(3)
119.5(3)
118.2(3)
118.2(3)
119.9(3)
119.5(3)
coordination plane extends to include the ligand back-
bone, the maximum deviations from planarity for the
(C5H3N)(MeCdN-Ci)2Co units being ca. 0.12 and 0.08
Å, respectively, cf. 0.03 and 0.08 Å in 3 and 13, respec-
tively. The pendant aryl rings are, as usual for bis-
(imino)pyridine ligands, inclined almost orthogonally to
the planes of their parent nitrogen atoms, the
N(7)-Ar, N(9)-Ar torsion angles being ca. 85°, 87° and
89°, 86° in 2 and 10, respectively. (The geometries at
these nitrogen centers are all trigonal planar, the cen-
tral nitrogen lying in each case no more than ca. 0.04 Å
out of its associated {C2Co} plane.) Inspection of the
In our earlier communication9a we reported the solid
state structures of 3 and 13 (see Supporting Informa-
tion, Figures S3 and S4), and we have now also
determined the X-ray structures of crystals of 2 and 10
(comparative selected bond lengths and angles for all
four complexes are given in Table 2). The core structures
of 2 (Figure 1) and 10 (Figure 2) are very similar to
those of 3 and 13, having approximate C2v symmetry
about the Co-N(py) bond. In both 2 and 10 the metal
(12) Knijnenburg, Q.; Hetterscheid, D.; Kooistra, T. M.; Budzelaar,
P. H. M. Eur. J. Inorg. Chem. 2004, 1204.
(13) While this description of the systems as cobalt(II) bound to a
ligand radical may be correct, we will continue to refer to these systems
as Co(I) in this article in order to avoid confusion with cobalt(II)
complexes such as 1. The distribution of electrons within the cobalt(I)
complexes does not affect our later mechanistic experiments and
conclusions.
Figure 1. Molecular structure of 2.

2042 Organometallics, Vol. 24, No. 9, 2005
Humphries et al.
one vinyl and one methyl end group per chain.3 This is
consistent with the major termination process being â-
hydrogen transfer, either via the metal center or directly
to monomer. It is thus clear that significant amounts
of chain-transfer to the aluminum cocatalyst do not
occur for cobalt, unlike for closely related iron systems.3
The well-established mode of activation for both early
and late transition metal complexes involves the trans-
formation of a catalyst precursor LnMX2 into an electron-
deficient metal alkyl species [LnM-R]+ by the action of
MAO.5 If this mode of activation were followed by the
cobalt system, the formation of a 15-electron cobalt(II)
species [LCo-R]+ would be anticipated. However, as
described in the previous section, attempts to synthesize
this putative species led to cobalt(I) products. Indeed,
activation of the cobalt(I) chloride complex 3 with MAO
affords a catalytically active species that displays the
same activity as 1/MAO; the resultant polymer also has
identical properties (microstructure, molecular weight,
PDI, end groups; see Figure S5 for GPC traces). The
same catalytically active species thus appears to be
present in each case, and in view of the reduction
chemistry, the first species formed in the activation
process is presumed to be [2,6-{MeCdN(2,6-iPr2-
C6H3)}2C5H3N]CoCl (3), which is then alkylated to form
[2,6-{MeCdN(2,6-iPr2C6H3)2C5H3N]CoMe (2); subse-
quent transformations then form the active species.
Similar conclusions were reached upon studying the
reaction between 1 with small amounts of MAO.9b
Addition of ethylene to a solution of 2 in toluene does
not lead to the formation of polymer, nor are any
changes observed in the NMR spectrum of 2 under an
atmosphere of ethylene. It is clear, therefore, that 2 is
not the active species. Those alkyl compounds that
have protons in the â-position of the chain (6 and 7)
undergo exchange reactions with ethylene, but not
insertion. Studies of this process have been reported
separately.10 The cobalt(I) hydride [2,6-{MeCdN(2,6-
iPr2C6H3)}2C5H3N]CoH (9) does react with ethylene, but
only undergoes a single insertion to afford 6.10 It is thus
clear that the Co(I) alkyl complexes are not the chain-
propagating species in ethylene polymerization.
Reactions of Cobalt(I) Alkyls with Lewis Acids.
Treatment of 2 with Al2Me6 does not activate it toward
ethylene polymerization. However, in situ NMR studies
reveal that alkyl group exchange occurs between the
aluminum and the cobalt centers. Although slow on the
NMR time scale, addition of Al2Me6-d18 to a sample of
2 afforded a statistical distribution of protio- and
deuterio-methyl groups between the two metals. Simi-
larly, addition of AlEt3 afforded a mixture of 2 and 6. It
would thus appear that Al2Me6 is not a sufficiently
strong Lewis acid to activate 2 toward ethylene polym-
erization, and a stronger Lewis acid (e.g., MAO) is
needed. We therefore decided to investigate the reaction
of 2 with the strong Lewis acid B(C6F5)3. Upon mixing,
an immediate color change occurred from claret to royal
blue; the product of this reaction was identified as [{2,6-
{MeCdN(2,6-iPr2C6H3)}2C5H3N}Co(N2)][MeB(C6F5)3],
13-MeBArf3.14 The molecular structure of this com-
pound has been reported previously and shows the
Figure 2. Molecular structure of 10.
comparative table of bond lengths and angles reveals
that there are few instances where statistically signifi-
cant differences of comparable parameters can be seen.
In particular, there is no obvious variation in the angle
of the aryl rings to the imine bond with variation of the
bulk of the fourth substituent at the cobalt center (see
the C(7)-N(7)-Ar and C(9)-N(9)-Ar angles). For the
Co-N(1) bond to the central pyridyl ring, however, the
distance seen in 2 is markedly longer than those seen
in 3, 10, and 13. The Co-N(imine) bonds are all
statistically the same, although the imine bonds them-
selves in 13 are noticeably (though not significantly)
shorter than in the other three species. For the bond
angles, the only feature of note is a slight lateral
displacement of the X ligand in 10 [where X ) Cl]; in
the other three structures the two N(imine)-Co-X
angles are the same to within ca. 0.72°, but in 10 they
are different by ca. 2.15°. Associated with this is a
markedly more distorted N(1)-Co-X trans angle of
176.65(11)°, cf. 178.94(17)°, 179.15(10)°, and 178.3(2)°
for 2, 3, and 13, respectively.
The packing of molecules in the crystals of 2 is
dominated by intermolecular C-H‚‚‚π contacts. One
of the meta protons on ring B in one molecule ap-
proaches the centroid of ring A in another (H‚‚‚A 2.84
Å, C-H‚‚‚A 136°, vector inclined by ca. 76° to the ring
plane), the para proton of ring C approaches ring B
(H‚‚‚B 3.01 Å, C-H‚‚‚B 165°, vector inclined by ca. 70°
to the ring plane), and one of the C(7)-Me protons
approaches ring C (H‚‚‚C 2.96 Å, C-H‚‚‚C 166°, vector
inclined by ca. 75° to the ring plane). The effect of these
interactions is to form a two-dimensional sheet of
molecules in the 101 plane. The most notable intermo-
lecular interactions in the structure of 10 involve the
pyridyl ring A. There is a weak π-π contact between
this ring in one molecule and ring B in a glide related
counterpart (centroid‚‚‚centroid and mean interplanar
separations of ca. 4.25 and 3.42 Å, respectively, rings
inclined by ca. 5°), and the disposition of the two rings
is such that one of the para methyl protons on ring B
comes closest to the centroid of ring A (H‚‚‚A 3.00 Å,
C-H‚‚‚A 124°, vector inclined by ca. 73° to the ring
plane). The opposite face of ring A is approached by one
of the meta methyl protons on ring C of a lattice
translated molecule (H‚‚‚A 3.09 Å, C-H‚‚‚A 123°, vector
inclined by ca. 73° to the ring plane), the two C-H‚‚‚π
contacts subtending an angle of ca. 147° at the centroid
of ring A. As was seen in 2, the combined effect of these
intermolecular interactions is the formation of a two-
dimensional sheet.
Mechanism of Catalyst Activation and Initiation
of Polymerization. Catalyst Activation. The poly-
ethylene produced by the cobalt system is linear, with
(14) This nomenclature is used so as to be able to refer to the cationic
cobalt fragment as 13, whether synthesized by the use of B(C6F5)3 or
MAO; the MAO-activated cation will be referred to as 13-MAO.

Bis(imino)pyridine Cobalt Catalysts
Organometallics, Vol. 24, No. 9, 2005 2043
presence of an end-on bonded dinitrogen ligand.9a The
[MeB(C6F5)3] anion does not show any close interaction
with the cobalt center in the solid state. This also
appears to be the case in solution, where the 19F NMR
data are indicative of an anion that is not coordinated
to the metal center.15
Reactions of Cobalt(I) Cations with Ethylene.
Exposure of a dark blue solution of 18 µmol of
13-MeBArf3 in toluene to ethylene resulted in a color
change to aquamarine, concomitant with the onset of
polymerization. Over a 20 min run 130 mg of polyeth-
ylene was produced, corresponding to an activity of 11
g/mmol bar h. Repeating this reaction in the presence
of 5 equiv of Al2Me6 as co-activator resulted in the
formation of 1.12 g of polymer (96 g/mmol bar h), an
order of magnitude increase in activity. The polymer
formed has properties identical to that produced by
1/MAO.
NMR spectroscopy is a useful tool for the study of
these cobalt(I) species due to their diamagnetic nature
and especially due to the aforementioned sensitivity of
certain ligand proton resonances to groups attached to
the cobalt center. The chemical shift of the ketimine
methyl group is particularly useful in this respect,
allowing even very closely related species to be distin-
guished. Addition of 1.0 equiv of ethylene to a solution
of the cation afforded a species identifiable from its 1H
NMR spectrum as [{2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N}-
Co(η-C2H4)][MeB(C6F5)3] (14-MeBArf3). The ketimine
resonance is now found at δ 0.75 (cf. δ 1.11 for the
dinitrogen adduct), with the resonance attributable to
bound ethylene occurring as a sharp singlet at δ 4.65
(cf. δ 5.24 for free ethylene). The small shift in the
ethylene resonance upon binding to the cobalt center
implies that the ethylene is only weakly bound; indeed,
it can be removed by degassing the NMR sample.
Addition of further equivalents of ethylene resulted in
a rapid exchange between free and bound ethylene,
affording an averaged chemical shift for all of the
ethylene present in the system. The excess ethylene was
consumed over time, resulting in the formation of
polyethylene on the walls of the NMR tube. Once all
excess ethylene had been consumed, the ethylene-
adduct 14-MeBArf3 remained. No other organometallic
species were discernible in the 1H NMR spectrum
during the polymerization.
The direct study of MAO-activated systems by NMR
spectroscopy is generally complicated by the presence
of an intense, broad resonance in the high-field region
of the spectrum attributable to a variety of averaging
Me-Al environments. Taking advantage of the highly
sensitive nature of the chemical shifts of the ketimine
methyl resonances, we decided to use ligands labeled
with deuterium at their ketimine positions. These
ligands are readily accessible from diacetylpyridine-d6,16
and 1-d6, 2-d6, and 3-d6, have been synthesized by
routes described for their nonlabeled analogues. The
activation of these compounds with MAO and the
subsequent interaction of the activated species with
ethylene can thus be conveniently studied using 2H
NMR spectroscopy, since only the ketimine resonance
is observed for each species formed.
The reaction between 2-d6 and B(C6F5)3 in 2-chloro-
toluene17 afforded 13-d6-MeBArf3, for which the 2H
NMR spectrum contained a sharp resonance at δ 1.2,
1
consistent with the H NMR spectrum of its per-protio
analogue. Introduction of an atmosphere of ethylene to
the NMR tube caused the resonance to shift to δ 0.83,
a shift indicative of the ethylene adduct 14-d6-MeBArf3.
The activation of 3-d6 in 2-chlorotoluene using dry
18
MAO/Al2Me6 (Al:Co ) 14:1) afforded a species that
displayed a ketimine methyl resonance at δ 0.2. A
slightly broader resonance but at a similar chemical
shift was also obtained upon activation of 1-d6 with the
same MAO solution and the same Al:Co ratio. A
ketimine methyl resonance at this chemical shift,
partway between that of the cobalt(I) chloride 3-d6 and
the cobalt(I) cation 13-d6-MeBArf3, is likely to be the
result of the interaction of the anion ([Cl-MAO]- or
[Me-MAO]- formed during the activation process) with
a cationic cobalt center. Addition of ethylene to the
NMR tube gave rise to a ketimine resonance at ca. δ
0.9, a chemical shift very close to that observed for
14-d6-MeBArf3. Thus, it can be reasonably concluded
that addition of ethylene to the MAO-activated cata-
lyst also results in the formation of the cationic eth-
ylene species [{2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N}Co-
(η-C2H4)]+, 14. These observations are summarized in
Scheme 2.
Unlike the B(C6F5)3-activated system, once the eth-
ylene in the MAO-activated system is consumed, the
initial species is regenerated; that is, the resonance at
δ 0.2 replaces the resonance at δ 0.9. This reflects
the stronger coordinating ability of [Cl-MAO]- or
[Me-MAO]- over [MeB(C6F5)3]-, the anions derived
from MAO being able to displace the bound ethylene,
whereas the more weakly coordinating [MeB(C6F5)3]-
counteranion cannot.
There are some initial conclusions that can be drawn
from these results. First, the species that is initially
formed upon activation, and thus the species from
which the initiation of polymerization must occur, is a
cobalt(I) cation, a species that contains no cobalt-
C(alkyl) σ-bond. Second, the MAO cocatalyst appears
to perform a number of important functions: (i) reduc-
tion of the cobalt(II) precatalyst to cobalt(I), (ii) alky-
lation of the cobalt(I) halide to cobalt(I) methyl, (iii)
abstraction of a chloride or methide unit to give a cobalt-
(I) cation, and (iv) stabilization of the active catalyst,
most likely through the scavenging of impurities. It is
also possible that the Al2Me6 present in the MAO may
enhance the rate of polymerization, as evidenced by the
greater activity found upon addition of Al2Me6 to
13-MeBArf3.
Mechanism of Initiation. Plausible pathways for
transforming a cobalt cation with a bound ethylene into
a complex with a Co-C(alkyl) bond are shown in
Scheme 3. Pathway I involves nucleophilic attack by
methide on the cobalt-bonded ethylene to give a cobalt-
(I) alkyl species. The methide may be derived from the
(17) The MAO-activated system was insufficiently soluble in toluene
to allow 2H NMR data to be obtained. Although the cobalt(I) complexes
are generally unstable in chlorinated and brominated solvents, leading
to unidentified paramagnetic products, no decomposition is seen in the
sterically protected 2-chlorotoluene over a period of at least 5 h.
(18) Dry MAO is the white solid obtained by removal of solvent and
trialkylaluminums from a solution of MAO, followed by repeated
slurrying in heptane and removal of the volatile components.
(15) Horton, A. D. Organometallics 1996, 15, 2675
(16) Clentsmith, G. K. B.; Gibson, V. C.; Hitchcock, P. B.; Kimberley,
B. C.; Rees, C. W. Chem. Commun. 2002, 1498.

2044 Organometallics, Vol. 24, No. 9, 2005
Scheme 2. Summary of H NMR Study of the Active System
Humphries et al.
2
[MeBArf3]- or [Me-MAO]- counteranion, or another
Al-Me bond. II is a closely related transformation, but
here the Lewis acid remains associated with the alkyl
group to form a bimetallic Co-Al/B complex, similar to
cobalt(I) propyl complex 7. It is already known that
cobalt(I) alkyls such as 7 do not polymerize ethylene.
Pathway I can thus be ruled out. In II, the active site
would effectively be a cobalt(I) alkyl interacting with a
Lewis acid. But, we already know that the cobalt(I)
alkyls are not active for ethylene polymerization in the
presence of Al2Me6. Rather, alkyl group exchange is
observed. Additionally, in the presence of B(C6F5)3 or
MAO, the cobalt-bonded alkyl group is abstracted to
afford cationic cobalt(I) products rather than engaging
in an interaction of the type found in II, and so this
pathway can also be excluded.
In addressing pathway III, separate solutions of
13-MeBarf3 were prepared via reactions of 18 µmol of
2 with 0.98 and 5.0 equiv of B(C6F5)3, respectively; these
were then exposed to 1 atm of ethylene. No difference
in activity was observed between the two; both afforded
130 mg of polymer over a 20 min run, corresponding to
activities of 11 g/mmol bar h. It is thus clear that the
presence of excess Lewis acid is not required to initiate
polymerization, implying that route III can be dis-
counted as a mechanism for the initiation of polymer-
ization. The low activities obtained may be due to the
lack of any scavenger to remove impurities in the
ethylene feed; the order of magnitude increase in
activity upon adding small amounts of alkyl aluminums
(vide supra) is consistent with this. A second cobalt
center acting as a Lewis acid is highly unlikely due to
the steric hindrance of the ligand system, the very low
concentrations of the cobalt species in solution, and the
the active species originally proposed by Natta.19
A
species of this type has recently been proposed to
explain the activity of cationic zirconium complexes with
no apparent Zr-alkyl bond.20 III involves electrophilic
attack of a Lewis acid on the bound ethylene in a fashion
analogous to Erker’s activation of zirconocene and nickel
butadiene complexes, leading to a cobalt(III) center.6
This Lewis acid may be the cocatalyst, B(C6F5)3 or MAO,
or possibly another cobalt cation; the latter would lead
initially to a bimetallic Co(I)/Co(III) species. In IV,
oxidative coupling of two molecules of ethylene occurs
to afford a cobaltacyclopentane species, where the metal
center is now in the +3 oxidation state. Such a pathway
would be analogous to the commonly postulated initial
step of the Phillips and other chromium-based polym-
erization systems.21 V involves the transfer of a methide
group to the cobalt center, concomitant with electro-
philic attack of the Lewis acid on the bound ethylene,
leading to a net oxidation to cobalt(III). Insertion could
potentially occur into either, or both, of the cobalt-
carbon bonds. In VI, C-H activation of the bound
ethylene molecule occurs to afford a cobalt(III) vinyl
hydride species, a reaction that has precedence in the
chemistry of the group 9 metals22 and that has been
proposed previously as an activation step in the dimer-
ization of ethylene by a Co(I) tris(triphenylphosphine)
cation.23
Studies to Distinguish among Pathways I-VI.
Nucleophilic attack by methide on the bound ethylene,
as shown in pathway I, leads to the formation of the
(22) See, for example: (a) Stoutland, P. O.; Bergman, R. G. J. Am.
Chem. Soc. 1985, 107, 4581. (b) Bely, S. T.; Duckett, S. B. Organome-
tallics 1989, 8, 748. (c) Ghosh, C. K.; Hoyano, J. K.; Krentz, R.; Graham,
W. A. G. J. Am. Chem. Soc. 1989, 111, 5480. (d) Tanke, R. S.; Crabtree,
R. H. Inorg. Chem. 1989, 28, 3444. (e) Papenfuhs, B.; Mahr, N.; Werner,
H. Organometallics 1993, 12, 4244. (f) Gutie´rrez-Puebla, E.; Monge,
A.; Nicasio, M. C.; Pere´z, P. J.; Poveda, M. L.; Rey, L.; Ru´ız, C.;
Carmona, E. Inorg. Chem. 1998, 37, 4538.
(19) Natta, G.; Mazzanti, G. Tetrahedron 1960, 8, 86.
(20) (a) Jin, J.; Wilson, D. R.; Chen, E. X.-Y. Chem. Commun. 2002,
708. (b) Cano, J.; Royo, P.; Lanfranchi, M.; Pellinghelli, M. A.;
Tiripicchio, A. Angew. Chem., Int. Ed. 2001, 40, 2495.
(21) McDaniel, M. P. Adv. Catal. 1985, 33, 47.
(23) Kawakami, K.; Mizoroki, T.; Ozaki, A. Bull. Chem. Soc. Jpn.
1978, 51, 21.

Bis(imino)pyridine Cobalt Catalysts
Organometallics, Vol. 24, No. 9, 2005 2045
Scheme 3. Possible Mechanisms for the Initiation
of Polymerization from the Cobalt(I) Cation 14 (LA
represents a Lewis acid)
Figure 3. Mass spectrum of the 1-hexene fraction obtained
from the co-oligomerization of C2H4 and C2D4 using 12/dry
MAO.
deuterium labels. 12/MAO was treated with a 50:50
mixture of C2H4 and C2D4, and the reaction terminated
after 75 s by vacuum transfer of all the volatile species.
1
The H NMR spectrum of the product mixture showed
the presence of R-olefins and unreacted ethylene, both
with complex coupling patterns consistent with H/D
coupling. Analysis of the solution by GC-MS confirmed
that the oligomers were R-olefins and showed that all
possible H/D isotopomers were present, with the dis-
tribution approaching the statistical ratios expected for
a 1:1 ratio of H and D. The mass spectrum of the C6
fraction is shown in Figure 3.
GC-MS analysis of the ethylene gas remaining above
the polymerization mixture (prior to transfer of the
volatiles) revealed peaks at m/z 29, 30, and 31 corre-
sponding to C2H3D, C2H2D2, and C2HD3, respectively.
This indicates that H/D scrambling between the protio
and deuterio ethylene had occurred. Two control experi-
ments were undertaken, one using pure C2H4, and the
other using pure C2D4. In neither of these experiments
were isotopic mixtures observed, ruling out the possible
incorporation of hydrogen isotopes from other sources,
e.g., from ligand C-H activation processes or from
contamination of the feedstock.
The presence of all possible isotopomers in the reac-
tion products, at first sight, seems to rule out the
possibility of initiation via the oxidative coupling route
(IV, Scheme 3) since oligomers generated via this route
should contain only even-numbered isotopomers.26 Iso-
topic scrambling on the other hand is readily accounted
for in a Cossee-Arlman-type mechanism. However, the
isotopic scrambling in the free ethylenes complicates
matters somewhat since oligomers generated from iso-
topically scrambled ethylene would lead to all possible
oligomer isotopomers. If isotopic scrambling in the
monomer was arising in a predominantly metallacyclic
mechanism, it would mean that a second process must
be occurring that serves to scramble H and D between
molecules of ethylene. In oligomerization systems with
low R-values, the rate of chain transfer is high relative
to the rate of propagation, meaning that the likelihood
of â-elimination from a Co-Et species is high. We
therefore carried out a similar experiment using the
polymerization system 13-MeBArf3 as the catalyst. In
this case no scrambling was observed in the free
repulsive electrostatic interactions associated with bring-
ing together positively charged cobalt centers.
We also examined the reaction of 59 µmol of 2-d3 with
1 equiv of [HNMe2Ph][B(C6F5)4], i.e., a Lewis acid-free
system. Exposure of this mixture to an atmosphere of
ethylene gave 42 mg of polymer, corresponding to an
activity of 0.5 g/mmol bar h.24 A similar activity was
observed by Erker and co-workers upon treatment of 3
with Li[B{3,5-(CF3)2C6H3}4].25
To probe pathway IV, we turned to the oligomeriza-
tion precatalyst 12, since this would generate liquid
products more readily analyzed for end groups and
(24) The lower activity is not due to the presence of dimethylaniline
competing with ethylene for the metal center; the 1H NMR spectra of
13-MeB(C6F5)3 and 13-B(C6F5)4 are identical, whereas differences
would be apparent due to the sensitivity of the ligand chemical shifts
to the environment at the cobalt center. In addition, it is possible to
make the ethylene adduct 14-B(C6F5)4 by the addition of 1 equiv of
ethylene to 13-B(C6F5)4; no equilibrium is observed. It thus appears
that the tertiary amine is too bulky to interact with the sterically
protected cobalt center.
(25) Steffen, W.; Blo¨mker, T.; Kleigrewe, N.; Kehr, G.; Fro¨hlich, R.;
Erker, G. Chem. Commun. 2004, 1188.
(26) Agapie, T.; Schofer, S. J.; Labinger, J. A.; Bercaw, J. E. J. Am.
Chem. Soc. 2004, 126, 1304

2046 Organometallics, Vol. 24, No. 9, 2005
Humphries et al.
Scheme 4. Hypothesized Route for the Formation
ethylene, consistent with the much lower rate of â-H
elimination from a Co-Et species for a polymerization
catalyst compared to an oligomerization catalyst. This
observation favors a Cossee-Arlman-type propagation
mechanism over the oxidative olefin coupling route since
a secondary process that causes isotopic scrambling
between the free ethylenes independent of the polym-
erization mechanism would be expected to operate in
both oligomerizations and polymerizations catalyzed by
these cobalt systems.
of 15 (Ar ) 2,6-iPr2C6H3)
Having ruled out pathways I-IV, pathways V and VI
are left as possibilities. V involves attack of an ab-
stracted methide group on the cobalt center with
concomitant electrophilic attack by the Lewis acid on
the bound ethylene to afford a zwitterionic cobalt(III)
dialkyl. The final, and substantially different, possibility
is a vinylic C-H activation to afford a cobalt(III) vinyl
hydride species (VI, Scheme 3).
To probe pathway V, we turned to deuterium labeling
studies involving (separately) the precatalyst and acti-
vator.
A polymerization using [2,6-{MeCdN(2,6-
iPr2C6H3)}2C5H3N]CoCD3 (2-d3) activated by B(C6F5)3
gave an activity of 22 g/mmol bar h over a 1 h run. The
polymer produced was analyzed by both 1H and 2H NMR
color change from dark blue to brown-green is observed,
and the NMR spectrum of the product was consistent
with the formation of [{2,6-{MeCdN(2,6-iPr2C6H3)}2-
C5H3N}Co(CCtBu)(C(H)dC(H)tBu)][MeB(C6F5)3] (15)
(Scheme 4). The 1H NMR spectrum contains two doublet
resonances each integrating for one proton and with a
3JH-H coupling constant of 5.6 Hz; this is at the low end
of the range for a cis arrangement about a double bond.
One of these doublets has a chemical shift of δ 3.12; the
other is shifted greatly to the high-field end of the
spectrum, at δ -0.83. The latter is reminiscent of the
chemical shift of the protons in the â-position of the
alkyl chains in complexes 6-8 and is similarly pre-
sumed to be due to the ring current of the aryl substit-
uents on the imine groups, particularly as the vinyl
group is too rigidly constrained to adopt an agostic
conformation. We were unable to obtain further char-
acterizing data for 15, as its high solubility and the
presence of byproducts have precluded its purification.
The formation of 15 is presumed to proceed via C-H
activation of the alkyne to afford a Co(III) alkynyl
hydride, followed by the insertion of a second equivalent
of alkyne into the Co-H bond (Scheme 4). It is thus
clear that it is possible for Co(III) species to be derived
from the Co(I) precursor 13 via the oxidative addition
of a C-H bond to the cobalt(I) center. A similar reaction
was observed for [Co(PMe3)3][BPh4], where the reaction
product, [Co(PMe3)3(CCPh)(C(H)dC(H)Ph)][BPh4], was
fully characterized.31
1
spectroscopy. The H NMR spectrum showed the poly-
mer to be linear with methyl and vinyl chain ends in
an approximately 1:1 ratio, as observed for all polymer-
izations using bis(imino)pyridine cobalt systems. More-
over, the 2H NMR spectrum showed the methyl end
groups of the polymer to be deuterium enriched. From
the Mn and yield of the polymer produced, it can be
calculated that 5% of the polymer chains contain a CD3
end group.27,28 If these arise from the initiation process,
it would mean that 6% of the cobalt centers were active,
producing an average of 20 chains per active center. A
similar experiment undertaken using 2-d3 activated by
[HNMe2Ph][B(C6F5)4] gave an activity of 0.5 g/mmol bar
h. In this case neither CD3 nor C6F5 end groups were
2
observable in H or 19F NMR spectra of the polymer.
In another series of polymerizations, 1 was activated
with perdeuterated MAO (synthesized by the controlled
hydrolysis of Al2Me6-d18), followed by termination of the
polymerization by the addition of MeOH/HCl. These also
yielded polymers that contained CD3 end groups, con-
firming that the abstracted methide is reincorporated
into the polymer in both the MAO- and B(C6F5)3-
activated systems, consistent with pathway V and
providing additional evidence that pathways III and IV
do not operate.29
Since it is not possible to observe the organocobalt
product of vinylic C-H activation (VI, Scheme 3) nor
to readily distinguish the polymer products of such a
pathway (except in certain specific instances),30 we have
probed the viability of obtaining Co(III) from Co(I) via
a C-H activation process through the reaction of
13-MeBArf3 with 3,3-dimethylbutyne. An immediate
Discussion
The observation of CD3 end groups in all of the PE
samples prepared using either CD3-labeled cobalt pre-
(30) Insertion into the Co-H bond that results from a vinylic C-H
activation (the less sterically demanding bond into which insertion
could occur) would result in a polymer or oligomer with one methyl
and one vinyl bond (as observed) whether the polymer was terminated
by â-hydrogen elimination or reductive elimination. Insertion into the
Co-vinyl bond would afford chains with one methyl and one vinyl
chain end only if termination occurred via reductive elimination. If
termination occurred by â-hydrogen elimination, dienes would result,
which would be detectable among the oligomeric products.
(27) No C6F5 groups were detectable in the polymer.
(28) Errors in all quantitative measurements obtained by NMR
spectroscopy are estimated to be (20%, arising mainly from errors in
the integration of NMR signals.
(29) It should be noted that it is possible that some of these CD3
chain end groups arise through small amounts of chain transfer to
aluminum. While chain transfer to aluminum does not occur on a large
scale in the cobalt system, as shown by the 1:1 ratio of methyl:vinyl
chain ends in the 1H NMR spectrum, the presence of very low amounts
of chain transfer cannot be ruled out completely.
(31) Habadie, N.; Dartiguenave, M.; Dartiguenave, Y.; Britten, J.
F.; Beauchamp, A. Organometallics 1989, 8, 2564.

Bis(imino)pyridine Cobalt Catalysts
Organometallics, Vol. 24, No. 9, 2005 2047
Table 3. Summary of Activities and
Summary
Polymerization Data Using Different Activatorsa
Cobalt bis(imino)pyridine/MAO ethylene polymeriza-
tion catalysts have been shown to operate by initial
reduction of a cobalt(II) precatalyst to a cobalt(I) halide,
followed by conversion to a cobalt(I) methyl, and ulti-
mately to a cobalt(I) cationic species. Addition of eth-
ylene to this system affords an ethylene adduct, which
is the immediate precursor to the active species. Deu-
terium labeling studies show that the methyl groups of
“noncoordinating” [MeB(C6F5)3] or [Me-MAO] anions are
incorporated at the saturated ends of the polyethylene
chains, consistent with an activation mechanism that
involves nucleophilic attack by an abstracted alkyl
group on the cationic ethylene species.
activity
polymer (g/mol
catalytic system
yield (g) bar h)
Mn
Mw/Mn
1a/100 MAO
6.12
6.14
0.13
0.13
1.12
350
350
11
11
96
6800
8300
14000
9600
8100
d
2.8
2.3
2.4
2.3
2.5
d
3a/100 MAO
2b/1 B(C6F5)3
2b/5 B(C6F5)3
2b/1 B(C6F5)3 + 5 Al2Me6
2-d3c/1 [HNMe2Ph][B(C6F5)4] 0.042
0.5
a 8.8 µmol of precatalyst, 100 mL of toluene, 2 bar of ethylene,
1 h, 298 K. b18 µmol of precatalyst, 15 mL of toluene, 2 bar of
c
ethylene, 20 min, 298 K. 59 µmol of precatalyst, 10 mL of toluene,
1.5 bar of ethylene, 1 h, 298 K. dNot analyzed.
catalysts or CD3-labeled activators points to the mech-
anism proceeding via either (or possibly both) path-
way II or V. The C-H activation of alkyne to afford a
cobalt(III) species suggests that the vinylic C-H activa-
tion reaction pathway VI is a possibility, but it does not
prove that the related reaction occurs for ethylene. A
key difference between the postulated pathways lies in
the involvement of the cocatalyst. The productivities of
the catalyst/cocatalyst combinations lie in the order
13/MAO . 13/MeB(C6F5)3 . 13/B(C6F5)4, with order
of magnitude differences between each system. The
oxidative addition of a vinylic C-H bond to the cobalt
center would be expected to be largely independent of
the nature of the cocatalyst, while the nucleophilic
attack of a methide group is clearly dependent on the
nature of the cocatalyst. When these observations are
combined with the appearance of CD3 end groups in the
resultant polymers, a pathway involving nucleophilic
attack of a methide anion on 14, i.e., pathways II and
V, seems the most plausible of the various possibilities.
However, II can be ruled out since the cobalt(I) alkyls
undergo alkyl group abstraction in the presence of
strong Lewis acids such as MAO or B(C6F5)3 and are
inactive for ethylene polymerization in the presence of
Al2Me6. This leaves pathway V, in which the Lewis acid
is attached to a carbon terminus, and therefore can be
more readily accommodated sterically. It also involves
a cobalt(III) center as opposed to cobalt(I), in common
with other known cobalt ethylene polymerization sys-
tems.32 Chain growth could occur via either of the
cobalt-carbon bonds, polymerization then being termi-
nated by â-H transfer.
On a final point, it should be noted that it is not
possible to explain the activity (albeit very low) observed
for 2/[HNMe2Ph][B(C6F5)4] on the basis of the above
observations, where there are no nucleophilic alkyl
groups present, and the polymer does not contain C6F5
end groups. That the activity in this system is far lower
than in the systems where nucleophilic alkyl groups are
present (see Table 3 for a summary of activities using
different activators) may indicate that a different mech-
anism is operating, possibly involving vinylic C-H
activation or the formation of metallacycles. However,
the difficulties of assessing the intricacies of what may
be a side reaction, with little net contribution to the
overall catalyst productivity, preclude conclusions on the
possible involvement of this process at this time.
Experimental Details
All reactions were carried out under an atmosphere of
dinitrogen gas. Standard Schlenk-line and associated tech-
niques were employed for the manipulation of air-sensitive
compounds. Conventional inert atmosphere dryboxes were
used for the preparation of the analytical and spectroscopic
samples, as well as for weighing and storage of air-sensitive
compounds. With the exception of pentane, heptane, and
toluene, all solvents were dried by refluxing over an appropri-
ate drying agent,33 distilled, and degassed prior to use.
Pentane, heptane, and toluene were dried and deoxygenated
according to a published procedure.34 1H, 13C, and 19F NMR
spectra were recorded at ambient temperature on a Bruker
AC-250 at 250.133, 62.896, and 235.318 MHz, respectively,
unless otherwise stated. Spectra at higher fields were recorded
on a Bruker DRX-400 (1H and 13C resonances at 400.129 and
100.613 MHz, respectively). Low-temperature 1H NMR and
2H NMR were recorded on a Bruker AM-500 at 500.133 and
76.774 MHz, respectively. Chemical shifts were referenced
internally to the residual protio impurity of the deuterated
solvents (1H), the solvent resonance (13C), or d6-benzene at 7.15
ppm (2H) and are quoted in ppm relative to tetramethylsilane.
Coupling constants are quoted in Hz. Chemical shifts for 19F
were referenced externally to CFCl3 at 0 ppm. Mass spectra
were recorded on either a VG Autospec or a VG Platform II
spectrometer. Microanalyses (C, H, N) were performed at
London Metropolitan University. GPC and NMR analyses of
polymers were carried out by BP Chemicals. The compounds
1,3 6,10 7,10 [2,6-{(CD3)CdN(2,6-iPr2C6H3)}2C5H3N],16 and tris-
(pentafluorophenyl)borane35 were prepared as described in the
literature. The compound Al2Me6-d18 was prepared from CD3I
and Al by a published procedure.36
Preparation of [2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N]-
CoMe, 2. A suspension of 1 (612 mg, 1.00 mmol) in diethyl
ether (30 mL) at -78 °C was treated dropwise with methyl-
magnesium bromide (0.84 mL of a 3.0 M solution in diethyl
ether, 2.52 mmol) and allowed to slowly warm to ambient
temperature, during which time a color change to red-purple
was observed. After stirring for 24 h at room temperature, the
volatiles were removed under reduced pressure and the red-
purple residue was extracted with toluene (30 cm3). Filtration
and removal of the volatile components under reduced pressure
afforded the crude product as a claret-colored powder. Yield:
465 mg, 81% based on 1. Crystals suitable for an X-ray
(33) Armarego, W. L. F.; Perrin, D. D. Purification of Laboratory
Chemicals, 4th ed.; Butterworth-Heinemann: Oxford, UK, 1998.
(34) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518.
(35) (a) Massey A. G.; Park, A. J. J. Organomet. Chem. 1964, 2, 245.
(b) Chernega, A. N.; Graham, A. J.; Green, M. L. H.; Haggitt, J.; Lloyd,
J.; Mehnert, C. P.; Metzler, N.; Souter, J. J. Chem. Soc., Dalton Trans.
1997, 2293.
(36) Yang, P.-W.; Liou, K.-F.; Lin, Y.-T. J. Organomet. Chem. 1986,
307, 273.
(32) (a) Schmidt, G. F.; Brookhart, M. J. Am. Chem. Soc. 1985, 107,
1443. (b) Daugulis, O.; Brookhart, M.; White, P. S. Organometallics
2003, 22, 4699.

2048 Organometallics, Vol. 24, No. 9, 2005
Humphries et al.
structure determination were grown via slow solvent evapora-
tion from a concentrated solution of 2 in diethyl ether.
1H NMR (C6D6): 10.19 (1H, t; p-H, py), 7.86 (2H, d; m-H,
py), 7.49 (2H, distorted triplet; p-H, Ar), 7.37 (4H, distorted
doublet; m-H, Ar), 3.13 (4H, septet; (CH3)CH(CH3)), 1.19 (12H,
d; (CH3)CH(CH3)), 0.62 (12H, d; (CH3)CH(CH3)), 0.58 (3H, s;
CoCH3), -1.14 ppm (6H, s; ArNdCCH3). 13C{1H} NMR (C6D6,
62.9 MHz): 166.1 (ArNdCCH3), 157.1 (o-C, py), 154.7 (i-C, Ar),
140.6 (o-C, Ar), 126.5 (p-C, Ar), 123.9 (m-C, Ar), 122.7 (m-C,
py), 117.4 (p-C, py), 28.4 ((CH3)CH(CH3)), 25.3 (ArNdCCH3),
24.1 (CH3)CH(CH3)), 23.2 (CH3)CH(CH3)), -15.5 ppm (CoCH3).
EI-MS (m/z): 555 [M]+, 28%; 540 [M - Me]+, 100%. Anal.
Calcd (%) for C34H46N3Co: C 73.49, H 8.34, N 7.56. Found: C
73.56, H 8.25, N 7.44.
Ar), 141.3 (o-C, Ar), 127.8 (m-C, CH2Ph; signal buried beneath
solvent resonance), 127.5 (o-C, CH2Ph), 126.9 (p-C, Ar), 124.8
(m-C, Ar), 124.0 (m-C, py), 119.9 (p-C, CH2Ph), 116.4 (p-C, py),
28.8 ((CH3)CH(CH3)), 25.9 (ArNdCCH3), 24.6 (CH3)CH(CH3)),
23.5 (CH3)CH(CH3)), -3.4 ppm (CoCH2Ph). EI-MS (m/z): 555
[M - C6H5]+, 12%; 482 [M - Co - CH2Ph]+, 26%; [M - Co -
CH2Ph - Me]+, 24%. Anal. Calcd (%) for C40H50N3Co: C 76.04,
H 7.98, N 6.65. Found: C 75.89, H 7.60, N 6.61.
Preparation of [2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N]-
CoH, 9. In a typical procedure, a solution of 2 (5.6 mg, 0.010
mmol) in benzene-d6 (0.7 cm3) was degassed by two freeze-
pump-thaw cycles. A stoichiometric quantity of room-tem-
perature hydrogen gas was then admitted into the vacuum
above the solution, frozen at 77 K. The solution was allowed
to thaw and mixed with the gas by repeated shaking. Yield >
98% by NMR.
[2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N]CoCD3 (2-d3) was pre-
pared in similar fashion using MgMe2-d6.
3
Preparation of [2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N]-
CoCl, 3. A suspension of 1 (917 mg, 1.50 mmol) and an excess
of zinc powder (3 g, 45 mmol) in toluene (75 cm3) were heated
to 50 °C for 15 h. During this time, the initial ochre slurry
became an intense red-purple-colored solution. The suspended
solids were allowed to settle, and the solution was then filtered.
The volatile components were removed under reduced pressure
to afford 3 as a red-purple solid, which was recrystallized from
toluene/pentane. Yield: 664 mg, 78% based on 1.
1H NMR (C6D6, 500 MHz): 10.83 (1H, t, JH-H ) 7.6 Hz;
p-H, py), 7.60 (2H, distorted triplet, 3JH-H ) 7.7 Hz; p-H, Ar),
3
7.59 (2H, d, JH-H ) 7.6 Hz; m-H, py), 7.46 (4H, distorted
3
3
doublet, JH-H ) 7.7 Hz; m-H, Ar), 3.42 (4H, septet, JH-H
)
3
6.8 Hz; (CH3)CH(CH3)), 1.31 (12H, d, JH-H ) 6.8 Hz;
(CH3)CH(CH3)), 0.28 (12H, d, 3JH-H ) 6.8 Hz; (CH3)CH(CH3)),
-1.66 ppm (6H, s; ArNdCCH3). No resonance for Co-H is
detectable over the range +100 to -150 ppm [I(59Co) ) 7/2].
13C{1H} NMR (C6D6, 100 MHz): 168.6 (ArNdCCH3), 160.1
(i-C, Ar), 156.5 (o-C, py), 140.2 (o-C, Ar), 126.3 (p-C, Ar),
124.2 (m-C, Ar), 123.7 (m-C, py), 118.0 (p-C, py), 28.9
((CH3)CH(CH3)), 25.6 (ArNdCCH3), 23.6 ((CH3)CH(CH3)),
22.7 ppm ((CH3)CH(CH3)). Infrared (Nujol): ν(Co-H) 2092
cm-1. The sensitivity of this complex precluded its isolation.
Preparation of [2,6-{MeCdN(2,4,6-Me3C6H2)}2C5H3N]-
CoCl, 10. The synthesis was as for 3, starting from [2,6-
{MeCdN(2,4,6-Me3C6H2)}2C5H3N]CoCl2. Yield: 56%.
1H NMR (C6D6): 9.54 (1H, t; p-H, py), 7.41 (2H, distorted
triplet; p-H, Ar), 7.27 (4H, distorted doublet; m-H, Ar), 6.91
(2H, d; m-H, py), 3.33 (4H, septet; (CH3)CH(CH3)), 1.18 (12H,
d; (CH3)CH(CH3)), 1.06 (12H, d; (CH3)CH(CH3)), 0.05 ppm
(6H, s; ArNdCCH3). 13C{1H} NMR (C6D6, 100 MHz): 167.3
(ArNdCCH3), 152.7 (o-C, py), 150.8 (i-C, Ar), 140.7 (o-C, Ar),
127.0 (p-C, Ar), 125.5 (m-C, py), 123.7 (m-C, Ar), 114.9 (p-C,
py), 29.3 (CH3)CH(CH3), 24.0 (CH3)CH(CH3), 23.8 (CH3)CH-
(CH3), 21.2 ppm (ArNdCCH3). EI-MS (m/z): 575 [M]+, 100%.
Anal. Calcd (%) for C33H43N3CoCl: C 68.80, H 7.52, N 7.29.
Found: C 68.82, H 7.57, N 7.22.
1H NMR (C6D6, 500 MHz): 9.49 (1H, t, 3JH-H ) 7.7 Hz; p-H,
3
py), 6.93 (4H, s; m-H, Ar), 6.90 (2H, d, JH-H ) 7.7 Hz; m-H,
py), 2.19 (12H, s; o-CH3, Ar), 2.18 (6H, s; p-CH3, Ar), -0.06
ppm (6H, s; ArNdCCH3). 13C{1H} NMR (C6D6, 100 MHz):
167.0 (ArNdCCH3), 152.9 (o-C, py), 151.8 (i-C, Ar), 135.0 (p-
C, Ar), 129.8 (o-C, Ar), 129.4 (m-C, Ar), 125.1 (m-C, py), 114.8
(p-C, py), 21.3 (p-CH3, Ar), 20.2 (ArNdCCH3), 19.4 ppm
(o-CH3, Ar). EI-MS (m/z): 491 [M]+, 72%. Anal. Calcd (%) for
C27H31N3CoCl: C 65.92, H 6.35, N 8.54. Found: C 66.06, H
6.19, N 8.43.
[2,6-{(CD3)CdN(2,6-iPr2C6H3)}2C5H3N]CoCl (3-d6) was pre-
pared similarly.
Preparation of [2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N]-
CoBr, 4. A solution 3 (14.8 mg, 25.7µmol) in benzene-d6 (0.7
cm3) was treated in situ with trimethylsilylbromide (3.4 mm3,
1
25.8 µmol) at room temperature. Following inspection by H
NMR spectroscopy, a further 4.0 equiv of trimethylsilylbromide
(13.6 mm3, 103 µmol) was added to the solution in order to
drive the reaction to completion.
Preparation of [2,6-{MeCdN(2,4,6-Me3C6H2)}2C5H3N]-
CoMe, 11. The synthesis was as for 2, using [2,6-{MeCdN-
(2,4,6-Me3C6H2)}2C5H3N]CoCl2 (580 mg, 1.10 mmol) and me-
thylmagnesium chloride (1.00 cm3 of a 3.0 M solution in THF,
3.0 mmol) in diethyl ether (35 cm3). Yield: 397 mg, 77%.
1H NMR (C7D8, 400 MHz): 10.08 (1H, t; p-H, py), 7.88 (2H,
d; m-H, py), 6.99 (4H, s; m-H, Ar), 2.28 (6H, s; p-CH3, Ar),
1.93 (12H, s; o-CH3, Ar), 0.46 (3H, s; CoCH3), -1.21 ppm
(6H, s; ArNdCCH3). 13C{1H} NMR (C7D8, 100 MHz): 165.2
(ArNdCCH3), 157.1 (o-C, py), 155.2 (i-C, Ar), 134.2 (p-C, Ar),
129.5 (o-C, Ar), 129.3 (m-C, Ar), 121.9 (m-C, py), 117.4 (p-C,
py), 24.1 (ArNdCCH3), 21.2 (p-CH3, Ar), 19.0 (o-CH3, Ar),
-17.3 ppm (br, CoCH3). EI-MS (m/z): 471 M+, 34%; 456 [M -
Me]+, 100%. Anal. Calcd (%) for C28H34N3Co: C 71.32, H 7.27,
N 8.91. Found: C 71.11, H 7.08, N 8.81.
1H NMR (C6D6): 9.79 (1H, t; p-H, py), 7.45 (2H, distorted
triplet; p-H, Ar), 7.30 (4H, distorted doublet; m-H, Ar), 6.88
(2H, d; m-H, py), 3.37 (4H, septet; (CH3)CH(CH3)), 1.18 (12H,
d; (CH3)CH(CH3)), 1.10 (12H, d; (CH3)CH(CH3)), -0.08 ppm
(6H, s; ArNdCCH3).
Preparation of [2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N]CoI,
5. The procedure was as for 4 above. Approximately 5.0 equiv
of trimethylsilylbromide was required to drive the equilibrium
over to 5 + Me3SiCl.
1H NMR (C6D6): 10.15 (1H, t; p-H, py), 7.44 (2H, distorted
triplet; p-H, Ar), 7.30 (4H, distorted doublet; m-H, Ar), 6.81
(2H, d; m-H, py), 3.37 (4H, septet; CH(CH3)2), 1.13 (24H, d;
CH(CH3)2), -0.37 ppm (6H, s; ArNdCCH3).
Preparation of [2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N]-
CoCH2Ph, 8. The synthesis was as for 2, using 1 (520 mg,
0.850 mmol) and benzylmagnesium bromide (2.8 cm3 of a
1.0 M solution in Et2O, 2.8 mmol). Yield: 310 mg, 54% based
on 1.
Preparation of [2,6-{MeCdN(2-EtC6H4)}2C5H3N]CoCl,
12. To a mixture of [2,6-{MeCdN(2-EtC6H4)}2C5H3N]CoCl2
(153 mg, 0.250 mmol) and potassium graphite (35.3 mg, 0.263
mmol) in a Schlenk tube at -78 °C was added precooled
toluene (15 cm3). Within seconds, the color of the mixture
changed to deep purple, and the solution was then allowed to
warm to room temperature slowly. After stirring for 16 h the
supernatant was separated by filtration. Removal of all
volatiles under reduced pressure afforded the title compound
as a claret solid. Yield: 102 mg, 71% based on 1.
1H NMR (C6D6): 10.26 (1H, t; p-H, py), 7.68 (2H, t; m-H,
py), 7.61 (2H, distorted triplet; p-H, Ar), 7.44 (4H, distorted
doublet; m-H, Ar), 6.79 (1H, t; p-H, CH2Ph), 6.56 (2H, d; m-H,
CH2Ph), 5.59 (2H, d; o-H, CH2Ph), 3.22 (4H, septet; (CH3)-
CH(CH3)), 2.50 (2H, s; CoCH2Ph), 1.14 (12H, d;
(CH3)CH(CH3)), 0.63 (12H, d; (CH3)CH(CH3)), -1.17 ppm
(6H, s; ArNdCCH3). 13C{1H} NMR (C6D6, 100 MHz): 165.9
(ArNdCCH3), 157.4 (i-C, CH2Ph), 155.5 (o-C, py), 154.8 (i-C,
1H NMR (C7D8), 388 K: 9.72 (1H, br; p-H, py), 7.67 (2H, m;
Ar), 7.3-6.8 (8H; Ar and m-H, py), 2.55 (4H, q; o-CH2CH3),
0.92 (6H, t; o-CH2CH3), -0.01 ppm (6H, br s; ArNdCCH3). The

Bis(imino)pyridine Cobalt Catalysts
Organometallics, Vol. 24, No. 9, 2005 2049
low solubility of 11 in hydrocarbons, combined with its
instability in other solvents, precluded acquisition of a resolved
low-temperature 1H NMR spectrum. EI-MS (m/z): 463 M+,
100%; 369, 46%; 354, 88%; 77, 60%. Anal. Calcd (%) for
C25H27N3CoCl: C 64.73, H 5.87, N 9.06. Found: C 64.60, H
5.71, N 8.93.
Preparation of [{{2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N}-
Co(η1-N2)}{MeB(C6F5)3}], 13-MeB(C6F5)3. A solution of
B(C6F5)3 (184 mg, 0.36 mmol) in pentane (10 cm3) was allowed
to diffuse slowly into a claret solution of 2 (200 mg, 0.36 mmol)
in benzene (5 cm3) at room temperature. After 4 days the
resulting blue solution was decanted, and the crystals were
washed with pentane (2 × 5 cm3). Yield: 310 mg, 81% based
on 2.
afford a viscous solution. The solution was then spiked with a
small amount of benzene-d6 as a reference.
2H NMR Study of MAO-Activated Systems. Either 1-d6
or 3-d6 (0.035 mmol) was dissolved in a solution of MAO (0.6
cm3, Al:Co ) 14:1) in 2-chlorotoluene to give a purple solution.
A 2H NMR spectrum of each activated complex was recorded.
The sample was then freeze-pump-thaw degassed twice and
ethylene condensed into the tube. The sample was allowed to
warm to room temperature and shaken, and the spectrum
recorded.
2H NMR Study of B(C6F5)3-Activated System. A sample
of 13-MeB(C6F5)3 in 2-chlorotoluene was prepared via the
reaction of 2-d6 (19.4 mg, 0.035 mmol) and B(C6F5)3 (18.0 mg,
0.035 mmol) as described above.
1H NMR (C7D8): 7.55 (1H, t; p-H, py), 7.00 (2H, t, p-H, Ar),
6.87 (4H, d; m-H, Ar), 6.69 (2H, pseudo-doublet; m-H, py), 2.85
(4H, septet; (CH3)CH(CH3)), 1.11 (6H, s; ArNdCCH3), 1.05
(12H, d; (CH3)CH(CH3)), 0.98 (12H, d; (CH3)CH(CH3)), -2.00
ppm (∼3H, br s; H3C-B(C6F5)3). 13C{1H} NMR (C7D8, 125
MHz): 173.8 (ArNdCCH3), 155.9 (o-C, py), 143.2 (i-C, Ar),
139.8 (o-C, Ar), 136.6 (p-C, py), 124.6 (m-C, Ar), 123.4 (m-C,
py), 29.1 ((CH3)CH(CH3)), 23.6 ((CH3)CH(CH3)), 23.3
((CH3)CH(CH3)), 17.1 (ArNdCCH3), 13.0 ppm (H3C-B(C6F5)3).
The 13C resonance signal for the para-C, Arsalthough ob-
scured by solvent peakssmay be located through the HSQC
spectrum and has a chemical shift in the region 128.0-
128.5 ppm. The 13C signals of the pentafluorophenyl groups
are too broad to be observed. 19F NMR (C7D8): -130.8
(2F, m; o-F), -163.2 (1F, m; p-F), -164.1 ppm (2F, m; m-F).
Infrared (Nujol): ν(NN) 2197 cm-1. Anal. Calcd (%) for
C52H46BCoF15N5: C 57.00, H 4.23, N 6.39. Found: C 56.90, H
4.17, N 6.29.
Ethylene Polymerizations. Polymerizations Using MAO
as Cocatalyst. To a magnetically stirred solution of a pre-
catalyst (0.0088 mmol) in toluene (100 cm3) was added
MAO (0.55 cm3 of a 1.6 M solution in toluene, Al:Co )
100:1) at room temperature. The solution first turned claret,
then blue within a few minutes. The solution was degassed,
and ethylene admitted (2.0 bar), the pressure being main-
tained for the full duration of the polymerization run. Poly-
merizations were terminated (typically after 60 min) by
addition of MeOH/1 M HCl (about 9:1). The precipitated
polyethylene was filtered, washed with methanol, and dried
in vacuo at 60 °C overnight.
Polymerizations Using B(C6F5)3 as Cocatalyst. Inside
a glovebox, freshly prepared solutions of 2 or 2-d3 (10.0 mg,
0.018 mmol) in toluene (10 cm3) and B(C6F5)3 (9.2 mg, 0.018
mmol) in toluene (15 cm3) were mixed at room temperature,
leading to a color change from claret to blue. Trimethylalu-
minum (13 mg, 0.18 mmol, 10 equiv) was added if required.
This solution was transferred into a Schlenk tube, diluted with
toluene (to 50 cm3 total volume), sealed, and removed from
the glovebox. The solution was degassed and subsequently
exposed to ethylene (2.0 bar). Runs were typically terminated
after 20 min because no further ethylene uptake was observed.
The polymer was isolated as described for the runs using MAO
as cocatalyst.
Preparation in situ. Solutions of 2 (5.0 mg, 9 µmol) in
toluene-d8 (0.35 cm3) and B(C6F5)3 (4.6 mg, 9 µmol) in toluene-
d8 (0.35 cm3) were mixed to give a blue solution. This was then
transferred to an NMR tube fitted with a Young’s tap. NMR
spectral data were the same as those given above.
Preparation of [{{2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N}-
Co(η-C2H4)}{MeB(C6F5)3}], 14-MeB(C6F5)3. A solution of
13-MeB(C6F5)3 prepared as described above was degassed by
two freeze-pump-thaw cycles, after which ethylene (V ) 5.0
cm3, p ) 41 mbar, 1.1 equiv) was condensed into the NMR
tube. Warming of the mixture to room temperature was
accompanied by a color change to aquamarine. Yield: quan-
titative by NMR spectroscopy.
Polymerization Using [PhNMe2H][B(C6F5)4] as Co-
catalyst. The general procedure was as described for the
B(C6F5)3-activated systems. A claret solution of 2-d3 (5.2 mg,
0.0094 mmol) in toluene (10 cm3) was added to [PhNMe2H]-
[B(C6F5)4] (7.5 mg, 0.0094 mmol) at room temperature. The
mixture was agitated for several minutes, during which the
anilinium salt slowly reacted to afford a deep blue solution.
This solution was transferred into a Schlenk tube, diluted with
toluene (to 50 cm3 total volume), sealed, degassed, and exposed
to ethylene (1.5 bar). The run was terminated after 60 min.
Polymer yield: 7 mg (activity ) 0.5 g mmol-1 h-1 bar-1).
Repeating this run on a larger scale (0.059 mmol) afforded
3
1H NMR (C7D8): 8.00 (1H, t, JH-H ) 7.8 Hz; p-H, py), ca.
7.0 (obscured by solvent resonance) (2H, t; p-H, Ar), ca. 7.0
3
(2H, d; m-H, py), 6.77 (4H, d, JH-H ) 7.5 Hz; m-H, Ar), 4.65
3
(4H, s; C2H4), 3.02 (4H, septet, JH-H ) 6.5; (CH3)CH(CH3)),
3
1.09 (12H, d, JH-H ) 6.5 Hz; (CH3)CH(CH3)), 0.76 (12H, d,
3JH-H ) 6.5 Hz; (CH3)CH(CH3)), 0.75 (6H, s; ArNdCCH3) 0.36
ppm (3H, br; B-CH3). 19F NMR (C7D8): -130.8 (2F, m; o-F),
-163.2 (1F, m; p-F), -163.8 ppm (2F, m; m-F). Low solubility
in noncoordinating solvents precluded the collection of 13C
NMR data.
sufficient polymer (42 mg, activity ) 0.5 g mmol-1 h-1 bar-1
for characterization by NMR.
)
Preparation of Deuterated MAO. Caution: care must
be taken to avoid the release of large volumes of methane and
potential explosion. A precooled (-20 °C) suspension of LiOH‚
H2O (147 mg, 3.50 mmol) in toluene (4.0 cm3) was transferred
in several portions onto a cold solution (-20 °C) of either
Al2Me6-d18 (505 mg, 3.50 mmol) in toluene (3.0 cm3). The
mixture was slowly allowed to warm to room temperature and
then left to stir overnight. The supernatant was separated by
filtration. [Al] ) 1.0 M.
X-ray Crystallography. Crystal data for 2: C34H46CoN3,
M ) 555.67, monoclinic, P21/n (no. 14), a ) 13.882(2) Å, b )
14.823(2) Å, c ) 15.492(6) Å, â ) 90.92(2)°, V ) 3187.4(14) Å3,
Z ) 4, Dc ) 1.158 g cm-3, µ(Mo KR) ) 0.563 mm-1, T ) 203 K,
very dark green platy blocks; 5600 independent measured
reflections, F2 refinement, R1 ) 0.054, wR2 ) 0.103, 3750
independent observed reflections [|Fo| > 4σ(|Fo|), 2θmax ) 50°],
346 parameters. CCDC 256785.
Further equivalents of ethylene were added in similar
fashion. The only change to the NMR spectra was in the
position of the resonance attributable to the ethylene protons
which moved to an averaged position of 5.24 ppm.
Preparation of [{{2,6-{MeCdN(2,6-iPr2C6H3)}2C5H3N}-
Co(CCtBu)(C(H)dC(H)tBu)}{MeB(C6F5)3}], 15. To a solu-
tion of 13-MeB(C6F5)3 prepared as described above was added
3,3-dimethylbutyne (2.2 µL, 18 µmol). An instant color change
from blue to dark green-brown occurred. The solution was
transferred to a NMR tube and the spectrum recorded.
Deuterium NMR Spectroscopy. Preparation of MAO
Solutions in 2-Chlorotoluene. All volatiles were removed
from a sample of commercial MAO. The resulting white solid
(146 mg, about 2.0-2.5 mmol) and trimethylaluminum (146
mg, 2.0 mmol) were dissolved in 2-chlorotoluene (5.0 cm3) to

2050 Organometallics, Vol. 24, No. 9, 2005
Humphries et al.
Crystal data for 10: C27H31ClCoN3, M ) 491.93, monoclinic,
C2/c (no. 15), a ) 36.945(11) Å, b ) 8.062(3) Å, c ) 17.718(5)
polyethylene samples; Crompton Corp. is thanked for
generous donation of MAO.
Å, â ) 101.91(5)°, V ) 5164(3) Å3, Z ) 8, Dc ) 1.265 g cm-3
,
µ(Mo KR) ) 0.786 mm-1, T ) 203 K, claret platy needles; 4500
independent measured reflections, F2 refinement, R1 ) 0.044,
wR2 ) 0.094, 3089 independent observed absorption-corrected
reflections [|Fo| > 4σ(|Fo|), 2θmax ) 50°], 298 parameters. CCDC
256786.
Supporting Information Available: Molecular struc-
1
tures of 2, 3, 10, and 13; CIF files for 2 and 10; H NMR and
2H NMR spectra for key reactions and products, and GPC
traces are available free of charge via the Internet at
http://pubs.acs.org.
Acknowledgment. BP Chemicals Ltd is thanked for
financial support, and GPC and NMR analyses of
OM049032B