Mono- and bimetallic ethylene polymerization catalysts having an azanickellacyclopentene skeleton

Article abstract of DOI:10.1016/j.poly.2009.09.007

A series of azanickellacyclopentene complexes having iodo, bromo, chloro, or triflate ligand on the Ni center were prepared, and are subjected to studies on ethylene polymerization catalysis. Activity of these mononuclear azanickellacyclopentene complexes was increased in the order, Ni-Cl < Ni-I ≈ Ni-Br < Ni-OTf; this is explained by the performance of (pseudo)halogeno ligand as a leaving group from the nickel center. Methylalminoxane (MAO) and inexpensive AlEt₂Cl can be used as the cocatalyst. Mechanistic consideration suggested the involvement of neutral Ni-alkyl intermediates as proposed in the SHOP type catalytic system. Interestingly, the catalytic activity is significantly increased by incorporating the second metals into the diimino moiety of mononuclear complex. Two factors should be considered to explain this activity enhancement. One is the increased rigidity of the azanickellacyclopentene ligand backbone, and the other is the possibility of the presence of the two active centers in one molecule.

Full text of DOI:10.1016/j.poly.2009.09.007

Polyhedron 28 (2009) 3935–3944
Contents lists available at ScienceDirect
Polyhedron
journal homepage: www.elsevier.com/locate/poly
Mono- and bimetallic ethylene polymerization catalysts
having an azanickellacyclopentene skeleton
Daisuke Noda b, Masao Tanabiki c, Kazuhiro Tsuchiya b, Yusuke Sunada a,b, Hideo Nagashima a,b,
*
a Division of Applied Molecular Chemistry, Institute for Materials Chemistry and Engineering, Kasuga, Fukuoka 816-8580, Japan
b Graduate School of Engineering Sciences, Kyushu University, Kasuga, Fukuoka 816-8580, Japan
c Yokkaichi Research Laboratory TOSOH Corporation, 1-8, Kasumi, Mie 510-8540, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 2 June 2009
Accepted 7 September 2009
Available online 16 September 2009
A series of azanickellacyclopentene complexes having iodo, bromo, chloro, or triflate ligand on the Ni cen-
ter were prepared, and are subjected to studies on ethylene polymerization catalysis. Activity of these
mononuclear azanickellacyclopentene complexes was increased in the order, Ni–Cl < Ni–I ꢀ Ni–
Br < Ni–OTf; this is explained by the performance of (pseudo)halogeno ligand as a leaving group from
the nickel center. Methylalminoxane (MAO) and inexpensive AlEt2Cl can be used as the cocatalyst. Mech-
anistic consideration suggested the involvement of neutral Ni-alkyl intermediates as proposed in the
SHOP type catalytic system. Interestingly, the catalytic activity is significantly increased by incorporating
the second metals into the diimino moiety of mononuclear complex. Two factors should be considered to
explain this activity enhancement. One is the increased rigidity of the azanickellacyclopentene ligand
backbone, and the other is the possibility of the presence of the two active centers in one molecule.
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Ethylene polymerization
Azanickellacyclopentene
Bimetallic catalyst
1. Introduction
In 2004 we found NiX2(CNAr)2 to be a unique ethylene poly-
merization catalyst in the presence of MAO [5]. This catalyst does
It is well known that organonickel complexes exhibit excellent
catalytic activity for oligomerization and polymerization of al-
kenes, in which appropriate selection of auxiliary ligands of nickel
is important for increasing the catalytic activity and molecular
weight of the polymer formed [1]. One of the typical ethylene poly-
not belong to a category of conventional nickel-based ethylene
polymerization catalysts such as the Brookhart catalyst or the
SHOP catalyst. In the course of our mechanistic considerations,
we got the idea that oligomerization of isocyanide ligands on the
nickel atom might be involved to generate catalytically active spe-
cies. It is known that activation of isocyanides by a catalytic
amount of nickel complexes leads to polymerization of isocyanides
[6]. A complex shown in Chart 1, D (R = tBu), which is formed by
merization catalysts is Brookhart’s
a-diimine complex, which
shows an excellent catalytic activity with the aid of methylalumi-
noxane (MAO), providing relatively high molecular weight poly-
ethylene (Mn = 30 000–1 000 000) [2]. Detailed experiments
suggest that a key to forming high molecular weight polyethylene
is introduction of sterically bulky 2,6-diisopropylphenyl groups as
the N-aryl group as shown in Chart 1, A. In contrast, the SHOP cat-
alysts typically shown in Chart 1, B, give oligomers of ethylene,
t
trimerization of BuNC on the nickel atom, is considered to be an
intermediate for the isocyanide polymerization. Complex D is pre-
pared by treatment of Ni(tBuNC)4 with CH3I or reaction of trans-
Ni(PMe3)2MeCl with tBuNC [7]. The complex has an interesting
structural feature in comparison with the Brookhart catalyst and
SHOP catalyst. First is that it is analogous to the N-type SHOP cat-
alyst C, which has a nickellacyclopentane structure with a Ni–O
and a Ni N bond. Complex D has a nickellacyclopentane struc-
ture with a Ni–C and a Ni N bond. In other words, D is a novel
class of the N-type SHOP catalyst having a Ni–C bond. The second
which are actually used in industrial production of
a-olefins [3].
Studies stimulated by success of the Brookhart catalyst revealed
that introduction of the sterically demanding substituents on the
phosphorus atom leads to the formation of high molecular weight
polyethylene with high efficiency. Nitrogen analogues of the SHOP
catalyst bearing sterically bulky substituents on the N atoms were
also investigated, giving polyethylene with high molecular weights
[4].
unique structural feature of D is an a-diimine substructure, which
potentially coordinates with other transition metal fragments.
Namely, introduction of nickel or other transition metal species
gives a new type of Brookhart’s a-diimine complex (Scheme 1).
With these concepts in mind, we examined catalytic ethylene
polymerization with azanickellacyclopentene complexes. Although
the original complex shown in Chart 1, D (R = tBu) did not exhibit
catalytic activity, we found that its C6H3-2,6-Me2 analogues, 1–4,
* Corresponding author. Address: Division of Applied Molecular Chemistry,
Institute for Materials Chemistry and Engineering, Kasuga, Fukuoka 816-8580,
Japan. Tel./fax: +81 92 583 7819.
E-mail address: nagasima@cm.kyushu-u.ac.jp (H. Nagashima).
0277-5387/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.poly.2009.09.007

3936
D. Noda et al. / Polyhedron 28 (2009) 3935–3944
R
X'
R
N
R2
Ni
R1
R3
N
M
R
R
R
R
X'
R1
R2
R'
N
R
N
N
R
N
P
R
O
L
N
N
N
O
Ni
D
Ni
R
Ni
B
R
M
A
C
X
C
X
R'
X
X
L
R
N
N
R
R
E
C
Brookhart-type catalysts
SHOP-type catalysts
Azanickellacyclopentene and its bimetallic complex
Chart 1. Previously reported ethylene polymerization catalysts.
R
R
N
N
M
R
Replacement
from "O" to "C"
Incorporation of the
second metal "M"
N
N
C
N
N
O
N
C
R
R
R
Ni
Ni
Ni
R
Scheme 1. Strategy for construction of the new ethylene polymerization catalysts.
behaved as a good catalyst for production of polyethylene with
ton was observed at 2.20 ppm (1H NMR) and 17.3 ppm (13
C
Mn = 61 000–130 000 in the presence of MAO or Et2AlCl. Further-
more, introduction of the second metal to the a-diimine part of 1
significantly enhanced the catalytic activity. In this paper, we wish
to report the catalytic behavior of these mono- and bimetallic cat-
alysts for ethylene polymerization activated by organoaluminum
cocatalysts in detail, and discuss the polymerization mechanisms
[8].
NMR). The IR spectrum shows a strong absorption band at
2161 cmꢁ1, which is assigned to the C„N stretching vibration of
coordinated CNC6H3-2,6-Me2 ligand, along with a strong bands at
1640 cmꢁ1 derived from the C@N stretching mode of imino groups
of the azanickellacyclopentene skeleton. These spectroscopic data
are consistent with the molecular structure of 1 determined by
X-ray diffraction analysis (vide infra). Spectroscopic data of bromo
and chloro homologues of 1, i.e. 3 and 4, are almost identical with
those of 1. The triflate 2 also gives similar spectra to 1; however, it
is worthwhile to point out that 1H resonance of the methyl group
of the azanickellacyclopentene backbone shows significant upfield
2. Results and discussion
2.1. Synthesis and characterization of azanickellacyclopentene
derivatives 1–4
shift, and mC
„
absorption is shifted to a higher wavenumber
NAr
compared with those of 1(0.48 ppm, and 14 cmꢁ1, respectively).
The dissociation of OTf anion from 2 in the solution state was sug-
gested by low temperature 19F NMR spectra. In fact, the single 19F
resonance at ꢁ78.8 ppm is consistent with a typical chemical shift
due to the uncoordinated OTf anion in the literature, which gener-
ally appears around ꢁ79 ppm in CD2Cl2 shifted to higher field by
1–2 ppm compared with the coordinated OTf [9]. Since no dynamic
behavior was observed at the temperature range between room
temperature to ꢁ90 °C, 2 is likely to exist in the solvent stabilized
form shown in Chart 2.
The azanickellacyclopentene complex bearing a C6H3-2,6-Me2
group on the nitrogen atoms was synthesized by treatment of
Ni(CNC6H3-2,6-Me2)4, which was generated in situ by the reaction
of Ni(COD)2 with 4 equivalent of CNC6H3-2,6-Me2, with excess
amounts of MeI in boiling C6H6 for 24 h (95% yield) (Scheme 2).
The OTf, Br, and Cl homologues of 1 were synthesized as follows:
treatment of 1 with 1 equivalent of AgOTf in CH2Cl2 for 3–4 h affor-
ded azanickellacyclopentene triflate 2 in 89% yield as orange pow-
der. The triflate 2 was allowed to react with LiBr or LiCl in acetone
for 16 h, and gave rise to the formation of bromo- and chloro-
homologue 3 and 4 in 72% and 63% yield, respectively (Scheme 2).
Spectroscopic data of 1–4 are summarized in Table 1. NMR
spectrum of 1 shows four 1H resonances as a singlet at 1.95,
2.09, 2.29, and 2.32 ppm and four 13C resonances at 18.6, 18.7,
19.1, and 19.2 ppm due to methyl groups of four C6H3-2,6-Me2
moieties. The methyl signal of the azanickellacyclopentene skele-
The molecular structures of 1, 2, 3, and 4 were unequivocally
determined by X-ray diffraction analysis; molecular structures of
2 and 3 are shown in Fig. 1, and representative bond lengths and
angles are summarized in Table 1. Although OTf ligand in 2 is dis-
sociated in the solution state, it is coordinated to the nickel center
in the solid state. All of these four complexes have an azanickella-
cyclopentene framework, in which one nitrogen atom of the imino
N
N
N
N
N
N
N
4 xylylNC
C6H6
N
AgOTf
N
LiX
excess MeI
C6H6
Ni
Ni
Ni(COD)2
Ni
Ni(xylylNC)4
C
OTf
CH2Cl2
acetone
r.t. 16 h
C
X
C
I
N
N
r.t. 3-4 h
N
3: X = Br
4: X = Cl
1
2
Scheme 2. Synthesis of azanickellacyclopentene complexes 1–4.

D. Noda et al. / Polyhedron 28 (2009) 3935–3944
3937
Table 1
Spectroscopic data of complexes 1–4.
1
2
3
4
1H
CH3 of xylyl
CH3 of
1.95, 2.09, 2.29, 2.32
2.20
1.97, 2.03, 2.21, 2.26
1.72
1.93, 2.05, 2.23, 2.35
2.18
1.94, 2.06, 2.23, 2.36
2.18
nickellacycle
Ph
6.53–6.58 (m, 1H), 6.62–6.65 (m,
2H), 6.88–6.94 (m, 3H), 7.00–7.09
(m, 3H), 7.15–7.16 (m, 3H)
6.43–6.47 (m, 3H), 6.51–6.58 6.50–6.56 (m, 1H), 6.60–6.67 (m,
(m, 1H), 6.62–6.66 (m, 1H), 2H), 6.86–6.95 (m, 3H), 6.98–7.03 2H), 6.87–7.11 (m, 6H), 7.13–7.14
6.83–6.94 (m, 5H), 6.96–7.00 (m, 2H), 7.05–7.11 (m, 1H), 7.13– (m, 3H)
6.48–6.54 (m, 1H), 6.61–6.67 (m,
(m, 1H)
7.14 (m, 3H)
17.2, 18.5, 18.6, 18.9, 19.2
123.8, 124.1, 124.4, 126.1, 127.2,
127.8, 127.9, 128.0, 128.4, 129.1,
13C (CH3)
17.3, 18.6, 18.7, 19.1, 19.2
16.1, 18.2, 18.3, 18.7, 19.1
124.1, 124.6, 125.8, 127.6,
128.1, 128.2, 128.9, 129.5,
135.8, 142.2, 148.6, 148.6,
163.7, 177.8, 179.9
2175
17.1, 18.5, 18.6, 18.8, 19.2
123.7, 124.0, 124.4, 126.0, 127.1,
127.7, 127.9, 128.4, 128.6, 129.1,
129.4, 134.9, 144.1, 148.7, 150.1,
164.9, 179.4, 184.1
Ph(C„N)(C@N) 123.8, 124.1, 124.4, 126.1, 127.2,
127.7, 127.8, 128.0, 128.5, 128.9,
129.2, 134.7, 146.8, 148.6, 149.4,
164.5, 179.4, 186.4
129.3, 134.9, 148.7, 149.8, 164.7,
179.5
2164
1628
IR
(C„N)
(C@N)
2161
1627
2175
1633
1640
torsion angle of N(2)–C(3)–C(4)–N(3) is 58.37° (1), 51.28° (2),
58.67° (3) and 42.52° (4), respectively. The Ni–X (X = I, OTf, Br,
Cl) bond length is decreased in the order, 1 [2.5770(8) Å] > 3
[2.3803(4) Å] > 4 [2.2287(5) Å] > 2 [1.988(2) Å]. Other bond lengths
and angles of 1–4 are nearly identical, and the C–N bond length of
the coordinated CNC6H3-2,6-Me2 ligand (ca. 1.16 Å) is similar to
those of the known Ni(II)-isocyanide complexes [11]. The C6H3-
2,6-Me2 ring on the N(1) atom of 4 is nearly perpendicular to the
square plane around nickel, and the dihedral angle is 87.39°, while
that of the others is slightly more acute than 4 (82.56° for 1, 75.04°
for 2, and 80.44° for 3). In all complexes, the phenyl ring of the
CNC6H3-2,6-Me2 moiety on the N(3) atom is oriented to the coor-
dinated CNC6H3-2,6-Me2 ligand (Table 2).
Chart 2. Possible dissociation/coordination process.
moiety and a carbon atom are bonded to the square planar nickel
center. The complex 1, 2, and 3 exhibit a small distortion from
the square planar geometry of nickel: the C(5) atoms of 1, 2, and
3 lie slightly out of the least square plane around Ni, and the devi-
ation from the plane is 0.2258 Å, 0.2183 Å, and 0.1521 Å, respec-
tively. In contrast, the C(5) atom of 4 lies completely on the
square plane around Ni, and the sum of the angles around Ni is
360.0o. Conformation of the five-membered azanickellacyclopen-
tene ring can be described as an envelope form, in which the
C(4) atom is out of the plane defined by Ni, N(1), C(2), and C(3)
(deviation from the plane is 0.6614 Å for 1, 0.6176 Å for 2,
0.6429 Å for 3, 0.5374 Å for 4, respectively). In the molecular struc-
ture of Carmona’s azanickellacyclopentene chloride complex bear-
ing tBu-imino moieties, the five-membered azanickellacycle adopts
the similar envelope form, however, the deviation of the C(4) atom
from the plane is more significant than our complexes 1–4 (Car-
mona’s complex reportedly contains two independent molecules
in the unit cell, and the deviations from the plane are 1.0193 and
1.0299 Å) [7]. This may be due to the more crowded steric environ-
ment given by the tBu group. In contrast, in the previously reported
azapalladacycle complexes bearing iodide or chloride ligands on
palladium prepared by triple insertion of CNC6H3-2,6-Me2 [10],
the five-membered azapalladacycles also have an envelope
conformation with the deviation of the C(4) atom ca. 0.55 Å. The
As mentioned in the introduction, the nickellacyclopentane
complexes have a diimino moiety of the azanickellacyclopentene
backbone, which is capable of coordinating to the second metal
‘‘MBr2” fragment; this leads to the formation of (hetero)bimetallic
complexes (Scheme 3). As a typical example, treatment of 1 with
ZnBr2 in CH2Cl2/EtOH at room temperature for 24 h, followed by
recrystallization from CH2Cl2/hexane gave rise to Ni–Zn (het-
ero)bimetallic complex 5a in 77% yield, which was characterized
by NMR, IR, X-ray crystallographic study, and elemental analysis
(vide infra). Similarly, the Ni–Zn (5b) homologue was synthesized
in 67% yield, and characterized by spectroscopy. Studies on the
reactions of the nickellacyclopentane with CoBr2, FeBr2, or NiBr2
were hampered due to the paramagnetism of the products; how-
ever, consumption of almost all of the starting complex 1 or 3
checked by NMR spectroscopy strongly suggests the formation of
the corresponding (hetero)bimetallic complexes. In the reaction
of 1 with CoBr2, a formed Ni–Co (6a) complexes was isolated and
characterized by crystallography.
In the 1H NMR spectrum of the Ni–Zn (hetero)bimetallic com-
plexes, four singlets due to the C6H3-2,6-Me2 groups appeared at
2.24, 2.34, 2.47, and 2.69 ppm for 5a, and 2.23, 2.33, 2.46, and
2.68 ppm for 5b, respectively, whereas four signals were observed
Fig. 1. Molecular structures of 1, 2, 3, and 4 with 50% probability ellipsoids. Hydrogen atoms were omitted for clarity.

3938
D. Noda et al. / Polyhedron 28 (2009) 3935–3944
to the halogen ligand, which appeared at 2165 cmꢁ1 for 5a and
2183 cmꢁ1 for 5b. Although the NMR data of the CoBr2 homologue
6a gave only broad signals, the formation of the expected bimetal-
lic complex was confirmed by IR and elemental analysis. IR spec-
trum of 6a shows a strong absorption band at 2159 cmꢁ1, which
is assigned to the C„N stretching vibration of coordinated
CNC6H3-2,6-Me2 ligand. The result of the elemental analysis is con-
sistent with the formulation of the Ni–Co (hetero)bimetallic
complex.
The crystals of 5a and 6a suitable for X-ray diffraction analysis
were obtained by recrystallization from CH2Cl2/hexane. The ORTEP
drawing (Fig. 2) reveals that the coordination geometry around the
nickel is square planar, and the arrangement around the zinc or co-
balt center can be described as slightly distorted tetrahedral. Inter-
estingly, the orientation of the C6H3-2,6-Me2 group on the N(2)
atom is flipped to the opposite side compared with that of the par-
ent complex 1, which leads to the effective shielding of the methyl
group of the azanickellacyclopentene backbone as observed in the
1H NMR spectrum. In addition, coordination of ZnBr2 or CoBr2 frag-
ments also triggers the structural change over the nickellacycl-
opentene framework, in which planarity of the five-membered
ring is increased. Thus, the C(4) atom lies nearly on the least-
square plane defined by Ni, N(1), C(2), and C(3) (deviation from
the plane is 0.1613 Å for 5a and 0.1510 Å for 6a) (Table 3). The
dihedral angle between the Ni–N(1)–C(2) and Ni–C(4)–C(3) plane
is 19.597° for 5a and 20.284° for 6a. The planarity around Ni is de-
creased by formation of the (hetero)bimetallic complex; the C(5)
atom is out of the least-square plane around Ni for 0.6606 Å (5a)
and 0.6411 Å (6a), and the sum of the angles around Ni is 366°
(5a) and 365.9° (6a), respectively. The torsion angle of N(2)–
C(3)–C(4)–N(3) is 2.45° (5a) and 1.11° (6a), respectively, both of
which are significantly decreased due to the complexation of the
Table 2
Representative bond lengths and angles for azanickellacyclopentene complexes 1–4.
1
2
3
4
Bond length (Å)
Ni–X
Ni–N(1)
Ni–C(4)
Ni–C(5)
N(1)–C(2)
C(2)–C(3)
C(3)–C(4)
N(2)–C(3)
N(3)–C(4)
N(4)–C(5)
2.5770(8)
1.969(4)
1.906(5)
1.821(4)
1.317(8)
1.480(6)
1.505(9)
1.274(8)
1.255(6)
1.159(6)
1.988(2)
1.942(2)
1.882(2)
1.838(2)
1.290(3)
1.484(4)
1.499(3)
1.272(3)
1.261(3)
1.154(3)
2.3803(4)
1.968(2)
1.892(2)
1.816(2)
1.284(3)
1.483(2)
1.504(3)
1.264(3)
1.269(2)
1.153(2)
2.2287(5)
1.937(2)
1.899(2)
1.814(2)
1.290(2)
1.480(2)
1.512(2)
1.273(2)
1.267(2)
1.161(2)
Bond angles (°)
X–Ni–N(1)
X–Ni–C(4)
97.2(1)
173.2(2)
91.4(2)
169.2(3)
113.9(3)
111.8(5)
111.0(5)
103.9(3)
178.1(6)
89.39(8)
171.54(9)
97.71(9)
167.12(11)
114.6(2)
112.4(2)
110.9(2)
105.4(2)
174.3(2)
96.41(5)
174.45(7)
92.23(8)
169.21(11)
114.71(12)
112.4 (2)
109.7(2)
104.99(13)
177.2(2)
94.15(5)
178.00(6)
87.73(6)
177.90(8)
116.36(13)
112.6(2)
110.8(2)
X–Ni–C(5)
N(1)–Ni–C(5)
Ni–N(1)–C(2)
N(1)–C(2)–C(3)
C(2)–C(3)–C(4)
Ni–C(4)–C(3)
Ni–C(5)–N(4)
106.62(14)
173.2(2)
at around 20 ppm in the 13C NMR spectrum. It should be noted that
1H resonance due to the methyl group of the azanickellacycl-
a
opentene backbone appeared at around 1.45 ppm for both 5a and
5b, which was higher by ca. 0.75 ppm than that of mononuclear
azanickellacyclopentene precursor 1. This significant upfield shift
is attributed to effective shielding of the methyl group of the nick-
ellacyclopentene skeleton by one of the phenyl rings of the C6H3-
2,6-Me2 groups of the diimino moiety (vide infra). The absorption
band due to the coordinated CNC6H3-2,6-Me2 ligand is sensitive
Ar
N
Ar
N
Br
M
MBr2
r.t.
N
Ar
Br
N
Ni
N
N
X
Ar
Ar
5a : M = Zn, X = I
5b : M = Zn, X = Br
6a : M = Co, X = I
6b : M = Co, X = Br
7a : M = Fe, X = I
7b : M = Fe, X = Br
8a : M = Ni, X = I
Ni
C
X
Ar
C
(Ar = 2,6-Me 2-C6H3)
N
N
Ar
Ar
1: X = I
3: X = Br
Scheme 3. Synthesis of bimetallic complexes 5–8.
Fig. 2. Molecular structures of 5a (left) and 6a (right) with 50% probability ellipsoids. Hydrogen atoms were omitted for clarity.

D. Noda et al. / Polyhedron 28 (2009) 3935–3944
3939
Table 3
polymerization reactions were carried out to figure out the effect of
halogeno ligand on the nickel and the aluminum activator as
shown in Table 4. Polymerization was typically performed in a
100 mL stainless autoclave fitted with a Teflon inner tube in tolu-
ene (20 mL) for 20 min under 8 atm of ethylene gas in the presence
of an Al activator at room temperature. When the polymerization
was carried out with Cp2ZrCl2 under similar conditions, the activity
reached ca 2.0 kg/mmolZr h. This suggests that the activity over
1.0 kg/mmolNi h is a benchmark for the polymerization catalyst
showing good activity. As shown in entries 1–3, Lewis acidic alu-
minum activators, Et2AlCl, Me2AlCl, and MAO, are well assist the
polymerization with 1 to form polyethylene with similar melting
points. Under similar conditions, almost no polymerization oc-
curred using AlEt3 (activity: 6 g/mmolNi h) or AlMe3 (activity:
69 g/mmolNi h). A halogen or OTf ligand on the nickel atom also af-
fects the activity. As shown in Table 4, the order of catalytic activity
is 2 > 3 ꢀ 1 > 4 with MAO, whereas it is 2 > 1 > 3 > 4 with Et2AlCl.
Activity of polymerization reached over 1 kg/mmolNi h in the case
of 2/MAO, 2/Et2AlCl, and 1/Et2AlCl. It is of interest that inexpensive
Et2AlCl showed better activity than MAO. The activity of polymer-
ization is reproducible (672–1059 g/mmolNi h) when 200 equiv. of
the aluminum activator is used based on 1. Careful removal of
moisture is necessary for reproducible ethylene polymerization;
however, it is worthwhile to point out that polymerization pro-
ceeds well in the presence of even 20 equiv. of the aluminum acti-
vator. In the polymerization assisted by 20 equiv. of MAO, the
activity of 1 is 242 g/mmolNi h, and that of 2 is 395 mmolNi h.
Polyethylene obtained in the polymerization summarized in Ta-
ble 4 has Mn = 61 000–130 000 with PDI = 2.4–4.2. The GPC profile
is monomodal. 13C{1H} NMR spectroscopy suggests that the poly-
ethylene formed has relatively low branching content (ꢂ10C/
1000C). Increase of polymer branches lowers Tm as shown in en-
tries 1, 7, and 8.
Representative bond lengths and angles for bimetallic complexes 5a and 6a.
5a
6a
Bond length (Å)
Ni–I
M–Br(1)
M–Br(2)
Ni–N(1)
Ni–C(4)
Ni–C(5)
M–N(2)
M–N(3)
N(1)–C(2)
C(2)–C(3)
C(3)–C(4)
N(2)–C(3)
N(3)–C(4)
N(4)–C(5)
2.481(1)
2.394(1)
2.354(1)
1.941(8)
1.863(8)
1.824(9)
2.115(7)
2.062(7)
1.27(1)
1.48(1)
1.53(1)
1.28(1)
1.29(1)
2.493(1)
2.387(1)
2.358(1)
1.918(7)
1.863(7)
1.821(8)
2.053(6)
2.034(7)
1.30(1)
1.47(1)
1.53(1)
1.28(1)
1.288(9)
1.17(1)
1.16(1)
Bond angles (°)
X–Ni–N(1)
X–Ni–C(4)
97.9(2)
162.8(2)
87.8(3)
159.6(3)
79.1(3)
112.18(5)
116.5(6)
113.6(8)
110.6(7)
111.5(5)
168.6(7)
97.7(2)
162.8(2)
87.6(3)
160.1(3)
79.9(2)
108.62(5)
117.6(5)
111.6(7)
111.4(6)
111.5(5)
168.5(7)
X–Ni–C(5)
N(1)–Ni–C(5)
N(2)–M–N(3)
Br(1)–M–Br(2)
Ni–N(1)–C(2)
N(1)–C(2)–C(3)
C(2)–C(3)–C(4)
Ni–C(4)–C(3)
Ni–C(5)–N(4)
diimino moieties with the second transition metal fragments. The
bond lengths of Ni–I (2.481(1) Å for 5a and 2.493(1) Å for 6a) are
significantly shorter than that of 1 (2.5770(8) Å), which may be
due to the effective delocalization of the p-electron over the nick-
ellacyclopentene skeleton containing a heterometal at the diimino
substructure. The C6H3-2,6-Me2 ring on the N(1) atom of 5a and 6a
is nearly perpendicular to the square plane around nickel (dihedral
angles are 88.409° (5a) and 81.592° (6a)). Two C6H3-2,6-Me2 rings
of the diimino moieties are also arranged perpendicularly to the
least-square plane defined by M–N(2)–C(3)–C(4)–N(3) atoms,
and the dihedral angles are 83.502° and 86.746° for 5a, and
82.767° and 86.712° for 6a, respectively. The bond lengths of
C„N bond (1.16(1) Å for 5a and 1.17(1) Å for 6a) are comparable
to that of 1.
In Table 5 are summarized the data for ethylene polymerization
catalyzed by (hetero)bimetallic complexes. In entries 1–5 of this
table, the experiments were performed under the same conditions
as used in the polymerization shown in Table 4. Introduction of a
second metal, ZnBr2, CoBr2, FeBr2, or NiBr2, to the diimino moiety
of 1 apparently increased the catalytic activity (1.2–3.6 times as
much as the activity of 1) as shown in entries 1–5. The high cata-
lytic activity of the bimetallic catalysts prompted us to lower the
amount of the aluminum activator (see Table 5 entry 6–13). Since
the catalyst is sensitive to moisture and oxygen, the experiments
with 20 equiv. of Al activator to the catalyst required higher bime-
tallic catalyst contents (10 lmol) to obtain reproducible TON data
2.2. Ethylene polymerization catalyzed by 1–8
for the experiment using a 100 mL autoclave system. The activity
of the (hetero)bimetallic catalysts [entries 7–9 (MAO), 11–13
(Et2AlCl)] was 2.5–5.2 times that of the monometallic catalyst [en-
tries 6 (MAO) and 10 (Et2AlCl)]. Thus, the bimetallic catalysts, in
particular those containing Fe as the second metal, apparently
show excellent catalytic activity even in the presence of smaller
As described earlier, the azanickellacyclopentene complex 1
having N–C6H3-2,6-Me2 groups showed good reactivity towards
catalytic ethylene polymerization in the presence of MAO, though
no polyethylene was formed with the N–tBu homologue of 1 re-
ported by Otsuka [7c] under similar conditions. A series of ethylene
Table 4
Catalytic activities of azanickellacyclopentene complexes using MAO as activatora.
Entry
Catalyst
Al-cocatalyst
Yield (g)
Activity (g/mmol h)
Mn (ꢃ104)
Mw (ꢃ104)
Mw/Mn
Tm (°C)
Branch/1000C
1
2
3
4
5
6
7
8
9
1
1
1
2
3
4
2
3
4
MAO
0.197
0.353
0.407
0.385
0.232
0.174
0.413
0.329
0.201
591
7.9
8.2
–
7.9
13.0
12.0
6.1
6.6
8.0
26
30
–
33
32
37
23
28
30
3.3
3.7
–
4.2
2.4
3.0
3.9
4.2
3.7
125.0
128.3
127.4
127.3
133.0
132.1
123.4
122.6
128.1
8.3
6.6
–
5.9
2.4
3.4
11.5
9.7
6.1
AlEt2Cl
AlMe2Cl
MAO
MAO
MAO
AlEt2Cl
AlEt2Cl
AlEt2Cl
1059
1221
1155
696
522
1239
987
603
a
All reactions were carried out in a 100 mL stainless-steel autoclave in the presence of 200 equiv. of Al-cocatalysts based on the charged nickel catalysts (1
temperature for 20 min. Ethylene (0.8 MPa) was applied.
lmol) at room

3940
D. Noda et al. / Polyhedron 28 (2009) 3935–3944
Table 5
Catalytic activities of bimetallic complexesa.
Entry
Catalyst
Cat (l
mol)
A-cocatalyst
Al/Ni
Yield (g)
Activity (g/mmol h)
Mn (ꢃ104)
Mw (ꢃ104)
Mw/Mn
Tm (°C)
Branch/1000C
1
2
3
4
5
6
7
8
1
1
1
1
1
MAO
MAO
MAO
MAO
MAO
MAO
MAO
MAO
200
200
200
200
200
20
20
20
20
20
0.197
0.344
0.234
0.540
0.709
0.329
0.918
0.946
1.545
0.314
0.946
0.833
1.756
591
1032
702
1620
2127
99
275
284
464
102
284
250
527
7.9
8.8
9.7
5.6
6.7
–
6.9
7.5
2.8
–
26
29
28
19b
24b
–
21
22
13
–
3.3
3.3
2.9
3.4
3.6
–
3.1
2.9
4.7
–
125.0
128.1
128.1
127.4
132.9
–
126.0
127.9
126.2
–
8.3
4.2
4.3
7.6
4.4
–
9.9
8.3
6.5
–
5a
6a
7a
8
1
3
10
10
10
10
10
10
10
10
5b
6b
7b
3
5b
6b
7b
9
MAO
10
11
12
13
AlEt2Cl
AlEt2Cl
AlEt2Cl
AlEt2Cl
20
20
20
6.3
7.3
3
18
20
15
2.9
2.8
4.9
122.3
120.4
121.3
28.1
24.1
17.6
a
All reactions were carried out in a 100 mL stainless-steel autoclave in the presence of 20 or 200 equiv. of Al-cocatalysts based on the charged nickel catalysts (1 or
10
l
mol) at room temperature for 20 min. Ethylene (0.8 MPa) was applied.
b
The GPC profile is bimodal.
amounts of Al activator. Use of Et2AlCl provides somewhat higher
activity than MAO. There is no significant difference in Mn
(28 000–97 000) among the polymers obtained by the experiments
shown in Table 5. Although the polyethylene formed in most of the
experiments in this table has only a small number of polymer
branches (<10, entries 1–5, 7–9), it may be worthwhile to point
out that polyethylene formed by the experiments shown in entry
11–13 has relatively lower Tm (120–122 °C), of which the 13C
NMR spectrum showed the existence of a significantly higher num-
ber of branches (17.6–28.1). Of interest is the bimodal GPC profile
of the polymer obtained in the experiments using Ni–Fe and Ni–Ni
bimetallic catalysts (See Supplementary data in detail), which will
be discussed later.
Reaction of the monometallic catalyst with the alkylaluminum
activator results in dissociation of an isocyanide ligand and dis-
placement of a halogeno ligand by an alkyl group to give a coord-
inatively unsaturated intermediate I. The dissociation of the
isocyanide would be assisted by Lewis acidity of MAO or Et2AlCl,
whereas the halogeno ligands are good leaving group that will be
readily displaced by Me or Et groups leading to formation of the
Ni–C bond. Coordination of ethylene to I forms an intermediate
II; insertion of ethylene between the Ni–C bond in II initiates the
polymerization. Although the chain propagation takes place on
the cationic nickel species in the Brookhart catalyst, the catalytic
species for the chain propagation is neutral in the mononuclear
nickel complexes presented in this paper. It is well known from
mechanistic studies on the Brookhart catalyst that steric demands
around the nickel center, which are produced by the ortho-substit-
uents of the N-aryl moieties nearby, are important for the smooth
propagation to form polymer. This clearly explains the dramatic in-
crease of the catalytic activity by changing the N–tBu groups in Ot-
suka’s complex [7c] by the N–C6H3-2,6-Me2 groups; a structure of
the intermediate II shows effective blocking of the coordination
sphere by 2,6-dimethyl groups in the N–C6H3-2,6-Me2 ligand.
One explanation for the rate enhancement is that the second metal
species acts as the promoter for the polymerization. Introduction
of ZnX2 or CoX2 species to the mononuclear nickel complex pro-
vides an activity enhancement, while the GPC profile of formed
polyethylene is monomodal. Since Zn and Co diimines showed no
catalytic activity towards ethylene polymerization in the presence
of MAO, both of Zn and Co center are not involved in the catalytic
cycle. It should be noted that the molecular structures of 5 and 6
were determined by crystallography, indicating that the incorpora-
tion of the second metal fragment to the diimino moiety makes the
ring structure more rigid. The increased rigidity of the azanickella-
cyclopentene backbone would lead to an effective stabilization of
the azanickellacyclopentene framework during polymerization.
Similar influence of the rigidity of the ligand backbone was re-
ported by Wass et al., in which the nickel complexes having
NMe, CH2, and phenylene-bridged diphosphine ligand show some
2.3. Mechanistic considerations
The above described features of mono- and bimetallic nickella-
cyclopentanes as an ethylene polymerization catalyst can be sum-
marized as follows: (1) catalytic activity of the monometallic
catalyst 3 (X = Br) is 380 times higher than trans-(CNC6H3-2,6-
Me2)NiBr2 under the same conditions; however, that is ca. one third
of the Brookhart catalyst, [(C6H3-2,6-Me2)N@C(Me)C(Me)@N(C6H3-
2,6-Me2)]NiBr2; (2) the N–tBu analogue of the monometallic cata-
lyst 1 (X = I) showed no catalytic activity towards ethylene poly-
merization, (3) activity is affected by the halogeno ligand of the
monometallic catalysts, 1–4, and the order is X = OTf > I ꢀ Br > Cl.
This order is consistent with the ease of dissociation of X from the
nickel center, which is estimated by solution dynamics using
variable temperature NMR; (4) enhancement of the catalytic activ-
ity is achieved by introduction of the second metal. The order is
Ni > Fe > Co = Zn; (5) In supporting experiments, it was confirmed
that iron, cobalt, and zinc homologues of the Brookhart catalyst,
[(C6H3-2,6-Me2)N@C(Me)C(Me)@N(C6H3-2,6-Me2)]- MBr2 (M@Fe,
Co, Ni), showed no catalytic activity towards ethylene polymeriza-
tion [12,13].
As a possible mechanism to explain all of the features described
above, we proposed catalytic mechanism shown in Scheme 4.
N
N
N
Al activator
N
alkyl
N
X
N
alkyl
polyethylene
N
N
N
Ni
II
Ni
Ni
I
C
N
Scheme 4. Possible reaction mechanism for ethylene polymerization.

D. Noda et al. / Polyhedron 28 (2009) 3935–3944
3941
catalytic activity for polymerization of ethylene, whereas ethyl- or
propyl-bridged diphosphine complexes exhibited no activities
[14]. The incorporation of the second metal results in increased
planarity of the azanickellacyclopentene skeleton and increased
steric repulsion between the C6H3-2,6-Me2 ring on N(3) and the
second metal. These two factors force two C6H3-2,6-Me2 rings con-
nected to N(3) and N(1) close to the nickel center. It is known that
ortho-substituents in the Brookhart catalyst play a key role in
increasing activities by blocking the apical coordination sites of
the square-planar nickel species. In similar fashion, the ortho-
methyl groups on the C6H3-2,6-Me2 rings close to the nickel center
could effectively produce the sterically suitable circumstances
around the nickel center in the active species.
4A). 1H, 13C, 19F NMR spectra were recorded on a JEOL Lambda
600 or a Lambda 400 spectrometer at ambient temperature unless
otherwise noted. 1H, 13C, and 19F NMR chemical shifts (d values)
were given in ppm relative to the solvent signal (1H, 13C) or stan-
dard resonances (19F; external CF3COOH). IR spectra were recorded
on a JASCO FT/IR-550 spectrometer. Melting points were measured
on a Yanaco micro melting point apparatus. HRMS spectrum was
recorded on a JEOL Mstation JMS-70 apparatus. Elemental analyses
were performed by a Perkin–Elmer 2400/CHN analyzer. Starting
material, Ni(COD)2 [15], CNC6H3-2,6-Me2 and complex 1 [8], were
synthesized by the method reported in the literature.
The rate enhancement mechanism in which the second metal
acts as the promoter also explains the second metal effect of NiX2
giving significantly higher activity. However, it may be worth-
while to speculate that the Ni2 bimetallic catalysts may have
two active centers inside and outside of the metallacylic skeleton;
the active species of the former is neutral similar to that pro-
duced from 1, whereas the cationic active center would be
formed from the second NiBr2 species deduced from the analogy
to the Brookhart catalyst. Bimodal GPC profiles may support this
speculation. The polymerization with the Ni–Fe bimetallic cata-
4.2. Preparation of Iodo[1,2,3-tris(2,6-
dimethylphenylimino)butyl](2,6-dimethylphenylisocyanide)nickel (1)
To a solution of Ni(COD)2 (5.8 g, 21.2 mmol) in benzene (50 mL)
was added a solution of 2,6-xylylisocyanide (11.1 g, 85.2 mmol) in
benzene (100 mL), the mixture was stirred at room temperature
for 2 h, and the resulting solution was concentrated in vacuo. The
obtained pale yellow residue was washed with hexane three times
(in total 150 mL), and dissolved in benzene (100 mL). To this ben-
zene solution of crude Ni(CN–C6H3-2,6-Me2)4 was added methyl
iodide (6.6 mL, 106 mmol). After heating under reflux for 24 h,
the reaction mixture was concentrated in vacuo, and the brown
residue was washed with hexane (250 mL). Purification by recrys-
tallization from dichloromethane/hexane gave the desired com-
plex 1 in 95% yield (11.1 g) as dark red platelets. Mp 141 °C
(dec). Anal. Calc. for C37H39N4I1Ni1: C, 61.27; H, 5.42; N, 7.72.
Found: C, 61.48; H, 5.44; N, 7.61%. 1H NMR (600 MHz, C6D6, room
temperature): d 1.95 (s, 6H, CH3 of xylyl), 2.09 (s, 6H, CH3 of xylyl),
2.20 (s, 3H, CH3), 2.29 (s, 6H, CH3 of xylyl), 2.32 (s, 6H, CH3 of xylyl),
6.53–6.58 (m, 1H, Ph), 6.62–6.65 (m, 2H, Ph), 6.88–6.94 (m, 3H,
Ph), 7.00–7.09 (m, 3H, Ph), 7.15–7.16 (m, 3H, Ph). 13C{1H} NMR
(150 MHz, C6D6, room temperature): d 17.3 (s, CH3), 18.6 (s, CH3),
18.7 (s, CH3), 19.1 (s, CH3), 19.2 (s, CH3), 123.8, 124.1, 124.4,
126.1, 127.2, 127.7, 127.8, 128.0, 128.5, 128.9, 129.2, 134.7,
146.8, 148.6, 149.4, 164.5, 179.4, 186.4. IR (KBr):
lyst
7 may involve similar bimetallic active centers, which
showed high catalytic activity to give polyethylene having a
bimodal GPC profile.
3. Conclusion
As described above, we have been successful in preparing new
ethylene polymerization catalysts, which have a special structural
feature with an azanickellacyclopentene skeleton. The catalytic
activity of the monometallic catalyst 1–4 is ca. one third of the
Brookhart catalyst, [(C6H3-2,6-Me2)N@C(Me)C(Me)@N(C6H3-2,6-
Me2)]NiBr2; however, it is significantly higher than trans-
(CNC6H3-2,6-Me2)NiBr2 under the same conditions. It should be
noted that inexpensive AlEt2Cl can be used as the cocatalyst of this
catalyst system. Activities of mononuclear azanickellacyclopenten-
e complexes is increased in the order, M–Cl < M–I ꢀ M–Br < M–
OTf; this is explained by the order of X as a leaving group from
the nickel center. The active species for the polymerization is sug-
gested to be neutral, in other words, the catalyst is more close to
the SHOP catalyst but the Brookhart catalyst. The catalytic activity
is significantly increased by introduction of the second metal to the
diimio moiety of 1, which forms a series of heterobimetallic com-
plexes. Order of the polymerization activity is Ni–Ni > Fe–Ni > Co–
Ni@Zn–Ni. Increased rigidity of the azanickellacyclopentene skele-
ton provided by incorporation of the second metals at the diimino
moiety is a key factor in understanding the catalytic activity. In the
case of Ni–Ni and Fe–Ni bimetallic catalysts, two metals in the cat-
alyst molecule may produce individual active centers leading to
high catalytic activity and formation of polyethylene having bimo-
dal GPC features. The present results clearly demonstrate the mono
and bimetallic new C–N SHOP type compounds as an unique post-
metallocene ethylene polymerization catalyst.
m
C„N(cmꢁ1) = 2161 (s), C@N(cmꢁ1) = 1627 (s).
m
4.3. Preparation of trifluoromethanesulfonate[1,2,3-tris(2,6-
dimethylphenylimino)butyl](2,6-dimethylphenylisocyanide)nickel (2)
In a 50 mL Schlenk tube were placed complex 1 (800 mg,
1.10 mmol) and AgOTf (330 mg, 1.28 mmol) in CH2Cl2 (15 mL).
The resulting solution was stirred for 4 h at room temperature,
during which the color of the solution turned red. The solution
was filtered through a pad of celite, then the solvent was removed
in vacuo. The remaining dark red solid was dissolved in toluene
(30 mL) and layered with hexane (60 mL), from which orange crys-
tals of 2 were obtained in 89% yield (730 mg). Mp 105 °C (dec).
Anal. Calc. for C38H39N4O3S1F3Ni: C, 61.27; H, 5.42; N, 7.72. Found:
C, 61.48; H,5.44; N, 7.61%. 1H NMR (600 MHz, C6D6, room temper-
ature): d 1.72 (s, 3H, CH3), 1.97 (s, 6H, CH3 of xylyl), 2.03 (s, 6H, CH3
of xylyl), 2.21 (s, 6H, CH3 of xylyl), 2.26 (s, 6H, CH3 of xylyl), 6.43–
6.47 (m, 3H, Ph), 6.51–6.58 (m, 1H, Ph), 6.62–6.66 (m, 1H, Ph),
6.83–6.94 (m, 5H, Ph), 6.96–7.00 (m, 1H, Ph). 13C{1H} NMR
(150 MHz, C6D6, room temperature): d 16.1 (s, CH3), 18.2 (s, CH3),
18.3 (s, CH3), 18.7 (s, CH3), 19.1 (s, CH3), 124.1, 124.6, 125.8,
127.6, 128.1, 128.2, 128.9, 129.5, 135.8, 142.2, 148.6,
148.6, 163.7, 177.8, 179.9. 19F{1H} NMR (565 MHz, C6D6, room
4. Experimental
4.1. General procedures
The manipulation of air and moisture sensitive organometallic
compounds was carried out under a dry argon atmosphere using
standard Schlenk tube techniques associated with a high-vacuum
line. All solvents were distilled over appropriate drying reagents
prior to use (toluene, ether, THF, hexane, Ph2CO/Na, acetone, MS
temperature): d ꢁ78.8 (s, CF3). IR (KBr):
m
C„N(cmꢁ1) = 2175 (s),
C
1H3914N459Ni1: 597.2528. Found:
37
m
C@N(cmꢁ1) = 1640(s), S@O(cmꢁ1) = 1320, 1233, 1016 (s).
m
12
ESI-(HI)-MS. Calc. for
597.2531.

3942
D. Noda et al. / Polyhedron 28 (2009) 3935–3944
4.4. Preparation of bromo[1,2,3-tris(2,6-dimethylphenylimino)-
butyl](2,6-dimethylphenylisocyanide)nickel (3)
129.2, 129.3, 129.4, 133.7, 142.7, 148.3, 167.9, 176.2. IR (KBr):
m
C„N(cmꢁ1) = 2165 (s).
In a 20 mL Schlenk tube were placed complex 2 (490 mg,
0.66 mmol) and LiBr (88 mg, 1.01 mmol) in acetone (5 mL). The
resulting solution was stirred for 16 h at room temperature. The
solution was filtered through a pad of celite, then the solvent
was removed in vacuo. The remaining red solid was dissolved in
CH2Cl2 (10 mL) and layered with hexane (30 mL), from which red
crystals of 3 were obtained in 72% yield (320 mg). Mp 156 °C
(dec). Anal. Calc. for C37H39N4Br1Ni1: C, 65.51; H, 5.80; N, 8.26.
Found: C, 65.48; H, 5.94; N, 8.13%. 1H NMR (600 MHz, C6D6, room
temperature): d 1.93 (s, 6H, CH3 of xylyl), 2.05 (s, 6H, CH3 of xylyl),
2.18 (s, 3H, CH3), 2.23 (s, 6H, CH3 of xylyl), 2.35 (s, 6H, CH3 of xylyl),
6.50–6.56 (m, 1H, Ph), 6.60–6.67 (m, 2H, Ph), 6.86–6.95 (m, 3H, Ph),
6.98–7.03 (m, 2H, Ph), 7.05–7.11 (m, 1H, Ph), 7.13–7.14 (m, 3H,
Ph). 13C{1H} NMR (150 MHz, C6D6, room temperature): d 17.2 (s,
CH3), 18.5 (s, CH3), 18.6 (s, CH3), 18.9 (s, CH3), 19.2 (s, CH3),
123.8, 124.1, 124.4, 126.1, 127.2, 127.8, 127.9, 128.0, 128.4,
129.1, 129.3, 134.9, 148.7, 149.8, 164.7, 179.5. IR (KBr):
4.7. Preparation of {bromo[1,2,3-tris(2,6-dimethylphenylimino)-
butyl](2,6-dimethylphenylisocyanide)nickel}(dibromo)zinc(II) (5b)
In a 20 mL Schlenk tube were placed complex 3 (71 mg,
0.11 mmol) and ZnBr2 (30 mg, 0.13 mmol) in Et2O (10 mL). The
resulting suspension was stirred for 8 h at room temperature.
The solvent was removed in vacuo, and the residue was washed
with hexane (30 mL). The remaining dark red solid was dissolved
in CH2Cl2 (10 mL), centrifuged to remove the remaining solid,
and layered with hexane (60 mL), from which red crystals of 5b
were obtained in 67% yield (68 mg). The crystals were used for
the ethylene polymerization. Mp 144 °C (dec). 1H NMR (600 MHz,
C6D6, room temperature): d 1.46 (s, 3H, CH3), 2.23 (s, 6H, CH3 of xy-
lyl), 2.33 (s, 6H, CH3 of xylyl), 2.46 (s, 6H, CH3 of xylyl), 2.68 (s, 6H,
CH3 of xylyl), 6.48–6.52 (m, 1H, Ph), 6.82–6.86 (m, 2H, Ph), 6.92–
7.01 (m, 2H, Ph), 7.09–7.14 (m, 4H, Ph), 7.15–7.16 (m, 3H, Ph).
13C{1H} NMR (150 MHz, C6D6, room temperature): d 17.1 (s,
CH3), 18.5 (s, CH3), 18.6 (s, CH3), 18.8 (s, CH3), 19.2 (s, CH3),
123.7, 124.0, 124.4, 126.0, 127.1, 127.7, 127.9, 128.4, 128.6,
129.1, 129.4, 134.9, 144.1, 148.7, 150.1, 164.9, 179.4, 184.1. IR
m
C„N(cmꢁ1) = 2164 (s),
m
C@N(cmꢁ1) = 1628 (s). HR-MS. Calc. for
12
C
37
1H3914N481Br159Ni1: 678.1690. Found: 678.1700.
4.5. Preparation of chloro[1,2,3-tris(2,6-dimethylphenylimino)butyl]-
(KBr): m m
C„N(cmꢁ1) = 2183 (s), C@N(cmꢁ1) = 1632 (s).
(2,6-dimethylphenylisocyanide)nickel (4)
4.8. Preparation of {iodo[1,2,3-tris(2,6-dimethylphenylimino)-
In a 50 mL Schlenk tube were placed complex 2 (1290 mg,
1.73 mmol) and LiCl (88 mg, 2.08 mmol) in acetone (40 mL). The
resulting solution was stirred for 16 h at room temperature. The
solution was filtered through a pad of celite, then the solvent
was removed in vacuo. The remaining red solid was dissolved in
toluene (20 mL) and layered with hexane (50 mL), from which
red crystals of 4 were obtained in 63% yield (639 mg). Mp 160 °C
(dec). Anal. Calc. for C37H39N4Cl1Ni1: C, 70.11; H, 6.20; N, 8.84.
Found: C, 70.10; H, 6.35; N, 8.82%. 1H NMR (600 MHz, C6D6, room
temperature): d 1.94 (s, 6H, CH3 of xylyl), 2.06 (s, 6H, CH3 of xylyl),
2.18 (s, 3H, CH3), 2.23 (s, 6H, CH3 of xylyl), 2.36 (s, 6H, CH3 of xylyl),
6.48–6.54 (m, 1H, Ph), 6.61–6.67 (m, 2H, Ph), 6.87–7.11 (m, 6H,
Ph), 7.13–7.14 (m, 3H, Ph). 13C{1H} NMR (150 MHz, C6D6, room
temperature): d 17.1 (s, CH3), 18.5 (s, CH3), 18.6 (s, CH3), 18.8 (s,
CH3), 19.2 (s, CH3), 123.7, 124.0, 124.4, 126.0, 127.1, 127.7, 127.9,
128.4, 128.6, 129.1, 129.4, 134.9, 144.1, 148.7, 150.1, 164.9,
butyl](2,6-dimethylphenylisocyanide)nickel}(dibromo)cobalt(II) (6a)
In a 100 mL Schlenk tube were placed complex 1 (240 mg,
0.33 mmol) and CoBr2 (72 mg, 0.33 mmol) in benzene (20 mL)
and THF (5 mL). The resulting mixture was stirred for 24 h at room
temperature, and the solvent was removed in vacuo. The obtained
crude product was used for ethylene polymerization without fur-
ther purification. The crystals of 6a suitable for X-ray diffraction
analysis can be isolated by recrystallization from CH2Cl2/hexane
as dark red crystals. Mp 193 °C (dec). Anal. Calc. for C37H39Br2Co1-
N4I1Ni1ꢄCH2Cl2: C, 44.35; H, 4.02; N, 5.44. Found: C, 44.59; H,
3.99; N, 5.47%. IR (KBr): m
C„N(cmꢁ1) = 2159 (s).
4.9. Preparation of Ni–Fe (7a, 7b) and Ni–Ni (8a)
(hetero)bimetallic complexes
179.4, 184.1. IR (KBr):
(s). ESI-(HI)-MS. Calc. for
m
C„N(cmꢁ1) = 2175 (s),
m
C@N(cmꢁ1) = 1633
As a typical example, to a solution of 1 (220 mg, 0.31 mmol) in
benzene was added a solution of FeBr2 (66 mg, 0.31 mmol) in THF.
After stirring for 3 h, the reaction mixture was evacuated under
vacuum. The resulting solid was dissolved in CH2Cl2 (20 mL), and
stirred at room temperature for 12 h. The solvent was removed un-
der reduced pressure, and the crude product was used for the eth-
ylene polymerization without further purification. The Ni(Br)–Fe
(7b), and Ni(I)–Ni (8) (hetero)bimetallic complexes were prepared
according to the similar procedure using complex 3 (100 mg,
0.15 mmol) and FeBr2 (30 mg, 0.14 mmol) in Et2O/THF (10 mL/
1 mL) for 7b, and complex 1 (131 mg, 0.18 mmol) and (DME)NiBr2
(56 mg, 0.18 mmol) in benzene (20 mL) for 8a.
C
1H3914N435Cl159Ni1: 632.2217.
37
12
Found: 632.2218.
4.6. Preparation of {iodo[1,2,3-tris(2,6-dimethylphenylimino)-
butyl](2,6-dimethylphenylisocyanide)nickel}(dibromo)zinc(II) (5a)
To a solution of ZnBr2 (170 mg, 0.75 mmol) in EtOH (10 mL) was
added a solution of 1 (530 mg, 0.73 mmol) in CH2Cl2 (10 mL). After
it was stirred at room temperature for 24 h, the reaction mixture
was concentrated under a reduced pressure. The remaining crude
product was used for ethylene polymerization without further
purification. For the characterization of the product, purification
by recrystallization from CH2Cl2/hexane gave the desired complex
5a in 77% (540 mg) yield as dark red crystals. Mp 139 °C (dec). Anal.
Calc. for C37H39N4Br2Cl1Ni1Zn1: C, 46.75; H, 4.53; N, 5.89. Found: C,
46.45; H, 4.53; N, 5.68%. 1H NMR (600 MHz, C6D6, room tempera-
ture): d 1.45 (s, 3H, CH3), 2.24 (s, 6H, CH3 of xylyl), 2.34 (s, 6H,
CH3 of xylyl), 2.47 (s, 6H, CH3 of xylyl), 2.69 (s, 6H, CH3 of xylyl),
6.48–6.52 (m, 1H, Ph), 6.82–6.86 (m, 2H, Ph), 6.95–6.97 (m, 2H,
Ph), 6.98–7.00 (m, 1H, Ph), 7.11–7.17 (m, 6H, Ph). 13C{1H} NMR
(150 MHz, C6D6, room temperature): d 18.5 (s, CH3), 19.2 (s, CH3),
19.4 (s, CH3), 20.0 (s, CH3), 20.8 (s, CH3), 127.2, 127.2, 127.5,
127.6, 127.7, 127.9, 128.1, 128.2, 128.4, 128.6, 128.7, 128.8,
4.10. Ethylene polymerization
In a 100 mL stainless autoclave fitted with a Teflon inner tube
was placed the precatalyst 1–8 (1 or 10 lmol) and toluene
(20 mL). A toluene solution of MAO (PMAO-S available from TOS-
OH-FINECHEM Corp. 2.85 M in toluene) or a toluene solution of
AlEt2Cl (1.0 M in toluene) was added via syringe (20 equiv. or
200 equiv. for precursor 1–8) and the mixture was stirred for 1 h
at room temperature. Then ethylene (0.8 MPa) was introduced,
and the mixture was stirred at room temperature for 20 min. Unre-
acted ethylene was released, and methanol (10 mL) was added to

D. Noda et al. / Polyhedron 28 (2009) 3935–3944
3943
Table 6
Crystallographic data for 1, 2, 3, 4, 5a and 6a.
1
2
3
4
5a
6a
Empirical formula
Formula weight
Crystal system
Lattice type
Space group
a (Å)
C37H39IN4Ni
725.35
C38H39F3O3SN4Ni
747.51
triclinic
C37H39BrN4Ni
678.35
C37H39ClN4Ni
633.89
C37H39Br2IN4NiZnꢄCH2Cl2 C37H39N4Br2CoINiꢄCH2Cl2
1035.47
monoclinic
primitive
P21/c (#14)
15.091(2)
11.585(2)
22.938(4)
90
1029.02
monoclinic
primitive
P21/c (#14)
15.211(1)
11.6798(8)
22.888(2)
90
monoclinic
C-centered
C2/c (#15)
29.374(9)
14.767(4)
17.319(5)
90
116.5987(11)
90
6718(3)
8
monoclinic
C-centered
C2/c (#15)
29.111(7)
14.872(3)
16.959(4)
90
116.4000(9)
90
6576(3)
8
orthorhombic
primitive
Pbca (#61)
17.791(2)
15.265(2)
24.375(3)
90
primitive
ꢀ
P1 (#2)
10.828(2)
13.441(2)
13.623(2)
72.866(5)
82.243(6)
74.594(6)
1823.2(4)
2
b (Å)
c (Å)
a
(°)
b (°)
90
90
97.1574(8)
90
96.590(2)
90
c
(°)
Volume, Å3
6619.9(13)
8
1.272
3979.1(11)
4
4039.5(5)
4
Z value
Dcalc, g/cm3
1.434
1.362
1.370
1.728
1.692
F(0 0 0)
2960.00
Dark red, platelet
780.00
brown, prism
2816.00
brown, prism
2672.00
red, platelet
2048.00
brown, chip
2036.00
dark red, platelet
0.50 ꢃ 0.50 ꢃ 0.03
8837
Crystal color, habit
Crystal dimensions, mm
No of observations (all
reflections)
0.24 ꢃ 0.08 ꢃ 0.06 0.15 ꢃ 0.15 ꢃ 0.10 0.21 ꢃ 0.12 ꢃ 0.11 0.25 ꢃ 0.10 ꢃ 0.03 0.15 ꢃ 0.10 ꢃ 0.05
7622
7982
7540
7572
5062
No of variables
427
490
427
427
483
443
Reflection/parameter ratio 17.85
16.29
0.0466
0.0628
0.1394
1.002
0.000
17.66
0.0659
0.0314
0.0276
0.876
0.000
17.73
0.0601
0.0406
0.1069
1.019
0.000
10.48
–
0.055(I > 3
0.161(I > 3
1.004
19.95
0.109
0.069
0.225
1.048
0.002
R (all reflections)
0.092
0.042
0.111
1.000
0.000
R1(I > 2
r
(I))a
r
r
(I))
(I))
wR2 (all reflections)b
Goodness of fit (GOF)
Maximum shift/error in
final cycle
0.000
Maximum peak in final
difference map, e/Å3
Minimum peak in final
difference map, e/Å3
1.06
0.98
0.75
0.77
3.52
1.66
ꢁ0.98
ꢁ0.86
ꢁ0.82
ꢁ0.37
ꢁ1.48
ꢁ2.09
a
R1
wR2 = [
=
R
|Fo|ꢁ|Fc|/
R|Fo|.
b
2
R
(w(Fo ꢁFc2)2)/
R
(w(Fo2)2)]1/2
.
stop the polymerization. Polyethylene was precipitated when the
reaction mixture was poured into a methanol solution of HCl.
The polymer was collected by filtration and dried in vacuo.
and 6a. These data can be obtained free of charge via http://
www.ccdc.cam.ac.uk/conts/retrieving.html, or from the Cambridge
Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ,
UK; fax: (+44) 1223-336-033; or e-mail: deposit@ccdc.cam.ac.uk.
4.11. X-ray data collection and reduction
Acknowledgement
X-ray crystallography for new complexes 2, 3, and 4 was
performed on a Rigaku Saturn CCD area detector with graphite
This work was supported by Grant-in-Aid for Science Research
on Priority Areas (No. 18064014, Synergy of Elements) from Minis-
try of Education, Culture, Sports, Science and Technology, Japan.
monochromated Mo K
a
radiation (k = 0.71070 A). The data were
scan in the h range of 3.0 6
collected at 123(2) K using
x
h 6 27.5° (2), 3.0 6 h 6 27.5° (3) and 3.0 6 h 6 27.5° (4). The data
obtained were processed using Crystal-Clear (Rigaku) on a Pentium
computer, and were corrected for Lorentz and polarization effects.
The structures were solved by direct methods [16], and expanded
using Fourier techniques [17]. The non-hydrogen atoms were re-
fined anisotropically. Hydrogen atoms were refined using the rid-
ing model. The final cycle of full-matrix least-squares refinement
on F2 was based on 7982 observed reflections and 490 variable
parameters for 2, 7540 observed reflections and 427 variable
parameters for 3, and 7572 observed reflections and 427 variable
parameters for 4. Neutral atom scattering factors were taken from
Cromer and Waber [18]. All calculations were performed using the
CRYSTALSTRUCTURE [19,20] crystallographic software package. Details
of final refinement as well as the bond lengths and angles are sum-
marized in Table 6 in the supporting information, and the number-
ing scheme employed is also shown in the supporting information,
which were drawn with ORTEP at 50% probability ellipsoid.
References
[1] (a) S.D. Ittel, L.K. Johnson, M. Brookhart, Chem. Rev. 100 (2000) 1169;
(b) V.C. Gibson, S.K. Spitzmesser, Chem. Rev. 103 (2003) 283;
(c) G.J.P. Britovsek, V.C. Gibson, D.F. Wass, Angew. Chem., Int. Ed. 38 (1999)
428;
(d) S. Mecking, Angew. Chem., Int. Ed. 40 (2001) 534.
[2] L.K. Johnson, C.M. Killian, M. Brookhart, J. Am. Chem. Soc. 117 (1995) 6414.
[3] (a) W. Keim, Angew. Chem., Int. Ed. 29 (1990) 235;
(b) M. Peuckert, W. Keim, Organometallics 2 (1983) 594;
(c) P. Kuhn, D. Sémeril, D. Matt, M.J. Chetcuti, P. Lutz, Dalton. Trans. (2007)
515;
(d) J. Skupin´ ska, Chem. Rev. 91 (1991) 613.
[4] (a) G.A. Luinstra, J. Queisser, B. Bildstein, H. -H. Görtz, C. Amort, M. Malaun, A.
Krajete, G. Werne, M.O. Kristen, N. Huber, C. Gernert, in: B. Rieger, L. Saunders
Baugh, S. Kacker, S. Striegler, (Eds.), Late Transition Metal Polymerization
Catalysis, Wiley-VCH, Weinheim, Germany, 2003, p. 59, and references
therein.;
(b) M.P. Batten, A.J. Canty, K.J. Cavell, T. Rüther, B.W. Skelton, A.H. White,
Inorg. Chim. Acta 359 (2006) 1710;
(c) J.D. Azoulay, K. Itigaki, G. Wu, G.C. Bazan, Organometallics 27 (2008) 2273;
(d) B.Y. Lee, G.C. Bazan, J. Vela, Z.J.A. Komon, X. Bu, J. Am. Chem. Soc. 14 (2001)
5352;
(e) S.J. Diamanti, V. Khanna, A. Hotta, D. Yamakawa, F. Shimizu, E. Kramr, G.H.
Fredrickson, G.C. Bazan, J. Am. Chem. Soc. 126 (2004) 10528;
(f) S.J. Diamanti, P. Ghosh, F. Shimizu, G.C. Bazan, Macromolecules 36 (2003)
9731;
Supplementary data
CCDC 198229, 689689, 689690, 689691, 265483 and 265482
contain the supplementary crystallographic data for 1, 2, 3, 4, 5a

3944
D. Noda et al. / Polyhedron 28 (2009) 3935–3944
(g) R.S. Rojas, J.-C. Wasilke, G. Wu, J.W. Ziller, G.C. Bazan, Organometallics 24
(2005) 5644;
(q) R.A.T.M. Abbenhuis, I. del Río, M.M. Bergshoef, J. Boersma, N. Veldman, A.L.
Spek, G. van Koten, Inorg. Chem. 37 (1998) 1749;
(h) R.S. Rojas, G.B. Galland, G. Wu, J.W. Ziller, G.C. Bazan, Organometallics 26
(2007) 5339;
(i) S.J. Diamanti, V. Khanna, A. Hotta, R.C. Coffin, D. Yamakawa, E.J. Kramer,
G.H. Fredrickson, G.C. Bazan, Macromolecules 39 (2003) 3270.
(r) J.-P. Sutter, S.L. James, P. Steenwinkel, T. Karlen, D.M. Grove, N. Veldman,
W.J.J. Smeets, A.L. Spek, G. van Koten, Organometallics 15 (1996) 941;
(s) J.O. Krause, O. Nuyken, K. Wurst, M.R. Buchmeiser, Chem. Eur. J. 10 (2004)
777;
[5] M. Tanabiki, K. Tsuchiya, Y. Kumanomido, K. Matsubara, Y. Motoyama, H.
Nagashima, Organometallics 23 (2004) 3976.
[6] (a) Y. Yamamoto, H. Yamazaki, Coord. Chem. Rev. 8 (1972) 225;
(b) P.C.J. Kamer, R.J.M. Nolte, W. Drenth, H.L.L. Nijs, J.A. Kanters, J. Mol. Catal.
49 (1988) 21;
(t) C.P. Casey, T.E. Vos, S.W. Singer, I.G. Guzei, Organometallics 21 (2002) 5038.
[10] (a) G.R. Owen, R. Vilar, A.J.P. White, D.J. Williams, Organometallics 21 (2002)
4799;
(b) T. Tanase, T. Ohizumi, K. Kobayashi, Y. Yamamoto, Organometallics 15
(1996) 3404;
(c) T.J. Deming, B.M. Novak, Macromolecules 26 (1993) 7092;
(d) F. Millich, Macromol. Rev. 15 (1980) 207;
(c) Y. Yamamoto, T. Tanase, T. Yanai, T. Asano, K. Kobayashi, J. Organomet.
Chem. 456 (1993) 287;
(e) T.J. Deming, B.M. Novak, J. Am. Chem. Soc. 115 (1993) 9101;
(f) W. Drenth, R.J.M. Noite, Act. Chem. Res. 12 (1979) 30.
[7] (a) E. Carmona, J.M. Marín, P. Palma, M.L. Poveda, J. Organomet. Chem. 377
(1989) 157;
(d) Y. Yamamoto, H. Yamazaki, Bull. Chem. Soc. Jpn. 43 (1970) 2653;
(e) Y. Yamamoto, H. Yamazaki, Inorg. Chem. 13 (1974) 438;
(f) J. Vicente, J. -A. Abad, E. Martínez-Viviente, P.G. Jonas, Organometallics 21
(2002) 4454;
(b) S. Otsuka, A. Nakamura, T. Yoshida, M. Naruto, K. Ataka, J. Am. Chem. Soc.
95 (1973) 3180;
(g) S.S.M. Hassa, W.H. Mohmoud, S.-S.S.H. Elgazwy, N.M. Badawy,
Electroanalysis 18 (2006) 2070.
(c) S. Otsuka, A. Nakamura, T. Yoshida, J. Am. Chem. Soc. 91 (1969) 7196;
(d) S. Otsuka, A. Nakamura, K. Ito, Chem. Lett. (1972) 939.
[8] A preliminary account of this work appeared in M. Tanabiki, K. Tsuchiya, Y.
Motoyama, H. Nagashima, Chem. Commun. (2005) 3409.
[9] (a) M.J.A. Kraakman, B. de Klerk-Engels, P.P.M. de Lange, K. Vrieze, W.J. Smeets,
A.L. Spek, Organometallics 11 (1992) 3774;
[11] [a] See Ref. [5].;
(b) Y. Yamazaki, K. Ehara, K. Takahashi, H. Yamazaki, Bull. Chem. Soc. Jpn. 64
(1991) 3376;
(c) Y. Yamamoto, H. Takahata, F. Takei, J. Organomet. Chem. 545 (1997) 369.
[12] The [(o-xylyl)N@C(Me)C(Me)@N(o-xylyl)]MBr2 (M = Fe, Co, Ni, and Zn)
complexes were generated in situ by treatment of FeBr2, CoBr2, NiBr2(dme),
and ZnBr2 with 1 equiv. of (o-xylyl)N@C(Me)C(Me)@N(o-xylyl) in THF for
overnight at room temperature. Ethylene polymerization catalyzed by [(o-
xylyl)N@C(Me)C(Me)@N(o-xylyl)]MBr2 (M = Fe, Co, Ni, and Zn) was carried out
in a 100 mL stainless steel autoclave fitted with a Teflon inner tube under the
same condition (catalyst: 1 lmol, MAO: 200 equiv.) described in this paper.
[13] V. Rosa, S.A. Carabineiro, T. Avilés, P.T. Gomes, R. Welter, J.M. Campos, M.R.
Ribeiro, J. Organomet. Chem. 693 (2008) 769.
(b) I. del Río, R.A. Gossage, M.S. Hannu, M. Lutz, A.L. Spek, G. van Koten,
Organometallics 18 (1999) 1097;
(c) B. de Klerk-Engels, J.H. Groen, M.J.A. Kraakman, J.M. Ernsting, K. Vrieze, K.
Goubitz, J. Fraanje, Organometallics 13 (1994) 3279;
(d) M.A.O. Volland, S.M. Hansen, F. Rominger, P. Hofmann, Organometallics 23
(2004) 800;
(e) J.M. Bikley, A.A. La Pensee, S.J. Higgins, C.A. Stuart, Dalton Trans. (2003)
4663;
(f) A. Svetlanova-Larsen, C.R. Zoch, J.L. Hubbard, Organometallics 15 (1966)
3076;
[14] (a) N.A. Cooley, S.M. Green, D.F. Wass, K. Heslop, A.G. Orpen, P.G. Pringle,
Organometallics 20 (2001) 4769;
[b] A. Carter, N.A. Cooley, P.G. Pringle, J. Scutt, D.F. Wass, PMSE-Preprints,
223rd ACS National Meeting, vol. 86, Orlando, FL, April 7–11, 2002, American
Chemical Society, Washington, DC, 2002, p. 314.
(g) R.M. Barns, J.L. Hubbard, J. Am. Chem. Soc. 116 (1994) 9514;
(h) A.C. Ontko, J.F. Houlis, R.C. Schnabel, D.M. Roddick, T.P. Fong, A.J. Lough,
R.H. Morris, Organometallics 17 (1998) 5467;
[15] R.A. Schunn, Inorg. Synth. 15 (1974) 5.
(i) D. Huang, J.C. Bollinger, W.E. Streib, K. Folting, V. Young Jr., O. Eisenstein,
K.G. Caulton, Organometallics 19 (2000) 2281;
(j) T. Funaioli, C. Cavazza, F. Marchetti, G. Fachinetti, Inorg. Chem. 38 (1999)
3361;
(k) P.W. Blosser, J.C. Gallucci, A. Wojcicki, Inorg. Chem. 31 (1992) 2376;
(l) T.P. Fong, C.E. Forde, A.L. Lough, R.H. Morris, P. Rigo, E. Rocchini, T. Stephan,
J. Chem. Soc., Dalton Trans. (1999) 4475;
[16] (a) G.M. Sheldrick, SHELX97 (1997);
[b] A. Altomare, M.C. Burla, M. Camalli, G. Cascarano, A. Guagliardi, A.
Moliterni, G. Polidori, R. Spagna, SIR97, 1999.
[17] P.T. Beurskens, G. Admiraal, G. Beurskens, W.P. Bosman, R. de Gelder, R. Israel,
J.M.M. Smits, DIRDIF99, The DIRDIF-99 program system; Technical Report of the
Crystallography Laboratory, University of Nijmegen, Nijmegen, The
Netherlands, 1999.
(m) C.J. den Reijer, M. Wörle, P.S. Pregosin, Organometallics 19 (2000) 309;
(n) F.A. Cotton, C.A. Murillo, J.H. Reinbenspies, D. Villagrán, X. Wang, C.C.
Wilkinson, Inorg. Chem. 43 (2004) 8373;
(o) C. Gemel, D. Kalt, K. Mareiter, V.N. Sapunov, R. Schmid, K. Kirchner,
Organometallics 16 (1997) 427;
[18] D.T. Cromer, J.T. Waber, International Tables for X-ray Crystallography, vol. 4,
Kynoch Press, Birmingham, UK, 1974.
[19] CRYSTALSTRUCTURE 3.8.0: Crystal Structure Analysis Package; Package, Rigaku and
Rigaku/MSC. 9009 New Trails Dr., The Woodlands, TX 77381, USA, 2000–2006.
[20] J.R. Carruthers, J.S. Rollett, P.W. Betteridge, D. Kinna, L. Pearce, A. Larsen, E.
Gabe, Chemical Crystallography Laboratory, Oxford, UK, 1999.
(p) M.D. Hagereaves, M.F. Mahon, M.K. Whittlesey, Inorg. Chem. 41 (2002)
3137;