New chromium complexes for ethylene oligomerization: Extended use of tridentate ligands in metal-catalyzed olefin polymerization

Article abstract of DOI:10.1021/ma035554b

A family of chromium complexes bearing tridentate pyridine-based ligands are disclosed as highly active precatalysts for the oligomerization of ethylene. The ligands are comprised of two distinct types: Type 1, in which both ketone groups of 2,6-diacetylpyridine are converted to imines to produce pyridine bisimine NNN ligands; and Type 2, in which only one ketone group of 2,6-diacetylpyridine is condensed with an aniline derivative to give monoimine NNO coordination sets. Ligands of either type are coordinated to chromium(II) or chromium(III) chlorides, and activation of the resultant complexes with methylaluminoxane (MAO) produces highly active ethylene oligomerization and polymerization catalysts. Catalysts of Type 1 (NNN set) generally produce 1-butene when only two ortho alkyl substituents are present but switch to making waxes or polyethylene when the size and/or number of ortho substituents are increased. Catalysts of Type 2 (NNO set) produce waxes and polyethylene under all of the substitution patterns studied. The butene-producing catalysts can make 1-butene with 99.5+% purity, and the wax-producing catalysts make highly linear to moderately branched waxes, depending on the presence of an α-olefin comonomer.

Full text of DOI:10.1021/ma035554b

Macromolecules 2004, 37, 4375-4386
4375
New Chromium Complexes for Ethylene Oligomerization: Extended
Use of Tridentate Ligands in Metal-Catalyzed Olefin Polymerization
Br ook e L. Sm a ll,*,† Mich a el J . Ca r n ey,‡ Da n a h M. Holm a n ,‡
Colleen E. O’Rou r k e,‡ a n d J a son A. Ha lfen ‡
Chevron Phillips Chemical Company, LP, 1862 Kingwood Drive, Kingwood, Texas 77339, and
Department of Chemistry, University of WisconsinsEau Claire, 105 Garfield Avenue,
Eau Claire, Wisconsin 54702
Received October 15, 2003; Revised Manuscript Received April 6, 2004
ABSTRACT: A family of chromium complexes bearing tridentate pyridine-based ligands are disclosed
as highly active precatalysts for the oligomerization of ethylene. The ligands are comprised of two distinct
types: Type 1, in which both ketone groups of 2,6-diacetylpyridine are converted to imines to produce
pyridine bisimine NNN ligands; and Type 2, in which only one ketone group of 2,6-diacetylpyridine is
condensed with an aniline derivative to give monoimine NNO coordination sets. Ligands of either type
are coordinated to chromium(II) or chromium(III) chlorides, and activation of the resultant complexes
with methylaluminoxane (MAO) produces highly active ethylene oligomerization and polymerization
catalysts. Catalysts of Type 1 (NNN set) generally produce 1-butene when only two ortho alkyl substituents
are present but switch to making waxes or polyethylene when the size and/or number of ortho substituents
are increased. Catalysts of Type 2 (NNO set) produce waxes and polyethylene under all of the substitution
patterns studied. The butene-producing catalysts can make 1-butene with 99.5+% purity, and the wax-
producing catalysts make highly linear to moderately branched waxes, depending on the presence of an
R-olefin comonomer.
In tr od u ction
The development of homogeneous chromium-based
catalysts for olefin polymerization has received consid-
erable attention,1 spurred by the extensive use of
chromium catalysts in commercial ethylene polymeri-
zation.2 Seemingly at the interface of late-transition-
metal and early-metal metallocene catalysts, chromium
has been used in a number of different polymerization
studies, employing cyclopentadienyl-type ligands, non-
Cp ligands possessing heteroatom donor sets and vari-
ous mixed systems. Figure 1 provides an overview of
some recently reported homogeneous chromium cata-
lysts.
As can be seen from Figure 1, the complexes vary not
only in their structures, but also in oxidation state (+2
vs +3), formal electron count, and overall charge
(neutral vs cationic).3 Furthermore, although it is
reasonable that more electrophilic cationic systems
should exhibit higher activities than neutral ones,
correlations between chromium oxidation state (or
formal electron count) and catalyst productivity are
vague at best. For example, neutral catalysts A and B,
reported by Theopold,4 and C, reported by Carney,5 have
relatively low activities, but potentially cationic (after
MAO activation) complexes D, E, and F are only
marginally more active.1,6 Since D only has only two
nitrogen donors, the catalyst might be too electron
deficient. Using this line of reasoning, one might expect
that tridentate monoanionic ligands would yield more
active catalysts. However, complex G, containing a
tridentate monoanionic ligand, only exhibits moderate
polymerization activity.7 One of the more promising
F igu r e 1. Selected examples of homogeneous chromium-
based polymerization catalysts. In each example the reported
precatalyst structure is shown. Many of these complexes
require an alkylating cocatalyst in order to achieve activation.
Speculation on the nature of the active species for each system
is well beyond the intent of this work.
results in homogeneous chromium catalysts, complex H,
was recently reported by J olly,8 who disclosed that by
changing the well-known ‘constrained geometry’ ligand’s
anionic amido ligand to a neutral amine, highly active
cationic chromium(III) catalysts could be produced upon
activation with MAO.
The performance of type H catalysts seems to indicate
that cationic chromium(III) complexes with carefully
chosen monoanionic ligands might be the best alterna-
tive for homogeneous chromium systems. Complex I,
though recently reported to polymerize propylene with
modest rates,9 caused us to further consider this issue
of catalyst design. Ethylene polymerization data for I
* To whom correspondence should be addressed. E-mail: small-
bl@cpchem.com.
† Chevron Phillips Chemical Co.
‡ University of WisconsinsEau Claire.
10.1021/ma035554b CCC: $27.50 © 2004 American Chemical Society
Published on Web 05/18/2004

4376 Small et al.
Sch em e 1. P r ep a r a tion of Ch r om iu m -Ba sed
Macromolecules, Vol. 37, No. 12, 2004
The majority of the polymerizations in this report were
carried out using the chromium(II) analogues of both
Type 1 and Type 2. The chromium(II) derivatives are
air-sensitive, making them difficult to isolate in analyti-
cally pure form. Indeed, on occasion we have noted that
small amounts of oxygen contamination during isolation
of the chromium(II) species results in green products,
presumably due to oxidation to chromium(III). We have,
however, obtained satisfactory elemental analyses for
the chromium(III) complexes, and these results are
shown in Table 6. In any event, the data and discussion
to follow will show that the chromium(II) and chro-
mium(III) complexes possess nearly identical activities,
as well as likely identical active site structures. The
chromium(II) complexes in this report are numbered
sequentially, and the chromium(III) analogues are
identified by the same numbers followed by the “prime”
superscript, e.g., 2′.
P r eca ta lysts
Discu ssion of P olym er iza tion Da ta . When com-
plexes of either Type 1 or Type 2 (Scheme 1) are
activated with modified methylalumoxane (MMAO),15
remarkably active catalystssfrom both the chromium-
(II) and chromium(III) precursorssare generated for the
oligomerization and polymerization of ethylene. Table
1, which includes representative data, illuminates sev-
eral interesting features of these new systems. Analysis
of the data for catalysts of Type 1, shown in entries 1-9,
indicates a sizable effect on catalyst activity with
increasing ligand steric bulk. However, the effect is far
from regular. In fact, it should be noted that the widely
differing activities exhibited by catalysts 1-13 made it
difficult to run all of the reactions under identical
conditions (catalyst amount, pressure, temperature, Al:
Cr ratio). In the paragraphs that follow, we have been
careful to only discuss and directly compare catalysts
run under comparable reaction conditions. For example,
experiments 1-4 all exhibited rapid exotherms from
room temperature to 35 °C (indicating high catalyst
activity) upon catalyst activation under 1.0 bar of
ethylene, but the reaction involving the 2-tert-butyl-
substituted complex 5 (entry 5) did not exceed room
temperature during the course of the reaction. In entry
6, complex 6, bearing four ortho methyl substituents,
caused a large exotherm and polymerized ethylene with
approximately five times higher productivity than 5, as
evidenced by the respective product yields.
were not reported in ref 9, but since even chromium(II)
chloride, when activated with MAO, shows some intrin-
sic ability to polymerize ethylene, we expected that
complexes of type I would display some activity for
ethylene polymerization. Additionally, complexes such
as J , which like I also bear tridentate neutral ligands,
yield relatively high activity polymerization catalysts.10
During the preparation of our manuscript, Esteruelas
et al. reported the synthesis of pyridine bisimine chro-
mium(III) complexes for the polymerization of ethylene.11a
Our studies augment this recent publication by focusing
on previously unreported chromium complexes sup-
ported by substituted and asymmetric tridentate pyri-
dine bisimine and monoimine ligands. The products
made by these chromium complexes vary dramatically
from those reported by Esteruelas, ranging from 1-butene
to heavy waxes and polyethylene. Moreover, our results
conclusively demonstrate that both chromium(II) and
chromium(III) complexes produce highly active catalysts
for the oligomerization and polymerization of ethylene.11b
This observation indicates that large tert-butyl groups
in two ortho positions (the 2-position of each aryl ring)
have a more detrimental effect on activity (presumably
steric in origin) than methyl groups in all four ortho
positions. However, entries 7 and 8, in which the 2-tert-
butyl and 2,5-di-tert-butyl complexes 5 and 7 are
compared under identical reaction pressure and Al:Cr
ratio, clearly show that steric arguments alone cannot
rationalize Type 1 catalyst activity.16 Comparing entries
7 and 8, complex 7 with 2,5-di-tert-butyl substituents
shows a high activity relative to complex 5 (the higher
reaction temperature of 100 °C in experiment 8 is due
to the catalyst exotherm). The relationship between the
increased activity and the presence of an additional tert-
butyl group in the 5-position of the aryl rings is not yet
understood. Increasing ethylene pressure also dramati-
cally increases catalyst activity, as shown by comparing
entries 2 and 9 (catalyst 2). At 27 bar of ethylene,
catalyst activation at room temperature results in a very
rapid exotherm to 80 °C. The temperature was main-
tained in this reaction by internal cooling, and even
Resu lts a n d Discu ssion
Tridentate pyridine bisimine ligands, which were first
reported many years ago in the chemical literature,12
have been much more recently documented as key
components in the preparation of highly active iron- and
cobalt-based catalysts for olefin polymerization.13 The
iron and cobalt precatalysts are easily prepared in high
yields by stirring a suspension of the ligand and the
appropriate metal salt in either hot n-butanol or THF.
Similarly, the analogous chromium precatalysts may be
prepared by simply stirring a suspension of CrCl2 or
CrCl3(THF)3 with the tridentate ligand in hot THF
(Scheme 1). In addition to the bisimine complexes (Type
1) of Scheme 1, tridentate complexes of Type 2, in which
one of the ketone groups is not converted to an imine,
can also be prepared.14
The chromium(II) complexes are blue or violet in
color, whereas the chromium(III) complexes are green.

Macromolecules, Vol. 37, No. 12, 2004
New Chromium Complexes for Ethylene Oligomerization 4377
Ta ble 1. Selected Da ta for th e P olym er iza tion of Eth ylen e a n d r-Olefin s Usin g Ch r om iu m (II) Com p lexes
cat. amt.
(mg)
PC2(bar),
rxn length
(min)
T
Yield
(g)
Prod.d
(g/g Cr complex)
major
GPC data
entry complex
Al:Cra com.(amt.)b
(°C)c
products/notese
(MWpeak × 103)f
1
2
1
2
5.7
6.1
5.4
4.6
5.2
6.7
5.0
5.0
5.0
18.0
21.0
6.3
5.0
4.0
10.0
13.6
3.0
4.0
4.0
500 1.0
500 1.0
500 1.0
500 1.0
500 1.0
500 1.0
500 27
30
60
80
35
35
35
35
25
35
60
100
80
25
25
25
n.d.
n.d.
n.d.
n.d.
1.9
10.4
19.9
57.3
n.d.
6.0
n.d.
n.d.
n.d.
n.d.
2-butene
1-butene
1-butene
1-butene
wax/PE
wax/PE
wax/PE
n.d.
n.d.
n.d.
n.d.
0.72
0.83
0.17
11.0
40.0
34.0
68.2
103
1.54
1.20
0.98
0.85
0.35
0.15
0.16
3
3
4
4
30
5
5
180
120
30
30
30
180
180
60
60
60
60
180
30
30
30
370
6
6
1550
4000
11 500
n.d.
330
60
7
5
8
7
500 27
500 27
PE
9
2
1-butene
10
11
12
13
14
15
16
17
18
19
8
9
10
8
8
9
8
11
12
13
160 1.0
160 1.0
500 1.0
1150 27
1150 27
1150 27
260 1.0, 1-hexene
500 1.0
1000 27
1000 27
PE (Figure 6b)
PE
1.2
1.1
180
PE
80 101
20 200
23 800
1490
1520
n.d.
wax/PE (Figure 6a)
wax/PE
100
45
25
40
60
60
95.0
4.9
22.2
n.d.
70.0
45.0
wax/PE
wax/PE (Figure 6c)
1-butene
linear R-olefins, K ) 0.60
linear R-olefins, K ) 0.65
17 500
11 300
a
b
Molar ratio of Al to Cr in the reaction. Ethylene pressure and amount of comonomer present (entry 16). For entry 16, the 1-hexene
was introduced as a 1:5 v:v mixture in n-heptane. c Temperatures shown are the maximum values reached due to the heat generated by
the reaction. Prod. ) productivity, which was not determined for reactions in which the major product was butene. e All of the reactions
d
produced small to moderate amounts of polyethylene byproducts. The reactions were run in n-heptane or cyclohexane. f GPC data are for
the isolated solid fraction of the product. Bimodal behavior, whether solely due to chain transfer to aluminum or the presence of multiple
catalyst active sites, makes the values for Mw/Mn less useful than the more straightforward peak molecular weights of the key GPC
signals. GPC traces can be found in the Supporting Information. Table 2 and Figure 3 illustrate the effect of changing activator
concentration.
though the amount of product was not accurately
quantified due to its volatility (butene), the degree of
the exotherm indicated that the turn-over frequency
(TOF) was likely approaching 106 mol ethylene/mol cat‚
h (productivity > 50 000 g of product/g of Cr complex).
Gas chromatography showed this butene to be very high
purity 1-butene, at least 99% within the C4 fraction,
with traces of hexene and octene present. In addition
to these olefins, a significant amount of polyethylene
was also produced. Due to the butene volatility, it was
not possible to determine the exact amounts of each
product. However, the polyethylene coproduct was
analyzed by gel-permeation chromatography (see Table
1), and the molecular weight data are discussed in the
following section.
The catalysts in entries 1-8, in addition to possessing
widely varying activities, also produce a highly diver-
gent product slate. Methyl-, ethyl-, and isopropyl-
substituted catalysts (entries 2-4) make predominantly
high-purity 1-butene (>99%, with polyethylene coprod-
uct), but entry 1, using the unsubstituted complex 1,
produces mainly 2-butene, via isomerization of 1-butene.17
The product composition becomes even more curious
upon comparing entries 1-4 to entry 5. While entries
1-4 all result in butene production, the 2-tert-butyl
catalyst 5 makes only waxes and polyethylene. Fur-
thermore, catalyst 6 (entry 6), bearing four ortho methyl
groups, makes waxes and polyethylene as the major
products. The effects of minor substituent changes on
product molecular weight have been well-documented
in tridentate pyridine bisimine iron and cobalt catalysts
as have the substituent effects for R-diimine-based
nickel and palladium systems.13,18 In general, increasing
the steric bulk at the 2- and 6-positions of the imino-
aryl rings results in higher molecular weight, but the
drastic effect on product molecular weight by slight
ligand modifications in these chromium catalysts is
unusual.
ortho substituents of the imino-aryl ring (Note: for
Type 2 complexes, there is only one ring to consider).
Entry 10, employing complex 8 with two methyl groups,
shows a much higher activity than either catalyst 9 with
two isopropyl groups (entry 11) or catalyst 10 with a
single tert-butyl group (entry 12). This trend is also seen
at higher pressures (entries 13 and 14 vs 15) as 8
remains much more active than 9.
In addition to their ability to rapidly oligomerize
ethylene, these new chromium catalysts are capable of
incorporating comonomers. Entry 16, in which dimethyl-
substituted 8 was used, demonstrates the ability of
these systems to form ethylene/R-olefin copolymers. The
branched wax produced in entry 16, run in a 1:5
1-hexene:heptane v:v ratio, indicates substantial comono-
mer incorporation, as demonstrated by examining the
copolymers via gas chromatography. Figure 2 shows a
comparison of the waxes made by catalyst system 8 in
entries 10, 13, and 16. Entry 13, in which elevated
ethylene pressure was used and no comonomer was
employed, shows only very low levels of branching. The
wax from experiment 10 is again mostly linear, with
the slightly increased level of branching due to the use
of lower ethylene pressure (product reincorporation
likely becomes more competitive with ethylene inser-
tion). Entry 16, in which 1-hexene comonomer was
added, shows high levels of branching, thus conclusively
demonstrating the ability to incorporate comonomer.
Worth noting, these GC traces only show the low-
molecular-weight (<1000) portions of the polymer
samples. For entries 10, 13, and 16, the GC traces
represent about 15%, 40%, and 50%, respectively, of the
total polymer formed. GPC data are discussed in the
following section.
Further comparisons between Type 1 and Type 2
catalysts highlight other differences between these
chromium systems. For example, upon comparing entry
2 (Type 1) to entry 10 (Type 2), the primary product
changes from 1-butene to linear wax as illustrated in
Scheme 2. The observation that the Type 2 ketone-
As with Type 1 complexes, catalyst activity for the
Type 2 systems is largely dependent on the size of the

4378 Small et al.
Macromolecules, Vol. 37, No. 12, 2004
F igu r e 2. Branching increases as the ratio of comonomer to ethylene increases: (A) wax produced at 27 bar ethylene (Table 1,
entry 13); (B) wax produced at 1.0 bar ethylene (entry 10); (C) wax produced with 20% 1-hexene comonomer present (entry 16).
See Table 1 for details.
Sch em e 2. Com p a r ison of P r od u cts for Typ e 1 (top )
a n d Typ e 2 Ca ta lysts
the base ligand of the Type 1 complexes. Activation of
this new catalyst under 1.0 bar of ethylene results in
almost exclusive and rapid dimerization of ethylene to
1-butene, a clear indicator of ‘Type 1’ behavior (Table
1, entry 17). This experiment demonstrates that steric
arguments made for the Type 1 systems are not predic-
tive of behavior for the Type 2 catalysts and vice versa.
The propensity of Type 1 catalysts bearing two total
ortho methyl groups to produce butene (entries 2 and
17), along with the tendency of the system with four
ortho methyl groups to make waxes (entry 6), suggested
that an intermediate degree of steric bulk might produce
catalysts for making R-olefins with a more commercially
typical Schulz-Flory constant (0.6-0.8). With that goal
in mind, asymmetric precatalysts 12 and 13 in Scheme
3 were synthesized and used to oligomerize ethylene.
As shown by Scheme 3, catalysts bearing three total
ortho alkyl groups produce linear R-olefins with high
productivity and selectivity and with product molecular
weights that are between the molecular weights made
by the less substituted and more highly substituted
analogues. Entry 18 (complex 12) provides more details
about the oligomerization reaction, and entry 19 con-
tains the results from using the slightly more bulky
ethyl-substituted 13. Note, the products made in entry
19 are slightly heavier (demonstrated by a slightly
larger K value) than those made in entry 18, indicating
that for some Type 1 catalysts rational ligand modifica-
tions can lead to predictable product properties.
containing system produces much higher molecular
weight oligomers than the pyridine bisimine catalyst is
somewhat counterintuitive. Entries 1-6 suggest that
ligand steric bulk plays a key role in determining
product molecular weight. Thus, one might have pre-
dicted that the inherently bulkier Type 1 catalysts
would make higher oligomers than Type 2. However,
the trend is surprisingly reversed, indicating that simple
steric arguments alone cannot be used to rationalize the
relative product distributions of the two catalyst sys-
tems. Close inspection of Scheme 2 shows that complex
2 (Type 1) and complex 8 (Type 2) possess the same total
number of ortho methyl groups, but the Type 2 ketone-
containing catalyst has both methyl groups on the same
ring and makes a much higher molecular weight
product. This observation seems to indicate that the
presence of two ortho substituents on a single aryl ring
results in polymer or heavy oligomer formation, regard-
less of the presence of a second imino-aryl moiety. This
same effect is observed for the isopropyl-substituted
systems (entry 4 vs entry 11) and for catalysts run at
elevated pressure (entry 9 vs entries 13 and 14).
To test whether the location of the ortho methyl
groupssi.e., both on the same aryl ring or one on each
ringswas the primary factor in determining product
molecular weight, a new Type 1 complex with an
asymmetric NNN donor set was synthesized. This
complex 11, shown in Scheme 3, may be viewed as a
hybrid between the two species shown in Scheme 2.
Although this new system possesses the same location
of the ortho methyl groups as the Type 2 catalyst, it has
Discu ssion of P olym er Molecu la r Weigh t Da ta .
Regardless of the product distribution, all of the cata-
lysts in this report, whether derived from the chromium-
(II) or the chromium(III) source, made significant
quantities of polyethylene coproduct. In the cases of
ethylene dimerization catalyzed by the Type 1 systems,
this contrast was stark; for the catalysts that made
higher oligomers, the distinction was less noticeable.
Gel-permeation chromatography (GPC) was used to
analyze the polymers made in experiments 5-19 in
Table 1. Upon examination of the GPC data, many of
the traces showed bimodal or even polymodal behavior.
For pyridine bisimine iron catalysts, it is known that
the MAO:metal ratio impacts product molecular weight,
with higher MAO loadings yielding lower molecular
weight product via chain transfer to aluminum.13d We

Macromolecules, Vol. 37, No. 12, 2004
New Chromium Complexes for Ethylene Oligomerization 4379
Sch em e 3. Ster ic Effects on Eth ylen e Oligom er iza tion for Typ e 1 Ca ta lysts
Ta ble 2. P olym er iza tion Da ta for 7′ w ith Va r yin g Al:Cr
The molecular weights reported in Table 1 are the
peak molecular weights of the largest peak in the GPC
trace. Since all of the catalysts that produce mainly low-
molecular-weight products also produce high-molecular-
weight coproducts, the peak values reported in Table 1
are more useful than the Mn and Mw values for the
purposes of the following discussion. However, the Mn
and Mw values, along with the GPC traces, are included
in the Supporting Information.
Ra tios
Al:Cr mol
ratio
productivity
(g of PE/g of Cr complex)
entry
Mw
Mw/Mn
20
21
22
23
100
200
400
800
730
830
1080
1450
55 000
37 700
32 200
33 700
7.9
9.0
10.1
12.5
a
Polymerizations were performed under the following condi-
tions: 10 µmol of chromium complex in 100 mL of toluene, 40 °C,
1.3 bar, 1 h.
As found for the iron and cobalt pyridine bisimine
catalysts, the molecular weights of polymers produced
by chromium pyridine bisimine compounds are largely
dependent on the steric bulk at the ortho positions of
the aryl rings. The effect of ligand sterics on the product
distribution for the NNN chromium complexes was
discussed earlier, but this effect is also seen for the NNO
catalysts. For example, comparing complexes 8 and 9
(entries 10 and 11), the methyl-substituted catalyst
makes lower molecular weight polyethylene than the
isopropyl-substituted system (peak MW’s 34 000 and
68 200, respectively), and both produce lower molecular
weight polymer than tert-butyl derivative 10 (peak MW
) 103 000). Also, the data clearly indicate that the NNO
catalysts (entries 10-12) make higher molecular weight
polymers than the NNN catalysts. In fact, according to
the GPC numbers, the NNN catalysts produced prima-
rily wax products, with peak molecular weights of
<1000 (entries 5-8). Complex 7 was the exception to
this trend, producing a polymer with a peak MW of
11 000 (entry 8). Due to the peculiar bimodal and
polymodal behavior of these catalysts, more study is
needed to more fully assess these molecular weight
trends.
Cr ysta llogr a p h ic Stu d ies. To better characterize
the molecular structures of the chromium(II) and chro-
mium(III) pyridine bisimine precatalysts, we examined
complexes 5′, 6′, and 6 using X-ray crystallography.
Complex 5′, containing the 2-tert-butyl-substituted ligand,
is of special interest since it displays low activity
compared to other 2-alkyl-substituted complexes, as well
as compared to 2,6-disubstituted derivatives such as 6′.
Furthermore, the structure of chromium(II) complex 6
allows us to directly compare chromium(II) and chro-
mium(III) complexes with the same ligand set. Unfor-
tunately, despite repeated attempts, we were unable to
were curious to see how the polymers made by the
chromium catalysts would be affected by the cocatalyst
level. We carried out a separate study examining the
effect of Al:Cr ratio on catalyst productivity and polymer
molecular weight for the 2,5-di-tert-butyl-substituted
complex 7′. The data are contained in Table 2. For this
study, complex 7′ is especially useful since it produces
predominantly moderate molecular weight polyethylene
that is readily analyzed by GPC. As expected, increasing
the Al:Cr ratio from 100 to 800 results in a gradual
increase in catalyst productivity. It is also clear from
the Mw and Mw/Mn data in Table 2 that lower molecular
weight product is formed at higher Al:Cr ratios. The
GPC traces, shown in Figure 3, are even more instruc-
tive as they distinctly show the dramatic increase in low
molecular weight material (peak centered at about MW
) 1000) as the Al:Cr ratio increases. The GPC traces
are reminiscent of those reported for iron(II)-based
catalysts and are consistent with chain transfer to
aluminum as a likely route for the formation of low-
molecular-weight polymers. In addition, 13C NMR data
(not shown) of the polymers produced in entries 20-23
indicate that substantial quantities of olefinic products
are formed for Al:Cr ratios up to 200 but that increasing
amounts of saturated products are formed at Al:Cr >
200, again consistent with chain transfer to aluminum
as a likely termination mechanism. Changing the Al:
Cr ratio clearly explains one way to generate bimodal
polymers, but polymodal behavior likely indicates that
more than one catalyst active species is present. The
possible nature of these species is discussed later in this
report.

4380 Small et al.
Macromolecules, Vol. 37, No. 12, 2004
F igu r e 3. GPC traces for polymer produced using 7′ with varying Al:Cr ratios.
Ta ble 3. Su m m a r y of X-r a y Cr ysta llogr a p h ic Da ta for Ch r om iu m Com p lexes 5′, 6′, a n d 6a
complex
5′
6′
6
empirical formula
fw
C
32H39.5Cl3CrN4.5
C25H27Cl3CrN3
527.85
C25H27Cl2CrN3
492.40
645.52
cryst syst
space group
a (Å)
b (Å)
c (Å)
triclinic
P1h
monoclinic
P21/c
11.148(5)
15.492(3)
14.393(4)
90
92.59(3)
90
2483.2(14)
4
monoclinic
P21/c
12.626(1)
11.740(2)
16.790(2)
90
98.362(6)
90
2462.3(5)
4
9.834(3)
12.942(3)
15.375(3)
112.99(4)
96.97(3)
105.19(3)
1682.4(7)
2
R (°)
â (°)
γ (°)
V (Å3)
Z
d
calcd (Mg‚m-3
)
1.271
1.412
1.328
cryst size (mm)
abs. coeff. (mm-1
2θ max (°)
0.40 × 0.40 × 0.10
0.35 × 0.32 × 0.06
0.38 × 0.20 × 0.12
)
0.606
0.802
0.698
50.08
1.0-0.5305
6311
5936
3102
377
0.0808 (0.1840)
1.003
1.121, -0.687
49.98
1.0-0.8879
4589
4353
3038
289
0.0397 (0.0985)
1.027
0.366, -0.270
50.08
1.0-0.9128
4554
4345
2678
280
0.0792 (0.1447)
1.029
0.343, -0.331
transmission range
no. of reflns collected
no. of indep reflns
no. of obsd reflections
no. of variables
R1 (wR2)b [I > 2σ(I)]
goodness-of-fit (F2)
diff. peaks (e-‚Å-3
)
a
b
See Experimental Section for additional data collection, reduction, and structure solution and refinement details. R1 ) Σ|Fo| -
|Fc|/Σ|Fo|; wR2 ) [Σ[w(Fo - Fc2)2]]1/2, where w ) 1/σ2(Fo2) + (aP)2 + b.
2
obtain suitable single crystals of Type 2 (NNO) com-
plexes. A summary of crystal data collection parameters
for 5′, 6′, and 6 is given in Table 3.
ported separately by Gambarotta19a and Esteruelas.11a
Similarities include the fact that the Cr-N(pyridyl)
distances are shorter (0.13-0.16 Å) than the Cr-
N(imino) bonds and the equatorial Cr-Cl bond (opposite
the pyridyl nitrogen) is the shortest of the three Cr-Cl
distances. Also, as expected, the phenyl ring planes in
6′ and 5′ are oriented nearly perpendicular to the plane
of the pyridine bisimine ligand (dihedral angles ranging
from 67.5° to 94.6°) to minimize steric interactions
between the aryl substituents and the remainder of the
complex.
Structures of {2,6-Bis[1-((2,6-dimethylphenyl)imino)-
ethyl]pyridine}CrCl3 (6′) and {2,6-Bis[1-((2-tert-
butylphenyl)imino)ethyl]pyridine}CrCl3‚1.5CH3CN (5′).
Crystals of 6′ and 5′ suitable for X-ray structure
determination were obtained by slow evaporation of
acetonitrile solutions of the respective compounds. The
molecular structures are shown in Figure 4, and selected
distances and angles are given in Table 4. The struc-
tures of 6′ and 5′ are similar to that of {2,6-bis[1-((2,6-
diisopropylphenyl)imino)ethyl]pyridine}CrCl3 (14′), re-
In addition, 6′, like 14′, possesses one long axial Cr-
Cl bond. For 6′, the long Cr-Cl bond (Cr-Cl(3) )

Macromolecules, Vol. 37, No. 12, 2004
New Chromium Complexes for Ethylene Oligomerization 4381
F igu r e 5. Thermal ellipsoid representation (35% probability
boundaries) of the X-ray crystal structure of 6. Hydrogen atoms
are omitted for clarity.
ents that are directed toward Cl(1), it simultaneously
increases the steric interactions between the remaining
two methyl substituents and Cl(3), lengthening the Cr-
Cl(3) bond. The structure of 14′ reported by Esteruelas
also displays an acute Cl-Cr-N(pyridyl) angle (82.27(8)°)
with attendant tilting of the pyridine bisimine ligand
toward the long Cr-Cl bond.11a While this distortion is
notable, at this time it is unclear whether the distortion
is retained in solution or whether it persists in the
activated catalyst.
In contrast to 6′ and 14′, complex 5′ displays nearly
equivalent N(pyridyl)-Cr-Cl angles (N(1)-Cr-Cl(1) )
87.16(16)° and N(1)-Cr-Cl(3) ) 85.38(16)°) and the
axial Cr-Cl bond lengths are also similar (2.308(3) Å
and 2.319(3) Å). Apparently, complex 5′ accommodates
the 2-tert-butyl substituents without significant tilting
of the pyridine bisimine ligand, in part by positioning
F igu r e 4. Thermal ellipsoid representation (35% probability
boundaries) of the X-ray crystal structures of 5′ (top) and 6′
(bottom). Hydrogen atoms are omitted for clarity.
Ta ble 4. Selected Dista n ces (Å) a n d An gles (d eg) for
Str u ctu r a lly Ch a r a cter ized Com p lexesa
complex 6′
Cr(1)-N(1)
1.993(3)
2.123(3)
2.140(3)
2.2996(13) N(3)-C(16)
75.75(11)
77.24(11)
Cr(1)-Cl(2)
Cr(1)-Cl(3)
N(2)-C(6)
2.2695(10)
the tert-butyl groups in an anti arrangement. We should
note that the anti arrangement appears to provide a
reasonably accessible metal center. Thus, the reasons
for the low activity of 5′ relative to other 2-substituted
derivatives are not readily apparent from the solid-state
structure.
Cr(1)-N(2)
2.3513(13)
1.291(4)
1.297(4)
98.57(7)
87.69(8)
Cr(1)-N(3)
Cr(1)-Cl(1)
N(1)-Cr(1)-N(2)
N(1)-Cr(1)-N(3)
N(2)-Cr(1)-N(3) 153.36(10)
N(1)-Cr(1)-Cl(1) 94.45(8)
N(1)-Cr(1)-Cl(2) 173.25(8)
N(1)-Cr(1)-Cl(3)
N(2)-Cr(1)-Cl(1)
N(2)-Cr(1)-Cl(2) 103.41(8)
N(2)-Cr(1)-Cl(3)
N(3)-Cr(1)-Cl(1)
N(3)-Cr(1)-Cl(2) 103.10(8)
N(3)-Cr(1)-Cl(3)
Cl(1)-Cr(1)-Cl(2)
Cl(1)-Cr(1)-Cl(3) 175.78(4)
Cl(2)-Cr(1)-Cl(3)
90.50(8)
92.30(4)
Structure of {2,6-Bis[1-((2,6-dimethylphenyl)imino)-
ethyl]pyridine}CrCl2 (6). Slow cooling of a saturated
CH2Cl2 solution deposited single crystals of 6. The
structure is shown in Figure 5, and selected distances
and angles are included in Table 4. The pseudo-square-
pyramidal structure of 6 is reminiscent of certain
5-coordinate iron(II) and cobalt(II) pyridine bisimine
derivatives reported by Gibson and Brookhart.13,14 For
6, the basal plane consists of the three nitrogen atoms
from the pyridine bisimine ligand and one equatorial
chloride, with the remaining chloride occupying the
apical position. In this geometry, the chromium atom
is raised 0.40 Å above the basal plane. In the iron(II)
and cobalt(II) pyridine bisimine derivatives, the metal-
Cl(apical) bonds are longer (0.04-0.06 Å) than the
corresponding equatorial M-Cl distances. The same
trend is observed for 6, with the apical Cr-Cl(2) bond
(2.406(2) Å) being 0.08 Å longer than its equatorial
counterpart (2.3215(18) Å). As expected, given the larger
atomic radius of chromium(II), the Cr-Cl bonds in 6
are longer than the average Cr-Cl distances in the
chromium(III) derivatives 6′ and 5′. However, the Cr-
N(pyridyl) and Cr-N(imino) distances in 6 are similar
to, if not slightly shorter than, those in 6′ and 5′. The
near equivalence of the Cr-N distances in the two
oxidation states is presumably due to the lower coordi-
nation number in 6 relative to 6′ and 5′. Finally, the
C-N(imino) distances in 6 (1.286(7) and 1.294(7) Å)
indicate that the imino groups have retained their
double-bond character and that the ligand binds as a
81.43(8)
88.68(8)
91.82(4)
complex 5′
Cr(1)-N(1)
1.981(5)
2.141(6)
2.127(6)
2.308(3)
77.9(2)
Cr(1)-Cl(2)
2.270(2)
2.319(3)
1.284(8)
1.288(8)
86.32(15)
87.04(15)
Cr(1)-N(2)
Cr(1)-Cl(3)
Cr(1)-N(3)
Cr(1)-Cl(1)
N(1)-Cr(1)-N(2)
N(1)-Cr(1)-N(3)
N(2)-Cr(1)-N(3) 154.81(19)
N(1)-Cr(1)-Cl(1) 87.16(16)
N(1)-Cr(1)-Cl(2) 178.28(17)
N(1)-Cr(1)-Cl(3)
N(2)-Cr(1)-Cl(1)
N(2)-Cr(1)-Cl(2) 102.33(14)
N(2)-C(7)
N(3)-C(19)
N(2)-Cr(1)-Cl(3)
N(3)-Cr(1)-Cl(1)
N(3)-Cr(1)-Cl(2) 102.85(15)
N(3)-Cr(1)-Cl(3)
Cl(1)-Cr(1)-Cl(2)
Cl(1)-Cr(1)-Cl(3) 172.54(8)
Cl(2)-Cr(1)-Cl(3)
76.9(2)
91.57(16)
94.54(9)
85.38(16)
91.82(15)
92.92(9)
complex 6
Cr(1)-N(1)
1.993(5)
2.128(5)
2.107(5)
2.3215(18)
76.11(19)
75.6(2)
Cr(1)-Cl(2)
N(2)-C(6)
N(3)-C(17)
2.406(2)
1.286(7)
1.294(7)
Cr(1)-N(2)
Cr(1)-N(3)
Cr(1)-Cl(1)
N(1)-Cr(1)-N(2)
N(1)-Cr(1)-N(3)
N(2)-Cr(1)-N(3) 148.5(2)
N(1)-Cr(1)-Cl(1) 155.92(15)
N(1)-Cr(1)-Cl(2) 104.12(15)
N(2)-Cr(1)-Cl(1) 102.14(14)
N(2)-Cr(1)-Cl(2)
N(3)-Cr(1)-Cl(1)
N(3)-Cr(1)-Cl(2) 101.99(15)
Cl(1)-Cr(1)-Cl(2) 99.92(7)
98.03(14)
98.09(15)
a
Estimated standard deviations are indicated in parentheses.
2.3513(13) Å) coincides with a subtle canting of the
pyridine bisimine ligand toward Cl(3), which manifests
itself in an acute N(1)-Cr-Cl(3) angle of 81.43(8)°.
While this canting accommodates the methyl substitu-

4382 Small et al.
Macromolecules, Vol. 37, No. 12, 2004
neutral pyridine bisimine to the chromium(II) centers
there is no structural evidence for significant electron
delocalization from the reduced metal center to the
ligand.
Na tu r e of th e Ca ta lyst Active Sp ecies. The ability
to draw a loose connection between the ligand sterics
and the product molecular weights within each catalyst
family, along with the inability to extend the steric
connection between the two catalyst types, suggests that
the Type 1 (NNN) and Type 2 (NNO) catalysts differ
from each other in charge, chain growth mechanism,
oxidation state, and/or some other feature. An examina-
tion of the literature shows that the tridentate ligand/
metal combination can lead to wide variations in the
electronics of the formed complexes; in that regard, it
is not too surprising that the Type 1 catalyst sterics are
not predictive for the Type 2 systems and vice versa.
A number of recent studies on pyridine bisimine
complexes have reported a wide variety of organome-
tallic species, covering a wide range of oxidations states,
responses to alkylating agents, and catalyst initiation
mechanisms. For example, in the case of pyridine
bisimine vanadium systems, ligand alkylation by the
cocatalyst causes the formation of an anionic ligand.20
Moving to the later transition metals, a neutral triden-
tate manganese(I) complex and an anionic manganese-
(II) complex have been isolated and characterized.19b
Neither of these species nor the manganese(II) dichlo-
ride precursor are initiators or precatalysts for olefin
polymerization. For the tridentate iron analogues, re-
cent studies using trimethylaluminum as the cocatalyst
for ethylene polymerization reactions led the authors
to propose a neutral iron(II) active species.21 For the
tridentate cobalt catalysts, two very recent studies
suggest that polymerization might be initiated by a
cationic cobalt(I) complex, which may undergo propaga-
tion via a cobalt(III) species.22 With ruthenium, a
cationic ruthenium(II) alkyl olefin complex has been
isolated, but this species is not active for polymeriza-
tion.23 For rhodium, yet another type of complex, a
dicationic rhodium(III) alkyl olefin species, has been
identified. This rhodium complex is also not active for
ethylene polymerization.23 None of these basic struc-
tures may at this point be ruled out for these new
chromium catalysts. Furthermore, we should note that
the ketone group in Type 2 systems may be especially
susceptible to attack by aluminum alkyls, perhaps
generating an in situ alkoxide or Lewis acid-base
adduct.
F igu r e 6. UV-vis spectra of 6′ (b), 6′ + 200 equiv of MAO
(O), 6 (2), 6 + 200 equiv of MAO (4). Samples are 5 mM in
CH2Cl2. Spectra of samples with added MAO were acquired
within 15 min after MAO addition.
similar as are the polymers made by 14 and 14′. This
evidence further suggests the identical nature of the
chromium(II) and chromium(III) active species.
In conclusion, a highly active family of chromium
catalysts for the dimerization and oligomerization of
ethylene has been introduced. Although these are not
the first chromium systems known to dimerize ethyl-
ene,26 the high purity of the 1-butene and the high
activity of the catalysts are remarkable. In addition to
the butene production, the catalysts also make highly
linear R-olefin waxes and readily incorporate comono-
mer to make branched waxes. As shown in Table 1,
these catalysts are quite robust, exhibiting high activity
up to at least 100 °C. Finally, the ligand substituent
effects on the product composition are profound and for
the most part predictable within each catalyst type
(NNN and NNO complexes). Further studies will focus
on optimizing the oligomerization reaction conditions,
gaining a better understanding of the relationship
between catalyst sterics and activity, studying the
copolymerization reactions in more detail, and gaining
more information about the nature of the catalytically
active species.
Exp er im en ta l Section
Ma ter ia ls a n d Meth od s. Anhydrous tetrahydrofuran,
toluene, heptane, and cyclohexane were purchased from Ald-
rich and stored over molecular sieves. Acetonitrile was distilled
from calcium hydride and stored over molecular sieves. 1-Hex-
ene was obtained as a commercial grade of Chevron Phillips’
Normal Alpha Olefins (NAO’s), degassed, and dried over 4A
molecular sieves. MMAO-3A (Al 7% by weight) was purchased
from Akzo Nobel. MAO (10 wt %) in toluene was provided by
Albemarle. 2,6-Diacetylpyridine, chromium(II) chloride, tris-
(tetrahydrofuran)chromium(III) chloride, and all substituted
anilines were purchased from Aldrich and used without further
purification. NMR spectra were obtained using a Varian Unity
Plus 300 MHz Spectrometer. Elemental analyses were per-
formed by Atlantic Microlabs of Norcross, GA. Gel-permeation
chromatography (GPC) data for Figure 3 were obtained on a
Waters 150 °C GPC operating at 140 °C using 1,2,4-trichlo-
robenzene as solvent. GPC data for Table 1 were obtained with
a PL 210 SEC high-temperature chromatography unit (Poly-
mer Laboratories) equipped with three PLgel 10 µm Mixed A
columns (Polymer Laboratories). A refractive index detector
Despite difficulties in preparing analytically pure
chromium(II) complexes, the chromium(II) and chro-
mium(III) catalysts exhibit equal activities and make
the same products upon activation. These results cer-
tainly imply that identical active species are generated
in the presence of MAO, but more evidence was sought
to justify this assertion.24 A UV-vis study25 was chosen,
in which the chromium(II) and chromium(III) Type 1
complexes 6 and 6′ were analyzed before and after the
addition of MAO (10 wt % in toluene). As Figure 6
indicates, the addition of MAO generates the same UV
spectrum for each complex, a result that agrees with
their nearly identical catalytic activities. To further
support the argument of identical catalyst active spe-
cies, GPC data was obtained for the polyethylenes made
by the activated complexes 6, 6′, 14, and 14′ under
identical polymerization conditions. As shown in Table
5 and Figure 7, the polymers made by 6 and 6′ are very

Macromolecules, Vol. 37, No. 12, 2004
New Chromium Complexes for Ethylene Oligomerization 4383
Ta ble 5. Com p a r a tive P olym er iza tion Da ta for Ch r om iu m (II) a n d Ch r om iu m (III) Com p lexesa
productivity
(g of PE/g of Cr complex)
entry
catalyst systemb
substituent Rn
substituent R′n
Mw
Mw/Mn
24
25
26
27
6, MAO
6′, MAO
14, MAO
14′, MAO
2,6-Me2
2,6-Me2
2,6-iPr2
2,6-iPr2
2,6-Me2
2,6-Me2
2,6-iPr2
2,6-iPr2
410
430
580
420
1300
1200
21 600
17 300
1.6
1.3
4.2
5.2
a
Polymerizations were performed under the following conditions: 10 µmol of chromium complex in 100 mL of toluene, Al:Cr ) 200, 40
b
°C, 1.3 bar, 1 h. Commercial MAO from Albemarle (10 wt % MAO in toluene) was used in these studies.
F igu r e 7. Gel permeation chromatography traces of polymer produced from chromium(II) and chromium(III) complexes under
low-pressure polymerization conditions. The chromium complexes used are identified as follows: 6 (upper left), 6′ (upper right),
14 (lower left), 14′ (lower right). Polymerization conditions: 40 °C, 1.3 bar, Al:Cr ) 200, 10 µmol of compound in 100 mL of
toluene, 1 h.
was used for molecular weight determinations. Samples (1 mg/
mL) were chromatographed at 1 mL/min using 1,2,4-trichlo-
robenzene as the mobile phase. Columns were calibrated using
a broad molecular weight, PE standard. 13C NMR measure-
ments on polymer samples were conducted at 120 °C in 1,1,2,2-
tetrachloroethane using a Bruker ACE 300 spectrometer.
Samples for IR were dispersed in KBr, and spectra were
recorded on a Nicolet 5DXC spectrometer with a diffuse
reflectance (DRIFTS) attachment. A Hewlett-Packard 6890
Series GC System with an HP-5 50m column with a 0.2 mm
i.d., an initial oven temperature of 35 °C, and a ramp rate of
2.4 °C/min. was used for analysis of the light oligomers. An
HP 5890 GC System with a Chrompak WCOT Ulti-Metal 10m
column with a 0.5 mm i.d., an initial oven temperature of 100
°C, a ramp rate of 8.0 °C/min, and a final temperature of 400
°C was used for analysis of the waxes. Helium was the carrier
gas for both instruments, and FID detection was used. Chem-
Station from Agilent Technologies was used to analyze the
collected data. MALDI mass spectrometry data were obtained
using a BiFlexIII MALDI mass spectrometer. ESI mass
spectrometry data were obtained using Micromass AutoSpec
Ultima Magnetic sector mass spectrometer.
diacetylpyridine and 2.3 mL (18.7 mmol) of 2,6-dimethyl
aniline were stirred together in flask with 20 mL of
a
anhydrous methanol. Several drops of glacial acetic acid were
added, and the reaction was heated with stirring for 3 days
at 55 °C. The reaction flask was then placed in a freezer at
-20 °C, resulting in the formation of yellow, needlelike
crystals. These crystals were removed by filtration and washed
1
with cold methanol (yield ) 1.15 g, 24%). H NMR (CDCl3) δ
2.12 (s, 6H, aryl CH3), 2.26 (s, 3H, acetyl CH3), 2.86 (s, 3H,
imino CH3), 7.01 (t, 1H, aryl CH), 7.17 (d, 2H, aryl CH), 8.00
(dd, 1H, pyr. CH), 8.18 (d, 1H, pyr. CH), 8.64 (d, 1H, pyr. CH).
6-[1-{(2,4,6-Tr im e t h ylp h e n yl)im in o}e t h yl]-2-a ce t yl-
p yr id in e (Typ e 2; 2,4,6-Me3, Mes NNO). This compound
may be prepared according to the method reported in the
literature.14e
2-[1-{(2,4,6-Tr im eth ylp h en yl)im in o}eth yl]-6-[1-{4-ter t-
b u t ylp h en yl)im in o}et h yl]p yr id in e (Typ e 1; 2,4,6-Me3;
4-tBu ). A 5.00 g (18 mmol) amount of the monoimine (Mes
NNO) was dissolved with 4.3 mL (27 mmol) of 4-tert-butyl-
aniline in ethanol (50 mL). Five drops of acetic acid were
added, and the reaction was heated with stirring at 60 °C for
18 h. The heating was discontinued, and the reaction was
allowed to sit at room temperature. After 2 days, 4.44 g (61%
yield) of needlelike crystals was isolated, washed with ethanol,
Syn th esis of Liga n d s. Syntheses of the tridentate ligands
used to prepare complexes of both Type 1 and Type 2 have
been previously reported.13,14 Modest adaptations of those
methods were used to prepare the ligands used herein. For
new ligands, complete synthetic and characterization details
are given. For previously reported ligands, only significant
synthetic modifications are presented.
1
and identified by H NMR as the pure desired product. This
compound has been previously reported.14e
2-[1-{(2,4,6-Tr im e t h ylp h e n yl)im in o }e t h yl]-6-[1-{2-
m eth ylp h en yl)im in o}eth yl]p yr id in e (Typ e 1; 2,4,6-Me3;
2-Me). A 1.0 g (3.6 mmol) amount of the monoimine (Mes
NNO) was dissolved with 0.57 g (5.3 mmol) of o-toluidine in
methanol (10 mL). Three drops of acetic acid were added, and
6-[1-{(2,6-Dim et h ylp h en yl)im in o}et h yl]-2-a cet ylp yr i-
d in e (Typ e 2; 2,6-Me2). A 3.0 g (18.4 mmol) amount of 2,6-

4384 Small et al.
Macromolecules, Vol. 37, No. 12, 2004
the reaction was heated with stirring at 60 °C for 16 h. The
reaction was allowed to cool, then was placed in a freezer at 0
°C, resulting in the formation and collection of 262 mg (20%)
of needlelike crystals that were identified as pure by NMR.
1H NMR (CDCl3) δ 1.87 (s, 6H, mes. o-CH3), 2.11 (s, 3H, o-tol.
CH3), 2.22 (s, 3H, mes. p-CH3), 2.28 (s, 3H, imino CH3), 2.34
(s, 3H, imino CH3), 6.63 (d, 1H, tol. ring), 6.84 (s, 2H, mes.
ring), 7.02 (m, 1H, tol. ring), 7.20 (m, 2H, tol. ring), 7.81 (dd,
1H, py. Hp), 8.41 (d, 1H, py. Hm), 8.43 (d, 1H, py. Hm). 13C NMR
(CDCl3) δ 16.0 (2), 17.5 (3), 20.5 (1), 118.1 (1), 122.3 (2), 123.4
(1), 124.9 (2), 126.2 (1), 127.0 (1), 128.4 (2), 130.3 (1), 132.1
(1), 136.6 (1), 145.8 (1), 150.1 (1), 155.2 (1), 155.4 (1), 166.7
(1), 167.8 (1); 25 chemically different carbons, but several show
chemical shift equivalence.
2-[1-{(2,4,6-Tr im e t h ylp h e n yl)im in o }e t h yl]-6-[1-{2-
et h ylp h en yl)im in o}et h yl]p yr id in e (Typ e 1; 2,4,6-Me3;
2-Et). A 1.0 g (3.6 mmol) amount of the monoimine (Mes NNO)
was dissolved with 1.30 g (10.7 mmol) of o-toluidine in
methanol (10 mL). Three drops of acetic acid were added, and
the reaction was heated with stirring at 60 °C for 16 h. The
reaction was allowed to cool, then was placed in a freezer at 0
°C, resulting in the formation and collection of 247 mg (20%)
of needlelike crystals that were identified as pure by NMR.
1H NMR (CDCl3) δ 1.20 (t, 3H, CH3 on ethyl), 1.98 (s, 6H, mes.
o-CH3), 2.23 (s, 3H), 2.29 (s, 3H), 2.34 (q, 2H, CH2 on ethyl),
2.35 (s, 3H), 6.62 (d, 1H, tol. ring), 6.87 (s, 2H, mes. ring), 7.07
(t, 1H, tol. ring), 7.24 (m, 2H, tol. ring), 7.88 (t, 1H, py. Hp),
8.41 (d, 1H, py. Hm), 8.43 (d, 1H, py. Hm). 13C NMR (CDCl3) δ
14.5 (1), 22.8 (1), 27.1 (1), 29.2 (1), 29.7 (2), 32.0 (1), 117.8 (1),
121.9 (1), 122.1 (1), 123.7 (1), 125.1 (2), 126.3 (1), 128.3 (2),
128.6 (1), 132.1 (1), 133.5 (1), 136.8 (1), 146.4 (1), 149.2 (1),
155.2 (1), 155.4 (1), 166.3 (1), 167.8 (1); 26 chemically different
carbons, but several show chemical shift equivalence.
Syn th esis of Ch r om iu m P r eca ta lyst Com p lexes. Chro-
mium complexes of both Type 1 and Type 2 were prepared
using slight modifications of the reported procedure for
preparation of analogous iron and cobalt complexes.13 In
general, a slight excess of the ligand (∼1.1equiv) was stirred
with the chromium source in hot THF for 1 day, followed by
precipitation with pentane, filtration under inert atmosphere,
and washing with ether and pentane. Yields of about 70% or
higher could be obtained for all complexes in this manner (see
Table 6). However, elemental analyses for the chromium(II)
precursors indicated substantial deviation from the calculated
values for many of the complexes. Bulkier ligands typically
resulted in poorer CHN analysis values (3% or more low for
carbon). Satisfactory elemental analyses for the chromium(III)
complexes are reported in Table 6.
Oligom er ization of Eth ylen e. In an oxygen- and moisture-
free environment, the chromium precatalyst, solvent, comono-
mer (if any), and a stir bar were added to a flask. The flask
was transferred to a Schlenk manifold and placed under a
continuous ethylene purge. The flask contents were stirred
rapidly for several minutes to saturate the solvent with
ethylene and to break apart any small chunks of precatalyst.
MMAO-3A was added via syringe. Optionally, a cooling bath
was used to maintain the desired reaction temperature. For
reactions in which light olefins were the primary product, the
ethylene was continuously purged through the reaction flask.
For the wax-producing reactions, the ethylene was fed “on
demand”.
For reactions at elevated pressure, a 500 mL Zipperclave
reactor from Autoclave Engineers was used. The catalyst was
dissolved in a small amount of dichloromethane in an NMR
tube, which was then sealed and bound to the stirrer shaft of
the clean, dry reactor. The reactor was evacuated, then
charged with the solvent and the alumoxane cocatalyst. The
desired pressure of ethylene was introduced, and the reactor
stirrer was started, resulting in breakage of the NMR tube
and catalyst activation. Ethylene was fed “on demand” using
a
TESCOM regulator, and the reactor temperature was
maintained by internal cooling. An external heating mantle
was used for reactions that were not sufficiently exothermic
to reach the desired temperature. All of the reactions were

Macromolecules, Vol. 37, No. 12, 2004
New Chromium Complexes for Ethylene Oligomerization 4385
C.; Dow, A. W.; J ohnson, R. N.; Carrick, W. L. J . Polym. Sci.,
Polym. Chem. Ed. 1972, 10, 2621-2637. (d) Karol, F. J .;
Brown, G. L.; Davison, J . M. J . Polym. Sci., Polym. Chem.
Ed. 1973, 11, 413-424.
run at least twice or until comparable results could be
obtained.
The studies examining the impact of Al:Cr ratio on product
properties (Table 2 and Figure 4) and the comparative reactiv-
ity between chromium(II) and chromium(III) complexes (Table
5 and Figure 7) were carried out in toluene and used com-
mercial MAO (10% in toluene) provided by Albemarle.
P r od u ct An a lysis. The aluminum cocatalysts were re-
moved by the addition of acidified methanol and subsequent
decantation and/or filtration of the products. After removal of
the cocatalysts, the products were analyzed by gas chroma-
tography (GC) using tridecane as internal standard.
(3) In addition to the examples of chromium polymerization
catalysts shown in Figure 1 in this report, the growing
number of chromium systems that are selective for producing
1-hexene must also be noted. Although such a discussion is
beyond the scope of this work, a few noteworthy references
are hereby included: (a) McGuinness, D. S.; Wasserscheid,
P.; Keim, W.; Morgan, D.; Dixon, J . T.; Bollmann, A.;
Maumela, H.; Hess, F.; Englert, U. J . Am. Chem. Soc. 2003,
125, 5272. (b) McGuinness, D. S.; Wasserscheid, P.; Keim,
W.; Hu, C.; Englert, U.; Dixon, J . T.; Grove, C. Chem.
Commun. 2003, 334. (c) Carter, A.; Cohen, S. A.; Cooley, N.
A.; Murphy, A.; Scutt, J .; Wass, D. F. Chem. Commun. 2002,
858. (d) Reagan, W. K.; Pettijohn, T. M.; Freeman, J . W.
(Phillips Petroleum) U.S. Patent 5523507, 1996. (e) Wu, F.-
J . (Amoco) U.S. Patent 5811618, 1998. (f) Briggs, J . R. (Union
Carbide) U.S. Patent 4668838, 1987.
(4) (a) Leeasubcharoen, S.; Lam, K.-C.; Concolino, T. E.; Rhei-
ngold, A. L.; Theopold, K. H. Organometallics 2001, 20, 182.
(b) Heintz, R. A.; Leeasubcharoen, S.; Liable-Sands, L. M.;
Rheingold, A. L.; Theopold, K. H. Organometallics 1998, 17,
5477. (c) Theopold, K. H. Eur. J . Inorg. Chem. 1998, 15. (d)
Theopold, K. H. Chemtech 1997, October, 26. (e) White, P.
A.; Calabrese, J .; Theopold, K. H. Organometallics 1996, 15,
5473. (f) Liang, Y.; Yap, G. P. A.; Rheingold, A. L.; Theopold,
K. H. Organometallics 1996, 15, 5284.
X-r a y Cr ysta llogr a p h y. Single-crystal specimens selected
for crystallographic characterization were mounted on thin
glass fibers and transferred to an Enraf-Nonius CAD4 (6′, 6)
or a Bruker-Nonius MACH3S (5′) diffractometer equipped with
graphite-monochromated Mo KR radiation (λ ) 0.71073 Å) for
data collections at ambient temperature. Unit cell constants
were determined from a least-squares refinement of the setting
angles of 25 machine-centered reflections located using an
automated search routine. Intensity data were collected using
the ω/2θ scan technique to a maximum 2θ value of 50°. When
necessary, absorption corrections were applied using azimuthal
scans of several intense reflections. The data were corrected
for Lorentz and polarization effects and converted to structure
factors using the teXsan for Windows crystallographic software
package.27 Space groups were determined based on systematic
absences and intensity statistics. Structures were solved by
direct methods or by the Patterson method, which provided
the positions of most of the non-hydrogen atoms. The remain-
ing non-hydrogen atoms were located after several cycles of
structure expansion and full-matrix least-squares refinement
(on F2). In addition to the chromium complex, two CH3CN
solvates were identified in the lattice of 5′. Hydrogen atoms
were added geometrically. All non-hydrogen atoms were
refined using anisotropic displacement parameters, while
hydrogen atoms were refined as riding atoms with group
isotropic displacement parameters. Structure solution and
refinement was performed using the SHELXTL suite of
programs running on a Pentium PC.28
(5) Baralt, E. J .; Carney, M. J .; Cole, J . B. (Chevron Chemical)
U.S. Patent 5780698, 1998.
(6) (a) Kim, W.-K.; Fevola, M. J .; Liable-Sands, L. M.; Rheingold,
A. L.; Theopold, K. H. Organometallics 1998, 17, 4541. (b)
Gibson, V. C.; Maddox, P. J .; Newton, C.; Redshaw, C.; Solan,
G. A.; White, A. J . P.; Williams, D. J . Chem. Commun. 1998,
1651. (c) Gibson, V. C.; Newton, C.; Redshaw, C.; Solan, G.
A.; White, A. J . P.; Williams, D. J . J . Chem. Soc., Dalton
Trans. 1999, 827. (d) Coles, M. P.; Dalby, C. I.; Gibson, V.
C.; Little, I. R.; Marshall, E. L.; Ribeiro da Costa, M. H.;
Mastroianni, S. J . Organomet. Chem. 1999, 591, 78. (e)
J ensen, V. R.; Børve, K. J . Organometallics 2001, 20, 616.
(7) Matsunaga, P. T. (Exxon Chemical) WO9957159, 1999.
(8) (a) J ensen, V. R.; Angermund, K.; J olly, P. W. Organometal-
lics 2000, 19, 403. (b) Do¨hring, A.; Go¨hre, J .; J olly, P. W.;
Kryger, B.; Rust, J .; Verhovnik, G. P. J . Organometallics
2000, 19, 388. (c) Emrich, R.; Heinemann, O.; J olly, P. W.;
Kru¨ger, C.; Verhovnik, G. P. J . Organometallics 1997, 16,
1511.
(9) Devore, D.; Feng, S. S.; Frazier, K. A.; Patton, J . T. (Dow
Chemical) WO0069923, 2000.
(10) Ko¨hn, R. D.; Haufe, M.; Mihan, S.; Lilge, D. Chem. Commun.
2000, 1927.
(11) (a) Esteruelas, M. A.; Lo´pez, A. M.; Me´ndez, L.; Oliva´n, M.;
On˜ate, E. Organometallics 2003, 22, 395. (b) Small, B. L.;
Marcucci, A. J . (Chevron Phillips Chemical) WO03054038,
2003.
(12) Alyea, E. C.; Merrell, P. H. Synth. React. Inorg. Metal-Org.
Chem. 1974, 4(6), 535.
(13) (a) Ittel, S. D.; J ohnson, L. K.; Brookhart, M. Organometallics
2000, 19, 1169. (b) Small, B. L.; Brookhart, M. Polym. Prepr.,
Am. Chem. Soc. Div. Polym. Chem. 1998, 39, 213. (c)
Britovsek, G. J . P.; Gibson, V. C.; Kimberley, B. S.; Maddox,
P. J .; McTavish, S. J .; Solan, G. A.; White, A. J . P.; Williams,
D. J . Chem. Commun. 1998, 849. (d) Small, B. L.; Brookhart,
M.; Bennett, A. M. A. J . Am. Chem. Soc. 1998, 120, 4049. (e)
Small, B. L.; Brookhart, M. J . Am. Chem. Soc. 1998, 120,
7143. (f) 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. (g)
Bennett, A. M. A. Chemtech 1999, 29 (7), 24. (h) 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, No. 12, 2221.
Ack n ow led gm en t. B.L.S. acknowledges Ray Rios,
Eric Fernandez, and A. J . Marcucci for performing many
of the batch reactor studies. M.J .C. acknowledges the
donors of the Petroleum Research Fund, administered
by the American Chemical Society (Grant 37885-GB3),
Research Corporation (Cottrell College Science Award),
and the University of WisconsinsEau Claire for finan-
cial support of this research. J .A.H. acknowledges the
support of the National Science Foundation (CHE-
0078746) and the University of WisconsinsEau Claire.
Instrumentation for infrared analyses was obtained
with the support of the National Science Foundation
MRI (CHE-0216058). We also thank Nick Robertson for
experimental assistance, Albemarle for kindly providing
MAO, W. R. Grace (Davison Catalyst Division) for
polymer GPC and 13C NMR analyses, and Paul De-
sLauriers (CPChem) for GPC analyses.
Su p p or tin g In for m a tion Ava ila ble: Complete X-ray
crystallographic information in CIF format; GPC traces for the
polymers reported in Table 1. This material is available free
of charge via the Internet at http://pubs.acs.org.
Refer en ces a n d Notes
(1) (a) Britovsek, G. J . P.; Gibson, V. C.; Wass, D. F. Angew.
Chem., Int. Ed. Engl. 1999, 38, 428 and references therein.
(b) Gibson, V. C.; Spitzmesser, S. K. Chem. Rev. 2003, 103,
283-316.
(2) Supported chromium oxide catalysts: (a) Hogan, J . P. J .
Polym. Sci., Polym. Chem. Ed. 1970, 8, 2637-2652. (b)
McDaniel, M. P. Adv. Catal. 1985, 33, 47-96. Supported
chromocene catalysts: (c) Karol, F. J .; Karapinka, G. L.; Wu,
(14) Syntheses for the Type 2 unsymmetrical ligands, as well as
their iron and cobalt transition-metal complexes, can be found
in the following references: (a) Bennett, A. M. A. (DuPont)
U.S. Patent 5955555, 1999. (b) Small, B. L.; Brookhart, M.
Macromolecules 1999, 32, 2120. (c) Gibson, V. C.; Kimberley,
B. S.; Solan, G. A. (BP) WO0020427, 2000. (d) Sommazzi, A.;
Milani, B.; Proto, A.; Corso, G.; Mestroni, G.; Masi, F.
(Enichem) WO0110875, 2001. (e) de Boer, E. J . M.; Deuling,

4386 Small et al.
Macromolecules, Vol. 37, No. 12, 2004
H. H.; van der Heijden, H.; Meijboom, N.; van Oort, A. B.
(Shell) W00158874, 2001.
(23) (a) Dias, E. L.; Brookhart, M.; White, P. S. Organometallics
2000, 19, 4995. (b) Dias, E. L.; Brookhart, M.; White, P. S.
Chem. Commun. 2001, 423.
(15) Modified methylalumoxane refers to MMAO-3A (Al 7% by
wt), an isobutyl-modified MAO supplied by Akzo Nobel.
(24) Batch studies of the Cr(II) and Cr(III) systems showed similar
activities, product distributions, and product compositions
when the same ligand was used with either metal source. To
further evaluate the similarity of the Cr(II) and Cr(III)
systems, a side-by-side comparison was performed using a
Multiclave 10× reactor from Autoclave Engineers. This
reactor possesses 10 individual 30 mL cells, with each cell
having stirring and temperature control capabilities. A 2.0
mg amount of five different Cr(II) and Cr(III) complexes was
activated with MMAO (1000:1 Al:Cr) in heptane (15 mL) and
exposed to 13.6 bar of ethylene, fed “on demand” for 30 min.
In each case, as in the larger batch studies, the products were
the same, depending entirely on the ligand rather than the
oxidation state of the precatalyst. The only difference ob-
served was that the Cr(III) complexes appeared to activate
slightly faster, which was noted by watching the comparative
temperature profiles for the Cr(II) and Cr(III) catalysts,
respectively.
(25) UV-vis studies have been used to study the reaction between
alumoxanes and transition-metal complexes in related Fe
systems: (a) Luo, H.-K.; Yang, Z.-H.; Mao, B.-Q.; Yu, D.-S.;
Tang, R.-G. J . Mol. Catal. A: Chem. 2002, 177, 195. (b)
Britovsek, G. J . P.; Gibson, V. C.; Spitzmesser, S. K.;
Tellmann, K. P.; White, A. J . P.; Williams, D. J . J . Chem.
Soc., Dalton Trans. 2002, 1159.
(26) Zuech, E. A. (Phillips Petroleum) U.S. Patent 3726939, 1973.
(27) TeXsan for Windows, V 1.02; Molecular Structure Corp.,
Inc.: The Woodlands, TX.
(16) The activity of the 2-tert-butyl complex 5 in entry 5 of Table
3 agrees reasonably well with the activity reported in ref 11a.
(17) To determine whether the 2-butene was being generated
directly by the dimerization reaction or in a separate isomer-
ization step, MMAO-activated 1 was tested for the isomer-
ization of 1-hexene and found to be a potent selective
isomerization catalyst for converting the 1-olefin to the
2-olefin.
(18) (a) J ohnson, L. K.; Killian, C. M.; Brookhart, M. J . Am. Chem.
Soc. 1995, 117, 6414. (b) Killian, C. M.; J ohnson, L. K.;
Tempel, D. J .; Brookhart, M. J . Am. Chem. Soc. 1996, 118,
11664.
(19) (a) Sugiyama, H.; Aharonian, G.; Gambarotta, S.; Yap, G. P.
A.; Budzelaar, P. H. M. J . Am. Chem. Soc. 2002, 124, 12268.
(b) Reardon, D.; Aharonian, G.; Gambarotta, S.; Yap, G. P.
A. Organometallics 2002, 21, 786.
(20) Reardon, D.; Conan, F.; Gambarotta, S.; Yap, G.; Wang, Q.
J . Am. Chem. Soc. 1999, 9318.
(21) Talzi, E. P.; Babushkin, D. E.; Semikolenova, N. V.; Zudin,
V. N.; Zakharov, V. A. Kinet. Catal. 2001, 42 (2), 147.
(22) (a) Gibson, V. C.; Humphries, M. J .; Tellman, K. P.; Wass,
D. F.; White, A. J . P.; Williams, D. J . Chem. Commun. 2001,
2252. (b) Kooistra, T. M.; Knijnenburg, Q.; Smits, J . M. M.;
Horton, A. D.; Budzelaar, P. H. M.; Gal, A. W. Angew. Chem.,
Int. Ed. 2001, 40 (24), 4719.
(28) SHELXTL V. 5.1; Bruker AXS: Madison, WI.
MA035554B