Tridentate cobalt catalysts for linear dimerization and isomerization of α-olefins

Article abstract of DOI:10.1021/om030210v

Upon activation with modified methylalumoxane (MMAO), tridentate pyridine bisimine cobalt catalysts dimerize α-olefins with lower productivity than analagous iron systems, as indicated by comparing the batch dimerization of 1-butene (TON ～42 000 for Co, ～147 000 for Fe). The cobalt-produced dimers are extremely linear (>97%) and contain only traces of trimeric species. The cobalt catalysts also have a tendency to isomerize α-olefins, as evidenced by the approximately equal levels of dimerization and isomerization achieved when 1-butene is dimerized. When the cocatalyst is changed to diethylaluminum chloride, isomerization occurs exclusively to give cis- and trans-2-olefins selectively. To mitigate the effects of feed isomerization, dimerization of propylene was also studied. GC analysis of the products reveals a stepwise oligomerization process that makes linear hexenes, nonenes, and dodecenes, with the hexenes comprising up to 70% of the product mix. The hexenes are over 99% linear and may contain over 50% 1-hexene. Catalyst productivity is high, with turnover numbers exceeding 200 000 mol propylene/mol Co (17 000 g oligomers/g Co complex).

Full text of DOI:10.1021/om030210v

3178
Organometallics 2003, 22, 3178-3183
Articles
Tr id en ta te Coba lt Ca ta lysts for Lin ea r Dim er iza tion a n d
Isom er iza tion of r-Olefin s
Brooke L. Small*
Chevron Phillips Chemical Company LP, 1862 Kingwood Drive, Kingwood, Texas 77339
Received March 19, 2003
Upon activation with modified methylalumoxane (MMAO), tridentate pyridine bisimine
cobalt catalysts dimerize R-olefins with lower productivity than analagous iron systems, as
indicated by comparing the batch dimerization of 1-butene (TON ∼42 000 for Co, ∼147 000
for Fe). The cobalt-produced dimers are extremely linear (>97%) and contain only traces of
trimeric species. The cobalt catalysts also have a tendency to isomerize R-olefins, as evidenced
by the approximately equal levels of dimerization and isomerization achieved when 1-butene
is dimerized. When the cocatalyst is changed to diethylaluminum chloride, isomerization
occurs exclusively to give cis- and trans-2-olefins selectively. To mitigate the effects of feed
isomerization, dimerization of propylene was also studied. GC analysis of the products reveals
a stepwise oligomerization process that makes linear hexenes, nonenes, and dodecenes, with
the hexenes comprising up to 70% of the product mix. The hexenes are over 99% linear and
may contain over 50% 1-hexene. Catalyst productivity is high, with turnover numbers
exceeding 200 000 mol propylene/mol Co (17 000 g oligomers/g Co complex).
In tr od u ction
The major byproduct in the reaction is the methyl-
branched dimer, which results from two successive 2,1
insertions followed by chain termination. In addition,
approximately 15% of the product is a lightly branched
olefin trimer, which possesses an average of g1.2
branches per molecule. The final product in the reaction
is the undimerized olefin substrate, which often contains
several percent of isomerized material due to chain
transfer following an initial 2,1 insertion (Scheme 2).
In an attempt to expand this dimerization chemistry
to other transition metals, several tridentate cobalt
complexes were synthesized and tested for their dimer-
ization ability.6 Remarkably, some of these cobalt
catalysts are much more selective for producing linear
dimers than their iron analogues, while other systems
are highly selective for isomerizing the starting mater-
ial. Herein are reported the details of this cobalt catalyst
study.
The dimerization of olefins by transition metal com-
plexes represents an important class of industrially
relevant chemistry.1 For example, ethylene dimerization
to 1-butene can provide a source of comonomer in the
production of polyethylene,2 and olefins such as propyl-
ene and butene are dimerized to give C6-C8 materials
that serve as feedstocks for gasoline blending or alcohol
production.3 While most dimerization catalysts produce
branched dimers from propylene and higher olefins, we
recently reported a family of iron-based catalysts that
make predominantly linear dimers (up to 80% linear-
ity).4 Due to a unique mechanism of dimerization in
which the regiochemistry of olefin insertion changes
from 1,2 to 2,1 between the first and second steps,5 the
major product is the head-to-head dimer (Scheme 1).
* Corresponding author.
(1) (a) Chauvin, Y.; Olivier, H. In Applied Homogeneous Catalysis
with Organometallic Compounds; Cornils, B., Herrmann, W., Eds.;
VCH: New York, 1996; Vol. 1, pp 258-268. (b) Skupinska, J . Chem.
Rev. 1991, 91, 613. (c) Parshall, G. W.; Ittel, S. D. In Homogeneous
Catalysis, The Applications and Chemistry of Catalysis by Soluble
Transition Metal Complexes, 2nd ed.; J ohn Wiley & Sons: New York,
1992; pp 72-85. (d) Bhaduri, S.; Mukesh, D. In Homogeneous Catalysis,
Mechanisms and Industrial Applications; J ohn Wiley & Sons: New
York, 2000; pp 142-147.
Resu lts a n d Discu ssion
Cobalt complexes 1-4, shown in Figure 1, were
synthesized by reported methods and tested for their
ability to dimerize R-olefins.7-9 In efforts to compare
(2) Al-J arallah, A. M.; Anabtawi, J . A.; Siddiqui, M. A. B.; Aitani,
A. M.; Al-Sa’doun, A. W. Catal. Today 1992, 14 (1).
(5) (a) Small, B. L.; Brookhart, M. Macromolecules 1999, 32, 2120.
(b) Pellecchia, C.; Mazzeo, M.; Pappalardo, D. Macromol. Rapid
Commun. 1998, 19, 651.
(6) Cobalt complexes containing pyrrole ligands have previously
been reported for the dimerization of ethylene to high purity
1-butene: Wu, A. U.S. Pat. 5414178 (Phillips Petroleum), 1995.
(7) For specific ligand syntheses, see the following references: (a)
Small, B. L.; Brookhart, M. J . Am. Chem. Soc. 1998, 120, 7143. (b)
Alyea, E. C.; Merrell, P. H. Synth. React. Inorg. Metal-Org. Chem. 1974,
4 (6), 535.
(3) (a) Olivier-Bourbigou, H.; Chodorge, J .-A.; Travers, P. Pet.
Technol. Q. 1999, Autumn, 141. (b) Chauvin, Y.; Gaillard, J . F.; Quang,
D. V.; Andrews, J . W. Chem. Ind. 1974, 375. (c) Commereuc, D.;
Chauvin, Y.; Gaillard, J .; Le´onard, J .; Andrews, J . W. Hydrocarbon
Process. 1984, 118.
(4) (a) Small, B. L.; Marcucci, A. J . Organometallics 2001, 20, 5738.
(b) Small, B. L.; Baralt, E. J .; Marcucci, A. J . (Chevron Phillips) U.S.
Patent 6291733, 2001.
10.1021/om030210v CCC: $25.00 © 2003 American Chemical Society
Publication on Web 07/04/2003

Tridentate Cobalt Catalysts
Organometallics, Vol. 22, No. 16, 2003 3179
Sch em e 1. Dim er iza tion of 1-Bu ten e Ca ta lyzed by Tr id en ta te Ir on Ca ta lysts
Ta ble 1. Dim er iza tion of 1-Bu ten e Usin g MMAO-Activa ted Coba lt Com p lexes 2-4
productivity
cat./mass Al:Co C4 mass
rxn length
(h)
prod.
(g dimer/g Co
complex)
% linear
dimer
entry
(mg)
ratio
(g)
T (°C)
mass (g) % conversiona
% dimer
D/I ratiob
1
2
3
4
5
2/25
3/25
3/82
4/25
2′/10c
500
500
200
500
500
250
250
1080
250
30
30
30
30
30
3
3
5
3
3
97.7
51.4
340
11.7
153
39.1
20.6
31.5
4.7
3910
2060
4150
470
15 300
99
99
99
97
82
97
98
98
98
70
0.70
0.71
0.72
0.47
250
61.2
>20
a
b
Does not include isomerized butene. D/I ratio ) mass ratio of (dimerized + trimerized butene)/isomerized butene. c 2′ ) Fe analogue
of Co complex 2.
Sch em e 2. F or m a tion of Br a n ch ed Dim er s a n d
Isom er ized Sta r tin g Ma ter ia ls
these systems to our earlier results using iron, 1-butene
was dimerized in liquid phase to assess the catalyst
activity. The results of these reactions are reported in
Table 1. Unlike the analogous tridentate iron systems,
these cobalt catalysts produce extremely low levels of
methyl-branched heptenes in the octene products, re-
sulting in 97%+ linearity in the dimers (entries 1-4).
Figure 2 shows the analysis of the C8 portion of entry 2
F igu r e 1. Cobalt complexes 1-4.
To better understand the remarkable selectivity of the
cobalt catalysts, it was useful to examine the undimer-
ized butene. With the iron-based catalysts, the R-olefin
feed was only lightly isomerized (entry 5). With cobalt,
however, complexes 1-4 tend to isomerize the substrate
heavily, resulting in the production of substantial
quantities of 2-butene in the undimerized olefin. As
noted earlier, isomerization occurs when an initial 2,1
(secondary) insertion of olefin is followed by â-elimina-
tion with opposite regiochemistry. For iron, initial 2,1
insertions tend to produce branched dimers, indicating
that propagation is preferred; cobalt, on the other hand,
undergoes chain transfer following a 2,1 insertion,
resulting in both highly linear dimers and high amounts
of isomerization in the feed. These observations show
that cobalt is likely not more regioselective than iron
by GC/MS. Also, the cobalt systems make only traces
of butene trimer, in comparison to the iron systems,
which produce about 15% trimer (entry 5). Comparing
the catalyst activities, the cobalt systems are slower
than iron, and the overall catalyst productivity of the
iron system is several times higher than cobalt under
similar operating conditions (Fe ∼15 300 g product/g Fe
complex; Co ∼4150 g product/g Co complex). Both
systems may be activated with relatively low amounts
of alumoxane cocatalysts (<200:1 Al:cat molar ratios,
entry 3).10
(8) For specific Co complex syntheses, see the following references:
(a) Edwards, D. A.; Edwards, S. D.; Martin, W. R.; Pringle, T. J .
Polyhedron 1992, 11 (13), 1569. (b) Britovsek, G. J . P.; Mastroianni,
S.; Solan, G. A.; Baugh, S. P. D.; Redshaw, C.; Gibson, V. C.; White,
A. J . P.; Williams, D. J .; Elsegood, M. R. J . Chem. Eur. J . 2000, 6 (12),
2221.
(9) For general synthetic details for preparing pyridinebisimine
cobalt complexes, see, for example, the following references: (a) Small,
B. L.; Brookhart, M.; Bennett, A. M. A. J . Am. Chem. Soc. 1998, 120,
4049. (b) Britovsek, G. J . P.; Gibson, V. C.; Kimberley, B. S.; Maddox,
P. J .; McTavish, S. J .; Solan, G. A.; White, A. J . P.; Williams, D. J .
Chem. Commun. 1998, 849. (c) Ittel, S. D.; J ohnson, L. K.; Brookhart,
M. Chem. Rev. 2000, 100, 1169. (d) Britovsek, G. J . P.; Gibson, V. C.;
Wass, D. F. Angew. Chem., Int. Ed. 1999, 38, 428.
(10) A ratio of 500:1 Al:Co was chosen to ensure rapid activation
and the scavenging of residual moisture in the reactions. However,
our studies have shown that this ratio can be greatly reduced when
the scale of the reaction is increased. Entry 3 illustrates this trend,
which was also observed for the Fe-based catalysts (ref 4a). Table 3,
in which polymer grade propylene (<1 ppm H2O) was used, demon-
strates that low Al:Co ratios can be achieved when the olefin is
substantially free of moisture.

3180 Organometallics, Vol. 22, No. 16, 2003
Small
F igu r e 2. GC trace of 1-butene dimers made by MMAO-activated cobalt catalyst 3 (Table 1, entry 2).
Ta ble 2. Oligom er iza tion of P r op ylen e Usin g MMAO-Activa ted Coba lt Com p lexes 2-4a
productivity
cat./ mass Al:Co C3 mass prod. mass
(g olig./g Co mass C6
mass C9 mass C12+
entry
(mg)
ratio
(g)
(g)
% Conv.
complex)
(g)
% linearb % 1-hexenec
(g)d
(g)
1
2
3
4
5
6
2/20
3/20
4/20
2/20
3/20
4/20
100
100
100
150
150
150
400
400
400
400
400
400
304
231
187
342
312
147
76
58
47
86
78
37
15 200
11 600
9400
17 100
15 600
7400
177
105
130
200
142
112
99.0
99.7
99.9
99.0
99.7
99.8
10
14
59
8
22
57
87
72
47
97
101
28
40
54
11
45
69
6
a
b
Reaction conditions: Each reaction was run for 20 h at 30 °C. % linear ) % linearity of the C6 fraction. c % 1-hexene ) % 1-hexene
d
in the C6 fraction. All of the nonene products contained significant quantities of 1-nonenes. Entry 6, for example, contained 34% 1-nonene
in the C9 fraction.
F igu r e 3. GC trace of propylene dimers made by MMAO-activated cobalt complex 4 (57% 1-hexene, Table 2, entry 6).
at the point of initial insertion, but rather that the cobalt
systems promote propagation (dimerization) only when
a cobalt-n-alkyl species is present. This implies that if
the isomerization side reaction could be suppressed, it
would likely be achieved at the expense of dimer
linearity, since initial 2,1 insertions would lead to
branched dimer formation.
To remove feed isomerization as a possibility, a study
of propylene dimerization was undertaken, the results
of which are summarized in Table 2. Propylene oligo-
merization with tridentate cobalt catalysts has been
reported by Bennett;11 however, the catalysts used in
this study are less sterically bulky than those described
in the earlier report. The experiments in Table 2 employ
cobalt catalysts bearing a single ortho substituent on
each aryl ring, and the results illustrate several unique
trends. First, the catalysts are highly active, with
catalyst productivities exceeding 17 000 g product/g Co
complex (entry 4). Second, not only do the catalysts
produce extremely linear dimers (Figure 3), but the
trimer products are also highly linear (Figure 4). GC/
MS analysis of the C9 fraction made by catalyst 4
revealed 95% linearity in the nonenes, a clear indicator
that the C9 and C12 byproducts are formed by a step
growth dimerization process (chain transfer following
each insertion). As further evidence for a step growth
process, analysis of the linear nonenes by GC/MS also
did not reveal any 3-nonenes, which agrees with the
proposed catalytic cycle shown in Scheme 3. Further-
more, the nonenes made by catalyst 4 were found to
contain over 30% 1-nonene (Table 2, entry 6), a result
of codimerization of 1-hexene and propylene, with
propylene involved in the second insertion step.
As shown in Scheme 3, perhaps the most interesting
feature of these catalysts is their ability to make
1-hexene from propylene. Under the conditions em-
ployed in this report, it was possible to isolate a
propylene-based oligomer in which 70% of the products
were n-hexenes with over 99.8% linearity. Of these
hexenes, 59% were the 1-hexene isomer, representing
an overall product distribution that contained 41%
1-hexene (catalyst 4, entry 3). Both the linearity of the
dimerization process and the selectivity for 1-hexene
(11) Bennett, A. M. A. (DuPont) U.S. Patent 6063881, 2000.

Tridentate Cobalt Catalysts
Organometallics, Vol. 22, No. 16, 2003 3181
F igu r e 4. GC trace of propylene trimers made by MMAO-activated cobalt complex 4 (34% 1-nonene, Table 2, entry 3).
Sch em e 3. P r op osed Mech a n ism for Coca ta lyzed P r op ylen e Oligom er iza tion a
a
Dodecene formation is omitted for clarity; [Co] ) unobserved cobalt initiating species.
a step growth rather than a chain growth mechanism,
the inverse trend may be rationalized by further inspec-
tion of Scheme 3. Further step growth following a
dimerization reaction requires an R-olefin such as
1-hexene, but 1-hexene’s increased size relative to
propylene promotes propylene uptake in the successive
catalytic steps. Even though complex 4 exhibits a high
selectivity for producing 1-hexene, this 1-hexene is
relatively inert toward further dimerization steps. As
also seen in Figure 5, this trend of hexene percentage
in the total product does not remain true over the series
F igu r e 5. Product distribution by carbon number for
catalysts 2-4, entries 1-3 in Table 2 (C12 refers to C12+).
i
of catalysts 2-4 (Me f Et f Pr). While complex 4
makes the most hexenes, methyl-substituted 2 makes
the second highest amount and ethyl-substituted 3
produces the lowest percentage of hexenes. This anomaly
in the trend is likely due to catalyst 3 (Et) having a
higher selectivity for making 1-hexene than catalyst 2
(Me) and by catalyst 3’s ability to incorporate 1-hexene
into nonenes and dodecenes more readily than catalyst
4. These observations readily illustrate that the product
composition in these types of step growth reactions will
be highly dependent on the interplay of factors such as
the selectivity toward 1-hexene, reaction length (resi-
dence time), and R-olefin concentrations.
are, to our knowledge, unprecedented for propylene
dimerization.
Figure 5 shows the product distribution by carbon
number for catalysts 2-4, taken from entries 1-3 of
Table 2. Interestingly, catalyst 4, the most sterically
bulky system, produces the largest percentage of hexenes,
which is counterintuitive when one considers the previ-
ous reports that have directly correlated product mo-
lecular weight to increasing ligand sterics.7a,8b,9 How-
ever, because the propylene oligomerization proceeds via

3182 Organometallics, Vol. 22, No. 16, 2003
Small
Ta ble 3. Com p a r ison of Non en e a n d Dod ecen e
Lin ea r ity of Oligom er s Ma d e by Ca ta lysts 2-4
provides more details on isomerization reactions using
1-hexene as substrate; other R-olefins may also be used.
C6 linearity C9 linearity C12 linearity C15 linearity
catalyst
(%)
(%)
(%)
(%)
Con clu sion
2/MMAO
3/MMAO
4/MMAO
99.0
99.7
99.9
96.0
94.9
94.8
94.5
93.5
87.2
93.5
91.6
75.9
The dimerization reactions reported herein represent
the most selective linear dimerization technology known.
1-Butene dimerization catalyzed by tridentate cobalt
complexes produces over 97% linear dimer, but the
overall catalyst activity is lower than the previously
reported iron systems. The butene dimers made by
cobalt are highly linear because chain transfer is fast
relative to chain propagation following an initial 2,1
insertion, which leads to a competing isomerization
reaction that rivals the rate of dimerization. Because
propylene cannot be isomerized to a somewhat inert
internal olefin, it undergoes dimerization to make up
to 99.9% linear hexenes, which may include over 40%
1-hexene in the final isolated product. The 1-hexene can
also be isomerized or dimerized during the course of the
reaction to form nonenes, dodecenes, and even small
amounts of higher olefins. Due to the remarkable step
growth oligomerization reaction, the nonenes may be
over 95% linear, and they may contain over 30%
1-nonene. Because the reaction is a step growth rather
than a chain growth process, the concentrations of the
various products depend greatly on the interaction of
many factors, some of which include the relative rates
of dimerization and isomerization, the catalyst selectiv-
ity for making R-olefins or internal olefins during the
chain transfer step, the overall level of conversion, and
the actual catalyst structure used. Future studies will
be directed toward understanding the interaction of
these various factors, with a view of optimizing the
process operating conditions toward the production of
specific products.
Ta ble 4. Isom er iza tion of 1-Hexen e Usin g Coba lt
Com p lexes 1 a n d 2a
rxn
product
cat./mass
(mg)
Al:Co olefin/amt
cocat. ratio (mL)
T
length
distribution
(°C) (h)
(% each isomer)
1/10
MMAO 115 1-hexene/50 35
2
1.9 1-hexene
62.4 t-2-hexene
34.2 c-2-hexene
1.0 other hexenes
0.6 dimer
1/28
MMAO 60 1-hexene/50
5
1
1.1 1-hexene
77.5 t-2-hexene
19.9 c-2-hexene
0.9 other hexenes
0.5 dimer
1/10
2/27
DEAC
DEAC
40 1-hexene/50 25
18
72
1.0 1-hexene
62.9 t-2-hexene
15.4 c-2-hexene
20.5 3-hexenes
1.0 1-hexene
40 1-hexene/100 25
61.2 t-2-hexene
15.6 c-2-hexene
22.0 3-hexenes
0.1 dimer
a
The effect of varying the Al:Co ratio with DEAC-activated
catalysts was not investigated and would not be expected to
significantly change the product distribution.
Another interesting trend was discovered upon ex-
amining the linearity of the propylene oligomers. To
analyze these oligomers, they were first hydrogenated
using 10% Pd/C and 1 atm of hydrogen pressure (see
Experimental Section). Catalyst 4, not surprisingly,
makes the most linear hexene products because its
steric bulk promotes higher regioregularity.4 However,
catalyst 4 produces the least linear nonenes, dodecenes,
and pentadecenes (Table 3). This increased branching
in the higher carbon numbers is attributable to the
competing chain growth mechanism, which becomes
more prevalent with increasing ligand sterics. Catalysts
2 and 3 undergo less of this chain growth propagation,
thereby explaining the higher linearity of their higher
carbon number products.
In our attempts to use other activating cocatalysts
with these cobalt complexes, it was discovered that the
ratio of dimerization to isomerization varies dramati-
cally depending on the activator used. For example,
when complexes 2-4 are activated with MMAO, dimer-
ization and isomerization of the feed are competitive.
When diethylaluminum chloride (DEAC) is used, isomer-
ization occurs almost exclusively, resulting in the selec-
tive isomerization of 1-olefins to 2-olefins. These data
are reported in Table 4. Rather than producing a
thermodynamic distribution of internal olefin isomers
from the R-olefin feed, the catalysts typically move the
double bond only one position. After extended reaction
times (of 10 days), the distribution is closer to thermo-
dynamic, but the predominant olefin isomer remains the
2-olefin. When complex 1, which bears no ortho alkyl
groups on the aryl rings, is used as the precatalyst,
selective isomerization occurs regardless of whether
MMAO or DEAC is employed as the activator. Table 4
Exp er im en ta l Section
Ma ter ia ls. Cobalt(II) chloride hexahydrate, 2,6-diacetylpy-
ridine, diethylaluminum chloride, 10% Pd/C, and all aniline
derivatives were purchased from Aldrich and used without
further purification. Polymer grade propylene in cylinders with
dip tubes for transfer of liquefied gas was purchased from
Matheson Gas Products, Inc. CPChem’s commercial grade of
1-butene was used without purification. CPChem’s 1-hexene
was degassed and dried over 3 Å molecular sieves prior to use.
MMAO 3A was purchased from Akzo Nobel.
Syn th esis of P r eca ta lyst Com p lexes 1-4. Precatalyst
complexes 1-4 were synthesized according to literature meth-
ods, as were the ligands used to make the complexes.7-9 In
general, the ligands were prepared by dissolving 2,6-di-
acetylpyridine and a slight excess (>2 equiv) of the appropriate
aniline in methanol, heating the solution for 1 day under inert
atmosphere with
a catalytic amount of acetic acid, and
recrystallizing the isolated solid from ethanol. The cobalt
complexes were prepared by stirring a slight excess of the
tridentate ligand with cobalt(II) chloride hexahydrate in THF
for at least 1 day, then adding pentane to the solution and
removing the precatalyst complexes by filtration. The com-
plexes were all isolated in near-quantitative yield. Elemental
analyses for complexes 1-4 were carried out to determine the
amount of THF in the isolated precatalysts.12 The solids were
heated under vacuum at 40 °C prior to analysis. Complexes
(12) Complex 1 has been previously reported (ref 8a), but was
prepared using ethanol instead of THF as solvent.

Tridentate Cobalt Catalysts
Organometallics, Vol. 22, No. 16, 2003 3183
1-3 tested positive for an equivalent of THF, but complex 4
contained only trace amounts. Elemental analyses are reported
as follows.
breakage of the NMR tube and activation of the catalyst.
Reactor temperatures were easily maintained by internal
cooling.
2,6-Bis[1-(phenylimino)ethyl]pyridinecobalt(II)chloride‚
THF (1). Anal. Calcd for C25H27N3Cl2OCo: C, 58.27; H, 5.28;
N, 8.15; O, 3.10. Found: C, 57.90; H, 5.00; N, 8.50; O, 2.71.
2,6-Bis[1-(2-m eth ylp h en ylim in o)eth yl]p yr id in ecoba lt-
(II) ch lor id e‚THF (2). Anal. Calcd for C27H31N3Cl2OCo: C,
59.68; H, 5.75; N, 7.73; O, 2.94. Found: C, 59.32; H, 5.57; N,
8.22; O, 2.33.
2,6-Bis[1-(2-et h ylp h en ylim in o)et h yl]p yr id in ecob a lt -
(II) ch lor id e‚THF (3). Anal. Calcd for C29H35N3Cl2OCo: C,
60.95; H, 6.17; N, 7.35; O, 2.80. Found: C, 60.02; H, 5.80; N,
8.06; O, 2.37.
P r oced u r e for Dim er iza tion /Isom er iza tion of Liqu id
Olefin s. Under inert conditions, the appropriate cobalt pre-
catalyst was added to a dry flask with a stirbar. The R-olefin
was then added, and rapid stirring was begun to slurry the
complex. The flask was placed under a slight argon purge, and
the cocatalyst was added via syringe. Temperatures were
maintained by use of a water cooling bath.
P r od u ct An a lysis. After slowly adding water to deactivate
the catalyst, tridecane was added as an internal standard.
Hydrogenation of olefinic products was achieved using a 10%
Pd/C catalyst with 1 atm of hydrogen gas slowly bubbling
through the reaction suspension. Total conversion to saturated
species was accomplished over several hours using 100 mg of
the catalyst and 5 mL of the olefin mixture dissolved in
tridecane. A Hewlett-Packard 6890 Series GC System with an
HP-5 50m column with a 0.2 mm inner diameter was used for
product characterization. Agilent ChemStation from Agilent
Technologies was used to analyze the collected data. GC/MS
data were obtained using an Agilent 5973 benchtop mass
spectrometer using electron impact ionization interfaced to an
HP 6890 gas chromatograph. The GC column was a J &W
Scientific DB-5MS, 60 m × 0.25 mm i.d.
2,6-Bis[1-(2-isopr opylph en ylim in o)eth yl]pyr idin ecobalt-
(II) ch lor id e (4). Anal. Calcd for C25H27N3Cl2Co: C, 61.49;
H, 5.92; N, 7.97; O, 0.00. Found: C, 60.91; H, 5.89; N, 7.43; O,
0.16.
P r oced u r e for Dim er iza tion of Liqu efied Ga ses. Under
inert conditions, the appropriate cobalt complex was weighed
out and added to an NMR tube. A small amount of methylene
chloride was added to solublize the complex, and the tube was
sealed. The sealed tube was then tied, using copper wire, to
the internal cooling coils of a clean, dry Zipperclave reactor.
The reactor was evacuated and then placed under static
vacuum. A glass charger was then used to transfer the
cocatalyst to the reactor, and the reactor was back-filled with
argon. The liquefied gas cylinder was pressurized with a head
pressure of argon and placed on a scale with (5 g accuracy.
Flexible hose was used to connect the gas cylinder to the
reactor, and the desired amount of olefin was delivered to the
reactor using the head pressure of the cylinder. The reactor
was pressurized further with argon to ensure that the olefin
remained in the liquid phase. Stirring was begun resulting in
Ack n ow led gm en t. Dr. Mark J eansonne is grate-
fully acknowledged for GC/MS analyses of the oligo-
merization products. A. J . Marcucci, Eric Fernandez,
and Ray Rios are acknowledged for carrying out many
of the dimerization reactions.
OM030210V