Selective dimerisation of α-olefins using tungsten-based initiators

Article abstract of DOI:10.1039/c0dt00106f

The selective dimerisation of the α-olefins 1-pentene through to 1-nonene is reported using an in situ-generated catalyst derived from tungsten hexachloride, aniline, triethylamine and alkylaluminium halide. The influence of reagent identity and reaction stoichiometry, along with activator, solvent and α-olefin substrate choice are probed. The catalyst is found to be highly selective towards dimerisation, minimising the formation of undesired heavier oligomers. Notably, the selectivity within the dimer fraction is found to favour the formation of products with methyl branches. The selectivity towards individual olefin isomers has been determined and the system is found to also produce trace levels of dienes and alkanes. A kinetic study of the system reveals a second order dependence on substrate. Comparison of the products observed, with those expected for metallacyclic and Cossee-type mechanisms, suggests that the latter is in operation, something confirmed by the results of a C₂H₄/C₂D₄ co-dimerisation experiment which showed full isotopic scrambling in the products. Thus a mechanistic proposal is made to account for the observed behaviour of the system, including the diene and alkane formation. The Royal Society of Chemistry 2010.

Full text of DOI:10.1039/c0dt00106f

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Selective dimerisation of a-olefins using tungsten-based initiators†
a
a
b
a
a
c
Martin J. Hanton,* Louisa Daubney, Tomas Lebl, Stacey Polas, David M. Smith and Alex Willemse
Received 10th March 2010, Accepted 30th April 2010
First published as an Advance Article on the web 25th May 2010
DOI: 10.1039/c0dt00106f
The selective dimerisation of the a-olefins 1-pentene through to 1-nonene is reported using an in
situ-generated catalyst derived from tungsten hexachloride, aniline, triethylamine and alkylaluminium
halide. The influence of reagent identity and reaction stoichiometry, along with activator, solvent and
a-olefin substrate choice are probed. The catalyst is found to be highly selective towards dimerisation,
minimising the formation of undesired heavier oligomers. Notably, the selectivity within the dimer
fraction is found to favour the formation of products with methyl branches. The selectivity towards
individual olefin isomers has been determined and the system is found to also produce trace levels of
dienes and alkanes. A kinetic study of the system reveals a second order dependence on substrate.
Comparison of the products observed, with those expected for metallacyclic and Cossee-type
mechanisms, suggests that the latter is in operation, something confirmed by the results of a
C
2
H
4
/C
2
4
D co-dimerisation experiment which showed full isotopic scrambling in the products. Thus a
mechanistic proposal is made to account for the observed behaviour of the system, including the diene
and alkane formation.
Introduction
internal olefins with high skeletal selectivity (>99% linear), and
much reduced formation of higher oligomers (15–20%).
3b
The homogeneous, metal-catalysed dimerisation of olefins is a
conversion of particular industrial interest. Accordingly, many
catalyst systems and several commercial processes, capable of
effecting this reaction selectively are known. Where the dimerisa-
The use of group six complexes with N-donor ligands in dimeri-
sation catalysis has been reported on several previous occasions,
1
7
in the form of a pre-catalyst formed in situ, or from activation
2
8
of tungsten imido complexes. However, this class of dimerisation
tion of a-olefins is concerned, the selectivity towards dimerisation,
but also towards specific branching in the skeleton of the
product must be considered. In this regard, catalysts displaying
catalyst is relatively unexplored in comparison to many others,
something that led us to re-examine systems of this type for
the dimerisation of a-olefins. We recently reported an in situ
catalyst system based upon tungsten hexachloride in combination
with varying amounts of aniline and tertiary amine; which when
activated with aluminium co-catalyst gives the highest selectivity
to mono-methyl branched product reported to date for group
3
4
enhanced selectivity towards linear, mono-branched and di-
5
branched products within the dimer fraction have been disclosed.
Furthermore, the location of the unsaturation in the dimer product
is dependent upon the catalyst system employed, and hence it is
desirable to have control of this additional degree of variation to
9
six-based olefin dimerisation catalysts. Furthermore, the catalyst
3c
tailor products to specific applications.
is readily prepared from economically attractive precursors and
the selectivity towards dimerisation (vs. higher oligomerisation) is
often >99%, making this system highly desirable in terms of atom
efficiency. Herein, we report full details of this initiator system and
demonstrate that through modification of the many experimental
parameters a degree of control can be exerted over the dimerisation
of a-olefins.
The advantages of highly selective dimerisation processes are
manifold, and include efficiency of feed usage via the avoidance
of by-product formation, and the reduction of isomerisation
1a
processes. For example, group four cyclopentadienyl ligand-
based catalysts when activated with aluminium co-catalysts pro-
duce methylidene (terminal) products from the dimerisation of
a-olefins, but these systems suffer from significant formation of
4,6
higher oligomers (up to 50%). Meanwhile, recent reports of
bis(imino)pyridyl ligand systems bound to iron, show these can
be used as pre-catalysts in the dimerisation of a-olefins to give
Results and discussion
In situ catalysis
The catalyst system studied can be described by the general
formula shown in Fig. 1. The catalyst is prepared by addition of a
primary amine, preferably an aniline, and optionally a base, to a
solution of tungsten hexachloride. The mixture is then left to stir
a
Sasol Technology UK, St Andrews Laboratory, Purdie Building, North
Haugh, St Andrews, Fife, UK KY16 9ST. E-mail: martin.hanton@
eu.sasol.com; Fax: +44(0)1334 460 939; Tel: +44 (0)1334 460 830
b
School of Chemistry, University of St Andrews, Purdie Building, North
Haugh, St Andrews, Fife, UK KY16 9ST
for 15–30 min at the ‘catalyst formation temperature’ (T ), before
F
c
Sasol Technology (Pty) Ltd., Klasie Havenga Street, PO Box 1, Sasolburg,
addition of olefin substrate, and finally, aluminium activator to
initiate the dimerisation catalysis.
Previous reports of this class of catalyst system do not include
the use of a base; the removal of HCl having been achieved via
1
947, South Africa
†
Electronic supplementary information (ESI) available: NMR analysis of
dienes, alternative metallocycle mechanism and synthesis of GC standards.
See DOI: 10.1039/c0dt00106f
This journal is © The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 7025–7037 | 7025

View Article Online
a
Table 1 Varying catalyst composition
Branching selectivity
Conv. (%) S (%) Linear Me Et MeMe MeEt EtEt
◦
b
c
d
Entry X Eq RNH
2
Y Eq Et
3
N
T
F
/T
C
( C) Olefin (eq)
Act
TON
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
0
0
0
0
0
0
0
0
0
0
1
2
4
3
60/60
60/60
60/60
60/60
60/60
60/60
60/60
60/60
60/60
60/60
60/60
60/60
60/60
60/60
60/60
60/60
20/20
20/20
20/20
20/20
20/20
20/20
20/20
100/100
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
1-C
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
5
5
5
5
5
5
5
5
5
6
6
6
0
0
0
—
nd
nd
nd
nd
nd
nd
nd
—
nd
nd
95.3
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
nd
60.1
97.1
80.0
62.6
—
0
0
0
0
0
0
0
—
0
0
0
0
0
0
0
0
0.2
0.2
0.2
0.2
0.2
0
0.5
0
0
—
—
0
0
0
0
0
0
0
—
—
6.3
6.8
10.9 1.3
9.4
29.4
28.9
44.7
—
—
7.1
7.7
—
0
0
0
0
0
0
0
—
0
0
0
0
0
0
0
0
0
0.7
0.7
0
0.5
0
0
0
0
0
0
g
g
g
g
1 (Ph)
1.5 (Ph)
2 (Ph)
2.5 (Ph)
3 (Ph)
4 (Ph)
6 (Ph)
8 (Ph)
1 (Ph)
2 (Ph)
2 (Ph)
3 (Ph)
1 (Ph)
2 (Ph)
3 (Ph)
1 (Ph)
2 (Ph)
3 (Ph)
4 (Ph)
1 (Ph)
1 (Ph)
2 (Ph)
2 (Ph)
1 (Ph)
2 (4-F-Ph)
51
46
60
57
117
44
146
0
54
58
108
45
46
205
183
240
228
468
177
584
0
325
349
432
270
275
338
336
630
870
904
845
835
808
595
321
116
408
95
41.0
36.6
48.0
45.6
46.8
35.4
58.4
0
65.1
69.8
43.2
54.0
55.0
67.6
67.2
48.7
87.0
90.4
84.5
83.5
80.8
59.5
32.1
11.6
40.7
9.5
86.6
85.5
87.8
86.1
70.6
71.1
55.3
—
4.5
0
0
0
—
g
g
h
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
80.4 0.1 18.7 0.8
72.3 0.2 27.5
57.1 42.9
63.1 0.1 36.8
88.1 0.1 9.7
79.2 0.3 20.5
73.8 0.2 26.0
57.8
55.7
58.4 1.9 36.6 2.2
59.4 1.8 35.6 2.3
57.1
61.5 1.2 35.3 1.4
56.1
g
h
0
0
0
2.0
0
0
g
g
g
g
0.5 (dabco)
1 (dabco)
1.5 (dabco)
56
56
0
0
0
0
0
1
3
4
4
0
4
4
4
157
218
226
211
209
202
149
160
23
0
0
42.0 0.2
44.0 0.1
0
42.7 0.1
0
43.9
0
0
0
0
0
0
e
50.9 0.5 48.1
f
132/20
53.9
58.2
57.5
61.1
0
0
0
0
44.1
41.8
42.5
38.9
60/60
60/60
60/60
102
24
14
2 (4-Me-Ph)
2 (4-Cl-Ph)
0
0
55
5.5
a
General conditions: 0.1 mmol WCl
6
, X mmol PhNH
2
, Y mmol Et
3
N, 1.2 mmol EtAlCl
2
(EADC), PhCl (20 mL), catalyst formation period (30 mins
◦
◦
b
-1
-1
c
-1
@
T
F
C), nonane standard (1 mL), catalysis temperature T
C
C, 1-olefin (0.1 mol), 240 mins. (mol olefin)(mol W) hr . (mol olefin)(mol W)
d
e
f
Selectivity to the dimer fraction. 120 mins. According to method of US5,059,739, HCl removed via reflux under flow of N
2
for 60 mins, then cooled
◦
g
h
to 25 C for catalysis, 300 mins. 1-olefin (50 mmol). 360 mins.
case. Thus the requirement for at least some aniline combined with
the previous observation of HCl evolution, suggests that the active
catalyst is constituted of a W-amido or -imido species, as has been
8
suggested elsewhere.
Initial examination of the base-free system at catalyst formation
◦
(
T
F
) and catalysis (T
C
) temperatures of 60 C (Table 1, entries
2–8), shows a general increase in activity and productivity as
the amount of aniline is increased from 1 to 6 equivalents.
However, it is strongly suspected that at least some of the aniline
is behaving as a Lewis base and sequestering the HCl evolved, as
in the case where 6 equivalents are employed, if all reacted with
Fig. 1 The dimerisation of a-olefins using a tungsten-based catalyst
prepared in situ.
WCl , there would presumably be no vacant sites left at which
6
catalysis could proceed. Indeed, when 8 equivalents of aniline are
utilised this situation maybe realised, as no catalysis is observed
(
entry 9). However, an alternative explanation maybe that excess
7
a,c
7b
using excess aluminium reagent
or via sparging. However,
aniline coordinates to the active species, blocking catalysis. In
terms of selectivity, it can be seen that a distinct swing in the
degree of branching within the dimer fraction is observed when
between 2.5 and 3 equivalents of aniline (entries 2–5 vs. 6–8)
are employed, and a second shift on moving to 6 equivalents,
suggesting significant changes to the structure of the active species
occur at these points—possibly from amido to imido, or from
mono- to bis-amido/imido tungsten species. Given the assumption
that at least some of the aniline added is behaving not as a ligand
precursor, but merely as a base scavenging HCl, the use of the
tertiary amine bases triethylamine and dabco in conjunction with
there is an economic penalty with both these methods due
to the cost of aluminium reagent or of sparging under reflux
conditions, which is energetically expensive in chlorobenzene. To
surmount these issues the incorporation of base into this catalyst
system was explored, in varying ratios, and in combination with
differing amounts of aniline (see Table 1). The use of a soluble,
organic, low cost base was imperative, thus triethylamine and 1,4-
diazabicyclo[2.2.2]octane (dabco) were examined. Table 1, entry
1, reveals that WCl
6
alone does not possess an inherent activity
towards olefin dimerisation, no conversion being observed in this
7
026 | Dalton Trans., 2010, 39, 7025–7037
This journal is © The Royal Society of Chemistry 2010

View Article Online
a
Table 2 Effect of aluminium activator on a-olefin dimerisation
Branching Selectivity
Lin
b
c
d
Entry
Al activator (eq)
Act
TON
Conv (%)
S
(%)
MB
DB
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
AlCl
AlCl
AlCl
EADC (5)
EADC (15)
EADC (30)
EASC (5)
EASC (10)
EASC (15)
EASC (30)
AlEt
AlEt
AlEt
AlEt
MAO (10)
MAO (25)
MAO (50)
MAO (100)
MAO (150)
EtOAlEt
EtOAlEt
EtOAlEt
3
3
3
(5)
(15)
(30)
—
—
—
—
21
32
6
25
14
12
—
—
—
—
—
—
—
—
—
—
—
—
—
4
—
—
—
—
427
636
110
496
272
238
—
—
—
—
—
—
—
—
—
—
—
—
—
73
35
—
10
0
—
0
0
—
—
—
—
1.1
0.8
0.5
0.4
0.4
0.3
—
—
—
—
—
—
—
—
—
—
—
—
—
2.8
3.9
—
4.9
1.6
—
—
0.6
—
—
—
—
58.3
54.2
62.9
61.0
60.3
59.5
—
—
—
—
—
—
—
—
—
—
—
—
—
63.4
63.5
—
55.8
56.6
—
—
—
—
—
40.5
45.0
36.6
38.6
39.3
40.2
—
—
—
—
—
—
—
—
—
—
—
—
—
33.9
32.6
—
35.8
38.3
—
0.5
0.6
0.5
42.7
63.6
11.0
49.6
27.2
23.8
0
0
0
0
0
0
98.4
91.3
99.3
98.8
98.3
91.3
—
—
—
—
—
0
0
0
0
—
—
0
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
3
3
3
3
(5)
(10)
(15)
(30)
0.5
0.7
0.8
0.7
0
2
2
2
(5)
(15)
(30)
0
0.5
0.5
7.3
3.5
0.5
1
13.8
0.5
3.8
8.1
(Me{Cl}Al)
(Me{Cl}Al)
(Me{Cl}Al)
2
2
2
O (5)
O (15)
O (30)
94.5
85.3
0
96.5
96.5
0
2
(Et{Cl}Al)
(Et{Cl}Al)
(Et{Cl}Al)
2
2
2
O (5)
O (15)
O (30)
—
0.5
7
—
2
138
—
38
(iBu{Cl}Al)
(iBu{Cl}Al)
(iBu{Cl}Al)
2
2
2
O (5)
O (15)
O (30)
0
89.3
—
57.3
—
31.4
4
81
a
◦
General conditions: 0.067 mmol WCl
6
, 0.13 mmol PhNH
2
, 0.27 mmol Et
3
N, Al activator, PhCl (8 mL), catalyst formation period (30 mins @60 C),
◦
b
-1
-1
c
-1
d
nonane standard (1 mL), catalysis temperature 60 C, 1-hexene (67 mmol), 20 h. (mol 1-C
dimer fraction.
6 6
)(mol W) hr . (mol 1-C )(mol W) . Selectivity to the
aniline were examined. It can be seen that triethylamine gave a
more productive and active catalyst as compared to that obtained
using an analogous loading of tertiary amine via dabco (entries
Previous literature reports have employed an elevated temper-
7
ature during the catalyst formation period (T
F
), so to probe the
◦
necessity of this a set of tests were undertaken with a T
F
of 20 C
1
0 and 14 vs. 11 and 15), although using dabco favoured the
(Table 1, entries 17–23). As can be seen, the observed activities and
selectivities are relatively insensitive to the wide variance in aniline
and triethylamine loading, suggesting that complete formation of
production of methyl-branched dimers. Comparing 2 equivalents
of aniline with 1 equivalent each of aniline and triethylamine
◦
◦
(
entry 4 vs. entry 10) reveals an increase in activity, but a drop in
the pre-catalyst species achieved at 60 C does not occur at 20 C.
selectivity to monomethyl-branched dimer. However, comparing
entries 7 and 11 (4 equivalents of aniline versus 2 equivalents
aniline and 2 equivalents of triethylamine) reveals an increase in
activity and doubling of productivity, but a remarkably similar
branching selectivity. Finally, comparing the system generated
using 6 equivalents of aniline (entry 8) with that using 2 equivalents
of aniline and 4 of triethylamine (entry 12), or 3 equivalents
of both (entry 13), reveals a drop in activity, and a swing away
from dimethyl-branched dimers to monomethyl-branched. Thus
it is difficult to draw general conclusions regarding the effect of
triethylamine. For entry 12, a mass balance was performed against
an internal standard (nonane) and this revealed a selectivity to
the dimer fraction (versus the formation of heavy products) of
In order to probe the other extreme, a run was undertaken with
◦
T
F
and T
C
at 100 C in an autoclave (Table 1, entry 24). This run
gave the lowest selectivity to methyl-branched dimers observed
and showed half the productivity of the equivalent run with T
F
◦
and T
C
of 20 C (entry 23). For the purposes of comparison
with the literature, a run was conducted under the conditions
7b
◦
employed by Hendrikson (entry 25); namely a T
F
of 132 C
(chlorobenzene at reflux) for 30 min with N sparging. As can
2
be seen this gave a poorly active catalyst as compared with the
protocols employed herein, and a lower selectivity to methyl-
branched dimer. Significantly, the selectivity to the dimer fraction
in this case, was found to be low at 60% (cf . entry 12, >95%).
The penultimate vector probed was the choice of aniline
(Table 1, entries 26–28), as a previous report using W-imido
complexes for propene dimerisation had clearly shown a rate
enhancement when the aniline ring was substituted by electron
withdrawing groups, and conversely a rate reduction for electron
9
5.3%, which is high compared to most catalysts available for
the dimerisation of a-olefins. The heavy products observed were
largely trimers and tetramers, with trace pentamers occasionally
present. No polymer products were ever observed.
This journal is © The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 7025–7037 | 7027

View Article Online
a
Table 3 Effect of aluminium activator on a-olefin dimerisation
At 30 min
At 300 min
Branching selectivity
b
c
d
b
c
d
Entry Al activator (eq)
Act
TON
Conv (%) S (%) Act
TON Conv (%) S (%) Linear
Me
Et MeMe MeEt EtEt
1
2
3
4
5
6
7
8
EADC (12)
EADC (12)
EASC (6)
EASC (12)
DEAC (12)
MADC (12)
EADB (12)
EADB (20)
1025 513
1019 510
57.3
57.8
23.7
39.7
1.9
3.0
8.6
>99.9 129
646
642
298
503
163
510
92
72.3
72.8
34.4
57.3
5.1
48.4
10.4
7.7
>99.9 0.2
>99.9 0.2
57.3
57.4
63.0
60.5
69.2
50.4
—
0
0
0
0
0
0
—
0
42.5
42.4
37.0
39.3
30.3
49.2
—
0
0
0
0
0
0
—
0
0
0
0
0
0
0
—
0
>99.9 128
e
410
696
59
205
348
29
69.7e
84.3
31.8
95.5
0
60
99.6
99.6
96.8
78.5
0
0
101
33
102
18
0.2
0.5
0.4
—
0.0
65
32
153
133
76
67
7.7
42.5
133
67
42.5
67.9
32.1
a
◦
General conditions: 0.1 mmol WCl
6
, 0.2 mmol PhNH
2
, 0.4 mmol Et
3
N, Al activator, PhCl (10 mL), catalyst formation period (30 mins @45 C), nonane
◦
b
-1
-1
c
-1
d
standard (1 mL), catalysis temperature 45 C, 1-hexene (~0.1 mol). (mol 1-C
6
)(mol W) hr . (mol 1-C
6
)(mol W) . Selectivity to the dimer fraction.
e
See ref. 12.
8
donating groups. It can be seen that introduction of a p-methyl
making this a very efficient catalyst in terms of substrate usage. The
results shown in Table 2 reveal that when EADC was employed
at low loadings (5 equivalents, entry 4), almost no activity was
observed; whereas EASC gave an active catalyst across the range
5–30 equivalents (entries 7–10). Thus the use of EASC was
examined at 6 and 12 equivalents under these conditions (Table 3).
As can be seen, the higher loading gave the more active catalyst,
but in both cases this was inferior to the EADC-based system,
in terms of both activity/productivity and selectivity to the dimer
fraction. It should be noted at this juncture that an examination of
entries 3 and 4 in Table 3, reveals a significant increase in selectivity
to the dimer fraction between 30 and 300 min. This is an artefact
of the way in which the intermediate samples are taken and worked
group on the phenyl ring leads to a drop in rate of one order of
magnitude, as expected from comparison with previous related
work. Surprisingly, however, the introduction of a p-fluorine
substituent had no effect on catalysis in terms of selectivity (to the
dimer fraction or within the dimer fraction) or activity, suggesting
that a more electrophilic tungsten centre does not accelerate the
rate determining step or influence the regiochemistry of coupling.
Curiously, the p-chloro analogue performed worse than p-methyl,
something that could be attributed to reaction with the aluminium
activator in situ to yield a p-ethyl group, which then behaves in a
8
10
very similar fashion to the p-methyl. Finally, the use of MoCl
with anilineandtriethylamine in place of WCl , was also examined;
however, in these cases less than 2% conversion of the a-olefin
5
6
12
up during catalysis. A comparison of entries 1, 4 and 5 reveals a
clear trend of increasing performance as one moves through the
substrate was observed.
From this initial screening the catalyst system employing WCl
6
,
series Et
2
AlCl, Et
3
Al
2
Cl
3
and EtAlCl , suggesting that only a single
2
PhNH and Et N in a 1 : 2 : 4 ratio was chosen for further study on
2
3
alkyl group is necessary to alkylate the tungsten centre, whilst
maximising the number of halide substituents enhances activity.
An explanation for this may lie with the enhanced Lewis acidity of
the aluminium species with more halide substituents. This elevated
Lewis acidity may in turn be important as it will lead to enhanced
interactions between the N-donor ligand and aluminium activator,
◦
account of good activity at both 20 and 60 C, and its selectivity
towards only methyl- and dimethyl-branched products. The nature
of the aluminium activator is well known to have a dramatic effect
upon the activity and selectivity of many polymerisation and
2-5,7-8,11
oligomerisation processes,
hence a number of alternatives
13
were examined with this system (see Tables 2 and 3). Inspection
of these results reveals a clear trend; the presence of both an
alkyl and a halide group at the aluminium centre are crucial to
such species having been observed experimentally, and suggested
14
computationally.
The use of MeAlCl
2
(MADC) was examined to probe the
catalysis, all of EtAlCl
2
(EADC), Et
3
Al
2
Cl
3
(EASC) and Et
, AlCl , EtOAlEt
2
AlCl
and
necessity of b-hydrogens on the alkyl group in the aluminium
activator. As can be seen from Table 3, entry 6 catalysis was
successfully initiated by MADC, although the performance was
significantly attenuated as compared to EADC. Nonetheless, this
result suggests that a b-hydrogen elimination pathway from a
tungsten alkyl generated in situ is not involved in the mechanistic
trajectory followed during catalyst activation. Given the obser-
vation that the presence of a halide in the aluminium activator
compound is essential, and the proclivity of such species to form
(
DEAC) giving active systems, whilst AlEt
3
3
2
methylaluminoxane (MAO) all failed to promote dimerisation.
Further substantiation of this conclusion comes from the use of
i
{
R(Cl)Al}
2
O (R = Me, Et, Bu) based activators, which also gave
catalysis, but were less active and productive than the simple alkyl
aluminium halide species.
Table 3 shows some further results with the optimal activators
(
entries 1–5), along with variations chosen to further probe the
15,16
constitutional requirements of the halide and alkyl substituents
at aluminium (entries 6–8). Entries 1 and 2 illustrate the typical
reproducibility of the catalysis reported herein; a variability
in branching selectivity of ±0.1% is representative, whilst the
variability in activity and turnover number (TON) is small enough
to be within the experimental error for these particular two
halide bridges with themselves and active catalyst species,
the impact of the nature of the halide on activity was probed
by investigating EtAlBr (EADB, entries 7–8). The effect was
2
dramatic; the resultant catalyst being far less active and much
less selective to the dimer fraction, suggesting that the halide
is indeed intimately involved in the active catalytic species, and
supporting a similar proposition resulting from computational
examples. It can be seen that under these conditions (T
F
= T =
C
◦
14
4
5 C) the selectivity to the dimer fraction with EADC is >99.9%,
studies.
7
028 | Dalton Trans., 2010, 39, 7025–7037
This journal is © The Royal Society of Chemistry 2010

View Article Online
a
Table 4 Effect of solvent choice upon a-olefin dimerisation
b
c
d
Entry
Solvent
Act
TON
S (%)
1
2
3
4
5
6
Toluene
Xylenes
s-Butylbenzene
t-Butylbenzene
Cumene
35.3
30.3
21.9
19.8
25.6
30.5
35.3
30.3
21.9
19.8
25.6
30.5
82.8
64.0
22.7
41.0
73.7
79.5
Heptane
a
General conditions: 0.1 mmol WCl
6
2 3
, 0.2 mmol PhNH , 0.4 mmol Et N,
EADC (1.2 mmol), solvent (10 mL), catalyst formation period (30 mins
◦
◦
@
60 C), nonane standard (1 mL), catalysis temperature 60 C, 1-hexene
b -1 -1 c -1
(
0.1 mol), 60 mins. (mol 1-C
Selectivity to the dimer fraction.
6
)(mol W) hr . (mol 1-C
6
)(mol W) .
d
The catalysis described thus far was conducted in chloroben-
zene, but the ability to use the most benign solvent possible for
catalysis is always desirable, thus the effect of different reaction
solvents was explored. Initial attempts to use pure hydrocarbon
solvents were unsuccessful (Table 4), largely due to the lack of sol-
ubility of the catalyst system. Thus a set of strong donor solvents
Fig. 2 Branching selectivity obtained with different olefin substrates (see
Table 6 for experimental details).
an examination of the data at 30 min reveals the expected trend of
decreasing activity as substrate molecular weight increases (Table 3
and Fig. 3). However, there are also some distinct anomalies;
it is noted that 1-nonene is converted particularly slowly and
that the selectivity to the dimer fraction is very poor (entry
4). This is rationalised as follows: 1-nonene is not substantially
different from its dimerisation product in terms of inherent activity
towards oligomerisation, and thus there is little differentiation
by the catalyst between binding and oligomerising 1-nonene or
its C18 dimer product, which is primarily an a,a¢-disubstituted
olefin. Perhaps a more significant anomaly is the lack of further
1-pentene conversion after 30 min. Despite a very fast initial
rate of reaction with 1-pentene being observed, this is followed
by complete catalyst deactivation, consistently and repeatedly.
This is in part rationalised as being due to poor solubility of
the catalyst in the chlorobenzene–1-pentene solvent mixture. To
investigate this, a run was performed at a higher dilution and with
much higher 1-pentene loading (under the conditions used for
Table 6, entry 1, the catalyst was 4.6 mM, in a total volume of
(
THF, PhOMe and MeCN) was explored (see Table 5, entries 2, 3
and 4). Despite very good solubility of the catalyst in these media,
only poor activity was observed, most probably due to competition
between the substrate olefin and the solvent for binding at the
metal centre. In order to probe this donor solvent inhibition
further, a run was performed using chlorobenzene spiked with
0
.5% by volume N,N,N¢,N¢-tetramethylethylenediamine (tmeda);
here again catalysis was effectively shutdown (entry 11). Catalysis
in chlorinated alkane solvents did proceed, but did not show
any improvement over chlorobenzene (entries 5–8). In DCM,
only slightly attenuated activity and productivity as compared
to chlorobenzene was observed. The selectivity to the dimer
fraction remained unchanged, whilst the selectivity within the
dimer fraction was perturbed towards dimethyl-branched dimers
(
entry 5). Reactions undertaken in chloroform showed a very
high activity, but extremely low selectivity to the dimer fraction
entry 7). Chlorinated ethanes showed a good selectivity to the
(
dimer fraction at 30 min, but after extended catalysis periods this
dropped significantly, suggesting that catalyst decomposition had
occurred (entries 6 and 8). Finally, the use of chlorinated additives
was examined, such species having been utilised in the Phillips
trimerisation system, where they are believed to interact with
21.6 mL of PhCl : 1-C
was 0.6 mM in a total volume of 618 mL of PhCl : 1-C
5
{1 : 0.73}, whereas in entry 5 the catalyst
5
{1 : 2.085}),
and under these conditions the catalyst does not deactivate after
30 min (entry 5). This unusual behaviour with 1-pentene remains
without good explanation, as when the volume of chlorobenzene
was dramatically increased, without also increasing the 1-pentene
loading, catalyst deactivation was still seen after 30 min, indicating
solubility alone was not the issue.
16
Et
Hexachlorobenzene had no effect within experimental error (entry
), whilst hexachloroethane was detrimental, especially at longer
reaction times, again confirming this trend for chlorinated ethanes
x
AlCl(3-x) type activators altering the nuclearity in solution.
9
(
entries 10, and 6, 8).
The catalysis described thus far, has variously used 1-pentene,
-hexene and 1-heptene substrates under slightly different reaction
Two runs were also performed at increased loadings of 1-heptene
whilst keeping the chlorobenzene solvent volume constant; these
gave higher activities and productivities as expected (entries
6 and 7), although upon inspection entry 7 does not appear
proportionately more active than entry 6. This observation is
assigned to catalyst poisoning by trace impurities in the 1-heptene
substrate, which at such high substrate loading begin to overwhelm
the active species. Notably, with both entries 6 and 7, the selectivity
to the dimer fraction and the branching selectivity remained
unchanged compared to entry 3.
1
conditions, hence it was of interest to make a meaningful
comparison under the same conditions, of the differences as the
a-olefin chain length increases (C
tabulated in Table 6, and shown graphically in Fig. 2–4. It is noted
that the catalyst behaves fairly uniformly in terms of branching
selectivity within the dimer fraction (Table 3, Fig. 2); there
being a trend towards increased monomethyl-branched product
formation as the olefin substrate mass increases, reaching a
maximum at 1-heptene, after which it subsides. In terms of activity,
5
–C ). The results of this are
9
In terms of longevity of the active catalyst, the reactions
described by Table 6 entries 2–7 were all allowed to run for a
This journal is © The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 7025–7037 | 7029

View Article Online
a
Table 5 Impact of reaction of solvent upon a-olefin dimerisation
At 30 min
At 240 min
Branching selectivity
b
c
d
b
c
d
Entry Solvent
Act
TON
Conv (%) S (%) Act
TON Conv (%) S (%) Linear
Me
Et MeMe MeEt EtEt
1
2
3
4
5
6
7
8
9
1
1
PhCl
THF
662
0
0
0
562
356
1923
838
553
328
331
0
0
34.6
0
0
99.7
—
—
149
0
0
594
0
0
142
574
671
975
750
597
772
62.2
0
0
15.5
61.6
71.7
97.9
78.5
61.5
80.1
>99.9 >0.1
62.1
—
—
0
37.9
—
—
0
0
—
—
0
—
—
—
—
—
—
0
0
—
0
0
0
0
—
—
—
0
0
—
0
0
0
0
—
—
—
0
0
—
0
0
0
0
e
PhOMe
MeCN
DCM
1,1,2,2-C
CHCl
1,2-C
PhCl + 2 g C
PhCl + 2 g C
PhCl + 0.5% tmeda 41
0
0
—
33
—
—
281
178
961
419
277
164
20
30.1
19.0
96.6
43.9
28.5
17.0
2.1
>99.9 133
94.7
5.3
>99.9 >0.1
55.4
69.6
—
62.4
62.2
75.9
64.0
44.6
30.4
—
37.6
37.8
24.1
36.0
2
H
2
Cl
4
155
244
14.6
5.0
22.3
>0.1
—
>0.1
3
2
H
4
Cl
2
>99.9 187
6
2
Cl
Cl
6
99.1
90.0
78.8
149
193
>99.9 >0.1
0
1
6
12.4
>0.1
>0.1
No further conversion
a
◦
General conditions: 0.2 mmol WCl
6
, 0.4 mmol PhNH
2
, 0.8 mmol Et
3
N, EADC (2.4 mmol), solvent (20 mL), catalyst formation period (30 mins @40 C),
◦
b
-1
-1
c
-1
d
nonane standard (2 mL), catalysis temperature 40 C, 1-heptene (0.2 mol), 240 mins. (mol 1-C
the dimer fraction. 1 mmol WCl
mol), 240 mins.
7
)(mol W) hr . (mol 1-C
7
)(mol W) . Selectivity to
N, EADC (12 mmol), PhOMe (200 mL), nonane standard (10 mL), 35 C, 1-pentene (1
e
◦
6
, 2 mmol PhNH
2
, 4 mmol Et
3
a
Table 6 The dimerisation of a-olefins with increasing molecular weight
At 30 min
At 360 min
Branching selectivity
b
c
d
b
c
d
Entry Olefin (eq)
Act
TON Conv (%) S (%) Act
TON Conv (%) S (%) Linear
Me
0
Et MeMe MeEt EtEt
0
e
1
1-Pentene (834)
1-Hexene (885)
379
183
189
22.7
15.2
No further >0.1 57.7
conversion
0
42.3
e
2
3
4
5
6
7
92
167
28
10.4
18.2
2.9
77.4e
85.7
20.6e
86.3
99
94
5
594
563
28
67.1
61.8
2.8
>99.9 >0.1
>99.9 >0.1
30.5
99.8
>99.9 >0.1
>99.9 >0.1
63.1
64.4
59.8
58.2
64.3
64.6
0
0
0
0
0
0
36.9
35.6
40.2
41.8
35.7
35.4
0
0
0
0
0
0
0
0
0
0
0
0
1-Heptene (920) 84
1-Nonene (972) f 14
1-Pentene (3818) 136
1-Heptene (2500) 433
1-Heptene (5000) 746
>0.1
>0.1
68
217
373
1.8
8.7
7.5
112
199
213
673
17.6
e
86.4e
1193 47.7
1279 25.6
90.5
a
◦
General conditions: 0.1 mmol WCl
6
, 0.2 mmol PhNH
2
3
, 0.4 mmol Et N, EADC (1.2 mmol), PhCl (12.5 mL), catalyst formation period (30 mins @30 C),
b -1 -1 c -1 d
◦
nonane standard (1 mL), catalysis temperature 20 C, a-olefin, 360 mins. (mol olefin)(mol W) hr . (mol olefin)(mol W) . Selectivity to the dimer
e
f
◦
fraction. See ref. 12 0.4 mmol WCl
nonane standard (1 mL), catalysis temperature 20 C, a-olefin, 360 mins.
6
, 0.8 mmol PhNH
2
, 1.6 mmol Et
3
N, EADC (4.8 mmol), PhCl (200 mL), catalyst formation period (30 mins @30 C),
◦
Fig. 3 TON’s obtained with different olefin substrates (see Table 6 for
Fig. 4 The extent of isomerisation of the substrate olefin (see Table 6 for
experimental details).
experimental details).
duration of 24 h. The catalyst remained active throughout this
time, but little further conversion occurred after the initial 6 h.
Considering the kinetics of dimerisation reactions, one would
expect catalysis of this type to be higher order in substrate, and
thus show a markedly reduced activity once high conversions
be apparent when kinetic analyses are applied. However, one more
vector must be considered, namely isomerisation of the a-olefin
substrate to internal olefins. The catalyst shows isomerisation
activity towards the a-olefin substrate to give a mixture of
terminal and internal olefins, but appears capable of dimerising
only a-olefins (vide infra). Hence, the effective concentration of
a-olefin substrate will be diminished not just by dimerisation,
17
are reached. However, catalyst decomposition is potentially a
further contributing factor to this type of behaviour and should
7
030 | Dalton Trans., 2010, 39, 7025–7037
This journal is © The Royal Society of Chemistry 2010

View Article Online
but by isomerisation processes as well. We thus undertook to
determine the order of reaction in substrate, and attempt to
deconvolute the effects of substrate isomerisation and catalyst
deactivation, the results being depicted graphically in Fig. 5. The
first reaction examined was the larger scale dimerisation of 1-
pentene (Table 6, entry 5), because this reaction was sampled with
comparable reaction conditions. These values follow the trend
expected, namely that as the substrate olefin increases in mass, its
inherent reactivity towards oligomerisation decreases. As can be
seen from Fig. 5a and 5b, once again the concentration of a-olefin
rather than that of all isomers of the substrate is the crucial factor,
and when this is accounted for, no signs of catalyst deactivation
during catalysis are evident. Lastly, the reactions with elevated
loadings of 1-heptene were examined (Table 6, entries 6 and 7),
increased frequency during catalysis. For plots of [1-C
5
], ln([1-
C
5
]
t
/[1-C ) and 1/[1-C ] versus time, only the latter showed an
5
]
0
5
-
5
-3
-1 -1
approximately linear relationship, indicative of a second order
dependence on the substrate. However, as can be seen from Fig. 5a,
there was a distinct deviation from linearity, especially after 6 h. It
should be noted that the substrate concentration values calculated
yielding second order rate constants of 1.56 ¥ 10 dm mol s
5
-
-3
-1 -1
and 2.10 ¥ 10 dm mol s , respectively. Fig. 5b shows that
for 2356 equivalents of 1-heptene only a very slight deviation
from linearity exists, whereas for 4728 equivalents of 1-heptene,
a more marked deviation is present, strongly suggesting that at
this substrate loading, catalyst poisoning is starting to become
for the plots in Fig. 5a were based upon the entire C fraction, i.e.
5
both unreacted and isomerised substrate. Thus, as the extent of
isomerisation was quantified from the GC samples (see Fig. 4), the
data was reprocessed based upon the concentration of 1-pentene
specifically, as this is believed to be the only true substrate. Indeed,
as can be seen from Fig. 5b, a linear plot was now obtained for
18
significant (vide supra).
The finding that the WCl
6
: PhNH
2
: Et N (1 : 2 : 4) catalyst
3
system shows some activity towards isomerisation of the feed
olefin (Fig. 4) is significant from a process perspective, as recycling
of unreacted olefin feed is greatly facilitated if it remains un-
isomerised. A further question is whether the catalyst is able to
process olefin substrate once it has been isomerised to internal
1
/[1-C ], which allowed derivation of a second order rate constant
5
-
5
-3
-1 -1
of 2.07 ¥ 10 dm mol s . It should be noted that a very small
deviation from linearity remains, which is therefore indicative of
trace catalyst decomposition.
isomers. The branching selectivity with C
5
–C a-olefin substrates
9
reveals no formation of branches longer than methyl which
suggests that internal olefins are not converted. However, in
order to probe this definitively, it was decided to investigate the
behaviour of the catalyst system towards internal olefins. Thus,
in six separate experiments, fresh batches of in situ prepared
catalyst were exposed to cis-2-pentene, trans-2-pentene, cis-2-
hexene, trans-2-hexene, cis-2-heptene and trans-2-heptene. In each
case, no dimerisation activity or isomerisation of the double bond
position was observed, although in the case of cis-2-pentene
and cis-2-hexene, there was a small degree of isomerisation in
the conformation of the double bond from cis to the more
thermodynamically stable trans isomer (~ 15%).
Whilst isomerisation of the relatively light substrate olefins
occurs, it was strongly suspected that the catalyst system had a very
low propensity towards isomerisation of the product olefins. This
inference is based upon the observation that the major products are
the mono- and di-branched methylidene species (vide infra), which
as terminal olefins are thermodynamically disfavoured. Thus, if
isomerisation processes were facile for this catalyst, these products
would not be isolated. However, to check this assertion, in two
separate experiments fresh batches of in situ prepared catalyst
were exposed to 4-methyl-4-nonene and 2-propyl-1-heptene. In
both cases no isomerisation was observed, however with the 2-
propyl-1-heptene (an a,a-substituted olefin) a small amount of
dimerisation did occur (17.8% conversion after 4 h).
Product identification
From the catalysis performed we have clearly identified the
selectivity towards dimerisation, and also identified the branching
selectivity within the dimer fraction. The latter parameter was
determined via hydrogenation of the product mixture using a
hydrogenative GC method (see experimental section), although
selected samples were also hydrogenated on a larger scale using
Pd/C to confirm the veracity of the hydrogenative GC technique.
The correlation between the results obtained by the two different
methods was always in good agreement (< ±0.1%). A typical GC
Fig. 5 (a) Plots of 1/[substrate] versus time for data from Table 6. (b)
Plots of 1/[substrate] versus time for data from Table 6 after correction of
[
substrate] by removal of the contribution from internal olefins.
A similar analysis was performed on the data presented
in Table 6, entries 2 and 3, in order to determine second
order rate constants for the dimerisation of 1-hexene (1.28 ¥
-
4
-3
-1 -1
-5
-3
-1 -1
1
0 dm mol s ) and 1-heptene (4.97 ¥ 10 dm mol s ) under
This journal is © The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 7025–7037 | 7031

View Article Online
trace of the product after hydrogenation can be seen in Fig. 6; the
double peak for 4,5-dimethyloctane arises due to separation of
the diastereoisomeric pairs present for this compound. As the EI
mass spectra for different isomers of heavy alkanes can be very
similar, identification via GC-MS alone can leave uncertainty,
hence the alkanes were also identified via comparison with
authentic standards, which were synthesised where necessary (4-
methylnonane, 4,5-dimethyloctane).
Fig. 7 The C10 region of a typical GC-trace from 1-pentene dimerisation
using WCl : PhNH : Et N : EADC (1 : 2 : 4 : 12).
6 2 3
more carefully for the alkane corresponding to the monomethyl-
branched skeleton (4-methylnonane), as the formation of one
without the other seemed curious. Based on retention times,
the GC methods available co-eluted 4-methylnonane with trans-
2
-propyl-1,4-heptadiene (vide infra), but from simple GC-MS
analysis this peak was known to be the diene. However, extractive
ion trace analysis (for m/z’s 138, 140 and 142 ±0.5) of the GC-
MS results revealed that underlying this was a trace amount of
alkane, but at such a small level as to be normally undetectable
and below the limits of accurate quantification (ꢀ 0.1%). When
catalysis was performed on a larger scale with 1-pentene (Table 6,
entry 5) we subjected the product mixture to an automated
distillation procedure, which allowed significant fractionation of
the products, some cuts being highly enriched in trace compounds.
These fractions were subjected to an NMR study and this led to the
identification of the previously unspecified diene products as cis-
and trans-2-propyl-1,4-heptadiene, which constitute ~1% and ~2%
of the dimer fraction respectively. It will be noted from Fig. 7 (data
obtained using a GC instrument equipped with a PONA column),
that the peaks for trans-6-methyl-4-nonene and 2-propyl-3-methyl-
1-hexene overlap. The amounts of these two products could be
determined via one of two methods; GC using a MDN-12 column
allowed separation of these two peaks, whilst hydrogenative GC-
analysis (see Fig. 6) allowed the overall branching profile to be
determined, and thus the relative amounts of these two products
to be deconvoluted. Generally, both these methods were applied
to provide a double check, and in all cases were in good agreement
(±0.1%).
Fig. 6 The C10 region of a typical GC-trace from 1-pentene dimerisation
using WCl : PhNH : Et N : EADC (1 : 2 : 4 : 12) after hydrogenation.
6 2 3
The further piece of analysis required is the selectivity towards
specific olefin isomers within the variously branched skeletons.
As with heavy alkanes, the rigorous identification of various
heavy alkene isomers can be troublesome via routine GC-MS
techniques due to the similarity of the compounds. This also
hinders NMR analysis, which is additionally complicated by the
complexity of the product mixture. Thus in order to probe the
products of 1-pentene dimerisation, and knowing the carbon
skeletons present, we undertook the synthesis of the relevant
isomers of C10H20 that could not be obtained commercially.
Specific focus was upon the 4-methylnonene skeleton, all nine
isomers proving accessible via either Wittig synthesis or lithium-
cuprate-based coupling (see ESI† for experimental details). A GC
methodology was established that allowed the separation of all
isomers, and hence, via comparison of GC traces of product
mixtures and authentic samples it was possible to elucidate the
identity of the majority of the olefin products. However, the
second most abundant product did not show a GC retention
time match with any 4-methylnonene standards and was known
from hydrogenation to be a dimethyl-branched species. It was
suspected to be a methylidene by comparison with the methyl-
branched skeleton and its identity as 2-propyl-3-methyl-1-hexene,
One final piece of analysis concerned identification of the trace
amounts of linear C10 species formed. This was performed on both
normal strength samples and those enriched in linear C10 products
after distillation. The analysis via both methods was in agreement,
but the quantification is taken from the enriched sample analysis.
1
3
1
was established from a C { H} NMR experiment of the product
mixture, resonances at d157.7 and 108.2 being characteristic, and
observed in addition to those for 3-propyl-1-heptene (d150.1 and
All linear isomers of C10H20 with the exception of cis-4-decene were
1
09.6). At this stage three trace products remained unassigned.
commercially available and hence via retention time matching, the
formation of 1-decene, cis/trans-2-decene and cis/trans-3-decene
could be eliminated. Five peaks were observed in the linear C10
region of the GC chromatogram, and GC-MS analysis revealed
one of these to be decane (also confirmed by retention time
match), two as linear C10 dienes and two as linear C10 olefins.
As a percentage of the linear C10 fraction (which comprises only
0.1% of the total dimer fraction), the components were: trans-5-
decene (20.1%), cis-4-decene (40.4%), diene (16.5%), diene (8.7%),
GC-MS analysis of the mixture proved to be decisive in identifying
these peaks, unambiguously confirming one to be a C10 alkane and
the other two as C10 dienes. By correlation of the numerical data
before and after hydrogenation, the C10 alkane was assigned as
4
,5-dimethyloctane (Fig. 7). This was verified by a retention time
match with an authentic sample and an EI mass spectra match
in the GC-MS. Given the detection of the alkane corresponding
to the dimethyl-branched C10 skeleton, we undertook to look
7
032 | Dalton Trans., 2010, 39, 7025–7037
This journal is © The Royal Society of Chemistry 2010

View Article Online
decane (14.3%). Due to the low concentrations of these species,
it proved impossible to determine the precise constitution of the
linear dienes. Of the two peaks confirmed by GC-MS to be linear
C10 olefins, one gave a retention time match most indicative of
trans-5-decene. However, this assignment is tentative as despite
prolonged method development, it proved impossible to achieve
baseline separation of trans-4-decene and cis/trans-5-decene, thus
there may also be a component of these species in this peak as
well. The other peak identified as a linear C10 olefin did not show
a retention time match with any of the standards available and is
thus assigned, via a process of elimination, as cis-4-decene.
As can be seen from Fig. 7 the catalyst has a clear preference
towards the formation of methylidene products, whether mono- or
dimethyl-branched, with the remainder featuring internal unsatu-
ration on the chain. The notable absence of the thermodynamically
most stable, tri- and tetra-substituted alkene isomers (4-methyl-
3
-nonene and 4-methyl-4-nonene) again illustrates the lack of
isomerisation activity of this catalyst towards the product. When
catalysis was performed with 1-hexene, 1-heptene or 1-nonene
as substrates, an identical GC fingerprint was obtained for the
dimer products, but was shifted to a higher retention time (due to
increased carbon number), suggesting that the catalysis selectivity
remains unchanged. The observation of alkane and diene products
may at first seem surprising, but their formation during a-olefin
19
dimerisation is not without precedent. Comparison with the
previous report (19% C alkane, 28% C diene), reveals that the
8
8
levels of formation seen here are significantly lower (0.5% C10
alkane, C10 3.0% diene), but that in both cases the amount of
alkane observed is in deficit compared to the amount of diene.
It would be expected however that the amounts of diene and
alkane should correlate quite closely, unless the excess hydrogen
20
is lost as H
2
. In agreement with this, inspection of the liquid
fraction in this case reveals the presence of hydrogenated substrate
21
at a level matching that expected, suggesting that equimolar
amounts of dienes and alkanes are formed. For example, absolute
amounts present at the end of a run: a) pentane (0.277 mmol), 4,5-
dimethyloctane (0.059 mmol), C10 diene (0.213 mmol), C10 diene
(
0.099 mmol), total alkane (0.336 mmol), total diene (0.312 mmol);
b) pentane (0.593 mmol), 4,5-dimethyloctane (0.126 mmol),
10 diene (0.438 mmol), C10 diene (0.237 mmol), total alkane
C
(
0.719 mmol), total diene (0.675 mmol).
In order to probe if there was a temporal dependence of the
selectivity to specific olefins, a 1-pentene dimerisation experiment
was sampled over time; as can be seen from Fig. 8a and 8b,
the catalyst selectivity remains constant. Fig. 8c shows the
accumulation of the products in mmols and confirms that catalysis
continued to proceed throughout the sampling run, and that the
constant selectivity was not due to a deactivated catalyst.
Fig. 8 (a) Product selectivity with time during a 1-pentene dimerisation
using WCl
6
: PhNH
2
: Et
3
N : EADC (1 : 2 : 4 : 12). (b) Expansion of ordi-
22
22
nate to show the levels of minor products more clearly. (c) Absolute
amount of olefins formed with time.
2b,24
type
mechanistic pathways. A metallacyclic pathway is often
invoked for highly selective catalysts and not unreasonable for a
23
group six metal given the precedents with chromium. Indeed,
Mechanistic considerations
Olivier et al. propose a metallacyclic pathway for dimerisation
8
Given the number of possible isomers of C10
cis/trans isomerisation) that could arise from the dimerisation of
-pentene, the primary selectivity of this catalyst to just 8 isomers
plus trace alkanes and dienes) is quite remarkable. In part this is
H
20 (57 including
catalysis with tungsten-imido complexes, although no experi-
mental evidence is available in support. Reference is made to the
isolation by Boncella et al. of an imido-tungstanacyclopentane
complex, which whilst confirming the existence of metalla-
cyclic species for tungsten, was the deactivation by-product
1
(
accounted for by the fact that the catalyst only dimerises terminal
olefins and does not show any isomerisation activity towards the
dimer products. This leads to mechanistic speculation, all of the
25
from metathesis. This, coupled with the ability to isolate and
characterise this species suggests that it is kinetically inert,
resisting breakdown of the metallacycle and hence perhaps not a
23
products observed being accessible via metallacyclic or Cossee-
This journal is © The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 7025–7037 | 7033

View Article Online
Fig. 9 Proposed mechanistic trajectories for the formation of the products observed.
catalytically important intermediate. Nonetheless, computational
studies have suggested that a metallacyclic pathway may be ener-
Considering the experimental evidence obtained in this study,
the observed isomerisation of the substrate olefin can be taken
as indicative of a metal hydride species, which in itself suggests a
14
getically feasible for tungsten-imido based catalysts. It was found
that stepwise metallacycle breakdown via b-hydrogen elimination
and subsequent reductive elimination was a lower energy pathway
than the concerted hydride shift. This could be taken to suggest
that b-hydrogen elimination and reductive elimination should also
be feasible in the form of a Cossee-type pathway, but significantly
a similar analysis of this mechanistic trajectory has not been
reported.
2b,24
Cossee-type mechanism.
Given the lack of isomerisation of the
products by the catalyst, a further potential clue to the mechanistic
pathway lies in a consideration of the products expected from both
metallacyclic and Cossee-type mechanisms. This reveals that the
linear C10 olefins formed should be diagnostic; 3- and 4-decenes
resulting from a metallacyclic pathway and 4- and 5-decenes from
a chain-growth trajectory (see Fig. S3 in the ESI† and Fig. 9,
7
034 | Dalton Trans., 2010, 39, 7025–7037
This journal is © The Royal Society of Chemistry 2010

View Article Online
respectively). As discussed (vide supra), in practice no cis or trans-3-
decene is observed, whilst analysis indicates the presence of trans-
chain undergoes a further b-hydrogen elimination step to yield the
diene.
5
-decene and cis-4-decene, along with possibly traces of cis-5-
Based on the products observed, there is a specificity of cy-
clometallation towards intermediate D only. This can be argued on
the basis that the methyl group is predisposed towards such C–H
activation, and hence no cyclometallation of intermediates A and
B occurs. However, such an argument suggests that intermediate C
should also cyclometallate, as it also features a methyl group on the
g-carbon of the pendant alkyl. The failure to observe any dienes of
the dimethyl-branched skeleton is thus perhaps surprising, but can
be explained in several ways. If cyclometallation of intermediate C
has a slightly higher activation energy as compared to D, then diene
formation may still occur, but the levels of diene formation could
be pushed below the limit of detection by GC. Alternatively, if
b-hydrogen elimination from intermediate C is significantly faster
than from D, then cyclometallation of C may simply be kinetically
disfavoured.
decene and trans-4-decene. It is thus suggested that catalysis with
this in situ tungsten-based catalyst system proceeds via a Cossee-
type mechanism, as depicted in Fig. 9.
In order to obtain a second verification of this result the in situ
catalyst system was screened with a mixture of C
the now well established isotopomer experiment for the discrimi-
nation of metallacyclic and Cossee-type mechanisms. Analysis of
the products from a dimerisation reaction using C /C (3.5 : 1
ratio) revealed 1-butene as the primary product (73.7 mol%) with
traces of internal butenes (2.5 mol%), methylpentenes (22.8 mol%)
2
H
4
/C
2
D
4
using
23
2
H
4
2
D
4
26
and linear internal hexenes (1.0 mol%). Extractive ion trace GC-
MS analysis reveals full isotopic scrambling of all the C and
products, which is indicative of the Cossee-type mechanism.
4
27
C
6
Thus, this experiment serves to confirm the mechanistic trajectory
indicated by analysis of the 1-pentene dimerisation experiments,
namely that a Cossee-type mechanism is in operation.
Considering this mechanism, and the observed levels of forma-
tion of the various products, it can be concluded that the first
olefin insertion step is almost equally likely to proceed in a 1,2
or 2,1 fashion but that there is an element of regioselectivity in
the subsequent insertion. Given the level of linear products versus
At this stage a consideration of alkane formation is timely, and
given the good correlation between diene and alkane formation
levels, the loss of the diene-derived-hydrogen as dihydrogen is
considered unlikely. This necessarily invokes a tungsten trihydride
species (Fig. 9, intermediate G), which is not unfeasible, tungsten
29
polyhydride species being well known, with tungsten-imido
30
polyhydride complexes specifically being reported. Given the
suspicion of imido functionality in the active catalyst species,
the formation of W–H species could be transient, with onwards
reaction to give a tungsten-amido species. This amido species
could then perform the hydrogenation, in the process regenerating
the imido moiety. Such a species could then readily hydrogenate
an olefin moiety via a normal coordination, insertion, reductive
elimination pathway. Indeed the observation of trace levels of 4,5-
dimethyloctane and decane fits with this hypothesis, but the very
low level of 4-methylnonane remains without good explanation.
However, the observation that the substrate olefin is the most
hydrogenated species is to be expected, being the least sterically
hindered olefin present and in the highest concentration for the
majority of the experiment.
2
-propyl-1-heptene (0.1% versus 41.9%), it is suggested that after
an initial 1,2 insertion, a further 1,2 insertion is ~420 times more
likely than a 2,1 insertion. Also, given the observation of 40.1% of
2
-propyl-3-methyl-1-hexene and in total 14.5% of 6-methyl-3/4-
nonene and 2-propyl-1,4-heptadiene, then following an initial 2,1
insertion, the subsequent insertion is ~2.8 times more likely to
occur in a 1,2 rather than a 2,1 fashion. These statistics indicate
that intermediates B and C, dominate over A and D, suggesting
that the second insertion preferentially occurs to place steric bulk
away from the tungsten centre.
The formation of dienes and alkanes also requires explanation,
especially given the specificity with which this occurs. Herein, we
invoke a C–H activation mechanism akin to that proposed by
19
Small, however there are differences due to the selectivity in this
case, versus the less selective formation of dienes in the previous
work. Most specifically, whilst Small invokes an intermolecular
C–H activation, in this case an intramolecular mechanism seems
most likely. In this case an intermolecular mechanism similar to
that proposed by Small, would lead to the conjugated diene 2-
propyl-1,3-heptadiene, not the 2-propyl-1,4-heptadiene observed.
The latter product could be formed by invoking an isomerisa-
tion of the W-alkenyl, but such an isomerisation would likely
be energetically unfavourable (see ESI† for further discussion).
Indeed, if intermediate D (Fig. 9) were to intramolecularly
cyclometallate the methyl group on the pendant alkyl chain to give
intermediate E, the formation of 2-propyl-1,4-heptadiene only,
and not 2-propyl-1,3-heptadiene, is readily explained. b-hydrogen
elimination from metallacyclopentane species is well documented
to be energetically disfavoured, and thus it is not surprising that
the product resulting from elimination involving the endocyclic
methylene protons (2-propyl-1,3-heptadiene) is not observed.
However, the exocyclic methylene protons are expected to be
significantly more available towards metal-mediated elimination
and would indeed give the observed 2-propyl-1,4-heptadiene.
After, this initial b-hydrogen elimination step, the resulting alkenyl
Conclusions
The ability to control selectivity whilst using an in situ
WCl
6
/PhNH
2
/Et N-based catalyst system via a choice of stoi-
3
chiometry has been demonstrated, allowing the exclusive forma-
tion of methyl and dimethyl-branched dimer products. Through
optimum choice of aluminium activator and solvent, a selectivity
to the dimer fraction of >99% has been demonstrated. The
presence of the chloride in the aluminium species has been found
to be crucial for successful initiation, suggesting the formation
of aluminium-tungsten chloride-bridged species, as has been
substantiated by us and co-workers elsewhere. A kinetic study has
revealed a second order dependence on substrate. The selectivity
of the catalyst with different a-olefins has been shown to remain
remarkably constant, giving methylidene species as the major
products in both the methyl- and dimethyl-branched skeletons.
Extensive product analysis has led to identification of all trace
products from dimerisation catalysis and reveals the additional
formation of dienes and alkanes in equimolar amounts. Following
the identification of trace linear C10 species, it is proposed
that this in situ catalyst operates via a Cossee-type mechanism.
13
28
This journal is © The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 7025–7037 | 7035

View Article Online
A C
2
H
4
/C
2
D
4
co-dimerisation experiment showed full isotopic
ratio between internal standard and olefin initially. The aluminium
activator was then added to initiate catalysis. When samples
were taken during catalysis, they were subjected to a micro-scale
acidified aqueous work-up and analysed by GC-FID as soon as
possible. At the end of catalysis MeOH (1 mL) was added to each
well, followed by 10% HClaq and after initial stirring, the mixture
allowed to separate and a sample of the organic layer taken for
GC-FID analysis. After GC-FID analysis using both PONA and
MDN columns, the samples were subjected to hydrogenative GC-
FID analysis to assist in the assignment of products.
scrambling of the products, and thus supports this mechanistic
proposal. The highly selective diene formation is readily explained
by this mechanism, as is the concomitant alkane formation.
Experimental
General procedures
All operations were conducted under dry nitrogen using standard
Schlenk and cannula techniques, or in a nitrogen-filled glove
box. Solvents were procured from Aldrich, purified using a
Solvent Purification System, and de-oxygenated prior to use.
Given that loss of substrate olefin by evaporation appears
as heavy oligomers when a mass balance is applied, ‘blank’
experiments were conducted to assess the loss of substrate by
evaporation during reaction. This loss was found to be very low
and consistent (< 0.5% per hour during a 6 h period) and was
corrected for in the mass balance.
31
All liquid reagents were dried and deoxygenated prior to use.
-Pentene, 1-hexene, 1-heptene and 1-nonene were stored over
1
˚
activated molecular sieves (4 A), TMEDA was distilled from
Na, aniline, triethylamine and 4-fluoroaniline from CaH
31
◦
.
All
The reaction conducted at 100 C (Table 1, entry 24) was
2
-
1
other chemicals were purchased from Aldrich or Akzo Nobel
and used without further purification unless stated otherwise. 4-
Nonanone was procured from Kodak Eastman. GC standards
were purchased from ChemSampCo (linear decenes, 3-ethyl-4-
methylheptane, 3-ethyloctane and 3,4-diethylhexane).
performed as normal, but in a 50 mL s autoclave with mechanical
stirring; no samples were taken in operando. Larger scale dimeri-
sation reactions were conducted as normal but in 3-neck round-
bottom flasks, equipped with thermometer, magnetic stirrer and
reflux condenser connected to a Schlenk line. The whole assembly
was maintained at constant temperature via a water bath.
NMR spectra were recorded on a Bruker Avance 300 MHz
spectrometer. Chemical shifts are reported in d (ppm) being
referenced to residual protio impurities in the deuterated solvent
The C
autoclave equipped with mechanical stirring and a fluid-filled
jacket for temperature control. The C /C was prepared by
condensing C into a 200 mL bomb, allowing it to warm to
RT and noting the pressure (5.5 bar). The C was then added
to the bomb, to give a total pressure of 25 bar (C /C ratio
3.5 : 1). The catalyst (0.1 mmol WCl , 0.2 mmol PhNH
Et
2
H
4
/C
2
D
4
dimerisation was performed in a 250 mL
1
13
13
1
(
H) or C shift of the solvent. All C { H} NMR spectra were
2
H
4
2
D
4
obtained as DEPT135 acquisitions to assist in assignment. GC-
FID analysis was performed on an Agilent Technologies 6890
N GC system equipped with PONA (50 m ¥ 0.20 mm ¥ 0.50
mm) and MDN-12 (60 m ¥ 0.25 mm ¥ 0.25 mm) columns. GC-
MS analysis was performed on an Agilent Technologies 6890 N
GC system equipped with PONA (50 m ¥ 0.20 mm ¥ 0.50 mm)
and MDN12 (60 m ¥ 0.25 mm ¥ 0.25 mm) columns, coupled
to an Agilent Technologies 5973 N MSD Mass Spectrometric
instrument equipped with EI source. Hydrogenative GC-FID
analysis was performed using an Agilent Technologies 6890 N GC
System equipped with an inlet liner packed with hydrogenating
2
D
4
2
H
4
2
H
4
2
D
4
6
2
, 0.4 mmol
◦
3
N, in 100 mL PhCl) was prepared in situ for 30 min at 45 C.
The vessel was briefly evacuated, then back filled to 1 bar with
premixed C /C . EADC (1.5 mmol, 1.8M in toluene) was
2
H
4
2
D
4
then added and the vessel pressurised to 7 bar with premixed
◦
C
2
H
4
/C
2
D
4
. An exotherm to 47 C was observed. Over the course
of 9 min the pressure dropped to 2 bar, and the reaction was
i
◦
quenched with Pr OH and H O (40 mL), and cooled to 1 C.
2
◦
catalyst (Pt on Chromosorb W at 200 C) and a PONA column
Toluene (100 mL) was added to assist in separation of the organic
and aqueous phases in situ. A sample was taken for GC-FID
and GC-MS analysis, extractive ion trace methods being used to
accurately quantify the amounts of each isotopomer present.
32
(
50 m ¥ 0.20 mm ¥ 0.50 mm). Elemental analyses were performed
by the Science Technical Support Unit of London Metropoli-
tan University. Fractional distillations were conducted using
an automated, custom built Fisher Technology Spaltr o¨ hr HMS
5
00 AC rig.
Acknowledgements
The authors would like to thank Sasol Technology UK Ltd and
Sasol Technology (Pty) Ltd for permission to publish this work. Dr
Philip Dyer (Durham University) is thanked for useful discussions.
Catalysis protocol
Catalysis was conducted in a Radleys 6-well (250 mL) carousel
reactor retrofitted in-house to allow rigorous exclusion of oxygen
and moisture when attached to a Schlenk line. Each well was
magnetically stirred, fitted with a cooled reflux tube and had a
Notes and references
side-arm to allow sampling via septa. The vessels were heated to
1
(a) “Applied Homogeneous Catalysis with Organometallic Compounds”,
2nd Ed., (Eds: B. Cornils and W. A. Herrmann), Wiley-VCH, Wein-
heim, 2002, Chapter 2.3.1.4, H. Olivier-Bourbigou, L. Saussine, 253-
◦
90 C under vacuum for 1 h, before cooling to the desired catalyst
formation temperature and back-filling with N
and nonane (internal standard) were added, followed by WCl
RNH and Et N as stock solutions. The vessel was then stirred
2
. Dry solvent
2
2
64; (b) M. Marchionna, M. Di Girolamo and R. Patrini, Catal. Today,
001, 65, 397–403.
6
,
2
3
2 (a) S. M. Pillai, M. Ravindranathan and S. Sivaram, Chem. Rev., 1986,
at the catalyst formation temperature for the desired period. If
necessary the vessel temperature was then adjusted to the catalysis
temperature. The olefin substrate was added and a sample taken
86, 353–399; (b) J. Skupinska, Chem. Rev., 1991, 91, 613–648.
(a) P. Wasserscheid and M. Eichmann, Catal. Today, 2001, 66, 309–
3
3
3
16; (b) B. Ellis, W. Keim and P. Wasserscheid, Chem. Commun., 1999,
37–338; (c) W. Keim, B. Hoffmann, R. Lodewick, M. Peukert, G.
for GC-FID analysis (t
0
) to allow accurate determination of the
Schmitt, J. Fleischhauer and U. Meier, J. Mol. Catal., 1979, 6, 79–9;
7
036 | Dalton Trans., 2010, 39, 7025–7037
This journal is © The Royal Society of Chemistry 2010

View Article Online
(
d) R. G. Komoto, US4,096,273, Chevron Res., 1978; (e) B. L. Small
occurs, it is immediately obvious as an abnormally low selectivity to
the dimer fraction.
13 W. R. H. Wright, A. S. Batsanov, J. A. K. Howard, R. P. Tooze, M. J.
Hanton and P. W. Dyer, J. Chem. Soc., Dalton Trans., 2010, 39, 7038.
14 S. Tobisch, Dalton Trans., 2008, 2120–2127; S. Tobisch,
Organometallics, 2007, 26, 6529–6532.
15 C. N. McMahon, J. A. Francis and A. R. Barron, J. Chem. Crystallogr.,
1997, 27, 191–194; A. Tarazona, E. Koglin, F. Buda, B. B. Coussens,
J. Renkema, S. van Heel and R. J. Meier, J. Phys. Chem. B, 1997, 101,
4370–4378; C. E. Keller, W. R. Carper and B. J. Piersma, Inorg. Chim.
Acta, 1993, 209, 239–240; G. J. Mains, C. W. Bock, M. Trachtman and
V. S. Mastryukov, J. Mol. Struct., 1992, 274, 277–287; G. J. Mains, E. A.
Natsis and W. R. Carper, J. Phys. Chem. A, 2001, 105, 4371–4378; R.
Benn, E. Janssen, H. Lehmkuhl and A. Rufinska, J. Organomet. Chem.,
1987, 333, 155–168; E. Rytter and S. Kvisle, Inorg. Chem., 1986, 25,
3796–3804.
and A. J. Marcucci, Organometallics, 2001, 20, 5738–5744; B. L. Small,
Organometallics, 2003, 22, 3178–3183; (f) R. P. Tooze, M. J. Hanton, K.
Tenza, WO2008038173, Sasol Technology UK, 2008; (g) E. J. Baralt,
B. L. Small, and A. J. Marcucci, US6,291,733, Chevron CC, 2001;
(
h) E. J. Baralt and B. L. Small, US2002/0177744, Chevron Phillips
CC, 2002; (i) B. L. Small, US2004/068072, Chevron Phillips CC, 2004;
j) P. Kuhn, D. S e´ meril, D. Matt, M. J. Chetcuti and P. Lutz, Dalton
Trans., 2007, 515–528; (k) S. T. Babik and G. Fink, J. Organomet. Chem.,
003, 683, 209–219; (l) A. Kermagoret, F. Tomicki and P. Braunstein,
(
2
Dalton Trans., 2008, 2945–2955; (m) Y. Yang, P. Yang, C. Zhang, G.
Li, X.-J. Yang, B. Wu and C. Janiak, J. Mol. Catal. A: Chem., 2008,
2
96, 9–17; (n) A. Kermagoret and P. Braunstein, Dalton Trans., 2008,
564–1573; (o) Q. Z. Yang, A. Kermagoret, M. Agostinho, O. Siri and
1
P. Braunstein, Organometallics, 2006, 25, 5518–5527; (p) F. Speiser, P.
Braunstein and L. Saussine, Acc. Chem. Res., 2005, 38, 784–793.
(a) J. Christoffers and R. G. Bergman, J. Am. Chem. Soc., 1996, 118,
4
16 Y. Yang, H. Kim, J. Lee, H. Paik and H. G. Jang, Appl. Catal., A, 2000,
4
715–4716; (b) H. von der Heijden, B. Hessen and A. G. Orpen, J. Am.
193, 29–38.
Chem. Soc., 1998, 120, 1112–1113; (c) C. Janiak, Coord. Chem. Rev.,
17 “Physical Chemistry”, 5th Ed., P. W. Atkins, Oxford University Press,
Oxford, 1994, p 873.
2
1
006, 250, 66–94; C. Janiak and F. Blank, Macromol. Symp., 2006, 236,
4–22; C. Janiak, K. C. H. Lange, P. Marquardt, R.-P. Kruger and R.
18 GC-MS analysis of the 1-heptene substrate revealed trace levels
of impurity which precluded identification. Notably, trace levels of
moisture would not have been detected by this method.
19 B. L. Small and R. Schmidt, Chem.–Eur. J., 2004, 10, 1014–1020.
20 In the previous report, it was noted that elevated butane levels were
detected, but this was not quantified (see ref. 19).
21 Although not isomerisation, the conversion of a portion of the substrate
to alkane is included in the calculations of isomerisation of substrate,
most notably when used to quantify the true amount of a-olefin
substrate remaining for kinetic analysis.
Hanselmann, Macromol. Chem. Phys., 2002, 203, 129–138; C. Janiak,
K. C. H. Lange and P. Marquardt, J. Mol. Catal. A: Chem., 2002, 180,
4
3–58; C. Janiak, K. C. H. Lange and P. Marquardt, Macromol. Rapid
Commun., 1995, 16, 643–650.
5
6
7
(a) A. H. Neal, P. T. Parker, US3,686,352, 1972; (b) A. G. Kent, M. J.
Lawrenson, K. D. Macalpine, EP0,288,295, BP, 1988; (c) M. Itagaki,
G. Suzukamo, M. Yamamoto, US6,388,160, Sumitomo Chemical Co.,
2
002.
Our own (unpublished) studies of Cp
2
2
ZrCl /MAO systems have
confirmed that as the MAO loading is increased to achieve acceptable
activities there is a concomitant increase in the degree of higher
oligomerisation.
22 The apparent perturbation in selectivity at early times is due to the
formation of trace products still being below the detection limit of GC-
FID analysis, thus as products are calculated as a percentage of the
total formed, those detected are artificially inflated in the amount.
23 T. Agapie, J. A. Labinger and J. E. Bercaw, J. Am. Chem. Soc., 2007,
129, 14281–14295; T. Agapie, S. J. Schofer, J. A. Labinger and J. E.
Bercaw, J. Am. Chem. Soc., 2004, 126, 1304–1305; D. S. McGuinness,
Organometallics, 2009, 28, 244–248; M. J. Overett, K. Blann, A.
Bollmann, J. T. Dixon, D. Haasbroek, E. Killian, H. Maumela, D. S.
McGuinness and D. H. Morgan, J. Am. Chem. Soc., 2005, 127, 10723–
10730.
(a) N. A. Maly, H. R. Menapace, G. S. Benner, US3,903,193, Goodyear,
1
975L. G. Wideman, US3,813,453, Goodyear, 1974N. A. Maly, H. R.
Menapace, G. S. Benner, US3,784,629, Goodyear, 1974N. A. Maly,
H. R. Menapace, G. S. Benner, US3,784,630, Goodyear, 1974N. A.
Maly, H. R. Menapace, G. S. Benner, US3,784,631, Goodyear,
1
1
974N. A. Maly, H. R. Menapace, M. Brown, US3,897,512 Goodyear,
975; (b) D. E. Hendriksen, Exxon, US5,059,739, 1991; (c) N. A. Maly,
H. R. Menapace, J. L. Wang and L. G. Wideman, J. Org. Chem., 1975,
4
0, 2983–2985.
H. Olivier and P. Laurent-G e´ rot, J. Mol. Catal. A: Chem., 1999, 148,
3–48.
M. J. Hanton and R. P. Tooze, WO2005089940, Sasol Technology UK,
005.
0 No literature precedent could be found for the reaction of an aromatic
aniline) halide with an aluminium alkyl to yield an alkyl aromatic
24 (a) P. Cossee, J. Catal., 1964, 3, 80–8; E. J. Arlman, J. Catal., 1964, 3,
89–98; P. Cossee and E. J. Arlman, J. Catal., 1964, 3, 99–104.
25 S.-Y. S. Wang, D. D. van der Lende, K. A. Abboud and J. M. Boncella,
Organometallics, 1998, 17, 2628–2635.
8
9
4
2
26 Total liquid product formed = 2.936 g; No solid product/polymer
-1
1
was observed; TON = 1,047 (mol =)(mol W) ; activity = 6,978 (mol
-1 -1
(
=)(mol W) hr .
species, and certainly aromatic halides are deactivated towards elec-
trophilic aromatic substitution, although the amine moiety would be
expected to have a countering activating effect. Alternatively, reaction
could occur once the aniline had ligated the tungsten. It should be noted
that Friedel–Crafts alkylation of aromatic species (solvent) during
catalysis activated with EADC has been observed previously: P. W.
Dyer, J. Fawcett and M. J. Hanton, Organometallics, 2008, 27, 5082–
27 A. K. Tomov, V. C. Gibson, G. J. P. Britovsek, R. J. Long, M. van
Meurs, D. J. Jones, K. P. Tellmann and J. J. Chirinos, Organometallics,
2009, 28, 7033–7040.
28 T. J. M. de Bruin, L. Magna, P. Raybaud and H. Toulhoat,
Organometallics, 2003, 22, 3404–3413; S. Tobisch and T. Ziegler,
Organometallics, 2003, 22, 5392–5405; A. N. J. Blok, P. H. M. Budzelaar
and A. W. Gal, Organometallics, 2003, 22, 2564–2570.
29 See for example: J. G. Hinman, A. J. Lough and R. H. Morris, Inorg.
Chem., 2007, 46, 4392–4401; D. M. Heinekey and W. J. Oldham, Chem.
Rev., 1993, 93, 913–926; M. H. Chisholm, K. S. Kramer and W. E.
Streib, J. Am. Chem. Soc., 1992, 114, 3571–3573; F. Maseras, A. Lled o´ s,
E. Clot and O. Eisenstein, Chem. Rev., 2000, 100, 601–636; D. Gregson,
S. A. Mason, J. A. K. Howard, J. L. Spencer and D. G. Turner, Inorg.
Chem., 1984, 23, 4103–4107; R. M. Bullock, J.-S. Song and D. J. Szalda,
Organometallics, 1996, 15, 2504–2516.
5
087.
1
1
1 M. J. Hanton and K. Tenza, Organometallics, 2008, 27, 5712–5716; K.
Tenza, M. J. Hanton and A. M. Z. Slawin, Organometallics, 2009, 28,
4
852–4867; J. T. Dixon, M. J. Green, F. M. Hess and D. H. Morgan,
J. Organomet. Chem., 2004, 689, 3641–3668.
2 For samples taken during catalysis an error in terms of selectivity to the
dimer fraction can sometimes result. As a small amount of material is
withdrawn via syringe and subjected to acidified aqueous work-up on
a very small scale, evaporative loss of the substrate olefin is sometimes
unavoidable. Due to the way in which selectivity is calculated, any
substrate lost to evaporation does not appear in the GC trace and is thus
inherently accounted for in the mass balance as heavies not observed
by GC (>C34). This therefore artificially decreases the selectivity to the
dimer fraction and increases the activity/productivity calculated. If this
30 J. M. Boncella, S.-Y. S. Wang and D. D. van der Lende, J. Organomet.
Chem., 1999, 591, 8–13.
31 “Purification of Laboratory Chemicals”, 4th Ed., W. L. F. Armarego,
D. D. Perrin, Butterworth Heinemann, Oxford, 2002.
32 T. E. Kuzmenko, A. L. Samusenko, V. P. Uralets and R. V. Golovnya,
J. High Resolut. Chromatogr., 1979, 2, 43.
This journal is © The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 7025–7037 | 7037