Carbon-bridged diphosphine ligands for chromium-catalysed ethylene tetramerisation and trimerisation reactions

Article abstract of DOI:10.1016/j.molcata.2007.11.036

The use of carbon-bridged diphosphine ligands in chromium-catalysed ethylene tri- and tetramerisation reactions has been investigated. Two- and three-carbon spacer ligands all showed activity for selective oligomerisation, with a structure-selectivity correlation between P-Cr-P bite angle and 1-octene:1-hexene ratio evident. Activated chromium complexes of single carbon spacer diphosphines were also shown to be effective tetramerisation catalysts, provided that the ligand is innocent under the conditions of catalyst activation. A catalyst with the bis(diphenylphosphino)benzene ligand was found to be exceptionally active, although the combined 1-hexene and 1-octene selectivity was lower than with the best diphosphinoamine (PNP) ligands. The yield losses to by-products can to an extent be minimised by the use of high reaction temperatures and pressures. Unlike with the PNP-based systems, attempts to activate the Cr/bis(diphenylphosphino)benzene catalyst in situ from a chromium salt and free ligand resulted in low activity and high polymer formation. The effect of different phosphine substitution on catalyst selectivity was explored. Steric constraints around the catalytic centre (ortho-alkylphenyl phosphines) resulted in a shift towards 1-hexene formation, as with PNP catalysts. Additionally, the basicity of the phosphines appears to influence catalyst selectivity, with alkyl phosphines favouring trimerisation. An interplay between phosphine basicity and bridge structure is in evidence, however, as a catalyst containing a ligand with both basic phosphine atoms and a small bite angle was shown to be selective towards 1-octene.

Full text of DOI:10.1016/j.molcata.2007.11.036

Available online at www.sciencedirect.com
Journal of Molecular Catalysis A: Chemical 283 (2008) 114–119
Carbon-bridged diphosphine ligands for chromium-catalysed
ethylene tetramerisation and trimerisation reactions
Matthew J. Overetta,∗, Kevin Blanna, Annette Bollmanna, Raylene de Villiersa,
John T. Dixona, Esna Killiana, Munaka C. Maumelaa, Hulisani Maumelaa,
David S. McGuinnessb, David H. Morgana,
Adam Rucklidgeb, Alexandra M.Z. Slawinc
a R&D Division, Sasol Technology (Pty) Ltd., 1 Klasie Havenga Road, Sasolburg 1947, South Africa
b Sasol Technology U.K. Ltd., Purdie Building, North Haugh, St. Andrews KY16 9ST, UK
c School of Chemistry, University of St. Andrews, Purdie Building, North Haugh, St. Andrews KY16 9ST, UK
Received 9 November 2007; accepted 17 November 2007
Available online 1 February 2008
Abstract
The use of carbon-bridged diphosphine ligands in chromium-catalysed ethylene tri- and tetramerisation reactions has been investigated. Two- and
three-carbon spacer ligands all showed activity for selective oligomerisation, with a structure–selectivity correlation between P–Cr–P bite angle and
1-octene:1-hexene ratio evident. Activated chromium complexes of single carbon spacer diphosphines were also shown to be effective tetramerisa-
tion catalysts, provided that the ligand is innocent under the conditions of catalyst activation. A catalyst with the bis(diphenylphosphino)benzene
ligand was found to be exceptionally active, although the combined 1-hexene and 1-octene selectivity was lower than with the best diphosphi-
noamine (PNP) ligands. The yield losses to by-products can to an extent be minimised by the use of high reaction temperatures and pressures.
Unlike with the PNP-based systems, attempts to activate the Cr/bis(diphenylphosphino)benzene catalyst in situ from a chromium salt and free
ligand resulted in low activity and high polymer formation. The effect of different phosphine substitution on catalyst selectivity was explored.
Steric constraints around the catalytic centre (ortho-alkylphenyl phosphines) resulted in a shift towards 1-hexene formation, as with PNP catalysts.
Additionally, the basicity of the phosphines appears to influence catalyst selectivity, with alkyl phosphines favouring trimerisation. An interplay
between phosphine basicity and bridge structure is in evidence, however, as a catalyst containing a ligand with both basic phosphine atoms and a
small bite angle was shown to be selective towards 1-octene.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Chromium; Phosphine ligands; Ethylene trimerisation; Ethylene tetramerisation; Homogeneous catalysis
1. Introduction
hexene have been reported, and a process is now practiced
commercially [2]. Recently, the first catalysts for the selec-
tive tetramerisation of ethylene to 1-octene were reported [3].
The active catalyst is typically generated by a combination of a
chromium salt and a diphosphinoamine ligand (Ar2PN(R)PAr2)
in the presence of an aluminoxane activator. Mechanistic stud-
ies [4] on these catalyst systems indicate that the 1-octene
is liberated from a metallacyclononane catalytic intermediate,
formed by an extension of the metallacyclic mechanism pre-
viously proven for ethylene trimerisation reactions [5]. The
exact nature of the active catalyst species is still not clear,
although there is evidence for a Cr(I)–Cr(III) redox couple
[6].
Traditional linear alpha olefin (LAO) production processes
typically produce a broad and relatively inflexible distribution
of LAO homologues [1]. Selective processes for the produc-
tion of 1-hexene and 1-octene will enable industry to meet
incremental demand growth for these valuable homologues
without co-producing other low-value olefinic products. To
this end, numerous catalysts for ethylene trimerisation to 1-
∗
Corresponding author. Fax: +27 11 522 9696.
E-mail address: matthew.overett@sasol.com (M.J. Overett).
1381-1169/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2007.11.036

M.J. Overett et al. / Journal of Molecular Catalysis A: Chemical 283 (2008) 114–119
115
The uniqueness of the PNP–Cr system in its ability to selec-
tively produce 1-octene (unlike numerous other catalysts for
1-hexene) has led to considerable interest in the features of
the ligand which confer the unusual selectivity. Several studies
have examined the influence of the ligand’s aryl (Ar) sub-
stituents on catalyst selectivity; in particular on the selectivity
split between 1-hexene and 1-octene [7]. In general, substitu-
tion at the ortho-position by coordinating groups (e.g.–OMe)
or sterically demanding groups (e.g. –Et) causes a switch in
selectivity from 1-octene to 1-hexene. Variations of the PNP
N-substituent have also been explored: increasing the steric
bulk of alkyl or aryl groups enhances the overall selectivity to
total alpha olefins (1-hexene plus 1-octene), while lowering the
1-octene:1-hexene ratio [8]. It has also been reported that poten-
tially coordinating ether groups on the PNP nitrogen substituent
may increase overall productivity by enhancing catalyst stability
[9].
The relationship between structure and selectivity is thus
well studied for diphosphinoamine (PNP) ligands. Relatively
little, by contrast, has been reported on the application of
diphosphine ligands with different spacers in tetramerisation
reactions, although it is known that hydrazyl-bridged diphos-
phines (PNNP) and even a classical carbon-bridged ligand such
as bis(diphenylphosphino)ethane can be used to obtain an active
tetramerisation catalyst [3]. With the aim of further elucidating
the role of ligand structure in these novel catalytic reactions,
we now report a detailed study on the use of carbon-bridged
diphosphine ligands in ethylene trimerisation and tetramerisa-
tion reactions.
Fig. 1. Carbon-bridged diphosphine ligands, complexed with Cr and evaluated
in selective ethylene oligomerisation reactions.
2.3. Ligand synthesis
Ligands 1–6 and 12 (see Figs. 1 and 3) were obtained from
commercial sources. Ligands 7 [10], 8 [11] and 11 [12] were
prepared by literature methods.
2. Experimental
2.1. General
1,2-Bis(di[2-ethylphenyl]phosphino)benzene (9). 1-Bromo-
ethylbenzene (5.0 g, 27 mmol) was added dropwise to a flask
containing excess magnesium turnings in THF (40 ml) at 0 ◦C.
A vigorous reaction ensued. After 3 h, the solution was added
dropwise to bis(dichlorophosphino)benzene (1.7 g, 6.1 mmol)
in THF (30 ml) at 0 ◦C. The mixture was warmed to room tem-
perature and stirred overnight. After filtration through celite, the
solvent was removed in vacuo and the crude product was recrys-
tallisedfromdichloromethane/ethanol. (2.2 g, 65%);δH (CDCl3,
400 MHz): 0.99 (12H, t, J = 7.2 Hz, –ArCH2CH3), 2.50–2.70
(8H, m, –ArCH2CH3), 6.77–6.80 (4H, m, Ar), 6.99–7.00 (2H,
m, Ar), 7.18–7.20 (4H, m, Ar), 7.21–7.24 (10H, m, Ar). δC{H}
(CDCl3, 100 MHz): 14.9 (–ArCH2CH3), 27.4 (–ArCH2CH3),
125.7 (Ar), 128.0 (Ar), 128.5 (Ar), 128.9 (Ar), 134.1 (Ar),
134.3 (Ar), 135.7 (Ar), 143.8 (Ar), 148.6 (Ar). δP{H} (CDCl3,
162 MHz) −30.3 ppm (s).
All solvents were purified by percolation under N2 through
activated alumina. MMAO-3A (modified methylaluminoxane)
was purchased from Akzo-Nobel.
2.2. Catalyst testing procedure
A 300 ml autoclave was charged with methylcyclohexane
(95 ml) and MMAO-3A (volume calculated for correct Al:Cr)
and heated to just below the desired reaction temperature.
The procatalyst (suspension or solution in methylcyclohex-
ane, 5 ml) was added to the autoclave, which was pressurised
with ethylene (50 bar) and stirred with a gas-entraining stir-
rer at 1000 rpm. Ethylene was fed on demand to maintain the
reactor pressure, and the uptake was monitored with a flow
meter. The temperature was maintained at reaction tempera-
ture with an internal cooling coil and external cooling. The
reaction was stopped by addition of ethanol (2 ml) after either
30 min or 110 g ethylene uptake. The reactor was cooled to
0 ◦C and slowly depressurised. A weighed aliquot of nonane
was added as an internal standard, and the liquid products
were analysed using GC-FID. The polymeric products were
recovered by filtration and overnight drying in an oven at
100 ◦C.
1,2-Bis(diisopropylphosphino)benzene
(10).
2-
Bromopropane (3.0 g, 24 mmol) was added dropwise to a
flask containing excess magnesium turnings in THF (40 ml)
at 0 ◦C. A vigorous reaction ensued. After 3 h, the solution
was added dropwise to bis(dichlorophosphino)benzene (1.5 g,
5.4 mmol) in THF (30 ml) at 0 ◦C. The mixture was warmed to
room temperature and stirred overnight. After filtration through
celite, the solvent was removed in vacuo and the crude product
was recrystallised from dichloromethane/ethanol. (1.2 g, 72%);

116
M.J. Overett et al. / Journal of Molecular Catalysis A: Chemical 283 (2008) 114–119
δH (CDCl3, 400 MHz): 0.88–0.93 (12H, m, –CH2CH3CH3),
1.12–1.18 (12H, m, –CH2CH3CH3), 2.09–2.14 (4H, m,
–CH2CH3CH3), 7.27–7.32 (2H, m, Ar), 7.50–7.52 (2H, m,
Ar). δC{H} (CDCl3, 100 MHz): 19.57 (–CH2CH3CH3), 20.3
(–CH2CH3CH3), 25.0 (–CH2CH3CH3), 127.8 (Ar), 132.5 (Ar),
144.3 (Ar). δP{H} (CDCl3, 162 MHz) −31.0 ppm (br s).
sured = 9764, independent reflections (Rint) = 5280(0.0162),
final R1 = 0.0297, wR2 [I > 2σ(I)] = 0.0708. Crystallographic
data (excluding structure factors) have been deposited with
the Cambridge Crystallographic Data Centre as supplementary
publication no. CCDC 212528. Copies of the data can be
obtained free of charge on application to CCDC, 12 Union
Road, Cambridge CB2 1EZ, UK [fax: (44) 1223-336-033;
e-mail: deposit@ccdc.cam.ac.uk].
2.4. Preparation of complexes
[L2CrCl2(␮-Cl)]2 (L2 = 1–7) was prepared in quantitative
yields according to literature procedure [3]. The complexes
were characterised by elemental analysis, or used directly
without characterisation. [(1)CrCl2(␮-Cl)]2: Found (calcu-
lated for C50H44Cl6P4Cr2). C, 55.48 (55.32); H, 3.91 (4.00).
[(3)CrCl2(␮-Cl)]2: Found (calculated for C52H48Cl6P4Cr2). C,
56.05 (56.09); H, 4.27 (4.34). [(4)CrCl2(␮-Cl)]2: Found (cal-
culated for C52H44Cl6P4Cr2). C, 56.15 (56.29); H, 3.85 (4.00).
[(5)CrCl2(␮-Cl)]2: Found (calculated for C60H48Cl6P4Cr2). C,
59.67 (59.58); H, 3.91 (4.00).
3. Results and discussion
A range of carbon-bridged bis-diphenylphosphine ligands
(Fig. 1) were obtained from commercial sources (1–6) or pre-
pared via literature methods (7 [10]) and complexes of the form
[L2CrCl2(␮-Cl)]2 were prepared by reaction of the ligand with
CrTHF3Cl3 [3].
These complexes were activated with MMAO-3A and
screened for ethylene oligomerisation using typical tri- or
tetramerisation conditions. The results, with a comparison
against two diphosphinoamine (PNP) ligands, are shown in
Table 1.
All of the complexes with 2- or 3-carbon spacer ligands
(3–7) were successfully activated for selective oligomerisation
reactions. Although various parameters are at play here (as
exemplified by the unselective catalysis with ligands 1 and 2),
there appears to be some general correlation between P–Cr–P
bite angle and the 1-octene:1-hexene ratio. The PNP ligands,
which have very small bite angles (ca. 67◦ [3]) have the high-
est 1-octene:1-hexene ratios, although clearly this ratio is also
strongly influenced by the N-substitution. The 2-carbon spacer
ligands (3–5), which typically coordinate to transition metals
with P–M–P bite angles of around 81–83◦ [13], produce slightly
more 1-hexene at the expense of 1-octene, while the 3-carbon
spacer ligands with still larger bite angles – typically ca. 91◦
for complexes of 6 [13] – gave the lowest 1-octene:1-hexene
ratios. Ligand 7, with a large (ca. 90◦) bite angle, unusual out-of-
plane binding mode and an inflexible backbone [14], produces
more 1-hexene than 1-octene. The poor activities obtained using
L2CrCl3(THF) (L2 = 5, 8, 11, 12). A mixture of CrTHF3Cl3
(1.0 g, 2.7 mmol) and ligand (1.1 equiv.) was stirred in THF
(30 ml) at room temperature. After stirring overnight, the com-
plex was precipitated by addition of pentane. The filtered
blue powders were washed with pentane and dried in vacuo.
(5)CrCl3(THF): Found (calculated for C34H32Cl3OP2Cr). C,
60.20 (60.33); H, 4.59 (4.76); (8)CrCl3(THF): Found (calcu-
lated for C17H38Cl3OP2Cr). C, 42.65 (42.49); H, 8.00 (7.94).
(11)CrCl3(THF): Found (calculated for C18H40Cl3OP2Cr). C,
43.52 (43.87); H, 8.45 (8.18). (12)CrCl3(THF): Found (calcu-
lated for C10H24Cl3OP2Cr). C, 31.29 (31.56); H, 6.30 (6.36).
Crystals of (5)CrCl3(THF) suitable for X-ray structure
solution were grown from evaporating THF. Crystal data and
structure refinements for (5)CrCl3(THF): C34H32Cl3CrOP2, Mr
676.89, crystal size 0.1 mm × 0.01 mm × 0.01 mm, blue plate,
˚
triclinic, space group P-1, T = 93(2) K, a = 9.768(3) A,
b = 10.257(3) A,
α = 101.123(6)◦,
˚
˚
c = 17.699(4) A,
3
β = 95.817(2)◦, γ = 115.983(5)◦, V = 1528.9(7) A , Z = 2,
˚
Dc = 1.470 Mg/m3,
μ = 0.769 mm−1
,
reflections
mea-
Table 1
Screening of complexes [L2CrCl2(␮-Cl)]2 (L2 = carbon-bridged diphosphine ligands) in chromium-catalysed ethylene oligomerisation reactions
Ligand (L2)
Activity
1-Hexene
(mass%)a
C6 cyclics
(mass%)a
1-Octene
(mass%)a
C10–C14
Polymer
1-Octene:1-hexene
ratio
(×103 g/g Cr/h)
(mass%)a
(mass%)b
PNPc
PNPd
1072
1868
21
<10
144
303
2240
13
6.7
14.1
9.5
4.4
61.4
69.3
8.5
7.7
1.7
0.8
9.1
4.9
1
2
3
4
5
6
7
Schulz–Flory distribution of LAO’s (α = 0.55)
Unselective distribution of oligomers
15.7
7.3
13.0
8.6
8.7
17.3
13.5
8.1
59.3
46.7
56.8
30.3
31.4
6.5
10.8
7.6
16.2
2.9
4.9
5.2
0.9
24.3
4.0
3.8
6.4
4.4
3.5
0.51
70
61.9
2.6
Conditions: 5–30 ␮mol Cr (for each reaction, amount of Cr chosen to ensure significant ethylene uptake but controllable exotherm), MMAO-3A activator
(Al:Cr = 500:1), 60 ◦C, 50 bar C2H4, methylcyclohexane (100 ml).
a
Mass% of liquid products (oligomers).
Mass% of total product (oligomers + polymer).
Ph2PN(Et)PPh2.
Ph2PN(iPr)PPh2.
b
c
d

M.J. Overett et al. / Journal of Molecular Catalysis A: Chemical 283 (2008) 114–119
117
Table 2
Evaluation of [(5)CrCl2(␮-Cl)]2/MMAO-3A for selective oligomerisation under different reaction conditions
C2H4 pressure
(bar)
Temperature
(◦C)
Activity
(×103 g/g Cr/h)
1-Hexene
(mass%)a
C6 cyclics
(mass%)a
1-Octene
(mass%)a
C10–C14
(mass%)a
Polymer
(mass%)b
1-Octene:1-hexene
ratio
50
50
70
70
70
60
80
60
80
100
2240
5930
3410
7200
4570
13.0
27.4
10.2
24.7
38.9
13.5
10.7
13.0
10.6
8.0
56.8
53.6
57.5
54.8
45.5
7.6
5.9
7.8
5.2
4.7
0.9
0.3
1.5
0.8
3.2
4.4
1.9
5.6
2.2
1.2
Conditions: 1.5 ␮mol Cr, MMAO-3A activator (Al:Cr = 500:1), methylcyclohexane (100 ml).
a
Mass% of liquid products (oligomers).
Mass% of total product (oligomers + polymer).
b
diphosphines with saturated linkers (3, 6) suggest that ligand
rigidity and/or electronic effects play an important role in cata-
lyst performance as well.
of preparing a procatalyst Cr–ligand complex while still obtain-
ing an active and selective catalyst. Attempts to apply similar in
situ complexation with the carbon-bridged diphosphine ligands
were however less successful. A typical comparative example
of this behaviour is shown in Table 3. While the product selec-
tivities within the liquid products were effectively identical to
those obtained with the activated complex, the catalyst activities
were dramatically attenuated and the amount of solids (poly-
mers and waxes) greatly increased. These are indications of
poor catalyst activation, where part of the incompletely activated
chromium acts as a linear chain growth catalyst. The reasons
for the poor in situ activation compared with PNP ligands are
unknown.
Dimeric complexes of the form [L2CrCl2(␮-Cl)]2 are typi-
cally highly insoluble in hydrocarbon solvents before activation.
A more soluble monomeric procatalyst complex which retains
excellent catalytic performance is thus desirable. Reaction of 5
with CrTHF3Cl3 in THF at room temperature gave the soluble
monomer complex (5)CrCl3(THF) in high yield. The complex
may be dissolved in the catalytic solvent at room temperature
without immediate dimerisation, and this procatalyst was then
evaluated under catalytic conditions (Table 3). Although the
amount of polymer obtained was slightly higher than when
the dimer procatalyst was used, the rates and selectivities are
comparable.
Very high catalyst activity – among the highest obtained
under comparable conditions – was obtained when the
chromium complex of bis(diphenylphosphino)benzene (5) was
evaluated. However, under the catalytic conditions employed,
the selectivity is inferior compared to the best PNP ligands in
two respects: increased formation of C6 cyclic products (methyl-
cyclopentane and methylenecyclopentane) and lower 1-octene
to 1-hexene ratio. In an effort to improve selectivity through
choice of reaction conditions while retaining the benefit of high
activity, this system was then selected for further study. The
results are shown in Table 2.
The selectivity trends for this catalyst systems are in line
with those reported for Cr/PNP catalyst systems [15]. The 1-
octene/1-hexene split is strongly influenced by temperature,
while the formation of C6 cyclics is suppressed at higher tem-
peratures. However, even at 100 ◦C, a substantial amount (8%)
of these undesirable cyclic products is still found. Reactions
at higher pressures and temperatures resulted in catalyst activity
increases. Theobservedratesundertheseconditionsareamongst
the highest ever reported, with a nominal hourly activity of more
than 7 million g ethylene/g Cr at 80 ◦C, 70 bar ethylene. Even at
100 ◦C, where catalyst deactivation and polymer formation are
problematicformanytri-andtetramerisationcatalysts, veryhigh
rates and relatively low polymer formation were nevertheless
obtained.
The molecular structure of (5)CrCl3(THF) is shown in Fig. 2.
The expected distorted octahedral structure is observed. The
P–Cr–P bite angle is 77.8◦, substantially larger than the 66.7◦
bite angle of a Cr–PNP procatalyst complex [3], but smaller than
the average for transition metal complexes of this ligand [13].
The Cr–P bond distances are similar to those in the Cr–PNP
complex [3].
Cr/PNP tri-/tetramerisation catalyst systems are often pre-
pared in situ by combining a Cr(III) salt such as Cr(acac)3, the
free ligand and aluminoxane at the beginning of the catalytic
reaction. This approach avoids the process complexity and cost
Table 3
Catalyst performance with different Cr/5 procatalyst sources
Procatalyst
Efficiency
1-Hexene
C6 cyclics
(mass%)a
1-Octene
(mass%)a
C10–C14
(mass%)a
Polymer
(mass%)b
1-Octene:1-hexene
ratio
(×103 g/g Cr/h) (mass%)a
[(5)CrCl2(␮-Cl)]2
Cr(acac)3/5c
(5)CrCl3(THF)
2240
48
1543
13.0
13.4
11.3
13.5
13.9
12.8
56.8
57.1
53.9
7.6
6.4
8.7
0.9
21.3
3.5
4.4
4.3
4.8
Conditions: MMAO-3A activator (Al:Cr = 500:1), 60 ◦C, 50 bar C2H4, methylcyclohexane (100 ml).
a
Mass% of liquid products (oligomers).
Mass% of total product (oligomers + polymer).
Non-reacting mixture.
b
c

118
M.J. Overett et al. / Journal of Molecular Catalysis A: Chemical 283 (2008) 114–119
Fig. 3. Carbon-bridged diphosphine ligands with varying phosphine substitu-
tion.
Fig. 2. Molecular structure of (5)CrCl3(THF). Selected bond distances and
angles: Cr(1)–P(1) 2.4619(7) A; Cr(1)–P(2) 2.5558(9) A; P(1)–Cr(1)–P(2)
77.82(2)◦.
˚
˚
isopropyl substituents; however it is noteworthy that the rela-
tively high 1-octene:1-hexene ratio is consistent with the narrow
bite angle expected for a single-atom spacer diphosphine ligand
(cf. typical bite angles of 72◦ for dppm) [13].
The failure to obtain selective oligomerisation catalysts with
the 1-carbon spacer ligands 1 and 2 (Table 1) – direct analogues
of PNP ligands – was ascribed to non-innocence of the ligands
under the reaction conditions. Bis(diphenylphosphino)methane
(dppm, 1) is known to be susceptible to deprotonation at the
bridging methylene by metal alkyl Lewis bases [16]. In the hope
that decreasing the acidity of the methylene hydrogens would
prevent ligand deprotonation, the more basic diphosphine lig-
and bis(diisopropylphosphino)methane (8, Fig. 3) was prepared
[11], complexed to chromium and evaluated under catalytic
conditions. The successful tetramerisation activity obtained
(Table 4) strongly supports our hypothesis of non-innocence of
dppm, and demonstrates that single-spacer diphosphines, other
than PNP ligands, may indeed be successfully applied in selec-
tive oligomerisation reactions. Selectivity comparisons of 8 with
the other carbon-bridged ligands have limited value due to the
very different steric and electronic environment caused by the
As with Cr–PNP catalysts [7], modification of the
phosphine substituents on carbon-bridged diphosphines can
affect the C6/C8 selectivity of the catalysts. Making use
of the 1,2-phenylene bridge successfully identified in the
screening experiments, diphosphines 1,2-bis(R2P)C6H4 (R = 2-
ethylphenyl, 9; R = isopropyl, 10; Fig. 3) were prepared.
Attempts to prepare chromium(III) complexes of 9 were unsuc-
cessful, so the ligands were tested in situ. The results, with
comparisonagainst5, areshowninTable4. Whiletheinsituacti-
vation method again resulted in poor activities and high polymer
formation, it is clear that the change in phosphine substitution
causes a swing from tetramerisation (5, phenyl substituents) to
trimerisation (9, 2-ethylphenyl or 10, isopropyl substituents).
The effect of ortho-alkylphenyl substitution has already been
noted with Cr/PNP catalysts [7b], but the dramatic selectivity
Table 4
The effect of variations in phosphine substitution on catalyst selectivity
Procatalyst
Activity
(×103 g/g Cr/h)
1-Hexene
(mass%)a
C6 cyclics
(mass%)a
1-Octene
(mass%)a
C10–C14
(mass%)a
Polymer
(mass%)b
1-Octene:1-hexene
ratio
(8)CrCl3(THF)
Cr(acac)3/5
Cr(acac)3/9
174
48
31
39
144
11
7
14.5
13.4
59.2
82.8
15.7
90.0
86.2
12.0
13.9
2.5
1.8
8.7
62.3
57.1
10.7
10.5
59.3
5.8
7.5
6.4
13.1
12.7
6.5
4.7
21.3
65.6
24.0
4.9
4.3
4.3
0.18
0.13
3.8
0.06
0.10
Cr(acac)3/10
[(3)CrCl2(␮-Cl)]2
(11)CrCl3(THF)
(12)CrCl3(THF)
0.8
0.9
4.5
4.3
2.1
20.4
9.0
Conditions: 20 ␮mol Cr, MMAO-3A activator (Al:Cr = 500:1), 60 ◦C, 50 bar C2H4, methylcyclohexane (100 ml).
a
Mass% of liquid products (oligomers).
Mass% of total product (oligomers + polymer).
b

M.J. Overett et al. / Journal of Molecular Catalysis A: Chemical 283 (2008) 114–119
119
change effected by the isopropyl substituents was unexpected.
This observation suggests that, apart from steric considerations,
electronic factors also play a prominent role in determining the
reaction selectivity.
[2] For a recent review on ethylene trimerisation see J.T. Dixon, M.J.
Green, F.M. Hess, D.H. Morgan, J. Organomet. Chem. 689 (2004)
3641.
[3] A. Bollmann, K. Blann, J.T. Dixon, F.M. Hess, E. Killian, H. Maumela, D.S.
McGuinness, D.H. Morgan, A. Neveling, S. Otto, M.J. Overett, A.M.Z.
Slawin, P. Wasserscheid, S. Kuhlmann, J. Am. Chem. Soc. 126 (2004)
14712.
[4] M.J. Overett, K. Blann, A. Bollmann, J.T. Dixon, D. Haasbroek, E. Killian,
H. Maumela, D.S. McGuinness, D.H. Morgan, J. Am. Chem. Soc. 127
(2005) 10723.
[5] (a) J.R. Briggs, J. Chem. Soc., Chem. Commun. (1989) 674;
(b) R. Emrich, O. Heinemann, P.W. Jolly, C. Kruger, G.P.J. Verhovnik,
Organometallics 16 (1997) 1511;
To investigate the interplay between steric and electronic
factors for carbon-bridged ligands further, alkyl substituted
diphosphines R2PCH2CH2PR2 (R = iPr, 11, R = Me, 12; Fig. 3)
were obtained (12) or prepared (11) [12]. Cr complexes of these
ligands were prepared and tested under catalytic conditions. The
results, withcomparisonagainstthephenyl-substitutedanalogue
3, are shown in Table 4. Given the propensity of even the unhin-
dered methyl-substituted diphosphine 12 for trimerisation, it
appears that the basicity of the phosphine plays a distinct role in
determining selectivity, with more basic phosphines favouring
hexene formation. A complex interplay between bridge structure
(bite angle) and phosphine bacisity is evident, however; witness
the tetramerisation selectivity with methylene bridged ligand 8
versus the trimerisation selectivity with ethyl-bridged 11.
(c) N. Meijboom, C. Schaveien, A.G. Orpen, Organometallics 9 (1990)
774;
(d) T. Agapie, S.J. Schofer, J.A. Labinger, J.E. Bercaw, J. Am. Chem. Soc.
125 (2004) 1304.
[6] (a) A.J. Rucklidge, D.S. McGuinness, R.P. Tooze, A.M.Z. Slawin,
J.D.A. Pelletier, M.J. Hanton, P.B. Webb, Organometallics 26 (2007)
2782;
(b) T. Agapie, M.W. Day, L.M. Henling, J.A. Labinger, J.E. Bercaw,
Organometallics 25 (2006) 2733.
[7] (a) C. Carter, S.A. Cohen, N.A. Cooley, A. Murphy, J. Scutt, D.F. Wass,
Chem. Commun. (2002) 858;
4. Conclusions
(b) K. Blann, A. Bollmann, J.T. Dixon, F.M. Hess, E. Killian, H. Maumela,
D.H. Morgan, A. Neveling, S. Otto, M.J. Overett, Chem. Commun. (2005)
620;
(c) M.J. Overett, K. Blann, A. Bollmann, J.T. Dixon, F.M. Hess, E. Killian,
H. Maumela, D.H. Morgan, A. Neveling, S. Otto, Chem. Commun. (2005)
622.
It is clear that a wide array of different structural parameters
of diphosphine ligands affect the catalyst activity, selectivity
and activation. In addition to steric and coordinative mediation
of catalyst selectivity by the phosphine aryl substituents [7], the
ligand backbone also influences the selectivity. None of these
parameters should be viewed in isolation; however, bite angle,
rigidity, phosphine basicity and bridge unsaturation all appear
to play a role in the catalyst selectivity.
An extremely active ethylene tetramerisation catalyst using
bis(diphenylphosphino)benzene as ligand was identified. While
this system displays unprecedented activity, even at high tem-
peratures, the selectivity is inferior to that obtained with the best
PNP ligands. The selectivity to valuable products (1-hexene and
1-octene) can be improved to some extent by the use of high
temperatures (≥80 ◦C) and pressures (70 bar). Efforts to develop
new, more selective ligands based on the 1,2-bis(phosphino)aryl
template are underway.
Variation of phosphine substitution can cause a switch
from 1-octene formation toward 1-hexene. Steric mediation of
the selectivity by ortho-alkylphenyl substituents was expected
based on PNP analogies; however the switch to trimerisation
caused by increasing phosphine basicity has not previously
been observed. Apart from other effects, a complex relation-
ship between bridge structure and phosphine basicity appears to
determine the catalyst selectivity.
[8] (a) K. Blann, A. Bollmann, H. de Bod, J.T. Dixon, E. Killian, P. Nongodl-
wana, M.C. Maumela, H. Maumela, A.E. McConnell, D.H. Morgan, M.J.
Overett, M. Pretorius, S. Kuhlmann, P. Wasserscheid, J. Catal. 249 (2007)
244;
(b) S. Kuhlmann, K. Blann, A. Bollmann, J.T. Dixon, E. Killian, M.C.
Maumela, H. Maumela, D.H. Morgan, M. Pretorius, N. Taccardi, P. Wasser-
scheid, J. Catal. 245 (2007) 277;
(c) E. Killian, K. Blann, A. Bollmann, J.T. Dixon, S. Kuhlmann, M.C.
Maumela, H. Maumela, D.H. Morgan, P. Nongodlwana, M.J. Overett, M.
Pretorius, K. Ho¨fener, P. Wasserscheid, J. Mol. Catal. A: Chem. 270 (2007)
214.
[9] P.R. Elowe, C. McCann, P.G. Pringle, S.K. Spitzmesser, J.E. Bercaw,
Organometallics 25 (2006) 5255.
[10] R.D. Jackson, S. James, A.G. Orpen, P.G. Pringle, J. Organomet. Chem.
458 (1993) C3.
[11] Z.S. Novikova, A.A. Prishchenko, I.F. Lutsenko, J. Gen. Chem. USSR 47
(1977) 707.
[12] M.D. Fryzuk, T. Jones, F.W.B. Einstein, Organometallics 3 (1984)
185.
[13] P. Dierkes, P.W.N.M. van Leeuwen, J. Chem. Soc., Dalton Trans. (1999)
1519.
[14] S. James, A.G. Orpen, P.G. Pringle, J. Organomet. Chem. 525 (1996)
299.
[15] (a) S. Kuhlmann, J.T. Dixon, M. Haumann, D.H. Morgan, J. Ofili, O.
Spuhl, N. Taccardi, P. Wasserscheid, Adv. Synth. Catal. 348 (2006)
1200;
(b) R. Walsh, D.H. Morgan, A. Bollmann, J.T. Dixon, Appl. Catal. A: Gen.
306 (2006) 184.
References
[16] (a) K. Issleib, H.P. Abicht, J. Prakt. Chem. 312 (1970) 456;
(b) S. Al-Jibori, B.L. Shaw, J. Chem. Soc., Chem. Commun. (1982)
286.
[1] An excellent general overview is provided by: D. Vogt, B. Cornills, W.A.
Herrmann (Eds.), Applied Homogeneous Catalysis with Organometallic
Compounds, vol. 1, Wiley–VCH, New York, 2002, p. 240.