The impact of P substituents on the oligomerization of ethylene with nickel 2-diphenyl and 2-dicyclohexylphosphinophenolate phosphine catalysts

Article abstract of DOI:10.1016/j.jcat.2004.03.029

Solutions of 2-diphenyl- or 2-dicyclohexylphosphinophenol (3a or 3b), nickel bis(1,5-cyclooctadiene), and a tertiary phosphine R₃P in toluene oligomerize ethylene in most cases with high selectivity to linear α-olefins. The 2-phosphinophenols formed the backbone of the Ni catalyst while the phosphine additives, competing with ethylene and olefins for coordination in the trans position to the 2-phosphino group, allowed tuning of the catalyst properties. A rough estimation of the effects of other substituents or additives and of the scope and limits for application of such catalysts in the ethylene oligomerization should be feasible based on the results and assumptions made to understand the in part controversial behavior of 3a/Ni/PR₃ and 3b/Ni/PR₃ catalysts.

Full text of DOI:10.1016/j.jcat.2004.03.029

Journal of Catalysis 225 (2004) 16–23
www.elsevier.com/locate/jcat
The impact of P substituents on the oligomerization of ethylene
with nickel 2-diphenyl and 2-dicyclohexylphosphinophenolate
phosphine catalysts
Joachim Heinicke,a,∗ Martin Köhler,a,b Normen Peulecke,a and Wilhelm Keim b
a
Institut für Chemie und Biochemie, Ernst-Moritz-Arndt-Universität Greifswald, Soldmannstr. 16, D-17487 Greifswald, Germany
b
Institut für Technische Chemie und Petrolchemie, Rheinisch-Westfälische Technische Hochschule Aachen, Worringer Weg 1, D-52056 Aachen, Germany
Received 28 November 2003; revised 3 March 2004; accepted 5 March 2004
Available online 27 April 2004
Dedicated to Professor Gerd-Volker Röschenthaler on the occasion of his 60th birthday
Abstract
Solutions of 2-diphenyl- or 2-dicyclohexylphosphinophenol, nickel bis(1,5-cyclooctadiene), and a tertiary phosphine R P in toluene
3
oligomerize ethylene in most cases with high selectivity to linear α-olefins. The molecular weights can be tuned by the catalyst compo-
∩
∩
sition. Molecular weights are generally lower with Ph P ONi than with cHex P ONi catalysts and in the presence of R P additives, but
2
2
3
basicity-dependent effects of various triarylphosphine additives on the molecular weights and relative reaction rates, displayed by pressure-
∩
∩
time plots, are rather different or opposite for Ph P ONi and cHex P ONi catalysts.
2
2
2004 Elsevier Inc. All rights reserved.
Keywords: Chelates; Homogeneous catalysis; Nickel; O, P ligands; Oligomerization
1. Introduction
The oligomerization of ethylene to linear α-olefins by the
Shell Higher Olefins Process (SHOP) is of industrial as well
as of academic interest and has stimulated intensive research
on nickel catalysts with P∩O- and related chelating lig-
ands [1–6]. It was shown that catalytically active and selec-
tive 2-phosphinoenolate and 2-phosphinophenolate nickel
complexes are chelates with square-planar geometry [7–12].
In the single-component catalysts the activation temperature
depends strongly on the nature of the ligand trans to oxygen
[1,13,14] while the molecular weights are particularly influ-
enced by substituents at phosphorus [12–16] and by solvents
or ligands interacting or bound in trans position to phos-
phorus, e.g., donor solvents or pyridine [13,16–20], phos-
phines [19,21], or ylides [22,23]. A study of the influence
of various phosphines R3P in diphenylphosphinoenolate
and -phenolate phenylnickel phosphine single-component
oligomerization catalysts 1 and 2, accessible from α-keto- or
o-chinone-triphenylphosphorus ylides, Ni(COD)2, and the
Scheme 1.
respective phosphine (Scheme 1) [7,8], demonstrated widely
varying activities and mass distributions of the resulting eth-
ylene oligomers, somewhat higher molecular weights, and
different response toward PMe3 and PcHex3 ligands on the
use of the phosphinophenolate as compared to the phosphi-
noenolate catalysts (1 C4–C30; 2 C4–C90) [21].
In investigations of phosphinophenolate methylnickel
trimethylphosphine complexes bearing varying R2P groups
we observed a strong influence of the P substituents in ad-
dition to the PMe3 content on the catalytic activity toward
ethylene and the molecular weights of the oligomers [12].
*
Corresponding author. Fax: +49 (0)3834/86-4319
E-mail address: heinicke@uni-greifswald.de (J. Heinicke).
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2004.03.029
2004 Elsevier Inc. All rights reserved.

J. Heinicke et al. / Journal of Catalysis 225 (2004) 16–23
17
To learn more on the interplay of the phosphino groups of
the P∩O chelate and the PR3 ligand in the trans position, we
studied the oligomerization of ethylene with catalysts pre-
pared in situ from 2-diphenyl- or 2-dicyclohexylphosphino-
phenol 3a or 3b, respectively, Ni(COD)2 (COD 1,5-cyclo-
octadiene), and a phosphine additive and report here on the
results.
mechanic or electronic pressure gauge, and a safety di-
aphragm. Ethylene was added (see Table 1, pstart ca. 50 bar),
and the autoclave was placed into a preheated bath (100 ◦C)
and heated overnight (ca. 15 h). After cooling unreacted
ethylene was allowed to escape, butenes were condensed
in a cooling trap (−78 ◦C). The remaining product, cooled
to −25 ◦C, was transferred to a flask, and volatiles were
flash-distilled (80–100◦C/1–10−2 Torr) into a cooling trap
(−196 ◦C). Residual waxes or polymers were stirred for 1
day with methanol/hydrochloric acid (1/1), washed with
2. Experimental
1
methanol, and dried. The mp and H NMR data were de-
2.1. General
termined after extraction of lower oligomers with 50 mL
of CH2Cl2; mass loss was usually < 2%. Conversion, iso-
lated products, and characteristic data of polymers are given
in Table 1, and the oligomer composition in Table 2 so far
analyzed by GC and exceeding 5% of the flash distillate.
(Differences between the amount of converted ethylene and
the amount of isolated products are due mainly to loss of
butenes, and in the case of polyethylene also to loss of solu-
ble oligomers.)
All operations with phosphines and catalyst solutions
were carried out under carefully dried, oxygen-free argon,
using Schlenk techniques. Toluene was ketyl-dried and dis-
tilled before use. Ni(COD)2 [24] and the phosphinophenols
3a [25] and 3b [13] were synthesized as reported, and other
chemicals were purchased. Ethylene (99.5%, Air Liquide)
was used without further treatment. GC analyses were car-
ried out using a Siemens SICHROMAT 1-4, column SE54-
CS(25m), 50–280 ◦C, 5 min isotherm, 8 or 12 ◦C/min, ref-
erence n-nonane, or a gas chromatograph Hewlett Packard
5890, column HP-5(30m) (crosslinked 5% PhMe silicone),
40–150 ◦C, 10 min isotherm, 4 ◦C/min. 1H NMR spectra of
polymers were measured on a multinuclear FT-NMR spec-
trometer ARX300 (Bruker) (300.1 MHz) at 100 ◦C, acquisi-
tion time 4.9–5.4 s, delay 1.0 s, using concentrated solutions
of the polyethylene samples prepared by swelling for 1 day
at 120 ◦C under argon in C6D5Br (p-CH, δ = 7.23 as refer-
ence). Melting points were determined with a Sanyo Gal-
lenkamp melting-point apparatus. The maximum turnover
frequency (TOFmax) was determined in experiments carried
out with online pressure registration (HEJU pressure sensor
1–100 bar from Juchheim connected to digital multimeter
and PC) from the slope of the steepest (nearly) linear pres-
sure decrease (pressure vs ethylene amount in 20 mL of
toluene calibrated at 100 ◦C after heating for 1 h; effects of
products on the solubility of ethylene neglected).
3. Results and discussion
3.1. Introductory remarks to the catalyst system
Recently we have shown that catalysts generated in situ
from equimolar amounts of 2-diphenyl- or 2-dicyclohexyl-
phosphinophenol (3a or 3b) and Ni(COD)2 polymerize eth-
ylene like methallylnickel 2-diphenyl- or 2-dicyclohexyl-
phosphinophenolate single-component catalysts with high
selectivity for linear α-olefins. The similarity is due to the
formation of allyl-type precursor complexes detected in so-
lution by 1H, 13C, and 31P NMR spectroscopy [13] and re-
lated activation pathways (Scheme 2). Both catalyst types
derived from 3a furnish low molecular weight polyethylenes
(Mw ca. 3000–5000 g mol−1) while catalysts based on 3b
give much higher chain lengths (Mw ca. 60,000 g mol−1
)
under comparable conditions. Replacement of the solvent
toluene by a variety of O-donor solvents improves the cat-
alyst lifetime and productivity but diminishes the molecular
weights and slightly the reaction rate while the selectivity is
not influenced. Additions of tetrahydrothiophene (THT) or
pyridine (pyr) to the precatalyst solution in toluene have a
much stronger impact on the molecular weight and the reac-
tion rate and improve the catalyst lifetime as expressed by
the pressure time plot (Fig. 1) [13]. The effects of phosphine
additives which are soft Lewis bases like THT are diverse
and dependent on the P substituents, and are reported here in
more detail.
2.2. Oligomerization: general procedure
3a (27.8 mg, 100 µmol) or 3b (29.0 mg, 100 µmol) and
Ni(COD)2 (27.5 mg, 100 µmol) were dissolved in toluene
(each 8 mL) at 0 ◦C, combined, and stirred for 5 min at
0 ◦C and for 10 min at 20 ◦C to give orange or yellow so-
lutions. Then, the corresponding amount of the phosphine
(Table 1) dissolved in toluene (4 mL) was added. Me3P (as
1 M solution in toluene), cHex3P, and TMOP did not in-
duce significant color changes while addition of oTol3P or
mTol3P furnished light (3a/Ni/R3P) or dark red (3b/Ni/R3P),
pTol3P darker and Ph3P deep red solutions for both 3a-
and 3b-based catalysts. The precatalyst solution was stirred
for 2 min and transferred by a syringe (Teflon canula) to a
stainless-steel autoclave (75 mL), equipped with a Teflon-
coated magnetic stirrer, gas inlet and sample inlet valves,
3.2. Influence of the molar ratio of 2-phosphinophenolate,
nickel, and triphenylphosphine
The molar ratio of 2-phosphinophenol and Ni(COD)2 (in
the following abbreviated as 3/Ni) influences strongly the

18
J. Heinicke et al. / Journal of Catalysis 225 (2004) 16–23
Table 1
a
Oligomerization of ethylene with catalysts formed from 3, Ni(COD) , and R P
2
3
c
Exp.
Ligands (µmol)
C H (g),
conversion (%),
TON (mol mol
TOF
Products (g)
Solid polymer:
max
2
4
−1 −1
◦
f
(mol mol
(min)
h
),
mp ( C), molar mass
−1
−1
)
t
(g mol
)
start
1
3a (100) [13]
12.6
8000–23,000
(22–16)
Oligomers (< 0.1)
HDPE (6.0–8.1)
mp 124–126
M 1300–2000
NMR
d
(50–66)
2250–2960
12.6
(76) 3420
14.6
2
3
3a (120)
3a (100)
26,000
(15)
28,000
(13)
Wax (0.5)
HDPE (9.1)
mp 119–120
1600–1700
d
M
NMR
C and C (2.5)
4 6
Oligomers (2.8)
Wax (8.6)
Ph P (100)
3
(95) 4950
4
5
3a (100)
Ph P (200)
3
13.6
(> 99) 4850
10,000
(13)
Butenes (1.0)
Oligomers (8.4)
Wax (3.2)
mp < 50
3a (100)
11.9
Butenes
Ph P (300)
3
(> 99) 4240
Oligomers (2.4)
Wax
6
7
3a (100)
18.0
(97) 6240
20.8
(96) 7130
18.7
(87) 5810
13.4
(95) 4530
13.3
(89) 4210
18.6
(71) 4700
14.0
(70) 3500
10.1
Butenes (> 3.1)
Oligomers (7.7)
Butenes (> 2.3)
Oligomers (11.7)
Butenes (> 4.6)
Oligomers (8.6)
Cond. butenes (1.8)
Oligomers (10.4)
Cond. butenes (1.5)
Oligomers (10.0)
Butenes (> 4.4)
Oligomers (5.4)
Cond. butenes (2.9)
Oligomers (6.9)
Butenes (> 1.1)
Oligomers (5.8)
Butenes (1.8)
Oligomers (4.7)
Wax (6.4)
Ph P (400)
3
3a (100)
6000
(12)
Ph P (500)
3
8
3a (100)
Ph P (700)
3
9
3a (100)
Ph P (800)
3
10
11
12
13
14
3a (100)
Ph P (900)
3
3a (100)
2000
(12)
Ph P (1000)
3
3a (100)
Ph P (1500)
3
3a (100)
8000
(10)
4000
(13)
pTol P (100)
(> 99) 3600
13.0
(99) 4600
3
3a (100)
mTol P (100)
3
15
3a (100)
14.0
(99) 4950
Butenes (7.4)
Oligomers (6.4)
Trace wax
mTol P (300)
3
16
17
18
3a (100)
13.2
(92) 4310
13.7
(51) 2500
9.5
(99) 3350
4500
(10)
12,000
(23)
6000
(23)
Butenes (8.8)
Oligomers (4.0)
Oligomers (0.1)
HDPE (6.5)
Oligomers (0.1)
HDPE (9.2)
(Me/Vin 2.6–3.0,
α-olefins 90%)
Oligomers (0.3)
HDPE 9.8
(Me/Vin 1.9–2.3,
α-olefins 92%)
Butenes (> 3.8)
Oligomers (1.5)
Butenes (> 4.8)
Oligomers (6.8)
Butenes (> 4.4)
Oligomers (2.3)
Oligomers (0.1)
Wax
mTol P (500)
3
3a (100)
mp 124–125
d
oTol P (100)
M
NMR
1900–2100
3
3a (100)
Mes P (100)
3
mp 120–122
M
M
M
4000–4300
4600
1950
NMR
w
n
19
3a (100)
11.5
(88) 3600
16,000
(23)
mp 122–124
b
TMOP (100)
M
M
M
2300–2700
3400
1300
NMR
w
n
20
21
22
23
3a (100)
11.3
(70) 2800
16.0
(72) 3600
15.0
(44) 2150
6.8
300
Me P (100)
(39)
1600
(22)
500
(36)
2000
(15)
3
3a (115)
Et P (115)
3
3a (110)
Et P (330)
3
3a (100)
Cy P (100)
3
(57) 1390
(continued on next page)

J. Heinicke et al. / Journal of Catalysis 225 (2004) 16–23
19
Table 1 (Continued)
c
Exp.
Ligands (µmol)
C H (g),
TOF
Products (g)
Solid polymer:
max
2
4
−1 −1
◦
f
conversion (%),
TON (mol mol
(mol mol
(min)
h
),
mp ( C), molar mass
−1
−1
)
t
(g mol
)
start
24
3a (120)
Cy P (120)
3
15.0
(77)
3560
16,000
(16)
Oligomers (0.5)
Wax (11.1)
(Me/Vin 1.3,
α-olefins 94%)
mp 112–115
840
M
NMR
25
26
27
3b (100) [13]
3b (100)
11.1–15.3
(62–72)
2460–3920
12.9–15.0
(60–70)
3200
15,000
(29)
Oligomers (< 0.1)
HDPE
mp 133–134
d
M
58,900
26,500
h
w
n
M
6000–8000
Butenes (ꢀ 0.3)
mp 96–98
h
h
Ph P (100)
3
(20 –24)
Oligomers (0.2–0.7 )
mp 124–126
h
d
HDPE (8.8–8.0)
M
NMR
800 –2500
3b (100)
Ph P (500)
3
12.8
(99) 4530
8000
(17)
Butenes (0.1)
mp 110–112
M 1100
NMR
Oligomers (0.8)
d
HDPE (11.8)
28
29
30
31
32
3b (100)
13.0
(0)
12.3
(81) 3565
13.8
(73) 3600
12.0
(0)
13.1
(70) 3280
Ph P (1000)
3
3b (100)
11,000
(20)
8000
(19)
Oligomers (< 0.1)
HDPE (9.8)
mp 128–130
2300
mp 126–128
M 2550
NMR
d
pTol P (100)
M
NMR
3
3b (100)
Oligomers (< 0.1)
d
mTol P (100)
HDPE (9.9)
3
3b (100)
mTol P (500)
3
3b (100)
9000
(21)
Butenes (0.6)
mp 131–132
M 6800
NMR
oTol P (100)
Oligomers (< 0.1)
3
d
HDPE (8.5)
33
34
3b (100)
TMOP (300)
13.0
(53) 2460
6000
Oligomers (0.2)
HDPE (6.7)
mp 126–128
2300
b
M
NMR
(Me/Vin 2.5,
d
α-olefins 66%)
3b (100)
12.6
(38) 1710
7000
(33)
Oligomers (0.1)
mp 132–133
4700
b
e
TMOP (700)
HDPE (4.7)
M
NMR
(Me/Vin 2.3,
α-olefins 88%)
g
35
36
3b (100)
Me P (100)
3
12.6
(58) 2600
500
(32)
Butenes
mp 89–93
Oligomers (0.7)
Wax
g
3b (100)
12.7
100
Butenes
Me P (300)
(44) 2000
(48)
Oligomers (2.6)
Wax
3
a
b
c
Ni(COD) 100 µmol, solvent toluene.
2
TMOP is [2,4,6-(MeO) C H ] P.
3
6 2 3
◦
Butenes: amount of condensed gaseous products (cooling trap −78 C) and of butenes in the flash-distilled oligomer solution; oligomers: amount of dist.
oligomers except butenes in the distilled oligomer solution analyzed by GC.
d
e
f
1
Content of α-olefins ꢁ 96%, Me/Vin ratio 1.2–2.0 (based on H NMR integration).
◦
Mass loss on extraction with CH Cl 7%, crude polymer mp 123–126 C.
2
2
1
M
based on H NMR integration; M and M based on GPC (HDPE not extracted with CH Cl ).
w n
2 2
NMR
g
h
Uncertain, low content.
Shorter formation time caused lower molecular weights.
properties of the resulting catalysts. Excess Ni(COD)2 with
respect to a 1/1 molar ratio of 3/Ni is decomposed and di-
minishes the productivity of the catalyst. A small excess of
2-phosphinophenol,e.g., 3a/Ni (1.2/1), is tolerated and even
increases the ethylene conversion (Table 1, Fig. 2) without
marked effects on the molecular weights of the oligomers
and the selectivity while a molar ratio of 2/1 3a/Ni leads
to inactivity. The phosphorus/nickel ratio is not the crucial
point in the suppression of catalytic activity. Triphenylphos-
phine, lacking the 2-hydroxyl group of 3a, can be added in
a 10–15 molar excess to the 3a/Ni (1/1) precatalyst with-
out cessation of the catalytic conversion of ethylene. The
deactivation is thus caused by formation of a stable bis(P∩O-
chelate) complex (cf. [25,26]). However, the conversion of
ethylene in the presence of triphenylphosphine is consid-
erably different from that in its absence. Using catalysts
formed from 3a, Ni(COD)2, and Ph3P in toluene in molar
ratios of 1/1/1 up to 1/1/15, the turnover numbers and cat-
alyst lifetime rise with the Ph3P content until a maximum at
a molar ratio of roughly 1/1/5, then they decrease, but even

20
J. Heinicke et al. / Journal of Catalysis 225 (2004) 16–23
Table 2
Oligomer composition C
4–20
ꢀ
Exp.
C , % (linear α-olefins, %)
n
oligomers
in distillate ,
% (g)
a
C
4
C
6
C
8
C
10
C
12
C
14
C
16
C
18
C
20
3
4
5
6
7
14.1 (2.8)
32.7 (9.4)
13.4 (2.4)
38.5 (10.6)
43.8 (13.4)
40.7 (11.7)
32.4 (8.0)
27.0 (6.8)
20.3 (4.8)
24.7 (6.5)
17.2 (4.0)
18.7 (4.1)
36.2 (9.8)
20.2 (4.5)
4.2 (0.7)
–
11.3 (95)
10.9 (78)
29.1 (56)
24.0 (80)
30.4 (91)
33.0 (95)
39.1 (96)
22.6 (46)
22.5 (58)
32.8 (38)
44.6 (88)
26.2 (94)
30.0 (94)
15.5 (90)
27.2 (83)
18.8 (94)
9.7 (94)
30.4 (96)
6.1 (80)
29.7 (94)
2.7 (78)
24.3 (54)
11.5 (74)
14.2 (85)
10.6 (88)
7.2 (88)
13.8 (44)
20.2 (59)
18.8 (35)
13.9 (78)
1.9 (84)
4.7 (93)
0.2
20.2 (94)
1.0 (80)
11.7 (57)
7.6 (74)
8.6 (85)
5.4 (88)
2.5 (92)
9.9 (44)
17.6 (58)
11.0 (38)
5.5 (79)
0.4
1.8
< 0.01
10.0 (96)
19.5 (94)
21.9
6.6 (95)
0.2
1.4
0.3
0.05
b
11.4 (90)
0.7 (68)
27.0 (90)
12.5
26.6 (> 99)
32.0 (> 99)
15.4 (70)
1.7
0.6
1.7
63.9 (91)
> 41 (99)
> 70 (99)
17.6 (72)
3.9 (98)
0.4
30.3 (60)
16.9 (81)
22.0 (91)
19.7 (93)
17.9 (95)
18.6 (44)
22.7 (61)
31.0 (45)
32.4 (89)
7.5 (89)
14.1 (96)
2.9 (82)
24.1 (99)
23.3 (94)
19.5 (91)
28.3 (98)
2.9 (59)
4.9 (72)
5.2 (81)
2.6 (90)
0.9 (96)
6.3 (44)
13.9 (56)
4.6 (36)
1.7 (76)
0.1
0.7 (65)
3.2 (69)
3.1 (78)
1.1 (96)
0.3
4.2 (41)
1.3 (54)
0.9 (34)
0.2
0.2
0.1
1.9 (70)
1.8 (78)
0.5
0.08
2.9 (40)
0.1
1.1 (73)
1.0 (80)
0.2
0.02
2.1 (40)
0.02
8
11
13
14
15
16
20
21
22
26
27
35
36
0.2 (40)
0.1 (75)
1.1
19.2 (97)
23.3 (93)
20.9
1.9
9.1 (93)
19.7
5.9 (1.2)
3.3 (0.6)
12.4 (2.6)
1.9
7.0
1.0
0.2
0.8
0.3
0.1
4
32 (98)
21.1 (98)
10.0 (97)
3.2
a
◦
−2
Solvent and volatile oligomers flash distilled at 80–100 C/1–10 Torr.
Higher boiling components not distilled (waxy residue, 3.2 g).
b
Fig. 1. Pressure–time plots for batch polymerization of ethylene with cat-
alysts formed in situ from 3a and Ni(COD) : (a, solid line) in toluene,
2
(b) in thf, (c) in toluene/pyr (9/1), (d, dotted line) in toluene/tht (9/1) (each
Scheme 2.
0.1 mmol 3a and Ni(COD) , solvents 20 mL, p
50 bar, bath temperature
start
2
◦
100 C) [13].
for a 15-fold molar excess of Ph3P the conversion of ethyl-
ene exceeds or reaches that in absence of Ph3P (Table 1). The
reaction rate, illustrated by the pressure-time plot (Fig. 2),
is increased in the presence of a small amount of triph-
enylphosphine (1/1/1) and likewise in the case of a small
excess of 3a (1.2 3a/Ni), probably by prevention or retarda-
tion of catalyst deactivation. Larger portions of Ph3P com-
pete with ethylene for coordination and result in an increas-
ing decrease of the reaction rate while the heating/induction
period1 is slightly shortened. The chain lengths of the result-
ing α-olefins decrease with growing Ph3P/3a ratio. The cat-
alyst 3a/Ni/Ph3P (1/1/1) converts ethylene mainly to waxy
oligomers while catalysts with higher portions of Ph3P give
rise to lower liquid oligomers and butenes (Table 2). For cat-
alysts 3a/Ni/Ph3P (1/1/n) n ꢁ 5 the oligomer distribution
fits approximately with a Schulz–Flory distribution [27–29].
The β value, the ratio of the rates of elimination and chain
growth reactions, is about 0.6, 0.9, 1.4 for n = 5, 7, 10.
The selectivity for linear α-olefins is lower than in the poly-
merization of ethylene by 3a/Ni [13] and ranges from 60 to
> 95% with a minimum for the catalyst 3a/Ni/Ph3P (1/1/3).
3.3. Influence of basicity and steric demand of the
phosphine additive
1
◦
The reaction starts around 70 C, reached 12–14 min after the begin-
ning of heating. If the precatalyst solution is added to a preheated solution
of ethylene in toluene (80 C) the induction period is very short but the life-
time of the catalyst is low.
◦
Studies with other phosphine additives reveal that the
effects on the conversion of ethylene in the presence of

J. Heinicke et al. / Journal of Catalysis 225 (2004) 16–23
21
Fig. 2. Pressure–time plots for batch oligomerization of ethylene, catalysts
formed from 3a, Ni(COD) , and Ph P in toluene: (a) 1/1.2/0, (b) 1/1/1,
2
3
Fig. 4. Pressure–time plots for batch oligomerization of ethylene with:
(c) 1/1/2, (d) 1/1/5, (e) 1/1/10.
(a) 3a/Ni (reference) and with 3a/Ni/trialkylphosphine, (b) Me P (1/1/1),
3
(c) Et P (1/1/1), (d) Et P (1/1/3), (e) cHex P (1/1/1).
3
3
3
nishes low molecular weight oligomers (Table 2). Increasing
concentration of the small trialkylphosphine increases the
effect on the reaction rate and molecular weights but dimin-
ishes the conversion of ethylene by deactivation at a higher
residual ethylene concentration (pressure). The rather more
basic but much bulkier tricyclohexylphosphine (pKa 9.7,
θ 170◦ [30,31]) behaves like more bulky triarylphosphines
as tris(o-tolyl)phosphine (θ 194◦), trimesitylphosphine (θ
212◦), or tris(2,4,6-trimethoxyphenyl)phosphine (PTMOP)
and allows maximum turnover frequencies and formation
of polymer chain lengths similar to those in the absence
of a phosphine additive. The formation time, however, is
extended and the more basic cHex3P, Mes3P, and PTMOP
additives stabilize the catalyst and cause higher productiv-
ity than in its absence (Figs. 3 and 4, Table 1). A lower
ethylene to catalyst ratio, e.g., conversion of ethylene by
3a/Ni/cHex3P (1/1/1) at pstart 30 bar, leads to slower re-
action rates and conversions. The R3P additives diminish
the selectivity for linear α-olefins in oligomers with short
chain lengths, but not to a significant extend in the polymers
(ꢁ 90% α-olefins). Finally, it should be noted that an addi-
tion of triethylphosphite, possessing low P basicity but good
π-acceptor properties, causes inactivity already at a molar
ratio 3a/Ni/(EtO)3P (1/1/1).
Fig. 3. Pressure–time plots for batch oligomerization of ethylene with
catalysts 3a/Ni/triarylphosphine (1/1/1); R P: (a) Ph P, (b) pTol P,
3
3
3
(c) mTol P, (d) oTol P, (e) Mes P, (f) [2,4,6-(MeO) C H ] P.
3
3
3
3 6 2 3
3a/Ni/R3P (1/1/1) catalysts depend strongly on the basic-
ity and bulkiness of the added phosphine R3P (Tables 1
and 2). Replacement of triphenylphosphine (pKa 2.73,
θ 145◦ [30,31]) by the more basic tris(p-tolyl)phosphine
(pKa 3.84, θ 145◦) leads to a strong decrease of the
oligomerization rate (Fig. 3) and a strong drop of the mole-
cular weights and the selectivity for linear α-olefins from
about 95 to 45%. Replacement by tris(m-tolyl)phosphine
(pKa 3.3, θ 165◦) which is less basic but bulkier than pTol3P
causes a slower conversion, similar low molecular weights,
and somewhat lower loss in selectivity (60–70% α-olefins).
An increase of the amount of mTol3P (to 1/1/3 and 1/1/5)
has no marked effect on the reaction rate but similar to Ph3P
the selectivity is lowest for the 1/1/3 molar ratio. The slow-
est reaction rates are exhibited in those cases where small
trialkylphosphines (Me3P, Et3P; pKa 8.65, 8.69; θ 118◦,
132◦ [30,31]) are added (Fig. 4, Table 1). The small space
demand along with the high basicity and soft character al-
lows rather strong coordination (for PMe3 stronger than
PEt3) on account of the interactions with ethylene and other
donors but does not block the catalytic conversion and fur-
3.4. Influence of the P basicity of the 2-phosphinophenol
The nickel catalyst formed from the more P-basic 2-
dicyclohexylphosphinophenol 3b and Ni(COD)2 gives rise,
as noted above, to much higher molecular weights in the
polymerization of ethylene than the 3a/Ni catalyst. This
feature is maintained, though in much lower measure, in
the presence of phosphine additives. The catalyst systems
3b/Ni/ar3P (1/1/1) (ar3P = Ph3P, pTol3P, mTol3P, oTol3P,
Mes3P, TMOP) convert ethylene mainly to low molecular
weight high-density polyethylenes (HDPE), and the molec-
ular weights of the HDPE formed with catalysts from 3b/Ni
and bulky arylphosphines, oTol3P or TMOP, are higher than

22
J. Heinicke et al. / Journal of Catalysis 225 (2004) 16–23
those of HDPE obtained with related catalysts based on
3a. Excess of triarylphosphines increases the productivity,
e.g., in 3b/Ni/Ph3P (1/1/5), or stops the catalytic activity of
3b/Ni catalysts, e.g., 3b/Ni/Ph3P (1/1/10) or 3b/Ni/mTol3P
(1/1/5), rather than to shift the conversion toward markedly
lower oligomers as observed for 3a/Ni/Ph3P (1/1/n) (n = 1
to 15) or 3a/Ni/mTol3P (1/1/n) (n = 1, 3, 5). Finally, even
the addition of the small and highly P-basic trimethylphos-
phine causes mainly formation of waxy polymers. However,
in this case the amount of lower oligomers is higher for
the catalyst 3b/Ni/Me3P (1/1/3) as compared to 3b/NiMe3P
(1/1/1) (Tables 1 and 2).
The catalyst systems 3b/Ni/R3P distinguish themselves
from 3a/Ni/R3P not only by different effects on the mole-
cular weights and catalyst deactivation, but also by the
impact of triarylphosphines on the reaction rate. Under
comparable conditions, Ph3P and mTol3P cause similar
drops of the maximum reaction rates (TOFmax 6000–9000
mol mol−1 h−1) while the bulkier phosphine oTol3P displays
a weaker influence (TOFmax 11,000 mol mol−1 h−1). Addi-
tion of the highly basic but even more bulky triarylphosphine
TMOP (three and seven molar excess) cause only slightly di-
minished reaction rates (6000–7000 mol mol−1 h−1). In all
cases, the reaction rates decrease strongly after a more or
less short period of a steep pressure drop in the beginning
(Fig. 5). As the rate of the polymerization of ethylene by
3/Ni in toluene is diminished by addition of olefins [13] or
by use of liquid α-olefins in place of toluene, in case of
3b/Ni much stronger than by 3a/Ni [32], this is attributed to
interactions of the catalyst 3b/Ni/R3P (R = aryl) with the α-
olefins formed. Excess of triphenylphosphinesuppresses this
influence as shown by the rapid conversion of ethylene (99%
within 1.1 h) to HDPE by 3b/Ni/Ph3P (1/1/5) (Table 1).
In contrast to the different response of 3a/Ni and 3b/Ni
to triarylphosphine additives, the influence of the small and
highly basic trimethylphosphine on the oligomerization rate
is similar and strong in both 3a- and 3b-based nickel cata-
lysts.
4. Discussion
The above results show that the two types of phos-
phine ligands in in situ-formed nickel phosphinophenol(ate)
phosphine catalysts, 2-R2PC6H4OH and R3P, have different
functions in the oligomerization of ethylene. The necessity
of one and the inactivity of systems with two phosphinophe-
nol per nickel give evidence that the P, O ligand forms the
essential structural feature, the stabilizing backbone of the
nickel catalyst as seen in the known square-planar single-
component organonickel phosphinophenolate phosphine
oligomerization catalysts. The strong trans effect of the
phosphino group allows rapid exchange of the trans-ligand
by the nucleophiles present in the solution, ethylene, phos-
phine additives, and solvent, deactivation by a free phos-
phinophenolate ligand, or interactions with a β-hydrogen
atom of the Ni-alkyl chain to start the olefin elimination
step. The different properties of the catalyst systems based
on 3a and 3b, respectively, indicate a distinct influence of
the 2-diphenylphosphino and the 2-dicyclohexylphosphino
group on the P–Ni bond in the P∩ONi chelate and the
interplay with the nucleophiles, controlled by σ-bonding,
π-bonding, and steric factors. Such differences were also
observed for defined variously P-substituted methylnickel 2-
phosphinophenolate trimethylphosphine single component
oligomerization catalysts [12], and a comparison may help
to understand the behavior of the in situ catalysts.
Diisopropylphosphino derivatives were found to be more
fluxional (sharp 31P NMR doublets only at −70 ◦C, diphe-
nylphosphino derivatives −40 to −30 ◦C), to be catalytically
more active (higher TON), and to give higher molecular
weight oligomers than diphenylphosphino derivatives [12].
It is supposed that the more rapid replacement of PMe3 in
the complex by solvent (in absence of ethylene) and the
longer oligomer chain lengths on use in ethylene oligomer-
ization (smaller relative rate of olefin elimination versus
chain growth) are due to weaker bonding or interactions
of σ-donors (PMe3, β-hydrogen atom of Ni-alkyl in tran-
sition state) opposite to the basic iPr2P ligand, and that
the higher catalytic productivity is due to stronger fixation
of the π-acidic ethylene (higher equilibrium concentrations
of intermediate ethylene complexes) and subsequent inser-
tion of ethylene into the Ni–C bond. The behavior of the in
situ-formed catalysts 3a/Ni/R3P or 3b/Ni/R3P fits similarly
with the assumption of higher equilibrium concentrations
of intermediates with balanced charge at nickel as a main
controlling factor (Scheme 3). Thus, the decreasing rate of
ethylene conversion by 3a/Ni/triarylphosphine with increas-
ing concentration or basicity of triarylphosphines (so far as
they are not too bulky) can be ascribed to favored coordina-
tion of phosphines (σ-donors, weak π-acceptors) opposite
to the Ph2P donor site, while the lower effect of triarylphos-
phines and stronger impact of olefins on the reaction rate
in catalyses with 3b/Ni/triarylphosphine (P basicity of 3b
higher than of 3a) points to a shift in favor of trans in-
teractions to ethylene or olefins (π-donors, π-acceptors).
Fig. 5. Pressure–time plots for batch oligomerization of ethylene with 3b/Ni
and (a) no additive (reference), (b) Ph P, (c) mTol P, (d) oTol P, (e) Me P,
3
3
3
3
each in 1/1/1 molar ratio.

J. Heinicke et al. / Journal of Catalysis 225 (2004) 16–23
23
Scheme 3.
The effect of small donors like trimethylphosphine remains
strong, however, and suggests that the steric situation is also
an important factor in the subtle balance of intermediates
controlling the conversion of ethylene. The lowered selec-
tivity of 3a/Ni/R3P catalysts containing pTol3P or mTol3P
additives might be induced by a competing trans effect of
the relatively basic triarylphosphine ligands which lower the
stability of the P∩O chelate.
[2] W. Keim, New J. Chem. 18 (1994) 93.
[3] K.A. Ostoja-Starzewski, J. Witte, in: R.P. Quirk (Ed.), Transition
Metal Catalyzed Polymerizations [Proc. Int. Symp.], second ed., Cam-
bridge Univ. Press, Cambridge, UK, 1988, p. 472.
[4] G.J.P. Britovsek, V.C. Gibson, D.F. Wass, Angew. Chem. Int. Ed. 38
(1999) 428.
[5] S.D. Ittel, L.K. Johnson, M. Brookhardt, Chem. Rev. 100 (2000) 1169.
[6] S. Mecking, Coord. Chem. Rev. 203 (2000) 325.
[7] W. Keim, F.H. Kowaldt, R. Goddard, C. Krüger, Angew. Chem. Int.
Ed. Engl. 17 (1978) 466.
[8] W. Keim, A. Behr, B. Gruber, B. Hoffmann, F.H. Kowaldt, U. Kür-
schner, B. Limbäcker, F.P. Sistig, Organometallics 5 (1986) 2356.
[9] K.A. Ostoja-Starzewski, L. Born, Organometallics 11 (1992) 2701.
[10] D. Matt, M. Huhn, J. Fischer, A. De Cian, W. Kläui, I. Tkatchenko,
M.C. Bonnet, J. Chem. Soc. Dalton Trans. (1993) 1173.
[11] P. Braunstein, Y. Chauvin, S. Mercier, L. Saussine, A. De Cian, J. Fi-
scher, Chem. Commun. (1994) 2203.
[12] J. Heinicke, M. He, A. Dal, H.-F. Klein, O. Hetche, W. Keim,
U. Flörke, H.-J. Haupt, Eur. J. Inorg. Chem. (2000) 431.
[13] J. Heinicke, M. Köhler, N. Peulecke, M. He, M.K. Kindermann,
W. Keim, G. Fink, Chem. Eur. J. 9 (2003) 6093.
[14] J. Heinicke, M. Koesling, R. Brüll, W. Keim, H. Pritzkow, Eur. J. In-
org. Chem. (2000) 299.
[15] K. Hirose, W. Keim, J. Mol. Catal. 73 (1992) 271.
[16] M.O. Kristen, J. Heinicke, W. Keim, M. Köhler, M. He, DE Patent
19,955,454 (1999), to BASF.
[17] A. Held, F.M. Bauers, S. Mecking, Chem. Commun. (2000) 301.
[18] F.M. Bauers, S. Mecking, Macromolecules 34 (2001) 1165.
[19] U. Klabunde, R. Mülhaupt, T. Herskovitz, A.H. Janowicz, J. Ca-
labrese, S.D. Ittel, J. Polym. Sci. Part A: Polym. Chem. 25 (1987)
1989.
[20] V.C. Gibson, A. Tomov, Chem. Commun. (2001) 1964.
[21] J. Pietsch, P. Braunstein, Y. Chauvin, New J. Chem. (1998) 467.
[22] K.A. Ostoja-Starzewski, J. Witte, Angew. Chem. Int. Ed. Engl. 24
(1985) 599.
[23] K.A. Ostoja-Starzewski, J. Witte, Angew. Chem. Int. Ed. Engl. 26
(1987) 63.
[24] B. Bogdanovic, M. Kröger, G. Wilke, Liebigs Ann. Chem. 699 (1966)
1.
5. Conclusions
Catalysts for the oligomerization of ethylene to α-olefins
of defined molecular weight ranges and selectivity can
be obtained from 2-diphenyl- or 2-dicyclohexylphosphino-
phenol, nickel bis(1,5-cyclooctadiene) and an appropriate
amount of a suitable tertiary phosphine additive in toluene.
The 2-phosphinophenols form the backbone of the nickel
catalyst while the phosphine additives, competing with eth-
ylene and olefins for coordination in the trans position to the
2-phosphino group, allow tuning of the catalyst properties.
A rough estimation of the effects of other substituents or ad-
ditives and of the scope and limits for application of such
catalysts in the ethylene oligomerization should be feasible
based on the above results and assumptions made to under-
stand the in part controversial behavior of 3a/Ni/PR3 and
3b/Ni/PR3 catalysts.
Acknowledgments
We are grateful to the Deutsche Forschungsgemeinschaft
for the support of this study and grants for M.K. and to the
Fonds der Chemischen Industrie for financial support. Fur-
thermore, we thank M.K. Kindermann and B. Witt (Greif-
swald) for NMR and P. Montag (PPS Mainz) for GPC
measurements, and H. Meltzow, G. Peters (Aachen), and
P. Lobitz (Greifswald) for GC analysis.
[25] T.B. Rauchfuss, Inorg. Chem. 16 (1977) 2966.
[26] J. Heinicke, M. Köhler, N. Peulecke, W. Keim, P.G. Jones, Z. Anorg.
Allg. Chem. (2004), in press.
[27] G.V. Schulz, Z. Phys. Chem. B 30 (1935) 379.
[28] G.V. Schulz, Z. Phys. Chem. B 43 (1939) 25.
[29] P.J. Flory, J. Am. Chem. Soc. 62 (1940) 1561.
[30] C.A. Tolman, Chem. Rev. 77 (1977) 313.
[31] C.A. McAuliffe, in: G. Wilkinson, R.D. Gillard, J.A. McCleverty
(Eds.), Comprehensive Coordination Chemistry, vol. II, Pergamon,
Oxford, 1987, p. 990.
References
[1] W. Keim, Angew. Chem. Int. Ed. Engl. 29 (1990) 235.
[32] J. Heinicke, N. Peulecke, unpublished results.