Phospholyl catalysts for olefin polymerization

Article abstract of DOI:10.1016/S1381-1169(97)00170-2

Seventeen different phospholyl ligands were incorporated in a total of 22 zirconium complexes, (Phos)₂ZrCl₂, (PhosXC₅H₅)ZrCl₂, investigated in propylene polymerization catalysis using methylaluminoxane as cocatalyst. Atactic polypropylene with M(n) varying from 450 to > 20000 and vinylidene end groups (CH₂=C(Me)R) was obtained with activities up to 170 kg/g Zr· h. For the 11 diphospholyls of structure (2,5-R₂C₄H₂P)₂ZrCl₂, catalytic activity was highest with substituents of moderate bulk adjacent to phosphorus (e.g., c-Pr, Ph), whereas complexes with two small (H) or two large (CMe₃, SiMe₃) ligand substituents were inactive. It is hypothesized that optimum activity with MAO requires selective blocking of phosphorus lone pair coordination to aluminum, whilst allowing free propylene approach to the active site. The degree of polymerization increased steadily in the series of 2,5-disubstituted phospholyl complexes, dialkyl < alkyl-phenyl < diphenyl, suggesting that electronic factors are more important than steric factors in determining M(n).

Full text of DOI:10.1016/S1381-1169(97)00170-2

Ž
.
Journal of Molecular Catalysis A: Chemical 128 1998 155–165
Phospholyl catalysts for olefin polymerization
a
b
a
Eric J.M. de Boer , Ian J. Gilmore , Frans M. Korndorffer ,
Andrew D. Horton a,), Arjan van der Linden , Bruce W. Royan , Bart J. Ruisch ,
Lodi Schoon , Robert W. Shaw
Shell Research and Technology Centre, Amsterdam, Postbus 38000, 1030 BN Amsterdam, The Netherlands
Shell AdditiÕes International Ltd., cro Shell Research and Technology Centre, Thornton, P.O. Box 1, Chester, Cheshire CH1 3SH, UK
a
b
a
a
b
a
b
Abstract
Ž
.
Seventeen different phospholyl ligands were incorporated in a total of 22 zirconium complexes, Phos 2ZrCl2,
.Ž
Ž
.
Phos C5H5 ZrCl2, investigated in propylene polymerization catalysis using methylaluminoxane as cocatalyst. Atactic
. .
Ž
Ž
polypropylene with Mn varying from 450 to )20 000 and vinylidene end groups CH2 5C Me R was obtained with
Ž
.
activities up to 170 kgrg ZrPh. For the 11 diphospholyls of structure 2,5-R2C4H2P 2ZrCl2, catalytic activity was highest
Ž
.
Ž .
with substituents of moderate bulk adjacent to phosphorus e.g., c-Pr, Ph , whereas complexes with two small H or two
Ž
.
large CMe3, SiMe3 ligand substituents were inactive. It is hypothesized that optimum activity with MAO requires selective
blocking of phosphorus lone pair coordination to aluminum, whilst allowing free propylene approach to the active site. The
degree of polymerization increased steadily in the series of 2,5-disubstituted phospholyl complexes, dialkyl-alkyl-phenyl
-diphenyl, suggesting that electronic factors are more important than steric factors in determining Mn. q 1998 Elsevier
Science B.V.
Keywords: Phospholyl; Zirconium; Atactic polypropylene; Oligomers; Polymerization; Vinylidene
1. Introduction
fragments, as well as hard heteroatom-based
w
x
ligands, such as amides 7–9 and alkoxides
w
x
The high activity and tunable stereo- and
regioselectivity of metallocene-based olefin
10 , have recently received much attention. The
interest in and potential of five or six-membered
heteroatom-containing rings is illustrated by
w
x
polymerization catalysts 1,2 has spawned an
intensive academic and industrial effort to de-
velop related catalysts in which one or both of
the cyclopentadienyl groups is replaced by an-
other ligand. Alternative unsaturated carbon-
Ž
.
studies of systems containing amino borolyl
Ž
.
w
x
Ž
. w x
C4B ring 11 , azaborolyl C3BN ring 12
Ž
.
w
x
and boratabenzene ligands C5B ring 13,14 .
We anticipated that phospholyl ligands, for
which a well-established synthetic methodology
w x
based ligands, based on the butadiene 3 , penta-
w x
w x
w x
w x
dienyl 4 , benzene 5 or cyclooctatetraene 6
and organometallic chemistry exists 15 , as well
as their analogues containing heavier group 15
elements, would allow a facile entry to a broad
new family of heteroring-based polymerization
catalysts.
)
Corresponding author. Tel.: q 31-20-6302930;
e-mail: horton1@sic.shell.nl
1381-1169r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved.
Ž
.
PII S1381-1169 97 00170-2

(
)
156
E.J.M. de Boer et al.rJournal of Molecular Catalysis A: Chemical 128 1998 155–165
Unlike pyrolyl ligands, which, unless con-
One of the goals of this work was to study
atactic oligo- and polypropylene formation. In
contrast to the effort expended on isotactic and
syndiotactic polypropylene, aspecific poly-
merization with metallocenes has received much
taining bulky substituents adjacent to the het-
eroatom, tend to distort towards h1-coordination
w
x
of nitrogen to electrophilic metals 16,17 , phos-
pholyl ligands generally adopt a cyclopentadi-
5
w
x
w
x
enyl-like h -coordination 18,19 . Further, theo-
retical studies have indicated that phospholyl
ligands exhibit similar electronic characteristics
less attention 1,2,36 . Metallocene-derived atac-
tic polypropylene might replace the material
previously formed as a by-product in Ziegler–
Natta polymerization and incorporated in bitu-
men and hot melt adhesives, whereas low
molecular weight atactic propylene oligomers
w
x
to cyclopentadienyl ligands 20,21 . A major
difference between metallocene and phospholyl
complexes is the presence of the phosphorus
lone pair, which could play a non-innocent role
in catalysis. For example, coordination to the
Lewis acidic alkylaluminoxane cocatalyst might
Ž
.
APO’s may afford interesting new materials
via the functionalization of the unsaturated end
w
x
group 36–44 . Regioselective propylene poly-
Ž
.
Ž .
block the cationic active site and alter the
merization 1,2-insertion most commonly af-
w
x
Ž
.
ligand steric properties 20–22 . Alternatively,
the formation of dimetallic complexes via well-
precedented h1,h5-bridging coordination of the
phospholyl might result in catalyst deactivation
fords vinylidene end groups, CH2 5C Me R, by
b-hydrogen eliminationrtransfer; vinyl end
groups, CH2 5CHCH2C Me R, are obtained by
b-methyl elimination with certain sterically
Ž
.
w
x
w x
23 . Variation of the ligand steric bulk in the
crowded metallocenes 42–44 . Functionaliza-
Ž
tion of vinylidene-capped APO’s with polar
vicinity of phosphorus might, therefore, be ex-
pected to effect catalytic behaviour signifi-
cantly. For this reason, and in order to allow a
systematic study of ligand effects in phospholyl
catalysis, a project was initiated to utilize exist-
ing methodology to synthesize a wide range of
different phospholyl ligands and corresponding
zirconium complexes.
.
w
x
functionalities or aluminum 37–41 has af-
forded oligomers, which may, for example, be
used for block copolymer formation. Vinyl-
capped APO’s have found use as monomers in
w
x
copolymerization 37–41 . We believed that by
appropriate control of the steric and electronic
environment in phospholyl complexes, we could
tune the activity, product molecular weight and
end groups in atactic propylene oligorpoly-
merization.
Despite the fact that a range of phospholyl
ligands is known, the number of different phos-
pholyl ligands applied in Early Transition Metal
chemistry is limited to three, C4Me4P, 3,4-
Ž
Me2C4H2 P and C4H4P the latter only in a
.
w
x
mono-phospholyl Ti complex 24–28 , the ma-
jor effort having been devoted to later transition
2. Preparation of phospholylzirconium com-
plexes
w
x
metals 23,24 . Since the completion of this
w
x
work 29 , other, primarily Japanese, workers
have reported studies and potential industrial
application of group 4 phospholyl complexes in
Phospholyl complexes of zirconium were ob-
tained in three steps from dienes or diynes
w
x
following modified literature procedures 15 .
The initially formed 1-phenylphosphole precur-
sor is reduced to the phospholyl anion, followed
by reaction with an appropriate zirconium di-
halide to give diphospholyl or mono-phospholyl
complexes. Although a full discussion of the
syntheses is outside the scope of this article, an
overview of the synthetic methods utilized is
w
x
catalytic olefin polymerization 30–35 . How-
ever, this work of limited scope utilized only
the narrow range of existing phospholyl com-
plexes, with the greatest emphasis on ethylene
co polymerization. Here we report the first sys-
tematic study of ligand effects in polymerization
Ž
.
w
x
catalysis with phospholyl complexes 29 .

(
)
E.J.M. de Boer et al.rJournal of Molecular Catalysis A: Chemical 128 1998 155–165
157
Fig. 1. Synthetic routes to 1-phenylphospholes.
given below. The preparation of four representa-
tive diphospholyl complexes is described in
Section 2 .
Anionic phospholyl ligands are most conve-
niently obtained from the corresponding 1-
phenylphosphole by reductive cleavage of the
advantage of the latter method is the use of mild
w
x
experimental conditions 45–47 . This obviates
the phenylralkyl scrambling between phospho-
1
Ž
rus and carbon via a concerted 1,5-hydrocarbyl
shift which can lead to mixtures of phospholes
at the high temperatures 150–2008C used in
the diene route 48 . An additional disadvantage
of the diene route is that the permitted sub-
stituents are limited to phenyl groups and small
.
Ž
.
w
x
w x
P–Ph bond 15 . Two general routes to substi-
tuted 1-phenylphospholes, which allow facile
Ž
.
substituent variation, have been utilized Fig. 1 :
Ž .
w x
i Reaction of 1,3-dienes with PhPCl2 under
forcing conditions with HCl elimination 15 .
ii Direct addition of PhPH2 to 1,3-diynes at
ambient temperatures using a lithium hydrocar-
alkyl groups 15 . Prior to this work, the diyne
route had been applied to the synthesis of only a
w
x
Ž .
w
x
limited range of phospholyl ligands 45–47 .
The known 1-phenylphospholes, C4Me4PPh
w
x
w
x
w x
byl catalyst 45–47 .
24 and C4Ph4PPh 15 , were obtained by al-
ternative routine literature procedures, without
general applicability.
Whereas the diene route permits the synthesis
of phospholyl ligands with up to four sub-
stituents, the diyne route is restricted to 2,5-dis-
ubstituted derivatives. However, a significant
Zirconium complexes with 17 different phos-
pholyl ligands were obtained, as shown in Fig.
2, by reductive cleavage of the P–Ph bond of
the 1-phenylphosphole, followed by reaction of
the phospholyl anion with ZrCl4 or, for mixed
complexes, C5H5 ZrCl3. The phenylalkali by-
product of the reaction of the phenylphosphole
with excess lithium or sodium was normally
1
w
x
Synthetic details are contained in Ref. 29 . Characterization
of new phospholyl complexes has included X-ray structural stud-
ies, which will be published elsewhere. These studies show that
complexes with two crowded 2,5-disubstituted phospholyl ligands
exhibit bent metallocene structures with the phosphorus atoms at
the open end of the wedge formed by the two rings. In contrast in
Ž
.
Ž
quenched by reaction with t-BuCl 1.0 equiva-
lent or AlCl3 1r3 equivalent . The crude
Ž
.
C4 Me4 P 2 ZrCl2 , the phosphorous atoms are located at the back
end of the wedge 24 .
. Ž .
w
x

(
)
158
E.J.M. de Boer et al.rJournal of Molecular Catalysis A: Chemical 128 1998 155–165
Fig. 2. Synthesis of zirconium phospholyl complexes.
Ž
phospholyl anion contaminated with LiCl or
group are of interest for a range of applications.
.
Ž
NaCl , was generally obtained as a spectroscop-
The oligorpolymerization of propylene 600
Ž31
.
.
ically pure P-NMR solid after extraction and
washing steps. Zirconium complex formation
was carried out in toluene, diethyl ether or
mixtures of these solvents. The generally spec-
kPa was investigated under standard conditions
in toluene solution 458C, 10–60 min using
Ž
.
MAO as cocatalyst. The amounts of zirconium
Ž
.
Ž
complex 10 mmol and MAO 5.0 mmol alu-
Ž31
.
.
troscopically pure
P-NMR solid yellow or
minum used resulted in an AlrZr ratio of 500
orange complexes, obtained by solvent extrac-
tion of the crude residue, reduction to dryness
and washing with hexane, were used directly for
polymerization testing, without further recrystal-
lization. Yields of isolated diphospholyl com-
and an aluminum concentration of 22.4 mM.
Under the polymerization conditions, the gener-
ally atactic product is soluble in toluene. The
advantage of studying a truly homogeneous
polymerization is that complications caused by
Ž
Ž
plexes in the range 10–50% based on starting
polymer precipitation e.g., catalyst deactivation
.
.
1-phenylphosphole were obtained.
on heterogenization , can be avoided and activ-
ity comparisons of different catalysts are ex-
pected to be more meaningful.
3. Propylene polymerization
3.1. Catalytic actiÕity
The major goal of this investigation was to
study the effect of ligand variation on the poly-
merization activity and product structure in
phospholylzirconium catalysis. As already dis-
cussed, APO’s containing a vinylidene end
Of the seventeen diphospholyl complexes ex-
Ž
amined, twelve afforded atactic or near-atactic,
.
see below polypropylene or propylene
oligomers and five were catalytically inactive.

(
)
E.J.M. de Boer et al.rJournal of Molecular Catalysis A: Chemical 128 1998 155–165
159
Table 1
Propylene polymerization activity of phospholylzirconium complexes
Complex
Type
Substituents
R2
Activity
Mn
R3
R4
R5
kgrgZrPh
Ž
.
1
2
P2 Zr
P2 Zr
P2 Zr
P2 Zr
P2 Zr
P2 Zr
P2 Zr
P2 Zr
P2 Zr
P2 Zr
P2 Zr
P2 Zr
P2 Zr
P2 Zr
P2 Zr
P2 Zr
P2 Zr
H
Me
Me
H
H
H
H
H
H
H
Me
Me
H
H
H
H
H
H
H
H
H
H
H
Me
Ph
H
H
—
—
8.1
36.4
104.2
115.3
—
—
—
Me
n-Pr
i-Pr
c-Pr
c-Pe
t-Bu
Me3Si
Ph
Ph
Ph
Ph
Ph
Ph
Ph
An
Phen
H
Me
t-Bu
Ph
Me
n-Pr
i-Pr
c-Pr
c-Pe
t-Bu
Me3Si
Me
n-Pr
c-Pe
Ph
Ph
Ph
Ph
Me
Me
H
Me
t-Bu
Ph
3
1 600
1 400
4 100
1 500
4a
5
6
7a
8
—
—
—
9
92.2
83.0
108.8
169.6
82.9
15.2
—
55.3
32.2
—
6 700
5 100
4 600
10
11
12
13
14
15
16
17
18
19
20
21
22
H
H
H
)20 000
20 000
450
Me
Me
Ph
H
—
Me
Me
H
—
7 500
)20 000
—
H
Ž
.
.
.
.
P Cp Zr
Me
Me
H
H
Me
Ž
P Cp Zr
—
7.8
64.5
4.6
—
Ž
P Cp Zr
1 400
1 700
3 000
Ž
P Cp Zr
H
Me
P2 Zr.W
Me
Me
4
Ž
.
CO
Ža
.
Standard conditions: 10.0 mmol Zr, 5.0 mmol MAO complexes 4 and 7: 20.0 mmol Zr, 10.0 mmol Al , total volume toluene 223.5 ml,
pressure propylene 600 kPa, temp. 458C, polymerization time 10–60 min. Mn and end group content determined by 1H-NMR studies of
isolated product.
Ž
.
Ž
.
Ž
.Ž
.
P2 Zrs Phos 2 ZrCl2; P Cp Zrs Phos C5H5 ZrCl2; c-Pescyclopentyl
Examination of the data in Table 1 shows that
the catalytic activity exhibits a remarkable de-
pendence on the steric bulk of the substituents
of the phospholyl rings. The presence of two
hydrocarbyl substituents adjacent to phospho-
rus, the larger of which is of intermediate steric
pholyl ligands, exhibit high activities in a nar-
Ž
row range in contrast to the greater variation
.
found for 3–6 . These complexes, are, however,
significantly less active than the 2,5-diphenyl-
phospholyl complex, 12, perhaps reflecting the
effect of differences in electron-donating abili-
ties of phenyl and alkyl substituents. Complex
12 exhibits the highest activity of all complexes
tested. Phospholyls with two very bulky sub-
Ž
.
bulk phenyl or CHR2 , appears to be necessary
for optimum catalytic activity. For example, the
increased activity in the series of complexes
3–6, with 2,5-dialkyl-phospholyls, 2,5-
Ž
w x
stituents adjacent to phosphorus t-butyl 47 ,
Ž
.
.
R2C4H2 P, R s n-Pr - i-Pr - c-Pr s c-Pe,
correlates with increased effective steric bulk.
Complexes 9–11, with 2-alkyl,5-phenylphos-
trimethylsilyl are inactive. Similarly, a continu-
ous decrease in activity as the degree and total
bulk of substitution increases is observed in the

(
)
160
E.J.M. de Boer et al.rJournal of Molecular Catalysis A: Chemical 128 1998 155–165
Ž
series of complexes 12–15 activity: Ph2 )
4100, compared to 1400–1600 for the other
2,5-dialkylphospholyl complexes, is consistent
with conjugative interaction of the cyclopropyl
ring with the heteropentadienyl ligand, which
affords to the cyclopropyl group electronic
properties intermediate between an alkyl and a
phenyl group. The p-interaction of two cyclo-
.
Ph2 Me)Ph2 Me2 )Ph4 . The result obtained
with crowded complex 15 is consistent with the
catalytic inactivity of the metallocene analogue,
C5Ph4H ZrCl2rMAO 49 . The inactivity of
the 3,4-dimethylphospholyl complex, 1, and the
2,3,4,5-tetramethylphospholyl complex, 2, again
underlies the apparent necessity for intermediate
steric bulk in the phospholyl adjacent to phos-
phorus; this seems to be more important than
Ž
.
w x
2
Ž
.
propyl ring- C–C bonding orbitals with the
aromatic ring is expected to be optimized by a
w
x
perpendicular orientation of the two rings 50 .
Stabilization of such a conformation may in-
crease the effective steric bulk of the ligand
compared to the di-isopropyl analogue, in which
the methyl substituents likely point away from
the zirconium atom. This in turn may partly
explain the three-fold increase in activity from
the isopropyl complex 4 to the cyclopropyl
analogue 5.
Ž
‘total’ ligand steric bulk for activity the
Me4C4P ligand may, in cyclopentadienyl terms,
.
be considered bulky .
Catalytic activity in the mixed phospholyl-
cyclopentadienyl complexes 18–21 shows a
similar sensitivity to the steric bulk adjacent to
phosphorus. The 2,5-diphenylphospholyl com-
plex is much more active than the t-butyl ana-
logue, whilst the methyl-substituted phospholyl
complexes are inert. Bridging coordination of
A decrease in Mn with increased methyl
substitution is observed in the series of com-
Ž
.
Ž
W CO between the two tetramethylphospholyl
plexes 12–14 substituents: Ph2, Ph2 Me,
4
Ž
.
ligands in complex 22 results in increased al-
beit low activity compared to complex 2. This
Ph2 Me2, respectively , but it is not clear whether
.
this is a steric or electronic effect. Interestingly,
dibenzophosphaindene complex 17 affords rela-
tively higher molecular weight polypropylene
than 9–11, despite the fact that it contains an
alkyl substituent adjacent to phosphorus.
Substituent effects on molecular weight in
the mixed phospholyl-cyclopentadienyl com-
plexes appear to be less pronounced, with com-
plexes 20 and 21, with t-butyl and phenyl sub-
stituents, respectively, at the 2,5-positions, af-
fording APO’s with similar Mn. The range of
complexes studied is, however, too small, to
allow firm conclusions to be drawn.
may reflect ‘protection’ of the phosphorus atom
from interaction with aluminum, although the
carbonyltungsten fragment might be expected to
react with MAO under the polymerization reac-
tion conditions.
3.2. Degree of polymerization
In contrast to the strong dependence of activ-
ity on the steric properties of the phospholyl
ligand, the degree of polymerization found for
the diphospholyl complexes appears to be
closely linked to the electronic properties of the
phospholyl substituents. For the nine active
2,5-disubstituted complexes, the molecular
weight increases in the series dialkyl-alkyl,
phenyl-diphenyl. This trend, with Mn increas-
3.3. End groups and steroregularity
Ž
.
The propylene oligomers Mn -20 000 ex-
hibit H-NMR resonances )95% for vinyli-
1
Ž
Ž
.
Ž
.
Ž
Ž
. .
ing from 1400–4100 3–6 , to 4600–7500 9–
dene end groups CH2 5C Me R , as expected
for a process of chain termination involving
b-hydrogen elimination or transfer to monomer.
R esonances for vinyl end groups
.
Ž
.
11 to values greater than 20 000 12, 17 , ap-
pears to reflect decreased electron-donation to
the heteroatom ring.
Ž
Ž . .
The one complex which, although fitting the
trend, does not do so elegantly, is 2,5-di-
cyclopropylphospholyl 5. The higher Mn of
CH2 5CHCH2CH Me R or internal olefins are
not observed. Certain sterically crowded metal-
locenes have been shown to undergo chain ter-

(
)
E.J.M. de Boer et al.rJournal of Molecular Catalysis A: Chemical 128 1998 155–165
161
mination by b-methyl elimination, with the pro-
portion of vinyl end groups increasing in the
when at least one, and preferably two of the
substituents at the 2,5-positions of the phospho-
lyl ring is a secondary hydrocarbyl group. Suffi-
cient steric bulk adjacent to phosphorus may be
required to suppress lone pair coordination to
Ž
.
o r d e r
Ž
w
C
M e
.
H
Z r C l
-
5
4
2
.Ž
x
2Ž
.
C5Me4 H C5Me5 2 ZrCl2 - C5Me5 2 ZrCl2
36,42–44 . The absence of vinyl end groups in
the phospholyl catalysis reported here shows
that, despite the high steric crowding of some of
w
Lewis acidic aluminum centers in MAO 20–
x
22 . Such coordination could block the approach
Ž
.
the active complexes e.g., 14, 22 , b-methyl
elimination does not occur to a significant de-
gree.
of propylene monomer to the active site, or,
alternatively, lead to phospholyl transfer to alu-
Ž
minum. Very large substituents tertiary atoms
.
C2-symmetric diphospholyl complexes might
be expected to exhibit some degree of isospeci-
ficity in propylene oligorpolymerization. Com-
plexes 9–11, 13 and 17 are present as mixtures
of rac and meso isomers, as reflected in the
observation of two distinct 31P-NMR reso-
nances. 13C-NMR analysis of the methyl region
showed that the propylene oligomers are, in
general, atactic with m:r dyad ratios close to
the ideal value of 50:50. In the case of complex
17, slight isospecificity was found, with a m:r
dyad ratio of 60:40. Given that 17 consists of an
approximately equimolar mixture of rac and
on ring shield not only the phosphorus, but also
the active zirconium center and lead to inactive
2
catalysts .
In c o n tra st to th e c a se fo r
Ž
.
C4Me4P ZrCl2rMAO, which is catalytically
2
Ž
.
inactive for propylene but active for ethylene ,
we have shown in other studies that the MAO-
w
Ž
.
xqw xy
free cationic analogue, C4Me4P ZrMe A
2
Ž
w xy
.
is a poorly coordinating anion , does
A
Ž
polymerize propylene albeit with only moder-
ate activity 29 3. This suggests that cationic
diphospholyl complexes, similar to cationic
metallocenes, should, intrinsically, exhibit activ-
ity for propylene polymerization, provided side
.
w x
Ž
.
meso isomers, and tentatively assuming equal
activities, it may be proposed that the rac iso-
mer affords polymer with a m:r ratio of 70:30.
The generally low or negligible stereospecificity
of the rac isomers in propylene oligorpoly-
merization compares with previous studies of
non-bridged metallocenes with 1-hydrocarby-
Ž
.
reactions such as aluminum coordination can
be blocked.
The clear effect of the electronic nature of
the phospholyl substituents on the degree of
polymerization of propylene, most clearly re-
flected in the series of 2,5-disubstituted-phos-
Ž
.
Ž
lindenyl ligands, in which limited sterospeci-
ficity was observed in certain cases in low
pholyls Mn increases: dialkyl-alkyl-phenyl
.
-diphenyl , is remarkable. Some previous
studies of electronic effects on metallocene
polymerization indicated that the presence of
w
x
temperature polymerizations 51 . It is possible
that the unsymmetrically substituted phospholyl
complexes may exhibit higher sterospecificity at
polymerization temperatures far below that used
in this study.
2 It might be argued that the inactivity of complexes with bulky
phospholyl ligands is due to a reduced rate of cationic alkylzirco-
nium complex formation. However, we have found that cationic
derivatives of complex 7 are instantaneously obtained on methyl-
abstraction from the dimethyl precursor using normal
boranerborate reagents: A.D. Horton and A.J. van der Linden, to
4. Discussion and conclusions
be published.
3
A certain degree of steric bulk adjacent to
phosphorus in the phospholyl ring appears to be
necessary for activity in phospholylzirconium
catalysis. Of the 17 diphospholyl complexes
tested to date, the highest activity is observed
Ž
.
Z rC l 2 r A lR
2
T h e s y s te m ,
xw Ž . x
C
M e 4
P
r
4
3
w
PhMe2 NH B C6 F5 , in which a cationic alkyl species is pre-
4
sumably generated in situ, is reported to exhibit propylene poly-
merization activity 33 . In contrast to our findings, other workers
have claimed low propene polymerization activity for the sys-
tem, C4 Me4 P 2 ZrCl2 rMAO 32,33 .
w
.
x
Ž
Ž
.
w
x

(
)
162
E.J.M. de Boer et al.rJournal of Molecular Catalysis A: Chemical 128 1998 155–165
electron-withdrawing substituents results in a
5. Experimental details
Ž
lowering of polymer molecular weight and ac-
.
tivity , provided the ligand steric bulk in the
immediate vicinity of the metal is relatively
5.1. General procedures
w
x
invariant 52–54 . There is no obvious explana-
tion for the observation of a reverse trend in Mn
in diphospholyl catalysis. It is especially sur-
prising that steric bulk plays such a subsidiary
role to the electronic nature of the substituents,
as shown by the observation of similar Mn
Preparative experiments were performed un-
der nitrogen in a Braun MB 200-G dry box or
under argon using standard Schlenk techniques.
Solvents were dried by refluxing over and dis-
tilling from standard reagents. NMR solvents
Ž
values for dialkyl complexes 3, 4 and 6 n-Pr,
˚
were dried over 4 A molecular sieves. Propy-
.
i-Pr and cyclopentyl substituents, respectively
lene was purified over AlEt3. Dienes were ob-
tained from commercial suppliers 1,4-diphenyl-
1,3-butadiene, 2,3-dimethyl-1,3-butadiene or
following literature procedures: in general, reac-
and the significantly higher Mn for 2,5-di-
Ž
Ž
cyclopropylphospholyl complex 5 proposed to
.
be electronically intermediate between alkyl and
.
phenyl .
This study has shown both the potential of
Ž
tion of trans-cinnamaldehyde or appropriate
.
substituted derivatives with a Grignard reagent
phospholyl complexes as alternatives to metal-
locenes in polymerization catalysis and the im-
portance of extensive ligand variation in tuning
the catalytic properties. In this context, the inac-
tivity of pre-existing diphospholylzirconium
afforded an unsaturated alcohol, which was then
dehydrated using para-toluenesulphonic acid
w
x
55,56 . 1,4-diphenyl-2,3-dimethyl-1,3-
butadiene was obtained from the reaction of
2,3-butadione with benzylmagnesium chloride,
followed by dehydration of the resulting diol
Ž
complexes for propylene polymerization with
3
.
MAO should be noted . Variation of the phos-
w
x
57 . 1,3-diynes, as well as other reagents, were
pholyl substitution pattern has allowed
polypropylene products to be obtained with
molecular weights varying over a wide range
and, in some cases, with high metallocene-like
activities. Comparison of this programme of
ligand variation to the much more extensive
investigations of cyclopentadienyl ligand modi-
fication suggests that further studies may lead,
for example, to stereospecific phospholyl cata-
lysts. Extension of this concept of ligand sub-
stituent variation to pyrolyl, arsolyl and other
hetero-ring ligands is also likely to lead to new
either obtained from commercial suppliers and
used without further purification or prepared by
straightforward literature routes. 31P-NMR spec-
tra were recorded on Varian XL-200 or VXR-
Ž
.
300 80.96 or 121.14 MHz, respectively . NMR
data were recorded at 258C and are listed in
parts per million, referenced relative to
Ž
.
Ž
.
P C6H5 3 in CDCl3 d y6.0 ppm .
5.2. Typical preparations of diphospholyl com-
plexes
4
catalyst families .
Four representative syntheses of diphospho-
lylzirconium complexes are presented below.
Full details of the preparation and characteriza-
Ž1
tion of the new complexes H, 13C, 31P-NMR,
.
elemental analysis, X-ray structural studies are
4 We have found that tetrahydrocarbazole C20 H16
N
com-
plexes, L2 ZrCl2 and L C5H5 ZrCl2 , and the arsolyl complex,
.Ž
1
Ž
.
w
x
Ž
.
discussed elsewhere 29 see . The syntheses
are non-optimized and the isolated yields vary
in the range 10–50%. Emphasis was placed on
Ž
.
Ž
.
C4 Me4 As C5H5 ZrCl2 , exhibit propylene polymerization activ-
w
x
ity with MAO 29 .

(
)
E.J.M. de Boer et al.rJournal of Molecular Catalysis A: Chemical 128 1998 155–165
163
( (
)
obtaining pure complexes for catalytic testing,
rather than on optimum yields. As the isolation
of the complexes in pure form often relies on
crystallization, the yields depend crucially on
the complex solubility.
5.4. Preparation of 2- 4-MeO-3,5-t-Bu2C6 H2 -
)
( )
5-MeC4 H2 P 2 ZrCl2 16
Ž
.
1- 4-MeO-3,5-t-Bu2C6 H 2 -1,3-pentadiene
5.3 g, 18.5 mmol , was allowed to react with
Ž
.
Ž
.
PhPCl2 5.0 g, 27.8 mmol for 16 h at 1408C,
followed by 5 h at 1508C, under a slow argon
flow. Subsequently, the reaction mixture was
cooled to 258C, dissolved in hexane and eluted
over a silica column with hexane. The solvent
was removed in vacuo and the volatile organic
impurities removed by distillation at 2008C,
(
)
5.3. Preparation of 2,5-Ph2-3-MeC4 HP 2 ZrCl2
( )
13
Ž
1,4-Diphenyl-3-methyl-1,3-butadiene 9.1 g,
.
41.4 mmol , was allowed to react with PhPCl2
11.2 g, 62.6 mmol for 17 h at 1708C under a
Ž
.
Ž
leaving 2.4 g of non-volatile product, 1-Ph-2- 4-
slow argon flow. Subsequently, the reaction
mixture was cooled to 258C and 50 ml of
dichloromethane was added. After further cool-
ing of the obtained reaction mixture to 08C,
water was added slowly and the organic layer
was separated and dried. The obtained residue
was dissolved in a small amount of toluene and
chromatographed over silica. Initially petroleum
ether 80r110 was used as eluent and during
elution progressively more toluene was intro-
duced. The first fractions were mainly organic
by-products, followed by a fraction containing
1.3 g of spectroscopically pure 1,2,5-Ph3-3-
.
MeO-3,5-t-Bu2C6H2 -5-MeC4H2 P.
The above phenylphosphole was converted
to 2- 4-MeO-3,5-t-Bu2C6H2 -5-MeC4H2 PNa
THF by reaction with excess sodium, fol-
lowed by quenching with t-BuCl. To a suspen-
Ž
.
Ž
.
n
Ž
.
sion of the anion 1.47 mmol in toluene at
Ž
258C was added a suspension of ZrCl4 0.17 g,
.
Ž
.
0.73 mmol in toluene 10 ml at 258C. Subse-
quently the mixture was centrifuged and the
clear orange solution decanted. The solution
was reduced to dryness in vacuo and the solid
extracted with hexane. On cooling the hexane
solution to y208C, orange crystals of spectro-
Ž31
Ž31
.
MeC4HP P-NMR, CDCl3, d 11.9 ppm .
scopically pure product were obtained
P-
Ž
.
1,2,5-Ph3-3-MeC4HP 0.64 g, 2.0 mmol was
.
NMR, CD2Cl2, d 83.0 ppm . It is not clear
whether the product is the rac or the meso
isomer.
Ž
.
dissolved in THF 40 ml and reacted with
excess lithium. The reaction mixture was stirred
for 45 min, excess lithium was removed by
Ž
.
filtration, and t-BuCl 0.5 ml then added, fol-
lowed by warming to 558C and stirring for 30
min. Subsequently the volatiles were removed
in vacuo and the residue dissolved in diethyl
(
)
( )
5.5. Preparation of 2,5-Ph2C4 H2 P 2 ZrCl2 12
1,2,5-Ph3C4H2 P was prepared by a literature
w
x
route
15
and converted to 2,5-
Ž
.
Ž
.
ether 30 ml . This solution was cooled to
y788C and then a suspension of ZrCl4 0.22 g,
Ph2C4H2 PNa THF by reaction with excess
n
Ž
sodium, followed by quenching with t-BuCl. To
a suspension of the anion 2.3 mmol in toluene
40 ml at 258C was added a suspension of
.
Ž
.
Ž
.
0.94 mmol in toluene 10 ml was added. The
stirred reaction mixture was allowed to warm to
room temperature, the solvent removed in vacuo
and the solids taken up in diethyl ether and
passed through a short silica column. The sol-
vent was removed and the residue washed with
hexane and dried under vacuum to yield a mix-
Ž
.
Ž
.
Ž
.
ZrCl4 0.26 g, 1.14 mmol in toluene 10 ml at
258C. After stirring for 2 h the mixture was
warmed to 508C and stirring was continued for
another 0.5 h. Subsequently the mixture was
centrifuged and the clear orange solution de-
canted and cooled to y208C, yielding orange
Ž
ture of rac and meso isomers of 2,5-Ph2-3-
Ž31
P-NMR, CD2Cl2, d 87.2
Ž
.
Ž31
.
MeC4HP ZrCl2
crystals 0.33 g of product P-NMR, CD2Cl2,
2
.
.
and 83.2 ppm .
d 83.1 ppm .

(
)
164
E.J.M. de Boer et al.rJournal of Molecular Catalysis A: Chemical 128 1998 155–165
(
)
5.6. Preparation of 2,5-t-Bu2-C4 H2 P 2 ZrCl2
ml of a toluene solution containing 10% MAO
( )
7
Ž
.
5.0 mmol , and then pressurized with 600 kPa
of propylene. The system was allowed to reach
equilibrium while the pressure was maintained
Ž
.
PhPH2 5.0 g, 45.5 mmol was added to a
Ž
.
Ž
Ž
solution consisting of benzene 25 ml , THF 4
ml and 4 ml of a 1.6 M solution of n-BuLi in
hexane 6.4 mmol . 2,2,7,7-Tetramethyl-3,5-oc-
tadiyne 7.4 g, 45.5 mmol was slowly added to
the reaction mixture at 08C. The resulting red
suspension was stirred at 258C for 24 h. Water
was added after which the organic layer was
separated and the volatiles removed in vacuo.
The resulting slightly impure oil was used for
at 600 kP a. S ubsequently,
2,5-
.
.
Ž
.
Ph2C4H2 P ZrCl2 10.0 mmol , dissolved in
2
Ž
Ž
.
Ž
.
toluene 10 ml , was added to the autoclave by
means of the catalyst injection system. The
injection system and catalyst supply vessel were
.
Ž
.
washed with toluene 10 ml , which was also
added to the autoclave. After 14 min the reac-
tion was stopped by rapid release of excess of
propylene, and opening of the autoclave. After
work-up, 38 g of product was obtained, shown
by 1H and 13C-NMR analysis to be atactic
polypropylene with Mn )20 000.
Ž31
further experiments P-NMR, C6D6, d y1.1
.
ppm .
To a solution of the 2,5-di-t-butyl-1-phenyl-
Ž
.
Ž
.
phosphole 4.5 ml in THF 50 ml was added
excess Li. After 2 h, 31P-NMR spectroscopy
showed the reaction to be complete and the
References
Ž
.
excess lithium was filtered off. t-BuCl 2.5 ml
was added and the solution was heated for 1 h
w x
1
H.H. Brintzinger, D. Fischer, R. Mulhaupt, B. Rieger, R.M.
¨
Ž
.
Waymouth, Angew. Chem. Int. Ed. Engl. 34 1995 1143.
Ž
.
at 608C. A yellow solid 2.6 g was obtained
after removal of the solvent in Õacuo and wash-
w x
2
Ž
.
P.C. Mohring, N.J. Coville, J. Organomet. Chem. 479 1994
¨
1.
w x
3
K. Mashima, S. Fujikawa, Y. Tanaka, H. Urata, T. Oshiki, E.
Ž
.
ing the resultant residue with hexane twice ,
Ž
.
Tanaka, A. Nakamura, Organometallics 14 1995 2633.
D. Devore, L.H. Crawford, J.C. Stevens, F.J. Timmers, R.D.
Mussels, D.R. Wilson, R.K. Rosen, World Patent Applica-
Ž
shown by NMR spectroscopy to be 2,5-t-Bu2-
C4H2 P Li THF contaminated with one equiv-
alent of LiCl P-NMR, THF-d8, d 62.5 ppm .
2,5-t-B u 2-
4.4 mmol in diethylether
20 ml at y408C, was added ZrCl4 0.46 g, 2.0
mmol and the suspension warmed to 258C and
stirred for 1 h. The yellow precipitate was sepa-
rated by filtration and then extracted with
CH2Cl2. After evaporation of the CH2Cl2,
w x
4
.
Ž
.
2
Ž31
.
Ž
.
tion No. 95r00526 Dow, 1995 .
w x
5
J.C. Flores, J.S. Wood, J.C.W. Chien, M.D. Rausch,
Ž
To
a
suspension of
Ž
.
Organometallics 15 1996 4944.
.
Ž
.
Ž
.
C4H2 P Li THF
w x
6
Ž
F.G. Cloke, European Patent Application No. 672676 BP,
2
.
1995 .
Ž
.
Ž
w x
7
J.D. Scollard, D.H. McConville, J. Am. Chem. Soc. 118
.
Ž
.
1996 10008, and references therein.
w x
8
A.D. Horton, J. de With, A.J. van der Linden, H. van de
Ž
.
Weg, Organometallics 15 1996 2672, and references therein.
w x
9
Ž
.
A.D. Horton, J. de With, Chem. Commun. 1996 1375, and
references therein.
Ž
.
Ž
.
2,5-t-Bu2-C4H2 P ZrCl2 0.12 g, 15% was
w
w
w
w
w
x
2
10 A.J. van der Linden, C.J. Schaverien, N. Meijboom, C.
Ž31
Ž
.
obtained as yellow microcrystals
CD2Cl2, d 100.7 ppm .
P-NMR,
Ganter, A.G. Orpen, J. Am. Chem. Soc. 117 1995 3008.
x
11 G.C. Bazan, S.J. Donnelly, G.J. Rodriguez, J. Am. Chem.
.
Ž
.
Soc. 117 1995 2671.
x
12 S. Nagy, R. Krishnamurti, B.P. Etherton, World Patent Ap-
5.7. Polymerization reactions
Ž
.
plication No. 96r34021 Occidental, 1996 .
x
13 G.C. Bazan, G. Rodriguez, A.J. Ashe III, S. Al-Ahmad, C.
Ž
.
Muller, J. Am. Chem. Soc. 118 1996 2291.
¨
A 1 l steel jacket-cooled autoclave, equiped
with magnetic stirring and a catalyst injection
device, was heated to 908C, evacuated for 2 h,
and subsequently pressurized and flushed three
times with dry nitrogen. The reactor was cooled
to 458C and charged with 200 ml of toluene, 3.5
x
14 R. Krishnamurti, S. Nagy, B.P. Etherton, World Patent Ap-
Ž
.
plication No. 96r23004 Occidental, 1996 .
w
w
x
Ž
.
15 F. Mathey, Chem. Rev. 88 1988 429.
x
16 N. Kuhn, S. Stubenrauch, R. Boese, D. Blaeser, J. Organomet.
Ž
.
Chem. 440 1992 289.
w
x
17 H. Schumann, E.C.E. Rosenthal, J. Winterfeld, R. Weimann,
Ž
.
J. Demtschuk, J. Organomet. Chem. 507 1996 287.

(
)
E.J.M. de Boer et al.rJournal of Molecular Catalysis A: Chemical 128 1998 155–165
165
w
x
w
w
x
Ž
.
18 F. Mathey, J. Fischer, J.H. Nelson, Structure Bonding 55
38 C. Janiak, P. Marquardt, GIT Fachz. Lab. 7 1996 716.
Ž
.
x
1983 153.
39 R. Mulhaupt, T. Duschek, D. Fischer, S. Setz, Pol. Adv.
¨
w
w
x
x
Ž
.
Ž
.
19 F. Mathey, J. Organomet. Chem. 475 1994 25.
20 C. Janiak, K.C.H. Lange, U. Versteeg, D. Lentz, P.H.M.
Technol. 4 1993 439.
w
x
40 E.J.M. de Boer, A.A. van der Huizen, F. Hage, C.L. Willis,
Ž
.
Ž
.
Budzelaar, Chem. Ber. 129 1996 1517.
European Patent Application No. 549064 Shell, 1994 .
w
w
w
w
w
w
w
w
w
w
x
w
x
21 C. Janiak, U. Versteeg, K.C.H. Lange, R. Weimann, E.
41 S. Seelert, D. Fischer, B.L. Marczinke, F. Langhauser, J.
Ž
.
Hahn, J. Organomet. Chem. 501 1995 219.
Kerth, H.-J. Mueller, German Patent Application No.
x
Ž
.
22 R.M.G. Roberts, J. Silver, A.S. Wells, Inorg. Chim. Acta.
4425787 BASF, 1996 .
Ž
.
w
w
w
w
w
w
w
w
w
w
w
w
w
x
119 1986 165.
42 L. Resconi, F. Piemontesi, G. Franciscono, L. Abis, T.
x
Ž
.
23 P. Gradoz, M. Ephritikhine, M. Lance, J. Vigner, M. Nier-
Fiorani, J. Am. Chem. Soc. 114 1992 1025.
Ž
.
x
lich, J. Organnomet. Chem. 481 1994 69.
43 T. Mise, A. Kageyama, S. Miya, H. Yamazaki, Chem. Lett.
x
Ž
.
24 F. Nief, F. Mathey, L. Ricard, F. Robert, Organometallics 7
1991 1525.
Ž
.
x
1988 921.
44 S. Miya, M. Harada, T. Mise, H. Yamazaki, European Patent
x
Ž
.
25 F. Nief, F. Mathey, L. Ricard, J. Organomet. Chem. 384
Application No. 283739 Chisso Corporation, 1988 .
Ž
.
x
1990 271.
45 W. Egan, R. Tang, G. Zon, K. Mislow, J. Am. Chem. Soc.
x
Ž
.
.
.
Ž
.
26 F. Nief, L. Ricard, F. Mathey, Organometallics 8 1989
93 1971 6205.
x
Ž
.
46 G. Markl, R. Potthast, Angew Chem. Int. Ed. Engl. 6 1967
¨
1473.
x
Ž
27 F. Nief, F. Mathey, J. Chem. Soc. Chem. Commun. 1988
86.
x
770.
47 D. Carmichael, L. Ricard, F. Mathey, J. Chem. Soc. Chem.
x
Ž
Ž
.
28 F. Meunier, B. Gautheron, J. Organomet. Chem. 193 1980
Commun. 1994 1167.
x
C13.
48 F. Mathey, F. Mercier, F. Charrier, C. Fisher, J. Mitscher, J.
x
Ž
.
29 E.J.M. de Boer, H.J.R. de Boer, H.J. Heeres, World Patent
Am. Chem. Soc. 103 1983 4595.
Ž
.
x
Application No. 95r04087 Shell, 1995 .
49 G. Conti, G. Arribas, A. Altomare, B. Mendez, F. Ciardelli,
x
Ž
.
30 M. Sone, A. Yano, European Patent Application No. 574794
Z. Naturforsch. 50b 1995 411.
Ž
.
x
Tosoh, 1993 .
50 J. March, Advanced Organic Chemistry, 4th ed., Wiley, New
w
w
x
x
31 European Chemical News, 20–26 January, 1997, p. 20.
32 K. Yokota, European Patent Application No. 590486
York, NY 1992, p. 151
x
51 T.K. Han, S.C. Yoon, Y.S. Ko, B.W. Wo, J.T. Park, S.I.
Ž
.
Ž
.
Idemitsu Kosan, 1994 .
Woo, Pol. Bull. 37 1996 35.
w
x
x
33 J. van Beek, G.H.J. van Doremaele, K.A. Vandewiele, Y.
52 I.-M. Lee, W.J. Gauthier, J.M. Ball, B. Iyengar, S. Collins,
Ž
.
van den Winkel, Belgian Patent Application No. 9300710
Organometallics 11 1992 11.
Ž
.
x
DSM, 1995 .
53 N. Piccolrovazzi, P. Pino, G. Consiglio, A. Sironi, M. Moret,
w
w
w
w
x
Ž
.
34 T. Aoki, O. Kurashiki-shi, European Patent Application No.
Organometallics 9 1990 3098.
Ž
.
x
617052 Asahi Kasei, 1994 .
54 E.J.M. de Boer and K. Steernberg, European Patent Applica-
x
Ž
.
35 T. Sugano, K. Yamamoto, European Patent Application No.
tion No. 540108 Shell, 1993 .
Ž
.
w
w
x
x
Ž
.
728773 Mitsubishi Chemical, 1996 .
55 K. Hess, W. Wustrow, Ann. N.Y. Acad. Sci. 437 1924 256.
56 K. Alder, M. Schumacher, Ann. N.Y. Acad. Sci. 571 1951
x
Ž
.
36 L. Resconi, R.L. Jones, A.L. Rheingold, G.P.A. Yap,
Ž
.
Organometallics 15 1996 998.
122.
x
w
x
Ž
.
37 K.P. Hungenberg, J. Kerth, H.-J. Muller, P. Muller, Angew.
57 F. Langer, F. Wessely, Monatsch. Chem. 86 1955 887.
¨
¨
Ž
.
Makromol. Chem. 227 1995 159.