High surface area polystyrene resin-supported Pt catalysts in room temperature solventless octene hydrosilylation using methyldichlorosilane

Article abstract of DOI:10.1039/b203899d

Eight macroporous styrene-divinylbenzene-vinylbenzyl chloride resins have been synthesised by suspension polymerisation. The first four employed toluene as the porogen and the second four n-butyl acetate, at a level of 1:1 v/v relative to the comonomers. In all cases a high level of divinylbenzene leads to resins with high surface area, ～ 500 m² g^-1, as determined from a BET treatment of N₂ sorption data. The functional group content of each group of four resins was varied from 5-25%. All resins were aminated to generate benzyltrimethylethylenediamine ligands on the polymer matrix, and then each was loaded with Pt(II) using KPtCl₄. The analytical data confirmed the formation of ligand PtCl₂ molecular complexes. Each of the resin immobilised Pt complexes has been assessed for catalytic activity in the room temperature, solventless, hydrosilylation of oct-1-ene using methyldichlorosilane, and a comparison made with soluble Speier's catalyst under identical conditions. Though less active than the soluble catalyst the activity of all the polymer catalysts is good, and of practical value, the activity being higher than we have previously reported in the case of supports with lower surface area. Furthermore while Speier's catalyst induces significant levels of oct-1-ene isomerisation, isomerisation in the case of the polymer catalysts is much lower, and indeed can be all but eliminated by appropriate washing. Extensive catalyst leaching and recycling studies have been carried out, with the best catalysts showing retention of useful activity after 10 cycles. Careful control experiments have provided strong circumstantial evidence that the isomerisation that does arise with the polymer catalysts can be attributed to a component of leached soluble Pt species. Overall the most active and stable polymer catalyst has the highest surface area (～550 m² g^-1) of those studied, along with the lowest ligand and Pt contents (each ～0.25 mmol g^-1). The surface area dependence confirms our earlier view that maximum accessibility to potential metal complex catalytic sites is vital in these systems, and the metal complex loading dependence suggests that generating discrete isolated ligand PtCl₂ species provides optimal use of the loaded Pt.

Full text of DOI:10.1039/b203899d

View Article Online / Journal Homepage / Table of Contents for this issue
High surface area polystyrene resin-supported Pt catalysts in room
temperature solventless octene hydrosilylation using
methyldichlorosilane
1
Robert Drake,a David C. Sherrington*b and Steven J. Thomsonb
a Dow Corning, Barry, S. Glamorgan, UK CF63 27L
b Department of Pure and Applied Chemistry, University of Strathclyde, Glasgow, UK G11 1XL
Received (in Cambridge, UK) 23rd April 2002, Accepted 7th May 2002
First published as an Advance Article on the web 31st May 2002
Eight macroporous styrene–divinylbenzene–vinylbenzyl chloride resins have been synthesised by suspension
polymerisation. The first four employed toluene as the porogen and the second four n-butyl acetate, at a level of
1 : 1 v/v relative to the comonomers. In all cases a high level of divinylbenzene leads to resins with high surface area,
∼500 m2 gϪ1, as determined from a BET treatment of N2 sorption data. The functional group content of each group
of four resins was varied from 5–25%. All resins were aminated to generate benzyltrimethylethylenediamine ligands
on the polymer matrix, and then each was loaded with Pt() using KPtCl4. The analytical data confirmed the
formation of ligand PtCl2 molecular complexes. Each of the resin immobilised Pt complexes has been assessed for
catalytic activity in the room temperature, solventless, hydrosilylation of oct-1-ene using methyldichlorosilane, and
a comparison made with soluble Speier’s catalyst under identical conditions. Though less active than the soluble
catalyst the activity of all the polymer catalysts is good, and of practical value, the activity being higher than we
have previously reported in the case of supports with lower surface area. Furthermore while Speier’s catalyst
induces significant levels of oct-1-ene isomerisation, isomerisation in the case of the polymer catalysts is much
lower, and indeed can be all but eliminated by appropriate washing. Extensive catalyst leaching and recycling
studies have been carried out, with the best catalysts showing retention of useful activity after 10 cycles. Careful
control experiments have provided strong circumstantial evidence that the isomerisation that does arise with the
polymer catalysts can be attributed to a component of leached soluble Pt species. Overall the most active and stable
polymer catalyst has the highest surface area (∼550 m2 gϪ1) of those studied, along with the lowest ligand and Pt
contents (each ∼0.25 mmol gϪ1). The surface area dependence confirms our earlier view that maximum accessibility
to potential metal complex catalytic sites is vital in these systems, and the metal complex loading dependence suggests
that generating discrete isolated ligand PtCl2 species provides optimal use of the loaded Pt.
chlorosilanes, and equally importantly meticulous evaluation
of the contribution from leached soluble catalyst has generally
Introduction
The hydrosilylation of alkenes is an important industrial
process for the production of organosilicon compounds, and is
also extremely valuable in laboratory scale synthesis. Many
metal complexes are known to be catalysts for the reaction,1 but
the discovery by Speier in 1957 that hexachloroplatinic acid is
a very active catalyst, even under ambient conditions,2 has
led to Pt complexes becoming the catalysts of choice for these
reactions. Pt-based species are not only useful when alkyl and
alkyloxysilanes are employed, but they also have the addi-
tional advantage of not being deactivated by chlorosilanes. This
versatility, often combined with remarkable turnover frequen-
cies, explains the dominance of these catalysts in industrial
processes. Paradoxically, however, these high activities are also
associated with potentially high exothermicity in the hydrosilyl-
ation reaction which can lead to a rapid temperature rise and
the occurrence of a significant level of alkene isomerisation.
The internal alkenes that are formed react much more slowly,
and so in effect can limit the conversion of alk-l-ene to useful
terminally silylated products.
been absent.
Recently we made a preliminary disclosure of the remarkable
activity, selectivity and stability of some poly(styrene) and
poly(methacrylate) resin-supported Pt catalysts synthesised in-
house and used under solvent-less conditions at ambient tem-
perature.13 More detailed studies revealed that gel-type resins
were not well suited as supports because of their incompati-
bility with the components of the reaction, undergoing only
poor swelling.14 Likewise the poly(methacrylate)-based resins
produced catalysts which performed more poorly than the
comparable polystyrene-based species probably because of
the higher polarity of the former. A series of macroporous
polystyrene-based resin Pt catalysts with dry state surface areas
ranging from 25–200 m2 gϪ1 demonstrated a good correlation
between catalyst activity and surface area;14 while further
studies in which the immobilised ligand–Pt mole ratio was
adjusted from ∼45 : 1 to ∼2 : 1 demonstrated that high ligand
ratios (designed to minimise Pt leaching) offered no advan-
tage.15 Typically macroporous resins loaded with ∼0.4 mmol Pt
Early studies on the heterogenisation of Pt-based catalysts on
both inorganic and polymer supports have been reviewed3,4 and
exploitation of this strategy has continued more recently as
well.5–12 While these various reports have demonstrated active
heterogeneous catalysts, data pertaining to the activity and
selectivity of these catalysts with prolonged use are rare,
particularly with the more experimentally difficult to handle
g
Ϫ1 with a corresponding ligand–Pt ratio of 2 : 1 and a dry state
surface area of ∼200 m2 gϪ1 were shown to be highly active and
stable catalysts capable of being re-cycled up to ∼10 times in
batch reactions.
Since it is well known that polystyrene-based resins with sur-
face areas well in excess of 200 m2 gϪ1 are readily accessible,16–18
we felt that further optimisation of our polymer-supported Pt
DOI: 10.1039/b203899d
J. Chem. Soc., Perkin Trans. 1, 2002, 1523–1534
This journal is © The Royal Society of Chemistry 2002
1523

View Article Online
Table 1 Surface area of VBC-based resins (PR1–7) suspension polymerised using various porogens
Precursor resina
Porogen
Solubility parameter of porogen (MPa)1/2
BET surface areab/m2 gϪ1
PR1
PR2
PR3
PR4
PR5
PR6
PR7
Toluene
18.2
—
20.5
17.4
—
19.4
18.0
532
374
501
541
332
343
545
1,2,4-Trichlorobenzene
1,2-Dichlorobenzene
Butyl acetate
2-Ethylhexanoic acid
2-Ethylhexan-1-ol
m-Xylene
a Each resin was produced using a 60 : 25 : 15 vol% mixture of divinylbenzene–vinybenzyl chloride–ethylstyrene monomers; the volume ratio of
monomer–porogen was 1 : 1. b 5-point BET surface area analysis.
catalysts might be possible, and we now report on our extended
studies of the activity, selectivity and stability of these systems.
these porogens yielded high quality resins with a surface area
>500 m2 gϪ1. Toluene and m-xylene were judged to be chem-
ically very similar porogens.
Results and discussion
Catalyst precursor VBC-resins
Two series of resins were synthesised using toluene and n-butyl
acetate respectively as porogens employing a comonomer–
porogen volume ratio of 1 : 1. Within each series the feed con-
tent of VBC was varied from 5 to 25 vol% in order to allow a
range of ligand, and hence catalyst, contents to be probed.
Table 2 summarises the data for these precursor resins. In all
cases there is good correlation between the VBC level used in
the polymerisation mixture and the content of –CH2Cl func-
tionality in each resin determined from Cl elemental micro-
analysis (Table 3). The dry state surface area is >500 m2 gϪ1 for
each resin with the biggest area arising with lowest level of
–CH2Cl functionality in both series of resins. The pores present
fall in the micropore to mesopore range (∼1–50 nm) with few
macropores. There is indeed a suggestion that the surface area
varies in an inverse linear fashion with the –CH2Cl content, but
more data over a broader range of functional group content is
really needed to confirm this. This correlation however would
imply that both of these porogens become more solvating
towards the polymer matrix as the VBC content is reduced. The
presence of the –CH2Cl functionality in each resin is confirmed
by the appearance of a band at 1268 cmϪ1 in the FTIR spectra,
with the intensity of this varying as expected with the level of
VBC used in the preparation of the resins.
Synthesis of high surface area polystyrene-based resins using
suspension polymerisation
As a general design criterion it is well documented that high
surface area resins can be produced by exploiting a combin-
ation of a high level of crosslinker and a porogen with good
thermodynamic compatibility with the polymer matrix being
synthesised.16–18 This allows a high conversion to crosslinked
polymer to be achieved before any phase separation from the
porogen occurs. In turn this results in the formation of a mass
of microgel particles which become relatively lightly fused, and
between which little in-filling of polymer occurs. Typically an
even volume balance of comonomers and porogen is required
to allow generation of significant pore volume in the product,
without producing a structure which is macroscopically fragile.
While compositional data for St-DVB resins with these charac-
teristics are fairly well established, this is not the case for
functional analogues of these prepared with an additional
functional comonomer component. The latter can change
significantly the solvation characteristics of the resultant poly-
mer matrix, and hence the conditions for achieving ‘late’ phase
separation during the polymerisation. Accordingly therefore a
number of control ethylstyrene–divinylbenzene–vinylbenzyl
chloride (EtSt–DVB–VBC) resins were synthesised by a stand-
ard suspension polymerisation procedure using a fixed comon-
omer composition of EtSt–DVB–VBC of 25 : 60 : 15 vol% and
a fixed comonomer–porogen volume ratio of 1 : 1 with various
porogens. The results obtained are summarised with the
data in Table 1. Typically excellent yields of beaded product
were obtained (>90%), with a broad particle size distribution
∼100–700 µm. Despite a common stabilising system, reactor,
stirrer etc. and polymerisation conditions being used through-
out, the actual particle size distribution varied with the nature
of the porogen. A larger fraction of smaller beads (<425 µm)
was obtained when toluene or xylene was used as the porogen,
whereas a large fraction of bigger beads (>425 µm) resulted
when n-butyl acetate 1,2-dichlorobenzene, 1,2,4-trichloro-
benzene, 2-ethylhexan-l-ol or 2-ethylhexanoic acid was the
porogen. The precise distribution varied with each porogen.
This illustrates just how sensitive suspension polymerisations
are to all the components used in the polymerisation mixture.
For consistency a particle size 212–300 µm was selected for use
throughout this work; this provided a good working quantity of
quality beads from all suspension polymerisations. The dry
state surface area of each resin was determined from a BET
treatment of 5-point N2 sorption data and the results are shown
in Table 1. Although the solubility parameters of the porogens
examined fall within a relatively narrow range the variation in
surface area achieved is quite high. Toluene and n-butyl acetate
were selected as two chemically contrasting porogens to carry
through in the synthesis of a broader range of resins. Both of
Aminated resins
Benzyltrimethylethylenediamine ligands were introduced onto
each resin by reaction of the –CH2Cl groups with the sodium
salt of trimethylethylenediamine (Scheme 1). The reactions
Scheme 1 Amination of chloromethylated precursor styrene-based
resins and loading of Pt species.
employed a –CH2Cl–diamine–NaH mole ratio of 1 : 3 : 3.3
based on the experimentally determined level of –CH2Cl
in each resin. Reactions were run for 48 h at 60 ЊC in THF.
The ligand content achieved was calculated from the experi-
mentally determined elemental N content (Table 3), and the
data are summarised in Table 2 along with an indication of
1524
J. Chem. Soc., Perkin Trans. 1, 2002, 1523–1534

View Article Online
Table 2 Composition and properties of VBC-based resins (PR1, 8–14) and corresponding aminated resins (AFR1–8)
–CH2Cl Contentb /
Comonomer mixture (vol%)a Porogen
mmol gϪ1
Diamine
contentd /
mmol gϪ1
–CH2Cl
conversione
(%)
VBC
resin
(porogen–monomer
volume ratio)
BET surface
areac /m2 gϪ1
Aminated
resin
DVB
ESt
St
VBC
Theory
Foundb
PR8
60
60
60
60
60
60
60
60
15
15
15
15
15
15
15
15
20
15
10
0
20
15
10
0
5
10
15
25
5
10
15
25
Toluene (1 : 1)
Toluene (1 : 1)
Toluene (1 : 1)
Toluene (1 : 1)
Butyl acetate (1 : 1)
Butyl acetate (1 : 1)
Butyl acetate (1 : 1)
Butyl acetate (1 : 1)
0.39
0.76
1.13
1.83
0.39
0.76
1.13
1.83
0.31
0.73
1.21
1.78
0.45
0.87
1.21
1.78
596 (55)
520 (55)
519 (55)
532 (5)
564 (5)
539 (5)
529 (5)
533 (5)
AFR1
AFR2
AFR3
AFR4
AFR5
AFR6
AFR7
AFR8
0.25
0.36
0.61
1.07
0.18
0.32
0.54
0.89
81
50
45
62
40
37
48
49
PR9
PR10
PR1
PR11
PR12
PR13
PR14
a DVB = divinylbenzene, ESt = ethylstyrene, St = styrene, VBC = vinylbenzyl chloride. b Calculated from Cl% found. c Figure in brackets indicates
whether 5-point or 55-point analysis was obtained. d Calculated from N% found. e Calculated from diamine content of resins.
Table 3 Elemental microanalytical data for VBC resins (PR1, 8–14) and corresponding aminated resins (AFR1–8)
VBC Resin
%C
%H
%Cl
Aminated resin
%C
%H
%N
%Cl
PR8
90.0
89.0
87.6
85.7
85.5
89.0
87.6
85.4
8.0
7.7
7.6
7.2
7.5
7.7
7.5
7.3
1.1
2.6
4.3
6.3
1.6
3.1
4.3
6.3
AFR1
AFR2
AFR3
AFR4
AFR5
AFR6
AFR7
AFR8
89.8
88.3
86.6
84.1
89.3
88.6
85.9
83.8
7.8
7.7
6.7
8.1
7.7
8.0
7.7
8.0
0.7
1.0
1.7
3.0
0.5
0.9
1.5
2.3
0.5
1.5
1.9
1.7
1.6
1.6
2.3
2.5
PR9
PR10
PR1
PR11
PR12
PR13
PR14
Table 4 Elemental microanalytical data and ligand–Pt content of catalyst resins (C1–11)
Microanalytical data (%)
Source resin
Catalyst resin
C
H
N
Cl
Diamine contenta /mmol gϪ1
Pt contentb /mmol gϪ1
Pt loadingc (%)
AFR1
AFR1
AFR2
AFR3
AFR4
AFR4
AFR5
AFR6
AFR7
AFR8
PR10
C1d
C2d, e
C3
83.9
83.6
77.8
75.6
66.7
66.0
85.1
80.9
75.7
69.9
83.4
7.3
7.2
6.9
6.8
6.6
6.5
7.5
7.3
7.2
6.6
7.2
0.7
0.6
0.9
1.4
2.2
2.1
0.5
0.8
1.3
2.1
—
2.3
2.0
3.7
5.3
7.2
7.3
1.6
3.4
4.8
6.4
4.8
0.25
0.21
0.32
0.50
0.79
0.75
0.18
0.29
0.46
0.75
0.00
0.25
0.27
0.30
0.46
0.73
0.76
0.15
0.27
0.45
0.64
0.18
92
97
94
90
87
91
94
92
94
97
–
C4
C5
C6e
C7
C8
C9
C10
C11
a Calculated from N% found. b Determined from AAS data for digested resins. c Based on original ligand content of aminated source resin. d Loaded
using ∼10 mol% excess of K2PtCl4 relative to ligand, remainder at equimolar with ligand. e Loaded using 5 equal portions of K2PtCl4 over extended
time period (see Experimental section).
the % conversion of –CH2Cl groups. The latter is typically
∼40–60% which is comparable with other data in the literature
for functional group chemical modification in highly cross-
linked macroporous resins. Overall the conversions achieved for
the resins prepared with toluene as the porogen are somewhat
higher than for the resins derived from n-butyl acetate as the
porogen. The FTIR spectra of all resins show a significant fall
in the intensity of the band at 1268 cmϪ1 assigned to the
–CH2Cl group, and the appearance of shoulders at 2774 and
2813 cmϪ1 on the main C–H bands. These were assigned to the
tertiary amine groups in the diamine functionality.19 In all cases
the aminated resins displayed good dry state surface area
(see later).
experimentally determined N% content and the Pt content
from atomic absorption spectrophotometric analysis of solu-
tions of digested catalyst resins. These data are also summar-
ised in Table 4. In each case there is a very good correspond-
ence, close to 1 : 1, between the ligand and the Pt content and
this provides good evidence for the formation of well defined
diamine PtCl2 complexes on all the resins. The correspondence
of N% with Cl% is however not as good, with the Cl% data
being consistently on the high side. The major reason for this is
almost certainly the residual Cl content in the aminated resins,
arising from incomplete conversion of the –CH2Cl groups.
When account is taken of this the correspondence of N% with
Cl% is much better. Furthermore the variation of Cl% with
ligand and Pt content is consistent. Overall therefore for both
series of catalysts there is good experimental evidence that the
construction of the required Pt complexes on the resins has
been successful in all cases.
Pt loaded resins
Each diamine functionalised resin was loaded with Pt by treat-
ment with an aqueous solution of K2PtCl4 employing a ligand–
Pt mole ratio of 1 : 1. Minor variations on the method are
indicated in Table 4 and in the Experimental section. The ligand
content of the derived resin catalysts was calculated from the
Surface area analysis
Despite the now considerable literature on polymer-supported
J. Chem. Soc., Perkin Trans. 1, 2002, 1523–1534
1525

View Article Online
Table 5 Changes in N2 sorption BET surface area on amination of VBC resins and loading with PtCl2
VBC resins
Aminated resins
Surface area (m2 gϪ1
Catalyst resins
Surface area (m2 gϪ1
VBC
content
(vol %)
Surface
area
)
)
Code
(m2 gϪ1
)
Code
Expecteda
Found
Differenceb
Code
Expectedc
Found
Differenceb
PR8
5
5
10
15
25
25
5
10
15
25
15
596 (55)
596 (55)
520 (55)
519 (55)
532
532
564
539
529
533
519 (55)
AFR1
AFR1
AFR2
AFR3
AFR4
AFR4
AFR5
AFR6
AFR7
AFR8
—
586
586
508
499
497
497
557
518
510
504
—
577
577
490
463
425
425
545
510
456
448
—
Ϫ9
Ϫ9
C1
C2
C3
C4
C5
C6
C7
C8
C9
C10
C11
538
536
449
405
351
348
523
473
401
370
494
542
527
355
396
368
325
515
440
387
280
501
ϩ4
Ϫ9
Ϫ94
Ϫ9
ϩ17
Ϫ23
Ϫ8
Ϫ33
Ϫ14
Ϫ90
ϩ7
PR8
PR9
Ϫ18
Ϫ36
Ϫ72
Ϫ72
Ϫ12
Ϫ8
Ϫ54
Ϫ56
—
PR10
PR1
PR1
PR11
PR12
PR13
PR14
PR10
a Calculated from experimental data for VBC resin accounting for mass increase on amination. b Difference between expected and measured (found)
surface area. c Calculated from experimental data for aminated resin accounting for mass increase on loading with PtCl2 Note: except where
otherwise indicated (in brackets) 5-point BET surface area analyses were obtained.
metal complex catalysts there remains a dearth of data on the
changes in surface area of resins, if any, that occurs during a
sequential chemical modification process. This factor was
deemed to be potentially important in the present work, and so
the change in surface area of each resin at each step of the
catalyst synthesis was monitored carefully. The data for the two
series of resins are shown in Table 5. In considering these it is
important to realise that chemical modification of a given mass
of a resin may change the mass (increase or decrease depending
on the chemical change). Computation of surface area from
BET treatment of N2 sorption data yields the surface area per
gramme of the resin under analysis, and in evaluating whether
the surface area has really changed during the chemical modifi-
cation process it is vital to account for any mass change arising
with the process. In the present case conversion of –CH2Cl to
ligand and thence to Pt complex causes a progressive mass
increase in any given sample of resin and so if no surface area
changes occur, the surface areas quoted per gramme of –CH2Cl
resin, aminated resin and Pt complexed resin should decrease
progressively. This is reflected in the data in the surface area
‘expected’ columns in Table 5. The actual experimentally
determined data corresponding to these ‘expected’ values are
listed in the ‘found’ columns. Clearly in the case of the conver-
sion of –CH2Cl resins to aminated resins there is a real and
consistent loss of surface area for all the resins examined as
indicated in the “difference” column. For the lightly functional-
decrease in surface area is the shear bulk of the Pt complex
formed. If generated in an appropriate micropore then this
could readily cause blockage and effective loss of surface area.
Explanations for an increase in surface area, such as Pt com-
plexation inducing local topographic rearrangement of the
polymer network, can be advanced, but are more difficult to
sustain. Overall, however, though there are clearly real changes
in surface area generally involving a loss, these are never more
than ∼25%, and indeed are usually very much less than this.
Hence it is possible to say with some confidence that a pre-
cursor resin designed and prepared with a high surface area (say
≥500 m2 gϪ1) will have this surface area substantially main-
tained when relatively bulky metal complex catalyst structures
are constructed on the polymer matrix, certainly in the case of
metal complex loading of ≤25% of polymer segments.
Octene hydrosilylation chemistry
The hydrosilylation of alkenes such as octene is complicated by
the possibility of concurrent alkene isomerisation,20 and hence
hydrosilylation of the in-situ formed isomers, and also by the
possibility of forming different regio-isomers of the alkylsilane
product. In the case of Pt-based catalysts hydrosilylation of
terminal alkenes is generally much faster than hydrosilylation
of any internal isomers,21,22 and also the addition reaction
is usually very regioselective in terms of yielding the linear
alkylsilane product (Scheme 2).
ised resins e.g. PR 8
AFR 1 and PR11
AFR 5 the loss of
surface area is very small, indeed probably within the experi-
mental error of the measurements. For the more heavily func-
tionalised resins the loss of surface area is larger and certainly
significant. A possible explanation for this is that minor levels
of additional crosslinking might occur in the amination reac-
tion. A bound ligand –N(CH3)CH2CH2 N(CH3)3 could under-
go a quaternisation reaction with a nearby –CH2Cl group. This
complication would be expected to occur with higher prob-
ability as the –CH2Cl content rises. The analytical data available
is not accurate enough to detect such reactions, which almost
certainly are not extensive. However, relatively few additional
crosslinks might easily give rise to blockage or loss of some
(micro) pore structure. The situation with the Pt loading step is
less clear. For the series of resins prepared with n-butyl acetate
as the porogen there seems to be a consistent further drop in
surface area (C7–C10) again with evidence that the fall is larger
for the more heavily loaded resins. However for the series of
resins prepared using toluene as the porogen (C1–C6) the
change in surface area is very variable, and indeed in two resins
(C1 and C5) there appears to be an increase rather than a
decrease. The only realistic explanation for Pt loading causing a
Scheme 2 Pt catalysed hydrosilylation and isomerisation of oct-1-ene.
Alkylsilane product
The reaction chosen for study in this work was that of methyl-
dichlorosilane with oct-1-ene and it was thought important to
confirm the regioselective nature of this particular Pt catalysed
reaction as performed by ourselves with our heterogeneous
catalysts. Accordingly a number of reaction mixtures from
catalyst recycling experiments (see later) were retained and
1526
J. Chem. Soc., Perkin Trans. 1, 2002, 1523–1534

View Article Online
Table 6 Relative GC elution times of octene isomers formed during
catalyst and
a polymer-supported Pt catalyst was also
oct-1-ene hydrosilylations
investigated. As indicated above the hydrosilylation reaction is
slower than with pure oct-1-ene but the product is the same,
l-octylmethyldichlorosilane. The ratio of trans-oct-2-ene to
cis-oct-2-ene rises in the reaction mixture indicating that the cis
isomer undergoes isomerisation more rapidly than the trans,
presumably because it is able to coordinate more easily to the Pt
centre. Interestingly the content of oct-1-ene in the mixture
also rises from its starting value (∼1.0%) confirming this isomer
to be formed in the isomerisation process, subsequently to
undergo hydrosilylation to 1-octylmethyldichlorosilane.
Elution timea /min
Octene isomer assignment
Ϫ0.04
0.0
ϩ0.01–0.02
ϩ0.03
ϩ0.06
ϩ0.08
ϩ0.11
Unknown A
Oct-1-ene
trans-Oct-4-eneb
trans-Oct-3-ene
trans-Oct-2-ene
Unknown B
cis-Oct-2-ene
a Relative to that of oct-1-ene. b Depending on dilution this isomer
sometimes merged with oct-1-ene.
Evaluation of catalyst performance in hydrosilylation of
oct-1-ene
stored under N2 before analysis by GC and FTIR and NMR
spectroscopy. A commercially available sample of 1-octyl-
methyldichlorosilane was analysed in parallel. In all cases GC
analysis (see Experimental section) showed the presence of
only one product with a retention time identical to that of
l-octylmethyldichlorosilane. The reaction mixture from the
eighth use of one catalyst was decanted from the polymer-
supported Pt complex, and unreacted oct-1-ene any octene
isomers and the original nonane GC standard were removed
using a Kügel-Röhr distillation apparatus. The residue was
analysed by FITR and NMR spectroscopy yielding the follow-
ing data. FTIR(cmϪ1) 550, Si–Cl; 1262, CH3–Si; 2800–3000,
C–H (Note absence of 2200 cmϪ1 band due to Si–H). 1H NMR
(CDCl3) δ: 1.51 (m, 2H), 1.35 (m, 10H), 1.13 (m, 2H), 0.90 (m,
3H), 0.78 (t, 3H). 13C NMR (CDCl3) δ: 32.60, 32.08, 29.33,
29.32, 22.86, 22.65, 21.88, 14.30, 5.42. In each case the spectral
data correlates excellently with the data for a genuine sample of
l-octylmethyldichlorosilane.
For completeness the reaction of oct-2-ene with methyl-
dichlorosilane catalysed by Speier’s catalyst and also by one of
the polymer-supported Pt catalysts was examined. The octyl-
silane product was isolated as described above and shown to be
l-octylmethyldichlorosilane. There was no evidence for any
silane product from direct hydrosilylation of the oct-2-ene.
Hence the slow hydrosilylation that does occur seems to take
place via the (unfavourable) equilibrium isomerisation of oct-2-
ene to oct-1-ene, and this confirms the observations in the
literature.21,22
As in our preliminary studies all reactions were carried out at
room temperature in the absence of solvent, although the level
of nonane present as an internal GC standard is significant.
The standard conditions employed a mole ratio of oct-1-ene–
silane–Pt of 2 : 1 : 1 × 10Ϫ3. The oct-1-ene–silane ratio of 2 : 1
was chosen in some respects to favour isomerisation and to
allow any isomerisation reaction to be monitored relatively
easily even if 100% conversion of the methyldichlorosilane to
the octylsilane were to occur.
Speier’s catalyst
With Speier’s catalyst and using the above conditions a short
induction period was observed (∼2 min) after which the reac-
tion accelerated rapidly with a sharp rise in temperature (to
∼90 ЊC). The conversion to 1-octylmethyldichlorosilane reached
∼90% (based on the silane feed) in 15 min and ∼30% of the
original oct-1-ene was found to be isomerised at this time. This
confirmed our earlier findings employing this homogeneous
catalyst.13,14 Since ∼1 mole equivalent of octenes was present at
the end of this reaction it was decided to repeat the reaction
under identical conditions and, when all the silane was con-
sumed, to add a second mole equivalent of silane. As before
the (first) conversion achieved in only 15 min was >90% with
simultaneously ∼25% of the original oct-1-ene converted to
isomers. On addition of the second mole equivalent of silane
the reaction accelerated again with ∼55% conversion of methyl-
dichlorosilane to the octylsilane in 15 min. This corresponds to
a little more than the oct-1-ene content (∼50%) of the octenes
remaining at the end of the first reaction (note 25% of original
oct-1-ene converted to isomers in the first reaction corresponds
to ∼50% oct-1-ene and 50% other isomers in the residue of
octenes). At the end of the second reaction the proportion
of internal octene isomers remaining was ∼18% based on the
original oct-1-ene feed, indicating that in the second reaction
(and presumably in the first as well) some of the internal iso-
mers convert back to oct-1-ene. However, the composition of
the residual octene mixture at the end of the second reaction is
still ∼80% internal alkenes indicating that it is overwhelmingly
the existing residual oct-1-ene at the end of the first reaction
that undergoes hydrosilylation in the second reaction. Overall
therefore the behaviour of Speier’s catalyst in our hands is
much as expected from the literature and the present catalyst
activity and selectivity is consistent with our own earlier
data.13,14
Alkene isomerisation
Since alkene isomerisation can be a very important side reac-
tion in hydrosilylation some effort was made to detect this in
our reactions and to identify the isomeric products. There are a
total of seven possible isomers of octene: oct-1-ene and the six
cis–trans stereoisomers of oct-2-ene, oct-3-ene and oct-4-ene.
Only five of the seven were found to be available from conveni-
ent commercial sources i.e. oct-1-ene, cis-oct-2-ene, trans-oct-2-
ene, trans-oct-3-ene and trans-oct-4-ene. GC analysis (see
Experimental section) of oct-1-ene reaction mixtures using
Speier’s catalyst and polymer-supported catalysts, where sig-
nificant alkene isomerisation occurred, yielded seven peaks in
the alkene region with the elution times relative to that of oct-1-
ene shown in Table 6. Comparison with the elution behaviour
of the available standards allowed the assignment indicated
in Table 6, and we assume that unknowns A and B are cis-oct-
4-ene and cis-oct-3-ene, although which is which is not
clear. No attempt was made to systematically quantify each
of the isomers but it is clear that where isomerisation does
arise equilibration occurs along the whole carbon backbone.
In all subsequent catalyses the summation of all alkenes
other than oct-1-ene was taken as a measure of the level of
isomerisation.
Polymer-supported catalysts
Our previous work13,14 with styrene and methacrylate-based
resin supported Pt catalysts showed that gel-type resins are
not suitable supports for these solventless alkene hydrosilyl-
ations since they do not swell in the reaction media. Styrenic
macroporous resin species also proved consistently better
than methacrylate derived species probably because the
latter are more polar and less compatible with the very hydro-
phobic components of the reaction mixtures. Of the styrenic
The isomerisation of a mixture of cis– (74.9%) and trans-
oct-2-ene (24.1%) containing ∼1.0% oct-1-ene during hydro-
silylation with methyldichlorosilane catalysed by both Speier’s
J. Chem. Soc., Perkin Trans. 1, 2002, 1523–1534
1527

View Article Online
resin-supported catalysts examined, two with a dry state surface
area of ∼100 m2 gϪ1 and a pore volume of ∼1 ml gϪ1, with a Pt
loading of 0.5 mmol gϪ1 and a ligand–Pt ratio of ∼2 : 1 dis-
played the best combination of activity, selectivity and stability.
Subsequent work demonstrated that higher ligand–Pt ratios
provided no advantage in terms of reducing Pt leaching.15 The
design criteria in the present two series of supported catalysts
C1–C6 and C7–C10 were therefore: i) to use styrenic-based
resins; ii) to employ a high level of crosslinker in resin prepar-
ations with two suitable porogens to general high surface area
resins; iii) to use a common level of porogen (1 : 1 v/v with
comonomers) to generate a common pore volume in the prod-
ucts; iv) to vary the Pt content of catalysts (0.25–0.75 mmol gϪ1
keeping the ligand–Pt ratio at 1 : 1 (See Tables 2, 4 and 5).
)
Fig. 2 Room temperature hydrosilylation of oct-1-ene by methyl-
Two series of catalytic experiments were carried out in
order to assess the relative activity, selectivity and recyclability
of each polymer catalyst. In the first a given sample of each
catalyst was used under standard conditions in the room tem-
perature solventless hydrosilylation of oct-1-ene using methyl-
dichlorosilane (see Experimental section). At the end of each
reaction the catalyst sample was recovered by decantation of
the reaction mixture and then re-used in a second catalytic
reaction. Each catalyst sample was used in up to 10 successive
cycles. Reactions were monitored in real time by GC and the
conversion of oct-1-ene to 1-octylmethyldichlorosilane calcu-
lated. The standard reaction conditions employed an oct-1-ene–
silane–Pt mole ratio of 2 : 1 : 1×10Ϫ3 and the conversion was
based on the feed of silane i.e. 100% corresponds to 50% con-
version of the starting oct-1-ene. At the same time the level of
isomerisation of the unused oct-1-ene was also monitored and
expressed as a % of the original oct-1-ene feed i.e. 50% isomeris-
ation corresponds to all the residual octene being present as
internal isomers with 0% oct-1-ene remaining. In order to dis-
play in a digestible form the large amount of data generated,
the activity of each catalyst has been expressed in terms of the
time (h) taken to achieve 90% conversion to 1-octylmethyl-
dichlorosilane, and the corresponding % isomerisation quoted
at this time. The results in histogram form for catalysts C1–C10
are shown in Figs. 1–10. For comparison similar data for
dichlorosilane catalysed by resin supported Pt complex C2. (Oct-1-ene–
silane–Pt = 2 : 1 : 1 × 10Ϫ3). Left-hand bar = time to achieve 90%
conversion; right hand bar = % isomerisation of alkene.
Fig. 3 Room temperature hydrosilylation of oct-1-ene by methyl-
dichlorosilane catalysed by resin supported Pt complex C3. (Oct-1-ene–
silane–Pt = 2 : 1 : 1 × 10Ϫ3). Left-hand bar = time to achieve 90%
conversion; right hand bar = % isomerisation of alkene.
Fig. 4 Room temperature hydrosilylation of oct-1-ene by methyl-
dichlorosilane catalysed by resin supported Pt complex C4. (Oct-1-ene–
silane–Pt = 2 : 1 : 1 × 10Ϫ3). Left-hand bar = time to achieve 90%
conversion; right hand bar = % isomerisation of alkene.
Fig. 1 Room temperature hydrosilylation of oct-1-ene by methyl-
dichlorosilane catalysed by resin supported Pt complex C1. (Oct-1-ene–
silane–Pt = 2 : 1 : 1 × 10Ϫ3). Left-hand bar = time to achieve 90%
conversion; right hand bar = % isomerisation of alkene.
The % isomerisation of residual octene was also calculated at
the time the polymer resin catalyst was removed and after 24 h.
Again this was expressed as a % of the original oct-1-ene feed.
The data for the odd numbered runs using catalyst samples
from C1–C10 are shown in Tables 7 and 8. The corresponding
even numbered catalyst runs were allowed to proceed in the
presence of polymer resin catalyst for 24 h and each mixture
was assayed at the end to ensure that reactions were proceeding
normally in the presence of heterogeneous polymer catalysts as
in the first series of experiments.
Comparison of the data from the even numbered catalyst
runs (not recorded here) in the second series of reactions with
those from the first series showed the same activity trends for
each catalyst, though the levels of oct-1-ene isomerisation were
generally lower. The reason for this (see long-term stability
catalyst C11 are shown in Fig. 11 (note shift in scale of vertical
axis) this control species being loaded with Pt as with the other
resins, but possessing no diamine ligand.
The second series of catalytic experiments was designed pri-
marily to detect and deconvolute any contribution to catalysis
from soluble leached Pt species. Again each sample of catalyst
was re-used up to 10 times under identical reaction conditions
to those previously employed. In each odd numbered experi-
ment (i.e. runs 1, 3, 5 etc.) the reaction was allowed to run and
was monitored by GC until a reasonable conversion (∼20–40%
if possible) was detected. The reaction mixture was then
decanted from the heterogeneous polymer resin catalyst and
monitoring of any subsequent additional conversion in the
absence of heterogeneous resin catalyst continued for 24 hours.
1528
J. Chem. Soc., Perkin Trans. 1, 2002, 1523–1534

View Article Online
Fig. 8 Room temperature hydrosilylation of oct-1-ene by methyl-
dichlorosilane catalysed by resin supported Pt complex C8. (Oct-1-ene–
silane–Pt = 2 : 1 : 1 × 10Ϫ3). Left-hand bar = time to achieve 90%
conversion; right hand bar = % isomerisation of alkene.
Fig. 5 Room temperature hydrosilylation of oct-1-ene by methyl-
dichlorosilane catalysed by resin supported Pt complex C5. (Oct-1-ene–
silane–Pt = 2 : 1 : 1 × 10Ϫ3). Left-hand bar = time to achieve 90%
conversion; right hand bar = % isomerisation of alkene.
Fig. 9 Room temperature hydrosilylation of oct-1-ene by methyl-
dichlorosilane catalysed by resin supported Pt complex C9. (Oct-1-ene–
silane–Pt = 2 : 1 : 1 × 10Ϫ3). Left-hand bar = time to achieve 90%
conversion; right hand bar = % isomerisation of alkene.
Fig. 6 Room temperature hydrosilylation of oct-1-ene by methyl-
dichlorosilane catalysed by resin supported Pt complex C6. (Oct-1-ene–
silane–Pt = 2 : 1 : 1 × 10Ϫ3). Left-hand bar = time to achieve 90%
conversion; right hand bar = % isomerisation of alkene.
Fig. 10 Room temperature hydrosilylation of oct-1-ene by methyl-
dichlorosilane catalysed by resin supported Pt complex C10. (Oct-1-
ene–silane–Pt = 2 : 1 : 1 × 10Ϫ3). Left-hand bar = time to achieve 90%
conversion; right hand bar = % isomerisation of alkene.
Fig. 7 Room temperature hydrosilylation of oct-1-ene by methyl-
dichlorosilane catalysed by resin supported Pt complex C7. (Oct-1-ene–
silane–Pt = 2 : 1 : 1 × 10Ϫ3). Left-hand bar = time to achieve 90%
conversion; right hand bar = % isomerisation of alkene.
potentially on a larger scale as well. Reactions showed no
induction period and there was no tendency to exotherm. For
many of the polymer catalysts activity was retained during
extensive recycling (up to 10 runs) and again this offers con-
siderable benefits, not the least in cost terms, for use on a
laboratory scale. The behaviour of catalyst C11 was in marked
contrast to the others (C1–C10). The data in Fig. 11 show that
while the activity in hydrosilylation is high in run 1 so is the level
of oct-1-ene isomerisation. Furthermore the activity falls off
dramatically by run 4 with oct-1-ene isomerisation remaining
high ∼50%. These results suggest that Pt is lost quickly from this
resin support and that much of the initially observed catalytic
activity arises from Pt species leached into the reaction
medium. Since this resin contains no immobilised diamine lig-
and this behaviour seems totally reasonable, and demonstrates
the vital role the ligand plays in the other resin catalysts in
retaining Pt as a discrete molecular complex.
discussion later) is almost certainly the time lapse between the
first series of experiments and the second (up to 8 months)
during which no special precautions were taken in storing the
polymer resin catalysts which were simply kept in air in screw-
top sample bottles. Despite this, however, consideration of the
results of both series of experiments taken together provides a
valuable in-sight into the relative behaviour of the different
catalysts.
General observations
As in our previous work all the polymer-supported catalysts
show lower activity than Speier’s catalyst. However, the activity
is very good, certainly in terms of use on a laboratory scale, and
J. Chem. Soc., Perkin Trans. 1, 2002, 1523–1534
1529

View Article Online
Table 7 Leaching evaluation of resin catalysts C1–C6 in the reaction of oct-l-ene with methyldichlorosilane at room temperaturea
Run 1 Run 3 Run 5 Run 7 Run 9
Conv. (%) Isom. (%) Conv. (%) Isom. (%) Conv. (%) Isom. (%) Conv. (%) Isom. (%) Conv. (%) Isom. (%)
Resin
catalyst
Time
C1
0 min
15 min
20 min
35 min
90 min
24 hb
3
36
—
—
—
95
2
33
—
—
89
0.6
2.6
—
—
—
39.5
0
0.9
—
3
—
39
—
—
96
1
—
36
—
50
2
—
27
—
—
74
0
—
15
—
—
13
0
—
36
—
—
45
0
—
34
—
—
38
0
—
2.7
6
—
30
—
—
33
2
—
—
20
26
0
—
18
—
—
23
Ϫ2
—
4
0.7
—
0.6
—
—
1
5
—
—
33
—
42
0
—
—
17
13
Ϫ3
—
—
8
—
13
1
—
—
1
0
0
—
—
—
46
54
—
—
—
—
—
3
—
—
—
6
12
1
0.9
—
—
—
0.6
0.7
—
—
—
—
—
0
—
—
—
0
0
0
—
—
—
1.1
0
0.3
—
—
—
0.8
0.8
1.2
—
—
—
0.6
0.6
—
—
0.3
—
0.4
0
—
—
40.8
0
—
0.9
—
0.8
0
—
1.4
—
—
32.7
0
—
0.5
—
—
0.5
0
—
1
—
—
10.5
0
—
0
—
—
0.8
C2
C3
0 min
0
20 min
25 min
35 min
24 hb
—
—
0.5
0.4
0
—
0.3
—
—
0.3
0
—
0
—
—
0
0.2
—
0.6
—
—
0.4
0.2
—
0.5
—
—
0.5
—
—
0.1
0.1
0.2
—
—
0.2
—
0.3
0.4
—
—
0.4
—
0.8
0.2
—
—
0.2
—
0.8
0
—
6.5
0.4
1.3
—
—
—
0 min Ϫ1
15 min
20 min
35 min
90 min
24 hb
21
—
—
—
70
32
C4
C5
C6
0 min Ϫ2
0.4
0.4
—
—
—
0.4
0.4
0.4
—
—
—
2.6
0.4
0.4
—
—
—
15 min
20 min
35 min
90 min
24 hb
17
—
—
—
20
—
—
—
Ϫ2
4
—
—
9
—
7
2
0 min Ϫ2
Ϫ1
—
29
—
—
25
Ϫ5
—
19
—
—
21
1
15 min
20 min
35 min
90 min
24 hb
29
—
—
—
36
0
17
—
—
—
18
—
—
11
—
21
4
—
—
16
—
17
—
—
—
24
28
Ϫ4
—
—
—
22
26
0 min
15 min
20 min
35 min
90 min
24 hb
—
—
0.2
—
0.2
0.4
a See Experimental for further details. b Analysis of supernatant 24 hours after being decanted from the catalyst.
Pt complexes are active selectively in hydrosilylation. The pic-
ture for catalyst C2 is similar, while for C3 the level of leaching
seems lower. In contrast the data for C4–C6 suggest that these
resin catalysts leach very little Pt, or at least active Pt species.
In these experiments therefore catalyst C6 is the best per-
former along with C5. These resins have the highest Pt loading
(∼0.75 mmol gϪ1) and a corresponding content of diamine
ligand. Interestingly these species are identical except for the
method of loading Pt. C5 was loaded in one-step, whereas C6
was loaded in 5 steps but using the same total quantity of Pt. It
seems that the latter method produces a more uniform distribu-
tion of Pt complex and overall better immobilisation of the Pt.
C1 and C2 are also identical species except for the same vari-
ation of method of loading. While Pt leaching is evident from
both of these, C2, loaded in multiple steps seems somewhat
better than C1.
The results using catalysts C7–C10 show similar trends
(Table 8), although the rather variable data for C10 spoils the
quality of the correlation. Why this is so is unclear. However,
C7 clearly shows evidence of leaching active Pt species in runs
1–3, while C8 and C9 display low or negligible leaching.
These results must be borne in mind when considering the
more extensive catalyst activity and selectivity data generated in
the first series of experiments.
Fig. 11 Room temperature hydrosilylation of oct-1-ene by methyl-
dichlorosilane catalysed by resin supported Pt complex C11. (Oct-1-
ene–silane–Pt = 2 : 1 : 1 × 10Ϫ3). Left-hand bar = time to achieve 90%
conversion; right hand bar = % isomerisation of alkene.
Leaching behaviour
Considering the data for catalyst C1 in Table 7, in runs 1 and 3
hydrosilylation of oct-1-ene continues significantly over 24 h
after the polymer catalyst is removed and the % isomerisation
observed is simultaneously high, ∼40%. In runs 5, 7 and 9 how-
ever hydrosilylation essentially stops when the heterogeneous
catalyst is removed, and at the same time the level of isomeris-
ation drops dramatically to ∼1%. These results suggest that Pt
leaching is significant in runs 1–3 (potentially also in run 4) and
that thereafter leaching is minimal or absent. The data also
provides powerful circumstantial evidence that it is leached
soluble Pt species that are largely responsible for catalysing oct-
1-ene isomerisation, and that the tightly bound heterogeneous
Catalyst activity and selectivity in extensive recycling
Catalysts C1–C6. Very rewardingly the extensive catalyst
recycling data for resin catalysts C1–C6 correlate well with the
experimental data from the leaching studies. Catalysts C1–C3
show good activity over 10 cycles, and in the case of C1 the
particularly impressive performance is retained throughout. All
1530
J. Chem. Soc., Perkin Trans. 1, 2002, 1523–1534

View Article Online
Table 8 Leaching evaluation of resin catalysts C7–C10 in the reaction of oct-l-ene with methyldichlorosilane at room temperaturea
Run 1 Run 3 Run 5 Run 7 Run 9
Conv. (%) Isom. (%) Conv. (%) Isom. (%) Conv. (%) Isom. (%) Conv. (%) Isom. (%) Conv. (%) Isom. (%)
Catalyst Time
C7
0 min
0
0
0.8
—
3
—
41
—
—
—
96
2
—
6
—
—
—
8
0
—
2.2
—
—
—
24.8
0.4
—
0.6
—
—
—
0.6
0
0.7
—
—
—
—
3
—
—
59
—
—
80
2
—
—
28
—
—
7
0.5
—
—
1.1
—
—
5.2
0.5
—
—
0.4
—
2
—
—
—
69
—
87
0
—
—
—
4
—
11
0
—
—
—
27
—
45
1
0.4
—
—
—
3.1
4
—
—
—
55
—
60
2
—
—
—
—
41
5
0
20 min 49
—
—
—
0.7
—
0.7
0
—
—
—
—
0
0
0
—
—
—
—
0
25 min
45 min
1 h
—
—
51c
—
79
3
—
2.2c
2 h
—
10.6
0
0.2
—
—
—
—
1.3
0
—
0.8
—
—
—
0.8
1.2
—
0.9
—
—
0.9
—
24 hb
0 min
13.4
0.2
—
—
—
0.7
—
0.9
0.5
—
—
—
0.5
—
2.3
0.9
—
—
—
C8
20 min 29
25 min
1 h
—
—
—
—
33
3
90 min
2 h
—
24 hb
0 min
25 min
0.5
0.6
—
—
0.5
—
C9
0
1
2
—
24
—
—
—
—
40
0
43
—
—
—
85
—
—
35
—
—
42
6
—
—
57
—
70
—
—
—
—
20
24
3
—
—
—
49
54
28 min 33
1 h
—
—
—
39
0
90 min
2 h
—
24 hb
0 min
25 min
3.4
0.6
0.8
—
—
—
2.4
0.2
—
—
1.1
—
0
0
—
—
—
0.8
0
C10
—
—
—
—
58
90
28 min 56
45 min
1 h
—
—
60
1.8
15
24 hb
21.3
2.9
a See Experimental for further details. b Analysis of supernatant 24 hours after being decanted from the catalyst. c Analysis carried out as beads were
noticed in the supernatant – beads removed thereafter.
three catalysts also show a significant level of isomerisation in
runs 1–3 and thereafter isomerisation declines considerably.
Clearly therefore in runs 1–3 catalysis arises primarily from
immobilised Pt complexes but there is a contribution from
leached soluble Pt species. Catalysts C1 and C2 were loaded
originally with a 20 mol% excess of Pt relative to polymer lig-
and, and this may well have contributed to the early leaching
with these two species. Beyond run 4, however, the evidence is
that catalysis is overwhelmingly via the polymer supported Pt
complexes. Catalyst C4 (Fig. 4) is again very active initially but
beyond run 5 the activity falls dramatically. It is not clear why
this is so but it may be associated with the onset of accessibility
problems. The performances of C5 and C6 (Fig. 5 and 6) are
very similar which is not surprising since they differ only in the
method of loading the Pt. They are a little less active than C1
but nevertheless the activity is good and is maintained over 9
runs, with a low level of isomerisation. Overall however, C1 is
the most active and long-lived of the C1–C6 catalyst series, and
performs consistently better than our previous best polymer-
supported species.14 It has significantly higher surface area than
the latter (∼550 versus 100 m2 gϪ1) and indeed has the highest
surface area of the C1–C6 series. This result confirms our
hypothesis15 that surface area is a particularly important
parameter in determining the activity of polymer-supported
catalysts in these solventless reactions. Catalyst C1 also has the
lowest Pt loading of the C1–C6 series (0.25 mmol gϪ1) which
suggests that generating discrete isolated ligand PtCl2 species is
important in optimising the use of loaded Pt. However we have
shown previously15 that catalyst activity declines when the Pt
loading becomes too low (∼0.03 mmol gϪ1), at least in the case
of a resin with modest surface area (∼100 m2 gϪ1). Clearly there-
fore resin surface area and Pt loading are both important, and
in the latter context it remains unclear what proportion of the
Pt complexes are actually active in catalysis. Some indirect evi-
dence relevant to this was gleaned from additional experiments
with catalysts C1 and C4. A sample of each of these was finely
dry crushed in a mortar and pestle, and then used repeatedly in
hydrosilylation reactions. With both catalysts the activity was
higher than when the beaded form was used, ∼100% con-
version being reached typically in 30 min. With C1 the level of
isomerisation was ∼15%, whereas with C4 it was ∼5%. With
both crushed catalysts the reaction mixtures warmed a little
when the catalyst was contacted, but the exotherm was by no
means as severe as that with Speier’s catalyst. The balance
between hydrosilylation and isomerisation for each crushed
catalyst sample is similar to that seen with the corresponding
intact beaded samples and it seems therefore that crushing does
expose catalyst sites previously inaccessible, but does not for
example increase significantly the level of Pt leaching.
Catalysts C7–C10. The results for the second series of poly-
mer catalysts, C7–C10, did not correlate as well with the data
from the leaching studies as did the first series. Why this is so is
not clear. Use of n-butyl acetate instead of toluene as the poro-
gen in synthesising the precursor resins would be expected to
change the finer detail of the pore structure of the resins, but
the dry state surface areas of the derived catalysts C7–C10
(Table 5) fall within a similar range to those of C1–C6, and so it
is difficult to believe that this is a major factor. The poor corre-
lation between the leaching studies and the recycling experi-
ments almost certainly arises in part because of the 8 months
storage time between the two series of experiments but why
should the correlation be poorer for C7–C10 than for C1–C6?
In the recycling experiments C7 is characterised by a high
hydrosilylation activity (Fig. 7) but also a consistently high level
of isomerisation. C8 on the other hand shows good catalytic
activity which falls off a little beyond run 6, accompanied by a
consistently low level of leaching (Fig. 8). C9 and C10 are very
similar with good activity maintained over 10 cycles but in each
case accompanied by significant levels of isomerisation, which
appears to grow to a maximum at run 5 and thereafter decline
again (Figs. 9, 10). Relative to the results of the leaching
behaviour studies the biggest difference is in the level of iso-
merisation detected which in the leaching studies is consistently
J. Chem. Soc., Perkin Trans. 1, 2002, 1523–1534
1531

View Article Online
low. A possible rationalisation lies in the mechanism of the
deactivation which arises when the catalysts are stored without
protection from the air.
activity is therefore similar to that in the earlier experiments
(Fig. 1) though the isomerisation is a little lower (the catalyst
having been stored for some months – see previous section).
The catalyst was then recovered and washed with three
successive mixtures of oct-1-ene and n-nonane for 20, 10 and
13 h respectively, the composition and quantity of each wash
mixture being the same as in a catalytic run (but with no silane
present). Wash mixtures 2 and 3 each then received the appro-
priate quantity of methyldichlorosilane used in a normal
reaction and the solutions monitored for any catalytic activity.
Solution 2 showed a conversion to 1-octylmethyldichlorosilane
of ∼90% in 4 h with a level of isomerisation of ∼50%. The data
for solution 3 were ∼50% conversion in 4 h and ∼23% isomeris-
ation. Clearly therefore these wash solutions contain sufficient
Pt to show reasonable hydrosilylation activity, and certainly
good isomerisation activity. The washed sample of C1 was then
used in a second reaction under standard conditions yielding
∼90% conversion in 2 h with ∼4% isomerisation. The washing
procedure therefore leaves C1 as active in hydrosilylation as
previously, but reduces the isomerisation very significantly (see
Fig. 1). This tends to confirm therefore that the easily leached
Pt in early runs contributes in a significant manner to the
observed isomerisation.
Long-term catalyst stability
Catalyst C10 was reassessed in extensive recycling tests 17
months after its synthesis and 14 months after the first extensive
recycling tests (Fig. 10). The results are shown in Fig. 12. Two
A second washing experiment was also performed using Cl.
Firstly a catalyst sample was used under standard conditions
and the conversion was 100% after 3 h with ∼16% isomeris-
ation. The catalyst sample was then recovered by decantation
and washed for 5 min with each of 10 batches of oct-1-ene–n-
nonane. The catalyst sample was then used in a second reaction
under standard conditions and the conversion was >90%
and isomerisation ∼2% after 2 h. Thus the washing pro-
cedure almost eliminates isomerisation completely. Each of
the wash solutions was also assayed for activity by addition
of the appropriate amount of methyldichlorosilane. All proved
to be active and though there was some variation in the data
wash solution 1 gave ∼100 conversion after 24 h with ∼50%
isomerisation, and these fell to ∼50% conversion and ∼20%
isomerisation after 24 h with wash solution 10. These results
therefore confirm those from the longer washing experi-
ments and show that the isomerisation observed in the early
runs with polymer-supported catalysts such as C1 arises mainly
from reaction involving leached soluble Pt species. Washing
tends to eliminate this component and at the same time reduces
isomerisation with the polymer catalysts to very low levels.
Fig. 12 Room temperature hydrosilylation of oct-1-ene by methyl-
dichlorosilane catalysed by resin supported Pt complex C10. (Oct-1-
ene–silane–Pt = 2 : 1 : 1 × 10Ϫ3). Left-hand bar = time to achieve 90%
conversion; right hand bar = % isomerisation of alkene. Reassessment
of recycling behaviour 14 months later than earlier assessment
(Fig. 10).
major differences are clear. Firstly the activity in hydrosilylation
is reduced a little and in particular falls off progressively with
recycling. Perhaps more remarkably, however, the level of
isomerisation is very low and remains so through all cycles.
Overall this behaviour is more consistent with the results of
the leaching experiments (Table 8) than with the data from the
first recycling experiments (Fig. 10). Significantly the leach-
ing experiments were also carried out 8 months later than the
first recycling experiments. A very similar pattern of results
was obtained when catalyst C7 was re-subjected to recycling
experiments again, and then likewise with catalysts C5 and C9
(data not shown).
It is fairly clear therefore that storing the polymer-supported
Pt catalysts (in air) in simple screw top sample bottles is not
sufficiently adequate to maintain the catalysts in their original
state. Most likely oxidation to PtO2 occurs fairly readily, and
interestingly one can speculate that PtO2 formation might be
faster with unbound or weakly bound Pt located on or near the
resin surface. If we hypothesise further that oct-1-ene isomeris-
ation occurs primarily via leached soluble Pt species, and that
the most easily leached species are those located on or near the
resin surface, then clearly selective oxidation of these species on
storage in contact with air will reduce the isomerisation activity
of the polymer catalyst. Intriguingly therefore the selectivity
towards hydrosilylation of polymer bound Pt catalysts might
well be improved by deliberate oxidation of surface located
species. Alternatively if the original state of these catalysts
needs preserving during long-term storage then the presence
of a suitable inert (e.g. N2) or reducing atmosphere would be
advisable.
Conclusions
Polystyrene resins prepared with vinylbenzyl chloride comon-
omer and designed to be permanently porous with a high dry
state surface area (∼500 m2 gϪ1) are readily aminated and yield
good supports for Pt() chloro complexes. The so-formed
polymer-supported Pt complexes are highly active and selective
heterogeneous catalysts in the room temperature, solvent-
less hydrosilylation of oct-1-ene using methyldichlorosilane.
Though the hydrosilylation activity is lower than that of soluble
Speier’s catalyst, the activity is high, and certainly high enough
to be a practical value in the laboratory, and potentially on a
large scale. The polymer-supported catalysts can be easily
recovered and re-used up to 10 tens with useful activity. They
also show a very much reduced tendency to isomerise oct-1-ene
during hydrosilylation and so are significantly more selective
than Speier’s catalyst under the conditions employed. Careful
control washing experiments have shown that the isomerisation
that does arise with the polymer catalysts seems to derive
primarily from Pt species leached from the support, and that
isomerisation can be almost eliminated by appropriate washing.
Polymer catalysts stored in contact with air in simple screw top
sample bottles for many months remain active but the activity is
reduced somewhat; most remarkably however such storage
reduces considerably the isomerisation activity of the catalysts
Catalyst washing experiments
In an attempt to confirm the picture that some weakly bound Pt
is rapidly lost from polymer-supported species and that this
component is largely responsible for much of the isomerisation
observed a number of careful catalyst washing experiments
were carried out using species C1. Firstly C1 was used under
identical conditions to those in the earlier recycling experiments
and the conversion to 1-octylmethyldichlorosilane was >90%
within 2 h with the level of isomerisation being ∼13%. The
1532
J. Chem. Soc., Perkin Trans. 1, 2002, 1523–1534

View Article Online
most likely as a result of oxidation of weakly surface bound Pt
to PtO2 which effectively eliminates leachable active Pt species.
The catalyst displaying the optimum balance of activity,
selectivity and recycling stability is C1. This has a surface area
of ∼550 m2 gϪ1, a Pt loading of ∼0.25 mmol gϪ1, a ligand–Pt
ratio of ∼1 : 1 and it out performs those polymer catalysts of
lower surface area that we have reported earlier.14,15 We believe
that we are now close to achieving the optimum combination of
physical and chemical parameters in these polymer-supported
systems. There remains, however, methodologies for pushing
condenser, containing dry THF (50 cm3). Under dry N2, the
required amount of NaH was added and the mixture stirred for
15 min using an overhead stirrer. The appropriate precursor
resin beads (4.0 or 5.0 g) were then added and the temperature
raised to 60 ЊC. The mixture was then left stirring for 48 h, at
this temperature, under N2 gas. The quantities of TrMEDA and
NaH used were such that a 1 : 3 : 3.3 molar ratio of resin benzyl
chloride functionality : TrMEDA : NaH was present in the
reaction mixture. The benzyl chloride contents of the precursor
resins are shown in Table 2. The reaction was quenched in
methanol (150 cm3), filtered and the beads washed with THF.
The beads were then placed in dilute NaOH (2 M) for 1 h, with
constant stirring, before extraction in a Soxhlet apparatus with
water (24 h), and then THF (24 h) before vacuum drying. The
FTIR spectra indicated a reduction in intensity of the CH2Cl
band at 1268 cmϪ1, and the appearance of shoulders at 2774
and 2813 cmϪ1 on the main CH bands assigned to the tertiary
amine groups. The corresponding elemental microanalytical
data are shown in Table 3.
the surface area of resins beyond even the figure of ∼550 m2 gϪ1
,
and we are currently exploring this possibility.
Experimental
Materials
Styrene was supplied by Fisher Scientific and vinylbenzyl chlor-
ide (VBC) was secured from the Dow Chemical Company as a
mixture of para- and meta-isomers in a ratio of 44% : 56%.
Divinylbenzene (DVB) was supplied as 80% technical grade by
Aldrich, again as a mixture of isomers (28% para- : 72% meta-),
with the bulk of the remaining 20% being ethylvinylbenzene.
The isomer ratios applicable to the VBC and DVB were deter-
mined via gas chromatography. Boric acid was supplied by
BDH, Cellosize (QP4400L) by Hythe Chemicals. Xanthan
Gum I (90/GV/233) by Kelco International and AIBN by
BDH. Toluene was from BDH, m-xylene, n-butyl acetate, 1,2-
dichlorobenzene, 1,2,4-trichlorobenzene, 2-ethylhexanoic acid
and 2-ethylhexan-1-ol were from Aldrich.
n-Nonane was supplied by Lancaster Chemicals, oct-1-ene
was obtained from Avocado Research Chemicals Ltd., oct-2-
ene was supplied by Lancaster Synthesis as a mixture of cis-
and trans-isomers in a ratio of 74.9%, : 24.1%, along with 1.0%
oct-1-ene (determined via gas chromatography), and trans-oct-
2-ene, trans-oct-3-ene and trans-oct-4-ene were from Aldrich.
Methyldichlorosilane was from Aldrich and 1-octylmethyl-
dichlorosilane from Fluorochem.
Loading of Pt to yield catalysts C1–C10 (Scheme 1)
Each aminated resin (2.0 g) was first wetted with THF (5 cm3)
before being placed in a poly(ethylene) bottle along with the
appropriate amount of potassium tetrachloroplatinate dis-
solved in doubly distilled water (200 cm3). The quantity of the
platinate used was stoichiometrically equivalent to the diamine
content of the aminated resin (AFR1–8) used, and was calcu-
lated from the values detailed in Table 2. The bottle was closed
and the resulting mixture was shaken for 24 h at room temper-
ature. The polymer-supported catalyst produced was collected
by filtration, washed with water (doubly distilled) and acetone,
and vacuum dried at 40 ЊC for 24 h. The elemental micro-
analytical data for all the supported catalysts (C1–10) are
shown in Table 4.
In the case of catalysts C2 and C6 the method of loading of
Pt was modified. The total amount of platinate used was calcu-
lated as before but the feed solution was divided into 5 equal
portions and each portion added to the resin over an extended
period of time. Successive portions were added when the solu-
tion cleared after the addition of the previous portion. Since
control precursor resin PR10 contained no ligand a typical
quantity of platinate was used in the loading procedure i.e.
0.498 g (1.20 mmol) for 2.0 g of resin. The experimental details
were otherwise as before.
Potassium tetrachloroplatinate was supplied by Johnson
Matthey and hexachloroplatinic acid was from Aldrich, as was
N,N,N Ј-trimethylethylenediamine. NaH (60%) stabilised in
mineral oil and isopropanol (propan-2-ol) (HPLC) grade were
from Aldrich. Cyclohexane, THF, acetone, dichloromethane
and methanol were all bench reagents. THF was distilled from
sodium–benzophenone. Cyclohexane, isopropanol, n-nonane
and oct-1-ene were dried over molecular sieves.
Platinium analysis
Synthesis of precursor resins PR1–PR14
A sample of polymer catalyst (∼100 mg) was weighed accur-
ately into either a silica or alumina crucible. The crucible, with-
out its lid, was then placed in a muffle furnace and the temper-
ature raised to 300 ЊC at a rate of 2.0 ЊC minϪ1. The temperature
was then raised further, at a slower rate (0.5 ЊC minϪ1), until it
reached 600 ЊC. This temperature was maintained for 6 h before
the crucible was allowed to cool. Aqua regia (5 cm3) was added
to the dark residue, at room temperature, and the crucible (with
its lid fitted) returned to the furnace and heated to 110 ЊC where
it was held for 1 h. The resulting clear yellow solution was
diluted to 25 cm3 then assayed by ICP AAS.
These were all synthesised via suspension polymerisation using
a procedure already reported in detail elsewhere.23 The aqueous
stabiliser system was based on a mixture of Cellusize, Xanthan
gum and boric acid. The organic phase (100 cm3) comprised
comonomers (50 cm3) and porogen (50 cm3) as indicated in
Tables 1 and 2. AIBN (1% wt) was used as initiator and the
polymerisations were carried out at 80 ЊC for 5 h while stirring
at 600 rpm. Resin beads (typically 85–95%) were recovered by
filtration and were washed copiously with water and acetone.
The air dried resin was separated into various size fractions
and that fraction in the range 212–300 µm was selected and
exhausted extracted with acetone in a Soxhlet for 24 h before
being vacuum dried at 40 ЊC for 24 h. VBC-containing prod-
ucts displayed a band at 1268 cmϪ1 in the FTIR spectrum
assigned to the –CH2Cl group. Elemental microanalytical data
are summarised in Table 3.
Polymer resin catalysts were also subjected to elemental
microanalysis and the results of this and the Pt analysis are
summarised in Table 4.
Homogeneous hydrosilylation reactions
Homogeneously catalysed hydrosilylations using Speier’s
catalyst (Scheme 2) were carried out in 8 cm3 Pyrex vials which
were equipped with tight fitting screw caps. The reaction mix-
tures consisted of methyldichlorosilane (1.15 g, 10 mmol),
n-nonane (1.0 g, 7.8 mmol) and oct-1-ene (and occasionally oct-
2-ene) (2.24 g, 20 mmol). The n-nonane was used as the internal
Amination of VBC resins
Initially the appropriate amount of N,N,N Ј-trimethylethylene-
diamine (TrMEDA) was charged into a cooled (0 ЊC) three-
necked round-bottomed flask (100 cm3), fitted with
a
J. Chem. Soc., Perkin Trans. 1, 2002, 1523–1534
1533

View Article Online
reference for GC analysis. The appropriate amount of catalyst
was added at time zero corresponding to the desired silane–Pt
molar ratio of 1 : 1 × 10Ϫ3. A 5% w/w H2PtCl6ؒH2O in isopro-
panol catalyst was used. The vial was then shaken using a
Stuart Scientific shaker and the mixture was left to react under
ambient conditions. The reaction was followed by gas
chromatography.
Impact 400D instrument using 10 scans at a resolution of
4 cmϪ1 1H and 13C NMR analyses were recorded using a
400 MHz Bruker DPX-400 spectrometer.
.
GC analysis of each hydrosilylation reaction was carried out
using initially a Fisons DP 1701 (30 × 0.32 × 0.25) silica capil-
lary column and later a Chrompack CP-SIL 19CB (30 × 0.32 ×
0.25) silica capillary column. These were mounted on a Carlo
Erba HRGC 5300 Mega Series instrument fitted with a flame
ionisation detector. Hydrogen was used as the mobile phase and
n-nonane was used as an internal reference standard. Samples
were prepared for injection by dripping one drop of the
appropriate reaction mixture into 1 cm3 of cyclohexane and
0.2 µl of this mixture was injected and analysis run using a two
Heterogeneous hydrosilylation reactions
These were carried out essentially as described for the homo-
geneous reactions although in this case the appropriate amount
of the resin catalyst corresponding to a molar ratio of silane–Pt
of 1 : 1 × 10Ϫ3 was weighed initially into the vial. The reaction
mixture was then added at time zero and the above procedure
followed.
minute temperature program [60 ЊC (1 min)
100 ЊC (at 40 ЊC
minϪ1)]. It was necessary to calibrate the GC instrument at
regular intervals to maintain accuracy.
Confirmation of the identity of the hydrosilylation product
was also carried out via GC analysis. Samples were prepared for
injection by dripping one drop of the appropriate spent reac-
tion mixture into 1 cm3 of cyclohexane; 0.5 µm of this mixture
was then injected and analyses run using a ten minute temper-
Catalyst recycling reactions
In these experiments the performance of a sample of each
catalyst C1–C10 was evaluated in up to 10 back-to-back
catalytic reactions. After each run was complete the super-
natant liquid phase was carefully decanted from the reaction
vial, taking care that all of the resin catalyst beads remained
behind in the vial. The sample of catalyst was then recycled by
adding a fresh reaction mixture. The quantities of materials
were standardised as described above and each reaction was
monitored by GC.
ature program [50 ЊC (1 min)
maintained for 3 min)].
200 ЊC (at 25 ЊC minϪ1
;
Acknowledgements
S. J. T. acknowledges the receipt of a studentship from
Dow Corning. The gift of vinylbenzyl chloride from the Dow
Chemical Co. is appreciated. We also thank Dr G Wiltshire,
University of Paisley, UK for the supply of ICP AAS data.
Catalyst leaching reactions
Initially, a fresh reaction mixture and catalyst resin were added
to a reaction vial (as detailed above) and then shaken for a
previously decided length of time, this being long enough for an
intermediate conversion to be achieved. At this point the reac-
tion mixture was sampled for analysis in order to assess the
progress of the reaction. The supernatant (liquid phase) was
then decanted into a second vial, taking care that no resin
catalyst beads were also removed. The second vial was then
shaken for 24 h before being analysed to check whether further
conversion had occurred. Simultaneously, over the same 24 h
period, the original vial containing the heterogeneous catalyst
(along with a fresh reaction mixture) was also put through
another catalysis cycle. This whole procedure was then repeated
until each catalyst sample had catalysed either eight or ten
different hydrosilylations.
References
1 B. Marciniec, J. Gulinski, W. Urbaniak and Z. W. Kornetka,
Comprehensive Handbook on Hydrosilylation, ed. B. Marciniec,
Pergamon Press, Oxford, UK, 1992, ch. 2, p. 8.
2 J. L. Speier, J. A. Webster and C. H. Barnes, J. Am. Chem. Soc., 1957,
79, 974.
3 For a summary of work published prior to 1985 see: F. R. Hartley,
Supported Metal Complexes – A New Generation of Catalysts,
Reidel, Dordrecht, Germany, 1985, ch. 7, p. 204.
4 For a further summary of work prior to 1992 see Ref. 1, p. 84.
5 C. Kan, Q. Yuan, A. Luo and X. Kong, Polym. Ad. Technol., 1995, 7,
76.
6 Z. M. Michalska, B. Ostaszewski and J. Zientarska, React. Polym.,
1991/92, 16, 213.
7 Z. M. Michalska and B. Ostaszewski, J. Organomet. Chem., 1986,
299, 256.
8 Z. M. Michalska and B. Ostaszewski, J. Mol. Catal., 1989, 55, 256.
9 Z. M. Michalska, B. Ostaszewski, K. Strzelec, R. Kwiatkowski and
A. Wlochowicz, React. Polym., 1994, 23, 85.
10 Z. M. Michalska, B. Ostaszewski and K. Strzelec, J. Organomet.
Chem., 1995, 496, 19.
Control reactions
In those recycling experiments using crushed resin catalysts
it was necessary to centrifuge reaction mixtures (3000 rpm,
30 min) to effect separation of the catalyst particles.
In those recycling experiments where the catalyst resin beads
were washed between successive runs two procedures were
employed. In the first, recovered catalyst from decantation of
the previous reaction mixture was washed successively with
three oct-1-ene–nonane mixtures of the same composition and
volume as those used in a reaction. The wash solutions were left
in contact with catalyst for 20, 10 and 13 hours respectively.
Again separation was by decantation. In the second method
recovered catalyst was washed successively with ten oct-1-ene–
nonane mixtures as above each for 5 min only.
11 Z. M. Michalska, K. Strzelec and J. W. Sobczak, J. Mol. Catal. A:
Chem., 2000, 156, 91.
12 H. S. Hilal, M. A. Suleiman, W. J. Jondi, S. Kalaf and M. M.
Masoud, J. Mol. Catal. A: Chem., 1999, 144, 47.
13 R. Drake, R. Dunn, D. C. Sherrington and S. J. Thomson, Chem.
Commun., 2000, 1931.
14 R. Drake, R. Dunn, D. C. Sherrington and S. J. Thomson, J. Mol.
Catal., A: Chem., 2001, 177, 49.
15 R. Drake, R. Dunn, S. J. Thomson and D. C. Sherrington, J. Comb.
Chem., 2002 in press.
16 D. C. Sherrington, Chem. Commun., 1998, 2275.
17 R. L. Albright, React. Polym., 1986, 4, 55.
18 O. Okay, Prog. Polym. Sci., 2000, 25, 711.
19 D. H. Williams and I. Fleming, Spectroscopic Methods in Organic
Chemistry, 4th edn., McGraw-Hill, London, 1989.
20 C. S. Cundy, B. M. Kingston and M. F. Lappert, Adv. Organomet.
Chem., 1973, 12, 297.
21 J. L. Speier, J. A. Webster and G. H. Barnes, J. Am. Chem. Soc., 1957,
79, 974.
Analytical methods and instrumentation
Elemental microanalyses were carried out by the Microanalysis
Laboratory at the University of Strathclyde. ICP AAS was
carried out by the Centre for Particle Characterisation and
Analysis (CPCA) at the University of Paisley, Scotland. N2
BET analyses were performed using a Micromeritics ASAP
2000 instrument. FTIR spectra were collected using a Nicolet
22 J. C. Saam and J. L. Speier, J. Am. Chem. Soc., 1958, 80, 4104.
23 P. M. van Berkel and D. C. Sherrington, Polymer, 1996, 37, 1431.
1534
J. Chem. Soc., Perkin Trans. 1, 2002, 1523–1534