Lightly fluorous [Pd(OAc)2{P(C6H4-p-SiMe2CH2CH2C6F13)3}2] in the methoxycarbonylation of styrene: Formation, performance and stability of the catalyst system

Article abstract of DOI:10.1016/j.molcata.2006.05.052

The fluorous palladium(II) complex, [Pd(OAc)₂{P(C₆H₄-p-SiMe₂CH₂CH₂C₆F₁₃)₃}₂], has been prepared and characterized. Its application in the catalytic methoxycarbonylation of styrene in an MeOH/CF₃C₆H₅ mixture (1/1 v/v) has been explored and its activity was compared to that of [Pd(OAc)₂(PPh₃)₂]. The fluorous complex showed a lower activity but a significantly higher selectivity towards the branched product. Investigation of both the conversion-versus-time and i:n ratio-versus-time profiles showed an unusual behaviour in the case of the fluorous complex, which has been ascribed to the formation of a dinuclear species for the fluorous complex.

Full text of DOI:10.1016/j.molcata.2006.05.052

Journal of Molecular Catalysis A: Chemical 258 (2006) 334–340
Lightly fluorous [Pd(OAc) {P(C H -p-SiMe CH CH C F ) } ]
2
6 4
2
2
2 6 13 3 2
in the methoxycarbonylation of styrene: Formation,
performance and stability of the catalyst system
a,b
a,c,∗
b,1
a
Jeroen J.M. de Pater , Berth-Jan Deelman
, Cornelis J. Elsevier , Gerard van Koten
a
Organic Chemistry and Catalysis, Utrecht University, Padualaan 8, NL-3584 CH Utrecht, The Netherlands
Van’t Hoff Institute for Molecular Sciences, Molecular Inorganic Chemistry, University of Amsterdam,
b
Nieuwe Achtergracht 166, NL-1018 WV Amsterdam, The Netherlands
ARKEMA Vlissingen B.V., P.O. Box 70, NL-4380 AB Vlissingen, The Netherlands
c
Received 14 March 2006; received in revised form 19 May 2006; accepted 22 May 2006
Available online 11 July 2006
Abstract
The fluorous palladium(II) complex, [Pd(OAc)
2
{P(C
6
H
4
-p-SiMe
2
5
CH
2
CH
2
C
6
F
13
)
3
}
2
], has been prepared and characterized. Its application in
the catalytic methoxycarbonylation of styrene in an MeOH/CF
Pd(OAc) (PPh
3
C
6
H
mixture (1/1 v/v) has been explored and its activity was compared to that of
[
2
3
)
2
]. The fluorous complex showed a lower activity but a significantly higher selectivity towards the branched product. Investigation
of both the conversion-versus-time and i:n ratio-versus-time profiles showed an unusual behaviour in the case of the fluorous complex, which has
been ascribed to the formation of a dinuclear species for the fluorous complex.
©
2006 Elsevier B.V. All rights reserved.
Keywords: Fluorous; FBS; Triarylphosphine; Palladium; Methoxycarbonylation; Styrene; Catalyst recycling
1
. Introduction
the linear (n) or branched (i) product. The branched esters are of
interest to the pharmaceutical industry because they are precur-
sors for an important class of non-steriodal, anti-inflammatory
drugs like naproxen [6] and ibuprofen [7]. In order to comply
with legislation, these compounds should be obtained in pure
form and be devoid of even traces of catalyst remaining in the
product. Furthermore a high selectivity towards the branched
product is desirable.
The methoxycarbonylation of vinyl-arenes is an interesting
topic for academic research and has been performed industrially
1–5]. In this process, a transition metal species catalyzes the
formation of methyl esters from alkenes, CO and MeOH (Eq.
1)):
[
(
A number of catalysts have been used for this reaction, some
of which give a high selectivity towards the desired branched
product. Generallythese reactions suffer from lowreaction rates,
because of the low temperatures required to obtain good selec-
tivities. Raising the reaction temperature leads to higher rates
indeed, but also to a complete loss or even a reversal in selectiv-
ity providing more of the linear product. A further aspect that
has been addressed in several studies during recent years is cat-
alyst recycling. This research is mainly stimulated to deal with
the high palladium loadings (0.1–5 mol%) that are needed to
get high conversions in reasonable reaction times. In most cases
catalyst recycling was attempted by performing the reaction in
an aqueous biphasic solvent system, in which the catalyst was
preferentially dissolved in the aqueous phase whereas the sub-
(
1)
The most active catalysts for this reaction are based on palla-
dium(II) and phosphine ligands. Both monodentate [1–4] and
bidentate phosphine ligands [5] can be used, while addition of
strong acids like p-toluenesulfonic acid or triflic acid is needed to
boost activity. An important issue is the selectivity towards either
∗
Corresponding author. Tel.: +31 113 617000; fax: +31 113 612984.
E-mail addresses: berth-jan.deelman@arkemagroup.com (B.-J. Deelman),
else@science.uva.nl (C.J. Elsevier).
1
Fax: +31 20 5256456.
1
381-1169/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2006.05.052

J.J.M. de Pater et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 334–340
335
strates and products are present in the second, mostly organic
phase. Consequently, alkoxycarbonylations under biphasic con-
sitions are limited to those substrates that have a relatively good
solubility in the aqueous phase [8,9].
An alternative biphasic separation technique based on fluo-
rous biphase systems (FBS) and fluorous catalysts [10–13] has
not yet been explored for metal-catalyzed alkoxycarbonylation
reaction. It is obvious that the dissolution of polar transition
metal catalysts in the fluorous phase requires their function-
alization with perfluoroalkyl groups, which will increase the
fluorophilicity of the precursor complexes. In this paper we
present our attempts to use the FBS technique for the methoxy-
carbonylation of styrene.
by H. Kolbe Mikroanalytisch Laboratorium, M u¨ lheim an der
Ruhr. P{C H4-p-SiMe2(CH2CH2C F13)}3 was synthesized as
6
6
reported [14].
Note. Little is known yet about the toxicology of fluorous
phosphines and their metal complexes. It is therefore recom-
mended that proper personal protection measures are taken as
a precaution when working with these kinds of compounds.
Because of their amphiphilic nature, especially contact with the
skin should be avoided.
2.2. Synthesis of [Pd(OAc)2(PPh3)2] (1) [15]
In a typical experiment, [Pd(OAc)2] (0.24 g, 1.06 mmol) was
dissolved in MeOH (20 mL). The [Pd(OAc)2] had dissolved in
less than 30 s. Immediate addition of PPh3 (0.56 g, 2.12 mmol)
to the orange solution resulted in the formation of a yellow sus-
pension. After stirring for 5 min at room temperature all volatiles
were evaporated in vacuo. The remaining orange-yellow solid
was washed twice with Et2O and dried, affording a yellow solid
To study the effect of perfluoroalkyl tail-functionalization
on the catalytic performance and stability of Pd-based alkoxy-
carbonylation reactions, a lightly fluorous palladium complex
[
(
Pd(OAc)2{P(C H4-p-R )3}2] (R = SiMe2CH2CH2C F13)
6
F6
F6
6
2) was prepared and its performance in the methoxycarbonyla-
tion of styrene was compared with the well-studied non-fluorous
analogue [Pd(OAc)2(PPh3)2] (1). A related fluorous Pd(0)
complex, [Pd(0)(P{C H4-p-SiMe2(CH2CH2C F13)}3)2(MA)]
1
13
31
(0.73 g, 0.98 mmol, 92%). NMR-data ( H, C, P) were in
agreement with literature values [15].
6
6
(MA: maleic anhydride), was found to be too unstable in the
methoxycarbonylation of styrene [28]. The successful applica-
tion of lightly fluorous P, N ligands in the Pd-catalyzed methoxy-
carbonylation of phenylacetylene, in regular solvents as well as
supercritical CO2, was recently published [29].
2.3. Synthesis of [Pd(OAc)2(P{C H4-p-SiMe2
6
(CH2CH2C F13)}3)2] (2)
6
In a typical experiment, P{C H4-p-SiMe2(CH2CH2C
6
6
F13)}3 (0.39 g, 0.26 mmol) was dissolved in a mixture of
dichloromethane and methanol (1:2 v/v, 21 mL total volume).
To this slightly turbid yellow solution was added [Pd(OAc)2]
(0.029 g, 0.13 mmol). The solution immediately became orange
and turbid. After 5 min, when the colour changed from orange to
yellow, all solvents were removed in vacuo. Thedesiredcomplex
was isolated as a light-yellow solid (0.41 g, 0.12 mmol, 85%)
2
. Experimental
2
.1. General remarks
All reactions were performed using standard Schlenk tech-
niquesunderdinitrogenatmosphere. Diethylether, n-hexaneand
n-pentane were distilled from Na/benzophenone prior to use.
CH2Cl2 was distilled from CaH2. MeOH was distilled from
1
after washing with n-pentane. H NMR (CDCl3, 300.1 MHz):
−
δ 0.29 (s, 36H, Si(CH3)2), 0.74 (s, 6H, H3CCOO ), 0.96
Mg(OMe)2. CF3C H was distilled from P2O and stored on
(m, 12H, SiCH2), 1.95 (m, 12H, SiCH2CH2), 7.62 (m, 24H,
aryl-H). 13C{ H} NMR (CDCl3, 125.7 MHz): δ −3.59 (s,
6
5
5
1
molsieves under a dinitrogen atmosphere. CDCl3 (Cambridge
Isotopes Laboratories, CIL) was degassed and stored under a
dinitrogenatmosphereonmolsieves. Deuteratedmethanol(CIL)
was distilled from NaOCD3 and stored on molsieves under a
dinitrogen atmosphere. [Pd(OAc)2] (Degussa), PPh3, p-toluene
sulfonic acid monohydrate and n-decane (Acros) were used as
received. Styrene was passed over a short column of neutral
alumina (Merck) immediately prior to use. CO (Hoekloos, Prax-
Si(CH3)2), 5.18 (s, SiCH2CH2), 21.38 (s, CH3C O), 25.98 (t,
2
JC–F = 25 Hz, SiCH2CH2), 114–121 (several m, C F13-tail),
6
130.63 (t, J = 25 Hz), 133.59 (m), 134.23 (m), 140.57 (s), 141.42
(s), 175.66 (s, C O). 31P{ H} NMR (CDCl3, 81.0 MHz): δ 15.6
(s, trans-complex); 17.3 (s, cis-complex). Elemental analyses:
Calc. C: 37.83, H: 2.86, F: 46.68, P: 1.95, Pd: 3.35. Found C:
37.81, H: 2.94, F: 46.78, P: 1.93, Pd: 3.28.
1
1
19
air, >99.9% purity) was used as received. Standard H, F and
3
1
1
P{ H} NMR spectra were recorded on a Varian Unity-INOVA
2.4. Synthesis of [Pd(OAc)(µ-OAc)(P{C H4-p-SiMe2(CH2
6
3
00 MHz, a Varian Mercury 200 MHz or a Varian Mercury-
CH2C F13)}3)]2 (3)
6
13
1
VxWorks 300 MHz spectrometer. C{ H} NMR spectra of
the fluorous complex were recorded on a Varian Unity-INOVA
A similar procedure as that for 2 was used, but now
a palladium-to-phosphine ratio of 1:1 was used. Starting
from P{C H4-p-SiMe2(CH2CH2C F13)}3 (0.36 g, 0.25 mmol)
1
13
1
5
00 MHz spectrometer. H and C{ H} NMR spectra were
31 1
referenced to the residual solvent peak, P{ H} NMR spectra
externally to H3PO4 (85%) and 19F NMR spectra externally to
CFCl3. Quantitative GC analysis were performed on a Varian
6
6
and [Pd(OAc)2] (0.058 g, 0.25 mmol) the final product was
1
obtained as a light-orange solid (0.25 g, 60%). H NMR
3
300 Gas Chromatograph, equipped with J&W Scientific Inc.
(CDCl3, 300.1 MHz): δ 0.29 (s, 36H, Si(CH3)2), 0.70 (s, 12H,
−
DB-5 column (30 m, 0.320 mm internal diameter) and a FID-
detector (280 C), using N2 as the carrier gas and attached to
a Varian 4400 integrator. Elemental analyses were carried out
H3CCOO ), 0.96 (m, 12H, SiCH2), 1.95 (m, 12H, SiCH2CH2),
◦
31
1
7.75 (m, 24H, aryl-H). P{ H} NMR (CDCl3, 81.0 MHz): δ
20.93. No suitable elemental analyses could be obtained.

3
36
J.J.M. de Pater et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 334–340
2
.5. Catalytic experiments
methoxycarbonylation of styrene [1]. We also used this approach
successfully in the methoxycarbonyation of phenylacetylene
by the related palladium acetate precursor [Pd(OAc)2{P(2-
C H N)(C H4-4-R )2}] [29].
Standard catalytic runs were performed by using two auto-
claves. One was a stainless steel autoclave with a total volume of
00 mL, equipped with a pressure gauge, a rupture disc assem-
6
5
6
F6
1
For the synthesis of the fluorous pre-catalyst, we used the
TM
bly and a Swagelok
connection towards the high pressure
fluorous phosphine P(C H4-p-R )3 that was already available
6 F6
system. To prevent contamination of the solution with residues
from the autoclave walls and bottom a glass insert was used
in all reactions. The second autoclave was an in-house built
stainless steel autoclave (V = 50 mL) [16], equipped with two
borosilicate (Maxos 200) windows to allow visual inspection of
the interior of the autoclave, a pressure transducer, temperature
gauges inside the autoclave and the wall, a rupture disc assembly
from previous studies [14]. The common route for the synthesis
of 1 involves the reaction of [Pd(OAc)2] and PPh3 in boiling
benzene for several hours [19]. When we initially employed this
synthetic route we obtained 1 in 80–90% yield after precipi-
tation and washing with n-hexane. This method, however, did
not work for 2. Instead, decomposition occurred and mixtures
of compounds were obtained at these temperatures, presumably
because of the low stability of compound 2 in boiling benzene
[20]. Conducting the synthesis at lower temperatures was also
not successful. In all cases mixtures of products were obtained,
containing the desired product in minor amounts.
We therefore adopted the method published by Toniolo
and co-workers in 1995, who prepared the complex [Pd(dppp)
(OAc)2] using MeOH as the solvent [15]. When applied to the
synthesis of 1, the desired complex could be obtained in a 92%
yield (Eq. (2)). However, the fluorous ligand did not completely
dissolve in MeOH. By changing the solvent to a 2:1 (v/v) mix-
ture of MeOH and CH2Cl2 (where CH2Cl2 ensured that all of
the fluorous phosphine had dissolved) and by applying short
reaction times, the desired compound 2 was obtained in good
yield (Eq. (3)) [22]:
(
set at 250 bar) and four heating elements in the wall for heat-
ing. In a typical experiment, a known amount of the appropriate
catalyst precursor, 30 equiv. of p-toluenesulfonic acid monohy-
drate (p-TsOH), 600 equiv. of degassed water and 150 equiv. of
styrene were mixed together in the appropriate solvent(s), giving
−
3
a palladium concentration of ∼2 × 10 M. This solution was
charged to the autoclave under nitrogen flow. After sealing the
autoclave it was flushed three times with N2, followed by three
times flushing with CO. The autoclave was then heated to the
desired temperature. The reaction was started by pressurizing
the autoclave with 30 bar of CO. The content of the reactor was
stirred using a magnetic stirring bar. When the reaction time
had passed the autoclave was cooled with an icebath for 15 min.
Subsequently, the pressure was carefully released during 10 min
and the liquid phase was quantitatively analysed for reagents and
products by GC using n-decane as an internal standard.
MeOH
[
Pd(OAc) ] + 2PPh3 −→ [Pd(OAc) (PPh3) ]
(2)
2
2
1
2
2
.6. HP-NMR experiments
[Pd(OAc) ] + 2P(C H -p-R )
2
6
4
F6 3
MeOH/CH2Cl2(1:1)
−→
[Pd(OAc) (P{C H4-p-R } ) ],
2
6
2
F6 3 2
An in-house built 10 mm sapphire HP-NMR-tube [17,18],
equipped with a titanium head which allows the introduction of
gasses, waschargedwith1.5 mLofa1:1solutionofd4-methanol
and CF3C H , containing the catalyst precursor, styrene and the
appropriate amounts of acid and H2O. After sealing the tube
and flushing with N2, the tube was charged with CO (30 bar).
The tube was transported to the NMR spectrometer, which was
already warmed to 80 C. After shimming H NMR spectra were
recorded with 5 min intervals. After the desired reaction time the
NMR-tube was ejected from the spectrometer and transported
to a fumehood where the excess pressure was released.
Note. Since high pressures of CO are used, work should be
conducted in a fumehood equipped with a CO detection sys-
tem and sufficient ventilation capacity. Furthermore, laboratory
personnel must never be directly exposed to the pressurized cell
and sufficient personal protection must be ensured by employing
proper safety shields.
RF6 = –SiMe2CH2CH2C F13
(3)
6
6
5
1
13
1
31
1
Complex 2 was characterized by H, C{ H} and P{ H}
NMRspectroscopyandbyelementalanalyses. IntheNMRspec-
tra some differences between 1 and 2 were apparent. Whereas for
31
1
◦
1
1 only the trans-complex could be detected in its P{ H} NMR
◦
spectrum at 25 C (in CDCl3), the solution of 2 appeared to con-
sistofamixtureoftrans-andcis-isomers, thetrans-isomerbeing
the major species (90:10 trans/cis) [21]. The different propen-
sity of 1 and 2 towards dimer formation is most interesting.
Facile formation of dimeric [Pd(OAc)2{P(C H4-p-R )3}]2 (3)
6
F6
already took place during the synthesis of 2. Complex 3 was also
synthesized independently (Eq. (4)) [23]:
[
Pd(OAc) ] + P{C H4-p-R }
2
6
F6 3
MeOH/CH2Cl2(1:1)
1
2
−
→
[Pd(OAc)(␮-OAc)(P{C H4-p-R } )] ,
6
F6 3
2
3
3
. Results
RF6 = –SiMe2CH2CH2C F13
(4)
6
3
.1. Synthesis of Pd-complexes
Solubility studies involving 2 showed that this complex is
soluble (more than 0.1 g/mL) in a variety of commonly used
organic solvents (e.g., pentane, hexane, chloroform and di-
chloromethane). Furthermore, this compound is also soluble
After the initial investigations by Drent and co-workers at
Shell, [Pd(OAc)2(PPh3)2] (1) has been successfully used by the
group of Toniolo and co-workers as a catalyst precursor in the

J.J.M. de Pater et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 334–340
337
Table 1
Results for methoxycarbonylation of styrene using non-fluorous 1, fluorous 2
and 3 as catalyst precursorsa
Entry
Precursor
Solvent (v/v)
Time
Conversion
(%)
i:n
(min)
1
2
3
4
5
6
7
8
9
1
2
1
2
2
2
2
2
2
2
3
3
MeOH
MeOH
60
60
60
60
90
120
180
240
180
180
105
150
95
0
0.67
–
MeOH/BTF (1:1)
MeOH/BTF (1:1)
MeOH/BTF (1:1)
MeOH/BTF (1:1)
MeOH/BTF (1:1)
MeOH/BTF (1:1)
MeOH/BTF (1:2)
MeOH/BTF (2:1)
MeOH/BTF (1:1)
MeOH/BTF (1:1)
85
17
25
45
68
70
15
0
0.67
1.17
1.11
1.11
1.11
1.11
1.17
–
Fig. 1. Conversion and i:n vs. time plots for the methoxycarbonylation of styrene
in MeOH/BTF using complex [Pd(OAc)2(PPh3)2] (1) or [Pd(OAc)2(P{C6H4-
p-RF6}3)2] (2) as catalyst precursor. For conditions see Table 1.
1
1
1
0
1
2
10
16
1.63
1.63
In Fig. 1 it can be seen that when using 1 as the
pre-catalyst, the reaction was fast right from the beginning
a
◦
General conditions: 80 C, 30 bar CO, 30 equiv. of p-toluene sulfonic
−
1
−1
(TOF50 = 250 molstyrene molPd
h ) [24], whereas with com-
acid (p-TsOH), Pd:styrene = 1:150, 600 equiv. of H2O, [Pd] = 2 ␮mol/mL,
BTF = CF3C6H5.
plex 2 during the first circa 90 min conversion was rather low.
Thereafter, the activity increased and leveled off when circa 70%
−
1
−1
conversion was obtained. A TOF50 of 36 molstyrene molPd
h
in several fluorous solvents (perfluorohexane, 1-(perfluorobutyl)
tetrahydrofuran, and perfluoro methylcyclohexane) as well as
scCO2. Also in these solvents the equilibrium involving 2 and 3
was calculated for 2. Although the TOF50 values differ signif-
icantly, the TOF values at maximum activity only differ by a
−1
−1
h , TOFmax
factor of 2.5 (TOFmax for 2 = 99 molstyrene mol
Pd
(
Eq. (5)) is shifted to the side of 3 and free phosphine:
−
1
−1
for 1 = 250 molstyrene molPd
h ). Fig. 1 also shows that for
[
Pd(OAc) (P{C H4-p-R } ) ]
complex 1 the i:n ratio remained constant during the entire reac-
tion time. For complex 2 again a completely different picture was
observed. During the first 90 min, when the conversion was low,
a high i:n ratio is observed (initially i:n = 1.82), which slowly
decreased during that time period to i:n = 1.11. After that the
i:n ratio remains constant. By monitoring the reaction by High
2
6
2
F6 3 2
ꢀ 1
[Pd(OAc)(␮-OAc)(P{C H4-p-R } )]
3
6
F6 3
2
2
+
P(C H4-p-R ) , RF6 = –SiMe2CH2CH2C F13 (5)
6
F6 3
6
1
Pressure H NMR spectroscopy, using an in-house built 10 mm
3
.2. Catalytic methoxycarbonylation of styrene
Palladium pre-catalysts 1 with 2 were first compared in pure
sapphire NMR-tube [17,18], essentially identical results were
obtained, excluding a possible influence from sampling of the
reaction mixture.
MeOH as the solvent, applying almost identical conditions as
reported in the literature (150 equiv. of styrene were used instead
of 200 equiv.) [1]. For 1 this gave 95% conversion in one hour
with a similar i:n ratio of the products as reported in literature
We also tested 2 in a 1:2 mixture of methanol and ␣,␣,␣-
trifluorotoluene (Table 1, entry 9). This medium, we reasoned,
might provide a better compatibility with the fluorous cata-
lyst 2 and could possibly lead to higher conversion. However,
the opposite effect was observed (conversion of 15% after
180 min). Using a 2:1 biphasic mixture of methanol and ␣,␣,␣-
trifluorotoluene (all other conditions equal) did not result in any
conversion (after 180 min).
(Table 1, entry 1). When 2 was used as pre-catalyst in MeOH as
the solvent, no conversion was observed (entry 2), most likely
because of the poor solubility of the pre-catalyst in this solvent.
However, a 1:1 (v:v) mixture of MeOH and CF3C H provided
6
5
a homogeneous reaction mixture in which both 1 and 2 were
soluble under the reaction conditions. In the further catalytic
studies this solvent system was used.
As can be clearly seen from Table 1 (entries 3 and 4), com-
plex 2 still had a significant lower activity than 1. However, a
significantlyhigherselectivitytowardsthebranchedproductwas
observed when compared with the non-fluorous complex. When
the dimeric fluorous complex 3 was used as the precursor, the
activity was even lower (entries 11 and 12) and a significantly
higher selectivity towards the branched product was observed
The use of the fluorous complex 2 as the precursor led to
appreciable amounts of palladium black as compared to the case
when complex 1 was used as the pre-catalyst. This indicates that
the in situ formed active fluorous species are less stable under the
reaction conditions than those originating from the non-fluorous
complex 1. In fact initial experiments under fluorous biphasic
conditions failed due to a combination of Pd black formation and
instability of the catalyst system combined with a low reaction
rate due to mass transport limitations.
Reduction of catalyst decomposition was possible by
decreasing the reaction temperature. When the reaction with
(i:n = 1.63), which is higher than that observed for 2 (i:n = 1.11
◦
after 150 min).
2 was performed at 60 C, no palladium black formation was
To better understand the differences between 1 and 2, the
conversion and i:n ratio of the products for both complexes 1
and 2 were monitored over the course of the reaction (Fig. 1).
observed. However, conversion dropped dramatically (15% after
180 min) and more significantly than for non-fluorous 1 (45%
◦
after 90 min compared to >95% conversion at 80 C). Also the

3
38
J.J.M. de Pater et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 334–340
−
Scheme 1. Catalytic cycle for methoxycarbonylation of styrene [1]. For all cationic palladium complexes the counteranion is p-TsO . L = CO, PAr3 or MeOH.
selectivity of the reactions was found to be highly temperature
dependent (i:n values of 1.86 (1) and 2.33 (2)), which is in line
with previous observations for non-fluorous catalysts [25,26].
The fluorous complex 2 maintained a higher selectivity than
non-fluorous 1 for the branched ester.
present which has a low activity, but a high selectivity towards
the i-isomer. Gradually, during the first 90 min of the reaction a
second species is formed, which has a higher activity but a lower
selectivity towards the i-isomer.
To offer an explanation for this behaviour one has to consider
the proposed catalytic cycle for the methoxycarbonylation of
styrene as proposed in the literature in more detail (Scheme 1)
4
. Discussion
[1].
Several differences can be noted between non-fluorous com-
Before the catalysis starts, a cationic palladium hydride com-
plex 1 and fluorous complex 2. Whereas 1 was observed as the
trans-isomer only, 2 consisted of a mixture of cis- and trans-
isomers. Furthermore 2 is less stable in solution than 1, leading
to formation of the dimeric complex 3 and free phosphine (Eq.
plex has to be formed, which takes several steps, all involving
mononuclear Pd species. Since we observed during the synthe-
sis of 2 that this compound easily forms the dimeric complex
[Pd(OAc)(␮-OAc){P(C H4-p-R )3}]2 (3), such a process may
6
F6
(
4)). For 1 the formation of the corresponding dimeric com-
also compete with palladium hydride formation during catalysis
(Scheme 2). Compound 3 is likely to be converted further to
analogous tosylate complexes under the action of the excess of
p-TsOH.
pound ([Pd(OAc)2(PPh3)]2) was not observed. Although this
was not further investigated, we think that the presence of the
perfluoroalkyl groups is probably responsible for this behaviour.
Formation of the dimeric compound in the case of 3 may lead
to more favorable solvent–solute interactions.
The use of the fluorous palladium complex 2 as a precursor
in the methoxycarbonylation of styrene gave rise to a number
of interesting phenomena. During the first 90 min of the pro-
cess, a remarkable difference in reactivity between 1 and 2 was
observed. Whereas for the former a high conversion (>95% after
When tested independently, pre-catalyst 3 gave a low conver-
sion but a high i:n ratio. The observed rate and i:n profiles for 2
in the first 90 min can be explained when the formation of com-
pound 3 as a competing catalyst precursor is taken into account.
We see, however, that after these 90 min, activity increases and
the i:n ratio remains constant when 2 is used as the precursor
complex. We can qualitatively account for this by assuming a
situation where mononuclear diphosphine species derived from
2 and monophosphine complexes derived from dinuclear 3 are
competing catalysts. The reaction profile for 2 reflects the differ-
ent stability of these species and is possibly further complicated
by residual activity of decomposition products. When the reac-
9
0 min), and a constant i:n ratio are observed, 2 suffers from an
induction period characterized by a lower conversion and a high
but decreasing i:n ratio. After this initial phase the selectivity
reaches a plateau at a level that is still higher than that of 1. These
results suggest that in the catalysis with 2 initially a species is

J.J.M. de Pater et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 334–340
339
Scheme 2. Possible reactions during formation of the active species from 2, leading to formation of 3.
tion is performed with pure 3 as the precursor, this is not possible
and we see only the poor activity but good selectivity of the
species derived from 3.
Another interesting feature of the use of fluorous complex 2
compared to 1 is that, even after the initial induction period, it
gave a higher selectivity for the branched product. This is rather
surprising, since one would expect that introduction of the flu-
orous phosphines in the active species would lead to increased
steric bulk around palladium, which would favor the linear ester
branched product when compared with non-fluorous 1. The
observed conversion-versus-time and i:n ratio-versus-time pro-
files showed a remarkably different behaviour between the flu-
orous and non-fluorous complex, especially in the early stages
of the reaction. The differences can be explained in terms of
competing species derived from 2 and 3 in case of the fluo-
rous complex 2. The present results demonstrate that fluorous
Pd-complexes that are active in the methoxycarbonylation of
styrene can be prepared. However, as initial experiments under
fluorous biphasic conditions failed due to Pd black formation,
catalyst stability and recycling need further investigation.
(Scheme 1). However, we saw earlier that complex 2 in solu-
tion consists of a mixture of cis- and trans-isomers, whereas 1
appears as a solution of purely the trans-isomer. This tendency
of the fluorous complexes to form the cis-isomer could very well
play a role during catalysis as well. Although both pathways in
Scheme 1 involve cis-bis(phosphine)palladium intermediates in
which the acetate anions are replaced, the tendency of the flu-
orous catalyst precursor 2 to form more of the cis-coordinated
complex, despite opposing steric considerations, may be indica-
tive of less steric crowding induced by the fluorous ligand and
could be the origin of the higher i:n ratio obtained with the cat-
alyst system derived from 2. Although only speculative at this
point, the minimization of the number of solvent–solute inter-
actions and consequently an optimization of energetically more
stabilizing solvent–solvent interactions that result from pairing-
up of the fluortails in a cis-coordination mode, may play a role
here [27].
Acknowledgements
The authors thank J.-M. Ernsting for assistance with the HP-
NMR experiments. This work was supported by the Council for
Chemical Sciences from the Dutch Organization for Scientific
Research (CW-NWO).
References
[
[
1] A. Seayad, S. Jayasree, K. Damodaran, L. Toniolo, R.V. Chaudhari, J.
Organomet. Chem. 601 (2000) 100.
2] C.W. Lee, J.S. Lee, N.S. Cho, K.D. Kim, S.M. Lee, J.S. Oh, J. Mol. Catal.
80 (1993) 31.
[
[
3] C.B. Dart, M.E. Davis, Ind. Eng. Chem. Res. 33 (1994) 2887.
4] S. Oi, M. Nomura, T. Aiko, Y. Inoue, J. Mol. Catal. A: Chem. 115 (1995)
Although there are significant differences between fluorous
289.
2
and conventional precursor 1, these do not rule out catalyst
[5] O.V. Gusev, A.M. Kalsin, P.V. Petrovskii, K.A. Lyssenko, Y.F. Oprunenko,
C. Bianchini, A. Meli, W. Oberhauser, Organometallics 22 (2003) 913.
recycling by the FBS techniques. However, thermal instability
and Pd black formation is a major obstacle for recycling of an
active catalyst system. Hence, Pd systems that are more resistant
to Pd(0) formation are clearly desirable.
[
6] H. Zhou, J. Cheng, S. Lu, H. Fu, H. Wang, J. Organomet. Chem. 556 (1998)
39.
7] V. Elango, M.A. Murphy, G.N. Mott, E.G. Zey, B.L. Smith, G.L. Moss, EP
00 (1990) 892.
2
[
4
[
[
8] G. Papadogianakis, Catal. Lett. 47 (1997) 43.
9] S. Jayasree, A. Seayad, B.R. Sarkar, R.V. Chaudhari, J. Mol. Catal. A:
Chem. 181 (2002) 221.
5
. Conclusions
[
[
10] I.T. Horv a´ th, J. R a´ bai, Science 266 (1994) 72.
11] I.T. Horv a´ th, Acc. Chem. Res. 31 (1998) 641.
[12] E. de Wolf, B.J. Deelman, G. van Koten, Chem. Soc. Rev. 28 (1999) 37.
[13] W. Keim, M. Vogt, P. Wasserscheid, B. Driessen-H o¨ lscher, J. Mol. Catal.
A: Chem. 139 (1999) 171.
We have synthesized the fluorous complex [Pd(OAc)2
{
P(C H4-p-SiMe2CH2CH2C F13)3}2] (2) in good yield and
6
6
purity. Insolutionthiscomplexislessstablethanitsnon-fluorous
analogue [Pd(OAc)2(PPh3)2] (1), leading to the formation of
[
14] B. Richter, A.C.A. de Wolf, G. van Koten, B.J. Deelman, J. Org. Chem. 65
2000) 3885.
[
Pd(OAc)(␮-OAc){P(C H4-p-SiMe2CH2CH2C F13)3}]2 (3).
6
6
(
The mononuclear complex 2 was employed as a catalyst pre-
cursor in the methoxycarbonylation of styrene where it gave
rise to a lower activity, but a higher selectivity towards the
[
15] The synthesis of [Pd(OAc)2(PPh3)2] was done following a similar proce-
dure as reported for [Pd(OAc)2(dppp)], see: F. Benetollo, R. Bertani, G.
Bombieri, L. Toniolo, Inorg. Chim. Acta 233 (1995) 5.

3
40
J.J.M. de Pater et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 334–340
[
[
[
16] For a detailed description of the autoclave see: A.M. Kluwer, Ph.D. Thesis,
University of Amsterdam, NL, 2004 (Chapter 2).
17] Such a HP-NMR tube has been first described in 1985, see: D.C. Roe, J.
Magn. Reson. 63 (1985) 388.
18] At the University of Amsterdam a modified version of the HP-NMR tube as
described in Ref. [17] has been developed and employed in several studies.
For a description of this NMR tube as well as some applications see: C.J.
Elsevier, J. Mol. Catal. 92 (1994) 285.
[22] The use of a 2:1 (v/v) mixture of methanol and CH2Cl2 in the synthesis of
[Pd(OAc)2(PPh3)2] resulted in a lower yield (80–85%). Also in this case
only the trans-isomer was observed in solution ( P{ H} NMR).
[23] T.A. Stephenson, G. Wilkinson, J. Inorg. Nucl. Chem. 29 (1967) 2122.
[24] TOF50 = turn over frequency at 50% conversion.
[25] J. Girones, J. Duran, A. Polo, J. Real, J. Mol. Catal. A: Chem. 198 (2003)
77.
[26] A. Seayad, A.A. Kelkar, R.V. Chaudhari, L. Toniolo, Ind. Eng. Chem. Res.
37 (1998) 180.
[27] For theoretical, treatments that explain phase separation phenomena in
fluorous biphasic systems based on intermolecular interactions see:
(a) J.H. Hildebrand, D.R.F. Cochran, J. Am. Chem. Soc. 71 (1949) 22;
(b) R.L. Scott, J. Phys. Chem. 62 (1958) 136;
(c) P. Lo Nostro, Adv. Colloid Interf. Sci. 56 (1995) 245.
[28] J.J.M. de Pater, D.S. Tromp, D.M. Tooke, A.L. Spek, B.-J. Deelman, G.
van Koten, C.J. Elsevier, Organometallics 24 (2005) 6411.
[29] J.J.M. de Pater, E.P. Maljaars, E. de Wolf, M. Lutz, A.L. Spek, B.-J. Deel-
man, C.J. Elsevier, G. van Koten, Organometallics 24 (2005) 5299.
3
1
1
[
[
19] T.A. Stephenson, S.M. Morehouse, A.R. Powell, J.P. Heffer, G. Wilkinson,
J. Chem. Soc. (1965) 3632.
20] Analysis, of the reaction mixture by 31P NMR and comparison with liter-
ature data [21] allowed for the identification of several of the compounds
present in the mixture. Thus it could be found that the desired complex had
indeed been formed, along with amounts of [Pd(OAc)(␮-OAc)P{C6H4-p-
SiMe2(CH2CH2C6F13}3)]2, phosphine-oxide and Pd(0) complexes with 3
or 4 phosphine ligands attached.
[
21] C.W. Lee, J.S. Lee, N.S. Cho, K.D. Kim, S.M. Lee, J.S. Oh, J. Mol. Catal.
8
0 (1993) 31.