Assessment of the reusability of Pd complexes supported on fluorous silica gel as catalysts for Suzuki couplings

Article abstract of DOI:10.1002/ejoc.200500281

A thorough investigation of different perfluoro-tagged Pd complexes supported on fluorous silica gel (FSG) as catalysts for Suzuki reactions is presented. The rates of reaction upon recycling of the supported complexes were measured in order to determine the reusability of the catalysts. Perfluoro-tagged and untagged complexes adsorbed on FSG and unmodified silica gel were also compared. While with dimethoxyethane (DME) as a solvent the catalytic activity decreased upon reuse, significant activity was retained with recycled catalysts for reactions in water. With catalyst loadings of 0.001 mol-%, TONs as high as 526000 were obtained. Wiley-VCH Verlag GmbH & Co. KGaA, 2005.

Full text of DOI:10.1002/ejoc.200500281

FULL PAPER
DOI: 10.1002/ejoc.200500281
Assessment of the Reusability of Pd Complexes Supported on Fluorous Silica
Gel as Catalysts for Suzuki Couplings
Carl Christoph Tzschucke,[a] Vasyl Andrushko,[a] and Willi Bannwarth*[a]
Keywords: C–C coupling / Palladium / Recycling / Solvent effects / Supported catalysts / Fluorous silica gel
A thorough investigation of different perfluoro-tagged Pd
complexes supported on fluorous silica gel (FSG) as catalysts
for Suzuki reactions is presented. The rates of reaction upon
recycling of the supported complexes were measured in or-
der to determine the reusability of the catalysts. Perfluoro-
tagged and untagged complexes adsorbed on FSG and un-
modified silica gel were also compared. While with dime-
thoxyethane (DME) as a solvent the catalytic activity de-
creased upon reuse, significant activity was retained with re-
cycled catalysts for reactions in water. With catalyst loadings
of 0.001 mol-%, TONs as high as 526000 were obtained.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2005)
consecutive runs as a criterion for the reusability of the cat-
alyst, we soon recognized that yields alone are insufficient
to evaluate the recycling of catalysts. Assuming that the re-
action times are rather long but that the reaction is finished
after a short period of time and that there would be a par-
tial loss of catalytic activity in each cycle, it would take
longer reaction times in the following cycles to complete the
reaction but this would still give the impression that the
catalyst is completely recyclable. This common misunder-
standing has been treated in detail by Gladysz.[11]
We therefore set out to investigate the rate of reaction in
repetitive cycles in addition to the conversion, since this
should provide a good measure of the efficiency of catalyst
reuse and therefore allow different pre-catalysts to be easily
compared. In these studies we employed a number of palla-
dium complexes with monodentate and bidentate phos-
phane ligands (Figure 1).
Introduction
Since the seminal work of Horváth and Rábai, fluorous
catalysis has received ever growing attention as a strategy
for the separation and recovery of catalysts.[1–6] This con-
cept is based on the phenomenon that perfluorinated sol-
vents exhibit a temperature-dependent immiscibility with
many organic solvents, which allows for the formation of
liquid–liquid biphasic systems. Since the solubility of or-
ganic compounds in perfluorinated solvents can be in-
creased by the attachment of perfluoroalkyl chains, the so-
called “fluorous tags”, compounds labelled in this way, in
our context catalysts, can be selectively separated from
other compounds by simple liquid–liquid extraction. As an
alternative, the isolation or separation of the catalyst can
be performed by chromatography on a plug of fluorous sil-
ica gel (FSG), which is silica gel that has been modified
with perfluoroalkyl entities.[7,8]
Complexes 1a–e and the required ligands were prepared
according to our previously published procedures.[12–14] The
straightforward synthesis of the bidentate phosphanes ac-
cording to Scheme 1 started from perfluoroalkyl-tagged
aryl bromide 6.[15] For the preparation of methylene-
bridged diphosphane 7a and ethylene-bridged diphosphane
7b lithium–halogen exchange, followed by condensation
with the appropriate bidentate chlorophosphane, gave the
diphosphanes in 57% and 67% yields, respectively.
We have recently reported another approach, in which
we have used perfluoro-tagged Pd–phosphane complexes
physisorbed on FSG for catalysis of C–C couplings in or-
ganic solvents and water. The separation of the catalyst was
performed by simple filtration and it was possible to re-use
the supported catalyst several times.[9,10] Here we present a
more detailed kinetic study and a comparison of different
catalysts, involving perfluoro-tagged diphosphane com-
plexes and different support materials.
The propylene-bridged diphosphane 7c was prepared ac-
cording to
a modified procedure of Curran et al.
Results and Discussion
While in our initial investigations of perfluoro-tagged Pd
catalysts on fluorous silica gel we employed the yield of
(Scheme 2).[16] Thus, aryl bromide 6 was converted into the
corresponding aryllithium derivative, which was treated
with freshly distilled Et2NPCl2. The phosphoramidite inter-
mediate was hydrolysed with concentrated aqueous hydro-
gen chloride to give diarylphosphane oxide 8 in 67% yield.
Alkylation of diarylphosphane oxide 8 with 1,3-dibro-
mopropane, and subsequent reduction with trichlorosilane
[a] Institut für Organische Chemie und Biochemie, Albert-Lud-
wigs-Universität Freiburg,
Albertstraße 21, 79104 Freiburg, Germany
Fax: +49-761-203-8705
E-mail: willi.bannwarth@organik.chemie.uni-freiburg.de
5248
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2005, 5248–5261

Pd-Catalysts for Suzuki Reactions
FULL PAPER
Figure 1. Palladium complexes employed in this study.
Scheme 1. Synthesis of bidentate phosphanes 7a and 7b.
Scheme 3. Synthesis of phosphane 10.
and triethylamine, provided the perfluoro-tagged propyl-
ene-bridged diphosphane 7c in an isolated yield of 55%.
The synthesis of the perfluoro-tagged phosphane 10 was
carried out with a yield of 58% as described by Genêt
(Scheme 3).[15]
Palladium complexes 4a–c of bidentate ligands were pre-
pared by addition of
a solution of Na2PdCl4 or
[PdCl2(CH3CN)2] to a solution of bidentate ligand in THF
at room temperature (Scheme 4).
The perfluoro-tagged palladacyclic complexes 3a and 3b
were prepared according to the procedure of Hermann
(Scheme 5).[17]
In our initial experiments we investigated the Suzuki
coupling of 4-nitrobromobenzene and phenylboronic acid
with different Pd complexes on FSG as catalyst (Scheme 6).
Scheme 4. Synthesis of palladium complexes 4a–c.
Scheme 2. Synthesis of bidentate phosphane 7c.
Eur. J. Org. Chem. 2005, 5248–5261
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5249

C. C. Tzschucke, V. Andrushko, W. Bannwarth
FULL PAPER
(ln[S]0 – ln[S])·Vrxn/nPd = kapp·t
(2)
where [S]0 is the initial concentration of the substrate. Ac-
cording to Equation (2), the apparent rate constant, kapp
,
can be obtained from a logarithmic plot of [S] vs. t.
To carry out the experiments, we employed a Chemspeed
Automated Synthesis Workstation, which allowed us to
carry out reactions in parallel. The reactions were run on a
0.3-mmol scale in a reaction volume of 3 mL. Aliquots of
the reaction mixture were withdrawn by the robotic syn-
thesiser and subsequently analysed by HPLC.
Three consecutive runs were performed automatically
with the same catalyst for every experiment. Palladium
complexes 1a–e and 3a of perfluoro-tagged monodentate
phosphanes and complexes 4a–c of bidentate phosphanes
on FSG were used with a catalyst loading of 0.1 mol-%.
Additionally, untagged complex 2 on FSG as well as on
unmodified silica gel (SG) and complexes 1a and 4a–c on
SG were used for comparison. Different from the previous
experiments, potassium carbonate was used as a base in
these experiments and the amounts of base and water were
reduced in order to avoid the formation of two distinct li-
quid phases in the reaction mixture, which would add fur-
ther complexity to the reaction system. In contrast to ear-
lier experiments with the very reactive 4-bromonitroben-
zene, differences in the reactivity of the pre-catalysts in
terms of conversion and rate of reaction, as illustrated in
Table 1, became recognizable with 4-bromobenzyl alcohol
as substrate.
Scheme 5. Synthesis of palladacyclic complex 3a and 3b.
Scheme 6. Suzuki coupling of 4-bromonitrobenzene and phenylbo-
ronic acid.
However, this substrate combination proved not to be
suitable for two reasons: the reaction is very fast, and hence
precise sampling is difficult to perform, and secondly, reli-
able automated workup cannot be ensured due to the ten-
dency of the product to crystallise. Nevertheless, these early
experiments showed that the rate of reaction is proportional
to the catalyst loading, as one would expect.
Therefore, we switched to a less-active substrate combi-
nation of 4-bromobenzyl alcohol and phenylboronic acid
(Scheme 7). The former is almost two orders of magnitude
less reactive than 4-bromonitrobenzene as starting material.
Furthermore, the product of the reaction, 4-phenylbenzyl
alcohol is better soluble, thus allowing for reliable auto-
mated workup. As a measure of the activity of the catalyst,
Table 1. Comparison of different supported catalysts, 4-bromoben-
zyl alcohol (0.3 mmol), phenylboronic acid (0.33 mmol), 0.1 mol-
% Pd, K2CO3 (0.6 mmol), DME, 80 °C, 16 h.
[a]
Entry Complex Support
Conversion[a] Cumulative kapp
[%]
TON[b]
[Lmol–1 min–1
]
1
2
3
4
5
6
7
8
2
2
SG
FSG
SG
FSG
FSG
FSG
FSG
FSG
SG
FSG
SG
FSG
SG
FSG
SG
FSG
63 (16, 10)
82 (18, 4)
61 (19, 9)
71 (36, 13)
44 (32, 27)
40 (26, 8)
36 (4, 3)
70 (31, 7)
62 (54, 38)
67 (37, 19)
66 (60, 35)
66 (64, 50)
53 (13, 5)
75 (55, 49)
62 (48, 27)
70 (35, 26)
893
1042
886
1205
1029
747
90 (1, 0)
313 (7, 0)
67 (6, 1)
111 (10, 1)
35 (6, 1)
35 (1, 0)
39 (0, 0)
76 (3, 1)
18 (14, 3)
24 (9, 3)
39 (18, 8)
27 (23, 15)
25 (2, 1)
42 (15, 8)
22 (13, 6)
41 (10, 7)
1a
1a
1b
1c
1d
1e
3a
3a
4a
4a
4b
4b
4c
4c
we chose the apparent rate constant, kapp
.
422
1073
1545
1226
1607
1809
700
1786
1376
1310
9
10
11
12
13
14
15
16
Scheme 7. Suzuki coupling of 4-bromobenzyl alcohol and phen-
ylboronic acid.
The reaction follows a first-order rate law of the form
d[S]/dt = –kapp·[S]·nPd/Vrxn
(1)
[a] Values for recycling in brackets. [b] (mol product)/(mol Pd com-
plex).
where [S] is the substrate concentration at a given time, nPd
is the molar amount of pre-catalyst and Vrxn is the reaction
When comparing the FSG-supported perfluoro-tagged
volume. The ratio nPd/Vrxn is used instead of the palladium complexes, the highest cumulative turnover number of ap-
concentration because the latter is not known. Thus, nPd prox. 1800 was achieved with 4a and 4b as pre-catalysts,
/
Vrxn is a substitute, which equals [Pd] only in the case that while the other perfluoro-tagged complexes gave TONs in
all Pd precatalyst is indeed dissolved. Equation (1) can be the range of 1000–1300; only 1c and 1d yielded lower
transformed by integration into
TONs. This order of reactivity was not reflected by the re-
5250
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2005, 5248–5261

Pd-Catalysts for Suzuki Reactions
FULL PAPER
spective rates of reaction. In the initial run, the highest rate all complexes, the rates of reaction as well as the final con-
constants, kapp, were observed for complexes 1a and 1e; the version decrease more or less sharply when the complexes
rate constants for the other catalysts were smaller by are reused. The probable causes for this loss of catalytic
roughly a factor 3. The differences in the yields arise less activity could be catalyst deactivation or leaching of the
from differences in the rate of reaction but rather from the complex. We had already observed in our earlier work that
degree to which activity is lost upon recycling. While all the activity of the catalyst decreases to a larger degree than
complexes exhibit a decrease of kapp, this is less marked for expected with less-reactive substrates.[9] This effect could be
4a and 4b; in the other cases kapp drops by a factor of 5– reduced by the addition of additional ligand, which is a
20 upon reuse of the supported catalyst. As one might ex- hint that decomposition might occur at the stage of the Pd0
pect, the loss of activity observed for the untagged complex intermediate. Extensive leaching, on the other hand, seems
2 is even more pronounced (Figure 2).
to be an unlikely reason for the decrease of activity, since,
These results are certainly not as clear-cut as we would based on the palladium content, we found less than 2% of
have desired, but nevertheless it is possible to sketch a gene- the total catalyst amount in the product.[9]
ral picture of these catalyst systems, which partly revises
We briefly compared four different organic solvents and
our previous conclusions.[9] Of the pre-catalysts examined, bases using 0.1 mol-% 1a on FSG as pre-catalyst. The re-
only the ones with bidentate phosphane ligands, namely 4a, sults did not deviate notably from the ones reported above.
show a limited reusability in DME as organic solvent. For Of the solvents used, we obtained the highest TONs with
Figure 2. Comparison of 0.1 mol-% of different supported complexes in the Suzuki coupling of 4-bromobenzyl alcohol and phenylboronic
acid in DME. a: conversion, b: rate constant.
Eur. J. Org. Chem. 2005, 5248–5261
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5251

C. C. Tzschucke, V. Andrushko, W. Bannwarth
FULL PAPER
ethanol; toluene and DME gave slightly lower conversions, dropped significantly upon recycling of the complex
while in acetonitrile we observed only low conversion. (Scheme 8).
Cs2CO3, CsF, NaOAc and LiOH were employed as bases.
We also turned our attention to Suzuki couplings of po-
Except for NaOAc, which resulted in low activity, all bases lar substrates in water as the solvent, since we felt that an
were equally well suited. Generally, the rate of reaction increase in the solvent polarity might influence the catalyst
Table 2. Comparison of different supported catalysts, 4-bromomandelic acid (0.3 mmol), phenylboronic acid (0.33 mmol), 0.1 mol-% Pd,
Na2CO3 (0.6 mmol), water, 80 °C, 16 h. Values are averages of at least two independent experiments.
[a]
Entry Complex
Pd
[mol-%]
Support
Conversion[a]
[%]
Cumulative
TON[b]
kapp
[Lmol–1 min–1]
1
2
3
4
5
6
7
8
1a
1a
1a
1a
1b
1c
1e
1a
2
0.001
0.01
0.1
1
0.1
0.1
0.1
0.1
0.1
FSG
FSG
FSG
FSG
FSG
FSG
FSG
SG
FSG
SG
FSG
FSG
FSG
SG
FSG
SG
FSG
FSG
FSG
SG
FSG
SG
FSG
SG
FSG
SG
FSG
FSG
FSG
SG
SG
SG
FSG
FSG
FSG
SG
27 (71, 54, 37, 22, 7)
4 (76, 84, 85, 85, 69)
99 (91, 88, 79, 80, 76)
98 (98, 95, 90, 83, 87)
88 (62, 83, 84, 86, 82)
94 (76, 83, 82, 74, 73)
85 (90, 89, 88, 85, 75)
88 (89, 87, 86, 84, 82)
90 (91, 90, 86, 85, 72)
52 (95, 95, 91, 85, 77)
1 (8, 6, 8, 8, 2)
1 (45, 38, 39, 21, 17)
58 (93, 93, 85, 90, 88)
77 (94, 87, 85, 86, 75)
69 (97, 96, 96, 97, 88)
56 (86, 98, 97, 94, 85)
24 (96, 95, 97, 99, 86)
60 (98, 98, 96, 99, 96)
4 (41, 82)
217388
40382
5130
551
4846
4820
5120
5170
5135
121 (1341, 1315, 921, 345, 68)
6 (416, 422, 424, 373, 141)
288 (37, 33, 31, 31, 22)
51 (14, 8, 4, 4, 3)
108 (23, 23, 24, 29, 33)
112 (32, 24, 23, 25, 23)
44 (50, 42, 38, 34, 22)
50 (46, 33, 27, 31, 34)
75 (58, 47, 38, 31, 21)
53 (68, 55, 38, 32, 33)
13 (130, 87, 153, 155, 20)
1 (72, 87, 81, 44, 36)
69 (67, 71, 72, 71, 70)
21 (90, 52, 62, 44, 42)
17 (106, 89, 123, 67, 63)
11 (95, 105, 92, 64, 48)
5 (90, 69, 82, 60, 51)
1 (7, 8, 11, 8, 10)
7 (285, 471)
535 (586, 752)
12 (452, 482)
568 (332, 270)
885 (418, 459)
688 (403, 315)
47 (1, 1)
9
10
2
0.1
4953
11[c]
12[c]
13[c]
14[c]
15[c]
16[c]
17[c]
18[c]
19
1a
1a
1a
1a
2
0.001
0.01
0.1
0.1
0.1
0.1
0.1
1
32317
16114
5071
5048
5431
5297
4970
547
2
1e
1a
1a
1a
2
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.1
0.01
0.001
0.1
0.01
0.001
0.1
0.01
0.001
0.1
0.01
0.001
0.01
0.001
0.01
0.001
0.1
0.01
0.001
0.1
0.01
0.001
0.1
12799[d]
26939[d]
18072[d]
24150[d]
26715[d]
15376[d]
2031[d]
3169[d]
5263
44669
483997
5541
54034
169058
5242
50966
497535
5395
51362
266682
23919
36278
21713
114459
5593
48936
389054
5335
53687
341896
5775
47059
234607
5690
53236
525865
20
21
22
85 (90, 94)
10 (86, 85)
84 (84, 74)
94 (88, 85)
61 (47, 46)
17 (2, 1)
24 (6, 2)
2
23[e]
24[e]
25[e]
26[e]
27[e]
28[e]
29[e]
30[e]
31[e]
32[e]
33[e]
34[e]
35[e]
36[e]
37[e]
38[e]
39[e]
40[e]
41[e]
42[e]
43[e]
44[e]
45[e]
46[e]
47[e]
48[e]
49[e]
50[e]
51[e]
52[e]
53[e]
54[e]
1a
1a
2
2
30 (8, 2)
4a
4a
4a
4a
4a
4a
4b
4b
4b
4b
4b
4b
5
93 (92, 92, 82, 86, 82)
55 (89, 85, 84, 87, 84)
82 (85, 88, 81, 81, 68)
95 (94, 92, 92, 91, 90)
96 (98, 92, 86, 84, 83)
32 (41, 29, 22, 26, 19)
89 (92, 90, 82, 86, 85)
77 (90, 88, 84, 86, 84)
86 (83, 91, 91, 84, 63)
92 (96, 91, 89, 87, 86)
85 (98, 95, 84, 77, 75)
49 (32, 36, 65, 27, 58)
32 (45, 41, 45, 32, 45)
1 (3, 4, 4, 19, 4)
48 (45, 43, 35, 14, 33)
37 (6, 3, 32, 3, 34)
98 (96, 92, 91, 96, 87)
96 (79, 77, 79, 79, 79)
89 (83, 70, 63, 50, 33)
98 (92, 88, 87, 85, 83)
98 (97, 89, 89, 88, 76)
82 (68, 65, 44, 48, 34)
98 (98, 97, 97, 94, 93)
43 (87, 85, 85, 81, 89)
53 (53, 36, 36, 28, 28)
99 (98, 96, 93, 91, 92)
92 (95, 93, 84, 83, 85)
86 (88, 93, 94, 78, 87)
108 (58, 56, 31, 30, 26)
81 (588, 428, 188, 495, 227)
5093 (4546, 4507, 2446, 3963, 1810)
57 (45, 40, 36, 31, 37)
1613 (1079, 561, 421, 378, 347)
1 (1243, 533, 214, 388, 213)
62 (59, 47, 30, 28, 30)
696 (492, 392, 176, 371, 260)
4664 (3137, 3170, 3254, 1926, 702)
62 (59, 47, 30, 28, 30)
696 (492, 392, 176, 371, 260)
4664 (3137, 3170, 3254, 1926, 702)
136 (181, 54, 171, 118, 180)
28 (50, 80, 21, 42, 48)
154 (199, 25, 107, 9, 104)
74 (91, 24, 1365, 10, 77)
52 (68, 49, 34, 43, 26)
485 (356, 293, 192, 271, 259)
5678 (2721, 2636, 1722, 1108, 611)
52 (49, 38, 10, 28, 34)
1307 (702, 407, 417, 371, 288)
3026 (1628, 1906, 1402, 1082, 462)
136 (97, 73, 44, 44, 62)
82 4(89, 401, 433, 312, 397)
2644 (2020, 863, 720, 498, 404)
85 (110, 73, 41, 38, 41)
807 (820, 510, 396, 356, 408)
1381 (3839, 4177, 5349, 3221, 4369)
SG
SG
FSG
FSG
SG
5
5
5
SG
4c
4c
4c
4c
4c
4c
3a
3a
3a
3a
3a
3a
FSG
FSG
FSG
SG
SG
SG
FSG
FSG
FSG
SG
SG
SG
0.01
0.001
0.1
0.01
0.001
[a] Values for recycling in brackets. [b] (mol product)/(mol Pd complex). [c] Filtration experiment. [d] Cumulative TON over three runs.
[e] Pre-wetting of support.
5252
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2005, 5248–5261

Pd-Catalysts for Suzuki Reactions
FULL PAPER
tude, which is the result of the rate of reaction being largely
independent of the initial pre-catalyst loading. Since it is
safe to assume that the same catalytic species is formed re-
gardless of the amount of pre-catalyst, the most obvious
interpretation of the results is that, independent of the
amount of pre-catalyst, approximately the same amount of
catalytic species is formed upon recycling. Although this is
surprising at first sight, a comparable phenomenon has
been reported recently in the case of a so-called ligand-free
Heck reaction.[18]
We also compared the perfluoro-tagged complexes 1a,
1b, 1c and 1e, and the untagged complex 2 at a loading
of 0.1 mol-% on FSG. The conversions exhibit an almost
random variation within a range of 85–100% in the first
run, decreasing to 70–80% in the sixth run (Figure 4, a).
While the rate constants, kapp, show a spread of nearly one
order of magnitude in the first run, the following order of
activity can be observed in the second, third and fourth
runs: 2 Ϸ 1e Ͼ 1a Ͼ 1c Ϸ 1b, although the differences
are small considering the variation of the results of single
experiments (Figure 4, b). In the fifth and sixth runs the
differences in reactivity are no longer significant.
Scheme 8. Suzuki coupling of bromomandelic acid and phenylbo-
ronic acid.
reactivity, reusability and leaching. Indeed, we observed a
markedly different behaviour of the catalyst in initial experi-
ments. With 4-bromomandelic acid as a water-soluble sub-
strate with a reactivity similar to 4-bromobenzyl alcohol,
high cross-coupling yields were obtained even after the fifth
recycling of the catalyst. The results for different catalyst
loadings and different catalysts, as well as for a number of
substrates, were communicated recently (Table 2).[10]
In addition, we measured the rates of reactions for these
examples. Upon comparing the rate constants, kapp, for the
coupling of 4-bromomandelic acid and phenylboronic acid
in water with catalyst loadings between 0.001 and 1.0 mol-
% of 1a we observed a striking dependence of kapp on the
catalyst loading (Figure 3, b). For 0.1 mol-% 1a, kapp starts
out at a value significantly higher than the analogous value
for the reaction in DME and remains on a fairly constant
lower level for five reuses of the catalyst. Similarly, for
1.0 mol-% 1a, kapp starts at a high value, which slowly de-
creases until it levels-off in the fourth run. In contrast to
this, the reaction with 0.01 mol-% 1a is unexpectedly slow
in the first run, but upon reuse of the catalyst kapp jumped
Similarly, complexes 1a and 2 were compared on FSG
and unmodified SG as support. Again, in terms of conver-
sion a difference could be hardly detected. In terms of the
rate of reaction, however, 2 is slightly more active than 1a
upon recycling, and no influence of the support is obvious
(Figure 5).
For this set of pre-catalysts (complexes 1a and 2 on FSG
to a high level, where it remains until the fifth run. A sim- and unmodified SG) we also carried out a filtration experi-
ilar behaviour was observed for 0.001 mol-% 1a: the reac- ment, where we filtered the hot reaction mixture and moni-
tion is slow in the first run but a high value of kapp is ob- tored the conversion in the filtrate (Figure 6). In the first
served in the second and third runs and decreases in the run, the conversions and the rates of reaction were signifi-
following reuses. The most prominent observation is that cantly lower than in the analogous experiments without fil-
the levels of reactivity differ by roughly one order of magni- tration. In the following runs, however, the conversions
Figure 3. Suzuki coupling of 4-bromomandelic acid and phenylboronic acid in water with different amounts of complex 1a on FSG. a:
conversion, b: rate constant.
Eur. J. Org. Chem. 2005, 5248–5261
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5253

C. C. Tzschucke, V. Andrushko, W. Bannwarth
FULL PAPER
Figure 4. Comparison of different supported complexes in the coupling of 4-bromomandelic acid and phenylboronic acid in water with
0.1 mol-% Pd. a: conversion, b: rate constant.
Figure 5. Comparison of perfluoro-tagged and untagged complexes on FSG and SG in the coupling of 4-bromomandelic acid and
phenylboronic acid in water with 0.1 mol-% Pd. a: conversion, b: rate constant.
were in the usual range again, and the rate constants were This should lead to an activation period or at least to non-
even higher than in the analogous series. No significant sys- first-order kinetics in the first run. In the event, we did not
tematic influence of either the fluorous tag or the support observe any activation period,[19] and the conversions
material is apparent. This result is strongly in favour of a closely followed first-order kinetics.
homogeneously dissolved catalytic species, and is in accord-
ance with the already reported results of a three-phase support might be the reason for lower conversions and rates
test.[10]
of reaction in the first run, a further series of experiments
Since we suspected that insufficient wetting of the solid
A puzzling result, however, is the observation that the was performed with pre-wetting of the supported complex.
reactions are slower with fresh complex than with reused Thus, the support was shaken with a small amount of meth-
catalyst, as had already been observed for smaller catalyst anol and then filtered prior to the actual first cross-coup-
loadings of 0.01 and 0.001 mol-%. The most obvious expla- ling reaction. We speculated that this procedure would re-
nation would be a slow catalyst-activation step, which gen- move the air from the pores of the support and cover it with
erates the catalytically active species from the Pd complex. a film of methanol, which should facilitate the subsequent
5254
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2005, 5248–5261

Pd-Catalysts for Suzuki Reactions
FULL PAPER
Figure 6. Filtration experiment with perfluoro-tagged and untagged complexes on FSG and SG in the coupling of 4-bromomandelic acid
and phenylboronic acid in water with 0.1 mol-% Pd. a: conversion, b: rate constant.
wetting of the support by the aqueous reaction mixture.
On the basis of these results, we carried out several ex-
Again, complexes 1a and 2 on FSG and SG were used as periments using bidentate perfluoro-tagged complexes 4a–c
catalyst sources for these comparative experiments, but this as pre-catalysts. All three complexes were tested with Pd
time with an initial loading of 0.01 mol-% (Figure 7). In the loadings of 0.001, 0.01 and 0.1 mol-% on both FSG and
control experiments without pre-wetting, both catalysts on SG using the pre-wetting procedure. A few experiments
FSG gave low conversions in the first runs, while with SG with the corresponding unmodified ligands were carried out
as support the conversion was above 80% already in the for comparison.
initial run. In the following runs the yields were in the range
For complex 4a, high conversions were generally ob-
between 75 and 95% and no influence of the support or tained in the range of 80 to 100%, the only exception being
pre-catalyst was apparent. The same picture also arises the lowest catalyst loading on SG. The cumulative TON
from the rates of reaction. The complexes on FSG were 50 over six runs for 0.001 mol-% 4a on FSG was 484000,
times slower than the complexes on SG during the first run. which is 2.5 times the TON for the monodentate complex
In the second and third runs, however, the rate constants 1a (Figure 8). With regard to the rate constants, the same
were similar for all four catalyst support combinations. In phenomenon as for complex 1a was observed: the rate con-
the pre-wetting experiment the reactivities were completely stants decreased only slightly upon recycling, and decreas-
altered. The conversion with the perfluoro-tagged complex ing the catalyst loading by one order of magnitude in-
1a on FSG was already high in the first run and decreased creased kapp by approximately one order of magnitude. Ex-
slightly to 85% in the third run. The conversions for the cept for the lowest catalyst loading on SG, which gave a
same complex on SG were considerably lower and remained lower rate of reaction, there is no influence of the support
in the range between 50 and 60%. The unmodified complex material. The rates are slightly higher than with the mono-
2 on either support led only to low conversions of between dentate complex 1a. The only pronounced difference was
0 and 20%. Also, in terms of the rate of reaction the influ- seen for the lowest catalyst loading of 0.001 mol-% 4a,
ence of the perfluoro tag is clearly visible. The rate constant where kapp was larger by a factor of 4 and remained at a
in the first run for the perfluoro-tagged complex 1a is 20 high level through all six runs, in contrast to 1a. Due to the
times larger than for complex 2, independent of the sup- pre-wetting, the results were more regular for the initial run,
port. This difference increased further upon recycling of the with the only exception being that for 0.01 mol-% 4a on
catalyst. Clearly, the pre-wetting procedure is beneficial FSG, where a lower activity was initially observed.
only for the perfluoro-tagged complex on FSG, but is detri-
Very similar results were obtained for complex 4b (Fig-
mental for the unmodified complex. A possible explanation ure 9): the yields remained at a high level of 75–95% over
for this observation could be that wetting of the FSG with six runs, again with the exception of the lowest catalyst
water is indeed facilitated by washing it with methanol, loading on SG. The rates of reaction were similar to the
whereas the unmodified complex is largely washed off the ones observed for 4a. Although less clear-cut, the same
support, thus leading to low catalytic activities.
pattern emerges, with lower loadings giving higher values of
Eur. J. Org. Chem. 2005, 5248–5261
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5255

C. C. Tzschucke, V. Andrushko, W. Bannwarth
FULL PAPER
Figure 7. Influence of pre-wetting of the supported complex on the coupling of 4-bromomandelic acid and phenylboronic acid in water
with 0.1 mol-% Pd. Comparison of perfluoro-tagged and untagged complexes on FSG and SG. a: conversion without pre-wetting; b:
conversion with pre-wetting with MeOH; c: rate constant without pre-wetting; d: rate constant with pre-wetting with MeOH.
Figure 8. Suzuki coupling of 4-bromomandelic acid and phenylboronic acid in water with different amounts of complex 4a on FSG and
SG with pre-wetting of the support with MeOH. a: conversion, b: rate constant.
kapp and no influence of the support material. Much lower 389000, almost twice as much as for the monodentate com-
conversions were obtained for the corresponding untagged plex 1a (Figure 10).
complex 5, and the rates of reactions were lower with a
In general, the reactions exhibit a considerably different
larger random variation. This, again, is in accordance with behaviour depending on whether water or DME is used as
our initial investigation of the influence of the pre-wetting solvent. In organic solvents there is a significant decrease of
with methanol.
activity upon recycling of the catalyst, while for very active
The results for complex 4c were similar to the ones ob- substrates, although the yields of cross-coupling product re-
tained with 4a and 4b. The conversions remained above mained high, the rates of reaction clearly decreased. For less
80%, and only with the lowest catalyst loading did the con- reactive substrates a drop in yield as well as rate of reaction
version decrease from approx. 85% in the first run to 40% was obvious when the catalyst was recycled. This could be
in the last run, which still led to a cumulated TON of a combined result of catalyst leaching and catalyst decom-
5256
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2005, 5248–5261

Pd-Catalysts for Suzuki Reactions
FULL PAPER
Figure 9. Suzuki coupling of 4-bromomandelic acid and phenylboronic acid in water with different amounts of complex 4b on FSG and
SG with pre-wetting of the support with MeOH. a: conversion, b: rate constant.
Figure 10. Suzuki coupling of 4-bromomandelic acid and phenylboronic acid in water with different amounts of complex 4c on FSG and
SG with pre-wetting of the support with MeOH. a: conversion, b: rate constant.
position, but since the Pd content of the products were influence of the support material (FSG or SG) and the na-
small, the latter process seems to be the most relevant one.
ture of the pre-catalyst is small. Only with 0.001 mol-% of
In water as solvent a completely different behaviour was complexes 4a and 4b were significantly higher rates of reac-
observed. The conversion to cross-coupling product re- tion and conversions observed in comparison to the com-
mained high over six runs, and the rate of reaction de- plexes with monodentate ligands. Filtration experiments
creased only slightly for catalyst loadings of 0.01 mol-% or and a three-phase test clearly indicated that the catalytic
above. Surprisingly, the rates of reaction showed only a reaction occurs homogeneously in solution and not hetero-
small dependence on the catalyst loading, which results in geneously on the surface of the support material. These re-
increasing observed rate constants with lower catalyst load- sults lead us to the assumption that the homogeneously dis-
ing. The conclusion from this observation is that the solved catalytic species no longer contains phosphane li-
amount of actual catalytically active species is almost inde- gands and is rather solvated by water molecules or anions
pendent of the initial pre-catalyst loading. Furthermore, the in the reaction mixture. Comparable ligand-free Pd cata-
Eur. J. Org. Chem. 2005, 5248–5261
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5257

C. C. Tzschucke, V. Andrushko, W. Bannwarth
FULL PAPER
lysts have been reported for Heck reactions with so called
homeopathic catalyst loadings.[18] Therefore no significant
difference between perfluoro-tagged complexes and un-
tagged complex 2 is observed. Likewise, the nature of the
support does not exert a detectable influence. We surmise
that different phosphane ligands only exert an influence on
the stability of the insoluble Pd reservoir. The process of
conversion of the initial complex into such a reservoir is
not well understood. However, the wetting of the support
seems to play an important role, at least for lower catalyst
loadings, since low activities are sometimes observed with
the very hydrophobic FSG as support. Pre-wetting with
methanol to remove air from the pores and cover the sur-
face with a hydrophilic film alleviates this problem for per-
fluoro-tagged complexes. A very interesting aspect of the
catalyst activation appears to be the influence of pre-wet-
ting with methanol on untagged complex 2. While 2 with-
out pre-wetting yields high conversions and reaction rates
upon reuse of the support, washing with methanol prior to
the first reaction largely removes the catalytic activity. This
latter observation can clearly be interpreted as the result
of washing the pre-catalyst off the support. However, since
Experimental Section
General: All reagents and solvents were obtained from either Fluka
or Aldrich and were used without further purification. 1-Bromo-4-
(3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptadecafluorodecyl)benzene
(6) and 1-bromo-4-(3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptade-
cafluorodecyl)-2-methylbenzene (10) were prepared according to
the literature.[15] Methylenebis(dichlorophosphane) (Cl2PCH2PCl2)
was prepared according to the literature.[20] (C6F13CH2CH2–)FSG
was prepared from SG (particle size: 100–300 µm; pore size: 500 Å;
specific surface: 70–90 m2 g–1), obtained from Grace, and
(C6F13CH2CH2)Si(OEt)3 as described earlier.[21] Melting points
were measured with an IA9000 apparatus from Electrothermal
Engineering Ltd. and are uncorrected. NMR Spectra: chemical
1
shifts δ are given in ppm relative to Si(CH3)4 (δ = 0 ppm) for H
NMR, CHCl3 (δ = 77 ppm) for 13C NMR and 85% aq. H3PO4 (δ
= 0 ppm) for 31P NMR. MS: Finnegan MAT8200 (EI), MAT312
(CI) and TSQ7000 (ESI) mass spectrometers. HPLC: Agilent 1100
system with binary pump, sample changer, column oven and diode
array detector.
1,2-Bis{bis[4-(3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptadecafluorode-
cyl)phenyl]phosphanyl}methane (7a): Under argon, a solution of n-
butyllithium (8.0 mL, 2.5  in hexane) was added dropwise to a
solution of 6 (12.12 g, 20.096 mmol) in anhydrous THF/Et2O (1:1,
methanol was also employed in the work-up in-between the 40 mL) keeping the bath temperature below –30 °C. The yellow-
green solution was stirred for 10 min, after which freshly distilled
Cl2PCH2PCl2 (1.09 g, 5.005 mmol) dissolved in 3 mL of anhydrous
THF was added dropwise at –30 °C and the resulting purple mix-
ture was stirred at room temperature for 3 h. The solution was fil-
tered through SiO2 under argon, eluting with anhydrous Et2O
(3×30 mL). The solvent was removed under reduced pressure to
afford a pale-yellow solid (10.4 g, 90%). The compound (contain-
ing around 7% of oxide) could be recrystallised from Et2O/MeOH
to give the product as a white powder (7.51 g, 65% yield) or puri-
fied by column chromatography [silica gel 0.063–0.2 mm, cyclohex-
ane/ethyl acetate (10:1), Rf = 0.32] under argon to give 7a (6.19 g,
reactions, it is obvious that the complex is converted into a
catalyst reservoir, which more strongly adheres to the sup-
port regardless of whether it is FSG or SG. However, on
the basis of these results we cannot distinguish whether the
catalyst is truly recycled or the support acts only as a Pd
reservoir.
1
Conclusions
57%). M.p. 89 °C. H NMR (300 MHz, C6F6/C6D6 capillary): δ =
2
2.8 (t, JP,H = 1.5 Hz, 2 H, CH2-P), 2.28–2.42 (m, 8 H, CH2-Ar),
2.89–2.96 (m, 8 H, CH2-CF2), 7.2–7.38 (m, 16 H, Ar) ppm. 13C
We have investigated a set of different perfluoro-tagged
Pd complexes with mono- and bidentate ligands on FSG or
SG as support as reusable catalysts for the Suzuki cross-
coupling reaction in organic solvents and water. In organic
solvents the activity decreases upon recycling, therefore the
catalyst is clearly not reusable under these conditions, which
we attribute to catalyst decomposition rather than catalyst
leaching. In water, the behaviour is completely different: sig-
nificantly higher cumulative TONs are observed, and the
activity remains at a high level upon reuse of the supported
complex. This clearly indicates, once again, that yields ob-
tained in repetitive cycles alone are not appropriate to judge
1
NMR [75 MHz, C6D6/C6F6, 1:1 (v/v)]: δ = 28.2 (t, JC,P = 24.2 Hz,
CH2-P), 32.6 (t, 2JC,F = 22.0 Hz, CH2CF2), 104.8–123.4 (m, C8F17),
3
2
128.4 (d, JC,P = 6.3 Hz, m-C, Ar), 133.4 (d, JC,P = 18.2 Hz, o-C,
1
Ar), 137.2 (d, JC,P = 12.3 Hz, C-P, Ar), 139.8 (s, p-C, Ar) ppm.
31P NMR (162 MHz, C6F6/C6D6 capillary): δ = –22.2 ppm. MS
(ESI-dir.): m/z (%) = 2168 (20) [M+], 2216 (35), 2215 (68), 1750
(44), 1749 (100), 1283 (21), 1167 (16), 596 (23).
1,2-Bis{bis[4-(3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptadecafluorode-
cyl)phenyl]phosphanyl}ethane (7b): Under argon, a solution of n-
butyllithium (8.0 mL, 2.5  in hexane) was added dropwise to a
solution of 6 (12.12 g, 20.096 mmol) in anhydrous THF/Et2O (1:1,
40 mL) keeping the bath temperature below –30 °C. The yellow-
the reusability of catalysts. Instead, it becomes obvious that green solution was stirred for 10 min, after which freshly distilled
Cl2PCH2CH2PCl2 (from ABCR; 1.16 g, 5.005 mmol) dissolved in
3 mL anhydrous THF was added dropwise at –30 °C and the re-
sulting purple mixture was stirred at room temperature for 3 h.
The solution was filtered through SiO2 under argon, eluting with
anhydrous Et2O (3×30 mL). The solvents were removed under re-
duced pressure to afford 10.4 g of a pale-yellow solid (10.7 g, 92%).
The product (containing around 9% of oxide) could be purified by
column chromatography [silica gel 0.063-0.2 mm, cyclohexane/ethyl
acetate (10:1), Rf = 0.34] under argon to give 7b (7.32 g, 67% yield).
the rates of reaction have also to be investigated in order
to assess the true potential of a catalytic system. Although
reusability of catalysts is frequently claimed, thorough in-
vestigations including rates of reaction are rarely published.
While in our case it still remains debatable whether the cata-
lytic species itself is truly recycled in the Suzuki reactions
in water, the high turnover numbers and the low Pd leach-
ing open interesting practical opportunities for Suzuki
couplings of polar substrates in water as the sole reaction
medium.
1
M.p. 82 °C. H NMR (300 MHz, C6F6/C6D6 capillary): δ = 2.1 (s,
4 H, CH2P), 2.3–2.5 (m, 8 H, CH2-Ar), 2.8–3.1 (m, 8 H, CH2-CF2),
5258
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2005, 5248–5261

Pd-Catalysts for Suzuki Reactions
FULL PAPER
CH2CH2CH2), 26,3 (s, CH2-Ar), 29.5 (t, JC,P = 14.3 Hz, P-CH2),
7.2–7.4 (m, 16 H, Ar) ppm. 13C NMR [75 MHz, C6D6/C6F6, 1:1
(v/v)]: δ = 26.3 (s, CH2-Ar), 27.4–29.2 (m, P-CH2), 32.6 (t, JC,F
1
3
2
=
=
32.7 (t, JC,F = 22.1 Hz, CH2CF2), 104.6–123.2 (m, C8F17), 128.6
3
3
2
22.1 Hz, CH2CF2), 104.5–123.3 (m, C8F17), 128.4 (d, JC,P
(d, JC,P = 6.4 Hz, m-C, Ar), 133.2 (d, JC,P = 18.6 Hz, o-C, Ar),
2
1
6.4 Hz, m-C, Ar), 133.3 (d, JC,P = 18.6 Hz, o-C, Ar), 137.2 (d,
1JC,P = 12.1 Hz, I-P, Ar), 140.1 (s, p-C, Ar) ppm. 31P NMR
(162 MHz, C6F6/C6D6 capillary): δ = –13.8 ppm. MS (ESI-dir.):
m/z (%) = 2182 (32) [M+], 2230 (25), 2229 (58), 1954 (14), 1876
(12), 1749 (44), 1748 (100), 1283 (11), 1187 (16), 596 (19).
137.1 (d, JC,P = 12.1 Hz, C-P, Ar), 139.9 (s, p-C, Ar) ppm. 31P
NMR (162 MHz, C6F6/C6D6 capillary): δ = –18.1 ppm.
Tris[4-(3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptadecafluorodecyl)-2-
methyl]phosphane (10): A solution of n-butyllithium (4.0 mL, 2.5 
in hexane) under argon was added dropwise to a solution of 9
(6.17 g, 10 mmol) in anhydrous THF/Et2O (1:1, 40 mL) keeping
the bath temperature below –30 °C. The yellow-green solution was
stirred for 15 min. Then, freshly distilled PCl3 (456 mg, 3.3 mmol)
in 10 mL of THF was added dropwise. The reaction was monitored
by 31P NMR spectroscopy. The solution was filtered through silica
gel under argon, eluting with anhydrous Et2O (3×30 mL). The sol-
vents were removed under reduced pressure and the product, which
contained approx. 10 % of oxide, was purified by column
chromatography [silica gel 0.063–0.2 mm, cyclohexane/ethyl acetate
(10:1), Rf = 0.36] under argon to give the title compound (3.14 g,
58% yield). M.p. 76 °C. 1H NMR (400 MHz, CDCl3): δ = 2.32–
2.52 (m, 6 H, CH2-Si), 2.5 (s, 9 H, CH3), 2.88–2.96 (m, CH2-CF2),
6.7 (dd, J = 7.8, 4.2 Hz, 3 H, H5), 7.0 (d, J = 7.8 Hz, 3 H, H6), 7.1
(d, J = 3.4 Hz, 3 H, H3) ppm. 13C NMR (75 MHz, CDCl3): δ =
Bis[4-(3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptadecafluorodecyl)phen-
yl]phosphane Oxide (8): A solution of n-butyllithium (4.0 mL, 2.5 
in hexane) was added dropwise, under argon, to a solution of 6
(6.06 g, 10.048 mmol), cooled to –40 °C, in anhydrous Et2O
(70 mL) keeping the bath temperature below –35 °C. The yellow-
green solution was stirred for 10 min. Then, freshly distilled
Et2NPCl2 (871 mg, 5.005 mmol) dissolved in 10 mL of anhydrous
Et2O was added slowly dropwise at –35 °C. The resulting purple
mixture was warmed to 0 °C over 30 min and was further stirred
at room temperature overnight. The reaction was monitored by 31
P
NMR spectroscopy (δ = –4.2 ppm). Concentrated aqueous hydro-
gen chloride (3.0 mL, 36 mmol) was then added and the mixture
was stirred at room temperature for an additional 3 h. Water
(50 mL) was added, the ethereal layer was separated and the aque-
ous layer was further extracted with diethyl ether (3×50 ml). The
combined diethyl ether layers were washed with brine and dried
with MgSO4. Diethyl ether was removed under reduced pressure
and the residue was purified by column chromatography (silica gel
0.063-0.2 mm, cyclohexane/ethyl acetate, 10:1), to give the product
(3.64 g, 66.5%). 1H NMR (300 MHz, C6F6/C6D6 capillary): δ =
2.28–2.49 (m, 8 H, CH2-Ar), 2.78–3.22 (m, 8 H, CH2-CF2), 7.4 (dd,
J = 8.2 Hz, 2.1 Hz, 4 H), 7.7 (dd, J = 13.6 Hz, 8.2 Hz, 4 H), 8.12
3
2
21.1 (d, JC,P = 21.2 Hz, CH3), 26.1 (s, CH2-Ar), 32.7 (t, JC,F
=
22.1 Hz, CH2CF2), 125.9 (s, C5, Ar), 130.2 (s, C3, Ar), 132.6 (d,
1JC,P = 10.6 Hz, C1, Ar), 133.4 (s, C6), 139.6 (s, C4, Ar), 143.2 (d,
2JC,P = 26.6 Hz, C2, Ar) ppm. 31P NMR (162 MHz, CDCl3): δ =
–30.44 ppm.
General Procedure for Preparation of PdCl2 Complexes 4a–c of Bi-
dentate Ligands. Method A: A solution of Na2PdCl4 (264.77 mg,
0.90 mmol) in 50 mL of MeOH was added dropwise, at room tem-
perature, to a solution of perfluoro-tagged bidentate phosphane
(0.922 mmol) in 200 mL of THF. The reaction mixture was then
stirred at room temperature for a further 12 h. The precipitated
yellow Pd complex was separated by filtration and washed with
distilled water (2 × 20 mL), MeOH (2 × 20 mL) and Et2O
(2×10 mL).
1
(d, JP,H = 483.1 Hz, 1 H, P-H) ppm. 31P NMR (162 MHz, C6F6/
C6D6 capillary): δ = 22.8 ppm.
1,2-Bis{bis[4-(3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptadecafluorode-
cyl)phenyl]phosphanyl}propane (7c): Sodium hydride (0.133 g, 60%
suspension in oil, 3.33 mmol) was added at room temperature, un-
der argon, to a solution of 8 (3.9 g, 3.33 mmol) in 20 mL of dry
THF. The resulting mixture was stirred at room temperature for
30 min. A solution of 1,3-dibromopropane (336.1 mg, 1.665 mmol)
in THF (2 mL) was then added slowly. After stirring at room tem-
perature for 4 h, water (2 mL) was added. The THF was partially
removed in vacuo and the residue was extracted with benzotrifluo-
ride (BTF; 3×10 mL). The combined BTF layers were dried with
MgSO4, the solvent was removed under reduced pressure and the
residue was dried under vacuum overnight. [31P NMR (162 MHz,
C6F6/C6D6 capillary): δ = 21.36 ppm]. The residue was then redis-
solved in BTF (50 mL) and trichlorosilane (2.4 mL) and triethyl-
amine (3.5 mL) were added under argon. The mixture was refluxed
for 8 h, and the reaction was monitored by 31P NMR spectroscopy.
The reaction mixture was then cooled to room temperature and
50 mL of 30% sodium hydroxide aqueous solution (previously de-
gassed with argon) was added. The resulting mixture was stirred at
room temperature under argon until both layers were clear. The
organic layer was separated and the aqueous layer was further ex-
tracted with BTF (3×30 mL). The combined organic layers were
dried with MgSO4, the solvent was removed under reduced pres-
sure, and the residue was dried under vacuum overnight and puri-
fied by column chromatography (silica gel 0.063–0.2 mm, cyclohex-
ane/ethyl acetate, 10:1), to give 7c (2.01 g, 55%) as a white solid.
Method B: A solution of [PdCl2(CH3CN)2] (233.47 mg, 0.90 mmol)
in 50 mL of CH2Cl2 was added dropwise, at room temperature, to
a solution of perfluoro-tagged bidentate phosphane (0.922 mmol)
in 200 mL of THF. The reaction mixture was then stirred at room
temperature for a further 12 h. The precipitated yellowish Pd com-
plex was separated by filtration and washed with CH2Cl2
(2×20 mL) and Et2O (2×10 mL).
[{Bis[P,PЈ-{4-(3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptadeca-
fluorodecyl)phenyl}phosphanyl]methane-P,PЈ}dichloropalladium(II)]
(4a): Yellow solid. Yield: 1.94 g, (92%). 1H NMR (400 MHz, C6F6/
C6D6 capillary): δ = 2.3–2.5 (m, 8 H, CH2-Ar), 2.9–3.2 (m, 8 H,
2
CH2-CF2), 4.3 (t, JP,H = 11.2 Hz, 2 H, P-CH2-P), 7.1–8.2 (m, 16
H, Ar) ppm. 31P NMR (162 MHz, C6F6/C6D6 capillary): δ =
–54.2 ppm.* MS (ESI-dir.): m/z (%) = 2343 (9) [M]+, 2325 (9) [M –
F + H]+, 2307 (21) [M – Cl]+, 2243 (15), 2242 (62), 2241 (100),
1617 (16), 700 (14), 699 (32), 671 (31), 663 (22). C65H34Cl2F68P2Pd
(2346.1): calcd. C 33.28, H 1.46, P 2.64; found C 33.26, H 1.57, P
2.48. * cf. 31P NMR chemical shift of the Pd complex of bis(di-
phenylphosphanyl)methane (δ = –53.7 ppm).[22]
[{Bis[P,PЈ-{4-(3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptadecafluoro-
decyl)phenyl}phosphanyl]ethane-P,PЈ}dichloropalladium(II)] (4b):
M.p. 78 °C. 1H NMR (300 MHz, C6F6/C6D6 capillary): δ = 1.64 Yellow solid. Yield: 1.91 g, (90%). 1H NMR (400 MHz, C6F6/C6D6
(m, 2 H, CH2CH2P), 2.23 (t, J = 7.5 Hz, 4 H, CH2CH2P), 2.38 (tt, capillary): δ = 2.2–2.5 (m, 8 H, CH2-Ar and 4 H P-CH2), 2.8–3.1
J = 18.2, 9.3 Hz, 8 H, CH2-Ar), 2.89–2.96 (m, 8 H, CH2-CF2), 7.18
(m, 8 H, CH2-CF2), 7.0–8.1 (m, 16 H, Ar) ppm. 31P NMR
(162 MHz, C6F6/C6D6 capillary): δ = 62.2 ppm. MS (ESI-dir.):
m/z (%) = 2357 (8) [M]+, 2341 (13) [MH – F + H]+, 2323 (91) [M –
(d, J = 7.5 Hz, 8 H, Ar), 7.33–7.39 (m, 8 H, Ar) ppm. 13C NMR
2
[75 MHz, C6D6/C6F6, 1:1 (v/v)]: δ = 22.3 (t, JC,P = 16.4 Hz,
Eur. J. Org. Chem. 2005, 5248–5261
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5259

C. C. Tzschucke, V. Andrushko, W. Bannwarth
Cl]+, 2215 (100) [M – Pd – Cl – H]+, 1749 (27), 1659 (22), 1617 The other complexes employed in this study were supported on
FULL PAPER
(34), 1179 (12), 1151 (13), 699 (65), 672 (25), 671 (55), 537 (45).
C66H36Cl2F68P2Pd (2360.2): calcd. C 33.59, H 1.54, P 2.62; found
C 33.43, H 1.60, P 2.29.
FSG or SG in a similar manner. Supported complexes were pre-
pared with palladium loadings of 0.03, 0.3, 3 and 30 µmolg–1.
Typical Procedure for Recycling Experiments in DME: A stock solu-
tion of the substrates in DME (3 mL, containing 0.3 mmol of 4-
bromobenzyl alcohol and 0.33 mmol of phenylboronic acid) and
an aqueous solution of potassium carbonate (0.5 mL, 0.6 mmol)
were added to the FSG-supported catalyst (100 mg) under argon.
The reaction mixture was shaken with a frequency of 900 min–1 at
80 °C for 16 h. Aliquots (20 µL) of the solution were withdrawn
after 15, 30, 45, 60, 75, 90, 105, 120, 240, 480 and 960 min, and
the conversion was measured by HPLC analysis. The reaction mix-
ture was then cooled to 20 °C and filtered. The FSG was washed
with DME (2×2 mL), water (2×2 mL) and DME (2×2 mL). The
immobilized catalyst was reused as such in further experiments. All
manipulations were carried out automatically (Automated Synthe-
sis Workstation ASW 2000, Chemspeed Ltd., Augst, Switzerland).
[{Bis[P,PЈ-{4-(3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptadecafluoro-
decyl)phenyl}phosphanyl]propane-P,PЈ}dichloropalladium(II)] (4c):
Yellow solid. Yield: 1.84 g, (86%). 1H NMR (400 MHz, C6F6/C6D6
capillary): δ = 1.8–2.6 [m, 6 H, P-(CH2)3 and 8 H, CH2-Ar], 2.89–
3.25 (m, 8 H, CH2-CF2), 7.12–8.07 (m, 16 H, Ar) ppm. 31P NMR
(162 MHz, C6F6/C6D6 capillary): δ = 11.4 ppm. MS (ESI-dir.):
m/z (%) = 2371 (7) [M]+, 2354 (12) [M – F + H]+, 2337 (41)
[M – Cl]+, 2310 (100), 1334 (38), 1333 (27), 1306 (19), 1300 (16),
1276 (26), 1228 (36), 1201 (47). C67H38Cl2F68P2Pd (2374.2): calcd.
C 33.90, H 1.61, P 2.61; found C 34.32, H 1.71, P 2.49.
P r e p a r a t i o n o f t r a n s - [ D i ( µ - a c e t a t o ) b i s { 4 - [ b i s -
{(3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-heptadecafluorodecyl)(2-
methyl)phenyl}phosphanyl]benzyl-P,C}dipalladium(II)] (3b): A solu-
tion of Pd(OAc)2 (134.7 mg, 0.600 mmol) in 20 mL of THF was
added dropwise to a solution of 10 (1.0 g, 0.609 mmol) in 20 mL
of THF and the reaction mixture was stirred for 12 h. The resulting
yellow-green solid was filtered and washed with distilled water
(2×10 mL), MeOH (2×10 mL) and Et2O (2×10 mL). The product
was dried in vacuo. Yield: 965 mg (89 %). 1H NMR (400 MHz,
CDCl3): δ = 1.98 (s, 6 H, CH3), 2.15 (br. s, 6 H, CH3), 2.32–2.52
(m, 6 H, CH2-Si and m, 2 H, CHaHb-Pd), 2.82–2.95 (m, 6 H, CH2-
CF2; m, 2 H, CHaHb-Pd; s, 6 H, CH3), 6.5–7.15 (br. m, 18 H, Ar)
ppm. 31P NMR (162 MHz, C6F6/C6D6 capillary): δ = 29.7 ppm.
MS (ESI-dir.): m/z (%) = 3995 (5), 3234 (5), 2570 (5), 1822 (8)
[1/2MH + O]+, 1821 (9) [1/2M + O]+, 1805 (4) [1/2M]+, 1722 (26),
1700 (10), 1669 (14), 1641 (25), 1640 (30), 1639 (13), 1623 (15),
1622 (45), 1618 (39), 1617 (75), 1547 (46), 1546 (100), 1545 (48),
1522 (44), 1521 (44), 1502 (34), 1501 (21), 1500 (19).
HPLC Analysis: Separations were carried out on a C-18 column
(Zorbax SB, 3 µm, 4.6×50 mm) with water/acetonitrile (1:1, v/v) as
isocratic eluent and detection at 210 nm. Conversions were calcu-
lated from the ratio of the peak areas of 4-bromobenzyl alcohol
and 4-phenylbenzyl alcohol corrected for the extinction coeffi-
cients.
Calculation of Rate Constants: ln(1 – x), where x is the conversion,
was plotted against time. The slope, K, was calculated by linear
regression. The apparent rate constant was calculated as kapp
=
–K Vrxn/nPd, where nPd is the molar amount of pre-catalyst and Vrxn
is the reaction volume [see Equation (2)].
Typical Procedure for Recycling Experiments in Water: An aqueous
stock solution of the substrates (3 mL, containing 0.3 mmol of 4-
bromomandelic acid, 0.33 mmol of phenylboronic acid and
0.6 mmol of Na2CO3) was added to the FSG-supported catalyst
(100 mg containing 0.003, 0.03, 0.3 or 3 µmol of Pd complex) under
argon. The reaction mixture was shaken with a frequency of
900 min–1 at 80 °C for 16 h. Aliquots (20 µL) of the solution were
withdrawn after 15, 30, 45, 60, 75, 90, 105, 120, 240, 480 and
960 min, and the conversion was measured by HPLC analysis. The
reaction mixture was then cooled to 20 °C and filtered. The FSG
was washed with methanolic HCl (2×1 mL, 1 , MeOH/H2O,
11:1), MeOH (2×1 mL) and water (2×1 mL). The immobilized
catalyst was reused as such in further experiments. All manipula-
tions were carried out automatically.
C106H62F102O4P2Pd2 (3614.3): calcd. C 35.23, H 1.78, P 1.71; found
C 35.62, H 1.97, P 1.71.
Preparation of trans-[Bis{4-[bis{(3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,10-
heptadecafluorodecyl)(2-methyl)phenyl}phosphanyl]benzyl-P,C}di(µ-
chloro)dipalladium(II)] (3a): A solution of LiCl (6.35 mg,
0.15 mmol) in MeOH (10 mL) was added dropwise to a solution
of 3b (216.8 mg, 0.06 mmol) in THF/Et2O (1:1, 40 mL). A yellow-
green solid precipitated immediately and the reaction mixture was
stirred for an additional 2 h. The yellow-green solid was filtered,
washed with water (2 × 10 mL), MeOH (2 × 10 mL) and Et2O
(2×10 mL), and dried under vacuum. Yield: 196.9 g, (92%). 31P
NMR (162 MHz, C6F6/C6D6 capillary): δ = 26.44 ppm. MS (ESI-
dir.): m/z (%) = 3563 (4) [M]+, 2570 (4), 1798 (8) [1/2MH + O]+,
1797 (7) [1/2M + O]+, 1781 (5) [1/2M]+, 1724 (13), 1723 (10), 1670
(12), 1641 (25), 1639 (20), 1639 (11), 1622 (14), 1621 (45), 1618
(12), 1617 (39), 1616 (65), 1563 (18), 1547 (46), 1546 (100), 1545
(48), 1522 (42), 1521 (43), 1502 (33), 1501 (22), 1500 (18).
C102H58Cl2F102P2Pd2 (3567.1): calcd. C 34.35, H 1.64, P 1.74;
found C 34.65, H 1.79, P 1.53.
HPLC Analysis: Separations were carried out on a C-18 column
(Zorbax SB, 3 µm, 4.6×50 mm) with water/acetonitrile/formic acid
(70:30:0.2, v/v) as isocratic eluent and detection at 210 nm. Conver-
sions were calculated from the ratio of the peak areas of 4-bromo-
mandelic acid and 4-phenylmandelic acid corrected for the extinc-
tion coefficients.
Typical Procedure for Filtration Experiments: The experiments were
carried out as described above for the recycling experiments in
water with the modification that the reaction mixture was filtered
(suction filtration) after withdrawal of the second sample. The fol-
lowing samples were withdrawn from the filtrate.
Typical Preparation of Supported Complexes: Palladium complex
1a (30.5 mg, 9.03 µmol) was dissolved in BTF (2 mL). The yellow
solution was diluted with diethyl ether (300 mL) and FSG
(3020 mg) was added. The solvent was removed by rotary evapora-
tion. Diethyl ether (300 mL) was added to the dry FSG and the
solvent was again removed by rotary evaporation. This was re-
peated once more. Drying in vacuo gave the FSG-supported com-
plex (3040 mg) as a pale-yellow, free-flowing solid with a calculated
loading of 2.97 µmolg–1.
Typical Procedure for Experiments with Pre-wetting of the Support:
The experiments were carried out as described above for the recycl-
ing experiments in water with the modification that the supported
complex was shaken with methanolic HCl (2×1 mL, 1 , MeOH/
H2O, 11:1) and MeOH (2×1 mL) and filtered prior to the first
reaction.
5260
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2005, 5248–5261

Pd-Catalysts for Suzuki Reactions
FULL PAPER
[12] S. Schneider, W. Bannwarth, Angew. Chem. 2000, 112, 4293–
4296; Angew. Chem. Int. Ed. 2000, 39, 4142–4145.
[13] S. Schneider, W. Bannwarth, Helv. Chim. Acta 2001, 84, 735–
742.
Acknowledgments
Generous support of this work by Fluka (Buchs, CH) is gratefully
acknowledged.
[14] C. Markert, W. Bannwarth, Helv. Chim. Acta 2002, 85, 1877–
1882.
[15]
[16]
[17]
S. Darses, M. Pucheault, J. P. Genêt, Eur. J. Org. Chem. 2001,
1121–1128.
K. S. A. Vallin, Q. Zhang, M. Larhed, D. P. Curran, A.
Hallberg, J. Org. Chem. 2003, 68, 6639–6645.
W. A. Hermann, C. Broßmer, K. Öfele, C.-P. Reisinger, T. Pri-
ermeier, M. Beller, H. Fischer, Angew. Chem. 1995, 107, 1989–
1992; Angew. Chem. Int. Ed. Engl. 1995, 34, 1844–1847.
[1] I. T. Horváth, J. Rábai, Science 1994, 266, 72–75.
[2] D. P. Curran, Angew. Chem. 1998, 110, 1230–1255; Angew.
Chem. Int. Ed. 1998, 37, 1174–1196.
[3] E. de Wolf, G. van Koten, B.-J. Deelman, Chem. Soc. Rev.
1999, 28, 37–41.
[4] A. Endres, G. Maas, Chem. Unserer Zeit 2000, 34, 382–393.
[5] C. C. Tzschucke, C. Markert, W. Bannwarth, S. Roller, A.
Hebel, R. Haag, Angew. Chem. 2002, 114, 4136–4173; Angew.
Chem. Int. Ed. 2002, 41, 3964–4001.
[6] Handbook of Fluorous Chemistry (Eds.: J. A. Gladysz, D. P.
Curran, I. T. Horváth), Wiley-VCH, Weinheim, 2004.
[7] G. E. Berendsen, L. De Galan, J. Liq. Chromatogr. 1978, 1,
403–426.
[18] A. H. M. de Vries, J. M. C. Mulders, J. H. M. Mommers,
H. J. W. Henderickx, J. G. de Vries, Org. Lett. 2003, 5, 3285–
3288.
[19] Since the first sample was withdrawn after 15 min, any acti-
vation period would have to have been longer then approx.
10 min in order to be detectable.
[8] D. P. Curran, Synlett 2001, 1488–1496.
[20] S. Hietkamp, H. Sommer, O. Stelzer, Inorg. Synth. 1989, 25,
[9] C. C. Tzschucke, C. Markert, H. Glatz, W. Bannwarth, Angew.
Chem. 2002, 114, 4678–4681; Angew. Chem. Int. Ed. 2002, 41,
4500–4503.
120–123.
[21] D. Schwinn, H. Glatz, W. Bannwarth, Helv. Chim. Acta 2003,
86, 188–195.
[10] C. C. Tzschucke, W. Bannwarth, Helv. Chim. Acta 2004, 87,
[22] C. T. Hunt, A. L. Balch, Inorg. Chem. 1982, 21, 1641–1644.
2882–2889.
Received: April 20, 2005
[11] J. A. Gladysz, Pure Appl. Chem. 2001, 73, 1319–1324.
Published Online: October 31, 2005
Eur. J. Org. Chem. 2005, 5248–5261
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5261