A recyclable nanosize aminoarenethiolato copper(I) catalyst for C-C coupling reactions

Article abstract of DOI:10.1016/j.jorganchem.2004.07.023

An aminoarenethiolato copper(I) catalyst was attached to a carbosilane dendritic wedge, which had been prepared via a novel convergent synthetic method. Compared with the unsupported complex, this novel dendritic copper(I) catalyst is more robust towards

Full text of DOI:10.1016/j.jorganchem.2004.07.023

Journal of Organometallic Chemistry 689 (2004) 3813–3819
www.elsevier.com/locate/jorganchem
A recyclable nanosize aminoarenethiolato copper(I) catalyst for
C–C coupling reactions
Anne M. Arink, Rob van de Coevering, Birgit Wieczorek, Judith Firet,
Johann T.B.H. Jastrzebski, Robertus J.M. Klein Gebbink, Gerard van Koten
*
Department of Metal-Mediated Synthesis, Debye Research Institute, Utrecht University, Padualaan 8, 3584 CH Utrecht, The Netherlands
Received 13 May 2004; accepted 14 July 2004
Available online 12 August 2004
Abstract
An aminoarenethiolato copper(I) catalyst was attached to a carbosilane dendritic wedge, which had been prepared via a novel
convergent synthetic method. Compared with the unsupported complex, this novel dendritic copper(I) catalyst is more robust to-
wards hydrolysis and oxidation and has increased solubility in common organic solvents. The catalytic activity of the dendritic cop-
per catalyst was tested in the 1,4-addition of Et2Zn to 2-cyclohexenone. In both polar (Et2O) and apolar (hexane) solvents excellent
activity was observed. The fact that the catalytic copper site remains attached to the nanosize dendritic aminoarenethiolate ligand
allows separation of this catalyst by means of nanofiltration.
ꢀ 2004 Elsevier B.V. All rights reserved.
Keywords: Aminoarenethiolate; Copper catalyst; Carbosilane dendrimer; Catalysis; Nanofiltration; Recycling
1. Introduction
studied, which appeared to have wide applicability as
stable catalysts in, e.g. 1,4-additions [6–9], 1,6-additions
[10] and allylic substitutions [11,12]. Although it is not
the most active copper catalyst reported to date, recent
studies showed that these aminoarenethiolato copper(I)
catalysts are very robust, i.e. thermally stable and have
considerable stability towards oxidation by air. Further-
more, mechanistic studies established that the covalent
thiolato-copper bond is maintained throughout the cat-
alytic cycle in processes involving the use of RLi [13],
RMgX [6] and R2Zn reagents [14]. A recent example,
shown in Fig. 1, is the successful use of 1a and 1b as cat-
alyst in the 1,4-addition of Et2Zn to 2-cyclohexenone.
In contrast to the general low solubility of copper(I)
salts as a consequence of their polymeric nature, the
thiolato copper complexes, of type 1, are well-defined
aggregated species, usually trimers [15,16], and have fav-
orable solubility properties, which can even be enhanced
by the introduction of silyl groups (1b). Here we report
the synthesis and catalytic properties of an aminoaren-
Continuous recycling of soluble transition-metal cat-
alysts in homogenous catalysis is possible by application
of catalysts immobilized on nanosize organic supports
(e.g. carbosilanes [1,2] or cartwheel-type molecules [3])
in a membrane reactor [4]. These catalysts must have a
stable metal–ligand bond otherwise leaching of the me-
tal will rapidly reduce the performance and recyclability
of the catalytic system. In copper-catalyzed reactions
this prerequisite is difficult to fulfill [5], because only a
few isolated cases are known in which the copper-ligand
bond remains intact throughout the catalytic cycle. One
of these cases involves the use of monoanionic aminoa-
renethiolato ligands which are N,S-bonded to copper(I)
in a bidentate fashion. In particular the catalytic activity
of aminoarenethiolato copper(I) 1a has been extensively
*
Corresponding author. Tel.: +31 30 2533120; fax: +31 30 2523615.
E-mail address: g.vankoten@chem.uu.nl (G. van Koten).
0022-328X/$ - see front matter ꢀ 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2004.07.023

3814
A.M. Arink et al. / Journal of Organometallic Chemistry 689 (2004) 3813–3819
2.2. Preparation of compounds 5–10
2.2.1. Tris[2-(dimethylphenylsilyl)ethyl]chlorosilane (5)
Solid [(Bu4N)2PtCl6] catalyst (1 mg) was added to a
solution of trivinylsilyl chloride (1.0 mL, 6.46 mmol)
and dimethylphenylsilane (3.3 mL, 21.31 mmol) in
THF (4 mL). The suspension was heated for 10 min in
an oil bath at 80 ꢁC and then the mixture was stirred
overnight at room temperature. All volatiles were re-
moved in vacuo yielding a crude yellowish viscous oil
(3.45 g, 6.24 mmol, 96%) which was used without fur-
Fig. 1. 1,4-Addition of Et2Zn to 2-cyclohexenone using copper(I)
complexes 1a and 1b as catalysts.
1
ethiolato copper(I) catalyst that has been connected to
the focal point of a dendritic wedge. This species exists
in the resting state at least as a trimer which has a nano-
size dimension. Consequently, this copper(I) catalyst
can be separated from products and reagents and iso-
lated by nanofiltration and subsequently reused.
ther purification. H NMR: d 0.18 (s, 18H, 3· SiMe2),
0.66–0.77 (m, 12H, 3· Si(CH2CH2Si)), 7.20 (m, 9H,
ArH(3,4,5)), 7.42 (m, 6H, ArH(2,6)). 13C NMR: d
À3.53 (3· SiMe2), 7.81 (3· Si(CH2CH2Si)), 128.2
(ArC(2,6)), 129.3 (ArC(4)), 134.0 (ArC(3,5)), 138.8
(ArC(1)). 29Si NMR: d À12.24 (3· SiMe2), 24.51 (SiCl).
2.2.2. {Tris[2-(dimethylphenylsilyl)ethyl]}{4-[(dimeth-
ylaminomethyl)phenyl]}silane (7)
2. Experimental
t-BuLi (7.5 mL, 11.2 mmol) was added to a solution of
4 (1.06 g, 4.96 mmol) in diethyl ether (17 mL) at À78 ꢁC.
After 30 min. a solution of 5 (2.88 g, 5.22 mmol) in diethyl
ether (5 mL) was added to the cold mixture. After 25 min.
the cloudy mixture was allowed to reach room tempera-
ture, stirred for another 18 h and subsequently quenched
with H2O (20 mL). The organic layer was separated and
the aqueous layer was extracted with pentane (4 · 10
mL). The combined organic layers were dried on Na2SO4
and filtered. All volatiles were removed in vacuo which
yielded the pure yellow-colored viscous oil (2.50 g, 3.85
mmol, 77%). 1H NMR: d 0.23 (s, 18H, 3· SiMe2), 0.69–
0.73 (m, 6H, 3· Si(CH2CH2Si)), 0.83–0.87 (m, 6H, 3·
Si(CH2CH2Si)), 2.07 (s, 6H, NMe2), 3.27 (s, 2H, CH2),
7.21–7.23 (m, 9H, SiMe2ArH(3,4,5)), 7.39 (d, 2H,
3J = 7.5 Hz, SiArH), 7.44–7.49 (m, 8H, SiMe2ArH(2,6)
and SiArH). 13C NMR: d À3.56 (3· SiMe2), 3.55
(Si(CH2CH2Si)3), 7.92 (Si(CH2CH2Si)3), 45.36 (NMe2),
64.40 (CH2N), 128.0 (SiMe2ArC), 128.6 (SiArC), 129.0,
133.9 (SiMe2ArC), 134.6, 135.4 (SiArC), 139.1 (Si-
Me2ArC), 140.6 (SiArC). 29Si NMR: d À12.42 (3·
SiMe2), À7.70 (Si(CH2CH2Si)3). MS (MALDI-TOF):
m/z 650.03 ([M + DHB]+) (calcd. 651.36). Anal. Calc.
for C39H57NSi4: C, 71.82; H, 8.81; N, 2.15; Si, 17.22.
Found: C, 71.96; H, 8.86; N, 2.07; Si, 17.11%.
2.1. Chemicals and equipment
All experiments were performed under an inert
nitrogen atmosphere using standard Schlenk tech-
niques. Solvents were carefully dried and distilled
prior to use. Standard chemicals were purchased from
Acros Organics or Aldrich, except for dimethylphen-
ylsilane (Lancaster) and trivinylchlorosilane (ABCR).
4-[(Dimethylamino)methyl]bromobenzene (4) was pre-
pared according to a previously reported procedure
[17–19]. 1H, 13C{1H} and 29Si{1H} NMR spectra of
all compounds were recorded in benzene-d6 on a 300
MHz spectrometer at 298 K unless stated otherwise.
GC measurements were carried out using a DB17
(30 m · 0.32 mm) column and GC-MS measurements
were carried out on a Turbo Mass spectrometer. Fast
Atom Bombardment (FAB) mass spectrometry was
carried out on a four-sector mass spectrometer. Sam-
ples were loaded in a matrix solution (3-nitrobenzyl
alcohol) on to a stainless steel probe and bombarded
with Xenon atoms with an energy of 3 keV. During
the high-resolution FAB-MS measurements a resolv-
ing power of 10,000 (10% valley definition) was used.
MALDI-TOF mass spectrometry measurements were
carried out on a mass spectrometer equipped with a
nitrogen laser emitting at 337 nm. External calibration
was performed using C60 and C70. Sample and matrix
(2,5-dihydroxybenzoic acid) solutions were prepared
with an approximate concentration of 30 mg/mL in
THF. A 0.2 lL sample of the solution and 0.2 lL
of the matrix solution were combined and placed on
the golden MALDI target and analyzed after evapora-
tion of the solvents. Elemental analyses were per-
formed by Kolbe, Mikroanalytisches Laboratorium,
2.2.3. {Tris[2-(dimethylphenylsilyl)ethyl]}{[4-(dimeth-
ylaminomethyl)-2-(trimethylsilylthio)phenyl]}silane (8)
t-BuLi (4 mL, 6.0 mmol, 1.5 M in pentane) was added
to a solution of 7 (1.94 g, 2.98 mmol) in pentane (40 mL)
at À78 ꢁC. After 1 h, the orange solution was allowed to
reach room temperature and was stirred for 2.5 days.
The solution was concentrated to 20 mL in vacuo. Sub-
sequently, cold THF (100 mL, 60 ꢁC) was added and the
solution was further cooled to À78 ꢁC. Sublimed sulfur
(0.19 g, 6.0 mmol) was added to the solution and the
Muhlheim/Ruhr, Germany.
¨

A.M. Arink et al. / Journal of Organometallic Chemistry 689 (2004) 3813–3819
3815
reaction was stirred at low temperature. After 7 h, the
reaction mixture was allowed to reach room tempera-
ture and was stirred overnight. Trimethylchlorosilane
(1 mL, 7.88 mmol) was added and the solution was stir-
red for 4.5 h. All volatiles were removed in vacuo and
toluene (50 mL) was added to the milky orange-colored
viscous oil. After centrifugation the supernatant was
separated and all volatiles were removed in vacuo which
yielded a red-brown-colored oil (1.53 g, 2.03 mmol,
was stirred overnight at room temperature. After filtra-
tion through Celite, the colorless filtrate was added to
HCl(aq) (100 mL, 2 M). The aqueous layer was extracted
with Et2O (2 · 100 mL) and the combined organic layers
were washed with H2O (2 · 50 mL) and with brine (100
mL). The mixture was dried over anhydrous Na2SO4
and filtered through a pad of silica gel. The solvents were
removed in vacuo which yielded a white milky oil (1.31 g,
1
2.52 mmol, 51%). H NMR: d 0.20 (s, 18H, 3· SiMe2),
1
68%). H NMR: d 0.22 (s, 9H, SSiMe3), 0.23 (s, 18H,
0.46–0.54 (m, 6H, Si(CH2CH2Si)3), 0.58–0.66 (m, 6H,
Si(CH2CH2Si)3), 7.13–7.22 (m, 9H, ArH), 7.41–7.47 (m,
6H, ArH). 13C NMR: d À2.98 (3 SiMe2), 7.01
Si(CH2CH2Si), 8.10 (3· Si(CH2CH2Si)), 128.5
(ArC(2,6)), 129.5 (ArC(4)), 134.3 (ArC(3,5)), 139.6
(ArC(1)). 29Si NMR: d À12.39 (3· SiMe2), À2.56 (SiH).
Anal. Calc. for C30H46Si4: C, 69.42; H, 8.93; Si, 21.64.
Found: C, 69.28; H, 9.06; Si, 21.61%.
3· SiMe2), 0.66–0.75 (m, 6H, 3· Si(CH2CH2Si)), 0.78–
0.89 (m, 6H, 3· Si(CH2CH2Si)), 2.19 (s, 6H, NMe2),
3.76 (s, 2H, CH2N), 7.16–7.26 (m, 9H, Si-
Me2ArH(3,4,5)), 7.38/7.70 ((AB, 2H, SiArH(3,4)),
JAB = 7.6 Hz), 7.44–7.53 (m, 6H, SiMe2ArH(2,6)), 7.84
(s, 1H, SiArH(6)). 13C NMR: d À2.99 (3· SiMe2), 1.76
(SSiMe3), 3.96 (3· Si(CH2CH2Si)), 8.25 (3· Si(CH2CH2-
Si)), 46.01 (NMe2), 62.97 (CH2N), 128.5, 129.5 (Si-
Me2ArC), 130.1, 131.9, 133.7 (SiArC), 134.3, 139.5
(SiMe2ArC), 139.0, 143.0, 144.3 (SiArC). 29Si NMR: d
À12.43 (3· SiMe2), À7.30 (Si(CH2CH2Si)3), À0.84
(SSiMe3). Anal. Calc. for C42H65NSSi5: C, 66.38; H,
8.57; N, 1.79; Si, 18.48. Found: C, 66.69; H, 8.66; N,
1.85; Si, 18.56%.
2.3. Membrane filtration
Crude 9 was dissolved in diethyl ether (10 mL) and
transferred into a dialysis tubing under a purge of nitro-
gen. The dialysis tubing was sealed and placed in a
closed system with diethyl ether (200 mL) under nitro-
gen. After stirring for 4 h the contents of the dialysis
tube was transferred to a Schlenk flask and all volatiles
were removed in vacuo.
2.2.4. {Tris[2-(dimethylphenylsilyl)ethyl]}{[4-(dimeth-
ylaminomethyl)-2-(copperthiolato)phenyl]}silane (9)
A solution of 8 (1.49 g, 1.97 mmol) in toluene (20 mL)
was added to a CuCl suspension (0.18 g, 1.87 mmol) in tol-
uene (10 mL). The brownish suspension was stirred until
all CuCl had reacted (1 h). All volatiles were removed in
vacuo which yielded the crude red-brown-colored viscous
oil (1.34 g, 1.80 mmol, 96%) which was purified by mem-
brane filtration. 1H NMR (toluene-d6, 358 K): d 0.21, 0.30
(2 · s, 18H, 3· SiMe2), 0.67–0.83 (m, 6H, 3 Si(CH2CH2-
Si)), 0.86–0.99 (m, 6H, 3· Si(CH2CH2Si)), 2.03–2.20 (m,
6H, NMe2), 3.31–3.82 (m, 2H, CH2), 6.78/7.14 ((AB,
2H, SiArH(3,4)), JAB = 7.6 Hz), 7.20–7.25 (m, 9H, Si-
Me2ArH(3,4,5)), 7.47–7.50 (m, 6H, SiMe2ArH(2,6)),
8.12 (s, 1H, SiArH(6)). 13C NMR (toluene-d8, 358 K): d
À3.95 (3· Si(CH3)2), 3.70 (3· Si(CH2CH2Si)), 8.40 (3·
Si(CH2CH2Si)), 47.79 (NMe2), 66.89 (CH2N), 127.8,
128.7 (Si(CH3)2ArC), 129.6, 131.7, (SiArC), 133.5, 133.6
(Si(CH3)2ArC), 134.3, 138.9, 139.2 (SiArC). 29Si NMR
(toluene-d8, 358 K): d À12.50 (3· SiMe2), À8.06 (Si
(CH2CH2Si)3). MALDI-TOF-MS: m/z 1558 ([Cu3
(SAr)2]+), 1492 ([Cu2(SAr)2]+), 808 ([Cu2(SAr)]+), 745
([Cu(SAr)]+). FAB-MS: Rel. Peak Int. m/z 1491 ([Cu2
(SAr)2]+), 0.7%), 808 ([Cu2(SAr)]+, 10%), 745 ([Cu
(SAr)]+), 8%). Anal. Calc. for C39H56CuNSSi4: C, 62.72;
H, 7.56; N, 1.88. Found: C, 61.83; H, 8.03; N, 1.48%.
2.4. General procedure for 1,4-addition of diethylzinc to 2-
cyclohexenone
The catalyst (20 lmol) was dissolved in dry solvent (2
mL) and cooled to À20 ꢁC. After 20 min, Et2Zn (1.0 M
solution in pentane, 1.2 mL, 1.2 mmol) was added. The
mixture was stirred for 5 min and then 2-cyclohexenone
(96 lL, 1.0 mmol) was added. After 4 h the reaction
mixture was quenched with 1 M HCl and the layers were
separated. The organic layer was filtered through silica
and reaction products were analyzed by GC.
2.5. Molecular modeling
Molecular modeling was performed using SPARTAN
5.1.1 running on a SGI-O2 computer and calculations
employed the MMFF84 forcefield. Parameters from
the crystal structure of 1a were used as the starting
point for the aminoarenethiolato Cu(I) portion of the
molecule. Solvent effects were excluded in the calcula-
tions.
2.2.5. Tris-[2-(dimethylphenylsilyl)ethyl]silane (10)
An ice-cooled solution of 5 (2.73 g, 4.94 mmol) in
diethyl ether (10.5 mL) was added to LiAlH4 (93.7 mg,
2.47 mmol) in diethyl ether (7.5 mL). The grayish mixture
3. Results and discussion
In an explorative study, we began with linking a
number of aminoarenethiolato copper(I) units to the

3816
A.M. Arink et al. / Journal of Organometallic Chemistry 689 (2004) 3813–3819
N
Fig. 2. A G0-carbosilane dendritic species with peripheral aminoaren-
ethiolato copper(I) units.
Cu
periphery of a small carbosilane dendrimer (Fig. 2).
Three aminoarenethiolato copper(I) units were attached
via the 5-arene position to the peripheral Si-groups of a
G0-dendritic wedge with the expectation that the three
aminoarenethiolato copper(I) units in 2 would trigger
the formation of a trimeric aggregate by intramolecular
S–Cu coordination. This assumption was based on the
fact that the parent aminoarenethiolato copper(I) com-
pound exists both in the solid state and in solution as
a trimeric aggregate [16], see Fig. 3. However, irrespec-
tive the method of preparation of 2, a highly insoluble
material was obtained, indicating that, instead of intra-
molecular aggregation, polymeric networks were formed
as a result of intermolecular S–Cu bond formation.
To avoid the formation of such (insoluble) polymer
networks and to obtain a soluble catalyst the strategy
was changed and a single aminoarenethiolato copper(I)
unit was linked to the focal point of a carbosilane dend-
ritic wedge. The choice of carbosilane dendritic materi-
als as support in these studies is based on its chemical
inertness which makes this material compatible with
the harsh reaction conditions involved in both catalyst
S
Fig. 3. X-ray structure [16] of the enantiopure a-methyl-substituted
analog of 1a.
preparation and catalytic C–C coupling reactions with
organometallic reagents [15–19].
G0-wedge 5 was prepared via a hydrosilylation reac-
tion of trivinylchlorosilane (3) with 3.3 equivalents of
dimethylphenylsilane (Scheme 1). Excess dimethyl-
phenylsilane could be removed in vacuo. 4-[(Dimethyl-
amino)methyl]-1-bromobenzene (4) was converted into
the corresponding lithium derivative (6) and subse-
quently coupled to carbosilane wedge 5 to afford 7 [20].
Scheme 1. Reagents and conditions: (i) 3.3 eq. HSiMe2Ph, (Bu4N)2PtCl6, THF, 10 min, 80 ꢁC; (ii) 2.2 eq. t-BuLi, Et2O, rt; (iii) Et2O, rt; (iv) 2 eq.
t-BuLi, pentane, rt; (v) 1/4 S8, THF, À60 ꢁC; (vi) excess SiMe3Cl; (vii) 0.95 eq. CuCl, toluene, rt; and (viii) LiAlH4, Et2O, rt.

A.M. Arink et al. / Journal of Organometallic Chemistry 689 (2004) 3813–3819
3817
Lithiation of 7 was accomplished by the addition of
a twofold excess of lithiating reagent, t-BuLi, to ensure
complete lithiation at the position ortho to the amino
substituent [21]. A stoichiometric amount of sulfur rel-
ative to t-BuLi was used for the subsequent sulfur
insertion reaction. The resulting lithium thiolate inter-
mediate was quenched with trimethylsilyl chloride
affording 8. The t-butyl thioether formed as byproduct
could be easily removed in vacuo. Reaction of the aryl
trimethylsilyl thioether 8 with 0.95 equivalents of CuCl
gave the desired aminoarenethiolato copper(I) complex
9, as well as Me3SiCl, which was removed in vacuo
(Scheme 1). Unreacted 8 was removed from 9 by pas-
sive dialysis [22] in diethyl ether (see Section 2). Com-
plex 9 was isolated as a reddish oil and is soluble in
common organic solvents ranging from tetrahydrofu-
ran to hexane. FAB-MS and MALDI-TOF-MS spectra
of 9 showed a peak at m/z = 745 ([monomeric 9]+, 8%),
as well as peaks at 808 ([monomeric 9 + Cu]+, 10%)
and 1491 ([dimeric 9]+, 0.7%). The resonances found
a molecular mass of 745 g molÀ1, separation from, for
example 8, can only occur when monomeric 9 has aggre-
gated at least to trimers (molecular mass of 2235
g molÀ1) [24–27].
Molecular mechanics calculations [28] were carried
out to gain more insight in the size and shape of the
molecular structure of 9 (Figs. 4 and 5). The structures
of both monomeric and trimeric 9 have been calculated
and both showed a relatively flat structure. They have
average diameters of 1.6 and 2.3 nm, and molecular vol-
3
˚
umes of 869 and 2583 A , respectively (Figs. 4 and 5)
[28].
Complex 9 was tested as a catalyst in the 1,4-addition
reaction of diethylzinc to 2-cyclohexenone in a variety of
solvents (Fig. 1). The results were compared to those ob-
tained with unsupported complexes 1a and 1b. The cat-
alytic performance of metallodendrimer 9 (entries 6–10)
is similar to those of unsupported copper(I) complexes
1a and 1b (entries 2 and 5) or even exceeds these (com-
pare entries 3 and 4 with entries 7 and 8) (Table 1). The
influence of para-substitution had already been studied
with catalyst 1b (entry 5) [14].
1
in the H, 13C and 29Si NMR spectra of 9 confirmed
its synthesis.
Most likely 9 has an aggregated structure like the par-
ent compounds 1a and 1b. These parent compounds ex-
ist in benzene solution as trimeric aggregates comprising
a six-membered Cu3S3 ring in a chair configuration with
alternating copper and sulfur atoms, see Fig. 3. Each
sulfur atom bridges two copper atoms with all arene
substituents taking equatorial positions [15,16]. Analysis
of models of 2 indicate that, for intramolecular aggrega-
tion, all the arene substituents would occupy axial posi-
tions. Due to steric considerations, this configuration is
disfavoured and intermolecular bonding is observed
resulting in formation of polymeric insoluble products.
However, the similar spectroscopic data of 9 and 1 with
respect to the central Cu3S3-core would suggest a trinu-
clear structure (m/z = 2235 g molÀ1) for 9.
The increase in activity in toluene and THF going
from 1a to 9 indicates that the presence of the wedge
has a positive influence on the catalytic activity. This
is most likely due to the improved physical properties
of 9, i.e. solubility, and better stability to hydrolysis
and oxidation. The improved solubility of 9 prompted
us to perform the same catalysis in less common solvents
for these reactions, such as hexane and dichloromethane
(entries 9 and 10). Interestingly, the activity and selectiv-
ity of the dendritic catalyst in hexane were similar to
those in the other more polar solvents, except THF.
After quenching of the reaction mixture with aqueous
HCl (entry 6), the organic layer was submitted to nano-
filtration in order to recover 9 as a solution in Et2O. The
fraction retained by the membrane was subjected to an-
other catalytic run and the 1,4-addition product was
formed quantitatively, proving the presence of active
catalyst after recycling. The stability of catalyst 9 in
solution towards repeated exposure to water demon-
strates its robustness as compared to 1a. Moreover,
exposure of a solution of 1a to air results in rapid
The proposed aggregation of a number of monomeric
units 9 is also in line with the successful purification of 9
from low molecular-weight impurities by using a passive
membrane filtration of crude 9. The membrane used sep-
arates molecules with a mass below 1200 Da from those
with a mass above 2000 Da [23]. Since monomeric 8 has
Fig. 4. Molecular modeling of monomeric 9; topview and sideview.

3818
A.M. Arink et al. / Journal of Organometallic Chemistry 689 (2004) 3813–3819
cies. First of all, the wedge has a stabilizing effect on
the resting state of the catalyst. Second, the presence
of a bulky carbosilane substituent could activate aggre-
gate dissociation to provide catalyst–substrate as well
as catalyst–substrate–reagent intermediates [29] from
which product formation occurs. Finally, site isolation
effects caused by the inert carbosilane wedge may be in
operation, making deactivation of the catalyst by
aggregation less likely.
4. Conclusions
An aminoarenethiolato copper(I) compound was at-
tached to a G0-carbosilane dendritic wedge, which had
been prepared via a novel convergent synthetic method.
The convergent synthesis of the supported catalyst al-
lows easy access to higher generations. By reduction of
chlorosilane 5 with LiAlH4, silane 10 was prepared. This
silane can be used to react with trivinylchlorosilane 3 to
obtain higher generation chlorosilane wedges, see
Scheme 1. The introduction of the actual catalytic cop-
per center can thus be postponed to the final step. Com-
pared with the unsupported complex, the novel dendritic
catalyst 9 is more robust towards water, has increased
solubility in common organic solvents and its catalytic
activity is comparable to that of the parent aminoaren-
ethiolato copper(I) compound.
The improved solubility of the dendritic Cu(I) com-
plex, even in apolar solvents such as hexane, allows
the use of a wide variety of substrates that are not solu-
ble in conventional solvents.
Fig. 5. Molecular modeling of trimeric 9; topview and sideview.
The increased stability and size of 9 allows separation
of this catalyst by means of a nanofiltration membrane.
Optimization of catalyst recycling requires synthesis of
copper(I) catalysts with higher generation dendritic
wedges. Research towards this goal is currently under-
way.
Table 1
Catalytic data for the Michael addition of Et2Zn to 2-cyclohexenone in
the absence and presence of 2 mol% of aminoarenethiolato copper(I)
catalysts 1a, 1b and 9a
Entry
Cu(I) complex
Solvent
Conversionb (%)
1
2
None
1a
1a
1a
1b
9
Et2O
Et2O
3
>99
95
3
Toluene
THF
Et2O
4
24
Acknowledgement
5
>99
>99
>99
62
6
Et2O
Toluene
THF
The authors acknowledge financial support by the
Council of Chemical Sciences (CW) of the Netherlands
Organization for Scientific Research (NWO).
7
9
8
9
9
9
Hexane
CH2Cl2
>99
>99
10
a
9
Reaction carried out in solvent (2 mL) at À20 ꢁC in the presence
of 1.2 eq. Et2Zn, 2 mol% of catalyst.
b
Determined by GC after 4 h.
References
[1] R. Kreiter, A.W. Kleij, R.J.M. Klein Gebbink, G. van Koten,
Top. Curr. Chem. 217 (2001) 163.
[2] R. van Heerbeek, P.C.J. Kamer, P.W.N.M. van Leeuwen, J.N.H.
Reek, Chem. Rev. 102 (2002) 3717.
decomposition of the catalyst, whereas a solution of 9 is
still active after exposure to air.
The improved catalytic performance of the dendritic
catalyst 9 as compared to 1a can be ascribed to the
presence of the dendritic wedge in this novel Cu(I) spe-
[3] H.P. Dijkstra, G.P.M. van Klink, G. van Koten, Acc. Chem. Res.
35 (2002) 798.
[4] I.F.J. Vankelecom, Chem. Rev. 102 (2002) 3779.
[5] C.C. Mak, H.-F. Chow, Macromolecules 30 (1997) 1228.

A.M. Arink et al. / Journal of Organometallic Chemistry 689 (2004) 3813–3819
3819
[6] M. van Klaveren, F. Lambert, J.F.M. Eijkelkamp, D.M. Grove,
G. van Koten, Tetrahedron Lett. 35 (1994) 6135.
[19] A.W. Kleij, R.A. Gossage, R.J.M. Klein Gebbink, N. Brinkmann,
E.J. Reijers, U. Kragl, M. Lutz, A.L. Spek, G. van Koten, J. Am.
Chem. Soc. 122 (2000) 12112.
[7] Q.L. Zhou, A. Pfaltz, Tetrahedron 50 (1994) 4467.
[8] M. Spescha, G. Rihs, Helv. Chim. Acta 76 (1993) 1219.
[9] D. Seebach, G. Jaeschke, A. Pichota, L. Audergon, Helv. Chim.
Acta 80 (1997) 2515.
[20] A.W. Kleij, H. Kleijn, J.T.B.H. Jastrzebski, W.J.J. Smeets, A.L.
Spek, G. van Koten, Organometallics 18 (1999) 268.
[21] Excess of lithiating reagent had also been required, when
fluorous tails were attached to the para-position: H. Kleijn, E.
Rijnberg, J.T.B.H. Jastrzebski, G. van Koten, Org. Lett. 1
(1999) 853.
[10] A. Haubrich, M. van Klaveren, G. van Koten, G. Handke, N.
Krause, J. Org. Chem. 58 (1993) 5849.
[11] G.J. Meuzelaar, A.S.E. Karlstro¨m, M. van Klaveren, E.S.M.
Persson, A. del Villar, G. van Koten, J.-E. Ba¨ckvall, Tetrahedron
56 (2000) 2895.
[22] Commercially available dialysis membranes were employed (ben-
zoylated cellulose, Sigma).
[12] M. van Klaveren, E.S.M. Persson, A. del Villar, D.M. Grove, J.-
E. Ba¨ckvall, G. van Koten, Tetrahedron Lett. 36 (1995) 3059.
[13] F. Lambert, D.M. Knotter, M.D. Janssen, M. van Klaveren, J.
Boersma, G. van Koten, Tetrahedron Lett. 2 (1991) 1097.
[14] A.M. Arink, T.W. Braam, R. Keeris, J.T.B.H. Jastrzebski, C.
Benhaim, S. Rosset, A. Alexakis, G. van Koten, Org. Lett. 6
(2004) 1959.
[23] Dialysis specifications were given for compounds in aqueous
solutions.
[24] C.B. Gorman, B.L. Parkhurst, W.Y. Su, K.-Y. Chen, J. Am.
Chem. Soc. 119 (1997) 1141.
[25] C.B. Gorman, J.C. Smith, M.W. Hager, B.L. Parkhurst, H.
Sierzputowska-Gracz, C.A. Haney, J. Am. Chem. Soc. 121 (1999)
9958.
[15] D.M. Knotter, M.D. Janssen, D.M. Grove, W.J.J. Smeets, E.
Horn, A.L. Spek, G. van Koten, Inorg. Chem. 30 (1991) 4361.
[16] D.M. Knotter, G. van Koten, H.L. van Maanen, D.M. Grove,
A.L. Spek, Angew. Chem. Int. Ed. Engl. 28 (1989) 341.
[17] N.J. Hovestad, J.T.B.H. Jastrzebski, G. van Koten, S.A.F. Bon,
C. Waterson, D.M. Haddleton, Polym. Prep. (1999) 393.
[26] C.B. Gorman, W.Y. Su, H. Jiang, C.M. Watson, P. Boyle, Chem.
Commun. (1999) 877.
[27] R. Wang, Z. Zheng, J. Am. Chem. Soc. 121 (1999) 3549.
[28] For molecular modeling the SPARTAN 5.1.1. (SGI) program was
used with MMFF94 as a force field. The molecular volumes where
determined with the same program, using the smallest spherical
shape to fit the molecule. In this way solvent effects are not
included.
[18] N.J. Hovestad, E.B. Eggeling, H.J. Heidbuchel, J.T.B.H.
¨
Jastrzebski, U. Kragl, W. Keim, D. Vogt, G. van Koten, Angew.
Chem. Int. Ed. 38 (1999) 1655.
[29] M. Enomoto, T. Aida, J. Am. Chem. Soc. 121 (1999) 874.