Palladium complexes of phosphane-functionalised carbosilane dendrimers as catalysts in a continuous-flow membrane reactor

Article abstract of DOI:10.1002/1099-0690(200203)2002:6<1085::AID-EJOC1085>3.0.CO;2-6

Phosphane-functionalised carbosilane dendrimers 9, 10, 11, 12, 13, 19 and 20 have been synthesised, and their palladium complexes have been used as catalysts in the allylic substitution reaction. The catalytic sites at the periphery of the dendrimer support are readily accessible to the substrate, which is reflected - also for the larger dendrimeric systems - in the high catalytic activity. Moreover, the higher generations are sufficiently large to be retained by a nanofilter, and dendrimeric catalysts 13 and 20 have been applied in a continuous-flow membrane reactor. The stability of the palladium complexes of the phosphane-functionalised dendrimers is crucial for application in a continuous process of this type, and appeared to be very sensitive to small changes in the dendrimeric structure. Wiley-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002.

Full text of DOI:10.1002/1099-0690(200203)2002:6<1085::AID-EJOC1085>3.0.CO;2-6

FULL PAPER
Palladium Complexes of Phosphane-Functionalised Carbosilane Dendrimers as
Catalysts in a Continuous-Flow Membrane Reactor
Debby de Groot,[a] Joost N. H. Reek,*[a] Paul C. J. Kamer,[a] and
Piet W. N. M. van Leeuwen*[a]
Keywords: Carbosilane dendrimers / Catalysis / Continuous process / Homogeneous catalysis / Membrane reactor /
Palladium / Phosphanes
Phosphane-functionalised carbosilane dendrimers 9, 10, 11,
12, 13, 19 and 20 have been synthesised, and their palladium
complexes have been used as catalysts in the allylic substitu-
dendrimeric catalysts 13 and 20 have been applied in a con-
tinuous-flow membrane reactor. The stability of the palla-
dium complexes of the phosphane-functionalised dendrimers
tion reaction. The catalytic sites at the periphery of the den- is crucial for application in a continuous process of this type,
drimer support are readily accessible to the substrate, which and appeared to be very sensitive to small changes in the
is reflected − also for the larger dendrimeric systems − in the
high catalytic activity. Moreover, the higher generations
are sufficiently large to be retained by a nanofilter, and
dendrimeric structure.
Wiley-VCH Verlag GmbH, 69451 Weinheim, Germany,
2002)
(
Introduction
generation phosphaneϪester-functionalised carbosilane
dendrimer in the palladium-catalysed hydrovinylation of
styrene in a continuous-flow membrane reactor.[51] They re-
ported precipitation of palladium metal in the reactor and
on the surface of the membrane. The deactivation was at-
tributed to the formation of multiply coordinated phos-
phane complexes and multinuclear phosphane-bridged
complexes, possibly as a result of the high local concentra-
tion of phosphane at the periphery of the dendrimer. Brink-
mann and co-workers reported leaching of palladium when
using a fourth generation poly(propylene imine) dendrimer
functionalised with diphenylphosphaneϪPd end groups in
a palladium-catalysed allylic substitution reaction per-
formed in a continuous-flow membrane reactor.[38] They
compensated for the loss of palladium by adding (allyl)pal-
ladium chloride in the feed, which resulted in product
formation for a longer period. The formation of (li-
gand)PdCl2 complexes was suggested as partly being re-
sponsible for the observed deactivation.
Since the first reports by Vögtle,[1] Newkome,[2] and Tom-
alia,[3] the field of dendrimers has been explored
intensively.[4Ϫ7] The synthesis of these highly ordered, tree-
like structures has been a challenge to many chemists, re-
sulting in new molecules with unique properties. Many dif-
ferent dendrimers have been synthesised, with a variety of
functional groups in the chains,[8Ϫ14] in the core[15Ϫ21] and
at the periphery.[22Ϫ28] One of the main applications of
dendrimers[29Ϫ32] is their use in catalysis,[16,33Ϫ35] enabling
recycling of the homogeneous catalysts by means of
nanofiltration.[36Ϫ40]
Palladium complexes have been used as catalysts in many
reactions that are of great importance to synthetic
chemists.[41Ϫ43] Key issues in the development of easily ap-
plicable palladium catalysts include catalyst stability and re-
cycling. Generally, palladium catalysts are not very stable,
and the formation of palladium metal is often observed.
Nevertheless, polymer-immobilised systems that can be
used in consecutive batchwise reactions have been
reported.[44Ϫ50] An even greater challenge would be the use
of supported palladium catalysts in continuous processes,
since small inactive palladium species would be washed out
of the reactor, whereas these may be reactivated in batch
reactions. Indeed, the so far reported palladium-func-
tionalised dendrimers applied in continuous processes
showed deactivation of the catalyst. Eggeling and co-
workers observed catalyst deactivation when using a first
The goal of the work presented here is the exploration of
palladium-functionalised dendrimers as catalysts for ap-
plication in a continuously operated membrane reactor. For
this
purpose
we
functionalised
carbosilane
dendrimers[52Ϫ57] with diphenylphosphane end groups, and
used the palladium complexes of these systems as catalysts.
The dendrimers are large enough to enable their use in a
continuous-flow membrane reactor. Previously, we had
found that the second generation dendrimer with a calcu-
3 [39]
˚
lated molecular volume of 2414 A
already displayed a
retention of 98.1% when THF was used as a solvent[39]
(99.7% for dichloromethane). The dendrimers used for cata-
lysis in the membrane reactor are much larger, as is clearly
[a]
Institute of Molecular Chemistry, University of Amsterdam,
Nieuwe Achtergracht 166, 1018 WV Amsterdam, The
Netherlands
Eur. J. Org. Chem. 2002, 1085Ϫ1095
WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002 1434Ϫ193X/02/0306Ϫ1085 $ 17.50ϩ.50/0
1085

D. de Groot, J. N. H. Reek, P. C. J. Kamer, P. W. N. M. van Leeuwen
FULL PAPER
shown in Figure 1, and so should display an even higher phenylphosphanyl)methyl]lithium (Scheme 1). The hydro-
retention. Catalyst stability is of crucial importance when silylation step was performed with an excess of chlorosilane
reactions are performed in a continuous-flow membrane re- as the solvent and tetrabutylammonium hexachloroplatin-
actor, and this appeared to be very sensitive to small ate as the catalyst.[59] Ph2PCH2Li·TMEDA was synthesised
changes in the dendrimeric structure.
as described by Schore et al.[60] and obtained as a yellow
solid after lithiation of Ph2PCH3 with n-butyllithium and
TMEDA. A solution of Ph2PCH2Li·TMEDA in THF was
added to the chlorosilane dendrimer. After stirring over-
night, the reaction mixture was filtered through silica, and
the excess of Ph2PCH3 was removed in vacuo. All the phos-
phane-functionalised dendrimers were obtained as oils or
1
wax-like solids. Characterisation by H and 31P{1H} NMR
spectroscopy and MALDI-TOF mass spectrometry showed
that the phosphane-functionalised dendrimers were ob-
tained in at least 95% purity (see Exp. Sect. for details).
Impurities were probably due to hydrolysis of the SiϪCl
bond to give SiϪOH and SiϪOϪSi. The MALDI-TOF
mass spectra of the larger dendrimers (11 and 13) showed
very broad signals in the expected range [for 11: m/z ϭ
8000Ϫ12000 (calcd. 11202); for 13: m/z ϭ 3000Ϫ6000
(calcd. 5888)]. Similar broad signals for larger func-
tionalised dendrimers have been observed by Van Koten
and
co-workers.[40,61]
A
model
compound,
(CH3)2Si(CH2PPh2)2 (14), was prepared by treatment of
bis(chloromethyl)dimethylsilane with potassium di-
phenylphosphide. Previously, Alyea et al. reported the syn-
thesis of this compound by treatment of chlorodiphenyl-
phosphane with the Grignard reagent of bis(chloromethyl)-
dimethylsilane.[62]
Scheme 1. Synthesis of diphenylphosphane-functionalised carbosil-
ane dendrimers with a methylene spacer between Si atom and P
atom
To obtain dendrimers with two carbon atoms between
the terminal silicon atom and the phosphorus atom, carbo-
silane dendrimers (G0, G1) were hydrosilylated with (chloro-
methyl)dimethylsilane, followed by treatment with
Ph2PCH2Li·TMEDA (Scheme 2).
Figure 1. Crystal structure of the second generation dendrimer 3
(a) and a modelled structure[58] of the (allyl)Pd complex of 13 (b);
hydrogen atoms and counter ions have been omitted for clarity
Dendrimers 12 and 13 have four and twelve end groups,
respectively, each containing an SiMe(CH2PPh2)2 bis(phos-
phane) ligand. These units can be viewed as analogues of
Me2Si(CH2PPh2)2 and dppp [ϭ bis(diphenylphosphanyl)-
Results and Discussion
Synthesis of the Dendrimeric Ligands
The phosphane-functionalised carbosilane dendrimers propane]. The dendrimers 9, 10, 11, 19 and 20 have end
were synthesised by hydrosilylation of the double bonds of groups with SiMe2(CH2)yPPh2 (y ϭ 1, 2) phosphane li-
the various generations (G0, G1, G2) with chlorodimethylsi- gands, and bis(coordinated) metal complexes can only be
lane or dichloro(methyl)silane, followed by treatment with formed between the phosphane units of different end
the tetramethylethylenediamine (TMEDA) complex of [(di- groups.
1086
Eur. J. Org. Chem. 2002, 1085Ϫ1095

Palladium Complexes of Phosphane-Functionalised Carbosilane Dendrimers
FULL PAPER
Scheme 3. Pd-catalysed allylic alkylation with sodium diethyl 2-
methylmalonate as the nucleophile
Scheme 2. Synthesis of diphenylphosphane-functionalised carbosil-
ane dendrimers with an ethylene spacer between the Si atom and
the P atom
The catalysts possessing dendrimeric ligands with large
bridges between the P atoms (9 and 11) had very low activ-
ity compared to the catalysts containing 12 and 13 as li-
gands (Table 1). The selectivity towards the linear trans
product was the same for all catalysts, and was also similar
to that induced by palladium complexes of dppp [ϭ bis(di-
phenylphosphanyl)propane] or dppb [ϭ bis(diphenylphos-
phanyl)butane].[64] The second generation dendrimer 13
gave a slightly slower catalyst for the reaction with crotyl
acetate than did its smaller analogue 12. For cinnamyl acet-
ate these dendrimeric catalysts showed similar activities.
Formation of the Palladium Complexes
In order to investigate complex formation between the
dendrimeric ligand and the palladium ion, the phosphane-
functionalised dendrimers were mixed with solutions of
(COD)MePdCl (COD ϭ 1,5-cyclooctadiene) to give the
(dendrimer)MePdCl complexes (P/Pd ϭ 2). The dendrimers
9, 10 and 11 formed trans complexes according to 31P{1H}
NMR spectroscopy, as singlets were observed at δ ϭ 19.
These signals were broader for the higher generations [the
signal for 11-(PdMeCl)18 was more than 165 Hz wide],
which was probably a result of the many different con-
formations of the larger system and the different PϪPdϪP
rings formed. Dendrimers 12 and 13 coordinated in a cis
fashion, similarly to the complexes formed by dppp.[63] The
two doublets in the 31P{1H} NMR spectrum of 12-
(PdMeCl)4 (at δ ϭ 31.3 and 2.75; J ϭ 39.0 Hz) clearly
showed the coupling of the inequivalent phosphorus atoms.
For 13-(PdMeCl)12 the coupling between these atoms could
not be observed, because the signals had broadened signi-
ficantly. However, the two signals found in the 31P{1H}
NMR spectrum at δ ϭ 30.8 and 2.38 suggested that cis
complexes were formed.
Table 1. Pd-catalysed allylic alkylation of crotyl acetate or cinnamyl
acetate and sodium diethyl 2-methylmalonate with (allyl)Pd com-
plexes of 9, 11, 12, 13
Substrate
Ligand Conversion trans cis Branched
(%)
(%)
(%) (%)
Crotyl acetate[a]
9
11
12
13
2
2
30
22
80
78
80
80
5
6
5
5
15
16
15
15
Cinnamyl acetate[b]
9
11
12
13
0.6
0.6
10
99
100
97
0
0
0
0
1
0
3
4
10
96
[a]
(Allyl)palladium complexes were synthesised by treat-
Room temperature; solvent: THF; volume: 5 mL; [crotyl acet-
ment of the dendrimers with [(η3-C3H7)PdCl]2. According ate] ϭ 0.2 ; [sodium diethyl 2-methylmalonate] ϭ 0.1 ; [Pd] ϭ
0.05 m; conversion after 2 h. [b] Room temperature; solvent: THF;
volume: 20 mL; [cinnamyl acetate] ϭ 46 m; [sodium diethyl 2-
methylmalonate] ϭ 25 m; [Pd] ϭ 12.5 µ; conversion after 1 h.
1
to the H and 31P{1H} NMR spectra, the phosphane den-
drimers 12 and 13 coordinated in a bidentate fashion, re-
sulting in well-defined Pd complexes (both products gave
broad signals at δ ϭ 11.1 in their 31P{1H} NMR spectra),
whereas the monodentate phosphane dendrimers 9 and 11
gave rise to mixtures of products (as indicated by the pres-
ence of several signals in the 31P{1H} NMR spectra).
Sodium acetate, which is not soluble in THF, was formed
as a side-product during the above alkylation reaction. The
precipitate formed might complicate the nanofiltration pro-
cess if the reaction were performed in a continuous-flow
membrane reactor. If allyl trifluoroacetate were used as a
substrate, soluble sodium trifluoroacetate would be formed,
which would prevent these potential complications. We
studied this substrate in a batch process, and all the den-
drimeric catalysts showed very high activity. With a sub-
strate/Pd ratio of 2000, the yield after 5 min was approxim-
Catalysis
The Allylic Alkylation Reaction
Batch Process
The (allyl)Pd complexes of dendrimers 9, 11, 12 and 13 ately 50% and only small differences in reaction rates were
were used as catalysts in the allylic alkylation[41] of substi- observed for the different generations: The degrees of con-
tuted allyl acetates with the sodium salt of diethyl 2-methyl- version after 5 min were 49%, 55%, 45% and 47% with li-
malonate (Scheme 3).
gands 9, 11, 12 and 13, respectively. The activity did not
Eur. J. Org. Chem. 2002, 1085Ϫ1095
1087

D. de Groot, J. N. H. Reek, P. C. J. Kamer, P. W. N. M. van Leeuwen
FULL PAPER
decrease with increasing generation, indicating that all act- lysed for palladium. ICP-AES showed that all the palla-
ive sites were acting as independent catalysts. From molecu- dium had passed through the membrane.
lar modelling (Figure 1b) it is clear that all the (allyl)Pd
groups indeed reside at the outer surface of the dendrimer in sharp contrast with the stability of the catalyst when ap-
and should be easily accessible to the nucleophile. plied in a batch reaction. This suggested that the deactiva-
The deactivation observed in the continuous process was
Addition of a second portion of substrate after a degree tion was induced either by interactions with the membrane
of conversion of over 90% had been reached showed that or by leaching of small inactive palladium species. The for-
the catalyst remained active. Turnover numbers of over mer explanation can be ruled out, since batch experiments
15,000 were reached in the batch process with 13-[Pd(al- performed in the presence of pieces of membrane material
lyl)Cl]12, indicating that this is a stable dendrimeric catalyst gave results similar to those without membrane. Thus, one
that might perform very well in a continuous process if a of the intermediate palladium complexes of the catalytic
membrane reactor was used.
cycle was less strongly bound to the phosphane ligand and
leached out of the reactor as a complex that showed little
or no activity.
Continuous Process
For the continuous process, a solution of allyl trifluo-
roacetate and sodium diethyl 2-methylmalonate in THF
Allylic Amination
(with n-decane as an internal standard) was pumped Batch Process
through the reactor and 13-[Pd(allyl)Cl]12 was used as a
We were interested in the effect of the reaction conditions
catalyst.
on the performance of these dendrimeric catalysts in the
continuous-flow membrane reactor. An allylic substitution
reaction was performed with an amine as the nucleophile
and dichloromethane as the solvent. When the
(allyl)PdϪdendrimer complexes of 9, 11, 12 and 13 were
used as catalysts in the allylic amination reaction[42] be-
tween crotyl acetate and piperidine (Scheme 4, Table 2),
dendrimers 12 and 13 gave more active catalysts than 9 and
11 did, as was also observed for the allylic alkylation reac-
tion between crotyl acetate with sodium diethyl 2-methyl-
malonate. The selectivity induced by the dendrimers with
SiMe(CH2PPh2)2 end groups (12 and 13) was slightly differ-
In Figure 2 the yield is plotted as a function of the
amount of substrate solution (expressed in reactor volumes)
pumped through the reactor. The reaction started immedi-
ately after addition of the catalyst and reached its maximum
yield after two reactor volumes. After this point the yield
dropped and was nearly zero after approximately fifteen
times the reactor volume of substrate solution had been
pumped through the reactor. In view of the size of the sys-
tem and the stability of the catalyst in batch reactions, this
decrease was unexpectedly rapid. The retention of the den-
drimeric catalyst was determined to be 99.7% in dichloro-
methane (vide infra). On the basis of this retention, the
yield should have decreased by less than 5%. The retention
of the dendrimeric catalyst in THF has not been deter-
mined, but the retention of the second generation den-
drimer 3, which is much smaller than the catalyst, was
98.1% with THF as a solvent. The observed rapid decrease
in yield was therefore ascribed to decomposition of the pal-
ladium catalyst and not to leaching of the dendrimer. This
was in agreement with the observation that samples taken
from the product flow were catalytically inactive, indicating
that no active palladium catalyst had passed through the
membrane. When the experiment was stopped after 39 re-
actor volumes of substrate solution had been flushed
through the reactor, the contents of the reactor and the so-
lution that had passed through the membrane were ana-
Scheme 4. Pd-catalysed allylic amination with piperidine as the nu-
cleophile
Table 2. Pd-catalysed allylic amination of crotyl acetate and piper-
idine with (allyl)Pd complexes of 9, 11, 12, 13
Ligand
Conversion[a][b]
(%)
trans
(%)
cis
(%)
Branched
(%)
9
11
12
13
26
24
31
42
44
6
5
5
5
70
64
53
51
26
59[c]
79
[a] Room temperature; solvent: CH2Cl2; volume: 5 mL; [crotyl acet-
Figure 2. Application of dendrimeric ligand 13 in the allylic al-
kylation of allyl trifluoroacetate and sodium diethyl 2-methylma-
lonate in a continuous-flow membrane reactor (yield in % conver-
sion of allyl trifluoroacetate in the product stream)
[b]
ate] ϭ 0.12 ; [piperidine] ϭ 0.25 ; [Pd] ϭ 2 m.
Conversion
[c]
after 1 h. The conversion is slightly lower due to partly decom-
posed catalyst.
1088
Eur. J. Org. Chem. 2002, 1085Ϫ1095

Palladium Complexes of Phosphane-Functionalised Carbosilane Dendrimers
FULL PAPER
ent from that generated by the dendrimers with Si- catalyst in the Pd0 state. By addition of 1 equiv. of di-
Me2CH2PPh2 end groups (9 and 11).
ethylamine[65] we expected that, after nucleophilic attack,
the PdII would have changed to Pd0. Palladium and phos-
phorus analysis revealed a retention of 99.5%. The retention
was still as high as 98.5% upon changing the solvent to a
mixture of diethylamine and dichloromethane.
Apparently, the catalyst was stable under conditions very
similar to those of the catalysis experiments; only the allyl
acetate substrate was absent. We therefore conclude that the
decomposition was stimulated by the presence of the allyl
acetate. One possible deactivation pathway is through the
formation of (allyl)Pd acetate complexes. Another possibil-
ity is that the alkene product or the alkene starting material
might be stabilising the zerovalent (alkene)(PϪP)Pd inter-
mediate and causing the formation of (alkene)2Pd com-
plexes.
Continuous Process
When palladium complexes of 13 were applied in a con-
tinuous process, deactivation of the catalyst during the ex-
periment was again observed (Figure 3). The reaction
started immediately after injection of the catalyst, and the
maximum yield was reached after approximately three re-
actor volumes of substrate solution had been pumped
through the reactor. The activity of the catalytic system de-
creased even more rapidly than that of the continuous al-
lylic alkylation in THF.
We also prepared dendrimers with ethylene spacers be-
tween the terminal silicon atom and the phosphorus atom
(19 and 20). The larger dendrimer (20) was applied as a
ligand in the continuous allylic amination reaction. Figure 4
shows that the maximum yield was reached after approxim-
ately five reactor volumes of substrate solution had been
pumped through the reactor. To our surprise, this dendrim-
eric catalyst was much more stable, and the formation of
product was fairly constant during the next ten reactor vol-
umes. After this period the degree of conversion was still
Figure 3. Application of dendrimeric ligand 13 in the allylic amin-
ation of crotyl acetate and piperidine in a continuous-flow mem-
brane reactor (yield in % conversion of crotyl acetate in the prod-
uct stream)
The retention of the second generation dendrimer (3) was more than 70% of the maximum reached. This decrease
high, also when the substrate solution was used as the solv- would correspond to a retention of the dendrimeric catalyst
ent (98.2% versus 99.7% for pure CH2Cl2). Therefore, wash- of more than 98%.
ing out of the dendrimeric system cannot account for the
fast decrease in catalytic activity during the continuous pro-
cess. Brinkmann and co-workers reported deactivation of
the palladium catalyst in a continuous allylic amination re-
action and observed the formation of an inactive (li-
gand)PdCl2 complex, probably resulting from a reaction
with dichloromethane.[38] To study this reaction, a model
compound {(CH3)2Si(CH2PPh2)2}(η3-C3H7)PdCl was pre-
pared by treatment of (CH3)2Si(CH2PPh2)2 (14) and 0.5
equiv. of [(η3-C3H7)PdCl]2. Decomposition of this complex
was monitored by 1H and 31P{1H} NMR spectroscopy, and
the formation of {(CH3)2Si(CH2PPh2)2}PdCl2 was ob-
served. However, {(CH3)2Si(CH2PPh2)2}PdCl2 was only
Figure 4. Application of dendrimeric ligand 20 in the allylic amin-
ation of crotyl acetate and piperidine in a continuous-flow mem-
brane reactor (P/Pd ϭ 2, yield in % conversion of crotyl acetate in
the product stream)
In order to increase the activity and stability of the cata-
slightly less active than {(CH3)2Si(CH2PPh2)2}(η3- lyst even further, the P/Pd ratio was increased from 2 to
C3H7)PdCl in the allylic amination of crotyl acetate and 4.[66] The batchwise allylic amination reaction between cro-
piperidine. Complete conversion was reached after 2 h. tyl acetate and piperidine with dendrimeric ligand 20 and
Therefore, the formation of {(CH3)2Si(CH2PPh2)2}PdCl2 [(crotyl)PdCl]2 as palladium precursor, with a P/Pd ratio of
by reaction with the solvent cannot account for the fast 4, appeared to be very fast: Over 65% conversion was ob-
deactivation of the catalyst.
tained after 5 min under the same conditions as for the
To study other possible deactivation pathways, we ex- other dendrimeric ligands. This was nearly five times as fast
plored the stability of the 13-[Pd(allyl)Cl)]12 catalyst system as for 13-[Pd(allyl)Cl]12 with a P/Pd ratio of 2, and 25 times
by performing retention measurements in dichloromethane. faster than for 9-[Pd(allyl)Cl]4 or 11-[Pd(allyl)Cl]36 (P/Pd ϭ
The reactor was flushed with 10 reactor volumes and the 2). We also performed the continuous process with a P/Pd
amounts of palladium and phosphorus on both sides of the ratio of 4 (Figure 5). The maximum yield obtained during
membrane were determined by ICP-AES. The retention of the reaction was higher for P/Pd ϭ 4 than for P/Pd ϭ 2,
13-[Pd(allyl)Cl]12 was 99.7%, indicating that the catalyst, showing that the catalyst was also more active under con-
which was in the PdII state, was stable under these condi- tinuous-flow conditions, since the palladium concentration
tions. During the catalytic cycle the catalyst also becomes was the same in both experiments. In addition, the stability
Pd0, and so we attempted a retention measurement of the of the catalysts was very high. The slight decrease in yield
Eur. J. Org. Chem. 2002, 1085Ϫ1095
1089

D. de Groot, J. N. H. Reek, P. C. J. Kamer, P. W. N. M. van Leeuwen
FULL PAPER
during this experiment (P/Pd ϭ 4) is attributed to very dendrimeric catalysts were sufficiently large to enable their
small amounts of dendrimeric catalyst passing through the application in a continuous-flow membrane reactor, as in-
membrane. From the curve in Figure 5, the retention of the dicated by the retention of 99.7% for the dendrimeric cata-
dendrimeric complex was estimated to be 98.5Ϫ99%, which lyst 13-[Pd(allyl)Cl]12 in dichloromethane. When dendrim-
is indeed in the range of the expected values. In contrast to eric ligands with Si(CH2PPh2)2 end groups were used, the
13-[Pd(allyl)Cl]12, the 20-[Pd(allyl)Cl]3 catalyst did not show yields for allylic substitution reactions dropped rapidly dur-
decomposition during the continuous allylic amination re- ing continuous experiments, due to catalyst deactivation.
action. The dendrimeric ligands with ethylene spacers be- When dendrimeric ligands with SiCH2CH2PPh2 end groups
tween the terminal Si atom and P atom give catalysts that were used, stable catalysts were obtained, that were success-
are more stable in the continuous process than the dendrim- fully applied in the continuous-flow membrane reactor.
eric ligands with methylene spacers between Si and P. This During the experiment the yield of the reaction decreased
might be due to the higher stability of the SiCH2CH2P unit to some extent, which was in line with a retention of
compared to SiCH2P (we have observed decomposition of 98.5Ϫ99%. The important message is that small changes in
this type of dendrimer to yield Ph2PCH3; see Exp. Sect.). the dendrimeric ligand had a large impact on the catalyst
More probably, the SiCH2CH2PPh2 dendrimers give stability. High catalyst stability is pivotal for successful ap-
more stable Pd complexes than the Si(CH2PPh2)2 dendri- plication in continuous-flow membrane reactors.
mers. If the reason for decomposition lies in an equilibrium
between coordination of the phosphane ligand in the zero-
Experimental Section
valent (alkene)Pd complex and stabilisation of this complex
by another alkene, then ligand 20 might possibly favour the
coordinated complex, while the equilibrium might be
shifted in the opposite direction for ligand 13.
General Data: All reactions were carried out under purified nitro-
gen by standard Schlenk techniques. Solvents were distilled under
N2 from sodium/benzophenone (THF, diethyl ether, hexane, pent-
ane) or calcium hydride (dichloromethane) prior to use. TMEDA
(N,N,NЈ,NЈ-tetramethylethylenediamine) was distilled from n-butyl-
lithium. Chemicals were purchased from Aldrich Chemical Co. and
Acros Chimica and were used without further purification. Silica
60 (SDS Chromagel, 70Ϫ200 µm) was used for filtration of the
reaction mixtures. The various generations of dendrimers (1, 2, 3,
15 and 16) were prepared as described by Van der Made et al.[52]
Tetrabutylammonium hexachloroplatinate,[59] methyldiphenylphos-
phane,[67] the tetramethylethylenediamine complex of (diphenyl-
Figure 5. Application of dendrimeric ligand 20 in the allylic amin-
ation of crotyl acetate and piperidine in a continuous-flow mem-
brane reactor (P/Pd ϭ 4, yield in % conversion of crotyl acetate in
the product stream)
phosphanyl)methyllithium,[60]
(COD)MePdCl,[68]
[(η3-
C4H7)(COD)Pd][BF4],[69] [(crotyl)PdCl]2
and [(allyl)PdCl]2
[70]
[70]
were prepared according to literature procedures. H and 31P{1H}
NMR spectra were recorded with a Bruker AMX 300 and a Varian
Mercury 300. 13C{1H} NMR spectra were measured with a Bruker
AMX 300 and a Varian Inova 500. The chemical shifts are given
in ppm relative to TMS for 1H and 13C NMR and relative to
H3PO4 for 31P NMR. Matrix Assisted Laser Desorption Ionisation
(MALDI) Time-of-Flight (TOF) mass spectrometry was performed
1
In general, the stability of the catalyst is very important
when considering its application in a continuous process.
We have shown that small changes in the periphery of the
dendrimer can have a huge impact on the stability of these
catalytic systems, possibly through the formation of more
stable palladium complexes.
with
a PerkinϪElmer/PerSeptive Biosystems Voyager-DE-RP
MALDI-TOF mass spectrometer (PerSeptive Biosystems, Inc.,
Framingham, MA, USA) equipped with delayed extraction. A 337-
nm UV nitrogen laser producing 3-ns pulses was used, and the
mass spectra were obtained in the linear and reflectron mode.
Samples were prepared in an Atmosbag (Aldrich) filled with argon,
by mixing 10 µL of dichloromethane solution of the sample with
30 µL of a solution of 3 mg/L 2,5-dihydroxybenzoic acid (DHB) or
Dithranol (DIT) in dichloromethane; 1 µL of the solution was
Conclusions
We have described several routes to phosphane-func-
tionalised carbosilane dendrimers, resulting in systems
that differ in the number of phosphane units per silicon
atom of the end group [SiCH3(CH2PPh2)2 versus loaded onto a gold sample plate, the solvent was removed, and the
sample was transferred to the vacuum of the mass spectrometer for
analysis. Fast atom bombardment (FAB) mass spectrometry was
carried out with a JEOL SX/SX 102A (Tokyo, Japan) four-sector
tandem mass spectrometer (B1E1B2E2 geometry), coupled to a
JEOL MS/MP9021D/UPD data system. The samples, in nitro-
benzyl alcohol solution, were loaded onto a stainless steel probe
and bombarded with xenon atoms with an energy of 8 keV. A re-
solving power of 5000Ϫ10000 (10% valley definition) was used dur-
ing the high-resolution FAB-MS measurements. Polyethylene glycol
(PEG) 300 and 600 were used to calibrate the mass spectrometer.
Si(CH3)2CH2PPh2] and the number of carbon atoms be-
tween the phosphane unit and silicon end groups
[Si(CH3)2CH2PPh2 versus Si(CH3)2CH2CH2PPh2]. Palla-
dium complexes of phosphane-functionalised carbosilane
dendrimers were active as catalysts in the allylic substitu-
tion reaction and, as expected, the number of phosphane
units per end group had an impact on the activity and se-
lectivity in the allylic amination reaction. Interestingly, the
activity was hardly influenced by the dendrimer size, mak-
ing these dendrimers attractive as catalyst supports. These Elemental analyses were measured with an Elementar Vario EL
1090 Eur. J. Org. Chem. 2002, 1085Ϫ1095

Palladium Complexes of Phosphane-Functionalised Carbosilane Dendrimers
apparatus. The amounts of palladium and phosphorus were deter-
cedure as for 7, to yield a white solid (0.562 g, 0.287 mmol, 100%).
FULL PAPER
mined with induced coupled argon plasma Ϫ atomic emission spec- 1H NMR (300 MHz, CDCl3): δ ϭ 0.80 (s, 36 H, SiCH3), 0.46
trometry (ICP-AES). The ICP-AES measurements were performed
0.77Ϫ0.70, 1.01Ϫ0.95 (m, 64 H, SiCH2CH2Si). 13C{1H} NMR
(75 MHz, CDCl3): δ ϭ 1.5 (s, SiCH3), 4.2 (s, SiCH2CH2SiCl),14.2
by a literature procedure,[71] with a sequential Jarrell Ash upgraded
(Model 25) Atomscan model 2400 ICP scanning monochromator (s, SiCH2CH2SiCl).
and a PerkinϪElmer Optima 3000 XL instrument. The measured
Typical Procedure for the Preparation of the Diphenylphosphane-
Terminated Dendrimers
atomic lines of Pd and P were 340.458 nm and 213.618 nm, respect-
ively. Gas chromatography was performed with an Interscience HR
GC Mega 2 apparatus (split/splitless injector, J&W Scientific, DB1
30-m column, film thickness 3.0 mm, carrier gas 70 kPa He, F.I.D.
detector). The GPC measurements were performed with a Shim-
adzu apparatus equipped with a Waters Styragel Column HR 1,
HR 2 and HR 4 in series and with a RID-10A refractive index
detector and a SPD-10A VP UV/Vis detector. The membrane re-
actor set-up contained a Gilson Piston Pump Model 303/10SC and
a Gilson Model 802C Manometric Module. The reactor volume
was 5 mL and the diameter of the membrane was 23 mm. A Koch/
SelRO MPF-60 NF membrane (Koch Membrane Systems, Düssel-
dorf, Germany) with a molecular weight cut-off (MWCO) of 400
Dalton was used for the continuous experiments.
Si[CH2CH2Si(CH3)2CH2PPh2]4
(9):
A
solution
of
Ph2PCH2Li·TMEDA (2.43 g, 7.54 mmol) in THF (40 mL) was ad-
ded to 4 (0.7565 g, 1.47 mmol). After stirring overnight, the reac-
tion mixture was filtered through silica, and the excess of Ph2PCH3
was removed in vacuo (100 °C, 10Ϫ5 mbar, overnight). The slightly
yellow liquid was obtained in 97% yield (1.67 g, 1.43 mmol). 1H
NMR (300 MHz, CDCl3): δ ϭ Ϫ0.10 (s, 24 H, SiCH3), 0.26 (s, 16
H, SiCH2CH2Si), 1.40 (s, 8 H, SiCH2P), 7.3 (m, 24 H, ArH), 7.5
(m, 16 H, ArH). 31P{1H} NMR (121.5 MHz, CDCl3): δ ϭ Ϫ20.8
(s). MALDI-TOF-MS: m/z ϭ 1170 [Mϩ] (calcd. 1169.8), 986 [Mϩ
Ϫ CH2PPh2 ϩ OH]. C68H88P4Si5: calcd. C 69.81, H 7.59; found C
69.33, H 7.59.
Typical Procedure for the Preparation of the Chloro-Terminated
Dendrimers
Estimation of the Purity: The vinyl-functionalised dendrimers were
all purified by silica column chromatography, and these compounds
were considered to be pure starting materials. The hydrosilylation
Si[CH2CH2Si(CH3)2Cl]4 (4): The reaction was performed under
N2. A few drops of a concentrated solution of [Bu4N]2[PtCl6] in
technical ethanol were added to a solution of tetravinylsilane (1;
0.526 g, 3.86 mmol) in chlorodimethylsilane. After this had been
stirred overnight, the solvent was evaporated. Product 4 was ob-
1
reaction was quantitative, as evidenced by H NMR (no vinyl sig-
nals were visible in any of the spectra). The reaction between the
chlorosilane-terminated dendrimers and the lithium reagents
{[(diphenylphosphane)methyl]lithium} was monitored by 1H
NMR, by monitoring the Si(CH3)2Cl signals (δ ϭ 0.8), which were
shifted to δ ϭ Ϫ0.10 [Ϫ0.20 for SiMe(CH2PPh2)2] after func-
tionalisation. Unchanged Si(CH3)2Cl gave a clear peak at δ ϭ 0.04
[δ ϭ Ϫ0.13 for SiMeOR(CH2PPh2)] after hydrolysis. According to
these NMR spectra at least 95% of the end groups had been func-
tionalised with phosphanes, and the remaining end groups were
hydrolysed chlorosilanes, giving Si(CH3)2OR. When dendrimers 9,
10, 11, 12 and 13 and model compound 14 were stored in glass
Schlenk flasks, either at room temperature or at Ϫ20 °C and with
or without exclusion of light, slow decomposition took place to
yield Ph2PCH3 and an insoluble polymeric substance.
1
tained as a white solid in 92% yield (1.82 g, 3.53 mmol). H NMR
(300 MHz, CDCl3): δ ϭ 0.7Ϫ0.5 (m, 16 H, SiCH2CH2Si), 0.42 (s,
24 H, SiCH3). 13C{1H} NMR (75 MHz, CDCl3): δ ϭ 2.34 (s,
SiCH3), 3.47 (s, SiCH2CH2SiCl), 12.7 (s, SiCH2CH2SiCl). No fur-
ther purification was performed, and product 4 was converted im-
mediately into 9.
Si{CH2CH2Si[CH2CH2Si(CH3)2Cl]3}4 (5): This compound was
prepared from 2 (0.510 g, 0.883 mmol) according to the same pro-
cedure as for 4, to yield a white solid (1.52 g, 0.887 mmol, 100%).
1H NMR (300 MHz, CDCl3):
δ ϭ 0.7Ϫ0.5 (m, 64 H,
SiCH2CH2Si), 0.42 (s, 72 H, SiCH3). 13C{1H} NMR (75 MHz,
CDCl3): δ ϭ 2.40 (s, SiCH3), 3.37 (s, SiCH2CH2SiCl), 12.9 (s,
SiCH2CH2SiCl).
Si{CH2CH2Si[CH2CH2Si(CH3)2CH2PPh2]3}4 (10): This compound
was
Ph2PCH2Li·TMEDA (3.38 g, 10.49 mmol) according to the same
procedure as for 9, to yield slightly yellow oil (0.944 g,
prepared
from
5
(0.5355 g,
0.313 mmol)
and
Si(CH2CH2Si{CH2CH2Si[CH2CH2Si(CH3)2Cl]3}3)4 (6): This com-
pound was prepared from 3 (0.322 g, 0.170 mmol) according to the
same procedure as for 4 (a longer reaction time was necessary; con-
version was monitored by 1H NMR), to yield a white solid (0.676 g,
a
1
0.257 mmol, 82%). H NMR (300 MHz, CDCl3): δ ϭ Ϫ0.20 (br. s,
72 H, SiCH3), 0.28 (s, 64 H, SiCH2CH2Si), 1.28 (br. s, 24 H,
SiCH2P), 7.2 (m, 72 H, ArH), 7.4 (m, 48 H, ArH). 31P{1H} NMR
(121.5 MHz, CDCl3): δ ϭ Ϫ21.7 (s). MALDI-TOF-MS: m/z ϭ
3682 [Mϩ] (calcd. 3677.7), 3498 [Mϩ Ϫ CH2PPh2 ϩ OH], 3322
[Mϩ Ϫ 2 CH2PPh2 ϩ 2 OH]. C212H280P12Si17: calcd. C 69.23, H
7.68; found C 68.51, H 7.66.
1
0.127 mmol, 75%). H NMR (300 MHz, CDCl3): δ ϭ 0.7Ϫ0.4 (m,
208 H, SiCH2CH2Si), 0.42 (s, 216 H, SiCH3). 13C{1H} NMR
(75 MHz, CDCl3): δ ϭ 2.69 (s, SiCH3), 3.41 (s, SiCH2CH2SiCl),
12.9 (s, SiCH2CH2SiCl).
Si[CH2CH2Si(CH3)Cl2]4 (7): The reaction was performed under
N2. A few drops of a concentrated solution of [Bu4N]2[PtCl6] in
technical ethanol were added to a solution of tetravinylsilane (1;
0.138 g, 1.01 mmol) in dichloromethylsilane. After this had been
stirred overnight, the solvent was evaporated. Product 7 was ob-
tained as a white solid in 97% yield (0.587 g, 0.985 mmol). 1H
NMR (300 MHz, CDCl3): δ ϭ 0.81 (s, 12 H, SiCH3), 0.78Ϫ0.71
(m, 8 H, SiCH2CH2Si), 1.01Ϫ0.95 (m, 8 H, SiCH2CH2Si). 13C{1H}
NMR (75 MHz, CDCl3): δ ϭ 1.5 (s, SiCH3), 4.2 (s, SiCH2CH2S-
iCl), 13.9 (s, SiCH2CH2SiCl). No further purification was per-
formed, and product 7 was converted immediately into 12.
Si(CH2CH2Si{CH2CH2Si[CH2CH2Si(CH3)2CH2PPh2]3}3)4 (11):
This compound was prepared from 6 (0.4625 g, 0.0872 mmol) and
Ph2PCH2Li·TMEDA (2.69 g, 8.33 mmol) according to the same
procedure as for 9, to yield a slightly yellow oil (0.8858 g,
1
0.0791 mmol, 91%). H NMR (300 MHz, CDCl3): δ ϭ Ϫ0.25 (br.
s, 216 H, SiCH3), 0.31 (br. s, 208 H, SiCH2CH2Si), 1.25 (br. s, 72 H,
SiCH2P), 7.5Ϫ7.0 (m, 360 H, ArH). 31P{1H} NMR (121.5 MHz,
CDCl3): δ ϭ Ϫ22.0 (s). MALDI-TOF-MS: m/z ϭ 8000Ϫ12000
(very broad) (calcd. 11201.6). C644H856P36Si53: calcd. C 69.05, H
7.71; found C 68.20, H 7.70.
Si{CH2CH2Si[CH2CH2Si(CH3)Cl2]3}4 (8): This compound was Si[CH2CH2SiCH3(CH2PPh2)2]4 (12): This compound was prepared
prepared from 3 (0.162 g, 0.281 mmol) according to the same pro-
from 7 (0.5716 g, 0.958 mmol) and Ph2PCH2Li·TMEDA (6.44 g,
Eur. J. Org. Chem. 2002, 1085Ϫ1095
1091

D. de Groot, J. N. H. Reek, P. C. J. Kamer, P. W. N. M. van Leeuwen
NMR (121.5 MHz, CDCl3)
Ϫ9.03. 13C{1H} NMR
FULL PAPER
20.0 mmol) according to the same procedure as for 9, to yield a
δ ϭ
yellow liquid (1.098 g, 0.576 mmol, 60%). 1H NMR (300 MHz, (125.8 MHz, CDCl3): δ ϭ Ϫ1.6 (s, SiCH3), 11.0 (d, SiCH2CH2P,
CDCl3): δ ϭ Ϫ0.29 (br. s, 12 H, SiCH3), 0.19 (m, 16 H, 2JC,P ϭ 9.3 Hz), 17.4 (s, CH2 of SiCH2CH2CH2Si), 18.5 (s, CH2
SiCH2CH2Si), 1.26 (br. s, 16 H, SiCH2P), 7.5Ϫ7.1 (m, 80 H, ArH). of SiCH2CH2CH2Si), 19.8 (s, CH2 of SiCH2CH2CH2Si), 21.7 (d,
1
31P{1H} NMR (121.5 MHz, CDCl3): δ ϭ Ϫ22.5 (s). MALDI-
SiCH2CH2P, JC,P ϭ 13.5 Hz), 128.5 (s, p-PhP), 128.3 (d, m-PhP,
TOF-MS: m/z ϭ 1907 [Mϩ] (calcd. 1906.5), 1707 [Mϩ Ϫ CH2PPh2, 3JC,P ϭ 6.3 Hz), 132.7 (d, o-PhP, JC,P ϭ 18.2 Hz), 138.8 (d, ipso-
2
fragmentation]. C116H124P8Si5: calcd. C 73.08, H 6.56; found C
72.59, H 6.75.
PhP, 1JC,P ϭ 14.2 Hz). FAB-MS: m/z ϭ 1283 [M ϩ H]ϩ, 1206 [Mϩ
Ϫ Ph], 1098 [Mϩ Ϫ PPh2], 1085 [Mϩ Ϫ CH2PPh2], 1069 [Mϩ
Ϫ
CH2PPh2-CH3]. MALDI-TOF-MS: m/z ϭ 1280.7 [Mϩ] (calcd.
1282.0), 1096 [Mϩ Ϫ PPh2, fragmentation]. C76H104P4Si5: calcd. C
71.20, H 8.18; found C 70.48, H 8.29.
Si{CH2CH2Si[CH2CH2SiCH3(CH2PPh2)2]3}4 (13): This compound
was
prepared
from
8
(0.5628 g,
0.288 mmol)
and
Ph2PCH2Li·TMEDA (5.11 g, 15.9 mmol) according to the same
procedure as for 9, to yield a yellow oil (1.368 g, 0.232 mmol, 81%).
1H NMR (300 MHz, CDCl3): δ ϭ Ϫ0.38 (br. s, 36 H, SiCH3), 0.35
(m, 64 H, SiCH2CH2Si), 1.12 (br. s, 48 H, SiCH2P), 7.5Ϫ7.0 (m,
240 H, ArH). 31P{1H} NMR (121.5 MHz, CDCl3): δ Ϫϭ 22.6 (s).
MALDI-TOF-MS: m/z ϭ 3000Ϫ6000 (very broad) (calcd. 5887.9).
C356H388P24Si17: calcd. C 72.6, H 6.65; found C 70.08, H 6.82. The
deviation was attributed to incomplete functionalisation of the den-
drimer.
Si{CH2CH2CH2Si[CH2CH2CH2Si(CH3)2CH2CH2PPh2]3}4 (20):
This compound was prepared from 18 (0.2153 g, 0.102 mmol) and
Ph2PCH2Li·TMEDA (0.4386 g, 1.36 mmol) according to the same
procedure as for 19, to yield
a colourless oil (0.1512 g,
0.0371 mmol, 36%, some product probably remained on the silica
1
during the filtration). H NMR (300 MHz, CDCl3): δ ϭ Ϫ0.09 (s,
24 H, SiCH3), 0.51 (m, 88 H, SiCH2CH2CH2Si ϩ SiCH2CH2P),
1.20 (m, 32 H, SiCH2CH2CH2Si), 1.95 (m, 24 H, SiCH2CH2P),
7.25Ϫ7.41 (m, 120 H, ArH). 31P{1H} NMR (121.5 MHz, CDCl3):
δ ϭ Ϫ9.16. 13C{1H} NMR (125.8 MHz, CDCl3): δ ϭ Ϫ3.2 (s,
(H3C)2Si(CH2PPh2)2 (14): KPPh2 in THF (0.5 , 20 mL, 10 mmol)
was added dropwise at Ϫ70 °C to a solution of (H3C)2Si(CH2Cl)2
(0.7867 g, 5.01 mmol) in THF (2 mL). After stirring for 3 h, during
which the temperature was increased to room temperature, the re-
action mixture was filtered through silica. The excess of PPh2 was
removed in vacuo (room temperature, 10Ϫ5 mbar, overnight). A
white oil was obtained in 64% yield (1.458 g, 3.19 mmol). 1H NMR
(300 MHz, CDCl3) δ ϭ Ϫ0.20 (s, 6 H, SiCH3), 1.27 (d, 4 H,
SiCH2P, 2JHP ϭ 1.2 Hz), 7.27 (m, 12 H, ArH), 7.36 (m, 8 H, ArH).
31P{1H} NMR (121.5 MHz, CDCl3) δ ϭ Ϫ22.8. FAB-MS: m/z ϭ
457.2 [M ϩ H]ϩ, 379.1 [Mϩ Ϫ Ph], 73.0 [Me3Si]. HRMS (FABϩ):
m/z calcd. for C28H31SiP2 [M ϩ H]ϩ: 457.1670; found 457.1676.
C28H30P2Si: calcd. C 73.65, H 6.63; found C 73.18, H 6.63.
2
SiCH3), 11.3 (d, SiCH2CH2P, JC,P ϭ 10 Hz), 17.6 (s, CH2 of
SiCH2CH2CH2Si),18.7 (s, CH2 of SiCH2CH2CH2Si), 20.1 (s, CH2
of SiCH2CH2CH2Si), 21.9 (d, SiCH2CH2P, 1JC,P ϭ 13.8 Hz), 128.6
3
(d, m-PhP, JC,P ϭ 6.8 Hz), 128.7 (s, p-PhP), 132.9 (d, o-PhP,
2JC,P ϭ 17.7 Hz), 138.9 (s, ipso-PhP). MALDI-TOF-MS: m/z ϭ
4064.8 [Mϩ] (calcd. 4070.5), 3880 [Mϩ Ϫ PPh2, fragmentation].
Si[CH2CH2SiCH3(CH2PPh2)2PdMeCl]4:
A
toluene solution
(4 mL) containing (COD)MePdCl (49.4 mg, 0.19 mmol ϭ 0.5
equiv. with respect to the phosphane concentration) was slowly ad-
ded to a solution of 12 (88.2 mg, 0.046 mmol), dissolved in 4 mL
of toluene. The white suspension was stirred for 2.5 h, and, after
filtration and subsequent washing with toluene (2 ϫ 2.5 mL) and
diethyl ether (4 ϫ 4 mL), 95.5 mg of a white product was isolated
(81%). 1H NMR (300 MHz, CDCl3): δ ϭ Ϫ0.41 (br. s, 12 H, Pd
CH3), Ϫ0.37 (br. s, 12 H, SiCH3), 0.57 (br., 16 H, SiCH2CH2Si),
1.24 and 1.61 (br., 16 H, SiCH2P), 7.66, 7.42, 7.19 (br., 80 H, ArH).
31P{1H} NMR (121.5 MHz, [D6]DMSO): δ ϭ 2.75 (d, J ϭ
39.0 Hz), 31.28 (d, J ϭ 39.0 Hz).
Si[CH2CH2CH2Si(CH3)2CH2Cl]4 (17): The reaction was performed
under N2. [Bu4N]2[PtCl6] solution in CH2Cl2 (0.134 , 3.3 µL) was
added to the dendrimer 15 (0.2105 g, 1.09 mmol), followed by
(chloromethyl)dimethylsilane (1.0 mL, 8.2 mmol). The reaction
mixture was stirred overnight at 45 °C. Excess (chloromethyl)dime-
thylsilane was removed in vacuo. Product 17 was obtained as a
1
slightly yellow oil. H NMR (CDCl3): δ ϭ 0.11 (s, 24 H, SiCH3),
Si{CH2CH2Si[CH2CH2SiCH3(CH2PPh2)2PdMeCl]3}4: A toluene
solution (4 mL) containing (COD)MePdCl (44 mg, 0.17 mmol ϭ
0.5 equiv. with respect to the phosphane concentration) was slowly
added to a solution of 13 (83 mg, 0.014 mmol) in 4 mL of toluene,.
The white suspension was stirred for 2 h, and, after filtration and
subsequent washing with toluene (2 ϫ 2.5 mL) and diethyl ether (4
0.58 (m, 8 H, SiCH2CH2CH2Si), 0.71 (m, 8 H, SiCH2CH2CH2Si),
1.33 (m,
8 H, SiCH2CH2CH2Si), 2.78 (s, 8 H, SiCH2Cl).
C24H56Cl4Si5: calcd. C 45.98, H 9.00; found C 45.64, H 9.01.
Si{CH2CH2CH2Si[CH2CH2CH2Si(CH3)2CH2Cl]3}4 (18): This
compound was prepared from 16 (0.809 g, 1.01 mmol) and (chloro-
methyl)dimethylsilane (3.0 mL, 24 mmol) according to the same
procedure as for 17, to yield a slightly yellow oil. 1H NMR
(300 MHz, CDCl3): δ ϭ 0.10 (s, 72 H, SiCH3), 0.59 (m, 32 H,
SiCH2CH2CH2Si), 0.72 (m, 32 H, SiCH2CH2CH2Si), 1.37 (m, 32
H, SiCH2CH2CH2Si), 2.77 (s, 24 H, SiCH2Cl). C84H192Cl12Si17:
calcd. C 47.92, H 9.19; found C 47.82, H 9.14.
1
ϫ 5 mL), 84 mg of a white powder was isolated (77%). H NMR
(300 MHz, [D6]DMSO): δ ϭ Ϫ0.65 (br. s, 36 H, Pd CH3), Ϫ0.4
(br. s, 36 H, SiCH3), 0.03 and 0.27 (br., 64 H, SiCH2CH2Si), 1.2
and 1.75 (br. s, 48 H, SiCH2P), 7.42 and 7.52 (br., 240 H, ArH).
31P{1H} NMR (121.5 MHz, [D6]DMSO): δ ϭ 2.38 (br.), 30.84
(br.).
Si[CH2CH2CH2Si(CH3)2CH2CH2PPh2]4 (19):
A
solution of
Si[CH2CH2SiCH3(CH2PPh2)2Pd(allyl)Cl]4: Compound 12 (62 mg,
Ph2PCH2Li·TMEDA (2.070 g, 6.422 mmol) in 15 mL of THF was 0.033 mmol) and [(allyl)PdCl]2 (24.1 mg, 0.07 mmol ϭ 0.5 equiv.
added dropwise at Ϫ70 °C to
a
solution of 17 (0.4166 g,
with respect to the phosphane concentration) were dissolved in
5 mL of dichloromethane. The yellow solution was stirred for 2 h,
the solvent was evaporated, and the solid was subsequently washed
with diethyl ether (2 ϫ 5 mL) to yield 80 mg of the product (92%).
1H NMR (300 MHz, CDCl3): δ ϭ Ϫ0.59 (br. s, 12 H, SiCH3), 0.08
0.665 mmol) in 5 mL of THF. The reaction mixture was allowed to
warm to room temperature overnight. After filtration through sil-
ica, the solvent was evaporated. The excess of Ph2PCH3 was re-
moved in vacuo (100 °C, 10Ϫ5 mbar, overnight). A slightly yellow
oil was obtained in 84% yield (0.7182 g, 0.560 mmol). 1H NMR and 0.23 (br., 16 H, SiCH2CH2Si), 1.71 (s., 16 H, SiCH2P), 3.56
(300 MHz, CDCl3): δ ϭ Ϫ0.07 (s, 24 H, SiCH3), 0.51 (m, 24 H,
(br., 16 H, allyl-H), 5.56 (br., 4 H, allyl H), 7.22, 7.37, 7.46, 7.68
SiCH2CH2CH2Si ϩ SiCH2CH2P), 1.23 (m, 8 H, SiCH2CH2CH2Si), (br., 80 H, ArH). 31P{1H} NMR (121.5 MHz, CDCl3): δ ϭ 11.1
1.95 (m, 8 H, SiCH2CH2P), 7.25Ϫ7.42 (m, 40 H, ArH). 31P{1H} (br. s).
1092
Eur. J. Org. Chem. 2002, 1085Ϫ1095

Palladium Complexes of Phosphane-Functionalised Carbosilane Dendrimers
Si{CH2CH2Si[CH2CH2SiCH3(CH2PPh2)2Pd(allyl)Cl]3}4: Com-
Et2O mixture. The conversion and product distribution were deter-
FULL PAPER
pound 13 (75 mg, 0.013 mmol) and [(allyl)PdCl]2 (28 mg, mined by GC analysis of the Et2O layer. The allylic alkylations
0.077 mmol ϭ 0.5 equiv. with respect to the phosphane concentra- with allyl trifluoroacetate as a substrate were performed in a total
tion) were dissolved in 5 mL of dichloromethane. The yellow solu- volume of 20 mL. The stock solutions were added to 14 mL of
tion was stirred for 2 h, the solvent was evaporated, and the solid
THF in the following order: 1.0 mL of n-decane solution, 1.0 mL
material was subsequently washed with diethyl ether (2 ϫ 5 mL) to
of sodium diethyl 2-methylmalonate solution, 2.0 mL of allyl tri-
yield 76 mg of the product (74%). 1H NMR (300 MHz, CDCl3): fluoroacetate solution and 2.0 mL of catalyst solution. Samples
δ ϭ Ϫ0.65 (br. s, 36 H, SiCH3), 0.1 and 0.4 (br., 64 H, were quenched in a H2O/Et2O mixture. The conversion was deter-
SiCH2CH2Si), 1.7 (br. s, 48 H, SiCH2P), 3.5 (br., 48 H, allyl-H), mined by GC analysis of the Et2O layer.
5.56 (br., 12 H, allyl H), 7Ϫ7.8 (br., 240 H, ArH). 31P{1H} NMR
Continuous Process: The continuous experiment was performed by
(121.5 MHz, CDCl3): δ ϭ 11.1 (br. s).
use of the reactor set-up described by Eggeling.[73] The membrane
(stored in ethanol) was rinsed with acetone before being transferred
into the membrane reactor. The membrane was first flushed over-
night with THF and then with substrate solution. The substrate
solution was prepared by mixing allyl trifluoroacetate (6.106 g,
39.6 mmol), sodium diethyl 2-methylmalonate (36.6 mL 0.547 ,
20.0 mmol) and n-decane (0.8 mL, 4.1 mmol, internal standard) in
THF (total volume: 800 mL) and was pumped through the reactor
with a flow rate of 43.6 mL/h. The catalyst (0.00025 mmol Pd) was
dissolved in 2 mL of CH2Cl2 and transferred into the membrane
reactor. Samples of the solution coming out of the reactor were
taken continuously, quenched in H2O/Et2O and analysed by GC.
Retention Measurements: The retention measurements with dichlo-
romethane, diethylamine in dichloromethane (0.26 ), or the sub-
strate solution for the allylic amination reactions were measured
with the continuous-flow membrane reactor set-up as described
elsewhere.[72] The retention measurement in THF was performed
by using the reactor set-up as described by Eggeling.[73]
Typical Procedure for a Retention Measurement: The membrane
was cut to the correct size for the reactor and stored in acetone for
one night before being stored in methanol (for at least one night).
After it had been adjusted in the membrane reactor, the membrane
was flushed overnight with CH2Cl2. (When the substrate solution
was used as a solvent, the reactor was first flushed overnight with
CH2Cl2 before flushing with the desired solvent for the amount of
time needed to flush the reactor with two reactor volumes.) The
dendrimer was transferred into the reactor and the solvent was
pumped through for a certain amount of time (e.g., 250 min at a
flow rate of 12 mL/h in order to flush the reactor 10 times). The
contents of the reactor and the solution that had been pumped
through were both then analysed by GPC or NMR. For the reten-
tion measurements of [13-Pd(allyl)Cl], ICP-AES analyses of palla-
dium and phosphorus were performed.
Allylic Amination Experiments: The allylic amination experiments
were performed under N2 at room temperature. The same catalysts
as for the allylic alkylation reactions were used. The catalyst solu-
tion for ligand 20 was prepared by mixing either 20 (0.0142 g,
0.00349 mmol) (for P/Pd ϭ 4) or 20 (0.0072 g, 0.00177 mmol) (for
P/Pd
ϭ
2) and [(crotyl)PdCl]2 (1.965 mg, 0.00499 mmol,
0.00998 mmol Pd) in 2 mL of CH2Cl2 and stirring for 1.5 h.
Batch Process: A substrate solution was prepared by mixing crotyl
acetate (0.351 g, 3.07 mmol), piperidine (0.538 g, 6.32 mmol, fil-
tered through neutral alumina) and n-decane (0.368 g, 2.59 mmol,
internal standard) in CH2Cl2 (total volume: 5.0 mL). The catalyst
(0.0100 mmol Pd) was dissolved in 4.0 mL of CH2Cl2, and 1.0 mL
of substrate solution was added. Samples were quenched in DBA/
Et2O solution (DBA ϭ dibenzylideneacetone). Conversion and
product distribution were determined by GC analysis.
Allylic Alkylation Experiments: The allylic alkylation experiments
were performed under N2 at room temperature. The catalyst was
prepared by mixing the diphenylphosphane-functionalised den-
drimer with [(η3-C3H7)PdCl]2, with CH2Cl2 as a solvent (P/Pd ϭ
2). After this had been stirred for 2 h, the solvent was evaporated
and the resulting yellow solid was washed with diethyl ether. The
sodium diethyl 2-methylmalonate was prepared by treatment of di-
ethyl 2-methylmalonate with sodium hydride. The sodium hydride
(dispersion in oil) was washed with pentane before use. The diethyl
2-methylmalonate (1.03 g, 5.9 mmol) was added dropwise to the
solution of sodium hydride (0.141 g, 5.9 mmol) in THF at 0 °C, to
obtain a 0.5  solution of the sodium diethyl 2-methylmalonate.
Continuous Process: The continuous experiment was performed
with the continuous-flow membrane reactor set-up as described
elsewhere.[72] The membrane was cut to the correct size for the re-
actor and stored in acetone for one night before being stored in
methanol (for at least one night). After it had been transferred into
the membrane reactor, the membrane was first flushed overnight
with CH2Cl2 and then with substrate solution (approximately two
reactor volumes). The substrate solution was prepared by mixing
1.5 mL of crotyl acetate, 2.47 mL of piperidine (after filtration over
neutral alumina) and 2.44 mL of n-decane (as internal standard) in
CH2Cl2 (total volume ϭ 100 mL) and was pumped through the
reactor with a flow rate of 9 mL/h. The catalyst (0.0100 mmol Pd)
was dissolved in 2 mL of CH2Cl2 and transferred into the mem-
brane reactor. Samples of the solution coming out of the reactor
were taken continuously, quenched in DBA/Et2O and analysed by
GC.
Batch Process: Before the reaction, several stock solutions were pre-
pared: a sodium diethyl 2-methylmalonate solution (0.5  in THF),
an n-decane solution (0.04  in THF), a crotyl acetate solution (0.5
 in THF), a cinnamyl acetate solution (0.5  in THF), an allyl
trifluoroacetate solution (0.5  in THF) and a catalyst solution
(0.125 m Pd in THF). The allylic alkylations with crotyl acetate
as a substrate were performed in a total volume of 5 mL. To 2.0 mL
of crotyl acetate solution, 1.0 mL of n-decane solution, 1.0 mL of
sodium diethyl 2-methylmalonate solution and 2.0 mL of catalyst
solution were added. Samples were quenched in an H2O/Et2O mix-
ture. The conversion and product distribution were determined by
GC analysis of the Et2O layer. The allylic alkylations with cinnamyl
acetate as a substrate were performed in a total volume of 20 mL.
Acknowledgments
The stock solutions were added to 14 mL of THF in the following We gratefully acknowledge Prof. Dr. D. Vogt and Dr. E. B. Eggeling
order: 1.0 mL of n-decane solution, 1.0 mL of sodium diethyl 2- from the Eindhoven University of Technology for the use of their
methylmalonate solution, 2.0 mL of cinnamyl acetate solution and continuous-flow membrane reactor for some of the experiments de-
2.0 mL of catalyst solution. Samples were quenched in an H2O/ scribed in this paper and their valuable assistance. We also would
Eur. J. Org. Chem. 2002, 1085Ϫ1095
1093

D. de Groot, J. N. H. Reek, P. C. J. Kamer, P. W. N. M. van Leeuwen
FULL PAPER
[32]
M. Enomoto, T. Aida, J. Synth. Org. Chem. Jpn. 1999, 57,
like to thank R. J. van Haaren for fruitful discussions, H. Peters
and R. Fokkens for the mass spectrometric analyses and J.
Elgersma for the ICP-AES analysis.
32Ϫ42.
[33]
G. E. Oosterom, J. N. H. Reek, P. C. J. Kamer, P. W. N. M.
van Leeuwen, Angew. Chem. Int. Ed. 2001, 40, 1828Ϫ1849.
[34]
D. Seebach, P. B. Rheiner, G. Greiveldinger, T. Butz, H. Sellner,
Top. Curr. Chem. 1998, 197, 125Ϫ164.
S. C. Bourque, F. Maltais, W.-J. Xiao, O. Tardif, H. Alper, P.
[35]
Arya, L. E. Manzer, J. Am. Chem. Soc. 1999, 121, 3035Ϫ3038.
U. Kragl, C. Dreisbach, Angew. Chem. Int. Ed. Engl. 1996,
35, 642Ϫ644.
N. Hovestad, E. B. Eggeling, H. J. Heidbüchel, J. T. B. H.
Jastrzebski, U. Kragl, W. Keim, D. Vogt, G. van Koten, Angew.
Chem. Int. Ed. 1999, 38, 1655Ϫ1658.
N. Brinkmann, D. Giebel, G. Lohmer, M. T. Reetz, U. Kragl,
J. Catal. 1999, 183, 163Ϫ168.
[1]
[36]
E. Buhleier, W. Wehner, F. Vögtle, Synthesis 1978, 155Ϫ158.
[2]
G. R. Newkome, Z.-Q. Yao, G. R. Baker, V. K. Gupta, J. Org.
[37]
Chem. 1985, 50, 2003Ϫ2004.
D. A. Tomalia, H. Baker, J. Dewald, M. Hall, C. Kallos, S.
Martin, J. Roeck, J. Ryder, P. Smith, Polym. J. (Tokyo) 1985,
17, 117Ϫ132.
J. Issberner, R. Moors, F. Vögtle, Angew. Chem. Int. Ed. Engl.
1994, 33, 2413Ϫ2420.
G. R. Newkome, C. N. Moorefield, F. Vögtle, Dendritic Molec-
[3]
[38]
[4]
[39]
D. de Groot, E. B. Eggeling, J. C. de Wilde, H. Kooijman, R.
[5]
J. van Haaren, A. W. van der Made, A. L. Spek, D. Vogt, J. N.
H. Reek, P. C. J. Kamer, P. W. N. M. van Leeuwen, Chem.
Commun. 1999, 1623Ϫ1624.
A. W. Kleij, R. A. Gossage, R. J. M. Klein Gebbink, N. Brink-
mann, E. J. Reijerse, U. Kragl, M. Lutz, A. L. Spek, G. van
Koten, J. Am. Chem. Soc. 2000, 122, 12112Ϫ12124.
B. M. Trost, D. L. Van Vranken, Chem. Rev. 1996, 96,
395Ϫ422.
M. Johannsen, K. A. Jørgensen, Chem. Rev. 1998, 98,
1689Ϫ1708.
F. Diederich, P. J. Stang, Metal-catalyzed Cross-coupling Reac-
tions, Wiley-VCH, Weinheim, 1998.
ules, Verlag Chemie, Weinheim, Germany, 1996.
[6]
J.-P. Majoral, A.-M. Caminade, Chem. Rev. 1999, 99, 845Ϫ880.
[7]
[40]
A. W. Bosman, H. M. Janssen, E. W. Meijer, Chem. Rev. 1999,
99, 1665Ϫ1688.
[8]
S. Serroni, G. Denti, S. Campagna, A. Juris, M. Ciano, V. Balz-
ani, Angew. Chem. Int. Ed. Engl. 1992, 31, 1493Ϫ1495.
G. R. Newkome, F. Cardullo, E. C. Constable, C. N. Moore-
field, A. M. W. Cargill Thompson, J. Chem. Soc., Chem. Com-
mun. 1993, 925Ϫ927.
T. Nagasaki, O. Kimura, M. Ukon, S. Arimori, I. Hamachi, S.
Shinkai, J. Chem. Soc., Perkin. Trans. 1 1994, 75Ϫ81.
W. T. S. Huck, F. C. J. M. van Veggel, D. N. Reinhoudt, Angew.
[41]
[9]
[42]
[10]
[43]
[11]
[44]
Y. Uozumi, H. Danjo, T. Hayashi, Tetrahedron Lett. 1997, 38,
Chem. Int. Ed. Engl. 1996, 35, 1213Ϫ1215.
3557Ϫ3560.
[12]
G.-X. Liu, R. J. Puddephatt, Organometallics 1996, 15,
[45]
´
´
N. Riegel, C. Darcel, O. Stephan, S. Juge, J. Organomet. Chem.
5257Ϫ5259.
1998, 567, 219Ϫ233.
D. E. Bergbreiter, Y.-S. Liu, P. L. Osburn, J. Am. Chem. Soc.
1998, 120, 4250Ϫ4251.
D. E. Bergbreiter, P. L. Osburn, Y.-S. Liu, J. Am. Chem. Soc.
1999, 121, 9531Ϫ9538.
S. Jayasree, A. Seayad, R. V. Chaudhari, Chem. Commun.
1999, 1067Ϫ1068.
H. Alper, P. Arya, S. C. Bourque, G. R. Jefferson, L. E.
Manzer, Can. J. Chem. 2000, 78, 920Ϫ924.
A. J. Sandee, D. Dimitrijevic, R. J. van Haaren, J. N. H. Reek,
P. C. J. Kamer, P. W. N. M. van Leeuwen, J. Mol. Catal. A,
in press.
E. B. Eggeling, N. J. Hovestad, J. T. B. H. Jastrzebski, D. Vogt,
G. van Koten, J. Org. Chem. 2000, 65, 8857Ϫ8865.
A. W. van der Made, P. W. N. M. van Leeuwen, J. Chem. Soc.,
Chem. Commun. 1992, 1400Ϫ1401.
[13]
M. Petrucci-Samija, V. Guillemette, M. Dasgupta, A. K. Kak-
kar, J. Am. Chem. Soc. 1999, 121, 1968Ϫ1969.
C.-O. Turrin, J. Chiffre, D. de Montauzon, J.-C. Daran, A.-M.
Caminade, E. Manoury, G. Balavoine, J.-P. Majoral, Macro-
molecules 2000, 33, 7328Ϫ7336.
P. J. Dandliker, F. Diederich, M. Gross, C. B. Knobler, A. Lou-
ati, E. M. Sandford, Angew. Chem. Int. Ed. Engl. 1994, 33,
1739Ϫ1742.
[46]
[47]
[48]
[49]
[50]
[14]
[15]
[16]
H. Brunner, J. Organomet. Chem. 1995, 500, 39Ϫ46.
P. Bhyrappa, J. K. Young, J. S. Moore, K. S. Suslick, J. Am.
[17]
Chem. Soc. 1996, 118, 5708Ϫ5711.
[18]
H.-F. Chow, C. C. Mak, J. Org. Chem. 1997, 62, 5116Ϫ5127.
[51]
[52]
[19]
G. E. Oosterom, R. J. van Haaren, J. N. H. Reek, P. C. J.
Kamer, P. W. N. M. van Leeuwen, Chem. Commun. 1999,
1119Ϫ1120.
R. van Heerbeek, J. N. H. Reek, P. C. J. Kamer, P. W. N. M.
[20]
[53]
[54]
L.-L. Zhou, J. Roovers, Macromolecules 1993, 26, 963Ϫ968.
D. Seyferth, D. Y. Son, A. L. Rheingold, R. L. Ostrander, Or-
ganometallics 1994, 13, 2682Ϫ2690.
van Leeuwen, Tetrahedron Lett. 1999, 40, 7127Ϫ7131.
[21]
C. Bolm, N. Derrien, A. Seger, Chem. Commun. 1999,
2087Ϫ2088.
[55]
[56]
[57]
[22]
´
B. Alonso, I. Cuadrado, M. Moran, J. Losada, J. Chem. Soc.,
J. W. J. Knapen, A. W. van der Made, J. C. de Wilde, P. W. N.
Chem. Commun. 1994, 2575Ϫ2576.
H. Frey, K. Lorenz, R. Muelhaupt, U. Rapp, F. J. Mayer-
Posner, Macromol. Symp. 1996, 102, 19.
E. V. Getmanova, T. B. Chenskaya, O. B. Gorbatsevich, E. A.
Rebrov, N. G. Vasilenko, A. M. Muzafarov, React. Funct. Po-
lym. 1997, 33, 289Ϫ297.
The structure was calculated by using Sybyl in Spartan 5.1.3
with a Unix workstation.
I. G. Iovel, Y. S. Goldberg, M. V. Shymanska, E. Lukevics,
Organometallics 1987, 6, 1410.
N. E. Schore, L. S. Benner, B. E. LaBelle, Inorg. Chem. 1981,
20, 3200.
A. W. Kleij, R. J. M. Klein Gebbink, P. A. J. van den Nieuwen-
huijzen, H. Kooijman, M. Lutz, A. L. Spek, G. van Koten,
Organometallics 2001, 20, 634Ϫ647.
M. van Leeuwen, P. Wijkens, D. M. Grove, G. van Koten, Na-
ture 1994, 372, 659Ϫ663.
D. Seebach, R. E. Marti, T. Hintermann, Helv. Chim. Acta
1996, 79, 1710Ϫ1740.
M. Bardaji, M. Kustos, A.-M. Caminade, J.-P. Majoral, B.
Chaudret, Organometallics 1997, 16, 403Ϫ410.
M. T. Reetz, G. Lohmer, R. Schwickardi, Angew. Chem. Int.
Ed. Engl. 1997, 36, 1526Ϫ1529.
R. Schneider, C. Köllner, I. Weber, A. Togni, Chem. Commun.
1999, 2415Ϫ2416.
A. W. Kleij, R. A. Gossage, J. T. B. H. Jastrzebski, J. Boersma,
G. van Koten, Angew. Chem. Int. Ed. 2000, 39, 176Ϫ178.
D. de Groot, P. G. Emmerink, C. Coucke, J. N. H. Reek, P. C.
[23]
[24]
[58]
[59]
[60]
[61]
[25]
[26]
[27]
[28]
J. Kamer, P. W. N. M. van Leeuwen, Inorg. Chem. Commun.
2000, 3, 711Ϫ715.
[62]
[63]
[29]
E. C. Alyea, R. P. Shakya, A. E. Vougioukas, Transition Met.
Chem. 1985, 10, 435.
G. P. C. M. Dekker, C. J. Elsevier, K. Vrieze, P. W. N. M. van
Leeuwen, Organometallics 1992, 11, 1598Ϫ1603.
D. Astruc, C. R. Acad. Sci. Ser. II 1996, 322, 757Ϫ766.
[30]
D. K. Smith, F. Diederich, Chem. Eur. J. 1998, 4, 1353Ϫ1361.
[31]
M. Fischer, F. Vögtle, Angew. Chem. Int. Ed. 1999, 38,
884Ϫ905.
1094
Eur. J. Org. Chem. 2002, 1085Ϫ1095

Palladium Complexes of Phosphane-Functionalised Carbosilane Dendrimers
FULL PAPER
[64]
[69]
[70]
R. J. van Haaren, H. Oevering, B. B. Coussens, G. P. F. van
Strijdonck, J. N. H. Reek, P. C. J. Kamer, P. W. N. M. van
Leeuwen, Eur. J. Inorg. Chem. 1999, 1237Ϫ1241.
Diethylamine was used instead of piperidine because it can eas-
ily be evaporated.
D. A. White, Inorg. Synth. 1972, 13, 55.
W. T. Dent, R. Long, A. J. Wilkinson, J. Chem. Soc. 1964,
1585.
A. Buhling, J. W. Elgersma, S. Nkrumah, P. C. J. Kamer, P. W.
N. M. van Leeuwen, J. Chem. Soc., Dalton Trans. 1996,
2143Ϫ2154.
[65]
[66]
[67]
[68]
[71]
M. Feuerstein, D. Laurenti, H. Doucet, M. Santelli, Chem.
Commun. 2001, 43Ϫ44.
J. E. Hoots, T. B. Rauchfuss, D. E. Wrobleski, Inorg. Synth.
1991, 28, 175.
R. E. Rülke, I. M. Han, C. J. Elsevier, K. Vrieze, P. W. N. M.
van Leeuwen, C. F. Roobeek, M. C. Zoutberg, Y. F. Wang, C.
H. Stam, Inorg. Chim. Acta 1991, 169, 5Ϫ8.
[72]
[73]
D. de Groot, PhD thesis, University of Amsterdam, Amster-
dam, The Netherlands, 2001.
E. Eggeling, PhD thesis, Rheinisch-Westfälischen Technischen
Hochschule Aachen, Aachen, Germany, 1999.
Received August 24, 2001
[O01411]
Eur. J. Org. Chem. 2002, 1085Ϫ1095
1095