Fluorous biphasic catalysis with dirhodium(II) perfluorocarboxylates: Selective silylation of alcohols under fluorous biphasic conditions

Article abstract of DOI:10.1016/S0040-4020(01)01086-9

Dirhodium(II) perfluorocarboxylates bearing C7-C13 perfluoroalkyl chains have been prepared and used as catalysts under fluorous biphasic conditions. They were found to be active and recyclable catalysts for the silylation of alcohols with triethylsilane.

Full text of DOI:10.1016/S0040-4020(01)01086-9

TETRAHEDRON
Pergamon
Tetrahedron 57 '2001) 10391±10394
Fluorous biphasic catalysis with dirhodiumꢀII)
per¯uorocarboxylates: selective silylation of alcohols under
¯uorous biphasic conditions
Andrea Bif®s,p Erika Castello, Marco Zecca and Marino Basato
Á Á
Dipartimento di Chimica Inorganica Metallorganica ed Analitica, Centro di Studio sulla Stabilita e Reattivita dei Composti di
Á
CoordinazioneÐC.N.R., Universita di Padova, via Marzolo 1, I-35131 Padova, Italy
Received 11 July 2001; accepted 9November 2001
AbstractÐDirhodium'II) per¯uorocarboxylates bearing C7±C13 per¯uoroalkyl chains have been prepared and used as catalysts under
¯uorous biphasic conditions. They were found to be active and recyclable catalysts for the silylation of alcohols with triethylsilane.
Hydrophobic, primary alcohols are preferentially silylated by the ¯uorous biphasic catalytic system in comparison with hydrophilic or
secondary ones. This opens the way to the development of selective silylation protocols. q 2001 Elsevier Science Ltd. All rights reserved.
1. Introduction
on the use of biphasic systems in which the catalyst is
soluble in one phase whereas reagents and products are
soluble in the other; in this case, simple phase separation
allows the isolation and recycling of the catalysts.10 The
major successes 'two large-scale industrial applications)
have been so far achieved with aqueous biphasic catalysis,
in which the catalyst is dissolved in an aqueous phase and
reagents and products in an organic phase.11 However, such
a strategy is neither applicable to substrates and reactions
which are water-sensitive, nor to hydrophobic reagents
which react too slowly with the water-con®ned catalyst. A
recently proposed alternative system is given by the so-
called ¯uorous biphasic catalysis.12±14 In this case, the
aqueous phase is substituted for by a per¯uorinated solvent
immiscible with the major part of organic solvents, in which
the catalyst is made soluble by functionalisation with
suitable per¯uoroalkyl chains.
The synthetic potential of dirhodium'II) carboxylates and
carboamidates as catalysts for organic syntheses, most
notably those involving diazocompounds, has been
disclosed in the course of the last years.1 There is an
impressive variety of chemical transformations that can be
mediated by such compounds, including cyclopropanations,
insertions into C±H and N±H bonds, ylide formations and
dipolar cycloaddition reactions; ef®cient asymmetric
variants of many of these reactions have been developed
as well.1,2 The chemo-, regio- and stereoselectivity of
these processes is chie¯y in¯uenced by the nature of the
coordination sphere surrounding the metal centre.3 In
particular, dirhodium'II) per¯uoroacetates and per¯uoro-
butyrates have been shown to be exceptionally useful for
insertions in C±H bonds3,4 and for the synthesis of 'Z)-a,b-
unsaturated carbonyl compounds.5 Furthermore, tetrakis
dirhodium'II) per¯uorobutyrate, introduced by Doyle
towards the end of the eighties, proved to be an excellent
catalyst not only for reactions involving diazocompounds,
but also for a series of other processes involving silanes,
such as silylcarbonylations of alkynes,6 hydrosilylations of
alkynes and alkenes,7,8 as well as silylations of alcohols.9
Dirhodium'II) pe¯uorocarboxylates appear particularly
suitable for ¯uorous biphasic catalysis. A simple extension
of the ligand per¯uoroalkyl chain should in fact be suf®cient
to impart the complex the proper solubility in a ¯uorous
phase. According to the current literature, suitable catalysts
for ¯uorous biphasic catalysis should contain a weight
percentage of ¯uorine of at least 60%. In the case of
dirhodium per¯uorocarboxylates, this percentage is already
achieved with per¯uoroalkyl chains six carbon atoms long.
In comparison with the known dirhodium catalysts with
short-chain per¯uorinated carboxylate substituents, the
elongation of the per¯uorinated chain should have minimal
in¯uence on the electronic properties of the ligand as well as
on the steric constraints around the active centre of the
catalyst 'the free apical position). Therefore, large varia-
tions in the catalytic performance of these compounds are
In spite of the potential of these catalysts, one of the funda-
mental problems for their industrial application is the need
for ef®cient catalyst recovery and recycling. In this connec-
tion, a novel approach has been recently developed, based
Keywords: alcohols; catalysis; rhodium and compounds; solvents and
solvent effects.
Corresponding author. Tel.: 139-49-8275216; fax: 139-49-8275223;
e-mail: andrea.bif®s@unipd.it
p
0040±4020/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020'01)01086-9

10392
A. Bif®s et al. / Tetrahedron 57 12001) 10391±10394
Table 1. Silylation of 1-octanol with triethylsilane under ¯uorous biphasic
conditions
not expected. Indeed, very recently it was demonstrated that
dirhodium'II) per¯uorooctanoate was an active catalyst for
ole®n cyclopropanations and that it could be easily
recovered by extraction into a ¯uorous phase after reaction
completion.15
Entry
Catalyst
Fluorous solvent
Alcohol/ Yield
'%)a
silane
ratio
1
2
3
4
5
6
Rh2'pfo)4 Per¯uorodecaline
Rh2'pfd)4 Per¯uorodecaline
1:1
1:1
1:1
1:1
1:1
1:2
2:1
1:1
1:1
82
69
52
68
68
64
87
60
45
In this paper, we aim at evaluating the reactivity of these
catalysts under ¯uorous biphasic conditions. For this
purpose, we have chosen as test reaction the silylation of
alcohols with silanes:
Rh2'pft)4
Rh2'pft)4
Rh2'pft)4
Rh2'pft)4
Rh2'pft)4
Rh2'pft)4
Rh2'pft)4
Per¯uorodecaline
Per¯uoro-methylcyclohexane
Fluorinertw FC-77
Fluorinertw FC-77
Fluorinertw FC-77
Fluorinertw FC-77
Fluorinertw FC-77
R3SiH 1 R0OH
R3SiOR0 1 H2
7
8b
9c
This reaction has practical interest as a valuable alternative
to the more classical silylation via stoichiometric reaction of
alcohols with chlorosilanes, which necessitates the presence
of a base and produces a stoichiometric amount of salt as
byproduct. Previous work by Doyle9 showed that tetrakis
dirhodium'II) per¯uorobutyrate is a very ef®cient catalyst
for this reaction. On the basis of spectroscopic and stereo-
chemical evidence, the following reaction mechanism was
proposed,9 in which the catalyst activates the silane for
nucleophilic attack by the alcohol at silicon to give a
dirhodium hydride complex and the protonated silyl ether;
subsequent protonation of the hydride forms the observed
products and releases the catalyst. The alcohol acts as an
inhibitor by competing with the silane for coordination to
the active site of the catalyst.
Reaction conditions: 1±2 mmols of each reagent, 0.01 mmol '1 mol%)
catalyst, 5 mL dichloromethane11 mL ¯uorous solvent, room temperature,
6 h.
a
GC yield referred to the limiting reagent.
Recycle of entry 5.
recycle of entry 8.
b
c
coloured, took place irrespective of the ¯uorous solvent
employed and was found to be irreversible. On the contrary,
with Rh2'pft)4 the dichloromethane phase remained
perfectly colourless. Since Rh2'pft)4 was the only catalyst
to remain satisfactorily con®ned in the ¯uorous phase under
the conditions employed, we concentrated our investiga-
tions on the reactivity of this complex.
ROH 1 Rh2L4 $ RhꢀL†4RhꢀROH†
The activity of the catalysts was signi®cantly decreased in
comparison with the corresponding homogeneous system
studied by Doyle '96% yield in 3 h).9 This is not surprising,
since it generally occurs when homogeneous catalysts are
employed under biphasic conditions. In this connection, the
higher reaction yields obtained with Rh2'pfo)4 and
Rh2'pfd)4 are probably due to the contribution of catalyti-
cally active species dissolved in the dichloromethane phase.
The reaction yield with Rh2'pft)4 can be increased by a
proper choice of the ¯uorous phase 'Table 1, entries 3±5),
as well as by using an excess of alcohol 'Table 1, entries
5±7). The reason for this behaviour will be discussed later.
Et3SiH 1 Rh2L4 $ RhꢀL†4RhꢀH±SiEt3†
2
RhꢀL†4RhꢀH±SiEt3† 1 ROH ! ‰RhꢀL†4Rh±HŠ
1Et3SiO1ꢀH†R ! Rh2L4 1 Et3SiOR 1 H2
2. Results and discussion
The preparation of the rhodium per¯uorocarboxylate
catalysts is straightforward.15 The necessary per¯uoro-
carboxylic acids RfCOOH 'Rfˆper¯uoroalkyl chain) are
among the most common per¯uorinated compounds and
are commercially available with a chain length up to 17
carbon atoms. Simple ligand exchange reactions with
dirhodium'II) acetate readily yield the catalyst of interest.
The Rh2'pft)4 catalyst can be conveniently recycled by
removing the dichloromethane phase after the reaction
and subsequently feeding the reactor with fresh reagent
solution. However, a decrease in the reaction yield, more
pronounced in the second recycle, was recorded 'Table 1,
entries 8 and 9). This is most likely attributable to catalyst
decomposition. We have been able to detect by GC/MS the
presence of the triethylsilyl ester of the per¯uorotetra-
decanoic acid ligand in the reaction mixture, which suggests
the existence of a decomposition pathway of the catalyst by
reductive elimination of the ester from a silylated inter-
mediate. The formation of traces of a brown deposit, pos-
sibly metallic rhodium, at the liquid±liquid interphase
supports this hypothesis. We are currently trying to optimise
the reaction procedure 'by e.g. controlled addition of the
silane to the reaction mixture) in order to minimize the
decomposition.
Rh2ꢀOAc†4 1 4RfCOOH
Rh2ꢀOOCRf†4 1 4AcOH
In this way, we have prepared three catalysts bearing
per¯uorinated chains with seven 'Rh2'pfo)4, with
pfoˆper¯uorooctanoate), nine 'Rh2'pfd)4, with pfdˆ
per¯uorodecanoate) and 13 'Rh2'pft)4, with pftˆper¯uoro-
tetradecanoate) carbon atoms, respectively.
Initial experiments were performed in a dichloromethane/
¯uorous solvent biphasic system using 1 mol% catalyst and
an equimolar mixture of 1-octanol and triethylsilane as
reagent 'Table 1). All catalysts were found to be active.
However, with Rh2'pfo)4 and Rh2'pfd)4 an appreciable
amount of catalyst was invariably extracted into the
dichloromethane phase in the course of the reaction. Leach-
ing into the dichloromethane phase, which became
We have extended our observations to other alcohol
substrates. The results are listed in Table 2. Interestingly,
the biphasic catalytic system with Rh2'pft)4 shows a marked
preference for primary alcohols; 1-octanol or benzyl alcohol

A. Bif®s et al. / Tetrahedron 57 12001) 10391±10394
10393
Table 2. Silylation of different alcohols with Rh2'pft)4 under ¯uorous
biphasic conditions
Entry
Alcohol
Alcohol/silane ratio
Yield '%)a
1
2
3
4
5
6
7
8
1-Octanol
1-Octanol
1-Octanol
2-Octanol
2-Octanol
2-Octanol
Benzyl alcohol
Cyclohexanol
1:1
1:2
2:1
1:1
1:2
2:1
1:1
1:1
68
64
87
22
22
33
81
28
9Propargyl alcohol
1:1
0
Reaction conditions: 1±2 mmol of each reagent, 0.01 mmol '1 mol%)
catalyst, 5 mL dichloromethane11 mL Fluorinertw FC-77, room
temperature, 6 h.
Figure 2. A possible reaction mechanism of the silylation of alcohols under
¯uorous biphasic conditions.
a
GC yield referred to the limiting reagent.
are much more readily silylated than 2-octanol or cyclo-
hexanol. Furthermore, hydrophilic alcohols, like propargyl
alcohol, are not reactive under biphasic conditions,
presumably because they are not able to diffuse in the
¯uorous phase.
ciency of the phase-transfer process depends on the stability
constant of the catalyst±alcohol complex. This appears to be
greater for primary alcohols, probably because of lower
steric limitations to coordination. A simple demonstration
of the higher coordination ability of primary alcohols is
given by the observation that addition of an excess of
1-octanol causes solubilisation of the per¯uorinated cata-
lysts even in dichloromethane, whereas this is not the case
with 2-octanol. Consequently, the phase-transfer effect is
expected to be more pronounced for primary alcohols.
The reported results suggest the possibility to perform
selective silylations of mixtures of alcohols with different
hydrophilicity and/or different degrees of branching at the a
position. The latter possibility has been tested by reacting a
1:1 mixture of 1-octanol and 2-octanol; the extent of
silylation of the primary alcohol greatly exceeded that of
the secondary one 'Fig. 1).
The hypothesis illustrated above is further supported by the
consideration that the presence of an excess of alcohol
invariably results in higher product yields. We have already
pointed out this for 1-octanol 'Table 1, entries 5±7), but this
is also true for 2-octanol 'Table 2, entries 4±6). In the latter
case, the reaction yield increases by about 50%, on going
from a stoichiometric 1:1 mixture to a 2:1 excess of the
alcohol.
The homogeneous catalyst investigated by Doyle also
showed higher activity with primary alcohols. However,
even if the reaction with secondary alcohols was slower,
high yields could be reached in the homogeneous system
at properly long reaction times '94% yield in 14 h for
2-octanol).9 In contrast, in the ¯uorous biphasic system
the reaction almost completely stops slightly above 20%
conversion for 2-octanol, whereas for 1-octanol it slows
down only above 70% conversion. This difference in
behaviour can be explained in terms of a reaction mechan-
ism in which the alcohol not only acts as a reagent, but also
as a phase-transfer agent towards the catalyst 'Fig. 2).
According to this hypothesis, the alcohol has ®rst to diffuse
in the ¯uorous phase where it binds to the rhodium catalyst.
The resulting complex adduct is less ¯uorous, hence more
prone to be partially transferred to the dichloromethane
phase, where it catalyzes the reaction homogeneously;
when the concentration of alcohol falls below a certain
value, the catalyst-alcohol adduct is no longer stable and
the catalyst diffuses back in the ¯uorous phase. The ef®-
3. Conclusions
We have shown that silylations of alcohols catalyzed by
dirhodium'II) per¯uorocarboxylates can be carried out in
a ¯uorous biphasic system, and that the utilization of
¯uorous biphasic conditions results in a peculiar selectivity
of the catalytic process. Hydrophobic, primary alcohols are
in fact preferentially silylated in comparison with hydro-
philic or secondary ones. We are currently engaged in the
optimisation of the performance of this catalytic system as
well as in its application to other reactions of practical
interest.
4. Experimental
All reagents and per¯uorinated solvents were purchased
from Aldrich or ABCR and used as received. Non¯uori-
nated solvents 'Carlo Erba RPE) were dried and distilled
according to standard procedures prior to use.
4.1. Preparation of the catalysts
4.1.1. tetrakis-DirhodiumꢀII)-per¯uorooctanoate, Rh2ꢀpfo)4.
The compound was prepared following a procedure similar
Figure 1. Competitive silylation of 1- and 2-octanol under ¯uorous bipha-
sic conditions.

10394
A. Bif®s et al. / Tetrahedron 57 12001) 10391±10394
to that adopted by Endres and Maas.15 0.750 g '1.81 mmol)
of per¯uorooctanoic acid are suspended in 40 mL of dry
toluene under N2. A suspension of 0.200 g '0.452 mmol)
of dirhodium'II) acetate in 40 mL of absolute ethanol is
then added, and the resulting mixture is heated at re¯ux.
The solvent is azeotropically removed down to a volume
of about 25 mL. The resulting solution is cooled to room
temperature and the residual solvent is stripped off in vacuo.
The waxy green residue is treated with a little acetonitrile.
The resulting suspension is ®ltered and the isolated purple
solid 'the complex-acetonitrile adduct) is heated at 1008C in
vacuo for 5 h to liberate the product. Anal. calcd for
Rh2C32F60O8: C 20.69, H absent; found C 20.49, H absent.
IR'KBr, cm21): 1651, 1424, 1244, 1204, 1146, 1018 'all s).
'25m OV-1701 capillary column; 1008C isotherm for 60 s
followed by heating at 168C min21 to 2008C). The GC
system was previously calibrated by determining the reten-
tion times and the response factors of the alcohol reagents
and of the silylated products. The latter were separately
prepared by reaction of the alcohol with chlorotriethylsilane
and pyridine in diethylether '1 h, rt) followed by ®ltration,
solvent evaporation and distillation under reduced pressure.
Acknowledgements
Financial support from MURST-PRIN 1999 'Project nr.
9903153427) and from the University of Padova-`Progetti
giovani ricercatori 1999' is gratefully acknowledged. We
wish to thank Dr Roberta Seraglia, C.N.R., for the
MALDI-TOF measurements.
4.1.2.
tetrakis-DirhodiumꢀII)-per¯uorodecanoate,
Rh2ꢀpfd)4. The compound was prepared from 0.465 g
'0.905 mmol) of per¯uorodecanoic acid and 0.100 g
'0.23 mmol) of dirhodium'II) acetate following a procedure
fully analogous to that reported above. Anal. calcd for
Rh2C40F76O8: C 21.28, H absent; found C 21.57, H absent.
IR'KBr, cm21): 1634, 1424, 1207, 1150, 804 'all s). MS
'MALDI-TOF, dihydroxybenzoic acid matrix): m/z 2648
'M112dihydroxybenzoic acid12K1).
References
1. Doyle, M. P.; McKervey, M. A.; Ye, T. Modern Catalytic
Methods for Organic Synthesis with Diazo Compounds;
Wiley: New York, 1997.
2. Doyle, M. P.; Forbes, D. C. Chem. Rev. 1998, 98, 911±935.
3. Doyle, M. P.; Ren, T. Prog. Inorg. Chem. 2001, 49, 113±168.
4. Padwa, A.; Austin, D. J.; Price, A. T.; Semones, M. A.; Doyle,
M. P.; Protopopova, M. N.; Winchester, W. R.; Tran, A. J. Am.
Chem. Soc. 1993, 115, 8669±8680.
4.1.3. tetrakis-DirhodiumꢀII)-per¯uorotetradecanoate,
Rh2ꢀpft)4. 1.29g '1.81 mmol) of per¯uorotetradecanoic
acid are suspended in 45 mL of dry toluene under N2. A
suspension of 200 mg '0.452 mmol) of dirhodium'II)
acetate in 40 mL of absolute ethanol is then added, and
the resulting mixture is heated at re¯ux. The solvent is
azeotropically removed down to a residual volume of
about 30 mL. The resulting suspension is cooled to room
temperature, ®ltered, and the isolated solid green product is
dried in vacuo. Anal. calcd for Rh2C56F108O8: C 21.99, H
absent; found C 21.64, H absent. IR 'KBr, cm21): 1634,
1424, 1204, 1150, 804 'all s). MS 'MALDI-TOF,
dihydroxybenzoic acid matrix): m/z 3252 'M11dihydroxy-
benzoic acid1K1).
5. Taber, D. F.; Herr, R. J.; Pack, S. K.; Geremia, J. M. J. Org.
Chem. 1996, 61, 2908±2910.
6. Doyle, M. P.; High, K. G.; Nesloney, C. L.; Clayton Jr, T. W.;
Lin, J.; Lin, J. Organometallics 1991, 10, 1225±1226.
7. Doyle, M. P.; Devora, G. A.; Nefedov, A. O.; High, K. G.
Organometallics 1992, 11, 549±555.
8. Doyle, M. P.; Shanklin, M. S. Organometallics 1993, 12, 11±
12.
9. Doyle, M. P.; High, K. G.; Bagheri, V.; Pieters, R. J.; Lewis,
P. J.; Pearson, M. W. J. Org. Chem. 1990, 55, 6082±6086.
10. Cornils, B., Herrmann, W. A., Eds.; Applied Homogeneous
Catalysis with Organometallic Compounds; VCH: Weinheim,
1996; Vol. II Chapter 3.1.
4.2. General procedure for the silylation reactions
11. Aqueous-Phase Organometallic Catalysis; Cornils, B.,
Herrmann, W. A., Eds.; Wiley-VCH: Weinheim, 1998.
12. Horvath, I. T.; Rabai, J. Science 1994, 266, 72±75.
13. Horvath, I. T. Acc. Chem. Res. 1998, 31, 641±650.
14. de Wolf, E.; Van Koten, G.; Deelman, B.-J. Chem. Soc. Rev.
1999, 28, 37±41.
The catalyst '0.01 mmol) is dissolved in 1 mL of ¯uorous
solvent in a glass vial under Ar. Dry dichloromethane
'5 mL) and the two reagents '1±2 mmol each) are then
added, and the resulting system is vigorously stirred at
room temperature. 0.5 mL samples of the dichloromethane
phase are withdrawn at regular intervals, diluted 1:1 with
dichloromethane, and analysed by gas chromatography
15. Endres, A.; Maas, G. Tetrahedron Lett. 1999, 40, 6365±6368.