Rhodium complexes non-covalently bound to cyclodextrins: Novel water-soluble supramolecular catalysts for the biphasic hydroformylation of higher olefins

Article abstract of DOI:10.1002/chem.200500337

A new class of cationic α-cyclodextrins bearing 2-hydroxy-3-trime- thylammoniopropyl groups has been synthesised. We investigated their efficiency as mass-transfer promoters in a biphasic hydroformylation reaction catalysed by a rhodium trisulfonated triphenylphosphine system. These cationic α-cyclodextrins greatly increased the reaction rate, the chemoselectivity, and, surprisingly, the linear-to-branched aldehyde ratio. We attributed this unexpected enhancement of the linear-to-branched aldehyde ratio to the in situ formation of new catalytic supramolecular species obtained by ion-exchange between the catalyst ligand and the cationic α-cyclodextrins.

Full text of DOI:10.1002/chem.200500337

DOI: 10.1002/chem.200500337
Rhodium Complexes Non-Covalently Bound to Cyclodextrins:
Novel Water-Soluble Supramolecular Catalysts for the
Biphasic Hydroformylation of Higher Olefins
BenoîtSueur, [a] Loïc Leclercq,[b] Mathieu Sauthier,[b] Yves Castanet,[b] AndrØ Mortreux,[b]
HervØ Bricout,[a] SØbastien Tilloy,[a] and Eric Monflier*[a]
Abstract: A new class of cationic a-cy-
clodextrins bearing 2-hydroxy-3-trime-
thylammoniopropyl groups has been
synthesised. We investigated their effi-
ciency as mass-transfer promoters in a
biphasic hydroformylation reaction cat-
alysed by a rhodium trisulfonated tri-
phenylphosphine system. These cation-
ic a-cyclodextrins greatly increased the
reaction rate, the chemoselectivity, and,
surprisingly, the linear-to-branched al-
dehyde ratio. We attributed this unex-
pected enhancement of the linear-to-
branched aldehyde ratio to the in situ
formation of new catalytic supramolec-
ular species obtained by ion-exchange
between the catalyst ligand and the cat-
ionic a-cyclodextrins.
Keywords: cyclodextrins · hydrofor-
mylation · phosphanes · rhodium ·
supramolecular chemistry
Introduction
(P(m-C6H4SO3Na)3 (TPPTS)).[3] The geometry of this inclu-
sion complex was deduced from two-dimensional NMR ex-
periments and it was demonstrated that one phenyl ring of
the TPPTS was included into the hydrophobic cavity of the
b-CD from the side of the secondary hydroxy groups.[4] The
affinity between the native b-CD and the ligand is very de-
pendent on the nature of the ligand. Thus, the value of the
association constant is particularly high when the ligand pos-
sesses a non-sulfonated aromatic ring as in the case of the
(3-sodium sulfonatophenyl)diphenylphosphine (TPPMS)[5]
The use of chemically modified b-cyclodextrins (b-CD) as
mass-transfer promoters is an attractive solution to extend
the scope of aqueous organometallic catalysis to water-in-
soluble substrates.[1] Indeed, the b-CD derivatives greatly
improve the solubility of highly hydrophobic substrates by
forming water-soluble inclusion complexes and, consequent-
ly avoid mass-transfer limitations.[2] Formation of inclusion
complexes with the phosphine used to solubilise the transi-
tion-metal in water can occur in parallel to the complexation
of the substrate. The first example of such a phenomenon
was reported with the native b-CD and the currently most-
used ligand in aqueous organometallic catalysis: the sodium
salt of the meta-substituted trisulfonated triphenylphosphine
and
the
bis(3-sodium
sulfonatophenyl)-[4-(4-tert-
butylphenyl)phenyl]phosphine.[6] Substitution of one or sev-
eral hydroxy groups of the b-CD by ionic groups has a
major influence on the original inclusion ability of the b-CD
towards the ligand. For example, the lowest association con-
stants are observed with b-CD derivatives bearing anionic
substituents. Indeed, electronic repulsions between the
anionic groups of these b-CD derivatives and the sulfonato
groups of the water-soluble phosphine impede the formation
of inclusion complexes.[7] Conversely, the presence of cation-
ic groups such as ammonium groups on the b-CD greatly in-
crease the affinity of the CD for the ligand due to attractive
electrostatic interactions between the cationic group of the
CD and the anionic group of the phosphine. For instance,
the value of the association constant for the TPPMS-
mono[2-O-(2-hydroxy-3-trimethylammoniopropyl)]-b-CD
inclusion complex was found to be 105500 versus 7110mÀ1
for the TPPMS-b-CD inclusion complex.[8] This inclusion
[a] B. Sueur, Dr. H. Bricout, Dr. S. Tilloy, Prof. E. Monflier
Laboratoire de Physico-Chimie des
Interfaces et Applications, FRE CNRS 2485
UniversitØ d’Artois
FacultØ des Sciences Jean Perrin
Rue Jean Souvraz, SP 18–62307 Lens CØdex (France)
Fax : (+33)321-791-755
E-mail: monflier@univ-artois.fr
[b] L. Leclercq, Dr. M. Sauthier, Prof. Y. Castanet, Prof. A. Mortreux
Laboratoire de Catalyse de Lille
UMR 8010, ENSCL, B.P.
108–59652 Villeneuve d’Ascq (France)
6228
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 6228 – 6236

FULL PAPER
complex was formed in situ by
an ion-exchange between the
mono[2-O-(2-hydroxy-3-trime-
thylammoniopropyl)]-b-CD
chloride and the sodium salt of
the TPPMS. The NMR experi-
ments demonstrated that one of
the non-sulfonated groups of
the TPPMS was included in the
host cavity of the CD and that
methyl groups of the 2-hy-
droxy-3-trimethylammonio-
propyl substituent were inequi-
valent due to the formation of
the ion pair.
The discovery of this ion-ex-
change gave us the idea to syn-
thesise water-soluble organome-
tallic catalysts possessing mass-
transfer functions by self-assem-
bly. Indeed, if the CD cavity is
not occupied by one phenyl
ring of the phosphine and re-
mains located near the catalyti-
cally active centre owing to the
formation of an ion pair be-
tween the ligand and the CD,
such an organometallic complex
could act as an efficient phase-
transfer catalyst and/or could
lead to unusual selectivity. As
we have recently demonstrated
that non-charged a-cyclodextrin
(a-CD) derivatives cannot form
inclusion complexes with the
TPPTS due to their cavities
being too small,[9] the a-CD de-
rivatives bearing 2-hydroxy-3-
trimethylammoniopropyl sub-
Scheme 1. Rhodium-catalysed hydroformylation of 1-decene in the presence of chemically modified a-cyclo-
dextrins.
stituents could be excellent candidates to obtain such cata-
lysts.
The native a-CD and 2,3-epoxypropyltrimethylammoni-
um chloride reacted in basic medium to afford the O-(2-hy-
Here we report the synthesis of new a-CD derivatives
bearing 2-hydroxy-3-trimethylammoniopropyl substituents.
The behaviour of these a-CD derivatives towards the
TPPTS ligand was investigated by 1H and 31P{1H} NMR
spectroscopy, and the ability of [Rh(acac)(CO)2]/TPPTS/cat-
ionic a-CD systems to catalyse the biphasic hydroformyla-
tion of 1-decene has been evaluated (Scheme 1).
droxy-3-trimethylammoniumpropyl)-a-CD
(mono-HTMAP-a-CD and tetra-HTMAP-a-CD;
derivatives
Scheme 2). For example, attachment of one HTMAP group
1
to the a-CD is evidenced by H NMR spectra showing the
presence of new signals assigned to H1’, H2’’, H3’, and
+
NMe3 (Figure 1a).
We checked the monosubstitution by comparing the inte-
grated areas of the resonance for the proton H2’’ and the
protons H1and H1 ’. The degree of substitution was also
confirmed by ES-MS. The substitution in position 2 of the
a-CD was indicated by the presence of H1’ corresponding
to the acetalic proton of the glucose unit substituted in posi-
tion 2. This position was in agreement with the data report-
ed by Deratani and co-workers for the HTMAP-b-CD. We
synthesised tetra-HTMAP-a-CD in the same way as mono-
HTMAP-a-CD by increasing the quantity of 2,3-epoxypro-
Results and Discussion
Synthesis of cationic a-CD derivatives: The a-CD deriva-
tives bearing the 2-hydroxy-3-trimethylammoniopropyl
(HTMAP) group were synthesised according to a procedure
analogous to that described by Deratani and co-workers for
b-CD derivatives bearing HTMAP groups.[10]
Chem. Eur. J. 2005, 11, 6228 – 6236
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6229

E. Monflier et al.
In the same way, the NMR
spectrum of the methylated de-
rivative of tetra-HTMAP-a-CD
indicates that methylation of
the hydroxy group of the
HTMAP group has occurred.
The average number of methyl
groups
and
methylated
HTMAP groups was deter-
mined by ES-MS and were
found to be 15 and 2, respec-
tively. The number of methylat-
ed HTMAP groups was surpris-
ingly lower than the initial
number of HTMAP groups in
the starting product (2 as op-
posed to 4). The loss of
HTMAP groups during the syn-
thesis was attributed to the high
concentration of base required
for the methylation. We assume
that the degradation proceeds
by a Hofmann elimination as
we detected the characteristic
odour of amine when the reac-
tion was complete.
1H and 31P{1H} NMR studies:
The behaviour of the cationic
a-CD derivatives toward the
TPPTS ligand has been investi-
Scheme 2. Synthesis of the cationic a-cyclodextrins used in this work.
gated by 1H and 31P{1H} NMR
spectroscopy. The 1H and
pyltrimethylammonium chloride (14 versus 0.7 equivalents).
The NMR and ES-MS analyses were in accordance with an
average degree of substitution of four. The regioselectivity
was not only in the 2 position but either in the 3 or 6 posi-
tions.
31P NMR spectra of a stoichiometric mixture of TPPTS and
1
mono-HTMAP-a-CD are displayed in Figures 2–4. H and
31P{1H} NMR spectra of TPPTS in the presence of mono-
HTMAP-b-CD are also given for comparison, as it was pre-
viously shown that the TPPTS ligand can be partially includ-
ed in the cavity of chemically modified b-CD derivatives.[6]
As expected, inclusion of one phenyl ring of the TPPTS
into the cavity of the mono-HTMAP-b-CD induces strong
As methylation of CD enhances the surface active proper-
ties[11] of CD and its affinity[12] for long-chain olefins, we en-
visaged the synthesis of methylated derivatives of the mono-
HTMAP-a-CD and the tetra-HTMAP-a-CD. Initial at-
tempts by adding 2,3-epoxypropyltrimethylammonium chlo-
ride to randomly methylated-a-CD (RAME-a-CD) in aque-
ous basic medium, or in N,N-dimethylformamide containing
barium oxide and barium hydroxide, failed due to the poor
reactivity of the hydroxy groups of the RAME-a-CD. Fortu-
nately, the target CD derivatives have been obtained from
mono-HTMAP-a-CD and tetra-HTMAP-a-CD by using di-
methylsulfate as a methylation reagent.[13] The 1H NMR
spectrum of the methylated mono-HTMAP-a-CD showed
the presence of methyl groups in the 2, 3, and 6 positions.
We obtained the average substitution-degree of 13 by ES-
MS. It must be pointed out that the hydroxy group of the
HTMAP group was also methylated as the signal of the H2’’
proton of the HTMAP group was found upfield in the CD
proton pattern (see Figure 1b).
1
modifications of the H NMR spectra of the TPPTS and b-
CD. We observed an overlap of the Hp and Ho’ signals and a
shift for Hm and Ho for the TPPTS protons (see Figure 2).
Concerning the b-CD protons, the signal at d=3.25 ppm
corresponding to the three methyl groups of ammonium was
split into several signals and the internal protons of the CD
(H3 and H5) were shifted to higher field (see Figure 3).
31P{1H} NMR spectra were also indicative of the inclusion
complex formation as an important variation of the chemi-
cal shift of the TPPTS phosphorus atom was observed in the
presence of mono-HTMAP-b-CD (d=À8.8 ppm as opposed
to À5.9 ppm) (Figure 4).
Conversely, when mono-HTMAP-a-CD was added to
TPPTS, the H and 31P{1H} spectra of the TPPTS were very
1
similar to those observed without CD. We observed a slight
change in chemical shift of the phosphorus atom (d=
6230
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 6228 – 6236

Rhodium Complexes as Supramolecular Catalysts
FULL PAPER
Figure 2. Effect of mono-HTMAP cyclodextrins on the 1H NMR spec-
trum of TPPTS (3 mm) in D2O at 298 K: a) 1H NMR spectrum of TPPTS
without cyclodextrin; b) in the presence of mono-HTMAP-b-CD (3 mm);
c) in the presence of mono-HTMAP-a-CD (3 mm).
Figure 1. 1H NMR spectrum of: a) mono-HTMAP-a-CD (3 mm);
b) mono-MTMAP/Me-a-CD (3 mm); in D2O at 298 K.
Figure 3. Effect of TPPTS on 1H NMR spectra of mono-HTMAP cyclodextrins (3 mm) in D2O at 298 K: a) mono-HTMAP-b-CD; b) mono-HTMAP-b-
CD with TPPTS (3 mm); c) mono-HTMAP-a-CD; d) mono-HTMAP-a-CD with TPPTS (3 mm).
Chem. Eur. J. 2005, 11, 6228 – 6236
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6231

E. Monflier et al.
cavity was also evidenced by two-dimensional T-ROESY ex-
periments on the stoichiometric mixture as no intense corre-
lation peak was observed between the internal protons of
the cavity and the TPPTS ligand. Although no definitive
evidence has been obtained, the slight modifications ob-
served in the NMR spectra could result from a partial ion-
exchange between the sodium salt of TPPTS and the mono-
HTMAP-a-CD. This hypothesis was supported by
a
31P{1H} NMR study conducted on various TPPTS/mono-
HTMAP-a-CD mixtures. Figure 5 displays the chemical
shifts of the TPPTS phosphorus atom as a function of the
mono-HTMAP-a-CD/TPPTS ratio.
The TPPTS chemical shift was strongly influenced by the
presence of mono-HTMAP-a-CD up to a ratio of mono-
HTMAP-a-CD/TPPTS of 11–12. At higher ratios, there is
almost no additional change in chemical shift suggesting
that all sodium cations of the TPPTS have been substituted
1
by the cationic a-CD. We carefully examined the H NMR
spectra of the above TPPTS/mono-HTMAP-a-CD mixtures.
We observed no change of the CD proton pattern confirm-
ing that inclusion of a phenyl ring of the TPPTS did not
occur even at high mono-HTMAP-a-CD/TPPTS ratios. We
should point out that the influence of the counterion of the
TPPTS on the chemical shift of the TPPTS phosphorus
atom has also been observed by Hanson and Riisager with
non-amphiphilic quaternary ammonium ions.[14] These au-
thors reported that the 31P{1H} NMR signal of TPPTS is
shifted toward high field in the presence of tetrabutylammo-
nium bromide due to ion-exchange between TPPTS and the
ammonium salt.
Figure 4. Effect of mono-HTMAP cyclodextrins on the 31P{1H} NMR
spectrum of TPPTS (3 mm) in D2O at 298 K with H3PO4 as internal refer-
ence: a) 31P{1H} NMR spectrum of TPPTS without cyclodextrin; b) in the
presence of mono-HTMAP-b-CD (3 mm); c) in the presence of mono-
HTMAP-a-CD (3 mm).
The stability of the mono-HTMAP-a-CD/TPPTS ion pair
was investigated in the presence of inorganic salts. In partic-
ular, it is expected that the addition of a large amount of
salt such as NaCl to a solution containing the mono-
HTMAP-a-CD/TPPTS ion pair could induce dissociation of
the ion pair and, consequently, formation of the sodium salt
of the TPPTS and the mono-HTMAP-a-CD with the chlo-
ride as counteranion. As shown in Figure 5, addition of
NaCl (100 equivalents relative to the mono-HTMAP-a-CD)
did not induce a total dissociation of the ion pair. Hence the
31P{1H} NMR signal of TPPTS for a CD/TPPTS ratio of nine
was d=À7.04 ppm without NaCl and À6.98 ppm in the pres-
ence of NaCl (100 equivalents/mono-HTMAP-a-CD). The
effect of NaCl was however more marked when the CD/
TPPTS ratio was lower. In fact, the chemical shift of the
phosphorus atom for a CD/TPPTS ratio of three was found
to be d=À6.32 ppm without NaCl and À6.11 ppm in the
presence of NaCl (100 equivalents/mono-HTMAP-a-CD).
A similar phenomenon was also observed when Na2SO4
(50 equivalents/CD) was added to the solution containing
the ion pair, confirming that the mono-HTMAP-a-CD/
TPPTS ion pair is relatively stable. Finally, it should be
noted that the high stability of TPPTS/ammonium ion pair
has also been mentioned in the literature. Numerous authors
have attributed the enhancement of activity and selectivity
observed when the hydroformylation reaction is conducted
in the presence of cetyltrimethylammonium surfactants to
Figure 5. Influence of salts on the 31P{1H} chemical shift of TPPTS:
(
)
31P{1H} chemical shift of the TPPTS as a function of the mono-
^
HTMAP-a-CD/TPPTS ratio. (+) 31P{1H} chemical shift of TPPTS for
mono-HTMAP-a-CD/TPPTS ratios of 3 and 9 when NaCl (100 equiv rel-
ative to the mono-HTMAP-a-CD) was added to the solution containing
the TPPTS and the mono-HTMAP-a-CD.
À0.2 ppm) and a modification of the signal corresponding to
the three methyl groups of ammonium (d=3.25 ppm) con-
firming that the cavity of the a-CD is not occupied by a
TPPTS phenyl ring. The lack of a phenyl ring in the host
6232
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 6228 – 6236

Rhodium Complexes as Supramolecular Catalysts
FULL PAPER
Table 1. Hydroformylation of 1-decene in the presence of different
As expected for a reaction under mass-transfer control,
the conversion was very low without CD (3%). Moreover,
the aldehyde selectivity was poor since an important part of
the terminal decene was converted into internal decene
(41%). Finally, the linear-to-branched (l/b) aldehyde ratio
was equal to 2.8. For comparison, we performed experi-
ments with non-charged CD derivatives such as the hydroxy-
propyl-a-CD (HP-a-CD) and the randomly methylated a-
CD (RAME-a-CD) (Table 1, entries 4 to 6). It has been re-
ported that these two CD derivatives greatly improved the
conversion and selectivity without decreasing the l/b alde-
hyde ratio.[17]
chemically modified a-cyclodextrins.[a]
Entry CD
Na2SO4 Time Conversion Selectivity l/b
[h]
[%]
[%]
ratio
1
–
–
–
(À)
(À)
(+)
(À)
(À)
(À)
6
6
6
6
6
24
6
3
3
59
67
2.8
88
912.8
92
2.8
2.7
2[b]
3[c]
4
160
24
43
65
13
HP-a-CD
RAME-a-CD
RAME-a-CD
3.0
5
6
7
2.8
3.6
mono-HTMAP- (À)
93
a-CD
8[d]
9[e]
10
11
12
13
mono-HTMAP- (+)
a-CD
mono-HTMAP- (+)
a-CD
6
9
93
93
98
99
88
94
3.6
3.6
3.1
3.1
5.4
4.7
6
4
Addition of mono-HTMAP-a-CD to the reaction medium
has a significant influence on the decene conversion as the
latter was enhanced by a factor of four in comparison with
the control experiment (compare entry 7 with entry 1in
Table 1). However, we observed a lower rate-increase than
with the HP-a-CD and RAME-a-CD, probably owing to
the ionic nature of the mono-HTMAP-a-CD. Our experi-
ment conducted in the presence of Na2SO4 (Table 1, entry 3)
and numerous literature data indicate that addition of a
non-amphiphilic salt to aqueous catalytic solution has a det-
rimental effect on the hydroformylation rate.[18] In fact, the
solubilising power of the mono-HTMAP-a-CD due to its
cavity is partly counterbalanced by a salting out effect. Fur-
thermore, tensiometric measurements indicated that this cat-
ionic CD adsorbs weakly at the liquid–liquid interface,
which is unfavourable to the mass-transfer between the
aqueous and organic phases.[11] The aldehyde selectivity in
the presence of mono-HTMAP-a-CD was similar to that
observed with the non-charged CD derivatives, confirming
the protective effect of the CD cavity towards the isomerisa-
tion reaction. Indeed, when the olefin is inside the hydro-
phobic host cavity of the a-CD, the formation of isomerising
species leading to internal olefins by b-hydride elimination
is prohibited by the steric hindrance. The l/b aldehyde ratio
obtained with the mono-HTMAP-a-CD was more unexpect-
ed and was notably increased: 3.6 versus 2.8–3.0 in the pres-
ence of non-charged CD derivatives. We experimented to
determine the origin of this increase by adding the ammoni-
um salt HOCH2CH(OH)CH2N(CH3)3+ClÀ to the catalytic
medium instead of the mono-HTMAP-a-CD (see Table 1,
entry 2). This ammonium salt was chosen because of its
structural similarity with the substituent group of the mono-
HTMAP-a-CD and its lack of molecular recognition ability.
No increase of the conversion and l/b aldehyde ratio was ob-
served in the presence of this additive, suggesting that the l/
b aldehyde ratio increase in the presence of mono-HTMAP-
a-CD is not due to a salt effect but to the presence of the
host cavity of mono-HTMAP-a-CD. We could explain the
beneficial effect of the cavity on the l/b aldehyde ratio by
two hypotheses. The first is the formation of a new catalytic
species by an in situ ion-exchange between the sodium of
TPPTS and the mono-HTMAP-a-CD. As the receptor site
of such a supramolecular catalyst is located near the coordi-
nation sphere of the catalytically active metal centre, the
olefin included into the cavity could adopt a preferential ori-
mono-MTMAP/ (À)
6
43
89
8
Me-a-CD
mono-MTMAP/ (À)
Me-a-CD
24
6
tetra-HTMAP-
a-CD
(À)
di-MTMAP/Me- (À)
24
16
a-CD
[a] Experimental conditions: [Rh(acac)(CO)2] (4.07”10À2 mmol), TPPTS
(0.21mmol), CD (0.48 mmol), H 2O (11.5 mL), 1-decene (20.35 mmol),
+
808C, CO/H2 (1:1, 50 atm). [b] The salt HOCH2CH(OH)CH2N(CH3)3
ClÀ (0.48 mmol) was added in the catalytic medium instead of the CD.
[c] Na2SO4 (2.4 mmol) was added to the solution. [d] Na2SO4 (1.2 mmol)
was added to the solution. [e] Na2SO4 (2.4 mmol) was added to the solu-
tion.
strong interactions of the TPPTS sulfonate groups with the
surface of micelles composed of surfactant ammonium
groups.[15] The strong affinity of the TPPTS for ammonium
groups has also been exploited to immobilise rhodium com-
plexes on structurally well-defined microparticles. Mecking
and Thomann have reported that the [HRh(CO)(TPPTS)3]
complex can be bound to polyelectrolyte-coated-latex parti-
cles through strong interactions between the TPPTS ligand
and the ammonium groups of the polyelectrolyte.[16]
Although the spectra are not displayed here, similar re-
sults were obtained with the other cationic a-CD deriva-
tives. The 1H NMR spectra of the tetra-HTMAP-a-CD,
mono-MTMAP/Me-a-CD and di-MTMAP/Me-a-CD in the
presence of TPPTS were very similar to those obtained
without TPPTS except for the signal of the trimethylammo-
nium group. As in the case of the mono-HTMAP-a-CD, the
31P NMR signal of TPPTS shifted slightly towards high field
in the presence of the three previous a-CD derivatives, sug-
gesting partial cation-exchanges.
Catalytic experiments: The catalytic properties of rhodium
catalysts prepared in situ from [Rh(acac)(CO)2], TPPTS,
and cationic a-CD derivatives were studied in the biphasic
hydroformylation of 1-decene. We performed preliminary
experiments to determine the precise influence of the vari-
ous cationic a-CD derivatives. Firstly, a control experiment
was performed in the absence of CD under standard reac-
tion conditions (Table 1, entry 1).
Chem. Eur. J. 2005, 11, 6228 – 6236
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6233

E. Monflier et al.
entation which is favourable to the formation of the linear
alkyl rhodium species leading to the linear aldehyde. The
second hypothesis is a steric stress of the CD cavity on the
substrate compelling the latter to react preferentially by its
terminal carbon. We favour the ion-exchange hypothesis as
being the most probable. It is difficult to imagine that the
presence of only one arm on the CD can constrain the
olefin to react by its terminal carbon whereas such a phe-
nomenon is not observed with HP-a-CD that has about 3.6
arms of similar structure without the ammonium group. Ex-
periments performed in the presence of increasing amounts
of sodium sulfate salt indicate that the catalytic species ob-
tained by in situ ion-exchange are stable in the reaction
medium. As shown in Table 1entries 7–9, addition of
Na2SO4 decreased the conversion by a salting out effect but
had no effect on the l/b aldehyde ratio. The ratio remained
high (3.6 as opposed to 2.8 without CD), suggesting that the
screening effect of Na2SO4 is lessened due to the strong
ionic bonding of the cationic CD to the TPPTS ligand. This
hypothesis was totally in agreement with our NMR study of
various mono-HTMAP-a-CD/TPPTS mixtures in the pres-
ence of salts.
The results obtained with the methylated form of the
mono-HTMAP-a-CD are illustrative of the difficulty to an-
ticipate the reactivity of the catalysts involving molecular
recognition processes. The conversion and the aldehyde se-
lectivity in the presence of mono-MTMAP/Me-a-CD were
much higher that those observed with the mono-HTMAP-a-
CD (compare entries 7 and 10) and, appeared even higher
than those with the RAME-a-CD, which is the most effi-
cient CD known to date (compare entries 6 and 11 in
Table 1). A decrease in the l/b aldehyde ratio was unfortu-
nately observed with this mono-MTMAP/Me-a-CD (3.1
versus 3.6 with the mono-HTMAP-a-CD) even though the
l/b aldehyde ratio was higher than that observed with the
RAME-a-CD. The conversion and selectivity increases ob-
served with the mono-MTMAP/Me-a-CD can be attributed
to the presence of an enlarged hydrophobic cavity able to
receive the olefin in a more efficient way and to enhance
surface active properties of this cyclodextrin. Unfortunately,
the l/b aldehyde ratio decrease cannot be easily rationalised.
Indeed, methylation of the mono-HTMAP should not
impede the cation exchange between the TPPTS and the
cationic CD and should not notably modify the orientation
of the olefin in the coordination sphere of catalytic centre.
The results from a-CD derivatives containing several cat-
ionic arms confirm the beneficial effect of ammonium
groups on the l/b aldehyde ratio (Table 1, entries 12 and 13).
Indeed, the highest l/b aldehyde ratios were obtained with
these polycationic a-CD derivatives (4.7 for the di-
MTMAP/Me-a-CD and 5.4 for of tetra-HTMAP-a-CD).
The remarkable l/b aldehyde ratio increase supports our hy-
pothesis concerning the origin of the increase. The presence
of several ammonium groups increases the probability of ex-
change between the sodium salt of TPPTS and the cationic
CD and so, the formation of new catalytic species which are
non-covalently connected to the CD. Furthermore, the pres-
ence of several cationic groups on the CD gives rise to more
sterically hindered a-CD derivatives which induce an orien-
tation of olefin in the cavity more efficiently, which is fa-
vourable to the formation of the linear alkyl rhodium spe-
cies. Finally, the poor conversions observed with the di-
MTMAP/Me-a-CD and the tetra-HTMAP-a-CD are relat-
ed to the polycationic nature of these cyclodextrins. Indeed,
polycharged CD derivatives do not adsorb strongly at the
liquid–liquid interface and consequently have poor phase-
transfer properties.
Conclusion
We have shown that a-CD derivatives bearing 2-hydroxy-3-
trimethylammoniopropyl groups are efficient mass-transfer
promoters for the hydroformylation of higher olefins in bi-
phasic systems as the conversion, the aldehyde selectivity
and especially the l/b aldehyde ratio can be significantly in-
creased. The l/b aldehyde ratio increase was attributed to
the in situ formation of new catalytic supramolecular species
by ion-exchange between the catalyst ligand and the cationic
a-CD. Contrary to supramolecular catalysts where the CD is
covalently bound to the catalysts,[19] ionic bonding of CD to
the organometallic catalyst offers the advantage of a reversi-
ble interaction with the catalyst. Consequently, when the
CD is used in large excess toward the catalyst, the binding
of the substrate–CD inclusion complex to the catalyst is
transient and CD could behave as a supramolecular shuttle
between the catalytically active centre and the bulk or inter-
face as schematically represented in Figure 6. We are cur-
rently investigating the properties of this new class of supra-
molecular catalytic systems.
Figure 6. Principle of supramolecular catalysis conducted in the presence
of an excess of CD non-covalently bound to the catalysts.
Experimental Section
General remarks: The 1H and 31P NMR spectra were recorded at 300.13
and 121.49 MHz on a Bruker Avance DRX300 spectrometer, respective-
ly. 1H and 31P{1H} chemical shifts are given in ppm relative to external
references: sodium [D4]3-(trimethylsilyl)propionate (98% atom D) in
D2O for 1H NMR and H3PO4 in D2O for 31P{1H} NMR. ES-MS was per-
formed on a Finnigan LCQ-DUO mass spectrometer. Gas chromatogra-
phy was carried out on a Shimadzu GC-17 A gas chromatograph equip-
ped with a methylsilicone capillary column (25 m”0.25 mm) and a flame
ionisation detector (GC:FID). The exchange of sulfate by chloride ions
was performed with a Dowex 1X8–100 ion-exchange resin. Ion chroma-
tography was conducted on a Dionex DX-120 ion chromatograph equip-
6234
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 6228 – 6236

Rhodium Complexes as Supramolecular Catalysts
FULL PAPER
ped with an anionic column (Ion Pac AS14 Anion-Exchange). Dicarbo-
nylacetylacetonatorhodium(i) and organic compounds (undecane, 1-
decene, 2,3-epoxypropyltrimethylammonium chloride, native a-CD) were
purchased from Strem Chemicals and Aldrich Chemicals in their highest
purity and used without further purification. Randomly methylated-a-cy-
clodextrin (RAME-a-CD) was prepared by adapting a procedure report-
ed by Kenichi and co-workers.[13] This cyclodextrin was partially methy-
lated. Methylation occurred at positions 2, 3, or 6 and 1.8OH groups per
glucopyranose unit were statistically modified. Hydroxypropylated a-cy-
clodextrin (HP-a-CD) was obtained from Aldrich Chemicals. This cyclo-
dextrin was partially O-2-hydroxypropylated; statistically 0.6OH groups
were modified per glucopyranose unit. Tris(3-sodium sulfonatophenyl)-
phosphine (TPPTS - P(m-C6H4SO3Na)3) was synthesised as reported by
Gärtner and co-workers.[20] The purity of the TPPTS was carefully con-
trolled. In particular, 31P{1H} NMR solution spectroscopy indicated that
the product was a mixture of phosphine (approximately 98%) and its
oxide (approximately 2%). Carbon monoxide/hydrogen mixtures (1:1)
were used directly from cylinders (>99.9% pure, Air Liquide). Distilled
deionised water was used in all experiments. All catalytic reactions were
performed under nitrogen using standard Schlenk techniques. All sol-
vents and liquid reagents were degassed by bubbling with nitrogen for
15 min or by two freeze-pump-thaw cycles before use.
ES-MS: m/z: 1229.1–1312.6 [M]+. The average substitution degree for
MTMAP groups and for methyl groups was 1and 13, respectively.
Methylated di-[O-(2-methoxy-3-trimethylammoniopropyl)]-a-cyclodex-
trin chloride (di-MTMAP/Me-a-CD): tetra-HTMAP-a-CD (3.5 g,
2.33 mmol) was dissolved in aqueous NaOH solution (22.6 g in 34 mL,
0.56 mol), then dimethyl sulfate (53 mL, 0.56 mol) was added dropwise
over 10 h at room temperature. The reaction mixture was stirred for 17 h.
At the end of reaction, unreacted dimethyl sulfate was decomposed by
addition of concentrated ammonia solution (83 mL) followed by mixing
at room temperature for 6 h. di-MTMAP/Me-a-CD was extracted with
chloroform, and the organic layer was washed with water until the wash-
ings were of neutral pH and dried over anhydrous sodium sulfate. After
evaporation of chloroform, the product was dissolved in water and
passed through a chloride exchange resin to remove sulfate ions. The
total exchange of sulfate by chloride anions was checked by ion chroma-
tography. The water was distilled off and the product was obtained as a
white solid; yield 0.6 g (17%); 1H NMR (D2O): d=5.11 and 5.00 (m,
+
À
H1), 3.27 (s, MeOC6), 3.09 (s, N Me3), 4.0–3.35 ppm (other protons);
ES-MS: m/z: 1463.7–1491.7 [M]+. The average substitution degree for
MTMAP groups and for methyl groups was 2 and 15, respectively.
Catalytic experiments: [Rh(acac)(CO)2] (4.07”10À2 mmol), TPPTS
(0.21mmol) and the required amount of chemically modified cyclodex-
trin were dissolved in water (11.5 mL). The resulting aqueous phase and
an organic phase composed of olefin (20.35 mmol) and undecane
(4 mmol, GC internal standard) were charged under an atmosphere of N2
into the 50 mL reactor which was heated at 808C. A mechanical stirrer
equipped with a multipaddle unit was then started (1500 rpm) and the
autoclave was pressurised with CO/H2 (1:1, 50 atm) from a gas reservoir
connected to the reactor through a high pressure regulator valve enabling
constant pressure in the reactor throughout the whole reaction. The reac-
tion medium was sampled during the reaction for GC analysis of the or-
ganic phase after decantation. For kinetic measurements the time corre-
sponding to the addition of the CO/H2 mixture was considered as the be-
ginning of the reaction.
Mono[2-O-(2-hydroxy-3-trimethylammoniopropyl)]-a-cyclodextrin chlo-
ride (mono-HTMAP-a-CD): An aqueous solution of 2,3-epoxypropyltri-
methylammonium chloride (0.53 g, 3.49 mmol, 5 mL) was added dropwise
with stirring to a solution of a-CD (5 g, 5.14 mmol) dissolved in aqueous
NaOH (0.20 g in 10 mL of water, 5.14 mmol). The reaction mixture was
stirred overnight at 508C. After the mixture had been allowed to cool,
the solution was neutralised with aqueous HCl and the water was evapo-
rated. The residue was taken up in methanol (60 mL) and then refluxed
for 1h. The white product was precipitated by addition of acetone
(80 mL) to the filtrate. The product was obtained as a white solid; yield
1.35 g (23%); 1H NMR (D2O): d=5.08 (m, H1’), 4.93 (m, H1), 4.30 (m,
H2’’), 3.98 (t, J=9 Hz, H3’), 3.86 (t, J=9 Hz, H3), 3.85–3.60 (m, H5,
H5’, H6, H6’), 3.60–3.30 (H2, H2’, H4, H4’, H3’’, H1’’), 3.11 ppm
+
(s, N Me3); ES-MS: m/z: 1088.76 [M]+. The average substitution
À
degree determined by ES-MS was found to be one.
[1] a) E. Karakhanov, A. Maximov, A. Kirillov, J. Mol. Catal. A: Chem.
2000, 157, 25–30; b) E. A. Karakhanov, T. Y. Filippova, S. A. Marty-
nova, A. L. Maximov, V. V. Predeina, I. N. Topchieva, Catal. Today
1998, 44, 189–198; c) P. Kalck, L. Moquel, M. Dessoudeix, Catal.
Today 1998, 42, 431–440; d) L. N. Lewis, C. A. Sumpter, J. Stein, J.
Inorg. Organomet. Polym. 1996, 6, 123–144; e) C. Pinel, N. Gen-
dreau-Diaz, A. BrØhØret, M. Lemaire, J. Mol. Catal. A: Chem. 1996,
112, L157 L161; f) L. N. Lewis, C. A. Sumpter, J. Mol. Catal. A:
Chem. 1993, 104, 293–297; g) J. T. Lee, H. Alper, Tetrahedron Lett.
1990, 31, 4101–4104; h) H. Arzoumanian, D. Nuel, C. R. Acad. Sci.
Paris 1999, SØrie IIc, 289–293; i) J. R. Anderson, E. M. Campi, W. R.
Jackson, Catal. Lett. 1991, 9, 55–58; j) J. T. Lee, H. Alper, Tetrahe-
dron Lett. 1990, 31, 1941–1942; k) J. T. Lee, H. Alper, J. Org. Chem.
1990, 55, 1854–1856; l) H. A. Zahalka, K. Januszkiewicz, H. Alper,
J. Mol. Catal. A: Chem. 1986, 35, 249–253; m) A. Harada, Y. Hu, S.
Takahashi, Chem. Lett. 1986, 2083–2084; n) H. A. Zahalka, H.
Alper, Organometallics 1986, 5, 1909–1911.
[2] a) F. Hapiot, J. Lyskawa, S. Tilloy, H. Bricout, E. Monflier, Adv.
Synth. Catal. 2004, 346, 83–89; b) C. Torque, H. Bricout, F. Hapiot,
E. Monflier, Tetrahedron 2004, 60, 6487–6493; c) S. Tilloy, H. Bric-
out, E. Monflier, Green Chem. 2002, 4, 188–193; d) M. Dessoudeix,
M. Urrutigoꢁty, P. Kalck, Eur. J. Inorg. Chem. 2001, 1797–1800; e) T.
Mathivet, C. MØliet, Y. Castanet, A. Mortreux, L. Caron, S. Tilloy,
E. Monflier, J. Mol. Catal. A 2001, 176, 105–116; f) E. Monflier, G.
Fremy, Y. Castanet, A. Mortreux, Angew. Chem. 1995, 107, 2450–
2452; Angew. Chem. Int. Ed. Engl. 1995, 34, 2269–2271; g) E. Mon-
flier, S. Tilloy, G. Fremy, Y. Castanet, A. Mortreux, Tetrahedron
Lett. 1995, 36, 9481–9484.
Tetra[O-(2-hydroxy-3-trimethylammoniopropyl)]-a-cyclodextrin chloride
(tetra-HTMAP-a-CD): An aqueous solution (5 mL) of 2,3-epoxypropyl-
trimethylammonium chloride (10.9 g, 71.9 mmol) was added dropwise
with stirring to a solution of a-CD (5 g, 5.14 mmol) dissolved in aqueous
NaOH (0.41g in 20 mL of water, 10.3 mmol). The reaction mixture was
stirred overnight at 508C. After the mixture had been allowed to cool,
the solution was neutralised with aqueous HCl and the water was evapo-
rated. The residue was taken up in methanol (60 mL), and then refluxed
for 1h. The white product was precipitated by addition of acetone
(80 mL) to the filtrate. The product was obtained as a white solid; yield
4.1g (54%); 1H NMR (D2O): d=5.09 and 4.94 (m, H1, H1’), 4.31(m,
H2’’), 4.00–3.25 (H3’, H3, H5, H5’, H6, H6’, H2, H2’, H4, H4’, H3’’,
+
H1’’), 3.10 ppm (s, N Me3); ES-MS: m/z: 602.6–981.6 [M]2+. The aver-
À
age substitution degree determined by ES-MS was found to be four.
Methylated mono[2-O-(2-methoxy-3-trimethylammoniopropyl)]-a-cyclo-
dextrin chloride (mono-MTMAP/Me-a-CD): mono-HTMAP-a-CD (5 g,
4.45 mmol) was dissolved in aqueous NaOH solution (32.4 g in 49 mL,
0.8 mol), then dimethyl sulfate (76 mL, 0.8 mol) was added dropwise
over 10 h at room temperature. The reaction mixture was stirred for 17 h.
At the end of reaction, unreacted dimethyl sulfate was decomposed by
addition of concentrated ammonia solution (120 mL) followed by mixing
at room temperature for 6 h. mono-MTMAP/Me-a-CD was extracted
with chloroform, and the organic layer was washed with water until the
washings were of neutral pH and was dried over anhydrous sodium sul-
fate. After evaporation of chloroform, the product was dissolved in water
and was passed through a chloride exchange resin to remove sulfate ions.
The total exchange of sulfate by chloride ions was assessed by ion chro-
matography. The water was distilled and the product was obtained as a
[3] E. Monflier, S. Tilloy, C. MØliet, A. Mortreux, S. Fourmentin, D.
Landy, G. Surpateanu, New J. Chem. 1999, 23, 469–472.
[4] a) E. Monflier, S. Tilloy, L. Caron, J. M. Wieruszeski, G. Lippens, S.
Fourmentin, D. Landy, G. Surpateanu, J. Inclusion Phenom. Macro-
white solid; yield 2.3 g (39%); 1H NMR (D2O): d=5.13 and 4.90 (m,
+
À
H1), 3.30 (s, MeOC6), 3.12 (s, N Me3), 4.0–3.35 ppm (other protons);
Chem. Eur. J. 2005, 11, 6228 – 6236
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6235

E. Monflier et al.
cyclic Chem. 2000, 38, 361–379; b) A. Da Costa, E. Monflier, D.
Landy, S. Fourmentin, G. Surpateanu, Surf. Sci. 2001, 470, 275–283.
[5] a) M. Canipelle, L. Caron, C. Caline, S. Tilloy, E. Monflier, Carbo-
hydr. Res. 2002, 337, 281–287; b) L. Caron, C. Christine, S. Tilloy, E.
Monflier, D. Landy, S. Fourmentin, G. Surpateanu, Supramol. Chem.
2002, 14, 11–20.
[6] L. Caron, H. Bricout, S. Tilloy, D. Landy, S. Fourmentin, E. Monfli-
er, Adv. Synth. Catal. 2004, 346, 1449–1456.
[7] P. Black, S. Fourmentin, D. Landy, G. Surpateanu, A. Ponchel, H.
Bricout, F. Hapiot, E. Monflier, Adv. Synth. Catal. 2005, in press.
[8] M. Canipelle, S. Tilloy, A. Ponchel, H. Bricout, E. Monflier, J. Inclu-
sion Phenom. Macrocyclic Chem. 2005, 51, 79–85.
Li, J. Mol. Catal. A: Chem. 2003, 194, 13–17; c) L. Wang, H. Chen,
Y. He, Y. Li, M. Li, X. Li, Appl. Catal. A 2003, 242, 85–88; d) M.
GimØnez-Petros, A. Aghmiz, C. Claver, A. Masdeu-Bulto, D. Sinou,
J. Mol. Catal. A: Chem. 2003, 200, 157–163; e) Y. Zhang, Z. S. Mao,
J. Chen, Catal. Today 2002, 74, 23–35; f) C. Yang, X. Bi, Z. S. Mao,
J. Mol. Catal. A: Chem. 2002, 187, 35–46; g) H. Chen, Y. Li, J.
Chen, P. Cheng, X. Li, Catal. Today 2002, 74, 131–135; h) H. Chen,
Y. Li, J. Chen, P. Cheng, Y. He, X. Li, J. Mol. Catal. A 1999, 149, 1 –
6.
[16] S. Mecking, R. Thomann, Adv. Mater. 2000, 12, 953–956.
[17] L. Leclercq, M. Sauthier, Y. Castanet, A. Mortreux, H. Bricout, E.
Monflier, Adv. Synth. Catal. 2005, 347, 55–59.
[9] C. Binkowski, J. Cabou, H. Bricout, F. Hapiot, E. Monflier, J. Mol.
Catal. A: Chem. 2004, 215, 23–32.
[10] A. Deratani, G. Leliꢂvre, T. Maraldo, B. SØbille, Carbohydr. Res.
1989, 192, 215–222.
[11] L. Leclercq, H. Bricout, E. Monflier, unpublished results.
[12] a) N. Tanaka, A. Yamaguchi, Y. Araki, M. Araki, Chem. Lett. 1987,
715–718; b) M. Czugler, E. Ecke, J. J. Stzzowski, J. Chem. Soc.
Chem. Commun. 1981, 1294–1295; c) K. B. Lipkowitz, Chem. Rev.
1998, 98, 1829–1873; d) M. V. Rekharsky, Y. Inoue, Chem. Rev.
1998, 98, 1875–1917.
[18] a) Y. Huang, Li. Min, Y. Li, R. Li, P. Cheng, X. Li, Catal. Commun.
2002, 3, 71–75; b) H. Ding, B. E. Hanson, J. Mol. Catal. A: Chem.
1995, 99, 131–137; c) H. Ding, B. E. Hanson, T. E. Glass, Inorg.
Chim. Acta 1995, 229, 329–333; d) H. Ding, B. E. Hanson, J. Chem.
Soc. Chem. Commun. 1994, 2747–2748.
[19] a) M. T. Reetz, Catal. Today 1998, 42, 399–411; b) M. T. Reetz, J.
Heterocycl. Chem. 1998, 35, 1065–1073; c) M. T. Reetz, Top. Catal.
1997, 4, 187–200; d) M. T. Reetz, S. R. Waldvogel, Angew. Chem.
1997, 109, 870–873; Angew. Chem. Int. Ed. Engl. 1997, 36, 865–867;
e) D. Armspach, D. Matt, Chem. Commun. 1999, 1073–1074.
[20] R. Gärtner, B. Cornils, H. Springer, P. Lappe, DE Pat. 3235030,
1982.
[13] Y. Kenichi, M. Atsushi, T. Yukio, S. Mitsukatsu, Y. Yoshiaki, I. To-
moyuki, JP Pat. 8333406, 1996.
[14] A. Riisager, B. E. Hanson, J. Mol. Catal. A: Chem. 2002, 189, 195–
202.
[15] a) H. J. V. Barros, B. E. Hanson, E. V. Gusevskaya, E. N. Dos Santos,
Appl. Catal. A 2004, 278, 57–63; b) M. Li, Y. Li, H. Chen, Y. He, X.
Received: March 24, 2005
Published online: August 1, 2005
6236
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2005, 11, 6228 – 6236