β-Cyclodextrins modified by alkyl and poly(ethylene oxide) chains: A novel class of mass transfer additives for aqueous organometallic catalysis

Article abstract of DOI:10.1016/j.molcata.2009.11.015

A novel class of β-cyclodextrins (β-CDs) bearing alkyl chains on the secondary face and poly(ethylene oxide) chains on the primary face was synthesized. Their interactions with two water-soluble derivatives of triphenylphosphane were investigated by ³¹P{¹H} and ¹H NMR. Their behaviour in rhodium-catalysed biphasic hydroformylation of 1-decene was evaluated and the best result was obtained with a β-CD bearing methyl groups on the secondary face and poly(ethylene oxide) chains on the primary face. This CD appeared to be more efficient than randomly methylated β-CD which is currently one of the best mass transfer additives for hydroformylation.

Full text of DOI:10.1016/j.molcata.2009.11.015

Journal of Molecular Catalysis A: Chemical 318 (2010) 8–14
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Editors Choice Paper
␤-Cyclodextrins modified by alkyl and poly(ethylene oxide) chains:
A novel class of mass transfer additives for aqueous organometallic catalysis
Nezha Badif, Philippe Guéganf, Franc¸ ois-Xavier Legranda,b,d, Loïc Leclercqa,c,e
,
Sébastien Tilloya,b,d, Eric Monfliera,b,d,∗
a Univ Lille Nord de France, F-59000 Lille, France
b UArtois, UCCS, F-62300 Lens, France
c USTL, LCOM, F-59650 Villeneuve d’Ascq, France
d CNRS, UMR 8181, F-59650 Villeneuve d’Ascq, France
e CNRS, UMR 8009, F-59650 Villeneuve d’Ascq, France
f
Université d’Evry Val d’Essonne, Laboratoire MPI, 91025 Evry, France
a r t i c l e i n f o
a b s t r a c t
Article history:
A novel class of ␤-cyclodextrins (␤-CDs) bearing alkyl chains on the secondary face and poly(ethylene
Received 3 September 2009
Received in revised form
20 November 2009
Accepted 23 November 2009
Available online 3 December 2009
oxide) chains on the primary face was synthesized. Their interactions with two water-soluble derivatives
1
of triphenylphosphane were investigated by 31P{ H} and 1H NMR. Their behaviour in rhodium-catalysed
biphasic hydroformylation of 1-decene was evaluated and the best result was obtained with a ␤-CD
bearing methyl groups on the secondary face and poly(ethylene oxide) chains on the primary face. This
CD appeared to be more efficient than randomly methylated ␤-CD which is currently one of the best
mass transfer additives for hydroformylation.
Keywords:
Aqueous biphasic catalysis
Cyclodextrin
© 2009 Elsevier B.V. All rights reserved.
Hydroformylation
Long-chain olefin
Poly(ethylene oxide)
1. Introduction
CDs could be successfully used as mass transfer additive in alkyla-
tion [11,12], metathesis [13], telomerization [14] or polymerization
Heterogeneous and homogeneous catalytic processes are used
to produce numerous value-added products. However, both
approaches suffer from drawbacks. Homogeneous catalysts can-
not be separated from the product and contaminate it whereas
heterogeneous catalysts had a lower catalytic activity. In a green
chemistry context, the replacement of organic solvents by water
and the minimization of metal leaching are essential. This has
led to the development of aqueous biphasic catalysis which can
gather the positive aspects of both homogeneous and heteroge-
neous catalytic processes [1,2]. Unfortunately, numerous organic
substrates display a weak solubility in water leading to a very
low catalytic activity. In order to circumvent this major problem,
the use of cyclodextrins (CDs) for aqueous biphasic organometallic
catalysis was proposed in 1986 and is still the object of inten-
sive research [3–10]. For instance, it was recently reported that
[15] reactions. Some of us had deeply studied the role played
by CDs in rhodium-catalysed biphasic hydroformylation of higher
olefins [16]. The efficiency of CDs was attributed to the forma-
tion of an inclusion complex between the hydrophobic substrate
and CD at the liquid/liquid interface which facilitates the reac-
tion between the substrate and the water-soluble organometallic
catalyst [17–19]. In addition, it was demonstrated that one of the
essential points is to synthesize a surface-active CD which does not
interact with the water-soluble phosphanes. From NMR and cat-
alytic studies, it was proposed that a not-interacting ␤-CD has to
be encumbered on the secondary face by alkyl groups and has to
possess hydrophilic groups on the primary face. For instance, some
␤-CDs bearing alkyl chains on the secondary face and sulfoalkyl
chains on the primary face were synthesized and had proved to be
a really efficient mass transfer additive [20,21].
In our continuing research of better modified CDs, we report
hereby a novel class of ␤-CDs bearing alkyl groups on the sec-
ondary face and poly(ethylene oxide) (PEO) chains on the primary
face (Scheme 1). Methyl or heptyl groups were covalently linked
on the secondary face of ␤-CD to evaluate the effect of chain length
on the complexing properties. On the primary face, PEO chains
∗
Corresponding author at: Université d’Artois, Faculté des Sciences Jean Perrin,
UCCS-Artois, Rue Jean Souvraz, SP 18, 62307 Lens Cedex, France.
Tel.: +33 3 21 79 17 72; fax: +33 3 21 79 17 55.
E-mail address: eric.monflier@univ-artois.fr (E. Monflier).
1381-1169/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2009.11.015

N. Badi et al. / Journal of Molecular Catalysis A: Chemical 318 (2010) 8–14
9
Scheme 1. Structures of ␤-cyclodextrin derivatives, linear oligo(ethylene oxide) derivative and water-soluble ligands used in this work. The poly(ethylene oxide)-alkyl-␤-CDs
have been named by acronyms of the type n-PEO-Cnꢀ-␤-CD where n is the average number of ethylene oxide units per branch and Cnꢀ is the length of the alkylchain.
were grafted to confer a high hydrophilic character. The poly-
merization degree was chosen in order to obtain a cloud point
higher than the temperature of the reaction (80 ◦C). The way these
PEO-alkyl-␤-CDs interacted with two water-soluble derivatives of
triphenylphosphane (Scheme 1) and the behaviour of these CDs in
rhodium-catalysed biphasic hydroformylation were investigated.
of the two water-soluble phosphanes was carefully controlled. In
1
particular, 1H and 31P{ H} NMR analysis indicated that the water-
soluble phosphanes were only sulfonated in meta position and
less than 1% of its oxide was present. Carbon monoxide/hydrogen
mixture (1:1) was used directly from cylinders (>99.9% pure; Air
Liquide). Distilled deionized water was used in all experiments.
All solvents and liquid reagents were degassed by bubbling nitro-
gen for 15 min or by two freeze–pump–thaw cycles before use.
All the high-pressure hydroformylation experiments were carried
out in a 25 mL stainless steel autoclave supplied by Parr. Gas chro-
matographic analyses were carried out on a Shimadzu GC-17A gas
chromatograph equipped with a methyl silicone capillary column
(30 m × 0.32 mm) and a flame ionization detector (GC:FID). The
2. Experimental
2.1. Materials and apparatus
All chemicals were purchased from Acros Organics and
Sigma–Aldrich in their highest purity. Dicarbonylacetylacetonato
1
1H and 31P{ H} NMR spectra performed to investigate the inter-
rhodium(I), polyethylene glycol methyl ether (MW = 2000 g mol−1
)
actions between CDs and phosphanes were recorded at 300.13
and 121.49 MHz, respectively, on a Bruker Avance DRX300 instru-
and deuterated solvents were purchased, respectively, from Strem
Chemicals, Sigma–Aldrich and Euriso-Top in their highest purity
and used without further purification. ␤-CD and RAME-␤-CD
were purchased from Roquette Frères and Wacker Chemie GmbH,
respectively. ␤-CD was dried under vacuum at 80 ◦C during 16 h
before using. RAME-␤-CD was of pharmaceutical grade (Cavasol®
W7M Pharma) and its degree of substitution was equal to
1.7 [22]. RAME-␤-CD was used as received. The sodium salt
of tris(3-sulfonatophenyl)phosphane (TPPTS—P(m-C6H4SO3Na)3)
was synthesized as reported by Gärtner et al. [23]. The sodium salt
of diphenyl(3-sulfonatophenyl)phosphane (TPPMS—(C6H5)2P(m-
C6H4SO3Na)) was prepared by a literature method [24]. The purity
1
ment. Chemical shifts in 1H and 31P{ H} NMR are given in parts
per million (ppm) relative to external references sodium [D4]3-
(trimethylsilyl)propionate (98% atom D) in D2O for 1H NMR and
1
H3PO4 in H2O for 31P{ H} NMR.
The 1H and 13C (data not showed) NMR performed to charac-
terize the CD derivatives were recorded on a Bruker spectrometer
(1H, 300.13 MHz and 13C, 75.5 MHz) using DMSO-d6 or chloroform
(CDCl3) as solvents at 20 ◦C. The molar masses were determined by
size exclusion chromatography (SEC) with THF as eluent. SEC was
calibrated using linear poly(ethylene oxide) samples. The molecu-

10
N. Badi et al. / Journal of Molecular Catalysis A: Chemical 318 (2010) 8–14
Scheme 2. Synthetic strategy for PEO-alkyl-␤-CDs from ␤-cyclodextrin: (i) tert-butylchlorodimethylsilane, pyridine; (iia) 1-bromoheptane, sodium hydride, THF/DMF
(50/50); (iib) iodomethane, sodium hydride, THF; (iiia) tetrabutylammonium fluoride, THF; (iiib) ammonium fluoride, methanol and (iv) ethylene oxide, DPMK, THF.
lar weights of the star polymers are then calculated considering the
functionality of the initiator [25].
2.3. General procedure for the hydroformylation reaction
All catalytic reactions were performed under nitrogen
using standard Schlenk techniques. All solvents and liquid
reagents were degassed by bubbling nitrogen for 15 min or
by two freeze–pump–thaw cycles before use. Rh(acac)(CO)2
(4.07 × 10−3 mmol), TPPTS (0.021 mmol), chemically modified CD
(0.0068 mmol) or polyethylene glycol methyl ether (0.0472 mmol)
were dissolved in 1.2 mL of water. The resulting aqueous phase
and an organic phase composed of olefin (5.50 mmol) and unde-
cane (0.203 mol—GC internal standard) were charged under an
atmosphere of nitrogen into the 25 mL reactor, which was heated
at 80 ◦C. Mechanical stirring equipped with a multipaddle unit
was then started (1500 rpm) and the autoclave was pressurized
with 50 atm of CO/H2 (1/1) from a gas reservoir connected to
the reactor through a high-pressure regulator valve allowing to
keep constant the pressure in the reactor throughout the whole
reaction. The reaction medium was sampled after 24 h of reaction
for GC analyses of the organic phase after decantation. For kinetic
measurements the time corresponding to the addition of CO/H2
was considered as the beginning of the reaction.
2.2. Synthesis of ˇ-CD derivatives
The first step of the chemical modification was the synthesis
of heptakis(6-O-tert-butyldimethylsilyl)-␤-CD. This protection of
the ␤-CD primary face with tert-butyldimethylsilane was com-
mon to both alkyl CD derivatives and was described by Fügedi
[26]. Heptakis(2,3-di-O-heptyl)-␤-CD was synthesized according
to the procedure described by some of us [27]. Heptakis(6-O-
tert-butyldimethylsilyl-2,3-di-O-heptyl)-␤-CD was deprotected by
using tetrabutylammonium fluoride (TBAF) [28]. Heptakis(2,3-
di-O-methyl)-␤-CD was synthesized according to the procedure
described by Takeo et al. [29,30]. The deprotection of the primary
hydroxyl functions of heptakis(6-O-tert-butyldimethylsilyl-2,3-di-
O-methyl)-␤-CD was successfully achieved by using ammonium
fluoride [31].
The polymerization conditions were previously described by
some of us in the case of the heptyl derivatives and the same con-
ditions were nearly used for the methyl derivatives [32].
The synthesis of heptakis(6-O-PEO-2,3-di-O-methyl)-␤-CD
from heptakis(2,3-di-O-methyl)-␤-CD is described here to illus-
trate a typical synthesis. The polymerization reactions were carried
out at 40 ◦C under high vacuum in a 250 mL four-neck flask
equipped with a magnetic stirrer, an inlet, and three burets con-
taining ethylene oxide, diphenylmethylpotassium (DPMK) and the
anhydrous precursor CD derivative dissolved in THF (the final
hydroxyl concentration was [OH−] = 3 × 10−2 mol L−1). CD was first
introduced and it was followed by the slow addition of DPMK. The
orange-red color of the deprotonating agent becomes yellow when
the alkoxides were formed. The obtained homogenous solution
was stirred during 48 h at room temperature. It was cooled to 0 ◦C
before adding ethylene oxide ([OE] = 3 mol L−1) and then warmed at
40 ◦C for 9 days. The alkoxides were deactivated by adding a small
quantity of HCl (10%) in methanol (3 mL). The solution was concen-
trated and then precipitated twice in cold diethyl ether and once in
cold pentane. The heptakis(6-O-PEO-2,3-di-O-methyl)-␤-CD was
obtained with a yield of 45%.
2.4. Recycling experiments
The initial run was carried out as described above. After reaction,
the autoclave was cooled to room temperature and CO/H2 pressure
was carefully evacuated. The biphasic system was transferred out of
the autoclave under a nitrogen atmosphere. After decantation, the
Table 1
Characteristics of PEO-␤-CD conjugates.
a
b
Cyclodextrin
MnSEC (g mol−1
)
MnNMR (g mol−1
)
PDIc
39-PEO-C7-␤-CD
47-PEO-C7-␤-CD
51-PEO-C7-␤-CD
53-PEO-C1-␤-CD
11 700
13 500
14 500
17 700
14 600
17 000
18 300
17 900
1.05
1.05
1.10
1.06
a
Calculated from SEC chromatograms taking into account the seven arms of the
PEO.
b
Determined by 1H NMR.
c
Determined by SEC chromatography.

N. Badi et al. / Journal of Molecular Catalysis A: Chemical 318 (2010) 8–14
11
Fig. 1. 1H NMR spectrum of 53-PEO-C1-␤-CD in DMSO-d6 at 25 ◦C.
bulk organic solution was removed. The catalytic aqueous layer and
a new organic phase (composed of 1-decene (5.50 mmol) and unde-
cane (0.203 mmol)) were introduced in the autoclave. Then, the first
recycling was carried out using the same procedure as described
above. The overall recycling procedure was then repeated for the
second recycling.
iodomethane to generate heptakis(6-O-tert-butyldimethylsilyl-
2,3-di-O-alkyl)-␤-CDs. The deprotection of the primary face was
conducted with ammonium fluoride for the methyl derivative [31]
and with tetrabutylammonium fluoride for the heptyl derivative
[28]. This procedure allowed generating the primary hydroxyl
functions which had the same reactivity toward the subsequent
ethylene oxide polymerization. The ethylene oxide polymerization
with heptakis(2,3-di-O-heptyl)-␤-CD was already reported and
depended on the polymerization time. The number of ethylene
oxide units per PEO branches might vary from 39 to 51 (Table 1)
[32]. The synthesis of heptakis(6-O-PEO-2,3-di-O-methyl)-␤-CD
was carried out under vacuum in THF. A key step of this polymer-
ization was to remove all the water molecules entrapped in the
CD derivative. The best results were obtained by distillation of
heptane followed by distillation of benzene. This polymerization
was undertaken by addition of a THF solution of heptakis(2,3-
di-O-methyl)-␤-CD. Then, a diphenylmethylpotassium (DPMK)
solution was added. The molar ratio hydroxyl/DPMK was set
to 5 in order to prevent the initiator precipitation. All the pri-
mary hydroxyl groups could take part to the polymerization
step thanks to the fast kinetic exchange between hydroxyl and
hydroxylate forms [33]. Forty-eight hours after this metallation
3. Results and discussion
3.1. Synthesis
The synthesis of poly(ethylene oxide)-alkyl-␤-cyclodextrins
(PEO-alkyl-␤-CDs) was achieved in a multi-steps procedure. First,
the synthesis of CD initiators was conducted by using already
reported procedures and then, the polymerization of ethylene
oxide was carried out thanks to an anionic ring opening polymer-
ization (Scheme 2).
The synthesis of heptakis(6-O-tert-butyldimethylsilyl)-␤-
CD was first carried out to obtain
a selective modification
of the primary face of ␤-CD [26]. Then, heptakis(6-O-tert-
butyldimethylsilyl)-␤-CD was reacted with bromoheptane or
Fig. 2. Size exclusion chromatogram of 53-PEO-C1-␤-CD carried out in THF at RT.

12
N. Badi et al. / Journal of Molecular Catalysis A: Chemical 318 (2010) 8–14
1
Fig. 3. Partial 31P{ H} and 1H NMR (aromatic part) spectra of TPPTS (1 mM) in D2O at 25 ◦C in the presence of various PEO-alkyl-␤-CDs (1 mM).
step, ethylene oxide was introduced in the reactor. The charac-
teristics of this polymerization and those obtained for the various
heptakis(6-O-PEO-2,3-di-O-heptyl)-␤-CDs are reported in Table 1.
The polymerization yield was low and witnessed a very slow
polymerization kinetic. The NMR (Fig. 1) and SEC (Fig. 2) analy-
ses provide similar molecular weights (taking into account the star
correction for the SEC) [25] and the polydispersity index (PDI) was
pretty low (Table 1).
peak corresponding to the expected molecule, and a small peak in
the low molecular weight region (Fig. 2). This last population had a
molecular weight close to 2000 g mol−1, and corresponded to ␣,␻-
dihydroxy-PEO. The molecular weight of this side population was
close to the one of each branch of the poly(ethylene oxide)-alkyl-
␤-CDs, indicating that water molecules might still be entrapped in
the CD derivatives and participated to the polymerization. It was
already reported that such ethylene oxide polymerizations initi-
ated with CD-based initiators occurred in living conditions [32].
We assumed that it took place in the same way as our conditions
as suggested by the low polydispersity indexes of the polymers.
This result was in accordance with the living ethylene oxide
polymerization initiated by heptakis(2,3-di-O-heptyl)-␤-CD.
A
careful analysis of the size exclusion chromatogram showed a large
1
Fig. 4. Partial 31P{ H} and 1H NMR (aromatic part) spectra of TPPMS (1 mM) in D2O at 25 ◦C in the presence of various PEO-alkyl-␤-CDs (1 mM) or linear oligo(ethylene
oxide) derivative (7 mM).

N. Badi et al. / Journal of Molecular Catalysis A: Chemical 318 (2010) 8–14
13
Table 2
Biphasic rhodium-catalysed hydroformylation of 1-decene with TPPTS as water-soluble ligand in the presence of different additivesa.
Entry
Additives
Conversion (%)b
Selectivity (%)c
l/b ratiod
Phase separatione
1
2
3
4
5
6
7
8
9
–
1
29
31
30
26
61
33
56
58
61
95
95
98
98
89
89
90
89
2.8
1.8
2.8
2.8
2.8
2.5
2.7
2.5
2.6
+
+
−
−
−
+
RAME-␤-CD
39-PEO-C7-␤-CD
47-PEO-C7-␤-CD
51-PEO-C7-␤-CD
53-PEO-C1-␤-CD
44-PEGME
−
53-PEO-C1-␤-CDf
53-PEO-C1-␤-CDg
+
+
a
Experimental conditions: Rh(acac)(CO)2 (4.07 × 10−3 mmol), TPPTS (0.021 mmol), cyclodextrin (0.0068 mmol) or polyethylene glycol methyl ether (0.0472 mmol), H2O
(1.2 mL), 1-decene (5.50 mmol), undecane (0.203 mmol) 1500 rpm, T: 80 ◦C, P(CO/H2: 1/1) = 50 bar, time: 24 h.
b
Olefin conversion after 24 h (calculated with respect to the starting olefin).
c
Aldehydes selectivity after 24 h, i.e. (mol. of aldehydes)/(mol. of converted olefins) × 100. The side products were mainly isomeric olefins.
d
Ratio of linear to branched aldehyde product after 24 h.
e
The separation between the aqueous and organic phases is reported according to the following convention: −, formation of an emulsion; +, good and fast phase separation.
f
First recycling.
Second recycling.
g
3.2. Interaction of PEO-alkyl-ˇ-CDs with water-soluble
triphenylphosphanes
CD and the protons of PEO chains. The modifications observed on
1
the 31P{ H} NMR spectra for TPPMS in the presence of PEO-alkyl-
1
␤-CDs were particularly informative. Indeed, the 31P{ H} signals
The behaviour of PEO-alkyl-␤-CDs with water-soluble phos-
phanes such as the sodium salt of meta-substituted trisul-
fonated triphenylphosphane (TPPTS) and the sodium salt of
meta-substituted monosulfonated triphenylphosphane (TPPMS)
appeared to be upfield or downfield shifts according to the alkyl
1
group linked on PEO-alkyl-␤-CDs. Actually, the 31P{ H} signal of
TPPMS was downfield shifted in the presence of the three PEO-
C7-␤-CDs whereas an upfield shift was observed in the presence
of PEO-C1-␤-CD. These results suggested that a part of TPPMS
was included into the PEO-C1-␤-CD cavity whereas TPPMS formed
external adduct with the heptyl chains of PEO-C7-␤-CDs. Indeed,
some of us had previously demonstrated that an upfield shift
(Scheme 1) was investigated by NMR spectroscopy. The 31P{ H}
1
and 1H NMR spectra of 1:1 mixture of PEO-alkyl-␤-CDs and TPPTS
were very similar to those obtained for free TPPTS (Ref.), indicat-
ing that no interaction between both constituents took place (see
Fig. 3).
in 31P{ H} signal was synonymous with a formation of genuine
1
The absence of interaction between these PEO-alkyl-␤-CDs and
TPPTS was attributed to the presence of alkyl groups on the CD
secondary face rather than the presence of PEO groups on the CD
primary face. Indeed, some of us shown that the steric hindrance
generated by the presence of methyl groups on the secondary face
was sufficient to impede TPPTS penetration into the ␤-CD cav-
inclusion complexes whereas a downfield shift corresponded to a
modification of the phosphane environment [21,34,35]. For PEO-
C7-␤-CDs, the accessibility of the cavities was strongly reduced for
TPPMS because of sterically demanding heptyl chains on the sec-
ondary face. The presence of these chains probably prevented from
the formation of genuine inclusion complexes. In fact, it was already
described in the case of amphiphilic CDs that the access to the cav-
ity depended on the grafted chain length and the substituted face
[36,37].
1
ity [21]. In the case of TPPMS, the 1H and 31P{ H} NMR spectra
of 1:1 mixture of PEO-alkyl-␤-CDs and TPPMS were different to
those obtained for free TPPMS (Ref.—Fig. 4). In order to ensure that
these modifications in NMR spectra were really due to CD skele-
ton and not to the PEO chains, the same NMR experiments were
performed in the presence of a polyethylene glycol methyl ether
(44-PEGME—the polymerization degree is equal to 44 and close to
that of PEO-alkyl-␤-CDs). Contrary to PEO-alkyl-␤-CDs, no change
was observed for TPPMS spectra in the presence of 44-PEGME.
Chemical shift changes on 1H NMR spectra were observed for
the aromatic protons of TPPMS in the PEO-alkyl-␤-CD/TPPMS mix-
tures compared to pure TPPMS. For each PEO-C7-␤-CD mixed with
TPPMS, the three obtained spectra were similar. It was not the
case for PEO-C1-␤-CD for which the spectrum was different from
the three others. These different results suggested that the inter-
actions between TPPMS and these PEO-alkyl-␤-CDs depended on
the CD nature. In order to illustrate these different behaviours, 2D
T-ROESY experiments were performed to observe 1H–1H homonu-
clear dipolar interactions between TPPMS and PEO-alkyl-␤-CDs.
Unfortunately, the interpretation of these spectra was impossi-
ble because of a spectral overlap between the internal protons of
3.3. Catalysis
The effect of these different PEO-alkyl-␤-CDs on the activity,
the chemoselectivity and the regioselectivity was evaluated in the
rhodium-catalysed hydroformylation reaction of 1-decene in an
organic–aqueous biphasic system using Rh(acac)(CO)2 as catalyst
precursor and TPPTS as water-soluble ligand. TPPTS was preferred
to TPPMS since this latter did not interact with the PEO-alkyl-␤-
CDs. As a comparison, the results without CD and with commercial
randomly methylated ␤-CD (RAME-␤-CD) are also presented in
Table 2. The presence of any chemically modified ␤-CDs had a
beneficial effect on the activity and the selectivity (Table 2).
Surprisingly, similar conversions were reached in the presence
of any PEO-C7-␤-CDs whatever the number of ethylene oxide units
was. Unfortunately, a formation of an emulsion at the end of reac-
tion was observed and the separation between the aqueous and
organic layers was impossible. A reduction of the alkyl chain length

14
N. Badi et al. / Journal of Molecular Catalysis A: Chemical 318 (2010) 8–14
on the secondary face of PEO-alkyl-␤-CD had a beneficial effect on
the conversion. Indeed, the conversion was twice higher in the pres-
ence of PEO-C1-␤-CD compared to PEO-C7-␤-CDs. In addition, in
the case of PEO-C1-␤-CD, no emulsion was observed at the end
of the reaction and the separation between the organic and the
aqueous layer was very fast. To ensure that this positive effect on
the conversion was well due to the CD skeleton, the same experi-
ment was performed by replacing PEO-C1-␤-CD by 44-PEGME. In
the presence of 44-PEGME, the conversion was lower (33% com-
pared to 61%) and at the end of the reaction, the separation between
the aqueous and organic layers was not easy. So, the positive effect
of PEO-C1-␤-CD was unambiguously attributed to its complexing
properties. Interestingly, PEO-C1-␤-CD appeared more efficient in
terms of activity than RAME-␤-CD which is one of the best mass
transfer additives in rhodium-catalysed biphasic hydroformylation
of higher olefins [16].
Acknowledgments
This work was supported by the Centre National de la Recherche
Scientifique (CNRS). F.X. Legrand is grateful to the Ministère de
l’Education et de la Recherche (2007–2010) and the Réseau de
Recherche (RDR2 – Chimie éco-compatible – 2008) for financial
support. Roquette Frères (Lestrem, France) is gratefully acknowl-
edged for generous gifts of cyclodextrins.
References
[1] B. Cornils, W.A. Herrmann, in: B. Cornils, W.A. Herrmann (Eds.), Aqueous-Phase
Organometallic Catalysis, Wiley-VCH, 2004, pp. 3–16.
[2] For a recent review on hydrophilic ligands and their application in aqueous-
phase metal-catalyzed reactions see: K.H. Shaughnessy, Chem. Rev. 109 (2009)
643.
[3] H.A. Zahalka, K. Januszkiewicz, H. Alper, J. Mol. Catal. 35 (1986) 249.
[4] A. Harada, Y. Hu, S. Takahashi, Chem. Lett. (1986) 2083.
[5] A. Benyei, F. Joo, J. Mol. Catal. 58 (1990) 151.
[6] J.R. Anderson, E.M. Campi, W.R. Jackson, Catal. Lett. 9 (1991) 55.
[7] E. Monflier, E. Blouet, Y. Barbaux, A. Mortreux, Angew. Chem. Int. Ed. Engl. 33
(1994) 2100.
[8] C. Pinel, N. Gendreau-Diaz, A. Bréhéret, M. Lemaire, J. Mol. Catal. A: Chem. 112
(1996) L157–L161.
[9] E.A. Karakhanov, T.Y. Filippova, S.A. Martynova, A.L. Maximov, V.V. Predeina,
I.N. Topchieva, Catal. Today 44 (1998) 189.
[10] P. Kalck, L. Miquel, M. Dessoudeix, Catal. Today 42 (1998) 431.
[11] K. Hirano, H. Yorimitsu, K. Oshima, Adv. Synth. Catal. 348 (2006) 1543.
[12] K. Hirano, H. Yorimitsu, K. Oshima, Chem. Commun. (2008) 3234.
[13] T. Brendgen, T. Fahlbusch, M. Frank, D.T. Schühle, M. Seßler, J. Schatz, Adv.
Synth. Catal. 351 (2009) 303.
[14] A. Behr, J. Leschinski, C. Awungacha, S. Simic, T. Knoth, ChemSusChem 2 (2009)
71.
[15] L. Ding, X. Jiao, J. Deng, W. Zhao, W. Yang, Macromol. Rapid Commun. 30 (2009)
120.
[16] F. Hapiot, L. Leclercq, N. Azaroual, S. Fourmentin, S. Tilloy, E. Monflier, Curr. Org.
Synth. 5 (2008) 162.
[17] N. Sieffert, G. Wipff, J. Phys. Chem. B 110 (2006) 4125.
[18] N. Sieffert, G. Wipff, Chem. Eur. J. 13 (2007) 1978.
[19] L. Leclercq, H. Bricout, S. Tilloy, E. Monflier, J. Colloid Interface Sci. 307 (2007)
481.
The effect of these CDs on the chemoselectivity (the selectiv-
ity in aldehydes versus isomerisation products) was also studied.
No marked difference among the various PEO-alkyl-␤-CDs was
observed since the values increased from 60% (without CD) to at
least 89% in the presence of any CDs.
Concerning the regioselectivity, the linear to branched aldehy-
des ratio (l/b) varied from 2.8 (without CD) to 1.8 (in the presence
of RAME-␤-CD). As already described, this decrease was attributed
to the formation of 1:1 inclusion complexes between TPPTS and
RAME-␤-CD [38,39]. Actually, these inclusion complexes induced
the formation of phosphane low-coordinated rhodium species
responsible for the decrease in linear to branched aldehydes ratio
[40]. In the presence of PEO-alkyl-␤-CDs, the value of the l/b ratio
varied from 2.5 to 2.8 for PEO-C1-␤-CD and PEO-C7-␤-CDs, respec-
tively. So, all these PEO-alkyl-␤-CDs gave similar proportions of
linear and branched aldehydes. Indeed, when the l/b ratio varied
from 2.5 to 2.8, the percentages were equal to 71.4% and 73.7%
for linear aldehydes and 28.6% and 26.3% for branched aldehy-
des, respectively. These data confirmed that no modification of
the catalytic species occurred during the reaction when these PEO-
alkyl-␤-CDs were used as mass transfer additives.
[20] D. Kirschner, T. Green, F. Hapiot, S. Tilloy, L. Leclercq, H. Bricout, E. Monflier,
Adv. Synth. Catal. 348 (2006) 379.
[21] D. Kirschner, M. Jaramillo, T. Green, F. Hapiot, L. Leclercq, H. Bricout, E. Monflier,
J. Mol. Catal. A: Chem. 286 (2008) 11.
[22] F.X. Legrand, M. Sauthier, C. Flahaut, J. Hachani, C. Elfakir, S. Fourmentin, S.
Tilloy, E. Monflier, J. Mol. Catal. A: Chem. 303 (2009) 72.
[23] R. Gärtner, B. Cornils, H. Springer, P. Lappe, DE Pat. 3235030 (1982).
[24] F. Ahrland, J. Chatt, N.R. Davies, A.A. Williams, J. Chem. Soc. (1958) 276.
[25] G.S. Grest, L.J. Fetters, S.J. Huang, in: I. Prigogine, S.A. Rice (Eds.), Advances in
Chemical Physics, Polymeric Systems, vol. 94, Wiley-VCH, 1997, pp. 67–163.
[26] P. Fügedi, Carbohydr. Res. 192 (1989) 366.
Finally, it is important to underline that the catalytic system was
totally recyclable in the presence of the best mass transfer additive
(PEO-C1-␤-CD). Indeed, conversion, chemoselectivity and regios-
electivity were nearly unchanged after two recycling experiments
(see Table 2; entries 6 and 8–9).
[27] N. Badi, N. Jarroux, P. Guégan, Tetrahedron Lett. 47 (2006) 8925.
[28] M. Lalonde, T.H. Chan, Synthesis (1985) 817.
4. Conclusion
[29] K. Takeo, K. Uemura, H. Mitoh, J. Carbohydr. Chem. 7 (1988) 293.
[30] K. Takeo, H. Mitoh, K. Uemura, Carbohydr. Res. 187 (1989) 203.
[31] W. Zhang, M.J. Robins, Tetrahedron Lett. 33 (1992) 1177.
[32] N. Badi, L. Auvray, P. Guégan, Adv. Mater. 21 (2009) 4054.
[33] S. Feng, D. Taton, E.L. Chaikof, Y. Gnanou, J. Am. Chem. Soc. 127 (2005) 10956.
[34] M. Canipelle, L. Caron, C. Christine, S. Tilloy, E. Monflier, Carbohydr. Res. 337
(2002) 281.
In our continuing research of an ideal mass transfer additive
based on CDs, a ␤-CD bearing methyl chains on the secondary face
and PEO chains on the primary face appears to be an attractive
candidate for rhodium-catalysed biphasic hydroformylation of 1-
decene since high activity and selectivity were reached. Indeed, the
cavity of this CD is sufficiently accessible for the substrate while it is
too narrow to interact with the water-soluble phosphane (TPPTS).
Interestingly, this CD appeared more efficient in terms of activity
and regioselectivity than randomly methylated ␤-CD which is one
of the best mass transfer additives for hydroformylation. In addi-
tion, the chemoselectivity obtained in the presence of both CD was
similar. Finally, the catalytic system was totally recyclable in the
presence of this CD. We are currently working to synthesize new
PEO-methyl-␤-CDs with shorter PEO chains.
[35] L. Caron, C. Christine, S. Tilloy, E. Monflier, D. Landy, S. Fourmentin, G. Sur-
pateanu, Supramol. Chem. 14 (2002) 11.
[36] E. Memis¸ og˘lu, A. Bochot, M. Özalp, M. S¸ en, D. Duchêne, A.A. Hincal, Pharm. Res.
20 (2003) 117.
[37] E. Memis¸ og˘lu, A. Bochot, M. S¸ en, D. Duchêne, A.A. Hincal, Int. J. Pharm. 251
(2003) 143.
[38] E. Monflier, S. Tilloy, C. Méliet, A. Mortreux, S. Fourmentin, D. Landy, G. Sur-
pateanu, New J. Chem. 23 (1999) 469.
[39] E. Monflier, S. Tilloy, L. Caron, J.M. Wieruszeski, G. Lippens, S. Fourmentin, D.
Landy, G. Surpateanu, J. Inclusion Phenom. Macrocyclic Chem. 38 (2000) 361.
[40] E. Monflier, H. Bricout, F. Hapiot, S. Tilloy, A. Aghmiz, A.M. Masdeu-Bultó, Adv.
Synth. Catal. 346 (2004) 425.