Low melting mixtures based on β-cyclodextrin derivatives and N,N′-dimethylurea as solvents for sustainable catalytic processes

Article abstract of DOI:10.1039/c4gc00591k

β-Cyclodextrin series and N,N′-dimethylurea formed low melting mixtures able to immobilize organometallic species based on sulfonated phosphanes. Hydroformylation and Tsuji-Trost reactions were efficiently performed in these new solvents which led to new recyclable catalytic systems. This journal is the Partner Organisations 2014.

Full text of DOI:10.1039/c4gc00591k

Green Chemistry
View Article Online
View Journal | View Issue
PAPER
Low melting mixtures based on β-cyclodextrin
derivatives and N,N’-dimethylurea as solvents for
sustainable catalytic processes†
Cite this: Green Chem., 2014, 16,
876
3
a
b
b
b
François Jérôme, Michel Ferreira, Hervé Bricout, Stéphane Menuel,
b
b
Eric Monflier and Sébastien Tilloy*
Received 3rd April 2014,
Accepted 30th May 2014
β-Cyclodextrin series and N,N’-dimethylurea formed low melting mixtures able to immobilize organo-
metallic species based on sulfonated phosphanes. Hydroformylation and Tsuji–Trost reactions were
efficiently performed in these new solvents which led to new recyclable catalytic systems.
DOI: 10.1039/c4gc00591k
www.rsc.org/greenchem
Introduction
Recently, a new family of solvents, so-called deep eutectic sol-
vents (DESs) or low melting mixtures (LMMs), have emerged in
1
–5
the current literature.
They are prepared by mixing high-
melting-point starting materials, which form a liquid by hydro-
gen-bond interactions. They are generally cheap and easy to
prepare from readily available materials. Among them, LMMs
Scheme 1 Structure of CDs used.
6
hydroxyl groups could form hydrogen bonds with DMU. In
addition, the CD family allowed increasing catalytic activity
and/or selectivity during organometallic processes by playing
based on carbohydrates have been also described. The latter
are environmentally benign, because they are easily biodegrad-
able, relatively non-toxic, and are available from bulk
renewable resources without numerous energy consuming
modification steps. Some LMMs based on carbohydrates and
organometallic catalysis have already been combined with
success. In 2006, König and co-workers investigated the palla-
dium-catalyzed Suzuki coupling of phenyl boronic acid with
aryl bromides in different carbohydrates–urea–inorganic salts
1
0,11
12–14
the role of mass transfer agents
or ligands.
In this
article, we described new LMM mixtures easily obtained by the
simple association of β-CD or its derivatives (Scheme 1) and
DMU. Some of these mixtures were evaluated as solvents
in rhodium-catalyzed hydroformylation reaction and in
palladium-catalyzed cleavage of allylcarbonates (Tsuji–Trost
reaction).
7
mixtures. In the same publication, they have also reported the
hydrogenation of methyl α-cinnamate in various carbohydrate–
urea melts in the presence of Wilkinson’s catalyst. The Stille
cross-coupling in various sugar–urea–salt melts was also investi- Results and discussion
8
gated. Next, the same group highlighted that palladium-
DMU was mixed with different CDs at a weight ratio of 70/30
catalyzed Heck coupling can be successfully performed in low
9
by using a mortar and pestle. Each preparation was then intro-
duced into a Schlenk tube with a stirring bar and gradually
heated with an oil bath until the melting of the mixture was
visually observed (see Fig. 1 as an illustration). Note that the
melting points of all the CDs are superior to 200 °C whereas
that of DMU is equal to 104 °C. Table 1 summarizes the
melting temperatures of these mixtures. For TRIME-β-CD, no
melt was observed. This behaviour could be due to the
inability of this CD to form hydrogen bonds with DMU. All
other mixtures gave a clear melt around 90 °C. These tempera-
tures were also determined by differential scanning calori-
metry (DSC; see Table 1 and ESI†). The obtained temperatures
by the two methods are in accordance. To evaluate the thermal
melt composed of D-mannose and N,N′-dimethylurea (DMU).
In this context, we propose a new family of LMMs based on
DMU and β-cyclodextrin (β-CD) derivatives for an organometal-
lic catalysis application. Indeed, β-CD is a cyclic carbohydrate
composed of seven D-glucopyranose units some of whose
a
Institut de Chimie des Milieux et Matériaux de Poitiers, Université de Poitiers-
ENSIP, 1 rue Marcel Doré, 86022 Poitiers, France
b
Unité de Catalyse et de Chimie du Solide (UCCS), CNRS, UMR 8181,
Université d’Artois, Rue Jean Souvraz, SP 18, F-62307 Lens, France.
E-mail: sebastien.tilloy@univ-artois.fr; Fax: +33 321791755; Tel: +33 321791754
†
Electronic supplementary information (ESI) available. See DOI:
0.1039/c4gc00591k
1
3876 | Green Chem., 2014, 16, 3876–3880
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
Fig. 1 DMU–RAME-β-CD (w/w (%): 70/30) at room temperature (left)
and at 90 °C (right).
Table 1 Description of the various solvents based on DMU–CDs
a
mixture (70/30, w/w (%))
b
b
Melting point
(°C) (visually)
Melting point
(°C) (DSC)
Viscosity (cP) Fig. 2 Viscosity of DMU–CD (w/w (%): 70/30) low melting mixtures at
CD
at 90 °C
9
0 °C vs. percentage of water in DMU–CD.
β-CD
90
90
90
No melt
85
91
92
91
No melt
91
1165
445
235
—
205
495
CRYSME-β-CD
RAME-β-CD
TRIME-β-CD
HP-β-CD
containing TPPTS was dissolved in D
2
O after 10 heating–
cooling cycles (from room temperature to 90 °C and then to
3
1
1
HTMAP-β-CD
80
87
room temperature) under nitrogen. The P{ H} NMR spectrum
exhibited a singlet at −6.9 ppm in the mixture DMU–RAME-
a
All these CDs are commercially available except HTMAP-β-CD
β-CD–D O (see ESI†). This spectrum showed no modification
synthesized as described in the literature. These CDs were used
2
b
without drying (see the Experimental part). Melting points are at of NMR signal before and after these heating–cooling cycles.
normal pressure in air.
These experiments demonstrated that TPPTS is not oxidized in
these media.
First, the role played by various solvents based on DMU and
stability of the melts, all the mixtures were heated overnight at CD (w/w (%): 70/30) was investigated in a hydroformylation
9
1
0 °C and then the mixtures were left at room temperature for reaction catalyzed by the Rh/TPPTS combination at 90 °C
0 hours. This cycle was repeated for 7 days without any (Table 2). 1-Decene was chosen as a solvent-insoluble substrate
1
6
decomposition of DMU or CD. The viscosities of the LMMs are in order to obtain a biphasic system. For the combination
important data that need to be addressed for application in DMU–CD (Table 2, entries 1–5), the conversions after one hour
a catalytic process. Consequently, these viscosities were deter- were situated between 37% and 99%. The values clearly
mined at 90 °C (Table 1).
depended on the nature of the CDs. In the presence of a
The highest value was obtained for the melt DMU–β-CD cationic CD (HTMAP-β-CD), the conversion was the lowest
whereas the lowest for DMU–HP-β-CD. These viscosities are (entry 1). The best results were obtained with native and
relatively high (>100 cP). These values are attributed to the methylated CDs. More precisely, the methylation degree of the
presence of an extensive hydrogen bond network between each CD affected the outcome of the reaction since the more the CD
component. As the viscosities are mainly affected by the water was methylated, the more the conversion was. The best result
content, viscosity measurements were performed by varying was obtained in the case of DMU–RAME-β-CD with a conver-
the percentage of water in the LMMs (up to 10% – Fig. 2). As sion of 99% (entry 5). It is worth mentioning that addition of
expected, the value decreased when water was added. All the water to the DMU–RAME-β-CD mixture has no beneficial effect
values of viscosity are in the same order above 8% of water in on the conversion (compare entry 5 with entries 6 and 7).
the LMMs.
Indeed, when 10% or 20% of the DMU–RAME-β-CD mixture
Hydroformylation and Tsuji–Trost reactions have been was replaced by water, the conversion decreased. This behavior
selected to demonstrate the interest of these new solvents. For was surprising since addition of water decreased the viscosity.
both these reactions, the widespread water-soluble ligand tris- The selectivity in aldehydes was high (91–95%) in each case
(
m-sulfonatophenyl)phosphine trisodium salt (TPPTS) has and the linear to branched aldehydes ratio (l/b) was equal to
been chosen to maintain the transition metal (rhodium or 2.3 (compare entries 1–5). This value is classical for hydro-
1
5
palladium) in the catalytic layer. Note that this polar ligand formylation experiments performed with TPPTS as a ligand in
is well soluble in all the mixtures based on DMU–CD (up to polar media and indirectly proved the integrity of the catalytic
1
7
3
% w/w). The integrity of this ligand in these solvents was con- system. In order to ensure that the reactants and/or products
3
1
1
trolled by P NMR experiments by heating–cooling cycles. did not react with the solvent, H NMR experiments were
As an illustration, a fraction of the solvent DMU–RAME-β-CD performed at the end of the reaction. The data clearly
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 3876–3880 | 3877

View Article Online
Paper
Green Chemistry
Table 2 Rhodium-catalyzed hydroformylation of 1-decene in various
a
solvents
Scheme 2 Palladium catalyzed cleavage of allyloctylcarbonate (n = 7)
or allyloctadecylcarbonate (n = 17).
b
c
d
Entry
Solvent (w/w (%): 70/30)
C (%)
S (%)
l/b
1
2
3
4
5
DMU–HTMAP-β-CD
DMU–HP-β-CD
37
43
66
76
99
76
34
99
36
27
17
9
91
92
93
94
95
96
96
90
90
86
88
42
2.3
2.3
2.3
2.3
2.3
2.4
2.3
2.6
2.4
2.6
1.9
2.9
DMU–β-CD
DMU–CRYSME-β-CD
DMU–RAME-β-CD
DMU–RAME-β-CD
DMU–RAME-β-CD
DMU–RAME-β-CD
DMU–choline chloride
DMU–water
e
6
f
7
8
g
9
10
11
12
Water–RAME-β-CD
Water
h
a
Fig. 3 Reaction media of Tsuji–Trost reaction at room temperature
Experimental conditions: Rh(acac)(CO)
05 μmol (5 eq.), 1-decene = 42 mmol (2000 eq.), solvent = 6 g, 90 °C,
0 bar CO–H (1/1), 1500 rpm, reaction time = 1 h. All CDs were used
2
= 21 μmol (1 eq.), TPPTS =
(left) and at 90 °C (right).
1
5
2
b
c
without drying (see Experimental part). C = 1-decene conversion. S =
d
e
aldehydes selectivity. Branched aldehyde (b) is racemic. 10% of the
f
DMU–RAME-β-CD mixture was replaced by 0.6 g of water. 20% of the DMU–RAME-β-CD as a solvent for catalytic processes. The clea-
g
DMU–RAME-β-CD mixture was replaced by 1.2 g of water. TPP was
vage of allyloctylcarbonate was carried out at 90 °C under
h
used instead of TPPTS – reaction time = 30 min. Only 6 g of water.
1
8
nitrogen with diethylamine as an allyl scavenger (Scheme 2).
Quantitative conversion of allyloctylcarbonate was observed in
only five minutes. NMR control experiments were performed
indicated that no cross reaction was observed. This point is and no side-product was observed. In order to evaluate the
very important since the integrity of each partner is preserved stability of the catalytic system, recycling experiments were
during the catalytic experiments in these new solvents. performed. The work-up was extremely simple. At the end of
Another experiment was performed in DMU–RAME-β-CD the reaction, the biphasic system was cooled down to room
solvent by using triphenylphosphane (TPP) instead of TPPTS temperature. The DMU–RAME-β-CD phase became solid and
as a ligand (entry 8). The reaction is completed after only the organic layer was simply removed under air-atmosphere
3
0 minutes instead of 1 hour in the case of TPPTS. Neverthe- without special precaution (Fig. 3). The new fresh organic layer
less, the organic layer was coloured, suggesting the presence of was added to the remaining DMU–RAME-β-CD phase and this
3
1
catalytic species. This observation was confirmed by P NMR mixture was heated at 90 °C under nitrogen to perform the
experiments performed on an organic layer which revealed recycling experiment (Fig. 4 – Exp. Cond. 1). The catalytic layer
that 30% of the initial amount of TPP was transferred to the remained active without loss of activity during 4 cycles. Then
organic layer. These experiments undoubtedly showed that a to test the robustness of the catalytic system, this DMU–RAME-
part of rhodium was lost in the organic layer. This behaviour β-CD phase containing the catalyst was left overnight in the
is not in accordance with the specifications required for an solid state under air-atmosphere and was used the next day for
application in a biphasic organometallic catalytic process another set of recycling experiments performed with allylocta-
since the product is polluted by the catalyst. This result decylcarbonate instead of allyloctylcarbonate (Fig. 4 – Exp.
showed that the use of a polar ligand such as TPPTS is prefer- Cond. 2). Interestingly, the catalytic layer remained active
able since the extraction of this ligand in the organic layer is without loss of activity during 4 cycles. This result showed not
3
1
completely avoided as demonstrated by P NMR experi- only the stability of the catalytic system under air-atmosphere
ments. As comparative data, another set of experiments in the solid state but also the efficiency of this system even if a
was performed in DMU–choline chloride, DMU–water, water– more apolar substrate was used. A third condition was tested
RAME-β-CD or water (entries 9–12) as a solvent. In each case, to evaluate the stability of the catalyst under more drastic con-
the conversion was lower compared to the experiments ditions. After removing the organic layer, the DMU–RAME-
performed in DMU–CD mixtures. So, the composition of the β-CD phase used for the first 8 runs and containing the cata-
solvent significantly affected the outcome of the reaction lyst was stirred and heated under air-atmosphere at 90 °C for
and the best result in terms of catalytic activity and selectivity 15 minutes. Then, a new fresh organic layer of allyloctadecyl-
was obtained in the case of the DMU–RAME-β-CD mixture carbonate was added to the DMU–RAME-β-CD phase and this
(
entry 5).
Secondly, the palladium-catalyzed cleavage of allylcarbo- nitrogen atmosphere (Fig. 4 – Exp. Cond. 3). The conversion
nates (Tsuji–Trost reaction) was chosen to extend the scope of dramatically decreased up to 15% and became null for the
mixture was heated at 90 °C under air-atmosphere instead of a
3878 | Green Chem., 2014, 16, 3876–3880
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
of water, respectively. These values corresponded to a percen-
tage of water equal to 3%, 3%, 2.1%, 1.8% and 3.6% for
the LMMs based on DMU–β-CD, DMU–CRYSME-β-CD,
DMU–RAME-β-CD, DMU–HP-β-CD and DMU–HTMAP-β-CD,
respectively.
Organic layers obtained after catalysis experiments were
1
analyzed by H-NMR on a Bruker DRX300 spectrometer
1
(
300 MHz for H nuclei) and by gas chromatography on a
Shimadzu GC-17A gas chromatograph equipped with a poly-
dimethylsiloxane capillary column (30 m × 0.32 mm) and a
flame ionization detector. Perkin-Elmer DSC Pyris 1 apparatus
was used for differential scanning calorimetry analysis. All
experiments were performed in the temperature range
−
1
6
5–105 °C, at a heating/cooling rate of 5 °C min . Three suc-
cessive cycles were done following the same program. Temp-
Fig. 4 Recycling experiment for palladium catalyzed cleavage of allyl- erature and heat flow rates were calibrated with a very pure
alkylcarbonates. Experimental condition 1: Pd(OAc)
TPPTS = 70 μmol (9 eq.), allyloctylcarbonate = 800 μmol (100 eq.),
diethylamine = 2.4 mmol (300 eq.), DMU–RAME-β-CD (w/w (%) = 70/
2
= 8 μmol (1 eq.), indium standard, whose melting temperature and enthalpy are
well known. Data acquisition (onset of the (solid + liquid) equi-
libria temperatures) and processing were done with Perkin-
Elmer’s Pyris™ software. A Brookfield Model LVDVII+Pro visco-
30) = 1.4 g, heptane = 1.4 g, 90 °C, 1200 rpm under N
2
, reaction time =
5
minutes. Experimental condition 2: The same conditions as in 1 except
allyloctadecylcarbonate instead of allyloctylcarbonate. Experimental meter with the SC4-25 spindle at 200 rpm was used in this
condition 3: The same conditions as in 2 except under air instead of study. Viscosity was measured at 90 °C in a SC4-13RPY sample
2
under N .
chamber with an embedded RTD temperature probe. The
temperature was controlled with a TC502 circulating bath
refrigerated using a programmable controller. The sample
volume was slightly adapted to the volume of the sample
chamber. The program parameters and results analysis were
controlled with the Rheocalc@PC software.
following experiment. In this case, due to the heating under
air, O dramatically diffused into the DMU–RAME-β-CD phase
2
and TPPTS was transformed to its oxide. Consequently, the
latter was unable to stabilize palladium species that led to a
deactivation of the catalytic system. These experiments high-
lighted two interesting behaviours. Firstly, the catalyst species
dissolved in DMU–RAME-β-CD is recyclable at least 8 times
without loss of activity and allows the change of substrates.
Secondly, at room temperature, the catalyst species maintained
in the solvent in the solid state was protected against degra-
dation. So, the recycling work-up is very convenient and this
kind of solvent offers very simple handling under protective
conditions for the catalytic system.
Catalytic experiments
In
a typical hydroformylation experiment, Rh(acac)(CO)
2
(
5.5 mg; 21 μmol; 1 eq.), TPPTS (60 mg; 105 μmol; 5 eq.) and
solvent constituents in the solid state (4.2 g of N,N′-dimethyl-
urea and 1.8 g of RAME-β-CD (w/w(%) = 70/30) for example)
were charged into a 25 mL autoclave. Air was replaced by 20
2
bar of CO–H (1/1), and after heating at 90 °C, the mixture was
stirred using a multipaddle unit (1500 rpm) for 1 hour for an
incubation period. The stirring was then stopped, and after
cooling and depressurization, 1-decene (6 g; 42 mmol;
2
000 eq.) was introduced under nitrogen into the autoclave.
The medium was heated at 90 °C and then stirred (1500 rpm)
for 1 hour at this temperature under 50 bar of CO–H (1/1).
In a typical Tsuji–Trost experiment, Pd(OAc) (1.75 mg;
Experimental
Materials
2
All chemicals were purchased from Fisher Scientific and
2
Aldrich Chemicals in their highest purity. All chemicals were 8 μmol; 1 eq.), TPPTS (40 mg; 70 μmol; 9 eq.), N,N′-dimethyl-
used as supplied without further purification. β-CD and urea (0.98 g) and RAME-β-CD (0.42 g) (w/w (%) = 70/30) were
CRYSME-β-CD were provided by Roquette Frères and RAME- introduced into a Schlenk tube. Then, the reaction mixture
β-CD was furnished by Wacker. HTMAP-β-CD was obtained by was stirred under nitrogen for 1 hour in an oil bath at 90 °C
1
9
a chemical modification of β-CD described by Deratani et al.
and at 1200 rpm for an incubation period during which the
HP-β-CD was purchased from Aldrich Chemicals. As commer- medium became fluid and limpid. After cooling at room temp-
cial CDs are known to contain water, we have determined the erature, this catalytic layer became quickly solid, and allyloctyl-
2
0
percentage of water for each CD. The percentage of water in carbonate (0.167 g; 800 μmol; 100 eq.) and diethylamine
each CD was determined by heating CD in a Schlenk tube (0.171 g; 2.4 mmol; 300 eq.) dissolved in heptane (1.4 g) were
under vacuum (2 h at 80 °C, 2 h at 100 °C and then 2 h at added to it. After immersion of the Schlenk tube in the oil
1
20 °C). The difference of the weight allowed obtaining the bath at 90 °C, the solid catalytic layer became quickly liquid
quantity of water. The β-CD, CRYSME-β-CD, RAME-β-CD, HP- again and the biphasic system was then stirred for 5 minutes
β-CD and HTMAP-β-CD contained 10%, 10%, 7%, 6% and 12% at 90 °C. After cooling, the organic layer was very easily recov-
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 3876–3880 | 3879

View Article Online
Paper
Green Chemistry
ered under air-atmosphere without special precaution and the
solid catalytic layer was then recycled after introduction under
nitrogen of a new solution of allyloctylcarbonate and diethyl-
amine in heptane.
2 Q. H. Zhang, K. D. Vigier, S. Royer and F. Jérôme, Chem.
Soc. Rev., 2012, 41, 7108.
3 C. Russ and B. König, Green Chem., 2012, 14,
2969.
4
M. Francisco, A. van den Bruinhorst and M. C. Kroon,
Angew. Chem., Int. Ed., 2013, 52, 3074.
Conclusions
5 B. K. Tang and K. H. Row, Monatsh. Chem., 2013, 144,
427.
1
To summarize, we have reported for the first time the use of
LMMs based on β-CD series and DMU as solvents for hydro-
formylation and Tsuji–Trost reactions. These solvents can be
formed by simply mixing the starting materials (commercially
available), thus by-passing all problems of purification and
waste disposal. They are not volatile solvents and not flam-
mable, making their storage convenient. They are easily re-
coverable due to the nonexistence of chemical reaction during
their formation. They can be prepared from cheap, readily
available and toxicologically well characterized starting
materials. Thanks to these positive points, these LMMs have
now become of growing interest in organometallic catalysis.
In both catalytic reactions, the transition metal is efficiently
maintained in the solvent by using TPPTS as a polar ligand. In
hydroformylation reaction, the use of the solvent DMU–RAME-
β-CD allowed to reach higher catalytic activities than those
reported for classical systems based on CD in water as a
solvent. In the case of Tsuji–Trost reaction, both high catalytic
activity and recyclability are obtained. In addition, this solvent
offered simple handling and protection of the catalytic species
confined in this solvent in the solid state. We are currently
further studying other LMMs based on various CDs by varying
the size (α- or γ-CD) and/or the nature of the substituents for
applications in organometallic catalysis. We will also explore
the capacity of CDs to form inclusion complexes with organic
substrates in these LMMs.
6
7
8
9
G. Imperato, E. Eibler, J. Niedermaier and B. König, Chem.
Commun., 2005, 1170.
G. Imperato, S. Hoger, D. Lenoir and B. König, Green
Chem., 2006, 8, 1051.
G. Imperato, R. Vasold and B. König, Adv. Synth. Catal.,
2
006, 348, 2243.
S. Gore, S. Baskaran and B. König, Green Chem., 2011, 13,
009–1013.
0 F. Hapiot, A. Ponchel, S. Tilloy and E. Monflier, C. R. Chim.,
011, 14, 149.
1
1
1
1
2
1 F. Hapiot, H. Bricout, S. Tilloy and E. Monflier, Top. Curr.
Chem., 2013, 342, 49.
2 C. Machut-Binkowski, F. X. Legrand, N. Azaroual,
S. Tilloy and E. Monflier, Chem. – Eur. J., 2010, 16,
10195.
1
1
3 F. X. Legrand, N. Six, C. Slomianny, H. Bricout,
S. Tilloy and E. Monflier, Adv. Synth. Catal., 2011, 353,
1325.
4 D. N. Tran, F. X. Legrand, S. Menuel, H. Bricout, S. Tilloy
and E. Monflier, Chem. Commun., 2012, 48, 753.
1
1
5 E. G. Kuntz, CHEMTECH, 1987, 570.
6 In the autoclave containing DMU–CD, the rhodium precur-
sor and TPPTS were added. Then, the reaction mixture was
stirred for 1 hour at 90 °C under 20 bar of CO–H mixture
2
(
1/1) for an incubation period. After cooling, the decene
was added and the biphasic system was stirred for 1 hour
at 90 °C under 50 bar of CO–H (1/1).
2
17 F. Hapiot, L. Leclercq, N. Azaroual, S. Fourmentin, S. Tilloy
and E. Monflier, Curr. Org. Synth., 2008, 5, 162.
Acknowledgements
1
8 In a Schlenk tube containing DMU–CD, the palladium pre-
cursor and TPPTS were added. Then, the reaction mixture
was stirred for 1 hour at 90 °C for an incubation period.
After cooling, allyloctylcarbonate and diethylamine dis-
solved in heptane were added and the biphasic system was
stirred for 5 minutes at 90 °C.
Roquette Frères (Lestrem, France) is acknowledged for gener-
ous gifts of cyclodextrins. Dominique Prevost and Jean Paul
Cavrot are acknowledged for their technical assistance and
their fruitful discussions.
1
9 A. Deratani, G. Lelièvre, T. Maraldo and B. Sébille, Carbo-
hydr. Res., 1989, 192, 215.
Notes and references
1
A. P. Abbott, G. Capper, D. L. Davies, R. K. Rasheed and 20 N. G. Hadaruga, D. I. Hadaruga and H. D. Isengard, Food
V. Tambyrajah, Chem. Commun., 2003, 70. Chem., 2012, 132, 1741.
3880 | Green Chem., 2014, 16, 3876–3880
This journal is © The Royal Society of Chemistry 2014