Ionic liquid binary mixtures: Promising reaction media for carbohydrate conversion into 5-hydroxymethylfurfural

Article abstract of DOI:10.1016/j.apcata.2014.05.039

The conversion of carbohydrates into 5-hydroxymethylfurfural (5-HMF) has been studied in binary mixtures of ionic liquids (ILs), using strongly acidic resin Amberlyst 15 as the catalyst. In particular, both mono- and disaccharides, such as fructose, glucose and sucrose have been investigated. Considering the favorable effect exerted by chloride-based ionic liquids in the dissolution of carbohydrates, we used binary mixtures of 1-butyl-3-methylimidazolium chloride ([bmim][Cl]) with [bmim⁺] based ionic liquids differing in size, shape and coordination ability of the anion ([bmim][BF₄], [bmim][N(CF₃SO₂)₂], [bmim][N(CN)₂], [bmim][SbF₆] and [bmim][CF₃SO₃]). Carbohydrate conversion in [bmim][BF₄]/[bmim][Cl] binary mixtures, has been studied under both magnetic stirring and ultrasound (US) activation. The catalytic system used led to the formation of 5-hydroxymethylfurfural in good yield under mild conditions. A significant influence of IL binary mixture composition on the outcome of the target processes was evidenced. Improvements in both reaction time and temperature have been observed, under US activation.

Full text of DOI:10.1016/j.apcata.2014.05.039

Applied Catalysis A: General 482 (2014) 287–293
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Ionic liquid binary mixtures: Promising reaction media for
carbohydrate conversion into 5-hydroxymethylfurfural
Francesca D’Anna∗, Salvatore Marullo, Paola Vitale, Carla Rizzo,
Paolo Lo Meo, Renato Noto∗
Dipartimento STEBICEF, Sezione di Chimica, Università degli Studi di Palermo, Viale delle Scienze, Parco d’Orleans II, Ed. 17, Palermo, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
The conversion of carbohydrates into 5-hydroxymethylfurfural (5-HMF) has been studied in binary
mixtures of ionic liquids (ILs), using strongly acidic resin Amberlyst 15 as the catalyst. In partic-
ular, both mono- and disaccharides, such as fructose, glucose and sucrose have been investigated.
Considering the favorable effect exerted by chloride-based ionic liquids in the dissolution of car-
bohydrates, we used binary mixtures of 1-butyl-3-methylimidazolium chloride ([bmim][Cl]) with
[bmim+] based ionic liquids differing in size, shape and coordination ability of the anion ([bmim][BF4],
[bmim][N(CF3SO2)2], [bmim][N(CN)2], [bmim][SbF6] and [bmim][CF3SO3]). Carbohydrate conversion in
[bmim][BF4]/[bmim][Cl] binary mixtures, has been studied under both magnetic stirring and ultrasound
(US) activation. The catalytic system used led to the formation of 5-hydroxymethylfurfural in good yield
under mild conditions. A significant influence of IL binary mixture composition on the outcome of the tar-
get processes was evidenced. Improvements in both reaction time and temperature have been observed,
under US activation.
Received 6 February 2014
Received in revised form 23 May 2014
Accepted 27 May 2014
Available online 13 June 2014
Keywords:
Ionic liquids
Carbohydrate conversion
Ultrasounds
5-Hydroxymethylfurfural
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
The conversion of mono- and disaccharides into FUR and its
derivatives has been widely investigated in both water [8–14]
Energy demand in modern society and the negative conse-
quences of global warming during the last decade have raised
a growing interest in the scientific community toward renew-
able energy sources. In this scenario, the production of energy
by means of solar cells or from bio-resources is an issue of
paramount importance. In particular, as far as vegetable biomass
is concerned, much attention has been paid to the production
of industrially relevant intermediates from carbohydrates. These
are a major component of biomass derived from wood process-
ing by-products, agricultural surplus and so on. The dehydration of
carbohydrates can give rise to important chemicals, such as furfural
(FUR) and 5-hydroxymethylfurfural (5-HMF), which are currently
obtained from petroleum manufacturing, and are used as starting
materials for many industrial processes [1–7]. Thus, the implemen-
tation of carbohydrate conversion processes might have significant
outcomes from both an economical and, most importantly, an envi-
ronmental point of view.
and conventional organic solvents [15–22]. Very recently, with the
introduction of ionic liquids (ILs) as alternative reaction media,
many studies have been directed toward the possibility to study
these processes changing the nature of both the solvent media
and the catalyst, or even varying other important operational
parameters such as the reaction time and temperature dependence
of different heating methods used [23–33]. Regarding this topic,
different reports have analyzed the effect of microwaves on car-
bohydrate conversion [34–38]. On the other hand, US activation
has been little investigated and it has been mostly applied only
to improve carbohydrate dissolution in the solvent media used
[39–42].
The use of ILs is advantageous because of their low flammability
and possible reuse of the reaction media. Reports have highlighted
that conversion of mono- and polysaccharides is favored in partic-
ular by chloride based ILs. Indeed, the high coordination ability of
the chloride anion allows the formation of hydrogen bonds with
the hydroxyl groups of the sugar units, favoring the dissolution of
the substrate and, on the whole, its dehydration [43–45]. However,
from a practical point of view the use of halide based ILs poses some
limitations. As a matter of fact, these ILs are usually solid at room
temperature and the liquid phases obtained on their melting show
∗
Corresponding authors. Tel.: +39 09123897540.
E-mail addresses: francesca.danna@unipa.it (F. D’Anna), renato.noto@unipa.it
(R. Noto).
http://dx.doi.org/10.1016/j.apcata.2014.05.039
0926-860X/© 2014 Elsevier B.V. All rights reserved.

288
F. D’Anna et al. / Applied Catalysis A: General 482 (2014) 287–293
ILs used:
high viscosities. Consequently, high reaction temperatures have
to be used in order to overcome these problems. This frequently
causes the reaction mixture to turn brown, with the concomitant
degradation of FUR and its derivatives into by-products of low
industrial value. Furthermore, as the degradation processes of FUR
are favored by water, an additional problem stems from the signif-
icant hygroscopicity of halide based ILs, making it difficult to carry
out the target process in perfectly anhydrous media. From an envi-
ronmental point of view, also the good water solubility and high
thermal stability of halide based ILs pose some problems. Indeed,
as a consequence of their possible release, the aqueous solubility
might induce unwanted effects in aquatic ecosystems. On the other
hand, the high thermal stability might favor their bioaccumulation
[46–49].
Several attempts have been made to lower the viscosity of halide
based ILs, and the use of conventional solvents/IL binary mixtures
has been suggested [50]. However, this option could increase the
environmental impact of the process. Therefore, to overcome the
aforementioned problems, and bearing in mind the role played
by chloride based ILs, we reasoned that the use of IL binary mix-
tures could be a valid alternative. Recent reports have highlighted
some peculiar features of IL binary mixtures, which frequently
show a strikingly different behavior with respect to the pure com-
ponents, in particular as far as mixtures of anions are concerned
[51–54]. Indeed, in the latter case, differences in size and coordi-
nation abilities of anions may give rise to either an ideal behavior
or to a solvent system characterized by the presence of different
micro-domains [55]. In this area, using the mononuclear rear-
rangement of heterocycles as a probe reaction, we have recently
demonstrated that mixing of cations having significantly different
electronic properties, induced a gradual change in the properties
of the binary mixtures which perfectly reside between those cor-
responding to pure components. On the other hand, mixing anions
differing in size, shape and coordination ability produced a dis-
continuous behavior, so that the properties of the binary mixtures
obtained could not be related to the pure ILs [56].
In view of this, we studied on the conversion of carbohydrates
into 5-HMF using IL binary mixtures such as [bmim][X]/[bmim][Cl],
i.e. ILs all sharing the imidazolium cation, but differing in size, shape
and coordination ability of the anion. This study was mainly aimed
at identifying the most suitable co-solvent able to decrease the
[bmim][Cl] viscosity, thus allowing the target process to be car-
ried out close to room temperature. Fructose, glucose and sucrose
were used as substrates (Scheme 1).
Carbohydrate conversion was carried out under acidic cataly-
sis; in particular, we tested both species derived from renewable
sources such as citric and tartaric acid, as well as strongly acidic
resins differing in particles size, such as Amberlyst 15 and Dowex
50W X8 (hereafter indicated as Dowex). Reactions were initially
carried out under magnetic stirring. However, taking into account
that the sustainability of a process heavily depends on the amount
of energy used, we also investigated the effect of US activation on
the target processes, with the aim to decrease the reaction temper-
ature and time.
[bmim][X] = [bmim][BF4]
[bmim][SbF6]
X
N
N
[bmim][N(CF3SO2)2]
[bmim][N(CN)2]
[bmim][CF3SO3]
[bmim][X]
X = Cl, BF4, SbF6, N(CF3SO2)2, N(CN)2, CF3SO3
The reactions studied:
HO
OH
O
H
OH
H
OH
H
HO
OH
H
Fructose
H
O
O
HO
HO
H
O
H
OH
OH
H
OH
OH
H
Glucose
5-HMF
H
O
HO
HO
HO
H
H
OH
H
O
O
H
OH
H
OH
H
HO
Sucrose
Scheme 1. Schematic representation of ILs used and reactions studied.
[bmim][Cl], [bmim][BF4] and [bmim][CF3SO3] were pur-
chased from Iolitec. [bmim][N(CF3SO2)2], [bmim][SbF6] and
[bmim][N(CN)2] were prepared according to previously reported
procedures [57–59]. Before use, all ILs were dried on a vacuum line
at 343 K for 2 h, lyophilized for 48 h at 0.06 mbar, and stored in a
desiccator under argon and over calcium chloride.
Each IL binary mixture was prepared by weighing the required
amount of IL into a round bottom flask. To favor mixing, each mix-
ture was vigorously stirred, sonicated for 1 min (45 kHz, 200 W) and
then left to equilibrate overnight. In all cases the mixtures appeared
to be homogeneous after this treatment.
2.2. General procedure for carbohydrate conversion under
magnetic stirring
A suitable amount of carbohydrate (0.025 g) was weighed in
a round bottom flask containing the pure IL or IL binary mix-
ture (0.5 g). To favor carbohydrate dissolution, the mixture was
stirred at 353 K for 30 min under argon. In all cases, the mixtures
appeared to be homogeneous after this treatment. After equili-
bration at the reaction temperature, the suitable amount of acidic
catalyst (0.025 g for fructose and glucose and 0.05 g for sucrose) was
added. After reaction time, each sample was diluted with 0.250 g
of ultra pure water, stirred at room temperature and diluted with
methanol in such a way to have a 5-HMF concentration ranging
from 7 × 10−6 M up to 7 × 10−5 M. The concentration of 5-HMF was
determined from UV absorbance recorded at 277 nm. The presence
of 5-HMF in the reaction mixture, as single UV absorbing product,
was further verified by means of TLC on silica gel by comparison
with a standard sample (eluent: ethyl acetate/methanol 5:1, v/v).
2. Experimental
2.1. Materials
2.3. General procedure for carbohydrate conversion under US
activation
Fructose (≥99.0), glucose (≥99.5), sucrose (≥99.5), citric and
tartaric acids (≥99.5) were purchased from Sigma–Aldrich and
used without further purification. Amberlyst 15 (hydrogen form)
and Dowex 50W X8 resins were purchased and kept before use
overnight in a desiccator at 373 K, under vacuum and over phos-
phorus pentoxide.
Sonochemical reactions were carried out in a thermostated
ultrasonic bath operating at a frequency of 45 kHz, with an out-
put power of 200 W. The bath was filled with a 1% (w:w) aqueous
solution of the surfactant Contrad 2000. Prior to each reaction

F. D’Anna et al. / Applied Catalysis A: General 482 (2014) 287–293
289
we identified the portions of the bath experiencing the maxi-
mum ultrasonic intensity by aluminum foil mapping [60]. For this
purpose, a sheet of aluminum foil (15 cm × 7 cm) was immersed
vertically in the bath and subjected to sonication for 30 s. The dis-
tance at which the sheet underwent the maximum damage was
chosen as the horizontal position of the reaction vessel. As for the
vertical position, the bottom of the reaction vessel was placed 5 cm
beneath the water surface. Samples were prepared in a tube having
a 10 cm length, and subjected to the same analysis reported above
for the reactions carried out under magnetic stirring.
The data show that the course of the target reaction is signifi-
cantly affected by the different nature of the IL used as co-solvent.
Indeed, yields significantly decreased on going from [bmim][BF4] to
[bmim][CF3SO3] according to the following trend: [bmim][BF4] >
[bmim][N(CN)2] > [bmim][N(CF3SO2)2] > [bmim][SbF6] > [bmim]
[CF3SO3]. Considering that a higher anion coordination ability
favors the dissolution of carbohydrates in ILs [43–45], it can easily
be noticed that experimental data cannot be correlated with the
anion ability of the co-solvent IL to accept a hydrogen bond, as
accounted for by ˇ parameter values. Indeed, for the ILs used, ˇ
decreases along the series: [bmim][N(CN)2] > [bmim][CF3SO3] >
[bmim][BF4] > [bmim][N(CF3SO2)2] > [bmim][SbF6] (ˇ = 0.71, 0.46,
0.38, 0.24 and 0.15, respectively) [66]. This suggests that carbohy-
drate dissolution is not the only factor determining the outcome
of the target reaction. Furthermore, according to previous reports
[55], this result indicates that the behavior of IL binary mixtures
cannot be simply related to the properties of the pure components.
2.4. General procedure for the determination of acid exchange
between the resin and ILs
The determination was carried out using a previously reported
procedure [61]. In a typical experiment, 0.025 g of Amberlyst
15 were added to the suitable [bmim][Cl]/[bmim][BF4] solvent
mixture and subjected to sonication or stirring under the best reac-
tion conditions found for each substrate. Then 5 mL of ultrapure
water were added to the mixture and the resin was filtered off.
The amount of acid dissolved in the resulting aqueous solution
was determined by potentiometric titration with standard NaOH
(0.02 M). No appreciable change in pH was observed in a blank test
after stirring or sonicating the same amount of resin in 5 mL of
ultrapure water, in the absence of ILs.
3.2. Experimental conditions
Having identified the most suitable co-solvent, we studied the
conversion of fructose in the presence of different acidic catalysts.
We reasoned that it could be interesting to investigate the use of
either carboxylic acids derived from renewable sources, such as
citric and tartaric acids or, on the other hand, solid recyclable cata-
lysts such as acidic resins. Focusing in particular on solid catalysts,
we considered that differences in size of resin particles could be
an important factor in determining the outcome of the process
studied. In this light, the target reaction was performed using the
strongly acidic resins Dowex and Amberlyst 15. The data collected
at 298 K by using a [bmim][BF4]/[bmim][Cl] binary mixture, at XCl
∼0.5, are reported in Table S2 of the Supplementary Material.
It was found that Amberlyst 15 was the most effective catalyst
under our experimental conditions. To explain these results, we
have to consider the nature of the resins, the amount of acidic sites
and particle size. Amberlyst 15 is a macroporous resin, whereas
Dowex is a gel resin. Our results account for a stronger activity of
the macroporous resin with respect to the gel one and this agrees
with recent reports [26]. Differently from previous observations
in literature [67,68], in our case HMF yields increased on increas-
ing the particle size of resin (Amberlyst 15: 18–23 mesh; Dowex:
20–50 mesh). However, the catalysts used have a different amount
of acidic sites (higher for Amberlyst 15 than for Dowex). To ascer-
tain if the results obtained could be rationalized as a function of
the latter parameter, we also performed fructose conversion using
equivalent amounts of the catalysts (i.e. 25 mg of Amberlyst 15 or
75 mg of Dowex). 5-HMF yields were very different (56% and 6%
in the presence of Amberlyst 15 or Dowex respectively), probably
indicating that the trend obtained was significantly affected by the
nature of the resin.
3. Results and discussion
3.1. The effect of co-solvent nature
Initially, we screened for the IL to be able to behave as the
most suitable co-solvent for the conversion of fructose into 5-
HMF. To pursue this goal, we used [bmim][BF4], [bmim][N(CN)2],
[bmim][N(CF3SO2)2], [bmim][SbF6] and [bmim][CF3SO3]. The ILs
used also have very different viscosities (ꢀ = 40,890, 154, 90, 52
and 39 cP at 293 K for [bmim][Cl], [bmim][BF4], [bmim][CF3SO3],
[bmim][N(CF3SO2)2] and [bmim][N(CN)2] respectively; ꢀ = 108 cP
for [bmim][SbF6] at 298 K) [62–65]. The viscosity of IL binary
mixtures could be significantly different than those of pure ILs.
However, in all cases, the addition of IL co-solvent should signif-
icantly decrease the viscosity of solvent mixtures.
Then, we performed the reaction without the acid catalyst. In
all cases we did not observe the formation of 5-HMF. In the pres-
ence of catalyst we carried out the reaction at 298 K, using various
binary mixtures of ILs having a 1:1 [Cl−]/anion molar ratio (here-
after expressed as [bmim][Cl] mole fraction, XCl ∼0.5). The results
collected are illustrated in Fig. 1, which reports 5-HMF yields as a
function of the different nature of ILs used (the data are summarized
in Table S1 of the Supplementary Material).
Previous studies report that in many cases the use of Brønsted
solid acids in ILs gives rise to homogeneous catalysis [61]. In order
to get deeper insights, we carried out investigations to verify the
presence of homogeneous acidic species using a method previously
reported [61]. Under both magnetic stirring and US activation, we
detected the presence of 0.07 mmol of acid dissolved in the sol-
vent mixture, evidencing homogeneous as well as heterogeneous
catalysis.
Of course, before studying the effect of IL binary mixtures on the
efficiency of glucose and fructose conversion, we tried to evaluate
in both cases the relevance of all other experimental parameters
in optimizing the target processes. So, we carried out the carbo-
hydrate conversion varying the reaction times and temperatures,
as well as the catalyst/substrate ratio and the solvent/substrate
weight ratio. Also in these cases, we used a [bmim][BF4]/[bmim][Cl]
binary mixture at XCl ∼0.5 as the solvent system. Relevant data
Fig. 1. 5-HMF yields as
0.5 g of [bmim][X]/[bmim][Cl] binary mixture (XCl = 0.5, mAmberlyst 15 = 0.025 g,
mfructose = 0.025 g, t = 3 h).
a function of different IL nature at 298 K, in

290
F. D’Anna et al. / Applied Catalysis A: General 482 (2014) 287–293
Table 1
Material). Indeed, both increasing the amount of catalyst and dilut-
ing the reaction mixture resulted in a lower yield. A larger amount
of catalyst may negatively affect the reaction favoring the hydration
and poly-condensation of 5-HMF. According to previous reports
[26] this may result in a more difficult mass transfer. On the other
hand, an increase in the initial carbohydrate loading may cause con-
comitant secondary reactions between sugar and FUR, giving rise
to the formation of mixed polymers such as humans [78] humins.
5-HMF yields obtained from fructose and glucose conversion in
[bmim][Cl]/[bmim][BF4], as a function of reaction time and temperature.a
Fructose
Glucose
Time (h) Temperature
(K)
Yield (%)b
Time (h) Temperature
(K)
Yield (%)b
1
2
3
3
4
5
298
298
298
313
298
298
7
34
56
52
38
35
0.5
1
1.5
2
3
3
298
298
298
298
313
298
363
298
0
0
5
32
0
15
14
5
3.3. The effect of IL binary mixture composition
3
4
The last factor to be analyzed was the effect of composition of
the IL binary mixture. As we pointed out above, IL binary mixtures
can give rise to different behaviors according to whether differ-
ent cations or anions are mixed. Literature reports underline that
mixtures of anions generally show the most distinct behavior. In
these cases, the largest deviations from the ideal behavior have
been observed when the anions possess large differences in size
[55].
a
Reaction conditions: XCl = 0.5, mcarbohydrate = 0.025 g, mAmberlyst 15 = 0.025 g.
Yields were reproducible within 2%.
b
are presented in Table 1, whereas results about the effect of sub-
strate/catalyst and substrate/solvent ratios are reported in Tables
S3 and S4 of the Supplementary Material.
On the grounds of this information, and using the optimized
experimental conditions identified above (see Section 3.2), we
studied both fructose and glucose conversion into 5-HMF across the
whole [Cl−]/anion mole fraction range. For the sake of clarity, the
data collected are presented in Fig. 2 (5-HMF yields as a function of
[bmim][Cl] mole fraction, for both fructose and glucose conversion
are reported in Table S5 of the Supplementary Material).
In both cases 5-HMF yields show a bell-shaped trend centered
at XCl ∼0.5. Low yields detected in pure [bmim][BF4] or in binary
mixtures richer in this component could be due to the lower ability
of [BF4−] in favoring the dissolution of the sugar. On the other hand,
low yields detected in pure [bmim][Cl], or in IL binary mixtures
richer in this component, could be ascribed to the large viscosity of
the solvent media, which probably hampers the approach between
the reagents.
It is noteworthy that, unlike fructose, few effective systems are
reported for glucose conversion to 5-HMF. In the case of fruc-
tose, conversion to 5-HMF has been already studied in [bmim][Cl]
and in the presence of Amberlyst 15, at 353 K (reaction time:
10 min, 5-HMF yield: 82%).[69] As far as glucose is concerned, it
has been frequently reported that best experimental conditions
for its dehydration require the use of a basic catalyst. This favors
the isomerization of glucose to fructose, which in turn more eas-
ily undergoes the dehydration process. Generally, strong mineral
or Lewis acids have been used at temperatures higher than 363 K
(reaction times: 1–4 h, 5-HMF yields: 17–81%) [36,70–73]. Under
our experimental conditions, the dehydration of glucose occurred
under much milder conditions. Furthermore, to the best of our
knowledge [74], this is one of the few examples in which glucose
conversion has been performed using an acidic resin.
Therefore, our results suggest that the outcome of the reaction
is a consequence of a balance between the favorable effect resulting
from a higher concentration of chloride ion, able to form hydrogen
bonds with sugar units, and the unfavorable one due to the higher
viscosity of the solvent medium.
The data show that, at equivalent reaction times (t = 3 h), a higher
temperature did not improve 5-HMF yields. Furthermore, in both
cases, an increase in temperature caused the browning of the reac-
tion system. According to previous reports [75–77], we ascribed
the observed browning to the presence of by-products such as
polymers and humins.
As far as the reaction time is concerned, our data indicate that
also under relatively mild conditions (T = 298 and 333 K for fructose
and glucose respectively), prolonged contact between catalysts and
5-HMF exerts a negative effect on the process efficiency. Indeed,
the 5-HMF yield gradually decreased after three hours for fruc-
tose, whereas in the case of glucose the best 5-HMF yield (32%)
was obtained after two hours.
Finally, for both fructose and glucose the best 5-HMF yields were
obtained using 25 mg of sugar, 25 mg of catalyst and a 20/1 sol-
vent/substrate weight ratio (Tables S3 and S4 of the Supplementary
3.4. Sucrose conversion
It is well known that the demanding energy of sucrose con-
version into 5-HMF is larger than the one of fructose or glucose.
Indeed, this reaction generally requires high reaction temperatures
and strong acidic catalysts. According to literature, the above pro-
cess has been studied both in conventional organic solvents and in
IL solution, in the temperature range 363–413 K and using acidic
catalysts such as HCl or Lewis acids [79–83]. In order to test the
efficiency of our reaction system in this process, we analyzed the
Fig. 2. 5-HMF yields as a function of IL binary mixtures composition: (a) from fructose at 298 K; (b) from glucose at 333 K (mcarbohydrate = 0.025 g, mAmberlyst 15 = 0.025 g,
tfructose = 2 h, tglucose = 3 h).

F. D’Anna et al. / Applied Catalysis A: General 482 (2014) 287–293
291
vapor pressure and network of cation–cation and cation–anion
interactions [93].
Initially, we investigated the possible 5-HMF formation under
US activation in the absence of the acidic catalyst. However, in all
the cases examined, we did not observe the presence of the desired
product. Furthermore, we analyzed the possible dispersion of the
catalyst particles as a consequence of US application, but we did
not observe particles dispersion or loss in weight. The images of
the resin before and after US action are reported in Fig. S1 of the
Supplementary Material.
We tried to convert all the carbohydrates used in this work
under US activation. However, under our experimental conditions
(ultrasonic bath operating at 45 kHz and 200 W), we did not detect
the formation of 5-HMF from glucose. In this case, after US action in
the temperature range 298–343 K, only a faint browning of the reac-
tion mixture was observed. HPLC analysis of the reaction mixture
evidenced only a slight decrease in glucose concentration; minor
amounts of possible by-products such as 5-HMF, levulinic acid, lac-
tic acid were not detected even in trace amount. Consequently, the
results discussed hereafter are exclusively referred to fructose and
sucrose.
We searched for the best experimental conditions to obtain
5-HMF from fructose and sucrose under US activation. The first
experimental parameter we optimized was the reaction time, using
the best experimental conditions found for the reaction under mag-
netic stirring and with a IL binary mixture at XCl ∼0.5 as solvent
(the data collected are reported in Table S8 of the Supplemen-
tary Material). The data (reported in the Table) evidence that US
activation is able to favor the outcome of the target reactions. Fur-
thermore, in the case of fructose, at equivalent temperature (298 K),
the best 5-HMF yield was obtained in shorter reaction time (3 h
and 1 h under magnetic stirring and US activation respectively).
On the other hand, changes in the optimal reaction time were
not observed for sucrose (2 h under both magnetic stirring and
US activation). Nevertheless, a good increase in 5-HMF yield was
obtained (20 and 34% under magnetic stirring and US activation
respectively) at a significantly lower reaction temperature (363
and 298 K under magnetic stirring and US activation respectively).
Moreover, the data collected demonstrate also the negative out-
come of a prolonged contact between the acidic catalyst and the
reaction mixture.
We also tried to evaluate possible changes in yields as a function
of different catalyst/substrate weight ratios (the data collected are
reported in Table S8 of the Supplementary Material). In this case,
different trends were obtained for fructose and sucrose. While in
the case of sucrose the best catalyst/substrate weight ratio was the
same as the one employed under magnetic stirring, for fructose the
US promoted reaction required a lower amount of resin than the
one used under magnetic stirring.
In the latter case, comparable yields were obtained by using
weight ratios as large as 0.5 and 1. Probably, as far as the cata-
lyst concentration is concerned, differences in the behavior of the
substrates could be once again ascribed to the occurrence of dif-
ferent reaction pathways giving rise to 5-HMF from either fructose
or sucrose, and to different effects that US activation can exert on
each step.
Fig. 3. 5-HMF yields as a function of IL binary mixtures composition from sucrose
at 363 K (msucrose = 0.025 g, mAmberlyst 15 = 0.05 g, t = 2 h).
effect of temperature, reaction time, catalyst/substrate weight ratio
and solvent mixture composition. The data collected are reported
in Tables S6 and S7 of the Supplementary Material.
In agreement with the large energy demand of this process,
the investigation carried out using a solvent/substrate weight ratio
equal to 20/1 and a catalyst/substrate weight ratio equal to 2/1
showed that the largest 5-HMF yield (17%) can be obtained only
at 363 K. Such a reaction temperature is higher than the one used
in the case of fructose or glucose conversion, and affords a lower
yield.
Data previously reported in literature cannot be directly com-
pared. Indeed, to the best of our knowledge, Amberlyst 15 has only
been used in DMF solution at 393 K [84]. Furthermore, in the lat-
ter case 5-HMF yield was larger than the one obtained under our
experimental conditions (58 and 17% in DMF and IL binary mix-
ture respectively). On the other hand, in one of the few examples
of sucrose conversion without catalyst, the target reaction was
performed in [hmim][Cl] at 363 K and gave rise to a modest 5-
HMF yield (37%) [85]. Notwithstanding the yield obtained, also in
this case we analyzed the effect of different operational parame-
ters, such as reaction time, solvent/substrate and catalyst/substrate
ratios. Analysis of the data reported in Table S6 of the Supple-
mentary Material shows that the best yield was obtained in 2 h
using a catalyst/substrate weight ratio as large as 2/1. Indeed, an
increase in both reaction time and amount of catalyst led to a lower
5-HMF yield. On the other hand, increasing the reaction tempera-
ture induced a browning of the reaction mixture and a decrease in
5-HMF yield.
Furthermore, 5-HMF yield from sucrose conversion increased
up to XCl ∼0.3 and, unlike fructose or glucose conversion, the sub-
sequent increase in XCl did not affect the yield (Fig. 3).
This result probably stems from the balance of the different and
contrasting effects that solvent composition can exert on the dif-
ferent steps of the sucrose conversion, namely the hydrolysis of
glycosidic bond, the glucose–fructose isomerization and finally the
fructose conversion to 5-HMF.
3.5. The effect of US activation
As we stated previously (see Section 1), we were also interested
in the possibility of using US activation to promote the target pro-
cesses. Indeed, this alternative methodology has seldom been used
in studying the conversion of carbohydrates. On the other hand,
different reports have highlighted the suitability of ILs to carry
out ultrasound assisted reactions [86–92]. In particular, we have
recently studied the formation of aryl azides under US activation.
We have demonstrated that ILs are able to induce highly efficient
cavitation processes, generating favorable physical conditions for
the occurrence of the reaction. This is a consequence of their low
Data discussed hitherto allow us to underline some important
advantages deriving from use of US activation. Indeed, according to
our previous report [93], a high efficiency of the cavitation process
in IL solution can be proposed. In some cases (namely fructose),
this results in a reduction of the amount of catalyst needed as well
as of the reaction time. In other cases (namely sucrose), the high
efficiency of the cavitation process induces a significant decrease in
the reaction temperature. The use of ultrasound reduces the waste
of materials and energy, which represents a benefit from both an
environmental and economic point of view.

292
F. D’Anna et al. / Applied Catalysis A: General 482 (2014) 287–293
Fig. 4. 5-HMF yields, obtained under US activation, as a function of IL binary mixtures composition at 298 K from: (a) fructose (mfructose = 0.025 g, mAmberlyst 15 = 0.0125 g,
t = 1 h); (b) sucrose (msucrose = 0.025 g, mAmberlyst 15 = 0.05 g, t = 2 h).
In view of these results, we analyzed the effect of IL binary mix-
tures composition also under US activation. The data collected as a
function of XCl are illustrated in Fig. 4, and are reported in Table S9
of the Supplementary Material.
taking advantage from the ability of the latter IL to improve the
dissolution of carbohydrates.
Investigations carried out by using this solvent system evidence
how the results obtained are affected by the catalyst/substrate
and the solvent/substrate weight ratios. In all cases the IL binary
mixture composition plays a significant role in determining the
outcome of the reaction.
Comparison between data collected for fructose and sucrose
highlights the different relevance of the solvent mixture composi-
tion even under US activation. Indeed, in the case of fructose, 5-HMF
yields regularly change with XCl, whereas for sucrose this param-
eter barely affects the outcome of the reaction. However, in this
latter case it is noteworthy that US activation allows us to obtain,
under mild conditions (298 K), 5-HMF yields slightly lower than
those observed at 363 K. Differently from magnetic stirring, the
best 5-HMF yields require the use of IL binary mixtures that are
richer in chloride anion (XCl ∼0.5). Furthermore, bearing in mind
the absence of US activation effect on the conversion of glucose, it
can be proposed that 5-HMF originated from the fructose moiety.
By analogy with data collected under magnetic stirring, in the
case of fructose, the trend of 5-HMF yields as a function of XCl
is represented by a bell-shaped curve. Comparison with the data
collected under magnetic stirring accounts also in this case for sig-
nificant catalytic efficiency in IL binary mixtures rich in chloride
anions. Indeed, the highest 5-HMF yield was obtained at XCl ∼0.7
and yields collected at larger XCl values were larger than those col-
lected under magnetic stirring. The above finding can be ascribed
to the synergistic action of different factors, such as the effect of
chloride anion concentration on the outcome of the reaction and
the effect of the solvent viscosity on the ultrasound efficiency.
As stated previously (see Section 3.3), a solvent mixture richer in
the chloride anion should improve the dissolution of carbohydrates
in ILs and, consequently, their conversion to 5-HMF. However, the
presence of a larger chloride concentration also boosts the viscos-
ity of the solvent, slowing the reaction rate under magnetic stirring
as a consequence of more difficult mass transfer. By contrast, bear-
ing in mind that a higher solvent viscosity positively acts on the
cavitation efficiency [60], the increase in the chloride mole frac-
tion enhances the efficacy of ultrasounds, allowing us to explain
the trend observed.
To the best of our knowledge, this is one of the few reports
analyzing the effect of US activation in the formation of 5-HMF
from both mono- and disaccharides. Our results shed light also
on the advantages deriving from US activation of reaction mix-
tures. Indeed, the US assisted reactions proceed with lower reaction
times and temperatures than the ones needed under magnetic stir-
ring. Furthermore, in the case of fructose, the US action induces a
change in the composition of the solvent mixture giving rise to
the best 5-HMF yield (XCl ∼0.5 and 0.7 under magnetic stirring
and US activation respectively). Probably, because this binary mix-
ture is richer in [bmim][Cl], this result could be a consequence of
the higher efficiency of cavitation process in more viscous solvent
systems.
Acknowledgements
We thank University of Palermo (2012-ATE-0405) and Italian
MUR (FIRB 2010RBFR10BF5V) for financial support.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.apcata.
2014.05.039.
References
[1] A.A. Rosatella, S.P. Simeonov, R.F.M. Frade, C.A.M. Afonso, Green Chem. 13
(2011) 754–763.
[2] A.J. Crisci, M.H. Tucker, J.A. Dumesic, S.L. Scott, Top. Catal. 53 (2010) 1185–1192.
[3] H.B. Zhao, J.E. Holladay, H. Brown, Z.C. Zhang, Science 316 (2007) 1597–1600.
[4] J.C. Serrano-Ruiz, R.M. West, J.A. Dumesic, Annu. Rev. Chem. Biomol. Eng. 1
(2010) 79–100.
[5] R. Karinen, K. Vilonen, M. Niemelä, ChemSusChem 4 (2011) 1002–1016.
[6] M. Bicker, D. Kaiser, H.V.L. Ott, J. Supercrit. Fluids 36 (2005) 118–126.
[7] C.V. McNeff, D.T. Nowlan, L.C. McNeff, B. Yan, R.L. Fedie, Appl. Catal., A 384
(2010) 65–69.
[8] T.S. Hansen, J. Mielby, A. Riisager, Green Chem. 13 (2011) 109–114.
[9] M.B. Turner, S.K. Spear, J.D. Holbrey, R.D. Rogers, Biomacromolecules 5 (2004)
1379–1384.
[10] C. Moreau, R. Durand, S. Razigade, J. Duhamet, P. Faugeras, P. Rivalier, P. Ros, G.
Avignon, App. Catal., A 145 (1996) 211–224.
[11] T. Okano, K. Qiao, Q. Bao, D. Tomida, H. Hagiwara, C. Yokoyama, Appl. Catal., A
451 (2013) 1–5.
4. Conclusions
All of the data collected sheds light on the suitability of IL binary
mixtures as reaction media for the conversion of carbohydrates to
5-HMF. According to previous reports, our data show that the prop-
erties and catalytic activity of IL binary mixtures cannot be related
to the behavior of pure ILs [56].
Among the binary mixtures investigated, the [bmim][BF4]/
[bmim][Cl] mixture has been proven able to give the best 5-
HMF yields under mild conditions. The use of [bmim][BF4] as
co-solvent allows [bmim][Cl] to be utilized at room temperature,
[12] R.P. Swatloski, S.K. Spear, J.D. Holbrey, R.D. Rogers, J. Am. Chem. Soc. 124 (2002)
4974–4975.
[13] S.H. Thomas, M. Jerrik, R. Anders, Green Chem. 13 (2011) 109–114.

F. D’Anna et al. / Applied Catalysis A: General 482 (2014) 287–293
293
[14] Q. Xinhua, W. Masaru, M.A. Taku, L.S. Richard Jr., Catal. Commun. 9 (2008)
2244–2249.
[55] D. Xiao, J.R. Rajian, S. Li, R.A. Bartsch, E.L. Quitevis, J. Phys. Chem. B 110 (2006)
16174–16178.
[15] C. Aellig, I. Hermans, ChemSusChem 5 (2012) 1737–1742.
[16] M. Benoit, Y. Brissonnet, E. Guélou, K. De Oliveira Vigier, J. Barrault, F. Jérôme,
ChemSusChem 3 (2010) 1304–1309.
[56] F. D’Anna, S. Marullo, P. Vitale, R. Noto, ChemPhysChem 13 (2012) 1877.
[57] L. Cammarata, S.G. Kazarian, P.A. Salter, T. Welton, Phys. Chem. Chem. Phys. 3
(2001) 5192–5200.
[17] B. Caes, R. Raines, ChemSusChem 4 (2011) 353–356.
[18] H. Zhang, J. Wu, J. Zhang, J. He, Macromolecules 38 (2005) 8272–8277.
[19] J. Wu, J. Zhang, J. He, Q. Ren, M. Guo, Biomacromolecules 5 (2004) 266–268.
[20] R. Langer, Science 249 (1990) 1527–1533.
[21] T. Heinze, K. Schwikal, S. Barthel, Macromol. Biosci. 5 (2005) 520–525.
[22] F. Yang, Q. Liu, X. Bai, Y. Du, Bioresour. Technol. 102 (2011) 3424–3429.
[23] Q. Cao, X. Guo, J. Guan, X. Mu, D. Zhang, Appl. Catal., A 403 (2011) 98–103.
[24] L.-M. Clément, M. Claude, Catal. Commun. 4 (2003) 517–520.
[25] E. Khokhlova, V. Kachala, V. Ananikov, ChemSusChem 5 (2012) 783–789.
[26] Y. Li, H. Liu, C. Song, X. Gu, H. Li, W. Zhu, S. Yin, C. Han, Bioresour. Technol. 133
(2013) 347–353.
[58] P.S. Kulkarni, L.C. Branco, J.G. Crespo, M.C. Nunes, A. Raymundo, C.A.M. Afonso,
Chem. Eur. J. 13 (2007) 8478–8488.
[59] A. Stoppa, J. Hunger, R. Buchner, J. Chem. Eng. Data 54 (2008) 472–479.
[60] T.J. Mason, D. Peters, Practical Sonochemistry Power Ultrasound Uses and
Applications, Woodhead Publishing, 2002, pp. 51–55.
[61] C. Hardacre, S.P. Katdare, D. Milroy, P. Nancarrow, D.W. Rooney, J.M. Thompson,
J. Catal. 227 (2004) 44–52.
[62] P. Bonhôte, A.-P. Dias, N. Papageorgiou, K. Kalyanasundaram, M. Grätzel, Inorg.
Chem. 35 (1996) 1168–1178.
[63] P.J. Carvalho, T. Regueira, L.M.N.B.F. Santos, J. Fernandez, J.A.P. Coutinho, J.
Chem. Eng. Data 55 (2010) 645–652.
[27] S. Lima, M.M. Antunes, M. Pillinger, A.A. Valente, ChemCatChem 3 (2011)
1686–1706.
[28] S. Lima, P. Neves, M.M. Antunes, M. Pillinger, N. Ignatyev, A.A. Valente, Appl.
Catal. A-Gen. 363 (2009) 93–99.
[29] D. Liu, E.Y.X. Chen, Appl. Catal., A 435–436 (2012) 78–85.
[30] M.E. Zakrzewska, E. Bogel-Lukasik, R. Bogel-Lukasik, Chem. Rev. 111 (2011)
397–417.
[31] T. Ståhlberg, W. Fu, J.M. Woodley, A. Riisager, ChemSusChem 4 (2011) 451–458.
[32] X. Tong, Y. Ma, Y. Li, Carbohydr. Res. 345 (2010) 1698–1701.
[33] H. Xie, Z.K. Zhao, Q. Wang, ChemSusChem 5 (2012) 901–905.
[34] F. Guo, Z. Fang, T.-J. Zhou, Bioresour. Technol. 112 (2012) 313–318.
[35] A. Hernoux-Villière, J.-M. Lévêque, J. Kärkkäinen, N. Papaiconomou, M. Lajunen,
U. Lassi, Catal. Today 223 (2014) 11–17.
[64] O.O. Okoturo, T.J. VanderNoot, J. Electroanal. Chem. 568 (2004) 167–181.
[65] K.R.S. Kenneth, A. Stark, M.-J. Torres, Clean Solvents, vol. 819, American Chem-
ical Society, 2002, pp. 34–49.
[66] L. Crowhurst, P.R. Mawdsley, J.M. Perez-Arlandis, P.A. Salter, T. Welton, Phys.
Chem. Chem. Phys. 5 (2003) 2790–2794.
[67] C. Aelling, I. Hermans, Green Chem. 12 (2012) 1493–1513.
[68] K. Shimizu, R. Uozumi, A. Satsuma, Catal. Commun. 10 (2009) 1849–1853.
[69] X. Qi, M. Watanabe, T.M. Aida, J.R.L. Smith, Green Chem. 11 (2009) 1327–1331.
[70] S. Lima, P. Neves, M.M. Antunes, M. Pillinger, N. Ignatyev, A.A. Valente, Appl.
Catal. A 363 (2009) 93–99.
[71] S.J. Dee, A.T. Bell, ChemSusChem 4 (2011) 1166–1173.
[72] C. Sievers, I. Musin, T. Marzialetti, M.B. Valenzuela Olarte, P.K. Agrawal, C.W.
Jones, ChemSusChem 2 (2009) 665–671.
[36] C. Li, Z. Zhang, Z.K. Zhao, Tetrahedron Lett. 50 (2009) 5403–5405.
[37] B. Liu, Z. Zhang, Z.K. Zhao, Chem. Eng. J. 215–216 (2013) 517–521.
[38] A. Szabolcs, M. Molnar, G. Dibo, L.T. Mika, Green Chem. 15 (2013) 439–445.
[39] N. Kardos, J.-L. Luche, Carbohydr. Res. 332 (2001) 115–131.
[40] W. Lan, C.-F. Liu, F.-X. Yue, R.-C. Sun, J.F. Kennedy, Carbohydr. Polym. 86 (2011)
672–677.
[41] J.-P. Mikkola, A. Kirilin, J.-C. Tuuf, A. Pranovich, B. Holmbom, L.M. Kustov, D.Y.
Murzin, T. Salmi, Green Chem. 9 (2007) 1229–1237.
[42] N. Sun, M. Rahman, Y. Qin, M.L. Maxim, H. Rodriguez, R.D. Rogers, Green Chem.
11 (2009) 646–655.
[73] T. Stahlberg, M.G. Sorensen, A. Riisager, Green Chem. 12 (2010) 321–325.
[74] M.E. Zakrzewska, E. Bogel-Łukasik, R. Bogel-Łukasik, Chem. Rev. 111 (2010)
397–417.
[75] B. Girisuta, L.P.B.M. Janssen, H.J. Heeres, Green Chem. 8 (2006) 701–709.
[76] X. Hu, C.-Z. Li, Green Chem. 13 (2011) 1676–1679.
[77] G.A. Kraus, T. Guney, Green Chem. 14 (2012) 1593–1596.
[78] H.B. Xie, Z.B.K. Zhao, Q. Wang, ChemSusChem 5 (2012) 901–905.
[79] J.N. Chheda, Y. Roman-Leshkov, J.A. Dumesic, Green Chem. 9 (2007) 342–350.
[80] F. Ilgen, D. Ott, D. Kralisch, C. Reil, A. Palmberger, B. Konig, Green Chem. 11
(2009) 1948–1954.
[43] D.A. Fort, R.C. Remsing, R.P. Swatloski, P. Moyna, G. Moyna, R.D. Rogers, Green
Chem. 9 (2007) 63–69.
[81] S. Lima, P. Neves, M.M. Antunes, M. Pillinger, N. Ignatyev, A.A. Valente, Appl.
Catal., A 363 (2009) 93–99.
[44] Q. Liu, M.H.A. Janssen, F. van Rantwijk, R.A. Sheldon, Green Chem. 7 (2005)
39–42.
[45] R.C. Remsing, R.P. Swatloski, R.D. Rogers, G. Moyna, Chem. Commun. (2005)
1271–1273.
[82] X. Tong, M. Li, N. Yan, Y. Ma, P.J. Dyson, Y. Li, Catal. Today 175 (2011) 524–527.
[83] Z. Zhang, Q. Wang, H. Xie, W. Liu, Z. Zhao, ChemSusChem 4 (2011) 131–138.
[84] A. Takagaki, M. Ohara, S. Nishimura, K. Ebitani, Chem. Commun. (2009)
6276–6278.
[46] R.N. Das, K. Roy, Toxicol. Res. 1 (2012) 186–195.
[85] C. Moreau, A. Finiels, L. Vanoye, J. Mol. Catal. A: Chem. 253 (2006) 165–169.
[86] S.P. Borikar, T. Daniel, Ultrason. Sonochem. 18 (2011) 928–931.
[87] G. Chatel, C. Goux-Henry, N. Kardos, J. Suptil, B. Andrioletti, M. Draye, Ultrason.
Sonochem. 19 (2012) 390–394.
[47] A. Efimova, G. Hubrig, P. Schmidt, Thermochim. Acta 573 (2013) 162–169.
[48] S. Viboud, N. Papaiconomou, A. Cortesi, G. Chatel, M. Draye, D. Fontvieille, J.
Hazard. Mater. 215–216 (2012) 40–48.
[49] C.P. Fredlake, J.M. Crosthwaite, D.G. Hert, S.N.V.K. Aki, J.F. Brennecke, J. Chem.
Eng. Data 49 (2004) 954–964.
[50] X. Qi, M. Watanabe, T.M. Aida, R.L. Smith, ChemSusChem 2 (2009) 944–946.
[51] A. Stoppa, R. Buchner, G. Hefter, J. Mol. Liq. 153 (2010) 46–51.
[52] D. Xiao, J.R. Rajian, L.G. Hines, R.A. Bartsch, E.L. Quitevis, J. Phys. Chem. B 112
(2008) 13316–13325.
[88] E. Kowsari, M. Mallakmohammadi, Ultrason. Sonochem. 18 (2011) 447–454.
[89] S. Narayanaperumal, R.C. da Silva, K.S. Feu, A.F. de la Torre, A.G. Corrêa, M.W.
Paixão, Ultrason. Sonochem. 20 (2013) 793–798.
[90] M. Shekouhy, A. Hasaninejad, Ultrason. Sonochem. 19 (2012) 307–313.
[91] N. Yadav, M.K. Hussain, M.I. Ansari, P.K. Gupta, K. Hajela, RSC Adv. 3 (2013)
540–544.
[53] J.N. Canongia Lopes, T.C. Cordeiro, J.M.S.S. Esperanca, H.J.R. Guedes, S. Huq, L.P.N.
Rebelo, K.R. Seddon, J. Phys. Chem. B 109 (2005) 3519–3525.
[54] K. Shimizu, M. Tariq, L.P.N. Rebelo, J.N.C. Lopes, J. Mol. Liq. 153 (2010) 52–56.
[92] S. Zhao, X. Wang, L. Zhang, RSC Adv. 3 (2013) 11691–11696.
[93] F. D’Anna, S. Marullo, P. Vitale, R. Noto, Ultrason. Sonochem. 19 (2012) 136–
142.