Fructose Transformations in Ethanol using Carbon Supported Polyoxometalate Acidic Solids for 5-Ethoxymethylfurfural Production

Article abstract of DOI:10.1002/cctc.201800657

A series of carbon supported polyoxometalates have been prepared and studied as acid catalysts for the fructose dehydration. The catalytic supports, microporous activated carbon (AC, S_BET=1190 m²/g) and high surface area graphite (HSAG, S_BET=400 m²/g), were loaded with 15 wt% of polyoxometalates: phosphotungstic acid (TPA) or tungstosilicic acid (STA). The four resulting catalysts were tested in the fructose reaction at moderate temperature 140 °C, using water and ethanol solvents. Catalytic properties have been compared with those of an acidic resin, Amberlyst 15. As relevant findings the specific interactions of carbon supports and polyoxometalates let the inhibition of active phase lixiviation. An improved catalyst (STA-HSAG) in terms of selectivity to valuable products (ethoxymethylfurfural and ethyl levulinate) and high specific catalytic activity using ethanol as solvent has been developed. This catalyst can be reused after regeneration by washing with organic solvents.

Full text of DOI:10.1002/cctc.201800657

Accepted Article
Title: Fructose transformations in ethanol using carbon supported
polyoxometalate acidic solids for 5-Etoxymethylfurfural production
Authors: Nadia García Bosch, Belén Bachiller-Baeza, Inmaculada
Rodriguez-Ramos, and Antonio Guerrero-Ruiz
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemCatChem 10.1002/cctc.201800657
Link to VoR: http://dx.doi.org/10.1002/cctc.201800657
A Journal of
www.chemcatchem.org

10.1002/cctc.201800657
ChemCatChem
FULL PAPER
Fructose transformations in ethanol using carbon supported
polyoxometalate
acidic
solids
for
5-Etoxymethylfurfural
production
Nadia García-Bosch[a,b], Belén Bachiller-Baeza[b], Inmaculada Rodríguez-Ramos[b] and Antonio
Guerrero-Ruiz*[a]
Abstract: A series of carbon supported polyoxometalates have been
quality products and even serve as reactants themselves. Thus,
depending on the solvent employed, fructose dehydration
reaction goes through different reaction routes, as illustrated in
the Scheme 1.
prepared and studied as acid catalysts for the fructose dehydration.
The catalytic supports, microporous activated carbon (AC, SBET
=
1190 m2/g) and high surface area graphite (HSAG, SBET = 400 m2/g),
were loaded with 15wt% of polyoxometalates: phosphotungstic acid
(TPA) or tungstosilicic acid (STA). The four resulting catalysts were
tested in the fructose reaction at moderate temperature 140ºC, using
water and ethanol solvents. Catalytic properties have been
compared with those of an acidic resin, Amberlyst 15. As relevant
findings the specific interactions of carbon supports and
polyoxometalaltes let the inhibition of active phase lixiviation. An
improved catalyst (STA-HSAG) in terms of selectivity to valuable
products (ethoxymethylfurfural and ethyl levulinate) and high specific
catalytic activity using ethanol as solvent has been developed. This
catalyst can be reused after regeneration by washing with organic
solvents.
When the solvent is water, there is only one possible way:
fructose is dehydrated to HMF (1) and later a second hydration
occurred to obtain LA (2) and formic acid. On the other hand, if
ethanol (EtOH) is used as solvent, the reaction involves two
possible pathways. One of this consists firstly in the dehydration
of fructose to HMF (1) and subsequently HMF is ethylated to
EMF (4) or 5-(diethoxymethyl)2-furanmethanol (DEFM) (5). Both
of these could be ethylated to 5-ethoxymethylfurfural diethyl
acetal (EMFDA) (7,8) or hydrated to ELA (9,10). Finally
ethylated products (EMFDA) could be hydrated to produce ELA
(11) and ethyl formate. Notice that this scheme is rather similar
to the reported in the reference.[4] Following the second
possibility of fructose dehydration in ethanol medium, this is
transformed into E-fructose by ethylation (3), thus this
compound goes to its dehydration to obtain EMF (6). The next
steps starting from EMF to yield DEMF, ELA and EMFDA are
also displayed in the Scheme 1.
Introduction
Biomass derived compounds are possible sources for
substitution of petroleum based products because they have a
high content of carbons. In particular sugars (fructose and
glucose mixtures), which are easily obtained after a cascade of
operations from biomass, can be converted into platform
molecules through acidic catalysed reactions.[1-3] For instance 5-
hydroxymethylfurfural (HMF) and levulinic acid (LA) products are
interesting platform molecules for the bio-refinery industry. Thus
by dehydration of hexoses (i.e. glucose or fructose) can be
produced HMF and later LA. For these catalyzed processes new
high performance acid solids are required. Furthermore, the
reaction media is also important, water as a green solvent
should be desirable against organic liquid. Alternatively,
bioethanol (another platform compound or raw material) may be
used as reaction medium being possible to obtain other valuable
products, instead of HMF and LA, ethoxymethylfurfural (EMF)
and ethyl levulinate (ELA) will be formed.
Scheme 1. Some possible pathways when fructose reacts over acid solids in
Solvents used for biomass processes may lead to different
the presence of ethanol or water.
[a]
Miss N. García-Bosch and Prof. A.Guerrero-Ruiz
Department of Inorganic and Technical Chemistry
Institution UNED
Address Facultad de Ciencias, Paseo Senda del Rey 9, 28040
Madrid, Spain
E-mail: aguerrero@ccia.uned.es
Miss N. García-Bosch, Dr. B. Bachiller-Baeza and Prof. I.
Rodríguez-Ramos
The requirement of developing new solid acidic catalysts for the
bio-refinery can be rationalized considering that at present used
catalysts in the petroleum industry (typically zeolites) could not
be efficient enough with biomass compounds, which can be
related with the higher content of oxygen in this second group of
raw materials.[5] In addition, due to the relatively low thermal
stability of the biomass feedstocks, compared with compounds
resulting from petroleum industry, lower reaction temperatures
have to be applied for biomass processing and consequently
stronger acid sites are required.
[b]
Instituto de Catálisis y Petroleoquímica
Institution CSIC
Address Marie Curie 2, Cantoblanco, 28049 Madrid, Spain
Supporting information for this article is given via a link at the end of
the document.((Please delete this text if not appropriate))
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800657
ChemCatChem
FULL PAPER
Ion exchange resins and zeolites are typically heterogeneous
acidic catalysts used for dehydration reactions.[6] Taking into
account the acidity of the ion exchange resins, i.e. commercial
Amberlyst-15, these protonated ion exchange solids appear to
be a suitable alternative[4,7]. As reported in different publications,
using ion exchange resin higher catalytic performances than on
zeolites can be achieved.[8,9] Of course, solved inorganic acid
can also be used in these reactions but their environmental
negative impact implies less sustainability.
activated carbon (AC)) and compared with another acidic
material of different nature (i.e. ion exchange resin). For this
study Amberlyst-15 (Amb-15) was selected as reference catalyst
due to its extensive use for dehydration biomass reactions.
Results and Discussion
All the catalytic experiments were carried out in a batch reactor
and the kinetic data were obtained at a constant reaction
temperature of 140 ºC and a pressure of 30 bar of inert gas.
These conditions were selected after different previous
experiments without catalyst exploring fructose reactivity at
higher reaction temperatures. We found that the reactive
fructose is intensively decomposed in the empty reactor with
increasing temperature (thermal degradation). Under the
indicated reaction conditions blank experiments with the carbon
supports, AC and HSAG, were performed to verify their intrinsic
catalytic activity. Less than 10% of fructose conversion was
obtained, mainly due to thermal decomposition, because only a
yield lower than 1% toward interest products was observed after
6 h in reaction. Consequently the yield of desirable products
using carbon support materials can be considered negligible.
In a first set of experiments, we have compared two catalysts,
Amberlyst-15 (Amb-15) and H40SiW12O4·nH2O (STA) supported
over a high surface area graphite HSAG (STA-HSAG), in order
to evaluate the effects of the solvents, water and ethanol. The
results are presented in Figure 1. In these experiments fructose
conversions and product yields evolution with the reaction time
are compared for both catalysts (Amb-15 and STA-HSAG).
On the other hand, heteropolyacid solids or polyoxometalates
(POMs) are another type of eco-friendly acid solids, and
furthermore with stronger acidic sites than mineral acids.[10,11]
Many applications of POM as catalysts for hydrocarbon
transformation processes have been reported[12-15] and some
excellent reviews have been devoted to these acid solids.[16-19]
However, less extensive are the reports where POM are used as
catalysts for transformations of biomass derived compounds. In
this line Palkovits et al.[20] have reported conversions of 90% and
HMF’s selectivity between 40-50% in fructose dehydration
reaction, using both commercial or modified POMs. Also,
dehydration of fructose has been studied by Chen et al.[21] over
cesium and vanadium modified POMs for whom conversions up
to 99% and HMF yields near 70% were reported. Unfortunately
many of these experiments were carried out with DMSO
(dimethyl sulfoxide) as solvent, giving rise to less green
processes. Relevantly, Ueda et al.[22] have reported the
dehydration of biomass raw material (i.e. cellulose) in an
aqueous medium employing POM molecular wires based on
tungstotellurate and tungstoselenate oxides as catalysts. Under
the reaction conditions of this latter study high cellulose
conversions and high glucose yields were achieved. As for the
solvents used in these processes catalyzed by POMS, it should
be remarked that commercial POMs are in general water soluble.
Therefore, to these acid materials be employed as viable
heterogeneous catalysts, anchoring of these structures over
stable metal oxide solids (i.e. silica, alumina or metal oxides) or
supporting them over highly hydrophobic materials (i.e.
carbonaceous materials) would be necessary. So, for example,
two phosphotungstic acids supported over silica (HPW/SiO2 and
HSiW/SiO2) were used by Yang et al.[23] for fructose dehydration
achieving conversion values of 70% and HMF yields near 37%.
Recently we synthesized POM catalysts supported on activated
carbon and high surface area graphite which were applied as
solid acids for the vapour phase dehydration of bioethanol[24], as
well as bifunctional (by adding Ru nanoparticles) catalysts for
the ethylene glycol and propylene glycol production from
cellulose.[25]
Herein, we report new catalyst series based on two commercial
POMs for studying fructose dehydration reaction. Seeing that in
previous reports organic solvents were often used, our first
objective is to do this reaction with green solvents (i.e. water or
bioethanol). Secondly, as the possibility of lixiviation exits when
supported POMs are employed as heterogeneous acid catalysts,
the stability and recyclability of these new acidic materials was
tested. The two commercial POMs (H3PW12O40 (TPA) and
H4[W12SiO40] (STA)) were supported over two different
carbonaceous materials (high surface area graphite (HSAG) or
Figure 1. Comparison of the conversion of fructose and the yields toward
main products when the catalysts are Amb-15 (a,c) or STA-HSAG (b,d) and
solvents are water (a,b) or ethanol (c,d).
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800657
ChemCatChem
FULL PAPER
Two main aspects can be featured: fructose is significantly more
rapidly transformed in the ethanol media and the product
selectivity (see Figure 1) changes from HMF (obtained by
reaction 1 of the Scheme 1) in the case of water to EMF and
EMFDA with ethanol, so reactions numbered as 1 or 3, and 4 or
6 followed of 7 in Scheme 1 are taking place. In Figure 2 yields
to the main products and other products (formic, acetic and
lactic acids in the case of water and ethyl formate and E-fructose
with ethanol) are compared at the end of the reaction time (6 h).
Clearly the best yield to EMF is achieved with ethanol and using
STA-HSAG as catalyst. Interestedly this preliminary comparison
with the Amb-15 reveals that our catalysts based on POMs can
be competitive for fructose transformation.
ethanol (total volume in the batch reactor 90 ml). These results
suggest that the dispersion of the POM on the carbon support
may influence its catalytic behavior. In addition, as we will
discuss below these solid materials can be reused in an ethanol
medium. The reasons for the improved catalytic performance
and stability can be associated to the tight anchoring of POMs
over the carbon support surfaces. So Qi et al.[27] demonstrated
that heteropolyacids can be efficiently dispersed on
N
functionalized carbon nanotubes achieving active and stable
catalysts in water media. Very recently, following this
investigation line, some different interactions have been
proposed to explain the anchoring of POMs over carbon
nanotubes surfaces.[28] These authors suggest that electrostatic
interactions of POMs with the surface pyridinic N groups of the
carbon nanotubes are one of the reasons for their irreversible
immobilization on the carbon nanotubes. In our catalysts, the
carbon supports are not functionalized with basic species, but
the delocalized p-electrons present on the basal planes of
carbons are assumed to have basic nature.[29] On the other
hand, the anchoring of POMs would take place at the graphite
crystallite edges or with defective carbon surface sites will be
inhibited by the oxygen surface groups, with acid character, that
exit in such as defective sites. So an effective interaction of
POMs with the basal graphitic surfaces would lead to a good
dispersion of the POM on the support, i.e., low sizes of
crystallites.
In order to determine the sizes of POM crystallites in the fresh
samples we have applied two techniques, Scanning Electron
Microscopy (SEM) and X-ray diffraction (XRD). In Figure S2, the
diffractograms of the four catalysts prepared as well as the two
commercial POMs used in their preparation are displayed. It is
clear that diffraction lines stemming from POM phases in the
carbon supported POMs are broad and practically undetectable.
The absence of significant peaks of POMs in the diffraction
patterns of the catalysts, in spite of their high 15wt% loadings,
reveals that crystallites are rather small, surely in the nanometric
scale. These facts are confirmed from the SEM images of the
four POM supported catalysts, which are shown in Figure 3 and
Figure S3.
Figure 2. . Comparison of catalytic yields to main products (Scheme 1) when
water or ethanol are used as fructose solvents. The studied catalysts were:
Amb-15 and STA-HSAG.
From Figures 1 and 2 we have selected ethanol as solvent
media for the subsequent reaction studies. Also we are
interested in the comparison of Amb-15 with the two POMs (STA
and TPA) supported over activated carbon and HSAG. At this
point it is necessary to remark that we have selected this
commercial Amb-15 catalyst just due facility of acquisition, so
we have not considered an improved acidic resin.[26] Given the
high solubility of the POM in the two solvents used in reaction
(water and ethanol), some leaching of the supported acid POM
into the reaction solution under the applied reaction conditions
could be envisaged. In order to check the extent of this lixiviation
effect, two complementary set of experiments were performed.
First the liquid after finishing the reactions was analysed by
inductively coupled plasma mass spectrometry to detect the
presence of solved POMs. When water is the solvent the
amount of POM lixiviated to the solution was of 50% (Table S1)
while in ethanol was lower than 15 % of the initial loading in the
catalyst. The second set of experiments was done to verify the
leached POMs (working as homogeneous catalyst) give place to
conversion and selectivities rather lower than our heterogeneous
catalysts. For instance, with pure STA in the fructose
dehydration in ethanol, we obtain conversions of 20%, with
carbon balances near 90%, but with yields of 3% of DEFM and
7% of E-fructose. These data correspond to 3 h reaction in an
experiment where a concentration of STA 3x10-5 molar has been
used, with an initial 0.01 molar concentration of fructose in
Figure 3. Representative SEM images of samples: HSAG (a), STA-HSAG (b),
AC (C) and STA-AC (d).
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800657
ChemCatChem
FULL PAPER
For HSAG supported samples practically the POM crystallites
are not distinguished from the bare support. In samples
supported on AC some big POM particles can be observed (Fig.
3-d) on the external surface, but the amount of such particles
seems to be rather low, so possibly the majority of POM
particles are trapped in the micropores or small mesopores of
the AC support, which are inaccessible to the SEM technique.
These smaller crystallite sizes are indicative of specific
interactions between carbon supports and POMs, which can
justify the hindered solubility of these supported POM samples.
Additionally, the sample STA-AC was studied by SEM after
being used in reaction, some representative images are given in
the Fig. S4. Interestingly most of the crystallites of STA with
higher sizes have disappeared likely by dissolution in the
ethanol reaction medium. The smaller POM crystallites remain
stabilized by interaction with the carbon supports.
In the case of pure POMs two weight losses can be
distinguished: the first about 100ºC, where the loss of hydration
water molecules takes place, and the second between 160 ºC
and 260ºC, corresponding to the removal of hydroxyl groups, as
previously discussed.[24] These weight losses are endothermic
processes. In addition, for STA and TPA DSC exothermic
signals at temperatures above 550 ºC are evidenced, which are
not accompanied of weight losses. These may probably
correspond to POM decomposition into single metal oxides.[24] In
all supported catalysts this exothermic decomposition at higher
temperatures is not observed, probably because parasite
contributions of the carbon supports. Relevantly the weight loss
associated to dehydration reactions are very diminished in the
supported catalysts. In particular the peak due to
dehydroxylation (near to 200 ºC) is practically undetectable (see
Fig. S5 where DTG profiles are presented). This fact can be
indicative of existing specific interactions among carbon
supports and POMs, which stabilize the properties of supported
POMs versus the bulk crystalline material, in particular with
regard to the hydroxyl group stability. These interactions of
POMs with carbon supports can probably be related, as
discussed above, with basal graphitic surface sites. In fact,
though small these peaks of dehydroxylation can be appreciated
for TPA and STA supported on AC whereas they entirely
disappear when POMs are supported over HSAG. According
XRD and SEM results, it seems that the smaller POM
nanoparticles in interaction with graphitic surfaces (HSAG) are
more stabilized in their hydroxylated forms than bigger POM
crystallites supported over AC. Unfortunately the first weight loss
associated to the dehydration of POMs can not be differenced of
desorption of water from the supports (see also Fig. S6).
In order to gain information about these specific interactions a
series of thermogravimetric analysis under inert gas flow have
been performed. In Figure
4
are represented the
thermogravimetric analysis curves and the differential scanning
calorimetry (DSC) profiles for the bulk POMs (STA or TPA) and
the supported catalytic systems. Moreover in the supplementary
information, Figure S5, are displayed the differential thermal
analysis (DTG) of all the samples and in Figure S6 the TG, DTG
and DSC of the supports (HSAG or AC).
In order to compare all the synthesized catalytic materials, they
were tested in the same reaction conditions: ethanol as solvent,
140ºC of reaction temperature, a pressure of 30 bar of inert gas
and a weight ratio of 15 of active phase to 100 of fructose.
Figure 5 shows the fructose conversion values at different
reaction times. Clearly from this Figure 5 and from data collected
in Fig. S7 can be seen that POMs supported on HSAG and AC
behave differently. In the presence of the POMs supported on
HSAG, fructose reacts slower than over AC supported POMs
but the production of target products seems to be quicker with
HSAG supported catalysts. Probably the adsorption of fructose
is more rapid over the higher sized crystallites of POM located at
the external surfaces of AC. Also the production of byproducts,
such as humic acids, is favored over the microporous AC
support in comparison with HSAG, so the production of
desirable products is delayed in the case of STA-AC and TPA-
AC samples. This behavior about the formation of non-desirable
products will be discussed below. As before mentioned under
these reaction conditions a maximum of fructose conversion of
10% is achieved after 6 hours in reaction and the production of
interest compounds is not significant for the bare supports.
Figure 4. Thermogravimetric analysis (TG) black lines and scanning
differential calorimetry (DSC) red lines of supported POMs, carbon supports
and POMs.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800657
ChemCatChem
FULL PAPER
flow operation. It is well known that Amberlyst materials can
suffers swelling phenomena in liquid medium modifying their
accessible surface area as well as the amount of exposed acid
sites.[30]
In general, from Figure 5, it appears that POMs supported on
AC are more active for the fructose transformation than when
supported on HSAG. At least two aspects could contribute to
these slight differences. As previously demonstrated by XRD
and SEM, POM crystallites are smaller in size when supported
over HSAG, which could explain the maximum of IPD activity
(Table 1) detected for STA-HSAG. On the other hand, in the
fructose dehydration reaction we could assume that some
diffusional effects can be occurring. In fact small POM
crystallites in the catalysts supported over HSAG are practically
invisible in the corresponding SEM images, thus on these
samples the smaller crystallites could be placed inside the
nanometric hollows (inter graphite particle spaces) and suffer a
restricted accessibility of fructose in comparison with the POMs
supported over AC which presents crystallites with higher
dimensions observable by SEM.
Figure 5. Evolution of fructose conversion with the reaction time over different
catalytic materials.
By the extrapolation of the conversion rates in Fig. 5 to time zero
it is possible to confirm that all the POM derived catalysts have
comparable initial activities. However, the commercial Amb-15
sample has a diminished catalytic activity. In order to check the
acidic properties of the different catalysts an additional test, the
decomposition of isopropanol (IPD) reaction, was performed.
For all the catalysts, the only detected product in this test was
propylene, which is derived from the dehydration reaction
catalyzed by acidic surface centers. Therefore, from the
determination of specific activities for IPD, we have been able to
obtain a direct and quantitative measure of acid sites. These
results are reported in Table 1. It is noteworthy that during this
IPD gas phase reaction no deactivation phenomena were
observed, so both catalytic activity and propylene selectivity
remain constant.
Also relevant is the comparison of catalytic yields to significant
reaction products. In Figure 6 are reported as bar diagrams the
yield values to different interesting compounds obtained from
fructose using ethanol as solvent-reactant.
These values
correspond to under reaction. In the supplementary
3
h
information is represented (Figure S7) the evolution of reaction
products with the time for all studied catalytic materials.
Interestingly, the yields to EMF+EMFDA depend on the active
acid component. Clearly these yield values are higher over STA
derived catalysts in comparison with TPA ones. Amb-15 sample
can be placed in an intermediate position from this point of view
of EMF productivity. In particular with the catalyst STA-HSAG,
which exposes the maximum of acidic site density (Table 1), is
possible to obtain up to near 50% yield to EMF+EMFDA under
our soft experimental conditions.
Table 1. Specific catalytic activities determined for the different catalysts in the
gas phase isopropanol decomposition reaction.
Catalyst
Activity (µmol conv./g·s)
Amb-15
STA-AC
STA-HSAG
TPA-AC
1.5
1.7
3.0
2.0
1.8
TPA-HSAG
From this Table 1 is apparent that the density of acid surface
sites exposed on Amb-15 material is slightly lower than on the
POM supported samples, but this slight difference could explain
the lower initial catalytic activity of this commercial material for
the fructose dehydration reaction. However, in the case of Amb-
15 catalyst is particularly relevant to note that the fructose
transformation reaction and IPD test are performed in different
media. The former process is studied in a batch reactor with
ethanol as solvent, while the second is a gas phase continuous
Figure 6. Interest products yield (%), carbon balance and fructose conversion
of all catalysts using EtOH as solvent.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800657
ChemCatChem
FULL PAPER
A preliminary comparison with published data, for the same
reaction and at comparable experimental conditions, evidences
that our catalytic results agree in terms of catalytic activities and
selectivities.[31] So, using zeolites and unsupported Amberlyst
similar catalytic yields have been reported.[32] Thus, our best
(STA-HSAG) catalytic system, based in supported POM,
achieves similar yields but in our case the ratio of
substrate/active catalytic phase is noticeably higher.
from the catalytic surfaces, as previously suggested in the
literature.[33,34] Second and third consist of thermal treatment
under air flow at 300ºC (T3.Activ) and 400ºC (T4.Activ) for two
hours.[35] As can be observed in Fig. 7 the only efficient
regeneration treatment is the one which involves washing.
These findings support our supposition that probably
deactivation is due to humic acids as deactivating species.
Moreover, humic acid byproducts cannot be eliminated by
thermal desorption but can be removed by dissolution in organic
solvents. Furthermore, the regeneration of the spent STA-HSAG
catalyst means that STA does not lixiviated either under reaction
condition or during the washing treatment. On the other hand, it
should be noted that the commercial Amberlyst-15 material also
suffers from deactivation by humic acid deposition, which is
apparent from the change in its color turning to dark brown, and
cannot be regenerated applying the washing treatment.
Apparently HMF+DEFM selectivity values, these two
compounds being intermediate reaction products (see Scheme
1), are higher when using Amb-15 or supported TPA catalysts.
Contrarily the production of final products, EMF+EMFDA, is
favoured with the STA derived catalysts. Taking into account
that the total number of acid sites (Table 1) cannot directly be
related with these findings, we could propose that the particular
property of the supported STA smaller crystallites combines their
surface acid sites and the neighbor atomic arrangements
surrounding them. The cooperative action of both types of sites
leads to EMF+EMFDA production. But we have no further
evidences on the nanostructure of these special surface sites.
On the other hand, it is relevant to compare fructose
conversions and carbon balances for the different studied
catalysts, these parameters are depicted in Fig. 6. Apparently
catalysts supported over AC give place to poor carbon balances,
comparing STA-AC and TPA-AC with the corresponding STA-
HSAG and TPA-HSAG counterparts. For a tentative explanation
of this support effect, we may consider that activated carbon
support possesses a rather high contribution of micropores in its
porous structure, and that these micropores can act as recipient
(by irreversible adsorption) of humic acid byproducts, where they
can polymerize or block the acid sites of POMs. Also the carbon
balances, or the production of desirable products, seems to be
higher with the catalytic samples prepared from STA versus TPA.
As significant conclusions, we have been able to distinguish
among two POMs, showing that comparing STA versus TPA the
former produces less quantity of byproducts (humic acids) and it
is better as catalyst for the fructose conversion with regard to the
yield to final products. Furthermore, the comparison of two
carbonaceous supports evidenced that carbon balances (or
carbon deposits) are worse in the case of activated carbon. So
we have been able to develop an improved catalyst in terms of
yield to desirable products, for the fructose dehydration in
ethanol media: STA-HSAG sample. Finally, this new
synthesized catalytic material exhibits catalytic performance
superior to commercial Amb-15 under the reaction conditions
studied. Another important aspect to be considered for the
application of this catalytic material (STA-HSAG) is its reusability.
In Figure 7 the performances (fructose conversion, catalytic
activities, selectivities and carbon balances) of the fresh catalyst
and after regeneration under different treatments are shown. It
can be seen that when the filtered catalytic material from the
reaction media was directly reused the activity was considerably
depleted (named D.Reuse). Therefore, as the spent catalyst
becomes severely deactivated, we have treated aliquots of the
used catalyst with three different activation protocols. First, the
solid was washed sequentially with cyclohexane and
diethylether (sample W.Activ) in order to remove humic acids
Figure 7. Catalytic properties of the reused STA-HSAG catalyst after
regeneration of the spent material by applying different treatments.
Conclusions
Polyoxometalate acid catalysts supported on two carbon
materials, activated carbon and high surface area graphite, have
been studied in the fructose dehydration reaction. From the
analysis and discussion of the catalytic performances and from
the concurrent characterization of the synthesized catalytic
materials, some relevant conclusions can be stablished. Under
moderate reaction temperature and using ethanol as solvent,
instead of water, increased values of yield to desirable products
can be obtained. Interestingly the anchoring of POM to basal
planes of graphite crystallites seem to takes place, thus these
solids are quite stable again lixiviation. Also as a consequence
of these interactions the resulting POM crystallites are rather low
in dimensions (nanometric scale).
The systematic study of these composite materials as catalysts
for the fructose transformation reaction reveals that product
yields depend on the selected carbon support and on the POM
active component. So it has been concluded that between the
two studied POMs, STA and TPA, STA is more suitable for the
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800657
ChemCatChem
FULL PAPER
fructose conversion to desirable final products. Furthermore, the
comparison of two carbonaceous supports evidenced that
carbon balance are worse in the case of activated carbon due to
higher accumulation of carbon deposits during reaction on this
support.
the reactor by bubbling a flow of nitrogen (80 cm3 min-1) through
a saturator-condenser maintained at 16,5ºC, which allowed a
constant isopropanol flow of 3 cm3 min-1 or 130 mmol min-1. In a
typical experiment, an aliquot of sample of 70 mg (sized at 0.35-
0.5 mm) was pre-treated inside the reactor at 140ºC during 1 h
under continuous flow of N2. After cooling at 100ºC in inert flow,
the reaction was started at this temperature by feeding the 2-
propanol mixture (vaporized alcohol and N2 gas). The reaction
products were analyzed by an on-line gas chromatograph
provided with a thermal conductivity detector or FID at 150ºC
and a Chromosorb W-HP column at 50ºC for the separation of
products.
Summarizing, an improved catalyst (STA-HSAG) in terms of
selectivity to valuable products (ethoxymethylfurfural and ethyl
levulinate) and specific catalytic activities using ethanol as
solvent has been developed. Furthermore, the generation of
undesirable sub-products, such as humic acids, detected firstly
by the loss of reaction carbon balances, causes severe
deactivation of the catalytic materials. In fact regeneration of this
more selective and active catalyst, STA-HSAG sample, can be
achieved by washing treatments with organic solvents to remove
strongly adsorbed sub-products. Comparison of STA-HSAG
catalytic properties with those of an acidic resin, Amberlyst 15,
reveals that our improved STA-HSAG catalyst is superior in
activity and yields to desired products. In addition acid resin
cannot be regenerated, either by thermal treatment or by
washing with organic solvents.
In order to obtain information on morphological characteristics of
the acidic solids supported over carbonaceous materials, the
catalysts were subjected to
a detailed scanning electron
microscopy (SEM) study. Two equipments were employed, one
of them was FE-SEM Hitachi S-4700 and the other apparatus
was Hitachi TM-1000. Through energy dispersive X-ray
spectroscopy (EDX) studies in both equipments, the presence of
metals (Si, P and W), as well as the atomic ratio between these
metals (1:12
– Si:W or P:W) of supported POMs were
corroborated. Some of these results are presented in the
supplementary information, Figure S1. The supported POM
samples were studied without previous metallization treatment.
Experimental Section
Catalyst preparation
Two carbon materials were studied as catalyst support: a
commercial activated carbon (AC), produced from olive stones
by Oleicola el Tejar, Córdoba Spain, and a commercial high
surface area graphite HSAG-400 (HSAG) provided by
TIMCAL.[24] The activated carbon (AC) was sieved to 1.25–0.8
mm of grain sizes. This raw material was subjected to a
treatment with hydrochloric acid solution 10% (v/v) at 100 ºC for
24 h, in order to remove residual inorganic components, and
after filtering and washing with distilled water until complete
removal of detectable Cl- ions. The BET areas of the supports
were: 1190 m2 g-1 (AC) and 400 m2 g-1 (HSAG).[36]
Thermogravimetric Analysis (TG) studies were carried out on a
model SDTQ600 5200 TA apparatus. Each sample were heated
from room temperature to 700 °C, at heating rate of 10 °C min− 1
under helium flow (100 cm3/min). These measures involved the
weight loss (TG), derivative weight loss curves (DTG) and the
differential scanning calorimetry (DSC) analysis. In order to
detect the possible leaching of POM metals (i.e. tungsten and
silicon or phosphorus) during reaction, the final reaction mixture
was analyzed by inductively coupled plasma mass spectrometry
(ICP-OES Optima 3300 DV Perkin Elmer). Determining tungsten
concentration made possible to know the percentage (%) of
POM is dissolved into the reaction medium.
The supported catalysts, with a 15 wt% POM loading, were
prepared by incipient impregnation of the supports, with an
ethanol/water solution of the solid acids H3PW12O4·H2O (TPA)
and H40SiW12O4·nH2O (STA) as reported previously.[24] Finally,
commercial acid resin (Amberlyst-15) catalyst was obtained from
Alfa-Aesar.
Catalytic study of fructose dehydration
The dehydration of fructose was studied in a 100 ml Teflon-lined
stainless steel autoclave reactor (Parr 4072, Parr Instrument
Co.). Fructose, solvent and catalyst were mixed and loaded in
the reactor that was flushed with He to exclude air. Then, the
autoclave was heated to the reaction temperature while stirring
the mixture (500 rpm). The reaction time was set to zero when
the reaction temperature was reached, and data were recorded
thereafter. Typical catalytic tests were carried out at 140ºC and
30 bar of inert He, using 90 ml of solvent (ethanol or water), and
a mass catalyst loading of 15 wt% of acid sites based on
fructose, mass/mass. Comparisons of the catalytic results were
made on the basis of similar mass of the catalysts. Liquid
samples were taken periodically, filtered (0.2 µm-PTFE
membrane) and analyzed. Quantitative analyses of fructose, E-
fructose, HMF, DEFM and LA were performed by HPLC with an
Agilent system equipped with a Hi-plex H column, eluting with an
Catalyst characterization
X-ray diffraction (XRD) was use to study the structural properties
of the synthetized catalysts. The patterns were obtained on a
Polycrystal X'Pert Pro PANalytical instrument, with Ni-filtered
CuKα X-rays (λ = 1.54 Å). Bragg's angles between 5 and 95
were scanned at a rate of 0.058 degree per second.
The catalytic activity and selectivity in the decomposition of
isopropanol (IPD) was taken as model reaction test for acid sites.
The reaction was performed in a fixed-bed tubular glass reactor
working at atmospheric pressure. The 2-propanol was fed into
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800657
ChemCatChem
FULL PAPER
[11] a) J. Zhang, Y. Huang, G. Li, Y. Wei, Coordination Chemistry Reviews
in press. DOI: 10.1016/j.ccr.2017.10.025; b) Y. Chen, C. Zhang, J.
Zhang, Z. Ye, K. Zheng, Q. Fang, G. Li, Inorg. Chem. Front. 2017,4,
1917-1922.
aqueous solution of 0.005 M sulphuric acid (rate 0.6 mL/min) as
mobile phase and an index refraction detector. Standard
solutions covering the concentration range of the samples were
used to obtain the calibration curve for the compounds of
interest. The yield to EMF, EMFDA and ELA was determined by
gas chromatography with a Varian 3350 system equipped with a
capillary column SPB-5 and a FID detector using decane as
internal standard.
[12] M.L. Sarazen, E. Doskocil, E. Iglesia, J. Catal. 2016, 344, 553-569.
[13] T. Pinto, Ph. Arquillière, V. Dufaud, F. Lefebvre, Appl. Catal. A: Gen.
2016, 528, 44-51.
[14] K. Fan, J. Liu, X. Yang, L. Rong, RSC Adv. 2015, 5, 37916-37924.
[15] S. Wang, E. Iglesia, ACS Catal. 2016, 6, 7664-7684.
[16] F. Lefebvre, Current Catalysis 2017, 6, 77-89.
[17] G. Busca, Chem. Rev. 2007, 107, 5366-5410.
[18] Y.B. Huang, Y. Fu, Green Chem. 2013, 15, 1095-1111.
[19] C. Antonetti, D. Licursi, S. Fulignati, G. Valentini, A.M. Raspolli Galletti,
Catalysts 2016, 6, 196.
The specific results were expressed as:
ꢉꢁ0ꢂꢃꢄꢅꢆꢇꢈ − ꢉꢁꢂꢃꢄꢅꢆꢇꢈ
ꢀꢁꢂꢃꢄꢅꢆꢇꢈ(%) =
× 100
ꢉꢁ0ꢂꢃꢄꢅꢆꢇꢈ
[20] X. Yi, I. Delidovich, Z. Sun, S. Wang, X., Wang, R.A. Palkovits, Catal.
Sci. Technol. 2015, 5, 2496–2502.
[21] R. Liu, J. Chen, L. Chen, Y. Guo, J. Zhong, ChemPlusChem 2014, 79,
1448-1454.
[22] Z. Zhang, M. Sadakane, N. Hiyoshi, A. Yoshida, M. Hara, W. Ueda,
Angew. Chem. Int. Ed. 2016, 128, 10390-10394.
ꢉꢋ
ꢉꢁ0ꢂꢃꢄꢅꢆꢇꢈ
ꢊ(%) =
ꢋ
× 100
[23] G. Lv, L. Deng, B. Lu, J. Li, X. Hou, Y. Yang, Journal of Cleaner
Production 2017, 142, 2244-2251.
[24] M. Almohalla, I. Rodríguez-Ramos, A. Guerrero-Ruiz, Catal. Sci.
Technol. 2017, 7, 1892-1901.
[25] M. Almohalla, I. Rodríguez-Ramos, L.S. Ribeiro, J.J.M. Órfão, M.F,R.
Pereira, A.; Guerrero-Ruiz, Catal. Today 2018, 301, 65-71.
[26] F.H. Richter, K. Pupovac, R. Palkovits, F. Schüth, ACS Catal. 2013, 3,
123-127.
Acknowledgements
We acknowledge financial support from the Spanish
Government (CTQ2014-52956-C3-2-R and -3-R, and CTQ2017-
89443-C3-1-R and -3-R projects).
[27] W. Qi, W. Liu, S. Liu, B. Zhang, X. Gu, X. Guo, D. Su, ChemCatChem
2014, 6, 2613 – 2620.
[28] V.Y. Evtushok, A. N. Suboch, O.Y. Podyacheva, O.A. Stonkus, V.I.
Zaikovskii, Y.A. Chesalov, L.S. Kibis, O.A.; Kholdeeva, ACS Catal.
2018, 8, 1297-1307.
Keywords: fructose transformation • heterogeneous catalytic
reaction • polyoxometalates • ethoxymethylfurfural • ethyl
levulinate
[29] M.A. Montes-Moran, D. Suarez, J.A. Menendez, E. Fuente, Carbon
2004, 42, 1219–1225.
[30] R. Bringué, E. Ramírez, M. Iborra, J. Tejero, F. Cunill, J. Catal. 2013,
304, 7-21.
[1]
[2]
[3]
[4]
P. Gallezot, Chem. Soc. Rev. 2012, 41, 1538-1558.
J.C. Serrano-Ruiz, J.A. Dumesic, Energy Environ. Sci. 2011, 4, 83-99.
A. Corma, S. Iborra, A. Velty, Chem. Rev.2007, 107, 2411-2502.
D. Song, Y. Sun, Q. Zhang, P. Zhang, Y. Guo, J. Leng, Appl. Catal. A:
Gen. 2017, 546, 36-46.
[31] X. Xiang, J. Cui, G. Ding, H. Zheng, Y. Zhu, Y. Li, ACS Sustainable
Chem. Eng. 2016, 4, 4506−4510.
[32] H. Li, S. Saravanamurugan, S. Yang, A; Riisager, Green Chem. 2016,
18, 726-734.
[33] P. Che, F. Lu, J. Zhang, Y. Huang, X. Nie, J. Gao, J. Xu, Bioresource
Technology 2012, 119, 433-436.
[5]
H. Li, S. Saravanamurugan, S. Yang, A. Riisager, Green Chem. 2016,
18, 726-734..
[34] S. Wang, H. Lin, J. Chen, Y. Zhao, B. Ru, K. Qiu, J. Zhou, RSC Adv.
2015, 5, 84014-84021.
[6]
[7]
Y.B. Huang, Y. Fu, Green Chem. 2013, 15, 1095-1111.
E. Ahmad, M.I. Alam, K.K. Pant, A. Haider, Green Chem. 2016, 18,
4804–4823.
[35]
C.H. Kuo, A.S. Poyraz, L. Jin, Y. Meng, L. Pahalagedara, S.Y. Chen,
D.A. Kriz, C. Guild, A. Gudza, S.L. Suib, Green Chem. 2014, 16, 785-
791.
[8]
[9]
J. Cui, J. Tan, T. Deng, X. Cui, Y. Zhu, Y. Li, Green Chem. 2016, 18,
1619–1624.
[36] A. Guerrero-Ruiz, E. Gallegos-Suarez, L. Gonzalo-Chacón, I.
D.R. Fernandes, A.S. Rocha, E.F. Mai, C.J.A. Mota, V. Teixeira da
Silva, Appl. Catal. A: General 2012, 425– 426, 199 – 204.
Rodríguez-Ramos, Thermochimica Acta 2013, 567, 112-117.
[10] W. Deng, Q. Zhang, Y. Wang, Dalton Trans. 2012, 41, 9817-9831.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800657
ChemCatChem
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 1:
FULL PAPER
A new acid catalyst is developed by
anchoring a polymetalate to basal
surfaces of graphite, with
Nadia García-Bosch, Belén Bachiller-
Baeza, Inmaculada Rodríguez-Ramos
and Antonio Guerrero-Ruiz*
applicability for the transformation in
ethanol solvent of fructose into
ethoxymethylfurfural and ethyl
levulinate.
Page No. – Page No.
Fructose transformations using
carbon supported polyoxometalate
acidic solids as heterogeneous
catalysts
This article is protected by copyright. All rights reserved.