Synergistic effect between CaCl2 and Γ-Al2O3 for furfural production by dehydration of hemicellulosic carbohydrates

Article abstract of DOI:10.1016/j.apcata.2019.117188

γ-Al₂O₃ has successfully been used as solid acid catalyst for xylose dehydration in a water:toluene biphasic system. The addition of alkaline earth metal chlorides improved the partition coefficient, thus facilitating the extraction of furfural from the aqueous phase. In this sense, the best catalytic performance was attained by using CaCl₂ and γ-Al₂O₃ with almost full xylose conversion and a furfural yield of 55% after 50 min, at 150 °C. A ¹H NMR study has demonstrated that CaCl₂ shifts the α-xylopyranose/β-xylopyranose anomeric equilibrium toward the α-form, more favorable for initiating the dehydration process, and γ-Al₂O₃ promotes the furfural production. γ-Al₂O₃ can be reused during ten catalytic cycles without any pretreatment. Finally, the synergistic effect between CaCl₂ and γ-Al₂O₃ could also be inferred from the increase in furfural yield, until a value of 83% after 50 min at 175 °C, by using a hemicellulosic liquor obtained from olive stones.

Full text of DOI:10.1016/j.apcata.2019.117188

Accepted Manuscript
Title: Synergistic effect between CaCl2 and ␥-Al2O3 for
furfural production by dehydration of hemicellulosic
carbohydrates
´
´ ˜
´
Authors: I. Funez-Nunez, C. Garcıa-Sancho, J.A. Cecilia, R.
´
Moreno-Tost, E. Perez-Inestrosa, L. Serrano-Cantador, P.
Maireles-Torres
PII:
DOI:
Article Number:
S0926-860X(19)30343-6
https://doi.org/10.1016/j.apcata.2019.117188
117188
Reference:
APCATA 117188
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
6 May 2019
31 July 2019
1 August 2019
´
´ ˜
´
Please cite this article as: Funez-Nunez I, Garcıa-Sancho C, Cecilia JA,
´
Moreno-Tost R, Perez-Inestrosa E, Serrano-Cantador L, Maireles-Torres P,
Synergistic effect between CaCl2 and ␥-Al2O3 for furfural production by
dehydration of hemicellulosic carbohydrates, Applied Catalysis A, General (2019),
https://doi.org/10.1016/j.apcata.2019.117188
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Synergistic effect between CaCl2 and γ-Al2O3 for furfural production by
dehydration of hemicellulosic carbohydrates
I. Fúnez-Núñeza, C. García-Sanchoa*, J.A. Ceciliaa, R. Moreno-Tosta, E. Pérez-
Inestrosab,c, L. Serrano-Cantadord, P. Maireles-Torresa
a Universidad de Málaga, Departamento de Química Inorgánica, Cristalografía y
Mineralogía (Unidad Asociada al ICP-CSIC), Facultad de Ciencias, Campus de
Teatinos, 29071 Málaga, Spain
b Andalusian Center for Nanomedicine and Biotechnology - BIONAND, Málaga,
Spain
c Department of Organic Chemistry, University of Málaga, IBIMA, Málaga, Spain
d Universidad de Córdoba. Departamento de Química Inorgánica e Ingeniería Química.
Campus Universitario de Rabanales, Edificio Marie Curie (C3). Ctra. (a) de Madrid,
km 396, 14071-Córdoba.
*Corresponding author: cristinags@uma.es
1

Graphical abstract
Highlights
 CaCl2 increases the -xylopyranose/-xylopyranose ratio favoring the xylose
dehydration with respect to other salts and -Al2O3 promotes the furfural
production due to their Lewis acid sites.
 The synergistic effect between CaCl2 and γ-Al2O3 gives full conversion and
furfural yield of 55% after 50 min at 150ºC and -Al2O3 can be reused at least
for 10 catalytic runs without regeneration.
 CaCl2 and -Al2O3 provided a furfural yield of 83% by liquors obtained from
olive stones demonstrating their applicability for furfural production from
lignocellulosic biomass.
Abstract
2

-Al2O3 has successfully been used as solid acid catalyst for xylose dehydration
in a water:toluene biphasic system. The addition of alkaline earth metal chlorides
improved the partition coefficient, thus facilitating the extraction of furfural from the
aqueous phase. In this sense, the best catalytic performance was attained by using CaCl2
and -Al2O3 with almost full xylose conversion and a furfural yield of 55% after 50 min,
1
at 150ºC. A H NMR study has demonstrated that CaCl2 shifts the -xylopyranose/-
xylopyranose anomeric equilibrium toward the -form, more favorable for initiating the
dehydration process, and -Al2O3 promotes the furfural production. -Al2O3 can be
reused during ten catalytic cycles without any pretreatment. Finally, the synergistic
effect between CaCl2 and -Al2O3 could also be inferred from the increase in furfural
yield, until a value of 83% after 50 min at 175ºC, by using a hemicellulosic liquor
obtained from olive stones.
Keywords: lignocellulose; alumina; calcium chloride; furfural; xylose; olive stones
1.
Introduction
The development of environmentally and economically viable synthetic routes
and technologies in order to obtain chemicals and fuels by using non-fossil carbon
sources has been reenergized in recent years in order to replace fossil fuels [1]. In this
context, biomass is a very promising sustainable feedstock due to its availability and it
can be employed as renewable carbon source [2-4]. In particular, lignocellulosic
biomass is the most abundant (2·1011 metric tons per year) and it is mainly composed by
three compounds (lignin, cellulose and hemicellulose), being necessary its full
fractioning and valorization for the profitability of its use as a raw material [5].
However, lignocellulose will be a sustainable resource for production of biofuels and
3

chemicals as long as the biomass does not interfere with the food chain. In particular,
hemicellulose is made up of amorphous heteropolysaccharides, containing five and six
carbon units, which can be substituted with phenolic, uronic or acetyl groups [6]. The
pentoses, xylose and arabinose, are the main and more abundant monomers, although
hexoses such as glucose, mannose and galactose are found in minor proportions. From
biomass waste, hemicellulosic carbohydrates can be recovered by different
physicochemical processes for the production of chemicals, such as xylitol, lactic acid
or furfural [7].
The acid hydrolysis of hemicellulose leads to pentoses, mainly xylose, whose
dehydration gives rise to furfural. Furfural is a key building block produced in
lignocellulose-based biorefineries, which could further be transformed to fuels and
useful chemicals, being widely used in oil refining, plastics, pharmaceutical and
agrochemical industries [8]. Considering that xylose is the most abundant carbohydrate
present in hemicellulose, great efforts have been carried out to study its dehydration into
furfural. One of the main problems in the production of furfural is related to the
existence of other unwanted side reactions, which may occur at the same time as
dehydration, decreasing the furfural yield. Thus, fragmentation reactions of furfural can
take place, leading to the formation of formaldehyde, formic acid, acetaldehyde,
dihydroxyacetone, glyceraldehyde, glycolaldehyde, lactic acid, acetol and
pyruvaldehyde [9, 10]. Likewise, condensation reactions between furfural and the
reaction intermediates from xylose have usually been observed, although in the absence
of xylose these reactions do not take place [11]. Besides, furfural reactions with itself
can give rise to resinification products [10]. On the other hand, it is well known that the
dehydration of xylose requires the presence of acid sites to yield furfural. Weingarten et
4

al. reported that both Brönsted and Lewis acid sites catalyze xylose dehydration into
furfural, being the latter more active but less selective since condensation reactions take
place preferentially on Lewis acid sites [12].
Nevertheless, both solvent and catalyst are considered as two key factors to
attain high furfural yields from xylose dehydration. The use of biphasic systems for
furfural production has gained attention, because this approach can allow achieving
higher furfural yields than those systems employing only water [13-15]. A biphasic
medium, formed by the addition of immiscible organic solvents, such as toluene [16],
methyl isobutyl ketone [17] and alkylphenols [15] to a diluted acid solution, or the
addition of miscible organic solvents such as -valerolactone (GVL) [18] or
tetrahydrofuran (THF) [19], to a saturated salt solution, allows to extract the target
product, i.e. furfural, from the aqueous phase, preventing its further degradation and
condensation.
Likewise, a large variety of solid acid catalysts as alternative to conventional
mineral acids has been studied for the dehydration of xylose to furfural: zeolites [20],
heteropolyacids [21], ion exchange resins [22] and metal oxides and phosphates [23,
24], either pure or supported to improve their textural properties. Several reviews have
been published dealing with the dehydration of xylose to furfural [10, 25-27]. Thus,
aluminum oxides have been traditionally employed as supports and catalysts due to
their abundance, price and availability. The catalytic behavior of alumina, as such or in
mixed oxides, has been evaluated for dehydration of different carbohydrates, such as
glucose, fructose or xylose [28-33]. In particular, in a previous study, it was
demonstrated that the use of γ-Al2O3 as support for Nb2O5 based catalysts enhanced
5

xylose conversion in its dehydration into furfural, although it promoted side reactions
and considerably decreased the furfural selectivity [28].
In order to prevent these side reactions, the use of inorganic salts has been
proposed in the literature. Marcotullio et al. [34, 35] found that halide anions promoted
the formation of 1,2-enediol as acyclic intermediate for xylose dehydration, which was
subsequently dehydrated to furfural. Enslow et al. [36] demonstrated that xylose is
quickly transformed in the presence of aqueous solutions of inorganic salts, in
comparison with pure water, and its rate depended on the nature of both cation and
anion. However, other conclusions about the effect of inorganic salts on the dehydration
of sugars have been reached by using other carbohydrates. Thus, Rasrendra et al. [37]
determined, from the study of different metal halide and sulfate salts, that the influence
of anion was less relevant than the effect of cations. Román-Leshkov et al. [38]
employed inorganic salts in a biphasic system in order to obtain HMF from fructose;
they concluded that the catalytic improvement attained in the presence of salts could be
attributed to the interaction of both cationic and anionic species. On the other hand, it
has been reported that divalent cations interact more strongly with saccharides than the
monovalent ones [39, 40]. Thus, Combs et al. [41] observed that alkaline earth metal
cations can interact with glucose, forming bidentate complexes, which accelerated its
transformation. Later, our research group studied the beneficial effects of CaCl2 on
glucose dehydration to HMF in the presence of Al2O3 as catalyst, in such a way that the
addition of CaCl2 notably enhanced the HMF production, even at very short reaction
times, since the presence of Ca2+ promoted the α-D-glucopyranose formation [42].
Therefore, it would be possible that alkaline earth cations exert a similar effect on the
dehydration of xylose.
6

In this work, the dehydration of xylose to furfural has been evaluated by using a
mesoporous γ-Al2O3 with acid character, in a biphasic water-toluene solvent system to
minimize side reactions involving furfural. The addition of inorganic salts, such as
CaCl2 and MgCl2, to the reaction medium has been carried out in order to maximize the
furfural yield. Likewise, the role of these alkaline earth metal salts into the mechanistic
1
pathways has been elucidated for the first time by H NMR. Different experimental
parameters, such as the reaction time, the temperature, the amount of catalyst and its
possible reuse and the influence of the amount of inorganic salts to the reaction
medium, have also been evaluated. In addition, the catalytic performance of γ-Al2O3
and/or CaCl2 has been studied with liquors obtained from the treatment of olive stones,
as lignocellulosic biomass.
2.
Experimental
2.1. Reagents
The following chemicals have been used for the catalytic tests: alumina (Alfa-
Aesar, Brockmann Grade I, 58 Å), xylose (Sigma-Aldrich, >99%), deionized water and
toluene (VWR, 98%) as solvents and CaCl2·2H2O (VWR, 97%) and MgCl2·6H2O
(VWR, 99%) as alkaline earth salts.
2.2. Characterization of catalysts
The textural parameters were determined from the N2 adsorption-desorption
isotherms at -196 ºC by using a gas adsorption analyser (ASAP, Micromeritics). The
sample was previously degassed at 200 ºC and 10-4 mbar for 10 h. The Brunauer-
Emmet-Teller (BET) equation was utilized to determine surface area taking a N2 cross
7

section of 16.2 Å2 and the BJH method was employed to calculate the pore size
distribution.
The total surface acidity of catalyst was evaluated by the temperature-
programmed desorption of ammonia (NH3-TPD). Firstly, the cleaning of catalyst (80
mg) was carried out under a helium flow at 550 ºC. Later, the adsorption of ammonia
took place at 100 ºC and the NH3-TPD was realized by increasing the temperature from
100 to 550 ºC (10 ºC min-1) and maintaining at 550 ºC for 15 min under a helium flow
of 40 mL min-1. The evolved ammonia was determined by using a Thermal
Conductivity Detector.
FTIR spectra of adsorbed pyridine were recorded on a Shimadzu Fourier
Transform Infrared Instrument (FTIR8300). Self-supported wafer of the sample with a
weight/surface radio of about 15 mg cm−2 was placed in a vacuum cell greaseless
stopcocks and CaF2 windows. The sample was evacuated at 300 ºC and 10−2 Pa
overnight, exposed to pyridine vapors at room temperature (vapor pressure of 200 mbar)
for 10 min and then outgassed at room temperature, 100 ºC and 200 ºC.
Thermogravimetric analyses (TGA) were performed under air flow of 50 mL
min-1 by a TGA/DSC 1 model (Mettler-Toledo) from 30 ºC until 900 ºC with a heating
ramp of 10 ºC min-1. The carbon content of the spent catalyst was determined with a
LECO CHNS 932 analyzer.
The leaching of Al species was carried out on an ICP-MS Nexión 300D by using
a software NexION (Perkin Elmer) to calculate the concentration.
2.3. Catalytic tests
8

The furfural production from the catalytic dehydration of xylose was carried out
under batch conditions with magnetic stirring, by using a glass pressure tube with thread
bushing (Ace, 15 mL). In a typical experiment, 0.05 g of catalyst, 0.15 g of xylose, 1.5
mL of deionized water and 3.5 mL of toluene as solvent and co-solvent respectively
were poured into the reactor. The reaction temperature was controlled with a
thermocouple in an oil bath. In order to avoid the secondary reactions, reactors were
always purged with N2 prior the experiments. The reaction was quenched by
submerging the reactor in a cool water bath. Both organic and aqueous phases were
separated, filtered and the products were analyzed by high performance liquid
chromatography (HPLC), being xylose and furfural the only detected products. A
JASCO instrument equipped with two detectors, multiwavelength detector (MD-2015)
and refraction index (RI-2031), a quaternary gradient pump (PU-2089), autosampler
(AS-2055), column oven (co-2065) and two columns, Phenomenex Luna C18 reversed-
phase column (250 mm × 4.6 mm,5 µm) and Phenomenex Rezex ROA-Organic Acid
H+ (8%) (300 mm × 7.8 mm, 5 µm), was employed. Xylose was monitored using a
refractive index detector in the aqueous phase, while furfural production was monitored
using multiwavelength detector in both phases, water and toluene (278 and 285 nm
respectively). The mobile phases consisted in pure methanol (flow rate 0.5 mL·min−1)
for Luna C18 and 0.005M H2SO4 aqueous solution (flow rate 0.35 mL·min−1) for Rezex
ROA-Organic Acid H+, being the columns at room temperature and 40 ºC, respectively.
Moreover, the catalyst was reused for several catalytic runs, at 150 ºC and 50
minutes, in the presence of 0.65 g.CaCl₂·g.aq.sol.-1. For this, the catalyst was kept inside the
reactor after each cycle, and, without any pre-treatment, a new reaction mixture (xylose,
CaCl2, water and toluene) was added to carry out the reaction under identical
9

experimental conditions. On the other hand, the catalyst was washed by vacuum
filtration, after the first reaction cycle, with a controlled amount of 10 mL of both water
and toluene for its characterization, and these solvents were analyzed after washing by
HPLC for analyzing both phases.
Finally, liquors from lignocellulosic biomass were prepared from olive stones by
using a solid:water mass ratio of 1:20, at 180 ºC for 30 min, in a temperature-controlled
stainless-steel reactor (2 L) under batch conditions. These extraction conditions were
based on results previously reported by Amendola et al. [43]. For the catalytic study, 1.5
mL of this liquor was taken as aqueous phase, and the same experimental conditions
previously described for the tests with xylose were employed.
3. Results and discussion
It is well known that both the textural properties and the acidity of catalysts are
key parameters to maximize the furfural yield [44]. Regarding the textural properties of
γ-Al2O3 obtained from nitrogen adsorption-desorption data, a Type IV isotherm which
trends to a Type II isotherm according to the IUPAC classification was found which
confirmed its mesoporous character (Fig. 1A). In addition, the catalyst shows a high
BET surface area (158 m2·g-1) and large average pore diameter (4.7 nm) (Table 1)
included in the range of mesopores, giving rise to a narrow pore size distribution (Fig.
1B).
On other hand the existence of acid sites with different strength, and a high total
acidity (478 μmoles NH3·gcat-1) were demonstrated by the NH3-desorption profile (Fig. 2
and Table 1). A maximum desorption at 180ºC demonstrates the presence of weak acid
sites, whereas the desorption at high temperatures is associated to strong acid sites [45].
10

The high values of BET surface area and total acidity make this catalyst displays higher
amount of available acid sites since it presents an acid density equal to 3.02 μmoles
NH3·m-2.
Likewise, it is relevant to know the nature of these acid centers which have been
elucidated by adsorption of pyridine coupled to FTIR spectroscopy (Fig. 3). Thus, the
absorption bands appearing at 1613 and 1449 cm-1 correspond to the 8a and 19b
vibration modes of pyridine coordinated to strong Lewis acid sites (s-LAS), respectively
[46]. In addition, it can be observed that although the outgassing temperature increases,
the characteristic signal of pyridine adsorbed on these Lewis acid centers remains,
although less intense, which corroborates the strength of Lewis acid sites. The band
located at 1578 cm-1 is attributed to pyridine coordinated on weak Lewis acid sites (w-
LAS), as was previously pointed out by Gallo et al. [46]. The intensity of this latter
band decreases with the outgassing temperature, confirming the weakness of these acid
sites. The existence of strong and weak Lewis acid sites could be inferred from its NH3-
TPD curve. The absence of peaks around 1550 cm-1 demonstrates that γ-Al2O3 does not
possess Brönsted acid sites (BAS). Another band is also detected at 1491 cm−1 which is
usually associated to the 19a vibration mode of pyridine associated to both Brönsted and
Lewis acid sites [47], being in this case only due to LAS. Finally, a band centered at
1594 cm−1 is also detected when pyridine is evacuated at room temperature, which can
be assigned to the 8a vibration mode of pyridine bonded by hydrogen bond to the
hydroxyl groups of the catalyst surface [48]. However, this band almost disappears after
thermal treatment at 100 ºC, evidencing the weak interaction of pyridine with surface
hydroxyl groups.
11

The γ-Al2O3 catalyst was employed for dehydration of xylose into furfural. It has
previously been demonstrated that Lewis acid sites are more active than Brönsted ones,
but less selective, since Lewis acid sites can promote condensation reactions [12]. -
Al2O3 possesses weak and strong Lewis acid sites, as inferred from NH3-TPD and
pyridine adsorption coupled to FTIR spectroscopy. However, Lewis acid sites can
promote secondary reactions, leading to a low furfural selectivity, in spite of a high
catalytic activity in xylose dehydration [28]. Therefore, it is necessary to optimize
experimental conditions in order to maximize the furfural yield. In this sense, the use of
biphasic systems can facilitate the continuous extraction of furfural formed from the
aqueous phase, thus reducing its residence time in the aqueous phase for preventing the
reaction with xylose and reaction intermediates [14, 49, 50]. In this work, a biphasic
water-toluene system has been employed to obtain furfural from xylose dehydration by
uing -Al2O3 as solid acid catalyst and compared with the non-catalyzed process at 175
°C and 50 min of reaction (Fig. 4A). It was observed that the addition of -Al2O3 to the
reaction medium considerably enhances the conversion of xylose and the furfural yield
(86% and 26%, respectively) with respect to non-catalyzed process (15% and 14%,
respectively). However, furfural yield is not very high in spite of the high catalytic
activity of -Al2O3.
Considering that the Lewis acid sites of γ-Al2O3 could favor the side reactions that
decrease the furfural selectivity, the optimization of other reaction parameters has been
performed to inhibit these side reactions, such as the addition of inorganic salts. In
recent years, different authors have demonstrated the positive effect of adding inorganic
salts for the production of HMF and furfural [35-42, 51, 52]. In this context, HMF yield
obtained from dehydration of glucose was successfully improved by using calcium
12

chloride, achieving a maximum yield in presence of 0.65 gCaCl ·gaq.sol.-1, since the Ca2+
₂
ions interacted with glucose molecules favoring the α-D-glucopyranose formation, the
most favorable anomer for enhancing HMF yield [42]. Thus, in the present work, a
similar concentration of CaCl2 has been utilized for dehydration of xylose to favor
furfural production, as well as another alkaline earth metal salt, MgCl2, under equimolar
concentration (0.39 g·gaq.sol.-1), in order to evaluate the effect of the cation nature (Fig.
4A). Although both salts improved the catalytic performance, the addition of CaCl2
demonstrated to be more beneficial than MgCl2, attaining a xylose conversion of 96%
and a furfural yield of 71%. In addition, it can be observed that the furfural production
is higher using only CaCl2 (71%) in the reaction medium than together with alumina
(55%). It should be noted that the use of toluene as co-solvent provided higher values of
furfural yield with respect to those attained for catalytic test by using only water as
solvent (Fig. 4A). In order to get insights into the effect of these salts, the partition
coefficients, R, defined as the furfural concentration in the organic phase with respect to
the furfural concentration in the aqueous phase, have been determined. Thus, a positive
effect has been detected increasing the R values in the presence of alumina and/or
alkaline earth salts with respect to the catalytic test carried out in the absence of salt (R
= 0.52) (Fig. 4B). Moreover, a direct relationship between the furfural yield and the
partition coefficient can be observed under these experimental conditions, achieving the
highest R value (4.4) in the presence of CaCl2, which provides the highest furfural yield
(71%). In addition, the partition coefficient is lower for MgCl2 (R = 2.8), although
higher than in the case of using Al2O3. It should also be noted that both furfural yield
and the partition coefficient are lower when a mixture of γ-Al2O3 and CaCl2 was used,
which might be due to alumina favoring side reactions, as previously commented.
13

Takin into account that γ-Al2O3 favors secondary reactions, the influence of the
reaction time has also been performed by using γ-Al2O3 without and with CaCl2 in the
reaction medium, and compared with the non-catalyzed process (in the absence of
catalyst), without and with salt (Fig. 5). In all cases, higher conversion and furfural
yield values were attained at 175ºC in the presence of γ-Al2O3 and/or calcium chloride
with respect to the non-catalyzed process. The use of this acid catalyst in the presence of
CaCl2 exhibited a much higher catalytic activity, even at short reaction times, attaining a
xylose conversion and furfural yield of 97% and 55%, respectively, after 50 min of
reaction (Fig. 5A). It can be observed that the furfural yield was lower than expected,
achieving better results using only calcium chloride without a solid acid catalyst, with a
conversion of xylose of 96% and a furfural yield of 71%, after 50 min at 175 ºC. This
fact would demonstrate that γ-Al2O3 promoted the side reactions of furfural at high
temperature, consequently decreasing the furfural yield (Fig. 5B) [28]. However, it
should be noted that both xylose conversion and furfural yield after 15 min of reaction
are higher by using a mixture of Al2O3 and CaCl2 (66% and 44%, respectively) than
those obtained by employing only CaCl2, thus confirming that the furfural yield
decreases along the reaction due to the side reactions. Thus, in order to minimize
secondary reactions and enhancing the furfural selectivity, the rest of the catalytic study
will be performed at a lower reaction temperature and shorter reaction time (150 ºC and
30 min).
Firstly, the concentration of calcium chloride has been optimized in the absence
of γ-Al2O3 (Fig. 6). According to the results previously obtained in glucose dehydration
to HMF [42], the higher CaCl2 concentration is employed, the higher value of both
xylose conversion and furfural yield are achieved, achieving the maximum values in the
14

presence of 0.65 g CaCl2 per g aqueous solution (48% and 21%, respectively), after 30
min at 150 ºC. However, the catalytic activity does not improve when the amount of
calcium chloride is near to the saturation value (0.85 gCaCl ·gaq.sol.-1). Therefore, the
2
presence of CaCl2 enhances the catalytic performance, being the optimal amount equal
-1
to 0.65 gCaCl ·gaq.sol.
.
2
Next, the influence of the catalyst concentration has also been evaluated by
varying the xylose:catalyst weight ratio (10:1-1:1) (Fig. 7), by using the previously
-1
optimized amount of 0.65 gCaCl2·gaq.sol. and increasing the amount of catalyst,
maintaining constant the concentration of xylose (0.15 g xylose in 1.5 mL water). It can
be observed that xylose conversion is higher than 80% when alumina was incorporated,
attaining similar values (90-94%) for xylose:catalyst weight ratios between 6:1 and 1:1.
With respect to the furfural yield, a decrease was found when the amount of catalyst
raised, which would corroborate that γ-Al2O3 enhances xylose conversion but favoring
side reactions, consequently decreasing the furfural yield. Thus, maximum furfural
yields (42-44%) were attained for xylose:catalyst weight ratio equal to 3:1 and 6:1. In
the case of a xylose:catalyst weight ratio of 1:1, the furfural yield achieved was 33% in
spite of xylose conversion also was high. Therefore, the optimum amount of catalyst
should be that corresponding to a weight ratio of xylose:catalyst between 3:1 and 6:1,
because a higher loading of alumina could favor secondary reactions, thus decreasing
furfural yield.
Therefore, the kinetics of the dehydration process was studied at 150 ºC, with
γ-Al2O3 (xylose:catalyst weight ratio of 3:1) and 0.65 gCaCl2·gaq.sol.-1 (Fig. 8). It should be
noted that unlike the results found at higher temperature (175 ºC), both xylose
conversion and furfural yield are considerably greater in presence of γ-Al2O3 and CaCl2
15

with respect to values obtained only with CaCl2. It should be mentioned that xylose
conversion and furfural yield values were 94% and 44% respectively after only 30
minutes of reaction, (Fig. 8), attaining almost full xylose conversion and a furfural yield
of 55% after 50 min at the same temperature. In order to rule out the homogeneous
contribution of soluble aluminum species, the analysis of Al in the reaction medium,
after 50 min of reaction at 150 ºC in the presence of CaCl2, has been carried out by ICP-
MS. Thus, only 1.5 wt% of the total Al initially present in the catalyst was detected in
the aqueous phase, so the contribution of soluble aluminum species was negligible in
the dehydration of xylose to furfural. These results indicate that CaCl2 improves the
catalytic behavior of γ-Al2O3 at lower temperatures, and there may be a synergistic
effect between both, minimizing side reactions, which would demonstrate that this salt
can play a key role. Indeed, unlike the effect observed at 175 ºC, the partition
coefficient after 30 min at 150 ºC was higher by using γ-Al2O3 and CaCl2 together
(R=2.4) than if only CaCl2 was employed (R=1.7). Thus, the combined use of alumina
and CaCl2 at 150 ºC improved the catalytic performance, enhancing furfural production
which was quickly extracted by toluene and consequently increasing the partition
coefficient. Considering that calcium cations could interact with the strongest acid sites
of alumina and consequently decrease the side reactions that lead to furfural yield
losses, γ-Al2O3 was put in contact with CaCl2 under the same reaction conditions (0.65
gCaCl ·gaq.sol−1, 150 ºC and 50 min) without xylose in the medium. This catalyst, denoted
2
as exCa-Al2O3, was recovered and characterized by XPS in such a way that very low
atomic concentration of Ca (0.35 at%) was detected on the catalyst surface. The
catalytic behavior of exCa-Al2O3 has been also evaluated after 50 min at 150 ºC without
adding more CaCl2 (Figure 8). It can be observed that this catalyst led to lower values of
16

xylose conversion and furfural yield (29 and 11%, respectively) than those found in
presence of CaCl2 and/or γ-Al2O3. Thus, the improvement of catalytic behavior is not
favored by this cation exchange. Therefore, it was necessary a more exhaustive analysis
to prove the existence of this synergistic effect between CaCl2 and catalyst.
In this sense, García-Sancho et al. [42] have previously found that calcium
cations can alter the anomeric equilibrium of D-gucose, enhancing its dehydration rate
to HMF since Ca2+ favored the formation of the α-D-glucopyranose anomer,. It was
observed that the combination of CaCl2 and an acid catalyst, γ-Al2O3, provided high
HMF yields at very short reaction times. These results coincided with those found by
Rasrendra et al. [34], who reported that anions had a less effect for glucose conversion
than cations. As it has been mentioned, alkaline earth cations can form bidentate
complexes with different carbohydrates [53, 54]. The stability of the sugar-cation
complex depends on both the stereochemical disposition of the hydroxyl groups of
sugar and the Lewis acid strength of the cation [55-59]. The most favorable
configuration of a carbohydrate to interact with a divalent cation is the ax.-eq.-ax.
(positions of OH groups at C1, C2 and C3 in pyranose form, respectively; ax: axial and
eq: equatorial positions) configuration, but other cyclic sugars without this sequence,
like D-glucose, can form weaker Ca2+ complexes by interaction of alkaline earth metal
cations with two oxygen atoms [42, 53, 54]. In the case of xylose, the presence of Ca2+
could influence on the anomeric distribution of xylose, which could follow a similar
trend as that found in the case of glucose. As mentioned, sugar complexes with alkaline
earth halides are weak and unstable in water, due to the competition of water and sugar
for the cation. The interaction of cations with the oxygen atom of hydroxyl groups is
regarded as the main driving force for complex formation [59]. Therefore, it would be
17

possible that CaCl2 accelerates xylose dehydration as was found in the case of glucose
through complexation with salts. It was previously demonstrated that the addition of
CaCl2 induced the displacement of the β-form of D-glucose towards the α-conformer, as
previously commented [39]. On the other hand, Enslow et al. [36] concluded that the
presence of alkaline salts increased the -xylopyranose/β-xylopyranose ratio, although
it has been reported that the interaction between carbohydrates and divalent cations is
much stronger than those with monovalent cations [39, 40].
In order to verify these assumptions, 1H-NMR spectroscopy at room temperature
has been determined for different solutions of D-xylose in presence of -Al2O3 and
1
alkaline earth metal chlorides in deuterated water (Fig. 10). The H-NMR spectrum of
D-xylose in D2O (Fig. 9A) shows the equilibrium of both conformers,  and β (Scheme
1), although the intensity of -xylopyranose is much higher. This same trend was
detected after incorporation of -Al2O3, CaCl2 or MgCl2 to the xylose solution (Fig. 9B,
C and D, respectively). It has also been reported that the furanose content at equilibrium
in water is usually less than 1%, at 30 ºC [60]. Thus, the -xylopyranose/β-
xylopyranose ratio is higher than 1 in all cases, which indicates that the -xylopyranose
is favored at room temperature in spite of the addition of alumina or alkaline earth salts.
This fact may be due to the interaction between xylose and cations is weaker than that
found for other carbohydrates containing ax.-eq.-ax. configurations, such as ribose and
galactose, since the hydroxyl groups in -xylopyranose are in eq.-eq.-eq. orientation and
those in -xylopyranose are in ax.-eq.-eq. orientation.
On the other hand, Mikkola et al. reported that the ratio of -xylopyranose/-
xylopyranose in aqueous solution increases with the temperature [60]. In order to check
18

this fact, the solutions previously measured at room temperature were heated in a sealed
NMR tube at 110 °C for 24 h, and the corresponding spectra were registered (Fig. 10).
In our case, differently to that observed by Mikkola et al., the form is favored with
respect to the  form in all cases, although the analysis conditions were not similar to
those used by Mikkola et al., since they obtained the 1H-NMR spectra after 1 h at each
temperature, between 20-100 ºC. However, the trend of -xylopyranose/-xylopyranose
ratio after heating is not the same in all cases, being considerably higher in the case of
xylose dissolved in D2O with CaCl2 (Fig. 10C). Thus, the presence of CaCl2 at high
temperature promotes the formation of the α anomer in comparison with the -
xylopyranose/-xylopyranose ratio obtained with γ-Al2O3, MgCl2 or in absence of
catalyst or salt (0.94 versus 0.53-0.57) (Figs. 10A, 10B and 10D). It should be also
noted that the peaks corresponding to the signal of the furan ring began to appear in the
case of xylose with γ-Al2O3 (Fig. 10B).
In a previous mechanistic study, Nimlos et al. concluded that the intramolecular
rearrangement is more favorable for furfural formation than the pyranose ring opening
[61]. They also affirmed that the dehydration reaction is favored at the C2OH position
of xylopyranose with respect to C1OH and the fragmentation products increased when
dehydration started at C3OH or C4OH. Later, Enslow et al. reported that dehydration of
β-xylopyranose at the C2 position was energetically more favorable than at the C1
position, but both C1OH and C2OH initiated dehydration was possible in the case of -
xylopyranose due to the hydroxyl group in the axial position of the conformer was
more easily protonated and removed as water than the equatorial one [36]. These
authors also observed that the initial dehydration at C1OH group occurred more quickly
than that initiated at C2OH group when alkaline halides were added to the reaction
19

1
medium. Taking into account these conclusions and the results obtained by H-NMR
spectroscopy, it would be possible that dehydration of xylose is considerably enhanced
in presence of CaCl2 because of this salt increases the -xylopyranose/-xylopyranose
ratio, in such a way that xylose dehydration could take place through both C1OH and
C2OH pathways, being faster in the first case. Thus, a similar effect to that found by
Enslow et al. [36] for alkaline halides would explain the improved xylose conversion
and furfural yield observed with CaCl2. However, this improvement was not detected in
the presence of MgCl2, since this salt did not rise the -xylopyranose/-xylopyranose
ratio, as inferred from its 1H-NMR spectrum (Fig. 10D), which would explain its worse
catalytic performance. Therefore, this fact could justify the improvement of catalytic
performance found in presence of CaCl2 at 175 ºC with respect to the rest of cases.
Once that first dehydration was carried out, the high reaction temperature, 175 ºC,
favors the next dehydration steps, and consequently the furfural production, increasing
the partition coefficient since the existence of side reactions is decreased due to the
absence of acid sites.
However, different results in the dehydration process were attained at lower
temperature (150 ºC) and the best catalytic performance was found for a cooperative
1
contribution of both CaCl2 and γ-Al2O3. Considering that H-NMR spectroscopy has
demonstrated that CaCl2 increases the -xylopyranose/-xylopyranose ratio and the
presence of furfural is only detected in the spectrum of γ-Al2O3, this synergistic effect
between salt and Lewis acid catalyst could be explained in such a way that CaCl2
provides the correct anomer for efficient dehydration and γ-Al2O3 accelerates and
promotes the transformation of D-xylose to furfural when the reaction is carried out at
20

150 ºC, since the losses of furfural due to the unwanted side reactions are minimized
under these reaction conditions.
On the other hand, the reuse of γ-Al2O3 has also been evaluated in the presence
-1
of 0.65 gCaCl ·gaq.sol. with a xylose:catalyst weight ratio of 3:1 at 150 ºC for 50 min,
2
without any type of treatment after each catalytic cycle (Fig. 11). The catalyst remained
inside the reactor, without any type of treatment, and after each catalytic run a new
mixture (xylose, CaCl2, water and toluene) was added to carry out the reaction under
identical experimental conditions. It can be observed that the catalyst maintains its
activity after ten catalytic runs in such a way that the conversion values only decrease
from 98% to 89%, after the tenth catalytic cycle, maintaining the furfural yield between
47%-66% throughout the cycles. However, it should be noted that a small increase for
furfural yield can be found between different catalytic runs, mainly between 1st-2nd and
7th-8th cycles. This fact is probably due to a fraction of furfural formed could be retained
on the catalytic surface after the first cycle, which could be desorbed after heating the
new reaction mixture in the next catalytic run, enhancing the furfural yield for this
second cycle. In order to corroborate this fact, the first reaction cycle was carried out
again, but the used γ-Al2O3 catalyst was washed firstly with toluene and then with
water, under vacuum filtration, to facilitate the desorption of furfural adsorbed on the
catalyst surface. The analysis of these solvents after washing by HPLC has allowed to
detect the presence of furfural in both phases. However, the absence of xylose in the
aqueous phase would indicate that this compound is not adsorbed on the surface of γ-
Al2O3, or the washing is not enough for its desorption because its interaction is stronger.
Therefore, we could conclude that these variations in furfural yields throughout the
cycles during the reuse of the catalyst, are due to the retention of the furfural formed on
21

the surface of γ-Al2O3. This fact has been corroborated by CHN analysis of spent
catalyst which showed a carbon content of 5.98 wt% higher than 0.05 wt% detected for
fresh alumina. Likewise, the catalyst has been recovered after 1st catalytic run without
washing (γ-Al2O3 (u)) and after washing with toluene and water (γ-Al2O3 (w)) for their
characterization. Thus, it can be checked that both BET surface area and pore volume
considerably decrease for γ-Al2O3 (u) (Table 1). Moreover, the average pore diameter
slightly increases which confirms that the smallest pores are blocked by adsorbed
species. However, this fact is much less abrupt in the case of γ-Al2O3 (w) which
corroborates again that the washing with toluene and water allows to extract an
important fraction of adsorbed species, furfural among them, as detected in the analysis
of washing solvents. It has also been demonstrated by XRD that the catalyst preserves
its structure after reaction (not shown), since typical diffraction lines associated to
gamma-aluminum oxide remain after the catalytic process (PDF 00-029-1486)
Finally, the use of γ-Al2O3 and CaCl2 has been evaluated to obtain furfural from
liquors obtained from olive stones, at 175 °C after 50 min (Fig. 12). This catalytic test
has been realized at higher temperature in order to facilitate hydrolysis reaction, since
the hemicellulose hydrolysis carried out in the previous treatment was not complete. It
can be checked that the synergistic affect between γ-Al2O3 and CaCl2 has allowed to
increase the furfural yield from 53% for the non-catalyzed process until 85%. Therefore,
the combination of γ-Al2O3 and CaCl2 can be also employed for furfural production
from lignocellulosic biomass.
4. Conclusions
-Al2O3 has demonstrated to be an active catalyst for the dehydration of xylose
to furfural, since it possesses weak and strong Lewis acid sites, as inferred from the
22

NH3-TDP and pyridine adsorption coupled to FTIR spectroscopy. The use of a
water:toluene biphasic system has allowed to attain xylose conversion and furfural yield
values of 86% and 26%, respectively, after 50 min at 175ºC in presence of -Al2O3. The
addition of different alkaline earth metal chlorides, such as CaCl2 and MgCl2, in order
to decrease secondary reactions has been evaluated, in such a way that an increase in
partition coefficient, which indicates the proportion of furfural between the organic and
aqueous phases, was observed. In this sense, CaCl2 was much more effective due to the
presence of calcium affected to the -xylopyranose/-xylopyranose anomeric
1
equilibrium and favored the formation of the α anomer, as revealed the H NMR
spectra, which facilitated the dehydration process. A synergistic effect between -Al2O3
and CaCl2 was found, attaining almost full xylose conversion and a furfural yield of
55% after 50 min at 150ºC since CaCl2 increased the -xylopyranose/-xylopyranose
ratio and γ-Al2O3 accelerated the furfural production in these reaction conditions in
which the unwanted side reactions were minimized. The catalytic activity of -Al2O3
was also stable and, without any pretreatment, for ten catalytic cycles, with almost full
xylose conversion and maintaining an average furfural yield of 57%. The synergistic
effect between CaCl2 and -Al2O3 was also found by using hemicellulosic liquors
obtained from olive stones as source of carbohydrates, achieving until a furfural yield of
83% after 50 min at 175ºC, in comparison with the use of the individual components.
Acknowledgements
The authors want express their gratitude to the Spanish Ministry of Economy
and Competitiveness (CTQ2015-64226-C03-3-R project) and FEDER (European
Union) funds about financial support. J.A.C. and C.G.-S. thank to the Malaga
23

University for the financial support. L.S.C. gratefully acknowledges support from
Spanish MINECO via the concession of a Ramon y Cajal contract (RYC-2015-17109).
References
[1] In T. Werpy and G. Peterson (Editors), Top Value Added Chemicals from Biomass,
Vol. Volumen 1, Pacific Northwestern National Laboratory; 2004.
[2] M. Stocker, Angew. Chem.-Int. Edit., 47 (2008), pp. 9200-9211.
[3] R. van Putten, J. van der Waal, E. de Jong, C. Rasrendra, H. Heeres, J. de Vries,
Chem. Rev., 113 (2013), pp. 1499-1597.
[4] D. Alonso, J. Bond, J. Dumesic, Green Chem., 12 (2010), pp. 1493-1513.
[5] T. Wang, M. Nolte, B. Shanks, Green Chem., 16 (2014), pp. 548-572.
[6] G. Garrote, E. Falqué, H. Domínguez, J.C. Parajó, Bioresource Technol., 98 (2007),
pp. 1951-1957.
[7] W.G. Glasser, R.K. Jain, M.A. Sjöstedt, Thermoplastic pentosan-rich
polysaccharides from biomass, Vol. US Patent 5430142, US Patent; 1995.
[8] R. Mariscal, P. Maireles-Torres, M. Ojeda, I. Sádaba, M. López Granados, Energ.
Environ. Sci., 9 (2016), pp. 1144-1189.
[9] M.J. Antal, T. Leesomboon, W.S. Mok, Carbohyd. Res., 217 (1991), pp. 71-85.
[10] R. Karinen, K. Vilonen, M. Niemelä, ChemSusChem, 4 (2011), pp. 1002-1016.
[11] In K.J. Zeitsch (Editors), The chemistry and technology of furfural and its many
by-products, Elsevier Science; 2000.
[12] R. Weingarten, G. Tompsett, W. Conner, G. Huber, J. Catal., 279 (2011), pp. 174-
182.
[13] A. Mittal, S. K. Black, T. B. Vinzant, M. O'Brien, M. P. Tucker, D. K. Johnson,
ACS Sustain. Chem. Eng., 5 (2017), pp. 5694-5701.
[14] J. N. Chheda, Y. Román-Leshkov, J. A. Dumesic, Green Chem., 9 (2007), pp. 342-
350.
[15] E. I. Gürbüz, S. G. Wettstein, J. A. Dumesic, ChemSusChem, 5 (2012), pp. 383-
387.
[16] C.L. Burket, S. Sabesan, Process for furfural production from biomass, Vol.
Volume US Patent 0157697A1, US Patent; 2012.
[17] T.Y. Zhang, R. Kumar, C.E. Wyman, RSC Adv., 3 (2013), pp. 9809-9819.
[18] N. Shi, Q.Y. Liu, Q. Zhang, T.J. Wang, L.L. Ma, Green. Chem., 15 (2013), pp.
1967-1974.
[19] Y. Yang, C. W. Hu, M. M. Abu-Omar, ChemSusChem, 5 (2012), pp. 405-410.
[20] A. Dias, M. Pillinger, A. Valente, J. Catal., 229 (2005), pp. 414-423.
[21] R. Weingarten, G. Tompsett, W. Conner, G. Huber, J. Catal., 279 (2011), pp. 174-
182.
[22] I. Agirrezabal-Telleria, A. Larreategui, J. Requies, M. Guemez, P. Arias,
Bioresource Technol., 102 (2011), pp. 7478-7485.
[23] A. Dias, M. Pillinger, A. Valente, Micropor. Mesopor. Mat., 94 (2006), pp. 214-
225.
[24] X. Shi, Y. Wu, H. Yi, G. Rui, P. Li, M. Yang, G. Wang, Energies, 4 (2011), pp.
669-684.
24

[25] P. Zhou, Z. Zhang, Catal. Sci. Technol., 6 (2016), pp. 3694-3712.
[26] J. He, H. Li, S. Saravanamurugan, S. Yang, ChemSusChem, 12 (2019), pp. 347-
378.
[27] J.E. Romo, N.V. Bollar, C.J. Zimmermann, S.G. Wettstein, ChemCatChem, 10
(2018), pp. 4819-4830.
[28] C. Garcia-Sa.ncho, I. Agirrezabal-Telleria, M. Guemez, P. Maireles-Torres, Appl.
Catal. B-Environ., 152 (2014), pp. 1-10.
[29] C. Song, H. Liu, Y. Li, S. Ge, H. Wang, W. Zhu, Y. Chang, C. Han, H. M. Li,
Chin. J. Chem., 32 (2014), pp. 434-442.
[30] J. Kruger, V. Choudhary, V. Nikolakis, D. Vlachos, ACS Catal., 3 (2013), pp.
1279-1291.
[31] W. Zeng, D. Cheng, F. Chen, X. Zhan, Catal. Lett., 133 (2009), pp. 221-226.
[32] S. You, N. Park, E. Park, M. Park, J. Ind. Eng. Chem., 21 (2015), pp. 350-355.
[33] S. You, Y. Kim, E. Park, Reac. Kinet. Mech. Catal., 111 (2014), pp. 521-534.
[34] G. Marcotullio, W. De Jong, Green Chem., 12 (2010), pp. 1739-1746.
[35] G. Marcotullio, W. de Jong, Carbohyd. Res., 346 (2011), pp. 1291-1293.
[36] K.R. Enslow, A.T. Bell, ChemCatChem, 7 (2015), pp. 479-489.
[37] C. Rasrendra, I. Makertihartha, S. Adisasmito, H. Heeres, Top. Catal., 53 (2010),
pp. 1241-1247.
[38] Y. Roman-Leshkov, J. Dumesic, Top. Catal., 52 (2009), pp. 297-303.
[39] J. Teychené, H. Roux-De Balmann, S. Galier, Carbohyd. Res., 448 (2017), pp.
118-127.
[40] P.K. Banipal, A.K.C.N. Hundal, T.S. Banipal, Carbohyd. Res., 345 (2010), pp.
2262-2271.
[41] E. Combs, B. Cinlar, Y. Pagan-Torres, J. Dumesic, B. Shanks, Catal. Commun., 30
(2013), pp. 1-4.
[42] C. Garcia-Sancho, I. Fúnez-Núnez, R. Moreno-Tost, J. Santamaría-González, E.
Pérez-Inestrosa, J.L.G. Fierro, P. Maireles-Torres, Appl. Catal. B-Environ., 206 (2017),
pp. 617-625.
[43] D. Amendola, D.M. De Faveri, I. Egües, L. Serrano, J. Labidi, G. Spigno,
Bioresource Technol., 107 (2012), pp. 267-274.
[44] M.J Campos Molina, M. López Granados, A. Gervasini, P. Carnit, Catal. Today,
254 (2015), pp. 90-98.
[45] P. Berteau, B. Delmon, A. Crucq, A. Frennet, Catal. Today, 5 (1989), pp. 121-137.
[46] J. Gallo, C. Bisio, G. Gatti, L. Marchese, H. Pastore, Langmuir, 26 (2010), pp.
5791-5800.
[47] A. Ramirez, B. Lopez, L. J. Sierra, Phys. Chem. B, 107 (2003), pp. 9275-9280.
[48] C. García-Sancho, J.A. Cecilia, A. Moreno-Ruiz, J.M. Mérida-Robles, J.
Santamaría-González, R. Moreno-Tost, P. Maireles-Torres, Appl. Catal. B-Environ.,
179 (2015), pp. 139-149.
[49] J. Chan, Y. Zhang, ChemSusChem, 2 (2009), pp. 731-734.
[50] C. McNeff, D. Nowlan, L. McNeff, B. Yan, R. Fedie, Appl. Catal. A-Gen., 384
(2010), pp. 65-69.
[51] C. Rasrendra, J. Soetedjo, I. Makertihartha, S. Adisasmito, H. Heeres, Top. Catal.,
55 (2012), pp. 543-549.
[52] X. Wu, J. Fu, X. Lu, Bioresource Technol., 119 (2012), pp. 48-54.
[53] S. Tyrlik, D. Szerszen, M. Olejnik, W. Danikewicz, J. Mol. Catal. A-Chem., 106
(1996), pp. 223-233.
25

[54] C. Yang, X. Lu, W. Lin, X. Yang, J. Yao, Chem. Res. Chin., U22 (2006), pp. 524-
532
[55] J. Angyal, Adv. Carbohyd. Chem. Biochem., 49 (1991), p. 19-35.
[56] J. Angyal, Adv. Carbohyd. Chem. Biochem., 47 (1989), p. 1-43.
[57] J. Angyal, Adv. Carbohyd. Chem. Biochem., 42 (1984), p. 15-68.
[58] K. Bock, C. Pedersen, Adv. Carbohyd. Chem. Biochem., 41 (1983), p. 27-66.
[59] A. Rendleman, Adv. Carbohyd. Chem., 21 (1966), p. 209-271.
[60] J.P. Mikkola, R. Sjoholm, T. Salmi, P. Maki-Arvela, Catal. Today, 48 (1999), pp.
73-81.
[61] M.R. Nimlos, X. Qian, M. Davis, M. E. Himmel, D. K. Johnson, J. Phys. Chem. A,
110 (2006), pp. 11824-11838.
26

List of captions
Figure 1. Nitrogen adsorption (solid symbols)-desorption (open symbols) isotherms (A)
and pore size distribution curve (B) of the -Al2O3 catalyst.
Figure 2. NH3-TPD curve of the γ-Al2O3 catalyst.
Figure 3. FTIR spectra of adsorbed pyridine over γ-Al2O3 outgassed at different
temperatures.
-1
Figure 4. (A) Influence of adding CaCl2 (0.65 g.·g.aq.sol.-1), MgCl2 (0.39 g.·g.aq.sol.
)
and/or γ-Al2O3 on xylose conversion and furfural yield at 175 °C after 50 min
(xylose:catalyst weight ratio= 3:1). (B) Furfural yield as function of the partition coefficient
of furfural in both phases (175 °C, 50 min).
Figure 5. Xylose conversion (A) and furfural yield (B) as a function of reaction time in
presence of γ-Al2O3 and/or CaCl2 (175 °C, 0.65 gCaCl₂ ·gaq.sol.-1, 0.05 g of γ-Al2O3 and
xylose:catalyst weight ratio = 3:1).
Figure 6. Effect of CaCl2 concentration on xylose conversion and furfural yield without
catalyst at 150 ºC after 30 min.
Figure 7. Xylose conversion and furfural yield as function of xylose:catalyst weight
-1
ratio for γ-Al2O3 at 150 ºC and 30 min in presence of 0.65 gCaCl₂ ·gaq.sol. (0.15, 0.05,
0.025 and 0.015 g of γ-Al2O3 for xylose:catalyst weight ratio = 1:1, 3:1, 6:1 and 10:1,
respectively).
Figure 8. Evolution of xylose conversion (solid symbols) and furfural yield (open
symbols) with the reaction time in presence of γ-Al2O3 and/or CaCl2 and exCa-Al2O3
(150 °C, 0.65 gCaCl₂ ·gaq.sol.-1, 0.05 g of γ-Al2O3 and xylose:catalyst weight ratio = 3:1).
Figure 9. 1H NMR spectra of xylose in D2O (A), before and after addition of Al2O3 (B),
CaCl2 (C) and MgCl2 (D) at room temperature.
1
Figure 10. H NMR spectra of xylose in D2O (A), with Al2O3 (B), CaCl2 (C) and
MgCl2 (D) after 24 h at 110 ºC.
Figure 11. Reutilization study for the γ-Al2O3 catalyst at 150ºC after 50 min in the
presence of 0.65 gCaCl2·gaq.sol.-1 (0.05 g of γ-Al2O3 and xylose/catalyst weight ratio of
3:1).
Figure 12. Influence of adding Al2O3 and/or CaCl2 (0.65 g.·g.aq.sol.-1) on furfural yield
for xylose solutions from olive stones at 175 °C after 50 min (0.05 g of γ-Al2O3 and
xylose:catalyst weight ratio= 3:1).
27

180
150
120
90
A
-Al O
2
3
60
30
0
0.0
0.2
0.4
P/P0
0.6
0.8
1.0
B
0.003
0
5
10 15 20 25
d (nm)
p
Figure 1
28

-Al O
2
3
100 200 300 400 500
T (ºC)
Figure 2
29