Dehydration of D-xylose into furfural over bimetallic salts of heteropolyacid in DMSO/H2O mixture

Article abstract of DOI:10.1016/j.apcata.2018.03.027

Dehydration of D-xylose to yield furfural was carried out using bimetallic salts of a heteropolyacid as the catalyst at 160–220 °C in the DMSO/H₂O mixtures. The effect of Sn/Cs molar ratio of the bimetallic salts of 12-tungstophosphoric acid (PW) obtained by ultrasound-assisted coprecipitation on dehydration of D-xylose was investigated. The resultant catalysts were characterized by X-ray diffraction, N₂ adsorption, field emission scanning electron microscopy and energy dispersive X-ray (EDX). It was found that the Sn-Cs codoped PW catalysts retained the Keggin crystal structure of PW. Sn_0.625Cs_0.5PW was the most active catalyst in the dehydration of D-xylose into furfural. The maximum D-xylose conversion (close to 100 wt %) and furfural yield (63 wt %) were achieved at 200 °C for 3 h in DMSO/H₂O mixtures. It was found that 16.7 wt % Sn_0.625Cs_0.5PW on a chitosan-derived support displayed similar catalytic activity to that of Sn_0.625Cs_0.5PW and good stability after recycling six times.

Full text of DOI:10.1016/j.apcata.2018.03.027

Accepted Manuscript
Title: Dehydration of D-xylose into furfural over bimetallic
salts of heteropolyacid in DMSO/H2O mixture
Authors: Xiaoqian Guo, Feng Guo, Yishan Li, Zhangqin
Zheng, Zhexu Xing, Zihan Zhu, Ting Liu, Xin Zhang, Ying Jin
PII:
S0926-860X(18)30146-7
DOI:
Reference:
https://doi.org/10.1016/j.apcata.2018.03.027
APCATA 16601
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
13-12-2017
7-3-2018
26-3-2018
Please cite this article as: Guo X, Guo F, Li Y, Zheng Z, Xing Z, Zhu Z, Liu
T, Zhang X, Jin Y, Dehydration of D-xylose into furfural over bimetallic salts
of heteropolyacid in DMSO/H2O mixture, Applied Catalysis A, General (2010),
https://doi.org/10.1016/j.apcata.2018.03.027
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.

Dehydration of D-xylose into furfural over bimetallic salts of heteropolyacid in DMSO/H2O
mixture
Xiaoqian Guo1, Feng Guo1*, Yishan Li2, Zhangqin Zheng1, Zhexu Xing1, Zihan Zhu1, Ting Liu1, Xin
Zhang2, Ying Jin1
1School of Food and Environment, Dalian University of Technology, Panjin Campus, 2 Dagong
Road, Liaodong Bay New District, Panjin 124221, China
2School of Petroleum and Chemical Engineering, Dalian University of Technology, Panjin Campus,
2 Dagong Road, Liaodong Bay New District, Panjin 124221, China
Revised for Applied Catalysis A: General
(Jan. 2018)
* Corresponding author:
Tel.: +86-427-2631892.
E-mail Address: 0411guofeng@dlut.edu.cn (F. Guo);
Graphical abstract
1

Research highlights
 ►A series of bi-metallic salts of heteropoly acid was synthesized for the conversion of xylose
to furfural.
 ►Introduction of Sn4+ and Cs+ did not change the Keggin-type structure of original
heteropoly acid.
 ►Sn4+ occupied the surface available space of Cs+, which helps improve the catalytic activity.
 ►The addition of DMSO can effectively improve the furfural selectivity.
 ►Sn0.625Cs0.5PW/CH displayed similar catalytic activities with Sn0.625Cs0.5PW and good
stability after six cycles of recycling.
Abstract
Dehydration of D-xylose to yield furfural was carried out using bimetallic salts of a heteropolyacid
as the catalyst at 160 – 220 °C in the DMSO/H2O mixtures. The effect of Sn/Cs molar ratio of the
2

bimetallic salts of 12-tungstophosphoric acid (PW) obtained by ultrasound-assisted coprecipitation
on dehydration of D-xylose was investigated. The resultant catalysts were characterized by X-ray
diffraction, N2 adsorption, field emission scanning electron microscopy and energy dispersive X-ray
(EDX). It was found that the Sn-Cs codoped PW catalysts retained the Keggin crystal structure of
PW. Sn0.625Cs0.5PW was the most active catalyst in the dehydration of D-xylose into furfural. The
maximum D-xylose conversion (close to 100 wt %) and furfural yield (63 wt %) were achieved at
200 °C for 3 h in DMSO/H2O mixtures. It was found that 16.7 wt % Sn0.625Cs0.5PW on a
chitosan-derived support displayed similar catalytic activity to that of Sn0.625Cs0.5PW and good
stability after recycling six times.
Keywords: D-xylose; furfural; dehydration; heteropolyacid salts; bimetallic salts
1. Introduction
The decrease of global fossil resources has become a driving force for the use of biomass as a
resource. Development of renewable energy and chemicals from biomass lower the reliance on oil
resources and can provide new products [1]. Rated as one of the top ten bio-based chemicals by the
US Department of Energy, furfural could be used to sustainably produce C4 and C5 chemicals [2].
D-Xylose and pentosans are promising non-toxic renewable feedstock for furfural production
through control of the dehydration conditions.
Researchers have conducted much work to improve the conversion efficiency of D-xylose to
furfural through dehydration. However, the industrial production of furfural lacks competitive edge
because the current technologies are inefficient with low yield of furfural [3]. Dehydration of
D-xylose or hemicelluloses by traditional catalysts, such as mineral acid, heteropolyacid and metal
salts, requiring a high temperature (~ 250 ℃) and long reaction time (24 h) [4]. In addition, these
catalysts have many shortcomings, including equipment corrosion, high pollution and high energy
consumption [5]. It has been suggested that development of dehydration technology and synthesizing
heterogeneous catalyst should be effective methods to overcome these problems.
Heterogeneous catalysts receive considerable attention because of their easy recycling through
simple filtration and ready operation as fixed beds [5]. The use of ionic liquids and high-pressure
CO2 for biomass conversion result in effective production of furfural without formation of humins
[6-9]. These processes show good development prospects, because it integrates the processes of
3

dissolution, fractionation, hydrolysis and conversion in one pot. Solvent recycling and decreased
energy demand should make this approach technologically and economically viable. It is speculated
that the combination of a heterogeneous catalyst with a suitable solvent will result in effective
conversion of D-xylose to furfural.
Many different heterogeneous catalysts, such as Nafion [10], zeolite [11], Amberlyst-15 [12],
SO3H-MCM-41 [13], and as well as resorcinol-formaldehyde resin carbon [14], have been explored.
However, most heterogeneous catalysts display deficiencies, such as low furfural yield (~ 30%), poor
thermal stability, and high price [12, 14, 15]. Heteropolyacid salts exhibit promising furfural
production performances. Heteropolyacid salts inherit most of the functions and characteristics of a
heteropolyacid, e.g., non-explosive, strong acidity, and strong oxidizing capability [16]. However,
apart from cesium salts (Cs) of heteropolyacid (denoted as CsxH3−xPW12O40), heteropolyacid salts
have many shortcomings, such as low surface area and high solubility in polar solvents [17]. It has
been reported that the acidity of heteropolyacid Cs has the order of H3PW12O40 > Cs2HPW12O40 >
Cs2.5H0.5PW12O40 [18]. However, the acidic Cs salt Cs3PW12O40 has a much higher surface area (~200
m2/g) than that of the initial heteropolyacid H3PW12O40 (~5 m2/g) [19]. Meanwhile, an increase in
the content of acidic sites on the surface induced by Cs+/H+ exchange changes the catalytic capability
of CsxH3−xPW12O40. The initial activities of these catalysts for D-xylose dehydration displayed the
order of H3PW12O40 < Cs2.0H1PW12O40 < Cs2.5H0.5PW12O40 [20]. Moreover, the increasing Cs+
substitution of CsxH3−xPW12O40 could increase the hydrophobicity of the heteropoly compounds.
This makes CsxH3−xPW12O40 exhibit high stability in aqueous media.
Heteropolyacid cesium salts usually forms colloidal suspensions in polar media (i.e., H2O)
making it difficult to reuse the solid catalyst [21]. One solution to this problem would be to support
the heteropolyacid cesium salts on a carrier such as SiO2, ZrO2, TiO2, MCM-41, or SBA-15 [22, 23].
These catalyst supports have been used extensively for various organic reactions such as
hydrogenation, dehydration, hydrolysis, alkylation, acylation and esterification reactions [24-28]. On
the other hand, these materials demonstrate lower activity and selectivity than those of for
CsxH3−xPW12O40. It has been reported that heteropolyacid catalysts with Lewis acidity exhibit high
catalytic efficiency and mild specificity [29]. Therefore, introducing a Lewis acid into
CsxH3−xPW12O40 would change its catalytic properties and selectivity considering that
CsxH3−xPW12O40 already has high surface area and Brønsted acidity.
4

In this study, some of water-soluble heteropolyacid salts are first prepared. The catalyst cations
are substituted by Cs+ to form heterogeneous bi-metallic salts of the heteropolyacid. The Keggin
crystal structures of the catalysts are investigated. Then, the dehydration of D-xylose to form furfural
catalyzed by the bimetallic heteropolyacid salts and the influence of reaction parameters on D-xylose
dehydration are studied. Catalyst stability is also studied by recycling experiments.
2. Experiment
2.1 Materials
12-tungstophosphoric acid (H3PW12O40·xH2O, 99+% purity) and cesium chloride (99+% purity)
used for the catalyst preparation were purchased from J&K® Scientific Ltd. (Beijing, China). Stannic
chloride pentahydrate (99% purity), Ferric chloride hexahydrate (99% purity), potassium chloride
(99.5% purity), nickel chloride hexahydrate (98% purity), cerium nitrate hexahydrate (99% purity),
ammonium sulfate (99% purity), rubidium chloride (99% purity), and dimethyl sulfoxide (99%
purity) were bought from Damao Chemical Reagent Factory (Tianjin, China) and used for catalyst
preparation without further purification. D-xylose in conversion experiments was obtained from
Aladdin Reagent Co., Ltd. (Shanghai, China). Standard furfural (99% purity) and D-xylose (99%
purity) were purchased from Sigma-Aldrich (Shanghai, China).
2.2 Preparation and characterization of catalysts
Catalysts were prepared in a 200-mL Teflon hydrothermal vessel. In the presence of ultrasonic
irradiation, 20 mL of 1.8 mM metal chloride (e.g. SnCl4, FeCl2, CuCl2, NiCl2, (NH4)2SO4, or KCl)
aqueous solution was added dropwise to 100 mL of a 1 mM aqueous solution of H3PW12O40.
Subsequently, 20 mL of 5 mM CsCl (or RbCl) aqueous solution was added dropwise, which gave a
milky precipitate. The reaction Teflon vessel was then immersed in a thermostat-controlled oil-bath
controlled at 200 °C with vigorous magnetic stirring. After a certain reaction time, the mixture was
cooled to room temperature, centrifuged, and then washed with deionized water until the rinsing
water reached a pH of 6–7. The resulting sample was dried by lyophilization and then used a catalyst.
In addition, a chitosan–derived carbon material (CH) supported catalyst was prepared via a similar
process except that a certain amount of chitosan was dissolved in the aqueous solution of H3PW12O40
solution in advance. The catalyst loading occurred in parallel with the preparation of the catalyst
carrier. The catalyst loading was 30 wt %. All experiments were repeated three times, and the
5

reported values are the averages of the three replicates.
The crystal structure of the catalysts was analyzed using X-ray diffraction (XRD; XRD-7000s,
Shimadzu, Japan), measured with CuKα (λ = 0.154 nm) radiation at 40 kV and 20 mA by scanning at
a rate of 5°/min in the 2θ range of 5 – 100°. Scans were collected by an X-ray diffractometer control
and data acquisition system (MDI, US). The microstructure of the catalysts was characterized using a
Nova NanoSEM 450 (FEI) field-emission scanning electron microscope (FE-SEM) equipped with an
energy dispersive X-ray (EDX) spectrometer operating at an acceleration voltage of 20 kV and
working distance of 5 mm. The catalysts were dried and sprayed with gold before FE-SEM analysis.
Specific surface areas were determined by N2 adsorption at 77 K with the Brunauer-Emmett-Teller
(BET) method using an Autosorb-iQ-C chemisorption analyzer (Quantachrome Instruments). Surface
functional groups of solid catalysts were characterized by Fourier transform infrared (FT-IR)
spectroscopy (ThermoFisher Scientific, USA) using an ATR cell with a scanning range from 700 to
1500 cm-1.
2.3 Experimental procedure
A 20-mL autoclave equipped with a Teflon-lined hydrothermal vessel and microwave reactor
(MDS-6G, Shanghai) equipped with eight 50-mL Teflon vessels were used for the dehydration
experiments. The Teflon-lined hydrothermal vessel was charged with D-xylose, catalyst, and reaction
medium. The vessel was subsequently sealed in the reactor and placed in a circulating
constant-temperature oil bath (or microwave reactor). In a typical procedure, 0.1 g D-xylose, 0.08 g
Sn0.625Cs0.5PW and 20 ml solvent were put in the Teflon vessel and reacted at the set temperature
(160–200 °C) with magnetic stirring at 400 rpm.
2.4 Analytical methods
D-xylose was analyzed by high-performance liquid chromatography (HPLC; Ultimate 3000, US)
fitted with an Aminex HPX-87H column (Bio-Rad, Richmond, CA) and refractive index (RI)
detector. Furfural was determined using the same column, but was measured with an ultraviolet (UV)
detector at 280 nm. The column temperature was set at 50ꢀ°C. The mobile phase was 5 mM of H2SO4
at a flow rate of 0.6 ml/min.
D-xylose conversion rate (X, mol%) and furfural yield (Y, mol%) were calculated as follows:
mol concentration of D-xylose in products
X= (1-
) ×100%
(1)
mol concentration of initial D-xylose
6

mol concentration of furfural
Y= (
) ×100%
(2)
mol concentration of initial D-xylose
3. Result and Discussion
3.1 FE-SEM analysis
The microtopography of the bimetallic salt-exchanged PW catalysts were observed by FE-SEM.
The morphology of Cs0.5H2.5PW was heterogeneous spherical particles, as shown in Fig. 1(a).
Spherical particles of submicron size with a particle size distribution between 100 and 500 nm were
obtained for Cs2.5H0.5 PW (Fig. 1b). The surface of these particles still remained rough. These
particles were composed of aggregates of small spherical particles obtained via re-crystallization
during hydrothermal preparation.
The introduction of Sn4+ into the mono-cationic salts caused the formation of agglomerates with
an irregular particle size of less than 250 nm (Fig. 1c and d). The average particle size and
distribution range of Sn0.625Cs0.5PW were similar to those of Sn0.5CsPW. In addition, both Sn0.5CsPW
and Sn0.625Cs0.5PW have rough surfaces. Such surface structure permits access of reactants to the
active sites, which is required to realize the catalytic activity of these heteropolyacids. SnxRb3−4xPW
was obtained via Sn4+ substitution for Rb+ and H+ in RbxH3−xPW. When Cs+ was substituted by Rb+,
the surface of Sn0.55Rb0.8PW became smoother than that of Sn0.625Cs0.5PW (Fig. 1e). In this study, a
chitosan-derived carbon material (denoted as CH) was used as a catalyst carrier to anchor catalytic
species. Fig. 1f shows that CH has layered-stacking arrangements because of the synergistic effects
of the hydrothermal process and heteropolyacid salts. The structures of Cs2.5H0.5PW and
Sn0.625Cs0.5PW were maintained after loading on the surface of the CH carrier (Fig. 1g and h). It can
be seen that submicron catalyst particles were uniformly dispersed on the surface of CH. CH has
abundant surface functional groups, such as C-OH, COOH and C=O [30, 31]. Such groups are
conductive to achieve rapid adsorption of substrates during the contact phase.
3.2 EDX study
Both solid bi-metallic and Cs of heteropolyacid can convert D-xylose into furfural effectively.
Besides number of surface active sites, catalytic species usually play a decisive role in reactions
catalyzed by solid catalysts. EDX analysis of the synthesized heteropolyacid salts was conducted to
assess their elemental distributions. The EDX spectra revealed the presence of P, W, O, Sn, Cs (or Rb)
elements in the heteropolyacid salts (Fig. S1).
7

According to the EDX area analysis of Sn0.625Cs0.5PW (Fig. S1b), the determined Sn/Cs atomic
ratio was 1.68, which is higher than that of the theoretical value (1.25). This indicates that excess
Sn4+ is distributed on the crystal surface of the heteropolyacid salt. Sn0.625Cs0.5PW displayed the
highest catalytic activity in the dehydration of D-xylose of the investigated catalysts. It is speculated
that Sn4+ occupied the available surface site of Cs+ and consequently achieved excellent dehydration
activity. Furthermore, the substitution of H+ by Rb+ was more stable than that of substitution by Cs+,
as confirmed by the EDX spectrum in Fig. S1c. It can be seen that Rb+ (EDX area of 34%) was
preferentially distributed on the surface of the solid catalyst compared with the case for Cs+ (EDX
area of 13%). However, Sn0.55Rb0.8PW showed lower dehydration activity than that of Sn0.65Rb0.5PW.
This further proves that a high distribution of Sn4+ helps to improve catalytic performance in
D-xylose dehydration. After Sn0.625Cs0.5PW was supported on CH, the Sn/Cs atomic ratio decreased
from 1.69 to 1.375. The presence of carbon in the EDX images reveals the interference of C with Sn
distribution.
3.3 XRD study
XRD analysis was used to evaluate the crystalline structure of the unsupported and CH-supported
bimetallic heteropolyacid salts, as shown in Fig. 2. All catalysts possessed the same crystalline
structure because all peaks existed at similar position. The Keggin structure of PW compounds was
confirmed by the presence of diffraction peaks at 2θof 8°–10°, 17°–20°, 26°–30° and 33°–35°. The
XRD patterns of Sn0.625Cs0.5PW showed diffraction peaks at 2θ=10.3°, 18.0°, 21.0°, 23.3° and 30.2°,
which are consistent with cubic system [32]. This was also confirmed by FT-IR spectroscopy (Fig.
S2a). The FT-IR spectra displayed the characteristic features of a Keggin-type structure of four bands
at 1077, 982, 885 and 778 cm−1, which corresponded to P-O, W = O, W – O – W, and W – O – W
asymmetric vibrations [33], respectively.
No diffraction peak confirming the existence of Sn-O or Sn-OH was observed, indicating that
Sn4+ substitution did not change the structure of the Keggin anion. The most intense peaks were
observed at around 27°, indicating that the Cs-substituted PW catalysts had crystalline structures.
Moreover, the diffraction peak at around 27° shifted to higher diffraction angle as the Cs+ content
increased (Fig. 2a-c). This peak shift behavior was caused by the decrease of lattice constant and loss
of water from the crystal [34]. The diffraction peaks for Sn0.625Cs0.5PW/CH suggested that the
Keggin structure remained after heteropolyacid immobilization. FT-IR data also indicated that the
8

Keggin structure was retained when loading Sn0.625Cs0.5PW on CH (Fig. S2b). After immobilization,
the three characteristic bands at 982, 885 and 778 cm−1 shifted to 986, 889 and 807 cm−1,
respectively. This may be partially caused by the formation of ion pair between Sn0.625Cs0.5PW and
the amine groups in CH [35].
3.4 N2 adsorption
Figure 3 shows the N2 adsorption–desorption isotherm and Barrett–Joyner–Halenda pore size
distribution curve of Sn0.625Cs0.5PW. Sn0.625Cs0.5PW presents typical Langmuir type–II
adsorption–desorption isothermal curves, which is a consistent with clustered particles. This
conclusion agrees with the morphology of Sn0.625Cs0.5PW observed by FE-SEM (Fig. 1(d)). It has
been reported that the specific surface area of PW and Cs+–doped PW are 5 and 156 m2/g,
respectively [36]. In this study, as seen in Table 2, the surface area of Sn0.625Cs0.5PW calculated from
the linear part of its BET plot is 140 m2/g. Obviously, Sn-Cs substitution of PW can increase its
surface area. The determined pore volume and diameter of Sn0.625Cs0.5PW are 0.32 cm3/g and 11 nm,
respectively. The mesoporous structure of Sn0.625Cs0.5PW is conducive to the ion-exchange between
Sn4+ and Cs+. After Sn0.625Cs0.5PW was supported on the surface of CH, the BET surface area
increased to 219 m2/g and the pore diameter increased to 9.6 nm. These results may be favorable
because mesoporous catalysts usually exhibit much higher catalytic activities, as a result of facile
substrate molecular diffusion to the catalytic sites in the internal pores.
3.5 Catalyst testing
3.5.1 Catalysts screening
All of the catalysts were prepared under the same conditions: hydrothermal treatment at 200 °C
for 2 h in a Teflon-lined vessel. Eight solid catalysts were synthesized and all exhibited D-xylose
conversion rates of nearly 100%. For comparison, H2SO4, HCl and CH3COOH were used to catalyze
D-xylose dehydration under the same conditions, as shown in Fig. 4. The catalytic activity of the
tested catalysts tested decreased with the order of H2SO4 > HCl > CH3COOH > Sn0.625Cs0.5PW >
H2RbPW > H0.5Cs2.5PW > HKCsPW > H0.5Fe0.5PW > HNi0.5CsPW > H0.5Ce0.5CsPW. Thus,
Sn0.625Cs0.5PW showed the highest activity among all the metallic ion exchanged PW catalysts, with
a furfural yield of 44%. This is because Sn4+ occupied the available surface sites of Cs+ on the
spherical particle surface of the Keggin crystal. Addition of transition metals could influence the
catalytic activity of PW when they are incorporated in the primary structure of the Keggin unit [37].
9

H2RbPW has a strong preference towards surface cation distribution but exhibited similar catalytic
activity (34.5%) to that of H0.5Cs2.5PW. This supports the notion that the cation distribution of Rb+
+
and Cs+ has little effect on structure and acidity of PW. Other M-Cs+ (M = Fe2+, Ni2+, Al3+, NH4 or
Ce4+) substituted HPW catalysts, displayed much lower the furfural yields of less than 11%.
A series of SnxCs3−4xPW (x = 0.25, 0.50, 0.55 and 0.625) catalysts were further synthesized to
study the effect of surface distribution of Sn4+ and Cs+ on furfural yield. Table 1 shows that furfural
yield indeed increased slightly with the content of Sn4+ on the surface of SnxCs3−4xPW. The optimal
attribution ratio of Sn4+ was 0.625, in which case a furfural yield of 44.4% was achieved in 3.5 h at
200 °C with 99% D-xylose conversion. A higher theoretical value of Sn4+ during SnxCs3−4xPW
preparation gave no further significant increase in surface Sn4+ content; instead, catalytic activity
decreased. Therefore, Sn0.625Cs0.5PW was selected for the following dehydration reactions.
3.5.2 Dehydration of D-xylose
The effect of variables on D-xylose dehydration with Sn0.625Cs0.5PW in different media was
studied (Fig. 5–7). Compared with water, DMSO/H2O mixtures lowered the required reaction
temperature to 180 °C (vs. 200 °C) and time to 3 h (vs. 3.5 h) while increasing the furfural yield from
44% to 63% (Fig. 5). This may be partially because DMSO effectively inhibited the progressive
transformation of the target furfural [30, 38]. DMSO has advantage of a high boiling point, making it
a useful solvent in industrial applications.
Increasing in the concentration of D-xylose from 5 to 10 g/L led to a remarkable decrease in the
furfural yield from 63% to 15%. It is known that the water negatively affects furfural formation
because water accelerates furfural degradation reactions [39]. Therefore, using water as the reaction
medium led to the low selectivity. The leading published researches on heterogeneous solid acid
catalyzed dehydration in aqueous media have wide initial D-xylose concentration ranges from 1 wt%
to 15 wt% [40, 41]. In this study, higher D-xylose concentrations caused a considerable decrease of
furfural selectivity. This issue could be solved by using ILs or supercritical CO2 as co-solvent [6-9],
but there are many theoretical and technical problems remain open, such as high cost of ILs and
product separation difficulty. In this study, lowering the reaction temperature and extending the
reaction time is one of the feasible ways. However, higher D-xylose concentration will result in an
increase yield of humins by-product [42], and definitely influence on the stability and activity of
Sn0.625Cs0.5PW. 5 g/L was selected as an acceptable D-xylose concentration in subsequent reactions.
10

It has been reported that the condensation of furfural and D-xylose causes a decrease of furfural
selectivity [30]. Furan and formic acid are the main by-products in D-xylose dehydration at high
temperature. Therefore, the addition of DMSO can effectively decrease the formation of by-products.
It should be noted that the addition of water is desirable for increasing the solubility of D-xylose [43].
In addition, water can act as an amphoteric substance capable of donating and accepting protons
during the dehydration process [44]. An appropriate amount of water is necessary to obtain a high
furfural yield. Moreover, the addition of organic solvent was beneficial to inhibit furfural yield-loss
reactions that lead to the polymerization of furfural and D-xylose.
To understand the effects of solvent on furfural selectivity, the conversion of D-xylose was
investigated in different solvents (Fig. 6). It was found that the furfural selectivity in solvent-H2O
mixtures decreased in the order of DMSO > butanol > ethanol > isopropanol > acetone. The solvents
used covered a wide range of polarity. Ketones (i.e., acetone) are very reactive and can react with
both D-xylose and furfural [45]; therefore, acetone gave the lowest yield of furfural (38%). The
furfural yield was greatly enhanced in alcohols/H2O mixtures. To a certain extent, alcohols can
prevent the polymerization of furfural by inhibiting aldol condensation [44]. DMSO achieved the
transformation of D-xylose into furfural with the highest selectivity of 63%. DMSO has the lowest
polarity and highest boiling point (189 °C) of the investigated solvent, and these properties would be
favorable for the dissolution of furfural and separation of products, respectively.
The reaction temperature also had a considerable effect on the degree of D-xylose dehydration. It
is generally known that furfural molecules in the vapor phase cannot undergo loss reactions [45]. The
influence of temperature on the dehydration reaction was studied in the range of 160–220 °C (Fig. 7).
At 160 °C for 3 h, a 26.5% yield of furfural was achieved. Furfural yield increased obviously as
temperature was increased from 160 °C to 200 °C, at which a maximum of 64% was achieved.
Further increase of temperature caused a decrease of furfural yield to 42%. The loss of furfural was
attributed to cross- and self-polymerization [46]. The resultant solid products are called insoluble
humins. Correspondingly, the color of the product varied from light yellow to brown, showing the
reaction of D-xylose and furfural to produce humins and other decomposition products. In this study,
an insoluble humins yield of 1.21 g/L was obtained at 220 °C. The increase in soluble humins yield
was higher than the decrease in furfural yield (1.1 g/L). The insoluble humins were deposited on the
outer surface of the solid catalyst, leading to a decrease in catalyst acidity [42]. Therefore, the
11

optimum reaction temperature of 200 oC was used in subsequent experiments.
In a review, different heterogeneous acid catalysts, especially mesoporous acid-catalysts, were
summarized to provide a general overview of heterogeneous catalysts studies [40]. Compared with
other heterogeneous acid catalysts with furfural yields of 10%–65% in organic solvent/H2O mixtures
in the above review [40], Sn0.625Cs0.5PW exhibited relatively high catalytic efficiency. Taking into
account the reaction media, the reaction in DMSO/H2O mixtures was less active than that in ionic
liquids. However, it should be noted that the DMSO/H2O system has a cost advantage over that of
ionic liquids.
3.6 Supported Sn0.625Cs0.5PW
A carrier to support Sn0.625Cs0.5PW was prepared by incomplete carbonization of chitosan via a
hydrothermal process at 220 °C for 8 h. The supported Sn0.625Cs0.5PW prepared by hydrothermal
deposition process was studied. The effects of Sn0.625Cs0.5PW addition on carbonization of chitosan
were determined by changing the Sn0.625Cs0.5PW loading, as shown in Fig. 8. The maximum furfural
yield of 62% was achieved using 16.7 wt % Sn0.625Cs0.5PW/CH. Increasing the Sn0.625Cs0.5PW
loading from 16.7 wt % to 50 wt % caused the furfural yield to drop from 62% to 28%. These results
indicated that 16.7 wt % Sn0.625Cs0.5PW/CH prossessed similar catalytic activity to that of
Sn0.625Cs0.5PW. Thus, 16.7 wt % Sn0.625Cs0.5PW/CH is active as a heterogeneous systems as well as
being recoverable and reusable.
3.7 Catalyst stability
Recycling of 16.7 wt % Sn0.625Cs0.5PW/CH was conducted in a DMSO/H2O mixture at 200 °C,
for 3 h with D-xylose/16.7 wt % Sn0.625Cs0.5PW/CH weight ratio of 1.25. After one batch reaction,
the solid Sn0.625Cs0.5PW/CH was separated from the mixture by centrifugation, washed with
deionized water, dried then used in the next dehydration process. Both D-xylose conversation rate
and furfural yield slightly decreased with an increasing number of recycling runs (Fig. 9). A furfural
yield of 63% was achieved for the first use of 16.7 wt Sn0.625Cs0.5PW/CH, which decreased to 57%
after the catalyst was recycled four times. After the 16.7 wt % Sn0.625Cs0.5PW/CH was reused in six
runs, the furfural yield was still 54.2%. The EDX pattern of 16.7 wt % Sn0.625Cs0.5PW/CH after five
cycling experiments is shown in Fig. S3. The atomic ratios of C and O increased considerably, which
can be explained by the adsorption and deposition of formed solid humins on the surface of
Sn0.625Cs0.5PW/CH. As a result, the atomic ratios of Sn4+ and Cs+ decreased accordingly. The cycling
12

experiments showed that the catalytic activity of the supported solid catalysts was stable. The good
stability of the catalyst is probably caused by its large surface area and pore volume, uniform
distribution of Sn0.625Cs0.5PW on CH, and resistance to leaching. The slight activity decrease was
possibly caused by the adsorption and deposition of humins.
This methodology was also suitable for the dehydration of xylose-rich liquors from biomass
pre-treatment. Renewable bio-mass, which contains rich cellulose and hemicellulose, are potential
starting materials for the manufacture of furans. The furfural yields obtained from xylan were usually
lower than those from D-xylose. The transformation of D-xylose to furfural depends on the real
concentration of D-xylose released during the biomass pre-treatment. One-pot process combining
excellent solubility (i.e., ionic liquids) and microwave-assisted heating will be attempted to improve
the furfural yield.
Conclusion
A series of bi-metallic salts of heteropoly acid was synthesized for the conversion of D-xylose to
furfural in a DMSO-H2O system. The introduction of Sn4+ into the structure of Cs0.5PW did not
change its Keggin structure, but caused the particle size to increase to 250 nm. Sn4+ occupied the
available surface sites of Cs+, which helped improve its catalytic performance in D-xylose
dehydration. The addition of DMSO effectively improved the furfural selectivity of the system. The
highest furfural yield of 63% with a conversion efficiency of 98% and furfural selectivity of 84.3%
was obtained after 3 h at 200 °C in a DMSO/H2O mixture with a 1:1 volume ratio. The supported
16.7 wt % Sn0.625Cs0.5PW/CH catalyst displayed similar catalytic activity to that of unsupported
Sn0.625Cs0.5PW and good stability after recycling six times.
Acknowledgement
The authors wish to acknowledge the financial support from Dalian University of Technology
(DUT15RC(3)114) and China National Natural Science Foundation (No: 31270620). We thank
Natasha Lundin, PhD, from Liwen Bianji, Edanz Editing China, for editing the English text of a draft
of this manuscript.
13

Reference
[1] O. Ellabban, H. Abu-Rub, F. Blaabjerg, Renew. Sust. Energ. Rev. 39 (2014) 748-764.
[2] J.J. Bozell, G.R. Petersen, Green Chem. 12 (2010) 539-554.
[3] X. Li, P. Jia, T. Wang, ACS Catal. 6 (2016) 7621-7640.
[4] I. Agirrezabal-Telleria, I. Gandarias, P.L.Arias, Catal. Today 234 (2014) 42-58.
[5] J.P. Lange, E. van der Heide, J. van Buijtenen, R. Price, ChemSusChem 5 (2012) 150-166.
[6] S. Peleteiro, G. Garrote, V. Santos, J.C. Parajó, Afinidad, 71 (2014) 202-206.
[7] S. Peleteiro, A.M. da Costa Lopes, G. Garrote, J.C. Parajꢁ, R. Bogel-Łukasik, Ind. Eng. Chem.
Res. 54 (2015) 8368-8373.
[8] Z. Fang, R.L.J. Smith, X. Qi (Eds.), Production of Platform Chemicals from Renewable
Resources, Springer, 2017, pp. 45-80.
[9] A.M. da Costa Lopes, R. Bogel‐Łukasik, Green Chem. 8 (2016) 2331-2334.
[10] E. Lam, E. Majid, A.C.W. Leung, J.H. Chong, K.A. Mahmoud, J.H.T. Luong, ChemSusChem 4
(2011) 535-541.
[11] S.M. Bruce, Z. Zong, A. Chatzidimitriou, L.E. Avci, J.Q. Bond, M.A. Carreon, S.G. J. Wettstein,
Mol. Catal. A-Chem 422 (2016) 18-22.
[12] W. Jeon, C. Ban, J.E. Kim, H.C. Woo, Kim D.H. J., Mol. Catal. A-Chem 423 (2016) 264-269.
[13] S. Kaiprommarat, S. Kongparakul, P. Reubroycharoen, G. Guan, C. Samart, Fuel 174 (2016)
189-196.
[14] Y. Zhu, W. Li, Y. Lu, T. Zhang, H. Jameel, H. Chang, L. Ma. RSC Adv. 7(2017) 29916-29924.
[15] M.J. Climent, A. Corma, S. Iborra, Chem. Rev. 111 (2011) 1072-1133.
[16] M.M. Heravi, Z. Faghihi, J. Iran. Chem. Soc. 11 (2014) 209-224.
[17] G. Raveendra, M. Srinivas, N. Pasha, A.V.P. Rao, P. S.S. Prasad, N. Lingaiah, React. Kinet.
Mech. Cat. 115 (2015) 663-678.
[18] G. Marcì, E.I. García-López, L. Palmisano, Eur. J. Inorg. Chem. 2014, 21-35.
[19] T. Okuhara, Chem. Rev. 102 (2002) 3641-3666.
[20] A.S. Dias, S. Lima, M. Pillinger, A.A. Valente, Carbohydr. Res. 341 (2006) 2946-2953.
[21] P.M. Rao, M.V. Landau, A. Wolfson, A.M. Shapira-Tchelet, M. Herskowitz, Microporous
Mesoporous Mat. 80 (2005) 43-55.
[22] A.M. Alsalme, P.V. Wiper, Y.Z. Khimyak, E.F. Kozhevnikova, I.V. Kozhevnikov, J. Catal. 276
14

(2010) 181-189.
[23] K.M. Parida, S. Rana, S. Mallick, D. Rath, J. Colloid Interface Sci. 350 (2010) 132-139.
[24] J. Geboers, S.V. Vyver, K. Carpentier, P. Jacobs, B. Sels, Green Chem. 13 (2011) 2167-2174.
[25] W. Alharbi, E. Brown, E.F. Kozhevnikova, I.V. Kozhevnikov, J. Catal. 319 (2014) 174-181.
[26] S. Tsubaki, K. Oono, T. Ueda, A. Onda, K. Yanagisawa, T. Mitani, J. Azuma, Bioresourc.
Technol. 144 (2013) 67-73.
[27] P.M. Rao, A. Wolfson, M.V. Landau, M. Herskowitz, Catal. Commun. 5 (2004) 327-331.
[28] M.A. Hanif, S. Nisar, U. Rashid, Catal. Rev. 59 (2017) 165-188.
[29] K. Shimizu, H. Furukawa, N. Kobayashi, Y. Itaya, A. Satsuma, Green Chem. 11 (2009)
1627-1632.
[30] F. Guo, Z. Fang, T. Zhou, Bioresourc. Technol. 112 (2012) 313-318.
[31] F. Guo, Z. Xiu, Z. Liang, Appl. Energ. 98 (2012) 47-52.
[32] K. Shimizu, K. Niimi, A. Satsuma, Appl. Catal. A-Gen. 349 (2008) 1-5.
[33] Q.Y. Liu, W.L. Wu, J. Wang, X.Q. Ren, Y.R. Wang, Microp. Mesopor. Mat. 76 (2004) 51-60.
[34] J.A. Dias, E.C. Sílvia, C.L. Dias, Microporous Mesoporous Mat. 76 (2004) 221-232.
[35] S. Wu, J. Wang, W. Zhang, X. Ren, Catal. Lett. 125 (2008) 308–314.
[36] K. Narasimharao，D.R. Brown，A.F. Lee，A.D. Newman，P.F. Siril，S.J. Tavener，K. Wilson，
J Catal. 248 ( 2007) 226-234
[37] N. Lingaiah, K. MohanReddy, N. SeshuBabu, K. NarasimhaRao, I. Suryanarayana, P.S.
SaiPrasad, Catal. Commun. 7 (2006) 245-250.
[38] X. Guo, C.H. Zhu, F. Guo, Bioresources 11 (2016) 2457-2469.
[39] V. Choudhary, S.I. Sandler, D.G. Vlachos, ACS Catal. 2 (2012) 2022-2028.
[40] I. Agirrezabal-Telleria, I. Gandarias, P.L. Arias, Catal. Today 234 (2014) 42-58.
[41] A.S. Dias, S. Lima, M. Pillinger, A.A. Valente, Catal. Lett. 114 (2007) 151-160.
[42] A.S. Dias, M. Pillinger, A.A. Valente, J. Catal. 229 (2005) 414-423.
[43] L. Zhang, H. Yu, P. Wang, Y. Li, Bioresource Technol. 151(2014)355-360.
[44] X. Hu, R.J.M. Westerhof, D. Dong, L. Wu, C.Z. Li, ACS Sustainable Chem. Eng., 11 (2014)
2562–2575.
[45] K.J. Zeitsch, Chem. Innovat. 30 (2000) 29-33.
[46] O. Yemiş, G. Mazza, Bioresource Technol. 102 (2011) 7371-7378.
15

Tables caption
Table 1 Effect of surficial Sn4+ and Cs+ attribution on furfural yield and D-xylose conversion.
Table 2 BET surface area and pore properties of the heteropolyacid salts
Figures captions
Figure 1. SEM images of heteropolyacid salts: (a) H2.5Cs0.5PW, (b) H0.5Cs2.5PW, (c) Sn0.5CsPW, (d)
Sn0.625Cs0.5PW, (e) Sn0.55Rb0.8PW, (f) CH, (g) 30% H0.5Cs2.5PW/CH, (h) Sn0.625Cs0.5PW/CH.
Figure 2. XRD pattern for (a) Sn0.625Cs0.5PW; (b) Sn0.5CsPW; (c) Sn0.55Rb0.8PW; (d) 16.7%
Sn0.625Cs0.5PW/CH.
Figure 3. Nitrogen adsorption–desorption isotherm of Sn0.625Cs0.5PW.
Figure 4. Dehydration of D-xylose catalyzed by different solid heteropoly salts: a, H2RbPW; b,
H0.5Cs2.5PW; c, HKCsPW; d, HNH4CsPW; e, H0.5Fe0.5PW; f, HNi0.5CsPW; g, Sn0.625Cs0.5PW; h:
H0.5Ce0.5CsPW; l, H2SO4; m, HCl; n, CH3COOH. Conditions: D-xylose 0.1 g, amount of catalyst
0.08 g, H2O 20 ml, 200 °C for 3 h.
Figure 5. Effect of initial concentration of D-xylose on dehydration: (a) in water (200 °C for 3.5 h);
(b) in DMSO/H2O mixture with 1:1 volume ratio (180 °C for 3 h).
Figure 6. Furfural selectivity (mol%) vs. solvent/H2O mixtures with volume ratio of 1:1 (D-xylose
0.1 g, Sn0.625Cs0.5PW 0.08 g, solvent 20 ml, 200 °C, 3h)
Figure 7. Effect of reaction temperature on dehydration in DMSO/H2O mixture with 1:1 volume
ratio (D-xylose 0.1 g, Sn0.625Cs0.5PW 0.08 g, DMSO/H2O 20 ml).
16

Figure 8. The effect of Sn0.625Cs0.5PW loading amount on D-xylose conversion and furfural yield.
(DMSO-H2O with 1:1 volume ratioratio 20 ml, D-xylose 0.1 g, Sn0.625Cs0.5PW 0.08 g, 200 °C, 3 h)
Fig. 9. Effect of recycling of 16.7 wt % Sn0.625Cs0.5PW/CH on D-xylose conversion and furfural
yield (D-xylose 0.1 g, Sn0.625Cs0.5PW 0.08 g, DMSO/H2O 20 ml, 200 °C, 3 h)
17

18

19

20

Table 1 Effect of surficial Sn4+ and Cs+ attribution on furfural yield and D-xylose conversion
21

Sn
Cs
Furfural yield
(mol %)
16.7
D-xylose conversion
Molecular formula
mol %
0.452
0.917
1.012
1.155
1.155
mol %
0.905
1.147
1.472
0.924
0.924
(mol %)
99
H1.5Sn0.25Cs0.5PW12O40
Sn0.5CsPW12O40
25.3
99
Sn0.55Cs0.8PW12O40
33.5
99
Sn0.625Cs0.5PW12O40
44.4
99
Sn0.625Cs0.5PW12O40
39.6
99
Error of measurement = ±3%.
Reaction conditions: D-xylose 0.2 g, Sn0.625Cs0.5PW 0.16 g, solvent 20 ml, 200 °C, 3.5 h, in water media.
Table 2 BET surface area and pore properties of the heteropolyacid salts
Surface area
(m2/g)a
Pore diameter
(nm)
> 48.0b
Pore volume
(cm3/g)
-
Ref.
H3PW
5
[34]
Cs3PW
156
140
219
19.4b
11c
9.6c
14.1
[34]
Sn0.625Cs0.5PW
16.7 wt% Sn0.625Cs0.5PW/CH
0.32
In this study
In this study
0.39
a BET equation.
b Estimated from XRD line width of peak ∼26°.
c BJH equation.
22