Aluminum doped solid acid with suitable ratio of Br?nsted and Lewis acid sites synthesized by electric-flocculation of phosphotungstic acid via hydrothermal treatment for producing 5-hydroxymethylfurfural from glucose

Article abstract of DOI:10.1016/j.apcata.2019.02.005

Novel solid acid catalyst PWAl-200 was synthesized by a new green method of electric-flocculation of phosphotungstic acid (PW) with aluminum as electrodes to transform glucose to 5-hydroxymethylfurfural (5-HMF). Through electric-flocculation, PW was deposited in the hydrated aluminum received from electrolysis. Then the obtained floc was treated by hydrothermal process to get solid acid. Characterization results showed that after electric-flocculation, the electron-withdrawing influence of terminal W[dbnd]O in the PW on hydrated aluminum formed Lewis (L) acid sites on these six-coordinated aluminum. The H⁺ derived from PW could supply Br?nsted (B) acid sites. In the further hydrothermal treatment, hydrated aluminum dehydrated to produce reformed six-coordinated aluminum. It was linked to heteropoly anions ligands through oxygen bridges, producing more L acid center. Besides, four-coordinated aluminum formed from the hydrated aluminum. The positive charge it produced increased L acid sites due to the strong effect of nearby tungsten species, further adjusting to a moderate ratio of Lewis/Br?nsted (L/B) on solid acid for catalytic synthesis 5-HMF from glucose. The highest yield was 61.7% at 170 ℃ for 4 h and the catalyst could be recycled for four times and tend to stabilize.

Full text of DOI:10.1016/j.apcata.2019.02.005

Accepted Manuscript
Title: Aluminum doped solid acid with suitable ratio of
Brønsted and Lewis acid sites synthesized by
Electric-flocculation of phosphotungstic acid via
hydrothermal treatment for producing
5-hydroxymethylfurfural from glucose
Authors: Xin Wang, Tao Lv, Minghui Wu, Junwei Sui, Qing
Liu, Huan Liu, Jiaojiao Huang, Lishan Jia
PII:
S0926-860X(19)30060-2
DOI:
Reference:
https://doi.org/10.1016/j.apcata.2019.02.005
APCATA 16975
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
24 October 2018
8 January 2019
4 February 2019
Please cite this article as: Wang X, Lv T, Wu M, Sui J, Liu Q, Liu H, Huang J, Jia
L, Aluminum doped solid acid with suitable ratio of Brønsted and Lewis acid sites
synthesizedbyElectric-flocculationofphosphotungsticacidviahydrothermaltreatment
for producing 5-hydroxymethylfurfural from glucose, Applied Catalysis A, General
(2019), https://doi.org/10.1016/j.apcata.2019.02.005
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.

Aluminum doped solid acid with suitable ratio of Brønsted and Lewis acid sites
synthesized by Electric-flocculation of phosphotungstic acid via hydrothermal
treatment for producing 5-hydroxymethylfurfural from glucose
Xin Wang, Tao Lv, Minghui Wu, Junwei Sui, Qing Liu, Huan Liu, Jiaojiao Huang,
Lishan Jia*
Department of Chemical and Biochemical Engineering, College of Chemistry and
Chemical Engineering, Xiamen University, Xiamen 361005, Fujian, China
*Corresponding author. Tel.: +86 0592 2188283; Fax: +86 0592 2184822.
E-mail address: jials@xmu.edu.cn
Graphical abstract
Highlights
 A novel solid acid catalyst PWAl-200 was prepared by Electric-flocculation

method.
 The introduction of aluminum to PW produced L acid promoting catalytic
reaction.
 The interaction between hydrated aluminum and W increased L acid and
adjusted L/B.
 The as-prepared catalyst had excellent performance for 5-HMF from glucose.
Abstract
Novel solid acid catalyst PWAl-200 was synthesized by a new green method of
electric-flocculation of phosphotungstic acid (PW) with aluminum as electrodes to
transform
glucose
to
5-hydroxymethylfurfural
(5-HMF).
Through
electric-flocculation, PW was deposited in the hydrated aluminum received from
electrolysis. Then the obtained floc was treated by hydrothermal process to get solid
acid. Characterization results showed that after electric-flocculation, the
electron-withdrawing influence of terminal W=O in the PW on hydrated aluminum
formed Lewis (L) acid sites on these six-coordinated aluminum. The H+ derived from
PW could supply Brønsted (B) acid sites. In the further hydrothermal treatment,
hydrated aluminum dehydrated to produce reformed six-coordinated aluminum. It was
linked to heteropoly anions ligands through oxygen bridges, producing more L acid
center. Besides, four-coordinated aluminum formed from the hydrated aluminum. The
positive charge it produced increased L acid sites due to the strong effect of nearby
tungsten species, further adjusting to a moderate ratio of Lewis/Brønsted (L/B) on
solid acid for catalytic synthesis 5-HMF from glucose. The highest yield was 61.7%

at 170℃ for 4 hours and the catalyst could be recycled for four times and tend to
stabilize.
Key words: aluminum electrode, phosphotungstic acid, 5-hydroxymethylfurfural,
ratio of B/L
1. Introduction
Nowadays, it is out of balance between increasing demand for resources and the
rapid consumption of fossil fuel reserves such as coal, oil, and natural gas that could
not be renewable and have harmful threat to environment [1-3]. In this respect,
development and utilization of sustainable resources draw more attention [4, 5]. Until
now, biomass has been regarded as a promising resource to produce sustainable
biofuel and valuable chemicals as it is abundant and renewable [6].
5-Hydroxymethylfurfural (5-HMF), a versatile intermediate from biomass, is a key
platform chemical [7, 8] owing to its high chemical reactivity for further desired
products like biofuel and value-added chemicals [9]. It has been found that the
monomers of bio-based polymers (2, 5-diformylfuran and 2, 5-furandicarboxylic acid),
energy products like 2, 5-dimethylfuran and levulinate acid could be obtained by
oxidation, reduction and rehydration reaction of 5-HMF [10, 11]. As the most
abundant source of biomass, the transformation of cellulose monomer (glucose) into
5-HMF has become one of the core technology for catalytic conversion of natural
biomass into chemicals. Hence, an efficient catalyst to increase 5-HMF yield means to
minimize side reactions that lead to humins and soluble or insoluble polymers.
The catalysts for producing 5-HMF involve homogeneous and heterogeneous

catalysts. Homogeneous catalysts like mineral acids and metal salts suffer from
corrosion and difficulty of separation and regeneration [1, 12]. Heterogeneous solid
acid as an environmentally benign alternative for biomass conversion has become a
research hotspot, which is attributed to its good catalytic performance, easy separation,
regeneration, and relatively low process cost. Glucose converting to 5-HMF catalyzed
by solid acid needs two stages. First, glucose isomerizes to fructose, which is a rate
determining step and requires suitable Lewis (L) acid for more efficient performance.
Second, the fructose further dehydrates to 5-HMF, which is catalyzed by Brønsted (B)
acid. In this case, some studies has adjusted effectively the L and B acid sites on the
surface of the zeolite by the metal oxide load and metal ion exchange to improve the
yield of the 5-HMF obtained with 27.5% and 42%, respectively [13, 14]. The
mesoporous tantalum phosphate catalyst is synthetized with presence of both B and L
acid sites to enhance 5-HMF yield and receives 32.8% [15]. These adjustments do not
always give a high yield. Therefore, designing L acid and B acid sites on the surface
of catalyst has always been a concern, which requires the synergistic effect of the
acidic (L/B acid) center. For this reason, the heteropoly acid such as H3PW12O40 could
be an excellent candidate owing to its well-defined structure and B acidity. The
introduction of metal cation could adjust and enhance its acidity. Several metal ions
are introduced into the PW to form solid acid as well as improve its acidity. A solid
heteropolyacid Cs2.5H0.5PW12O40 catalyst is applied for converting fructose to 5-HMF
with a 74% yield [16]. Besides, 93% and 86% yield of 5-HMF are received from
fructose by using CePW12O40 and Cu1.5PW12O40, respectively [12]. These

modifications enhance the acidity, which is beneficial for fructose converting.
However, they could not be suitable for conversion of glucose. Only 0.7% 5-HMF
yield is obtained from glucose while 81.1% is received from fructose over
Cs2.3H0.7PW12O40 catalyst [10]. Moreover, the presence of Al3+ provides some Lewis
acidity due to its own electron deficient properties and easy formation of weak ligand
structure. There are numerous studies preparing Al-doping PW by the partial
exchange of protons in H3PW12O40. Replacing hydrogen with aluminum ion in
AlPW12O40 catalyst introduces the L acid sites, which is applied in the esterification,
ring opening and tosylation reaction [17, 18]. But, the fact that Al3+ ion hydrolyzes to
Al(OH)3 in aqueous hampers its application in H2O systems [19]. Therefore, we
design a novel Al-modified phosphotungstic acid (PW) catalyst using electric
flocculation and hydrothermal treatment method. Reformed PW could coordinate with
alumium species and hydrated aluminum substitute partial position of H+ in PW,
which could solve this problem and also adjust the ratio of L and B acid.
In this paper, we report the synthesis of the novel Al-modified phosphotungstic
acid using electric flocculation and hydrothermal treatment method and its activity to
catalyze glucose converting to 5-HMF. In the electrolysis process, the four PW
+
monomer units connected by H5O2 are destroyed to shape monomer molecules which
flocculate with hydrated aluminum. In flocculation process, the electron-withdrawing
influence of W=O in PW on hydrated aluminum leads to formation of L acid sites on
the six-coordinated aluminum. During the further hydrothermal treatment, hydrated
aluminum dehydrates and connects with the heteropoly anion to form more L acid

sites. Besides, dehydrated hydrated aluminum forms positive charge on the
four-coordinated aluminum due to the nearby tungsten species, further increasing L
acid sites. The adjustment of the ratio of L acid and B acid shows good catalytic
performance in converting glucose to 5-HMF.
2. Experiment
2.1. Catalyst preparation
2.1.1. Preparation of PWAl-200 and PWAl
(1). The sheet aluminum metal (99.999%) was cut into a sample of 30 mm × 30
mm × 1 mm in size and cleaned surface with dilute acid. One side of the Al electrode
was immersed into electrolyte and another side was connected with an experimental
power supply (Lodestar DC.POWER supply LP3005D) via a slide rheostat using a
copper wire encapsulated with epoxy resin. 1.0 g phosphotungstic acid hydrate (AR)
(Aladdin Industrial Corporation) dissolved into 80 ml deionized water obtained the
solution as electrolyte. In the course of the experiment, the current kept at 20 mA, and
the electric voltage increased gradually to 29.5V. Then, end the electrolysis until the
current tended to 0 A. The process lasted about 6 hours. Furthermore, the different
electro-flocculation time was set at 40, 45, 50, 60 and 120 min. respectively, and the
obtained samples were used to investigate the changes in the structure of Al modified
heteropoly acids with time.
(2). After stirring for 5 min. to mix evenly, the flocculating product was
transferred into a Teflon-lined stainless autoclave reactor and heated to 200℃ for 12 h.
After filtered, the sample was dried in the oven at 80℃ and named PWAl-200. The

sample which directly dried without hydrothermal treatment was named PWAl.
2.1.2. Prepration of AlOOH-PW
NaCl solution (2 mol/L) replaced the PW solution as the electrolyte. The
flocculation obtained through the electric flocculation process was filtered as well as
dried and named AlOOH. Then, PW and AlOOH mixed by stirring for 5 min. and
transferred into a Teflon-lined stainless autoclave reactor and heated to 200℃ for 12 h.
After filtered, the sample was dried in the oven at 80℃and named AlOOH-PW.
2.1.3. Prepration of PWAl-C
An appropriate amount of metal salt (0.1 mol/L) aqueous solution was added into
10 mL phosphotungstic acid (0.08 mol/L) and stirred 2 hours at room temperature.
Then it was dried at 80℃ overnight and calcined at 200℃ for 4 hours. The catalyst
was named PWAl-C.
2.2. Catalytic test
In a typical reaction procedure, 0.2 g glucose (Sinopharm Chemical Reagent Co.
Ltd) and 0.05 g catalyst were poured into a Teflon-lined stainless steel autoclave
reactor, and then added into 20 ml solvent (12 ml DMSO (Sinopharm Chemical
Reagent Co. Ltd) and 8 ml water). The reactor was heated in an oven at a certain
temperature. After reaction, the reactor was cooled to room temperature and the solid
residue was separated by filtration. The liquid product was collected for analysis. The
solid residue was washed by water and ethanol several times. The catalyst was reused
after drying at 80℃.
2.3. Characterization of catalyst

Powder X-ray diffraction (XRD) patterns were determined using a PANalytical
X’pert PRO X-ray diffractometer, which operated at 40 kV and 30 mA, with Cu Kα as
a radiation source in the range of 2θ = 5° to 90° at a scanning rate of 10° min-1.
Scanning electron microscope (SEM) and the EDS energy spectrum was performed
using Zeiss Sigma with an accelerating voltage of 15 kV. The element distribution
was carried out by using high resolution transmission electron microscopy (HRTEM),
performed on a Philip Analytical FEI Tecnai 30 electron microscope at an accelerating
voltage of 300ꢀkV. The Brunauer-Emmett-Teller (BET) surface area and BJH pore
sizes were measured by high purity nitrogen gas adsorption in a Micro-meritics
TristarⅡ3020 instrument and samples were pretreated at a condition of outgassing at
150℃ for 4 h. The Fourier transform infrared (FT-IR) spectra were performed using
Nicolet Avatar 380 Fourier transform spectrometer with a 4 cm-1 resolution from 400
cm-1 to 4000 cm-1. The Pyridine-adsorbed FT-IR spectra were recorded using Thermo
Scientific Nicolet iS 50 with a ZDF-5227B composite vacuum system. Before
pyridine adsorption, the sample prepared as a thin self-supporting wafer was heated to
150℃ and degassed. Then pyridine was introduced to absorb for 5 min, followed by
removing physisorbed pyridine. The resulting pyridine-adsorbed sample was
subjected to desorption at 150℃, which was in situ recorded with FT-IR. The Raman
shift was measured on an Xplora spectrometer (Horiba, Japan) from 200 cm-1 to 1000
cm-1. X-ray photoelectron spectroscopy (XPS) analysis was tested with brand PHI
Quantum 2000 Scanning ESCA microprobe consisting of a monochromatized
microfocused Al X-ray source. UV-visible diffused reflectance spectra (UV-vis-DRS)

was recorded using a brand Varian Carry 5000 UV-vis spectrophotometer from 200
31
27
nm to 800 nm with basement of BaSO4. The P, Al magic angle spinning-nuclear
magnetic resonance (MAS NMR) spectra were recorded with Bruker 400 MHz
Avance III SS-NMR equipment, working at a frequency of 121.4 MHz and 104.3MHz
31
27
for P and Al at room temperature. The spin rate was 12 kHz using 4 mm MAS
rotors. The TG patterns were determined using TG-209F1 in the condition of N2 in the
range of 35℃ to 700℃.
2.4. Analysis of production
The concentration of 5-HMF was measured at 284 nm using Agilent 1260 HPLC
equipped with a ZORBAX 300-SCX column and a MWD detector. Glucose
concentration and fructose yield were determined by Agilent 1260 HPLC with a RID
detector using ZORBAX NH2 column. Levulinate acid (LA) was detected by Gas
Chromatography (Agilent 6890N, CA)-Mass Spectrometry (Agilent 5973, CA) using
a capillary column. The 5-HMF yield, glucose conversion, and selectivity were
calculated according to the following equations:
moles of produced 5−HMF
5-HMF yield (mol%) =
Glucose conversion (mol%)
× 100%
moles of initial glucose
(moles of initial glucose − moles of remaining glucose)
moles of initial glucose
=
× 100%
moles of produced 5−HMF
5-HMF selectivity (mol%) =
× 100%
(moles of initial glucose − moles of remaining glucose)
moles of produced fructose
moles of initial glucose
Fructose yield (mol%) =
× 100%
moles of produced LA
moles of initial glucose
LA yield (mol%) =
× 100%
Humin and other by-products yield (mol%) =

Glucose conversion − 5 − HMF yield − Frutose yield − LA yield
3. Result and discussion
3.1. Characterization of catalysts
3.1.1. FT-IR
FT-IR spectroscopy was measured to analyze the structure properties of different
samples in the range of 400-4000 cm-1. As could be seen in Fig. 1, the spectrum
showed a wide absorption band in the range of 3300-3700 cm-1, which was related to
the tensile -OH vibration. The absorption band in the 1500-1700 cm-1 region was
originated from the bending vibration of physically adsorbed water molecules.
Several typical absorption bands of Keggin structure in which a PO4 tetrahedron
surrounded by four W3O13 groups with an edge-sharing octahedron could be observed.
The band at 1080 cm-1, 982 cm-1, 887 cm-1 and 805 cm-1 with hard intensity in PW
indicated the P-Oa equivalent mode in the free PO4, W=Od (terminal oxygen bond)
stretch mode, W-Ob-W (inter bridges between corner-sharing WO6 octahedra) stretch
mode and W-Oc-W (intra bridges between edge-sharing WO6 octahedra) stretch mode,
respectively [20, 21]. The band at 1080 cm-1, 949 cm-1, 887 cm-1 and 805 cm-1 in the
PWAl-200 and PWAl smaples indicated the Keggin structure basically remained.
However, the band at 982 cm-1 shifted to a low wavelength 949 cm-1 and the intensity
of these main peaks decreased, indicating the interaction of terminal W=O bond in the
reformed PW and hydrated aluminum (Al3+(H2O)6 in PWAl and Al3+(H2O)3 in
PWAl-200 ). This was similar to the effect in Zr-bonded sulfate groups [22]. In
AlOOH-PW, only the band at 1080 cm-1 remained while some peaks in the range of

500-750 cm-1 belonging to AlOOH also existed, indicating that Keggin structure was
destroyed. PW fragment and the AlOOH physically mixed without electrolysis
process.
Fig. 1. FT-IR patterns of PWAl-200, PWAl, PW, AlOOH-PW and AlOOH
3.1.2. Raman
The Raman analysis of PW, PWAl and PWAl-200 samples were shown in Fig. 2.
It could be observed that 1011 and 990 cm-1 in the PW were originated from the
symmetric and asymmetric stretching modes of W=O, respectively. However, the
width of the peak at 974 cm-1 was broadened in PWAl-200 and PWAl samples. A shift
to low wavelength compared with the pure PW indicated the weakening of W=O
bonds due to interactions with hydrated aluminum, which was similar to the FT-IR
characterization [23]. Besides, board shoulders at 893 and 925 cm-1 were
corresponded to the stretching mode of W-Ob-W bond while the W-Ocentral stretching
peaked at 219 cm-1 [21], which was almost invariant. The band at 540 cm-1 was

attributed to the symmetric stretching mode of W-Oc-W and it shifted to a low
wavelength at 516 cm-1 after introducing hydrated aluminum [24, 25]. The band at
348 cm-1 in PWAl-200 gave a shift compared with 360 cm-1 in PWAl due to modified
six-coordinated or four-coordinated aluminum in the hydrothermal process as shown
in NMR analysis.
Fig. 2. Raman patterns of PW, PWAl and PWAl-200
3.1.3. 31P and 27Al MAS NMR
31P MAS NMR was an efficient method to study the P state in PW. It could be
observed the 31P MAS NMR of PW, PWAl and PWAl-200 samples from Fig. 3a.
There was a sharp peak at -15 ppm in pure PW. The W=O terminal oxygens were
more accessible to water molecules and linked four neighbouring heteropoly anions to
+
form H5O2 by hydrogen bonds [21]. A peak at -10.6 ppm in PWAl-200 and PWAl
indicated the presence of monomer PW, in which 3 H+ protons directly attached to
+
polyanions instead of the existence mode of H5O2 cation [26, 27]. A signal at -8 ppm

in both PWAl and PWAl-200 implied the presence of partial fragmentation of the
7-
Keggin units to produce the lacunary heteropoly anions PW11O39 species [23, 28].
However, the main resonance turned to -12.9 ppm in PWAl-200 indicated that the
strong interaction such as electrostatic attraction made the partly dehydrated
aluminum species connected with the heteropoly anion ligand to form
[Al3+(OH2)3]2[HPW12O402-]3 species, in which one isolated acidic proton directly
+
+
attach to polyanions [29] such as [Ti-OH2 ]2[HPW12O402-] and [≡Zr-OH2]n
[H3−nW12PO40]n −3 [23, 28, 30]. In this way, less electron density was on this reformed
six-coordinated aluminium leading to more L acid sites. The added width of the peak
in PWAl and PWAl-200 was also attributed to the Keggin units interacting with
hydrated aluminum [23, 31]. The peak at -4ppm and -8 ppm in AlOOH-PW implied
that aluminum species adsorbed phosphorus species in fragmented Keggin units [28].
The 27Al at Fig. 3b showed a main peak at 0 ppm in PWAl which was assigned to
the six-coordinated aluminum like the hydrated aluminum (Al3+(H2O)6) [32]. In the
PWAl-200, the peak shifting to a low field at 4 ppm also proved that the present of
[Al3+ (OH2)3]2[HPW12O402-]3 could induce more electron-withdrawing effect of W=O
to the reformed six-coordinated aluminum and thus provide with more L acid sites.
The peak at about 73 ppm in PWAl-200 was attributed to the four-coordinated
aluminum which shifted to the lower field [33]. The reason of this phenomenon might
be electron-withdrawing effect of the surrounding tungsten species which also
improved the symmetry and the sharpness of this peak. However, there was no
four-coordinated aluminum in the PWAl. This phenomenon combined with the result

of UV-vis implied that the four-coordinated aluminum might coexist and gather with
the tetrahedral coordinated tungsten forming a -W-O-Al-O-W- arrangement. The
surrounding tungsten species gave strong electron-attracting influence on these Al-O
bonds, forming positive charge as L acid sites on the four-coordinated aluminum as
shown in Scheme S1. In AlOOH-PW, the peak at 3 ppm showed there was only
six-coordinated aluminum without electrolysis.
Fig. 3. (a) 31P patterns of PW, PWAl, PWAl-200 and AlOOH-PW and (b) 27Al

patterns of PWAl-200, PWAl and AlOOH-PW
3.1.4. UV-vis
The UV–vis DRS spectra of catalysts were shown in Fig. 4. The absorbance
bands were originated from the charge-transfer (O 2p-W 5d) of ligand to metal [34].
The required energy depended primarily on the coordination structure and preparation
procedure. In the pure PW, the band at 260 nm referred to the octahedral tungsten in
the Keggin structure and the band at about 350 nm was attributed to the crystallines
WO3 species which came from the secondary structure of PW [35]. The PWAl
without hydrothermal treatment remained the octahedral tungsten while the crystalline
WO3 disappeared, which was due to the interaction between terminal W=O with
hydrated aluminum. However, partly octahedral tungsten remained and there was a
new band at 220 nm in the PWAl-200 catalyst, which assigned to the tetrahedral
tungsten species [34]. These unsaturated tungsten species had the capacity to attract
electrons from the nearby Al-O bond in the four-coordinated aluminum, where
positive charge formed and L acid sites further increased. The AlOOH-PW remained
the peak at 260nm in PW and the peak at 208 nm in AlOOH.

Fig. 4. UV-vis DRS patterns of PW, PWAl, PWAl-200, AlOOH-PW and AlOOH
3.1.5. XRD
As shown in Fig. 5a, XRD patterns of electro-flocculation with different time
were displayed. It could be found that the peaks at the range from 20° to 40°gradually
weaken as the electrolytic time prolonged, which disappeared and a broad peak
completely formed when the time increased to 2 h. This phenomenon explained that
hydrated aluminum formed by electrolyzing. They could connect with PW and enter
among the Keggin monomers resulting in breaching the long-range order of PW as
the time prolonged. Thus, it emerged the trending of forming an amorphous structure.
Patterns of PW, PWAl and PWAl-200 were shown in Fig. 5b, the main
characteristic lines of pure PW were at 2θ= 6.62° and 26.47° with sharp peaks.
However, the broad peaks of samples formed after electro-flocculation, indicating that
the samples emerged an amorphous state. The peak shapes before and after
hydrothermal treatment were basically unchanged. The peaks of AlOOH remained in

AlOOH-PW, which implied that PW might deposite on the surface of AlOOH rather
than interact with it. This demonstrated that the hydrated aluminum could not interact
or enter into PW without electric flocculation. This might because electric
flocculation process made the PW molecules unstable owing to reduced tungsten with
negative electricity, which promoted the interaction between PW and aluminum
species.
Fig. 5. XRD diffraction patterns of (a) PWAl-200 of 40min, 45min, 50min, 1h, 2h and

(b) PW, PWAl, PWAl-200, AlOOH-PW and AlOOH
3.1.6. XPS
XPS was employed to determine the tungsten oxidation state. Fig. 6a showed the
evolution of W 4f spectras. The W 4f7/2 and W 4f5/2 binding energy at around 35.9 eV
and 37.9 eV in the PW were characteristic of W (VI). The peaks shifted to 35.5 eV
and 37.5 eV after electro-flocculation, respectively [35], indicating that there was a
raise of electron cloud density around the W (VI) because it could attract the electrons
from hydrated aluminum and thus positive charge formed on these six-coordinated
aluminum producing L acid sites. This matched with the result of FT-IR. However, the
peaks at 35.6 eV and 37.6 eV in PWAl-200 basically kept almost the same with PWAl.
This meant that although coordination number of tungsten was reduced, the electron
cloud of Al=O bond could be inclined to tungsten, which compensated the electrons
decrease caused by its unsaturated coordination. The result was fit with the 27Al MAS
NMR, which made the four-coordinated aluminum positively charged and further
increased the L acid sites.
Fig. 6b showed the evolution of Al 2p spectra. The main peak in PWAl (74.5 eV)
shifted to high binding energy (74.7 eV) after hydrothermal treatment. This
phenomenon demonstrated that PWAl-200 sample had lower electron density
compared with PWAl sample, which also explained the more L acid sites. This was
because some electrons of new-formed four-coordinated aluminum were attracted by
surrounding tungsten species.

Fig. 6. (a) W 4f XPS patterns of PW, PWAl and PWAl-200 and (b) Al 2p XPS patterns
of PWAl and PWAl-200
3.1.7. Morphology and physicochemical properties
The SEM images of PW, PWAl and PWAl-200 were shown in Fig. 7. It could be
observed that PW was a bulk aggregation. PWAl sample without hydrothermal
treatment appeared some small pellet. However, the PWAl-200 sample presented a

small fragment aggregating to form flower-like structure, which was related to the
formation of amorphous structure by the introduction of the hydrated aluminum.
Hydrothermal treatment changed the aggregation state of the sample particles which
exposed more active sites that could be easier to touch and react with glucose.
Fig. 7. SEM patterns of PW(a), PWAl(b), PWAl-200(c,d,e) and EDS of the
PWAl-200(f)
The element distribution in Fig. 8 reveals that Al, O, P, W elements were
well-distributed, which indicated that the aluminum ion entered the PW with a
homogeneously distribution. Al element existed on the AlOOH as could be seen in
Fig.S1. Nervertheless, P and W elements distributing in the whole space implied that
PW precipitated on the AlOOH instead of interacting with it.

Fig. 8. TEM Mapping of O (b), Al (c), P (d), W-L (e) and W-M (f) in PWAl-200
Table S1 summarized BET surface areas, pore volumes and average pore
diameters of PW, PWAl, PWAl-200 and AlOOH-PW. PW showed small BET surface
area and pore volume. The pore volume and diameter of PWAl increased due to the
large crystals breaking and the hydrate lost water produced pores after the
introduction of aluminum ion [36]. The pore volume was further increased with the
dehydration effect by hydrothermal treatment while the diameter of pore decreased,
which contribute to enhancing the surface area. This result was consistent with SEM.
In comparison, AlOOH-PW sample had large pore diameter while the pore volume
and surface area were really small. Therefore, the contact area between reactant and
catalyst was small, which was adverse to reaction.
The ratio of P/Al in PWAl was almost the same as it was in PWAl-200 as
displayed in Table 1. As for the sample without electrolysis, the ratio of P/Al was

much higher. This might because a small number of PW were absorbed by aluminum
species while some PW dissolved.
3.1.8. Acidity
As could be seen in Fig. 9, the acid property was analyzed by pyridine absorption.
The peaks at 1437 cm-1were assigned to the vibrations characteristic of pyridine
absorption on L acid sites while the B acid signal was located at the band of 1577 cm-1.
The peak at 1485 cm-1 indicated the presence of both B and L acid sites [37, 38]. Pure
PW was mainly a kind of B acid [39], which catalyzed the dehydration of fructose. In
PWAl, the terminal W=O in PW had electron-absorption effect on hydrated aluminum
forming L acid sites on these six-coordinated aluminum [22]. It was beyond doubt to
be beneficial for the reaction because the isomorization process which catalyzed by L
acid was the rate-determining step [40].The ratio of L/B was shown in Table 1. L acid
sites in PWAl-200 were much more than the PWAl. There were two aspects that might
enhance the L acid sites. On the one hand, L acid sites produced on the reformed
six-coordinated aluminum. On the other hand, the positive charge on four-coordinated
aluminum also increased the L acid sites due to the impacts of nearby tungsten species.
Besides, without the electro-flocculation process of PW, there was a great deal of L
acid sites produced in AlOOH-PW due to plenty of aluminum species. The ratio of
L/B acid was 1.58 and excessive L acid sites promoted the side reactions. Moreover,
the ratio of L/B in PWAl-C sample was 1.07, which was a little higher than PWAl-200.
The activity results showed that the electroflocculation method gave a more suitable
acid property than the conventional method.

Table 1. P/Al and acid site analysis of PW, PWAl, PWAl-200, AlOOH-PW and
PWAl-C
Catalyst
P/Al a
PW
PWAl
0.113
0.61
PWAl-200
0.1
AlOOH-PW
0.018
PWAl-C
0.128
1.07
N/A
0.017
b
CL/CB
0.85
1.58
a The ratio of P and Al
b
CL/CB was calculated based on the integrated areas of the absorption bands at
1577 cm-1 (B acid) and 1437 cm−1 (L acid) by Lambert–Beer equation A=ε▪ W▪ C / S
[39]，where A is the absorbance (area in cm−1), ε is the extinction coefficient(εB = 1.67
cm μmol−1 and εL = 2.22 cm μmol−1) [36], W is the sample weight(kg), c is the
concentration of acid (mol/kg or mmol/g) and S is the sample disk area (m2),
respectively.
Fig. 9. Py-IR patterns of PW-200, PWAl, PW, PWAl-C and AlOOH-PW
3.1.8. TG analysis

The thermal stability of PWAl-200 and PWAl catalysts were compared by TG
characterization as shown in Fig.10. Mass loss below 125℃ were 3.7% and 4.4% for
PWAl-200 and PWAl, respectively, which was due to the evaporation of physically
adsorbed water. Mass loss between 350℃ and 450℃ might be owing to the
decomposition of phosphatotungstic acid. Throughout the whole heating process, the
mass loss of PWAl-200 under the same temperature was less than that of PWAl. This
was probably due to the dehydration effect during the hydrothermal process.
Moreover, there was only a slight loss below 17ꢀ0°C, indicating that the as-prepared
catalyst had better thermal stability under the reaction temperatures [41, 42].
Fig. 10. TG analysis of PWAl and PWAl-200
Combined with the above characterization analysis, the possible catalyst
formation process was as follows: Through electroflocculation, the Al anode lost
electrons to form hydrated aluminum (Al3+ (H2O)6). Besides, the four keggin structure
+
units connected by H5O2 (Scheme S1a) were destroyed to form monomer molecules

(Scheme S1b). The monomer molecules obtained electrons in cathode giving a dark
blue turbid solution. The terminal W=O of these monomer molecules absorbed the
hydrated aluminum to form floc as shown in Scheme S1c. Meanwhile, partial
fragmentation from Keggin structure formed the lacunary heteropoly anions
7-
PW11O39 with a few W-O groups dropping out (Scheme S1d). Duing further
hydrothermal treatment, the hydrated aluminum gradually dehydrated and interacted
with terminal W=O of heteropoly anions due to the electron-attracting effect, which
made the hydrated aluminum substitute partial position of H+ as the charge
2-
compensation of Keggin polyoxo anion HPW12O40 forming a renormalized
six-coordinate aluminium [Al3+(OH2)3]2[HPW12O402-]3. Besides, the falling W-O
fragment might gather with some dehydrated aluminum owing to the
electron-attracting effect of tungsten species to form a -W-O-Al-O-W- arrangement
with four-coordinated aluminum and tungsten species shown in Scheme S1e. In
addition, the PW11O397-, monomer PW, [Al3+(OH2)3]2[HPW12O402-]3 and the
-W-O-Al-O-W- structure were connected by H-bond to form an amorphous state.
However, without electrolysis, a small number of PW were absorbed by aluminum
species while some PW dissolved. Comparing with the characteristics of AlOOH-PW,
electrolysis process made the PW molecules unstable and promoted the interaction
between PW and aluminum species.
3.2. Catalytic study
3.2.1. Screening of catalysts for conversion of glucose to 5-HMF
Glucose converting to 5-HMF catalyzed by PW, PWAl, PWAl-200, AlOOH-PW

and PWAl-C catalysts were operated at 170℃ for 4 hours. The results of 5-HMF
yields were concluded in Fig. 11. Pure PW showed 25.3% 5-HMF yield and 72%
glucose conversion. Although samples added aluminum all had higher glucose
conversion between 82% and 95%, PWAl, AlOOH-PW and PWAl-C exhibited
5-HMF yield between 32% and 45%，with 5-HMF selectivity between 39% and 48%.
PWAl-200 produced highest 5-HMF yield of 61.7%, with 5-HMF selectivity of 64.9%.
Fructose was not detected in the presence of PW. For other catalysts, there were also
low yield of fructose between 1.1% and 3.0%. This was because glucose isomerising
to fructose was the rate-determining step for this reaction. In addition, there was a
small number of LA (1%-4.3%) generating. Others were some humins and
by-droducts.
Fig. 11. Yield of products and conversion of glucose with different catalysts (reaction
conditions: glucose 0.2 g, catalyst 0.05 g, solvent 20 ml ( DMSO : water=3 : 2), 170℃
for 4 h )

Fig. 12. (a) Effect of reaction temperature on 5-HMF yield (reaction conditions:
glucose 0.2 g, PWAl-200 catalyst 0.05 g, solvent 20 ml (DMSO : water=3 : 2), 4
hours ) and (b) effect of reaction time on 5-HMF yield by PWAl-200 catalysts
(reaction conditions: glucose 0.2 g, PWAl-200 catalyst 0.05 g, solvent 20 ml (DMSO :
water=3 : 2), 170℃ )
3.2.2. Effect of reaction temperature and time
The yield and selectivity of 5-HMF showed a trend of climbing up and then

declining in the reaction temperature ranges from 150 to 190℃ (see Fig. 12a). The
highest yield with 61.7% was at 170℃for 4hours. After reacting at 150℃, only 12.7%
yield of 5-HMF with around 50% glucose conversion. Glucose conversion and yield
of 5-HMF could reach 88% and 35% at a higher temperature of 160℃, respectively. It
could be found that molecules gained more energy to sport and combine with the acid
sites as temperature rose. However, the side reaction of 5-HMF as an intermediate like
dehydration increased when the temperature was higher than 170℃. The yield of
5-HMF had a reduction of 38.3% at 180℃ . Lots of black humins were formed as
temperature increased to 190℃ while the yield of 5-HMF was only 17.2% with
complete conversion of glucose.
Fig. 12b showed the effect of reaction time on yield of 5-HMF. It revealed a rise
with the lapse of the reaction time, 13.7%, 59.2% and 61.7% yields with 52% to 97%
glucose conversions were given as reaction duration gradually extended at 170℃for 2
h, 3 h and 4 h, respectively. The yield of 5-HMF gradually depressed from 53.9% to
50.3% with almost total conversions and the side reaction in which humins were
produced increaced when the reaction time continued prolonging from 5 h to 6 h.
Therefore, the reusable study was selected at 170℃for 4 hours.
3.2.3. Mechanism of glucose converting to 5-HMF reaction
The above results showed that the aluminum modified PW solid catalyst
PWAl-200 obtained by electrocoagulation and hydrothermal treatment had excellent
catalytic performance. The possible catalytic mechanisms (Scheme S2) of the
prepared catalysts for the synthesis of 5-HMF from glucose could be summarized as

follows: the L acid center, deriving from the coordinated aluminum in Al-modified
PW, could promote glucose isomerization to fructose [39, 43]. Then, the H+ stemming
from Keggin structure of PW offered the B acid center which catalyzed dehydration
of fructose for 5-HMF. The increased L acid sites promoted the adjustment of the ratio
of L acid to B acid and their synergism, which was beneficial for the isomerization
and dehydration process to synthesize 5-HMF.
3.3. Comparison with other catalyst
Several reports of converting glucose to 5-HMF under optimal conditions were
listed in Table.2 comparing with this work. The HPA/TiO2–ZrO2 catalyst obtained
considerable yield of 47.6% at 160℃ for 4 hours using 26.5% NaCl (aq) with
tetrahydrofuran (THF) as solvent. When Sn-beta catalyst was applied in this reaction,
13.5% and 12.5% yield of 5-HMF was obtained by using 1-butanol with NaCl (aq)
and DMSO with water, respectively. In addition, 53% yield could be achieved at 140℃
for 10 hours by employing P-BnNH3Cl. The CSZA-3 catalyst produced 47.6%
5-HMF at 130℃for 15 hours. However, these catalysts needed a long reaction time.
The Sn-CP catalyst exploited with liquid acid HCl as a co-catalyst gave a high yield
of 63.9% in oil bath, which was unavoidable for defects of the liquid acids. H-ZSM-5
and 10Al-MCM were also used in the reaction, 5-HMF could reach 42% and 36% at
195℃ for 0.5 hours and 2.5 hours in the glass reactor with stirring magnetically,
respectively. Nevertheless, the reaction temperature was a little higher. In this work, a
Teflon-lined stainless autoclave reactor was used. The PWAl-200 catalyst could
produce 61.7% 5-HMF compared with the PWAl without hydrothermal treatment of