Effect of WOx on Bifunctional Pd-WOx/Al2O3 Catalysts for the Selective Hydrogenolysis of Glucose to 1,2-Propanediol

Article abstract of DOI:10.1021/acscatal.5b00800

A series of Pd-WO_x/Al₂O₃ catalysts with different contents of WO_x were prepared by stepwise incipient wetness impregnations. The influence of WO_x on the physicochemical properties of Pd-WO_x/Al₂O₃ catalysts, as well as their catalytic performance for the hydrogenolysis of glucose to 1,2-propanediol (1,2-PDO), was investigated. At low surface W density (0.3-2.1 W nm^-2), distorted isolated WO_x and oligomeric WO_x are present on the Pd-WO_x/Al₂O₃ catalysts. Furthermore, isolated WO₄ are the dominating species on the Pd-WO_x(5%)/Al₂O₃ catalyst. When the W density increased to 3.1 W nm^-2, polymeric WO_x species are dominant on the Pd-WO_x(30%)/Al₂O₃ catalyst. The Pd surface area decreased while the acid amount increased with increasing W density. Furthermore, increased Lewis acid sites are provided by isolated WO₄ and oligomeric WO_x species whereas increased Bronsted acid sites exist on polymeric WO_x species. Lewis acid sites promote glucose isomerization to fructose, which is an intermediate in glucose hydrogenolysis to 1,2-PDO. Metal sites catalyze C=O hydrogenation and C-C hydrogenolysis, which avoid the coke formation on catalysts. 1,2-PDO selectivity is dependent on the synergy of Lewis acid and metal sites; however, Bronsted acid sites have no contribution to the 1,2-PDO production. Typically, the Pd-WO_x(5%)/Al₂O₃ catalyst possessing the optimal balance of Lewis acid and the metal site shows a 1,2-PDO selectivity of 60.8% at a glucose conversion of 92.2% and has a lifetime of over 200 h.

Full text of DOI:10.1021/acscatal.5b00800

Research Article
pubs.acs.org/acscatalysis
Effect of WOx on Bifunctional Pd−WOx/Al2O3 Catalysts for the
Selective Hydrogenolysis of Glucose to 1,2-Propanediol
Chengwei Liu,†,§ Chenghua Zhang,*,†,‡ Sikai Sun,‡ Kangkai Liu,‡ Shunli Hao,‡ Jian Xu,‡ Yulei Zhu,*,†,‡
and Yongwang Li†,‡
†State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, People’s
Republic of China
‡National Energy Center for Coal to Liquids, Synfuels CHINA Co., Ltd., Huairou District, Beijing, 101400, People’s Republic of
China
§University of Chinese Academy of Sciences, Beijing 100049, People’s Republic of China
S
* Supporting Information
ABSTRACT: A series of Pd−WOx/Al2O3 catalysts with
different contents of WOx were prepared by stepwise incipient
wetness impregnations. The influence of WOx on the
physicochemical properties of Pd−WOx/Al2O3 catalysts, as
well as their catalytic performance for the hydrogenolysis of
glucose to 1,2-propanediol (1,2-PDO), was investigated. At
low surface W density (0.3−2.1 W nm−2), distorted isolated
WOx and oligomeric WOx are present on the Pd−WOx/Al2O3
catalysts. Furthermore, isolated WO4 are the dominating
species on the Pd−WOx(5%)/Al2O3 catalyst. When the W
density increased to 3.1 W nm−2, polymeric WOx species are
dominant on the Pd−WOx(30%)/Al2O3 catalyst. The Pd surface area decreased while the acid amount increased with increasing
W density. Furthermore, increased Lewis acid sites are provided by isolated WO4 and oligomeric WOx species whereas increased
Brønsted acid sites exist on polymeric WOx species. Lewis acid sites promote glucose isomerization to fructose, which is an
intermediate in glucose hydrogenolysis to 1,2-PDO. Metal sites catalyze CO hydrogenation and C−C hydrogenolysis, which
avoid the coke formation on catalysts. 1,2-PDO selectivity is dependent on the synergy of Lewis acid and metal sites; however,
Brønsted acid sites have no contribution to the 1,2-PDO production. Typically, the Pd−WOx(5%)/Al2O3 catalyst possessing the
optimal balance of Lewis acid and the metal site shows a 1,2-PDO selectivity of 60.8% at a glucose conversion of 92.2% and has a
lifetime of over 200 h.
KEYWORDS: biomass, WOx, Pd−WOx/Al2O3, hydrogenolysis, glucose, 1,2-propanediol, fructose, isomerization
polyols from biomass resources instead of petroleum-based
ethylene or propylene oxide.
1. INTRODUCTION
Declining fossil fuel reserves, increasing demand for energy, and
ongoing global climate change have stimulated intensive
interest in developing sustainable resources.1,2 In this context,
biomass, an inexpensive, renewable, and abundant organic
carbon resource, has great potential for use as an alternative
feedstock to petroleum in producing liquid fuels and valuable
chemicals.1,3−5 Accordingly, significant work has been done on
the conversion of biomass by various catalytic processes, such
as hydrogenolysis,6−8 dehydration,9,10 oxidation,11,12 aqueous-
phase reforming,13,14 and oligomerization.15 Among them, one
of the most promising processes for biomass transformation is
the catalytic hydrogenolysis of biomass to lower polyols.
Furthermore, lower polyols, such as 1,2-propanediol (1,2-
PDO) and ethylene glycol (EG), are important polyols widely
used for pharmaceutical, polyester resins, paints, cosmetics,
antifreeze, etc.16 This process offers a sustainable and
economically competitive route for the synthesis of lower
The hydrogenolysis of sugar and sugar alcohols to lower
polyols has been reported since the 1930s. A large number of
patents have subsequently appeared.17−20 Despite many
metallic components and base additives such as Ca(OH)2
and NaOH being used in the reaction system, the products
have contained mainly a mixture of glycerol, 1,2-PDO, and EG.
Nowadays, with the abundance of glycerol from biodiesel
production, the importance of polyols-related value-added
chemicals has shifted toward EG and 1,2-PDO.21 Some efforts
have been devoted recently to the hydrogenolysis of biomass to
1,2-PDO. Much more recent patents reported by Werpy et al.
have described the hydrogenolysis of sugar and sugar alcohol to
lower polyols.22−24 For example, a supported rhenium-
Received: February 10, 2015
Revised: June 16, 2015
© XXXX American Chemical Society
4612
DOI: 10.1021/acscatal.5b00800
ACS Catal. 2015, 5, 4612−4623

ACS Catalysis
Research Article
Scheme 1. Possible Reaction Pathway of Glucose Hydrogenolysis
containing multimetallic catalyst was used. The additional
metals included Ni, Pd, Ru, Co, Ag, Au, Rh, Ir, Os, and Cu, of
which Ni, Pd, and Ru are preferred. Especially, 30% selectivity
of 1,2-PDO was obtained in sorbitol hydrogenolysis at 220 °C
and 4.14 MPa hydrogen over a Ni−Re/C catalyst. The group of
Zhang2 has studied the hydrogenolysis of Jerusalem artichoke
tuber to 1,2-PDO over Ni−W2C/AC catalysts and obtained a
high yield of 38.5% at 245 °C and 6 MPa H2. Highly efficient
CuCr catalysts have been developed for the conversion of
cellulose to 1,2-PDO.5 A maximum 1,2-PDO yield of 42.6% at
100% conversion was obtained with Ca(OH)2 as a cocatalyst.
Liu25 obtained 40.9% of PDO yield from cellulose on the Ru/C
catalyst combined with 50%WO3/Al2O3 and activated carbon.
Apparently, the catalytic process still encountered some
challenges, such as environmental problems, poor selectivity,
and an unclear reaction mechanism.
Glucose is the most abundant biomass platform molecule in
nature and easily soluble in water. Compared with cellulose,
glucose is simple in molecular structure, convenient for
characterization, and suitable for the study of the catalytic
mechanism in catalytic reactions.26 The hydrogenolysis of
glucose in aqueous solution can avoid fatal weaknesses in
biomass conversion, such as the feedstock transportation, batch
reaction, coking, organic solution, environment pollution, etc.
All these merits make glucose an ideal feedstock in future large-
scale biorefining. However, the aqueous hydrogenolysis of
glucose has rarely been studied in recent years. Moreover,
highly efficient and stable catalysts are the key factors for the
selective conversion of glucose to 1,2-PDO in aqueous solution.
The main reaction pathway of glucose hydrogenolysis to
lower polyols has been depicted in the literature.25,27−32 As
shown in Scheme 1, for glucose hydrogenolysis, hydrogenation,
isomerization, retro-aldol condensation, and dehydration
reaction occur mainly under hydrothermal and hydrogen
atmosphere conditions. It is generally accepted that the
hydrogenation of glucose to sorbitol can easily occur under
mild conditions over transition metal sites, especially Ni and
Ru. The hydrogenation of fructose often produces mannitol.
Using several substrates, Wang et al. showed that the retro-aldol
condensation reaction is the dominant pathway for C−C bond
cleavage under hydrogenation conditions.33 The retro-aldol
reaction happens mainly at temperatures higher than 160 °C
with and without base catalysts.34−37 The cleavage of C−O
bonds is generally considered to occur via dehydration on acid
sites.38,39 The hydrogenolysis of glucose to 1,2-PDO under
hydrothermal conditions contains two steps: (1) the isomer-
ization of glucose to fructose on Lewis acid/base sites and (2)
the retro-aldol condensation reaction of fructose to glycer-
aldehyde and dihydroxyacetone, followed by dehydration and
hydrogenation on metal and acid sites.40−42 However, glucose
can also undergo retro-aldol condensation to form glycolalde-
hyde and erythrose, which can be hydrogenated to EG and
erythritol, respectively. This reaction route is competitive with
glucose isomerization, which is the first step for 1,2-PDO
production from glucose. Therefore, a bifunctional catalyst with
proper acid and metal sites is required for the isomerization of
glucose and the following hydrogenolysis of fructose to 1,2-
PDO.
It was reported that supported Pt and Pd catalysts with
mineral acids were effective in C−O bond cleavage.43,44 Pd has
been widely used as the active metal in the selective
hydrogenation catalysts.45 γ-Al2O3 is widely used in industry
because of the abundant surface acid sites. Furthermore, γ-
Al2O3 can promote the isomerization of glucose to fructose.46,47
Because of the outstanding catalytic performance, tungsten
oxide-based catalysts have been used in many applications,
including metathesis and isomerization of alkenes, selective
oxidation of unsaturated compounds, selective reduction of
nitric oxide, dehydrogenation of alcohols, and hydrodesulfuri-
zation and hydrocracking of heavy fractions in the petroleum
chemistry. The molecular and electronic structures of the
surface WOx species as well as surface chemical characteristics
in supported WOx catalysts have been extensively inves-
tigated.48 Moreover, tungsten oxide on alumina was used for
hydrotreating and hydrocarbon cracking reactions and emerged
as an efficient solid acid catalyst for its changeable acid
properties and the strong interaction between WOx and γ-
Al2O3.49 Therefore, a bifunctional catalyst containing Pd, WOx,
4613
DOI: 10.1021/acscatal.5b00800
ACS Catal. 2015, 5, 4612−4623

ACS Catalysis
Research Article
Table 1. Physicochemical Properties of the Pd−WOx/Al2O3 Catalysts
metal loading
(mmol
a
g cat.−1
)
W surface density
WOx (W nm−2
dPdO
dPd
d
chemisorption (μmol CO NH3 (μmol
Brønsted/
f
b
c
Pd d
d
e
catalyst
Pd/SiO2
Pd
)
(nm)
(nm)
(nm)
g cat.−1
)
g−1
0.35
53.7
)
Lewis acid
0.102
0.072
0.075
0.072
0.079
0.076
0.080
12.9
4.2
9.8
19.3
18.1
15.3
12.8
4.5
0.0
Pd/Al2O3
4.9
5.0
0.0
Pd−WOx(2%)/Al2O3
Pd−WOx(5%)/Al2O3
Pd−WOx(10%)/Al2O3
Pd−WOx(20%)/Al2O3
Pd−WOx(30%)/Al2O3
0.09
0.21
0.41
0.84
1.20
0.3
0.6
1.0
2.1
3.1
4.6
59.9
65.0
73.9
77.0
84.4
0.0
5.5
5.3
0.0
6.9
6.9
0.0
10.9
11.5
11.8
13.4
19.0
70.1
0.06
0.51
1.2
a
b
c
Metal loading determined using ICP. Calculated from Scherrer equation for the peak of PdO (111). The XRD results of spent catalysts,
d
e
calculated from Scherrer equation for the peak of PdO (111). Calculated from the CO-pulse chemisorption results. Calculated from the NH3-TPD
results. The amount of acid sites was determined by quantifying the desorbed pyridine from FT-IR.
f
Figure 1. XRD patterns of (A) calcined and (B) spent catalysts: (a) Pd/Al2O3, (b) Pd−WOx(2%)/Al2O3, (c) Pd−WOx(5%)/Al2O3, (d) Pd−
WOx(10%)/Al2O3, (e) Pd−WOx(20%)/Al2O3, and (f) Pd−WOx(30%)/Al2O3.
and Al2O3 components was designed for the hydrogenolysis of
glucose to 1,2-PDO.
were reduced at 500 °C and atmospheric pressure for 1 h with a
ramp of 2 °C/min and a hydrogen flow rate of 100 mL/min.
2.2. Catalyst Characterization. The chemical composi-
tions of prepared samples were determined by ICP optical
emission spectroscopy (Optima2100DV, PerkinElmer). N2
physisorption isotherms were measured on a micromeritics
ASAP 2420 instrument. Powder X-ray diffraction (XRD)
patterns were recorded with a D2/max-RA X-ray diffractometer
(Bruker, Germany), operating with Cu Kα radiation at 30 kV
and 10 mA. The dispersion and surface area of the palladium
were determined by CO pulse chemisorption, which was
performed on Auto Chem II 2920 equipment (Mircromeritics,
USA) with a TCD detector. HRTEM was performed with a
Tecnai G2 F30 electron microscope operating at 300 kV.
Raman spectra were recorded at room temperature with a
LabRAM HR800 system equipped with a CCD detector. The
325 nm of the He−Cd laser was used as the excitation source
with a power of 30 MW. UV−vis DRS spectra were collected
on an Agilent cary 5000 UV−vis spectrometer with BaSO4 as a
reference. NH3 temperature-programmed desorption (NH3-
TPD) measurements were conducted on Auto Chem II 2920
equipment (Mircromeritics, USA). The Py-FTIR spectra were
obtained using a VERTEX70 Bruker FT-IR spectrometer. X-ray
photoelectron spectroscopy (XPS) was recorded on a Thermal
XPS ESCALAB 250Xi spectrometer with Al Kα radiation
(1486.6 eV) and a multichannel detector. Prior to the test, the
calcined sample was reduced in H2 at 500 °C for 1 h. The
obtained binding energies were calibrated using the C 1s peak
In this context, a series of bifunctional Pd−WOx/Al2O3
catalysts were first reported for selective hydrogenolysis of
glucose to 1,2-PDO. The structure−performance relationship
of Pd−WOx/Al2O3 catalysts in glucose hydrogenolysis to 1,2-
PDO have been discussed in detail. The role of WOx in glucose
hydrogenolysis to 1,2-PDO was revealed.
2. EXPERIMENTAL SECTION
2.1. Catalyst Preparation. WOx-modified Pd/Al2O3
(denoted as Pd−WOx/Al2O3) catalysts with different WOx
nominal weight loadings in the range of 0−30% were prepared
by a stepwise impregnation method. First, the support, γ-Al2O3
(purchased from Sinopharm Chemical Reagent Co., Ltd.,
China), was impregnated with an aqueous solution of
(NH4)6W7O24·6H2O (Sinopharm Chemical Reagent Co.,
Ltd., China). The impregnated sample was dried overnight at
120 °C and then calcined at 600 °C for 4 h. The obtained
sample was then impregnated with an aqueous solution of
Pd(NO3)2 (Shanxi Kaida Chemical Engineering Co., Ltd.).
After impregnation, the sample was dried overnight at 120 °C
and then calcined at 500 °C for 4 h. The catalysts were labeled
as Pd−WOx(M)/Al2O3, in which M stands for the nominal
weight loading of WOx. The Pd/Al2O3 catalyst without WOx
was prepared as the method mentioned above. Prior to the
catalytic tests, all the Pd−WOx/Al2O3 and Pd/Al2O3 catalysts
4614
DOI: 10.1021/acscatal.5b00800
ACS Catal. 2015, 5, 4612−4623

ACS Catalysis
Research Article
Figure 2. HRTEM images of (A) Pd−WOx(2%)/Al2O3, (B) Pd−WOx(5%)/Al2O3, and (C) Pd−WOx(20%)/Al2O3.
Figure 3. Raman spectra of (A) calcined and (B) reduced catalysts: (a) Pd/Al2O3, (b) Pd−WOx(2%)/Al2O3, (c) Pd−WOx(5%)/Al2O3, (d) Pd−
WOx(10%)/Al2O3, (e) Pd−WOx(20%)/Al2O3, and (f) Pd−WOx(30%)/Al2O3.
at 284.6 eV as the reference. The experiment error reported in
this study is within 0.1 eV. Details are in the Supporting
Information.
BET data (Table S1) varied from 0.3 to 3.1 W nm−2
(calculation methods are described in the Supporting
Information).
Figure 1A displays the XRD patterns of the calcined Pd−
WOx/Al2O3 catalysts. All the samples exhibited the character-
istic peaks of γ-alumina at 2θ = 37.4, 46.0, and 67.0°.50
Obviously, the intensity of the Al2O3 characteristic peaks
decreased as a result of WOx loading, reflecting the decrease in
the Al2O3 molecule coverage on the surface of the catalysts. No
diffraction peaks of crystalline WO3 were observed for Pd−
WOx/Al2O3 catalysts with a W density below 3.1 W nm−2,
indicating the formation of highly dispersed WOx species.51 In
addition, the diffraction peaks at 2θ = 23.3, 23.8, and 24.5° were
observed on the Pd−WOx(30%)/Al2O3 catalyst, which could
be attributed to crystalline phase of WO3,52,53 indicating the
agglomeration and low dispersion of tungsten species on the
surface of Pd−WOx(30%)/Al2O3. There was also a significant
diffraction peak at 33.8°, which is attributed to PdO (JCPDS
75−0584). Furthermore, the intensity of the peak increased
obviously with increasing W density to 2.1 and 3.1 W nm−2,
indicating the increased PdO crystallite size and the decreased
dispersion. The PdO crystallite sizes of the Pd/Al2O3, Pd−
WOx(2%)/Al2O3, Pd−WOx(5%)/Al2O3, Pd−WOx(10%)/
Al2O3, Pd−WOx(20%)/Al2O3, and Pd−WOx(30%)/Al2O3,
2.3. Catalytic Reaction Process. The hydrogenolysis of
glucose was carried out in a tubular fixed-bed reactor (i.d. 12
mm, length 600 mm) with a cold trap. In a typical run, 2.0 g of
catalyst was packed at the isothermal zone and reduced. After
reduction, a 5 wt % glucose aqueous solution was introduced
into the reactor using a pump along with H2 cofeeding. The
reaction conditions were set as follows: 5 wt % of glucose
aqueous solution, 4 MPa H2 (45 mL/min), 160−220 °C, molar
ratio of H2 to glucose = 22.6:1, WHSV (reactant per gram of
catalyst per hour) = 0.48 h−1. The liquid products were
collected after 1 h of stable reaction in each temperature point.
The total reaction time on-stream was 10 h with temperature
increasing from 160 to 220 °C.
3. RESULTS AND DISCUSSION
3.1. Catalyst Characterization. 3.1.1. Structural Proper-
ties of Catalysts. As shown in Table 1, the concentrations of
Pd and W were close to the nominal loading as designed,
despite a small discrepancy owing to the loss in preparation
process. The W surface density determined with ICP-OES and
4615
DOI: 10.1021/acscatal.5b00800
ACS Catal. 2015, 5, 4612−4623

ACS Catalysis
Research Article
estimated from the Scherrer equation, were 4.9, 5.0, 5.5, 6.9,
10.9, and 11.5 nm, respectively.
WO3 under reduction conditions. The new Raman bands at
844 and 578 cm−1 on Pd−WOx(30%)/Al2O3 reflect the
presence of bridging W−O−W bonds, suggesting the higher
polymerization of WOx species.
UV−vis spectra are often used to discriminate between
different tungsten oxide species in combination with Raman
spectroscopy, such as isolated surface WOx, oligomeric surface
WOx, and polymeric WOx (WO3 NPS or large bulk-like WO3
particles). The UV−vis spectra of reduced catalysts are shown
in Figure 4. A weak band around 271 nm is displayed on the
CO pulse chemisorption was used to determine the Pd
dispersion, specific surface area, and particle size. The Pd
particle size increased and the dispersion of Pd decreased with
increasing W density. CO pulse chemisorption shows that the
amount of CO adsorption decreased with increasing W density.
The results were consistent with XRD. Compared with XRD
results, Pd particle sizes in reduced Pd−WOx(30%)/Al2O3
catalyst were much larger than that in its calcined precursor.
It indicates that Pd particles sintered after reduction. To
confirm the dispersion and morphology of Pd particles,
HRTEM images of reduced samples were obtained (Figure
2). The lattice distances of 0.227, 0.228, and 0.220 nm could be
attributed to the (111) planes of face-centered cubic (fcc) Pd
(JCPDS 46-1043). It can be observed that the Pd particles with
diameters of ∼5 nm are uniformly dispersed in the Pd−
WOx(2%)/Al2O3 and Pd−WOx(5%)/Al2O3 catalysts; however,
large NPs with diameters of ∼20 nm are observed on Pd−
WOx(20%)/Al2O3, indicating severe agglomeration occurred
during reduction. Therefore, the CO pulse chemisorption,
XRD, and HRTEM results are in good agreement with each
other.
Raman spectroscopy is often used to determine the
molecular structures of the bulk and supported tungsten
oxide catalysts. Figure 3 shows the Raman spectra of calcined
and reduced Pd−WOx/Al2O3 catalysts. The UV excitation was
used because of sample fluorescence when excited in the visible
region. It is known that the γ-Al2O3 support has no Raman
resonance. As shown in Figure 3A, the Raman bands at 427 and
626 cm−1 are associated with the Raman-active Eg and B1g
vibrational modes of PdO, respectively.54 The Raman band
around 322 cm−1 originates from a resonance effect induced by
the use of laser excitation near the excitation state of PdO.54
The broad bands at 980−1002 cm−1 can be ascribed to the
symmetric stretching vibration mode of the terminal WO
bonds of well-dispersed tetrahedral WOx species.54−57 The
band intensity increases with the increase of the W density.
This behavior is indicative of a progressive buildup of WOx
surface species.56 Furthermore, the band at 980 cm−1 in the
spectra of Pd−WOx/Al2O3 catalysts in the 0.3−2.1 W nm−2
range is characteristic of highly dispersed WOx species, such as
isolated WO4 or low-condensed oligomeric (e.g., dipolymer)
WOx species.48
Figure 4. UV−vis DRS spectra of the reduced catalysts.
spectrum of Pd/Al2O3. It is negligible compared with the strong
absorption of the WOx species. Crystalline WO3 is a three-
dimensional structure composed of distorted WO6 units.48 Bulk
WO3 contains two ligand-to-metal charge transfer (LMCT)
bands at 252 and 343 nm with the corresponding UV−vis DRS
Eg value of 2.9 eV (Figure S1 of the Supporting Information).
The UV−vis spectra of Pd−WOx/Al2O3 catalysts are
completely different from that of crystalline WO3, indicating
that WOx species on the Pd−WOx/Al2O3 catalysts are in a
highly dispersed form. A band at 240 nm was detected in the
spectra of Pd−WOx/Al2O3 catalysts, which could be ascribed to
the slightly distorted WO4 species.48 Furthermore, a weak
shoulder band at 252 nm was shown in the spectra of Pd−
WOx(20%)/Al2O3 and Pd−WOx(30%)/Al2O3 catalysts. How-
ever, it is not clear whether the shoulder band exists in the
spectra of the catalysts in the 0.3−1.0 W nm−2 range. As
pointed out by Ross-Medgaarden,48 several types of tungstate
compounds give rise to LMCT band at about 250 nm, such as
distorted monotungstate and polytungstates, especially oligo-
meric unit and polymeric polytungstate species. Therefore,
polytungstate species may also exist on the catalysts in the 0.3−
1.0 W nm−2 range. The UV−vis DRS Eg value allows for the
discrimination of slightly distorted isolated WO4/WO6 (Eg >
4.4 eV) and polymeric structure (Eg < 4.0 eV). Both highly
distorted WO4 and polytungstate species could exist in Eg
values of 4.0−4.4 eV. The Eg value for Pd−WOx(30%)/Al2O3
of 3.8 eV is lower than that for other catalysts, reflecting that
tungsten species in Pd−WOx(30%)/Al2O3 are polymeric
polytungstate. The Eg value for Pd−WOx(5%)/Al2O3 of 4.4
is the highest in these catalysts, which is closer to the value of
distorted isolated WO4/WO6 species. Meanwhile, the shoulder
peak at 252 nm was not obvious. It suggests that the WOx
species on the Pd−WOx(5%)/Al2O3 catalyst are present mainly
as isolated WO4. The Eg values for Pd−WOx(2%)/Al2O3, Pd−
A band around 860 cm−1, which can be assigned to the
symmetric stretching mode of the W−O−Al,48 was also
observed. The peak intensity also increased with increasing
W density. For Pd−WOx(30%)/Al2O3 catalyst, three new
Raman bands at 805, 710, and 267 cm−1 appeared, which could
be assigned to W−O stretching, W−O bending, and W−O−W
deformation modes of crystalline WO3 nanoparticles, respec-
tively.48 Accordingly, the Raman band shifted from 980 to 1002
cm−1, indicating an increased polymerization of the surface
WOx species.49,58 For the reduced samples in Figure 3B, we can
see that Raman bands at 980 cm−1 were also observed for the
samples in the 0.3−1.0 W nm−2 range, indicating that the WOx
also presented as tetrahedral coordinated surface species on the
reduced samples. However, two bands at 989 and 1006 cm−1
were observed in the spectra of samples in the 2.1−3.1 W nm−2
range, indicating the present of a mixture of two types of W
surface species with different degrees of condensation.48 For
the reduced Pd−WOx(30%)/Al2O3 catalyst, the Raman bands
corresponding to WO3 disappeared, indicating the reduction of
4616
DOI: 10.1021/acscatal.5b00800
ACS Catal. 2015, 5, 4612−4623

ACS Catalysis
Research Article
WOx(20%)/Al2O3, and Pd−WOx(10%)/Al2O3 catalysts are
4.0, 4.1, and 4.2 eV, respectively, indicating that tungsten
species are present in a mixture of polytungstate and highly
distorted isolated WO4 in these catalysts.
By combining the Raman and UV−vis results, it can be
concluded that highly dispersed WOx species such as highly
distorted isolated WO4 or oligomeric WOx species are present
on the catalyst surface mainly in the 0.3−2.1 W nm−2 range.
Especially, the Pd−WOx(5%)/Al2O3 surface was occupied by
isolated WO4 species. Tungsten species in Pd−WOx(30%)/
Al2O3 are present mainly as polymeric polytungstate.
3.1.2. Acid Properties of the Reduced Pd−WOx/Al2O3
Catalysts. The acid properties of the Pd−WOx/Al2O3 catalysts
were measured by NH3-TPD and Py-FTIR spectra. NH3-TPD
profiles and the total acid amount of the reduced samples are
shown in Figure S2 and Table 1, respectively. Figure S2A shows
that Pd−WOx/Al2O3 catalysts possess broad ammonia
desorption curves between 185 and 400 °C, with the central
peak at ∼187 °C, suggesting the presence of main weak acid
sites on the sample surface. The acid amounts based on the
desorbed NH3 amounts of these catalysts increase monotoni-
cally with increasing surface W density.
are provided mainly by low-condensed polytungstate, especially
the isolated WO4 species, whereas Brønsted acid sites exist
mainly on polytungstate, which is consistent with the
literature.49 Wu reported that the increased Lewis acid sites
were related to the coordinatively unsaturated Wx+ cations, and
the Brønsted acid sites arose from partially hydrated tungsten
species, including OH groups associated with WO, W−O−
W, and W−O−Al linkages in WOx clusters.51
Chen50 has studied the correlation between the WOx
structure and the acidity of WOx/Al2O3 catalysts. Brønsted
acid was negligible for the alumina support and low for a W
surface density below 1.4 W nm−2. The amount of Lewis acid
decreased, and that of Brønsted acid sites increased with
increasing W density in the 1.4−4 W nm−2 range. It was
reported that a direct relationship between the amount of
Brønsted acid sites and that of the polymeric WOx surface
species exists. A similar relationship was also found on WOx/
ZrO2.60,61 Moreover, Onfroy60 reported that a minimum of the
surface W density (above 1.6 W nm−2) of the supported phase
was required for the formation of Brønsted acid sites. These
results are consistent with the present results in this work.
3.1.3. Surface Chemical States of WOx Species before and
after Reaction. The XPS spectra of W 4f and the deconvoluted
ones of the supported WOx catalysts are shown in Figure 6 and
Table 2. As shown in Figure 6A, for the reduced Pd−
WOx(2%)/Al2O3 catalyst, the W 4f doublets at 35.4 and 37.5
eV are attributed to W5+ atoms, and peaks at around 37.0 and
39.1 eV are ascribed to W6+ atoms.62 The molar ratios of W5+/
The FTIR spectra of adsorbed pyridine were used to study
the types of acid sites, as shown in Figure 5. The IR band at
W
6+ on Pd−WOx/Al2O3 catalysts in the 0.3−0.6 W nm−2 range
were lower than that in the 1.0−3.1 W nm−2 range, indicating
that WOx in the higher W density is easily reduced. After
reaction, the ratios of W5+/W6+ of the Pd−WOx(2%)/Al2O3
and Pd−WOx(5%)/Al2O3 catalysts did not change. It suggests
that the valence state of the highly isolated tungsten species was
stable during the reaction. However, the ratios of W5+/W6+ of
Pd−WOx(10%)/Al2O3, Pd−WOx(20%)/Al2O3 and Pd−
WOx(30%)/Al2O3 decreased obviously after reaction, indicat-
ing that the reduced polymeric tungsten species are susceptible
to oxidation in steam atmosphere during the glucose
hydrogenolysis.
3.2. Catalytic Performance of Pd−WOx/Al2O3 in
Glucose Hydrogenolysis. The activity and selectivity of the
Pd−WOx/Al2O3 catalysts for glucose hydrogenolysis are shown
in Figure 7 and Table 3. For the product distribution, only a
trace amount of CH4 was detected in the gas phase, with the
yield below 0.2%. No CO or other gas products were detected.
Except for the large amount of isomerization and hydro-
genation products of C6 species that were fructose and sorbitol
in liquid phase, a wide variety of hydrogenolysis products,
including ethanol, EG, glycerol, 1,2-PDO, isopropyl alcohol,
propanol, erythritol, 1,2-butanediol (1,2-BDO), 2-butanol,
xylitol, 1,2-pentanediol, and 1,2-hexanediol, were detected.
Table 3 shows that the selectivities of EG and 1,2-BDO
changed slightly with the loading of WOx species. Glucose can
undergo retro-aldol condensation to form glycolaldehyde and
erythrose, which can be hydrogenated to EG and erythritol,
respectively. Furthermore, erythrose can also undergo further
dehydration and hydrogenation reaction to form 1,2-BDO. The
results suggest that WOx species have nearly no effect on the
retro-aldol condensation of glucose or the dehydration of the
intermediates. Furthermore, fructose and 1,2-PDO were the
predominant products. However, only fructose was the main
product in the liquid phase on WOx(5%)/Al2O3, with the
Figure 5. Py-FTIR spectra of the reduced catalysts: (a) Pd/Al2O3, (b)
Pd−WOx(2%)/Al2O3, (c) Pd−WOx(5%)/Al2O3, (d) Pd−
WOx(10%)/Al2O3, (e) Pd−WOx(20%)/Al2O3, and (f) Pd−
WOx(30%)/Al2O3.
1540 cm−1 is ascribed to pyridinium ions adsorbed on protonic
acid sites, and the band at 1450 cm−1 is attributed to pyridine
coordinated to Lewis acid sites.59 The Lewis acid is present on
all the Pd−WOx/Al2O3 catalysts but the protonic acid is
present only on catalysts with higher W densities (2.1−3.1 W
nm−2). The intensities of the band at 1450 cm−1 decreased
significantly as the desorption temperature increased from 30 to
350 °C, indicating the acid sites are mainly composed of weak
acid sites. When W density increased to 2.1 W nm−2, a weak
band at 1540 cm−1 was observed when outgassing at 30 °C and
almost disappeared when outgassing at 200 °C, indicating that a
small amount of weak Brønsted acid appeared on the catalyst
surface. Furthermore, the mole ratio of Brønsted to Lewis acid
increased from 0.02 in Pd−WOx(20%)/Al2O3 to 0.51 in Pd−
WOx(30%)/Al2O3. It suggests that increased Lewis acid sites
4617
DOI: 10.1021/acscatal.5b00800
ACS Catal. 2015, 5, 4612−4623

ACS Catalysis
Research Article
Figure 6. W 4f XPS spectra of the (A) reduced and (B) spent catalysts: (a) Pd−WOx(2%)/Al2O3, (b) Pd−WOx(5%)/Al2O3, (c) Pd−WOx(10%)/
Al2O3, (d) Pd−WOx(20%)/Al2O3, and (e) Pd−WOx(30%)/Al2O3.
Table 2. XPS Parameters of the W 4f Region in Pd−WOx/Al2O3 Samples
a
binding energy for W 4f (eV)
W5+/W6+
sample
W6+ W 4f5/2
W6+ W 4f7/2
W5+ W 4f5/2
W5+ W 4f7/2
after reduced
after reaction
Pd−WOx(2%)/Al2O3
Pd−WOx(5%)/Al2O3
Pd−WOx(10%)/Al2O3
Pd−WOx(20%)/Al2O3
Pd−WOx(30%)/Al2O3
39.1 (39.8)
39.8 (39.6)
40.5 (39.7)
40.6 (40.0)
40.9 (40.1)
37.0 (37.7)
37.7 (37.5)
38.4 (37.6)
38.5 (37.9)
38.8 (38.0)
37.5 (37.2)
37.6 (37.2)
37.7 (37.3)
37.8 (37.4)
37.8 (37.5)
35.4 (35.1)
35.5 (35.1)
35.6 (35.2)
35.7 (35.3)
35.7 (35.4)
1.3
0.9
2.2
2.1
2.2
1.3
1.0
1.0
1.0
1.1
a
The B.E. values outside parentheses are ascribed to W4f of the reduced Pd−WOx/Al2O3 catalysts. The B.E. values in parentheses are assigned to W4f
of the spent Pd−WOx/Al2O3 catalysts.
selectivity of 67.9% at 160 °C. Furthermore, the fructose
selectivity decreased to 24.1% when the temperature increased
to 180 °C, with most of the feedstock cooking in the reactor. It
is known that glycerol and 1,2-PDO easily undergo further
degradation under an acid environment.63 In this work, except
for glycerol and 1,2-PDO, only a trace amount of propanol with
a selectivity lower than 1% was detected; however, no other C3
products, such as propanal, propylene, or propane, were
detected. The results show that 1,2-PDO showed very low
hydrogenolysis activity under this condition, which must be
related to the reaction conditions and catalyst properties. The
carbon balance, in the percentage of carbon accounted (both in
liquid and gas phase), decreased with increasing temperature. It
suggests that glucose was coked at a higher temperatures.
Furthermore, the carbon balance was above 85% for all Pd−
WOx/Al2O3 catalysts at 160 and 180 °C.
Figure 7A shows that the activity of glucose hydrogenolysis
increased with the increase in the temperature. The selectivity
of fructose decreased sharply with increasing temperature;
however, the selectivity of 1,2-PDO increased first and then
decreased with an increase in the temperature. It displayed that
180 °C was an optimal temperature for 1,2-PDO production.
The doping of WOx influences not only the catalytic activity
but also the product selectivity. The conversion of glucose
increased first and then decreased with increasing W density at
180 °C. For the product selectivity shown in Figure 7B,
fructose selectivity decreased first and then increased with
increasing W density at 180 °C. The selectivity of 1,2-PDO
showed a trend completely the opposite of fructose, as shown
in Figure 7C. 1,2-PDO selectivity increased first and then
decreased with increasing W density. The maximum selectivity
was obtained on Pd−WOx(5%)/Al2O3 at 180 °C, with the
highest value being 60.8%, which is one of the best
performances reported in this reaction.
3.3. The Property-Performance Relationship of Pd−
WOx/Al2O3 Catalysts in Glucose Hydrogenolysis to 1,2-
PDO. H2 pressure plays an important role in glucose
hydrogenoysis. In the pressure range of 2.0−7.5 MPa,
hydrogenation is considered to be the first-order kinetics of
hydrogen pressure.64 Furthermore, lowering the hydrogen
pressure is expected to decrease the activity of glucose
hydrogenation to sorbitol, increase the activity of retro-aldol
reaction, and improve the selectivity to lower polyols.41,65
Ooms reported that when the hydrogen pressure was lowered
from 60 to 45 bar, the sorbitol yield decreased and the ethylene
glycol yield slightly increased. However, a further decrease in
hydrogen pressure to 30 bar resulted in a lower ethylene glycol
yield and an unfavorable carbon mass balance. To inhibit the
hydrogenation of glucose to sorbitol and improve the carbon
mass balance, 4 MPa was applied for glucose hydrogenolysis.
Usually, a small metal size or large specific area means a high
hydrogenation capability; however, the glucose conversions do
not strictly follow the monotonic decreasing tendency of Pd
specific area in these Pd−WOx/Al2O3 catalysts, proving that Pd
is not the sole influencing factor in glucose conversion.
Unfortunately, the variation of acidity does not coincide with
the tendency of activities and 1,2-PDO selectivities of catalysts.
Combined with the 1,2-PDO selectivity, metal surface area, and
acid amount, we can see that 1,2-PDO was dependent on the
synergy of the metal and acid sites, and an excellent catalytic
4618
DOI: 10.1021/acscatal.5b00800
ACS Catal. 2015, 5, 4612−4623

ACS Catalysis
Research Article
results show that the Pd metal sites on this acid-free catalyst
surface mainly act as a hydrogenation function but have quite
low C−C bond cleavage capability.
To understand the intrinsic function of the acid site in the
reaction, model reactions of glucose hydrogenolysis on Al2O3
and WOx(5%)/Al2O3 in H2 and N2 atmosphere were
investigated (Table S3 of the Supporting Information). The
results indicate that the pure acid catalysts show an isomer-
ization selectivity of glucose to fructose but quite low
hydrogenation activity. However, a large amount of humins
was observed on both catalysts. Furthermore, the exact humin
products are too complicated to be identified, which inhibits us
from understanding the role of acid sites in glucose conversion.
When the 1,2-PDO selectivity and the characterization results
of acidity are combined, it can be found that the increased
Lewis acid on the Pd−WOx/Al2O3 catalysts with lower W
densities (0−1.0 W nm−2) could promote glucose hydro-
genolysis to 1,2-PDO and the increased Brønsted acid has no
additional contribution to 1,2-PDO selectivity. This phenom-
enon indicates that a Lewis acid plays important roles in
catalyzing glucose to 1,2-PDO. Therefore, it is worthwhile to
study the exact role of these structures that result in the Lewis
acid in glucose hydrogenolysis.
3.4. Effect of the Structure of WOx on Glucose
Hydrogenolysis to 1,2-PDO. As indicated by the above
results, the Pd−WOx(5%)/Al2O3 catalyst shows excellent
selectivity to 1,2-PDO. Meanwhile, the W species in this
catalyst are present as isolated WO4. It indicates that the
isolated WO4 species plays the vital role in the 1,2-PDO
production.
The ratios of W5+/W6+ of the Pd−WOx/Al2O3 after reaction
were nearly the same, especially in the 0.6−3.1 W nm−2 range.
However, these catalysts showed obvious difference on glucose
conversion and 1,2-PDO selectivity. It suggests that the
catalytic performances were independent of the valence state
of tungsten. To further reveal the function of the WOx structure
in glucose conversion, the hydrogenolysis of glucose on bulk
WO3 crystal and Pd/WO3 was conducted; the results are
shown in Table 4. WO3 showed nearly no activity for glucose
conversion. Furthermore, both WO3 and Pd/WO3 showed very
low selectivity for 1,2-PDO. It suggests that the crystalline WO3
or highly polymerized WOx species have no contribution to the
hydrogenolysis of glucose and 1,2-PDO production. Indeed, the
highly dispersed WOx species, especially the isolated WO4
species, promote glucose conversion and improve the 1,2-PDO
selectivity. However, it is unclear how these structures catalyze
the reaction. It is worth studying and discussing them in depth.
3.5. The Mechanism of the Promoting Effect of WOx
for Glucose Hydrogenolysis to 1,2-PDO. To study the
mechanism of glucose hydrogenolysis to 1,2-PDO on Pd−
WOx/Al2O3, the hydrogenolysis of fructose on Pd/Al2O3 and
Pd−WOx(5%)/Al2O3 catalysts was performed; the results are
shown in Table 5. It is shown that fructose hydrogenolysis
showed higher conversion and higher 1,2-PDO selectivity than
glucose hydrogenolysis on Pd/Al2O3 catalyst (as shown in
Table 5). It indicates that glucose hydrogenolysis is a cascade
reaction. This reaction contains at least two consecutive
reactions: glucose isomerization to fructose and fructose
hydrogenolysis to 1,2-PDO. Moreover, glucose isomerization
is a rate-determining step for glucose conversion. Sasaki41 has
conducted retro-aldol condensation of glucose in supercritical
water. They pointed out that retro-aldol condensation was a
key reaction for saccharide degradation in high-temperature
Figure 7. Glucose conversion and product distribution on Pd−WOx/
Al2O3 catalysts: (A) glucose conversion, (B) fructose selectivity, (C)
1,2-PDO selectivity. Reaction conditions: 4 MPa H2, 5 wt % glucose
aqueous solution, molar ratio of H2 to glucose = 22.6:1, WHSV = 0.48
h−1.
performance was reached on Pd−WOx(5%)/Al2O3. To reveal
the mechanism of the synergy of metal and acid sites, their
individual roles in the reaction route should be investigated.
A Pd/SiO2 catalyst with nearly no acidity was designed to
study the effect of metal sites on glucose conversion and 1,2-
PDO selectivity (Figure S2B and Table S2 of the Supporting
Information and Table 1; a lower reduction temperature of 200
°C was used for the catalytic tests and characterization to
inhibit the sintering of Pd particles). The results showed that
Pd/SiO2 had very poor activity for the conversion of glucose,
with values of 3.6 and 19.6% at 160 and 180 °C, respectively.
Furthermore, sorbitol was the main product with the selectivity
of 74.6% and 57.4% at 160 and 180 °C, respectively. In
contrast, 1,2-PDO selectivity was only 11.0 and 16.1%. The
4619
DOI: 10.1021/acscatal.5b00800
ACS Catal. 2015, 5, 4612−4623

ACS Catalysis
Research Article
a
Table 3. Activity and Selectivity of Pd−WOx/Al2O3 Catalysts for Glucose Hydrogenolysis at 160 and 180 °C
selectivity (%)
b
sample
T (°C)
conv (%)
sor
fru
gly
EG
1,2-PDO
1,2-BDO
CH4
others
carbon balance
WOx(5%)/Al2O3
160
180
160
180
160
180
160
180
160
180
160
180
160
180
26.7
67.6
58.6
73.8
54.1
87.3
74.0
92.2
73.7
90.5
60.7
84.9
58.4
84.6
0.0
0.0
6.8
3.6
7.3
2.4
6.1
2.3
5.7
2.2
4.0
2.0
0.0
0.0
67.9
24.1
23.6
12.2
24.8
8.2
5.3
1.5
9.1
8.8
9.8
7.3
8.8
7.2
8.1
6.6
7.9
6.5
9.4
6.1
0.9
0.9
4.0
5.1
3.7
4.0
5.2
5.7
4.7
5.2
6.1
5.5
5.8
4.6
6.5
6.4
0.0
1.3
4.2
6.7
4.3
8.1
5.9
9.9
7.1
10.9
6.9
10.7
7.0
9.9
0.0
0.0
0.2
0.2
0.2
0.2
0.2
0.1
0.2
0.1
0.2
0.2
0.2
0.2
19.4
65.8
11.2
16.7
6.9
99.5
41.7
99.5
95.3
99.2
93.7
99.4
97.8
99.6
98.6
95.3
94.8
95.5
85.7
Pd/Al2O3
40.9
46.7
43.0
55.9
52.3
60.8
51.7
59.4
41.3
49.3
38.6
42.1
Pd−WOx(2%)/Al2O3
Pd−WOx(5%)/Al2O3
Pd−WOx(10%)/Al2O3
Pd−WOx(20%)/Al2O3
Pd−WOx(30%)/Al2O3
13.9
8.1
13.4
4.1
9.9
11.3
4.1
11.2
11.5
21.4
20.0
19.6
12.2
5.8
19.4
8.2
28.9
a
b
Reaction conditions: 4 MPa H2, 5 wt % glucose aqueous solution, molar ratio of H2 to glucose = 22.6:1, WHSV = 0.48 h−1. Others: ethanol, EG,
isopropyl alcohol, propanol, erythritol, 2-butanol, xylitol, 1,2-pentanediol, and 1,2-hexanediol, etc. sor = sorbitol, fru = fructose, gly = glycerol.
a
Table 4. Glucose Conversion over Model Catalysts
yield (%)
catalyst
atmosphere
conv. (%)
fru
ery
EG
gly
1,2-PDO
1,2-BDO
2-butanol
others
blank
H2
H2
H2
0.0
3.7
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.2
0.0
0.0
1.4
0.0
2.5
3.9
0.0
0.5
1.1
0.0
0.2
0.2
0.0
0.5
8.3
WO3
Pd/WO3
16.1
a
Reaction conditions: 4 MPa H2, 180 °C, 5 wt % glucose aqueous solution, molar ratio of H2 to glucose = 22.6:1, WHSV = 0.48 h−1. fru = fructose,
ery = erythritol, gly = glycerol.
a
Table 5. Activity and Selectivity of Pd−WOx/Al2O3 Catalysts for Fructose Hydrogenolysis at 160 and 180 °C
selectivity (%)
b
sample
T (°C)
conv (%)
glu
sor
ery
EG
gly
1,2-PDO
1,2-BDO
others
Al2O3
180
180
160
180
160
180
68.8
80.3
78.8
95.0
88.3
97.7
15.2
8.7
4.7
3.9
8.2
2.7
0.0
0.0
7.6
2.9
5.8
2.0
3.7
2.5
5.8
3.7
6.2
3.0
0.5
1.4
5.8
8.1
5.0
4.5
0.9
2.3
5.9
6.8
0.0
0.0
4.3
6.8
5.0
73.8
78.3
8.3
WOx(5%)/Al2O3
Pd/Al2O3
10.9
9.3
52.6
57.0
56.8
62.2
8.3
Pd−WOx(5%)/Al2O3
8.8
4.2
6.5
7.7
11.4
a
b
Reaction conditions: 4 MPa H2, 5 wt % fructose aqueous solution, molar ratio of H2 to fructose = 22.6:1, WHSV = 0.48 h−1. Others: ethanol, EG,
isopropyl alcohol, propanol, 2-butanol, xylitol, 1,2-pentanediol, and 1,2-hexanediol, humins, etc. glu = glucose, sor = sorbitol, ery = erythritol, gly =
glycerol.
water. Glycolaldehyde and erythrose, intermediates of C2 and
C4 polyol, could be obtained from retro-aldol condensation of
glucose in supercritical water, whereas glyceraldehyde and
dihydroxyacetone, intermediates of C3 polyols, could be
obtained from retro-aldol condensation of fructose. The
isomerization of glucose to fructose also occurred in super-
critical water. Furthermore, they also proposed that fructose
was the intermediate for the conversion of glucose to C3
products (glyceraldehyde and dihydroxyacetone). Kanie42
found that fructose hydrogenolysis displayed higher 1,2-PDO
selectivity than glucose hydrogenolysis on Pt nanoparticles.
Isomerization between glucose and fructose proceeded at a
temperatures as low as 403 K and could be accelerated by Pt
nanoparticles and H2. They suggested that the isomerization of
glucose to fructose was the intermediate reaction step for the
hydrogenolysis of glucose to 1,2-PDO.
when glucose is the feedstock (as shown in Table 3); however,
they are only slightly increased by WOx when fructose is the
feedstock (as shown in Table 5). It suggests that the tungsten
addition improves the isomerization of glucose to fructose, but
it has no obvious effect on the hydrogenolysis of fructose to
1,2-PDO. These results indicate that fructose is highly active
and acts as the key intermediate for the production of 1,2-PDO.
Combined with all these results, it can be concluded that the
exact role of the dispersed WOx in glucose hydrogenolysis to
1,2-PDO is catalyzing the isomerization step from glucose to
fructose and having a slightly promoting effect on the selective
conversion of fructose to 1,2-PDO. As we have known, the
highly dispersed WOx or isolated WO4 species provide the
Lewis acid sites in catalysts; therefore, it is proposed that the
Lewis acid-related structures could also catalyze the isomer-
ization of glucose to fructose. This speculation is also supported
by the model reaction on Al2O3 for glucose hydrogenolysis,
which shows a fructose selectivity of 42.7% (as shown in Table
Comparing Pd/Al2O3 and Pd−WOx(5%)/Al2O3, the activity
and 1,2-PDO selectivity are largely improved by WOx addition
4620
DOI: 10.1021/acscatal.5b00800
ACS Catal. 2015, 5, 4612−4623

ACS Catalysis
Research Article
S2). Davis et al.66 and Choudhary67 have reported that the
isomerization reaction can be promoted by a Lewis acid in the
catalyst. The Lewis acid could polarize the carbonyl group in
the ketone, which follows an intramolecular hydride shift
mechanism between the carbonyl-containing C-1 and the
hydroxyl-bearing C-2 of glucose.
that they play similar roles and run in parallel in the
hydrogenolysis of glucose. Moreover, WOx species could also
inhibit the reverse conversion of glucose isomerization to
fructose and promote the conversion of fructose to lower
polyols, which makes it an efficient component in catalysts for
glucose hydrogenolysis to 1,2-PDO.
As shown in Scheme 2A, isolated tetrahedral WO4 species
provide the Lewis acid sites. The structure of isolated WO4
3.6. The Stability of Pd−WOx/Al2O3 Catalyst in
Glucose Hydrogenolysis to 1,2-PDO. To investigate the
stability of the catalysts, their structural properties after reaction
were also studied by ICP, BET, CO pulse chemisorption, and
XRD (Table 1, Figure 1, and Table S3). The BET, ICP, and
CO pulse chemisorption results of Pd−WOx(5%)/Al2O3 are
nearly the same before and after reaction, which indicates that
structure damage, leaching, and sintering of the active sites were
negligible. From Figure 1 we can see the XRD patterns of Al2O3
were nearly the same as that of the calcined samples. It
suggested that Al2O3 was stable in the hydrothermal condition
during the reaction. The diffraction peak of PdO disappeared,
whereas the peak at 40.0° was detected for all the spent
catalysts that could be ascribed to Pd0 (JCPDS 46-1043).
However, the Pd peaks were not clear enough for the
calculation of the Pd particle size. The peak intensity of Pd
in the 2.1−3.1 W nm−2 range was also higher than that in the
lower W density. The estimated Pd particle sizes in Table 1 are
11.8 and 13.4 nm for Pd−WOx(20%)/Al2O3 and Pd−
WOx(30%)/Al2O3, respectively, which are slightly higher than
that of the PdO in the calcined samples. The stability of the
Pd−WOx(5%)/Al2O3 catalyst was also tested at different
WHSVs (Figure 8). The activities and selectivities remained
stable over 200 h. It turned out that Pd−WOx(5%)/Al2O3 had
good stability. Metal leaching into the liquid phase after 12 h
on-stream was detected. The results showed that no leaching of
Pd was detected, and only a trace amount of W, 1.48 mg/L, was
detected in the liquid phase. The results confirmed the stability
of the catalyst.
Scheme 2. Probable Structures of (A) Isolated WO4 or (B)
Tetrahedral AlO4 Species on Pd−WOx/Al2O3
species has also been proposed by many researchers.48,68 It is
well-known that the γ-Al2O3 surface is composed of mainly
octahedral and tetrahedral Al sites.69,70 Furthermore, it is
suggested experimentally and theoretically that a tetrahedral
surface Al site has a stronger Lewis acidity than an octahedral
one.69,70 The tetrahedral AlO4 structure on γ-Al2O3 surface is
shown in Scheme 2B. In addition, Davis et al.66 found that Sn
incorporated in the framework of zeolite Beta, which is also
present as tetrahedral coordination and acts as Lewis acid,
performed the isomerization reaction following an intra-
molecular hydride shift mechanism between the carbonyl-
containing C-1 and hydroxyl-bearing C-2 of glucose by a way of
a 5-member complex. Therefore, a similar promoting
mechanism of WO4 or AlO4 is proposed in Scheme 3,
considering the isomerization on Lewis acid sites provided by
isolated WO4 or tetrahedral AlO4 species. Furthermore, the
conversions of fructose on Al2O3 and WOx(5%)/Al2O3 were
also studied to distinguish the roles of acid sites provided by
WOx species and alumina, which is shown in Table 5. The
results showed that fructose conversion was higher on
WOx(5%)/Al2O3 than that on Al2O3. It suggests that WOx
species could promote fructose conversion. However,
WOx(5%)/Al2O3 showed lower glucose selectivity than
Al2O3. It indicates that WOx species could promote the
isomerization of glucose to fructose and inhibit the reverse
reaction. Compared with Al2O3, both of them can promote the
isomerization of glucose to fructose. In this respect, we suppose
CONCLUSION
■
An efficient and stable Pd−WOx/Al2O3 catalyst was designed
for the selective hydrogenolysis of glucose to 1,2-PDO. Highly
dispersed WOx species, such as distorted isolated WO4 and
oligomeric WOx species, are present in the 0.3−2.1 W nm−2
range, and polymeric WOx species are present mainly on the
Pd−WOx(30%)/Al2O3 catalyst. Furthermore, distorted isolated
Scheme 3. Mechanism of Promoting Effect of Isolated WO4 or AlO4 Species on Pd−WOx/Al2O3 for the Isomerization of
Glucose to Fructose
4621
DOI: 10.1021/acscatal.5b00800
ACS Catal. 2015, 5, 4612−4623

ACS Catalysis
Research Article
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was financially supported by the Major State Basic
Research Development Program of China (973 Program) (No.
2012CB215305).
REFERENCES
■
(1) Jimen
́
ez-Morales, I.; Moreno-Recio, M.; Santamaría-Gonzal
́
ez, J.;
Maireles-Torres, P.; Jimen
́
ez-Lopez, A. Appl. Catal., B 2015, 164, 70−
́
76.
(2) Zhou, L.; Wang, A.; Li, C.; Zheng, M.; Zhang, T. ChemSusChem
2012, 5, 932−938.
(3) Sun, R.; Wang, T.; Zheng, M.; Deng, W.; Pang, J.; Wang, A.;
Wang, X.; Zhang, T. ACS Catal. 2015, 5, 874−883.
(4) Huang, Z.; Chen, J.; Jia, Y.; Liu, H.; Xia, C.; Liu, H. Appl. Catal., B
2014, 147, 377−386.
Figure 8. Stability of the Pd−WOx(5%)/Al2O3 catalyst in the
hydrogenolysis of glucose at different values of WHSV. Reaction
conditions: 180 °C, 4 MPa H2, 5 wt % glucose aqueous solution, molar
ratio of H2 to glucose = 22.6:1.
(5) Xiao, Z.; Jin, S.; Pang, M.; Liang, C. Green Chem. 2013, 15, 891−
895.
̌ ́
(6) Fabicovicova, K.; Malter, O.; Lucas, M.; Claus, P. Green Chem.
WO4 species are present mainly on the Pd−WOx(5%)/Al2O3
catalyst. Only Lewis acid sites are present in the 0.3−1.0 W
nm−2 range provided by the alumina support and highly
dispersed WOx species. Brønsted acid sites also appear in the
2.1−3.1 W nm−2 range provided by polytunstate species. Both
the Lewis acid sites and Pd metal sites are active sites for
glucose hydrogenolysis to 1,2-PDO. The hydrogenolysis of
glucose to 1,2-PDO is strongly dependent on the synergy of the
Lewis acid sites and Pd metal sites. Pd metal catalyzes the
hydrogenation of CO in glucose and intermediates and has
weak C−C bond cleavage capability. Lewis acids can promote
the isomerization of glucose to fructose and show a slight
promoting effect on the selective conversion of fructose to 1,2-
PDO. Fructose is the key intermediate for highly selective
production of 1,2-PDO. Lewis acid is provided by highly
dispersed WOx species, such as isolated WO4 and oligomeric
WOx species, and by tetrahedrally coordinated AlO4 species.
Furthermore, highly distorted isolated WO4 species showed
excellent promoting effect for the hydrogenolysis of glucose to
1,2-PDO. By the promotion of WOx, Pd−WOx(5%)/Al2O3
showed a 60.8% of 1,2-PDO selectivity at glucose conversion of
92.2%. The catalyst also showed good stability, with a lifetime
of more than 200 h.
2014, 16, 3580−3588.
(7) Grilc, M.; Veryasov, G.; Likozar, B.; Jesih, A.; Levec, J. Appl.
Catal., B 2015, 163, 467−477.
(8) Zhang, J.; Teo, J.; Chen, X.; Asakura, H.; Tanaka, T.; Teramura,
K.; Yan, N. ACS Catal. 2014, 4, 1574−1583.
(9) Dabbawala, A. A.; Mishra, D. K.; Huber, G. W.; Hwang, J.-S. Appl.
Catal., A 2015, 492, 252−261.
(10) Ordomsky, V. V.; Sushkevich, V. L.; Schouten, J. C.; van der
Schaaf, J.; Nijhuis, T. A. J. Catal. 2013, 300, 37−46.
(11) Siankevich, S.; Savoglidis, G.; Fei, Z.; Laurenczy, G.; Alexander,
D. T. L.; Yan, N.; Dyson, P. J. J. Catal. 2014, 315, 67−74.
(12) Albonetti, S.; Lolli, A.; Morandi, V.; Migliori, A.; Lucarelli, C.;
Cavani, F. Appl. Catal., B 2015, 163, 520−530.
(13) Kirilin, A.; Warna, J.; Tokarev, A.; Murzin, D. Y. Ind. Eng. Chem.
̈
̊
Res. 2014, 53, 4580−4588.
(14) van Haasterecht, T.; Ludding, C. C. I.; de Jong, K. P.; Bitter, J.
H. J. Catal. 2014, 319, 27−35.
(15) Liu, D.; Chen, E. Y.-X. ACS Catal. 2014, 4, 1302−1310.
(16) Miranda, B. C.; Chimentao, R. J.; Santos, J. B. O.; Gispert-
̃
Guirado, F.; Llorca, J.; Medina, F.; Bonillo, F. L.; Sueiras, J. E. Appl.
Catal., B 2014, 147, 464−480.
(17) Lautenschlager, C. L.; Bockmuhl, M.; Ehrhart, G.; Kroβ, W.
̈
̈
German Patent 541,362, 1932.
(18) Lautenschlager, K. L.; Bockmuhl, M.; Ehrhart, G.; Krohs, W.
̈
̈
U.S. Patent 1,915,431, 1933.
(19) Larcher, A. W. (to Du Pont Co.) U.S. Patent 1,963,997, 1934.
(20) Rothrock, H. S. (to Du Pont Co.), U.S. Patent 2,004,135, 1935.
(21) Ruppert, A. M.; Weinberg, K.; Palkovits, R. Angew. Chem., Int.
Ed. 2012, 51, 2564−2601.
(22) Chopade, S. P.; Miller, D. J.; Jackson, J. E.; Werpy, T. A.; Frye, J.
G., Jr.; Zacher, A. H. U.S. Patent 6,291,725 B1, 2001.
(23) Werpy, T. A.; Frye, J. G., Jr.; Zacher, A. H.; Miller, D. J. U.S.
Patent 6,677,385 B2, 2004.
(24) Werpy, T. A.; Frye, J. G., Jr.; Zacher, A. H.; Miller, D. J. U.S.
Patent 6,841,085 B2, 2005.
(25) Liu, Y.; Luo, C.; Liu, H. Angew. Chem., Int. Ed. 2012, 51, 3249−
3253.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.5b00800.
■
S
Details of catalyst preparation and product analysis
methods; UV−vis DRS spectra and Eg values of the Pd−
WOx/Al2O3 catalysts and WO3 (Figure S1); NH3-TPD
patterns of the reduced catalysts and Py-FTIR pattern of
reduced Pd/SiO2 (Figure S2); The BET surface area of
the Pd−WOx/Al2O3 catalysts (Tables S1); glucose
conversion over different model catalysts and in different
atmospheres (Table S2); comparison of the structure
properties of Pd−WOx(5%)/Al2O3 before and after
reaction (Table S3) (PDF)
(26) Yepez, A.; Pineda, A.; Garcia, A.; Romero, A. A.; Luque, R. Phys.
Chem. Chem. Phys. 2013, 15, 12165−12172.
(27) Ooms, R.; Dusselier, M.; Geboers, J. A.; Op de Beeck, B.;
Verhaeven, R.; Gobechiya, E.; Martens, J. A.; Redl, A.; Sels, B. F. Green
Chem. 2014, 16, 695−707.
(28) Song, J.; Fan, H.; Ma, J.; Han, B. Green Chem. 2013, 15, 2619−
2635.
AUTHOR INFORMATION
Corresponding Authors
*E-mail: zhangchh@sxicc.ac.cn.
*E-mail: zhuyulei@sxicc.ac.cn.
■
(29) Kobayashi, H.; Fukuoka, A. Green Chem. 2013, 15, 1740−1763.
(30) Miao, G.; Zhu, C. C.; Wang, J. J.; Tan, Z. C.; Wang, L.; Liu, J. L.;
Kong, L. Z.; Sun, Y. H. Green Chem. 2015, 17, 2538−2544.
4622
DOI: 10.1021/acscatal.5b00800
ACS Catal. 2015, 5, 4612−4623

ACS Catalysis
Research Article
(31) Wang, X.; Meng, L.; Wu, F.; Jiang, Y.; Wang, L.; Mu, X. Green
Chem. 2012, 14, 758−765.
(65) Watanabe, M.; Aizawa, Y.; Iida, T.; Nishimura, R.; Inomata, H.
Appl. Catal., A 2005, 295, 150−156.
́
(66) Roman-Leshkov, Y.; Moliner, M.; Labinger, J. A.; Davis, M. E.
(32) Zhang, J.; Hou, B.; Wang, A.; Li, Z.; Wang, H.; Zhang, T. AIChE
J. 2014, 60, 3804−3813.
Angew. Chem., Int. Ed. 2010, 49, 8954−8957.
(67) Choudhary, V.; Mushrif, S. H.; Ho, C.; Anderko, A.; Nikolakis,
V.; Marinkovic, N. S.; Frenkel, A. I.; Sandler, S. I.; Vlachos, D. G. J.
Am. Chem. Soc. 2013, 135, 3997−4006.
(33) Wang, K.; Hawley, M. C.; Furney, T. D. Ind. Eng. Chem. Res.
1995, 34, 3766−3770.
(34) Zhao, G.; Zheng, M.; Zhang, J.; Wang, A.; Zhang, T. Ind. Eng.
Chem. Res. 2013, 52, 9566−9572.
(68) Liu, N.; Ding, S.; Cui, Y.; Xue, N.; Peng, L.; Guo, X.; Ding, W.
Chem. Eng. Res. Des. 2013, 91, 573−580.
(35) Zheng, M.; Wang, A.; Ji, N.; Pang, J.; Wang, X.; Zhang, T.
ChemSusChem 2010, 3, 63−66.
(69) Samain, L.; Jaworski, A.; Eden
́
, M.; Ladd, D. M.; Seo, D.-K.;
Javier Garcia-Garcia, F.; Haussermann, U. J. Solid State Chem. 2014,
̈
(36) Tai, Z.; Zhang, J.; Wang, A.; Zheng, M.; Zhang, T. Chem.
Commun. 2012, 48, 7052−7054.
217, 1−8.
(70) Dou, X.; Mohan, D.; Zhao, X.; Pittman, C. U., Jr. Chem. Eng. J.
2015, 264, 617−624.
́
(37) Sifontes Herrera, V. A.; Oladele, O.; Kordas, K.; Eranen, K.;
̈
Mikkola, J.-P.; Murzin, D. Yu.; Salmi, T. J. Chem. Technol. Biotechnol.
2011, 86, 658−668.
(38) Li, N.; Huber, G. W. J. Catal. 2010, 270, 48−59.
(39) You, S. J.; Baek, I. G.; Park, E. D. Appl. Catal., A 2013, 466,
161−168.
(40) Holm, M.; Saravanamurugan, S.; Taarning, E. Science 2010, 328,
602−605.
(41) Sasaki, M.; Goto, K.; Tajima, K.; Adschiri, T.; Arai, K. Green
Chem. 2002, 4, 285−287.
(42) Kanie, Y.; Akiyama, K.; Iwamoto, M. Catal. Today 2011, 178,
58−63.
(43) Vispute, T. P.; Huber, G. W. Green Chem. 2009, 11, 1433−1445.
(44) Zhao, C.; Kou, Y.; Lemonidou, A. A.; Li, X.; Lercher, J. A.
Angew. Chem., Int. Ed. 2009, 48, 3987−3990.
(45) Hong, Y. K.; Lee, D. W.; Eom, H. J.; Lee, K. Y. Appl. Catal., B
2014, 150−151, 438−445.
(46) Roman
1580.
́
-Leshkov, Y.; Davis, M. E. ACS Catal. 2011, 1, 1566−
(47) Ekeberg, D.; Morgenlie, S.; Stenstrøm, Y. Carbohydr. Res. 2005,
340, 373−377.
(48) Ross-Medgaarden, E. I.; Wachs, I. E. J. Phys. Chem. C 2007, 111,
15089−15099.
(49) Kim, T.; Burrows, A.; Kiely, C. J.; Wachs, I. E. J. Catal. 2007,
246, 370−381.
(50) Chen, X.; Clet, G.; Thomas, K.; Houalla, M. J. Catal. 2010, 273,
236−244.
(51) Wu, X.; Zhang, L.; Weng, D.; Liu, S.; Si, Z.; Fan, J. J. Hazard.
Mater. 2012, 225−226, 146−154.
(52) Ghosh, S.; Acharyya, S. S.; Tiwari, R.; Sarkar, B.; Singha, R. K.;
Pendem, C.; Sasaki, T.; Bal, R. ACS Catal. 2014, 4, 2169−2174.
(53) Fu, L.; Xia, T.; Zheng, Y.; Yang, J.; Wang, A.; Wang, Z. Ceram.
Int. 2015, 41, 5903−5908.
(54) Mamede, A.-S.; Leclercq, G.; Payen, E.; Granger, P.; Grimblot, J.
J. Mol. Struct. 2003, 651−653, 353−364.
(55) Moraes, R.; Thomas, K.; Thomas, S.; Van Donk, S. V.; Grasso,
G.; Gilson, J- P.; Houalla, M. J. Catal. 2012, 286, 62−77.
(56) Taylor, M. N.; Zhou, W.; Garcia, T.; Solsona, B.; Carley, A. F.;
Kiely, C. J.; Taylor, S. H. J. Catal. 2012, 285, 103−114.
(57) Contreras, J. L.; Fuentes, G. A.; García, L. A.; Salmones, J.;
Zeifert, B. J. Alloys Compd. 2009, 483, 450−452.
́
(58) García-Fernandez, S.; Gandarias, I.; Requies, J.; Guemez, M. B.;
̈
Bennici, S.; Auroux, A.; Arias, P. L. J. Catal. 2015, 323, 65−75.
(59) Karim, A. H.; Triwahyono, S.; Jalil, A. A.; Hattori, H. Appl.
Catal., A 2012, 433−434, 49−57.
(60) Onfroy, T.; Clet, G.; Houalla, M. J. Phys. Chem. B 2005, 109,
3345−3354.
(61) Baertsch, C. D.; Soled, S. L.; Iglesia, E. J. Phys. Chem. B 2001,
105, 1320−1330.
(62) Kim, T. Y.; Park, D. S.; Choi, Y.; Baek, J.; Park, J. R.; Yi, J. J.
Mater. Chem. 2012, 22, 10021−10028.
(63) Sun, D.; Yamada, Y.; Sato, S. Appl. Catal., A 2014, 487, 234−
241.
(64) Ooms, R.; Dusselier, M.; Geboers, J. A.; Op de Beeck, B.;
Verhaeven, R.; Gobechiya, E.; Martens, J. A.; Redl, A.; Sels, B. F. Green
Chem. 2014, 16, 695−707.
4623
DOI: 10.1021/acscatal.5b00800
ACS Catal. 2015, 5, 4612−4623