Mesoporous Ti-W oxide: Synthesis, characterization, and performance in selective hydrogenolysis of glycerol

Article abstract of DOI:10.1039/c3ta10217c

Mesoporous Ti-W oxides, bearing high surface area, large pore volume, uniform pore size and tunable W/Ti ratios in a wide range (10-40 mol%), were successfully fabricated via an evaporation-induced self-assembly (EISA) strategy. In this approach, the incorporation of W species not only effectively resulted in well-ordered mesoporous structures when calcined below 400 °C but also modified the acidic properties of the obtained oxide composites. The optimal acid amounts (0.47-0.67 mmol g^-1 for 400 °C calcinations, 0.25-0.27 mmol g^-1 for 500 °C calcinations) were obtained when the W concentration was between 10 and 20 mol%. When calcined at 500 °C, Bronsted acids were generated in Ti₉₀W₁₀-500 and Ti₈₀W₂₀-500. The catalytic performance of these mesoporous solid acids in glycerol hydrogenolysis was studied with a loading of 2 wt% Pt. Pt/Ti_100-nW_n-500s exhibited high selectivity to 1,3-propanediol (33.5% and 40.3%) and promising catalytic activities (18.4% and 24.2% glycerol conversion) when n is 10 and 20, respectively. This work presents a step forward in the development of highly efficient glycerol hydrogenolysis catalysts and a new understanding of the reaction mechanism of glycerol hydrogenolysis to 1,3-propanediol.

Full text of DOI:10.1039/c3ta10217c

Journal of
Materials Chemistry A
View Article Online
View Journal | View Issue
PAPER
Mesoporous Ti–W oxide: synthesis, characterization,
and performance in selective hydrogenolysis of
glycerol†
Cite this: J. Mater. Chem. A, 2013, 1,
3724
Yanhua Zhang,ab Xiao-Chen Zhao,a Yao Wang,a Likun Zhou,a Junying Zhang,a
a
a
a
*
Jia Wang, Aiqin Wang and Tao Zhang
Mesoporous Ti–W oxides, bearing high surface area, large pore volume, uniform pore size and tunable
W/Ti ratios in a wide range (10–40 mol%), were successfully fabricated via an evaporation-induced self-
assembly (EISA) strategy. In this approach, the incorporation of W species not only effectively resulted in
well-ordered mesoporous structures when calcined below 400 ꢀC but also modified the acidic properties
of the obtained oxide composites. The optimal acid amounts (0.47–0.67 mmol gꢁ1 for 400 ꢀC
calcinations, 0.25–0.27 mmol gꢁ1 for 500 ꢀC calcinations) were obtained when the W concentration was
between 10 and 20 mol%. When calcined at 500 ꢀC, Brønsted acids were generated in Ti90W10-500 and
Ti80W20-500. The catalytic performance of these mesoporous solid acids in glycerol hydrogenolysis was
studied with a loading of 2 wt% Pt. Pt/Ti100ꢁnWn-500s exhibited high selectivity to 1,3-propanediol
(33.5% and 40.3%) and promising catalytic activities (18.4% and 24.2% glycerol conversion) when n is
10 and 20, respectively. This work presents a step forward in the development of highly efficient
glycerol hydrogenolysis catalysts and a new understanding of the reaction mechanism of glycerol
hydrogenolysis to 1,3-propanediol.
Received 28th October 2012
Accepted 18th January 2013
DOI: 10.1039/c3ta10217c
www.rsc.org/MaterialsA
typically combined with oxygen to form tungsten oxide with
highly acidic nature, played specic and signicant roles in the
1 Introduction
selective conversion of biomass.15–18 The acidity of W oxide
species, which is highly dependent on their dispersion, was
proposed to play a key role in the selective production of 1,3-
PD9–13 and propylene glycol.17 Therefore, design and fabrication
of highly dispersed W containing materials is of great value in
both the materials science and the practical application of
catalytic biomass conversion.
As a consequence of increasing manufacture of biodiesel, the
production of glycerol has enormously increased at a speed of
10 wt% as the main byproduct.1 The surplus glycerol, nomi-
nated as one of the top-12 potential biomass-based chemicals
by the DoE,2 can be valorized by transformation into value-
added propanediols (PD), including 1,2-PD and 1,3-PD.3 In
particular, 1,3-PD is of great value in the fabrication of polytri-
methylene terephthalate (PTT),4,5 a kind of ber with promising
chemical resistance, light stability, elastic recovery and
dyeability. However, for steric reasons, the primary hydroxyl
group is cleaved more readily and most reports favored the
formation of 1,2-PD,3,6–8 except for the case employing catalysts
comprised of tungsten oxides9–13 or Ir–Re catalysts.14 In previous
studies, tungsten species, bearing an unsaturated d-orbital,
In general, there are two typical approaches to obtain
dispersed tungsten oxide species: conventional wetness
impregnation and a sol–gel strategy. The former method is
simple but less controllable in the tungsten oxide dispersion
and the pore structures. Besides, due to their weak interaction
with the support, the dispersed W species grievously aggregated
during reaction, and in turn limited the practical application of
these catalysts. Nevertheless, the sol–gel strategy, directly
incorporating tungsten species into metal oxide composites, is
supposed to afford a homogenous distribution of tungsten
oxide, as well as a high stability for anti-aggregation.18–22 Mes-
oporous TiO2 is widely studied due to the variety of its potential
applications.23 It was reported that the crystallization tempera-
ture of TiO2 was much lower than that of tungsten oxide,24 and
aState Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese
Academy of Sciences, Dalian 116023, China. E-mail: taozhang@dicp.ac.cn; Fax: +86
411 84691570; Tel: +86 411 84379015
bKey Laboratory of Green Chemical Technology and High Efficient Energy Saving of
Hebei Province, Hebei University of Technology, Tianjin 300130, China
† Electronic supplementary information (ESI) available: Characterization
methods, TEM images of Ti90W10-400 and Ti80W20-400 obtained without the
addition of extra HCl; wide-angle XRD patterns of mesoporous Ti100ꢁnWn-500;
nitrogen adsorption isotherms and the corresponding pore size distribution
curves for Ti100ꢁnWn-500 (n ¼ 0 to 50); NH3-TPD curves of Ti100ꢁnWn-400s. See
DOI: 10.1039/c3ta10217c
the solid solution of TiO2 and WO3 was formed over 700 C.25
ꢀ
Therefore, if uniformly isolated by titanium precursors, tung-
sten oxide species can be expelled to the boundary and the
surface of TiO2 during calcination, in the form of highly
3724 | J. Mater. Chem. A, 2013, 1, 3724–3732
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Journal of Materials Chemistry A
dispersed molecular species and clusters without intensive sol was cast into Petri dishes and exposed under ambient
interaction with TiO2 or forming a solid solution.26 On one conditions for ca. 1 h till the transparent solgradually transformed
hand, the less agglomerated and more stable tungsten oxide into gel. The as-synthesized gel was cured at 40 ꢀC for 2–4 days and
species greatly increased the surface acidity, responsible for the then calcined in owing air at 400 ꢀC for 4 h to remove the
high selectivity of 1,3-PD.16 Moreover, the incorporation of Ti surfactant, with a ramping rate of 1 C minꢁ1, and followed by
ꢀ
and W will strengthen the mechanical and chemical stability of calcinations at higher temperature, if designated, for another 1 h,
the resultant mesoporous composites, and accordingly will with a ramping rate of 5 C minꢁ1. By varying the amounts of
ꢀ
provide promising catalyst supports for glycerol hydrogenolysis. Ti(OBu)4 and WCl6, mesoporous Ti–W oxides with different
In this work, we developed a facile method to synthesize compositionswerefabricatedanddenotedasTi100ꢁnWn-T, wheren
mesoporous Ti–W oxides with a wide range of tungsten intro- and T refer to the mol% of W in Ti–W composites and the calci-
duction and investigated how tungsten oxide species impact on nation temperature, respectively. It should be noted that, in order
the structure, the chemistry, and as a consequence the catalytic to moderate the hydrolysis rate, 0.6 g of hydrochloric acid (37 wt%)
performance of the resultant composites. Though mesoporous was added to the solutions together with the inorganic precursors,
Ti–W oxides have been obtained before, the content of W incor- when synthesizing Ti100W0, Ti90W10 and Ti80W20 oxides.
poration is extremely lowandtheordered mesoscopicstructure is
difficult to maintain.26–28 In Pan's protocols, they used 2,4-pen-
2.3 Catalyst preparation
tadione stabilized tungsten(V) pentaethoxide as a W precursor
The catalysts (Pt/Ti100ꢁnWn-T) were prepared by impregnating
and synthesized a series of ordered cubic mesoporous Ti–W
Ti100ꢁnWn-500s with aqueous solution of H2PtCl6 and then
oxides, however, the maximum W introduction is only 10 mol%.
calcined at 400 ꢀC for 1 h with a ramping rate of 5 ꢀC minꢁ1. The
Besides, to preserve the fragile ordered meso-structure, a relative
loading amount of Pt is stated as 2 wt%. Prior to the glycerol
humidity of 60% was required in closed chambers during the
hydrogenolysis reaction, the catalysts were reduced in owing
aging period.26 In contrast, our strategy is more facile and more
H2 (60 ml minꢁ1) at 300 ꢀC for 1 h, and passivated with a ow of
efficient to handle, i.e. by selecting tungsten chloride (WCl6) as
1% O2/N2 for 4 h at room temperature.
the W precursor and controlling the addition of HCl, we easily
fabricated 2D ordered mesoporous Ti–W oxides with a wide range
of W incorporation from 0 to 40 mol%. To the best of our
2.4 Characterization
The synthesized mesoporous Ti–W oxides and catalysts (Pt/
Ti100ꢁnWn-T) were characterised by nitrogen adsorption–desorp-
tion measurements, powder X-ray diffraction (XRD), trans-
mission electron microscopy (TEM), thermogravimetric analysis
(TGA) and differential thermal analysis (DTA), temperature-pro-
grammed desorption of NH3 (NH3-TPD), pyridine adsorption
Fourier transform infrared (FT-IR) spectroscopy, temperature-
programmed reduction (TPR), H2 and CO chemisorption uptakes
and electron paramagnetic resonance (EPR) spectra. These
methods are described in detail in the ESI†.
knowledge, this is the rst report on successful synthesis of well-
ordered Ti–W oxides with such a wide Ti/W ratio. The catalytic
performances of the resultant materials were evaluated on glyc-
erol hydrogenolysis when supporting 2 wt% Pt species. Employ-
ing such sol–gel derived mesoporous W-containing composites
as catalyst supports has never been reported in the application of
glycerol hydrogenolysis before. In our research, the mesoporous
Ti–W oxides exhibited promising catalytic activity and 1,3-PD
selectivity, and the optimal performance was achieved when an
appropriate amount (10–20 mol%) of tungsten was incorporated.
2.5 Catalytic hydrogenolysis of glycerol
2 Experimental section
2.1 Chemicals
Catalytic hydrogenolysis of glycerol was carried out in a stain-
less autoclave (Parr Instrument Company, 100 ml) typically at
5.5 MPa H2 pressure (measured at room temperature) and
180 ꢀC for 12 h, stirring at 700 rpm. For each reaction, 1.0 g of
catalyst and 40 ml of glycerol aqueous solution (10 wt%) were
put into the autoclave, which then underwent several cycles of
ushing with H2 ow to drive off the air in the autoclave. Aer
the reaction, the gas products and liquid products were
collected separately. The liquid-phase products were analyzed
by high-performance liquid chromatography (HPLC) equipped
with a Sugar SC 1011 column (Shodex). The conversion of
glycerol and the selectivity of each liquid product were calcu-
lated by the following equations:
Amphiphilic poly(ethylene oxide)-b-poly(propylene oxide)-b-
poly(ethylene oxide) triblock copolymer Pluronic P123 (Mw
¼
5800, EO20PO70EO20) and tungsten hexachloride (WCl6, 99.9%)
were purchased from Sigma Aldrich Corp. Titanium n-butoxide
(Ti(OBu)4, 99%) was provided by J&K Chemical Corp. HCl (37
wt%) and n-propanol were all supplied by chemical plants in
China. All chemicals were used as received without any further
purication. Deionized water was used in all experiments.
2.2 Preparation of mesoporous Ti–W mixed oxides
Mesoporous Ti–W oxides were synthesized via a modied
evaporation-induced self-assembly (EISA) approach, employing
n-propanol as the solvent. With a total Ti–W molar of 9 mmol,
Ti(OBu)4 and WCl6 were rst dissolved in a solution containing
10 g of n-propanol and 1 g of P123, and kept stirring for over 2 h at
room temperature in a sealed beaker. The obtained homogeneous
conversion of glycerol (%) ¼ (mole of glycerol consumed)/
(mole of glycerol initially added) ꢂ 100
selectivity (%) ¼ (mole of carbon in specific product)/(mole of
carbon in consumed glycerol) ꢂ 100
This journal is ª The Royal Society of Chemistry 2013
J. Mater. Chem. A, 2013, 1, 3724–3732 | 3725

View Article Online
Journal of Materials Chemistry A
Paper
wormlike mesoporous structures, accordingto Grosso's previous
results.29 Moreover, further increasing the W content to above
20 mol%, the crystal phase of anatase was completely swallowed,
and instead, the amorphous composite of Ti–W oxides was
engendered. This implied that the presence of W species
somehow retarded the crystallization of TiO2 during calcinations
and hence beneted the preservation of a well-ordered structure,
consistent with Fig. 1a and previous studies.30–33
3 Results and discussions
3.1 Structures of mesoporous Ti100ꢁnWn-400 oxides
Fig. 1a presents the low-angle XRD patterns of Ti100ꢁnWn-400s
(n ¼ 0–50). When the W content is between 10 and 40 mol%, a
distinct peak at around 1.0ꢀ can be observed which is indexed as
(100) reections of the 2D p6mm hexagonal meso-structure.
Besides, another weak diffraction peak at 2q ¼ 1.5–2.5ꢀ indi-
cates a long range mesoscopic order of Ti100ꢁnWn-400s, when n
is between 10 and 40. However, further increasing the W
content to 50 mol%, no obvious diffraction peak was observed
in a low-angle range and suggested that the overdose W incor-
poration destroyed the well-ordered mesoporous structure. In
contrast to the W-containing Ti–W oxides, only a very weak and
broad peak was identied for the W-free TiO2, indicating its less
ordering of the mesoporous structure. Therefore, the intro-
duction of an appropriate amount of W is benecial for main-
taining the well-ordered mesoporous structure.
Fig. 1b shows the wide-angle XRD patterns of Ti100ꢁnWn-400s
(n ¼ 0–50). Though both Ti100W0-400 and Ti90W10-400 displayed
a typical anatase crystalline phase (JCPDS Card no. 01-083-2243),
the half width of the (101) peak for Ti90W10-400 was much
broader than that of Ti100W0-400. According to the Scherrer
formula, the nanocrystal sizes were estimated to be 9 nm and 6
nm for Ti100W0-400 and Ti90W10-400, respectively. Compared
with Ti90W10-400, the W-free TiO2 exhibited a well-crystallized
anatase phase, however, sacriced the long-range mesoporous
order. This was attributed to the excessive crystallization by
thermal treatment, which tended to result in collapsed or
As aforementioned, one of the key factors to successfully
obtain a specic mesoporous metal oxide is the stability of
inorganic frameworks upon calcinations for surfactant
removals. Besides, the inceptive acidity of the nonaqueous
system is also an essential factor in preparing mesoporous
metal oxides and it is reported that a highly ordered meso-
structure was formed at a pH value of 1–2.34 In our work, dis-
solving WCl6 in n-propanol generated large amounts of HCl,
thus the acidity of the resultant solution can be effectively
mediated by varying the proportion of WCl6 added. However,
when the W concentration is below 20 mol%, the amount of
HCl generated from WCl6 dissolution is insufficient, thus only
wormlike meso-structures can be obtained in Ti90W10-400 and
Ti80W20-400 (as shown in Fig. S1†). In contrast, hydrochloric
acid was rationally added to make up the acid deciency and
consequently resulted in well-ordered mesoporous Ti90W10-400
and Ti80W20-400, implying an essential role of acidity in the
architecture of the structure. According to the previous study,
the amount of chloric acid inuenced not only the hydrolysis
process but also the interaction between surfactants and inor-
ganic precursors.35–38 Therefore, in this regard, the addition of
extra HCl on one hand suppressed the hydrolysis rate of Ti–W
species, and on the other hand protonated the inorganic
precursors and enhanced its affinity to the hydrophilic blocks
of P123; and in turn collaboratively helped the formation of
arranged mesoscale pores in Ti–W oxides.35 In the case of
Ti70W30-400 and Ti60W40-400, the high proportion of W species
provided sufficient acid to form and maintain ordered meso-
structures, thus addition of HCl is no longer necessary when
the W content is above 30 mol%. However, the absence of the
ordered mesostructure in Ti50W50-400 may be related to the
excessive incorporation of W species, as well as the large
amount of HCl generated, which induced destructive cross-
linkage of the inorganic framework and interrupted periodic
assembly of surfactant micelles.34
The textural properties of these Ti–W oxides were analyzed
by N2 adsorption–desorption isotherms (Fig. 2a). Ti100ꢁnWn-
400s (n ¼ 0 to 50) showed typical type IV isotherms with sharp
capillary condensation steps at a relative pressure (p/p0) of 0.4–
0.8, representative of mesoporous materials. The average pore
size of all these samples was narrowly distributed at around
3–5 nm (Fig. 2b), implying a uniform mesopores system in the
obtained materials. The detailed textual parameters are
summarized in Table 1. Compared with W-free TiO2, a small
amount of W (Ti90W10-400) inuenced little on the resultant
pore volume and surface area. However, due to the large density
of tungsten oxide (7.16 g cmꢁ3) and high proportion of W
introduction, the pore volume and the surface area of
Fig. 1 (a) Low- and (b) wide-angle XRD patterns of mesoporous Ti100ꢁnWn-400. Ti100ꢁnWn-400s decreased from 0.25 cm3 gꢁ1 to 0.11 cm3 gꢁ1
3726 | J. Mater. Chem. A, 2013, 1, 3724–3732
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Journal of Materials Chemistry A
(n ¼ 10–40) show typical morphologies of the ordered 2D
hexagonal mesostructure with uniform channels in large
domains (as shown in Fig. 3b–e), which corresponds well with
the XRD and N2 adsorption results. However, when beyond the
appropriate range of W incorporation, the well-structured
mesopores of Ti100ꢁnWn-400s were destroyed or distorted into
wormlike ones, as exemplied by W-free Ti100W0-400 (Fig. 3a)
and W-excess Ti50W50-400 (Fig. 3f).
3.2 Thermal stability
To monitor the calcination process and evaluate the thermal
stability of the obtained Ti–W oxides, TG and DTA proles (Fig. 4)
were acquired on the as-prepared Ti100ꢁnWn gel and Ti100ꢁnWn-
400s in the air atmosphere. The TG curves of the as-prepared
Ti70W30 and Ti60W40 gel, as displayed in Fig. 4a, comprised three
obvious stages of weight loss during the temperature increase
from 25 to 800 ꢀC. The rst weight loss occurred at around 150 ꢀC
and was designated to the loss of physically adsorbed species
such as water and other volatile species (n-propanol, butanol,
and/or HCl). The following sharp weight loss (ca. 33 wt%), along
with drastic exothermic signals, happened between 200 ꢀC and
260 ꢀC, indicating the decomposition of the Pluronic P123
template. The last weight loss lasted from 300 ꢀC to 450 ꢀC arose
from the dehydroxylation of OHꢁ groups and crystallization of
the inorganic framework. Besides, though no obvious weight
loss was observed over 500 ꢀC, exothermic signals were detected.
Combined with the enlarged TG spectra of Ti100ꢁnWn-400
(Fig. 4b), these peaks were also found accompanied by slight
weight loss, reecting the release of relatively stable groups. In
Fig. 4b, distinct exothermic peaks generated as W increased,
and the correlated weight loss temperature increased along with
the W content, implying that a large amount of W introduction,
per contra, also suppressed this weight losing process. With the
help of on-line mass spectrometer signals in NH3-TPD
measurements (shown in Fig. S2†), the weight loss along with
the sharp exothermic peak over 500 ꢀC was assigned to the
release of carboxylate species, which are strongly interacted
with tungsten oxides. Similar results have been reported by
Okumura et al.39
Fig. 2 (a) Nitrogen adsorption isotherms and (b) the corresponding PSD curves
for Ti100ꢁnWn-400 (n ¼ 0 to 50). Traces in adsorption isotherms are vertically
shifted for clarity.
and from 155 m2 gꢁ1 to 78 m2 gꢁ1, respectively, along with the
increase of W proportion from 10 to 40 mol%. Besides, the
increase in the average pore size of Ti–W oxides aer W incor-
poration demonstrated that the presence of tungsten oxide
might inhibit the shrinkage of the mesoporous framework
during calcinations.
TEM images of mesoporous Ti–W oxides with a variable Ti/W
ratio are displayed in Fig. 3. It is apparent that Ti100ꢁnWn-400s
Table 1 Textural parameters and acid concentrations of Ti100ꢁnWn-400 (n ¼ 0 to 50)
SBETa (m2 gꢁ1
)
VPb (cm3 gꢁ1
)
DPb (nm)
d100 (nm)
a0d (nm)
Acid concentratione (mmol gꢁ1
)
c
Ti100ꢁnWn
Ti100W0-400
Ti90W10-400
Ti80W20-400
Ti70W30-400
Ti60W40-400
Ti50W50-400
Ti100W0-500
Ti90W10-500
Ti80W20-500
Ti50W50-500
172
178
155
111
78
0.24
0.25
0.20
0.14
0.11
0.14
0.19
0.13
0.19
0.11
3.0
3.9
3.7
3.1
4.1
3.6
3.7
2.5
3.6
4.6
—
—
0.28
0.67
0.47
0.47
0.32
0.37
0.09
0.25
0.27
0.19
7.8
8.2
7.2
7.4
—
—
—
—
—
9.0
9.5
8.3
8.5
—
—
—
—
—
99
115
128
123
60
a
SBET is the surface area calculated by the Brunauer–Emmett–Teller (BET) method. b Mesopores volume (Vp) and the average pore diameter (Dp)
c
were estimated with adsorption branches of the isotherms using the Barrett–Joyner–Halenda (BJH) method. d-spacing is calculated from the
d
e
(100) diffraction peak in low-angle XRD patterns. a0, unit cell parameter calculated as a0 ¼ 2 ꢂ d(100)/O3. Acid concentrations were
calculated according to the amount of adsorbed NH3.
This journal is ª The Royal Society of Chemistry 2013
J. Mater. Chem. A, 2013, 1, 3724–3732 | 3727

View Article Online
Journal of Materials Chemistry A
Paper
Fig. 4 TG (solid) and DTA (short dash) curves for the as-prepared Ti100ꢁnWn gels
(a) and Ti100ꢁnWn-400s (b).
Fig. 3 TEM images of (a) Ti100W0-400, (b) Ti90W10-400, (c) Ti80W20-400, (d)
Ti70W30-400, (e) Ti60W40-400, and (f) Ti50W50-400, and the fast Fourier transform
(FFT) diffractograms (inset of c).
oxides. As listed in Table 1, with a calcination temperature of
400 ꢀC, the acid concentration of Ti90W10-400 greatly increased
from 0.28 to 0.67 mmol gꢁ1 aer W incorporation. Though
To investigate the stabilities of the pore ordering and the bearing more acid sites than the W-free one, the acid concen-
crystal structure against temperature, Ti–W oxides were treated tration of the mesoporous Ti–W oxides decreased from 0.67
at higher temperature and both low- and wide-angle XRDs were back to 0.32 mmol gꢁ1 as the proportion of W increased from 10
conducted. Aer calcination at 500 ꢀC, amorphous Ti80W20-400 to 40 mol%. According to the previous study,26,40 the increased
(Fig. 1b) was crystallized into a composition of anatase and acid concentrations could be ascribed to the enrichment of
rutile phase (Fig. S3†), while Ti50W50-500 remained amorphous hydroxyl groups or the enhanced polarizability.
due to the ubiquity of W species. Meanwhile, no reection peak
Besides the amount of acid sites, the acid strength was also
was detected in low-angle XRD patterns, illustrating that the evaluated via NH3-TPD based on desorption temperatures. In
well-ordered mesoporous structures were seriously degraded Fig. 5a, the acid intensity increased drastically aer W intro-
due to the crystal growth and the phase transition. Nevertheless, duction, especially the weak acid indicated by NH3 desorption
these Ti100ꢁnWn-500s still presented typical type IV isotherms at low temperature (150 to 300 ꢀC). Surprisingly, additional
ꢀ
with well-dened H1 hysteresis loops and narrow pore size sharp peaks were detected over 500 C when increasing the W
distributions (Fig. S4†), indicating that uniform mesopores content to over 20 percent in Ti100ꢁnWn-400s (details in Fig. 5a
were retained. According to the textural parameters listed in and S5†). Correlated with the TGA results, these peaks origi-
Table 1, the surface area and pore volume of Ti100ꢁnWn-500 had nated from the fragment of CO2 due to the decomposition of
some decrease compared with their counterparts calcined at carboxylate species, rather than indicated the character of acid
ꢀ
400 C, but in general, maintained at a high standard, e.g. the sites. To rule out the inuences of the caꢀrbon residue, the
surface area and the pore volume of Ti90W10-500 and Ti80W20
500 are still over 120 m2 gꢁ1 and 0.13 cm3 gꢁ1, respectively.
-
Ti100ꢁnWn-400s were further calcined at 500 C for another 1 h
and nominated as Ti100ꢁnWn-500s. As shown in Fig. 5b, the
fragments of CO2 disappeared over Ti100ꢁnWn-500s, indicating
a complete decomposition of carbon residues. However, the
pure surface was obtained at the cost of total acid amounts.
3.3 Acidic properties
To study the acidic properties, NH3-TPD and pyridine adsorp- When increasing the calcination temperature from 400 to
tion FT-IR were employed on Ti100ꢁnWn-Ts (n ¼ 0, 10, 20, 50; T ¼ 500 ꢀC, the acid concentration of Ti100ꢁnWn-T (n ¼ 10, 20)
400, 500), as representatives of the obtained mesoporous Ti–W decreased from ca. 0.47 mmol gꢁ1 to 0.25 mmol gꢁ1
3728 | J. Mater. Chem. A, 2013, 1, 3724–3732
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Journal of Materials Chemistry A
Fig. 6 Pyridine adsorption FT-IR spectra of Ti100ꢁnWn-500 (n ¼ 0, 10, 20 and 50).
3.4 Hydrogenolysis of glycerol
In the present work, Pt was supported on the mesoporous Ti–W
oxides and the resultant catalysts were evaluated for glycerol
hydrogenolysis. Considering the properties of high surface
areas, crystallized pore walls, high porosities, relative high acid
amounts, as well as high thermal stabilities, Ti90W10-500 and
Ti80W20-500 were selected as the catalyst supports, while
Ti100W0-500 and Ti50W50-500 were employed as their counter-
parts. Table 2 summarizes the activity and selectivity results of
Pt/Ti100ꢁnWn-500s for glycerol hydrogenolysis. The predomi-
nant products were 1,3-PD, 1,2-PD and 1-propanol. Over the W-
free support, Pt/Ti100W0-500 exhibited a low conversion of
glycerol (7.8%), and was quite selective to 1,2-PD (83%) rather
Fig. 5 NH3-TPD (m/e ¼ 16) curves of Ti100ꢁnWn-400s (a) and Ti100ꢁnWn-500s (b)
(n ¼ 0, 10, 20 and 50).
substantially (listed in Table 1). This was supposed due to the than to 1,3-PD (7.2%). Over the W-overdose support, Pt/Ti50W50-
reduction of the surface area and the covering of acid sites due 500, in contrast, presented a high selectivity of 40.9% to 1,3-PD
to the blocked pores by the agglomerate particles.41 Besides, with a ratio of 1,3-PD/1,2-PD as high as 5.7, but the conversion is
during the severe pyrolysis, most of the weak acids were still very low (6.7%). Nevertheless, when employing Pt/Ti90W10-
removed (shown in Fig. 5b), which also accounted for the 500 or Pt/Ti80W20-500 as the catalyst, both the conversion of
reduction of acid amounts.
glycerol (18.4% and 24.2%, respectively) and the selectivity of
Fig. 6 showed the pyridine adsorption FT-IR spectra of 1,3-PD (40.3% and 33.5%, respectively) were signicantly
Ti100ꢁnWn-500 and the peaks at 1537 cmꢁ1 indicated the pres- enhanced.
ence of Brønsted acid sites in Ti90W10-500 and Ti80W20-500.
To date, one of the best-performance solid catalysts for
´
´
Cortes-Jacome et al. reported that, for ZrO2 supported WO3 solid selective hydrogenolysis of glycerol into 1,3-PD in an aqueous
acid catalysts, Brønsted acid sites were formed in the most system has been reported by Gong et al.13 In their report, a
exposed or well-dispersed W6+ atoms.42 Therefore, in the case of glycerol conversion of 15.3% and a 1,3-PD yield of 7.7% were
Ti–W oxides, we suggested that WOx species were formed or obtained over SiO2-supported Pt/WO3/TiO2 catalysts, employing
expelled to the surface when calcined at 500 ꢀC, and this led to the same reaction conditions as we do except for a duple
the formation of well-dispersed WOx clusters bearing Brønsted amount of catalysts. Therefore, to compare with their work, we
acid centers. Further increase of W proportion to 50 mol%, doubled the amount of Pt/Ti80W20-500 and conducted the
however, resulted in re-disappearance of the Brønsted acid site. reaction under exactly the same conditions. Excitingly, a much
This is probably due to the agglomeration of overdose W better glycerol conversion (55.4%) and a doubled yield of 1,3-PD
introduction, which hence restrain the formation of Brønsted (14.9%) were achieved on 2.0 g of Pt/Ti80W20-500, as listed in
acid sites. On the other hand, according to Tanabe's hypothesis Table 2.
which was supported by numerous reports,43–45 the titania-rich
Associated with the above characterizations, the presence of
Ti–W oxide was speculated to exhibit Lewis acid sites due to the strong Brønsted acid sites, when incorporating an appropriate
excess of positive charges over the W–O bond.46 In this regard, amount of W, i.e. 10–20 mol% (Fig. 6b), was speculated to be
the introduction of W species also enhances the strong Lewis responsible for the superior performance in selective
acid peaks centered at 1445 cmꢁ1 and 1608 cmꢁ1, in good hydrogenolysis of glycerol into 1,3-PD. To clarify and conrm
agreement with previous reports.46,47
the catalytic mechanism, a series of physical and chemical
This journal is ª The Royal Society of Chemistry 2013
J. Mater. Chem. A, 2013, 1, 3724–3732 | 3729

View Article Online
Journal of Materials Chemistry A
Paper
Table 2 Summary of glycerol hydrogenolysis results over different catalysts in aqueous mediaa
Selectivity (%)
Conversion (%)
1,3-PD
1,2-PD
1-PrOH
Othersc
1,3-PD/1,2-PD
Yield (%) 1,3-PD
Pt/Ti100W0
Pt/Ti90W10
Pt/Ti80W20
Pt/Ti50W50
Pt/Ti80W20
7.8
18.4
24.2
6.7
7.2
40.3
33.5
40.9
26.9
83.0
13.7
11.6
7.2
9.6
32.5
36.6
44.3
37.1
0.1
13.5
18.3
7.6
0.1
2.9
2.9
5.7
2.3
0.6
7.4
8.1
2.7
14.9
b
55.4
11.8
24.2
a
Reaction conditions: catalyst (1.0 g), 40 ml of glycerol aqueous solution (10 wt%), and H2 (5.5 MPa) at 180 ꢀC for 12 h. b 2 g of the catalyst was used.
Others include ethylene glycol, ethanol, methanol, methane, ethane and propane.
c
characterizations were conducted on the catalysts. Firstly, H2 function of W introduction. With regard to Pt/Ti80W20-500, this
and CO site titrations were conducted over Pt/Ti80W20-500 (aer reduction temperature was as low as 11 ꢀC. Besides lowered
H2 reduction) to evaluate Pt dispersion. The results showed that reduction temperatures, broad shoulder peaks were also found
the uptakes of hydrogen (91.6 mmol gcatꢁ1) substantially excee- around 100 ꢀC aer W incorporation, indicating that strong
ded those of CO (35.2 mmol gcatꢁ1), and suggested that hydrogen interactions were formed between Pt species and supports.
ad-atoms spillover onto WOx clusters and generated, as well as Thus, the reduction behavior was signicantly inuenced by
maintained, Brønsted acid sites on the catalysts under reaction varying the support compositions. The H2 consumptions were
conditions.48 This is in good agreement with and further calculated to be stoichiometric to reduce PtO2 into Pt. Besides
intensied the previous report that the addition of solid acid the reduction of Pt, another distinct H2 consumption was
can improve glycerol conversion.49
observed on Pt/Ti50W50-500 at ca. 480 ꢀC. This peak was
XRD patterns of Pt/Ti100ꢁnWn are presented in Fig. 7. Besides assigned to the reduction of large amorphous WO3 particles
a tiny peak at 2q ¼ 39.88ꢀ on Pt/Ti50W50-500, no obvious comprising of Ti50W50-500,48 consistent with the above analysis
metallic Pt peak was detected on the other Pt/Ti–W oxides, of Ti–W oxide supports.
suggesting that Pt species were highly dispersed on Pt/
With promising glycerol conversion and 1,3-PD selectivity,
Ti100ꢁnWn-500s when n is below 50. Meanwhile, in the TEM Pt/Ti80W20-500 was selected as representative to conduct EPR in
analysis (Fig. 8), Pt species with a particle size of ca. 1 nm were the calcined, as-prepared and used forms. In Fig. 10, no signal
highly dispersed on all the Ti–W oxides and no large Pt nano- of tungsten can be found in the calcined and as-prepared
particles were observed in the detected domains, which corre- passivated Pt/Ti80W20-500s. However, the used catalyst exhibi-
sponded well with the XRD results (Fig. 7). The high dispersion ted an EPR signal of W5+ species (g ¼ 1.73)50–52 aer the rst run
of Pt nanoparticles is considered as one of the reasons for the of glycerol hydrogenolysis. The intense peak with a g-value of
high activity in the hydrogenolysis of glycerol.
H2-TPR (Fig. 9) and EPR (Fig. 10) measurements were per-
formed to study the chemistry of the catalysts. As shown in
Fig. 9, all catalysts exhibited a sharp reduction peak at low
temperature (below 100 ꢀC) and this temperature decreased as a
Fig. 7 XRD patterns of the supported Pt catalysts (#) anatase; (*) rutile of TiO2;
Fig. 8 TEM images of (A) Pt/Ti100W0-500, (B) Pt/Ti90W10-500, (C) Pt/Ti80W20-500
and (+) Pt.
and (D) Pt/Ti50W50-500.
3730 | J. Mater. Chem. A, 2013, 1, 3724–3732
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Journal of Materials Chemistry A
4 Conclusions
Mesoporous Ti–W oxides (Ti100ꢁnWn-T) with a wide tungsten
content range (n ¼ 10–50) were prepared via an EISA strategy
with the help of a sol–gel reaction. The obtained mesoporous
materials were of high surface area, uniform pore size, as well as
large pore volume. When calcined at 400 ꢀC, the mesostructure
of Ti100ꢁnWn-400 was highly ordered when n is below 50, while
when calcined at 500 ꢀC, the ordering was destroyed but the
high porosity and uniform mesopores were retained. Compared
with W-free TiO2, the incorporation of tungsten greatly
increased the surface acid amounts, thus the obtained meso-
porous Ti–W oxides were considered as typical solid acids. The
optimal W proportion for acidic properties is between 10 and 30
mol%. Besides, Brønsted acid sites were generated on Ti90W10-
500 and Ti80W20-500 aer being annealed at 500 ꢀC. Employing
the obtained materials as catalyst supports, Pt/Ti100ꢁnWn-500s
were used in glycerol hydrogenolysis and exhibited promising
catalytic activity (18.4–24.2% glycerol conversion) and high
selectivity to 1,3-propanediol (33.5–40.3%) when n is between 10
and 20. The excellent performance was attributed to the pres-
ence of a large amount of acid sites, in particular the Brønsted
acid site, and the strong synergy between the well-isolated WOx
species and the highly dispersed Pt nanoparticles.
Fig. 9 Temperature programmed reduction profiles of the catalysts.
Acknowledgements
This work was nancially supported by Natural Science Foun-
dation of China (NSFC no. 21176235).
Notes and references
1 M. Pagliaro, R. Ciriminna, H. Kimura, M. Rossi and
C. D. Pina, Angew. Chem., Int. Ed., 2007, 46, 4434.
2 T. Werpy and G. Petersen, in Top Value Added Chemicals from
Biomass. Results of Screening for Potential Candidates from
Sugars and Synthesis Gas, US Department of Energy, 2004,
vol. 1.
Fig. 10 EPR spectra at ꢁ173 ꢀC for calcined Pt/Ti80W20-500 (a), as-prepared Pt/
Ti80W20-500 catalyst (b) and the used Pt/Ti80W20-500 catalyst aftera reaction run (c).
ꢁ
2.00 is characteristic of adsorbed O2 ions which are probably
produced via the following electron transfer reaction:53
3 A. M. Ruppert, K. Weinberg and R. Palkovits, Angew. Chem.,
Int. Ed., 2012, 51, 2564.
ꢁ
W5+ + O2 / W6+ + O2
4 Y. Nakagawa, Y. Shinmi, S. Koso and K. Tomishige, J. Catal.,
2010, 272, 191.
We infer that part of tungsten(VI) oxide is in situ reduced into
W5+ species by the collective reducing power of glycerol and H2
during the reaction. Aer the reaction, the used catalyst was
exposed to the atmosphere and adsorbed oxygen, which resul-
5 J. Oh, S. Dash and H. Lee, Green Chem., 2011, 13, 2004.
6 C. H. Zhou, J. N. Beltramini, Y. X. Fan and G. Q. Lu, Chem.
Soc. Rev., 2008, 37, 527.
7 T. Miyazawa, S. Koso, K. Kunimori and K. Tomishige, Appl.
Catal., A, 2007, 318, 244.
8 J. X. Zhou, L. Y. Guo, X. W. Guo, J. B. Mao and S. G. Zhang,
Green Chem., 2010, 12, 1835.
9 J. Chaminand, L. Djakovitch, P. Gallezot, P. Marion, C. Pinel
and C. Rosier, Green Chem., 2004, 6, 359.
ꢁ
ted in the formation of O2 and made the intensity of the W5+
signal decrease.53 The formation of O2ꢁ by oxygen adsorption at
RT gave an indirect evidence of W5+ species.51
This result suggests that the redox of W6+/W5+ may play an
important role in the selective hydrogenolysis of glycerol to 1,3-
PD. Further study on the reaction mechanism is still ongoing, 10 T. Kurosaka, H. Maruyama, I. Naribayashi and Y. Sasaki,
however, we can propose that the dispersion of tungsten oxide Catal. Commun., 2008, 9, 1360.
in the Ti–W composite oxide affects signicantly the dispersion 11 L. Z. Qin, M. J. Song and C. L. Chen, Green Chem., 2010, 12,
of Pt nanoparticles as well as the interplay between Pt and WO3, 1466.
which in turn affects the redox of W6+/W5+ and nally the 12 S. Zhu, Y. Zhu, S. Hao, L. Chen, B. Zhang and Y. Li, Catal.
activity and selectivity.
Lett., 2011, 142, 267.
This journal is ª The Royal Society of Chemistry 2013
J. Mater. Chem. A, 2013, 1, 3724–3732 | 3731

View Article Online
Journal of Materials Chemistry A
Paper
13 L. F. Gong, Y. Lu, Y. J. Ding, R. H. Lin, J. W. Li, W. D. Dong, 33 J. Y. Kim, S. H. Kang, H. S. Kim and Y. E. Sung, Langmuir,
T. Wang and W. M. Chen, Appl. Catal., A, 2010, 390, 119. 2010, 26, 2864.
14 Y. Nakagawa, Y. Shinmi, S. Koso and K. Tomishige, J. Catal., 34 B. Z. Tian, H. F. Yang, X. Y. Liu, S. H. Xie, C. Z. Yu, J. Fan,
2010, 272, 191. B. Tu and D. Y. Zhao, Chem. Commun., 2002, 1824.
15 N. Ji, T. Zhang, M. Zheng, A. Wang, H. Wang, X. Wang and 35 J. N. Kondo, T. Yamashita, K. Nakajima, D. Lu, M. Hara and
J. G. Chen, Angew. Chem., Int. Ed., 2008, 47, 8510. K. Domen, J. Mater. Chem., 2005, 15, 2035.
16 Z. Tai, J. Zhang, A. Wang, M. Zheng and T. Zhang, Chem. 36 J. Fan, S. W. Boettcher and G. D. Stucky, Chem. Mater., 2006,
Commun., 2012, 48, 7052. 18, 6391.
17 Y. Liu, C. Luo and H. Liu, Angew. Chem., Int. Ed., 2012, 51, 37 A. Firouzi, D. Kumar, L. M. Bull, T. Besier, P. Sieger, Q. Huo,
3249.
S. A. Walker, J. A. Zasadzinski, C. Glinka, J. Nicol,
D. Margolese, G. D. Stucky and B. F. Chmelka, Science,
1997, 267, 1138.
18 C. Tagusagawa, A. Takagaki, A. Iguchi, K. Takanabe,
J. N. Kondo, K. Ebitani, S. Hayashi, T. Tatsumi and
K. Domen, Angew. Chem., Int. Ed., 2010, 49, 1128.
19 C. Tagusagawa, A. Takagaki, A. Iguchi, K. Takanabe,
38 X. Zhao, A. Wang, J. Yan, G. Sun, L. Sun and T. Zhang, Chem.
Mater., 2010, 22, 5463.
J. N. Kondo, K. Ebitani, T. Tatsumi and K. Domen, Chem. 39 K. Okumura, K. Yamashita, M. Hirano and M. Niwa, J. Catal.,
Mater., 2010, 22, 3072. 2005, 234, 300.
20 Y. X. Rao, J. J. Kang and D. Antonelli, J. Am. Chem. Soc., 2008, 40 J. Papp, S. Soled, K. Dwight and A. Wold, Chem. Mater., 1994,
130, 394. 6, 496.
21 K. Nakajima, T. Fukui, H. Kato, M. Kitano, J. N. Kondo, 41 Z. Ma, W. M. Hua, Y. Tang and Z. Gao, J. Mol. Catal. A: Chem.,
S. Hayashi and M. Hara, Chem. Mater., 2010, 22, 3332. 2000, 159, 335.
22 Y. Zhang, A. Wang and T. Zhang, Chem. Commun., 2010, 46, 42 M. A. Cortes-Jacome, C. Angeles-Chavez, E. Lopez-Salinas,
´
´
´
862.
J. Navarrete, P. Toribio and J. A. Toledo, Appl. Catal., A,
2007, 318, 178.
43 S. Li, H. Zhou, B. Han, F. Deng, X. N. Liu, L. P. Xiao and
J. Fan, Catal. Sci. Technol., 2012, 2, 719.
44 J. Ren, Z. Li, S. S. Liu, Y. L. Xing and K. C. Xie, Catal. Lett.,
2008, 124, 185.
45 B. Bonelli, M. Cozzolino, R. Tesser, M. Di Serio, M. Piumetti,
E. Garrone and E. Santacesaria, J. Catal., 2007, 246, 293.
23 J. C. Yu, X. C. Wang and X. Z. Fu, Chem. Mater., 2004, 16,
1523.
24 A. Rampaul, I. P. Parkin, S. A. O'Neill, J. DeSouza, A. Mills
and N. Elliott, Polyhedron, 2003, 22, 35.
25 S. Komornicki, M. Radecka and P. Sobaoe, Mater. Res. Bull.,
2004, 39, 2007, and references therein.
26 J. H. Pan and W. I. Lee, Chem. Mater., 2006, 18, 847.
27 M. J. Yuan, Z. Shan, B. Z. Tian, B. Tu, P. Y. Yang and 46 T.Yamaguchi,Y.TanakaandK.Tanabe, J.Catal., 1980,65, 442.
D. Y. Zhao, Microporous Mesoporous Mater., 2005, 78, 37.
28 Y. D. Lu, M. J. Yuan, Y. Liu, B. Tu, C. H. Xu, B. H. Liu,
D. Y. Zhao and J. L. Kong, Langmuir, 2005, 21, 4071.
47 G. Ramis, G. Busca, C. Cristiani, L. Lietti, P. Forzatti and
F. Bregani, Langmuir, 1992, 8, 1744.
48 D. G. Barton, S. L. Soled, G. D. Meitzner, G. A. Fuentes and
E. Iglesia, J. Catal., 1999, 181, 57.
29 D. Grosso, G. J. A. A. Soler-Illia, F. Babonneau, C. Sanchez,
P. A. Albouy, A. Brunet-Bruneau and A. R. Balkenende, Adv. 49 I. Furikado, T. Miyazawa, S. Koso, A. Shimao, K. Kunimori
Mater., 2001, 13, 1085. and K. Tomishige, Green Chem., 2007, 9, 582.
30 Q. Yuan, Y. Liu, L. L. Li, Z. X. Li, C. J. Fang, W. T. Duan, 50 A. Punnoose and M. S. Seehra, Catal. Lett., 2002, 78, 157.
X. G. Li and C. H. Yan, Microporous Mesoporous Mater., 51 M. Occhiuzzi, D. Cordischi, D. Gazzoli, M. Valigi and
2009, 124, 169.
31 G. Calleja, D. P. Serrano, R. Sanz and P. Pizarro, Microporous 52 D. Bravo, X. Ruiz, F. Dıaz and F. J. Lopez, Phys. Rev. B:
Mesoporous Mater., 2008, 111, 429. Condens. Matter Mater. Phys., 1995, 52, 3159.
32 K. L. Frindell, J. Tang, J. H. Harreld and G. D. Stucky, Chem. 53 S. Kuba, P. C. Heydorn, R. K. Grasselli, B. C. Gates, M. Che
P. C. Heydorn, Appl. Catal., A, 2004, 269, 169.
´
´
¨
Mater., 2004, 16, 3524.
and H. Knozinger, Phys. Chem. Chem. Phys., 2001, 3, 146.
3732 | J. Mater. Chem. A, 2013, 1, 3724–3732
This journal is ª The Royal Society of Chemistry 2013