Promoting effect of WOx on selective hydrogenolysis of glycerol to 1,3-propanediol over bifunctional Pt-WOx/Al2O3 catalysts

Article abstract of DOI:10.1016/j.molcata.2014.12.021

Despite 1,3-propanediol possessing high economic value, its production from glycerol hydrogenolysis is a challenging task. Herein, a series of WO_x promoted Pt/Al₂O₃ catalysts with various WO_x contents were prepared and investigated for selective production 1,3-propanediol from glycerol hydrogenolysis. To explore the structure feature, these catalysts were fully characterized by BET, CO chemisorption, HRTEM, XRD (in situ XRD), Raman, NH₃-TPD, Py-IR, H₂-TPR, and XPS. Among them, Pt-10WO_x/Al₂O₃ achieved the highest 1,3-propanediol yield up to 42.4%, which was ascribed to the large concentration of Br?nsted acid sites, strong electronic interaction between Pt with WO_x and hydrogen spillover. The strong correlation between 1,3-propanediol yield and Br?nsted acid site indicated its essential role for the formation of 1,3-propanediol. Meanwhile the linear correlation between 1,2-propanediol yield and Lewis acid site gave direct evidence that Lewis acid site preferentially generated 1,2-propanediol.

Full text of DOI:10.1016/j.molcata.2014.12.021

Journal of Molecular Catalysis A: Chemical 398 (2015) 391–398
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Promoting effect of WO on selective hydrogenolysis of glycerol to
x
1
,3-propanediol over bifunctional Pt–WO /Al O catalysts
x
2 3
a,∗
b
a,b,∗
, Yongwang Lia,b
Shanhui Zhu , Xiaoqing Gao , Yulei Zhu
a
State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, PR China
Synfuels China Co., Ltd, Taiyuan 030032, PR China
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 10 November 2014
Received in revised form
Despite 1,3-propanediol possessing high economic value, its production from glycerol hydrogenolysis
is a challenging task. Herein, a series of WOx promoted Pt/Al2O3 catalysts with various WOx contents
were prepared and investigated for selective production 1,3-propanediol from glycerol hydrogenolysis.
To explore the structure feature, these catalysts were fully characterized by BET, CO chemisorption,
HRTEM, XRD (in situ XRD), Raman, NH3–TPD, Py–IR, H2–TPR, and XPS. Among them, Pt–10WOx/Al2O3
achieved the highest 1,3-propanediol yield up to 42.4%, which was ascribed to the large concentration
of Brønsted acid sites, strong electronic interaction between Pt with WOx and hydrogen spillover. The
strong correlation between 1,3-propanediol yield and Brønsted acid site indicated its essential role for
the formation of 1,3-propanediol. Meanwhile the linear correlation between 1,2-propanediol yield and
Lewis acid site gave direct evidence that Lewis acid site preferentially generated 1,2-propanediol.
© 2014 Elsevier B.V. All rights reserved.
2
4 December 2014
Accepted 29 December 2014
Available online 31 December 2014
Keywords:
Glycerol
Hydrogenolysis
1
,3-Propanediol
WOx
Pt
1
. Introduction
The use of renewable resources as feedstocks for the produc-
work has achieved 98.0% yield of 1,2-PDO over B O3 modified
Cu/SiO2 catalyst [15]. On the contrary, the direct hydrogenolysis
of glycerol to 1,3-PDO is still a challenging task. One of the key
problems is concerned with the 1,3-PDO selectivity, which requires
careful design of catalyst.
2
tion of fuels, chemicals and materials has been attracting much
attention, as the fossil resources will be exhausted in a few decades
[
1,2]. In this context, biomass is a desirable candidate as an alterna-
There have been several reports on homogeneous and hetero-
geneous catalysts for the catalytic hydrogenolysis of glycerol to
1,3-PDO [16–21]. Nevertheless, most of previous studies have been
performed in organic solvent [16,19,22]. Considering that crude
glycerol from biodiesel production contains water unavoidably and
it is also a by-product of the reaction sequence, water is the desired
solvent for glycerol hydrogenolysis from the standpoint of environ-
mental and economic viability. Among them, the relatively effective
processes using water primarily employed Ir–ReOx/SiO2 [20,23] or
tive and carbon neutral resource. As a biomass-derivative, glycerol
is currently produced in a large amount as a by-product in man-
ufacturing biodiesel by transesterification of vegetable oils with
methanol or ethanol [3]. Therefore, it is urgent to effectively uti-
lize the renewable glycerol for the sustainable development of
biodiesel industry. Significant research efforts have been focused
on the transformation of glycerol by various catalytic processes,
such as reforming [4], oxidation [5], dehydration [6], hydrogenoly-
sis [7,8], esterification [9,10] and polymerization [11].
Pt-based catalysts modified by tungsten species (WO [16,24], WOx
3
The catalytic hydrogenolysis of glycerol to 1,2-propanediol (1,2-
PDO) and 1,3-propanediol (1,3-PDO), which are widely used as
versatile specialty chemicals, is quite promising. 1,3-PDO owes
much higher economic value than 1,2-PDO, in particular, as an
important monomer in the synthesis of polyester fibers [7]. Conver-
sion of glycerol to 1,2-PDO has been extensively studied, and high
yields have been obtained in previous reports [12–15]. Our recent
[21,25], H WO4 [22] and H SiW12O40 [26,27]). Nakagawa et al.
2
4
developed Ir–ReOx/SiO2 catalyst with H SO4 as an additive in a
2
batch reactor, over which the yield of 1,3-PDO reached 38% [20].
To overcome the corrosive liquid H SO , H-ZSM-5 was consid-
2
4
ered as the most suitable solid acid co-catalyst and the yield of
1,3-PDO declined to 33% moderately [23]. Zhang et al. [21] per-
formed glycerol hydrogenolysis over mesoporous WO –TiO with
3
2
◦
a loading of 2 wt% Pt at 180 C and 5.5 MPa, which gave 40.3%
1
3
,3-PDO selectivity and 24.2% conversion. Qin et al. [24] achieved
2.0% yield of 1,3-PDO over Pt/WO /ZrO catalyst in glycerol
3
2
∗ Corresponding author. Tel.: +86 351 7117097.
E-mail addresses: zhushanhui@sxicc.ac.cn (S. Zhu), zhuyulei@sxicc.ac.cn
◦
hydrogenolysis at 130 C and 4 MPa. The other relatively feasi-
ble catalysts included Pt/WO /TiO /SiO [28], Pt-H SiW12O40/ZrO2
3
2
2
4
(
Y. Zhu).
http://dx.doi.org/10.1016/j.molcata.2014.12.021
381-1169/© 2014 Elsevier B.V. All rights reserved.
1

3
92
S. Zhu et al. / Journal of Molecular Catalysis A: Chemical 398 (2015) 391–398
[
27,29], Pt/Al O + H SiW12O40 [26], Pt/WOx/Boehmite [25], and
ichiometry of adsorbed CO molecule to surface platinum atom was
one.
2
3
4
Pt/WOx/SiO /ZrO [30]. These results show that the acidic tungsten
2
2
species appear as a key to the selective formation of 1,3-PDO, but
the role of tungsten species is still unclear. Despite growing exper-
imental work, a systematic understanding remains unavailable on
the type of acid sites in governing the rate and cleavage C O bond
selectively over tungsten-containing catalysts.
Temperature-programmed reduction of hydrogen (H –TPR)
2
was conducted in the same apparatus as CO chemisorption. For
◦
each run, about 0.10 g sample was firstly pretreated in Ar at 150 C
◦
for 30 min and then cooled to 30 C. After that, 10% H diluted in Ar
2
was introduced into the system and a cold trap of 2-propanol-liquid
nitrogen slurry was provided to condense the water gas. The sample
In the present work, we reported a detailed study for continuous
hydrogenolysis of glycerol over Pt-WOx/Al O catalysts in aqueous
◦
◦
was heated up to 900 C at a ramp of 10 C/min and simultaneously
recorded by a thermal conductivity detector.
2
3
media. A 42.4% yield of 1,3-PDO was achieved at 64.2% conversion,
◦
1
60 C and 5.0 MPa over Pt–10WOx/Al O catalyst. Special empha-
Temperature-programmed desorption of ammonia (NH –TPD)
2
3
3
sis would be focused on the role of WOx and structure-behavior
correlation.
was measured in the same apparatus as CO chemisorption. At first,
◦
0.3 g catalyst sample was pretreated in He at 400 C for 1 h and
◦
then cooled to 100 C followed by saturating with pure NH for
3
◦
3
1
0 min. Finally, the sample was heated to 600 C at a ramping rate of
0 C/min and the desorbed NH was monitored with a MS detector
2
. Experiment
◦
3
2.1. Catalyst preparation
(Agilent, USA).
IR spectra of adsorbed pyridine (Py–IR) were performed in
Vertex 70 (Bruker) FT-IR spectrophotometer with a deuterium
triglycine sulfate (DTGS) detector. For each run, the sample was
pressed into self-supporting wafers and degassed in vacuum at
The Pt–WOx/Al O catalysts were prepared by sequential
2
3
impregnation method. Specifically, ␥-Al O3 (China Research
2
Institute of Daily Chemical Industry) was impregnated with
aqueous solutions containing the desired amount of ammo-
◦
300 C for 1 h followed by exposure to pyridine vapor. Subse-
◦
nium paratungstate (Sinopharm Chemical Reagent Co., Ltd., China,
quently, the Py–IR spectra were measured at 200 C after applying
◦
(
SCRC)). These WOx/Al O samples were dried overnight at 110 C
vacuum for 30 min. The quantification of Brønsted and Lewis acid
sites was estimated from the integrated area of adsorption bands
2
3
◦
and calcined at 600 C in static air for 4 h. The Pt–WOx/Al O cata-
2
3
−
1
lysts were fabricated by impregnation of WOx/Al O3 samples with
at ca. 1540 and 1450 cm , respectively, described elaborately in
our previous reports [10,27].
2
an aqueous solution of H PtCl ·6H O (SCRC). Impregnated sam-
2
6
2
◦
ples were dried overnight at 110 C and then calcined in static air
at 400 C for 4 h. According to the tungstate oxide mass (y in wt%,
X-ray photoelectron spectroscopy (XPS) was carried out on a
VG MiltiLab 2000 spectrometer with Mg K␣ radiation and a mul-
tichannel detector. The catalyst was reduced in flowing hydrogen
◦
quantified in the form of WO ), the final catalysts were designated
3
◦
as Pt–yWOx/Al O (y = 5 ∼ 20). The Pt/Al O without WOx was also
at 200 C for 2 h before the measurement. The obtained binding
2
3
2
3
prepared as reference sample by impregnation of ␥-Al O with an
energy values were calibrated by the C 1s peak at 284.6 eV.
2
3
aqueous solution of H PtCl ·6H O. The Pt/Al O sample was dried
2
6
2
2
3
◦
and then calcined at 400 C for 4 h. The loading of Pt were fixed at
2.3. Catalytic reaction
2
wt% in all catalysts.
The catalytic test was performed in a vertical fixed-bed stain-
less steel reactor with an ice–water trap. Typically, 2.0 g catalyst
2.2. Catalyst characterization
(20–40 mesh) was placed into the constant temperature section of
◦
the reactor possessing quartz sand to fix it in both ends. Prior to
the test, the catalyst was reduced in flowing H2 (100 ml/min) at
200 C for 2 h. Subsequently, a glycerol aqueous solution (1.8 ml/h)
N2 adsorption–desorption isotherms were recorded at −196 C
on a Micromeritics ASAP 2420 instrument. Prior to the measure-
◦
◦
ment, each sample was degassed under vacuum at 300 C for 8 h.
was continuously pumped into the reactor inlet through an HPLC
pump. The liquid and gas products were condensed and col-
lected in a gas–liquid separator immersed in an ice–water trap.
Powder X-ray diffraction (XRD) patterns were collected on a
D2/max-RA X-ray diffractometer (Bruker, Germany) with Cu K␣
radiation operated at 30 kV and 10 mA. The X-ray patterns were
◦
◦
◦
The standard reaction conditions were as follows: 160 C, 5.0 MPa,
recorded in 2ꢀ values ranging from 10 to 90 . For in situ XRD mea-
1
0 wt% glycerol aqueous solution, H /glycerol = 137:1 (molar ratio),
surement, these samples were flushed in pure H at a flow rate
2
2
−
1
3
−1
◦
WHSV = 0.09 h
.
of 30 cm min and heated from 30 C to 150, 250, 300, 350, 400,
◦
◦
−1
The liquid products were analyzed by a gas chromatography
Ruihong chromatogram analysis Co., Ltd., China) with a FID using
4
50, 500, 550 and 600 C at a ramping rate of 5 C min . The XRD
patterns were recorded after the preset temperature attained for
0 min.
Transmission electron microscopy (TEM) and high-resolution
(
a DB-WAX capillary column. The outlet gas was off-line analyzed
using a gas chromatograph (Huaai chromatogram analysis Co., Ltd.,
China) equipped with a TCD and OV-101 column. Concurrently, the
assignment of the products was identified by GC–MS (Agilent, USA).
The identified products were 1,3-PDO, 1,2-PDO, 1-propanol (1-
PO), 2-propanol (2-PO), ethanol, acetol, propionic acid, acetic acid,
3
transmission electron microscopy (HRTEM) images of the samples
were obtained in JEM 2011F apparatus operating at 200 kV voltages.
The samples after reduction were suspended in ethanol with an
ultrasonic dispersion for 30 min and deposited on carbon-coated
copper grids.
Raman spectroscopy was conducted on a LabRAM HR800 Sys-
tem using a CCD detector at room temperature. The 325 nm of the
He–Cd laser was employed as the exciting source with a power of
ethylene glycol, methanol, propane, methane and CO . The conver-
2
sion of glycerol and selectivity of products were calculated using
the following expressions:
moles of glycerol(in) − moles of glycerol(out)
Conversion(%) =
3
0 MW.
CO chemisorption was performed in Auto Chem. II 2920 equip-
moles of glycerol(in)
×
100
ment (Mircromeritics, USA). Prior to adsorption measurement, 0.2 g
◦
catalyst sample was reduced in H2 for 2 h at 200 C and then
◦
flushed with He for 1 h followed by cooling down to 30 C. The
moles of one product
moles of all products
CO chemisorption was measured by pulse injection of pure CO at
Selectivity(%) =
× 100
◦
3
0 C. The Pt particle size was determined by assuming that the sto-

S. Zhu et al. / Journal of Molecular Catalysis A: Chemical 398 (2015) 391–398
393
Table 1
Physicochemical properties and acidities of Pt/Al2O3 and Pt–WOx/Al2O3 catalysts.
Catalyst
Surface
area
Pore size
(nm)
Pore
Pt size (nm)
Pt
Acidity (mmol
NH3/gcat.)
Lewis acidity
(␮mol/gcat.)
Brønsted
acidity
a
b
volume
dispersion
(%)
2
−1
3
−1
)
(␮mol/gcat.)b
(m
g
)
(cm
g
Pt/Al2O3
179.1
160.3
145.0
133.8
121.5
11.6
11.9
12.2
12.8
13.6
0.593
0.550
0.481
0.434
0.401
2.1
1.8
1.5
1.8
1.9
53.5
64.7
78.1
63.2
59.0
0.17
0.33
0.41
0.40
0.38
47.5
85.7
141.2
138.0
129.1
1.7
29.4
39.1
37.5
35.4
Pt–5WOx/Al2O3
Pt–10WOx/Al2O3
Pt–15WOx/Al2O3
Pt–20WOx/Al2O3
a
The amount of acid sites was determined by quantifying the desorbed NH3 from NH3–TPD.
The amount of acid sites was determined by quantifying the desorbed pyridine from Py–IR.
b
3
3
3
. Results
effects of WOx promoter, namely bimetallic coverage effect and dis-
persion effect. Previous research on Ru–MoOx/ZrO2 has confirmed
that the presence of minor MoOx species promotes the dispersion
of Ru sites, while redundant MoOx can weaken the dispersion effect
and cover Ru sites [34].
.1. Catalyst characterizations
.1.1. Physicochemical properties of catalysts
Table 1 lists the textural properties for Pt–WOx/Al O catalysts
Fig. 1a and c displays representative HRTEM images of reduced
Pt/Al O3 and Pt–10WOx/Al O . The dispersed Pt nanoparticles
2
3
with different WOx loading. Only a minor decrease in surface area
was observed with an increase in WOx content. The Pt particle size
and dispersion can be determined by measuring the amount of CO
adsorbed because CO is adsorbed more preferentially on the sur-
face of Pt atoms than of WOx species at room temperature [31]. As
shown in Table 1, the irreversible CO uptake was an indication of Pt
dispersion of 53.5% for Pt/Al O catalyst. Despite good Pt dispersion
2
2
3
were surrounded by acicular ␥-Al O for these two samples, which
2
3
can be clearly verified by the lattice fringes of Pt (d = 0.223 nm)
[35]. Addition of WOx to Pt/Al O3 moderately enhanced Pt disper-
2
sion and declined Pt particle size correspondingly (Fig. 1b and d).
As shown in the size distribution histogram of Pt–10WOx/Al O ,
2
3
most of Pt nanoparticles are in the range of 0.8–2.0 nm, and no
particles larger 3 nm were detected. Notably, no concrete evidence
pointed to the formation of WOx clusters over Pt–10WOx/Al O ,
2
3
in Pt/Al O , it still enhanced with the increasing WOx loading up to
2
3
1
0 wt%, implying that WOx species had positive impact on the dis-
2
3
persion of Pt. However, further increase in WOx content declined
the dispersion mildly. Such a decrease in Pt dispersion was virtually
not ascribed to the change in Pt particles sizes, which may originate
from partial blockage of Pt sites by excessive WOx species. Similar
coverage effects have been widely reported in the cases of ReOx
speculating that the TEM-invisible WOx species primarily existed
as polytungstate (see below).
The XRD patterns of Pt–WOx/Al O3 catalysts with various WOx
2
◦
◦
loading are displayed in Fig. 2. The diffraction lines at 37.4 , 46.0
and 67.0 were attributed to support ␥-Al O3 [14]. For catalysts
◦
2
[
32], SnOx [33], and MoOx [34]. This can be explained by the double
with WOx loading lower than 10 wt%, the absence of any diffrac-
Fig. 1. HRTEM images of reduced (a) Pt/Al2O3 and (c) Pt–10WOx/Al2O3 with Pt particle size distribution of (b) Pt/Al2O3 and (d) Pt–10WOx/Al2O3.

3
94
S. Zhu et al. / Journal of Molecular Catalysis A: Chemical 398 (2015) 391–398
Pt-20WO /Al O
3
x
2
Pt-15WO /Al O
3
x
2
Pt-10WO /Al O
3
x
2
Pt-5WO /Al O
x
2
3
Pt/Al O3
2
200
400
Temperature ( C)
600
o
Fig. 2. XRD patterns of Pt/Al2O3 and Pt–WOx/Al2O3 catalysts.
Fig. 4. NH3–TPD profiles of Pt/Al2O3 and Pt–WOx/Al2O3 catalysts.
tion peaks related to WOx species was detected, implying the
homogenous dispersion of WOx species on the Al O surface. How-
ever, further increase in WOx loading improved polymerization of
2
3
bands at 271, 712 and 807 cm 1 to crystalline m-WO3 phase [31].
Apparently, no crystalline m-WO3 bands were detected for these
catalysts below the dispersion threshold (10 wt%). These character-
−
tungstic species and resulted in the formation of monoclinic m-
◦
◦
WO3 (23.6 and 33.5 ) [36]. The intensity of XRD diffraction peaks
for m-WO3 increased obviously with the increasing WOx loading.
istic bands emerged at Pt–10WOx/Al O and then became intensive
with an increase in WOx content, in support of the XRD results.
2
3
It can be reasonably inferred that Pt–10WOx/Al O3 should be the
2
sample with dispersion threshold of WOx, namely reaching sub-
monolayer coverage [36]. On the other hand, no diffraction peaks
of Pt were observed, reflecting that Pt clusters were uniformly
dispersed on the support surface, in agreement with CO chemisorp-
tion and TEM results. Moreover, in situ XRD (Fig. S1 and Fig. S2)
results verified that the Pt/Al O and Pt–10WOx/Al O3 catalysts
3.1.2. Acidic properties of catalysts
The NH -TPD technique was employed to probe the avail-
3
able surface acidic sites for Pt–WO /Al O catalysts, together with
x
2
3
acidic strength. As displayed in Fig. 4, all the catalysts presented
broad TPD profiles, reflecting that the surface acid strength was
widely distributed. The NH –TPD profile of Pt/Al O catalyst exhib-
2
3
2
were pretty stable upon high temperature treatment.
3
2
3
Raman spectroscopy is well-known to be suitable for examin-
ing the surface structure of WOx species. As illustrated in Fig. 3,
the Raman spectra of Pt/Al O displayed two typical bands at 369
ited poorly resolved desorption peaks in two distinct temperature
regions, centering at ca. 187 and 324 C, which was ascribed to weak
and moderate acidic sites, respectively [29]. These peaks shifted
toward higher temperature and improved with increasing WOx
◦
2
3
−
1
and 562 cm ascribed to the ␥-Al O [37]. All these Pt–WOx/Al O
2
3
2
3
−
1
samples gave a broad band at about 972–995 cm
correspond-
content, reflecting that addition of WO facilitated to increase acidic
x
ing to the stretch mode of surface monooxo W O species [37].
strength. The total acidity can be calculated on the basis of NH3
−
1
−1
Furthermore, this band shifted from 972 cm to 995 cm with
increasing WOx loading, suggesting the increase of WOx domain
size, namely, transformation from isolated monotungstate to poly-
tungstate species and further aggregation to m-WO3 clusters
injection experiment and the amount of acidic sites is summa-
rized in Table 1. The total acidity enhanced continuously with the
increasing WO loading, maximized on Pt–10WO /Al O and then
x
x
2
3
did not change the acid site distribution any more.
[
36,38]. It has been identified and attributed Raman characteristic
According to earlier study on WO /ZrO2 [36,38], the acid sites
x
improved steadily with increasing WOx loading and attained maxi-
mum at the monolayer coverage (dispersion threshold). The surface
monotungstate species at low WOx content showed small acid-
ity while the polytungstate became dominant at the monolayer
WOx coverage and possessed strong acid sites due to the genera-
tion of interconnected WOx species. Nevertheless, further increase
in WOx loading led to the formation of crystalline m-WO3 phase,
which made no contribution to the total acidity. Consequently, the
1
0 wt% WOx loading reached dispersion threshold and obtained the
maximum acidity, consistent well with the results from XRD and
Raman.
FTIR spectra of pyridine adsorption were exhibited to determine
the nature of acid sites. As illustrated in Fig. 5, the bands centered
−
1
at ca. 1445 and 1610 cm
nated to Lewis sites [27]. The bands at ca. 1540 and 1640 cm
were attributed to pyridine adsorbed at Brønsted acid sites. The
were typical pyridine bands coordi-
−
1
−
1
adsorption around 1489 cm was assigned to the combined con-
tribution of pyridine adsorbed on Lewis and Brønsted acid sites
[
10]. The concentration of Lewis and Brønsted acid sites was cal-
culated from the integrated area of adsorption bands at ca. 1450
−
1
Fig. 3. Raman spectra of Pt/Al2O3 and Pt–WOx/Al2O3 catalysts.
and 1540 cm , respectively. As displayed in Table 1, Pt/Al2O3 pos-

S. Zhu et al. / Journal of Molecular Catalysis A: Chemical 398 (2015) 391–398
395
Pt 4d3/2
Pt 4d5/2
(A)
Pt-20WO /Al O
x
2
3
Pt-15WO /Al O
x
2
3
Pt-10WO /Al O
3
x
2
Pt-5WO /Al O
x
2
3
Pt/Al O3
2
310
315
320
325
330
335
340
345
Binding energy (eV)
5
+
Fig. 5. FTIR spectra of pyridine adsorption of Pt/Al2O3 and Pt–WOx/Al2O3 catalysts.
W
6+
W
(B)
Pt-20WO /Al O
3
x
2
sessed some Lewis acid sites and hardly contained Brønsted acid
sites. Addition of WOx led to a noticeable increase in both Lewis and
Brønsted acid sites. Triwahyono et al. [39] suggested that Brønsted
acid sites were stemmed from surface OH groups bonded on the
exposed or well-dispersed WOx of WO /ZrO while Lewis acid sites
Pt-15WO /Al O
3
3
2
x
2
were generated by the removal of surface OH groups. As a result,
Pt–10WOx/Al O existed as surface monolayer WOx coverage and
2
3
exposed the most surface WOx species, resulting in the highest acid
sites in terms of Lewis and Brønsted acid sites, in agreement with
Pt-10WO /Al O
3
x
2
NH –TPD results.
3
Pt-5WO /Al O
x
2
3
3
.1.3. Reducibility property and chemical state
Fig. displays the H –TPR profiles of Pt/Al O3 and
Pt–WOx/Al O catalysts. Pt/Al O3 showed a broad PtOx reduction
6
2
2
3
2
34
36
38
40
2
3
◦
2
Binding energy (eV)
at around 155 C, in line with the previous report [40]. The incorpo-
ration of WOx remarkably decreased the reduction temperature of
PtOx because of the strong electron interaction between PtOx and
Fig. 7. (A) Pt 4d and W4f (B) XPS photoemission peaks of reduced Pt/Al2O3 and
Pt–WOx/Al2O3 catalysts.
WOx species (see below). The WOx-containing catalysts presented
◦
other two peaks at ca. 320 and 620 C, which were corresponding
to the reduction of WO → WO and WO2.9 → WO , respectively
3
2.9
2
[
41]. There was a general enhancement in the total hydrogen
consumption with the increasing WOx loading, indicating the
increasing amount of potential reducible species as WOx loading
improved.
XPS analysis is used to identify surface chemical states after
◦
reduction at 200 C and the results are depicted in Fig. 7. The Pt 4d
lines (Fig. 7A) were analyzed instead of the most intense Pt 4f lines
due to their serious overlapping with the strong Al 2p peaks [33].
The two peaks at ca. 314.5 and 331.3 eV were ascribed to Pt 4d5/2
and Pt 4d3/2 of metallic Pt, respectively, suggesting that the sur-
face Pt species had been reduced to zero-valent Pt completely [22].
Fig. 7B illustrates XPS spectra of W 4f consisting of two spin–orbit
components. In the W 4f region, the doublet at binding energies of
ca. 34.9 and 37.0 eV was corresponding to W4f
and W 4f5/2 of
7/2
5+
W
, and another one at 35.9 and 38.1 eV was assigned to W4f7
/2
and W 4f5/2 of W6+ [42]. As shown in Table S1, the Pt 4d peak posi-
tion moved to higher binding energy with increasing WOx content,
while the W 4f peak shifted to lower binding energy simultane-
0
ously. The declining electron density in Pt nucleus originated from
the electron scavenging property of WOx [41]. Accordingly, an elec-
0
tronic donor-acceptor interaction between Pt and WOx might take
place on ␥-Al O3 surface. The above results revealed the existence
2
0
Fig. 6. H2–TPR profiles of Pt/Al2O3 and Pt–WOx/Al2O3 catalysts.
of strong interaction between Pt and WOx species wherein the

3
96
S. Zhu et al. / Journal of Molecular Catalysis A: Chemical 398 (2015) 391–398
0
electron was prone to transfer from Pt to WOx species. In addition,
some partially reduced W5+ species were formed for all the sam-
ples. Kuba et al. [43] observed similar phenomenon in Pt–WOx/ZrO2
and argued that the hydrogen atoms from dissociative adsorption
hydrogen molecule on Pt surface could spillover to WO3 surface
and result in its reduction.
WOx and results in promoting the protonation of hydroxyl group
of glycerol via withdrawing electrons from the oxygen atom in its
hydroxyl group [46]. Such enhanced protonation favors glycerol
dehydration to form intermediate. In sum, compared to Pt/Al2O3,
this electron effect between Pt and WOx species is instrumental
in elevating the catalytic behavior of glycerol hydrogenolysis for
Pt–WOx/Al O .
2
3
As PtOx is reduced to Pt by H , it is speculated that H2
2
3
.1.4. Catalytic performance of Pt–WOx/Al O
2 3
spillover can occur in glycerol hydrogenolysis for Pt–WOx/Al O .
2
3
Table 2 lists the catalytic activity and product distributions
for Pt–WOx/Al O catalysts. The Pt/Al O3 catalyst exhibited poor
activity; the conversion was merely 9.3%. In contrast to Pt/Al O ,
glycerol conversion enhanced greatly for WOx promoted catalysts.
Regarding the selectivity, 1,3-PDO selectivity on Pt/Al O3 was only
9
The other products were main 1,2-PDO (67.6%), and minor 1-PO
with 2-PO derived from sequential hydrogenolysis of propanediols.
Conversely, with the promotion of WOx, even in a minor amount,
H2 molecule can dissociative chemisorb on Pt sites and the formed
2
3
2
hydrogen atoms tend to spillover to the nearby WO surface. These
3
2
3
hydrogen atoms are very active and induce the partial reduc-
tion of W6+ to W [43]. The formation of abundant W species
was demonstrated by XPS results. Furthermore, the “hydrogen
spillover” has been well established on the basis of the reduc-
5+
5+
2
.4%, similar to the result (12.1%) reported by Delgado et al. [40].
tion of Pt/WO3 system by H , in which tungsten bronzes such as
2
HxWO3 are formed [47]. R. Prins [48] insist that hydrogen spillover
from a metal surface to the surface of non-reducible support such
as Al O , SiO , MgO, and zeolites is energetically impossible. As
1
,3-PDO selectivity increased considerably. For example, the selec-
2
3
2
tivity of 1,3-PDO was up to 54.3% over Pt–5WOx/Al O , whereas
2
3
a result, hydrogen spillover cannot occur over Pt/Al O3 catalyst.
2
that of 1,2-PDO declined to a lower level (8.4%). 1,3-PDO selectivity
enhanced with increasing WOx loading and attained a maximum
at 10 wt% WOx content with a ratio of 1,3-PDO/1,2-PDO as high
as 11.6. Nevertheless, further increase in WOx loading led to its
mild decrease. Of the catalysts tested, Pt–10WOx/Al O achieved
the excellent performance, up to 42.4% 1,3-PDO yield, which was
among the best reported results [8,21–25,28]. Additionally, 1-PO
selectivity improved drastically for WOx modified Pt/Al O cata-
Notably, the WOx modified Pt/Al O3 catalysts endow “hydrogen
2
spillover” ability and lead to the formation of large amount of active
hydrogen and W5+ species simultaneously, which should greatly
enhance the reaction rate of hydrogenation of intermediates in
glycerol hydrogenolysis.
2
3
4
.2. Structure-behavior relationship
2
3
lysts while 2-PO selectivity varied slightly. This can be explained
by the fact that 1,2-PDO primarily tended to proceed sequential
hydrogenolysis to form 1-PO [24,29]. Taken together, addition of
WOx significantly enhanced glycerol conversion and switched the
main product from 1,2-PDO to 1,3-PDO by changing the dissocia-
tion of C O bond.
Production of propanediols (1,2-PDO and 1,3-PDO) from glyc-
erol hydrogenolysis involves acid-catalyzed dehydration to form
intermediates (acetol and 3-hydroxypropaldehyde (3-HPA)) and
subsequent hydrogenation on metal sites [29,49–51]. Therefore,
higher acid sites, metals sites and active hydrogen species are
favorable to convert glycerol. The WOx modified Pt/Al O3 cata-
2
lysts enhanced the total acidity, electronic interaction between Pt
with WOx and hydrogen spillover, which was responsible for their
superior catalytic performance.
4
. Discussion
4
.1. The structure feature of Pt–WOx/Al O catalysts
The nature of acid site indeed plays a key role in determining
the product distribution. Brønsted acid site is prone to removal of
secondary hydroxyl group of glycerol to generate 1,3-PDO while
Lewis acid sites favors to form 1,2-PDO [19,52]. As shown in Fig. 8,
it is evident that 1,3-PDO yield is approximately proportional to
the concentration of Brønsted acid, reflecting the preferential gen-
eration of 1,3-PDO on Brønsted acid sites. Apparently, addition of
WOx species significantly improves the amount of Brønsted acid,
2
3
It is well-known that WOx has been extensively employed as
an acidic additive, such as glycerol dehydration [44], hydrolysis of
cellobiose [38] and alkane isomerization [43]. As deduced by XRD,
Raman and NH –TPD results, the molecular structures of WOx on
Pt/Al O catalysts consisted of monotungstate, polytungstate and
crystalline WO3 clusters, which depended on WOx loading. The
WOx on Pt–5WOx/Al O primarily existed in the form of mono-
3
2
3
rendering the superior 1,3-PDO selectivity. The Pt–10WOx/Al O3
2
3
2
tungstate due to its low WOx loading in highly dispersed state
on support surface. When WOx content reached its dispersion
threshold over Pt–10WOx/Al O , namely sub-monolayer coverage,
catalyst in the series has the largest amount of Brønsted acid sites
and exhibits the highest 1,3-PDO yield, nearly 50 times that of
Pt/Al O3 under the same conditions.
2
3
2
polytungstate became dominant and held the maximum acidity.
Theoretical calculation also reveals that at monolayer coverage
octahedral polytungstate species has the most abundant edge sites
owing to condensation process, and the edge sites are the most
acidic units [45]. The superfluous WOx in Pt–15WOx/Al O and
Regarding the relationship between 1,2-PDO yield and Lewis
acid sites, the correlation is not good, which is mainly ascribed
to the instability of 1,2-PDO [26,29]. Our earlier report [29] has
confirmed that the formed propanols (1-PO + 2-PO) are primar-
ily derived from sequential hydrogenolysis of 1,2-PDO in glycerol
hydrogenolysis, as also elucidated on Pt–WOx/Al O . The reaction
2
3
Pt–20WOx/Al O existed as crystalline WO3 phase that made no
2
3
2
3
contribution to the number of acid sites.
results for hydrogenolysis of glycerol, 1,2-PDO and 1,3-PDO over
XPS results indicated that the interaction between Pt and WOx
diminished the electron density of Pt, which was stemmed from
the scavenging property of WOx. The interaction also enhanced
the reducibility of PtOx, and increased dispersion of Pt for WOx
promoted Pt/Al O catalysts. TEM and CO-chemisorption results
Pt–10WOx/Al O3 is displayed in Table S2. The conversion of 1,3-
2
PDO was much lower than that of 1,2-PDO and glycerol, which
can explain the high 1,3-PDO selectivity in glycerol hydrogenolysis
over Pt–10WOx/Al O . Additionally, the 1-PO/2-PO ratio in glyc-
2
3
erol hydrogenolysis was similar to that in 1,2-PDO hydrogenolysis.
Thus, it is reasonably to infer that 1-PO and 2-PO are predominantly
derived from 1,2-PDO in glycerol hydrogenolysis. Moreover, simi-
2
3
confirmed that Pt dispersion of Pt–WOx/Al O3 was high, with only
2
some detectable nanoclusters. Moreover, the deficient electron
density of Pt sites facilitates to adsorb H and promotes their disso-
lar behavior has been disclosed in the cases of Pt/WO /ZrO2 [24],
2
3
ciation, essential for the hydrogenation of intermediates in glycerol
hydrogenolysis. The electron transfer enhances electron density of
Pt/Al O + H SiW12O40 [26] and ReOx modified Ir or Rh based cata-
lysts [53–56]. Accordingly, the calculated 1,2-PDO yield should be
2
3
4

S. Zhu et al. / Journal of Molecular Catalysis A: Chemical 398 (2015) 391–398
397
Table 2
Catalytic performance of glycerol hydrogenolysis over Pt/Al2O3 and Pt–WOx/Al2O3 catalysts.
a
Catalyst
Conversion (%)
Selectivity (%)b
1,3-PDO/1,2-PDO 1,3-PDO yield (%)
molar ratio)
(
1,3-PDO
1,2-PDO
1-PO
2-PO
Pt/Al2O3
9.3
9.4
67.6
8.4
5.7
6.5
6.0
3.7
2.9
6.5
3.7
4.3
4.0
0.14
6.5
11.6
9.4
0.87
21.9
42.4
34.3
31.5
Pt–5WOx/Al2O3
Pt–10WOx/Al2O3
Pt–15WOx/Al2O3
Pt–20WOx/Al2O3
40.4
64.2
56.0
53.4
54.3
66.1
61.2
58.9
24.6
20.3
22.7
23.1
9.8
a
◦
−1
Reaction conditions: 160 C, 5.0 MPa, H2/glycerol = 137:1 (molar ratio), WHSV = 0.09 h .
1
b
,3-PDO: 1,3-propanediol, 1,2-PDO: 1,2-propanediol, 1-PO: 1-propanol, 2-PO: 2-propanol.
◦
Fig. 8. Correlation between the amount of acids sites and product yield over Pt/Al2O3 and Pt–WOx/Al2O3 catalysts. Reaction conditions: 160 C, 5.0 MPa, H2/glycerol = 137:1
−
1
(
molar ratio), WHSV = 0.09 h
.
the sum of detected 1,2-PDO and the part that has been converted
to propanols. Interestingly, a roughly linear correlation between
the calculated 1,2-PDO yield and the concentration of Lewis acid
sites is also given in Fig. 8, revealing the essential active sites of
Lewis acid sites for glycerol hydrogenolysis to 1,2-PDO.
Previous reports [57–61] showed that the main product was
not 1,3-PDO but 1,2-PDO when using Ru/C and Brønsted acidic
Amberlyst catalyst. Because Ru/C and Amberlyst was only mixed
mechanically, the formed intermediate 3-HPA in Brønsted acid sites
cannot immediately hydrogenate to 1,3-PDO. Meanwhile, large
amount of 1,2-PDO was produced over Ru/C due to the weak Lewis
acidic sites. Compared to Ru/C alone, the main product was still
to 42.4%, nearly 50 times that of Pt/Al O , which was attributed
2 3
to the abundant Brønsted acid sites, strong electronic interaction
between Pt with WOx, and hydrogen spillover.
The amount of Brønsted acid sites is proportional to 1,3-PDO
yield, giving direct evidence that Brønsted acid sites facilitate to
cleave secondary C O bond of glycerol to produce 1,3-PDO. Mean-
while, the linear correlation between Lewis acid sites with 1,2-PDO
yield confirms the importance of Lewis acid sites in controlling
cleavage primary C O bond to form 1,2-PDO. Accordingly, these
correlations will provide guidance for designing effective catalysts
to produce 1,2-PDO and 1,3-PDO, selectively.
1
,2-PDO over Ru/C + Amberlyst. Contrarily, the formed 3-HPA in
Acknowledgements
Brønsted acid sites can be quickly converted into 1,3-PDO in inter-
face between WOx and Pt species over Pt–WOx/Al O .
2
3
The authors gratefully acknowledge financial support from the
Major State Basic Research Development Program of China (973
Program) (No. 2012CB215305) and the National Natural Science
Foundation of China (No. 21403269). We also thank Dr. Shunli Hao
for valuable suggestions in catalyst preparation.
5
. Conclusions
In this work, addition of WOx to Pt/Al O3 has demon-
2
strated to be a powerful approach to tune acidic properties with
electronic structure, and to control the catalytic performance
in glycerol hydrogenolysis. The WOx species has been identi-
fied as monotungstate, polytungstate, and crystalline m-WO3
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.molcata.
2014.12.021.
phase over Pt–WOx/Al O3 catalysts. Of the catalysts examined,
2
Pt–10WOx/Al O attained the superior 1,3-propanediol yield up
2
3

3
98
S. Zhu et al. / Journal of Molecular Catalysis A: Chemical 398 (2015) 391–398
[32] Y. Amada, Y. Shinmi, S. Koso, T. Kubota, Y. Nakagawa, K. Tomishige, Appl.
References
Catal. B 105 (2011) 117–127.
[
[
[
33] T. Deng, H. Liu, Green Chem. 15 (2013) 116–124.
34] L. Chen, Y. Zhu, H. Zheng, C. Zhang, Y. Li, Appl. Catal. 411–412 (2012) 95–104.
35] C. Wen, Y. Zhu, Y. Ye, S. Zhang, F. Cheng, Y. Liu, P. Wang, F. Tao, ACS Nano 6
[
1] J.N. Chheda, G.W. Huber, J.A. Dumesic, Angew. Chem. Int. Ed. 46 (2007)
7
164–7183.
[
[
2] A. Corma, S. Iborra, A. Velty, Chem. Rev. 107 (2007) 2411–2502.
3] C.H.C. Zhou, J.N. Beltramini, Y.X. Fan, G.Q.M. Lu, Chem. Soc. Rev. 37 (2008)
(
2012) 9305–9313.
[
[
[
36] K. Song, H. Zhang, Y. Zhang, Y. Tang, K. Tang, J. Catal. 299 (2013) 119–128.
37] X. Chen, G. Clet, K. Thomas, M. Houalla, J. Catal. 273 (2010) 236–244.
38] R. Kourieh, S. Bennici, M. Marzo, A. Gervasini, A. Auroux, Catal. Commun. 19
5
27–549.
4] L. Zhang, A.M. Karim, M.H. Engelhard, Z. Wei, D.L. King, Y. Wang, J. Catal. 287
2012) 37–43.
5] P. Lakshmanan, P.P. Upare, N.-T. Le, Y.K. Hwang, D.W. Hwang, U.H. Lee, H.R.
Kim, J.-S. Chang, Appl. Catal. A 468 (2013) 260–268.
[
[
(
(
2012) 119–126.
[
[
[
[
39] S. Triwahyono, T. Yamada, H. Hattori, Appl. Catal. A 242 (2003) 101–109.
40] S.N. Delgado, D. Yap, L. Vivier, C. Especel, J. Mol. Catal. A 367 (2013) 89–98.
41] J.M. Grau, C.R. Vera, V.M. Benitez, J.C. Yori, Energy Fuels 22 (2008) 1680–1686.
42] T.Y. Kim, D.S. Park, Y. Choi, J. Baek, J.R. Park, J. Yi, J. Mater. Chem. 22 (2012)
[
[
[
[
6] W. Suprun, M. Lutecki, T. Haber, H. Papp, J. Mol. Catal. A 309 (2009) 71–78.
7] S. Zhu, Y. Zhu, S. Hao, L. Chen, B. Zhang, Y. Li, Catal. Lett. 142 (2012) 267–274.
8] R.V. Sharma, P. Kumar, A.K. Dalai, Appl. Catal. A 477 (2014) 147–156.
9] S. Zhu, Y. Zhu, X. Gao, T. Mo, Y. Zhu, Y. Li, Bioresource Technol. 130 (2013)
1
0021–10028.
43] S. Kuba, M. Che, R.K. Grasselli, H. Knözinger, J. Phy. Chem. B 107 (2003)
459–3463.
[
4
5–51.
3
[
[
10] S. Zhu, X. Gao, F. Dong, Y. Zhu, H. Zheng, Y. Li, J. Catal. 306 (2013) 155–163.
11] A. Behr, J. Eilting, K. Irawadi, J. Leschinski, F. Lindner, Green Chem. 10 (2008)
[
[
44] A. Ulgen, W.F. Hoelderich, Appl. Catal. A 400 (2011) 34–38.
45] A. Galano, G. Rodriguez-Gattorno, E. Torres-Garcia, Phys. Chem. Chem. Phys.
1
3–30.
1
0 (2008) 4181–4188.
46] M.L. Barbelli, G.F. Santori, N.N. Nichio, Bioresource Technol. 111 (2012)
00–503.
[
[
[
12] D. Sun, Y. Yamada, S. Sato, Appl. Catal. A 475 (2014) 63–68.
13] S.N. Delgado, D. Yap, L. Vivier, C. Especel, J. Mol. Catal. A 367 (2013) 89–98.
14] J.B. Salazar, D.D. Falcone, H.N. Pham, A.K. Datye, F.B. Passos, R.J. Davis, Appl.
Catal. A 482 (2014) 137–144.
15] S. Zhu, X. Gao, Y. Zhu, Y. Zhu, H. Zheng, Y. Li, J. Catal. 303 (2013) 70–79.
16] T. Kurosaka, H. Maruyama, I. Naribayashi, Y. Sasaki, Catal. Commun. 9 (2008)
[
5
[
[
[
47] Y. Xi, Q. Zhang, H. Cheng, J. Phy. Chem. C 118 (2013) 494–501.
48] R. Prins, Chem. Rev. 112 (2012) 2714–2738.
49] I. Gandarias, P.L. Arias, J. Requies, M.B. Guemez, J.L.G. Fierro, Appl. Catal. B 97
[
[
(
2010) 248–256.
1
360–1363.
[
[
[
[
[
[
50] M. Balaraju, V. Rekha, P.S.S. Prasad, B.L.A.P. Devi, R.B.N. Prasad, N. Lingaiah,
Appl. Catal. A 354 (2009) 82–87.
51] M.G. Musolino, L.A. Scarpino, F. Mauriello, R. Pietropaolo, ChemSusChem 4
[
[
17] L. Ma, D.H. He, Catal. Today 149 (2010) 148–156.
18] O.M. Daniel, A. DeLaRiva, E.L. Kunkes, A.K. Datye, J.A. Dumesic, R.J. Davis,
ChemCatChem 2 (2010) 1107–1114.
19] J. Oh, S. Dash, H. Lee, Green Chem. 13 (2011) 2004–2007.
20] Y. Nakagawa, Y. Shinmi, S. Koso, K. Tomishige, J. Catal. 272 (2010) 191–194.
21] Y. Zhang, X.-C. Zhao, Y. Wang, L. Zhou, J. Zhang, J. Wang, A. Wang, T. Zhang, J.
Mater. Chem. A 1 (2013) 3724–3732.
22] J. Chaminand, L. Djakovitch, P. Gallezot, P. Marion, C. Pinel, C. Rosier, Green
Chem. 6 (2004) 359–361.
(
2011) 1143–1150.
52] A. Alhanash, E.F. Kozhevnikova, I.V. Kozhevnikov, Appl. Catal. A 378 (2010)
1–18.
53] Y. Amada, S. Koso, Y. Nakagawa, K. Tomishige, ChemSusChem 3 (2010)
28–736.
54] S. Koso, H. Watanabe, K. Okumura, Y. Nakagawa, K. Tomishige, Appl. Catal. B
11–112 (2012) 27–37.
55] Y. Shinmi, S. Koso, T. Kubota, Y. Nakagawa, K. Tomishige, Appl. Catal. B 94
2010) 318–326.
[
[
[
1
7
[
[
1
23] Y. Nakagawa, X. Ning, Y. Amada, K. Tomishige, Appl. Catal. A 433–434 (2012)
1
28–134.
(
[
[
24] L.-Z. Qin, M.-J. Song, C.-L. Chen, Green Chem. 12 (2010) 1466–1472.
25] R. Arundhathi, T. Mizugaki, T. Mitsudome, K. Jitsukawa, K. Kaneda,
ChemSusChem 6 (2013) 1345–1347.
[
[
56] Y. Nakagawa, M. Tamura, K. Tomishige, J. Mater. Chem. A 2 (2014) 6688–6702.
57] Y. Kusunoki, T. Miyazawa, K. Kunimori, K. Tomishige, Catal. Commun. 6
(
2005) 645–649.
58] T. Miyazawa, S. Koso, K. Kunimori, K. Tomishige, Appl. Catal. A 329 (2007)
0–35.
59] T. Miyazawa, S. Koso, K. Kunimori, K. Tomishige, Appl. Catal. A 318 (2007)
44–251.
[
[
[
26] J. Dam, K. Djanashvili, F. Kapteijn, U. Hanefeld, ChemCatChem 5 (2013)
[
[
[
[
4
97–505.
27] S. Zhu, X. Gao, Y. Zhu, Y. Zhu, X. Xiang, C. Hu, Y. Li, Appl. Catal. B 140–141
2013) 60–67.
28] L. Gong, Y. Lu, Y. Ding, R. Lin, J. Li, W. Dong, T. Wang, W. Chen, Appl. Catal. A
90 (2010) 119–126.
3
(
2
60] I. Furikado, T. Miyazawa, S. Koso, A. Shimao, K. Kunimori, K. Tomishige, Green
Chem. 9 (2007) 582–588.
61] T. Miyazawa, Y. Kusunoki, K. Kunimori, K. Tomishige, J. Catal. 240 (2006)
3
[
[
29] S. Zhu, Y. Qiu, Y. Zhu, S. Hao, H. Zheng, Y. Li, Catal. Today 212 (2013) 120–126.
30] S. Zhu, X. Gao, Y. Zhu, J. Cui, H. Zheng, Y. Li, Appl. Catal. B 158–159 (2014)
213–221.
3
91–399.
[31] M.N. Taylor, W. Zhou, T. Garcia, B. Solsona, A.F. Carley, C.J. Kiely, S.H. Taylor, J.
Catal. 285 (2012) 103–114.