Hydrogenolysis of Glycerol to 1,3-propanediol under Low Hydrogen Pressure over WOx-Supported Single/Pseudo-Single Atom Pt Catalyst

Article abstract of DOI:10.1002/cssc.201501506

Single/pseudo-single atom Pt catalyst was prepared on mesoporous WO_x. The large surface area and abundant oxygen vacancies of WO_x improve the Pt dispersion and stabilize the Pt isolation. This newly prepared catalyst exhibited outstanding hydrogenolysis activity under 1 MPa H₂ pressure with a very high space-time yield towards 1,3-propanediol (3.78 g g_Pt^-1 h^-1) in Pt-W catalysts. The highly isolated Pt structure is thought to contribute to the excellent H₂ dissociation capacity over Pt/WO_x. The high selectivity towards 1,3-propanediol is attributed to the heterolytic dissociation of H₂ at the interface of Pt and WO_x (providing specific Br?nsted acid sites and the concerted dehydration-hydrogenation reaction) and the bond formation between glycerol and WO_x, which favors/stabilizes the formation of a secondary carbocation intermediate as well as triggers the redox cycle of the W species (W⁶⁺?W⁵⁺).

Full text of DOI:10.1002/cssc.201501506

DOI: 10.1002/cssc.201501506
Communications
Hydrogenolysis of Glycerol to 1,3-propanediol under Low
Hydrogen Pressure over WOx-Supported Single/Pseudo-
Single Atom Pt Catalyst
Jia Wang+,[a, b] Xiaochen Zhao+,[a] Nian Lei,[a, b] Lin Li,[a] Leilei Zhang,[a] Shutao Xu,[c]
Shu Miao,[a] Xiaoli Pan,[a] Aiqin Wang,*[a] and Tao Zhang*[a]
Single/pseudo-single atom Pt catalyst was prepared on meso-
porous WOx. The large surface area and abundant oxygen va-
cancies of WOx improve the Pt dispersion and stabilize the Pt
isolation. This newly prepared catalyst exhibited outstanding
hydrogenolysis activity under 1 MPa H2 pressure with a very
1,3-PD.[1a] In aqueous-phase glycerol hydrogenolysis, Pt-W[4]
and Ir-Re[5] catalysts appear to be so far the only effective cata-
lysts with high selectivity to 1,3-PD instead of 1,2-PD.
Generally, H2 homolytically dissociates to two hydrogen
atoms on noble metal sites, which subsequently hydrogenate
the dehydrated intermediate formed during hydrogenolysis to
target products. Owing to the oxophilic nature of tungsten
oxide, the LUMOs of tungsten can capture electrons from H,
which results in reduced tungsten and Hd+ and gives rise to
Brønsted acidity.[6] This redox cycle[2b,7] is the source of the
unique characteristics of tungsten oxide in hydrogen-related
reactions. Accordingly, H2 is expected to heterolytically dissoci-
ate to Hd+ (over W) and HdÀ (over Pt) at the interface of Pt and
WOx, respectively, which in turn may be the cause for both de-
hydration and hydrogenation occurring simultaneously during
the reaction. It should be noted that during dehydration, the
secondary carbocation is more stable than the primary one;
thus, fast stabilization of the secondary carbocation with HdÀ is
the key step to achieve a high selectivity to 1,3-PD. Therefore,
maximizing the interface between Pt and WOx may induce op-
timized heterolytic dissociation of H2 and lead to a high 1,3-PD
selectivity under low H2 pressure. In other words, the hydroge-
nolysis performance can be optimized using a good, rational
use of H2, which in turn strongly depends on the chemistry of
catalysts and their capacity for H2 dissociation.
high space–time yield towards 1,3-propanediol (3.78 ggPtÀ1 hÀ1
)
in Pt–W catalysts. The highly isolated Pt structure is thought to
contribute to the excellent H2 dissociation capacity over Pt/
WOx. The high selectivity towards 1,3-propanediol is attributed
to the heterolytic dissociation of H2 at the interface of Pt and
WOx (providing specific Brønsted acid sites and the concerted
dehydration–hydrogenation reaction) and the bond formation
between glycerol and WOx, which favors/stabilizes the forma-
tion of a secondary carbocation intermediate as well as trig-
gers the redox cycle of the W species (W6+QW5+).
Glycerol, a byproduct of the biodiesel and soap synthesis, is
available in surplus amounts and requires efficient valorization
to high value-added chemicals.[1] Owing to its high oxygen
content (C/O=1), selective hydrogenolysis, whereby CÀO
bonds are cleaved or undergo lysis by hydrogen, is considered
an atom-economical and cost-competitive process.[2] Among
the products, 1,3-propanediol (1,3-PD) is the most desirable
one due to its wide application as monomer in the polyester
industry [polytrimethylene terephthalate (PTT)].[3] However, this
reaction is rather challenging as the formation of 1,2-propane-
diol (1,2-PD) is thermodynamically favored in comparison to
In co-supported Pt–W catalysts, the dispersion of Pt and W
species (key for high 1,3-PD selectivity)[4c,e,8] and the acidity of
the catalysts (key for high activity) strongly depend on the se-
lected support, which in turn interferes with the elucidation of
the true active center and reaction mechanism. For this reason,
we employed tungsten oxide alone as the support to rule out
subordinate/interferential factors from additional supports;
thus, the direct interaction between Pt and W species can be
detected, allowing mechanistic insights into structure–activity
relationships. Mesoporous WOx,[9] rich in oxygen vacancies and
hydroxyl functionalities, was employed as the support instead
of well-crystallized WO3 because 1) high surface areas may im-
prove Pt dispersion and 2) abundant oxygen vacancies may
strengthen the interaction between WOx and Pt, thereby stabi-
lizing Pt isolation.[10] Herein, we developed a WOx-supported
single/pseudo-single atom Pt catalyst, which exhibited excep-
tionally high hydrogenolysis activity under low H2 pressure
[a] J. Wang,+ Dr. X. Zhao,+ N. Lei, Dr. L. Li, Dr. L. Zhang, Dr. S. Miao, X. Pan,
Prof. Dr. A. Wang, Prof. Dr. T. Zhang
State Key Laboratory of Catalysis
Dalian Institute of Chemical Physics
Chinese Academy of Sciences
Dalian 116023 (PR China)
E-mail: aqwang@dicp.ac.cn
taozhang@dicp.ac.cn
[b] J. Wang,+ N. Lei
University of Chinese Academy of Science
Beijing 100049 (PR China)
[c] Dr. S. Xu
National Engineering Laboratory for Methanol to Olefins
Dalian National Laboratory for Clean Energy
Dalian Institute of Chemical Physics
Chinese Academy of Sciences
(1 MPa, other reported catalysts demand
a H2 pressure
Dalian 116023 (PR China)
[+] These authors contributed equally to this work.
>4 MPa), and the best space–time yield towards 1,3-PD
(3.78 ggPtÀ1 hÀ1) among all reported results for Pt–W catalysts.
Because lower H2 pressure is desirable in view of reducing the
Supporting Information for this article can be found under http://
dx.doi.org/10.1002/cssc.201501506.
ChemSusChem 2016, 9, 784 – 790
784
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
related capital cost in industrial operation, this new Pt/WOx
catalyst, as well as the preparation strategy, is of great signifi-
cance in H-related reactions.
The textual structures of Pt/WOx, WOx, their reference cata-
lyst Pt/WO3, and the support WO3 were analyzed using N2 ad-
sorption–desorption isotherms (Figure S1 in the Supporting In-
formation). The type-IV isotherms with H4 hysteresis indicated
that slit mesopores were formed in WOx and basically main-
tained in Pt/WOx. These mesopores were formed by stacks of
needle-shaped tungsten suboxide crystals (Figure 1b), which
endowed a larger surface area to WOx (126 m2 gÀ1) and the re-
sulting Pt/WOx catalyst (82 m2 gÀ1). According to XRD results
(Figure S2), the obtained tungsten suboxide was assigned to
WO2.8 (PDF 00-041-0745) whereas Pt/WOx and Pt/WO3 shared
a similar crystal structure corresponding to WO2.92 (PDF 00-030-
1387). Unlike Pt/WO3, Pt signals could not be detected in the
XRD patterns of Pt/WOx, indicating that smaller Pt particles
were formed over WOx, which is consistent with the TEM re-
sults (Figure 1c and d). To further identify the dispersion and
particle size of Pt, high angle annular dark field (HAADF)–STEM
was performed; an average particle size of 2.3 nm was estimat-
ed for Pt in Pt/WO3 (Figure 1e). Surprisingly, although a loading
of 2.59 wt% homogeneously dispersed Pt was detected in the
mapping area (Figure 1g), no clear subnano clusters, or even
bright single atoms, of Pt were observed in Pt/WOx (Figure 1 f
and Figure S3), suggesting that the Pt species were dispersed
at a single or/and pseudo-single atom scale (less than ten
atoms Pt-cluster). However, due to the close atomic numbers
of Pt and W (only 4 apart), their contrast in HAADF–STEM was
too low to obtain an exact number of atom numbers for Pt. In
good agreement with H2 chemisorption results, a microcalori-
metric study for H2 adsorption (Figure 2a) revealed that over
Pt/WOx the initial H2 differential heat (108 kJmolÀ1) is higher
than over Pt/WO3 (92 kJmolÀ1) whereas the H2 uptake over Pt/
WOx is also larger than that over Pt/WO3.[11] This further con-
vinced us that Pt is highly dispersed on WOx. As most studies
suggested that H2 was dissociated on PtÀPt sites, we further
investigated why an enhanced H2 dissociation capacity is ob-
tained on this single/pseudo-single atom Pt catalyst. The as-
prepared catalyst without reduction was first characterized as
reference using in situ X-ray photoelectron spectroscopy (XPS;
Figure 2b); following an in situ reduction at 573 K for 1 h, the
Pt/WOx catalyst was again analyzed to monitor the effect of H2
during the reduction. Clearly, two new W species were gener-
ated after H2 reduction; W3 is assigned to the oxidation of W1
whereas W2 is assigned to the reduction of W1. We attribute
the appearance of these two new W species to the heterolyti-
cal H2 dissociation over tungsten oxide whereas dissociated
Hd+ and HdÀ gave rise to the oxidation and the reduction reac-
tions, respectively.
Figure 1. SEM images for (a) WO3 and (b) WOx; TEM images for (c) Pt/WO3
and (d) Pt/WOx; HAADF–STEM images for (e) Pt/WO3 and (f) Pt/WOx; (g) EDS-
mapping images of W, O, Pt for Pt/WOx.
whether the dispersion of Pt is on a single-atom scale. There-
fore, we called the catalyst WOx-supported single/pseudo-
single atom Pt catalyst.
In addition, we also attempted to use in situ CO-adsorption
FTIR spectroscopy to distinguish between linear and bridge
bond formation of CO adsorption[10c] to identify single/pseudo-
single atom Pt. However, for Pt/WOx, the adsorption of CO was
too weak to withstand evacuation and only gas-phase CO was
detected. This indirectly confirmed that Pt was highly dis-
persed on WOx, but it would still be adventurous to clarify
This high dispersion of Pt suggests a strong interaction be-
tween Pt and WOx, which could be attributed to the presence
of oxygen vacancies in WOx (Figure S2 XRD and Figure S4 UV/
Vis). As dissociated Pt4+ species come in contact with W5+
,
they tended to remain at the vacancies and were reduced to
ChemSusChem 2016, 9, 784 – 790
www.chemsuschem.org
785
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
Figure 2. (a) Isotherms and differential heat of H2 adsorption on Pt/WOx and
Pt/WO3; (b) in situ XPS results for Pt/WOx and reduced Pt/WOx.
Figure 3. (a) XPS patterns of Pt4f for Pt/WOx and Pt/WO3. (b) NH3-TPD (m/
e=16) curves of Pt/WOx, Pt/WO3, WOx, and WO3. Enlarged signals of WO3
and Pt/WO3 are on the right.
less positive Pt species whereas W5+ was oxidized to W6+ (evi-
denced by the color change of WOx after Pt impregnation, Fig-
ure S5).[9] This in situ redox reaction occurred at the interface
simultaneously as the impregnation began, and in turn, signifi-
cantly influenced the subsequent reduction behavior.[9] Accord-
ing to H2 temperature-programmed reduction (TPR) results
(Figure S6), the reduction temperature of Pt in Pt/WOx (355 K)
is much higher than that in Pt/WO3 (276 K). Generally, a small
particle size will facilitate the kinetics of reduction and allow
a lower reduction temperature unless the strong metal support
interaction (SMSI) plays an overwhelming role in suppressing
the Pt reduction. Combined with the electron microscopy
images in Figure 1, we attribute the higher Pt reduction tem-
perature in Pt/WOx to the SMSI effect. For Pt/WO3, H2 con-
sumed for the sharp peak centered at 275 K is calculated to
stoichiometrically reduce PtO2 to Pt; for Pt/WOx, the first peak
centered at 355 K takes up almost twice the amount of H2 as
Pt/WO3, implying that besides Pt some adjacent W species
were reduced simultaneously.
In the patterns of Pt4f, only one species (Pt0) could be detect-
ed in Pt/WOx, which in contrast to Pt/WO3 for which both Pt0
and PtII species were found. We attributed the homogeneous
presence of Pt0 to the SMSI that prevented the oxidation of
metallic Pt into PtII even after exposure to air. The surface acid
sites present on 2 wt% Pt/WOx and Pt/WO3 were characterized
by NH3 temperature-programmed desorption (TPD; Figure 3b).
After loading of Pt, clear weak-acid peaks appeared at around
430 K over both WOx and WO3, which can be attributed to the
hydroxyl groups on tungsten atoms associated with Pt.[12] Esti-
mated from the NH3 adsorption amount, the acid-site densities
of Pt/WOx increased from 6.38 (WOx) to 7.95 mmolmÀ2 after
loading with Pt and were slightly larger than that of Pt/WO3
(Table 1). All acid sites over Pt/WOx and Pt/WO3 were assigned
to weak–middle strength acids, and most of them were con-
firmed as Brønsted acids by 31P NMR spectroscopy. To evaluate
the performance for glycerol hydrogenolysis over the as-pre-
pared Pt/WOx catalysts, first conventional reaction conditions[13]
(413 K, 5.5 MPa H2) were employed over Pt/WOx and its refer-
encePt/WO3. Although the conversion of glycerol and the ratio
of 1,3-PD/1,2-PD over Pt/WOx is improved compared to that
over Pt/WO3 (Table 2, entries 3 and 4), the absolute conversion
of glycerol and the yield of 1,3-PD is relatively low compared
to the best results reported so far.
The first peak is directly followed by a distinct reduction oc-
curring at roughly 367 K, attributed to the reduction of less
strongly interacting W species. Unlike previously reported,
single-atom/subnano cluster catalysts bearing very small
amounts of Pt group metals,[10] the Pt loading on WOx was as
high as 2 wt% and could thus be analyzed by XPS (Figure 3a).
ChemSusChem 2016, 9, 784 – 790
www.chemsuschem.org
786
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
Table 1. Textural parameters, acid properties, and H2 uptakes of Pt/WOx,
Pt/WO3, and their supports.
[a]
[b]
Sample
SBET
Dpore
Acid density[c]
H2 uptake
[mmolgÀ1
]
[m2 gÀ1
]
[nm]
[mmolmÀ2
]
Pt/WOx
WOx
Pt/WO3
WO3
82
126
16
6.5
3.4
20.4
18.5
7.95
6.38
6.06
6.88
0.137
–
0.096
–
17
[a] Brunauer–Emmett–Teller surface area. [b] The average pore diameters.
[c]Acid densities were calculated according to the amount of adsorbed
NH3.
However, optimization of reaction conditions revealed an ex-
citing catalytic behavior over Pt/WOx: the conversion of glycer-
ol greatly increased, without compromising the 1,3-PD selectiv-
ity, as a negative function of H2 pressure [up to 1.0 MPa;
Tables 2 entry 6 and S1; the high ratio of 1,3-PD/1,2-PD was at-
tributed to the likely secondary reaction of 1,2-PD to 1-propa-
nol (1-PO) (Table S1)]. This behavior was not observed on Pt/
WO3.Therefore, compared with Pt/WO3, the 1,3-PD/1,2-PD ratio
increased from 3.4 to 13.1 over Pt/WOx when reducing the H2
pressure from 5.5 to 1.0 MPa (Table 2, entries 5 and 6). The cal-
culated space–time yield of 1,3-PD for Pt/WOx is approximately
3.78 ggPtÀ1 hÀ1, to the best of our knowledge the highest
space–time yield reported on Pt–W catalysts to date (Tables 2,
entry 10 and S2). Moreover, this value was obtained under
a H2 pressure of 1 MPa, much milder than other reported reac-
tion conditions, which would drastically reduce the capital cost
in practical industry. To rule out the influence of surface areas
and consequently variance of Pt dispersion and acid-site
amount, a control experiment was conducted employing
8 times of WO3 as support but keeping the amount of Pt
loaded constant (Figure S8 and Table 2, entry 8). The results
showed that the surface area and the amount of acid sites
play very limited roles in reactivity improvement; however, the
Pt dispersion and its associated acid sites are crucial in this re-
action. Taking advantage of the mild reaction conditions, Pt/
Figure 4. (a) Stability test of Pt/WOx was conducted in fixed-bed reactor at
413 K, 1.0 MPa, gas hourly space velocity GHSV=1000 hÀ1, liquid hourly
space velocity LHSV=1 hÀ1, with 5 wt% glycerol. (b) TEM (left) and HAADF-
STEM images of the used Pt/WOx catalyst (middle and right; scale bars corre-
spond to 200 and 50 nm).
WOx exhibited a promising stability in 75 h run without notice-
able agglomeration of Pt (Figure 4).
Generally, a higher hydrogen pressure is favorable for hydro-
genolysis. In this work, however, the best performance was ob-
tained under circa 1 MPa H2; increasing H2 pressure afterwards
resulted in a decreased glycerol conversion. According to the
H2 chemisorption results (Table 1), we attribute this uncommon
shift to lower pressure to the superior dissociating H2 capacity
over the WOx-supported single/pseudo-single atom Pt catalyst,
which facilitated the hydrogenolysis of glycerol under low H2
pressure and, conversely, “poisoned” the catalysts under high
H2 pressure. Comparison of the used catalysts under 1 and
5.5 MPa showed that higher H2 pressure may further reduce
W5+ to W4+ (Figure 5), which limits the catalytic redox cycle of
Table 2. Hydrogenolysis of glycerol over tungsten oxide supported Pt catalysts.[a]
Entry
Catalyst
T
[K]
PH
[MPa]
Conv.[b]
[%]
Selectivity [%]
1,2-PD
1,3-PD/1,2-PD
Yield1,3-PD
[%]
2
2-PO
1-PO
1,3-PD
others[c]
1
2
3
4
5
6
7
8[d]
9[e]
10[e]
11[f]
WO3
WOx
413
413
413
413
413
413
433
413
413
433
413
5.5
5.5
5.5
5.5
1.0
1.0
1.0
1.0
1.0
1.0
1.0
–
–
3.2
14.2
2.2
37.4
49.5
6.0
6.6
16.1
59.8
2%Pt/WO3
2%Pt/WOx
2%Pt/WO3
2%Pt/WOx
2%Pt/WOx
2%Pt/WO3
2%Pt/WOx
2%Pt/WOx
2%Pt/WOx
4.9
5.2
3.9
5.1
4.1
7.2
0.0
4.3
5.8
31.2
47.4
36.2
50.3
52.3
52.8
40.9
45.8
48.2
9.9
1.1
5.9
2.3
6.5
43.7
33.0
45.7
35.1
28.1
27.8
43.1
34.0
36.3
10.3
13.3
8.3
7.3
9.0
7.6
5.7
5.3
6.6
4.4
29.6
7.7
15.3
4.3
6.0
4.2
3.2
11.7
1.4
4.7
1.0
13.1
13.9
1.6
2.9
5.5
21.7
4.6
10.3
10.6
3.1
[a] The conversion deviation is Æ4.5%, the 1,3-PD selectivity deviation is Æ6.3%. [b] Typical reaction conditions: 12 g 5 wt% glycerol aqueous solution in
75 mL autoclave with Teflon lining, 0.3 g of catalyst, stirring at 800 rpm, 413 K, 12 h. [c] Others include propylene, ethylene glycol, ethanol, methanol, meth-
ane, and ethane. [d] 2.4 g 0.25%Pt/WO3 catalysts were employed under the reaction conditions of 413 K, 12 h, and H2 pressure 1.0 MPa. [e] 12 g 50 wt%
glycerol aqueous solution was used as the substrate. [f] the amount of catalysts is doubled to 0.6 g.
ChemSusChem 2016, 9, 784 – 790
www.chemsuschem.org
787
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
Figure 5. W4f XPS spectra of fresh and used Pt/WOx under 1 and 5.5 MPa
H2, respectively.
W6+QW5+ and in turn decreases the activity under high H2
pressure.
To further understand the hydrogenolysis process, a series of
in situ characterizations were conducted on Pt/WOx as a func-
tion of reaction time to monitor the variance of W state
(Figure 6). In Raman spectra, three strong bands at 706, 785,
and 959 cmÀ1 were detected in Pt/WOx (Figure 6a). The former
two were assigned as n(OÀWÀO) modes of the bridging
oxygen atoms in the WO6 octahedra, whereas the latter one
was assigned to the stretching mode of the terminal W=O
bond on the surface, which can extract protons in many sur-
face reactions.[14] The terminal-W=O band (959 cmÀ1) decreased
dramatically after coming in contact with glycerol whereas the
785 cmÀ1 band was shifted to 796 cmÀ1 and the 706 cmÀ1
band, sensitive to cation intercalation,[6,15] decreased slightly.
This change implies that the W=O bond was cleaved after the
adsorption of glycerol whereas WOx was reduced by the insert-
ed proton. In situ X-ray absorption near edge structure
(XANES) spectra further evidenced this implication, in which
the general valence of W was shown to gradually decrease
from the beginning of the reaction (Figure 6b).
Figure 6. (a) Raman spectra of Pt/WOx, Pt/WOx impregnating in glycerol, and
in situ Raman spectra of Pt/WOx in glycerol at 413 K under H2 atmosphere
after 3 h. (b) in situ XANES spectra at W L3 edge; the sample after 12 h
(purple) corresponds to the used catalysts.
WÀOH (Figure 6a) whereas W6+ was partially reduced (Fig-
ure 6b), which is consistent with our previous study employing
Ti–W composite oxide as the support.[13] After the “adsorption”
of glycerol, Brønsted acid sites were consumed, catalyzing the
dehydration of glycerol (Figure S7). In contrast to the usually
formed H bond between glycerol and metal oxide supports,
the strengthened bonding with WOx favored/stabilized the for-
mation of the secondary carbocation from glycerol. Moreover,
owing to the oxophilic nature of W, H2 is assumed to be heter-
olytically dissociated on the interface between Pt and WOx, ex-
hibiting both acid sites (Hd+) and hydrogenation sites (HdÀ).
This characteristic speeds up the hydrogenation of 3-hydroxy-
propanal before it transforms to the thermodynamically fa-
vored acetol and gives rise to a high selectivity to 1,3-PD. In
this work, the interface between Pt and WOx was maximized
by the design of a single/pseudo-single atom Pt catalyst, and
thus high H2 pressure was unnecessary to increase the hydro-
genation rate.
Inspired by the photochromism of tungsten oxide[16] and
taken all current characterization and performance results to-
gether, we, therefore, proposed a reaction scheme as following
(Scheme 1):
First, the unshared-pair electrons of oxygen atoms in glycer-
ol were trapped by the unoccupied d orbital of W6+, forming
an ether-like bond with a W atom. The strong interaction with
the W atom, in turn, weakened the bond between the O and
H atoms and facilitated the oxidation of a H atom. The proton
was then extracted by the terminal O atom of W=O to form
In summary, a new WOx-supported single/pseudo-single
atom Pt catalyst was designed by employing large surface
area, mesoporous tungsten suboxide as the support. In the se-
lective hydrogenolysis of glycerol,
a
productivity of
3.78 ggPtÀ1 hÀ1 was achieved for 1,3-PD over Pt/WOx, to the
ChemSusChem 2016, 9, 784 – 790
www.chemsuschem.org
788
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
new Pt/WOx catalyst can be further extended to photocatalysis
and photo-electrocatalysis.
Experimental Section
WOx was prepared using a modified procedure according to Xi
et al.[9] In detail, WCl6 (3 g) (Aladdin) was added to ethanol
(100 mL) while stirring mildly for 20 min and then transferred to
a Teflon-lined autoclave and heated in an oven at 433 K for 36 h.
After cooling to room temperature, dark blue products were recov-
ered, which were washed with ethanol and water. WOx was ob-
tained by drying the as-prepared samples at 323 K for 6 h in
vacuum. Pt/WOx was prepared by impregnating WOx with H2PtCl6
solution (6.7 wt%; Tianjin Fengchuan Chemical Reagent Technolo-
gies Co., Ltd., Pt >37.0%) overnight before drying at 323 K for 6 h
in vacuum and then reducing in flowing H2 at 573 K for 1 h; the
catalyst was then passivated using 1% O2/N2 for 4 h at room tem-
perature. Pt/WO3 was prepared by impregnating WO3 with H2PtCl6
(same conditions as above) overnight and drying at 373 K over-
night, which was followed by calcination at 673 K for 1 h. The cata-
lyst was then reduced in flowing H2 at 573 K for 1 h, and passivat-
ed with 1% O2/N2 for 4 h at room temperature.
Catalytic activities were carried on in a 75 mL of autoclave with
Teflon lining. Typically, catalyst (0.3 g ) and glycerol aqueous solu-
tion (12 g; 5 wt%; Tianjin Kermel Chemical Reagent Co., Ltd.
>99.0%) were put into the autoclave and flushed with H2 several
times. The reaction was conducted at 413 K and 1 MPa H2 for 12 h.
The gas and the liquid products were collected and analyzed sepa-
rately. The conversion (conv.) of glycerol and the selectivity (sel.) of
each liquid product were calculated using Equations (1) and (2), re-
spectively:
Conv: ½%Š ¼ ðmolglycerol consumedÞ=ðmolglycerol initially addedÞ Â 100
ð1Þ
ð2Þ
Sel: ½%Š ¼ ðmolcarbon in specific productÞ=ðmolcarbon in consumed glycerolÞ Â 100:
Acknowledgements
The authors appreciate financial support by the National Natural
Science Foundation of China (21176235, 21373206, and
21303187). The authors greatly thank Leifeng Gong and Hua
Wang for product analysis, Xiong Su for solid-acid-analysis dis-
cussions, Botao Qiao for single-atom-catalysis discussions, and
Guanfeng Liang, Longjie Liu, Chaojun Yang, and Yao Wang for
fruitful discussions.
Scheme 1. Proposed reaction scheme for hydrogenolysis of glycerol to 1,3-
PD over Pt/WOx.
best of our knowledge, the highest space–time yield among all
reported results over Pt–W catalysts. Moreover, this per-
formance was obtained under mild conditions (433 K, 1 MPa
H2) using a relatively high glycerol concentration (50%), which
means a great cost reduction in industrial operation. The
trends in characterizations and performances allowed mecha-
nistic insight into the structure–activity relationship and further
enabled the optimization of reaction conditions and catalyst
preparation. In the current study, precise identification of Pt
atoms are still challenging owing to the close atomic number
of W and Pt; therefore, efficient characterization and theoreti-
cal calculations to clarify/support the isolation of Pt over WOx
are highly desired. In addition, due to the unique photochro-
mic characteristics of tungsten oxide, the application of this
Keywords: 1,3-propanediol
· glycerol · hydrogenolysis ·
supported catalysts · tungsten oxide
[1] a) J. ten Dam, U. Hanefeld, ChemSusChem 2011, 4, 1017–1034; b) Y. Na-
kagawa, M. Tamura, K. Tomishige, J. Mater. Chem. A 2014, 2, 6688; c) C.-
H. Zhou, J. N. Beltramini, Y.-X. Fan, G. Q. Lu, Chem. Soc. Rev. 2008, 37,
527–549.
[2] a) A. M. Ruppert, K. Weinberg, R. Palkovits, Angew. Chem. Int. Ed. 2012,
51, 2564–2601; Angew. Chem. 2012, 124, 2614–2654; b) N. Ota, M.
Tamura, Y. Nakagawa, K. Okumura, K. Tomishige, Angew. Chem. Int. Ed.
2015, 54, 1897–1900; Angew. Chem. 2015, 127, 1917–1920.
[3] A. Martin, M. Richter, Eur. J. Lipid Sci. Technol. 2011, 113, 100–117.
ChemSusChem 2016, 9, 784 – 790
www.chemsuschem.org
789
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
[4] a) S. Zhu, X. Gao, Y. Zhu, Y. Li, J. Mol. Catal. A 2015, 398, 391–398; b) S.
García-Fernµndez, I. Gandarias, J. Requies, M. B. Güemez, S. Bennici, A.
Auroux, P. L. Arias, J. Catal. 2015, 323, 65–75; c) R. Arundhathi, T. Mizu-
gaki, T. Mitsudome, K. Jitsukawa, K. Kaneda, ChemSusChem 2013, 6,
1345–1347; d) J. ten Dam, K. Djanashvili, F. Kapteijn, U. Hanefeld, Chem-
CatChem 2013, 5, 497–505; e) L.-Z. Qin, M.-J. Song, C.-L. Chen, Green
Chem. 2010, 12, 1466–1472.
[5] a) Y. Nakagawa, Y. Shinmi, S. Koso, K. Tomishige, J. Catal. 2010, 272,
191–194; b) Y. Amada, Y. Shinmi, S. Koso, T. Kubota, Y. Nakagawa, K. To-
mishige, Appl. Catal. B 2011, 105, 117–127; c) Y. Nakagawa, X. Ning, Y.
Amada, K. Tomishige, Appl. Catal. A 2012, 433–434, 128–134; d) J.
Guan, X. Chen, G. Peng, X. Wang, Q. Cao, Z. Lan, X. Mu, Chin. J. Catal.
2013, 34, 1656–1666.
[6] D. G. Barton, M. Shtein, R. D. Wilson, S. L. Soled, E. Iglesia, J. Phys. Chem.
B 1999, 103, 630–640.
[7] T. Yang, Y. Zhang, C. Li, J. Alloys Compd. 2014, 584, 546–552.
[8] a) T. Kurosaka, H. Maruyama, I. Naribayashi, Y. Sasaki, Catal. Commun.
2008, 9, 1360–1363; b) L. Gong, Y. Lu, Y. Ding, R. Lin, J. Li, W. Dong, T.
Wang, W. Chen, Appl. Catal. A 2010, 390, 119–126.
5634; c) B. Qiao, A. Wang, X. Yang, L. F. Allard, Z. Jiang, Y. Cui, J. Liu, J.
Li, T. Zhang, Nat. Chem. 2011, 3, 634–641; d) X.-F. Yang, A. Wang, B.
Qiao, J. Li, J. Liu, T. Zhang, Acc. Chem. Res. 2013, 46, 1740–1748.
[11] a) L. Li, X. Wang, X. Zhao, M. Zheng, R. Cheng, L. Zhou, T. Zhang, Ther-
mochim. Acta 2005, 434, 119–124; b) D. Uner, M. Uner, Thermochim.
Acta 2005, 434, 107–112.
[12] a) D. Hibbitts, Q. Tan, M. Neurock, J. Catal. 2014, 315, 48–58; b) M. Chia,
Y. J. Pagµn-Torres, D. Hibbitts, Q. Tan, H. N. Pham, A. K. Datye, M. Neu-
rock, R. J. Davis, J. A. Dumesic, J. Am. Chem. Soc. 2011, 133, 12675–
12689.
[13] Y. Zhang, X.-C. Zhao, Y. Wang, L. Zhou, J. Zhang, J. Wang, A. Wang, T.
Zhang, J. Mater. Chem. A 2013, 1, 3724–3732.
[14] C. Santato, M. Odziemkowski, M. Ulmann, J. Augustynski, J. Am. Chem.
Soc. 2001, 123, 10639–10649.
[15] a) S. Balaji, A.-S. Albert, Y. Djaoued, R. Brüning, J. Raman Spectrosc.
2009, 40, 92–100; b) S. Balaji, Y. Djaoued, A.-S. Albert, R. Bruning, N.
Beaudoin, J. Robichaud, J. Mater. Chem. 2011, 21, 3940–3948.
[16] a) U. Tritthart, W. Gey, A. Gavrilyuk, Electrochim. Acta 1999, 44, 3039–
3049; b) N. Tahmasebi Garavand, M. Ranjbar, S. M. Mahdavi, A. Irajizad,
Appl. Surf. Sci. 2012, 258, 10089–10094.
[9] G. Xi, J. Ye, Q. Ma, N. Su, H. Bai, C. Wang, J. Am. Chem. Soc. 2012, 134,
6508–6511.
[10] a) J. Lin, B. Qiao, J. Liu, Y. Huang, A. Wang, L. Li, W. Zhang, L. F. Allard, X.
Wang, T. Zhang, Angew. Chem. Int. Ed. 2012, 51, 2920–2924; Angew.
Chem. 2012, 124, 2974–2978; b) H. Wei, X. Liu, A. Wang, L. Zhang, B.
Qiao, X. Yang, Y. Huang, S. Miao, J. Liu, T. Zhang, Nat. Commun. 2014, 5,
Received: November 8, 2015
Revised: December 19, 2015
Published online on February 23, 2016
ChemSusChem 2016, 9, 784 – 790
www.chemsuschem.org
790
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim