Catalysis Today 212 (2013) 120–126

Contents lists available at SciVerse ScienceDirect

Catalysis Today
journal homepage: www.elsevier.com/locate/cattod

Hydrogenolysis of glycerol to 1,3-propanediol over bifunctional catalysts

containing Pt and heteropolyacids
Shanhui Zhu a,b , Yanan Qiu c , Yulei Zhu a,d,∗ , Shunli Hao d , Hongyan Zheng d , Yongwang Li a,d
a
State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, PR China
b
Graduate University of Chinese Academy of Sciences, Beijing 100039, PR China
c
Institute of Science & Technology Information of Shanxi, Taiyuan 030001, PR China
d
Synfuels China Co. Ltd., Taiyuan 030032, PR China

a r t i c l e i n f o a b s t r a c t

Article history: Hydrogenolysis of glycerol to 1,3-propanediol was conducted over zirconia supported bifunctional cata-
Received 11 March 2012 lysts containing Pt and heteropolyacids using H4 SiW12 O40 (HSiW), H3 PW12 O40 (HPW) and H3 PMo12 O40
Received in revised form 27 June 2012 (HPMo) as active compounds. Pt/ZrO2 was also examined for comparison. The as-prepared catalysts were
Accepted 2 September 2012
characterized by BET, CO chemisorption, XRD, Raman spectra, NH3 -TPD and FTIR of adsorbed pyridine.
Available online 18 October 2012
Compared with Pt/ZrO2 , heteropolyacids modified Pt/ZrO2 catalysts showed higher acidity and better
catalytic performance of glycerol hydrogenolysis to 1,3-propanediol. Among them, Pt-HSiW/ZrO2 exhib-
Keywords:
ited superior performance due to the high Brønsted acid sites and good thermal stability. Independent
Glycerol
Hydrogenolysis of the heteropolyacid type, the concentration of Brønsted acid sites appeared as a key to the selective
1,3-Propanediol formation of 1,3-propanediol from glycerol hydrogenolysis, whereas the concentration of Lewis acid sites
Heteropolyacids was related to the formation of 1,2-propanediol. We also investigated the reaction network and proposed
a possible reaction pathway.
© 2012 Elsevier B.V. All rights reserved.

1. Introduction synthesis via hydration of acrolein or hydroformylation of ethyl-

The catalytic transformation of biomass-derived feedstocks for
the synthesis of value-added fuels, chemicals and materials has
been attracting much attention, due to the diminishing reserves
and rising price of fossil resources [1,2]. Recent development of
biodiesel production by transesterification of plant oils and animal
fat has made a large amount of glycerol available as a renewable
feedstock for the production of various fuels and valuable chemicals
[2–5]. Significant research efforts have been focused on the conver-
sion of glycerol by various catalytic processes such as reforming,
oxidation, dehydration, hydrogenolysis, etherification, esterifica-
tion and so on [3–10].
Hydrogenolysis of glycerol to 1,2-propanediol (1,2-PDO) and
1,3-propanediol (1,3-PDO) is a noticeable pathway due to its
production of renewable value-added chemicals. 1,3-PDO owes
much higher economical value than 1,2-PDO, in particular, as an
important monomer in the synthesis of polyester fibers [4,5,11].
Industrial production of 1,3-PDO is currently based on chemical
∗ Corresponding author at: State Key Laboratory of Coal Conversion, Institute of
Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, PR China.
Tel.: +86 351 7117097; fax: +86 351 7560668.
E-mail address: zhuyulei@sxicc.ac.cn (Y. Zhu).
ene oxide [12]. The development of an alternative environmentally
benign process for the production of 1,3-PDO based on the biomass-
derived glycerol as a sustainable feedstock is a real challenge.
Supported Ir, Pt, Ru, Rh, Pd and Cu catalysts have been used
for glycerol hydrogenolysis to 1,3-PDO [9,10,13–21]. Among these
catalysts, supported Pt catalysts are more selective for this reac-
tion than others. To obtain 1,3-PDO effectively, it is necessary
to add acid component to the catalysts due to the bi-functional
mechanism. Previous research [9,22] shows that hydrogenoly-
sis of glycerol to 1,3-PDO proceeds via dehydration of glycerol
to 3-hydroxypropaldehyde (3-HPA) followed by hydrogenation.
According to the above mechanism, acid catalysts play an important
role in glycerol hydrogenolysis. Various acid catalysts including
H2 WO4 , H2 SO4 , sulfated zirconia, WO3 /ZrO2 , and WO3 /TiO2 /SiO2
have been used [9,10,18,21,23,24]. Studies by Chaminand et al. [23]
showed that hydrogenolysis of glycerol to 1,3-PDO on Rh/SiO2 cat-
alyst achieved 4% yield in the presence of H2 WO4 at 200 ◦ C and
8.0 MPa. Tomishige and co-workers [18] reported that the yield of
1,3-PDO reached 38% over Ir-ReOx /SiO2 catalyst with H2 SO4 as an
additive in a batch reactor. However, the liquid acids would cause
several problems in terms of catalyst separation, reactor corro-
sion and environmental protection. Recently, Oh et al. [24] have
communicated the use of Pt-sulfated zirconia with 1,3-dimethyl-2-
imidazolidinone in glycerol hydrogenolysis but such strong acidic

0920-5861/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cattod.2012.09.011
 S. Zhu et al. / Catalysis Today 212 (2013) 120–126
catalysts deactivate quickly as carbonaceous deposits block the cat-
alyst surface. Additionally, the usage of organic solvent will greatly
reduce the environmental and economical viability.
Heteropolyacids (HPAs) have demonstrated to display out-
standing catalytic performance for both homogeneous and
heterogeneous acid-catalyzed reactions because of its well-defined
structure and widely tunable acidity [8,25–27]. HPAs possess
unique physicochemical properties, with their structural mobil-
ity and multifunctionality, which is useful in industrial processes
such as the liquid–phase dehydration of alcohols and hydration of
esters or olefins [26–28]. Additionally, heterogeneous acid catal-
ysis by HPAs has attracted much interest due to its potential of
great economic rewards and green benefits. Compared to con-
ventional solid acid catalysts such as oxides and zeolites, HPAs
possess strong acidity, uniform acid sites and easily tunable acid-
ity. Generally, tungsten HPAs owns stronger acidity, higher thermal
stability and lower oxidation potential than molybdenum HPAs,
which makes tungsten HPAs used more widely [29]. However, HPAs
lack thermal stability and leach easily in polar solvents. To solve
this problem, HPAs are often immobilized on supports such as
SiO2 or ZrO2 . Kozhevnikov and co-workers [27] have studied the
interaction between HPW and support with ammonia adsorption
calorimetry, MAS NMR and FTIR and suggested increasing interac-
tion in the following order: SiO2 < TiO2 < Nb2 O5 < ZrO2 . Moreover,
our previous study [13] showed that the Pt-HSiW/ZrO2 catalyst
was rather durable during one-step hydrogenolysis of glycerol to
biopropanols. Taken together, it is worthwhile to study supported
HPAs as acid catalysts for glycerol hydrogenolysis.
In the present investigation, several zirconia supported Pt-
HPAs (HSiW, HPW and HPMo) bi-functional catalysts were
prepared, characterized, and tested in aqueous solution for glyc-
erol hydrogenolysis. Special emphasis was also laid on formation
route of glycerol hydrogenolysis and degradation based on the
catalytic performance in the reaction of products and detected
intermediates.
2. Experimental
2.1. Catalyst preparation
ZrO2 (Jiangsu Qianye Co., Ltd., China) was used as support of
the catalysts. H2 PtCl6 ·6H2 O (AR), HSiW (AR), HPW (AR) and HPMo
(AR) were purchased from Sinopharm Chemical Reagent Co., Ltd.
All the catalysts were prepared by incipient wetness impregnation
method. Typical procedures for the preparation of Pt-HSiW/ZrO2
catalyst are as follows. The Pt/ZrO2 catalyst was prepared by
impregnation of zirconia support with an aqueous solution of
H2 PtCl6 ·6H2 O. The impregnated sample was dried overnight at
110 ◦ C and then calcined at 400 ◦ C in static air for 4 h. After impreg-
nated HSiW onto Pt/ZrO2 , Pt-HSiW/ZrO2 was dried overnight at
110 ◦ C and then calcined at 350 ◦ C in static air for 4 h. Pt-HPW/ZrO2
and Pt-HPMo/ZrO2 were prepared in the similar way to that of
Pt-HSiW/ZrO2 . The Pt and HPAs loadings were fixed at 2 wt.% and
15 wt.%, respectively.
2.2. Catalyst characterization
N2 adsorption–desorption isotherms were recorded at
−196 ◦ C using a Micromeritics ASAP 2420 instrument after
degassing at 300 ◦ C for 8 h in vacuum. BET surface area and
BJH pore size distribution were calculated by desorption
isotherms.
CO chemisorption was measured at 50 ◦ C in Auto Chem. 2920
equipment (Mircromeritics, USA). Prior to the measurement, about
0.2 g catalyst sample was first reduced in situ for 2 h in pure H2 at
121
200 ◦ C, then purged with He at the same temperature for 1 h and
finally cooled down to 50 ◦ C. The CO chemisorption was operated
by pulse injection of pure CO at 50 ◦ C. The Pt particle size was cal-
culated by assuming an adsorption of one CO molecule per surface
platinum atom.
Powder X-ray diffraction (XRD) patterns were conducted at
room temperature on a D2/max-RA X-ray diffractometer (Bruker,
Germany) using Cu K␣ radiation at 30 kV and 10 mA. The X-ray
patterns were recorded in 2 values ranging from 10◦ to 90◦ at the
scanning rate of 5◦ /min.
Raman spectra were recorded with a LabRAM HR800 System
equipped with a CCD detector at room temperature. The 325 nm of
the He-Cd laser was used as the exciting source with a power of
30 MW.
NH3 -TPD was carried out in the same apparatus as CO
chemisorption. Prior to each run, the catalyst sample (0.3 g) was
pretreated in He at 350 ◦ C for 1 h, then cooled to 100 ◦ C and was
saturated with pure NH3 for 30 min. After being purged with He
for 30 min, the sample was heated to 700 ◦ C at a heating rate
of 10 ◦ C/min and the NH3 desorption was monitored with a TCD
detector.
To evaluate and analyze the nature and acid amount of cata-
lysts, FTIR spectra of adsorbed pyridine (Py-IR) of solid samples
were carried out on a VERTEX70 (Bruker) FT-IR spectrophotome-
ter, equipped with a deuterium triglycine sulphate (DTGS) detector.
Prior to pyridine adsorption, the samples were heated to 300 ◦ C
under vacuum, and then cooled to 30 ◦ C. The Py-IR spectra were
then recorded at 200 ◦ C after applying vacuum for 30 min. The
quantitative calculation of Brønsted and Lewis acid sites was deter-
mined by the integrated area of the adsorption bands at about 1540
and 1450 cm−1 , respectively. The acidity of Brønsted and Lewis
acid sites was calculated by using the adsorption coefficient values
reported in the previous paper [30].
2.3. Catalytic tests
Hydrogenolysis of glycerol was conducted in a vertical fixed-
bed reactor (i.d. 12 mm, length 600 mm) with an ice-water trap. In
a typical run, 2.0 g catalyst (20–40 mesh) was charged in the con-
stant temperature section of the reactor, with quartz sand packed
in both ends. Prior to the test, the catalyst sample was in situ
reduced in a stream of pure H2 (100 ml/min) at 200 ◦ C for 2 h. After
reduction, the system was cooled to the desired temperature. A
10 wt.% glycerol aqueous solution was fed continuously into the
reactor with an HPLC pump along with a co-feed H2 of gas flow-
ing at 100 cm3 min−1 . The liquid and gas products were cooled and
collected in a gas–liquid separator immersed in an ice-water trap.
The products were obtained when the reaction reached the steady
state. Additionally, the reaction performance of 1,3-PDO, 1,2-PDO
and ethylene glycol (EG) was also evaluated in order to reveal the
reaction pathway of glycerol hydrogenolysis.
The liquid products were analyzed by gas chromatography using
a capillary column (DB-WAX, 30 m × 0.32 mm) equipped with a FID
detector. The tail gas was off-line analyzed by a gas chromatogra-
phy with a capillary column (OV-101, 60 m × 0.25 mm) and a TCD
detector. The products obtained were also identified by GC–MS. The
conversion of glycerol and selectivity of products were calculated
as follows:
Conversion (%)
moles of glycerol (in) − moles of glycerol (out)
= × 100
moles of glycerol (in)
moles of one product
Selectivity (%) = × 100.
moles of all products
122 S. Zhu et al. / Catalysis Today 212 (2013) 120–126

Table 1
Physicochemical properties and acidities of Pt/ZrO2 and Pt-HPAs/ZrO2 catalysts.

Catalyst Surface area Pore size Pore volume Pt size (nm) Pt dispersion Total acidity Brønsted acidity Lewis acidity Total acidity
(m2 /g−1 ) (nm) (cm3 g−1 ) (%) (mmol NH3 /gcat )a (␮mol/gcat. )b (␮mol/gcat )b (␮mol/gcat )b

Pt/ZrO2 61.0 9.7 0.202 3.01 37.6 0.49 0 164.82 164.82

Pt-HSiW/ZrO2 54.8 9.8 0.164 3.16 35.8 0.53 86.56 85.59 172.15
Pt-HPW/ZrO2 57.0 9.4 0.171 3.06 37.0 0.63 62.89 117.18 180.07
Pt-HPMo/ZrO2 55.3 8.9 0.170 3.69 30.4 0.66 20.09 179.03 199.12
a
The amount of acid sites was determined by quantifying the desorbed NH3 from NH3 -TPD.
b
The amount of acid sites was determined by quantifying the desorbed pyridine from Py-IR.

3. Results and discussion

Pt-PMo/ZrO2
3.1. Catalyst characterizations

3.1.1. Physicochemical properties of catalysts

Table 1 lists the textural properties derived from nitrogen Pt-HPW/ZrO2
physisorption isotherms and dispersion of Pt determined by CO

Intensity
chemisorption of as-prepared catalysts. As can be seen, both surface
area and pore volume of Pt-HPAs/ZrO2 samples decreased as com- Pt-HSiW/ZrO2
pared to Pt/ZrO2 . This decrease may be attributed to pore blockage
caused by heteropolyacids (HPAs) as their Keggin units (diameter Pt/ZrO2
≈1.2 nm) are small enough to enter mircropores of ZrO2 [31]. As
shown in Table 1, all the catalysts have small Pt particle size and
good dispersion of Pt. Compared with Pt/ZrO2 sample, no remark- ZrO2
able changes of Pt particle size are observed for any supported HPAs
catalysts, implying that the HPAs have little effect on the disper-
sion of Pt. The slight decrease in Pt dispersion of Pt-HPMo/ZrO2 10 20 30 40 50 60
can be probably related to the partial coverage by HPMo or decom- 2θ (º)
position component. Cui et al. [32] have studied the effect of Mo
promoter in iron-based catalysts by high-resolution transmission Fig. 1. XRD patterns of bulk ZrO2 , Pt/ZrO2 and Pt-HPAs/ZrO2 catalysts.

electron microscopy and suggested that the presence of high Mo-

content would weaken the dispersion of iron active sites due to the
coverage effect.
Assuming that a cross section of the heteropolyacid (HPA)
molecule is 144 Å2 , the 15% HPA loading will correspond to
45 m2 /g monolayer coverage [27]. Due to the 61.0 m2 /g surface
area of Pt/ZrO2 , the supported HPA catalysts can be adequately
viewed as having sub-monolayer HPA coverage. The XRD patterns
(discussed below) confirmed the high dispersion of HPA on the sup-
port surface. The sub-monolayer HPA coverage can ensure strong
HPA/support interaction which can strengthen the thermal stabil-
ity and avoid the leaching of HPA during the reaction. Additionally,
the 15 wt.% loading is sufficient to provide efficient acid catalysis for
glycerol hydrogenolysis. Chai et al. [33] have investigated glycerol
dehydration over a series of HPW loading on zirconia and found
that the 15 wt.% HPW/ZrO2 showed superior catalytic performance
and most of the Keggin HPW remained intact after calcination up
to 650 ◦ C.
The XRD patterns of bulk ZrO2 and as-prepared catalysts are
illustrated in Fig. 1. ZrO2 presented in all the samples exhibited
characteristic peaks corresponding to monoclinic phase of ZrO2
(e.g., 24.1◦ , 28.3◦ , 31.5◦ , 34.3◦ , 49.4◦ ) [25]. It is of interest to note
that neither diffraction peaks of Keggin structure nor decomposi-
below 700 cm−1 interfere with the characteristic Raman bands
of HPAs species, while the range above 700 cm−1 is free from
characteristic bands of ZrO2 and hence relevant to the structure
determination of HPAs species. The typical Raman spectra bands of
Pt-HSiW/ZrO2 appeared at 998 and 975 cm−1 , which were assigned
to W O symmetric and asymmetric stretching vibrations of HSiW
Keggin structure, respectively [13]. The Pt-HPW/ZrO2 sample also
displayed two overlapping bands at around 1008 and 990 cm−1 ,
which can be ascribed to the symmetric and asymmetric stretch-
ing vibrations of the terminal W O bond of HPW Keggin anions,
respectively [33]. For Pt-HPMo/ZrO2 , Raman band for the Keggin
structure was observed at 996 cm−1 but a weak band characteris-
tic of MoO3 was also identified at 818 cm−1 [31]. Furthermore, a
HPMo
1/4
Pt-HPMo/ZrO2
HPW 1/4
Intensity

tion products WO3 or MoO3 were detected, most probably because

Pt-HPW/ZrO2
of the fine dispersion of HPAs or their decomposition species on
the support surface. As discussed above, the sub-monolayer HPA
coverage can avoid HPA crystallites on the support surface. On the HSiW 1/4
other hand, there were also no diffraction peaks of Pt, indicating Pt-HSiW/ZrO2
that Pt was highly dispersed on the support surface, which was
well in accordance with the CO chemisorption results. ZrO2
Raman spectroscopy is known to be well suited for observa-
tion the state of HPAs in the catalysts. As shown in Fig. 2, it has 200 400 600 800 1000 1200
been identified and attributed Raman bands at 222, 310, 337, -1
Raman shift (cm )
382, 474, 499, 540, 559, 620 and 636 cm−1 to monoclinic zirco-
nia [25], as evidenced by XRD results. The strong Raman bands Fig. 2. Raman spectra of bulk heteropolyacids and different catalysts.
 S. Zhu et al. / Catalysis Today 212 (2013) 120–126
broad band in the range of 980–750 cm−1 was detected, which was
assigned to the stretching vibrations of the bridges between tri-
molybdic groups of Keggin structure caused by dehydration process
[31]. The broadening of this region and presence of MoO3 indi-
cated a partial decomposition of HPMo Keggin structure during the
high temperature calcination. These observations are consistent
with earlier literature that HPMo would undergo decomposition
at around 350 ◦ C to form molybdenum and phosphorous oxides
[29,31].
In comparison with the bulk HPAs, a decrease in the intensity
and a broadening of the Keggin bands over supported HPAs is also
observed. This is probably due to the strong HPAs interaction with
Zr OH groups which is an acid-base reaction where the Zr OH
groups act as a base and HPAs as a Brønsted acid. Some NMR could
also support these findings. Oliveira et al. [25] reported in their
studies on ZrO2 supported HPW catalyst two lines present at −15.0
and −13.0 ppm. The first chemical shift is related to the Keggin unit
and the second is assigned to zirconia surface bridges. A similar
shift was also observed by Alsalme et al. [27] that indicated strong
interaction between finely dispersed amorphous HPW species and
ZrO2 .
In summary, it can be concluded that the thermal stability of
HSiW and HPW was stronger than HPMo; the Keggin structure of
HSiW and HPW was well preserved on the support surface, but that
of HPMo was partially destroyed.
3.1.2. Acidic properties of catalysts
Appropriate acidity is essential for the selective hydrogenol-
ysis of glycerol to 1,3-PDO due to the bi-functional mechanism
[9,14,22,34]. The NH3 -TPD technique has been extensively
employed to probe the acidity of solid acid catalysts. Meanwhile,
the FT-IR after pyridine adsorption is a useful tool to determine the
nature and amount of acid sites.
The NH3 -TPD profiles of the as-prepared samples are pre-
sented in Fig. 3. Generally the strength of solid acid sites
with TPD profiles can be classified into three types by the
temperature of desorption NH3 weak (150–300 ◦ C), moderate
(300–450 ◦ C) and strong (450–650 ◦ C) [35]. The NH3 -TPD result
of Pt/ZrO2 showed poorly resolved desorption peaks in two dis-
tinct temperature regions, centering about 255 ◦ C and 437 ◦ C,
respectively. On the other hand, Pt-HPMo/ZrO2 exhibited an over-
lapping peak including one main peak and two shoulders. In
the case of Pt-HPW/ZrO2 , two main peaks and a weak shoulder
were observed, around 222 ◦ C, 479 ◦ C and 157 ◦ C, respectively.
Pt-HPMo/ZrO2
Intensity
Pt-HPW/ZrO2
Pt-HSiW/ZrO2
Pt/ZrO2
100 200 300 400 500
o
Temperature ( C)
Fig. 3. NH3 -TPD profiles of Pt/ZrO2 and Pt-HPAs/ZrO2 catalysts.
123
L B
Absorbance
Pt-HPMo/ZrO2
Pt-HPW/ZrO2
Pt-HSiW/ZrO2
Pt/ZrO2
1400 1450 1500 1550 1600 1650 1700
-1
Wavenumber (cm )
Fig. 4. Py-IR spectra of Pt/ZrO2 and Pt-HPAs/ZrO2 catalysts.
Similar patterns were also obtained for Pt-HSiW/ZrO2 . The max-
imum TPD peak position of supported HPAs catalysts decreased
in the following order: Pt-HPW/ZrO2 (479 ◦ C) > Pt-HSiW/ZrO2
(400 ◦ C) > Pt-HPMo/ZrO2 (337 ◦ C). The results indicates that the acid
strength of HPAs is in the order of HPW > HSiW > HPMo, consistent
well with the previous report [36]. The amount of acid sites can be
calculated from TPD desorption peak area and the results are listed
in Table 1. The total acidities of the as-prepared catalysts were in an
order of Pt-HPMo/ZrO2 > Pt-HPW/ZrO2 > Pt-HSiW/ZrO2 > Pt/ZrO2 .
Fig. 4 illustrates the FTIR spectra after pyridine adsorption on
Pt/ZrO2 and Pt-HPAs/ZrO2 . Pyridine adsorption at Brønsted acid
sites and Lewis acid sites exhibited typical bands centering at
1540–1548 cm−1 and 1445–1460 cm−1 , respectively [13]. The con-
centration of Brønsted acid and Lewis acid are obtained from the
integral intensities of the typical bands and the results are pre-
sented in Table 1. The results show that Pt/ZrO2 possessed only
Lewis acid sites. The loading of HPAs resulted in a noticeable
increase in the Brønsted acid on all the samples except that on
HPMo, in which HPMo partially decomposed to MoO3 . On the
other hand, the amount of Lewis acid sites diminished after HPAs
impregnation except HPMo. The decrease can be related to block-
age of Lewis acid sites due to deposition of HPAs. The increase in
Lewis acid sites of Pt-HPMo/ZrO2 sample may be derived from the
decomposition product MoO3 . It can be seen that the total acidity
determined by Py-IR correlated well with the NH3 -TPD results and
HPAs had significant influence on the relative amount of Brønsted
acid and Lewis acid.
3.2. Effect of supported heteropolyacids on glycerol
hydrogenolysis
The hydrogenolysis of glycerol in aqueous solution was car-
ried out over Pt/ZrO2 and Pt-HPAs/ZrO2 . Fig. 5 illustrates glycerol
conversion and 1,3-PDO selectivity as a function of reaction
temperature. Generally glycerol conversion increased remarkably
whereas 1,3-PDO selectivity decreased with the rising temperature
from 160 to 240 ◦ C over all catalysts. It can be seen that temperature
has a positive effect on glycerol conversion, as claimed by previous
literature [9,10,13]. However, higher temperature disfavored the
formation of 1,3-PDO and increased by-products 1,2-PDO and 1-
600 propanol (1-PO). Gong et al. [9] reported similar phenomenon over
bi-functional Pt/WO3 /TiO2 /SiO2 catalyst during hydrogenolysis of
glycerol in a batch reactor and supposed that elevated temperature
can be favorable to activate terminal hydroxyl groups of glycerol.
124 S. Zhu et al. / Catalysis Today 212 (2013) 120–126

2
100 Lewis acid sites/[µmol/m ]
Pt/ZrO2 90 120 150 180
Pt-HSiW/ZrO2
80 Pt-HPW/ZrO2 20 20

Pt-HPMo/ZrO2
Conversion (%)

60 15 15

1,3-PDO yield/%
1,2-PDO yield/%
40 10 10

20 5 5

0 0 0
160 180 200 220 240
o
Temperature ( C ) 0 20 40 60 80
2
Brønsted acid sites/[µmol/m ]

Fig. 6. Correlation between the amount of acids sites and product yield obtained at
180 ◦ C of the reaction.
50 Pt/ZrO2
Pt-HSiW/ZrO2 In comparison to Pt/ZrO2 , supported HPAs catalysts enhanced the
Pt-HPW/ZrO2 catalytic activity considerably and the most prominent benefit
40
Selectivity of 1,3-PDO (%)

rose from the significant impact on product distribution. For sup-

Pt-HPMo/ZrO2 ported HPAs catalysts, large amounts of sequential hydrogenolysis
30
products (1-PO/2-PO) were formed and the selectivity to degrada-
tion products (EG and ethanol) diminished moderately. Among the
four catalysts, Pt-HSiW/ZrO2 was the most selective catalyst as it
20 showed 48.1% selectivity to 1,3-PDO at 180 ◦ C.
It is well known that the dispersion of active metal and acidity of
catalyst play a key role in the bi-functional hydrogenolysis mech-
10 anism of glycerol [13,14]. It was found that the addition of HPAs
to Pt/ZrO2 slightly decreased the surface area and the dispersion of
Pt. These results inferred that the increased hydrogenolysis ability
0 of Pt-HPAs/ZrO2 was related to the acidity of catalyst. As shown
in Tables 1 and 2, it can be observed that the conversion increased
160 180 200 220 240
o
gradually with the total amount of acid sites over these catalysts. In
Temperature ( C ) lower reaction temperature, Pt-HPMo/ZrO2 exhibited the highest
activity related to high amount of acid sites. However, the catalytic
Fig. 5. Effect of reaction temperature on glycerol conversion and 1,3-PDO selec-
tivity over Pt/ZrO2 and Pt-HPAs/ZrO2 catalysts. Reaction conditions: 5.0 MPa, 2.0 g activity became less than other catalysts at 220 and 240 ◦ C, prob-
catalysts; 10 wt.% glycerol aqueous solution, H2 /glycerol = 137:1 (molar ratio), ably due to the bad hydrothermal stability of HPMo. Varisli et al.
WHSV = 0.09 h−1 . [37] reported that HPMo was less stable than HSiW and HPW at
temperatures higher than 200 ◦ C in dehydration of ethanol.
On the other hand, the increase of reaction temperature promoted It is clearly stated in literature that Brønsted acid sites favors
the excessive hydrogenolysis of propanediols (PDOs) to propanols the removal of the secondary hydroxyl group from glycerol to gen-
(1-PO + 2-PO). erate 1,3-PDO [9,13,24]. As presented in Tables 1 and 2, the yield
The activity and main products selectivity of Pt/ZrO2 and Pt- to 1,3-PDO is proportional to the concentration of Brønsted acid
HPAs/ZrO2 in glycerol hydrogenolysis at 180 ◦ C are summarized sites. A close correlation between the 1,3-PDO yield at 180 ◦ C and
in Table 2. The Pt/ZrO2 catalyst produced lower glycerol conver- the concentration of Brønsted acid sites can be derived, as dis-
sion, a high 1,2-PDO selectivity and very poor 1,3-PDO selectivity, played in Fig. 6. Although Brønsted acid sites were not available in
suggesting that Pt/ZrO2 was not effective for 1,3-PDO formation. the fresh Pt/ZrO2 , they could be generated by interaction of Lewis

Table 2
Catalytic performance of glycerol hydrogenolysis over Pt/ZrO2 and Pt-HPAs/ZrO2 catalysts at 180 ◦ C.

Catalyst Conversion (%) Selectivity (%)a

1,3-PDO 1,2-PDO 1-PO 2-PO EG Ethanol Acetone Propanal Othersb

Pt/ZrO2 18.7 3.9 76.1 3.2 0.6 6.6 3.9 1.2 0.5 4.0
Pt-HSiW/ZrO2 24.1 48.1 16.5 21.8 4.5 0.7 0.9 0.2 1.1 6.2
Pt-HPW/ZrO2 25.5 32.9 10.9 37.9 5.2 0.6 4.2 0.7 5.3 2.3
Pt-HPMo/ZrO2 27.1 7.8 39.2 30.4 3.2 1.0 1.9 1.1 1.4 14.0

Reaction conditions: 180 ◦ C, 5.0 MPa, 2.0 g catalysts; 10 wt.% glycerol aqueous solution, H2 /glycerol = 137:1 (molar ratio), WHSV = 0.09 h−1 .
a
1,3-PDO, 1,3-propanediol; 1,2-PDO, 1,2-propanediol; 1-PO, 1-propanol; 2-PO, 2-propanol; EG, ethylene glycol.
b
Others: acetol, methanol, acetic acid, propionic acid, propane, methane, and so on.
 S. Zhu et al. / Catalysis Today 212 (2013) 120–126 125

Table 3
Reaction results of various reactants over Pt-HSiW/ZrO2 .

Reactant Conversion (%) Selectivity (%)a

1,3-PDO 1,2-PDO 1-PO 2-PO EG Ethanol Acetone Propanal Othersb

Glycerol 24.1 48.1 16.5 21.8 4.5 0.7 0.9 0.2 1.1 6.2
1,2-PDO 47.1 0.0 – 85.2 8.9 0.0 1.8 2.0 0.7 1.4
1,3-PDO 9.1 – 0.0 96.5 1.0 0.0 1.5 0.0 0.6 0.4
EG 53.7 0.0 0.0 0.0 0.0 – 97.7 0.0 0.0 2.3

Reaction conditions: 180 ◦ C, 5.0 MPa, 2.0 g catalysts; 10 wt.% aqueous solution of reactant, WHSV = 0.09 h−1 .
a
1,3-PDO, 1,3-propanediol; 1,2-PDO, 1,2-propanediol; 1-PO, 1-propanol; 2-PO, 2-propanol; EG, ethylene glycol.
b
Others: acetol, methanol, acetic acid, propionic acid, propane, methane, and so on.

acid sites with steam to give rise to 1,3-PDO. Among the tested
supported HPAs catalysts, HSiW led to the highest 1,3-PDO yield
whereas HPMo displayed the lowest 1,3-PDO yield. The excellent
performance of HSiW was attributed to high Brønsted acid sites and
hydrolytic stability even in hydrothermal environment. Atia et al.
[31] reported that supported HSiW was more selective than HPW
and HPMo in glycerol dehydration to acrolein which has the same
intermediate 3-HPA as 1,3-PDO. Tsukuda et al. [36] also acclaimed
that HSiW showed high performance in glycerol dehydration due
to high Brønsted acid strength and resistance to water. On the other
hand, HPMo always showed the lowest 1,3-PDO selectivity whereas
high 1,2-PDO selectivity due to partial decomposition of HPMo,
high Lewis acid sites and bad thermal stability.
According to the literature [8,38], Lewis acid sites are respon-
sible for the glycerol dehydration to acetol which can further
hydrogenation to form 1,2-PDO over metal catalysts. Therefore, we
attempted to correlate the concentration of Lewis acid sites with
1,2-PDO yield at 180 ◦ C. Unfortunately, the correlation is not good,
which could be mainly attributed to sequential hydrogenolysis of
1,2-PDO to propanols. Qin et al. [10] proposed that propanols are
mainly formed via 1,2-PDO in glycerol hydrogenolysis, and it was
also verified by our separate experiment, as discussed later. Thus,
the calculated 1,2-PDO yield should include the observed 1,2-PDO
and the part converted to propanols. It is interesting to find that
a close correlation between the calculated 1,2-PDO yield at 180 ◦ C
and the concentration of Lewis acid sites was obtained, as illus-
trated in Fig. 6.
3.3. Reaction mechanism of glycerol hydrogenolysis over
Pt-HSiW/ZrO2
Activity tests of reaction products 1,2-PDO, 1,3-PDO and
EG over Pt-HSiW/ZrO2 were also examined under similar
conditions as used for glycerol to elucidate the reaction
routes. The reaction results are summarized in Table 3. Regarding
the reaction activity, 1,3-PDO was much lower than 1,2-PDO and
glycerol. Additionally, the selectivity to 1-PO was much higher than
that to 2-PO in the hydrogenolysis of 1,3-PDO or 1,2-PDO, which
was related to the high selectivity of 1-PO formation in the glycerol
hydrogenolysis. The 1-PO/2-PO ratio in the glycerol hydrogenolysis
was similar to that in the 1,2-PDO hydrogenolysis. The combined
results indicate that 1-PO and 2-PO are mainly formed via 1,2-PDO
in the glycerol hydrogenolysis, consistent well with the previous
report [10].
A plausible reaction pathway for glycerol hydrogenolysis and
degradation is presented in Fig. 7. Reaction routes of hydrogenoly-
sis products will be discussed firstly. Previous research [22] shows
that PDOs formation proceeds via dehydration of glycerol to ace-
tol and 3-HPA on acid catalyst and consecutive hydrogenation to
1,2-PDO and 1,3-PDO over metal catalysts, respectively. Alhanash
et al. [8] have suggested that acetol was formed on Lewis acid
sites whereas acrolein on Brønsted acid sites in glycerol dehydra-
tion, which is well in agreement with our results. In the case of
Pt/ZrO2 (Table 2), the main product was 1,2-PDO due to its large
Lewis acid sites. However, addition of HSiW to Pt/ZrO2 catalysts
improved the catalytic activity and 1,3-PDO selectivity remarkably
because of the enhanced Brønsted acid. To corroborate this point
further, we also investigated the Pt-free HSiW/ZrO2 catalyst for
glycerol hydrogenolysis under identical conditions as Pt/ZrO2 at
180 ◦ C. The result showed that glycerol conversion was only 2.3%
and the primary product was acrolein with a selectivity of 66.5%,
whereas the selectivity of acetol was only 13.3%, consistent well
with the conclusion reported by Alhanash et al. [8]. The 3-HPA
was not detected in any experiment whereas acetol was observed,
implying that acetol is much less reactive than 3-HPA. 3-HPA can
dehydrate into acrolein or hydrogenate to 1,3-PDO. Additionally,
3-HPA can further decompose to form acetaldehyde and formalde-
hyde through a retro-aldol condensation reaction. Then acetic

OH + CH4

O O OH
+H2 -H2O +H2
OH OH
OH
-H2O -H2O
+H2 OH
O

-H2O
-H2O [O]
HO OH +H2
HO O HO OH OH
OH
O O
-H2 [O]
-H2 O + O
OH

HO OH HO -CO -H2O
O HO +H2
OH OH
O OH O

Fig. 7. Reaction scheme of glycerol hydrogenolysis and degradation reactions.

126 S. Zhu et al. / Catalysis Today 212 (2013) 120–126
acid can be formed by the oxidation of acetaldehyde in aqueous
solution.
Propanal and acetone are formed through PDOs dehydra-
tion and then hydrogenate to 1-PO and 2-PO, respectively. It is
reported that propanal can be further oxidized to produce pro-
pionic acid in aqueous solution [39]. Ethanol and methane are
produced by C C cleavage of 1,2-PDO. Additionally, trace propane
is formed by sequential hydrogenolysis of propanols. Methane
can be formed from methanol hydrogenation or from PDOs
degradation.
Degradation product EG showed higher hydrogenolysis activity
than glycerol and PDOs over Pt-HSiW/ZrO2 . The fact that EG was
not formed in the reaction of PDOs hydrogenolysis indicated
that it was formed directly from glycerol. The EG formation
proceeds via dehydrogenation of glycerol to glyceraldehyde and
subsequent C C bond cleavage. The mechanism by which the
C C bond cleavage occurs on Pt sites is subject to some debate
that it proceeds via decarbonylation or retro-aldol condensation.
In the presence of base, Maris and Davis [40] have asserted that
EG formation is attributed to retro-aldol reaction rather than
through decarbonylation. King et al. [7] proposed that C C bond
cleavage was formed through glyceraldehyde decarbonylation
under an acidic condition. It seems that both offer plausible
pathways which are highly dependent on the operation conditions
and catalytic system used. Very recently, Vlachos and co-works
[41] have identified the most favorable decomposition pathway
of glycerol on Pt with quantum mechanics calculations and the
pathway is shown as follows: C3 H8 O3 → CHOHCHOHCH2 OH →
CHOHCHOHCHOH → CHOHCOHCHOH → COHCOHCHOH →
COCOHCHOH → CO + COHCHOH. These above results indicate
that EG formation in our reaction chemistry proceeded most
probably via decarbonylation of glyceraldehyde. However, not CO
but CO2 was detected in glycerol hydrogenolysis. Wawrzetz et al.
[42] have observed similar phenomenon during aqueous-phase
hydrogenolysis of glycerol over Pt/Al2 O3 and argued that CO2
is generated by decarbonylation of glyceraldehyde followed by
water gas shift. Additionally, glyceraldehyde is also not detected
because it is very reactive. Interestingly, we observed dihydroxy-
acetone which can be produced from glycerol dehydrogenation or
glyceraldehyde transformation. Liang et al. [6] have disclosed that
glyceraldehyde is more reactive than dihydroxyacetone over Pt
based catalyst. Glyceraldehyde was not detected in our products
most likely because it can undergo decarbonylation or convert to
dihydroxyacetone rapidly. As shown in Table 3, the major product
of EG hydrogenolysis was ethanol and trace of acetaldehyde.
The EG can undergo dehydration to acetaldehyde intermediate
followed by subsequent hydrogenation to ethanol.
4. Conclusions
The results of this study indicate that heteropolyacids mod-
ified Pt/ZrO2 catalysts were effective in glycerol hydrogenolysis
to 1,3-PDO. Addition of HPAs to Pt/ZrO2 catalysts increased the
catalytic activity and 1,3-PDO selectivity considerably due to the
enhanced acidity. Compared to Pt-HPW/ZrO2 and Pt-HPMo/ZrO2 ,
Pt-HSiW/ZrO2 obtained the highest selectivity to 1,3-PDO, which
was probably related to the high Brønsted acid sites and good
thermal stability. The 1,3-PDO yield was proportional to the con-
centration of the Brønsted acid sites, whereas the 1,2-PDO yield
was proportional to the concentration of Lewis acid sites. Finally,
the reaction pathways have been proposed, and the results indi-
cate that 1-PO and 2-PO are mainly formed via 1,2-PDO in glycerol
hydrogenolysis. In addition, the formation of by-product EG is con-
firmed to originate from decarbonylation of glyceraldehyde under
given experimental conditions.
Acknowledgements
This work was financially supported by the Natural Science
Foundation of China (No. 20976185), and the Major State Basic
Research Development Program of China (973 Program) (No.
2012CB215305).
References
[1] A. Corma, S. Iborra, A. Velty, Chemical Reviews 107 (2007) 2411–2502.
[2] G.W. Huber, S. Iborra, A. Corma, Chemical Reviews 106 (2006) 4044–4098.
[3] M. Pagliaro, R. Ciriminna, H. Kimura, M. Rossi, C. Della Pina, Angewandte Chemie
International Edition 46 (2007) 4434–4440.
[4] C.H.C. Zhou, J.N. Beltramini, Y.X. Fan, G.Q.M. Lu, Chemical Society Reviews 37
(2008) 527–549.
[5] A. Behr, J. Eilting, K. Irawadi, J. Leschinski, F. Lindner, Green Chemistry 10 (2008)
13–30.
[6] D. Liang, J. Gao, H. Sun, P. Chen, Z. Hou, X. Zheng, Applied Catalysis B: Environ-
mental 106 (2011) 423–432.
[7] D.L. King, L. Zhang, G. Xia, A.M. Karim, D.J. Heldebrant, X. Wang,
T. Peterson, Y. Wang, Applied Catalysis B: Environmental 99 (2010)
206–213.
[8] A. Alhanash, E.F. Kozhevnikova, I.V. Kozhevnikov, Applied Catalysis A: General
378 (2010) 11–18.
[9] L. Gong, Y. Lu, Y. Ding, R. Lin, J. Li, W. Dong, T. Wang, W. Chen, Applied Catalysis
A: General 390 (2010) 119–126.
[10] L.-Z. Qin, M.-J. Song, C.-L. Chen, Green Chemistry 12 (2010) 1466–1472.
[11] A. Behr, J. Eilting, K. Irawadi, J. Leschinski, F. Lindner, Chimica Oggi 26 (2008)
32–36.
[12] G.A. Kraus, Clean-Soil Air Water 36 (2008) 648–651.
[13] S. Zhu, Y. Zhu, S. Hao, H. Zheng, T. Mo, Y. Li, Green Chemistry 14 (2012)
2607–2616.
[14] M. Balaraju, V. Rekha, P.S.S. Prasad, B.L.A.P. Devi, R.B.N. Prasad, N. Lingaiah,
Applied Catalysis A: General 354 (2009) 82–87.
[15] L. Ma, D.H. He, Catalysis Today 149 (2010) 148–156.
[16] S. Panyad, S. Jongpatiwut, T. Sreethawong, T. Rirksomboon, S. Osuwan, Catalysis
Today 174 (2011) 59–64.
[17] M.G. Musolino, L.A. Scarpino, F. Mauriello, R. Pietropaolo, ChemSusChem 4
(2011) 1143–1150.
[18] Y. Amada, Y. Shinmi, S. Koso, T. Kubota, Y. Nakagawa, K. Tomishige, Applied
Catalysis B: Environmental 105 (2011) 117–127.
[19] Y. Shinmi, S. Koso, T. Kubota, Y. Nakagawa, K. Tomishige, Applied Catalysis B:
Environmental 94 (2009) 318–326.
[20] L. Huang, Y.L. Zhu, H.Y. Zheng, G.Q. Ding, Y.W. Li, Catalysis Letters 131 (2009)
312–320.
[21] T. Kurosaka, H. Maruyama, I. Naribayashi, Y. Sasaki, Catalysis Communications
9 (2008) 1360–1363.
[22] I. Gandarias, P.L. Arias, J. Requies, M.B. Guemez, J.L.G. Fierro, Applied Catalysis
B: Environmental 97 (2010) 248–256.
[23] J. Chaminand, L. Djakovitch, P. Gallezot, P. Marion, C. Pinel, C. Rosier, Green
Chemistry 6 (2004) 359–361.
[24] J. Oh, S. Dash, H. Lee, Green Chemistry 13 (2011) 2004–2007.
[25] C.F. Oliveira, L.M. Dezaneti, F.A.C. Garcia, J.L. de Macedo, J.A. Dias, S.C.L. Dias,
K.S.P. Alvim, Applied Catalysis A: General 372 (2010) 153–161.
[26] O. Toshio, Catalysis Today 73 (2002) 167–176.
[27] A.M. Alsalme, P.V. Wiper, Y.Z. Khimyak, E.F. Kozhevnikova, I.V. Kozhevnikov,
Journal of Catalysis 276 (2010) 181–189.
[28] M. Misono, Catalysis Today 144 (2009) 285–291.
[29] A. Micek-Ilnicka, Journal of Molecular Catalysis A: Chemical 308 (2009) 1–14.
[30] C.A. Emeis, Journal of Catalysis 141 (1993) 347–354.
[31] H. Atia, U. Armbruster, A. Martin, Journal of Catalysis 258 (2008)
71–82.
[32] X. Cui, J. Xu, C. Zhang, Y. Yang, P. Gao, B. Wu, Y. Li, Journal of Catalysis 282 (2011)
35–46.
[33] S.H. Chai, H.P. Wang, Y. Liang, B.Q. Xu, Applied Catalysis A: General 353 (2009)
213–222.
[34] M. Chia, Y.J. Pagán-Torres, D. Hibbitts, Q. Tan, H.N. Pham, A.K. Datye, M. Neurock,
R.J. Davis, J.A. Dumesic, Journal of the American Chemical Society 133 (2011)
12675–12689.
[35] A.S. de Oliveira, S.J.S. Vasconcelos, J.R. de Sousa, F.F. de Sousa, J.M. Filho, A.C.
Oliveira, Chemical Engineering Journal 168 (2011) 765–774.
[36] E. Tsukuda, S. Sato, R. Takahashi, T. Sodesawa, Catalysis Communications 8
(2007) 1349–1353.
[37] D. Varisli, T. Dogu, G. Dogu, Chemical Engineering Science 62 (2007)
5349–5352.
[38] Y.T. Kim, K.-D. Jung, E.D. Park, Applied Catalysis B: Environmental 107 (2011)
177–187.
[39] F. Wang, J.-L. Dubois, W. Ueda, Journal of Catalysis 268 (2009) 260–267.
[40] E.P. Maris, R.J. Davis, Journal of Catalysis 249 (2007) 328–337.
[41] Y. Chen, M. Salciccioli, D.G. Vlachos, Journal of Physical Chemistry C 115 (2011)
18707–18720.
[42] A. Wawrzetz, B. Peng, A. Hrabar, A. Jentys, A.A. Lemonidou, J.A. Lercher, Journal
of Catalysis 269 (2010) 411–420.