Comparative study of WCX-based catalysts for aqueous phase hydrogenolysis of glycerol into bioadditives

Article abstract of DOI:10.1039/c7nj04844k

Hydrogenolysis of glycerol has been a subject of intensive research in recent decades. Great efforts have been made investigating the mechanism of bond cleavage using metal or metallic oxide catalysts. In this work, effects of a series of WC_x-based catalysts, including WC_x, Cu/WC_x, Pt/WC_x, Ru/WC_x, CuRu/WC_x and CuPt/WC_x, were studied on hydrogenolysis of glycerol to investigate the role of each active phase. WO₃-Based catalysts were employed as a comparison with the WC_x-based catalysts. Hydrogenolysis of model compounds (1,2-PDO and EG) of glycerol was also studied using these catalysts, in attempts to determine the mechanism of the whole reaction. The presence of metal elements on WC_x or WO₃ can significantly improve the conversion of reactant. In addition, the ability of WC_x to promote cleavage of C-C was verified. Ru is more beneficial for formation of EG than Pt, whereas Pt is favorable for formation of 1,2-PDO. It had been supposed that hydrogenolysis of glycerol by WC_x-based catalysts occurred via two different but overlapping routes: WC_x-related reactions caused by C-C cleavage and metal-related reactions caused by C-O cleavage, which happen simultaneously in the reaction. The current study not only demonstrates that the catalytic performance of WC_x can be manipulated by varying metal loading, but also offers a possible mechanism for C-C and C-O cleavage in hydrogenolysis of glycerol, which could be beneficial when focusing on the target product.

Full text of DOI:10.1039/c7nj04844k

NJC
View Article Online
View Journal | View Issue
PAPER
Comparative study of WCx-based catalysts for
aqueous phase hydrogenolysis of glycerol into
bioadditives†
Cite this: New J. Chem., 2018,
42, 3633
Lingfei Wei,a Rehana Bibi,a Wei Tian,a Lingyu Chen,a Yu Zheng,a Naixu Li *a and
abc
Jiancheng Zhou
*
Hydrogenolysis of glycerol has been a subject of intensive research in recent decades. Great efforts
have been made investigating the mechanism of bond cleavage using metal or metallic oxide catalysts.
In this work, effects of a series of WCx-based catalysts, including WCx, Cu/WCx, Pt/WCx, Ru/WCx,
CuRu/WCx and CuPt/WCx, were studied on hydrogenolysis of glycerol to investigate the role of each
active phase. WO3-Based catalysts were employed as a comparison with the WCx-based catalysts.
Hydrogenolysis of model compounds (1,2-PDO and EG) of glycerol was also studied using these catalysts,
in attempts to determine the mechanism of the whole reaction. The presence of metal elements on WCx
or WO3 can significantly improve the conversion of reactant. In addition, the ability of WCx to promote
cleavage of C–C was verified. Ru is more beneficial for formation of EG than Pt, whereas Pt is favorable
for formation of 1,2-PDO. It had been supposed that hydrogenolysis of glycerol by WCx-based catalysts
occurred via two different but overlapping routes: WCx-related reactions caused by C–C cleavage
and metal-related reactions caused by C–O cleavage, which happen simultaneously in the reaction. The
current study not only demonstrates that the catalytic performance of WCx can be manipulated by
varying metal loading, but also offers a possible mechanism for C–C and C–O cleavage in hydrogenolysis
of glycerol, which could be beneficial when focusing on the target product.
Received 7th December 2017,
Accepted 24th January 2018
DOI: 10.1039/c7nj04844k
rsc.li/njc
2-PO) and methanol are indispensible in daily life as coatings,
solvents, cosmetics, pharmaceuticals, and pesticides.
Introduction
As a byproduct of the biodiesel industry, a huge amount of
Hydrogenolysis of glycerol involves either a C–C or a C–O
glycerol is produced. The properties of this biomass-derivated cleavage. It is mostly acknowledged that C–C cleavage occurs
substance, such as nontoxicity and sustainability, have driven through a retro-aldol mechanism whereas C–O cleavage occurs
research into treatments such as hydrogenolysis,1–3 dehydration,4–6 by dehydration, which is always affected by base catalysis.17
reforming,7–9 acetalization,10,11 and esterification,12–14 in attempts Control of the PDOs/EG/monohydric alcohol ratio requires high
to transfer glycerol into important bioadditives thus giving them selectivity of C–O or C–C cleavage of glycerol, which is strongly
higher value. Among these, catalytic hydrogenolysis is a promising dependent on the catalyst used.18 During the reaction, products
method for conversion of glycerol into propanediols (PDOs), such as PDOs and EG can be further degraded into substances
namely 1,2-propanediol (1,2-PDO) and 1,3-propanediol (1,3-PDO), with smaller molecular weights (such as POs, methanol etc.), thus
ethylene glycol(EG), and monohydric alcohol (propanols, methanol, reducing the productivity of the polyalcohol. This reaction process
etc.). PDOs and EG are industrially important chemicals used as is shown in Scheme 1.19–21 In this process, glyceraldehyde is an
raw materials in production of polymers, surfactants, func- important intermediate that is unstable and elusive, and can
tional fluids (antifreeze, deicing, dehydrant), resins, cosmetics, subsequently transfer into different products through various path-
and foods.15,16 Furthermore, propanols (POs, including 1-PO and ways (dehydration, hydration etc.). The transition of glycerol into
glyceraldehyde has been considered to be a metal-promoted reac-
a School of Chemistry and Chemical Engineering, Southeast University, Nanjing,
tion, with noble metals such as Pt and Ru showing outstanding
performance with the addition of alkali such as NaOH and CaO.19
EG can be formed from glyceraldehyde through a retro-aldol process
or by C–C cleavage of 1,2-PDO. Ru was found to be a good choice
by Maris et al.19 and Jian Feng et al.,22 as it shows high selectivity
211189, P. R. China. E-mail: jczhou@seu.edu.cn
b Department of Chemical and Pharmaceutical Engineering, Southeast University,
Chengxian College, Nanjing, 210088, P. R. China
c Jiangsu Province Hi-Tech Key Laboratory for Bio-medical Research,
Southeast University, Nanjing 211189, P. R. China
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7nj04844k of EG in hydrogenolysis of glycerol.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018
New J. Chem., 2018, 42, 3633--3641 | 3633

View Article Online
Paper
NJC
Table 1 Physical properties of the catalysts
Catalyst
SBET (m2 gꢀ1
)
Dpore (nm)
Vpore (cm3 gꢀ1
)
WCx
34.7
28.0
21.6
25.7
26.7
19.6
153.7
40.2
2.804
5.185
9.771
4.455
8.024
6.025
3.250
5.865
0.024
0.036
0.053
0.029
0.053
0.030
0.124
0.059
Cu/WCx
Pt/WO3
Pt/WCx
Ru/WO3
Ru/WCx
CuPt/WCx
CuRu/WCx
The BET surface area, pore diameter, and pore volume of WCx
are 34.7 m2 gꢀ1, 2.804 nm, and 0.024 cm3 gꢀ1, respectively. For
Cu/WCx, Pt/WCx, and Ru/WCx, the BET surface areas are much
lower than that of WCx, whereas the pore diameters and pore
volumes are larger than those of WCx. This may related to the
pore size distribution.28 For CuPt/WCx and CuRu/WCx, the BET
surface area, pore diameter, and pore volume are both larger
than those of WCx, and, notably, CuPt/WCx possesses the largest
BET surface area and pore volume, which can be explained by its
apparent porous structure and the plicated surface as shown in
the SEM images in Fig. S1 (ESI†). Comparing the physical
properties of WCx-based and WO3-based catalysts, it is found
that WCx-based catalysts have larger BET surface but lower pore
diameter and pore volume than WO3-based catalysts. Composi-
tion of each element was measured by ICP-OES (Table S1, ESI†).
Differences between theoretical value and effective value may
result from experimental bias.
Structure properties of the catalysts. Fig. 1 shows XRD patterns
of the as-synthesized WCx under different calcination tempera-
tures (800, 1000, 1200, 1400 1C). Apparently, no crystal forms exist
under 800 1C. When the calcination temperature was changed to
1000 1C, there were obvious peaks at around 31.61, 35.61, 38.11,
39.61, 48.31, 61.61, and 73.21, consistent with the JCPDS cards
No. 89-2727 (pure hexagonal WC) and No. 89-2371 (pure ortho-
rhombic W2C). As no peaks for other impurities are found in
the pattern, it can be confirmed that the catalyst is a composite
of WC and W2C. Diffraction peaks at around 31.61, 35.61, 48.31,
and 73.21 are assigned to the (001), (100), (101), and (111) faces
of pure hexagonal WC, respectively. Diffraction peaks at around
38.11, 39.61, and 61.61 are assigned to the (200), (121), and (040)
faces of pure orthorhombic W2C, respectively. Moreover, typical
Scheme 1 Mechanism and process of C–C and C–O cleavage in glycerol
hydrogenolysis. a A, B, C in the scheme shows three main parts in this
reaction, which is corresponding to Scheme 2 and will be illustrated in
detail hereinafter.
In recent years, transition metal carbides such as tungsten
carbides (WCx, x = 0.5, 1) have been identified to have excellent
performance in selective C–C and C–O breakage. Zhang Tao’s
group have done much research on hydrogenation of cellulose
with tungsten carbides.23–25 Ren’s team investigated tungsten
carbides as selective deoxygenation catalysts, using DFT calcu-
lations, surface science experiments, and reactor evaluations.
They proved the high activity of the WC(0001) surface towards
selective C–O/CQO bond breakage.26 Bifunctional nickel-
containing tungsten carbides have also been found to be very
effective for conversion of glucose. High selectivity is derived
from the fed-batch system to maintain a certain concentration
of sugar solutions, and excellent behavior of metallic nickel
component and tungsten component in retro-aldol and hydro-
genation processes, respectively.27
Accordingly, to better understand how catalysts influence
product distribution, investigation is required on the role of
each active phase in different reaction parts (part A, B, C).
Because of the excellent performance of noble metals Pt and
Ru, and non-noble metal Cu in glycerol hydrogenolysis, and the
effective bond-breaking ability of WCx and WO3, this study
looks at different loadings (Pt, Ru, Cu or their bimetal) on WCx
or WO3, synthesized by a sol–gel or hydrothermal method, and
use of these products in hydrogenolysis of glycerol to explore
the role of each active phase. Hydrogenolysis of the model
compounds of glycerol (EG and 1,2-PDO) is also discussed to
better understand each reaction part. Moreover, on the basis of
a previously defined mechanism provided, the effect of each
catalyst in the reaction is discussed and a new mechanism
based on WCx is proposed. It should be noted that some side
reactions occurred under high temperature, such as aqueous
phase reforming and hydrothermal reaction, and these are not
considered in the whole study.
Results and discussion
Characterization of the catalysts
Physical properties of the catalysts. The physical properties
Fig. 1 XRD patterns of WCx under different temperatures.
of WCx-based and WO3-based catalysts are shown in Table 1.
3634 | New J. Chem., 2018, 42, 3633--3641
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018

View Article Online
NJC
Paper
diffraction peaks of pure orthorhombic W2C at 34.51, 52.41, load of Cu is favorable for growing crystalline grains and
and 69.81 all missing under each condition, indicate that the yielding a better degree of crystallinity. Comparing the patterns
high energy surfaces of W2C phase fall away gradually as of Cu/WCx, CuRu/WCx, and CuPt/WCx, Ru-loaded catalyst
the temperature rises, leaving only some low-energy surfaces. shows a relatively weaker peak intensity at 38.11 assigned to
The sharpest peaks present at 35.61 and 48.31 of each sample the (200) face of pure orthorhombic W2C, indicating that Ru
indicate that WC grains synthesized by the sol–gel method under can impair growth of the (200) face of W2C. There are some
such conditions show (100) and (101) preferred orientations. The peaks belonging to carbon shown in the patterns of Cu/WCx,
intensity of each peak of the WC phase becomes stronger at CuRu/WCx, and CuPt/WCx, illustrating that Cu leads to carbon
1200 1C than at 1000 1C, indicating growing of crystalline grains deposition in the calcination process.
and a better degree of crystallinity. However, at further increased
The morphologies of the prepared WCx were characterized
temperature (1400 1C), peak intensity becomes weak again. This by SEM and TEM (Fig. 3). As shown in Fig. 3a, WCx prepared by
may because of transition of the crystalline phase from WC to the sol–gel method accumulate in clumps and exhibit a scaly-
W2C at such high temperatures, speculated from the sharpening like surface. Fig. 3b shows a more detailed picture, with orderly
of the peak at 38.11 assigned to the (200) face of pure orthor- wrinkles on the surface. Such a blocky structure is stacked up
hombic W2C. A detailed study of the predominant 301–501 with lattice fringes of 0.284 nm, corresponding to the (001)
reflections showed that rising temperature resulted in a slight plane of WC and indicating that WCx mostly consists of WC in
shift of WC diffraction peaks toward higher 2y values. This this area. Fig. 3c and d give the TEM image and SAED pattern of
behavior is indicative of structure defects (lattice contraction) another area of the synthetic WCx. Wrinkles can also be seen in
on the surface of WC, caused by the high temperature. This can Fig. 3c, and electron diffraction (Fig. 3d) confirms the presence
be corroborated by calculation of the grain size using the Debye– of the W2C phase in this area as the single crystal form. This
Scherrer equation.29 The grain size of WC changed from 16.4 Å result that both WC phase and W2C phase exist in the synthesized
to 20.7 Å, and 25.7 Å with rising calcination temperature (1000– catalyst is consistent with results from the XRD experiment. Fig. 4
1400 1C). To obtain a relatively small crystal grain and take gives TEM images of WCx-based catalysts. From Fig. 4a and b,
energy conservation into consideration, a temperature of 1200 1C Cu particles are distributed on the surface of WCx and have an
was chosen to calcine catalysts in the following steps.
average size of 17.22 nm. The lattice fringes exhibited in Fig. 4b
Fig. 2 shows XRD patterns of metal-loaded WCx catalysts are about 0.130 nm, corresponding to the (220) plane of Cu.
synthesized at 1200 1C. Clearly, when Cu, Ru, CuRu, or CuPt were Similarly, Fig. 4c–f show the particle distribution of Pt and Ru
loaded onto WCx, WCx remained a composite of WC and W2C, on WCx. EDS spectra also give a rough element content, which
whereas from the Pt-loaded WCx, showing no diffraction peaks of is approximate to the theoretical value, implying that the metal
W2C, only the pure WC phase was left. Interestingly, peak has been successfully loaded onto the WCx. TEM images of
intensity of WC and W2C in the WCx-based catalysts showed the CuPt/WCx and CuRu/WCx are given in Fig. 4g and h. It is
following trend: Cu/WCx 4 CuRu/WCx, CuPt/WCx 4 Ru/WCx 4 interesting that no wrinkles on the surface of WCx are seen,
Pt/WCx, indicating that the load of noble metals can significantly which may related to the loaded metal. The element content
reduce crystallinity and degree of order of WC or W2C, while the also proved to be approximate to the theoretical value, detected
by the EDS spectra. Further confirmation of the existence
of each element in CuPt/WCx and CuRu/WCx was characterized
Fig. 3 (a) SEM image of WCx, (b and c) TEM images of WCx. The inset in
(b) shows the corresponding HRTEM image to (d) the SAED patterns from
Fig. 2 XRD patterns of WCx-based catalysts (calcined at 1200 1C).
the region of WCx (c).
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018
New J. Chem., 2018, 42, 3633--3641 | 3635

View Article Online
Paper
NJC
Fig. 5 (a) XPS survey spectra of sample CuRu/WCx and CuPt/WCx; (b–f)
high resolution XPS spectra of O, W, Cu, Ru, and Pt.
attributed to Ru(0). The Ru 3d3/2 and 3d5/2 peaks appear
separated by approximately 4.2 eV, consistent with the litera-
ture records.34 Fig. 5f shows the Pt 4f peaks of CuPt/WCx. Two
peaks at around 72.1 and 74.2 eV correspond to Pt(0) and Pt(IV),
respectively,35,36 which illustrates that the metallic Pt is not the
only species to exist in the catalyst.
Fig. 4 TEM images of (a and b)a Cu/WCx, (c and d) Pt/WCx, (e and f) Ru/WCx,
(g)a CuPt/WCx, and (h)a CuRu/WCx. Insets are the corresponding EDS spectra
and HRTEM image. a The EDS spectra of Cu/WCx was ignored because of the
use of coated copper grids.
Catalytic performance of the catalysts
by STEM-EDX mapping (Fig. S2 and S3, ESI†). Highly dispersed The carbon mass balance was calculated (Table S2, ESI†), and
Cu, Ru, and Pt were observed, and the existence of O could be the conversion calculated by glycerol weight (con. by weight%).
attributed to adventitious oxygen from the air.
The difference between the two conversions is tentatively regarded
The surface chemical compositions and electronic structure as the amount of carbon-containing gases produced from glycerol
of WCx-based catalysts were investigated by X-ray photoelectron hydrogenolysis. Results show that about 60–80% of the total
spectroscopy (XPS) (Fig. 5). Fig. 5a provides a comparison of carbon put into the reactor remains in the liquid product, demon-
survey spectra of CuRu/WCx and CuPt/WCx. Both samples show strating that the cellulose was predominantly converted into liquid
binding energies originating from O 1s, C 1s, W 4f, and Cu 2p. products. Table 2 summarizes the results of hydrogenolysis of
The W 4f spectra for the catalysts (Fig. 5c) are characteristic glycerol/EG/1,2-PDO under WCx-based and WO3-based catalysts
of WC and W2C phases. Peaks at binding energies of 31.2 and (WCx, Cu/WCx, Pt/WO3, Pt/WCx, Ru/WO3, Ru/WCx, CuPt/WCx, and
29.0 eV belong to the WC phase, corresponding to W 4f5/2 CuRu/WCx), including the conversion of reactant and selectivity
and W 4f7/2, respectively. Peaks at binding energies of 32.9 and of each product. As shown by runs #1, 4, 7, 10, 13, 16, 19, and
35.0 eV belong to the W2C phase, corresponding to W 4f5/2 and 22 in Table 2, in the reaction of hydrogenolysis of glycerol,
W 4f7/2, respectively.30, 31 This result confirms the fact that both tungsten-based catalysts showed good performance in convert-
WC and W2C phases exist in the catalyst, consistent with the ing glycerol into dihydric alcohol or monohydric alcohol. In
XRD and TEM results. Two peaks in Fig. 5d at binding energies each reaction, conversion of glycerol was no less than 64.9%
around 931.8 and 954.7 eV could be assigned to Cu 2p3/2 and and reached up to 84.5% under the catalysis of CuRu/WCx,
Cu 2p1/2, respectively, corresponding to the metallic Cu (Cu(0)) suggesting that CuRu/WCx may be the preferred choice to
species.32, 33 The Ru 3d and Pt 4f spectra for the CuRu/WCx and convert glycerol. However, it is more complicated when the
CuPt/WCx catalysts are shown in Fig. 5e and f, respectively. In selectivity of each product is considered. It can be seen in
the XPS of Ru systems, the referencing of the binding energy run #1 that when WCx is taken as the single active phase, the
scale is complicated by the superposition of the C 1s and the main products of hydrogenolysis of glycerol are methanol and
Ru 3d signals. Fig. 5e shows a Ru 3d5/2 signal at 280.3 eV, EG, occupying selectivity of 19.6% and 25.2%, respectively,
3636 | New J. Chem., 2018, 42, 3633--3641
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018

View Article Online
NJC
Paper
Table 2 Summary of reactant conversion and product selectivity during hydrogenolysis of glycerol/EG/1,2-PDOa,b,c,d
Selectivity (%)
Run#
Catalyst
Reactant
Conversion (%)
Methanol
1-PO
2-PO
1,2-PDO
EG
1,3-PDO
1
2
3
4
5
6
7
8
WCx
WCx
WCx
Cu/WCx
Cu/WCx
Cu/WCx
Pt/WO3
Pt/WO3
Pt/WO3
Pt/WCx
Glycerol
EG
1,2-PDO
Glycerol
EG
1,2-PDO
Glycerol
EG
1,2-PDO
Glycerol
EG
1,2-PDO
Glycerol
EG
1,2-PDO
Glycerol
EG
1,2-PDO
Glycerol
EG
1,2-PDO
Glycerol
EG
64.9
43.7
41.3
72.0
60.2
54.3
74.4
83.0
57.8
79.2
80.7
57.8
73.5
67.4
48.9
84.3
30.3
66.2
80.2
84.2
48.9
84.5
86.5
68.4
19.6
16.7
7.0
13.3
19.3
12.2
3.1
2.9
—
1.2
1.1
—
29.5
3.0
—
22.3
4.7
—
30.2
0.2
—
14.7
2.5
—
28.1
3.8
—
32.8
13.3
—
3.0
—
2.8
3.8
—
11.1
4.6
—
19.8
0.9
—
8.0
4.9
—
18.7
2.2
—
4.2
5.3
—
10.9
9.2
—
10.3
—
—
10.1
—
—
30.8
—
—
39.3
—
—
22.9
—
—
10.5
—
—
49.7
—
—
45.6
—
—
25.2
—
24.0
19.6
—
15.4
8.6
—
12.5
19.7
—
13.9
20.0
—
21.9
61.6
—
33.2
21.7
—
11.4
—
—
37.2
—
—
8.1
—
—
19.9
—
—
5.4
—
—
3.3
—
—
13.2
—
—
8.4
—
—
9.8
9
18.9
12.3
19.3
18.3
0.3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Pt/WCx
Pt/WCx
Ru/WO3
Ru/WO3
Ru/WO3
Ru/WCx
Ru/WCx
Ru/WCx
CuPt/WCx
CuPt/WCx
CuPt/WCx
CuRu/WCx
CuRu/WCx
CuRu/WCx
4.9
22.3
10.7
24.5
10.7
3.7
10.6
17.7
0.5
17.2
19.8
—
14.4
1.5
1,2-PDO
44.4
11.9
24.9
a
b
Each value was obtained after 8 h reaction with the same H2 pressure (4 MPa) and catalyst amount (0.15 g). Each reaction was performed three
c
d
times and values were obtained as the mean value. Gas products were released after the reaction, but ignored in calculation. Data in the table
include only the six main products, therefore numbers not totalling 100% can be explained by several unnameable products in trace amounts,
which were ignored intentionally.
indicating WCx itself has good C–C cleavage ability. Compare Compared with run #1, a relatively high selectivity of 1,3-PDO
run #1 with runs #4, 10, and 16, and a remarkable enhance- (37.2%) was obtained in run #4, indicating that when Cu
ment in glycerol conversion is found, illustrating that metal participates in the form of a single loaded metal (Cu/WCx), it
loading has a positive promotion effect on glycerol conversion can significantly improve the ability of glycerol transfer into
and that Ru/WCx appears to be more active than Pt/WCx and 1,3-PDO. In runs #10 and 19, runs #16 and 22, in which Cu and
Cu/WCx for hydrogenolysis of glycerol. This result is consistent a noble metal (Pt and Ru) were added as a bi-metal loaded onto
with other studies comparing the activity of Ru- and Pt-based WCx, selectivity of 1,2-PDO all increased more than 10%.
catalysts for hydrogenolysis of glycerol.19,37–39 By analyzing data Furthermore, selectivity of EG showed a sharp decline when
on selectivity, Ru/WCx exhibited a better tendency to transfer CuRu/WCx was used in the reaction compared with Ru/WCx.
glycerol into EG than Pt/WCx and Cu/WCx, while Pt/WCx and Clearly, a favorable C–O cleavage pathway exists on Cu. In
Cu/WCx revealed significant 1,2-PDO and 1,3-PDO formation. addition, comparing run #4 with run #19 and 22, participation
Selectivity of methanol also occupied an appreciable portion when of Pt or Ru, respectively, in the catalyst system caused higher
Cu/WCx, Pt/WCx, or Ru/WCx were used as the catalyst, which may selectivity of 1,2-PDO, and decline of selectivity of methanol
because of the ability of transition metal carbides to cleave C–C and 1,3-PDO at the same time. This could be because of
bonds. Comparing runs #7 and 13 in Table 2, having almost the subsequent hydrogenolysis of EG and 1,2-PDO under the
same glycerol conversion, Pt/WO3 promoted 30.76% selectivity to different catalysts, and will be discussed in the following
1,2-PDO with only 8.6% selectivity to EG, while Ru/WO3 promoted text. Results from use of WCx-based catalysts and WO3-based
22.9% selectivity to 1,2-PDO with 20.0% selectivity to EG. Runs catalysts were also considered. From runs #7 and 10, conver-
#10 and 16 showed that selectivity of 1,2-PDO and EG was 39.3% sion of glycerol is much higher with Pt/WCx, and selectivity of
and 19.7%, respectively, with catalysis of Pt/WCx, 10.5% and methanol and EG were also higher, compared with the Pt/WO3
61.6%, respectively, with catalysis of Ru/WCx. The obvious differ- catalyzed reaction. Similar results were also obtained from runs
ences also imply that Ru is more beneficial for formation of EG #13 and 16, confirming that WCx has much better ability to
than Pt, whereas Pt is favorable for formation of 1,2-PDO. When transfer glycerol and promote cleavage of C–C over C–O.
comparing runs #19 and 22, selectivity of 1,2-PDO remained high
Using the same method, the catalytic influence on hydro-
both with CuPt/WCx (49.7%) or CuRu/WCx (45.6%), but much of genolysis of EG and 1,2-PDO was studied. Results show that
the EG and 1,2-PDO were transferred into POs with catalysis of (i) metal-loaded WCx are preferable to enhance conversion of
CuRu/WCx, compared with CuPt/WCx.
EG/1,2-PDO. Among these, the highest conversions were obtained
The catalytic activity of Cu in hydrogenolysis of glycerol was by CuRu/WCx, 86.5% for hydrogenolysis of EG and 68.4% for
studied in runs #1 and 4, runs #10 and 19, runs #16 and 22 (Table 2). hydrogenolysis of 1,2-PDO; (ii) Cu/WCx, Pt/WCx, and Ru/WCx
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018
New J. Chem., 2018, 42, 3633--3641 | 3637

View Article Online
Paper
NJC
Fig. 6 (a) Selectivity of 1,2-PDO or 1-PO with different catalysts, (b) selectivity
of 1,2-PDO or 2-PO with different catalysts. A, B and C refer to selectivity
of 1,2-PDO in glycerol hydrogenolysis, selectivity of 1-PO (a) or 2-PO
(b) in 1,2-PDO hydrogenolysis and selectivity of 1-PO (a) or 2-PO (b) in glycerol
hydrogenolysis, respectively.
Fig. 7 Recycling of WCx-based catalysts in hydrogenolysis of glycerol.
crystallinity. Recycling stability was also examined (Fig. 7).
Although the general tendency of glycerol conversion with each
catalyst is decline within five runs, the drop rate is quite slight
compared with the highest value. Among the five catalysts, Cu/WCx
seems to have the poorest reusability (declined about 7.3%), whereas
Pt/WCx and CuPt/WCx possess the best reusability (declined
about 3.1% and 3.9%, respectively). It was considered that
deactivation of catalysts results from loss of the W element in
the reused catalysts,27 as a decline of W content was detected by
ICP-OES (Fig. S6, ESI†). Results also show that the amount of
loaded metal decreased after each run. It was assumed that
loaded metal elements can drop out easily from the original
support and fall into solution during reaction, which may be
another reason for deactivation of catalysts.
exhibited higher selectivity of methanol in hydrogenolysis of
EG compared with other catalysts, which may because of the
good C–C cleavage ability of WCx. CuPt/WCx and CuRu/WCx did
not show such tendencies, perhaps because of the large area
coverage of metals on the surface of WCx, which can decrease
the active sites of WCx; (iii) the main product of hydrogenolysis
of 1,2-PDO is 1-PO, metal-loaded catalysts can significantly
improve the selectivity of 1-PO. Among these, CuRu/WCx
showed the most inclination to transfer 1,2-PDO into 1-PO.
As discussed, WCx specializes in C–C cleavage, thus showing
high selectivity of methanol both in EG and 1,2-PDO hydro-
genolysis. In this case, selectivity of methanol in glycerol
hydrogenolysis exhibits a relatively high value. The same ten-
dency is found when referring to selectivity of 1-PO or 2-PO. As
shown in Fig. 6, high selectivity of 1,2-PDO in glycerol hydro-
genolysis and high selectivity of 1-PO (in Fig. 6a) or 2-PO (in
Fig. 6b) in 1,2-PDO hydrogenolysis caused high selectivity of 1-
PO (in Fig. 6a) or 2-PO (in Fig. 6b) in glycerol hydrogenolysis.
On the contrary, when Ru/WCx was used as the catalyst, 33.2%
selectivity of EG was obtained in hydrogenolysis of 1,2-PDO,
and with a relatively high conversion of 1,2-PDO (66.2%), a
considerable amount of the product 1,2-PDO was subsequently
transferred into EG in glycerol hydrogenolysis with Ru/WCx.
Meanwhile, only 30.3% of EG was transferred into methanol,
thus the selectivity of EG can reach up to 61.6% in hydrogeno-
lysis of glycerol while selectivity of 1,2-PDO reaches only 10.5%.
This trend is found with other catalysts (Fig. 6), highlighting
the importance of consideration of catalytic effects on side
reactions in order to get a certain product.
Discussion
From the results, it can be concluded that metal and WCx both
play efficient roles in the reactions. The metal and WCx may have
distinct routes to interact with the reactant, and a suggested
mechanism is shown in Scheme 2. Hydrogenolysis of glycerol
can be divided into three parts: A, B, and C, which represent
reactions in glycerol, 1,2-PDO, and EG, respectively. Each part
includes WCx-related reactions, i.e. C–C cleavage (route A), and
metal-related reactions, i.e. C–O cleavage (route B).
Active hydrogen species can be produced under high tempera-
ture with a metal catalyst, plenty of free H+ filled in the reaction
system and bonded with –OH that link to the terminal carbon or
the center carbon in glycerol (called C1 in this article). In route A
processes, the increase of electron cloud density of C1 caused
another shift of electron cloud between two adjacent carbon
atoms (called C2 or C3 in this study, shift of electron cloud in
Recycling stability
After each reaction, catalysts were recollected and washed with the way of C1–C2–C3 or C1–C2). In this process, a carbon with
deionized water several times, followed by drying at 80 1C for positive electricity must be produced and this would be attracted
12 h. X-ray diffraction was used to characterize the structure of to the carbon with negative electricity in the WCx-based catalyst
catalysts after the reaction (Fig. S5, ESI†). It is clear that all exposed on the catalyst surface. Because of the strong attraction
catalysts after reaction retain a fine lattice structure, including between the carbon atom and tungsten atoms in the catalyst
diffraction peaks both of WC and W2C, along with several small system, cleavage of C–C can only occur in glycerol thus forming a
peaks of graphite carbon. Peak intensity of Pt/WCx and Ru/WCx dihydric alcohol (EG) and a monohydric alcohol (methanol).
both increased compared with intensity before the reaction, Similarly, a route B process occurs simultaneously, indicating a
indicating growth of crystalline grains and a better degree of C–O cleavage process through dehydration, and forming 1,2-PDO
3638 | New J. Chem., 2018, 42, 3633--3641
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018

View Article Online
NJC
Paper
Scheme 2 Possible mechanism for hydrogenolysis of glycerol with WCx-based catalysts. a Parts A, B, C represent three different reactions with reactants
of glycerol, 1,2-PDO, and EG, respectively. b Route A represents WCx-related reactions caused by C–C cleavage and route B represents metal-related
reactions caused by C–O cleavage. c Green dots in the scheme represent the loaded metals on the surface of WCx.
39, 40
and 1,3-PDO.37,
Part B shows a reaction of 1,2-PDO, to methanol) was catalyzed mainly by WCx through route A,
including route A and route B, and forms products with small thus WO3-based catalysts did not show such tendencies. How-
molecules such as 1-PO, 2-PO, EG, ethanol, and methanol. ever, the reasons CuPt/WCx and CuRu/WCx did not show such
Similarly, part C shows a subsequent reaction of EG, which tendencies remain unclear. This may be because of competition
produced methanol as its main product.
between WCx-related reactions and metal-related reactions.
As discussed, Pt/WCx and Ru/WCx showed good performance Further factors in the two routes are under study.
in hydrogenolysis of glycerol into EG (line A in Fig. 8). Cu/WCx,
Pt/WCx, and Ru/WCx all exhibited good performance in hydro-
genolysis of EG into methanol (line B in Fig. 8). Correspondingly,
selectivity of methanol also showed a relatively high value in
Experimental
Chemicals and materials
hydrogenolysis of glycerol under Cu/WCx, Pt/WCx, and Ru/WCx
(line C in Fig. 8). This process (from glycerol to EG and then
Ammonium tungstate hydrate (H40N10O41W12ꢁH2O, 99.0%), citric
acid monohydrate (C6H8O7ꢁH2O, 99.8%), sucrose (C12H22O11,),
ammonium platinic chloride ((NH4)2PtCl6, 99.8%), ruodium
chloride hydrate (RuCl3ꢁxH2O, 99.6%), copper nitrate trihydrate
(Cu(NO)3ꢁ3H2O, 99.8%), and sodium borohydride (NaBH4, 99.8%)
were all purchased from the Sinopharm Chemical Reagent
Company in analytical grade, and used as received without
further purification. Deionized water, with a resistivity of
18.2 MO cm, was used throughout the whole experiment.
Catalyst preparation
Synthesis of WCx. WCx was prepared by a simple sol–gel
method, in two parts: preparation of gel precursor and reduction
carbonization under high temperature. Ammonium tungstate
hydrate (3.0 g, 1 mmol) was dissolved in 50 ml deionized water
and stirred for 10 min, followed by addition of a certain amount
of citric acid monohydrate under continuous stirring until it
Fig. 8 Selectivity of EG in hydrogenolysis of glycerol (A), selectivity of
became transparent and homogeneous. This solution was than
blended with a sucrose aqueous solution to provide carbon in
methanol in hydrogenolysis of EG (B), and selectivity of methanol in
hydrogenolysis of glycerol (C).
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018
New J. Chem., 2018, 42, 3633--3641 | 3639

View Article Online
Paper
NJC
the reduction carbonization process and mixed uniformly (ON-WAX, 30 m ꢂ 0.32 mm) and a flame ionization detector.
using an ultrasonic machine. The obtained solution was then Similarly, products were also identified by GC-MS (Shimadzu,
evaporated off the solvent at 80 1C and dried at 140 1C overnight GC-MS QP 2010 Plus). The conversion of glycerol/EG/1,2-PDO
to get an expanded crumbly solid, which was subsequently and selectivity of each product were calculated based on the
ground into a fine powder that was called the gel precursor. following equations:
After that, the gel precursor underwent a reduction carboniza-
tion process, that was calcined under a specific temperature
Conversion ð%Þ
(800, 1000, 1200, 1400 1C) in a temperature programmed tube
moles of reactant ðinÞ ꢀ moles of reactant ðoutÞ
¼
(1)
furnace for 3 h under an Ar and Ar/H2 atmosphere. After
cooling to room temperature, the solid sample was collected
and stored in a desiccator for further characterization and use.
Synthesis of WCx-based catalysts. Similar processes to those
used for WCx were used to synthesize monometallic/bimetallic-
loaded WCx, except a specific metallic salt solution or mixture
was added and reduced by a NaBH4 solution before blending
with the sucrose aqueous solution. All WCx-based catalysts were
1% mass content of each metal element. WO3-Based catalysts
were prepared by a hydrothermal method according to a previous
report,41,42 which is not described here.
moles of reactant ðinÞ
ꢂ 100%
moles of one product
Selectivity ð%Þ ¼
ꢂ 100% (2)
moles of all liquid products
Conclusions
In this work, a series of WCx-based catalysts were prepared and
applied to hydrogenolysis of glycerol, as well as its model
compounds 1,2-PDO and EG. The presence of metal elements
on WCx or WO3 was proven to significantly improve the con-
version of reactant. Ru is more beneficial for formation of EG,
whereas Pt is favorable for formation of 1,2-PDO. WCx was also
verified to possess good ability for C–C cleavage. WCx-Related
reactions caused by C–C cleavage and metal-related reactions
caused by C–O cleavage occurred simultaneously in this process.
Catalyst characterization
Surface area (SBET) was calculated using the Brunauer, Emmett,
and Teller (BET) method. Pore size distribution and average
pore diameter were calculated using the desorption branch of the
isotherm according to the Barrett–Joyner–Halenda (BJH) method.
The chemical composition of the samples was analyzed using
an inductively coupled plasma optical emission spectrometer
(ICP-OES, Varian ICP-OES 720). The structure and morphology
of the samples were characterized by X-ray diffraction (XRD,
Bruker D8-Discover with Cu-Ka radiation, l = 1.5478 Å), scanning
electron microscopy (SEM, JSM-7800F), and transmission electron
microscopy (TEM, Hitachi H-600, FEI Tecnai G2 F30). X-ray
photoelectron spectroscopy (XPS) was measured on a 2000 XPS
system with a monochromatic Al-Ka source and a multichannel
detector to determine the surface chemical states.
Conflicts of interest
There are no conflicts to declare.
Acknowledgements
Activity test
This work was financially supported by the Fundamental
Research Funds for the Central Universities of China (No.
3207045403, 3207045409, 3207046414), National Natural Science
Foundation of China (No. 21576050 and 51602052), Jiangsu
Provincial Natural Science Foundation of China (BK20150604),
and a Project Funded by the Priority Academic Program Develop-
ment of Jiangsu Higher Education Institutions (PAPD).
Hydrogenolysis of glycerol and its model compounds (EG and
1,2-PDO) were performed in a stainless-steel micro-autoclave
(YZ-MR-100ML) reactor. Typically, 40 ml of 20 wt% glycerol/EG/
1,2-PDO aqueous solution was prepared in a quartz cup and
0.15 g catalyst was added to the solution, which was sub-
sequently mixed homogeneously using a ultrasonic machine
for 10 min. After that, the quartz cup was put into a stainless-
steel micro-autoclave reactor, the reactor was sealed and
flushed three times with N2. Then 4 MPa of H2 was injected
into the reaction vessel and heated to 220 1C by increasing the
temperature gradually, then maintained for 8 h under magnetic
stirring at 800 rpm. After the reaction, the autoclave was cooled
to room temperature naturally, followed by the release of gas
and collection of the liquid product. The liquid product was
centrifuged at 8000 rpm and filtrated using the 0.22 mm PES
filter membrane. Solid catalyst was collected for further char-
acterization and recycling tests. The pure liquid product was
injected into a gas chromatographer (GC, Shanghai Qiyang Info
Technology Co., Ltd, China) equipped with a capillary column
Notes and references
1 Z. Yuan, L. Wang, J. Wang, S. Xia, P. Chen, Z. Hou and
X. Zheng, Appl. Catal., B, 2011, 101, 431–440.
2 M. L. Barbelli, G. F. Santori and N. N. Nichio, Bioresour.
Technol., 2012, 111, 500–503.
3 S. Zhu, Y. Zhu, S. Hao, H. Zheng, T. Mo and Y. Li, Green
Chem., 2012, 14, 2607–2616.
4 C. W. Chiu, M. A. Dasari, G. J. Suppes and W. R. Sutterlin,
AIChE J., 2006, 52, 3543–3548.
5 W. Suprun, M. Lutecki, T. Haber and H. Papp, J. Mol.
Catal. A: Chem., 2009, 309, 71–78.
3640 | New J. Chem., 2018, 42, 3633--3641
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018

View Article Online
NJC
Paper
6 H. Atia, U. Armbruster and A. Martin, J. Catal., 2008, 258, 25 S. Van de Vyver, J. Geboers, P. A. Jacobs and B. F. Sels,
71–82. ChemCatChem, 2011, 3, 82–94.
7 P. D. Vaidya and A. E. Rodrigues, Chem. Eng. Technol., 2009, 26 H. Ren, Y. Chen, Y. Huang, W. Deng, D. G. Vlachos and
32, 1463–1469. J. G. Chen, Green Chem., 2014, 16, 761–769.
8 E. L. Kunkes, R. R. Soares, D. A. Simonetti and J. A. Dumesic, 27 R. Ooms, M. Dusselier, J. A. Geboers, B. O. de Beeck,
Appl. Catal., B, 2009, 90, 693–698.
R. Verhaeven, E. Gobechiya and B. F. Sels, Green Chem.,
2014, 16, 695–707.
28 C. Leyva, M. S. Rana and J. Ancheyta, Catal. Today, 2008,
130, 345–353.
29 L. K. De Souza, J. R. Zamian, G. N. da Rocha Filho, L. E.
Soledade, I. M. dos Santos, A. G. Souza and C. E. da Costa,
Dyes Pigm., 2009, 81, 187–192.
30 M. B. Zellner and J. G. Chen, Catal. Today, 2005, 99,
299–307.
9 A. O. Menezes, M. T. Rodrigues, A. Zimmaro, L. E. Borges
and M. A. Fraga, Renewable Energy, 2011, 36, 595–599.
10 M. S. Khayoon and B. H. Hameed, Appl. Catal., A, 2013, 464,
191–199.
11 S. B. Umbarkar, T. V. Kotbagi, A. V. Biradar, R. Pasricha,
J. Chanale, M. K. Dongare and E. Payen, J. Mol. Catal. A:
Chem., 2009, 310, 150–158.
12 S. Zhu, Y. Zhu and X. Gao, et al., Bioresour. Technol., 2013,
130, 45–51.
˜
´
31 M. D. Abad, M. A. Munoz-Marquez, S. El Mrabet, A. Justo
´
´
13 S. Zhu, X. Gao, F. Dong, Y. Zhu, H. Zheng and Y. Li, J. Catal.,
2013, 306, 155–163.
and J. C. Sanchez-Lopez, Surf. Coat. Technol., 2010, 204,
3490–3500.
14 M. S. Khayoon and B. H. Hameed, Appl. Catal., A, 2012, 433, 32 T. Ghodselahi, M. A. Vesaghi, A. Shafiekhani, A. Baghizadeh
152–161. and M. Lameii, Appl. Surf. Sci., 2008, 255, 2730–2734.
15 S. Zhu, X. Gao, Y. Zhu, Y. Zhu, H. Zheng and Y. Li, J. Catal., 33 M. C. Biesinger, L. W. Lau, A. R. Gerson and R. S. C. Smart,
2013, 303, 70–79. Appl. Surf. Sci., 2010, 257, 887–898.
16 X. Luo, X. Ge, S. Cui and Y. Li, Bioresour. Technol., 2016, 215, 34 V. Mazzieri, F. Coloma-Pascual, A. Arcoya, P. C. L’Argentiere
`
´
and N. S. Fı, Appl. Surf. Sci., 2003, 210, 222–230.
144–154.
17 D. G. Lahr and B. H. Shanks, Ind. Eng. Chem. Res., 2003, 42, 35 F. B. Li and X. Z. Li, Chemosphere, 2002, 48, 1103–1111.
5467–5472.
36 C. Nethravathi, E. A. Anumol, M. Rajamathi and N. Ravishankar,
´
´
18 S. Garcıa-Fernandez, I. Gandarias, J. Requies, M. B. Gu¨emez,
Nanoscale, 2011, 3, 569–571.
S. Bennici, A. Auroux and P. L. Arias, J. Catal., 2015, 323, 37 M. A. Dasari, P. P. Kiatsimkul, W. R. Sutterlin and G. J. Suppes,
65–75.
Appl. Catal., A, 2005, 281, 225–231.
19 E. P. Maris and R. J. Davis, J. Catal., 2007, 249, 328–337.
20 D. K. Sohounloue, C. Montassier and J. Barbier, React. Kinet.
Catal. Lett., 1983, 22, 391–397.
21 C. Montassier, J. C. Menezo, L. C. Hoang, C. Renaud and
J. Barbier, J. Mol. Catal., 1991, 70, 99–110.
22 J. Feng, H. Fu, J. Wang, R. Li, H. Chen and X. Li, Catal.
Commun., 2008, 9, 1458–1464.
23 N. Ji, T. Zhang, M. Zheng, A. Wang, H. Wang, X. Wang and
J. G. Chen, Catal. Today, 2009, 147, 77–85.
38 Y. Kusunoki, T. Miyazawa, K. Kunimori and K. Tomishige,
Catal. Commun., 2005, 6, 645–649.
39 T. Miyazawa, Y. Kusunoki, K. Kunimori and K. Tomishige,
J. Catal., 2006, 240, 213–221.
40 T. Miyazawa, S. Koso, K. Kunimori and K. Tomishige,
Appl. Catal., A, 2007, 318, 244–251.
41 G. Zhang, W. Guan, H. Shen, X. Zhang, W. Fan, C. Lu and
W. Shi, Ind. Eng. Chem. Res., 2014, 53, 5443–5450.
42 N. Li, H. Teng, L. Zhang, J. Zhou and M. Liu, RSC Adv., 2015,
5, 95394–95400.
24 A. Wang and T. Zhang, Acc. Chem. Res., 2013, 46, 1377–1386.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018
New J. Chem., 2018, 42, 3633--3641 | 3641