Aqueous-phase reforming of ethylene glycol on Co/ZnO catalysts prepared by the coprecipitation method

Article abstract of DOI:10.1016/j.molcata.2010.11.024

Co/ZnO catalysts with different Co/Zn ratios have been prepared by the coprecipitation method. It is revealed that below the nominal Co/Zn molar ratio of 2, the calcined Co/ZnO samples were constituted by ZnO and ZnCo ₂O₄ spinel. At the nominal Co/Zn ratio of 2, a small amount of Co₃O₄ spinel emerged. After reduction, the catalysts were composed of fcc Co and ZnO. In aqueous-phase reforming (APR) of ethylene glycol, it is found that the intrinsic activity and selectivity to H₂ increase with the increment of the ZnO content. H₂ selectivities over the Co/ZnO catalysts ranged from 52% to 89%, which are substantially higher than that of the Raney Ni catalyst at similar conversion. Moreover, the Co/ZnO catalysts produced much less CO in the product gas. With the aid of literature works as well as additional experiments on APR of acetic acid, methanol, and ethanol, the reaction pathways of ethylene glycol in the presence of water on the Co/ZnO catalysts were discussed.

Full text of DOI:10.1016/j.molcata.2010.11.024

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Journal of Molecular Catalysis A: Chemical 335 (2011) 129–135
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Aqueous-phase reforming of ethylene glycol on Co/ZnO catalysts prepared by the
coprecipitation method
Xianwen Chu, Jun Liu, Bo Sun, Rui Dai, Yan Pei∗, Minghua Qiao∗∗, Kangnian Fan
Department of Chemistry and Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials, Fudan University, No. 220 Handan Road, Shanghai 200433, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
Co/ZnO catalysts with different Co/Zn ratios have been prepared by the coprecipitation method. It is
revealed that below the nominal Co/Zn molar ratio of 2, the calcined Co/ZnO samples were constituted
by ZnO and ZnCo2O4 spinel. At the nominal Co/Zn ratio of 2, a small amount of Co3O4 spinel emerged. After
reduction, the catalysts were composed of fcc Co and ZnO. In aqueous-phase reforming (APR) of ethylene
glycol, it is found that the intrinsic activity and selectivity to H2 increase with the increment of the ZnO
content. H2 selectivities over the Co/ZnO catalysts ranged from 52% to 89%, which are substantially higher
than that of the Raney Ni catalyst at similar conversion. Moreover, the Co/ZnO catalysts produced much
less CO in the product gas. With the aid of literature works as well as additional experiments on APR of
acetic acid, methanol, and ethanol, the reaction pathways of ethylene glycol in the presence of water on
the Co/ZnO catalysts were discussed.
Received 15 September 2010
Received in revised form
19 November 2010
Accepted 22 November 2010
Available online 2 December 2010
Keywords:
Co/ZnO
Coprecipitation
Ethylene glycol
Aqueous-phase reforming
Hydrogen
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
metals [6,19], so the Co-based catalyst is anticipated to be promis-
ing for the APR process.
Approximately 80% of the present world energy demand comes
from fossil fuels [1], which are exhaustible and cause seri-
ous environmental problems. Recently, Dumesic and coworkers
demonstrated the new process of aqueous-phase reforming (APR)
of biomass-derived polyols such as ethylene glycol, glycerol, and
sorbitol to H2 and hydrocarbons [2–4]. The APR process is operated
at temperatures near 500 K, at which the water–gas shift (WGS)
reaction is thermodynamically favored, making it possible to gener-
ate H2 and hydrocarbons in a single reactor with trace amount of CO
[5]. This process is energy-efficient and green-house gas-neutral,
thus opening a new opportunity for the utilization of readily avail-
able renewable biomass.
It has been shown that the Co-based catalysts are efficient in
steam reforming (SR) of ethanol [20]. Homs and coworkers have
prepared Co catalysts by impregnating Co2(CO)8 on a series of sup-
ports, and found that the ZnO-supported Co catalyst exhibited the
best catalytic performance in SR of ethanol [21]. Motivated by their
work, we prepared Co/ZnO catalysts with different Co/Zn ratios
by the coprecipitation method using less hazardous cobalt nitrate.
Their catalytic behaviors in APR of ethylene glycol were investi-
gated and correlated with the characterization results. The reaction
pathways of ethylene glycol on the Co/ZnO catalysts in the presence
of water were discussed with the aid of additional experiments on
APR of acetic acid, methanol, and ethanol.
Catalysts that have been studied in the APR of ethylene gly-
col include Pt-based catalysts [2,6–13], SiO2-supported Ni, Pd, Ru,
Rh, and Ir catalysts [6], Sn-modified Raney Ni [3,8,14,15], Ni/Al2O3
catalysts [8,14], rapidly quenched skeletal Ni [16] and NiMo cata-
lysts [17], and non-pyrophoric Ni catalyst derived from traditional
Ni50Al50 alloy [18]. However, the APR performance of Co-based
catalyst has not been explored as far as we are aware of. It is
acknowledged that Co exhibits appreciable activities for C–C bond
scission, WGS, and Fischer–Tropsch synthesis (FTS) among VIIIB
2. Experimental
2.1. Catalyst preparation
The Co/ZnO catalysts with four different Co/Zn molar ratios were
prepared by the coprecipitation method. An aqueous solution fixed
at 0.4 M containing a mixture of Zn(NO3)2 and Co(NO3)2 was added
drop by drop to an aqueous solution of Na2CO3 (0.4 M) as the pre-
cipitant at 333 K under stirring. After being aged at 333 K for 2 h,
the solid was filtered, washed with deionized water to neutrality,
dried at 373 K overnight, and then calcined at 723 K for 4 h at a heat-
ing rate of 10 K min−1 in static air. The catalysts were obtained by
reduction under 5 vol.% H2/Ar at 723 K for 2 h at a heating rate of
∗
Corresponding author. Tel.: +86 21 55664679; fax: +86 21 6562978.
Corresponding author. Tel.: +86 21 55664679; fax: +86 21 6562978.
∗∗
E-mail addresses: peiyan@fudan.edu.cn (Y. Pei), mhqiao@fudan.edu.cn (M. Qiao).
1381-1169/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2010.11.024

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
130
X. Chu et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 129–135
Table 1
Physicochemical properties of the Co and Co/ZnO samples before and after reduction.
a
a
b
Sample
Co/Zn (molar ratio)
SBET (m2 g−1
)
dZnCo 4 (nm)a
dCo (nm)
dZnO (nm)
SH (m2 g−1
)
O
2
Nominal
Exp.
Cal.
Red.
Cal.
Red.
Co/ZnO-13
Co/ZnO-12
Co/ZnO-11
Co/ZnO-21
Co3O4
0.33
0.5
1.0
2.0
/
0.5
0.8
1.3
2.2
/
41
44
45
47
42
12
11
11
9
12
13
13
12
20
24
27
28
71
14
13
10
/
23
28
27
27
/
1.1
1.3
1.8
2.3
0.2
2
20 (Co3O4)
/
a
Crystallite size determined by XRD.
Active surface area determined by H2 chemisorption.
b
1 K min−1. The nominal Co/Zn molar ratios were 1/3, 1/2, 1/1, and
2/1 for catalysts labeled Co/ZnO-13, Co/ZnO-12, Co/ZnO-11, and
Co/ZnO-21, respectively. Co3O4 was prepared in a way similar to
that of the calcined Co/ZnO samples in the absence of Zn(NO3)2.
We have also prepared the Co/ZnO-13 catalyst by the impregna-
tion method (Co/ZnO-13-imp). However, it is identified that the
catalytic performance in APR of ethylene glycol on the Co/ZnO-13-
imp catalyst was much inferior to that on the Co/ZnO-13 catalyst
prepared by the coprecipitation method.
tion) × (weight fraction of ethylene glycol in the feed)/(weight of
the catalyst)) of 0.59 h−1 if not specified. Argon was used to reg-
ulate the system pressure. The reaction course was monitored by
sampling the gas and liquid products at intervals, followed by gas
chromatographic analysis. For the gas effluent, H2, CO, CH4, and
˚
CO2 were separated by a 5 A molecular sieve-packed column, and
examined by TCD. CH4, C2H4, C2H6, C3H6, C3H8, C4H8, and C4H10
were separated by a Porapak R packed column, and examined by a
flame ionization detector (FID). The liquid effluent was separated
by a CP-Wax 52 CB capillary column (30 m × 0.25 mm × 0.25 ␮m),
and examined by FID [16]. The selectivities to H2 and hydrocarbons
were calculated based on independent hydrogen and carbon bal-
ances, respectively [14]. Steady-state product compositions were
achieved after 6 h on stream.
2.2. Catalyst characterization
The bulk composition was determined by inductively coupled
plasma-atomic emission spectroscopy (ICP-AES; Thermo Elemen-
tal IRIS Intrepid). The BET surface area was measured on a
Micromeritics TriStar3000 adsorption apparatus by N2 physisorp-
tion at 77 K. Prior to the measurement, the sample in the glass
adsorption tube was degassed at 423 K under a N2 flow for 2 h.
X-ray diffraction patterns (XRD) were acquired on a Bruker D8
Advance X-ray diffractometer using Ni-filtered Cu K␣ radiation
3. Results and discussion
3.1. Physicochemical properties
Table 1 lists the bulk compositions and BET surface areas of the
Co/ZnO samples. It is found that the experimental Co/Zn molar
ratios always exceed the nominal values, suggesting that zinc
cations are more reluctant to precipitate than cobalt cations under
the present preparation conditions. With the increment of the
Co/Zn ratio, the experimental ratios approach the nominal values.
For the calcined Co/ZnO samples, the BET surface area increases
slightly from 41 to 47 m2 g−1 with the increment of the nominal
Co/Zn ratio from 1/3 to 2/1. After reduction, the BET surface areas
◦
˚
(1.5418 A) with a scanning angle (2ꢀ) of 10–90 . The voltage was
40 kV, and the current was 40 mA. The Scherrer equation was used
to estimate the crystallite size. Temperature-programmed reduc-
tion (TPR) profiles were collected on a home-made apparatus with
the 5 vol.% H2/Ar flow rate of 40 ml min−1 and the heating rate of
10 K min−1. The amount of H2 consumed was quantified with a
thermal conductivity detector (TCD).
H2 chemisorption was carried out on a home-made apparatus
equipped with a TCD detector. A catalyst sample of ca. 300 mg
was reduced at 723 K for 2 h in 5 vol.% H2/Ar. After reduction, the
temperature was kept at 723 K for 2 h and then cooled to room tem-
perature in an Ar flow. The chemisorption was carried out at 423 K,
and the metal surface area (SH) of the catalyst was calculated based
on the assumption of a H:Co stoichiometry of 1:1 according to Reuel
and Bartholomew [22].
The surface morphology was observed by transmission electron
microscopy (TEM; JEOL JEM2011). X-ray photoelectron spectrum
(XPS; Perkin Elmer PHI5000C) was recorded using Mg K␣ radiation
as the excitation source (1253.6 eV). The sample, pressed into a self-
supported disc, was mounted on the sample plate and degassed in
the pretreatment chamber in vacuo at room temperature for 4 h.
The sample was then transferred to the analyzing chamber with a
background pressure better than 2 × 10−9 Torr. All binding energy
(BE) values were referenced to the C 1 s peak of contaminant carbon
at 284.6 eV with an uncertainty of 0.2 eV.
of all the Co/ZnO catalysts are decreased to about 10 m2 g−1
.
The XRD patterns of the calcined Co/ZnO samples are shown in
Fig. 1a. For the Co/ZnO-13 sample with the lowest Co/Zn ratio, the
diffraction peaks can be assigned to a combination of ZnO (JCPDS
36-1451) and ZnCo2O4 (JCPDS 23-1390) or Co3O4 (JCPDS 42-1467).
It is noted that the last two spinel phases are indistinguishable
by XRD, because their standard diffractograms are similar. How-
ever, one can find in Fig. 1a that with the increment of the Co/Zn
ratio, the diffraction peaks of ZnO are attenuated, and nearly dis-
appear for the Co/ZnO-21 sample with the nominal Co/Zn molar
ratio of 2, a stoichiometry identical to that in ZnCo2O4. This phe-
nomenon strongly suggests the formation of ZnCo2O4 rather than
Co3O4 in the calcined Co/ZnO samples. For the Co/ZnO-21 sam-
ple, Table 1 shows that the experimental Co/Zn molar ratio is 2.2,
implying that in this sample about 10% of Co is in the form of Co3O4.
For comparison, Homs and coworkers identified exclusively Co3O4
on their calcined Co/ZnO sample prepared by the impregnation
method [23], suggesting that the coprecipitation method results
in a stronger interaction between the oxides of cobalt and zinc.
After reduction, Fig. 1b shows that ZnCo2O4 as well as Co3O4 for
the Co/ZnO-21 sample originally in the calcined Co/ZnO samples
are diminished. Instead, new features at 2ꢀ of 44.3, 51.5, and 75.9◦
ascribable to fcc Co (JCPDS 15-0806) emerge. The diffraction peaks
of ZnO, which are not available on the calcined Co/ZnO-21 sam-
ple, are identified after reduction, signifying the decomposition of
2.3. Activity test and product analysis
A reactor system similar to that of Shabaker et al. [7] was used
for APR of ethylene glycol. An aqueous solution containing 5 wt%
of ethylene glycol was fed to the reactor in an up-flow configura-
tion. The reforming was operated at a temperature of 498 K, system
pressure of 2.58 MPa, and WHSV ((weight flow rate of the feed solu-

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
X. Chu et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 129–135
131
a
ZnCo2O4
ZnO
Co3O4
Co/ZnO-21
Co/ZnO-13
Co/ZnO-12
Co/ZnO-11
Co/ZnO-11
Co/ZnO-12
Co/ZnO-13
Co/ZnO-21
80
20
40
60
350
450
550
650
750
850
950
2θ (degree)
T (K)
b
Fig. 2. H2-TPR profiles of Co3O4 and the Co/ZnO samples.
Co
ZnO
dent that there is an additional shoulder peak at 647 K superposing
on the high-temperature peak. This shoulder peak appears at a
temperature close to that of the high-temperature peak of Co3O4,
suggesting their similar origin. We also integrated the areas of the
low-temperature peak and the high-temperature peak including
the shoulder peak, and obtained a ratio of 1/2.22. This ratio is
much larger than that of Co3O4, while slightly smaller than that
of ZnCo2O4, signifying again that the major phase in the calcined
Co/ZnO-21 sample is ZnCo2O4.
The TEM images of the reduced Co/ZnO catalysts are presented
in Fig. 3. The dark metallic Co nanoparticles are dispersed on the
platelet-like ZnO. The average particle size of metallic Co measured
in the TEM images increases from about 25 to 50 nm with the nom-
inal Co/Zn molar ratio from 1/3 to 2/1, while the particle size of
ZnO is in the range of 45–80 nm for all the catalysts. The particle
sizes of Co and ZnO observed in Fig. 3 are much larger than the
crystallite sizes derived from XRD, suggesting their polycrystalline
nature.
Co/ZnO-13
Co/ZnO-12
Co/ZnO-11
Co/ZnO-21
20
40
60
80
2θ (degree)
Fig. 1. XRD patterns of the Co/ZnO samples (a) before and (b) after reduction.
ZnCo2O4 to ZnO and Co. Moreover, the diffraction peaks of ZnO are
very sharp for the reduced catalysts. Based on X-ray line broad-
ening, the crystallite sizes of ZnO, ZnCo2O4, and metallic Co in the
calcined and reduced Co/ZnO samples are calculated. Table 1 shows
that the crystallite sizes of ZnO and metallic Co in the reduced cat-
alysts are about twice as those of ZnO and ZnCo2O4 in the calcined
counterparts, which can be an important reason for the decreased
BET surface areas after reduction.
Fig. 4a presents the HRTEM image of the calcined Co/ZnO-13
sample. The lattice fringes in direct space corresponding to the
˚
(1 1 1) plane of ZnCo2O4 of 4.66 A and the (1 0 1) plane of ZnO
˚
of 2.48 A are readily visible. Fig. 4b shows the HRTEM image of
the reduced Co/ZnO-13 catalyst. The particle labeled ‘A’ on top of
the ZnO platelet displays a complex high-resolution image. The
Fourier transform image of this area affords spots corresponding
Fig. 2 shows the H2-TPR profiles of the Co/ZnO samples. The
reduction degree of the Co/ZnO-13, Co/ZnO-12, Co/ZnO-11, and
Co/ZnO-21 catalysts are 86%, 91%, 94%, and 95%, respectively.
The TPR profile of a Co3O4 sample is also presented in the
figure for comparison. Co3O4 shows two distinct H2 consump-
tion peaks centered at 575 and 644 K, with the intensity ratio
between the low- and high-temperature peaks being ca. 1/3,
which corresponds exactly to a two-step reduction mechanism of
Co3O4 (Co3O4 + H2 → 3CoO + H2O, 3CoO + 3H2 → 3Co + 3H2O) [24].
On the other hand, the TPR profiles of the calcined Co/ZnO-13,
Co/ZnO-12, and Co/ZnO-11 samples are similar to each other.
These samples also give two H2 consumption peaks, but the
peak maxima (573 and 700 K) deviate noticeably from those of
Co3O4. Moreover, the intensity ratio of the peak area between
the low- and high-temperature peaks for these samples is ca.
1/1.99, which conforms to the stoichoimetry of the two-step reduc-
tion mechanism of ZnCo2O4 (ZnCo2O4 + H2 → 2CoO + ZnO + H2O,
2CoO + 2H2 → 2Co + 2H2O) [25–27].
˚
to the Co(1 1 1) diffraction at 2.04 A and the ZnO(1 0 1) diffraction
˚
at 2.48 A, which are consistent with the assignments based on XRD.
The highly defective structure of area ‘A’ may be an indication of
the out-growth of metallic Co from the bulk of the ZnCo2O4 spinel
during the reduction process.
Fig. 5 shows the Co 2p spectra of the calcined and the reduced
Co/ZnO samples. For the calcined samples (Fig. 5a), the Co 2p3/2 BE
is 780.4 eV, and the BE difference between the 2p1/2 and 2p3/2 lev-
els is 15.0 eV, which can be ascribed either to ZnCo2O4 or to Co3O4
which are not distinguishable if only based on these two parame-
ters [28,29]. However, compared with the standard Co 2p spectra of
ZnCo2O4 and Co3O4 reported by Zsoldos and Guczi [28], the pres-
ence of a weak satellite peak at ca. 9 eV higher than the Co 2p3/2
peak observed in Fig. 5a supports the formation of predominantly
ZnCo2O4.
After reduction, Fig. 5b shows that there are two Co 2p3/2 peaks
with BEs of 778.0 and 780.2 eV, respectively. The former is readily
assignable to metallic Co [30], while the latter, combining with the
strong satellite peak at ca. 6 eV higher BE, is assigned to CoO [31].
For the calcined Co/ZnO-21 sample, although its TPR profile
resembles more the profiles of other Co/ZnO samples, it is evi-

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
132
X. Chu et al. / Journal of Molecular Catalysis A: Chemical 335 (2011) 129–135
Fig. 3. TEM images of Co/ZnO catalysts after reduction. (a) Co/ZnO-13, (b) Co/ZnO-12, and (c) Co/ZnO-11, and (d) Co/ZnO-21.
It is observed that the proportion of metallic Co increases with the
increment of the Co/Zn ratio, which is consistent with the ascend-
ing active surface area (SH, Table 1) and reduction degree with the
increment of the Co/Zn ratio.
3.2. APR performance
Table 2 compiles the selectivities to H2 and hydrocarbons and
kinetic data for APR of ethylene glycol over the Co/ZnO catalysts
Fig. 4. HRTEM images of the Co/ZnO-13 sample (a) before and (b) after reduction. The inset corresponds to the Fourier-transformed image of area marked as ‘A’.