Journal of Power Sources 343 (2017) 437e445

Contents lists available at ScienceDirect

Journal of Power Sources

journal homepage: www.elsevier.com/locate/jpowsour

Rapid synthesis of Fe-doped CuO-Ce0.8Zr0.2O2 catalysts for CO

preferential oxidation in H2-rich streams: Effect of iron source and the
ratio of Fe/Cu
Jing Wang a, Caiyun Han a, Xiaoya Gao a, Jichang Lu a, Gengpin Wan a, b, Dedong He a,
Ran Chen a, Kezhen Chen a, Sufang He c, Yongming Luo a, *
a
Faculty of Environmental Science and Engineering, Kunming University of Science and Technology, Kunming 650500, PR China
b
Research Center for Analysis and Measurement, Hainan University, Haikou 570228, PR China
c
Research Center for Analysis and Measurement, Kunming University of Science and Technology, Kunming 650093, PR China

h i g h l i g h t s

Fe doped Cu-Ce-Zr (CuCZ) catalysts were rapidly synthesized within 30 min.

Both iron source and the ratio of Fe/Cu have significant effects on catalytic behaviors.

Fe(NO3)3 is superior to FeCl3 and Fe2(SO4)3, and the optimal ratio of Fe/Cu is 1/10.

1/10Fe(N)-CuCZ exhibits excellent stability in the 228 h of time-on-stream test.

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

Article history: A facile route (urea grind combustion method) is described for the rapid synthesis of Fe-doped Cu-Ce-Zr
Received 26 May 2016 catalysts within 30 min through simple grinding and combustion. The effects of iron source and Fe/Cu
Received in revised form mass ratio on the performances of the catalysts for CO preferential oxidation (CO-PROX) are evaluated.
16 January 2017
The influences of H2O, CO2, and their mixture on the activity as well as stability of the catalysts are also
Accepted 19 January 2017
investigated. The samples are characterized by XRD, N2 adsorption-desorption, H2-TPR, TEM, Raman and
XPS. Fe(NO3)3 is found to be superior to FeCl3 and Fe2(SO4)3 as the iron source for Fe-CuCZ catalyst.
Among the different synthesized catalysts, 1/10Fe(N)-CuCZ is found to be the most active catalyst, indi-
Keywords:
CO-PROX
cating that the optimal Fe/Cu mass ratio is 1/10. The influences of H2O, CO2, and H2O þ CO2 on the
Fe-doped Cu-Ce0.8Zr0.2O2 catalytic performance of 1/10Fe(N)-CuCZ are in the order of CO2 < CO2 þ H2O < H2O. 1/10Fe(N)-CuCZ
Iron source exhibits excellent stability during a 228 h time-on-stream test. 1/10Fe(N)-CuCZ shows the highest cata-
Fe/Ce ratio lytic activity and excellent stability even in the presence of H2O and CO2. The excellent catalytic per-
Stability formance can be attributed to the synergy between the highly dispersed copper species and ceria, as well
as the formation of more oxygen vacancies and reduced copper species.
© 2017 Elsevier B.V. All rights reserved.

1. Introduction oxygen and produces water (the anode reaction: 2H2 / 4Hþ þ 4e;
the cathode reaction: O2 þ 4Hþ þ 4e / 2H2O) [1]. However,
Hydrogen (H2), as a clean fuel, has been a subject of growing 0.5e2.0 vol% of carbon monoxide (CO) is found in the corre-
interest in the recent decades. The H2-rich stream used as fuel for sponding gas products. This amount of CO (10 ppm and above)
polymer exchange membrane fuel cells (PEMFCs) is typically would poison the Pt-based electrode of PEMFCs, which would
generated by steam reforming of hydrocarbons. PEMFCs make use seriously affect the service life of the fuel cell seriously. Therefore, it
of the chemical conversion of energy when hydrogen reacts with is essential to develop efficient techniques to reduce the CO con-
centration to a trace level (below 10 ppm [2]).
It has been demonstrated that CO preferential oxidation (CO-
PROX) is one of the most promising, cost-effective and straight-
* Corresponding author.
forward methods for eliminating CO from H2-rich streams [3].
E-mail address: [email protected] (Y. Luo).

http://dx.doi.org/10.1016/j.jpowsour.2017.01.084
0378-7753/© 2017 Elsevier B.V. All rights reserved.
438 J. Wang et al. / Journal of Power Sources 343 (2017) 437e445

The selection of catalyst is the most important factor for CO-PROX. XFe(N)-CuCZ, in which the loading of CuO was kept at a constant of
The ideal catalyst for CO-PROX should have good activity at low 10 wt% and X is the mass ratio of Fe/Cu (X ¼ 1/20, 1/15, 1/10 and 1/
reaction temperature (<120  C), stability, and tolerance to the 5). In order to investigate the influence of iron source,
presence of CO2 and H2O in the feed gas [3]. Various catalysts such Fe(NO3)3$9H2O was replaced with Fe2(SO4)3$9H2O or FeCl3$6H2O,
as CuO [4,5], a-Fe2O3 [6], TiO2 [7,8], Co3O4 [9e11] and CeO2 [12e14] which were then used to prepare 1/10Fe(S)-CuO-Ce0.8Zr0.2O2 and 1/
have been investigated extensively for CO preferential oxidation. 10Fe(C)-CuO-Ce0.8Zr0.2O2, named as 1/10Fe(S)-CuCZ and 1/10Fe(C)-
Among these, CuO-CeO2 has become the most attractive choice due CuCZ respectively. For comparison, CuO-Ce0.8Zr0.2O2 was also pre-
to its high activity together with high selectivity for CO-PROX in H2- pared by urea grind combustion method, and denoted as CuCZ.
rich gas [14e16]. Nevertheless, CuO-CeO2 catalysts exhibit a rela-
tively narrow temperature window (5-20  C) full CO conversion
[17e20]. Many efforts have been devoted to widening the tem- 2.2. Catalyst characterization
perature window for 100% conversion of CO and further improving
low-temperature activity by adding transition metals into CuO- N2 adsorption-desorption isotherms were measured by a V-Sorb
CeO2 catalyst [17e19]. For example, it was found that the incor- 2800P instrument at 196  C after the pretreatment of catalyst
poration of Zr can improve the mobility of lattice oxygen and samples by degassing at 200  C for 3 h. Specific surface areas were
enhance the redox properties of the Cu/Ce pair, which will help obtained by the Brunauer-Emmett-Teller (BET) method and the
improve the catalytic performances [17]. Iron and nickel oxides adsorption data. Pore sizes were calculated according to the
were used to improve catalytic activities of CuO-CeO2 catalysts with desorption data by Barrett-Joyner-Halenda (BJH) method. XRD
low loading of CuO (<5 wt%), by facilitating the formation of patterns were recorded by means of the Rigaku D/max-1200
additional amount of active sites related to copper oxide [21]. diffractometer with Cu Ka radiation (l ¼ 1.5406 Å) at 2q range
Moreover, it was also demonstrated that the combination of Fe and between 10 and 90 . The surface element compositions and the
Ce oxide in the catalyst could help to increase the surface area and corresponding chemical states of all the samples were determined
pore volume. Also, the addition of a-Fe2O3 could promote the cat- by XPS performed on a PHI 5000 VersaProbe II with non-
alytic activities of Cu-Ce-Fe-O catalysts [22]. The preparation monochromatic Al Ka radiation (1486.6 eV) as the excitation X-
method of catalyst is one of the key factors affecting the catalytic ray source. The C 1s photoelectron peak at 284.6 eV was used as a
performance of Cu-based catalysts, which would result in different reference to calibrate binding energy. Raman spectra were obtained
structural properties of the catalyst and the changes in dispersion by a Via reflex confocal Raman microscope with a laser (514 nm)
of copper species [23]. So, many investigators have developed working at 5 mV power. Transmission electron microscopy (TEM)
various preparation methods to enhance catalytic behaviors of Cu- and high resolution transmission electron microscopy (HR-TEM)
based catalysts for CO-PROX [14,24]. The common preparation images were obtained by the JEOL JEM-2100 microscope with an
methods involve deposition-precipitation, impregnation, co- accelerating voltage of 200 kV.
precipitation, sol-gel method and so on [25]. However, the cur- Hydrogen temperature programmed reduction (H2-TPR) was
rent synthesis procedures for obtaining high performance CuO- performed in a fixed-bed reactor as follows: prior to H2-TPR mea-
CeO2 catalysts are generally complex and time-consuming. surement, 50 mg of catalyst sample was pretreated at 400  C for 1 h
Recently, our group reported a facile route to synthesize CuO- in 5 vol% O2/Ar gas mixture. Then, the pretreated sample was
Ce0.8Zr0.2O2 with high performance for CO-PROX [26]. However, the cooled to 100  C under the flow of ultra-pure argon. After that, TPR
low-temperature CO conversion together with the temperature experiment was carried out in the flow rate of 10% H2/Ar (30 cm3/
window for full conversion still needs to be improved. Therefore, in min), and the temperature was increased from 100  C to 500  C
order to improve the catalytic performance of CuO-Ce0.8Zr0.2O2 for with a heating ramp of 10  C/min. H2 consumption was measured
CO-PROX, it has been doped with Fe for the first time. by an on-line thermal conductivity detector (TCD).
In this paper, Fe-doped CuO-Ce0.8Zr0.2O2 catalysts were rapidly
synthesized within 30 min and were evaluated in the CO-PROX
reaction for the first time. Even, the synthesis route is not limited
2.3. Catalytic activity measurements
to the use of Cu-Ce catalyst system for CO-PROX. It can be used to
synthesize a large number of other catalysts for a variety of re-
CO preferential oxidation was performed in a tubular reactor at
actions. The Fe-doped CuO-Ce0.8Zr0.2O2 catalysts were character-
atmospheric pressure, and 300 mg of the catalyst with 40e60 mesh
ized by N2 adsorption-desorption, XRD, H2-TPR, TEM, Raman and
size particles was loaded in each test. The typical composition of
XPS. The influences of iron source and the mass ratio of Fe/Cu on
the feed gas mixture was 1% CO, 1% O2, 50% H2 in volume and
the catalysts for CO-PROX were investigated in detail. Moreover, the
balanced with He. The flow rate of the feed gas was fixed at 80 ml/
effects of CO2, H2O, and the mixture of CO2 and H2O on the catalytic
min, corresponding to 24000 h1 GHSV. The oxygen content in the
behaviors for CO-PROX as well as the stability of the catalyst were
feed was denoted in terms of the parameter l defined as 2  O2/CO
studied by a long-term (228 h) time-on-stream test.
mole ratio (l ¼ 2). To study the effect of CO2 and H2O on catalytic
activities, the corresponding experiments were performed with
2. Materials and methods
addition of 15% CO2 and/or 10% H2O into the feed gas, respectively.
The inlet and outlet concentrations of reactants and products were
2.1. Catalyst preparation
analyzed by an on-line gas chromatography system with two de-
tectors (one TCD and one FID) together with a methanation reactor.
XFe(N)-CuO-Ce0.8Zr0.2O2 catalysts were synthesized by using the
The catalytic performances are expressed in terms of the CO con-
urea grind combustion (UGC) method, which can be briefly
version (%) and the CO2 selectivity (%), which were calculated based
described as follows: specific calculated amounts of Ce(NO3)3$6H2O
on the consumption of CO and O2, as shown below:
(0.048 mol), Cu(NO3)2$3H2O (0.0169 mol), Zr(NO3)4$5H2O
(0.012 mol), Fe(NO3)3$9H2O and (NH2)2CO (0.152 mol) were mixed ½COin  ½COout
and grinded in agate mortar at room temperature until a trans- CO conversionð%Þ ¼  100 (1)
½COin
parent viscous gel was obtained. After that, the gel was calcined at
300  C for 20 min. The as-obtained catalyst was designated as
 J. Wang et al. / Journal of Power Sources 343 (2017) 437e445 439

volume is observed for 1/10Fe(N)-CuCZ, which indicates increase in

0:5  ½COin  ½COout porosity and surface area. The corresponding texture properties of
CO2 selectivity ð%Þ ¼  100 (2)
½O2 in  ½O2 out 1/10Fe(C)-CuCZ, 1/10Fe(S)-CuCZ, 1/10Fe(N)-CuCZ and CuCZ are sum-
marized in Table 1. As can be seen in Table 1, BET specific surface
where [CO]in and [CO]out are the inlet and outlet concentrations of area of 1/10Fe(N)-CuCZ (107 m2 g1) is far larger than that of CuCZ
CO (%(v/v)), and [O2]in and [O2]out correspond to the inlet and outlet (22 m2 g1), 1/10Fe(C)-CuCZ (30 m2 g1) and 1/10Fe(S)-CuCZ
concentrations of O2 (%(v/v)), respectively. (26 m2 g1). This higher surface area is beneficial for the dispersion
of copper species and the formation of strong interaction between
3. Results and discussion active copper species and support. Combined with the results of
catalytic activities, it can be concluded that the addition of iron
3.1. Effect of iron sources on CuCZ and Fe-doped CuCZ catalysts nitrate increases the specific surface area and catalytic activity of
the catalyst.
3.1.1. Catalytic activities
Fig. 1 shows the CO conversion and CO2 selectivity of CuCZ, 1/ 3.1.3. XRD and TEM characterizations
10Fe(C)-CuCZ, 1/10Fe(S)-CuCZ and 1/10Fe(N)-CuCZ. Compared with XRD patterns of CuCZ, 1/10Fe(C)-CuCZ, 1/10Fe(S)-CuCZ and 1/
CuCZ, a significant inhibition in the CO conversion is observed for 1/ 10Fe(N)-CuCZ are shown in Fig. 2A. The corresponding crystallite
10Fe(C)-CuCZ and 1/10Fe(S)-CuCZ, whereas a remarkable improve- dimensions were calculated according to the (111) plane (located at
ment is observed for 1/10Fe(N)-CuCZ. The maximum CO conversion 2q ¼ 28.7 ) by the Scherrer's equation and are tabulated in Table 1.
of 1/10Fe(S)-CuCZ and 1/10Fe(C)-CuCZ is 75.7% (at 180  C) and 88.6% Diffraction peaks indexed to fluorite-type cubic structure of CeO2
(at 140  C), respectively. However, 100% CO conversion of 1/10Fe(N)- (JCPDS No. 28-0271) are clearly observed in the XRD patterns of all
CuCZ and CuCZ is achieved at 120  C and 140  C, respectively. the catalysts. However, none of the diffraction peaks of iron species
Moreover, the temperature window for 100% CO conversion of 1/ were detected, implying that the doped iron species were highly
10Fe(N)-CuCZ is wider than that of CuCZ. As shown in Fig. 1B, for all dispersed or inserted into ceria-zirconia to form a solid solution
the four catalysts, there is almost no difference in the CO2 selec- [26]. Two weak diffraction peaks (located at 2q ¼ 35.5 and 38.8 )
tivity when the reaction temperature is below 100  C. With further assigned to the bulk or crystalline Cu oxides (JCPDS No.45-0937) are
increase in the reaction temperature, the CO2 selectivity of the observed in the patterns of 1/10Fe(C)-CuCZ and 1/10Fe(S)-CuCZ, but
catalysts follows the order: 1/10Fe(N)-CuCZ < CuCZ < 1/10Fe(C)- not in the patterns of CuCZ and 1/10Fe(N)-CuCZ. The presence of
CuCZ < 1/10Fe(S)-CuCZ. diffraction peaks indexed to the bulk or crystalline Cu oxides within
1/10Fe(C)-CuCZ and 1/10Fe(S)-CuCZ might be ascribed to the ag-
3.1.2. N2 adsorption-desorption gregation of CuO species on their surfaces [28], which suggests that
N2 adsorption-desorption isotherms were used to investigate both the doped iron chloride and iron sulfate are not conducive to
the influences of iron source on the surface area and pore structure the dispersion of CuO species, thus preventing the interaction be-
of CuCZ, 1/10Fe(C)-CuCZ, 1/10Fe(S)-CuCZ and 1/10Fe(N)-CuCZ. As tween Cu active species and support. On the other hand, Cu species
shown in Fig. S1, all the four catalysts exhibit type IV isotherms with of CuCZ and 1/10Fe(N)-CuCZ were highly dispersed on their surface
one hysteresis loop, which is characteristic of mesoporous mate- or incorporated into the solid solution of ceria-zirconia [20,29e31].
rials. It is observed that the hysteresis loop of 1/10Fe(N)-CuCZ is As can be seen by comparing the patterns of CuCZ and 1/10Fe(N)-
considerably different from that of 1/10Fe(C)-CuCZ, 1/10Fe(S)-CuCZ CuCZ (Fig. 2A, inset), the position of (111) plane reflection of CeO2
and CuCZ. The hysteresis loop of 1/10Fe(N)-CuCZ is composed of the for 1/10Fe(N)-CuCZ shifts slightly toward the larger 2q value (from
type-H2 and type-H3, suggesting that the sample has channel-like 28.8 to 28.9 ) and the corresponding diffraction peak becomes
or ink-bottle pores [27]. The hysteresis loop of the latter three broader and weaker than that of CuCZ. This observation indicates
samples is of type-H3, which indicates the presence of slit-shaped that a smaller sized ceria-zirconia based solid solution was gener-
pores among plate-like particles. The comparison of N2 ated within 1/10Fe(N)-CuCZ (listed in Table 1), which is beneficial to
adsorption-desorption isotherms of the catalysts indicates that forming large specific surface area, in agreement with the results of
doping CuCZ with iron nitrate would change the type of hysteresis N2 adsorption-desorption characterization.
loop and the structure of pore. Moreover, a greater N2 adsorbed TEM characterizations were performed in order to complement

Fig. 1. (A) CO conversion and (B) CO2 selectivity of CuCZ, 1/10Fe(C)-CuCZ, 1/10Fe(S)-CuCZ and 1/10Fe(N)-CuCZ. Reaction conditions: catalyst ¼ 0.3 g, total flow rate ¼ 80 ml/min
(CO ¼ 1%, O2 ¼ 1%, H2 ¼ 50%, balanced with He), GHSV ¼ 24000 h1.
440 J. Wang et al. / Journal of Power Sources 343 (2017) 437e445

Table 1
Textural properties of CuCZ and CuCZ doped with various iron source.

Samples Lattice parameter (nm)a Crystallite size of CeO2 d(111) (nm)b BET surface area (m2 g1)c Vp (cm3 g1)c H2 consumption (mmol/gcat)d

CuCZ 0.53549 12.52 22 0.05 1295

1/10Fe(C)-CuCZ 0.53496 11.24 30 0.04 1124
1/10Fe(N)-CuCZ 0.53378 6.53 107 0.11 3887
1/10Fe(S)-CuCZ 0.53516 10.75 26 0.06 1069
a
Lattice parameter of CeO2 was obtained by using MDI Jade 5.0 software according the data of XRD.
b
Crystallite size of CeO2 was determined from XRD by using the Scherrer's equation.
c
N2 adsorption-desorption data.
d
Calculated based on the H2 uptake before 400  C for the reduction of copper species.

Fig. 2. (A) XRD patterns and (B) H2-TPR profiles of CuCZ, 1/10Fe(C)-CuCZ, 1/10Fe(S)-CuCZ and 1/10Fe(N)-CuCZ.

the structural analyses of CuCZ, 1/10Fe(C)-CuCZ, 1/10Fe(S)-CuCZ and larger than that of CuCZ, combining the result of N2 adsorption-
1/10Fe(N)-CuCZ, as shown in Fig. S2. The TEM images show the desorption characterization. The difference between the H2-TPR
presence of agglomerates of near-spherical particles. The electron profile of CuCZ and 1/10Fe(N)-CuCZ is likely because the larger BET
diffraction (ED) data of catalysts show nearly pure CeO2-type re- surface is beneficial for promoting the dispersion of CuO species
flections. Moreover, the HR-TEM images of all the catalysts show and inhibiting the aggregation of CuO species to form bulk CuO. The
lattice fringes with interplanar distances about 0.31e0.32 nm, H2-TPR results are in good agreement with the XRD data. As shown
corresponding to the (1 1 1) facets of CeO2. The particle sizes of the in Fig. 2B, all the reduction peaks of 1/10Fe(C)-CuCZ and 1/10Fe(S)-
four samples are in the following order: CuCZ >1/10Fe(C)-CuCZ > 1/ CuCZ with smaller H2 consumption shift to the higher temperature,
10Fe(S)-CuCZ > 1/10Fe(N)-CuCZ (in agreement with Table 1). In implying that the addition of iron chloride and iron sulfate into
addition, it is clear that the dispersion of 1/10Fe(N)-CuCZ is higher CuCZ significantly inhibits the reduction of CuO species, due to the
than that of the other three catalysts. formation of bulk CuO with large particle size on the surface of the
two catalysts. This is in good agreement with the XRD results
(Fig. 2A). Moreover, the quantitative results of H2-TPR are provided
3.1.4. H2-TPR study
in Table 1. As iron chloride and iron sulfate are added, there is a
Fig. 2B presents H2-TPR profiles of CuCZ and Fe doped CuCZ
decrease in the H2 consumption, suggesting that iron chloride and
catalysts (1/10Fe(C)-CuCZ, 1/10Fe(S)-CuCZ and 1/10Fe(N)-CuCZ).
iron sulfate can lower the reducibility of the catalyst. On the other
Three reduction peaks (centered at 205, 220 and 246  C) are
hand, the H2 consumption is much higher for 1/10Fe(N)-CuCZ
detected clearly for CuCZ, which are attributed to the reduction of
compared to CuCZ. Thus, considering these results together with
highly dispersed CuO species strongly interacted with the solid
the activity results, it can be concluded that the iron source has a
solution of Ce0.8Zr0.2O2 [32,33], copper ions entered into the lattices
significant effect on the CO PROX reaction, and the optimal iron
of Ce0.8Zr0.2O2 [34] and three-dimensional clusters and/or bulk CuO
source, iron nitrate, is essential. In addition, for 1/10Fe(S)-CuCZ, the
[35], respectively. Only two reduction peaks (situated at 201 and
absorption bands corresponding to SO2 4 are detected at 1058, 1130
220  C) are observed in the profile of 1/10Fe(N)-CuCZ. Compared
and 1219 cm1 (Fig. S3). Therefore, the highest temperature of
with CuCZ, the low-temperature reduction peak of 1/10Fe(N)-CuCZ
reduction peaks should be correlated to the presence of SO2 4 ,
with larger H2 consumption shifts slightly to the lower temperature
which blocks the interaction of active copper species and CeO2. A
(from 205 to 201  C), indicating that large numbers of CuO species
small reduction peak around 450  C ascribed to the reduction of
with high dispersion and smaller size were formed within 1/
surface ceria [36,37] is also observed in the H2-TPR profile of 1/
10Fe(N)-CuCZ. Furthermore, it is also observed that the H2 con-
10Fe(S)-CuCZ, which might be due to the weak interaction between
sumption peak (at 220  C) for reducing copper ions (Cu2þ) within
CuO species and supports.
the lattice of Ce0.8Zr0.2O2 is far larger than that of CuCZ, which
On the basis of the above activity measurements and charac-
suggests that more Cu2þ was incorporated into the lattices of
terization results, it can be concluded that 1/10Fe(N)-CuCZ is su-
Ce0.8Zr0.2O2. In addition, the H2 consumption peak of reducing
perior to CuCZ, 1/10Fe(C)-CuCZ and 1/10Fe(S)-CuCZ. Therefore, in
three-dimensional clusters and/or bulk CuO was detected in CuCZ,
order to optimize the CuCZ catalyst doped with iron nitrate, the
but not in 1/10Fe(N)-CuCZ. The BET surface of 1/10Fe(N)-CuCZ is far
 J. Wang et al. / Journal of Power Sources 343 (2017) 437e445 441

effect of Fe/Cu ratio on Fe doped CuCZ catalyst was investigated and related closely to the fact that more Ce4þ (0.97 Å) ions were
characterized in the following sections of this paper. replaced with Cu2þ (0.73 Å) [26] and Fe3þ (0.64 Å) [38] ions, and
these smaller size ions were incorporated into the lattices of
Ce0.8Zr0.2O2 solid solution. When the ratio of Fe/Cu increases up to
3.2. Effect of the ratio of Fe/Cu on Fe(N)-CuCZ
1/5, the corresponding lattice parameter of 1/5Fe(N)-CuCZ is smaller
than that of CuCZ but larger than that of 1/10Fe(N)-CuCZ. This might
3.2.1. Catalytic activities of XFe(N)-CuCZ
be due to the fact that more Fe3þ ions preferentially entered into
Fig. 3 shows both the CO conversion and CO2 selectivity of
the lattices of Ce0.8Zr0.2O2 rather than Cu2þ ions, resulting in the
XFe(N)-CuCZ catalysts for CO-PROX reaction. It is clear that the CO
aggregation of copper oxide species on the surface. Moreover, the
conversion, especially in low-temperature range, increases as the
crystallite sizes of XFe(N)-CuCZ are smaller than that of CuCZ
mass ratio of Fe/Cu increases from 1/20 to 1/10. With further in-
(Table 2), which is comparable to the particle size calculated from
crease in the ratio of Fe/Cu from 1/10 to 1/5, there is a decrease in
TEM images (Fig. S4). Moreover, the morphologies of XFe(N)-CuCZ
the CO conversion. The CO2 selectivity of all XFe(N)-CuCZ catalysts
catalysts are similar to that of CuCZ.
decreases when the reaction temperature is higher than 120  C.
When the reaction temperature is below 100  C, CO2 selectivity of
1/10Fe(N)-CuCZ is comparable to that of 1/20Fe(N)-CuCZ and is 3.2.3. H2-TPR study of XFe(N)-CuCZ
higher than that of 1/15Fe(N)-CuCZ and 1/5Fe(N)-CuCZ. Therefore, 1/ H2-TPR profiles of all the four XFe(N)-CuCZ catalysts are dis-
10Fe(N)-CuCZ exhibits the highest catalytic activity among the played in Fig. 4B. Three H2 reduction peaks (a, b, g) are detected in
different catalysts, indicating that the optimal mass ratio of Fe/Cu is the profiles of 1/15Fe(N)-CuCZ and 1/5Fe(N)-CuCZ, which have been
1/10. ascribed to the reduction of highly dispersed CuO species strongly
interacted with the solid solution of Ce0.8Zr0.2O2 [32,33], Cu2þ ions
entered in the lattices of the Ce0.8Zr0.2O2 solid solution and bulk
3.2.2. XRD characterization of XFe(N)-CuCZ
CuO, respectively [34,35]. Only a and b reduction peaks are well
XRD patterns of XFe(N)-CuCZ are displayed in Fig. 4A, and the
expressed in the profile of 1/20Fe(N)-CuCZ. The occurrence of a
corresponding crystallite sizes and lattice parameters of XFe(N)-
reduction peak together with the absence of g reduction peak
CuCZ in comparison with those of CuCZ are tabulated in Table 2. As
suggests that the highly dispersed CuO species instead of bulk CuO
shown in Fig. 4A, a weak diffraction peak of crystalline CuO is
were formed within 1/20Fe(N)-CuCZ. It is in excellent agreement
detected for 1/15Fe(N)-CuCZ and 1/5Fe(N)-CuCZ, but not for 1/
with the XRD results, where the diffraction peak of CuO phase was
20Fe(N)-CuCZ and 1/10Fe(N)-CuCZ. The absence of CuO-related XRD
not observed. As can be seen by comparing Fig. 4B, two reduction
diffraction peak in 1/10Fe(N)-CuCZ might be due to the following
peaks (a and b) for 1/10Fe(N)-CuCZ are found to shift to the lower
two reasons: i) CuO species were incorporated into the solid solu-
temperature range. It indicates that the reducibility of 1/10Fe(N)-
tion of Ce0.8Zr0.2O2 with a cubic structure [26,36], ii) highly
CuCZ was improved by the addition of Fe (Fe/Cu ¼ 1/10) into CuCZ.
dispersed CuO species were present with very small crystallite size
Moreover, the disappearance of g reduction peak accompanied
and amorphous structure beyond the detection limit of XRD. As for
with the remarkable increase of a and b reduction peaks has been
1/15Fe(N)-CuCZ, the characteristic diffraction peak of CuO (located
attributed to the fact that the suitable addition of Fe into CuCZ
at 2q ¼ 38.8 ) is detected, which might be assigned to the aggre-
would not only promote the insertion of Cu2þ ions into the lattices
gation of copper oxides on its surface, implying that the low loading
of Ce0.8Zr0.2O2 but also facilitate the dispersion of CuO species as
of Fe is not beneficial to the dispersion of copper species. Further-
well as inhibit the formation of bulk CuO phase via the aggregation
more, since it was found in our previous work that Cu2þ ions
of CuO species.
(0.73 Å) partly substituted the Ce4þ ions (0.97 Å) and were incor-
porated into the solid solution of Ce0.8Zr0.2O2 [26], the increase in
the lattice parameter of 1/20Fe(N)-CuCZ and 1/15Fe(N)-CuCZ 3.2.4. Raman characterization of XFe(N)-CuCZ
compared to CuCZ indicates that the lower mass ratio of Fe/Cu Raman spectra of CuCZ and XFe(N)-CuCZ catalyst samples are
would restrict the insertion of Cu2þ ions into the lattices of illustrated in Fig. 5A. All the five samples show an intense band
Ce0.8Zr0.2O2 solid solution. The lattice parameter of 1/10Fe(N)-CuCZ around 458 cm1, assigned to the Raman active F2g vibrational
is the smallest among the four catalyst samples, which should be mode, which is an indication of the typical cubic fluorite structure

Fig. 3. (A) CO conversion and (B) CO2 selectivity of 1/20Fe(N)-CuCZ, 1/15Fe(N)-CuCZ, 1/10Fe(N)-CuCZ, 1/5Fe(N)-CuCZ. Reaction conditions: catalyst ¼ 0.3 g, total flow rate ¼ 80 ml/min
(CO ¼ 1%, O2 ¼ 1%, H2 ¼ 50%, balanced with He), GHSV ¼ 24000 h1.
442 J. Wang et al. / Journal of Power Sources 343 (2017) 437e445

Fig. 4. (A) XRD patterns and (B) H2-TPR profiles of 1/20Fe(N)-CuCZ, 1/15Fe(N)-CuCZ, 1/10Fe(N)-CuCZ, 1/5Fe(N)-CuCZ.

Table 2
Characteristics of CuCZ and XFe(N)-CuCZ catalysts.

CuCZ 1/20Fe(N)-CuCZ 1/15Fe(N)-CuCZ 1/10Fe(N)-CuCZ 1/5Fe(N)-CuCZ

Lattice parameter (nm)a 0.53549 0.53646 0.53583 0.53378 0.53440

Crystallite size of CeO2 (nm) d(111)b 12.52 11.44 10.29 6.53 11.96
a
Lattice parameter of CeO2 was obtained by using MDI Jade 5.0 software according the data of XRD.
b
Crystallite size was determined from XRD by using the Scherrer's equation.

Fig. 5. (A) Raman spectra and (B) the ratio of A599/A458 of CuCZ, 1/20Fe(N)-CuCZ, 1/15Fe(N)-CuCZ, 1/10Fe(N)-CuCZ, 1/5Fe(N)-CuCZ.

Ce0.8Zr0.2O2 [39,40]. The broad peak at about 599 cm1 is closely Moreover, the concentration of oxygen vacancies within catalyst
related to the oxygen vacancies within the lattice of Ce0.8Zr0.2O2 sample is related to the particle size. It has been reported that the
[41] and ascribed to the presence of oxygen defects in Ce0.8Zr0.2O2 amount of oxygen vacancies increases with decreasing particle size
[42]. The ratio of Raman band area between 599 and 458 cm1 for ceria-based materials [47,48].
indicates the relative content of oxygen vacancies within the ma-
terials (denoted as A599/A458) [43e45]. In general, the amount of 3.2.5. XPS characterization of 1/10Fe(N)-CuCZ
oxygen vacancies within materials has a significant effect on the Surface chemical composition of 1/10Fe(N)-CuCZ was obtained
performances of catalysts for CO-PROX [46], and the higher number from XPS analysis. For comparison, it is helpful to refer to CuCZ. The
of oxygen vacancies are beneficial to improving the catalytic per- Cu 2p, O 1s and Ce 3d core-level spectra of CuCZ and 1/10Fe(N)-CuCZ
formance [28,40]. As shown in Fig. 5B, the value of A599/A458 in- were used to determine the corresponding species present on their
creases from 0.3 to 0.77 as the ratio of Fe/Cu increases from 0 to 1/ surface, as displayed in Fig. 6. Generally, the signals of Cu 2p core-
10, and the maximum A599/A458 (0.77) is achieved under the Fe/Cu level are complex owing to the simultaneous presence of two
ratio of 1/10. The value of A599/A458 decreases from 0.77 to 0.46 different copper species at least. Representative Cu 2p spectra of 1/
with further increase in the ratio of Fe/Cu to 1/5. Therefore, it can be 10Fe(N)-CuCZ and CuCZ are presented in Fig. 6A. It is well docu-
concluded that doping a small amount of Fe into CuCZ would mented that higher binding energy (933.0e933.8 eV) for the Cu2p3/
promote the formation of oxygen vacancies while the addition of 2 peak and the existence of the shake-up peak are characteristics of
excessive amount of Fe would reduce the oxygen vacancies. CuO, whereas lower binding energy (932.2e933.1 eV) for the Cu
 J. Wang et al. / Journal of Power Sources 343 (2017) 437e445 443

Fig. 6. X-ray photoelectron spectra in the (A) Cu 2p, (B) Ce 3d and (C) O 1s regions of 1/10Fe(N)-CuCZ and CuCZ.

2p3/2 peak together with the absence of the shake-up peak are energy centered around 529.5 eV (OI) and a shoulder peak at
features of Cu2O [49e51]. The main photoelectron peak of Cu2p3/2 approximately 531.6 eV (OII) (Fig. 6C), which have been attributed
(centered around 932.6e993.1 eV) and a weak shake-up peak were to the lattice oxygen and the chemisorbed oxygen on the surface,
observed for both1/10Fe(N)-CuCZ and CuCZ, indicating that Cuþ and respectively [36,55]. The latter was confirmed to be the main active
Cu2þ species coexisted on their surface. Moreover, the existence of (reactive) oxygen species, playing a predominant role in deter-
reduced Cu species can be further confirmed by investigating the mining the catalytic activity of oxidation reactions [56,57]. The
intensity ratio of shake-up peak (satellite) at high binding energy to relative amounts of OII species within CuCZ and 1/10Fe(N)-CuCZ
main photoelectron peak (Isat/Imp). If only CuO is present, then the were calculated and summarized in Table 3. The corresponding
ratio would be 0.55 [52]. As listed in Table 3, both 1/10Fe(N)-CuCZ concentration of OII species in 1/10Fe(N)-CuCZ (43.11%) is higher
and CuCZ exhibit lower values of Isat/Imp than that observed nor- than that in CuCZ (38.01%), which is likely responsible for the
mally for CuO (0.55), suggesting the presence of some Cuþ species improvement of catalytic performance of 1/10Fe(N)-CuCZ for CO-
within them. The formation of reduced Cu species might be PROX.
ascribed to the strong interactions between Ce0.8Zr0.2O2 and active Moreover, the Cu contents on the surface of CuCZ and 1/10Fe(N)-
copper components [24]. The Isat/Imp ratio of 1/10Fe(N)-CuCZ is CuCZ are listed in Table 3. On one hand, 1/10Fe(N)-CuCZ shows
lower than that of CuCZ, indicating the presence of more reduced lower Cu content among the two samples, indicating that the
copper species within 1/10Fe(N)-CuCZ, which would favor the addition of iron facilitates the incorporation of more copper ions
improvement of catalytic activities. into the ceria lattice. On the other hand, the lattice parameter of 1/
Fig. 6B shows Ce 3d core level spectra of CuCZ and 1/10Fe(N)- 10Fe(N)-CuCZ is smaller than that of CuCZ (Table 1), which may be
CuCZ. Minuscule v and u refer to 3d5/2 and 3d3/2 spin-orbit com- the consequence of the incorporation of copper or iron species into
ponents, respectively. The photoelectron peaks marked as v the CeO2 lattice, due to the smaller Cu(II) and Fe (III) ionic radius
(882.4 eV), vˊˊ(888.7 eV), vˊˊˊ (898.3 eV), u (900.9 eV), uˊˊ (907.8 eV) compared to that of Ce(IV). For the 1/5Fe(N)-CuCZ catalyst, Fe3þ ions
and uˊˊˊ (916.7 eV) have been attributed to Ce4þ, and the peaks preferentially entered into the lattices of Ce0.8Zr0.2O2 rather than
labeled as vˊ (ca. 884.5 eV) and uˊ (ca. 903.1 eV) are assigned to the Cu2þ ions. Based on these points, it can be inferred that iron species
Ce3þ species [53]. As can be seen from Fig. 6B, the cerium atom of not only enter into CeO2 lattice, but also the appropriate iron
CuCZ and 1/10Fe(N)-CuCZ was mainly present in the tetravalent content could facilitate the incorporation of Cu species into CeO2.
form (Ce4þ). It has been previously demonstrated that the higher This result is consistent with the results of Raman spectroscopy.
relative quantity of Ce3þ/(Ce4þþCe3þ) would promote oxygen The lattice distortion of CeO2 is caused by the incorporation of
mobility and improve the activity of the catalyst [54]. Furthermore, metal ions and is associated with the number of defects in the
it was also reported that the occurrence of Ce3þ is correlated closely catalyst.
with the formation of oxygen vacancy [53]. The percentage of Ce3þ/
(Ce4þþCe3þ) in CuCZ (listed in Table 3) is lower than that in 1/ 3.3. Effects of H2O, CO2, H2O and CO2 on catalytic activity as well as
10Fe(N)-CuCZ, indicating the formation of more oxygen vacancies catalytic stability of 1/10Fe(N)-CuCZ
within 1/10Fe(N)-CuCZ [54]. In addition, 1/10Fe(N)-CuCZ has a slight
increase in the binding energies of Ce 3d and Cu 2p, compared to It is quite clear that the catalytic activity of 1/10Fe(N)-CuCZ is the
CuCZ, suggesting the existence of stronger interaction between best among all the tested samples. In addition, the comparison of
Ce0.8Zr0.2O2 and active copper species. catalytic performance with various similar catalysts is displayed in
The O 1s spectrum of CuCZ exhibits the main peak of binding Table S1. It seems difficult to make a definite evaluation for catalytic

Table 3
XPS data of CuCZ and 1/10Fe(N)-CuCZ.

Sample Binding energy (eV) Ce3þ 3d5/2 in Ce (%) OII/(OI þ OII) (%) Isat/Imp Cu/(Ce þ Cu) (at.%)

Cu 2p3/2 Ce 3d5/2 OI OII

CuCZ 932.6 882.4 529.5 531.6 12.6 38.01 0.44 32.6

1/10Fe(N)-CuCZ 933.1 882.8 529.6 531.8 13.6 43.11 0.41 26.7
444 J. Wang et al. / Journal of Power Sources 343 (2017) 437e445

performance due to the relatively different test conditions. But it conversion. However, the CO conversion is still more than 94% of
can be found that the 1/10Fe(N)-CuCZ catalyst show a relatively the original conversion of CO even after 80 h time-on-stream test in
better catalytic activity. Both H2O and CO2 are generally coexist the presence of H2O and CO2. Moreover, CO2 selectivity increases
with CO and H2 in the H2-rich streams. Therefore, it is important to slightly somewhat with the reaction time. Subsequently, H2O and
investigate the performances of catalysts for CO-PROX in the CO2 were removed from the reaction system, and 1/10Fe(N)-CuCZ
presence of CO2 and H2O. So, the effects of H2O, CO2, and H2O þ CO2 was further subjected to another 76 h of time-on-stream test. As
on the catalytic performances of 1/10Fe(N)-CuCZ for CO-PROX were can be seen from Fig. 7, once H2O and CO2 were withdrawn from the
investigated, and the corresponding CO conversion and CO2 selec- reaction system, the conversion of CO was restored rapidly to
tivity are displayed in Fig. S5. The CO conversion of 1/10Fe(N)-CuCZ almost 100% and the corresponding CO2 selectivity was also un-
was suppressed to some degree by the addition of H2O, CO2, and expectedly increased from 50% to about 82%. Furthermore, both the
H2O þ CO2 into the feed gases. The inhibitory effects of these gases conversion and selectivity remain relatively stable during the
follow the order: CO2 < CO2 þ H2O < H2O. Similar behaviors have following 76 h of time-on-stream test. The difference in the CO2
been reported by previous studies for other catalysts [19,58]. For selectivity over 1/10Fe(N)-CuCZ before and after the addition of H2O
instance, the full CO conversion (100% CO conversion) was obtained and CO2 into the reaction system might be ascribed to the accu-
with the addition of CO2 (15%) only, whereas the full CO conversion mulation of water on the surface of catalyst [59,60], which would
was not achieved if either H2O (10%) or CO2 (15%) þ H2O (10%) was inhibit the reaction between hydrogen and oxygen. Based on the
added into the reaction system. Moreover, 100% CO conversion was results of time-on-stream experiment, it can be concluded that 1/
achieved at the reaction temperature of 120  C without addition of 10Fe(N)-CuCZ exhibits excellent stability even in the presence of
CO2 and H2O, while the corresponding temperature was 140  C H2O and CO2.
with the addition of CO2. In addition, it was found that the tem-
perature window for high CO conversion becomes narrower. When 4. Conclusions
the reaction temperature is below 100  C or above 180  C, the CO2
selectivity of 1/10Fe(N)-CuCZ under all the four conditions is the In summary, Fe-doped CuO-Ce0.8Zr0.2O2 (CuCZ) catalysts are
same or comparable; when reaction temperature is between 100 synthesized rapidly (within 30 min) by a simple route and evalu-
and 180  C, the corresponding CO2 selectivity is in the following ated for the CO-PROX reaction. The source of iron and the mass ratio
order: with H2O > with CO2 þ H2O > with CO2 > without of Fe/Cu both play important roles in determining the catalytic
CO2 þ H2O. performances of these Fe-doped CuCZ catalysts for CO-PROX. The
As reported in previous literature, the inhibitory effect of CO2 at catalyst doped with iron nitrate (1/10Fe(N)-CuCZ) is superior to the
low temperature has mainly been ascribed to the formation of catalysts doped with iron chloride and iron sulfate (1/10Fe(C)-CuCZ
carbonates on the surface of catalysts as well as the competitive and 1/10Fe(S)-CuCZ). Moreover, the catalytic performance of CuCZ
adsorption between CO2 and CO on the catalyst surface [19,29,58]. catalyst is improved by the addition of Fe, which can be attributed
It has been demonstrated [59] that the negative impact of H2O on to the enhancement of the lattice distortion due to the entry of
CO-PROX is because the adsorbed molecular water blocked the more metal ions into the CeO2 lattice. The optimal Fe/Cu mass ratio
active sites of the catalysts. Moreover, the occurrence of reverse of iron nitrate doped XFe(N)-CuCZ catalysts is found to be 1/10.
reaction of water gas shift (WGS) between CO2 and H2 could pro- According to the results of XRD, H2-TPR, Raman and XPS charac-
duce a small amount of H2O and CO to prevent oxygen from terization, 1/10Fe(N)-CuCZ exhibits the highest catalytic activity due
reacting with carbon monoxide [20]. to the synergistic effect of highly dispersed Cu species on the sur-
The stability of catalysts is an essential criteria for their practical face, formation of more oxygen vacancies and Cuþ species. The
application. In order to investigate the stability of 1/10Fe(N)-CuCZ, influences of H2O, CO2, and their mixture on the conversion of CO
the time-on-stream (228 h) experiment was carried out, as shown over 1/10Fe(N)-CuCZ are in the order of CO2 < CO2 þ H2O < H2O. The
in Fig. 7. There are no remarkable changes in either the CO con- catalyst 1/10Fe(N)-CuCZ exhibits excellent stability during 228 h
version and the CO2 selectivity, after the first 72 h of time-on- time-on-stream test even in the presence of H2O and CO2. More-
stream test in the absence of H2O and CO2. Subsequently, H2O over, the difference in the CO2 selectivity before and after the
and CO2 were added into the reaction system simultaneously, and addition of H2O and CO2 into the feed gas is likely due to the
the second 80 h of time-on-stream test was performed over 1/ accumulation of water on the surface of 1/10Fe(N)-CuCZ.
10Fe(N)-CuCZ continuously, which resulted in a decrease in the CO
Acknowledgements

We gratefully acknowledge the financial supports of the Na-

tional Natural Science Foundation of China (Grant Nos. U1402233,
21267011 and 21367015).

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://

dx.doi.org/10.1016/j.jpowsour.2017.01.084.

References

[1] S.G. Ruz, University of Notre Dame, Notre Dame, Indiana, 2007.
[2] A. Horne s, A.B. Hungría, P. Bera, A. Lo  pez-Camara, M. Fern andez-García,
A. Martínez-Arias, L. Barrio, M. Estrella, G. Zhou, J.J. Fonseca, J.C. Hanson,
J.A. Rodriguez, J. Am. Chem. Soc. 132 (2010) 34e35.
Fig. 7. Catalytic stability test of 1/10Fe(N)-CuCZ in H2-rich streams (a) without CO2 and [3] E.C. Njagi, H.C. Genuino, C.K. King’ondu, C.H. Chen, D. Horvath, S.L. Suib, Int. J.
H2O at 120  C, (b) with 15% CO2 and 10% H2O at 155  C and (c) without 15% CO2 and Hydrogen energy 36 (2011) 6768e6779.
10% H2O at 120  C. GHSV ¼ 24000 h1. [4] S. Zeng, L. Zhang, N. Jiang, M. Gao, X. Zhao, Y. Yin, H. Su, J. Power Sources 293
 J. Wang et al. / Journal of Power Sources 343 (2017) 437e445 445

(2015) 1016e1023. 121e131.

[5] S. Zeng, Y. Wang, S. Ding, J.J.H.B. Sattler, E. Borodina, L. Zhang, [33] M.F. Luo, J.M. Ma, J.Q. Lu, Y.P. Song, Y.J. Wang, J. Catal. 246 (2007) 52e59.
B.M. Weckhuysen, H. Su, J. Power Sources 256 (2014) 301e311. [34] J.L. Cao, Q.F. Deng, Z.Y. Yuan, J. Mater. Sci. 44 (2009) 6663e6669.
[6] M.M. Schubert, A. Venugopal, M.J. Kahlich, V. Plzak, R.J. Behm, J. Catal. 222 [35] W.P. Dow, Y.P. Wang, T.J. Huang, J. Catal. 160 (1996) 155e170.
(2004) 32e40. [36] J.L. Cao, Y. Wang, T.Y. Zhang, S.H. Wu, Z.Y. Yuan, Appl. Catal. B Environ. 78
[7] Z. Ma, S. Brown, S.H. Overbury, S. Dai, Appl. Catal. A Gen. 327 (2007) 226e237. (2008) 120e128.
[8] D. Widmann, R.J. Behm, Angew. Chem. Int. Ed. 50 (2011) 1e6. [37] J.L. Ayastuy, N.K. Gamboa, M.P. Gonz alez-Marcos, M.A. Gutie rrez-Ortiz, Chem.
[9] Y. Liu, B. Liu, Q. Wang, C. Li, W. Hua, Y. Liu, P. Jing, W. Zhao, J. Zhang, J. Catal. Eng. J. 171 (2011) 224e231.
296 (2012) 65e76. [38] H. Bao, X. Chen, J. Fang, Z. Jiang, W. Huang, Catal. Lett. 125 (2008) 160e167.
[10] T. Bao, Z. Zhao, Y. Dai, X. Lin, R. Jin, G. Wang, T. Muhammad, Appl. Catal. B [39] X. Li, X.Y. Quek, D.A.J.M. Ligthart, M. Guo, Y. Zhang, C. Li, Q. Yang,
Environ. 119e120 (2012) 62e73. E.J.M. Hensen, Appl. Catal. B Environ. 123e124 (2012) 424e432.
[11] P. Gawade, B. Bayram, A.M.C. Alexander, U.S. Ozkan, Appl. Catal. B Environ. [40] O.H. Laguna, M.A. Centeno, M. Boutonnet, J.A. Odriozola, Appl. Catal. B Envi-
128 (2012) 21e30. ron. 106 (2011) 621e629.
[12] E. Moretti, M. Lenarda, L. Storaro, A. Talon, T. Montanari, G. Busca, [41] Z.Y. Pu, X.S. Liu, A.P. Jia, Y.L. Xie, J.Q. Lu, M.F. Luo, J. Phys. Chem. C 112 (2008)
E. Rodríguez-Castello n, A. Jime
nez-Lopez, M. Turco, G. Bagnasco, R. Frattini, 15045e15051.
Appl. Catal. A Gen. 335 (2008) 46e55. [42] J.R. McBride, K.C. Hass, B.D. Poindexter, W.H. Weber, J. Appl. Phys. 76 (1994)
[13] I. Moreno, N. Navascues, S. Irusta, J. Santamaria, J. Catal. 329 (2015) 479e489. 2435e2441.
[14] D. Gamarra, A.L. Ca mara, M. Monte, S.B. Rasmussen, L.E. Chinchilla, [43] Q. Liang, X. Wu, D. Weng, H. Xu, Catal. Today 139 (2008) 113e118.
A.B. Hungría, G. Munuera, N. Gyorffy, Z. Schay, V.C. Corber an, J.C. Conesa, [44] W.Y. Hernandez, M.A. Centeno, F. Romero-Sarria, J.A. Odriozola, J. Phys. Chem.
A. Martínez-Arias, Appl. Catal. B Environ. 18e131 (2013) 224e238. C 113 (2009) 5629e5635.
[15] J. Li, Y. Han, Y. Zhu, R. Zhou, Appl. Catal. B Environ. 108e109 (2011) 72e80. [45] O.H. Laguna, F. Romero Sarria, M.A. Centeno, J.A. Odriozola, J. Catal. 276 (2010)
[16] D. Gamarra, G. Munuera, A.B. Hungría, M. Fern_andez-García, J.C. Conesa, 360e370.
P.A. Midgley, X.Q. Wang, J.C. Hanson, J.A. Rodríguez, A. Martínez-Arias, J. Phys. [46] W.P. Dow, Y.P. Wang, T.J. Huang, J. Catal. 160 (1996) 171e182.
Chem. C 111 (2007) 11026e11038. [47] G.W. Graham, W.H. Weber, C.R. Peters, R. Usmen, J. Catal. 130 (1991)
[17] Y.Z. Chen, B.J. Liaw, H.C. Chen, Int. J. Hydrogen Energy 31 (2006) 427e435. 310e313.
[18] G. Avgouropoulos, T. Ioannides, Appl. Catal. B Environ. 67 (2006) 1e11. [48] S. Deshpande, S. Patil, S.V. Kuchibhatla, S. Seal, Appl. Phys. Lett. 87 (2005)
[19] D. Gamarra, A. Martínez-Arias, J. Catal. 263 (2009) 189e195. 133113e133115.
[20] J. Li, P.F. Zhu, R.X. Zhou, J. Power Sources 196 (2011) 9590e9598. [49] E. Moretti, L. Storaro, A. Talon, M. Lenarda, P. Riello, R. Frattini, M.D.M. deYuso,
[21] A.A. Firsova, A.N. Il’ichev, T.I. Khomenko, L.V. Gorobinskii, Yu. V. Maksimov, A. Jimenez-Lo  pez, E. Rodríguez-Castello n, F. Ternero, A. Caballero, J.P. Holgado,
I.P. Suzdalev, V.N. Korchak, Kinet. Catal. 48 (2007) 282e291. Appl. Catal. B Environ. 102 (2011) 627e637.
[22] K. Sirichaiprasert, A. Luengnaruemitchai, S. Pongstabodee, Int. J. Hydrogen [50] L. Kundakovic, M.F. Stephanopoulos, Appl. Catal. A Gen. 171 (1998) 13e29.
Energy 32 (2007) 915e926. [51] W. Liu, M.F. Stephanopoulos, J. Catal. 153 (1995) 317e332.
[23] A. Go  mez-Corte s, Y. M arquez, J. Arenas-Alatorre, G. Díaz, Catal. Today [52] E. Moretti, L. Storaro, A. Talon, P. Riello, A.I. Molina, E. Rodríguez-Castello  n,
133e135 (2008) 743e749. Appl. Catal. B Environ. 168 (2015) 385e395.
[24] G. Avgouropoulos, T. Ioannides, H. Matralis, Appl. Catal. B Environ. 56 (2005) [53] J. Fan, X. Wu, X. Wu, Q. Liang, R. Ran, D. Weng, Appl. Catal. B Environ. 81
87e93. (2008) 38e48.
[25] Z.G. Liu, R.X. Zhou, X.M. Zheng, J. Nat, Gas. Chem. 17 (2008) 125e129. [54] L.F. Liotta, D.G. Carlo, G. Pantaleo, A.M. Venezia, G. Deganello, Appl. Catal. B
[26] J. Wang, L. Deng, D. He, J. Lu, S. He, S. He, Y. Luo, Int. J. Hydrogen Energy 40 Environ. 66 (2006) 217e227.
(2015) 12478e12488. [55] A. Platau, L.I. Johansson, A.L. Hagstrom, S.E. Karlsson, S.B.M. Hagstrom, Surf.
[27] T. Alammar, H. Noei, Y. Wang, W. Grünert, A.V. Mudring, ACS Sustain. Chem. Sci. 63 (1977) 153e161.
Eng. 3 (2015) 42e54. [56] H. Wang, J. Wang, Z. Wu, Y. Liu, Catal. Lett. 134 (2010) 295e302.
[28] J. Li, P. Zhu, S. Zuo, Q. Huang, R. Zhou, Appl. Catal. A Gen. 381 (2010) 261e266. [57] D. Shang, W. Cai, W. Zhao, Y. Bu, Q. Zhong, Catal. Lett. 144 (2014) 538e544.
[29] X. Du, Z. Yuan, L. Cao, C. Zhang, S. Wang, Fuel Process. Technol. 89 (2008) [58] J.W. Park, J.H. Jeong, W.L. Yoon, Y.W. Rhee, J. Power Sources 132 (2004)
131e138. 18e28.
[30] J.L. Cao, Y. Wang, X.L. Yu, S.R. Wang, S.H. Wu, Z.Y. Yuan, Appl. Catal. B Environ. [59] Z. Wu, H. Zhu, Z. Qin, H. Wang, L. Huang, J. Wang, Appl. Catal. B Environ. 98
79 (2008) 26e34. (2010) 204e212.
[31] C.R. Jung, A. Kundu, S.W. Nam, H.I. Lee, Appl. Catal. A Gen. 331 (2007) [60] C.R. Jung, J. Han, S.W. Nam, T.H. Lim, S.A. Hong, H.I. Lee, Catal. Today 93e95
112e120. (2004) 183e190.
[32] M.F. Luo, Y.J. Zhong, X.X. Yuan, X.M. Zheng, Appl. Catal. A Gen. 162 (1997)