Low-temperature PROX (preferential oxidation) on novel CeO 2-supported Cu-cluster catalysts under fuel-cell operating conditions

Article abstract of DOI:10.1039/b709176a

Cu¹⁺-clusters on a CeO₂ support, which were prepared by hydrothermal synthesis using cetyltrimethylammonium bromide (CTAB), were found to be highly active and selective for preferential oxidation (PROX) of CO in excess H₂ with H₂O and CO₂ under practical fuel-cell operating conditions. The Royal Society of Chemistry.

Full text of DOI:10.1039/b709176a

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Low-temperature PROX (preferential oxidation) on novel CeO2-
supported Cu-cluster catalysts under fuel-cell operating conditions{
Mizuki Tada,a Rajaram Bal,a Xindong Mu,a Rudy Coquet,a Seitaro Nambab and Yasuhiro Iwasawa*a
Received (in Cambridge, UK) 18th June 2007, Accepted 20th August 2007
First published as an Advance Article on the web 31st August 2007
DOI: 10.1039/b709176a
Cu1+-clusters on a CeO2 support, which were prepared by
washed with distilled H2O and EtOH, and dried at ambient
temperature for 10 h and then at 100 uC for 8 h, followed by
heating at 500 uC for 6 h under a He flow (30 ml min21). The Cu
content of the final catalyst was determined by XRF. The metallic
Cu clusters on MoO2 were completely inactive for CO PROX at
90 uC as shown in Table 1. Metallic Cu clusters on ZnO (Cu/Zn–
CTAB) and SiO2 (Cu/Si–CTAB) were also inactive (Table 1).
Similarly prepared Cu/Zr–CTAB, Cu/Fe–CTAB, and Cu/Al–
CTAB catalysts were also inactive. It was, however, found that the
hydrothermal synthesis using cetyltrimethylammonium bro-
mide (CTAB), were found to be highly active and selective for
preferential oxidation (PROX) of CO in excess H2 with H2O
and CO2 under practical fuel-cell operating conditions.
Hydrogen as the most efficient and cleanest energy source for fuel-
cell power is produced by partial oxidation followed by water gas-
shift reaction and reforming of hydrocarbons or methanol.1
A
small amount of CO (0.3–1%) present in the produced H2 must be
selectively removed because CO is highly poisonous to electro-
catalysts in proton-exchange-membrane (PEM) fuel cells.2 PROX
of CO in excess H2 is a key reaction in the practical use of H2 in
PEM fuel-cell systems.
Table 1 Catalytic performances of Cu and Ce catalysts for the PROX
of CO in excess H2 at 90 uCa
Specific
rate/mmolCO
Cu
wt%
CO
conv.%
O2
selec.%
Catalyst
gCu s21
21
The most important requirement for PROX catalysts is high
CO oxidation activity accompanied by prohibition of undesirable
H2 oxidation. Various precious metal catalysts such as Pt,3,4
Au,4a,5–7 Rh,4b Ru,4b,8 Pt–Sn,4c,d Pt–Fe,4e etc. with good PROX
activities have been utilized as candidates for PROX catalysts.
However, there are few non-precious metal catalysts which possess
sufficient PROX activities at low temperatures. CuO/CeO2 was
reported to be an active catalyst at temperatures .150 uC,4a,9 but
non-precious metals rarely show good CO conversions at lower
temperatures, ¡120 uC, under the practical PEM fuel-cell
operating conditions. We have succeeded in preparing novel
CeO2-supported Cu1+-cluster catalysts which are highly active and
selective for the PROX reaction.
CeO2
Ce2O3
Ce–CTAB
0
0
0
100
10
10
10
0
0
0
—
—
—
—
100
100
100
—
0
0
0
0
1.1
1.6
0.9
0
Cu–CTAB
0
Cu/Ce2O3 (impreg.)b
Cu/Ce2O3 (impreg.)b,c
Cu/CeO2 (impreg.)b
Cu/CeO2 (post-impreg.
with CTAB)b
8.9
12.8
7.2
0
7.5
CuBr/CeO2 (impreg.)
CuBr2/CeO2 (impreg.)
Cu–Ce (co-precip.)d
Cu–Ce (co-precip.)c,d
Cu/Ce–noCTABe
Cu/Ce–noCTABc,e
Cu/Ce–CTAB
7.5
7.5
10
10
9.6
9.6
7.5
7.5
4.5
0
0
—
—
0
0
5.6
7.3
10.2
15.9
91.9
72.1
32.0
53.9
16.5
14.8
58.4
4.2
0
0
0
0
0
0
0
0
100
100
100
100
99.8
100
100
100
100
100
100
100
—
—
—
—
—
—
—
—
0.7
0.9
1.3
2.1
15.2
23.8
17.6
8.9
5.4
4.9
19.3
1.4
0
0
0
0
0
0
0
0
Cu/Ce–CTABf
The supported Cu/CeO2 catalyst (denoted as Cu/Ce–CTAB)
was hydrothermally prepared using Ce and Cu nitrates as
precursors with a surfactant CTAB in a manner similar to that
we reported for preparation of metallic Cu clusters on MoO2
(Cu/Mo–CTAB).10 In a typical synthesis method, 5 g of
Ce(NO3)3?6H2O were dissolved in 8 ml of hot distilled water, to
which 0.61 g of Cu(NO3)2?3H2O in H2O (2 ml) was added
dropwise. Then, 0.54 g of CTAB was dissolved in a mixture of
H2O (5 ml) and ethanol (2 ml), and the obtained solution was
added to the Cu + Ce solution. A typical molar composition is Cu/
CTAB/H2O = 1.0/0.55/325. The homogeneous slurry mixture was
hydrothermally treated at 175 uC for 24 h in a Teflon-lined
autoclave vessel under an autogeneous pressure. The product was
Cu/Ce–CTABf
Cu/Ce–CTABf
15
Pt–Cu/Ce–CTABf,g
Pd–Cu/Ce–CTABf,g
Au–Cu/Ce–CTABf,g
In–Cu/Ce–CTABf,g
Cu/Ce–Pluorunic1h
Cu/Ce–dodecyl sulfatei
Cu/Zr–CTABj
7.5
7.5
7.5
7.5
7.5
7.5
10
10
7.8
Cu/Fe–CTABk
Cu/Mo–CTABl
Cu/Zn–CTABm
Cu/Al–CTABn
3.7
5.3
8.9
Cu/Si–CTABo
a
Catalyst: 0.4 g, time on stream: 2 h, W/F = 2.24 gcat h mol21, CO/
b
O2/H2/He
=
1/1/50/48 (mol%).
Impregnation of Cu nitrate.
d
c
Reduced with H2 (15% in He) at 500 uC. Co-precipitation.
Hydrothermal synthesis without CTAB. Catalyst: 0.2 g, W/F =
aDepartment of Chemistry, Graduate School of Science, The University
of Tokyo, Hongo, Bunkyo-ku, Tokyo 113-0033, Japan.
E-mail: iwasawa@chem.s.u-tokyo.ac.jp; Fax: +81 3-5800-6892;
Tel: +81 3-5841-4363
e
f
1.12 gcat h mol21, 80 uC. An additional metal was impregnated on
Cu/Ce–CTAB.
g
Pluorunic1 was used as
a neutral surfactant.
j
Dodecyl sulfate was used as an anionic surfactant. Prepared
h
i
bDepartment of Materials, Teikyo University of Science & Technology,
Uenohara, Yamanashi 409-0193, Japan
k
l
with Zr nitrate.
(NH4)6Mo7O24.
Al2(SO4)3. Prepared with Si(OC2H5)4.
Prepared with Fe nitrate. Prepared with
Prepared with Zn nitrate.
m
n
Prepared with
o
{ Electronic supplementary information (ESI) available: Experimental
details. See DOI: 10.1039/b709176a
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 4689–4691 | 4689

View Article Online
the metal oxides: Cu oxide (most reducible) , Mo oxide , Zn
oxide , Si oxide , Al oxide y Zr oxide y Ce oxide (hard to
reduce). Further, chemical interaction of the Cu1+ clusters with
CeO2 surfaces may also be the key to stabilize the clusters on the
support.
Table 1 shows the catalytic performances of various Cu and Ce
catalysts for CO PROX reactions in excess H2 at 90 uC. Ce oxides,
Ce–CTAB, and Cu–CTAB were completely inactive for CO
oxidation at 90 uC. On the other hand, the hydrothermally-
prepared Cu/Ce–CTAB catalyst (7.5 Cu wt%) exhibited good
catalytic performances for the CO PROX with 91.9–96.1% CO
conversion and 99.4–99.8% O2 selectivity at 90 uC in a feed of CO/
O2/H2 = 1/1/50 (Table 1). The performances of the Cu/Ce–CTAB
catalyst under the various reaction conditions at different W/F,
reaction temperatures, and feed compositions are summarized in
Table 2. It is to be noted that high CO conversions and O2
selectivities were also achieved in the reactant feeds containing
substantial amounts of H2O and CO2. The CO conversions and O2
selectivities at W/F = 2.24 gcat h mol21 and 90 uC were 85.7% and
98.7%, respectively when H2O (10%) existed, and 81.4% and
98.2%, respectively when H2O (10%) and CO2 (20%) co-existed.
The CO conversion increased significantly with an increase in
W/F, while the O2 selectivity decreased a little from 100%. The
catalytic performances of the Cu/Ce–CTAB catalyst (Cu 7.5 wt%)
Fig. 1 SEM photographs of Cu/Ce catalysts. (a) Fresh Cu/Ce–CTAB
(7.5 Cu wt%), (b) after the PROX reaction in the presence of H2O, (c)
impregnated Cu/Ce2O3 (10 Cu wt%).
were examined more systematically at W/F = 1.12 gcat h mol21
.
new Cu/Ce–CTAB catalyst exhibited high activities with the feed
of CO/O2/H2/He = 1/1/50/48 (mol%) (Table 1), which are
contrasted to the much lower performances of conventional
impregnated Cu/Ce-oxide catalysts and co-precipitated Cu–Ce
catalysts as shown in Table 1.
The conversion increased with increasing temperature from 80 to
120 uC, while keeping good selectivities (96.9–100%); for example,
the conversions (selectivities) at 90 uC were 82.5% (100%), 75.3%
(100%), and 62.9% (100%) for the feeds of CO/O2/H2 (1/1/50), CO/
O2/H2/H2O (1/1/50/10), and CO/O2/H2/H2O/CO2 (1/1/50/10/20),
respectively, and at 120 uC the performances increased respectively
to 100% (99.2%), 100% (97.1%), and 97.3% (96.9%) (Table 2). Au/
Fe2O3 catalysts are highly active for CO oxidation at 80 uC, but O2
selectivity is as low as 51%5a (,60%)4a,5c in the presence of H2O
and CO2. A conventional CuO/CeO2 catalyst is more selective, but
The Cu/Ce–CTAB catalyst hydrothermally prepared in the
presence of the CTAB surfactant showed a single-phase morphol-
ogy reflecting an assembly effect of the surfactant as imaged by
SEM in Fig. 1(a), which was completely different from the
inhomogeneous disordered morphology of a conventional Cu/Ce-
oxide catalyst as shown in Fig. 1(c). The SEM image did not
change after the PROX reaction (Fig. 1(b)), which indicates a high
stability under the PROX conditions.
high temperature (.150 uC) is required for sufficient CO
exhibits average perfor-
conversion (90%).4a,9a,9b Pt/Al2O3
3b,3d,4a
mances. The Cu/Ce–CTAB catalyst is the first example with
remarkable performances in both activity and selectivity for the
CO PROX with H2O and CO2 at low temperatures (¡120 uC).
XRD showed the formation of fluorite CeO2, while no XRD
peaks attributed to Cu metal or oxides were observed (ESI{ 1),
indicating that Cu species were dispersed as small units on the
CeO2 surfaces. XPS Cu 2p3/2 binding energy of 932.4 eV (referred
to C 1s of 284.6 eV) suggested that the Cu2+ precursor was reduced
to Cu0 or Cu1+ under the hydrothermal conditions. The reduction
of the Cu2+-nitrate precursor similarly occurred on other oxide
surfaces such as MoO2, ZnO, and SiO2 reported previously.10 In
contrast to these previous supports, however, the Cu2+ precursor
along with the Ce-nitrate precursor was not fully reduced to
metallic Cu0 under the hydrothermal synthesis conditions. The Cu
atoms of Cu/Ce–CTAB were bonded to oxygen atoms making
Cu–O at 0.194 nm besides Cu–Cu bonding at 0.256 nm as proved
by Cu K-edge EXAFS (ESI{ 2). The XRD, XPS, and XAFS data
reveal the formation of Cu1+ clusters rather than Cu0 metallic
clusters on CeO2 prepared from the Ce-nitrate precursor.
Table 2 PROX performances (conv.% and O2 selec.%) of the Cu/Ce–
CTAB catalyst (Cu 7.5 wt%) at different W/F (gcat h mol21), reaction
temperature, and reactant composition
CO/O2/H2
CO/O2 (1/1/50)
Temp./ (1/1) conv.
conv. (selec.)
CO/O2/H2/H2O CO/O2/H2/H2O/
CO2 (1/1/50/10/20)
conv. (selec.)
(1/1/50/10)
conv. (selec.)
W/F uC
0.56 80
90
100
1.12 80
90
100
76.8 64.5 (100) 54.1 (100)
82.4 71.9 (100) 66.7 (100)
94.0 87.5 (100) 82.3 (100)
83.4 72.1 (100) 64.8 (100)
91.2 82.5 (100) 75.3 (100)
99.2 93.2 (100) 88.2 (100)
55.3 (100)
62.9 (100)
83.9 (99.4)
89.3 (98.9)
97.3 (96.9)
110
120
100
100
96.3 (100) 93.8 (99.2)
100 (99.2) 100 (97.1)
2.24 80
90
100
3.36 80
90
100
87.9 75.8 (100) 64.9 (100)
97.8 91.9 (99.8) 85.7 (98.7)
The detailed mechanism for the formation of reduced Cu1+
species under the hydrothermal synthesis conditions in the
presence of CTAB without any additional reducing reagent is
not clear at present, but the degree of the reduction of the Cu-
precursor and oxide-precursor may depend on the oxophilicity of
81.4 (98.2)
100
98.5 (99.8) 91.5 (96.5)
89.2 77.9 (99.5) 67.2 (98.5)
100
100
96.1 (99.4) 88.4 (95.1)
100 (98.2) 93.7 (90.2)
4690 | Chem. Commun., 2007, 4689–4691
This journal is ß The Royal Society of Chemistry 2007

View Article Online
Cu/Ce–noCTAB hydrothermally prepared in the absence of
CTAB was much less active (10.2–15.9% CO conv.), similar to
conventional impregnated and co-precipitated catalysts as shown
in Table 1. The reduction of these catalysts with hydrogen did not
efficiently improve their catalytic activities. The high PROX
activity was characteristic of the hydrothermal synthesis, which
in-situ produced reductive Cu species dispersed on CeO2. The
conventional impregnated Cu/CeO2 catalyst was post-impregnated
with CTAB, but no activity appeared at 90 uC as shown in Table 1.
Other surfactants, neutral Pluorunic1 and anionic dodecyl sulfate,
did not produce any catalytically active Cu species by the
hydrothermal procedure (Table 1). Additional metals (Pt, Pd,
Au, or In) supported on the Cu/Ce–CTAB catalyst inhibited the
PROX activity more or less as shown in Table 1. Only the Cu/Ce–
CTAB among the examined catalysts exhibited such good PROX
performances at ¡120 uC. The catalytic performances were
maintained for at least 10 h in the presence of H2O and CO2.
Active Cu species in the Cu/Ce–CTAB catalyst could not be
imaged by TEM, while the EXAFS analysis (ESI{ 2) provided
structural parameters (bond distance and coordination number)
for Cu–O and Cu–Cu. The small coordination number (0.9) of the
Cu–Cu bond together with the small coordination number (2.4) of
the Cu–O bond indicates that the hydrothermal synthesis
However, impregnated CuBr as well as CuBr2 on CeO2 had no
PROX activity (Table 1). The Cu(I)Br species were also observed
after the PROX reaction in excess H2 at 90 uC as shown in ESI{
3-II (b). When the Cu/Ce–CTAB catalyst was exposed to O2 at
90 uC, the three patterns of CuBr completely disappeared (ESI{
3-II (d)) and they were reversibly recovered by subsequent reaction
with CO at 90 uC (ESI{ 3-II (e)). There may be a positive
effect of the remaining Cu bromide species on the PROX catalysis
of Cu/Ce–CTAB.
In conclusion, we successfully prepared Cu1+-oxide clusters on
CeO2 by the hydrothermal synthesis method using a surfactant,
CTAB. These small Cu clusters on CeO2 exhibited good CO
PROX performances under the PEM fuel-cell operating condi-
tions. This is the first report of a non-precious metal catalyst with
good performances for the CO PROX in the presence of H2O and
CO2. The PROX performances of the Cu1+-oxide clusters may be
promoted by traces of bromides. We also applied LDI-MASS to
characterization of the composition of supported metal species for
the first time.
Notes and references
1 J. R. Rostrup-Nielson and T. Rostrup-Nielson, Cattech, 2002, 6, 150.
2 (a) R. A. Lemons, J. Power Sources, 1990, 29, 251; (b) H. Igarashi,
T. Fujino and M. Watanabe, J. Electroanal. Chem., 1995, 391, 119; (c)
A. F. Ghenciu, Curr. Opin. Solid State Mater. Sci., 2002, 6, 389.
3 (a) M. J. Kahlich, A. Gasteiger and R. J. Behm, J. Catal., 1997, 171, 93;
(b) O. Korotkikh and R. Farrauto, Catal. Today, 2000, 62, 249; (c)
D. H. Kim and M. S. Lim, Appl. Catal., A, 2002, 224, 27; (d)
A. Manaslip and E. Gulari, Appl. Catal., B, 2002, 37, 17; (e) A. Wootsch,
C. Descorme and D. Duprez, J. Catal., 2004, 225, 259; (f) A. Fukuoka
and M. Ichikawa, Top. Catal., 2006, 40, 103.
4 (a) G. Avgouropoulos, T. Ioannides, C. Papadopoulou, J. Batista,
S. Hocevar and H. Matralis, Catal. Today, 2002, 75, 157; (b)
M. M. Schubert, M. J. Kahlich, H. A. Gasteiger and R. J. Behm,
J. Power Sources, 1999, 84, 175; (c) S. H. Oh and R. M. Sinkevitch,
J. Catal., 1993, 142, 254; (d) S. Ozkara and A. E. Aksoylu, Appl. Catal.,
A, 2003, 251, 75; (e) M. M. Shubert, M. J. Kahlich, G. Feldmeyer,
M. Huttner, S. Hackensberg, H. A. Gasteiger and R. J. Behm, Phys.
Chem. Chem. Phys., 2001, 3, 1123; (f) M. Kotobuki, A. Watanabe,
H. Uchida, H. Yamashita and M. Watanabe, J. Catal., 2005, 236, 262.
5 (a) P. Landon, J. Ferguson, B. E. Solsona, T. Garcia, A. F. Carley,
A. A. Herzing, C. J. Kiely, S. E. Golunski and G. J. Hutchings, Chem.
Commun., 2005, 3385; (b) B. T. Qiao and Y. Q. Deng, Chem. Commun.,
2003, 2192; (c) M. J. Kahlich, A. Gasteiger and R. J. Berm, J. Catal.,
1999, 182, 430; (d) M. M. Schubert, A. Venugopal, M. J. Kahlich,
V. Plzak and R. J. Berm, J. Catal., 2004, 222, 32; (e) B. T. Qiao and
Y. Q. Deng, Chem. Commun., 2003, 2192; (f) B. Grigorova, J. Mellor,
A. Palazov and F. Greyling, W. O. Pat. 00/59631/2000.
6 (a) S. Carrettin, P. Conception, A. Corma, J. M. L. Nieto and
V. F. Puntes, Angew. Chem., Int. Ed., 2004, 43, 2538; (b) G. Panzera,
V. Modafferi, S. Candamano, A. Donato, F. Frusteri and
P. L. Antonucci, J. Power Sources, 2004, 135, 177; (c)
A. Luengnaruemitchai, S. Osuwan and E. Gulari, Int. J. Hydrogen
Energy, 2004, 29, 429; (d) W. Deng, J. De Jesus, H. Saltsburg and
M. F. Stephanopoulos, Appl. Catal., A, 2005, 291, 126.
7 (a) G. K. Bethke and H. H. Kung, Appl. Catal., A, 2000, 194, 43; (b)
R. J. H. Grisel and B. E. Nieuwenhuys, J. Catal., 2001, 199, 48.
8 Y. F. Han, M. J. Kahlich, M. Kinne and R. J. Behm, Phys. Chem.
Chem. Phys., 2002, 4, 389.
prohibited the growth of Cu species and produced small Cu1+
-
oxide clusters, which did not significantly change in their sizes after
the PROX reaction (ESI{ 2). CO of 5.75 6 1024 mol was
adsorbed on 1 g of Cu/Ce–CTAB (0.49 CO/Cu), but no CO2
formation was observed. The results indicate that neither water-gas
shift reaction nor CO oxidation with lattice oxygen proceeded
on the Cu/Ce–CTAB catalyst. On the other hand, O2 of 2.40 6
1024 mol was adsorbed on 1 g of the fresh Cu/Ce–CTAB catalyst
(0.20 O2/Cu) and the stoichiometric amount of CO2 (0.39 CO2 per
Cu) was produced when this surface was subsequently exposed to
CO, which suggests high oxidation activity of the Cu1+-oxide
cluster species on the CeO2 surface. XRF analysis showed that a
small amount of Br (Br/Cu atomic ratio less than 0.26), derived
from the surfactant CTAB, remained on the hydrothermally
prepared Cu/Ce–CTAB catalysts. However, no Cu–Br contribu-
tion was observed by Cu K-edge EXAFS.
LDI-MASS (laser-desorbed-ionization mass spectroscopy)
detected typical mass signals of Cu(I)Br as shown in ESI{ 3. We
measured LDI-MASS using MALDI-MASS equipment for
characterization of the supported metal species on solid catalyst
surfaces for the first time. The mass signals of Cu clusters were
detected in a negative reflectron mode without any matrices for
ionization. A UV/VIS spectrum of the Cu/Ce–CTAB catalyst
suggested that the N2 laser of the MALDI-MASS apparatus
(337 nm) excited a band around 310 nm of the Cu catalyst. Three
distinct components were detected for the hydrothermally-
prepared Cu/Ce–CTAB (7.5 Cu wt%) by LDI-MASS (ESI{ 3-I
(g)), whose mass numbers and patterns implied the atomic
compositions of (CuBr)Br2, (CuBr)2Br2, and (CuBr)3Br2. These
mass numbers were nominally identical masses to the Cu-oxide
clusters of Cu3O22, Cu5O32, and Cu7O42, but all the intensity
patterns of isotopes were fitted to theoretical patterns of the Cu-
bromide anions rather than the Cu-oxide cluster anions. Cu(I)Br
gave similar mass patterns with the three components, while those
of Cu(II)Br2 were entirely different from the observed ones.
9 (a) Y. Liu, L. Q. Fu and M. F. Stephanopoulos, Catal. Today, 2004, 93–
95, 241; (b) G. Avgouropoulos, T. Ioannides, H. K. Matralis, J. Batista
and S. Hocevar, Catal. Lett., 2001, 73, 33; (c) M. F. Luo, J. M. Ma,
J. Q. Lu, Y. P. Song and Y. J. Wang, J. Catal., 2007, 246, 52.
10 (a) R. Bal, M. Tada and Y. Iwasawa, Chem. Commun., 2005, 3433; (b)
M. Tada, R. Bal, S. Namba and Y. Iwasawa, Appl. Catal., A, 2006, 307,
78.
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 4689–4691 | 4691