Alumina-supported LaCoO3 perovskite for selective CO oxidation (SELOX)

ww.sciencedirect.com

i n t e r n a t i o n a l j o u r n a l o f h yd r o g e n e n e r g y x x x ( 2 0 1 2 ) 1e1 0

Available online at w

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

Alumina-supported LaCoO3 perovskite for selective COoxidation (SELOX)

Carlos Alberto Chagas, Fabio Souza Toniolo, Robert Newton S.H. Magalhaes,Martin Schmal*

Federal University of Rio de Janeiro, Chemical Engineering Program, NUCAT/PEQ/COPPE e Centro de Tecnologia, Bl. G 128, C.P. 68502,

CEP. 21941-914 Rio de Janeiro, Brazil

a r t i c l e i n f o

Article history:

Received 14 September 2011

Received in revised form

8 December 2011

Accepted 9 December 2011

Available online xxx

Keywords:

SELOX

Perovskites

Supported perovskites

Purification

Hydrogen

* Corresponding author.E-mail address: [email protected]

Please cite this article in press as: ChagasInternational Journal of Hydrogen Energy

0360-3199/$ e see front matter Copyright ªdoi:10.1016/j.ijhydene.2011.12.052

a b s t r a c t

Perovskite-type LaCoO3 oxide was prepared using Pechini’s method and supported on

alumina using a physical mixture and thermal treatment in order to obtain catalysts with

different perovskite loadings (10, 20 and 40 wt.%). The catalysts were characterized by

different methods and their catalytic potential was tested in the selective CO oxidation

reaction (SELOX). Characterizations indicated that structural properties of LaCoO3 did not

change after supporting on the alumina and that the perovskite structure is resistant to

reduction in the temperature range of SELOX reaction. The most active catalysts were the

supported 40% LaCoO3/Al2O3 and LaCoO3. The supported catalyst presented ca. twice as

many metallic surface area than the unsupported sample, which suggests also twice as

many perovskite at the surface and therefore higher activity. These results evidence that

the supported perovskite oxides are very good alternative for SELOX reaction when

compared to noble metal supported catalyst.

Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights

reserved.

1. Introduction in the outlet streammay be low but still unsuitable for PEMFC

The research for new eco-friendly technologies able to provide

energy more efficiently has led the scientific community to

develop a series of novel apparatus, catalysts and processes.

The proton exchange membrane fuel cell (PEMFC) belongs to

this promising generation of technologies and presents a great

potential for transport and portable applications operating at

low temperature and pressure ranges. Hydrogen is the fuel for

PEMFC which requires a feed stream with restrictive CO

content (<10 ppm) in order to avoid irreversible deactivation

of the platinum-based electrodes. Such a feed stream usually

comes from reforming of hydrocarbons, followed by pro-

cessing in water gas shift reaction (WGSR), whose CO content

r (M. Schmal).

CA, et al., Alumina-sup(2012), doi:10.1016/j.ijh

2011, Hydrogen Energy P

applications. In this sense selective CO oxidation (SELOX)

arises as an interesting and economic approach to remove CO

from H2-rich gas streams.

A selective catalyst is demanded in order to avoid H2

consumption since CO and H2 oxidations are competitive

reactions [1]. Pt-supported catalysts are typically used because

CO is more strongly adsorbed than H2 on the Pt metal surface

[2e4], but other noble metals have been studied and the

catalytic activity was reported to be Ru> Rh> Pt> Pd [5].

Investigations with different supports like zeolites MOR, ZSM-

5 and FAU have also been cited [6]. Other classes of CO

selective oxidation catalysts comprise non-noble metals such

as Ag [7], Au [8e10] and also oxides, e.g., CuOeCeO2 [11],

ported LaCoO3 perovskite for selective CO oxidation (SELOX),ydene.2011.12.052

ublications, LLC. Published by Elsevier Ltd. All rights reserved.

0.0 0.2 0.4 0.6 0.8 1.00

50

100

150

200

250

300

Qua

ntit

y ad

sorb

ed (

cm3 /g

ST

P)

Relative Pressure (P/Po)

AL

10LC

20LC

40LC

Fig. 1 e Nitrogen adsorption/desorption isotherms of the

samples with various LaCoO3 contents and the support.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y x x x ( 2 0 1 2 ) 1e1 02

CuMnOx [12], CeO2, CueLaO2eCeO2 [13]. The development of

non-noble metal catalysts for SELOX with high activity and

selectivity is of great importance and represents a meaningful

challenge in industrial application. In this sense,mixed oxides

with perovskite-like structure have attracted extensive

attention in recent years because of their high efficiency in

both oxidation catalysis and membrane technology due to

surprisingly high mobility of structural oxygen, however

scarce applications of this class of oxides are reported in the

SELOX reaction.

Perovskite-type oxides are effective in catalytic oxidation

and particularly the LaCoO3 perovskite has shown activity

comparable to that of Pt/Al2O3 [14]. However, the bulk perov-

skites prepared via conventional procedures exhibit rather

surface area lower than 30 m2/g [15] which strongly limits the

application of these materials as catalysts. The LaCoO3 re-

ported by Taguchi et al. [16] showed exceptionally high CO

oxidation activity; however the authors observed a decrease in

the catalytic activity to those catalysts synthesized at high

temperatures, which was ascribed to the abrupt loss of

surface area under such synthesis conditions. An alternative

to overcome this situation is to support perovskites on porous

solid matrices and the literature reports some studies using

traditional oxide supports such as Al2O3 [17], Al2O3eLa2O3 [18],

SiO2 [19], Ce1�xZrxO2 [20], ZrO2 [21] and molecular sieves [22],

nevertheless no report exists to perovskite supported on

alumina applied in the SELOX reaction.

The aim of this research was to prepare well-defined

lanthanum cobalt perovskites on alumina and evaluate their

catalytic properties for SELOX reaction by correlating the

Table 1 e Physicochemical properties of the samples with var

Sample Chemical composition (wt.%) Mesopore volu

LaCoO3 oxide content

AL e 0.45

10LC 9.3 0.40

20LC 17.4 0.36

40LC 34.4 0.27

LC 100 0.01

Please cite this article in press as: Chagas CA, et al., Alumina-supInternational Journal of Hydrogen Energy (2012), doi:10.1016/j.ijh

activity with their physicochemical properties. Various

LaCoO3 contents on alumina were prepared in order to

determine the most efficient formulation for SELOX reaction.

2. Experimental

2.1. Synthesis of the supported perovskites

LaCoO3 perovskites supported onto alumina were prepared

according to the following methodologies.

2.1.1. Bulk perovskite preparationPerovskite-type oxide LaCoO3 (denoted as LC) was synthesized

by the polymerizable complex method based on the poly-

esterification between citric acid (CA) and ethylene glycol (EG)

[23,24], also called Pechini method. The synthesis procedure

consisted in dissolving La(NO3)3$6H2O and Co(NO3)2$6H2O in

deionized water followed by the addition of CA to give amolar

ratio of total metals and CA 1:1. The mixture was stirred for

30 min at 90 �C on a thermal plate to form stable metal

complexes and after EG was added in the molar ratio

CA:EG¼ 3:2 to promote the polyesterification with the metal

complexes. The polymeric resin obtained after some hours in

constant heating and stirring was heat-treated into two steps,

the first calcination was performed at 450 �C for 2 h in air to

eliminate the organic constituents and the second at 900 �Cfor 3 h in air to generate a pure phase metal oxide.

2.1.2. Supported perovskites preparationThe LaCoO3 powder previously synthesized and g-Al2O3 (deno-

ted asAL)wereused to obtain supportedcatalysts.Aluminawas

firstpretreatedinamufflefurnaceat500 �Cfor6 hunderheating

rate of 10 �Cmin�1 and air atmosphere. Afterwards, a physical

mixture between LaCoO3 and alumina was produced by hand-

grinding in a mortar for 10 min to give x LaCoO3/g-Al2O3

(x¼ 10, 20, 40 wt.%), now designated as 10LC, 20LC and 40LC

respectively, and the resulting powders were heated in muffle

furnace under air atmosphere at 500 �C for 24 h.

2.2. Catalyst characterization

Nitrogen adsorption/desorption isotherms were carried out at

77 K using a Micromeritics ASAP2010 instrument. The

samples were degassed at 300 �C prior to the adsorption

experiments. The specific surface area (SBET) was determined

from the linear part of the BET (Brunauer curve in 0.05e0.35

partial pressure range) and the pore size distribution was

ious LaCoO3 contents.

me (cm3 g�1) Pore diameter (A) BET area (m2 g�1)

6 77.8 164

0 75.2 146

6 74.5 131

7 73.2 96

37 156.4 <10

ported LaCoO3 perovskite for selective CO oxidation (SELOX),ydene.2011.12.052

10 20 30 40 50 60 70 80 90

AL

LC

10LC

20LC

40LC

Inte

nsit

y (a

. u.)

Bragg's angle 2θ (deg.)

(0 1 2)

Fig. 2 e X-ray diffraction patterns of the samples with

different LaCoO3 contents and the support.

Table 2e Crystallite structure properties of the perovskitecatalysts obtained by Rietveld refinement.

Sample Lattice parameters (A) FWHM (deg.) Lhkl* (A)

a¼ b c

LC 5.43963 13.12015 0.24447 L012¼ 520

40LC 5.44229 13.12428 0.18360 L012¼ 850

20LC 5.44359 13.12491 0.21365 L012¼ 650

10LC 5.44319 13.12186 0.28808 L012¼ 400

* Mean crystallite size estimated by means of Scherrer’s equation

using the peak referent to the hkl plane.

i n t e r n a t i o n a l j o u r n a l o f h yd r o g e n e n e r g y x x x ( 2 0 1 2 ) 1e1 0 3

determined by BarretteJoynereHalenda (BJH) method from

the desorption branch of the isotherm. Chemical composition

analysis was performed by X-ray fluorescence (XRF) using

a Rigaku spectrometer RIX 3100 model apparatus.

10 20 30 40 50 60 70 80

Al2O

3

Inte

nsity

(a.

u.)

Bragg's angle 2θ (deg.)

Yobs.

Ycalc.

Yobs. - Ycalc.

Bragg position

LaCoO3

Fig. 3 e Rietveldrefinementof thecatalyst40LC.Experimental

data are indicated by circles and the calculated values by

refinement form the continuous line. The difference between

the experimental data and the calculated curve is represented

by the lowest curve.

Please cite this article in press as: Chagas CA, et al., Alumina-supInternational Journal of Hydrogen Energy (2012), doi:10.1016/j.ijh

The X-ray powder diffraction measurements (XRD) were

performed using a Rigaku Miniflex diffractometer operated at

30 kV and 15 mA, using CuKa radiation (l¼ 1.5406 A) with

a step size of 0.05� and counting time of 1 s per step. Crystal-

line structures were refined with the Rietveld technique using

Fig. 4 e SEM micrographs of (a) LC and (b) 40LC catalysts

after synthesis.

ported LaCoO3 perovskite for selective CO oxidation (SELOX),ydene.2011.12.052

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y x x x ( 2 0 1 2 ) 1e1 04

FULLPROF� 98 code. The Rietveld refinement provided

composition of each phase, crystallite size and lattice

parameters of the crystalline structures.

For TPR, 50 mg of sample was placed in a quartz micro-

reactor coupled to a quadrupole mass spectrometer (Bal-

zersePfeiffer). The pretreatment consisted in dehydrating the

powder at 300 �C for 1 h in He flow 50 cm3 min�1 prior the

reduction. After cooling to room temperature the gas mixture

was switched to 10 vol.% H2/He at 50 cm3min�1 and the

samplewas heated up to 650 �C at 10 �Cmin�1 with a 1 h dwell

at 650 �C. Calibration with CuO powder was performed to

quantify the reduction extent of the samples.

Diffuse reflectance spectra were recorded using an

UVeViseNIR spectrophotometer (Cary 5E Varian) equipped

with an integrating sphere (Harrick) and the analyses were

performed at room temperature in the wavelength range of

200e800 nm; spectra were obtained using the

SchulzeKubelkaeMunk’s equation ( f(R)).

Surface composition was analyzed by X-ray photoelectron

spectroscopy (XPS) using a XR50 SPECS electron spectrometer

with a PHOIBOS hemisphere analyzer and AlKa radiation.

Each spectrumwas calibrated by using Al 2p and peak position

of (74.3� 0.1 eV). Spectra were analyzed and peak-fitted after

subtraction of a Shirley background using a Gaussian-

Lorenzian peak shape obtained from CasaXps software

package.

Hydrogen chemisorption was carried out for LC and 40LC

samples in order to obtain the metal specific area as an indi-

rect measurement of the perovskite area, exclusively to

compare both catalysts. The measurements were obtained

Fig. 5 e EDS/SEM analysis of 40

Please cite this article in press as: Chagas CA, et al., Alumina-supInternational Journal of Hydrogen Energy (2012), doi:10.1016/j.ijh

with an apparatus model ASAP-2020 (Micromeritics�). Before

H2-chemisorption, ca. 500 mg of each sample was evacuated

at 300 �C/30 min, cooled to room temperature and reduced in

a 10 vol.% H2/Ar flow under heating rate of 10 �C min�1 up to

650 �C, remaining at this temperature for 30 min. Then the

sample was cooled to 400 �C and evacuated for 1 h at

1.0�10�6 Torr, and again cooled to 150 �C, which was the

temperature of H2 adsorption isotherm, varying the pressure

from 50 to 515 Torr. The first isotherm represented the total

amount of H2 adsorption and the second one the reversible

physisorbed hydrogen. The total H2 uptake was determined

(mmol H2 per gram of catalyst or gram of cobalt), by extrapo-

lation to null pressure. The adsorption stoichiometry was H/

Co¼ 1/1 and the metal specific area (Sm) was calculated

according to:

Sm ¼ 2� 10�6 �HT �NA

NS

where HT is the total H2 uptake (mmol/gCo); NA, Avogadro

constant and Ns, 1.51� 1019 atomsCo/m2.

It must be emphasized that the total-H2 isotherm was

employed and not the irreversible chemisorption. Such

procedure is adopted in the literature since the H2-chemi-

sorption over cobalt is usually activated (at 150 �C) [25e28].

2.3. Catalytic tests

The catalytic performance tests were carried out with a fixed

bed flow reactor under atmospheric pressure. Approximately

150 mg of the catalysts were dried in situ under nitrogen flow

LC catalyst after synthesis.

ported LaCoO3 perovskite for selective CO oxidation (SELOX),ydene.2011.12.052

200 400 600

90

10LC

40LCInte

nsity

(a.

u.)

Temperature (oC)

650

LC

419

572

420

583

400 580

91 403 586

20LC

isothermal

Fig. 6 e Temperature-programmed reduction profiles

under 10 vol.% H2/He for the samples.

i n t e r n a t i o n a l j o u r n a l o f h yd r o g e n e n e r g y x x x ( 2 0 1 2 ) 1e1 0 5

at 200 �C for 30 min. The reactor was then cooled to 100 �C.The reaction gas mixture consisted in 1 vol.% CO, 1 vol.% O2,

60 vol.%H2 andN2 balance at a flow rate of 100 cm3min�1. The

outlet gas mixture from the reactor was analyzed with a gas

chromatograph system equipped with TCD and FID detectors.

A nickel catalytic converter was used for detecting trace

amount of CO (below 10 ppm). The activity tests were

200 300 400 500 600 700

0.0

1.0

2.0

3.0

4.0

5.0

20LC

40LC

AL

10LC

F (r

)

λ (nm)

LC

800

Fig. 7 e UV-DRS spectra of the samples with different

LaCoO3 contents and the support.

Please cite this article in press as: Chagas CA, et al., Alumina-supInternational Journal of Hydrogen Energy (2012), doi:10.1016/j.ijh

performed at different temperatures, ranging from 100 to

300 �C, in step increments of 20 �C that were maintained for

30 min at each temperature. The conversions of CO and O2

and the CO2 selectivity were defined as follows:

CO conversion ð%Þ ¼ ½CO�in � ½CO�out½CO�in

� 100

O2 conversion ð%Þ ¼ ½O2�in � ½O2�out½O2�in

� 100

CO2 selectivityð%Þ ¼ 0:5� ½CO2�out½O2�in � ½O2�out

� 100

Equilibrium compositions were calculated by the minimi-

zation of the Gibb’s free energy. The calculations were done

for products as a function of temperature, pressure and

composition, based on the Lagrange multipliers principle and

ideal gas hypothesis. In this study, blank experiments were

carried out with g-Al2O3 support and no oxidation of CO was

detected under the adopted reaction conditions. All the cata-

lytic tests were carried out in duplicates and the values ob-

tained for CO conversion showed a standard deviation below

3.0%.

3. Results and discussion

3.1. Characterization of the catalysts

The N2 adsorption/desorption isotherms of the supported

catalysts and alumina are shown in Fig. 1. All adsorption

isotherms are similar to type VI isotherm according to the

IUPAC classification, which are typical of mesoporous solids

(20e500 A) [19]. The type H1 hysteresis obtained is character-

istic of solids consisting of particles crossed by nearly cylin-

drical channels or made by aggregates (consolidated) or

agglomerates (unconsolidated) of spherical particles [29].

Table 1 shows the physicochemical properties of the

samples with different LaCoO3 contents. All of the catalysts

showed lower LaeCo content than the nominal values, which

can be attributed to the loss of material during the synthesis

process, especially in the grinding step as previously

described in Section 2.

The gradual loss of BET area, as perovskite content rises

(Table 1), can be associated to an increase of density of the

materials after LaCoO3 oxide deposition. The decrease in

mesopore volume (0.456e0.277 cm3 g�1) and in pore diameter

(from 77.8 to 73.2 A) observed for the supported catalysts as

a function of LaCoO3 loading is ascribed to the perovskite

blocking the pores (therefore limiting the accessibility of

sorptive molecules) and also to the loading inside the chan-

nels. Similar phenomenon has been observed by Deng et al.

[30] over x LaCoO3/SBA-15 (x¼ 10e50 wt.%) catalysts. As seen

in Table 1, the specific surface area for bulk perovskite (LC)

was quite low owing to the high calcination temperature.

Fig. 2 illustrates the XRD patterns of the supported cata-

lysts 10LC, 20LC and 40LC, as well as the pure perovskite (LC)

and support. Diffraction lines of LC agree with the JCPDS 48-

0123 file and display the formation of a single crystalline

ported LaCoO3 perovskite for selective CO oxidation (SELOX),ydene.2011.12.052

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y x x x ( 2 0 1 2 ) 1e1 06

phase of the LaCoO3 perovskite with rhombohedral unit cell

and space group R-3c. Besides a doublet of themain perovskite

peak in the 2q region, 32.8e33.3�, usually indicates formation

of a perovskite A3þB3þO3 with large A and small B ions [31]. On

the other hand, alumina support showed to be amorphous

whilst the supported catalysts displayed characteristic peaks

of the perovskite structure. Even the sample with the lowest

LaCoO3 content (10LC) showed the striking presence of

perovskite peaks, suggesting indirectly that the catalysts

presented low dispersion.

The Rietveld refinement was applied in XRD data from LC,

10LC, 20LC and 40LC and resulted in a very good correspon-

dence between experimental data and simulated curve, as

exemplified in Fig. 3 with 40LC. For the other samples, the

results were as satisfactory as to 40LC. Some structural

properties inferred by Rietveld refinement are presented in

Table 2 and revealed that changes in the lattice parameters of

perovskite phase were insignificant (<0.1%) upon supporting

it over alumina. On the other hand, the average crystallite size

estimated by Scherrer’s equation for the supported samples

showed that the highest the perovskite loading the largest the

mean crystallite size. Furthermore, the mass quantification of

LaCoO3 phase provided by Rietveld method for the samples

10LC, 20LC and 40LC resulted in 9.9, 17.7 and 34.6 wt.%,

respectively, showing good agreement with the XRF results

(Table 1).

Fig. 4 shows the SEM images of the samples LC and 40LC.

The bulk catalyst LC exhibited a homogeneousmicrostructure

with grain size and diameter ranging from 300 to 2000 A

(Fig. 4a). This result is in part consistent with that obtained by

XRD, i.e., 520 A (estimated by Scherrer’s equation) but shows

the presence of many grains resulting from agglomeration of

simple crystallites. On the other hand, the supported catalyst

40LC (Fig. 4b) showed different morphological and textural

characteristics from the bulk LC, with clusters of LaCoO3

perovskite heterogeneously spread on alumina surface. A

representative EDS result for 40LC is illustrated in Fig. 5 sug-

gesting that the elements La and Co are concentrated in the

regions where there are agglomerates (perovskite agglomer-

ates), which are surrounded by alumina. As a previous

conclusion, the synthesis method employed for the supported

perovskites (physical mixture followed by thermal treatment)

800 790 780 770

780.2 eV795.5 eVCo 2p3/2Co 2p1/2

ΔE = 15.3 eV

C.P

.S. (

a.u.

)

Binding Energy (eV)

a

Fig. 8 e (a) Co 2p and (b) La 3d X

Please cite this article in press as: Chagas CA, et al., Alumina-supInternational Journal of Hydrogen Energy (2012), doi:10.1016/j.ijh

favored a heterogeneous distribution of perovskite on the

support, what agrees with the XRD results that suggest a low

dispersion.

TPR experiments were performed in order to examine the

influence of reducing conditions on the catalyst structure and

perovskite-support interactions, given the H2-rich environ-

ment used in catalytic tests. Previous research has shown that

cobalt oxide is quite active for CO oxidation, but the presence

of H2-rich atmosphere under reaction conditions may lead

cobalt to reduce to lower valences, including metallic cobalt

[32], which tends to catalyze the methanation and H2 oxida-

tion reactions. The reduction profiles are shown in Fig. 6.

All catalysts containing perovskite phase showed similar

reduction profiles with two peaks. The first reduction peak

begins at about 300 �C and ends around 500 �C, which is

ascribed to the reduction of Co3þ to Co2þ, agreeing with the

quantitative analysis and theoretical calculation of LC, corre-

sponding to 33% reduction; however, the first peak of the

supported catalysts indicates 37e39% of reduction, suggesting

that a small fraction of Co2þ is reduced to Co0. The second

reduction peak between 500 and 650 �C indicates complete

reduction (100%), which is ascribed to the reduction of Co2þ to

metal cobalt. The TPR results are consistent with the reported

results in the literature [25,33,34]. The TPR-profiles for sup-

ported catalysts and pure perovskite (LC) are very similar and

demonstrate that there are not other Co3þ coordination types

in the crystallite structure of LaCoO3 phase. The H2 amounts

confirm complete reduction of Co species on all catalysts. The

absence of other peaks suggests similar interaction of the

perovskite phase with the support on all the supported

samples.

The DRS spectra in the UVevisible range can reveal the

existence of ded transitions and oxygen-metal charge trans-

fer, and based on these informations it is possible to infer

about the geometry of the metal centers existing in the

materials [35]. The spectra obtained for LC catalyst (Fig. 7)

revealed a band in the ultraviolet region (260 nm) character-

istic of charge transfer metal-ligand which is attributed to the

Co3þ ion in octahedral environment, and another band in the

visible region (560 nm) characteristic of ded transitions of

cobalt. Evidences of broad bands at 340 and 408 nm are

ascribed to distortions of Co3þ in octahedral coordination

860 850 840 830

La 3d3/2

833.9 eV

C.P

.S. (

a.u.

)

Binding Energy (eV)

837.7 eV

La 3d5/2

ΔE = 3.8 eV

b

PS spectra of 40LC catalyst.

ported LaCoO3 perovskite for selective CO oxidation (SELOX),ydene.2011.12.052

100 120 140 160 180 200 220 240 260 280 3000

20

40

60

80

100 Therm. equilibrium

CO

Con

vers

ion

(%)

Temperature (°C)

10LC

20LC

40LC

LC

Fig. 9 e Catalytic activity in terms of CO conversion as

a function of reaction temperature.

i n t e r n a t i o n a l j o u r n a l o f h yd r o g e n e n e r g y x x x ( 2 0 1 2 ) 1e1 0 7

[22,36]. Alumina support spectrum showed a band at 227 nm

attributed to Al-OH bond characteristic of this material [35]

and the catalyst 10LC presented no significant change in the

ultraviolet region regarding alumina. In contrast, for 40LC the

band at 227 nmdisappeared and the spectral profile resembles

that of the non-supported sample. The DRS spectra of LC and

40LC did not exhibit significant change in the bands assigned

to the Co3þ in octahedral coordination, providing one more

evidence that the method to support LaCoO3 on alumina did

not cause structural modifications in the perovskite structure.

The XPS spectrum of Fig. 8 shows the binding energies at

780.2 and 795.5 eV which are associated to the doublet 2p3/2and 2p1/2 of cobalt on the 40LC sample. The energy difference

of the doublet is 15.3 eV, which is ascribed to Co3þ species of

the mixed LaCoO3, according to Natile et al. [37]. We did not

see the satellite peak around 785e788 eV related to Co2þ, assuggested by Merino et al. [38]. Indeed it evidences that the

Pechini preparation method was efficient forming pure

LaCoO3 perovskite over the alumina support, in contradiction

to Villoria et al. [39] asserting that the surface chemical state of

pure perovskite is constituted by Co3þ/Co2þ species. Fig. 8b

evidences the presence of La3þ at the surface in agreement

with the literature [37,38].

100 120 140 160 180 200 220 240 260 280 3000

20

40

60

80

100

O2 C

onve

rsio

n (%

)

Temperature (°C)

10LC

20LC

40LC

LC

Fig. 10 e Catalytic activity in terms of O2 conversion as

a function of reaction temperature.

Please cite this article in press as: Chagas CA, et al., Alumina-supInternational Journal of Hydrogen Energy (2012), doi:10.1016/j.ijh

3.2. Catalytic evaluation

The CO and O2 conversions as a function of temperature are

shown in Figs. 9 and 10, respectively. The catalytic activity is

usually associated to the temperature corresponding to 50%

conversion of CO to CO2 (T50%CO), which can be obtained

from the light-off curve [40]. In this work, the T50%CO was

used to compare the activity of the catalysts and the

results are summarized in Table 3. The 40LC catalyst exhibited

the highest activity for SELOX reaction characterized by a

T50%CO¼ 168 �C.On the other hand, 10LC (catalyst with the lowest LaCoO3

loading) was the less active, since it presented conversion

below 50% over the entire investigated temperature range.

Both LC and 20LC catalysts presented an intermediate

performance between 10LC and 40LC (Fig. 9) and T50%CO¼240 �C and 230 �C for LC and 20LC, respectively. The

increasing performance obtained for the supported catalysts

(10LC< 20LC< 40LC) is explained through the higher effective

exposition of the perovskite active phase over the supports

when the loading increases from 10 to 40 wt.%. However, the

bulk catalyst (LC) did not present better activity than 40LC,

probably due to its severely limited specific surface area,

restricting its performance. A probable explanation for the

difference in performances between LC and 40LC may be that

the latter one has more perovskite area for SELOX than the

bulk LC.

In order to obtain evidence about this hypothesis, H2-

chemisorption was carried out on LC and 40LC samples and

themetal cobalt specific areas were determined from the total

H2-uptakes, resulting in 6.2 and 13.5 m2/gCo for LC and 40LC,

respectively. Therefore, the supported catalyst 40LC pre-

sented ca. twice as many metallic area than LC, which

suggests also twice as many perovskite at the surface avail-

able to SELOX reaction. Therefore, the higher activity of 40LC

in the range of temperature studied may be attributed to its

higher perovskite surface exposed area.

The O2 conversion shown in Fig. 10 presented an important

remark: the most active catalyst, 40LC, was the only one to

lead to total O2 conversion (from 220 �C) and this happened

while CO had not been fully converted at this temperature. It

must be pointed out that the relationship CO/O2 of the feed

stream was 1/1, which represents an excess of oxygen con-

cerning the theoretical reaction stoichiometry, i.e., 1/0.5.

Table 3e Temperature of half-conversion, temperature ofmaximum conversion and CO2 selectivity.

Catalyst CO conversion CO2 selectivity (%)a

T50%CO (�C)a Tmax (�C)b

LC 240 280 82

10LC ec 300 ec

20 LC 230 280 35

40LC 168 280 75

a T50%CO, temperature at which 50% of CO was converted.

b Tmax, temperature corresponding to the maximum CO conver-

sion reached in the SELOX reaction.

c It showed conversion below 50% for all temperatures tested.

ported LaCoO3 perovskite for selective CO oxidation (SELOX),ydene.2011.12.052

100 120 140 160 180 200 220 240 260 280 30020

40

60

80

100

CO

2 S

elec

tivity

(%

)

Temperature (°C)

10LC

20LC

40LC

LC

Fig. 11 e Catalytic selectivity to CO2 as a function of

reaction temperature.

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y x x x ( 2 0 1 2 ) 1e1 08

Therefore, the total O2 conversion reached while a fraction of

CO remained not reacted suggests the presence of side reac-

tions such as H2 oxidation on 40LC catalyst. This hypothesis is

supported by a previous work [41] in which we verified by

means of temperature-programmed SELOX reaction (with

similar feed stream) that LaCoO3 perovskite started catalyzing

H2 oxidation above 130 �C, since the formation of H2O and

decrease of CO2 signal were clearly evidenced to take place

from this temperature on.

Fig. 11 shows CO2 selectivity as a function of temperature

for the catalysts investigated in the CO oxidation. The selec-

tivities were compared under CO isoconversion conditions,

i.e., at T50%CO (Table 3) and it was verified that LC obtained the

highest CO2 selectivity (82%) followed by 40LC (75%). A

Table 4 e Comparison between data from literature and this w

Catalyst(Pt and Co in wt.%)

Feed molar composition (vol. %)

CO O2 He or N2 H2

40LC (8.3% Co) 1 1 38 60

LC (24.1% Co)

6.0% Pt/Al2O3 1 1 e 98

6.0% Pt-A

5.9% Pt-MOR

5.8% Pt-X

1% Pt/Al2O3 1 1 38 60

2% Pt/Al2O3

0.5% Pt/SiO2-TiO2 1 1 e 98

1% Pt/ZrO2 5 5 78 12

1% Pt/Al2O3

1% Pt/SiO2

10%Co/ZrO2 1 1 58 40

10.6%Co/ZrO2/monolith

10%Co/CeO2 1 1 38 60

10%Co/ZrO2 1 1 38 60

10%Co/CeO2

10%Co/Al2O3

a Spaceetime is defined as the ratio catalyst mass (mg)/feed flow rate (c

b Pt catalysts supported on A, mordenite and X zeolites.

c Previous studies of our research group with lower H2 concentrations.

d CO conversion below 10% in the temperature range 50e225 �C investig

Please cite this article in press as: Chagas CA, et al., Alumina-supInternational Journal of Hydrogen Energy (2012), doi:10.1016/j.ijh

comparison between the data of this work and some data re-

ported in the literature on SELOX employing Pt-based catalysts

and cobalt supported catalysts is shown in Table 4. Some

measurements were carried out at different feed conditions

and spaceetimes, thus a more appropriate comparison would

require the same ones, since CO2 and H2O can affect the

catalyst performance. In spite of that, it is possible to notice

the catalysts 40LC and LC, investigated in this work, are very

promising for use in the SELOX reaction, since they showed

notable higher CO2 selectivity than noble metal-based cata-

lysts. In particular, 40LC was tested under similar reaction

conditions to 1 wt.% Pt/Al2O3 of [42] andboth presented similar

T50%CO, but the supported 40LC exhibited superior selectivity.

According to the literature [43e45] the supported cobalt

oxides showed T50%CO values lower than the most active

catalyst (40LC), and similar high CO2 conversions, however,

less stability of the oxide structure in the presence of H2

atmosphere leading to side reactions. Woods and coworkers

[44] claim that for 10%CoOx/CeO2, CO competes with H2

oxidation between 175 and 275 �C while the methanation

reaction dominates above 275 �C. An interesting result was

reported by Zhao et al. [45], employing 10%Co/Al2O3 under

similar feed composition, showing very low CO conversion in

the temperature range of 50e225 �C, contrasting with the

supported perovskite 40LC (8.3 wt.%Co).

Some previous works of our research group [10,46,47]

employing supported noble metals and gold-based catalysts

showed excellent results in the selective CO oxidation and this

experience motivated the application of perovskite. Recently

other publications about selective CO oxidation (also called

preferential oxidation) using perovskites [48,49] have been

reported, however, most investigated this reaction employing

an H2-free feed stream. Comparing our results to those

ork about SELOX reaction with H2-rich feed stream.

Space-timea

(mgmin cm�3)T50%CO (�C) CO2 seletivity (%) Reference

1.5 168 75 This work

240 82

1.0 154 32 [50] b

210 55

172 47

190 31

1.75 165 46 [42]

150 43

3.0 215 25 [51]

1.75 120 28 [46] c

131 56

190 43

2.1 167 84 [43]

185 70

4.0 122 96 [44]

4.0 130 86 [45]

149 72

ed e

m3min�1 STP).

ated and CO2 selectivity below 37% in the same range.

ported LaCoO3 perovskite for selective CO oxidation (SELOX),ydene.2011.12.052

i n t e r n a t i o n a l j o u r n a l o f h yd r o g e n e n e r g y x x x ( 2 0 1 2 ) 1e1 0 9

obtained with materials recognized as the most active for

SELOX reaction it seems that the LaCoO3 perovskite-type

oxide supported on Al2O3 is promising in the SELOX reaction

with H2-rich feed stream.

4. Conclusions

Cobalt-based perovskites showed to be highly effective for

selective CO oxidation (SELOX) in a H2-rich stream. The

characterization results showed that the Pechini method was

suitable for obtaining single phase perovskite, but the

sequential procedure to support the perovskite onto alumina

did not lead to high LaCoO3 dispersion, though such a method

has not caused structural changes in the active phase after

supporting and heat-treatment steps. XPS results showed that

the active surface species of the mixed LaCoO3 are preferen-

tially Co3þ when supported on alumina. Furthermore, LaCoO3

perovskite structure presented reductive behavior only above

300 �C, which is an interesting feature since the reduction of

cobalt in the temperature range for SELOX may lead to

undesirable reactions (methanation, reverse water gas shift,

H2 oxidation).

Both, the 40 wt.% LaCoO3/Al2O3 and bulk LaCoO3, catalysts

showed high performance for CO oxidation, especially con-

cerning the selectivity. The higher activity presented for the 40

wt.% LaCoO3/Al2O3 catalyst may be attributed to its larger

exposed perovskite area available. These results evidence that

the supported perovskite oxides are very good alternative for

SELOX reaction when compared to noble metal supported

catalyst.

Acknowledgements

The authors gratefully acknowledge Carlos Andre de Castro

Perez for technical support in Rietveld refinement, Marta M.M.

Amorim for SEM/EDS analysis, as well as CNPq, CAPES, FAPERJ

and FINEP for financial support.

r e f e r e n c e s

[1] Bernardo P, Algieri C, Barbieri G, Drioli E. Catalytic (PteY)membranes for the purification of H2-rich streams. CatalToday 2006;118:90e7.

[2] Sakamoto Y, Higuchi K, Takahashi N, Yokota K, Doi H,Sugiura M. Effect of the addition of Fe on catalytic activitiesof Pt/Fe/geAl2O3 catalyst. Appl Catal B 1999;23:159e67.

[3] Mozer TS, Passos FB. Selective CO oxidation on Cu promotedPt/Al2O3 and Pt/Nb2O5 catalysts. Int J Hydrogen Energy 2011;36:13369e78.

[4] Lu S, Zhang C, Liu Y. Carbon nanotube supported PteNicatalysts for preferential oxidation of CO in hydrogen-richgases. Int J Hydrogen Energy 2011;36:1939e48.

[5] Oh SH, Sinkevitch RM. Carbon monoxide removal fromhydrogen-rich fuel cell feedstreams by selective catalyticoxidation. J Catal 1993;142:254e62.

[6] Sebastian V, Irusta S, Mallada R, Santamarıa J. Selectiveoxidation of CO in the presence of H2, CO2 and H2O, on

Please cite this article in press as: Chagas CA, et al., Alumina-supInternational Journal of Hydrogen Energy (2012), doi:10.1016/j.ijh

different zeolite-supported Pt catalysts. Appl Catal A 2009;366:242e51.

[7] Hu R, Yan C, Xie L, Cheng Y, Wang D. Selective oxidation ofCO in rich hydrogen stream over Ag/OMS-2 catalyst. Int JHydrogen Energy 2011;36:64e71.

[8] Luengnaruemitchai A, Thoa DTK, Somchai S, Gulari E. Acomparative study of Au/MnOx and Au/FeOx catalysts for thecatalytic oxidation of CO in hydrogen rich stream. Int JHydrogen Energy 2005;30:981e7.

[9] Yang YF, Sangeetha P, Chen YW. Au/TiO2 catalysts preparedby photo-deposition method for selective CO oxidation in H2

stream. Int J Hydrogen Energy 2009;34:8912e20.[10] Ribeiro NFP, Mendes FMT, Perez CAC, Souza MMVM,

Schmal M. Selective CO oxidation with nano gold particles-based catalysts over Al2O3 and ZrO2. Appl Catal A 2008;347:62e71.

[11] Park JW, Jeong JH, Yoon WLY, Kim CS, Lee DK, Park YK, et al.Selective oxidation of CO in hydrogen-rich stream overCueCe catalyst promoted with transition metals. Int JHydrogen Energy 2005;30:209e20.

[12] Njagi EC, Genuino HC, Kingondu CK, Chen CH, Horvath D,Suib SL. Preferential oxidation of CO in H2-rich feeds overmesoporous copper manganese oxides synthesized bya redox method. Int J Hydrogen Energy 2011;36:6768e79.

[13] Gosavi PV, Biniwale RB. Effective cleanup of CO in hydrogenby PROX over perovskite and mixed oxides. Int J HydrogenEnergy; 2011:1e6. doi:10.1016/j.ijhydene.2011.04.110.

[14] Hackenberger M, Stephan K, Kießling D, Schmitz W,Wendt G. Influence of the preparation conditions on theproperties of perovskite-type oxide catalysts. Solid StateIonics 1997;101:1195e200.

[15] Tejuca LG, Fierro JLG, Tascon JMD. Structure and reactivity ofperovskite-type oxides. Adv Catal 1989;36:237e328.

[16] Taguchi H, Yamada S, Nagao M, Ichikawa Y, Tabata K.Surface characterization of LaCoO3 synthesized using citricacid. Mater Res Bull 2002;37:69e76.

[17] Rosmaninho MG, Tristao JC, Moura FCC, Lago RM,Araujo MH, Fierro JLG. Structural and surface analysis ofunsupported and alumina-suppored La(Mn, Fe, Mo)O3

perovskite oxides. Anal Bioanal Chem 2010;396:2785e95.[18] Liotta LF, Di Carlo G, Longo A, Pantaleo G, Deganello G,

Marcı G, et al. Structural and morphological properties ofCoeLa catalysts supported on alumina/lanthana forhydrocarbon oxidation. J Non-Cryst Solids 2004;345:620e3.

[19] Nguyen SV, Szabo V, Trong On D, Kaliaguine S. Mesoporoussilica supported LaCoO3 perovskites as catalysts for methaneoxidation. Micropor Mesopor Mat 2002;54:51e61.

[20] Alifanti M, Florea M, Parvulescu VI. Ceria-based oxides assupports for LaCoO3 perovskite: catalysts for total oxidationof VOC. Appl Catal B 2007;70:400e5.

[21] Villoria JA, Alvarez-Galvan MC, Navarro RM, Briceno Y,Gordillo Alvarez F, Rosa F, et al. Zirconia-supported LaCoO3

catalysts for hydrogen production by oxidative reforming ofdiesel: optimization of preparation conditions. Catal Today2008;138:135e40.

[22] Makshina EV, Nesterenko NS, Siffert S, Zhilinskaya EA,Aboukais A, Romanovsky BV. Methanol oxidation on LaComixed oxide supported onto MCM-41 molecular sieve. CatalToday 2008;131:427e30.

[23] Popa M, Kakihana M. Synthesis of lanthanum cobaltite(LaCoO3) by the polymerizable complex route. Solid StateIonics 2002;151:251e7.

[24] Lin J, Yu M, Lin CK, Liu X. Multiform oxide optical materialsvia the versatile pechini-type sol-gel process: synthesis andcharacteristics. J Phys Chem C 2007;111:5835e45.

[25] Echchahed B, Kaliaguine S, Alamdari H. Well dispersed Co0

by reduction of LaCoO3 perovskite. Int J Chem Reactor Eng2006;4:A29.

ported LaCoO3 perovskite for selective CO oxidation (SELOX),ydene.2011.12.052

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n en e r g y x x x ( 2 0 1 2 ) 1e1 010

[26] Xiong J, Borg Ø, Blekkan EA, Holmen A. Hydrogenchemisorption on rhenium-promoted c-alumina supportedcobalt catalysts. Catal Commun 2008;9:2327e30.

[27] Reuel RC, Bartholomew CH. The stoichiometries of H2 andCO adsorptions on cobalt: effect of support and preparation. JCatal 1984;85:63e77.

[28] Silva RRCM, Schmal M, Frety R, Dalmon JA. Effect of thesupport on the FischereTropsch synthesis with Co/Nb2O5

catalysts. J Chem Soc Faraday Trans 1993;89:3975e80.[29] Leofanti G, Padovan M, Tozzola G, Venturelli B. Surface area

and pore texture of catalysts. Catal Today 1998;41:207e19.[30] Deng J, Zhang L, Dai H, Au CT. In situ hydrothermally

synthesized mesoporous LaCoO3/SBA-15 catalysts: highactivity for the complete oxidation of toluene and ethylacetate. Appl Catal A 2009;352:43e9.

[31] Tai LW, Nasrallah MM, Anderson HU, Sparlin DM, Sehlin SR.Structure and electrical properties of La1�xSrxCo1�yFeyO3.Part 1: the system La0.8Sr0.2Co1�yFeyO3. Solid State Ionics1996;76:259e71.

[32] Omata K, Kobayashi Y, Yamada M. Artificial neural network-aided development of supported Co catalyst for preferentialoxidation of CO in excess hydrogen. Catal Commun 2005;6:563e7.

[33] Tien-Thao N, Alamdari H, Zahedi-Niaki MH, Kaliaguine S.LaCo1�xCuxO3�d perovskite catalysts for higher alcoholsynthesis. Appl Catal A 2006;311:204e12.

[34] Bedel L, Roger AC, Estournes C, Kiennemann A. Co0 frompartial reduction of La(Co, Fe)O3 perovskites for FischerTropsch synthesis. Catal Today 2003;85:207e18.

[35] Martra G, Gianotti E, Coluccia S. In: Jackson SD,Hargreaves JSJ, editors. Metal oxide catalysis. Weinheim:Wiley-VCH; 2009.

[36] Finocchio E, Montanari T, Resini C, Busca G. Spectroscopiccharacterization of cobalt-containing solid catalysts. J MolCatal A 2003;204, 205:535e44.

[37] Natile MM, Ugel E, Maccato C, Glisenti A. LaCoO3: effect ofsynthesis conditions on properties and reactivity. App CatalB 2007;72:351e62.

[38] Merino NA, Barbero BP, Grange P, Cadus LE. La1�xCaxCoO3

perovskite-type oxides: preparation, characterisation,

Please cite this article in press as: Chagas CA, et al., Alumina-supInternational Journal of Hydrogen Energy (2012), doi:10.1016/j.ijh

stability, and catalytic potentiality for the total oxidation ofpropane. J Catal 2005;231:232e44.

[39] Villoria JA, Alvarez-Galvan MC, Al-Zahrani SM, Palmisano P,Specchia S, Specchia V, et al. Appl Catal B 2011;105:276e88.

[40] Haruta M, Yamada N, Kobayashi T, Iijima S. Gold catalystsprepared by coprecipitation for low-temperature oxidationof hydrogen and of carbon monoxide. J Catal 1989;115:301e9.

[41] Magalhaes RNSH, Toniolo FS, Silva VT, Schmal M. SelectiveCO oxidation reaction (SELOX) over cerium-doped LaCoO3

perovskite catalysts. Appl Catal A 2010;388:216e24.[42] Manasilp A, Gulari E. Selective CO oxidation over Pt/alumina

catalysts for fuel cell applications. Appl Catal B 2002;37:17e25.[43] Gomez LE, Tiscornia IS, Boix AV, Miro EE. Co/ZrO2 catalysts

coated on cordierite monoliths for CO preferential oxidation.App Catal A 2011;401:124e33.

[44] Woods MP, Gawade P, Tan B, Ozkan US. Preferentialoxidation of carbon monoxide on Co/CeO2 nanoparticles.App Catal B 2010;97:28e35.

[45] Zhao Z, Yung MM, Ozkan US. Effect of support on thepreferential oxidation of CO over cobalt catalysts. CatalCommun 2008;9:1465e71.

[46] Souza MMVM, Ribeiro NFP, Schmal M. Influence of thesupport in selective CO oxidation on Pt catalysts for fuel cell:applications. Int J Hydrogen Energy 2007;32:425e9.

[47] Marques P, Ribeiro NFP, Schmal M, Aranda DAG,Souza MMVM. Selective CO oxidation in the presence of H2

over Pt and PteSn catalysts supported on niobia. J PowerSources 2006;158:504e8.

[48] Maluf SS, Assaf EM. CO preferential oxidation (CO-PROx) onLa1�xCexNiO3 perovskites. Catal Commun 2011;12:703e6.

[49] Gosavi PV, Biniwale RB. Effective cleanup of CO in hydrogenby PROX over perovskite and mixed oxides. Int J HydrogenEnerg; 2011. doi:10.1016/ j.ijhydene.2011.04.110.

[50] Igarashi H, Uchida H, Suzuki M, Sasaki Y, Watanabe M.Removal of carbon monoxide from hydrogen-rich fuels byselective oxidation over platinum catalyst supported onzeolite. Appl Catal A 1997;159:159e69.

[51] Zheng T, Dong YR, Nishiyama N. Selective CO oxidation overcarbon-coated Pt/SiO2eTiO2 particles. Appl Catal A 2006;308:210e5.

ported LaCoO3 perovskite for selective CO oxidation (SELOX),ydene.2011.12.052