Applied Catalysis B: Environmental 55 (2005) 229–241

www.elsevier.com/locate/apcatb

Production of hydrogen by oxidative reforming of ethanol over

Pt catalysts supported on Al2O3 modified with Ce and La
R.M. Navarroa, M.C. Álvarez-Galvána, M. Cruz Sánchez-Sáncheza,
F. Rosab, J.L.G. Fierroa,*
a
Instituto de Catálisis y Petroleoquı́mica (CSIC), C/Marie Curie s/n. Cantoblanco 28049, Madrid, Spain
b
Centro de Experimentación de ‘‘El Arenosillo’’ (CEDEA), Instituto Nacional de Técnica Aeroespacial (INTA),
Ctra. de San Juan del Puerto-Matalascañas km 31, 21130 Mazagón-Moguer, Huelva, Spain
Received 29 June 2004; received in revised form 31 August 2004; accepted 1 September 2004
Available online 13 October 2004

Abstract

A series of Pt catalysts supported on alumina modified by Ce and/or La were prepared and tested in the reaction involving the production of
hydrogen by oxidative reforming of ethanol. Characterization of the catalysts by N2 adsorption, X-ray diffraction, TPR and X-ray
photoelectron spectroscopy (XPS) revealed differences in the interaction between platinum and the supports. These differences have
important implications in the behaviour of the catalysts for the ethanol-reforming reaction. Ethanol conversion on the Pt/(La)Al2O3 catalyst
was found to be close to that of the Pt/Al2O3 catalyst. However, modification of the Al2O3 substrate by ceria led to a more active Pt/(Ce)Al2O3
catalyst. A lower promotion effect was observed when both ceria and lanthana were present on the support substrate. TPR and XPS analyses of
ceria-containing samples revealed the existence of a strong interaction between platinum and ceria. This interaction between Pt crystallites at
the ceria surface seemed to be responsible for the better performance of the cerium-containing catalysts. It is suggested that the platinum–ceria
interaction affects the adsorption–decomposition of ethanol to CH4 and CO products and their subsequent reforming reactions with steam.
When both ceria and lanthana were present on the support substrate the platinum–ceria interaction was diminished, reducing the promoter
effect in the production of hydrogen by oxidative reforming of ethanol.
# 2004 Elsevier B.V. All rights reserved.

Keywords: Platinum; Alumina; Ceria; Lanthana; Hydrogen production; Ethanol

1. Introduction solutions to remove the huge amounts of CO2 emissions co-

produced in the reforming processes. Additionally, the use of
The environmental problems caused by the intensive use fuels with limited reserves does not overcome the non-
of fossil fuels—as well as their limited availability—are sustainable nature of current energy systems. In this context,
increasingly demanding a more rational use of our energy there is growing interest in the study of methods for the
resources. Among all the different alternatives available, production of hydrogen from renewable sources as effective
hydrogen is an excellent energy vector when used directly as solutions for the reduction of CO2 emissions and for the
a fuel in internal combustion engines or, indirectly, to supply preservation of fossil energy resources. Among the bio-fuel
electricity through the use of fuel cells. Currently, the steam candidates for carriers of hydrogen, ethanol is of particular
reforming of methane is the most used and generally the interest because of: (i) its low toxicity; (ii) its low production
most economical way to obtain hydrogen. However, the use costs; (iii) the fact that is a relative clean fuel in terms of
of fossil fuels as a primary hydrogen source fails to provide composition; (iv) its relatively high hydrogen content and
(v) its availability and ease of handling.
* Corresponding author. Fax: +34 915854760. Hydrogen can be obtained directly from ethanol by two
E-mail address: jlgfierro@icp.csic.es (J.L.G. Fierro). main processes: partial oxidation (POE, Eq. (1)) and steam

0926-3373/$ – see front matter # 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcatb.2004.09.002
230 R.M. Navarro et al. / Applied Catalysis B: Environmental 55 (2005) 229–241

reforming (SRE, Eq. (2)). The overall processes summarised which facilitates coke gasification [15]. From a practical point
in Eqs. (1) and (2) are a complex convolution of elementary of view, and in order to maximize the exposed surface of
steps that involve several organic intermediates. platinum and cerium, both elements are added to high-surface
area substrates such as Al2O3. As stated above, the thermal
CH3 CH2 OH þ 1:5O2 ! 3H2 þ 2CO2 stability of the support is another requirement to be fulfilled by
ðDH  ¼ 545:2kJ=molÞ (1) the catalyst for the ethanol-reforming application. Thermal
stabilization of the support is important since of g-Al2O3
between 873 and 1373 K suffers sintering processes via
CH3 CH2 OH þ 3H2 O ! 6H2 þ 2CO2 surface diffusion with a phase change [16]. Many cations [17],
ðDH  ¼ 207:7kJ=molÞ (2) such as La3+, Ba2+, Zr4+, have positive effects as inhibitors on
the sintering and phase transformations of alumina. Among
Whereas POE offers exothermicity and a rapid response, these cations, lanthanum is known to act as an excellent
SER is endothermic and produces greater amounts of hydro- thermal stabilizer of alumina supports [18].
gen, resulting in higher system efficiencies. A third option With this background, the aim of the present work was to
combines the advantages of both approaches by co-feeding study the role and effect of La and Ce on the behaviour of
oxygen, steam and ethanol simultaneously through oxidative alumina-supported Pt catalysts in the production of hydrogen
reforming processes (ORE, Eq. (3)) by the oxidative reforming of ethanol. Careful investigations
of the structure of the catalysts were performed in an attempt to
CH3 CH2 OH þ xO2 þ ð3  2xÞH2 O ! ð6  2xÞH2
   understand the relationship between activity and their
 3  2x structural and surface characteristics.
þ 2CO2 0 < x < 0:5 DH ¼
3
 x  
 207:7   545:2 kJ =mol (3) 2. Experimental
1:5
According to the literature consulted, the ethanol-reforming 2.1. Catalyst preparation
reaction has been performed on several catalyst systems using
Ni [1,2], Co [3,4], Ni/Cu [5,6] and noble metals (Pd, Pt, Rh) A lanthana-containing alumina support (3 wt.% La2O3,
[7–9] as active phases deposited on oxide supports promoted hereafter abbreviated A–L) was prepared by impregnation of a
with alkali or alkali-earth compounds to improve both thermal commercial alumina (Alfa Aesar, 225 m2/g) with aqueous
stability and the resistance to coke formation. Ethanol-reform- solutions of lanthanum nitrate (Alfa Aesar, 99.9%). Following
ing reactions include several steps that involve the need for impregnation, the sample was dried at 383 K for 8 h and, finally
catalytic surfaces able to: (i) dehydrogenate ethanol; (ii) break calcined in air at 1173 K for 8 h. Similarly, the ceria-containing
the carbon–carbon bonds of surface intermediates to produce alumina support (10 wt.% CeO2, hereafter abbreviated A–C)
CO and CH4 and (iii) water reforms these C1 products to was prepared by impregnation with cerium nitrate (Alfa Aesar,
generate hydrogen. Additionally, catalytic systems must pos- Reacton 99.5%) of the above commercial alumina previously
sess high thermal stability due to the relative high temperature stabilized by thermal treatment at 1173 K for 8 h. The
range (823–1073 K) at which reforming reactions take place. impregnate was dried under air at 383 K for 8 h and
On the basis of the influence of the nature of both the metal and subsequently calcined in air at 773 K for 8 h. The support
support on the catalytic characteristics of supported metals, containing Ce and La (10 wt.% CeO2, 3% La2O3, abbreviated
choice of these elements is a key factor in developing sup- A–L–C) was prepared by impregnation of the A–L support
ported catalysts that fulfil the above requirements. with cerium nitrate using the same methodology as that
Among the noble metals, the good thermal stability and employed for the preparation of the A–C sample. Supported Pt
relatively high C–C bond breaking [10] and WGS [11] catalysts (0.75 wt.% metal) were prepared by impregnating
activities of Pt make it a suitable active phase for ethanol- each of the above supports, as well as the bare alumina calcined
reforming reactions. Additionally, the catalytic activity of at 1173 K, under stirring at 353 K for 2 h using aqueous
dispersed Pt may be strongly influenced by the presence of solutions of H2PtCl6 metal precursor (Aldrich, 99.9%). After
basic promoters. In particular, the addition of CeO2 to Pt-loading, the samples were dried in air at 383 K, followed by
platinum has been reported to be beneficial to Pt for both calcination in air at 773 K for 3 h. The samples were designated
ethanol decomposition [12] and WGS [13] reactions. The as Pt/A, Pt/A–L, Pt/A–C and Pt/A–L–C for platinum-
promoting effect of CeO2 has been attributed [14] to: (i) its supported on bare alumina, lanthanum–alumina, ceria–
higher reducibility in the presence of Pt; (ii) the higher alumina and lanthana–ceria–alumina, respectively.
dispersion of Pt over CeO2 and (iii) prevention of the
sintering of Pt metal particles. The presence of ceria also 2.2. Catalyst characterization
improves the catalytic performance of the systems with
regard to their resistance to coke deposition due to the The chemical composition of the catalysts was deter-
oxygen storage-release capability associated with CeO2, mined by inductively coupled plasma atomic emission
 R.M. Navarro et al. / Applied Catalysis B: Environmental 55 (2005) 229–241 231

spectroscopy (ICP-AES), using a Perkin-Elmer Optima GHSV = 19,492 h1) at atmospheric pressure using a fixed-
3300DV apparatus. The samples were first dissolved in acid bed flow reactor in the 848–989 K temperature range.
solutions (a mixture of HF, HCl and HNO3), microwaved for
15 min, and diluted to concentrations within the detection CH3 CH2 OH þ 2:28H2 O þ 0:36O2 ! 5:28H2 þ 2CO2 (4)
range of the instrument.
N2 adsorption–desorption isotherms were obtained at Activity tests were performed using 0.1 of catalyst diluted
77 K over the whole range of relative pressures, using a with SiC (both in the 0.4–0.5 mm particle size range and
Micromeritics ASAP 2100 automatic device on samples selected after preliminary mass transport experiments to
previously outgassed at 423 K for 12 h. BET specific areas minimize diffusional resistances) at a volume ratio of 3:1
were calculated from these isotherms using the BET method to avoid adverse thermal effects. The catalyst bed was placed
and taking a value of 0.162 nm2 for the cross-section of the in a 6 mm ID quartz tubular reactor with a coaxially centred
physically adsorbed N2 molecule. thermocouple. Prior to reaction, the catalysts were flushed in
Temperature-programmed reduction experiments were nitrogen at 473 K, followed by reduction in situ at 573 K for
carried out on a semiautomatic Micromeritics TPD/TPR 2 h (heating rate 5 K/min) with 50 mL (STP)/min of a
2900 apparatus equipped with a TC detector. Prior to 10 vol.% H2/N2 mixture. The pretreating gases were flushed
reduction experiments, the samples (about 30 mg) were from the reactor with N2 before admission of the ethanol/air/
treated thermally under an air stream at 573 K to remove water reaction mixture. Ethanol and water were fed inde-
water and other contaminants. TPR profiles were pendently into the pre-heater by means of syringe pumps
obtained by heating the samples under a 10% H2/Ar flow (Becton-Dickinson) before mixing with an air stream
(50 mL/min) from 233 to 1173 K at a linearly programmed adjusted with Brooks model 5850E mass flow controllers.
rate of 10 K/min. The effluent gas was passed through a cold The reaction products were analysed on-line by GC with
trap to remove water before measuring the amount of TCD (Varian chromatograph Model Star 3400 CX) equipped
hydrogen consumed during reduction by the TC detector. with Porapack Q (CO2, C2H6, C2H4, water, acetaldehyde,
X-ray powder diffractograms were recorded following ethanol, acetone, acetic acid, diethyl ether, ethyl acetate, and
the step-scanning procedure (step size 0.028, 2u scanning crotonaldehyde) and molecular sieve 5A (H2, O2, N2, CO)
from 208 to 808) using a computerised Seifert XRD 3000P packed columns connected in series, using He as carrier gas.
diffractometer (Cu Ka radiation, l = 0.15418 nm) equipped Activity was measured at reaction temperatures between 848
with a PW Bragg-Brentano u/2u goniometer and a bent and 989 K, running from the lowest to the highest tempera-
graphite monochromator and automatic slit. Volume- ture and maintaining, in order to check if there was any
averaged crystallite sizes were determined by applying deactivation of the catalysts, the reaction for 8 h at each
the Debye–Scherrer equation. temperature. For all the tested samples, carbon balances
X-ray photoelectron spectroscopy (XPS) was used to (close to 100%) and differences in ethanol conversion
study the chemical composition and oxidation state of the between subsequent analyses (lower than 2%) indicated
catalyst surfaces. Photoelectron spectra were recorded with the absence of catalysts deactivation phenomena in the time
a VG Escalab 200R electron spectrometer equipped with a on-stream studied.
Mg Ka X-ray source (hn = 1253.6 eV) and a hemispherical
electron analyser operating at constant transmission energy
(50 eV). The X-ray source was operated at low X-ray fluxes 3. Results and discussion
in order to minimize X-ray-induced reduction of Pt and Ce
species. The reduction treatment was carried out in situ at 3.1. Composition and textural properties of the catalysts
573 K for 1 h. After outgassing at 105 mbar, the calcined or
reduced samples were transferred to the ion-pumped Table 1 shows the chemical compositions obtained from
analysis chamber, whose the residual pressure was kept ICP analyses, expressed as weight percentages, of the
below 7  109 mbar during data acquisition. The C 1s peak calcined supports and catalysts. The textural data offered in
at 284.6 eV was used as an internal standard for peak Table 2 show that the specific areas for the lanthanum-
position measurement. The areas of the peaks were containing support are higher than those corresponding to
estimated by calculating the integral of each peak after bare alumina. This agrees with previous findings reported in
subtracting a Shirley background and fitting the experi- the literature [18] indicating the inhibiting effect of La3+ on
mental peak to a combination of Lorentzian/Gaussian lines the loss of surface area associated with the sintering of
of variable proportions. alumina due to the temperature-induced transition phases of
alumina. In contrast to A–L, the specific area of the A–C
2.3. Catalytic activity measurements support (Table 2) did not differ very much from that of the
bare Al2O3. The same trend was observed on comparing the
The samples in reduced form were tested in the oxidative specific areas of La-modified alumina with/without added
reforming of ethanol (Eq. (4)) with air (O2/CH3CH2OH = Ce (113 and 107 m2/g for A–L–C and A–L, respectively).
0.36, H2O/CH3CH2OH = 2.28, [CH3CH2OH] = 20.1 vol.%, The incorporation of ceria, however, elicited a slight
232 R.M. Navarro et al. / Applied Catalysis B: Environmental 55 (2005) 229–241

Table 1
Chemical composition of calcined supports and platinum-supported cata-
lysts
% Pt % CeO2 % La2O3
A–C – 9.7 –
A–L – – 2.8
A–L–C – 9.8 2.8
Pt/A 0.7 – –
Pt/A–L 0.8 – 2.7
Pt/A–C 0.7 9.6 –
Pt/A–L–C 0.8 9.9 3.1

decrease in pore volume, confirming the expected partial

blocking of the alumina pores by ceria owing to the
relatively high cerium loading used in the study and the ease
with which CeO2 is aggregated to Al2O3 [19]. Accordingly, Fig. 1. Powder X-ray diffraction patterns of calcined Al2O3 (A), La/Al2O3
the similarity between the specific areas for loaded/unloaded (A–L), Ce/Al2O3 (A–C) and Ce–La/Al2O3 (A–L–C) supports. (~) a-
Al2O3, (*) u-Al2O3, (*) CeO2.
Ce supports can be taken as indicating a certain degree of
thermal stabilization of the support. Notwithstanding,
according to Piras et al. [20] at the calcination temperature
used in the present study, the stabilizing effects of Ce under pattern of the A–L support (A–L in Fig. 1) exhibited lines
oxidizing conditions can be considered negligible. The only attributable to the crystalline u-Al2O3 phase. This
almost constant specific area and the simultaneous pore confirms the above BET results indicating the prevention of
blocking following the addition of ceria suggests that the the thermally induced phase transformations of transition
ceria phase deposited on the alumina support develops some aluminas by the presence of lanthanum on the surface of the
porosity and contributes to the measured specific area. The support. No diffraction peaks corresponding to crystalline
specific area and pore volume data for the Pt-loaded samples species of either La2O3 (JCPDS 83-1355) or LaAlO3
indicate that the textural properties of the bare supports do (JCPDS 85-1071) were observed in the A–L support. This is
not change significantly upon incorporating Pt to the not surprising because lanthanum exists in the form of a two-
supports. The slight decrease in pore volume indicates some dimensional overlayer on alumina, which cannot be detected
filling of the pores by platinum species. by XRD for lanthanum loadings up to 8.5 mmol La/m2 [22].
Besides the Al2O3 phases detected for the corresponding Ce-
3.2. X-ray diffraction (XRD) free supports (A and A–L, respectively), the diffraction
patterns of the Ce-containing supports (A–C and A–L–C in
Fig. 1 shows the X-ray diffraction patterns of the fresh Fig. 1) showed diffraction peaks characteristic of bulk
calcined supports. The characteristic peaks of both crystal- crystalline CeO2 (28.58, 47.58 and 56.38 JCPDS 34-394),
line a-Al2O3 (at 25.68, 43.38 and 57.58 JCPDS 83-2080) and with an average crystallite size of ca. 10 nm for both
u-Al2O3 (at 31.28, 36.68 and 67.48 JCPDS 86-1410) were supports. The appearance of crystalline ceria structures
observed in the thermally treated alumina (A in Fig. 1). This indicates the saturation of the capacity of the alumina to
agrees with the notion that g-aluminas are metastable at high accommodate the ceria loading. This result is consistent
temperatures and transform step-wise into the thermody- with the above BET results and is also in accordance with
namically more stable u-and a-structures [16,21]. The XRD previous observations indicating the greater tendency of La

Table 2
Textural properties of calcined supports and platinum-supported catalysts
BET (m2 g1) BET (m2 g1
Al2 O3 )
Pore volume (cm3/g) Pore diameter (nm)

A 237 0.81 19.8

A-calca 83 83 0.59 21.2
A–L 104 107 0.71 20.1
A–C 80 89 0.49 19.8
A–L–C 100 113 0.61 19.1
Pt/A 83 0.59 21.2
Pt/A–L 108 112 0.69 19.8
Pt/A–C 77 86 0.48 20.7
Pt/A–L–C 101 117 0.60 19.7
a
Calcined in air at 1173 K for 8 h.
 R.M. Navarro et al. / Applied Catalysis B: Environmental 55 (2005) 229–241 233

than Ce to disperse on alumina upon impregnation from absence of the reduction peak at 663 K and the lower
aqueous nitrate solutions [23]. intensity of the reduction peak corresponding to the
The XRD patterns of the platinum-containing samples, formation of CeAlO3 would be the main differences
not presented here, showed the crystalline phases already observed when lanthanum is incorporated on alumina
detected in the Pt-free supports with no detectable phases before cerium loading. The absence of the reduction peak at
due to platinum species. 663 K, ascribed to surface ceria reduction, indicates the
lower degree of dispersion of CeO2 entities on the A–L–C
3.3. Temperature-programmed reduction (TPR) support with respect to the values observed for the A–C
counterpart. In contrast, the lower intensity of the reduction
The results of the TPR experiments performed on the peak associated with CeAlO3 implies a decrease in the
supports are shown in Fig. 2. The TPR profiles correspond- highly dispersed ceria entities generated from strong
ing to both bare and La-loaded alumina (A and A–L in Fig. interactions with alumina. Both these facts indicate that
2) did not show any hydrogen uptake indicative of reduction the presence of lanthanum in the alumina prior to
phenomena. In contrast, the A–C support (Fig. 2) showed impregnation with ceria modifies the Al–Ce interactions
two broad peaks at approximately 663 and 890 K, together observed in the case of bare alumina.
with a sharp peak at 1181 K. This reductive behaviour can be The TPR profiles of the platinum-containing samples are
attributed to changes in the extent and the nature of the depicted in Fig. 3. The reduction profiles of platinum
interaction between the CeO2 and the Al2O3. In accordance species are rather complex and indicate differences in the
with the TPR literature addressing ceria–alumina systems interaction between Pt species and the supports. For
[24,25], the peaks observed can be assigned, respectively, to: platinum deposited on bare alumina (Pt/A in Fig. 3), two
(i) the reduction of the most easily reducible surface oxygen main reduction peaks at 383 and 512 K and a minor peak at
of highly dispersed ceria species, (ii) the removal of oxygen 650 K were observed. The low-temperature peak, only
from bulk ceria and (iii) the reduction of dispersed ceria observed for this platinum-supported sample, can be
species, with the formation of a CeAlO3 phase. The TPR
profile for the A–L–C sample (Fig. 2) displayed a broad
reduction peak at 873 K and a sharp peak at 1173 K that,
according to the previous assignment, corresponded to the
reduction of bulk oxygen anions in CeO2 and to the
formation of CeAlO3 from dispersed ceria entities,
respectively. In comparison with the A–C support, the

Fig. 3. Temperature-programmed reduction profiles for calcined Pt cata-

Fig. 2. Temperature-programmed reduction profiles for calcined Al2O3 (A), lysts supported on Al2O3 (Pt/A), La/Al2O3 (Pt/A–L), Ce/Al2O3 (Pt/A–C)
La/Al2O3 (A–L), Ce/Al2O3 (A–C) and Ce–La/Al2O3 (A–L–C) supports. and Ce–La/Al2O3 (Pt/A–L–C).
234 R.M. Navarro et al. / Applied Catalysis B: Environmental 55 (2005) 229–241

assigned to the reduction of PtO2 species [26] and the peaks Table 3
at 512 and 650 K are attributed to oxychlorinated species in Binding energies (eV) of core electrons and surface atomic ratios of
calcined supports
the bulk and two-dimensional phase, with a strong
Sample Al 2p La 3d5/2 Ce 3d5/2 u000 (% in Ce 3d) La/Al at Ce/Al at
interaction with alumina, respectively [26–28]. The Pt/
A–L sample (Fig. 3) shows only H2-consumption peaks, at A–L 73.9 835.2 0.019
A–C 73.9 882.9 12.1 0.053
477 and 652 K, corresponding to the reduction of the above
A–L–C 74.0 835.6 882.9 12.8 0.011 0.040
oxychloride platinum species. The intensity of the reduction
peak at 652 K indicates the high concentration of PtOxCly
precursors in close contact with the support present in this
Pt/A–L sample. In contrast, the reduction profile associated (834.3 eV) and by Haack et al. [30] for La2O3 (833.2 eV)
with platinum species on ceria-containing supports (Pt/A–C and LaAlO3 (833.8 eV). The observed binding energies of
and Pt/A–L–C in Fig. 3) exhibited a major peak at 440– La 3d5/2 for both the A–L and A–L–C supports are close to
450 K and a small peak at 580 K, pointing to the presence of the values reported in those studies for well-dispersed
a high proportion of bulk oxidized platinum precursors. lanthanum species on alumina. This observation confirms
Besides the reduction peaks at 906 and 1160 K observed for the strong tendency of La to become dispersed on alumina,
the A–C support and assigned to the formation of CeAlO3 as observed by XRD. The homogeneous dispersion of
and Ce2O3 from CeO2 species, the Pt/A–C sample showed lanthanum on the alumina surface observed for the A–L
three reduction peaks at 468, 484 and 591 K. These support is also corroborated by the close La/Al atomic ratios
reduction peaks include both platinum species and partial obtained from XPS with respect to the predicted value
surface reduction of ceria species, since the measured value (0.010) derived from their nominal composition. The surface
of hydrogen consumption was higher than the amount XPS La/Al ratio for the A–L–C support was substantially
expected for the complete reduction of platinum precursors lower than in the A–L sample, indicating preferential
to Pt0. The reduction of platinum species (peak at 468 K) coverage of the lanthanum with ceria or an increase in the
and the concomitant ceria surface reduction (peak at 484 K) particle size of the lanthanum phases. The Ce 3d spectra of
indicate a close interaction between Pt and ceria surface the cerium-containing supports consisted of several peaks
phases, in agreement with previous findings [24]. In the case resulting from either multiplet splitting [31] or shake-up
of the sample supported on alumina containing both processes [32]. The isolated peak in the Ce 3d spectra at ca.
lanthana and ceria (Pt/A–L–C profile in Fig. 3), the presence 917 eV (u000 using the nomenclature of Burroughs [33]) was
of a small shoulder at 486 K, corresponding to the used as identification of Ce4+ species and its relative peak
aforementioned platinum-induced surface ceria reduction, area contribution to total Ce 3d area as a quantitative
indicates the lower number of Pt–CeO2 interactive species. measure of the amount of Ce4+. The fraction of the total Ce
The reduction profile corresponding to Pt/A–L–C also signal attributable to the u000 peak in the A–C sample (12.1%,
showed a broad peak consumption at temperatures higher Table 3) is lower than the 13.4% reported for pure, fully
than 700 K that included the reduction of bulk cerium oxidized ceria [34], implying the presence of Ce3+ species
oxides already detected in the free-Pt support. However, the (ca. 10%). This is consistent with the findings of other
formation of CeAlO3 (around 1150 K), also detected in the studies [24] indicating the presence of Ce3+ in cerium phases
A–L–C support, was not evident in the TPR curve of the Pt/ supported on alumina prior to any reduction treatment. Some
A–L–C sample. authors [34,35] have proposed the formation of CeAlO3 to
explain the existence of these Ce3+ species. However, taking
3.4. X-ray photoelectron spectroscopy (XPS) into account the relatively low temperature used in the
calcination of the A–C support, the presence of these Ce3+
The chemical species present on the catalyst surfaces and species was probably due to highly dispersed ceria entities in
their proportions were evaluated by XPS. The energy close contact with the alumina support. Previous TPR
regions of Al 2p, La 3d, Ce 3d and Pt 4d5/2 core levels in analyses performed on this A–C support corroborate this
oxidised and reduced samples were recorded. Although the interpretation. For the A–L–C support, the percentage
most intense photoemission lines of platinum were those contribution of u000 to the total Ce 3d area was higher (12.8%,
arising from Pt 4f levels, this energy region became Table 3), thus pointing to a lower concentration of Ce ions in
overshadowed by the presence of a very strong Al 2p peak. the 3+ state associated, as stated above, with ceria entities in
Consequently, the energy region of the less intense Pt 4d close contact with the alumina support. In the A–C support,
peak was recorded for all the Pt-containing samples. The the calculated XPS Ce/Al atomic ratio was higher than the
binding energies of core electrons and the surface atomic expected value derived from its chemical composition
ratios of the calcined supports are summarized in Table 3. (0.033), indicating the presence of an overlayer of cerium
For La-containing aluminas, the La 3d5/2 peak envelopes species. The value of this XPS Ce/Al ratio decreased for the
were centred at 835.2 eV for the A–L support, and at A–L–C support, indicating the lower degree of surface
835.6 eV for the A–L–C support. These binding energies are concentration of cerium oxide particles present in this
higher than the values reported by Chen et al. [29] for La2O3 sample. As in the case of the A–C sample, the higher value
 R.M. Navarro et al. / Applied Catalysis B: Environmental 55 (2005) 229–241 235

of the surface Ce/Al ratio for the A–L–C sample with respect covering of the alumina support by Pt species can be ruled
to the bulk Ce/Al proportion indicates the presence of out. The disagreement between the XPS and theoretical Pt/
aggregates of ceria. Al ratios may be the result of error in the area measurement
The fresh calcined samples containing platinum showed a of the Pt 4d5/2 peaks due to the low signal-to-noise ratio in
broad and asymmetric Pt 4d5/2 peak, which could be this region. In spite of this, the variations in the calculated
resolved, after curve fitting procedures, into two components XPS Pt/Al ratios point to changes in surface concentration
with binding energies of 318.3–318.8 and 315.4–316.0 eV occurring in the samples studied. When platinum was
(Fig. 4). Both components can be ascribed to Pt4+ species in deposited on the A–L support, the lower percentage of the
oxide [36] or in incompletely decomposed platinum– component at high BE, together with the shift of the second
chlorine-containing species [37] with different interactions component to a lower BE (315.4 eV in Table 4), point to the
with the support. Higher binding energies have been stronger interaction of the platinum precursors and the
observed for small clusters of Pt precursors interacting support containing lanthanum. As a consequence of this
strongly with supports [38]. The peak percentages corre- interaction, the Pt/Al XPS ratio calculated for this sample
sponding to the above two components in the Pt 4d5/2 level, (Pt/A–L in Table 4) indicates that the dispersion of platinum
summarised in Table 4, indicate that Pt supported on bare species is lower than that obtained on the bare alumina. As a
alumina has the higher proportion of Pt4+ species consequence of the low Pt surface coating, the calculated
(component at 318 eV) interacting weakly with the alumina. XPS atomic La/Al ratio for this sample did not differ from
The calculated XPS Pt/Al surface ratio for this sample, also the values corresponding to the calcined support. The Pt 4d5/
summarized in Table 4, is somewhat higher than the 2 BE values of the samples containing ceria shifted—
expected value derived from its chemical composition (Pt/Al especially in the case of Pt/A–C—to lower values (317.5 and
= 0.0021). This may be indicative of a platinum enrichment 314.7 eV, Table 4). The observed shift in binding energies
on the external surface of the support particles. However, may be the consequence of either the smaller size of the
taking into account the low metal loading used, the possible particles of platinum precursors or the strong interaction
with the cerium phase, leading to a profound electronic
polarization of the clusters of platinum precursors [39].
Electronic perturbation associated with changes in particle
sizes can be ruled out because the calculated XPS Pt/Al
atomic ratio for this sample was not lower than those
obtained for the Pt/A–L sample. The Pt/A–L–C sample
showed a more moderate shift in Pt 4d5/2 peaks (318.8 and
316.0 eV) than those observed for the Pt/A–C sample, thus
indicating a diminished Pt–Ce interaction with respect to
that observed in the Pt/A–C sample.
The chemical changes in the catalyst after reduction in H2
at 573 K were also investigated by XPS. As shown in Fig. 5,
the Pt 4d5/2 peaks for the Pt/A, Pt/A–L, Pt/A–C and Pt/A–L–
C samples shifted, after the reduction treatment, to binding
energies of 316.1–317.2 and 314.0–314.3 eV. According to
the data in the literature, the component at higher BE is
associated with Pt particles in an electron-deficient state,
whereas those at lower binding energies indicate the
presence of Pt0 species [40]. As seen in Table 4, the sample
supported on bare alumina and lanthanum modified
aluminas had the higher proportion of platinum species
maintaining some d+ character. This may be the consequence
of a strong Al–Pt–O interaction between dispersed platinum
species and the Al2O3 support, which has been reported to
stabilise surface Pt in an oxidised state [41]. In contrast, the
sample supported on A–C showed a higher proportion of
metallic Pt after reduction. For this sample, a small shift in
the BE of the peak associates with Pt in the electron-
deficient state was observed with respect to that of the Pt/A
sample. Some authors [42] have interpreted this shift as
Fig. 4. Pt 4d5/2 core-level spectra of calcined Pt catalysts supported on being due to the existence of Pt–Ce interactions. After
Al2O3 (Pt/A), La/Al2O3 (Pt/A–L), Ce/Al2O3 (Pt/A–C) and Ce–La/Al2O3 reduction, the XPS Pt/Al ratios were lower than those
(Pt/A–L–C). obtained from the fresh calcined samples. This points to a
236 R.M. Navarro et al. / Applied Catalysis B: Environmental 55 (2005) 229–241

Table 4
Binding energies (eV) of core electrons and surface atomic ratios of calcined and reduced Pt-supported catalysts
Sample Pt 4d5/2 La 3d5/2 Ce 3d5/2 (% u000 in Ce3d) Pt/Al at La/Al at Ce/Al at
Pt-A
Fresh 318.8 (58.5), 316.4 (41.5) 0.0091
Reduced 317.2 (45.4), 314.3 (54.6) 0.0068
Pt/A–L
Fresh 318.3 (52.4), 315.4 (47.6) 835.4 0.0076 0.020
Reduced 316.9 (44.4), 314.0 (55.6) 835.1 0.0059 0.018
Pt/A–C
Fresh 317.5 (53.4), 314.7 (46.6) 882.6 (12.2) 0.0073 0.050
Reduced 316.1 (39.7), 314.0 (60.3) 881.8 (9.5) 0.0055 0.053
Pt/A–L–C
Fresh 318.8 (29.1), 316.0 (70.9) 835.6 882.5 (12.1) 0.0048 0.014 0.036
Reduced 316.5 (51.7), 314.0 (48.3) 835.9 882.0 (11.8) 0.0045 0.016 0.046

transformation in particle size or shape of the platinum percentages corresponding to the Pt0 state, the surface
precursors during the reduction process. The degree of concentration of metallic platinum on reduced samples
platinum sintering during reduction depended on the varied in the following order: Pt/A > Pt/A–L Pt/A–C >
support, the sample supported on bare alumina showing Pt/A–L–C. The changes in the chemical state of the Ce and
greatest loss of surface platinum during reduction while in La promoters after reduction were also monitored by XPS.
the case of the sample supported on A–L–C this loss was After the H2 treatment, the binding energies of La did not
negligible. Combining the XPS Pt/Al ratios and the peak change with respect to the fresh counterparts, indicating that
these species were scarcely affected by reduction at 623 K.
Treatment under hydrogen, however, produced changes in
the Ce state, as indicated by the XPS spectra of the Ce 3d
region acquired from the Pt/A–C and Pt/A–L–C samples
when calcined and after reduction (Fig. 6). Reduction under
H2 at 573 K resulted in Ce species with a u000 percentage in
the total Ce 3d region, indicating the reduction of Ce4+ to
Ce3+. As shown in Table 4, a substantial reduction of cerium
occurred in both samples, being especially important in the
case of the Pt/A–C catalyst (u000 percentage equivalent to
30% of Ce3+ species). The increase in the degree of
reduction of ceria as compared to the support alone
indicates, as previously observed by TPR, that platinum
facilitates ceria surface reduction. This phenomenon is well
documented in the literature [43] and implies a close contact
between metallic Pt and ceria surfaces. In the case of the Pt/
A–C–L sample, its higher u000 percentage in the total Ce 3d
region indicates a lower degree of ceria reduction (ca. 14%
of Ce3+ species). This lower ceria reduction corroborates the
lower degree of Pt–Ce interaction already indicated in the
above analysis of Pt BE as well as from the TPR profiles
obtained for this sample.

3.5. Catalytic activity

To evaluate the contribution of the thermal decomposi-

tion of ethanol, first the homogeneous (non-catalytic)
oxidative reforming of ethanol was investigated. Fig. 7
and Table 5 show the conversion and product selectivities,
respectively, corresponding to this blank experiment. As
Fig. 5. Pt 4d5/2 core-level spectra of reduced Pt catalysts supported on shown in Table 5, ethanol was thermally converted to a
Al2O3 (Pt/A), La/Al2O3 (Pt/A–L), Ce/Al2O3 (Pt/A–C) and Ce–La/Al2O3 considerable extent in the 848–948 K temperature range.
(Pt/A–L–C). The main products observed from the thermal conversion of
 R.M. Navarro et al. / Applied Catalysis B: Environmental 55 (2005) 229–241 237

Fig. 7. Temperature evolution of ethanol conversion (*) and reaction

products: (!) CO, (&) H2, (~) CH4, (&) CH3CHO, (*) C2H4, (^)
C2H6, (*) CO2 during the homogeneous (non-catalytic) oxidative reform-
ing of ethanol.

as indicated in Fig. 8, in the order: Pt/A–C > Pt/A–L–C >

Pt/A > Pt/A–L. Significant differences in product selectiv-
ities with the type of support used were also found (Fig. 9).
Fig. 6. Ce 3d core-level spectra of Pt/A–C and Pt/A–L–C catalysts after:
calcination (a) and reduction (b).
Similarly to homogeneous oxidative reforming, the reaction
of ethanol over the platinum alumina-supported catalysts
proceeded through dehydrogenation and dehydration path-
ethanol were CO, H2, CH4, CH3CHO, C2H4 and C2H6. The ways, as suggested by the presence of acetaldehyde and
presence of acetaldehyde, ethene and ethane suggests the
participation of dehydrogenation and dehydration processes.
Besides these primary ethanol reactions, secondary reac-
tions comprising the oxidation of hydrogen and the
decomposition of acetaldehyde to CO and CH4 are also
involved in view of the evolution of the concentrations of
CO, CH4, and H2 observed in the product stream (Fig. 7). It
is interesting to note the low level of total oxidation of
ethanol as revealed by the low CO2 selectivity observed in
products. Ethanol conversion over platinum catalysts
depended strongly on the type of support used increasing,

Table 5
Ethanol conversion and hydrogen yield as function of temperature in
homogeneous (non-catalytic) reaction and over Pt-supported catalysts
Ethanol conversion (%) H2 yield (NL/h)
848 K 898 K 948 K 848 K 898 K 948 K
Blank 63.8 62.8 69.5 0.67 0.69 0.79
Pt/A 64.8 63.4 73.3 0.65 0.96 1.66
Pt/A–L 65 63.1 69.1 0.69 1.06 1.86
Fig. 8. Ethanol conversion rate as function of temperature during the
Pt/A–C 66.9 72.9 94.8 1.22 2.72 5.36
oxidative reforming of ethanol over Pt/A, Pt/A–L, Pt/A–C and Pt/A–L–
Pt/A–L–C 65.6 68.1 87.2 0.96 1.56 2.58
C catalysts.
238 R.M. Navarro et al. / Applied Catalysis B: Environmental 55 (2005) 229–241

Fig. 9. Temperature evolution of ethanol conversion (*) and reaction products: (!) CO, (&) H2, (~) CH4, (&) CH3CHO, (*) C2H4, (*) CO2 during the
oxidative reforming of ethanol over: Pt/A (a), Pt/A–L (b), Pt/A–C (c) and Pt/A–L–C (d) catalysts.

ethene in the products (Fig. 9a). Comparison of the evolution distributions while maintaining a fraction of unconverted
of product selectivities on the Pt/A sample with respect to O2 in the gas phase. The selectivity patterns of the catalyst
those observed for the homogeneous reaction indicates that using the A–L support are shown in Fig. 9b. The product
the presence of platinum does not significantly promote distribution obtained for this Pt/A–L sample was similar to
ethanol bond breaking, following the sequence ethoxy ! that obtained for the Pt/A sample. This shows that similar
acetaldehyde ! CH4 + CO suggested by mechanistic studies reaction pathways operate in the ethanol-reforming reaction
of ethanol on Pt surfaces [44,45]. The dehydration pathway over both platinum-supported samples. The product
towards ethylene formation, possible on alumina surfaces distribution obtained over the platinum deposited on
[46], was also not favoured on the Pt/A sample. Therefore, ceria–alumina is depicted in Fig. 9c. The selectivity patterns
the higher H2 and CO2 selectivities, with a concomitant of the Pt/A–C sample were different from those described
decrease in CH4 and CO, observed on the Pt/A sample before corresponding to the Pt/A and Pt/A–L samples. The
indicate that the presence of this catalyst mainly promotes reaction on Pt/A–C showed high initial selectivity towards
water-reforming reactions over CO and, to a lesser extent, H2 and CO2. The increase in temperature was accompanied
over CH4, both derived from the breakage of ethanol with an by an increase in the selectivity of H2 and CO at the expense
important contribution of the thermal homogeneous path. of CH4 and CO2. This evolution of products suggests the
Under the temperature range studied in the present work, participation of the reverse water-gas shift upon increasing
total oxygen consumption was achieved. Therefore, the role the temperature. From the above observations it is clear that
of oxygen in the reaction mechanism cannot be established the presence of the Pt/A–C catalyst seems to promote the
because the inability to observe changes in product adsorption–decomposition of ethanol to CH4 and CO and
 R.M. Navarro et al. / Applied Catalysis B: Environmental 55 (2005) 229–241 239

reforming with water over these latter products. Ethanol reduced reactivity associated with Pt species strongly
adsorption via the formation of ethoxy species has been interacting with the support may be responsible for the
observed on unreduced CeO2 surfaces [12] and the observed differences in activity and hydrogen selectivity
promotion of platinum by ceria for the WGS reaction has between the Pt/A and Pt/A–L samples. Comparison of the
also been described [13]. The Pt/A–L–C catalyst (Pt/A–C– catalytic activities of Pt/A–C and Pt/A–L shows that the
L, Fig. 9d) followed a similar reaction behaviour to that addition of cerium strongly enhances both ethanol conver-
observed for the ceria-modified support. However, the Pt/A– sion and selectivity towards H2. Taking into account that
L–C catalyst achieved lower H2 and CO2 selectivities, both samples had similar XPS surface Pt/Al ratios, the
indicating a lower CH4 and CO water reforming activity than variation in activity associated with changes in particle size
that observed for the ceria–alumina sample. suggested above cannot be the reason for the observed
differences in activity. The XPS spectra of the Pt/A–C
3.6. Catalyst structure and activity correlations sample indicate a strong interaction of Pt with cerium, which
is absent in the Pt/A–L sample. The TPR profile of the Pt/A–
It is clear that the addition of cerium and/or lanthanum C sample also indicates the existence of a direct Pt–CeO2
modifies the base alumina support at different levels. As may interaction upon reduction. The intimate contact between Pt
be seen from the BET and XRD analyses, lanthanum and cerium may be the cause of the promoter effect observed
interacted strongly with alumina increasing its thermal in the Pt/A–C catalyst. Why the Pt–Ce interaction enhances
resistance to sintering. The incorporation of La3+ to alumina hydrogen production from ethanol is a key question. Study
before cerium loading led to a decrease in both the of the reaction pathway of ethanol decomposition over the
interaction between cerium ions and alumina and in the Pt/CeO2 system has shown [12] that Pt–Ce interactions
dispersion of the ceria entities relative to those achieved on contribute to the efficient carbon–carbon bond dissociation
bare alumina. The above modifications in the properties of of ethoxy species adsorbed on its surfaces. In addition, the
the alumina support due to the addition of Ce and/or La beneficial effect observed on the Pt/A–C sample was also
directly affected the dispersion and stability of the deposited associated, as indicated in Fig. 9c, with an enhanced
platinum species. The XPS and TPR results showed that the reforming capacity of CO and CH4 formed during the
surface concentration of metallic platinum on the reduced reaction of ethanol on the catalyst surface. It is known [50]
catalysts varied in the following order: Pt/A > Pt/A–L > Pt/ that Pt/CeO2 systems promote water-gas shift and steam
A–C > Pt/A–L–C. Clearly, the presence of platinum is an reforming reactions. Although there is no complete
important parameter for the catalytic performance, as consensus about the role of ceria and the metal as promoters
indicated by the higher hydrogen production obtained over for the above reactions, it appears that the improvements in
the Pt-containing catalyst in comparison with that corre- reforming capacity are related to ceria-mediated redox
sponding to the homogeneous thermal reaction. However, on process in which CO/CH4 adsorbed on Pt reduces the ceria
comparing the above sequence of Pt dispersion with the surface to generate CO2, and water reoxidizes the ceria
order of intrinsic catalytic activity of these samples derived surface to CeO2, releasing hydrogen. As discussed above,
from Fig. 8, it is evident that properties other than the the TPR experiments carried out in the current study
number of metallic platinum sites affect the catalytic activity revealed that the intimate Pt–Ce contact promoted the
of the samples when ceria and/or lanthanum are present. The reduction of surface CeO2 by H2. In keeping with the above
introduction of lanthanum to the alumina did not promote explanation, the poorer catalytic behaviour of the Pt/A–L–C
the degree of ethanol conversion with respect to that catalyst may thus be related to the lower extent of the Pt–Ce
obtained with bare alumina but increased selectivity towards interaction. The TPR profile of this sample revealed that Pt
hydrogen. Changes in activity derived from changes in the reduced the surface of CeO2 to a lesser extent. XPS analysis
acidity of Al2O3 due to the incorporation of lanthanum can of the Pt 4d5/2 level of the Pt/A–L–C sample also
be ruled out, since the selectivity to dehydration product, corroborates the decrease in the Pt–Ce interaction and the
ethylene, did not change (see Fig. 9a and c). Taking into lower surface reduction of cerium phases. Both the XPS and
account that the XRD and XPS analyses indicated the TPR analyses revealed that the presence of lanthanum on
absence of La2O3, which according to the data in the alumina blocked the surface anchoring sites for cerium ions,
literature promotes ethanol adsorption via the dehydrogena- leading to a drop in ceria dispersion. Therefore, taking into
tion reaction [47], the observed changes in activity/ account that Pt preferentially interacts with Ce even in the
selectivity on Pt/A–L may be due to changes in the atomic presence of a much larger Al2O3 surface [51], the lower Pt–
structure of Pt. The XPS spectra and TPR profile of Pt/A–L Ce interaction observed for the Pt/A–L–C sample may be
agree with this, since they indicate changes in the Pt–support related to the lower extent of interactive surfaces of ceria
interaction and a lower platinum dispersion on the Pt/A–L able to interact with Pt. In spite of the lower interaction
catalyst with respect to those observed in the Pt/A sample. between Pt and Ce observed for this sample, as stated above
Changes in either the relative abundance of different these surface interactions play a key role in the promotion of
crystallographic Pt metal planes with different reactivity ethanol-reforming reaction, as indicated by the important
[48,49] associated with the increase in particle size or a improvement in activity for water-reforming reactions over
240 R.M. Navarro et al. / Applied Catalysis B: Environmental 55 (2005) 229–241

CH4 and CO observed when Ce was present in samples References

containing La (Pt/A–L–C, Fig. 9d) with respect to that
obtained by the counterpart without Ce in the support (Pt/A– [1] A.N. Fatsikostas, D.I. Kondarides, X.E. Verykios, Catal. Today 75
L, Fig. 9b). (2002) 145.
In sum, the observed improvements in hydrogen yield [2] S. Freni, S. Cavallaro, N. Mondello, L. Spadaro, F. Frusteri, J. Power
Sour. 108 (1–2) (2002) 53.
are closely related to the increase in the capacity for water
[3] F. Haga, T. Nakajima, H.B. Miya, S. Mishima, Catal. Lett. 48 (1997)
reforming of the CO and CH4 products arising from ethanol 223.
carbon–carbon bonds associated with the presence of Pt– [4] J. Llorca, N. Homs, J. Sales, P. Ramı́rez de la Piscina, J. Catal. 209
Ce interactions in the catalysts. The question as to how (2002) 306.
these Pt–Ce interactions lead to higher catalytic activities [5] F. Mariño, M. Boveri, G. Baronetti, M. Laborde, Int. J. Hydr. Energy
26 (2001) 665.
in ethanol reforming remain obscure and will require
[6] V. Fierro, V. Klouz, O. Akdim, C. Mirodatos, Catal. Today 75 (2002)
further investigation. 141.
[7] J.P. Breen, R. Burch, H.M. Coleman, Appl. Catal. B: Environ. 39
(2002) 65.
[8] D.K. Liguras, D.I. Kondarides, X.E. Verykios, Appl. Catal. B:
4. Conclusions
Environ. 43 (4) (2003) 345.
[9] V. Fierro, O. Akdim, C. Mirodatos, Green Chem. 5 (2003) 20.
The oxidative reforming of ethanol over Pt catalysts [10] J.H. Sinfelt, Adv. Catal. 23 (1973) 91.
supported on alumina modified by Ce and/or La has been [11] D.C. Grenoble, M.M. Estat, D.F. Ollis, J. Catal. 67 (1981) 90.
studied. Ethanol-reforming reactions over platinum-sup- [12] A. Yee, S.J. Morrison, H. Idriss, J. Catal. 191 (2000) 30.
[13] T. Bunluesin, R.J. Gorte, G.W. Graham, Appl. Catal. B: Environ. 15
ported on metal oxides include several pathways that
(1988) 107.
contribute to the overall reaction. On the basis of the activity [14] A. Trovarelli, Catal. Rev.-Sci. Eng. 38 (1996) 439.
results, the catalytic behaviour of supported Pt catalysts [15] J. van Doorn, J. Varloud, P. Meriaudeau, V. Pericón, M. Chevrier, C.
towards the production of hydrogen from ethanol is strongly Gauthier, Appl. Catal. B: Environ. 1 (1992) 117.
promoted by the presence of Ce on the support, in contrast [16] H. Schaper, E.B.M. Doesburg, L.L. Van Reijen, Appl. Catal. 7 (1983)
211.
with the presence of lanthanum, which did not promote the
[17] P. Burtin, J.P. Brunelle, M. Pijolat, M. Soustelle, Appl. Catal. 34
degree of ethanol conversion. The poorer performance of (1987) 225.
the La-containing sample is associated with the lower Pt [18] M. Bettman, R.E. Chase, K. Otto, W.H. Weber, J. Catal. 117 (1989)
dispersion achieved on this sample. The use of cerium as an 447.
additive was found to be beneficial for the production of [19] C. Morterra, G. Magnacca, V. Bolis, G. Cerrato, M. Barico, A.
Giachello, M. Fucale, Stud. Surf. Sci. Catal. 96 (1995) 361.
hydrogen from oxidative ethanol reforming. On the basis of
[20] A. Piras, A. Trovarelli, G. Dolcetti, Appl. Catal. B: Environ. 28 (2000)
the TPR and XPS analyses, it is apparent that Pt and CeO2 L77–L81.
interact strongly after calcination at 773 K. The interaction [21] B.C. Lippens, J.H. de Boer, Acta Crystalogr. 17 (1964) 1312.
between Pt and cerium is proposed as being responsible for [22] M. Bettman, R.E. Chase, K. Otto, W.H. Weber, J. Catal. 117 (1989)
the better activity observed for samples containing cerium. 447.
[23] G.W. Graham, P.J. Scmitz, R.K. Usmen, R.W. McCabe, Catal. Lett. 17
It is suggested that the Pt–Ce interaction affects the
(1993) 175.
adsorption–decomposition of ethanol to CH4 and CO and [24] J.Z. Shyu, W.H. Weber, H.S. Gandhi, J. Phys. Chem. 92 (1988) 4964.
the reforming of these products with water. The combina- [25] H.C. Yao, Y.F. Yao, J. Catal. 86 (1984) 254.
tion of both effects is interpreted as being responsible for the [26] C. Hwang, C. Yeh, J. Mol. Catal. A: Chem. 112 (1996) 295.
higher hydrogen production observed over the cerium- [27] H. Lieske, G. Lietz, H. Spindler, J. Wölter, J. Catal. 81 (1983) 8.
[28] G. Lietz, H. Lieske, H. Spindler, W. Hamker, J. Wölter, J. Catal. 81
containing sample. When cerium and lanthanum were
(1983) 17.
present in the support, a poorer catalytic behaviour was [29] X. Chen, Y. Liu, G. Niu, Z. Yang, M. Bian, A. He, Appl. Catal. A: Gen.
observed as a consequence of the lower Pt–Ce interaction 205 (2001) 159.
derived from the decrease in surface ceria dispersion with [30] L.P. Haack, J.E. de Vries, K. Otto, M.S. Chatta, Appl. Catal. A: Gen.
respect to that achieved on the lanthanum-free ceria– 82 (1992) 199.
[31] G. Thornton, M.J. Dempsey, Chem. Phys. Lett. 77 (1981) 409.
alumina support.
[32] A. Fujimori, A. Fujimori, J. Magn. Magn. Mater. 47–48 (1985) 243.
[33] P. Burroughs, A. Hammet, A.F. Orchard, G. Thornton, J. Chem. Soc.,
Dalton Trans. 17 (1976) 1686.
Acknowledgement [34] J.Z. Shyu, K. Otto, W.L.H. Watkings, G.W. Graham, R.K. Belitz, M.S.
Gandi, J. Catal. 114 (1988) 23.
This work was carried out within the framework of the [35] P.J. Schmitz, R.K. Usmen, C.R. Peters, G.W. Graham, R.W. McCabe,
Project ‘‘Development of 5 kW fuel processors’’ carried Appl. Surf. Sci. 72 (1993) 181.
out by the Instituto de Técnica Aeroespacial (INTA), [36] R. Bouwman, P. Biloen, J. Catal. 48 (1977) 209.
[37] H. Karhu, A. Kalantar, I.J. Väyrynen, T. Salmi, D. Yu Murzin, Appl.
Centro de Investigación y Desarrollo en Automoción Catal. A: Gen. 247 (2003) 283.
(CIDAUT) and the Instituto de Catálisis y Petroleoquı́mica [38] J. Goetz, M.A. Volpe, A.M. Sica, C.E. Gigola, R. Toroude, J. Catal.
(CSIC). RMN is grateful to MCYT for a Ramon y Cajal 153 (1995) 6.
research programme. [39] J.C. Summers, S.A. Ausen, J. Catal. 58 (1979) 131.
 R.M. Navarro et al. / Applied Catalysis B: Environmental 55 (2005) 229–241 241

[40] J.M. Thomas, W.J. Thomas, Principles and Practice of [45] Y. Cong, V. van Spaendonk, R.I. Masel, Surf. Sci. 385 (1997) 246.
Heterogeneous Catalysis, VCH, Verlag Gesellschaft mbH, Weinheim, [46] H.H. Kung, Stud. Surf. Sci. Catal. 45 (1989) 146.
Germany, 1997. [47] M.P. Rosynek, R.J. Koprowski, G.N. Dellisante, J. Catal. 122 (1990)
[41] J.Z. Shyu, K. Otto, Appl. Surf. Sci. 32 (1988) 246. 80.
[42] C. Serre, F. Garin, G. Belot, G. Maire, J. Catal. 141 (1993) 1. [48] B.A. Sexon, K.D. Rendulic, A.E. Hughes, Surf. Sci. 121 (1982) 181.
[43] S. Bernal, J.J. Calvino, G.A. Cifredo, J.M. Rodrı́guez Izquierdo, V. [49] Y. Cong, V. van Spaendonk, R.I. Masel, Surf. Sci. 395 (1998) 248.
Perrinchon, A. Laachir, J. Catal. 137 (1992) 1. [50] G. Jacobs, P.M. Patterson, U.M. Graham, D.E. Sparks, B.H. Davis,
[44] A.F. Lee, D.E. Gawthrope, N.J. Hart, K. Wilson, Surf. Sci. 548 (2004) Appl. Catal. A: Gen. 269 (1–2) (2004) 63.
200. [51] J.Z. Shyu, K. Otto, J. Catal. 115 (1989) 16.