Selective Production of Hydrogen by Partial Oxidation of Methanol over ZnO-Supported Palladium Catalysts

Article abstract of DOI:10.1006/jcat.1998.2184

Hydrogen production by partial oxidation of methanol (CH₃OH + 1/2 O₂ ? 2 H₂ + CO₂) was studied over zinc oxide-supported Pd catalysts. Catalyst performance was investigated as a function of the Pd loading and pretreatment in hydrogen environment under feed ratios O₂/CH₃OH (molar) of 0.3 and 0.5 at 503-543 K. High yields of hydrogen were obtained under integral operation regime. The hydrogen selectivity showed a strong dependence on the methanol conversion, which suggests that oxidation and reforming steps take place consecutively. Catalyst characterization by temperature-programmed reduction, X-ray diffraction, and X-ray photoelectron spectroscopy revealed that PdZn alloys can be formed upon reduction at moderate temperatures. A shift of +0.7 eV has been observed in the binding energy of Pd 3d_5/2 core level spectrum of catalyst 1% Pd/ZnO prereduced at temperatures as low as 373 K and catalyst 5% Pd/ZnO reduced at 573 K and above. The relative ease with which PdZn alloy is formed in 1% Pd/ZnO catalyst is explained in terms of a stronger metal-metal oxide interactions of the smaller Pd particles. PdZn alloys were also detected by X-ray diffraction on the 2% Pd/ZnO catalyst after on-stream operation. Pretreatments of the catalysts in hydrogen at high temperatures led to sintering of metallic particles with the subsequent drop in methanol conversion.

Full text of DOI:10.1006/jcat.1998.2184

JOURNAL OF CATALYSIS 179, 150–162 (1998)
ARTICLE NO. CA982184
Selective Production of Hydrogen by Partial Oxidation of Methanol
over ZnO-Supported Palladium Catalysts
,1
ꢀ
M. L. Cubeiro and J. L. G. Fierro†
ꢀ
Centro de Cata´lisis, Petro´leo y Petroqu´ımica, Escuela de Qu´ımica, Facultad de Ciencias, Universidad Central de Venezuela, A.P. 47102, Los
Chaguaramos, Caracas, Venezuela; and †Instituto de Cata´lisis y Petroleoqu´ımica, CSIC, Campus UAM, Cantoblanco, 28049 Madrid, Spain
Received September 9, 1997; revised October 24, 1997; accepted October 29, 1997
be achieved, in comparison with the 25–30% typical of a
normal petrol fuelled engine (7).
Hydrogen production by partial oxidation of methanol
→
(CH3OH + 1/2 O2 2 H + CO ) was studied over zinc oxide-
The second option are fuel cells. Fuel cells are attrac-
tive in terms of potentially high energy conversion efficien-
cies, and for these hydrogen is orders of magnitude more
reactive than any other fuel (8). The major impediment
to the wider use of hydrogen as an energy source are the
difficulties inherent in storage, distribution and transporta-
tion. One solution to the storage problem is the on-board
hydrogen generation from a suitable high energy density
fuel such as methanol. Hydrogen can be obtained directly
from methanol according to three different processes: par-
tial oxidation, thermal decomposition (4,9) and steam re-
forming. When used in engine and vehicle technologies,
methanol partial oxidation offers some advantages in re-
spect to methanol steam reforming (1,6) as it requires oxy-
gen (air) instead water, and it is an exothermic reaction
which does not require heat supply.
ꢁ
2
2
supported Pd catalysts. Catalyst performance was investigated as
a function of the Pd loading and pretreatment in hydrogen envi-
ronment under feed ratios O2/CH3OH (molar) of 0.3 and 0.5 at
503–543 K. High yields of hydrogen were obtained under integral
operation regime. The hydrogen selectivity showed a strong depen-
dence on the methanol conversion, which suggests that oxidation
and reforming steps take place consecutively. Catalyst character-
ization by temperature-programmed reduction, X-ray diffraction,
and X-ray photoelectron spectroscopy revealed that PdZn alloys
can be formed upon reduction at moderate temperatures. A shift of
+0.7 eV has been observed in the binding energy of Pd 3d5/2 core
level spectrum of catalyst 1% Pd/ZnO prereduced at temperatures
as low as 373 K and catalyst 5% Pd/ZnO reduced at 573 K and
above. The relative ease with which PdZn alloy is formed in 1%
Pd/ZnO catalyst is explained in terms of a stronger metal–metal
oxide interactions of the smaller Pd particles. PdZn alloys were
also detected by X-ray diffraction on the 2% Pd/ZnO catalyst af-
ter on-stream operation. Pretreatments of the catalysts in hydrogen
at high temperatures led to sintering of metallic particles with the
c
Partial oxidation of methanol with oxygen or air (reac-
tion [1]) and steam reforming (reaction [2]), under specific
reaction conditions produces almost quantitatively H2 and
CO2:
subsequent drop in methanol conversion.
ꢂ 1998 Academic Press
→
ꢁ
CH3OH + 1/2 O2
CH OH + H O
2 H + CO
[1]
[2]
2
2
INTRODUCTION
→
ꢁ
3 H + CO .
3
2
2
2
The scarcity of fossil fuels and the associated pollutions
problems during their combustion have attracted the atten-
tion towards the search of alternative fuels (1–3). Hydrogen
gas, whether used directly as a fuel in internal combustion
engines (2,4) or indirectly to supply electricity using fuel
cells (5,6), is inherently clean burning and highly reactive
so that high thermal efficiencies can be obtained. When
used in internal combustion engines the compression ratio
can be increased markedly and because of its wide combus-
tion limits hydrogen can be burnt in mixtures much leaner
than is possible using conventional hydrocarbon/air mix-
tures. Thermal efficiencies in the order 35–40% may thus
While reaction [1] is exothermic, reaction [2] is endother-
mic and produces more favourable H2/CO2 ratios. Cu-ZnO
based methanol synthesis catalysts have been extensively
studied in the steam reforming (5,6,10,11) and to lesser ex-
tent in the partial oxidation reaction (6,12–14). Consider-
ing the different processes involved in the partial oxida-
tion, two options have been considered: (i), combination
of combustion-reforming through consecutive beds of no-
ble metal and copper catalyst, respectively (10); and (ii),
oxygen incorporation during steam reforming over copper
catalysts (12,13).
Group VIII noble metals such as palladium and plat-
inum have been found active in methanol transformation
although they are less selective in the steam reforming,
1 To whom correspondence should be addressed.
150
0021-9517/98 $25.00
c
Copyright ꢂ 1998 by Academic Press
All rights of reproduction in any form reserved.

PARTIAL OXIDATION OF METHANOL
151
yielding primarily the decomposition products CO and H2. Experimental Techniques
Apart from this general trend, the behaviour of Pd-ZnO
The palladium content of the catalysts was determined
by atomic absorption spectrometry using a Perkin Elmer
3030 instrument. The solids were kept in contact with a
HCl + HNO3 mixture in a hermetical container and di-
gested in a microwave oven over 1 h.
catalyst seems to be close to that of the conventional Cu-
ZnO methanol synthesis ones. In this respect, Takezawa
et al. (15,16) reported exceptional performance of Pd-ZnO
catalysts, not only in the steam reforming, but also in the
dehydrogenation of methanol to methyl formate, reactions
otherwise conducted with high selectivities over copper
catalysts. These authors attributed this particular behaviour
of Pd-ZnO systems to the formation of PdZn alloys at the
metal–support interface.
These peculiar characteristics of Pd-based catalysts men-
tioned above, together with our experience on Cu-ZnO
systems (1,14,17,18), prompted us to study the effect of
Pd loading and reaction conditions on the performance of
Pd-ZnO for the partial oxidation of methanol. Therefore,
we prepared ZnO-supported Pd catalysts with Pd loadings
rangingfrom 1to 5% and then tested in the partialoxidation
of methanol at atmospheric pressure. Parallel characteriza-
tion ofcatalyst structureshave alreadybeen performed with
X-ray diffraction, temperature-programmed reduction and
photoelectron spectroscopy.
Powder X-ray diffraction (XRD) patterns of the sam-
ples were recorded on a Seifert 3000P diffractometer with a
nickel-filtered CuKꢀ1 (ꢁ = 0.15406 nm) radiation scanning
2ꢂ angles ranging from 5 to 75ꢄ. Thermogravimetric analysis
of the Zn-oxalate precursor was performed with a Perkin
Elmer TGA 7 instrument working at a heating rate of
5 K/min and under air flow (60 ml/min). Specific areas of the
catalysts were calculated by applying the BET method to
the nitrogen adsorption isotherms obtained at liquid nitro-
gen temperature on outgassed samples at 413 K using an au-
tomatic ASAP 2000 adsorption instrument. Temperature-
programmed reduction (TPR) experiments were carried
out in a semiautomatic Micromeritics TPD/TPR 2900 ap-
paratus interfaced to a microcomputer. TPR profiles were
obtained by passing a 10% H2/Ar flow (50 ml/min) through
the sample (about 70 mg). The temperature was increased
from 305 to 1000 K at a rate of 10 K/min, and the amount of
H2 consumed was determined with a thermoconductivity
detector (TCD); the effluent gas was passed through a cold
trap placed before the TCD in order to remove water from
the exit stream.
Photoelectron spectra (XPS) were acquired with a VG
ESCALAB 200R spectrometer equipped with a hemispher-
ical electron analyzer and Mg Kꢀ (hv = 1253.6 eV, 1 eV =
1.6302 ꢅ 10ꢆ19 J) 120-W X-ray source. The powder samples
were evacuated in the pretreatment chamber of the instru-
ment at 295 and 573 K for 2 h and reduced in H2 at temper-
atures ranging from 305 to 773 K for 2 h. The residual pres-
sure in the ion-pumped analysis chamber was maintained
below 5 ꢅ 10ꢆ9 Torr during data acquisition. The intensities
of Zn 2p3/2 and Pd 3d5/2 peaks were estimated by calculating
the integral of each peak after smoothing and subtraction
of the “S-shaped” background and fitting the experimental
curve to a combination of Gaussian and Lorentzian lines of
variable proportion. The binding energies (BE) were ref-
erenced to the C 1s peak at 284.9 eV, this reference giving
BE values with an accuracy of ꢇ0.1 eV.
EXPERIMENTAL
Catalyst Preparation
The ZnO carrier was prepared by thermal decomposition
of Zn oxalate. Zinc oxalate was prepared by precipitation
of a Zn(NO3)2 ꢃ 6H2O (Aldrich, purity >98% ) solution by
dropwise adding at 313 K of another H2C2O4 ꢃ 2H2O (Pan-
reac, pa) solution (16.5% in excess). The initial concen-
trations of zinc nitrate and oxalic acid were 18.7 and 1.8 M,
respectively. After complete precipitation, the excess water
was removed under vacuum at 353 K. Thermal decompo-
sition of ZnC2O4 was performed in air in two steps: 418 K
for 0.5 h; then the temperature was increased at a rate of
5 K/min up to 653 K and maintaining this temperature for
0.5 h. Since the product exhibited a grey colour, probably
because of ZnO defect structure, the oxide was subsequ-
ently calcined in air at 673 K for 4 h while its colour turned
white.
Four Pd-ZnO catalysts were prepared by impregnation
of ZnO (particle size 70 nm) with aqueous solutions of
Pd(NO3)2 ꢃ 2H2O (Fluka, purum >98% ). Two catalysts,
nominal Pd contents of 1 and 2% , were prepared by the in-
cipient wetness impregnation method using a solution/solid
ratio = 1(ml/g), and two other catalysts, nominalPd content
Activity Measurements
Catalytic experiments for methanol oxidation were per-
1 and 5% , by the impregnation in excess of solution (10 and formed at atmospheric pressure in a continuous flow system
7 ml solution/g ZnO, respectively). After contacting ZnO with a stainless steel tubular reactor 6.3 mm ID, using 0.2
with the Pd solution at pH close to 3, a brown precipitate and 0.5 g of catalyst (particle size 0.42-0.59 mm) held be-
was formed. Although this precipitate was solubilized by tween quartz glass wool portions with a preheating section
dropwise adding dilute nitric acid, it persists at the end of packed with SiC at the top and an Inconel filter at the bot-
the impregnation. The impregnates were dried at 383 K for tom. The temperature of the reactor was maintained by
8 h and then calcined in air at 623 K for 3 h.
an electronic controller which powers an electrical furnace

152
CUBEIRO AND FIERRO
and uses a thermocouple as reference, placed 12 mm up the into ZnO. These TGA and DTGA profiles clearly indicate
catalytic bed. The temperature at the catalyst bed was mea- that a minimum decomposition temperature of 673 K is
sured by a second thermocouple placed inside the catalyst required to obtain ZnO. As the BET area is expected to
bed. Methanol (Scharlau, HPLC grade) was fed by means decrease with increasing decomposition temperature, the
of a liquid pump (Becton-Dickinson) at a rate of 2 ml/h ZnO support was prepared by thermal decomposition of
and then vaporized in a preheater. The oxygen and nitro- the Zn-oxalate hydrate precursor at 673 K.
gen (diluent) flows were adjusted by mass flow controllers.
The impregnation of the resulting ZnO with an aque-
The total flow was 100 ml(STP)/min, with a methanol mo- ous solution of Pd nitrate appear quite complex due to
lar concentration of 21.2% and O2/CH3OH ratios of 0.3 the tendency of Pd(NO3)2 to be hydrolyzed and also the
and 0.5. The samples were reduced in situ in a 10% H2/N2 ability of ZnO to be solubilized in acid solution (19). The
stream starting at room temperature for 0.5 h, then heating Pd(NO3)2 undergoes hydrolysis with the subsequent in-
→
at a rate of 5 K/min up to 573 K and keeping this tempera- crease of the acidity of the solution (Pd(NO3)2 + H2O
ꢁ
ture for 0.5 h. Reaction temperatures explored covered the Pd(OH)(NO3) + NOꢆ3 + H+). As the solubility of the Pd
range 503–543 K. The analysis of the effluents of the reactor hydroxinitrate is very low, some precipitation is expected
was performed with an on-line gas chromatograph (Varian to occur. In our case, the extent of this process may be even
3400 CX) provided with a TC detector and two Porapack larger because some solubilization of ZnO takes place upon
N and molecular 5A packed columns in series, using Ar contactingthe support and the impregnant solution with the
as the carrier gas. Data over periods of 8–12 h on-stream subsequent pH increase.
were obtained for each set of fixed reaction conditions. The
same sequence of variation on the O2/CH3OH ratio and the
temperature for each catalyst sample was followed.
Chemical Composition and Textural Properties
Chemical analyses by atomic absorption spectroscopy re-
vealed that Pd content is close to the nominal one (Table 1).
The shape of the N2 adsorption isotherms and BET areas
were conclusive that no microporosity developed. The BET
area of the catalysts prepared according to the wetness
impregnation method (1% Pd/ZnO and 2% Pd/ZnO) is
similar to that of the starting ZnO carrier (Table 1). How-
ever, the BET area decreased drastically in the catalysts
prepared by impregnation in excess solution (1% Pd/ZnO
(ES) and 5% Pd/ZnO). This result suggests, as already ad-
vanced above, that a small fraction of ZnO is solubilized in
acid medium during impregnation and again precipitated
over the ZnO substrate at the end of impregnation when
RESULTS AND DISCUSSION
Characterization of the Precursor
TGA and its derivative DTGA profiles of the Zn-oxalate
carrier precursor are shown in Fig. 1. Two major decompo-
sition peaks centred at 418 and 652 K can be observed.
The weight loss of the former peak, which amounts 18.7% ,
fits well with the 19.0% expected for water removal from
the Zn-oxalate hydrate [Zn(C2O4 ꢃ 2H2O], while the 37.9%
weight loss of the latter closely corresponds to the theo-
retical one expected for the decomposition of Zn(C2O4)
FIG. 1. Thermogravimetric (TG) and first derivative (DTG) curves of the zinc oxalate precursor under air flow (60 cm3/min) at a heating rate of
5 K/min.

PARTIAL OXIDATION OF METHANOL
153
reflection of Pd metal would be positioned at 40.1ꢄ and the
PdO(110) peak at 41.9ꢄ. In this latter case other more in-
tense reflections of PdO would be observed. In the light of
these observations and considering data from the literature
(16,20–22), the peak at a Bragg’s angle of 41.2ꢄ can be rea-
sonably assigned to a PdZn alloy. The presence of this alloy
in the catalyst during on-stream operation suggests that Pd
when alloyed is resistant toward oxidation.
TABLE 1
Chemical Composition and BET Area of Calcined Samples
Sample
Pd (wt% )
ZnO (wt% )
SBET (m2/g)
ZnO
—
—
10
11
9
4
4
1% Pd/ZnO (IW)
2% Pd/ZnO (IW)
1% Pd/ZnO (ES)
5% Pd/ZnO (ES)
1.13
2.43
1.24
5.20
83.8
89.9
83.8
84.8
Temperature-Programmed Reduction (TPR)
Note. IW = incipient wetness; ES = excess of solution.
Since PdO is an easily reducible oxide, even at room
temperature, the TPR experiments were carried out af-
ter sample conditioning and further exposure to the H2/Ar
mixture at room temperature. Hydrogen consumption was
observed at room temperature in several experiments al-
though its quantification became difficult, if not impossi-
ble, due to some fluctuations of the signal baseline just
at the beginning of the experiment. The TPR profiles of
ZnO-supported palladium catalysts are displayed in Fig. 4.
Three regions can be distinguished: (i), the first one ranging
from room temperature to 420 K; (ii), the second one in the
500–630 K range; and (iii) a steady increase above 670 K,
with a clear peak at about 900 K in catalyst 5% Pd/ZnO.
In general, it is apparent from Fig. 4 that the extent of
H2-consumption increased with increasing Pd loading. The
the water excess is completely removed. As the precipitated
phase, pressumably constituted by Zn(NO3)2, can be acco-
modated within the macropores of the original ZnO carrier,
a fraction of the pores of the original ZnO sample can be
blocked by this solubilized and precipitated ZnO phase.
X-Ray Diffraction (XRD)
XRD patterns of ZnO carrier and 1% Pd/ZnO catalysts
exhibited diffraction lines of ZnO phase but in no case that
of Pd metal or its compounds. Considering the low Pd load-
ing, the absence of diffraction lines of PdO may be due to
its high dispersion degree on the ZnO surface. It is also em-
phasized in this point that the most intense diffraction line
of PdO (PdO(101)) is positioned at a Bragg’s angle of 33.9ꢄ
which would be overshadowed by the strong ZnO(002) re-
flection at a Bragg’s angle of 34.4ꢄ. However, as the other
less intense Pd reflections which do not overlap with any
reflections from ZnO were absent, no bulk Pd is found by
XRD. The XRD diffraction patterns of calcined Pd/ZnO
catalysts with different Pd loadings are displayed in Fig. 2.
The diffraction lines of catalysts 1% Pd/ZnO (ES) and 5%
Pd/ZnO (ES) appeared substantially narrower than that
of the ZnO carrier and their 1% Pd/ZnO (IW) and 2%
Pd/ZnO counterparts, prepared by the incipient wetness
method. This finding indicates that the crystal size of ZnO
particles is larger in the samples prepared in excess of so-
lution, which agrees with their lower BET areas.
A complete study by XRD was carried out on the cata-
lyst 2% Pd/ZnO (IW). Figure 3 includes the XRD patterns
of this catalyst after hydrogen pretreatments and used in
the methanol partial oxidation. The catalyst 2% Pd/ZnO
(IW) calcined and reduced in hydrogen at 573 K (Figs. 3a
and b, respectively) basically exhibit the pattern of ZnO.
However, the pattern 3c of the catalyst reduced at 573 K
and used in the reaction for 160 h on-stream at a maximum
temperature of 563 K, and pattern 3d of the catalyst re-
duced at 623 K and then exposed to the reaction mixture
for 90 h at temperatures ranging from 503 to 543 K display
a broad and small peak at a Bragg’s angle of 41.2ꢄ. A similar
peak at the same position is clearly discerned in the catalyst
5% Pd/ZnO prereduced at 573 K (Fig. 3e). This peak can-
FIG. 2. XRD patterns of calcined Pd/ZnO catalysts with different Pd
loadings: (b), 1% Pd/ZnO (IW); (c) 2% Pd/ZnO (IW); (d), 5% Pd/ZnO
(ES). For comparative purpose the pattern of the ZnO carrier is also
not be indexed to either Pdꢄ or PdO because the Pd(111) included (a).

154
CUBEIRO AND FIERRO
As the Pd hydride is a bulk compound, the extent of hy-
dride formation can be strongly affected by Pd dispersion.
As confirmed by XPS measurements, the intensity of Pd
3d core level is rather high, even for the catalysts with 1%
Pd loading, suggesting that metal is highly dispersed at the
ZnO surface. Another alternate explanation is that a hy-
dride can be decomposed once it is formed; therefore the
negative peak of the hydride would be masked by the positi-
ve one due to the reduction of PdO. It is emphasized in
this point that H2-consumption at room temperature for
the catalyst 1% Pd/ZnO (IW) was somewhat higher than
the theoretical one expected for the quantitative reduction
to Pdꢄ, suggesting the formation of a certain proportion of
Pd-like hydride species or even ZnO reduction. For catalyst
1% Pd/ZnO (ES) the first H2 consumption peak at 343 K
was 1.7 times greater than the theoretical one for the PdO
reduction. This result suggests that some reduction of the
ZnO support can occur at low temperatures.
The progressive increase of TPR baseline for catalyst 2%
Pd/ZnO at temperatures above 700 K and the observation
of a broad and strong peak at ca 900 K for catalyst 5%
Pd/ZnO indicate that another reduction process is involved
in that temperature region. A blank TPR experiment using
pure ZnO (not shown here) revealed the appearance of a
broad feature somewhat above 800 K, as well as a sharp
increase on the H2 consumption at ca 1000 K, indicative
of the partial reduction of the ZnO substrate. In the light
FIG. 3. XRD patterns of catalyst 2% Pd/ZnO subjected to various
pretreatments: (a) calcined sample; (b) reduced in hydrogen at 573 K;
(c) used catalyst reduced at 573 K; (d) used catalyst reduced at 623 K. For
comparison, catalysts 5% Pd/ZnO reduced in hydrogen at 573 K is also
included (e).
impregnation procedure does appear to have some influ-
ence on the redox properties of the catalysts. Thus the cata-
lyst 1% Pd/ZnO (ES) was reduced at about 350 K, whereas
its counterpart 1% Pd/ZnO (IW) was essentially reduced
at ambient temperature (Figs. 4a and 4b). The interpreta-
tion of these profiles may be complicated not only by the
observation of H2-consumption peaks from both PdO and
ZnO oxides but also by the formation of bulk palladium
hydride. The first H2-consumption peak at 300–420 K can
be assigned to the reduction of PdO to Pd metal, whereas
the broad peaks at 600 and 563 K observed for catalysts
1% Pd/ZnO (ES) and (IW) and at 617 and 553 K for 2%
Pd/ZnO and 5% Pd/ZnO, respectively, may be associated
to some reduction of ZnO substrate. As can be seen below
in the XPS section, palladium became completely reduced
at somewhat lower temperatures. Using Pd/ZnO catalysts,
Hong et al. (20) found reduction of PdO, and also mildly of
ZnO, to occur at room temperature, then Pd metal forms
a hydride (HxPd) which decomposes at higher tempera-
tures. Iwasa et al. (16) also observed that this hydride de-
composes (negative peak in TPR profiles) at lower temper-
atures in ZnO-supported catalysts than in other carriers.
Surprisingly, no hydride decomposition peak was observed
in the Pd/ZnO catalysts reported in this study, which con-
FIG. 4. TPR profiles of ZnO-supported palladium catalysts: (a) 1%
Pd/ZnO (ES); (b) 1% Pd/ZnO (IW); (c) 2% Pd/ZnO (IW); and (d) 5%
trasts with Pd/AlO3, Pd/ZrO2, and Pd/SiO2 systems (16,20). Pd/ZnO (ES).

PARTIAL OXIDATION OF METHANOL
155
TABLE 2
Binding Energies (eV) of Core Electrons and Surface (XPS)
Atomic Ratios of Catalyst 1% Pd/ZnO Subjected to Different
Pretreatments
Pretreatment
O 1s
Zn 2p3/2
Pd 3d3/2
Pd/Zn atom
530.9(77)
532.4(23)
Vacuum, RT
Vacuum, 573 K
H2, RT
1022.2
1022.2
1022.2
1022.2
1022.2
1022.2
336.9
0.047
0.045
0.036
0.034
0.033
0.031
530.9(76)
532.3(24)
336.7(100)
530.9(69)
532.6(31)
337.0(41)
335.5(59)
530.9(66)
532.2(34)
336.9(18)
335.7(82)
H2, 373 K
H2, 573 K
H2, 773 K
531.0(73)
532.5(27)
335.7(100)
335.7(100)
531.0(74)
532.5(26)
of this result, it is inferred that H2-consumption at high
temperatures is associated to the reduction of bulk ZnO.
Photoelectron Spectroscopy (XPS)
Photoelectron spectroscopy was used to study the chem-
ical state of the elements and their relative abundance at
catalyst surfaces. Two representative catalysts 1% Pd/ZnO
(IW) and 5% Pd/ZnO (ES) were selected for this purpose,
and the core level spectra recorded included O 1s, Zn 2p3/2
,
Pd 3d, and C 1s (as internal reference). The binding ener-
gies of core electrons and surface atomic ratios for catalysts
1% Pd/ZnO (IW) and 5% Pd/ZnO (ES) are summarized
in Tables 2 and 3, respectively.
FIG. 5. Pd 3d core levelspectra ofcatalyst 1% Pd/ZnO (IW) subjected
to treatments of outgassing and H2-reduction at different temperatures.
For the sake of clarity and in order to have an idea on
peak profile, the Pd 3d core level of these two catalysts are
displayed in Figs. 5 and 6, respectively. The profile of Pd
3d doublet was complex because of the overlaping ZnLMM
Auger peak and the less intense Pd 3d3/2 component. To
avoid this difficulty, attention will be thereupon exclusively
focused on the most intense Pd 3d5/2 peak. Calcined samples
exhibit a single Pd 3d5/2 component with a BE of 336.8–
336.9 eV (Figs. 5 and 6) which is typical of PdO species
(16,23,24). Upon exposure to H2 at room temperature, this
oxide becomes reduced to a significant extent. By applying
curve fitting procedures to the Pd 3d5/2 envelope, a compo-
nent at a BE of 335.5 eV, with a 59% of the whole area,
for the sample 1% Pd/ZnO (IW) (Fig. 5), and at BE of
335.0 eV, with 65% of the area, for the sample 5% Pd/ZnO
(ES) (Fig. 6) were observed. The extent of Pd reduction
increased upon increasing the reduction temperature up to
373 K. Under these conditions, the area of the reduced Pd
increased up to 82% for catalyst 1% Pd/ZnO (IW) (BE of
Pd 3d5/2 = 335.7 eV) (Fig. 5) and to 100% for catalyst 5%
Pd/ZnO (ES) (Fig. 6). Reduction of catalysts at 573 and
773 K led to a single Pd 3d5/2 component (Figs. 5 and 6).
However, comparison of BEs of Pd 3d5/2 peaks for sam-
ples reduced at higher temperatures reveals important dif-
ferences. Thus, while catalyst 1 Pd/ZnO (IW) shows BEs
at 335.7 eV when reduced at 373, 573, and 773 K, its 5%
TABLE 3
Binding Energies (eV) of Core Electrons and Surface (XPS)
Atomic Ratios of Catalyst 5% Pd/ZnO Subjected to Different
Pretreatments
Pretreatment
O 1s
Zn 2p3/2
Pd 3d3/2
Pd/Zn atom
531.0(67)
533.0(33)
vacuum, RT
H2, RT
1022.2
1022.2
1022.2
1022.2
1022.2
336.8(100)
0.182
0.494
0.333
0.242
0.207
530.8(63)
533.0(37)
336.8(36)
335.0(64)
530.8(65)
532.7(35)
H2, 373 K
H2, 573 K
H2, 773 K
335.0(100)
335.2(100)
335.7(100)
530.9(72)
532.7(28)
530.9(74)
532.9(26)

156
CUBEIRO AND FIERRO
of 373, 573, and 773 K for catalyst 1% Pd/ZnO (IW) and
the shift from 335.2 eV at 573 K to 335.7 eV at 773 K for
the parent catalyst 5% Pd/ZnO indicate that not only Pd
loading but pressumably Pd crystal size and metal–support
interaction, among other factors, would be responsible for
alloy formation.
The O 1s core level exhibited two components irrespec-
tive of Pd content and catalyst pretreatments. Catalyst 1%
Pd/ZnO (IW) showed a major peak at a BE of 530.9–
531.0 eV corresponding to lattice oxygen and a minor one
at 532.7–533.0 eV assigned to surface hydroxyl groups. The
relative proportion of these components changed with cata-
lyst pretreatments (Tables 2 and 3). Thus, the intensity
of the peak at ca 532.7 eV increased upon exposure to
H2 at room temperature and then decreased at higher re-
duction temperatures. This phenomenon can be explained
on the basis of water adsorption/reaction at the ZnO in-
terface and the further decomposition of the product at
higher temperatures. The water generated in the course of
→
PdO reduction (PdO + H2 Pd + H O) is adsorbed on the
ꢁ
2
ZnO surface where it is dissociated, leading to OH groups
→
(ZnOZn + H O
2Zn-OH). While this last equilibrium
ꢁ
2
may be shifted to the right side at lower temperatures, the
decomposition of hydroxyl groups becomes a favored pro-
cess at temperatures above 373 K.
Atomic Pd/Zn ratios have been also included in Tables 2
and 3. These ratios were calculated from peak area mea-
surements and atomic sensitivity factors (24). For catalyst
1% Pd/ZnO the Pd/Zn XPS ratio is a maximum for the cal-
cined and outgassed sample (0.047 and 0.045, respectively),
but it decreases by about a 25% upon exposure to H2 at-
mosphere at room temperature and even more (ca 34% )
at 773 K. The dependence of the Pd/Zn ratio on pretreat-
ments is somewhat different for the parent 5% Pd/ZnO
catalyst. Upon H2-reduction at room temperature this ra-
tio increases by a factor of about 3 with respect that of the
calcined sample, and then it decreases drastically with in-
creasing reduction temperature (58% decrease at 773 K).
The decay of Pd/Zn XPS ratios upon reduction can be pri-
marily associated to the well-known tendency of Pd to sin-
ter. Moreover, if PdZn alloys are formed, the dilution effect
of Zn in the alloy could explain the decrease of the Pd sig-
nal. As pointed out by Rodriguez (30), zinc tends to be
segregated towards the surface of PdZn alloys. Therefore,
one would expect that Zn-rich alloys or Zn segregation to
the surface may occur upon increasing the reduction tem-
perature. On the other hand, if alloy formation is favored at
low temperatures, as is the case of 1% Pd/ZnO catalyst, no
so-marked changes in the Pd/Zn ratio would be observed
upon increasing the temperature of reduction.
FIG. 6. Pd 3d core level spectra of catalyst 5% Pd/ZnO (ES) subjected
to treatments of outgassing and H2-reduction at different temperatures.
Pd/ZnO (ES) counterpart displays BEs at 335.2 eV for the
catalyst reduced at 573 K and increased up to 335.7 eV for
the catalyst reduced at 773 K.
From an inspection of the BE values of Pd 3d5/2 core level
of several supported-palladium catalysts (16,23,25–27), it
becomes apparent that BE close to 335 eV are responsible
of Pdꢄ species (16,23–29). Slightly higher BE values than
this for the Pd 3d5/2 core level have been reported for Pd
supported in redox oxides (16) or for Pd-deposited on single
crystals (26) and asigned to alloy formation (PdMx, where
M isthe redoxmetalofthe carrier or ofthe promoter). Thus,
Rodriguez (30) reported that a chemical shift of +0.7 eV
with respect to the metal would be expected if it is alloyed.
In line with this and considering the XRD diffraction pat-
terns of the used and reduced catalysts (Figs. 3c–e), it can be
argued that a PdZn alloy is pressumably formed upon H2-
reduction at temperatures above 573 K. The observation
Catalytic Activity
In order to explore the activity and selectivity behavior
of the Pd 3d5/2 peak at 335.7 eV at reduction temperatures of Pd/ZnO catalysts, operation conditions under integral

PARTIAL OXIDATION OF METHANOL
157
regime and O2/CH3OH ratios in the feed close to the stoi-
chiometry of Eq. [1] were selected. Preliminary runs were
performed with catalyst 1% Pd/ZnO, using0.5-gsample and
flow rates ranging from 50 to 100 ml/min. As mass transport
limitations were suggested to be involved in the oxidation
reaction when working with the lowest flow rate, a feed rate
of 100 ml/min was selected for activity tests. Two methanol
compositions of 21.2 and 42.4% (molar) were explored as
well. Due to the high exothermicity of the reaction, the oc-
currence of heat transport limitation may have increased
the temperature of the catalyst particles significantly; thus,
the real temperature is higher than that measured by the
thermocouple inside the catalyst bed. In order to allevi-
ate this uncertainty, the former composition (21.2% molar)
was used in most experiments. It is emphasized at this point
that a CH3OH concentration in the feed stream of 21.2%
is quite similar to that of a O2/CH3OH = 0.5 which can be
reached using air instead oxygen as an oxidant.
Catalyst 1% Pd/ZnO (IW)
Activity and product distributions in the methanol partial
oxidation over catalyst 1% Pd/ZnO (0.5 g) at O2/CH3OH
feed ratios of 0.3 and 0.5 are displayed in Figs. 7a and 7b,
respectively. Within the temperature range 503–543 K, oxy-
gen consumption was complete, while methanol conver-
sions reached 40–80% . Hydrogen, water, carbon oxides,
and unconverted methanol were the only products detected
at the outlet of the reactor. Upon increasing the reaction
temperature, methanol conversion increased with a simul-
taneous increase in H2 selectivity at the expense of water.
Selectivity of CO2 also increased, although its increment
was less marked (Fig. 7). Since O2 was completely con-
sumed, this selectivity trend suggests some contribution
of the methanol steam reforming produced by the water
by-product. Furthermore, as illustrated in Table 4, the se-
lectivities to the highest methanol conversions and fixed
O2/CH3OH ratio can be predicted, considering that the ad-
ditional methanol converted disappeared according to the
steam reaction (Eq. [2]). Calculated data in Table 4 were
obtained from experimental compositions at 503 K, using
the stoichiometry of reaction [2] to predict the compositions
to be attained if conversion increases up to the experimen-
tal value at 543 K. Table 4 also includes calculated values
assuming CH3OH decomposition (Eq. 3):
FIG. 7. Dependence of methanol conversion and selectivities on
the temperature for the partial oxidation of methanol over catalyst 1%
Pd/ZnO using feed ratios O2/CH3OH of 0.3 (a) and 0.5 (b).
temperature, revealed that very high H2 selectivities can be
reached at high methanol conversions. This is clear for a ra-
tio O2/CH3OH = 0.3 in the feed, with which H2 selectivity
of 96% can be attained at a CH3OH conversion of 70% .
Looking at the carbon oxide distribution, CO2 is dominant
TABLE 4
Product Distributions in the Partial Oxidation of Methanol
(O2/CH3OH = 0.3 M) on Catalyst 1% Pd/ZnO Assuming the Con-
tribution of Reactions [2] and [3]
Experimental dataa
Calculated datab
503 K
543 K
Eq. [2]
Eq. [3]
70(82)
→
ꢁ
methanol conv (% )
carbon selectivity (% )
38(66)
70(82)
70(82)
CH OH
2 H + CO.
[3]
3
2
CO
CO2
34(25)
66(75)
19(20)
81(80)
18(20)
82(80)
64(39)
36(61)
If the decomposition reaction (Eq. [3]) is involved, the cal-
culated hydrogen proportion becomes much lower with,
simultaneously, a much higher CO percentage than the ex-
perimental ones.
hydrogen selectivity (% )
H2
H2O
50(58)
50(42)
96(78)
4(22)
96(76)
4(24)
73(66)
27(34)
As illustrated in Fig. 8, the product distributions exhib-
ited a marked dependence on the CH3OH conversion. The
a Oxygen conversion = 100% .
b Starting from experimental data at 503 K. Data in parentheses are for
data reported in Fig. 8, obtained by changing the reaction O2/CH3OH = 0.5 M.

158
CUBEIRO AND FIERRO
over copper catalysts. On the other hand, Jiang et al. (32)
and Takezawa et al. (15) excluded that reaction scheme and
proposed a newone for the CH3OH reforming, consistingin
the reaction of oxygenate intermediates, such as formalde-
hyde or methyl formate with water vapor. These authors
also argued that the WGS does not occur in the presence
of methanol over copper catalysts. Furthermore, Takezawa
et al. (15,16) proposed a reactional scheme based on the re-
forming of methyl formate or formaldehyde over Pd/ZnO
catalysts. In our case, the observation of a certain propor-
tion of CO and water among the reaction products indicate
that the rate of WGS reaction is not fast. This finding also
supports the hypothesis that the oxidation reaction might
well occur through reforming of methanol or certain oxy-
genate intermediates, at least under the conditions used in
the present study.
Figure 9 displays the changes in CH3OH conversion pro-
duced upon catalyst exposure to the reaction mixture for
long on-stream operation. At the two O2/CH3OH ratios
(0.3 and 0.5), methanol conversion decreased around 8%
for reaction times longer than 48 h. Moreover, selectivity
does not change at a significant extent, as these data fit well
with the whole set of data. It is worth noting that selectivity
results from these life tests have been included in Fig. 8. The
exothermicity of the reaction could be perceived by mea-
suring the temperature at the catalyst bed and at the oven
wall (Tbed ꢆ Twall). It can be seen in Fig. 9 that the slow de-
activation observed in the catalyst subjected to on-stream
operation for periods longer than 48 h is accompanied by
a parallel increase of the Tbed ꢆ Twall. While this difference
amounts 9 K for a feed ratio O2/CH3OH = 0.3 and reaction
temperature of 503 K, it increases up to 28 K for a feed ra-
tio O2/CH3OH = 0.5 and reaction temperature of 543 K. If
the reaction involves oxidation (exothermic) and reform-
ing (endothermic) steps, higher heat evolution would be
expected at lower rather than at the higher CH3OH con-
versions.
FIG. 8. Selectivities as a function of methanol conversion for the par-
tial oxidation of methanol over catalyst 1% Pd/ZnO using feed ratios
O2/CH3OH of 0.3 (a) and 0.5 (b).
although a certain proportion of CO is observed. From the
data reported in Fig. 10 it becomes apparent that, upon in-
creasing O2 in the feed stream, larger CH3OH conversions
are required to obtain the same H2 selectivity. The less-
marked effect on carbon oxides points to a faster oxidation
of H2 than CO.
The methanol steam reforming reaction (Eq. [2]) can be
considered as a combination of the methanol decomposi-
tion (Eq. [3]) and water gas shift (WGS) reaction (Eq. [4]):
Catalyst Weight and Dilution
→
With the aim to analyse the selectivity–conversion re-
lationship an additional experiment was carried out using
0.20 g of catalyst 1% Pd/ZnO. However, conversion was
similar to that using 0.50 g of catalyst (Fig. 10). There are no
thermodynamic restrictions, although, as can be seen below,
the strong exothermicity of the reaction could be respon-
CH OH
2 H + CO
[3]
[4]
ꢁ
3
2
→
CO + H O
CO + H
ꢁ
2
2
2
→
ꢁ
CH OH + H O
3 H + CO
[2]
3
2
2
2
The H2O (or the CO2) required for the WGS reaction sible for this behavior. Figure 10 also shows the increase
comes from the deep oxidation of a fraction of methanol fed of CH3OH conversion when increasing the O2/CH3OH ra-
→
(CH3OH + O2
H O + CO ). According to this reaction tio from 0.3 to 0.5. As this increase represents an increase
ꢁ
2
2
sequence, if the whole CO formed in the decomposition in the oxygen partial pressure (PO2) and a positive depen-
reaction (Eq. [3]) is consumed in the WGS reaction, coinci- dence order between PO2 and reaction rate is expected, the
dence in the proportion of products by steam reforming and observed trend is reasonable. As is well known, methanol
by decomposition-WGS reactions could be expected. The conversion over Cu/ZnO catalysts is accelerated in the
scheme decomposition-WGS has been proposed by sev- presence of oxygen (33–35). Although a similar mecha-
eral authors (5,12,31) for the methanol steam reforming nism could be involved in the oxidation of methanol over

PARTIAL OXIDATION OF METHANOL
159
FIG. 9. Changes of the methanol conversion and Tbed ꢆ Twall difference on the time on-stream for the partial oxidation of methanol over catalyst
1% Pd/ZnO: (a) 0–19 h; (b) 68–76 h; (c) 44–53 h; (d) 93–99 h.
Pd/ZnO catalysts; heat transport would influence the reac- methanol and oxygen conversions were lower in the di-
tion when using higher PO2 in the feed stream.
luted catalyst, the average partial pressure of the reactants
Upon catalyst dilution in SiC (SiC/Cat = 2/1), methanol over the catalyst surface was higher. The hypothesis of a
and oxygen conversions decreased with respect to undi- faster deactivation of the catalyst in the presence of oxy-
luted catalyst under the same experimental conditions gen is not discarded. With respect to the convenience of
(Fig. 10). This finding suggests that in the undiluted catalyst operating under reaction conditions far from the isother-
heat transport effects, pressumably with hot spots within micity Jenkins and Shutt (36) pointed out that the use of a
the catalyst bed, are involved in the oxidation reaction. Be- reactor, properly designed to operate with hot spots within
sides that, the heat transfer limitation can also cause sig- the catalyst bed, could be appropriate to generate H2 from
nificant radial temperature gradients which also influence methanol according to Eq. [1]. Furthermore, in the compar-
the results. Catalyst dilution led to a less marked change in ison of isothermal and hot spot reactors these authors ob-
the methanol conversion upon increasing the temperature served poorer performances in the conventional isothermal
of reaction or O2/CH3OH ratio in the feed. Deactivation reactor.
was also higher for the diluted bed (Fig. 12). One differ-
The H2 selectivity during the partial oxidation of metha-
ence between the runs, with and without dilution, lies in nol over 1% Pd/ZnO catalyst for the three catalytic runs
the environment to which the catalyst was subjected. Since (0.5, 0.2, and 0.2 g with dilution in SiC) are displayed in
Fig. 11. H2 selectivity was quite low at low CH3OH con-
versions when O2 was not completely consumed, and it in-
creased with increasing CH3OH conversion. The converted
O2/CH3OH ratio was higher than 1 at low conversions and
approached the stoichiometric value of 0.5 (see Eq. [1])
with increasing conversion (Fig. 12).
In addition to the major products of the reaction (H2,
H2O, CO2, and CO), a small proportion of HCHO was de-
tected mainly at the lowest CH3OH conversions. Therefore,
the participation of the following reactions is envisaged to
a certain extent when oxygen is present in the gas phase:
→
ꢁ
CH3OH + 1/2 O2
HCHO
HCHO + 1/2 H + 1/2 H O [5]
2
2
→
ꢁ
CO + H
[6]
[7]
[8]
2
→
ꢁ
H2 + 1/2 O2
CO + 1/2 O2
H O
2
FIG. 10. Methanol conversion on catalyst 1% Pd/ZnO using different
weights and SiC dilution: (a) 0.5 g; (b) 0.2 g; (c) 0.2 g + 0.4 g SiC.
→
ꢁ
CO .
2

160
CUBEIRO AND FIERRO
FIG. 13. Influence of the Pd content on the methanol conversion at
503 K and feed ratio O2/CH3OH of 0.3 and reduction temperatures of
473 K (ᮀ), 573 K (ꢅ), and 623 K (4).
the Pd/ZnO catalysts prereduced at 573 K are displayed
in Fig. 13. The CH3OH conversion is essentially constant
between 1 and 2% Pd and then drops slightly for catalyst
5% Pd/ZnO. With respect to product distribution some dif-
ferences were observed among the catalysts. Catalysts 1%
and 2% Pd/ZnO, which exhibited comparable conversion
levels, displayed similar H2 selectivity (Fig. 11). The excep-
tion of this tendency was catalyst 5% Pd/ZnO in which CO
selectivity was enhanced with respect to the 1% and 2% Pd
counterpart. In addition, HCHO selectivity over catalyst
5% Pd/ZnO was much higher. As a general tendency, H2
(and CO) selectivity is directly related to CH3OH conver-
sion; however, the behavior of catalyst 5% Pd/ZnO cannot
be explained in terms of this trend. It is likely that processes
such as methanol decomposition, unability to oxidize the
intermediate HCHO, and low oxidation rate of CO might
be involved in large PdZn alloy (and Pd) particles as that
found on prereduced catalyst 5% Pd/ZnO.
FIG. 11. Hydrogen selectivity over Pd/ZnO catalysts of different Pd
loading for O2/CH3OH ratios of 0.3 (a) and 0.5 (b). (♦) 1% Pd; (4) 2%
Pd; (᭹) 5% Pd.
The low H2 selectivity observed at low methanol conversion
may be due to the fast oxidation of hydrogen (Eq. [7]),
produced by the other reactions (Eqs. [1], [2], [5], and [6]).
On the other hand, as illustrated in Table 4, the contribution
ofthe steam reformingreaction seemsto be important when
oxygen becomes completely consumed.
Palladium Loading
Catalyst Pretreatments
The dependence of methanol conversion at a reaction
temperature of 503 K and feed ratio O2/CH3OH of 0.3 on
There are literature reports which document the im-
portance of the activation step on the performance of
Pd/ZnO catalysts when used in the methanol steam re-
forming (15,16). It has been emphasized that catalyst ac-
tivity is enhanced if PdZn alloy is allowed to be formed
upon hydrogen pretreatment. In order to study the influ-
ence of H2-pretreatment on the performance for partial
oxidation of methanol, catalyst 2% Pd/ZnO prereduced at
three different temperatures was selected for this purpose.
As already shown in Fig. 4, the selected temperatures of
reduction of 473, 573, and 623 K represent different extents
or levels of PdZn alloy formation. Therefore, this explored
temperature range is expected to cover surface composi-
tions from Pd-rich particles (473 K) to well-defined PdZn
structures (623 K). Catalyst prereduced at 473 and 573 K
yielded the same activity and decreased upon reduction
at 623 K (Fig. 13). The selectivity of hydrogen followed a
FIG. 12. Dependence of the consumed O2/CH3OH ratio on the
methanol conversion over catalyst 1% Pd/ZnO for feed ratios O2/CH3OH
of 0.3 (᭜) and 0.5 (ᮀ).

PARTIAL OXIDATION OF METHANOL
161
metal–metal oxide interactions of the smaller Pd particles.
The reactivity of the alloy seems to be somewhat different
from that of Pd particles as for catalyst 5% Pd/ZnO the
HCHO selectivity was much higher. The presence of PdZn
alloys was also detected by X-ray diffraction on the 2%
Pd/ZnO catalyst used on-stream. It is likely that processes
such as methanol decomposition, inability to oxidize the in-
termediate HCHO and low oxidation rate of CO might be
involved in large PdZn alloy particles since CO and HCHO
selectivities were much higher for catalyst 5% Pd/ZnO. The
reduction temperature appears as a critical parameter for
the catalytic behavior. A compromise between enhance-
ment in the H2 selectivity and activity drop at increasing
reduction temperatures was apparent.
FIG. 14. H2 selectivity (O2/CH3OH = 0.5) for catalyst 2% Pd/ZnO
subjected to different reduction temperatures: (♦) 473 K; (ᮀ) 573 K;
(4) 623 K.
ACKNOWLEDGMENT
similar trend (Fig. 14), although the comparison of H2 se-
lectivity of this catalyst prereduced at 623 K with that of the
parent 1% Pd/ZnO prereduced at 573 K (Fig. 11b) reveals
a substancial improvement in the former. It appears that
H2 selectivity became enhanced in the catalysts in which
PdZn alloy formation is favored. Even considering this
beneficial effect, care must be taken to avoid high temper-
ature pretreatments to minimize the sintering of metallic
particles.
M.L.C. thanks the Central University of Venezuela and its CDCH
Council for a sabbatical leave. Finantial support from CICYT, Spain, un-
der Projects MAT95-0894, JOE3-CT97-0049, and MAT96-2056CE is also
acknowledged.
REFERENCES
1. Pen˜a, M. A., Gomez, J. P., and Fierro, J. L. G., Appl. Catal. A: General
144, 7 (1996).
2. Jamal, Y., and Wyszynski, Int. J. Hydrogen Energy 19, 557 (1994).
3. Veziroglu, T. N., and Barbir, F., Int. J. Hydrogen Energy 17, 391
(1992).
CONCLUSION
4. Petterson, L., and Sjo¨stro¨m, K., Int. J. Hydrogen Energy 16, 671
(1991).
5. Amphlett, J. C., Creber, K. A., Davis, J. M., Mann, R. F., Peppley,
B. A., and Stokes, D. M., Int. J. Hydrogen Energy 19, 131 (1994).
6. Kumar, R., Ahmed, S., and Yu, M., Preprints, Am. Chem. Soc., Div.
Fuel Chem. 38, 1471 (1993).
7. Lindner, B., and Sjo¨mstro¨m, K., Fuel 63, 1485 (1984).
8. Srinivasan, S., J. Electrochem. Soc. 136, 41C (1989).
9. Schmitz, A. D., Eyman, D. P., and Gloer, K., Energy & Fuels 8, 729
(1994).
10. Jiang, C. J., Trimm, D. L., and Wainwright, M. S., Chem. Eng. Technol.
18, 1 (1995).
11. Jiang, C. J., Trimm, D. L., Wainwright, M. S., and Cant, N. W., Appl.
Catal. A: General 93, 24 (1993).
12. Huang, T. J., and Wang, S. W., Appl. Catal. 24, 287 (1986).
13. Huang, T. J., and Chren, S. L., Appl. Catal. 40, 43 (1988).
14. Alejo, L., Lago, R., Pen˜a, M. A., and Fierro, J. L. G., Appl. Catal. A:
General 162, 281 (1997).
ZnO-supported Pd catalysts are very active and selec-
tive toward hydrogen production in the partial oxidation
of methanol. Catalyst performance has been investigated
as a function of the Pd loading and pretreatment in hydro-
gen environment under feed ratios O2/CH3OH (molar) of
0.3 and 0.5 at 503–543 K. The dependences of the selectivi-
ties, the converted ratio O2/CH3OH, and, also, the evolved
heat on methanol conversion were consistent with the re-
action scheme: oxidation → reforming of methanol. As the
oxidation reaction is exothermic and the operation was con-
ducted, in general, using quite concentrated feed, important
thermal effects were observed, with the undiluted catalyst
bed showing a better performance than the diluted one.
Catalyst characterization by TPR, X-ray diffraction, and
XPS revealed that PdZn alloys can be formed upon reduc-
15. Takezawa, N., and Iwasa, N., Catal. Today 36, 45 (1997).
tion at moderate temperatures. A shift of +0.7 eV has been 16. Iwasa, N., Masuda, S., Ogawa, N., and Takezawa, N., Appl. Catal. A:
General 125, 145 (1995).
17. Alejo, L., Pen˜a, M. A., and Fierro, J. L. G., in “Proceedings of the 15th
observed in the binding energy of Pd 3d5/2 core level spec-
trum of the catalyst 5% Pd/ZnO prereduced at tempera-
tures above 373 K, which agrees with that reported in the lit-
Iberoamerican Symposium on Catalysis, Cordoba, Argentina, 1996,”
Vol. 3, p. 1661.
erature for alloyed palladium (16). Moreover, PdZn alloys
appear to be formed even easily on catalyst 1% Pd/ZnO.
This is confirmed not only by the observation of a compo-
nent at 335.7 eV in the Pd 3d5/2 core level at lower tempera-
tures of reduction but also by the rather low surface-to-bulk
Pd/Zn ratios in catalyst 1% Pd/ZnO compared to that of 5%
18. Lago, R., Alejo, L., Pen˜a, M. A., and Fierro, J. L. G., Stud. Surf. Sci.
Catal. 110, 623 (1997).
19. “The Merck Index. An Encyclopedia of Chemicals, Drugs and Bio-
logicals,” 12th ed. Merck, Rahway, NJ, 1996.
20. Hong, Ch. T., Yeh, Ch. T., and Yu, F. H., Appl. Catal. 48, 385 (1989).
21. Sarkany, A., Zsoldos, Z., Furlong, B., Hightower, J. W., and Guczi, L.,
J. Catal. 141, 566 (1993).
Pd/ZnO. This is pressumably a consequence of the stronger 22. Zsoldos, Z., Sarkany, A., and Guczi, L., J. Catal. 145, 235 (1994).

162
CUBEIRO AND FIERRO
23. Kim, K., Gossmann, A., and Winograd, N., Anal. Chem. 46, 197 (1976).
24. Briggs, D., and Seah, M. P. (Eds.), “PracticalSurface AnalysisbyAuger
and X-Ray Photoelectron Spectroscopy. Wiley, Chichester, 1985.
29. Pitchon, V., Guenin, M., and Praliaud, H., Appl. Catal. 63, 333 (1990).
30. Rodriguez, J. R., J. Phys. Chem. 98, 5758 (1994).
31. Santacesaria, E., and Carra, S., Appl. Catal. 5, 345 (1983).
25. Goetz, J., Volpe, M. A., Sica, A. M., Gigola, C. E., and Touroude, R., 32. Jiang, C. J., Trimm, D. A., Wainwright, M. S., and Cant, N. W., Appl.
J. Catal. 153, 86 (195).
Catal. A: General 97, 81 (1994).
26. Hermann, P., Guigner, J. M., Tardy, B., Jugnet, Y., Simon, D., and
Bertolini, J. C., J. Catal. 163, 169 (1996).
27. Venezia, A. M., Rossi, A., Duca, D., Martorana, A., and Deganello,
G., Appl. Catal. A: General 125, 113 (1995).
33. Wachs, I. E., and Madix, R. J., J. Catal. 53, 208 (1978).
34. Clarke, D. B., Lee, D. K., Sandoval, M. J., and Bell, A. T., J. Catal. 150,
81 (1994).
35. Fu, S. S., and Somorjai, G. A., J. Phys. Chem. 96, 4542 (1992).
36. Jenkins, J. W., and Shutt, E., Plat. Met. Rev. 33, 118 (1989).
28. Jones, M. G., and Nevell, T. G., Appl. Catal. 70, 277 (1991).