Production of hydrogen from oxidative steam reforming of methanol II. Catalytic activity and reaction mechanism on Cu/ZnO/Al2O3 hydrotalcite-derived catalysts

Article abstract of DOI:10.1016/j.jcat.2004.08.029

Cu/ZnO/Al₂O₃ catalysts derived from a hydrotalcite precursor were studied for oxidative steam reforming of methanol (OSRM), with O₂/H₂O/CH₃OH molar ratios = 0.12:1.1:1. The results were compared with those obtained in nonoxidative steam reforming and in partial oxidation of methanol. Good results, in terms of CH₃OH conversion and H₂ yield, together with low CO production, were obtained, showing that the method of homogeneous precipitation of hydrotalcite-like precursors is suitable for preparing high activity catalysts. Methanol decomposition appeared a reaction slower than steam reforming and partial oxidation and was greatly depressed by water vapor. A noticeable catalytic activity for methanation was evidenced in high temperature tests. The reaction network involved also dehydrogenation to formaldehyde and dehydration to dimethylether, the water gas shift reaction, besides oxidation reactions. A study on the reaction mechanism showed different species formed by adsorption of methanol on the catalyst. The hypothesized mechanism pointed to an important role of Cu(I) and Cu(II) species: Cu(I) strongly adsorbs CO, while Cu(II) catalyzes CO oxidation to CO₂. The oxidation state of Cu species is influenced by the presence of O₂ and H₂O.

Full text of DOI:10.1016/j.jcat.2004.08.029

Journal of Catalysis 228 (2004) 56–65
www.elsevier.com/locate/jcat
Production of hydrogen from oxidative steam reforming of methanol
II. Catalytic activity and reaction mechanism on Cu/ZnO/Al2O3
hydrotalcite-derived catalysts
M. Turco a,∗, G. Bagnasco a, U. Costantino b, F. Marmottini b, T. Montanari c, G. Ramis c,
G. Busca c
a
Dipartimento di Ingegneria Chimica, Università di Napoli Federico II, P.le V. Tecchio 80, 80125 Napoli, Italy
b
CEMIN – Centro Eccellenza Materiali Innovativi Nanostrutturati, Dipartimento di Chimica, Università di Perugia, Via Elce di Sotto 8, 06123 Perugia, Italy
c
Dipartimento di Ingegneria Chimica e di Processo “G.B. Bonino”, Università di Genova, P.le J.F. Kennedy 1, 16129 Genova, Italy
Received 10 May 2004; revised 30 July 2004; accepted 17 August 2004
Available online 25 September 2004
Abstract
Cu/ZnO/Al O catalysts derived from a hydrotalcite precursor were studied for oxidative steam reforming of methanol (OSRM), with
2
3
O :H O:CH OH molar ratios = 0.12:1.1:1. The results were compared with those obtained in nonoxidative steam reforming (SRM) and
2
2
3
◦
in partial oxidation (POM) of methanol. The catalysts were highly active for OSRM, giving total CH OH conversion at 350–400 C with
3
5
−1
GHSV = 0.6–1.2 × 10 h . They also showed high selectivity, giving H yield of 2.5 mol per mole CH OH, with CO concentration
2
3
below 500 ppm. Besides the main products, CH O and (CH ) O were also observed at all temperatures, while some CH was produced
2
3 2
4
◦
at T > 300 C. The catalytic activity increased with increasing heating rate of the catalyst precursor. High selectivity was observed also in
SRM, but with lower conversion. On the other hand, POM appeared faster than SRM, but produced CO concentrations of a few percent
◦
at T > 300 C, due to CH OH decomposition. The production of CO was strongly reduced in the presence of water vapor. The reaction
3
network was described in detail. The interaction methanol catalyst was studied by IR spectroscopy. IR spectra showed the presence of
adsorbed methoxy groups, that were converted into formate groups at high temperature, and then decomposed into H and CO in the absence
2
of O and/or H O, while in their presence CO was probably oxidized to CO before desorption, due to the action of Cu(II) species.
2
2
2
2004 Elsevier Inc. All rights reserved.
Keywords: Methanol steam reforming; Methanol partial oxidation; Hydrogen production; Cu/ZnO/Al O catalysts; Hydrotalcite; Methanol adsorption;
2
3
Reaction mechanism; Reaction network
1. Introduction
Hydrogen can be produced from methanol by different
reactions. The most simple reaction is endothermic decom-
position that allows the production of 1:2 mol/mol CO and
H2 mixtures: these must be further treated with water to con-
vert CO to CO2 through the water gas shift reaction (WGS).
Methanol appears to be a suitable liquid source for on-
board production of hydrogen in the case of fuel cell electric
engine cars and boats. A suitable process should produce H2
in high concentration to obtain high performances of the fuel
cells [1]. Additionally CO impurity in H2 must be very low
(< 20 ppm) because it poisons the platinum anodes of fuel
cells [2–4].
CH3OH(g) = CO + 2H2, ꢀH0 = 90.6 kJ mol−1
,
(1)
CO + H2O(g) = CO2 + H2, ꢀH0 = − 41.1 kJ mol−1
.
(2)
The second exothermic step can be performed at low temper-
atures in order to limit the CO content in the gas; however,
this reaction does not allow reduction of the CO content
*
Corresponding author. Fax: +39 081 5936936.
E-mail address: turco@unina.it (M. Turco).
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2004.08.029
2004 Elsevier Inc. All rights reserved.

M. Turco et al. / Journal of Catalysis 228 (2004) 56–65
57
below 5000 ppm [5]. The two reactions (1) + (2) can be ob-
tained in one step with the socalled methanol catalytic steam
reforming (SRM), according to [3,6–15]
Some aspects of the OSRM process need further investi-
gation. The reaction network of OSRM is very complex and
not yet defined. In the reacting system, besides SRM and
POM, other reactions, such as combustion and decomposi-
tion of methanol, can also occur [8,24,27,28]. The pathway
of CO formation is widely debated [8,22,27,32,33]. More-
over different points of view are reported regarding the na-
ture of the active sites and reaction mechanisms [6,7,34,35].
In Part I of this paper the preparation and wide charac-
terization of Cu/Zn/Al hydrotalcite-like precursor, obtained
by the method of homogeneous precipitation induced by
urea hydrolysis, and of the derived catalysts have been re-
ported [31]. In this paper we will describe the catalytic ac-
tivity of the materials obtained and the effects of operating
conditions. Reaction paths and mechanistic studies based on
IR experiments will also be described and discussed.
CH3OH(g) + H2O(g) = CO2 + 3H2,
ꢀH0 = 49.5 kJ mol−1
.
(3)
This process allows high hydrogen yield and high con-
centration of H2 in the gaseous mixture (up to 75%). How-
ever, since the reaction is endothermic, the reformer needs
to be heated, and this leads to some drawbacks, such as
starting delay and difficulty in obtaining rapid variations
of speed [3]. Also, due to the relatively high temperature
needed, residual CO cannot be pushed very low because of
the equilibrium (2).
Another widely studied reaction for H2 production is
methanol partial oxidation (POM), according to [16–22]
1
CH3OH(g) + O2 = CO2 + 2H2,
2
2. Experimental
ꢀH0 = −192.3 kJ mol−1
.
(4)
The POM process has the advantages of being fast and
suitable for rapid speed variations. However this process,
compared to SRM, leads to a lower yield to H2 and to lower
H2 concentration. The maximum theoretical H2 concentra-
tion is 67%, but since the use of pure oxygen is not techni-
cally feasible for car traction, air must be used as oxidant, so
the maximum hydrogen concentration is 41%. Other draw-
backs are related to the exothermicity of the reaction: high
release of heat decreases power efficiency, temperature con-
trol of the reactor is difficult, and hot spots can occur in the
catalyst, leading to reduction of activity [23]. On the other
hand, with this process residual CO could be reduced to trace
amounts if the catalyst were highly active for the CO + O2
reaction. In practice, both SRM and POM produce appre-
ciable CO amounts, strongly depending on the nature of the
catalyst and operating conditions [7,24].
Recently a new process based on combination of SRM
and POM reactions, named oxidative steam reforming of
methanol (OSRM) has been proposed [1,23–30]. By feed-
ing CH3OH, H2O, and O2 (air) in proper amounts, the heat
released by POM balances that taken up by SRM and re-
quired for heating the reactants. In this way, an autothermal
process is obtained, thus leading to a more easily controlled
system. This allows delivery of high hydrogen concentra-
tion gas (up to 65%, if air is used) [23] and can be suited
for rapid startup and power variations by properly varying
the methanol/oxygen ratio [1]. Moreover, under proper con-
ditions, OSRM can produce hydrogen containing very low
CO concentrations [24,26,27].
The catalyst precursor was prepared via precipitation of
Zn, Cu, and Al chlorides in the presence of urea, accord-
ing to the previously described method [31]. The precipitate
consisting of Cu/Zn/Al hydrotalcite and Zn/Cu paratacamite,
in the mass weight ratio of 3 to 1, was heated in dry air flow
at rate of 2 or 10 ◦C min−1 up to 450 ◦C and maintained at
this temperature for 12 h. Then the samples were reduced in
situ in 2% H2/He mixture at 450 ◦C for 2 h. The two cat-
alysts, consisting mainly of Cu, Zn, and Al oxides, were
named 2-Cu/Zn/Al and 10-Cu/Zn/Al, respectively, accord-
ing to the heating rate of the precursor. Catalytic activity
measurements were carried out in a laboratory flow ap-
paratus with a fixed-bed reactor operating at atmospheric
pressure. The catalyst (size = 90–110 µm) was diluted
with fused-quartz powders in 1:10 ratio to obtain isother-
mal conditions. The reactor was equipped with two lines
for liquid and gaseous feeding. Liquid feed (H2O/CH3OH)
was introduced by metering pumps and gaseous feed (O2,
He) by electronic flow controllers. A gas chromatograph
HP 5890 equipped with a Porapak/molecular sieve double-
packed column and a TCD detector allowed analysis of H2,
CO (detection limit = 0.05%), CO2, O2, CH4, CH3OH,
and H2O. From the concentrations and flow rates of the
effluent stream, total and partial conversions were calcu-
lated. A Hiden mass spectrometer was also employed for
identification of products not detected by GC. The used
catalyst samples were analyzed for organic compounds
by an LECO Corporation CHN 2000 elemental analyzer.
The OSRM tests were carried out with catalyst mass of
0.045 or 0.09 g (apparent density = 1 g cm−3), at T =
200–400 ◦C, H2O/CH3OH/O2 molar ratios = 1.1/1/0.12
(CH3OH concentration = 17.8%), GHSV = 0.6 × 105,
1.2 × 105 h−1. SRM and POM tests were carried out un-
der the same conditions, but without feeding O2 or H2O,
respectively; methanol decomposition (DEC) was tested ex-
cluding both H2O and O2. Each test lasted 1.5–2 h and in
As described in the first part, the OSRM catalysts pro-
posed in the literature are mostly based on copper dispersed
on metal oxides such as Al2O3 or ZrO2, often with the pres-
ence of ZnO that acts as a promoter [29,31]. Systems derived
from mixed hydroxycarbonates with hydrotalcite structure
appear very promising for high H2 yield and low CO pro-
duction [26,27].

58
M. Turco et al. / Journal of Catalysis 228 (2004) 56–65
this time the products were sampled and analyzed two or
three times: it was verified that conversion values were con-
stant. It was ascertained in preliminary tests that diffusive
resistances were negligible.
The IR spectra were recorded with a Nicolet Protégé
460 Fourier transform instrument. A conventional manipu-
lation/outgassing ramp connected to the IR cell was used.
Pressed disks of pure catalyst powder (15 mg, 2 cm diam-
eter) were used. The sample was thermally pretreated by
outgassing at 450 ◦C in the IR cell. The adsorption proce-
dure involves contact of the activated sample disk with the
methanol vapor at 250 ◦C in situ and outgassing in steps
from 250 ◦C to higher temperatures. In order to better eval-
uate the surface species from all the spectra herein reported
the spectrum of the pretreated catalyst has been subtracted.
3. Results and discussion
3.1. Catalytic activity measurements
In preliminary tests it was ascertained that no reactions
occurred in the absence of catalysts under all the experimen-
tal conditions investigated.
In OSRM tests the reaction products analyzed by GC are
H2, CO2, O2, CH4, CH3OH, and H2O. Carbon monoxide
was in a concentration below the detection limit of our ap-
paratus (0.05%). In addition, MS analysis identified CH2O
and (CH3)2O, while HCOOCH3, alkenes, and higher alka-
nes were not detected. Taking into account that CH2O and
(CH3)2O are not quantitatively analyzed by our GC, the total
amount CH2O + (CH3)2O is evaluated from a mass balance.
Elemental analysis of carbon, carried out on the used cata-
lysts, ascertained that no carbon compounds were present.
The results of OSRM tests on 2-Cu/Zn/Al are reported in
Figs. 1 and 2. Catalytic activity is markedly influenced by
temperature and space velocity. At GHSV = 1.2 × 105 h−1
(Fig. 1a) CH3OH conversion appears negligible at T <
300 ◦C but it increases rapidly with temperature, reaching
90% at 400 ◦C. H2 yield, defined as H2 moles per CH3OH
moles, is negligible at low temperatures and reaches a value
of about 2 at 400 ◦C. Conversions of methanol to differ-
ent products are reported in Fig. 1b, where CxHyO means
the total amount (CH3)2O + CH2O. At low temperatures
conversion to CO2 is negligible while some conversion to
CxHyO is observed, suggesting that methanol dehydration
and/or dehydrogenation is prevailing under these conditions.
No other products are observed up to 300 ◦C. By increas-
ing temperature, conversion to CO2 markedly increases up
to 80% at 400 ◦C while conversion to CxHyO reaches a max-
imum at 350 ◦C and sharply decreases at 400 ◦C. CH4 is
formed in an appreciable amount (5%) only at 400 ◦C. CO
is less than the detection limit (0.05%) at all temperatures.
At GHSV = 0.6 × 105 h−1 (Fig. 2a) CH3OH conversion
shows a similar trend but with higher values in compari-
son to GHSV = 1.2 × 105 h−1, being appreciable even from
Fig. 1. (a) Methanol conversion and hydrogen yield; (b) conversions to dif-
ferent products as a function of temperature under OSRM conditions for the
5
−1
.
catalyst 2-Cu/Zn/Al, GHSV = 1.2 × 10
h
Fig. 2. (a) Methanol conversion and hydrogen yield; (b) conversions to dif-
ferent products as a function of temperature under OSRM conditions for the
5
−1
.
catalyst 2-Cu/Zn/Al, GHSV = 0.6 × 10
h

M. Turco et al. / Journal of Catalysis 228 (2004) 56–65
59
Fig. 3. (a) Methanol conversion and hydrogen yield; (b) conversions to dif-
ferent products as a function of temperature under OSRM conditions for the
Fig. 4. (a) Methanol conversion and hydrogen yield; (b) conversions to dif-
ferent products as a function of temperature under OSRM conditions for the
catalyst 10-Cu/Zn/Al, GHSV = 0.6 × 10
5
−1
5
−1
h .
catalyst 10-Cu/Zn/Al, GHSV = 1.2 × 10
h
.
200 ◦C and complete at 350 ◦C. At the same time H2 yield,
that was negligible up to 300 ◦C with the higher GHSV,
gives a value of about 2 at 300 ◦C, and at 350 ◦C reaches
a value of 2.5: this is close to the theoretical value (2.8) as-
suming a total conversion of CH3OH and O2. Conversion
to CO2 (Fig. 2b) parallels H2 yield, since it is very low
at T < 300 ◦C and sharply increases with temperature up
to about 90% at 350–400 ◦C. CxHyO are prevailing at low
temperatures, showing a maximum at 250 ◦C. It is remark-
able that the formation of these compounds increases by
decreasing GHSV at low temperatures, while the opposite
behavior is observed at high temperatures: this results into
shifting of the maximum of the CxHyO conversion curve
to lower temperatures with the lower GHSV. CH4 conver-
sion is appreciable only at high temperatures and reaches
a value of about 9% at 400 ◦C, which is higher than that
observed at GHSV = 1.2 × 105 h−1. The CO concentration
is below the detectio−n1limit also in the tests carried out at
The results of the OSRM tests on 10-Cu/Zn/Al are re-
ported in Figs. 3 and 4. The reaction products are the same
as those observed for 2-Cu/Zn/Al. Methanol conversions,
hydrogen yield, and partial conversions are comparable to
those of 2-Cu/Zn/Al and appear similarly influenced by tem-
perature and space velocity. However, it can be noted that
the 10-Cu/Zn/Al sample is more active at low temperatures,
as shown by the higher values of CH3OH conversion and
H2 yield: methanol conversion reaches about 50% with H2
yield of 0.7 at 250 ◦C.−I1t is worth noting that at 200 ◦C and
GHSV = 0.6 × 105 h 15% methanol conversion is ob-
served, without any H2 production, CxHyO being the only
reaction products, suggesting that at low temperatures only
dehydration and dehydrogenation occur. A similar behavior
is exhibited by 2-Cu/Zn/Al, but to a lower extent.
These results show that the behavior of the two samples is
not substantially different, although the sample 10-Cu/Zn/Al
appears more active at low temperatures. This could be re-
lated to different redox properties, as shown by TPR and
TPO measurements [31]. In fact the sample 10-Cu/Zn/Al
has a larger fraction of Cu species that undergo reduction
and oxidation at low temperatures and can be catalytically
active. This is probably related to the higher Cu dispersion
of this catalyst [31]. The correlation between the Cu reduc-
tion temperature and the catalytic activity was reported by
other authors [29]. Another effect could be taken into ac-
count to explain the different activity at low temperatures.
The samples 2-Cu/Zn/Al and 10-Cu/Zn/Al have markedly
different surface areas, 73 and 111 m2 g−1, respectively [31].
The higher surface area, which is probably due mainly to
amorphous alumina, corresponds to a higher amount of sur-
face acid sites: this condition could lead to higher activ-
ity for dehydration, thus explaining the higher activity at
200 ◦C. The activity of the present catalysts, evaluated as H2
yield and CO production, can be compared with literature
data on similar systems. Velu et al. [34] studied a Cu/Zn/Al
catalyst (at. ratio = 37/51/12) derived from hydrotalcite–
GHSV = 0.6 × 105 h
.

60
M. Turco et al. / Journal of Catalysis 228 (2004) 56–65
Fig. 5. Methanol conversion, hydrogen yield, and conversions to different
products as a function of temperature under SRM conditions for the catalyst
5
−1
.
10-Cu/Zn/Al, GHSV = 0.6 × 10
h
aurichalcite phases and tested under similar operating condi-
tions and found methanol conversion of about 60–65%, with
H2 yield of about 2 mol mol−1 in the temperature range 200–
260 ◦C. The CO concentration in the product stream was
300–1200 ppm. With a catalyst of similar composition but
under somewhat different conditions Agrell et al. [24] re-
port CH3OH conversions of 50–90% and CO concentrations
of 300–1300 ppm at temperatures of 270–325 ◦C. Murcia-
Mascarós et al. [26] report data for a hydrotalcite-derived
Cu/Zn/Al catalyst with compositions more similar to the
present system (at. ratio = 22/37/41), tested under simi-
lar conditions, obtaining methanol conversions of about 50–
70% and H2 yield of 1.6 mol mol−1 at 270–325◦C: the CO
content in the effluent stream was 1500–4000 ppm. Shen
and Song [29] studied a coprecipitated Cu/Zn/Al catalyst
(at. ratio = 45/43/12) with a similar feed composition, but
much lower GHSV than ours (∼22,000 h−1) obtaining con-
version of 99% at 230 ◦C and CO concentration of about
500 ppm. Also supported systems were studied: alumina
supported Cu/Zn catalysts [23] gave high methanol conver-
sion (up to 90% at 300 ◦C) and H2 yield (up to 2.8) with
GHSV = 25,000 h−1, however, with higher CO concen-
trations (1.1%). On the basis of these results, taking into
account differences in catalyst composition and experimen-
tal conditions, it may be concluded that Cu/Zn/Al catalysts
obtained from hydrotalcite-like precursors compare favor-
ably, in terms of high H2 yield and low CO production,
with those obtained by impregnation or hydroxide precip-
itation. It is likely that catalytic properties are improved
by the method of homogeneous precipitation employed for
preparation of the precursor, because this method leads to
higher surface areas and more homogeneous materials, as
suggested by the previous XRPD and TPR characterization
of these systems [31]. A recent paper [36] dealing with the
SRM reaction has evidenced better catalytic performances
of Cu/Zn/Al catalysts prepared by the method of urea hy-
drolysis in comparison with similar catalysts obtained by the
coprecipitation method.
Fig. 6. (a) Methanol conversion and hydrogen yield; (b) conversions to dif-
ferent products as a function of temperature under DEC conditions for the
5
−1
.
catalyst 10-Cu/Zn/Al, GHSV = 0.6 × 10
h
For a better understanding of the reaction path, tests
of simple steam reforming (SRM), decomposition (DEC),
and partial oxidation (POM) of methanol are carri−ed1 out
on the catalyst 10-Cu/Al/Zn at GHSV = 0.6 × 105 h and
the results are plotted in Figs. 5–7. As expected, in SRM
tests (Fig. 5) lower conversions of methanol than under
OSRM conditions are found at all temperatures; CO2 and
H2 are selectively produced due to very limited formation
of methane and CxHyO that are produced in some amount
only at 250 ◦C. The concentrations of CO are below the
detection limit also under these conditions. In DEC tests
(Fig. 6) the catalyst appears less active than in SRM tests.
The decomposition products, H2 and CO, are formed in ap-
preciable amounts only at T > 300 ◦C. The maximum H2
yield, attained at 400 ◦C, is about 1, which is lower than
that expected from methanol conversion, taking into account
the decomposition stoichiometry. This is explained by the
occurrence of side reactions. In fact, besides CO and H2,
also CxHyO, CO2, and traces of CH4 are produced. CxHyO
are the only reaction products at T ꢀ 300 ◦C, while small
amounts of CO2 are formed at high temperatures. The cat-
alyst is tested also for POM reaction, and the results are
shown in Fig. 7. The production of CO2 and H2 starts at
300 ◦C, while at lower temperatures CxHyO are the only
reaction products. CO2 and H2 are formed at 300 ◦C in
amounts corresponding to complete O2 consumption and
stoichiometry of Eq. (2). By increasing temperature H2 yield
increases up to 1.6 mol mol−1 at 400 ◦C: at the same time

M. Turco et al. / Journal of Catalysis 228 (2004) 56–65
61
Fig. 8. TPR profile of fresh (oxidized) and used 10-Cu/Zn/Al sample.
Methanol total oxidation,
3
CH3OH + O2 = CO2 + 2H2O,
(5)
2
can occur on CuO-based catalysts [37]. It is hypothesized
that in OSRM with Cu catalysts methanol combustion oc-
curs rapidly, so oxygen is consumed in the first region of
the reactor, and the SRM reaction begins when oxygen has
been consumed [24,28]. Previous TPR/TPO measurements
[31] suggest that under OSRM conditions Cu(0), Cu(I), and
Cu(II) species can be present, depending on temperature and
O2 and H2 concentrations. Therefore in the zone where O2
concentration is high, copper is in an oxidized state, and
methanol oxidation prevails. On the other hand, in the zone
where O2 is low and H2 concentration is high, copper is in
a reduced form and steam reforming becomes the dominant
reaction. Reaction (5) is not a parasitic reaction because, if it
is followed by reaction of produced H2O with CH3OH, the
net result is POM reaction, as already observed [1,22].
With the aim to investigate the Cu oxidation state of the
catalysts under OSRM conditions, TPR measurements have
been carried out on the samples after OSRM tests. In Fig. 8
the TPR profile of the sample 2-Cu/Zn/Al is reported to-
gether with that of the oxidized sample, for comparison. It
can be observed that the spectrum of the used sample shows
a low intensity signal at 240 ◦C. This signal contains differ-
ent components that cannot be explained only by stepwise
reduction of some Cu(II) to Cu(0), but are probably related
to the presence of copper species with different oxidation
states. The total H2 consumption is 0.2 mol mol−1, which
corresponds to a mean Cu oxidation state of 0.4, as if 20%
of Cu were oxidized to Cu(II) (but probably Cu(I) and Cu(II)
are present together). This indicates that under OSRM con-
ditions copper is present in both oxidized and reduced states,
but the Cu(0) state prevails. These data seem to confirm that
there is a zone of the catalytic bed in which the catalyst is
oxidized, and another zone in which it is reduced, and the
latter zone is larger. This agrees with the faster kinetics of
the oxidation reactions compared to that of steam reforming.
Moreover it is worth noting that the TPR peak of the used
samples appears at lower temperatures and with a shape dif-
ferent from that of the fresh calcined sample, showing that
Fig. 7. (a) Methanol conversion and hydrogen yield; (b) conversions to dif-
ferent products as a function of temperature under POM conditions for the
5
−1
.
catalyst 10-Cu/Zn/Al, GHSV = 0.6 × 10
h
conversion to CO increases, while conversion to CO2 is al-
most constant in the range 300–400◦C. This suggests that at
T > 300 ◦C reaction (2) is complete and residual methanol
undergoes decomposition into CO and H2.
From these results we can observe that SRM is faster
than decomposition, being activated at lower temperature
(Figs. 5 and 6) and POM is faster than SRM, since at 300 ◦C
conversion to CO2 is higher in POM than in SRM tests
(Figs. 5 and 7), although the low O2 partial pressure lim-
its the extent of the reaction. The products CxHyO prevail
at low temperatures, suggesting faster dehydrogenation and
dehydration reactions. Therefore it can be supposed that the
activation energies of SRM, POM, methanol decomposition
(DEC), dehydration (DEHD), and dehydrogenation(DEHG)
increase in the following order:
DEHD, DEHG < POM < SRM < DEC.
At low temperatures (200–250◦C) the kinetics of SRM
and POM is low and only dehydration and dehydrogenation
occur to some extent. By increasing temperature, first POM
and after SRM reactions are activated. POM can go forward
to a small extent due to the limited concentration of O2, and
then residual methanol reacts through SRM or DEC.
3.2. The reaction network
The above data show that OSRM involves a complex net-
work of parallel and series reactions.

62
M. Turco et al. / Journal of Catalysis 228 (2004) 56–65
Cu(II) species formed by oxidation of Cu(0) have properties
different from those present in the original oxidized sam-
ple. A similar effect due to the reduction–reoxidation cycle
was already observed for CuO/ZnO/Al2O3 catalysts and can
be attributed to a structural modification that improves cop-
per dispersion [24]. However the precise mechanism of such
modification is not clear.
As observed in the Section 1, the steam reforming reac-
tion (3) is formally the sum of methanol decomposition (1)
and water gas shift (2): it is debated if SRM really occurs
through the sequence of reactions (1) and (2) [34,38] or is
a single reaction [1,7,8]. It is well known that both reac-
tions (1) and (2) can be catalyzed by copper-based catalysts
[7,39]; however, if this sequence were the main pathway for
SRM, on the basis of equilibrium calculation, the expected
CO concentration would be about 1% at 200 ◦C and 5% at
400 ◦C under our experimental conditions. Since all OSRM
and SRM tests produced CO concentrations well below these
values, we conclude that the sequence of reactions (1) and
(2) is not the main pathway for SRM reaction, and this is
probably a single reaction, in agreement with other studies
[1,7,8].
Fig. 9. Reaction network under OSRM conditions.
has been reported [42]. This reaction can contribute to keep
the CO concentration at a low level. Methane is formed in
appreciable amounts in OSRM tests at high temperatures
(Figs. 1–4), while it is negligible in DEC and POM tests, al-
though CO production is high under these conditions. This is
probably due to the low H2 concentrations that reduce the ki-
netics of methanation. Also under SRM conditions methane
is negligible, probably due to very low CO production, as ob-
served above. The formation of appreciable amounts of CH4
in OSRM tests can be explained by the higher methanol con-
versions, in comparison with SRM tests. This leads to lower
H2O concentrations: under these conditions the unfavorable
effect of steam on CO production is reduced, so that appre-
ciable CO amount can be formed. This is readily converted
into CH4, due to the high H2 concentration.
An other open question is the pathway for CO formation.
As above observed, CO is formed in appreciable amounts in
POM and DEC tests at T > 300 ◦C. It is evident that in DEC
tests CO is formed by methanol decomposition (1), possibly
through intermediate dehydrogenation to formaldehyde (6),
CH3OH = CH2O + H2,
and probably the same reaction(s) occur in POM tests.
(6)
On the other hand, in SRM and OSRM tests the formation
of CO is reduced below the detection limit: this behavior can
be attributed to the presence of water vapor. A decrease of
CO production from methanol due to the presence of water
vapor is reported in recent works on Cu catalysts [24,40] and
this effect is explained by the WGS reaction (2) consuming
CO [40]. However, as observed above, the WGS reaction is
unable to reduce the CO concentration to very low values as
found in our SRM and OSRM tests. As a possible explana-
tion, we hypothesize that the presence of water vapor shifts
the methanol reaction toward CO2 rather than CO by modi-
fying the catalyst properties, that is causing some oxidation
of Cu, as discussed in the following (see reaction mecha-
nism).
According to some authors [1,24,32] CO can be produced
under SRM or OSRM conditions by the reverse of the WGS
reaction (2). It is obvious that this reaction is very limited
under the present conditions. Probably this reaction is slower
than SRM, as also suggested by [1,41].
The formation of methane as by-product of OSRM was
not considered in previous works. Methane can be formed
by the reaction (7)
On the basis of the above discussion, the following net-
work for OSRM can be proposed (Fig. 9). The scheme also
includes methanol condensation to dimethylether that oc-
curs probably on acid sites of Al2O3, being a well-known
acid-catalyzed reaction, and oxidative dehydrogenation to
formaldehyde (8), which is known to occur on Cu-based cat-
alysts [43]:
1
CH3OH + O2 = CH2O + H2O.
(8)
2
Reaction (8) could explain the results obtained in OSRM
tests at low temperatures, that is, some methanol conver-
sion with no H2 or CO2 production (Figs. 2 and 4). More-
over reactions consuming CH2O and (CH3)2O are also taken
into account. In fact the maxima observed in the curves of
conversion to CxHyO (Figs. 1–4) may be due to reactions
consuming these products, that prevail at high temperature.
These can be steam reforming reactions (9) and (10):
CH2O + H2O = CO2 + 2H2,
(CH3)2O + 3H2O = 2CO2 + 6H2.
(9)
(10)
Reaction (9) can occur on Cu-based catalysts, as sug-
gested by mechanistic studies [32]. Reaction (10) is also
reported for Cu-based catalysts [44].
CO + 3H2 = CH4 + H2O
(7)
that is highly favored in the whole temperature range. The
activity of CuO/ZnO-based catalysts for CO methanation
Obviously, the scheme in Fig. 9 does not include oxi-
dation reactions, that can involve all combustible species:

M. Turco et al. / Journal of Catalysis 228 (2004) 56–65
63
Fig. 10. FTIR subtraction spectra; the spectrum of the 10-Cu/Zn/Al catalyst after reduction was subtracted from the spectra of pure powder: (a) surface species
◦
◦
◦
after contact of the catalyst with CH OH at 250 C for 10 min; (b) surface evolution after 10 min at 300 C; (c) surface evolution after 10 min at 350 C;
(d) surface evolution after 10 min at 400 C.
3
◦
CH3OH, CH2O, (CH3)2O, H2, CO, CH4. These reactions
are limited by the defective O2 amount and can occur only
in the reactor zone where O2 is still present.
Eventual reactions producing carbonaceous compounds,
such as polymerization or condensations, can be excluded
since the elemental analysis of the catalysts used in the tests
has shown the absence of carbon compounds.
This spectrum is typical of formate ions [45] and have to
be assigned to asymmetric stretching of the O–C–O group
(1598 cm−1), the CH deformation mode (1381 cm−1), and,
finally, to the symmetric O–C–O stretching (1376 cm−1).
Under the same conditions of Fig. 10a, the gas phase
species evolved from the catalyst have been also studied. The
spectrum of the gas in contact with the catalyst at 250 ◦C is
dominated by the features of gaseous methanol, although the
features of gaseous water are clearly also present. Actually,
a very weak band at 2140 cm−1 due to gaseous CO and two
weak features near 1178 and 1103 cm−1, likely belonging to
dimethylether gas, can also be detected.
3.3. FTIR study of methanol adsorption and reaction
mechanism
The IR spectrum of the adsorbed species arising from
the interaction of methanol at 250 ◦C with the prereduced
catalyst (10-Cu/Zn/Al sample) is shown in Fig. 10a. In the
low-frequency side of the spectrum a complex band with
two main maxima at 1031 and 1088 cm−1 is observed.
However, a well-defined component is observed, weak, at
1191 cm−1, while other components are envisaged near
1120 and 950 cm−1. In the CH stretching region weak
bands with two sharp maxima at 2953 and 2839 cm−1 and
a weaker broad maximum in the middle (2889 cm−1) are
also found, corresponding to a sharp CH3 deformation band
at 1469 cm−1. These features can be assigned to adsorbed
methoxy groups. The multiplicity of the band in the re-
gion 1200–900 cm−1 is due very likely to the presence of
two or more kinds of methoxy groups. As for comparison,
we can note that the C–O stretching of methoxy groups on
γ -Al2O3 is found at 1095 cm−1 [45], and those on ZnO- and
ZnO-based systems [46] at 1060 cm−1. Those on CuO are
reported to be very labile, whose C–O stretching is found
at 1055–1070 cm−1, and finally, those on Cu(111) [47] at
1036 cm−1. The sharper component at 1191 is most likely
due to the rocking CH3 mode of the same methoxy groups.
However, already at low temperatures we can find sharp
By increasing temperature in the IR cell, the spectra of
adsorbed species change as reported in Figs. 10b–d. It is
evident that the spectra of both methoxy groups and for-
mate groups decrease progressively in intensity. However,
while the bands of formate groups are practically fully dis-
appeared at 400 ◦C, the main band of methoxy groups at
1100–900 cm−1 is still present, although much weaker and
with the main maximum shifted down to 1016 cm−1. In par-
allel the band now observed near 1458 cm−1 seems to have
grown, likely due to the formation of carbonate species. Af-
ter heat treatment at 300 and 350 ◦C a weak sharp band
at 1690 cm−1 is evident. It can be assigned to the C=O
stretching of an adsorbed formaldehyde species [45]. The
presence of dioxymethylene species, a likely intermediate
in the conversion to methoxy groups into formates, cannot
be excluded, due to the presence of weak additional compo-
nents at 1245 cm−1 and in the range 1180–1090 cm−1 [48].
Looking at the gas evolving from the surface (Fig. 11),
only very small amounts of methanol are desorbed at 300 ◦C,
CO being mostly produced up to 400 ◦C. As it is well known,
hydrogen gas is IR undetectable.
The IR spectra reported above show that the adsorption
of methanol on the catalyst is dissociative. Methoxy groups
strong bands at 1598, 1381 (shoulder), and 1376 cm−1
.

64
M. Turco et al. / Journal of Catalysis 228 (2004) 56–65
◦
◦
◦
◦
Fig. 11. FTIR spectra of the gas phases evolved after 10 min at (a) 250 C; (b) 300 C; (c) 350 C; (d) 400 C.
other hand, both CO and CO2 are produced in noticeable
amounts under POM conditions. To explain this we can re-
sume the data concerning CO adsorption on Cu species, as
determined by IR spectroscopy. Previous data [31] show that
CO is adsorbed very weakly on Cu metal particles, while it is
adsorbed very strongly on Cu(I) centers. On the other hand,
Cu(II) easily oxidizes CO to CO2. It seems quite likely that
upon DEC experiments, the catalyst is fully reduced. Under
these conditions CO is formed by formate species decompo-
sition and is easily desorbed as the product. On the contrary,
under the conditions of POM the catalyst is, at least par-
tially, oxidized. In this case Cu(I) strongly adsorbs CO while
Cu(II) allows its oxidation to CO2 before its desorption. The
true product of POM is consequently CO2. However, due to
the substoichiometric feed in POM tests, it can be supposed
that oxygen is consumed in the upper reactor zone, so the
catalyst in the lower zone is reduced. This means that af-
ter O2 is consumed, excess methanol decomposes (partially)
producing CO. A somewhat different behavior is observed
under SRM conditions, since CO production is negligible in
this case. We can suppose that the presence of excess wa-
ter vapor oxidizes the catalyst, at least partially, leading to
formation of Cu(II) species that produce CO2 at the expense
of CO. Such oxidation should be only partial, according to
TPR measurements on the used catalysts that show a low
mean oxidation state of Cu, as seen above.
Fig. 12. Reaction mechanism of CH OH decomposition.
3
are formed that apparently adsorb both on the active sites
(which are likely associated to Cu species) and over inert
regions of the catalyst surface, likely mainly associated to
alumina. Part of these methoxy groups, likely those located
near or on copper sites, transform easily to formate species,
possibly with the intermediacy of dioxymethylene species.
Traces of formaldehydeadsorbed as such are also detectable.
At 300 ◦C or above formate species decompose and, under
the conditions of our experiment, mainly give rise to gas-
phase CO. Dimethylether is also found in the gas phase.
The mechanism of methanol decomposition, consequently,
is reasonably proposed as formed by the reaction scheme of
Fig. 12, as proposed previously [46].
It seems quite reasonable to suppose that the mechanisms
for DEC, POM, and SRM are closely related, and that the
surface phenomena are similar. However, the flow catalytic
data give some interesting information. In particular, the
conversion of methanol in the presence of both water (SRM)
and oxygen (POM) is higher than under the conditions of
DEC. This could be due to an effect of these substances
on the oxidation state of Cu. Moreover, CO is produced in
relative low amounts in the presence of water (SRM condi-
tions), CO2 being the largely predominant product. On the
4. Conclusions
Cu–Zn–Al catalysts derived from a hydrotalcite-like pre-
cursor have been studied for oxidative reforming of CH3OH.
Good results, in terms of CH3OH conversion and H2 yield,
together with low CO production, have been obtained,
showing that the method of homogeneous precipitation of
hydrotalcite-like precursors is suitable for preparing high

M. Turco et al. / Journal of Catalysis 228 (2004) 56–65
65
activity catalysts. Methanol decomposition appears a reac-
tion slower than steam reforming and partial oxidation and
is greatly depressed by water vapor. A noticeable catalytic
activity for methanation has been evidenced in high temper-
ature tests: this property appears unfavorable for hydrogen
yield, but could be advantageous for developing catalysts
with very low CO selectivity.
The reaction network involves also dehydrogenation to
formaldehyde and dehydration to dimethylether, the WGS
reaction, besides oxidation reactions.
[14] Y. Liu, T. Hayakawa, K. Suzuki, S. Hamakawa, T. Tsunoda, T. Ishii,
M. Kumagai, Appl. Catal. A 223 (2002) 137.
[15] Y.H. Chin, R. Dagle, J. Hu, A.C. Dohnalkova, Y. Wang, Catal. To-
day 77 (2002) 79.
[16] M.L. Cubeiro, J.L.G. Fierro, J. Catal. 179 (1998) 150.
[17] J. Agrell, K. Hasselbo, K. Jansson, S.G. Järas, M. Boutonnet, Appl.
Catal. A 211 (2001) 239.
[18] Z. Wang, W. Wang, G. Lu, Int. J. Hydrogen Energy 28 (2002) 151.
[19] R.M. Navarro, M.A. Pena, J.L.G. Fierro, J. Catal. 212 (2002) 112.
[20] Z. Wang, W. Wang, G. Lu, J. Mol. Catal. A: Chem. 191 (2003) 123.
[21] J. Agrell, G. Germani, S.G. Jaras, M. Boutonnet, Appl. Catal. A 242
(2003) 233.
The reaction mechanism has been studied by IR tech-
nique, that evidenced different species formed by adsorption
of methanol on the catalyst. The hypothesized mechanism
points to an important role of Cu(I) and Cu(II) species: Cu(I)
strongly adsorbs CO, while Cu(II) catalyzes CO oxidation to
CO2. The oxidation state of Cu species is influenced by the
presence of O2 and H2O.
Work is in due course in our Laboratories to study the ef-
fect of the preparation methods of the Cu–Zn–Al precursors
on the selectivity and activity of the OSRM catalysts thereby
obtained.
[22] J. Agrell, M. Boutonnet, J.L.G. Fierro, Appl. Catal. A 253 (2003) 213.
[23] B. Lindström, L.J. Pettersson, P.G. Menon, Appl. Catal. A 234 (2002)
111.
[24] J. Agrell, H. Birgersson, M. Boutonnet, I. Meliàn-Cabrera, R.M. Na-
varro, J.L.G. Fierro, J. Catal. 219 (2003) 389.
[25] S.H. Chan, H.M. Wang, J. Power Sources 126 (2004) 8.
[26] S. Murcia-Mascarós, R.M. Navarro, L. Gómez-Sainero, U. Costantino,
M. Nocchetti, J.L.G. Fierro, J. Catal. 198 (2001) 338.
[27] S. Velu, K. Suzuki, M.P. Kapoor, F. Ohashi, T. Osaki, Appl. Ca-
tal. A 213 (2001) 47.
[28] T.L. Reitz, P.L. Lee, K.F. Czaplewski, J.C. Lang, K.E. Popp, H.H.
Kung, J. Catal. 199 (2001) 193.
[29] J.-P. Shen, C. Song, Catal. Today 77 (2002) 89.
[30] S. Liu, K. Takahashi, M. Ayabe, Catal. Today 87 (2003) 247.
[31] M. Turco, G. Bagnasco, U. Costantino, F. Marmottini, T. Montanari,
G. Ramis, G. Busca, J. Catal. 228 (2004) 43.
Acknowledgments
[32] J.P. Breen, F.C. Meunier, J.R.H. Ross, Chem. Commun. (1999) 2247.
[33] F. Raimondi, K. Geissler, J. Wambach, A. Wokaun, Appl. Surf.
Sci. 189 (2002) 59.
[34] S. Velu, K. Suzuki, M. Okazaki, M.P. Kapoor, T. Osaki, F. Ohashi,
J. Catal. 194 (2000) 373.
[35] I.A. Fisher, A.T. Bell, J. Catal. 184 (1999) 357.
[36] T. Shishido, Y. Yamamoto, H. Morioka, K. Takaki, K. Takehira, Appl.
Catal. A 263 (2004) 249.
This research was founded by Ministero Istruzione Uni-
versità e Ricerca and Genova, Napoli and Perugia Univer-
sities, in the framework of Programmi di Ricerca di Inter-
esse Nazionale, 2002 No. 038793. The authors acknowledge
Mr. Luciano Cortese for his contribution to catalytic activity
tests.
[37] T.L. Reitz, S. Ahmed, M. Krumpelt, R. Kumar, H.H. Kung, J. Mol.
Catal. A: Chem. 162 (2000) 275.
[38] J.C. Amphlett, M.J. Evans, R.F. Mann, R.D. Weir, Can. J. Chem.
Eng. 66 (1988) 950.
References
[39] B.A. Peppley, J.C. Amphlett, L.M. Kearns, R.F. Mann, Appl. Ca-
tal. A 179 (1999) 31.
[40] Y. Choi, H.G. Stenger, Appl. Catal. B 38 (2002) 259.
[41] H. Purnama, T. Ressler, R.E. Jentoft, H. Soerijanto, R. Schlögl, R.
Schomäcker, Appl. Catal. A 259 (2004) 83.
[42] V.I. Sharkina, G.I. Salomatin, N.N. Aksenov, V.S. Sobolevskii, Kinet.
Katal. 25 (1984) 765.
[43] C.N. Satterfield, Heterogeneous Catalysis in Practice, McGraw-Hill,
New York, 1980, p. 195.
[44] K. Nakamura. Y. Hiramatsu, JP 2003038957 (2003), to Mitsubishi Gas
Chemical Co., Ltd., CAN: 138:156226.
[45] G. Busca, P.F. Rossi, V. Lorenzelli, M. Benaissa, J. Travert, J.C. Laval-
ley, J. Phys. Chem. 89 (1985) 5433.
[46] A. Riva, F. Trifirò, A. Vaccari, L. Mintchev, G. Busca, J. Chem. Soc.,
Faraday Trans. 1 84 (1988) 1423.
[47] M.A. Chester, E.M. McCash, Spectrochim. Acta A 43 (1987) 1625.
[48] G. Busca, J. Lamotte, J.C. Lavalley, V. Lorenzelli, J. Am. Chem.
Soc. 109 (1987) 5197.
[1] K. Geissler, E. Newson, F. Vogel, T.B. Truong, P. Hottinger, A. Wo-
kaun, Phys. Chem. Chem. Phys. 3 (2001) 289.
[2] X. Liu, O. Korotkikh, R. Farrauto, Appl. Catal. A 226 (2002) 293.
[3] P.J. de Wild, M.J.F.M. Verhaak, Catal. Today 60 (2000) 3.
[4] D.J. Moon, K. Sreekumar, S.D. Lee, B.G. Lee, H.S. Kim, Appl.
Catal. A 215 (2001) 1.
[5] B. Lindström, L.J. Pettersson, J. Power Sources 118 (2003) 71.
[6] R.O. Idem, N.N. Bakhshi, Chem. Eng. Sci. 51 (1996) 3697.
[7] J.P. Breen, J.R.H. Ross, Catal. Today 51 (1999) 521.
[8] B.A. Peppley, J.C. Amphlett, L.M. Kearns, R.F. Mann, Appl. Ca-
tal. A 179 (1999) 21.
[9] L. Ma, B. Gong, T. Tran, M.S. Wainwright, Catal. Today 63 (2000)
499.
[10] Y.M. Lin, M.H. Rei, Catal. Today 67 (2001) 77.
[11] A.P. Tsai, M. Yoshimura, Appl. Catal. A 214 (2001) 237.
[12] T. Takahashi, M. Inoue, T. Kai, Appl. Catal. A 218 (2001) 189.
[13] M.M. Günter, T. Ressler, R.E. Jentoft, B. Bems, J. Catal. 203 (2001)
133.