Mechanistic aspects of the steam reforming of methanol over a CuO/ZnO/ZrO2/Al2O3 catalyst

Article abstract of DOI:10.1039/a906393e

CO is not a primary but is a secondary product of the steam reforming of methanol over a CuO/ZnO/ZrO₂/Al₂/O₃ catalyst, CO is formed as a secondary product by the reverse WGS reaction only when the methanol has reacted.

Full text of DOI:10.1039/a906393e

Mechanistic aspects of the steam reforming of methanol over a
CuO/ZnO/ZrO /Al O catalyst
2
2
3
John P. Breen,* Frederic C. Meunier and Julian R. H. Ross
Centre for Environmental Research, University of Limerick, Limerick, Ireland. E-mail: john.paul.breen@ul.ie
Received (in Cambridge, UK) 5th August 1999, Accepted 11th October 1999
CO is not a primary but is a secondary product of the steam
reforming of methanol over a CuO/ZnO/ZrO /Al cata-
lyst, CO is formed as a secondary product by the reverse
WGS reaction only when the methanol has reacted.
reaction were analysed by GC equipped with a Porapak Q
column. This column gave satisfactory separation of CO and N
2
at a column temperature of 25 °C, the limit of detection for CO
was ca. 1000 ppm with an accuracy of ±2%. The reaction was
carried out in a quartz plug-flow microreactor, the composition
2
2
O
3
The steam reforming of methanol is attracting growing interest
as a means of producing hydrogen for use in fuel cells. CO
produced in the steam reforming reaction is currently a poison
to the promising proton-exchange membrane (PEM) fuel cells
and as a result much attention has been focused on the
mechanism of CO formation during the reaction. The formation
of carbon monoxide as a product and the fact that copper is a
of the reactant gases being 28.5% H
balance of N . Prior to introduction of the reaction mixture, the
catalyst was reduced in situ in a 5% H –N mixture at a
temperature of 240 °C for 4 h, ramping to this temperature at a
rate of 5 °C min . IR measurements were carried out using a
Nicolet Magna 550 FTIR fitted with a diffuse reflectance cell
and incorporating a high-pressure, high-temperature catalytic
chamber.
2 3
O, 21.5% CH OH, with a
2
2
2
2
1
1
good catalyst for the water–gas shift (WGS) reaction has led
–5
some workers2 to speculate that a decomposition/WGS
reaction sequence is involved [eqn. (1) and (2)]:
The effect of contact time on the distribution of reaction
products was studied by varying the W/F ratio while using a
constant feed composition at two different temperatures (200
and 300 °C). Fig. 1 shows the methanol conversions and the
percentages of hydrogen, carbon dioxide and carbon monoxide
in the product stream as a function of the (pseudo-)contact time,
CH
CO + H
CH OH + H
3
OH ? CO + 2H
O " H + CO
O ? CO + 3H
2
(1)
(2)
(3)
2
2
2
3
2
2
2
2
3
W/F (g min cm ), at 300 °C. Two equilibrium lines for the
carbon dioxide and hydrogen production are also shown as
dotted lines; the top line (line ‘m’) in each case represents the
equilibrium values obtained by excluding carbon monoxide
from the calculations whereas the bottom line (line ‘n’) shows
calculated equilibrium values with the inclusion of carbon
monoxide as a product. It must be noted that the equilibrium
lines ‘m’ and ‘n’ correspond to the situation when complete
conversion of methanol is assumed; the equilibrium values
shown in Fig. 1 are the results of thermodynamic calculations
carried out using the HSC Chemistry for Windows software
package (Outokumpu Research Oy). This software uses the
Gibbs Free Minimisation Method to calculate equilibrium
However, Amphlett et al.6,7 have more recently noted that
carbon monoxide production is less than that predicted from
equilibrium. As a result, they have revised their reaction scheme
to one containing parallel reactions, the steam reforming of
methanol [eqn. (3)] occurring in parallel to the decomposition of
methanol to CO and H
of the CO/CO ratio also occurs as a result of the WGS reaction
eqn. (2)]. Peppley et al.8 found that the methanol decomposi-
2
[eqn. (1)]; they suggest that adjustment
2
,9
[
tion reaction is much slower than the steam reforming reaction
but argued that it must be included in the overall reaction
scheme. Takahashi et al.10 found that methyl formate was
formed at low water/methanol ratios and so they incorporated
methyl formate into their reaction scheme [eqn. (4)–(6)]:
14
compositions.
At 200 °C, 100% conversion of methanol was not attained
and carbon monoxide was not detected in the product stream at
any of the pseudo-contact times.
2
CH
OCHO + H
HCOOH ? CO
3
OH ? CH
O ? HCOOH + CH
+ H
3
OCHO + 2H
2
(4)
(5)
(6)
CH
3
2
3
OH
2
2
Jiang et al.11,12 carried out a detailed kinetic analysis of the
steam reforming reaction and found that the reaction over a Cu/
ZnO/Al O catalyst could also be explained by eqn. (4)–(6),
2 3
they suggested that the carbon monoxide observed in the
product stream is produced via the decomposition of methyl
formate. Studies of the surface mechanism of the steam
9
reforming reaction have been limited and authors have instead
tended to use the large quantity of data published on the
methanol synthesis reaction to support their mechanistic surface
9,12
reaction models.
We have previously reported13 that copper–zirconia contain-
ing catalysts are very active for the steam reforming of methanol
2
and that they give good selectivities towards CO production.
We have now obtained clear evidence from both kinetic
measurements and in situ DRIFTS experiments for the
formation of CO as a secondary product during the reforming
reaction over a CuO/ZnO/ZrO
surface species involved in the reaction are given here.
The CuO/ZnO/ZrO /Al catalyst was prepared by sequen-
2 2 3
/Al O catalyst and details of the
Fig. 1 The influence of W/F on the product compositions at T = 300 °C over
a CuO/ZnO/ZrO /Al O catalyst, m = equilibrium excluding CO from
2 2 3
2
2 3
O
tial precipitation, the order and method of precipitation having
calculations, n = equilibrium including CO in calculations (H O/CH OH =
2
3
been described elsewhere.13 The reactants and products of the
1.3, P = 101 kPa).
Chem. Commun., 1999, 2247–2248
2247
This journal is © The Royal Society of Chemistry 1999

catalyst and then flushing through with Ar. The peaks at 1590
2
1
and 1380 cm can be attributed to a formate species, the peak
2
1
at 1150 cm to p bonded formaldehyde and the peak centred
2
1
15
around 1050 cm to a methoxy species. Water was then
introduced and the spectrum [Fig. 2, spectrum (2)] taken after
the introduction of the water showed that neither the methoxy
nor the formaldehyde peaks were evident but that the intensity
of the formate peaks had increased. This is illustrated in
spectrum (3) of Fig. 2 which shows the difference between
spectra (2) and (1). This indicates that the methoxy and/or the
formaldehyde species were readily converted to formates by
reaction either with gas phase water or more probably with OH
groups generated by the dissociative adsorption of water. In situ
DRIFTS analysis of species during the steam reforming reaction
showed that in addition to surface formates, there was an
2
1
absorption centred at 1033 cm
which can probably be
attributed to gas phase methanol in addition to some adsorbed
methoxy species. Gas phase CO (2364 cm ), was also
2
2
1
detected but there was no evidence of either gas phase CO or
adsorbed CO species. Formates are known to decompose on
CuO/ZnO/ZrO aerogels according to eqn. (7) to give either
2
CO and surface OH groups or CO and H :
2 2
Fig. 2 FTIR spectra of surface species generated by methanol adsorption
followed by water adsorption on a CuO/ZnO/ZrO /Al catalyst at
40 °C.
16
2
2 3
O
2
OH(s) + CO(g)/HCOO2(s)?CO2(g) + 1/2H2(g)
(7)
The results of the kinetic study, however, showed that CO did
not form when methanol was present and that CO and H were
At 300 °C, the percentages of hydrogen and carbon dioxide
2
2
produced were higher than those at 200 °C for all values of
contact time. Carbon monoxide was not detected at the shorter
contact times; however, as the value of W/F was increased,
carbon monoxide was detected and its molar percentage
increased steadily with increasing contact time, approaching,
but not attaining, the predicted equilibrium value at higher
values of W/F. It is evident from Fig. 1 that carbon monoxide
was a secondary product; this effectively rules out the
possibility of the occurrence of a mechanism involving
methanol decomposition followed by the water–gas shift
the primary products, but that when methanol was fully
converted, CO was evident in the gas stream. This points to a
mechanism whereby the decomposition of formates to CO is
either inhibited in the presence of methanol, or more probably,
by the methoxy and/or formaldehyde species formed from the
adsorption of methanol on the catalyst. The mechanism of the
inhibition of the formation of CO by methanol is the subject of
further research.
reaction as proposed by several authors over their cata-
Notes and references
lysts.2
–5
.
1
M. V. Twigg, Catalysis Handbook, Wolfe Publishing, London, 2nd
When 100% methanol conversion was achieved at 300 °C
Fig. 1), the hydrogen and carbon dioxide molar compositions
edn., 1989, p. 283.
(
2
V. Pour, J. Barton and A. Benda, Coll. Czech. Chem. Commun., 1975,
40, 2923.
exceeded the equilibrium line ‘n’; at a value of W/F = 0.00389
2
3
g min cm , there was very little carbon monoxide production
and the hydrogen and carbon dioxide compositions were very
close to the equilibrium values calculated for the steam
reforming reaction when carbon monoxide was excluded from
the product stream (line ‘m’). At values of W/F > 0.00389 g
3 J. C. Amphlett, M. J. Evans, R. F. Mann and R. D. Weir, Can. J. Chem.
Eng., 1985, 63, 605.
4
5
6
J. Barton and V. Pour, Coll. Czech. Chem. Commun., 1980, 45, 3402.
E. Santacesaria and S. Carra, Appl. Catal., 1983, 5, 345.
J. C. Amphlett, R. F. Mann and B. A. Peppley, in Proc. 3rd Natural Gas
Conversion Symp., Sydney, Australia, 7–9 Apr, 1993; Stud. Surf. Sci.
Catal., 1994, 81, 409.
2
3
min cm , the proportions of hydrogen and carbon dioxide
decreased and that of carbon monoxide increased with increas-
ing contact time, approaching equilibrium ‘n’ values (a
sequence including CO formation) with high contact times. At
7
J. C. Amphlett, K. A. M. Creber, J. M. Davis, R. F. Mann, B. A. Peppley
and D. M. Stokes, in Proc. 9th World Hydrogen Energy Conf., Paris,
1992, p. 1541.
1
00% methanol conversion and a reaction temperature of 300
8 B. A. Peppley, J. C. Amphlett, L. M. Kearns and R. F. Mann, Appl.
Catal. A, 1999, 179, 21.
°
C, the carbon dioxide and hydrogen percentages were greater
9
B. A. Peppley, J. C. Amphlett, L. M. Kearns and R. F. Mann, Appl.
Catal. A, 1999, 179, 31.
than those predicted by equilibrium (n) whereas the carbon
monoxide percentages were less than equilibrium predictions
by a similar amount at the different values of W/F.
1
1
1
0 K. Takahashi, N. Takezawa and H. Kobayashi, Appl. Catal., 1982, 2,
83.
3
These results provide evidence of a consecutive reaction
scheme in which methanol and water react first to produce
carbon dioxide and hydrogen [eqn. (3)] and the carbon dioxide
and hydrogen then react via the reverse WGS reaction [eqn. (2)]
to produce CO. There was no evidence of the occurrence of the
methanol decomposition reaction [eqn. (1)].
In order to investigate the nature of the surface species
involved in the steam reforming reaction over the CuO/ZnO/
2 2 3
ZrO /Al O catalyst, in situ IR measurements were carried out.
1 C. J. Jiang, D. L. Trimm, M. S. Wainwright and N. W. Cant, Appl. Catal.
A, 1993, 93, 245.
2 C. J. Jiang, D. L. Trimm, M. S. Wainwright and N. W. Cant, Appl. Catal.
A, 1993, 97, 145.
13 J. P. Breen and J. R. H. Ross, Catal. Today, 1999, 51, 521.
14 W. B. White, S. M. Johnson and G. B. Dantzig, J. Chem. Phys., 1958,
2
8, 751.
1
1
5 C. Schild, A. Wokaun and A. Baiker, J. Mol. Catal., 1990, 63, 243.
6 D. Bianchi, T. Chafik, M. Khalfallah and S. J. Teichner, Appl. Catal. A,
1
995, 123, 89.
Fig. 2 [spectrum (1)] shows the surface species generated on the
surface of the catalyst at 240 °C after passing methanol over the
Communication 9/06393E
2248
Chem. Commun., 1999, 2247–2248