Applied Catalysis A: General 281 (2005) 19–24
www.elsevier.com/locate/apcata
Autothermal reforming of methane over Pt/ZrO2/Al2O3 catalysts
Mariana M.V.M. Souzaa,b, Martin Schmala,b,*
aNUCAT/PEQ/COPPE, Universidade Federal do Rio de Janeiro, C.P. 68502, 21945-970, Rio de Janeiro, BrazilbEscola de Quımica, Universidade Federal do Rio de Janeiro, C.P. 68542, 21940-900, Rio de Janeiro, Brazil
Received 9 July 2004; received in revised form 28 October 2004; accepted 5 November 2004
Available online 15 December 2004
Abstract
Autothermal reforming of methane, combining steam reforming and partial oxidation was carried out with Pt/Al2O3, Pt/ZrO2 and Pt/ZrO2/
Al2O3 catalysts, in the temperature range of 400–900 8C. The Pt/ZrO2/Al2O3 catalyst was found to be the most active and stable at 800 8C due
to the higher resistance to coke formation. The reaction occurs in two simultaneous stages: total combustion of methane and reforming of the
unconverted methane with steam and CO2, with the O2 conversion of 100% starting from 450 8C. The addition of O2 to the feed increases
methane conversion and the catalyst stability, decreasing the H2 and CO yields due to the enhancement of methane combustion. By
manipulating the O2/CH4 ratio of the feed it is possible to achieve the H2/CO ratio that is optimal to the GTL processes.
# 2004 Elsevier B.V. All rights reserved.
Keywords: Methane; Steam reforming; Partial oxidation; Autothermal reforming; Pt/ZrO2/Al2O3 catalysts
1. Introduction
There has been substantial interest in recent years in
alternative routes for conversion of natural gas (methane) to
synthesis gas, a mixture of CO and H2, which can be used to
produce chemical products with high added values, such as
hydrocarbons and oxygenated compounds. In GTL (gas-to-
liquid) plants, where natural gas is firstly converted to
synthesis gas, which is the feedstock for Fischer–Tropsch
synthesis of hydrocarbons, above 60–70% of the cost of the
overall process is associated with syngas production [1,2].
Therefore, reduction in syngas generation costs would have
a large and direct influence on the overall economics of these
downstream industrial processes.
Steam reforming is the established process for converting
natural gas into synthesis gas. SRM is a very energy-
intensive process because of the highly endothermic
property of the reaction and the H2/CO ratio obtained
(about 3) is only suitable for processes requiring a H2-rich
feed (such as ammonia syntheis and petroleum refining
process), but it is too high for fuel synthesis via Fischer–
* Corresponding author. Tel.: +5521 2562 8352; fax: +5521 2562 8300.
E-mail address: [email protected] (M. Schmal).
0926-860X/$ – see front matter # 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2004.11.007
Tropsch reaction [2,3]. Moreover, steam reformers are
inadequate for hydrogen generation in fuel cell electric
vehicles because of their high thermal inertia for frequent
start-up and shut-down operation condition [4,5].
Autothermal reforming (ATR), a combination of steam
reforming and partial oxidation reactions, is an advanta-
geous route for syngas production for both economical and
technical reasons. It has low-energy requirements due to the
opposite contribution of the exothermic methane oxidation
and endothermic steam reforming. The combination of these
reactions can improve the reactor temperature control and
reduce the formation of hot spots, avoiding catalyst
deactivation by sintering or carbon deposition. Moreover,
ATR allows the production of syngas with a wider range of
H2/CO ratio by manipulating the relative concentrations of
H2O and O2 in the feed [6–8]. All these advantages indicate
that ATR should be the technology of choice for large-scale
GTL plants [9]. In addition, a fuel processor based on
autothermal reforming of methane could provide a low cost
and compact system, with fast start-up and capability to
follow load variations, more adequate for fuel cell electric
vehicles [10,11].
We have previously reported that Pt/ZrO2/Al2O3 systems
are effective formulations for CO2 reforming of methane
M.M.V.M. Souza, M. Schmal / Applied Catalysis A: General 281 (2005) 19–2420
[12,13], partial oxidation [14,15] and steam reforming [15].
The aim of this work is to investigate the coupling between
the steam reforming of methane and partial oxidation, with
varying O2/CH4 ratio in the feedstream, over the Pt/ZrO2/
Al2O3 catalyst, comparing with the catalytic behavior of
Pt/A2O3 and Pt/ZrO2.
2. Experimental
2.1. Catalyst preparation
The g-Al2O3 support (Harshaw, Al 3996) was calcined in
flowing air at 550 8C for 2 h (BET surface area = 200 m2/g).
The ZrO2 support was prepared by calcination of zirconium
hydroxide (MEL Chemicals) using the same conditions
(BET surface area = 62 m2/g). The 10% ZrO2/Al2O3 was
obtained by impregnation of alumina with a nitric acid
solution of zirconium hydroxide (BET surface area =
180 m2/g), as described elsewhere [13]. The supports were
impregnated with an aqueous solution of chloroplatinic acid
(H2PtCl6, Aldrich) by incipient wetness technique. The
catalysts were subsequently dried overnight at 120 8C and
calcined in air at 550 8C for 2 h. The platinum content was
around 1 wt.%. The prepared catalysts will be referred to
as PtAl for Pt/Al2O3, PtZr for Pt/ZrO2 and Pt10ZrAl for
Pt/ZrO2/Al2O3.
2.2. Catalyst testing
The reaction was carried out in a U-tube fixed-bed flow-
type quartz reactor, with length of 15 cm and 6 mm of
external diameter (the bed height is 4–5 mm), loaded with
20 mg of catalyst, under atmospheric pressure. The catalyst
bed is supported by quartz wool. The total feed flow rate was
held constant at 200 cm3/min (WHSV = 160 h�1), with
flowing He. Steam was added to the system using a saturator
at controlled temperature. The gas compositions are listed in
Table 1. The activity tests were performed at different
temperatures, ranging from 400 to 900 8C in steps of 50 8Cthat were kept for 30 min at each temperature. The loss in
catalyst activity at 800 8C was monitored up to 60 h on
stream. Blank experiments were also carried out but did not
show conversions above 5% even at high temperatures. The
reaction products were analyzed by on-line gas chromato-
graph (CHROMPACK CP9001), equipped with a Hayesep D
column and a thermal conductivity detector.
Table 1
Flow rates for activity and deactivation tests (cm3/min)
O2/CH4 H2O/CH4 CH4 O2 H2O He Total
0.25 0.5 40 10 20 140 200
0.5 0.5 40 20 20 120 200
1.0 0.5 20 20 10 150 200
2.0 0.5 10 20 5 165 200
2.3. Carbon deposition measurements
The amount of coke formed over the catalysts after
deactivation tests at 800 8C was examined by thermogravi-
metric analysis (TGA), using a RIGAKU thermoanalyzer
(model TAS 100). The samples were pretreated at 150 8Cunder flowing nitrogen and then heated at a rate of 10 8C/min
to 800 8C in a flow of 15% O2/N2 (50 cm3/min).
3. Results and discussion
3.1. Effect of temperature
The comparison of catalyst activities for autothermal
reforming of methane is displayed in Fig. 1A and B, in terms
of CH4 conversion and H2/CO product ratio, respectively.
The O2 conversion is 100% starting from 450 8C for PtAl
and Pt10ZrAl catalysts and 400 8C for PtZr. The PtZr
showed the highest activity at 400 8C, with 20% of CH4
Fig. 1. Catalytic activities in terms of CH4 conversion (A) and H2/CO
product ratio (B) of Pt catalysts for autothermal reforming of methane as a
function of temperature.
M.M.V.M. Souza, M. Schmal / Applied Catalysis A: General 281 (2005) 19–24 21
Fig. 2. Composition profiles for autothermal reforming of methane as a
function of temperature over the Pt10ZrAl catalyst.
conversion, maintaining this conversion up to 550 8C. The
Pt10ZrAl was the most active catalyst over the mid-
temperature range (450–600 8C) while at higher tempera-
tures PtZr exhibited slightly better activity than PtAl and
Pt10ZrAl.
The activity was almost constant in temperature range
between 450 and 650 8C, which can be related to the
combustion of methane to CO2 and H2O, a very exothermic
reaction:
CH4 þ 2O2 ,CO2 þ 2H2O;
DH298 K ¼ �802 kJ=mol(1)
The observed high H2/CO ratio (>6.0) at low tempera-
tures, especially for PtZr catalyst, suggests that water–gas
shift (WGS) reaction occurs to a great extent with reforming
of methane, as already reported in [7,16]. The WGS
equilibrium was reached, according to thermodynamic
calculations. At the same time, the decrease in H2/CO ratio
with increasing reaction temperature is consistent with the
fact that WGS reaction is thermodynamically unfavorable at
higher temperatures:
CO þ H2O,CO2 þ H2; DH298 K ¼ �41 kJ=mol (2)
At high temperatures, the H2/CO ratio lower than 2.0 (this
ratio is 1.4 for PtZr and 1.8 for Pt10ZrAl at 800 8C) indicates
that the reverse of WGS reaction takes place simultaneously
with reforming of methane:
CH4 þ H2O,CO þ 3H2; DH298 K ¼ 206 kJ=mol (3)
CH4 þ CO2 ,CO þ H2; DH298 K ¼ 247 kJ=mol (4)
Fig. 3. Deactivation test in terms of CH4 conversion as a function of time on
stream at 800 8C.
Several authors have proposed a two-step mechanism for
the partial oxidation of methane. In the first step combustion
of methane takes place, producing CO2 and H2O; in the
second step, synthesis gas is produced via CO2 and steam
reforming reactions of unreacted methane [17–19].
This indirect mechanism is also clearly evidenced for
autothermal reforming, as can be seen by the composition
profiles shown in Fig. 2 for the Pt10ZrAl catalyst. The
O2 conversion is 100% starting from 450 8C. At this
temperature, there is the onset of CO2 production and the
water conversion is lower than methane conversion due to
the combustion of methane, which produces CO2 and H2O.
The low production of H2 and CO up to 600 8C is also related
to the occurrence of total combustion. The H2 and CO
contents increase sharply with increasing temperature from
650 8C, which can be explained in terms of the contribution
of methane reforming with steam and CO2, in agreement
with thermodynamic calculations.
The reaction mechanism, according to the literature, was
related to the change in the nickel valence state during the
partial oxidation of methane. The rapid transformation of the
reaction from complete oxidation of methane to partial
oxidation was associated to the reduction of NiO to Ni0 [17].
However, the surface modification for the platinum catalysts
can be disregarded, because after pretreatment with H2 flow
up to 700 8C the catalytic activity did not change (not
shown).
3.2. Effect of time on stream
Comparison of catalyst stability at 800 8C is shown in
Fig. 3. The equilibrium conversion at this temperature is
95%. The catalysts exhibited different initial activities and
deactivation rates. The PtZr showed the highest initial
activity, but deactivated at a mean rate of 0.86%/h during
28 h on stream. PtAl was the least active catalyst and the
mean deactivation rate was 0.42%/h. The catalyst with
10 wt.% of ZrO2 exhibited the best performance, with CH4
conversion decreasing from 82 to 68% in the first 10 h,
maintaining this conversion for over 60 h (mean deactivation
rate of 0.24%/h).
The deactivation is related to the deposition of inactive
carbon over the active surface and the amount of coke on
these catalysts was quantified by thermogravimetric
M.M.V.M. Souza, M. Schmal / Applied Catalysis A: General 281 (2005) 19–2422
Fig. 4. TGA of Pt catalysts after deactivation test at 800 8C. Conditions:
15% O2/N2, 10 8C/min, and feed flow rate = 50 cm3/min.
analysis, carried out in an oxygen-containing atmosphere
(Fig. 4). The stability of the Pt10ZrAl catalyst is really
associated with the observation of little coke formation
during the reaction (about 0.2 mg coke/gcat h). On the other
hand, TGA experiment showed a weight loss of about 17%
on treating the PtAl and PtZr catalysts in oxygen, indicating
a significant amount of carbon deposition during 30 h on
stream (approximately 5.7 mg coke/gcat h).
The deposition of inactive carbon during methane
reforming can be originated from either methane decom-
position (reaction (5)) or CO disproportionation (Boudouard
reaction (6)), which are thermodynamically favorable below
900 8C [3,20]:
CH4 ,C þ 2H2; DH298 K ¼ 75 kJ=mol (5)
2CO,C þ CO2; DH298 K ¼ �172 kJ=mol (6)
Thermodynamic calculations showed that the extent of
carbon deposition during reforming decreases at higher reac-
tion temperatures, in agreement with several experimental
Fig. 5. CH4 conversion as a function of time on stream, at 800 8C, varying
the O2/CH4 feed ratio, for Pt10ZrAl catalyst.
observation [21,22]. These results suggest that CO dispro-
portionation is the main contributor to carbon deposition
because it is exothermic and the equilibrium constant
decreases with increasing temperature.
The higher stability of Pt10ZrAl has already been verified
for steam reforming and partial oxidation separately [14,15]
as well as for CO2 reforming [13]. The higher stability is
closely related to its coking resistivity, which has been
attributed to Pt–Zrn+ interactions at metal-support interface.
The number of interfacial sites is higher on the Pt10ZrAl
catalyst as compared to PtZr itself, because the particle sizes
of ZrO2 on Al2O3 are very small compared to the bulk ZrO2
particles. Therefore, the Pt particles are better dispersed on
highly dispersed ZrO2 particles over AlsO3 [13,14]. The
interfacial sites on Pt-support promote CO2 dissociation,
improving the catalyst stability by shifting the Boudouard
reaction. Our previous FT-IR analysis [13] showed that the
Pt–Zrn+ interface is active for CO and CO2 adsorption, with a
decrease in the Pt–CO bond strength, inhibiting C–O bond
breaking and consequently producing less carbon formation
Fig. 6. Catalytic selectivities as a function of time on stream, at 800 8C, in
terms of H2/CO (A) and CO2/CO (B) product ratios, for Pt10ZrAl catalyst.
M.M.V.M. Souza, M. Schmal / Applied Catalysis A: General 281 (2005) 19–24 23
on the catalyst surface. Moreover, zirconia is a well-know
oxygen supplier, and its oxygen mobility is fast, which helps
to keep the metal surface free of carbon.
3.3. Effect of O2/CH4 feed ratio
Fig. 5 presents the CH4 conversion at 800 8C as a function
of time on stream, varying the O2/CH4 ratio and at a constant
H2O/CH4 ratio of 0.5, for the Pt10ZrAl catalyst. The O2/CH4
ratio of 0.25 corresponds to the stoichiometric amount of
autothermal reforming. The increase in the O2/CH4 feed
ratio not only increases CH4 conversion but also decreases
the catalyst deactivation. When the O2/CH4 ratio is 0.5 there
is a slight activation of the catalyst with time on stream and
when this ratio is increased to 1.0 the CH4 conversion is
100% throughout the duration of the experiment (60 h).
The increase in the catalytic stability with the addition of
O2 to the feed has already been observed in the literature, for
Pt/ZrO2 catalysts, during either steam reforming [16] or CO2
reforming of methane [23]. The excess of O2 probably favors
the reoxidation of carbonaceous residues formed over the
catalytic surface, avoiding the catalyst deactivation. The
increase in CH4 conversion with O2 addition can be related
to two distinct effects: the fast combustion of part of
methane with all O2 in the feed and the increase in the
Fig. 7. Composition profiles as a function of time on stream at 800 8C for Pt10Z
(D) 2.0.
reforming reaction rate due to the increase in temperature
caused by the exothermicity of combustion reaction. This
second effect is less significant in our case because all the
streams are greatly diluted by He (Table 1).
The comparison of catalyst selectivities at 800 8Cwith different O2/CH4 feed ratios is displayed in Fig. 6A
and B, in terms of H2/CO and CO2/CO product ratios,
respectively. The H2/CO ratio is minimum (about 1.5) when
O2/CH4 = 0.5; this ratio reaches the value of 2.0 when O2/
CH4 = 1.0. The CO2/CO ratio is about 0.2 for O2/CH4 � 0.5
and 0.9 for O2/CH4 = 1.0.
The role of oxygen addition on the catalytic selectivity is
fairly complicated due to the different contributions of
reactions (1)–(4). When O2/CH4 � 0.5 the CO2/CO ratios
have about the same value, which indicates that the total
combustion proceeds to the same extent, but the reforming
reaction is enhanced by decreasing O2/CH4 ratio due to the
higher relative availability of methane under these condi-
tions. When the ratio O2/CH4 is 1.0, the H2/CO ratio
obtained (about 2–2.1) has the optimum value for GTL
processes; however, the H2 and CO yields are lower, with
relative higher formation of CO2 and H2O by methane
combustion. The composition profiles of H2, CO, CO2 and
H2O as a function of O2/CH4 feed ratio are represented in
Fig. 7. It can be clearly observed that as O2/CH4 ratio
rAl catalyst in different O2/CH4 feed ratios: (A) 0.25, (B) 0.5, (C) 1.0, and
M.M.V.M. Souza, M. Schmal / Applied Catalysis A: General 281 (2005) 19–2424
increases, the relative formation of CO2 and H2O increases
until the limit case (O2/CH4 = 2) when the production of H2
and CO is completely inhibited because of methane total
consumption by combustion.
4. Conclusions
The support plays a decisive role on the catalytic behavior
during autothermal reforming of methane, in which steam
reforming and partial oxidation take place simultaneously.
The Pt/10% ZrO2/Al2O3 catalyst showed the highest activity
and stability during 70 h on stream at 800 8C, when
compared to Pt/Al2O3 and Pt/ZrO2. The composition
profiles when the feed is in stoichiometric ratio showed
that the reaction proceeds via a two-step mechanism with the
total combustion of methane followed by reforming of
unreacted methane with CO2 and H2O; the heat released by
combustion favors the reforming reaction. The higher
stability of the Pt/10% ZrO2/Al2O3 catalyst is closely related
to its coking resistance due to Pt–Zrn+ interactions at metal-
support interface.
The increase in O2/CH4 feed ratio not only increases the
CH4 conversion but also improves the catalyst stability. The
excess of O2 avoids the deposition of carbonaceous residues
over the catalytic surface; however, the H2 and CO yields
also decrease due to the enhancing of total combustion.
The H2/CO product ratio can be manipulated by varying the
O2/CH4 feed ratio and it is possible to achieve the optimum
ratio for GTL processes (H2/CO = 2 when O2/CH4 = 1.0).
References
[1] J.R. Rostrup-Nielsen, Catal. Today 21 (1994) 257.
[2] M.A. Pena, J.P. Gomez, J.L.G. Fierro, Appl. Catal. A 144 (1996) 7.
[3] S.C. Tsang, J.B. Claridge, M.L.H. Green, Catal. Today 23 (1995) 3.
[4] L. Pino, V. Recupero, S. Beninati, A.K. Shukla, M.S. Hegde, P. Bera,
Appl. Catal. A 225 (2002) 63.
[5] L. Pino, A. Vita, M. Cordaro, V. Recupero, M.S. Hegde, Appl. Catal. A
243 (2003) 135.
[6] S. Liu, G. Xiong, H. Dong, W. Yang, Appl. Catal. A 202 (2000) 141.
[7] Z.-W. Liu, K.-W. Jun, H.-S. Roh, S.-E. Park, J. Power Sources 111
(2002) 283.
[8] S. Ayabe, H. Omoto, T. Utaka, R. Kikuchi, K. Sasaki, Y. Teraoka, K.
Eguchi, Appl. Catal. A 241 (2003) 261.
[9] D.J. Wilhelm, D.R. Simbeck, A.D. Karp, R.L. Dickenson, Fuel Proc.
Technol. 71 (2001) 139.
[10] S.H. Chan, H.M. Wang, Int. J. Hydrogen Energy 25 (2000) 441.
[11] S. Ahmed, M. Krumpelt, Int. J. Hydrogen Energy 26 (2001) 291.
[12] M. Schmal, M.M.V.M. Souza, D.A.G. Aranda, C.A.C. Perez, Stud.
Surf. Sci. Catal. 132 (2001) 695.
[13] M.M.V.M. Souza, D.A.G. Aranda, M. Schmal, J. Catal. 204 (2001)
498.
[14] M.M.V.M. Souza, M. Schmal, Appl. Catal. A 255 (2003) 83.
[15] M.M.V.M. Souza, M. Schmal, Stud. Surf. Sci. Catal. 147 (2004) 133.
[16] M.E.S. Hegarty, A.M. O’Connor, J.R.H. Ross, Catal. Today 42 (1998)
225.
[17] D. Dissanayaki, M.P. Rosynek, J.H. Lunsford, J. Phys. Chem. 97
(1993) 3644.
[18] E.P.J. Mallens, J.H.B. Hoebink, G.B. Martin, Catal. Lett. 33 (1995)
291.
[19] F. van Looij, E.R. Stobbe, J.W. Geus, Catal. Lett. 50 (1998) 59.
[20] A.M. Gadalla, B. Bower, Chem. Eng. Sci. 43 (1988) 3049.
[21] M.C.J. Bradford, M.A. Vannice, Catal. Rev. Sci. Eng. 41 (1) (1999) 1.
[22] J.T. Richardson, S.A. Paripatyadar, Appl. Catal. 61 (1990) 293.
[23] A.M. O’Connor, J.R.H. Ross, Catal. Today 46 (1998) 203.