Kinetic of methane steam reforming reaction over nickel- and rhodium-based catalysts

Applied Catalysis A: General 387 (2010) 147–154

Contents lists available at ScienceDirect

Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata

Kinetic of methane steam reforming reaction over nickel- and rhodium-based catalysts M. Zeppieri a , P.L. Villa b , N. Verdone a , M. Scarsella a , P. De Filippis a,∗ a b

Dipartimento di Ingegneria Chimica Materiali Ambiente, Università “La Sapienza” Via Eudossiana 18,00184 Roma, Italy Dipartimento di Chimica, Ingegneria Chimica e Materiali, Università dell’Aquila Monteluco di Roio, 67040 L’Aquila, Italy

a r t i c l e

i n f o

Article history: Received 7 June 2010 Received in revised form 30 July 2010 Accepted 10 August 2010 Available online 17 August 2010 Keywords: Methane steam reforming Reaction kinetics Rhodium-perovskite

a b s t r a c t Kinetics of methane steam reforming over a commercial nickel-based catalyst and over an innovative rhodium-perovskite catalyst of formula BaRhx Zr(1−x) O3 was studied at atmospheric pressure and in the temperature range 723–1023 K. Extensive experimental runs were firstly carried out in a micro-reactor, with the catalysts in powder form, to evaluate their performances in steam methane reforming process. The behaviour of rhodium-perovskite catalyst was analyzed as function of steam-to-methane ratio and temperature, comparing the obtained performances of this catalyst with that showed by the commercial nickel-based one. Rhodium-perovskite catalyst shows higher activity for steam methane reforming reaching methane conversions higher than those obtained with the nickel-based catalyst and closer to the thermodynamic equilibrium value. Based on a detailed review of the kinetic models proposed for steam reforming, the experimental data were fitted using a simple kinetic model where the conversion of methane is proportional only to the methane partial pressure. Good agreement was obtained between the experimental data and the kinetic model prediction. © 2010 Elsevier B.V. All rights reserved.

1. Introduction Currently almost 96% of the world’s hydrogen demand is satisfied by fossil fuels, with about half of it being generated by catalytic steam reforming (SR) of natural gas [1]. Supported nickel catalysts are actually the most widely used in the steam reforming processes but also supported noble metal (as rhodium, palladium, platinum and ruthenium) catalysts have been extensively investigated for SR [2]. In comparison with noble metals-based catalysts, supported nickel catalysts are less active and usually more prone to deactivation by carbon formation or oxidation [3]. However, owing to the nickel low cost (100–150 times less expensive than noble metals) nickel-based catalysts are now the preferred option. The selection of the catalyst support as well the dispersion of the active metal also play an important role in the reactions steps involved in the catalytic process. Alternative catalysts have been actively explored and more recently, the research on catalysts has paid much attention to systems with perovskitic structure. Perovskites are mixed metal oxides which may fit the catalytically active element into the crystal lattice. This structure can improve thermal stability, reducing sintering and avoiding carbon formation. Recent works [4,5] concern the synthesis of some

∗ Corresponding author. E-mail address: paolo.defi[email protected] (P. De Filippis). 0926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2010.08.017

innovative systems with a perovskite structure based on barium zirconate, BaZrO3 . Supported and unsupported catalysts have been prepared in which the active phase consists of an oxide with a perovskitic structure of formula BaMex Zr(1−x) O3 in which the zirconium has been partially replaced by a noble metal (Me = Rh, Pd, Pt) [4,6,7]. These catalysts with perovskitic structure guarantee a good compromise between stability and activity and have a relatively low cost, thereby constituting a valid alternative to supported noble metals with particular reference to the reactions of partial or total oxidation of hydrocarbons and of steam reforming. Scope of this work is to analyze and evaluate, by means of a kinetic model, the performance of a rhodium-perovskite catalyst compared with a commercial nickel-based catalyst in a steam methane reforming (SMR) process carried out in a lab-scale microreactor.

2. Experimental 2.1. Materials and methods SMR experimental runs were performed over two different catalysts: the commercial nickel-based catalyst ICI Katalco 559 25-4MQ, supplied by Johnson Matthey, and an innovative rhodiumperovskite catalyst. The rhodium-perovskite catalyst was prepared using a variant of the citrate route [6,7], obtaining an oxide with a perovskitic structure in which zirconium has been partially

148

M. Zeppieri et al. / Applied Catalysis A: General 387 (2010) 147–154

Fig. 1. Schematic layout of experimental SMR system.

replaced by rhodium, of formula BaRhx Zr(1−x) O3 , and a Rh concentration of around 5% by weight. BET analysis of both catalysts was performed using a Micrometrics ASAP 2000 V 2.05, in order to determine the surface area. X-ray diffraction and SEM analysis and microanalysis were used to characterize the phase composition and to evaluate the surface morphology and composition of rhodium-perovskite. SEM was equipped also with EDS. The absence of possible unwanted phases was checked by XRD with profile fitting [4,5]. For the SMR experimental runs pure methane (>99.995%), pure nitrogen (>99.9995%), pure hydrogen (>99.995%), all supplied by SIAD, and demineralised water were used.

control the preheating-vaporizer and the heating of reaction zone. The process was controlled by a control system with a computer interface for monitoring and setting operation parameters. After leaving the reformer, the reformate gas passed through a condenser, a gas–liquid separator and a gas dryer to remove all the water before going to a gas mass flow rate meter and to the gas analyzer where its composition was determined. The gas analysis was performed by means of on-line analyzers (ABB AO2020 Uras 26 infrared analyzer; Ultramat 23 infrared gas analyzer; Siemens Calomat 6E). Spot gas analyses were done in a gas chromatography system (DANI 3800 gas chromatograph, carrier gas Argon).

2.3. Experimental procedure 2.2. Micro-reactor equipment description The SMR experimental runs were conducted in the equipment schematically reported in Fig. 1. The system consists of a feed section where the gas from cylinders and deionised water from a pressurized system, both regulated by means of mass flow controllers, were mixed in a mixing valve and sent to an evaporator. The reaction section is constituted by a quartz tube of 8 mm inner diameter and total length of 500 mm, equipped with a quartz septum at 290 mm height from the bottom. The reactor is divided into two zones: a preheating zone of 90–105 mm length, and a reaction zone of 5–20 mm length. Preheating and reaction zones were heated by a cable heater. In order to keep isothermal conditions both pre-heater and reactor were insulated by ceramic fibres. Two thermocouples, sliding in a central tube, were used to measure catalyst temperature (T1 and T2 ) while other two thermocouples (T3 and T4 ) were used to

The perovskitic catalyst was used as produced, while the commercial nickel-based catalyst, provided in form of 13 mm × 10.5 mm, 2,7-hole quadrupoles, was used after crushing and sieving. The sieved particles had a diameter lower than 200 ␮m to avoid intra-particle diffusion effects [8] and to allow a better comparison with the perovskitic catalysts obtained in powder form. Catalytic rates were measured by placing in the quartz tube 5–30 mg of catalyst samples diluted with 500–1500 mg of acid washed silicon dioxide (grain size 0.1–0.3 mm) to minimize temperature gradients. In the case of nickel-based catalyst it was necessary to have a preliminary reduction step once loaded in the reforming reactor. The reduction was carried out with the following procedure: the catalyst was heated to 873 K in nitrogen atmosphere and maintained at this temperature until steady state was reached, then the inlet gas was switched to hydrogen. The catalyst was

M. Zeppieri et al. / Applied Catalysis A: General 387 (2010) 147–154

149

Table 1 Experimental conditions for SMR over rhodium-perovskite catalyst. Sample

Contact time (ms)

Rh (A) Rh (B) Rh (C) Rh (D)

50.5 12.8 7.4 3.2

a

Contact timea (gcat min m−3 ) 168 83 49 19

W /FCH4 (gcat min mol−1 )

Linear velocity (m/s)

H2 O/CH4 molar ratio

Temperature range (K)

14.40 6.48 4.24 1.62

0.1981 0.2020 0.3438 0.1955

3 3 3 3

782–920 782–920 782–920 782–920

Based on catalyst weight.

then maintained at this temperature in a hydrogen atmosphere for 1 h. Preliminary tests showed that this procedure was adequate to activate the catalyst. Afterwards, the catalyst was maintained in nitrogen flow. The perovskitic catalyst did not require preliminary reduction and the only needed pre-operation was heating the catalyst in nitrogen atmosphere at the desired temperature. At the beginning of each test, the temperature was set to the desired reforming conditions and, once the operation conditions were reached, steam was initially supplied with nitrogen and then nitrogen was switched to methane. Experimental kinetic data were collected in a range of temperatures between 723 and 1023 K at atmospheric pressure and in a wide range of contact times, expressed in terms of grams of catalyst loaded and methane molar flow ratio (W /FCH4 in ) or in terms of physical volume of catalyst and volumetric flow rate of feed ratio (Vcat /QCH4 in ). In order to approach plug flow conditions and to minimize back mixing, following Froment and Bischoff [9] the ratio between catalyst bed length and catalyst particle size (L/dp ) was 400–1000, the ratio between the inner diameter of the reactor and the particle size (Drea /dp ) was 40–55, and Reynolds number was greater than 30. Preliminary tests on both catalysts, conducted to check the transient time, the deactivation and the degree of carbon deposition, showed that no deactivation occurred during the reforming experiments. On the basis of these evidences and considering that the experiments were conducted quite fast, the effect of catalyst deactivation can be neglected. Based on the catalysts stability experiments, complete runs were conducted within 8 h, using fresh catalyst for each run. The occurrence of deactivation was further checked at the end of each experiment by restoring the initial operative conditions and verifying the CH4 conversion. The amount of catalyst used for each test was adjusted to maintain CH4 conversions far from those predicted by the thermodynamic equilibrium. Experiments were performed with a steam-to-methane ratio equal to 3. Table 1 summarizes the experimental conditions using rhodium-perovskite catalyst, while Table 2 summarizes the experimental conditions using commercial nickel-based catalyst. Additional tests were performed on the rhodium-perovskite catalyst in order to evaluate its behaviour for different H2 O/CH4 ratios. Experimental data were compared with the thermodynamic equilibrium conditions calculated by minimizing Gibbs free energy using the NASA computer program CEA (Chemical Equilibrium with Applications) [10].

3. Reaction mechanisms and kinetic of steam reforming One of the earliest proposed expressions for SR reactions, based on a detailed mechanism, was provided by Xu and Froment [11] which extensively studied the kinetic of steam methane reforming (SMR) over a Ni/MgAl2 O4 catalyst. The mechanism proposed by Xu and Froment indicates that the reactions of carbon intermediates with adsorbed oxygen are rate determining, suggesting that the concentration of the oxygen may determine to a large extent the reaction kinetics. This model is very complex and illustrates the many possible steps involved in the SMR. Many other research groups followed the work of Xu and Froment and proposed their own mechanism. Rostrup-Nielsen [3] proposed a four steps mechanism and argued that the model by Xu and Froment is not consistent with the current understanding of methane dissociation [12], which had been shown not to proceed via an adsorbed precursor state. Another complex mechanism for SMR over a nickel surface was proposed by Compton [13]. One disadvantage of some of the proposed models is their complexity that makes difficult to determine reliable reactions rate parameters. A different approach has been proposed by Wei and Iglesia, who investigated the mechanisms for the reactions of CH4 with CO2 and H2 O on different metal clusters [14–17]. They found that reaction rates are proportional to the CH4 partial pressure and independent on co-reactants partial pressures, leading to the conclusion that only C–H bond activation steps are kinetically relevant. Their data indicate that the activation of co-reactant and its kinetic coupling with CH4 activation via scavenging of chemisorbed carbon intermediates are the fast steps and therefore metal surfaces are essentially uncovered by reactive intermediates. It was also shown that C–H bond activation elementary steps are irreversible and that re-combinative desorption steps of H atoms with OH groups to form H2 or H2 O are quasi-equilibrated. The quasi-equilibrated nature of these and other steps confirms that the water gas shift reaction is also at equilibrium. Wei and Iglesia mechanism indicates that the reactivity of the metal towards the C–H bond breaking governs the overall reaction kinetics. This emphasizes the importance of the catalytic activity of the metal. In general, a well-balanced interplay between the metal and support will undoubtedly lead to the best catalytic performance. Several scientists agreed with Wei and Iglesia results, considering their mechanism the more realistic. It is generally accepted among experimentalists and theoreticians that the activation of

Table 2 Experimental conditions for SMR over commercial nickel-based catalyst. Sample

Contact time (ms)

Contact timea (gcat min m−3 )

W /FCH4 (gcat min mol−1 )

Linear velocity (m/s)

H2 O/CH4 molar ratio

Temperature range (K)

Ni (A) Ni (B) Ni (C) Ni (D) Ni (E)

32.7 25.1 18.6 7.6 5.9

108.0 83.0 61.0 51.0 35.4

9.74 7.43 5.59 4.68 3.19

0.2521 0.3984 0.5363 0.3059 0.5185

3 3 3 3 3

783–958 783–958 783–958 783–958 783–958

a

Based on catalyst weight.

150

M. Zeppieri et al. / Applied Catalysis A: General 387 (2010) 147–154

methane is the rate-determining step. This is also supported by the fact that Xu and Froment found for steam an unlikely negative heat of adsorption. Furthermore the experiments were performed in a range of temperatures below the traditional temperatures of SMR [3]. Wei and Iglesia have established that dissociation of methane is the rate-determining step for SMR, so this reaction is limited kinetically by the initial activation of C–H bond in CH4 . Indeed, the relevant forward rate constants for these three reactions were observed to be the same with respect to a given metal, the metal dispersion, and the reaction conditions, being unaffected by the H2 O concentration. 3.1. Kinetic model Under steady state conditions only one continuity equation is required for completely describing methane conversion: the mass balance for CH4 . During steam methane reforming, the total gas volume increases and at the steady state, considering a very small height of bed catalyst, the mass balance can be written as: FCH4 in − FCH4 out = (−rn ) × Q

(1)

where Q is calculated with the following relation: Q =

catalyst bed volume × active phase weight catalyst bed weight

(2)

allowing to take into account the dilution of the active phase with respect to the overall quantity loaded. Using Eq. (1) it is possible to calculate the apparent reaction rate of methane steam reforming expressed in units of [mol/(g s)] or [mol/(m3 s)]. In absence of detectable deactivation and of transport disguises forward reforming rates can be rigorously obtained using the equilibrium approach, by defining the ˇ and  parameters from steam reforming thermodynamic reaction data and the measured partial pressures [Pj ] of reactants and products: ˇ =1− =

[PCO ][PH2 ]3 [PCH4 ][PH2 O ]

(3) ×

1 Keq

(4)

The values of ˇ ranges from 0.01 to 0.20 in this study. The measured net reaction rates, rn , are used to obtain the forward rates, rf , using: rf =

rn ˇ

(5)

Table 3 Main physical proprieties of catalysts. Properties 2

Surface area (BET, m /g) ˚ Average pore diameter (A) Pore volume (cm3 /g) Particle size ␮m

Katalco 25-4MQ

BaRhx Zr(1−x) O3

14.30 – 0.246 <200

7.33 122 0.0225 <150

For the water gas shift reaction (WGSR), the reaction rate could not be determined because in the operative conditions applied the WGSR is at equilibrium. For the purpose of this work, using the simple and unified model proposed by Wei and Iglesia it was possible to calculate the main parameters of reaction rate, the pre-exponential factor k0 and the activation energy Ea . 4. Experimental results 4.1. Catalysts characterization The results of the physical characterization of the catalysts are listed in Table 3. As shown in table, rhodium-perovskite catalyst has a BET specific surface lower than that of commercial nickelbased catalyst. The obtained value of surface area for the perovskite catalyst of about 7 m2 /g is consistent with the values reported in the literature for barium zirconate (10 m2 /g) [4] and for perovskite obtained with the citrate method (5–20 m2 /g) [18]. X-ray diffraction analysis (XRD) revealed that the perovskite structure was the only phase existing in the material. XRD of the rhodium-perovskite (Fig. 2) shows slight shift of peaks on the right related to the red line at 2 equal to 30, suggesting that a contraction of crystalline lattice occurred with respect to the original crystalline lattice of BaZrO3 , a clear evidence of the rhodium incorporation in the original structure. In the XRD the only detected phases were BaZr(1−x) Rhx O3 and BaCO3 , present in a very low amount. Fig. 3 shows the results of microanalysis. The most significant peaks are due to the perovskite and to the incorporated metal. SEM analysis was performed using different scanning resolution, as shown in Fig. 4, highlighting a homogeneous structure for the rhodium-perovskite catalyst. 4.2. Rhodium-perovskite catalyst The trend of CH4 conversion as function of steam-to-methane molar ratio for a contact time of 80 ms and a catalyst temperature of 1073 K is shown in Fig. 5a, where the theoretical equilibrium values

Eq. (5) accurately describes all the effects on measured reaction rates that occur in the reactor and are related to residence time and conversion. The determination of the dependencies of the forward reaction rate is based on the work of Wei and Iglesia [17], where the C–H bond activation is the only kinetically relevant step in SMR. As a consequence, forward rates are not influenced by CO or H2 concentrations, whether they are varied by external addition or by changes in residence times and CH4 conversion. CH4 steam reforming rates become simply first order in CH4 and zero order in H2 O: rf = k × PCH4

(6)

Additional effects of product concentrations on  value are taken into account using Eqs. (3)–(5). Steam reforming reaction rate has temperature dependence in accordance with the Arrhenius law. If rate constants can be calculated without ambiguity from rate data over a wide range of temperature, the plot ln(k) versus 1/T provides the true activation energy.

Fig. 2. XRD of rhodium-perovskite.

M. Zeppieri et al. / Applied Catalysis A: General 387 (2010) 147–154

151

Fig. 3. Microanalysis of rhodium-perovskite.

Fig. 4. SEM analysis of rhodium-perovskite.

at the same conditions are also reported. Rhodium catalyst shows a typical behaviour in steam reforming reactions, with an improved conversion increasing the H2 O/CH4 molar ratio. The CO selectivity (Fig. 5b) decreases at increasing steam-to-methane molar ratio. This is likely due to the water gas shift reaction at higher steam concentration. At steam-to-methane ratio equal to 3, the WGSR is at equilibrium conditions.

The methane conversion versus residence time of methane, expressed as mass of catalyst per mole of inlet CH4 flow rate, in the different experimental conditions is reported in Fig. 6. In the investigated temperature range, the CH4 conversion increases at increasing residence time and for a residence time greater than about 10 gcat min / molinlet CH4 1 the CH4 conversion no longer increases but reaches a quite stable value. Rhodium-perovskite

Fig. 5. Trend of CH4 conversion (a) and CO selectivity (b) versus H2 O/CH4 ratio for rhodium-perovskite at 1073 K and comparison with the equilibrium values.

152

M. Zeppieri et al. / Applied Catalysis A: General 387 (2010) 147–154

Fig. 8. Molar fractions obtained with commercial Ni-based catalyst in different conditions of temperature and contact time. Fig. 6. Methane conversion versus residence time at different catalyst temperatures and H2 O/CH4 = 3 for rhodium-perovskite catalyst.

catalyst showed high activity in the steam reforming reaction, reaching stable value of CH4 conversion with low residence times. In Fig. 7 the measured percentage molar fractions of the gaseous species on a gas dry basis are reported. For the different experimental conditions reported in Table 1, increasing contact time increases the methane conversion as well as the hydrogen molar fraction, as expected. For temperature greater than 819 K, CO2 molar fraction resulted higher than the theoretical equilibrium value in the same conditions. These results are due to the relatively high rate of WGSR in a wide range of temperatures. The micro-reactor cooling system was probably not as quick as WGSR rate, and as a consequence further CO2 was produced after the steam reforming section. This can also explain the CO selectivity trend reported in Fig. 5b, where for H2 O/CH4 values greater than 2.5 CO selectivity is almost constant and equal to 20%, meaning that a further CO conversion occurs during the cooling phase and the equilibrium conditions for WGSR are reached. 4.3. Commercial catalyst Katalco 25-4MQ The methane conversions obtained in the different experimental conditions reported in Table 2 are shown in Fig. 8. The conversions are compared to the values in the same conditions at the thermodynamic equilibrium. In all the cases the measured

values are far from the equilibrium, with higher conversions corresponding to higher contact times.

4.4. Comparison of catalysts performances The experimental SMR tests for the two different active phases allow to select the better one for steam reforming processes. As shown in Fig. 9, where a comparison between the performances of the commercial Katalco 25-4MQ and the rhodium-perovskite catalysts is reported, the rhodium-perovskite phase shows to be the most active and allows to obtain higher conversions also at lower temperatures and with lower contact times (7.4 ms for Rh-perovskite and 7.6 ms for Ni-based catalyst). The methane conversions using rhodium-perovskite are closer to the theoretical thermodynamic values. These results suggest that higher conversion could be obtained with a lower quantity of catalyst. Experimental tests clearly demonstrate that methane conversion improve from 23% to 68% in similar experimental conditions. Rhodium-perovskite, by means of SEM analysis, showed a lower carbon deposition than the nickel-based commercial catalyst, suggesting that the use of this innovative active phase on long time operations could prevent catalyst deactivation, enhancing the steam reforming process.

Fig. 7. Percentages of gaseous species on dry basis at different contact time as function of temperature for rhodium-perovskite catalyst.

M. Zeppieri et al. / Applied Catalysis A: General 387 (2010) 147–154

153

Fig. 12. Calculated forward reaction rates for commercial Ni-based catalyst. Fig. 9. Methane conversion for commercial Ni-based catalyst (contact time 7.6 ms) and rhodium-perovskite catalyst (contact time 7.4 ms).

Fig. 13. Arrhenius plot for commercial Ni-based catalyst.

Fig. 10. Calculated forward reaction rates for rhodium-perovskite.

Fig. 11. Arrhenius plot for rhodium-perovskite.

4.5. Kinetic analysis The experimental results at different contact times were used to perform kinetic analyses based on the selected model. Figs. 10–13 show the forward reaction rates and the Arrhenius plots for

rhodium-perovskite catalyst and for the commercial catalyst, respectively. In both cases, the model completely fits the experimental data and allows to calculate activation energy and pre-exponential values at the different contact times. The resulting average values of activation energy and preexponential factor for rhodium-perovskite and nickel-based catalyst are reported in Table 4. The obtained kinetic parameters were used to validate the model with a new set of experimental data. The parity chart, for both rhodium-perovskite and commercial nickel-based catalysts, reported in Fig. 14 for methane conversion, shows that the calculated kinetic parameters well fit the experimental data, collected in the same range of temperature used to determine kinetic values, with H2 O/CH4 equal to 3 and at atmospheric pressure. The found kinetic values in the case of the commercial nickelbased catalyst, 96.1 kJ/mol and 7.26E+07 s−1 for the activation energy and for the pre-exponential factor, respectively, are in good agreement with literature data. The activation energy measured for the steam reforming after Wei and Iglesia is 102 kJ/mol [15], similar to those previously reported for steam reforming on Ni/YZrO2 of 95 kJ/mol [19] and within a broad range of values reported for Ni supported on Al2 O3 (74–118 kJ/mol [20–23]) and modified Al2 O3 (90–107 kJ/mol [24,25]). The density-functional theory estimated the activation energy of 85 kJ/mol [26], 93 kJ/mol [27], and

Table 4 Calculated activation energy and pre-exponential factors for rhodium-perovskite and commercial nickel-based catalysts.

Rhodium-perovskite Ni-based catalyst

Ea (kJ/mol) average value

k0a (s−1 ) average value

k0b (mol m−3 s−1 bar−1 ) average value

69.1 ± 1 96.1 ± 2

2.88E+06 7.26E+07

1.72E+07 9.46E+07

154

M. Zeppieri et al. / Applied Catalysis A: General 387 (2010) 147–154

ues of the activation energy, lower for rhodium-perovskite catalyst than for nickel-based one, are in the range of those reported in the literature. In the case of rhodium-perovskite, the main differences between the data found in the present work and the value found by Wei and Iglesia could be explained with the different nature of the support analyzed, which was Al2 O3 . Perovskite catalyst could interact differently with methane molecules, and the dispersion of the metal could affect activation energy as well as the metal cluster dimensions. Further analysis of the characteristics of perovskite structures could explain some of this points and make clear the phenomena that could affect the activity of this innovative active phase. Acknowledgements Fig. 14. Parity chart for rhodium-perovskite and commercial Ni-based catalysts.

100 kJ/mol [28] for the C–H bond activation of CH4 on Ni(1 1 1) surfaces, while embedding methods led to slightly lower values (72 kJ/mol [29]). These theoretical estimates are in reasonable agreement with experiments and consistent with the conclusion that the dynamics of CH4 reforming and its decomposition reflects only the rate constants for the C–H bond activation elementary steps on Ni surfaces. For rhodium-perovskites the calculated Ea value of 69.1 kJ/mol differs from those reported in the literature. For methane reforming over Rh/Al2 O3 Wei and Iglesia [17] calculated an activation energy of 109 kJ/mol. In contrast, using field emission and molecular beam methods the dissociative adsorption of CH4 on rhodium films is reported to occur with an activation energy of 29 kJ/mol [28], while measures of thermally averaged dissociation probabilities for CH4 molecular beams on rhodium films gave an activation energy of 46.4 kJ/mol [29]. By density-functional theory calculations for CH4 reactions on flat Rh(1 1 1), stepped Rh(2 1 1), and kinked Rh surfaces reported C–H activation energies of 67, 32, and 20 kJ/mol, respectively [30]. These theoretical estimates are in agreement with the activation energy found in this work for rhodium-perovskite catalyst, considering the non-homogeneity of the comparison, in view of the perovskitic structure of the studied catalyst, for which no activation energy data are actually available in the literature. 5. Conclusion In this study the performances of a rhodium-perovskite catalyst in a SMR application have been evaluated and compared to those obtained with a commercial nickel-based SMR catalyst. In the range of temperature and steam-to-methane molar ratio investigated the perovskite catalyst shows higher methane conversion. The kinetic study confirms that, under low methane conversions and in the region of intrinsic kinetics, the CH4 rate of reforming reaction with steam is first order with regard to methane and zero order with regard to steam. Methane conversion is proportional to the partial pressure of methane and to the contact times. The model used for the calculation of the kinetic parameters, developed by Wei and Iglesia, has demonstrated to be adequate in the interpretation of the experimental results. In particular the val-

This work has been carried out within the framework of the project “Pure hydrogen from natural gas through reforming up to total conversion obtained by integrating chemical reaction and membrane separation” financially supported by the Italian Ministry of University and Research (FISR DM 17/12/2002). References [1] J.R. Rostrup-Nielsen, Catal. Today 63 (2000) 159–164. [2] C.H. Barthlomew, R.J. Farrauto, Fundamentals of Industrial Catalytic Processes, second ed., Wiley Interscience, Hoboken, NJ, 2006. [3] J.R. Rostrup-Nielsen, Catal. Today 18 (1993) 305–324. [4] P. Viparelli, P. Villa, F. Basile, F. Trifirò, A. Vaccari, P. Nanni, M. Viviani, Appl. Catal. A 280 (2005) 225–232. [5] F. Cifà, P. Dinka, P. Viparelli, S. Lancione, G. Benedetti, P.L. Villa, Appl. Catal. B 46 (2003) 463–471. [6] P.L. Villa, Soluzioni solide a struttura perovskitica comprendenti metalli nobili, utili come catalizzatori, Italian Patent Application no. MI2001A 001519 of 17.07.01, granted on 21.12.04, no. 1325822. [7] P.L.Villa, Solid solutions, applicable as catalysts, with a perovskite structure comprising noble metals, PCT/EP02/07906 on the 17.07.02 and USA Patent no. 7,166,267 (granted 23.01.07), European Patent Application (filed on 16.07.04). [8] K. Hou, R. Hughes, Chem. Eng. J. 82 (2001) 311–328. [9] G.F. Froment, K.B. Bischoff, Chemical Reactor Analysis and Design, second ed., Wiley, New York, 1990. [10] http://www.grc.nasa.gov/WWW/CEAWeb/. [11] J. Xu, G.F. Froment, AIChE J. 35 (1989) 88–96. [12] J.R. Rostrup-Nielsen, J. Sehested, J.K. Nørskov, Adv. Catal. 47 (2002) 65–139. [13] R.G. Compton, Kinetic Models of Catalytic Reactions, Elsevier, New York, 1991. [14] J. Wei, E. Iglesia, J. Phys. Chem. B 108 (2004) 4094–4103. [15] J. Wei, E. Iglesia, J. Catal. 224 (2004) 370–383. [16] J. Wei, E. Iglesia, J. Phys. Chem. B 108 (2004) 7253–7262. [17] J. Wei, E. Iglesia, J. Catal. 225 (2004) 116–127. [18] M. Alifanti, J. Kirchnerova, B. Delmon, Appl. Catal. A 245 (2003) 231–243. [19] K. Ahmed, K. Foger, Catal. Today 63 (2000) 479–487. [20] L.M. Aparico, A.G. Ruiz, I.R. Ramoa, Appl. Catal. A 170 (1998) 177–187. [21] O. Tokunaga, S. Ogasawara, React. Kinet. Catal. Lett. 39 (1989) 69–74. [22] T. Horiuchi, K. Sakuma, T. Fukui, Y. Kubo, T. Osaki, T. Mori, Appl. Catal. A 144 (1996) 111–120. [23] A.A. Lemonidou, L.A. Vasalos, Appl. Catal. A 228 (2002) 227–235. [24] I.M. Ciobîc, F. Frechard, R.A. van Santen, A.W. Kleyn, J. Hafner, J. Phys. Chem. B 104 (2000) 3364–3369. [25] A.P.J. Jansen, H. Burghgraef, Surf. Sci. 344 (1995) 149–158. [26] P. Kratzer, B. Hammer, J.K. Nørskov, J. Chem. Phys. 105 (1996) 5595–5604. [27] H. Yang, J.L. Whitten, J. Chem. Phys. 96 (1992) 5529–5537. [28] C.N. Stewart, G. Ehrlich, J. Chem. Phys. 62 (1975) 4672–4682. [29] S.G. Brass, G. Ehrlich, J. Chem. Phys. 87 (1987) 4285–4293. [30] Z. Liu, P. Hu, Am. Chem. Soc. 125 (2003) 1958–1967.