Aging of Premixed Metal Fiber Burners for Natural Gas Combustion

Aug 3, 2007 - A novel generation of catalytic premixed fiber burners for low-environmental- impact natural gas combustion are based on Pd/LaMnO3‚2Zr...
1 downloads 0 Views 1MB Size
Download PDF
6666

Ind. Eng. Chem. Res. 2007, 46, 6666-6673

Aging of Premixed Metal Fiber Burners for Natural Gas Combustion Catalyzed with Pd/LaMnO3‚2ZrO2 Stefania Specchia,* Marcela A. Ahumada Irribarra, Pietro Palmisano, Guido Saracco, and Vito Specchia Dipartimento di Scienza dei Materiali ed Ingegneria Chimica, Politecnico di Torino Corso Duca degli Abruzzi 24, 10129 Torino, Italy

The aim of this work is the study of the aging effect induced by S-compounds added as odorants in the natural gas network over a previously developed catalysts now deposited on a FeCrAl alloy fiber mat (NIT 200/S by ACOTECH BV). A novel generation of catalytic premixed fiber burners for low-environmentalimpact natural gas combustion are based on Pd/LaMnO3‚2ZrO2 catalyst, where ZrO2 acts as a structural promoter and the noble metal/perovskite synergism is effectively exploited, leading to enhanced performances with respect to noncatalyzed burners. The catalytic premixed fiber burners were characterized by means of field emission scanning electron microscope/energy dispersive spectrometer (FESEM/EDS), temperature programmed desorption/reduction/oxidation (TPD/R/O), and thermogravimetric/mass spectroscopy (TG/MS) techniques. The combustion runs were accomplished in a domestic boiler pilot plant under realistic operating conditions. An unexpected improvement was found in the overall performance of catalyzed burners after sulfur aging carried out up to 3 weeks (the catalyzed burners were kept in an electric oven at 800 °C under a N2 flow containing 200 ppmv of SO2). Despite the effect of catalyst thermal stress and aging, the overall performance, in terms of pollutant emissions, was preserved after accelerated lifetime utilization. This is probably due to the presence of microfractures on the aged burners, accomplished by an increase of active sites for β-type oxygen desorption, able to improve methane combustion. 1. Introduction In the past decade, the use of methane as a source of thermal energy in hydrothermal plants for domestic application has known a relevant increase to respect the ultimate modern climate policy requirements. Natural gas combustion is widely studied and analyzed in relation to the concentration of pollutants like NOx and CO, and several solutions have been evaluated to reach new highly efficient and clean combustion appliances.1 A good solution in this field has been reached in premixed combustion within porous media coupled with a catalytic system based on perovskite-based catalyst, as demonstrated in numerous extensive experimental research works:2-10 the improvement of the fuel flow rate fraction burnt within or just downstream the catalytic burner surface is able to maximize the heat fraction transferred by radiation, cooling the flame, and favoring the combustion completion with lower CO and unburned hydrocarbon emission levels.4 Moreover, this solution allows an increase in the overall thermal efficiency and a reduction of flame temperature, which results in a limited production of thermal NOx. In perovskite-based catalysts, the catalytic activity is generally determined by the B cation. Substitution of A or B ions with cations having a lower oxidation state results in the formation of oxygen vacancies and/or in the oxidation of a fraction of B cations to a higher valence state.8,11,12 The presence of the transition metal is essential for a high activity; in fact, it enhances redox properties (easy change of cation oxidation state with fast bonding/release of surface oxygen) which are beneficial for the CO oxidation reaction.10 Moreover, the high stability of the perovskite structure allows the partial substitution of both A and B site cations by other metals with different oxidation * To whom correspondence should be addressed. Tel.: +39.011. 5644608. Fax: +39.011.5644699. E-mail: [email protected].

states and consequent creation of structural defects such as anionic or cationic vacancies.10,13 A widely variable perovskite characterized by a nonstoichiometric O2 (reductive or oxidative, ABO3) structure can thus be achieved with a high ease of removing the O2 but still preserving the original lattice framework. The mechanisms of redox processes on these materials (both on the surface and in the bulk) account for the reversible loss and uptake of O2 and/or for the creation and filling up of vacancies. The effect of the nature of the B cation on the catalytic properties of La-based perovskites was widely studied: perovskites containing Mn, Co, and Fe were found to behave as the most active in CH4 combustion.9,13 In this work, the effect of aging induced by S-compounds (added as odorants in the natural gas network) on catalytic burners prepared by the authors was evaluated in a pilot plant under realistic operating conditions. The FeCrAlloy fiber-mat burners were catalyzed by means of an ad-hoc developed deposition technique based on solution combustion synthesis (SCS).5,6 An optimized pretreatment favoring the formation of an R-Al2O3 layer on the fibers’ surface was designed, both to protect the metal alloy from further oxidation when the burner is operated and to allow an optimal catalyst anchoring.7 SCS operating conditions on the FeCrAlloy fiber-mat burners were adjusted and tailored in order to find the best compromise between adhesion and specific surface area of the deposited catalyst layer. A zirconia-stabilized perovskite catalyst, LaMnO3‚ 2ZrO2, was chosen, in line with earlier investigations,5,14,15 and an attempt was made to exploit its potential synergism with palladium. ZrO2 acts as a structural promoter, in that its presence limits the specific surface area loss caused by prolonged exposure to high temperatures.16,17 Moreover, the potential benefits of ZrO2 to enhance the temperature stability of LaMnO3 perovskite has been broadly demonstrated.15 Therefore, the dispersion of a noble metal over a support like perovskite,

10.1021/ie061665y CCC: $37.00 © 2007 American Chemical Society Published on Web 08/03/2007

Ind. Eng. Chem. Res., Vol. 46, No. 21, 2007 6667

Figure 1. Scheme (cross section and frontal view) of the combustion chamber used during the experimental runs.

intrinsically active toward methane oxidation, is expected to have favorable effects on the overall catalytic activity. 2. Experimental 2.1. Catalytic Fiber-Mat Burners Preparation. Three catalytic fiber-mat burners were prepared by the authors by using commercial knitted fiber mats (Acotech NIT 200/S) made of FeCrAlloy. The catalytic panel supports were first kept at 1200 °C for 10 min under O2 flow (0.5 vol % in N2) so as to favor the regular growth of the R-Al2O3 grains into a uniform protective layer and develop an external surface morphology able to ensure a good adherence of the catalytic phase to be deposited on the metallic mat.7 The LaMnO3‚2ZrO2 catalyst was deposited via SCS-spray pyrolysis. The aqueous solution of the perovskite/zirconia precursors was sprayed over the surface of the FeCrAlloy panels, previously heated at 400 °C. Due to the in situ pyrolysis/combustion synthesis occurring on the hot panel’s surface, catalyst formation was obtained. The panels were then placed back into the hot oven to stabilize the coating. The spray deposition cycle was repeated several times in order to achieve the desired catalyst load (namely, 2% w/w).14 For a further stabilization and complete crystallization of the catalytic phase, the burners were finally calcined at 900 °C for 2 h in still air. The Pd deposition was performed by means of further spray-pyrolysis runs over the perovskite/zirconia layer, by employing a diluted Pd(NO3)2 aqueous solution. Calcination in air at 600 °C followed, in order to promote the full decomposition of Pd(NO3)2 into the oxidized form PdO.14 2.2. Aging Procedure. Since Pd/perovskite-based catalysts are prone to be poisoned by sulfur compounds, the catalyzed panels were kept in an electric tubular oven at 800 °C (a thermocouple was used to monitor the furnace temperature) under a N2 flow containing 200 ppmv of SO2, chosen several times higher than the odorant concentration added to commercial natural gas (about 8 ppmv, in Italy, as THT) so as to accelerate any possible poisoning effect.6,18,19 Earlier sulfur poisoning studies on perovskite catalysts showed that the basic poisoning mechanism under catalytic combustion conditions is chemisorption of SO2/SO3 species generated by combustion of whatever sulfur-organic compound present in the feed (e.g., odorants).19 Particularly, it was demonstrated that the direct presence of SO2 or of the THT odorant in the feed does not lead to a significantly different aging behavior, provided the overall sulfur content is the same. SO2/SO3 species, in fact, result from the total oxidation of methyl-mercaptan during methane and natural gas combustion.20 For this reason only, aging runs with a SO2-laden flow were accomplished. The catalyzed burners were continuously aged up to 3 weeks: for a domestic

boiler, such a time under 200 ppm SO2 atmosphere may be considered equivalent to an utilization lifetime of approximately 3 years under real operation. 2.3. Catalytic Activity Tests. Tests under realistic operating conditions were performed on a partially modified commercial condensing boiler rig for domestic application (production of hot water for sanitary and heating purposes). CH4 was fed to a modulating electrovalve, with which its volumetric flow rate could be varied (max power Q ∼ 30 kW). Air coming from a blower was mixed with CH4 in a Venturi positioned so that a proper mixing was achieved before entering the burner. The fiber-mat burner, fitted vertically in the combustion chamber, fired horizontally through the heat exchanger coils. The burner diameter was approximately 10 cm. A scheme of the combustion chamber used for the experimental runs is represented in Figure 1. Tests were carried out over a wide range of operating conditions by varying the nominal power Q from 10 to 25 kW and the air excess Ea from 2 to 45% (from λ ) 1.02 to λ ) 1.45). A K-type thermocouple measured the temperature on the downstream surface of the burner deck. The flue gases composition (O2, CO2, CO, and NO) was monitored by means of a multiple gas continuous analyzer (from ABB, Uras 14 infrared analyzer and Magnos 16 paramagnetic sensor). Tests were performed both on fresh and aged burners, after 1 and 3 weeks of aging, respectively. The obtained results in terms of CO and NO emissions and temperature are shown in Figures 2-4. 2.4. Field Emission Scanning Electron Microscope/Energy Dispersive Spectrometer Analysis. A field emission scanning electron microscope (FESEM Leo 50/50 VP with Gemini column) coupled with an energy dispersive spectrometer (EDAX 9900 EDS by Oxford Instrument combined with INCA energy software) was employed to analyze the microstructure of the catalytic surface deposited over the as-prepared burners before and after each aging treatment, in order to assess the influence of both the exposure to S-compounds and high temperatures on the catalytic layer morphology. The obtained results are plotted in Figures 5 and 8. 2.5. Temperature Programmed Desorption/Temperature Programmed Reduction Analysis. The capability of catalyzed fibers to act as O2 pumps was investigated by temperature programmed desorption (TPD) of O2 carried out by means of a Termoquest TPD/R/O 1100 Series Thermo Finningan analyzer, equipped with a thermal conductivity detector (TCD). A fixed bed of catalyzed fibers (about 500 mg in weight) was enclosed in a quartz tube and sandwiched between two quartz layers; then, an oxidation pretreatment was carried out by heating the fixed bed under an O2 flow (40 N mL/min) up to 750 °C. Then,

6668

Ind. Eng. Chem. Res., Vol. 46, No. 21, 2007

Figure 2. CO emissions (dry gases) vs Ea, at three different power output values, for the fresh catalytic burner and the 1 and 3 week aged catalytic burners.

tion over the catalyst. Afterward, He was fed to the reactor at 25 N mL/min flow rate for 1 h at room temperature, thus purging out any excess of gaseous O2. TPD tests on the fresh, 1 week, and 3 week aged burners were then performed by heating the catalyst up to 1100 °C at a 10 °C/min rate under He flow. The O2 desorbed during heating was detected in the outlet flow by the TCD detector after proper calibration. The obtained results are plotted in Figure 6. Temperature programmed reduction (TPR) experiments, carried out to evaluate the fraction of Pd present as metal or oxide in the fresh burner, were carried out in the same apparatus. After the same oxidation pretreatment adopted for the TPD runs, the samples were reduced with a 4.95% H2/He mixture (10 N mL/min) as they were heated at a rate of 10 °C/min to 1100 °C. Once again, the amount of converted H2 was monitored via the TCD detector after proper calibration. 2.6. TG/MS Analysis. Finally, to evaluate the temperature range in which the SO2 can be adsorbed and desorbed from the burner’s catalytic layer and try to better understand the role and interactions of S-compounds with the catalyst, three different aging treatments, respectively at 400, 600, and 800 °C for 24 h19-21 under a N2 flow containing 200 ppmv of SO2, were conduced on three small cut samples of the as-prepared fresh catalyzed burners. Thermogravimetic analysis (TGA MettlerToledo) runs, carried out from 25 to 1100 °C, with a heating rate of 10 °C/min in a nitrogen flow of 50 mL/min, were conducted over a sample of each treated cloth. The released gases from the TGA cell were analyzed by a quadrupole mass spectrometer (MS Balzers), in order to detect simultaneously during a temperature scan the samples’ weight variation and SO2/SO3 gaseous desorption. The obtained results are shown in Figure 7. 3. Results and Discussion

Figure 3. NO emissions (dry gases) vs Ea, at three different power output values, for the fresh catalytic burner and the 1 and 3 week aged catalytic burners.

Figure 4. Temperature values (measured on the downstream surface of the burner deck) vs Ea, at three different power output values, for the fresh catalytic burner and the 1 and 3 week aged catalytic burners.

after a 30-min stay at this temperature under O2 flow, the reactor temperature was lowered down to room temperature under oxidizing atmosphere, thereby allowing for complete O2 adsorp-

Figures 2 and 3 show the CO and NO emissions of NIT 200/S fiber-mat burners catalyzed with Pd/LaMnO3‚2ZrO2 during the test runs in the combustion chamber in the fresh and the aged status, after 1 and 3 weeks, respectively, at three different power levels (low, medium, and high): 12, 18, and 22 kW, respectively. By increasing the power output, i.e., the fuel fed to the combustion chamber, the CO emission levels increased whereas the increase in NO emissions was very slight. Of course, by increasing the air excess, the overall values of both CO and NO emissions decreased due to the effect of air dilution. Moreover, the fiber-mat burners exposed to prolonged contact with S-compounds at high temperature performed better in terms of CO and NO reduction. Despite the aging of the tested catalysts, the more aged the catalyst, the more evident the CO reduction. The same is also valid for the NO emissions, but in a less noticeable way. Also, the temperature level on the burner deck showed a slight decrease with the aging time (Figure 4). It was quite evident, especially for CO emissions, that the aged burners performed better in term of emission reduction. Concerning the aged burners, especially the most aged one, when the excess of air approached the stoichiometric conditions, the catalytically assisted combustion was strongly favored, thus lowering significantly the CO emissions. With the catalyst promoting the complete methane oxidation, and increasing the CO oxidation, a greater portion of the combustion heat can be released within the porous medium; this increases the radiant output and the overall thermal efficiency and in the meantime decreases the burner’s deck temperature, as pointed out in Figure 4. If the catalyst can stabilize the oxidation reactions deeper inside the metal fiber mat, the gas phase remains cooler since

Ind. Eng. Chem. Res., Vol. 46, No. 21, 2007 6669

Figure 5. FESEM analysis for the fresh (A.1/A.2), 1 week (B.1/B.2), and 3 week (C.1/C.2) aged burners at 2500× (.1 images) and 40 000× (.2 images). In C.2, a microfracture is pointed out.

the porous medium absorbs and radiates directly to the heat sink part of the combustion heat. Calcination at high temperature (750 °C and more) decreases the dispersion of Pd for deep oxidation, but its specific activity increases.22 The raise in specific activity can compensate for the decrease in supported metal dispersion, and as a whole, the catalytic activity increases: this is evident especially for the most aged burner.

In the present case, NO2 is practically absent. NO originates mainly from the Zeldowich thermal process,23-25 since the CO molecule does not contain N atoms, and when the gas mixture is sufficiently lean (Ea ) 20% or more), the formation of radicals CH and HCN is not significant (Fenimore mechanism).23-29 Moreover, thermal NO emissions are expected to decrease with lower flame temperatures. Catalytic combustion is operated at

6670

Ind. Eng. Chem. Res., Vol. 46, No. 21, 2007

Figure 6. TPD analysis for the fresh, 1 week, and 3 week aged burners as concerns the released oxygen (R-type at T < 600 °C; β-type at T > 600 °C).

lower temperature compared to the conventional flame combustion, thus reducing the formation of thermal NOx, which starts at about 1500 °C.30,31 This is confirmed by the decreasing trend of NO emissions shown in Figure 3 coupled with the decrease of the temperatures on the burner deck (see Figure 4). As general statement, the temperature level increases by increasing the power output and decreases by increasing the excess of air and the aging of the catalyst deposited over the fiber-mat burner. Another interesting effect linked to the aging of the metal fiber-mat burners is related to the catalyst morphology.32-34 The positive effect of aging might be due to the morphological modification of the fiber-mat surface that could to some extent cause a reduction of the local void fraction; this allowed it to maintain the same local gas momentum intensity with decreased flow rates. This was quite evident in Figure 5, where FESEM micrographs of the fresh and aged burners were reported at 2500 and 40 000 magnification, respectively: the fresh catalyst over each fiber shows a spongy structure, typically related to the perovskite phase (Figure 5A.1 and A.2) in analogy with that found in a perovskites powder prepared by the SCS method.8 Micrographs of the 1 week aged burner (Figure 5B.1 and B.2) evidenced a drastic reduction of porosity in the catalytic layer with some isolated superficial microfractures. Micrographs of

the 3 week aged burner (Figure 5C.1 and C.2) showed the appearance of a new and completely different morphology, characterized by a globular structure, no longer spongy but with more microfractures (pointed out at higher magnification) compared to the 1 week aged burner. The observed porosity reduction and the growth of elementary agglomerates is perfectly in line with data reported by other authors,17,33,34 but not the catalytic activity, which is not only preserved but also enhanced on our structured catalysts (Figures 2 and 3). All the aged fibers, regardless of the aging time, preserved a good catalyst adhesion and an atomic sulfur presence of about 1%, measured via EDS analysis. It is well-known that, when a perovskite is heated at high temperature, oxygen vacancies can be formed,35 depending of the number of structural and electronic defects and corresponding cation/anion vacancies due to nonstoichiometry;28,36 for LaMn-based perovskites, these vacancies contribute to the catalytic activity in full oxidation.18 As thoroughly discussed in a number of papers,37-43 two types of chemisorbed oxygen species, and the related desorption peaks, were defined as follows: a lowtemperature species, the R-type, desorbed in the 300-600 °C range, and a high-temperature one, the β-type, desorbed at 600900 °C. The R peak is not always observable and strongly depends on the concentration of surface oxygen vacancies,44 depending also in part on the nature of the metal B of the ABO3 structure but mainly on the degree of substitution of the A ion with ions of lower valence.32 It also well-known in the literature that methane oxidation at low temperature is a suprafacial reaction involving oxygen coming from the gas phase or sitting at the oxygen vacancies of the catalysts.18,44,45 The β peak, characterized by a higher onset temperature, is strictly related to the nature of the B ion, and its occurrence is strictly linked to redox transitions of the valence state of this ion. The most likely explanation is that the perovskite can act as an oxygen pump toward the methane molecule. Oxygen molecules coming from the gaseous atmosphere can reoxidize the perovskite. Such an intrafacial mechanism should mostly regard β-type oxygen species.46,47 It was also demonstrated how this β-type oxygen could be responsible for ignition of natural gas combustion through an Elay-Rideal mechanism.48 Such a mechanism leads,

Figure 7. TG/MS analysis for burners aged 24 h at 400, 600, and 800 °C: burners relative weight loss on the left side; released SO2 and SO3 from the burners on the right side. No SO2 and SO3 signals are detected for burners aged 24 h at 400 and 800 °C.

Ind. Eng. Chem. Res., Vol. 46, No. 21, 2007 6671

Figure 8. FESEM analysis for burners aged 24 h at 400 (A), 600 (B), and 800 °C (C).

as a first step, to hydrogen abstraction from the methane molecule by reaction with monatomic intrafacial oxygen. Moreover, it was widely demonstrated that the CO oxidation is strongly favored by the presence of R-type oxygen.8 H2 TPR analysis showed that on the fresh burner approximately 30% of deposited Pd is under the oxidized form, whereas the remaining 70% is metallic Pd. O2 TPD spectra (Figure 6) showed peaks of O2 released at low or high temperature, depending on the aging of the burners: the fresh and the 3 week aged burners release only intrafacial β-type oxygen (46.9 and 179.8 µmol/g, respectively), in particular a higher quantity for the most aged one, whereas the 1 week aged sample showed also the desorption of

suprafacial R-type oxygen (30.8 and 70.1 µmol/g as R- and β-type oxygen, respectively). The β-type oxygen released by the most aged burner is more than 3 times that of the fresh one, i.e., such a burner has the capability to better reduce the emissions during the combustion of natural gas. This is in line with the results obtained with the tests on the boiler rig (lower emissions are released by increasing the aging of the burners) and with the results obtained with the FESEM micrographs (increasing the aging of the burners causes more microfractures to appear, favoring thus the release of larger amounts of intrafacial β-type oxygen from the deepest parts of the burner). The values of O2 released at low temperature suggest that the oxygen evolution must be related only to the catalyst surface since the larger amounts of O2 released at high temperature should be expected for a phenomenon involving the bulk of the catalyst.8,30 This is in line with the presence of the microfractures on the 3 week aged burner (coupled with a wide exposition of active points with the methane/air mixturessee Figure 5C.2), which showed the highest values of intrafacial β-type oxygen released, accompanied by an optimal performance of the catalytic burner during the bench tests (Figures 2 and 3), thus confirming the importance of catalyst morphology in the catalytic mechanism of methane combustion.20,33 Such intrafacial β-type oxygen is presumably released from the catalytic internal bulk material now exposed through the new surface of the microfractures.49 Consequently, the enhancement on the actives sites for oxygen adsorption (oxygen vacancies) is likely to allow the in-plane oxygen from the catalyst bulk material be more easily available.44 Previous papers8,9,50 demonstrated that in perovskites of the family AMnO3 there is a substantial fraction of manganese as Mn(IV) (35% for LaMnO3): the reduction steps Mn(IV) f Mn(III) f Mn(II) occurring in LaMnO3 perovskite, the former at low temperature and the latter at high temperature, are mainly responsible for the CO oxidation in the low-temperature range and for the CH4 oxidation at high temperature, respectively. LaMnO3 perovskite, where Mn is more stable toward reduction compared to other rare elements such as Sm or Nd, is capable to desorb larger quantities of β-type oxygen, as confirmed also by Figure 6 showing the TPD results, thus confirming it to be a very active catalyst at high temperature.13 The TG/MS measurements coupled with FESEM/EDS analysis, conducted on small cuts of fresh catalytic burners specifically aged for 24 h, were aimed to investigate the role of S adsorbed over the catalyst surface. The TG/MS results, reported in Figure 7, show how the burner aged at 400 °C did not lose weight during the programmed temperature scaling up; moreover, the MS did not detect a variation for the SO2 and SO3 signals. The sample aged at 800 °C lost only 0.4% of its initial weight, but the SO2 and SO3 signals remained again undetected during the whole test. Conversely, the sample aged at 600 °C lost about 1.5% of its initial weight, and simultaneously, the MS detected a significantly variation of the SO2 and SO3 intensity signals, equivalent to approximately 127 and 80 µmol/g of SO2 and SO3 released, respectively. FESEM/EDS analysis (Figure 8) performed in parallel on the same samples leads to the same conclusions: samples aged at 400 and 800 °C did not show any variation in the catalyst morphology, with the typical spongy structure of perovskite well preserved, with no sulfur at all deposited over the fibers. On the contrary, the sample aged at 600 °C had a completely different morphology, and the presence of sulfur was detected by EDS sensor (approximately 8% atomic). Such a difference was also visible with the naked eye: in fact, such a sample

6672

Ind. Eng. Chem. Res., Vol. 46, No. 21, 2007

appeared to be a red color whereas the other two samples aged at 400 and 800 °C appeared to be a gray color, exactly as the fresh catalytic burner. In the literature, all the studies on S-compound poisoning were carried out on powdered catalysts and not on structured catalysts: deactivation of transition metal oxide based catalysts by sulfur oxides is mainly caused by active site obstruction due to the formation of surface or subsurface and bulk sulfates, where surface sulfates are formed at first, followed by bulk surface sulfates.20,32,33 Depending on the experimental conditions and catalyst morphology, this process may vary from a completely reversible phenomenon (due to surface sulfates) to an irreversible poisoning of the catalyst (due to bulk sulfates),33 resulting in a complete loss of catalytic activity in a relatively short time by transformation of the perovskite structure into metal oxides and sulfates. In agreement with the relative thermodynamic stability of various transition metal sulfates, as well as kinetic data on their formation, surface sulfates, formed after adsorption, are thermally unstable and easily desorb as SO2/SO3 (the last as an oxidation product of SO2) at approximately 500 °C, whereas bulk sulfates are more resistant to thermal effects.32,33,51 This is in line with our TG/MS measurements (Figure 7), where SO2/SO3 from the aged sample at 600 °C were released from 500 to 600 °C. Such adsorption/ desorption phenomena of S-compounds from the catalytic layer for Pd/LaMnO3‚2ZrO2 catalysts occurred around 550 °C, well below the boiler working temperature range (raising from 750 to 1150 °C). Probably, on our structured catalysts, where the catalytic layer completely covers the metal fibers with a thickness of a few micrometers (1 µm of R-Al2O3 from the preoxidation treatment plus another 1-2 µm of perovskite),5,7 only surface sulfates are present on the catalyst layer, which are volatile at such temperatures, thus resulting in a limited presence of S-compounds in the 1 and 3 week aged burners (as mentioned before, EDS analysis revealed the presence of only 1% atomic S on the burners compared to the 8% atomic S on the burner aged for 24 h at 600 °C) and consequently in an increase of the catalysts resistance to poison. Moreover, S-compound poisoning is limited when SO2 is added at high temperature due to the weak adsorption of SO2 at these temperatures, which additionally arises from the competitive adsorption of H2O produced in the oxidation reaction, and the most thermally unstable species desorb at high temperature.52,53 According to these experimental results, and moreover to the higher catalytic activity registered for the aged burners, the role of S-compounds might be considered as not relevant in the combustion of CO and CH4 on structured catalysts at high temperature because they do not determine either the coating or poisoning of the active sites. Further investigations are in progress to determine an eventual promoting role of ZrO2 toward the oxidation reactions, or a protecting role against sulfur poisoning, and to compare the aging effects on powdered and structured catalysts. 4. Conclusions The developed methane combustion catalyst (Pd/LaMnO3‚ 2ZrO2) and the adopted technique for its deposition allowed the design of a premixed metal fiber burner with very interesting performance, especially in the aged state: it was able, over a wide range of operating conditions, to give limited pollutant emissions, in terms of both CO and NO. The accelerated aging treatment carried out at 800 °C with a SO2 concentration 25 times higher than the real THT concentration employed for the natural gas network in Italy (conditions

equivalent to an utilization lifetime of about 1 year of normal operation) did not produce any poisoning effect on the catalyst. The obtained results in terms of pollutant emissions seem, in fact, to guarantee that the aged burner can still contribute to an environmentally cleaner combustion, compared to the noncatalytic one. This is probably due to compensation effects produced by the increase of both the available catalytic surface derived from induced surface microfractures and the actives sites for β-type oxygen adsorption/desorption (oxygen vacancies) in the fiber burners substrate; more surface and active sites seem to be created under the thermal/aging treatment that is likely to let the in-plane catalyst oxygen be more easily available. Acknowledgment Funding of the European Community is gratefully acknowledged (EU project CAT-NAT nr G5RD-CT2001-00567: Costeffective and durable nanostructured Pd catalysts for natural gas vehicle and premixed burners applications). Literature Cited (1) Forzatti, P.; Groppi, G. Catalytic combustion for the production of energy. Catal. Today 1999, 54, 165. (2) Saracco, G.; Cerri, I.; Specchia, V.; Accornero, R. Catalytic premixed fiber burners. Chem. Eng. Sci. 1999, 54, 3599. (3) Cerri, I.; Saracco, G.; Specchia, V.; Trimis, D. Improved-performance knitted fiber mats as supports for pre-mixed natural gas catalytic combustion. Chem. Eng. J. 2001, 82, 73. (4) McCarty, J. G.; Wise, H. Perovskite catalysts for methane combustion. Catal. Today 1990, 8, 231. (5) Specchia, S.; Civera, A.; Saracco, G. In situ combustion synthesis of perovskite catalysts for efficient and clean methane premixed metal burners. Chem. Eng. Sci. 2004, 59, 5091. (6) Specchia, S.; Civera, A.; Saracco, G.; Specchia, V. Palladium/ Perovskite/Zirconia catalytic premixed fiber burners for efficient and clean natural gas combustion. Catal. Today 2006, 117, 427. (7) Ugues, D.; Specchia, S.; Saracco, G. Optimal micro-structural design of a catalytic premixed FeCrAlloy fiber burner for methane combustion. Ind. Eng. Chem. Res. 2004, 43, 1990. (8) Ciambelli, P.; Cimino, S.; De Rossi, S.; Faticanti, M.; Lisi, L.; Minelli, G.; Pettiti, I.; Porta, P.; Russo, G.; Turco, M. AMnO3 (A ) La, Nd, Sm) and Sm1-xSrxMnO3 perovskites as combustion catalysts: structural, redox and catalytic properties. Appl. Catal. B: EnViron. 2000, 24, 243. (9) Ciambelli, P.; Cimino, S.; De Rossi, S.; Lisi, L.; Minelli, G.; Porta, P.; Russo, G. AFeO3 (A ) D, La, Nd, Sm) and LaFe1-xMgxO3 perovskites as methane combustion and CO oxidation catalysts: structural, redox and catalytic properties. Appl. Catal. B: EnViron. 2001, 29, 239. (10) Ciambelli, P.; Cimino, S.; Lasorella, G.; Lisi, L.; De Rossi, S.; Faticanti, M.; Minelli, G.; Porta, P. CO oxidation and methane combustion on LaAl1-xFexO3 perovskite solid solutions. Appl. Catal. B: EnViron. 2002, 37, 231. (11) Arai, H.; Yamada, T.; Eguchi, K.; Seiyama, T. Catalytic combustion of methane over various perovskite-type oxides. Appl. Catal. 1986, 26, 265. (12) Zhong, Z.; Chen, K.; Ji, Y.; Yan, Q. Methane combustion over B-site partially substituted perovskite-type LaFeO3 prepared by sol-gel method. Appl. Catal. A: Gen. 1997, 156, 29. (13) Futai, M.; Yonghua, C.; Louhui, L. Characterization of perovskitetype oxide catalysts RECoO3 (RE ) La, Pr, Nd, Sm, Eu, Gd, Tb, Dy) by TPR React. Kinet. Catal. Lett. 1986, 31, 47. (14) Civera, A.; Negro, G.; Specchia, S.; Saracco, G.; Specchia, V. Optimal compositional and structural design of a LaMnO3/ZrO2/Pd-based catalyst for methane combustion. Catal. Today 2005, 100, 275. (15) Cimino, S.; Pirone, R.; Lisi, L. Zirconia supported LaMnO3 monoliths for the catalytic combustion of methane. Appl. Catal. B: EnViron. 2002, 35, 243. (16) Mu¨ller, C. A.; Maciejewski, M.; Koeppel, R. A.; Baiker, A. Combustion of methane over palladium/zirconia derived from a glassy PdZr Alloy: effect of Pd particle size on catalytic behavior. J. Catal. 1997, 166, 36. (17) Gonza´lez-Velasco, J. R.; Gutie´rrez-Ortiz, M. A.; Marc, J. L.; Botas, J. A.; Gonza´lez-Marcos, M. P.; Blanchard, G. Effects of redox thermal treatments and feedstream composition on the activity of Ce/Zr mixed oxides for TWC applications. Appl. Catal. B: EnViron. 2000, 25, 19.

Ind. Eng. Chem. Res., Vol. 46, No. 21, 2007 6673 (18) Zhang-Steenwinkel, Y.; Castricum, H. L.; Beckers, J.; Eiser, E.; Bliek, A. Dielectric heating effects on the activity and SO2 resistance of La0.8Ce0.2MnO3 perovskite for methane oxidation. J. Catal. 2004, 221, 523. (19) Rosso, I.; Garrone, E.; Geobaldo, F.; Onida, B.; Saracco, G. Sulfur poisoning of LaMn1-xMgxO3 catalysts for natural gas combustion. Appl. Catal. B: EnViron. 2001, 30, 61. (20) Klvana, D.; Delval, J.; Kirchnerova, J.; Chaouri, J. Deactivation of fiber-supported La0.65Sr0.29Ni0.29Co0.69Fe0.02O3 catalyst by mercaptan during combustion of methane or natural gas. Appl. Catal. A: Gen. 1997, 165, 171. (21) Rosso, I.; Garrone, E.; Geobaldo, F.; Onida, B.; Saracco, G. Sulfur poisoning of LaMn1-xMgxO3‚yMgO catalysts for methane combustion. Appl. Catal. B: EnViron. 2001, 34, 29. (22) Tsyrulnikov, P. G.; Kovalenko, O. N.; Gogin, L. L.; Starostina, T. G.; Noskov, A. S.; Kalinkin, A. V.; Krukova, A. V.; Tsybulya, S. V.; Kudrya, E. N.; Bubnov, A. V.; Behavior of some deep oxidation catalysts under extreme conditions. 1. Comparison of resistance to thermal shock and SO2 poisoning. Appl. Catal. A: Gen. 1998, 167, 31. (23) Turns, S. R. An introduction to combustion, concepts and applications; McGraw-Hill: New York, 1996. (24) Miller, J. A.; Bowman, C. T. Mechanism and modeling of nitrogen chemistry in combustion. Prog. Energy Combust. 1989, 15, 287. (25) Wang, Z.; Zhou, J.; Zhang, Y.; Lu, Z.; Fan, J.; Cen, K. Experiment and mechanism investigation on advanced reburning for NOx reduction: influence of CO and temperature. J. Zhejiang UniV. Sci. B 2005, 6, 187. (26) Costa, F. de S.; Cardoso, J.; Villela, T. E. A.; Veras, C. A. G. Effects of wet CO oxidation on the operation of engines and power generators. J. Braz. Soc. Mech. Sci. Eng. 2003, 25, 341. (27) Bhargava, A.; Colket, M.; Sowa, W.; Casleton, K.; Maloney, D. An experimental and modeling study of humid air premixed flames. J. Eng. Gas Turbines Power 2000, 122, 3. (28) Rogg, B. Response and flamelet structure of stretched premixed methane-air flames. Combust. Flame 1988, 73, 45. (29) Heywood, J. B. Internal combustion engine fundamentals; McGrawHill: New York, 1988. (30) Tejuca, L. G., Fierro, J. L. G., Eds. Properties and applications of peroVskite-type oxides; Marcel Dekker: New York, 1993. (31) Kasuya, F.; Glarborg, P. The thermal DeNOx process: influence of partial pressures and temperature. Chem. Eng. Sci. 1995, 50, 1455. (32) Rosso, I.; Fiorilli, S.; Onida, B.; Saracco, G.; Garrone, E. Sulfate Species in MgO-Supported LaMn0.5Mg0.5O3 Perovskites: An Insight into the Chemistry of MgO. J. Phys. Chem. B 2002, 106, 11980. (33) Royer, S.; Van Neste, A.; Davidson, R.; McIntyre, S.; Kaliaguine, S. Methane oxidation over nanocrystalline LaCo1-xFexO3: resistance to SO2 poisoning. Ind. Eng. Chem. Res. 2004, 43, 5670. (34) Hietikkoa, M.; Lassib, U.; Kallinen, K.; Savima¨kid, A.; Ha¨rko¨nen, M.; Pursiainena, J.; Laitinena, R. S.; Keiski, R. L. Effect of the ageing atmosphere on catalytic activity and textural properties of Pd/Rh exhaust gas catalysts studied by XRD. Appl. Catal. A: Gen. 2004, 277, 107. (35) Fino, D.; Russo, N.; Saracco, G.; Specchia, V. The role of suprafacial oxygen in some perovskites for the catalytic combustion of soot. J. Catal. 2003, 217, 367. (36) Forni, L.; Rossetti, I. Catalytic combustion of hydrocarbons over perovskites. Appl. Catal. B: EnViron. 2002, 38, 29. (37) Seyama, T. Total oxidation of hydrocarbons on perovskite oxides. Catal. ReV. Sci. Eng. 1992, 34, 281.

(38) Teraoka, Y.; Yoshimatsu, M.; Yamazoe, N.; Seyama, T. Oxygensorptive properties and defect structure of perovskite-type oxides. Chem. Lett. 1984, 893. (39) Yamazoe, N.; Teraoka, Y. Oxidation catalysis of perovskiterelationships to bulk structure and composition (valency, defect, etc.). Catal. Today 1990, 8, 175. (40) Baythoun, M. S. G.; Sale, F. R. Production of strontium-substituted lanthanum manganite perovskite powder by the amorphous citrate process. J. Mater. Sci. 1982, 17, 2757. (41) Leanza, R.; Rossetti, I.; Fabbrini, L.; Oliva, C.; Forni, L. Perovskite catalysts for the catalytic flameless combustion of methane: preparation by flame-hydrolysis and characterisation by TPD-TPR-MS and EPR. Appl. Catal. B: EnViron. 2000, 28, 55. (42) Rossetti, I.; Forni, L. Catalytic flameless combustion of methane over perovskites prepared by flame-hydrolysis. Appl. Catal. B: EnViron. 2001, 33, 345. (43) Zhang, H. M.; Shimizu, Y.; Teraoka, Y.; Miura, N.; Yamazoe, N. Oxygen sorption and catalytic properties of La1-xSrxCo1-yFeyO3 perovskitetype oxides. J. Catal. 1990, 121, 432. (44) Cox, D. F.; Fryberger, T. B.; Semancink, S. Oxygen vacancies and defect electronic states on the SnO2. Phys. ReV. 1998, 38, 2072. (45) Pen˜a, M. A.; Fierro, J. L. G. Chemical structures and performance of perovskite oxides. Chem. ReV. 2001, 101, 1981. (46) Voorhoeve, R. J. H.; Remeika, J. P.; Johnson, D. W. Rare-Earth Manganites: Catalysts with low ammonia yield in the reduction of nitrogen oxides. Science 1973, 180, 62. (47) Voorhoeve, R. J. H. AdVanced materials in catalysis; Academic Press: New York, 1977. (48) Saracco, G.; Geobaldo, F.; Baldi, G. Methane combustion on Mgdoped LaMnO3 perovskite catalysts. Appl. Catal. B: EnViron. 1999, 20, 277. (49) Marchetti, L.; Forni, L. Catalytic combustion of methane over perovskites. Appl. Catal. B: EnViron. 1998, 15, 179. (50) Collongue, B.; Garbowski, E.; Primet, M. Catalytic combustion of methane over bulk and supported LaCrO3 perovskites. J. Chem. Soc. Faraday Trans. 1991, 87, 2493. (51) Alifanti, M.; Auer, R.; Kirchnerova, J.; Thyrion, F.; Grange, P.; Delmona, B. Activity in methane combustion and sensitivity to sulfur poisoning of La1-xCexMn1-yCoyO3 perovskite oxides. Appl. Catal. B: EnViron. 2003, 41, 71. (52) Koponen, M. J.; Vena¨la¨inen, T.; Suvanto, M.; Kallinen, K.; Kinnunen, T. J. J-; Ha¨rko¨nen M.; Pakkanen, T. A. Effect of A-site metal on methane combustion on 2% Pd/AMn1-xFexO3 (A ) Ba, La, Pr; x ) 0.4, 0.6, 1) perovskites. Catal. Lett. 2006, 111, 75. (53) Koponen, M. J.; Vena¨la¨inen, T.; Suvanto, M.; Kallinen, K.; Kinnunen, T. J. J.; Ha¨rko¨nen, M.; Pakkanen, T. A. Methane conversion and SO2 resistance of LaMn1-xFexO3 (x ) 0.4, 0.5, 0.6, 1) perovskite catalysts promoted with palladium. J. Mol. Catal. A: Chem. 2006, 258, 246.

ReceiVed for reView December 22, 2006 ReVised manuscript receiVed June 15, 2007 Accepted June 19, 2007 IE061665Y