- Email: [email protected]
The role of support and promoter on the oxidation of sulfur dioxide using platinum based catalysts
Applied Catalysis A: General 306 (2006) 142–148 www.elsevier.com/locate/apcata
The role of support and promoter on the oxidation of sulfur dioxide using platinum based catalysts Sotirios Koutsopoulos a,b,*, Søren B. Rasmussen a,b, K. Michael Eriksen a,b, Rasmus Fehrmann a,b a
Department of Chemistry, Building 207, Technical University of Denmark, DK-2800 Lyngby, Denmark b Interdisciplinary Research Center for Catalysis (ICAT), DK-2800 Lyngby, Denmark Received 18 November 2005; received in revised form 16 March 2006; accepted 20 March 2006 Available online 27 April 2006
Abstract The catalytic oxidation of SO2 to SO3 was studied over platinum based catalysts in the absence and the presence of dopants. The active metal was supported on silica gel or titania (anatase) by impregnation. The activities of the silica supported catalysts were found to follow the order Pt– Rh/SiO2 > Pt/SiO2 > Pt–Al/SiO2. For the samples supported on titania the respective order was Pt/TiO2 > Pt–Rh/TiO2 > Pt–Al/TiO2. The size of the particles of the active phase, the presence of dopants and the purity of the catalyst were found to affect the catalytic performance. A careful selection of the pH of the impregnation solution and of the reduction temperature of the precursor salts resulted in a very active catalyst with average particle size of 1.7 nm. # 2006 Elsevier B.V. All rights reserved. Keywords: Silica; Titania; Platinum; Palladium; Rhodium; Catalyst; Sulphur dioxide oxidation
1. Introduction Sulfur dioxide emission to the atmosphere is an increasing concern worldwide. In order to meet legislative demands to reduce release of SO2, more efficient catalysts need to be developed. Platinum has been used as a catalyst for the oxidation of SO2 in the industry of sulfuric acid production. Before the Second World War platinum was exclusively used in this process however, the immense requirements of sulfuric acid caused such a demand in platinum that resulted in difficulties in supply. Moreover, cost limitations and deactivation of the catalyst by e.g., arsenic compounds stimulated the invention of alternatives. Nowadays, the industrial use of platinum as a catalyst for the SO2 oxidation – the key step in the sulfuric acid production and in catalytic flue gas desulfurization by the SNOX process – is essentially replaced by catalysts based on vanadium. However, the oxidation of SO2 by
* Corresponding author at: Center for Biomedical Engineering, Massachusetts Institute of Technology, NE47-Room 307, 500 Technology Square, Cambridge, MA 02139-4307, USA. Tel.: +1 617 324 7612; fax: +1 617 258 5239. E-mail address: [email protected] (S. Koutsopoulos). 0926-860X/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2006.03.041
platinum-based catalysts continues to attract interest [1–5]. This is mainly due to the fact that platinum particles can be redistributed and regain high catalytic activity whenever that drops as a result of prolonged use and exposure to O2 [6] or chlorine gas [7]. Rhodium is a hardening agent of platinum alloys that maintain structural and textural characteristics even after long exposure at high temperatures. The contribution of Rh towards hardening of platinum alloys significantly exceeds the properties of any other metal alloyed with platinum. The use of the Pt–Rh alloy as a catalyst for the SO2 oxidation is advantageous over other platinum based catalysts in terms of speed of activation, efficiency of conversion, duration of catalytic performance and stability at high temperatures [8]. Moreover, Rh strongly interacts with O2 molecules and thus increases their surface concentration. This mechanism is responsible for enhancing the rate of the oxidation reaction. The idea of doping Pt with Al originated from electrochemical studies [9]. Although the formation of an alloy between Pt and Al is ambiguous, we will investigate if simple contact between the two metals would result in interesting catalytic properties. Silica and titania were chosen to support the platinum particles. Silica has been reported to give insufficient stabilization of precious metal particles against sintering
S. Koutsopoulos et al. / Applied Catalysis A: General 306 (2006) 142–148
[10], and low dispersion of the active phase [11]. Nevertheless, it has some advantages, such as poisoning resistance against arsenic compounds, high conversion rates at high space velocities and extremely low surface oxygen mobility [12]. Especially, the low surface mobility of oxygen is very important because the catalyst is immunized against prolonged decreased oxygen content in the reacting gases; at such conditions the vanadium based catalysts undergo significant decrease in catalytic activity. Titania was chosen because it is one of most commonly used carriers in a variety of catalytic processes such as flue gas denitrification, hydrocarbon selective oxidation, CO hydrogenation, and NOx reduction with hydrocarbons [13–15]. The surface properties of titania, its poisoning resistance features and the strong metal support interactions, especially when the active phase is Pt, are additional reasons for the wide use of titania as a carrier [16]. In the present work, we investigated the Pt based catalysts supported on silica gel and titania (anatase). Catalyst samples were prepared in the presence of Rh and Al dopants and their activity was tested for the oxidation of SO2. 2. Experimental 2.1. Chemicals and gases Silica gel 100, SiO2, (Merck) and titania (anatase), TiO2, (Alfa Aesar Chemicals, >99.99%) were used as supports. The specific surface area of silica and titania are 290 and 53 m2/g, respectively as measured by the triple point BET method. The average pore size is 100 and 250 nm for silica gel and titania, respectively. Platinum chloride, PtCl4, (H. Drijfhout & Zoon’s, Amsterdam, The Netherlands) was chosen as a precursor salt for the synthesis. Dopant solutions were prepared from the respective extra pure solids, i.e., RhCl33H2O (Aldrich), and AlCl36H2O (Merck) using deionized water. We have chosen the chloride salts of the respective metals because they are very soluble in water. Commercial gases SO2 (>99.99%), O2 (99.8% + 0.2%N2 and Ar) and N2 (<40 ppm O2 + H2O) were used in the catalytic activity tests. All gases were dried through P2O5 columns. 2.2. Catalyst preparation The silica gel contained up to 0.01%Zn, which is suggested to have a poisoning effect on platinum catalysts used in the SO2 oxidation process [17]. To remove Zn, silica gel particles were suspended in 1 M HCl solution and allowed to equilibrate for 2 h at 60 8C under continuous stirring and ultrasonication. Then the particles were filtered and washed through a 0.22 mm filter. The process was repeated three times and finally the particles were dried overnight at 100 8C. After treatment, Zn could not be detected by chemical and EDAX analysis. The Zn-free silica gel was impregnated with the PtCl4 solution aiming at 2 wt.% loading with the active phase. This was confirmed by combined TEM/EDAX and chemical analysis. In the doped samples, the silica gel was suspended and coimpregnated in a solution, which contained the precursor chloride salts of Pt and Rh or Al.
143
The content of the dopant was meant to be 10% of the total metal content of the catalyst. After equilibration for 2 days at 50 8C the solvent was slowly evaporated in a rotary evaporator and the solid material was dried overnight at 110 8C. Then the sample was reduced at 500 8C with formier gas (90%N2, 10%H2). The temperature increased stepwise (i.e., each step was 1 h at 100, 200, 300, and 400 8C, and 3 h at 500 8C). Then, the catalyst was allowed to cool to room temperature under continuous formier gas flow. Two series of silica-supported catalysts were prepared keeping the pH of the impregnation solution above (i.e., pH 7) and below (i.e., pH 1) the point of zero charge (p.z.c.) of silica gel, which is at pH 2.7. The pH was adjusted by adding 0.01 M NaOH or 0.01 M HCl. To test the effect of the reduction temperature on the catalytic activity, the catalyst sample prepared from a solution with pH 1, was divided in two parts that were reduced with formier gas at 500 and 700 8C, respectively. Titania anatase, TiO2, was also impregnated by a PtCl4 solution with pH 6.7, which is slightly higher than the p.z.c. of the surface (i.e., at pH 6–6.5). In the doped samples, the dissolved PtCl4 and one of the RhCl3 or AlCl3 salts were coimpregnated in the titania particles. The final composition of the dopant in the total metal content of the catalyst was 10%. After impregnation the sample was dried overnight and the precursor salts were reduced to the active metal by formier gas at 500 8C (as before the temperature was increased stepwise). 2.3. Catalyst characterization The size of the platinum particles supported on silica or titania was measured by transmission electron microscopy, TEM, (Philips EM 430) using an image analysis software (average of ca. 80 particles). The microscope was calibrated using an internal standard (asbestos lines) to avoid the common source of error, where the display magnification is higher than the real and hence, the particles appear smaller. Thermogravimetric–differential thermal analysis, TGA– DTA, (Mettler TGA cell) and temperature-programmed reduction, TPR, experiments were carried out using formier gas as reducing agent (gas flow 110 ml/min). The set-up was coupled with a mass spectrometer to analyze the composition of the gas outlet. Before analysis all samples were kept overnight under vacuum to remove adsorbed surface water and other molecules. The samples were heated with a rate of 1.5 8C/min and a step at 300 8C for 1 h, before reaching the maximum at 650 8C. X-ray powder diffraction patterns (Philips PW 3710, Cu Ka radiation, scan rate 0.048/min) of the catalyst were recorded to analyze the crystallographic characteristics of the active phase. 2.4. Catalytic activity test The experimental set-up used for the catalytic activity tests consisted of a U-shaped capillary reactor made of Pyrex glass, which was placed in a temperature-controlled furnace. The synthesis gas passed through the catalyst bed, which was located between two plugs of quartz wool. The temperature in the furnace was measured with a thermocouple in contact with
144
S. Koutsopoulos et al. / Applied Catalysis A: General 306 (2006) 142–148
3. Results and discussion 3.1. Catalyst characterization
Fig. 1. UV absorbance of SO2 in the gas outlet at 250 8C (&) and 480 8C (*) after passing through the PtpH > p.z.c.,500 8C/SiO2 catalyst (see Table 1).
the reactor and was accurately controlled with a EUROTHERM regulator. The concentration of SO2 in the gas outlet was monitored spectrophotometrically by measuring the absorbance peak at 300.4 nm, which is the maximum absorption of SO2 in the UV region (Fig. 1). The conversion of SO2 was calculated from the concentration change between inlet and outlet of the reactor. The absorption spectra were recorded by a modified Jasco V-570 UV–vis spectrophotometer equipped with a 0.5 cm quartz cell. From the calibration curves the SO2 partial pressure was correlated with the measured absorbance. Before and after the reactor the gas was dried through P2O5 columns. For the catalytic activity tests a gas mixture with composition 10%O2, 11%SO2 balanced with N2 (i.e., simulated sulfuric acid synthesis gas) was used, under atmospheric pressure. All conversions were measured at constant gas flow of 30 ml/min at ambient temperature and pressure using a mass flow controller (Brooks, 5810). The rate and the concentration of the gases feeding the reactor were kept constant in all tests. No heating device was installed to pre-heat the reacting gases prior to entering the reactor.
From the DTA–TGA analysis, it was concluded that the reduction of the platinum precursor salt to the metal form was completed at approximately 250–270 8C. This was confirmed by X-ray diffraction analysis. However, EDAX and chemical analysis of the catalyst samples showed that chloride ions were present even after reduction at 500 8C. TPR analyses revealed two peaks for HCl; a strong peak at 160 8C and a weak one at 600 8C. Chloride has been suggested to have a poisoning effect on platinum based catalysts, when used in the SO2 oxidation reaction. To remove the chloride ions in one of the preparations we performed the reduction step at 700 8C. TEM image analysis showed that the average size of the platinum particles supported on silica (average of 80 particles) was ca. 5 nm (Fig. 2a). When titania was used as a carrier platinum particles formed crystals with a hexagonal shape and an average size of 11 nm (Fig. 2c). The difference in the particle size between the two preparations was attributed to the differences of the surface characteristics of the two carriers. The average pore size of silica gel is half the size of the pores of titania. Hence, it is reasonable to correlate the particle size of the supported particles with the dimensions of the pores; bigger crystals of the precipitated precursor salt grow in the wider cavities of titania while smaller crystals are limited by the walls of the thinner pores of silica. During reduction the counterions of the precursor salts were removed and the pure metal, which is catalytically active, remained in the pores of the carrier. The Pt– Rh and Pt–Al particles were found to be slightly bigger (i.e., ca. 6 and 13 nm when supported on silica and titania, respectively). It should be noted that in all cases, AFM analyses showed the presence of a significant number of particles as small as 2 nm supported on both carriers. For the silica-supported catalysts the surface distribution of the Pt nanoparticles was poor when the preparation was made in a solution with pH above the p.z.c. of silica. In this case, the surface and platinum ions of the precursor salt were oppositely charged. At conditions of electrostatic repulsion between
Fig. 2. TEM pictures of platinum catalysts supported on (a) silica gel (impregnation at pH > p.z.c. and reduction at 700 8C, i.e., Pt/SiO2 in Table 1), (b) silica gel (after impregnation at pH < p.z.c. and reduction at 500 8C, i.e. PtpH < p.z.c.,500 8C/SiO2 in Table 1), and (c) titania.
S. Koutsopoulos et al. / Applied Catalysis A: General 306 (2006) 142–148
145
Table 1 Catalytic activities of the platinum based catalysts supported on silica gel and titania (anatase) Sample
Reactor loading (mg)
Particle size (nm)
Maximum activity (%Conversion)
Turnover rate ðmolSO2 ;converted =ðmolPt sÞÞ
Temperature of maximum activity (8C)
Space velocity (h1)
Pt/SiO2 Pt–Rh/SiO2 Pt–Al/SiO2 Pt/TiO2 Pt–Rh/TiO2 Pt–Al/TiO2 PtpH > p.z.c./SiO2 PtpH < p.z.c./SiO2 Pt/SiO2 (3 mm catalyst bed) Pt/TiO2 (3 mm catalyst bed) Pt/SiO2 (1 cm catalyst bed) Pt/TiO2 (1 cm catalyst bed)
1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 3.8 4.8 21.3
5.1 5.4 6.0 11.0 12.5 12.9 4.2 1.7 5.1 11.0 5.1 11.0
43 50 43 46 43 39 50 61 43 59 71 80
9.8 11.0 9.6 10.2 9.6 8.6 11.2 13.6 9.8 2.5 2.4 0.6
550 510 510 510 510 550 550 550 550 500 520 480
590000 510000 570000 2600000 2500000 2500000 560000 560000 590000 690000 120000 130000
Feed gas 11%SO2, 10%O2 and 79%N2, flow rate 30 ml/min, catalyst loading 2 wt.%.
surface and platinum ions (i.e., when the pH of the precipitating solution was below the p.z.c. of silica) the surface distribution of the active phase was very good and moreover, the size of the nanoparticles very small ca. 1.7 0.4 nm (Fig. 2b). The formation of bimetallic alloys could not be detected by X-ray diffraction analysis. The XRD patterns of the alloyed catalysts resembled those of the pure Pt catalyst (JCPDS card file 04-0802). This is often the case where the dopant concentration in the alloy is low and when the active phase consists of nanoparticles [18,19]. Peaks assigned to pure Rh or Al (JCPDS 05-0685 and 04-0787) could not be detected in the diffractograms. In the case of particles supported on titania (titanium dioxide, anatase, JCPDS card file 83-2243) the spectra were dominated by the titania diffraction peaks which often overlapped with the weak peaks of the active phase. Peaks assigned to the precursor salts were present the diffraction pattern before reduction but they were completely absent after heat-treatment with formier gas indicating formation of the pure metals.
in Table 1. Maximum activities are also reported as turnover rates (i.e., mol of SO2 converted per mol of Pt in the reactor and per second). Silica gel supported catalysts gave higher conversion rates than those supported on titania. The latter operated at very high space velocities (e.g., 2,500,000 h1), which render them more attractive in terms of industrial selection. Such a high space velocity means that the necessary contact time for conversion of the SO2 molecule over the catalyst surface is extremely small (i.e., 0.0014 s). Activity tests performed at different loading of the reactor showed that the catalytic reaction of SO2 depends not only on the temperature but also on the mass of catalyst. Experiments performed using larger amount of catalyst resulted in higher SO2 conversion and in lower temperature of maximum activity [20]. The conversion of SO2 to SO3 as a function of the operating temperature of the catalyst is shown in Figs. 3–7. The oxidation profile of SO2 as a function of temperature shows the typical
3.2. SO2 oxidation catalysis A typical UV-spectrophotometer output representing the SO2 depletion from the synthesis gas over the catalyst is shown in Fig. 1. The conversion of SO2 was calculated from the difference between the moles of SO2, nSO2 , before and after passing through the catalyst bed in unit time, using the equation: nSO2 ;before nSO2 ;after 100 (1) ð%ConversionÞ ¼ nSO2 ;before which can be transformed to: ð%ConversionÞ ¼
ð1 PSO2 ;after Þ=PSO2 ;before 100 ð1 PSO2 ;after Þ=2PT
(2)
where, PSO2 is the partial pressure of SO2 in the reactant gas mixture before and after the reactor and PT is the total pressure over the set-up equal to 1 atm. The results from the catalytic activity tests and the experimental conditions are summarized
Fig. 3. Conversion of SO2 over 1.0 mg of platinum based catalysts supported on silica gel: (&) Pt/SiO2, (*) Pt–Rh/SiO2, and (^) Pt–Al/SiO2. Space velocities are shown in Table 1. The broken line represents the calculated equilibrium conversion for the oxidation of SO2 to SO3 with O2, (11%SO2, 10%O2, balance N2, total pressure 1 atm). The reproducibility of the catalytic activity measurements was better than 10%.
146
S. Koutsopoulos et al. / Applied Catalysis A: General 306 (2006) 142–148
Fig. 6. Long-term experiment of the Pt–Al/TiO2 catalyst at 400 8C. Fig. 4. Conversion of SO2 over 1.0 mg of platinum based catalysts supported on titania: (&) Pt/TiO2, (*) Pt–Rh/TiO2, and (^) Pt–Al/TiO2. Space velocities are shown in Table 1. The broken line represents the calculated equilibrium conversion for the oxidation of SO2 to SO3 (11%SO2, 10%O2, balance N2, total pressure 1 atm). The reproducibility of the catalytic activity measurements was better than 10%.
pattern in which the activity gradually increased by increasing the temperature. After reaching a maximum the activity decreased as a result of the reaction approaching equilibrium conversion. A temperature gradient in the reactor, which is an important implication for exothermic reactions, is not a significant reason for the decreased activity observed here because of the capillary dimensions of the reactor and the relatively high flow rate of the reactant gases. In a subsequent step the catalytic performance was monitored upon decreasing the temperature. From these measurements it was revealed that the activity decreased 5–20% as compared to the up-scan measurements, depending on the composition of the active phase and on the carrier. This may be attributed to sintering of the particles of the active phase as a result of operating the catalyst at high temperatures. Sintering leads to irreversible changes in the structure and the shape of the particles. It is
possible that at high temperatures (e.g., 600 8C) some of the platinum atoms on the surface of the nanoparticles are oxidized by the oxygen molecules, which are present in the gas mixture and therefore become catalytically inactive. The maximum catalytic activity depends on the dopant. The aluminum doped platinum catalyst supported on silica gel showed maximum activity at 460 8C. This is slightly lower than the respective maximum of the catalyst promoted with rhodium (i.e., at 480 8C) and that of the undoped pure platinum catalyst (i.e., at 500 8C), (Fig. 3). The titania supported catalysts did not show the same behavior (Fig. 4) and the Pt–Al/ TiO2 catalyst showed maximum activity at significantly higher temperatures (at 550 8C). For the silica supported catalysts the temperature of maximum conversion follows the order: TPt=SiO2 > TPtRh=SiO2 > TPtAl=SiO2 while for the titania supported catalysts the respective temperatures follow the reverse order: TPt=TiO2 < TPtRh=TiO2 < TPtAl=TiO2 : Further work is required to interpret these findings. The electronic properties of the carrier, the type of the supported metal and of the alloys and the still uncertain mechanism of SO2 oxidation over platinum based catalysts need to be considered. Moreover, the structural characteristics and the physicochemical properties
Fig. 5. Conversion of SO2 over platinum catalysts supported on titania as a function of the reactor loading: (&) 1.0 mg (*) 3.8 mg, and (~) 21.3 mg of Pt/ TiO2. The broken line represents the calculated equilibrium conversion for the oxidation of SO2 to SO3 (11%SO2, 10%O2, balance N2, total pressure 1 atm). The reproducibility of the catalytic activity measurements was better than 10%.
Fig. 7. Effect of pH during impregnation in the catalytic activity of platinum catalysts supported on silica gel (Pt/SiO2): (^) pHimpr > p.z.c. Tred 500 8C (*) pHimpr > p.z.c. Tred 700 8C, and (&) pHimpr < p.z.c. Tred 500 8C. The broken line represents the calculated equilibrium conversion for the oxidation of SO2 to SO3 (11%SO2, 10%O2, balance N2, total pressure 1 atm). The reproducibility of the activity measurements was better than 10%.
S. Koutsopoulos et al. / Applied Catalysis A: General 306 (2006) 142–148
of the active phase consisting of nanoparticles are expected to play a role in the performance of the catalyst. The titania supported catalysts have lower mass densities as compared to those supported on silica and therefore, the catalyst bed in the reactor was longer. Activity tests were performed at conditions where the reactor was loaded with such amount of catalyst supported on silica or titania that resulted in equally long catalyst beds. For a 3-mm long catalyst bed in the reactor 1 mg of silica gel and 3.8 mg of the titaniasupported Pt catalysts were required. In this case, the SO2 conversion was 43 and 59%, respectively; these values represent turnover rates of 9.8 and 2.5 mol of SO2 converted per mol of Pt and per second. For a catalyst bed of 1-cm length 4.8 and 21.3 mg of the silica and titania supported catalysts, respectively were required. At these conditions the SO2 conversion increased to 71 and 80% (Fig. 5). Interesting features were revealed from the long-term activity tests. As may be seen in Fig. 6, the activity of the aluminum-promoted catalyst decreased ca. 40% in 4.2 days. The test was performed at 420 8C, which is lower than the temperature of maximum activity (Fig. 4). The activity drop coincides with the activity of the undoped Pt/TiO2 catalyst at the same temperature. This suggests that the Pt–Al interaction is interrupted as a function of time probably due to formation of aluminum sulfates. Aluminum forms sulfate salts in the presence of SO2 and for this reason alumina, Al2O3, is not commonly used as a carrier for the SO2 oxidation process. The effect of pH during impregnation and of the reduction temperature of the precursor salts on the dispersion of the active phase and catalytic activity were investigated for platinum catalysts supported on silica gel. When the impregnation was done in an acidic solution at pH 1, which is below the p.z.c. of silica surface, the resulting particles of the active phase were remarkably smaller and the dispersion significantly better as compared to that of the samples prepared when the pH of the impregnating solution was higher than the p.z.c. of silica. Electrostatic repulsions between the like-charged surface of silica and Pt4+ ions resulted in good dispersion of the crystallized particles of the precursor phase (sample PtpH p.z.c.silica and reduced at 500 8C (average size of Pt particles is ca. 5 nm), respectively. Hence, removal of chloride ions that are suspected to poison the catalyst does not significantly increase the activity. Instead, the size of the particles seems to be a more critical factor to increase the catalytic activity.
147
The oxidation of SO2 is exothermic meaning that at high temperatures the equilibrium is shifted to low conversion. For the composition of the synthesis gas mixture used in the present work, the equilibrium is 99.98% to the side of the oxidized products when the temperature is 270 8C. At 400 8C the conversion decrease to 99.1% and drops rapidly from 95.6 to 83.8% in the temperature range from 470 to 550 8C (where most of the activity maximums were observed for the catalyst tested). The conversion of SO2 was low at 270 8C. However, for the very low loading of the reactor and of the catalyst with the active phase the activity is high when the temperature increased to 400 8C. Since the SO2 oxidation reaction is exothermal, when the temperature exceeds 550 8C the equilibrium is significantly shifted to the side of the reactants. At such high temperatures the thermodynamic factors prevail over the kinetic. The thermodynamics of the system and sintering of the particles of the active phase are reasons for the decreased catalytic activity at high temperatures. 4. Conclusions The catalytic oxidation of SO2 over platinum based catalysts takes place in the temperature range between 250 and 600 8C. At higher temperatures the catalytic activity drops as a result of sintering of the platinum particles and thermodynamic limitations. Physical parameters such as the pore size of the carrier and the molecular volume of the impregnated precursor salt influence the characteristics of the produced catalyst and the catalytic activity. Silica gel gave good particle distribution when impregnated in solution of pH below the p.z.c. of the surface. Furthermore, the particle size of the active phase is critical and more important than the poisoning effect of traces of chloride ions. Titania supported samples exhibit better catalytic properties as suggested from the high space velocities obtained even if the particle size of the active phase was twice as big as those supported on silica gel. This confirms the importance of the properties of the carrier in the catalystassisted SO2 oxidation process. Acknowledgements The NATO, Science for Peace Program, (SfP 971984) has supported this work. We also thank two of the reviewers for insightful suggestions and useful comments. References [1] Y. Nagai, H. Shinjoh, K. Yokota, Appl. Catal. B 39 (2002) 149. [2] E. Xue, K. Seshan, J.R.H. Ross, Appl. Catal. B 11 (1996) 65. [3] L.G. Simonova, V.V. Barelko, A.V. Toktarev, A.F. Chernyshov, V.A. Chumachenko, B.S. Bal’zhinimaev, Kinet. Catal. 43 (2002) 61. [4] M. Deeba, R.J. Farrauto, Y.K. Lui, Appl. Catal. A 124 (1995) 339. [5] D. Uy, A. Dubkov, G.W. Graham, W.H. Weber, Catal. Lett. 68 (2000) 25. [6] B.C. Gates, Catalytic Chemistry, John Wiley and Sons, Inc., New York, 1992, p. 401. [7] J.M. Parera, N.S. Fagoli, in: J.J. Spivey (Ed.), Catalysis, vol. 9, Royal Society of Chemistry, 1992, pp. 65–125. [8] D. McDonald, Chem. Ind. (1931) 1031. [9] J. Lee, S.H. Langer, J. Appl. Electrochem. 25 (1995) 353.
148
S. Koutsopoulos et al. / Applied Catalysis A: General 306 (2006) 142–148
[10] R. Lamber, N. Jaeger, G. Schultz-Ekloff, J. Catal. 23 (1990) 285. [11] C.R. Adams, H.A. Benesi, R.M. Curtis, R.G. Meisenheimer, J. Catal. 1 (1962) 336. [12] D. Martin, D. Duprez, Stud. Surf. Sci. Catal. 77 (1993) 201. [13] G. Ramis, G. Busca, V. Lorenzelli, P. Forzatti, Appl. Catal. 64 (1990) 243. [14] M.A. Vannice, J. Catal. 74 (1982) 199. [15] A. Bahamonde, C. Knapp, P. Avila, J. Blanco, Appl. Catal. B 19 (1998) 1.
[16] P. Meriaudeau, O.H. Ellenstad, M. Dufaux, C. Naccache, J. Catal. 75 (1982) 243. [17] E.B. Maxted, A. Marsden, J. Chem. Soc. (1940) 469. [18] P.K. Gallagher, M.E. Gross, J. Therm. Anal. 31 (1986) 1231. [19] H.P. Klugg, L.E. Alexander, X-Ray Diffraction Procedures, John Wiley and Sons, Inc., New York, 1970. [20] H.N. Holmes, J. Ramsay, A.L. Elder, Ind. Eng. Chem. 21 (1929) 850.
The role of support and promoter on the oxidation of sulfur dioxide using platinum based catalysts Sotirios Koutsopoulos a,b,*, Søren B. Rasmussen a,b, K. Michael Eriksen a,b, Rasmus Fehrmann a,b a
Department of Chemistry, Building 207, Technical University of Denmark, DK-2800 Lyngby, Denmark b Interdisciplinary Research Center for Catalysis (ICAT), DK-2800 Lyngby, Denmark Received 18 November 2005; received in revised form 16 March 2006; accepted 20 March 2006 Available online 27 April 2006
Abstract The catalytic oxidation of SO2 to SO3 was studied over platinum based catalysts in the absence and the presence of dopants. The active metal was supported on silica gel or titania (anatase) by impregnation. The activities of the silica supported catalysts were found to follow the order Pt– Rh/SiO2 > Pt/SiO2 > Pt–Al/SiO2. For the samples supported on titania the respective order was Pt/TiO2 > Pt–Rh/TiO2 > Pt–Al/TiO2. The size of the particles of the active phase, the presence of dopants and the purity of the catalyst were found to affect the catalytic performance. A careful selection of the pH of the impregnation solution and of the reduction temperature of the precursor salts resulted in a very active catalyst with average particle size of 1.7 nm. # 2006 Elsevier B.V. All rights reserved. Keywords: Silica; Titania; Platinum; Palladium; Rhodium; Catalyst; Sulphur dioxide oxidation
1. Introduction Sulfur dioxide emission to the atmosphere is an increasing concern worldwide. In order to meet legislative demands to reduce release of SO2, more efficient catalysts need to be developed. Platinum has been used as a catalyst for the oxidation of SO2 in the industry of sulfuric acid production. Before the Second World War platinum was exclusively used in this process however, the immense requirements of sulfuric acid caused such a demand in platinum that resulted in difficulties in supply. Moreover, cost limitations and deactivation of the catalyst by e.g., arsenic compounds stimulated the invention of alternatives. Nowadays, the industrial use of platinum as a catalyst for the SO2 oxidation – the key step in the sulfuric acid production and in catalytic flue gas desulfurization by the SNOX process – is essentially replaced by catalysts based on vanadium. However, the oxidation of SO2 by
* Corresponding author at: Center for Biomedical Engineering, Massachusetts Institute of Technology, NE47-Room 307, 500 Technology Square, Cambridge, MA 02139-4307, USA. Tel.: +1 617 324 7612; fax: +1 617 258 5239. E-mail address: [email protected] (S. Koutsopoulos). 0926-860X/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2006.03.041
platinum-based catalysts continues to attract interest [1–5]. This is mainly due to the fact that platinum particles can be redistributed and regain high catalytic activity whenever that drops as a result of prolonged use and exposure to O2 [6] or chlorine gas [7]. Rhodium is a hardening agent of platinum alloys that maintain structural and textural characteristics even after long exposure at high temperatures. The contribution of Rh towards hardening of platinum alloys significantly exceeds the properties of any other metal alloyed with platinum. The use of the Pt–Rh alloy as a catalyst for the SO2 oxidation is advantageous over other platinum based catalysts in terms of speed of activation, efficiency of conversion, duration of catalytic performance and stability at high temperatures [8]. Moreover, Rh strongly interacts with O2 molecules and thus increases their surface concentration. This mechanism is responsible for enhancing the rate of the oxidation reaction. The idea of doping Pt with Al originated from electrochemical studies [9]. Although the formation of an alloy between Pt and Al is ambiguous, we will investigate if simple contact between the two metals would result in interesting catalytic properties. Silica and titania were chosen to support the platinum particles. Silica has been reported to give insufficient stabilization of precious metal particles against sintering
S. Koutsopoulos et al. / Applied Catalysis A: General 306 (2006) 142–148
[10], and low dispersion of the active phase [11]. Nevertheless, it has some advantages, such as poisoning resistance against arsenic compounds, high conversion rates at high space velocities and extremely low surface oxygen mobility [12]. Especially, the low surface mobility of oxygen is very important because the catalyst is immunized against prolonged decreased oxygen content in the reacting gases; at such conditions the vanadium based catalysts undergo significant decrease in catalytic activity. Titania was chosen because it is one of most commonly used carriers in a variety of catalytic processes such as flue gas denitrification, hydrocarbon selective oxidation, CO hydrogenation, and NOx reduction with hydrocarbons [13–15]. The surface properties of titania, its poisoning resistance features and the strong metal support interactions, especially when the active phase is Pt, are additional reasons for the wide use of titania as a carrier [16]. In the present work, we investigated the Pt based catalysts supported on silica gel and titania (anatase). Catalyst samples were prepared in the presence of Rh and Al dopants and their activity was tested for the oxidation of SO2. 2. Experimental 2.1. Chemicals and gases Silica gel 100, SiO2, (Merck) and titania (anatase), TiO2, (Alfa Aesar Chemicals, >99.99%) were used as supports. The specific surface area of silica and titania are 290 and 53 m2/g, respectively as measured by the triple point BET method. The average pore size is 100 and 250 nm for silica gel and titania, respectively. Platinum chloride, PtCl4, (H. Drijfhout & Zoon’s, Amsterdam, The Netherlands) was chosen as a precursor salt for the synthesis. Dopant solutions were prepared from the respective extra pure solids, i.e., RhCl33H2O (Aldrich), and AlCl36H2O (Merck) using deionized water. We have chosen the chloride salts of the respective metals because they are very soluble in water. Commercial gases SO2 (>99.99%), O2 (99.8% + 0.2%N2 and Ar) and N2 (<40 ppm O2 + H2O) were used in the catalytic activity tests. All gases were dried through P2O5 columns. 2.2. Catalyst preparation The silica gel contained up to 0.01%Zn, which is suggested to have a poisoning effect on platinum catalysts used in the SO2 oxidation process [17]. To remove Zn, silica gel particles were suspended in 1 M HCl solution and allowed to equilibrate for 2 h at 60 8C under continuous stirring and ultrasonication. Then the particles were filtered and washed through a 0.22 mm filter. The process was repeated three times and finally the particles were dried overnight at 100 8C. After treatment, Zn could not be detected by chemical and EDAX analysis. The Zn-free silica gel was impregnated with the PtCl4 solution aiming at 2 wt.% loading with the active phase. This was confirmed by combined TEM/EDAX and chemical analysis. In the doped samples, the silica gel was suspended and coimpregnated in a solution, which contained the precursor chloride salts of Pt and Rh or Al.
143
The content of the dopant was meant to be 10% of the total metal content of the catalyst. After equilibration for 2 days at 50 8C the solvent was slowly evaporated in a rotary evaporator and the solid material was dried overnight at 110 8C. Then the sample was reduced at 500 8C with formier gas (90%N2, 10%H2). The temperature increased stepwise (i.e., each step was 1 h at 100, 200, 300, and 400 8C, and 3 h at 500 8C). Then, the catalyst was allowed to cool to room temperature under continuous formier gas flow. Two series of silica-supported catalysts were prepared keeping the pH of the impregnation solution above (i.e., pH 7) and below (i.e., pH 1) the point of zero charge (p.z.c.) of silica gel, which is at pH 2.7. The pH was adjusted by adding 0.01 M NaOH or 0.01 M HCl. To test the effect of the reduction temperature on the catalytic activity, the catalyst sample prepared from a solution with pH 1, was divided in two parts that were reduced with formier gas at 500 and 700 8C, respectively. Titania anatase, TiO2, was also impregnated by a PtCl4 solution with pH 6.7, which is slightly higher than the p.z.c. of the surface (i.e., at pH 6–6.5). In the doped samples, the dissolved PtCl4 and one of the RhCl3 or AlCl3 salts were coimpregnated in the titania particles. The final composition of the dopant in the total metal content of the catalyst was 10%. After impregnation the sample was dried overnight and the precursor salts were reduced to the active metal by formier gas at 500 8C (as before the temperature was increased stepwise). 2.3. Catalyst characterization The size of the platinum particles supported on silica or titania was measured by transmission electron microscopy, TEM, (Philips EM 430) using an image analysis software (average of ca. 80 particles). The microscope was calibrated using an internal standard (asbestos lines) to avoid the common source of error, where the display magnification is higher than the real and hence, the particles appear smaller. Thermogravimetric–differential thermal analysis, TGA– DTA, (Mettler TGA cell) and temperature-programmed reduction, TPR, experiments were carried out using formier gas as reducing agent (gas flow 110 ml/min). The set-up was coupled with a mass spectrometer to analyze the composition of the gas outlet. Before analysis all samples were kept overnight under vacuum to remove adsorbed surface water and other molecules. The samples were heated with a rate of 1.5 8C/min and a step at 300 8C for 1 h, before reaching the maximum at 650 8C. X-ray powder diffraction patterns (Philips PW 3710, Cu Ka radiation, scan rate 0.048/min) of the catalyst were recorded to analyze the crystallographic characteristics of the active phase. 2.4. Catalytic activity test The experimental set-up used for the catalytic activity tests consisted of a U-shaped capillary reactor made of Pyrex glass, which was placed in a temperature-controlled furnace. The synthesis gas passed through the catalyst bed, which was located between two plugs of quartz wool. The temperature in the furnace was measured with a thermocouple in contact with
144
S. Koutsopoulos et al. / Applied Catalysis A: General 306 (2006) 142–148
3. Results and discussion 3.1. Catalyst characterization
Fig. 1. UV absorbance of SO2 in the gas outlet at 250 8C (&) and 480 8C (*) after passing through the PtpH > p.z.c.,500 8C/SiO2 catalyst (see Table 1).
the reactor and was accurately controlled with a EUROTHERM regulator. The concentration of SO2 in the gas outlet was monitored spectrophotometrically by measuring the absorbance peak at 300.4 nm, which is the maximum absorption of SO2 in the UV region (Fig. 1). The conversion of SO2 was calculated from the concentration change between inlet and outlet of the reactor. The absorption spectra were recorded by a modified Jasco V-570 UV–vis spectrophotometer equipped with a 0.5 cm quartz cell. From the calibration curves the SO2 partial pressure was correlated with the measured absorbance. Before and after the reactor the gas was dried through P2O5 columns. For the catalytic activity tests a gas mixture with composition 10%O2, 11%SO2 balanced with N2 (i.e., simulated sulfuric acid synthesis gas) was used, under atmospheric pressure. All conversions were measured at constant gas flow of 30 ml/min at ambient temperature and pressure using a mass flow controller (Brooks, 5810). The rate and the concentration of the gases feeding the reactor were kept constant in all tests. No heating device was installed to pre-heat the reacting gases prior to entering the reactor.
From the DTA–TGA analysis, it was concluded that the reduction of the platinum precursor salt to the metal form was completed at approximately 250–270 8C. This was confirmed by X-ray diffraction analysis. However, EDAX and chemical analysis of the catalyst samples showed that chloride ions were present even after reduction at 500 8C. TPR analyses revealed two peaks for HCl; a strong peak at 160 8C and a weak one at 600 8C. Chloride has been suggested to have a poisoning effect on platinum based catalysts, when used in the SO2 oxidation reaction. To remove the chloride ions in one of the preparations we performed the reduction step at 700 8C. TEM image analysis showed that the average size of the platinum particles supported on silica (average of 80 particles) was ca. 5 nm (Fig. 2a). When titania was used as a carrier platinum particles formed crystals with a hexagonal shape and an average size of 11 nm (Fig. 2c). The difference in the particle size between the two preparations was attributed to the differences of the surface characteristics of the two carriers. The average pore size of silica gel is half the size of the pores of titania. Hence, it is reasonable to correlate the particle size of the supported particles with the dimensions of the pores; bigger crystals of the precipitated precursor salt grow in the wider cavities of titania while smaller crystals are limited by the walls of the thinner pores of silica. During reduction the counterions of the precursor salts were removed and the pure metal, which is catalytically active, remained in the pores of the carrier. The Pt– Rh and Pt–Al particles were found to be slightly bigger (i.e., ca. 6 and 13 nm when supported on silica and titania, respectively). It should be noted that in all cases, AFM analyses showed the presence of a significant number of particles as small as 2 nm supported on both carriers. For the silica-supported catalysts the surface distribution of the Pt nanoparticles was poor when the preparation was made in a solution with pH above the p.z.c. of silica. In this case, the surface and platinum ions of the precursor salt were oppositely charged. At conditions of electrostatic repulsion between
Fig. 2. TEM pictures of platinum catalysts supported on (a) silica gel (impregnation at pH > p.z.c. and reduction at 700 8C, i.e., Pt/SiO2 in Table 1), (b) silica gel (after impregnation at pH < p.z.c. and reduction at 500 8C, i.e. PtpH < p.z.c.,500 8C/SiO2 in Table 1), and (c) titania.
S. Koutsopoulos et al. / Applied Catalysis A: General 306 (2006) 142–148
145
Table 1 Catalytic activities of the platinum based catalysts supported on silica gel and titania (anatase) Sample
Reactor loading (mg)
Particle size (nm)
Maximum activity (%Conversion)
Turnover rate ðmolSO2 ;converted =ðmolPt sÞÞ
Temperature of maximum activity (8C)
Space velocity (h1)
Pt/SiO2 Pt–Rh/SiO2 Pt–Al/SiO2 Pt/TiO2 Pt–Rh/TiO2 Pt–Al/TiO2 PtpH > p.z.c./SiO2 PtpH < p.z.c./SiO2 Pt/SiO2 (3 mm catalyst bed) Pt/TiO2 (3 mm catalyst bed) Pt/SiO2 (1 cm catalyst bed) Pt/TiO2 (1 cm catalyst bed)
1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 3.8 4.8 21.3
5.1 5.4 6.0 11.0 12.5 12.9 4.2 1.7 5.1 11.0 5.1 11.0
43 50 43 46 43 39 50 61 43 59 71 80
9.8 11.0 9.6 10.2 9.6 8.6 11.2 13.6 9.8 2.5 2.4 0.6
550 510 510 510 510 550 550 550 550 500 520 480
590000 510000 570000 2600000 2500000 2500000 560000 560000 590000 690000 120000 130000
Feed gas 11%SO2, 10%O2 and 79%N2, flow rate 30 ml/min, catalyst loading 2 wt.%.
surface and platinum ions (i.e., when the pH of the precipitating solution was below the p.z.c. of silica) the surface distribution of the active phase was very good and moreover, the size of the nanoparticles very small ca. 1.7 0.4 nm (Fig. 2b). The formation of bimetallic alloys could not be detected by X-ray diffraction analysis. The XRD patterns of the alloyed catalysts resembled those of the pure Pt catalyst (JCPDS card file 04-0802). This is often the case where the dopant concentration in the alloy is low and when the active phase consists of nanoparticles [18,19]. Peaks assigned to pure Rh or Al (JCPDS 05-0685 and 04-0787) could not be detected in the diffractograms. In the case of particles supported on titania (titanium dioxide, anatase, JCPDS card file 83-2243) the spectra were dominated by the titania diffraction peaks which often overlapped with the weak peaks of the active phase. Peaks assigned to the precursor salts were present the diffraction pattern before reduction but they were completely absent after heat-treatment with formier gas indicating formation of the pure metals.
in Table 1. Maximum activities are also reported as turnover rates (i.e., mol of SO2 converted per mol of Pt in the reactor and per second). Silica gel supported catalysts gave higher conversion rates than those supported on titania. The latter operated at very high space velocities (e.g., 2,500,000 h1), which render them more attractive in terms of industrial selection. Such a high space velocity means that the necessary contact time for conversion of the SO2 molecule over the catalyst surface is extremely small (i.e., 0.0014 s). Activity tests performed at different loading of the reactor showed that the catalytic reaction of SO2 depends not only on the temperature but also on the mass of catalyst. Experiments performed using larger amount of catalyst resulted in higher SO2 conversion and in lower temperature of maximum activity [20]. The conversion of SO2 to SO3 as a function of the operating temperature of the catalyst is shown in Figs. 3–7. The oxidation profile of SO2 as a function of temperature shows the typical
3.2. SO2 oxidation catalysis A typical UV-spectrophotometer output representing the SO2 depletion from the synthesis gas over the catalyst is shown in Fig. 1. The conversion of SO2 was calculated from the difference between the moles of SO2, nSO2 , before and after passing through the catalyst bed in unit time, using the equation: nSO2 ;before nSO2 ;after 100 (1) ð%ConversionÞ ¼ nSO2 ;before which can be transformed to: ð%ConversionÞ ¼
ð1 PSO2 ;after Þ=PSO2 ;before 100 ð1 PSO2 ;after Þ=2PT
(2)
where, PSO2 is the partial pressure of SO2 in the reactant gas mixture before and after the reactor and PT is the total pressure over the set-up equal to 1 atm. The results from the catalytic activity tests and the experimental conditions are summarized
Fig. 3. Conversion of SO2 over 1.0 mg of platinum based catalysts supported on silica gel: (&) Pt/SiO2, (*) Pt–Rh/SiO2, and (^) Pt–Al/SiO2. Space velocities are shown in Table 1. The broken line represents the calculated equilibrium conversion for the oxidation of SO2 to SO3 with O2, (11%SO2, 10%O2, balance N2, total pressure 1 atm). The reproducibility of the catalytic activity measurements was better than 10%.
146
S. Koutsopoulos et al. / Applied Catalysis A: General 306 (2006) 142–148
Fig. 6. Long-term experiment of the Pt–Al/TiO2 catalyst at 400 8C. Fig. 4. Conversion of SO2 over 1.0 mg of platinum based catalysts supported on titania: (&) Pt/TiO2, (*) Pt–Rh/TiO2, and (^) Pt–Al/TiO2. Space velocities are shown in Table 1. The broken line represents the calculated equilibrium conversion for the oxidation of SO2 to SO3 (11%SO2, 10%O2, balance N2, total pressure 1 atm). The reproducibility of the catalytic activity measurements was better than 10%.
pattern in which the activity gradually increased by increasing the temperature. After reaching a maximum the activity decreased as a result of the reaction approaching equilibrium conversion. A temperature gradient in the reactor, which is an important implication for exothermic reactions, is not a significant reason for the decreased activity observed here because of the capillary dimensions of the reactor and the relatively high flow rate of the reactant gases. In a subsequent step the catalytic performance was monitored upon decreasing the temperature. From these measurements it was revealed that the activity decreased 5–20% as compared to the up-scan measurements, depending on the composition of the active phase and on the carrier. This may be attributed to sintering of the particles of the active phase as a result of operating the catalyst at high temperatures. Sintering leads to irreversible changes in the structure and the shape of the particles. It is
possible that at high temperatures (e.g., 600 8C) some of the platinum atoms on the surface of the nanoparticles are oxidized by the oxygen molecules, which are present in the gas mixture and therefore become catalytically inactive. The maximum catalytic activity depends on the dopant. The aluminum doped platinum catalyst supported on silica gel showed maximum activity at 460 8C. This is slightly lower than the respective maximum of the catalyst promoted with rhodium (i.e., at 480 8C) and that of the undoped pure platinum catalyst (i.e., at 500 8C), (Fig. 3). The titania supported catalysts did not show the same behavior (Fig. 4) and the Pt–Al/ TiO2 catalyst showed maximum activity at significantly higher temperatures (at 550 8C). For the silica supported catalysts the temperature of maximum conversion follows the order: TPt=SiO2 > TPtRh=SiO2 > TPtAl=SiO2 while for the titania supported catalysts the respective temperatures follow the reverse order: TPt=TiO2 < TPtRh=TiO2 < TPtAl=TiO2 : Further work is required to interpret these findings. The electronic properties of the carrier, the type of the supported metal and of the alloys and the still uncertain mechanism of SO2 oxidation over platinum based catalysts need to be considered. Moreover, the structural characteristics and the physicochemical properties
Fig. 5. Conversion of SO2 over platinum catalysts supported on titania as a function of the reactor loading: (&) 1.0 mg (*) 3.8 mg, and (~) 21.3 mg of Pt/ TiO2. The broken line represents the calculated equilibrium conversion for the oxidation of SO2 to SO3 (11%SO2, 10%O2, balance N2, total pressure 1 atm). The reproducibility of the catalytic activity measurements was better than 10%.
Fig. 7. Effect of pH during impregnation in the catalytic activity of platinum catalysts supported on silica gel (Pt/SiO2): (^) pHimpr > p.z.c. Tred 500 8C (*) pHimpr > p.z.c. Tred 700 8C, and (&) pHimpr < p.z.c. Tred 500 8C. The broken line represents the calculated equilibrium conversion for the oxidation of SO2 to SO3 (11%SO2, 10%O2, balance N2, total pressure 1 atm). The reproducibility of the activity measurements was better than 10%.
S. Koutsopoulos et al. / Applied Catalysis A: General 306 (2006) 142–148
of the active phase consisting of nanoparticles are expected to play a role in the performance of the catalyst. The titania supported catalysts have lower mass densities as compared to those supported on silica and therefore, the catalyst bed in the reactor was longer. Activity tests were performed at conditions where the reactor was loaded with such amount of catalyst supported on silica or titania that resulted in equally long catalyst beds. For a 3-mm long catalyst bed in the reactor 1 mg of silica gel and 3.8 mg of the titaniasupported Pt catalysts were required. In this case, the SO2 conversion was 43 and 59%, respectively; these values represent turnover rates of 9.8 and 2.5 mol of SO2 converted per mol of Pt and per second. For a catalyst bed of 1-cm length 4.8 and 21.3 mg of the silica and titania supported catalysts, respectively were required. At these conditions the SO2 conversion increased to 71 and 80% (Fig. 5). Interesting features were revealed from the long-term activity tests. As may be seen in Fig. 6, the activity of the aluminum-promoted catalyst decreased ca. 40% in 4.2 days. The test was performed at 420 8C, which is lower than the temperature of maximum activity (Fig. 4). The activity drop coincides with the activity of the undoped Pt/TiO2 catalyst at the same temperature. This suggests that the Pt–Al interaction is interrupted as a function of time probably due to formation of aluminum sulfates. Aluminum forms sulfate salts in the presence of SO2 and for this reason alumina, Al2O3, is not commonly used as a carrier for the SO2 oxidation process. The effect of pH during impregnation and of the reduction temperature of the precursor salts on the dispersion of the active phase and catalytic activity were investigated for platinum catalysts supported on silica gel. When the impregnation was done in an acidic solution at pH 1, which is below the p.z.c. of silica surface, the resulting particles of the active phase were remarkably smaller and the dispersion significantly better as compared to that of the samples prepared when the pH of the impregnating solution was higher than the p.z.c. of silica. Electrostatic repulsions between the like-charged surface of silica and Pt4+ ions resulted in good dispersion of the crystallized particles of the precursor phase (sample PtpH
147
The oxidation of SO2 is exothermic meaning that at high temperatures the equilibrium is shifted to low conversion. For the composition of the synthesis gas mixture used in the present work, the equilibrium is 99.98% to the side of the oxidized products when the temperature is 270 8C. At 400 8C the conversion decrease to 99.1% and drops rapidly from 95.6 to 83.8% in the temperature range from 470 to 550 8C (where most of the activity maximums were observed for the catalyst tested). The conversion of SO2 was low at 270 8C. However, for the very low loading of the reactor and of the catalyst with the active phase the activity is high when the temperature increased to 400 8C. Since the SO2 oxidation reaction is exothermal, when the temperature exceeds 550 8C the equilibrium is significantly shifted to the side of the reactants. At such high temperatures the thermodynamic factors prevail over the kinetic. The thermodynamics of the system and sintering of the particles of the active phase are reasons for the decreased catalytic activity at high temperatures. 4. Conclusions The catalytic oxidation of SO2 over platinum based catalysts takes place in the temperature range between 250 and 600 8C. At higher temperatures the catalytic activity drops as a result of sintering of the platinum particles and thermodynamic limitations. Physical parameters such as the pore size of the carrier and the molecular volume of the impregnated precursor salt influence the characteristics of the produced catalyst and the catalytic activity. Silica gel gave good particle distribution when impregnated in solution of pH below the p.z.c. of the surface. Furthermore, the particle size of the active phase is critical and more important than the poisoning effect of traces of chloride ions. Titania supported samples exhibit better catalytic properties as suggested from the high space velocities obtained even if the particle size of the active phase was twice as big as those supported on silica gel. This confirms the importance of the properties of the carrier in the catalystassisted SO2 oxidation process. Acknowledgements The NATO, Science for Peace Program, (SfP 971984) has supported this work. We also thank two of the reviewers for insightful suggestions and useful comments. References [1] Y. Nagai, H. Shinjoh, K. Yokota, Appl. Catal. B 39 (2002) 149. [2] E. Xue, K. Seshan, J.R.H. Ross, Appl. Catal. B 11 (1996) 65. [3] L.G. Simonova, V.V. Barelko, A.V. Toktarev, A.F. Chernyshov, V.A. Chumachenko, B.S. Bal’zhinimaev, Kinet. Catal. 43 (2002) 61. [4] M. Deeba, R.J. Farrauto, Y.K. Lui, Appl. Catal. A 124 (1995) 339. [5] D. Uy, A. Dubkov, G.W. Graham, W.H. Weber, Catal. Lett. 68 (2000) 25. [6] B.C. Gates, Catalytic Chemistry, John Wiley and Sons, Inc., New York, 1992, p. 401. [7] J.M. Parera, N.S. Fagoli, in: J.J. Spivey (Ed.), Catalysis, vol. 9, Royal Society of Chemistry, 1992, pp. 65–125. [8] D. McDonald, Chem. Ind. (1931) 1031. [9] J. Lee, S.H. Langer, J. Appl. Electrochem. 25 (1995) 353.
148
S. Koutsopoulos et al. / Applied Catalysis A: General 306 (2006) 142–148
[10] R. Lamber, N. Jaeger, G. Schultz-Ekloff, J. Catal. 23 (1990) 285. [11] C.R. Adams, H.A. Benesi, R.M. Curtis, R.G. Meisenheimer, J. Catal. 1 (1962) 336. [12] D. Martin, D. Duprez, Stud. Surf. Sci. Catal. 77 (1993) 201. [13] G. Ramis, G. Busca, V. Lorenzelli, P. Forzatti, Appl. Catal. 64 (1990) 243. [14] M.A. Vannice, J. Catal. 74 (1982) 199. [15] A. Bahamonde, C. Knapp, P. Avila, J. Blanco, Appl. Catal. B 19 (1998) 1.
[16] P. Meriaudeau, O.H. Ellenstad, M. Dufaux, C. Naccache, J. Catal. 75 (1982) 243. [17] E.B. Maxted, A. Marsden, J. Chem. Soc. (1940) 469. [18] P.K. Gallagher, M.E. Gross, J. Therm. Anal. 31 (1986) 1231. [19] H.P. Klugg, L.E. Alexander, X-Ray Diffraction Procedures, John Wiley and Sons, Inc., New York, 1970. [20] H.N. Holmes, J. Ramsay, A.L. Elder, Ind. Eng. Chem. 21 (1929) 850.