- Email: [email protected]
NO conversion using binary vanadium mixtures supported on activated carbon
Applied Catalysis B: Environmental 44 (2003) 227–235
NO conversion using binary vanadium mixtures supported on activated carbon Sónia A. Carabineiro a , F. Brás Fernandes b , Joaquim S. Vital a , Ana M. Ramos a , Isabel M. Fonseca a,∗ a
Departamento de Qu´ımica, C.Q.F.B, Faculdade de Ciˆencias e Tecnologia, Universidade Nova de Lisboa, 2825–114 Caparica, Portugal b Departamento de Ciˆ encia dos Materiais, CENIMAT, Faculdade de Ciˆencias e Tecnologia, Universidade Nova de Lisboa, 2825–114 Caparica, Portugal Received 2 August 2002; received in revised form 29 January 2003; accepted 31 January 2003
Abstract The kinetics of the reaction of NO with activated carbon without catalyst and impregnated with precursor salts of Ba, Co, Cu, Fe, Mg, Mn, Ni and Pb were investigated. Binary mixtures of those metals with V were also studied. The conversion of NO was studied (300–900 ◦ C) using a TGA apparatus and a fixed bed reactor. The reactor effluents were analysed using a GC/MS on line and showed the presence of N2 O, N2 , CO2 and CO together with some unreacted NO. The best synergetic effects were observed for samples doped with V + Cu and V + Fe mixtures. This synergetic effect observed with addition of vanadium seems to be related with its ability to wet and spread on carbon surface and with enhancement of oxygen–carbon complexes on carbon surface increasing CO2 evolution, thus generating new active sites. © 2003 Elsevier Science B.V. All rights reserved. Keywords: NO carbon gasification; Catalysis; Kinetics; X-ray diffraction
1. Introduction NO removal from exhaust streams of various combustion sources has become increasingly important in the past few years [1–14]. Recently catalytic reduction of NO with carbons as reducing reagents has been intensively investigated [2–8,10–12,14]. The reaction depends significantly on the composition of the catalyst and nature of the reacting carbon [8]. In attempt to understand and propose a mechanism for catalytic carbon gasification with oxygen, carbon dioxide, water vapour and hydrogen, McKee stud∗ Corresponding author. Tel.: +351-21-2948385; fax: +351-21-2948385. E-mail address: [email protected] (I.M. Fonseca).
ied the effects of alkaline, alkaline-earth and transition metals using TGA/DTA and hot stage microscopy [15,16]. The author concluded that the catalytic process involves an oxidation–reduction cycle. The oxidation state of the metal determines its performance and the ability of the precursor to be reduced by carbon to a lower oxidation state is an important factor [15,16]. Also the ability of the catalyst to wet and spread on carbon surface promotes catalyst/carbon contact and reactivity is enhanced [1,15–21]. According to earlier studies of Aarna and Suuberg at low temperatures dissociative NO chemisorption results in the formation of a oxygen–carbon complex and the decomposition of these complexes (at higher temperatures) leads to production of CO2 and CO [13].
0926-3373/$ – see front matter © 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0926-3373(03)00064-X
228
S.A. Carabineiro et al. / Applied Catalysis B: Environmental 44 (2003) 227–235
Illán-Gomez et al. reported that alkaline, alkalineearth and transition metal catalysts seem to enhance NO chemisorption to different extents and they have an important role in the redox cycle, which transfers oxygen from the catalyst surface to the carbon surface to produce CO2 and CO [4–8]. This mechanism is analogous to that one occurring in other carbon gasification reactions [15,16]. The same authors reported that the analysis of the reaction products reveals a mechanism for catalytic NO reduction with three stages [4–8]: (i) At low temperatures, N2 and or N2 O are the only products. Oxygen is retained on the catalyst/carbon surface. (ii) Above 300 ◦ C, N2 continues to evolve but CO2 evolution occurs and the rate of NO reduction increases. (iii) Above 500 ◦ C, N2 evolution becomes constant and CO becomes dominant. The active sites that retain and transfer the oxygen to carbon are different, depending on the nature of the chemical species involved in each case [5–8]. However, when NO reduction is carried out over palladium supported on activated carbon, in the presence of stoichiometric amounts of CO, it is not observed carbon gasification, being proposed the following mechanism [22]: NO + CO → 21 N2 + CO2 2NO + CO → N2 O + CO2 A possible reaction between NO and CO evolved from the carbon surface is also reported by Mehandjiev and Bekyarova [23] for the NO reduction over cobalt catalysts supported on activated carbon. The aim of the present work is to study the behaviour of several binary vanadium mixtures for the C–NO reaction. In situ XRD is used to identify the phases present under reaction conditions. This technique combined with kinetics measurements is a useful tool for interpreting catalyst behaviour.
2. Experimental Charcoal activated GR MERCK (powder) was used to carry out isothermal kinetic studies of NO. This
activated charcoal had a surface area of 1011 m2 /g and a pore total volume of 0.54 cm3 /g. The charcoal was impregnated with diluted solutions of Ba, Co, Cu, Fe, Mg, Mn, Ni, Pb and V precursors (acetates). Metal salt was diluted in a small amount of distilled water and the obtained solution was added dropwise to the carbon, mixing thoroughly and the obtained product was evaporated in hot plate. The metal loading was 4 wt.%. Binary vanadium mixtures of these catalysts were also prepared. Isothermal kinetic studies were carried out by using a thermobalance system (CI Electronics MKII Microbalance), with continuous recording of change in weight between 300 and 900 ◦ C, under Ar + 0.5% NO. In order to prevent diffusional limitations, sample weights of 2–3 mg and a flow rate of 3.3 cm3 /s, were used. For comparison purposes, initial rates obtained at initial burn off, were used. The gasification rates at each temperature were derived from the relation: rate =
−w/t w0
where w0 is the initial weight of sample. Before each experiment, Ar was introduced with a flow rate of 3.3 cm3 /s, for 30 min, in order to eliminate traces of air. The samples were heated to 500 ◦ C, under Ar flow, to decompose precursor salts and to eliminate volatiles adsorbed on the carbon surface, until constant sample weight. The temperature was then adjusted to the desired value. NO conversion was studied using a fixed bed reactor. The reactor effluents were analysed using a GC/MS (Fisons MD800) apparatus, equipped with a column of GS-Molesieve (30 m × 0.541 mm). These essays included heating of sample at 2 ◦ C/min (TPR) up to 950 ◦ C in a mixture of 0.5% NO in He. The flow rate was 3.3 cm3 /s. All samples (20 mg) were pre-treated in He at 500 ◦ C for 30 min. Tests were performed to check the absence of mass transfer limitations. In situ XRD studies were conducted in a Rigaku D/max III C diffractometer with a Cu K␣ radiation source (50 kV, 30 mA), equipped with a high temperature special chamber. All the experiments were carried out at exactly the same experimental conditions as in the kinetic studies.
S.A. Carabineiro et al. / Applied Catalysis B: Environmental 44 (2003) 227–235
229
Table 1 Initial gasification rates in NO (0.5% NO in Ar) of carbon impregnated with V binary mixtures, at 450 ◦ C (for V catalysed reaction, the rate was 0.0083 × 10−5 mol/(s gcat )), compared to NO conversion rates System 4% 4% 4% 4% 4% 4% 4% 4%
V V V V V V V V
+ 4% + 4% + 4% + 4% + 4% + 4% + 4% + 4%
Ba Co Cu Fe Mg Mn Ni Pb
Initial carbon gasification rate (r × 105 mol/(s gcat ))
Sum of initial carbon gasification rates of individual catalysts
0.014 0.125 0.419 0.253 0.083 0.067 0.024 0.013
0.015 0.129 0.125 0.013 0.081 0.065 0.024 0.015
3. Results and discussion Tables 1 and 2 show the reaction rates obtained at 450 and 800 ◦ C, respectively, for the vanadium mixtures. At higher temperature almost all mixtures exhibited synergetic effects as shown in Table 2. But at lower temperature (Table 1) the highest synergetic effects observed were for the mixtures V + Cu and V + Fe. Figs. 1 and 2 show the Arrhenius plots for NO reaction for these V + Fe and V + Cu mixtures. The values obtained for the apparent activation energies range from 43 to 139 kJ/mol and lie relatively well within those reported in the literature [13]. A substantial decrease in the apparent activation energy for the uncatalysed (138 kJ/mol) and Cu (41 kJ/mol) and V (43 kJ/mol) catalysed reaction is observed. In the presence of iron (single catalyst) there was no significant decrease on the apparent
NO conversion rate (r × 105 mol/(s gcat ))
0.953 0.903
activation energy in comparison to the uncatalysed reaction (131 kJ/mol). Above 700 ◦ C, internal diffusional limitations to mass transfer seem to take place because the activation energy is reduced to one half (55 kJ/mol). In the case of the reaction catalysed by Fe, addition of vanadium decreased significantly the apparent activation energy (61 kJ/mol). Synergetic effects of binary mixtures were reported for catalytic gasification of carbon with steam, CO2 , and N2 O [20,21,24]. Melting of the catalyst and wetting of the carbon surface are important factors that affect the catalytic activity. In previous works it has been shown that addition of V or Ba increased reactivity at low temperatures, due to the presence of the catalyst as melted phase [20,21,24]. The synergetic effects of binary mixtures can also result from other beneficial interactions between the mixture components and those between reactants and the individual components [19].
Table 2 Initial gasification rates in NO (0.5% NO in Ar) of carbon impregnated with V binary mixtures, at 800 ◦ C (for V catalysed reaction, the rate was 0.4308 × 10−5 mol/(s gcat )), compared to NO conversion rates System 4% 4% 4% 4% 4% 4% 4% 4%
V V V V V V V V
+ 4% + 4% + 4% + 4% + 4% + 4% + 4% + 4%
Ba Co Cu Fe Mg Mn Ni Pb
Initial carbon gasification rate (r × 105 mol/(s gcat ))
Sum of initial carbon gasification rates of individual catalysts
6.008 4.578 4.333 6.963 0.833 1.154 4.694 2.768
1.108 2.376 1.548 1.744 0.782 1.168 4.883 0.655
NO conversion rate (r × 105 mol/(s gcat ))
1.308 1.314
230
S.A. Carabineiro et al. / Applied Catalysis B: Environmental 44 (2003) 227–235
Fig. 1. Arrhenius plots for carbon gasification in NO for carbon parent sample and samples doped with V, Cu and V + Cu mixture.
Figs. 3 and 4 show the catalysed program temperature reaction profile for NO conversion. These results are in agreement with the kinetic ones, as the mixture converts the NO at lower temperature than the individ-
ual components. The analysis of the reaction products for the reaction with NO (Fig. 5a–f) using a GC/MS on line show the presence of N2 O, N2 , CO2 and CO together with some unreacted NO.
Fig. 2. Arrhenius plots for carbon gasification in NO for carbon parent sample and samples doped with V, Fe and V + Fe mixture.
S.A. Carabineiro et al. / Applied Catalysis B: Environmental 44 (2003) 227–235
Fig. 3. NO conversion as a function of temperature (TPR) for samples doped with V, Cu and V + Cu mixture.
Fig. 4. NO conversion as a function of temperature (TPR) for samples doped with V, Fe and V + Fe mixture.
231
232
S.A. Carabineiro et al. / Applied Catalysis B: Environmental 44 (2003) 227–235
Fig. 5. Reaction products in TPR for carbon parent sample (a) and samples doped with V (b), Cu (c), Fe (d) and Cu + V (e) and Fe + V (f) mixtures.
The analysis of the reaction products for the uncatalysed reaction (Fig. 5a) and the vanadium catalysed reaction (Fig. 5b), showed that between 400 and 500 ◦ C only CO2 , N2 O and N2 were detected. Above 500 ◦ C minor quantities of CO were observed. In the case of the uncatalysed reaction CO2 evolution was observed at higher temperatures (above 500 ◦ C). A substantial increase in CO2 evolution is observed for the vanadium catalysed reaction compared with the uncatalysed one. These results suggest that dissociative NO chemisorption occurs and the catalyst enhances the formation of oxygen–carbon groups. Decomposi-
tion of these complexes leads to production of CO2 and CO. In the case of Fe and Cu, only N2 and N2 O were detected in the low temperature range (20–300 ◦ C). Like in the mechanism proposed by Illán-Gomez et al. [4–8] N2 and N2 O were observed at low temperatures. To observe N2 formation it is necessary to adsorb and dissociate NO, i.e. to have sites to retain the oxygen atoms. The N2 O observed at low temperatures seems to suggest that the oxygen is irreversibly adsorbed on the catalyst surface. Above 300 ◦ C, CO2 is also observed. For the Fe catalysed reaction CO2 evolution
S.A. Carabineiro et al. / Applied Catalysis B: Environmental 44 (2003) 227–235
is observed at 250 ◦ C (Fig. 5d) suggesting that the oxygen–carbon complexes are less stable and more abundant. In the case of the binary mixtures (Fig. 5e and f), a substantial increase in CO2 evolution was observed above 300 ◦ C and it is delayed with respect to N2 . The delay may be due to the fact that the catalyst retains the oxygen produced by dissociative NO chemisorption. The evolution of CO2 allows the catalyst to continue to transfer oxygen from NO to carbon, generating new active sites on the carbon surface. The catalytic effect observed with addition of vanadium seems to be related with an enhancement of oxygen–carbon complexes and increase in CO2 evolution. The greater the rate of CO2 evolution the greater the NO conversion. Similar results were obtained by other authors [3–7]. Active phases play also an important role on the catalytic conversion of NO [4–8,20,21]. Figs. 6 and 7 show XRD patterns obtained in NO at several temperatures for carbon doped with V + Cu and V + Fe mixtures. The peaks show shifts to 2θ with increasing temperature, which reflects expansion of the crystal lattices. Pt peaks appear in the spectra resulting from exposure of the sample holder to the X-ray beam as carbon burned away.
233
In situ XRD shows clearly that metal oxides are reduced to lower oxidation states such as V2 O5 , V6 O13 ; CuO, Cu; Fe2 O3 , Fe3 O4 . Same phases were detected under inert atmosphere. No evidence of a new phase responsible for the synergetic effect was observed in the mixtures by in situ XRD since the same phases were observed both in the XRD patterns for the samples with two catalyst species and in the spectra for the samples bearing single catalyst species. However the addition of vanadium seems to increase carbon reactivity at low temperatures. As temperature increases some peaks become less intense suggesting that the catalyst is present as a melted phase. This is the case of Fe2 O3 and V2 O5 /V6 O13 . The catalytic effect observed for the NO conversion can be explained by the occurrence of redox processes in which the oxide particles are reduced by reaction with carbon at points of contact with the carbon substrate to form lower oxides according to a redox mechanism such as: y Mx + yNO → Mx Oy + N2 2 y y Mx Oy + C → Mx + CO2 2 2
Fig. 6. In situ XRD data obtained in NO on heating sample doped with Cu + V mixture at several temperatures.
234
S.A. Carabineiro et al. / Applied Catalysis B: Environmental 44 (2003) 227–235
Fig. 7. In situ XRD data obtained in NO on heating sample doped with Fe + V mixture at several temperatures.
The parallel reaction between NO and CO evolved from the carbon surface, as proposed for Pd or Co catalysts supported on activated carbon [22,23], is not likely to take place in the present reaction conditions, not only because CO is only detected in minor amounts (Fig. 5), but also because the carbon gasification rates have the same order of magnitude than the NO conversion rates (Tables 1 and 2). At 800 ◦ C the carbon gasification rate is even 3–5 times higher then the NO conversion rate. This is an indication that the parallel reaction between NO and CO does not take place in a significant extension. It is known that V2 O5 is a good catalyst for carbon gasification because of its low Tamman temperature. Various authors reported that V2 O5 has the ability to spread and wet carbon surfaces [15,16,19,25,26]. The reduction of V2 O5 to V6 O13 further reduces the melting point of the catalyst and could lead to more extensive spreading on carbon surface [15,16,19,25,26]. In situ XRD carried out both in NO and N2 atmospheres showed that both V2 O5 and V6 O13 are present under reaction conditions and with increasing temperature the peaks become less intense to both oxides suggesting that the catalyst is spreading to form an amorphous thin film on carbon surface. It has been reported that melting of the catalysts improves wetting of the
carbon surface promoting catalyst/carbon interaction [1,15,16,19–21]. Also Pan and Yang indicated that pitting and channelling of graphite was caused by the reaction between V2 O5 and carbon in which V2 O5 was reduced to V6 O13 . Also a liquid-like behaviour of the vanadium oxide particles was observed [25]. Thus the reduction of the metal oxide seems to be a key factor that controls the catalytic activity of the metal. Illán-Gomez et al. have reported that in an NO atmosphere a relationship exists between the ease of the oxidation–reduction reaction (to the metallic or a lower oxidation state) and the lattice energy of the oxide or the free energy for oxide formation [7]. In the same work the catalyst became inactive due to its conversion to the oxidised state and negligible reduction of the gas took place [7,8]. In this work it was shown by XRD that after pre-treatment at 500 ◦ C in inert atmosphere the catalysts were already reduced by carbon to lower oxidation states. It has been reported recently that alkali, alkaline-earth and transition metals enhance NO chemisorption. The catalytic system is more or less active depending on the ability of the oxide to be reduced by carbon and undergo redox reactions on the carbon surface [1,4–8,15,20,21]. The same mechanism has
S.A. Carabineiro et al. / Applied Catalysis B: Environmental 44 (2003) 227–235
been proposed to explain carbon catalysed reactions in several gaseous atmospheres such as CO2 reaction [7,8,15–18,20,21,24,27–30]. Another important factor is the catalyst dispersion. As dispersion increases the oxidation–reduction process of the catalysts is enhanced as the redox mechanism occurs at the catalyst/carbon surface. Thus in the case of the samples doped with vanadium, this interaction is promoted because the catalyst seems to be presented as a thin film, wetting the carbon surface. In situ XRD showed the probable species involved in the reaction conditions. It seems that a reduced catalyst surface is required for NO reduction. Active catalysts seem to act as an oxygen acceptor from NO, transferring it to the carbon surface and recovering reduced state. Similar behaviour has been reported by several authors [1,4,5,6,10,11,15,21,27,29].
4. Conclusions This study has indicated that kinetic measurements combined with in situ XRD are useful tools for interpreting catalyst behaviour. In the C–NO reaction a reduced catalyst surface is required to NO reduction. The ability of the catalyst to chemisorb NO going through redox transference of oxygen to the carbon reactive sites seems to explain catalytic activity. The synergetic effect observed with addition of vanadium seems to be related not only with its ability to wet and spread on the carbon surface but also with the enhancement of oxygen–carbon complexes. This last effect leads to the increase of CO2 evolution thus generating new active sites.
Acknowledgements Sónia Carabineiro is grateful to Fundação para a Ciˆencia e Tecnologia for Ph.D. Grant BD1275/95. References [1] R.T.K. Baker, C.R.F. Lund, J.J. Chlundzinski Jr., J. Catal. 87 (1984) 255.
235
[2] T. Okuhara, K. Tanaka, J. Chem. Soc. Faraday Trans. 1 (1986) 2357. [3] J. Rodriguez-Mirasol, A.C. Ooms, J.R. Pels, F. Kapteijn, J.A. Moulijn, Combust. Flame 99 (1994) 499. [4] M.J. Illán-Gomez, C. Salinas-Martinez de Lecea, A. Linares-Solano, in: Proceedings of the Extended Abstracts of Carbon’94, Granada, Spain, 1994, p. 460. [5] M.J. Illán-Gomez, A. Linares-Solano, L. Radovic, C. SalinasMartinez de Lecea, Energy Fuels 9 (1995) 97. [6] M.J. Illán-Gomez, A. Linares-Solano, L. Radovic, C. Salinas-Martinez de Lecea, Energy Fuels 9 (1995) 540. [7] M.J. Illán-Gomez, A. Linares-Solano, C. Salinas-Martinez de Lecea, Energy Fuels 9 (1995) 976. [8] M.J. Illán-Gomez, A. Linares-Solano, L. Radovic, C. SalinasMartinez de Lecea, Energy Fuels 10 (1995) 158. [9] J. Rodriguez-Mirasol, J.R. Pels, F. Kapteijn, J.A. Moulijn, in: Proceedings of the 22nd Biennial Conference on Carbon, San Diego, CA, USA, 1995, 620. [10] D. Mehandjiev, M. Khristova, E. Bekyarova, Carbon 34 (1996) 757. [11] C. Marquez-Alvarez, I. Rodrigues-Ramos, A. Guerrero-Ruiz, Carbon 34 (1996) 339. [12] S. Carabineiro, I.F. Silva, F.B. Fernandes, in: Proceedings of the 9th International Conference of Coal Science, Essen, Germany, 1997, p. 1103. [13] I. Aarna, E.M. Suuberg, Fuel 76 (1997) 475. [14] S.A. Carabineiro, I.F. Silva, F.B. Fernandes, A.M. Ramos, J. Vital, in: Proceedings of the Eurocabon’98, Strasbourg, France, 1998, p. 411. [15] D.W. McKee, Chemistry and Physics of Carbon, vol. 16, Marcel Dekker, New York, 1981, p. 1. [16] D.W. McKee, Carbon 8 (1970) 623. [17] I.F. Silva, L.S. Lobo, J. Catal. 126 (1990) 489. [18] I.F. Silva, D.W. McKee, L.S. Lobo, J. Catal. 170 (1997) 54. [19] I.F. Silva, C. Palma, M. Klimkiewicz, S. Eser, Carbon 36 (1998) 861. [20] S.A. Carabineiro, F.B. Fernandes, A.M. Ramos, J. Vital, I.F. Silva, Catal. Today 54 (1999) 559. [21] S.A. Carabineiro, F.B. Fernandes, A.M. Ramos, J. Vital, I.F. Silva, Catal. Today 57 (2000) 305. [22] M.A. Salas-Peregrin, M. Primet, H. Praliaud, Appl. Catal. B 8 (1996) 79. [23] D. Mehandjiev, E. Bekyarova, J. Colloid Interf. Sci. 166 (1994) 476. [24] S.A. Carabineiro, D.W. McKee, I.F. Silva, Carbon 39 (2001) 451. [25] Z. Pan, R. Yang, J. Catal. 130 (1991) 161. [26] X. Chu, L. Schmidt, S. Chen, R. Yang, J. Catal. 140 (1993) 543. [27] D.W. McKee, Fuel 59 (1980) 308. [28] B.J. Wood, K.M. Sancier, Catal. Rev.-Sci. Eng. 26 (1984) 233. [29] H. Ohme, T. Suzuki, Energy Fuels 10 (1996) 980. [30] F. Kapteijn, H. Porre, J.A. Moulijn, AIChE J. 31 (1986) 691.
NO conversion using binary vanadium mixtures supported on activated carbon Sónia A. Carabineiro a , F. Brás Fernandes b , Joaquim S. Vital a , Ana M. Ramos a , Isabel M. Fonseca a,∗ a
Departamento de Qu´ımica, C.Q.F.B, Faculdade de Ciˆencias e Tecnologia, Universidade Nova de Lisboa, 2825–114 Caparica, Portugal b Departamento de Ciˆ encia dos Materiais, CENIMAT, Faculdade de Ciˆencias e Tecnologia, Universidade Nova de Lisboa, 2825–114 Caparica, Portugal Received 2 August 2002; received in revised form 29 January 2003; accepted 31 January 2003
Abstract The kinetics of the reaction of NO with activated carbon without catalyst and impregnated with precursor salts of Ba, Co, Cu, Fe, Mg, Mn, Ni and Pb were investigated. Binary mixtures of those metals with V were also studied. The conversion of NO was studied (300–900 ◦ C) using a TGA apparatus and a fixed bed reactor. The reactor effluents were analysed using a GC/MS on line and showed the presence of N2 O, N2 , CO2 and CO together with some unreacted NO. The best synergetic effects were observed for samples doped with V + Cu and V + Fe mixtures. This synergetic effect observed with addition of vanadium seems to be related with its ability to wet and spread on carbon surface and with enhancement of oxygen–carbon complexes on carbon surface increasing CO2 evolution, thus generating new active sites. © 2003 Elsevier Science B.V. All rights reserved. Keywords: NO carbon gasification; Catalysis; Kinetics; X-ray diffraction
1. Introduction NO removal from exhaust streams of various combustion sources has become increasingly important in the past few years [1–14]. Recently catalytic reduction of NO with carbons as reducing reagents has been intensively investigated [2–8,10–12,14]. The reaction depends significantly on the composition of the catalyst and nature of the reacting carbon [8]. In attempt to understand and propose a mechanism for catalytic carbon gasification with oxygen, carbon dioxide, water vapour and hydrogen, McKee stud∗ Corresponding author. Tel.: +351-21-2948385; fax: +351-21-2948385. E-mail address: [email protected] (I.M. Fonseca).
ied the effects of alkaline, alkaline-earth and transition metals using TGA/DTA and hot stage microscopy [15,16]. The author concluded that the catalytic process involves an oxidation–reduction cycle. The oxidation state of the metal determines its performance and the ability of the precursor to be reduced by carbon to a lower oxidation state is an important factor [15,16]. Also the ability of the catalyst to wet and spread on carbon surface promotes catalyst/carbon contact and reactivity is enhanced [1,15–21]. According to earlier studies of Aarna and Suuberg at low temperatures dissociative NO chemisorption results in the formation of a oxygen–carbon complex and the decomposition of these complexes (at higher temperatures) leads to production of CO2 and CO [13].
0926-3373/$ – see front matter © 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0926-3373(03)00064-X
228
S.A. Carabineiro et al. / Applied Catalysis B: Environmental 44 (2003) 227–235
Illán-Gomez et al. reported that alkaline, alkalineearth and transition metal catalysts seem to enhance NO chemisorption to different extents and they have an important role in the redox cycle, which transfers oxygen from the catalyst surface to the carbon surface to produce CO2 and CO [4–8]. This mechanism is analogous to that one occurring in other carbon gasification reactions [15,16]. The same authors reported that the analysis of the reaction products reveals a mechanism for catalytic NO reduction with three stages [4–8]: (i) At low temperatures, N2 and or N2 O are the only products. Oxygen is retained on the catalyst/carbon surface. (ii) Above 300 ◦ C, N2 continues to evolve but CO2 evolution occurs and the rate of NO reduction increases. (iii) Above 500 ◦ C, N2 evolution becomes constant and CO becomes dominant. The active sites that retain and transfer the oxygen to carbon are different, depending on the nature of the chemical species involved in each case [5–8]. However, when NO reduction is carried out over palladium supported on activated carbon, in the presence of stoichiometric amounts of CO, it is not observed carbon gasification, being proposed the following mechanism [22]: NO + CO → 21 N2 + CO2 2NO + CO → N2 O + CO2 A possible reaction between NO and CO evolved from the carbon surface is also reported by Mehandjiev and Bekyarova [23] for the NO reduction over cobalt catalysts supported on activated carbon. The aim of the present work is to study the behaviour of several binary vanadium mixtures for the C–NO reaction. In situ XRD is used to identify the phases present under reaction conditions. This technique combined with kinetics measurements is a useful tool for interpreting catalyst behaviour.
2. Experimental Charcoal activated GR MERCK (powder) was used to carry out isothermal kinetic studies of NO. This
activated charcoal had a surface area of 1011 m2 /g and a pore total volume of 0.54 cm3 /g. The charcoal was impregnated with diluted solutions of Ba, Co, Cu, Fe, Mg, Mn, Ni, Pb and V precursors (acetates). Metal salt was diluted in a small amount of distilled water and the obtained solution was added dropwise to the carbon, mixing thoroughly and the obtained product was evaporated in hot plate. The metal loading was 4 wt.%. Binary vanadium mixtures of these catalysts were also prepared. Isothermal kinetic studies were carried out by using a thermobalance system (CI Electronics MKII Microbalance), with continuous recording of change in weight between 300 and 900 ◦ C, under Ar + 0.5% NO. In order to prevent diffusional limitations, sample weights of 2–3 mg and a flow rate of 3.3 cm3 /s, were used. For comparison purposes, initial rates obtained at initial burn off, were used. The gasification rates at each temperature were derived from the relation: rate =
−w/t w0
where w0 is the initial weight of sample. Before each experiment, Ar was introduced with a flow rate of 3.3 cm3 /s, for 30 min, in order to eliminate traces of air. The samples were heated to 500 ◦ C, under Ar flow, to decompose precursor salts and to eliminate volatiles adsorbed on the carbon surface, until constant sample weight. The temperature was then adjusted to the desired value. NO conversion was studied using a fixed bed reactor. The reactor effluents were analysed using a GC/MS (Fisons MD800) apparatus, equipped with a column of GS-Molesieve (30 m × 0.541 mm). These essays included heating of sample at 2 ◦ C/min (TPR) up to 950 ◦ C in a mixture of 0.5% NO in He. The flow rate was 3.3 cm3 /s. All samples (20 mg) were pre-treated in He at 500 ◦ C for 30 min. Tests were performed to check the absence of mass transfer limitations. In situ XRD studies were conducted in a Rigaku D/max III C diffractometer with a Cu K␣ radiation source (50 kV, 30 mA), equipped with a high temperature special chamber. All the experiments were carried out at exactly the same experimental conditions as in the kinetic studies.
S.A. Carabineiro et al. / Applied Catalysis B: Environmental 44 (2003) 227–235
229
Table 1 Initial gasification rates in NO (0.5% NO in Ar) of carbon impregnated with V binary mixtures, at 450 ◦ C (for V catalysed reaction, the rate was 0.0083 × 10−5 mol/(s gcat )), compared to NO conversion rates System 4% 4% 4% 4% 4% 4% 4% 4%
V V V V V V V V
+ 4% + 4% + 4% + 4% + 4% + 4% + 4% + 4%
Ba Co Cu Fe Mg Mn Ni Pb
Initial carbon gasification rate (r × 105 mol/(s gcat ))
Sum of initial carbon gasification rates of individual catalysts
0.014 0.125 0.419 0.253 0.083 0.067 0.024 0.013
0.015 0.129 0.125 0.013 0.081 0.065 0.024 0.015
3. Results and discussion Tables 1 and 2 show the reaction rates obtained at 450 and 800 ◦ C, respectively, for the vanadium mixtures. At higher temperature almost all mixtures exhibited synergetic effects as shown in Table 2. But at lower temperature (Table 1) the highest synergetic effects observed were for the mixtures V + Cu and V + Fe. Figs. 1 and 2 show the Arrhenius plots for NO reaction for these V + Fe and V + Cu mixtures. The values obtained for the apparent activation energies range from 43 to 139 kJ/mol and lie relatively well within those reported in the literature [13]. A substantial decrease in the apparent activation energy for the uncatalysed (138 kJ/mol) and Cu (41 kJ/mol) and V (43 kJ/mol) catalysed reaction is observed. In the presence of iron (single catalyst) there was no significant decrease on the apparent
NO conversion rate (r × 105 mol/(s gcat ))
0.953 0.903
activation energy in comparison to the uncatalysed reaction (131 kJ/mol). Above 700 ◦ C, internal diffusional limitations to mass transfer seem to take place because the activation energy is reduced to one half (55 kJ/mol). In the case of the reaction catalysed by Fe, addition of vanadium decreased significantly the apparent activation energy (61 kJ/mol). Synergetic effects of binary mixtures were reported for catalytic gasification of carbon with steam, CO2 , and N2 O [20,21,24]. Melting of the catalyst and wetting of the carbon surface are important factors that affect the catalytic activity. In previous works it has been shown that addition of V or Ba increased reactivity at low temperatures, due to the presence of the catalyst as melted phase [20,21,24]. The synergetic effects of binary mixtures can also result from other beneficial interactions between the mixture components and those between reactants and the individual components [19].
Table 2 Initial gasification rates in NO (0.5% NO in Ar) of carbon impregnated with V binary mixtures, at 800 ◦ C (for V catalysed reaction, the rate was 0.4308 × 10−5 mol/(s gcat )), compared to NO conversion rates System 4% 4% 4% 4% 4% 4% 4% 4%
V V V V V V V V
+ 4% + 4% + 4% + 4% + 4% + 4% + 4% + 4%
Ba Co Cu Fe Mg Mn Ni Pb
Initial carbon gasification rate (r × 105 mol/(s gcat ))
Sum of initial carbon gasification rates of individual catalysts
6.008 4.578 4.333 6.963 0.833 1.154 4.694 2.768
1.108 2.376 1.548 1.744 0.782 1.168 4.883 0.655
NO conversion rate (r × 105 mol/(s gcat ))
1.308 1.314
230
S.A. Carabineiro et al. / Applied Catalysis B: Environmental 44 (2003) 227–235
Fig. 1. Arrhenius plots for carbon gasification in NO for carbon parent sample and samples doped with V, Cu and V + Cu mixture.
Figs. 3 and 4 show the catalysed program temperature reaction profile for NO conversion. These results are in agreement with the kinetic ones, as the mixture converts the NO at lower temperature than the individ-
ual components. The analysis of the reaction products for the reaction with NO (Fig. 5a–f) using a GC/MS on line show the presence of N2 O, N2 , CO2 and CO together with some unreacted NO.
Fig. 2. Arrhenius plots for carbon gasification in NO for carbon parent sample and samples doped with V, Fe and V + Fe mixture.
S.A. Carabineiro et al. / Applied Catalysis B: Environmental 44 (2003) 227–235
Fig. 3. NO conversion as a function of temperature (TPR) for samples doped with V, Cu and V + Cu mixture.
Fig. 4. NO conversion as a function of temperature (TPR) for samples doped with V, Fe and V + Fe mixture.
231
232
S.A. Carabineiro et al. / Applied Catalysis B: Environmental 44 (2003) 227–235
Fig. 5. Reaction products in TPR for carbon parent sample (a) and samples doped with V (b), Cu (c), Fe (d) and Cu + V (e) and Fe + V (f) mixtures.
The analysis of the reaction products for the uncatalysed reaction (Fig. 5a) and the vanadium catalysed reaction (Fig. 5b), showed that between 400 and 500 ◦ C only CO2 , N2 O and N2 were detected. Above 500 ◦ C minor quantities of CO were observed. In the case of the uncatalysed reaction CO2 evolution was observed at higher temperatures (above 500 ◦ C). A substantial increase in CO2 evolution is observed for the vanadium catalysed reaction compared with the uncatalysed one. These results suggest that dissociative NO chemisorption occurs and the catalyst enhances the formation of oxygen–carbon groups. Decomposi-
tion of these complexes leads to production of CO2 and CO. In the case of Fe and Cu, only N2 and N2 O were detected in the low temperature range (20–300 ◦ C). Like in the mechanism proposed by Illán-Gomez et al. [4–8] N2 and N2 O were observed at low temperatures. To observe N2 formation it is necessary to adsorb and dissociate NO, i.e. to have sites to retain the oxygen atoms. The N2 O observed at low temperatures seems to suggest that the oxygen is irreversibly adsorbed on the catalyst surface. Above 300 ◦ C, CO2 is also observed. For the Fe catalysed reaction CO2 evolution
S.A. Carabineiro et al. / Applied Catalysis B: Environmental 44 (2003) 227–235
is observed at 250 ◦ C (Fig. 5d) suggesting that the oxygen–carbon complexes are less stable and more abundant. In the case of the binary mixtures (Fig. 5e and f), a substantial increase in CO2 evolution was observed above 300 ◦ C and it is delayed with respect to N2 . The delay may be due to the fact that the catalyst retains the oxygen produced by dissociative NO chemisorption. The evolution of CO2 allows the catalyst to continue to transfer oxygen from NO to carbon, generating new active sites on the carbon surface. The catalytic effect observed with addition of vanadium seems to be related with an enhancement of oxygen–carbon complexes and increase in CO2 evolution. The greater the rate of CO2 evolution the greater the NO conversion. Similar results were obtained by other authors [3–7]. Active phases play also an important role on the catalytic conversion of NO [4–8,20,21]. Figs. 6 and 7 show XRD patterns obtained in NO at several temperatures for carbon doped with V + Cu and V + Fe mixtures. The peaks show shifts to 2θ with increasing temperature, which reflects expansion of the crystal lattices. Pt peaks appear in the spectra resulting from exposure of the sample holder to the X-ray beam as carbon burned away.
233
In situ XRD shows clearly that metal oxides are reduced to lower oxidation states such as V2 O5 , V6 O13 ; CuO, Cu; Fe2 O3 , Fe3 O4 . Same phases were detected under inert atmosphere. No evidence of a new phase responsible for the synergetic effect was observed in the mixtures by in situ XRD since the same phases were observed both in the XRD patterns for the samples with two catalyst species and in the spectra for the samples bearing single catalyst species. However the addition of vanadium seems to increase carbon reactivity at low temperatures. As temperature increases some peaks become less intense suggesting that the catalyst is present as a melted phase. This is the case of Fe2 O3 and V2 O5 /V6 O13 . The catalytic effect observed for the NO conversion can be explained by the occurrence of redox processes in which the oxide particles are reduced by reaction with carbon at points of contact with the carbon substrate to form lower oxides according to a redox mechanism such as: y Mx + yNO → Mx Oy + N2 2 y y Mx Oy + C → Mx + CO2 2 2
Fig. 6. In situ XRD data obtained in NO on heating sample doped with Cu + V mixture at several temperatures.
234
S.A. Carabineiro et al. / Applied Catalysis B: Environmental 44 (2003) 227–235
Fig. 7. In situ XRD data obtained in NO on heating sample doped with Fe + V mixture at several temperatures.
The parallel reaction between NO and CO evolved from the carbon surface, as proposed for Pd or Co catalysts supported on activated carbon [22,23], is not likely to take place in the present reaction conditions, not only because CO is only detected in minor amounts (Fig. 5), but also because the carbon gasification rates have the same order of magnitude than the NO conversion rates (Tables 1 and 2). At 800 ◦ C the carbon gasification rate is even 3–5 times higher then the NO conversion rate. This is an indication that the parallel reaction between NO and CO does not take place in a significant extension. It is known that V2 O5 is a good catalyst for carbon gasification because of its low Tamman temperature. Various authors reported that V2 O5 has the ability to spread and wet carbon surfaces [15,16,19,25,26]. The reduction of V2 O5 to V6 O13 further reduces the melting point of the catalyst and could lead to more extensive spreading on carbon surface [15,16,19,25,26]. In situ XRD carried out both in NO and N2 atmospheres showed that both V2 O5 and V6 O13 are present under reaction conditions and with increasing temperature the peaks become less intense to both oxides suggesting that the catalyst is spreading to form an amorphous thin film on carbon surface. It has been reported that melting of the catalysts improves wetting of the
carbon surface promoting catalyst/carbon interaction [1,15,16,19–21]. Also Pan and Yang indicated that pitting and channelling of graphite was caused by the reaction between V2 O5 and carbon in which V2 O5 was reduced to V6 O13 . Also a liquid-like behaviour of the vanadium oxide particles was observed [25]. Thus the reduction of the metal oxide seems to be a key factor that controls the catalytic activity of the metal. Illán-Gomez et al. have reported that in an NO atmosphere a relationship exists between the ease of the oxidation–reduction reaction (to the metallic or a lower oxidation state) and the lattice energy of the oxide or the free energy for oxide formation [7]. In the same work the catalyst became inactive due to its conversion to the oxidised state and negligible reduction of the gas took place [7,8]. In this work it was shown by XRD that after pre-treatment at 500 ◦ C in inert atmosphere the catalysts were already reduced by carbon to lower oxidation states. It has been reported recently that alkali, alkaline-earth and transition metals enhance NO chemisorption. The catalytic system is more or less active depending on the ability of the oxide to be reduced by carbon and undergo redox reactions on the carbon surface [1,4–8,15,20,21]. The same mechanism has
S.A. Carabineiro et al. / Applied Catalysis B: Environmental 44 (2003) 227–235
been proposed to explain carbon catalysed reactions in several gaseous atmospheres such as CO2 reaction [7,8,15–18,20,21,24,27–30]. Another important factor is the catalyst dispersion. As dispersion increases the oxidation–reduction process of the catalysts is enhanced as the redox mechanism occurs at the catalyst/carbon surface. Thus in the case of the samples doped with vanadium, this interaction is promoted because the catalyst seems to be presented as a thin film, wetting the carbon surface. In situ XRD showed the probable species involved in the reaction conditions. It seems that a reduced catalyst surface is required for NO reduction. Active catalysts seem to act as an oxygen acceptor from NO, transferring it to the carbon surface and recovering reduced state. Similar behaviour has been reported by several authors [1,4,5,6,10,11,15,21,27,29].
4. Conclusions This study has indicated that kinetic measurements combined with in situ XRD are useful tools for interpreting catalyst behaviour. In the C–NO reaction a reduced catalyst surface is required to NO reduction. The ability of the catalyst to chemisorb NO going through redox transference of oxygen to the carbon reactive sites seems to explain catalytic activity. The synergetic effect observed with addition of vanadium seems to be related not only with its ability to wet and spread on the carbon surface but also with the enhancement of oxygen–carbon complexes. This last effect leads to the increase of CO2 evolution thus generating new active sites.
Acknowledgements Sónia Carabineiro is grateful to Fundação para a Ciˆencia e Tecnologia for Ph.D. Grant BD1275/95. References [1] R.T.K. Baker, C.R.F. Lund, J.J. Chlundzinski Jr., J. Catal. 87 (1984) 255.
235
[2] T. Okuhara, K. Tanaka, J. Chem. Soc. Faraday Trans. 1 (1986) 2357. [3] J. Rodriguez-Mirasol, A.C. Ooms, J.R. Pels, F. Kapteijn, J.A. Moulijn, Combust. Flame 99 (1994) 499. [4] M.J. Illán-Gomez, C. Salinas-Martinez de Lecea, A. Linares-Solano, in: Proceedings of the Extended Abstracts of Carbon’94, Granada, Spain, 1994, p. 460. [5] M.J. Illán-Gomez, A. Linares-Solano, L. Radovic, C. SalinasMartinez de Lecea, Energy Fuels 9 (1995) 97. [6] M.J. Illán-Gomez, A. Linares-Solano, L. Radovic, C. Salinas-Martinez de Lecea, Energy Fuels 9 (1995) 540. [7] M.J. Illán-Gomez, A. Linares-Solano, C. Salinas-Martinez de Lecea, Energy Fuels 9 (1995) 976. [8] M.J. Illán-Gomez, A. Linares-Solano, L. Radovic, C. SalinasMartinez de Lecea, Energy Fuels 10 (1995) 158. [9] J. Rodriguez-Mirasol, J.R. Pels, F. Kapteijn, J.A. Moulijn, in: Proceedings of the 22nd Biennial Conference on Carbon, San Diego, CA, USA, 1995, 620. [10] D. Mehandjiev, M. Khristova, E. Bekyarova, Carbon 34 (1996) 757. [11] C. Marquez-Alvarez, I. Rodrigues-Ramos, A. Guerrero-Ruiz, Carbon 34 (1996) 339. [12] S. Carabineiro, I.F. Silva, F.B. Fernandes, in: Proceedings of the 9th International Conference of Coal Science, Essen, Germany, 1997, p. 1103. [13] I. Aarna, E.M. Suuberg, Fuel 76 (1997) 475. [14] S.A. Carabineiro, I.F. Silva, F.B. Fernandes, A.M. Ramos, J. Vital, in: Proceedings of the Eurocabon’98, Strasbourg, France, 1998, p. 411. [15] D.W. McKee, Chemistry and Physics of Carbon, vol. 16, Marcel Dekker, New York, 1981, p. 1. [16] D.W. McKee, Carbon 8 (1970) 623. [17] I.F. Silva, L.S. Lobo, J. Catal. 126 (1990) 489. [18] I.F. Silva, D.W. McKee, L.S. Lobo, J. Catal. 170 (1997) 54. [19] I.F. Silva, C. Palma, M. Klimkiewicz, S. Eser, Carbon 36 (1998) 861. [20] S.A. Carabineiro, F.B. Fernandes, A.M. Ramos, J. Vital, I.F. Silva, Catal. Today 54 (1999) 559. [21] S.A. Carabineiro, F.B. Fernandes, A.M. Ramos, J. Vital, I.F. Silva, Catal. Today 57 (2000) 305. [22] M.A. Salas-Peregrin, M. Primet, H. Praliaud, Appl. Catal. B 8 (1996) 79. [23] D. Mehandjiev, E. Bekyarova, J. Colloid Interf. Sci. 166 (1994) 476. [24] S.A. Carabineiro, D.W. McKee, I.F. Silva, Carbon 39 (2001) 451. [25] Z. Pan, R. Yang, J. Catal. 130 (1991) 161. [26] X. Chu, L. Schmidt, S. Chen, R. Yang, J. Catal. 140 (1993) 543. [27] D.W. McKee, Fuel 59 (1980) 308. [28] B.J. Wood, K.M. Sancier, Catal. Rev.-Sci. Eng. 26 (1984) 233. [29] H. Ohme, T. Suzuki, Energy Fuels 10 (1996) 980. [30] F. Kapteijn, H. Porre, J.A. Moulijn, AIChE J. 31 (1986) 691.