- Email: [email protected]
Isocyanic acid hydrolysis over Fe-ZSM5 in urea-SCR
Catalysis Communications 7 (2006) 600–603 www.elsevier.com/locate/catcom
Isocyanic acid hydrolysis over Fe-ZSM5 in urea-SCR Gaia Piazzesi, Mukundan Devadas, Oliver Kro¨cher *, Martin Elsener, Alexander Wokaun Paul Scherrer Institute, CH-5232 Villigen PSI, Aargau, Switzerland Received 25 October 2005; received in revised form 20 January 2006; accepted 25 January 2006 Available online 6 March 2006
Abstract Two different iron exchanged zeolites were studied for the hydrolysis of isocyanic acid (HNCO) to ammonia, which is an important intermediate step in the selective catalytic reduction (SCR) of NO with urea. The hydrolysis reaction proceeded with very high activity and selectivity on over-exchanged as well as low-exchanged Fe-ZSM5 catalyst samples. The hydrolysis activity of the catalysts was inversely correlated with their Brønsted acidity, which decreased with the degree of iron exchange and the degree of ageing. Thus, overexchanged Fe-ZSM5 had the highest activity, which even increased when the catalyst was aged. Ó 2006 Elsevier B.V. All rights reserved. Keywords: HNCO hydrolysis; Fe-ZSM5; Urea-SCR; NOx
1. Introduction The selective catalytic reduction with ammonia (NH3 SCR) is widely used to reduce NOx emissions from stationary combustion processes [1]. This DeNOx technology is also suitable for diesel vehicles, but ammonia has been substituted by an aqueous urea solution (urea SCR) due to safety reasons [2]. In the hot exhaust gas of the diesel engine, injected urea is decomposed to isocyanic acid (HNCO) which is a stable compound in the gas phase. Therefore, a catalyst is required to hydrolyse the intermediate compound HNCO to the actual reducing agent ammonia in the humid exhaust gas. If no specific hydrolysis catalyst shall be used in front of the SCR catalysts, the SCR catalyst itself must exhibit high hydrolysis activity. For the well-established V2O5/WO3–TiO2 SCR catalysts it has actually been shown that their hydrolysis activity is much higher than their SCR activity [3], which allows a vehicle manufacturer to omit a specialized hydrolysis catalyst. However, in order to avoid problems with the decomposition of urea droplets on the SCR catalyst and to avoid *
Corresponding author. Tel.: +41 0 56 310 2066; fax: +41 0 56 310 2323. E-mail address: [email protected] (O. Kro¨cher). 1566-7367/$ - see front matter Ó 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.catcom.2006.01.022
slip of isocyanic acid through the catalyst, system designs including a specialised hydrolysis catalyst have been developed [4]. In recent years, Fe-ZSM5 was extensively investigated as potential alternative to vanadia based SCR catalysts [5]. However, since ammonia is exclusively used as reducing agent in laboratory investigations, nothing is known about the activity of this material for the preceding hydrolysis of HNCO in the urea SCR process. We studied the HNCO hydrolysis over Fe-ZSM5 in order to relate this reaction to the selective catalytic reduction of NO. 2. Experimental Two different Fe-ZSM5 samples were prepared by solid state ion exchange, starting from H-ZSM5 with Si/Al = 28. In one preparation method the ion exchange was performed in the typical way with FeCl2 in a ball mill (FeZSM51) [6] and in another method with FeCl3 in a mortar, called ‘‘mechanochemical’’ route (Fe-ZSM52) [7]. In both methods the zeolite to iron chloride weight ratio was 2:1. Fe-ZSM51 was prepared by mixing H-ZSM5 and FeCl2 in a ball mill for 1 h. The resulting powder was calcined at 550 °C for 5 h, washed with deionised water to remove the anions (AgNO3 test) and then dried at 100 °C for
G. Piazzesi et al. / Catalysis Communications 7 (2006) 600–603
3. Results and discussion The hydrolysis activity of the tested catalysts is shown in Fig. 1. For T < 300 °C the iron containing samples (FeZSM51 and Fe-ZSM52) had higher catalytic activity than the base material (H-ZSM5). 100% HNCO conversion to NH3 was observed at T P 200 °C for both Fe-ZSM51 and Fe-ZSM52, pointing out that even higher space velocities may be applied. Fe-ZSM51 was so active, that already at 150 °C over 70% conversion was reached. At this temperature, Fe-ZSM52 exhibited a conversion of 48% and H-ZSM5 of 27%. No formation of by-products (NO, NO2, N2O) was observed for all three catalysts over the entire temperature range, i.e. 100% selectivity to NH3 was reached. The DeNOx performance of the catalysts was measured at 10 ppm ammonia slip through the catalyst, which is regarded to be an acceptable emission level for the practical application. Fe-ZSM51 performed much better than the other catalysts with DeNOx values over 75% at T P 350 °C (Fig. 2). Fe-ZSM52 exhibited a DeNOx of 60% for
100
HNCO C onvers ion (%)
80
60
40
20
0 100
150
200
250
300
350
400
450
T/ ˚C Fig. 1. HNCO hydrolysis as a function of temperature over (j) FeZSM51, (m) Fe-ZSM52, (d) H-ZSM5, (h) aged Fe-ZSM51, and (s) aged H-ZSM5, coated on cordierite monoliths. SNH3 = 100%.
100
80
DeNOx (% )
10 h. Fe-ZSM52 was prepared by mixing H-ZSM5 and FeCl3 in a mortar for 1 h. The excess FeCl3 was washed out with water. Then the sample was dried at 100 °C for 10 h and finally calcined at 550 °C for 5 h. About 0.80 g of each powder sample and of the base material H-ZSM5 were coated on cordierite monoliths with a volume of 7.5 cm3. The coating procedure is described in [8] and the experimental setup and the procedure to generate isocyanic acid are published in [3]. The HNCO hydrolysis activity was studied from 150–450 °C at a GHSV of 52,000 h 1. The inlet flow was composed of 1000 ppm HNCO; 5% H2O; 10% O2; and balance N2. For investigations of the SCR reaction, the composition of diesel exhaust gas was approximated by a model feed gas containing 10% O2, 5% H2O, 1000 ppm of NOx and balance N2. Ammonia was added in the range 100–1000 ppm. The results are expressed in terms of NOx reduction efficiency (DeNOx). Elemental analysis by ICP-AES confirmed that the Si/Al ratio of the zeolite remained the same after ion exchange and thermal treatment. With the first method a high iron loading of 11.4 wt% was achieved for Fe-ZSM51, whereas the mechanochemical route results in 0.3 wt% of iron in Fe-ZSM52. The acidity of the samples was characterized by 27Al MAS NMR spectroscopy and NH3 TPD. NH3 TPD was carried out with 50 mg of sample in a TPD/ TPR 2900 analyser of Micromeritics. Adsorption of NH3 was carried out at 100 °C until saturation. Afterwards, the catalyst was flushed with He for 30 min and then TPD was performed up to 600 °C with a heating rate of 20 °C/min. 27Al MAS NMR was performed with a Bruker Ultrashield 500 spectrometer at a magnetic field of 11.7 T equipped with a 4 mm MAS head probe. The aluminium resonance frequency at this field is 130 MHz. The sample rotation speed was 12.5 kHz. The 27Al chemical shifts were referenced to a saturated Al(NO3)3 solution.
601
60
40
20
0 150
200
250
300
350
400
450
500
T / ˚C
Fig. 2. DeNOx at 10 ppm NH3 slip as a function of temperature. [NOx]inlet = 1000 ppm NO: (j) Fe-ZSM51, (m) Fe-ZSM52 and (d) HZSM5. [NOx]inlet = 500 ppm NO + 500 ppm NO2: (h) Fe-ZSM51, (n) Fe-ZSM52, and (s) H-ZSM5.
T P 400 °C, whereas H-ZSM5 showed almost no activity over the entire temperature range. Again, no formation of unwanted products was observed for all the three catalysts. These results go in parallel with those observed for the HNCO hydrolysis, where the catalyst activity showed the same order (Fe-ZSM51 > Fe-ZSM52 > H-ZSM5). With NO2 in the feed, very high SCR activities were found for Fe-ZSM51 and Fe-ZSM52 (Fig. 2), indicating that NO2 is required for the SCR reaction [5]. In the absence of NO2 in the feed, it has first to be produced from NO over the iron centers in the zeolite, which reduces the observed DeNOx values. The role of iron to oxidize NO to NO2 is also the reason for the inactivity of the base material H-ZSM5 for the SCR reaction with NO, as it has no oxidizing functionality, and its strongly enhanced SCR activity in the presence of NO2 in the feed (Fig. 2). In the further course of the SCR reaction the NO2 is
G. Piazzesi et al. / Catalysis Communications 7 (2006) 600–603
believed to react with the NH3 stored on both Lewis and Brønsted acid sites present on the surface [9]. With increasing iron content, the number of Brønsted acid sites present on the catalyst surface is decreased. This could be evinced by NH3 TPD experiments which resulted in a low temperature peak at 200–220 °C attributed to physisorbed ammonia [10] or ammonia from a non-zeolitic impurity [11] and a high temperature peak at 420–440 °C attributed to ammonia adsorbed on Brønsted acid sites [10] (Fig. 3A). The high temperature peak showed a decrease in Brønsted acidity in the order H-ZSM5 > FeZSM52 > Fe-ZSM51. The same order was found by 27Al MAS NMR spectroscopy, where the area of the signal at 55–60 ppm attributed to tetrahedrally coordinated alumina Al(OSiO)4 in the framework lattice [12] was used to quantify the Brønsted acidity of the sample (Fig. 3B). The signal at 0 ppm is due to aluminium in octahedral AlO6 groups [13]. The resonance around 55 ppm, which was very sharp for H-ZSM5 changed into a broad signal after Fe loading. This effect was more evident for Fe-ZSM51 since the amount of Fe loaded was higher. Apart from dealumination of the zeolite the broadening of the resonance signals at 0 ppm might also be due to dislodgement of framework iron, which dampens the Al signal because of its paramagnetic effect [14]. The comparison of the Brønsted acidity scale with the HNCO hydrolysis activity scale suggests that Brønsted acidity does not favour the HNCO hydrolysis, supported by the finding that HNCO adsorbs mainly on Lewis acid
(a) (b) (c) 100 150
200
250
300
350 400
450 500
550
600
T/ ˚ C
A
100 95 HNCO Conversion ( %)
602
90 85 80 75 70 65 60 100
150
200
250
300
350
400
450
500
T / ˚C Fig. 4. HNCO hydrolysis over (j) Fe-ZSM51, (n) 1% V2O5/WO3–TiO2, and (X) 3% V2O5/WO3–TiO2 powder samples. SNH3 = 100%.
sites of metal oxides [15,16]. This conclusion was also confirmed by ageing experiments. Additional 27Al MAS NMR investigations (not shown here for brevity) revealed that ageing of Fe-ZSM51 in a nitrogen flow with 5% H2O and 10% O2 at 650 °C for 5 h and of H-ZSM5 in air at 700 °C causes a decrease in Brønsted acidity for both samples. However, the HNCO conversion obtained with the aged samples was even higher than the conversion obtained with the fresh catalysts (Fig. 1). Therefore, we suggest that the hydrolysis activity is correlated with the Lewis acidity of the catalysts. In order to compare the hydrolysis activity of Fe-ZSM5 with conventional V2O5/WO3–TiO2 catalysts, Fe-ZSM51 was tested in powder form according to the procedure described in [3]. The results are plotted in Fig. 4 and compared with two V2O5/WO3–TiO2 catalysts containing 1% and 3% V2O5, respectively. At high temperatures all samples gave 100% HNCO conversion, but the samples started to differ below 350 °C. Fe-ZSM51 showed a higher hydrolysis activity than 3% V2O5/WO3–TiO2, but was slightly less active than 1% V2O5/WO3–TiO2. It is interesting to mention that the hydrolysis activity of V2O5/WO3– TiO2 decreased with increasing V2O5 contents. This activity order shows once more that Brønsted acidity does not favour the HNCO hydrolysis, since the Brønsted acidity of V2O5/WO3–TiO2 increases with the V2O5 loading [17,18].
(a) 4. Conclusions
(b)
(c) -200 B
-150
-100
-50
0
50
100
150
200
Chemical Shift (ppm)
Fig. 3. (A) NH3 TPD profiles and (B) 27Al MAS NMR spectra of (a) HZSM5, (b) Fe-ZSM52, and (c) Fe-ZSM51.
The combination of the excellent HNCO hydrolysis activity of Fe-ZSM5, which is even better after ageing, with its very good SCR performance, especially in the presence of NO2, makes it a highly interesting catalyst for the practical application in urea-SCR systems in diesel vehicles. In particular, Fe-ZSM5 prepared by solid state ion exchange showed the best catalytic properties. The high SCR activity of Fe-ZSM5 is correlated with the presence of iron, which
G. Piazzesi et al. / Catalysis Communications 7 (2006) 600–603
is required to oxidize NO to NO2. The hydrolysis activity of Fe-ZSM5 seems to be related to the Lewis acidity of the catalyst, but further investigations are being carried out to find out how the HNCO adsorption takes place on Fe-ZSM5 zeolites. Acknowledgements We thank Wacker-Chemie GmbH, Germany, and Su¨dChemie AG, Germany, for their financial support. D. Poduval (Schuit Institute of Catalysis, University of Eindhoven) is gratefully acknowledged for carrying out the 27Al MAS NMR measurements. References [1] P. Forzatti, Catal. Today 62 (2000) 51. [2] M. Koebel, M. Elsener, M. Kleemann, Catal. Today 59 (2000) 335. [3] M. Kleemann, M. Elsener, M. Koebel, A. Wokaun, Ind. Eng. Chem. Res. 39 (2000) 4120.
603
[4] E. Jacob, in: A. Leipertz, (Ed.), Proceedings of the Conference Motorische Verbrennung – aktuelle Probleme und moderne Lo¨sungsansa¨tze, in Berichte zur Energie- und Verfahrenstechnik, Heft 99.1, Essen, 1999. [5] R.Q. Long, R.T. Yang, J. Catal. 198 (2001) 20. [6] I.O.Y Liu, N.W. Cant, M. Ko¨gel, T. Turek, Catal. Lett. 63 (1999) 214. [7] F. Heinrich, C. Schmidt, E. Lo¨ffler, W. Gru¨nert, Catal. Commun. 2 (2001) 317. [8] M. Devadas, O. Kro¨cher, A. Wokaun, React. Kin. Catal. Lett. 86 (2005) 347. [9] G.T. Went, L.-J. Leu, R.R. Rosin, A.T. Bell, J. Catal. 134 (1992) 492. [10] N.Y. Topsøe, K. Pedersen, E.G. Derouane, J. Catal. 70 (1981) 41. [11] K. Chao, B. Chiou, C. Cho, S. Jeng, Zeolites 4 (1984) 2. [12] G. Engelhardt, D. Michel, High-Resolution Solid State NMR of Silicates and Zeolites, Wiley, New York, 1987. [13] J.A. van Bokhoven, D.C. Koningsberger, P. Kunkeler, H. van Bekkum, A.P.M. Kentgens, J. Am. Chem. Soc. 122 (2000) 12842. [14] J.P. Ramirez, J.C. Groen, A. Bru¨ckner, M.S. Kumar, U. Bentrup, M.N. Debbagh, L.A. Villaescusa, J. Catal. 232 (2005) 318. [15] F. Solymosi, T. Ba´nsa´gi, J. Catal. 156 (1995) 75. [16] G. Piazzesi, O. Kro¨cher, M. Elsener, A. Wokaun, Appl. Catal. B 65 (2006) 55–61. [17] C. Lin, H. Bai, Appl. Catal. B 42 (2003) 279. [18] N. Topsøe, J. Catal. 128 (1991) 499.
Isocyanic acid hydrolysis over Fe-ZSM5 in urea-SCR Gaia Piazzesi, Mukundan Devadas, Oliver Kro¨cher *, Martin Elsener, Alexander Wokaun Paul Scherrer Institute, CH-5232 Villigen PSI, Aargau, Switzerland Received 25 October 2005; received in revised form 20 January 2006; accepted 25 January 2006 Available online 6 March 2006
Abstract Two different iron exchanged zeolites were studied for the hydrolysis of isocyanic acid (HNCO) to ammonia, which is an important intermediate step in the selective catalytic reduction (SCR) of NO with urea. The hydrolysis reaction proceeded with very high activity and selectivity on over-exchanged as well as low-exchanged Fe-ZSM5 catalyst samples. The hydrolysis activity of the catalysts was inversely correlated with their Brønsted acidity, which decreased with the degree of iron exchange and the degree of ageing. Thus, overexchanged Fe-ZSM5 had the highest activity, which even increased when the catalyst was aged. Ó 2006 Elsevier B.V. All rights reserved. Keywords: HNCO hydrolysis; Fe-ZSM5; Urea-SCR; NOx
1. Introduction The selective catalytic reduction with ammonia (NH3 SCR) is widely used to reduce NOx emissions from stationary combustion processes [1]. This DeNOx technology is also suitable for diesel vehicles, but ammonia has been substituted by an aqueous urea solution (urea SCR) due to safety reasons [2]. In the hot exhaust gas of the diesel engine, injected urea is decomposed to isocyanic acid (HNCO) which is a stable compound in the gas phase. Therefore, a catalyst is required to hydrolyse the intermediate compound HNCO to the actual reducing agent ammonia in the humid exhaust gas. If no specific hydrolysis catalyst shall be used in front of the SCR catalysts, the SCR catalyst itself must exhibit high hydrolysis activity. For the well-established V2O5/WO3–TiO2 SCR catalysts it has actually been shown that their hydrolysis activity is much higher than their SCR activity [3], which allows a vehicle manufacturer to omit a specialized hydrolysis catalyst. However, in order to avoid problems with the decomposition of urea droplets on the SCR catalyst and to avoid *
Corresponding author. Tel.: +41 0 56 310 2066; fax: +41 0 56 310 2323. E-mail address: [email protected] (O. Kro¨cher). 1566-7367/$ - see front matter Ó 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.catcom.2006.01.022
slip of isocyanic acid through the catalyst, system designs including a specialised hydrolysis catalyst have been developed [4]. In recent years, Fe-ZSM5 was extensively investigated as potential alternative to vanadia based SCR catalysts [5]. However, since ammonia is exclusively used as reducing agent in laboratory investigations, nothing is known about the activity of this material for the preceding hydrolysis of HNCO in the urea SCR process. We studied the HNCO hydrolysis over Fe-ZSM5 in order to relate this reaction to the selective catalytic reduction of NO. 2. Experimental Two different Fe-ZSM5 samples were prepared by solid state ion exchange, starting from H-ZSM5 with Si/Al = 28. In one preparation method the ion exchange was performed in the typical way with FeCl2 in a ball mill (FeZSM51) [6] and in another method with FeCl3 in a mortar, called ‘‘mechanochemical’’ route (Fe-ZSM52) [7]. In both methods the zeolite to iron chloride weight ratio was 2:1. Fe-ZSM51 was prepared by mixing H-ZSM5 and FeCl2 in a ball mill for 1 h. The resulting powder was calcined at 550 °C for 5 h, washed with deionised water to remove the anions (AgNO3 test) and then dried at 100 °C for
G. Piazzesi et al. / Catalysis Communications 7 (2006) 600–603
3. Results and discussion The hydrolysis activity of the tested catalysts is shown in Fig. 1. For T < 300 °C the iron containing samples (FeZSM51 and Fe-ZSM52) had higher catalytic activity than the base material (H-ZSM5). 100% HNCO conversion to NH3 was observed at T P 200 °C for both Fe-ZSM51 and Fe-ZSM52, pointing out that even higher space velocities may be applied. Fe-ZSM51 was so active, that already at 150 °C over 70% conversion was reached. At this temperature, Fe-ZSM52 exhibited a conversion of 48% and H-ZSM5 of 27%. No formation of by-products (NO, NO2, N2O) was observed for all three catalysts over the entire temperature range, i.e. 100% selectivity to NH3 was reached. The DeNOx performance of the catalysts was measured at 10 ppm ammonia slip through the catalyst, which is regarded to be an acceptable emission level for the practical application. Fe-ZSM51 performed much better than the other catalysts with DeNOx values over 75% at T P 350 °C (Fig. 2). Fe-ZSM52 exhibited a DeNOx of 60% for
100
HNCO C onvers ion (%)
80
60
40
20
0 100
150
200
250
300
350
400
450
T/ ˚C Fig. 1. HNCO hydrolysis as a function of temperature over (j) FeZSM51, (m) Fe-ZSM52, (d) H-ZSM5, (h) aged Fe-ZSM51, and (s) aged H-ZSM5, coated on cordierite monoliths. SNH3 = 100%.
100
80
DeNOx (% )
10 h. Fe-ZSM52 was prepared by mixing H-ZSM5 and FeCl3 in a mortar for 1 h. The excess FeCl3 was washed out with water. Then the sample was dried at 100 °C for 10 h and finally calcined at 550 °C for 5 h. About 0.80 g of each powder sample and of the base material H-ZSM5 were coated on cordierite monoliths with a volume of 7.5 cm3. The coating procedure is described in [8] and the experimental setup and the procedure to generate isocyanic acid are published in [3]. The HNCO hydrolysis activity was studied from 150–450 °C at a GHSV of 52,000 h 1. The inlet flow was composed of 1000 ppm HNCO; 5% H2O; 10% O2; and balance N2. For investigations of the SCR reaction, the composition of diesel exhaust gas was approximated by a model feed gas containing 10% O2, 5% H2O, 1000 ppm of NOx and balance N2. Ammonia was added in the range 100–1000 ppm. The results are expressed in terms of NOx reduction efficiency (DeNOx). Elemental analysis by ICP-AES confirmed that the Si/Al ratio of the zeolite remained the same after ion exchange and thermal treatment. With the first method a high iron loading of 11.4 wt% was achieved for Fe-ZSM51, whereas the mechanochemical route results in 0.3 wt% of iron in Fe-ZSM52. The acidity of the samples was characterized by 27Al MAS NMR spectroscopy and NH3 TPD. NH3 TPD was carried out with 50 mg of sample in a TPD/ TPR 2900 analyser of Micromeritics. Adsorption of NH3 was carried out at 100 °C until saturation. Afterwards, the catalyst was flushed with He for 30 min and then TPD was performed up to 600 °C with a heating rate of 20 °C/min. 27Al MAS NMR was performed with a Bruker Ultrashield 500 spectrometer at a magnetic field of 11.7 T equipped with a 4 mm MAS head probe. The aluminium resonance frequency at this field is 130 MHz. The sample rotation speed was 12.5 kHz. The 27Al chemical shifts were referenced to a saturated Al(NO3)3 solution.
601
60
40
20
0 150
200
250
300
350
400
450
500
T / ˚C
Fig. 2. DeNOx at 10 ppm NH3 slip as a function of temperature. [NOx]inlet = 1000 ppm NO: (j) Fe-ZSM51, (m) Fe-ZSM52 and (d) HZSM5. [NOx]inlet = 500 ppm NO + 500 ppm NO2: (h) Fe-ZSM51, (n) Fe-ZSM52, and (s) H-ZSM5.
T P 400 °C, whereas H-ZSM5 showed almost no activity over the entire temperature range. Again, no formation of unwanted products was observed for all the three catalysts. These results go in parallel with those observed for the HNCO hydrolysis, where the catalyst activity showed the same order (Fe-ZSM51 > Fe-ZSM52 > H-ZSM5). With NO2 in the feed, very high SCR activities were found for Fe-ZSM51 and Fe-ZSM52 (Fig. 2), indicating that NO2 is required for the SCR reaction [5]. In the absence of NO2 in the feed, it has first to be produced from NO over the iron centers in the zeolite, which reduces the observed DeNOx values. The role of iron to oxidize NO to NO2 is also the reason for the inactivity of the base material H-ZSM5 for the SCR reaction with NO, as it has no oxidizing functionality, and its strongly enhanced SCR activity in the presence of NO2 in the feed (Fig. 2). In the further course of the SCR reaction the NO2 is
G. Piazzesi et al. / Catalysis Communications 7 (2006) 600–603
believed to react with the NH3 stored on both Lewis and Brønsted acid sites present on the surface [9]. With increasing iron content, the number of Brønsted acid sites present on the catalyst surface is decreased. This could be evinced by NH3 TPD experiments which resulted in a low temperature peak at 200–220 °C attributed to physisorbed ammonia [10] or ammonia from a non-zeolitic impurity [11] and a high temperature peak at 420–440 °C attributed to ammonia adsorbed on Brønsted acid sites [10] (Fig. 3A). The high temperature peak showed a decrease in Brønsted acidity in the order H-ZSM5 > FeZSM52 > Fe-ZSM51. The same order was found by 27Al MAS NMR spectroscopy, where the area of the signal at 55–60 ppm attributed to tetrahedrally coordinated alumina Al(OSiO)4 in the framework lattice [12] was used to quantify the Brønsted acidity of the sample (Fig. 3B). The signal at 0 ppm is due to aluminium in octahedral AlO6 groups [13]. The resonance around 55 ppm, which was very sharp for H-ZSM5 changed into a broad signal after Fe loading. This effect was more evident for Fe-ZSM51 since the amount of Fe loaded was higher. Apart from dealumination of the zeolite the broadening of the resonance signals at 0 ppm might also be due to dislodgement of framework iron, which dampens the Al signal because of its paramagnetic effect [14]. The comparison of the Brønsted acidity scale with the HNCO hydrolysis activity scale suggests that Brønsted acidity does not favour the HNCO hydrolysis, supported by the finding that HNCO adsorbs mainly on Lewis acid
(a) (b) (c) 100 150
200
250
300
350 400
450 500
550
600
T/ ˚ C
A
100 95 HNCO Conversion ( %)
602
90 85 80 75 70 65 60 100
150
200
250
300
350
400
450
500
T / ˚C Fig. 4. HNCO hydrolysis over (j) Fe-ZSM51, (n) 1% V2O5/WO3–TiO2, and (X) 3% V2O5/WO3–TiO2 powder samples. SNH3 = 100%.
sites of metal oxides [15,16]. This conclusion was also confirmed by ageing experiments. Additional 27Al MAS NMR investigations (not shown here for brevity) revealed that ageing of Fe-ZSM51 in a nitrogen flow with 5% H2O and 10% O2 at 650 °C for 5 h and of H-ZSM5 in air at 700 °C causes a decrease in Brønsted acidity for both samples. However, the HNCO conversion obtained with the aged samples was even higher than the conversion obtained with the fresh catalysts (Fig. 1). Therefore, we suggest that the hydrolysis activity is correlated with the Lewis acidity of the catalysts. In order to compare the hydrolysis activity of Fe-ZSM5 with conventional V2O5/WO3–TiO2 catalysts, Fe-ZSM51 was tested in powder form according to the procedure described in [3]. The results are plotted in Fig. 4 and compared with two V2O5/WO3–TiO2 catalysts containing 1% and 3% V2O5, respectively. At high temperatures all samples gave 100% HNCO conversion, but the samples started to differ below 350 °C. Fe-ZSM51 showed a higher hydrolysis activity than 3% V2O5/WO3–TiO2, but was slightly less active than 1% V2O5/WO3–TiO2. It is interesting to mention that the hydrolysis activity of V2O5/WO3– TiO2 decreased with increasing V2O5 contents. This activity order shows once more that Brønsted acidity does not favour the HNCO hydrolysis, since the Brønsted acidity of V2O5/WO3–TiO2 increases with the V2O5 loading [17,18].
(a) 4. Conclusions
(b)
(c) -200 B
-150
-100
-50
0
50
100
150
200
Chemical Shift (ppm)
Fig. 3. (A) NH3 TPD profiles and (B) 27Al MAS NMR spectra of (a) HZSM5, (b) Fe-ZSM52, and (c) Fe-ZSM51.
The combination of the excellent HNCO hydrolysis activity of Fe-ZSM5, which is even better after ageing, with its very good SCR performance, especially in the presence of NO2, makes it a highly interesting catalyst for the practical application in urea-SCR systems in diesel vehicles. In particular, Fe-ZSM5 prepared by solid state ion exchange showed the best catalytic properties. The high SCR activity of Fe-ZSM5 is correlated with the presence of iron, which
G. Piazzesi et al. / Catalysis Communications 7 (2006) 600–603
is required to oxidize NO to NO2. The hydrolysis activity of Fe-ZSM5 seems to be related to the Lewis acidity of the catalyst, but further investigations are being carried out to find out how the HNCO adsorption takes place on Fe-ZSM5 zeolites. Acknowledgements We thank Wacker-Chemie GmbH, Germany, and Su¨dChemie AG, Germany, for their financial support. D. Poduval (Schuit Institute of Catalysis, University of Eindhoven) is gratefully acknowledged for carrying out the 27Al MAS NMR measurements. References [1] P. Forzatti, Catal. Today 62 (2000) 51. [2] M. Koebel, M. Elsener, M. Kleemann, Catal. Today 59 (2000) 335. [3] M. Kleemann, M. Elsener, M. Koebel, A. Wokaun, Ind. Eng. Chem. Res. 39 (2000) 4120.
603
[4] E. Jacob, in: A. Leipertz, (Ed.), Proceedings of the Conference Motorische Verbrennung – aktuelle Probleme und moderne Lo¨sungsansa¨tze, in Berichte zur Energie- und Verfahrenstechnik, Heft 99.1, Essen, 1999. [5] R.Q. Long, R.T. Yang, J. Catal. 198 (2001) 20. [6] I.O.Y Liu, N.W. Cant, M. Ko¨gel, T. Turek, Catal. Lett. 63 (1999) 214. [7] F. Heinrich, C. Schmidt, E. Lo¨ffler, W. Gru¨nert, Catal. Commun. 2 (2001) 317. [8] M. Devadas, O. Kro¨cher, A. Wokaun, React. Kin. Catal. Lett. 86 (2005) 347. [9] G.T. Went, L.-J. Leu, R.R. Rosin, A.T. Bell, J. Catal. 134 (1992) 492. [10] N.Y. Topsøe, K. Pedersen, E.G. Derouane, J. Catal. 70 (1981) 41. [11] K. Chao, B. Chiou, C. Cho, S. Jeng, Zeolites 4 (1984) 2. [12] G. Engelhardt, D. Michel, High-Resolution Solid State NMR of Silicates and Zeolites, Wiley, New York, 1987. [13] J.A. van Bokhoven, D.C. Koningsberger, P. Kunkeler, H. van Bekkum, A.P.M. Kentgens, J. Am. Chem. Soc. 122 (2000) 12842. [14] J.P. Ramirez, J.C. Groen, A. Bru¨ckner, M.S. Kumar, U. Bentrup, M.N. Debbagh, L.A. Villaescusa, J. Catal. 232 (2005) 318. [15] F. Solymosi, T. Ba´nsa´gi, J. Catal. 156 (1995) 75. [16] G. Piazzesi, O. Kro¨cher, M. Elsener, A. Wokaun, Appl. Catal. B 65 (2006) 55–61. [17] C. Lin, H. Bai, Appl. Catal. B 42 (2003) 279. [18] N. Topsøe, J. Catal. 128 (1991) 499.