Catalysis Today,17 (1993)131-140 Publishers B.V.,Amsterdam
Surface and catalytic properties of Vanadia-Titania and Tungsta-Titania systems in the Selective Catalytic Reduction of nitrogen oxides L. Liettia, J. Svachulaa, P. Forzattia, G. Buscab, G. Ramisb and F. BreganiC a Dipartimento di Chimica Industriale ed Ing. Chimica "G. Natta", Politecnico di Milano, P.zza L. da Vinci 32, 20133 Milan0 (Italy) b Istituto di Chimica, Facolth di Ingegneria, P.le Kennedy, 16129 Genova (Italy) c CRTN-ENEL, Via Rubattino 54, 20134 Milan0 (Italy) Abstract The chemistry of the Selective Catalytic Reduction (SCR) of with ammonia over VZ05-Ti02 and W03-Ti02 catalysts has NoX been investigated by Fourier-Transform Infrared spectroscopy, Temperature Programmed Desorption, Temperature Programmed catalytic Reaction and activity measurements. Surface coordinated and molecularly ammonia species, Protonated associated with Brijnsted and Lewis acid sites respectively and with different thermal stability, are observed upon NH show stronger Lewis and BrZjnsted acl*a adsorption. W03-TiO sites than *;b Upon heating in NO atmosphere, a reaction betwVe2eOnSeT1 adsorbed 2' ammonia and gas-phase or weakly adsorbed NO is monitored by FT-IR and TPSR measurements. The results indicate that V205-TiO2 is significantly more active than W03-Ti02, due to its superior redox properties. Monomeric vanadyls and meta-vanadate polymers are proposed as the active sites in the SCR reaction, the former sites showing lower reactivity. Oxygen is involved in the reaction and plays a crucial role in determining the reactivity of the catalysts. 1. INTRODUCTION The Selective Catalytic used for the control of plants. thermal power homogeneous mixture of vanadium oxide along with as mechanical promoters. Many efforts have been vanadia-titania DeNO ing mechanism of NO reduction
Reduction of NO, by NH3 is widely NO, emission in stack gases from Commercial consist of catalysts anatase TiO tungsten oxide and minor amoun2' s of silica-alluminates devoted to the characterization of catalysts and to elucidate the by NH3, but the debate is still open
0920-5861/93/$6.00 0 1993Elsevier Science PublishersB.V. Allrightsreserved.
concerning the nature of the reactive surface species and the role of catalyst acid-base and redox properties. In the present paper results obtained from Fourier-Transform Infrared (FT-IR) spectroscopy, Temperature Programmed Desorption (TPD), and Temperature Programmed Surface Reaction (TPSR) experiments of ammonia adsorbed over vanadia-titania and tungsta-titania oxide catalysts and catalytic data collected over honeycomb commercial SCR catalysts are reported and discussed with the aim to clarify the chemistry of the process and the role of the different catalyst's components. 2. EXPERIMENTAL The preparation and characterization of the catalysts used in this study as well as details on the FT-IR experimental setup and procedures can be found elsewhere [1,2]. TPD and TPSR experiments were performed in a quartz fixed bed microreactor (I.D. 7 mm) connected with a quadrupole mass spectrometer (UT1 model 1OOC). 200 mg of the catalyst (40-60 mesh) where loaded in the reactor and oxidized in He + 20% 0 at 773 K for 30 min. Then the catalyst was cooled down to 313 K and saturated with an He + 0.2% NH3 stream for 30 min. After 1 h He purge at the adsorption temperature, the catalyst was linearly heated un to 773 K at 15 K/min in He (TPD) , _ or in He + 600 ppm_NO (TPSR): Catalytic activity tests on monolithic catalysts under SCR conditions were performed in an experimental apparatus described elsewhere . 3. RESULTS AWD DISCUSSION 3.1. FT-IR study of NH3 adsorption
Figure 1 show the spectra of the adsorbed species upon contact of ammonia with W03-Ti02. The sharp band at 1603 cm-l is due to the asymmetric deformation of NH3 coordinatively held over Lewis acid sites. However, the corresponding symmetric deformation mode is clearly complex showing the main maximum at 1215 cm-l which shifts progressively to 1240 cm-l upon heating in vacuum, while a shoulder near 1280 cm-l is evident only at the highest coverages. The frequency of the symmetric deformation mode is higher than in the case of NH coordinative1 held on V205-TiOz (1230 cm-l) and TiO2 (117a and 1215 cm- 9 ), while approxima es that observed on pure W03 (1275 and 1222 cm-l). These data and the strong perturbation of the W=O fundamental and first overtone stretching bands after adsorption of ammonia indicate that wolframyl species act as Lewis acid sites for ammonia coordination. From these results, as well as from isopropanol decomposition experiments it appears that the Lewis sites of W03-Ti02 are stronger [II, than those of V205-Ti02.
Figure 1. FT-IR spectra of the adsorbed species arising from contact of WO3(W=6.2% w/w)-Ti02 with ammonia (50 Torr) and following evacuation at 300 K (a), 400 K (b), 450 K (c), and 520 K (d). The band observed at 1660 cm-l(weak) and 1440 cm-l (strong) are characteristic of ammonium ions produced by ammonia protonation over WOH Bronsted acid sites. These sites are also found on pure W03( while they are not observed on TiO . Outgassing at increasing temperatures causes 8 irst the desorption of ammonia from Bronsted sites, and then from Lewis sites. This indicates the greater thermal stability of the molecularly adsorbed NH3 than of the corresponding protonated species. Worth of note is that upon heating in vacuum a shoulder appears at 1485 cm-l, that can be due to a splittin of the degenerate asymmetric NH4 deformation mode (1440 cm-? ) or to the formation of a new species. A possible assignment may be the NH2 scissoring mode of an amide species, similar to that detected at 1550 cm-1 on V205-Ti02  and at 1485 cm-l on Mo03-Ti02 . When the same experiment is carried out in the presence of NO gas (instead of vacuum), the behaviour of adsorbed ammonia species looks completely different (Figure 2). In this case the bands of the coordinatively held species disappear fast (shoulder at 1280 and band at 1215 cm-l) while those of ammonium ions (band at 1440 cm-l) at first grow and then disappear only partially. Moreover water is formed starting from 423 K (bands at 1625, 3200 and 3600 cm-l). This may eventually indicate that a reaction has occurred between adsorbed ammonia and NO producing water and N2, which is not
Figure 2. FT-IR spectra of the adsorbed species arising from with ammonia (50 Torr) and contact of WO3(W=6.2% w/w)-TiO following contact with NO gas ($00 Torr) at 300 K (a), 373 K (b), 400 K (c), 420 K (d), 450 K (e), and 520 K (f). detectable. The reaction appears to involve coordinated ammonia which is activated via the amide species. However the participation of the protonated species cannot be excluded due to the possible interconversion of one species into the other. Indeed adsorption of water over an ammonia-covered surface formed by reaction of showed that ammonium ions are coordinated ammonia species with water. On the other hand, coordinatively held species may originate from ammonium ions via ammonia desorption from Brijnstedacid sites and subsequent adsorption over Lewis acid sites. Formation of water is representative of the occurrence of the SCR reaction: indeed no water has been observed during adsorption and subsequent heating in the absence of NO (see Figure l), whereas formation of water via oxidation of NH occurred only at higher T (573 K) upon adsorption of NH3 and 3heating in oxygen. NO does not produce any detectable adsorbed species over W03TiO2 when ammonia is preadsorbed. Only after desorption or reaction of ammonia, NO adsorbed species become detectable (surface nitrosyls and traces of nitrates). The nitrosyl species are very weakly bonded to the surface and desorb immediately by outgassing at r.t.. The above data are similar to the results previously reported concerning analogous experiments carried out over a V205-Ti02 catalyst [Z]. However, the following differences can be stressed: i) W03-Ti02 shows stronger Lewis and Brijnsted acid sites than
V205-TiO2 (see above); ii) the reaction between adsorbed ammonia and gas-phase (or weakly adsorbed) NO is observed at lower temperature on V 05TiO2 than on W03-Ti02; this compares well with the cata1yst activities in flow reactor and to the TPSR data shown below; iii) the bands attributed to the amide species are more evident and are detected at lower temperatures on V205-Ti02 than on W03-Ti02. 3.2. TPD of NH3 and TPSR of NH3 with NO FT-IR measurements have been performed over V205(V=5% w/w)Ti02 and WO (W=6% w/w)-TiO2. While the loading of W03 closely approaches t3hose usually employed in commercial SCR catalysts, that of vanadia is much greater. In order to allow for a comparison with catalysts with a lower vanadia content the TPD-TPSR study was performed on VJ.~S% 0 -TiO wtw. samples with different vanadia loadings down to V =
Figure 3. TPSR in NO (solid lines) and TPD (dotted lines) of ammonia over 2% V205-Ti02.
Figure 4. TPSR in NO (solid lines) and TPD (dotted lines) of ammonia over 6% W03-Ti02.
136 The results of TPD (dotted lines) and TPSR (solid lines) experiments performed over V205(V=Z% w/w)-TiO2 and WO (W=6% w/w)-TiO2 are reported in Figures 3 and 4, respective3y. In the case of the vanadia-titania catalyst (Figure 3), TPSR data indicate that a significant amount of adsorbed ammonia is consumed by reaction with NO. The SCR reaction occurs to a significant extent already from T = 360 K. Both the peak of N2 corresponding negative the perfectly (and of H20) and consumption peak of NO show two distinct maxima at 473 K and 590 K. The reaction is observed up to 650 K till complete ammonia. In contrast, over adsorbed consumption of a NO consumption peak with a at 590 K has been observed , corresponding to the higher temperature peak of the higher V20 loading catalyst. Tl?e different behaviour of the two vanadia-titania catalysts can be explained by considering that different vanadia species are present on the catalyst surface. In line with previous literature data [6-83, both isolated vanadyls species (IR and Raman band near 1030 cm-l on dry sample) and polymeric metavanadate type species (band at 940 cm-l) have been detected. It is reported that the relative abundance of isolated vanadyls decreases on increasing the V-loadings and that isolated vanadyls are less active than polymeric meta-vanadate groups [6,7]. Accordingly the low temperature TPSR peak of associated to meta-vanadate can be peak centred at T= 590 K and observed over bdth 2% and 0.78% V205-Ti02 is likely associated with monomeric vanadyls. The SCR reaction has been monitored also over the tungstatitania catalyst (Figure 4), but at higher temperatures than over vanadia-titania. In fact consumption of NO and NH (and evolution of N2 and water) occurs above 550 K. This ora er of reactivity agrees with FT-IR data and with the different reactivity of the two catalysts in oxidation reactions. This points to the major role of the catalyst redox properties (i.e. the ability to undergo easily redox cycles) in the SCR reaction. It is also worth of note that W03-Ti02 is a stronger solid acid (and a stronger adsorbant for ammonia) than V205Ti02. To assess the role of gas-phase oxygen and the nature of the active sites on the V 05-TiO been carried out on 25 and a.7"8"%'a~~~~~T~~~c~~~~~~~~n~~s~a~~ the presence of oxygen (2% v/v). As repor ed in Figure 5 for the V2O5(2% w/w)-Ti02 catalyst, only one peak for both NO consumption and N2 formation are evident. NO consumption is monitored already starting from 323 K with a maximum near 460 K. Also in the case of 0.78% V205-Ti02 catalyst the temperature of the peak is markedly lowered in the presence of oxygen (T =540K vs. 590K) , in line with flow reactor data (reported%elow) obtained over low-vanadia loading catalysts. TPSR data clearly indicate a direct participation of oxygen in the reaction, in line with previous literature results and
137 with the different reaction stoichiometries observed in the presence and in the absence of oxygen 121. Furthermore, results show that also in the presence of oxygen the the high-vanadia reactivity of loading catalyst is significantly higher than that of the sample with a lower vanadia loading. This confirms the higher reactivity of metavanadate species with respect to isolated vanadyls. However, results of TPSR experiments reported in Figure 3 and Temperature Programmed Reaction data [S] further indicate that meta-vanadate species are less selective than isolated vanadyls, since formation of significant amounts of N20 have been observed in the case of the 2% V OS-TiO catalyst at high temperatures and in the presence o f2 gas-phase oxygen. This indicates that the greater reactivity of meta-vanadate species with respect to isolated vanadyls is associated by a much lower selectivity in the formation of N2.
3.3. Flow reactor experiments Figure 6 shows the effects of oxygen and water in the SCR reaction performed over honeycomb comme cial catalysts. In agreement with TPSR data, an incre$se in NO conversion is observed by increasing oxygen concentration up co 24% because limited by the catalyst redox is the overall process properties. The results compare well with data collected by several other authors over catalysts in the powder form [9,10,11]. Figure 6 shows that water inhibits the reaction. A decrease in NO, conversion is apparent at low water contents, which however levels off above 5% v/v water Icontent. This indicates that NO, conversion is independent o water content in the concentration range of practical int rest (S-15% v/v). The inhibition of water may be explainEd either in terms of competition with ammonia for (water is a less strong base than ammonia but it also present in much greater concentration) or in of the active sulphated vanadyl sites . 4. CONCLUSIONS
On the basis of the data presented1 above and of the data already published concerning the characterization of W03-Ti02 [l] and V205-W03-Ti02  the following conclusions can be derived: i) monomeric vanadyls and meta-vanadate species are likely po;ymeri;,;p;P,P,ecigerse,"t', present ~~rh~~5-~_i~~,~~l~~~ly~~~ present activity but lower selectivity in the SCR reaction than monomeric vanadyls; ii) the redox properties of the catalyst are the major factor
Figure 5. TPSR of NH3 in NO+02 over 2% V205TiO2
Figure 6. Effect of 03 and H20 concentration on NO, conversion over commercial honeycomb catalyst. NO, = 500 ppm, SO = 500 ppm, 02= 2% v/v (curve b 5 , H20 = 10% v/v (curve a), N2 balance, T=593 K.
governing catalyst activity; this eventually accounts for the higher reactivity of V205-TiOZ if compared to W03-Ti03 catalysts; iii) tungsta-titania is characterized by a larger number and by stronger Lewis and Bronsted acid sites than vanadiatitania: this eventually ensures higher surface ammonia coverages even at high temperatures (larger T window for the reaction); iv) the mechanism of the reaction appears to involve coordinated ammonia which is activated via the amide species and gas-phase (or weakly bonded) NO. However the participating of the protonated species cannot be excluded due to the possible interconversion of the two species. 5.
G. Ramis, G. Busca, C. Cristiani, L. Lietti, P. Forzatti and F. Bregani, Langmuir, 8 (1992) 1744. G. Ramis, G. Busca, F. Bregani and P. Forzatti, Appl, Catal., 64 (1990) 259. J. Svachula, N. Ferlazzo, P. Forzatti, E. Tronconi and F. Bregani, Ind. Eng. Chem. Res., submitted.
H. Miyata, T. Mukai, T. Ono and Y. Kubokawa, J. Chem. Faraday Trans. I, 84 (1988) 4137. Manuscript in preparation G.T. Went, L.-J. Leu and A.T. Bell, J. Catal., 134 (1992) 479. L.-J. Leu, R.R. Rosin and A.T. Bell, J. Catal., sot.,
5 6 7
8 9 10 11 12 13
C. Cristiani, P. Forzatti and G. Busca, J. Catal., 116 (1989) 586; G. Ramis, C. Cristiani, P. Forzatti and G. Busca, J. Catal., 124 (1990) 574. M. Inomata, A. Miyamoto, Y. Murakami, J. Catal., 62 (1980) 140. M. Inomata, A. Miyamoto, T. Ui, K. Kobayashi, and Y. Murakami, Ind. Eng. Chem. Prod. Res. Dev. 21 (1982) 424. H. Bosch, F.J.J.G. Janssen, F.H.G. Van der Kerkhof, J. Oldenzie, J.G. Van Ommen, J.R.H. Ross, Appl. Catal., 25 (186) 239. J. Svachula, L.J. Alemany, N. Ferlazzo, P. Forzatti, E. Tronconi and F. Bregani, Ind. Eng. Chem. Res., submitted. G. Ramis, G. Busca and P. Forzatti, Appl. Catal. B, 1 (1992) L9.