Transient Kinetics of SO2 Oxidation Over SCR-DeNOx Monolith

This work was performed under contract with ENEL Ricerca−Polo termico (Pisa, Italy). The authors thank Dr. Natale Ferlazzo for his valuable contribu...
0 downloads 0 Views 77KB Size
Download PDF
Ind. Eng. Chem. Res. 1999, 38, 2593-2598

2593

Transient Kinetics of SO2 Oxidation Over SCR-DeNOx Monolith Catalysts Enrico Tronconi, Andrea Cavanna, Carlo Orsenigo, and Pio Forzatti* Dipartimento di Chimica Industriale e Ingegneria Chimica “G. Natta”, Politecnico di Milano, Piazza L. da Vinci 32, 20133 Milano, Italy

A dynamic kinetic study of the undesired oxidation of SO2 over a commercial V2O5-WO3/TiO2 monolith catalyst for the SCR-DeNOx process is herein presented. Transient SO2 conversion data were systematically collected during experiments involving step changes of temperature, area velocity, and feed composition (SO2, O2, H2O, and NH3). Characteristic times of the system response were of a few hours: peculiar SO3 emission peaks were apparent upon increasing the reaction temperature and the H2O feed content. The data were well interpreted by a dynamic kinetic model assuming that buildup-depletion of surface sulfate species is rate controlling. The model accounts successfully also for the transient effects resulting from the interaction between the SCR-DeNOx reaction (NOx reduction with NH3) and the oxidation of SO2. Introduction The selective catalytic reduction (SCR) of NOx with NH3 over V2O5-WO3/TiO2 monolith catalysts is widely used worldwide to control the emissions of nitrogen oxides in the flue gases from power plants and stationary sources.1-3 In the case of sulfur-bearing fuels, the unfavorable oxidation of SO2

SO2 + 1/2 O2 f SO3 occurs simultaneously with the DeNOx reaction.4-6 The present commercial SCR catalysts are highly selective, yielding typical SO2 conversions of 1-2%. Even such small conversions are detrimental, however, because SO3 reacts with NH3 to form ammonium bisulfate (NH4HSO4) which may then deposit both on the catalyst surface, causing catalyst deactivation, and on other downstream parts of the SCR plant, causing corrosion problems and pressure drop increments. Deposition of NH4HSO4 occurs below a critical temperature which depends on the local gas-phase concentrations of NH3 and SO3:7 thus, it is of interest to model accurately the spatial and temporal evolutions of such concentrations. Notably, SCR-DeNOx reactors are often involved in transient operations associated, e.g., with startup, shutdown, or load variations: under such conditions peaks of SO3 emissions have been reported.8 The unsteady-state behaviors of NOx reduction9-11 and SO2 oxidation12 over V2O5-WO3/TiO2 catalysts have been systematically studied in our laboratory, both during conditioning of commercial SCR catalysts under a SO2-containing atmosphere and upon step changes of the operating conditions. It was found that the characteristic response time is as long as a few hours in the case of the SO2 oxidation, being controlled by the buildup of sulfates on the catalyst surface, whereas it is much shorter for the DeNOx reaction. A dynamic kinetic model of SO2 oxidation over SCR monolith catalysts is herein presented, based on a detailed mechanism of the catalyst sulfate coverage * Corresponding author. E-mail: [email protected]. Fax: (++39)-02-7063 8173. Tel.: (++39)-02-2399 3238.

which accounts for the interaction with NOx reduction. The model is validated against transient experiments both in the absence of the DeNOx reaction and with its simultaneous occurrence. Experimental Section The data presented in this work refer to a commercial low-dust SCR monolith honeycomb V2O5-WO3/TiO2 catalyst with medium V loading (0.6% w/w as V2O5), with V being homogeneously distributed across the whole thickness of the catalyst walls; the WO3 loading was about 10% w/w. The BET analysis resulted in a specific surface area of 63 m2/g and a pore volume of 0.28 cm3/g in the pore radius range 15-300; a monomodal pore size distribution was revealed by the Hg penetration method. Monolith samples with nine channels, 15 cm in length, were cut from commercial modules, wrapped with quartz wool, and forced into the test reactor to prevent bypass. The pitch and the wall thickness of the tested monolith catalyst were ≈4.95 and ≈0.8 mm, respectively. The experimental apparatus used for the transient kinetic runs has been described in detail elsewhere.4 The experimental plan was designed in order to cover the dynamic effects associated with step changes of the major operating variables, namely, temperature (T), area velocity (AV), and feed concentrations of SO2, H2O, and O2, over an experimental field representative of industrial SCR operation. The following variable ranges were investigated: T ) 350-380 °C, C°SO2 ) 160-1917 ppm, H2O feed content ) 3.39.7% v/v, O2 feed content ) 0.1-5% v/v, AV ) 5-7.8 Nm3/(m2 h). In line with the SCR technical literature, AV is herein defined as the ratio of the volumetric feed flow rate to the geometric surface area of the catalyst: in our runs the above AV values implied feed flows in the range 1900-2900 Ncm3/min (at 273 K, 1 atm). In the runs with simultaneous occurrence of SO2 oxidation and NOx reduction, NO and NH3 feed concentrations in the range 200-400 ppm were also used. The SO3 content of the outlet gases was determined by condensing sulfuric acid at 90 °C in a glass spiral, followed by off-line analysis with an ionic chromatograph Dionex

10.1021/ie980673e CCC: $18.00 © 1999 American Chemical Society Published on Web 05/25/1999

2594 Ind. Eng. Chem. Res., Vol. 38, No. 7, 1999

Model Quick. The mean sampling time was 45 min, much shorter than the typical system response time; sampling was more frequent during fast transients. Results and Discussion Derivation of the Dynamic Kinetic Model. Assumptions and Reaction Scheme. In previous papers4-6 we have systematically investigated the effects of the operating conditions, feed composition, and catalyst design parameters in the steady-state oxidation of SO2 to SO3 over honeycomb DeNOxing catalysts. It has been shown that: (i) the catalyst active sites for SO2 oxidation are not the same active sites for the DeNOx reaction, (ii) the slow SO2 oxidation reaction occurs in a chemical regime and is not limited by diffusional resistances, (iii) SO2 reacts with gaseous oxygen on a free active site, originating an adsorbed SO3 molecule, kSO

2

SO2 + 1/2O2 + ϑf98ϑSO3

(1)

(iv) SO3 undergoes desorption-readsorption on the active sites, kdes

ϑSO3 79 8SO3 + ϑf k ads

(2)

(v) there is competitive adsorption of water from the gas phase onto the catalyst active sites,

H2O + ϑf S ϑH2O

(3)

(vi) if ammonia is present, it can react with adsorbed SO3 to give a surface ammonium sulfate, which blocks further desorption of SO3,

NH3 + ϑSO3 S ϑSO3,NH3

(4)

and (vii) adsorption-desorption dynamics of water and ammonia are much faster than SO3 desorption, whose characteristic time is on the order of several hours. Under such assumptions, four types of catalytic sites are identified: (i) sites occupied by SO3 alone (also referred to as sites with free SO3), (ii) sites occupied by SO3 and ammonia (ammonium sulfates), (iii) sites occupied by water, (iv) free sites. The corresponding surface coverages are indicated in the following as ϑSO3,f, ϑSO3,NH3, ϑH2O, and ϑf, respectively. Sites occupied by SO3 alone and by SO3 plus NH3 can be collected as SO3 sites, with overall surface coverage. The site balance is then given by the following equations:

ϑH2O + ϑSO3,f + ϑSO3,NH3 + ϑf ) 1

(5)

ϑSO3,f + ϑSO3NH3 ) ϑSO3

(6)

Assuming adsorption-desorption equilibrium of H2O, the water surface coverage ϑH2O is calculated from eq 7, where the fraction of SO3-occupied sites is assumed to be negligible in comparison to ϑH2O as the H2O gasphase concentration is greater than the SO3 concentration by at least 3 orders of magnitude:

ϑH2O )

KH2OCH2O 1 + KH2OCH2O

(7)

In eq 7 KH2O is an H2O adsorption equilibrium constant.

Furthermore, considering equilibrium adsorption of NH3 on SO3-occupied sites, the volume-averaged fraction of SO3-occupied sites free from adsorbed ammonia, ϑ h SO3,f, can be computed as a function of the overall fraction of SO3-occupied sites according to the following equation, as outlined in the Appendix:

ϑ h SO3,f )

ϑSO3

(8)

W KNH3,SO3CNH 3

ΦNH3 On an integral basis, eq 8 accounts for the variation of both CNO and CNH3 along x, the transverse coordinate across the monolith wall thickness, due to strong intraporous concentration gradients caused by the fast W is the NH3 concentraDeNOx reaction.11 In eq 8, CNH 3 tion at the catalyst wall and ΦNH3 a suitable Thiele modulus correcting for intraporous diffusional resistances.11 Rate Expressions. According to previous steady-state results,4,5 the SO2 oxidation rate, eq 9, is first order in SO2 concentration, exhibits a fractional order in O2, and accounts for the promoting effect of NOx and the inhibiting effects of water and ammonia RO2 -RW CH2O kSO2CSO2CO 2 rSO2 ) [1 + kNOCNO(x)] 1 + KNH3CNH3(x)

(9)

Again, notice that both CNO and CNH3 vary along x, the transverse coordinate across the monolith wall thickness. Thus, eq 9 provides the local rate: the overall rate of reaction (rjSO2) is obtained upon integrating this expression over the wall half thickness of the catalytic monolith, with the overbar representing an average quantity. Along similar lines, eqs 10 and 11 provide the average

h SO3 - ϑH2O) rads,SO3 ) kads,SO3CSO3(1 - ϑ rjdes,SO3 ) kdes,SO3ϑ h SO3,f ) kdes,SO3

(10)

ϑ h SO3 W 1 + KNH3,SO3CNH /ΦNH 3

3

(11) SO3 adsorption and desorption rates, respectively, where in eq 10 a linear dependence on the gaseous SO3 concentration and on the fraction of free sites has been adopted and, making use of eq 8, eq 11 includes a linear dependence on the fraction of free SO3 sites. Again, the overbars indicate that eqs 10 and 11 represent average rates of SO3 adsorption and desorption over the catalyst wall half-thickness. Mass balance Equations. Given the very limited SO2 conversion (≈1%), the SO2 concentration can be regarded as constant along the whole reactor. Water is in large excess, and its concentration is considered constant, too. So only unsteady mass balances for gaseous and adsorbed SO3 are required, which take into account however the distributions of NH3 and NOx both along the monolith axial coordinate z and across the monolith walls, with the latter gradients resulting from the intraporous diffusional resistances. The gas-phase SO3 mass balance yields accordingly

Ind. Eng. Chem. Res., Vol. 38, No. 7, 1999 2595

∂CSO3 ∂t

)-

vz ∂CSO3 1- - ΩSO3 (rjads,SO3 - rjdes,SO3) L ∂z*  (12)

whereas eq 13 provides the mass balance for adsorbed SO3 In eqs 12 and 13 ΩSO3 is the catalyst capacity for

∂ϑ h SO3 ∂t

) rjads,SO3 - rjdes,SO3 + rjSO2

(13)

SO3 adsorption,  the void fraction of the honeycomb monolith, vz is the linear flow velocity in the monolith channels, L the channel length and z* ) z/L a dimensionless axial coordinate. The overbars imply that the rates are averaged over the catalyst wall thickness at every axial location z. Relevant initial and boundary conditions are

at t ) 0 at z ) 0

0 0 h SO (z); CSO3 ) CSO (z) ϑ h SO3) ϑ 3 3

CSO3(t) ) 0

Table 1. Parameter Estimates, with 95% Confidence Limits, for the Unsteady Kinetic Model of SO2 Oxidation over a Commercial Honeycomb SCR-DeNOx Catalyst in the Absence of NOx Reductiona Ω′SO3 ) Ω(1 - )/ ) 8470 ( 2313 kad ) (2.16 ( 0.754) × 10-3 kdes ) (2.08 ( 1.50) × 10-4 exp[-(11750 ( 1331)(1/T - 1/653)] kSO2 ) (8.02 ( 0.476) × 10-6 exp[-(5550 ( 534)(1/T - 1/653)] KH2O ) 0.828 ( 0.048 Rox ) 0.064 ( 0.018 Rw ) 0.211 ( 0.071 aIn

[mol of SO3/m3 of gas] [m3 of gas /mol/s] [1/s] [(mgas3/mol)1+aO2-Rw/s] [m3/mol]

the expressions for kdes and kSO2, T is in degrees Kelvin.

(14) (15)

The model equations were solved numerically by polynomial approximation13 along the axial coordinate z, while the integration in time was carried out by the library routine LSODI.14 Data Fit. Transient Runs with SO2 Oxidation Only. A set of 13 transient kinetic runs of SO2 oxidation was carried out in the absence of the DeNOx reaction, i.e., with no NO and NH3 in the reactor feed stream. In this case the parameters kNO, KNH3, and KNH3,SO3 included in the unsteady kinetic model, eqs 5-15, were not significant, being related to the presence of NO and NH3 in the reaction system. The nine remaining adaptive parameters were estimated by global nonlinear regression, using 164 values of the experimental response, namely, the outlet SO3 concentration. The parameter estimates are reported in Table 1 along with 95% confidence limits. For proper evaluation of the regression results, one should consider that (a) the parameter estimates rely on 155 degrees of freedom and (b) the nine parameters are used to describe the T-dependent dynamic influence of five independent variables on the experimental response. The largest absolute value of the off-diagonal terms in the correlation matrix was 0.948, indicating that the parameters are correlated to some extent, though not dramatically: this prevents discussion of the individual parameter estimates. We may notice however that (a) the estimated SO3 adsorption capacity is consistent with a monolayer coverage of the catalyst, (b) the relative magnitudes of kad and kdes agree with previous indications that SO3 desorption is much slower than readsorption,12 (c) the activation energy of SO3 desorption (≈98 kJ/mol) is in line with the expected strong interaction of SO3 with the catalyst surface, (d) the activation energy of the surface reaction (SO2 oxidation) (≈46 kJ/mol) is close to the apparent activation energy estimated over this catalyst from steadystate data (∼60 kJ/mol), (e) the small value of the kinetic order Rox results from operation under O2 excess, and (f) the extent of H2O inhibition is consistent with previous steady-state data.5 The goodness of fit was generally satisfactory. The average absolute error was 5.7% on the experimental response, namely, the outlet SO3 concentration, which is certainly in line with the corresponding experimental uncertainty. As an a posteriori check, the results of the

Figure 1. Experimental and calculated temporal evolution of the SO3 outlet concentration upon step changes of the reaction temperature. Reaction conditions: C°SO2 ) 1278 ppm, AV ) 7.8 Nm3/(m2 h), C°O2 ) 2.6% v/v, C°H2O ) 12.8% v/v.

transient runs were simulated also by considering the integral of the outlet SO3 concentration over each sampling interval as the experimental response rather than the instantaneous value. Negligible deviations were observed, indicating that the sampling intervals were sufficiently short not to interfere with the fitting of the transient response. Figures 1-5 illustrate the measured and fitted temporal evolutions of the gaseous SO3 concentration at the reactor outlet during different types of transient experiments. The effects of positive and negative step changes of the reaction temperature are shown in Figure 1. A reduction of T from 380 to 360 °C caused an initial, sudden decrease of the outlet SO3 concentration, followed however by a slower recovery of the SO3 emission level. Such a characteristic response is related to the accumulation of sulfates onto the catalyst: in fact, a temperature reduction leads to a decrease of the rate of SO2 oxidation but to a more significant decrease of the rate of SO3 desorption in view of the activation energies of the two reactions. The net result is a buildup of sulfates at the catalyst surface, responsible for the observed negative overshoot of the SO3 concentration level. In the following stages of the transient, however, the system slowly approached a new ultimate value of SO3 concentration through reequilibration of the rates of SO3 generation, desorption, and readsorption, resulting eventually in a greater SO3 surface coverage at steady state. This behavior was confirmed in the case of a further T reduction to 350 °C. When the temperature was suddenly raised to 370 °C, on the contrary, a peak of SO3 emission was observed, resulting from decomposition of the surface sulfates, followed by a slow

2596 Ind. Eng. Chem. Res., Vol. 38, No. 7, 1999

Figure 2. Experimental and calculated temporal evolution of the SO3 outlet concentration upon step changes of the area velocity. Reaction conditions: T ) 380 °C, C°SO2 ) 1278 ppm, C°O2 ) 2.6% v/v, C°H2O ) 12.8% v/v.

Figure 4. Experimental and calculated temporal evolution of the SO3 outlet concentration upon step changes of the H2O concentration. Reaction conditions: T ) 380 °C, C°SO2 ) 1278 ppm, AV ) 7.8 Nm3/(m2 h), C°O2 ) 2.6% v/v.

Figure 3. Experimental and calculated temporal evolution of the SO3 outlet concentration upon step changes of the SO2 feed concentration. Reaction conditions: T ) 380 °C, AV ) 7.8 Nm3/ (m2 h), C°O2 ) 2.6% v/v, C°H2O ) 12.8% v/v.

Figure 5. -Experimental and calculated temporal evolution of the SO3 outlet concentration upon step changes of the O2 concentration. Reaction conditions: T ) 380 °C, C°SO2 ) 1278 ppm, AV ) 7.8 Nm3/(m2 h), C°H2O ) 12.8% v/v.

decline of CSO3 toward the new stationary value associated with a smaller sulfate coverage. The approach to steady state required over 10 h in all cases. The system response to step changes of the feed flow rate is shown in Figure 2. Again we observed a sudden response of the SO3 emission followed by a much slower approach to the new steady state. The fast initial transient was associated with the response of the gasphase concentrations, whereas the longer subsequent evolution resulted from the buildup/depletion dynamics of the surface sulfates. The negative step change of AV led in fact to a higher gaseous concentration of SO3, which led in turn to a higher ultimate coverage. The situation was reversed upon restoring the initial AV value of 7.8 Nm3/(m2 h). Along similar lines one can interpret the system behavior observed during positive or negative step changes of the SO2 feed concentration (Figure 3), of the H2O feed concentration (Figure 4), and of the O2 feed concentration (Figure 5). In all such cases, in fact, the step change induces a perturbation in the rates of SO3 formation, desorption, and readsorption, which is immediately reflected by the gas-phase SO3 concentration. The response of the SO3 coverage, however, is considerably slower: the approach to a new steady state, in fact, involves either the accumulation of sulfates, as, e.g., in the case of positive step changes of CSO2 and CO2 and

negative step changes of CH2O, or their decomposition, as in the case of the opposite step changes. In view of the measurement precision, the accuracy of the dynamic kinetic model in reproducing all of the observed trends was acceptable. Particularly, the occurrence of SO3 emission peaks upon step increments of the reaction temperature (Figure 1) and of the H2O feed content (Figure 4) was successfully predicted. SO3 emission peaks were reported by Gutberlet et al.8 for an industrial SCR installation following a load increase of the power station from 250 to 450 MW and a corresponding increase of the operating temperature from 340 to 390 °C. Figures 1-5 indicate that typical SO3 transients extend over a few hours. Again, this roughly corresponds to the duration of the transient in SO3 emissions reported by Gutberlet et al.8 in industrial practice (over 5 h). Notably, the parameter estimates in Table 1 assess that SO3 desorption is the rate-determining step of the whole process, controlling the buildup or depletion of sulfate species on the catalyst surface. Transient kinetics of SO2 oxidation with simultaneously occurring DeNOx reaction-SO2 oxidation kinetics was addressed also in the presence of the DeNOx reactants, though only in a limited number of transient experiments. For this purpose, the previous estimates of the adaptive parameters (Table 1) were retained, and

Ind. Eng. Chem. Res., Vol. 38, No. 7, 1999 2597

Figure 6. Experimental and calculated temporal evolution of the SO3 outlet concentration upon step changes of the NH3 feed concentration. Reaction conditions: T ) 380 °C, C°SO2 ) 1278 ppm, AV ) 7.8 Nm3/(m2 h), C°H2O ) 12.8% v/v, C°O2 ) 2.6% v/v, C°NO ) 200 ppm.

the additional kinetic parameters kNO and KNH3 in eq 9 were estimated from data obtained under steady-state conditions.5 The NH3 equilibrium constant of adsorption over sulfated sites, KNH3,SO3, was simply set to a large value (106) because of the lack of adequate data for its accurate estimation. This is in line with the expectation that ammonia adsorption onto sulfated sites is largely favored and was found adequate to reproduce qualitatively the observed trends of the data. The model equations for SO3 gas- and solid-phase balances were then included in the existing 1D dynamic model of the SCR-DeNOx monolith reactor,11 in order to couple the simulations of the temporal evolution of NOx reduction and SO2 oxidation. Eventually, the complete dynamic kinetic model was validated for simultaneously occurring DeNOx and SO2 oxidation reactions by predictive comparison with transient data following step changes of the NH3 inlet concentration. Figure 6 shows one of such comparisons for the temporal evolution of the SO3 outlet concentration profile: both model and data exhibit a maximum in the SO3 emission after a sudden drop in the feed concentration of ammonia. This response is associated with the release of SO3 originally blocked as surface ammonium sulfates. The same effect shown in Figure 6 was observed also during transients involving changes of the NH3 feed over other commercial catalysts. Although a quantitative comparison of the model predictions with data was prevented in these cases because of the unknown reaction kinetics over the tested catalysts, a qualitative agreement was always obtained. Notably, the introduction of SO2 oxidation dynamics does not significantly increment the computational load for numerical model solution with respect to the dynamic model of the DeNOx reaction only. Accordingly, the simulation of typical SCR reactor transients can still be carried out in a fraction of the actual transient time.11 This is a prerequisite for application to predictive control systems of SCR reactors. Conclusions The unsteady kinetics of the undesired SO2 oxidation reaction has been investigated over commercial SCRDeNOx catalysts. On the basis of previous results concerning the mechanism and the active sites for this reaction, a kinetic model has been derived which ac-

counts satisfactorily for the observed transient effects resulting from step changes of the main variable settings in the absence of NOx and NH3. The treatment has then been extended to account also for the simultaneous occurrence of the DeNOx reaction, which interacts with the mechanism of SO2 oxidation. Accordingly, the kinetic influence of the major reactants in the SCR process has been introduced both in the rate expression for SO2 oxidation (promoting action of NO, inhibition due to NH3) and in that for SO3 desorption (formation of surface ammonium sulfates). The peculiar transients observed upon varying the NH3 feed content can thus be reproduced. The occurrence of SO3 emission peaks following either sudden increments of temperature and H2O feed content or sudden drops of the NH3 feed concentration is worth noting. When coupled with a previous dynamic model of NOx reduction in monolith catalysts,11 the present kinetic treatment completes the description of the unsteady behavior of industrial SCR reactors with respect to both the DeNOx reaction and the parasite SO2 f SO3 reaction: combined predictions of NOx reduction efficiency, NH3 slip, and SO3 emission levels for typical industrial operating transients are thus provided, corresponding to changes in reactor load, flue-gas temperature, and feed composition. Finally, the model computing times are consistent with its use in predictive systems for optimized control of the off-gas emissions. Acknowledgment This work was performed under contract with ENEL Ricerca-Polo termico (Pisa, Italy). The authors thank Dr. Natale Ferlazzo for his valuable contributions to the present work. Notation AV ) area velocity [Nmgas3/(m2 h)]. Ci ) gas-phase concentration of species i [mol/mgas3] dh ) hydraulic diameter of the monolith channels [m] kads ) rate constant for SO3 adsorption [mgas3/mol/s] kdes ) rate constant for SO3 desorption [1/s] kNO ) parameter for NO promotion, eq 9 [mgas3/mol] kSO2 ) rate constant for SO2 oxidation [(mgas3/ (mol)1+RO2-Rw/s] KH2O ) equilibrium constant for H2O adsorption, eq 3 [mgas3/ mol] KNH3 ) parameter for NH3 inhibition, eq 9 [mgas3/mol] KNH3, SO3 ) equilibrium constant for NH3 adsorption on sulfated sites, eq 4 [mgas3/mol] L ) monolith length [m] rads,NH3,SO3 ) rate of NH3 adsorption on sulfated sites [s-1] rdes,NH3,SO3 ) rate of NH3 desorption from sulfated sites [s-1] rads,SO3 ) rate of SO3 adsorption [s-1] rdes,SO3 ) rate of SO3 desorption [s-1] rSO2 ) rate of SO2 oxidation [s-1] t ) time [s] T ) temperature [K] vz ) gas linear velocity [m/s] x ) dimensionless monolith transverse (intraporous) coordinate z ) monolith axial coordinate z* ) z/L, dimensionless monolith axial coordinate Greek Letters Ro2, Rw ) kinetic orders for O2 and H2O

2598 Ind. Eng. Chem. Res., Vol. 38, No. 7, 1999 ϑ ) surface coverage ΦNH3 ) Thiele modulus for NH3 adsorption Ω ) catalyst SO3 adsorption capacity [mol/mcat3] b ) bulk gas conditions eff ) effective rate f ) free sites g ) gas phase s ) solid phase W ) conditions at the monolith wall ° ) conditions at the reactor inlet

Literature Cited

In view of the difference in time scales between the DeNOx process and the SO2 oxidation,6,12 we introduce a pseudo-steady-state assumption for adsorption-desorption of NH3 on sulfated sites,

0 ≈ rads,NH3,SO3 - rdes,NH3,SO3

(A1)

Then we express the local rates as follows:

rads,NH3,SO3 ) rads,NH3CNH3(x) ϑSO3,f

(A2)

rdes,NH3,SO3 ) kdes,NH3ϑSO3,NH3

(A3)

where CNH3(x) indicates that the ammonia concentration varies along the intraporous coordinate x because of the strong diffusional limitations associated with the DeNOx reaction. Introducing eq 6 and defining KNH3,SO3 ) kads,NH3/ kdes,NH3, we obtain

KNH3,SO3CNH3(x) ϑSO3,f ) ϑSO3 - ϑSO3,f

(A4)

Introducing average coverages, represented by overbars, and integrating eq A4 over the catalyst wall halfthickness, we get

∫01CNH (x) dxj ) ϑh SO

3

-ϑ h SO3,f (A5)

Then we make use of the approximate intraporous NH3 concentration profile adopted by Tronconi et al.11 in their dynamic study of the SCR-DeNOx reaction over commercial V2O5-WO3/TiO2 catalysts, W CNH3(x) ) CNH 3

(A7)

Finally, eq 8 is derived from eq A7 by considering h NH3 . 1 and that, under typical SCR conditions,11 Φ tanh(ΦNH3) ≈ 1.

Appendix: Derivation of Equation 8

3

W CNH 3

1 + KNH3SO3 tanh(ΦNH3) ΦNH3

Superscripts

KNH3,SO3 ϑ h NH3,f

ϑ h SO3

ϑ h SO3,f )

cosh(ΦΝΗ3x) cosh(ΦΝΗ3)

(1) Bosch, H.; Janssen, F. Catalytic Reduction of Nitrogen Oxides. Catal. Rev. 1988, 2, 369-531. (2) Nakajima, F.; Hamada, I. The state-of-the-art technology of NOx control. Catal. Today 1996, 29, 109-115. (3) Forzatti, P.; Lietti, L. Recent Advances in De-NOxing Catalysis for Stationary Applications Heterog. Chem. Rev. 1996, 3, 33-51. (4) Svachula, J.; Ferlazzo, N.; Forzatti, P.; Tronconi, E.; Bregani, F. Oxidation of SO2 to SO3 over Honeycomb DeNOxing Catalysts. Ind. Eng. Chem. Res. 1993, 32, 826-834. (5) Tronconi, E.; Beretta, A.; Elmi, A. S.; Forzatti, P.; Malloggi, S.; Baldacci, A. A Complete Model of SCR Monolith Reactors for the Analysis of Interacting NOx Reduction and SO2 Oxidation Reactions. Chem. Eng. Sci. 1994, 49, 4277-4287. (6) Orsenigo, C.; Beretta, A.; Forzatti, P.; Svachula, J.; Tronconi, E.; Bregani, F.; Baldacci, A. Theoretical and experimental study of the interaction between NOx reduction and SO2 oxidation over DeNOx SCR catalysts. Catal. Today 1996, 27, 15-21. (7) Matsuda, S.; Kamo, T.; Kato, A.; Nakajima, F. Deposition of Ammonium Bisulfate in the Selective Catalytic Reduction of Nitrogen Oxides with Ammonia. Ind. Eng. Chem. Res. 1982, 21, 48-52. (8) Gutberlet, H.; Dieckmann, A.; Merz, A.; Schreiber, L. SO2Konversionrate von DENOx-Katalysatoren. VGB Kraftwerkstech. 1990, 70, 959-968. (9) Tronconi, E.; Lietti, L.; Forzatti, P.; Malloggi, S. Experimental and Theoretical Investigation of the Dynamics of the SCRDeNOx Reaction. Chem. Eng. Sci. 1996, 51, 2965-2970. (10) Lietti, L.; Nova, I.; Camurri, S.; Tronconi, E.; Forzatti, P. Dynamics of the SCR-DeNOx Reaction by the Transient Response Method. AIChE J. 1997, 43, 2559-2570. (11) Tronconi, E.; Cavanna, A.; Forzatti, P. Unsteady Analysis of NO Reduction over Selective Catalytic Reduction-DeNOx Monolith Catalysts. Ind. Eng. Chem. Res. 1998, 37, 2341-2349. (12) Orsenigo, C.; Lietti, L.; Tronconi, E.; Forzatti, P.; Bregani, F. Dynamic investigation of the role of surface sulfates in NOx reduction and SO2 oxidation over V2O5-WO3/TiO2 catalysts. Ind. Eng. Chem. Res. 1998, 37, 2350-2359. (13) Finlayson, B. Nonlinear Analysis in Chemical Engineering; McGraw-Hill: New York, 1980. (14) Hindmarsh, A. C. Odepack: a systematized collection of ODE solvers. In Scientific Computing; Stepleman, R. S., et al., Eds.; North-Holland: Amsterdam, The Netherlands, 1983.

(A6)

W where CNH is the NH3 concentration at the gas/solid 3 interface and ΦNH3 a suitable Thiele modulus.11 Solving the integral in eq A5 and rearranging, we get

Received for review October 28, 1998 Revised manuscript received February 24, 1999 Accepted March 23, 1999 IE980673E