A Kinetic Model for the Selective Catalytic Reduction of NO

Ind. Eng. Chem. Res. 2010, 49, 39–52

39

A Kinetic Model for the Selective Catalytic Reduction of NOx with NH3 over an Fe-zeolite Catalyst Hanna Sjo¨vall,† Richard J. Blint,‡ Ashok Gopinath,§ and Louise Olsson*,† Chemical Reaction Engineering and Competence Centre for Catalysis, Chalmers UniVersity of Technology, SE-412 96 Go¨teborg, Sweden, General Motors R&D Center, Chemical and EnVironmental Sciences Laboratory, 30500 Mound Rd, Warren, Michigan 48090-9055 and General Motors R&D India Science Lab, Creator Building, 2nd Floor, ITPB, Whitefield Road, Bangalore 560066, India

The selective catalytic reduction of NOx with ammonia over an Fe-zeolite catalyst was investigated experimentally and a transient kinetic model was developed. The model includes reactions that describe ammonia storage and oxidation, NO oxidation, selective catalytic reduction (SCR) of NO and NO2, formation of N2O, ammonia inhibition and ammonium nitrate formation. The model can account for a broad range of experimental conditions in the presence of H2O, CO2, and O2 at temperatures from 150 to 650 °C. The catalyst stores ammonia at temperatures up to 400 °C and shows ammonia oxidation activity from 350 °C. The catalyst is also active for the oxidation of NO to NO2 and the oxidation reaches equilibrium at 500 °C. The SCR of NO is already active at 150 °C and the introduction of equal amounts of NO and NO2 greatly enhances the conversion of NOx at temperatures up to 300 °C. The formation of N2O is negligible if small fractions of NO2 are fed to the reactor, but a significant amount of N2O is formed at high NO2 to NO ratios. An ammonia inhibition on the SCR of NO is observed at 200 °C. This kinetic model contains 12 reactions and is able to describe the experimental results well. The model was validated using short transient experiments and experimental conditions not used in the parameter estimation and predicted these new conditions adequately. 1. Introduction Pollutants such as nitrogen oxides (NOx) are formed during the combustion of fossil fuels, and there are several possible ways to reduce these pollutants. Two major techniques have been developed that are able to reduce NOx in a lean environment, one in which NOx may be stored and reduced in a cyclic operation, and one in which the nitrogen oxides undergo selective catalytic reduction (SCR) in a continuous operation by the addition of either hydrocarbons (HC-SCR) or ammonia (NH3-SCR). For a vehicle application, ammonia is produced from the decomposition of urea (NH2-CO-NH2). High conversions of NOx can be achieved at a wide temperature range using ammonia SCR over various types of catalysts, such as catalysts containing vanadia,1-4 zeolite-based catalysts,5-12 or combinations of selected catalysts.13 One of the most extensively studied catalyst for this application is the iron exchanged zeolite which have shown to be a highly efficient catalyst for NOx depletion.14-20 An efficient SCR catalyst facilitates the reduction of NO and NO2 with NH3 and suppresses other side reactions such as NH3 oxidation. The reaction between NO and NH3 occurs in the presence of oxygen over the catalyst.1,2,4,21-24 The NOx reduction activity increases if NO and NO2 are fed in about equimolecular amounts, compared to if only NO is present.1,2,21,22 Reduction of NO2 without the presence of NO is also possible, but the overall reaction between NO2 and NH3 is not clear. It is however often presented as a global reaction without oxygen.1,23,24 The role of NO2 in the SCR of NOx with NH3 has been extensively studied and several reaction pathways are proposed in the literature. It has been suggested that in the absence of NO2, the oxidation of NO is the rate determining step on * To whom correspondence should be addressed. E-mail: louise. [email protected]. Fax: +46-31-772 3035. † Chalmers University of Technology. ‡ General Motors R&D Center. § General Motors R&D India Science Lab.

H-ZSM-5 catalysts.25,26 This was also concluded by Devadas et al.,16 who observed that zeolite catalysts require NO2 in the feed gas, or an activity for oxidation of NO to attain high SCR activity. Komatsu et al.5 present a mechanism for copperexchanged zeolites which involves the formation of a bridging NO3. The nitrate reacts with NO to form NO2 which reacts further with NH3 to produce N2 and H2O. Sun et al.14 conclude in their work on Fe/MFI catalysts that the preferred path for NOx reduction with ammonia occurs via ammonium nitrite which decomposes to N2 and H2O. Several suggested mechanisms for the reduction of NO and NO2 by NH3 involve formation of HNO2 and HNO3 and include reaction steps similar to the ones involved in the nitric acid chemistry27 as well as in the HNO2 atmospheric chemistry.28-30 Yeom et al.31 also include ammonium nitrite formation in a more detailed reaction scheme over a BaNa-Y catalyst. This mechanism involves the formation of both HNO2 and HNO3 which may react independently with ammonia to produce NH4NO2 and NH4NO3, respectively. Ammonium nitrite decomposes readily to N2 and H2O. Ammonium nitrate is more stable and was suggested to either decompose to NH3 and HNO3 or react with NO to form ammonium nitrite and NO2. Nitric oxide was able to react with a number of additional species that facilitates formation of N2 and H2O in this reaction scheme. A similar mechanism was also suggested by Nova et al.32 and Ciardelli et al.3 for vanadiabased catalysts. This mechanism includes a reaction pathway for the fast SCR reaction that involves formation of ammonium nitrite and nitrate via HNO2 and HNO3. In a recent paper, Grossale et al.33 suggest a reaction scheme for the fast SCR chemistry over an Fe-ZSM-5 catalyst, which stresses the key role of the reaction between nitrates and NO. There are several kinetic models that describe the selective catalytic reduction of NOx over various zeolite-based catalysts. Baik et al.34 developed a model that describes the reduction of NO with urea at steady-state conditions over a Cu-ZSM-5 catalyst. Other steady-state models have been developed to

10.1021/ie9003464  2010 American Chemical Society Published on Web 11/20/2009

40

Ind. Eng. Chem. Res., Vol. 49, No. 1, 2010

account for the reduction of NO with NH3 over a Cu-faujasite catalyst35 and a H-ZSM-5 catalyst.25 Models that describe transient catalytic effects have also been developed. Malmberg et al.36 presented a transient kinetic model for the SCR of NO over an Fe-zeolite catalyst and Chatterjee et al.37 developed a model for a commercial zeolite SCR catalyst. This model is based on kinetics developed in earlier papers by the same group for a vanadia-based catalyst38-41 and accounts for the NH3 SCR of NO and the SCR of NOx mixtures containing NO to NO2 ratios up to 1. We present, in a recent paper, a kinetic model for a Cu-ZSM-5 catalyst, which describes the transient selective catalytic reduction of NOx at various mixtures of NO and NO2.42 In a series of three papers we also present a more detailed kinetic model for the SCR of NOx for the same catalyst.43-45 This detailed model describes the ammonia storage and oxidation, the formation of nitrites and nitrates on the surface, and the reduction of NOx via intermediate reaction steps. However, there are no kinetic models for the Fe-zeolite catalyst that consider NOx mixtures with excess NO2 and the formation of N2O and that is valid at temperatures up to 650 °C. The objective of this work was to develop a global transient kinetic model for a commercial Fe-zeolite catalyst that accounts for ammonia storage, ammonia oxidation, NO oxidation, the SCR of NO and mixtures of NO and NO2 including excess NO2, ammonia inhibition, and N2O formation. The model describes the catalytic activity in various feed mixtures at temperatures from 150 to 650 °C. 2. Experimental Section Flow reactor experiments were carried out over a commercial Fe-zeolite catalyst provided by Johnson Matthey. The catalyst used was a washcoated monolith, with length and diameter of 22 mm. The cell density was 400 cpsi. The catalyst was placed in a flow reactor which consisted of a horizontal quartz tube

equipped with a heating coil and insulation. Two thermocouples connected to a Eurotherm instrument were used to control and measure the temperature. One thermocouple was placed in front of the catalyst to control the gas phase temperature, and a second thermocouple was placed inside the sample to measure the catalyst temperature. All gases except H2O were introduced to the reactor via an Environics 2000 gas mixer. The water was added downstream of the mixed gases in an evaporator which was heated to 150 °C. The amount of H2O added to the feed was controlled by exerting pressure on the container holding the water. The total flow rate was 3500 mL/min, which corresponds to a space velocity of about 25 000 h-1. All experiments were performed at atmospheric pressure, and the carrier gas was argon. The outlet gases were analyzed by a MultiGas 2030 HS FTIR instrument. The catalyst was aged for 16 h at 700 °C in a flow of 1000 mL/min containing 20% O2, 5% CO2, 5% H2O, 250 ppm NO, 250 ppm NH3, and 28 ppm SO2 prior to the experimental study. The catalyst was exposed to 8% O2 in Ar for 20 min at 650 °C prior to each experiment. Temperature-programmed desorption experiments (TPD) were carried out storing ammonia in the presence of H2O and CO2. The catalyst was exposed to 510 ppm NH3, 5% H2O, and 5% CO2 for 40 min at 159 °C, followed by a 60 min period of 5% H2O in argon, and a temperature ramp (10 °C/min) in a feed containing 5% H2O in argon. A second TPD experiment was carried out in a similar way. Ammonia was stored in the presence of 5% H2O and 5% CO2 for 40 min at 210 °C, followed by a 60 min period of 5% H2O in argon, and a temperature ramp (10 °C/min) in a feed containing 5% H2O in argon. The ammonia oxidation was investigated at temperatures from 150 to 650 °C. The feed contained 510 ppm NH3, 8% O2, 5% H2O, and 5% CO2. The temperatures investigated were 150, 200, 250, 300, 350, 400, 500, and 650 °C and each temperature was kept for 20 min. The NO oxidation was also studied at

Table 1. Summary of Experiments experiment NH3 TPD NH3 TPD NH3 oxidation NO oxidation NH3 SCR of NO NH3 SCR of NOx (75% NO, 25% NO2) NH3 SCR of NOx (50% NO, 50% NO2) NH3 SCR of NOx (25% NO, 75% NO2) NH3 SCR of NO (high NH3 conc) NH3 SCR of NO (high NH3 conc) NH3 inhibition “total” ammonia storage

“dynamic” ammonia storage transient NO SCR experiments transient NOx SCR experiments

composition

temperature

storage: 510 ppm NH3, 5% H2O and 5% CO2 desorption: 5% H2O storage: 510 ppm NH3, 5% H2O and 5% CO2 desorption: 5% H2O 510 ppm NH3, 8% O2, 5% H2O and 5% CO2 495 ppm NO, 15 ppm NO2, 8% O2, 5% H2O and 5% CO2 510 ppm NH3, 495 ppm NO, 15 ppm NO2, 8% O2, 5% H2O and 5% CO2 510 ppm NH3, 380 ppm NO, 130 ppm NO2, 8% O2, 5% H2O and 5% CO2 510 ppm NH3, 255 ppm NO, 255 ppm NO2, 8% O2, 5% H2O and 5% CO2 510 ppm NH3, 130 ppm NO, 380 ppm NO2, 8% O2, 5% H2O and 5% CO2 1070 ppm NH3, 495 ppm NO, 15 ppm NO2, 8% O2, 5% H2O and 5% CO2 1570 ppm NH3, 495 ppm NO, 15 ppm NO2, 8% O2, 5% H2O and 5% CO2 300-700 ppm NH3, 495 ppm NO, 15 ppm NO2, 8% O2, 5% H2O and 5% CO2 constant feed of 8% O2, 5% H2O and 5% CO2 switch between 500 ppm NH3 and 490 ppm NO + 14 ppm NO2 and 250 ppm NO + 275 ppm NO2 500 ppm NH3, 490 ppm NO+14 ppm NO2, 8% O2, 5% H2O and 5% CO2 NH3 switched off at steady-state 200-900 ppm NH3, 200-900 ppm NO, 8% O2, 5% H2O and 5% CO2 500 ppm NH3, 0-800 ppm NOx, 8% O2, 5% H2O and 5% CO2

storage at 159 °C temperature ramp: 10 °C/min. storage at 210 °C temperature ramp: 10 °C/min. 150-650 °C 150-650 °C 150-650 °C 150-650 °C 150-650 °C 150-650 °C 150-650 °C 150-650 °C 200 °C 200 and 400 °C

comment used in model development (Figure 1) used in model development (Figure 2) used in model development (Figure 3) used in model development (Figure 4) used in model development (Figure 5) used in model validation (Figure 10) used in model development (Figure 7) used in model development (Figure 8) used in model validation (Figure 11) used in model validation (Figure 12) used in model development (Figure 6) used in model validation (Figure 13)

200 and 400 °C

used in model validation (Figure 14)

200 and 400 °C

used in model validation (Figure 15) used in model validation (Figures 16-18)

200, 300, and 400 °C


temperatures from 150 to 650 °C. The feed contained 495 ppm NO, 15 ppm NO2, 8% O2, 5% H2O, and 5% CO2. The temperatures investigated were 150, 200, 250, 300, 350, 400, 500, and 650 °C, and each temperature was kept for 20 min. The SCR of NOx is influenced by the temperature as well as the feed composition and six experiments were carried out in order to investigate the activity at temperatures from 150 to 650 °C. One experiment was conducted using 495 ppm NO, 15 ppm NO2, 510 ppm NH3, 8% O2, 5% H2O, and 5% CO2 in argon. The NO conversion was studied while increasing the temperature stepwise from 150 to 650 °C, each temperature kept constant for 20 min. The temperatures investigated were: 150, 175, 200, 250, 300, 350, 400, 500, and 650 °C. This experiment was repeated using three additional NOx compositions. The new NOx mixtures contained 380 ppm NO + 130 ppm NO2, 255 ppm NO + 255 ppm NO2, and 130 ppm NO + 380 ppm NO2. The experiment using 380 ppm NO + 130 ppm NO2 was not used in the parameter tuning, but was instead used to validate the model. Two additional experiments were used to validate the model. They were carried out to study the activity when exposing the catalyst to a high NH3 to NO concentration ratio using either 1070 or 1570 ppm NH3 in combination with 495 ppm NO, 15 ppm NO2, 8% O2, 5% H2O, and 5% CO2 in argon. The catalyst was heated to 150 °C and kept at that temperature for 20 min in order to reach steady-state. The temperature was then increased stepwise to: 175, 200, 250, 300, 350, 400, 500, and 650 °C. The NO reduction activity was studied for 20 min at each temperature. One additional experiment was included in the model development in order to account for the inhibiting effect by ammonia. This effect was investigated at 200 °C exposing the catalyst to 495 ppm NO, 15 ppm NO2, 8% O2, 5% H2O, 5% CO2, and changing the concentration of ammonia in steps from 300 ppm via 400, 510, and 600 ppm, up to 700 ppm, and down in steps of 100 ppm to 300 ppm NH3 again. The amount of ammonia stored on the catalyst surface was validated in the presence of CO2, H2O, O2, and NOx. The catalyst was exposed to a mixture of 500 ppm NH3, 8% O2, 5% H2O, and 5% CO2 until steady-state was reached. Ammonia was then switched off at the same time as the 490 ppm NO + 14 ppm NO2 was switched on. The flow was fed to the reactor until all stored ammonia was consumed. The same procedure was then repeated, but instead adding 250 ppm NO + 275 ppm NO2 as NH3 was switched off. These experiments are referred to as “total” ammonia storage experiments. A similar experiment was carried out investigating the “dynamic” ammonia storage. The catalyst was exposed to 490 ppm NO + 14 ppm NO2, 500 ppm NH3, 8% O2, 5% H2O, and 5% CO2 until steady-state was reached. Ammonia was then switched off and the flow was fed to the reactor until all stored ammonia was consumed. The experiments were carried out at both 200 and 400 °C. Short transient NO SCR experiments were performed at both 200 and 400 °C in order to validate the model. A feed consisting of 490 ppm NO, 14 ppm NO2, 500 ppm NH3, 8% O2, 5% H2O, and 5% CO2 was fed to the reactor for 15 min. The concentration of either NO or NH3, or both, was then varied and the length of each sequence was 2 min. Three additional transient experiments were included to investigate to catalyst activity at several levels of total NOx using various NO to NO2 ratios. The catalyst was initially exposed to 500 ppm NH3 and 500 ppm NO in the presence of 8% O2, 5% H2O, and 5% CO2, and the NO to NO2 ratio was then changed every 2 min. The total NOx concentration and the NO to NO2 ratio were varied in the later part of the experiment to study the activity at high and low

41

levels of NOx. The conditions used in each experiment are summarized in Table 1. 3. Model 3.1. Mathematical Model. The material balances were solved using a Fortran code. The heat of reaction associated with NH3 SCR is very low and was therefore neglected in the model. Further, the catalyst was placed inside the heating zone and the measured catalyst temperature was used in the simulations. The film model was used to describe the mass-transfer between the gas and the catalyst surface, but mass-transfer limitations in the washcoat were neglected. The mass-transfer in the washcoat of a Cu-ZSM-5 catalyst was investigated in our previous study, and the result showed that there are no such limitations in that catalyst during SCR of NO.42 The results from a study by Chatterjee et al.37 also indicate that the masstransfer limitations in the washcoat of a zeolite-based catalyst is of minor importance. The main governing equation for trace species i in the gas phase is w ∂xg,i ) -km,iS(xg,i - xs,i) ) Atot ∂z

nr

∑ a s r (T , c , θ) j ij j

s

s

(1)

j)1

and the mass transfer coefficient was calculated using Sherwood number (Sh ) 3) km,i )

Sh c D Dh g,tot i,m

(2)

and the ratio between S and Dh is calculated as S ) 4 × (cell density) Dh

(3)

where the cell density is given in cells/m2. The surface coverage of component k is solved by dθk ) dt

nr

∑ s r (T , c , θ) kj j

s

s

(4)

j)1

and the correlation between the concentration and the molar fraction is cs,i ) cs,totxs,i

(5)

where cs,tot is given by the ideal gas law. P RTs

cs,tot )

(6)

3.2. Kinetic Model. We have previously developed a kinetic model for ammonia SCR over Cu-ZSM-5,42 which accounts for effects such as ammonia storage, NO and NH3 oxidation, NH3 SCR of mixtures of NO + NO2 and N2O formation. In this work, we develop the model on the Cu-ZSM-5 catalyst further in order to describe a commercial Fe-zeolite. In addition, the temperature window is increased to 650 °C and the experimental base for the NO to NO2 ratio and ammonia storage is broadened. Further, the model is also validated using total and dynamic storage data. The model is based on 12 reaction rates which describe various catalytic processes. The rate constants are described by the Arrhenius expression.

( )

kj ) Aj exp

-Ea,j RTs

(7)

Temperature programmed desorption (TPD) experiments showed that the NH3 desorption peaks were broad, which is indicative of repulsive interactions between the adsorbed species on the

42


surface, or different strength of the adsorption sites. This is modeled by coverage dependent activation energies for the desorption steps. Such coverage dependence has earlier been described in several similar models for ammonia desorption.36,39,40,42 Ea,j ) Ea,j(0)(1 - Rjθk)

(8)

The nature of the active sites in Fe-zeolites has been a subject to several studies in the literature since different Fe species may coexist at various positions in the catalyst.46-48 However, due to the uncertainty in the actual nature of these active sites, and in order to keep the model as simple as possible, only one type of site was included for the SCR activity, denoted S1. Kro¨cher et al.15 investigated the desorption of NH3 from an Fe-zeolite catalyst and concluded that the adsorbed ammonia can be divided into two parts, one part which may react with ammonia and one part which is chemically inaccessible. Iwasaki et al.48 observed ammonia desorption from Fe-ZSM-5 catalysts at both high and low temperatures and assigned the low temperature peak to physisorbed ammonia and the high temperature desorption peak to NH3 strongly bound on Bro¨nsted acid sites. Weakly bound ammonia is observed at low temperatures in this study, which is in line also with the experimental and simulation results for the Cu-ZSM-5 catalyst,43 and a second type of site (S2) was therefore introduced in this model. The S2 sites are included to account for the physisorbed ammonia which is inaccessible for NOx reduction. The number of S1 and S2 sites and the parameters for adsorption and desorption of ammonia were then tuned using two ammonia TPD experiments. The reactions and the rate expressions for adsorption and desorption of ammonia on the sites S1 and S2 are given below (r1 and r2). r1

NH3 + S1 798 NH3 - S1 r1 ) k1fcNH3θS1-vacant - k1bθNH3-S1 and r2

NH3 + S2 798 NH3 - S2 r2 ) k2fcNH3θS2-vacant - k2bθNH3-S2 One rate, r3, was added to account for ammonia oxidation. However, this reaction was not sufficient in order to describe the ammonia oxidation at higher temperatures. The ammonia desorption is rapid at 650 °C, causing a very low coverage of ammonia on the surface. However, the oxidation of ammonia is close to 100%, and when using only r3, the oxidation becomes too small due to the very low coverage of ammonia. No significant ammonia oxidation occurs in the empty reactor at this temperature, and it was therefore concluded that the high temperature oxidation takes place in the zeolite channels. The Fe-zeolite includes a number of different sites and it has been reported that the Fe-ZSM-5 catalysts contain three different Fe species: aggregated R-Fe2O3, FexOy, oligomer in the extraframework, and oxo-Fe3+ at ion-exchanged sites.48 The authors further suggest that the ion-exchanged oxo-Fe3+ are the active sites for the NH3 SCR reaction, and that these sites only represent a small part of the total Fe content. Schwidder at al.47 also conclude that the Fe-ZSM-5 catalysts contain different types of Fe species and they further suggest that the standard SCR proceed on both isolated and oligomeric Fe oxo sites, but the fast SCR occur on the isolated Fe sites only. In addition, Schwidder et al.11 conclude that oligomers and aggregate

surfaces are active in the unselective oxidation of NH3 at high temperatures. It is likely that the catalyst used in this study also consists of several types of Fe sites and that the high temperature activity can be attributed to one site that is present in a small amount. However, this model does not distinguish between the different sites on the catalyst surface and in order to develop a model as simple as possible, this high temperature activity is modeled using one additional rate (r4) which depends on the gas phase concentration of ammonia and oxygen. r3 3 2NH3 - S1 + O2 98 N2 + 3H2O 2

r3 ) k3cO2θNH3-S1 and r4 3 2NH3 + O2 98 N2 + 3H2O 2

r4 ) k4cO2cNH3 Nitric oxide is also oxidized on the surface of the Fe-zeolite catalyst. The oxidation is described in the model by a global reversible reaction rate which includes a concentration dependency of NO, O2, and NO2. A similar expression has also been included in global models developed for zeolite-based catalysts, such as a commercial washcoated zeolite SCR catalyst37 and in our work on a Cu-ZSM-5 catalyst.42 Using such a global rate however does not imply that the NO oxidation is a gas phase reaction, but rather that a global rate using gas phase concentration dependency is sufficient to describe the experimental observations made for this catalyst. r5 1 NO + O2 798 NO2 2

r5 ) k5,fcO21/2cNO - k5,bcNO2 The rate constant for the dissociation of NO2, k5,b, is calculated from the thermodynamic restriction for the oxidation of NO (NO + 1/2O2 T NO2) using ∆H ) -58.279 kJ/mol and ∆S ) -76.1 J/(mol K). Several reactions for the selective catalytic reduction of NO and NO2 were included in the model. The standard SCR reaction is included (r6), but with the use of a modified stoichiometry based on the observed overconsumption of NH3 over this Fe-zeolite catalyst. Other authors have also reported an overconsumption of ammonia compared to NO over Fe-zeolite catalysts. It has been suggested that the overconsumption of ammonia is due to side reactions, such as ammonia oxidation.15,18 The ammonia oxidation experiment in the absence of NO performed over the Fe-zeolite catalyst used in this study shows that ammonia oxidation starts at higher temperatures than the SCR of NO, see Figure 3 and Figure 5. However, the rate of ammonia oxidation in the presence of NO may not be the same as the rate of oxidation by oxygen only. The reason for the overconsumption during the SCR of NO is not well understood and to develop a model as simple as possible, a rate based on an increased NH3 to NO stoichiometry was chosen. The rate expression for NH3 SCR of NO contains a dependence on the number of vacant sites in order to explain the inhibiting effect of high ammonia concentrations. Tronconi and co-workers38,39 also observed an ammonia blocking effect on a vanadia-based catalyst, and developed an inhibition expression based on a redox mechanism for that catalyst. This redox mechanism has also been applied to models on zeolite-

Ind. Eng. Chem. Res., Vol. 49, No. 1, 2010 36,37

17

based catalysts. Devadas et al. propose that the inhibition effect of ammonia on the SCR activity is attributed to the reduction of Fe3+ to Fe2+ by ammonia. Another possible explanation to the inhibiting effect by ammonia at low temperatures is competitive adsorption.15,25,49 Grossale et al.33 suggest in a recent paper that ammonia likely blocks the surface nitrates and inhibits the oxidation of NO over an Fe-ZSM-5 catalyst. The inhibition expression developed on the vanadia-based catalyst by Tronconi and co-workers38,39 was simplified in this model, since it was observed that multiplying the standard rate expression with the number of vacant sites could satisfy the experimental results accurately over this Fe zeolite. The vacant site dependency is included in the rate expression in order to include the inhibiting effect caused by ammonia at high surface coverage. Without the vacant site dependency, the calculated conversion of NO would increase with increased ammonia coverage, and not decrease as experimentally observed at low temperatures. High ammonia coverage is thus expected to block sites which are needed for formation of other species involved in the SCR of NO on the catalyst surface. The SCR of NO is active also at the highest temperatures in this study, but due to the rapid ammonia desorption and the subsequent low ammonia coverage, the conversion of NO is difficult to describe from 500 °C. As described above for the ammonia oxidation, a corresponding reaction rate based on the gas phase concentration of ammonia was included in the model, which is active mainly at 650 °C. The reaction is incorporated as a 1:1 stoichiometry between NH3 and NO, since it is difficult to determine the NH3 consumption in the SCR reaction due to rapid ammonia oxidation at this high temperature.

The reaction rates for the reduction of NO only, NO + NO2, NO2 only, and the rate for N2O formation are all based on global reactions. These rates can be divided into several steps to include formation of intermediate species. Formation and build up of NH4NO3 is expected if a high concentration of NO2 is present in the feed at low temperatures, and the global rate is often described as a reaction between NO2 and NH3 according to the stoichiometry below (r11).1,19,32,33 Ammonium nitrate will deposit as a solid salt on the catalyst at temperatures below 170 °C,1 or in cold parts along the lines in the experimental setup.3 Since ammonium nitrate can easily be deposited on any surface, the formation is assumed to occur on S2 in this model. The formation of NH4NO3 may also take place on the S1 sites, but due to the small number these sites, ammonium nitrate deposition on S1 sites does not contribute significantly to the total amount stored, and was therefore not included in the model. Ammonium nitrate is a solid salt and its deposition may block active sites in the catalyst.3 The reaction on S2 contains an inhibition expression which accounts for the build up of high levels of ammonium nitrates. These nitrates may block the pores in the zeolite structure and hinder further ammonium nitrate formation. This is observed in the experiment as a slow increase in outlet concentrations of NH3 and NO2 at 150 °C (Figure 8). At this temperature the storage is high during the first 5 min, but as the catalyst becomes covered with ammonium nitrate, less of it is formed and more NH3 and NO2 leave the catalyst unconverted. r11

2NH3 - S2 + 2NO2 98 NH4NO3 - S2 + N2 + H2O

r6

6NH3 - S1 + 5NO + 2O2 98 5.5N2 + 9H2O r6 ) k6cNOθNH3-S1(1 - θNH3-S1) and r7

4NH3 + 4NO + O2 98 4N2 + 6H2O r7 ) k7cNOcNH3 The fast SCR consuming equal amounts of NO and NO2 and the SCR of NO2 are also included in the model and global rates are used for these reactions (r8 and r9).42 r8

2NH3 - S1 + NO + NO2 98 2N2 + 3H2O r8 ) k8cNOcNO2θNH3-S1 and

43

r11 ) k11

cNO2θNH3-S2 1 + K11θNH4NO3-S2

Ammonium nitrate may decompose to N2O and H2O or dissociate to HNO3 and NH3.1 A third possibility is the reaction between ammonium nitrate and NO.3,31 The experimental result in this study shows however that the main decomposition products at each temperature increase up to 200 °C, where large amounts of ammonium nitrate are available, are NO2 and NH3 (Figure 8). Still, a small amount of N2O may be formed, but the main part of the N2O formation is accurately described by the model using the rate reported above (r10). The ammonium nitrate is therefore believed to mainly decompose into HNO3 and NH3. Further, nitric acid may decompose into NO2, O2, and H2O.31,50 One global rate, r12, which is a combination of ammonium nitrate decomposition to NH3 and HNO3 and the subsequent decomposition of HNO3 into NO2, O2, and H2O, was used to describe the large amounts of NO2 and NH3 observed: r12

r9

4NH3 - S1 + 3NO2 98 3.5N2 + 6H2O r9 ) k9cNO2θNH3-S1 The experimental results show that a significant amount of N2O is formed only if the feed contains NO2, thus, the N2O formation is based on a reaction including NO2, but no NO.42 r10

2NH3 - S1 + 2NO2 98 N2 + N2O + 3H2O r10 ) k10cNO2θNH3-S1

NH4NO3 - S2 98 NH3 - S2 + NO2 + 0.25O2 + 0.5H2O r12 ) k12θNH4NO3-S2 4. Results and Discussion 4.1. Ammonia Storage. Two ammonia TPD experiments were used in the parameter tuning. In the first TPD experiment, the catalyst was exposed to 510 ppm NH3, 5% H2O, and 5% CO2 for 40 min at 159 °C, followed by a period of 5% H2O in argon and a temperature ramp (10 °C/min). The result from both the experiment and the simulation are shown in Figure 1. There is a total uptake of ammonia for 1.5 min and then the

44


Figure 1. Ammonia storage for 40 min at 159 °C using 510 ppm ammonia in the presence of 5% H2O and 5% CO2, followed by a period of 5% H2O in argon and a heating ramp (10 °C/min). The solid line shows the calculated concentration, and the dotted line shows the measured concentration.

Figure 2. Ammonia storage for 40 min at 210 °C using 510 ppm ammonia in the presence of 5% H2O and 5% CO2, followed by a period of 5% H2O in argon and a heating ramp (10 °C/min). The solid line shows the calculated concentration, and the dotted line shows the measured concentration.

concentration of outlet ammonia increases up to feed concentration. After 40 min the ammonia is shut-off and the concentration of ammonia slowly decreases down to zero. The temperature is then increased 10 °C/min, and ammonia starts to desorb. The measured concentration of ammonia shows that at temperatures above 400 °C all ammonia is desorbed from the surface. The data points up to zero coverage were included in the model fit, which occurs at about 430 °C. Empty reactor data was used as inlet concentration, and the model overestimates the ammonia desorption after shut-off and releases a too small amount of ammonia during the heating ramp, but the total amount stored at this temperature is well described by the model. Ammonia desorbs from the sites for weakly bound species (S2) after ammonia shut-off, and the fraction released during the heating ramp is mainly desorbed from the active sites (S1). The storage of ammonia at 210 °C was also included in the tuning of the parameters, and the result from the second TPD experiment is presented in Figure 2. The model is able to describe the adsorption and desorption of ammonia well throughout the whole experiment. 4.2. NH3 Oxidation. Ammonia is consumed by oxygen over the catalyst at high temperatures. An experiment which was made at temperatures from 150 to 650 °C in the presence of 510 ppm NH3, 8% O2, 5% H2O, and 5% CO2 was modeled. The result is shown in Figure 3. Ammonia is stored on the surface for about 1.5 min until the surface is saturated at 150 °C, and no ammonia oxidation is observed at this temperature. Small ammonia desorption peaks are observed each time the temperature is increased, but ammonia oxidation is not active until the temperature reaches 350 °C. At 650 °C almost all

Figure 3. Ammonia oxidation experiment at temperatures from 150 to 650 °C. Feed gas composition of 510 ppm NH3, 8% O2, 5% CO2, and 5% H2O. The solid line shows the calculated concentration and the dotted line shows the measured concentration.

Figure 4. NO oxidation experiment at temperatures from 150 to 650 °C. Feed gas composition of 495 ppm NO, 15 ppm NO2, 8% O2, 5% CO2, and 5% H2O. The solid lines show the calculated concentration and the dotted lines show the measured concentration.

ammonia is consumed. A negligible amount of NOx is observed at all temperatures investigated. It has been suggested that ammonia is oxidized to NO which is reduced by ammonia to N2.51 Such a two-step reaction is possible, but since very little NO is detected, a global rate for NH3 oxidation directly to N2 is sufficient to describe the experimental results in this temperature region. A high selectivity toward N2 at temperatures up to 600 °C has also been observed by Kro¨cher et al.15 The calculated result from our model is shown in Figure 3. Desorption of ammonia is overestimated by the model at the first temperature increase, but it is well described at each additional temperature increase. The total ammonia storage and the rate of oxidation are also well correlated to the experimental result. At the higher temperatures, especially at 650 °C, the introduction of the second reaction for ammonia oxidation was critical. Without the second reaction, the model would predict a decreased ammonia oxidation at 650 °C since the coverage of ammonia is very low at this high temperature. 4.3. NO Oxidation. The Fe-zeolite catalyst is able to oxidize NO to NO2. The catalyst was exposed to 495 ppm NO, 15 ppm NO2, 8% O2, 5% CO2, and 5% H2O, and the temperature was stepwise increased from 150 to 650 °C. The parameters for the reversible NO oxidation reaction were tuned and the result is presented, together with the measured concentrations, in Figure 4. The oxidation of NO increases with temperature and reaches equilibrium above 400 °C. The measured concentration of NOx rapidly reaches the inlet concentration at the beginning of the experiment and there is no desorption of NOx observed as the temperature is raised, which indicates that very little, or no NOx,


Figure 5. Experimental and calculated concentrations from 150 to 650 °C, using a SV of 25 000 h-1 and inlet concentrations of 495 ppm NO, 15 ppm NO2, 510 ppm NH3, 5% H2O, 5% CO2, and 8% O2. The solid lines show the calculated concentration and the dotted lines show the measured concentration.

is stored on the surface. The model contains one reversible rate for the NO oxidation, which is sufficient to describe the experimental result well at all temperatures investigated. 4.4. NH3 SCR of NOx. The NH3 SCR experiment reducing NO is shown in Figure 5. The feed consisted of 495 ppm NO, 15 ppm NO2, 510 ppm NH3, 8% O2, 5% H2O, and 5% CO2, and the temperature was increased stepwise from 150 to 650 °C. The conversion of NO is about 10% at 150 °C and increases each time the temperature is raised. The NOx reduction reaches 92% at 500 °C but decreases again at higher temperature due to the rapid NH3 oxidation that occurs at 650 °C. The selective catalytic reduction of NO with ammonia consumes more NH3 than NO over this Fe-zeolite catalyst. Such an overconsumption was not observed in our earlier studies on Cu-ZSM-5.9,42 Other studies on Fe-zeolite catalysts have on the other hand also shown an overconsumption of ammonia in the reduction of NO due to side reactions such as ammonia oxidation.15,18 However, in this study, as shown in Figure 3, the oxidation of NH3 in the absence of NO does not start until the temperature reaches 350 °C, but the overconsumption of ammonia is obvious already at 250 °C (Figure 5). It is thus not clear why the reduction of NO consumes more than stoichiometric concentration of ammonia over this catalyst. However, the model can account for the overconsumption because of a modified reaction stoichiometry as discussed earlier, and is therefore able to describe both consumption of NH3 and reduction of NO well at all temperatures. The calculated and experimental results differ about 30 ppm at the most, and both modeled and measured formation of N2O is less than 4 ppm. The second SCR experiment used in the model development was performed at 200 °C exposing the catalyst to 495 ppm NO, 15 ppm NO2, 8% O2, 5% H2O, 5% CO2, and the concentration of ammonia changed in steps from 300 to 700 ppm and down to 300 ppm again. The results from the experiments and the model are shown in Figure 6. The catalyst is exposed to 300 ppm NH3 initially, but after 30 min the inlet NH3 concentration is changed to 400 ppm, resulting in a reduced NOx reduction activity. Thus, a clear ammonia blocking effect is observed already for a ratio of NH3 to NO below 1. Ammonia inhibition effects have been observed earlier on iron exchanged zeolites,15,17,33,36 Cu-ZSM-5,49 H-ZSM-5,25 and on vanadia-based catalysts.38 Figure 6 further shows that the activity is reduced each time the ammonia concentration is raised, but the higher NOx reduction activity is recovered as soon as the feed NH3

45

Figure 6. Experimental and calculated concentrations at 200 °C, using a SV of 25 000 h-1 and inlet concentrations of 495 ppm NO, 15 ppm NO2, 300-700 ppm NH3, 5% H2O, 5% CO2, and 8% O2. The solid lines show the calculated concentration and the dotted lines show the measured concentration.

Figure 7. Experimental and calculated concentrations from 150 to 650 °C, using a SV of 25 000 h-1 and inlet concentrations of 255 ppm NO, 255 ppm NO2, 510 ppm NH3, 5% H2O, 5% CO2, and 8% O2. The solid lines show the calculated and the dotted lines show the measured concentrations.

level is decreased. The experimental and calculated NOx conversions are well correlated. The vacant site dependence in the reaction rate for the standard SCR (r6) is responsible for the prediction of the decreased NO reduction with increased NH3 feed concentration. Without the vacant site dependence, the model would predict an enhanced activity with higher feed concentrations of NH3. The modeling result shows that NH3 is blocking sites needed for the SCR of NO. The model was developed to account also for the effect of varying the NO to NO2 ratio. It is well-known that the conversion of NOx is strongly influenced by the NO2 content in the feed. The optimum selective catalytic reduction of NOx typically occurs around equimolecular amounts of NO and NO2 over various catalyst formulations such as vanadia-based catalysts,3,8,52 Cu-ZSM-5 catalyst,9 and Fe-zeolite catalysts.8,16,19 Figure 7 shows the result from an experiment where the catalyst was exposed to 510 ppm NH3, 255 ppm NO, 255 ppm NO2, 8% O2, 5% H2O, and 5% CO2. The temperature was increased stepwise from 150 to 650 °C. The SCR activity is high already at 150 °C and stays at about 95% conversion up to 500 °C where the NOx conversion decreases due to ammonia oxidation. Moreover, at 650 °C almost all unconverted NOx exit the catalyst as NO. The model is able to describe the enhanced activity with equal amounts of NO and NO2 at all temperatures investigated, and the rate for fast SCR (r7) is responsible for this high NOx reduction activity at lower temperatures.

46


Figure 8. Experimental and calculated concentrations from 150 to 650 °C, using a SV of 25 000 h-1 and inlet concentrations of 130 ppm NO, 380 ppm NO2, 510 ppm NH3, 5% H2O, 5% CO2, and 8% O2. The solid lines show the calculated concentration, and the dotted lines show the measured concentration.

A similar experiment was also modeled using a higher NO2 to NO ratio, which is shown in Figure 8, where the feed consisted of 510 ppm NH3, 130 ppm NO, 380 ppm NO2, 8% O2, 5% H2O, and 5% CO2. The conversion of NOx is lower than if equal amounts of NO and NO2 is used, which is in agreement with the results reported by other authors for both zeolite and vanadia based catalysts.3,8,9,16,19 The SCR of equimolecular amounts of NO and NO2 is active to a smaller extent since less NO is available, but higher formation and storage of ammonium nitrate is expected at the lower temperatures. The reduction of NO2 consumes more ammonia than the simultaneous reduction of NO and NO2 does, which explains why the conversion of NO2 is lower also at the higher temperatures. Further, at 650 °C the conversion of NOx drops rapidly and the main fraction of unconverted NOx consists of NO, even though the main feed fraction is NO2, which shows that the NO2 dissociation is rapid over the catalyst at this temperature. The catalyst stores a large amount of ammonia and NOx at 150 °C, which probably is due to the formation of ammonium nitrate. As the temperature is increased to 175 and 200 °C, desorption of both ammonia and NO2 is observed. At these temperatures, ammonium nitrate is expected to decompose into HNO3 + NH3 or N2O + H2O. The result shows that the main decomposition products are NH3 + NO2, where NO2 likely is formed from the subsequent decomposition of HNO3. The model is able to describe the ammonium nitrate formation and decomposition as well as the NOx conversion at all temperatures using summation reaction rates only. The summary rates for the SCR reactions and the reaction for formation of N2O describe the catalytic activity, and the rate for ammonium nitrate formation accounts for the possibility for NH3NO3 to deposit on any part of the catalytic surface (S2 sites) at low temperatures if high fractions of NO2 are available. The result in Figure 8 shows that the main decomposition products are NH3 and NO2, but a minor fraction is also forming N2O. The formation of N2O is included in the model as a global reaction between NO2 and NH3, and the result shows that such an expression can describe the N2O formation well. A comparison between the experimental and calculated conversions using various NO to NO2 feed ratios is shown in Figure 9. It is clear that the NOx reduction activity increases if NO2 is introduced to the feed and that a NOx composition of about equal amounts of NO and NO2 is beneficial at temperatures up to 500 °C. Both NO and NO2 are thus of importance

Figure 9. Experimental and calculated steady-state conversions of NH3 and NOx. The catalyst was exposed to 510 ppm NH3, 510 ppm total NOx, 8% O2, 5% H2O, 5% CO2. The composition of total NOx was varied using 0, 50%, or 75% NO2. Markers show the experimental conversions, and the lines show the calculated conversions.

in order to reach the highest conversion possible. It has been reported that the oxidation of NO is the rate determining step on H-ZSM-5 catalysts in absence of NO2.25,26 Devadas et al.16 also observed that zeolite catalysts require NO2 in the feed, or an activity for oxidation of NO to attain high SCR activity. On the other hand, NO is also required and Yeom et al.31 proposed a mechanism which involves NO in several reaction steps. Grossale et al.33 also suggest a reaction scheme for the fast SCR chemistry over an Fe-ZSM-5 catalyst where the role of NO is to reduce surface nitrates to nitrites. At 650 °C, the highest conversion is observed if NO is used. Ammonia oxidation is rapid at this temperature, which results in a decreased conversion of NOx in all three cases, but the NOx conversion drops more if NO2 is present. The conversion of NH3 and NOx may occur in the catalyst front, leaving a large fraction of the surface for the dissociation of unconverted NO2, which can explain why only NO is observed even though NO2 is fed to the reactor. The model is able to describe the steady-state conversion of both NOx and NH3 well throughout all experiments. 4.5. Model Validation. Several additional experiments were used to validate the model. The first validation experiment investigated if the model can predict one additional NO to NO2 ratio. The temperature was stepwise increased from 150 to 650 °C, and the feed consisted of 510 ppm NH3, 380 ppm NO, 130 ppm NO2, 8% O2, 5% H2O, and 5% CO2. The result is shown in Figure 10. The NOx reduction is higher than if only NO is used at temperatures up to 400 °C, which was expected since the NO + NO2 SCR is more rapid than the SCR of NO only. The model describes the experimental result adequately. Two additional experiments using high fractions of ammonia were predicted by model. The ammonia concentration was two or three times the NO concentration, respectively. Figure 11 shows the result from an experiment and simulation using 495 ppm NO, 15 ppm NO2, 1070 ppm NH3, 5% H2O, 5% CO2, and 8% O2, while Figure 12 shows the result from a simulation and experiment exposing the catalyst to 495 ppm NO, 15 ppm NO2, 1570 ppm NH3, 5% H2O, 5% CO2, and 8% O2. The temperature was increased stepwise from 150 to 650 °C in both runs. High fraction of NH3 does not result in an improved NOx conversion at low temperatures, but generates an enhanced NOx reduction


Figure 10. Experimental and calculated concentrations from 150 to 650 °C, using a SV of 25 000 h-1 and inlet concentrations of 380 ppm NO, 130 ppm NO2, 510 ppm NH3, 5% H2O, 5% CO2 and 8% O2. The solid lines show the calculated concentration, and the dotted lines show the measured concentration.



at high temperatures where ammonia oxidation is important since more of it becomes available for the reaction with NOx (compare with Figure 5). However, a higher feed concentration of NH3 also results in an increased ammonia slip. Also, a total

47

Figure 13. Total ammonia storage at 200 and 400 °C. The catalyst was initially exposed to 500 ppm NH3, 8% O2, 5% H2O, and 5% CO2. Ammonia was switched off after 30 min and NO or NO + NO2 was switched on. (a) NH3 off and 490 ppm NO + 14 ppm NO2 on at 200 °C, (b) NH3 off and 250 ppm NO + 275 ppm NO2 on at 200 °C, (c) NH3 off and 490 ppm NO + 14 ppm NO2 on at 400 °C, and (d) NH3 off and 250 ppm NO + 275 ppm NO2 on at 400 °C. The solid lines show the calculated concentration, and the dotted lines show the measured concentration.

Figure 14. Dynamic ammonia storage at 200 and 400 °C. The catalyst was initially exposed to 500 ppm NH3, 490 ppm NO+14 ppm NO2, 8% O2, 5% H2O, and 5% CO2. Ammonia was switched off after 30 min, (a) experiment carried out at 200 °C and, (b) experiment performed at 400 °C. The solid lines show the calculated concentration, and the dotted lines show the measured concentration.

conversion of NOx is achieved already at 1070 ppm NH3 (Figure 11), and very little effect is gained from adding additional 500 ppm NH3 (Figure 12). The model is able to accurately describe the enhanced NOx reduction activity as well as the increased ammonia slip. The total and the dynamic ammonia storage were validated using experiments where ammonia is shut off and NOx is used to titrate off the adsorbed ammonia. The total ammonia storage experiment investigates the amount of ammonia adsorbed in the presence of 500 ppm NH3, 5% H2O, 5% CO2, and 8% O2, while the dynamic ammonia storage experiments investigates the amount of stored ammonia in the presence of 500 ppm NOx, 500 ppm NH3, 5% H2O, 5% CO2, and 8% O2. The experiments were carried out at both 200 and 400 °C. The results from the model validation on the total ammonia storage experiments are shown in Figure 13, and the results from the model validation on the dynamic ammonia storage are shown in Figure 14. No oxidation of ammonia occurs at 200 °C (see Figure 13a,b), and the surface coverage of ammonia is expected to be high. As ammonia is switched off, NO (Figure 13a) is added, and a reaction between NOx and adsorbed NH3 takes place. The

48


Figure 15. Transient experiments carried out at both 200 and 400 °C. The catalyst was exposed to a feed of 5% H2O, 5% CO2, 8% O2, and various mixtures of NH3 and NO. (a) The inlet concentration of NH3 and NOx, (b) experimental and calculated concentrations at 200 °C, and (c) experimental and calculated concentrations at 400 °C.

activity increases after about 30 s because of the reduced NH3 inhibition effect as the coverage of ammonia decreases initially. The coverage will then reach a low level where it causes a decreased NOx reduction. Eventually all ammonia is consumed and no reduction can take place. The model is able to predict an enhanced NO reduction as ammonia is consumed as well as the decreased NOx conversion as the coverage of ammonia becomes too low. If NO + NO2 is added as ammonia is switched off (Figure 13b), a more rapid reaction between NOx and ammonia takes place which results in a decreased period for the ammonia consumption. The model predicts the total titration time adequately and is able to accurately describe the amount of N2O formed after the switch is carried out. At 400 °C, very little ammonia is adsorbed on the surface and some ammonia is oxidized. As a result, NOx is observed almost instantly after the switch (see Figure 13c,d). The calculated result is well correlated to the experimental observations at this temperature, both concerning the rate of ammonia oxidation and the amount of ammonia stored on the surface. Similar results are observed if ammonia is removed from a mixture of NOx and NH3 in the presence of oxygen. The NOx conversion at 200 °C increases initially as ammonia is shut off (Figure 14a). This effect is caused by the reduced inhibition effect as surface coverage of ammonia is decreased. However, eventually all ammonia is consumed and no reduction can take place. The transient effect caused by ammonia inhibition at low temperatures has been reported to occur over various types of catalysts during ammonia SCR.19,26,38 At 400 °C (Figure 14b), a high conversion of NOx is expected and the coverage of ammonia is low causing a rapid change in outlet NOx as ammonia is switched off. The model is able to predict these changes adequately at both 200 and 400 °C. Transient experiments where the NO and/or NH3 concentration was changed in sequences of 2 min were validated, and the result is shown in Figure 15. The feed consisted of a constant concentration of 5% H2O, 5% CO2, and 8% O2,

but the concentration of NO and NH3 was varied. The catalyst was initially exposed to 500 ppm NH3 and 500 ppm NO for 15 min. The concentration of NO was then increased and decreased in sequences of 2 min, followed by a similar change in NH3 concentration and the corresponding change in both NO and NH3. The same procedure was then repeated but the concentrations were varied to higher and lower levels. The experiment was carried out at both 200 and 400 °C. The inlet concentrations of NH3 and NOx are shown in Figure 15a. The catalyst stores ammonia at 200 °C for almost 3 min but NO is detected as soon as the experiment is started (see Figure 15b). About 100 ppm NO is reduced by ammonia during the first 15 min, which is successfully described by the model. Inlet NO is then changed up and down by 200 ppm which is well captured by the model. The corresponding change was then carried out for ammonia, followed by a combination of both ammonia and NO. An adequate correlation between the calculated and the experimental results is predicted by the model during these sequences. The change in concentration is then increased, but the model has some difficulties describing the sequences where the NO to NH3 ratio is the highest. Such conditions were not considered during the model development and are likely due to an overestimated ammonia inhibition. However, the main catalytic performance and responses during the transient cycle are well described by the model. The experiment was repeated at 400 °C and used for model validation and the results are shown in Figure 15c. The transient performance of the catalyst activity during NH3 SCR of NO is adequately described by the model at this temperature where the reduction activity is rapid. The last three experiments used in the model validation were included to study the catalyst activity as the NO to NO2 ratio and the total NOx fractions were varied while keeping the NH3 concentration at a constant level. Figure 16 shows the experimental and calculated results at 200 °C. The catalyst was initially exposed to 500 ppm NH3 and 500 ppm NO in


49

Figure 16. Transient experiment carried out at 200 °C. The catalyst was exposed to a feed of 500 ppm NH3, 5% H2O, 5% CO2, 8% O2, and various mixtures of NO and NO2.

Figure 17. Transient experiment carried out at 300 °C. The catalyst was exposed to a feed of 500 ppm NH3, 5% H2O, 5% CO2, 8% O2 and various mixtures of NO and NO2.

the presence of O2, H2O, and CO2, and the NO to NO2 ratio was then changed every 2 min. The total NOx concentration and the NO to NO2 ratio were varied in the later part of the experiment to study the activity at high and low levels of NOx. The conversion of NOx increases from about 20% to close to 100% as the NO2 fraction is changed from 0 to 50%, which is well described by the model. After 18 min, the NO2 fraction is increased to 100% and experimental result shows that the conversion of NO2 is high and the formation of N2O increases up to about 100 ppm. This change is partly described by the model, but the conversion of NO2 and the formation of N2O are underestimated. The NO2 is then switched off while NO is switched on, and the measured concentration of NOx is zero for about 20 s. This temporary high conversion is likely due to a reaction with NOx species formed on the surface during the former sequence where excess NO2 was used. Such a delay cannot be described by this global model since reaction steps for storage of NO and NO2 were excluded in this study. The model would require several additional steps to describe the formation of surface NOx species, and the objective of this work was to use a global model with few reaction steps. After 24 min, the

catalyst was exposed to a NOx composition of 25% NO and 75% NO2. The experimental result shows that about 60 ppm N2O is formed, but as the catalyst was exposed to the same condition in an earlier experiment (see Figure 8), only 35 ppm N2O was detected. These last three experiments (Figure 16-18) were performed a long time after the other experiments, and the catalyst has been used in other studies during this period. The activity for N2O formation is slightly enhanced, but the activity for the SCR of NO is preserved. The discrepancy between experimental and calculated NO2 and N2O concentrations at excess NO2 is therefore partly due to changed catalytic activity. In the later part of the experiment the NOx levels were varied between 0 and 800 ppm, and the model is able to predict these new conditions adequately. The same conditions were also validated at 300 and 400 °C and the results are shown in Figures 17 and 18. At these temperatures, the activity is less affected by the ammonia storage and the formation of surface NOx species, and the conversion of NOx changes more rapidly with the changes in feed

50


Figure 18. Transient experiment carried out at 400 °C. The catalyst was exposed to a feed of 500 ppm NH3, 5% H2O, 5% CO2, 8% O2, and various mixtures of NO and NO2.

concentration. The measured and calculated concentrations correlate well throughout the experiment at both 300 and 400 °C. 5. Conclusions The selective catalytic reduction of NOx with ammonia over a commercial Fe-zeolite catalyst was investigated using flow reactor experiments and kinetic modeling. The model includes reactions that descibe ammonia storage and oxidation, NO oxidation, SCR of NO, SCR of NO + NO2, SCR of NO2, formation of N2O, ammonia inhibition, and ammonium nitrate formation. All experiments were carried out in presence of O2, H2O, and CO2, and the model was developed to account for a broad range of experimental conditions at temperatures from 150 to 650 °C. Ammonia TPD experiments were used to investigate the ammonia storage, and the results show that the catalyst stores ammonia at temperatures up to 400 °C. The catalyst is active for the oxidation of ammonia from 350 °C, and at 650 °C almost all ammonia is consumed. The NO oxidation is also included in the model, and it is observed that NO oxidation increases with the temperature and reaches equilibrium above 400 °C. The SCR of NO occurs already at 150 °C, and the introduction of equal amounts of NO and NO2 greatly enhances the conversion of NOx at temperatures up to 300 °C. At 650 °C the highest NOx conversion is instead achieved if NO is fed to the reactor. The formation of N2O is negligible if small fractions of NO2 are used, but a significant amount of N2O is formed at high NO2 to NO ratios. An ammonia inhibition on the SCR of NO is observed at 200 °C, where the activity is reduced by ammonia already below stoichiometric NH3 to NO fractions. At high temperatures, increased ammonia to nitric oxide stoichiometry is beneficial for the NOx reduction since the coverage of ammonia is low and the oxidation of ammonia competes with the reduction of NO. More ammonia becomes available for the reduction but a large ammonia concentration also results in an increased ammonia slip. The model contains 12 reactions and 8 experiments were used for the model development. The calculated results show that the model is able to describe the experimental observations well. There are many similarities between the kinetic models on the Cu-ZSM-546 and the Fe-zeolite catalyst. Mostly, it is the rate of the reactions that changes. However, for iron we observe

an overconsumption of ammonia for the standard SCR reaction, which was not observed for the Cu-ZSM-5. Other differences are that the Fe-zeolite shows (i) lower NH3 storage, (ii) lower NO reduction activity at low temperatures, (iii) lower NO oxidation, and (iv) lower NH3 oxidation which results in increased NO conversion at higher temperatures. In addition, since temperatures up to 650 °C were used in the model development for the Fe-zeolite catalyst, we also needed to add extra reaction steps for high temperature performance. The model was validated by adequately predicting shorter transient experiments, ammonia storage experiments, and experimental conditions not used in the parameter estimation. Acknowledgment The work was performed at Competence Centre for Catalysis (KCK), Chalmers and at General Motors Research and Development Center. The authors would like to acknowledge helpful discussions with Se Oh, Ed Bissett, Jong-Hwan Lee, and Byong Cho of the General Motors Research and Development Center. We would also like to acknowledge Johnson Matthey for providing us the catalyst on which to do the kinetic measurements. L. Olsson would also like to acknowledge the Swedish Research Council (Contract No. 621-2003-4149 and No. 6212006-3706) for additional support. The financial support for the reactor equipment from Knut and Alice Wallenberg Foundation, Dnr KAW 2005.0055, is gratefully acknowledged. Nomenclature aj ) active site density for reaction j, molsites/m3 Aj ) pre-exponential factor, unit depends on rate expression Atot ) front area of the monolith, m2 cg,tot ) total concentration in the gas bulk, mol/m3 cs,i ) molar concentration of gas species i at the catalyst surface, mol/m3 cs,tot ) total concentration at the catalyst surface, mol/m3 Dh ) hydraulic diameter of channel, m Di,m ) binary diffusion coefficient of species i in the mixture, m2/s Ea,j(0) ) activation energy for zero coverage, J/mol Ea,j ) activation energy, J/mol kj ) rate constant, unit depends on rate expression km,i ) mass transfer coefficient for species i, mol/(m2 s)

Ind. Eng. Chem. Res., Vol. 49, No. 1, 2010 nr ) number of reactions P ) total pressure, Pa rj ) reaction rate for reaction j, mol/(molsites · s) R ) gas constant, J/(mol · K) S ) surface area per reactor volume, m-1 Sh ) Sherwood number sij ) stoichiometric coefficient of species i in reaction j skj ) stoichiometric coefficient of surface species k in reaction j t ) time, s Ts ) temperature at catalyst surface, K w ) molar flow rate, mol/s xg,i ) mole fraction of gas species i in the bulk xs,i ) mole fraction of gas species i at the surface z ) axial position, m Rj ) constant θk ) coverage of component k

Literature Cited (1) Koebel, M.; Madia, G.; Elsener, M. Selective catalytic reduction of NO and NO2 at low temperatures. Catal. Today 2002, 73, 239. (2) Koebel, M.; Elsener, M.; Madia, G. Reaction pathways in the selective catalytic reduction process with NO and NO2 at low temperatures. Ind. Eng. Chem. Res. 2001, 40, 52. (3) Ciardelli, C.; Nova, I.; Tronconi, E.; Chatterjee, D.; Bandl-Konrad, B.; Weibel, M.; Krutzsch, B. Reactivity of NO/NO2-NH3 SCR system for diesel exhaust aftertreatment: Identification of the reaction network as a function of temperature and NO2 feed content. Appl. Catal., B 2007, 70, 80. (4) Busca, G.; Lietti, L.; Ramis, G.; Berti, F. Chemical and mechanistic aspects of the selective catalytic reduction of NOx by ammonia over oxide catalysts: A review. Appl. Catal., B 1998, 18, 1. (5) Komatsu, T.; Nunokawa, M.; Moon, I. S.; Takahara, T.; Namba, S.; Yashima, T. Kinetic studies of reduction of nitric oxide with ammonia on 2+ Cu -exchanged zeolites. J. Catal. 1994, 148, 427. (6) Baik, J. H.; Yim, S. D.; Nam, I. S.; Mok, Y. S.; Lee, J. H.; Cho, B. K.; Oh, S. H. Control of NOx emissions from diesel engine by selective catalytic reduction (SCR) with urea. Top. Catal. 2004, 30-31, 37. (7) Rahkamaa-Tolonen, K.; Maunula, T.; Lomma, M.; Huuhtanen, M.; Keiski, R. L. The effect of NO2 on the activity of fresh and aged zeolite catalysts in the NH3-SCR reaction. Catal. Today 2005, 100, 217. (8) Schmieg, S. J.; Lee, J.-H. Evaluation of supplier catalyst formulations for the selective catalytic reduction of NOx with ammonia. SAE Tech. Pap. 2005, 2005-01-3881. (9) Sjo¨vall, H.; Olsson, L.; Fridell, E.; Blint, R. J. Selective catalytic reduction of NOx with NH3 over Cu-ZSM-5sThe effect of changing the gas composition. Appl. Catal., B 2006, 64, 180. (10) Park, J.-H.; Park, H. J.; Baik, J. H.; Nam, I.-S.; Shin, C.-H.; Lee, J.-H.; Cho, B. K.; Oh, S. H. Hydrothermal stability of Cu-ZSM-5 catalyst in reducing NO by NH3 for the urea selective catalytic reduction process. J. Catal. 2006, 240, 47. (11) Schwidder, M.; Kumar, M. S.; Klementiev, K.; Pohl, M. M.; Bruckner, A.; Grunert, W. Selective reduction of NO with Fe-ZSM-5 catalysts of low Fe content: I. Relations between active site structure and catalytic performance. J. Catal. 2005, 231, 314. (12) Qi, G.; Wang, Y.; Yang, R. T. Selective catalytic reduction of nitric oxide with ammonia over ZSM-5 based catalysts for diesel engine applications. Catal. Lett. 2008, 121, 111. (13) Kro¨cher, O.; Elsener, M. Combination of V2O5/WO3-TiO2, Fe-ZSM5, and Cu-ZSM-5 catalysts for the selective catalytic reduction of nitric oxide with ammonia. Ind. Eng. Chem. Res. 2008, 47, 8588. (14) Sun, Q.; Gao, Z.-X.; Chen, H.-Y.; Sachtler, W. M. H. Reduction of NOx with ammonia over Fe/MFI: Reaction mechanism based on isotopic labeling. J. Catal. 2001, 201, 89. (15) Kro¨cher, O.; Devadas, M.; Elsener, M.; Wokaun, A.; So¨ger, N.; Pfeifer, M.; Demel, Y.; Mussmann, L. Investigation of the selective catalytic reduction of NO by NH3 on Fe-ZSM-5 monolith catalysts. Appl. Catal., B 2006, 66, 208. (16) Devadas, M.; Kro¨cher, O.; Elsener, M.; Wokaun, A.; So¨ger, N.; Pfeifer, M.; Demel, Y.; Mussmann, L. Influence of NO2 on the selective catalytic reduction of NO with ammonia over Fe-ZSM-5. Appl. Catal., B 2006, 67, 187.

51

(17) Devadas, M.; Kro¨cher, O.; Elsener, M.; Wokaun, A.; Mitrikas, G.; So¨ger, N.; Pfeifer, M.; Demel, Y.; Mussmann, L. Characterization and catalytic investigation of Fe-ZSM-5 for urea-SCR. Catal. Today 2007, 119, 137. (18) Devadas, M.; Kro¨cher, O.; Wokaun, A. Catalytic investigation of Fe-ZSM-5 in the selective catalytic reduction of NOx with NH3. React. Kinet. Catal. Lett. 2005, 86, 347. (19) Grossale, A.; Nova, I.; Tronconi, E. Study of a Fe-zeolite-based system as NH3-SCR catalyst for diesel exhaust aftertreatment. Catal. Today 2008, 136, 18. (20) Qi, G. S.; Yang, R. T. Ultra-active Fe/ZSM-5 catalyst for selective catalytic reduction of nitric oxide with ammonia. Appl. Catal., B 2005, 60, 13. (21) Koebel, M.; Elsener, M.; Kleemann, M. Urea-SCR: A promising technique to reduce NOx emissions from automotive diesel engines. Catal. Today 2000, 59, 335. (22) Koebel, M.; Madia, G.; Raimondi, F.; Wokaun, A. Enhanced reoxidation of vanadia by NO2 in the fast SCR reaction. J. Catal. 2002, 209, 159. (23) Stevenson, S. A.; Vartuli, J. C. The selective catalytic reduction of NO2 by NH3 over HZSM-5. J. Catal. 2002, 208, 100. (24) Parvulescu, V. I.; Grange, P.; Delmon, B. Catalytic removal of NO. Catal. Today 1998, 46, 233. (25) Stevenson, S. A.; Vartuli, J. C.; Brooks, C. F. Kinetics of the selective catalytic reduction of NO over HZSM-5. J. Catal. 2000, 190, 228. (26) Wallin, M.; Karlsson, C. J.; Skoglundh, M.; Palmqvist, A. Selective catalytic reduction of NOx with NH3 over zeolite H-ZSM-5: Influence of transient ammonia supply. J. Catal. 2003, 218, 354. (27) Hocking, M. B. Handbook of Chemical Technology and Pollution Control, 3rd ed.; Academic: San Diego, CA, 2005. (28) Goodman, A. L.; Underwood, G. M.; Grassian, V. H. Heterogeneous reaction of NO2: Characterization of gas-phase and adsorbed products from the reaction, 2NO(g) + H2O(a) f HONO(g) + HNO3(a) on hydrated silica particles. J. Phys. Chem. A 1999, 103, 7217. (29) Kalberer, M.; Ammann, M.; Arens, F.; Ga¨ggeler, H. W.; Baltensperger, U. Heterogeneous formation of nitrous acid (HONO) on soot aerosol particles. J. Geophys. Res. 1999, 104, 13825. (30) Ammann, M.; Kalberer, M.; Jost, D. T.; Tobler, L.; Rossler, E.; Piguet, D.; Gaggeler, H. W.; Baltensperger, U. Heterogeneous production of nitrous acid on soot in polluted air masses. Nature 1998, 395, 157. (31) Yeom, Y. H.; Henao, J.; Li, M. J.; Sachtler, W. M. H.; Weitz, E. The role of NO in the mechanism of NOx reduction with ammonia over a BaNa-Y catalyst. J. Catal. 2005, 231, 181. (32) Nova, I.; Ciardelli, C.; Tronconi, E.; Chatterjee, D.; Bandl-Konrad, B. NH3-NO/NO2 chemistry over V-based catalysts and its role in the mechanism of the fast SCR reaction. Catal. Today 2006, 114, 3. (33) Grossale, A.; Nova, I.; Tronconi, E.; Chatterjee, D.; Weibel, M. The chemistry of the NO/NO2-NH3 “fast” SCR reaction over Fe-ZSM-5 investigated by transient reaction analysis. J. Catal. 2008, 256, 312. (34) Baik, J. H.; Yim, S. D.; Nam, I. S.; Mok, Y. S.; Lee, J. H.; Cho, B. K.; Oh, S. H. Modeling of monolith reactor washcoated with Cu-ZSM-5 Catalyst for removing NO from diesel engine by urea. Ind. Eng. Chem. Res. 2006, 45, 5258. (35) Delahay, G.; Kieger, S.; Tanchoux, N.; Trens, P.; Coq, B. Kinetics of the selective catalytic reduction of NO by NH3 on a Cu-faujasite catalyst. Appl. Catal., B 2004, 52, 251. (36) Malmberg, S.; Votsmeier, M.; Gieshoff, J; So¨ger, N.; Muβmann, L.; Schuler, A.; Drochner, A. Dynamic phenomena of SCR-catalysts containing Fe-exchanged zeolitessExperiments and computer simulations. Top. Catal. 2007, 42-43, 33. (37) Chatterjee, D.; Burkhardt, T.; Weibel, M.; Nova, I.; Grossale, A.; Tronconi, E. Numerical simulation of zeolite- and V-based SCR catalytic converters. SAE Tech. Pap. 2007, 2007-01-1136. (38) Nova, I.; Ciardelli, C.; Tronconi, E.; Chatterjee, D.; Bandl-Konrad, B. NH3-SCR of NO over a V-based catalyst: Low-T redox kinetics with NH3 inhibition. AlChE J. 2006, 52, 3222. (39) Tronconi, E.; Nova, I.; Ciardelli, C.; Chatterjee, D.; Bandl-Konrad, B.; Burkhardt, T. Modelling of an SCR catalytic converter for diesel exhaust after treatment: Dynamic effects at low temperature. Catal. Today 2005, 105, 529. (40) Ciardelli, C.; Nova, I.; Tronconi, E.; Konrad, B.; Chatterjee, D.; Ecke, K.; Weibel, M. SCR-DeNOx for diesel engine exhaust aftertreatment: Unsteady-state kinetic study and monolith reactor modelling. Chem. Eng. Sci. 2004, 59, 5301. (41) Chatterjee, D.; Burkhardt, T.; Bandl-Konrad, B.; Braun, T.; Tronconi, E.; Nova, I.; Ciardelli, C. Numerical simulation of ammonia SCRcatalytic converters: Model development and application. SAE Tech. Pap. 2005, 2005-01-0965.

52


(42) Olsson, L.; Sjo¨vall, H.; Blint, R. J. A kinetic model for ammonia selective catalytic reduction over Cu-ZSM-5. Appl. Catal., B 2008, 81, 203. (43) Sjo¨vall, H.; Olsson, L.; Blint, R. J. Detailed kinetic modeling of NH3 adsorption and NH3 oxidation over Cu-ZSM-5. J. Phys. Chem. C 2009, 113, 1393. (44) Olsson, L.; Sjo¨vall, H.; Blint, R. J. Detailed kinetic modeling of NOx adsorption and NO oxidation over Cu-ZSM-5. Appl. Catal., B 2009, 87, 200. (45) Sjo¨vall, H.; Blint, R. J.; Olsson, L. Detailed kinetic modeling of NH3 SCR over Cu-ZSM-5. Appl. Catal., B 2009, 92, 138. (46) Schwidder, M.; Kumar, M. S.; Bentrup, U.; Perez-Ramirez, J.; Bruckner, A.; Grunert, W. The role of Bro¨nsted acidity in the SCR of NO over Fe-MFI catalysts. Microporous Mesoporous Mater. 2008, 111, 124. (47) Schwidder, M.; Heikens, S.; De Toni, A.; Geisler, S.; Berndt, M.; Bruckner, A.; Grunert, W. The role of NO2 in the selective catalytic reduction of nitrogen oxides over Fe-ZSM-5 catalysts: Active sites for the conversion of NO and of NO/NO2 mixtures. J. Catal. 2008, 259, 96. (48) Iwasaki, M.; Yamazaki, K.; Banno, K.; Shinjoh, H. Characterization of Fe/ZSM-5 DeNOx catalysts prepared by different methods: Relationships between active Fe sites and NH3-SCR performance. J. Catal. 2008, 260, 205.

(49) Sjo¨vall, H.; Fridell, E.; Blint, R. J.; Olsson, L. Identification of adsorbed species on Cu-ZSM-5 under NH3 SCR conditions. Top. Catal. 2007, 42-43, 113. (50) Johnston, H. S.; Foering, L.; Thompson, R. J. Kinetics of the thermal decomposition of nitric acid vapor. II. Mechanism. J. Phys. Chem. 1953, 57, 390. (51) Qi, G.; Gatt, J. E.; Yang, R. T. Selective catalytic oxidation (SCO) of ammonia to nitrogen over Fe-exchanged zeolites prepared by sublimation of FeCl3. J. Catal. 2004, 226, 120. (52) Chatterjee, D.; Burkhardt, T.; Weibel, M.; Tronconi, E.; Nova, I.; Ciardelli, C. Numerical simulation of NO/NO2/NH3 reactions on SCRcatalytic converters: Model development and applications. SAE Tech. Pap. 2006, 2006-01-0468.

ReceiVed for reView March 7, 2009 ReVised manuscript receiVed October 15, 2009 Accepted November 2, 2009 IE9003464

A Kinetic Model for the Selective Catalytic Reduction of NO

Recommend Documents