`
`www.elsevier.com/locate/apcatb
`
`Selective catalytic reduction of NOx with NH3 over
`Cu-ZSM-5—The effect of changing the gas composition
`
`Hanna Sjo¨vall a,b,*, Louise Olsson a,b, Erik Fridell b, Richard J. Blint c
`a Chemical Reaction Engineering, Chalmers University of Technology, SE-412 96 Go¨teborg, Sweden
`b Competence Centre for Catalysis, Chalmers University of Technology, SE-412 96 Go¨teborg, Sweden
`c General Motors R&D Center, Chemical and Environmental Sciences Laboratory, 30500 Mound Road, Warren, MI 48090-9055, United States
`
`Received 11 August 2005; received in revised form 29 November 2005; accepted 1 December 2005
`Available online 19 January 2006
`
`Abstract
`
`The selective catalytic reduction of nitrogen oxides (NOx) with ammonia over ZSM-5 catalysts was studied with and without water vapor. The
`activity of H-, Na- and Cu-ZSM-5 was compared and the result showed that the activity was greatly enhanced by the introduction of copper ions. A
`comparison between Cu-ZSM-5 of different silica to alumina ratios was also performed. The highest NO conversion was observed over the sample
`with the lowest silica to alumina ratio and the highest copper content. Further studies were performed with the Cu-ZSM-5-27 (silica/alumina = 27)
`sample to investigate the effect of changes in the feed gas. Oxygen improves the activity at temperatures below 250 8C, but at higher temperatures
`O2 decreases the activity. The presence of water enhances the NO reduction, especially at high temperature. It is important to use about equal
`amounts of nitrogen oxides and ammonia at 175 8C to avoid ammonia slip and a blocking effect, but also to have high enough concentration to
`reduce the NOx. At high temperature higher NH3 concentrations result in additional NOx reduction since more NH3 becomes available for the NO
`reduction. At these higher temperatures ammonia oxidation increases so that there is no ammonia slip. Exposing the catalyst to equimolecular
`amounts of NO and NO2 increases the conversion of NOx, but causes an increased formation of N2O.
`# 2005 Elsevier B.V. All rights reserved.
`
`Keywords: Ammonia; Ammonia oxidation; Cu-ZSM-5; Nitrogen oxide; NO reduction; Selective catalytic reduction (SCR); Water; Zeolite
`
`1. Introduction
`
`(NOx)
`source of nitrogen oxides
`One major
`the
`is
`combustion of fossil fuel. Nitrogen oxides may cause formation
`of ground-level ozone, production of acid rain and respiratory
`problems to mankind [1]. Oxides of nitrogen are difficult to
`reduce in the presence of excess oxygen that occurs in diesel
`exhaust. There is currently a need for a solution for NOx
`abatement in light duty diesel engines. One possible approach
`for reduction of NOx to N2 is selective catalytic reduction
`(SCR) with urea or ammonia. The use of NH3-SCR has been
`investigated for several years and is today a well established
`technique for DeNOx in stationary applications [2].
`The catalysts studied in the literature for this reaction can be
`divided in three groups that are active at different temperatures
`[3]. Noble metals, like platinum, were first considered for the SCR
`
`* Corresponding author. Tel.: +46 31 7723028; fax: +46 31 7723035.
`E-mail address: hanna.sjovall@chalmers.se (H. Sjo¨vall).
`
`0926-3373/$ – see front matter # 2005 Elsevier B.V. All rights reserved.
`doi:10.1016/j.apcatb.2005.12.003
`
`of NOx. They are active in the selective reduction of NOx at low
`temperatures, but the selectivity is poor at higher temperatures
`[1]. The second type of catalysts is metal oxides. Among the
`various investigated metal oxide mixtures, those based onvanadia
`supported on titania are commonly used. The catalyst is active at
`350–450 8C, at higher temperatures the catalyst loses selectivity
`due to enhanced oxidation of ammonia [4]. For application in a
`wider temperature range, zeolite based catalysts have been
`developed, which are active at high temperatures, to a maximum
`of about 600 8C [5]. Among these, copper exchanged zeolites
`such as Cu-ZSM-5 have been widely studied for SCR with
`ammonia and also for applications such as NO decomposition and
`selective catalytic reduction by hydrocarbons [6–9].
`NOx in the exhaust gas from a diesel engine is usually
`composed of more than 90% NO. The overall SCR reaction
`with ammonia is usually assumed to involve stoichiometric
`amounts of NO and NH3 in presence of oxygen to produce
`nitrogen and water [10]:
`4NH3 þ 4NO þ O2 ! 4N2 þ 6H2O
`
`(1)
`
`Exhibit 2023.001
`
`
`
`H. Sjo¨vall et al. / Applied Catalysis B: Environmental 64 (2006) 180–188
`
`181
`
`(3)
`
`(4)
`
`(2)
`
`The overall reaction of a mixture composed of equimolecular
`amounts of NO and NO2 may also be important, since it is
`suggested to occur much faster than the main reaction [2]:
`2NH3 þ NO þ NO2 ! 2N2 þ 3H2O
`At high temperatures the oxidation of ammonia may lead to
`formation of additional NO over the catalyst [11]:
`4NH3 þ 3O2 ! 2N2 þ 6H2O
`4NH3 þ 5O2 ! 4NO þ 6H2O
`Other reactions may be taking place over the catalyst that could
`result in the reduction of NO2 without concurrent NO reduc-
`tion. In addition, formation of other products such as nitrous
`oxide and ammonium nitrate may be occurring.
`The importance of the different reactions above varies with
`both gas composition and temperature. Due to the transient
`conditions in vehicle application it is important to investigate
`the SCR activity at several settings to reach a high average NOx
`conversion and to avoid ammonia slip. There are several studies
`that investigate zeolites such as Cu-ZSM-5 and their catalytic
`behavior in ammonia SCR [11–17]. However, neither of them
`investigates changes in the feed composition in order to
`evaluate the possible reactions
`that occur over coated
`monoliths. In this study, Cu-ZSM-5 coated catalysts of three
`different silica/alumina ratios have been evaluated. The catalyst
`that produced the highest conversion has been studied further.
`The aim is to investigate the catalytic activity at various
`concentrations of oxygen, ammonia, nitrogen oxides and water
`to provide knowledge about the different reactions that occur at
`the surface of a Cu-ZSM-5 catalyst.
`
`2. Experimental
`
`2.1. Catalyst preparation
`
`Catalysts were prepared from zeolite powder of three different
`SiO2/Al2O3 ratios obtained from Alsi-Penta. The starting
`material for the zeolites with SiO2/Al2O3 ratios of 27 and 55
`was H-ZSM-5 and the starting material for the zeolite with ratio
`300 was Na-ZSM-5. Five different catalysts, H-ZSM-5-27, Na-
`ZSM-5-27, Cu-ZSM-5-27, Cu-ZSM-5-55 and Cu-ZSM-5-300,
`were prepared according to the method described below.
`To ion exchange the H-ZSM-5 powder, a 108 mM NaNO3
`solution was stirred for 30 min and the pH was adjusted by
`adding NH3. The zeolite was then added. This ion exchange was
`performed twice. The total amount of Na+ in the slurries was two
`times the number of aluminum in the zeolite. The Na-ZSM-5 was
`placed in an oven and dried for 1 h at 80 8C, followed by a second
`drying at 125 8C for 30 min. The copper was introduced into the
`zeolite by exchange in an 11 mM Cu(CH3COO)2 solution at
`ambient temperature and the pH was adjusted with ammonia.
`The slurry was stirred for 14 h, followed by a second (8 h)
`exchange and then a third (14 h) exchange. The total amount of
`Cu2+ in the slurries was one and a half times the number of
`aluminum in the zeolite. After the last exchange, the powder was
`filtered and washed with 1 l distilled water. The Cu-ZSM-5 was
`placed in an oven and dried for 1 h at 80 8C, followed by a second
`drying at 125 8C for 5 h. For detailed information of the different
`catalysts, see Tables 1A and 1B.
`The zeolite powder was washcoated on monoliths. The
`respective slurry was composed of a liquid phase of equal
`amounts of distilled water and ethanol and a solid phase of
`
`Table 1A
`Information about the ion exchange from H+ to Na+. The volume of NaNO3 solution used in the ion exchange was determined from the number of Al present in the
`zeolite. The ion exchange was performed twice
`
`SiO2/Al2O3
`
`27
`
`55
`
`Zeolite
`weight (g)
`
`45.2
`
`31.6
`
`Volume, NaNO3
`solution (ml)
`
`Concentration,
`NaNO3 (mM)
`
`485
`485
`
`170
`170
`
`108
`108
`
`108
`108
`
`Time (h)
`
`0.5
`0.5
`
`0.5
`0.5
`
`pH
`
`7.06
`6.90
`
`7.10
`7.06
`
`Table 1B
`Information about the ion exchange from Na+ to Cu2+. The volume of Cu(Ac)2 solution used in the ion exchange was determined from the number of Al present in the
`zeolite. The ion exchange was performed three times
`
`SiO2/Al2O3
`
`Zeolite
`weight (g)
`
`Volume, Cu(Ac)2
`solution (ml)
`
`Concentration,
`Cu(Ac)2 (mM)
`
`Time (h)
`
`27
`
`55
`
`300
`
`31.2
`
`23.2
`
`31.1
`
`1643
`1643
`1643
`
`618
`618
`618
`
`151
`151
`151
`
`11
`11
`11
`
`11
`11
`11
`
`11
`11
`11
`
`14
`8
`14
`
`14
`8
`14
`
`14
`8
`14
`
`pH
`
`5.64
`5.75
`5.73
`
`5.80
`5.84
`5.85
`
`5.80
`5.70
`5.80
`
`Exhibit 2023.002
`
`
`
`182
`
`H. Sjo¨vall et al. / Applied Catalysis B: Environmental 64 (2006) 180–188
`
`Table 2
`The amount of washcoat on the five catalysts prepared. The catalysts were first
`immersed in a 5% binder slurry, then in a 20% zeolite/binder slurry and finally
`in a slurry of 5% zeolite/binder to adjust the amount of washcoat on the catalyst.
`Drying and heating were performed between all steps in the washcoating
`procedure
`
`Catalyst
`
`H-ZSM-5-27
`Na-ZSM-5-27
`Cu-ZSM-5-27
`Cu-ZSM-5-55
`Cu-ZSM-5-300
`
`Weight alumina
`layer (g)
`
`Weight zeolite layer
`(80% zeolite, 20% alumina) (g)
`
`0.092
`0.101
`0.073
`0.065
`0.078
`
`1.002
`1.018
`1.038
`1.004
`1.018
`
`20 wt.% boehmite (Disperal D) and 80 wt.% Cu-ZSM-5. Two
`slurries with different liquid/solid ratios were prepared from
`each zeolite powder
`to prepare monoliths with similar
`washcoats for each of the zeolites. The liquid/solid weight
`ratios used in the slurries were 80/20 and 95/5. The monoliths
`consisted of 188 channels (1 mm 1 mm) and the length and
`the diameter of
`the monolith were 30 mm and 22 mm,
`respectively. All monoliths were heated to 550 8C for 1 h
`prior to washcoating. A thin layer of alumina was washcoated
`on the catalyst to generate an improved surface for the zeolite
`attachment. This sample was then calcined for 1 h at 550 8C
`before introducing the zeolite layer. The slurry used to prepare
`the thin layer of alumina was composed of distilled water and
`ethanol (50:50 mixture) and 5 wt.% binder (boehmite). The
`monolith was coated with the zeolite slurry by:
` immersing the monolith in the slurry;
` blowing away the excess slurry;
` drying in air at 85 8C for 30 s;
` heating in air at 500 8C for 1.5 min.
`
`This procedure was repeated until the monolith was coated
`with the desired amount of washcoat. This catalyst was then
`calcined at 550 8C for 3 h. Properties of the washcoats of the
`different catalysts are shown in Table 2.
`
`2.2. Catalyst characterization
`
`The aluminum and copper amounts were determined using
`inductively coupled plasma and atomic emission spectro-
`metry (ICP–AES), and the result
`is shown in Table 3.
`(<0.09%), K2O (<0.08%), Na2O
`Impurities of Fe2O3
`(<0.07%) and TiO2 (<0.04%) were also identified in the
`
`Table 3
`The alumina content, copper-ion-exchange level based on atomic ratio and the
`copper loading in the washcoat of the three Cu-ZSM-5 catalysts
`
`Catalyst
`
`Al2O3 content in
`zeolite (wt.%)
`
`Ion-exchange
`level (Cu/Al)
`
`Cu loading in
`washcoat (wt.%)
`
`Cu-ZSM-5-27
`Cu-ZSM-5-55
`Cu-ZSM-5-300
`
`5.9
`4.1
`1.1
`
`0.35
`0.27
`0.10
`
`2.03
`1.09
`0.11
`
`Table 4
`The BET surface of the zeolite starting material, the copper exchanged powder
`and the copper exchanged zeolites washcoated on monoliths
`
`SiO2/Al2O3
`
`Powder
`
`Monolith
`
`H- or Na-ZSM-5
`(m2/g)
`
`Cu-ZSM-5
`(m2/g)
`
`Cu-ZSM-5
`(m2/g washcoat)
`
`27
`55
`300
`
`313
`344
`413
`
`319
`308
`412
`
`304
`299
`349
`
`three Cu-ZSM-5 powders. The specific surface areas of the
`zeolite powder and the catalysts were determined by nitrogen
`adsorption according to the BET method using a Digisorb
`2600 (Micromertics) instrument. The specific surface areas
`are shown in Table 4.
`
`2.3. Activity measurements
`
`The activity of the catalysts was tested in a flow reactor
`consisting of a horizontal quartz tube, 800 mm long with an
`inner diameter of 22 mm. The heating unit consisted of a
`heating coil, a power supply and a Eurotherm controller. The
`catalyst was placed in the quartz tube with a thermocouple
`placed about 1 cm in front of the catalyst
`to control
`the
`temperature. A second thermocouple was placed inside the
`sample to measure the catalyst temperature. The catalyst was
`sealed in the tube with quartz wool. The inlet gas composition
`was controlled by an Environics 2000 gas mixer. An FTIR (Bio-
`Rad FTS 3000 Excalibur spectrometer with a Specac Sirocco
`series heatable gas cell, P/N 24102, with a 2 m pathlength and a
`volume of 0.19 l) was used to measure the concentration of NO,
`NO2, N2O and NH3. The nitric oxide concentration was also
`measured by a chemiluminescence detector
`(CLD 700)
`connected to the system. The gas flow and the space velocity
` 1,
`respectively,
`in all
`were 3500 ml/min and 18,400 h
`experiments. Prior to each experiment, the catalyst was pre-
`treated with 8% O2 in Ar for 20 min at 500 8C and cooled in Ar.
`All experiments were performed at atmospheric pressure and
`the inert balance was argon.
`Steady-state activity tests were performed for all catalysts in
`the absence of water using a feed gas composition of 500 ppm
`NO, 500 ppm NH3 and 8% O2 in Ar. The catalyst was exposed
`to the gas mixture at 100 8C for 40 min, at 150 8C and 200 8C
`for 30 min and at 250 8C, 300 8C, 350 8C, 400 8C, 450 8C and
`500 8C for 20 min.
`The Cu-ZSM-5-27 catalyst was used in further studies as
`described below, to investigate the catalytic activity at different
`concentrations of O2, NO, NO2, NH3 and H2O. Two tests were
`performed keeping the NO and NH3 concentrations at 500 ppm
`but changing the O2 concentration to either 1% or 4%.
`Additional steady-state experiments were performed in the
`same way using 500 ppm NO, 500 ppm NH3 and 8% O2 but in
`presence of 1% or 5% H2O in order to investigate the effect of
`water.
`In another set of experiments the gas contained fixed NO and
`O2 concentrations (500 ppm NO and 8% O2) but the NH3
`
`Exhibit 2023.003
`
`
`
`H. Sjo¨vall et al. / Applied Catalysis B: Environmental 64 (2006) 180–188
`
`183
`
`concentration was varied. The experiment was performed at
`175 8C fixing the NH3 concentration at 200 ppm, 300 ppm,
`400 ppm, 500 ppm, 600 ppm, 700 ppm and 800 ppm for
`60 min. The experiment was performed both with and without
`5% H2O in the gas mixture. An additional experiment was
`performed at 175 8C changing the inlet NO concentration and
`keeping the concentration of ammonia and oxygen constant at
`500 ppm NH3 and 8% O2. Initially the catalyst was exposed to
`the mixture using 50 ppm NO for 30 min and then it was
`exposed to the mixture using 100 ppm, 150 ppm, 200 ppm,
`300 ppm, 400 ppm, 500 ppm, 600 ppm, 700 ppm and 800 ppm
`NO for 20 min at each concentration. Similar experiments were
`performed changing either the NH3 or the NO concentration
`at 350 8C.
`Experiments were also performed to investigate the NOx
`reduction at various NO2/NOx ratios at both 175 8C and
`350 8C. The gas composition was 500 ppm NOx, 500 ppm NH3
`and 8% O2 but the NO2/NOx ratio was changed. At 175 8C the
`activity was measured as the catalyst was exposed to the gas
`mixture with a NO2/NOx ratio equal to zero for 40 min. Then
`the NOx reduction activity was measured at NO2/NOx ratios
`of 0.2, 0.4, 0.5, 0.6, 0.8 and 1 for 20 min at each ratio. The
`activity was also measured at 350 8C as the catalyst was
`exposed to the gas mixture with a NO2/NOx ratio equal to zero
`for 30 min. Then the NOx reduction activity was measured at
`NO2/NOx ratios of 0.2, 0.4, 0.5, 0.6, 0.8 and 1 for 15 min at
`each ratio.
`The ammonia oxidation was investigated in a steady-state
`experiment at temperatures from 100 8C to 400 8C with a feed
`gas mixture containing 500 ppm NH3 and 8% oxygen. The
`temperature was held constant at 100 8C for 50 min, at 150 8C
`for 20 min and at 200–400 8C for 10 min. The experiment was
`repeated using 500 ppm NH3, 8% O2 and 5% H2O to investigate
`the effect of water.
`An experiment was performed to oxidize NO to NO2 at
`steady-state conditions. The catalyst was exposed to 500 ppm
`NO, 8% O2 and 5% H2O at 100 8C for 60 min. Thereafter, the
`activity for NO oxidation was measured at 150 8C, 200 8C,
`250 8C, 300 8C, 350 8C, 400 8C, 450 8C and 500 8C for 20 min
`at each temperature. The experiment was also performed
`without water present in the feed.
`
`3. Results and discussion
`
`3.1. NOx conversion over ZSM-5 catalysts
`
`The result of activity measurements over the Cu-ZSM-5-27
`catalyst is shown in Fig. 1. The catalytic activity was measured
`using a feed gas composition of 500 ppm NO, 500 ppm NH3
`and 8% O2 at temperatures from 100 8C to 500 8C. The catalyst
`was kept at each temperature for 20–40 min as shown in the
`figure. The concentration of measured NOx increases to
`500 ppm after a few minutes, but the concentration of NH3 does
`not increase until after approximately 20 min. This indicates
`that there is more ammonia than nitrogen oxides stored on the
`surface. There is ammonia slip as the temperature increases
`from 100 8C to 300 8C. At higher temperatures, ammonia is
`
`Fig. 1. Catalyst out concentrations over Cu-ZSM-5-27 from 100 8C to 500 8C
`are shown from a feed gas composition of 500 ppm NO, 500 ppm NH3 and 8%
`O2.
`
`oxidized and there is no ammonia leaving the catalyst. There is
`almost no conversion of NOx at 100 8C, but the conversion
`increases and reaches 100% at 200 8C. The NOx conversion
`decreases at temperatures from 300 8C to 500 8C, which may be
`due to the oxidation of ammonia. Less than 20 ppm of N2O is
`formed during the experiment. The NO/NO2 ratio of NOx
`exiting the catalyst decreases above 300 8C because the NO
`oxidation increases with the temperature and reaches equili-
`brium at 400 8C.
`Similar experiments were performed for the Cu-ZSM-5-55
`and the Cu-ZSM-5-300 catalysts. The steady-state conversions
`of NH3 and NOx as a function of temperature for the three Cu-
`ZSM-5 catalysts are compared in Fig. 2.
`The analysis of NO reduction by NH3 in the presence of O2
`shows a continuous NO reduction. The Si/Al ratio determines
`the number of Brønsted acid sites in ZSM-5 and therefore also
`the capacity to introduce copper ions at exchangeable sites [18].
`The Cu-ZSM-5-27 catalyst holds the largest number of acid
`sites and the result shows that this catalyst, with the lowest ratio,
`reveals the highest NOx conversion, and, as the ratio increases,
`the conversion decreases. Similar Si/Al effects were shown by
`Long and Yang [19], who reported decreasing NO conversions
`with increasing Si/Al ratio over Fe-ZSM5 catalysts with similar
`ion-exchange levels. Komatsu et al. [11] reported that a
`decreased Si/Al ratio improves the specific activity (per Cu2+
`ion) over Cu-ZSM-5, because the aluminum concentration
`increases resulting in an increase in the copper concentration.
`They also reported an increased specific activity to an ion-
`
`Fig. 2. Steady-state conversions at temperatures from 150 8C to 500 8C over
`Cu-ZSM-5 with different silica to alumina ratios. The catalysts were exposed to
`500 ppm NO, 500 ppm NH3 and 8% O2.
`
`Exhibit 2023.004
`
`
`
`184
`
`H. Sjo¨vall et al. / Applied Catalysis B: Environmental 64 (2006) 180–188
`
`exchange level of a Cu/Al ratio equal to 1, and concluded that
`the specific activity is governed by the concentration of the
`copper ions in Cu-ZSM-5. At higher Cu/Al ratios they suggest
`that aggregates of CuO may be formed which are much less
`active than Cu2+ and may block the pores of ZSM-5, resulting in
`decreased activity. The ion-exchange level differs between the
`Cu-ZSM-5 samples in this study, increasing in an order of Cu-
`ZSM-5-300 < Cu-ZSM-5-55 < Cu-ZSM-5-27. All three cata-
`lysts have a Cu/Al ratio < 0.5, which implies that the samples
`are under-exchanged, i.e. no formation of CuO is expected, and
`an enhancement in NO conversion with increased ion-exchange
`level can thus be expected. Since the Cu-ZSM-5-27 sample
`contains both the largest number of Brønsted acid sites and
`copper ions, the highest activity was observed over this catalyst.
`It can also be seen in Fig. 2 that the conversions of NO and
`NH3 are similar for the respective catalysts at low temperatures.
`This indicates that the reduction of NO occurs selectively at low
`temperatures according to reaction (1), consuming equimole-
`cular amount of NH3 and NO. At higher temperatures the
`conversion of ammonia and nitric oxide deviate, indicating that
`some ammonia is consumed without
`simultaneous NO
`reduction. It is noted that all three catalysts show decreased
`conversion of nitric oxide after reaching a maximum, but the
`conversion of ammonia shows a continuous increase to 100%
`with increasing temperature. One explanation for the decrease
`in NO reduction activity suggests that the selectivity for NO
`reduction decreases at high temperatures and a competition
`between NO reduction and NH3 oxidation occurs. If NH3 is
`oxidized, a lower amount of NH3 is available for SCR and more
`NO may be produced resulting in a lower conversion of NO.
`Several authors have observed similar behavior over various
`catalysts and attributed the decreased NOx conversion at
`high temperatures to an effect caused by ammonia oxidation
`[11,13,20,21].
`The copper exchanged catalyst was prepared from H-ZSM-5
`via Na-ZSM-5 and activity tests were performed to study the
`NOx conversion over both the H-ZSM-5 and the Na-ZSM-5 as
`described above. The steady-state results from the H-ZSM-5-
`27, Na-ZSM-5-27 and the Cu-ZSM-5-27 catalysts are
`compared in Fig. 3. The activity for reduction of nitrogen
`oxides over the H- and the Na-ZSM-5-27 catalysts is similar but
`introducing copper ions into the zeolite greatly enhances the
`catalytic activity. The result is in accordance with Komatsu
`et al. [11], who found that the conversion of nitric oxide was
`
`Fig. 3. Steady-state conversions at temperatures from 150 8C to 500 8C over
`ZSM-5-27 with different ions. The catalysts were exposed to 500 ppm NO,
`500 ppm NH3 and 8% O2.
`
`Fig. 4. Steady-state NOx reduction over the Cu-ZSM-5-27 catalyst performed
`at different O2 concentrations and temperatures from 150 8C to 500 8C. The
`catalyst was exposed to 500 ppm NO, 500 ppm NH3 and 1–8% O2.
`
`much lower for the H-ZSM-5 than for the Cu-ZSM-5 catalyst,
`and,
`that
`the reduction is catalyzed by the copper ions
`exchanged into the zeolite. Ammonia oxidation occurs in all
`catalysts but the ammonia conversion does not reach 100% over
`the H and Na catalysts.
`
`3.2. Influence of gas composition
`
`The presence of oxygen has been reported to influence both
`the reaction mechanism and the nature of the active copper
`species [6], which in turn may influence the catalytic activity. In
`this study, changing the concentration of O2 from 8% to 1% and
`4% and performing similar tests as described above examined
`the influence of O2. A comparison between the steady-state
`activity at 1%, 4% and 8% oxygen is shown in Fig. 4. The result
`reveals that the O2 concentration affects the NO conversion and
`even though the separation between the curves is rather small,
`the difference between the experiments is consistent with the
`change in oxygen concentration. Increased O2 concentration
`results in decreased conversion at high temperatures, but at low
`temperatures the opposite behavior
`is observed. Several
`explanations have been given to the observation that an
`increased oxygen concentration enhances the NO reduction rate
`at low temperatures. Two possible explanations may be that O2
`can be needed to activate NO by oxidation to nitrites and/or
`nitrates, and secondly, oxygen may be needed to maintain the
`proper oxidation state of the copper ions. Ham et al. [22]
`reported that oxygen plays an important role in the reoxidation
`of cuprous ions to cupric ions in copper exchanged mordenite.
`However, Eng and Bartholomew [23] studied NO reduction
`over H-ZSM-5 and concluded that the main role of oxygen is to
`react with NO to form an active intermediate species that can
`adsorb on the surface. Komatsu et al. [24] concluded that
`
`can be one of the key intermediates for the reduction of
`NO3
`nitric oxide with ammonia over Cu-ZSM-5. The decrease in
`conversion with increasing oxygen concentration at high
`temperatures seen in this study may be due to NH3 oxidation.
`The NH3 oxidation rate increases with higher O2 concentrations
`resulting in decreased conversion of NOx.
`Additional experiments were performed as described above
`with water present in the feed gas (Fig. 5). The NO conversion
`is enhanced by the presence of water both at low and high
`temperatures. The improved conversion is particularly pro-
`nounced at high temperatures in the region where ammonia
`
`Exhibit 2023.005
`
`
`
`H. Sjo¨vall et al. / Applied Catalysis B: Environmental 64 (2006) 180–188
`
`185
`
`Fig. 5. Steady-state NOx reduction over the Cu-ZSM-5-27 catalyst performed
`at different H2O concentrations and temperatures from 150 8C to 500 8C. The
`catalyst was exposed to 500 ppm NO, 500 ppm NH3, 8% O2 and 0–5% H2O.
`
`oxidation is important. The improved performance with the
`water concentration at high temperatures is caused by a higher
`adsorption of water, which then reduces ammonia adsorption
`and consequently results in less ammonia oxidation. This is in
`accordance with results by Sullivan et al. [12], who also
`reported an enhancing effect of water vapor on the SCR activity
`over Cu-ZSM-5 at temperatures from 100 8C to 500 8C. Sun
`et al.
`[25]
`found that water vapor suppresses ammonia
`oxidation at high temperatures, so that the temperature window
`of high NO conversion over Fe/MFI becomes wider. Salker and
`Weisweiler [14] also reported an enhanced NO reduction with
`water vapor over a Cu-ZSM-5 catalyst and concluded that the
`lower reduction without water could be due to the formation of
`a copper ammonia complex that may not be participating in
`NOx reduction but might favor ammonia oxidation. Introduc-
`tion of water would then decrease the formation of the copper
`ammonia complexes and also facilitate strong chemisorption of
`ammonia.
`It is noted that ammonia oxidation does not occur at low
`temperatures but the NOx conversion is still enhanced by water
`vapor at 150 8C. The increased low-temperature activity can,
`however, not be explained by an enhanced formation of nitrite
`and/or nitrate species as in the case with increased oxygen
`concentration, since NO oxidation is higher in dry than in
`humid feed (see Fig. 11). Even though water has been reported
`not to directly cause oxidation of copper ions in zeolites such as
`mordenite and ZSM-5 at room temperature, it has been reported
`to facilitate the reoxidation of the system by oxygen [26,27].
`However, as will be discussed below, high concentrations of
`ammonia will reduce the reaction rate due to competition for
`adsorption sites. Further, this blocking effect is influenced by
`the presence of water.
`NO reduction experiments at 175 8C with 500 ppm NO, 8%
`O2 and NH3 concentrations increasing from 200 ppm to
`800 ppm were performed with and without water (5%) to
`investigate the NO conversion at various ammonia concentra-
`tions. The time at each ammonia concentration was 60 min in
`order to reach steady state. In absence of water (Fig. 6A), the
`conversion of NO initially increases as NH3 is added to the inlet
`feed, but above 500 ppm NH3, the NOx reduction activity
`decreases. NH3 is strongly adsorbed on the surface and high
`ammonia coverage may cause the decrease in NO conversion,
`since fewer active sites become available for the reduction. A
`similar inhibition effect by ammonia over H-ZSM-5 has been
`
`Fig. 6. NOx reduction over the Cu-ZSM-5-27 at 175 8C using various con-
`centrations of ammonia. Experiment performed with inlet concentrations of:
`(A) 500 ppm NO, 8% O2 and 200–800 ppm NH3; (B) 500 ppm NO, 8% O2, 5%
`H2O and 200–800 ppm NH3.
`
`reported by Stevenson et al. [28], who found that the effect was
`most likely due to competitive adsorption, i.e. ammonia is
`blocking sites needed for the SCR reaction to take place.
`A similar experiment was also performed in the presence of
`5% water (Fig. 6B). The NOx reduction at high ammonia
`concentrations is higher as water is present in the gas mixture.
`This may be related to that the blocking effect of ammonia is
`partially suppressed by the presence of water. The low blocking
`effect
`in the presence of water may be caused by lower
`adsorption of ammonia. The NOx conversion with and without
`water is comparable at ammonia concentrations up to 400 ppm,
`indicating that the blocking effect caused by ammonia is of
`importance only at higher ammonia concentrations. The
`blocking effect at 500 ppm NH3 in dry feed causes a decreased
`NO conversion compared to the conversion in humid feed,
`which may explain the difference in NO conversion seen at low
`temperatures in Fig. 5. The introduction of water may thus
`suppress the blocking effect of NH3 and facilitate higher NO
`conversion. However,
`the exact nature of the interaction
`between ammonia and water and the effect on NO adsorption is
`unclear.
`An additional experiment was performed by varying the
`inlet NO concentration from 50 ppm to 800 ppm and exposing
`the catalyst to 500 ppm NH3 and 8% O2 at 175 8C. Fig. 7 shows
`an ammonia slip when the catalyst is exposed to more ammonia
`than nitrogen oxide. As the inlet NO concentration increases,
`more NH3 is consumed and at 700 ppm NO in the feed, all
`ammonia is consumed. Note that any inlet NO that exceeds the
`inlet NH3 cannot be reduced.
`The effect of changing NO or NH3 was also investigated at
`350 8C. Fig. 8A shows the result as the catalyst was exposed to
`500 ppm NO, 8% O2 and 50–800 ppm NH3 and in Fig. 8B the
`result from a measurement as the catalyst was exposed to
`500 ppm NH3, 8% O2 and 50–800 ppm NO is revealed. Both
`experiments indicate that a high concentration of ammonia
`compared to nitric oxide generates a high NO conversion. There
`
`Exhibit 2023.006
`
`
`
`186
`
`H. Sjo¨vall et al. / Applied Catalysis B: Environmental 64 (2006) 180–188
`
`Fig. 7. NOx reduction over the Cu-ZSM-5-27 catalyst with various NO con-
`centrations at 175 8C. The experiments were performed with the inlet con-
`centrations of 50–800 ppm NO, 500 ppm NH3 and 8% O2. Initially the catalyst
`was exposed to a mixture with 50 ppm NO for 30 min and then to 100 ppm,
`150 ppm, 200 ppm, 300 ppm, 400 ppm, 500 ppm, 600 ppm, 700 ppm and
`800 ppm for 20 min at each NO concentration.
`
`is no ammonia slip or inhibition caused by ammonia adsorption
`observed, since ammonia easily is oxidized and less NH3
`adsorption occurs at this high temperature. An increased NH3/
`NO feed ratio generates improved NO conversion since more
`NH3 is available for the reduction without causing a blocking
`effect at this high temperature.
`Experiments were performed by varying both NO and NO2
`in the feed to investigate how the conversion of NOx is affected.
`The catalyst was exposed to 8% O2, 500 ppm NOx and 500 ppm
`NH3 at either 175 8C or 350 8C. The catalyst out concentration
`measured is reported in Fig. 9. At the lower temperature steady
`state was not reached but the result still shows that the NOx
`reduction is enhanced by the introduction of NO2. The NOx
`
`Fig. 8. NO reduction over the Cu-ZSM-5-27 catalyst at 350 8C using various
`NH3 or NO concentrations. (A) Experiment performed with the inlet concen-
`trations of 50–800 ppm NH3, 500 ppm NO and 8% O2. Initially the catalyst was
`exposed to a mixture with 50 ppm NH3 for 30 min and then to 100 ppm,
`150 ppm, 200 ppm, 300 ppm, 400 ppm, 500 ppm, 600 ppm, 700 ppm and
`800 ppm for 20 min at each NH3 concentration. (B) Experiment performed
`with the inlet concentrations of 50–800 ppm NO, 500 ppm NH3 and 8% O2.
`Initially the catalyst was exposed to a mixture with 50 ppm NO for 30 min and
`then to 100 ppm, 150 ppm, 200 ppm, 300 ppm, 400 ppm, 500 ppm, 600 ppm,
`700 ppm and 800 ppm for 20 min at each NO concentration.
`
`Fig. 9. Experiments performed over the Cu-ZSM-5-27 catalyst with the inlet
`concentrations of 500 ppm NOx, 500 ppm NH3 and 8% O2. (A) The activity
`measured at 175 8C as the catalyst was exposed to the gas mixture with a NO2/
`NOx ratio equal to zero for 40 min. Then the NOx reduction activity was
`measured at 0.2, 0.4, 0.5, 0.6, 0.8 and 1 for 20 min at each NO2/NOx ratio. (B)
`The activity measured at 350 8C as the catalyst was exposed to the gas mixture
`with a NO2/NOx ratio equal to zero for 30 min. Then the NOx reduction activity
`was measured at 0.2, 0.4, 0.5, 0.6, 0.8 and 1 for 15 min at each NO2/NOx ratio.
`
`conversion reaches 100% at NO2/NOx ratios from 0.4 to 0.6 but
`decreases rapidly at higher ratios. The formation of N2O
`increases with increasing NO2/NOx ratio, but at high ratios
`where the NOx reduction decreases, the formation of N2O also
`decreases. The highest N2O formation observed at 175 8C was
`50 ppm. At 350 8C the highest conversion of NOx is reached
`when the concentration of NO equals the concentration of NO2,
`but the introduction of NO2 c