This article was downloaded by: [Basf Corporation]
`On: 08 June 2012, At: 04:50
`Publisher: Taylor & Francis
`Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office:
`Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
`
`Catalysis Reviews: Science and Engineering
`Publication details, including instructions for authors and subscription
`information:
`http://www.tandfonline.com/loi/lctr20
`The State of the Art in Selective Catalytic
`Reduction of NOx by Ammonia Using
`Metal‐Exchanged Zeolite Catalysts
`Sandro Brandenberger a , Oliver Kröcher a , Arno Tissler b & Roderik Althoff b
`a Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerland
`b Süd‐Chemie AG, Bruckmühl, Germany
`Available online: 07 Jan 2009
`
`To cite this article: Sandro Brandenberger, Oliver Kröcher, Arno Tissler & Roderik Althoff (2008): The State
`of the Art in Selective Catalytic Reduction of NOx by Ammonia Using Metal‐Exchanged Zeolite Catalysts,
`Catalysis Reviews: Science and Engineering, 50:4, 492-531
`
`To link to this article: http://dx.doi.org/10.1080/01614940802480122
`
`PLEASE SCROLL DOWN FOR ARTICLE
`
`Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions
`
`This article may be used for research, teaching, and private study purposes. Any substantial
`or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or
`distribution in any form to anyone is expressly forbidden.
`
`The publisher does not give any warranty express or implied or make any representation that the
`contents will be complete or accurate or up to date. The accuracy of any instructions, formulae,
`and drug doses should be independently verified with primary sources. The publisher shall not
`be liable for any loss, actions, claims, proceedings, demand, or costs or damages whatsoever or
`howsoever caused arising directly or indirectly in connection with or arising out of the use of this
`material.
`
`Exhibit 2022.001
`
`

`
`Catalysis Reviews, 50:492–531, 2008
`Copyright # Taylor & Francis Group, LLC
`ISSN: 0161-4940 print 1520-5703 online
`DOI: 10.1080/01614940802480122
`
`The State of the Art
`in Selective Catalytic
`Reduction of NOx
`by Ammonia Using
`Metal-Exchanged
`Zeolite Catalysts
`
`Sandro Brandenberger,1 Oliver Kro¨ cher,1 Arno Tissler,2 and
`Roderik Althoff2
`1Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerland
`2Su¨ d-Chemie AG, Bruckmu¨ hl, Germany
`
`An overview is given of the selective catalytic reduction of NOx by ammonia (NH3-SCR)
`over metal-exchanged zeolites. The review gives a comprehensive overview of NH3-SCR
`chemistry, including undesired side-reactions and aspects of the reaction mechanism
`over zeolites and the active sites involved. The review attempts to correlate catalyst
`activity and stability with the preparation method, the exchange metal, the exchange
`degree, and the zeolite topology. A comparison of Fe-ZSM-5 catalysts prepared by differ-
`ent methods and research groups shows that the preparation method is not a decisive
`factor in determining catalytic activity. It seems that decreased turnover frequency
`(TOF) is an oft-neglected effect of increasing Fe content, and this oversight may have
`led to the mistaken conclusion that certain production methods produce highly active
`catalysts. The available data indicate that both isolated and bridged iron species partici-
`pate in the NH3-SCR reaction over Fe-ZSM-5, with isolated species being the most
`active.
`
`Keywords NH3-SCR chemistry, Selective catalytic reduction of NO with ammonia,
`Metal-exchanged zeolites, Fe-ZSM-5, Active site, Hydrothermal aging,
`Preparation method, Exchange degree, Activity, Stability
`
`Received 24 January 2008; accepted 25 March 2008.
`Address correspondence to Oliver Kro¨cher, Paul Scherrer Institute, CH-5232 Villigen
`PSI, Switzerland. E-mail: oliver.kroecher@psi.ch
`
`492
`
`Downloaded by [Basf Corporation] at 04:50 08 June 2012
`
`Exhibit 2022.002
`
`

`
`Selective Catalytic Reduction of NOx by Ammonia
`
`493
`
`INTRODUCTION
`Nitrogen oxides (NOx, x ¼ 1, 2), which result from the combustion of fossil
`fuels, are a major source of air pollution (1, 2). Most of the NOx (about 60%
`in Europe) is produced from combustion processes in engines (thermal NOx)
`by the oxidation of atmospheric nitrogen at very high temperatures (1, 3):
`N2 þ O2 ! 2NO
`
`ð1Þ
`
`NOx formed by oxidation of organic nitrogen present in fuel is less significant
`nowadays, because the nitrogen content in gasoline and diesel has fallen signifi-
`cantly over the last 10 years. NOx from diesel engines contributes about 75% of
`the total NOx emissions of road traffic (3). Many efforts have been made to
`minimize NOx emission either by combustion control or by post-combustion
`abatement technologies. While combustion control and engine management
`are adequate for compliance with current NOx emission limits (4–6), legislation
`will become more stringent in the future, such that engine management alone
`will no longer be sufficient, and techniques to treat exhaust gas will become
`mandatory (4, 6).
`Catalytic technologies are preferred for emission abatement because of
`their low costs and high efficiency (7). Although NO is thermodynamically
`unstable with respect to the decomposition to N2 and O2 (reverse of reaction
`1), this reaction is kinetically very difficult to achieve, and to date no catalyst
`has been found that supports appreciable conversions under the conditions
`found in a vehicle (4, 8). Although NOx from gasoline is very efficiently
`reduced by means of a three-way catalyst (8, 9), this technology cannot be
`applied in a diesel engine because it operates under oxygen excess. There
`exists, however, a very elegant reaction to remove nitrogen oxides from lean
`exhaust gases: the selective catalytic reduction with NH3 to nitrogen (NH3-
`SCR) (10). The special feature of this reaction is that a stoichiometric dosage
`of ammonia is sufficient for total NOx conversion. The SCR process has been
`used for several decades to reduce NOx emitted from stationary power plants
`(11, 12), and it has now matured enough to be applied to diesel vehicles as
`well (10, 13–15).
`The abbreviation “SCR” is also widely used for the reduction of NOx with
`hydrocarbons (HC-SCR) or hydrogen (H2-SCR) (16, 17). This use of the label
`is misleading, however, since these reducing agents have to be applied in
`several-fold excess relative to the stoichiometry of underlying reaction
`equations, which should therefore not be called “selective.” Nevertheless,
`strenuous efforts have been made to develop these alternative NOx reduction
`technologies, because the required reducing agents are either available as
`diesel fuel (HC-SCR) or may be easily produced from fuel by steam reforming
`(H2-SCR). Another technology for the reduction of nitrogen oxides in lean
`exhaust gases is the NOx storage and reduction (NSR) catalyst. It utilizes a
`
`Downloaded by [Basf Corporation] at 04:50 08 June 2012
`
`Exhibit 2022.003
`
`

`
`494
`
`S. Brandenberger et al.
`
`basic oxide (e.g., barium oxide), which stores NOx in the form of nitrates during
`the normal, lean operation mode of the engine (18). For a short period, the
`exhaust is made rich, thereby allowing the desorption and reduction of NOx
`with CO and hydrocarbons. In stationary SCR applications, ammonia is
`usually used as the reducing agent.
`One of the first articles describing an SCR system to be applied in a vehicle
`was published by Held et al. (19), who suggested using non-toxic urea as the
`reducing agent, from which ammonia could be released under hydrothermal
`conditions. Urea-SCR is a rather complex technology due to the difficulties of
`using the right amount of urea as a function of constant changes in NOx emis-
`sions, catalyst activity, and ammonia stored on the catalyst (6, 20). The process
`has been improved step by step in the last decade (13, 21, 22), and reduction
`efficiencies in the range of 80–90% or even higher have been demonstrated
`(14, 15, 23–25). The catalysts used for SCR in industry are based mainly on
`TiO2-supported V2O5, promoted with WO3 (26, 27). This catalyst type has
`also been used since 2005 for HD diesel vehicles in Europe (4). Although
`SCR technology based on vanadia catalysts has been introduced into the
`market for diesel vehicles, problems remain due to the high activity for oxi-
`dation of SO2 to SO3, the rapid decrease in activity and selectivity at 5508C,
`and the toxicity of the vanadia species, which begin to volatilize above 6508C
`(24, 28, 29). Hence, researchers continue to work to develop new SCR catalysts;
`among these, metal-exchanged zeolites and pillared clays have received much
`attention in recent years. Zeolite types with narrow pores, such as MOR, FER,
`BEA, and ZSM-5 have proved to be especially suited for SCR applications.
`In this study, we summarize the literature on NH3-SCR over metal-
`exchanged zeolites, focusing on mechanistic aspects and the influence of
`various parameters on the activity and stability of the catalysts.
`
`CHEMISTRY OF THE SCR PROCESS AND MECHANISTIC
`ASPECTS
`
`Basics of SCR Chemistry
`The main feature of the SCR process is the use of a reducing agent to react
`specifically with nitrogen oxides, but not with the excess oxygen in the lean
`exhaust gas. N-containing compounds like ammonia or urea are well suited
`to this purpose. Because of its negligible toxicity, urea is used in diesel
`vehicles as a storage compound for ammonia. The decomposition reactions of
`urea in the presence of water are (30):
`NH2-CO-NH2 !DT
`NH3 þ HNCOðisocyanic acidÞ; and
`HNCO þ H2O ! NH3 þ CO2:
`
`ð2Þ
`ð3Þ
`
`Downloaded by [Basf Corporation] at 04:50 08 June 2012
`
`Exhibit 2022.004
`
`

`
`Selective Catalytic Reduction of NOx by Ammonia
`
`495
`
`In the following discussion, we restrict ourselves to ammonia as the selec-
`tive reducing agent, though all statements about the SCR mechanism can also
`be applied to urea, except for the preparation of the reducing agent.
`It is well known that the exhaust gas of today’s diesel engines contains
`nitrogen oxides mainly in the form of nitrogen monoxide (NO); only a minor
`fraction is nitrogen dioxide (NO2) (20). Therefore, the basic reaction on SCR
`catalysts is (15, 20, 30):
`4NH3 þ 4NO þ O2 ! 4N2 þ 6H2O:
`
`ð4Þ
`
`The reaction of a 1:1 mixture of NO and NO2 is definitely faster than reaction
`(4) (15, 30, 31):
`
`4NH3 þ 2NO þ 2NO2 ! 4N2 þ 6H2O:
`
`ð5Þ
`
`This reaction will become more important in future diesel engines, since
`new engine technologies, such as exhaust gas recycling (EGR) and especially
`the homogenous charge compression ignition (HCCI), will increase the NO2
`share of NOx emissions.
`If the NO2/NOx fraction exceeds 50%, an SCR reaction with pure NO2
`takes place (20, 30):
`
`4NH3 þ 3NO2 ! 3:5N2 þ 6H2O:
`
`ð6Þ
`
`There is also an SCR reaction with NO in the absence of O2, which changes
`the stoichiometry of NH3:NO from the usual 1:1 ratio to 2:3 (20, 30):
`4NH3 þ 6NO ! 5N2 þ 6H2O:
`
`ð7Þ
`
`The rate of reaction (7) is very slow and can be neglected in lean combustion
`gases (20).
`
`Mechanistic Aspects of SCR Reactions
`Experiments with isotopically labeled reactants using noble metals (32) and
`a Ba-Na-Y zeolite catalyst (33) have shown that N2 takes one nitrogen atom
`from a molecule of NOx and the other nitrogen from ammonia, which is in agree-
`ment with SCR reactions (4) and (5). However, the uniqueness of the SCR
`process over zeolites is that NO first has to be oxidized to NO2, which is the
`rate-determining step of the mechanism (15, 28, 34–36). Thus, reaction (5)
`can be regarded as the actual and general SCR reaction over zeolites, and
`NO2 has to be provided by the input gas or must be produced by the oxidation
`of NO. This is evident in the fact that H-ZSM5 shows negligible SCR activity
`if only NO is present, and high SCR activity after 50% NO2 is added to the
`feed (15, 28, 37). On metal-exchanged zeolites the NO2 is produced at the
`metal centers, while the SCR reaction itself takes place within the zeolite
`framework (15, 28). The NO2 is consumed immediately after production in
`
`Downloaded by [Basf Corporation] at 04:50 08 June 2012
`
`Exhibit 2022.005
`
`

`
`attice NO
`
`l
`
`496
`
`S. Brandenberger et al.
`
`the SCR process (as NO2,ads), so it does not show up as gas-phase NO2. Devadas
`et al. (15) noticed that SCR activity is higher on Fe-ZSM-5 than on H-ZSM-5,
`even when the optimum NO22NO2 ratio of 50% is used over both, and they
`suggested that iron not only facilitates NO oxidation, but also promotes the
`SCR reaction over the zeolite framework. This contradicts the results of Long
`þ
`and Yang (36), who found the same reactivity of NH4
`ions with an equimolar
`NO22NO2 ratio on H-ZSM-5 and Fe-ZSM-5 at 2008C.
`The adsorption of NO on Fe-ZSM-5 has been investigated by Diffuse Reflec-
`tance Infrared Fourier Transformation DRIFT spectroscopy, and these studies
`22NO species (n ¼ 2, 3), Fe2þ
`revealed the formation of intermediate Fenþ
`(NO)2
`þ
`þ
`complexes, and NO
`(38–40). The NO
`is formed by metal ions on cationic
`exchange sites [Eq. (8)] (39, 41) or on Brønsted acid sites (40) [Eq. (9)]:
`NO þ Fe3þ ! NO
`þ þ Fe2þ
`; and
`2NO þ 1=2O2 þ 2Olattice H ! 2O
`
`þ þ H2O:
`
`ð8Þ
`ð9Þ
`
`Based on the abovementioned results, the mechanism of NO oxidation over
`iron-exchanged zeolites can be formulated as follows (34):
`NO þ Fe3þ O ! Fe3þ
`22O-NO;
`22O22NO ! Fe2þ þ NO2;ads;
`Fe3þ
`NO2;ads ! NO
`; and
`! Fe3þ
`Fe2þ þ 1
`22O:
`2O2
`
`ð10Þ
`ð11aÞ
`ð11bÞ
`ð12Þ
`
`"2
`
`Delahay et al. (34) suggested that the rate of the NO oxidation reaction is
`controlled by the desorption of NO2 [Eq. (11)b)], which is consistent with the
`immediate consumption of the NO2,ads formed in the SCR process, such that
`no observable gas-phase NO2 is produced. Koebel et al. (42) argued that for
`vanadia-based SCR catalysts, NO2 rather than O2 regenerates the reduced
`metal site (V4þ
`) because of the stronger oxidizing power of the former; the
`same may be expected to be true for Fe2þ
`-based catalysts [Eq. (12)]. For Cu-
`FAU, on the other hand, Delahay et al. (43) suggested that NO2 contributes
`less than O2 to the oxidation of CuI to CuII. However, it is obvious that under
`oxygen-free conditions, the metal remains in the reduced state (e.g., Fe2þ
`)
`and the catalytic cycle is not closed (34, 42). In fact, Long and coworkers
`(24, 35) noticed only a minor SCR activity over Fe-ZSM-5 without oxygen.
`Moreover, it is known that while simple evacuation at elevated temperature
`suffices to reduce Fe3þ
`in Fe-ZSM-5 to Fe2þ, the reverse process takes place
`
`only in oxidizing media (44, 45).
`Another important feature of the SCR reaction over iron zeolites is the fact
`that the NO2-SCR reaction (6) is only slightly slower than the NO/NO2-SCR
`
`Downloaded by [Basf Corporation] at 04:50 08 June 2012
`
`Exhibit 2022.006
`
`

`
`Selective Catalytic Reduction of NOx by Ammonia
`
`497
`
`reaction (5), but much faster than the NO-SCR reaction (4). This can be attrib-
`uted to the facile formation of reactive nitrates on the catalyst surface in the
`presence of gaseous NO2 (15):
`3NO2 þ ½O2Š ! NO þ 2NO
`
`:
`
`3
`
`ð13Þ
`
`However, a surface reaction involving three NO2 molecules is rather unlikely.
`The formation of surface nitrates might be better described by reactions (14)
`and (15) (15, 46, 47), since gaseous NO2 readily dimerizes at low temperatures
`to form N2O4 (39, 48). Yeom et al. (47) suggested that N2O4 can disproportion-
`þ
`2) ions:
`ate on BaNa-Y to nitrosyl (NO
`) and nitrate (NO3
`2NO2 ! N2O4; and
`N2O4 ! NOþ þ NO
`
`
`:
`
`3
`
`ð14Þ
`ð15Þ
`
`However, they suggested the alternative reaction of NO with NO2 to N2O3,
`which might also easily adsorb on the zeolite surface (39, 48). N2O3 was
`proposed as a reactive intermediate during NO-SCR on Cu-ZSM-5 and
`Fe-ZSM-5 (49, 50):
`
`NO þ NO2 ! N2O3
`
`ð16Þ
`
`Chen et al. (51) explained the function of N2O3 as direct precursor of
`NH4NO2, which easily decomposes to the desired product nitrogen already
`below 1008C. These researchers argued that with a ratio of NO/NO2 ¼ 1:1
`the highest possible yield of N2O3 (reaction 16) is achieved and the SCR
`activity is maximized (see reaction 5). Moreover, the nitrogen in this
`compound has the appropriate oxidation state (N3þ
`) to react directly with the
`N32 nitrogen of the ammonia-forming NH4NO2 (reaction 20). However, since
`according to Delahay (34) the oxidation of NO to NO2 is controlled by the
`rate of NO2 desorption, NO-SCR must involve the formation of N2O3 from
`NOads and NO2,ads on the catalyst surface, and not in the gas phase from NO
`and desorbed NO2.
`As the reaction proceeds, HNO2 is formed by interaction of NO
`or by the reduction of HNO3 with NO:
`þ þ H2O ! HNO2 þ H
`þ
`NO
`; and
`HNO3 þ NO ! HNO2 þ NO2:
`
`þ
`
`with H2O
`
`ð17Þ
`ð18Þ
`
`2 or the decompo-
`The HNO3 in reaction (18) results from the protonation of NO3
`þ
`sition of NH4NO3 (reaction 25). The NO
`in reaction (17) occurs by dispropor-
`tionation of N2O4 (reaction 15) or on metal ions (reaction 8) and Brønsted-acid
`þ
`sites (reaction 9). The production of NO
`via reaction 8 would explain the above
`mentioned findings from Devadas et al. (15), who reported that on Fe-ZSM-5,
`iron facilitates not only NO oxidation, but also promotes the SCR reaction
`
`Downloaded by [Basf Corporation] at 04:50 08 June 2012
`
`Exhibit 2022.007
`
`

`
`498
`
`S. Brandenberger et al.
`
`over the zeolite framework. It is well accepted that in the SCR process over
`metal-exchanged zeolites, NH3 is activated on Brønsted acid sites to yield
`þ
`NH4
`ions. In an SCR mechanism first proposed by Sun et al. (33) and
`further developed by Yeom et al. (48), the interaction of NH3 with HNO2 or
`N2O3 leads to the formation of NH4NO2 (33, 48):
`HNO2 þ NH3 ! NH4NO2; and
`N2O3 þ 2NH3 þ H2O ! 2NH4NO2:
`
`ð19Þ
`ð20Þ
`
`It is well known that ammonium nitrite (NH4NO2) decomposes quickly at
`temperatures below 1008C to yield the SCR products N2 and H2O (48, 52, 53):
`ð21Þ
`NH4NO2 ! N2 þ 2H2O:
`
`Li et al. (54) showed that reaction (21) is very efficiently catalyzed by
`zeolites due to the protonation of NH4NO2 by the Brønsted acid sites, leading
`to an intermediate product that decomposes even at room temperature.
`Besides ammonium nitrite (NH4NO2), ammonium nitrate (NH4NO3) is also
`formed (15, 55):
`
`ð22Þ
`ð23Þ
`
`2NO2 þ 2NH3 ! NH4NO3 þ N2 þ H2O, and
`þ NH
`! NH4NO3:
`NO
`
`3
`
`þ4
`
`NH4NO3 forms deposits on the catalysts at low temperatures, thereby
`blocking the active sites. Upon heating above 2008C, ammonium nitrate decom-
`poses to NH3 and HNO3 or nitrous oxide (N2O), a highly undesirable gas due to
`its greenhouse activity and its ability to destroy ozone in the upper atmosphere
`(15, 30, 48):
`
`NH4NO3 ! N2O þ 2H2O; and
`NH4NO3 ! NH3 þ HNO3:
`
`ð24Þ
`ð25Þ
`
`Centi et al. (56) explained the formation of N2O over Cu/Al2O3 at 250–
`3008C by the formation of ammonium nitrate, followed by its decomposition
`to N2O. However, experiments in our lab (57) showed that N2O is formed
`from NH4NO3 only in the absence of water, not under SCR conditions. There-
`fore, we assume that N2O is formed if an activated complex, such as the one
`shown in Figure 2, is directly attacked by NO2. Delahay et al. (58) reported
`that aggregates of CuO lead to a significant formation of N2O over Cu-FAU
`at low temperature, and they assumed that neighboring Cu ions are required
`for the formation of N2O. However, while the decomposition of ammonium
`nitrate is relatively fast at temperatures above 2008C, at lower temperatures
`it is insufficient to unblock active sites (15, 55).
`The HNO3 that evolves from the decomposition of ammonium nitrate is
`very efficiently reduced by NO over the catalyst according to reaction (18)
`
`Downloaded by [Basf Corporation] at 04:50 08 June 2012
`
`Exhibit 2022.008
`
`

`
`Selective Catalytic Reduction of NOx by Ammonia
`
`499
`
`(59) which “cleans” the zeolite surface by pushing the equilibrium of (25) to the
`right. Note that the direct reduction of ammonium nitrate (reaction 26) to
`ammonium nitrite as stated by Yeom et al. (48) presumably does not take
`place on zeolites (59).
`NH4NO3 þ NO ! NH4NO2 þ NO2
`
`ð26Þ
`
`With rising temperature, and therefore rising oxidation activity, metal-
`exchanged zeolites tend to form N2O, which decreases the selectivity of the
`catalyst. Five reactions could conceivably lead to the formation of N2O:
`2NH3 þ 2NO2 ! N2O þ N2 þ 3H2O
`3NH3 þ 4NO2 ! 3:5N2O þ 4:5H2O
`2NH3 þ 2O2 ! N2O þ 3H2O
`4NH3 þ 4NO2 þ O2 ! 4N2O þ 6H2O
`4NH3 þ 4NO þ 3O2 ! 4N2O þ 6H2O
`
`ð27Þ
`ð28Þ
`ð29Þ
`ð30Þ
`ð31Þ
`
`It should, however, be mentioned that reactions (29) to (31) have not
`been observed on metal-exchanged zeolites under typical SCR conditions
`(15, 23, 60, 61).
`The N2O formed on zeolites containing transition metals decomposes to
`nitrogen and oxygen at higher temperatures due to reaction (32) or to an
`even easier reaction, namely, reduction by ammonia to nitrogen (33). Thus,
`N2O appears only at low or intermediate temperatures over different iron
`and copper zeolites (15, 62–66):
`2N2O ! 2N2 þ O2
`3N2O þ 2NH3 ! 4N2 þ 3H2O
`
`ð32Þ
`ð33Þ
`
`The rate constant for reaction (32) is very low for ZSM-5 zeolites at low
`metal loading, but it jumps to a much higher rate for M/Al  0.4 (M ¼ Fe,
`Cu) (67, 68). More N2O is formed over Cu-ZSM-5 in comparison to Fe-ZSM-5,
`and zeolites promoted with noble metals (Pt, Rh, Pd) are even worse (69–71)
`which is why overly strong oxidizing elements should be avoided. For the
`decomposition of N2O over metal-exchanged zeolites, a mechanism has been
`proposed in which N2O is dissociatively adsorbed on an active metal site ( )
`to form N2 and an activated surface oxygen O (bounded to a metal atom)
`(72, 73):
`
`N2Oþ ! N2 þ O
`
`
`
`ð34Þ
`
`This surface oxygen is able to attack a second N2O molecule [reaction (35a)]
`or to recombine to O2 [reaction (35b)]. In the presence of NH3 the active oxygen
`
`Downloaded by [Basf Corporation] at 04:50 08 June 2012
`
`Exhibit 2022.009
`
`

`
`500
`
`S. Brandenberger et al.
`
`is removed by reaction 35c (62).
`N2O þ O
` ! N2 þ O2þ
` ! O2 þ 2
`2O
` þ 2NH3 ! N2 þ 3H2O þ 3
`3O
`
`ð35aÞ
`ð35bÞ
`ð35cÞ
`
`Combination of reactions (35a) and (35c) results in the stoichiometry
`shown in reaction (33). Reaction (33) implies that the N2O decomposition is
`accelerated by NH3, which has been confirmed by Mauvezin et al. (74).
`At still higher temperatures, the oxidizing properties of the catalyst become
`even more pronounced. This results in a direct oxidation of ammonia to N2 or
`NO, thus limiting the maximum possible NOx conversion (30, 56, 75, 76):
`4NH3 þ 5O2 ! 4NO þ 6H2O
`4NH3 þ 3O2 ! 2N2 þ 6H2OðSCO ¼ selective catalytic oxidationÞ
`
`ð36Þ
`ð37Þ
`
`Long and Yang (71) have found that Fe-, Cu- and Cr-ZSM-5 catalyze reaction
`(37) quite efficiently, whereas Mn, Co, and Ni are relatively inactive exchange
`metals. Among the transition metal oxides supported on Al2O3, the following
`were reported to be active for the SCO reaction: Mn, Co, Fe, and Ni (77).
`Reactions (4) to (37) are all connected with one another: for example, the
`stoichiometry of reactions (16, 20), and (21) is consistent with that of reaction
`(5), and this stoichiometry explains why an optimal N2 yield is obtained for
`an equimolar mixture of NO/NO2. Figure 1 summarizes the state of knowledge
`about the reaction scheme of the SCR process over metal-exchanged zeolites,
`
`Figure 1: Reaction pathways for SCR over metal-exchanged zeolites (numbers in brackets
`correspond to equations in the text).
`
`Downloaded by [Basf Corporation] at 04:50 08 June 2012
`
`Exhibit 2022.010
`
`

`
`Selective Catalytic Reduction of NOx by Ammonia
`
`501
`
`illustrating the role of nitrates and nitrites as key compounds in the SCR
`mechanism. Note that we decided not to include gas-phase chemistry reactions
`from NO2 (e.g., reaction 14) in Fig. 1 because NO2 likely remains adsorbed on
`the catalyst during the SCR reaction (78).
`Alternatively, as depicted in Fig. 2, it has been proposed that once NH3 has
`þ
`been activated, NO2 reacts with two adjacent NH4
`ions to form an (NH4)xNO2
`complex (with x ¼ 1, 2) (79, 80), which is similar to the key compound proposed
`by Sun et al. (33) and Yeom et al. (48). This requirement of two adjacent
`Brønsted acid sites has been confirmed by transient tests, which show that
`approximately half of the adsorbed NH3 is active for N2 formation (80). Next,
`NO from the gas-phase reacts with the activated complex to form an intermedi-
`ate that decomposes further to N2 and H2O (Fig. 2).
`
`REACTION KINETICS
`
`A detailed description of the kinetic models proposed for the SCR reaction is
`beyond the scope of this review. However, some general results will be briefly
`discussed here.
`In the SCR reaction as described in Fig. 1, the NO conversion rate is
`thought to depend on the concentration of NO, NH3, and O2. Water should
`also be taken into consideration, however, since it is a reaction product and
`
`Figure 2: Reaction mechanism of the SCR process over metal-exchanged zeolites with two
`adjacent Brønsted acid sites according to Long et al. (79) and (80) Eng et al.
`
`Downloaded by [Basf Corporation] at 04:50 08 June 2012
`
`Exhibit 2022.011
`
`

`
`502
`
`S. Brandenberger et al.
`
`is also present in the feed under actual SCR conditions. Therefore the reaction
`rate can be described by the following simple power rate law:
`
`rNO ¼ k caNO cb
` cc
` cd
`H2O:
`
`ð38Þ
`
`NH3
`
`O2
`
`The reaction order with respect to NO (a) has been measured by many
`research groups to be 1 on metal-exchanged zeolites (35, 43, 81–83), with
`activation energies ranging between 44 and 61 kJ/mol (35, 80–81, 84–88)
`for H, Fe, and Cu-MOR and -ZSM-5 zeolites, respectively. These activation
`energies are typical for the reaction-controlled regime and imply that diffusion
`is not rate-limiting in the SCR reaction with NOx ¼ 100% NO (reaction 4), as
`proposed in (24) and verified in (35). However, Devadas et al. (15) measured
`an activation energy of only 7 kJ/mol for SCR reaction (5) with NOx ¼ 50%
`NO and NO2, which is typical for a diffusion-controlled process (89) resulting
`from the very high rate of the NO/NO2-SCR reaction.
`According to many authors the reaction order with respect to NH3 (b) is
`about zero (35, 43, 81, 82). The observed reaction order of NO and NH3 can
`be interpreted in such a way that ammonia reacts from a strongly adsorbed
`state and NO from the gaseous, or weakly adsorbed, state. This is consistent
`with findings of Eng et al. (80), who found that an intermediate NOx species
`þ
`forms a complex with NH4
`and does not merely adsorb onto adjacent sites.
`However, Choi et al. (87) favored the Langmuir-Hinshelwood mechanism,
`which posits a reaction between NH3 and NO adsorbed on the surface of the
`mordenite catalyst. Delahay et al. (43) studied SCR kinetics on a Cu-FAU
`catalyst between 185 and 2408C and derived a rate law based on a Mars and
`van Krevelen mechanism. This model involved three sequential reactions: (i)
`the oxidation of CuI to CuII-oxo by O2, (ii) the reaction of CuII-oxo with NO to
`CuII-NxOy, and (iii) the reaction of CuII-NxOy with NH3 to N2 and H2O and
`the regeneration of CuI. Using this model the authors postulated that CuI is
`oxidized by O2 rather than by NO2, which is itself formed from a gas-phase
`reaction between NO and O2, and that the reaction of NO with CuII-oxo is irre-
`versible, as indicated by the formation of nitrite/nitrate-like species. Results
`from Coq et al. (90) showed that at 2508C the oxidation of CuI to CuII-oxo is
`the rate-determining step on Cu-FAU, but at 4508C the reduction of CuII to
`CuI becomes rate-determining. These findings are contradictory to results on
`Fe-ZSM-5 zeolites (45), which showed an increase in the amount of reduced
`Fe2þ
`species when heated to 7008C. This indicates that the reduction of Fe3þ
`to Fe2þ
`may be not the rate-determining step in the process at high
`temperature.
`Results in our lab (23) and from De Toni et al. (78) show that over Fe-ZSM-5,
`the measured SCR rates are much higher than the measured rate of NO oxi-
`dation to NO2. However, since NO oxidation is the essential first step in the
`SCR mechanism on Fe-ZSM-5 (15, 28, 34, 35), the reaction rate of the overall
`
`Downloaded by [Basf Corporation] at 04:50 08 June 2012
`
`Exhibit 2022.012
`
`

`
`Selective Catalytic Reduction of NOx by Ammonia
`
`503
`
`SCR process should be equal to the slower reaction rate of the NO oxidation
`to NO2. This apparent conflict leads to the conclusion that the desorption
`of NO2 from the catalyst surface is the true rate-determining step in the
`NO oxidation reaction. This desorption of NO2 does not play a role in
`the SCR reaction, since NO2 is immediately consumed after its production as
`NO2,ads.
`Taking a closer look at the reaction order of NH3, many authors found a
`slightly negative value (b , 0) at low temperatures (T  3508C), which is
`typical for an inhibition of the SCR reaction by NH3. This effect is explained
`by competitive adsorption of NH3 and NO on the same sites (23, 64, 88), or
`by competitive consumption of adsorbed O2 by gaseous NH3, which decreases
`the NO oxidation activity and consequently the SCR activity (35). In addition
`to these explanations, Devadas et al. (91) proposed that the inhibition effect
`might also be due to an enhanced reduction of Fe3þ
`to Fe2þ
`at higher NH3 con-
`centrations, but they could not unequivocally distinguish between the expla-
`nations. They argued that at high temperatures Fe2þ
`might be reoxidized by
`oxygen from the zeolite matrix, as proposed by Voskoboinikov et al. (92),
`which would therefore eliminate the observed inhibition. In fact, Battiston
`et al. (93) observed a decrease in the oxidation state of iron from 2.9 to 2.3
`after adding NH3 to a gas feed containing exclusively NO in N2.
`Under real-life conditions, O2 and H2O are present in large excess (94), and
`their rate dependencies can be neglected. At low concentrations, however, the
`reaction order with respect to O2 has been found to be 0.5 on metal-exchanged
`zeolites (35, 81). H2O inhibits the SCR reaction (d , 0) and reduces N2O
`formation, particularly at high temperatures. Therefore its presence
`improves catalyst selectivity, which has been related to a reduced oxidation
`activity (15, 23, 33).
`In view of the above findings, SCR kinetics can be modeled with a simplified
`rate law also proposed in (10, 87):
`rNO  k cNO:
`
`ð39Þ
`
`NUCLEARITY OF THE ACTIVE SITE
`
`There is still debate in the literature about the identity and nuclearity of sites
`active for SCR chemistry on metal-exchanged zeolites. In copper zeolites,
`binuclear species are clearly favored. Kieger et al. (95) proposed that for
`NO-SCR over Cu-FAU at temperatures below 3008C, the active sites are
`formed by neighboring Cu ions, which may be [CuOCu]2þ
`dimer species;
`above 3508C, however, all Cu ions become active. Iwamoto et al. (96) concluded
`that paired Cu2þ
`species are the active sites in Cu-ZSM-5 at 4508C, which
`seems very likely in view of the strong increase in activity with rising Cu
`content above a certain Cu/Al ratio. Moreover, Komatsu et al. (81) observed
`
`Downloaded by [Basf Corporation] at 04:50 08 June 2012
`
`Exhibit 2022.013
`
`

`
`504
`
`S. Brandenberger et al.
`
`the same relation between the specific activity and the copper concentration
`and an increasing specific activity with decreasing Si/Al ratio of the parent
`ZSM-5 zeolite. They also proposed paired Cu2þ
`to be the active copper
`species. In contrast to copper zeolites, several active sites for iron zeolites
`have been postulated: small FexOy clusters like Fe4O4 or oxygen bridged binuc-
`lear iron species like [HO-Fe-O-Fe-OH]2þ
`(33, 93, 97–102), isolated Fe2þ
`and
`Fe3þ
`ions (97–98, 103–105), and extra-framework Fe-O-Al species (106).
`However, Krishna et al. (105) were able to rule out the extensive formation
`of Fe-O-Al species even under extreme conditions (7008C and high concen-
`tration of gas-phase HCl). Therefore, for Fe-zeolites as well, the question can
`be reduced to whether isolated metal ions or ion pairs catalyze SCR chemistry
`and especially its key step, NO oxidation to NO2.
`The group of Gru¨ nert (98, 103) assumed that for Fe-zeolites, isolated Fe
`species as well as bridged Fe species contribute to NH3-SCR. An analysis of
`the available data from the literature supports this assumption. Table 1
`shows the turnover frequency (TOF) values of different Fe-ZSM-5 zeolites,
`and the calculated probabilities (P) for the formation of binuclear oxygen
`bridged Fe-O-Fe species. TOF was defined as the number of NO molecules con-
`verted per Fe per second (based on total Fe content) and represents the SCR
`activity of a single Fe center. P reflects the probability of finding at least one
`Al atom within a sphere of radius r centered around an arbitrary Al center of
`the zeolite, whereas both Al centers are occupied with an Fe specie. To calculate
`P, the average number of Al and Si-T sites within a sphere of radius r was deter-
`mined by means of the method of Feng and Hall (107), which was slightly
`modified according to Rice et al. (108). Feng and Hall (107) estimated a
`
`Table 1: TOF and calculated probability P for the formation of dinuc