Progress on Selective Catalytic Ammonia Oxidation (NH3‐SCO) over Cu−Containing Zeolite‐Based Catalysts

The selective catalytic oxidation of ammonia into nitrogen and water vapor (NH3‐SCO) has proven to be the most efficient technique for the removal of NH3 from the oxygen‐containing diesel exhausts. Until now, two reviews over investigated catalysts have been published in 2015–2016 (Jabłońska et al., RCS. Adv. 2015; Appl. Catal. B: Environ. 2016). Among various NH3‐SCO catalysts, Cu‐ and/or Fe‐containing zeolites are considered as promising candidates due to their benign nature, NH3 storage capacity, very high NH3 conversion and N2 selectivity. The present review focusses only on the Cu‐containing zeolites (mainly on chabazite (CHA): SSZ‐13, SAPO‐34) catalysts) applied in NH3‐SCO starting from 2015. The synthesis/preparation methods and the state of Cu in catalysts on the activity and selectivity are highlighted. The advantages of the understanding of the reaction mechanism are reviewed. Finally, an outlook is given together with expected inspirations for future research directions in developing zeolite‐based catalysts.


Introduction
Ammonia (NH 3 ) is dangerous to health, life and causes serious environmental problems. [1,2] In diesel engine aftertreatment systems, NH 3 served as NO x reducer and is generated on-board from aqueous urea solution. To achieve the desired NO x conversion, a stoichiometric or even an excess quantity of NH 3 is required, which results in an NH 3 slip. An NH 3 oxidation catalyst (ASC, guard catalysts, AMOX) is usually employed to selectively oxidize the unreacted NH 3 into N 2 and H 2 O as well as provide flexibility for controlling the dose of urea introduced before the NH 3 -SCR (the selective catalytic reduction of NO x by NH 3 ) catalyst. Ammonia slip catalyst is expected to balance high activity, N 2 selectivity and stability (hydrothermal stability even under harsh treatment conditions at 1073 K) across a broad range of operating conditions and in the presence of the typical components of exhausts, i. e., CO x , SO x , H 2 O. [3,4] The Cu-based catalysts are one of the most efficient systems in NH 3 -SCO and have attracted more and more attention. [5,6] Gang et al. [7] reported full NH 3 -conversion temperature at 623 K and 90 % N 2 selectivity over Cu/Al 2 O 3 . 10 wt.-% of copper was considered to be the most appropriate loading and this applied also for other types of catalysts (e. g., supported oxides, hydrotalcite-derived mixed metal oxides, etc.). [5] Both redox and acid properties are critical factors that affect catalyst activity, N 2 selectivity and stability in NH 3 -SCO. Thus, mainly alumina and even more frequently zeolites are utilized as supports. The copper-containing zeolites showed higher activity also in the presence of H 2 O, compared to alumina supported copper oxides with the same metal loading, [7,8] due to the high dispersion of metal species together with acid sites of high strength. Nevertheless, concerning the discussed material requirements, continuous studies and development of new types of the oxidation catalysts are still required. Investigation in reaction mechanisms themselves is still highly appealed. Many researchers discussed mainly NH 3 -SCO (as side process) in the research papers focusing mainly on the NH 3 -SCR process itself. Thus, finding literature focusing only on zeolite-based catalysts dedicated to NH 3 -SCO may cause difficulties. Table 1 summarizes the currently developed zeolite-based catalysts with various topology structures. The applied catalysts, their synthesis/preparation procedure and set-ups were generally different from each other, thus a rough comparison is feasible. The catalytic activity of the samples can be also compared based on T 50 and/or T 100 (temperature needed to achieve 50 and 100 % conversion, respectively).

Microporous and hierarchical micro-/mesoporous catalysts
The studies over zeolite-based catalysts focus mainly on ZSM-5 (MFI topology), Beta (BEA), Y (FAU) and SSZ-13 and SAPO-34 (CHA). In 2010, Kwak et al. [9] revealed the difference in activity and N 2 selectivity of the investigated copper-containing ZSM-5, Beta and SSZ-13, and assigned such results to different pore sizes of zeolites and location of the copper ions. The smallmedium size pores were preferred for high activity and N 2 selectivity: SSZ-13 having the smallest pores (~0.4 nm, 8membered ring) and ZSM-5 with medium pore opening (0 .55 nm, 10-membered ring). The Cu-beta catalyst with the largest pores (~0.7 nm and~0.55 nm, 12-membered ring) produced significant amounts of by-products (NO x = NO, NO 2 , and N 2 O), with a maximum of about 55 ppm at 623 K. Another factor, i. e., the content of copper, species and aggregation state of Cu-containing zeolites were not discussed in detail hampering property-activity relationships. Furthermore, Leistner et al. [10] confirmed that catalytic properties not only depend on the exchange level but also the structure of zeolite/silicoaluminophosphate (among Beta, SSZ-13, and SAPO-34). The (2.5 wt.-%) Cu-Beta catalyst in the form of monolith revealed higher activity than other materials above 673 K as a result of its significantly higher n(Cu)/n(Al) ratio of 56 (compared to 0.16-0.20 for CuÀ CHA). Again, the selectivity to N 2 O over the Cu-Beta catalyst was high and reach 20 ppm (during NH 3 -SCR). Surprisingly, Cucontaining Y zeolite was less intensively investigated in NH 3 -SCO, besides the studies of Gang et al. [7,26] in 1999-2000 over (3.7 or 8.7 wt.-%)Cu/Y. Kwak et al. [27] revealed a significant amount of N 2 O produced over (7.2 wt.-%)CuÀ Y (n(Si)/n(Al) = 2.65) accounting for its poor catalytic properties above 573 K in NH 3 -SCR. After hydrothermal treatment in 10 Vol.-% H 2 O in the air at 1073 K for 16 h, CuÀ Y lost completely its activity. In recently published work, Wang et al. [13] showed that T 50 [28] reported that silver-containing USY (ultrastable Y) zeolite with n(Si)/n(Al) = 2.5 as highly active, N 2 selective and stable catalysts in NH 3 -SCO. The materials revealed also high resistance in the presence of H 2 O. Thus, more current research should focus on high aluminacontaining zeolite USY to develop property-activity relationship over Cu-USY catalysts. The resistance in the presence of H 2 O should be always approved because water emits from internal combustion engines and the exhaust temperature can reach 973 K during diesel particle filter regeneration.
Although Cu-ZSM-5 and Fe-ZSM-5 zeolites have been most extensively studied because of their superior NH 3 -SCO activity and relatively high resistance to H 2 O, [5,6,29] only a limited number of studies focus on hierarchical micro-/mesoporous catalysts. Modification of the zeolite porous structure strongly influences the form (coordination and aggregation) on transition metals introduced by the ion-exchange method. E.g., Góra-Marek et al. [30] reported superior catalytic activity and N 2 selectivity (below 723 K) of (0.68-0.92 wt.%)Fe-containing mesostructured ZSM-5 (n(Si)/n(Al) = 32). The mesoporosity in ZSM-5 was generated during post-synthetic treatment (with NaOH/TBAOHtetrabutylammonium hydroxide) and direct synthesis route with amphiphilic organosilanes. The uniform micro-/mesoporous structure prevented the clustering of iron oxide species and preserved Fe 3 + species isolated inside the zeolite channels leading to enhanced catalytic activity and N 2 selectivity. Afterward, similar studies were conducted over Cu-containing ZSM-5. [11] The studies cover an influence of the mesoporosity generation in commercial ZSM-5 with NaOH and a mixture of NaOH/TPAOH (tetraproylamminum hydroxide) with different ratio (TPA + /OHÀ = 0.2, 0.4, 0.6, 0.8 and 1) and for different duration time (1, 2, 4 and 6 h). Again, the catalytic tests revealed improved activity and N 2 selectivity over materials with the generated mesoporosity (improved hierarchy factor (HF) of the samples). The effect was enhanced especially under condition treatment: concentration of TPA + /OHÀ = 0.2 M and duration treatment of 2 h. Increased acidity of the micro-/mesoporous samples, as well as the content of easily reducible copper species resulted in a significant improvement of catalytic properties of (2.5-4.1 wt.%)Cu-ZSM-5 not only in NH 3 -SCO but also in NH 3 -SCR. Only a slight difference in activity and N 2 selectivity was observed between the materials treated with NaOH or NaOH/TPAOH. Besides mesostructured zeolites, superior activity and N 2 selectivity were reported earlier over mesoporous CuFe 3 O 4 , [31] CuO/RuO 2 [32] or RuO 2 -CuO/Al-ZrO 2 , [33] [34] and facilitates the transport of the reactants to and from the active surface sites and thus enhances ammonia adsorption and oxidation. Thus, discussed results suggest that for the ZSM-5 system pore diffusion limitations should be considered for optimum practical applications.
Ideally, the components of one catalyst simultaneously perform superior NH 3 oxidation and SCR activity. Otherwise, the Cu-containing zeolites are also applied in the bi-functional NH 3 -SCO catalysts in the form of a mixed (washcoat) catalyst or as a dual-layer. A limited number of studies refer to catalysts containing both noble and transition metal as a one system, e. g., PtÀ Cu/ZSM-5. [35] Mainly such systems include a combination of an ammonia oxidation catalysts (usually Pt/Al 2 O 3 , bottom layer) and a SCR (Cu-and/or Fe-containing, top layer) catalyst.  E.g., Shrestha et al. [36] investigated the selective oxidation of NH 3 to N 2 on mixed and dual-layer Pt/Al 2 O 3 and Cu-SSZ-13 washcoated monolith catalysts under several reaction conditions to probe the effects of the added mass transfer barrier, in terms of the SCR loading (Cu-SSZ-13 layer thickness) on the activity and selectivity of the final catalyst. Still, N 2 O selectivity was rather high and thus, unacceptable for commercial applications. Thus, a full understanding of the system is required through the optimization of the catalyst architecture [36,37] and the development of the kinetic model. [37][38][39] E.g., Dhillon et al. [37] applied a hybrid layer design with a bottom mixed layer of Pt/Al 2 O 3 and Cu-SSZ-13, and a top Cu-SSZ-13 layer to obtain high activity and N 2 selectivity in NH 3 -SCO. Still, N 2 yield below 80 % was found in the range of 623-773 K.
Recently, Ghosh et al. [40] proposed a Pt/Al 2 O 3 @Cu-ZSM-5 core-shell catalyst. Evaluation of such catalyst shows excellent NH 3 oxidation activity (full NH 3 conversion at ca. 573 K) and 100 % N 2 selectivity (up to 548 K). Above 548 K, the combined yield of NO x and N 2 O was less than 10 %. The core-shell catalyst showed improved activity and N 2 selectivity than a physical mixture of Pt/Al 2 O 3 and Cu-ZSM-5. The thicker Cu-ZSM-5 shell (1.2 microns versus 0.5 microns, Figure 1) provides a longer diffusion length enabling additional Cu sites for SCR to occur, resulting in a higher N 2 selectivity. Interestingly, the material was also stable even after hydrothermal aging at 823 K. Certainly, further research in such direction could produce stable NH 3 slip catalysts applicable in industry.

Chabazite-based catalysts
Cu-containing medium and large pore zeolites, as Cu-ZSM-5 or Cu-Beta were described as efficient catalysts for NH 3 -SCR. However, those zeolites do not meet the most demanding hydrothermal stability targets when treated in the presence of steam at high temperatures as a result of dealumination causing the skeleton collapse and agglomeration of active sites. [41,42] The application of CuÀ CHA (Cu-SSZ-13 and Cu-SAPO-34) represents the great breakthrough in the field of NH 3 -SCR in the past decade due to its excellent catalytic activity, N 2 selectivity, and high hydrothermal stability at 1023 K. Both of the SSZ-13 and SAPO-34 zeolites have the same chabazite (CHA) framework structure. The CHA framework is composed of 4-, 6-, and 8-membered rings arranged to form a tridimensional system of channels perpendicular to each other (0.38 × 0.38 nm; R3 m (#166) space group). [43] The majority of the researchers working in the field agree that the unique properties of CuÀ CHA are due to its small-pore structure of the CHA (0.38 nm) that prevent the Al(OH) 3 (0.503 nm) from exciting the pores of the framework (i. e., eliminate the dealumination process). The detached Al(OH) 3 cannot exist in the pores of the framework and Al may even reattach back to the framework during the hydrothermal aging, arising from its relatively large kinetic diameter. [12,22,44,45] Otherwise, Fickel and Lobo [46] claimed that hydrothermal stability of CuÀ CHA arises due to the preferential coordination of cationic copper species to the double 6-ring (D6R) cages located in the CHA cavities. In recent years, some CuÀ CHA synthesis and preparation methods have been extensively developed, including aqueous ion-exchange method, solid-state ion-exchange method, onepot synthesis method, etc. SSZ-13 and SAPO-34 (preferentially in the NH 4 -form) are copper-exchanged with an aqueous solution of copper salt. Although a lot of studies focused on the activity-property relationship in NH 3 -SCR of CuÀ CHA zeolites prepared by ion-exchange procedure, less attention was paid for the process of NH 4 + replacement by copper ion during the Cu-exchanged process. E.g., Xu et al. [47] investigated adsorbed NH 3 (through gaseous NH 3 adsorption or traditional liquid exchange method) on Brönsted acid sites (Si-OH-Al) and reported three different forms: monodentate, bidentate and tridentate ammonium species (Figure 2). Cu + and Cu 2 + respectively replaced one or two Brönsted protons, lowering their amount. During the liquid Cu exchange process, polydentate ammonium species, occupying three neighboring Si-O-Al sites, would prompt more Cu species (Cu 2 + and/or Cu(OH) + ) to exchange onto Si-O-Al sites in a fixed region. These excessive Cu species would aggregate into (CuO) x species after calcination. Therefore, the different NH 4 + species and their contents in NH 4 + -CHA affected the Cu loadings and distributions in the final catalysts.
Gao et al. [12,48,49] employed Cu-SSZ-13 catalysts with varying Cu loadings and n(Si)/n(Al) ratios to examine their catalytic properties using NH 3 and NO oxidation and standard NH 3 -SCR. At higher Cu loadings (ion-exchange, IE) level of 90 % versus Reprinted from [40] with the permission of American Chemical Society. Reprinted from [47] with permission of Elsevier Science Publisher B.V.

ChemCatChem
Minireviews doi.org/10.1002/cctc.202000649 23 %), NH 3 conversion was shown to increase due to: i) weaker interactions between Cu ions and the zeolite framework (more facile Cu 2 + $ Cu + redox-cycling), ii) more accessible to reactants catalytic centers (located closer to pore openings). Interestingly, NH 3 oxidation at low temperature for Cu-SSZ-13 followed by a minimum in conversion at ca. 673 K and thereafter, an increase at high temperature. A similar profile was reported later by Leistner et al.. [10] A kinetic model for the NH 3 oxidation over these catalysts [50] revealed that when temperature was raised the coverage of NH 3 on Cu in the 6-membered ring significantly decreased in the model, resulting in the ammonia oxidation approaching a minimum. Afterward, NH 3 oxidized was assumed to occur on Cu sites in the large chabazite cages and/or Cu x O y species. A similar phenomenon was observed in NH 3 -SCR, and it is known as a seagull profile. Further discussion is provided in Chapter 4.
Zhang et al. [15] investigated a one-pot synthesized Cu-SSZ-13 catalyst dedicated as catalyst to NH 3 -SCO. No apparent dealumination occurred during the process of post-treatment with dilute HNO 3 solution. Otherwise, the authors pointed out its positive effect: i) eliminating excess Cu species in the initial Cu-SSZ-13; ii) boosting the quantity of Cu 2 + ions; iii) optimizing the special distribution of Cu 2 + ions in SSZ-13 structure on enhanced NH 3 oxidation. Furthermore, recent studies of Wang et al. [12] revealed comparable activity and selectivity over Cu-SSZ-13, Cu-SSZ-16 and Cu-SSZ-17. Guo et al. [18] compared CuO x / CHA (SSZ-13, SAPO-34) prepared by the impregnation method. CuO x /SSZ-13 exhibited superior NH 3 activity and N 2 selectivity compared with CuO x /SAPO-34 below 623 K under both dry and wet conditions (3.3 Vol.-% H 2 O). Hydrothermal aging (973 K, 10 h, 10 Vol.-% H 2 O) lead to significant loss of activity for CuO x / SAPO-34, while CuO x /SSZ-13 revealed high stability. The higher activity and N 2 selectivity of CuO x /SSZ-13 were assigned to its more plentiful acid sites and enhanced redox properties than for CuO x /SAPO-34.
One aspect of Cu-containing SAPO-34 is the difficulty in producing it using conventional aqueous ion-exchange. [20,51] Thus, some authors modify the ion-exchange procedure to obtain higher Cu loadings of Cu-SAPO-34 (e. g., 1.13-2.60 wt.-%). [23] Also, several studies have been published on Cu-SAPO-34 prepared via other methods, including precipitation, one-pot synthesis or solid-state ion-exchange versus ion-exchange. E.g., Wang et al. [19] examined Cu-SAPO-34 prepared by ion-exchange and precipitation method (precipitation of CuC 2 O 4 on H/SAPO-34 powder with Cu(CH 3 COO) 2 as the precursor and H 2 C 2 O 4 as the precipitator). The precipitated sample shows a stronger NH 3 oxidation ability, but a lower N 2 selectivity over CuO x clusters dispersed on the external surface of the zeolite. Yu et al. [21] by applying H 2 -TPR, EPR and CO-DRIFT analyses and the kinetic results further confirmed that CuO is the active site for NH 3 oxidation in impregnated Cu/SAPO-34. Also, NH 3 oxidation activity extensively increased with increasing Cu loading from 2 to 3 wt.-% in Cu-SAPO-34 (prepared by solid-state ion-exchange method), which correlates to the evolution in the amount of Cu x O y clusters (dimeric or oligomeric Cu species, for Cu-SAPO-34 with > 3 wt.-%). Cu species existed as isolated Cu 2 + cations inside the SAPO-34 pores in the ion-exchange samples are active in NH 3 -SCR. [22] Recently, Han et al. [24] demonstrated that CuO can be homogenously confined in small zeolitic materials SAPO-34 by the Trojan Horse approach. Copper ions were introduced into SAPO-34 by the one-pot process first and the sequential diffusion of Cu 2 + ions from the outer surface in SAPO-34 (after the subsequent impregnation). Superoxo species (O 2 À , based on Raman DRIFT and XAFS analyses; Figure 3) produced on CuO nanoclusters were reported to cause enhanced NH 3 oxidation and high N 2 selectivity below 523 K. This new approach for the combination of two preparation procedures should be considered in the future to obtain highly active and stable NH 3 -SCO catalysts. Considering the different Cu species may be present in the CuÀ CHA samples due to the different preparation methods, it seems that the location and the status of Cu determine catalyst activity and N 2 selectivity. Still, the studies of NH 3 -SCO over CuÀ CHA provides a basic overview of the catalyst preparation and their influence on the catalytic properties. There is a lack of the in deeps investigation for property-activity relationships. Therefore, it is necessary to comprehend the influence not only for H 2 O (including hydrothermal treatment) but also the effect of CO x and SO x on the NH 3 -SCO catalytic activity and its poisoning mechanism. As reported before, NH 3 -SCO catalysts deactivate after contact with SO 2 . [5,52,53]

Identification of copper species
The monomeric cations: isolated Cu + , Cu 2 + and/or aggregated species: Cu x O y may exist in Cu-SSZ-13, while for Cu-SAPO-34 this copper species co-exist regardless of the Cu loading. [54,55] The preparation methods mainly affect the distribution of Cu species, e. g., the CuO cluster can migrate into the framework of CuÀ CHA through hydrothermal treatment. [55,56] An introduction of copper ions in the desired form can be facilitated by the modification of the zeolite porous structure, e. g., via the generation of mesoporosity. A high dispersion of copper, i. e., in the form of uniform species and without blocking of narrow pores guarantees high activity and N 2 selectivity of the obtained catalysts in NH 3 -SCO. A clear correlation exists between the higher reducibility of Cu species and the enhanced activity [e. g., 57,58] , demonstrating that NH 3 oxidation over Cu-containing catalysts follows a redox mechanism. Yu et al. [21] investigated the Cu-containing SAPO-34 prepared by ion-exchange and impregnation. The copper-exchanged SAPO-34 contained more Figure 3. Formation mechanism of Cu(II) superoxo species in CuO@SAPO-34 prepared by the Trojan Horse approach. Reprinted from [24] with the permission of American Chemical Society.

ChemCatChem
Minireviews doi.org/10.1002/cctc.202000649 isolated Cu 2 + species, while the CuO species was dominant in impregnated one. Furthermore, H 2 -TPR, EPR, and CO-DRIFT analyses and the kinetic results revealed that CuO is the active site for NH 3 oxidation over Cu/SAPO-34. General, CuO x served as an active phase for the ammonia oxidation. [5] While the NH 3 -SCO proceed in the steps, according to the internal selective catalytic reduction (in situ SCR (i-SCR)) mechanism: i) NH 3 is oxidized to NO over CuO sites, and ii) the produced NO x is reduced by unreacted NH 3 over Cu 2 + sites in NH 3 -SCR (see Chapter 5). Recently, Han et al. [24] reported superoxo species (O 2 À ; Figure 3) as key active sites produced on CuO nanoclusters, which lead to enhanced NH 3 oxidation and high N 2 selectivity below 523 K. Certainly, more effort is required for identification of the variety of oxygen species, which can exist on the surface of catalysts. At this stage, none of the studies precisely identify the nature of oxygen species responsible for the NH 3 oxidation. The basis for the controversial point of view is the nature of species and the difficulties of monitoring the active sites under reaction conditions.
In Cu-exchanged SSZ-13, in addition to necked Cu 2 + (i. e., ions coordinate only with lattice oxygen but no other extraframework ligands), [Cu II (OH)] + species were reported as an active centers in O 2 activated catalyst. [59] For NH 3 -SCO over Cu-SSZ-13, reaction kinetic studies revealed that Cu-ion dimers were the actual active centers at low temperatures. Above 623 K, these Cu-dimers are unstable and split into Cu-ion monomers ([Cu I (NH 3 ) 2 ] + , [Cu(OH)(NH 3 ) x ] + ), which become the active centers. [46] Such transportation of Cu ion centers is responsible for a peculiar dual-maxima (seagull) NH 3 conversion profile. An atypical phenomenon recognized for Cu-SSZ-13 during NH 3 -SCO and NH 3 -SCR is the decrease followed by the increase of the NH 3 or NO conversion, generally between 523-623 K for intermediate Cu loadings (0.5-3 wt.%). Otherwise to NH 3 -SCO, for the seagull profile of NO conversion in NH 3 -SCR, Gao et al. [60] thought that the low-temperature NH 3 -SCR proceeds on this intermediate (formed from two isolated [Cu I (NH 3 ) 2 ] + ions), whereas isolated Cu 2 + ions acted mainly as individual active sites at high temperatures. Paolucci et al. [61] revealed that transient ion pairs [(NH 3 ) 2 Cu II ]-O 2 -[(NH 3 ) 2 Cu II ] (formed from mobilized Cu ions) acted as the active centers by participating in an O 2 -mediated Cu + $ Cu 2 + redox step. Recently, Fahami et al. [62] applied operando X-ray absorption, X-ray emission and in situ electron paramagnetic resonance spectroscopy measurements, including novel photon-in/photon-out techniques and reported that the activity decreased around 623 K that gives rise to the seagull shaped NO conversion profile could be caused by the formation of a more localized structure of mono (μ-oxo)dicopper [CuÀ OÀ Cu] 2 + complexes ( Figure 4). Above this temperature, which corresponds to partial NH 3 desorption from Cu sites, the isolated Cu sites migrate to form additional dimeric entities thus recovering the NH 3 -SCR activity. As NH 3 -SCO proceeds according to the i-SCR mechanism with NH 3 -SCR as the second step of NH 3 oxidation, a similar phenomenon can be considered for a description of the acidity of Cu-SSZ-13.

Reaction mechanisms
Three major NH 3 -SCO mechanisms were proposed in the literature, including (i) the imide mechanism, (ii) the hydrazine mechanism, and (iii) the internal selective catalytic reduction (i-SCR) mechanism. Details about these mechanisms can be found in two previous reviews. [5,6] Briefly, the majority of the researchers working in the field agree that NH 3 -SCO followed the i-SCR mechanism, consisting of two main steps. In the first step, part of ammonia is oxidized to NO x , while in the second step NO x is reduced by unreacted ammonia to N 2 , N 2 O, and H 2 O. The i-SCR mechanism was proposed for a large number of catalysts, such as Pt [e. g. 63 3 is activated by abstraction of hydrogen atoms to form -NH 2 , -NH or -N intermediates, which depending on the availability of oxygen can be converted with the formation of N 2 , NO x and N 2 O. E.g., Guo et al. [18] investigated that -NH 2 and -NH were observed on the spectra of CuO x /SSZ-13. Only -NH 2 species were obtained on the spectra of CuO x /SAPO-34, indicating that its redox property was not strong enough to Figure 4. Transformations of Cu species in Cu-SSZ-13 under different SCR-related conditions. Reprinted from [62] with the permission of Royal Society of Chemistry. further dehydrogenate -NH 2 into -NH species (in agreement with H 2 -TPR results). The highly dispersed and easily reducible transition/noble metal species play a role in the main adsorbed sites of NH 3 . Furthermore, a high Brönsted acidity strength in the zeolites favors ammonia adsorption as NH 4 + (an additional reservoir of chemisorbed NH 3 , especially in the high-temperature range). High N 2 selectivity was correlated to the protection of NH 4 + against oxidation, and thus their availability for reduction of NO x (NH 3 -SCR). [28,68] Therefore, for a complete understanding of the reaction mechanism of NH 3 -SCO the influence of both active sites -transition/noble metal and Lewis/Brönsted acid sites -needs to be determined and investigated separately.
Early studies over Fe-modified zeolites revealed that the higher N 2 selectivity for NH 3 -SCO correlates with the higher activity for NH 3 -SCR. [69,70] For Cu-ZSM-5 -including micro-/ mesoporous Cu-ZSM-5, [11] the comparison of the results of NH 3 -SCR and NH 3 -SCO, i. e., the NO and NH 3 conversion revealed NH 3 mining step of NH 3 -SCO. Thus, NH 3 -SCO produces NO x (the main by-products, [Eq. (1-2)]), which are reduced in the next step of NH 3 -SCR with formation of N 2 [Eq. ] or N 2 O [Eq. (4)]: NH 3 -SCO with different space velocities (15,400,30,800 and 61,600 h À 1 ) was carried out over PdÀ Y zeolites. [68] As the space velocity increased, the contact time was too short for the effective reduction of NO by ammonia, and thus, selectivity to NO increased. On the contrary, for tests carried out with a relatively low space velocity, the contact time of reactants with the catalyst surface was long enough for both NH 3 oxidation to NO x , and subsequent reduction of NO x by unreacted ammonia into N 2 and/or N 2 O. Furthermore, an FT-IR study of ammonia adsorbed on PdÀ Y zeolites confirmed the presence of NO, together with small amounts of NO 2 . Thus, NO x (NO and NO 2 ) species appeared as an intermediate in NH 3 -SCO leading to the formation of N 2 and/or N 2 O. [68] In situ-produced NO x À species (NO 3 À and/or NO 2 À ) on the surface of AgÀ Y and Ag-USY zeolites were believed to interact with adsorbed ammonia (NH x species) with the formation of N 2 and/or N 2 O, according to the i-SCR mechanism. [28] However, an appearance of hydrazine and nitroxyl species on the ammonia pre-adsorbed PdÀ Y zeolites at 523 K suggested that the mechanism of NH 3 -SCO could follow different routes, possibly depending on the reaction temperature. [65] For example, Zhang and He [71] applied in situ DRIFT spectroscopy of NH 3 adsorption and oxidation, and claimed that NH 3 -SCO over Ag/Al 2 O 3 proceeds according to an imide mechanism with -NH and -HNO as the key intermediates below 413 K. Above 413 K reaction proceeds according to the i-SCR mechanism, as follows [Eq. (5)(6)(7)(8) On the other hand, Chen et al. [72,73] claimed that -HNO appeared as an intermediate in the i-SCR mechanism of NH 3 -SCO over Ru/Ce 0.6 Zr 0.4 O 2 (PVP) or IrO 2 /Ce 0.6 Zr 0.4 O 2 (PVP). The formed < C-HNO interacted with atomic oxygen with the formation of NO, which furthermore reacts with < C-NH x (< C-NH 2 and -NH) species with the formation of N 2 and N 2 O (minor by-product). Sun et al. [35] proposed also based on in situ DRIFT investigations over PtÀ Cu/ZSM-5 that first NH 3 is oxidized to NO and then NO is reduced over the Cu species by the stored ammonia in ZSM-5 to form N 2 according to elementary steps as follows [Eq. (9-14)]: Thus, which steps best describes the NH 3 -SCO mechanism is still an open question and so far these proposals should not be judged as conclusive. In situ DRIFT spectroscopy of NH 3 adsorption and desorption/oxidation was mainly applied to determine the composition of the adsorbed intermediates at different temperatures, however, this information is insufficient for developing a detailed reaction mechanism of such complex reaction as NH 3 -SCO: a key reaction intermediate may very well be below the detection limit and that a detectable species may very well be a spectator. Also, the investigation of the reaction mechanism including information on catalytically active sites should be improved. E.g., recently, Wang et al. [13] reported based on the in situ DRIFT study, that NH 3 -SCO on the Cucontaining zeolites (Cu/ZSM-5, Cu/Beta, Cu/Y, Cu/MOR, Cu/FER; n(Si)/n(Al) = 10-15) take place via the i-SCR mechanism (presence of -NH 2 , nitrate, N 2 O 4 and NH 4 + species). The Si-OH-Al acid sites were assumed to help in the activation of NH 3 during NH 3 -SCO. NH 4 + ad-species reacted with O 2 to form -NH 2 and nitrate intermediates. The isolated Cu 2 + and Cu + species were identified as the main active sites over all catalysts. On the other hand, they showed that NH 3 adsorbed on Cu 2 + sites appeared only on particular catalysts (Cu/ZSM-5, Cu/Beta, Cu/ FER). Similarly, the [Cu 2 + (OH) À ] + species appeared only on Cu/ ZSM-5 and Cu/Y. Otherwise, the presence of the unique Cu (H 2 O) complex sites in the vicinity of the framework oxygen (based on EPR analysis) of Cu/MCM-49 evidenced that NH 3 -SCO followed the imide mechanism (presence of -HNO species). The role of CuO (present in all catalysts) was not specified. Otherwise, Guo et al. [17] and Yu et al. [21] reported that NH 3 is oxidized to NO x by surface CuO, and subsequently NO x are reduced to N 2 and H 2 O by unreacted NH 3 on isolated Cu 2 + sites of Cu/SSZ-13 and Cu/SAPO-34, respectively.
Furthermore, Sjövall et al. [74] using detailed kinetic modeling for NH 3 oxidation over Cu-ZSM-5 (n(Si)/n(Al) = 27, 2.03 wt.-% of Cu) assumed that NH 3 is oxidized directly to N 2 . They included four sites in the model: Cu-sites (including [Cu(NH 3 ) 4 ] 2 + , based on EPR analysis and DFT calculations), acid sites and weak adsorption sites. Mainly NH 3 adsorbed on Cu species in the low temperatures. When the temperature is raised weakly bonded NH 3 is replaced by O 2 , which dissociates on the surface. H 2 O and N 2 are produced during NH 3 oxidation (ca. at 523 K). On the active copper sites, the formed H 2 O reacts with oxygen to produce OH groups up to 673 K. Wang et al. [13] observed also the presence of an OH group on Cu-ZSM-5. Moreover, it was shown that the applied model was able to predict NH 3 storage, desorption and NH 3 oxidation, also in the presence of H 2 O. Therefore, computer simulation can be used as a useful tool for designing of Cu-containing zeolites. Furthermore, in addition to DRIFT analysis, determination of surface coverage of the main intermediates (in contrast to spectators) and their turnover frequency could play a turning role to elucidate the reaction mechanism of selective ammonia oxidation. Furthermore, to rationally develop NH 3 -SCO over transition/noble metal-containing zeolites, the reaction mechanism must be clarified not only in the presence of NH 3 and O 2 , but also under real conditions, including CO x , SO x , and H 2 O.
For Cu-SSZ-13 (n(Si)/n(Al) = 6 and varying n(Cu)/n(Al) = ca. 0.11 to 0.45), Gao et al. [14] found two kinetic regimes, separated at a common temperature of about 523 K. The apparent activation energies in the lower temperature regime (448-523 K) are 130 kJ mol À 1 and those in the higher-temperature regime (523-573 K) are 60 kJ mol À 1 ( Figure 5). This is due to a transformation of Cu ion centers in this temperature range discussed above. Thus, the reaction limiting step varies certainly as a function of temperature.

Summary and outlook
Cu-containing zeolites, mainly CuÀ CHA (Cu-SSZ-13, Cu-SSZ-16, Cu-SSZ-17, Cu-SAPO-34) are promising candidates for the development of active, N 2 selective a stable NH 3 -SCO catalysts for industrial applications. Mainly the catalytic tests are carried out in the presence of NH 3 , O 2 diluted in the inert gas (e. g., Ar, He or N 2 ). There is a lack of catalytic investigations conducted in Reprinted from [14] with permission of Elsevier Science Publisher B.V.