In this study, Fe-zeolites prepared by different techniques (CVD, SSIE, MR, LIE, ILIE, hydrothermal synthesis followed by steam activation and incipient wetness impregnation) using different framework compositions and structures and different pretreatment strategies have been used to probe their influence on the nature and distribution of Fe species. To this end, these samples have been characterized by EPR and UV/VIS-DRS-ex situ (after preparation, after calcination, after use in the selective catalytic reduction of NO) and in situ (during calcination and interaction of feed components with Fe sites of the catalysts at 293 K and at typical SCR temperature (623 K)). Iron redox kinetics of different isolated Fe³⁺ species and Fe³⁺ _x O _y clusters have been studied by UV/VIS-DRS to evaluate the impact of redox properties of different Fe species on the SCR of NO and N₂O as well as on the decomposition of N₂O. Acidity of selected Fe-zeolites was analyzed by FT-IR of pyridine adsorption. This information has been used to study the role of acidity on the SCR of NO either by NH₃ or by isobutane and SCR of N₂O with CO as well as N₂O decomposition. In situ FT-IR was also used to investigate the surface species formed during the adsorption of NO on oxidized and reduced surfaces at 293 K as well as during isobutane-SCR of NO at 623 K. It has been found that combined EPR and UV/VIS-DRS techniques are effective tools for elucidating the nature of coexisting Fe species formed in the Fe-containing catalysts investigated. The results have been discussed in relation to the catalytic performance. In this section, the main conclusions for structure-reactivity relationships will be summarized.

In the studied Fe-zeolites, at least three different isolated sites have been identified by EPR spectroscopy, which have been assigned to Fe³⁺ in tetrahedral (g' ≈ 4.3) and higher coordination (g' ≈ 6), and to Fe³⁺ in a highly symmetric environment (g' ≈ 2). The latter has been separately observed only in the samples A(MR)0.5, A(ILIE)0.2, A(ILIE)0.3, ex-Fe-silicalite, c-Fe-beta, (Fe-SBA-I)0.95. In the remaining samples (CVD, SSIE, ex-Fe-ZSM-5, ex-Fe-beta) an intense signal of aggregated iron oxide species is superimposed at the same g' value of 2. UV/VIS-DRS differentiates between isolated Fe³⁺ sites of different coordination, small (oligomeric) Fe oxide clusters and large Fe oxide aggregates. The degree of aggregation and the magnetic ordering of the aggregates were studied by temperature-dependent EPR measurements. Based on these results, the following conclusions could be derived:

All Fe-MFI preparations (including MR, ILIE and hydrothermal synthesis followed by steam activation) lead to the coexistence of different iron species. In all cases, isolated Fe sites are found. They are predominant after the MR, ILIE and hydrothermal synthesis (only in Al free silicalite) followed by steam activation while, in the other preparations, they are formed together with Fe _x O _y clusters of different size.

In the CVD preparation, clustered species occur already after the washing step, but their quantity and size increase strongly during calcination, with low heating rates resulting in higher Fe dispersions. Intense washing of the material after the CVD step also favors higher Fe dispersion but does not prevent clustering. With a matrix of low Al content and low density of Brønsted sites but high defect density, clusters are preferentially created in both uncalcined and calcined samples. However, the cluster size is somewhat restricted in the calcined sample as compared to samples with a matrix containing high Al content and low defect density, probably due to the internal defects providing additional aggregation nuclei. By the SSIE technique, clustered species are formed already during the preparation and their quantity and size increase even more after calcination. Remarkably, the MR preparation method leads to formation of mainly isolated Fe³⁺ species and iron association is largely suppressed even after calcination. Similarly, the ILIE method favours high Fe dispersion for low Fe content (≤ 0.3 wt.%) for which Fe is mainly in the form of mononuclear Fe sites. With increasing Fe content from 0.3 to 1.2 wt.% formation of isolated Fe sites level off at a certain Fe content, after which oligomeric iron-oxo species and poorly ordered iron oxide aggregates are formed. Contrarily, the conventional LIE method favours the formation of iron oxide clusters of different size, particularly the larger ones already at rather low Fe content (Table 4.1).

Samples with desired framework composition can be prepared using hydrothermal synthesis and framework cations can be extracted to extraframework positions by steam activation. Framework Fe³⁺ species either in the MFI or in the beta are well shielded in the zeolite framework and are not reducible upon reductive treatment with H₂ (Fig. 4.10). The presence of Al in the zeolite framework destabilizes framework iron and promotes extraction and clustering of extraframework iron to oxide clusters upon steaming or even after calcination to some extent. Accordingly, Al free ex-Fe-silicalite shows almost exclusively isolated Fe³⁺ ions even after steam activation and iron association is largely prevented. Al containing ex-Fe-ZSM-5 and ex-Fe-beta with relatively low Fe content show extensive degree of iron clustering in the form of oligomers and large iron oxide particles (Table 4.1).

The very high surface area of mesoporous silica supports such as SBA-15 favours the formation of highly dispersed, almost exclusively isolated Fe species upon incipient wetness preparation (Table 4.1).

In cluster free samples, extraframework isolated Fe⁺³ ions in tetrahedral and octahedral coordination possess different redox properties as evidenced by the time dependence of reduction and reoxidation at 673 K by UV/VIS-DRS (Table 4.2). The former species are slightly slower reduced and faster reoxidized than the latter. In contrast, iron oxide clusters are hardly reduced but fastly reoxidized in cluster containing samples. Interestingly, Fe³⁺ sites that remain isolated in samples dominated by Fe _x O _y clusters, change their redox behaviour and become highly reduction resistant in comparison to those in samples without oxide clusters. Additionally, isolated Fe sites reflected by EPR signals at g' ≈ 6, g' ≈ 4.3 and g' ≈ 2 are easily,moderately and hardly reduced respectively, whereas the reoxidation takes place the other way around, i.e., hardly, moderately and easily respectively as evidenced by in situ EPR studies.

Brønsted acidity of the samples decreases with increasing Si/Al ratio as evidenced by FT-IR studies of adsorbed pyridine over catalysts with different Si/Al ratio (Table 4.3). It was also found that iron oxide clusters provide additional Lewis acidity.

In situ FT-IR spectroscopy revealed that adsorption of NO on the pre-oxidized surface of Fe-ZSM-5 zeolites forms mainly adsorbed nitrate species presumably on Fe sites. However, the adsorbed nitrate species are also found over H-ZSM-5(A) support which might proceed over Fe sites that are present as impurities and/or extraframework Al sites. The formation of N₂O₄ species was observed. Therefore, besides oxidation of NO by Fe species, the disproportionation reaction of N₂O₄ to NO⁺ and NO₃ ^- was proposed as one of the reaction pathways for the formation of nitrates. This was further supported by the formation NO⁺ (2134 cm^-1) and [NO⁺][N₂O₄] (2198 cm^-1) species. Differently, adsorption of NO on pre-reduced surface of Fe-ZSM-5 zeolites lowers the formation of nitrate species by oxidation as evidenced by the lower intensity of the respective FT-IR bands. This leads to formation of nitrito species on the cluster containing samples. Thus, it can be concluded that Fe³⁺ sites are essential for the oxidation/activation of NO species.

Under typical NH₃ and isobutane-SCR conditions, the different isolated iron ions show a different sensitivity versus reduction/reoxidation. Accordingly, in cluster free samples (e.g., A(ILIE)0.3) isolated Fe³⁺ ions are partially reduced under steady-state SCR conditions, however, to different extents. Octahedral Fe³⁺ reflected by EPR signals around g' ≈ 6 and a UV/VIS band at 291 nm are most sensitive to reduction followed by tetrahedral Fe³⁺ (g' ≈ 4.3, 241 nm) while Fe³⁺ ions evidenced by the EPR signal at g' ≈ 2, for which the coordination geometry cannot be easily specified, are hardly reduced. With increasing Fe content, the amount of these hardly reducible isolated Fe³⁺ sites increases as evidenced by a comparison of the UV/VIS intensity below 300 nm under steady-state conditions. Regarding the relative intensities of the different EPR signals, it appears that hardly reducible sites comprise the majority of the isolated Fe species and are mainly reflected by the signal at g' ≈ 2. In contrast, easily reducible sites give rise to the line at g' ≈ 6 and partly also to the signal at g’ ≈ 4.3 and seem to be much less abundant. Taking account of the in situ FT-IR results that revealed a preferred reaction of NO with oxidized Fe³⁺ species, it is straightforward to conclude that, among the isolated Fe species, it might be just the hardly reducible Fe³⁺ ions which are essential for the oxidation of NO to form NO₂/NO₃ as intermediates for the reaction with the reductant. In contrast, Fe sites that are irreversibly reduced under SCR conditions do probably not belong to active sites. This is also supported by the fact that in used Fe-ZSM-5 catalysts the g' ≈ 6 signal disappeared completely, however, no deactivation was observed during the SCR reaction [55]. Since the amount of hardly reducible isolated Fe increases with rising Fe content, it is not surprising that the activity of the Fe-samples has been observed to increase, too (section 4.4).

However, attempts to correlate the rate of the SCR reaction with the number of Fe sites in the Fe-catalysts (Figs. 4.62 and 4.63) show the involvement of isolated Fe ions and oligonuclear Fe _x O _y clusters in both SCR reactions. Additionally, it was also found that in NH₃-SCR probably even Fe ions accessible on the surface of oxide particles also participate. In situ UV/VIS studies including those of the redox kinetics have shown that Fe _x O _y clusters are much faster reoxidized than isolated Fe sites and, thus, can immediately enter in another redox cycle. It is therefore plausible to assume that they contribute to the selective catalytic process at lower reaction temperature, too. At higher temperature these agglomerates, due to their higher oxidation potential in comparison to isolated Fe³⁺ species, give rise to unselective total oxidation of the reductant, thus, limiting the temperature window of selective NO reduction. This effect is much more pronounced for isobutane-SCR since Fe _x O _y clusters oxidize the isobutane to CO_x already at temperatures as low as 623 K, whereas in NH₃-SCR they are selective up to temperatures of about 700 K. This is also evident from carbonyl-containing species detected by in situ FT-IR preferentially on cluster containing A(ILIE)1.2 but not on cluster free A(ILIE)0.2. These species are regarded as intermediates in the total oxidation of isobutane. As a result of this unselective oxidation behaviour of the Fe _x O _y agglomerates, the NO conversion drops dramatically above 600 K in case of isobutane. Hence, the catalyst performing best in this reaction (A(ILIE)0.3) is almost void of clusters. With the NH₃ reductant, the unselective attack occurs at much higher temperature and to a much lower extent. Thus, the limitation of the selective temperature region is of little practical importance. Hence, the best catalyst for NH₃-SCR was the one with the highest number of accessible Fe sites (A(ILIE)1.2).

The acidity of the zeolite is essential for both NH₃ and isobutane-SCR as evidenced by the catalytic performance of cluster free ex-Fe-silicalite and A(ILIE)0.3 and cluster containing B(CVD,W1,C5) and A(ILIE)1.2 (Figs. 4.55 and 4.57). In situ FT-IR studies show that isobutane-SCR of NO over Fe-ZSM-5 mainly proceeds via nitriles, cyanates and/or isocyanates. Interestingly, the formation of nitriles, cyanates and/or isocyanates seem to proceed preferentially on isolated Fe sites, the amount of which increases from sample A(ILIE)0.2 to A(ILIE)1.2 (Fig. 4.45). Therefore, it can be regarded as another reason for the superior catalytic performance of isolated Fe sites in the isobutane-SCR. Finally, the microporous MFI structure of the Fe-catalysts is much more favorable for the SCR of NO than the mesoporous SBA-15 structure, since the confined pore structure of the former favors the intimate contact between active Fe sites and reactants (Figs. 4.55 and 4.57).

The following preparation strategies are necessary for tailoring Fe-catalysts for isobutane- and NH₃-SCR. For the former reaction the Fe-ZSM-5 with mainly isolated Fe sites is desirable and the amount of Fe sites should be increased without imposing the formation of clusters. The same strategy applies also for designing Fe-ZSM-5 catalyst for NH₃-SCR, however, in the course of increasing the amount of isolated Fe sites the formation of highly dispersed oligomeric Fe sites may be acceptable to some extent. The rational catalyst for both SCR reactions is the one with strong Brønsted and Lewis acidity.

By correlation of N₂O conversions in direct decomposition and SCR of N₂O with different iron species detected by UV/VIS-DRS it was found that in direct N₂O decomposition oligomers are preferred over isolated Fe³⁺ ions in view of the easier oxygen recombination (rate determining step in the process) of two iron centers that are close together. Hence, the best catalyst for the decomposition of N₂O was the one with high amount of oligomers, A'(CVD,W1,C2) (Fig. 4.58).

For SCR of N₂O by CO, isolated Fe³⁺ are the active iron centers. The importance of the mononuclear iron sites in the reduction of N₂O by CO was strongly evidenced by the high specific activity of A(ILIE)0.2 and ex-Fe-silicalite, with a remarkable uniform distribution of isolated iron ions. Furthermore, a correlation between the N₂O conversion and the fraction of isolated Fe³⁺ ions in the catalysts was found (Fig. 5.3). In situ UV/VIS-DRS and EPR studies further evidenced the participation of mononuclear iron ions in the SCR of N₂O with CO, however, they also support the involvement of oligomeric species. The interaction of N₂O and CO and the reaction mechanism is iron site dependent. Over tetrahedral isolated iron sites, which are coordinatively unsaturated and, thus, can extend their coordination sphere, the reduction of N₂O with CO occurs via coordinated CO species on Fe³⁺ ions, not involving change of oxidation state (Eqs. (5.7) and (5.8)). In contrast, octahedrally coordinated isolated Fe³⁺ sites convert N₂O in the presence of CO by involving a Fe³⁺/Fe²⁺ redox process (Eqs. (5.5) and (5.6)). The reaction over oligomers proceeds, too, via a redox Fe³⁺/Fe²⁺ process and involves the intermediate formation of O¯ radicals.

The acidity of the zeolite has probably no effect on both the decomposition and SCR of N₂O as evidenced by the catalytic performance of ex-Fe-silicalite and A(ILIE)0.2 (Figs. 4.58, 4.60 and 4.61).

In contrast, pore structure of the Fe-catalyst strongly influences the catalytic activity as shown by ex-Fe-silicalite and (Fe-SBA-I)0.95 (Figs. 4.59 and 4.61). Thus, it was found that microporous structure such as MFI is more favorable than mesoporous material.

Finally, for tailoring Fe-MFI catalysts for SCR and direct decomposition of N₂O, the same strategies can be followed as drawn for isobutane and NH₃-SCR respectively. Thus, Fe-MFI catalysts for SCR of N₂O should contain mainly extraframework isolated Fe sites, while for decomposition highly dispersed oligomeric moieties are required.