5. Discussion

5.1 Influence of synthesis conditions, framework composition and SCR reaction on the nature of Fe species

↓145

Fe-containing zeolites prepared by various techniques can be classified into two categories: (i) postsynthesis insertion of iron via ion exchange in liquid and solid state as well as by CVD and by impregnation and (ii) isomorphous substitution of Fe+3 into the framework of MFI and beta structure followed by extraction by steaming to extrafamework positions. The samples have been characterized by EPR and UV/VIS-DRS after synthesis, calcination, steaming and use in catalysis as well as in situ during calcination to assess the nature and distribution of iron species and their role in the catalytic reaction. In this section, the EPR and UV/VIS-DRS results of the influence of the mode of Fe insertion, washing and calcination procedure, Al content of the parent zeolite matrix and use in the SCR reaction on the iron species are discussed.

5.1.1 Mode of Fe incorporation

↓146

The results of UV/VIS-DRS and EPR characterization show that the distribution of iron species, as isolated iron ions, oligomeric iron oxo species and iron oxide particles, is a function of the sample genesis. According to EPR results it can be concluded that almost all samples contain at least two kinds of strongly distorted isolated Fe+3 ions, probably in tetrahedral and higher coordination as reflected by EPR signals at g' ≈ 4.3 and g' ≈ 6. A third kind of isolated Fe+3 ions in less distorted environment contributes to the EPR signal at g' ≈ 2. In UV/VIS-DRS rather the different coordination geometry (tetrahedral and octahedral) but not the distortion of the isolated Fe site is reflected by two signals below 300 nm. Beside the different isolated Fe+3 ions, Fe x O y aggregates are formed, the amount and size of which varied from sample to sample. In UV/VIS-DRS they are reflected by bands above 300 nm while in EPR spectra they give rise to a signal around g' ≈ 2.

Chemical vapor deposition (CVD) is one of the most effective postsynthesis methods, since it is easy to obtain high iron loading (Fe/Al ≈ 1). A complete insertion of iron at ion exchanged positions has been claimed, forming diferric (hydr)oxo-bridge binuclear clusters [31,56,57]. The presence of almost exclusively such Fe dimers in Fe-ZSM-5 zeolites prepared by CVD has been claimed by several authors using different spectroscopic techniques, mainly EXAFS [56,57,110,111]. However, recently, Heinrich et. al. [36] have shown that EXAFS seriously underestimates scattering contributions from higher shells. Thus, it is much less sensitive to FeOx clusters (dimers being the smallest possible cluster species) than to isolated Fe. Moreover, EXAFS gives average values for the scattering contributions from all Fe species making the identification of particular species in samples with several types of coexisting Fe species difficult. In an EPR study, Chen et al. assigned such dimer species to the signal at g' ≈ 2.03 in the spectra recorded at 293 K. It is unlikely that such dimer species gives an EPR signal at this temperature since, it requires much lower temperatures (e.g. 8 K) to be detected [55,186].

In the present study, the UV/VIS-DRS and EPR results in sections 4.1.1 and 4.1.2 respectively show the presence of iron in the sample as isolated, oligomeric iron oxo species and large iron oxide particles. Furthermore, the semi-quantitative data obtained by UV/VIS-DRS show the occurrence of these three kinds of Fe species in almost equal amounts (Table 4.1). Hence, a uniform distribution of iron species was certainly not achieved in the studied CVD samples, prepared according to the procedure described by Chen and Sachtler [31]. Therefore, the almost exclusive presence of Fe dimers in Fe-ZSM-5 as claimed by some authors is not appropriate. The hydrolysis and calcination steps were proven to be the key steps in order to control the formation of large iron oxide particles after sublimation of FeCl3 into zeolite matrix. The results will be discussed in detail in the corresponding sections.

↓147

Similar to CVD, solid-state ion exchange (SSIE) leads to the formation of both isolated iron ions and iron oxo clusters of different sizes. However, large iron oxide particles with long range magnetic ordering as evidenced by temperature dependent EPR results (Fig. 4.20) are markedly more pronounced than with CVD. These particles are at the external surface of the zeolite as evidenced by XRD and EXAFS [36].

UV/VIS-DRS and EPR results show that the mechanochemical route (MR) leads to the formation of highly dispersed isolated Fe+3 ions besides a small amount of oligomeric Fe x O y moieties. Semi-quantitative UV/VIS-DRS data show that almost 83% of the Fe is distributed as isolated iron sites and 17% of the Fe is present as oligomeric Fe x O y . Interestingly, XAFS data of sample A(MR)0.5 do not reflect any Fe x O y clusters although they are clearly detected by EPR and UV/VIS-DRS. This illustrates the benefit of the two techniques for the characterization of complex solid materials like Fe-ZSM-5.

The sample prepared by conventional liquid ion exchange (LIE) shows extensive formation of iron oxide clusters and particles as evidenced by UV/VIS-DRS and EPR results, which is also in line with HRTEM results [58] although the Fe content is rather low. On the other hand, improved liquid ion exchange (ILIE) in dilute HCl with iron powder was proven to be a more effective post synthesis method, since iron clustering into large particles is largely suppressed as compared to samples obtained by other techniques. UV/VIS-DRS and EPR data revealed that at low iron content in the zeolite matrix iron is mainly in the form of isolated Fe+3 ions and a uniform distribution of iron species was achieved in A(ILIE)0.2 and A(ILIE)0.3. At an iron content ≤ 0.3 wt-%, ca. 95 % of the iron is present as mononuclear sites. This conclusion rests mainly on the UV/VIS-DRS data (Fig. 4.5, Table 4.1). It is, however, confirmed by the EPR results (Fig. 4.23) where all signals observed in A(ILIE)0.2 and A(ILIE)0.3 could be assigned to isolated sites and exhibit paramagnetic behavior. Previous EXAFS results also support this conclusion qualitatively by a very low sum coordination number of the neighboring Fe shells [64]. On increasing iron percentage more and more oligomeric Fe x O y moieties are formed in A(ILIE)0.6 and A(ILIE)0.7. Increasing clustering tendency is also supported by EXAFS, where the Fe sum coordination number increases above 3 [64]. The coexistence of isolated Fe3+ ions and weakly interacting Fe sites within low oligomers is also suggested from the temperature dependence of the EPR spectra (Fig. 4.24). However, the missing ferrimagnetic/super-paramagnetic behaviour confirms that the formation of extended oxide particles does not take place at Fe contents of 0.6 %. At a Fe content of 1.2 %, the aggregation into larger particles is obvious as evidenced by UV/VIS-DRS and EPR results. The temperature dependence of EPR spectra of A(ILIE)1.2 is typical for ferrimagnetic/super-paramagnetic Fe2O3 nanoparticles (Fig. 4.24), the formation of which is further supported by TEM measurements [64].

↓148

Finally, the observed differences in the distribution of Fe species between samples prepared by LIE and ILIE can be explained as: In LIE, Fe3+ ions are exchanged with NH4 + ions (NH4-ZSM-5 with Si/Al ≈ 37). Thus, one Fe3+ ion compensates three spatially separated negative charges of the zeolite matrix which were balanced by NH4 + ions. Accordingly, only a small amount of Fe3+ could be deposited at cationic sites (ca. Fe/Al ≈ 0.33). The additional Fe3+ ions tend to precipitate and form Fe x O y clusters upon subsequent washing and calcination which leads to a highly heterogeneous distribution of Fe species in the final catalyst. Differently, in ILIE procedure Fe2+ ions replace two H+ (H-ZSM-5 with Si/Al ≈ 14). Obviously, this leads to the incorporation of more Fe ions at cationic sites in the zeolite matrix. Accordingly, the possibility for the formation of Fe x O y clusters is reduced and a more uniform distribution of Fe species was achieved, in particular at low Fe content. Furthermore, low Al content NH4-ZSM-5 support (Si/Al ≈ 37) used for LIE provides lower number of cation exchange sites than high Al content H-ZSM-5 (Si/Al ≈ 14) used for ILIE. This can be another reason for the formation of more pronounced clusters in LIE than in ILIE samples.

Steam activated ex-Fe-silicalite, ex-Fe-ZSM-5 and ex-Fe-beta are highly diluted samples, with Si/Fe ≈ 150. The migration of iron from framework to extraframework positions upon steaming is complex and can be envisaged as a clustering process in the latter two samples [155]. An appropriate comparison of the iron constitution in these samples can be established, since the preparation and activation procedures, as well as the iron content in the catalysts are similar. A better-defined distribution of iron species was attained upon steam treatment of Fe-silicalite. The color of this sample was nearly white, suggesting the more isolated nature of the iron species in the catalyst, while ex-Fe-ZSM-5 and ex-Fe-beta were light brownish. Indeed, the majority of iron in ex-Fe-silicalite was found as isolated Fe3+ ions in extraframework positions. A very minor degree of iron aggregation is present, as concluded from UV/VIS-DRS. EPR spectroscopy evidences the paramagnetic behaviour of Fe3+ species in ex-Fe-silicalite following the Curie-Weiss law, as typically observed for highly symmetric isolated species (g' ≈ 2). However, a certain degree of weak dipolar interactions between Fe3+ sites is also identified, indicative of certain iron association. The UV/VIS-DRS and EPR results of ex-Fe-ZSM-5 and ex-Fe-beta are similar, showing a variety of iron species including the formation of small iron oxide nanoparticles. Hence a remarkable uniformity of iron species and the absence of extensive clustering are a priori not guaranteed by this method. Of course, the size of the particles is significantly smaller than in the Fe-zeolites prepared by post-synthesis methods as evidenced by (HR)TEM [58,64], which indicates a higher iron dispersion. However, this is also associated to the lower iron content in the sample (e.g.,≈ 7 times lower than in the catalyst prepared by CVD and SSIE).

Additionally the above results indicate the importance of framework composition of the zeolite in determining the iron distribution. As shown in results (section 4.1.1 and page no. 62), the presence of Al in the framework structure favors the formation of oligonuclear Fe x O y clusters and even Fe2O3 particles upon steaming. By comparing the UV/VIS-DR spectra of c-Fe-beta and ex-Fe-beta (Fig. 4.9) it is evident that this undesired process takes partly place already during synthesis and/or calcination. However, it is largely suppressed in Al-free Fe-silicalite. This suggests that Al and Fe might compete for the same Si lattice positions, whereby the isomorphous substitution of Al is obviously favored. This agrees with previous observations [96,61] on Fe-MFI and also with recent results of Berlier et al. [187] who observed that the presence of Al in the framework of Fe-ZSM-5 promotes the formation of extraframework Fe upon vacuum treatment at higher temperature in comparison to Al-free Fe-silicalite. The ex-Fe-silicalite results presented here and in previous studies on this system [155], using steam as activation atmosphere (more effective for iron dislodgment than vacuum treatment), show that the higher stability of iron in the framework of silicalite enables a better control of iron extraction upon steam activation, since iron clustering can be largely prevented. Still, by a proper selection of the steaming temperature (873 K), a substantial degree of Fe extraction can be accomplished, although not being complete. This has been demonstrated by the different reduction characteristics of framework and extraframework isolated iron species (irrespective of the zeolite structure) during in situ UV/VIS-DRS in H2 at 773 K (Fig. 4.10), and previously by voltammetric response studies [155].

↓149

The Fe content as well as the BET surface area of ex-Fe-silicalite and (Fe-SBA-I)0.95 are rather similar. Significant differences exist in the pore structure. While ex-Fe-silicalite is dominated by micropores of 0.55 nm diameter resulting from the well-known MFI structurewhich is confirmed by the XRD powder pattern [161], (Fe-SBA-I)0.95 contains mesopores, the mean diameter of which amounts to 7.5 nm as derived from N2 adsorption measurements [161]. Taking into account the results obtained by UV/VIS-DRS and EPR and also by other characterization techniques obtained over these two samples [161], it can be concluded that both Fe-silicalite and (Fe-SBA-I)0.95 contain almost exclusively isolated Fe3+ sites of very similar structure. Moreover, very similar reduction/reoxidation behaviour of iron species in H2 and in air is observed, too, for both samples (Fig. 4.11). This suggests that Fe sites of almost the same structure and very similar redox properties can be created in matrices of very different pore structure by using suitable preparation techniques.

In summary, the characterization results indicate that improved liquid ion exchange and the mechanochemical route are the most effective postsynthesis techniques to introduce preferably isolated Fe+3 species into pore positions of ZSM-5. Chemical vapor deposition and solid-state ion exchange methods are also effective to create isolated Fe sites but, besides, cluster formation cannot be avoided. Differently, conventional liquid ion exchange creates preferably large iron oxide clusters/particles. On the other hand, hydrothermal synthesis followed by steam activation leads to extensive clustering of Fe in Al containing Fe-ZSM-5 and Fe-beta but remarkably, this method leads to formation of exclusively extraframework isolated Fe+3 species in Al free Fe-silicalite. Finally, mesoporous supports like SBA-15 are certainly favorable for the formation of highly dispersed Fe species.

5.1.2 Washing intensity

The washing procedure after sublimation of FeCl3 into the ZSM-5 matrix, certainly, has influence on the distribution of iron species. As evidenced by UV/VIS-DRS spectra of samples A(CVD,W1) and A(CVD,W10), intense washing diminishes the amount of large Fe x O y clusters slightly (Fig. 4.2 and Table 4.1). This has been supported earlier from TPR and Mössbauer measurements although these measurements had been performed with the respective A(CVD,W1,C0.5) and A(CVD,W10,C0.5) samples after use in the SCR reaction [36]. For comparison, these (used) samples were also studied by UV/VIS-DRS (Fig. 4.2b) and EPR spectroscopy (Fig. 4.14). Thermal stress (calcination or use in the SCR reaction, which is initiated by a thermal activation/stabilization process) favors aggregation of isolated Fe sites markedly. However, even after use in catalysis the better dispersion of the iron species in the intensely washed material can be traced by a decreased absorption above 400 nm (solid line) in UV/VIS-DRS (Fig. 4.2) and a less intense g' ≈ 2 signal above 373 K in EPR spectra (Fig. 4.14). The latter effect indicates a weaker antiferromagnetic coupling due to a smaller cluster size. This confirmed that intense washing diminishes the amount of large Fe x O y clusters slightly and leads to slightly higher dispersion as compared to the washing procedure using a 10-fold smaller amount of water.

5.1.3 Calcination procedure

↓150

Heat treatment, e.g. the calcination in air at 873 K, usually performed before catalytic experiments [32,36] causes significant structural changes. It is seen from the UV/VIS-DRS and EPR spectra that initially isolated Fe sites aggregate to form Fe x O y clusters (see also Table 4.1). This effect seems to be slightly favored by higher heating rates. It has been reported that mild calcination (0.5 K/min and 1000 ml/min O2/He (20% O2 + 80% He)) favours high Fe dispersion and formation of large iron oxide clusters is strongly suppressed, while severe calcination (5 K/min and 200 ml/min O2) leads to large Fe oxide clusters [57]. These calcination procedures are similar to those used for the samples in this thesis (e.g. samples A(CVD,W1,C0.5) and A(CVD,W1,C5). Different from the above authors observations, in both cases extensive iron oxide clusters including larger ones are formed as evidenced by EPR and UV/VIS but the clustering slightly lowered in A(CVD,W1,C0.5) (Figs. 4.3 and 4.15). Clustering of isolated Fe species upon calcination in air has been observed for almost all Fe-containing samples (Table 4.1). It appears that Fe x O y cluster formation is favored by moisture remaining in the zeolites after washing as evidenced by dehydration and re-hydration experiments (Fig. 4.16). After deposition of [FeCl2]+ at cationic sites of the zeolite matrix, the chloride ligands from the coordination sphere of Fe sites are replaced by OH or H2O during the washing procedure (Eq. (4.2)). The latter can be reversibly removed from the Fe coordination either by evacuation or by mild thermal treatment (< 423 K) as evidenced in Fig. 4.16a [55]. EPR signal at g' ≈ 6 which is assigned to isolated Fe sites with 5 or 6 coordinating ligands (Eq. (4.2)) lose intensity to g' ≈ 4.3 (due to isolated Fe sites with 3 or 4 ligands) upon room temperature evacuation (Fig. 4.16a). This is due to the removal of OH or H2O ligands from the coordination sphere of Fe sites that are responsible for the g' ≈ 6 signal. These effects are almost completely reversible upon contact with ambient atmosphere due to the adsorption of moisture. Calcination at 873 K initiates the condensation of these ligands (OH or H2O) between different Fe coordinations as well as between Fe and framework OH groups. The former leading to Fe association as evidenced by decrease in the amount of isolated Fe sites after calcination (Fig. 4.15 and Table 4.1) and the latter attaches the Fe site to the framework via oxygen bridge (Fig. 4.16b and c). Accordingly, it is also likely that iron oxy hydroxo clusters, which most probably form after the washing step, lose moisture (H2O/OH) upon calcination and form Fe x O y clusters as evidenced UV/VIS (Table 4.1).

5.1.4 Al content and defect density of the parent zeolite matrix

The effect of Si/Al ratio in the zeolite matrix on the distribution of iron species was studied using two different zeolite matrices with Si/Al ≈ 40 and 14. ZSM-5(A) with Si/Al ≈ 14 shows higher acidity than ZSM-5(B) with Si/Al ≈ 40 as evidenced by FT-IR of pyridine adsorption (Fig. 4.28 and Table 4.3). UV/VIS-DRS and EPR results of uncalcined samples B(CVD,W1) and A(CVD,W1) show that the ZSM-5(B) matrix favours the formation of iron oxide clusters (Fig. 4.3b and 4.17a). Surprisingly, the opposite is observed after calcination as evident from both UV/VIS-DRS and EPR results (Fig. 4.3c and 4.17b). This might be due to the presence of silanol nests in the defective matrix (ZSM-5(B)) which serve as additional nuclei for iron condensation and keep the iron inside the pores. In contrast, in well-structured ZSM-5(A), iron oxide species may migrate out of the pores toward the external crystal surface where they can grow further in size [57]. Hence, in calcined B(CVD,W1,C5) the formation of large iron oxide clusters is less pronounced compared to sample A(CVD,W1,C5).

5.1.5 Use in the SCR reaction

UV/VIS-DR spectra of A(CVD,W1,C0.5) and A(CVD,W1,C5) before and after use in the SCR of NO with isobutane have shown that further structural changes occur in a precalcined catalyst under reaction conditions (Fig. 4.3d and Table 4.1). The relative intensity of bands above 400 nm increases after use in the SCR reaction mainly at the expense of bands between 300 and 400 nm, while bands below 300 nm decrease only slightly. This is particularly evident for sample A(CVD,W1,C0.5) (Table 4.1). It suggests that, under reaction conditions, the growth of iron oxide particles is supported mainly by further agglomeration of oligonuclear clusters [55]. However, it cannot be excluded, that some isolated Fe sites are also involved in this agglomeration process.

↓151

In the EPR spectra (Fig. 4.18), those isolated Fe sites might be reflected by the signal at g' ≈ 6 which is completely missing after use in catalysis. Alternatively, the disappearance of the g' ≈ 6 signal may also be due to a reduction and/or a change of the coordination of the respective isolated Fe sites during reaction. In any case, the change of the isolated Fe sites reflected by the vanishing g' ≈ 6 signal does obviously not lead to a loss of catalyst activity since in the catalytic tests no deactivation was observed with time on stream.

By comparing the EPR and UV/VIS spectra for sample A(CVD,W1,C0.5) before and after use in the SCR of isobutane an apparent contradiction is evident (Figs. 4.3d and 4.18). While a huge increase of the EPR signal at g' ≈ 2 is observed after reaction, UV/VIS intensity in the range characteristic of agglomerated iron oxide species rises only slightly. This suggests that the observed changes are rather due to modified magnetic interactions within the oxidic clusters than to a marked increase of their concentration. The fact that the amount of small oligonuclear clusters decreases in favour of large oxide particles while isolated Fe sites mainly persist during reaction suggests that the latter are the active Fe sites of the SCR reaction. However, the changes between the amounts of oligomers and large aggregates (Table 4.1) are too small to safely exclude a contribution of the clustered phases. On the other hand, since EPR reflects a strong ordering tendency in the clustered phase and it is unlikely that the catalytic activity of clusters will be improved by increasing structural perfection, the structural changes observed on the whole support the view that isolated Fe sites play a deciding role in the SCR reaction as proposed, for instance, in [32,36,91]. However, this aspect will be discussed in detail in the following sections.

5.2 Structure-reactivity relationships in Fe-containing zeolites

Spectroscopic ex situ studies described in section 4.1 revealed that a variety of Fe species ranging from isolated ions via low oligonuclear clusters to large Fe2O3 particles may be formed in zeolites depending on peculiarities of the synthesis procedures and the nature of the matrices. Catalytic results described in section 4.4 together with findings from in situ EPR, UV/VIS-DRS and FT-IR studies suggest that the various Fe species play a different role in the three reactions studied in this work. Those relations between the nature of the Fe sites and their catalytic role are discussed separately for the abatement of NO and N2O in this section.

5.2.1 Low temperature interaction of NH3and NO with Fe-ZSM-5 zeolites

Interaction of NH 3 with samples A(MR)0.5 and A(CVD ,W1,C0.5 ) at 293 K

↓152

Interaction of NH3 with Fe-ZSM-5 zeolites at 293 K leads to the disappearance of EPR signals at g' ≈ 6 and shifts the UV/VIS band to a lower wavelength as observed in Fig. 4.31. The observed effects are most likely due to linking of NH3 to coordinatively unsaturated isolated Fe3+ ions without considerable reduction of the latter. This is evidenced by UV/VIS-DR spectra (Fig. 4.31c), which show similar spectral intensity before and after interaction of NH3. This indicates that isolated Fe3+ ions, in particular those reflected by EPR signal at g' ≈ 6 can extend their coordination while this is less pronounced for those iron species reflected by the signal at g' ≈ 4.3. It is difficult to discern between these two kinds of isolated Fe3+ ions by UV/VIS-DRS but it can be clearly seen from Fig.36c that part of the isolated Fe3+ ions are coordinated with NH3 as evidenced by a shift in the band position. However, considering differences in the area of UV/VIS spectra before and after NH3 treatment, reduction of a small amount of Fe3+ spices cannot be completely ruled out. Kucherov et al. have observed similar effects during in situ EPR studies over Fe-ZSM-5 prepared by sublimation [59]. In that study the EPR signals at g' ≈ 6.5 and g' ≈ 5.6 were suppressed upon adsorption of H2O, NO or NO2 at 293 K while the signal at g' ≈ 4.3 did virtually not change. They ascribed this behavior to a change in the symmetry of iron species, reflected by EPR signals at g' ≈ 6.5 and g' ≈ 5.6, from lower to higher symmetry and designated them as active Fe sites in the SCR reaction.

Adsorption of NO on oxidized and reduced surfaces of Fe-ZSM-5 at 293 K studied by FT-IR spectroscopy

FT-IR spectra in Fig. 4.32 and 4.33 show the formation of [NO+][N2O4], NO+, Fe3+-NO, N2O4, N2O3 and differently bound nitrato species upon adsorption of 1% NO/He on pre-oxidized Fe-ZSM-5 samples. This suggests the occurrence of oxidation, dimerization and disproportionation reactions over oxidized surface of the catalyst. The fact that they are found, though weakly, also after adsorption of NO on the bare H-ZSM-5 suggests that these processes can occur even in the presence of residual traces of gaseous O2 and/or Fe impurities (ca. 500 ppm) in the commercial H-ZSM-5 used for the preparation of the Fe-samples. The formation of the above mentioned species could follow as:

NO is obviously oxidized by Fe3+ ions as shown by Eq. (2.3) (page no. 18). The other possible route for the oxidation of NO is by traces of oxygen which might be present on the catalyst surface (Eq. (5.1)).

↓153

2NO + O2 → 2NO2 (5.1)

It is well known that NO2 can undergo dimerization to form N2O4 [86,180] as evidenced by a band at 1743 cm-1 in Fig. 4.32 (Eq. (5.2)).

2NO2 → N2O4 (5.2)

↓154

Subsequently, N2O4 disproportionates to form NO+, as evidenced by a band at 2134 cm-1, and NO3 - (Eq. (5.3)) [86,180]. However, the formation of NO3 - cannot be seen within the spectral region recorded for this study (Fig. 4.32).

N2O4 → NO+ + NO3 - (5.3)

Note that the reaction via the intermediate formation of N2O4 with subsequent disproportionation is only one of the routes for the formation of NO+ and NO3 - species. The other routes for the formation of these speciesare well documented in the literature [86-90,180]

↓155

The disproportionation of NO and NO2 leads to formation of N2O3 (Eq. (5.4)) [180].

NO + NO2 → N2O3 (5.4)

However, upon evacuation the weakly adsorbed [NO+][N2O4], NO+, Fe3+-NO, N2O4, N2O3 species are removed. In contrast, differently bound nitrato species (bands at 1633, 1620 and 1575 cm-1)remain even after evacuation indicating the higher stability of these species. This is likely due to the involvement of oxygen species that are directly connected to the Fe sites. Furthermore, the band at 1575 cm-1 is missing on the bare H-ZSM-5 and increases with rising Fe content from sample A(ILIE)0.3 to A(CVD,W1,C5) via A(ILIE)1.2. This indicates that NO3 species reflected by this band are preferably adsorbed on Fe species. The presence of traces of Fe might be responsible for the observed nitrate band around 1620 cm-1 on H-ZSM-5.

↓156

Adsorption of NO on reduced surface leads to different spectral features as compared to NO adsorption on oxidized surface (Figs. 4.32, 4.33, 4.34 and 4.35). Here, [NO+][N2O4], NO+, Fe3+-NO, N2O4, N2O3 species are almost completely absent and nitrato species are only weakly present. In contrast, the nitrito (1465 cm-1) species are formed over cluster containing sample A(CVD,W1,C5). This indicates the strong suppression of oxidation, dimerization and disproportionation of NO and NO2 reactions over reduced catalyst surface (Eqs. (2.3),(5.1)-(5.4)). Hence, the bands for nitrato species are only weakly present in A(CVD,W1,C0.5), A(ILIE)1.2 and A(MR). However, this is different over A(ILIE)0.3 as evidenced by the corresponding FT-IR bands in Fig. 4.34. This could be due to the possible contact of prereduced sample with air. Interestingly, the only sample with high iron content and with extensive iron oxide clusters (A(CVD,W1,C0.5)) shows nitrito species, at 1462 cm-1, upon adsorption of NO over reduced surface. This may be due to the differently linked NO or NO2 species to reduced iron species, presumably to reduced iron oxide clusters. However, it has to be taken into account that the severity of the reductive pretreatment for FT-IR measurements was only moderate (NH3/He at 673 K for 1 h) as compared for in situ UV/VIS and EPR (NH3/He or H2/He at 773 K for 1 h).In situ UV/VIS and EPR experiments revealed that even after 1 h at 773 K in NH3/He or H2/He flow no complete reduction of Fe3+ was possible (Figs. 4.10, 4.37 and 4.39). Nevertheless, the drop in band intensities in the in situ FT-IR spectra after reductive pretreatment is obvious. This suggests that, in the SCR reaction, trivalent Fe species are preferred as active sites in comparison to Fe2+ ions due to their ability to activate/oxidize NO to the corresponding NO2 or NO3. In agreement, it has been reported that reaction between NO and activated reductant (NH3 or isobutane) is almost negligible [37,46]. The activation of the reductant proceeds via surface acidity of the zeolite [37,46,49,162]. In contrast, oxidized NO (NO2 or NO3) readily reacts with activated reductants to form N2. Thus, it can be proposed that Fe3+ sites and acidity of the zeolite is essential for SCR of NO by NH3 and isobutane. The possible reaction between activated NO and reductants proceeds similarly as that described earlier in section 2.1.1 (Eqs. (2.1)-(2.10)).

5.2.2 SCR of NO with NH3and isobutane

Relation between structure of Fe sites and catalytic performance in SCR of NO

Ex situ UV/VIS-DRS and EPR results show that the series of ILIE samples comprises both samples with only isolated Fe sites (A(ILIE)0.2 and A(ILIE)0.3) as well as samples that are characterized by an increasing amount of additional Fe x O y clusters (A(ILIE)1.2) (Table 4.1). Therefore, it is possible to investigate the catalytic role of isolated Fe3+ sites discretely and iron oxide clusters, however, along with isolated Fe sites, by comparing the catalytic activity of these samples.

In NH3-SCR, activity increases with increasing amount of isolated Fe sites in samples A(ILIE)0.2 to A(ILIE)0.3 (Fig. 4.55), which contain almost exclusively isolated Fe sites (Table 4.1). This indicates the participation of isolated sites in the catalytic reaction. Accordingly, A(MR)0.5 with mainly isolated Fe sites showed similar catalytic behaviour as sample A(ILIE)0.3. However, the dramatic increase of the concentration of oligomeric sites from A(ILIE)0.3 to A(ILIE)0.6 improves the activity much more than the gradual growth of the isolated site concentration (Table 4.1 and Fig. 4.55). This suggests a remarkable contribution of the oligomers to the SCR of NO with NH3. This is further supported by correlating the percentage of different Fe sites (isolated, oligomers and particles) as derived by UV/VIS-DRS (Table 4.1) and first order rate constants of the SCR reaction [64]. Fig. 5.1 shows the results for different correlation attempts for NH3-SCR at 523 K, which is well below the conversion maximum.

↓157

Fig. 5.1. Correlation of SCR rates (first order rate constants) with the concentration of different Fe species detected by UV/VIS-DRS (Table 4.1) [64].

It is clear from Fig. 5.1 that the rate constants do not correlate neither with the amount of isolated Fe sites nor with the amount of Fe x O y oligomers alone. There is a much better correlation with the percentage of both isolated and oligomeric Fe sites. However, the best correlation is obtained with the total Fe content. This suggests that all accessible Fe sites including, besides isolated and low oligomeric Fe, also Fe sites on the surface of oxidic particles participate in selective reduction at temperatures below 700 K (Fig. 4.55). At higher temperatures (T > 700 K) NO conversion was observed to level off or even to drop down for those samples that are characterized by a remarkable percentage of oxidic clusters (Fig. 4.55, sample A(CVD,W1,C5)). Since this undesired behavior is not observed for Fe samples free of clusters, it suggests that Fe x O y agglomerates play a detrimental role at high temperatures by favoring the total oxidation of the NH3 reductant which is then no longer available for the SCR of NO. Differently, isolated Fe3+ species of different coordination are efficient for reducing NO in a whole range of temperature as evidenced by the catalytic behavior of cluster free samples (Fig. 4.55). However, these Fe species behave differently under steady-state SCR conditions as evidenced by in situ UV/VIS and in situ EPR studies. Thus, it was found that in cluster free samples, octahedral Fe3+ sites reflected by EPR signal around g' ≈ 6 and UV/VIS band around 290 nm are more sensitive to reduction. In comparison, tetrahedral Fe3+ sites shown by EPR signal at g' ≈ 4.3 and a UV/VIS band around 230 nm are less sensitive. Differently, Fe3+ sites reflected by EPR signal at g' ≈ 2, for which coordination state cannot be simply specified, are hardly reduced.The amount of these Fe species preferably increases with rising Fe content from sample A(ILIE)0.3 to A(ILIE)1.2 as evidenced by a comparison of the UV/VIS spectral intensities below 300 nm under steady-state SCR conditions (Figs. 4.38 and 4.40). On the other hand, in situ FT-IR results show preferential reaction of NO with Fe3+ to form oxidized products which are important intermediates in the SCR reaction (Fig. 4.32). Based on these considerations it can be concluded that among isolated Fe sites, the hardly reducible Fe sites are the active sites for SCR of NO. In agreement, the NO conversion increases with increasing the amount of hardly reducible Fe sites from samples A(ILIE)0.2 to A(ILIE)1.2 (Fig. 4.55).

By comparing the isobutane-SCR activity data of A(ILIE)0.2 and A(ILIE)0.3 the participation of isolated sites in the reaction is obvious (Fig. 4.56). In particular, A(ILIE)0.3, which does not contain particles and only a very small amount of oligomers, shows even the highest NO conversions in a broad temperature range. Apparently, particles can be therefore safely rejected as active sites which is in agreement with observations made earlier by other research groups [36,86]. At the same time the extremely low quantity of oligomers (more than an order of magnitude less than in the cluster-containing catalysts such as A(ILIE)1.2) strongly discourages an assignment of the catalytic activity exclusively to oligomers. The data imply rather that mononuclear Fe species govern the activity in SCR catalysis, probably together with a certain contribution of oligomeric entities. This is also suggested by a correlation of the SCR rate with the concentration of different Fe species (Fig. 5.2). It is clear from this figure that the rate constant correlates best with the sum of the iron in isolated and oligomeric sites, whereas there is no proportionality with the concentration of the oligomeric sites alone. Therefore, a major contribution of the isolated Fe sites can be identified. The correlation of the rate constant with the concentration of isolated sites yields a bent curve, which implies the participation of a second type of Fe site, the oligomers. Apparently, the surface of the particles does not contribute to the SCR reaction as evidenced by the correlation of the rate constants with the total Fe content, which does not go through the origin. The increase of the concentration of oligomeric sites from A(ILIE)0.3 to A(ILIE)1.2 (Table 4.1) enhances the activity below 600 K but above this temperature NO conversion drastically deteriorates (Fig. 4.56). This suggests that oligomeric sites, which are present in highly oxidized state under typical SCR conditions as evidenced by in situ spectroscopic studies (section 4.3.2), do contribute to SCR reaction below 600 K but above this temperature they promote the unselective total oxidation of isobutane. This tendency has been observed, too, for NH3-SCR. However, it is much more pronounced for isobutane-SCR.

↓158

Fig. 5.2. Correlation of SCR rates (first order rate constants) with the concentration of different Fe species detected by UV/VIS-DRS (Table 3) [64].

In summary, hardly reducible isolated Fe sites reflected by EPR signals atg' ≈ 2 and partly also by g' ≈ 4.3 are the active Fe sites for both the SCR reactions. Whereas, easily reducible isolated Fe sites reflected by the signal at g' ≈ 6 are already reduced at 623 K under typical SCR conditions as evidenced by in situ EPR. Therefore, it cannot be excluded that these sites are only spectator species at and/or above 623 K. Consequently, the role of isolated Fe sites in the SCR may increase in the following order: g' ≈ 6 < g' ≈ 4.3 < g' ≈ 2. Oligomers including dimers are also active sites for both SCR reactions, in particular for NH3-SCR. However, these species are only selective at lower temperatures but at higher temperatures they are unselective and contribute to total oxidation of the reductant, which is much more pronounced for isobutane than for NH3. Therefore, for isobutane-SCR clustered-type active sites are not essential and must be avoided, while these species play, to a certain extent, a constructive role in NH3-SCR.

Influence of the acidity on the SCR of NO

The role of acidity on the SCR of NO either by NH3 or by isobutane is not much studied. Hence, in this work an effort is made to shed some light on this aspect as well [161,162]. For this purpose, the catalytic behavior of samples ex-Fe-silicalite, A(ILIE)0.2, A(ILIE)0.3, A(ILIE)1.2 and B(CVD,W1,C5) was compared. Samples ex-Fe-silicalite, A(ILIE)0.2 and A(ILIE)0.3 contain almost exclusively isolated Fe sites, while A(ILIE)1.2 and B(CVD,W1,C5) are dominated by clusters (Table 4.1). However, framework composition of the samples is different (Table 3.1). Ex-Fe-silicalite does not contain any Al in the framework whereas B(CVD,W1,C5) is prepared with silica rich matrix H-ZSM-5(B) (Si/Al ≈ 40). Accordingly, ex-Fe-silicalite shows very poor acidity as evidenced by FT-IR of pyridine adsorption (Fig. 4.30 and Table 4.3). As expected, B(CVD,W1,C5) contains a very low Brønsted acidity but considerable amount of Lewis acidity. The latter acidity is mainly due to the presence of iron oxide clusters in the sample as concluded in the results section (Fig. 4.28 and Table 4.3). Samples A(ILIE)0.2 and A(ILIE)1.2 with Si/Al ≈ 14 show much higher acidity (Fig. 4.29. Although the acidity was measured for sample A(ILIE)0.2, the same acidity can be expected for A(ILIE)0.3 since both the samples contain similar Si/Al ratio (14) and show similar nature and distribution of Fe species (section 4.1).

↓159

In both NH3- and isobutane-SCR, in spite of lower amount of isolated Fe entities with respect to the iron content (Table 4.1), A(ILIE)0.2 and A(ILIE)0.3 exhibit the highest activity. Contrarily, ex-Fe-silicalite, which contains a higher amount of isolated Fe sites than the most active A(ILIE)0.3 (Table 4.1), shows a very poor activity in both SCR reactions (Fig. 4.55 and 4.57). This strongly indicates that the presence of Fe sites in high dispersion is not a sufficient criterion for high activity. Definitely, acidic sites are inevitable as well. Additionally, by comparing the catalytic behavior of B(CVD,W1,C5) and A(ILIE)1.2, which both contain heterogeneous distribution of Fe sites, the importance of Brønsted and Lewis acidity even in the presence of active sites of the isolated and clustered-type for the SCR is discussed. Despite containing double the amount of Fe content (Table 3.1) as compared to A(ILIE)1.2, B(CVD,W1,C5) shows moderate activity in NH3-SCR while it completely failed in isobutane-SCR. This suggests that the presence of clusters, which can also act as Lewis sites, compensates the low Brønsted acidity to some extent in NH3-SCR but this is not the case in isobutane-SCR. However, the observed NH3-SCR activity for sample B(CVD,W1,C5) is still markedly lower than that of A(ILIE)1.2 despite the higher Fe content of the former due to the absence of sufficient acidity.

In summary, on the basis of these observations it can be concluded that for both NH3-and isobutane-SCR, acidity (Brønsted and Lewis acidity) is compulsory even in the presence of active sites of isolated and clustered-type. Since, acidic sites are required for the activation of reductants which can only then further react with NOx (x > 2) and NO to form N2.

Influence of the pore structure on the SCR of NO

The influence of the pore structure on the SCR of NO is evident by comparing the behavior of ex-Fe-silicalite and (Fe-SBA-I)0.95. These two catalysts have comparable Fe contents, comprise almost exclusively isolated Fe sites and show almost no acidity (section 4.1 and 4.2). Due to this low acidity, both catalysts are much less active than ZSM-5 based materials (Fig. 4.55 and 4.57). Nevertheless, a clear difference is seen that reveals higher catalytic activity for ex-Fe-silicalite than (Fe-SBA-I)0.95. This is line with the earlier studies on this topic which revealed that open mesoporous framework structures such as SBA-15 are less favorable for the SCR of NO as compared to microporous structures like MFI [158]. Therefore, it is reasonable to conclude that the mesoporous structure such like SBA-15 is not favorable for this reaction. In contrast, the unique pore structure of the MFI is certainly an additional asset for this reaction.

Optimization strategies for preparing Fe-ZSM-5 zeolites for SCR of NO

↓160

On the basis of the above discussion and observations, an uniform concept for tailoring new Fe-ZSM-5 zeolites for SCR of NO can be drawn. The increase in the concentration of clustered sites from samples A(ILIE)0.2 and A(ILIE)0.3 to samples A(ILIE)0.7, A(ILIE)1.2 and CVD leads to a continuous deterioration of the catalytic behavior in isobutane-SCR at temperatures above 600 K (Fig. 4.56). The bridging oxygen, be it in binuclear complexes as favored by [31,56,57] or in oligomers of a wide range of nuclearities as suggested in [36], probably participates in the SCR at low temperatures. However, at high temperatures, it tends to attack the reductant unselectively leading to its total oxidation. This can be easily understood from reduction/reoxidation kinetic measurements of A(ILIE)0.3 and A(ILIE)1.2 (see section 4.1.1) which show that iron oxide clusters are hardly reduced but fastly reoxidized. In agreement with these results, in situ spectroscopic studies (sections 4.3.1 and 4.3.2) show that iron oxide clusters are essentially in +3 oxidation state under typical SCR conditions which has a higher oxidation potential than reduced Fe valence state. Therefore, on the basis of these experimental evidences it is reasonable to ascribe the unselective total oxidation of isobutane to iron oxide clusters.

In NH3-SCR, the same trend can be seen, however at much higher temperatures. With the catalysts containing almost exclusively isolated sites, unselective NH3 oxidation cannot be detected up to 873 K. With increasing content of clustered sites, the NO conversion starts to level off under more and more moderate conditions (Fig. 4.55). In situ spectroscopic studies show that under SCR conditions the iron sites are much less reduced in NH3-SCR than in isobutane-SCR feed. Furthermore, studies of subsequent interaction of isobutane or NH3 and NO with Fe-zeolites show that Fe species are stronger reduced by isobutane than by NH3 as clearly evidenced by in situ EPR spectroscopy (sections 4.3.1 and 4.3.2). These results indicate that NH3 is not as easily oxidized than isobutane. This could explain why clustered sites play a rather constructive role in NH3-SCR but not in isobutane-SCR.

The above observations form the basis for optimization strategies for Fe-ZSM-5 in these reactions. For isobutane-SCR, the concentration of isolated sites has to be increased avoiding clustered sites. For NH3-SCR, a high dispersion of iron should be sought as well to increase the number of accessible Fe sites. However, in contrast to isobutane-SCR, these sites may be part of low oligomeric clusters, too. Acidity is essential for both SCR reactions to activate the reductant.

5.2.3 Decomposition and SCR of N2O with CO

↓161

For this study, differently prepared Fe-catalysts (ex-Fe-silicalite, ex-Fe-ZSM-5, A'(CVD,W1,C2), Fe-ZSM-5(LIE)1.4, A(ILIE)0.2, A(ILIE)1.2 and (Fe-SBA-I)0.95) with different framework composition, nature and distribution of Fe sites and pore structures have been used to investigate the role of these properties on the decomposition and SCR of N2O. The nature and distribution of Fe species in these materials strongly depend on the framework composition and preparation route as evidenced by UV/VIS-DRS and EPR (section 4.1).

In direct N2O decomposition, O2 desorption is the rate-determining step, which can be accelerated at a high temperature, in the presence of NO [41], and also by addition of reductants. Interestingly, in spite of containing similar Fe content ex-Fe-ZSM-5 and ex-Fe-silicalite show significant differences in the activity of N2O decomposition. This is attributed to the distinct Fe constitution in the samples and, thus, ex-Fe-ZSM-5 with extensive clusters was found to be a more effective catalyst than ex-Fe-silicalite with exclusively isolated Fe sites [96,155]. In this case, the isolated nature of iron in ex-Fe-silicalite makes the recombination of N2O deposited oxygen (O*), which is the rate-determining step, difficult according to Eq. 2.13, even though the rate of Eq. 2.11 is higher than over ex-Fe-ZSM-5 [58]. Differently, the oligomeric Fe sites facilitate the recombination of O* due to the presence of Fe ions in a closer vicinity. In agreement with this, A'(CVD,W1,C2) which contains both isolated and oligomeric Fe sites in a larger quantity than the former two samples (Table 4.1) shows the best activity.

In the SCR of N2O, CO efficiently removes atomic oxygen from the catalyst surface and leads to a substantially decreased operation temperature with respect to the direct N2O decomposition. The reaction of N2O with CO is stoichiometric as shown by Eq. 4.4 and the N2O conversion increases linearly with the molar feed CO/N2O ratio (Fig. 4.60) [58].

↓162

The remarkably uniform distribution of iron in ex-Fe-silicalite as isolated Fe ions connected to its high catalyst activity is a strong indication of the importance of mononuclear iron ions in the reduction of N2O with CO. Fig. 5.3 shows the derived correlation between the relative fraction of isolated Fe sites estimated from the relative intensity of the UV/VIS-DRS subbands below 300 nm (Table 4.1) and the relative N2O conversion derived from the catalytic activity at 600 K.

Fig. 5.3. Correlation between the relative N2O conversion at 600 K (CO/N2O=1 and W/F (N2O) = 9 X 105 g s mol-1) and the relative fraction of isolated Fe3+ sites (from Table 4.1) in the Fe-MFI zeolites investigates. Values are referred to ex-Fe-silicalite [58].

The N2O conversion and the percentage of isolated Fe3+ sites of the catalysts (normalized to the values of ex-Fe-silicalite) follow the sequence:

↓163

A'(CVD,W1,C2) > ex-Fe-silicalite > ex-Fe-ZSM-5 ≈ Fe-ZSM-5(LIE)1.4

The poor catalytic performance of Fe-ZSM-5(LIE)1.4 and ex-Fe-ZSM-5 is due to the formation of iron oxide particles which render the majority of Fe sites inaccessible and, thus, inactive. Iron clustering also occurs in the catalyst prepared by sublimation (A'(CVD,W1,C2)), but the high iron content together with a relatively high fraction of iron in intrazeolitic positions makes it the most active formulation in terms of absolute N2O conversions. The activity of ex-Fe-ZSM-5 is largely reduced compared to ex -Fe-silicalite, which correlates with the decreased concentration of isolated iron species due to clustering in the former sample.

Finally, it should be noted that the correlation in Fig. 5.3 assumes that all isolated sites in the steam-activated samples are in extraframework positions. This is the case for ex-Fe-ZSM-5, as concluded in [154,155]. For ex-Fe-silicalite, a certain (small) fraction of redox inactive framework iron cannot be excluded. This fraction was not quantified here, since framework and extraframework species cannot be discriminated in UV/VIS-DRS. In rigorous terms, the fraction of remaining framework iron in ex-Fe-silicalite should be subtracted from the total fraction of isolated sites in Fig. 5.3, since Fe3+ ions in tetrahedral framework positions do not contribute to the catalytic activity [188].

↓164

However, the participation of cluster species in the SCR of N2O cannot be excluded. In situ UV/VIS-DRS and EPR spectroscopic studies revealed important differences in the redox behaviour of the iron species in the catalysts upon interaction with O2, N2O, and CO, suggesting that both isolated Fe3+ sites as well as oligonuclear Fe3+ x O y species are active in the reaction of N2O with CO at 623 K. The reaction mechanism associated with these species differs substantially. In situ UV/VIS-DRS analyses have clearly shown that Fe3+ in oligonuclear Fe3+ x O y clusters can be preferably reduced by CO and can be completely reoxidized by N2O. In combination with EPR results, it can be proposed that the reduction of N2O with CO over these sites occurs according to Eqs. (5.5) and (5.6), i.e., via intermediate formation of O- radicals. This mechanism agrees with the classical hypothesis stating that CO purely acts as an O-scavenger leading to CO2 (Eq. 5.5) and regenerating the active site (ڤ) for subsequent N2O activation [96,117,118].

(5.5)

(5.6)

↓165

Differently, the (dominant) tetrahedral isolated Fe 3+ ions in extraframework positions of ex -Fe-silicalite suffer no reduction by CO at 623 K (Fig. 4.48). The relatively weak interaction between CO and iron ions is commonly accepted [62]. However, the changes of the local symmetry and altered position of the signals in the EPR spectra suggest that CO coordinates to ferric ions in ex-Fe-silicalite.

Attending to these results, the mechanism represented by Eqs. (5.7) and (5.8) can be proposed for the N2O reduction with CO over tetrahedral isolated Fe sites.

(5.7)

↓166

(5.8)

This mechanism implies that the involved isolated Fe species must be coordinatively unsaturated in order to chemisorb CO. This holds for species associated with the g' ≈ 2 as well as the g' > 5 signals in ex-Fe-silicalite (Fig. 4.48). The observations related to the latter group of signals are in excellent agreement with various EPR studies, where the reactivity of the EPR signals around g' ≈ 6 is elaborated. Kucherov et al. [59] determined by in situ EPR that the signals at g' ≈ 5.6 and 6.5 observed in Fe-ZSM-5 catalysts prepared by sublimation disappear upon contact with NO or NO2. The corresponding signals were assigned to very reactive coordinatively unsaturated Fe3+ species. Volodin et al [189] also attributed the disappearance of this line (in vacuum and in the presence of H2O and NO) as an indication for the reactivity of an ion-exchanged Fe-ZSM-5 with N2O. Ribera et al. [60] found that the signals at g' ≈ 6.4 and 5.7 in steam-activated Fe-ZSM-5 disappear upon addition of β-mercaptoethanol, concluding their participation in redox processes. In a recent study, Kubánek et al [190] have correlated the intensity of the signals at g' ≈ 6.0 and 5.6 and the activity of H-ZSM-5 zeolites with low Fe concentrations in the oxidation of benzene to phenol with N2O. These signals were assigned to distorted Td-coordinated isolated ions with a complex oxo-structure in cationic sites.

In agreement with this discussion, A(ILIE)0.2 which contains almost exclusively isolated Fe sites shows similar behavior to that of ex-Fe-silicalite. Furthermore, with increasing amount of isolated and oligomeric sites in the A(ILIE)1.2 the activity increases considerably than the former two samples. This confirms that isolated Fe sites are determining the SCR activity. However, besides isolated Fe sites oligomeric sites are also participating in the reaction.

↓167

In conclusion, isolated Fe sites and oligomeric sites are active Fe centers for both decomposition and SCR of N2O. However, the isolated Fe sites are more effective in the latter reaction. Differently, oligomeric Fe sites are more active in the decomposition due to the easier oxygen recombination over these species, which is the rate-determining step. However, these two types of active Fe sites involve differently in the SCR of N2O. Coordinatively unsaturated isolated Fe sites coordinate the CO without being reduced. The reaction over oligomeric sites proceeds via a redox Fe3+/Fe2+ process probably via intermediate formation of O- radicals.

Influence of the acidity on the decomposition and SCR of N 2 O

To study the role of acidity on these reactions the catalytic performance of ex-Fe-silicalite, A(ILIE)0.2, A'(CVD,W1,C2) and ex-Fe-ZSM-5 is compared. Ex-Fe-silicalite does not contain any lattice Al sites and thus a weak surface acidity was measured by FT-IR spectroscopy (Fig. 4.30). A(ILIE)0.2 with Si/Al ≈ 14 shows markedly higher acidity (Fig. 4.29). Although the acidity was not measured for A'(CVD,W1,C2) and ex-Fe-ZSM-5, similar acidity can be expected as observed for A(CVD,W1,C5) and ex-Fe-beta respectively (Table 4.3 and Figs. 4.28, 4.30). Since, A'(CVD,W1,C2) and ex-Fe-ZSM-5 contain similar Si/Al ratio, Fe content and nature and distribution of Fe sites as that of A(CVD,W1,C5) and ex-Fe-beta respectively (section 4.1). Accordingly, A'(CVD,W1,C2) should contain much higher acidity than ex-Fe-ZSM-5. In spite of containing roughly ten times lower Fe content, ex-Fe-ZSM-5 shows comparable activity in direct N2O decomposition to that of A'(CVD,W1,C2) (Fig. 4.58). In SCR of N2O with CO, ex-Fe-silicalite shows similar activity to that of A(ILIE)0.2 (Fig. 4.61). Thus, it is clear that the surface acidity of the catalyst is not essential for these reactions.

Influence of the pore structure on the decomposition and SCR of N 2 O

The effect of micro and mesoporous framework structure on the catalytic N2O decomposition and N2O reduction with CO has been nicely illustrated by the performance of ex-Fe-silicalite and (Fe-SBA-I)0.95. From UV/VIS-DRS and EPR characterization results it is evident that content and nature of iron sites in the two catalysts are very similar (section 4.1). Both ex-Fe-silicalite and (Fe-SBA-I)0.95 contain almost exclusively isolated Fe3+ sites of very similar structure but differ markedly in their pore structure. While ex-Fe-silicalite is characterized by the well known MFI structure consisting of intersecting straight and sinusoidal channels of 0.55 nm, the structure of (Fe-SBA-I)0.95 is dominated by parallel linear pores, the diameter of which is more than ten times as large as in ex-Fe-silicalite [161]. Ex-Fe-silicalite revealed to be much more active than (Fe-SBA-I)0.95 in both direct N2O decomposition and N2O reduction with CO. This is certainly due to the difference in the pore structure. Obviously, the large pores in (Fe-SBA-I)0.95 do not support an intimate contact between active iron sites and reactant molecules, most of the latter passing the pore system of (Fe-SBA-I)0.95 without approaching active sites and, consequently, give rise to low activity. Furthermore, ex-Fe-ZSM-5 and ex-Fe-beta with similar Fe content, nature and distribution of iron sites and pore structure (section 4.1) exhibit similar activity in both SCR and direct decomposition of N2O (not shown) as reported elsewhere [159]. Therefore, it can be concluded that the micropore structure like MFI is favorable for these reactions since it facilitates intimate contact between active Fe sites and reactants.

Optimization strategies for preparing Fe-MFI zeolites for decomposition and SCR of N 2 O

↓168

On the basis of these results it is possible to draw an unified concept for the behaviour of Fe-zeolites in both reactions. The high specific activity of ex-Fe-silicalite, with a remarkable uniform distribution of isolated iron ions, is essential to conclude the importance of mononuclear iron ions in the reduction of N2O by CO. In agreement, in situ UV/VIS-DRS and in situ EPR studies evidenced the participation of isolated Fe sites in the reaction, however, they also support the involvement of oligomeric Fe sites (section 4.3.3).

In direct N2O decomposition,samples A'(CVD,W1,C2) and ex-Fe-ZSM-5 which are dominated by oligomeric clusters show higher activity than ex-Fe-silicalite. It indicates the preference of bridging oxygens, be it in binuclear complexes or in oligomers of a wide range of nuclearities, over isolated iron ions in the reaction. This is obviously due to the easier oxygen recombination, which is the rate-determining step in the reaction, of two iron centers that are close together. On the basis of the poor performance of Fe-ZSM-5(LIE)1.4 with extensive clustering, mainly large iron oxide particles, particles can be safely rejected as active sites in SCR and direct decomposition of N2O.

The results are evidence for the requirement of quite different optimization strategies for the tailoring of Fe-MFI zeolites for these two reactions. For the reduction of N2O with a reducing agent, the concentration of extraframework isolated iron ions has to be increased avoiding clustered sites as concluded for isobutane-SCR of NO. For N2O decomposition, a high dispersion of iron should be required as well but along with small oligomeric iron oxide clusters. This is a similar strategy as concluded for NH3-SCR of NO. Finally, the acidity is not mandatory for both reduction and direct decomposition of N2O.


© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
DiML DTD Version 4.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML generated:
09.11.2005