Properties of Iron-Modiﬁed-by-Silver Supported on Mordenite as Catalysts for NO x Reduction

: A series of mono and bimetallic catalysts based on a Fe-Ag mixture deposited on mordenite was prepared by ion-exchange and evaluated in the catalytic activity test of the de-NO x reaction in the presence of CO / C 3 H 6. The activity results showed that the most active samples were the Fe-containing ones, and at high temperatures, a co-promoter e ﬀ ect of Ag on the activity of Fe catalysts was also observed. The inﬂuence of the order of cation deposition on catalysts formation and their physicochemical properties was studied by FTIR (Fourier Transform Infrared Spectroscopy) of adsorbed NO, XANES (X-ray Absorption Near-Edge Structure), and EXAFS (Extended X-ray Absorption Fine Structure) and discussed in terms of the state of iron. Results of Fe K-edge XANES oscillations showed that, in FeMOR catalysts, iron was present in a disordered state as Fe 3 + and Fe 2 + . In FeAgMOR, the prevailing species was Fe 3 + , while in the AgFeMOR catalyst, the state of iron was intermediate or mixed between FeMOR and FeAgMOR. The Fe K-edge EXAFS results were characteristic of a disordered phase, the ﬁrst coordination sphere being asymmetric with two di ﬀ erent Fe-O distances. In FeAgMOR and AgFeMOR, coordination of Fe-O was similar to Fe 2 O 3 with a few amount of Fe 2 + species. We may conclude that, in the bimetallic FeAgMOR and AgFeMOR samples, a certain amount of tetrahedral Al 3 + ions in the mordenite framework is replaced by Fe 3 + ions, conﬁrming the previous reports that these species are active sites for the de-NO x reaction. Based on the thermodynamic analysis and experimental data, also, it was conﬁrmed that the order of deposition of the components inﬂuenced the mechanism of active sites’ formation during the two steps ion-exchange synthesis.


Introduction
Reducing NO x emissions (de-NO x ) is an important problem facing modern technology. NO x plays a critical role in air pollution and in the physicochemical processes in the troposphere due to its chemical activity in the environment, which causes photochemical smog and acid rain [1,2]. Moreover, the role of NO x emissions in global warming is also of paramount importance [3] since N 2 O has an absorption and Ag 0 nanoparticles formed by redox cycles. The same has been observed by other authors [45][46][47]. Therefore, Ag has been studied as a promoter in de-NO x catalysts to increase the operating temperature range [48]. Although Ag is well known as an iron modifier, their interaction can affect structural, adsorptive, electronic, redox, and catalytic properties in various reactions. Given the individual catalytic properties of silver and iron, the interaction between these metals allows us to develop more efficient catalytic centers for the de-NO x reaction in order to achieve high NO x conversions.
The order of deposition of the metal components in the preparation of bimetallic catalysts has been shown to strongly influence the redox interaction of metals, contributing to the formation of sites of interest for HC-SCR of NO x [49,50]. Based on these premises, the goal of this work was to study the formation of Fe-Ag species deposited on mordenite and their physicochemical properties, focusing on a two steps ion-exchange procedure with a different order of insertion. The catalytic properties of the studied materials were tested for the environmentally important reaction of de-NO x by HC-SCR.

Results and Discussion
Mordenite is a medium pore size zeolite, where the negative charge induced by the presence of aluminum in the neutral silicate framework needs to be compensated by cations. Charge compensating exchangeable cations are located in the zeolite channels and pores. The sum of the charges of all counterions divided by the aluminum content (called the equilibrium ion-exchange modulus (EIEM)) should be equal to 1. In this study, the cations used are Ag + , Fe 2+ , Fe 3+ , and H + (from the acidified iron precursor solution), along with residual Na + of the initial mordenite (if any): We have to underline that the EIEM calculations based on the integral concentration of the components are complicated due the presence of some amount of reduced Ag 0 , which are forming clusters and nanoparticles, as well as the dual role of Fe, which partially balances the charge of Al-containing tetrahedra, but also partially substitutes Al in tetrahedral coordination, and also needs to be equilibrated by some extra lattice cations.
The chemical composition of the samples, as measured by ICP-OES (Inductively Coupled Plasma-Optical Emission Spectroscopy), is presented in Table 1. By comparing the Si/Al ratio of monometallic and bimetallic samples with the original value for the NaMOR, it can be seen that all samples, except AgMOR, have presented a weak dealumination, which is attributed to the acidic pH of the iron-containing solution. The EIEM is calculated from the ICP-OES results for all samples (see Table 1). Different values are observed between NaMOR, AgMOR, and Fe-containing samples. Since it is impossible to estimate the number of iron cations those who enter in the framework and evaluate the ratio between the coexisting Fe 2+ and Fe 3+ cations [50] from the ICP-OES data, the limit values of EIEM-Fe 2+ and EIEM-Fe 3+ are calculated, considering that all the iron is entirely as cations of one of the possible oxidation states (Fe 2+ or Fe 3+ ). The real EIEM for our samples should approximate to an intermediate value between the two calculated idealized estimations: For most samples, EIEM is close to 1, with the exception of NaMOR and FeMOR. The high EIEM value for NaMOR has been previously associated with the fact that NaOH, which is present during the synthesis of zeolite in the industrial process, is not completely removed by washing of the resulting material [50]. For FeMOR, the EIEM has a value of less than 1. This can be explained by the participation of H + in the process of ion exchange. Since it is not possible to evaluate the amount of H + by ICP-OES, therefore, its content is not considered in the EIEM calculations.
It is also seen that despite the same nominal relative concentration per unit charge of both solutions of Ag + and Fe 2+ precursors, the ion exchange of sodium for iron in FeMOR is significantly lower than for silver in AgMOR. In addition, a lower Fe content is observed compared to Ag in both bimetallic samples. This can be explained by the difference in the selectivity of the exchange of Na + ions for Ag + and Fe 2+ cations. The second step of ion exchange during the preparation of bimetallic samples do not have a significant effect on the sodium content but decreases the content of the first exchangeable cation; that is, the second stage of exchange occurs due to the leaching of cations that entered in mordenite during the first stage of the exchange.
In multimetallic materials based on zeolites, cation competition for ion exchange sites can take place. The difference in the adsorption capacity of zeolites for heavy metal ions may be attributed to several factors, including hydrated ion diameters, hydration enthalpies, charge-to-radius (Z/r) ratio, and cation solubility, as discussed by Motsi et al. [51].
In addition to those mentioned above, it has been reported that the adsorption of nickel (II) ions on Na-mordenite depends on pH [52]. Both Zn 2+ [53] and Co 2+ [54] exhibit practically the same behavior on different zeolites. Nakamura et al. [55] listed the requirements that determine the adsorption capacity of zeolites:

•
The radius of micropores must be comparable to the ionic radius of the target cation.

•
The Si/Al ratio should not be very high; moderate ratios are preferred; in such case, the zeolite framework becomes more negatively charged, and then the cation adsorption capacity is enhanced.

•
A uniform distribution of ion-exchange centers (i.e., Al atoms) around channels and cavities is required [56].
To evaluate the adsorption capacity of mordenite with respect to target cations, we have examined the ratio of charge to radius Z/r, enthalpy of hydration [57], and dissociation energy of cation-oxygen bonds [58] (see Table 2). The enthalpy of hydration is the amount of energy released due to the interaction of one mole of ions with water dipoles and characterizes the solubility of ions. The dissociation energy of the cation-oxygen bond is one of the indicators of the strength of this chemical bond. Thus, the difference between these two energetic contributions can characterize the transition from the state associated with the zeolite to the dissolved state. Based on these theoretical parameters, we might suggest the mechanism of adsorption. As can be seen from Table 2, for all considered cations, this transition is exothermic. −∆H • changes in the sequence: Na + > Ag + > H + > Fe 2+ > Fe 3+ , which means that all additional cations are able to replace Na + , which balances the negative charge of the zeolite structure. Thus, based on the data in Table 2, it can be assumed that during the preparation of FeMOR (treatment of NaMOR in a solution containing Fe 2+ and H + ), Na + is exchanged for both H + and Fe 2+ . This is in good agreement with the ICP-OES data presented in Table 1, where the values of both EIEM-Fe 2+ and EIEM-Fe 3+ are below 1, which confirms the presence of H + in FeMOR. From Table 2, it can also be assumed that H + can replace Ag + , also during the synthesis of AgFeMOR (AgMOR treatment in a solution containing Fe 2+ and H + , EIEM-Fe 2+ = 0.95), but Ag + could not replace H + during the preparation of FeAgMOR. The sequence to form this system is Fe(H + )MOR in the Ag + solution (EIEM-Fe 2+ = 1.09). Since the EIEM-Fe 2+ for both bimetallic samples is close to 1, we cannot assume the contribution of H + to the second ion-exchange step of the synthesis. According to Table 2, iron ions have the greatest exothermic effect; however, among all the studied cations, these two are multi charged, and their interaction with singly charged Al-tetrahedrons in the zeolite lattice may have steric limitations.
The treatment of FeMOR with an Ag + solution mainly leads to the exchange of Na + remaining in the sample for Ag + . At the same time, the contact between Ag + and Fe 2+ on the surface creates a redox interaction, leading to the formation of Ag 0 and Fe 3+ in MOR. A zero-charged silver atom could not displace Fe 3+ from its position. Thus, it is expected that a mixture of Ag 0 and Ag + species will form in FeAgMOR, with the simultaneous formation of a significant amount of Fe 3+ . Treatment of AgMOR in the Fe 2+ plus H + solution can lead to the exchange of the Na + residue with H + and Fe 2+ . Otherwise, the redox interaction between AgMOR and Fe 2+ cations causes Ag 0 to leave the negatively charged site. At this point, the vacant negatively charged site [-]MOR may be occupied by oxidized Fe 3+ or Fe 2+ or H + from the solution. Thus, in the case of AgFeMOR, a large number of Ag 0 -species and a mixture of Fe 2+ and Fe 3+ are expected. In the future, an attempt will be made to further refine the suggested mechanism for preparing samples with a different order of cation deposition, using additional data from various experimental methods, but such study is beyond the scope of this work and will be published elsewhere.
The textural properties of the samples are listed in Table 3. A comparison of the properties (surface area, pore-volume, and pore diameter) of the synthesized samples with the initial NaMOR indicates that no significant variations have been observed. It can be noted that the FeMOR sample presents a slight increase in the specific surface area, about 60 m 2 ·g −1 , but this is exactly the sample in which there is a significant fraction of ion-exchanged protons (see EIEM value, Table 1), which probably reduces steric hindrances to adsorption. Qualitative and quantitative analysis of the Brønsted surface acidity of the synthesized materials has been performed by the method of ammonia TPD (Temperature Programmed Desorption) in the range of 100-550 • C, and the data are presented in Table 4. Desorption of ammonia for all the samples is characterized by two main peaks: one at about 180 • C (l-peak), which can be correlated with weak acid sites, and another about 250 • C (h-peak), which corresponds to strong acid sites. In the case when silver is exchanged at the first stage (AgMOR and AgFeMOR), the ratio of weak/strong acid centers does not change and is in good agreement with the initial data for NaMOR (~30:70). However, when the iron is exchanged the first (FeMOR and FeAgMOR), the mentioned ratio changes significantly: the amount of weak acid sites decreases, and the amount of strong acid sites increases (16:84 and 19:81 for FeMOR and FeAgMOR, respectively). When iron is deposited in the last stage of ion exchange in AgFeMOR, despite maintaining the volume exchange ratio, the total acidity increases significantly due to the presence of protons (from the precursor solution acidified with H 2 SO 4 ). These protons can replace cationic sites, as suggested by ICP-OES data ( Table 1) and thermodynamic parameters ( Table 2). Thus, the order of deposition of the components affects the surface acidity, which, in turn, affects the catalytic properties.
To estimate the ratio of Fe 2+ and Fe 3+ in as-prepared samples, FTIR measurements of adsorbed NO have been performed; the results are shown in Figure 1. Adsorption of NO on the initial mordenite and on AgMOR is negligible, while all Fe-containing samples have shown intense absorption bands corresponding to the nitrosyl complexes of Fe: Fe 3+ (NO) at 1883 cm −1 and Fe 2+ (NO) n , where n = 1 at 1852 cm −1 , n = 2 at 1766 cm −1 , n = 3 at 1813 and 1919 cm −1 [59]. All Fe-containing samples have the presence of both nitrosyl complexes, based on Fe 2+ and Fe 3+ . As mentioned above, the Fe content in FeAgMOR is lower than in the other two Fe-containing samples; accordingly, the intensity of absorption of NO on FeAgMOR is also approximately two times lower.
Catalysts 2020, 10, x FOR PEER REVIEW 6 of 17  NaMOR  32  68  1359  AgMOR  33  67  1662  FeMOR  16  84  2970  AgFeMOR  29  71  2264  FeAgMOR  19  81  1251 When iron is deposited in the last stage of ion exchange in AgFeMOR, despite maintaining the volume exchange ratio, the total acidity increases significantly due to the presence of protons (from the precursor solution acidified with H2SO4). These protons can replace cationic sites, as suggested by ICP-OES data ( Table 1) and thermodynamic parameters ( Table 2). Thus, the order of deposition of the components affects the surface acidity, which, in turn, affects the catalytic properties.

Sample Amount of the Acid Sites, % Total Acidity, mmol/g l-peak h-peak
To estimate the ratio of Fe 2+ and Fe 3+ in as-prepared samples, FTIR measurements of adsorbed NO have been performed; the results are shown in Figure 1. Adsorption of NO on the initial mordenite and on AgMOR is negligible, while all Fe-containing samples have shown intense absorption bands corresponding to the nitrosyl complexes of Fe: Fe 3+ (NO) at 1883 cm −1 and Fe 2+ (NO)n, where n = 1 at 1852 cm −1 , n = 2 at 1766 cm −1 , n = 3 at 1813 and 1919 cm −1 [59]. All Fe-containing samples have the presence of both nitrosyl complexes, based on Fe 2+ and Fe 3+ . As mentioned above, the Fe content in FeAgMOR is lower than in the other two Fe-containing samples; accordingly, the intensity of absorption of NO on FeAgMOR is also approximately two times lower. As seen from Figure 1c-e, bands of nitrosyl complexes Fe 2+ (NO)n with n = 1 and 2 appea Then, with an increase of NO pressure, both of them transform into the adsorption corresponding to Fe 2+ (NO)n with n = 3. Complex adsorption band Fe 3+ (NO) shows a maxim intensity at an intermediate level of pressure, and then with increasing pressure, the in decreases. For AgFeMOR, the Fe 2+ (NO)3 adsorption band increases significantly with increasi pressure, while the intensity of the Fe 3+ (NO) band remains virtually unchanged. The intensitie two mentioned bands for FeMOR and FeAgMOR change more or less the same with the incr pressure. It could mean that the adsorption sites on the surface of AgFeMOR are different from on FeMOR and FeAgMOR.
The FTIR spectra of NO are measured after thermal treatment under vacuum, which is as a mildly reducing activation condition for Fe-zeolites. On all Fe-containing samples, NO the co-existence of Fe 2+ and Fe 3+ . This does not allow us to quantify the amount of Fe 2+ and Fe before thermal treatment but gives an indication of the effect of Ag on the redox properties o Since only one band at 1883 cm −1 is due to the Fe 3+ state, by comparing its intensity for FeMO AgFeMOR samples with similar content of Fe, we can conclude that the Fe 3+ -content in AgFeM higher than in FeMOR. This can be explained by the redox reaction between Ag + and Fe 2+ , as a of which Ag 0 and Fe 3+ are formed, and which is present only in bimetallic samples. Besid interaction of NO is stronger for Fe-containing samples (compared to AgMOR and NaMOR), allows one to expect higher catalytic activity in the de-NOx reaction.
The XANES and EXAFS methods are applied to shed light on the mechanism of formation from the point of view of the immediate atomic environment of iron. Fe K-edge X spectra (Figure 2a) for all Fe-containing samples are similar to the reference α-Fe2O3, indicati Fe is mostly present in the Fe 3+ oxidized state. However, compared to the reference, the inten oscillations is reduced, especially for FeMOR, which means that the structure is disordered possibly, another oxidized state of Fe (Fe 2+ ) is present. EXAFS Fourier transforms (Figure 2b shown that the spectra of FeMOR are closer to the FeO reference, while the shapes of EXAFS F transforms for both bimetallic samples remain partially similar to those of the α-Fe2O3 refere could be explained by the redox reaction of Fe 2+ with Ag + , leading to the formation of Fe 3+ [5 the presence of iron in both oxidation states simultaneously. In the case of AgFeMOR, the fr present in the solution undergoes a redox reaction with Ag + MOR, and consequently, the re Ag 0 diffuses from the mordenite channels, agglomerating into nanoparticles in the mesopores the surface of crystals. Negatively charged vacancies are compensated by Fe 3+ ions. In the FeAgMOR synthesis, free Ag + , which is present in excess in the exchange solution, undergoes a The FTIR spectra of NO are measured after thermal treatment under vacuum, which is known as a mildly reducing activation condition for Fe-zeolites. On all Fe-containing samples, NO reveals the co-existence of Fe 2+ and Fe 3+ . This does not allow us to quantify the amount of Fe 2+ and Fe 3+ ions before thermal treatment but gives an indication of the effect of Ag on the redox properties of iron. Since only one band at 1883 cm −1 is due to the Fe 3+ state, by comparing its intensity for FeMOR and AgFeMOR samples with similar content of Fe, we can conclude that the Fe 3+ -content in AgFeMOR is higher than in FeMOR. This can be explained by the redox reaction between Ag + and Fe 2+ , as a result of which Ag 0 and Fe 3+ are formed, and which is present only in bimetallic samples. Besides, the interaction of NO is stronger for Fe-containing samples (compared to AgMOR and NaMOR), which allows one to expect higher catalytic activity in the de-NO x reaction.
The XANES and EXAFS methods are applied to shed light on the mechanism of species formation from the point of view of the immediate atomic environment of iron. Fe K-edge XANES spectra (Figure 2a) for all Fe-containing samples are similar to the reference α-Fe 2 O 3 , indicating that Fe is mostly present in the Fe 3+ oxidized state. However, compared to the reference, the intensity of oscillations is reduced, especially for FeMOR, which means that the structure is disordered, and, possibly, another oxidized state of Fe (Fe 2+ ) is present. EXAFS Fourier transforms (Figure 2b) have shown that the spectra of FeMOR are closer to the FeO reference, while the shapes of EXAFS Fourier transforms for both bimetallic samples remain partially similar to those of the α-Fe 2 O 3 reference. It could be explained by the redox reaction of Fe 2+ with Ag + , leading to the formation of Fe 3+ [50] and the presence of iron in both oxidation states simultaneously. In the case of AgFeMOR, the free Fe 2+ present in the solution undergoes a redox reaction with Ag + MOR, and consequently, the resulting Ag 0 diffuses from the mordenite channels, agglomerating into nanoparticles in the mesopores and on the surface of crystals. Negatively charged vacancies are compensated by Fe 3+ ions. In the case of FeAgMOR synthesis, free Ag + , which is present in excess in the exchange solution, undergoes a redox reaction with Fe 2+ MOR to form Fe 3+ MOR, and, as a result, Ag 0 is leached out from the mordenite channels, while other Ag + ions from the solution balance the negative charge of the framework. 4+2). Moreover, for iron in Fe-containing samples, the maxima, which corresponds to the outer coordination spheres (Fe-Fe, as it is shown in Figure 2a), is either weak or absent. Therefore, based on the EXAFS data, we may conclude that Fe in FeAgMOR and AgFeMOR is partially included in the zeolite framework as Fe 3+ instead of Al 3+ . Previously, the mechanism of the isomorphous substitution of Al 3+ in the mordenite framework by tetrahedrally coordinated Fe 3+ , which is promoted by the presence of Ag + ions, has been proposed and confirmed by Mössbauer spectroscopy, and it has also been shown that the presence and concentration of the framework Fe 3+ ion in various samples correlate with their catalytic activity in de-NOx [33]. In the FeMOR sample, the coordination numbers are even lower, 0.6+2, which, along with a decrease in the intensity of XANES oscillations, indicates a disordered or mixed state of Fe, which is different from the states of Fe in bimetallic samples. Taking into account the reduced values of coordination numbers, we assume that this state may include other oxidized states of Fe (i.e., Fe 2+ ). Similar to XANES, Fe K-edge EXAFS (Figure 2b) demonstrates that iron in MOR is oxidized. The main maximum of EXAFS Fourier transforms for all MOR samples is very similar to Fe-O maximum for oxide references and, therefore, corresponds to the oxygen surroundings of Fe. However, this Fe-O maximum tends to be asymmetric. To account for this asymmetry, the range of Fourier transforms from 1.2 to 2.1 Å is modeled with two different Fe-O distances; the obtained parameters are shown in Table 5. The fit results suggest that in FeAgMOR and AgFeMOR catalysts, the coordination of Fe-O is similar; for both samples, the coordination numbers are 2+2 (two O atoms each, at two different distances), with Fe-O distances of~1.9 and~2.1 Å. Despite the fact that XANES data are similar to the reference α-Fe 2 O 3 for these samples, the coordination of Fe in α-Fe 2 O 3 is different (octahedral, i.e., 4+2). Moreover, for iron in Fe-containing samples, the maxima, which corresponds to the outer coordination spheres (Fe-Fe, as it is shown in Figure 2a), is either weak or absent. Therefore, based on the EXAFS data, we may conclude that Fe in FeAgMOR and AgFeMOR is partially included in the zeolite framework as Fe 3+ instead of Al 3+ . Previously, the mechanism of the isomorphous substitution of Al 3+ in the mordenite framework by tetrahedrally coordinated Fe 3+ , which is promoted by the presence of Ag + ions, has been proposed and confirmed by Mössbauer spectroscopy, and it has also been shown that the presence and concentration of the framework Fe 3+ ion in various samples correlate with their catalytic activity in de-NO x [33]. The samples are tested in the de-NO x reaction in the presence of a CO/C 3 H 6 mixture, after oxidation in a flow of a mixture of oxidants (4.4 vol.% of O 2 /N 2 balanced) at a heating rate of 5 • C/min from room temperature up to 550 • C. The catalytic evaluation results are presented in Figure 3. Mordenite itself has shown catalytic activity with a maximum at 312 • C that is attributed to the active sites of the Na + cation that acts as a Lewis acid site, while the oxygen framework with partial negative charge acts as a Lewis base. The AgMOR catalyst presents low-temperature activity (120-150 • C), which is associated with the Ag + , and two peaks at 260 and 377 • C. The latter has shown an increased conversion at temperatures in the range of 320-520 • C, which could be due to the reduction of Ag + →Ag 0 under the reaction conditions. According to the de-NOx, XANES, and EXAFS data, and our previous results [33], the de-NOx activity can be provided by two types of active centers: Fe 3+ compensating cation at the vacant negatively charged site of mordenite, or Fe 3+ in the mordenite framework instead of Al 3+ . Due to the complex correlation of activity with temperature, until now, it has not been possible to determine a Since all samples are subjected to preliminary oxidative treatment, the observed activity should be attributed to the oxidized states of the active components (Fe 3+ and Ag + ) present in the catalyst. However, as it has been confirmed by EXAFS (see above) for AgFeMOR and FeAgMOR catalysts, the presence of Fe 3+ species is possible in two forms: as a compensation cation on a vacant negatively charged site of mordenite or in the mordenite framework instead of Al 3+ . The FeMOR catalyst shows 100% conversion of NO in the range 290-380 • C, and it contains only Fe 3+ in the ion-exchange sites of mordenite, which are active centers for the de-NO x .
AgFeMOR shows at least two types of active centers with pronounced maxima at 312 and 402 • C. Those centers, which are active at 312 • C, may be related to Fe 3+ species interacting with NaMOR since FeMOR shows activity at the same temperature. The active center at about 402 • C can be related to Ag 0 centers, similar to those of the AgMOR sample; however, the activity of these centers on the AgFeMOR is much higher than the ones on the monometallic AgMOR. This can be explained by the activity of Ag 0 formed by the redox interaction of Ag + with Fe 2+ during preparation and by the reduction of Ag + residue under the reaction conditions. Two types of active centers represent the operating temperature window for the conversion of NO over FeAgMOR: the peaks at 312 and 463 • C correspond to Fe and Ag species, respectively. Since the Fe content in FeAgMOR is half that in FeMOR and AgFeMOR in accordance with the ICP-OES results, the FeAgMOR activity at 312 • C is lower. Activity after 350 • C of FeAgMOR has a complex temperature correlation, most likely due to the co-existence of different active centers.
According to the de-NO x , XANES, and EXAFS data, and our previous results [33], the de-NO x activity can be provided by two types of active centers: Fe 3+ compensating cation at the vacant negatively charged site of mordenite, or Fe 3+ in the mordenite framework instead of Al 3+ . Due to the complex correlation of activity with temperature, until now, it has not been possible to determine a reliable difference between them. The most promising catalysts for the de-NO x reaction in the presence of CO/C 3 H 6 are FeMOR and AgFeMOR. These two samples also have the highest total acidity. The activity profiles of the two mentioned samples are close to each other, which allows us to conclude that silver does not improve the catalytic properties of Fe by itself. However, Ag provides the effect of modifying the catalytic activity of Fe at high temperatures (after 350 • C) by changing the structure of the active centers of iron.
Thus, the experimental results confirm the differences in the mechanism of formation of diverse active species with different orders of cation deposition, proposed on the basis of theoretical parameters (see Table 2). According to EXAFS and NO adsorbed FTIR data for a monometallic FeMOR sample, it has been shown that the most common species are in the Fe 2+ state (Scheme 1, Step 1A). Treatment of FeMOR with an Ag + solution results in the replacement of Na + by Ag + (Scheme 1, Step 2A). Contact between Ag + and Fe 2+ leads to a redox interaction, resulting in the formation of Ag 0 and Fe 3+ MOR. A neutrally charged silver atom is unable to displace Fe 3+ from its position. Thus, the FeAgMOR is expected to have a mixture of Ag 0 and Ag+ species, while Fe 3+ is the predominant state of iron. Moreover, for bimetallic samples, it has been confirmed by EXAFS that Fe 3+ is partially incorporated into the framework of the mordenite.
When AgMOR is treated with Fe 2+ solution containing H + protons, residual Na + can be exchanged with both H + and Fe 2+ (Scheme 1, Step 2B). However, the redox interaction between AgMOR and Fe 2+ leads to the fact that the reduced neutral Ag 0 leaves the negatively charged site. Simultaneously, the vacant negatively charged site [-]MOR can be occupied by oxidized Fe 3+ , or either Fe 2+ or H + from solution. Thus, in the case of AgFeMOR, it is expected that it will contain a large number of Ag 0 -species and a mixture of Fe 2+ and Fe 3+ , plus Fe 3+ embedded into the framework of the mordenite, which is confirmed by the EXAFS data. The results obtained allow us to outline the next step of this study, which will be the study of the formation and state of silver in these samples, which will additionally confirm the presented mechanism. The received data will be published elsewhere. Catalysts 2020, 10, x FOR PEER REVIEW 11 of 17 Step 0. Original NaMOR zeolite. The Si atoms in NaMOR are omitted for clarity of mechanism (highlighted in navy blue).

Scheme 1.
Mechanism of the formation of the studied samples during the ion exchange process with a different order of incorporation.
When AgMOR is treated with Fe 2+ solution containing H + protons, residual Na + can be exchanged with both H + and Fe 2+ (Scheme 1, Step 2B). However, the redox interaction between AgMOR and Fe 2+ leads to the fact that the reduced neutral Ag 0 leaves the negatively charged site. Simultaneously, the vacant negatively charged site [-]MOR can be occupied by oxidized Fe 3+ , or either Fe 2+ or H + from solution. Thus, in the case of AgFeMOR, it is expected that it will contain a large number of Ag 0 -species and a mixture of Fe 2+ and Fe 3+ , plus Fe 3+ embedded into the framework of the mordenite, which is confirmed by the EXAFS data. The results obtained allow us to outline the next Scheme 1. Mechanism of the formation of the studied samples during the ion exchange process with a different order of incorporation.

Samples Preparation
The sodium form of mordenite (NaMOR) with an atomic ratio Si/Al = 6.5 was supplied by Zeolyst Int. (Product CBV10A, Conshohocken, PA, USA) and was used without any pretreatment/purification. Metals (Ag, Fe) were introduced into the starting mordenite by the conventional ion exchange method from solutions for 24 h at 60 • C, in a ratio of 1 g of NaMOR per 35 mL of the precursor solution.
The precursor of Ag was an aqueous solution of AgNO 3 (0.03 N). The Fe precursor was an aqueous solution of FeSO 4 (0.03 N), stabilized from hydrolysis by the addition of H 2 SO 4 until pH = 2 was adjusted; this value was chosen in accordance with the data [60]. After each ion exchange procedure, the samples were filtered, washed, and dried at 110 • C for 20 h. For the preparation of bimetallic samples, the obtained monometallic samples were processed under similar conditions in appropriate solutions. All ion exchange steps were carried out under conditions that prevent direct light from entering the samples and reducing the illumination time to a minimum in order to avoid the potential reduction of silver ions under the action of light.
For each sample preparation, 2 g of NaMOR was processed using solutions containing the same total metal equivalent concentration (normality) of each of both metals. The preparation of monometallic samples included one ion exchange step of NaMOR in an appropriate solution. Preparation of bimetallic samples included two steps of ion-exchange; labels of the bimetallic samples reflected the order of the cations incorporation during the ion exchange: AgFeMOR (first Ag + , then Fe 2+ ) and FeAgMOR (first Fe 2+ , then Ag + ).

Characterization Methods
Quantitative chemical analysis of the samples was carried out by Inductively Coupled Plasma-Optical Emission Spectroscopy (ICP-OES) method by means of a VARIAN VISTA-MPX CCD SIMULTANEOUS spectrometer (Varian Inc., Palo Alto, CA, USA). The samples were pretreated by degassing, followed by dissolution in a mixture of HNO 3 and HF at 40 • C overnight, and adding an H 3 BO 3 solution, followed by further keeping at 40 • C for 5 h.
Textural properties were determined from nitrogen adsorption-desorption isotherms at −196 • C, recorded with a Micromeritics TriStar 3000 apparatus (Micromeritics Instruments Corp., Norcross, GA, USA). Prior to experiments, the samples were degassed at 300 • C in a vacuum for 5 h. The volume of the adsorbed N 2 was normalized to a standard temperature and pressure. The specific surface area (S BET ) of the samples was calculated by applying the Brunauer-Emmett-Teller (BET) method to the nitrogen adsorption data within the P/P 0 range of 0.005-0.250. The average pore diameter was calculated by applying the Barret-Joyner-Halenda (BJH) method to the adsorption and desorption branches of the N 2 isotherms. The cumulative pore volume was obtained from the isotherms at P/P 0 = 0.99.
Surface acidity was measured in a fixed bed continuous flow quartz micro-reactor, using approximately 50 mg of the sample over a temperature range of 100-550 • C at a heating rate ramp of 10 • C/min. The acidity of the samples was measured by the analysis of TCD-TPD (Thermal Conductivity Detector) with a Quantachrome Chem BET Pulsar TPD/TPD (Quantachrome, Boynton Beach, FL, USA). First, in order to remove moisture, the samples were pre-treated in He (99.99% purity, 120 mL/min, temperature range 100-550 • C, a heating rate of 10 • C/min). The sample was then left at 550 • C for 1 h and then cooled down to 100 • C. For the ammonia chemisorption probe, anhydrous ammonia with a purity of 99.98% Aldrich was used at a flow rate of 120 mL/min for 10 min at 100 • C. After that, adsorbed molecules were removed by applying a flow of He for 40 min.
Fe K-edge EXAFS spectra were measured in transmission mode at the Structural Materials Science end-station of the Kurchatov Synchrotron Radiation source [61]. To scan the energy, Si (111) channel-cut monochromator (Shubnikov Institute of Crystallography RAS, Moscow, Russia) was used. The X-ray beam intensities before and after the sample were measured with ionization chambers filled with air and Ar for Fe K-edge at atmospheric pressure. The primary processing of the XAFS spectra was carried out in terms of the IFEFFIT software package (version 0.9.26) [62,63]. Normalized EXAFS oscillations were k 3 -weighted and analyzed in the 2÷12 Å −1 k-range in the case of Fe K-edge spectra.
Fourier transformed infrared (FTIR) spectra of adsorbed NO were carried out in situ on an FTIR-66 spectrometer (Bruker, Manning Park Billerica, MA, USA) equipped with a cryogenic MCT (Mercury-Cadmium-Telluride) detector in transmittance mode with a resolution of 4 cm −1 over 64 total scans per spectrum. Samples in the form of thin self-supporting pellets (5-10 mg/cm 2 ) were directly treated at temperatures from room temperature to 500 • C and pressures from 10 −2 to 760 Torr in a quartz cell equipped with KBr windows, designed for room temperature studies in a controlled atmosphere. First, the samples were degassed under a dynamic vacuum of 10 −3 mBar at 400 • C for 1 h, then cooled down to room temperature and then evacuated. The spectra were measured by dosing NO (70 mBar), and subsequently reducing the NO pressure by stepwise expansions in the vacuum line, followed by gradual opening to dynamic vacuum (final pressure 5.0 × 10 −4 mBar). The spectra reported after normalization with respect to the intensity of the overtone framework modes at 1872 cm −1 .
Catalytic activity was measured in a fixed bed continuous flow quartz micro-reactor inside a vertical furnace, using 10 mg of the catalyst. Before the catalytic reaction, the catalysts were pretreated in a flow of an oxidant mixture (70 mL/min of 4.4 vol.% of O 2 /N 2 balanced) at a heating rate of 5 • C/min from room temperature up to 550 • C. The catalytic activity of the samples was measured for NO reduction in the presence of CO, propene, and oxygen. Catalytic runs were performed within the temperature range of 25-500 • C at a rate of 10 • C/min. The reaction gas phase mixture consisted of NO (650 ppm), C 3 H 6 (550 ppm), O 2 (2 vol.%), CO (0.4 vol.%) in N 2 balance and a total gas flow of 100 mL min -1 . This gas mixture emulated the vehicle exhaust gases from internal combustion engines. Inlet and outlet gas concentrations were measured in a ZAI gas analyzer. Data acquisition of NO concentration during the catalytic reaction was obtained continuously. Conversion of NO was calculated from the difference of the reactor inlet and outlet NO concentration, divided by the inlet NO concentration.

Conclusions
Monometallic and bimetallic Ag-Fe-based materials deposited on a mordenite carrier were prepared by the ion exchange method, applying different orders of metal deposition. It is shown that ion-exchange treatment very slightly changes the mordenite stoichiometry and textural properties. The second step of ion exchange may wash out some amount of the cations of the first step of ion-exchanged.
The results of Fe K-edge XANES oscillations showed that in FeMOR, both Fe 3+ and Fe 2+ were in a disordered state. In FeAgMOR, the prevailing state was Fe 3+ , while in AgFeMOR, the iron states were intermediate, or mixed between FeMOR and FeAgMOR catalysts. The results of the Fe K-edge EXAFS study revealed the formation of a disordered state in which the first coordination sphere was asymmetric and was characterized by two different Fe-O distances. In FeAgMOR and AgFeMOR, the coordination of Fe-O was similar. Therefore, based on the EXAFS data, we may conclude that Fe is incorporated in FeAgMOR and AgFeMOR into the zeolite framework as Fe 3+ instead of Al 3+ , and in accordance with our knowledge, this is a new type of active center in the de-NO x reaction.
The catalytic activity results for the de-NO x reaction in the presence of CO/C 3 H 6 showed that the most active samples were the Fe-containing ones. The mordenite carrier and the monometallic silver sample AgMOR did not show high activity in the studied reaction. The promoter effect of Ag on iron-containing catalysts took place at the high-temperature activity.
Based on the thermodynamic analysis and experimental data, a mechanism for the formation of active centers at two steps of ion-exchange synthesis with a different order of deposition of components was proposed and confirmed.