Elucidating the reaction mechanism of SO2 with Cu-CHA catalysts for NH3-SCR by X-ray absorption spectroscopy

The application of Cu-CHA catalysts for the selective catalytic reduction of NOx by ammonia (NH3-SCR) in exhaust systems of diesel vehicles requires the use of fuel with low sulfur content, because the Cu-CHA catalysts are poisoned by higher concentrations of SO2. Understanding the mechanism of the interaction between the Cu-CHA catalyst and SO2 is crucial for elucidating the SO2 poisoning and development of efficient catalysts for SCR reactions. Earlier we have shown that SO2 reacts with the [Cu2II(NH3)4O2]2+ complex that is formed in the pores of Cu-CHA upon activation of O2 in the NH3-SCR cycle. In order to determine the products of this reaction, we use X-ray absorption spectroscopy (XAS) at the Cu K-edge and S K-edge, and X-ray emission spectroscopy (XES) for Cu-CHA catalysts with 0.8 wt% Cu and 3.2 wt% Cu loadings. We find that the mechanism for SO2 uptake is similar for catalysts with low and high Cu content. We show that the SO2 uptake proceeds via an oxidation of SO2 by the [Cu2II(NH3)4O2]2+ complex, resulting in the formation of different CuI species, which do not react with SO2, and a sulfated CuII complex that is accumulated in the pores of the zeolite. The increase of the SO2 uptake upon addition of oxygen to the SO2-containing feed, evidenced by X-ray adsorbate quantification (XAQ) and temperature-programmed desorption of SO2, is explained by the re-oxidation of the CuI species into the [Cu2II(NH3)4O2]2+ complexes, which makes them available for reaction with SO2.


Introduction
The current technology to reduce the harmful NO x emissions from diesel-powered vehicles is based on the selective catalytic reduction of nitrogen oxides (NO x ) to N 2 and H 2 O by ammonia (NH 3 -SCR). 1,2Cu-exchanged chabazite zeolites (Cu-CHA) are preferred catalysts in diesel exhaust systems, due to their high activity in the low-temperature region (150-350 °C) and hydrothermal stability above 500 °C.The low-temperature activity of Cu-CHA-based catalysts, however, is strongly reduced in the presence of SO 2 , and therefore, application of such catalysts in exhaust systems requires the use of ultra-low sulfur diesel fuel. 3,4e mechanism of the NH 3 -SCR reaction in Cu-CHA based catalysts is a redox cycle, 5,6 in which the oxidation state of Cu changes between Cu I and Cu II .The reaction proceeds via a number of Cu-complexes formed by adsorption and reaction of NO, NH 3 and O 2 as ligands on the Cu-ions inside the CHA zeolite. 7The oxidation from Cu I to Cu II occurs by activation of O 2 , which is a crucial step in the reaction cycle.2][13][14][15]  (NH 3 ) 4 O 2 ] 2+ complexes, thus limiting the mobility of the [Cu I (NH 3 ) 2 ] + species, and the Cupair formation necessary for O 2 activation. 17][20] Even though the formation of ammonium bisulfate can explain certain aspects of the deactivation by SO 2 , other observations point towards the formation of species that contain both S and Cu.Indeed, the uptake of SO 2 is oen saturated at S/ Cu ratios below 1, 4,19 which suggests that the uptake of SO 2 is limited by the amount of Cu in the catalyst.This implies a direct interaction between SO 2 and Cu, such that a further reaction with SO 2 is not possible.Furthermore, the release of SO 2 from a Cu-CHA catalyst exposed to SO 2 occurs at a slightly higher temperature as compared to a Cu-CHA catalyst with ammonium bisulfate deposited on it via impregnation. 16This indicates that a sulfate-or sulte-like compound is formed, that is more stable than ammonium bisulfate, which would be consistent with a direct interaction of SO 2 with Cu.Indeed, direct interaction of SO 2 with Cu was reported to result in the formation of Cusulfate 19,20 and Cu-bisulfate, 18 or similar species.Finally, Cu-CHA catalysts typically show 10-20 times lower lowtemperature activity for NH 3 -SCR when saturated with SO 2 , 3,4,20,21 but never a complete deactivation.If deactivation were caused by accumulation of ammonium sulfate, it would be expected to be complete upon saturation with SO 2 , at least for temperatures up to the onset of the decomposition of ammonium bisulfate.
Since the rst step of the SO 2 uptake is a reaction with the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complexes, the uptake of SO 2 may be affected by the Cu content in the catalyst and the gas atmosphere.The propensity toward the formation of the [Cu II 2 (NH 3 ) 4 O 2 ] 2+ complex is determined by the Cu content and partial pressure of oxygen. 4,13,22Furthermore, ammonia is required as well, in order to form the mobile [Cu I (NH 3 ) 2 ] + and the [Cu II 2 (NH 3 ) 4 O 2 ] 2+ complexes.The effect of ammonia is clearly demonstrated by the observation that the Cu II species formed upon exposure of Cu-CHA to O 2 at 500 °C is much less reactive towards SO 2 than the same species exposed to NH 3 . 16Therefore, the impact of SO 2 on the activity will depend on the reaction conditions. 4,13,22n this work, we use in situ X-ray absorption spectroscopy (XAS) to monitor the reaction of SO   2+ complexes have been formed, the sample is exposed to either 400 ppm SO 2 for 3 hours, or to a mixture of 360 ppm SO 2 and 10% O 2 at 200 °C, while monitoring the changes in the XAS spectra.All the experiments were performed under ow conditions with He as carrier gas.Total ow was 100 ml min −1 (for stages without SO 2 ) or 50 ml min −1 (for stages with SO 2 ).
The whole procedure was followed by in situ Cu K-edge XAS at the BM23 beamline of the European Synchrotron Radiation Facility (ESRF, Grenoble, France). 24Exposure of the [Cu 2 -II (NH 3 ) 4 O 2 ] 2+ dimer to 400 ppm SO 2 was also monitored by in situ S K-edge HERFD XANES spectroscopy during a separate experiment at the ID26 beamline of the ESRF. 25S K-alpha XES spectra were also recorded at ID26 for the stationary points of the treatment protocol.
The evolution of S content in the catalyst during the reaction with SO 2 was evaluated by in situ XAQ 23 measurements.Total SO 2 uptake was also independently determined by temperature programmed desorption of SO 2 (SO 2 -TPD).
Experimental procedures are reported in more details in the ESI.†

Exposure to SO 2 without O 2
To determine the effect of Cu content on the reaction of SO 2 with the Cu-CHA, we compared the Cu K-edge spectra for the low-Cu catalyst aer exposure to SO 2 using the same pretreatment protocols to form specic different Cu I and Cu II species earlier reported for the high-Cu catalyst. 16We observed the same general trend, which means that the [Cu 2 complex is the most sensitive to SO 2 for the low-Cu catalyst as well.The detailed results of these measurements are reported in Fig. S4  To develop a better understanding of the reaction of SO 2 with the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex, we identify the reaction intermediates and reaction products by applying a combination of multivariate curve resolution alternating least squares method (MCR-ALS) 26,27 and linear combination tting (LCF).This hybrid approach consists of three stages.First, we apply MCR-ALS to the ensemble of the experimental spectra to deduce the shape of principal components.Then, the MCR components that can be readily associated with the known Cu species whose experimental spectra are available are substituted by the experimental Cu K-edge spectra of these species.Finally, LCF is performed over the same experimental dataset, using the experimental spectra selected at the previous step and the remaining MCR components.It allows at the same time to minimize the spectral artefacts induced by the MCR algorithm, and get a reasonable estimate of the spectra for the species that cannot be readily identied.
In the reported procedures, it was possible to select as references the experimental spectra for framework-bound Cu II (fw-Cu II ), for the [Cu I (NH 3 ) 2 ] + complex, and for the [Cu 2 -II (NH 3 ) 4 O 2 ] 2+ complex.Conversely, Cu I directly bound to the zeolite framework (fw-Cu I ) was represented by calculated MCR component.The second MCR component that was used in the LCF was designated "sulfated component", since it appeared only aer the samples were exposed to SO 2 -containing mixtures.for both the low-Cu and the high-Cu catalysts, the fw-Cu II component appears at the nal stage of the process.

Alternating exposure to SO 2 and O 2
To discern the individual effects of SO 2 and O 2 , a third experiment was conducted for both the low-Cu and high-Cu samples, wherein the dimers were exposed to alternating switches between SO 2 and O 2 ; the results of the LCF are shown in Fig. 6 component forms aer the rst exposure of the sulfated sample to oxygen and, aer subsequent exposure to SO 2 , stabilizes at the same level as aer exposure of the high-Cu catalyst to SO 2 + O 2 (Fig. 5).At each subsequent exposure to SO 2 , less and less Cu I is formed, concomitantly with a decreasing amount of the "reactive" [Cu 2 II (NH 3 ) 4 O 2 ] 2+ component.

Uptake of SO 2 monitored by XAQ and SO 2 -TPD
The presence of oxygen during the SO 2 exposure not only affects the nal oxidation state of the Cu, but also leads to an increased uptake of SO 2 .This increased amount of SO 2 in the catalyst can be quantied by X-ray adsorbate quantication (XAQ) 23 and SO 2 -TPD.
The XAQ technique relies on the phenomenon that the total absorption of X-rays depends on the composition of the sample.Therefore, the increase in total X-ray absorption during SO 2 exposure reects the increase in the amount of SO 2 in the sample.This quantitative information was used to calculate the S/Cu ratios in situ during the exposure of the samples to the gas mixtures containing SO 2 .Fig. 7 shows the measured XAQ signals for the low Cu and high Cu catalysts during exposure to SO 2 and SO 2 + O 2 .All S/Cu curves have similar shape, exhibiting rather fast changes in the rst 30-40 minutes of the measurement, followed by a much slower growth at a later stage.The nal S/Cu levels reached in the presented experiments are 0.32 for the low-Cu sample and 0.22 for the high-Cu sample.In the presence of O 2 , the nal levels increase to S/Cu = 1.04 and 0.63 for the low-Cu and high-Cu catalysts, respectively.
The second method that was employed to measure SO 2 uptake is SO 2 -TPD, where the desorption of SO 2 is recorded as a function of temperature during heating of the catalyst.In contrast to XAQ, where the amount of SO 2 adsorbed on the catalyst is monitored in situ, SO 2 -TPD measures the amount of SO 2 released from a saturated sample.Fig. 8 shows the SO 2 -TPD data for the two catalysts aer exposure to SO 2 and SO 2 + O 2 ; the SO 2 -TPD of a Cu-free CHA impregnated with (NH 4 ) 2 SO 4 is included for comparison.All desorption curves for the Cu-CHA catalysts show a desorption feature in the range 400-600 °C, and one in the range 750-1000 °C.Because these proles are clearly different from that of the adsorbed (NH 4 ) 2 SO 4 on the Cufree zeolite reference, we conclude that the SO 2 in the Cu-CHA catalysts predominantly interacts with the Cu, without a significant amount of free (NH 4 ) 2 SO 4 .This is in good agreement with the conclusion that SO   presence of O 2 leads to a larger amount of SO 2 in the Cu-CHA catalysts, and that SO 2 binds predominantly to the Cu ions in the zeolite, in agreement with our earlier results. 16AQ and TPD show that exposure to SO 2 + O 2 leads to a greater sulfur uptake compared to the exposure to only SO 2 (Fig. 9).Importantly, the S/Cu molar ratio is in good correspondence with the concentration of the sulfated component in the XANES LCF, which conrms the assignment of the latter mainly to sulfated species.

Sulfur K-edge XANES and Ka XES
To resolve the oxidation state of sulfur and the conguration of sulfur species in the sulfated Cu-CHA catalyst we measured S Kedge XANES during the exposure of the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex to SO 2 and Ka-XES spectra at the end of the exposure.coordinated to the Cu in the structure of the [Cu II (NH 3 ) 4 ]SO 4 , and the Cu-S distance is about 4.6 Å, it is not possible to determine the precise location of the S atom in the sulfated species in the zeolite based on XANES data alone.
To locate the S atoms in the sulfated species, we extracted the EXAFS part of the sulfated component by subtracting the weighted reference spectrum for fw-Cu II from the spectrum aer exposure of the low-Cu catalyst to SO 2 and O 2 (see Fig. 4  and 5), using the weight coefficients derived from the LCF.We then tted the EXAFS spectrum of the sulfated component with two N and two O in the rst shell and S in the second shell.The tting results are reported in Table 1.The resulting t in comparison with the EXAFS of the sulfated component with a highlighted contribution of the Cu-S path is presented in Fig. 12. From this analysis, we nd a Cu-S distance of 2.58 Å, which is much shorter than in the Cu(NH 3 ) 4 SO 4 $H 2 O reference compound (4.65 Å).This means that the structure of the sulfated species is different from Cu(NH 3 ) 4 SO 4 $H 2 O.

Discussion
The unreactive part of the [Cu 2 For the high-Cu catalyst, the concentration of the component assigned to [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complexes does not go to zero during repeated exposures to SO 2 followed by exposure to O 2 , remaining above 18% aer each exposure to SO 2 (see Fig. 6).The rst explanation for such behavior may be that a fraction of the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complexes become inaccessible for SO 2 and therefore do not react.This interpretation resembles the mechanism for deactivation of Cu-CHA catalysts as proposed by Bjerregaard et al. 17 A further consequence of this interpretation is that it limits the uptake of SO 2 in a Cu-CHA catalyst with a sufficiently high Cu content (ca. 3 wt%), while, at the same time, a part of the Cu does not react with SO 2 .Such a conclusion is in good agreement with earlier measurements of a limited uptake of SO 2 in Cu-CHA catalysts, 3,4 and the observation that Cu-CHA catalysts show a residual activity aer saturation with SO 2 . 3With such an interpretation, the structure of the [Cu 2 -II (NH 3 ) 4 O 2 ] 2+ complex remains unchanged, which is directly reected in the XANES spectra.The second explanation for the persistence of the component assigned to the  3 is not yet available and the temperature is too high.As such moieties are essentially fw-Cu II species, we expect that their reactivity towards SO 2 is very low, as we have shown in our previous work, 16 thus resulting in an accumulation of these species in the sample.

Unraveling the mechanism of the sulfation reaction
The results presented above have some implications for a mechanism describing the reaction of SO 2 with Cu-CHA catalysts.Sulfur XES and XANES show that S is stored in the sample as S 6+ within (SO 4 )   Bringing these considerations together, we arrive at a reaction pathway consisting of two steps.In the rst step (eqn (1)), SO 2 reacts with a [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex, the peroxo bond breaks, a mobile SO a X intermediate forms, and Cu II reduces to Cu I .The Cu I appears in the form of [Cu I (NH 3 ) 2 ] + and fw-Cu I , as indicated by the LCF (Fig. 3).A possible candidate for the mobile intermediate SO a X is SO 3 , formed by oxidation of SO 2 , upon the formation of Cu I -species.It is known that SO 3 has a similar effect on the NH 3 -SCR activity of Cu-CHA as SO 2 , 3,20,29,30 which would be consistent with the proposed reaction scheme.However, our data do not allow the exact structure of the mobile SO a X complex to be determined, so the formation of SO 3 remains to be proven (or ruled out) experimentally.phases and their reconstruction from the Cu I during the exposures to O 2 .However, the exposure of the sample to alternating SO 2 /O 2 cycles (Fig. 6) also reveals that the growth of the sulfated component occurs not only in SO 2 phases but also when the sample is exposed to O 2 .Moreover, during the later stages of the cycles (3rd and 4th cycles), the growth of the sulfated component is observed only in the presence of O 2 , while the exposure to SO 2 at this stage does not result in the formation of the sulfated component at all.Nonetheless, even at this stage the reactive [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complexes do undergo decomposition in the presence of SO 2 , resulting in sulfur uptake, as evidenced by the evolution of the S/Cu ratio obtained from the in situ XAQ signal (Fig. 13).A possible explanation for this effect is that at the late stages of the cycles, when the reactive Cu species are few, it is less likely to have two

Conclusions
In this study, we applied in situ XAS at Cu and S K-edges, S Ka XES, XAQ and SO 2 -TPD to investigate the interaction mechanism between the [Cu  accumulated in the sample was uncovered.Copper in the sulfated species exists as Cu 2+ and adopts a square-planar coordination with four light ligands in the rst coordination shell, which, most probably, are NH 3 and O. Sulfur in the sulfated species is in the S 6+ oxidation state, forming an SO 4 group.The sulfur atom is located in the second shell of Cu at an approximate distance of 2.6 Å, suggesting that Cu and S are connected through two oxygen ligands.
in the ESI.† Fig. 1 shows the Cu K-edge XANES and EXAFS FT data collected during the exposure of the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex to SO 2 at 200 °C for the low-Cu/CHA and high-Cu/CHA catalysts.The observed trends in the XANES and EXAFS spectra in these measurements are quite similar, indicating that the reaction of SO 2 proceeds in a similar way for both catalysts.There is a clear increase of the XANES peak at 8983 eV, indicating the partial reduction of Cu II to Cu I , and a decrease in the intensity of the rst shell in the EXAFS FT, indicating a reduction of the coordination number, due to decomposition of the [Cu 2 II (NH 3 ) 4 -O 2 ] 2+ complex.

Fig. 2
shows all the reference components used in the linear combination ts (three experimental spectra and two MCR components).More details on the choice of the reference spectra for the linear combination tting procedure are given in the ESI.† Fig. 3 presents the concentration proles of the reference components for both the low-Cu and the high-Cu catalysts during the pre-treatment and exposure to SO 2 .The concentration of the sulfated component increases when the [Cu 2 -II (NH 3 ) 4 O 2 ] 2+ -complex is exposed to SO 2 .The nal fraction of the sulfated component is 17% in the high-Cu sample, and 22% in the low Cu/CHA sample.Aer the exposure to SO 2 , we nd around 50% of the Cu present as the linear [Cu I (NH 3 ) 2 ] + complex, and 25% as fw-Cu I , indicating that 75% of the Cu in

Fig. 2
Fig. 2 Components used as references in the linear combination fit.The fw-Cu II , [Cu I (NH 3 ) 2 ] + and [Cu 2 II (NH 3 ) 4 O 2 ] 2+ components are experimental spectra.The spectra of fw-Cu I and the sulfated component are calculated by MCR-ALS.

2 Fig. 3
Fig. 4 shows the evolution of XANES and EXAFS FT when exposing the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex to a mixture of 360 ppm SO 2 and 10% O 2 at 200 °C.XANES spectra reveal a slight initial increase and subsequent decrease of the peak at 8983 eV.This indicates that transient Cu I species are formed in the process, which react further with the oxygen to form Cu II , resulting in the nal oxidation state of Cu being +2.This is further corroborated by the 1s-3d transition at 8978 eV indicating the presence of Cu II , as this transition is not present in Cu I .The intensity of the rst peak of the EXAFS FT for the nal spectrum aer exposure to SO 2 + O 2 is close to the initial intensity, indicating that the average coordination number of the rst shell of Cu remains close to four.That shows that 4-

Fig. 5
Fig. 5 Quantification of Cu compounds during the in situ Cu K-edge XANES measurements from LCF, using the reference spectra shown in Fig. 2. Upper panels: R-factors of the linear combination fits.Lower panels: concentration profiles for the different Cu species during reduction in NH 3 + NO, the formation of the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ -complex, and exposure of the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ -complex to SO 2 + O 2 , for the low-Cu (left) and high-Cu (right) catalysts.

Fig. 6
Fig. 6 Quantification of Cu compounds during the in situ Cu K-edge XANES measurements from linear combination fits, using the reference spectra shown in Fig. 2. Upper panels: R-factors of the linear combination fits.Lower panels: concentration profiles for the different Cu species during reduction in NH 3 + NO, the formation of the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex, and consequent exposures of the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex to SO 2 and O 2 , for the low-Cu (left) and high-Cu (right) catalysts.
2 mainly reacts with the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex.The features in the range 750-1000 °C remain largely unaffected by the presence of O 2 , while the peak around 420 °C becomes larger, and shows a slight shi towards higher temperatures.Overall, SO 2 -TPD results indicate that the

Fig. 8
Fig. 8 SO 2 -TPD profiles collected after exposure of the catalysts to SO 2 (red and blue lines) and SO 2 + O 2 (orange and light blue lines) in comparison to a reference SO 2 -TPD curve of a Cu-free CHA zeolite impregnated with 20 wt% (NH 4 ) 2 SO 4 , downscaled ×10.Pre-treatment is the same as for the procedures followed by in situ XAS.The curves for the high Cu catalysts (3.2 wt% Cu/CHA) (blue and light blue lines) are downscaled ×4.

Fig. 10
shows the evolution of S K-edge XANES spectra during exposure of the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex to SO 2 .The spectra were collected in HERFD mode.The increase of the edge jump corresponds to the increasing concentration of sulfur, which means that S is accumulated in the sample.From the position of the edge and the shape of the spectrum we can identify the oxidation state of S and possible local environment by comparing with references.Fig. 11a shows that the sulfur in the sample predominately exists in the S 6+ oxidation state, forming an SO 4 2− group.Fig. 11b presents Ka XES spectra of the [Cu 2 -II (NH 3 ) 4 O 2 ] 2+ complex aer exposure to SO 2 in comparison with reference compounds.The positions of the two features at 2.3066 and 2.3078 keV in the Ka XES of the complex agree well with the references containing S 6+ in the SO 4 2− group, which is in line with the ndings from S K-edge XANES.The sulfated component From our measurements, we can identify the [Cu I (NH 3 ) 2 ] + , fw-Cu II , fw-Cu I and reactive [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complexes, even though the exact structure of the framework-bound complexes is still under debate.Following the exposure of the Cu-CHA catalysts to SO 2 , both in the presence and in the absence of O 2 , a new species, whose spectrum we designated a "sulfated component", appears.In order to elucidate the structure of the species associated with the sulfated component, we compared the spectrum generated by MCR-ALS with the experimental spectra of three references: [Cu II (NH 3 ) 4 ]SO 4 $H 2 O, the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex formed in Cu-CHA and [Cu 2 II (NH 3 ) 4 ] 2+ complex in solution (Fig. 12).The sulfated component has striking similarity to the spectrum of Cu(NH 3 ) 4 SO 4 $H 2 O. Therefore, we propose that the species that give rise to the sulfated component contain 4coordinated Cu II in the square-planar environment similar to the one of Cu(NH 3 ) 4 SO 4 , where a square-planar Cu II ion is coordinated to four NH 3 ligands.It is also possible that a geometrically similar Cu II conguration with mixed NH 3 /O ligands is realized in the zeolite, since the XANES spectrum is expected to be very similar.Because the S atoms are not directly

Fig. 9 S
Fig. 9 S/Cu ratios in the low-Cu (0.8 wt% Cu/CHA) and high-Cu (3.2 wt% Cu/CHA) samples after exposure to SO 2 and SO 2 + O 2 obtained from XAQ and SO 2 -TPD compared to the concentration of the sulfated component obtained from the XANES LCF.

Fig. 10 S
Fig. 10 S K-edge XANES spectra of the high-Cu catalyst (3.2 wt% Cu/ CHA) collected in situ during exposure of the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex to SO 2 at 200 °C and in He after exposure to SO 2 .

[Cu 2 II(
NH 3 ) 4 O 2 ] 2+ complex involves a transformation of the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex to a different structure, having similar coordination of the Cu-ions.A possibility is the formation of peroxo-dicopper complexes Cu x O y attached to the framework, 28 which can be formed by exposure of the Cu-CHA to O 2 at 400-500 °C.Such species are expected to have a very similar XANES spectrum to the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complexes, because coordination of the Cu-ions is similar, and oxidation state is the same.Therefore, XANES spectra of such framework-bound complexes may be difficult to distinguish from those of the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex by MCR-ALS analysis, which merges them into a single component.Such an interpretation is supported by the observation that the component assigned to [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complexes appears also during the initial oxidation of the high-Cu catalyst at 500 °C (green curve in Fig. 6, right panel, at t = 0) with the concentration of around 20%, which is very close to the concentration observed aer exposure to SO 2 + O 2 .Clearly, it is not possible to form [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complexes at the activation stage, since the required NH

Fig. 12
Fig. 12 Comparison of the sulfated component obtained with MCR-ALS with the Cu K-edge XANES spectra of the [Cu(NH 3 ) 4 ]SO 4 $H 2 O, [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex inside the CHA and Cu II 2 (NH 3 ) 4 in solution (left); the fitting results for the reconstructed EXAFS of the sulfated component (right).

[Cu 2 II(
NH 3 ) 4 O 2 ] 2+ complexes close enough to perform the second step of the sulfation reaction (eqn (2)).Therefore, the mobile sulfur species SO a X formed in the rst step (eqn (1)) can be converted into the sulfated Cu species only aer the stock of [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complexes is replenished upon the exposure to O 2 .It is possible that in such regime SO a X undergo further transformations (e.g.reacting with NH 3 or NH 4 + ) before reacting with newly formed [Cu 2 II (NH 3 ) 4 O 2 ] 2+ and yielding the sulfated Cu species, but the obtained data do not allow unambiguous iden-tication of the corresponding reaction pathways.Nonetheless, the observed effect serves as indirect conrmation of the multistep nature of the sulfation process involving at least two [Cu 2 -II (NH 3 ) 4 O 2 ] 2+ complexes per SO 2 .

Fig. 13
Fig. 13 The evolution of the S/Cu ratio during the SO 2 /O 2 -cycles for the high-Cu sample (3.2 wt% Cu/CHA) deduced from the XAQ signal.
These [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complexes are then reduced back to the original [Cu I (NH 3 ) 2 ] + complexes by NH 3 and NO, under the formation of the reaction products N 2 and H 2 O.The mobility of the [Cu I (NH 3 ) 2 ] + complexes is important, as it facilitates the formation of the required Cu I pairs for the O 2 activation, enabling the NH 3 -SCR reaction at low temperatures.Because the presence of SO 2 results in a signicantly lower activity of the Cu-CHA catalysts below 300 °C, the SO 2 must affect the NH 3 -SCR reaction cycle.Wehave recently shown that SO 2 reacts with [Cu 2 decomposition of the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex, and a partial reduction of Cu II to Cu I .To determine the effect of this reaction of SO 2 with the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex on the rate of the NH 3 -SCR reaction, we need to know how this reaction proceeds and what reaction products are formed.A recent theoretical study proposes that deactivation by SO 2 occurs via the accumulation of ammonium bisulfate (NH 4 )HSO 4 in the zeolite aer initial reaction with the [Cu 2 II in a mixture of NH 3 and NO at the same temperature.This procedure leads to an almost complete conversion of Cu to [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complexes.
complexes in Cu-CHA catalysts with an Si/Al ratio of 6.7, and Cu contents of 3.2 wt% and 0.8 wt%.To maximize the amount of the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complexes in the catalysts prior to the reaction with SO 2 , we expose the catalyst to O 2 at 200 °C for 60 minutes aer reduction 11,14,16A Cu content of 3.2 wt% is typical for technical Cu-CHA based NH 3 -SCR catalysts.At 0.8 wt%, the low-temperature activity is signicantly lower, and therefore this sample represents a situation in which the formation of the [Cu II 2 (NH 3 ) 4 O 2 ] 2+ complex becomes less efficient.The XAS at the Cu K-edge is used to obtain information on the Cu-species formed upon exposure to SO 2 .Since the reaction of the [Cu 2 . Upon exposure to SO 2 in the rst two cycles, the [Cu 2 II (NH 3 ) 4 -O 2 ] 2+ complexes underwent decomposition, resulting in the formation of fw-Cu I , [Cu I (NH 3 ) 2 ] + and the species corresponding to the sulfated component.However, in the subsequent 3rd and 4th cycles, the sulfated component did not exhibit signicant growth during the exposure to SO 2 .In the case of the low-Cu sample, the conversion of [Cu 2 -II (NH 3 ) 4 O 2 ] 2+ complexes at each step was nearly complete.In contrast, for the high-Cu sample, almost complete conversion was observed only aer the initial exposure to SO 2 .Conversely, during each consequent exposure to SO 2 , a signicant amount of the component previously assigned to the [Cu 2 II (NH 3 ) 4 O 2 ] 2+

Table 1
Results of the EXAFS fitting for sulfated component the reaction between SO 2 and the rst complex, which then reacts with a second [Cu 2 II (NH 3 ) 4 O 2 ] 2+ .Since the mobility of the bulky [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex is expected to be limited, the formation of a smaller mobile intermediate product seems more likely. in Cu 2 II (NH 3 ) 4 O 2 + SO 2 / Cu I (NH 3 ) 2 + fw-Cu I + SO a X + .(1)SO a X + Cu 2 II (NH 3 ) 4 O 2 / Cu I (NH 3 ) 2 + Cu II SO 4 Z + .(2)In the second step (eqn (2)), the mobile SO a X intermediate reacts with another [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex, breaking the peroxo-bond to form the (SO 4 ) 2− group within the sulfated Cuspecies associated with sulfated XAS component (Cu II SO 4 Z) and another linear diamino complex [Cu I (NH 3 ) 2 ] + .Z in the sulfated species Cu II SO 4 Z may comprise O, framework O or NH 3 to result in a square-planar coordination of Cu with 4-ligands proven by our EXAFS and XANES results (Fig.12).This two-step reaction scheme is in line with the Cu concentration proles obtained by the LCF analysis in the case of exposure to SO 2 in the absence of O 2 .Note that eqn (1) and (2) aim to summarize the experimental ndings in this article, and therefore do not represent a complete mechanism of the reaction of SO 2 with the Cu-CHA catalyst.When the sample is exposed to SO 2 in the presence of O 2 , we observe an initial transient formation of the same Cu I intermediates, as in case of the exposure to SO 2 alone (Fig.5).This indicates that O 2 reoxidizes the [Cu I (NH 3 ) 2 ] + complexes into new [Cu 2 II (NH 3 ) 4 O 2 ] 2+ species.In this way, the Cu I -species formed in the reaction with SO 2 become available again for further reaction with SO 2 , which explains the increased SO 2 uptake compared to exposure to SO 2 without oxygen.
2 II (NH 3 ) 4 O 2 ] 2+ complex in the Cu-CHA catalyst and SO 2 .Upon reacting the [Cu 2 II (NH 3 ) 4 O 2 ] 2+ complex with SO 2 , a mixture of fw-Cu I (approximately 1/4 of total Cu), [Cu I (NH 3 ) 2 ] + complexes (approximately 1/2) and a new sulfated Cu II compound (approximately 1/4) are formed.The presence of oxygen in the gas mixture with SO 2 enhances the reaction, leading to higher concentrations of the sulfated species and an increased S/Cu ratio in the sample.This effect is explained by reoxidation of the [Cu I (NH 3 ) 2 ] + species to the reactive [Cu 2 -II (NH 3 ) 4 O 2 ] 2+ complexes.The catalysts with 0.8 wt% Cu/CHA and 3.2 wt% Cu/CHA demonstrated similar results.Following a multi-technique experimental approach, the structure of the Cu and S local environment of sulfated species