Recent Understanding of Low-Temperature Copper Dynamics in Cu-Chabazite NH 3 -SCR Catalysts

: Dynamic motion of NH 3 -solvated Cu sites in Cu-chabazite (Cu-CHA) zeolites, which are the most promising and state-of-the-art catalysts for ammonia-assisted selective reduction of NO x (NH 3 -SCR) in the aftertreatment of diesel exhausts, represents a unique phenomenon linking heterogeneous and homogeneous catalysis. This review ﬁrst summarizes recent advances in the theoretical understanding of such low-temperature Cu dynamics. Speciﬁcally, evidence of both intra-cage and inter-cage Cu motions, given by ab initio molecular dynamics (AIMD) or metadynamics simulations, will be highlighted. Then, we will show how, among others, synchrotron-based X-ray spectroscopy, vibrational and optical spectroscopy (diffuse reﬂection infrared Fourier transform spectroscopy (DRIFTS) and diffuse reﬂection ultraviolet-visible spectroscopy (DRUVS)), electron paramagnetic spectroscopy (EPR), and impedance spectroscopy (IS) can be combined and complement each other to follow the evolution of coordinative environment and the local structure of Cu centers during low-temperature NH 3 -SCR reactions. Furthermore, the essential role of Cu dynamics in the tuning of low-temperature Cu redox, in the preparation of highly dispersed Cu-CHA catalysts by solid-state ion exchange method, and in the direct monitoring of NH 3 storage and conversion will be presented. Based on the achieved mechanistic insights, we will discuss brieﬂy the new perspectives in manipulating Cu dynamics to improve low-temperature NH 3 -SCR efﬁciency as well as in the understanding of other important reactions, such as selective methane-to-methanol oxidation and ethene dimerization, catalyzed by metal ion-exchanged zeolites.


Introduction
Nitrogen oxide emissions (mainly NO and NO 2 ) from power plants and automobiles not only harm the human respiratory system but also participate in atmospheric reactions to form, among others, fine particulate matter (PM), ground-level ozone (O 3 ), and photochemical smog [1]. Cu-loaded small-pore chabazite (Cu-CHA) zeolites, including Cu-SSZ-13 and Cu-SAPO-34, have been successfully commercialized as catalysts to reduce nitrogen oxide emissions (NO x ) from automotive exhausts by selective catalytic reduction with ammonia (NH 3 -SCR) [2][3][4][5][6] via the following route: 4 NO + 4 NH 3 + O 2 → 4 N 2 + 6 H 2 O (standard NH 3 -SCR) The unique small-pore feature of the CHA framework allows for the prevention of hydrocarbons (HCs) in exhausts from entering the zeolite cages, thereby avoiding damage to the framework structure caused by high temperatures due to HC combustion.

Nature of Dynamic Cu Species
In Cu-CHA, the presence of Brønsted acidity, which originates from the Al substitution of Si T-atoms in the CHA framework and the subsequent charge compensation by protons, allows for strong NH 3 adsorption even at high temperatures (above 400 • C) [38][39][40][41][42][71][72][73]. The introduction of Cu ions, enabled by electrostatic attraction between metal ions and the polyanionic zeolite framework, provides additional sites for NH 3 adsorption with a relatively lower binding enthalpy compared with that on Brønsted sites [74]. The coexistence of Brønsted sites, Cu sites, and other weaker Lewis acidic sites (such as extra-framework Al sites) leads to a condensed phase of adsorbed NH 3 molecules within the zeolite structure, generating a unique solution-like environment at low temperatures [50,75,76].
Even if only NH 3 adsorption is considered, the Cu species formed at two sites of different Al coordination are still very complex in nature, as shown in Figure 1. Paolucci et al. calculated the free energy of NH 3 adsorption in the presence of NH 3 (300 ppm) and H 2 O (2%) using first-principles thermodynamic analysis and obtained phase diagrams of Cu speciation by taking the coordination Al of the framework, temperature, and O 2 partial pressure into consideration [38]. It can be clearly seen that, with the increase in temperature, the NH 3 ligands are gradually desorbed and Cu ions eventually bind to the framework in the form of isolated Cu ions (Z represents a negative charge of the zeolite framework). Further calculations show that the proportion of Cu + on the 1Al site (a single Al atom in the 6-MR) is higher than that in the 2Al (paired Al atoms in proximity of the 6-MR) in reducing atmosphere, indicating that the Cu ions in the former are more easily reduced, which is consistent with temperature-programmed reduction with H 2 (H 2 -TPR) results in the literature [77][78][79][80]. During low-temperature NH 3 -SCR reactions (200 • C), the major species at 1Al sites were found to be [Cu + (OH)(NH 3 ) 3 ] + and [Cu + (NH 3 ) 2 ] + , while [Cu 2+ (NH 3 ) 4 ] 2+ species dominated at 2Al sites. When NH 3 ligands pull the Cu species from the framework oxygen coordination positions into the CHA cage, the NH 3 -solvated Cu species migrate dynamically between cages due to the low inter-cage diffusion barrier, as shown in Figure 2. Considering that the diameter of 4-MR is too small to allow molecules  [81]. Paolucci et al. [38] performed AIMD simulations to visualize the diffusion trajectories of [Cu + (NH 3 ) 2 ] + and [Cu 2+ (NH 3 ) 4 ] 2+ species in CHA cages and confirmed that both NH 3 -ligated species are mobile enough to allow Cu to move away from the equilibrium position. Relatively, [Cu + (NH 3 ) 2 ] + species move within a larger volume and thus farther from the original equilibrium position, suggesting that the linear and less sterically hindered [Cu + (NH 3 ) 2 ] + species are more mobile and thus could diffuse through the 8-MR windows between adjacent CHA cages.
sts 2021, 11, x FOR PEER REVIEW 4 of 20 the framework oxygen coordination positions into the CHA cage, the NH3-solvated Cu species migrate dynamically between cages due to the low inter-cage diffusion barrier, as shown in Figure 2. Considering that the diameter of 4-MR is too small to allow molecules to pass through, Chen et al. investigated the possible diffusion limitations for the complex encountering through 8-MRs by DFT calculations and obtained a diffusion barrier of only 0.29 eV [81]. Paolucci et al. [38] performed AIMD simulations to visualize the diffusion trajectories of [Cu + (NH3)2] + and [Cu 2+ (NH3)4] 2+ species in CHA cages and confirmed that both NH3-ligated species are mobile enough to allow Cu to move away from the equilibrium position. Relatively, [Cu + (NH3)2] + species move within a larger volume and thus farther from the original equilibrium position, suggesting that the linear and less sterically hindered [Cu + (NH3)2] + species are more mobile and thus could diffuse through the 8-MR windows between adjacent CHA cages.

Inter-Cage Motion of Cu Sites
To further assess the characteristics of [Cu + (NH3)2] migration between cages, Paolucci et al. calculated the free energy changes of [Cu + (NH3)2] at different positions by the meta-

Inter-Cage Motion of Cu Sites
To further assess the characteristics of [Cu + (NH3)2] migration between cages, Paolucci et al. calculated the free energy changes of [Cu + (NH3)2] at different positions by the meta-

Inter-Cage Motion of Cu Sites
To further assess the characteristics of [Cu + (NH 3 ) 2 ] migration between cages, Paolucci et al. calculated the free energy changes of [Cu + (NH 3 ) 2 ] at different positions by the metadynamics method and by taking Cu-Al distance as a variable [56]. As shown in Figure 3, in the simulation of [Cu + (NH 3 ) 2 ] diffusion between two adjacent CHA cages, the free energy was found to reach a minimum at a Cu-Al distance of 4.7 Å, that is, at the initial equilibrium position marked as (1). When the [Cu + (NH 3 ) 2 ] passes through 8-MR, the free energy value reaches a peak and decreases immediately after the crossing. With the further increase of Cu-Al distance to more than 9 Å, free energy increases due to an increase of electrostatic tethering of [Cu + (NH 3 ) 2 ] by the framework Al center, which means that the maximum distance for dynamic diffusion of Cu species was about 9 Å [56].
Catalysts 2021, 11, x FOR PEER REVIEW 5 of 20 dynamics method and by taking Cu-Al distance as a variable [56]. As shown in Figure 3, in the simulation of [Cu + (NH3)2] diffusion between two adjacent CHA cages, the free energy was found to reach a minimum at a Cu-Al distance of 4.7 Å, that is, at the initial equilibrium position marked as (1). When the [Cu + (NH3)2] passes through 8-MR, the free energy value reaches a peak and decreases immediately after the crossing. With the further increase of Cu-Al distance to more than 9 Å, free energy increases due to an increase of electrostatic tethering of [Cu + (NH3)2] by the framework Al center, which means that the maximum distance for dynamic diffusion of Cu species was about 9 Å [56]. This inter-cage dynamic motion of Cu sites is considered a key step affecting the NH3-SCR efficiency of low Cu-loaded CHA zeolites because two Cu sites are required for dissociation of O2 in the re-oxidation half-cycle [38,56,57,69]. Therefore, the higher the transport efficiency in the cage, the easier it is to form dimeric Cu species, which can promote NH3-SCR activity at low temperatures. As proposed by Paolucci et al., the inter-cage Cu motion is determined by electrostatic tethering of Cu to the framework. Therefore, change in the coordinative environment around Cu sites may cause losses in transport efficiency, which will eventually affect the low-temperature performance of the Cu-CHA catalyst. Recently, O'Malley et al. investigated the diffusion trajectory of Cu species in zeolite with different topologies (LEV and CHA) by molecular dynamics (MD) simulations and quasi-elastic neutron scattering (QENS) experiments [82]. Although QENS experiments indicated that the rates of NH3 migration within the two zeolite cages are the same, MD simulations (see Figure 4 for the molecular diffusion trajectories of NH3 in LEV and CHA) suggested that the inter-cage motion frequency of NH3 molecules in the CHA cage is twice that in the LEV because the former contains more 8-MR windows for NH3 transport between chabazite cages. Although only NH3 molecules were considered in this study, the results provide a theoretical and experimental basis for a deeper understanding of local dynamics of NH3-solvated Cu species. This inter-cage dynamic motion of Cu sites is considered a key step affecting the NH 3 -SCR efficiency of low Cu-loaded CHA zeolites because two Cu sites are required for dissociation of O 2 in the re-oxidation half-cycle [38,56,57,69]. Therefore, the higher the transport efficiency in the cage, the easier it is to form dimeric Cu species, which can promote NH 3 -SCR activity at low temperatures. As proposed by Paolucci et al., the inter-cage Cu motion is determined by electrostatic tethering of Cu to the framework. Therefore, change in the coordinative environment around Cu sites may cause losses in transport efficiency, which will eventually affect the low-temperature performance of the Cu-CHA catalyst. Recently, O'Malley et al. investigated the diffusion trajectory of Cu species in zeolite with different topologies (LEV and CHA) by molecular dynamics (MD) simulations and quasi-elastic neutron scattering (QENS) experiments [82]. Although QENS experiments indicated that the rates of NH 3 migration within the two zeolite cages are the same, MD simulations (see Figure 4 for the molecular diffusion trajectories of NH 3 in LEV and CHA) suggested that the inter-cage motion frequency of NH 3 molecules in the CHA cage is twice that in the LEV because the former contains more 8-MR windows for NH 3 transport between chabazite cages. Although only NH 3 molecules were considered in this study, the results provide a theoretical and experimental basis for a deeper understanding of local dynamics of NH 3 -solvated Cu species.

Experimental Detection
The dynamic formation of multi-nuclear active sites, such as NH3-solvated dimeric Cu complexes, provides different and important perspectives for understanding the NH3-SCR reaction mechanisms at low temperatures. However, challenges remain in the experimental study of such dynamic reaction processes and such transiently existing active sites. The ligating details of the NH3-solvated Cu sites are too complicated ( Figure 1) to be resolved unambiguously, which results in an obstacle to explore the dynamic reaction process. Recently, substantial advances have been achieved by applying combinatory and complementary characterizations techniques (such as XAS, DRIFTS, and EPR) to track the dynamic change in Cu coordination environment, which will be discussed in the following.

X-ray Absorption Spectroscopy Based Techniques
XAS mainly obtains relevant coordination information of active Cu sites, e.g., the type and number of ligands, by monitoring absorption at the K-edges [48,[83][84][85][86], and is one of the few techniques that can accurately analyze the complex structures of NH3-solvated Cu species at an atomic scale. Considering that NH3 desorption on Cu ions at high temperature is expected to affect SCR reaction activity, Lomachenko et al. explored the dependence between temperature and migration molecules in NH3-SCR atmosphere in the temperature range of 150~400 °C by operando XAS (Figure 5a) [84]. Characteristic peaks in the XAS spectra were identified and explained by comparing with standard references, and the Cu speciation was further quantified by linear combination fitting (LCF) analysis. The XANES spectra showed a significant edge-rising peak at 8982.5 eV during exposure at 150 °C in NH3-SCR atmosphere, which was assigned to a linear [Cu + (NH3)2] complex. LCF analysis showed that, at low temperature (150 °C), two transient species, m-Cu + and m-Cu 2+ (m represents "mobile"), were the dominant active sites, accounting for 46% and 25% of the total Cu species, respectively, while Z-Cu (Z represents framework-interacting) species account for only 26%. With the increase in temperature, NH3 molecules adsorbed on the Cu ions were gradually desorbed, and more and more Cu ions were finally anchored on the framework, corresponding to the reduction in dynamic species. The two different active Cu sites at high temperatures and low temperatures contributed to two different mechanisms for the NH3-SCR reaction on Cu-CHA, resulting in a "seagull" type of temperature-activity curve [56][57][58]84,87].

Experimental Detection
The dynamic formation of multi-nuclear active sites, such as NH 3 -solvated dimeric Cu complexes, provides different and important perspectives for understanding the NH 3 -SCR reaction mechanisms at low temperatures. However, challenges remain in the experimental study of such dynamic reaction processes and such transiently existing active sites. The ligating details of the NH 3 -solvated Cu sites are too complicated ( Figure 1) to be resolved unambiguously, which results in an obstacle to explore the dynamic reaction process. Recently, substantial advances have been achieved by applying combinatory and complementary characterizations techniques (such as XAS, DRIFTS, and EPR) to track the dynamic change in Cu coordination environment, which will be discussed in the following.

X-ray Absorption Spectroscopy Based Techniques
XAS mainly obtains relevant coordination information of active Cu sites, e.g., the type and number of ligands, by monitoring absorption at the K-edges [48,[83][84][85][86], and is one of the few techniques that can accurately analyze the complex structures of NH 3 -solvated Cu species at an atomic scale. Considering that NH 3 desorption on Cu ions at high temperature is expected to affect SCR reaction activity, Lomachenko et al. explored the dependence between temperature and migration molecules in NH 3 -SCR atmosphere in the temperature range of 150~400 • C by operando XAS (Figure 5a) [84]. Characteristic peaks in the XAS spectra were identified and explained by comparing with standard references, and the Cu speciation was further quantified by linear combination fitting (LCF) analysis. The XANES spectra showed a significant edge-rising peak at 8982.5 eV during exposure at 150 • C in NH 3 -SCR atmosphere, which was assigned to a linear [Cu + (NH 3 ) 2 ] complex. LCF analysis showed that, at low temperature (150 • C), two transient species, m-Cu + and m-Cu 2+ (m represents "mobile"), were the dominant active sites, accounting for 46% and 25% of the total Cu species, respectively, while Z-Cu (Z represents framework-interacting) species account for only 26%. With the increase in temperature, NH 3 molecules adsorbed on the Cu ions were gradually desorbed, and more and more Cu ions were finally anchored on the framework, corresponding to the reduction in dynamic species. The two different active Cu sites at high temperatures and low temperatures contributed to two different mechanisms for the NH 3 -SCR reaction on Cu-CHA, resulting in a "seagull" type of temperature-activity curve [56][57][58]84,87]. In addition to the significant influence of temperature on NH3-solvated copper, there are other factors that can cause notable interference in the distribution of dynamic species, such as pretreatment atmosphere and the properties of the material itself [87,88]. Borfecchia's group analyzed the XANES spectra of NH3-solvated species in samples activated in N2 or in vacuum and found that the latter contained more Cu + species [88]. This is because some Cu sites (mainly [Cu 2+ (OH)]) were reduced to Cu + when pretreated in vacuum or under high-temperature reducing atmosphere (i.e., so-called self-reduction), and the proportion of NH3-solvated species in the catalyst changed accordingly.

Optical Spectroscopy Techniques
DRIFTS is generally used to identify the characteristic vibrations of probe molecules (such as NH3, NO, and CO) adsorbed on the surface of catalysts. Additionally, DRIFTS can be also used to monitor vibrations of the zeolite framework, such as the -Al-O(H)-Siand asymmetric T-O-T vibrations [55,89]. The two IR features of framework T-O-T vibrations, at ca. 900 cm −1 and ca. 950 cm −1 , respectively, are sensitive to the existence and perturbation of Cu ions, i.e., Cu 2+ in the 6-MR and [Cu 2+ (OH)] in 8-MR [38,43,46,49]. After NH3 solvation, the Cu ions gradually migrate away from the original equilibrium position, weakening the interaction between the cations and zeolite framework and, consequently, leading to an obvious change in the T-O-T vibration peaks. On the basis of T-O-T vibration change, Yasser et al. tracked the poisoning of SO2 to NH3-solvated Cu active sites at low temperature by DRIFTS [90]. Two negative peaks located at 900 cm −1 and 950 cm −1 assigned to T-O-T vibration around Cu 2+ and [Cu 2+ (OH)], respectively, increased in intensity versus time ( Figure 6A), indicating that the two Cu active sites were gradually solvated by the adsorbed NH3 molecules. In the subsequent step of NO + O2 adsorption ( Figure  6B), the NH3 ligands on the dynamic Cu species were consumed by the SCR reaction and the negative IR bands returned to baseline because of the weakened local motion. The dynamic local migration process still existed even after suffering from SO2 poisoning (Figure 6C). However, the migration of [Cu 2+ (OH)] species was obviously inhibited. The SCR rates over isolated Cu 2+ and [Cu 2+ (OH)] sites can be calculated by integrating the timedependent evolution of respective IR band area. In addition to the significant influence of temperature on NH 3 -solvated copper, there are other factors that can cause notable interference in the distribution of dynamic species, such as pretreatment atmosphere and the properties of the material itself [87,88]. Borfecchia's group analyzed the XANES spectra of NH 3 -solvated species in samples activated in N 2 or in vacuum and found that the latter contained more Cu + species [88]. This is because some Cu sites (mainly [Cu 2+ (OH)]) were reduced to Cu + when pretreated in vacuum or under high-temperature reducing atmosphere (i.e., so-called self-reduction), and the proportion of NH 3 -solvated species in the catalyst changed accordingly.

Optical Spectroscopy Techniques
DRIFTS is generally used to identify the characteristic vibrations of probe molecules (such as NH 3 , NO, and CO) adsorbed on the surface of catalysts. Additionally, DRIFTS can be also used to monitor vibrations of the zeolite framework, such as the -Al-O(H)-Siand asymmetric T-O-T vibrations [55,89]. The two IR features of framework T-O-T vibrations, at ca. 900 cm −1 and ca. 950 cm −1 , respectively, are sensitive to the existence and perturbation of Cu ions, i.e., Cu 2+ in the 6-MR and [Cu 2+ (OH)] in 8-MR [38,43,46,49]. After NH 3 solvation, the Cu ions gradually migrate away from the original equilibrium position, weakening the interaction between the cations and zeolite framework and, consequently, leading to an obvious change in the T-O-T vibration peaks. On the basis of T-O-T vibration change, Yasser et al. tracked the poisoning of SO 2 to NH 3 -solvated Cu active sites at low temperature by DRIFTS [90]. Two negative peaks located at 900 cm −1 and 950 cm −1 assigned to T-O-T vibration around Cu 2+ and [Cu 2+ (OH)], respectively, increased in intensity versus time ( Figure 6A), indicating that the two Cu active sites were gradually solvated by the adsorbed NH 3 molecules. In the subsequent step of NO + O 2 adsorption ( Figure 6B), the NH 3 ligands on the dynamic Cu species were consumed by the SCR reaction and the negative IR bands returned to baseline because of the weakened local motion. The dynamic local migration process still existed even after suffering from SO 2 poisoning ( Figure 6C). However, the migration of [Cu 2+ (OH)] species was obviously inhibited. The SCR rates over isolated Cu 2+ and [Cu 2+ (OH)] sites can be calculated by integrating the time-dependent evolution of respective IR band area. DRUVS is often used as an auxiliary technique for exploring the dynamics of transient Cu species. Unlike DRIFTS, DRUVS does not monitor the vibration of probe molecules adsorbed on the surface but directly detects the charge transfer around the active site and, thus, can work even at high temperatures. Many researchers used the DRUVS technique to obtain information about the coordination and oxidation state of active sites during dehydration [91][92][93]. The gradual loss of H2O ligands in hydrated Cu ions upon hightemperature pretreatment, which makes the electronic exchange between Cu ion and framework oxygen more frequent, can be followed by monitoring the typical DRUVS spectral features in the range of 50,000-30,000 cm −1 (200-350 nm) originating from ligandto-metal charge transfer (LMCT) processes. Recently, Oda et al. demonstrated that the characteristic DRUVS features at 370 nm were assigned to Cu dimerization of two mobile NH3-solvated Cu ions [94]. The features of Cu dimers were observed in different zeolites, including Cu-CHA and Cu-AEI with different Si/Al ratios, as shown in Figure 7, indicating that the formation and dynamic migration of dimeric Cu species is a common phenomenon in Cu-zeolite materials. Apart from the LMCT feature, the characteristic d-d electron transition bands of Cu 2+ ions at 400~600 nm are often employed to identify the oxidation states of Cu ions [95]. The O2-activated Cu-CHA catalyst shows a broad d-d band, even after the adsorption of NH3. Nevertheless, the d-d transition peak only disappears after continuous exposure to NO, indicating that the reduction of Cu 2+ ions in Cu- DRUVS is often used as an auxiliary technique for exploring the dynamics of transient Cu species. Unlike DRIFTS, DRUVS does not monitor the vibration of probe molecules adsorbed on the surface but directly detects the charge transfer around the active site and, thus, can work even at high temperatures. Many researchers used the DRUVS technique to obtain information about the coordination and oxidation state of active sites during dehydration [91][92][93]. The gradual loss of H 2 O ligands in hydrated Cu ions upon hightemperature pretreatment, which makes the electronic exchange between Cu ion and framework oxygen more frequent, can be followed by monitoring the typical DRUVS spectral features in the range of 50,000-30,000 cm −1 (200-350 nm) originating from ligandto-metal charge transfer (LMCT) processes. Recently, Oda et al. demonstrated that the characteristic DRUVS features at 370 nm were assigned to Cu dimerization of two mobile NH 3 -solvated Cu ions [94]. The features of Cu dimers were observed in different zeolites, including Cu-CHA and Cu-AEI with different Si/Al ratios, as shown in Figure 7, indicating that the formation and dynamic migration of dimeric Cu species is a common phenomenon in Cu-zeolite materials. Apart from the LMCT feature, the characteristic d-d electron transition bands of Cu 2+ ions at 400~600 nm are often employed to identify the oxidation states of Cu ions [95]. The O 2 -activated Cu-CHA catalyst shows a broad d-d band, even after the adsorption of NH 3 . Nevertheless, the d-d transition peak only disappears after continuous exposure to NO, indicating that the reduction of Cu 2+ ions in Cu-CHA at low  . Reproduced with permission from [96], copyright ChemCatChem, 2020.

Electron Paramagnetic Resonance (EPR)
EPR is an ideal technique to investigate the nature of Cu 2+ species even in trace amounts and can provide valuable information about oxidation state and local coordination of the Cu sites. The intensity of an EPR signal has a significant linear relationship with Cu loading, and different Cu species can be identified and even quantified according to the hyperfine splitting peaks in low magnetic fields. However, the resolution of low-field characteristic peaks was significantly reduced by the increased mobility of Cu 2+ , as confirmed by a comparison of the EPR spectra of the dehydrated and hydrated samples (Figure 8a), which means that it might be inaccurate to attribute the reduced EPR signal to Cu 2+ → Cu + reduction after NH3 adsorption onto Cu-CHA [93,[97][98][99]. Based on DFT calculations with the Breit-Pauli Hamiltonian, Fernandez et al. simulated EPR spectra by adding adsorbate molecules to optimized geometric structures of zeolites and by comparing it with the experimental data [100]. The results show that the addition of a NH3 molecule can break a Cu-Of bond (Of represents oxygen in the zeolite framework) and can drive the Cu ion moving away from the 6-MR plane. With the increase in adsorbed NH3 molecules, the connection between the Cu ions and framework gradually weakens, which causes Cu ions to migrate further into the CHA cage. Such an atomic study reveals that the restriction of zeolite framework on Cu ions is an important factor affecting the dynamic motion of Cu-NH3 complexes. In a recent study, Godiksen et al. found that the different features in the hyperfine quadruplet of EPR spectra in Figure 8b (A and B are assigned to para-and meta-configurations, respectively, originating from different Al distributions) can be clearly visualized after the dehydration of Cu-CHA under O2/He atmosphere but were quickly transformed into a single feature of new species after NH3 adsorption [101]. This observation indicates that, despite the existence of various active sites, the EPR spectra of transient Cu-NH3 species formed upon NH3 solvation are identical.

Electron Paramagnetic Resonance (EPR)
EPR is an ideal technique to investigate the nature of Cu 2+ species even in trace amounts and can provide valuable information about oxidation state and local coordination of the Cu sites. The intensity of an EPR signal has a significant linear relationship with Cu loading, and different Cu species can be identified and even quantified according to the hyperfine splitting peaks in low magnetic fields. However, the resolution of lowfield characteristic peaks was significantly reduced by the increased mobility of Cu 2+ , as confirmed by a comparison of the EPR spectra of the dehydrated and hydrated samples (Figure 8a), which means that it might be inaccurate to attribute the reduced EPR signal to Cu 2+ → Cu + reduction after NH 3 adsorption onto Cu-CHA [93,[97][98][99]. Based on DFT calculations with the Breit-Pauli Hamiltonian, Fernandez et al. simulated EPR spectra by adding adsorbate molecules to optimized geometric structures of zeolites and by comparing it with the experimental data [100]. The results show that the addition of a NH 3 molecule can break a Cu-O f bond (O f represents oxygen in the zeolite framework) and can drive the Cu ion moving away from the 6-MR plane. With the increase in adsorbed NH 3 molecules, the connection between the Cu ions and framework gradually weakens, which causes Cu ions to migrate further into the CHA cage. Such an atomic study reveals that the restriction of zeolite framework on Cu ions is an important factor affecting the dynamic motion of Cu-NH 3 complexes. In a recent study, Godiksen et al. found that the different features in the hyperfine quadruplet of EPR spectra in Figure 8b (A and B are assigned to paraand meta-configurations, respectively, originating from different Al distributions) can be clearly visualized after the dehydration of Cu-CHA under O 2 /He atmosphere but were quickly transformed into a single feature of new species after NH 3 adsorption [101]. This observation indicates that, despite the existence of various active sites, the EPR spectra of transient Cu-NH 3 species formed upon NH 3 solvation are identical.

Impedance Spectroscopy
In impedance spectroscopy (IS), the electrical properties of the sample are studied by applying an alternating voltage (U*) of known amplitude and frequency (f) and by measuring the sample electric response in terms of alternating electrical current [102,103]. When applied to solids such as zeolites, in which extra-framework ions are present, the measured current arises from ion conduction within the zeolite crystal while the zeolite lattice itself is an electronic insulator with a wide bandgap of several electron volts [104][105][106]. Depending on the applied frequency, different ion motion modes are activated and can be followed by means of complex impedance-based in situ modulus spectroscopy ( Figure 9). While the low-frequency range (LF, 10 −1 -10 3 Hz) is dominated by the longrange ion transport (e.g., NH3-supported proton mobility), resonance peaks due to local displacement of charges or short-range ion movement can be detected in the high-frequency range (HF, 10 3 -10 6 Hz) [106,107]. According to the abovementioned AIMD results [56], the inter-cage mobility of NH3-solvated Cu species belongs to this second class of ion motion modes, leading to a resonance peak in the frequency range of 10 4 -10 6 Hz. The highfrequency ion motion can be tracked by means of multi-frequency in situ IS measurements, which allow the collection of modulus plots, as shown in Figure 10 (more details about the so-called modulus spectroscopy may be found elsewhere in the literature) [103,106]. Investigations on a series of Cu-SSZ-13 samples prepared by ion exchange with the same H-form starting material revealed a correlation between Cu ion exchange level and the position of the HF resonance peak and that a higher Cu/Al ratio corresponds to a more pronounced HF shift of the resonance peak corresponding to enhanced ion mobility [108]. This trend was only observed when the samples were flushed with NH3 and not during exposure to an inert atmosphere (i.e., N2). Therefore, it could be concluded that the ion-motion phenomenon causing the peak shift is strictly correlated to NH3 solvation of the Cu ions. The following IS studies confirmed the contribution of the dynamic Cu sites to HF resonance by performing in situ IS under SCR-related gas and temperature conditions [74,109]. During standard SCR (200 °C, NH3/NO/O2), the Cu sites are present both as Cu 2+ and Cu + species, and the latter one is expected to be more mobile than the former. The formation of such highly mobile NH3-solvated complexes can be detected by IS according to the shift in the HF resonance peak compared to the same sample exposed to NH3 after oxidative pretreatment (Figure 10a,b, red and black). In the absence of O2, the re-oxidation of Cu + to Cu 2+ was instead inhibited, and therefore, a complete reduction of

Impedance Spectroscopy
In impedance spectroscopy (IS), the electrical properties of the sample are studied by applying an alternating voltage (U*) of known amplitude and frequency (f ) and by measuring the sample electric response in terms of alternating electrical current [102,103]. When applied to solids such as zeolites, in which extra-framework ions are present, the measured current arises from ion conduction within the zeolite crystal while the zeolite lattice itself is an electronic insulator with a wide bandgap of several electron volts [104][105][106]. Depending on the applied frequency, different ion motion modes are activated and can be followed by means of complex impedance-based in situ modulus spectroscopy (Figure 9). While the low-frequency range (LF, 10 −1 -10 3 Hz) is dominated by the long-range ion transport (e.g., NH 3 -supported proton mobility), resonance peaks due to local displacement of charges or short-range ion movement can be detected in the high-frequency range (HF, 10 3 -10 6 Hz) [106,107]. According to the abovementioned AIMD results [56], the inter-cage mobility of NH 3 -solvated Cu species belongs to this second class of ion motion modes, leading to a resonance peak in the frequency range of 10 4 -10 6 Hz. The high-frequency ion motion can be tracked by means of multi-frequency in situ IS measurements, which allow the collection of modulus plots, as shown in Figure 10 (more details about the so-called modulus spectroscopy may be found elsewhere in the literature) [103,106]. Investigations on a series of Cu-SSZ-13 samples prepared by ion exchange with the same H-form starting material revealed a correlation between Cu ion exchange level and the position of the HF resonance peak and that a higher Cu/Al ratio corresponds to a more pronounced HF shift of the resonance peak corresponding to enhanced ion mobility [108]. This trend was only observed when the samples were flushed with NH 3 and not during exposure to an inert atmosphere (i.e., N 2 ). Therefore, it could be concluded that the ion-motion phenomenon causing the peak shift is strictly correlated to NH 3 solvation of the Cu ions. The following IS studies confirmed the contribution of the dynamic Cu sites to HF resonance by performing in situ IS under SCR-related gas and temperature conditions [74,109]. During standard SCR (200 • C, NH 3 /NO/O 2 ), the Cu sites are present both as Cu 2+ and Cu + species, and the latter one is expected to be more mobile than the former. The formation of such highly mobile NH 3 -solvated complexes can be detected by IS according to the shift in the HF resonance peak compared to the same sample exposed to NH 3 after oxidative pretreatment (Figure 10a,b, red and black). In the absence of O 2 , the re-oxidation of Cu + to Cu 2+ was instead inhibited, and therefore, a complete reduction of the Cu sites was favored. As a consequence, the overall ion conductivity substantially increased, causing a further HF shift in the resonance peak (Figure 10a,b, blue). Similar effects were observed not only on Cu-SAPO-34 but also on other zeolites (e.g., Cu-ZSM-5 and Cu-SSZ-13) [74,109]. The simultaneous measurement of IS and DRIFTS (IS-DRIFTS) also allowed for the interpretation of the dielectric behaviors of different zeolites by the identification of key SCR-related intermediates, such as NH 4 + intermediates [74].
Catalysts 2021, 11, x FOR PEER REVIEW 11 of 20 the Cu sites was favored. As a consequence, the overall ion conductivity substantially increased, causing a further HF shift in the resonance peak (Figure 10a,b, blue). Similar effects were observed not only on Cu-SAPO-34 but also on other zeolites (e.g., Cu-ZSM-5 and Cu-SSZ-13) [74,109]. The simultaneous measurement of IS and DRIFTS (IS-DRIFTS) also allowed for the interpretation of the dielectric behaviors of different zeolites by the identification of key SCR-related intermediates, such as NH4 + intermediates [74].

Regulating Low-Temperature Cu Redox
It is widely recognized that the NH3-SCR reaction mechanisms on Cu-CHA catalysts are different at low and high temperatures. As verified by in situ XAS, the dominant Cu species at low temperatures are in the form of Cu-NH3 complexes [84]. Gao et al. proposed that two Cu sites were needed to participate in the rate-limiting step for SCR reaction at low temperatures, based on the facts that the SCR rate was correlated linearly with the square of Cu/Al ratio at a low temperature of 200 °C and linearly with the Cu/Al ratio at a high temperature of 350 °C (Figure 11a). In the re-oxidation half-cycle (Cu + → Cu 2+ ) of NH3-SCR catalysis, O2 dissociation is well recognized as a rate-determining step the Cu sites was favored. As a consequence, the overall ion conductivity substantially increased, causing a further HF shift in the resonance peak (Figure 10a,b, blue). Similar effects were observed not only on Cu-SAPO-34 but also on other zeolites (e.g., Cu-ZSM-5 and Cu-SSZ-13) [74,109]. The simultaneous measurement of IS and DRIFTS (IS-DRIFTS) also allowed for the interpretation of the dielectric behaviors of different zeolites by the identification of key SCR-related intermediates, such as NH4 + intermediates [74].

Regulating Low-Temperature Cu Redox
It is widely recognized that the NH3-SCR reaction mechanisms on Cu-CHA catalysts are different at low and high temperatures. As verified by in situ XAS, the dominant Cu species at low temperatures are in the form of Cu-NH3 complexes [84]. Gao et al. proposed that two Cu sites were needed to participate in the rate-limiting step for SCR reaction at low temperatures, based on the facts that the SCR rate was correlated linearly with the square of Cu/Al ratio at a low temperature of 200 °C and linearly with the Cu/Al ratio at a high temperature of 350 °C (Figure 11a). In the re-oxidation half-cycle (Cu + → Cu 2+ ) of NH3-SCR catalysis, O2 dissociation is well recognized as a rate-determining step

Regulating Low-Temperature Cu Redox
It is widely recognized that the NH 3 -SCR reaction mechanisms on Cu-CHA catalysts are different at low and high temperatures. As verified by in situ XAS, the dominant Cu species at low temperatures are in the form of Cu-NH 3 complexes [84]. Gao et al. proposed that two Cu sites were needed to participate in the rate-limiting step for SCR reaction at low temperatures, based on the facts that the SCR rate was correlated linearly with the square of Cu/Al ratio at a low temperature of 200 • C and linearly with the Cu/Al ratio at a high temperature of 350 • C (Figure 11a). In the re-oxidation half-cycle (Cu + → Cu 2+ ) of NH 3 -SCR catalysis, O 2 dissociation is well recognized as a rate-determining step [38,49,97,[111][112][113] [56,114,115]. Using DFT calculations, Lin et al. proved that the energy barrier for O 2 dissociation is much lower on a NH 3 -solvated Cu pair than on a single NH 3 -solvated Cu site [111]. However, the structure of the Cu dimer formed by the combination of O 2 molecules with two [Cu + (NH 3 ) 2 ] + complexes is still controversial. Two highly similar structures, namely mono-(µ-oxo) dicopper(II) and bis(µ-oxo) dicopper(III), are frequently proposed in the literature. Based on operando spectroscopic evidence and DFT calculations, several recent studies proposed the bis(µ-oxo) dicopper(III) structure as a model, which means that the O 2 molecule is dissociated after adsorption by the Cu centers [56,94,111,114,115]. As suggested by the simulated diffusion area of dynamic Cu species within the CHA unit (the green spherical space shown in the Figure 11b) [56], the transport and pairing of [Cu + (NH 3 ) 2 ] + between CHA cages is the rate-limiting step for Cu-CHA catalysts with low Cu loadings. As described in Section 2.2, due to electrostatic tethering to the framework Al centers, the dynamic diffusion distance of individual Cu-NH 3 species is limited to less than 9 Å, so only in the overlapping parts of the diffusion space is it feasible to form Cu dimers for O 2 activation. It is speculated that the lowtemperature NH 3 -SCR efficiency at low Cu loadings can be improved by reinforcing the dynamic local motion of Cu-NH 3 species and by accordingly promoting the dissociation of O 2 [56,57,68]. With the increase of Cu loading, however, formation of Cu dimers is enabled by high Cu density and no longer limited by the long-distance migration process.
Catalysts 2021, 11, x FOR PEER REVIEW 12 of 20 [38,49,97,[111][112][113]. Recently, it was found that the [Cu + (NH3)2] + −O2−[Cu + (NH3)2] + intermediate as the most likely transient Cu dimer species plays a crucial role in O2 activation and dissociation [56,114,115]. Using DFT calculations, Lin et al. proved that the energy barrier for O2 dissociation is much lower on a NH3-solvated Cu pair than on a single NH3-solvated Cu site [111]. However, the structure of the Cu dimer formed by the combination of O2 molecules with two [Cu + (NH3)2] + complexes is still controversial. Two highly similar structures, namely mono-(μ-oxo) dicopper(II) and bis(μ-oxo) dicopper(III), are frequently proposed in the literature. Based on operando spectroscopic evidence and DFT calculations, several recent studies proposed the bis(μ-oxo) dicopper(III) structure as a model, which means that the O2 molecule is dissociated after adsorption by the Cu centers [56,94,111,114,115]. As suggested by the simulated diffusion area of dynamic Cu species within the CHA unit (the green spherical space shown in the Figure 11b) [56], the transport and pairing of [Cu + (NH3)2] + between CHA cages is the rate-limiting step for Cu-CHA catalysts with low Cu loadings. As described in Section 2.2, due to electrostatic tethering to the framework Al centers, the dynamic diffusion distance of individual Cu-NH3 species is limited to less than 9 Å, so only in the overlapping parts of the diffusion space is it feasible to form Cu dimers for O2 activation. It is speculated that the low-temperature NH3-SCR efficiency at low Cu loadings can be improved by reinforcing the dynamic local motion of Cu-NH3 species and by accordingly promoting the dissociation of O2 [56,57,68].
With the increase of Cu loading, however, formation of Cu dimers is enabled by high Cu density and no longer limited by the long-distance migration process.

Promoting Solid-State Cu Ion Exchange
The dynamics of Cu-NH3 complexes can also be utilized to facilitate Cu ion exchange in solid-state synthesis [116]. Compared with the aqueous-phase method in the synthesis of metal ion-exchange zeolites, solid-state ion-exchange (SSIE) is relatively simple and generates no wastewater and other pollutants. Traditional SSIE synthesis of Cu-CHA catalysts requires thermal treatment of zeolite-copper oxide mixture powders at high temperatures of 700~800 °C for prolonged duration to introduce Cu ions into zeolites. Zeolite frameworks might be damaged during this continuous high-temperature treatment, which restricts the wide application of SSIE. Vennestrøm et al. found that the temperature required to initiate the exchange process can be significantly reduced in NH3-SCR-related atmosphere (such as NH3, NO, etc.) and that the Cu loading in the SSIE-synthesized Cu-

Promoting Solid-State Cu Ion Exchange
The dynamics of Cu-NH 3 complexes can also be utilized to facilitate Cu ion exchange in solid-state synthesis [116]. Compared with the aqueous-phase method in the synthesis of metal ion-exchange zeolites, solid-state ion-exchange (SSIE) is relatively simple and generates no wastewater and other pollutants. Traditional SSIE synthesis of Cu-CHA catalysts requires thermal treatment of zeolite-copper oxide mixture powders at high temperatures of 700~800 • C for prolonged duration to introduce Cu ions into zeolites. Zeolite frameworks might be damaged during this continuous high-temperature treatment, which restricts the wide application of SSIE. Vennestrøm et al. found that the temperature required to initiate the exchange process can be significantly reduced in NH 3 -SCR-related atmosphere (such as NH 3 , NO, etc.) and that the Cu loading in the SSIE-synthesized Cu-zeolites are similar to those obtained by the traditional aqueous-phase ion-exchange method [117]. In the SSIE synthesis of Cu-CHA using Cu 2 O and CuO as precursors, a mixture of NH 3 + NO was found to promote the Cu ion introduction process, especially more pronounced in the case of CuO as a precursor, which indicates that the [Cu + (NH 3 ) x ] + (x ≥ 2) complexes are the main "transport carriers" for Cu ions to migrate into the zeolite framework [116]. They also found that the SSIE rate, i.e., the diffusion rate of transient Cu species, varied in different zeolite structures, in agreement with the simulation results described in Section 2.2. To explore the nature of SSIE-introduced Cu ions and their effects in NH 3 -SCR catalysis, Clemens et al. synthesized a series of Cu-SSZ-13 with different Cu loadings and characterized them using multiple techniques such as HD-XRD, XPS, and DRUVS [118]. They concluded that Cu + and Cu 2+ are present inside zeolite pores at the same time and that there are more Cu + species in the SSIE-synthesized samples than those obtained by the aqueous ion-exchange method. Meanwhile, there are also more CuO species on the external surface of SSIE-synthesized zeolites, which can migrate into the framework pores during NH 3 -SCR to increase NO x conversion. Recently, Vennestrøm et al. verified the formation and dynamic migration of key "transport carriers" [Cu + (NH 3 ) 2 ] + during the NH 3 -SSIE process by XANES [119].

Monitoring NH 3 Storage and Conversion
In addition to their traditional function as catalysts for NH 3 -SCR, zeolites can play another important role as impedimetric sensors for monitoring NH 3 storage and conversion levels. Such application of zeolites, which is based on NH 3 -supported proton transport, has been widely studied and even practically developed [61,120,121]. Precise sensing devices are needed in NH 3 -SCR systems in order to achieve the highest possible NO x conversion without NH 3 slip [122]. The NH 3 -sensing properties of zeolites are of particular interest for collecting real-time information about the storage and conversion levels of NH 3 within SCR catalysts. The NH 3 -storage capacity, for example, decreases with the increase in temperature due to thermal desorption. Consequently, proton conductivity of the zeolite catalyst, which can be measured by impedance spectroscopy, decreases due to a lower content of NH 3 serving as "vehicle-like" molecules for proton transport [123,124]. However, in light of the recently discovered Cu mobility in Cu-CHA, the influence of mobilized Cu-NH 3 complexes must be taken into account. A comparison of Cu-SSZ-13 and Cu-SAPO-34 revealed that the NH 3 -solvated and -mobilized Cu species actually considerably altered the NH 3 -sensing response [109]. In absence of NO, the ionic conductivity signals (I IS ) of both catalysts followed the stepwise drop of NH 3 concentration, demonstrating a semiquantitative relationship between the two variables ( Figure 12a). However, while the overall ion conductivity of Cu-SSZ-13 was stepwise lowered with the increase of the NO/NH 3 ratio in the gas mixture at 350 • C, the formation of highly mobile Cu + -NH 3 complexes appeared to compensate NH 3 consumption by SCR conversion and led to a delayed decrease of I IS signal in the case of Cu-SAPO-34 (Figure 12b). In situ modulus spectroscopy studies confirmed that the prevalence of short-range ion conduction (to which the Cu ions contribute) over the long-range one (dominated by the NH 3 -supported proton transport) was responsible for the unexpected NH 3 -SCR sensing behavior of Cu-SAPO-34. The mechanistic insights into the Cu-CHA catalyst as a sensor implied that tracking the condition-dependent Cu dynamics also provides new perspectives to understand the NH 3 -SCR reaction mechanisms.

Summary and Perspectives
Under low-temperature NH3-SCR reaction conditions, the Cu active sites in Cu-CHA catalysts mainly exist as NH3-solvated and -mobilized complexes, which are different from traditional homogeneous or heterogeneous catalysts. Following the dynamics of these transiently formed Cu-NH3 species can help to reconcile some of the controversies in NH3-SCR catalysis, including the "seagull" profile of NO conversion as a function of temperature and the nonlinear relationship of Cu density and SCR rates. By summarizing recent studies on the unique reaction-driven Cu dynamics in Cu-CHA during low-temperature NH3-SCR catalysis, this minireview aims to inspire new considerations for improving the DeNOx activity of Cu-zeolite catalysts. Theoretical simulations by AIMD, DFT, etc. are essential to unveil the dynamics of Cu active species at the molecular level. According to the simulation results, the diffusion of mobile Cu species is determined by electrostatic tethering generated by the zeolite framework, and the maximum diffusion distance of the Cu-NH3 species is predicted to be about 9 Å. This explains why the zeolite topology significantly influences the local movements of Cu species. While conventional spectroscopic techniques, such as in situ or operando XANES, DRIFTS, DRUVS, EPR, etc., could provide indirect evidence for the NH3 solvation of Cu sites by detecting changes in the Cu coordination environment, the recently applied dielectric measuring technique, namely impedance-based in situ modulus spectroscopy, allows for direct tracking of the relaxation of Cu sites under relevant reaction conditions. Interestingly, the NH3-solvated Cu species have been identified as essential participants not only in low-temperature NH3-SCR catalysis but also in solid-state ion-exchange synthesis and have complicated the sensing properties of Cu-CHA zeolite as a sensor for direct monitoring of NH3 storage and conversion.
In addition to being active centers in NH3-SCR reaction, the dynamically formed, reversible, and temperature-dependent Cu sites have also been found to be potentially important in the selective methane-to-methanol oxidation reaction [125]. In addition, such transiently formed dynamic metal sites were also observed in ethene dimerization reactions over Ni ion-exchanged zeolites and may even exist more broadly in a wide variety of chemical reactions [126]. Although several advanced techniques, such as operando optical spectroscopy or X-ray-based micro-spectroscopy, have been developed to detect, mainly indirectly, the formation of transient active sites, their dynamic nature in zeolite cavity poses a challenge for the design of computational models and precise identification by spectroscopic techniques. Due to the complex dynamics, the transient species cannot be treated either as fixed sites or as unbound free molecules in calculations; therefore, more suitable models for mimicking such reaction systems need to be explored. A more accurate computational model can also help to eliminate the interference of ion mobility

Summary and Perspectives
Under low-temperature NH 3 -SCR reaction conditions, the Cu active sites in Cu-CHA catalysts mainly exist as NH 3 -solvated and -mobilized complexes, which are different from traditional homogeneous or heterogeneous catalysts. Following the dynamics of these transiently formed Cu-NH 3 species can help to reconcile some of the controversies in NH 3 -SCR catalysis, including the "seagull" profile of NO conversion as a function of temperature and the nonlinear relationship of Cu density and SCR rates. By summarizing recent studies on the unique reaction-driven Cu dynamics in Cu-CHA during low-temperature NH 3 -SCR catalysis, this minireview aims to inspire new considerations for improving the DeNO x activity of Cu-zeolite catalysts. Theoretical simulations by AIMD, DFT, etc. are essential to unveil the dynamics of Cu active species at the molecular level. According to the simulation results, the diffusion of mobile Cu species is determined by electrostatic tethering generated by the zeolite framework, and the maximum diffusion distance of the Cu-NH 3 species is predicted to be about 9 Å. This explains why the zeolite topology significantly influences the local movements of Cu species. While conventional spectroscopic techniques, such as in situ or operando XANES, DRIFTS, DRUVS, EPR, etc., could provide indirect evidence for the NH 3 solvation of Cu sites by detecting changes in the Cu coordination environment, the recently applied dielectric measuring technique, namely impedance-based in situ modulus spectroscopy, allows for direct tracking of the relaxation of Cu sites under relevant reaction conditions. Interestingly, the NH 3 -solvated Cu species have been identified as essential participants not only in low-temperature NH 3 -SCR catalysis but also in solid-state ionexchange synthesis and have complicated the sensing properties of Cu-CHA zeolite as a sensor for direct monitoring of NH 3 storage and conversion.
In addition to being active centers in NH 3 -SCR reaction, the dynamically formed, reversible, and temperature-dependent Cu sites have also been found to be potentially important in the selective methane-to-methanol oxidation reaction [125]. In addition, such transiently formed dynamic metal sites were also observed in ethene dimerization reactions over Ni ion-exchanged zeolites and may even exist more broadly in a wide variety of chemical reactions [126]. Although several advanced techniques, such as operando optical spectroscopy or X-ray-based micro-spectroscopy, have been developed to detect, mainly indirectly, the formation of transient active sites, their dynamic nature in zeolite cavity poses a challenge for the design of computational models and precise identification by spectroscopic techniques. Due to the complex dynamics, the transient species cannot be treated either as fixed sites or as unbound free molecules in calculations; therefore, more suitable models for mimicking such reaction systems need to be explored. A more accurate computational model can also help to eliminate the interference of ion mobility on the spectrum and can improve the resolution of spectroscopic identification. It is expected that detailed understanding of dynamic species will help establish more reliable guidelines for the development of low-temperature NH 3 -SCR catalysts or even develop new concepts for a certain class of reaction systems with dynamic metal species as active centers.
Author Contributions: H.L. and V.R. wrote the draft and improved the manuscript based on the reviewers' comments; The manuscript was reviewed and edited by P.C., A.G., D.Y. and U.S. All authors have read and agreed to the published version of the manuscript.

Data Availability Statement:
No data, models, or code were generated or used during the study.

Conflicts of Interest:
The authors have no known conflict of interest to declare.