Iron-Based Composite Oxide Catalysts Tuned by CTAB Exhibit Superior NH 3 –SCR Performance

: Iron-based oxide catalysts for the NH 3 –SCR (selective catalytic reduction of NO x by NH 3 ) reaction have gained attention due to their high catalytic activity and structural adjustability. In this work, iron–niobium, iron–titanate and iron–molybdenum composite oxides were synthesized by a co-precipitation method with or without the assistance of hexadecyl trimethyl ammonium bromide (CTAB). The catalysts synthesized with the assistance of CTAB (FeM 0.3 O x -C, M = Nb, Ti, Mo) showed superior SCR performance in an operating temperature range from 150 ◦ C to 400 ◦ C compared to those without CTAB addition (FeM 0.3 O x , M = Nb, Ti, Mo). To reveal such enhancement, the catalysts were characterized by N 2 -physisorption, XRD (Powder X-ray diffraction), NH 3 -TPD (temperature-programmed desorption of ammonia), DRIFTS (Diffuse Reﬂectance Infrared Fourier Transform Spectroscopy), XPS (X-ray Photoelectron Spectroscopy), and H 2 -TPR (H 2 -Total Physical Response). It was found that the crystalline phase of Fe 2 O 3 formed was inﬂuenced by the presence of CTAB in the preparation process, which favored the formation of crystalline γ -Fe 2 O 3 . Owing to the changed structure, the redox-acid properties of FeM 0.3 O x -C catalysts were modiﬁed, with higher exposure of acid sites and improved ability of NO oxidation to NO 2 at low-temperature, both of which also contributed to the improvement of NO x conversion. In addition, the weakened redox ability of Fe prevented the over-oxidation of NH 3 , thus accounting for the greatly improved high-temperature activity as well as N 2 selectivity.


Introduction
Selective catalytic reduction of NO x by NH 3 (NH 3 -SCR) is an effective means of NO x abatement. Various catalysts have been developed for NH 3 -SCR, including metal oxide catalysts (VO x -based, CeO 2 -based, Fe 2 O 3 -based, MnO x -based catalysts, etc.) and zeolite catalysts (Cu-and Fe-exchanged zeolite catalysts, etc.) [1,2]. Among them, iron-based oxide catalysts have received much attention due to their high catalytic activity as well as non-toxicity, low cost, and accessibility [1,3,4]. As for NH 3 -SCR catalysts, both acid sites and redox sites are necessary to guarantee efficient NO x reduction [1,5]. High dispersion of sites having the same function and close coupling of redox-acid sites are important for the design of catalysts with superior NH 3 -SCR performance [2]. Taking this principle into account, large numbers of iron-based composite oxide catalysts have been prepared and investigated for NH 3 -SCR, including Fe-Ti [6][7][8], Fe-Mo [9][10][11], Fe-W [12][13][14], Fe-Nb [15], and so on, in which the Fe component provides redox sites while a second component supplies acid sites. For instance, Liu et al. [6,16] reported a novel iron titanate catalyst with outstanding performance in the medium-temperature range (200-400 • C), in which a specific Fe-O-Ti structure acted as the main active phase. Recently, Qu et al. [11] identified a di-nuclear entity (an isolated Mo ion and one adjacent surface Fe ion) as the active site in a Mo 1 /Fe 2 O 3 single-atom catalyst, thus giving an explanation for the improved SCR reaction at acid-redox interfaces.
For iron-based oxides, generally, the γ-Fe 2 O 3 crystalline phase shows better activity than α-Fe 2 O 3 in NH 3 -SCR [7,17,18] and other catalytic reactions, such as CO 2 hydrogenation, photodecomposition of H 2 S, and NO reduction by CO [19][20][21]. However, the γ-Fe 2 O 3 active phase is thermally unstable and undergoes irreversible transformation to α-Fe 2 O 3 at elevated temperatures. Such transformation generally occurs between 300 • C and 400 • C [18,[22][23][24], related to several factors, including the particle size, morphology of γ-Fe 2 O 3 and presence of a coating layer or dopants [23,25]. As a result, maintaining the γ-Fe 2 O 3 active phase in iron-based oxide catalysts is of great importance to the durability of catalysts in the NH 3 -SCR reaction. To restrain such phase transformation, previous researchers found it effective to introduce another metal into iron oxides, such as Mn, Ti or W, or to fabricate coating layers [21,22,24,[26][27][28].
Hexadecyl trimethyl ammonium bromide (CTAB) is a kind of amphiphilic organic compound containing both hydrophobic and hydrophilic ends [29], which are often used as template-directing agents that can function as both "structural" and "chemical" promoters. It was reported that CTAB could adjust the crystalline phase of metal oxides [30][31][32]; for instance, the steric hindrance of CTAB was found to prevent the transition of anatase to rutile for TiO 2 [31]. In this study, Nb, Ti and Mo were chosen to couple with Fe in order to provide sufficient acidity for NH 3 -SCR [9,18,[33][34][35][36]. During the preparation process of Fe-M (M = Nb, Ti, Mo) catalysts meanwhile, CTAB was added into the precursor solution.
With the assistance of CTAB, the obtained FeM 0.3 O x -C catalysts showed much better SCR performance than unassisted FeM 0.3 O x catalysts. It was identified that CTAB adjusted the structural properties, contributing to the formation of the γ-Fe 2 O 3 phase. Further investigation revealed that the modified catalysts showed high exposure of acid sites and an enhanced ability for low-temperature NO oxidation, but restrained behavior for NH 3 oxidation at high temperatures, thus contributing to improved NO x conversion and N 2 selectivity.

NH 3 -SCR Performance
The NO x conversion in the NH 3 -SCR reaction over all the prepared catalysts under the high GHSV (gaseous hourly space velocity) of 500,000 h −1 is shown in Figure 1. At 150 • C, the NO x conversion over FeNb 0.3 O x -C was 9%, which was slightly higher than that of FeNb 0.3 O x (7%). With the rising temperature, the SCR activity of FeNb 0.3 O x -C increased significantly, with NO x conversion of 73% at 300 • C, which was much higher than the value of 54% for FeNb 0.3 O x . Such an increase in NO x conversion was also observed over Fe-Ti and Fe-Mo catalysts, especially at higher temperatures. Meanwhile, FeM 0.3 O x -C catalysts showed higher N 2 selectivity than FeM 0.3 O x catalysts, as shown in Figure S1.

Kinetic Studies
The Arrhenius plots of the reaction rates for the reduction of NO x in the range of 200-320 • C are shown in Figure 2. The apparent activation energies (E a ) were then calculated from the fitted curves and are shown in Table 1. It was noted that the E a of FeNb 0.3 O x -C was 34.5 kJ mol −1 , similar to that of FeNb 0.3 O x (40.0 kJ mol −1 ). This was further confirmed by the reaction results of Fe-Ti and Fe-Mo catalysts with similar values of E a (shown in Table 1). Figure 2 also shows that the three FeM 0.3 O x -C catalysts exhibited almost the same reaction rate per square meter of surface area, which is further listed in Table 1 (reaction rate at 260 • C). In agreement with Figure 1, the activities of FeM 0.3 O x -C rate at 260 °C). In agreement with Figure 1, the activities of FeM0.3Ox-C were higher than those of FeM0.3Ox. Among the three FeM0.3Ox samples, interestingly, FeTi0.3Ox and FeMo0.3Ox exhibited similar reaction rates.   rate at 260 °C). In agreement with Figure 1, the activities of FeM0.3Ox-C were higher than those of FeM0.3Ox. Among the three FeM0.3Ox samples, interestingly, FeTi0.3Ox and FeMo0.3Ox exhibited similar reaction rates.     The surface areas and pore volumes of all the samples are summarized and displayed in Table 2. The FeNb 0.3 O x -C catalyst showed a smaller surface area than that of FeNb 0.3 O x (172 m 2 g −1 vs. 210 m 2 g −1 ) while exhibiting a larger average pore size. A decrease in surface area induced by the assistance of CTAB in the preparation process was also observed for Fe-Ti and Fe-Mo, with detailed information shown in Table 2.

XRD Analysis
The XRD patterns in Figure 3 illustrate the crystalline phases of the synthesized samples. Diffraction peaks associated with γ-Fe 2 O 3 appeared for the FeNb 0.3 O x -C catalyst, while no peak due to Nb species was observed. For the FeNb 0.3 O x catalyst, no characteristic peaks due to Fe or Nb components were observed, indicating their high dispersion or amorphous state [37]. For FeTi 0.3 O x , only peaks attributed to α-Fe 2 O 3 could be observed, while for FeTi 0.3 O x -C, besides the peaks of α-Fe 2 O 3 with lower intensity, peaks of γ-Fe 2 O 3 appeared. As for the Fe-Mo catalysts, interestingly, only peaks due to γ-Fe 2 O 3 were observed for the FeMo 0.3 O x -C catalyst, whereas all the peaks were assignable to α-Fe 2 O 3 in the case of FeMo 0.3 O x . As γ-Fe 2 O 3 and other iron oxides (such as Fe 3 O 4 ) show very similar diffraction patterns [38], 57 Fe Mössbauer spectroscopy measurements were carried out to confirm the crystalline phase of the iron oxides present, with results shown in Figure S2. The data were fitted using Moss Winn with parameters including isomer shift (mm s −1 ), quadrupole splitting (mm s −1 ), internal hyperfine field (T), and area listed in Table S1. According to previous research [39][40][41][42][43][44], the isomer shift values of doublets indicated the signals from Fe 3+ for each sample. In other words, no signal of Fe 2+ was observed, which ruled out the possibility of  Table S1 with the value of about 3:1 [44].

Acidic Properties
In the NH 3 -SCR reaction, acid sites on the catalysts are responsible for the adsorption of ammonia and play an important role in NO x reduction. To investigate the surface acidity, NH 3 -TPD experiments were carried out, and the amount of NH 3 desorption that occurred was calculated. For all the catalysts, one peak at around 150 • C and another peak around 300 • C were observed, which correspond to weakly adsorbed and strongly adsorbed ammonia species, respectively [45]. To reduce errors in the determination of the NH 3 desorption amount, the NH 3 -TPD experiments were repeated two times ( Figure S3a,b), with results shown in Figure S3c and Table 3. The total amount of NH 3 desorbed from FeNb 0.3 O x -C was around 260 µmol g −1 , less than that for FeNb 0.3 O x (around 289 µmol g −1 ). After normalization by surface area (that is, the NH 3 adsorption amount divided by the surface area), the calculated NH 3 desorption value was 1.5 µmol m −2 for FeNb 0.3 O x -C, which was larger than that for FeNb 0.3 O x (1.4 µmol m −2 ). For Fe-Ti and Fe-Mo, FeM 0.3 O x -C catalysts also showed lower NH 3 desorption but a higher value after normalization by surface area compared with FeM 0.3 O x catalysts. In addition, the percentages of weakly adsorbed ammonia species were calculated and are shown in Table 3, with the values of FeM 0.3 O x -C catalysts being slightly higher than those of the FeM 0.3 O x catalysts, respectively. around 300 °C were observed, which correspond to weakly adsorbed and strongly adsorbed ammonia species, respectively [45]. To reduce errors in the determination of the NH3 desorption amount, the NH3-TPD experiments were repeated two times ( Figure S3a and S3b), with results shown in Figure S3c and Table 3. The total amount of NH3 desorbed from FeNb0.3Ox-C was around 260 μmol g −1 , less than that for FeNb0.3Ox (around 289 μmol g −1 ). After normalization by surface area (that is, the NH3 adsorption amount divided by the surface area), the calculated NH3 desorption value was 1.5 μmol m −2 for FeNb0.3Ox-C, which was larger than that for FeNb0.3Ox (1.4 μmol m −2 ). For Fe-Ti and Fe-Mo, FeM0.3Ox-C catalysts also showed lower NH3 desorption but a higher value after normalization by surface area compared with FeM0.3Ox catalysts. In addition, the percentages of weakly adsorbed ammonia species were calculated and are shown in Table 3, with the values of FeM0.3Ox-C catalysts being slightly higher than those of the FeM0.3Ox catalysts, respectively.   Figure S3.
To identify the types of acid sites, DRIFT studies of NH3 adsorption at 150 °C were performed and shown in Figure 4. The bands at 1672 cm −1 and 1430 cm −1 were attributed to NH4 + species adsorbed on Brønsted acid sites, and bands at 1606 cm −1 and 1209 cm −1 were ascribed to NH3 species adsorbed on Lewis acid sites [46][47][48]. By comparing the intensity of bands at 1430 cm −1 , one can easily observe that more NH4 + species adsorbed on Brønsted acid sites on FeM0.3Ox-C catalysts than on FeM0.3Ox catalysts. In contrast, as indicated by the intensity of the band at 1606 cm -1 , the amount of NH3 species adsorbed on Lewis acid sites on FeM0.3Ox-C catalysts was slightly lower than that for FeM0.3Ox catalysts.   Figure S3.
To identify the types of acid sites, DRIFT studies of NH 3 adsorption at 150 • C were performed and shown in Figure 4. The bands at 1672 cm −1 and 1430 cm −1 were attributed to NH 4 + species adsorbed on Brønsted acid sites, and bands at 1606 cm −1 and 1209 cm −1 were ascribed to NH 3 species adsorbed on Lewis acid sites [46][47][48]. By comparing the intensity of bands at 1430 cm −1 , one can easily observe that more NH 4 + species adsorbed on Brønsted acid sites on FeM 0.3 O x -C catalysts than on FeM 0.3 O x catalysts. In contrast, as indicated by the intensity of the band at 1606 cm -1 , the amount of NH 3 species adsorbed on Lewis acid sites on FeM 0.3 O x -C catalysts was slightly lower than that for FeM 0.3 O x catalysts.

XPS Analysis
To explore the electronic states and atomic concentrations of surface atoms, XPS analysis was carried out for all the samples. As shown in Figure 5, the O 1 s spectra could be fitted into two peaks, with the peak at around 530.1 eV corresponding to lattice oxygen (denoted as Oβ) and the peak at around 531.4 eV assignable to surface oxygen (denoted as Oα). Compared with FeM0.3Ox catalysts, the Oβ species in FeM0.3Ox-C catalysts showed higher binding energy. Meanwhile, the ratios of Oα/(Oα + Oβ), as well as the proportion of Oα (that is, the value of Oα/(Oα + Oβ) plus the value of the O atom ratio on the surface), were calculated. As listed in Table 4, each of the FeM0.3Ox-C catalysts showed a higher proportion of Oα than the corresponding FeM0.3Ox catalysts, indicative of a higher content of surface O species on FeM0.3Ox-C.
The XPS Spectra of Nb 3d, Ti 2p, and Mo 3d are shown in Figure S4, showing Nb, Ti, and Mo elements in their highest valence state [49][50][51][52][53][54]. In addition, it was observed that the acid components (Nb, Ti, and Mo) in the FeM0.3Ox-C catalysts showed higher binding energy than those in FeM0.3Ox catalysts. The spectra of Fe 2p are shown in Figure S5, with the peaks located around 724.5 eV, 718.9 eV and 710.7 eV, which correspond to Fe 2p1/2, Fe 2p3/2 satellite and Fe 2p3/2, respectively, corresponding to Fe 3+ [45,46]. As shown in Table  4, for a given acid component, the ratio of M/Fe in FeM0.3Ox-C catalysts was always higher than that in FeM0.3Ox catalysts. This result indicated that a surface enrichment of acid components was induced by CTAB addition during the process of catalyst preparation, which was consistent with the results of NH3-TPD.

XPS Analysis
To explore the electronic states and atomic concentrations of surface atoms, XPS analysis was carried out for all the samples. As shown in Figure 5, the O 1 s spectra could be fitted into two peaks, with the peak at around 530.1 eV corresponding to lattice oxygen (denoted as O β ) and the peak at around 531.4 eV assignable to surface oxygen (denoted as O α ). The XPS Spectra of Nb 3d, Ti 2p, and Mo 3d are shown in Figure S4, showing Nb, Ti, and Mo elements in their highest valence state [49][50][51][52][53][54]. In addition, it was observed that the acid components (Nb, Ti, and Mo) in the FeM 0.3 O x -C catalysts showed higher binding energy than those in FeM 0.3 O x catalysts. The spectra of Fe 2p are shown in Figure S5, with the peaks located around 724.5 eV, 718.9 eV and 710.7 eV, which correspond to Fe 2p 1/2 , Fe 2p 3/2 satellite and Fe 2p 3/2 , respectively, corresponding to Fe 3+ [45,46]. As shown in Table 4, for a given acid component, the ratio of M/Fe in FeM 0.3 O x -C catalysts was always higher than that in FeM 0.3 O x catalysts. This result indicated that a surface enrichment of acid components was induced by CTAB addition during the process of catalyst preparation, which was consistent with the results of NH 3 -TPD.

H 2 -TPR Analysis
To further investigate the reducibility of the prepared catalysts, H 2 -TPR experiments were carried out. As shown in Figure 6,

H2-TPR Analysis
To further investigate the reducibility of the prepared catalysts, H2-TPR experiments were carried out. As shown in Figure 6, the peaks below 450 °C correspond to the reduction of Fe2O3 to Fe3O4. Further reduction of Fe3O4 to FeO and Fe occurred at higher temperatures (above 450 °C) [13,55,56]. Over the FeNb0.3Ox-C catalyst, it was observed that the peaks due to the reduction of Fe2O3 to Fe3O4 were centered at higher temperatures (348 °C and 413 °C) compared with FeNb0.3Ox (centered at 312 °C and 397 °C). Such weakening of the redox ability of Fe by CTAB addition was also observed during the reduction of Fe3O4 to FeO and Fe. As for FeTi0.3Ox-C and FeMo0.3Ox-C, similarly, the reduction peaks of Fe2O3 occurred more clearly at higher temperatures compared with those of FeTi0.3Ox and FeMo0.3Ox, respectively.

Direct Oxidation of NH3 and NO
Direct oxidation reactions of NH3 and NO were also conducted, with results shown in Figures S6 and S7, respectively. At temperatures below 250 °C, the direct oxidation of NH3 hardly occurred over any of the samples, benefiting the NH3-SCR reaction. At temperatures above 250 °C, NH3 conversion increased with rising temperature, during which the FeM0.3Ox-C always exhibited lower activity for NH3 oxidation than FeM0.3Ox (except for FeNb0.3Ox-C at the temperature of 400 °C). This was possibly due to the weakened redox ability induced by CTAB addition. By contrast, the direct oxidation of NO occurred over the whole temperature range, during which the NO conversion increased with rising temperature, reached a maximum value, and then decreased. Interestingly, FeM0.3Ox-C always showed higher activity for NO2 formation (except for FeMo0.3Ox-C at temperatures above 250 °C). These results suggested that the intrinsic properties governing the direct oxidation of NO were different from that of NH3 oxidation. Combined with the results listed in Table 4, it can be deduced that surface oxygen plays a crucial role in NO oxidation to NO2 over the Fe-M catalysts [6,57,58].

Direct Oxidation of NH 3 and NO
Direct oxidation reactions of NH 3 and NO were also conducted, with results shown in Figures S6 and S7, respectively. At temperatures below 250 • C, the direct oxidation of NH 3 hardly occurred over any of the samples, benefiting the NH 3 -SCR reaction. At temperatures above 250 • C, NH 3 conversion increased with rising temperature, during which the FeM 0.3 O x -C always exhibited lower activity for NH 3 oxidation than FeM 0.3 O x (except for FeNb 0.3 O x -C at the temperature of 400 • C). This was possibly due to the weakened redox ability induced by CTAB addition. By contrast, the direct oxidation of NO occurred over the whole temperature range, during which the NO conversion increased with rising temperature, reached a maximum value, and then decreased. Interestingly, FeM 0.3 O x -C always showed higher activity for NO 2 formation (except for FeMo 0.3 O x -C at Catalysts 2021, 11, 224 9 of 14 temperatures above 250 • C). These results suggested that the intrinsic properties governing the direct oxidation of NO were different from that of NH 3 oxidation. Combined with the results listed in Table 4, it can be deduced that surface oxygen plays a crucial role in NO oxidation to NO 2 over the Fe-M catalysts [6,57,58].

Discussion
The SCR performance results (Figure 1 and Figure S1) and kinetic studies (Figure 2) showed that FeM 0.3 O x -C catalysts showed higher SCR activity as well as N 2 selectivity than FeM 0.3 O x catalysts over the whole temperature range. It can be concluded from XRD ( Figure 3) and Mössbauer spectra ( Figure S2 and Table S1) that the addition of CTAB into solutions of precursors promoted the formation of γ-Fe 2 O 3 , which was more active in terms of SCR activity than α-Fe 2 O 3 ( Figure S8). Specifically, combining the XRD results with the kinetic studies, it was clear that samples containing the γ-  Figure 1 with those of Figure S8, it can be easily found that, at a given temperature, the activity of the pure γ-Fe 2 O 3 sample was much lower than FeNb 0.3 O x -C and FeTi 0.3 O x -C. These results, in turn, indicate that other factors also have a great influence on the catalytic performance of the Fe-M system.
It is well-accepted that optimizing the acid-redox properties is crucial to designing SCR catalysts with high catalytic performance. In our research, the alteration of the crystalline phase modified both the acidity and reducibility of the Fe-M catalysts, which played a significant role in the improved SCR performance. As for acidity, despite the smaller surface area (Table 2) and lower total desorption amount of NH 3 (Table 3) (Table 4), consistent with higher exposure of acid components. The NH 3 -TPD results ( Figure 4) indicated more weakly adsorbed ammonia on FeM 0.3 O x -C catalysts. As previous research indicated the weaker stability of NH 4 + species adsorbed on Brønsted acid sites compared to NH 3 species adsorbed on Lewis acid sites [59,60], this suggested that more NH 4 + species adsorbed on Brønsted acid sites existed on FeM 0.3 O x -C catalysts. This was further confirmed by the DRIFT results ( Figure 5). In conclusion, increased exposure of acid components, especially Brønsted acid sites, was observed over FeM 0.3 O x -C catalysts, which benefited SCR activity [45,61,62].
For metal oxide NH 3 -SCR catalysts, generally, surface oxygen (O α ) is important for the oxidation of NO to NO 2 , thus promoting the "fast-SCR" reaction at low-temperature [6,57,58]. In addition, the over-oxidation of NH 3 accounted for the decreased NO x conversion at high temperatures [1,13,16,63,64]. In our research, on one hand, the increased percentage of surface O α boosted the oxidation of NO to NO 2 (as shown in Table 4 and Figure S7), thus benefiting the activity at low temperatures. On the other hand, the weaker reducibility of Fe species revealed by H 2 -TPR ( Figure 6) suppressed the over-oxidation of NH 3 over FeTi 0.3 O x -C and FeMo 0.3 O x -C catalysts at high temperatures ( Figure S6), explaining the improved NO x conversion and N 2 selectivity. As a result, the SCR performance of FeM 0.3 O x -C catalysts over the whole temperature range was improved. In addition, the shift of binding energy for O 1 s, Nb 3d, Ti 2p, and Mo 3d in XPS results associated with the addition of CTAB was possibly due to deviation of the electron cloud [65][66][67], which indicated enhanced interaction between Fe and acid components and was related to the observed change in the crystalline phase.  (Table 4) and a higher ability for NO oxidation to NO 2 ( Figure S7). As discussed above, the surface oxygen species were active for the oxidation of NO to NO 2 , thus promoting the "fast-SCR" reaction at low-temperature. With this in mind, it is reasonable that FeTi 0.3 O x -C shows similar intrinsic activity to FeNb 0.3 O x -C, even though the latter exhibits a 100% γ-Fe 2 O 3 phase. As shown in Table 3, the desorption amount of NH 3 over FeMo 0.3 O x -C was slightly lower than that of FeTi 0.3 O x -C, suggesting that a smaller amount of acid sites was available for NH 3 -SCR. Compared with FeTi 0.3 O x -C, FeMo 0.3 O x -C also exhibited a lower proportion of surface oxygen (Table 4), resulting in lower activity for NO oxidation to NO 2 ( Figure S7), which is not beneficial for the occurrence of fast-SCR. Taking these facts into account, it is reasonable that FeMo 0.3 O x -C shows similar intrinsic activity to FeTi 0.3 O x -C, even though the former contains 100% γ-Fe 2 O 3 .

Catalyst Preparation
The CTAB-assisted co-precipitation process was inspired by previous research on FeMnTi catalysts by Wu et al. [68]. Typically, CTAB was first dissolved into deionized water, and then precursors containing Fe(NO 3

Activity Test
The NH 3 -SCR activities were tested in a fixed-bed quartz tube flow reactor with an inner diameter of 4 mm (at GHSV of 500,000 h −1 , the mass of each catalyst was around 60 mg). The feed gas consisted of 500 ppm of NO, 500 ppm of NH 3 , 5 vol % O 2 , balanced by N 2 , with a gas flow rate of 500 mL min −1 . The concentrations of NH 3 , NO, NO 2 , and N 2 O were continually monitored by FTIR spectrometer (IS10 Nicolet) (City, State, Abbr(if has), Country)(Thermo, Waltham, MA USA), which was equipped with a multiple path gas cell (2 m).
The NO x conversion and N 2 selectivity were calculated as follows [69,70]:

Kinetic Study
The apparent activation energy (E a ) and reaction order for NO x reduction were measured in a fixed-bed quartz tube flow reactor with an inner diameter of 4 mm. In this case, the mass of each catalyst was 20 mg, and the conversion of NO x was controlled below 25%. The feed gas composition was 500 ppm NO, 500 ppm NH 3 , 5% O 2 , and N 2 balance. The reaction rate of NO x conversion was calculated as follows: where F NOx is the molar flow rate of NO x , X NOx is the conversion of NO x , W is the weight of the catalyst, and S is the BET (Brunauer−Emmett−Teller) surface area.

Catalyst Characterization
N 2 -physisorption analysis was obtained at 77 K using a Quantachrome Autosorb-1C instrument (Anton Paar, Graz, Austria) at liquid nitrogen temperature. The specific surface areas were calculated by the BET equation in the 0.05-0.30 partial pressure range. The pore volumes and average pore diameters were determined by the BJH method from the desorption branches of the isotherms. Prior to each N 2 -physisorption analysis, the samples were degassed at 300 • C for 3 h. Powder X-ray diffraction (XRD) patterns of the samples were conducted on a Brucker D8 diffractometer (Brucker, Karlsruhe, Germany) with Cu Kα (λ = 0.15406 nm) radiation. The scan range of 2θ range was from 20 • to 90 • with a step size of 0.02 • .
The temperature-programmed desorption of ammonia (NH 3 -TPD) experiments were carried out on a fixed-bed quartz tube flow reactor with an inner diameter of 4 mm. The concentration of NH 3 was continually monitored by an FTIR spectrometer (IS10 Nicolet), which was equipped with a multiple path gas cell (2 m). Prior to each TPD experiment, the 100 mg samples were pretreated in 20% O 2 /N 2 at a flow rate of 300 mL min -1 at 350 • C for 0.5 h, and then cooled down to 50 • C and purged by N 2 for 0.5 h. The samples were then exposed to a flow of 500 ppm NH 3 /N 2 (500 mL min −1 ) at 50 • C for 0.5 h, followed by N 2 purging for 0.5 h. Finally, the temperature was raised to 800 • C in N 2 with the rate of 10 • C min −1 . In situ DRIFTS experiments were performed on an FTIR spectrometer (Nicolet IS50) equipped with an MCT/A detector cooled by liquid nitrogen. Each catalyst was pretreated in 20 vol % O 2 /N 2 at 400 • C for 0.5 h and then cooled down to 150 • C. The samples were exposed to a flow of 500 ppm NH 3 with an N 2 balance for 0.5 h and purged by N 2 for another 0.5 h. XPS measurements were carried out on an X-ray photoelectron spectrometer (Thermo Fisher Scientific K-Alpha, Thermo, Waltham, MA, USA) with Al Kα radiation (1486.8 eV) at an energy resolution of 0.05 eV (Ag 3d 5/2 ). The binding energies of Fe 2p, Nb 3d, Ti 2p, Mo 3d, and O 1 s were calibrated using the C 1 s peak (BE = 284.8 eV) as standard. The temperature-programmed reduction of hydrogen (H 2 -TPR) experiments were carried out on a Auto Chem 2920 chemisorption analyzer (Micromeritics, Aachen, Germany). In a typical measurement, 150 mg of the sample was first, preprocessed in a flow of N 2 with the total flow rate of 50 mL min −1 at 300 • C for 1 h, and then cooled to 50 • C, followed by N 2 purging for another 0.5 h. Then the temperature was linearly increased from 50 to 900 • C at the heating rate of 10 • C min −1 in a flow of 10 vol % H 2 /N 2 (50 mL min −1 ), during which the H 2 consumption was continuously recorded by a thermal conductivity detector (TCD).

Conclusions
In our research, FeM 0.3 O x -C catalysts synthesized with the assistance of CTAB exhibited higher SCR performance compared with FeM 0.3 O x . Characterization revealed that the presence of CTAB in the preparation process of Fe-M (M= Nb, Ti, Mo) composite oxides adjusted the crystalline phase of iron oxides and modified both the reducibility and acidity. The optimized catalysts exposed more surface acid sites, which was beneficial to the SCR activity. In addition, surface oxygen was increased, which benefited the NO oxidation and the "fast SCR" reaction. In addition, the redox ability of Fe was weakened by the addition of CTAB, thus restraining the over-oxidation of NH 3 and improved NO x conversion as well as N 2 selectivity at high temperatures. Overall, both the acidity and reducibility were tuned by CTAB addition during the process of catalyst preparation; thus, FeM 0.3 O x -C catalysts achieved higher catalytic performance over the whole temperature range.  Table S1: Isomer shift (mm s −1 ), quadrupole splitting (mm s −1 ), internal hyperfine field (T), and area of sub-spectra from Mössbauer.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.