Remarkable enhancement in the N2 selectivity of NH3-SCR over the CeNb3Fe0.3/TiO2 catalyst in the presence of chlorobenzene

The simultaneous removal of NOx and dioxins is the frontier of environmental catalysis, which is still in the initial stage and poses several challenges. In this study, a series of CeNb3Fex/TiO2 (x = 0, 0.3, 0.6, and 1.0) catalysts were prepared by the sol–gel method and examined for the synergistic removal of NOx and CB. The CeNb3Fe0.3/TiO2 catalyst exhibits an optimum catalytic performance, with an NOx conversion greater than 95% at 260–380 °C. It also exhibits an optimal CB oxidation activity, in which CB promoted both the NOx conversion and N2 selectivity below 250 °C. Moreover, the more favorable ratios of Ce4+ to Ce3+ and plentiful surface-adsorbed oxygen species are the reasons why CeNb3Fe0.3/TiO2 catalyst has better catalytic activity than other catalysts at the lower temperature. Simultaneously, owing to the modulation of Fe to the redox properties of Ce and Nb, the large number of oxygen vacancies and acid sites was generated, and the CeNb3Fe0.3/TiO2 catalyst is beneficial to NOx reduction and CB oxidation. Furthermore, the results of in situ DRIFTS study reveal the NH3-SCR reactions over CeNb3Fe0.3/TiO2 catalysts are mainly conformed to by the L–H mechanism (< 350 °C) and E-R mechanism (> 350 °C), respectively, and the multi-pollutant conversion mechanism in the synergistic reaction was systematically studied.


Introducton
With the acceleration in the industrialization in China, air pollution has attracted national attention Li et al. 2015). Nonetheless, metallurgical sintering plants, coke ovens, and municipal solid waste incinerators emit nitrogen oxides (NO x ) and dioxins that pose a serious threat to human health and cause some serious environmental problems (Fu et al. 2014;Li et al. 2018;Liu et al. 2013;Wang et al. 2020). Catalytic removal is an effective technology for NO x reduction and/or oxidation of chlorobenzene (CB), thereby reducing air pollution at the source. The individual installations are used for NO x and chlorobenzene treatment, which not only take up a lot of space but are also costly. NH 3 selective catalytic reduction (NH 3 -SCR) was a mature and an effective technology to reduce NO x emissions from stationary sources (Jiang et al. 2019). V 2 O 5 supported on TiO 2 are the most widely used industrial catalysts for SCR with a reaction window of 350-420 °C (Gan et al. 2018b). Gallastegi-Villa reported that the V 2 O 5 /TiO 2 catalysts were used for the simultaneous elimination of NO x and CB, but their high removal rates of NO x and CB pollutants could not be obtained in the same temperature range (Gallastegi-Villa However, vanadium has high toxicity and low N 2 selectivity at high temperatures. Therefore, it is imperative to develop high-efficiency technologies that can simultaneously control the release of dioxins and NO x . Currently, some studies have shown that it is possible to achieve the synergistic removal of NO x and CB with some specific catalysts (Huang et al. 2021;Martin-Martin et al. 2021). In particular, the key to achieve the synergistic removal of NO x and CB was by regulating the redox ability and surface acidity of the catalyst Long et al. 2021). Redox properties play a key role in improving low-temperature SCR and CB oxidation of catalysts; meanwhile, the NH 3 adsorption, NO x reducion, and CB oxidation activity are also well related with the surface acidity over catalysts. Among them, Ce oxide (CeO x ) has good NO x removal efficiency at low temperatures, due to the excellent oxygen storage capacity and high redox ability of cerium (Lee et al. 2019;Lei et al. 2020). Furthermore, some researchers have applied Ce-based catalysts in the field of CB removal. Gan reported that the MnO x -CeO 2 composite oxide exhibited good NH 3 -SCR and CB oxidation activity due to the good redox properties and surface acidity (Gan et al. 2018a). The CB removal efficiency reached 100% at 300 °C. However, the free Cl forms metal chlorides with the metal sites leading to catalyst deactivation. Weng prepared Mn x Ce 1−x O 2 /H-ZSM5 catalysts by loading MnO x -CeO 2 with H-ZSM5, which exhibited 90% CB removal efficiency near 250 °C (Sun et al. 2016). Furthermore, H-ZSM5 has abundant Brønsted and Lewis acid sites to promote oxidative activity and resistance to the chlorine poison. Therefore, the activity of the catalyst could be improved by adjusting the solid-phase structure and solid acidity of the catalyst. Niobium oxide (Nb 2 O 5 ) can be used as an active component and carrier in catalysts (Tanabe 2003;Zhao et al. 2017b). Zhao introduced Nb into Cu-BEA to prepare Cu-BEA-Nb catalyst to significantly improve the catalyst surface acidity (Zhao et al. 2017b). Ali reported that Nb additives could improve the Lewis acid and Brønsted acid sites of catalysts simultaneously (Ali et al. 2018). In addition, the metal-supported TiO 2 catalyst exhibits good CB removal efficiency due to the oxygen vacancies (V o ) on the TiO 2 surface (Cheng et al. 2020;Martin-Martin et al. 2021). Li showed that the Pd 0.12 V 4 /TiO 2 catalyst exhibited good synergistic CB/NO removal conversion and target product selectivity . The O vacancy (TiO x−1 ) generated by the reaction process promoted the CBCO and NH 3 -SCR reactions. In particular, Xia reported that Fe-Ti mixed oxides exhibited good activity and stability, attributed to the structure-activity relationship established between Ti 3+ -V o and CB oxidation (Xia et al. 2021). Nevertheless, there are few studies on the catalytic performance of CeNbFe/TiO 2 catalysts prepared by sol-gel method for the synergistic removal of NO x and CB. Therefore, in this study, a series of CeNb 3 Fe x /TiO 2 (x = 0, 0.3, 0.6, and 1.0) catalysts were prepared by the sol-gel method and investigated for the synergistic removal of NO x and CB. The effect of Fe doping on the performance of the CeNb 3 /TiO 2 catalyst for simultaneous removal of NO x and CB and its mechanism were systematically studied.

Experimental
The information on a detailed description of the characterization techniques used in the passage is given in the supplementary materials.

Catalyst preparation
The series CeNb 3 Fe x /TiO 2 (x = 0, 0.3, 0.6 and 1.0) catalysts were prepared by the sol-gel method, and the molar ratio of Ce, Nb, Fe, and Ti was 1:3 First, cerium nitrate, niobium chloride, and ferric nitrate were added in the designed ratio to prepare the precursor solution A. Next, ethanol, acetic acid, and tetrabutyl titanate were added to prepare precursor solution B. Subsequently, solution B was slowly added into solution A to form the gel. The resulting a gel was dried at 110 °C, calcined at 450 °C, and ground for further use.

Catalyst activity
The catalytic performance for NO x and CB removal was examined in a fixed-bed quartz tube flow reactor in the steady-flow mode. Under atmospheric pressure, 0.3 g of 40-60 mesh catalyst was placed between glass wool plugs into a fixed-bed quartz tube reactor (inner diameter = 9 mm). The feed gas composition was 500 mg/L of NO, 500 mg/L of NH 3 , 50 mg/L of CB, and 5 vol.% of O 2 , with N 2 as the balance gas. Under ambient conditions, the space velocity (SV) was 250,000 mL/(g·h). The experiment was performed at 60-500 °C. At each temperature, when the reaction reached a steady state (at least 40 min), the corresponding data were collected.
The removal efficiencies of NO x , chlorobenzene, and N 2 selectivity were defined by Eqs. (1)-(3), respectively: where NO x contains NO and NO 2 . C in and C out are the related concentrations of the inlet and outlet gas phase, e.g., NO x , NH 3 , N 2 O, and CB. Figure 1 shows the results of NO x conversion, CB conversion, and N 2 selectivity over CeNb 3 Fe x /TiO 2 (x = 0, 0.3, 0.6, and 1.0) catalyst. After the NO x conversion rate of the catalyst first increased, it decreased later due to the limitation by the thermodynamics and kinetics of the NH 3 -SCR reaction ( Fig. 1a) . The denitrification efficiency of the

Catalytic efficiency
(2) catalysts between 220 and 420 °C exhibited excellent results. Notably, after the addition of Fe, the NO x removal efficiency was significantly improved at all the reaction temperature ranges. Especially for CeNb 3 Fe 0.3 /TiO 2 , these catalysts exhibited highest catalytic activity, and the NO x conversion rate exceeded 90% at 250-400 °C. However, at the Fe/Ce molar ratio of 1, the activity at 350-500 °C decreased significantly that excess Fe can inhibit the SCR activity over the CeNb 3 / TiO 2 catalyst. Figure 1 b shows the results of N 2 selectivity. Clearly, the N 2 selectivity over the CeNb 3 Fe x /TiO 2 (x = 0.3, 0.6, and 1) catalyst was higher than that over CeNb 3 /TiO 2 , indicative of its good SCR performance. With the increasing of temperature, the conversion rate of CB over different catalysts increased (Fig. 1c). CeNb 3 Fe 0.3/ TiO 2 exhibited the highest CB conversion rate in the entire active window. However, with the increase in the Fe content, the CB conversion rate gradually decreases further confirming that excess Fe can inhibit the CB catalytic oxidation of CeNb 3 Fe x /TiO 2 catalysts. Figure 2 shows the effect of CB on the NO x conversion and N 2 selectivity over the CeNb 3 Fe 0.3 /TiO 2 catalyst. The presence of CB considerably affected the activity of the catalyst. In Fig. 2 a, by the introduction of CB into the flue gas at a temperature of greater than 250 °C, the NO x conversion rate of the catalyst started to decrease, with only a maximum decrease of 4.3% at 340 °C. Notably,the N 2 selectivity over the CeNb 3 Fe 0.3 /TiO 2 catalyst at low temperature was promoted by the process of CB oxidation Fig. 1 a NO x conversion, b CB conversion, and c N 2 selectivity over CeNb 3 Fe x /TiO 2 (Fig. 2b). The results revealed that CB exerts an inhibitory effect on the SCR activity over the CeNb 3 Fe 0.3 /TiO 2 catalyst at temperatures greater than 250 °C, but with the increase in the reaction temperature, the catalytic performance of CB can be improved.
In addition, the effect of internal/external diffusion on the catalysts was investigated. The effect of vertical diffusion was eliminated or mitigated by adding quartz wool to the fixedbed quartz reactor to fix the catalyst. Moreover, the particle sizes of the catalysts tested were all 40-60 mesh, which avoided the effect of internal diffusion. On the other hand, to find out the effect of external diffusion, the performance of the catalyst (CeNb 3 Fe 0.3 /TiO 2 ) was tested by halving or doubling the amount of catalyst and gas flow rate at the same SV (250,000 mL/(g·h)). Figure S1 shows the NO x conversion and CB conversion of the CeNb 3 Fe 0.3 /TiO 2 catalyst under different reaction conditions. As can be seen in Fig. S1, there was a good overlap of the test results, which indicated that external diffusion had almost no effect on the reactivity of the catalyst.

XRD and Raman
Figure 3 a shows the XRD patterns of the CeNb 3 Fe x /TiO 2 catalyst. Characteristic diffraction peaks of anatase TiO 2 (PDF#71-1169), rutile TiO 2 (PDF#75-1748), and Nb 2 O 5 (PDF#26-0885) can be detected in the CeNb 3 /TiO 2 catalyst. The absence of characteristic diffraction peaks of CeO x species may be related to highly dispersed or amorphous phase. Compared with CeNb 3 /TiO 2 , only rutile TiO 2 peaks (PDF#75-1748) and CeO 2 peaks were present in the Fe-containing catalysts. This indicated that the addition of Fe inhibited the formation of anatase TiO 2 phase and promoted the formation of rutile TiO 2 phase. The Fe and Nb species were not detected in the CeNb 3 Fe x /TiO 2 , indicating that Fe and Nb oxide species are highly dispersed or exist in an amorphous form, demonstrating the strong interaction between Ce, Nb, and Fe oxide caused by the introduction of Fe. Figure 3 b shows the Raman spectra of the CeNb 3 Fe x / TiO 2 catalyst. The Raman spectrum of the Fe-containing catalysts revealed a characteristic peak near 462 cm −1 , corresponding to the F 2g vibration mode of the octahedron around the CeO 2 cubic fluorite structure (Fan et al. 2020). In addition, the CeNb 3 Fe x /TiO 2 catalysts exhibited similar peaks at 509 (A 1g ) and 632 cm −1 (E g ), while the CeNb 3 /TiO 2 catalyst only a peak at 632 cm −1 (E g ) Tian et al. 2012). It exhibited a typical rutile TiO 2 phase. For CeNb 3 / TiO 2 catalyst, the intensity of rutile TiO 2 phase characteristic peak was significantly lower than that of Fe-containing Fig. 2 Interaction effects of CB and SCR performance on a NO x conversion and CB conversion over the CeNb 3 Fe 0.3 /TiO 2 catalyst and b N 2 selectivity

Fig. 3 XRD patterns (a) and
Raman spectra (b) of the the CeNb 3 Fe x /TiO 2 catalysts. This indicated that the addition of Fe promoted the formation of rutile TiO 2 phase. Notably, the wavenumber of TiO 2 with Fe-containing catalyst shifts to a higher direction compared with that of CeNb 3 /TiO 2 , possibly related to the rearrangement of the electron configuration. Moreover, a weak band corresponding to oxygen vacancies at 245 cm −1 was observed (Yao et al. 2014). This suggests that the addition of Fe can promote the formation of oxygen vacancies, thus facilitating the transportation of O species (Ma et al. 2020a). In addition, NO can be oxidized by reactive oxygen species to produce NO 2 (Bertinchamps et al. 2005;Zhang et al. 2015). The formed NO 2 is a stronger oxidant than O 2 , thus promoting the oxidation reaction of CB.

N 2 physisorption
To further investigate the structural characteristics of the CeNb 3 Fe x /TiO 2 catalyst, N 2 adsorption-desorption isotherm results were recorded (Fig. S2). All samples exhibited a type IV isotherm, which is related to the capillary condensation that occurs in the mesopore. Table 1 summarizes the specific structure information. The specific surface area of these samples decreased in the order of CeNb 3 /TiO 2 > CeNb 3 Fe 0.3 / TiO 2 > CeNb 3 Fe 0.6 /TiO 2 > CeNb 3 Fe/TiO 2 . Although CeNb 3 / TiO 2 has largest specific surface area among all the samples, its catalytic activity is relatively poor due to its single active component.

SEM and TEM
The structural morphologies of different samples were visualized by SEM and TEM. The results show that the catalyst has an irregular surface and porous multi-metal oxides (Fig. S3). With the addition of Fe in the CeNb 3 /TiO 2 catalyst, the catalyst exhibited a small number of non-porous blocked structures over the surface with fewer pore structures, reducing the reaction units on the catalysts surface. For the TEM results of CeNb 3 Fe x /TiO 2 ( Fig. 4a-d), with the increase of the amount of incorporated Fe, the iron species were loaded on the catalyst surface, thereby reducing the pore size and increasing the catalyst particle size. This result was consistent with the N 2 -BET and catalytic performance. In Fig. 4 e, the TEM image of CeNb 3 Fe 0.3 /TiO 2 clearly showed that the lattice fringes with an interplanar distance of 3.260, 3.218, and 1.695 nm can be assigned to the (112), (110), and (211) plane of CeO 2 and TiO 2 , respectively. For CeNb 3 Fe 0.3 /TiO 2 , the lattice fringes with an interplanar distance of 1.953, 3.124, and 3.622 nm can be assigned to the (420), (400), and (101) plane of Nb 2 O 5 and TiO 2 , respectively. Therefore, the highly dispersed Fe and Nb species in CeNb 3 Fe x /TiO 2 catalysts should be considered the major reason for the difference in catalytic performance.

XPS
The surface states and atomic concentrations of the catalysts were investigated by X-ray photoelectron spectroscopy (XPS), and the corresponding results are shown in Fig. 5 and Table 1. To distinguish and understand the surface species, by fitting Gaussian peaks after Shirley background subtraction, the XPS spectra of Ce 3d, Nb 3d, Fe 2p, and O1s were deconvolved into several peaks. As displayed in Fig. 5 a, the XPS of Ce 3d exhibited eight components, corresponding to four pairs of spin-orbit doublets (Perret et al. 2014;Romeo et al. 1993). The peaks of Ce 3d 5/2 at 882.4 (v), 889.1 (v″), and 898.5 (v‴) eV corresponded to Ce 4+ species, and the peak at 885.7 (v′) and 903.7 (u′) eV indicates that the initial electronic state of 3d 10 4f 1 corresponded to Ce 3+ species (Gomez et al. 2013). Similarly, the Ce3d 3/2 peaks at Ce 901.0 (u), 907.3 (u″), and 916.8 (u‴) eV corresponded to Ce 4+ species (Perret et al. 2014;Yao et al. 2017). Clearly, CeNb 3 Fe x /TiO 2 comprised a mixture of Ce 3+ and Ce 4+ oxidation states over the catalyst surface. Meanwhile, by calculating the Ce 3+ /(Ce 3+ + Ce 4+ ) ratio of different samples, the Ce 3+ /(Ce 3+ + Ce 4+ ) ratio of CeNb 3 Fe 0.3 / TiO 2 (0.44) was found to be greater than that of CeNb 3 Fe 0.6 / TiO 2 (0.32) and CeNb 3 Fe/TiO 2 (0.28), indicating the presence of more oxygen vacancies on the CeNb 3 Fe 0.3 /TiO 2 surface. These unsaturated chemical bonds caused by Ce 3+ could help for denitrification and CB decomposition activity (Ma et al. 2020a). Furthermore, the presence of oxygen vacancy is beneficial to the dissociation of the NO and CB , thereby promoting the NO x reduction and CB oxidation. Figure 5 b shows the deconvolution XPS spectra of Fe 2p for different catalysts. Bands characteristic of Fe2p 1/2 and Fe2p 3/2 were observed at 724 eV and 711 eV, respectively (Liu et al. 2017). Gauss fitting revealed that the peaks observed at 709.1-709.3 eV are assigned to typical Fe 2+ cations, while peaks at 712.2-713.3 eV correspond to characteristics Fe 3+ (Fan et al. 2020). From the above results, iron in our sample was present as Fe 2+ and Fe 3+ . The Fe 3+ / (Fe 3+ + Fe 2+ ) of the CeNb 3 Fe 0.3 /TiO 2 (0.43) catalyst was slightly greater than that of CeNb 3 Fe 0.6 /TiO 2 (0.41) and CeNb 3 Fe/TiO 2 (0.39), indicating the introduction of an appropriate amount of Fe can promote the formation of Fe 3+ content on the surface. According to previously reported study, Fe 3+ species could promote the SCR reaction (Liu et al. 2009). Therefore, this may be one of the reasons for the good SCR activity of the CeNb 3 Fe 0.3 /TiO 2 catalyst.
The Nb 3d and Ti 2p XPS spectra were deconvoluted into two separate peaks ( Fig. 5c and Fig. S4). The spin-orbit splitting peaks at 206.9-207.2 eV and 209.7-210.1 eV corresponded to Nb3d 3/2 and Nb3d 5/2 , respectively ( Fig. 5c) (Ali et al. 2018;Qu et al. 2013). The Nb 5+ /(Nb 5+ + Nb 4+ ) of the CeNb 3 Fe x /TiO 2 catalysts was greater than that of the CeNb 3 /TiO 2 catalyst. Compared with the CeNb 3 /TiO 2 sample, the binding energy of the CeNb 3 Fe x /TiO 2 shifted to a high value, indicating that the mixing of Fe and Nb oxides together changes the chemical environment around the Nb species. As displayed in Fig. S4, after the introduction of Fe, the binding energy increased to higher values, with an increase to 458.6 and 464.2 eV, and Ti still existed as Ti 4+ (Du et al. 2020), indicating that there is a strong interaction between Ce, Nb, Fe, and Ti species. Combining the Ce 3d, Fe 2p, and Nb 3d XPS results (Table 1), Fe-doping may cause redox cycling (Ce 3+ + Nb 4+ = Ce 3+ + Nb 5+ and Fe 3+ + Ce 3+ = Fe 2+ + Ce 4+ ) to promote the generation of oxygen vacancies on the catalyst surface, thereby improving the catalytic activity of NO x reduction and CB oxidation to a certain extent. The related studies indicated that O 2 2− or O − forms O β with a higher mobility and a better oxygen carrying capacity than those of O α , indicating that a catalyst with O β than O α can clearly improve its catalytic activity (Jawaher et al. 2018). Therefore, the high O β ratio is beneficial for the oxidation of NO by CB and oxidation to NO 2 in the SCR reaction, promoting "fast SCR": NO + NO 2 + 2NH 3 = 2N 2 + 3H 2 O (Ma et al. 2020b;Yan et al. 2020). The change trend of O β /(O α + O β ) decreased in the order of CeNb 3 Fe 0.3 / TiO 2 (0.43) > CeNb 3 /TiO 2 (0.40) > CeNb 3 Fe 0.6 /TiO 2 (0.39) > CeNb 3 Fe/TiO 2 (0.37) (Table 1). Therefore, the high active oxygen level of CeNb 3 Fe 0.3 /TiO 2 can promote the oxidation of chlorinated aromatic compounds and the NH 3 -SCR reaction.

H 2 -TPR analysis
H 2 -TPR was employed to investigate the effect of Fe doping on the reducibility of the CeNb 3 /TiO 2 catalyst (Fig. 6). The CeNb 3 /TiO 2 exhibits four TPR signals, which can be assigned to the reduction of the high dispersed surface CeO 2 (460 °C) (Jawaher et al. 2018), Nb 2 O 5 to NbO 2 (535 °C) (Lian et al. 2019), the reduction of the Ce-O-Ti species (645 °C) (Zhao et al. 2017a) and the reduction of the bulk CeO 2 (801 °C), respectively (Fan et al. 2020). After addition of Fe, the peak shape of CeNb 3 Fe x /TiO 2 changed considerably. The new reduction peak at 504 °C for the CeNb 3 Fe/TiO 2  (Xu et al. 2015). In addition, it can be observed that the reduction peak of surface CeO 2 → Ce 2 O 3 and Nb 2 O 5 → NbO 2 obviously shifts to low temperature when Fe ions are added to the CeNb 3 /TiO 2 catalyst, which may be related to better synergy between Ce, Nb, and Fe. Moreover, the temperatures of the surface CeO 2 and Nb 2 O 5 reduction peaks of CeNb 3 Fe 0.3 /TiO 2 catalyst are lower than those of CeNb 3 Fe x / TiO 2 (x > 0.3) catalysts, indicating that CeNb 3 Fe 0.3 /TiO 2 has better reduction activity. Considering that the CeNb 3 Fe 0.3 / TiO 2 catalyst has a lower reduction temperature, it exhibits the best reduction performance, which is obviously the main factor affecting the catalytic activity.

NH 3 -TPD analysis
As the major processes in the NH 3 -SCR reaction, NH 3 adsorption and activation play key roles in the generation of reactive intermediates. NH 3 was used as a probe molecule to detect surface acid sites, acid strength, and acid content. As shown in Fig. 7, the NH 3 -TPD curves exhibited a wide range of desorption temperatures, indicative of the presence of different acid sites with different intensities. Current studies have reviewed the desorption signals of physically adsorbed ammonia and NH 4+ adsorbed on the weak acid sites at 100-200 °C, as well as the desorption signals of NH 3 on the catalyst surface at the acid site at 200-350 °C and the strong acid site at 350-600 °C (Kim et al. 2018;Wang et al. 2019). Based on the desorption peak area, the percentages of acid sites with different intensity distributions and the desorption temperature of desorption peak were calculated for the CeNb 3 Fe x /TiO 2 catalysts (Table 2). Compared to CeNb 3 / TiO 2 , the introduction of Fe increased the ammonia desorption amount at the lower and medium temperature (A1 and A2) which corresponded to weak acid sites. The results indicate that the CeNb 3 Fe x /TiO 2 catalyst comprises a higher number of Bronsted acids after loading with Fe. The adsorbed ammonia coordinated to the Bronsted acid site exhibited a lower thermal stability compared to the Lewis acid site . Nevertheless, after the addition of Fe, the total amount desorbed ammonia (A total ) decreased slightly, and the ammonia desorption peak that corresponded to the strong acid site at the high temperature (T 3 ) shifted to a low temperature range. This phenomenon indicates that the addition of Fe leads to the partial destruction or occupation of the partial acid sites on the CeNb 3 /TiO 2 catalyst, as well as to the transformation of some strong acid sites to intermediate acid sites (Zi et al. 2019). This may result from that the introduction of Fe can improve the redox ability of the catalyst slightly, and low-medium acid can effectively promote the adsorption process (Liu et al. 2016;Shu et al. 2020). Compared to strong acid sties, medium-low strong acid sites can prevent excess oxidation performance, thereby inhibiting N 2 O formation and promoting NO x conversion (Niu et al. 2016). In conclusion, after the addition of Fe, the NO x and CB removal efficiencies increased (Sec 3.1 for evaluation results), which is related to the increase in the weak acid and middle acid. Therefore, CeNb 3 Fe 0.3 /TiO 2 catalyst has the largest amount of low-and medium-temperature acid sites, which explains why the catalyst exhibits the best reactivity.

Adsorption of NH 3
To further distinguish between Lewis and Brønsted acid types, in situ DRIFTS was conducted to investigate the adsorption/desorption behavior of NH 3 over the catalyst. Figure 8 shows the DRIFT spectra of NH 3 adsorption on CeNb 3 Fe 0.3 /TiO 2 catalyst at different reaction temperatures. The peaks at 1654 cm −1 and 1210 cm −1 could be assigned to the adsorption of NH 3 on Lewis acid sites (Zawadzki and Wiśniewski 2003), while peaks at 1595 cm −1 and 1403 cm −1 to the adsorption of NH 3 on Brønsted acid sites. Several peaks were observed at 3360, 3260, and 3135 cm −1 in the N-H region that belonged to the coordinated NH 3 on the Lewis acid sites (Brandenberger et al. 2009;Wang et al. 2013). As the temperature increased, the intensity of the band at the Lewis acid site decreased significantly compared to the band at the Brønsted acid site. This suggests that the NH 3 bound to the Brønsted acid site is more stable than the ammonia bound to the Lewis acid site.

Co-adsorption of NO and O 2
The variation of adsorption/desorption of NO + O 2 over CeNb 3 Fe 0.3 /TiO 2 with the increasing of temperature was studied using DRIFTS (Fig. 9). The band at 1576 and 1383 cm −1 was the bridging nitrate and bidentate nitrate species (Brandenberger et al. 2009;Wang et al. 2013), while the band at 1360 and 1122 cm −1 were the monodentate nitrate species (Zawadzki and Wiśniewski 2003). Similarly, the bands at 1655 cm −1 belonged to the absorbed NO 2 species (Chen et al. 2010b;Larrubia et al. 2001). The adsorbed NO 2 content gradually decreased with the increasing temperature, which was related to the bridging nitrate and bidentate nitrate species. This indicates that the bridging (1576 cm −1 ) and bidentate nitrates (1383 cm −1 ) are unstable, and their intensity decreases with the increasing temperature.

Reactions between NO + O 2 and pre-adsorbed NH 3
The effect of the adsorption of NH 3 species over CeNb 3 Fe x / TiO 2 in the NH 3 -SCR reaction was investigated, and the in situ DRIFTS for the reaction of NO + O 2 and pre-adsorbed NH 3 at 250 °C were recorded (Fig. 10). Figure 10 a shows the different NH 3 species with Lewis acid sites and Brønsted acid sites on the CeNb 3 /TiO 2 surface after exposure to NH 3 for 60 min. NH 3 coordination peaks (1558 and 1223 cm −1 ), NH 4 + coordination peaks (1409 cm −1 ), and several N-H peaks (3336, 3254, and 3155 cm −1 ) were observed for the CeNb 3 /TiO 2 catalyst (Zawadzki and Wiśniewski 2003). After the introduction of NO + O 2 , some bands of surface nitrate substances were observed, namely bidentate and bridging nitrates (1538, 1303 cm −1 ), monodentate nitrates (1130 cm −1 ), and absorbed NO 2 species (1610 cm −1 ) (Brandenberger et al. 2009;Wang et al. 2013). The results revealed that the NH 4 + on the Brønsted acid sites and the coordinated NH 3 on the Lewis acid sites are the main active sites in the NH 3 -SCR reaction over CeNb 3 /TiO 2 . For CeNb 3 Fe x /TiO 2 catalysts, the same results were also gotten as shown in Fig. 10 b-d, respectively. However, the characteristic peaks of the coordinated NH 3 disappeared after NO + O 2 were introduced into the IR cell for 3 min over the CeNb 3 Fe 0.3 /TiO 2 catalyst, while the corresponding bands were still prominent over the CeNb 3 /TiO 2 catalyst. The above results indicate that the addition of Fe species accelerate the reaction between surface NO and the adsorbed NH 3 species.

Reactions between NH 3 and pre-adsorbed NO + O 2
On the other hand, the role of adsorbed NO x species in the NH 3 -SCR reaction over CeNb 3 /TiO 2 , CeNb 3 Fe 0.3 /TiO 2 , CeNb 3 Fe 0.6 /TiO 2 , and CeNb 3 Fe/TiO 2 was also studied. The in situ DRIFT spectra of the reaction between NH 3 and preadsorbed NO x species over the three catalysts at 250 °C are shown in Fig. 11. As can be seen in Fig. 11 a, when the NO + O 2 pretreated catalyst was exposed to NH 3 , the decrease in the intensity of the NO 2 band at 1648 cm −1 and the shift of the band at 1564 cm −1 to 1592 cm −1 were observed, which could be attributed to the deformation of the nitrate species. Meanwhile, the bands belonging to the asymmetric and symmetric stretching patterns of NH 4 + species bound to Brønsted acid sites were observed in the range 1672 and 1423 cm −1 (Chen et al. 2010b;Larrubia et al. 2001), and several peaks at 1592, 3346, 3249, and 3154 cm −1 ascribed to NH 3 -coordinated Lewis acid sites were observed (Zawadzki and Wiśniewski 2003). After the introduction of NH 3 , the bands at 1343, 1378, and 1648 cm −1 vanished in 3 min. The results indicate that the absorbed NO x species can participate in the NH 3 -SCR reaction over CeNb 3 /TiO 2 . A similar phenomenon could also be observed on CeNb 3 Fe 0.3 /TiO 2 , CeNb 3 Fe 0.6 /TiO 2 , and CeNb 3 Fe/TiO 2 as displayed in Fig. 11 b-d, respectively. Interestingly, the band centered at 1573 cm −1 disappeared in 5 min, and the band intensity of the coordinated NH 3 bound to Lewis acid sites (1575 cm −1 ) increased gradually over CeNb 3 Fe 0.3 /TiO 2 , which is the highest of all the samples. From the results of in situ DRIFTS, we can conclude that promoted adsorption of NH 3 species over CeNb 3 Fe 0.3 /TiO 2 should also be responsible for the NH 3 -SCR reaction. Figure 12 shows in the situ DRIFTS for the co-adsorption of NH 3 + NO + O 2 on all samples at different temperatures. For both catalysts below 350 °C, numerous bands coordinated NH 3 on the Lewis acid sites (3389-3043 cm −1 ) in the N-H region were observed (Xiong et al. 2013). The bands at 1666-1718 cm −1 corresponded to the σ s and σ as of NH 4 + at the Brønsted acid site (Chen et al. 2010a;Jiang et al. 2018). The band at 1225-1240 cm −1 could be ascribed to overlapping of coordinated NH 3 and bridging nitrate. The band at 1599-1614 cm −1 belonged to the overlapping of bands of NH 3 and coordinated NO 2 . For CeNb 3 Fe 0.3 /TiO 2 , a new peak appeared at 1715 cm −1 compared with CeNb 3 /TiO 2 at below 200 °C (Lei et al. 2020). The new peak was ascribed to the surface nitrosyl of Fe 3+ -NO type (1715 cm −1 ) . In case of CeNb 3 /TiO 2 , NH 3 -related band at 1599 cm −1 decreased with the increasing of temperature, while the nitrate-related band at 1228 cm −1 increased in intensity and persistent even at 350 °C. However, the band at 1614 cm −1 on the CeNb 3 Fe 0.3 /TiO 2 catalyst disappeared Fig. 10 In situ DRIFT spectra of NO + O 2 and pre-adsorbed NH 3 species at 250 °C of a CeNb 3 /TiO 2 , b CeNb 3 Fe 0.3 / TiO 2 , c CeNb 3 Fe 0.6 /TiO 2 , and d CeNb 3 Fe/TiO 2 Fig. 11 In situ DRIFT spectra of NH 3 and pre-adsorbed NO + O 2 species at 250 °C of a CeNb 3 /TiO 2 , b CeNb 3 Fe 0.3 / TiO 2 , c CeNb 3 Fe 0.6 /TiO 2 , and d CeNb 3 Fe/TiO 2 Fig. 12 In situ DRIFT spectra of 1000 mg/L of NH 3 + 1000 mg/L of NO + 5 vol.% O 2 at various temperatures over a CeNb 3 /TiO 2 , b CeNb 3 Fe 0.3 /TiO 2 , c CeNb 3 Fe 0.6 / TiO 2 , and d CeNb 3 Fe/TiO 2 with increasing temperature, while the nitrate-related band decreased with increasing temperature, and no band was observed at 350 °C. These results point out that the CeNb 3 Fe 0.3 /TiO 2 catalyst shows much better NH 3 -SCR activity than CeNb 3 /TiO 2 .

Reaction of NH 3 + NO + O 2
For the NH 3 -SCR reaction, the Langmuir-Hinshelwood (L-H) and Eley-Rideal (E-R) mechanisms are usually proposed at low and high temperatures, respectively (Chen et al. 2010a;Jiang et al. 2018). A simplified reaction mechanism is proposed over CeNb 3 Fe 0.3 /TiO 2 catalyst. The adsorption of NH 3 on the catalyst surface is the initial step of the NH 3 -SCR reaction. In the NH 3 -SCR reaction below 350 °C, the adsorbed NH 3 species undergo L-H reaction with the adsorbed bridging nitrate and bidentate nitrate. Above 350 °C, the liganded NH 3 adsorbed on the catalyst surface is further oxidized to NH 2 species. However, no band of NH 2 species is detected on the catalyst surface because of its rapid decomposition. The residual adsorbed NH 3 indicated reacts with adsorbed NO to form N 2 and H 2 O, and the reaction follow the E-R mechanism.

Reaction of NH 3 + NO + O 2 + CB
In order to gain insight into the multi-pollutant conversion mechanism in the synergistic reaction, in situ DRIFTS was subsequently measured at 250 °C as well as in situ DRIFTS at different temperatures. As shown in Fig. 13, a series of IR bands appeared after CB, NH 3 , NO, and O 2 adsorptions. Several vibration bands could be observed in the range of 1100-2000 cm −1 , including bidentate nitrates (1550 and 1578 cm −1 ), and monodentate nitrates (1295 and 1384 cm −1 ). The peak centered at 1110 cm −1 was assigned to symmetric bending vibrations of NH 3 coordinated with Lewis acid sites, while the band at 1695 cm −1 could be attributed to the asymmetric bending vibration of N-H bond in NH 4 + species on Brønsted acid sites. The peak centered at 1245 cm −1 was assigned to C-O stretching of phenolate . The band at 1335 cm −1 was the carbonate bidentate species (Zasada et al. 2017), while the band at 1364 cm −1 was the maleate . Similarly, the band at 1795 cm −1 belonged to the chlorinated species (Long et al. 2021). The band at 1470 cm −1 detected on catalysts was assigned to gas-phase CB (Huang et al. 2015). Based on the in situ DRIFTS analyses, it is noted that the CB molecules are first adsorbed on the catalyst surface (1470 cm −1 ) and the Cl atoms are nucleo-philically substituted, followed by the formation of phenol (1245 cm −1 ). In the presence of oxygen on the catalyst surface, phenolic compounds are produced as maleates (1364 cm −1 ) or carboxylic acid (1335 cm −1 ) compounds, which are subsequently deeply oxidized to CO 2 and H 2 O, while the Cl atoms remain on the catalyst surface or combine with the surface acidic sites to form HCl, which are released to the atmosphere.

Conclusion
In this study, a series of CeNb 3 Fe x /TiO 2 catalysts were prepared by the sol-gel method, which exhibited good catalytic activity with the simultaneous removal of NO x and chlorobenzene. The NO x removal rate of greater than 95% at 260-380 °C is observed over CeNb 3 Fe 0.3 /TiO 2 . Notably, in the presence of CB, the CeNb 3 Fe 0.3 /TiO 2 catalyst exhibits > 95% N 2 selectivity over the entire temperature range (140-500 °C). The CeNb 3 Fe 0.3 /TiO 2 presents better NO x reduction and CB oxidation, which might be related to its more favorable ratios of Ce 3+ /Ce, Fe 3+ /Fe, Nb 5+ /Nb, and abundant surface adsorbed oxygen species, which afford optimal reduction ability, suitable NO x , and CB adsorption capacity. Owing to the modification of Fe to the redox properties and acid sites of the CeNb 3 /TiO 2 catalyst, the CeNb 3 Fe 0.3 /TiO 2 catalyst is beneficial to NO x reduction and CB oxidation. Finally, the results of in situ DRIFTS study reveal the NH 3 -SCR reactions over CeNb 3 Fe 0.3 /TiO 2 catalysts are mainly controlled by the L-H mechanism (< 350 °C) and E-R mechanism (> 350 °C), respectively, and explored the multi-pollutant conversion mechanism in the synergistic reaction. Fig. 13 a In situ DRIFT spectra of CBCO + SCR on CeNb 3 Fe 0.3 / TiO 2 at different temperatures; b CBCO + SCR on CeNb 3 Fe 0.3 / TiO 2 at 250° C for 30 min. (Conditions: 1000 mg/L NH 3 + 1000 mg/L NO + 5% O 2 + 50 mg/L CB)