NO_x emissions from industries and vehicles not only impair human health but also cause serious environmental issues such as PM 2.5 and O₃ [1-3]. NO_x from power plants could be effectively reduced by NH₃ selective catalytic reduction (NH₃-SCR) using V₂O₅-WO₃/TiO₂ catalysts [4-6]. Although commercial V₂O₅-based catalysts have excellent activity and SO₂ resistance at high temperatures (>300 ℃), the unsatisfactory alkali-resistant capacity restricts their broader applications in steel plants, biomass burning, waste incineration plants, etc. [7-10]. Therefore, a large number of studies have been conducted to explore effective strategies to ameliorate the alkali-resistance of catalysts.

In general, the deactivation mechanisms of alkali poisoning include the blockage of catalyst channels and the decrease in surface acidity, which diminishes the adsorption and activation of NH₃. Extensive research has been studied on the improvement of alkali tolerance via increasing more acidic sites. Sulfated TiO₂ and ZrO₂ improved the acidity of catalysts and thus effectively improved alkali resistance [11-13]. The ion-exchanged titanate nanotubes (TNTs) have a number of OH groups with adequate acid content and could neutralize the basicity of alkali. Some TNTs-based catalysts such as CeO₂-doped sulfated TNTs [14], Mo modified V-TNTs [15] and Nb-doped Ce nanotubes [16] showed excellent alkali-resistance. Furthermore, it is effective to enhance the alkali-resistance of catalysts via constructing alkali-trapping sites to protect active sites. It has been demonstrated that BO₃³⁻ and PO₄³⁻ as capturing sites could preferentially bond with alkali metals and protect active sites of CeO₂ and Fe₂O₃ based catalysts [17,18]. Zhou et al. [9] found that the SCR activity was well maintained over the K-poisoned V₂O₅/CeO₂ because the in situ constructed SO₄^2– sites could capture K and retain a high adsorption rate and reactivity of NH₃. Fe₂(SO₄)₃/TiO₂ catalysts are also highly resistant to alkali due to the protective effects of SO₄^2– species [19]. The reactivity of K-poisoned catalysts could be recovered via SO₂ treatment because SO₂ preferred to combine with K, thus releasing the active sites poisoned by K [6]. These results demonstrated the protective effects of SO₄^2– species against alkali-poisoning. Recently, we found that Ce⁴⁺–SO₄²⁻ pair sites on S doped CeO₂ catalysts not only increased the Brønsted acid sites but also greatly improved the SO₂-resistance of catalysts resulting from suppressing the adsorption and oxidation of SO₂. It is promising to develop alkali-resistant catalysts by constructing Ce⁴⁺–SO₄²⁻ pair sites, which is of significance to the practice application of catalysts in complex working conditions.

In this study, we demonstrated that Ce(SO₄)₂ modified V₂O₅/TiO₂ (denoted as CeSV/Ti) catalysts exhibited improved NO_x reduction against alkali poisoning than V₂O₅/TiO₂ (denoted as V/Ti). K poisoned V/Ti showed poor activity with NO_x conversion below 70%. In comparison, K-poisoned CeSV/Ti showed much higher activity with NO_x conversion above 80% within ~270–450 ℃. For revealing the promoting effects of Ce(SO₄)₂ on the K-resistant capacity of V/Ti catalysts, the structure features of catalysts are investigated by X-ray diffraction (XRD); the acidic amounts of catalysts were determined by NH₃ temperature-programmed desorption combined with online mass spectrum (NH₃-TPD-MS), and redox properties probed of H₂ temperature-programmed reduction (H₂-TPR), X-ray photoelectron spectroscopy (XPS), etc. In addition, the adsorption and activation of NH₃ and NO as well as the pathways were deeply investigated by in situ diffuse reflectance infrared fourier transform spectra (DRIFTS).

The catalysts were prepared via an impregnation method. Take the preparation of 10 wt% Ce(SO₄)₂ modified V₂O₅/TiO₂ catalyst as an example. First, 2 g of TiO₂ was dispersed in 10 mL of deionized water and sonicated for 5 min. 0.2 g Ce(SO₄)₂ and 0.051 g NH₄VO₃ were dissolved in 6 mL of deionized by a heating method, respectively. After the mixed solution was cooled, the Ce(SO₄)₂ solution was added by drop to the TiO₂ suspension and stirred for 1 h and then the NH₄VO₃ solution was added by drop. After stirring for 30 min, the samples are dehydrated using a rotary evaporator and then dried overnight in an oven at 80 ℃. After the dried sample was calcined in a muffle furnace at 450 ℃ for 4 h with a ramping rate of 2 ℃/min, 10 wt% Ce(SO₄)₂ modified V₂O₅/TiO₂ was obtained, named as CeSV/Ti. Other 5, 15 and 20 wt% Ce(SO₄)₂ modified V₂O₅/TiO₂ catalysts were prepared using the same method. The Ce⁴⁺-modified and SO₄²⁻-modified V/Ti catalysts were prepared by impregnating 0.2614 g of Ce(NO₃)₃·6H₂O and 0.1591 g of (NH₄)₂SO₄, respectively. The preparation of V₂O₅/TiO₂ catalysts is similar to CeSV/Ti without the addition of Ce(SO₄)₂, named as V/Ti. K-poisoned CeSV/Ti and V/Ti catalysts were prepared by the impregnation method. 0.017 g of KNO₃ was dissolved in 10 mL of deionized water, and 0.5 g CeSV/Ti and V/Ti catalysts were added and stirred for 1 h. The mixture was dehydrated by a rotary evaporator, dried overnight in an oven at 80 ℃, and then calcined in a muffle furnace under 450 ℃ for 4 h with a ramping rate of 2 ℃/min. The obtained catalysts were named as K-CeSV/Ti and K-V/Ti, respectively. Characterizations of the obtained catalysts can be found in Supporting information.

SCR activity was characterized by using a fixed-bed quartz flow reactor (inner diameter: 8 mm) with the catalyst (40–60 meshes). SCR activity was tested from 150 ℃ to 420 ℃. The gas mixture composition: 500 ppm of NO, 500 ppm of NH₃, 5 vol% of O₂, and N₂ as the balance gas. The gas hourly space velocity (GHSV) was controlled at 50,000 h⁻¹ corresponding to the total flow rate of 257.7 mL/min. FTIR spectrometer (Protea atmosFIR) was used to simultaneously monitor the concentrations of NO, NO₂, NH₃, H₂O, and N₂O. The SCR activity data were recorded after the reaction system reached a stable state.

The NO_x conversion, N₂ selectivity, and GHSV were calculated by using the following formulas:

where NO_x represents the total concentration of NO and NO₂. [NO_x]_in, [NO_x]_out, [N₂O]_out, [NH₃]_in, and [NH₃]_out stood for the inlet and outlet concentrations of corresponding gas, respectively.

The GHSV was calculated by the following equation: GHSV=q_v/πhr²q_v meant the total flow rate; h represented the height of the catalyst in the reactor; and r stood for the inner radius of the reactor.

Firstly, the SCR activities of fresh and K-poisoned V/Ti catalysts were tested, while different amounts of Ce(SO₄)₂ modified V/Ti catalysts were compared (Figs. S1 and S2 in Supporting information). It can be seen that V/Ti shows NO_x conversion above 90% within 240–450 ℃ (Fig. 1 and Fig. S1). Different amounts of Ce(SO₄)₂ modified V/Ti catalysts all show decreased activities to some extent, in which 10–15 wt% Ce(SO₄)₂ modified ones show satisfactory activity with NO_x conversion above 90% from 270 ℃ to 450 ℃. After K-poisoning, the activity of K-V/Ti decreases markedly with the highest NO_x conversion of only 72% at 330 ℃ (Fig. 1 and Fig. S2). For K-poisoned Ce(SO₄)₂ modified V/Ti catalysts, the activity increases and then decreases with increasing Ce(SO₄)₂ amount from 5 wt% to 20 wt%, and the 10 wt% Ce(SO₄)₂ modified V/Ti exhibits the highest activity with NO_x conversion above 90% from 300 ℃ to 450 ℃ (Fig. 1). Besides, the fresh and K-poisoned CeSV/Ti show good N₂ selectivity above 95% (Fig. S3 in Supporting information). The CeSV/Ti catalyst exhibits excellent H₂O tolerance at 300 ℃ (Fig. S4 in Supporting information). Clearly, Ce(SO₄)₂ modification notably improves NO_x reduction against alkali poisoning over V/Ti catalysts. The Ce(SO₄)₂ amount more than 10 wt% may led to the poor dispersion of Ce(SO₄)₂, that is unfavorable for the alkali resistance. Besides, we also supplemented the SCR activity for Ce⁴⁺-modified and SO₄²⁻ modified V/Ti catalysts and their K-poisoned ones (Fig. S3). Although the low-temperature activities of CeV/Ti and SV/Ti are better than that of CeSV/Ti, the activity of K-poisoned CeSV/Ti is higher than that of K-poisoned CeV/Ti and SV/Ti catalysts. These results indicate that Ce⁴⁺–SO₄²⁻ pair sites play crucial roles in the strong alkali resistance of CeSV/Ti.

Firstly, the thermal stability of CeSV/Ti catalyst was tested via the N₂ temperature-programmed decomposition (TPDC (Fig. S5 in Supporting information). A small quantity of SO₂ (m/z = 64) was detected below 500 ℃ that is derived from the decomposition of surface adsorbed SO₄²⁻ species. The decomposition peak of Ce(SO₄)₂ occurs above 500 ℃ associated with the large formation of SO₂. The thermogravimetry (TG) analysis also shows the similar results (Fig. S6 in Supporting information). These results indicate the good thermal stability of CeSV/Ti within the temperature range of the activity test of catalysts. In order to probe the nature reason for the stronger alkali-resistance of CeSV/Ti catalysts, the structure features of fresh and K-poisoned CeSV/Ti and V/Ti catalysts were investigated by XRD (Fig. S7 in Supporting information). All catalysts display characteristic diffraction peaks ascribed to anatase TiO₂ (PDF #21–1272). There are no characteristic diffraction peaks related to Ce and V species, implying the highly dispersed state of Ce and V species. From the N₂ adsorption-desorption isotherm, all fresh and K-poisoned catalysts show mesoporous features due to the unique capillary adsorption phenomenon (Figs. S8–S11 in Supporting information). It can be seen that the pore volumes and pore sizes of both K-poisoned catalysts decrease compared to the fresh ones (Table S1 in Supporting information), indicating that the K₂O nanoparticles enter the pore channels and occupy partial space of pores of fresh catalysts. However, the specific surface areas of CeSV/Ti and V/Ti catalysts both increase to some extent after K-poisoning, which is likely owing to that K₂O nanoparticles in the pore channels contribute to the specific surface areas. Besides, the pore volume and pore diameter of catalysts both decrease to some extent after K-poisoning, which results from that K₂O nanoparticles occupy the partial position of pores surface. These results indicate that there are no correlations between textual features and the activity of catalysts.

Acid sites of catalysts play crucial roles in NH₃ adsorption/activation, and NH₃-TPD-MS was performed to study the acidic properties of fresh and K-poisoned catalysts. As shown in Fig. 2a, both CeSV/Ti and V/Ti catalysts show two NH₃ desorption peaks, where the low-temperature peaks are attributed to the weak acid sites and the high-temperature one ascribed to strong acid sites, respectively [20,21]. It is notable that the total acid amount of V/Ti increases from 20.2 µmol/g to 26.0 µmol/g after Ce(SO₄)₂ modification. After K-poisoning, the total acid amount of K-CeSV/Ti and K-V/Ti both decreases to some extents. The total acid amount (12.0 µmol/g) of K-CeSV/Ti is higher than that (7.4 µmol/g) of K-V/Ti (Table S2 in Supporting information), which contributes to the SCR activity of K-CeSV/Ti. The reducibility of fresh and K-poisoned CeSV/Ti and V/Ti catalysts were probed by H₂-TPR. As shown in Fig. 2b, the reduction peak at 447 ℃ on V/Ti is attributed to the reduction of VO_x species. For CeSV/Ti, the reduction peak at 489 ℃ is ascribed to the reduction of VO_x and Ce⁴⁺ species while the peaks at 546 and 640 ℃ are related to the reduction of SO₄²⁻ species [22]. Compared with V/Ti, the higher reduction temperature of V species over CeSV/Ti indicates the strong interaction between Ce and V. After K-poisoning, the reduction peak of VO_x species shifts to a higher temperature of 503 ℃ for K-V/Ti, indicating the K bonded VO_x species become more difficult to reduce. The reduction temperatures of VO_x, Ce⁴⁺ and SO₄²⁻ species over K-CeSV/Ti have no large changes compared with CeSV/Ti. However, it is notable that the reduction peak intensity of Ce⁴⁺ and SO₄²⁻ species markedly reduces over the K-poisoned CeSV/Ti, implying that SO₄²⁻ species as sacrificial sites bond with K.

To investigate the redox properties of fresh and K-poisoned catalysts, XPS was applied to probe the electron states of V and O on the surface of catalysts. As shown in Fig. 2c, the XPS spectra of V 2p show two characteristic binding energies of V 2p_3/2 and V 2p_1/2, where the peaks around 516 and 524 eV are attributed to V⁴⁺ species while the peaks around 517 and 525 eV are ascribed to V⁵⁺ [23]. It can be found that the V⁴⁺/(V⁴⁺+V⁵⁺) ratio (47.5%) on CeSV/Ti is higher than that (40.0%) on V/Ti, evidencing the strong interaction between Ce and V species over CeSV/Ti. The lower valence of V over CeSV/Ti indicates that the electron could transfer from Ce to V. After K-poisoning, the V⁴⁺/(V⁴⁺+V⁵⁺) of K-V/Ti increases markedly to 72.4% while that (50.3%) of K-CeSV/Ti hardly changes compared to the fresh catalysts. The significant increase in V⁴⁺ fraction of K-V/Ti is attributed to the electron-donating effects of K that increase the electron density of V species. This also demonstrates that K strongly interacts with V species after K-poisoning. Differently, the maintenance of V valence over K-CeSV/Ti implies that K strongly interacts with Ce(SO₄)₂ rather than V. The O 1s XPS spectra were used to investigate the varieties and contents of surface oxygen species. The surface oxygen atomic percent (53.8%) of CeSV/Ti is much higher than that (27.0%) of V/Ti, which is attributed to the oxygen species of Ce(SO₄)₂ (Table S3 in Supporting information). After K-poisoning, the surface oxygen atomic percent of K-CeSV/Ti slightly decreases to 46.5% while that of K-V/Ti increases to 45.6%. These results imply that K₂O is mainly bonded on Ce(SO₄)₂ sites of K-CeSV/Ti but cover on the surface of K-V/Ti. As shown in Fig. 2d, there are adsorbed oxygen (O_α) species around 531 eV and lattice oxygen (O_β) species around 530 eV [24-31]. Ce(SO₄)₂ modification increases the O_α fraction of V/Ti from 18.1% to 27.8%, likely due to the contribution of surface SO₄²⁻. After K-poisoning, the O_α species of K-CeSV/Ti decrease notably to 16.3% while that of K-V/Ti keeps unchanged at 18.9%, respectively. It can be inferred that SO₄²⁻ of Ce(SO₄)₂ strongly interacts with K over K-CeSV/Ti, which leads to the decrease of O_α species. By the contrast, K₂O covers on the surface of K-V/Ti that results in the unchanged amount of O_α species because of the O_α species of K₂O. In order to probe the interaction between Ce and K species on the K-CeSV/Ti catalysts, the XPS spectra of Ce 3d and K 2p are investigated over CeSV/Ti, K-CeSV/Ti and K-V/Ti catalysts (Figs. 2e and f). The peaks labelled as v, v'', v''' and u, u'' and u''' at 882, 888.9, 898, 901, 907 and 916.8 eV are attributed to the 3d¹⁰4f⁰ Ce⁴⁺ state, while the peaks labeled v′ and u′ observed at 885 and 903 eV are ascribed to the 3d¹⁰4f¹ Ce³⁺ state (Fig. 2e) [28]. It can be found that the ratio of Ce³⁺/(Ce³⁺+Ce⁴⁺) (31.1%) of K-CeSV/Ti is higher than that (27.8%) of CeSV/Ti, indicating the more electron-rich state of Ce in K-CeSV/Ti. On the other hand, the XPS spectra of K 2p were also investigated over K-CeSV/Ti and K-V/Ti catalysts. As seen in Fig. 2f, the binding energy of K on K-CeSV/Ti is higher than that on K-V/Ti, indicating the electron deficient state of K on the former one. Based on these results, it can be inferred that K is bonded on Ce(SO₄)₂ sites, leading to the transfer of electron from K to Ce species.

To explore the effects of K-poisoning on the adsorption and activation behavior of NH₃ and NO_x, the in situ DRIFTS of adsorption and desorption for NO_x and NH₃ species were investigated over fresh and poisoned catalysts. As shown in Fig. 3a, after introducing NH₃, the band at 1429 cm⁻¹ [32] ascribed to NH₄⁺ and the bands around 1324 [33] and 1257 cm⁻¹ [34] ascribed to NH₃ species are observed on CeSV/Ti catalysts. These NH₃/NH₄⁺ species gradually decrease and disappear above 300 ℃ with increasing the desorption temperature. Similarly, the NH₃ species (1328 and 1229 cm⁻¹) [33,35] and NH₄⁺ species (1470 and 1411 cm⁻¹) [36,37] appear on V/Ti after the adsorption of NH₃ (Fig. 3b). With the increasing desorption temperature, the NH₃/NH₄⁺ species gradually decrease and disappear above 300 ℃. Notably, there are more NH₄⁺ species on CeSV/Ti catalysts and a larger number of NH₃ species on V/Ti catalysts. After K-poisoning, the adsorption amount of NH₃/NH₄⁺ species both decrease on K-CeSV/Ti and K-V/Ti catalysts compared with the fresh catalysts. As seen in Fig. 3c, the NH₃ species (1228 and 1321 cm⁻¹) [33,35] and NH₄⁺ species (1423 and 1467 cm⁻¹) [38,39] appear on K-CeSV/Ti after the adsorption of NH₃. These NH₃/NH₄⁺ species decrease and disappear above 260 ℃ with increasing the desorption temperature. As for K-V/Ti (Fig. 3d), NH₃ species (1217 cm⁻¹) [40] and NH₄⁺ species (1398 and 1468 cm⁻¹) [36,41] appear after the adsorption of NH₃. With increasing desorption temperature, these NH₃/NH₄⁺ species decrease and disappear above 230 ℃. These results imply that the adsorption strength of NH₄⁺ species on K-CeSV/Ti is stronger than that of NH₃ species on K-V/Ti.

As shown in Fig. 3e, the adsorbed NO_x species such as the bridging nitrate species (1626 and 1225 cm⁻¹) and M-NO₂ (1379 cm⁻¹) appear on CeSV/Ti after the adsorption of NO+O₂ [42]. With increasing desorption temperature, these NO_x species gradually decrease and disappear above 300 ℃. As for V/Ti (Fig. 4f), the bridging nitrate species (1624 and 1183 cm⁻¹), M-NO₂ (1375 cm⁻¹) and gaseous NO₂ species (1324 cm⁻¹) appear after the adsorption of NO+O₂ [42]. These NO_x species decrease and disappear above 260 ℃ with increasing the desorption temperature. It is notable that the amount of NO_x species over V/Ti is much higher than that over CeSV/Ti. Besides, the NO_x species adsorbed much more strongly on CeSV/Ti than that on V/Ti, implying that these NO_x species on CeSV/Ti are not easy to participate in the SCR reaction. Therefore, the SCR reaction may occur on the CeSV/Ti catalyst between the adsorbed NH_x species and gaseous NO following the Eley-Rideal (E-R) mechanism while the reaction might procced between the adsorbed NH_x and NO_x species following the Langmuir-Hinshelwood (L-H) mechanism over V/Ti. After K-poisoning, the adsorption strength of NO_x species changes compared to fresh catalysts. As shown in Fig. 4g, the bidentate nitrate species (1543 cm⁻¹), M-NO₂ (1423 and 1466 cm⁻¹) and monodentate nitrates (1264 cm⁻¹) appear on K-CeSV/Ti after the adsorption of NO+O₂ [42]. The amount of these NO_x species is much more than that on fresh CeSV/Ti while these NO_x species adsorb strongly and only disappear above 300 ℃ with increasing the desorption temperature. This result implies that K-poisoning improves the adsorption of NO_x species but these species are relatively inactive. The reaction might still follow between NH₄⁺ species and gaseous NO via the E-R mechanism over K-CeSV/Ti. As for K-V/Ti (Fig. 4h), the bridging nitrate species (1601 cm⁻¹), bidentate nitrate species (1582 cm⁻¹), M-NO₂ (1493 cm⁻¹) and monodentate nitrates (1307 cm⁻¹) appear after the adsorption of NO+O₂ [42]. The amount of these NO_x species is much less than fresh V/Ti. With the increase of desorption temperature, these NO_x species decrease and disappear above 200 ℃. This result implies that K-poisoning impairs the adsorption of NO_x species and the reaction might also occur between NH₄⁺ species and NO_x species via the L-H mechanism over K-V/Ti.

To further explore the alkali-resistant mechanism, in situ DRIFT spectra of transient reactions were performed at 200 ℃. Firstly, the in situ DRIFT spectra of transient reactions between NO+O₂ and pre-adsorbed NH₃ were probed for fresh catalysts, as shown in Figs. 4a and b. As for CeSV/Ti (Fig. 4a), the NH₄⁺ (1428 cm⁻¹) [43] and NH₃ (1249 cm⁻¹) [44] species appear after the adsorption of NH₃, in which the amount of NH₄⁺ species is more than NH₃ species. With the introduction of NO+O₂, these NH₃/NH₄⁺ species gradually decrease within 20 min and a small quantity of nitrates including M-NO₂ nitrate species (1346 cm⁻¹) and bridging nitrate (1237 cm⁻¹) emerged [42]. For V/Ti (Fig. 4b), the NH₃ species (1233 cm⁻¹) [44] and NH₄⁺ species (1418 cm⁻¹) [45] appear after the adsorption of NH₃, where the amount of NH₃ species is more than NH₄⁺ species. With the introduction of NO+O₂, these NH₃/NH₄⁺ species consume gradually within 40 min, and the M-NO₂ (1373 cm⁻¹), bridging species (1623 cm⁻¹) and gaseous NO₂ (1320 cm⁻¹) increase [42]. It is notable that more NH₃/NH₄⁺ species are generated over CeSV/Ti than V/Ti. Besides, the reactivity of NH₃/NH₄⁺ species over CeSV/Ti is also higher than that on V/Ti. Figs. 4c and d show the in situ DRIFT spectra of transient reactions for K-poisoned catalysts between NO+O₂ and pre-adsorbed NH₃ at 200 ℃. As for K-CeSV/Ti (Fig. 4c), the NH₃ species (1230 and 1595 cm⁻¹) [38,44] and NH₄⁺ species (1418, 1470 and 1684 cm⁻¹) [36,45,46] appear after the adsorption of NH₃. Similar to the fresh CeSV/Ti, the amount of NH₄⁺ species is still more than NH₃ species. With the introduction of NO+O₂, these NH₃/NH₄⁺ species gradually decrease and a small quantity of bidentate nitrates species (1551 cm⁻¹) and M-NO₂ species (1340 cm⁻¹) emerge [42]. For K-V/Ti (Fig. 4d), the NH₃ species (1217 and 1603 cm⁻¹) [37,40] and NH₄⁺ species (1732 and 1379 cm⁻¹) [43,47] appear after the adsorption of NH₃. Meanwhile, the amount of NH₃ species is more than NH₄⁺ species. With the introduction of NO+O₂, these NH₃/NH₄⁺ species gradually decrease while the bidentate nitrates species (1579 cm⁻¹) and M-NO₂ species (1359 cm⁻¹) gradually emerge [42]. Clearly, the adsorption of NO_x species on V/Ti catalysts is much stronger than CeSV/Ti, indicating that the gaseous NO mainly participates the SCR reaction over CeSV/Ti catalysts.

Figs. 4e and f show in situ DRIFT spectra of the transient reactions at 200 ℃ between NH₃ and pre-adsorbed NO+O₂ for fresh catalysts. For CeSV/Ti (Fig. 4e), a small quantity of bridging nitrate species (1626 cm⁻¹) and M-NO₂ (1381 cm⁻¹) appear after the adsorption of NO+O₂ [42]. With the introduction of NH₃, these adsorbed NO_x species decrease while the NH₄⁺ species (1469 and 1429 cm⁻¹) [36,48] and NH₃ species (1250 cm⁻¹) [34] gradually increase. From the above results, it can be inferred that the reaction mainly proceeds between NH₄⁺ species and gaseous NO via the E-R reaction pathway over CeSV/Ti. For V/Ti (Fig. 4f), the bridging nitrates (1628 cm⁻¹), M-NO₂ species (1377 cm⁻¹) and adsorbed NO₂ (1320 cm⁻¹) appear after the adsorption of NO+O₂ [42]. After the introduction of NH₃, these NO_x species gradually decrease while the NH₃ species (1235 cm⁻¹) [44] and NH₄⁺ (1468 and 1410 cm⁻¹) [36,37] increase. Based the above results, the reaction mainly proceeds between NH₃ species and NO_x species via the L-H reaction pathway over V/Ti.

Figs. 4g and h show the in situ DRIFTS of transient reactions for K poisoned-catalysts between NH₃ and pre-adsorbed NO+O₂ at 200 ℃. As for K-CeSV/Ti (Fig. 4g), only a small quantity of M-NO₂ species (1338 cm⁻¹) are detected after the adsorption of NO+O₂ [42]. With introduction of NH₃, these NO_x species were consumed within 10 min and corresponding NH₃ (1324 and 1237 cm⁻¹) [33] and NH₄⁺ (1421 and 1466 cm⁻¹) [39,49] species gradually increase. Therefore, the SCR reaction could still proceed via the reaction between NH₄⁺ species and gaseous NO over K-CeSV/Ti (Fig. 4). For K-V/Ti (Fig. 4h), the bridging nitrates (1601 cm⁻¹), bidentate nitrates (1578 cm⁻¹), M-NO₂ (1484 cm⁻¹) and monodentate nitrates (1307 cm⁻¹) species are accumulated after the adsorption of NO+O₂ over the K-poisoned V/Ti [42]. However, after the introduction of NH₃, these NO_x species consumed slowly while the NH₃ species (1397 and 1219 cm⁻¹) [40,49] are detected. This result indicates that K-poisoning decreases the reactivity of the adsorbed NO_x with NH₃ species and blocked the L-H reaction pathway (Fig. 4).

In summary, improved NO_x reduction against alkali poisoning over CeSV/Ti catalysts has been originally demonstrated. K poisoned V/Ti shows NO_x conversion below 70% while K-poisoned CeSV/Ti shows NO_x conversion above 80% from 270 ℃ to 450 ℃. It has been demonstrated that Ce⁴⁺–SO₄²⁻ pair sites play crucial roles in improving the K-resistance of V/Ti catalysts. Here, we propose a deactivation mechanism over K-V/Ti and alkali-resistant mechanism over K-CeSV/Ti catalysts, as seen in Fig. 5. V/Ti catalysts mainly proceed the reaction between NH₃ species and NO_x species via the L-H reaction pathway. After Ce(SO₄)₂ modification, the strong interaction between V and Ce sites of Ce⁴⁺–SO₄²⁻ pair leads to the change of L-H reaction pathway into the E-R reaction pathway via the reaction between NH₄⁺ species and gaseous NO. After K-poisoning, the reactivity between the adsorbed NO_x species and NH₃ species is impaired over K-V/Ti and thus the L-H reaction pathway is blocked. In comparison, K-CeSV/Ti could still proceed the reaction effectively via the E-R reaction pathway because the SO₄²⁻ sites of Ce⁴⁺–SO₄²⁻ pairs strongly bond with K and protect active sites. This work evidenced an effective strategy to enhance NO_x reduction against alkali poisoning over catalysts via constructing Ce⁴⁺–SO₄²⁻ pair sites, contributing to developing alkali-resistant SCR catalysts for practical application in nonelectrical industries.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

We acknowledge the National Natural Science Foundation of China (Nos. 22125604, 22106100, 21976117), Shanghai Rising-Star Program (No. 22QA1403700) and Chenguang Program supported by Shanghai Education Development Foundation and Shanghai Municipal Education Commission (No. 22Z00354).

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2023.108240.