English

• The emissions of nitrogen oxides (NO_x) from coal-fired power stations and steel plants can trigger a series of environmental issues [1-3]. Selective catalytic reduction of NO_x with NH₃ (NH₃-SCR) is the most effective technology to eliminate NO_x currently [4,5]. Due to the satisfied SCR performance, V₂O₅-WO₃(MoO₃)/TiO₂ catalysts have been widely applied in stationary sources [6-8]. However, there still exist some drawbacks, such as the narrow operating temperature window (300–400 ℃) and the over-oxidation of SO₂ [9,10]. Furthermore, the presence of alkali metals in the exhaust gas can lead to severe deactivation of the catalysts [11-13]. Therefore, it is quite urgent to enhance the activity and alkali resistance of V-based catalysts to broaden the application among NH₃-SCR [14-16].

  So far, the FeVO₄ based catalyst has aroused extensive research interests because of its excellent low temperature activity and SO₂ resistance [17-19]. Liu et al. proposed that the well-dispersed FeVO₄ catalysts exhibited high activity and N₂ selectivity. They also revealed that the true active sites in FeVO₄ catalysts were VO_x species [20]. Mu et al. constructed a wide range of Fe_1-xV_xO_δ catalysts with various Fe to V ratios and found that the electron-induced effects between Fe and V together with the interaction between FeVO₄ and Fe₂O₃ species would greatly enhance the SCR performance [21]. However, after K₂O poisoning, the activity of FeVO₄ would decreased dramatically because of the destruction of acidity and redox properties [22,23]. Traditional strategies for improving the resistance to alkali metals could be divided into two routes: (1) Enhancing the surface acidity and (2) separating the active sites and the poisoning sites [24-27]. Basically, alkali metals are kinds of electron donors and their poisoning effects on catalysts mainly rely on the electron impact via reducing the valence of the boned active metal elements, thus undermining the Lewis acidity and redox capacity of the catalysts. Hence, weakening the electronic effect of alkali metals on the active species plays a crucial role in the resistance promotion of alkali metals on SCR catalysts [28,29]. Li et al. modified commercial V₂O₅-WO₃/TiO₂ catalysts with Ce and Cu and verified that the interaction between V, Ce, and Cu increased the ratio of V⁵⁺/(V⁵⁺+V⁴⁺), which contributed to the excellent K₂O resistance of the catalysts [30]. However, the effect of electronic interactions between Fe and V on the alkali resistance of catalysts as well as the specific electron effect of alkali metals on the active V species had still not been clearly defined.

  In this work, the FeVO₄ based catalysts with strong or weak interactions between Fe and V have been fabricated to investigate the role of electronic effects on the alkali resistance of SCR catalysts. The FeVO₄ catalyst prepared via impregnation method (i-FeV/Ti) is characterized as tightly bounded Fe and V with strong interaction, which shows superior SCR performance and alkali resistance than the catalyst with weaker interaction between Fe and V via co-precipitation preparation approach (denoted as p-FeV/Ti). The strong electronic withdraw ability of Fe increases the content of V⁵⁺ on the surface, thus leading to promoted Lewis acidity and redox ability, which is favorable for the adsorption and activation of NH₃ and NO. After K₂O poisoning, the electron-absorbing effects of Fe have offset the electron-donating effects of K₂O on active V sites, therefore protecting the active VO_x species to express the superior catalytic performance of NO_x reduction. Subsequently, the adsorbed and activated NH₃ on the surface could react with gaseous NO and O₂ following Eley-Rideal mechanism even in the presence of K₂O. This work provides a deep understanding of the effects of electronic interactions on the alkali resistance of FeVO₄ catalysts and paves a promising way to further design SCR catalysts with improved alkali resistance.

  The i-FeV/Ti catalyst was prepared through a traditional impregnation method. First, 0.3000 g oxalic acid was dissolved in 50 mL deionized water, then 0.0684 g ammonium metavanadate was added and stirred for 0.5 h, followed by the addition of 0.2363 g Fe(NO₃)₃·9H₂O and 1 g TiO₂. After stirring for another 2 h, they were dried using a rotary evaporator at 50 ℃ and then calcined at 500 ℃ for 4 h with a ramping rate of 5 ℃/min. The resulting samples were labeled as i-FeV/Ti. The p-FeV/Ti catalyst was prepared by a co-precipitation method and alkali poisoned samples were obtained by impregnating 1 wt% K₂O onto fresh catalysts. All details can be found in the Supporting information.

  SCR performance was tested through a fixed-bed quartz flow reactor with 0.3 g catalysts. The reactant gas was composed of 500 ppm NO, 500 ppm NH₃, 5 vol% O₂ and the carrier gas N₂. The total gas hourly space velocity (GHSV) was controlled to be 50,000 h^–1. NO_x conversion and N₂ selectivity were calculated by the following equations (Eqs. 1 and 2):

  (1)

  (2)

  [NO_x] represented the total concentration of NO and NO₂, while [NO_x]_in, [NO_x]_out, [NH₃]_in, [NH₃]_out, [N₂O]_out indicated the concentrations of the corresponding gas at the inlet and outlet, respectively.

  More detailed experimental parameters and characterizations could be found in Supporting information.

  The i-FeV/Ti catalyst was prepared by impregnation method, which was characterized as tightly bounded Fe and V. Fig. 1 shows the NO_x conversion and the N₂ selectivity of i-FeV/Ti, K-i-FeV/Ti, p-FeV/Ti, and K-p-FeV/Ti catalysts under a gaseous hourly space velocity (GHSV) of 50,000 h^–1. The i-FeV/Ti catalyst exhibits a broad temperature window of 210–390 ℃ with NO_x conversion over 90% and N₂ selectivity nearly 100%. By contrast, the NO_x conversion of p-FeV/Ti is lower within 150–300 ℃. After K₂O poisoning, two catalysts show quite different resistance performance. The SCR performance of i-FeV/Ti declines slightly within the whole temperature region. In contrast, the NO_x conversion of p-FeV/Ti decreases greatly from 96% to 31% at 300 ℃ and stays below 45% within 120–390 ℃. We also adjusted the content of Fe and V in i-FeV/Ti and found that when the ratio of Fe to V was 1:1, the catalyst showed the best activity and alkali resistance (Fig. S1 in Supporting information). In addition, we also test the SCR performance of i-FeV/Ti and p-FeV/Ti in the copresence of 1% K and 3% Pb and found that the performance of K&Pb-i-FeV/Ti remains superior than that of K&Pb-p-FeV/Ti (Fig. S2 in Supporting information). Besides, the activation energy of i-FeV/Ti and p-FeV/Ti was also tested under high GHSV (200,000 h^–1) with NO_x conversion below 20% (Figs. S3 and S4 in Supporting information). It was indicated that the i-FeV/Ti catalyst shows lower activation energy (39.37 kJ/mol) than p-FeV/Ti (51.07 kJ/mol), demonstrating its higher intrinsic activity. Furthermore, after K₂O poisoning, the activation energy of K-i-FeV/Ti (55.33 kJ/mol) remains lower than that of K-p-FeV/Ti (64.76 kJ/mol). The tightly bounded structure between Fe and V is more conducive to induce strong electronic interactions, that should be the reason why i-FeV/Ti exhibited superior low temperature activity and alkali resistance [20,31].

  Figure 1

  

  Figure 1.  NO_x conversion and N₂ selectivity of i-FeV/Ti, K-i-FeV/Ti, p-FeV/Ti and K-p-FeV/Ti catalysts. Reaction conditions: 500 ppm of NO, 500 ppm of NH₃, 5 vol% O₂, N₂ as the balance gas, GHSV of 50,000 h^–1.

  DownLoad: Full-Size Img PowerPoint

  In order to explore the intrinsic reasons for the different alkali resistance of i-FeV/Ti and p-FeV/Ti catalysts, the structural characteristics of catalysts were investigated. X-ray diffraction (XRD) patterns reveal the same diffraction peaks belonged to anatase and FeVO₄ for all catalysts (Fig. S5 in Supporting information). There are no obvious changes after K₂O poisoning, indicating that K₂O makes no obvious effects on crystal phase. Besides, the content of Fe and V in all catalysts was investigated by inductively coupled plasma-atomic emission spectrometry (ICP-AES) (Table S1 in Supporting information). The results suggest that i-FeV/Ti and p-FeV/Ti possess nearly the same amount of Fe and V, well excluding the effects of different amount of active species on the differential alkali resistance. Moreover, N₂ adsorption-desorption experiments were performed to characterize the pore structure of catalysts and the specific surface area were calculated by the Brunauer-Emmett-Teller model. It could be seen that the specific surface area of p-FeV/Ti is slightly larger than that of i-FeV/Ti (Table S2 in Supporting information). After K₂O poisoning, the specific surface area values for both samples decrease slightly, probably due to the blockage of pore channels [32,33].

  To probe the element distribution on the surface of these catalysts, the high-resolution transmission electron microscopy (HRTEM) and energy-dispersive X-ray spectroscopy (EDS) were carried out. The HRTEM image in Fig. 2A shows that there are no large FeVO₄ crystalline particles on the TiO₂ support for i-FeV/Ti, indicating that the active components are more likely to present in an amorphous form. Besides, the Fe and V elements are uniformly dispersed on the i-FeV/Ti catalyst, as seen from the elemental distribution mapping results in Fig. 2B. However, there are some aggregations of Fe species on p-FeV/Ti, exhibiting grain-like characteristics in Fig. 2B. It could also be found that distribution tendencies of Fe and V are almost the same on the surface of i-FeV/Ti in Fig. 2C, indicating that Fe element is closely combined with V species. However, the line sweep results for Fe on p-FeV/Ti catalysts show a wave-like trend in Fig. 2C, further demonstrating the inhomogeneous distribution of Fe. The poor distribution of Fe leads to limited bindings with V species, resulting in weaker electron interactions between Fe and V on p-FeV/Ti catalyst, which is the main reason for the worse activity and alkali resistance.

  Figure 2

  

  Figure 2.  (A) HRTEM image of i-FeV/Ti. (B) EDX mapping of i-FeV/Ti and p-FeV/Ti. (C) EDX Line sweep of i-FeV/Ti and p-FeV/Ti. (D) UV–vis spectra and (E) Raman spectra of i-FeV/Ti, K-i-FeV/Ti, p-FeV/Ti, and K-p-FeV/Ti catalysts.

  DownLoad: Full-Size Img PowerPoint

  In addition, UV–vis spectra were performed to further characterize the coordination structure of these catalysts. As shown in Fig. 2D, the peaks at 234 and 311 nm are attributed to Fe³⁺←O charge transfer of isolated Fe ions in tetrahedral and octahedral coordination, respectively. While the d-d transition band located at 799 nm is the fingerprint of V⁴⁺ [34,35]. Additionally, p-FeV/Ti exhibits an obvious band at 501 nm which should be ascribed to Fe₂O₃ nanoparticles [36,37]. Similar bands are not observed on i-FeV/Ti, which is consistent with the EDX mapping and line sweep results. Moreover, the in situ UV–vis DRS spectra were also performed at 240 ℃ to verify the active species on i-FeV/Ti and p-FeV/Ti (Figs. S6 and S7 in Supporting information). After the introduction of NH₃ or NO+O₂, the variations mainly occur in the range of 510–800 nm, corresponding to the d-d transition bands of V³⁺/V⁴⁺ [38]. As for the bands at 234, 311 and 501 nm corresponding to the Fe sites, there is no obvious change after the flowing of reaction gasses. This result demonstrates that the real active sites in the FeVO₄ catalysts are polymeric VO_x species, which is consistent with the reported literatures [20,39].

  Raman spectra were also collected to further identify the structure of all catalysts in Fig. 2E. All the spectra show the second-order feature of TiO₂ at 795 cm^–1 [40]. The band at 923 cm^–1 should be ascribed to V=O stretches at polymeric vanadyl species for i-FeV/Ti [6]. As for p-FeV/Ti and K-p-FeV/Ti catalysts, the peaks corresponding to the polymeric V species are significantly weaker. Due to the overlay of TiO₂ peaks, the bands related to Fe species could not be observed. Therefore, to further exclude the influence of support on the observation of Fe and V species among i -FeV/Ti and p-FeV/Ti catalysts, we continued to prepare bulk i-FeV and p-FeV materials to perform the Raman spectroscopy. It can be observed that the bands at 320, 363, 454, 645, 719, 753, 821, 882, 915, 945 cm^–1 should be attributed to the representative FeVO₄ and bands at 212, 272 cm^–1 are ascribed to hematite phase (Fig. S8 in Supporting information) [41]. It is obvious that p-FeV shows much stronger characteristic of Fe₂O₃ than i-FeV, and the peaks at 645 cm^–1 corresponding to the bridging Fe-O-V bands are more prominent on i-FeV, indicating tighter bonds between Fe and V [42]. The Raman spectra and EDS results demonstrate that i-FeV/Ti catalysts possessed more Fe-O-V sites than p-FeV/Ti catalysts.

  To clarify the electronic interactions between Fe and V species, X-ray photoelectron spectroscopy (XPS) was performed to probe the valence states of elements on the catalysts surface. As shown in Fig. 3A, the V 2p_3/2 characteristic peak of i-FeV/Ti is centered at 517.39 eV [43]. Compared with p-FeV/Ti (517.19 eV), V 2p peaks of i-FeV/Ti show higher binding energies. This variation can be caused by the transfer of more electrons from V to Fe. After K poisoning, the location of V 2p_3/2 characteristic peak on K-i-FeV/Ti barely changes but that on K-p-FeV/Ti shifts to the lower binding energy, indicating that the introduction of K₂O reduces the valence states of V element on the surface for p-FeV/Ti but barely effect i-FeV/Ti. Besides, the V⁵⁺/(V⁵⁺+V⁴⁺) ratio of i-FeV/Ti (89.31%) is higher than that of p-FeV/Ti (86.48%), which is in favor of NH₃ adsorption due to the stronger Lewis acidity [44]. Furthermore, the V⁵⁺/(V⁵⁺+V⁴⁺) ratio slightly changes after K₂O is introduced (88.24%) for i-FeV/Ti. As for p-FeV/Ti, the ratio decreases from 86.48 to 81.84% after K₂O poisoning. This phenomenon indicates that K⁺ imposes less influence on the electronic property of VO_x species of i-FeV/Ti [39]. The ratios of Fe³⁺/(Fe³⁺+Fe²⁺) were also calculated based on the Fe 2p XPS results shown in Fig. 3B. There are less Fe³⁺ on i-FeV/Ti (57.91%) than on p-FeV/Ti (63.57%). The different ratio of Fe³⁺ on two samples exactly proves the electron transfer from V to Fe over i-FeV/Ti, which has mentioned above. When K₂O is introduced, the amount of Fe³⁺ declined from 57.91% to 51.03% in i-FeV/Ti. As for p-FeV/Ti, it changes slightly from 63.57% to 60.26%. For K 2p XPS, the peaks located on the same place on both i-FeV/Ti and p-FeV/Ti (Fig. S9 in Supporting information). K₂O is a kind of electron donor that would damage the redox capacity of SCR catalysts, but the electron withdrawing function of Fe protect the true active sites of VO_x species from the attacking of K₂O [33]. As for the O 1s XPS spectra of all catalysts (Fig. S10 in Supporting information), the peaks located at ~531.46 and ~530.16 eV are attributed to the surface absorbed oxygen (denoted as O_α) and the lattice oxygen (denoted as O_β), respectively [45]. Generally, O_α is easier to be involved in the SCR reaction [46]. The O_α/(O_α+ O_β) ratios of i-FeV/Ti, K-i-FeV/Ti, p-FeV/Ti, K-p-FeV/Ti are 19.98%, 15.99%, 12.68%, and 7.61%, respectively. The higher amounts of O_α on the surface of i-FeV/Ti can be resulted from the more O vacancies on the surface of i-FeV/Ti, which may be resulted from the strong interaction between Fe and V species. The introduction of K₂O would certainly cause the loss of O vacancies, leading to the decline of O_a ratios [22]. While, the loss of O vacancies among K-i-FeV/Ti is less than that of K-p-FeV/Ti, further confirming the strong interaction between Fe and V species as well as the electron withdrawing function of Fe protect the active VO_x sites from K₂O poisoning.

  Figure 3

  

  Figure 3.  (A) XPS spectra of V 2p and (B) Fe 2p of i-FeV/Ti, K-i-FeV/Ti, p-FeV/Ti and K-p-FeV/Ti catalysts. (C, D) H₂ temperature-programmed reduction (H₂-TPR) profiles and (E) NH₃-TPD-MS profiles of i-FeV/Ti, K-i-FeV/Ti, p-FeV/Ti, and K-p-FeV/Ti catalysts. (F) In situ DRIFTS of NH₃ desorption of i-FeV/Ti and p-FeV/Ti catalysts.

  DownLoad: Full-Size Img PowerPoint

  In NH₃-SCR reaction, the acidity and redox properties of the catalyst are critical to its performance. The redox ability of catalysts was tested by H₂ temperature-programmed reduction (H₂-TPR). As illustrated in Fig. 3C, the reduction peaks at 346, 372, and 387 ℃ for the i-FeV/Ti catalyst should be assigned to the reduction of Fe³⁺ to Fe^(3–δ)⁺, Fe^(3–δ)⁺ to Fe²⁺, and V⁵⁺ to V⁴⁺, respectively [22]. After K₂O poisoning, these peaks migrate to higher temperature and the reduction peak of Fe³⁺ to Fe^(3–δ)⁺, Fe^(3–δ)⁺ to Fe²⁺ overlap, indicating that the redox capacity of catalysts are affected. Besides, the peak attributed to V⁵⁺ to V⁴⁺ shifts from 387 ℃ to 414 ℃ after K₂O poisoning. Compared with p-FeV/Ti, it can be found that the redox capacity of V species on i-FeV/Ti is less affected. For p-FeV/Ti shown in Fig. 3D, the peaks located at 543 ℃ should be ascribed to V⁴⁺ to V³⁺, which is not apparent on i-FeV/Ti, further evidencing the presence of more V⁴⁺ on the surface of p-FeV/Ti [47]. Furthermore, this peak becomes more pronounced after K₂O poisoning, indicating that K₂O poisoning exerts a greater influence on the V species of p-FeV/Ti. Besides, the H₂ consumption of i-FeV/Ti (1.252 mmol/g) and K-i-FeV/Ti (1.375 mmol/g) are higher than p-FeV/Ti (0.909 mmol/g) and K-p-FeV/Ti (0.986 mmol/g), further proving the superior redox ability of i-FeV/Ti.

  In addition to redox capacity, the acidity is also an important factor affecting the performance of SCR catalysts. Therefore, the NH₃ temperature-programmed desorption mass spectrometry (NH₃-TPD-MS) was undertaken to examine the acidity of all catalysts. As shown in Fig. 3E, the peak at 191 ℃ assigned to weak acid, the peak at 238 ℃ attributed to medium strong acid, and the other one at 318 ℃ presenting strong acid are observed on i-FeV/Ti. The NH₃ desorption amount of i-FeV/Ti (19.19 µmol/g) is much higher than that of p-FeV/Ti (12.19 µmol/g). Actually, the NH₃ absorption ability of Fe species is quite poor [48], but the strong electron-withdrawing property enhances the acidity of VO_x species, which endows i-FeV/Ti with better NH₃ absorption. After K₂O poisoning, The NH₃ desorption amount of K-i-FeV/Ti (7.38 µmol/g) remains higher than that of K-p-FeV/Ti (5.21 µmol/g). In situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) of NH₃ desorption was also used to analyze the adsorption and desorption behavior of NH₃ as well as the specific desorbed NH₃ species in Fig. 3F. The bands at 1665, 1437 cm^–1 are ascribed to NH₄⁺ absorbed on Brønsted acid sites, and bands located at 1602, 1235, 1183 cm^–1 are attributed to NH₃ absorbed on Lewis acid sites [49-55]. It can be observed that the intensity of the desorption peaks of i-FeV/Ti are significantly higher than those of p-FeV/Ti. These results demonstrate that i-FeV/Ti possess much stronger surface acidity than that of p-FeV/Ti. When the K₂O was corporate, the adsorbed species as well as the peak intensities on i-FeV/Ti barely change at low temperatures (< 200 ℃) (Fig. S11 in Supporting information). As for p-FeV/Ti, there are dramatic reductions in peak intensities (Fig. S12 in Supporting information). The effect of K₂O on the surface acidity can be divided into two aspects: (1) occupying the Brønsted acid sites and (2) reducing the elemental valence thus weakening the Lewis acid strength through electronic effects. Therefore, due to the offset effect of the electron interactions caused by strong interaction of Fe and V on i-FeV/Ti, the Lewis acidity of active VO_x species is maintained after K₂O poisoning.

  Furthermore, in situ DRIFTS were also displayed to clarify the NO adsorption behaviors over fresh and K₂O-poisoned catalysts. For i-FeV/Ti, the bands at 1625, 1577, 1364 and 1294 cm^–1 are attributed to bridged nitrate, bidentate nitrate, cis-N₂O₂^2– and monodentate nitrate, respectively (Fig. S13 in Supporting information) [50,51,53,55]. As the temperature raised, the peaks at 1625 and 1577 cm^–1 gradually decline and those at 1364 cm^–1 increase, indicating that bridged nitrate and bidentate nitrate are progressively transformed to cis-N₂O₂^2–. It is reported that cis-N₂O₂^2– generated on the surface would greatly enhance the NO conversion [56]. That should be one reason why i-FeV/Ti exhibits superior SCR activity. As for p-FeV/Ti, the peaks at 1241 and 1623 cm^–1 ascribed to bridged nitrate are more stable (Fig. S14 in Supporting information) [51,52]. Compared with i-FeV/Ti, the peaks attributed to cis-N₂O₂^2– are weaker and appear at higher temperature, demonstrating its inferior redox ability. What is more, for p-FeV/Ti, K₂O addition promoted the formation of thermally stable absorbed NO_x species, which would cover the active sites, including bridged nitrate (1605 cm^–1), bidentate nitrate (1576 cm^–1), and monodentate nitrate (1300 cm^–1), leading to inferior K₂O-resistance (Fig. S15 in Supporting information) [57,58]. However, the intensities of absorbed nitrate on i-FeV/Ti are significantly reduced after K₂O poisoning (Fig. S16 in Supporting information). The negative effect of K₂O is greatly alleviated by the strong electronic interaction between Fe and V over i-FeV/Ti.

  In order to clarify the whole reaction pathways over all catalysts, in situ DRIFTs of the transient reaction between NO+O₂ and preabsorbed NH₃ species were carried out at 240 ℃, at which the activity difference between K-i-FeV/Ti and K-p-FeV/Ti is greatest. After NH₃ is absorbed on i-FeV/Ti for an hour, the surface absorbed ammonia species mainly include ionic NH₄⁺ (1421 cm^–1), coordinated NH₃ (1601, 1255 cm^–1) and -NH₂ amide species (1326 cm^–1) (Fig. S17 in Supporting information) [53,59-61]. They are consumed in 2 min with the addition of NO+O₂. Then the peaks ascribed to bridged nitrate (1619 cm^–1), bidentate nitrate (1256 cm^–1) and cis-N₂O₂^2– (1376 cm^–1) appear [57,62,63]. It is obvious that the absorbed ammonia species would easily participate in the SCR reaction. As for p-FeV/Ti, the peak intensities of ammonia species absorbed on the surface, which should be attributed to NH₃ (1602, 1243, and 1204 cm^–1) on the Lewis acid sites, are relatively lower, proving its inferior acidity (Fig. S18 in Supporting information) [51,52,55]. For the transient reaction at 240 ℃ between NH₃ and preadsorbed NO+O₂ in Fig. 4A, cis-N₂O₂²⁻ (1376 cm^–1) and bridged nitrate (1623, 1216 cm^–1) are formed on i-FeV/Ti after adsorption of NO+O₂ [51,57,63]. After the introduction of NH₃, these peaks vanish within two minutes and the bands belong to NH₄⁺ (1421 cm^–1), NH₃ (1605, 1252 cm^–1) and -NH₂ amide species (1329 cm^–1) appear, proving that these nitrates possess high reactivity with NH₃ [53,55,60,61]. For p-FeV/Ti in Fig. 4B, after injecting NO+O₂ for 1 h, the peaks attributed to bridged nitrate (1619 cm^–1), bidentate nitrate (1558 cm^–1), cis-N₂O₂^2– (1362 cm^–1), and monodentate nitrate (1291 cm^–1) form and these adsorbed nitrate species could contributed to the SCR reaction [50,53,61]. However, the intensity of these peaks is quite lower than those observed in i-FeV/Ti, indicating its inferior redox ability. According to the above phenomena, we could find that the preabsorbed NH₃ species could react with NO+O₂ and the preabsorbed NO_x species could also react with NH₃ on these fresh catalysts, demonstrating that the reactions on the surface would both follow Langmuir-Hinshelwood mechanism and Eley-Rideal mechanism.

  Figure 4

  

  Figure 4.  In situ DRIFTS of the transient reactions at 240 ℃ between NH₃ and preadsorbed NO+O₂ of (A) i-FeV/Ti, (B) p-FeV/Ti, (C) K-i-FeV(p)/Ti, and (D) K-p-FeV/Ti catalysts.

  DownLoad: Full-Size Img PowerPoint

  To elucidate the effect of K₂O addition on the reaction pathway, in situ DRIFTs of the transient reactions were also carried out on K₂O poisoned catalysts. Compared with fresh i-FeV/Ti, similar species such as NH₃ (1605, 1240 cm^–1), NH₄⁺ (1425 cm^–1) and -NH₂ (1332 cm^–1) form on K-i-FeV/Ti after the injection of NH₃ for 1 h (Fig. S19 in Supporting information), but the intensities of these bands are reduced, which indicate that the addition of K₂O suppress the acidity on the surface [49,53,55,61]. A similar phenomenon also occur over K-p-FeV/Ti, but the differences are that the peak intensities are quite lower on K-p-FeV/Ti. After the introduction of NO+O₂, the absorbed NH₃ species are quickly consumed and the subsequent peaks attributed to absorbed NO_x species are quite weak on i-FeV/Ti, while a large amount of inactive nitrate species such as bidentate nitrate (1597, 1574 cm^–1), linear nitrite (1484 cm^–1) and monodentate nitrate (1300 cm^–1) accumulate on K-p-FeV/Ti (Fig. S20 in Supporting information) [52,55,60,61]. Similar tendency also occur on the transient reaction at 240 ℃ between NH₃ and preadsorbed NO+O₂. After preadsorption of NO+O₂ for 1 h, only quite few amounts of NO_x species such as cis-N₂O₂^2– (1356 cm^–1), monodentate nitrite (1311 cm^–1) and chelating nitrite (1174 cm^–1) are preserved on i-FeV/Ti in Fig. 4C. With the addition of NH₃, they are consumed in 5 min and corresponding NH₃/NH₄⁺ species form on the surface [50,55,58]. In contrast, after the flowing of NO+O₂ for 1 h, a large number of stable nitrate species cover the surface of p-FeV/Ti in Fig. 4D. It was reported that K would not affect the formation of ammonium nitrate species, but it would restrain the decomposition of ammonium nitrate species to N₂ and H₂O [28]. With the introduction of NH₃, these peaks do not change at all while some bands (1254 cm^–1, 1193 cm^–1) attributed to the absorbed NH₃ appear. It indicates that these stable nitrates cover the active V sites and hinder the process of SCR reaction on K-p-FeV/Ti, while on K-i-FeV/Ti, the strong electron interaction between Fe and V protect the active V species and promote SCR reaction [55,60]. Based on these results, after K poisoning, the reactions occur on i-FeV/Ti mainly follow the Eley-Rideal mechanism while the Langmuir−Hinshelwood mechanism is weakened, which is similar with other K poisoned V-based catalysts [11].

  In summary, the effects of electron interactions on alkali resistance over FeVO₄/TiO₂ for NH₃-SCR are revealed. From Fig. 5, it can be found that due to the strong electron-withdrawing effect of Fe, the valence of V element on the surface is increased, which facilitates the adsorption and activation of NH₃. After alkali poisoning, the electron-absorbing effect of Fe offset the electron-giving effect of K₂O on V element, thus protecting the real active sites of VO_x species. It allows the catalysts to participate in the SCR reaction through E-R mechanism even in the presence of K₂O. This work provides a profound comprehension of the electron interactions effects on the alkali resistance of catalysts and paves a promising way for the design of highly alkali-resistant SCR catalysts in the future.

  Figure 5

  

  Figure 5.  Possible reaction pathway of K poisoned i-FeV/Ti and p-FeV/Ti catalysts.

  DownLoad: Full-Size Img PowerPoint

  Declaration of competing interest

  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.

  Acknowledgments

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

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2023.108635.

  
  
• 1. [1]

     S. Beirle, K.F. Boersma, U. Platt, et al., Science 333 (2011) 1737–1739. doi: 10.1126/science.1207824

  2. [2]

     L. Han, S. Cai, M. Gao, et al., Chem. Rev. 119 (2019) 10916–10976. doi: 10.1021/acs.chemrev.9b00202

  3. [3]

     H. Wang, T. Zhu, Y. Qiao, et al., Chin. Chem. Lett. 33 (2022) 5223–5227. doi: 10.1016/j.cclet.2022.01.075

  4. [4]

     A. Marberger, D. Ferri, M. Elsener, et al., Angew. Chem. Int. Ed. 55 (2016) 11989–11994. doi: 10.1002/anie.201605397

  5. [5]

     W. Song, J. Ji, K. Guo, et al., Chin. Chem. Lett. 33 (2022) 935–938. doi: 10.1016/j.cclet.2021.07.022

  6. [6]

     G. He, Z. Lian, Y. Yu, et al., Sci. Adv. 4 (2018) eaau4637. doi: 10.1126/sciadv.aau4637

  7. [7]

     N.R. Jaegers, J.K. Lai, Y. He, et al., Angew. Chem. Int. Ed. 58 (2019) 12609–12616. doi: 10.1002/anie.201904503

  8. [8]

     J. Ji, L. Han, W. Song, et al., Chin. Chem. Lett. 23 (2022) 107769.

  9. [9]

     X. Liu, P. Wang, Y. Shen, et al., Environ. Sci. Technol. 56 (2022) 11646–11656. doi: 10.1021/acs.est.2c01812

  10. [10]

      Y. Pu, X. Xie, W. Jiang, et al., Chin. Chem. Lett. 31 (2020) 2549–2555. doi: 10.1016/j.cclet.2020.04.012

  11. [11]

      G. Zhou, P. Maitarad, P. Wang, et al., Environ. Sci. Technol. 54 (2020) 13314–13321. doi: 10.1021/acs.est.0c04536

  12. [12]

      F. Tang, B. Xu, H. Shi, et al., Appl. Catal. B 94 (2010) 71–76. doi: 10.1016/j.apcatb.2009.10.022

  13. [13]

      Z. Shen, X. Liu, S. Impeng, et al., Environ. Sci. Technol. 56 (2022) 5141–5149. doi: 10.1021/acs.est.1c08096

  14. [14]

      Z. Jia, Y. Shen, T. Yan, et al., Environ. Sci. Technol. 56 (2022) 2647–2655. doi: 10.1021/acs.est.1c06932

  15. [15]

      Y. Zhao, L. Shi, Y. Shen, et al., Environ. Sci. Technol. 56 (2022) 4386–4395. doi: 10.1021/acs.est.1c07996

  16. [16]

      Z. Liu, S. Zhang, J. Li, et al., Appl. Catal. B 158-159 (2014) 11–19.

  17. [17]

      N. Zhu, W. Shan, Z. Lian, et al., J. Hazard. Mater. 382 (2020) 120970. doi: 10.1016/j.jhazmat.2019.120970

  18. [18]

      G. Wu, J. Li, Z. Fang, et al., Chem. Eng. J. 271 (2015) 1–13. doi: 10.1016/j.cej.2015.02.012

  19. [19]

      S. Li, W. Hu, Z. Xu, et al., ChemCatChem 14 (2022) e202200476. doi: 10.1002/cctc.202200476

  20. [20]

      F. Liu, H. He, Z. Lian, et al., J. Catal. 307 (2013) 340–351. doi: 10.1016/j.jcat.2013.08.003

  21. [21]

      J. Mu, X. Li, W. Sun, et al., ACS Catal. 8 (2018) 6760–6774. doi: 10.1021/acscatal.8b01196

  22. [22]

      Z. Si, Y. Shen, J. He, et al., Environ. Sci. Technol. 56 (2022) 605–613. doi: 10.1021/acs.est.1c05686

  23. [23]

      Y. Li, S. Cai, P. Wang, et al., Environ. Sci. Technol. 55 (2021) 9276–9284. doi: 10.1021/acs.est.1c01722

  24. [24]

      P. Zhang, P. Wang, A. Chen, et al., Environ. Sci. Technol. 55 (2021) 11970–11978. doi: 10.1021/acs.est.1c02882

  25. [25]

      C. Feng, P. Wang, X. Liu, et al., Environ. Sci. Technol. 55 (2021) 11255–11264. doi: 10.1021/acs.est.1c02061

  26. [26]

      K. Zha, C. Feng, L. Han, et al., Chem. Eng. J. 381 (2020) 122764. doi: 10.1016/j.cej.2019.122764

  27. [27]

      Z. Zhou, J. Lan, L. Liu, et al., Catal. Commun. 149 (2021) 106230. doi: 10.1016/j.catcom.2020.106230

  28. [28]

      S. Xiong, J. Chen, H. Liu, et al., Environ. Sci. Technol. 56 (2022) 3739–3747. doi: 10.1021/acs.est.2c00113

  29. [29]

      Y. Zheng, Y. Guo, W. Shan, et al., Appl. Catal. A 637 (2022) 118606. doi: 10.1016/j.apcata.2022.118606

  30. [30]

      H. Li, J. Miao, Q. Su, et al., J. Mater. Sci. 54 (2019) 14707–14719. doi: 10.1007/s10853-019-03919-5

  31. [31]

      M. Casanova, L. Nodari, A. Sagar, et al., Appl. Catal. B 176-177 (2015) 699–708.

  32. [32]

      S. Li, W. Huang, H. Xu, et al., Appl. Catal. B 270 (2020) 118872. doi: 10.1016/j.apcatb.2020.118872

  33. [33]

      S. Zhang, S. Liu, W. Hu, et al., Appl. Surf. Sci. 466 (2019) 99–109. doi: 10.1016/j.apsusc.2018.09.261

  34. [34]

      M. Zhu, J.K. Lai, U. Tumuluri, et al., ACS Catal. 7 (2017) 8358–8361. doi: 10.1021/acscatal.7b03149

  35. [35]

      H. Liu, J. Wang, T. Yu, et al., Catal. Sci. Technol. 4 (2014) 1350–1356. doi: 10.1039/C3CY01026K

  36. [36]

      F. Gao, Y. Zheng, R.K. Kukkadapu, et al., ACS Catal. 6 (2016) 2939–2954. doi: 10.1021/acscatal.6b00647

  37. [37]

      N. Zhu, Z. Lian, Y. Zhang, et al., Chin. Chem. Lett. 30 (2019) 867–870. doi: 10.1016/j.cclet.2019.03.011

  38. [38]

      L.J. Burcham, G. Deo, X.T. Gao, et al., Top. Catal. 11 (2000) 85–100.

  39. [39]

      A. Marberger, M. Elsener, D. Ferri, et al., ACS Catal. 5 (2015) 4180–4188. doi: 10.1021/acscatal.5b00738

  40. [40]

      Z. Lian, J. Wei, W. Shan, et al., J. Am. Chem. Soc. 143 (2021) 10454–10461. doi: 10.1021/jacs.1c05354

  41. [41]

      K. Routray, W. Zhou, C.J. Kiely, et al., ACS Catal. 1 (2010) 54–66.

  42. [42]

      H. Tian, I.E. Wachs, L.E. Briand, J. Phys. Chem. B 109 (2005) 23491–23499. doi: 10.1021/jp053879j

  43. [43]

      B. Liu, X. Li, Q. Zhao, et al., J. Mater. Chem. A 3 (2015) 15163–15170. doi: 10.1039/C5TA02295A

  44. [44]

      T.H. Vuong, J. Radnik, J. Rabeah, et al., ACS Catal. 7 (2017) 1693–1705. doi: 10.1021/acscatal.6b03223

  45. [45]

      L. Kang, L. Han, P. Wang, et al., Environ. Sci. Technol. 54 (2020) 14066–14075. doi: 10.1021/acs.est.0c05038

  46. [46]

      L. Yan, F. Wang, P. Wang, et al., Environ. Sci. Technol. 54 (2020) 7697–7705. doi: 10.1021/acs.est.0c01749

  47. [47]

      G. Wu, J. Li, Z. Fang, et al., Catal. Commun. 64 (2015) 75–79. doi: 10.1016/j.catcom.2015.01.035

  48. [48]

      H. Wang, Z. Qu, S. Dong, et al., Environ. Sci. Technol. 50 (2016) 13511–13519. doi: 10.1021/acs.est.6b03589

  49. [49]

      W. Chen, Z. Li, F. Hu, et al., Appl. Surf. Sci. 439 (2018) 75–81. doi: 10.1016/j.apsusc.2018.01.057

  50. [50]

      Y. Zhang, X. Yue, T. Huang, et al., Materials 11 (2018) 1307. doi: 10.3390/ma11081307

  51. [51]

      Q. Zhang, J. Fan, P. Ning, et al., Appl. Surf. Sci. 435 (2018) 1037–1045. doi: 10.1016/j.apsusc.2017.11.180

  52. [52]

      L. Huang, K. Zha, S. Namuangruk, et al., Catal. Sci. Technol. 6 (2016) 8516–8524. doi: 10.1039/C6CY02026G

  53. [53]

      Z. Liu, S. Zhang, J. Li, et al., Appl. Catal. B 144 (2014) 90–95. doi: 10.1016/j.apcatb.2013.06.036

  54. [54]

      L. Wei, S. Cui, H. Guo, et al., J. Mol. Catal. A: Chem. 421 (2016) 102–108. doi: 10.1016/j.molcata.2016.05.013

  55. [55]

      H. Hu, S. Cai, H. Li, et al., ACS Catal. 5 (2015) 6069–6077. doi: 10.1021/acscatal.5b01039

  56. [56]

      K. Cheng, B. Liu, W. Song, et al., Ind. Eng. Chem. Res. 57 (2018) 7802–7810. doi: 10.1021/acs.iecr.8b01441

  57. [57]

      H. Liang, K. Gui, X. Zha, Can. J. Chem. Eng. 94 (2016) 1668–1675. doi: 10.1002/cjce.22546

  58. [58]

      W. Mu, J. Zhu, S. Zhang, et al., Catal. Sci. Technol. 6 (2016) 7532–7548. doi: 10.1039/C6CY01510G

  59. [59]

      J. Liu, J. Meeprasert, S. Namuangruk, et al., J. Phys. Chem. C 121 (2017) 4970–4979. doi: 10.1021/acs.jpcc.6b11175

  60. [60]

      Y. Yu, J. Wang, J. Chen, et al., Ind. Eng. Chem. Res. 53 (2014) 16229–16234. doi: 10.1021/ie502065b

  61. [61]

      Z. Ma, X. Wu, H. Härelind, et al., J. Mol. Catal. A: Chem. 423 (2016) 172–180. doi: 10.1016/j.molcata.2016.06.023

  62. [62]

      M.M. Azis, H. Härelind, D. Creaser, Catal. Sci. Technol. 5 (2015) 296–309. doi: 10.1039/C4CY00816B

  63. [63]

      Z. Zhang, L. Chen, Z. Li, et al., Catal. Sci. Technol. 6 (2016) 7151–7162. doi: 10.1039/C6CY00475J

• 

  Figure 1  NO_x conversion and N₂ selectivity of i-FeV/Ti, K-i-FeV/Ti, p-FeV/Ti and K-p-FeV/Ti catalysts. Reaction conditions: 500 ppm of NO, 500 ppm of NH₃, 5 vol% O₂, N₂ as the balance gas, GHSV of 50,000 h^–1.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (A) HRTEM image of i-FeV/Ti. (B) EDX mapping of i-FeV/Ti and p-FeV/Ti. (C) EDX Line sweep of i-FeV/Ti and p-FeV/Ti. (D) UV–vis spectra and (E) Raman spectra of i-FeV/Ti, K-i-FeV/Ti, p-FeV/Ti, and K-p-FeV/Ti catalysts.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (A) XPS spectra of V 2p and (B) Fe 2p of i-FeV/Ti, K-i-FeV/Ti, p-FeV/Ti and K-p-FeV/Ti catalysts. (C, D) H₂ temperature-programmed reduction (H₂-TPR) profiles and (E) NH₃-TPD-MS profiles of i-FeV/Ti, K-i-FeV/Ti, p-FeV/Ti, and K-p-FeV/Ti catalysts. (F) In situ DRIFTS of NH₃ desorption of i-FeV/Ti and p-FeV/Ti catalysts.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  In situ DRIFTS of the transient reactions at 240 ℃ between NH₃ and preadsorbed NO+O₂ of (A) i-FeV/Ti, (B) p-FeV/Ti, (C) K-i-FeV(p)/Ti, and (D) K-p-FeV/Ti catalysts.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  Possible reaction pathway of K poisoned i-FeV/Ti and p-FeV/Ti catalysts.

  下载: 全尺寸图片 幻灯片
•