During the past several years, it has been proved that the selective catalytic reduction (SCR) process with NH₃ is one of the most prospective methods of removing NO_x in flue gases [1]. Among various transition-metal oxides (e.g., V₂O₅, CuO, Fe₂O₃, Cr₂O₃, MnO_x, etc.) [2-7], titania-based catalysts have been considered the most suitable catalysts for widespread SCR application due to their thermal stability and better resistance to sulfur dioxide poisoning [8, 9]. Recently, TiO₂ doped with ceria has attracted significant attention because of its excellent SCR catalytic performance and environmentally friendly properties [10-12].

At present, many researchers have focused on the anatase TiO₂ micro/nanostructures with (001) high-energy facets, and their applications in many fields, such as catalysts, photocatalysts, catalyst supports, sensors, photoanodes, etc. [13-18]. The TiO₂ (001) surface exposed all coordinatively unsaturated atoms five-fold Ti (5c-Ti), and this particular surface structure gave rise to an abundant oxygen deficiency and favors the dissociative adsorption for water and other molecules [19, 20]. Moreover, other researchers reported that the reactive oxygen atoms on the (001) facets played an important role in high photocatalytic selectivity [21]. It has been reported that TiO₂-NS with (001) facets showed much better catalytic performance in many cases than TiO₂-NP with (101) facets due to the more active sites and higher active energy. Liu et al. [22] investigated that, upon altering the ratio of (001), (101), and (010) facets, the photocatalytic reactivity would be correspondingly changed. Furthermore, Deng et al. [23] studied that manganese oxide doped on an anatase TiO₂ with (001) high-energy facets could promote the SCR process. Shi et al. [24] studied that octahedral vanadia species was the dominant species on TiO₂ (001) and was very active, while the tetrahedral vanadia species was prone to form on TiO₂ (101) and was inactive.

For restraining the growth of the stable (101) facets, some capping agents could be chosen to cause an exceptional stabilization of the (001) facets in anatase TiO₂ crystals, for instance, HF and SO₄^2- [18]. It is well known that SO₄^2- species was a very promising promoter of the high reactivity of the SCR reaction without by-products, especially for sulfated TiO₂ [25]. The introduction of sulfation could generate surface acidity, which leads to the enhancement of adsorption and activation of ammonia at high temperature [26]. Moreover, sulfation improves the redox properties of vanadate at the surface because of the strong electronic interaction between vanadia and sulfates on the sulfated V/TiO₂ catalysts [27]. In addition, it was also reported that catalytic activity was significantly affected by the different sulfation preparation methods [28]. Compared with CeO₂-S prepared by precipitation methods, the sulfated-CeO₂ prepared by a hydrothermal method with cerium(IV) sulfate as a precursor showed excellent SCR activity and high N₂ selectivity due to the intimate binding between sulfate species and Ce⁴⁺ [29].

In conclusion, until now, less attention has been paid to the influence of TiO₂ with high-energy (001) facets on catalytic performance, especially for the influence of sulfated TiO₂ with (001) facets. In this paper, sulfated titanium with exposed (001) facets was prepared by a simple one-step hydrothermal route using SO₄^2- as a morphology-controlling agent. After doping with ceria, Ce/TiO₂-001 was used as the catalyst for NH₃-SCR of NO, and its catalytic performance was tested. Meanwhile P25 and P25-S (prepared by another sulfation method, i.e., precipitation) were available as a reference. Furthermore, these catalysts were characterized by using various physico-chemical techniques, such as Brunauer-Emmett-Teller (BET) methods, X-ray diffraction (XRD), transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS), temperature programmed desorption (TPD), in situ diffuse reflectance infrared Fourier-transform spectroscopy (DRIFTS), etc., to discuss the relationship between the surface chemical properties and catalytic performance.

The precursors, containing TiOSO₄ (15 wt% solution in dilute sulfuric acid, Sigma Aldrich, USA) and tert-butyl alcohol (molar ratio 1:165), were placed in autoclave at 110 ℃ for 48 h [30]. The products were filtered and washed with distilled water and ethanol. After being dried and calcined, the required TiO₂ was prepared, and denoted TiO₂-001.

The other nanoparticle TiO₂ supports used in the present study were commercial P25 and P25-S. To obtain the sulfated P25, the P25 powder was added to the calculated H₂SO₄ solution (0.0625 mol/L) and stirred. The obtained solids were then dried and calcined.

The introduction of ceria (Ce/Ti molar ratio 1:19) into the supports was conducted by impregnating the titanium support with the required amount of cerium nitrate solution and the mixture stirred for 6 h, and then dried and calcined to give the final metal-oxide-loaded catalysts for further analysis and examination. The catalysts were designated Ce/TiO₂-001, Ce/P25, and Ce/P25-S.

The crystal phases of the samples were analyzed using an X-ray diffractometer (X'Pert3 Powder, PANalytical B.V., The Netherlands). XPS with Al K_α X-ray (hν = 1486.6 eV) radiation operated at 150 W (Thermo ESCALAB 250, USA) was used to investigate the surface properties. The morphology, structure, and grain size of the samples were examined by TEM (FEI F20, ThermoFisher, USA). Specific surface areas were determined by the BET methods on a nitrogen-adsorption apparatus (ASAP 2020, Micromeritics, Inc., USA). NH₃-TPD and O₂-TPD experiments were carried out on a custom-made thermal conductivity detector (TCD) setup using 100-mg catalysts. Prior to the experiments, catalysts were pre-treated in pure He at 400 ℃ for 1 h. For NH₃-TPD experiments, catalysts were saturated with anhydrous NH₃ (4% in He) at a flow rate of 30 mL/min for approximately 30 min. For O₂-TPD experiments, catalysts were saturated with anhydrous O₂ (4% in He) at a flow rate of 30 mL/min at 350 ℃ for approximately 30 min. Desorption of NH₃ or O₂ was carried out by heating the catalyst in He (50 mL/min) with a heating rate of 10 ℃/min.

In situ DRIFTS spectra were measured on a Fourier-transform infrared (FTIR) spectrometer (TENSOR 27, Bruker Corp., USA) with a mercury-cadmium-telluride (MCT) detector. Prior to each experiment, the sample was pretreated at 450 ℃ in a flow of He for 60 min and cooled to target temperature to obtain the background spectra. All spectra were recorded from 4000 to 600 cm^-1 by accumulating 64 scans with a resolution of 4 cm^-1.

For the desorption of NO+O₂ as a function of temperature, the catalysts were saturated by absorbed-NO_x species under a flow of 1000 ppm NO+3% O₂ at 30 ℃. The desorption process DRIFT spectra were then collected from 30 to 350 ℃ under a He flow. Then, in situ DRIFTS were recorded as a function of time at the reaction temperature of 250 ℃, the catalysts were exposed to a flow of 1000 ppm NO+3% O₂ for 60 min, and the reaction process was recorded at different periods with a flow of 1000 ppm NH₃.

SCR activity measurements were carried out in a fixed-bed reactor at 80–550 ℃ containing 0.5 g catalyst with a gas hourly space velocity (GHSV) of 100 000 h^-1. The typical reactant gas composition was as follows: 600 ppm NO, 600 ppm NH₃, 3.5% O₂, and balance N₂. NO, NO₂, and O₂ concentrations were monitored by a flue-gas analyzer (Testo 335, Testo, Inc., USA).

Based on the XRD patterns shown in Fig. 1, it was confirmed that all of the identified peaks could be perfectly indexed to anatase TiO₂ (JCPDS Card No. 21-1272) for TiO₂-001. The three strong peaks at 2θ= 25.3°, 37.8°, and 48.1° were attributed to anatase (101), (004), and (200) diffractions, respectively. P25 was a mixed-phased TiO₂ containing anatase and rutile phases. Few changes have been observed for the sample P25-S, except for the weaker intensity of the diffraction peaks, which indicated that sulfation had little influence on crystal-phase transformation. In addition, for TiO₂-001, the intensity of the diffraction peaks was much weaker and broader than that of P25, suggesting that the synthesized TiO₂ with exposed (001) facets had smaller crystal size. In addition, no diffraction peaks of cerium oxide were found on the catalysts Ce/TiO₂-001, Ce/P25-S, and Ce/P25, which might be due to good dispersion of cerium oxide on the supports.

From Table 1 and Fig. 2(a), it is obvious that P25 consists of mixed-phased nanoparticles with a surface area of 52.4 m²/g and an average diameter of 20 nm, matching the XRD definition. For P25-S, the average pore size and pore volume increased, and the BET surface areas decreased due to sulfation, but the influence was slight.

Because of the difference in the preparation method, the sample TiO₂-001 exerted a significant change on the structure and BET surface areas. The low-magnification TEM (left-hand inset of Fig. 2(b)) and nitrogen adsorption-desorption experience (Fig. 3 and Table 1) demonstrates the formation of mesoporous TiO₂ single crystals with a high surface area of 107 m²/g and a uniform pore diameter of 4.9 nm. From the high-resolution TEM (HRTEM) images in Fig. 2(b), it was found that a perpendicular lattice spacing of 0.19 nm was consistent with the (200) and (020) atomic planes of the anatase TiO₂. The corresponding selected-area electron-diffraction (SAED) pattern (right-hand inset of Fig. 2(b)) further confirms that the exposed crystal facet was a (001) facet. This is because of the addition of H₂SO₄ in the precursor solutions [30].

The XPS spectra of S 2p, Ce 3d, and O 1s levels are displayed in Fig. 4. The binding energies of S 2p for the catalysts Ce/TiO₂-001 and Ce/P25-S were 168.6 and 168.2 eV, which suggested the presence of S⁶⁺ and S existing as sulfate [25]. From Table 2, it was found that the surface sulfate content for both catalysts was similar to each other.

The XPS spectra of Ce 3d of all catalysts are shown in Fig. 4(b). The peaks at 916.4, 905.0, 900.0, 896.3, 886.3, and 882.7 eV correspond to the Ce⁴⁺ species, while those at 904, 898, 885.1, and 879.1 eV correspond to the Ce³⁺ species [31, 32]. The curve fittings in the XPS peaks of Ce 3d demonstrate that Ce existed as a mixture of Ce⁴⁺ and Ce³⁺. Incorporating XRD investigations confirms that the cerium existed well dispersed on catalysts as an amorphous nature of CeO₂ and Ce₂O₃ for the three catalysts. For the sulfated catalysts Ce/TiO₂-001 and Ce/P25-S, the peaks at binding energies of 885.1, 898, and 904 eV increased compared with Ce/P25, which contributed to the binding energy of Ce³⁺. Thus, it could be seen that sulfation was beneficial to the formation of Ce³⁺. It was reported that Ce³⁺ species could lead to the increase of chemisorbed oxygen on the surface, which was very active and played an important role in the reaction [10].

The XPS spectra for O 1s show that there are three peaks for the samples, which could be attributed to O_γ type (527.0–531.5 eV, lattice oxygen), O_β type (529.0–533.0 eV, chemisorbed oxygen) oxygen, and O_α type (530.7–534.7 eV, physically absorption oxygen) on the catalyst surface [11]. As shown in Fig. 4(c) and Table 2, compared with Ce/P25, the O_α concentration of Ce/TiO₂-001 and Ce/P25-S decreased slightly, but that of O_β type increased, which indicates the increase of chemisorbed oxygen. These results are in accord with the status of Ce³⁺ in Fig. 4(b).

In Fig. 5, the feature of NH₃ desorption of Ce/TiO₂-001 was quite different from Ce/P25 and the surface-sulfated Ce/P25-S. From 130 to 500 ℃, a wide and large ammonia adsorption peak was observed along with a maximum at 300 ℃, suggesting that the Ce/TiO₂-001 catalyst was composed of abundant weak, medium, and strong acid sites, especially the strong ones.

First, for Ce/TiO₂-001, the number of total acid sites might be relative to the sulfation and the high surface area, which could carry more acid sites. Sulfation could create abundant Brönsted/Lewis acid sites and strengthen the acidity [33]. However, as we know, half-fivefold-(5c-Ti) and half-sixfold-(6c-Ti) coordinated Ti atoms were present on the (101) ideal surface, while only 5c-Ti atoms were present on the (001) surface. Therefore, the water molecules were energetically favored to be dissociated on the (001) facets compared with (101) facets [20]. Consequently, more hydroxyls were terminally bonded to adjacent Ti sites and favored the formation of acid sites, especially for the medium-and high-strength acid sites. Therefore, it could be concluded that the TiO₂ with exposed (001) facets might affect the type of acid sites for the catalyst Ce/TiO₂-001.

TPD of O₂ is a direct method of measuring the oxygen species. As we know, the peak at 200–600 ℃ is due to chemisorbed oxygen, which plays an important role for medium-and high-temperature SCR performance [8]. From Fig. 6, the amount of chemisorbed oxygen increased successively as follows: Ce/P25 < Ce/P25-S < Ce/TiO₂-001. The reasons are twofold. First, the sulfation favored the generation of Ce³⁺, which was close to the chemisorbed oxygen, already mentioned above. Second, high-energy (001) facets played an important role. The surfaces of the (001) facets were composed of a high density of under-coordinated Ti atoms and very large Ti-O-Ti bond angles, resulting in abundant oxygen deficiency. In addition, the surface twofold-coordinated O (O_2c) atoms on (001) facets were quite different from the state of O_2c atoms on (101) facets and were very active [19, 21].

In situ DRIFTS experiments of NO+O₂ desorption were carried out to investigate the formation and transformation of NO_x species on the catalyst surface. As shown in Fig. 7(a), the adsorbed NO_x species were the gaseous NO₂ at 1614 cm^-1, the bidentate nitrates at 1581/1548 cm^-1, the linear nitrites at 1488 cm^-1, the monodentate nitrates at 1290 cm^-1, and the bridged nitrates at 1245 cm^-1 over Ce/P25 catalyst at 30 ℃ [34-39]. With increasing temperature, the linear nitrites, monodentate nitrates, and bridged nitrates gradually decreased and vanished at 350 ℃, while the gaseous NO₂ and bidentate nitrates were still stable and became the dominant species at high temperature.

In comparison with the in situ DRIFT spectra of Ce/P25, two obvious differences appeared in the spectra of Ce/P25-S and Ce/TiO₂-001 (Figs. 7(b) and 7(c)). (1) The main NO_x species were monodentate nitrates and monodentate nitrites on the Ce/P25-S and Ce/TiO₂-001 catalysts at 30 ℃, with bands at 1528/1520 cm^-1 [35] and 1321/1307 cm^-1 [38]. Additionally, the bands at 1633/1650 cm^-1 appeared at 30 ℃, but disappeared at temperatures higher than 100 ℃. Yang et al. [40] proposed that this band was to the absorbed H₂O. (2) The bands of monodentate nitrates/nitrites became weaker with increasing temperature and disappeared at temperatures higher than 150 ℃. Meanwhile, for the Ce/P25-S catalyst, the gaseous NO₂ (1603 cm^-1) and bidentate nitrates (1573 cm^-1) turned into the dominant species, and could be observed even at 350 ℃. Moreover, the bands at 1398 and 1380 cm^-1 assigned to ionic nitrates [36] were also detected at 350 ℃ over Ce/P25-S. In case of Ce/TiO₂-001 catalyst, the gaseous NO₂ (1610 cm^-1) and bidentate nitrates (1575 cm^-1) appeared at higher temperature, but only gaseous NO₂ was still obvious at 350 ℃. In brief, the evident differences might be related to the acidic properties and the high-energy (001) facet on the catalyst surface.

In order to identity the reactive species in the SCR reactions, the in situ DRIFTS spectra of NH₃ reacted with the pre-absorbed NO+O₂ species at 250 ℃ as a function of time were studied. As illustrated in Fig. 8(a), the bands at 1602 and 1570/1548 cm^-1 attributed to the gaseous NO₂ and bidentate nitrates hardly changed with the introduction of NH₃ in the first 5 min. Afterwards, the band at 1570 cm^-1 shifted to 1558 cm^-1 in 10 min. Simultaneously, the bands at 3380/3267/1174 cm^-1 started to be observed, which were attributed to the accumulation of NH₃ absorbed on the Lewis acid sites [41, 42]. However, the intensity of bands ranging from 1610 to 1500 cm^-1 remained and maintained stability longer. Zhan et al. [43] proposed that the bands at 1550 cm^-1 may be related to the intermediate of oxidation of ammonia. Then, it could be assumed that the gaseous NO₂ and bidentate nitrates were inactive on the Ce/P25 catalyst and not or partially reactive to NH₃. They might coexist with the intermediate of oxidation of ammonia.

In the case of Ce/P25-S (Fig. 8(b)), the band at 1604 cm^-1 ascribed to the gaseous NO₂ disappeared in 5 min. In addition, the band ascribed to bidentate nitrates (1573 cm^-1) decreased with increasing exposure duration, and then was replaced by the band at 1554 cm^-1 originating from the intermediate of oxidation of ammonia.

Fig 8(c), in contrast, indicates that the gaseous NO₂ (1612 cm^-1) disappeared completely after introducing NH₃ for only 1 min. Then, the bands at 3358/3263/1590/1320 and 1678/1445 cm^-1 appeared, which were attributed to the accumulation of NH₃ absorbed on the Lewis acid sites and Brönsted acid sites, respectively [41, 44-46].

The above results strongly demonstrate that the activity of the same NO_x-adsorption species on different catalysts varies widely. The activity of gaseous NO₂ and bidentate nitrates over the three catalysts were as follows: Ce/TiO₂-001 > Ce/P25-S > Ce/P25. The gaseous NO₂ was consumed faster over the Ce/TiO₂-001 catalyst than over the other two catalysts, which could be due to the special structure of the high-energy (001) facet of TiO₂ and the high concentration of surface chemically absorbed oxygen.

NO conversions for the prepared catalysts are shown in Fig. 9. It was observed that the catalytic activity of pure P25 was quite low, and the maximum value of NO conversion was only 38% at 525 ℃. After sulfation, the NO conversion reached 85% at 450 ℃ for the sample P25-S, due to the abundant acid sites, while a better activity was obtained for TiO₂-001 in the high-temperature range, the NO conversion of which reached up to 99% at 480 ℃ and was maintained above 85% in the range 420–530 ℃.

When introducing ceria, it was clear that a high NO conversion rate was obtained for Ce/TiO₂-001, whose maximum value was up to 99% at 290 ℃ and remained in the temperature range 290–480 ℃. It was found that the SCR activity decreased in the following sequence: Ce/TiO₂-001>Ce/P25-S>Ce/P25.

Sulfation created abundant acid sites, and the influence of acidity on SCR catalytic performance at high temperature, which played an important role in the adsorption and activation of ammonia at high temperature, was discussed in previous studies [26]. In this paper, according to the NO conversion for the P25-S and TiO₂-001 catalysts detailed above, it was further proved that sulfation was directly responsible for the SCR reaction at high temperature.

The high surface area was not only beneficial to the good dispersion of sulfate content and ceria, but could also carry more of these active components, which are crucial for SCR performance. Ceria doped on titania provides redox active sites for oxygen circulation and adsorption/activation of NO [10, 26]. Therefore, it was easy to understand the excellent performance for the catalyst Ce/TiO₂-001 with a high surface area 94.2 m²/g, approximately twice as much as that of Ce/P25 and Ce/P25-S.

It was obvious that the (001) facets were also responsible for TiO₂-001 catalyst activity, except for the acidity and high surface area, which affected the formation of the medium-and high-strength acid sites, active chemisorbed oxygen, and activated NO_x-adsorption species.

In short, the catalytic performance for SCR of NO on the catalyst Ce/TiO₂-001 was obviously superior to Ce/P25 and Ce/P25-S because of the high surface area, sulfation, and excellent properties of active-energy (001) facets.

For a series of samples, NO reduction activity followed the trend Ce/TiO₂-001>Ce/P25-S>Ce/P25. The use of Ce/TiO₂-001 revealed an excellent catalytic reaction for the SCR of NO by NH₃ at medium and high temperature, which was discussed from three main aspects. First, there was no doubt that high surface area was an important factor influencing its catalytic performance, which facilitated the dispersion and amount of the sulfate SO₄^2- and active-site ceria. Second, sulfation was an important factor that provided abundant acid sites for the good SCR reaction at high temperature and modulated the valence state of ceria. Third, the superior properties of high-energy (001) facets should not be ignored because of the specific surface reconstruction and abundant oxygen deficiency.