Enhanced low-temperature NH₃-SCR performance of Ce/TiO₂ modified by Ho catalyst

Abstract

Holmium was used as a dopant to boost the low-temperature NH₃-selective catalytic reduction (SCR) performance of Ce/TiO₂ catalyst. It was ascertained that certain amount of Ho-doping species could exceedingly improve the low-temperature SCR activity under 60 000 h⁻¹ of Ce/TiO₂, accompanied with the improvement of tolerance to H₂O and SO₂ at 200°C. Characterization results manifested that Ho modification could not only result in inhibiting the growth of TiO₂ crystals and the enlargement of specific surface area but also lead to the enhanced redox ability and the increased amount of surface-adsorbed substances, all of which could promote the low-temperature NH₃-SCR performance of Ce/TiO₂.

1. Introduction

Lately, NOx has become one of the significant sources of air pollution. The over-standard concentration of NOx emission was mainly caused by the combustion process of fossil fuel, which has caused many environmental problems such as city smog and pollution [1–4]. Selective catalytic reduction (SCR) is the widely accepted de-NOx technology, and V₂O₅-WO₃ (MoO₃)/TiO₂ catalyst is the most commercially used catalyst in this SCR system [5]. However, there are still some disadvantages in the SCR system with V₂O₅-WO₃ (MoO₃)/TiO₂, such as the high operating temperature (300–400°C), the toxicity of vanadium species, low N₂ selectivity in the working temperature range [6–9]. Based on these practical disadvantages, it is necessary to study non-vanadium catalysts with better low-temperature SCR performance.

Cerium, one of the most abundant rare-earth metals, has drawn attention due to the high oxygen storage capacity and good redox property. It has been widely applied in catalysis, such as carbon monoxide oxidation and reforming reactions [10–12]. Results of previous research proved that cerium-based oxide catalysts had a good SCR performance. Gao et al. [13] reported that Ce/TiO₂ by the sol–gel method possesses high surface area and good redox ability, contributing to its high SCR activity. Vuong et al. [14] reported that V/CeTiO₂ catalysts showed excellent de-NOx activity at low temperature. Notably, the best one of these V/CeTiO₂ catalysts showed almost 100% NO conversion at 190°C. It was also found [15] that doping certain quantity of Ca would increase Ce³⁺ and surface-adsorbed oxygen. Meanwhile, the Brønsted acidity and redox ability were also greatly enhanced. All these factors may be responsible for the enhanced activity. Mosrati [16] recently reported that an impregnated Ce/Ti oxide catalyst with Nb modification presents a 95% NOx conversion at 200°C. Relevant characterization results proved that the Nb introduction decreases the surface area and strengthens the surface acidity. A Ce–Ti oxide catalyst with Cu addition could promote the SO₂ resistance of Ce–Ti oxide [17]. Although several catalysts, such as V/CeTiO₂ and Ce–Cu–TiO₂, have been successfully applied in NH₃-SCR, enhanced low-temperature NH₃-SCR performance and SO₂ resistance of Ce/TiO₂ modified by Ho have never been reported.

As a rare earth metal, Ho has been successfully applied for improving the photocatalytic activity of TiO₂ [18]. Owing to its electron trap effect of Ho²⁺$\leftrightarrow$Ho³⁺, the doping of Ho could efficiently enhance the photocatalytic ability of TiO₂. Gamal et al. [19] reported that the surface of Ho₂O₃ exposes more Lewis acid sites, which play a vital role in NH₃-SCR reaction. It was also reported [20] that Ho-modified Fe–Mn/TiO₂ catalyst shows a larger specific area of Fe₂O₃ phase compared with that of Fe–Mn/TiO₂, which results in a board temperature window and high SO₂ tolerance in NH₃-SCR reaction. However, the investigation of Ce/TiO₂ catalyst with Ho addition has not been reported. In this work, Ho is used for improving the low-temperature NH₃-SCR activity of Ce/TiO₂, and several characterization methods were applied for investigating the promotion mechanism. Furthermore, SO₂ + H₂O tolerance of the best catalyst was also studied.

2. Experimental

2.1. Catalyst preparation

The impregnation method was used to prepare the catalysts. Titanium dioxide (anatase, 0.05 mol) was impregnated with cerium nitrate (0.0175 mol) and holmium nitrate in 100 ml deionized water, followed by stirring at 20°C for 3 h. The obtained mixture was dried for 12 h at 100°C and then calcined at 500°C for 4 h. The prepared samples were labelled as Ce_0.35/TiO₂ and Ho_xCe_0.35/TiO₂ (the molar ratios of Ho/Ti and Ce/Ti were x and 0.35, respectively).

2.2. Catalyst characterization

Powder X-ray diffraction (XRD) patterns were obtained on a Philips X'pert Pro diffractometer with Ni-filtered Cu Kα radiation (0.15408 nm). 2θ ranged from 10° to 80° with a step size of 0.02°.

The specific surface area was measured by N₂ adsorption at −196°C, using an ASAP 2020 volumetric adsorption analyser. Before each precise test, the catalysts were evacuated for 3 h at 300°C. The specific surface area and the pore volume of the samples were calculated by the Brunauer–Emmett–Teller (BET) method and the pore size distributions were derived from the adsorption branches of the isotherms using the Barrett–Joyner–Halenda model.

The H₂ temperature-programmed reduction (TPR) experiments were performed on a Micromeritics AutoChem 2920 chemisorption analyser. Typically, 0.1 g sample was pretreated in pure N₂ at 400°C for 0.5 h and then cooled to 20°C followed by N₂ purging for 0.5 h. The temperature was heated by 10°C min⁻¹ from 100 to 800°C in 10 vol% H₂/Ar. Thermal conductivity detector monitored H₂ consumption in this progress.

The NH₃ temperature-programmed desorption (TPD) experiments were carried out on the same equipment as the TPR experiment. As a pretreatment step, 0.1 g samples were purged at 400°C in N₂ for 0.5 h and cooled to 30°C. Then the samples were purged in NH₃ for 1.0 h. At last, the programmed desorption was carried out at the rate of 10°C min⁻¹ (100–500°C) in Ar.

In situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) experiments were carried out on a Nicolet 6700 FTIR spectrometer with an MCT/A detector. As a pretreatment step, the catalysts were treated at 450°C in N₂ for 0.5 h and cooled to 50°C. Background spectra were recorded in the N₂ flow and automatically subtracted from the corresponding spectra. The spectra were recorded by accumulating 64 scans at a 4 cm⁻¹ resolution.

2.3. Catalytic activity test

SCR activity experiments were performed in a fixed-bed reactor containing 0.4 g catalysts (40–60 mesh) with a GHSV of 60 000 h⁻¹. The total gas flow was 200 ml min⁻¹, which was premixed in a gas mixer to obtain the simulated gas of [NO] = [NH₃] = 500 ppm, [O₂] = 3 vol.%, [H₂O] = 8 vol.% (when used), [SO₂] = 200 ppm (when used) and balanced by N₂. Then the mixed gas went into the reactor and the NO and NO₂ concentrations were monitored by a 350-XL flue gas analyser. The experiment data were recorded from 100 to 400°C at a steady state. The formulae for NO_x conversion and N₂ selectivity were as follows:

$${\mathrm{NO}}_{\mathrm{x}}\hspace{0.17em}\mathrm{conversion}(\text{\%})=\frac{{[{\mathrm{NO}}_{\mathrm{x}}]}_{\mathrm{in}}-{[{\mathrm{NO}}_{\mathrm{x}}]}_{\mathrm{out}}}{{[{\mathrm{NO}}_{\mathrm{x}}]}_{\mathrm{in}}}\times 100\text{\%}$$ 2.1

$${\mathrm{N}}_2\mathrm{selectivity}=\left(1-\frac{2{[{\mathrm{N}}_2\mathrm{O}]}_{\mathrm{out}}}{{[\mathrm{NOx}]}_{\mathrm{in}}+{[\mathrm{N}{\mathrm{H}}_3]}_{\mathrm{in}}-{[\mathrm{NOx}]}_{\mathrm{out}}-{[\mathrm{N}{\mathrm{H}}_3]}_{\mathrm{out}}}\right)\times 100\text{\%}.$$ 2.2

Also, NO oxidation conversion was also tested in the same fixed-bed reactor in the same simulated flue gas components without NH₃.

3. Results and discussion

3.1. Catalytic performance

The NOx conversions of various catalysts are plotted as a function of temperature, as exhibited in figure 1a. Among the prepared catalysts, Ce_0.35/TiO₂ and Ho_0.35/TiO₂ showed a limited de-NOx activity (less than 80%) in the entire temperature scope. It is notable that the low-temperature (less than 200°C) catalytic activity of Ce_0.35/TiO₂ was much improved when small amounts of Ho species are doped, as evidenced by the NO conversion of Ho_0.15Ce_0.35/TiO₂. When the Ho/Ti molar ratio rises to 0.45, the NOx conversion over Ce_0.35/TiO₂ at 150°C was also increased from 22% to 56%. However, further increasing of Ho/Ti molar ratio to 0.6 led to a slight decrease of de-NOx activity in the whole temperature range. Figure 1b shows the N₂ selectivity as a function of temperature over Ho_xCe_0.35/TiO₂ catalysts. It could be readily observed that the addition of Ho could enhance the N₂ selectivity of Ce_0.35/TiO₂ catalyst. Although all prepared catalysts showed high N₂ selectivity in the temperature range of 100–300°C, Ce_0.35/TiO₂ added with Ho exhibited relatively better N₂ selectivity above 300°C compared with Ce_0.35/TiO₂ catalyst.

Figure 1.

(a) NOx conversion and (b) N₂ selectivity in the NH₃-SCR reaction over Ho_xCe_0.35/TiO₂ catalysts (500 ppm NO, 500 ppm NH₃, 3 vol.% O₂, total flow rate 200 ml min⁻¹ and GHSV = 60 000 h⁻¹).

3.2. Tolerance of SO₂ and H₂O

In practical applications, trace amounts of sulfur dioxide and water are still contained in the exhaust gas through the desulfurization unit, which may result in the deactivation of the catalyst. Therefore, the effect of SO₂ and water on the SCR activity of the catalyst was studied. Figure 2 depicts the catalytic performance of Ce_0.35/TiO₂ and Ho_0.45Ce_0.35/TiO₂, as a function of time in the presence of 200 ppm SO₂ and 8 vol.% water at 200°C. As exhibited in figure 2, the NO conversion over Ce_0.35/TiO₂ decreased from 52% to 33% after introducing SO₂ + H₂O for 200 min, then gradually recovered (37%) after the cut off of SO₂ + H₂O and kept stable during the following test period. By contrast, the presence of SO₂ + H₂O in the feed gas induced a dramatic decrease of NO conversion over Ho_xCe_0.35/TiO₂ by 10%. After eliminating SO₂ + H₂O from the feed gas, the conversion of NO over Ho_xCe_0.35/TiO₂ was gradually restored to a certain extent but is less than the initial value (about 72%). All these analyses implied that a better resistance of SO₂ + H₂O could be achieved by Ho modification.

Figure 2.

NOx conversion of Ce_0.35/TiO₂ and Ho_0.45Ce_0.35/TiO₂ in the presence of SO₂ and H₂O at 200°C (500 ppm NO, 500 ppm NH₃, 3 vol.% O₂, 8 vol.% H₂O, 200 ppm SO₂, total flow rate 200 ml min⁻¹ and GHSV = 60 000 h⁻¹).

3.3. Brunauer–Emmett–Teller results

BET surface area, total pore volume and average pore were tested. As listed in table 1, the specific surface areas of Ce_0.35/TiO₂, Ho_0.35/TiO₂, Ho_0.15Ce_0.35/TiO₂, Ho_0.3Ce_0.35/TiO₂, Ho_0.45Ce_0.35/TiO₂, and Ho_0.6Ce_0.35/TiO₂ are 189.61, 157.34, 196.33, 198.34, 204.56 and 203.65 m² g⁻¹, respectively. It is obvious that the specific surface area of Ho_xCe_0.35/TiO₂ became larger as the Ho/Ti molar ratio increased from 0.15 to 0.45. However, doping excess Ho species to Ce/TiO₂ (Ho/Ti molar ratio = 0.6) may result in a decrease in BET surface area. Considering the SCR activity results from figure 1a, Ce/TiO₂ with proper Ho species modification may possess higher active surface area, which is beneficial for the effective contacts with reactants.

Table 1.

Textural parameters of the catalysts.

samples	BET surface area (m2 g−1)	pore volume (cm3 g−1)	average pore diameter (nm)
Ce0.35/TiO2	189.61	0.612	9.55
Ho0.35/TiO2	157.34	0.424	7.89
Ho0.15Ce0.35/TiO2	196.33	0.608	9.43
Ho0.3Ce0.35/TiO2	198.34	0.627	9.57
Ho0.45Ce0.35/TiO2	204.56	0.628	9.61
Ho0.6Ce0.35/TiO2	203.65	0.611	9.47

3.4. Powder X-ray diffraction results

XRD patterns of Ce_0.35/TiO₂ and Ho_xCe_0.35/TiO₂ are shown in figure 3. Only diffraction peaks assigned to TiO₂ are detected. Specifically, much anatase-phase TiO₂ (PDF-ICDD21-1272) and a little rutile-phase TiO₂ (PDF-ICDD21-1276) are observed. A similar phenomenon was also reported by Liu et al. [21]. It means that Ce and Ho species are highly dispreading on the surface of TiO_2. With the increase of Ho-doping amount, the intensities of all diffraction peaks became weak, suggesting that the introduction of Ho could further reduce the crystallization of TiO₂. All of the factors above are favourable to a good SCR performance.

Figure 3.

XRD patterns of Ho_xCe_0.35/TiO₂ catalysts.

3.5. X-ray photoelectron spectroscopy results

Figure 4 exhibits the X-ray photoelectron spectroscopy (XPS) spectra of Ce 3d and O 1s over Ce_0.35/TiO₂ and Ho_0.45Ce_0.35/TiO₂ catalysts. In addition, the XPS spectrum of Ho 4d over Ho_0.45Ce_0.35/TiO₂ has been given in figure 4c. Table 2 lists the surface element compositions and their chemical states by the XPS technique.

Figure 4.

XPS spectra of Ce_0.35/TiO₂ and Ho_xCe_0.35/TiO₂ catalysts.

Table 2.

Surface elemental analysis by XPS.

samples	atomic concentration (%)				Ce3+/(Ce3+ + Ce4+)	Oβ/(Oβ + Oα)
	Ce	Ti	O	Ho
Ce0.35/TiO2	4.63	13.22	82.15	—	22.33	23.42
Ho0.45Ce0.35/TiO2	4.56	12.98	80.35	2.11	31.45	33.25

As seen in figure 4a, the complicated Ce 3d XPS curves of different samples were made up of eight peaks. u and v peaks belonged to 3d_3/2 and 3d_5/2 spin–orbit components, respectively. u’ and v’ peaks could be attributed to Ce³⁺ and the other peaks could be assigned to Ce⁴⁺ [22]. These Ce³⁺ /Ce⁴⁺ pairs over the catalyst surface were beneficial for not only the storage and release of active oxygen species but also the oxidation of NO to NO₂ [23]. Additionally, more Ce³⁺ would promote the generation of more oxygen vacancies, which help to adsorb reactants [24,25]. The factors mentioned above proved to contribute to the progress of the SCR reaction. Thus, it is necessary to study the ratio of Ce³⁺/(Ce³⁺ + Ce⁴⁺) over the selected catalysts. The ratio of Ce³⁺/(Ce³⁺ + Ce⁴⁺) was calculated according to the peak area ratio of the Ce³⁺ and Ce⁴⁺ peaks. The corresponding results are listed in table 2: Ho_0.45Ce_0.35/TiO₂ (31.45%) and Ce_0.35/TiO₂ (22.33%). Thus, Ho-doping could promote the transformation of Ce⁴⁺ to Ce³⁺ over the catalyst surface, which could also effectively improve the SCR activity of Ce_0.35/TiO₂.

Figure 4b shows that the O 1s XPS spectra of Ce_0.35/TiO₂ and Ho_0.45Ce_0.35/TiO₂ consisted of two peaks, lattice oxygen (binding energy = 529.8 eV, labelled as O_α) and chemisorbed oxygen (binding energy = 532 eV, labelled as O_β) [26,27]. It is well recognized that O_β is more active than O_α in the oxidation reactions of NO to NO₂ [28], which is beneficial for the ‘fast SCR’ reaction (NO + NO₂ + 2NH₃ = 2N₂ + 3H₂O). ‘Fast SCR’ reaction has been proved conducive to the improvement of the low-temperature SCR activity [29]. The O_β/(O_α + O_β) ratio was calculated and is presented in table 2. It could be observed that Ho_0.45Ce_0.35/TiO₂ has a bigger O_β/(O_α + O_β) ratio than Ce_0.35/TiO₂, which meant that chemisorbed oxygen over the catalyst surface of Ce_0.35/TiO₂ with Ho modification was obviously improved. Considering the results of the SCR activity and Ce 3d XPS, the O_β ratio result is corresponding with the Ce³⁺ ratio and SCR activity. It may be concluded that more Ce³⁺ was accompanied by an increment of oxygen vacancies and active oxygen species, which played a positive role in the SCR activity. Finally, the XPS spectrum of Ho 4d over Ho_0.45Ce_0.35/TiO₂ is exhibited in figure 4c.

3.6. H₂temperature-programmed reduction results

H₂-TPR was performed for studying the redox ability of catalysts. In figure 5, no obvious reduction peak of Ho_0.35/TiO₂ is observed. The reduction peak of Ce_0.35/TiO₂ at about 530°C belonged to the reduction of Ce⁴⁺ to Ce³⁺ [30,31]. With the introduction of Ho to Ce_0.35/TiO₂, the reduction peak of surface Ce⁴⁺ moved to lower temperature, which could significantly improve the mobility of surface O owing to the strong synergetic effect between Ti, Ce and Ho species. It was also reported that the synergetic effect could lead to the rise of abundant O defects [32,33]. More O defects were beneficial for the improvement of SCR activity because they could promote O diffusion from the subsurface layer and progressively proceed more in-depth into the bulk [34,35]. It could also be observed that Ho_0.45Ce_0.35/TiO₂ showed the lowest reduction temperature at 446°C and this result is corresponding with its best SCR performance. It seems that further increasing the Ho amount would increase the catalyst reduction temperature. In conclusion, the stronger oxidation–reduction ability of Ho_0.45Ce_0.35/TiO₂ is beneficial for the outstanding SCR reaction performance.

Figure 5.

H₂-TPR patterns of Ho_xCe_0.35/TiO₂ catalysts.

3.7. NH₃ temperature-programmed desorption

Figure 6 shows the effect of Ho modification on NH₃ desorption behaviour of the prepared samples. From figure 6, no obvious desorption peak of Ho_0.35/TiO₂ was observed and the peak area of Ce_0.35/TiO₂ is shallow. After the introduction of Ho, the peak surface area gradually increases and the NH₃-TPD profiles existed as a broad peak with the full range of 120–450°C, which included physically adsorbed NH₃, the chemically adsorbed species including adsorbed NH₃ species on Brønsted acid sites and strongly adsorbed on Lewis acid sites [33,36,37]. Thus, more surface sites were available on the Ce_0.35/TiO₂ surface for NH₃ adsorption after introducing Ho, which could be evidenced by the largest desorption peak area of Ho_0.45Ce_0.35/TiO₂. The phenomenon could also indicate that Ho_0.45Ce_0.35/TiO₂ possesses the most potent surface acidity. Thus the adsorption of NH₃ over it could be boosted and the SCR activity could be promoted correspondingly.

Figure 6.

NH₃-TPD patterns of Ho_xCe_0.35/TiO₂ catalysts.

3.8. NO oxidation

Figure 7 exhibits the NO conversion of NO oxidation reaction over the prepared catalysts. It could be easily seen that the NO oxidation conversions over Ce_0.35/TiO₂ and Ho_0.35/TiO₂ are very low (below 25%) during 100–400°C, which is consistent with the lowest SCR activity due to the inefficient conversion from NO to NO₂. The activity curves of other catalyst samples demonstrate a parabolic trend, which is an indication of the conversion from the kinetically controlled regime to thermo-dynamically controlled regime [38]. Especially, Ho_0.45Ce_0.35/TiO₂ has a more significant effect on NO oxidation than other samples. The formation of more NO₂ on the catalyst surface facilitates NOx reduction in the low-temperature range, which was also corresponding with the XPS results. Although Ho_0.6Ce_0.35/TiO₂ had the highest oxidation activity of NO to NO₂ of all samples, it exhibited a relatively lower de-NOx activity compared with Ho_0.45Ce_0.35/TiO₂, which may be attributed to its decreased specific surface area leading to the decreased adsorbed NH₃ species.

Figure 7.

Oxidation activity of NO to NO₂ by O₂ over different catalysts (500 ppm NO, 3 vol.% O₂ and 200 ml min⁻¹ total flow rate).

3.9. In situ diffuse reflectance infrared Fourier transform spectroscopy results

3.9.1. NH₃ adsorption

Figure 8a shows the DRIFT spectra of NH₃ adsorption over Ce_0.35/TiO₂ at different temperatures. The bands at 1599, 1161 cm⁻¹ with a shoulder at 1109 cm⁻¹ attributed to the coordinated NH₃ linked to Lewis acid sites (NH₃-L) [39,40] could be observed. The band at 1418 cm⁻¹ could be assigned to NH₄⁺ species on Brønsted acid sites (NH₄⁺-B). Notably, all the bands linked to NH₃ species decrease with the temperature increasing owing to the desorption effect.

Figure 8.

In situ DRIFTS of NH₃ adsorption with increasing temperature from 50 to 350°C: (a) Ce_0.35/TiO₂ and (b) Ho_0.45Ce_0.35/TiO₂.

Figure 8b exhibits the DRIFT spectra of NH₃ adsorption over Ho_xCe_0.35/TiO₂. Similar to the spectra over Ho_xCe_0.35/TiO_2, the NH₃-L bands (1599 and 1143 cm⁻¹) and the NH₄⁺-B band (1432 cm⁻¹) could also be seen. However, the band intensity of adsorbed NH₃ over Ho_xCe_0.35/TiO₂ was much stronger than that over Ce_0.35/TiO_2, which indicated that the introduction of Ho species could greatly increase the quantity of both Lewis acid sites and Brønsted acid sites. Previous study by Chen et al. [41] and Zhou et al. [42] reported that more Brønsted acid sites could help in the generation of adsorbed NH₃ species, thus promoting the low-temperature SCR performance. It should also be noted that the intensity of the bands at 1432 cm⁻¹ assigned to Brønsted acid sites in figure 8b decreases faster with temperature rising in comparison with those assigned to Lewis acid sites, suggesting NH₃ bonded to Lewis acid sites possessed a better thermostability than that bonded to Brønsted acid sites [41].

3.9.2. NO + O₂ adsorption

Figure 9a shows the DRIFT spectra of NO + O₂ adsorption over Ce_0.35/TiO₂ at different temperatures. The bands at 1577, 1536 cm⁻¹ attributed to bidentate nitrate could be clearly observed; the band at 1599 cm⁻¹ could be assigned to ad-NO₂ and the band at 1241 cm⁻¹ could be assigned to bridging nitrates [43–45]. It could be observed that all the bands decrease with the temperature increasing owing to the drop in thermal stability.

Figure 9.

In situ DRIFTS of NO + O₂ adsorption with increasing temperature from 50 to 350°C: (a) Ce_0.35/TiO₂ and (b) Ho_0.45Ce_0.35/TiO₂.

Figure 9b exhibits the DRIFT spectra of NO + O₂ adsorption over Ho_0.45Ce_0.35/TiO₂ at different temperatures. As shown in figure 9b, the peaks at 1600 cm⁻¹ and 1564 cm⁻¹ belonged to ad-NO₂ and bidentate nitrate. The peak at 1232 cm⁻¹ belonged to bridging nitrates [44]. In comparison with that shown in figure 9a, the peak intensity of Ho_0.45Ce_0.35/TiO₂ was stronger than that of Ce_0.35/TiO₂, which meant that Ho-doping could greatly improve NOx adsorption of Ce_0.35/TiO₂ catalyst.

3.10. Promotion mechanism

As evidenced by electronic supplementary material, figure S1, all the adsorbed reactants, including ad-NH₃ and ad-NO_X on Ho_0.45Ce_0.35/TiO₂, could participate in the NH₃-SCR reaction. Considering all analysis results given above, doping proper amount of Ho into Ce_0.35/TiO₂ could generate more active NH₃ and NO_x species on its surface. After adding Ho species, the generation of more Ce³⁺ and O_β over Ho_0.45Ce_0.35/TiO₂ has a facilitation effect on the conversion from NO to NO₂. Thus, the Langmuir–Hinshelwood (L–H) mechanism and Eley–Rideal (E–R) mechanism should be mainly responsible for the promoted low-temperature NH₃-SCR activity over Ho_0.45Ce_0.35/TiO₂, which could be described by the following processes:

• (1) L–H mechanism:

  $$\mathrm{NO}+{\mathrm{O}}_2(\mathrm{g})\to {\mathrm{NO}}_2(\mathrm{ad})$$ 3.1

  $$\mathrm{N}{\mathrm{H}}_3(\mathrm{g})\stackrel{\mathrm{C}{\mathrm{e}}^{4+}}{\to }\mathrm{N}{\mathrm{H}}_3(\mathrm{ad})(\mathrm{Lewis}\hspace{0.17em}\mathrm{acid}\hspace{0.17em}\mathrm{sites}).$$ 3.2

‘Fast SCR’ reaction:

$${\mathrm{NO}}_2(\mathrm{ad})+2\mathrm{N}{\mathrm{H}}_3(\mathrm{ad})+\mathrm{NO}(\mathrm{g})\to 2{\mathrm{N}}_2(\mathrm{g})+3{\mathrm{H}}_2\mathrm{O}(\mathrm{g})$$ 3.3

$$\mathrm{N}{\mathrm{H}}_3(\mathrm{g})\stackrel{\mathrm{C}{\mathrm{e}}^{3+}}{\to }{\mathrm{NH}}_4^{+}\hspace{0.17em}(\mathrm{ad})\hspace{0.17em}\text{over Brønsted acid sites}$$ 3.4

$$\hspace{0.17em}{\mathrm{NH}}_4^{+}+{\mathrm{e}}^{-}+{\mathrm{NO}}_2(\mathrm{ad})\to \mathrm{N}{\mathrm{H}}_4{\mathrm{NO}}_2(\mathrm{ad})\to {\mathrm{N}}_2+{\mathrm{H}}_2\mathrm{O}.$$ 3.5

• (2) E–R mechanism:

  $$\mathrm{N}{\mathrm{H}}_3(\mathrm{g})\stackrel{\mathrm{C}{\mathrm{e}}^{4+}}{\to }\mathrm{N}{\mathrm{H}}_3(\mathrm{ad})(\mathrm{on}\hspace{0.17em}\mathrm{Lewis}\hspace{0.17em}\mathrm{acid}\hspace{0.17em}\mathrm{sites})$$ 3.6

  $${\mathrm{O}}_2(\mathrm{g})\to 2\mathrm{O}(\mathrm{ad})$$ 3.7

  $$\mathrm{N}{\mathrm{H}}_3(\mathrm{ad})+\mathrm{O}(\mathrm{ad})\to \mathrm{N}{\mathrm{H}}_2(\mathrm{ad})+\mathrm{OH}(\mathrm{ad})$$ 3.8

  $$\mathrm{NO}(\mathrm{g})+\mathrm{N}{\mathrm{H}}_2(\mathrm{ad})\to \mathrm{N}{\mathrm{H}}_2\mathrm{NO}(\mathrm{ad})$$ 3.9

  $$\mathrm{N}{\mathrm{H}}_2\mathrm{NO}(\mathrm{ad})\to {\mathrm{N}}_2(\mathrm{g})+{\mathrm{H}}_2\mathrm{O}.$$ 3.10

4. Conclusion

In summary, Ce_0.35/TiO₂ modified with a certain amount of Ho shows an outstanding low-temperature SCR performance and superior SO₂ + H₂O durability, which could boost the practical application of Ce/TiO₂. In situ DRIFTS results revealed that the introduction of Ho species could efficiently promote both active ad-NH₃ and ad-NOx species on Ce_0.35/TiO_2. Moreover, all of these could contribute to the low-temperature SCR activity of Ce_0.35Ho_0.45/TiO₂ through L–H route and E–R route.

Ethics

Shanghai University Academic Committee approved the study, and the study was also approved by the National Key Research and Development Program of China (no. 2017YFB0404503). Informed consent for the participants to participate in the study has been received. All authors have been personally and actively involved in substantive work leading to the report, and will hold themselves jointly and individually responsible for its content.

Data accessibility

Our data are deposited at: http://dx.doi.org/10.5061/dryad.c86d5m0 [46].

Authors' contributions

T.-t.Z. designed the study, performed the laboratory experiment and wrote the manuscript. L.-m.Y. assisted in analysing experimental data and editing the manuscript for important intellectual content, and gave the final approval for publication.

Competing interests

We declare we have no competing interests.

Funding

Financial support came from the National Key Research and Development Program of China, no. 2017YFB0404503.