Substitution of WO₃ in V₂O₅/WO₃–TiO₂ by Fe₂O₃ for selective catalytic reduction of NO with NH₃

Abstract

To improve N₂ selectivity and lower the cost, WO₃ in V₂O₅/WO₃–TiO₂ was substituted by a low cost composition Fe₂O₃ for selective catalytic reduction (SCR) of NO with NH₃. The SCR reaction over V₂O₅/Fe₂O₃–TiO₂ mainly followed the Eley–Rideal mechanism (i.e. the reaction between activated ammonia species and gaseous NO). There were two active components on V₂O₅/WO₃–TiO₂ for the activation of adsorbed NH₃ (i.e. V⁵⁺ and Fe³⁺). The acid sites on V₂O₅/Fe₂O₃–TiO₂ mainly resulted from the support Fe₂O₃–TiO₂, so the adsorbed NH₃ preferred to be activated by Fe³⁺ rather than by V⁵⁺. V⁵⁺ on V₂O₅/Fe₂O₃–TiO₂ could accelerate the regeneration of Fe³⁺ on Fe₂O₃–TiO₂ due to the rapid electron transfer between V⁵⁺ and Fe²⁺ on the surface, so the activation of adsorbed NH₃ by Fe³⁺ was promoted. As some NH₃ adsorbed on V₂O₅/Fe₂O₃–TiO₂ was not activated by Fe³⁺, the inactivated NH₃ could then be activated by V⁵⁺ on the surface. As a result, 2% V₂O₅/Fe₂O₃–TiO₂ showed excellent SCR activity, N₂ selectivity and H₂O/SO₂ durability at 300–450 °C. Furthermore, the emission of 2% V₂O₅/Fe₂O₃–TiO₂ to the fly ash can be prevented by an external magnetic field due to its inherent magnetization. Therefore, 2% V₂O₅/Fe₂O₃–TiO₂ could be a promising low-cost catalyst in NO emission control.

---

1. Introduction

Nitrogen oxides (NO and NO₂), which result from automobile exhaust gas and industrial combustion of fossil fuels, have been a major contribution to air pollution.¹ They contribute to photochemical smog, acid rain, ozone depletion and the greenhouse effect. So far, the most efficient technology for the removal of nitrogen oxides from coal-fired power plants has been selective catalytic reduction (SCR) of NO with NH₃.² The standard SCR process is based on the following reaction between NH₃ and NO: Many metal oxide based systems, such as V₂O₅/WO₃–TiO₂,³ CeO₂/WO₃–TiO₂,⁴ iron titanate⁵ and Fe–Ti spinel,⁶ have been extensively studied. Among these catalysts, V₂O₅/WO₃–TiO₂ has been widely accepted as an industrial catalyst. Although V₂O₅/WO₃–TiO₂ has been employed as a commercial SCR catalyst to control the emission of NO_x from coal-fired power plants for several decades, some problems still exist, such as the relatively narrow temperature window of 300–400 °C, the low N₂ selectivity at high temperature, and the toxicity of vanadium pentoxide to the environment.⁷ Especially, the price of W resources is now soaring. Therefore, a more cost-effective and more environmentally-friendly SCR catalyst with better N₂ selectivity should be developed.

In this study, WO₃ in V₂O₅/WO₃–TiO₂ was substituted by a low cost composition Fe₂O₃ to improve its N₂ selectivity and lower its cost for SCR of NO with NH₃. The support Fe₂O₃–TiO₂ was prepared using a co-precipitation method at room temperature, and V₂O₅ was supported on Fe₂O₃–TiO₂ by the conventional impregnation method. Then, V₂O₅/Fe₂O₃–TiO₂ was characterized using nitrogen physisorption, X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), magnetization and H₂ temperature-programmed reduction (TPR). Subsequently, the SCR performance of V₂O₅/Fe₂O₃–TiO₂ was estimated using a packed-bed microreactor. At last, the mechanism of the SCR reaction over V₂O₅/Fe₂O₃–TiO₂ was studied using in situ DRIFTS.

2. Experimental

2.1 Catalyst preparation

The support Fe₂O₃–TiO₂ was prepared using a co-precipitation method at room temperature.⁸ Suitable amounts of ferrous sulfate, ferric trichloride, and titanium tetrachloride (Fe²⁺ : Fe³⁺ : Ti⁴⁺ = 3 : 2 : 1) were dissolved in a HCl solution. This mixture was added to an ammonia solution leading to an instantaneous precipitation. During the reaction, the system was continuously stirred at 800 rpm. The particles were then separated by centrifugation at 4500 rpm for 5 min and washed with distilled water followed by a new centrifugation. After 4 washings, the particles were collected and dried at 105 °C for 12 h.

V₂O₅/Fe₂O₃–TiO₂ catalysts with 1 wt% and 2 wt% V₂O₅ were prepared by the conventional impregnation method using NH₄VO₃ and H₂C₂O₄·2H₂O as precursors, and Fe₂O₃–TiO₂ as a support. After the impregnation, excess water was removed in a rotary evaporator at 80 °C. The sample was dried at 105 °C overnight and then calcined at 500 °C for 3 h under air. Meanwhile, Fe₂O₃–TiO₂ was calcined at 500 °C for 3 h under air as a comparison. Furthermore, conventional vanadium-based catalysts (V₂O₅/WO₃–TiO₂) with 1 wt% V₂O₅ and 10 wt% WO₃, and 2 wt% V₂O₅ and 10 wt% WO₃ were prepared by the conventional impregnation method using NH₄VO₃, (NH₄)₁₀W₁₂O₄₁ and H₂C₂O₄·2H₂O as precursors, and anatase TiO₂ as a support.

2.2 Catalyst characterization

Crystal structure was determined using an X-ray diffractionmeter (Rigaku, D/max-2200/PC) between 10° and 80° at a step of 7° min⁻¹ operating at 30 kV and 30 mA using Cu Kα radiation. BET surface area was determined using a nitrogen adsorption apparatus (Quantachrome, Autosorb-1). The catalyst was outgassed at 200 °C before BET measurement. H₂-TPR was recorded on a chemisorption analyzer (Micromeritics, ChemiSorb 2720 TPx) under a 10% hydrogen–90% nitrogen gas flow (50 cm³ min⁻¹) at a rate of 10 °C min⁻¹. Saturation magnetization was determined using a vibrating sample magnetometer (VSM, Model JDM-13) at room temperature. X-ray photoelectron spectroscopy (Thermo, ESCALAB 250) was used to determine the Fe 2p, V 2p, Ti 2p and O 1s binding energies with Al Kα (h_v = 1486.6 eV) as the excitation source. The C 1s line at 284.6 eV was taken as a reference for the binding energy calibration.

2.3 Catalytic test

SCR reaction was performed on a fixed-bed quartz tube reactor (6 mm of internal diameter) containing 100–200 mg of the catalyst (40–60 mesh). The typical reactant gas composition was as follows: 500 ppm of NO, 500 ppm of NH₃, 2% of O₂, 10% of H₂O (when used), 200 ppm of SO₂ (when used), and the balance of N₂. The total flow rate was 200 mL min⁻¹, and the gas hourly space velocity (GHSV) varied from about 60 000 to 120 000 cm³ g⁻¹ h⁻¹. The concentrations of NO, NO₂, NH₃ and N₂O were continually monitored by an FTIR spectrometer (Gasmet FTIR DX4000).

As the SCR reaction reached the steady state, the ratios of NO conversion and N₂ selectivity were calculated according to the following equations:

2.4 In situ DRIFTS study

In situ DRIFT spectra were recorded on a Fourier transform infrared spectrometer (FTIR, Nicolet NEXUS 870) equipped with a smart collector and an MCT detector cooled by liquid N₂. The diffuse reflectance measurement was carried out in situ in a high temperature cell with a ZnSe window. The FTIR spectra were recorded by accumulating 100 scans with a resolution of 4 cm⁻¹.

3. Results

3.1 Characterization

3.1.1 XRD and BET. XRD patterns of synthesized catalysts are shown in Fig. 1. The characteristic peaks of the support Fe₂O₃–TiO₂ corresponded very well to Fe–Ti spinel, and the characteristic peaks corresponding to rutile and anatase did not appear. It indicates that Ti could be introduced into the spinel structure.⁹ Furthermore, electron energy loss spectroscopy (EELS), X-ray adsorption near edge structure (XANES) and extended X-ray adsorption fine structure (EXAFS) also demonstrated that Ti could be incorporated into the spinel structure.¹⁰

Fig. 1 XRD patterns of: (a) Fe₂O₃–TiO₂; (b) 1% V₂O₅/Fe₂O₃–TiO₂; (c) 2% V₂O₅/Fe₂O₃–TiO₂.

After the loading of V₂O₅, the characteristic peaks of V₂O₅/Fe₂O₃–TiO₂ still corresponded very well to Fe–Ti spinel, and additional reflections that would indicate the presence of other crystalline vanadium oxides, such as V₂O₅, VO₂, V₂O₄, V₂O₃ or FeVO₄, were not present in the diffraction scan. It indicates that V cations could mainly present as an amorphous phase of VO_x, which was well dispersed on Fe₂O₃–TiO₂.

The BET surface areas of Fe₂O₃–TiO₂, 1% V₂O₅/Fe₂O₃–TiO₂ and 2% V₂O₅/Fe₂O₃–TiO₂ were 81.4, 73.9 and 70.0 m² g⁻¹, respectively.

3.1.2 XPS. Surface information of V₂O₅/Fe₂O₃–TiO₂ was analyzed by XPS. XPS spectra over the spectral regions of Fe 2p, Ti 2p, O 1s and V 2p were evaluated (shown in Fig. 2).

Fig. 2 XPS spectra of V₂O₅/Fe₂O₃–TiO₂ over the spectral regions of Fe 2p, Ti 2p, O 1s and V 2p.

The Fe peaks of Fe₂O₃–TiO₂ were assigned to oxidized Fe species, more likely Fe³⁺ type species. The binding energies centered at about 709.9 eV and 711.0 eV could be assigned to Fe³⁺ in the spinel structure, and the binding energy centered at about 712.5 eV could be ascribed to Fe³⁺ bonded with hydroxyl groups.^11,12 The Ti peaks of Fe₂O₃–TiO₂ were assigned to Ti 2p 1/2 (464.0 eV) and Ti 2p 3/2 (458.3 eV) of Ti⁴⁺.¹³ The O 1s peaks of Fe₂O₃–TiO₂ mainly centered at about 529.8 eV, as expected for the transition metal oxides. Another oxygen species centered at about 532.0 eV was also observed, which was assigned to −OH.⁸

After the loading of V₂O₅ on Fe₂O₃–TiO₂, no obvious changes happened to XPS spectra of V₂O₅/Fe₂O₃–TiO₂ over the spectral regions of Fe 2p, Ti 2p and O 1s (shown in Fig. 2). The V 2p peak mainly centered at about 516.8 eV, which was assigned to V⁵⁺.¹⁴

The ratios of Fe³⁺, Ti⁴⁺, O²⁻ and V⁵⁺ on Fe₂O₃–TiO₂, 1% V₂O₅/Fe₂O₃–TiO₂ and 2% V₂O₅/Fe₂O₃–TiO₂ collected from XPS spectra are shown in Table 1. The percentage of V⁵⁺ on the surfaces of 1% V₂O₅/Fe₂O₃–TiO₂ and 2% V₂O₅/Fe₂O₃–TiO₂ was much more than the content of V⁵⁺ in 1% V₂O₅/Fe₂O₃–TiO₂ and 2% V₂O₅/Fe₂O₃–TiO₂, respectively. It indicates that V⁵⁺ could mainly present as an amorphous phase of V₂O₅ on Fe₂O₃–TiO₂. As V₂O₅ was loaded on Fe₂O₃–TiO₂, the percent of Fe³⁺ on V₂O₅/Fe₂O₃–TiO₂ obviously decreased (shown in Table 1). It suggests that the loaded V⁵⁺ could mainly cover Fe³⁺ on Fe₂O₃–TiO₂.

Table 1 Data of atomic ratios on V₂O₅/Fe₂O₃–TiO₂ collected from XPS/%

	Fe^3+	Ti^4+	O^2−	V^5+
Fe2O3–TiO2	28.5	9.5	62.0	—
1% V2O5/Fe2O3–TiO2	26.0	9.7	62.6	1.7
2% V2O5/Fe2O3–TiO2	22.0	7.9	64.1	6.0

---

3.1.3 TPR. H₂-TPR profiles of V₂O₅/Fe₂O₃–TiO₂ are illustrated in Fig. 3. TPR profiles of Fe₂O₃–TiO₂ showed two obvious reduction peaks. The first peak centered at about 442 °C corresponded to the reduction of non-stoichiometric (Fe_2.5Ti_0.5)_1-δO₄ to stoichiometric Fe_2.5Ti_0.5O₄, and the broad peak at higher temperature was attributed to the reduction of Fe_2.5Ti_0.5O₄ to Fe and TiO₂.⁸ Our previous study demonstrated that the reduced Fe³⁺ corresponding to the first reduction peak of Fe₂O₃–TiO₂ mainly located in the octahedral site, which resulted from the oxidization of Fe²⁺ in the octahedral site of stoichiometric Fe_2.5Ti_0.5O₄.⁸

Fig. 3 H₂-TPR profiles of: (a) Fe₂O₃–TiO₂; (b) 1% V₂O₅/Fe₂O₃–TiO₂; (c) 2% V₂O₅/Fe₂O₃–TiO_2.

After the loading of V₂O₅, a strong displacement of the first reduction peak to low temperature happened in the TPR profiles (shown in Fig. 3). It suggests that the oxidization ability of Fe₂O₃–TiO₂ was gradually enhanced with the increase of V₂O₅ loading.

Because V₂O₅ was loaded on Fe₂O₃–TiO₂, the first step of V₂O₅/Fe₂O₃–TiO₂ reduction was the reduction of V⁵⁺ on the surface (reaction (4)), and the next step could be the reduction of V⁴⁺ on the surface (reaction (5)).

2≡V_Surf⁵⁺ + H₂ → 2≡V_Surf⁴⁺ + 2H⁺ (4)

2≡V_Surf⁴⁺ + H₂ → 2≡V_Surf³⁺ + 2H⁺ (5)

The electrical property of Fe–Ti spinel is similar to that of magnetite, and its band gap is very small.¹⁵ Hence, it has the lower resistivity and its conductivity is almost metallic. As a result, the electron can migrate easily from the bulk to the surface. The redox potential of the reduction of Fe³⁺ by V³⁺ is 0.41 V, so reaction (6) is thermodynamically favorable.¹⁶ Therefore, the formed V³⁺ on the surface could be reoxidized to V⁴⁺ by the reducible Fe³⁺ in the bulk. Then, the formed V⁴⁺ on the surface was reduced by gaseous H₂ again. Through the recycle, the reduction of Fe³⁺ in the octahedral site of Fe₂O₃–TiO₂ was promoted. As a result, the first reduction peak of V₂O₅/Fe₂O₃–TiO₂ obviously shifted to low temperature although the amount of V₂O₅ loaded was very small.

Fe_bulk³⁺ + ≡V_Surf³⁺ → Fe_bulk²⁺ + ≡V_Surf⁴⁺ E⁰ = 0.41 V (6)

3.1.4 Magnetization. Vanadium(V) species (e.g., vanadium pentoxide, V₂O₅) are considered particularly toxic with mutagenic effects, respiratory tract toxicity, and possible carcinogenic activity.¹⁷ SCR catalyst can be abraded by the fly ash, and then emits to the fly ash.¹⁸ Therefore, the emission of vanadium based catalyst to the environment during the SCR reaction is a serious concern.

As is well known, the support Fe₂O₃–TiO₂ is a magnetic material.¹⁹ 2% V₂O₅/Fe₂O₃–TiO₂ had the super-paramagnetism with a minimized coercivity and a negligible magnetization hysteresis, and its saturation magnetization was about 43 emu g⁻¹ (shown in Fig. 4). The magnetization characteristics ensure that the emission of 2% V₂O₅/Fe₂O₃–TiO₂ to the fly ash during the SCR reaction can be effectively prevented by exposure to an external magnetic field.

Fig. 4 Magnetization characteristics of 2% V₂O₅/Fe₂O₃–TiO_2.

3.2 SCR performance

3.2.1 SCR activity and selectivity. 1% V₂O₅/WO₃–TiO₂ showed a moderate SCR activity at 200–250 °C, and an excellent SCR activity at 300–500 °C (shown in Fig. 5a). As the loading of V₂O₅ increased from 1% to 2%, the SCR reaction over V₂O₅/WO₃–TiO₂ was obviously promoted at 150–300 °C. However, NO conversion over 2% V₂O₅/WO₃–TiO₂ gradually decreased with the further increase in reaction temperature from 350 to 500 °C. N₂ selectivity of V₂O₅/WO₃–TiO₂ gradually decreased above 300 °C, and N₂ selectivity of 2% V₂O₅/WO₃–TiO₂ was much worse than that of 1% V₂O₅/WO₃–TiO₂ (shown in Fig. 5b).

Fig. 5 SCR performance of synthesized catalysts: (a) NO conversion; (b) N₂ selectivity. Reaction conditions: [NO] = [NH₃] = 500 ppm, [O₂] = 2 vol%, catalyst mass = 100 mg, total flow rate = 200 mL min⁻¹, GHSV=120 000 cm³ g⁻¹ h⁻¹.

Fe₂O₃–TiO₂ showed a moderate SCR activity at 200–300 °C, and an excellent SCR activity at 350–450 °C (shown in Fig. 5a). Meanwhile, little N₂O can be observed during the SCR reaction over Fe₂O₃–TiO₂ (shown in Fig. 5b). The SCR activity of Fe₂O₃–TiO₂ was obviously promoted due to the loading of V₂O₅, and NO conversion over V₂O₅/Fe₂O₃–TiO₂ increased obviously with the increase of V₂O₅ loading (shown in Fig. 5a). Although there was a large amount of V⁵⁺ on V₂O₅/Fe₂O₃–TiO₂ (shown in Table 1), V₂O₅/Fe₂O₃–TiO₂ showed an excellent N₂ selectivity. N₂ selectivity of 2% V₂O₅/Fe₂O₃–TiO₂ only slightly decreased to about 90% as the reaction temperature increased to 500 °C, which was much better than those of 1% V₂O₅/WO₃–TiO₂ and 2% V₂O₅/WO₃–TiO₂ (shown in Fig. 5b).

3.2.2 Effect of H₂O and SO₂. Water vapor and SO₂ in the flue gas often lead to the deactivation of the SCR catalyst.²⁰ Therefore, the effect of 10% of H₂O and 200 ppm of SO₂ on the SCR activity of 2% V₂O₅/Fe₂O₃–TiO₂ was investigated with a 24 h test. As shown in Fig. 6a, the presence of H₂O and SO₂ showed a severe interference with the SCR reaction over 2% V₂O₅/Fe₂O₃–TiO₂ at 150–250 °C. However, the ratio of NO conversion over 2% V₂O₅/Fe₂O₃–TiO₂ in the presence of H₂O and SO₂ can reach above 95% at 300–450 °C, and it did not decrease in the 24 h test. This performance was close to that of 1% V₂O₅/WO₃–TiO₂ (shown in Fig. 6b). It indicates that 2% V₂O₅/Fe₂O₃–TiO₂ had an excellent H₂O and SO₂ durability in the temperature range of the SCR unit of the coal-fired power plant.

Fig. 6 Effect of 10% of H₂O and 200 ppm of SO₂ on the SCR reaction over: (a) 2% V₂O₅/Fe₂O₃–TiO₂; (b) 1% V₂O₅/WO₃–TiO₂. Reaction condition: [NO] = [NH₃] = 500 ppm, [O₂] = 2 vol%, catalyst mass = 200 mg, total flow rate = 200 mL min⁻¹, GHSV=60 000 cm³ g⁻¹ h⁻¹.

3.3 In situ DRIFTS study

3.3.1 In situ DRIFTS study on V₂O₅/WO₃–TiO₂. After 2% V₂O₅/WO₃–TiO₂ was treated with NH₃/N₂ at 250 °C, the characteristic vibrations mainly centered at 1425 and 1245 cm⁻¹ (shown in Fig. 7a). The band at 1425 cm⁻¹ was assigned to ionic NH₄⁺ bound to the Brønsted acid sites, and the band at 1245 cm⁻¹ was attributed to coordinated NH₃ bound to the Lewis acid sites.²¹ After NO + O₂/N₂ passed over NH₃ pretreated 2% V₂O₅/WO₃–TiO₂ at 250 °C, both ionic NH₄⁺ and coordinated NH₃ diminished (shown in Fig. 7a). Meanwhile, adsorbed H₂O, which is a product of the SCR reaction, can be clearly observed at 1620 cm⁻¹.²² They both suggest that adsorbed ammonia species on 2% V₂O₅/WO₃–TiO₂ can react with NO.

Fig. 7 (a) In situ DRIFT spectra taken at 250 °C upon passing NO + O₂ over NH₃ presorbed 2% V₂O₅/WO₃–TiO₂; (b) in situ DRIFT spectra taken at 250 °C upon passing NH₃ over NO + O₂ presorbed 2% V₂O₅/WO₃–TiO₂.

After 2% V₂O₅/WO₃–TiO₂ was treated with NO + O₂/N₂ at 250 °C, the bands corresponding to adsorbed NO_x species were hardly observed (shown in Fig. 7b). It suggests that gaseous NO can hardly adsorb on 2% V₂O₅/WO₃–TiO₂ to form adsorbed NO_x. After NH₃/N₂ passed over NO + O₂ pretreated 2% V₂O₅/WO₃–TiO₂, its surface was mainly covered by ionic NH₄⁺ bound to the Brønsted acid sites and coordinated NH₃ bound to the Lewis acid sites. Furthermore, adsorbed H₂O, which is a product of the SCR reaction, can hardly be detected at about 1620 cm⁻¹. It suggests that the reaction between ammonia and adsorbed nitrogen oxide species cannot happen on 2% V₂O₅/WO₃–TiO₂. Therefore, the SCR reaction over V₂O₅/WO₃–TiO₂ mainly followed the Eley–Rideal mechanism.

3.3.2 In situ DRIFTS study on Fe₂O₃–TiO₂. Fe₂O₃–TiO₂ was mainly covered by ionic NH₄⁺ bound to the Brønsted acid sites (1430 cm⁻¹) and coordinated NH₃ bound to the Lewis acid sites (1221 and 1606 cm⁻¹) after the treatment of NH₃/N₂ at 250 °C (shown in Fig. 8a). After NO + O₂/N₂ passed over NH₃ pretreated Fe₂O₃–TiO₂ at 250 °C, both ionic NH₄⁺ and coordinated NH₃ diminished (shown in Fig. 8a). It suggests that adsorbed ammonia species on Fe₂O₃–TiO₂ can react with NO. Meanwhile, three new vibrations at 1614, 1576 and 1551 cm⁻¹ appeared, which were assigned to monodentate nitrite and monodentate nitrate respectively.²²

Fig. 8 (a) In situ DRIFT spectra taken at 250 °C upon passing NO + O₂ over NH₃ presorbed 2% Fe₂O₃–TiO₂; (b) in situ DRIFT spectra taken at 250 °C upon passing NH₃ over NO + O₂ presorbed Fe₂O₃–TiO₂; (c) in situ DRIFT spectra taken at 250 °C upon passing NO + O₂ + NH₃ over 2% Fe₂O₃–TiO₂.

After the adsorption of NO + O₂ at 250 °C, Fe₂O₃–TiO₂ was mainly covered by monodentate nitrite and monodentate nitrate (1614, 1576, 1551 and 1371 cm⁻¹). After NH₃/N₂ passed over NO + O₂ pretreated Fe₂O₃–TiO₂ at 250 °C, the bands corresponding to adsorbed NO_x species disappeared. Meanwhile, the bands at 1606, 1430 and 1221 cm⁻¹ corresponding to adsorbed ammonia species appeared. It suggests that adsorbed NO_x could take part in the SCR reaction over Fe₂O₃–TiO₂.

At last, the IR spectra during the SCR reaction (i.e. NH₃ and NO + O₂ were simultaneously introduced) over Fe₂O₃–TiO₂ at 250 °C were recorded. As shown in Fig. 8c, Fe₂O₃–TiO₂ was mainly covered by ionic NH₄⁺ bound to the Brønsted acid sites (1430 cm⁻¹) and coordinated NH₃ bound to the Lewis acid sites (1221 and 1606 cm⁻¹), and the characteristic vibrations corresponding to adsorbed NO_x species were hardly observed. It suggests that the adsorption of NO + O₂ on Fe₂O₃–TiO₂ could not happen in the presence of NH₃. There is general agreement that the SCR reaction starts with the adsorption of NH₃, which is very strong compared to the adsorption of NO + O₂ and the reaction products.²³ Therefore, The SCR reaction over Fe₂O₃–TiO₂ mainly followed the Eley–Rideal mechanism.

3.3.3 In situ DRIFTS study on V₂O₅/Fe₂O₃–TiO₂. After 2% V₂O₅/Fe₂O₃–TiO₂ was treated with NH₃/N₂ at 250 °C, V₂O₅/Fe₂O₃–TiO₂ was mainly covered by ionic NH₄⁺ bound to the Brønsted acid sites (1430 cm⁻¹) and coordinated NH₃ bound to the Lewis acid sites (1221 and 1606 cm⁻¹). The characteristic vibration corresponding to NH₃ adsorbed on V₂O₅/Fe₂O₃–TiO₂ was the same as that on Fe₂O₃–TiO₂ (shown in Fig. 8a and 9a), and it was quite different from that on V₂O₅/WO₃–TiO₂ (shown in Fig. 7a and 9a). It suggests that the acid sites (i.e. Brønsted acid sites and Lewis acid sites) on V₂O₅/Fe₂O₃–TiO₂ mainly resulted from the support Fe₂O₃–TiO₂, and the effect of V₂O₅ loading on the acid sites can be approximately neglected. After NO + O₂/N₂ passed over NH₃ pretreated 2% V₂O₅/Fe₂O₃–TiO₂ at 250 °C, both ionic NH₄⁺ and coordinated NH₃ diminished (shown in Fig. 9a). Meanwhile, adsorbed H₂O, which is a product of the SCR reaction, was clearly observed at 1620 cm⁻¹.²² They both suggest that adsorbed ammonia species on 2% V₂O₅/Fe₂O₃–TiO₂ can react with NO.

Fig. 9 (a) in situ DRIFT spectra taken at 250 °C upon passing NO + O₂ over NH₃ presorbed 2% V₂O₅/Fe₂O₃–TiO₂; (b) in situ DRIFT spectra taken at 250 °C upon passing NH₃ over NO + O₂ presorbed 2% V₂O₅/Fe₂O₃–TiO₂.

After 2% V₂O₅/Fe₂O₃–TiO₂ was treated with NO + O₂/N₂ at 250 °C, the bands corresponding to adsorbed nitrite and nitrate cannot be observed (shown in Fig. 9b). It indicates that the active sites on Fe₂O₃–TiO₂ for the adsorption of gaseous NO were covered by the loaded V₂O₅. Hence, the adsorption of gaseous NO on Fe₂O₃–TiO₂ was restrained due to the loading of V₂O₅. After NH₃/N₂ passed over NO + O₂ pretreated 2% V₂O₅/Fe₂O₃–TiO₂, its surface was mainly covered by ionic NH₄⁺ bound to the Brønsted acid sites and coordinated NH₃ bound to the Lewis acid sites. Furthermore, adsorbed H₂O, which is a product of the SCR reaction, can hardly be detected at about 1620 cm⁻¹. It suggests that the reaction between adsorbed nitrogen oxide species and ammonia cannot happen on 2% V₂O₅/Fe₂O₃–TiO₂. Therefore, the SCR reaction over V₂O₅/Fe₂O₃–TiO₂ mainly followed the Eley–Rideal mechanism.

4. Discussion

The SCR reaction over Fe₂O₃–TiO₂ can be approximately described as follows:⁶

NH_3(g) ⇌ NH_3(ad) (7)

NH_3(ad) + ≡Fe_Surf³⁺ → −NH₂ + ≡Fe_Surf²⁺ + H⁺ (8)

−NH₂ + NO_(g)→N₂ + H₂O (9)

≡Fe_Surf²⁺ + ¼O₂ → ≡Fe_Surf³⁺ + ½≡O (10)

Meanwhile, the SCR reaction over V₂O₅/WO₃–TiO₂ can be approximately described as follows:²⁴

NH_3(g) ⇌ NH_3(ad) (7)

NH_3(ad) + ≡V_Surf⁵⁺ → −NH₂ + ≡V_Surf⁴⁺ + H⁺ (11)

−NH₂ + NO_(g) → N₂ + H₂O (9)

≡V_Surf⁴⁺ + ¼O₂ → ≡V_Surf⁵⁺ + ½≡O (12)

Reaction (7) was the adsorption of gaseous ammonia on the acid sites (i.e. Brønsted acid sites and Lewis acid sites) to form adsorbed ammonia species including ionic NH₄⁺ and coordinated NH₃. Reactions (8) and (11) were the activation of adsorbed ammonia species by Fe³⁺ and V⁵⁺ on the surface to form amide species (−NH₂), respectively. Then, gaseous NO was reduced by –NH₂ on the surface to form N₂ and H₂O (reaction (9)). Reactions (10) and (12) were the reoxidization of formed Fe²⁺ and V⁴⁺, respectively. Therefore, the SCR reaction over V₂O₅/Fe₂O₃–TiO₂ can be approximately described using reactions (7–12) and both V⁵⁺ and Fe³⁺ could be the active components for the activation of adsorbed NH₃.

In situ DRIFTS study demonstrated that the acid sites on V₂O₅/Fe₂O₃–TiO₂ mainly resulted from Fe₂O₃–TiO₂, and gaseous NH₃ mainly adsorbed onto Fe₂O₃–TiO₂. Therefore, the adsorbed NH₃ on V₂O₅/Fe₂O₃–TiO₂ preferred to be activated by Fe³⁺ rather than V⁵⁺ on V₂O₅/Fe₂O₃–TiO₂. Moreover, the oxidization of Fe²⁺ by V⁵⁺ is thermodynamically favorable because the redox potential of reaction (13) is 0.23 V. Therefore, the regeneration of Fe³⁺ on Fe₂O₃–TiO₂ could be accelerated due to the rapid electron transfer between V⁵⁺ and Fe²⁺ on V₂O₅/Fe₂O₃–TiO₂.¹⁶

≡Fe_Surf²⁺ + ≡V_Surf⁵⁺ → ≡Fe_Surf³⁺ + ≡V_Surf⁴⁺ E⁰ = 0.23 V (13)

At 150–300 °C, a large amount of NO cannot be reduced over Fe₂O₃–TiO₂ (shown in Fig. 5a). It indicates that adsorbed NH₃ cannot be completely activated by Fe³⁺ on Fe₂O₃–TiO₂ at 150–300 °C. The loading of V₂O₅ accelerated the regeneration of Fe³⁺ on Fe₂O₃–TiO₂ through reaction (13), which promoted the activation of adsorbed NH₃ by Fe³⁺ on Fe₂O₃–TiO₂. Furthermore, NH₃ adsorbed on V₂O₅/Fe₂O₃–TiO₂, which was not activated by Fe³⁺, could then be activated by V⁵⁺ on the surface. As a result, the SCR activity of Fe₂O₃–TiO₂ improved due to the loading of V₂O₅. However, the oxidization ability of Fe³⁺ is much less than that of V⁵⁺, so the activation of NH₃ by Fe³⁺ was much slower than that by V⁵⁺. As a result, the SCR activity of 2% V₂O₅/Fe₂O₃–TiO₂ was much less than that of 2% V₂O₅/WO₃–TiO₂ at 150–300 °C.

Some adsorbed NH₃ could be over-oxidized by V⁵⁺ cations on the surface to form −NH or −N above 300 °C. Then, −NH or −N reacted with gaseous NO to form N₂O.²⁵ Therefore, N₂ selectivity of V₂O₅/WO₃–TiO₂ obviously decreased above 300 °C (shown in Fig. 5b). If a large amount of NH₃ adsorbed on V₂O₅/Fe₂O₃–TiO₂ was activated by V⁵⁺ above 300 °C, some N₂O should form. However, little N₂O formed over V₂O₅/Fe₂O₃–TiO₂ (shown in Fig. 5b). It suggests that NH₃ adsorbed on V₂O₅/Fe₂O₃–TiO₂ was mainly activated by Fe³⁺ above 300 °C.

NO conversion over Fe₂O₃–TiO₂ obviously increased as the reaction temperature increased from 200 to 400 °C (shown in Fig. 5a). It suggests that the activation of adsorbed NH₃ by Fe³⁺ on Fe₂O₃–TiO₂ (reaction (8)) was obviously promoted with the increase in reaction temperature. As shown in Fig. 5a, more than 80% of gaseous NO can be reduced over Fe₂O₃–TiO₂ at 350–450 °C. It suggests that most of the adsorbed NH₃ could be activated by Fe³⁺ on V₂O₅/Fe₂O₃–TiO₂ even if V⁵⁺ did not take part in the activation of adsorbed NH₃. Because gaseous NH₃ mainly adsorbed on Fe₂O₃–TiO₂, the adsorbed NH₃ preferred to be activated by Fe³⁺ rather than by V⁵⁺ on V₂O₅/Fe₂O₃–TiO₂. Meanwhile, the ratios of Fe³⁺ to V⁵⁺ on 1% and 2% V₂O₅/Fe₂O₃–TiO₂ were 15.3 and 3.7, respectively. Therefore, reaction (8) predominated over the activation of adsorbed NH₃ during the SCR reaction over V₂O₅/Fe₂O₃–TiO₂ at 350–500 °C. As a result, V₂O₅/Fe₂O₃–TiO₂ showed an excellent N₂ selectivity at 350–500 °C. Because the regeneration of Fe³⁺ on Fe₂O₃–TiO₂ was accelerated due to the loading of V₂O₅, the SCR activity of V₂O₅/Fe₂O₃–TiO₂ was generally better than that of Fe₂O₃–TiO₂ at 350–450 °C.

5. Conclusion

2% V₂O₅/Fe₂O₃–TiO₂ showed excellent SCR activity, N₂ selectivity and H₂O/SO₂ durability at 300–450 °C. Meanwhile, the emission of 2% V₂O₅/Fe₂O₃–TiO₂ to the fly ash can be prevented by an external magnetic field due to its inherent magnetization. Therefore, 2% V₂O₅/Fe₂O₃–TiO₂ could be a promising low-cost SCR catalyst to control of the emission of NO.

Acknowledgements

This study was financially supported by the National Natural Science Fund of China (Grant Nos. 51078203 and 21207067), the National High-Tech Research and Development (863) Program of China (Grant Nos. 2010AA065002 and 2012AA062506).

Notes and references

1. G. S. Qi, R. T. Yang and R. Chang, Appl. Catal., B, 2004, 51, 93–106.
2. Z. Wu, B. Q. Jiang and Y. Liu, Appl. Catal., B, 2008, 79, 347–355.
3. N. Y. Topsoe, Science, 1994, 265, 1217–1219.
4. L. Chen, J. H. Li and M. F. Ge, Environ. Sci. Technol., 2010, 44, 9590–9596.
5. F. D. Liu, H. He, C. B. Zhang, Z. C. Feng, L. R. Zheng, Y. N. Xie and T. D. Hu, Appl. Catal., B, 2010, 96, 408–420.
6. S. J. Yang, J. H. Li, C. Z. Wang, J. H. Chen, L. Ma, H. Z. Chang, L. Chen, Y. Peng and N. Q. Yan, Appl. Catal., B, 2012, 117, 73–80.
7. L. Chen, J. H. Li and M. F. Ge, J. Phys. Chem. C, 2009, 113, 21177–21184.
8. S. Yang, Y. Guo, N. Yan, D. Wu, H. He, Z. Qu, C. Yang, Q. Zhou and J. Jia, ACS Appl. Mater. Interfaces, 2011, 3, 209–217.
9. S. Yang, H. He, D. Wu, D. Chen, X. Liang, Z. Qin, M. Fan, J. Zhu and P. Yuan, Appl. Catal., B, 2009, 89, 527–535.
10. P. Perriat, E. Fries, N. Millot and B. Domenichini, Solid State Ionics, 1999, 117, 175–184.
11. S. Yang, Y. Guo, N. Yan, Z. Qu, J. Xie, C. Yang and J. Jia, J. Hazard. Mater., 2011, 186, 508–515.
12. S. J. Yang, Y. F. Guo, N. Q. Yan, D. Q. Wu, H. P. He, Z. Qu and J. P. Jia, Ind. Eng. Chem. Res., 2011, 50, 9650–9656.
13. S. Yang, Y. Guo, N. Yan, D. Wu, H. He, J. Xie, Z. Qu and J. Jia, Appl. Catal., B, 2011, 101, 698–708.
14. S. Yang, Y. Guo, N. Yan, D. Wu, H. He, J. Xie, Z. Qu, C. Yang and J. Jia, Chem. Commun., 2010, 46, 8377–8379.
15. N. Guigue-Millot, Y. Champion, M. J. Hytch, F. Bernard, S. Begin-Colin and P. Perriat, J. Phys. Chem. B, 2001, 105, 7125–7132.
16. X. L. Liang, S. Y. Zhu, Y. H. Zhong, J. X. Zhu, P. Yuan, H. P. He and J. Zhang, Appl. Catal., B, 2010, 97, 151–159.
17. M. M. Shafer, B. M. Toner, J. T. Oyerdier, J. J. Schauer, S. C. Fakra, S. H. Hu, J. D. Herner and A. Ayala, Environ. Sci. Technol., 2012, 46, 189–195.
18. S. Yang, C. Wang, J. Chen, Y. Peng, L. Ma, H. Chang, H. Chang, C. Liu, J. Li and N. Yan, Catal. Sci. Technol., 2012, 2, 915–917.
19. S. J. Yang, H. P. He, D. Q. Wu, D. Chen, Y. H. Ma, X. L. Li, J. X. Zhu and P. Yuan, Ind. Eng. Chem. Res., 2009, 48, 9915–9921.
20. F. D. Liu, K. Asakura, H. He, W. P. Shan, X. Y. Shi and C. B. Zhang, Appl. Catal., B, 2010, 103, 369–377.
21. G. S. Qi and R. T. Yang, J. Phys. Chem. B, 2004, 108, 15738–15747.
22. S. Yang, C. Wang, J. Li, N. Yan, L. Ma and H. Chang, Appl. Catal., B, 2011, 110, 71–80.
23. W. S. Kijlstra, D. S. Brands, H. I. Smit, E. K. Poels and A. Bliek, J. Catal., 1997, 171, 219–230.
24. G. Busca, L. Lietti, G. Ramis and F. Berti, Appl. Catal., B, 1998, 18, 1–36.
25. X. F. Tang, J. H. Li, L. A. Sun and J. M. Hao, Appl. Catal., B, 2010, 99, 156–162.

---