Selective catalytic oxidation of ammonia over MnO_x–TiO₂ mixed oxides

Abstract

The selective catalytic oxidation of ammonia to nitrogen (NH₃–SCO) was investigated over MnO_x–TiO₂. The physicochemical properties of MnO_x–TiO₂ were characterized by XRD, O₂-TPD, NH₃-TPD, H₂-TPR and XPS, and the reaction mechanism was studied by in situ DRIFTS. The addition of Mn into TiO₂ accelerated the support phase transformation and the formation of Mn–O–Ti bonds. MnO_x(0.25)–TiO₂ showed the best performance in NH₃–SCO, for which complete conversion of NH₃ was obtained at 200 °C with the temperature window (180–300 °C) for N₂ yield > 80%. The formation of Mn–O–Ti provided abundant oxygen vacancies which promoted the adsorption and dissociation of molecular oxygen to form active oxygen species. The finely dispersed MnO_x species on the support favored NH₃ adsorption. N₂O was produced over the whole temperature range while NO_x was produced only at high temperatures (>250 °C). N₂O was formed from the combination of two HNO species at low temperatures, whereas it was formed from NH₃ and nitrate/nitrite species reaction at high temperatures.

---

1. Introduction

With the increasing use of automobiles, NO_x as one of the exhaust gases of vehicles leads to a worsening atmosphere in the case of photochemical smog and acid rain. The efficient removal of NO_x has become urgent for environmental protection. Selective catalytic reduction (SCR) of NO_x with NH₃ is a mature technology for reducing NO_x emission. In a typical SCR, the main reaction proceeds in the following way:

4NH₃ + 4NO + O₂ → 4N₂ + 6H₂O

In order to efficiently remove NO, the application of over stoichiometric NH₃ results in NH₃ slip at low temperature. The concentration of NH₃ slip of SCR units has been regulated to less than 10 ppm in China. Catalytic oxidation, with a low level of energy consumption and high purification efficiency, is a promising technology for NH₃ emission control. Due to the variety of products in NH₃ oxidation, such as N₂, NO_x (NO, NO₂), N₂O, the selective catalytic oxidation (SCO) of NH₃ (NH₃–SCO) to N₂ is a potential technology to control NH₃ slip without secondary pollutants.

Supported noble metals and transition metal oxides work well in NH₃–SCO. Higher NH₃ conversion at less than 180 °C could be obtained on Ag/Al₂O₃,¹ CuO/RuO₂,² Pt/CuO/Al₂O₃,³ and Ir,⁴ but the N₂ selectivity needs to be further improved. Compared with supported noble metal catalysts, transition metal oxide catalysts, CuO,⁵ Fe₂O₃,⁶ NiO,⁷ V₂O₅,⁸ exhibit higher N₂ selectivity at rather high temperatures (300–400 °C). Exploring the activity of metal oxides at low temperatures makes possible their practical application in the future.

The presence of Mn⁴⁺/Mn³⁺ type redox couples,⁹ the tunnel structure and active lattice oxygen species¹⁰ contribute to the excellent performance of manganese oxides in oxidation. Due to excellent SO₂ tolerance¹¹ and oxygen sensitivity,¹² TiO₂ is usually used as the support. So MnO_x/TiO₂ catalysts have been studied for low-temperature selective catalytic reduction of NH₃ (NH₃–SCR).¹³

Several mechanisms of the SCO process have been suggested. Three major reaction pathways have been proposed. The imido mechanism suggests that the intermediate species are the imido (NH) and nitrosyl (HNO) species,¹⁴ and N₂ comes from their reaction. The hydrazine (N₂H₄) mechanism proposes that N₂H₄ arises from two amide species (NH₂), and then N₂ from N₂H₄ oxidation.¹⁵ The key step of the above mechanisms is NH₃ dehydrogenation. The internal selective catalytic reduction (i-SCR) mechanism focuses on the reaction between NH₃ and NO_x species.¹⁶ The mechanism varies with the properties of the catalysts.

In this paper, a series of MnO_x–TiO₂ catalysts were prepared by the sol–gel method and their catalytic performance in NH₃ selective oxidation was investigated. XRD, Raman, N₂ adsorption–desorption, ICP, O₂-TPD, H₂-TPR and XPS were used to characterize the physicochemical properties of the samples, and in situ DRIFT was applied to understand the oxidation reaction process. The correlation between the activity and physicochemical properties was investigated.

2. Experimental

2.1 Catalyst preparation

The catalyst was obtained by the sol–gel method. All the chemicals used were from Shanghai Lingfeng Chemical Reagent Co. Butyl titanate (0.02 mol) and diethanolamine (0.02 mol) solution were added into 30 ml ethanol denoted as solution A. Mn(OAc)₂·4H₂O and acetic acid were dissolved in distilled water to get solution B. Then solution B was added dropwise into solution A under vigorous stirring to form a brownish red homogeneous sol. After aging from sol to gel, the gel was dried at 353 K for 12 h and calcined at 773 K for 4 h in air. The obtained sample was denoted as MnO_x(z)–TiO₂, where z represents the weight ratio of Mn/TiO₂.

2.2 Activity test

Catalytic activity measurements were carried out in a tubular quartz fixed-bed reactor (4 mm internal diameter). 500 ppm ammonia, 5 vol% oxygen and helium at flow rates of 200 ml min⁻¹ were passed through the catalyst bed (50 mg, 40–60 mesh). The inlet and outlet gases were analyzed by a Nicolet FTIR 6700 spectrometer (Thermo Fisher Scientific, USA) and a 42i-HL NO-NOx chemiluminescence analyzer (Thermo Fisher Scientific, USA). The NH₃ conversion, and N₂, NO_x, and N₂O yields were calculated from the concentrations of the gases according to eqn (1)–(4):

2.3 Catalyst characterization

The N₂ adsorption–desorption isotherms were measured with an ASAP2020 (Micromeritics, USA) at −196 °C. All samples were degassed at 180 °C before testing.

Powder X-ray diffraction (XRD) measurements of the catalysts were carried out using a Bruker/D8 X-ray diffractometer (Bruker, Germany) with Cu Kα radiation (λ = 0.154056 nm) at 40 kV and 40 mA. The patterns were taken over the 2θ range from 10° to 80° at a scan speed of 6° min⁻¹.

X-ray photoelectron spectroscopy (XPS) analysis was performed using an ESCALAB 250Xi (Thermo Fisher Scientific, USA) with a monochromatic Al Kα source (hν = 1486.6 eV), operating at 150 W. The binding energy of adventitious C 1s (284.6 eV) was used as a reference.

Hydrogen temperature-programmed reduction (H₂-TPR) experiments were carried out using a commercial system (PX200, PengXiang, China) equipped with a thermal conductivity detector. Ammonia temperature-programmed desorption (NH₃-TPD) and oxygen temperature-programmed desorption (O₂-TPD) were performed with the same apparatus. In H₂-TPR, the samples (50 mg) were heated in 5 vol% H₂/N₂ (40 ml min⁻¹) from room temperature to 850 °C at a ramp of 10 °C min⁻¹. In NH₃-TPD, the samples (50 mg) were preheated in N₂ at 500 °C for 1 h before testing. After 5% NH₃ adsorption saturation at 30 °C, N₂ was introduced to remove physically adsorbed NH₃. In O₂-TPD, the outlet gas was recorded with an online mass spectrometer apparatus (HIDEN QIC-20, England). The samples (100 mg) were pretreated in 3% O₂ in He stream (30 ml min⁻¹) at 450 °C for 1 h, then cooled down to ambient temperature in He to get a stable MS baseline. In all the experiments the heating rate (10 °C min) was the same from 50 to 800 °C.

Raman spectra were obtained using a Renishaw inVia Reflex (Renishaw, England) spectrometer equipped with a CCD detector at ambient temperature and under moisture-free conditions. The excitation source was the 514.5 nm line of an Ar ion laser. The laser power was set at 3 mW.

The in situ DRIFTS was performed using a Nicolet 6700 FT-IR spectrometer (Thermo Fisher Scientific, USA) with an MCT detector. In the DRIFT cell with ZnSe windows connected to a gas flow system, the sample was pretreated at 400 °C in Ar for 1 h before each test. The background spectra were recorded at each temperature. 500 ppm NH₃, 5% O₂ and Ar balance were used in the experiments at a total flow rate of 50 ml min⁻¹.

3. Results

3.1 NH₃ oxidation over MnO_x–TiO₂ catalysts

The catalytic performance of the SCO of NH₃ on MnO_x–TiO₂ catalysts at various temperatures is shown in Fig. 1. Temperature had positive effects on NH₃ conversion: the higher the temperature, the higher the NH₃ conversion. The complete NH₃ conversion temperature was almost 400 °C, and 80% N₂ yield was obtained in a sharp temperature window from 290 to 340 °C on TiO₂. With the introduction of Mn, the low-temperature activity was enhanced and the operation window of high N₂ yield broadened accordingly. The optimum Mn loading amount was 25 wt%, NH₃ could be completely converted at 200 °C and the temperature range of N₂ yield above 80% was 180–300 °C at a weight space velocity of 240 000 ml (min⁻¹ g⁻¹). At 150 °C, NH₃ conversion on MnO_x(0.25)–TiO₂ reached 78%, while on TiO₂ it only reached 10%. The increase in NH₃ conversion evidenced the promotion effects of Mn species.

Fig. 1 The results of activity tests over TiO₂ and MnO_x(z)–TiO₂ catalysts.

In order to exclude the effect of surface area on activity, the specific activities were calculated and are shown in Table 1. With the Mn loading increasing from 0.1 to 0.25, the specific activity increased from 0.7 to 3.6 (mmol (min⁻¹ m⁻²) × 10⁻⁶) at 130 °C, while further increasing Mn loading had less effect on specific activity. Therefore, there was an optimum Mn amount (25%) for selective NH₃ oxidation.

Table 1 The textural data of samples

Catalysts	SBET (m^2 g^−1)	Pore size (nm)	Crystallite size^a (nm)	Mn content^b (wt%)	Specific activity (mmol (min^−1 m^−2)) × 10^−6
					130 °C
a Calculated from the Scherer equation according to the (110) diffraction peak of TiO2.b Mn content of the catalyst determined by ICP-AES.
Pure TiO2	45.1	8.7	21.0	—	0.5
MnOx(0.1)–TiO2	74.2	5.0	8.8	8	0.7
MnOx(0.2)–TiO2	113.2	4.2	10.4	18	0.8
MnOx(0.25)–TiO2	46.6	6.2	11.7	27.8	3.6
MnOx(0.3)–TiO2	41.7	4.8	11.1	33.7	3.1

---

The N-containing product distribution (N₂, NO_x and N₂O) is also shown in Fig. 1. N₂O was produced even at low temperatures (>100 °C) with a yield lower than 10%; while the appearance of NO_x occurred at high temperatures (>250 °C) with a yield of up to 20% at 350 °C. The yield of N₂O decreased with the increase of NO_x yield.

3.2 Catalyst characterization

3.2.1 BET surface area and XRD. The textural data of the samples are shown in Table 1. All the samples were of a meso-pore structure because the pore size was within 4–9 nm. The surface area of the support was only 45 m² g⁻¹, and the surface area of MnO_x–TiO₂ increased up to an Mn loading of 25 wt%. The highest value of surface area (113 m² g⁻¹) was obtained on MnO_x(0.2)–TiO₂ with the addition of Mn efficiently suppressing the growth of TiO₂ crystallites, which resulted in an increase of surface area. Further increasing the loading of Mn led to a decrease of surface area because of the aggregation of Mn species on the support.

The XRD patterns of MnO_x–TiO₂ catalysts are presented in Fig. 2. This suggests that the support was in the form of anatase. With the introduction of Mn, the diffraction peaks of anatase weakened and the characteristic peaks of rutile gradually intensified. The MnO_x(0.2)–TiO₂ material exhibited complete rutile phase which indicated that Mn loading speeded up the anatase-to-rutile (A–R) transition. The doping of Mn caused the diffraction peak (2θ = 36.2°) to shift to higher angles compared with TiO₂, which was attributed to lattice shrinkage due to the lower ionic radius of Mn⁴⁺ (0.060 nm) with respect to Ti⁴⁺ (0.068 nm).

Fig. 2 XRD patterns of MnO_x(z)–TiO₂ catalysts.

New characteristic peaks assigned to Mn₂O₃ and Mn₃O₄ were detected in the pattern of MnO_x(0.3)–TiO₂. The appearance of MnO_x was due to the separation of Mn from TiO₂.

3.2.2 Raman spectroscopy. Raman spectroscopy has become an increasingly valuable tool for the investigation of structural variance. The spectra of MnO₂, Mn₂O₃, Mn₃O₄, TiO₂ and MnO_x(z)–TiO₂ are shown for comparison in Fig. 3. Peaks for MnO₂ were observed at 760, 664, 578 and 535 cm⁻¹ as well as Mn₂O₃ at 700, 646, 308 and 197 cm⁻¹, and Mn₃O₄ at 640, 360 and 300 cm⁻¹. The support was characterized by anatase (bands at 143, 193, 396, 512 and 638 cm⁻¹) and rutile (band at 446 cm⁻¹),¹⁷ which corresponded to the phase transition of the support in XRD results.

Fig. 3 Raman spectra of MnO_x, TiO₂ and MnO_x(z)–TiO₂.

In the typical sample of MnTiO₃, the characteristic peaks could be observed at 684, 611, 465, 360, 336, 264, 236, and 202 cm⁻¹ separately, and the intense Raman band at 684 cm⁻¹ is regarded as a typical feature for MnTiO₃.¹⁸ The strongest Raman band around 684 cm⁻¹ is assigned to the highest frequency vibrational mode of octahedral MnO₆, while the weak phonon modes below 300 cm⁻¹ arise from lattice vibrations.¹⁹

With the addition of Mn, the characteristic peak of MnTiO₃ at 684 cm⁻¹ appeared. The peak intensified with increasing Mn loading amount up to 25 wt%; and a further increase in Mn loading, such as 30 wt%, led to the peak at 684 cm⁻¹ being broadened which was due to overlapping with MnO_x. Eight Raman active fundamentals, at 684, 610, 460, 360, 335, 262, 234, and 202 cm⁻¹, were clearly observed for MnO_x(0.25)–TiO₂, which indicated the formation of MnTiO₃. When taking the sensitivity of the Raman technique to the vibration of chemical bonds into account, it can be understood why MnTiO₃ species could not be detected by XRD. It was seen from the Raman results that the formation of Mn–O–Ti supported the strongest interaction between TiO₂ and Mn on MnO_x(0.25)–TiO₂.

3.2.3 H₂-TPR. H₂-TPR was used to investigate the oxidation states of the samples. The H₂-TPR profiles of the catalysts and the H₂ consumption amounts are presented in Fig. 4 and Table 2.

Fig. 4 H₂-TPR profiles of MnO_x(z)–TiO₂ catalysts.

Table 2 The data obtained from H₂-TPR, O₂-TPD and XPS

Catalysts	H2 consumption (mmol g^−1)						(Oα + Oβ)^a	Mn^3+/Mntotal	OS/(OL + OS)	Surface atom ratio^b of Mn/Ti
	Total	a	b	c	d	e
a Obtained from O2-TPD.b Obtained from XPS.
TiO2	0.04	—	—	—	0.04	—	—	—	0.16	—
MnOx(0.1)–TiO2	0.75	0.1	0.33	0.27	—	—	1	0.58	0.19	0.3
MnOx(0.2)–TiO2	2.11	0.3	0.99	0.75	—	0.05	2.9	0.59	0.25	0.7
MnOx(0.25)–TiO2	3.13	0.43	1.42	1.0	—	0.27	5.9	0.61	0.3	1.2
MnOx(0.3)–TiO2	3.20	0.34	1.38	0.98	—	0.5	2.6	0.56	0.27	1.7

---

A weak peak was observed for the reduction of TiO₂ in the temperature range of 500–800 °C with 0.04 mmol g⁻¹ total H₂ consumption. With the presence of Mn, a peak with a shoulder at 100–450 °C was observed and intensified with an increase of Mn concentration. Combined with the ICP and the integral H₂ consumption, it was known that the oxidation state of Mn was around +3 which indicated that MnO_x was mainly in the form of Mn₂O₃. By de-convoluting the main reduction peak, three peaks were observed which represented the reduction of different species. Ettireddy²⁰ assigned the peak at 215 °C to the reduction of anatase TiO₂ closely interacted with MnO_x, and the peak area correlated with the anatase TiO₂ concentration. Although XRD results presented the A–R transformation of the support with the increase of Mn, the H₂ consumption amount of peak a still kept increasing as shown in Table 2. So it excluded the possibility that peak a was from the reduction of the support, and it should be assigned to the reduction of Mn₂O₃ close to TiO₂. The reduction of bulk Mn₂O₃ took place through a distinct two-step process: the first step involved the reduction of Mn₂O₃ to Mn₃O₄, and the second step was the reduction of Mn₃O₄ to MnO.²¹ The H₂ consumption ratio confirmed the dominant Mn₂O₃ species, so peaks b and c were assigned to the reduction of bulk Mn₂O₃. The shoulder peak e at 400–550 °C was assigned to the reduction of Mn₃O₄, and XRD also confirmed the existence of Mn₃O₄ with the introduction of Mn loading up to 30 wt%.

3.2.4 O₂-TPD. The O₂-TPD technique is useful to study oxygen species of metal oxides, characterizing the type of oxygen species and their properties. The O₂-TPD results are shown in Fig. 5. Three desorption peaks were observed in the temperature ranges of 200–300 °C, 400–650 °C and 650–800 °C respectively. The weak peak (O_α) at 200–300 °C is related to desorption of superoxide ion O₂⁻ weakly bound to the surface of Mn₂O₃. The broad peak (O_β) at 400–600 °C with shoulder is ascribed to desorption of peroxide ion O₂²⁻/O⁻ bound to oxygen vacancies. The high-temperature peak at 600–800 °C is assigned to the desorption of lattice oxygen ion O²⁻ in Mn₂O₃.^22,23

Fig. 5 O₂-TPD profiles of MnO_x(z)–TiO₂ catalysts.

Compared with the support, the intensity of the peaks (O_α, O_β) increased with the amount of Mn up to 25%, and the decrease in intensity was also observed on MnO_x(0.3)–TiO₂. The increased area of O_β implied a strong interaction between TiO₂ and MnO_x which led to the enhancement in lattice defects and oxygen vacancies and promoted the adsorption/mobility/desorption of O₂²⁻/O⁻. The formation of Ti–O–Mn or MnTiO₃ detected by Raman spectroscopy also confirmed the strong interaction between TiO₂ and MnO_x. Meanwhile, the peak at 200–300 °C attributed to the desorption of O₂⁻ becomes a little stronger, which was consistent with the XPS results that more surface oxygen species existed on the surface of MnO_x(0.25)–TiO₂.

The temperatures of the O²⁻ desorption peaks increased as Mn substitution amounts increased. This demonstrated that the interactions of Mn–O bonds were strengthened so the desorption of O²⁻ became difficult, which was in agreement with the result of H₂-TPR.

3.2.5 NH₃-TPD. The NH₃-TPD profiles of the catalysts are shown in Fig. 6A. The desorbed temperature and desorbed ammonia amount directly relate to the kind and amount of acid sites on the samples. Lower temperature desorption meant weaker interaction between NH₃ and acid sites and vice versa. An NH₃ desorption peak was found at 100–500 °C which was related to NH₃ desorption at weak acid sites.

Fig. 6 NH₃-TPD profiles (A) and acid density (B) of MnO_x(z)–TiO₂ catalysts.

Compared with TiO₂, the NH₃ desorption temperature at weak acid sites shifted to lower temperatures. The presence of Mn promoted the amount of acid sites which suggested that Mn was the source of acid. The maximum acid sites seemed to be obtained on MnO_x(0.2)–TiO₂, while further increasing Mn led to the decrease of the amount of acid sites which was attributed to the aggregation of manganese oxide.

By excluding the effect of surface area, the acid density is calculated and the results are shown in Fig. 6B. Unlike the results in Fig. 6A, the presence of Mn had no obvious effects on acid density until Mn loading was up to 0.25. A further increase of Mn loading up to 0.3 led to the decrease of acid density on the catalyst. The highest acid density was obtained on MnO_x(0.25)–TiO₂ which favored NH₃ adsorption and activation.

3.2.6 X-ray photoelectron spectroscopy. The MnO_x–TiO₂ catalysts were investigated by XPS to understand the surface atomic concentration and oxidation states of Mn in the catalysts. The XPS spectra of O 1s, Mn 2p and Ti 2p_3/2 are shown in Fig. 7A–(C), and the molar ratios are summarized in Table 2.

Fig. 7 XPS spectra of (A) O 1s, (B) Ti 2p and (C) Mn 2p.

As shown in Fig. 7A, two peaks can be distinguished in the XPS O 1s spectra. The peak at 529.5 eV was assigned to the lattice oxygen O²⁻ (O_L), and the peak at 531.5 eV was assigned to the surface oxygen (O_S), such as O₂⁻, O₂²⁻ and O⁻, belonging to defect oxide and hydroxyl-like groups.²⁴ The O_S/(O_L + O_S) ratio of TiO₂ was only 16%; with the introduction of Mn, the O_S/(O_L + O_S) ratios of the catalysts increased to 30%. This indicated that Mn promoted the amount of surface oxygen on the catalysts. The O_S/(O_L + O_S) ratio on MnO_x(0.25)–TiO₂ reached 30% which was 2 times higher than that on TiO₂.

Fig. 7C shows the XPS spectra of Mn 2p in the MnO_x–TiO₂ samples. The Mn 2p peaks could be de-convoluted to three peaks, the peaks at 643.0–644.0, 641.0–641.9 and 640.2–640.7 eV corresponding to Mn⁴⁺, Mn³⁺ and Mn²⁺, respectively.²⁵ It was interesting to note that Mn³⁺ was the dominant species in the samples with small amounts of Mn⁴⁺ and Mn²⁺ species. The ratio of Mn³⁺/Mn_total increased with the addition of Mn and reached a maximum on MnO_x(0.25)–TiO₂. The Mn⁴⁺ species were in the form of Mn–O–Ti, confirmed by Raman spectroscopy; while the Mn²⁺ species originated from Mn₃O₄ detected by XRD. With the increase of Mn loading, the surface atomic ratio of Mn/Ti increased accordingly which suggested the enrichment of Mn species on the catalyst surface. Mn 2p_3/2 binding energy slightly shifted to high binding energy (641.2 eV to 641.7 eV) because the Mn species showed deviation of electron cloud by interacting with titanium species, leading to an enhancement in oxidative ability.²⁶

According to the law of conservation of charge, a binding energy increase of one element must be accompanied by a binding energy decrease of another element in an isolated system. The shift of Ti 2p_3/2 (458.3 eV to 457.8 eV) and O 1s (529.5 eV to 529.2 eV) also confirmed the existence of interaction between Mn and Ti.

3.3 Studies of reaction mechanism

3.3.1 DRIFT of NH₃ adsorption. NH₃ adsorption measurements were performed in a feed of 500 ppm NH₃ in Ar at various temperatures, and the spectra are shown in Fig. 8. The bands at 1178 and 1603 cm⁻¹ are ascribed to the coordinated NH₃ on Lewis acid sites²⁷ while the bands at 1440 and 1694 cm⁻¹ are ascribed to NH⁴⁺ on Bronsted acid sites.²⁸ In the high-frequency region, bands at 3134, 3227 and 3377 cm⁻¹ are attributed to the NH stretching coordinated to NH₃ from Lewis acid sites.²⁹ The band at 1630 cm⁻¹ is ascribed to the O–H stretching vibration modes due to the interaction of surface hydroxyls with NH₃ (ref. 30) and the one at 3600 cm⁻¹ is due to the hydroxyl consumption through interaction with NH₃ to form NH₄⁺.³¹

Fig. 8 In situ DRIFT spectra of catalyst treated with NH₃ at various temperatures.

The intensity of Lewis acid site bands became strong when the temperature increased from 50 °C to 140 °C, while further elevating the temperature up to 300 °C caused the disappearance of the bands. By comparing the band intensities of Lewis acid sites and Bronsted acid sites, it was found that ammonia mainly existed on Lewis acid sites and kept stable even at 250 °C.

When the temperature increased from 100 °C to 140 °C, two new weak bands at 1456 and 1510 cm⁻¹ were observed which are ascribed to –NH deformation modes of imido (=N–H) species³² and amide (–NH₂),³³ respectively. The appearance of these intermediates indicated adsorbed NH₃ could be activated through the abstraction of hydrogen.

Meanwhile, two weak bands at 2213 and 2240 cm⁻¹ were observed which can be assigned to N–N stretching modes of adsorbed N₂O from the oxidation of ammonia.³⁴ The appearance of N₂O in the absence of O₂ demonstrated the oxidation of NH₃ by lattice oxygen.

3.3.2 The interaction of NH₃ with O₂. The sample behavior in NH₃ and O₂ was evaluated and the spectra are shown in Fig. 9. By comparing the intensity of NH₃ adsorbed on Lewis acid sites and Bronsted acid sites with or without O₂, the weaker intensities of these bands were ascribed to the reaction between oxygen and adsorbed NH₃.

Fig. 9 In situ DRIFT spectra of catalyst treated with NH₃ + O₂ at various temperatures.

The presence of NH₃ and O₂ also led to the appearance of bridge nitrite (1234 cm⁻¹) and monodentate nitrate (1277 cm⁻¹)³⁵ at 100 °C and bidentate nitrate^36,37 (1540 cm⁻¹) at even higher than 250 °C. A weak band at 1832 cm⁻¹ assigned to nitrosyl (–HNO)³⁸ species was observed which was from the interaction between –NH and oxygen atoms. Combined with the appearance of nitrite and nitrate species, it is proposed that nitrosyl (–HNO) could be further oxidized to nitrite and nitrate.

Gas-phase N₂O (2213, 2240 cm⁻¹) was found in the whole temperature range, while the band at 1612 cm⁻¹ ascribed to the weakly adsorbed gas NO₂ (ref. 39) was only observed at 250 °C. The activity test results also confirmed that the product distribution of N-containing species depended on the temperature. There are many researches on the form of N-containing species in exhaust. It is usually considered that N₂ arises from the reaction between NH and HNO.⁴⁰ There are two main ways to get N₂O, one is from the interaction between two HNO species,⁴¹ and the other way is by the reaction between nitrate species and ammonia.⁴² NO_x is usually considered as the decomposition product of adsorbed nitrate/nitrite species.⁴³

The interaction of O₂ with the in situ-formed NH₃ species at 230 °C was investigated. The experiment was conducted as follows: prior to the introduction of O₂, the catalyst was exposed to 500 ppm NH₃ at 230 °C for 1 h, followed by purging with Ar for 30 min. The in situ DRIFT spectra of the catalyst were recorded as a function of time, and the results are shown in Fig. 10.

Fig. 10 In situ DRIFT spectra of catalyst exposed to O₂ after NH₃ adsorption at 230 °C.

The intensities of Lewis acid site and Bronsted acid site bands decreased even after introducing O₂ for 1 min; and this became obvious after 5 min. With further exposure for 10 min, the nitrate species (1234, 1276, 1544 cm⁻¹) appeared. The changes in the first 5 min were ascribed to the reaction between adsorbed NH₃ and O₂, and the appearance of nitrate species showed that adsorbed NH₃ could be further oxidized to nitrate species. Meanwhile, the bands (2211, 2240 cm⁻¹) assigned to N₂O were observed after introducing O₂ for 1 min which then decreased with further exposure time. With the appearance of nitrate species, the weakly adsorbed gas NO₂ species (1612 cm⁻¹) were observed.

3.3.3 The interaction of NH₃ and O₂ with formed nitrate species. The interaction of NH₃ with in situ-formed nitrate species over the catalysts was investigated to understand the reaction mechanism. The experiment was conducted as follows: prior to the introduction of NH₃ and O₂, the catalyst was exposed to 500 ppm NO at 160 °C or 250 °C for 1 h, followed by purging with Ar for 30 min. The in situ DRIFT spectra of the catalyst were recorded as a function of time, and the results are shown in Fig. 11.

Fig. 11 In situ DRIFT spectra of catalyst exposed to NH₃ + O₂ after NO adsorption at 160 °C or 250 °C.

Bridge nitrite (1234 cm⁻¹), monodentate nitrate (1270 cm⁻¹), and bidentate nitrate (1547 cm⁻¹) were observed after NO pretreatment, which indicated that NO could be oxidized to nitrite and nitrate by lattice oxygen at 160 and 250 °C.

After exposure to NH₃ and O₂ for 10 min at 160 °C, the bands of NH₃ adsorbed on Lewis acid sites (1178, 1602 cm⁻¹) appeared and the intensity increased with the exposure time. Meanwhile, the intensity of the bands of nitrite and nitrate species kept stable in the presence of NH₃ and O₂.

Unlike the experiment conducted at 160 °C, an obvious disappearance of nitrite and nitrate species was observed after adding NH₃ and O₂ for 20 min at 250 °C, and the appearance of N₂O (2211, 2240 cm⁻¹) after 15 min was observed. The findings suggested that N₂O was the reaction product of NH₃ and nitrite/nitrate species.

Based on the above results, NH₃ was mainly adsorbed on Lewis acid sites (Fig. 8) and activated through hydrogen abstraction. The intermediates (–NH₂, –NH) were further oxidized to HNO species, and two HNO species can convert to N₂O at low temperatures (Fig. 9). Meanwhile adsorbed NH₃ can be directly oxidized to nitrite/nitrate species (Fig. 10).

The performance of nitrite and nitrate species depended on the temperature (Fig. 11). With the nitrite and nitrate species disappearing, N₂O was observed at 250 °C which suggested that N₂O was formed from the reaction between NH₃ and nitrite/nitrate species. Combined with the in situ DRIFT results, the reaction mechanism of NH₃ selective oxidation is shown in Fig. 12.

Fig. 12 The reaction mechanism of NH₃ selective oxidation on MnO_x–TiO₂.

4. Conclusions

The introduction of Mn can effectively promote the selective oxidation of NH₃ to N₂. The finely dispersed MnO_x on the support can increase the surface acid density, especially weak acid density, which benefited NH₃ adsorption. Meanwhile, the formation of Mn–O–Ti provided abundant oxygen vacancies that promoted the adsorption and dissociation of oxygen species. MnO_x(0.25)–TiO₂ showed the best performance in NH₃–SCO. The temperature window for N₂ yield higher than 80% was in the range of 180–300 °C. N₂O was produced in the whole temperature window while the production of NO_x was mainly at higher temperatures (>250 °C).

NH₃ adsorbed on the Lewis acid sites can convert to NH and NH₂ species by dehydrogenization and HNO species by further oxidation; meanwhile it also can be oxidized to nitrite and nitrate species deposited on the surface. N₂ was produced from the reaction between HNO and NH species. N₂O was formed from the combination of two HNO species at low temperatures, while from the reaction between adsorbed NH₃ and nitrite/nitrate species at high temperatures.

Acknowledgements

This work was financially supported by the National Basic Research Program of China (2013CB933200), National High Technology Research and Development Program of China (2015AA034603), National Natural Science Foundation of China (21333003, 21577034), and Fundamental Research Funds for the Central Universities (WJ1514020).

Notes and references

1. L. Zhang, C. Zhang and H. He, J. Catal., 2009, 261, 101–109.
2. X. Cui, J. Zhou, Z. Ye, H. Chen, L. Li, M. Ruan and J. Shi, J. Catal., 2010, 270, 310–317.
3. G. Olofsson, L. Reinewallenberg and A. Andersson, J. Catal., 2005, 230, 1–13.
4. S. A. C. Carabineiro and B. E. Nieuwenhuys, Surf. Sci., 2002, 505, 163–170.
5. L. Gang, J. van Grondelle, B. G. Anderson and R. A. van Santen, J. Catal., 1999, 186, 100–109.
6. R. Q. Long and R. T. Yang, J. Catal., 2002, 207, 158–165.
7. M. Amblard, R. Burch and B. W. L. Southward, Catal. Today, 2000, 59, 365–371.
8. S. M. Lee and S. C. Hong, Appl. Catal., B, 2015, 163, 30–39.
9. A. R. Gandhe, J. S. Rebello, J. L. Figueiredo and J. B. Fernandes, Appl. Catal., B, 2007, 72, 129–135.
10. J. Zhang, Y. Li, L. Wang, C. Zhang and H. He, Catal. Sci. Technol., 2015, 5, 2305–2313.
11. I. H. Fumito Nakajima, Catal. Today, 1996, 29, 109–115.
12. B. M. Reddy, K. J. Ratnam, P. Saikia and G. Thrimurthulu, J. Mol. Catal. A: Chem., 2007, 276, 197–204.
13. J. Xie, D. Fang, F. He, J. Chen, Z. Fu and X. Chen, Catal. Commun., 2012, 28, 77–81.
14. J. Zawadzki, Discuss. Faraday Soc., 1950, 8, 140.
15. L. I. Darvell, K. Heiskanen, J. M. Jones, A. B. Ross, P. Simell and A. Williams, Catal. Today, 2003, 81, 681–692.
16. L. Chmielarz, A. Węgrzyn, M. Wojciechowska, S. Witkowski and M. Michalik, Catal. Lett., 2011, 141, 1345–1354.
17. Y. Djaoued, S. Badilescu, P. V. Ashrit, D. Bersani, P. P. Lottici and J. Robichaud, J. Sol-Gel Sci. Technol., 2002, 24, 255–264.
18. L. Kernazhitsky, V. Shymanovska, T. Gavrilko, G. Puchkovska, V. Naumov, T. Khalyavka, V. Kshnyakin, V. Chernyak and J. Baran, J. Mater. Sci. Eng. B, 2010, 175, 48–55.
19. Y. Brik, M. Kacimi, M. Ziyad and F. Bozon-Verduraz, J. Catal., 2001, 202, 118–128.
20. P. R. Ettireddy, N. Ettireddy, S. Mamedov, P. Boolchand and P. G. Smirniotis, Appl. Catal., B, 2007, 76, 123–134.
21. E. R. Stobbe, B. A. de Boer and J. W. Geus, Catal. Today, 1999, 47, 161–167.
22. H. Liang, Y. Hong, C. Zhu, S. Li, Y. Chen, Z. Liu and D. Ye, Catal. Today, 2013, 201, 98–102.
23. Y. Wei, J. Liu, Z. Zhao, A. Duan, G. Jiang, C. Xu, J. Gao, H. He and X. Wang, Energy Environ. Sci., 2011, 4, 2959.
24. D. Meng, W. Zhan, Y. Guo, Y. Guo, L. Wang and G. Lu, ACS Catal., 2015, 5, 5973–5983.
25. D. A. Peña, B. S. Uphade and P. G. Smirniotis, J. Catal., 2004, 221, 421–431.
26. F. Liu, H. He, C. Zhang, Z. Feng, L. Zheng, Y. Xie and T. Hu, Appl. Catal., B, 2010, 96, 408–420.
27. Z. B. Wu, B. Q. Jiang, Y. L. Jiang, H. Q. Wang and R. B. Jin, Environ. Sci. Technol., 2007, 41, 5812–5817.
28. S. D. Lin, A. C. Gluhoi and B. E. Nieuwenhuys, Catal. Today, 2004, 90, 3–14.
29. W. Chen, Z. Qu, W. Huang, X. Hu and N. Yan, Fuel, 2016, 166, 179–187.
30. G. Y. Zhou, B. C. Zhong, W. H. Wang, X. J. Guan, B. C. Huang, D. Y. Qi and H. J. Wu, Catal. Today, 2011, 175, 157–163.
31. W. Chen, Y. Ma, Z. Qu, Q. Liu, W. Huang, X. Hu and N. Yan, Environ. Sci. Technol., 2014, 48, 12199–12205.
32. L. Zhang and H. He, J. Catal., 2009, 268, 18–25.
33. G. Qi and R. T. Yang, J. Phys. Chem. B, 2004, 108, 15738–15747.
34. J. G. Amores, V. S. Escribano, G. Ramis and G. Busca, Appl. Catal., B, 1997, 13, 45–58.
35. Q. Li, H. Gu, P. Li, Y. Zhou, Y. Liu, Z. Qi, Y. Xin and Z. Zhang, Chin. J. Catal., 2014, 35, 1289–1298.
36. Z. X. Ma, H. S. Yang, Q. Li, J. W. Zheng and X. B. Zhang, Appl. Catal., A, 2012, 427–428, 43–48.
37. B. Li, Z. Y. Ren, Z. X. Ma, X. D. Huang and H. S. Yang, Catal. Sci. Technol., 2016, 6, 1719.
38. M. A. Debeila, N. J. Coville, M. S. Scurrell and G. R. Hearne, Appl. Catal., A, 2005, 291, 98–115.
39. T. Chen, B. Guan, H. Lin and L. Zhu, Chin. J. Catal., 2014, 35, 294–301.
40. J. Barbier Jr, L. Oliviero, B. Renard and D. Duprez, Catal. Today, 2002, 75, 29–34.
41. G. Ramis, M. A. Larrubia and G. Busca, Top. Catal., 2000, 11/12, 161–166.
42. L. Qiu, D. Pang, C. Zhang, J. Meng, R. Zhu and F. Ouyang, Appl. Surf. Sci., 2015, 357, 189–196.
43. F. Can, X. Courtois, S. Royer and G. Blanchard, Catal. Today, 2012, 197, 144–154.

---