The effects of manganese precursors on Mn-based/TiO₂ catalysts for catalytic reduction of NO with NH₃

Abstract

Manganese acetate (MnAc) and manganese nitrate (MnN) were employed as precursors for the preparation of Mn(Ac)/TiO₂, Mn(N)/TiO₂, Mn(Ac)-Ce/TiO₂, and Mn(N)-Ce/TiO₂ by impregnation. These complexes were used as catalysts in the low-temperature selective catalytic reduction of NO with NH₃. The influence of manganese precursors on catalyst characteristics, the reduction activity, and the stability of the catalysts to poisoning by H₂O and SO₂ were studied. Experiments showed that Mn(N) produced MnO₂ with large grain sizes in Mn(N)/TiO₂ catalyst. On the contrary, Mn(Ac) led to highly dispersed and amorphous Mn₂O₃ in Mn(Ac)/TiO₂ catalyst, which had better catalytic activity and stability to SO₂ at low temperatures. The doping of cerium reduced the differences in catalytic performance between the catalysts derived from different Mn precursors.

Implications

Developing catalyst for selective catalytic reduction (SCR) of NO_x from stationary sources at low reaction temperature is an attractive and key area of alternative process for existing SCR technology due to economical and environmental reasons. Manganese-based catalysts were proven very active at low temperature. In this paper, comparative studies were conducted on how manganese precursors were affecting the properties of resultant Mn-based/TiO₂ catalysts. Different precursors could lead to different grain size of active components, manganese oxide dispersion, and surface phase state of Mn/TiO₂ catalyst, whereas these influences are reduced with the doping of cerium oxides into Mn/TiO₂.

Introduction

Nitrogen oxides (NO_x) remain a major source of air pollution. Over 90% NO_x in the atmosphere comes from fossil fuel combustion and about 95% of NO_x in flue gas is present as NO. Currently, selective catalytic reduction of NO with NH₃ (NH₃-SCR) is the most efficient technology in industry for the control of the emission of nitrogen oxides from stationary sources. Typically, the catalysts used in industrial applications are V₂O₅/TiO₂ and V₂O₅-WO₃/TiO₂, which require operating temperatures of 300–400 °C (Busca et al., 1998; Heck, 1999). In order to meet the temperature requirements, SCR units are usually installed between a boiler's economizer and air pre-heater. Due to the position of the SCR unit, the catalyst's lifetime is unavoidably reduced due to attrition and the plugging of the micropores of the catalyst, caused by high concentrations of dust and other contaminants (Parvulescu et al., 1998; Tang et al., 2005; Valdés-Solýs et al., 2003). In order to move the SCR unit behind the pre-heater, it is important to develop low-temperature SCR catalysts with high activity and strong stability to SO₂ and H₂O poisoning.

In recent years, SCR catalysts based on MnO_x/TiO₂ have attracted a great deal of attention due to their high activity at low reaction temperatures (Ettireddy et al., 2007; Jiang et al., 2009; Smirniotis et al., 2006; Wu et al., 2008; Wu et al., 2009). In our previous work, it was observed that catalyst preparation from manganese acetate and a calcination temperature of 500 °C led to a Mn-Ce/TiO₂ catalyst with good activity, even at reaction temperatures below 200 °C (Liu et al., 2006). Under the experimental conditions of 10,000 hr⁻¹ GHSV and 2000 ppm NO, the catalyst was able achieve 90% NO conversion at 120 °C and 95% at 150 °C. The catalysts in the experiment also showed good stability to poisoning by SO₂. However, the catalyst's performance became poor when both SO₂ and H₂O coexisted in the feed gas.

Manganese precursors for Mn-based SCR catalysts have great influence on the activity of the catalyst when prepared by impregnation. Kapteijn et al., (1994a, 1994b) found that a Mn/Al₂O₃ catalyst obtained from manganese acetate had better activity than a catalyst produced from manganese nitrate. This was attributed to highly dispersed manganese oxides on the catalyst surface. Similar work by (Peña Donovan et al., 2004) showed that a Mn/TiO₂ catalyst derived from manganese nitrate and calcinated at 400 °C had better activity at lower temperatures than a catalyst obtained from manganese acetate. However, Li et al., ( 2007) study reached an opposite conclusion and found that a Mn/TiO₂ catalyst with manganese acetate as a precursor had better activity. Currently, few studies have focused on the effect of the precursor on the catalyst's stability to poisoning by both SO₂ and H₂O. In this work, we prepared Mn/TiO₂ and Mn-Ce/TiO₂ catalysts with manganese acetate and manganese nitrate as precursors, respectively, and examined the effect of the precursors on the catalyst's performances. The effects of Ce doping in Mn/TiO₂ were also investigated.

Experimental

Catalyst preparation

Mn/TiO₂ and Mn-Ce/TiO₂ catalysts from different manganese precursors were prepared by an impregnation method. An anatase-type TiO₂ (Nanjing Haitai Co., Nanjing City, China; SA = 80 m²/g, average size = 10 nm) was employed as a catalyst support, whereas analytical-reagent (AR) cerium nitrate (Ce(NO₃)₃·6H₃O) was used as a source of cerium in the catalyst. Manganese precursors included manganese acetate (Mn(CH₃COO)₂·4H₂O, AR) and manganese nitrate (Mn(NO₃)₂, AR). The amount of Mn and Ce that was used was based on the mass of TiO₂ and was 15% and 50% by weight, respectively. Manganese precursor, cerium nitrate (if employed), and titanium dioxide were simultaneously mixed in distilled water, stirred for 1 hr, dried at 120 °C in an oven for 12 hr and calcined at 500 °C for 6 hr in an air atmosphere. After cooling to ambient temperature, the product was milled and filtrated to achieve catalyst particles between 40 to 80 meshes. Catalysts obtained from manganese acetate and manganese nitrate are denoted as Mn(Ac) and Mn(N), respectively, in the following text.

Catalyst activity and characterization

Figure 1 is the schematic diagram of the experimental installation. Catalyst evaluation was conducted in a tubular fixed bed reactor made of stainless steel 20 mm in internal diameter [i.d.]). H₂O and SO₂ were introduced into the reactor as needed. Ten milliliters of catalyst powder was placed on a sieve-like distributor in the middle of the reactor, which was settled in a tubular furnace equipped with a programmable temperature controller. Gas reactants were mixed online and fed into the reactor continually. NH₃ was introduced into the reactor to initiate the reaction only when the NO concentrations at the inlet/outlet of the reactor were identical. This was done to remove the influence of catalyst adsorption on NO conversion. The NO and SO₂ concentrations at the inlet/outlet of the reactor were continually monitored by a chemiluminescent 42C-HL NO-NO₂-NO_x analyzer and a 43C-HL SO₂ analyzer (Thermo Environmental Instruments Inc., Franklin, MA). To avoid errors caused by the oxidation of ammonia by ozone in the converter of the NO-NO₂-NO_x analyzer, an ammonia trap filled with a phosphoric acid solution was installed at the sample inlet of the chemiluminescent analyzer.

Figure 1. Schematic diagram of the experimental installation. 1 = N₂; 2 = O₂; 3 = 10% NO/N₂; 4 = 1.25% SO₂/N₂; 5 = 10% NH₃/N₂; 6 = rotor flowmeters; 7 = mass flowmeters; 8 = thermometers; 9 = temperature controllers; 10 = wash bottle (water); 11 = mix bottle; 12 = ammonia trap bottle (phosphoric acid); 13 = stainless steel tubular reactor; 14 = tube furnace; 15 = 42C-HL NO-NO₂-NO_x analyzer; 16 = 43C-HL SO₂ analyzer.

N₂, O₂, SO₂, and NO used in the experiments were commercial- grade gases of 99.99%, 99.99%, 99.9%, and 99.5% purity, respectively (Beijing Huayuan Gas Chemical Industry Co. Ltd., Beijing, China). The NH₃ used in the experiment was liquid ammonia of 99.99% purity. Three mixed gases of 10% NO/N₂, 10% NH₃/N₂, and 1.25% SO₂/N₂ were prepared and placed into a separate steel cylinder for the following experiments. Water vapor was generated by passing N₂ through a gas-wash bottle containing deionized water. The water temperature in gas-wash bottle could be adjusted to meet the desired water content in the fed gas.

Table 1 lists the standard reaction conditions for the evaluation of catalyst activity. The flow rates of N₂ and O₂ gases were regulated by rotameters. The control of the flow rates of NH₃/N₂ and NO/N₂ was conducted by two mass flowmeters. N₂, O₂, and NO/N₂ gases were mixed in a chamber before being supplied to the reactor. The gas flow of NH₃/N₂ was fed directly into the reactor to avoid a reaction between SO₂ and NH₃. The experimental data were collected when the SCR reaction had reached a steady state for at least 2 hr.

Table 1. Standard reaction conditions for catalyst activity measurement

Parameter	Value
[NO] (ppm)	1000
[NH3] (ppm)	1100
[O2] (%, ϕ)	6
[SO2] (ppm)	200 (if employed)
N2	Balance
Total flow rate (mL·min^−1)	1200
GHSV (hr^−1)	5000

Catalyst characterization

The specific surface areas of the fresh catalysts (Mn(Ac)/TiO₂, Mn(N)/TiO₂, Mn(Ac)-Ce/TiO₂, and Mn(N)-Ce/TiO₂) were measured by the nitrogen adsorption method using a NOVA 4000 (Quantachrome Instruments, Boynton Beach, Florida) automated gas sorption system at −196 °C. The value of the BET surface area (specific surface area based on Brunauer-Emmett-Teller equation) was determined by the two-parameter BET equation. Prior to the surface area measurement, catalyst samples were degassed under vacuum at 400 °C for 24 hr.

Powder X-ray diffraction measurements (XRD) were carried out on a Rigaku D/max-3C system (Rigaku Corporation, Tokyo, Japan) with Cu Kα radiation under the following conditions: a lamp voltage of 40 kV, a current of 100 mA, a step of 0.02, a scan rate of 6°/min and a scan range 2θ = 10–70°.

X-ray photoelectron spectra (XPS) were recorded in a PHI-1600 ESCA system (PerkinElmer Inc., Waltham, MA) spectrometer with Mg Kα as a source of X-rays (E _B = 1653.6 eV) and under a residual pressure of 5.0 × 10⁻⁶ Pa. The error of the binding energy (BE) was ±0.2 eV. Atomic sensitivity factors from Wagner (1977, 1972) were applied to the calibration of peak areas of binding energies. The peak areas were normalized to obtain the surface atom percentage. The charge effect was calibrated with contaminated carbon (C 1s peak at 284.6 eV).

Temperature-programmed reduction (TPR) measurements were conducted on a NOVA (Quantachrome Instruments, Boynton Beach, FL) physics chemistry adsorption system that required 30 mg of catalyst sample and H₂/N₂ mixed gas, in which H₂ accounted for 5%. The mixed gas was used as a reductant and supplied to the system at a flow rate of 20 mL/min. During the experiment, the sample was heated from room temperature to 800 °C and the heating rate was set at 10 °C/min. A thermal conductivity detector (TCD) was used for the evaluation of the composition of the gas.

A Bio-Rad FTS 165 analyzer (Bio-Rad Laboratories, Hercules, CA) was employed for Fourier transform infrared spectrometer (FT-IR) analysis. Before characterization, catalyst samples were cleaned for 1 hr with N₂ gas at the same temperature as the SCR reaction to wipe off physically adsorbed species. After being cooled to room temperature, the catalyst sample was mixed with KBr at a weight ratio of 1:150, ground, and pelleted. The FT-IR spectra were obtained from 4000 to 400 cm⁻¹.

Results and Discussion

BET testing

The results of surface area measurements of fresh Mn/TiO₂ and Mn-Ce/TiO₂ catalysts are listed in Table 2. The surface areas of the products from manganese acetate precursors are larger than those obtained from manganese nitrate. It was also noted from Table 2 that Mn-Ce/TiO₂ has larger surface area than Mn/TiO₂ due to CeO₂ in the catalysts.

Table 2. BET results of Mn/TiO₂ and Mn-Ce/TiO₂ with different precursors

Catalyst	SABET (m^2/g)
Mn(Ac)/TiO2	68.96
Mn(N)/TiO2	62.31
Mn(Ac)-Ce/TiO2	121.89
Mn(N)-Ce/TiO2	108.64

XRD analysis

Figure 2 is the XRD patterns of fresh Mn(Ac)/TiO₂, Mn(N)/TiO₂, Mn(Ac)-Ce/TiO₂, and Mn(N)-Ce/TiO₂ catalysts. The Mn(N)/TiO₂ catalyst has clear diffraction peaks from MnO₂ and TiO₂ of anatase type. However, Mn(Ac)/TiO₂ only has diffraction peaks from TiO₂, whereas the peaks attributed to MnO₂ are not observed. This indicates that the manganese oxides in the Mn(Ac)/TiO₂ catalyst are amorphous and well dispersed. After doping Ce into the Mn/TiO₂ catalysts, the XRD spectra of Mn(Ac)-Ce/TiO₂ and Mn(N)-Ce/TiO₂ are almost identical, in which the only diffraction peaks observed are attributed to TiO₂ and CeO₂. Obviously, there is no manganese oxide in the crystalline phase on the catalysts.

Figure 2. XRD patterns of Mn/TiO₂ and Mn-Ce/TiO₂ with different precursors.

XPS analysis

Figure 3 shows the Mn 2p binding energy (BE) spectra of fresh Mn(Ac)/TiO₂, Mn(N)/TiO₂, Mn(Ac)-Ce/TiO₂, and Mn(N)-Ce/TiO₂ catalysts. The Mn 2p_3/2 binding energy of the Mn(Ac)/TiO₂ and Mn(N)/TiO₂ catalysts are 641.3 and 641.9 eV, respectively. Taking into account the XRD results and BE data, the majority of the manganese oxides present in the Mn(Ac)/TiO₂ and Mn(N)/TiO₂ catalysts are in the form of Mn₂O₃ and MnO₂, respectively. The BE data of manganese oxides in literature were referenced and validated this conclusion (Alvarez-Galván et al., 2004; Kapteijn et al., 1994). This result is also in accordance with Kapteijn et al., (1994) work on Mn/Al₂O₃ catalysts prepared from different manganese precursors.

Figure 3. Spectra of Mn 2p binding energy of Mn/TiO₂ and Mn-Ce/TiO₂ catalysts derived from different precursors.

The Mn 2p_3/2 BEs of Mn(Ac)-Ce/TiO₂ and Mn(N)-Ce/TiO₂ are similar to each other. Each Mn 2p region consists of a spin-orbit doublet, with Mn 2p_1/2 having a BE of nearly 653.1 eV and Mn 2p_3/2 with a BE of nearly 641.7 eV. These values are characteristic of a mixed-valence manganese system (Mn⁴⁺ and Mn³⁺) (Wu et al., 2008; Zhang et al., 2007). This phenomenon suggests that the doping of Ce reduced the valence state of manganese in Mn(N)/TiO₂ catalyst and some of the MnO₂ was converted to Mn₂O₃. The interaction between manganese and cerium forces an electron to move from cerium to manganese, thus reducing the valence state of manganese and enhancing the concentration of Ce⁴⁺ (Chen et al., 2001). Figure 4 shows the Ce3d BE spectra of Mn(Ac)-Ce/TiO₂ and Ce-Mn(N)/TiO₂ catalysts. The BEs of cerium in the catalysts obtained from the two manganese precursors are almost identical, indicating that the cerium oxides in the two catalysts are CeO₂.

Figure 4. Spectra of Ce 3d binding energy of Mn(Ac)-Ce/TiO₂ and Mn(N)-Ce/TiO₂.

Table 3 gives the surface atom ratios of manganese to titanium in the catalysts according to XPS analysis. Compared with the catalysts derived from Mn(N), catalysts obtained from Mn(Ac) have a higher concentration of surface manganese.

Table 3. XPS results of the catalysts

	Binding Energy (eV)
Catalyst	Ti 2p3/2	Ti 2p1/2	O 1s	Ce 3d3/2	Ce 3d3/2	Ce 3d3/2	Mn 2p3/2	Mn 2p1/2	Mn/Ti
Mn(Ac)/TiO2	458.4	468.1	529.8	—	—	—	641.3	653.0	45.2%
Mn(N)/TiO2	458.4	468.1	529.8	—	—	—	641.9	653.4	40.6%
Mn(Ac)-Ce/TiO2	458.4	468.1	529.8	882.7	888.7	898.5	641.7	653.0	32.1%
Mn(N)-Ce/TiO2	458.4	468.1	529.8	882.8	888.7	898.6	641.8	653.1	26.7%

H₂-TPR analysis

According to the literature (Liang et al., 2008; Liu et al., 2007; Yao et al., 2006), in the reduction of MnO_x and CeO₂, pure MnO₂ has reduction peaks at 370 and 604 °C. For pure Mn₂O₃, reduction peaks appear at 468 and 534 °C, which are attributed to the reduction of MnO₂/Mn₂O₃ to Mn₃O₄ and further reduction of Mn₃O₄ to MnO. Reduction peaks of pure CeO₂ are observed at 501 and 746 °C. Thus, the peak near 500 °C is assigned to reduction of CeO₂ at the surface. The peak near 750 °C represents the reduction of lattice CeO₂ in bulk phase. The reduction peak of TiO₂ is very weak, and it is observed at 600 °C (Liang et al., 2008; Liu et al., 2007; Yao et al., 2006).

Figure 5 shows the TPR patterns of fresh Mn(Ac)/TiO₂, Mn(N)/TiO₂, Mn(Ac)-Ce/TiO₂, and Mn(N)-Ce/TiO₂ catalysts. According to above analysis from the literature, the two reduction peaks below 500 °C in the TPR patterns of Mn(N)/TiO₂ and Mn(Ac)/TiO₂ of Figure 5 are assigned to reduction of manganese oxides. In Mn(N)/TiO₂ of Figure 5, a reduction peak in the low temperature range appears at 380 °C, which is the characteristic of the reduction of MnO₂ (Liang et al., 2008; Liu et al., 2007; Yao et al., 2006). This peak is sharp and has a strong intensity, indicating that most manganese species within the Mn(N)/TiO₂ catalyst are MnO₂ in an aggregated state, with large grain sizes. The reduction peak of Mn(N)/TiO₂ at 440 °C is weak, and may result from the reduction of a small amount of Mn₂O₃.

Figure 5. H₂-TPR patterns of Mn/TiO₂ and Mn-Ce/TiO₂ catalysts from different precursors.

Mn(Ac)/TiO₂ yields weak reduction peaks within wide temperature range of 170∼510 °C. Taking into account the results of XPS analysis, these reduction peaks are attributed to the reduction of Mn₂O₃. The starting temperature of 170 °C and the peak point of low-temperature reduction at 305 °C are much lower than pure Mn₂O₃. This suggests a highly dispersed manganese species on TiO₂ support, and the interaction between TiO₂ support and manganese makes Mn–O bond easier to reduce.

The effects of Ce in Ce-doped Mn(Ac)/TiO₂ and Mn(N)/TiO₂ catalysts are different. On one hand, the spectra of Mn(Ac)-Ce/TiO₂ and Mn(Ac)/TiO₂ are similar, with the exception that the initial temperature of reduction reaction decreases to 150 °C and the peak point of low-temperature reduction shifts up to 350 °C after Ce doping into the Mn(Ac)/TiO₂ catalyst. H₂ consumption peak at 450∼600 °C is the result of overlap of further reduction of manganese oxides and reduction of surface CeO₂. The peak near 750 °C related to the reduction of lattice oxygen of CeO₂ in bulk phase was not observed. From these results and the XRD data, it is deduced that cerium in the doped catalyst (Mn(Ac)-Ce/TiO₂) interacts with manganese oxides and decreases reducibility of manganese. However, cerium doping increases the amount of CeO₂ at the surface and enhances the reducibility of lattice oxygen of CeO₂ in bulk phase.

On the other hand, the TPR pattern of the Mn(N)-Ce/TiO₂ catalyst is quite different from that of Mn(N)/TiO₂. Reduction peaks of the Mn(N)-Ce/TiO₂ catalyst become much wider and weaker. The initial temperature of the reduction reaction is lower than that of the catalyst without incorporation of Ce. This suggests that doping of cerium greatly improves dispersity of manganese on the catalyst. The peak point of low-temperature reduction appears at 380 °C, which is at the same position as observed in Mn(N)/TiO₂, suggesting that there is a small amount of MnO₂ in Mn(N)-Ce/TiO₂. In comparison with Mn(Ac)-Ce/TiO₂ catalyst, reduction peaks of Mn(N)-Ce/TiO₂ have shifted to higher temperatures, and a new but weak reduction peak appears near 720 °C. This peak could be attributed to the reduction of lattice oxygen of CeO₂ in bulk phase, suggesting that dispersity of manganese and reducibility of cerium in Mn(N)-Ce/TiO₂ are lower than that in Mn(Ac)-Ce/TiO₂.

Catalyst activity

The activity tests showed that Mn(Ac)/TiO₂, Mn(N)/TiO₂, Mn(Ac)-Ce/TiO₂, and Mn(N)-Ce/TiO₂ catalysts were stable when there are no H₂O and SO₂ in the feed gas. Figure 6 presents NO conversions for the four catalysts at reaction temperatures ranging from 100 to 210 °C. Figure 6 shows that Mn(Ac)/TiO₂ reaches 78% of NO conversion at 100 °C and 93% at 210 °C, which is much higher than that of Mn(N)/TiO₂. Lietti et al., (1996) (Lietti and Forzatti, 1994) stated that NH₃ is prone to adsorb on catalyst at low temperatures, but activation of adsorbed NH₃ is restricted by the low temperature. However, catalyst surface with strong oxidizability could help to promote the activation of adsorbed NH₃ to from NH_2(ads) by dehydrogenation. Therefore, the catalyst's redox capability is crucial for its activity at low temperatures. According to the results of catalyst characterization, Mn(Ac)/TiO₂ possesses good dispersity of manganese species, and high reducible surface. Both do contribute to the higher activity of Mn(Ac)/TiO₂ for NH₃-SCR at low temperatures than that of Mn(N)/TiO₂.

Figure 6. NO conversion over Mn/TiO₂ and Mn-Ce/TiO₂ catalysts from different precursors. NO = 1000 ppm, NH₃ = 1100 ppm, O₂ = 6%, GHSV = 5000 hr⁻¹, balance N₂.

The curves of Mn(Ac)-Ce/TiO₂ and Mn(N)-Ce/TiO₂ are similar to each other, suggesting that activity difference between the two catalysts has become smaller after doping with cerium. This is considered to be due to an increase in the dispersity of manganese species in the catalysts and enhanced catalytic activity, which is a result of the interaction between cerium and manganese. According to the TPR results, the activity of the surface species of CeO₂ in Mn(Ac)-Ce/TiO₂ catalyst is higher than that in Mn(N)-Ce/TiO₂. Mn(Ac)-Ce/TiO₂ catalyst still shows higher activity than Mn(N)-Ce/TiO₂.

The effects of H₂O and SO₂ on catalyst activity

The effects of 200 ppm SO₂ and 10% H₂O on activity of Mn(Ac)/TiO₂, Mn(N)/TiO₂, Mn(Ac)-Ce/TiO₂, and Mn(N)-Ce/TiO₂ catalysts were investigated at 150 °C, as shown in Figure 7. The NO conversion data were collected 2 hr after introduction of SO₂ or SO₂ and H₂O.

Figure 7. The impact of the introduction of SO₂/H₂O for 2 hr on Mn/TiO₂ and Mn-Ce/TiO₂ catalysts from different precursors. NO = 1000 ppm, NH₃ = 1100 ppm, O₂ = 6 %, GHSV = 5000 hr⁻¹, SO₂ = 200 ppm, H₂O = 10% (if used), balance N₂, T = 150 °C.

With SO₂ present in the feed gas, the decrease in catalytic activity was observed in all the tests. Compared with Mn(Ac)/TiO₂, Mn(N)/TiO₂ had much greater loss in activity, from 75% to 42%. However, Mn(Ac)-Ce/TiO₂ and Mn(N)-Ce/TiO₂ catalysts demonstrated similar behaviors. They were less affected, and maintained NO conversions over 85%.

When both SO₂ and H₂O were conducted into the SCR system, the behaviors of Mn(Ac)/TiO₂ and Mn(N)/TiO₂ were similar. Both suffered severe deactivation or inhibition. It was observed that NO conversion dropped to 35% with Mn(Ac)/TiO₂ catalyst. The effects of H₂O and SO₂ on Mn(Ac)-Ce/TiO₂ and Mn(N)-Ce/TiO₂ catalysts were analogous. Activities of the two catalysts fell to around 70% in experiments. Catalyst activity tests were also conducted at extended reaction times to verify the effects of SO₂ and H₂O on Mn(Ac)-Ce/TiO₂ and Mn(N)-Ce/TiO₂, as shown in Figure 8. The two catalysts behaved almost identically in the presence of both SO₂ and H₂O, except that Mn(Ac)-Ce/TiO₂ showed a faster decrease in NO conversion than Mn(N)-Ce/TiO₂ after 500 min. As a comparison, another two sets of NO conversion data over Mn(Ac)-Ce/TiO₂ catalyst, with SO₂ only, and without SO₂ and H₂O in the feed gas, are also given in Figure 8. The data show that Mn(Ac)-Ce/TiO₂ could maintain its activity for a long time without the influence from SO₂ and H₂O. However, when there was SO₂ in the feed gas, NO conversion fell sharply in 9 hr, which is similar to that observed with SO₂ and H₂O in the feed gas.

Figure 8. The effect of SO₂ and H₂O for extended times on Mn-Ce/TiO₂ catalysts. NO = 1000 ppm, NH₃ = 1100 ppm, O₂ = 6 %, GHSV = 5000 hr⁻¹, SO₂ = 200 ppm, H₂O = 10 %, balance N₂, T = 170 °C.

In order to investigate the cause for the difference between Mn(Ac)/TiO₂ and Mn(N)/TiO₂ catalysts in their stabilities to SO₂ poisoning, the oxidation of SO₂ over Mn(Ac)/TiO₂ and Mn(N)/TiO₂ was simulated under conditions with/without H₂O in the feed gas. These experiments were conducted at reaction temperature of 150 °C. Reaction gases (N₂, NO, O₂, and SO₂) with the same concentrations in previous experiments were fed into the reactor. Three hours later, the catalysts were removed from the reactor for FT-IR analysis, and the results are shown in Figure 9.

Figure 9. FT-IR spectra of spent Mn/TiO₂ from different precursors under the influence of SO₂ and H₂O.

Hug (Hug, 1997) and others (Nakamoto, 1986) reported that FTIR spectrum of SO₄ ²⁻ has characteristic bands near 1200, 1130, 1050, and 970 cm⁻¹, whereas HSO₄ ²⁻ gives characteristic bands near 1200, 1040, and 900 cm⁻¹. Obviously, in a FTIR spectrum of (bi)sulfate on catalyst, the band near 1130 cm⁻¹ could be assigned to SO₄ ²⁻, and the band near 900 cm⁻¹ is HSO₄ ²⁻. In Figure 9, there are absorption peaks near 1130 cm⁻¹ (1124 cm⁻¹, 1139 cm⁻¹, and 1130 cm⁻¹), but no peak near 900 cm⁻¹ could be found. This suggests that the S-containing salt on the four tested catalysts is sulfate salt, not bisulfate salt.

In Figure 9, curves a, b, and d have clear absorption peaks near 1130 cm⁻¹, which is attributed to SO₄ ²⁻. However, this peak does not appear in curve c. Obviously, the oxidation of SO₂ over Mn(Ac)/TiO₂ is very weak when H₂O is not present in the feed gases. There should be smaller amount of SO₄ ²⁻ formed on the catalyst than others. On the other hand, Mn(N)/TiO₂ has a catalytic effect on the oxidization of SO₂, as shown in curve d. This is related to MnO₂ in aggregation state in Mn(N)/TiO₂ catalyst. Compared with curve d, the stronger peaks near 1130 cm⁻¹ in curves a and b indicate that H₂O promotes the oxidation of SO₂ over the two Mn/TiO₂ catalysts. In particular, this is observed especially over Mn(Ac)/TiO₂ catalyst. Vadjić and Gentilizza (Vadjić and Gentilizza, 1985) reported similar phenomenon and found that the high relative humidity of ambient air increased the oxidization of SO₂ over MnO₂.

Conclusions

1.	Manganese precursors have great effects on the state and dispersity of manganese species in resultant Mn/TiO2 catalysts. Manganese species in Mn(N)/TiO2 is present as MnO2 with aggregated state and large grain size, whereas that in Mn(Ac)/TiO2 catalyst is amorphous Mn2O3 with high dispersity. Thus, Mn(Ac)/TiO2 catalyst has a higher surface concentration of active constituents, and is easier to be reduced, leading to better performance at low temperature in the SCR process.
2.	Doping of cerium into Mn/TiO2 catalysts can eliminate performance differences between the catalysts derived from manganese acetate and manganese nitrate precursors. Due to the higher activity of surface oxygen species from CeO2, Mn(Ac)-Ce/TiO2 shows better activity than Mn(N)-Ce/TiO2.
3.	Mn(Ac)/TiO2 catalyst shows better stability to SO2 than Mn(N)/TiO2. The impact of SO2 on Mn(Ac)-Ce/TiO2 and Mn(N)-Ce/TiO2 catalyst are similar, which is weaker than on Mn/TiO2 catalyst. When SO2 and H2O coexist in the feed gas, all four catalysts suffered severe deactivation. This is due to the promotion effect of H2O on the oxidization of SO2, no matter what kind of manganese precursor was taken.

Acknowledgments

This work was supported by the Science Foundation for the Excellent Youth Scholars of Ministry of Education of China (20070010017) and the Foundation for Key Discipline Construction of Beijing.