Enhancement of Ce_1−xSn_xO₂ support in LaMnO₃ for the catalytic oxidation and adsorption of elemental mercury

Abstract

Mn-based perovskite oxide was used as the active site for elemental mercury (Hg⁰) removal from coal-fired flue gas. Ce_1−xSn_xO₂ binary oxides were selected as the catalyst supports for LaMnO₃ to enhance the catalytic oxidation and adsorption performance. Ce_0.7Sn_0.3O₂ had the best Hg⁰ removal performance among the as-prepared Ce_1−xSn_xO₂ binary oxides; the Hg⁰ removal efficiency was 95.2% at 350 °C. LaMnO₃ had better performance at low temperatures (<200 °C). LaMnO₃/Ce_0.7Sn_0.3O₂ enlarged the reaction temperature window and enhanced the Hg⁰ removal efficiencies. The correlation between the physicochemical properties and the catalytic removal performance was investigated by XRD, BET surface area measurements, Raman spectroscopy, H₂-TPR and XPS analysis. With the addition of Ce–Sn binary oxides as catalyst support, the surface areas of LaMnO₃ was enlarged, the reducibility was enhanced and the oxygen mobility was improved. In addition, the Hg⁰ removal mechanism was illustrated on the basis of the experimental results. The roles of Ce, Sn and LaMnO₃ were also discussed in this study.

---

1. Introduction

Elemental mercury (Hg⁰) emitted from coal-fired power plants is hazardous pollutant in the atmosphere.^1–3 Hg⁰ is hardly removed due to its insolubility in water and high volatility.⁴ Currently, the efficient methods for the removal of Hg⁰ make full use of the existing air pollution control devices (APCDs) to remove Hg⁰ and NO_x/particles/SO₂.^2,5 In general, catalytic oxidation and adsorption are two primary mechanisms for Hg⁰ removal. Hg⁰ transformed to oxidized state (Hg²⁺) by catalytic oxidation could be easily removed in the wet flue gas desulfurization (WFGD) system. Hg⁰ changed to particle-bound mercury (Hg^p) can be captured by dust-cleaning apparatus. In the dust cleaning process, the catalytic oxidation of Hg⁰ to Hg²⁺ was also beneficial for chemical adsorption. Therefore, catalytic oxidation of Hg⁰ to Hg²⁺ was the key to controlling Hg⁰ emission.

Among the catalysts or sorbents in recent studies, Mn-based metal oxides were proven to be efficient materials for Hg⁰ removal from coal-fired flue gas.^6–12 The high redox potential makes them possible for Hg⁰ oxidation to Hg²⁺, along with the adsorption on the surface of materials. In this process, higher catalytic performance and sufficient surface oxygen were beneficial for Hg⁰ removal. However, Mn-based metal oxides often suffer two problems for Hg⁰ removal: (1) particle aggregation resulted in the low stability, and it is hard to make full use of MnO_x particles; and (2) MnO_x often loses its high activity when the reaction temperature is higher than >200 °C. To solve the first problem, perovskite oxides attracted our attention due to their excellent catalytic oxidation performance. Higher dispersion of Mn ions occupies the B sites in ABO₃ perovskite structure.^13–15 LaMnO₃ indicated high Hg⁰ removal performance in our recent study. However, it also suffers from low activity at high temperature. In our previous studies, doping of SnO₂ was employed to enhance the Hg⁰ removal performance at high temperature.^7,16 Sn–Mn binary metal oxides enlarged the reaction temperature window. In fact, SnO₂ had excellent electron transfer and O₂ capture performance.^17,18 O₂ in the gas can be adsorbed on the surface of SnO₂. Then, the adsorbed O₂ changed to O²⁻ by SnO₂ offering electrons on its surface. The sufficient O²⁻ was favorable for Hg²⁺ capture. In addition, CeO₂ has large oxygen storage capacity and oxygen conversion ability. Ce–Mn mixed metal oxides may have high catalytic oxidation performance because it makes good use of O₂ in the simulated flue gas.¹⁹ Liu et al. reported that doping with Ce into SnO₂ increased the surface area, decreased the crystallite sizes, and it showed higher catalytic activity for the catalytic combustion of methane.¹⁸ It was interesting to determine whether Ce–Sn binary metal oxides can be used for Hg⁰ catalytic oxidation.

In view of the complete perovskite structure, LaMnO₃ had higher catalytic performance. Ce–Sn binary oxides were impossible to dope into the perovskite crystal lattice. To date, few reports have discussed the Ce–Sn binary metal oxides used as catalyst supports. In this study, we first optimized the Ce–Mn binary oxides for Hg⁰ removal. Subsequently, the Ce–Mn binary oxides were used as a catalyst support for LaMnO₃. The performances for Hg⁰ removal over these as-prepared materials were evaluated in a fixed-bed adsorption system. The roles of LaMnO₃ and the Ce–Sn support were discussed in our study. On this basis, the relationship between the active center and catalyst support was also discussed by physico-chemical characterization. The possible Hg⁰ removal mechanism, as well as the application prospect, was proposed accordingly.

2. Experimental section

2.1 Materials preparation and characterization

Ce_1−xSn_xO₂ (x = 0.1, 0.3, 0.5, 0.7, and 0.9) supports were prepared by a co-precipitation method. Suitable amounts of Ce(NO₃)₄ and SnCl₄ dissolved in distilled water. A stoichiometric amount of ammonia was added to the mixture as the precipitation agent under strong stirring for 2 h. The precipitate was then filtered and washed with deionized water three times to remove Cl⁻ from the water. Lastly, the precipitate was transferred to a muffle furnace and calcined at 500 °C for 5 h. All the samples used as catalysts were ground to a 40–60 mesh size.

Ce_1−xSn_xO₂ was selected as a catalyst support for LaMnO₃. The prepared Ce_1−xSn_xO₂ supports were ground to <100 mesh size. The LaMnO₃/Ce_1−xSn_xO₂ was prepared using a sol–gel method. The required amount of La(NO₃)₃ and Mn(NO₃)₃ was dissolved together in Ce_1−xSn_xO₂ solution, followed by addition of citric acid (CA) in the mixed solution. The temperature of the aqueous solution was maintained constant at 80 °C. The molar ratio for each component was La/Mn/CA = 1 : 1 : 2. After vigorous stirring and evaporation, a transparent gel was formed, which was then dried at 90 °C overnight. The obtained precursor was first calcined at 400 °C for 1 h in air to decompose citric acid totally and was then calcined at 750 °C for 5 h with a rate of 10 °C min⁻¹. All the samples were grounded to 40–60 mesh.

The as-prepared materials were characterized by means of BET surface area measurements, XRD, Raman spectroscopy, H₂-TPR and XPS. The detail analysis can be found in ESI.†

2.2 Activities test

The performance for Hg⁰ removal over the as-prepared materials was evaluated using a fixed-bed reactor. A Hg⁰ permeation tube was used to generate Hg⁰ vapor carried by pure N₂, which was introduced to the inlet of the gas mixer. Other gases such as O₂ were introduced to the gas mixer at constant flows. The mass flow rate was controlled by mass flow controllers (MFC). A fixed-bed reactor system was used to investigate the Hg⁰ adsorption performance. The reaction temperature was controlled from 100 to 350 °C using a temperature controller tubular furnace. In each test, the as-prepared materials were placed into the reaction tube, which was placed to control the reaction temperature. The cold vapor atomic absorption spectroscopy (CVASS) analyzer was used as the online continuous detector. The concentration of Hg⁰ was calculated by Lumex RA 915+. The inlet concentration of Hg⁰ was 500 ± 50 μg m⁻³. At the beginning of each test, the simulated gas bypassed the reactor and the inlet gas was detected to ensure a stable Hg⁰ concentration. The simulated gas passed the samples and the Hg⁰ concentration was detected by CVASS online. The mass of the material for each test was 30 mg.

The Hg⁰ removal efficiency was calculated according to eqn (1):

where η_x is the removal efficiency, Hg⁰_in is the inlet concentration of Hg⁰, and Hg⁰_out is the outlet concentration of Hg⁰; the reaction time was 600 min for each test.

2.3 Hg-TPD method

Mercury temperature programmed desorption (Hg-TPD) method was built to evaluate the desorption performance of the as-prepared materials. Before each test, the sorbents were first under adsorption for 20 min at 150 °C with 4% O₂ balanced with N₂ (total flow rate = 500 mL min⁻¹). After the furnace was cooled to 100 °C, the materials were regenerated by heating from 100 °C to 700 in a pure N₂ carrier gas. The heating rate was set as 5 °C min⁻¹. The mercury signal was recorded by CVASS online system.

3. Results & discussion

3.1 Hg⁰ removal performance

3.1.1 Hg⁰ removal performance over Ce_1−xSn_xO₂ mixed oxides. The activities of Ce_1−xSn_xO₂ mixed oxides are shown in Fig. 1. CeO₂ showed nearly no activities for Hg⁰ removal at 100–350 °C. For SnO₂, it exhibited no activities when the temperature was less than 200 °C. However, the removal efficiencies were enhanced when the temperature was higher than 200 °C. The Hg⁰ removal efficiency was approximately 50% when the temperature was 300 °C. The results indicated that SnO₂ had high catalytic activity at high temperatures. SnO₂ could adsorb O₂ in the simulated gas to form O²⁻ on the its surface, which was beneficial for Hg⁰ oxidation.^20,21 For Ce_1−xSn_xO₂ mixed oxides, they all showed high activities compared with pure SnO₂. Ce_0.1Sn_0.9O₂ enhanced about 30% Hg⁰ removal efficiencies; the highest removal efficiency was higher than 70% at 350 °C. Among the mixed oxides, Ce_0.7Sn_0.3O₂ showed the best Hg⁰ removal performance. The Hg⁰ removal efficiency was about 90% at 200 °C and the highest removal efficiency was 95.2% at 350 °C. Ce_0.3Sn_0.7O₂, Ce_0.5Sn_0.5O₂ and Ce_0.9Sn_0.1O₂ showed similar performances for Hg⁰ removal. Based on these results, Ce_1−xSn_xO₂ mixed oxides enhanced the Hg⁰ removal efficiency significantly. Among the Ce–Sn binary oxides, Ce_0.7Sn_0.3O₂ showed the highest activities for Hg⁰ oxidation.

Fig. 1 Hg⁰ removal efficiencies of the Ce_1−xSn_xO₂ mixed oxides. Mass of catalyst: 30 mg; total flow rare: 500 mL min⁻¹; gas components: 4% O₂; reaction time: 600 min.

3.1.2 Hg⁰ removal performance using Ce_0.7Sn_0.3O₂ as catalyst support. It was indicated that Ce_0.7Sn_0.3O₂ had the highest activity for Hg⁰ removal. Ce_0.7Sn_0.3O₂ was used as a catalyst support for the modification of LaMnO₃. For comparison, the Hg⁰ removal performance of LaO_x/Ce_0.7Sn_0.3O₂ and MnO_x/Ce_0.7Sn_0.3O₂ were also tested. The reaction temperature was set to 150 °C because the material used for Hg⁰ adsorption is usually in the dust removal unit wherein the temperature ranged from 100 to 200 °C. As shown in Fig. 2, the performance of LaO_x/Ce_0.7Sn_0.3O₂, MnO_x/Ce_0.7Sn_0.3O₂ and LaMnO₃/Ce_0.7Sn_0.3O₂ was different. LaO_x/Ce_0.7Sn_0.3O₂ had almost no activity for Hg⁰ removal. MnO_x/Ce_0.7Sn_0.3O₂ had the highest Hg⁰ removal efficiency (∼80%) at the initial minutes. However, it lost the activity gradually and it completely had no activity until the reaction time was 600 min. Among these three materials, LaMnO₃/Ce_0.7Sn_0.3O₂ had the best performance for Hg⁰ removal. The Hg⁰ removal efficiency was higher than 90% even after a 600 min reaction. This indicates that the perovskite crystal structure was beneficial for Hg⁰ oxidation. In previous studies, the perovskite oxides had been proven to have high catalytic oxidation performance.¹⁴ The Hg⁰ removal mechanism over LaMnO₃ perovskite oxides would be discussed in the following section.

Fig. 2 Hg⁰ removal efficiencies of LaO_x/Ce_0.7Sn_0.3O₂, MnO_x/Ce_0.7Sn_0.3O₂ and LaMnO₃/Ce_0.7Sn_0.3O₂. Mass of catalyst: 30 mg; total flow rare: 500 mL min⁻¹; gas components: 4% O₂; reaction temperature: 150 °C, reaction time: 600 min.

To further identify the effects of the Ce_0.7Sn_0.3O₂ catalyst support for LaMnO₃ on the Hg⁰ removal efficiency, the performances of LaMnO₃/CeO₂, LaMnO₃/SnO₂ and LaMnO₃/Ce_0.7Sn_0.3O₂ were evaluated at a wide temperature window. As shown in Fig. 3, LaMnO₃/CeO₂ had the highest Hg⁰ removal efficiency (94.5%) at 150 °C, while the removal efficiency decreased to only 38.5% when the temperature was increased to 300 °C. Higher temperatures resulted in low activity over Mn-based material. However, when SnO₂ acted as catalyst support for LaMnO₃, the Hg⁰ removal efficiency increased, especially at high temperatures (>200 °C). The Hg⁰ removal efficiency was 85.63% at 250 °C, which was higher than that compared with 50.3% of LaMnO₃/CeO₂ at the same temperature. For LaMnO₃/Ce_0.7Sn_0.3O₂, the performance was further enhanced compared to that of LaMnO₃/SnO₂. The Hg⁰ removal efficiency was greater than 90% even when the temperature was as high as 250 °C, and the Hg⁰ removal efficiency can maintain about 60% at 300 °C. These results indicated that SnO₂ support can enlarge the reaction temperature window and Ce_0.7Sn_0.3O₂ acted as LaMnO₃'s support showed excellent Hg⁰ removal performance.

Fig. 3 Hg⁰ removal efficiencies of LaMnO₃/SnO₂, LaMnO₃/CeO₂ and LaMnO₃/Ce_0.7Sn_0.3O₂ at a wide temperature window (100–300 °C). Mass of catalyst: 30 mg; total flow rare: 500 mL min⁻¹; gas components: 4% O₂; reaction temperature: 100–300 °C, reaction time: 600 min.

3.1.3 Effects of O₂, SO₂ and H₂O on LaMnO₃/Ce_0.7Sn_0.3O₂. As shown in Fig. 4, the effects of various gas components on Hg⁰ removal efficiencies over LaMnO₃/Ce_0.7Sn_0.3O₂ were investigated. The reaction temperature was 150 °C and the reaction time was 600 min. It was obvious that the Hg⁰ removal efficiency was dropped by about 40% under pure N₂ compared to 4% O₂. When there was 8% O₂ in the simulated gas, the Hg⁰ removal efficiency increased. This indicates that O₂ participated in the reaction, O₂ acted as oxidant for Hg⁰ oxidation.²² The effects of H₂O and SO₂ on Hg⁰ removal were also tested. The results showed that 4% H₂O in the simulated gas was unfavorable for Hg⁰ removal, the removal efficiency decreased to 70.7%. In addition, 500 ppm SO₂ dropped the Hg⁰ removal efficiency. The Hg⁰ removal efficiency decreased sharply when 4% O₂ and 500 ppm SO₂ co-existed in the simulated gas. For LaMnO₃/Ce_0.7Sn_0.3O₂, it also suffered from poisoning by SO₂ and H₂O, as well as other Mn-based materials for Hg⁰ removal.¹⁶

Fig. 4 Effects of O₂, SO₂, and H₂O on the Hg⁰ removal efficiencies.

3.2 Characterization of the materials

X-ray power diffraction (XRD) was utilized to clarify the crystal phases of the as-prepared Ce_1−xSn_xO₂ binary metal oxides and their supported material. As depicted in Fig. 5(a), it showed the XRD patterns of Ce_1−xSn_xO₂. For SnO₂, the diffraction peaks of can be well indexed to SnO₂ (JPCDS no. 21-1250). The diffraction peaks of CeO₂ were ascribed to CeO₂ (JPCDS no. 43-1002). With the addition of CeO₂ in SnO₂, the Ce_0.1Sn_0.9O₂ had the CeO₂ phase. However, there were no SnO₂ phases for Ce_0.9Sn_0.1O₂ binary oxides. It could be speculated that the Sn was dispersed on the CeO₂ surface and/or incorporated into CeO₂ lattice.¹⁸ As for the Sn-rich materials, the XRD patterns of the Ce_0.1Sn_0.9O₂ and Ce_0.3Sn_0.7O₂, the patterns existed as a mixed form of CeO₂ and SnO₂. Even for the Ce_0.5Sn_0.5O₂ and Ce_0.7SnO_0.3O₂, the mixing phase co-existed, but the intensity of the SnO₂ phase was not as strong as CeO₂. It could be speculated that, for Ce_0.7Sn_0.3O₂, some of the SnO₂ entered into the lattice of CeO₂ and some of SnO₂ existed on the surface of the mixed metal oxides. Furthermore, as shown in Fig. 5(b), the XRD patterns of LaMnO₃/CeO₂, LaMnO₃/CeO₂ and LaMnO₃/Ce_0.7Sn_0.3O₂ were also represented. The peaks at 22.8, 32.6, 40.2, 46.7, 52.8, 58.1, 68.3 and 77.8° 2θ, all the characteristic peaks were indexed to a perovskite phase (PDF-88-0633). For LaMnO₃/CeO₂, the primary crystal was the perovskite structure. Some peaks can be ascribed to CeO₂.²³ No other peaks were generated on the composite surface. Similarly, the peaks of LaMnO₃/SnO₂ had the crystal phase of perovskite oxide and SnO₂. For LaMnO₃/Ce_0.7Sn_0.3O₂, perovskite oxide phase was also the primary crystal structure, LaMnO₃ and Ce_0.7Sn_0.3O₂ maintained their respective crystal phase.

Fig. 5 XRD patterns of (a) Ce_1−xSn_xO₂ mixed oxides, and (b) LaMnO₃/SnO₂, LaMnO₃/Ce_0.7Sn_0.3O₂ and LaMnO₃/CeO₂.

BET analysis was conducted for the as-prepared samples. The BET surface area and the BJH total pore volume are listed in Table 1. Pure LaMnO₃ had a surface area of 16.60 m² g⁻¹, which is lower than pure SnO₂ (29.89 m² g⁻¹) and CeO₂ (46.32 m² g⁻¹). For Ce_0.7Sn_0.3O₂, the surface area was 35.74 m², which was larger than SnO₂ but smaller than CeO₂. The addition of SnO₂ to CeO₂ increased the pore volume, resulting in a larger surface area. Similarly, the surface area of LaMnO₃/SnO₂ (21.77 m²) was smaller than LaMnO₃/CeO₂ (66.35 m²). However, when Ce_0.7Sn_0.3O₂ acted as the catalyst support, MnO_x/Ce_0.7Sn_0.3O₂ had surface areas of only 4.37 m² and a very small pore volume (0.056 m³). This was the reason why the MnO_x particles was easily aggregated.⁸ LaO_x/Ce_0.7Sn_0.3O₂ had a larger surface area of 74.21 m² and the larger pore volume of 0.625 m³. LaO_x can form porous structures combined with Ce–Sn binary oxides. For LaMnO₃/Ce_0.7Sn_0.3O₂, it had the largest surface area (86.08 m²) among the as-prepared materials. Mn-based oxides often suffer from the problem of particle aggregation. CeO₂ can enlarge the surface area of MnO_x and SnO₂ particles addition could make a larger pore volume, resulting in a larger surface area. The larger surface area was favorable for the Hg⁰ reaction on its surface. According to XRD analysis, the LaMnO₃ perovskite oxide and Ce–Sn binary oxide maintained had their own crystal structure. The cooperation of them benefited the building the larger pore and the dispersion of Mn-based material.

Table 1 BET surface areas and the pore volume of the as-prepared materials

Materials	BET surface area (m^2 g^−1)	BJH pore volume (m^3)
LaMnO3	16.60	0.118
SnO2	29.89	0.141
CeO2	46.32	0.130
Ce0.7Sn0.3O2	35.74	0.152
LaOx/Ce0.7Sn0.3O2	74.21	0.625
MnOx/Ce0.7Sn0.3O2	4.37	0.056
LaMnO3/SnO2	21.77	0.201
LaMnO3/CeO2	66.35	0.302
LaMnO3/Ce0.7Sn0.3O2	86.08	0.434

---

To reveal the effect of Ce–Sn binary oxide support on LaMnO₃, Raman spectroscopy technique was used for structure analysis. For comparison, LaO_x/Ce_0.7Sn_0.3O₂ and MnO_x/Ce_0.7Sn_0.3O₂ were also collected. As shown in Fig. 6, a wide peak at about 500–750 cm⁻¹ was observed for LaMnO₃/SnO₂. This was attributed to the surface vibrational mode of Sn–O–Sn (A_1g) in nanocrystalline SnO₂ and to volume modes of SnO₂.²⁴ The peak at about 648 cm⁻¹ was ascribed to the overlap of MnO₂ with SnO₂.²⁵ There were no obvious peaks in the LaMnO₃/CeO₂ spectrum under the same intensity. With the addition of Sn to the CeO₂ support, the A_1g peak widened and vanished due to the presence of bulk oxygen vacancies, resulting in a distortion of the Sn–O bond.¹⁸ The peak at about 461, 462 or 468 cm⁻¹ can be ascribed to the F_2g vibration mode of CeO₂ in the fluorite structure, the slight shift from 461 to 468 cm⁻¹ in the F_2g mode was due to nanosize sample.²⁶ Obviously, MnO₂ overlapped on SnO₂ from the spectrum of LaMnO₃/Ce_0.7Sn_0.3O₂, the addition of CeO₂ increased more oxygen vacancies.¹⁷ The addition of Sn enhanced the role of MnO₂ which was the main active site for Hg⁰ catalytic oxidation. Significantly, the addition of Ce to catalyst support generated more oxygen vacancies, which could enhance the redox ability of the material, resulting in the higher catalytic performance.

Fig. 6 Raman spectra of LaO_x/Ce_0.7Sn_0.3O₂, MnO_x/Ce_0.7Sn_0.3O₂, LaMnO₃/Ce_0.7Sn_0.3O₂, LaMnO₃/SnO₂, and LaMnO₃/CeO₂.

The reducibility of the samples was assessed by H₂-TPR. Fig. 7 shows the qualitative profiles recorded over the as-prepared materials. For LaO_x/Ce_0.7Sn_0.3O₂, it presented a wide peak centered at about 531.1 °C. In general, Sn⁴⁺ can be only reduced to Sn²⁺ below 500 °C, and Sn²⁺ reduced to Sn⁰ at the temperature higher than 500 °C. The shoulder peak over LaO_x/Ce_0.7Sn_0.3O₂ could indicate the role of Ce, the reducibility of Ce⁴⁺ and Sn⁴⁺ could be promoted simultaneously according to the equilibrium of Sn⁴⁺ + 2Ce³⁺ ↔ Sn²⁺ + 2Ce⁴⁺.¹⁸ MnO_x/Ce_0.7Sn_0.3O₂ presented two character peaks at 488.9 and 595.6 °C, which were assigned to Mn⁴⁺ → Mn³⁺ and Mn³⁺ → Mn²⁺, respectively.²⁷ In the LaMnO₃/SnO₂ profile, there were four peaks at 325.2, 547.5, 632.3 and 799.5 °C. The reduction peak of Mn⁴⁺ → Mn³⁺ moved to a lower temperature (325.2 °C). It was speculated that the SnO₂ support was beneficial for the redox of Sn²⁺ + Mn⁴⁺ ↔ Sn⁴⁺ + Mn³⁺. In addition, the peak of Mn³⁺ → Mn²⁺ moved to a higher temperature compared to MnO_x/Ce_0.7Sn_0.3O₂. However, with the addition of Ce to the SnO₂ support, for LaMnO₃/CeO₂, a wide shoulder peak contributed to Mn⁴⁺ → Mn³⁺ → Mn²⁺ at temperatures lower than 500 °C. Moreover, in the spectrum of LaMnO₃/Ce_0.7Sn_0.3O₂, two peaks at 344.8 and 530.4 °C can be ascribed to Mn⁴⁺ → Mn³⁺ and Mn³⁺ → Mn²⁺, respectively.²⁸ The reduction of Mn⁴⁺ → Mn³⁺ was beneficial for Hg⁰ oxidation. Ce–Sn binary oxides improved the reducibility compared with MnO_x/Ce_0.7Sn_0.3O_2.

Fig. 7 H₂-TPR profile of the as-prepared materials: (a) LaO_x/Ce_0.7Sn_0.3O₂, (b) MnO_x/Ce_0.7Sn_0.3O₂ and (c) LaMnO₃/Ce_0.7Sn_0.3O₂, (d) LaMnO₃/CeO₂, (e) LaMnO₃/SnO₂.

Based on the H₂-TPR results, it indicated that the perovskite structure (LaMnO₃/Ce_0.7Sn_0.3O₂) showed higher reducibility than pure MnO_x/Ce_0.7Sn_0.3O₂, which was beneficial for catalytic oxidation. Furthermore, Sn²⁺ can easily offer electrons, which also can enhance the reducibility of Mn⁴⁺ to Mn³⁺. The addition of Ce enhanced the reducibility of SnO₂ and MnO_x. It was speculated that the reduction of Mn⁴⁺ → Mn³⁺ → Mn²⁺ can be enhanced by Sn²⁺ → Sn⁴⁺ or Ce²⁺ → Ce³⁺ → Ce⁴⁺. CeO₂ and SnO₂ can also interact with each other. The interaction among the catalytic support and active sites resulted in high catalytic oxidation performance.

To discuss the Hg⁰ removal mechanism over LaMnO₃/Ce_0.7Sn_0.3O₂, the XPS spectra of O, Mn and Hg over the fresh and after adsorption of LaMnO₃/Ce_0.7Sn_0.3O₂ are shown in Fig. 8. For the fresh LaMnO₃/Ce_0.7Sn_0.3O₂ sample, Fig. 8(a) shows two peaks of O 1s at 531.1 and 529.6 eV, which corresponded to the adsorbed surface oxygen (O_ads) and the lattice oxygen (O_latt), respectively.⁸ The O_ads/O_latt ratio was 51.04/48.96. After adsorption, as shown in Fig. 8(b), two peaks were detected at the same position over the used LaMnO₃/Ce_0.7Sn_0.3O₂ sample. However, the ratio of O_ads/O_latt decreased to 49.98/50.02. The surface oxygen was more active than the lattice oxygen based on previous studies.²⁸ The oxygen took part in the Hg⁰ oxidation process, which is in agreement with the experiment results.

Fig. 8 XPS spectra of LaMnO₃/SnO₂ in the O 1s region: (a) fresh sample and (b) after adsorption sample; Mn 2p region: (c) fresh sample and (d) after adsorption sample; and Hg 4f region: (e) after adsorption sample.

Fig. 8(c) displays the Mn 2p XPS spectra over the fresh LaMnO₃/Ce_0.7Sn_0.3O₂ sample, the double main peaks were ascribed to Mn 2p_1/2 and Mn 2p_3/2. Based on the analysis, the fitting peaks at 644.0 and 642.0 were ascribed to Mn⁴⁺ species and Mn³⁺ species, respectively.⁸ No peak could be fitted to Mn²⁺ by analysis calculation. The ratio of Mn⁴⁺/Mn³⁺ was 31.88/68.12. The previous studies indicated that the higher valance of Mn, the higher the catalytic oxidation for Hg⁰.²⁸ MnO₂ was considered the primary phase of Mn-based material for Hg⁰ oxidation. After adsorption, as the spectra shown in Fig. 8(d), it was interesting that, there was only one fitting peak at 642.3 eV, which corresponded to Mn³⁺. Obviously, the used material lost Mn⁴⁺ species due to the reduction of Mn⁴⁺ to Mn³⁺ (Mn⁴⁺ + e⁻ → Mn³⁺).

In addition, the Hg 4f spectra of the used LaMnO₃/Ce_0.7Sn_0.3O₂ sample were shown in Fig. 8(e). Two character peaks at 100 to 110 eV can be ascribed to Hg–O on the surface of the material.²⁹ This further indicated that the mercury removal process was primarily due to chemical adsorption.

The Hg-TPD curves were collected to present the property of mercury binding on the surface of LaMnO₃-based material. As shown in Fig. 9, the Hg-TPD curves were collected under 5 °C min⁻¹. The desorption peaks of LaMnO₃/SnO₂ and LaMnO₃/CeO₂ were centered at 371.4 and 342.3 °C, respectively. Without CeO₂ or SnO₂ as the catalyst support, the desorption peak of LaMnO₃ was at 360.8 °C. However, it was obvious that the desorption peak of LaMnO₃/Ce_0.7Sn_0.3O₂ was quite different. There were two peaks at about 300.0 and 697.5 °C, suggesting that two forms of mercury combined with LaMnO₃/Ce_0.7Sn_0.3O₂. The higher desorption temperature indicated stronger interaction forces between the mercury and adsorption sites. The definite bonding styles of mercury can be calculated based on the first-principle theory, which will be performed in our further study. The Hg-TPD results also provided us regeneration method for mercury desorption.

Fig. 9 Hg-TPD curves of LaMnO₃/SnO₂, LaMnO₃/CeO₂, LaMnO₃ and LaMnO₃/Ce_0.7Sn_0.3O₂ at the heating rate of 5 °C min⁻¹.

3.3 Hg⁰ removal mechanism over LaMnO₃/Ce_0.7Sn_0.3O₂

3.3.1 The role of LaMnO₃ perovskite oxide structure. Based on the abovementioned discussions, LaMnO₃ was the main active site for Hg⁰ removal. In general, Hg⁰ removal can be simply described to two steps: first, Hg⁰ in the simultaneous flue gas was adsorbed on the surface of LaMnO₃, followed by catalytic oxidation of Hg⁰ to Hg²⁺; second, the Hg²⁺ existed on the surface in the form of the Hg–O state. According to the XPS analysis, the mechanism can be explained as follows:

Hg⁰(g) → Hg⁰(ads) (2)

Hg⁰(ads) + 2≡Mn⁴⁺ + ≡O²⁻ → 2≡Mn³⁺ + ≡Hg–O (3)

where ≡ represented the adsorption state.

As the active center of the catalyst, LaMnO₃ showed high catalytic activity for Hg⁰ oxidation due to its special crystal structure.¹⁴ Lanthanum-based perovskite oxides had demonstrated remarkable catalysis performance due to the higher redox behavior, oxygen mobility and ionic conductivity.³⁰ As illustrated in Fig. 10, oxygen vacancies (where “■” represented oxygen vacancies) can be introduced to the structure to facilitate oxygen transfer and thus increased oxygen mobility.³¹ O₂ introduced to oxygen vacancies and formed O²⁻, which was favorable for Hg⁰ oxidation and adsorption. This mechanism also indicated that the Hg⁰ removal efficiency decreased when there were no O₂ in the simulated gas.

Fig. 10 Possible Hg⁰ removal pathways over LaMnO₃/Ce_0.7Sn_0.3O₂ material.

3.3.2 The role of Ce–Sn binary oxides catalyst support. With the addition of Ce–Sn binary oxides as the catalyst support, the Hg⁰ removal performances were enhanced. Based on the physical characterization results, the surface areas were enlarged and the oxygen vacancies were increased. Hg⁰ can easily react with the active Mn sites due to the larger surface area. However, the enhancement for Hg⁰ removal at high temperatures should be explained mainly as chemical-adsorption mechanism.

With the addition of SnO₂, the performance at high temperature was enhanced. SnO₂ was believed to be an efficient catalyst that makes full use of O₂ in the simulated gas at high temperatures (>200 °C).^20,21 O₂ in the simulated gas can be adsorbed and exist as physical-adsorption (p-ads) O₂. SnO₂ had super electron transfer performance, which could offer electrons for O₂, the O₂ (p-ads) can be formed O₂⁻ then to O⁻ and O²⁻ by getting electrons. The O₂ (p-ads) changed to chemical-adsorption O₂, which was beneficial for Hg⁰ capture. During this process, the electrons came from the reduction of Mn⁴⁺ to Mn³⁺. The adsorbed O²⁻ on the surface of catalysts was binding site for the oxidized mercury. The mechanism for O₂ transformation can be illustrated as follows:

O₂(g) → O₂(p-ads) (4)

O₂(p-ads) + e⁻ → O₂⁻(c-ads) (5)

O₂⁻(c-ads) + e⁻ → 2O⁻(c-ads) (6)

O⁻(c-ads) + e⁻ → O²⁻(c-ads) (7)

The abovementioned reactions occurred at the temperatures higher than 200 °C. Therefore, the performance of LaMnO₃ at high temperatures (>200 °C) can be enhanced after SnO₂ acted as catalyst support.

Moreover, CeO₂ was believed to have high oxygen storage capacity. The adsorbed oxygen can be stored in CeO₂. On the one hand, the sufficient oxygen was beneficial for mercury adsorption. On the other hand, SnO₂ can adsorb O₂ from simulated gas for CeO₂ utilization. In addition, CeO₂ can re-oxidize the reduced Mn (Mn³⁺) to Mn⁴⁺, and Ce₂O₃ can be re-oxidized to CeO₂ by O₂. So the catalyst can be re-used. The equation can be illustrated as follows:

2Mn³⁺ + 2CeO₂ → 2Mn⁴⁺ + Ce₂O₃ + ≡O²⁻ (8)

Ce₂O₃ + 1/2O₂(g) → 2CeO₂ (9)

The interaction between Sn and Ce oxides is illustrated in Fig. 10. O₂ in the simulated gas can be captured by Ce–Sn binary oxides when O²⁻ is given to LaMnO₃ for the capture of mercury. The reduced Ce (Ce³⁺) or Sn (Sn²⁺) can be re-oxidized by the adsorbed-O₂. The catalyst support is regenerated. Furthermore, based on the H₂-TPR results, Ce and Sn also enhanced the reducibility of LaMnO₃. CeO₂ and SnO₂ benefited the regeneration of reduced Mn. The interaction between Sn, Ce and Mn can be illustrated as follows:

Sn⁴⁺ + 2Ce³⁺ ↔ Sn²⁺ + 2Ce⁴⁺ (10)

Sn²⁺ + Mn⁴⁺ ↔ Sn⁴⁺ + Mn³⁺ (11)

Ce³⁺ + Mn⁴⁺ ↔ Ce⁴⁺ + Mn³⁺ (12)

4. Conclusions

Herein, Ce_1−xSn_xO₂ binary metal oxides were synthesized and evaluated for the catalytic oxidation removal of Hg⁰. The results indicated that Ce_0.7Sn_0.3O₂ had the best performance for Hg⁰ removal. LaMnO₃/Ce_0.7Sn_0.3O₂ presented the highest performance among the as-prepared materials. The surface area was enlarged, the oxygen vacancies were promoted and the reducibility was enhanced when Ce_0.7Sn_0.3O₂ acted as the catalyst support. LaMnO₃ were the main active sites for Hg⁰ removal, but it lost its activity when the temperature was higher than 200 °C. SnO₂ enhanced the high-temperature performance because it can adsorb O₂ in the simulated gas to generate active O²⁻. CeO₂ enhanced the oxygen storage capacity and enhanced the catalyst performance. The Ce–Sn catalyst support-modified materials will be a potential catalyst not only for Hg⁰ removal, but also for other catalytic oxidation reactions.

Acknowledgements

This study was supported by the Major State Basic Research Development Program of China (973 Program, No. 2013CB430005), the National Natural Science Foundation of China (No. 51478261 and No. 51278294). This work was also funded by Collaborative Innovation Center of Environmental Pollution Control and Ecological Restoration, Henan Province (XTCX-002). Thanks for Shanghai Tongji Gao Tingyao Environmental Science and Technology Development Foundation.

References

1. Z.-w. Wang, Z.-s. Chen, D. Ning and X.-s. Zhang, J. Environ. Sci., 2007, 19, 176–180.
2. J. H. Pavlish, E. A. Sondreal, M. D. Mann, E. S. Olson, K. C. Galbreath, D. L. Laudal and S. A. Benson, Fuel Process. Technol., 2003, 82, 89–165.
3. S. Wang, L. Zhang, G. Li, Y. Wu, J. Hao, N. Pirrone, F. Sprovieri and M. Ancora, Atmos. Chem. Phys., 2010, 10, 1183–1192.
4. A. L. Soerensen, R. P. Mason, P. H. Balcom, D. J. Jacob, Y. Zhang, J. Kuss and E. M. Sunderland, Environ. Sci. Technol., 2014, 48, 11312–11319.
5. L. Zhao, C. Li, X. Zhang, G. Zeng and J. Zhang, Catal. Sci. Technol., 2015, 5, 3459–3472.
6. J. Li, N. Yan, Z. Qu, S. Qiao, S. Yang, Y. Guo, P. Liu and J. Jia, Environ. Sci. Technol., 2009, 44, 426–431.
7. J. Xie, H. Xu, Z. Qu, W. Huang, W. Chen, Y. Ma, S. Zhao, P. Liu and N. Yan, J. Colloid Interface Sci., 2014, 428, 121–127.
8. H. Xu, Z. Qu, C. Zong, W. Huang, F. Quan and N. Yan, Environ. Sci. Technol., 2015, 49(11), 6823–6830.
9. L. Xing, Y. Xu and Q. Zhong, Energy Fuels, 2012, 26, 4903–4909.
10. H. Li, C.-Y. Wu, Y. Li and J. Zhang, Appl. Catal., B, 2012, 111, 381–388.
11. D. Jampaiah, K. M. Tur, P. Venkataswamy, S. J. Ippolito, Y. M. Sabri, J. Tardio, S. K. Bhargava and B. M. Reddy, RSC Adv., 2015, 5, 30331–30341.
12. D. Jampaiah, S. J. Ippolito, Y. M. Sabri, B. M. Reddy and S. K. Bhargava, Catal. Sci. Technol., 2015, 5, 2913–2924.
13. S. Cimino, S. Colonna, S. De Rossi, M. Faticanti, L. Lisi, I. Pettiti and P. Porta, J. Catal., 2002, 205, 309–317.
14. S. Royer, D. Duprez, F. Can, X. Courtois, C. Batiot-Dupeyrat, S. Laassiri and H. Alamdari, Chem. Rev., 2014, 114, 10292–10368.
15. H. Najjar and H. Batis, Appl. Catal., A, 2010, 383, 192–201.
16. H. Xu, J. Xie, Y. Ma, Z. Qu, S. Zhao, W. Chen, W. Huang and N. Yan, Fuel, 2015, 140, 803–809.
17. H. Chang, X. Chen, J. Li, L. Ma, C. Wang, C. Liu, J. W. Schwank and J. Hao, Environ. Sci. Technol., 2013, 47, 5294–5301.
18. C. Liu, H. Xian, Z. Jiang, L. Wang, J. Zhang, L. Zheng, Y. Tan and X. Li, Appl. Catal., B, 2015, 176, 542–552.
19. C. He, B. Shen, J. Chen and J. Cai, Environ. Sci. Technol., 2014, 48, 7891–7898.
20. S. Park, S. An, Y. Mun and C. Lee, ACS Appl. Mater. Interfaces, 2013, 5, 4285–4292.
21. J. Yan, E. Khoo, A. Sumboja and P. S. Lee, ACS Nano, 2010, 4, 4247–4255.
22. Y. Gao, Z. Zhang, J. Wu, L. Duan, A. Umar, L. Sun, Z. Guo and Q. Wang, Environ. Sci. Technol., 2013, 47, 10813–10823.
23. T. Meng, M. Ara, L. Wang, R. Naik and K. S. Ng, J. Mater. Sci., 2014, 49, 4058–4066.
24. L. Abello, B. Bochu, A. Gaskov, S. Koudryavtseva, G. Lucazeau and M. Roumyantseva, J. Solid State Chem., 1998, 135, 78–85.
25. F. Buciuman, F. Patcas and T. Hahn, Chem. Eng. Process., 1999, 38, 563–569.
26. Z.-Y. Pu, J.-Q. Lu, M.-F. Luo and Y.-L. Xie, J. Phys. Chem. C, 2007, 111, 18695–18702.
27. J. Nie and H. Liu, J. Catal., 2014, 316, 57–66.
28. H. Xu, Z. Qu, S. Zhao, J. Mei, F. Quan and N. Yan, J. Hazard. Mater., 2015, 299, 86–93.
29. S. Yang, Y. Guo, N. Yan, D. Wu, H. He, Z. Qu and J. Jia, Ind. Eng. Chem. Res., 2011, 50, 9650–9656.
30. H. Zhu, P. Zhang and S. Dai, ACS Catal., 2015, 5(11), 6370–6385.
31. J. T. Mefford, W. G. Hardin, S. Dai, K. P. Johnston and K. J. Stevenson, Nat. Mater., 2014, 13, 726–732.

---

Footnote

† Electronic supplementary information (ESI) available: The detailed information about materials characterization can be found. See DOI: 10.1039/c6ra10006f

---