Removal of gas-phase Hg⁰ by Mn/montmorillonite K 10

Abstract

Mn/montmorillonite K 10 (Mn/MK10) prepared by impregnation method was studied to remove Hg⁰ in simulated coal-fired flue gas. The samples were characterized by Brunauer–Emmett–Teller (BET), X-ray diffraction (XRD), scanning electron microscopy (SEM), X-ray photoelectron spectroscopy (XPS) and temperature-programmed desorption (TPD). Effects of manganese loading value, reaction temperature and flue gas components on Hg⁰ removal efficiency were investigated. The results indicated that 4% Mn/MK10 was the optimal sample with outstanding Hg⁰ removal efficiency over the temperature range of 100–400 °C. The characteristic analysis demonstrated that amorphous MnO₂ and active oxygen were crucial for Hg⁰ removal. Besides, NO had a promoting effect due to the formation of Hg(NO₃)₂. The addition of only 5 ppm HCl led to excellent Hg⁰ removal performance as HCl enhanced Hg⁰ conversion to HgCl_x. The inhibition effect of SO₂ could be counteracted in the presence of NO and/or HCl. The Hg⁰ removal capacity showed a relative decrease when H₂O (g) was added to simulated flue gas. Moreover, Hg⁰ removal performance was maintained at 80–99% (NO/SO₂ = 0.26–1.71) in simulated flue gas without HCl, which appeared to be promising in industrial application.

---

1. Introduction

Mercury emissions have become a major environmental issue that attracts considerable attention owing to its extreme toxicity, volatility, persistence, and bioaccumulation.¹ As coal is a predominant energy source in China, coal combustion is a primary source of emitted mercury. There are three major forms of mercury emitted during coal combustion: oxidized mercury (Hg²⁺), particle-bound mercury (Hg_P) and elemental mercury (Hg⁰).² Hg²⁺ is water-soluble, enabling its removal through wet flue gas desulfurization devices (WFGD). Hg_p can be easily captured by electrostatic precipitators or fabric filters. However, Hg⁰ is much harder to remove due to its low water solubility and high vapor pressure.³ Thus, the control of Hg⁰ in coal-fired flue gas is particularly difficult and challenging.⁴

Numerous technologies have been developed for removing Hg⁰, and one of the most effective is activated carbon injection. However, some shortcomings limit its development, such as high cost, poor capacity, low temperature, high carbon-to-mercury ratios, and slow regeneration.^5,6 Transition metal oxides, proposed as one of the most promising alternatives to activated carbon sorbents, have been extensively examined in recent years.^7–17 In particular, manganese oxide (MnO_x) has been shown to have a good mercury capture potential through effective catalytic oxidation.^8,9 Manganese, which has multiple oxidation states, offers the ability for Hg⁰ oxidation to be achieved. According to the literature, Mn/MgO is effective for Hg⁰ removal at low temperatures, with amorphous MnO₂ and chemisorbed O₂ playing a crucial role in Hg⁰ oxidation.¹⁰ Xu et al.¹¹ found that MnO_x/graphene exhibits excellent Hg⁰ removal efficiency at 150 °C. Not only is MnO_x considered the main active site, but also it is often an additive for producing a synergistic effect for Hg⁰ removal. In Li's study,¹² Mn–Ce/Ti catalysts were highly active for Hg⁰ removal because of the coexistence of Mn⁴⁺ and Ce³⁺. Consequently, most previous studies have shown the positive effects of manganese for Hg⁰ removal.

Materials, like TiO₂ and Al₂O₃, were often employed as effective carries for Hg⁰ removal.^13–17 However, previous studies revealed some significant drawbacks for Hg⁰ removal over these catalysts, such as the inefficiency at high temperature,^12,13 the low resistance to SO₂ poisoning^14,15 and the dependence on high HCl concentrations.¹⁶ In recent years, scientists have shown much interest in minerals, such as magnetite,¹⁸ zeolite,^19,20 and clay,^21–23 principally to take advantage of their natural, abundant, and unique structural features. Montmorillonite, a type of natural Si–Al-based mineral, is widely used in adsorption and catalysis fields such as the removal of VOCs,^24,25 CO₂ ^26,27 and SO₂.^28,29 Up to now, however, it has hardly been used in gas-phase Hg⁰ removal.

As an effective adsorbent and catalyst, the montmorillonite owns several advantages in Hg⁰ removal. Firstly, its layer-structure and abundance of porosity are good for Hg⁰ adsorption.³⁰ Secondly, the good thermal stability of montmorillonite makes it possible to remove Hg⁰ at high temperature.³¹ Thirdly, its high surface acidity benefits the catalytic activity, especially for heavy metal removal.³² Finally, its cost is only 1/10–1/20 of that of activated carbon.³³ Therefore, it is interesting and promising to combine the advantages of Mn and montmorillonite for synthesizing a novel sample to effectively remove Hg⁰.

In this work, we synthesized a novel sample based on manganese oxide impregnated on montmorillonite K 10 (MK10: montmorillonite prepared by acid activation), and the Hg⁰ removal performance in simulated flue gas was investigated over the temperature range of 100–400 °C. Brunauer–Emmett–Teller (BET), X-ray diffraction (XRD), scanning electron microscopy (SEM), X-ray photoelectron spectroscopy (XPS), and temperature-programmed desorption (TPD) analysis were used to characterize the samples. The effects of flue gas components on Hg⁰ removal were examined, especially SO₂ resistance and the effect of low/zero HCl concentration. Further, a reaction mechanism of Hg⁰ removal over the prepared samples was proposed. This work also provides information on the industrial application of Mn/MK10 samples.

2. Experimental

2.1. Preparation of samples

The Mn/MK10 samples were prepared by an impregnation method. In a typical procedure, requisite quantities of MK10 and Mn(NO₃)₂ precursors were mixed in deionized water and stirred for 4 h at 25 °C. Afterwards, a rotary evaporator at 55 °C was used to evaporate excess water from the mixtures, and the obtained solid was dried at 110 °C and calcined at 500 °C for 4 h. The derived samples were crushed and sieved to 60–80 mesh (180–250 μm) particles for the sample activity tests. The samples are denoted as x% Mn/MK10, where x represents the weight ratio of Mn to MK10.

2.2. Characterization of samples

Nitrogen adsorption–desorption isotherms were recorded with a nitrogen adsorption apparatus (Autosorb-iQ, Quantachrome, USA). The specific surface areas and pore volumes were calculated using the BET and Barrett–Joyner–Halenda (BJH) methods, respectively. XRD patterns were determined using a Rigaku D/Max-RA powder diffractometer with a Ni-filtered Cu Kα radiation source (40 kV and 150 mA) and a scanning range from 4° to 70°. The physical morphology of the samples was observed using SEM (JSM 6700F, JEOF) with a field emission gun. XPS was carried out on an ESCALAB 250Xi employing Al Kα radiation (hν = 1486.6 eV) as the excitation source. The binding energies were corrected using the binding energy of adventitious carbon (284.8 eV).

Before the Hg temperature-programmed desorption (Hg-TPD) tests, about 50 mg samples were first treated with a gas flow of Hg⁰ balanced in N₂ for 3 h at 150 °C and then purged with N₂ at a flow rate of 500 mL min⁻¹ until the outlet of the Hg⁰ concentration decreasing to zero. The sample was heated from 100 °C to 700 °C at a heating rate of 10 °C min⁻¹ in pure N₂.

2.3. Sample activity test

The sample activity test was performed in a fixed-bed quartz tube reactor (4 mm i.d.) containing 0.05 g of sample. A gas comprising 120 μg m⁻³ Hg⁰ with a balance of N₂ was generated using an Hg⁰ permeation tube. The simulated flue gas comprising 0–5% O₂, 0–500 ppm NO, 0–1000 ppm SO₂, 0–5 ppm HCl, 0–5% H₂O and balanced N₂ was introduced into the reactor at the steady state. All lines that Hg⁰ passed through were heated to 90 °C to prevent mercury deposition on the inner surface. The Hg⁰ concentration was continuously monitored using a Lumex RA915M Zeeman mercury analyzer. In all tests, the total gas flow rate was 500 mL min⁻¹, the space velocity was 478 000 h⁻¹, and the steady reaction time was more than 3 h. A mercury concentration mass balance was conducted before each test. The Hg⁰ removal efficiency (η) is defined aswhere Hg⁰_inlet (μg m⁻³) and Hg⁰_outlet (μg m⁻³) are the concentrations of Hg⁰ measured at the inlet and the outlet of the reactor, respectively.

3. Results and discussion

3.1. Activity test for Hg⁰ removal

A series of experiments were conducted to examine the effect of the Mn contents on Hg⁰ removal over virgin MK10 and Mn/MK10 at various temperatures. Fig. 1 reveals no apparent Hg⁰ removal efficiency is observed over pure MK10, while the efficiency of Hg⁰ removal is gradually enhanced to over 90% with increasing Mn loading and temperature. When the Mn content was below 4%, the Hg⁰ removal performance at low temperature (<200 °C) was relatively poor, but the efficiency significantly increased with increasing temperature. However, at high temperature (>200 °C), the performance was comparably high and stable. The 4% Mn/MK10 sample showed the highest Hg⁰ removal efficiency and stability across a wide temperature window from 100 to 400 °C. Unlike other Mn-loaded materials, the outstanding performance of 4% Mn/MK10 at high temperatures is partly a result of the thermal and chemical stability of MK10.³² Further loading with Mn did not obviously heighten the sample efficiency. Thus, all subsequent tests were conducted utilizing the 4% Mn/MK10 samples.

Fig. 1 Hg⁰ removal efficiency over various Mn/MK10 samples at different temperatures (reaction conditions: O₂ = 5%, balance N₂).

3.2. Characterization of samples

3.2.1. BET analysis. The textural properties, including specific surface areas and pore volumes, of virgin MK10 and the impregnated MK10 (x% Mn/MK10) samples are summarized in Table 1. Compared with virgin MK10, the MK10 samples impregnated with MnO_x possess larger BET surface areas and pore volumes, similar to results found in other literature.^34,35 For one thing, as shown in Fig. 2, the impregnation and calcination process generated some new pores and increased the number of small pores (pore size < 4 nm) due to the powerful oxidizing property of Mn. For another, MnO_x was relatively well dispersed on the surface of the samples.³⁵ With an increase of the Mn content from 1% to 8%, the BET surface area first increased from 69.90 to 88.39 m² g⁻¹ and then decreased to 81.16 m² g⁻¹. This tendency was likely due to the formation and aggregation of MnO_x crystal particles on the sample surface, which caused pore blockages at high Mn loadings. Moreover, this phenomenon partly accounts for the trend observed in Fig. 1, in which the Hg⁰ removal efficiency did not increase further at high Mn loadings.

Table 1 Textural properties of various Mn/MK10 samples

Sample	SBET (m^2 g^−1)	Vtotal (cm^3 g^−1)
Virgin MK10	64.16	0.108
1% Mn/MK10	69.93	0.115
2% Mn/MK10	76.59	0.122
3% Mn/MK10	80.47	0.130
4% Mn/MK10	88.39	0.141
6% Mn/MK10	86.40	0.139
8% Mn/MK10	81.16	0.132

---

Fig. 2 Pore size distribution of MK10 and 4% Mn/MK10.

3.2.2. XRD analysis. To identify the structure of Mn species on the sample surface, XRD measurements were obtained for virgin MK10 and various Mn-loaded MK10 samples. The XRD patterns in Fig. 3 show characteristics peaks of MK10 at 2θ values of 5.9°, 19.8°, and 62.3° for all of the samples. However, the intensity of these peaks gradually weakened with increasing MnO_x content, indicating that MnO_x strongly interacted with MK10 in these samples. No obvious characteristic peaks of MnO_x were observed when Mn loading was below 6%, which indicated that MnO_x might be highly dispersed or present as an amorphous phase on sample surface, which can promote the catalytic oxidation activity efficiently.³⁶ However, on further increasing the Mn content, peaks were observed at the 2θ values of 28.7°, 37.4°, and 56.8°, corresponding to MnO₂. This result confirmed the BET results, in which too high a loading led to the formation and aggregation of crystalline phases. In addition, no Mn₂O₃ or MnO characteristic peaks were detected in any of these samples, signifying that no or little Mn³⁺ and Mn²⁺ were generated during the impregnation process.

Fig. 3 XRD pattern of various Mn/MK1O samples.

3.2.3. SEM and EDS analysis. SEM was used to analyze the alteration of the sample prior to and after impregnation. As shown in Fig. 4a (MK10) and 4c (4% Mn/MK10), the differences in the shape and size of the whole structure under low magnification (×2000) were very minute, illustrating that the addition of 4% Mn during the synthetic process did not substantially change the particle size. Moreover, the increase in the amount of small pores, as shown in Fig. 2, could not be observed under low magnification. Notably, under high magnification (×40 000), the “flower-like” smooth layer structure clearly observed in Fig. 4b (MK10) was destroyed after Mn impregnation, as seen in Fig. 4d (4% Mn/MK10), which could indicate a reaction between MK10 and the active components as well as MnO_x is likely highly dispersed on the surface of 4% Mn/MK10, consistent with the XRD results. Similarly, the EDS spectrum shown in Fig. 4e confirmed that the pure MK10 mainly contained elements of Si, Al, O. And Fig. 4f clearly shows the existence of Mn on the 4% Mn/MK10, which demonstrates the successful impregnating manganese oxide into MK10.

Fig. 4 SEM images (low and high magnification) and EDS patterns of MK10 (a, b, e) and 4% Mn/MK10 (c, d, f).

3.2.4. XPS analysis. To determine the chemical state and the relative proportion of the main elements on the surface of the samples, XPS was used to investigate the 4% Mn/MK10 sample before and after Hg⁰ adsorption under pure N₂, as seen in Fig. 5 and Table 2.

Fig. 5 O1s and Mn2p XPS spectra of fresh and spent 4% Mn/MK10 samples.

Table 2 The XPS data of O1s and Mn2p species in 4% Mn/MK10 samples

Sample	Oα (%)	Oβ (%)	Oγ (%)	Mn^4+ (%)	Mn^3+ (%)
Fresh 4% Mn/MK10	42.24	55.20	2.56	89.97	10.03
Spent 4% Mn/MK10	41.36	44.07	14.57	58.57	41.43

---

The O1s XPS spectra revealed three types of oxygen on the 4% Mn/MK10 sample, with characteristic peaks at 530.1–530.3 eV, 531.1–531.3 eV, and 532.3–532.7 eV ascribed to lattice oxygen (denoted as O_α), chemisorbed oxygen/OH groups (denoted as O_β), and molecular water (denoted as O_γ), respectively.^37,38 Clearly, O_α and O_β were the primary oxygen species in the fresh sample. However, both of these species were diminished after Hg⁰ absorption, while the amouqf O_γ increased from 2.56% to 14.57%, which suggested that both lattice oxygen and chemisorbed oxygen or OH groups participated in the Hg⁰ oxidation process. The reduction of O_β was much more obvious than that of O_α (Table 2); chemisorbed oxygen has a highly active and important role in oxidation reactions because it has higher mobility than lattice oxygen.³⁹ Hence, the consumption of more chemisorbed oxygen than lattice oxygen after Hg⁰ adsorption was reasonable.

Similar to other Mn-containing samples,^38,40 the Mn2p region of the samples in this work included Mn2p_3/2 with a binding energy of 641–644 eV and Mn2p_1/2 with a binding energy of 654.4 eV. The Mn2p_3/2 peak could be further fitted into two peaks around 643.0 eV and 641.7 eV, corresponding to Mn⁴⁺ and Mn³⁺, respectively. For the fresh sample, Mn⁴⁺ was the dominant Mn valence state, accounting for 89.97% (Table 2), which demonstrated that MnO₂ is the main active species, consistent with the XRD results. However, the amount of Mn³⁺ increased from 10.03% to 41.43% after Hg⁰ adsorption, indicating that Mn⁴⁺ took part in Hg⁰ oxidation and was converted to Mn³⁺.⁴¹

3.3. Effect of individual gas components

To better understand Hg⁰ removal capacity over 4% Mn/MK10, it is necessary to explore the roles of individual flue gas components on Hg⁰ removal efficiencies, especially NO, HCl and SO₂. Experiments were conducted by mixing Hg⁰ with individual flue gas components, with or without O₂, and the results are shown in Fig. 6. To explain the proposed reaction mechanism of each flue gas components, the TPD tests were carried out and the results are shown in Fig. 7.^42,43

Fig. 6 Hg⁰ removal efficiency of 4% Mn/MK10 over individual gas components at 150 °C (reaction conditions: O₂ = 0/5%, NO = 500 ppm, HCl = 5 ppm, SO₂ = 1000 ppm, balance).

Fig. 7 TPD profiles of 4% Mn/MK10 under N₂ with various gas components.

3.3.1. Effect of O₂. As seen in Fig. 6, the Hg⁰ removal efficiency on 4% Mn/MK10 was 36.3% under pure N₂ atmosphere, and the TPD profiles in Fig. 7a indicate the primary Hg⁰ desorption peaks appeared at approximately 210 °C was attributed to HgO.^44,45 Therefore, the loss of mercury could be explained by Mars–Maessen mechanism, where Hg⁰ bonded with lattice oxygen to form weakly bonded Hg–O–Mn–O_x−1 or reacted with surface oxygen (including chemisorbed oxygen and lattice oxygen) to form HgO directly. Apparently, surface oxygen played a key role in Hg⁰ conversion under pure N₂ flue gas. When 5% O₂ was added to the gas flow, the performance rapidly improved to 90.9%. Serving as the Hg⁰ oxidant, O₂ could regenerate the lattice oxygen and replenish the consumed chemisorbed oxygen, and hence improve Hg⁰ oxidation.

3.3.2. Effect of NO. To determine the effect of NO on Hg⁰ removal over Mn/MK10, 500 ppm NO was added into pure N₂ gas flow, the change of Hg⁰ removal efficiency is shown in Fig. 6. A prominent promotional effect was observed with the removal efficiency rapidly increased to 96.6%. This might partly be due to the addition of MnO₂ in the sample which can adsorb and oxidize NO.⁴⁶ Moreover, as shown in Fig. 7b, the TPD peaks appearing at around 213 °C and 435 °C indicated the production of Hg(NO₃)₂.⁴² It has been reported that,⁴⁷ NO would be weakly adsorbed on the surface of the metal oxide samples in the absence of O₂, and a fraction of it would react with the surface oxygen to generate NO₂, which are more active for Hg⁰ oxidation. Adding 5% O₂ resulted in 99.6% Hg⁰ removal efficiency, demonstrating that gas-phase O₂ regenerates and replenishes surface oxygen to facilitate the conversion of NO into NO₂ and Hg(NO₃)₂.

3.3.3. Effect of HCl. Chlorine, the major halogen specie in coals, is considered to be present chiefly as HCl in coal derived flus gases. As the main oxidized mercury specie in coal combustion flue gas is HgCl₂, HCl is regarded as the most important flue gas component affecting Hg⁰ removal. Therefore, the effect of HCl on Hg⁰ removal was investigated. In accordance with the low concentration of HCl in actual flue gas in China, only 5 ppm HCl was added to pure N₂ gas flow.⁴⁸ As shown in Fig. 6, the Hg⁰ removal efficiency increased to 98.9%, which is much higher than that observed under N₂ atmosphere. Lattice oxygen and chemisorbed oxygen supported the transformation of HCl to either Cl₂ or other chlorine species,^49,50 which are responsible for Hg⁰ oxidation in the presence of HCl. Additionally, the TPD results in Fig. 7c showed two desorption peaks at 220 °C and 370 °C. Because the Hg⁰ removal performance under HCl atmosphere is evidently superior to that of the N₂ atmosphere, these two peaks are most likely related to the decomposition of HgCl_x species.⁵¹ This is consistence with the result of other literatures that HCl could enhance Hg⁰ conversion to HgCl₂ over manganese based catalysts.⁵²

3.3.4. Effect of SO₂. Promotional,⁵³ negligible⁵⁴ or inhibitive^38,55 effects of SO₂ have all been reported. There was not consistent conclusion concerning the effect of SO₂ on Hg⁰ removal in flue gas over metal oxide samples. In this study, the results in Fig. 6 showed the efficiency decreased from 36.3% to 20.7% when 1000 ppm SO₂ was added into pure N₂, which revealed a suppressive effect of SO₂ on Hg⁰ removal. Two possible reasons are responsible for this result. On the one hand, SO₂ could compete with Hg⁰ for active sites on Mn/MK10. On the other hand, the adsorbed SO₂ could react with the surface oxygen to form SO₃ and thus consumed reactive oxygen which is beneficial for Hg⁰ removal.⁵³ If the deactivation of SO₂ was mainly attributed to the consumption of reactive oxygen, the addition of O₂ would apparently relieve the inhibitive effect. However, the presence of 5% O₂ only increased the removal efficiency from 20.7% to 29.1%, which still greatly inhibited Hg⁰ removal. As the TPD results shown in Fig. 7d, the slightly increased efficiency was probably due to the formation of Hg_xSO₄ in the presence of O₂.^42​ That is, the addition of O₂ only resulted in slight remission, with the removal efficiency still significantly suppressed compared with the efficiency under pure N₂. Therefore, it was not the consumption of surface oxygen but the competition of SO₂ and Hg⁰, the dominant factor in the inhibition by SO₂.

3.4. Effect of multiple gas components

3.4.1. Effect of SO₂ in the presence of NO and/or HCl. As discussed above, SO₂ had an inhibitive effect on Hg⁰ removal, regardless of the presence or absence of O₂. Thus, it is important to investigate whether the suppressive effect of SO₂ on Mn/MK10 sample can be offset in the presence of NO and/or HCl. As shown in Fig. 8, the efficiencies significantly increase from 29.1% to over 97% only if NO and/or HCl are added to the gas flow, which means the presence of NO and/or HCl can effectively counteract the inhibiting effect of SO₂. This phenomenon was probably due to the different active sites between SO₂ and NO/HCl, and Hg⁰ could be preferentially absorbed on NO/HCl. Thus, the promoting role of NO and/or HCl could offset the inhibitory effect of SO₂ on Hg⁰ removal. In general, NO always exists in actual coal-fired flue gases, and there is often a low concentration of HCl in flue gases in China. Therefore, it was reasonable and acceptable to conclude that the SO₂-poisoning phenomenon would be avoided for Mn/MK10 samples in simulated flue gases, even in the absence of HCl. Notably, under a typical-simulated flue gas including 5% O₂, 1000 ppm SO₂, 500 ppm NO, and 5 ppm HCl, the Hg⁰ removal performance of 4% Mn/MK10 was maintained at >97% for more than 140 h, which shows the excellent stability of 4% Mn/MK10 for Hg⁰ removal.

Fig. 8 Hg⁰ removal efficiency of 4% Mn/MK10 over multiple gas components at 150 °C (reaction conditions: O₂ = 5%, SO₂ = 1000 ppm, NO = 500 ppm, HCl = 5 ppm, H₂O = 5%, balance N₂).

3.4.2. Effect of H₂O in simulated flue gas. Water vapor, which is unavoidable in coal-fired flue gases, has been reported to suppress Hg⁰ oxidation and removal over metal oxide catalysts due to competitive adsorption.^35,55 Experiments were conducted to determine the effect of H₂O (g) on Hg⁰ removal over 4% Mn/MK10, and the results in Fig. 8 indicate a suppression of the removal efficiency from 99.1% to 78.6% on introduction of 5% H₂O (g) into the dry simulated flue gas. The effect is likely the result of the competition of water vapor on active sites inhibited the adsorption of reactive species that have promotional effect on Hg⁰ removal such as HCl and NO.⁵⁴ Additionally, H₂O (g) might react with SO₂ to form H₂SO₄ and further react with MnO₂ to form Mn(SO₄)₂,⁵⁶ which could cover the sample surface and prevent contact between Hg⁰ and the active components.

Nevertheless, it should be noted that only 0.05 g of sample (with the high GHSV of 487 000 h⁻¹) was used in these experiments to highlight the effect of H₂O (g) on Hg⁰ removal. According to the literature,⁵⁵ the suppressant effect of H₂O (g) should be mitigated by using a larger amount of sample. In this work, the inhibition by H₂O (g) was lessened significantly when 0.30 g of the 4% Mn/MK10 sample was used under the same wet flue gas conditions (Fig. 8). Thus, the industrial application of the Mn/MK10 sample is advantageous and promising for Hg⁰ removal in coal-fired flue gases.

3.5. Effect of NO/SO₂ in the absence of HCl

For coal-fired flue gas in china, it is more useful to investigate Hg⁰ removal efficiency of the samples without the assistance of HCl. Therefore, in this work, experiments were carried out to study the effect of SO₂ on Hg⁰ removal when adding various concentration of NO into the simulated flue gases. The effect of the NO/SO₂ ratio on Hg⁰ removal efficiency is shown in Fig. 9. In comparison with the efficiency of 29.1% under O₂ and SO₂, Hg⁰ removal performance was enhanced with increasing NO/SO₂ ratios. And when the NO/SO₂ ratio was beyond 0.33, the efficiency was maintained at 99%. There are three possible reasons accounting for this phenomenon. First, the promotion role of NO was much remarkable than the suppression effect of SO₂. In contrast to the situation under SO₂ and O₂, even at a low concentration, NO would react with Hg⁰ to promote oxidation, and when the concentration of NO exceeded a threshold value, the facilitation effect would play the dominant role. Second, intermediate products derived from the reaction between NO and SO₂ could be formed, which would consume some SO₂ to alleviate its inhibitory effect.⁵⁷ Thirdly, as discussed above, the active sites of SO₂ and NO might be different, and Hg⁰ could be preferentially absorbed on NO. Generally, the concentrations of SO₂ and NO_x in coal-fired flue gases are 340–1120 ppm and 290–580 ppm,⁵⁷ respectively, with actual NO/SO₂ volume ratios of 0.26–1.71, as shown in the shaded area in Fig. 9. The removal efficiency of Hg⁰ in this region is 80–99%. Thus, the efficiency of the 4% Mn/MK10 sample will not be notably affected in simulated coal-fired flue gases without the assistance of HCl and this sample has great potential for industrial applications.

Fig. 9 Hg⁰ removal efficiency of 4% Mn/MK10 over various NO/SO₂ ratios at 150 °C (reaction conditions: NO/SO₂ < 0.5, SO₂ = 1000 ppm, NO = 125–500 ppm; NO/SO₂ > 0.5, NO = 500 ppm, SO₂ = 250–1000 ppm. O₂ = 5%, balance N₂).

4. Conclusions

4% Mn/MK10 sample exhibited the highest activity (>90%, 100–400 °C) and stability (>140 h, 150 °C) for Hg⁰ removal in simulated coal-fired flue gas. Characterization of the sample indicated that amorphous MnO₂ and active oxygen were crucial for Hg⁰ removal. The effects of various gas components revealed that NO plays a promoting role in Hg⁰ removal and the addition of O₂ provided enough surface oxygen for formation of Hg(NO₃)₂. Moreover, as little as 5 ppm HCl prominently improved the Hg⁰ removal performance as HCl enhanced Hg⁰ conversion to HgCl_x. However, SO₂ had a strong inhibitory effect on Hg⁰ removal. The addition of O₂ only resulted in slight remission of the SO₂ inhibition effect, but NO and HCl effectively counteracted the observed suppression. H₂O also slightly inhibited Hg⁰ removal, mostly due to competitive adsorption. The Hg⁰ removal performance was maintained at 80–99% (NO/SO₂ = 0.26–1.71) in simulated flue gas without HCl, which appeared to be promising in industrial application.

Acknowledgements

This work was supported by the National Basic Research Program (973) of China (no. 2013CB430005) and the Special Research Funding for Public Benefit Industries from the National Ministry of Environmental Protection (201309018).

References

1. M. Brauer, M. Amann, R. T. Burnett, A. Cohen, F. Dentener, M. Ezzati, S. B. Henderson, M. Krzyzanowski, R. V. Martin, R. Van Dingenen, A. van Donkelaar and G. D. Thurston, Environ. Sci. Technol., 2011, 46, 652–660.
2. Y. Liu, C. Tian, B. Yan, Q. Lu, Y. Xie, J. Chen, R. Gupta, Z. Xu, S. M. Kuznicki, Q. Liu and H. Zeng, RSC Adv., 2015, 5, 15634–15640.
3. D. Jampaiah, S. J. Ippolito, Y. M. Sabri, B. M. Reddy and S. K. Bhargava, Catal. Sci. Technol., 2015, 5, 2913–2924.
4. Y. Yang, W. Xu, Y. Wu, J. Xiong, T. Zhu, X. Zhou and L. Tong, RSC Adv., 2016, 6, 59009–59015.
5. Y. Gao, Z. Zhang, J. Wu, L. Duan, A. Umar, L. Sun, Z. Guo and Q. Wang, Environ. Sci. Technol., 2013, 47, 10813–10823.
6. W. Xu, L. Tong, H. Qi, X. Zhou, J. Wang and T. Zhu, Ind. Eng. Chem. Res., 2014, 54, 146–152.
7. D. Jampaiah, K. M. Tur, S. J. Ippolito, Y. M. Sabri, J. Tardio, S. K. Bhargava and B. M. Reddy, RSC Adv., 2013, 3, 12963–12974.
8. Y. Shao, J. Li, H. Chang, Y. Peng and Y. Deng, Catal. Sci. Technol., 2015, 5, 3536–3544.
9. S. Cavallaro, N. Bertuccio, P. Antonucci, N. Giordano and J. C. J. Bart, J. Catal., 1982, 73, 337–348.
10. Y. Xu, Q. Zhong and X. Liu, J. Hazard. Mater., 2015, 283, 252–259.
11. H. Xu, Z. Qu, C. Zong, W. Huang, F. Quan and N. Yan, Environ. Sci. Technol., 2015, 49, 6823–6830.
12. H. Li, C. Y. Wu, Y. Li and J. Zhang, Appl. Catal., B, 2012, 111, 381–388.
13. S. Zhao, Z. Qu, N. Yan, Z. Li, W. Zhu, J. Xu and M. Li, RSC Adv., 2015, 5, 30841–30850.
14. S. Zhao, Z. Qu, N. Yan, Z. Li, H. Xu, J. Mei and F. Quan, Environ. Sci. Technol., 2015, 5, 2985–2993.
15. J. He, G. K. Reddy, S. W. Thiel, P. G. Smirniotis and N. G. Pinto, Energy Fuels, 2013, 27, 4832–4839.
16. W. Du, L. Yin, Y. Zhuo, Q. Xu, L. Zhang and C. Chen, Fuel Process. Technol., 2015, 131, 403–408.
17. X. Li, Z. Liu, J. Kim and J. Y. Lee, Appl. Catal., B, 2013, 132, 401–407.
18. S. Yang, Y. Guo, N. Yan, D. Wu, H. He, Z. Qu and J. Jia, Ind. Eng. Chem. Res., 2011, 50, 9650–9656.
19. Z. Wei, Y. Luo, B. Li, Z. Cheng, J. Wang and Q. Ye, Atmos. Pollut. Res., 2015, 6, 45–51.
20. C. H. Chiu, H. C. Hsi and H. P. Lin, Fuel Process. Technol., 2015, 126, 138–144.
21. D. Shi, Y. Lu, Z. Tang, F. Han, R. Chen and Q. Xu, Korean J. Chem. Eng., 2014, 31, 1405–1412.
22. C. He, B. Shen and F. Li, J. Hazard. Mater., 2016, 304, 10–17.
23. F. Ding, Y. Zhao, L. Mi, H. Li, Y. Li and J. Zhang, Ind. Eng. Chem. Res., 2012, 5, 3039–3047.
24. I. Fatimah and T. Huda, Appl. Clay Sci., 2013, 74, 115–120.
25. S. Zuo, F. Liu, R. Zhou and C. Qi, Catal. Commun., 2012, 22, 1–5.
26. M. Tahir and N. S. Amin, Appl. Catal., B, 2013, 142, 512–522.
27. L. Stevens, K. Williams, W. Y. Han, T. Drage, J. Wood and J. Wang, Chem. Eng. J., 2013, 215–216, 699–708.
28. D. Olszewska, Fuel Process. Technol., 2011, 92, 2412–2419.
29. C. Volzone and J. Ortiga, Appl. Clay Sci., 2009, 44, 251–254.
30. T. Munusamy, J. Ya-Ting and L. Jiunn-Fwu, RSC Adv., 2015, 5, 23957.
31. R. Zhu, Q. Chen, Q. Zhou, Y. Xi, J. Zhu and H. He, Appl. Clay Sci., 2016, 123, 239–258.
32. B. S. Kumar, A. Dhakshinamoorthy and K. Pitchumani, Catal. Sci. Technol., 2014, 4, 2378–2396.
33. J. Qiu, S. Dong, H. Wang, X. Cheng and Z. Du, RSC Adv., 2015, 5, 58284–58291.
34. L. Tian, C. Li, Q. Li, G. Zeng, Z. Gao, S. Li and X. Fan, Fuel, 2009, 88, 1687–1691.
35. J. Yang, Y. Zhao, J. Zhang and C. Zheng, Environ. Sci. Technol., 2014, 48, 14837–14843.
36. Z. Wu, B. Jiang and Y. Liu, Appl. Catal., B, 2008, 79, 347–355.
37. J. Xie, Z. Qu, N. Yan, S. Yang, W. Chen, L. Hu and W. H. P. Liu, J. Hazard. Mater., 2013, 261, 206–213.
38. Y. Xie, C. Li, L. Zhao, J. Zhang, G. Zeng, X. Zhang, W. Zhang and S. Tao, Appl. Surf. Sci., 2015, 333, 59–67.
39. H. Li, C. Y. Wu, Y. Li and J. Zhang, Environ. Sci. Technol., 2011, 45, 7394–7400.
40. Y. Guo, N. Yan, S. Yang, P. Liu, J. Wang, Z. Qu and J. Jia, J. Hazard. Mater., 2012, 213, 62–70.
41. S. Cimino and F. Scala, Ind. Eng. Chem. Res., 2016, 55, 5133–5138.
42. M. Rumayor, M. Diaz-Somoano, M. A. Lopez-Anton and M. R. Martinez Tarazona, Chemosphere, 2015, 119, 459–465.
43. A. Murakami, M. A. Uddin, R. Ochiai, E. Sasaoka and S. Wu, Energy Fuels, 2010, 24, 4241–4249.
44. S. Wu, R. Katayama, M. A. Uddin, E. Sasaoka and Z. Xie, Energy Fuels, 2015, 29, 6598–6604.
45. R. Liu, W. Xu, L. Tong and T. Zhu, J. Environ. Sci., 2015, 36, 76–83.
46. L. Zhao, C. Li, J. Zhang, X. Zhang, F. Zhan, J. Ma, Y. Xie and G. Zeng, Fuel, 2015, 153, 361–369.
47. G. Q. And and R. T. Yang, J. Phys. Chem. B, 2004, 108, 15738–15747.
48. W. Xu, H. Wang, X. Zhou and T. Zhu, Chem. Eng. J., 2014, 243, 380–385.
49. L. Tong, W. Q. Xu, X. Zhou, R. H. Liu and T. Y. Zhu, Energy Fuels, 2015, 29, 5231–5236.
50. Y. Zhuang, J. Laumb, R. Liggett, M. Holmes and J. Pavlish, Fuel Process. Technol., 2007, 88, 929–934.
51. J. Yang, Y. Zhao, J. Zhang and C. Zheng, Fuel, 2015, 167, 366–374.
52. P. Wang, S. Su, J. Xiang, H. You, F. Cao, L. Sun, S. Hu and Y. Zhang, Chemosphere, 2014, 101, 49–54.
53. S. Eswaran and H. G. Stenger, Energy Fuels, 2005, 19, 2328–2334.
54. Y. Li, P. D. Murphy, C. Y. Wu, K. W. Powers and J. C. Bonzongo, Environ. Sci. Technol., 2008, 42, 5304–5309.
55. H. Li, C. Y. Wu, Y. Li, L. Li, Y. Zhao and J. Zhang, J. Hazard. Mater., 2012, 243, 117–123.
56. S. Tao, C. Li, X. Fan, G. Zeng, P. Lu, X. Zhang, Q. Wen, W. Zhao, D. Luo and C. Fan, Chem. Eng. J., 2012, 210, 547–556.
57. Y. Guo, Y. Li, T. Zhu and M. Ye, Energy Fuels, 2012, 27, 360–366.

---