当前位置:首页 >> 能源/化工 >>

Gaseous Elemental Mercury Capture from Flue Gas Using Magnetic


ARTICLE pubs.acs.org/est

Gaseous Elemental Mercury Capture from Flue Gas Using Magnetic Nanosized (Fe3-xMnx)1-δO4
Shijian Yang,? Naiqiang Yan,?,* Yongfu Guo,? Daqing Wu,? Hong

ping He,? Zan Qu,? Jianfeng Li,? Qin Zhou,§ and Jingping Jia ?
?

School of Environmental Science and Engineering, Shanghai Jiao Tong University, 800 Dong Chuan Road, Shanghai, 200240 P. R. China ? Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, 511 Kehua Street, Wushan, Tianhe District, Guangzhou, 510640 P. R. China § State Key Laboratory of Environmental Aquatic Chemistry, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, 18 Shuangqing Road, Haidian District, Beijing, 100085 P. R. China
S b Supporting Information

ABSTRACT: A series of nanosized (Fe3-xMnx)1-δO4 (x = 0, 0.2, 0.5, and 0.8) were synthesized for elemental mercury capture from the ?ue gas. Cation vacancies on (Fe3-xMnx)1-δO4 can provide the active sites for elemental mercury adsorption, and Mn4? cations on (Fe3-xMnx)1-δO4 may be the oxidizing agents for elemental mercury oxidization. With the increase of Mn content in the spinel structure, the percents of Mn4? cations and cation vacancies on the surface increased. As a result, elemental mercury capture by (Fe3-xMnx)1-δO4 was obviously promoted with the increase of Mn content. (Fe2.2Mn0.8)1-δO4 showed an excellent capacity for elemental mercury capture (>1.5 mg g-1 at 100-300 °C) in the presence of SO2 and HCl. Furthermore, (Fe2.2Mn0.8)1-δO4 with the saturation magnetization of 45.6 emu g-1 can be separated from the ?y ash using magnetic separation, leaving the ?y ash essentially free of sorbent and adsorbed Hg. Therefore, nanosized (Fe2.2Mn0.8)1-δO4 may be a promising sorbent for the control of elemental mercury emission.

’ INTRODUCTION Mercury is a major pollutant because of its toxicity, mobility, and bioaccumulation in the ecosystem and food chain. The emission of mercury from anthropogenic activities is a serious concern in both the developed and developing countries. Coal-?red utility boilers are currently the largest single-known source of anthropogenic mercury emissions. Mercury exists in three forms in the coal-derived ?ue gas: elemental mercury (Hg0), oxidized mercury (Hg2?), and particle-bound mercury (Hgp).1 Because elemental mercury is di?cult to be removed by currently available pollution control devices, it is the major mercury species emitted in the ?ue gas. Many technologies have been investigated to capture elemental mercury from the ?ue gas. Sorbents/catalysts for elemental mercury capture studied to date mainly fall into one of three groups: carbon-based sorbents, selective catalytic reduction catalysts, and metals and metal oxides.1 Now, the mercury-sorbent materials are extremely restricted in the application for at least three reasons: sorbent recovery, removal of toxin from the industrial waste, and operation cost.2-5 The separation of sorbent from the ?y ash can be solved by the magnetic property of sorbent materials.2-5 A magnetic sorbent MagZ-Ag0 has been investigated for elemental mercury capture,6,7 but lower cost sorbents would be more attractive. As is well-known, maghemite (γ-Fe2O3) is one of the cheapest magnetic materials. Furthermore, an interesting feature of γ-Fe2O3 is the possibility of replacing Fe3? cations by other metal cations while maintaining the spinel structure. Its physicochemical property is strongly dependent on the nature, amount, and site of metal incorporated
r 2011 American Chemical Society

into the spinel structure. Our previous research demonstrated that Ti4? in γ-Fe2O3 can strongly improve its ability for elemental mercury capture, but the presence of a high concentration of SO2 resulted in a severe interference.3 Herein, Mn4? cations were incorporated into γ-Fe2O3 to form (Fe3-xMnx)1-δO4 using a coprecipitation method. Then, (Fe3xMnx)1-δO4 was characterized using X-ray di?raction (XRD), H2 temperature programmed reduction (TPR), N2 adsorption/desorption isotherm, X-ray photoelectron spectroscopy (XPS), and magnetization measurement. At last, a packed-bed reactor system was used to estimate the performance of (Fe3-xMnx)1-δO4 for elemental mercury capture.

’ EXPERIMENTAL SECTION
Samples Preparation. Nanosized Fe3-xMnxO4, the precursor of (Fe3-xMnx)1-δO4, was prepared using a coprecipitation method:8 (1) Suitable amounts of ferrous sulfate, ferric chloride, and manganese sulfate were dissolved in distilled water (total cation concentration = 0.30 mol L-1). (2) The mixture was added to a sodium hydroxide solution (about 1.20 mol L-1), leading to an instantaneous
Received: May 11, 2010 Accepted: December 17, 2010 Revised: December 15, 2010 Published: January 5, 2011
1540
dx.doi.org/10.1021/es103391w | Environ. Sci. Technol. 2011, 45, 1540–1546

Environmental Science & Technology Table 1. Crystal Size, Lattice Parameter, and BET Surface Area of Synthesized (Fe3-xMnx)1-δO4
crystal (Fe3-xMnx)1-δO4 x=0 x = 0.2 x = 0.5 x = 0.8 size/nm 12 14 18 31 lattice parameter/nm 0.8326 0.8324 0.8332 0.8346 BET surface area/m2 g-1 101 82.9 69.4 37.8

ARTICLE

precipitation of manganese ferrite according to the following equation 2Fe3? ? ?1-x?Fe2? ? xMn2? ? 8OH - f Fe3-x Mnx O4 ? 4H2 O ?1?

During the reaction, the system was continuously stirred at 800 rpm. (3) The particles were then separated by centrifugation at 4500 rpm for 5 min and washed with distilled water followed by a new centrifugation. After 3 washings, the particles were collected and dried in a vacuum oven at 105 °C for 12 h. γ-Fe2O3 was obtained after the thermal treatment of Fe3O4 under air at 250 °C for 3 h. (Fe3-xMnx)1-δO4 (x = 0.2, 0.5, and 0.8) were obtained after the thermal treatment of Fe3-xMnxO4 (x = 0.2, 0.5, and 0.8) under air at 400 °C for 3 h. During the oxidization of Fe3-xMnxO4 to (Fe3-xMnx)1-δO4, some cation vacancies (0) were introduced to sustain the spinel structure. Samples Characterization. Powder XRD pattern was recorded on an X-ray diffractionmeter (Rigaku, D/max-2200/PC) between 10° and 80° at a step of 7° min-1 operating at 30 kV and 30 mA using Cu KR radiation. BET surface area was determined using a nitrogen adsorption apparatus (Micromeritics, ASAP 2010 M?C). The sample was outgassed at 200 °C before BET measurement. TPR profile was recorded on a chemisorption analyzer (Micromeritics, ChemiSorb 2720 TPx) with a gas flow of 20 cm3 min-1 (10% of hydrogen and 90% of nitrogen) at a rate of 10 °C min-1. Saturation magnetization was determined using a vibrating sample magnetometer (VSM, Model JDM-13) at room temperature. XPS (Thermo, ESCALAB 250) with Al KR (hv = 1486.6 eV) as the excitation source was used to determine the binding energies of Fe 2p, Mn 2p, S 2p, O 1s, and Hg 4f. The C 1s line at 284.6 eV was taken as a reference for the binding energy calibration. Elemental Mercury Capture. The assembly used for elemental mercury capture (shown in Figure S1 in the SI) was similar to that described in our previous research.2-5 A flow of air passed through the permeation tube and yielded a stable concentration of elemental mercury. A temperature control device was employed to keep the reactor at the desired temperatures. The gas containing elemental mercury first passed through the blank tube and then entered the CVAAS to determine the baseline. When the concentration of elemental mercury had fluctuated within (5% for more than 30 min, the gas was diverted to pass through the adsorbent bed. An exact amount of sorbent was inserted in the middle of the column reactor and then packed with quartz wool to support the sorbent layer and to avoid its loss. It was demonstrated that quartz wool has no ability for elemental mercury capture. To preliminarily estimate the performance for elemental mercury capture, (Fe3-xMnx)1-δO4 was ?rst tested under air. The

inlet gas contained a stable concentration of elemental mercury (shown in Table S1 in the SI) with a feed of 12 L h-1. The test time for (Fe3-xMnx)1-δO4 (x6?0) was about 10 h. Because the breakthrough ratios of γ-Fe2O3 for elemental mercury capture were more than 90% within 4 h, the test time for γ-Fe2O3 was about 4 h. For each test, the sorbent mass was about 25.0 mg (the gas space velocity was about 1.2 ? 106 h-1), and the reaction temperature varied from 100 to 300 °C. The e?ect of a high concentration of SO2 on elemental mercury capture was investigated. The inlet gas contained about 2.8 g Nm-3 (1000 ppmv) of SO2 and 10% of O2. Furthermore, the e?ect of HCl on elemental mercury capture was investigated. The inlet gas contained about 2.8 g Nm-3 of SO2, 8.1 mg Nm-3 (5 ppmv) of HCl, and 10% of O2. The concentration of elemental mercury in the gas was analyzed using a SG-921 CVAAS. Meanwhile, Hg2? in the gas at the exit of reactor was absorbed by 1.0 mol L-1 of KCl. Then, the amount of Hg2? in the KCl solution was determined using the CVAAS. Breakthrough curve was generated by plotting the concentration of elemental mercury in the gas at the exit of reactor.

’ RESULTS AND DISCUSSION
Characterization. The characteristic reflections of synthesized samples (shown in Figure S2 in the SI) corresponded very well to the standard card of maghemite (JCPDS: 39-1346). Additional reflections that would indicate the presence of other crystalline manganese oxides, such as Mn3O4, Mn2O3, or MnO2, were not present in the diffraction scan. Furthermore, the lattice parameter of synthesized Fe2.2Mn0.8O4 was 0.8456 nm (XRD pattern is not shown), which was much larger than that of magnetite (0.8396 nm). These indicate that Mn cations were incorporated into the spinel structure. Crystal sizes of synthetic samples were calculated with the Scherrer’s equation.9 As shown in Table 1, the crystal size increased with the increase of Mn content in (Fe3-xMnx)1-δO4. TPR pro?le of γ-Fe2O3 showed two obvious reduction peaks (shown in Figure 1A). The peak centered at about 370 °C corresponded to the reduction of γ-Fe2O3 to Fe3O4, and the broad peak at the higher temperature was attributed to the reduction of Fe3O4 to Fe0.10 TPR pro?les of (Fe3-xMnx)1-δO4 (x6?0) showed three groups of reduction peaks (shown in Figure 1A). The peaks centered at about 345-362 °C corresponded to the reduction of (Fe3-xMnx)1-δO4 to Fe3-xMnxO4, the peaks centered at about 520-598 °C were attributed to the reduction of Fe3-xMnxO4 to manganowustite (Fe1-yMnyO), and the last peaks were assigned to the reduction of Fe1-yMnyO to Fe0 and MnO.11 As Mn was introduced into the spinel structure, the ?rst peak shifted to a lower temperature. Meanwhile, the area of the ?rst peak decreased with the increase of Mn content in (Fe3-xMnx)1-δO4. It may be related to the decrease of the BET surface area (shown in Table 1). The reduction of (Fe3-xMnx)1-δO4 to Fe3-xMnxO4 involved the reduction of Mn4? to Mn3?, Mn3? to Mn2?, and partial Fe3? to Fe2?. As shown in Figure 1B, the H2 consumption at <300 °C corresponding to the reduction of Mn4? cations on the surface obviously increased with the increase of Mn content in (Fe3-xMnx)1-δO4. It indicates that the amount of Mn4? cations on (Fe3-xMnx)1-δO4 increased with the increase of Mn content. A key feature of the novel sorbent is its magnetic property, which makes it possible to separate the sorbent from the ?y ash mixture. The saturation magnetizations of (Fe3-xMnx)1-δO4 (x = 0, 0.2, 0.5, and 0.8) were 59.0, 48.6, 45.4, and 45.6 emu g-1,
1541
dx.doi.org/10.1021/es103391w |Environ. Sci. Technol. 2011, 45, 1540–1546

Environmental Science & Technology

ARTICLE

Figure 1. (A) TPR pro?les of synthesized (Fe3-xMnx)1-δO4 and (B) expansion of TPR pro?les of synthesized (Fe3-xMnx)1-δO4 from 150 to 400 °C: (a), x = 0; (b), x = 0.2; (c), x = 0.5; (d), x = 0.8.

for the transition metal oxides. Another oxygen species at about 531.3 eV was also observed, which was assigned to -OH. Elemental Mercury Capture under Air. The determination of oxidized mercury concentration at the exit of reactor showed that there was little oxidized mercury in the gas after passing through the reactor tube with (Fe3-xMnx)1-δO4. It indicates that the reduced amount of elemental mercury in the breakthrough curve (shown in Figure S3a-c in the SI) was captured by the magnetic sorbent. The mass of elemental mercury captured per unit mass of sorbent (capacity) can be calculated from the breakthrough curve. As shown in Table 3, the capacity of (Fe3-xMnx)1-δO4 for elemental mercury capture generally increased with the increase of Mn content. With the continuous increase of reaction temperature from 100 to 300 °C, the capacities of (Fe3-xMnx)1-δO4 for elemental mercury capture showed the same variation tendency, and the optimal reaction temperatures all centered at about 250 °C. Elemental mercury capture by metal oxides in the absence of halogen is generally attributed to the Mars-Maessen mechanism.1,12 The mechanism for elemental mercury capture by (Fe3-xMnx)1-δO4 was studied using XPS analysis. In comparison with fresh (Fe2.2Mn0.8)1-δO4 (shown in Figure 3a-c), no obvious changes happened in the XPS spectra over the spectral regions of Fe 2p and O 1s (shown in Figure 3d and e). As shown in Figure 3f, the component centered at about 640.4 eV corresponding to Mn2? cations did not appear. Meanwhile, the ratio of Mn4? cation to Mn3? cation decreased from 1.57 to 1.37 after elemental mercury capture. They suggest that some Mn4? cations were reduced to Mn3? cations during elemental mercury capture. Taking account of the binding energy of Hg 4f7/2 at 100.1 eV and the absence of Hg 4f 5/2 at about 105 eV corresponding to Hg2? (shown in Figure 3g), the oxidized mercury formed may be mercurous oxide. Mercurous oxide has been previously observed on (Fe2Ti)0.8O4 and (Fe2Ti0.8Mn0.2)1-δO4 in our previous research.3,4 Therefore, the mechanism of elemental mercury capture by (Fe3-xMnx)1-δO4 can be described as Hg0 ? 0 f 0 - Hg0 ?g? ?ad? 0 - Hg0 ?  MnIV f  MnIII HgI ?2? ?3?

Figure 2. Magnetization characteristics of synthesized (Fe3-xMnx)1-δO4: (a) x = 0; (b) x = 0.2; (c) x = 0.5; (d) x = 0.8.

respectively. (Fe3-xMnx)1-δO4 showed the superparamagnetism with a minimized coercivity and a negligible magnetization hysteresis (shown in Figure 2). The magnetization characteristic ensures that the magnetic sorbents cannot be permanently magnetized after being exposed to an external magnetic ?eld. Therefore, the sorbent particles can be redispersed without aggregation when the magnetic ?eld is removed.6 The peaks of Fe species on (Fe2.2Mn0.8)1-δO4 were assigned to oxidized Fe species, more likely Fe3? type species. The binding energies centered at about 709.9 and 711.1 eV may be assigned to Fe3?cations in the spinel structure, and the binding energy centered at about 712.3 eV may be ascribed to FeIII-OH (shown in Figure 3a). The Mn peaks at 641.1 and 642.4 eV were assigned to Mn3? and Mn4?, respectively (shown in Figure 3b). As shown in Table 2, the percent of Mn4? on (Fe3-xMnx)1-δO4 obviously increased with the increase of Mn content. The O peak mainly centered at about 529.9 eV (shown in Figure 3c), as expected

Reaction 2 was the collision of elemental mercury with the surface, resulting in a physical adsorption on the cation vacancies. Cation vacancies on the surface are typical Lewis acid sites.2-5 Gaseous elemental mercury is a Lewis base because it can be an electron-pair donor. The term Lewis base is more general and refers to the propensity to complex with a Lewis acid. If the concentration of elemental mercury in gas phase was su?ciently high for the surface to be saturated with physically adsorbed elemental mercury, the concentration of physically adsorbed elemental mercury on the surface ([0-Hg0]) can be described as ?0 - Hg0 ? ? k1 ?0? ?4?

where [0] and k1 were the percent of cation vacancies on the surface and the constant, respectively. Reaction 2 was an exothermic reaction, so k1 would rapidly decrease with the increase of reaction temperature. Reaction 3 was the oxidation of physically adsorbed elemental mercury to a Mn-Hg bimetal oxide by Mn4? cations on the surface.12 As is well-known, mercury is a heavy metal, and its atomic radius (1.76 ?) is much bigger than the radiuses of Mn4? (0.60 ?), Mn3? (0.66 ?), Fe3? (0.64 ?), and O2- (1.32 ?). When a mercury atom is
1542
dx.doi.org/10.1021/es103391w |Environ. Sci. Technol. 2011, 45, 1540–1546

Environmental Science & Technology

ARTICLE

Figure 3. XPS spectra of (Fe2.2Mn0.8)1-δO4 over the spectral regions of Fe 2p, Mn 2p, O 1s, Hg 4f, and S 2p.
1543
dx.doi.org/10.1021/es103391w |Environ. Sci. Technol. 2011, 45, 1540–1546

Environmental Science & Technology Table 2. Data of Atomic Ratios on the Surface Collected from XPS Analysis/%
BET[0] (Fe3-xMnx)1-δO4 x = 0.2 x = 0.5 x = 0.8 0 Mn Mn2? Mn3? Mn4? [MnIV]0/m2 g-1 1.9 1.6 7.6 7.2 1.6 6.0 11.3 0.062 0.27 0.35

ARTICLE

Table 3. Capacity of (Fe3-xMnx)1-δO4 for Elemental Mercury Capture mg g-1
(Fe3-xMnx)1-δO4 x=0 x = 0.2 x = 0.5 x = 0.8 x = 0.8 with SO2 x = 0.8 with SO2 and HCl 100 °C 150 °C 200 °C 250 °C 300 °C <0.20 1.92 2.90 2.86 2.38 2.01 <0.20 1.80 2.92 3.20 1.92 1.92 0.26 1.60 2.42 4.44 1.72 2.07 0.44 2.20 4.26 5.10 2.48 1.54 0.34 0.84 1.74 1.04 0.96 2.21

4.7 5.1 6.4 13.6 8.1 18.5

physically adsorbed on the cation vacancy, several ions including Mn4?, Mn3?, Fe3?, and O2- around the cation vacancy may be covered. Once the adsorbed elemental mercury contacts Mn4? cation on the surface, the adsorbed elemental mercury will be oxidized. The array of cation vacancies, Mn3?/Mn4?, Fe3?, and O2- in/ on (Fe3-xMnx)1-δO4 was well-proportioned even at the atomic scale due to the incorporation of Mn cations into the spinel structure. The near two Mn cations on (Fe3-xMnx)1-δO4 were spaced at least by two Fe cations and four oxygen anions, so the distance between two Mn4? cations was much more than the diameter of the Hg atom. When a mercury atom was physically adsorbed on the active site (i.e., 0), at most one Mn4? cation can be covered. Therefore, reaction 3 happened. As is well-known, Hg2O is not stable and it can self-decompose to HgO and Hg at a high temperature. Because the oxidized mercury formed on (Fe3-xMnx)1-δO4 was isolated MnIIIHgIO2, the near two oxidized mercury formed were spaced at least by two Fe cations and four oxygen anions. As a result, two mercurous cations cannot collide to transform to one Hg atom and one Hg2? cation. Therefore, the formed mercurous oxide on (Fe3-xMnx)1-δO4 was stable (shown in Figure S3). The kinetic equation of reaction 3 can be described as d? MnIV ? d?0 - Hg0 ? d? MnIII HgI ? ? ? dt dt dt ? k? MnIV ??0 - Hg0 ? ?5?

where [MnIV], [MnIIIHgI], and k were the percent of Mn4? cation, the percent of the bimetal oxide on the surface, and the kinetic constant, respectively. Because reaction 3 was promoted with the increase of reaction temperature, k would increase with the increase of reaction temperature. According to eq 5, [MnIV] may be approximately described as ? MnIV ? ? ? MnIV ?0 exp? - k?0 - Hg0 ?t? ? ? MnIV ?0 exp? - kk1 ?0?t? Then, d? MnIII HgI ? dt ? kk1 ?0?? MnIV ?0 exp? - kk1 ?0?t? ? MnIII HgI ? Z t exp? - kk1 ?0?t?dt ? kk1 ?0?? MnIV ?0
0

?6?

?7?

where Q was the amount of elemental mercury captured, which can be described as the product of [MnIIIHgI] and BET surface area. As shown in eq 9, Q should be approximately proportional to the product of BET, [0], and [MnIV]0. There generally was a positive correlation between BET[0][MnIV]0 and the capacity of (Fe3-xMnx)1-δO4 for elemental mercury capture (shown in Tables 2 and 3), so the increase of Mn4? cations and cation vacancies on the surface may mainly account for the prominent promotion of elemental mercury capture by (Fe3-xMnx)1-δO4 due to the increase of Mn content. Although reaction 3 was promoted with the increase of reaction temperature, elemental mercury capture reached the optimal condition at a speci?c temperature, in most cases not the highest temperature due to the in?uence of reaction temperature on the physical adsorption (reaction 2). Mn4? cation on (Fe2.2Mn0.8)1-δO4 may be easier to be reduced by 0 to form Mn3? cation at higher temperatures,13 so the capacity of (Fe2.2Mn0.8)1-δO4 for elemental mercury capture at 300 °C was much less than that of (Fe2.5Mn0.5)1-δO4. Effect of SO2 on Elemental Mercury Capture by (Fe2.2Mn0.8)1-δO4. The chemical composition in the flue gas significantly affects elemental mercury capture by sorbents. The components in the real coal-fired flue gas which can interfere with elemental mercury capture are mainly a high concentration of SO2/SO3.14,15 SO2 gas molecules may compete with gaseous elemental mercury for the active sites. The concentration of SO2 in the real flue gas is about 104-105 times that of elemental mercury (v/v).14 The determination of oxidized mercury concentration at the exit of reactor showed that there was little oxidized mercury in the gas after passing through the reactor tube with (Fe2.2Mn0.8)1δO4 in the presence of 1000 ppmv of SO2. It indicates that the reduced amount of elemental mercury in the breakthrough curve (shown in Figure S3d in the SI) was captured by (Fe2.2Mn0.8)1-δO4. Table 3 shows that the presence of a high concentration of SO2 resulted in an obvious interference with elemental mercury capture by (Fe2.2Mn0.8)1-δO4. However, (Fe2.2Mn0.8)1-δO4 still showed an excellent capacity for elemental mercury capture (>1.70 mg g-1 at 100-250 °C) in the presence of 1000 ppmv of SO2. Previous research postulated a mechanism for the heterogeneous uptake and oxidization of SO2 on iron oxides,16 and the reactions can be described as
 FeIII - OH ? SO2?g? f  FeIII OSO2 ? H?  FeIII OSO2 f  FeII ? SO? 3

?8?

?10? ?11? ?12?

So, Z Q ? BETkk1 ?0?? MnIV ?0
0 t

exp? - kk1 ?0?t?dt

?9?
1544

 FeIII - OH ? SO? - f  FeII ? HSO-4 3

dx.doi.org/10.1021/es103391w |Environ. Sci. Technol. 2011, 45, 1540–1546

Environmental Science & Technology As shown in reactions 10-12, the uptake of SO2 on iron oxides may involve hydroxyl groups on the surface. Furthermore, SO2 can also react with Mn4? cations on the surface,17 and the reaction can be described as SO2?g? ?  MnIV ? 2  O f  MnII ? SO2 4 ?13?

ARTICLE

The possible routes of HCl uptake on (Fe2.2Mn0.8)1-δO4 can be described as  FeIII -OH ? HCl?g? f  FeIII ? Cl- ? H2 O  Mn4? ? HCl?g? f  MnIII ? Cl?ad? ? H?
?

?14? ?15?

If reaction 13 happened, elemental mercury capture by (Fe2.2Mn0.8)1-δO4 would be interfered. Elemental mercury capture by (Fe2.2Mn0.8)1-δO4 in the presence of a high concentration of SO2 at 150 °C was also studied using XPS (shown in Figure 3h-l). The S peaks mainly centered at 168.7 and 169.9 eV, which may be assigned to SO42- and HSO4-, respectively. The formation of SO42- can also be supported by the spectra of Fe 2p, Mn 2p, and O 1s. Three new peaks appeared in the spectra of Fe 2p (713.5 eV), Mn 2p (642.8 eV), and O 1s (532.3 eV), which may be assigned to Fe3? in Fe2(SO4)3, Mn2? in MnSO4, and O2- in SO42-, respectively. XPS analysis showed that 39% of Mn cations on (Fe2.2Mn0.8)1-δO4 transformed to MnSO4. It demonstrates that SO2 reacted with Mn4? on (Fe2.2Mn0.8)1-δO4 to form a surface sulfate species and then interfered with elemental mercury capture. XPS analysis also showed that 65% of the transformation of SO2 to SO42- involved Fe cations on the surface, and only 35% involved Mn cations on the surface. As a result, the high concentration of SO2 showed a moderate e?ect on elemental mercury capture by (Fe2.2Mn0.8)1-δO4, which was much less than it on elemental mercury capture by (Fe2Ti)0.8O4.3 Taking into account the binding energy of Hg 4f7/2 at 100.7 eV and the absence of Hg 4f 5/2 at about 105 eV corresponding to Hg2? (shown in Figure 3k), the oxidized mercury formed in the presence of SO2 may be mercurous sulfate. Mercurous sulfate has been previously observed during the photochemical removal of elemental mercury from the ?ue gas.18,19 Our previous research also demonstrated that mercurous sulfate formed during elemental mercury capture by (Fe2Ti)0.8O4 and (Fe2Ti0.8Mn0.2)1-δO4 in the presence of a high concentration of SO2.3,4 Effect of HCl on Elemental Mercury Capture by (Fe2.2Mn0.8)1-δO4. The presence of HCl in the flue gas may enhance elemental mercury oxidization, and the formed HgCl2 may sublime into the flow at 300 °C.20 So the effect of HCl on elemental mercury capture by (Fe2.2Mn0.8)1-δO4 was investigated. The determination of oxidized mercury concentration at the exit of reactor still showed that there was little oxidized mercury in the gas after passing through the reactor tube with (Fe2.2Mn0.8)1-δO4 in the presence of 1000 ppmv of SO2 and 5 ppmv of HCl. It indicates that the reduced amount of elemental mercury in the breakthrough curve (shown in Figure S3e in the SI) was captured by (Fe2.2Mn0.8)1-δO4. As shown in Table 3, the presence of HCl resulted in an insigni?cant e?ect on elemental mercury capture at 100-200 °C. But it showed a moderate interference and an obvious promotion at 250 and 300 °C, respectively. Taking account of the binding energy of Hg4f7/2 at about 100.7 eV and the absence of Hg 4f 5/2 at about 105 eV corresponding to Hg 2?, the oxidized mercury formed in the presence of SO 2 and HCl may still be mercurous sulfate or HgCl (shown in Figure 3m-o). The binding energy centered at about 103.2 eV was attributed to Si 2p of SiO 2 in quartz wool.

Reaction 14 may predominate over the uptake of HCl on (Fe2.2Mn0.8)1-δO4 at 100-200 °C, so the presence of 5 ppmv of HCl showed an insigni?cant e?ect on elemental mercury capture by (Fe2.2Mn0.8)1-δO4. The amount of FeIII-OH would decrease with the increase of reaction temperature due to the dehydration, so reaction 15 may predominate over the uptake of HCl on (Fe2.2Mn0.8)1-δO4 at 250-300 °C. There may be a large number of SO2 adsorbed on (Fe2.2Mn0.8)1-δO4 at 250 °C, so the formed Cl(ad)* may be eliminated by the adsorbed SO2. As a result, the presence of 5 ppmv of HCl showed a moderate interference with elemental mercury capture at 250 °C. Furthermore, the oxidized mercury formed may still be mercurous sulfate at 100-200 °C. Most of SO2 would desorb from (Fe2.2Mn0.8)1-δO4 with the further increase of reaction temperature, so the Langmuir-Hinshelwood mechanism may mainly account for the oxidization of elemental mercury at 300 °C. The near two Mn4? cations on (Fe2.2Mn0.8)1-δO4 were spaced at least by two Fe cations and four oxygen anions, so the formed Cl*(ad) may also be spaced by two Fe cations and four oxygen anions. When a mercury atom was physically adsorbed on the active site (i.e., 0), at most one Cl*(ad) can be covered, so reaction 16 happened and the oxidized mercury formed was HgCl 0 - Hg ? Cl?ad? f HgI Cl?ad?
?

?16?

The kinetic constant of reaction 16 may be much more than that of reaction 3, so elemental mercury capture by (Fe2.2Mn0.8)1-δO4 at 300 °C was promoted by 5 ppmv of HCl. As well-known, Hg2Cl2 is not stable, and it can self-decompose to HgCl2 and Hg at a high temperature. Because HgCl formed on (Fe2.2Mn0.8)1-δO4 was spaced at least by two Fe cations and four oxygen anions, the near two HgCl cannot collide to transform to one Hg and one HgCl2. As a result, the formed HgCl was stable and Hg2? was undetected at the exit of the reactor. Magnetic Separation. Although the crystal sizes of synthesized (Fe3-xMnx)1-δO4 were less than 50 nm, their particulate sizes were higher than 100 μm due to the agglomeration after the thermal treatment. The magnetic sorbent can be recovered in situ by a two-step process. The magnetic sorbent can first be removed from the flue gas as a mixture with the fly ash particles by an electrostatic precipitator or fabric filter, followed by the magnetic separation of the sorbent and adsorbed mercury from the fly ash. Previous research demonstrated that the magnetic sorbent MagZ-Ag0 with the saturation magnetization of 40 emu g-1 and the BET surface area of 164 m2 g-1 can be easily separated from the fly ash.6,7 The photograph inserted in Figure 2 shows the result of separating (Fe2.2Mn0.8)1-δO4 from the mixture with 10 g of fly ash and 1 g of (Fe2.2Mn0.8)1-δO4 by a normal magnet. After (Fe2.2Mn0.8)1-δO4 was separated from the mixture, the contents of Mn and Fe in the fly ash did not increase. It indicates that (Fe2.2Mn0.8)1-δO4 can be separated from the fly ash using magnetic separation, leaving the fly ash essentially free of sorbent and adsorbed mercury. Magnetic separation has been widely used in the mineral processing,21 and the device of the equipment for the magnetic separation of the sorbent from the fly ash may be modeled on that used in the mineral processing.
1545
dx.doi.org/10.1021/es103391w |Environ. Sci. Technol. 2011, 45, 1540–1546

Environmental Science & Technology In summary, (Fe2.2Mn0.8)1-δO4 showed an excellent capacity for elemental mercury capture. Meanwhile, its inherent magnetization made it possible to separate (Fe2.2Mn0.8)1-δO4 from the ?y ash, leaving the ?y ash essentially free of sorbent and adsorbed mercury. Therefore, (Fe2.2Mn0.8)1-δO4 may be a promising sorbent for the control of elemental mercury emission. In our future work, (Fe2.2Mn0.8)1-δO4 will be investigated to capture elemental mercury from the ?ue gas at a pilot scale, in which the separation of sorbent from the ?y ash and sorbent regeneration will be further studied.

ARTICLE

’ ASSOCIATED CONTENT
S b

Supporting Information. Text, Figures S1-S3, and Tables S1 and S2. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION
Corresponding Author

*Phone: 86-21-54745591; e-mail: nqyan@sjtu.edu.cn.

’ ACKNOWLEDGMENT This study was supported by the High-Tech R&D Program of China (No. 2007AA06Z340) and Shanghai Tongji Gao Tingyao Environmental Science & Technology Development Foundation. ’ REFERENCES
(1) Presto, A. A.; Granite, E. J. Survey of catalysts for oxidation of mercury in ?ue gas. Environ. Sci. Technol. 2006, 40, 5601–5609. (2) Yang, S. J.; Guo, Y. F.; Yan, N. Q.; Qu, Z.; Xie, J. K.; Yang, C.; Jia, J. P. Capture of gaseous elemental mercury from ?ue gas using a magnetic and sulfur poisoning resistant sorbent Mn/γ-Fe2O3 at lower temperatures. J. Hazard. Mater. 2010, doi:10.1016/j.jhazmat.2010.11.034. (3) Yang, S. J.; Guo, Y. F.; Yan, N. Q.; Wu, D. Q.; He, H. P.; Qu, Z.; Yang, C.; Zhou, Q.; Jia, J. P. Nanosized cation-de?cient Fe-Ti Spinel: A novel magnetic sorbent for elemental mercury capture from ?ue gas. ACS Appl. Mater. Inter. 2010, under revision. (4) Yang, S. J.; Guo, Y. F.; Yan, N. Q.; Wu, D. Q.; He, H. P.; Xie, J. K.; Qu, Z.; Jia, J. P. Remarkable e?ect of the incorporation of titanium on the catalytic activity and SO2 poisoning resistance of magnetic Mn-Fe spinel for elemental mercury capture. Appl. Catal., B 2010, 101, 698-708. (5) Yang, S. J.; Guo, Y. F.; Yan, N. Q.; Wu, D. Q.; He, H. P.; Xie, J. K.; Qu, Z.; Yang, C.; Jia, J. P. A novel muti-functional magnetic Fe-Ti-V spinel catalyst for elemental mercury capture and callback from ?ue gas. Chem. Commun. 2010, 46, 8377–8379. (6) Dong, J.; Xu, Z. H.; Kuznicki, S. M. Magnetic multi-functional nano composites for environmental applications. Adv. Funct. Mater. 2009, 19, 1268–1275. (7) Dong, J.; Xu, Z. H.; Kuznicki, S. M. Mercury removal from ?ue gases by novel regenerable magnetic nanocomposite sorbents. Environ. Sci. Technol. 2009, 43, 3266–3271. (8) Kodama, T.; Ookubo, M.; Miura, S.; Kitayama, Y. Synthesis and characterization of ultra?ne Mn(II)-bearing ferrite of type MnxFe3-xO4 by coprecipitation. Mater. Res. Bull. 1996, 31, 1501–1512. (9) Yang, S. J.; He, H. P.; Wu, D. Q.; Chen, D.; Liang, X. L.; Qin, Z. H.; Fan, M. D.; Zhu, J. X.; Yuan, P. Decolorization of methylene blue by heterogeneous Fenton reaction using Fe3-xTixO4 (0exe0.78) at neutral pH values. Appl. Catal., B 2009, 89, 527–535. (10) Ayub, I.; Berry, F. J.; Crabb, E.; Helgason, O. Titanium-doped γ-Fe2O3: Reduction and oxidation properties. J. Mater. Sci. 2004, 39, 6921–6927.

(11) Oliveira, L. C. A.; Fabris, J. D.; Rios, R.; Mussel, W. N.; Lago, R. M. Fe3-xMnxO4 catalysts: phase transformations and carbon monoxide oxidation. Appl. Catal., A 2004, 259, 253–259. (12) Granite, E. J.; Pennline, H. W.; Hargis, R. A. Novel sorbents for mercury removal from ?ue gas. Ind. Eng. Chem. Res. 2000, 39, 1020– 1029. (13) Gillot, B.; Laarj, M.; Kacim, S. Reactivity towards oxygen and cation distribution of manganese iron spinel Mn3-xFexO4 (0exe3) ?ne powders studied by thermogravimetry and IR spectroscopy. J. Mater. Chem. 1997, 7, 827–831. (14) Presto, A. A.; Granite, E. J. Impact of sulfur oxides on mercury capture by activated carbon. Environ. Sci. Technol. 2007, 41, 6579–6584. (15) Presto, A. A.; Granite, E. J.; Karash, A. Further investigation of the impact of sulfur oxides on mercury capture by activated carbon. Ind. Eng. Chem. Res. 2007, 46, 8273–8276. (16) Fu, H. B.; Wang, X.; Wu, H. B.; Yin, Y.; Chen, J. M. Heterogeneous uptake and oxidation of SO2 on iron oxides. J. Phys. Chem. C 2007, 111, 6077–6085. (17) Yu, J.; Guo, F.; Wang, Y. L.; Zhu, J. H.; Liu, Y. Y.; Su, F. B.; Gao, S. Q.; Xu, G. W. Sulfur poisoning resistant mesoporous Mn-base catalyst for low-temperature SCR of NO with NH3. Appl. Catal., B 2010, 95, 160–168. (18) McLarnon, C. R.; Granite, E. J.; Pennline, H. W. The PCO process for photochemical removal of mercury from ?ue gas. Fuel Process. Technol. 2005, 87, 85–89. (19) Granite, E. J.; Pennline, H. W.; Ho?man, J. S. E?ects of photochemical formation of mercuric oxide. Ind. Eng. Chem. Res. 1999, 38, 5034–5037. (20) Scho?eld, K. Fuel-mercury combustion emissions: An important heterogeneous mechanism and an overall review of its implications. Environ. Sci. Technol. 2008, 42, 9014–9030. (21) Shen, H. T.; Forssberg, E. An overview of recovery of metals from slags. Waste Manage. 2003, 23, 933–949.

1546

dx.doi.org/10.1021/es103391w |Environ. Sci. Technol. 2011, 45, 1540–1546


相关文章:
光腔衰荡光谱(CRDS)技术
gaseous elemental mercury (Hg0) mass concentration... techniques were used to analyse ammonia gas NH3...[J].Applied Magnetic Resonance,2006,21(3-4) [...
文献翻译 毕业论文 关于气相二价汞在钙基湿法烟气脱硫...
elemental form (Hg°), which was then partly ...mercury capture from coal derived flue gases had ...2 L/min of synthetic gas containing80|xg/m3 ...
更多相关标签: