Dielectric Barrier Discharge Coupling Catalytic Oxidation for Highly Efficient Hg0 Conversion.

In this work, we prepared CuCe/Ti catalysts in a dielectric barrier discharge (DBD) reactor and proposed a new method for flue gas mercury oxidation using DBD coupling CuCe/Ti catalyst. Our experiments verified the oxidation efficiency of flue gas Hg0 (ηHg) and clarified the influence of O2 content, NO concentration, SO2 concentration, water vapor content, and discharge voltage on ηHg. The oxidation mechanism of Hg0 in the DBD-CuCe/Ti reactor was also illustrated. The Hg0 oxidation experiment on the simulated flue gas (70 μg/m3 Hg0 + 300 mg/m3 NO + 1000 mg/m3 SO2 + 6%O2) with a flow rate of 1 L/min showed that when the amount of catalyst was 1.25 g and the discharge voltage was 9.5 kV, a ηHg of 93% can be achieved, which indicates that the DBD coupling CuCe/Ti technology is suitable for Hg0 conversion and flue gas mercury removal.

Introduction

The mercury in the environment can be divided into naturally released mercury and human released mercury. Eighty percent of the mercury in the atmosphere is in the form of vapor, and mainly from the burning of fossil fuels. Gaseous mercury is mercury that can pass through a 0.45 μm pore filter membrane or other filtering devices. It has the characteristics of long retention time, so it can not only carry out the long-distance transmission but also participate in the global mercury cycle, form particles, and settle in situ. China is the world’s largest coal consumer. The average mercury content (approximately 0.15–0.20 μg/g) of coal in China is higher than the world averaged content (0.13 μg/g), thus the mercury pollution problem in China is more serious.[1−4]

The Hg in coal mainly exists in the form of mercury–sulfur bonds. When the boiler burns at high temperatures, the mercury–sulfur bond breaks, and most of the Hg enters into the flue gas in the form of gaseous element mercury Hg0, which may be chlorinated or oxidized to form Hg2+ or be catalytically oxidized on the surface of fly ash. Gaseous Hg2+ is adsorbed on the surface of fly ash to form particulate mercury Hgp.[5,6][5,6] At present, there are three ways to remove mercury after combustion. One is to modify the existing flue gas purification equipment to achieve the combined removal of multiple pollutants and improve the removal efficiency; the other is to add absorbents to the tail flue gas to remove Hg; the third is to use a catalyst to oxidize Hg in the flue gas to ease its removal.[7−10] For example, selective catalytic reduction (SCR) technology can oxidize Hg0 to Hg2+ and the conversion rate reaches 30–80%.[11] The generated Hg2+ and particulate mercury Hgp can be further removed by electrostatic precipitator (ESP) or bag filter and wet desulfurization unit. However, these traditional mercury removal methods have too long and complicated processes, numerous equipment, and a large area occupation, and the operation and equipment maintenance costs are too high.[12] Therefore, the development of new technologies and the use of as few equipment as possible to efficiently remove multiple pollutants in a relatively short process are important development directions for flue gas purification.[10,13]

Hg0 has a low melting point and is not easily soluble in water, acid, and alkali. Therefore, the removal of Hg0 depends on its efficient and quick oxidation of Hg0 to Hg2+, which is easily removed.[14] The use of nonthermal plasma (NTP) to treat the catalyst and then coordinate Hg0 removal will be one of the new development trends of flue gas Hg0 removal in the future. Compared with the traditional Hg0 removal technology, this method can not only save the cost of the precious metal catalyst but also efficiently remove a variety of pollutants in the flue gas. Moreover, it has a small equipment area, simple procedures, operation convenience, and other advantages.[15]

NTP uses corona discharge to generate •O and •OH free radical particles and O3, which has a good effect on the oxidation of pollutants such as NO, SO2, and Hg0. Studies have shown[16,17] that NTP effectively removes NO, SO2, and Hg0 in the flue gas, but there are also some problems, such as large power consumption, low selectivity of active free radicals, and low energy utilization efficiency.[8,18,19] Therefore, other approaches must be combined to promote the application of NTP technology in the pollutant removal process. Catalytic oxidation is one of the commonly used approaches because it can increase the reaction rate and promote the conversion of pollutants. In this approach, the role of the catalyst is very critical.[20,21] Commonly used catalysts include precious metals, transition-metal oxides, activated carbon, and molecular sieves. In recent years, catalysts loading transition-metal oxides such as iron, copper, manganese, cobalt, and cerium have been extensively studied.[22,23] An et al.[24] prepared a 3% CeO2–WO3/TiO2 catalyst and found that the oxidation efficiency of Hg0 (ηHg) can reach 86.6% after combining with NTP.

Plasma technology has the advantage of the ability to produce nanostructured catalysts, low energy consumption, environmentally friendly properties, ability to use a wide range of substrates, and a high degree of universality in catalyst preparation compared with calcination.[25,26]

Liu et al.[12] and Huang et al.[27] showed that the catalyst with good performance could be prepared in a dielectric barrier discharge (DBD) reactor. When the catalyst is filled in the discharge area of DBD, the performance of the catalyst can be further improved in the discharge process of DBD.[28] Our previous work[29−31] showed that DBD combined with the catalyst can effectively promote simultaneous oxidation. Zhang et al.[32] and Zhou et al.[33] showed that the DBD coupling catalyst can significantly improve ηHg. Previous studies[29−31] have also shown that, compared with other metal oxides, Cu oxide has a strong ability to oxidize Hg0, while Ce oxide has good oxygen storage capacity. Titanium dioxide (TiO2) is rich in resources, has low cost, has stable chemical properties, and resistant to acid and alkali corrosion, thus it is often used as a synthetic material for commercial catalysts. It can be expected that the TiO2 catalyst loaded with Cu oxide and Ce oxide has a good oxidation ability for Hg0. Therefore, the combination of DBD-CuCe/Ti catalyst is worthy of further study.[34−36]

The main contents of this work are as follows: (1) use DBD reactor to prepare a CuCe/Ti catalyst suitable for Hg0 oxidation and perform X-ray diffraction (XRD) pattern analysis, X-ray photoelectron (XPS) energy spectrum analysis, and scanning electron microscopy (SEM) analysis; (2) determine the oxidation and conversion efficiency of Hg0 by the DBD-CuCe/Ti reactor and show the influence of O2 content, NO concentration, SO2 concentration, and water vapor content on ηHg; and (3) clarify the oxidation mechanism in the DBD-CuCe/Ti reactor.

Experimental Section

Chemicals and Materials

Cerium nitrate solution (Ce(NO3)3·6H2O) and copper nitrate (Cu(NO3)2·3H2O) were of AR grade and purchased from Shanghai Macklin Biochemical Co., Ltd. Titanium dioxide (TiO2, P25) was purchased from Degussa. All chemicals were used as received without further purification. The deionized water used in the experiment was prepared with an ultrapure system. High-purity N2, high-purity O2, 5% NO/N2, and 5% SO2/N2 were used as plasma-catalytic reaction reactants.

Catalyst Preparation

Using TiO2 (Degussa P25) as a carrier, the required catalyst was prepared by the impregnation method. To prepare Cu/Ti (mass ratio Cu/TiO2 = 0.1), Ce/Ti (mass ratio Ce/TiO2 = 0.1), and CuCe/Ti (molar ratio Cu/Ce = 1:1, mass ratio (Cu + Ce)/TiO2 = 0.1) catalysts, Cu(NO3)2·3H2O and Ce(NO3)3·6H2O were needed. First, certain amounts of cerium nitrate and copper nitrate were dissolved in deionized water then the dried TiO2 (Degussa P25) powder was added into the solution. The mixture was stirred using a thermostatic magnetic stirrer (DF101S, Yuhua Instrument Company, Gongyi) at a constant temperature of 60 °C for 30 min. Subsequently, the sample was sonicated with an ultrasonic oscillator (PS-30T, Kangjie Electric Appliance Company, Shenzhen) for 30 min. Finally, the catalyst precursor was packed in the sealed discharge area of the DBD reactor, then N2 and O2 were passed into the reactor through a gas distribution system. DBD discharged the catalyst precursor under N2–O2 atmosphere to prepare the required catalyst.

Catalyst Characterization

The crystal morphology of the catalyst was determined by an X-ray diffractometer (D8, Bruker, Germany). The diffractometer used Cu Kα (λ = 0.1542 nm) as the radiation source, with a diffraction angle of 2θ ranging from 10 to 90° and a resolution of 0.02°. The results were analyzed by MDI Jade 6.0 software. X-ray photoelectron spectroscopy (XPS) analysis was performed on an electron spectrometer (ESCALAB 250 Xi, Thermo Scientific). All binding energies were calibrated by the C 1s line of 284.6 eV. The relative content of the elements on the catalyst surface was deduced by analysis and fitting of the Gaussian–Lorentzian curves using XPS PEAK software. Scanning electron microscopy (SEM) images were taken on a JEOL JSM-7500F instrument equipped with an OXFORD energy-dispersive spectrometer at an acceleration voltage.

Experimental Setup

The system shown in Figure can be used to prepare the catalysts and oxidize pollutants. The whole system is divided into the gas distribution system, DBD reactor, discharge control and monitoring system, analysis and test system, and exhaust gas treatment system.

Figure 1

Experimental system schematic.

The simulated flue gas was prepared by the gas distribution system, and the N2, NO, SO2, and O2 gases were mixed according to the required concentration under the control of the flowmeter. H2O(g) was produced by a water vapor generator. The mercury permeation tube was placed in a U-shaped glass tube filled with glass beads, which ensured uniform heating of the mercury permeation tube and a uniform gas flow. The U-shaped tube was heated in a water bath at a constant temperature (50 °C) to ensure a stable amount of elemental mercury (70 μg/m3). The catalyst (1.25 g) was placed in the discharge space of the DBD reactor and fixed by quartz wool. The flow rate of the inlet flue gas was 1 L/min. The reacted gas was analyzed by a mercury analyzer and a flue gas analyzer and finally emptied after being absorbed by an exhaust gas treatment device.

Results and Discussion

XRD, XPS, and SEM Analysis

The XRD patterns of the Cu/Ti, Ce/Ti, and CuCe/Ti catalysts are shown in Figure . The observed characteristic peaks of CuO/Cu2O (Cu/Ti catalyst), CeO2 (Ce/Ti catalyst), CuO/Cu2O/CeO2 (CuCe/Ti catalyst) indicated that the CuCe/Ti catalyst had a higher dispersion of Cu and Ce oxides. Moreover, the crystallinity of Cu and Ce oxides in the CuCe/Ti catalyst was lower. A study showed that the active component oxide with low crystallinity is more favorable for the catalytic reaction.[37] Therefore, the catalytic oxidation activity of the catalyst containing CuCe may be higher.

Figure 2

X-ray diffraction patterns of catalysts.

The photoelectron spectra of Cu 2p, Ce 3d, O 1s, and Ti 2p are shown in Figure . From Figure a, it can be observed that Cu2+/(Cu+ + Cu2+) in Cu/Ti and CuCe/Ti were 70.6 and 76.2%, respectively. Cu2+ is an active site, which can strengthen the oxidation of Hg0, thereby promoting the catalytic oxidation under low-temperature conditions.[38] The proportions of Ce3+/(Ce3+ + Ce4+) in Cu/Ti and in CuCe/Ti obtained from Figure b are 28.3 and 31.1%, respectively. The doped Cu and Ce may decrease the energy required for the formation of oxygen vacancies on the catalyst surface, increasing Ce3+/(Ce3+ + Ce4+), which is conducive to the formation of chemisorbed oxygen. From Figure c, it can be calculated that Oα/(Oα + Oβ) in CuCe/Ti is 25.4%, while these values in Ce/Ti and Cu/Ti are 15.3 and 16.9%, respectively. Therefore, CuCe/Ti may have higher catalytic activity in the catalytic oxidation of Hg0.[39] It can be calculated from Figure d that the values of Ti3+/(Ti3+ + Ti4+) in Ce/Ti, Cu/Ti, and CuCe/Ti are 68.6, 70.6, and 77.2%, respectively. As active O species have mainly resulted from the conversion of Ti3+ to Ti4+,[40][40] one concludes that among the three catalysts, the CuCe/Ti catalyst may have the highest catalytic activity.

Figure 3

X-ray photoelectron energy spectrum analysis of the catalysts: (a) Cu 2p, (b) Ce 3d, (c) O 1s, and (d) Ti 2p.

The surface morphology in Figure shows that the surfaces of Cu/Ti and Ce/Ti catalysts are loose, the regularity of morphology is poor, and the particle size distribution is not uniform enough; however, the CuCe/Ti catalyst has a smoother surface and more uniform particle size distribution. The above three characterization results indicated that CuCe/Ti may have better catalytic performance.[41−43] Therefore, in this work, we selected CuCe/Ti as the catalyst and cooperated with the DBD reactor to oxidize Hg0.

Figure 4

Scanning electron microscope analysis of the catalysts: (a) Cu/Ti (50 000:1); (b) Ce/Ti (50 000:1); (c) CuCe/Ti (50 000:1); (d) Cu/Ti (20 000:1); (e) Ce/Ti (20 000:1); and (f) CuCe/Ti (20 000:1).

Effects of O2 Content and Discharge Voltage on ηHg

The energy density can be calculated by SED (J/L) = P (J/s)/Q (L/min) × 60 s/min, in which P represents the power consumed and Q represents the flow of gas.[33] We recorded the voltage and current data from the oscilloscope. The relationship between energy density and voltage thus obtained is shown in Figure , indicating that when the voltage increased from 3 to 7.5 kV, SED only slightly increased from 4.6 to 7.8 J/L. When the voltage was within the range of 8–10 kV, SED increased rapidly from 38.1 to 48.5 J/L. When the voltage continued to increase, SED increased faster.

Figure 5

Voltage dependence of SED.

O3 and free radicals are produced during the DBD discharge, as shown in eqs –4. Figure shows the effect of the O2 content on the ηHg at different discharge voltages. In the range of 3–8 kV, the variation of the Hg0 content was very small, so it is not plotted in Figure .

Figure 6

O2 content dependence of ηHg at different discharge voltages.

The reason for this can be explained in Figure . When the voltage was less than 8 kV, the change in energy density was small, so the number of generated free radicals was small and the ηHg low. When the voltage was greater than 8 kV, the energy density increased rapidly as the voltage increased, and the ηHg also increased rapidly. When the discharge voltage was constant, the higher the O2 content, the higher the collision probability of O2 and higher the energy of the electrons, thus more O3 and •O were generated to participate in the reaction, and the ηHg was increased. For the four kinds of O2 contents, the ηHg reached the maximum (85.1, 90.7, 93.5, and 96.5%, respectively) in the range of 9.5–10 kV. In the case of too high voltage, because N2 in the flue gas as the carrier gas and its content far exceeds O2, a large amount of •N produced by high-voltage discharge would preferentially consume O3 and •O, resulting in a decrease in the ηHg. Moreover, in the case of too high voltage, part of HgO (as shown in eqs and 6) adsorbed on the inner wall of the quartz tube would be reduced to elemental mercury under the influence of discharge arc, thus reducing the ηHg.

Effects of NO, SO2, and H2O(g) on ηHg

The composition of flue gas is very complex, including O2, CO2, N2, SO2, NOx, H2O, and HCl. When DBD is used to treat flue gas, multiple components can react with high-energy electrons to generate different active substances and free radicals and undergo redox reactions with O3 and •O, thereby affecting the conversion of Hg0. In this section, we used the controlled variable method to study the influences of NO, SO2, and H2O(g) on the ηHg.

NO will be oxidized by active materials and affect the ηHg. An et al.[44] found that the presence of NO will reduce the ηHg because the reaction rate constant of the reaction of NO and O3 was much greater than that of NO and Hg0, so NO would compete with Hg0 for active substances. However, some others[27,34] believed that NO provides additional O atoms to increase the active materials and promote the oxidation of Hg0. The NO concentration dependence of ηHg was shown in Figure (O2 concentration was 6%).

Figure 7

NO concentration dependence of ηHg.

The experimental results supported the conclusion of An et al.[44] that increasing NO would reduce the ηHg. The ηHg showed an upward trend with the increase of voltage and reached the maximum at 10 kV. The concentration of NO at the outlet also reached the minimum value at 10 kV, which proved that NO also participates in the reaction and is converted during the DBD discharge process. The main reactions in the DBD reactor include eqs –9[45−47]The effect of SO2 concentration on ηHg is shown in Figure (O2 concentration was 6%).

Figure 8

SO2 concentration dependence of ηHg.

Different from Figure , the ηHg quickly reached about 95% in the range of 3–8 kV, and the concentration of SO2 at the outlet also decreased continuously with the increase in voltage. Wang et al.[48] found that SO2 consumes O atoms and inhibits the conversion of Hg0. However, some researchers[20,24,33] believed that SO2 is mainly oxidized by •OH and •O to HSO3– and SO3, which had little effect on the conversion of Hg0. Through our experimental data, we found that the generated HSO3– and SO3 react with HgO to generate HgSO4, thereby oxidizing part of Hg0. The main reactions include eqs –13; however, there is still controversy about how SO2 affects the oxidation of Hg0.[49,50]To clarify the effect of H2O(g) content on the ηHg, we generated a certain amount of water vapor by bubbling and heating, and introduced the water vapor into the DBD reactor through N2, and obtained the corresponding absolute humidity based on the flow rate. The result is shown in Figure (O2 concentration was 6%).

Figure 9

Water content dependence of ηHg.

When there was no H2O in the reactor, the ηHg reached a peak of 93.5% at 9.5 kV. When there was H2O, the ηHg decreased with the increase in H2O(g). The chemical process in the DBD reactor is shown in eqs –17, indicating that water vapor can consume part of O3, thereby inhibiting the oxidation of Hg0 [44]The effect of the coexistence of NO and SO2 on the ηHg was shown in Figure . Under 6% O2 concentration, the ηHg was the highest when only SO2 was present, which can reach 95.6% at 4 kV and 100% at 9 kV. The ηHg was the lowest when only NO was present, which indicated again that SO2 promotes the oxidation of Hg0, while NO inhibits the oxidation of Hg0. When the inlet gas was composed of 70 μg/m3 Hg0 + 6% O2 + 300 mg/m3 NO + 1000 mg/m3 SO2, a ηHg of 93% was achieved at 9.5 kV. However, adding H2O(g) to this inlet gas would reduce the ηHg. For example, when the 2% H2O was introduced in the above inlet gas, the ηHg rapidly dropped from 93 to 86.3%; when the introduced H2O content was 5%, the ηHg dropped to 79.9%. Therefore, it is best to conduct dehydration first when using DBD-CuCe/Ti reactor to oxidize Hg0 or simultaneously oxidize multiple pollutants in the flue gas. Besides, the flue gas HCl affects the redox reaction, although the chemical properties of the flue gas CO2 are very stable, the concentration of CO2 is relatively high. Hence, it is necessary to investigate the influence of HCl and CO2 on the oxidation of multipollutants in future work.

Figure 10

NO and SO2 concentrations dependences of ηHg.

Conclusions

To efficiently oxidize Hg0 in the flue gas, we prepared the CuCe/Ti catalyst and proposed a new method of DBD coupling the CuCe/Ti catalytic oxidation. The experiment showed that DBD coupling CuCe/Ti can lead to high ηHg, so it has great application potential. Our results showed that

the oxides of Cu–Ce had higher catalytic activity than the single metal oxides;

when the discharge voltage was about 9.5 kV, the maximum ηHg reached 93%, which means both the energy density and ηHg were satisfactory;

the oxidation of Hg0 was mainly due to the oxidation from •O, •OH, and O3, but the intermediate products also affected the ηHg; and

O2 and SO2 were favorable for the oxidation of Hg0, but increasing NO and H2O tended to decrease ηHg.