Simultaneous Desulfurization and Denitrification of Flue Gas by ·OH

Aug 14, 2012 - The O2+ ions are generated by a strong ionization dielectric barrier discharge and then injected into the duct. Compared with conventio...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/est

Simultaneous Desulfurization and Denitrification of Flue Gas by ·OH Radicals Produced from O2+ and Water Vapor in a Duct Mindi Bai,* Zhitao Zhang, and Mindong Bai Institute of Environmental Engineering, Department of Physics, Dalian Maritime University, Dalian 116026, China ABSTRACT: In the present study, simultaneous flue gas desulfurization and denitrification are achieved with ·OH radicals generated from O2+ reacting with water vapor in a duct. The O2+ ions are generated by a strong ionization dielectric barrier discharge and then injected into the duct. Compared with conventional gas discharge treatment, the present method does not need a plasma reaction reactor, additional catalysts, reductants, or oxidants. The main recovered products are the liquids H2SO4 and HNO3, which can be used in many processes. Removal rates of 97% for NO and 82% for SO2 are obtained under the following optimal experimental conditions: molar ratio of reactive oxygen species (O2+, O3) to SO2 and NO, 5; inlet flue gas temperature, 65 °C; reaction time, 0.94 s; and H2O volume fraction, 8%. Production of O2+ and the plasma reaction mechanisms are discussed, and the recovered acid is characterized. The experimental results show that the present method performs better for denitrification than for desulfurization. Compared with conventional air discharge flue gas treatments, the present method has lower initial investment and operating costs, and the equipment is more compact. mean electron energy needs to be above 2 eV for efficient flue gas treatment. Electron beam irradiation provides a mean electron energy of 33 eV,7 which is more than 10 times higher than that of a pulse corona discharge. Industrial installations using electron beam irradiation have been built at coal-fired power plants in Japan, Poland and China, and desulfurization and denitrification rates can reach 85−95%.8 Chang et al.9 proposed that the high capital costs and X-ray hazards of electron beam irradiation have discouraged its use in many pollution control applications. To overcome the disadvantages of electron beam irradiation, many studies have investigated flue gas treatment using DBD. Wang et al.10 investigated simultaneous flue gas desulfurization and denitrification by injecting O3 produced by DBD. About 97% of NO and nearly 100% of SO2 were removed simultaneously with injection of 360 ppm O3. An alkaline solution added during the denitrification process, adsorbed the NO2 which was produced by the O3 oxidation of NO. The alkaline solution absorbed nearly 100% of the SO2, irrespective of whether O3 was injected or not. Wang et al.11 found that there was no obvious removal of NOx and NO when either a C3H6−SCR or DBD system was used, but the NOx removal rate was high (88.5%) in the combined system. In the present study, a high concentration of reactive oxygen species, produced from a strong ionization DBD, were injected into a duct (length = 1 m). The following plasma reactions took

1. INTRODUCTION Combustion processes in coal-fired power plants emit NOx and SO2 that cause air pollution, reduce air quality, and are a health risk. Wet flue-gas desulfurization technology can be used for efficient and cheap removal of SO2, but cannot be used to remove NOx. The largest component of NOx is NO, which has a very low solubility in aqueous solutions.1 Denitrification by selective catalytic reduction (SCR) is effective for NOx treatment. Excellent and extensive reviews of the current technologies used to remove SO2 and NOx from flue gas are provided by Srivastava et al.2,3 Compared with a combined system, the use of separate wet flue-gas desulfurization and SCR units requires higher investment and running costs,4 which developing countries want to avoid. With nonthermal plasma processes, including pulse corona discharge, electric beam irradiation, and dielectric barrier discharge (DBD), several pollutants can be treated simultaneously at atmospheric pressure. The operating costs for these processes are typically less than one-quarter than that of a conventional SCR system.5 Huang et al.6 combined pulsed corona with in situ Ca(OH)2 absorption for removal of SO2 and NOx, and achieved removal rates of 75% (SO2) and 40% (NO). By comparison, removal rates were 18% (SO2) and 30% (NO) with pulsed corona discharge alone. The low desulfurization and denitrification rates with pulse corona discharge can be attributed to a low mean electron energy (2 eV) in the discharge channel of the plasma source, which indicates that most electrons have lost considerable energy through vibrational excitation and only a few high-energy electrons are present for the desired reactions. Therefore, the © 2012 American Chemical Society

Received: Revised: Accepted: Published: 10161

April 7, 2012 August 14, 2012 August 14, 2012 August 14, 2012 dx.doi.org/10.1021/es3013886 | Environ. Sci. Technol. 2012, 46, 10161−10168

Environmental Science & Technology

Article

place in the duct: (1) O3 oxidation of NO to NO2, and SO2 to SO3; (2) reaction of O2+ with H2O from the flue gas, to produce ·OH radicals; and (3) oxidation of NO2 to HNO3, and SO3 to a H2SO4 submicrometer mist by ·OH radicals. The mist passed through a high-voltage DC field, which captured and collected the mist as a liquid. To our knowledge, this is the first report of simultaneous flue gas desulfurization and denitrification, and production of an acidic liquid, in a duct using ·OH radicals. Compared with simultaneous desulfurization and denitrification using conventional gas ionization discharge techniques, this method has the following advantages: (1) the entire plasma reaction occurs in the duct and a plasma reaction reactor is not required, therefore the investment cost and system size are reduced; (2) no additional catalysts, reductants, oxidants, or the use of other technologies are required because the ·OH radicals are strong oxidants with high reaction rate coefficients; and (3) the recovered products are H2SO4 and HNO3 liquids, which are useful products. Herein, the basic principles for O2+ production are described, and the plasma chemistry mechanisms for simultaneous flue gas desulfurization and denitrification, and acid production are presented. The recovered acids are characterized and the effects of the molar ratios between the reactive oxygen species (O2+, O3) and SO2 and NO, inlet flue gas temperature, reaction time, and H2O volume fraction on the SO2 and NO removal rates are investigated.

2. EXPERIMENTAL SECTION 2.1. Procedure and Apparatus. The experimental setup for simultaneous desulfurization and denitrification in the duct is shown in Figure 1 (a). The temperature and humidity values of the air (1 in Figure 1) were regulated, to match the values of flue gases from a coal-fired power plant, before the air reached the duct (2 in Figure 1). The duct was a stainless tube (I.D. Twenty mm, length 1 m). The experimental gas was formed by mixing NO and SO2 from compressed gas cylinders with the air in the duct. The reactive oxygen species (O2+ and O3) produced by DBD in the reactive oxygen species generator (3 in Figure 1) were injected into the center of the duct. Figure 1 (b) shows the major components such as the reactive oxygen species and high-voltage high-frequency power supply in the plasma reaction duct in the experimental system. The plasma reactions of simultaneous desulfurization and denitrification proceeded to completion between the area where the reactive oxygen species were injected (16 in Figure 1) and the position of the electric acid mist remover (17 in Figure 1). In the plasma reaction duct, O2+ and O3 oxidized SO2 to H2SO4, and NO to HNO3, through the following three reactions. The first of these is the reaction of O2+ with H2O to produce ·OH radicals. The second reaction involves O3 oxidation of SO2 to SO3, and NO to NO2, and the third reaction is ·OH radical oxidation of SO3 to a H2SO4 mist, and NO2 to a HNO3 mist. The mists then passed through a high-voltage direct current electric field in an electric acid mist remover (5 in Figure 1) and were converted to liquids HNO3 and H2SO4. The electric acid mist remover was a grounded, titanium steel cylinder that contained a starshaped electrode wire. The electric field strength in the middle of the acid mist remover was 14 kV/cm. Residual reactive oxygen species in the experimental gas were removed by heat treatment, and the final purified experimental gas was discharged from the duct using an induced-draft fan (7 in Figure 1).

Figure 1. (a) Flow diagram.(b) Photograph of the experimental setup for simultaneous flue gas desulfurization and denitrification in a duct. (1) temperature and humidity regulator; (2) duct; (3) reactive oxygen species generator; (4) high-voltage high-frequency power supply; (5) electric acid mist remover; (6) high-voltage DC power supply; (7) induced-draft fan; (8) peristaltic pump; (9) acid storage tank; (10) reactive oxygen species detector; (11) flue gas analyzer; (12) ion chromatograph; (13) injection NO; (14) injection SO2; (15) removal of acid liquid; (16) injection the reactive oxygen species; (17) electric acid mist remover.

2.2. Analysis Methods. A flue gas analyzer (11 in Figure 1) (Photon + PGD-100, Madur, Vienna, Austria) was used for online monitoring of the volume fraction, temperature, and pressure of each gas (SO2, NO, NOx, and O2). An ion chromatograph (12 in Figure 1) (ICS-1500, Dionex, Sunnyvale, CA) was used for quantitative analyses of the compositions of the recovered acids. The reactive oxygen species detector (10 in Figure 1) included an O2+ density tester and detector (BMT964, BMT Messtechnik GmbH, Stahnsdorf, Germany). We developed a ball probe (ø 6 mm) for determining the O2+ density in the high velocity duct. The density was measured at output ranges from 104 to 1010 cm−3 and air velocities from 0.2 to 50 m/s, as discussed in detail by Bai et al.12 The electron charges collected using the ball probe were converted into a microcurrent, and the O2+density was calculated as follows: ni = 10162

I eVoπd 2 dx.doi.org/10.1021/es3013886 | Environ. Sci. Technol. 2012, 46, 10161−10168

Environmental Science & Technology

Article

where ni is the O2+density (cm−3), I is the microcurrent, Vo is the gas flow, e is the electron charge, and d is the diameter of the ball probe. 2.3. Preparation of O2+. The concentration of O2+ is important in simultaneous desulfurization and denitrification because it determines the concentration of ·OH radicals produced. The O2+ output depends mainly on the concentration of reactive oxygen species. High concentrations of reactive oxygen species, such as O2+, O, O(1D), O2, O2(a1Δg), and O3,13 were produced by a strong ionization DBD in the reactive oxygen species generator (3 in Figure 1). The rectangular (L × W × H, 280 × 50 × 220 mm) reactive oxygen species generator (Figure 2) contained discharge

Figure 4. Atomic force microscopy image of the dielectric layer.

height The uniform and dense distribution of powder particles on a dielectric layer effectively inhibit electric sparks and arc discharges upon the application of increasing discharge currents, and this increases the breakdown voltage. The mean electron energy depends primarily on E/N. When E/N reached 350 Td in the discharge channel, the mean electron energy increased to 8 eV,7 which was 6 eV higher than the mean electron energy for pulsed corona discharge and 3 eV higher than that for conventional DBD. The strong ionization discharge formed by a large number of discharge streams is shown in Figure 5(a), and a single discharge stream is shown in

Figure 2. Diagram of the reactive oxygen species generator.

electrodes, grounding electrodes, spacers, and dielectric layers, with a discharge gas spacing of 0.1 mm between the discharge and grounding electrodes. A high-voltage, high-frequency discharge output from the high-voltage, high-frequency power supply (4 in Figure 1) was applied to the discharge electrodes, and the peak voltage (6 kV), current (100 mA) and waveform (11.5 kHz) were measured by a oscilloscope (TDS3032, Tektronix, Beaverton, OR) (Figure 3). The discharge strength

Figure 5. (a) Discharge image; (b) a single discharge space.

Figure 5(b). The E/N in the discharge channel reached 350 Td, and at the top of the discharge stream in the discharge path the value was 900 Td, resulting from the combined space charge and deposited charge effects with the applied electric field. The electron energy range was 8.4 eV (ionization energy) to 12.5 eV (dissociation energy). The high energy electrons ionized or dissociated O2 into reactive oxygen species, such as O2+, O, O(1D), O, O2, O2(a1Δg), and O3, in the discharge channel. The remaining low energy electrons did not contribute to the desired reactions. The reactive oxygen species O, O, O(1D), O2, and O2(a1Δg) are short-lived (approximately 1 × 10−8 s).13 Consequently, only O2, O2+ and O3 were injected into the duct for oxidation of SO2 and NO. The plasma chemistry reactions and their rate coefficients are presented in Table 1. The rate coefficients of reactions 1−6 were taken into account when selecting E/N = 150 Td in the discharge channel.14 The energy consumption rate affects the concentrations of ions produced by the reactive oxygen species generator (Figure 6). When the energy consumption rate increased from 0.3 Wh/ m3 to 0.7 Wh/m3, the concentration of O2+ quickly increased from 1.56 × 104 /cm3 to 5.57 × 109 /cm3 and that of O2 quickly increased from 7.64 × 103 /cm3 to 2.38 × 109 /cm3. The ion concentration increased by 1 order of magnitude with every 0.1 Wh/m3 increase in the energy consumption rate. When the energy consumption rate increased from 0.9 Wh/m3

Figure 3. Voltage, current and waveform output from the high-voltage high-frequency power supply.

and the mean electron energy in the discharge channel depended on the ratio E/N (1 Td=10−17V cm2, E = electric field intensity, N = concentration of neutral particles).7 Two methods for increasing the concentration of the reactive oxygen species were investigated. The first involved decreasing the gap between the discharge and grounding electrodes from 0.5 mm to 0.1 mm, which increased E/n from 230 to 350 Td in the discharge channel.7 Improved fabrication of the dielectric layers resulted in the application of greater breakdown voltages and ionization discharges. A dielectric layer of high-purity αAl2O3 powder (thickness = 0.47 mm) was analyzed using atomic force microscopy (Bruker, Billerica, MA) (Figure 4). The powder particles were 2−5 μm in diameter and 0.5 μm in 10163

dx.doi.org/10.1021/es3013886 | Environ. Sci. Technol. 2012, 46, 10161−10168

Environmental Science & Technology

Article

Table 1. Plasma Reactions Involved in Reactive Oxygen Species Generation electron impact reaction O2 + e →O2+ + 2e O2 + e → O + O + e O2 + e → O− + O O2 + e → O2 O2 + e → O + O(1D) + e O2 + e → O2(a1Δg) + e

rate coefficient 150 Td k1 k2 k3 k4 k5 k6

= = = = = =

8.0 7.9 3.2 5.6 2.3 7.2

× × × × × ×

10−10 cm3/s 10−10 cm3/s 10−11 cm3/s 10−12 cm3/s 10−9 cm3/s 10−10 cm3/s

Table 3. Plasma Reactions Involved in Reactive Oxygen Species Generation

reference (1) (2) (3) (4) (5) (6)

electron impact reaction SO2 + O3 → SO3 + O2 NO + O3 → NO2 + O2

14 14 14 14 14 14

rate coefficient 150 Td k9 = 3.0 × 10−12 cm3/s k10 = 1.8 × 10−14 cm3/s

(9) (10)

reference 16 14

discharged into the atmosphere along with the purified flue gas from the duct using an induced-draft fan (7 in Figure.1). Because the reaction rate coefficient and oxidization potential of ·OH radicals were 10−12 ∼10−10 cm3/s and 2.80 V, respectively, the ·OH radicals could rapidly and nonselectively oxidize SO3 to H2SO4 and NO2 to HNO3 (Table 4. Table 4. Plasma Reactions Involved in Reactive Oxygen Species Generation electron impact reaction SO2 + ·OH → HSO3 HSO3+ ·OH → H2SO4 NO2 + ·OH → HNO3

Figure 6. Effect of the energy consumption rate on ion concentration. The gas particle momentum is 1.09 × 10−20 g·cm/s at atmospheric pressure.

rate coefficient 150 Td k11 = 7.5 × 10−12 cm3/s k12 = 1.0 × 10−12 cm3/s k13 = 1.0 × 10−10 cm3/s

(11) (12) (13)

reference 14 14 14

This method for simultaneous desulfurization and denitrification is very different to conventional gas ionization discharge methods, in which the entire flue gas passes through the plasma source.17−19 Compared with conventional gas ionization discharge, the present method requires a smaller volume of plasma, is simpler, has reduced investment costs and a lower energy consumption, and does not require additional catalysts, reductants, oxidants, or the use of other technologies. This method also produces the liquids H2SO4 and HNO3, which can be used in other areas.

to 2.18 Wh/m3, the concentration of O2+ only increased from 8.79 × 109 /cm3 to 1.38 × 1010 /cm3, and that of O2 only increased from 2.38 × 109 /cm3 to 7.78 × 109 /cm3. From these results, the optimal energy consumption rate was 1−1.4 Wh/ m3.

3. PLASMA REACTION MECHANISMS FOR REMOVAL OF SO2 AND NO IN A DUCT In the reactive oxygen species generator, O2 was produced by the addition of electrons to O2. However, O2 has a low electron energy, and cannot oxidize SO2 and NO. Consequently, from O2+, O2 and O3 species injected into the duct, only O2+ and O3 oxidized SO2 to H2SO4, and NO to HNO3. Oxidation of SO2 and NO occurs through the following three reactions. The first of these is the reaction of O2+ with H2O to produce ·OH radicals. The second reaction involves O3 oxidation of SO2 to SO3, and NO to NO2, and the third reaction is the ·OH radical oxidation of SO3 to a H2SO4 mist, and NO2 to a HNO3 mist. The plasma reactions proceeded to completion in the 1 m long duct (Figure 1) between the injection of the reactive oxygen species and the position of the electric acid mist remover. First, O2+ reacted with H2O to form water cluster ions, O2+·H2O, and then these dissociated to form ·OH radicals (see Table 2. Next, O3 oxidized SO2 to SO3, and NO to NO2 (Table 3). According to eqs 7−10, a large amount of O2 is produced during oxidation of SO2 and NO by O3 and O2+. This O2 is

4. RESULTS AND DISCUSSION 4.1. Effect of the Molar Ratios between the Reactive Oxygen Species and SO2 and NO on the SO2 and NO Removal rates. The molar ratio of reactive oxygen species (O3, O2+) to SO2 and NO is an important index for simultaneous desulfurization and denitrification technology. It is used to determine the concentration of reactive oxygen species and electrical energy required to produce the radicals. The influence of the molar ratio on the SO2 and NO removal rates is shown in Figure 7. When the molar ratio increases from 0 to 2, the NO removal rate increases rapidly from 0 to 97.2%, and the SO2 removal rate increases slowly from 0 to 17%. During this time, the ·OH radical oxidation of NO2 to produce HNO3 is the main reaction occurring in the duct, and the ·OH

Table 2. Plasma Reactions Involved in Reactive Oxygen Species Generation rate coefficient 150 electron impact reaction Td O2+ + H2O + M →O2+·H2O + M k7 = 2.5 × 10−28 cm6/s + + O2 ·H2O + H2O →H3O + ·OH k8 = 1.2 × 10−9 + O2 cm3/s

(7)

reference 15

(8)

15

Figure 7. Effect of the molar ratio n between reactive oxygen species (O2+, O3) and SO2 and NO on SO2 and NO removal rates. The flow rate is 1.2 m3/h, the inlet flue gas temperature is 65 °C, the reaction time is 0.94s, the H2O volume is 8%, and the concentrations of SO2 and NO are and 2285 mg/m3 and 603 mg/m3, respectively. 10164

dx.doi.org/10.1021/es3013886 | Environ. Sci. Technol. 2012, 46, 10161−10168

Environmental Science & Technology

Article

7 and 8, Section 2.3), the H2O volume fraction is important in determining the denitrification and desulfurization rates. The influence of the H2O volume fraction on SO2 and NO removal rates is shown in Figure 8. When the H2O volume fraction

radical oxidation of SO2 to produce H2SO4 is a secondary reaction. When the molar ratio increases to 5, there is no significant increase in the NO removal rate (97.2−97.4%), and the SO2 removal rate increases rapidly from 17 to 83.2%. These results indicate that the ·OH radical oxidation of NO2 reaches completion at a molar ratio of two, and ·OH radical oxidation of SO2 is the main reaction occurring in the duct when the molar ratio is between two and five. The reaction rate depends primarily on the reaction rate coefficient for ·OH radicals reacting with NO2, SO2 or HSO3. From eqs 11−13 (Section 2.3), the reaction rate coefficients were 1.0 × 10−10 for NO2, 7.5 × 10−12 for SO2, and 1.0 × 10−12 cm3/s for HSO3. The reaction rate coefficient for the ·OH radical reaction with NO2 is 2 orders of magnitude higher than that of the ·OH radical reaction with SO2 or HSO3. Consequently, the denitrification rate is faster than that of desulfurization, and desulfurization occurs after denitrification, which is consistent with the experiment results. While the removal rates at a molar ratio of five are satisfactory (NO, 97.2%; SO2, 83.2%), the present method is more suitable for NO removal rather than SO2 removal. The removal rate for SO2 using this method does not meet the current emission standards.20 Further research will investigate if the addition of Na/Mg mixed oxidation catalysts, will improve the reaction potential and reaction rate of SO2 removal to meet the current standards (SO2 removal rate = 95.6%). The energy consumption per gram of NO or SO2 removed was calculated. With a molar ratio of five, the energy consumption rates are 9 Wh/g (NO) and 17 Wh/g (SO2), and the corresponding removal rates are 97.2% (NO) and 83.2% (SO2). The energy consumption rate of the present method is compared with conventional methods to gauge its feasibility for large-scale applications. Tseng et al. 21 have investigated the combined removal of NOx and SO2 from simulated flue gas in a bench-scale pulsed-corona enhanced wet electrostatic precipitator with the addition of ammonia (NH3) and ozone (O3). The maximum removal rates for NOx and SO2 were 77% and 99%, respectively, with injection of 312 ppm O3 and 2900 ppm NH3. The energy consumption rates were as high as 38.6 Wh/g (NOx) and 0.42 Wh/g (SO2). The total energy consumption in the study by Tseng et al. (39.02 Wh/g for NOx and SO2) is 1.5 times higher than that obtained in the simultaneous flue gas desulfurization and denitrification with O2+ in the present study (26 Wh/g). In addition, the method used by Tseng et al. has higher investment costs than the present method because it uses a wet electrostatic precipitator with a pulse corona discharge, NH3 storage tanks, O3 generator, and a sewage treatment system. Saavedra et al. 13 investigated NOx treatment using DBD without additional catalysts or oxidants, and the NOx removal rate reached 98%. It should be noted that only denitrification was performed in this study, and no desulfurization occurred. The NOx energy consumption rate (875 Wh/g) was 34 times higher than the simultaneous flue gas desulfurization and denitrification method with O2+ in the present study (26 Wh/g). The increased energy consumption is attributed to the requirement for all the flue gas to pass through a DBD reactor. We conclude that compared with conventional air discharge flue gas treatments, the present method has lower initial investment and operating costs, and the equipment is more compact. 4.2. Effect of H2O Volume Fraction on SO2 and NO Removal Rates. Because H2O reacts with O2+ from the reactive oxygen species generator to produce ·OH radicals (eqs

Figure 8. Effect of H2O volume fraction on SO2 and NO removal rates. The reaction time is 0.94 s, the inlet flue gas temperature is 65 °C, the molar ratio between the reactive oxygen species (O2+,O3) and SO2 and NO is 5, the flow rate is 1.2 m3/h, and the concentrations of SO2 and NO are 2285 mg/m3 and 603 mg/m3, respectively.

increases from 2% to 8%, the NO removal rate increases from 78.1% to 97% and that of SO2 increases from 43.9% to 78.5%. When the H2O volume fraction increases from 8% to 12%, the NO and SO2 removal rates increase by 0.4% and 2.7%, respectively. The maximum SO2 and NO removal rates are obtained with a H2O volume fraction of >8%. Because the typical H2O volume fraction in flue gas from a coal-fired power plant is approximately 10%,22 simultaneous flue gas desulfurization and denitrification can occur in these plants without the addition of water vapor. These results differ from those obtained by Saavedra et al.13 and Motret et al.23 Saavedra et al. infer that the NOx removal rate is higher with a lower H2O volume fraction (at 1% compared with 5%); and Motret et al. suggest that the increase in the water volume fraction accelerates ·OH radical decay. This may occur because when all the flue gas passes through the discharge channel of the plasma source, the electric field strength is reduced, and it is not enough to produce reactive species. By contrast, in the present method, the reactive oxygen species (O2+, O3) from the reactive oxygen species generator are injected into the duct to remove SO2 and NO, and this does not change the strength of the electric field or the formation of the reactive oxygen species. In addition to H2O, there are many other contaminants, such as HCl, particulates, CO, unburned hydrocarbons, and SO3, in the flue gas.24−27 Future research will investigate the impacts of these contaminants on the performance for removing SO2 and NOx using O2+. 4.3. Effect of Inlet Flue Gas Temperature on SO2 and NO Removal Rates. The inlet flue gas temperature affects the decomposition of reactive oxygen species. Therefore, we investigated the effect of temperature on the SO2 and NO removal rates (Figure 9). When the temperature increases from 30 to 110 °C, the SO2 removal rate decreases rapidly from 93.2% to 55.8% in a linear manner, with a decrease of 4.6% per 10 °C. Between 30 and 80 °C, the temperature has little effect on the NO removal rate. However, above 80 °C, the NO removal rate decreases rapidly by 10.6% per 10 °C. When the temperature increases many reactive oxygen species are lost in the duct because the kinetic energies of the gas molecules increase, which leads to greater thermal motions. Our results show that the removal rates of NO and SO2 are higher at lower 10165

dx.doi.org/10.1021/es3013886 | Environ. Sci. Technol. 2012, 46, 10161−10168

Environmental Science & Technology

Article

entire plasma reaction reaches completion within the duct, a plasma reaction chamber is not required; therefore this method has lower operational and initial investment costs, and the equipment is more compact than conventional technology. 4.5. Analysis of the Recovered Acid. The acid was collected in an acid storage tank (9 in Figure 1). Constant volume samples were removed from the tank, diluted by a factor of 5000, and then injected into the ion chromatograph (12 in Figure 1). The type of acid was determined from its peak retention time in the chromatogram, and the acid concentration was calculated from the peak area using an external standard method. Ion chromatography was performed for the dilute acid sample from the acid storage tank (Figure 11(b)), and a diluted

Figure 9. Effect of the inlet flue gas temperature on SO2 and NO removal rates. The flow rate is 1.2 m3/h, the molar ratio between the reactive oxygen species (O2+,O3) and SO2 and NO is five, the reaction time is 0.94 s, the H2O volume fraction is 8%, and the concentrations of SO2 and NO are 2285 mg/m3 and 603 mg/m3, respectively.

flue gas temperatures. With a flue gas temperature of 30 °C, the removal rates of NO and SO2 are 97.5% and 93.2%, respectively. The flue gas emission temperature from the chimney in a coal-fired power plant is calculated using Ts = Ta + ΔT),28 where Ts is the flue gas emission temperature, Ta is the ambient temperature near the chimney, and ΔT is the temperature difference between the flue gas emission temperature and the ambient temperature. A ΔT of above 35 °C is required so that the flue gas passing through the chimney can discharge into atmosphere to avoid low-altitude flue gas emissions, which cause environmental pollution. Therefore, when Ta is 25 °C, Ts must be above 60 °C. Consequently, in the present study, the optimum temperature range is 60−65 °C. 4.4. Effect of the Plasma Reaction Time on SO2 and NO Removal Rates. The plasma reaction time is the time taken for the reactive oxygen species to move from their duct injection point to the position of the electric acid mist remover (Figure 1, duct length 1 m). The influence of the reaction time on the SO2 and NO removal rates is shown in Figure 10.

Figure 11. Chromatogram for NO2, NO3, and SO42. The reaction time was 0.94 s, the inlet flue gas temperature is 65 °C, the molar ratio between the reactive oxygen species (O2+,O3) and SO2 and NO is 5, the flow rate was 1.2 m3/h, the concentration of NO and SO2 are 536 mg/m3 and 1142 mg/m3, the H2O volume fraction is 8%.

standard sample was analyzed for comparison (Figure 11(a)). The peak areas of NO3, NO2, and SO42 in the standard sample (Figure 11(a)) were 22.1735, and the corresponding concentrations were 10 mg/L. In the experimental sample, the peak areas of NO3, NO2, and SO42 (Figure 11(b)) were 20.08, 0, and 22.83, respectively. The NO3, NO2, and SO42 concentrations were calculated for this sample by dividing the peak areas in Figure 11(b) by those in Figure 11(a) and were 9.1, 0, and 10.3 mg/L, respectively. Based on the dilution factor (5000), the concentrations of NO3, NO2 and SO42 in the recovered acid in the storage tank were 45.5, 0, and 51.5 g/L, respectively. These results show that NO2 was not present in the recovered acid, which indicates that while HNO3 and H2SO4 were present in the recovered acid, HNO2 was not. This suggests that the ·OH radicals have sufficient strength to oxidize NO2 to HNO3, and SO2, and HSO3 to H2SO4. This result differs from that of Saavedra et al.,13 where DBD was used to remove NO from flue gas. They detected both HNO3 and HNO2 in the recovered acid, and the HNO2 content was 7.9−34.8% that of the HNO3. Recovery of these acids is important, because if they are released to the atmosphere they can photolytically degrade to NO and NO2 and cause secondary pollution. The recovery rates of HNO3 and H2SO4 were calculated using a mass balance of input and output flue gases with the formation of HNO3 and H2SO4. Theoretically, HNO3 and H2SO4 should be generated at 20 mg/min and 28.9 mg/min,

Figure 10. Effect of the reaction time on SO2 and NO removal rates. The H2O volume fraction is 8%, the inlet flue gas temperature is 65 °C, the molar ratio between the reactive oxygen species (O2+,O3) and SO2 and NO is 5, and the concentrations of SO2 and NO are 2285 mg/m3 and 603 mg/m3, respectively.

Increasing the reaction time from 0.19 to 1.52 s has little effect on the SO2 and NO removal rates. For these reaction times, the removal rate of SO2 was between 81.2% and 82.5%, and that of NO was between 96.5% and 97.2%. These results show that a reaction time of less than 0.19 s is sufficient for completion of the plasma reaction processes for simultaneous removal of SO2 and NO. The high reaction rate coefficients for the reactions of ·OH radicals with NO2, SO2 and HSO3 mean the entire plasma reaction reaches completion within 0.19 s. Taking into account the collision times of O2+ with H2O and ·OH radicals with SO2 and NO2, the optimal reaction time is about 1 s.29 Because the 10166

dx.doi.org/10.1021/es3013886 | Environ. Sci. Technol. 2012, 46, 10161−10168

Environmental Science & Technology

Article

(10) Wang, Z. H.; Zhou, J. H.; Zhu, Y. Q.; Wen, Z. C.; Liu, J. Z.; Cen, K. F. Simultaneous removal of NOx, SO2 and Hg in nitrogen flow in a narrow reactor by ozone injection: Experimental results. Fuel Process. Technol. 2007, 88, 817−823. (11) Wang, X. Q.; Li, Y.; Chen, W.; Lv, G. H.; Huang, J.; Zhu, G. X.; Zhang, X. H.; Wang, D. C.; Feng, K. C.; Yang, S. Z. Characteristics of NOx removal combining dielectric barrier discharge plasma with selective catalytic reduction by C3H6. Jpn. J. Appl. Phys. 2010, 49 (8), 086201(1)−086201(4). (12) Bai, M. D.; Yang, B.; Xue, X. H.; Liu, D.; Sun, J. Studies on the measuring method of number density of ions in the high speed flow field. IEEE Trans. Plasma Sci. 2009, 37 (11), 2257−2260. (13) Saavedra, H. M.; Pacheco, M. P.; Paceco-Sotelo, J. O.; Reyes, C. E. T.; Gomez, J. A. D. Modeling and experimental study on nitric oxide treatment using dielectric barrier discharge. IEEE Trans. Plasma Sci. 2007, 35 (5), 1533−1540. (14) Lowke, J. J.; Morrow, R. Theoretical analysis of removal of oxides of sulfur and nitrogen in pulsed operation of electrostatic precipitators. IEEE Trans. Plasma Sci. 1995, 23 (4), 661−671. (15) Person, J. C.; Ham, D. O. Removal of SO2 and NOx from stack gases by electron-beam irradiation. Radiat. Phys. Chem. 1988, 31 (1− 3), 1−8. (16) Demore, W. B.; Sander, S. P.; Golden, D. M.; Hampson, R. F.; Kurylo, R. F.; Howard, C. J.; Ravishankara, A. R.; Kolb, C. E.; Molina, M. J. Chemical kinetics and photochemical date for use in stratospheric modeling. Evaluation number12. JPL Publ. 1997. (17) Zhang, Z. T.; Bai, M. D.; Bai, M. D.; Bai, X. Y.; Pan, Q. Y. Removal of SO2 from simulated flue gases using non-thermal plasmabased microgap discharge. J. Air Waste Manage. Assoc. 2006, 56 (6), 810−815. (18) Wang, M. Y.; Zhu, T. L.; Wang, H. Oxidation and removal of NO from flue gas by DC corona discharge combined with alkaline absorption. IEEE Trans. Plasma Sci. 2011, 39 (2), 704−710. (19) Yin, S. E.; Sun, B. M.; Gao, X. D.; Xiao, H. P. The effect of oxygen and water vapor on nitric oxide conversion with a dielectric barrier discharge reactor. Plasma Chem. Plasma Process. 2009, 29 (6), 421−431. (20) Emission Standard of Air Pollutants for Thermal Power Plants. (GB13223-2011). http://www.mep.gov.cn/gkml/hbb/bgg/201109/ t20110921_217529.htm (accessed July 29, 2011). (21) Tseng, C. H.; Keener, T. C. Enhanced effect of in-situ generated ammonium salts aerosols on the removal of NOx from simulated flue gas. Environ. sci. Technlol. 2001, 35 (15), 3219−3224. (22) Lin, H.; Gao, X.; Luo, Z. Y.; Cen, K. F.; Huang, Z. Removal of NOx with radical injection caused by corona discharge. Fuel 2004, 83 (10), 1349−1355. (23) Motret, O.; Hibert, C.; Pellerin, S.; Pouvesle, J.M. Rotational temperature measurements in atmospheric pulsed dielectric barrier discharge. J. Phys. D, Appl. Phys. 2000, 33 (12), 1493−1498. (24) Granite, E. J.; Pennline, H. W.; Hoffman, J.S. Effects of photochemical formation of mercuric oxide. Ind. Eng. Chem. Res. 1999, 38 (12), 5034−5037. (25) Granite, E. J.; Pennline, H. W. Photochemical removal of mercury from flue gas. Ind. Eng. Chem. Res. 2002, 41 (22), 5470−5476. (26) Granitempm, E. J.; Freeman, M. C.; Hargis, R. A.; O’Dowd, W. J.; Pennline, H. W. The thief process for mercury removal from flue gas. J. Environ. Manage. 2007, 84 (4), 628−634. (27) Granitempm, E. J.; King, W. P.; Stanko, D. C.; Pennline, H. The implications of mercury interactions with band-gap semiconductor oxides. Main Group Chem. 2008, 7 (3), 227−237. (28) Qi, J. D. Modern Flue Gas Control Technology; Chem. Ind. Press, ISBN 978−7-122-01980-6 (in chinese). (29) Roth, J. R. Industrial Plasma Engineering, Vol. 2: Applications to Nonthermal Plasma Processing; Taylor & Francis. ISBN10:0750303174, ISBN-13:978-0750303170. (30) Podlinski, J.; Niewulis, A.; Mizeraczyk, J. Electrohydrodynamic flow and particle collection efficiency of a spike-plate type electrostatic precipitator. J. Electrostat. 2009, 67 (2−3), 99−104.

respectively, with removal rates of 97.2% (NO) and 83.2% (SO2). HNO3 and H2SO4 were actually generated at 10.6 mg/ min (HNO3) and 14.3 mg/min (H2SO4), as calculated from the concentrations of NO3 and SO42 in the ion chromatogram (Figure 11). The HNO3 and H2SO4 recovery rates (53% and 49%, respectively) were obtained by dividing the actual generation by the theoretical generation. The low HNO3 and H2SO4 recovery rates may be caused by the high-voltage direct current electric field in the electric acid mist remover (5 in Figure 1), which can efficiently remove large HNO3 and H2SO4 drops (ø >10 μm) but not the small drops (ø