Environ. Sci. Technol. 2003, 37, 2563-2567
Application of Pulsed Corona Induced Plasma Chemical Process to an Industrial Incinerator YONG-HWAN LEE,* WON-SUK JUNG, YU-RI CHOI, JONG-SEOK OH, SUNG-DUCK JANG, YOON-GYU SON, MOO-HYUN CHO, AND WON NAMKUNG School of Environmental Science and Engineering & Pohang Accelerator Laboratory, Pohang University of Science and Technology, San 31, Hyoja dong, Pohang 790-784, Republic of Korea DONG-JUN KOH Air Protection Research Team, Research Institute of Industrial Science and Technology, San 32, Hyoja dong, Pohang 790-784, Republic of Korea YOUNG-SUN MOK Department of Chemical Engineering, Cheju National University, Cheju, 690-756, Republic of Korea JAE-WOO CHUNG Department of Environmental Engineering, Jinju National University, 150 Chilamdong, Jinju, Kyungnam 660-758, Republic of Korea
Pulsed corona induced plasma chemical process (PPCP) has been investigated for the simultaneous removal of NOx (nitrogen oxides) and SO2 (sulfur dioxide) from the flue gas emission. It is one of the world’s largest scales of PPCP for treating NOx and SO2 simultaneously. A PPCP unit equipped with an average 120 kW modulator has been installed and tested at an industrial incinerator with the gas flow rate of 42 000 m3/h. To improve the removal efficiency of SO2 and NOx, ammonia (NH3) and propylene (C3H6) were used as chemical additives. It was observed that the pulsed corona induced plasma chemical process made significant NOx and SO2 conversion with reasonable electric power consumption. The ammonia injection was very effective in the enhancement of SO2 removal. NO removal efficiency was significantly improved by injecting a C3H6 additive. In the experiments, the removal efficiencies of SO2 and NOx were approximately 99 and 70%, respectively. The specific energy consumption during the normal operation was approximately 1.4 Wh/m3, and the nanopulse conversion efficiency of 64.3% was achieved with the pulsed corona induced plasma chemical process.
or dry scrubber for SO2 removal. The economical and simultaneous removal of SO2 and NOx still represents a significant technical challenge that could ultimately prevent the use of certain types of fossil fuels for energy production. Lately, alternative technologies have been developed including the application of pulsed corona discharge for the simultaneous removal of SO2 and NOx (1). Furthermore, pulsing also led to a higher overall electrical efficiency and is capable of producing the higher energy electrons. The possibility of removing SOx and NOx in the flue gas from the coal burning power plant by means of corona discharge was experimentally confirmed by Civitano at ENEL (2). The pulsed corona discharge process has been showing many encouraging results for the removal of gaseous pollutants based on small-scale experiments (3-5). Our previous experiment for the simultaneous removal of SO2 and NOx in the flue gas from an iron-ore sinter plant demonstrated the possibility of further scale-up to the practical application (6). However, there have been difficulties in applying pulsed corona discharge process to a large scale due to the lack of reliable high power nanopulse generator. Pulse modulator with a magnetic pulse compression (MPC) circuit is considered to satisfy long lifetime and high reliability requirements. It is pointed out earlier by E. L. Neau (7) that very fast rising (less than 100 ns rise time) repetitive (a few 100 Hz) short pulses (less than 500 ns full width half-maximum) with the peak voltage of approximately 200 kV and the peak current near 10 kA (this corresponds to the peak power of ∼2 GW per pulse) cannot be readily achieved with standard pulse generator using switches such as hydrogen filled Thyratron tube or spark gap switches. Spark gap switches are capable of handling such high power. However, due to the electrode erosion they bear problems in high repetition operation. MPC utilizes special reactors in conjunction with energy storage capacitors to compress input pulses into the narrower output pulses with the capability of handling very high peak current and voltage. The MPC is thought to be the only solution to obtain repetitive, short, and high peak power pulses. We have developed an average 120 kW pulse modulator and a plasma reactor for the simultaneous removal of SO2 and NOx in the flu gas exhausted from a large-scale industrial incinerator that has burning capacity of 30 000 ton/year. The plasma reactor has wire-plate type electrodes similar to the electrostatic dust precipitator and capable of treating the gas flow rate up to 50 000 Nm3/h at the gas flow speed of ∼1 m/s. This experiment has 2-fold objects; one is to find out the scale parameters for the SO2 and NOx removal efficiencies when it is applied to a large gas flow plant with highly fluctuating pollutants concentration during the normal operation, the other is to find out whether there is unforeseen technical issues which may jeopardize the reliability of high power nanopulse generator. This paper describes the removal characteristics of SO2 and NOx by PPCP at the incinerator plant including the test results and characteristics of the MPC type pulse modulator operated with the plasma reactor.
Experimental Section Introduction Sulfur dioxide (SO2) and nitrogen oxides (NOx, sum of NO and NO2) are major sources of acid rain and fine aerosols. NOx is also a primary precursor of photochemical smog. Currently many power stations and industries utilize selective catalytic reduction process (SCR) for NOx removal and wet * Corresponding author phone: +82-54-279-2758; fax: +82-54279-1191; e-mail:
[email protected]. 10.1021/es0261123 CCC: $25.00 Published on Web 04/18/2003
2003 American Chemical Society
Pulse Power Supply for PPCP. In many environmental applications, the key feature of plasma process is that it generates a large flux of energetic electrons, which can then be used to initiate or enhance a variety of chemical, biological, and physical reactions. The electrons created by the positive high voltage pulse discharge have higher energy than the energy obtained from DC or AC discharges. The discharge has a large number of intense streamers: thin plasma channels propagating in a discharge gap. Radicals relevant VOL. 37, NO. 11, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2563
FIGURE 1. Simplified equivalent circuit diagram of the MPC modulator. to the chemical processes are produced in the regions of high electric fields in the streamer head. For the generation of the intense streamers, it is necessary to develop the high power pulse generator that will allow an efficient generation of energetic electrons. In the appropriate conditions, the major part of electrical energy is transferred to electrons without heating the ambient gases in the reactor. For this experiment, we built a pulse modulator using the MPC (Magnetic Pulse Compression) circuit. This system utilizes a two-stage MPC that can generate 150 kV pulses with 500 ns (fwhm) pulse width and the 300 Hz repetition rate. The simplified equivalent circuit diagram for the two-stage MPC is shown in Figure 1. It can be subdivided in several sections such as a DC power supply, a semiconductor switch assembly, a pulse transformer, magnetic switching sections (MS1 and MS2), and the reactor (CL and RL). Four units of average 33 kW high voltage inverter power supplies provide DC high voltage with 0.5% fine regulation for the charging of the energy storage capacitor C1. The energy stored in C1 is transferred to C2 by a semiconductor switch (T1 and D1), and pulses are compressed by MS1 and MS2 magnetic switches sequentially. During the charging up period of C2, MS1 acts as an open switch till the core saturates. When the core of MS1 saturates, the circuit inductance drops down to a very small value that is equivalent to a closed switch. Then, the current flows through MS1 and charges C3. When the core of MS2 saturates, the energy stored in C3 is delivered to the reactor. The pulse corona reactor can be expressed as a capacitor before the corona onset and represented as a nonlinear resistor in parallel with a varying capacitor after corona onset (8, 9). To keep the streamer corona from transforming to an arc, duration of the high voltage pulse must be restricted to less than 1 µs (10). This pulse modulator has been tested with a plasma reactor for an industrial incinerator. The geometric capacitance of the reactor before the corona onset is 10 nF which corresponds to CL in the Figure 1. Figure 2 shows the voltages measured at CVD1, CVD2, and CVD3 indicated in Figure 1. Figure 3 shows the corresponding current waveforms measured at CT2, CT3, CT4, and CT5. The peak voltage and current reach up to 119 kV and 5.01 kA, respectively, which correspond to a peak power level of 596 MW. The pulse width (fwhm) is about 700 ns, and the voltage rise time is about 68 ns. When a high voltage is applied to the discharging electrode, the first capacitive current flows to charge the reactor, and the corona discharge occurs when the voltage passes over the corona onset value. At this time, one can observe a voltage drop caused by the reduction in gas 2564
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 11, 2003
FIGURE 2. Voltage waveforms.
FIGURE 3. Current waveforms. resistance between the discharging electrode and the grounded plate. In a high repetition rate system, even small increase of the energy efficiency is very important. The energy efficiency in the MPC loop is defined as the ratio between the accumulated energy in the final capacitor and the initial
Experimental Conditions and Analyses. We analyzed the inlet and outlet concentrations of SO2 and NOx with portable gas analyzers (Green MK2, Eurotron Italiana s.r.l). The actual flow rate of the flue gas exhausted from the incinerator is 42 000 m3/h at 170 °C. The energy density defined as the ratio of the electric power divided by the flow rate was analyzed to be 1.4 Wh/m3 from the measurement of the discharge power. The peak voltage and the pulse repetition rate were 160 kV and 270 pps, respectively.
Results and Discussion Pulsed corona discharges generate energetic electrons. The low-energy electrons are accelerated as they drift along the high voltage region (corona region) until they collide with a gas molecule and immediately lose energy by excitation, attachment, dissociation, collisions, or ionization. After transferring its energy to molecules, the electron is reenergized by the electrical field. Pulsed corona discharges have high power efficiencies in producing useful radicals such as O, OH, HO2, N, etc. It is reported that O, OH, and HO2 radicals contribute to oxidation process, and N radical contributes to reduction process (11-13). The reactions and the rate constants of NOx and SO2 with radicals are as follows:
SO2 + OH f HSO3
k1 ) 5.0 × 10-31 (T/300)-3.3 [M] (2)
HSO3 + OH f H2SO4 SO2 + O f SO3
FIGURE 4. Picture of PPCP reactor for an industrial incinerator plant. energy in the first capacitor bank. The energy delivered per pulse was found to be about 213.5 J. When the capacitor C1 is charged to the voltage of VC, the stored energy is expressed as follows: 2
E ) 0.5C1Vc
k4 ) 6.0 × 10-15
(5)
k5 ) 1.0 × 10-31 (T/300)-1.6 [M]
(6)
NO + HO2 f NO2 + OH k6 ) 3.7 × 10-12 exp(-240/T) (7) NO + OH f HNO2
k7 ) 7.4 × 10-31 (T/300)-2.4 [M] (8)
HNO2 + OH f NO2 + H2O
k8 ) 1.8 × 10-11 exp(-390/T) (9)
(1)
Since the value of C1 is 2 µF and VC is 18.2 kV, the energy stored is equal to 332 J, and the energy transfer efficiency of the MPC modulator becomes 64.3%. Actually, the high voltage inverter power supply also has some loss in the power conversion. Considering this loss, the overall energy transfer efficiency from the wall plug to the reactor load is approximately 50%. Plasma Reactor. Figure 4 shows the picture of PPCP installed at an industrial incinerator. The reactor is composed of two series reactors. The first one is the pulsed corona reactor to induce reactions of SO2 and NOx with radicals, and the second one is the electrostatic precipitator to collect byproducts such as ammonium sulfate and ammonium nitrate (white powders). The electrostatic precipitator uses a negative pulse voltage superimposed on DC voltage. Main parameters of the corona reactor are shown in Table 1. This reactor is designed to be able to treat the gas flow rate of 50 000 Nm3/h from the incinerator. The reactor structure is identical to a typical electrostatic precipitator. It has 10 gas passages composed of 12 discharging electrodes per frame and 11 collecting plates. The height and length of each collecting frame are 4096 mm and 3964 mm, respectively, and the gap distance between two adjacent plates is 300 mm.
(3)
k3 ) 4.0 × 10-32 exp(-1000/T) [M] (4)
SO3 + H2O f H2SO4 NO + O f NO2
k2 ) 9.8 × 10-12
NO2 + OH f HNO3
k9 ) 2.6 × 10-30 (T/300)-2.9 [M] (10)
SOx Removal Characteristics in PPCP. O, OH, and HO2 radicals produced in the pulsed corona reactor can convert SO2 into sulfuric acid (8, 9), and the sulfuric acid can be neutralized by ammonia. Figure 5 shows the SO2 removal characteristics in PPCP. First, we investigated the plasma process effect on the SO2 removal without NH3 addition. Previous studies reported that pulsed corona did remove SO2 without NH3 addition (9, 14). In the experimental conditions of 160 kV and 270 pps, approximately 25% of SO2 was removed in the pulsed corona reactor without NH3 addition. The condition of exhaust gases from the incinerator was very unstable, and the removal efficiency was also changed a little bit even though the operating conditions were the same. Generally, 10-25% of SO2 was removed in PPCP without NH3 injection. To improve the removal efficiency of SO2, ammonia (NH3) was added to the flue gas as a neutralizing chemical additive. As shown in Figure 5, the removal of SO2 was mainly affected by NH3 additive. As soon as NH3 was added to the flue gas, the concentration of SO2 dramatically dropped. In the presence of 370 ppm of NH3, the inlet SO2 concentration of 100-200 ppm decreased to almost 0 ppm. This result corresponds to the removal VOL. 37, NO. 11, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2565
TABLE 1. Design Parameters of Pulsed Corona Induced Plasma Chemical Process gas flow rate (Nm3/h) gas velocity (m/s) gas temperature (°C) gas passage (GP) no.
50 000 1.2-1.4 193 10
total wire length (m) reactor size, W × H × L (m) wire diameter (mm) distance between plates (cm)
FIGURE 6. Effect of propylene on NOx removal.
FIGURE 5. Effect of ammonia on SO2 removal. efficiency of 99% for SO2. Although the pulsed corona process promotes the removal of SO2, such a high removal efficiency seems to be caused by a chemical reaction between SO2 and NH3 in PPCP. It was also observed that most of SO2 was removed without the pulsed corona process as shown in the points between 3 and 4. The result indicates that SO2 can be removed effectively in the presence of a proper amount of NH3. Even though the reaction mechanism of SO2 in the pulsed corona process can be explained by the reactions 2-5, it can be changed by addition of ammonia. Ammonia (NH3) is widely used as a chemical additive in the pulsed corona method. Reactivity of SO2 with NH3 is higher than that with radicals produced in the reactor. Therefore, most of SO2 reacts with NH3 to form ammonium salts (11) as shown in the reactions 11 and 12.
SO2 + NH3 + H2O + 1/2O2 f NH4HSO4 (s)
(11)
SO2 + 2NH3 + H2O + 1/2O2 f (NH4)2SO4 (s) (12) NOx Removal Characteristics in PPCP. As shown in previous results, SO2 can be easily removed by NH3. However, NOx removal requires a large amount of energy. Therefore, one of the main problems to be solved is to minimize the energy consumption. One possible way of reducing the energy consumption is the injection of proper additive into PPCP. It is previously shown that the NO conversion rate can be improved significantly when the hydrocarbon gas is added (15). The reactions of hydrocarbons such as propylene and ethylene with OH, O, and O3 can produce R• (alkyl radical), RCO• (acyl radical), and RO• (alkoxy radical) as shown in the following reactions (16).
C2H4, C3H6 + O f R• + RCO•
(13)
C2H4, C3H6 + OH• f R•
(14)
C2H4, C3H6 + O3 f RO• + RCO• + aldehyde
(15)
Radicals produced from reactions of hydrocarbons with OH, O, and O3 can increase the NO conversion efficiency into NO2. Once NO is converted to NO2, it can be removed 2566
9
480 3.2 × 6.2 × 4.5 10 300
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 11, 2003
by NH3 injection as shown in the reaction 16. The role of hydrocarbons in NOx removal is well described in the previous study (15).
2NO2 + 2NH3 f N2 + H2O + NH4NO3 (s)
(16)
Figure 6 shows the NOx removal characteristics in PPCP. Most of NOx produced from the incinerator was in the form of NO. That is to say, NO2 production was negligible in the incinerator. Propylene (C3H6) was added to improve the removal efficiency of NOx from the flue gas. As shown in Figure 6, the removal of NO was increased significantly when propylene was added. For the case without propylene, the NO concentration was slightly decreased even at a high energy density. The NO removal can be explained by the oxidation of NO to NO2 in the pulsed corona reactor. Approximately 5% of NO was removed without propylene. On the other hand, the input NOx concentration of 70-75 ppm was decreased to 20 ppm when propylene was added to the gas stream. This result corresponds to the removal efficiency of approximately 70% for NOx on the basis of the initial concentration. When the plasma (pulsed corona discharge) was turned off, the NO concentration at the outlet increased to 54 ppm. The removal efficiency was approximately 20% in the point between 4 and 5. It means that propylene addition without the plasma has little effect on the removal of NOx. In the point between 4 and 5, both NH3 and propylene were injected to the reactor without plasmas. Generally, NH3 does not contribute to remove NOx in the pulsed corona reactor. The previous study indicated that the NOx removal is not nearly affected by NH3 (17). It is believed that C3H6 mainly affects the NO removal in the point between 4 and 5. Although the increase of the C3H6 concentration gives rise to an enhancement of the NOx removal, it should be limited to a proper amount because of C3H6 slip. In the experiments, the concentration of C3H6 was limited to 55 ppm. However, it is also believed that the NO removal efficiency can be greatly improved by using a proper amount of C3H6 additive. The concentration of NH3 used as an additive should also be considered. In the experiments, the average concentrations of SO2 and NOx were 150 and 73 ppm, respectively. Stoichiometric equivalent concentrations of NH3 are 150 ppm for SO2 and 146 ppm for NOx. The concentration of NH3 injected to the pulsed corona reactor was 370 ppm, corresponding to 1.25 times the stoichiometric equivalence. Even though we did not measure the concentration of NH3 at the
outlet of the plasma process, it is assumed that a small amount of ammonia slip enters the exhaust stream. Ammonia slip must be limited due to downstream impacts associated with corrosion and fouling. Therefore, PPCP should be set to minimize the amount of ammonia slip. We are making experiments to find out the optimal concentration of ammonia without NH3 slip. Cost Comparison. S. Masuda made an economical comparison between PPCP and other DeNOx/DeSOx technologies, and he reported that PPCP would provide the most cost-effective means of DeNOx/DeSOx as the future technology (18). PPCP was compared with E-beam and calciumgypsum process combined with ammonia catalytic process in the report. In the comparison of annual cost, the cost of PPCP was lower than that of the combined process by 17%. In his report, the ratio of investment cost for PPCP was approximately 39%, and the cost for pulse generator accounts for a large portion. The calculation of the power consumption for the removal of SO2 and NO was carried out in the experiments. The estimation is based on the rate in Korea. The power consumption required for the removal of SO2 and NO was 6.22 kWh/kgSO2 and 53.4 kWh/kgNO, respectively. The values correspond to the energy cost of $0.31 and $2.67 for removing 1 kg of SO2 and NO, respectively. The costs of NH3 and C3H6 required for the removal of SO2 and NO were $0.29 /kgSO2 and $1.95 /kgNO, respectively. Therefore, the total costs required for the removal of SO2 and NO were $0.6 /kgSO2 and $4.62 /kgNO, respectively. A further economical advantage of PPCP will be achieved by developing reliable and cheaper pulse power supply. This paper reports a full-scale industrial application of the pulsed plasma chemical process for the simultaneous removal of SOx and NOx in the flue gas. Our next goals in the test operation are to demonstrate the reliability of main HV (high voltage) components and the MPC system for an industrial application. Also, further study is ongoing to find out the removal mechanisms of SO2, NOx, CO, and HCl including the study on effective means of collecting the dust produced in the large-scale pulsed corona induced plasma chemical process.
and Technology (RIST), Korea Ministry of Science and Technology (MOST), and Korea Institute of Environmental Science and Technology (KIEST).
Acknowledgments
Received for review August 31, 2002. Revised manuscript received January 25, 2003. Accepted March 24, 2003.
This work is partially supported by Pohang Iron and Steel Company (POSCO), Research Institute of Industrial Science
Literature Cited (1) Tseng, C.-h. Air & Waste Management Association’s 92nd Annual Meeting & Exhibition, St. Louis, MO, U.S.A. 1999, Poster Number: DP015. (2) Dinelli, G.; Civitano, L.; Rea, M., IEEE Trans. Ind. Applicat. 1990, 26, 535. (3) Bhasavanish, D.; Ashby, S.; Deeney, C.; Schlitt, L. Ninth IEEE Pulsed Power Conf. 1993, 441. (4) Mizuno, A.; Clements, J. S.; Davis, R. H. IEEE Trans. Ind. Applicat. 1986, 22(3), 516. (5) Gallimberti, I. Pure Appl. Chem. 1988, 60(5), 663. (6) Mok, Y. S.; Nam, I. S. IEEE Trans. Plasma Sci. 1999, 27(4), 1188. (7) Neau, E. L.; Harjes, H. C.; et al. Non-Thermal Plasma Techniques for Pollution Control: Part A; NATO ASI Series, Berlin, Germany, Springer-Verlag: 1993; p 367. (8) Mok, Y. S.; Ham, S. W.; Nam, I. Plasma Chem. Plasma Proc. 1998, 18(4), 535. (9) Dinelli, G.; Civitano, L.; Rea, M. IEEE Trans. Ind. Applicat. 1990, 26, 535. (10) Lowke, J. J.; Morrow, R. IEEE Trans. Plasma Sci. 1995, 23, 661. (11) Civitano, L.; Non-Thermal Plasma Techniques for Pollution Control: Part B; NATO ASI Series, Berlin, Germany, SpringerVerlag: 1993; p 103. (12) Dahiya, R. P.; Mishra, S. K.; Veefkind, A. IEEE Trans. Plasma Sci. 1993, 21, 346. (13) Amirov, R. H.; Asinovsky, E. I.; Samoilov, I. S.; Shepelin, A. V. Non-Thermal Plasma Techniques for Pollution Control: Part B; NATO ASI Series, Berlin, Germany, Springer-Verlag: 1993; 149. (14) Clements, J. S.; Davis, R. H.; Finney, W. C.; Mizuno, A. IEEE Trans. Ind. Applicat. 1989, 25(1), 62. (15) Mok, Y. S.; Nam, I.; Chang, R. W.; Ham, S. W.; Kim, C. H.; Jo, Y. M. Proc. 7th Int. Conf. Electrostatic Precipitation; Hilton, Kyungju, Korea, 1998; p 270. (16) Seinfeld, J. H. Air Pollution: Physical and Chemical Fundamentals; McGraw-Hill: New York, 1975. (17) Mok, Y. S.; Nam, I. IEEE Trans. Plasma Sci. 1999, 27(4), 1188. (18) Masuda, S. Non-Thermal Plasma Techniques for Pollution Control: Part B; NATO ASI Series, Berlin, Germany, SpringerVerlag: 1993; p 131.
ES0261123
VOL. 37, NO. 11, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2567
