Application of Acoustic Agglomeration to Reduce Fine Particle

Jun 26, 1997 - ASINEL, Apartado 233-28930 Móstoles, Madrid, Spain. Removal of fine particles (smaller than 2.5 μm) from industrial flue gases is, at p...
0 downloads 0 Views 250KB Size
Download PDF
Environ. Sci. Technol. 1999, 33, 3843-3849

Application of Acoustic Agglomeration to Reduce Fine Particle Emissions from Coal Combustion Plants J U A N A . G A L L E G O - J U AÄ R E Z , * ENRIQUE RIERA-FRANCO DE SARABIA, G E R M A N R O D R IÄ G U E Z - C O R R A L , THOMAS L. HOFFMANN,# AND J U A N C . G AÄ L V E Z - M O R A L E D A Instituto de Acu ´ stica, CSIC, Serrano 144, 28006 Madrid, Spain J E S U S J . R O D R IÄ G U E Z - M A R O T O , FRANCISCO J. GO Ä MEZ-MORENO, ALBERTO BAHILLO-RUIZ, AND M A N U E L M A R T IÄ N - E S P I G A R E S CIEMAT, Av. Complutense 22, 28040 Madrid, Spain MIGUEL ACHA ASINEL, Apartado 233-28930 Mo´stoles, Madrid, Spain

Removal of fine particles (smaller than 2.5 µm) from industrial flue gases is, at present, one of the most important problems in air pollution abatement. These particles, which are hazardous because of their ability to penetrate deeply into the lungs, are difficult to remove by conventional separation technology. Sonic energy offers a means to solve this problem. The application of a high-intensity acoustic field to an aerosol induces agglomeration processes which changes the size distribution in favor of larger particles, which are then easier to precipitate with a conventional separator. In this work, we present a semiindustrial pilot plant in which this process is applied for reduction of particle emissions in coal combustion fumes. This installation basically consists of an acoustic agglomeration chamber with a rectangular cross-section, driven by four high-power and highly directional acoustic transducers of 10 and/or 20 kHz, and an electrostatic precipitator (ESP). In the experiments, a fluidized bed coal combustor was used as fume generator, and a sophisticated air sampling station was set up to carry out measurements with fume flow rates up to about 2000 m3/h, gas temperatures of about 150 °C, and mass concentrations in the range 1-5 g/m3. The fine particle reduction produced by the acoustic filter was about 40% of the number concentration.

Introduction Removal of fine particles smaller than 2.5 µm from industrial flue gases is one of the most important problems in air pollution control. In fact, these particles constitute a major health hazard because of their ability to penetrate deeply into the respiratory system. In addition, these particles are * Corresponding author phone: +34-91.561.88.06; fax: +3491.411.76.51; e-mail: [email protected]. # Present address: Fraunhofer Technology Center Hialeah, 601 W 20th Street, Hialeah, FL 33010. 10.1021/es990002n CCC: $18.00 Published on Web 09/29/1999

 1999 American Chemical Society

generally hazardous due to their origin (condensation of chemically active elements produced during combustion processes) and their capability to act as a vehicle for noxious agents (e.g. viruses, bacteria, dioxin). Lately, a renewed interest has arisen in the control of such fine particle emission from industrial combustion plants, reflected in new more stringent legislation which has been introduced in the U.S. (1). Fine particles make up a significant percentage of the particulate emission in coal burning processes such as coal fired power plants, in the steel-melting industry (open-hearth and converter), and in the cement industry. At the same time, the retention efficiency of conventional filters such as electrostatic precipitators (ESPs) drops off steeply for particles under 2 or 3 µm (2). As a consequence, an improved technology is needed to obtain higher fine-particle retention efficiencies. By forming larger particles, agglomeration of micron and submicron particles is a promising process to be used for preconditioning of the particles before they enter into a conventional filtration system. The preconditioned aerosol can then be filtered with a higher collection efficiency. Particle agglomeration by means of acoustic and electric fields have both been investigated. Agglomeration studies and experiments with alternating electric fields were recently carried out by Eliasson et al. (3), Gutsch and Lo¨ffler (4), and Kildesø et al. (5). The general principle of this technique was to split the particle flow into two different streams, charge the particles of the streams electrically opposite, and mix the charged particles back into a single stream where they are agglomerated. The main problem of the method seemed to be the difficulty to charge the fine particulates, which were the ones that needed to be agglomerated. Nevertheless some authors provided promising results with this technique. Acoustic agglomeration is another procedure to be considered for aerosol preconditioning. High-intensity acoustic fields applied to an aerosol may induce interaction effects among suspended particles, giving rise to collisions and agglomerations. The principle of acoustic agglomeration is primarily based on relative motion between suspended particles of different sizes which promote particle collisions (orthokinetic effect) (6). However, particles may also agglomerate due to the action of hydrodynamic forces resulting from mutual distortions of the flow fields around particles (hydrodynamic effect) (7-11). Acoustic agglomeration has been widely studied, but the development of this process into an industrial application has been slow. This is probably because of the lack of suitable high-intensity, high-efficiency sound sources and appropriate full scale agglomeration chambers. Economical reasons have also been blamed. Scott (12) investigated these problems by comparing progressive saw-tooth waves with conventionally used standing waves for acoustic agglomeration applications. He claimed improvements through saw-tooth excitation which supported his proposal for deployment of resonant pulse-jets as acoustic sources for sonic agglomeration. Nevertheless, the tests carried out at the Ontario Research Foundation did not lead to subsequent industrial developments. For many years, the Power Ultrasonics group at the Instituto de Acu´stica in Madrid has worked to overcome these problems (13-15). This paper presents the characteristics and performance of a new pilot scale, multifrequency acoustic agglomerator which was tested in connection with an electrostatic precipitator for the treatment of fumes from a 0.5 MW fluidized bed coal combustor. The novelty of this agglomerator concept is the deployment of a new type of macrosonic generator (16) which implements high-power VOL. 33, NO. 21, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3843

FIGURE 1. Scheme of the experimental facilities. capacity, efficiency, and directivity to allow for treatment of large gas volumes. The acoustic chamber is designed to optimize the spatial homogeneity and intensity of the acoustic field distribution. In addition, the new acoustic agglomerator allows for different frequencies along the path of the fume through the system. This feature represents an important advance because the frequency of the sound field is linked to the particle size which changes during the agglomeration process. The chamber dimensions are determined from the residence time requirements and the given flow rate of the aerosol to be treated. The acoustic agglomerator was studied in terms of its acoustic characteristics as well as its filtration performance. The improvement of the fine particle retention efficiency of the electrostatic filter obtained by acoustic preconditioning of the particles was evaluated by means of a sampling station to measure particle mass and number concentrations as well as particle size distributions.

Experimental Facilities The experimental facilities basically consist of a circulating fluidized bed combustion plant (0.5 MW), a conventional electrostatic precipitator (ESP), and the acoustic agglomeration chamber. All the components are connected through air ducts in which the test aerosols are controlled and measured (Figure 1). The facility includes three sampling points (SPs) along the air duct: SP1 upstream of the acoustic chamber, SP2 between the acoustic chamber and the ESP, and SP3 downstream of the ESP. A mobile sampling station permits the operator to perform a complete characterization of the aerosols along the process line. The acoustic agglomeration chamber has a length of 3.56 m and a rectangular cross-section of 0.7 × 0.5 m2. Four specially designed steppedplate, high-intensity transducers are arranged along the elongated side of the agglomeration chamber to achieve a homogeneous distribution of the sound field as well as sufficient residence time of the aerosol in the chamber (2-3 s). The ESP’s main characteristics are listed in Table 1. The fume emissions of the combustion plant (approximately 140 m3/h) were diluted with air previously heated in a propane burner to reach a total flow rate of approximately 1600 m3/h. The diluted fume is treated in the acoustic agglomeration chamber before it enters into the ESP. A scrubber unit is placed downstream of the ESP as an absolute filter for environmental safety. It should be noticed that the values of the gas volume are given at 0 °C and 1 atm. 3844

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 21, 1999

TABLE 1. Specifications of the Electrostatic Precipitator no. of mechanical fields in series no. of electrical fields total active length number of streets street width projected precipitation surface transformer-rectifier input voltage input current current frequency output voltage output current total power consumption

2 1 3.0 m 4 200 mm 33.6 m2 380 V 4A 50 Hz 35 kV 25 mA 1.5 kW

Requirements for Efficient Acoustic Preconditioning. Previous theoretical and experimental studies have shown that for the sound pressure levels (SPL) applied in an acoustic agglomerator the rule “more-is-always-better” holds almost without constraint. For this reason, powerful sound sources for the treatment of large gas volumes at high sound pressure levels are indispensable for an efficient acoustic preconditioning system. Another factor that strongly influences the resulting sound field patterns is the geometry of the agglomeration chamber. Here special attention has to be given to optimize for best possible homogeneity and maximum spatial mean level within the chamber. Moreover, the frequency of the exciting sound field determines in which particle size range the preconditioning system works most efficiently. Whereas lower frequencies in the audible range favor the agglomeration of micron sized particles, higher frequencies in the ultrasonic range deliver better results in the submicron range. Therefore, the sound frequency has to be adjusted in a way that the performance of the acoustic preconditioning process is optimized with respect to a given aerosol size distribution. This becomes even more important when bimodal or multimodal size distributions are considered (as apparent in many fly ash distributions after coal combustion). The efficiency of the agglomeration process is also influenced by parameters such as temperature and pressure, residence time in the agglomerator, and aerosol concentration. Experiments show that the latter two quantities are crucial for efficient acoustic preconditioning (13, 14). Characteristics of the Acoustic Agglomeration Chamber. As mentioned before the agglomeration chamber is designed

FIGURE 2. Macrosonic stepped radiator. with a rectangular cross-section and is axially traversed by the flow of aerosols to be agglomerated. The high-power acoustic transducers, mounted along the bottom wall, have their radiating surfaces aligned parallel to the opposing wall, which serves as a reflector for the generation of standing wave patterns (15). The transducer support features an adjustment device to permit variations of the radiatorreflector distance to achieve an optimum stationary field. High-Power Transducers. Two different models of stepped-plate transducers permit the operation of the acoustic preconditioner at sonic (10 kHz), ultrasonic (20 kHz), or mixed frequencies (10 and 20 kHz) (see Figure 2). The radiating plate of these transducers is used to increase the radiation impedance and the power capacity. The directivity is controlled through steps in the plate geometry. In this application, we worked with axisymmetrically vibrating stepped-plates with five and seven nodal circles for 10 and 20 kHz and plate diameters from 48 to 67 cm. The transducers have beamwidths of 1.5° at 3 dB, electroacoustic efficiencies of about 80%, and maximum power capacities of about 1 kW. The electronic driving system is designed to operate continuously at the resonant frequency of the transducer. The device is based on an analogue type oscillator assembly, consisting of a power amplifier which is feedback by the transducer itself by means of a tuned bridge circuit, a phase shifter, and a band-pass filter (16). Heating problems of the

transducer radiating surfaces due to the high temperatures of the gas flow are avoided through an air cooling system in which an air layer between the radiators and the fume flow provides a temperature gradient between the transducer surface and the flow itself (15). Acoustic Characteristics of the System. A high-intensity acoustic standing wave field was established within the agglomeration chamber to achieve a good balance between the applied energy and the volume of aerosol to be treated. The acoustic field inside the agglomerator was studied in the laboratory with a model of the agglomeration chamber, made of PVC-glass with the same dimensions as the actual chamber. As shown in Figure 3, four transducers were located at the bottom of the chamber to generate an intense acoustic standing wave field in the direction perpendicular to the gas flow. The dimensions of the chamber were chosen in a way that the treated gas flow always passed through the near field of the transducers. The acoustic field inside the agglomeration chamber was evaluated by measuring the sound pressure on a large number of cross-sectional x-y planes. The measurements were performed with a 1/8” B&K microphone following x-y raster scans. The sound pressure distribution within each plane was calculated by averaging for each x-coordinate all acoustic pressure values measured along a line parallel to the y-axis. Finally, a representative spatial mean pressure distribution for the whole chamber was obtained by averaging the sound pressure distributions of all the measured planes along the x-axis. Figure 4 shows, as an example, the profile of the spatial mean sound pressure level (SPL) generated by four transducers, distributed over the length of the chamber in a two-frequency setup (10 and 20 kHz) with an applied input power of 400 W per transducer. The maximum measured SPLs were higher than 165 dB, while the minimum values stayed above 145 dB. These levels are sufficient for efficient acoustic agglomeration. In the agglomeration tests presented in this paper, the chamber was operated at two different frequencies (10 and 20 kHz) in separate experiments. Three different acoustic field distributions were applied at each frequency. First, four transducers were operated along the entire length of the chamber with an applied power of 400 W of each transducer (4T 400 W). Second, two transducers, covering one-half of the chamber length, were operated with an applied power of 400 W of each transducer (2T 400 W). And third, four transducers were operated along the entire length of the chamber with an applied power of 80 W of each transducer (4T 80 W). Sampling Station. A particle sampling and measurement system was set up to characterize the suspended particles in the combustion gases. The system was capable of measuring online the number concentration and particle size distribu-

FIGURE 3. Acoustic agglomeration chamber. VOL. 33, NO. 21, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3845

FIGURE 6. Mass and number particle size distributions measured with three different instruments. The cascade impactor data were obtained for a particle density of 1 g/cm3 and a particle shape factor of 1. FIGURE 4. Mean SPL profile of the macrosonic field along the axis of the agglomeration chamber (CW excitation).

FIGURE 5. Schematic diagram of the sampling and measurement station. tion in the micron and submicron range. In addition, mass concentration, particle morphology, and chemical composition were obtained. Figure 5 schematically depicts this sampling station (17). The aerosol is first extracted from the air duct and then led to the sampling station through an isokinetic probe equipped with an automatic control unit. It was also possible to dilute the sample with dry, filtered air to avoid overloading of the measuring devices. Several instruments were used to perform simultaneous measurements, covering the entire particle size range of the aerosol. For the micron size range an optical on-line particle sizer was used, functioning on the principle of white light scattering on single particles (18). This instrument provided number concentration and size distribution measurements in the range 0.3-30 µm, with a maximum allowable particle number concentration of 105 particles/cm3. After diluting the sample, the flow was split into four separate streams using an isokinetic flow splitter. These streams were then conducted to different measurement instruments. Information on the number concentration and size distribution of submicron particles (0.005-0.5 µm) was obtained with a TSI Scanning Mobility Particle Sizer (SMPS) consisting of a differential electrical mobility analyzer coupled with a condensation nuclei counter (19-21). Because of the limitations imposed by this instrument, the sample has to be preconditioned before entering into the SMPS. A cyclone and a preimpactor 3846

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 21, 1999

removed particles larger than 1 µm before the measurement. The sample was then drawn into a chamber from which a suitable amount of sample was taken and conducted to the SMPS. This procedure was necessary because the sampling flow rates of the isokinetic probe were much larger than the maximum flow rate allowed for the SMPS. A special device was designed to collect particle samples deposited on glass slides. The aerosol sample was drawn into a cylindrical chamber which supported the slides. The particles were deposited on the slides by impaction, sedimentation, and diffusion. Subsequently, the samples were analyzed by scanning electron microscopy (SEM). Information on morphology and size distribution was obtained from image analysis, while the elemental composition of particles was determined by electron probe X-ray microanalysis (EPXMA). In addition, the aerodynamic particle size distribution was measured with an eight stage cascade impactor (22) using glass fiber paper substrates and a backup filter. Additionally, 47 mm glass fiber paper filters were used to measure total mass concentration of the aerosol. The filters and substrates were gravimetrically and chemically analyzed with different techniques such as atomic absorption spectrometry (AAS), inductively coupled plasma atomic emission spectrometry (ICP-AES), and inductively coupled plasma mass spectrometry (ICP-MS) (20). The flow rate through the different sampling instruments was controlled and measured with needle valves and mass flow-meters. After passing through the measuring and sampling devices the aerosol sample was filtered and drawn through a flow rate control unit to adjust and monitor for isokinetic sampling conditions.

Results and Discussion Some initial characterization tests were carried out to verify the performance of the combustion plant and the sampling and measurement station. These tests included measurements of the mass and number particle size distribution and the concentration and served for comparison of the results of the different instruments within the range of operation for each one of them. Figure 6 shows the results of one of these measurements. It is found that the results of two different measurement techniquesscascade impactor and optical instrumentscoincide in detecting the peak of the particle distribution around 1 µm in terms of mass and number concentration, respectively. Conversely, it should be noted that a mode observed with the cascade impactor around 25 µm could not be resolved by the optical instrument.

TABLE 2. Input Conditions of the Diluted Wasteair 100-200 °C 1200-1600 m3/h 5 g/m3 7-10 m/s

gas temperature duct flow rate initial particle mass concn duct gas velocity composition of gas (dry) O2 CO2 N2 CO NOX SO2 HC

6.0% 13.7% 80.2% 105 ppm 80 ppm 900 ppm 2 ppm

TABLE 3. Text Matrix and Summary Results operating conditions of the acoustic agglomeration chamber no. of transducers and power per test transducer (W) 1 2 3 4 5 6

4T 400 2T 400 4T 80 4T 400 2T 400 4T 80

total power/ flow rate (W h/m3) 1 0.5 0.2 1 0.5 0.2

reduction of the particulate matter freq mass particle no. (%) (kHz) (%) micron submicron 10 10 10 20 20 20

32 29 26 37 29 24

37 33 26 42 31 14

39 21 16 39 31 24

Note: During all experiments, the ESP voltage was fixed at 30 kV. The total power/flow rate value for the ESP remained at 1.5 kW/1600 m3/h ) 0.94 W h/m3.

With the different systems in proper working order, the performance of the agglomeration-retention process was studied under different acoustic conditions. The operational conditions at the inlet of the air-diluted fume are listed in Table 2. During the experiments, the aerosol was characterized only at the sampling point SP3, because of the high costs of the sampling station it was not possible to use two units. Also, moving the sampling station from SP3 to SP1, to measure downstream and upstream, was not practical because of calibration problems. Consequently, the methodology employed was to compare the measurements made at SP3 with the acoustic chamber turned “off” or “on”. Each experiment was carried out in a time period of about 8 h. During this time, three or four pairs of trials were conducted by turning the acoustic chamber “on” or “off”. The results presented in this paper are mean values of all measurements from selected pairs of trials. The trials were selected with the goal to filter out occasional instabilities of the combustion plant. A summary of the results obtained with the acoustic chamber is shown in Table 3. In this table the column “reduction of the particulate matter” includes values for reduction in terms of mass and number concentration. The subcolumn “mass” indicates the additional percentage of particle mass retained by the ESP when the acoustic preconditioner is operational with respect to the particle emission when ESP functions alone. The reduction of the particle number concentration is also included in Table 3 for the micron and submicron size ranges as they were measured by the optical sizer and the SMPS. These results are graphically represented in Figure 7 indicating that the experiments performed at 20 kHz (Figure 7b) give slightly better results than those at 10 kHz (Figure 7a). The three bars refer to the application of four transducers with 400 W (4T 400 W), two transducers with 400 W (2T 400 W), and four transducers with 80 W (4T 80 W).

FIGURE 7. Reduction of the particle emissions expressed as mass concentration and number concentration (for both ranges, micron and submicron sized particles) under the conditions studied: (a) 10 kHz and (b) 20 kHz. The three bars refers to the application of four transducers with 400 W each (4T 400 W), two transducers with 200 W each (2T 200 W), and four transducers with 80 W each (4T 80 W). In Figure 8, the evolution of the particle size distribution at different frequencies and with varying acoustic power can be found for experiments conducted with four transducers. It can be seen that in all cases the micron sized particles have a higher number concentration than the submicron sized particles, an effect that is inherent to fluidized bed combustion processes. It is shown that the peak of the distribution decreases when the acoustic chamber is operational and that this decrease is larger if more power is applied at both frequencies. Agglomeration also causes a shift of the particle size distribution toward larger sizes. However, this larger particle mode cannot be observed in Figure 8 because of its effective removal by the ESP. Although it is evident that the agglomeration effect increases with the acoustic intensity, it is difficult to establish a clear dependence on the power applied to the transducers. In fact, acoustic agglomeration is a phenomenon that occurs only at SPL values high enough to produce nonlinear behavior in the acoustic field. Consequently, nonlinear effects such as acoustic saturation play a determining role which limits the applied acoustic power to a saturation value. At the best acoustic power/frequency combination tested (4T × 400 W at 20 kHz), the improvement of the emission reduction can reach values up to 37 and 42% of the mass and number concentration, respectively, compared to the emissions when the electrostatic filter is operating alone. This result is insofar significant as the collected particles are very small in size, the applied energy levels low, and the acoustic treatment times relatively short. Also, in the light of the narrow gain margin left by the VOL. 33, NO. 21, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3847

FIGURE 8. Evolution of the particles size distribution as a function of the acoustic frequency and power. The curves labeled with OFF refer to no acoustic excitation and those labeled with ON that the acoustic chamber was operational: (a) 10 kHz and four transducers at 80 W, (b) 10 kHz and four transducers at 400 W, (c) 20 kHz and four transducers at 80 W, and (d) 20 kHz and four transducers at 400 W. electrostatic filter the improvement of the retention efficiency can be considered substantial. The new system presents several advantages. It may be applied to any industrial process where control of fine particles is required. The preconditioning procedure is equally useful in combination with other filtering systems such as cyclone filters, bag houses, or ceramic filters. Another important benefit of this process is its applicability to highpressure and high-temperature environments (e.g. for coal gasification). As a consequence, the prospects for large scale 3848

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 21, 1999

acoustic agglomeration applications are very positive. Presently large transducers with power capacities of about 6 kW are being developed for scale-up, together with improved agglomeration chambers for treatment of flow rates as high as 10 000 m3/h.

Acknowledgments This work has been supported by the research project PIE131.095 sponsored by companies ENDESA/OCIDE and the project CICYT AMB96-1211-CO2-01.

Literature Cited (1) See the U.S. air quality standards approved by President Clinton on June 1997 (The New York Times, June 26, 1997). (2) Riehle, C. In Basic and Theoretical operation of ESPs; Parker, K. R., Ed.; Blackie Academic & Professional: London, 1997; Chapter 3. (3) Eliasson, B.; Egli, W.; Ferguson, J. R.; Jodeit, H. O. J. Aerosol Sci. 1987, 18(6), 869. (4) Gutsch, A.; Lo¨ffler, F. J. Aerosol Sci 1944, 25(S1), S307. (5) Kildesø, J.; Bhatia, V. K.; Lind, L.; Johnson, E.; Johansen, A. J. Aerosol Sci 1995, 23, 603. (6) Mednikov, E. P. Acoustic coagulation and precipitation of aerosols; Consultants Bureau: New York, 1965. (7) Song, L. Ph. D. Dissertation, The Pennsylvania State University, 1990. (8) Dianov, D. B.; Podol’skii, A. A.; Turubarov, V. I. Soviet Physics - Acoustics 1968, 13, 314. (9) Tiwary, R.; Reethof, G. J. Sound Vibration 1986, 108, 33. (10) Hoffmann, T. L. Ph. D. Dissertation, The Pennsylvania State University, 1993. (11) Hoffmann, T. L.; Koopmann, G. H. Rev. Sci. Ins. 1994, 65, 1527. (12) Scott, D. S. J. Sound Vibration 1975, 43, 607 (13) Gallego-Jua´rez, J. A.; Riera-Franco de Sarabia, E.; RodriguezCorral, G. Ultrasonics International 79 Conference Procedings; IPC Science and Technology Press: Guildford, England, 1979; pp 267-271.

(14) Riera-Franco de Sarabia, E.; Gallego-Jua´rez, J. A. J. Sound Vibration 1986, 110, 413. (15) Gallego-Jua´rez, J. A.; Riera-Franco de Sarabia, E.; RodriguezCorral, G. Multifrequency Acoustic Chamber for the Agglomeration and Separation of Suspended Particles in Gas Effluents. U.S. Patent 5,769,913, June 1998. (16) Gallego-Jua´rez, J. A.; Rodriguez-Corral, G.; San Emeterio-Prieto, J. L.; Montoya-Vitini, F. Electroacoustic Unit for Generating High Sonic and Ultrasonic Intensities in Gases and Interfaces. U.S. Patent 5.2999.175, 1994. (17) Rodrı´guez-Maroto, J. J.; Go´mez-Moreno, F. J.; Martı´n-Espigares, M.; Bahillo, A. J. Aerosol Sci. 1995, 26, S685. (18) Bernhard, R. Rapid Measurement of Particle Size Distributions by use of Light Scattering Methods; Paper presented at PARTEC Nuremberg May 6-9 1981 (POLYTEC). (19) Knudson, E. O.; Whitby, K. T. J. Aerosol Sci. 1975, 6, 443. (20) Wang, S. C.; Flagan, R. C. Aerosol Sci. Technol. 1990, 13, 230. (21) Agarwal, J. K.; Sem, G. J. J. Aerosol Sci. 1980, 11, 343. (22) Cascade Impactors, Sampling & data Analysis; Lodge, J. P., Chan, T. L., Eds.; American Industrial Hygiene Association: Akron, OH, 1986.

Received for review January 4, 1999. Revised manuscript received June 24, 1999. Accepted July 22, 1999. ES990002N

VOL. 33, NO. 21, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3849