Aerosol Counterflow Two-Jets Unit for Continuous Measurement of the

Jul 28, 2005 - A new type of aerosol collector employing a liquid at laboratory temperature for continuous sampling of atmospheric particles is descri...
1 downloads 0 Views 168KB Size
Download PDF
Anal. Chem. 2005, 77, 5534-5541

Aerosol Counterflow Two-Jets Unit for Continuous Measurement of the Soluble Fraction of Atmospheric Aerosols Pavel Mikusˇka* and Zbyneˇk Vecˇerˇa

Institute of Analytical Chemistry, Academy of Sciences of the Czech Republic, Veverˇ´ı 97, CZ-61142 Brno, Czech Republic

A new type of aerosol collector employing a liquid at laboratory temperature for continuous sampling of atmospheric particles is described. The collector operates on the principle of a Venturi scrubber. Sampled air flows at high linear velocity through two Venturi nozzles “atomizing” the liquid to form two jets of a polydisperse aerosol of fine droplets situated against each other. Counterflow jets of droplets collide, and within this process, the aerosol particles are captured into dispersed liquid. Under optimum conditions (air flow rate of 5 L/min and water flow rate of 2 mL/min), aerosol particles down to 0.3 µm in diameter are quantitatively collected in the collector into deionized water while the collection efficiency of smaller particles decreases. There is very little loss of fine aerosol within the aerosol counterflow two-jets unit (ACTJU). Coupling of the aerosol collector with an annular diffusion denuder located upstream of the collector ensures an artifact-free sampling of atmospheric aerosols. Operation of the ACTJU in combination with on-line detection devices allows in situ automated analysis of water-soluble aerosol species (e.g., NO2-, NO3-) with high time resolution (as high as 1 s). Under the optimum conditions, the limit of detection for particulate nitrite and nitrate is 28 and 77 ng/m3, respectively. The instrument is sufficiently rugged for its application at routine monitoring of aerosol composition in the real time. Atmospheric aerosols play an important role in various environmental issues.1,2 The aerosol particles affect the Earth’s climate by changing the radiation budget of the atmosphere. By scattering UV radiation back to space, they reduce solar irradiance (so-called direct aerosol effect). In clouds, the aerosols may serve as cloud condensation nuclei affecting thus the formation, albedo, and occurrence of clouds (indirect aerosol effect). The aerosols also contribute to acidification and eutrophication of land and water resources via wet and dry deposition, take part in smog production and visibility reduction over large portions of the globe, or serve as the sites on which heterogeneous reactions of gaseous trace constituents occur. In addition, epidemiological studies report a * Corresponding author. E-mail: [email protected]. Phone: +420 532290167. Fax: +420 541212113. (1) Seinfeld, J. H.; Pandis, S. N. Atmospheric Chemistry and Physics. From Air Pollution to Climate Change; Wiley & Sons: New York, 1998. (2) Finlayson-Pitts B. J., Pitts J. N., Jr. Chemistry of the Upper and Lower Atmosphere. Theory, Experiments, and Applications; Academic Press: San Diego, CA, 2000.

5534 Analytical Chemistry, Vol. 77, No. 17, September 1, 2005

correlation between increased adverse health effects and mortality, respectively, and high concentration of ambient particulate matter.2-4 The exact role that atmospheric aerosols play depends primarily on aerosol number concentration, on size distribution, and especially on chemical composition of aerosols. The measurement of aerosol composition is very difficult because of high complexity of atmospheric aerosol components and their considerable variations with time and place. Although various constituents of atmospheric aerosols are frequently measured within specialized studies, in a long-term period, the chemical composition of atmospheric aerosols is not routinely measured because, at present, no reliable and low-cost automated instrumentation with a reasonable time resolution is commercially available. Measurement of the chemical composition of aerosols is usually carried out by collection of aerosols at filters or impactors with subsequent off-line analysis of collected particles providing time-integrated results. However, these procedures are known to be subject to significant sampling artifacts. Particles collected on filters can be lost by volatilization during and after collection (negative artifacts) while positive artifacts may arise from gas absorption on filter or collected particles.5 Another approach, an electrostatic collection of aerosols,6,7 allows high efficiency of particle collection but the necessity of regular washing of collected particles makes the measurement discontinuous. Various liquidbased collectors working on different principles such as absorption,8 impaction,9,10 or collection on frit11 or filter12 have been described for aerosol sampling, but efficiency of absorption decreases when particle diameter goes below 1 µm. Application of filters suffers from problems with disintegration of filter and gradual accumulation of water-insoluble substances on the filter (3) Dockery, D. W.; Pope, C. A., III; Xu, X.; Spengler, J. D.; Ware, J. H.; Fay, M. E.; Ferris, Jr., B. G.; Speizer, F. E. N. Engl. J. Med. 1993, 329, 17531759. (4) Wilson, W. E.; Suh, H. H. J. Air Waste Manag. Assoc. 1997, 47, 12381249. (5) Chow, J. C. J. Air Waste Manag. Assoc. 1995, 45, 320-382. (6) Liu, S.; Dasgupta, P. K. Anal. Chem. 1996, 68, 3638-3644. (7) Romay, F. J.; Pui, D. Y. H.; Smith, T. J.; Ngo, N. D.; Vincent, J. H. Atmos. Environ. 1996, 30, 2607-2613. (8) Spanne, M.; Grzybowski, P.; Bohgard, M. Am. Ind. Hyg. Assoc. J. 1999, 60, 540-544. (9) Karlsson, A.; Irgum, K.; Hansson, H. J. Aerosol Sci. 1997, 28, 1539-1551. (10) Stolzenburg, M. R.; Hering, S. V. Environ. Sci. Technol. 2000, 34, 907914. (11) Buhr, S.; Buhr, M.; Fehsenfeld, F.; Holloway, J.; Karst, U.; Norton, R.; Parrish, D.; Sievers, R. Atmos. Environ. 1995, 29, 2609-2624. (12) Boring, C. B.; Al-Horr, R.; Genfa, Z.; Dasgupta, P. K. Anal. Chem. 2002, 74, 1256-1268. 10.1021/ac050343l CCC: $30.25

© 2005 American Chemical Society Published on Web 07/28/2005

during long-term use. There is also a danger of leaching of fibers from filters with subsequent blocking of downstream devices. To avoid all these difficulties, there is a need for automated continuous devices that can analyze the aerosol composition online without manual treatment. On-line size-resolved analysis of single particles in real time can be accomplished by mass spectrometry,13,14 but the quantitative determination of particle components is still problematic. In addition, the instrumentation is costly and very complex requiring sophisticated operation, which limits its widespread use. Presently, aerosol collectors based on a condensation principle are most commonly used for continuous sampling of atmospheric aerosols.15-22 Sampled air stream is mixed with a water steam15-21 or saturated over a pool with warm water,22 cooling of the mixture results in high supersaturation, and subsequent condensation of water vapor causes aerosol particles to grow into large droplets. Although the steam collectors seem to be very advantageous tools for aerosol sampling, use of a hot steam does not allow reliable sampling of volatile aerosol species such as semivolatile organic compounds or ammonium nitrate. In addition, reaction of steam with atmospheric NO2 can result in significant artifacts during particulate nitrite and nitrate analysis, for example. Various kind of wet scrubbers23 are widely used for the control of air pollution mainly because of their low cost and their ability to remove particulate and gaseous contaminants. Recently, an aerosol collector operating on the principle of Venturi scrubber has been described,24 where aerosol particles are collected mainly in a water film formed on a hydrophobic membrane filter by impaction of a fine mist that is created by aerosolization of water by high-velocity air. Although optimization of Venturi scrubber efficiency has been frequently studied,25-36 a major drawback of (13) Gard, E.; Mayer, J. E.; Morrical, B. D.; Dienes, T.; Fergenson, D. P.; Prather, K. A. Anal. Chem. 1997, 69, 4083-4091. (14) Noble, C. A.; Prather, K. A. Mass Spectrom. Rev. 2000, 19, 248-274. (15) Simon, P. K.; Dasgupta, P. K. Anal. Chem. 1995, 67, 71-78. (16) Khlystov, A.; Wyers, G. P.; Slanina, J. Atmos. Environ. 1995, 29, 22292234. (17) Ito, K.; Chasteen, C. C.; Chung, H.; Prouthoor, S. K.; Genfa, Z.; Dasgupta, P. K. Anal. Chem. 1998, 70, 2839-2847. (18) Weber, R. J.; Orsini, D.; Daun, Y.; Lee, Y. N.; Klotz, P. J.; Brechtel, F. Aerosol Sci. Technol. 2001, 35, 718-727. (19) Lo ¨flund, M.; Kasper-Giebl, A.; Tscherwenka, W.; Schmid, M.; Giebl, H.; Hitzenberger, R.; Reischl, G.; Puxbaum, H. Atmos. Environ. 2001, 35, 28612869. (20) Sierau, B.; Stratmann, F.; Pelzing, M.; Neusu ¨ ss, C.; Hofmann, D.; Wilck, M. Aerosol Sci. 2003, 34, 225-242. (21) Orsini, D. A.; Ma, Y.; Sullivan, A.; Sierau, B.; Baumann, K.; Weber, R. J. Atmos. Environ. 2003, 37, 1243-1259. (22) Sioutas, C.; Koutrakis, P. Aerosol Sci. Technol. 1996, 25, 424-436. (23) Calvert, S. In Handbook of Air Pollution Technology; Calvert, S., Englund, H. M., Eds.; Wiley & Sons: New York, 1984; pp 215-248. (24) Al-Horr, R.; Samanta, G.; Dasgupta, P. K. Environ. Sci. Technol. 2003, 37, 5711-5720. (25) Calvert, S. AIChE J. 1970, 16, 392-396. (26) Boll, R. H. Ind. Eng. Chem. Fundam. 1973, 12, 40-50. (27) Behie, S. W.; Beeckmans, J. M. Can. J. Chem. Eng. 1973, 51, 430-433. (28) Goel, K. C.; Hollands, K. G. T. Atmos. Environ. 1977, 11, 837-845. (29) Leith, D.; Cooper, D. W. Atmos. Environ. 1980, 14, 657-664. (30) Placek, T. D.; Peters, L. K. AIChE J. 1981, 27, 984-993. (31) Crowder, J. W.; Noll, K. E.; Davis, W. T. Atmos. Environ. 1982, 16, 20092013. (32) Tigges, K. D.; Mayinger, F. Chem. Eng. Process. 1984, 18, 171-179. (33) Miller, R. L.; Jain, D. M.; Sharma, M. P. Chem. Eng. Commun. 1990, 89, 101-112. (34) Ananthanarayanan, N. V.; Viswanathan, S. AIChE J. 1998, 44, 2549-2560. (35) Lehner, M. Aerosol Sci. Technol. 1998, 28, 389-402. (36) Shyan, L. D.; Viswanathan, S. Environ. Sci. Technol. 2000, 34, 5007-5016.

present day commercial scrubbers is decreased effectiveness of collection mechanisms for submicrometer aerosol particles. In this paper, the new type of aerosol collector, an aerosol counterflow two-jets unit (ACTJU), employing a liquid at room temperature for continuous sampling of atmospheric aerosol particles is described. The ACTJU is destined for automated measurement (i.e., sampling and subsequent on-line analyses) of soluble aerosol constituents of both inorganic and organic nature. EXPERIMENTAL SECTION Apparatus. The employed apparatus consists of an impactor (2.5-µm aerodynamic cutoff diameter), an annular diffusion denuder, a particle collector (i.e., ACTJU), and various detection systems. The analyzed air aspirated into the collector by membrane pump first passes through the stainless steel single-stage impactor where particles with aerodynamic diameters larger than 2.5 µm are removed and then goes through a “dry” annular diffusion denuder to remove interfering gases. Particle Collector. The ACTJU, schematically shown in Figure 1, consists of two Venturi nozzles, a collision chamber, and a cyclone. The collision chamber has a cylindrical design with an inner diameter of 4 mm and a length of 10 mm. The analyzed air is aspirated into the collision chamber through two identical Venturi nozzles screwed against each other in opposite sides of the collision chamber along a longitudinal axis at a distance of 0.75 mm between nozzle throats. The inner diameter of the Venturi nozzle throat is 1 mm and the throat length is 2 mm; the distance between the throat of the Venturi nozzle and the throat of the capillary for liquid feed is 2.5 mm. A taper angle at the converging zone of the Venturi nozzle is 40°. Liquid (deionized water) is delivered into the nozzles through stainless steel capillaries (i.d. 0.6 mm, o.d. 0.9 mm) by means of a peristaltic pump (Ismatec, type ISM 852) or, alternatively, is aspirated due to an ejection effect of streaming air. High-velocity air stream sprays introduced liquid into fine droplets. Formed counterflow jets of fine droplets collide with each other inside the collision chamber, and within this process, together with other mechanisms, the aerosol particles are captured into dispersed liquid. The inhomogeneous mixture in water with collected particles is aspirated out from the collision chamber through a hole (1.5-mm diameter) drilled in the middle of the chamber perpendicularly versus the longitudinal axis and enters tangentially into the cyclone (i.d. 7 mm, 15 mm long) where air and liquid are separated. Collector effluent, i.e., aqueous concentrate of compounds of interest together with insoluble aerosol spices, is continuously aspirated from the bottom of the cyclone by a peristaltic pump for subsequent analysis, while air is aspirated out by the membrane pump through a stainless steel capillary (i.d. 1.0 mm, o.d. 1.5 mm). The body of the aerosol counterflow two-jets unit consisting of the collision chamber and the cyclone is made from Ketron PEEK-1000; Venturi nozzles are made of stainless steel. The dimensions of the ACTJU body are 40 × 40 × 50 mm. Analytical System. Nitrite and nitrate as representatives of water-soluble constituents of collected aerosols are on-line analyzed in the ACTJU effluent using a continuous detection system. The collector effluent is continuously pumped out from bottom part of the cyclone into a glass debubbler. After that, the debubbled solution is aspirated by means of the peristaltic pump Analytical Chemistry, Vol. 77, No. 17, September 1, 2005

5535

Figure 1. Aerosol counterflow two-jets unit. (A) front view; (B) side view. All dimensions are given in millimeters.

Figure 2. Scheme of nitrite and nitrate continuous detection system. ACTJU, particle collector; E, effluent; DB, debubbler; CEC, cation exchange column; NC, photolytic converter; D, detector; PP, peristaltic pump; PC, computer; CL, chemiluminescent solution; RE, reagent solution.

through the cation exchange column to remove interfering cations, and then it splits into two streams for nitrate and nitrite analysis. Both nitrite and nitrate are detected at the same time via a continuous method (Figure 2) with 1-s time resolution based on FIA methods reported previously.37-39 Content of particulate nitrite is detected directly in the first ACTJU stream that is continuously merged with a reagent solution consisting of 4 mM H2O2 and 3 mM EDTA in 0.3 M H2SO4 . During the flow through a reaction coil (PTFE, 0.5 mm i.d. × 30 cm length), H2O2 in acid medium oxidizes nitrous acid to form peroxynitrous acid that directly inside a chemiluminescent (CL) detector is merged with the CL solution containing 2 mM luminol and 3 mM EDTA in 0.6 M KOH. Luminol is oxidized by peroxynitrite and emitted CL light is detected by a photomultiplier tube (model 65 PK 518, TeslaVacuum Technique, Prague, Czech Republic).37 (37) Mikusˇka, P.; Zdra´hal, Z.; Vecˇerˇa, Z. Anal. Chim. Acta 1995, 316, 261-268. (38) Mikusˇka, P.; Vecˇerˇa, Z. Anal. Chim. Acta 2002, 474, 99-105. (39) Mikusˇka, P.; Vecˇerˇa, Z. Anal. Chim. Acta 2003, 495, 225-232.

5536

Analytical Chemistry, Vol. 77, No. 17, September 1, 2005

Particulate nitrate in the second ACTJU stream is first photolytically on-line reduced to nitrite by the absorption of UV light,38 and the sum of the original and reduced nitrite is determined in the parallel detection system via the way as described above for nitrite alone. Nitrate content is calculated from the difference of signals for nitrite and the sum of nitrite and nitrate.39 Peristaltic pumps (Ismatec, type ISM 852 and ISM 597A) are used for transportation of the sample and the reagent solutions. All liquids flow through Teflon tubings with i.d. 0.5 mm (o.d. 1/16 in.) or i.d. 0.25 mm (o.d. 1/16 in.) to minimize transport delay between the cyclone and the debubbler. Standard Aerosol. Performance of the ACTJU was tested under laboratory conditions with both polydisperse and monodisperse aerosols. A fine polydisperse sodium nitrate aerosol was generated by pneumatic atomization of a NaNO3 solution by a high-velocity air stream in a concentric nebulizer.40 The aerosol spray leaving the atomizer passed through a water trap (TSI), where large droplets were collected. The spray was then mixed in a glass tube with dry filtered air to evaporate water from the droplets to form dry aerosol particles. By changing the concentration of sodium nitrate in the sprayed solution, the NaNO3 particle in different size ranges can be produced. For the NaNO3 concentration in the sprayed solution, 20 g/L, NaNO3 particles in the size range of 10-750 nm are obtained. The geometric mean diameter (GMD) of produced polydisperse aerosol is 56.9 nm, and a geometric standard deviation is 2.30. For the NaNO3 concentration in the 0.2 g/L sprayed solution, the aerosol particles in the size range of 10-133 nm are produced, geometric mean diameter is 34.2 nm, and a geometric standard deviation of produced polydisperse aerosol is 1.61. Monodisperse aerosol was produced at a condensation generator (MAG-2010, Palas) operating on the principle of heterogeneous condensation of vaporized particle material (i.e., bis(2-ethylhexyl) sebacate, DEHS) at the nuclei produced by nebulization of NaCl (0.02 g/L) and uranine (0.3 g/L) solution by a nitrogen stream. The temperature of the saturator and nitrogen flow rate through (40) Mikusˇka, P. Collect. Czech. Chem. Commun. 2004, 69, 1453-1463.

the saturator were changed in order to obtain the monodisperse aerosol with particles of GMD values of 0.12, 0.19, 0.28, 0.43, 0.56, 0.74, 1.03, 1.34, and 2.26 µm, respectively. The parameters of standard aerosols were measured with an electrical aerosol size analyzer (EASA; model 3030, TSI) and with an optical aerosol spectrometer Welas 2000 (Palas). The Welas operated on the principle that the single particle’s scattering of white light measures number concentration of particles in the size range 0.3-17 µm. The particles are classified into 60 size channels on the basis of polystyrene latex equivalent diameters. Maximum concentration is 104 (sensor type 2200) or 105 particles/cm3 (sensor type 2100), respectively. Collection Efficiency and Loss of Particles. The collection efficiency (CE) of aerosol in the ACTJU as a function of particle size was determined using the standard monodisperse aerosol. The DEHS aerosol doped with uranine was sampled on a backup filter placed downstream of the ACTJU and simultaneously on a reference filter placed in parallel with the collector. After sampling, the filters (quartz microfiber filter, Whatman QM-A, 4.7 cm) were extracted in the ultrasonic bath with 8 mL of deionized water for 10 min at ambient temperature. The uranine concentration in the filter extract was analyzed using a fluorometer Fluoromat (model FS 950, Kratos Analytical Instruments). The CE was calculated as a ratio of uranine mass collected on both filters. The CE as a function of water flow rate into the ACTJU was determined by sampling the standard NaNO3 aerosol on a backup filter as a ratio of mass of nitrate collected in the collector to the nitrate mass collected both in the collector and on the backup filter. After sampling, the filter (quartz microfiber filter, Whatman QM-A, 4.7 cm) was extracted in the ultrasonic bath with 8 mL of deionized water for 15 min at ambient temperature. The nitrate concentration in the filter extract was determined by means of the FIA technique with chemiluminescent detection.38 The efficiency of aerosol collection in the ACTJU as a function of air flow rate was measured on-line using the spectrometer Welas. The CE of particles on a number basis is expressed as the fraction of entering particles that is retained in the ACTJU, CE ) (Nin - Nout)/Nin, where Nin is particle number concentration (#/cm3) at air entering into the ACTJU and Nout is particle number concentration (#/cm3) at the output from the ACTJU. The loss of aerosols in the ACTJU was determined by washing out the deposited monodisperse DEHS aerosol (doped with uranine) from the ACTJU with subsequent fluorescent analysis of uranine at extract. Annular Diffusion Denuder. The annular diffusion denuder41 consists of a stainless steel tube (470 mm length × 60 mm i.d. × 63 mm o.d.) in which two tubes (with diameters of 46 and 51 mm, respectively, and length of 450 mm) from copper wire net are coaxially placed to form an annulus width of 2.5 mm. The space within of the inner cuprous net and the space between the outer cuprous net and the stainless steel tube are filled with activated charcoal. Particle losses during passing of air through the “dry” annular denuder were measured with the spectrometer Welas and the electrical aerosol size analyzer. (41) Mikusˇka, P.; Vecˇerˇa, Z.; Brosˇkovicˇova´, A.; Sˇ teˇpa´n, M.; Chi, X.; Maenhaut, W. J. Aerosol Sci. 2003, S761-S762.

Reagents. All solutions are prepared with a distilled deionized water. NaNO3 (Aldrich, Milwaukee, WI), uranine (disodium salt of fluorescein; Aldrich), and other chemicals (Lachema, Brno, Czech Republic) are of analytical grade; luminol (p.a., SigmaAldrich, Praha, CR) is used without further purification. RESULTS AND DISCUSSION Optimization and Efficiency of the Collector. The collector operates on the principle of a Venturi scrubber when analyzed air flowing at high linear velocity through the Venturi nozzle atomizes the liquid to form a polydisperse aerosol of small droplets. Aerosol particles are collected by interactions with droplets. Droplets with collected aerosol particles are separated on-line from air in the small cyclone. A few collectors with different geometry have been developed and tested during the work. At the beginning, we employed configurations with a single Venturi nozzle in combination with subsequent multistage inputs of deionized water. The final collector version applied two Venturi nozzles located at opposite positions. All versions with a single Venturi nozzle were much larger in size, and in addition, they did not reach as high an efficiency of aerosol collection as did the two-nozzle version. For separation of droplets from air, the cyclone was used in all tested versions. At the two-nozzle version of the collector, so-called aerosol counterflow two-jets unit, formed counterflow jets of fine droplets colliding inside the collision chamber with each other and within this process, together with other mechanisms, the aerosol particles are captured into dispersed liquid. Overall efficiency of aerosol collection in the ACTJU depends on the collector parameters (diameter and position of capillary at Venturi nozzle, throat diameter and length, axial distance of nozzles, etc.) and primary operating variables (gas flow rate, liquid flow rate, throat gas velocity, liquid-to-gas ratio). Selection of the liquid applied to the collection of aerosols plays an important role. We use deionized water as the liquid for the collection of aerosol particles from air in the ACTJU. The longitudinal distance between the throats of the Venturi nozzles was optimized by sampling of NaNO3 aerosol in the size range of 0.3-0.7 µm. The optimum distance between the Venturi nozzles is 0.5-1.0 mm. At a larger distance, the efficiency of aerosol collection decreases while at a smaller distance the pressure drop acros the ACTJU increases. The collection efficiency of aerosol as a function of the air flow rate in the range 1-5 L/min was studied by sampling of NaNO3 aerosol (size range, 0.3-0.7 µm) followed by measurement of particle concentration upstream and downstream of the collector by means of the Welas spectrometer (Figure 3, water flow rate of 0.5 mL/min). It is evident that nearly quantitative aerosol capture (CE > 99.7%) in the ACTJU is accomplished for air flow rates of 4 (CE ) 99.7%) and 5 (CE ) 99.9%) L/min for the aerosol in the size range of 0.3-0.7 µm (the number concentration of NaNO3 particles in this size range is 3.2 × 105 #/cm3), and even at the air flow rate of 3 L/min, the collection efficiency of nitrate particles is as high as 98.3%. Similar results were obtained also for the DEHS aerosol in the size range of 0.3-3.4 µm although the DEHS is not miscible with water, hence producing very small oil drops inside the water stream. The effect of liquid flow rate was examined in the range 0.14.0 mL/min for equal liquid inputs into both nozzles and for Analytical Chemistry, Vol. 77, No. 17, September 1, 2005

5537

Figure 3. Dependence of aerosol collection efficiency on air flow rate. NaNO3 aerosol (the size range of 300-750 nm); water flow rate 0.5 mL/min.

Figure 4. Dependence of aerosol collection efficiency on water flow rate. NaNO3 aerosol (the size range of 10-133 nm); air flow rate 5 L/min.

various ratios of inputs into nozzles from ratio 1:1 up to 40:1. For a sampling of aerosols with diameters larger than 0.3 µm, water flow rate in the stated range has no effect on the efficiency of aerosol collection. The only problem arises at a water flow rate below 0.2 mL/min when an evaporation loss of water inside the ACTJU at low humidity of analyzed air results in too low output of the effluent from the collector, which may subsequently cause problems within the analysis of the effluent components. Quite a different situation exists at the collection of aerosols with particle diameters smaller than 0.3 µm when water flow rate plays a crucial role. Figure 4 shows the dependence of CE of the NaNO3 aerosol (in the size range of 10-133 nm) on the water flow rate (at air flow rate of 5 L/min). The CE increases from 26% at 0.5 mL/min to a maximum of 77% at 2 mL/min; at higher water flow rates, the CE decreases, probably due to reduced efficiency in the production of water droplets at the Venturi nozzles. The effluent flow rate (i.e., flow rate of water with collected aerosols out of the ACTJU; FLeff) is smaller than the water flow rate into the collector because of evaporation of water inside the collector; nevertheless, the dependence is linear. For the optimum water flow rate into the ACTJU (2 mL/min), the effect of relative humidity (RH) of sampled air was investigated but the changes in the effluent flow rate as a function of relative humidity of sampled air are negligible when FLeff increases from 1.75 mL/ min at RH ) 13% to 1.88 mL/min at RH ) 86%. In further work, 5538 Analytical Chemistry, Vol. 77, No. 17, September 1, 2005

Figure 5. Size-dependent collection efficiency of DEHS aerosol. Air flow rate 5 L/min; water flow rate 2 mL/min.

deionized water at normal temperature was used as the liquid at the flow rate of 2.0 mL/min while sampled air passed through the ACTJU at the flow rate of 5 L/min. During study of aerosol CE, we first sampled the polydisperse NaNO3 aerosol in the ACTJU (at the air flow rate of 5 L/min and the water flow rate of 2 mL/min). The penetration of aerosols through the ACTJU was on-line checked by means of the Welas spectrometer placed after the ACTJU. We found that no particles larger than 0.3 µm passed through the ACTJU. However, we do not have any information about the CE for smaller particles. Subsequent measurement was performed with the monodisperse uranine-doped DEHS aerosol. The CE was evaluated using aerosol sampling simultaneously on a filter after the collector and on a parallel reference filter. The dependence of aerosol collection efficiency on the size of DEHS particles is shown in Figure 5. The obtained curve confirmed the quantitative collection of particles with a diameter larger than 0.3 µm and determined the CE of 67, 87, and 98% for particles of GMD values 0.12, 0.19, and 0.28 µm, respectively. These results thus prove the quantitative collection of particles larger than 0.3 µm in the ACTJU while below this size limit the efficiency of particle collection decreases. Loss of aerosol within the ACTJU (at the air flow rate of 5 L/min and the water flow rate of 2 mL/min) measured with the uranine-doped DEHS aerosol was found to be 3.07, 2.75, 2.65, 2.59, 2.44, 2.35, 2.22, 2.40, and 3.15% for particles of GMD values 0.12, 0.19, 0.28, 0.43, 0.56, 0.74, 1.03, 1.34, and 2.26 µm, respectively. From the obtained results, it is evident that there is very little loss of fine particles inside the ACTJU. Particle Collection Mechanism. The mechanism of particle collection in the Venturi scrubbers has been subject of many papers.23,25-36 These studies indicate that the inertial impaction is the primary collection mechanism responsible for particle collection from an air stream by water droplets due to the relative velocity difference between the particle-laden gas stream and the liquid droplets. In the ACTJU, the air stream with particles flowing through the Venturi nozzle is accelerated in the converging entrance section. Water introduced into coaxially flowing air is atomized at the throat inlet by the high-velocity air stream to produce a spectrum of fine droplets. We propose this mechanism for the particle collection in the ACTJU: the faster moving particles are captured by impaction and interception with the slower moving

Figure 6. Record of on-line particulate nitrate concentration (1-s time resolution), Brno, 15 June 2004. Air flow rate 5 L/min; water flow rate 2 mL/min.

water droplets included at two counterflow streams of liquid droplets that are produced at two Venturi nozzles located at opposite sites of collision chamber. Moreover, the collision of both jets in the middle of the chamber significantly increases the number of impacts between particles and droplets with consequent considerable improvement in overall collection efficiency of particulate matter in the ACTJU. The aerosol collection efficiency in the ACTJU is further enhanced by adiabatic expansion of both air streams at the exit of the Venturi throats resulting in cooling, which causes a highly humid environment inside the collector. Under these wet conditions, growth of particles in size may occur by water uptake due to the deliquescence effect of aerosols1,42-45. Relative humidity corresponding to the deliquescence point of individual aerosol components seems to be high sometimes, but the deliquescence relative humidity of particles in a multicomponent aerosol mixture is lower than that of its individual components.1,44 Moreover, recent studies46,47 indicate that particles may start to grow in size even below the reported deliquescence relative humidity without any distinct step-size changes. Grown particles are then easily collected by impaction with water droplets. Analytical Performance. The collector effluent (i.e., the liquid with dissolved aerosol species as well as the nondissolved parts) is permanently aspirated out from the ACTJU for subsequent on/ off-line analysis of particulate components. Nitrite and nitrate as representatives of water-soluble inorganic species are analyzed on-line by means of a continuous method37,38 that allows fast and sensitive detection of both ions directly in the collector effluent without need of sample preconcentration. The detection limits (S/N > 3) of aqueous nitrite and nitrate are 1.5 and 3.1 nM, (42) Tang, I. N. In Generation of Aerosols and Facilities for Exposure Experiments; Willeke, K., Ed.; Ann Arbor Science: Ann Arbor, MI, 1980; pp 153-167. (43) Pilinis, C.; Seinfeld, J. H.; Grosjean, D. Atmos. Environ. 1989, 23, 16011606. (44) Wexler, A. S.; Seinfeld, J. H. Atmos. Environ. 1991, 25A, 2731-2748. (45) Tang, I. N.; Munkelwitz, H. R. Atmos. Environ. 1993, 27A, 467-473. (46) Ha¨meri, K.; Va¨keva¨, M.; Hansson, H. C.; Laaksonen, A. J. Geophys. Res. 2000, 105, 22231-22242. (47) Hoffman, R. C.; Laskin, A.; Finlayson-Pitts, B. J. J. Aerosol Sci. 2004, 35, 869-887.

respectively, and the corresponding detection limits of particulate nitrite and nitrate are 28 and 77 ng/m3, respectively (calculated for the water flow rate of 2 mL/min and the air flow rate of 5 L/min). Negligible inner volume of the ACTJU (0.8 cm3) enables a short residence time (12 ms), and the response time (95% of steady-state detector signal) of the whole system is then limited by analysis time of detection method used, which depends on the sample (i.e., the ACTJU effluent stream) flow rate (see Figure 2). When the NaNO3 aerosol concentration in air entering the ACTJU was steeply switched between 0 and 45 µg/m3, the response time of the system was 145, 100, and 82 s for the sample flow rates of 100, 200, and 300 µL/min, respectively. The corresponding limits of detection of nitrate in solution at the continuous detection system (CDS) are 3.1, 4.4, and 7.5 nM and detection limits of particulate nitrate are 77, 109, and 186 ng/m3. The times obtained when aerosol concentration was changed to zero are about 5-10 s longer than corresponding rise times, which still provides rapid response time. It is evident that with increasing sample flow rate the response time of the system decreases; however, simultaneously sensitivity of nitrite and nitrate determination goes down too. In further experiments, the sample flow rate of 100 µL/min at the CDS was mostly employed because the response time in the order of 2 min is still quite satisfactory for most field applications where high sensitivity rather than high response time is preferred. These results offer application of the collector for the measurement of aerosol chemical composition in the real time. Time resolution for on-line analysis of nitrite and nitrate is 1 s. It is obvious that this very high resolution has no practical use for most field applications, but there are a few special utilizations such as specialized kinetic studies within atmospheric chemistry, for example, where such time resolution may be useful. The example of particulate nitrate measurement (with time resolution of 1 s) is presented in Figure 6, where the ability of the ACTJU and connected analytical system to detect quick changes in particulate nitrate concentration is clearly demonstrated. In the course of 15 min, we observed frequent variations in nitrate Analytical Chemistry, Vol. 77, No. 17, September 1, 2005

5539

Figure 7. Parallel measurement of particulate nitrate and particle number concentration. Brno, 14-18 June 2004. Air flow rate 5 L/min; water flow rate 2 mL/min. (A) On-line measurement of particulate nitrate concentration, 3-min averages, PM 2.5. (B) Total particle number concentration, 5-min measuring interval, size range 0.3-2.5 mm.

concentrations with a few peaks of particulate nitrate. The other sampling systems15,16,18,19,21 combined with ion chromatography during the same time interval could give only a single integrated average of nitrate concentration over this time interval. The particle impaction system designed specifically for nitrate analysis10,48 provides a time resolution of 10 min, which is also much slower than that of the presented system. Interferences. Atmospheric water-soluble gases (e.g., HNO3, HONO, SO2, HCl, and NH3) coexisting with aerosols in analyzed air have to be removed from the sampled air prior its entrance into the ACTJU otherwise they are efficiently collected in the presented collector into deionized water. To eliminate these positive artifacts, the aerosol collector during its operation is coupled with the annular diffusion denuder41 placed between the impactor and the ACTJU. During passing of air through the denuder, the gaseous interferences are removed by their collection onto a solid absorbing layer (i.e., activated charcoal) while aerosols (48) Stolzenburg, M. R.; Dutcher, D. D.; Kirby, B. W.; Hering, S. V. Aerosol Sci. Technol. 2003, 37, 537-546.

5540

Analytical Chemistry, Vol. 77, No. 17, September 1, 2005

pass through the denuder without any change. Outgoing air containing aerosol particles then enters the aerosol collector where particles are collected. The combination of the annular denuder and the ACTJU thus provides artifact-free sampling of atmospheric aerosols. The used “dry” annular diffusion denuder41 also removes NO2 from passed air very efficiently, which is the source of positive artifacts at steam-based collectors15-17,19 even if wet denuders (although parallel plate15,17,19 or annular16 design are employed) are used for the sampling of gaseous interferences. Moreover, this dry denuder version, contrary to wet denuders,19 does not cause hygroscopic growth of passed particles. Losses of particles due to gravitational sedimentation are minimized by vertical orientation of the denuder during the measurement. Diffusion losses of particles in the size range 0.1-2.5 µm were not observed; losses of particles with diameters smaller than 0.1 µm are 2-3%. Ambient Air Measurement. The performance of the ACTJU was verified during measurement of atmospheric aerosols in air in the city of Brno. The sampling place was the roof of the Institute

of Analytical Chemistry at the fifth floor. A street with relatively heavy traffic (cars and trams) is ∼30 m away from the building. Ambient air with aerosol particles (PM 2.5) was continuously sampled in the ACTJU (at the air flow rate of 5 L/min and the water flow rate of 2 mL/min), and particulate nitrite and nitrate at the collector effluent were analyzed on-line. The ambient particulate nitrate measurement was performed with a 4-min time delay between entrance of aerosol into the ACTJU and obtaining of corresponding detector signal. Concentrations of particulate nitrite during the whole measurement period were very low, mostly 1-2 orders below concentration of nitrates without any episodes or correlation with corresponding nitrates or day time; therefore, nitrites were not included in the next discussion. The concentration of particulate nitrate during 14-18 June 2004 is shown in Figure 7A as 3-min averages. During the measurements, we observed a few episodes when the concentration of particulate nitrates reached almost up to levels of 1.5 µg/m3. In parallel with aerosol sampling in the ACTJU and subsequent particulate nitrite and nitrate analysis, the particle number concentration was measured with the Welas spectrometer (Figure 7B). Results depicted in Figure 7A and B demonstrate similar trends in the course of concentration of particulate nitrates (3-min averages, PM 2.5) and the total particle number concentration (5-min averages) in the size range 0.3-2.5 µm. CONCLUSIONS A new type of wet collector for the continuous automated sampling of atmospheric aerosols has been developed. Coupling of the collector with an on-line detection device was verified at

the analysis of particulate inorganic ions. Consequently, we suppose that the ACTJU can also be applied to on-line analysis of aerosol water-soluble components of organic origin. No steam is required for aerosol sampling, which offers a lot of advantages in comparison with widely used steam-based aerosol collectors. Particularly, application of water at laboratory temperature is assumed to eliminate sampling artifacts due to loss of particulate semivolatile compounds of organic as well as inorganic nature that are observed if steam collectors or classical filter techniques are used. High collection efficiency, simple design, small dead volume, and easy maintenance of ACTJU offer an excellent opportunity to employ this kind of collector for the long-term unattended monitoring of atmospheric aerosols in the frame of pollution control of ambient air and in the study of long-term impact of aerosols on environment. ACKNOWLEDGMENT This work was supported by the Grant Agency of the Academy of Sciences of the Czech Republic under Grant IAA4031105 and by the Grant Agency of the Czech Republic under Grants GA CR 526/03/1182 and 525/04/0011. We thank Dr. Smolı´k, Institute of Chemical Process Fundamentals, Academy of Sciences of the Czech Republic, Prague for lending EASA to us.

Received for review February 25, 2005. Accepted July 4, 2005. AC050343L

Analytical Chemistry, Vol. 77, No. 17, September 1, 2005

5541