An aqueous solution of HNO3 is introduced drop wise intothe volatilization chamber (Figure 2). The chamber is filled with glass beads andheated to 85 to 90 ◦C using a water bath. Dry air is passed through the volatilizationchamber, constructed from standard glass components, and the air stream carriesthe volatilized HNO3 and water vapor to the CSTRs.The concentrations of aqueous HNO3 solutions used varied between 10:1 (v/v)and 50:1 (v/v) deionized distilled water: concentrated HNO3. One-liter batcheswere large enough to provide for up to two month-long exposures in four CSTRsat 25 to 150 µg – HNO3 m−3without replacement. The solution pump used inthe HNO3 delivery system was a piston-type manufactured by Fluid Metering Inc. Figure 2. Diagram of the volatilization chamber. The volatilization chamber is immersed in a waterbath at 92 ◦C. Dilute aqueous solutions of HNO3 are dripped into the chamber onto a bed of glassbeads. Scrubbed, dry air is pumped through the chamber, picking up the volatilized HNO3 anddelivering it to the CSTRs. All components of the chamber are made from glass.(Oyster Bay, N.Y., U.S.A.,) model QG6, which has a range of 1.0 to 25 mL perhour. In general, very low flow rates of 2.2 to 3.0 mL hr−1were used in this system.Ambient air was dried to a relative humidity less that 1%, with a heatlessair drier (Purgas Heatless Air Dryer model HF200-12-143, General Cable Corp.,Westminster CO, U.S.A.). Before being introduced into the HNO3 volatilizationchamber, the air stream passed through an activated charcoal canister and a HEPAfilter capsule (model 12144 Gelman Sciences, Ann Arbor, MI, U.S.A.). Becausethe air compressor and dryer are a serious noise hazard, they were installed in aprotective structure outside the greenhouse and headhouse. Air was delivered from the air dryer at 65 psi but reduced by a pressure-reducing valve to about 3 psi beforeentering the volatilization chamber.The purified air stream is introduced into the bottom of the HNO3 volatilizationchamber (Figure 2). As the air flows upward to the exit port at the top of the cham-ber, the air stream scavenges the volatilized HNO3. Once out of the volatilizationchamber the HNO3 vapor is distributed to the inpidual CSTRs through a glassmanifold, fabricated from a standard borosilicate 20×150 mm screw cap glasstest tube. A professional glassblower added ten screw-type fittings to the side ofthe tube to enable connection of the Teflon delivery tubes. One of the ports isconnected to a needle valve that regulates direct evacuation of excess HNO3 vapor.The concentrations of HNO3 delivered to the chambers were regulated by (i) vary-ing the concentration of the aqueous solution, (ii) adjusting the evacuation needlevalve, releasing HNO3 prior to delivery into the CSTRs, (iii) changing solutionpump speed or (iv) restricting flows with a needle valve installed in-line of eachdelivery tube at the CSTRs. The strength of the aqueous HNO3 solutions has signi-ficant effects on the efficiency of vapor delivery; solutions that are too dilute tendto saturate the air stream with H2O causing condensation along the walls of theglass and Teflon tubing. This is a particular problem during cool, cloudy weather.Once water droplets form, HNO3 dissolves in the condensate and little is deliveredto the CSTRs. Solutions that are too concentrated require ultra low pump speedsand volatilization tends to occur in pulses resulting in inconsistent delivery. Theaddition of the evacuation valve enabled higher concentrations of solutions to beused. Automatic timers were added to the HNO3 delivery system to control thediurnal pattern of HNO3 concentrations.2.1.2. Continuously Stirred Tank Reactors (CSTRs)The CSTRs are housed in a multiple use greenhouse equipped with particulateand charcoal filtration of incoming air. An independent blower installed in thegreenhouse coupled with an exhaust blower provides the chambers with 1.5 airexchanges per minute under slightly negative pressure.
Air supplied to the CSTRs from the greenhouse is further purified by perman-ganate embedded chemisorbent/absorbent filters installed on the intake duct ofthe blower (model: 4-inch Purafilter B-850-4404, Purolator Products Air FiltrationCo., Henderson NC, U.S.A.). Permanganate traps HNO3, a potential contaminantdue to normal greenhouse operations. All chambers were connected independentlyto the same blower; there are no connections between CSTRs. Nitric acid is intro-duced into the CSTRs through a port in the air duct 0.5 m upstream of the CSTRs(Figure 3).The CSTRs are similar to those originally described by Heagle and Philbeck(1979) (Figure 3). They are constructed of wood and metal covered with Teflonfilm. All exposed surfaces inside the tanks are also coated in Teflon and the bottomof the tank is lined with a 2 mm Teflon sheet. The tanks are 1.35 m dia and 1.35 mht. Each is fitted with a 0.6 m by 1.2 m hinged door. Internal air circulation is provided by continuous speed impellers (Dayton, Model 22811A) mounted in thetop of the tanks.2.1.3. Monitoring SystemThe monitoring system provides continuous sampling from each of the CSTRs.A sampling port was installed approximately one-third of the way up the wall ofthe chamber, opposite the air-supply vent. The air sample from the CSTR is feddirectly into a molybdenum converter (‘Molycon’ Monitor Labs Inc., Englewood,CO, U.S.A.) mounted next to the chambers (Figure 3). The reduction of HNO3 toNO before transport to the monitoring instruments in the headhouse is critical formonitoring stability because of the very high deposition velocity of the pollutant.As with the delivery lines, all sample lines were of equal length.Nitric acid vapor concentrations in the inpidual chambers are monitored witha nitrogen oxide monitor model 8840 (Monitor Labs Inc., Englewood, CO, U.S.A.)using a chemiluminescence method. Nitrogen oxide monitors measure bulk NOxand can only separate and identify NO from most other forms of oxidized N.Therefore, under ambient conditions nitrogen oxide monitors cannot be used formonitoring HNO3 directly. However, because of the air purification systems inthe bulk air supply, ambient levels of NO, NO2, HNO3, are very low and there- fore oxidized N compounds detected within the CSTRs as NO are presumed tobe HNO3. Simultaneous monitoring of the bulk greenhouse air with a separatesystem aided in identifying species and attribution of the NO readings. Additionalverification was accomplished by monitoring air delivered to the 0-HNO3, con-trol chambers. A modified scanivalve (Scanivalve Corp., San Diego CA, U.S.A.)directs the incoming sample to the nitrogen analyzer. Air in the samples lines ismeasured sequentially as the scanivalve rotates through the chambers and the bulkgreenhouse air. A Campbell CR21X datalogger (Campbell Scientific, Inc. LoganUtah, U.S.A.) controls the sampling sequence and duration. Continuously stirredtank reactors and the ambient greenhouse air are normally monitored for 6 to11 min per sample line, however all sample lines are continuously purged. Aftera 1 min equilibration period, concentrations are recorded every minute and thenaveraged. As each chamber is monitored, the concentrations are recorded on a stripchart and a Campbell CR21X datalogger. The scanivalve also allows for manualcontrol over the chamber sequence, or for extended monitoring of single CSTRs.Manual controls were used for determination of monitoring consistency and theeffects of altering condition such as opening chamber doors or changes in HNO3concentrations.2.2. COMPARISON OF THE REAL TIME NITROGEN MONITOR SYSTEM TOHONEYCOMB DENUDERSFor evaluation of the efficiency of the NOx monitoring system in determiningHNO3 concentrations, recorded concentrations were compared to concentrationsmeasured by honeycomb denuder (Koutrakis et al., 1993). Denuders systems arestill the best-accepted method of determining HNO3 and NO−3 concentrations inde-pendent of the other oxide forms (Sickles, 1992; Slanina et al., 1992). The denuderswere mounted outside of the chambers with short sample tubes extended througha sampling port on the door. Denuders were exchanged twice daily in coordin-ation with the delivery system: just before the delivery system began producingHNO3 vapor in the morning and just after the delivery system was turned off in theevening. The denuders were extracted and atmospheric concentrations calculatedas described by Ogawa & Company (1995). The comparison study was conductedfor 5 days and tested 3 CSTRs set for high, moderate and zero pollutant concen-trations as well as the bulk greenhouse air. Evaluations have been repeated severaltimes. Total exposures (cumulative doses) were calculated by integration under theconcentration curves for each method. The use of a cumulative dose allowed for ac-curate comparisons of exposure intensity among the different treatment chambersand across independent experiments.2.3. DEPOSITION STUDIESDeposition rates can be quantified by several means (Seinfeld and Pandis, 1998).Because of the large volume of air (2.65 m−3) and the rapid air exchange (1.5 exchanges min−1), the common inlet/outlet method is inappropriate for the CSTRs.Therefore, deposition is typically measured by surface accumulation. Two examplesare displayed here: (i) inert soil surfaces and (ii) biologically active plant material.In both cases the approach is similar. Samples are placed in the chambers andsubsampled periodically during the experiment. For soils, samples are weighed intoaluminum weigh boats having an exposed surface area of 20 cm2. At each samplingdate, 3 replicate weigh boats were removed and the samples extracted by standardmethod (Maynard and Kalra, 1993). For plant tissue studies, four shrub speciesnative to southern California were tested, Artemisia (Artemisia californica), Brit-tlebush (Encelia farinose), buckwheat ((Eriogonum fasciculatum), and white sage(Salvia apiana). Whole leaves were removed and placed into 50 mL plastic cent-rifuge tubes with 20 mL nano-pure water. The contents are shaken for 30 sec andthe extract or wash solution is analyzed for NO−3 by continuous flow analyzer. Theleaf area for each sample is measured and deposition is calculated on an area basis.Both studies were run for 4 weeks with samples collected weekly. Two chamberswere employed. The data shown are the washable NO−3 concentrations from eachsurface at a calculated dose. Atmospheric concentrations as determined by honeycomb denuder were comparedto values determined by the nitrogen oxide monitors from a high HNO3 treat-ment chamber are shown in Figure 4. The apparent square waves generated by thedenuders are an artifact of the 12 hr air sampling period. The calculated concentra-tions are hourly averages over the exposure time. The actual, real-time values de-termined by the Nitrogen oxide monitor were much less consistent as demonstratedby the nitrogen monitor-derived data. However, when each curve was integratedand the areas compared, the total exposures, or doses, were similar, 21 vs. 29 µg-Nm−3×hr, (Figure 4, legend and Table I). The integrated concentrations for thehigh chamber were 36% higher by denuder method (Table I). 连续搅拌釜反应器系统英文文献和中文翻译(2):http://www.751com.cn/fanyi/lunwen_39742.html