In the moderate doseCSTR the denuder value was only 3% higher, and 38% higher in the control cham-ber. The source of higher denuder value in the high dose CSTR appears to havebeen caused by a single day’s sample at day 30 (Figure 4). The cumulative dosewas recalculated without day 30, the difference between methods was negligible(data not shown).The HNO3 concentrations from the denuder method shown in Figure 4 are thecombination of HNO3 vapor and NO−3 fine particles. When these two componentsare examined separately, HNO3 vapor was present only during the daylight hourswhen the volatilization system was producing HNO3 vapor, but the particles werepresent at all times (Figure 5). The pattern of NO−3 particulate concentration in thetwo treatment chambers was very similar, even though the vapor HNO3 concentra-tions were about twice as high in the high chamber (Figures 5a and b). ParticulateNO−3 was also present in the control chambers (Figure 5c), although at lower con-centrations than in the treatment chambers. This information suggests a portion ofthe particulate fraction in each of the HNO3 treatment chambers was from generalgreenhouse environment that had not been removed from the air by the filtration system. Another portion of the particles in the HNO3 treatment chambers appearto be the result of HNO3 vapor reacting with ambient NH3 vapor to form NH4NO3aerosols (details are provided later in this section).The two methods of measuring HNO3 vapor seem to be disagreement on thepresence of HNO3 vapor in the CSTRs during the evening hours once the volatiliz-ation system has been turned off.Without the benefit of the denuder measurements,the detection of NO by the nitrogen oxide monitor during the off-hours was origin-ally hypothesized to be due to remobilization and reemission of HNO3 that haddeposited on the chamber walls and tubing during the on-hours. However, data ac-quired from the honeycomb denuders indicate that the vapor phase HNO3 was notpresent during the evening, but particulates containing NO−3 were. This informationsuggests that the NO detected by the on-line nitrogen monitoring system during theoff-hours may be due to particulates, rather that vapor HNO3.The behavior of the molycons and the nitrogen oxide monitors in the presenceof fine particles is unknown, but in the catalytic process the air sample is heatedto over 150 ◦C before passing it over the molybdenum catalyst. It is reasonableto speculate that this could cause thermal decomposition of NO−3 particles intosome nitrogen-oxide containing vapor. The second molycon integral to the nitrogenanalyzer would insure that any non-NO decomposition product released from thefirst molycon associated with the inpidual chambers would be measured as NOonce it reached the detector.Attempts to deconstruct concentration data from the nitrogen analyzer and cor-relate it with honeycomb data by diurnal cycles were not successful, but deductivereasoning suggest that the most likely the source of the NO reading form the Nitro-gen oxide analyzers during off hours are the result of thermal decomposition of theNOx-containing particles rather than the result of remobilization and reemission ofHNO3 vapor deposited on the walls of the chamber or delivery lines.Further indication of the complexity of the HNO3 fumigation systems was re-vealed by investigation of other nitrogenous atmospheric components (Table II).Of all the nitrogenous pollutants, ammonia vapor dominated the nitrogen profile.On average 81% of the measured nitrogen was contributed by NH3 in the con-trol chamber, 66% in the moderate HNO3 chamber and 53% in the high HNO3chamber. Unlike HNO2, which was also present in small amounts (Table II), theNH3 concentrations indicated diurnal fluctuations. Ammonia fluctuations in thetreatment chambers were in opposition to the fluctuations recorded for HNO3.The honeycomb denuder evaluations were repeated several times and each timeNH3 concentrations were similarly high. Ammonia is assumed to be from the bulkgreenhouse air and therefore they would be expected to be identical in all chambers.However, the NH3 concentrations in the control chambers were, on average, 1.5times higher than in the moderate HNO3 chamber and 3 times higher that in thehigh HNO3 chamber. A regression analysis of NH3 levels across HNO3 levels sug-gests that HNO3 acts as a titrant for NH3, resulting in reduced atmospheric concen-trations of NH3 in the presence of high HNO3 and supporting the notion of titration of NH3 gases by HNO3 to form NH4NO3 particles (Figure 6). The stoichiometry ofammonia vapor in comparison to ammonium nitrate particles has been difficult toreconstruct because of the differences in measurement methodologies (real-time asopposed to 12 hr collection periods), but an operational hypothesis of spontaneousparticulate formation under the chamber conditions described seems reasonable.These data suggests that in the treatment chambers, the NO−3 particles meas-ured with the honeycomb denuders were from two sources: spontaneous formationof NH4NO3 in the chambers via titration processes (Figure 6) and dust from thegeneral greenhouse environment. The presence of NO−3 particulates, but no HNO3vapor in the control chambers indicates that the source of the contamination wasfrom the greenhouse air. Particulate concentrations in the two treatment chambersappeared to have similar patterns of high daytime levels and low nighttime con-centrations. The notion is also supported by higher NH+4 particles in the treatmentchambers (Figure 5 and Table II). The effect of airborne NH+4 ,NO−3 or NH4NO3particles on biological systems has not been fully explored, however, should not beignored. In other fumigation systems, e.g. Marshall and Cadle (1989), Norby et al.(1989), Hanson and Garten (1992), or Krywult et al. (1996), complete assessmentsof the atmospheric environment of the exposure chambers were not performed, soit is unknown whether this is a typical phenomenon.Two examples of experimental results are included to demonstrate utility of thesystem. Figure 7a shows the relationship between HNO3 dose and deposition asmeasured by extractable NO−3 in clay-sized particles. The relationship is highly Figure 7. A comparison of nitric acid deposition (as measured by surface accumulation and sub-sequent extraction) on biologically inert and biologically active materials. The top panel indicateswashable nitrate form clay-sized soil particles. The lower panel indicates a similar experiment usingactively growing plant material. 连续搅拌釜反应器系统英文文献和中文翻译(3):http://www.751com.cn/fanyi/lunwen_39742.html