Field test of four methods for gas-phase ambient nitric acid

Three semi-continuous methods for detecting nitric acid (HNO3) were tested against the annular denuder + filter pack (ADS) integrated collection technique at the Tampa Bay Regional Atmospheric Chemistry Experiment (BRACE) Sydney research station downwind of the Tampa, Florida, urban core. The semi-continuous instruments included: two slightly differing implementations of the NOY-NOY* (total oxides of nitrogen minus that total denuded of HNO3) denuder difference technique, one from the NOAA Air Resources Lab (ARL), and one from Atmospheric Research and Analysis, Inc. (ARA); the parallel plate wet diffusion scrubber + online ion chromatography technique from Texas Tech University (TTU); and the chemical ionization mass spectrometer from the Georgia Institute of Technology (GIT). Twelve hour ADS samples were collected by the University of South Florida (USF). Results for 10 min samples computed from the various higher sampling frequencies of each semi-continuous instrument showed good agreement (R2>0.7) for afternoon periods of the highest production and accumulation of HNO3. Further, agreement was within ±30% for these instruments even at HNO3 concentrations . The USF ADS results were biased low, however, by 44%, on average, compared to the corporate 12 h aggregated means from the semi-continuous methods, and by >60% for the nighttime samples; ADS results were below the corporate mean maximum HNO3 concentration by >30% as well. The four instruments using semi-continuous methods, by contrast, were all within 10% of each otherʹs 12 h mean mixing ratios. While only ARA employed a formal minimum detection limit at 0.050 ppb, error analysis with the other techniques established that at the same level of precision, TTUʹs effective limit was approximately the same as ARAʹs and that ARLʹs limit was 0.030 ppb; analysis for GIT showed no apparent effective limit at the levels of HNO3 encountered in this field study. The importance of sample inlet height for HNO3 measurements was indirectly shown through comparison to previous field work at this site when sample inlet heights ranged from 1.5–10 m and produced systematic discrepancies in HNO3 concentrations correlated with height of more than a factor of 2.