A passive sampler for the determination of airborne ammonia concentrations near large-scale animal facilities

Few data are available on the ammonia emissions of large-scale outdoor animal facilities in arid climates such as those found in California's San Joaquin Valley. Passive samplers provide an ideal tool for studying such large and heterogeneous area sources, because they are inexpensive, portable, and fully self-contained. UC Davis passive ammonia samplers incorporate modifications on a previous design, the Willems Badge, for ease of analysis. Citric acid was chosen as a coating medium though it performed as well as oxalic, sulfuric, and tartaric acids. Zefluor PTFE prefilters were used instead of Teflo though both showed the same resistance to diffusion. Citric acid-coated filters were stable for up to 10 weeks, though more so if stored in Petri dishes rather than in the sampling cassettes themselves. The most effective sampler position was found to be in a face-down configuration fixed into the wind to avoid debris and sensitivity to wind shifts. A new method of rinsing the filters within the cassettes by dropwise elution proved highly effective, with 85% of the ammonium being removed in the first 3 mL of the 10-mL rinse volume. Application of the sampler at a dairy in the Joaquin Valley revealed large variations in concentrations at different locations along the downwind fenceline, which correlated with animal populations and activities directly upwind. In addition, large variations in ammonia concentrations were observed in relation to time of day and animal activity. Field blank loadings were of 1.40 microg NH4-N/filter (SD = 0.74 microg NH4-N/filter). Replicate passive samplers placed side-by-side during sampling episodes agreed with a slope of 1.010 (standard error = 0.028). Impingers were used as a reference method to obtain the correlation between filter loadings and air concentrations, yielding an "effective sampling rate" for the passive samplers of 6.18 L/h (error = 0.23 L/h). Using a theoretical calculation, that "effective flow rate" was calculated to be 6.29 L/h. The method's limit of detection was found to be 82.5 microg NH4-N/m3. Wind speed was found to theoretically affect linearity of sampler response only for speeds less than 0.92 m/s.