Systematic low bias of passive samplers in characterizing nitrogen isotopic composition of atmospheric ammonia

Abstract

Identifying the sources of increased ammonia (NH3) in the atmosphere has important implications for air quality, ecosystems, and climate. While analysis of the nitrogen isotopic composition (δ15N) can aid in the apportionment of NH3 emission sources, the widely-used passive collection methods have not been verified for their ability to accurately characterize δ15N-NH3. In this study, three types of passive samplers (ALPHA, Analyst, and Radiello) were employed and collocated with a denuder based active sampler (DELTA, serves as a reference method) to compare and contrast their ability in collecting atmospheric NH3 for subsequent concentration and δ15N-NH3 analysis in urban Beijing during the summer of 2018. The results showed that NH3 concentrations determined using passive collection samplers that included ALPHA (n = 11), Analyst (n = 12), and Radiello (n = 7) samplers were 13.4%, 14.4%, and 27.5% lower than that of the active sampler (n = 12), respectively. The low NH3 concentration bias in the passive collection samplers was suggested to be the result of inaccurate effective sampling rates due to incorrect mass transfer correction factors for the environmental conditions of this study. The δ15N-NH3 values did not show significant differences between passive samplers, with weekly means of −31.0 ± 1.0‰ (n = 4), −28.4 ± 5.5‰ (n = 4), and −29.1 ± 0.9‰ (n = 4) for the ALPHA, Analyst, and Radiello, respectively. Notably, the overall mean δ15N-NH3 values collected by these passive samplers (−29.5 ± 3.2‰, n = 12) was significantly lower than that of the active sampler (−14.1 ± 1.6‰, n = 4), with a difference of 15.4 ± 3.5‰. The large underestimation of δ15N-NH3 values by passive samplers can be explained by the diffusive isotope fractionation of NH3 isotopologues, which is driven by relative mass differences of 14NH3 and 15NH3. Overall, these findings highlight a potential low bias in NH3 concentrations and a substantial uncertainty in previous source interpretation using δ15N-NH3 as previous studies have almost exclusively characterized δ15N-NH3 using passive samplers. We suggest that the denuder-based active sampler should be used in the future for characterizing δ15N-NH3 whether at the source or in an atmospheric environment with mixed sources. This work also has important implications for characterizing δ15N of other gaseous nitrogen compounds.