Characterizing the spatiotemporal nitrogen stable isotopic composition of ammonia in vehicle plumes

Abstract

Abstract. Vehicle emissions have been identified as an important urban source of ammonia (NH3). However, there are large uncertainties regarding the contribution of vehicle emissions to urban NH3 budgets, as well as the role of NH3 in spatiotemporal fine particulate matter (PM2.5) formation and nitrogen (N) deposition patterns. The N stable isotopic composition (δ15N) may be a useful observational constraint to track NH3 emission sources and chemical processing, but previously reported vehicle δ15N(NH3) emission signatures have reported a wide range of values, indicating the need for further refinement. Here we have characterized δ15N(NH3) spatiotemporal variabilities from vehicle plumes in stationary and on-road measurements in the USA and China using an active NH3 collection technique demonstrated to accurately characterize δ15N(NH3) on the order of hourly time resolution. Significant spatial and temporal δ15N(NH3) variabilities were observed and suggested to be driven by vehicle fleet composition and influences from NH3 dry deposition on tunnel surfaces. Overall, a consistent δ15N(NH3) signature of 6.6±2.1 ‰ (x‾±1σ; n=80) was found in fresh vehicle plumes with fleet compositions typical of urban regions. Our recommended vehicle δ15N(NH3) signature is significantly different from previous reports. This difference is due to a large and consistent δ15N(NH3) bias of approximately −15.5 ‰ between commonly employed passive NH3 collection techniques and the laboratory-tested active NH3 collection technique. This work constrains the δ15N(NH3) urban traffic plume signature, which has important implications for tracking vehicle NH3 in urban-affected areas and highlights the importance of utilizing verified collection methods for accurately characterizing δ15N(NH3) values.

δ15N-stable isotope analysis of NHx: An overview on analytical measurements, source sampling and its source apportionment

“…Thus, passive samplers can provide confidence in the reproducibility of collected samples. However, recent studies show that δ 15 N(NH 3 ) values measured with the ALPHA passive sampler could be lower than actual values based on the comparison of weekly ALPHA sampling with denuder based active sampler (DELTA) (Pan et al, 2020b), and with honeycomb denuders in Walters et al (2020). In this case, passive samplers may still supplement field sampling of NH 3 measurement for δ 15 N(NH 3 ) analysis after correction by adding 15 ‰ for samples collected using passive diffusion samplers (Pan et al, 2020a).…”

“…Notably, field-based δ 15 N(NH 3 ) source signature for livestock waste were more depleted (-41.1 ‰ to -7.8 ‰) than the instantaneous δ 15 N(NH 3 ) values of liquid manure (Felix et al, 2013), livestock waste (active) (-15.2 ‰ to -8.9 ‰), livestock waste (passive) (-56.1 ‰ to -22.8 ‰), fertilizer use (passive) (-52.0 ‰ to -35.0 ‰) (Felix et al, 2013;Chang et al, 2016;Bhattarai et al, 2020), urban waste (passive) including human excreta (-39.6 ‰ to -37.3 ‰), solid waste (-37.6 ‰ to -29.9 ‰) and waste water (-42.0 ‰ to -39.2 ‰) (Chang et al, 2016), coal-fired power plants (CFPP/active) which includes NH 3 slip from SCR equipped CFPP (-16.1 ‰ to -5.6 ‰) (Felix et al, 2013;Bhattarai et al, 2020) and coal combustion (-7.2 ‰ to 2 ‰) (Freyer, 1978), tailpipe emissions (active) (-9.3 ‰ to 9.0 ‰) (Berner and Felix, 2020), urban traffic (tunnel/active) (2.1 ‰ to 9.2 ‰) (Walters et al, 2020), and urban traffic (tunnel/passive) (-17.8 ‰ to -2.2 ‰) (Felix et al, 2013;Chang et al, 2016;Walters et al, 2020). δ 15 N (NH 3 ) values of different sources were corrected for low bias in passive samplers by adding 15 ‰ to measured δ 15 N(NH 3 ) values (Pan et al, 2020b;Walters et al, 2020). (-31.0 ‰ (Hristov et al, 2009)/-22.5 ‰ (Lee et al, 2011)/ -15.1 ‰ (Lee et al, 2011)) measured on the first or second day during various incubation experiments (Fig.…”

“…This is without consideration of potential fractionation in δ 15 N(NH 3 ) values characterized for ambient NH 3 samples collected with passive diffusion samplers. Notably, based on recent studies (Pan et al, 2020b;Walters et al, 2020), the contribution of nonagricultural sources (e.g., fossil fuel related sources, urban waste) is likely underestimated with isotope-based source apportionment of NH 3 when δ 15 N(NH 3 ) values in ambient NH 3 and δ 15 N(NH 3 ) source signatures are corrected for systematic low bias (Bhattarai et al, 2020;Pan et al, 2020a). However, it is not clear if the bias is significantly different in the shorter sampling intervals (e.g., few hours) for various sources when the concentration is high (e.g., livestock waste, urban wastes).…”

“…3 Isotope source signature of major NH 3 emission sources 3.1 Fossil fuel combustion related sources δ 15 N(NH 3 ) values of fossil fuel combustion-related sources (including emissions from coal-fired power plants and urban traffic) (0.7 ‰AE6.5 ‰) and agricultural sources (-21.9 ‰AE8.7 ‰) were significantly different (two-sample Welch's t-test, p < 0.00001) after correction of δ 15 N (NH 3 ) values for systematic bias in passive diffusion samplers (Pan et al, 2020b;Walters et al, 2020), allowing its potential use in tracing fossil fuel-related sources and agricultural sources of NH 3 in the atmosphere. The heavier δ 15 N(NH 3 ) from fossil fuel sources (-17.8 ‰ to 9.2 ‰ (measured) or -16.1 ‰ to 12.8 ‰ (after correction)) (Freyer, 1978;Heaton, 1987;Felix et al, 2013;Chang et al, 2016;Bhattarai et al, 2020;Walters et al, 2020) is potentially due to NH 3 formation during high-temperature combustion and associated fractionation (Felix et al, 2013), compared to lighter δ 15 N(NH 3 ) values observed in agricultural sources (-56.1 ‰ to -8.9 ‰ or -41.1 ‰ to -7.8 ‰ after correction) (Freyer, 1978;Heaton, 1987;Felix et al, 2013;Chang et al, 2016;Bhattarai et al, 2020) (Fig. 1).…”

“…The variability observed in δ 15 N(NH 3 ) source signatures for urban traffic (-2.8 ‰ to 12.8 ‰) could be likely due to the difference in active and passive collection methods. However, there is still mismatch between δ 15 N(NH 3 ) values of tailpipe emissions in field study (-9.3 ‰ to 9.0 ‰, -1.9 ‰AE5.5 ‰) (Berner and Felix, 2020) and validated sampling techniques using citric-acid coated honeycomb denuders (2.1 ‰ to 9.2 ‰, 4.2 ‰AE1.8 ‰) (two-sample Welch's t-test, p < 0.01) (Walters et al, 2020) (Fig. 1).…”