Closure between measured and modelled particle hygroscopic growth during TORCH2 implies ammonium nitrate artefact in the HTDMA measurements
1Atmospheric Sciences Group, SEAES, University of Manchester, P.O. Box 88, Manchester, M60 1QD, UK
2Labor für Atmsphärenchemie, Paul Scherrer Institut, 5232 Villigen PSI, Switzerland
*now at: Cooperative Institute for Research in Environmental Sciences, University of Colorado, Boulder, CO 80309, USA
Abstract. Measurements of aerosol properties were made in aged polluted and clean background air masses encountered at the North Norfolk (UK) coastline during the second field campaign of the Tropospheric ORganic CHemistry project (TORCH2) in May 2004. Hygroscopic growth factor (GF) measurements were performed at 90% relative humidity (RH) for D0=27–217 nm particles using a Hygroscopicity Tandem Differential Mobility Analyser (HTDMA), while the aerosol composition was simultaneously measured with an Aerodyne aerosol mass spectrometer (Q-AMS). During the clean background events the aerosol was characterised by little size dependence of properties with generally large GFs and inorganic sulphate being the dominant compound. In aged polluted air masses the particles were dominated by inorganic sulphate and nitrate at larger sizes, whereas organics were the largest fraction in smaller particles, thus explaining the trend of smaller GFs at smaller sizes. Organics do contribute to the hygroscopic growth, particularly at small sizes, but generally the dominant contribution to growth at 90% RH comes from inorganic salts. The ZSR mixing rule was used to predict GFs based on the chemical composition, theoretical GFs of pure inorganic salts and a "bulk" GF of ~1.20 for the organics. Good quantitative closure with HTDMA measurements as a function of both particle size and time was achieved in the absence of nitrate. However, GFs were clearly overpredicted at times when a significant fraction of nitrate was present. After careful considerations we attribute the overprediction to substantial evaporation losses of ammonium nitrate in the HTDMA instrument. If true, this implies that the ZSR predictions based on composition might be more representative of the actual "bulk" behaviour of undisturbed ambient particles than the HTDMA measurements.
The simplified model approach using the ZSR rule and a constant organic growth factor made high size and time resolution possible, which has proven to be essential for a valid closure study. The ZSR mixing rule appears to be sufficiently accurate, as the GF predictions are more sensitive to the exact GFs of the inorganic compounds than to the growth factor of the moderately hygroscopic organics. Therefore a more detailed analysis and modelling of the organic fraction at the expense of time and size resolution is not worth the effort for an aged aerosol and discrepancies in either direction might even be cancelled out by averaging.