The size-resolved cloud condensation nuclei (CCN) activity and its prediction based on aerosol hygroscopicity and composition in the Pearl Delta River (PRD) region during wintertime 2014

A hygroscopic tandem differential mobility analyzer (HTDMA), a scanning mobility cloud condensation nuclei (CCN) analyzer (SMCA), and an Aerodyne high-resolution time-of-flight aerosol mass spectrometer (HR-ToF-AMS) were used to, respectively, measure the hygroscopicity, condensation nuclei activation, and chemical composition of aerosol particles at the Panyu site in the Pearl River Delta region during wintertime 2014. The distribution of the size-resolved CCN at four supersaturations (SSs of 0.1 %, 0.2 %, 0.4 %, and 0.7 %) and the aerosol particle size distribution were obtained by the SMCA. The hygroscopicity parameter κ ( κ CCN , κ HTDMA , and κ AMS ) was, respectively, calculated based upon the SMCA, HTDMA, and AMS measurements. The results showed that the κ HTDMA value was slightly smaller than the κ CCN one at all diameters and for particles larger than 100 nm, and the κ AMS value was significantly smaller than the others ( κ CCN and κ HTDMA ), which could be attributed to the underestimated hygroscopicity of the organics ( κ org ). The activation ratio (AR) calculated from the growth factor – probability density function (Gf-PDF) without surface tension correction was found to be lower than that from the CCN measurements, due most likely to the uncorrected surface tension ( σ s∕a ) that did not consider the surfactant effects of the organic compounds. We demonstrated that better agreement between the calculated and measured ARs could be obtained by adjusting σ s∕a . Various schemes were proposed to predict the CCN number concentration ( N CCN ) based on the HTDMA and AMS measurements. In general, the predicted N CCN agreed reasonably well with the corresponding measured ones using different schemes. For the HTDMA measurements, the N CCN value predicted from the real-time AR measurements was slightly smaller (∼6.8 %) than that from the activation diameter ( D 50 ) method due to the assumed internal mixing in the D 50 prediction. The N CCN values predicted from bulk chemical composition of PM 1 were higher (∼11.5 %) than those from size-resolved composition measured by the AMS because a significant fraction of PM 1 was composed of inorganic matter. The N CCN values calculated from AMS measurement were underpredicted at 0.1 % and 0.2 % supersaturations, which could be due to underestimation of κ org and overestimation of σ s∕a . For SS values of 0.4 % and 0.7 %, slight overpredicted N CCN values were found because of the internal mixing assumption. Our results highlight the need for accurately evaluating the effects of organics on both the hygroscopic parameter κ and the surface tension σ in order to accurately predict CCN activity.