Search

Accurate physical and chemical characterization of the aerosols from various sources, such as urban/industrial emissions, biomass burning to dust intrusion events, paramount to unveiling the impact on air quality, radiative forcing and public health, still presents a big challenge. The first step to assess the impact of aerosols is real-time light absorption measurements, typically characterizing spectral dependence with an absorption Angstrom exponent (AAE) approach (Liu et al 2018). Using new model of Aethalometer, AE36s (Aerosol Magee Scientific), with an enhanced spectral resolution, further helps with improved characterization and distinction between different collected aerosols revealing several specific events and sources of aerosol emission. Unfortunately, filter photometers cannot be equipped with analysers which would uncover the distribution and physicochemical properties of collected aerosols on single particle scale, important for an accurate interpretation of the real-time measurements. In the presented study we show for the first time capability of successfully implementing machine learning-based smart characterization of collected aerosols from dust intrusion events in Europe (Barcelona, Ljubljana) and the Middle East (Qatar), to support real-time Aethalometer measurements indicating significant black carbon (BC) and absorbing fraction of organic aerosols (brown carbon, BrC) presence. Briefly, quartz fiber filters with collected aerosols were transferred from the measuring sites to the imaging instrument, Helium Ion Microscope (HIM), which provides unique properties: sub-nm lateral resolution, nm surface sensitivity and high depth-of-field (Hlawacek et al 2014), enabling imaging of aerosols deep into the fibers (Figure 1), not capable with Scanning Electron Microscope (SEM). Imaging at different magnifications enabled accurate analysis of aerosol concentration, size distribution and detection of morphologies at a single particle scale. Imaging contra

Accurate physical and chemical characterization of the aerosols from various sources, such as urban/industrial emissions, biomass burning to dust intrusion events, paramount to unveiling the impact on air quality, radiative forcing and public health, still presents a big challenge. The first step to assess the impact of aerosols is real-time light absorption measurements, typically characterizing spectral dependence with an absorption Angstrom exponent (AAE) approach (Liu et al 2018). Using new model of Aethalometer, AE36s (Aerosol Magee Scientific), with an enhanced spectral resolution, further helps with improved characterization and distinction between different collected aerosols revealing several specific events and sources of aerosol emission. Unfortunately, filter photometers cannot be equipped with analysers which would uncover the distribution and physicochemical properties of collected aerosols on single particle scale, important for an accurate interpretation of the real-time measurements. In the presented study we show for the first time capability of successfully implementing machine learning-based smart characterization of collected aerosols from dust intrusion events in Europe (Barcelona, Ljubljana) and the Middle East (Qatar), to support real-time Aethalometer measurements indicating significant black carbon (BC) and absorbing fraction of organic aerosols (brown carbon, BrC) presence. Briefly, quartz fiber filters with collected aerosols were transferred from the measuring sites to the imaging instrument, Helium Ion Microscope (HIM), which provides unique properties: sub-nm lateral resolution, nm surface sensitivity and high depth-of-field (Hlawacek et al 2014), enabling imaging of aerosols deep into the fibers (Figure 1), not capable with Scanning Electron Microscope (SEM). Imaging at different magnifications enabled accurate analysis of aerosol concentration, size distribution and detection of morphologies at a single particle scale. Imaging contra