Performance evaluation of four cascade impactors for airborne ultrafine-particle (UFP) collection: the influence of particle type, concentration, mass, and chemical nature

Abstract. Ultrafine particles (UFPs) have aerodynamic diameters of 100 nm or less. As UFPs potentially impact human and environmental health, their chemical composition is of interest. However, their small mass presents challenges for sampling and chemical characterization methods. Therefore, we conducted a comprehensive characterization and comparison of four cascade impactors suitable for separating and collecting UFPs – namely, the 120R Micro-Orifice Uniform Deposit Impactor (120R MOUDI-II), ultraMOUDI, electrical low-pressure impactor (ELPI), and personal nanoparticle sampler (PENS) – under controlled laboratory conditions and in a field application. In the laboratory, we evaluated pressure drops, cutoff diameters, the steepness of the cutoff curve, losses, particle bounce, and transmitted particle mass. We observed that the performance of the impactors varied between 59 and 116 nm in cutoff diameter (electromobility diameter), depending on the impactor's design and the type of test aerosol mixture – salt particles (NaCl), simulated secondary organic aerosol (SimSOA), or soot. All impactors separated UFPs, with the best agreement in cutoff diameters for SimSOA, which showed maximum deviations of about 4 nm. The cutoff curve was steeper for soot compared to SimSOA and NaCl. Pressure drops were measured at 260 ± 1 hPa (PENS), 420 ± 2 hPa (ultraMOUDI), 600 ± 3 hPa (120R MOUDI-II), and 690 ± 3 hPa (ELPI). Losses were assessed as maximum transmissions in the ultrafine fraction at 30 nm, yielding 83 ± 8 % for the PENS, 77 ± 8 % for the ultraMOUDI, 75 ± 8 % for the 120R MOUDI-II, and 69 ± 7 % for the ELPI. We compared two additional impactor-specific factors crucial for mass-based analyses of organic marker compounds: the evaporation of semi-volatile compounds due to a high-pressure drop across the impactor and material addition from larger particles bouncing off upper stages. “Bounce-off” was influenced by the particle number concentration in the sampled air and could be partially mitigated by applying a coating to the upper impaction plates. In the field application, we deployed the four cascade impactors side by side under environmental conditions to sample urban air. We analyzed six markers representing typical UFP sources and various molecular properties using HPLC-MS/FLD (high-performance liquid chromatography with mass spectrometry and fluorescence detection). These markers comprised benzo[a]pyrene (BaP), benzo[b]fluoranthene (BbF), levoglucosan (Levo), pinic acid (PA), terpenylic acid (TA), and N-(1,3-dimethylbutyl)-N'-phenyl-p-phenylenediamine (6PPD). The impactors showed the best agreement for BaP and BbF. BaP had an average mass concentration of 175 ± 25 pg m−3 across all impactors and sampling days. However, concentrations were about 29 % higher when sampled with the PENS and 30 % lower when sampled with the 120R MOUDI-II, indicating a maximum disagreement of nearly 60 %. The mass concentrations of the semi-volatile markers (PA, TA, and Levo) decreased on average from the PENS to the ultraMOUDI, then to the 120R MOUDI-II, and finally to the ELPI. We attributed this tendency to the following two effects. (1) Evaporation losses of these markers were likely driven by the pressure drop within the impactor, which increased from the PENS to the ELPI. (2) Despite the applied coating, bounce-off might have affected the smallest impactors (i.e., the PENS and ultraMOUDI) the most as they have fewer stages for retaining larger particles and fragments.