Sodium Thiosulfate-Coated Ceramic Denuders for Ozone Removal in Ultrafine Particle Sampling

Abstract. Ozone (O₃) remaining in sampling air can artefactually alter the chemical composition of collected ultrafine particles (UFP), biasing quantitative analysis of the chemical composition. In this study, we developed and evaluated a sodium-thiosulfate O₃ denuder (TSOD) specifically tailored for UFP sampling and assessed its O₃ scrubbing efficiency, particle losses, and chemical selectivity. In laboratory tests under controlled relative humidity and inlet O₃ levels up to 200 ppbV, the outlet concentration remained consistently between 0 and 0.3 ppbV, demonstrating the O₃ removal efficiency of the TSOD. During an urban field deployment over 7 days O₃ downstream of the TSOD consistently remained at 0 ppbV while ambient O₃ varied between 0 and 65 ppbV. Moreover, for particles with mobility diameters ranging from 10 to 1000 nm, we did not observe any significant losses in particle number concentrations. Using a parallel two-channel UFP sampler (with vs. without upstream TSOD), we quantified O₃-driven sampling artefacts in UFP mass focussing on three types of organic markers. (1) Firstly, we targeted polycyclic aromatic hydrocarbons (PAH), particularly chrysene (Chry), benz[a]anthracene (BaA), benzo[a]pyrene (BaP), indeno[1,2,3-cd]pyrene (IcdP), benzo[k]fluoranthene (BkF), and benzo[b]fluoranthene (BbF). Without upstream O₃ removal, the individual concentration of the PAH were 15–46 % lower. (2) Secondly, for the tire and road wear marker, the antioxidant N-(1,3-dimethylbutyl)-N′-phenyl-p-phenylenediamine (6PPD) and its oxidation product 6PPD-quinone (6PPDq), we observed in-situ ozonation of 6PPD to 6PPDq with transformation yields of about 13 to 20 %. (3) In contrast, biogenic organic acids (bOAs) did not show differences when sampled with or without O₃, as their O₃ reactivity is much lower than the one of the PAH. Moreover, this test indicated that the TSOD did not perturb the gas–particle partitioning of these semi-volatile species. Our results demonstrate that the TSOD (i) efficiently scrubs atmospheric O₃ at relevant mixing ratios, (ii) does not introduce measurable particle losses across 10–1000 nm, and (iii) preserves semi-volatile partitioning.