Treatment of losses of ultrafine aerosol particles in long sampling tubes during ambient measurements

Abstract Long sampling tubes are often required for particle measurements in street canyons. This may lead to significant losses of the number of ultrafine (those below 100 nm) particles within the sampling tubes. Inappropriate treatment of these losses may significantly change the measured particle number distributions (PND), because most of the ambient particles, by number, exist in the ultrafine size range. Based on the Reynolds number (Re) in the sampling tubes, most studies treat the particle losses using the Gormley and Kennedy laminar flow model (Gormley, P.G., Kennedy, M., 1949. Diffusion from a stream following through a cylinderical tube. Proceedings of Royal Irish Academy 52, 163–169.) or the Wells and Chamberlain turbulent flow model (Wells, A.C., Chamberlain, A.C., 1967. Transport of small particles to vertical surfaces. British Journal of Applied Physics 18, 1793–1799.). Our experiments used a particle spectrometer with various lengths (1.00, 5.47, 5.55, 8.90 and 13.40 m) of sampling tube to measure the PNDs in the 5–2738 nm range. Experiments were performed under different operating conditions to measure the particle losses through silicone rubber tubes of circular cross-section (7.85 mm internal diameter). Sources of particles included emissions from an idling diesel engine car in a street canyon, emissions from a burning candle and those from the generation of salt aerosols using a nebuliser in the laboratory. Results showed that losses for particles below ≈20 nm were important and were largest for the smallest size range (5–10 nm), but were modest for particles above ≈20 nm. In our experiments the laminar flow model did not reflect the observations for small Re. This may be due to the sampling tubes not being kept straight or other complications. In situ calibration or comparison appears to be required.