Abstract

Five different instruments for the determination of the mass concentration of PM₁₀ in air were compared side-by-side for up to 33 days in an undisturbed indoor environment: a tripod mounted BGI Inc. PQ100 gravimetric sampler with a US EPA certified Graseby Andersen PM₁₀ inlet; an Airmetrics Minivol static gravimetric sampler; a Casella cyclone gravimetric personal sampler; an Institute of Occupational Medicine gravimetric PM₁₀ personal sampler; and two TSI Inc. Dustrak real-time optical scattering personal samplers. For 24 h sampling of ambient PM₁₀ concentrations around 10 µg m⁻³, the estimated measurement uncertainty for the two gravimetric personal samplers was larger (∼±20%) compared with estimated measurement uncertainty for the PQ100/Graseby Andersen sampler (<±5%). Measurement uncertainty for the Dustraks was lower (∼±15% on average) but calibration of the optical response against a reference PM₁₀ method is essential since the Dustraks systematically over-read PM₁₀ determined gravimetrically by a factor ∼2.2. However, once calibrated, the Dustrak devices demonstrated excellent functionality in terms of ease of portability and real-time data acquisition. Estimated measurement uncertainty for PM₁₀ concentrations determined with the Minivol were ±5%. The Minivol data correlated well with PQ100/Graseby Andersen data (r = 0.97, n = 18) but were, on average, 23% greater. The reason for the systematic discrepancy could not be traced. Intercomparison experiments such as these are essential for assessing measurement error and revealing systematic bias. Application of two Dustraks demonstrated the spatial and temporal variability of exposure to PM₁₀ in different walking and transport microenvironments in the city of Edinburgh, UK. For example, very large exposures to PM₁₀ were identified for the lower deck of a double-decker tour bus compared with the open upper deck of the same vehicle. The variability observed emphasises the need to determine truly personal exposure profiles of PM₁₀ for quantifying exposure–response relationships for epidemiological studies.