Choosing the number of images and image position when analyzing the UNC Passive Aerosol Sampler for occupational exposure assessment

Abstract

The University of North Carolina passive aerosol sampler (UNC sampler) could be an alternative when measuring occupational dust exposure, but the time required for microscopic imaging of the sampler needs to be reduced to make it more attractive. The aims of this study were to (1) characterize the effect on precision when reducing imaging, in order to shorten analysis time and (2) assess if the position of the images makes a difference. Eighty-eight samplers were deployed in different locations of an open pit mine. Sixty images were captured for each UNC sampler, covering 51% of its collection surface, using scanning electron microscopy. Bootstrapped samples were generated with different image combinations, to assess the within-sampler coefficient of variation (CV) for different numbers of images. In addition, the particle concentration relative to the distance from the center of the sampler was studied. Reducing the number of images collected from the UNC sampler led to up to 8.3% CV for 10 images when calculating respirable fraction. As the overall CV has previously been assessed to 36%, the additional contribution becomes minimal, increasing the overall CV to 37%. The mean concentrations of the images were modestly related to distance from the center of the sampler. The CV changed from 8.26% to 8.13% for 10 images when applying rules for the image collection based on distance. Thus, the benefit of these rules on the precision is small and the images can therefore be chosen at random. In conclusion, reducing the number of images analyzed from 60 to 10, corresponding to a reduction of the imaged sampling area from 51% to 8.5%, results in a negligible loss in precision for respirable fraction dust measurements in occupational environments.