Filtration behavior of silver nanoparticle agglomerates and effects of the agglomerate model in data analysis

Abstract

In many data evaluation procedures for particle measuring devices and in filtration models, spherical particles are assumed. However, significant fractions of aerosol particles are agglomerates of small primary spheres. The morphology of particles in filtration processes may not be known a priori and if the filtration data are processed with wrong assumption, errors can be induced. In this work, we have quantified such errors for the case of open-structured agglomerates. Filtration efficiency tests with polydisperse silver nanoparticle agglomerates and their sintered spheres were performed. After the sintering process, particles with a compact structure with the shape close to a sphere are obtained, which are referred to as sintered spheres in the present study. The testing method involved generation of particulate forms, passing the particles through the testing section, and measurement of the particle number concentrations and size distributions before and after the filter. Measurements of the aerosols upstream and downstream of the filter were conducted using scanning mobility particle sizers (SMPS, TSI Inc.), which covered the rage from 10 to 480 nm. Particles were additionally characterized from the electron microscopic images and the average primary particle size was determined to be 16.8 nm. The number-size distribution curves were obtained and used for penetration calculation. The penetration was dependent on the particle size and morphology. Silversintered spheres were captured with a lower efficiency than agglomerates with the same mobility diameter because of the stronger interception effect for agglomerates. Data analysis of the number-size distribution for agglomerates was processed based on sphere assumption and using the model for open structured agglomerates developed by Lall and Friedlander. The efficiencies based on total concentrations of number, surface and volume were affected when the agglomerate model was used. The effect was weakest for the total number efficiency and strongest for the total volume efficiency. Filtration efficiency curves for agglomerates were additionally plotted as a function of the volume equivalent diameter or number of primary particles per agglomerate, because the latter parameters provide more information for the agglomerate structure.