Abstract
Particulate matter (PM) has been sampled from a compression ignition engine using a differential mobility spectrometer (Cambustion DMS 500) and for imaging in a transmission electron microscope (TEM) with the aim of coupling these two measuring techniques. A known issue when coupling these two methods is that a devise like the DMS samples all PM, and the TEM only soot. To help resolve this issue, a thermal denuder was designed and built to remove all volatile organic compounds (VOC) from the sample prior to entering the DMS. For TEM imaging, soot was either collected directly onto a TEM grid using the thermophoretic effect or collected onto quartz filters with the soot then transferred onto the TEM grids. The direct to grid technique did not work after the denuder due to the gas temperature being too low for the thermophoretic effect; hence the reason to collect some soot using the quartz filters. Soot was removed from the filters using an ethanol wash/sonication technique. Morphology; diameter of gyration, projected area, primary particle size and fractal dimension have been compared between the two TEM sampling techniques, with or without the denuder. Denuder effectiveness has been assessed using TGA analysis of sampled soot. Issues concerning the sampling process itself are outlined. A comparison between the TEM and the DMS results is conducted with the discrepancies between them discussed. Direct and filter sampling gave similar results as long as the sonication process and grid prep is done properly, otherwise the filter wash technique results in a number of clusters of agglomerates which distorts the post processing and morphological data.