Nanoscale characterisation of soot particulates from gasoline direct injection engines

Understanding the intricacies of particulate emissions from internal combustion engines is important due to their harmful impact on human health and the environment, as well as their contribution to engine wear and performance. Particulate emissions were historically associated with diesel engines which are predominantly using direct-injection systems. Thus, diesel soot is widely covered in the literature. However, the increasing market share of vehicles with gasoline direct-injection (GDI) engines over recent years raises the question of how GDI soot compares to diesel soot and if existing knowledge can be applied. The aim of this work was to close this gap in knowledge between GDI and diesel engine soot. Fringe analysis was assessed as a tool for quantifying graphitic nanostructures from transmission electron microscopy (TEM) images. A considerable influence of the processing parameters on the produced metrics was demonstrated, and optimised parameters were proposed. Moreover, the importance of the TEM focus point and the role of image quality was outlined. Subsequent thermogravimetric analysis of soot-in-oil samples suggested that the soot deposition rate into the lubricating oil is similar for GDI engines compared to diesel engines, even though their exhaust particulate emissions are generally considered to be one order of magnitude lower. Direct comparison of GDI and diesel engine soot samples and a carbon black identified primary particles with similar core-shell nanostructures in TEM images. However, for the GDI samples, also particles with surrounding amorphous layer were observed along with entirely amorphous particulates and traces of wear and oil chemistry. Fringe analysis revealed that fringes of GDI soot were distinctly shorter compared to the other soot types. This finding was confirmed by Raman spectroscopy, indicating that GDI soot is more disordered. Electrical mobility measurements of particulate emissions were acquired for a GDI engine with a differential mobility spectrometer (DMS). As additional processing is required to compare the detailed particle size distributions to the regulatory solid particle number (SPN), different methods were assessed. While lognormal function fitting can be sufficient for SPN23 measurements, modelling of counting efficiencies by applying digital filtering functions is required for measurements below 23 nm. A new function was designed to match the proposed counting efficiencies for SPN10 of upcoming regulations. Measurements with the DMS combined with a catalytic stripper showed an increase of up to 11.2% using this new function compared to the closest previous sub-23 nm function. However, the results are highly dependent on the shape of the particle size distribution. For a matrix of test conditions, the shift from SPN23 to SPN10 was observed to result in increases of 27% to 390%. Furthermore, soot particulates were sampled from the exhaust gas on TEM grids for three operating conditions. Core-shell primary particles were observed for all conditions. In addition, some particles at 1500 rpm fast-idle exhibited a surrounding amorphous layer. For 1500 rpm with 40 Nm brake torque, crystalline features within agglomerates and entirely amorphous/crystalline particulates could be found. Fringe analysis of the graphitic primary particle nanostructures did not find significant differences between the operating conditions; however, longer fringes than for the soot-in-oil samples were identified. An additional feature observed in all samples were separate sub-10 nm particulates of non-volatile nature. The average diameter of these particulates was below the lower detection size limit of the DMS