Characterization of soot is of interest to many researchers working on environmental aerosol and combustion diagnostics over decades. In this study the morphology of soot aggregates issued from incomplete combustion is investigated based on the need of their radiative properties. The state-of-art of the work on soot from incomplete combustions comprises the characterization techniques based on in situ and ex situ techniques. Due to the limitations and uncertainties arising from in situ measurements in a combustion flame for the definition of morphology for a dispersed group of aggregates, ex situ observations under electron microscopy are often referred to in the literature for aggregate samples collected inside the flame. The latter ex situ observations include the fractal analysis of 2D SEM/TEM images, the derivation of 3D geometry obtained by TEM tomography and the generation of numerical aggregates based on common fractal parameters. In our study, soot aggregates are extracted from a laboratory scale rich propane-air flame by thermophoretic sampling onto metal thin plates. Different from the previous researches, 3D geometry of soot is obtained here by SEM tomography: the geometry reconstruction techniques are successfully applied to a relatively small series of tilted SEM images of high resolution using the "seeing through" phenomenon occurring in nano sized material samples with low atomic number. The observation of soot monomer size under TEM microscopy allows us to test the errors that may arise from the SEM imaging technique (like the pollution deposition on the substrate due to charging, the image artifacts due to the nature of secondary electron emissions) and from the simplified sampling procedure (not using the conventional holey carbon films but metal thin plates). The obtained 3D soot geometry is promising in terms of reducing the analysis time, simplifying the sampling procedure and expanding the 3D observation applicability by not limiting the material sampling to depositing soot and aerosols on TEM grids. (C) 2015 Elsevier Ltd. All rights reserved.