A TEM-based method as an alternative to the BET method for measuring off-line the specific surface area of nanoaerosols

At the present time, no stabilised method exists allowing an estimation of the specific surface area for airborne nanostructured particles (nanoaerosols). Recent toxicological studies have, however, revealed biological effects linked to the surface area of these particles. Only the BET method, which can determine the specific mass surface area of a powder, constitutes a reference both in toxicology and in the materials domain. However, this technique is not applicable to nanostructured aerosols given the mass quantities of particles required (between approximately some mg to hundreds of mg taking into account the limit of quantification of existing BET instruments). To characterise the specific surface area of airborne nanostructured particles, a method based on analysing transmission electron microscopy (TEM) images is proposed. This has recourse in particular to previous work carried out in the area of nanoparticles originating from combustion (soot), and takes into account structural parameters of nanostructured particles including the number distribution of primary particles, their overlap coefficient and the fractal dimension of agglomerates and aggregates. The approach proposed in this work was applied to five commercially-available nanostructured powders of differing natures (SiO 2 , ZrO 2 , Al 2 O 3 , Fe 2 O 3 and Fe 3 O 4 ). This first involved their prior analysis by the BET method and then being placed in suspension in aerosol form using a vortex-type shaker system. The procedure to calculate the specific surface area using image analysis was then applied to the sampled aerosols and compared to the BET measurements. The experimental results obtained on the five nanostructured powders cover a range of specific surface areas from 20 to 200 m 2 /g, the primary particles having mean diameters varying from 7 to 47 nm. Close agreement was observed between the two approaches which, taking into account measurement uncertainties, are statistically equivalent at significance level α = 0.05.