Discussion about the use of the volume specific surface area (VSSA) as a criterion to identify nanomaterials according to the EU definition. Part two: experimental approach.

The first part of this study was dedicated to the modelling of the influence of particle shape, porosity and particle size distribution on the volume specific surface area (VSSA) values in order to check the applicability of this concept to the identification of nanomaterials according to the European Commission Recommendation. In this second part, experimental VSSA values are obtained for various samples from nitrogen adsorption isotherms and these values were used as a screening tool to identify and classify nanomaterials. These identification results are compared to the identification based on the 50% of particles with a size below 100 nm criterion applied to the experimental particle size distributions obtained by analysis of electron microscopy images on the same materials. It is concluded that the experimental VSSA values are able to identify nanomaterials, without false negative identification, if they have a mono-modal particle size, if the adsorption data cover the relative pressure range from 0.001 to 0.65 and if a simple, qualitative image of the particles by transmission or scanning electron microscopy is available to define their shape. The experimental conditions to obtain reliable adsorption data as well as the way to analyze the adsorption isotherms are described and discussed in some detail in order to help the reader in using the experimental VSSA criterion. To obtain the experimental VSSA values, the BET surface area can be used for non-porous particles, but for porous, nanostructured or coated nanoparticles, only the external surface of the particles, obtained by a modified t-plot approach, should be considered to determine the experimental VSSA and to avoid false positive identification of nanomaterials, only the external surface area being related to the particle size. Finally, the availability of experimental VSSA values together with particle size distributions obtained by electron microscopy gave the opportunity to check the representativeness of the two models described in the first part of this study. They were also used to calculate the VSSA values and these calculated values were compared to the experimental results. For narrow particle size distributions, both models give similar VSSA values quite comparable to the experimental ones. But when the particle size distribution broadens or is of multi-bimodal shape, as theoretically predicted, one model leads to VSSA values higher than the experimental ones while the other most often leads to VSSA values lower than the experimental ones. The experimental VSSA approach then appears as a reliable, simple screening tool to identify nano and non-nano-materials. The modelling approach cannot be used as a formal identification tool but could be useful to screen for potential effects of shape, polydispersity and size, for example to compare various possible nanoforms.