Behaviour of environmental aquatic nanocolloids when separated by split-flow thin-cell fractionation (SPLITT)

A split-flow thin-cell (SPLITT) system in different operating modes was used to size fractionate colloids and particles in a lake water. The unperturbed lake water sample and eluent fractions (nominally alpha<1 mu m; b>1 mu m) were analysed with atomic force microscopy (AFM) to assess the quality of the separation, in particular to quantify the contamination of the b (>1 mu m) fraction with nanocolloids (defined here as material <100 nm in size). Particle size distribution (PSD) results from AFM indicated that there was substantial contamination with nanocolloids. This contamination was, most likely, from diffusive transport across flow regimes within the SPLITT and this is supported by the fact that vertical distances between laminar flow regimes within the SPLITT channel are similar in magnitude, but slightly larger than the mean distances travelled by diffusion during the residence time of particles within the channel. Nevertheless, AFM surface density data showed that the concentration of nanoparticles in the a fraction was 6-9 times higher than in the b fraction, depending on the SPLITT mode, indicating that contamination of the b fraction was limited. Fluorescence data using monodisperse, low molar mass standards confirm the AFM results, with substantial contamination of the b fraction by the fluorescent molecular probes. The increased contamination of the b fraction of the standard molecular probes compared to natural nanocolloids is likely because they are smaller and more diffusive than the average of the natural material. Due to this contamination and the likelihood that these small colloids bind a large fraction of metals, SPLITT can only be used for metal fractionation and speciation in combination with other methods capable of performing further metal speciation analysis. (C) 2008 Elsevier B.V. All rights reserved.