Membrane-particle interactions in an asymmetric flow field flow fractionation channel studied with titanium dioxide nanoparticles

; ; ; ; (). Membrane-particle interactions in an asymmetric flow field flow fractionation channel studied with titanium dioxide nanoparticles. Journal of Chromatography A, 1334 92-100. Peer reviewed.

dx.doi.org/10.1016/j.chroma.2014.01.066

Asymmetric flow field flow fractionation operated in a multidetector approach (A4F-MDA) is a powerful tool to perform size-classified nanoparticle analysis. Recently several publications mentioned insufficient recovery rates and even retention time shifts attributed to unspecific membrane–particle interactions. One hypothesis to explain this phenomenon is based on the surface charge (zeta-potential) of the membrane material and the particle.

In this study, we investigated in how far the ζ-potential of A4F membrane and particles would determine the outcome of A4F in terms of feasibility, separation efficiency, retention time, and recovery rate, or whether other factors such as membrane morphology and particle size were equally important. We systematically studied the influence of the ζ-potential on the interactions between the most commonly used A4F membrane materials and two representative types of titanium dioxide nanoparticles (TiO2 NP). Furthermore the effect of different carrier media and additional surfactants on the surface charge of membranes and particles was investigated and the influence of the particle size and the particle concentration on the recovery rate was evaluated.

We found that the eligibility of an A4F method can be predicted based on the ζ-potential of the NPs and the A4F membrane. Furthermore knowing the ζ-potential allows to tuning the separation efficiency of an A4F method. On the other hand we observed significant shifts in retention time for different membrane materials that impede the determination of particle size based on the classical A4F theory. These shifts cannot be attributed to the ζ-potential. Also the ζ-potential does not account for varying recovery rates of different particle types, instead the particle size seems to be the limiting factor. Therefore, the proper characterization of a polydisperse sample remains a challenge.