A critical evaluation of short columns for estimating the attachment efficiency of engineered nanomaterials in natural soils

Short, saturated packed columns are used frequently to estimate the attachment efficiency (α) of engineered nanomaterials (ENMs) in relatively homogeneous porous media, but a combined experimental and theoretical approach to obtain α-values for heterogeneous natural soils has not yet been agreed upon. Accurately determined α-values that can be used to study and predict ENM transport in natural soils should vary with ENM and soil properties, but not with experimental settings. We investigated the effect of experimental conditions, and used different methods to obtain soil parameters, essential to calculate α. We applied 150 different approaches to determine α-values from 52 transport experiments using short columns with 5 different natural soils and 20 and 80 nm gold- or 27 nm silver sulphide ENMs. The choice of column end-filter material and pore size appeared critical to avoid overestimating α owing to filter – ENM interactions and/or incomplete saturation of the column. Using a low ionic strength (4.4 × 10–5 mol L−1) artificial rain water as an aqueous medium avoided ENM homo- or heteroaggregation in all soils, as confirmed by single-particle inductively coupled plasma – time of flight mass spectrometry. ENM breakthrough curves could be modelled using colloid filtration theory assuming irreversible attachment only. α-Values calculated from this model, having the grain size represented by a single average size, accounting for dispersivity and effective porosity based on a prior inert tracer test, explained up to 42% of the variance in α as revealed by partial least squares analysis. However, column length and dispersivity remained as important experimental parameters, which calls for further standardisation efforts of column tests with ENMs in natural soils, preferably cross-validated with batch tests.