An increasing amount of TiO2 engineered nanoparticles (TNP) is released into soils and sediments, increasing the need for dedicated detection methods. Titanium is naturally present in soils at concentrations typically much higher than the estimated concentrations for TNP. Therefore, a precise knowledge of this natural background, including the colloidal fraction, is required for developing adapted strategies for detecting TNP. In this study, we characterized the natural colloidal Ti-background by analyzing eight soils with different properties and origins. A combination of X-ray fluorescence analysis and ICP-OES was used for determining the silicate bound fraction, which was a minor fraction for all soils (0–32%). The colloidal fraction obtained by extracting colloids from soil prior to ICP-OES measurements ranged between 0.3% and 7%. Electron microscopy and hydrodynamic chromatography confirmed that Ti in the form of colloids or larger particles was mostly present as TiO2 minerals with a fraction smaller than 100 nm. The size distribution mode of the extracted colloids determined using hydrodynamic chromatography ranged between 80 and 120 nm. The chromatograms suggested a broad size distribution with a significant portion below 100 nm. In addition to these data, we also discuss possible implications of our findings for the method development for detecting TNP in soils.

Phosphate fertilization contributes to an input of uranium (U) in agricultural soils. Although its accumulation and fate in agricultural soils have been previously studied, its colloidal transport and accumulation along slopes through erosion have been studied to a lesser extent in viticulture soils. To bridge this gap, the contents and potential mobility of U were investigated in vineyard model soils in the Rhineland-Palatinate region, Germany. In addition to elevated U contents, U was expected to associate with colloids and subject to erosion, thus accumulating on slope foots and in soils with fine structure, and reflecting a greater variability. Moreover, another expectation was the favorable erosion/mobility of U in areas with greater carbonate content. This was tested in three regional locations, at different slope positions and through soil horizon depths, with a total of 57 soil samples. The results show that U concentrations (0.48–1.26 ppm) were slightly higher than proximal non-agricultural soils (0.50 ppm), quite homogenous along slope positions, and slightly higher in topsoils. Assuming a homogeneous fertilization, the vertical translocation of U in soil was most probably higher than along the slope by erosion. In addition, carbonate content and soil texture correlated with U concentrations, whereas other parameters such as organic carbon and iron contents did not. The central role of carbonate and soil texture for the prediction of U content was confirmed using decision trees and elastic net, although their limited prediction power suggests that a larger sample size with a larger range of U content is required to improve the accuracy. Overall, we did not observe neither U nor colloids accumulating on slope foots, thus suggesting that soils are aggregate-stable. Lastly, we suggested considering further soil parameters (e.g., Ca2+, phosphorus, alkali metals) in future works to improve our modelling approach. Overall, our results suggest U is fortunately immobile in the studied locations.

No abstract