Demonstrating sorption analogy of lanthanides in environmental matrices for effective decision-making: The case of carbon-rich materials, clay minerals, and soils

Abstract

Examining the effect of lanthanide-contaminated wastes, which have the potential to impact to other environmental compartments, requires conducting interaction studies with soils, as feasible first receptors of lanthanide leachates, and, if necessary, with sorbent materials, such as clay minerals and carbon-rich materials, which can serve as natural barriers and immobilisation agents used in remediation strategies. In this context, it is relevant to have available and reliable data on solid–liquid distribution coefficients (Kd) to understand the lanthanide sorption in these environmental matrices. Moreover, confirming lanthanide sorption analogies permits filling data gaps and data extrapolation among different contaminated scenarios, and thus facilitate to have available input data for decision-making related to the impact of a contaminated site. In this study, we demonstrate for the first time an analogous sorption of La, Sm, and Lu in carbon-rich materials (i.e., biochar and activated charcoal), clay minerals and soils, through laboratory batch experiments. The obtained sorption Kd values revealed similar sorption patterns among the three lanthanides for each matrix tested, even at different initial lanthanide concentrations. In all matrices, the maximum Kd values exceeded 104 L kg−1, with a significant decrease when testing high lanthanide concentrations. The analogy was first confirmed by examining the Kd correlations for the La-Sm, Lu-Sm, and La-Lu pairs within each matrix, for which strong linear correlations were obtained in all cases. Data compilations were built with own and literature data, and derived cumulative distribution functions revealed statistically equal lanthanide distributions and Kd best estimates. In addition to this, Kd variability decreased when grouping the data according to significant material properties. For the first time, Kd (Ln) best-estimates for different scenarios and materials were proposed as input data for risk assessment models.