Rare earth elements binding humic acids: NICA–Donnan modelling

Abstract

Environmental context Rare earth elements (REEs) are technologically critical elements released into the environment by various anthropogenic activities, and whose ecotoxicological impacts are still largely unknown. REE binding to natural organic matter (NOM) is key to understand their fate and bioavailability in the environment. With this work, it is now possible to predict REE binding to NOM in various environments using various speciation software (ECOSAT, ORCHESTRA, Visual MINTEQ). Rationale Understanding rare earth element (REE) speciation in different natural environments is important to evaluate their environmental risks because different chemical species of an element may have different bioavailability and toxicity. REEs have a great affinity for particulate and dissolved organic matter, particularly fulvic and humic acids (HAs). Thus, the use of humic ion binding models may help to understand and predict the behaviour and speciation of these species in surface waters, groundwaters and soils. Methodology In this work, we used previously published experimental datasets to parameterise the NICA–Donnan model for REEs binding with HAs, using the model optimisation tool PEST-ORCHESTRA. We propose using linear free energy relationships (LFERs) to constrain the number of parameters to optimise. Results We determined a coherent NICA–Donnan parameter set for the whole REEs series being compatible with available generic NICA–Donnan parameters for other metals. The impact of pH, ionic strength and REE/HA ratio as well as the presence of competitors (Fe3+, Al3+ and Cu2+) on model results is analysed. Discussion We consolidate confidence in our derived NICA–Donnan parameters for REEs by comparing them with the Irving–Rossotti LFER. We also show the general applicability of this relationship to predict and constrain metal-binding parameters for the NICA–Donnan model. We discuss observed shortcomings and provide suggestions for potential improvement of NICA–Donnan modelling.