Kinetics of trace metal interactions with lake sediments

Abstract

<p>This study evaluates: (1) the use of a radiotracer technique to obtain kinetic data on trace metal interactions with sediments in a freshwater environment, and (2) the importance of kinetic considerations in trace metal partitioning between sediment and water in a lake. The technique consists of measuring the distribution of a small radiotracer spike between the sediment and the aqueous phase over time. The system is assumed to be at equilibrium prior to spiking. The radiotracer uptake by sediments is first determined, the aqueous phase is then replaced by a non-tracer solution, and finally the tracer release is measured. Preliminary studies with cadmium and zinc showed: (1) stable solution conditions (no tracer added) with negligible change in trace metal concentrations and master parameters (pH, DOC) over several weeks, (2) precision in replicate analysis and experiments, and (3) confirmation of the technique as demonstrated with a synthetic TiO₂ suspension. A novel experimental design was devised, consisting of a kinetic model fitting using the "adsorption" kinetic data, followed by confirmation comparing the model data to "desorption" kinetic data. The technique was applied to surficial samples from three lacustrine systems using ⁶⁵Zn and ¹⁰⁹Cd. The kinetics of metal exchange are very similar for all samples for both "adsorption" and "desorption". The general kinetic behaviour was a fast reaction (hours) followed by a slower interaction (weeks). Three kinetic models were investigated to explain the "adsorption-desorption" data. A two-box model (TBM), describes the system as a reversible solid-liquid first-order reaction. It does not, however, explain the slower exchange well. Two other models include two distinctive kinetic reactions. They differ mostly by the nature of the "desorption" reaction prediction. A "two-box model with a leak" (TBL) considers a fast reversible first-order reaction followed by a slower first-order reaction (leak). A "restricted three-box model" (RTB) considers two independent reactions, both first-order and reversible, but of different rates. The TBL and RTB models can be fitted to the "adsorption" data very well. The "desorption" data is, however, better predicted by the TBL model revealing the slower short-term (hours-days) interactions of trace metals with lake sediments. Trace metal partitioning determined by short-term (hours-days) interactions gives higher solution concentration than longer-term (weeks-months) "equilibrium" partitioning. Thus short-term prediction of metal partitioning based on long-term "equilibrium" distribution would overestimate the role of natural sediments as metal sinks. This kinetic approach gives good estimates of the uptake and release of trace metals between sediment and water in lacustrine systems.</p>