DGT-labile As, Cd, Cu and Ni monitoring in freshwater: Toward a framework for interpretation of in situ deployment

•Defining a limit of quantification is more robust than blank subtraction.•The filter membrane used alters diffusion coefficients in the device.•Membrane fouling does not alter diffusion coefficients in this river system.•Not considering the DBL is satisfactory if its thickness is lower than 0.5 mm.•A framework is proposed to limit biases in DGT-labile concentration estimate.

The use of the Diffusive Gradient in Thin Film sampler (DGT) as a monitoring tool for regulatory programs is currently evaluated. In this context, the impact of commonly followed procedures on the accuracy of DGT-labile As, Cd, Cu, and Ni quantification was studied. Initial sampler contamination yields to define quantification limits instead of using blank subtraction, thus avoiding artifact concentrations. Not considering the alteration of element diffusion by the filter membrane leads to significant underestimation. However, diffusion coefficients determined on a non-fouled membrane were found to be suitable for the studied site, making it possible to use data from the literature. When diffusive boundary layer formation is neglected, no loss of accuracy is recorded provided the layer is thinner than 0.5 mm. Finally, exploration of potential biases allowed initiating a framework that might help limit inaccuracies in DGT-labile concentration estimation and interpretation, especially in a low contamination context.