Diagnosing trace metals contamination in ageing stormwater constructed wetlands by portable X-ray fluorescence analyzer (pXRF)

Abstract

Stormwaters in urban catchments are known to be contaminated with fine particles, trace metals, nutrients, hydrocarbons, etc., which are a source of environmental degradation. Yet, the Water Framework Directive (2000/60/CE) imposes to meet a “good status” for every water body in Europe. That is why the city of Strasbourg (Eastern France) has built in 2012 three stormwater constructed wetlands (SCW#1, 2 and 3) between the outflow of three urban residential catchments and the Ostwaldergraben, an urban stream achieving a mediocre ecological status at the time. The goals of this project are to smooth out the peak flows and to mitigate stormwater micropollution. Each SCW is composed of a stormwater pond followed by either a vertical flow reed bed filter (RBF) respectively 15x6 m and 25x4 m for RBF#1 and RBF#3 or an horizontal flow reed bed filter for RBF#2 (40x6 m). Substrates in the RBF are resp. 20 cm deep of sand 0/4 (RBF#1), 60 cm deep of gravel 4/10 (RBF#2) and 30 cm deep of 0/4 sand (RBF#3). Those infrastructures face ageing issues. While stormwater ponds show sediments accumulation, RBF are more likely to undergo clogging or chemical saturation. Moreover, metallic micropollution is mainly particle-bound and storage in sediments is conservative. Thus, sediment acts as the sink compartment of the SCW. Yet, the dynamics and the long term accumulation of the trapped trace metals (TMs) remains poorly known. The aim of our work is to study the distribution of TMs in the solid matrix (ponds sediments and RBF substrate) thanks to a portable X-ray Fluorescence (pXRF) analyzer. This technology will allow to multiply in situ measurement points and non-destructive campaigns, together with a reasonable cost. The objective is to illustrate spatial and temporal distributions of TMs in the three SCW. We work with an Olympus C-series Vanta pXRF (silicon diffusion detector, 8-40 keV X-ray tube, Rh anode). Two beams are emitted: beam 1 (B1) at 40 keV to analyze elements from V to U and beam 2 (B2) at 10 keV to analyze light element from Be to Ti, at concentrations ranging from ppm to a few %. Many parameters related to the sample or the pXRF can significantly distort the measurement such as the water content, the homogeneity and texture, the shooting time of the beams, etc. This is why it was important to conduct a preliminary step to calibrate the instrument. No standard specifies a method for the measurement and most existing studies recommend a site-specific and matrix-specific calibration. Appropriated statistical tests using R software (version 4.0.2) allowed to choose the best shooting time for B1 and B2 and the best texture among gravel, rough crushed gravel or sieved samples (250 µm and 125 µm) collected from RBF#2. As RBF#1 and RBF#3 have fine (sand) substrate, no crushing step would be necessary. Thereafter, onsite campaings were conducted. We set a regular meshing of resp. 16 (RBF#1), 24 (RBF#2) and 16 (RBF#3) meshes. On each mesh we collected a triplicate at four deepths (organic horizon if present, surface, 10 cm depth, and 22 cm depth). Then, a pretreatment was applied in the laboratory to the samples (drying and, in case of RBF#2, crushing) before shooting them with the pXRF in optimal conditions. The influence of water content and organic matter on analyzes accuracy was carried out. Finally, three-dimensionnal views of TMs infiltration through aged stormwater wetland may provide quantitative and qualitative informations about micropolluants fates on sediments, sand, soils and roots.