• It is difficult to measure the different chemical forms of As and Hg using a single DGT.
• DGTs were used to assess the spatiotemporal variability of concentrations at various scales.
• In the presence of Hg0, DGT concentrations of Hg are overestimated due to the trapping of Hg0.

Developed around 20 years ago, passive samplers are an alternative to discrete sampling. The sampling of contaminants is realized in-situ and allows to integrate the contamination over the exposure period. Passive samplers have the advantage to lower the quantification limits, to avoid matrix effects during the analysis and to facilitate the sample preservation. For metals, DGT is the most employed technique and allow in its classic configuration to sample cationic metals. These tools were adapted in order to sample mercury and arsenic. Nevertheless, the chemical forms of these contaminants influence their toxicity and their bioavailability which have to be taken into account when characterizing the exposition of aquatic environments. Thus, this thesis aimed to: i) develop and validate, in the laboratory, DGT techniques for the measurement of arsenic and mercury chemical speciation in freshwaters; ii) assess the spatio-temporal variability of mercury and arsenic concentrations and their chemical speciation at different scales in surface freshwaters by passive sampling. Arsenic speciation (AsIII + AsV) with a unique DGT could not be assessed according to the usual procedure since arsenic elution from ferrihydrite binging gel cause an oxidation of AsIII to AsV and therefore a loss of information on the speciation. The setting up of screening plans allowed to optimize this elution step and to elute 22 and 32% AsIII and AsV, respectively, from the ferrihydrite binding gel, with ammonium dihydrogenophosphate (NH4H2PO4) 0.5 M at 75°C with a conversion of 30% of AsIII to AsV. Through the use of corrective factors, AsIII and AsV could be monitored from 0.24 et 0.33 μg.L-1 in average on the exposure period for DGT (7 days) in surface freshwaters at 20°C. For the monitoring of mercury speciation (HgII + MeHg) with a unique DGT, the analysis step proved to be problematic since it necessitates to adapt double isotopic dilution technique for the DGT. While MeHg could be measured by DGT from 0.08 ng.L-1 in surface freshwaters, laboratory tests have highlighted specific difficulties in the measurement of HgII certainly in relation to 3M binding gel properties that could not be identified during this thesis. DGT were then applied in-situ in 4 different contexts in order to evaluate their relevance for integrating spatio-temporal variations of metals concentrations. First, during dam flushing operations on the Rhône River during ~15 days, we have showed that DGT were highly representative of the dynamic of metals concentrations and As speciation in comparison with discrete sampling, notably integrating the increase of Mn, Ni, Co and As (AsIII and then AsV) dissolved concentrations. Then, the use of DGT on the Gier and Deûle sites for the monitoring of mercury speciation have evidenced the consistency between mercury and methylmercury concentrations measured by passive and discrete sampling measurements in dynamic surface freshwaters from weekly to annual scales. Nevertheless, in environmental conditions which favor dissolved gaseous mercury production, Hg concentrations estimated by DGT would be overestimated as a reason of an uptake of these mercury chemical forms by DGT, necessitating to interpret the results with caution. Finally, DGT applications on the Rapel lake watershed in Chile have highlighted DGT capacity to identify the spatial variation of contamination as well as DGT relevance to integrate temporal variations of Hg concentrations linked with hydropower production plant operations

Figure 1: Dissolved concentrations obtained from discrete sampling (M(d)) and DGT concentrations (DGT_M) for AsIII, AsV, and As before, during, and after sediment management operations at Pougny, Seyssel, and Jons, respectively 10, 45, and 160 km downstream of the dam. The average dissolved concentrations (Moy_M(d)) were calculated for each period (7, 14, and 13 days) and for each station, based on the dissolved concentrations from discrete sampling weighted by time. Results below the LQ are represented as half of the LQ.

Publications and communications

• Bretier M., Dabrin A., Bessueille-Barbier F., Coquery M. The impact of dam flushing event on dissolved trace elements concentrations: Coupling integrative passive sampling and discrete monitoring. Sci. Total Environ. 2019, 656, 433-446
• Bretier M., Dabrin A., Panay J., Le Bescond C., Grisot G., Dherret L., Gahou J., Bessueille-Barbier F., Coquery M. Is DGT an efficient tool to integrate transient trace metals concentrations during a major hydrological event on the Rhone River? DGT2017 Conference, Gold Coast, Australia, 6-8 September 2017
• Bretier M., Dabrin A., Panay J., Le Bescond C., Grisot G., Dherret L., Bessueille-Barbier F., Coquery M. Intégration des concentrations métalliques dissoutes lors d’un évènement hydrologique transitoire sur le Rhône : cas des opérations de gestion sédimentaires. 25ème réunion des Sciences de la Terre, Caen, France, 24-28 Octobre 2016
   

Cite the thesis

Marie Bretier. Évaluation de la variabilité spatio-temporelle du mercure et de l'arsenic dans les eaux de surface par échantillonnage passif. Autre. Université de Lyon, 2019. Français. ⟨NNT : 2019LYSE1073⟩. ⟨tel-02495347v2⟩

Access manuscript on HAL thèses