Transport avec échange gazeux du trichloroéthylène vers une nappe aquifère

Two large-scale experiments were conducted on a controlled artificial aquifer referred to as SCERES (Site Contrôlé Expérimental de Recherche pour la Réhabilitation des Eaux et des Sols: 25 × 12 × 3 m). The experimental tool SCERES was completely buried in the subsurface in order to get stable temperature conditions in the aquifer. The hydraulic gradient, flow rate, visualization of the water table and water sampling were managed and monitored in two technical pits located at the upstream and downstream ends of the SCERES basin. It was packed with a main layer of uniform quartz sand and a 0.5 m-deep drainage layer at the bottom of the basin having hydraulic conductivities of 8 × 10^-4 and 6 × 10^-3 m/s respectively. The quartz sand had a mean grain diameter of 0.45 mm, a total porosity of 0.4 and a uniformity coefficient of 2. 1; its longitudinal dispersivity was determined in laboratory column experiments to be approximately 1 mm. The aquifer is composed of a 1 m-thick saturated zone and a 2 m-thick unsaturated zone, making it possible to monitor the propagation of vapours in this zone. The hydraulic gradient of the groundwater was fixed at 0.003 m/m, which corresponded to a flow rate of 0.5 m³/h and an average velocity of approximately 0.4 m/day. The thickness of the capillary fringe was estimated to be approximately 0.25 m as deduced from water profile measurements by exploration with a TDR (Time Domain Reflectometry) probe. In the capillary fringe, water saturation at depths of 1.85 and 1.95 m was approximately 57% and more than 90%, respectively. The contaminant chosen for these experiments was trichloroethylene (TCE), because it is among the most frequently detected volatile organic compounds (VOCs) in subsurface environments. TCE as the pure phase was injected 0.35 m beneath the soil surface of SCERES with an experimental design that maintained a uniform infiltration of TCE in the vadose zone for 15 min. The chosen injection device was built from a stainless steel tank (0-58 m diameter and 0.15 m height), on which 31 screened brass rods were fixed at the bottom and separated by a distance of 0.1 m. Each rod contained four injection holes of 0.2 mm diameter in the lower extremity (JELLALI, 2000; JELLALI et al., 2001). In order to prevent TCE volatilization from the device during injection, the TCE volume in the injection tank was covered with a 0.02 m thick water layer. The resulting vapour and aqueous phases were monitored along with temperature and moisture content in order to investigate the mass transfer of VOC from the unsaturated zone to groundwater. The originality of this research was based on the fact that the controlled SCERES site allowed the study of transfer phenomena in the capillary fringe, which is difficult to reach in a reduced laboratoty physical model or in a real contaminated site. This characteristic offers exceptional opportunity for data acquisition in controlled conditions between laboratory and site scales. This aspect is of importance due to the fact that in this capillary zone, water content varies with depth and this situation causes changes in flow rate affecting the intensity of the pollution flux. This study aimed to quantify, in an experimental way, the pollution flux from the unsaturated zone towards the groundwater. In the first experiment, carried out in the summer (July-September), the infiltrated TCE volume was 5 L. In the second experiment, carried out in autumn (October-December), the TCE volume was 3 L in order to increase the distance between the pollution source and the water-table. These volumes were selected on the basis of previous knowledge of TCE residual saturations determined in laboratory column tests (JELLALI, 2000) and in order to obtain a pollution source limited to the unsaturated zone. The first experiment was conducted without water infiltration to study the dispersion of TCE vapours across the capillary fringe, while the second experiment was carried out with a limited rain infiltration in order to investigate the effect of vapour leaching on groundwater pollution. The transport of TCE was monitored in the vadose zone, the capillary fringe and the groundwater where a comparative analysis of two mechanisms of TCE transfer from the unsaturated zone to groundwater via the capillary fringe was carried out: dispersion and dissolution. In both cases, the coupling of measurements of pollutant concentrations in the unsaturated zone, the capillary fringe and the groundwater of SCERES allowed us to take into account the mechanisms intervening near the source area, on a scale close to that of a real pollution problem. The concentration of dissolved TCE was analyzed by a gas chromatograph equipped with a flame ionization detector (GC-FID) after fiquid-liquid extraction with hexane. The online quantitative analysis of TCE vapours was performed by a multigas monitor equipped with a photoacoustic infrared detector. For the first experiment, the TCE mass fluxes from the vadose zone to groundwater were quantified using the method of JOHNSON and PANKOW (1992) where we applied the analytical solution of GRATHWOHL (1998) and incorporated parameters due to the capillary fringe. For the second experiment, we used the numerical code HYDRUS that allowed the simulation of one-dimensional flow in the unsaturated zone. In this example, the TCE mass fluxes leaving the unsaturated zone to the groundwater due to rain infiltration were obtained from the knowledge of the infiltration flow rate and the measured concentrations at the top of the capillary fringe. The observed results indicate that the hydrodynamic dispersion of TCE vapours within the capillary fringe (vertical dispersion) can cause significant groundwater pollution despite the slowness of the aqueous diffusion in the lower region of the highly water saturated capillary fringe. Vapour leaching due to controlled water infiltration causes more significant groundwater pollution in degree and extent than vertical dispersion. The experimental and analytical results demonstrate, on the one hand, the role of the capillary fringe as a barrier against pollution transfer to groundwater when the only mechanism is hydrodynamic dispersion, and on the other hand, significant enhancement of groundwater contamination due to the capture and leaching of vapours from the vadose zone by infiltrating water.