In France and more generally speaking in Europe, passive sampling is an emerging method to measure groundwater quality and as a consequence to monitor contaminated sites. Coupled with natural vertical flow measurements in monitoring wells, it appears very efficient for a vertical characterization of the pollution in groundwater. In the USA, the scientific literature is abundant on the subject and a lot of studies were conducted to assess their efficiency to measure groundwater quality. Nevertheless, in the frame of a previous project, a survey was conducted amongst French consultants and showed that passive samplers were not widely used in France because there was a need to provide feedback and guidelines to consultants to promote the use of passive samplers in a regulatory context. In this context, the work presented here aimed at testing passive samplers to measure groundwater quality at a site contaminated with chlorinated solvents and heavy metals, in order to write codes of best practices for their use. This work was conducted in the frame of the CityChlor project. The aim of this project is to improve the quality and minimize the pollution of soil and groundwater by developing an integrated approach to tackle the threats caused by contamination with chlorinated solvents in urban areas. One of the goals of the project is to define the most reliable, fast and cost effective techniques to detect, characterize and delineate chlorinated solvent pollution in groundwater. In this context and thanks to the literature on the subject, passive samplers seem to be an appropriate tool to reach this goal. Prior to the use of passive samplers, the groundwater contamination on site was characterized and 5 monitoring wells were installed at different depths. 4 passive samplers were tested in these wells: PDBs (Polyethylene Diffusion Bags), Ceramic Dosimeters, Gore Sorber Modules and Dialysis Membranes. The tests consisted in comparing the concentrations in groundwater given by these passive samplers to those obtained with the conventional sampling method, that is to say well purging prior to groundwater sampling with a pump. The number of generated data was quite large that is why a statistical analysis was conducted on them. Starting from the principal component analysis theoretical approach, we have developed an original statistical method consisting of a simultaneous analysis of the spatio-temporal correlations between pollutants. Results allowed the characterization of the relations between the passive and the conventional sampling methods. Chlorinated solvent concentrations in groundwater given by the tested passive samplers were consistent with the ones obtained from the conventional sampling method. The statistical analysis allowed us to identify the efficiency of the passive samplers as regard to the pollutant type, the concentration ranges and the measurement depth. Our results showed that passive samplers were very interesting to monitor groundwater at a contaminated site. In addition, they were very easy to use, generally cheaper than the conventional sampling method (particularly the PDBs) and cross-contamination was avoided. They could as well offer complementary information compared with traditional sampling method because they allowed depth discrete and multi-level sampling in a well. Consequently, they seem to be very promising tools for the long term monitoring of groundwater on well-characterized contaminated sites.