Influence de la turbidite sur l'efficacite de la desinfection par rayons UV par l'analyse de la cinetique d'inactivation de spores bacteriennes indigenes.

摘要

The use of UV light to disinfect drinking water is now considered a safe and efficient technique, particularly for the inactivation of the two protozoan parasites Cryptosporidium and Giardia. It is generally accepted that the presence of particles impairs UV disinfection processes. Therefore, most systems are preceded by a filtration step to reduce particle concentration. Nevertheless, the interaction between particles and microorganisms, as well the impacts of the presence of particles on UV disinfection, are not well understood. Also, very few models have been developed to describe UV inactivation kinetics in the presence of particles in drinking water.; A UV irradiation protocol has been developed and applied on samples from a high-quality surface water (St-Lawrence River), a surface water impacted by wastewaters (Mille-Iles River), and a wastewater (City of Montreal). The samples were divided in three fractions. A first fraction was blended at high speed (8000 rpm during 4 minutes) with a surfactant (100 mg/L Zwittergent 3-12) prior to monochromatic (254 nm) UV irradiation. A second fraction was blended following irradiation and the last fraction was filtered on 8mum, using a nitrocellulose membrane, prior to irradiation, and then blended. The choice of this specific pore size comes from published data showing 8mum to be the minimal particle size to offer bacterial protection. Also, the pores are large enough to let non-aggregated bacterial spores (1 a 2 mum) through. Inactivation of spores of aerobic spore-forming bacteria (SASFB) was assessed for each part of every sample. This allowed the investigation of the influence of particle on the kinetics of UV disinfection.; Three models, available from the published literature, were used to describe the results obtained from the application of the protocol described in the last paragraph. The models were evaluated in relation to the following criteria: (1) good correlation and precision of the predicted values when compared to experimental results; (2) relative simplicity; (3) potential for an interpretation of the model parameters. The models investigated were Scheible's (1986), Gehr and Wright's (1998) and Mamane-Gravetz and Linden's (2005). This last model was found to be the most adequate in regards to the three criteria.; A parallel Chick-Watson model was then used to study in detail the impacts of blending and filtration on 8 mum on the disinfection kinetic parameters. The results showed that no tailing in inactivation of spores was present for the samples from the high-quality source water, contrarily to samples from the other source water. High-speed blending decreased the aggregation responsible for the tailing and increased the inactivation rates significantly, with the exception of the wastewater. Filtration on 8 mum removed tailing on all samples and increased significantly the inactivation rate of spores, as did blending. Blending prior to irradiation allowed for greater inactivation when compared to blending after irradiation. Filtration increased inactivation rates even more so. The absence of tailing in the St-Lawrence River's water and the water filtered on 8 mum could be the result of low spore concentration which limits the study of inactivation to 1 to 2 log.; These findings strongly suggest that particles > 8 mum have a significant impact on the presence of tailing. Although particle distribution and turbidity are not related, no tailing was observed for samples with turbidity values under 10 NTU. Therefore, turbidity does not seem to have a significant impact on UV inactivation of spores when it is less than 10 NTU.; A linear model was applied using all the data. Results from the application of linear models to the combined data reveal significant differences (increases) between the fluence based rates associated with the raw, blended and filtered samples of the Mille-Iles River water but not from the St-Lawrence River, because of low aggregation.; The ob