Effet du chlore sur la colonisation bactérienne d'un réseau expérimental de distribution d'eau Effect of chlorine on the bacterial colonisation of a model distribution system

Résumé

La contamination bactérienne de la phase eau d'un réseau de distribution résulte d'une multiplication des bactéries sur les parois des canalisations d'eau (biofilms) suivie de leur arrachage et de leur transport dans le flux circulant. Ce travail met en évidence l'effet du chlore, d'une part, sur la formation des biofilms et, d'autre part, sur des biofilms déjà constitués. Des éprouvettes de matériaux neufs introduites dans des eaux présentant des concentrations en chlore total variant de 2,4 à 0,02 mg/l et véhiculant entre 0,5 x 10⁶ et 5 x 10⁵ cellules bactériennes/mi (dont 1 à 10 % de bactéries cultivables) sont rapidement colonisées (10⁶ à 10⁸ cellules/cm²). L'effet du chlore est sensible sur les cellules totales pour des concentrations de l'ordre de 1 à 2,4 mg/l. Sur les bactéries cultivables, un ralentissement de la croissance du biofilm est observé dès 0,3 mg/1 de chlore total. Par contre, des résiduels de 0,02 ou 0,05 mg/l sont sans effet sur la cinétique de formation des biofilms. Des résiduels moyens de chlore total compris entre 2,3 et 3,4 mg/l appliqués en continu pendant 14 jours sur un biofilm constitué d'environ 8,7 x 10⁶ cellules par cm² (1,7 % de bactéries cultivables), entraînent l'élimination d'environ 90 % des bactéries fixées (abattement d'1 logarithme) durant les premiers jours d'exposition. L'altération du biofilm exposé à un résiduel de chlore total de l'ordre de 1,3 mg/l est identique, mais toutefois plus étalée dans le temps. Ces essais réalisés sur des éprouvettes de PVC, PE et mortier de ciment n'ont pas permis la mise en évidence de comportements différents de ces 3 supports..

Mots-clés:

• Biofilm,
• désinfection,
• chlore,
• réseau de distribution d'eau

Abstract

Bacterial accumulation in drinking water systems results both of cell deposition on the pipe walls and attached bacteria growth. The presence of a complex biofilm (cells embedded in a matrix of exopolymers) leads to a continuous contamination of the water phase resulting from the erosion of the attached growing biomass. Then, many tentatives to lmit the formation of such a biofilm have been suggested as the removal of biodegradable organic matter fram water or as the application of disinfectant. However, the efficiency of chlorination of the distribution system is debatable. Indeed, adhesion is often described as a factor of protection of attached bacteria which counterbalances the expected effect of disinfectant. Then, the aim of this experimental work is using a model distribution system to evaluate (i) the kinetics of biofilm accumulation on coupons of new materials (Polyvinyl chlorure : PVC, polyethylene : PE, cement) disposed in a constantly chlorinated system (residual total chlorine from 0.021o 2.4 mg. l^-1), (ii) the effect of chlorination on previously accumulated biofilms.

The industrial pilot plant used in this study is comprised of five loops serially disposed (fig. 1). From previous study of simulation, one may assume that each loop works like a perfectly mixed reactor when the whole pilot plant is equivalent to an infinite tubular reactor with high axial dispersion coefficient. During the experiment, the pilot was continuously fed with finished drinking water front the surface water treatment plant of city of Nancy (i.e. natural finished water with its own chlorine demand, organic nutrients and heterotrophic bacteria).

Total number of cells (epifluorescence counts) and heterotrophic plate count bacteria (15 days of incubation at 20 °C) were enumerated both in the water and, after sonication, on the surface of the coupons of tested materials.

The first experimentations show that chlorine slows clown the kinetic of deposition of bacteria onto the pipe wall but never prohibits biofilm formation. When the drinking waters carried from 2.4 to 0.02 mg.1^-1 of chlorine and from 0.5 to 5 x 10⁵ ml^-1 bacterial cells, biofilm is observed after 24 hours of immersion of the coupons with at least 10¹ to 10⁶ bacteria/cm². Respectively, the deposition or/and growth rates of total cells are drastically affected only for chlorine residual as high as 1 to 2.4 mg. 1^-1. The number of heterotrophic plate count of the biofilm is affected with lower chlorine residual (around 0.3 mg.1^-1) but residual concentration as low as 0.05 mg.1^-1 are ineffective.

The tentatives carried out in the second experience on preformed biofilms (2 months old biofilms, 8.7 x 106 cells/cm²) show that the continuous application of 2.3 to 3.4 mg. 1^-1 of residual chlorine for 14 days, leads to the removal of only 90 % of attached total cells without modifications of the proportion of attached alive bacteria (around 1.7 %) into the biofitm. In other wards, a highly chlorinated networks shows at minima 10⁶ attached cells/cm². Its generally takes several days to reply to the chlorine demand of the system and to have a quasi steady state reactor in terms of residual chlorine.

These assays carried out with three types of coupons (PVC, PE, cement lined cast iron) did not show any difference between the tested materials.

The limited efficiency of chlorine against the biofilm can be explained by transfert limitations within the visquous layer, high consumption of chlorine by the biopolymers of the attached matrix (proteins...) or low sensitivity to the disinfectant of the slow growing attached bacteria. Then chlorination is really not a panacea in biofilm war but has to be applied in combination with other methods as biodegradable organic matter removal, hydraulic regime improvement...

Keywords:

• Biofilm,
• disinfection,
• chlorine,
• drinking water distribution system