The efficacy of chlorine dioxide as a disinfectant, in case of bioterrorism attack, in water networks was evaluated using Bacillus globigii as a surrogate to Bacillus anthracis, the main cause of anthrax. The efficacy of chlorine, the traditional disinfectant, was tested earlier and was shown to be deficient as a disinfectant in such a scenario. In a previous study, free available chlorine (FAC) concentrations were assessed to concentrations up to 75 mg/L and it was reported that the residual spore attachment after disinfection was 3.8x103 CFU/cm2.
In this work, experiments in similar conditions to those carried out earlier by Szabo (2006) were conducted using chlorine dioxide as a disinfectant. The sectioning of biofilm depth into layers parallel to the substratum was also performed, prior to and during disinfection at concentrations of 15 and 25 mg/L chlorine dioxide, to study the attached spore stratification and change in viable spore fraction in each layer during disinfection. Preceding this, experiments were performed to assess the hypothesis of chlorine dioxide being a better disinfectant than chlorine and to establish a basis of comparison between chlorine and chlorine dioxide concentrations and attributes.
Aqueous chlorine dioxide films on glass, plastic and aluminum, with and without surfactant, showed that the rate at which chlorine dioxide leaves the liquid phase to the gaseous phase was found to correlate with the film thickness. Chlorine dioxide was able to partition with and break through the caprylic acid layer at a very fast rate. These results supported the hypothesis that chlorine dioxide will not be hindered by extracellular polymeric substances (EPS) when disinfecting a biofilm layer.
The bulk chlorine and chlorine dioxide experiments were performed with 1.0, 1.5 and 2.0 mg/L of both disinfectants in duplicates. Chlorine dioxide was able to achieve up to three log reductions of B. globigii in about two hours while chlorine reached 3 log reductions in 6 hours for 1.5 mg/L dosages. Data analysis using the Delayed Chick Watson model showed that both disinfectants experience a lag shoulder followed by a pseudo first order inactivation kinetics. The chlorine's lag shoulder, however, was one order of magnitude higher than that of chlorine dioxide. Watson plots showed that, for both disinfectants, exposure time and concentration play similar roles.
Biofilm disinfection experiments were performed in duplicate at concentrations of 5, 10, 15 and 25 mg/L chlorine dioxide. The reduction in viable spore count reached 4 log reductions in periods up to 8 days, with residual spore count between 50 and 350 CFU/cm2 of coupon surface. The micro slicing study was performed before disinfection and during the disinfection phase with the 15 and 25 mg/L chlorine dioxide experiments. Prior to disinfection, the spores were almost equally distributed throughout the biofilm thickness. The slicing technique showed the ability of chlorine dioxide to start the disinfection simultaneously in all layers. The study also showed that a persistent population of spores was contained within the fourth layer in the 25 mg/L profile and within the second and fourth layers in the 15 mg/L profile.
Chlorine dioxide performed as a better disinfectant in the scenario of a bioterrorism attack on drinking water networks. Better performance was shown in the bulk phase through a shorter lag period. It also showed the ability to act simultaneously on all biofilm layers and thus was less affected by rust and biofilm debris.