Rapid Detection of Bacteria in a Water Distribution System
Abstract
A water distribution system may be contaminated by bacteria through cross-connections, by intrusion of soil water after a pressure loss due to a power blackout, or by the intentional contamination of the system in a bio-terrrorism event. The problem a utility faces is to detect it in a timely fashion, and to take immediate action to correct the problem and at least warn consumers.
Currently approved microbiological methods are all culture based. That is, after filtration the water sample is placed on an agar plate and then incubated at varying temperatures for one to seven days. A determination of the indicator bacteria E. coli takes 24 hours, and the most sensitive test for total bacteria is an HPC count of the colonies growing on R2A agar for three to seven days. These tests are clearly not useful for protection of the consumers in a timely fashion.
We have worked on a rapid test method whereby we capture the bacteria from a water sample on a 0.45 micron filter and then lyse the bacteria. The released adenosine triphosphate (ATP) from the bacteria reacts with added luciferin/luciferase and then light emission takes place that can be measured with a luminometer. The whole procedure takes less than five minutes. The amount of ATP is proportional to the number of bacteria.
To test if a sewage intrusion can be detected, experiments have been conducted in beakers where a small quantity of sewage (secondary effluent of the Ann Arbor Wastewater Treatment Plant (AAWWTP) was added to water samples, and the bacterial counts were determined by ATP analysis and the traditional culture methods. Three types of water were used: (i) drinking water that is chloraminated, (ii) water that is chlorinated, and (iii) water that carries a chlorine dioxide residual. As expected, chlorine dioxide will eliminate the bacteria very fast, followed by the chlorinated and chloraminated waters. In all cases the contamination of water can be clearly shown through the ATP analysis in less than five minutes. The test can be automated so that a nearly continuous monitoring is possible.
This paper is only available in PDF Format:
View full text PDF