Perchlorate (ClO4-) is a known endocrine disruptor that inhibits the normal function of the
thyroid gland and is a possible carcinogen. Due to these health concerns and the presence of
perchlorate in drinking water sources, the California Department of Health Services set the
action level for perchlorate in drinking water at 4 ug/l. Biological treatment using fixed bed
biofilm reactors, such as Biological Activated Carbon (BAC) filters, has proven to be an
effective and economical method for the removal of perchlorate. By adding an electron
donor, such as acetate, an existing granular activated carbon (GAC) bed can be retrofitted
to remove perchlorate to below 4 ug/l. Depending on raw water characteristics, it may even
be possible to achieve perchlorate removal to below this level using the natural organic
matter (NOM) present in the raw water as the electron donor (Lin et al., 2002). BAC filters
contain biofilms that are growing attached to the surfaces of GAC. These biofilms consist
of mixed communities, which contain mainly perchlorate reducing bacteria (PRB) and
other heterotrophic bacteria. PRB are ubiquitous (Coates et al., 1999) and remove
perchlorate using perchlorate reductase and chlorite dismutase enzymes as described by
Rikken et al (1996). When oxygen is present, most PRB use oxygen as their electron
acceptor. They switch to using perchlorate as their electron acceptor when oxygen levels
are very low. Therefore, control of the dissolved oxygen (DO) concentration is important for
the successful reduction of perchlorate.
Although the microbial aspects of biological perchlorate reduction have been relatively
well studied (Chaudhuri et al., 2002), no general design and operational criteria have been
established for BAC filters. The main objective of this paper is to provide a framework that
can be used to predict the patterns of BAC reactor behavior under varying influent DO
concentrations and loading rates. This information can then be used to develop better
design and operational guidelines for BAC filters.
Includes 9 references, figures.