Recently the use of low levels of medium- and low-pressure ultraviolet (UV) light for successful
inactivation of Cryptosporidium parvum oocysts has generated tremendous excitement in the
water industry. While these findings have been corroborated by numerous independent studies at
the bench-scale level, there is little experience in the U.S. with full scale disinfection of finished
water, using UV light. Concerns exist with respect to the reliability of UV technology and, in
addition to biodosimetry experiments to validate reactor performance for organism inactivation,
additional issues include determination of the performance efficiency of the reactors, effects of
suspended metals or other water characteristics on lamp sleeve fouling, effectiveness and
reliability of lamp cleaning mechanisms, UV measurement sensors' stability, costs associated
with retrofitting UV systems into existing water treatment plants, operation and maintenance
costs associated with employment of UV disinfection, impact of lamp aging on delivery of target
UV doses and impact of UV on disinfection byproduct (DBPs) formation or degeneration.
Although it is anticipated that experience with UV disinfection will provide information on some
of these issues, certain parameters need to be investigated before water utilities can commit to
using UV disinfection as one of the multiple barriers for protection of public health from
waterborne disease causing organisms. To address some of these operational issues, American
Water, has been actively involved in
ascertaining the long-term feasibility of applying UV for treatment of finished water. A 12 inch
diameter 4 x 1 KW, closed chamber UV reactor was installed after granular activated carbon
filtration at the Pennsylvania American Water treatment plant at Hayes Mine and was operated
continuously with a finished water flow rate of 600 gpm. Over a 12 month period, various
chemical (THM, HAA, UV254, DOC, TOC, metals, nitrate, nitrites) and physical measurements
(lamp voltage, current, sensor measurements) were monitored. Also parameters such as power
consumption and other operational issues associated with reactor function (i.e. frequency of
power outages, failing components, etc.) were recorded. Reactor validation studies were performed
using Bacillus subtilis spores, MS2 bacteriophage and Deinococcus radiodurans as a
biodosimetry surrogates and utilized various lamp configurations as well as lamp age. Includes 7 references, figures.