Degree Name

Master of Science (MS)

Semester of Degree Completion


Thesis Director

William A. Weiler


The purpose of this investigation was to determine the effectiveness of conventional water-treatment methods in removing viruses from raw water sources. Those processes studied included alum flocculation, rapid sand filtration of fleeced and unflocced water samples, and chlorine disinfection.

This study was approached using coliphage f2 as a seeded indicator of treatment efficiency. Coliphage f2 was chosen as an indicator because of its similarities to enteric viral pathogens in both morphology and sensitivity to water-treatment processes. Coliphage f2 was found to occur naturally in surface waters, but not in numbers sufficient to interfere with calculations.

A sand filter for laboratory use was constructed with thin-walled PVC pipe, which was supported in a vertical position. Filter columns of 10-, 20-, and 30-inch depths were constructed either entirely of sand or as dual-media filters with a 1:2 ratio of sand to anthracite coal. Flow rate in the 30in column was adjusted to that reported for rapid sand filtration by adjusting the head of water above the filter bed.

Duplicate one-liter samples of chlorinated final effluent were seeded with 10ml of an f2 stock (ca. 106pfu/ml). To one sample was added 0.12g alum and the mixture was stirred and allowed to settle. Phage titer of the supernatant was determined by the agar overlay method.

The samples were then run through duplicate filters and the titers of the filtrates were determined. Resulting titers were compared to the original titer levels, and the efficiency of the processes were calculated as percent recovery. Turbidity was also measured on both samples before and after filtration.

Alum flocculation followed by settling resulted in a wide range of reduction of the seeded f2. In most cases at least a one-log (90%) reduction was obtained and in several cases at least a two-log (99%) reduction occurred. The variability in phage reduction was attributed to the quality and quantity of floc produced.

Filtration of unflocced, seeded water through sand and dual-media filters of various lengths resulted in generally low f2 removal. In only a few instances did such treatment result in a one-log (90%) reduction and in most cases the removal by filtration was negligible.

Flocculation without settling, followed by filtration, removed significant numbers of the seeded f2. In several instances a three-log (99.9%) reduction occurred and generally a two-log (99%) reduction was observed.

No correlation of f2 removal vs. column depth was observed. In addition, neither matrix was shown to remove f2 better than the other.

The removal of turbidity by filtration roughly paralleled f2 removal. However, the large variations in turbidity and phage removal levels precludes the use of turbidity reductions as a satisfactory criterion for judging the efficiency of the process in removing viruses.

Chlorine disinfection at 2ppm total residual chlorine resulted in an 80% reduction in f2 after a 2hr contact time; most of this reduction (68-79%) occurred within the first two minutes of chlorine contact, and corresponded with a rapid decrease in the free residual chlorine level. Further disinfection occurred throughout the 2hr contact time, but at a markedly slower rate.

It was concluded that, connected in series in a water-treatment system, flocculation, filtration, and chlorination could be expected to reduce indigenous viral populations by at least 99.8-99.98%.