Formation and removal of ozonation by-products in drinking water biofilters

Abstract

The removal of organic ozonation by-products in biologically active filters was studied in this research. In particular, the effect of the presence of specific ozonation by-products on the removal rate of a given ozonation byproduct in biofilters, was studied. The ozonation by products examined were acetate, formate, pyruvate, oxalate, formaldehyde, and methylglyoxal.

A conceptual metabolic model was developed in this research, which assembled and simplified a number of relevant metabolic pathways. This conceptual model was used to demonstrate that the metabolism of acetate, pyruvate, and methylglyoxal were closely linked, and that the metabolism of formate, formaldehyde, and oxalate were similarly linked.

In batch experiments it was found that these ozonation by-products were readily biodegraded, as expected. In most instances biodegradation rates were independent of the other added ozonation by-products. An exception was the transient accumulation of measurable amounts of extracellular pyruvate during methylglyoxal utilization.

Further experiments were carried out in bench-scale biofilters, operated under conditions typical of drinking water filtration practice. Neither internal nor external mass transfer rates were found to limit the removal of the ozonation by-products studied in these biofilters.

Pyruvate was observed at intermediate filter depths in biofilters for which methylglyoxal was the sole feed compound. At steady-state the formed pyruvate was not observed in the filter effluents. However, both methylglyoxal and pyruvate were observed in the filter effluents upon increasing the influent methylglyoxal concentration by a factor of four. The time required to reestablish complete removals was found to be dependent on the hydraulic loading rate of the filter. Also, the removal of methylglyoxal was not found to be impaired following backwashing nor a 48 hour filter shut-down.

In a biofilter fed both pyruvate and methylglyoxal, pyruvate was found to have a lower removal rate than when it was the sole feed compound. This was postulated to have been caused by the formation of pyruvate during methylglyoxal biodegradation, as observed in previous experiments.

In a two-level factorial experiment it was found that neither the presence of two amino acids (serine and glycine) nor temperature (14 and 25°C) affected formaldehyde utilization rates. Serine and glycine were shown to be readily biodegradable in biofilters in this experiment, as expected.

A zero-order kinetic model, which took into account the biomass profile with filter depth, adequately described the concentrations of these components in the biofilters. Unexpectedly, kinetic parameter estimates calculated using the biofilter data were lower than the corresponding estimates based on the batch data.