Shahram Tabe, Taher Jamal, Rajesh Seth, Chaoyang Yue, Paul Yang, Xiaoming Zhao, and Linda Schweitzer

OBJECTIVES:

The first objective of this project was to investigate the seasonal loadings (occurrence) of pharmaceuticals and personal care products (PPCPs) and endocrine disrupting compounds (EDCs) discharged from a local sewage treatment plant (STP) into the Detroit River Watershed, at the intakes of three water treatment plants (WTPs) located at both sides of the river, and in the treated waters from those plants. The second objective was to study the efficiency of the ozonation process in removing PPCPs and EDCs from water and to determine optimized operating conditions for such removal. A major condition for meeting the objectives was the availability of valid and accurate analytical methods capable of detecting the target pollutants at extremely low concentration levels of a few ng/L.

BACKGROUND:

PPCPs are emerging contaminants in surface waters and, to a lesser extent, groundwater. This category of chemicals includes hydrophilic drugs that are not traditional chemicals of concern because they do not bioaccumulate and are not considered to be persistent in the environment. The concept of persistence, however, is shifting from those chemicals with long half-lives to those that are consistently being discharged into the environment and are thus ubiquitous.

A second category of emerging contaminants of concern include EDCs. Impacts of EDCs on ecosystems have been documented and are easier to substantiate than impacts on humans. Humans can rely upon resources far beyond the area in which they live, thus it is harder to control exposure variables in order to tease apart the factors that may be resulting in potential health effects. We do not know whether the problem can be attributed to our diet, lifestyle factors, or the chemicals that we are being exposed to.

The formulation of globally accepted treatment processes to remove target substances from water is necessary. Although several promising technologies exist for this purpose, they all have to be optimized according to the specific characteristics of the water as well as the type of pollutants found in the water.

APPROACH:

The project was completed in three phases. In Phase One, Occurrence Study, water samples were taken from different locations in Windsor, Ontario, and Detroit, Michigan, over a 13 month period and were analyzed for target substances. The sampling locations included effluent from the Little River STP, the Detroit River and its confluence with the Little River, and drinking water plants that draw from the Detroit River in Canada and the United States.

In Phase Two, Bench-Scale Experiments, the most frequently occurring chemicals detected in the Detroit River were chosen according to the data gathered from the first phase of the project for the bench-scale experiments. Bench-scale experiments utilized laboratory water spiked with the select PPCPs and EDCs to approximately 10 times their actual concentrations in the raw water. Two categories of target compounds were studied: ones that were expected to be highly refractory (Group A), and ones that were expected to be less refractory upon ozonation (Group B). The spiked water samples were treated by ozone under a variety of water quality and operating parameters pre-defined according to a five-factor fractional factorial design. From the bench-scale experiments, the optimum conditions for each set of water quality scenarios were determined using statistical analysis of the experimental results.

The outcomes of the bench-scale experimental stage were further tested and optimized in Phase Three, Pilot-Plant Experiments. The experiments were designed to simulate real water treatment operations including using the actual raw water from the Detroit River and the range of operating conditions that are usually encountered in a routine setup. The results from these experiments were compiled and used for the final recommendations on the optimized application of ozonation process for removal of PPCPs and EDCs from water.

RESULTS/CONCLUSIONS:

The outcomes of the research project can be summarized and concluded in the following list:

1. Of 51 target PPCPs and EDCs, 12 were not detected in any of the samples, 15 were detected in 1-10% of the total samples, and 24 were detected in at least 10% of the samples. Of the latter 24 substances, 13 appeared in 10-25% of the samples, 7 in 25-50% of the samples, 2 in 50-90% of the samples, and 2 in 90+% of the samples. The most occurring substances were perfluorooctanoic acid (PFOA) and perfluorooctyl sulfonate (PFOS) with occurrence frequencies of 94% and 97%, respectively.

2. Concentrations of a majority of the substances dropped two to three orders of magnitude in treated water samples compared to the effluent of the STP. Both natural removal and water treatment processes were responsible for this reduction. The existing water treatment processes were effective in removing a number of the target substances to levels below the instrument detection limits (IDLs).

3. Comparing the raw and treated water samples, of the 51 target substances, 22 were detected in at least 10% of the raw water samples, while 10 were detected in 10% or more of the treated water samples. The remaining compounds were either completely removed or were detected in lower frequencies.

4. The persistent substances included bisphenol A, erythromycin, ibuprofen, PFOA, PFOS, roxithromycin, and tylosin. Eight more substances were also detected in the treated water samples at similar concentrations as in the raw water but at very low frequencies, which made it difficult to evaluate their removal efficiencies. Other target chemicals were not detected in any of the treated water samples.

5. Sixteen of the most frequently occurring substances were chosen for the bench-scale experiments. A five-factor fractional design was used to study the individual and interactive effects of five parameters on the removal efficiency of the ozonation process applied to water samples containing a mixture of the target compounds. The parameters included ozone dosage, dissolved organic carbon (DOC) loading, contact time, temperature, and pH.

6. The target substances were divided into two groups according to their rates of reaction with ozone. The fast-reacting compounds in group A included bisphenol A, carbamazepine, erythromycin, gemfibrozil, indomethacin, lincomycin, naxproxen, sulfachloropyradizine, sulfamethazin, sulfamethoxazol, tetracycline, and tylosin. The slow-reacting compounds in Group B included bezafibrate, clofibric acid, ibuprofen, and monensin.

7. The results indicated that under a wide range of optimized water quality and operating conditions, close to complete transformation of the target substances is possible.

8. The statistical analysis of the results indicated that DOC loading and ozone dosage as well as their interaction had the greatest influence on removal efficiency. DOC loading had a negative effect, meaning that an increase in its level decreased the efficiency. Ozone dose had a positive effect. The secondary influential parameters were contact time and temperature. It was also observed that pH had a negligible effect on the efficiency within the experimental range.

9. The bench-scale findings were confirmed by a series of pilot-plant experiments. It was determined that group A substances were removed effectively even at low ozone exposure levels. Transformation of group B chemicals was enhanced at increased contact time. They also responded significantly to the ozone dosage.

10. Ibuprofen and clofibric acid were found to be the most difficult to transform. Their transformation efficiencies were limited to an average of 50% under the range of experimental conditions. Bezafibrate was responsive to ozone dose and showed a transformation efficiency increase from approximately 50% to 90+% when the ozone dose was increased from 0.3 mg/L to 1.5 mg/L at a contact time of 8.6 minutes. The experiments with monensin were not conclusive.

11. Presently, only limited information exists on the nature and toxicity of the products of the oxidation reactions of the target substances. Further investigation is necessary to understand the chemistry of the reactions and to determine the type of by-products of ozonation treatment of PPCPs and EDCs.

APPLICATIONS/RECOMMENDATIONS:

This project shows that ozone can be very effective at removing EDCs and PPCPs, with some exceptions. During ozonation, the micropollutants may react directly with molecular ozone (*OH) radicals formed as the ozone decomposes. Ozone is a selective oxidant that preferentially reacts rapidly with a limited number of compounds, characterized by a high value of the second-order reaction rate constant. The Group A compounds in the present study generally reacted quickly with ozone. Such compounds are expected to be significantly eliminated with ozone doses typically applied in drinking water treatment processes. Such compounds were efficiently transformed during both the bench-scale and the pilot-scale experiments. The transformation rate was correlated with ozone exposure (i.e., contact time [CT]) values, which are routinely monitored as a control parameter for the drinking water ozonation process. Transformation efficiencies exceeding 90% were observed for 12 group A compounds during the bench-scale experiments at ozone exposures exceeding 1.3 mg/L·min over a wide range of simulated raw water conditions, including a temperature range of 5°-23°C. These results were confirmed during pilot-scale experiments where transformation exceeding 87% was observed at ozone exposures exceeding 0.5 mg/L·min.

RESEARCH PARTNERS:

Windsor Utilities Commission

Ontario Ministry of the Environment