Optimisation of organic compound removal in artificial recharge systems by redox control and enhanced oxidation
Subsurface passage as utilized during river bank filtration and artificial groundwater recharge has shown to be an effective barrier for multiple substances present in surface waters during drinking water production. Additionally it is widely used as polishing step after wastewater treatment. However, there are limitations concerning the removal of DOC and specific trace organics. The project ”OXIRED“ aims at assessing possibilities to overcome these limitations by combining subsurface passage with pre-oxidation by ozone. In the first phase of the project, laboratory-scale column experiments were conducted in order to quantify removal for different settings under varying conditions. In a previous study different combinations of advanced oxidation and subsurface passage were evaluated concerning their potential removal efficiency and practical implementation on the basis of existing, published experiences and theoretical considerations. Two different scenarios were identified as promising for experiments in laboratory-scale columns with surface water and sewage treatment plant effluent: (A) surface water - oxidation - groundwater recharge and (B) surface water - short bankfiltration - oxidation - groundwater recharge. The investigations were designed to lead to recommendations for the implementation of a combined system of subsurface passage and advanced oxidation in pilot scale experiments that will be carried out in the second phase of the project. Prior to column experiments, batch tests following the RCT-concept by Elovitz and von Gunten (1999) were carried out to characterize the reaction of ozone with the investigated water qualities . Additional batch ozonation tests with subsequent analysis of biodegradable dissolved organic carbon (BDOC) were conducted in order to determine optimal ozone doses for DOC removal in column experiments. For laboratory-scale experiments a set of 8 soil columns (length: 1 m; diameter: 0.3 m) was operated at TUB to evaluate the effects of pre-ozonation of different source waters (secondary effluent, surface water, bank filtrate). Ozonation was conducted with gaseous ozone in a 13-L stirred tank reactor. Specific ozone doses of 0.7 mg O3/mg DOC0 and 0.9 mg O3/mg DOC0 were investigated. Trace organic compounds to be targeted were identified in a prior literature study on existing data on subsurface removal. Results from laboratory-scale soil column experiments led to recommend specific ozone doses (z) of 0.7 mg O3/mg DOC0 for the following technical- and pilot-scale applications. Removal of surface water DOC in the soil columns was increased from 22% without ozonation to 40% (z = 0.7) and 45% (z = 0.9) with preozonation and the DOC in the column effluent reached the level of tap water in Berlin within less than one week of retention time. At bank filtration and artificial recharge sites in Berlin similar removal rates were only observed within 3 - 6 months of retention . The transformation of many trace compounds was efficient with specific ozone doses of 0.6-0.7 mg O3/mg DOC0. Realistic surface water concentrations of carbamazepine,sulfamethoxazole, diclofenac and bentazone were reduced below the limits of quantification (LOQ). The pesticides diuron and linuron were reduced close to LOQ. The substances MTBE, ETBE and atrazine were only partly transformed during ozonation. For efficient transformation of these substances, higher ozone doses or an optimisation of the oxidation process, for example as advanced oxidation process (AOP), should be considered. Operating a preceding bank filtration (scenario B) will enhance the transformation efficiency of MTBE and ETBE. With similar ozone consumption the transformation of MTBE and ETBE was increased by 27-31% and 28-33% of the original removal, respectively. Other investigated compounds were efficiently transformed during ozonation of surface water independently of the preceding bank filtration step. For the removal of bulk organic carbon only little improvement was observed for scenario B. Overall DOC removal increased from 45% with direct ozonation of surface water to up to 50% with a preceding soil column. Despite the presence of relevant bromide concentrations (~ 100 µg/L) formation of the oxidation by-product bromate was not observed (< 5 µg/L). However, this could also be a result of analytical problems, as later spiking tests showed. Formation of brominated organic compounds was also not observed. Adsorbable organic bromide (AOBr) even decreased by 50 - 60% for secondary effluent and 80 - 90% for surface water. The reduction of AOBr concentrations was accompanied by an increase of inorganic bromide by up to 40 µg/L during ozonation of surface water. In the two conducted in vitro genotoxicity tests (Ames test, micronucleus assay) no genotoxicity caused by ozonation of water samples was observed. Testing for cytotoxicity (glucose consumption rate, ROS generation) showed positive results in several samples. However, a systematic attribution of toxic effects to ozonation or subsequent soil passage was not possible. Reasons for cytotoxic effects were not evaluated within the scope of this project but it is assumed that they were caused by unknown cofactors. These results show that the objectives of enhanced removal of trace organics and DOC by combining ozonation and subsurface passage are well met. Further investigations need to confirm this for the pilot scale, especially taking into account the formation, retention and toxicity of oxidation by-products.