Identifying technical synergy effects for organic micro-pollutants removal

Effluents of municipal wastewater treatment plants (WWTPs) represent the main pathway for organic micro-pollutants (OMP), micro-plastics and antibiotic resistant bacteria into waterbodies, such as surface waters. OMPs comprise synthetic persistent substances (Ternes 1998; Jekel & Reemtsma 2006; Brezonik & Arnold 2011), which remain in the environment and (bio)accumulate through the food chain, thus present a risk to human health and the ecosystems (Track & Kreysa 2003). OMPs are characterized by their critical substance properties: (high) polarity, (low) sorption tendency and (high) biological and chemical persistence, whereby these substances can even be detected in drinking water (Frimmel & Müller 2006). OMPs can be detected in low concentrations in the range of nanograms up to micrograms per liter. Currently used physical, biological and chemical processes in WWTPs are not designed for the removal of OMPs (Virkutyte et al. 2010). In recent years, various processes, which are established in drinking water treatment for OMP removal have been investigated for their potential application in WWTPs. High temporal variability and fluctuations of the wastewater composition lead to qualitative and quantitative dynamics that impact the wastewater treatment process. Thus, technologies from the drinking water sector have no or only limited direct transferability (Sontheimer et al. 1988; Virkutyte et al. 2010). Nevertheless, oxidative processes and/or adsorptive methods, such as powdered activated carbon (PAC) or granular activated carbon (GAC), haven been proven to be efficient methods for advanced wastewater treatment. The removal strongly depends on the properties of the OMPs and e.g. the presence of organic matter (Virkutyte et al. 2010). So far, the methods are applied individually, without the selection of any system and systematic approach regarding the selection of the appropriate process(combinations) is missing.

The main focus of this study is to provide an overview of the current available information on pilot-scale and full-scale advanced wastewater treatment plants (aWWTPs) equipped with oxidation, adsorption and filtration processes for OMP removal. This study aims to 1) select and prioritization OMPs according to state-of-the-art literature 2) develop a matrix of notation based on evaluation of 52 studies (see Table 3 in supplementary data) with the focus on GAC processes 3) define synergy effects of aWWTPs on conventional parameters. Furthermore, the present study comprises an evaluation of individual processes and their synergies for separation of particulate and dissolved compounds by a matrix notation of the processes. The selection of the OMPs depends on quality criteria and operates as basis for the selection of advanced wastewater treatment processes.

This study focuses on relevant OMPs which have been selected according to abundance and ecotoxicological risk. Table 1 summarizes predicted environmental concentrations in effluents of WWTPS (PECs) (see Database ‘Pharmaceuticals in the environment’ (aus der Beek et al. 2016)), predicted no-effect concentrations (PNECs) and PEC/PNEC ratios for selected pollutants. The ratio of PEC and PNEC is defined as Risk Quotient for waterbodies. For PEC/PNEC ratios ≥1, risk to the aquatic environment is suspected (EC 1996). Ratios between 0 and 4.69 for OMPs were chosen to represent a broad range of OMPs present in the environment. Furthermore, conventional water chemical parameters are listed, which show higher PEC/PNEC ratios (e.g. sP: 28.57) and are therefore, target parameters in WWTP. In case of missing legislative quantitative targets for OMPs in the European Union, these parameters help to evaluate the relevance of OMPs. Due to continuous improvements in the sensitivity of analytical methods, the complexity in terms of number and type of trace elements increases. When selecting and prioritizing relevant OMPs, not the number but rather their chemical-physical properties (e.g. non-polar/polar, degradable/persistent) and their environmental behavior (e.g. ecotoxicology, transport, interactions) are crucial to consider. Furthermore, the lists of substances reported by BAFU (2015), EU (2015), OGewV (2016) and Jekel et al. (2015) were also taken into account for the selection. The selected substances represent a broad PEC/PNEC spectrum and are representative for municipal wastewaters.

The input data of the matrix of processes was based on 52 studies in Germany, Switzerland, Austria, The Netherlands, France and Australia (see Table 3 in supplementary data). Furthermore, results of a pilot-scale aWWTPs (Hesse, Germany) were included to identify synergy effects for post-filtration. The selection of the aWWTPs was made according to the following criteria (Benstoem et al. 2016; Benstoem et al. 2017):

• pilot-scale and full-scale aWWTPs (Q_min = 0.5 m³/d), characterized by continuous stable operating conditions for extended periods of time

• real concentrations of OMPs in the μg/L or ng/L range in a municipal sewage water matrix measured with common analysis

• evaluation of OMP removal with regard to 24-h, 48-h, 72-h or 1-week mixed samples

• determination of OMP removal efficiency based on breakthrough curves

• consistent reference scale of normalized effluent concentration c/c₀ (effluent concentration relative to influent concentration), referred to the effluent of the WWTP and the influent to the further/advanced treatment process, respectively

For this purpose, specific operation parameters were collected, e.g. treated bed volumes (BV), carbon usage rate for PAC, and specific ozone consumption. For some studies data was derived from diagrams or data points of breakthrough curves. Relevant OMPs were selected according to Jekel et al. (2015), BAFU (2015) and EU (2015), as for example, pharmaceuticals (carbamazepine, diclofenac, clarithromycin, metoprolol, sulfamethoxazole), diagnostics (e.g. iopromide) and industrial chemicals (e.g. benzotriazole). The selected OMPs can serve as indicator substances, which can be reduced (low, medium and high) by adsorptive or/and oxidative processes (see Table 2 in supplementary data). Additionally, conventional chemical parameters (e.g. chemical oxygen demand (COD), total organic carbon (TOC) and dissolved organic carbon (DOC), total and soluble phosphorus, suspended solids) were also collected to identify synergetic effects in aWWTPs. In general, aWWTPs are operated to reach 80% reduction of OMPs based on the wastewater treatment inlet. Removal criteria of 80% reduction of OMPs for evaluating advanced processes are mostly defined as effluent concentration relative to influent concentration (c/c₀) = 0.2. This approach is applied in Switzerland through statutory requirements (BAFU 2015). For reduction of DOC c/c₀ = 0.8 is applied (Benstoem et al. 2017).

The data was analyzed by common statistical methods, i.e. calculating maximum, 75th percentile, median, mean, 25th percentile, and minimum values.

The removal efficiency of all processes applied for aWWTP depends on the chemical properties of OMPs, the performance of the WWTP and pre-treatment of the aWWTP. Due to the complex nature of the wastewaters – aWWTPs need to remove a wide variety of OMPs and other critical pollutants – a process combination of filtration and adsorption or ozonation is recommended. Therefore, GAC can serve as key technology in aWWTP. Figure 1 shows the collected GAC process input data for treated BV until the breakthrough criteria of c/c₀ = 0.2 for OMPs and c/c₀ = 0.8 for DOC are reached. The span widths for the breakthrough criteria are large, e.g. for diclofenac (including 13 studies) the criteria are met between 1,000 and 35,000 treated BV. This can be explained by different process parameters and water matrices used in the studies. Benstoem et al. (2017) considered influent DOC (DOC₀) and EBCT as the most relevant parameters. In addition, the TSS concentration in the influent of the collected GAC process was between 0 mg/L (effluent of membrane bioreactor or pre-treatment with microfiltration) and 30 mg/L (effluent of a conventional activated sludge process with secondary clarifier (CAS)), which might be correlated with OMP concentration.

Figure 2 illustrates the relative residuals of DCF over the course of treated BV. This shows the necessity of pre-treatment filtration step for adsorptive process as aWWTP for the first time, since the particulate matter disturbs them. DCF was chosen as target OMP due to low PNEC, high PEC/PNEC (see Table 1) and high removal by adsorptive and oxidative processes. The analysis of various GAC configurations with/without different types of filtration (e.g. sand filtration (SF), flocculation filtration (FF) and microfiltration (MF)) clearly demonstrates the improvement in DCF removal by addition of filtration. The OMP removal by GAC was shown in a lot of studies (see Table 3 in supplementary data). Nevertheless, the synergies for filtration were not established. The best performance was achieved for MF + GAC (Knopp et al. 2016a) because MF prevents blocking of the activated carbon by particulate and dissolved organic matter (Sontheimer et al. 1988) and therefore, running times for DCF were 3–5 times higher than for GAC only. According to Zietzschmann et al. (2014) higher concentrations of DOC cause significant reduction in the adsorption capacity of GAC. GAC filters which were operated directly in the effluent of CAS, reached the limiting criterion of c/c₀ = 0.2 or removal of 80% at <1,000 BV in comparison to 10,000–30,000 BV commonly found for MF + GAC. For the identification of these benefits, longer operation times >30,000 BV for GAC processes with filtration as a pre-treatment are necessary. Due to longer operation time, it is also possible to make statements about the backflushing intervals, which would be characterized as being required for pre-filtration. The question of economic efficiency remains unacknowledged and requires further research.

Observed breakthrough behavior can also be demonstrated for other selected OMPs and transferred to further OMPs in different wastewater matrices. The results demonstrate that by combining GAC and filtration pre-treatments, not only higher BV can be achieved, but also additional synergies with respect to conventional parameters are possible. The retention also has a positive effect on particulate nutrients (phosphorous), since these are additionally reduced. As a result, the effluent concentrations can be significantly reduced, and peaks of particulate matter can be avoided (overloaded secondary clarifier or poor settling characteristics of activated sludge), and the target values for TSS (<1–2 mg/L) can be accomplished. A reduction of 1 g TSS approximately decreases the particulate COD by 0.8–1.6 g and the particulate P by 0.02–0.05 g (Schröder 1998). Figure 3 displays the relative residuals by using pre-filtration processes (e.g. cloth media filtration (CF) and MF) without using coagulation agents to remove TOC, COD, TP and TSS. In recent years the elimination of TP became increasingly important due to the exceedance of thresholds in the water bodies, e.g. Germany (OGewV 2016). Required TP values of <0.2 mg/L can be achieved by using coagulation in combination with filtration (FF) as shown by other studies (Altmann et al. 2015). In this case, an increased use of coagulants (>2 mol Me³⁺/mol P) is expected. Furthermore, Figure 3 shows that the reduction of the particulate matter is almost equally (for considered parameters TOC, COD and TP) by means of both filtration processes. Here, it should be noted that CF consumes less than 10% energy compared to MF.

As already mentioned, particulate compounds and suspended solids significantly influence further processes for OMP removal, e.g. competing ozone-depleting substances (Huber et al. 2005), shorter lifetimes for backwashing of filters loading and/or biomass growth (Sontheimer et al. 1988). In addition, very low solid concentration (<1–2 mg TSS/l) lead to an increase in process stability.

Figure 4 shows a simplified matrix of processes for aWWTPs for the removal for various OMPs, which are indicated by the colorbar. The matrix proves that there are significant differences in the removal depending on the process combination and nature of the OMP. DCF can efficiently be removed by GAC, PAC and O₃, whereas IPM is only removed up to 40% by GAC (15,000 BV), but to >70% by PAC (15 mg_PAC/L). The overall removal efficiency for the OMPs type of contrast media (e.g. IMP) is (very) poor. Single-stage adsorption or oxidation processes can be classified as not efficient for elimination of X-ray contrast media (see Figure 4). Compared to OMPs with a PEC/PNEC ratio ≥1 (see Table 1), these can be removed up to 80% from wastewater. Very polar OMPs, such as IMP, can be eliminated up to 80% only by increased resource use or by the combination of methods (Knopp et al. (2016a). For example, at higher specific ozone consumption (>0.7 g O₃/g DOC), higher reductions of individual poorly removable substances (e.g., IPM) can be achieved. The removal of IPM may not be necessary on the basis of the risk assessment, however long term effects are still unclear.

The evaluation of the studies showed that there are no uniform standards for sampling of OMPs, e.g. studies report on 24-hour mixed samples up to 1-week mixed samples. The type of sampling has a considerable impact on the assessment of the methods (Ort et al. 2010). Furthermore, parameters such as DOC, TSS, TP are not always stated or recorded and therefore, evaluation of the process efficiency and comparability is limited. The (relevant) OMPs and purification targets determine the structural elements implemented in aWWTP. Basically, all processes are suitable to reduce the (selected) OMPs to a necessary degree (see Table 1). Considering that wastewater is a multicomponent mixture and contains a large number of constantly changing OMPs (only a small number of analytical data are collected) further tests are necessary to evaluate the processes. For a comprehensive evaluation, microbiological and ecotoxicological investigations are necessary in addition to chemical analyzes (Ternes et al. 2017). In this case, a low solids concentration (≤1 mg TSS/L) is advantageous.

The matrix serves as a decision-making tool for selecting the ideal solution for aWWTPs by prioritization of OMPs and identifying technical synergy effects. In order to ensure the proportionality of effort and benefit, also from an ecological point of view, the wastewater has to be treated in order to fulfill the requirements. In Switzerland 80% of the concentration of selected OMPs compared to raw wastewater has to be reduced from the effluent (BAFU 2015). The application of Risk Quotients can be used independently from the specific performance to review or select advanced treatment processes. The matrix contains only a section of OMPs (with different properties, see Table 2 in supplementary data), for a comprehensive consideration, further (locally specific/relevant) pollutants are to be considered. It guarantees the comparison of the procedures and gives an overview of the reduction of OMPs to achieve the good quality standards in the waterbodies.

The selection of the processes and process combinations is based on case-by-case decision. The construction and maintenance costs are mainly determined by the size of the WWTP (cf. Mertsch 2017) and the desired purification target. In this respect, the costs for the implementation of aWWTPs depend on whether the WWTP is already equipped with advanced process steps and the feasibility of potential for optimization exists. Does this not apply costs for a new implementation of aWWTPs has to be considered. In terms of operation costs, electrical energy demand for ozonation (depending on the ozone dosage) is 2 to 4 times higher than the energy demand of PAC or GAC filtration (Mousel et al. 2017). Energy requirements are plant-specific, and depend on existing plant components, geographical location with regard to pump energy and pre- and post-treatment steps. For a detailed economical assessment more large-scale implementations are necessary.

Available literature was systematically reviewed with the aim to consolidate studies dealing with the adsorption/oxidation of OMPs from municipal wastewater with GAC, PAC, O₃ and combinations of them (in pilot- and full-scale). The summary of the studies revealed that there are obvious differences between the advanced processes when combined with filtration. The evaluation of aWWTP for removal of OMP should include conventional parameters (e.g. TOC, TP, TSS) in addition to the Risk Quotient (PEC/PNEC). The matrix can be used as decision support for the choice of suitable procedures for individual requirements of effluent quality. GAC processes are suitable for removal of relevant OMPs (see Table 1) to achieve the target concentrations with respect to PEC/PNEC. Based on the presented evaluation the necessity of reduction of OMPs with a risk quotient ≥1 to an indispensable level (according to PNEC) can be declared. Furthermore, advanced processes for OMP removal result in additional synergies for the reduction of new target values for TP (<0.1 mg/L) and TOC (<7 mg/L) in receiving waterbodies.

The present work was supported by the German Federal Ministry of Education and Research (BMBF) as part of the funding measure ReWaM (NiddaMan, project number 02WRM1367). We also thank Eva-Maria Frei and staff (Abwasserverband Langen, Egelsbach, Erzhausen) for general and financial support at the pilot-plant.