The experimental design approach to removal of endocrine disrupting compounds from domestic wastewater by electrooxidation process

In this study, the treatment performance of the process in the removal of Endocrine Disrupting Compounds (EDCs) from domestic wastewater by a laboratory-scale electrooxidation process using Ti/IrO2/RuO2 electrodes as an anode was evaluated using the response surface method (RSM). The effect of pH (3.00–9.00), current density (10–20 A), and flow rate (8–16 mL/min) on the electrochemical removal of 17α-ethinylestradiol, β-estradiol, triclosan, and estrone has been studied. The Box-Behnken Design (BBD) was used to optimize the parameters that affect the removal efficiencies of the Electrooxidation process (EOP), and second-order quadratic models were developed for the EOP process. Under optimum conditions of current density = 10 A, pH = 3.00 and flow rate = 13.93 mL/min, the maximum removal efficiencies of triclosan, 17α-ethinylestradiol and β-estradiol, and minimum energy consumption are 91.65, 96.43 and 96.65% and 41.606 kWh/m3, respectively. The Response Surface Method predicted values that reasonably agreed with the experimental values. At the same time, the electrooxidation method is not successful in completely removing estrone from wastewater.


Introduction
One of the most significant environmental problems of this century is how to prevent contamination of water, which is essential for all living things and many industries, and how to guarantee its quality and safety. In recent years, there has been an upsurge in demand for water, and as a result, it has become abundantly evident that water scarcity poses a genuine risk to the long-term sustainability of societies [1,2]. The treatment and reuse of wastewater have evolved into a need rather than a luxury in recent years, shifting the focus away from the search for new supplies of clean water to fulfil growing demand [3]. Recently, there has been a growing global concern about "endocrine disrupting compounds" (EDCs), which can occur naturally or be synthetic, and their potential harm to the environment and human health. These detection methods that can measure EDCs in water and wastewater at very low concentrations, it is also exceedingly challenging to get EDCs removal from wastewater [11,12]. Since conventional wastewater treatment plants (WWTP) are intended to remove various pollutants from wastewater, including nutrients, dissolved organics, pathogens, and colloidal and suspended particles, it is not possible to completely remove EDCs using these systems due to the complex chemical structure of EDCs [13][14][15][16]. As a result, these compounds have been detected that they exit the WWTP systems without being treated, and they are present in extremely low amounts even in the treated effluent [17][18][19][20][21][22][23]. Different treatment methods, including membrane bioreactors, chemical precipitation, advanced oxidation processes, membrane filtration, and adsorption, have been developed considering the varying concentrations of existing EDCs in aquatic environments and the water quality standards for the reuse of treated water for different purposes [24][25][26][27]. The suitability of treatment processes for the reuse of water depends on the intended use of the treated water, and other factors such as the treatment cost, energy and land requirements [28,29]. Adopting proper treatment procedures that can remove EDCs and other pertinent compounds from water and wastewater is vital to protect the environment.
Researchers have recently paid attention to the electrooxidation process (EOP) because of its numerous advantages over conventional treatment methods, such as ease of use, safety, versatility, short treatment time, cost-effectiveness, environment friendly, adaptability and automation, and high removal efficiency of toxic and resistant organic substances [30][31][32][33]. The EOP process is predicated on the direct (electrooxidation or electro-reduction on the electrode surface) or indirect (through chemical oxidants generated in treated water) oxidation of organic substances using graphite, stainless steel, dimensionally stable electrodes (Ti/ PtO 2 -IrO 2 , Ti/RuO 2 -TiO 2 , Ti/RuO 2 -IrO 2 , Ti/IrO 2 /RuO 2 , and Ti/IrO 2 -Ta 2 O 5 ) and an insoluble anode material such as platinum and boron-doped diamond (BDD) [33][34][35][36]. After adsorption to the anode surface of the compounds, direct anodic oxidation (through direct electron transfer to the anode) comprises electron exchange between the compounds and the anode surface without the presence of other substances [37,38]. Theoretically, such oxidation is conceivable at more negative potentials than are required for the hydrolysis and the creation of oxygen. The EOP process is based on the on-site creation of highly reactive hydroxyl radicals ( • OH), which may even remove very persistent organic molecules that react with most organic compounds unselectively (1, 2 reactions).
(1) H 2 O → * OH + H + + e − This radical is the second strongest oxidant after fluorine [39,40]. It possesses a high standard redox potential of E ( • OH/H 2 O) = 2.80 V/SHE [41,42] and a 10 6 -10 10 M −1 s −1 rate constant for oxidation of the pollutants [43][44][45]. Additionally, this radical may react non-selectively with most organics until complete mineralisation occurs via hydroxylation or dehydrogenation [46,47]. It can be considered as a heterogeneous-like indirect oxidation process due to the short average lifetime of • OH radicals and their direct contribution to the anodic oxidation process being limited by the closeness of the electrode surface [48,49]. However, owing to the creation of polymeric layers on the surface of the electrode, this process often results in contamination and passivation of the electrode, and as a direct consequence, very weak chemical treatment is achieved [50,51]. These technologies hold great promise for the treatment of bio-refractory chemicals in water because of the nonselective nature of the • OH radical, which helps to avoid the creation of undesirable by-products [52,53]. Due to strong interactions (reactions 3 and 5) between the electrochemically generated • OH radical and the anode surface (M), the organic matter is directly oxidised via the oxidation mechanism through the following reactions (reactions 4 and 6). The direct oxidation rate of organic EDCs is determined by the anode's catalytic activity, which is aided by the diffusion rate of organic compounds to the anode's active points and the current density applied [54]. There are two major ways to decrease wastewater pollution using EOP given the existence of several heterogeneous species in water discharge [55]: i. Electrochemical transformation (reaction 4) wherein refractory organics (R) are selectively transformed into biodegradable compounds under the effect of chemically adsorbed "active oxygen" (MO) (reaction 3).
When treating a toxic and non-biodegradable endocrinedisrupting compound, electrooxidative reduction transforms the organic substrate into numerous biodegradable metabolites [56], and biological treatment is required for complete mineralisation following electrooxidation process [57]. On the other hand, if electrooxidative degradation yields H 2 O and CO 2 , no further procedure is needed. A variety of organic substrates can respond to anodic mineralisation in complex ways. In these situations, it is vital to consider the critical role played by physically adsorbed • OH radicals and the adsorption of organic species. Understanding the function of the oxidation processes that were indirectly generated during the treatment is one of the most critical aspects of explaining the high efficiencies attained by EOPs in the removal of EDCs. Indirect oxidation is the oxidation of EDCs via a chemical reaction with oxidants such as • OH, • CI, SO 4 ·− , O 3 , H 2 O 2, and peroxosulphates formed in wastewater based on the composition of the electrolyte and electrode materials [58,59].
Optimisation of a process seeks to change the functioning of a system in the most efficient manner possible using mathematical methods. Optimisation aims to get the best outcomes possible under the conditions. Certain systemrelated decisions need to be made for the conditions that need to be optimised. Such decisions aim to maximise the intended benefit or minimise the required time. Following system optimisation on laboratory scales, real pilot scale applications are applied. Every consumable material, process, and period employed in the laboratory setting is crucial for this reason. Using optimisation, the intended responses are maximised while using fewer chemicals, processes, and timeframes. Electricity is a component that when compared to other methods, offers an economic advantage in electrochemical processes. However, this factor becomes a disadvantage if the process takes a long time. To reduce the amount of electrical current used, it is essential to optimise the process parameters. Nowadays, several other mathematical and statistical programs for optimisation have been created. The response surface method (RSM), an efficient technique for optimising process conditions, is a commonly adopted statistical method for developing an empirical model [60]. To model complex systems, evaluate the concurrent impacts of numerous components, and identify the ideal circumstances for the intended responses, RSM is a potent statistical-based approach [61][62][63][64]. The ability to work with several parameters simultaneously has been optimised in this one-stage procedure, which results in time savings. It may also be used because it shows the potential interactions between different variable factors, which is an advantage of the method.
Dimensionally stable electrodes are often used to treat both domestic and industrial wastewater, but there is not enough research on effectively removing specific chemicals, like 17α-ethinylestradiol, β-estradiol, triclosan, and estrone using the EOP method in continuous mode with the dimensionally stable. This study aims to fill this gap by examining the treatment of these specific pollutants in wastewater by the EOP method in continuous mode using dimensionally stable Ti/IrO 2 /RuO 2 anode and the optimization of the parameters affecting the performance of the EOP system with the Response Surface Method.

Wastewater characterisation, reagents, and analyses
In this study, EDCs targeted for treatment were chosen based on their frequency in the sewage system and their specific physicochemical features. Analytical grade EDCs with greater than 98% purity were purchased from Sigma-Aldrich (St. Louis, MO, USA). All EDCs stock solutions were prepared using 250 mL of methanol (purity ≥ 99.9% CH 3 OH) purchased from Merck (Germany). 40 L of domestic wastewater was spiked to provide the desired concentrations. During the experiments, approximately 2.0 L of wastewater was used for each step of optimization of the parameters affecting the electrooxidation system. The wastewater's pH was adjusted by adding diluted aqueous solutions of HCl (0.1-1 N) or NaOH (0.1-1 N). All samples were stored in the refrigerator at + 4 °C. As a result of the methanol used to dissolve the EDCs, the initial COD of the wastewater increased to 1080.36 ± 45 mg/L. EDCs concentrations and wastewater properties added to wastewater are given in Table 1.

Electrolytic reactor installation
EOP runs were carried out in an electrolytic cell with a heating jacket made of glass [11 cm (radius) × 20 cm (height)]. 5 different Ti/IrO 2 /RuO 2 mesh plates were used as the anode, and 5 Ti mesh plates were used as the cathode. To avoid ohmic losses, the anode and cathode electrodes are placed parallel inside the reactor with a gap of 0.5 cm. The total surface area of 10 rectangular mesh plates is about 3300 cm 2 . Electricity power is supplied by a direct current power source (Chorome 62024P- . No additional external support electrodes were added for electrical conductivity, and the temperature was kept constant during all experiments at (25 ± 1 °C). After each run, the electrodes were cleaned with acetone and distilled water, and damaged electrodes were periodically replaced to maintain the same surface area throughout the study. The effects of current density (10-15 and 20 A), pH (3.0-6.0 and 9.0), and flow rate (8-12 and 16 mL/min) were examined. The EOP unit is shown in Fig. 1.

Analytical techniques
During the experiments, samples were collected after each procedure and analysed for COD, pH, and EDCs. For COD analysis, wastewater and effluent samples were measured without filtration. COD analysis was performed following the reflux colorimetric method. Electrical conductivity (EC), temperature, and pH measurements were performed using a multimeter (WTW mult340i multimeter). In this study, EDCs were simultaneously preconcentrated using the sensitive and precise dispersive liquid-liquid microextraction (DLLME) method developed by Kocoglu et al. (2019), and quantitative measures of EDCs were performed with the GC-MS system [65]. COD and EDCs removal efficiencies were determined by Eq. (7). The Eq. (8) was used to calculate the energy consumption.
(Energy consumed/m 3 of wastewater treatment) where C o and C e represent to the influent and effluent concentrations of EDC and COD.
where I = applied current density (A); V = potential difference in the EOP system (volt); t = reaction time (hour); ν = total wastewater volume (m 3 ).

Statistical analysis and modelling using RSM
Mathematical and statistical tools have been used successfully to explain and confirm the outcomes of many studies for a very long time. The primary objective of these methods is to identify driving factors and describe their quantitative relationship, as well as the link between factors and results. The Box-Behnken Design (BBD) and the Central Composite Design (CCD) are two statistical experiment designs used to study the relationship between variables and a response variable. The choice of design depends on the research question and variables under study. BBD is preferred in some cases because it requires fewer design points, is easier to set up and analyse, and is suitable for fitting quadratic models. CCD requires more design points to achieve precision, but it includes a wider design space and is more flexible, allowing for the exploration of nonlinear relationships between variables and responses [66,67]. This study investigated the effects of three independent variables on response functions (17α-ethinylestradiol and β-estradiol, triclosan removal efficiencies, and energy consumption. The determination of optimum conditions among the variables of pH (X 1 ), current density (X 2 ), and flow rate (X 3 ) that maximises the per cent removal of EDCs while simultaneously minimising time and energy consumption was accomplished by Box-Behnken Design (BBD) methodology. The study was completed with 15 experimental sets. Table 2 presents the experimental range and values of independent variables utilised in EOP system trials. Equation 9 illustrates the extended quadratic polynomial model used to assess the connection between independent variables (factors) and the dependent variable (response variable (Y)).
where Y (% EDCs removal efficiency and energy consumption refers to the estimation response, β 0 is the constant term, β i is the linear regression coefficient, β ii is the quadratic regression coefficient, and β ij refers to inter-factor interaction regression coefficients. At the same time, Ɛ is associated with experiments or random errors. X i and X j are coded independent variables that affect the response. Analysis of variance (ANOVA) was used to verify the statistical significance of second-order polynomial models and evaluate the system's removal performance.

COD removal results of the EOP system
The EOP reactor operating in a continuous mode was sampled at various times, and COD analyses were performed. In the experiments conducted at varied pH levels and a constant flow rate (12 mL/min), it is evident from Figs. 2, 3, 4 that there was little variation in COD removal during the first 10 min of the reaction at current intensities of 10-15 and 20 A. According to Fig. 4, there was a rise in COD removal after 15 min, and depending on the pH, the removal efficiency increased from roughly 66.41% to 85 ± 5% at a 20 A current density. It was found that COD removal has also been present in acidic pH, contrary to basic pH. In all experimental  studies that looked into the impact of pH change at 20 A current density and 12 mL/min flow rate, it was evident from Fig. 5 that at the end of the first 15 min of the electrooxidation process, the pH was between 2.8 and 4.0. Adding NaOH to the medium to raise the initial pH boosted the formation of • OH radicals and maintained acidic conditions at the conclusion of the reaction period. For all initial pH values, acidic conditions began to develop after the first 15 min of the reaction and kept on until the completion of the reaction period. The system's capacity to remove COD is observed to be more efficient at low pH than at its initial pH level. Removal effectiveness rises after 15 min and reaches its optimal level after 120 min.

Statistical analysis and interpretation with BBD
In the present study, the effect of operational parameters including current density (10-20 A), pH (3)(4)(5)(6)(7)(8)(9), and flow rate (8-16 mL/min), was explored in the treatment of EDCs using the EOP system running in continuous mode. As shown in Table 2, RSM was used through three levels of three factors to examine and optimize the effects of operating parameters on EDC removal and energy consumption. As indicated in Table 3, EOP system data were put into two second-order (quadratic) models, and the models' adequacy and significance were assessed using ANOVA. There is only a 0.01% for β-estradiol, triclosan and energy consumption and 0.05% for the 17α-ethinylestradiol chance that an F-value this large could occur due to noise. In case P value is < 0.0500; X 1 , X 2 , X 3 and X 1 2 are important model terms for triclosan while X 1 , X 2 , X 3 , X 1 X 3 , X 1 2 and X 2 2 are for 17α-ethinylestradiol, X 1 , X 2 , X 3 , X 1 X 2 , X 1 2 and X 3 2 are significant model terms for β-estradiol and lastly; X 1 , X 2 , X 3 , X 1 X 3 , X 2 X 3 , X 2 2 and X 3 2 are vital model terms for energy consumption. If there are many insignificant model terms (not counting those required to support hierarchy), model reduction may improve your model. The high squared correlation coefficient (R 2 ) and the adjusted coefficient of determination (R 2 Adj ) (triclosan: R 2 = 0.998-R 2 Adj = 0.996, β-estradiol: R 2 = 0.998-R 2 Adj = 0.994, 17α-ethinylestradiol: R 2 = 0.984-R 2 Adj = 0.956 and energy consumption = R 2 = 0.999-R 2 Adj = 0.998) indicate that the model can explain the total variance and the study parameters such as pH, current density, and flow rate are significant for the removal of triclosan, 17α-ethinylestradiol, β-estradiol, and energy consumption. In addition, the high significance of the models, the high accuracy, and the reliability of the experiments were validated by the low coefficient of variation (< %2) (triclosan = 0.8090, β-estradiol = 0.5202, 17α-ethinylestradiol = 0.8379, and energy consumption = 1.83). *Lack of Fit F-values* (triclosan = 5.79, 17α-ethinylestradiol = 1.86, β-estradiol = 2.51, and energy consumption = 2.02) mean that lack of fit is not significant compared to the pure error. The probability of such a great lack of fit F-values due to noise were 15.07, 36.82, 29.80, and 34.84% for triclosan, 17α-ethinylestradiol, β-estradiol, and energy consumption, respectively. *Adeq Precision* measures the signal-to-noise ratio. A ratio greater than 4 is desirable. Signalto-noise ratios for triclosan, 17α-ethinylestradiol, β-estradiol, and energy consumption were 70.060, 21.860, 56.952, and 122.424, respectively, indicating an adequate signal. These models may be utilised to navigate the design area. The model's estimated values for the removal of triclosan, 17-ethynylestradiol, and estradiol fitted to the experimental data. The equations show that it is feasible to discuss both an antagonistic and a synergistic impact of the variables coded as X 1 , X 2 , and X 3 in the removal of triclosan, 17α-ethinylestradiol and β-estradiol. Positive coefficients between factors and responses show the synergistic impact, and negative coefficients indicate the antagonistic effect. The fact that the variable significantly impacts the results may be used to explain why numerical expressions were significant. Equations provide response functions expressed in terms of coded factors.
Triclosan Removal (% ) (Coded)  RSM Models established for the removal of EDCs are suited for EPO procedures because the estimated values vs the experimental values in the removal approach in a straight line with high correlation, as shown in Fig. 6. Figure 7 Externally studentised residual test (also known as an outlier-t-test) examines whether one run is consistent with previous runs, provided the specified model is valid. Model coefficients were calculated based on all design points except one. At this point, an estimate of the response is determined. The residual is utilised via the t-test. A value larger than the estimated limits (often between 3 and 4) indicates that this point should be investigated as a potential outlier. Most data points fall on the line that provides a sufficient normality assumption with a valid model Figures 8, 9, 10 show 3D response surfaces relating to the interaction between independent variables. In these figures, 3D response surface plots illustrate independent variables' main and interaction impacts on response.
The figures show that flow rate, pH, and current density all have an impact on the removal efficiencies in EOP systems. As seen in Fig. 8a, the maximum triclosan removal efficiency at pH 3 and 20 A was more than 99.9%. The removal efficiency rose from 63 to 92% when the pH was lowered from 9 to 3 at a lower current density (10 A). In general, the pH of the electrolyte impacts indirect oxidation but does not affect direct oxidation [68]. The formation of oxygen in alkaline conditions is relatively high (0.401 V for alkaline pH, 1.229 V for acidic pH) [69]. And this may prevent the polluting molecule from diffusing to the electrode surface [70,71]. Although the pH effect is a complicated problem involving the ionisation state of each material and its potential interactions with the anode's surface, it is wellacknowledged that acidic environments promote the production of physically adsorbed · OH radicals on the anode surface [72,73]. Acidic pH may reduce the concentration of carbonate (CO 3 2− ) and bicarbonate (HCO 3 − ) ions, which are known as effective · OH radical scavengers [74,75], and it has also been determined that an inactive hydroperoxyl anion (HO 2 − ), which acts as an · OH scavenger per reaction, was formed in alkaline pH [68]. Consequently, the rate of  the oxidation process will likely be faster at lower pH levels. However, it was shown that flow rate had a lesser role in triclosan removal across all pH levels and current densities ( Fig. 8b-c). Increasing the flow rate from 8 mL/min to 16 mL/min at 20 A current density at pH 6 reduced the removal efficiency from 91 to 89%. Another set of studies found that increasing the flow rate from 8 mL/min to 16 mL/ min at a current density of 15 A at pH 9 decreased removal efficiency from 67 to 65%. The change of 17α-ethinylestradiol removal based on pH, current density, and the flow rate was given in Fig. 9d-f. Removal effectiveness of over 99% was achieved in investigations with a current density of 20 A in both the experimental study with pH 3 and a flow rate of 12 mL/min and the study with pH 6 and a flow rate of 8 mL/min. The impact of flow rate on 17-ethinylestradiol removal at acidic pH is relatively minor, as shown in Fig. 9d, but it is a vital marker under basic conditions. The removal efficiency of 17α-ethinylestradiol was found to be 96% at both flow rates when the flow rate was raised from 8 mL/min to 16 mL/ min at pH 3 and 15 A current density. However, as seen in Fig. 9e-f, the removal efficiency decreased from 92 to 85% when the pH was 9, and the current density was 15A when it was increased from 8 mL/min to 16 mL/min. The flow rate is one of the parameters of an EOP system influencing the mass-transfer coefficient [75,76]. Enhancing the masstransfer coefficient may help accelerate the transfer of EDCs from the solution to the anode surface, which will improve the reaction rate of direct and · OH radical-mediated oxidation. In the mass transport-controlled oxidation process, the flow rate of the electrolyte solution is crucial [77]. A low flow rate during electrolysis may result in a rise in the temperature of the solution, particularly when high current intensities are used, while a low flow rate of the solution in the reactor during electrolysis can prevent gas formation [78]. Consequently, determining the optimum flow rate is crucial in terms of both removal efficiency and cost. The effect of pH, flow rate, and current density on β-estradiol removal efficiency is given in Figs. 10g-i. At low pH levels, the change in current density at the same flow rate had no impact on the removal efficiency (Fig. 10g). When the pH was 3 and flow rate was 12 mL/min, and the current density was increased from 10 to 20 A, the removal efficiency increased from 96 to 98%. Nonetheless, the variation in current density at neutral and basic pH levels resulted in a substantial change in removal efficiency. In an experiment with pH 6 and a flow rate of 16 mL/min, where the intensity of the current was raised from 10 to 20 A, the removal efficiency increased from 88 to 93%. In a different study, the removal efficiency rose from 76 to 86% when the current density was raised from 10 to 20 A under pH 9 and a flow rate of 12 mL/min ( Fig. 10h and i). The oxidation of EDCs at basic and neutral pH values is positively impacted by current density. This is because active oxygen species ( · OH, H 2 O 2 and O 3 ) and other active oxidants (HOCI and OCI − ) form simultaneously depending on the electrolyte [79][80][81]. Nevertheless, the removal efficiency may not always benefit from consistently raising the current density. When the current density is higher than the limiting current density, it has been shown that oxygen formation sometimes outpaces the formation of active oxidants and even slows it down [82]. Generally, EDCs can be completely mineralised in EOP systems, but studies have shown that the time required for complete mineralisation is directly proportional to the increase in current density [83]. Table 4 shows the studies on electrochemical oxidation of EDCs and a summary of the most prominent findings of these studies.
While optimising operating parameters, the aim is to achieve the highest possible removal efficiency levels for triclosan, 17α-ethinylestradiol, and β-estradiol while minimizing energy consumption within defined constraints. Various constraints for the EOP process were applied, as specified in Table 5. In the study was determined that Fig. 8 Interaction effect of pH, current density, and flow rate on triclosan removal in the electrooxidation process: a pH/Current Density, b pH/ Flow Rate, c Current Density/Flow Rate under optimal conditions, the removal rates of triclosan, 17α-ethinylestradiol, and β-estradiol were 91.65, 96.43, and 96.66%, respectively. In addition, the energy consumption was found to be 41.606 kWh/m 3 . The overall desirability was calculated to be 0.862. The averages of the results of these experiments were found to closely match the values that were predicted by the RSM optimization, indicating a high level of agreement between the predicted and experimental outcomes (triclosan = 89.83%, 17α-ethinylestradiol = 93.23%, β-estradiol = %94.86, and energy consumption = 42.021kWh/m 3 ) ( Table 6).
Estrone could not be removed using the EOP system. In experimental studies, higher estrone concentrations were identified in treated effluent than in influent. Various processes, including photolysis and biological and chemical degradation, may result in the complete mineralisation of EDCs and, in certain instances, their transformation into intermediate conjugates. While β-estriol and 17α-ethinylestradiol are eliminated as conjugates, several other metabolic transformations occur [92,95]. It can be explained by converting β-estriol and 17α-ethinylestradiol to estrone during oxidation [70]. The concentrations of β-estriol and 17α-ethinylestradiol decreased over time, while the concentration of estrone increased in the experiments. While the influent concentration of estrone ranged between 188.9 and 211.1 µg/L, the output concentration was always higher than the concentration after all experiments. Figures 11, 12, 13 represents the pH-dependent influent and effluent values of triclosan, 17α-ethinylestradiol, β-estradiol, and estrone, which are removed by the EOP system. According to the Figs. 11, 12 and 13, the EO process has been effective in eliminating triclosan, 17α-ethinylestradiol, and β-estradiol, depending on the process variables; however, estrone removal has not been entirely accomplished.

Energy consumption
Operational energy consumption analysis is crucial when considering the application of the electrooxidation process at Fig. 9 Interaction effect of pH, current density and flow rate on 17α-Ethinylestradiol removal in the electrooxidation process: a pH/Current Density, b Flow Rate/pH, c Current Density/Flow Rate a real-scale because it helps to determine the energy requirements and associated costs of the process. The electrooxidation process is an energy-intensive process that requires a significant amount of electrical energy to operate. By analyzing the energy consumption, it is possible to estimate the total energy requirements of the electrooxidation process and evaluate the energy efficiency of the process. This analysis can help to identify areas where energy consumption can be reduced and optimize the process to minimize energy use. Therefore, including operational energy consumption analysis as a criterion for the application of the electrooxidation process at a real scale is important to ensure that the process is both environmentally sustainable and economically feasible in the long run. The flow rate at which the wastewater is fed into the EOP system is an important factor that affects both the removal efficiency of pollutants and the energy consumption of the system. If the retention time of the wastewater in the system is increased, the energy consumption and cost of the system may also increase. On the other hand, at high flow rates, the pump may require more power, leading to higher energy consumption. Therefore, it's crucial to find an optimal balance between the flow rate and retention time to achieve high removal efficiency without significantly increasing energy consumption or system cost. By optimizing the system for these conditions, it's possible to achieve efficient and cost-effective wastewater treatment. In the study, the flow rate was found to be 13.93 mL/min in order to provide high removal efficiency and low energy consumption as a result of RSM optimization. The relationship between energy consumption and pH in electrooxidation varies depending on the specifics of the process, including the type of electrode, the electrolyte solution, and the operating conditions. However, in general, higher pH levels lead to increased energy consumption due to the competition of hydroxide ions with anodic oxidation reactions [96,97], requiring higher overpotentials and longer reaction times. In contrast, lower pH  The results of the study showed that electrooxidation is an effective technique for obtaining satisfactory removal efficiencies of β-estradiol [94] levels can enhance anodic oxidation reactions and improve current efficiency, leading to lower energy consumption [98]. The optimal pH for electrooxidation depends on the specific application and pollutant being treated, with neutral or slightly acidic pH being preferred in some cases [99], while more alkaline pH may be necessary in others. In the study, at the end of the RSM optimization, it was found that pH 3 was the lowest energy consumption for maximum removal yield and minimum energy consumption. In electrooxidation processes, current density is critical to energy consumption. If the current density is too high, it can cause an increase in potential difference, leading to a rise in the temperature of the wastewater [100]. This can result in energy wastage and inefficiency, making the process economically unfeasible. High current density can also support the electrolysis of water to O 2 , which reduces the generation of radical OH • radicals [97,101]. Furthermore, the life of the electrodes may decrease, and the energy consumption of the system may increase at higher current densities [102]. Therefore, optimizing the current density is a vital parameter in electrooxidation processes. The highest removal and lowest energy consumption were obtained at 10 A (3.03 mA/cm 2 ) current density under optimum operating conditions. The amount of energy used per m 3 of wastewater at pH 3.00 at optimum flow rate (13.93 mL/min) and current density (10 A) was found to be 41.606 kWh. The calculation of energy price was carried out using market values specific to Turkey. In April 2023, the average price of electricity in Turkey was 0.266 dollars per kilowatt-hour (kWh). The relatively high energy consumption may be due to the low conductivity of domestic wastewater and the absence of supporting electrolytes. While the supporting electrolyte provides good conductivity, it can also reduce energy consumption by reducing the reaction time by producing oxidants that provide indirect oxidation [35,99]. Considering the treatment of endocrine disruptors with electrooxidation, both the 10 A current density (3.03 mA/cm 2 ) is quite low and the reaction time is quite short, 13.93 mL/min (1.19 h) [103][104][105].

Conclusion
RSM results revealed substantial impacts of factors, such as the linear and quadratic interactive effects of pH, current density, and flow rate on the removal of EDCs. RSM modelling results showed that high R 2 values (Triclosan = 0.998, 17α-Ethinylestradiol = 0.984, β-Estradiol = 0. 998, and Energy Consumption = 0.999) could accurately anticipate the electrochemical process behaviour for EDCs removal. The second-order (quadratic) model well fits the experimental data. Under optimum experimental parameters for the EOP system (Flow Rate = 13.93 mL/min, Current Density = 10 A, and pH = 3.00), the maximum removals of Triclosan, 17α-ethinylestradiol and β-estradiol and Energy Consumption were 91.65%, 96.43%, and 96.66%, and 41.606 kWh/ m 3 , respectively. Simultaneously, it was determined that Estrone could not be completely removed using the EOP system and that it was present in high concentrations in treated water. This study demonstrates that the EOP process may be an effective alternative to conventional treatment methods for the removal of EDCs. Considering the EDCs that restrict the reuse of water in terms of energy consumption and are very difficult to remove, it should not be seen as a very costly treatment method in terms of real-scale use. To reduce energy consumption, studies on the support electrode type and concentration can be considered.