A combined treatment system of O₃/UV oxidation and activated carbon adsorption: emerging contaminants in hospital wastewater

Combining UV irradiance with chemicals such as peroxides can improve the scope of UV irradiance, resulting in synergistic oxidation processes (Masschelein & Rice 2002; Sanei & Mokhtarani 2022). Indeed, hydroxyl radicals, produced upon bubbling of ozone through water, make UV-oxidation more effective (Safarzadeh-Amiri et al. 1997; Mehrjouei et al. 2015; Dehkordi et al. 2022). Thus, it is thought that combining ozone and UV irradiance will improve EC degradation in wastewater. Through ^•OH radicals in the wastewater, ozone reacts quickly with a variety of recalcitrant compounds and ECs, eventually destroying them. This process can help synergize the therapeutic benefits of CGCT + UV and overcome the respective limitations. The performance of the CGCT + UV method for hospital wastewater in Iran was evaluated in this study through laboratory tests. The removal of recalcitrant pollutants and ECs from wastewater was improved through GAC level oxidization before ozone treatment using catalytic ozonation with GAC playing the role of a catalyst. In this study, the CGCT + UV method was investigated to remove amoxicillin as a representative of pharmaceutical compounds in the wastewater. Additionally, the removal rates of Total Organic Carbon (TOC), Chemical Oxygen Demand (COD), and Total Dissolved Solids (TDS) were measured and compared to other physical-chemical treatments in terms of retention time and ozone dose. The CGCT + UV method also includes a mechanism for removing recalcitrant compounds.

The MLSW was immediately characterized for NH₄-N, BOD5, pH, COD, TOC, TN, NO₃-N, alkalinity (CaCO3), conductivity, and alkali metal cations, by employing standard methods. An Orion 710A pH meter was used to measure the MLSW pH before treatment. 0.1 N HCl and 0.1 N NaOH solutions were utilized to adjust pH. A spectrophotometer of Thermo Fisher Scientific type Genesys 20 (4001/4) was used to measure NH4-N and COD concentrations (USA). TOC was measured using a Shimadzu 5000 A TOC analyzer (USA). The conductivity meter type YSI 63/25 (USA) was used to measure conductivity. Amoxicillin concentrations were determined using Thermo UltiMate 3000 HPLC (USA). The content of aromatic compounds and olefins (alkenes), which, in general, react quickly with ozone, was estimated using UV₂₅₄ on diluted samples. The measurements were carried out using a Thermo Spectronic Helios spectrophotometer as well as 1 cm quartz cells (Beltrán et al. 2009). Additionally, amoxicillin concentration was measured using a Knauer HPLC (C18; 250 × 4.6 × 5) with a UV detector at 190 nm. Buffer phosphate and acetonitrile with ratio of 60:40 (V/V) were used as the mobile phase at an injection flow rate of 1 mL/min. It should be noted that prior to testing, samples were filtered using 0.2 μm cellulose acetate filters to avoid the entry of particles into HPLC.

Adsorption tests were carried out for 240 min to study the adsorption of amoxicillin (AMX) (Fig. S1, Supplementary Information). According to the results, for all concentrations, adsorption increased significantly up to the first 60 min of the experiments. However, when the time increased four times, the adsorption increased only 20%. Therefore, the following experiments were done in 60 min. As can be seen from Fig. S2, Supplementary Information, the pH_pzc was 8.4 which shows that AC, at pH higher than pH_pzc has an anionic surface while at pH lower than pH_pzc the surface is cationic which facilitates adsorption of anions. Since the pH_pzc of the AC was in a basic range, it had significant catalytic potential (Sanchez-Polo & Rivera-Utrilla 2003).

Adsorption onto activated carbon is a simple way to remove many ECs from aqueous solutions. The appropriateness of utilizing activated carbon (charcoal) in an adsorption application, on the other hand, is determined not only by its uptake capacity but also by its ability to be reused multiple times. Thermal reactivation is the most widely used technology for regenerating spent activated carbons, but it has some disadvantages, such as high energy demand and regeneration costs and carbon loss caused by attrition and oxidation (San Miguel et al. 2001). Thus, it is desirable to conduct R&D on alternative methods of regeneration that can be carried out in situ to save money on shipping. In this case, ozone-assisted regeneration of the spent activated carbon (SAC) could be a viable option. The method entails the long-term adsorption of wastewater pollutants onto the activated carbon surface until reaching the saturation point. The next step involves the in situ regeneration of the SAC via a short-term reaction with gaseous ozone. This method has already been tested on activated carbons that have been exposed to phenol and benzothiazole (Álvarez et al. 2004). In this study, to achieve an affordable method which prepares GAC particles to be used several times two strategies were used. GAC was placed in a glass jar before being used. Through ozone oxidization and NaOH treatment of the GAC surface, the functional group comprising oxygen was improved. Approximately 0.5 l of 0.1 N NaOH were poured into the glass, and the suspension was stirred for 2 h at 150 revolutions per minute. The specimens were then removed from the solution and dried for 24 h at 115 °C in the oven. GAC samples were placed in a glass column with various heights and diameters, resulting in various GAC densities. After that, gaseous ozone was used to treat the adsorbent. After that, the samples were rinsed with DIW and then dried in a 115 °C furnace for 24 h before being stored in a dryer. Secondly, pumping ozone from the bottom of the GAC tank, besides increasing system efficiency which greatly reduces the consumption of GAC particles, makes catalytic ozonation using GAC as a catalyst which greatly increases the useful life of GAC particles. The physical properties of the adsorbent are shown in Table 1.

Analysis of variance (ANOVA) was used to examine the results statistically. The determined R² and coefficient of determination R² were used to represent the model's proper value. F-tests were used to assess the effects of linear and quadratic terms. The final subset of variables was carefully chosen based on a 95% confidence level for the P-value. The R² coefficient was predicted to determine the planned model's prognostic ability. The predicted residual error sum of squares (PRESS) was used to carry out these procedures. The relationships between the empirical levels of each of the responses and variables were indicated using surface plots and contour plots of the fitted polynomial equation. The statistical model validated all variables in the designed space. Thus, a batch test was conducted under ideal conditions, and the results were compared to the expected values. To create 3D designs of responses and their outlines, a traditional program developed in Design Expert was employed. It was used to show the relationships between two variables (Dezvareh et al. 2022).

The features of MLSW are shown in Table 2. MLSW samples were taken in April 2021 (spring) and October 2021 (fall) to compare COD, TDS, and TOC concentrations. Despite this, no significant seasonal variations were observed. This occurrence suggests that the physical and chemical properties of MLSW have little impact on the amount of waste generated. This was consistent with Hong Kong's and Taiwan's (Fan et al. 2006) studies, in which there was no significant seasonal variation in leachate quality. Seasonal changes in the quantity and quality of wastewater have a different impact in different regions, depending on the seasonal climate/hydrology of the site (Kurniawan et al. 2006; Dezvareh et al. 2023).

MLSW, as opposed to organic compounds, contained trace amounts of minerals like heavy metals, as evidenced by the low measured conductivity. The color of the MLSW was dark-brown, with a Chroma of around 2000, and a foul odor, owing to the presence of organic acids decomposed from high organic matter concentrations. Funky and dark colors fade away or turn into light as wastewater ages; otherwise, such a change could be attributed to precipitation characteristics and the quality/quantity of wastes.

Pezoti et al. studied the adsorption of AMX with an AC from Guava seed which was modified with NaOH. They reported that the adsorption isotherm followed Langmuir and Redlich-Peterson models with R² of 0.96 and 0.93, respectively. Additionally, the free adsorption energy was measured 18.14 kJ/mol which indicated a chemical adsorption of AMX on the synthetic AC (Pezoti et al. 2016).

Thermodynamic studies were done at temperatures from 25 to 65 °C in 60 min. The effect of temperature on the adsorption process is presented in Fig. S8, Supplementary Information. According to the results, as the temperature increased the adsorption efficiency decreased. This shows that the adsorption process at temperatures higher than 25 °C was exothermic. Also, the results from thermodynamic studies are shown in Table 4. As represented in the table, the negative value for enthalpy approves the exothermicity of the process. Moreover, the low value of entropy (40 kJ/mol) indicates the physical adsorption of AMX on the AC (Zuim et al. 2011; Rathod et al. 2015). The reaction was also spontaneous because the free Gibbs energy was negative.

In another study, Saucier et al. investigated adsorption of amoxicillin and acetaminophen using modified AC with Iron (III) and cobalt benzoate (MAC-1) and Iron (III) and cobalt oxalate (MAC-2). They reported that adsorption of AMX was endothermic. Also, the enthalpy was measured lower than 40 kJ/mol, suggesting the physical adsorption of AMX on both modified ACs (Saucier et al. 2017).

According to Figure 3(a) as the ozone dose increased from 0.54 to 1.06 g/h, AMX removal increased from 30 to 82% in 30 min. However, by increasing ozone dose from 1.06 to 1.29 g/h, the removal efficiency increased only 6% which can be attributed to the presence of compounds with low reactivity with ozone. Hence, ozone dosage of 1.06 g/h was selected as the optimum value. Additionally, according to the figure, the removal efficiency after 30 min of the reaction did not change significantly. Therefore, the following tests were carried out in 30 min. Figure 3(b) demonstrates the effects of ozone dosage on AMX removal in catalytic ozonation. According to the results, as the ozone dose increased from 0.54 to 0.82 g/h, the removal efficiency increased 20% in 30 min. However, the removal efficiency only increased 6% when the dosage increased from 0.82 to 1.29 g/h. Thus, 0.82 g/h was selected as the optimum ozone dosage. Additionally, longer treatment time (60 min) did not facilitate the removal of AMX indicating that 30 min is the optimal treatment period. Based on the results, it is crystal clear that addition of AC facilitated the removal of AMX and decreased the ozone dose. This can be attributed to higher concentration of reactive radicals in the solution due to addition of the catalyst (Moussavi et al. 2015).

Figure 3(c) presents ozone dosage on AMX removal in CGCT + UV. According to the figure, by increasing ozone dosage from 0.54 to 0.82 g/h, AMX removal enhanced from 73 to 92 in 30 min while the removal did not exceed 6% when the dose increased from 0.82 to 1.29 g/h. Therefore, 0.82 g/h was selected as the optimum ozone dose in this process. Also, increasing time over 30 min had a negligible effect of AMX removal, so the 30 min was selected as the optimum time. UV facilitates ozone decomposition into reactive radicals, so the removal efficiency improved in the presence of UV source (Chávez et al. 2016).

Various GAC concentrations for MLSW treatment were explored in this study to see if they had any direct effects on amoxicillin removal. GAC was poured at various weights per litter of MLSW for this purpose, resulting in different granular carbon concentrations.

Solution pH is an important factor for adsorption of compounds because changes in pH result in the change of surface charge of the adsorbent (Samarghandi et al. 2015; Arya & Philip 2016; Shan et al. 2016; Zhao et al. 2016; Saucier et al. 2017). The optimum pH for adsorption of organic matters depends on the chemical characteristics of the adsorbent and solubility of organic compounds in the solution pH (Saucier et al. 2017). In this regard, pH_pzc is an influential factor indicating the surface charge. At pH = pH_pzc, the surface charge is neutral, while at pH > pH_pzc and pH < pH_pzc, we have negative and positive surface charge, respectively. AMX has various forms in aqueous solutions due to ionization of its functional groups. AMX has pKa₁ = 2.68, pKa₃ = 9.63, and pKa₂ = 7.49 for carboxyl, phenol, and amine functional groups, respectively. At pH < pKa₁ functional groups in AMX are protonated (–COOH, –NH₃⁺, –OH), while at pKa₁ < pH < pKa₂, carboxyl group will be deprotonated and converted into carboxylate (–COO⁻ , –NH₃⁺, –OH). Also, at pKa₂ < pH < pKa₃, amine group will be deprotonated. Finally, at pH > pKa₃, the phenol group loses its proton (–COO − , –NH₂, –O⁻) (Pezoti et al. 2016). At pKa₂ < pH < pKa₃ (7.49 < pH < 9.63) due to the presence of carboxylate ions in AMX and pH is lower than pH_pzc, the surface of AC is positive. Thus, at pH = 8, the maximum adsorption was achieved. It should be noted that the initial solution pH was 7.5 and because of its proximity to 8, the initial pH was selected as the optimum pH.

In Xing Zha et al.'s study, the removal efficiency of AMX with modified bentonite was 81.9%. The authors reported that as pH increased from 3 to 7, the removal enhanced due to the charge of AMX molecules becoming neutral. Also, as the pH increased from 7 to 11, the removal efficiency did not change significantly (Zha et al. 2013).

Figure 5 shows the changes in pH during a single ozonation process. As shown in the figure, pH played a key role in neutral and alkaline condition rather than acidic condition. As the pH increased from acidic to basic condition, the removal efficiency enhanced dramatically. However, the removal did not change noticeably at pH values above 8. Therefore, pH = 8 was chosen as the optimal value. In addition, the final pH of each sample is depicted in Fig. S5, Supplementary Information.

Ozone oxidizes compounds directly and indirectly. In the direct way, ozone molecules oxidize organic matters while in the indirect way, reactive radicals such as hydroxyl radicals which are the byproducts of ozone decomposition are responsible for oxidation of organic compounds. Solution pH has an essential role in the performance ozonation. At basic pH, higher ^•OH radicals can be generated which increases the removal efficiency. In single ozonation at acidic pH, ozone directly and slowly interacts with specific groups such as aromatics with double carbon bonds. The final products will be aldehydes and carboxylic acids (Kurniawan et al. 2006). In contrast, ^•OH radicals are 10⁶–10¹² times faster than ozone molecules resulting in faster oxidation of organic compounds. This is because ozone is a selective molecule and only interacts with compounds rich in electron (Munter 2001). At basic conditions, in addition to ozone molecules, reactive radicals are present in the solution which accelerate the removal process. Furthermore, AMX at basic pH is in its ionic forms which facilitates its reactivity with other molecules (Moussavi et al. 2015).

As shown in Figure 6(a), by increasing pH from 2 to 10, the removal of AMX increased significantly from 22 to 92% in the catalytic ozonation process. Ozone-based reactions occur slowly at acidic conditions whereas at basic conditions, due to accelerated ozone decomposition into hydroxyl radicals, the oxidation rate increases (Chaturapruek 2003). It is worth mentioning that the increase of pH above 8 only enhanced the removal efficiency by 7%. In this regard, the optimum pH of 8 was selected for further experiments. In addition, the final pH at each sample was measured and is shown in Fig. S6, Supplementary Information. During ozonation, organic acids are produced resulting in a lower final pH at the end of the process which is confirmed in Fig. S6, Supplementary Information (Zhang et al. 2013). Figure 6(b) illustrates the effect of pH during the CGCT + UV process. The results indicate that the process performed better at basic conditions, but at high alkaline conditions, the removal efficiency is not significant. Thus, like catalytic ozonation, pH of 8 was selected as the optimal value. Also, Fig. S7, Supplementary Information shows the final pH values at the end of the process.

In alkaline and neutral solutions, the dominant ozone reaction is shown in Equation (11). This reaction limits the occurrence of other reactions shown in Equations (9) and (10). In contrast, at pH values below 7, UV light plays an essential role in ozone decomposition (36). The increased final pH at acidic conditions can be attributed to enhanced generation of hydroxyl radicals in the presence of UV light which results in the oxidation of carboxylic acids and increase in the final pH.

Figure 7(b) shows that in the first 15 min of the processes, ozone consumption was 4.56, 4.73, and 4.88 mg/mg AMX for CGCT + UV, catalytic ozonation, and single ozonation, respectively. As the process time increased to 60 min the ozone consumption increased. Nevertheless, the ozone consumption in the CGCT + UV process was 3.13 and 0.59 mg/mg AMX lower than single and catalytic ozonation, respectively. It is crystal clear that addition of AC and UV improved the generation of reactive radicals and decreased ozone consumption rate.

In addition, the generation of byproducts is another possible reason for the higher ozone consumption in the solution. In addition to AC, UV facilitates ozone decomposition rate resulting in an efficient O₃ consumption. Also, the ozone consumption as the removed TOC is shown in Fig. S9, Supplementary Information. Based on the results, CGCT + UV had the lowest ozone consumption rate. In another study on the removal of ECs, Quiñones et al. reported that ozone consumption was 1.47, 0.19, 0.205 mg O₃/mg TOC in single ozonation, UV light assisted ozonation at pH = 3, and UV light assisted ozonation at pH = 7, respectively (Quiñones et al. 2015).

The statistical design of experiments and data analysis were conducted using the Design Expert Software. Besides, CCD designs and second-order RSM were utilized to examine the effects of the three independent variables on the performance of responses. Thirty-nine normalized observations were utilized to generate experimental data. The variables that were measured were ozone concentration (A) and GAC density (B). The interaction between various factors was identified using CCD. −, −1, 0, +1, and + were the levels of the independent variables. Due to previous and pilot studies, the range was identified. According to Lak et al. (2012), such codes can be used to fit regression models with variables ranging from – to +.

For TDS, COD, and TOC, Table 5 shows regression parameters’ ANOVA for the estimated quadratic response surface models, along with other statistical parameters. The total variation in response predicted by this model (TDS = 0.8525, COD = 0.8677, and TOC = 0.8792) is represented by the R² coefficient (TDS = 0.8525, COD = 0.8677, and TOC = 0.8792), which equals the ratio of the sum of squares of regression to their sum total. TDS, COD, and TOC have high R² values (85.25, 86.77, and 87.92%, respectively), implying that the results are consistent. The correlation coefficients should be at least 0.80 for a good model fit. A rational modification of the quadratic model was confirmed by a big R² coefficient to the empirical data. The R² values of 78.88, 85.22, and 82.62% in this study indicate that the regression model was able to shed light on the relationship between the independent variables and the response. The statistical significance of the model was established by determining the model's coefficients. For COD, TDS, and TOC of MLSW, the R² values were 0.8677, 0.8525, and 0.8792, respectively. With ‘Adj R²’ of 0.7888, 0.8922, and 0.8262, the ‘Pred R²’ of 0.6834, 0.8373, and 0.7268 indicated a good arrangement. The signs to noise ratio was calculated using ‘Adeq Precision.’ For this criterion, a ratio of >4 represented favorability. The ratios of 16.007, 19.681, and 17.758 showed a positive sign, indicating that the model can be employed in design space planning. The models were improved by removing the study's irrelevant model terms with limited effect. Thus, the COD, TDS, and TOC response surface models were thought to be reasonable.

Plots of expected vs. actual values for removal of parameters are shown in Fig. S10a, S11a, and S12a, Supplementary Information. The Design-Expert software displays normal probability plots (NPPs) of diagnostics and ‘studentized’ residuals to demonstrate if the selected model could properly estimate the real-world system. Data regularity regarding estimated values was investigated using SDs of actual values. Following that, the NPP indicated that the residuals were on the basis of the normal curve distribution, i.e., an important assumption to consider when evaluating a statistical model's suitability. In case of the plot points focusing on a straight line (Fig. S10b, S11b, and S12b, Supplementary Information), the residuals are normally distributed. The standardized residuals for TDS, TOC, and COD are plotted in normal probability plots. The stronger link between predicted and experimental data indicated minor differences. The diagnostics’ standard probability outlines were studentized and residuals were presented to demonstrate an adequate estimate of the real-world system using a given model. Thus, the data were normally disseminated within the responses of the selected models. Partially residual plots show the linearity extent and direction, as well as deviations from linearity such as heteroskedasticity, curvilinear associations, and outliers.

TDS was reduced by 31.8–58.2% using the CGCT + UV method. Lower ozone inlet concentration (X₁) and GAC density (X₂) were found to have lower TDS removal efficiencies. The findings were evaluated using ANOVA results, and the ‘fitness of fit’ was investigated. The empirical findings were used to create an empirical formula for the removal of TDS through the CGCT + UV procedure based on the response to the variables. The overall error in the prediction of TDS using the RSM method was 4.9% when compared to experimental data, according to the findings. Furthermore, the experimental data and formula were found to be very consistent in terms of TDS removal. This agreement indicates that the experimental formula presented was adequate for providing an estimate of the removal of TDS while maintaining a reasonable level of logical consistency.

Figure 9(b) shows the effects of parameters A and B on the removal of COD. According to Equation (17), since the coefficient of B is smaller than A, the effect of B on COD removal would be lower than A. Thus, as shown in the figure, the ozone concentration graphs have steeper slopes in comparison to GAC density graphs. This trend can be attributed to higher efficiency of chemical treatment methods such as ozonation in comparison with physical treatments like adsorption. According to Figure 9(a), since the sign of both A and B parameters is positive, the removal efficiency will enhance by increasing the value of these parameters. At ozone dose of 150 mg/l and GAC density between 1.1 and 1.4 g/cm³, the graph becomes concave which shows the optimal condition for COD removal. It is noteworthy that increasing ozone in this range is neither economical nor operational. Also, comparing with TDS removal, less GAC is needed that can affect filling of GAC tanks and economic aspects. By utilizing activated carbon and adjusting the ozone dose from 10 to 75 l/h in 30 min, Chunmao Chen et al. managed to increase the removal efficiency of COD from 18.8 to 34.44% (Chen et al. 2014a). Chunmao Chen et al. investigated the effects of changing the ozone dosage in the presence of CAC-supported manganese oxides for 0–120 min by changing the ozone dosage from 0.795 to 3.150 g/h in a similar study (Chen et al. 2014b). The removal efficiency was nearly fixed in the plot of COD removal in Lecheng Lei et al.'s study, and the system attained a peak efficiency from 40 min on at a 60 l/h ozone flow rate in the presence of a GAC adsorbent (Lei et al. 2007).

Figure 10(b) presents the interactions of parameters A and B on the removal of TOC. According to graphs’ slopes and coefficients of A and B in Equation (18), ozone had greater influence on removal of TOC. Also, by increasing the values of A and B, TOC removal enhanced. Based on Figure 10(a), the optimum ranges for GAC and ozone dose were selected 1.4–1.7 g/cm³ and 150 mg/l, respectively. In Moreira et al.'s study, the mineralization % of diclofenac and amoxicillin after 1 h of single ozonation was 24 and 32%, respectively. However, when UV was added to the system, the mineralization enhanced to 48 and 64%, respectively (Moreira et al. 2016).

The Design-Expert software was utilized to perform the optimization procedure and determine the optimum values of TDS, TOC, and COD removal performance. The optimum conditions were predicted using the numerical optimization section (Nehra et al. 2008). The optimal values of input parameters were acquired by solving the quadratic model based on experimental results and analysis of response surface plots. The desired objective was chosen ‘within’ the range for each practical state (GAC density and ozone concentration) in the software optimization stage. Responses (TOC, TDS, and COD) were determined as the maximum removal for obtaining the highest efficiency. The software combined the individual desirability into a singular number. Then, the optimization is searched in terms of the response, after achieving the optimum removal efficiencies and working conditions (Table 6). As shown in Table 6, the removal percentages of 62.34, 61.63, and 55.13% are forecasted for COD, TDS, and TOC, respectively. The model's improved practical circumstances in this regard are the ozone dosage of 151.49 mg/l, and the GAC density of 1.39 g/cm³. The desirability function value was 1.0 for these optimum circumstances. Then, a further experiment was performed to confirm the desired results. There is a consistency between the experimental results and the estimated response values.

To assess the performance of different treatments assessed in this paper, a comparative study was performed based on the essential dose (g/l), pH, biodegradability, and range of primary COD concentrations (mg/l) for wastewater. Considering the various test conditions, e.g., temperature, pH, wastewater resistance, hydrological site, and seasonal climate, it has only one relative meaning. However, such a comparison is still effective in the evaluation and assessment of each technique's effectiveness for wastewater treatment. Hence, the comparison contributes to evaluating each method's overall performance. Table 7 shows the removal efficiency of various methods (individuals and integrated) on COD in some other studies. Considering the high ratio of BOD₅/COD of the local wastewater, it seems using a biological process in combination with the CGCT + UV process could be useful to enhance the wastewater biodegradability and enhance the performance of removal of effluent parameters (TDS, COD, TOC). Therefore, despite indicating a lower removal of effluent parameters in comparison with other methods, the performance of the CGCT + UV treatment is still auspicious. This treatment can treat wastewater of changing strengths and change recalcitrant compounds in the wastewater into more biodegradable compounds via oxidation procedures. Subsequently, through biological treatments, the target compounds’ degradation can be complemented in the wastewater before their discharge. Moreover, the H₂O₂ decomposition by GAC is activated with no energy input, thus, treatment costs are reduced. In Table 8, comparison between some studies which used adsorption and ozonation methods is shown as well. Moreover, evaluation of the performance of methods in some studies in purifying the wastewater and pollutants by considering ozone and UV methods is depicted in Table 9.

Pursuant to the results, the performance of the CGCT + UV method is acceptable to treat the MLSW, and utilizing this process is highly recommended to purify all kinds of wastewaters that carry the same characteristics of MLSW. The CGCT + UV process has superiorities for wastewater treatment containing the structure simplicity and its susceptibility to adopt various strong compounds with seasonal changes. The RSM model used in the optimization of the production process caused increasing energy efficiency and made the process energy-efficient and sustainable by considering operating variables, such as ozone concentration and GAC density. The optimal conditions of the RSM-based model are the ozone dose of 151.49 mg/l, and the GAC density of 1.39 g/cm³ with the forecasted removal percentage of 61.63, 62.34, and 55.13% for TDS, COD and, TOC, respectively. In local condition experiments, the CGCT + UV process showed an acceptable efficiency by the removal percentage of TDS = 59.50%, COD = 67.40%, and BOD = 56.70% among the other individual and combined methods, nonetheless, this method was not sufficiently capable to treat the wastewater meeting discharge standards. Thus, other biological methods such as nitrification or activated sludge are required to supplement leachate treatment and meet environmental directives.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.