Response surface modeling for the treatment of methylene blue from aqueous media using electro-Fenton process before determination by UV-Vis spectrometer: Kinetic and degradation mechanism

In the present study, response surface methodology was employed to investigate the effects of main variables, including the initial MB concentration, hydrogen peroxide dosage, current density, and electrolysis time on the removal efficiency of MB using the electro-Fenton (EF) process. The MB concentration determination by UV-Vis spectrometer . The EF process degrades the MB contaminant molecules by the highly oxidizing species of the • OH. A quadratic regression model was developed to predict the removal of MB, where the R 2 value was found to be 0.9970, which indicates the satisfactory accuracy of the proposed model. ANOVA analysis showed a non-significant lack of fit value (0.0840). Moreover, the predicted correlation coefficient values (R 2 =0.9915) were reasonably in agreement with the adjusted correlation coefficient value (R 2 =0.9958), demonstrating a highly significant model for MB dye removal. In addition, the obtained results showed 95.8% MB was removed in the optimum removal efficiency, including the initial MB concentration of 20 mg L -1 , H 2 O 2 dosage of 400 μL, and the current density of 7.0 mA cm -2 , and electrolysis time of 10 min which was agreed with the predicted removal efficiency of 98.3%. Electrical energy consumption was found to be 0.163 kWh m -3 . The constant rate value of K app at the optimum operating condition was 0.3753 min -1 .


Introduction
Due to the increasing use of tens of thousands of various types of dyes, which are employed in plastic products, cosmetics, leather, and food industries, and their release of about over 100 tons into the aqueous environments, a severe global concern has been raised for efficient removal of such organic micropollutants before releasing the industrial wastewater into the environmental medium [1,2]. The adverse effects of toxic dyes on human beings, biota, microorganisms, and environmental health, even at deficient concentrations on the one hand, and their high resistance to photochemical and biological degradation, on the other hand, the necessity has required to eliminate the synthetic dyes contaminant from effluents [3][4][5]. Among the high-consumption dyes, methylene blue (MB), a derivative of phenothiazine, is widely used as a redox indicator, biological staining reagent, medication to treat methemoglobinemia, and adjunct employed in endoscopic polypectomy. Besides, it is used extensively in textiles dyeing industries of cotton, wool, and silk. Exposure of methylene blue to the eyes causes irreversible damage of partial blindness [6,7]. Moreover, its inhalation causes respiratory problems. Its ingestion orally causes a burning sensation in the gastrointestinal tract and induces nausea, vomiting, profuse sweating, as well as mental confusion, and methemoglobinemia [8][9][10][11].
There are many techniques such as UV-VIS spectrometer, gas chromatography, and liquid chromatography for determining MB, dye, tyrosine kinase, toluene, benzene, and organic materials in different matrixes [12][13][14][15][16][17]. During the last decades, several treatments techniques, including coagulation, adsorption, and filtration processes, have been developed for the removal of dye contaminants which suffered from the disadvantages such as high operational costs due to the production of a massive amount of sludge or solid waste and transferring the dyes to another phase by concentrating them rather than eliminating them [18,19]. The advanced oxidation processes (AOPs) as an economical and high efficient approach provide a valuable opportunity to overcome the limitations of the conventional treatment process by generating a robust oxidative agent of hydroxyl radicals, which degrade and mineralize the dye molecules into their less-toxic forms and harmless inorganic products such as water or carbon dioxide [20][21][22]. As seen from the proposed reaction in Equation 1, Fenton's reagent of Fe 2+ and H 2 O 2 generates the oxidative agents of hydroxyl radicals under the acidic medium of solution pH less than 3. The Fe 2+ species are generated from the oxidation of iron anode electrode and reduction of Fe 3+ ions through cathodic reaction as described by Equation 2 [23,24].
In the current work, response surface methodology (RSM) was employed as a computational approach to optimize the removal efficiency of MB using the EF process. The effect of main operational variables, including electrolysis time, MB initial concentration, H 2 O 2 dosage, and current density throughout the treatment process, were evaluated, and the kinetics of the process were studied.

Chemical
All chemicals used in the MB removal study are analytical (high purity) grade without further purification. C 16

EF process system
The Pyrex cylindrical reactor with dimensions of 9.0 cm (height) × 7.0 cm (diameter) and a volume capacity of 300 mL was used to carry out MB removal studies using the EF process. The iron electrodes with a thickness of 0.1 cm and an immersed area of 12 cm 2 were used as cathode and anode electrodes. The distance between electrodes (anode and cathode) was considered constant and equal to 3.0 cm. In a typical run, MB solution was degraded in the reactor with 50 mM Na 2 SO 4 as the supporting electrolyte. The pH of the MB solution was adjusted to 3.0 by 0.01 M of H 2 SO 4 . The rotation speed was kept at 200 rpm during the experiment.

Analytical methods
The MB concentration of samples before and after the treatment process was measured by a UV-Vis spectrometer (OPTIZEN 3220 model) at 670 nm [1]. Evaluations for the degradation and MB removal efficiency and total amount of electrical energy consumption (kWh m -3 ) were calculated from the Equation (3) and (4) [25][26][27]: where C 0 (MB concentration before EF process; mg L -1 ), Ct (MB concentration after EF process; mg L -1 ), U (applied voltage; V), I (electrical current; A), t (electrolysis time; h), and V (volume of sample;L).

Methodology
Response Surface Methodology (RSM) is a widely used mathematical and statistical approach employed to model, design, and evaluate the relationship between several independent variables and responses of the proposed model. The goal is to optimize the applied reaction in a short time and reduce the costs of the process. Unlike conventional methods for data analysis, RSM analyzes data using simple techniques based on the mathematical model. In the RSM, to optimize the studied variables, a polynomial function (often a quadratic polynomial model) can be used, as given in Equation 5 [27,28] (Eq. 5) where Y is the predicted response (MB removal %), k is the number of independent factors, β 0 , β i , β ii , and β ij are the constant, linear, quadratic, and interaction model coefficients, respectively, also, X i , X j , and ɛ are the independent factors and the error. Analysis of variance (ANOVA) with a 95% confidence interval was used to determine the significance of the parameters. In the current study, the RSM approach was used to investigate four variables' effects (initial MB concentration, H 2 O 2 dosage, current density, and electrolysis time) and identify the optimal condition for MB removal using the EF process employing Design-Expert Software Version 11.0.4.0. Some valuable parameters, including lack of fit of the model, the significance of linear and interaction effects of operating variables, and coefficient of determination, were evaluated by analysis of variance (ANOVA). As seen in Table 2, the p-value less than 0.0001 corresponds to the significant coefficients.

Development of models and analysis of variance (ANOVA) using RSM
As seen, the obtained results from ANOVA analysis confirmed that the developed quadratic regression model was satisfactorily fitted to the removal efficiency of the treatment process with the F-value of 858. Since the calculated p-value of the Lack of Fit was less than 0.05, it confirmed that the lack of fit was insignificant, corresponding to the pure error. The predicted and adjusted R 2 of 0.9915 and 0.9958 are in reasonable agreement. Furthermore, the fitted model's prediction ability over the experimental condition range is acceptable due to the obtained predicted R 2 of about 100%. The normality of the residuals was analyzed and demonstrated in Figure 1(a). As seen, all the demonstrated data are relatively near to a straight line with the R 2 of 0.9970. Moreover, the association between the obtained values of actual removal efficiencies and predicted removal efficiencies were revealed in Figure 1(b). As seen, since the residual results were distributed near the diagonal line, it can be concluded that the proposed treatment model successfully predicted the removal efficiencies. The regression model developed for the removal efficiency of MB is represented in Equation 6. The proposed second-order model in the term of coded factors by eliminating the insignificant terms is expressed as follows (6): (Eq. 6) Here, X 1 represents initial MB concentration, X 2 is H 2 O 2 dosage, X 3 is current density, and X 4 is electrolysis time. As seen, the intensity of each particular variable on the removal efficiency as the response of the proposed model is identified by the related magnitude of each variable's coefficient. Each coefficient's positive or negative value indicates the synergistic or antagonistic effect of the related variable on the response. As seen, the coefficients   of X 2 (H 2 O 2 dosage), X 3 (current density), and X 4 (electrolysis time) are all positive. Moreover, it can be concluded that the variables of X 1 and X 2 with much larger coefficients play more significant roles in MB's removal efficiency as the model's response.

Effect of parameter on the MB removal efficiency
In the present study, the effect of initial MB dye concentration on removal efficiency using the EF process was investigated. As Figure 2 shows, the initial MB dye concentration is inversely related to the MB removal efficiency, when increasing the initial MB dye concentration from 10 mg L-1 to 500 mg L -1 , and removal efficiency decreases from 99.8% to 68.2%. This phenomenon can be due to the decrease in the ratio of oxidant agents to MB dye molecules. For example, if all the parameters studied in the EF process, are constant (at high MB dye concentrations), the number of dye molecules is greater than the oxidizing agents. Therefore, the number of oxidizing agents is insufficient to remove high concentrations of the dye molecule, and consequently, the removal efficiency decreases [29,30].
Optimizing the amount of H 2 O 2 in all treatment methods in which H 2 O 2 plays a key role is very important because it affects the removal efficiency, the management of the treated effluent, and the cost. The effects of H 2 O 2 dosage in the range of 50-500 µL on MB removal were investigated at the initial MB concentration of 30 mg L -1 , the current density of 6.5 mA cm -2 , and electrolysis time of 8.5 min, the result is shown in Figure 3 the obtained results show that the increase in degradation percentage from 40.7% to 78.1% whenever H 2 O 2 increased from 50 to 500 µL, at the initial MB concentration of 30 mg L -1 , the current density of 6.5 mA cm -2 , and electrolysis time of 8.5 min.  (7), as well as the decomposition of • OH, produced during the EF reaction to radicals with lower oxidation strength (HO 2 • ions) according to Equation (8) and (9) which reduces the removal efficiency [31,32]. Current density is one of the most effective parameters that affects the kinetic rate and removal efficiency. Increasing the current density more than the optimal value causes more electrical energy consumption and heat generation and consequently adverse effects on the removal process, so its value must be optimized. Due to the importance and influence of current density in the EF process, the effect of applied current density was studied in the range of 3-10 mA cm -2 . The removal of MB dye depended on the current density. The increase in current density from 2 to 5.5 mA cm -2 offered faster MB dye degradation (MB removal efficiency of 47.5% to 66.9%). Increasing the dye removal efficiency with increasing current density can be attributed to the fact that increasing the current density increases the amount of iron generated by the anode electrode. The increase in removal and degradation rates from 3 mA cm -2 to the optimum value of 7.0 mA cm -2 can be related to the acceleration of Fenton's Reaction (Equation1). Therefore, increasing the concentration of Fe 2+ produced could lead to an increase in • OH generated through (Equation 1). These • OH generated reacts immediately with the MB dye, increasing the MB removal efficiency. However, more than the optimal value of current density leads to a decrease in the MB removal efficiency in the EF reactor. This negative effect is due to the role of Fe 2+ as the scavenger of • OH (Equation 10) [32,33]. Hence, 7.0 mA cm -2 was selected as the optimal value of current density.
(Eq. 10) The effect of electrolysis time in the range of 2-15 min on MB removal was studied (Fig. 4)

Optimization process
The optimization process was carried out over the upper to lower limits of each operating parameter's value to determine the best treatment condition for maximum MB removal at a reasonable cost. The highest MB removal efficiency of 98.3% was predicted at the optimum condition of initial MB concentration of 20 mg L -1 , H 2 O 2 dosage of 400 μL, the current density of 7.0 mA cm -2 , and electrolysis time of 10 min, which was in good accordance with the obtained experimental MB removal efficiency.

Kinetic model of hydroxide radical assisted EF process
The kinetic studies of the treatment process revealed that hydroxyl radical reaction with MB contaminant follows the pseudo-first-order reaction, which indicates that • OH directly attacked the contaminant molecules. The hydroxyl radicals are generated and consumed continuously at a similar rate to provide a steady-state concentration of active radicals throughout the treatment process. The apparent constant rate of K app could be evaluated by the pseudo-first-order reaction equation and corresponds to the slope of demonstrated plot in Figure 5, with the value of 0.3753 min -1 [35].

Conclusion
Herein, to achieve an economical treatment process for the efficient removal of MB, the response surface methodology was employed. The analysis of variance (ANOVA) at the confidence level of 95% was carried out to evaluate the significance of the independent variables and their interactions. The obtained results revealed that the initial MB concentration and H 2 O 2 dosage affected the removal efficiency significantly. The obtained experimental removal efficiency of 95.8% by the EF process was in satisfactory agreement with the predicted removal efficiency of 98.3% by the