Accelerated degradation of groundwater-containing malathion using persulfate activated magnetic Fe₃O₄/graphene oxide nanocomposite for advanced water treatment

2.1 Preparation and characterization of Fe₃O₄-GO MNCs

2.2 Experimental procedures

A 300 mL sealed Erlenmeyer flask with a rotary shaker (200 rpm, 20 ± 2 °C) was used in all experiments. The reactions involved were initiated by adding the indicated amount of Fe₃O₄-GO-MNCs to 250 mL of a mixed solution containing MT and PS. Moreover, to determine the effect of different parameters on MT degradation, several sets of experiments were performed. The solution pH was adjusted operating NaOH or H₂SO₄. To identify the influence of the reactive species in PS activated by Fe₃O₄-GO MNCs process (PS/ Fe₃O₄-GO MNCs), ethanol (EtOH) and tertiary butanol (TBA) were utilized as radical scavengers. Also, to verify the ability to reuse Fe₃O₄-GO MNCs, the MNCs were collected with a magnet and washed with Milli-Q water after each run, followed by reuse with the addition of fresh PS and MT. All experiments were performed in triplicate, and the standard deviation of all experiments was given. Collect 2 mL samples at regular intervals and pass through a 0.45 µm PTFE filter. Then 10 µL of 1 M Na₂S₂O₃ was added to quench the reaction. The filtered solution was analysed for MT concentration.

The detection of MT residual in samples after the treatment process were further validated by using High-performance liquid chromatography (HPLC) (KNAUER HPLC system model Smartline, Germany) having a UV detector with 220 nm wavelength. For this purpose, a Reliant C18 column (250 × 4.6 mm) was used and maintained at a constant column temperature of 40 °C. A solution of acetonitrile and water (75:25) was used as eluent at a flow rate of 1 mL min⁻¹ with a retention time of 25 min. The structure and morphology of the Fe₃O₄-GO MNCs were characterized by Field Emission scanning electron microscopy (FE-SEM, LEO-Germany).

2.3 Statistical model

Response surface methodology (RSM) stands for statistically identifying important factors (independent variables) in a model and generating a response surface model, which is used to predict outcomes by providing mathematical equations. In this study, experiments were designed using Design Expert version 11 software.

For this purpose, four independent variables are considered in the experimental design as follows concentration of MT, the dosage of PS, the dosage of Fe₃O₄-GO MNCs and reaction time. The dependent variable (response) is the degradation efficiency of MT. Each of the independent variable (X₁, X₂, X₃, and X₄) was varied numerically over five levels and coded as -α,-1, 0, +1 and + α Analysis of variance (ANOVA) and regression analysis were performed to determine the statistical significance of the model conditions, with regression correlation fit between experimental data and independent variables. Assuming that the changes in Y, (MT degradation efficiency) are consistent with a second order equation of the form Eq.1:

Y denotes the response value estimated by the model, β₀ is a constant, β_i denotes that the shift value is linear, β_ii is quadratic, and β_ij is the regression coefficient of correlation. X_i and X_j denote the coded independent variables. The validity of the model was determined by model summary, no-fit model test, and coefficient of determination (R²) analysis The terms were removed from the initial model for all statistically insignificant terms (p-value > 0.05), and the empirical data were re-fitted only for significant (p-value < 0.05) result variables to obtain the final reduced model.

3.1 Characterization of Fe₃O₄-GO MNCs catalyst

For the purpose of investigating the morphology and structure of the catalyst, there was FE-SEM image taken for the fabricated Fe₃O₄-GO MNCs and shown in Fig. 1-(A). In addition, GO distributed between Fe₃O₄ nanoparticles can prevent the aggregation of Fe₃O₄ nanoparticles to some extent resulting in a potentially beneficial effect on the reaction Even more relevant, after a long sonication time during FE-SEM sample preparation, the Fe₃O₄ nanoparticles were still tightly attached on the GO surface with high density, suggesting a strong interaction between Fe₃O₄ nanoparticles and GO. Also, it shows that the nanoparticles are agglomerated due to magnetic dipolar interactions between the magnetite nanoparticles.

Close

Fig. 1

(A) The FE-SEM image, (B) XRD pattern, and (C) FTIR spectra of Fe₃O₄-GO MNCs.

Export to PPT

The crystalline structure of Fe₃O₄-GO MNCs was identified with XRD technique. The XRD pattern of Fe₃O₄-GO MNCs is shown in Fig. 1-(B). The peaks at 2θ values of 17.9° (1 1 1), 30.7° (2 2 0), 35.9° (3 1 1), 42.8° (4 0 0), 57.3^° (4 2 2), and 63.1^° (5 1 1) are in agreement with standard XRD data for an inverted Fe₃O₄ spinel structure with lattice constants of θ = 8.397 ˚A (JCPDS file no. 19–0629). Computations with the Scherrer equation reveal an 18 nm equivalent particle size, as confirmed by FE-SEM. Also, the Fe₃O₄ diffraction peaks are wider, suggesting that the crystal sizes are quite small for Fe₃O₄ particles. Moreover, a diffraction peak marked by a very slight oval shape at 2θ = 9.7^° belongs to (0 0 1) crystal of GO. The obtained results correlate well with the previous studies (Kassaee et al., 2011; Yao et al., 2012).

In Fig. 1-(C), the fourier transform infrared spectroscopy (FTIR) spectra of Fe₃O₄-GO MNCs have been provided. The intense peaks at 3427 cm⁻¹ is associated with the stretching of O—H groups. The peak at 1638 cm⁻¹ (aromatic C⚌C) can be attributed to the skeletal vibrations of the unoxidized graphite domains (Kassaee et al., 2011). The peak at 1363 cm⁻¹ relates to the bending absorption of carboxyl group O⚌C (Amani et al., 2022), while the two characteristic absorption peaks of Fe-O in Fe₃O₄ nanoparticles appear at 572 cm⁻¹ and 628 cm⁻¹. The characteristic peaks at 1571 and 1251 cm⁻¹ indicate the interaction of carbonyl and epoxy groups of GO and Fe on the surface of magnetic particles, and establish the binding of Fe₃O₄ nanoparticles to GO (Umar et al., 2020).

3.2 Establishment for a new MT degradation activation system

Initially, the effects of numerous PSs including (NH₄)₂S₂O₈, K₂S₂O₈ and Na₂S₂O₈ on MT degradation efficiency were checked, and no noticeable differences were observed between these PSs; hence, K₂S₂O₈ was selected for further study. Prior to an extensive study on the MT degradation yield using PS/ Fe₃O₄-GO MNCs, the MT degradation yield using PS and Fe₃O₄-GO MNCs was investigated independently in control experiments. When ${\mathrm{S}}_2{\mathrm{O}}_8^{2-}$ or Fe₃O₄-GO MNCs were added lonely, the MT degradation yields were 2.6 % and 21 %, respectively, both of these cases were insignificant compared to the almost complete degradation yield of MT in the presence of both ${\mathrm{S}}_2{\mathrm{O}}_8^{2-}$ and Fe₃O₄-GO MNCs. As a result, the significant MT degradation efficiency of PS/Fe₃O₄-GO MNCs may be due to the activation of ${\mathrm{S}}_2{\mathrm{O}}_8^{2-}$ by Fe²⁺ dissolved from Fe₃O₄-GO MNCs (Shang et al., 2019; Zhen et al., 2018).

3.3 Developed model and effect of operational parameters

Aqueous solution pH has a major impact on oxidation processes through changes in the degree of ionization of impurities and oxidants, surface properties, and the degree of activity and solubility of reactants. Therefore, it is based on the important role of solution pH on degradation efficiency in chemical reactions, among them advanced oxidation, the efficiency of PS/Fe₃O₄-GO MNCs process for MT degradation yield from groundwater was studied at a pH range of 2–12 and the results are given in Fig. 2.

Close

Fig. 2

Effect of pH solution for degradation yield of MT using PS/Fe₃O₄-GO MNCs process: MT concentration of 5.0 mg/L, PS dosage of 2 mM, pH of 4, and Fe₃O₄-GO MNCs dosage of 150 mg/L.

Export to PPT

The degradation efficiency of MT showed an increasing trend from 26.7 % to 90.3 % when the pH was downregulated from 12 to 4. This result indicates that an acidic environment is more favourable for MT degradation than neutral and alkaline conditions. Any change in solution pH, whether below or above 4, leads to a decrease in MT degradation efficiency. Hence, the solution pH of 4 was preferred as the optimum solution pH and was applied in further experiments. If the pH is very low (strongly acidic), a large number of hydrogen ions (H⁺) are present in solution, allowing them to scavenge both sulphate and hydroxyl radicals. as shown in Eqs. (2) and (3).

${\mathit{SO}}_4^{-\bullet }$ mostly occurs in acidic and neutral conditions, and ${\mathrm{O}\mathrm{H}}^{\bullet }$ mostly is found in high pH conditions. Moreover, ${\mathrm{O}\mathrm{H}}^{\bullet }$ radicals can be generated from the outcomes of ${\mathit{SO}}_4^{-\bullet }$ with ${\mathrm{H}}_2\mathrm{O}$ / ${\mathrm{O}\mathrm{H}}^{-}$ (Eqs. (4) and (5)):

However, the radical ${\mathrm{S}\mathrm{O}}_4^{-\bullet }$ dominates at acidic pH, but both ${\mathrm{S}\mathrm{O}}_4^{-\bullet }$ and ${\mathrm{O}\mathrm{H}}^{\bullet }$ radicals are present, and in alkaline pH and near to 12, ${\mathrm{O}\mathrm{H}}^{\bullet }$ is the dominant radical. Consequently, when PS/Fe₃O₄-GO MNCs is treated as an oxidant, corresponding to Eq. (4), including ${\mathrm{S}\mathrm{O}}_4^{-\bullet }$ , ${\mathrm{O}\mathrm{H}}^{\bullet }$ is also generated. Hence, the yield decreases at high pH (Eq. (6)) due to the reaction of ${\mathrm{S}\mathrm{O}}_4^{-\bullet }$ , a radical by ${\mathrm{O}\mathrm{H}}^{\bullet }$ , which leads to higher consumption of ${\mathrm{S}\mathrm{O}}_4^{-\bullet }$ , and subsequent decrease in MT degradation yield (Liang et al., 2020).

Furthermore, the reactions described in Eqs. (7) and (8) can also follow in the company of ${\mathrm{O}\mathrm{H}}^{\bullet }$ radicals at pH levels of 8 to 10 and 12, respectively. The reaction of ${\mathrm{O}\mathrm{H}}^{\bullet }$ with ${\mathrm{O}\mathrm{H}}^{\bullet }$ may result in rapid loss of ${\mathrm{O}\mathrm{H}}^{\bullet }$ in basic alkaline solution. Therefore, the decomposition efficiency of MT under alkaline conditions will be less as compared to acidic and neutral conditions. The ${\mathrm{O}\mathrm{H}}^{\bullet }$ redox potential also becomes lower under alkaline conditions (about 2.0 V), leading to a decrease in the oxidation potential of ${\mathrm{O}\mathrm{H}}^{\bullet }$ radicals (Norzaee et al., 2018).

For the optimization purpose, an experimental study on the MT degradation performance using PS/Fe₃O₄-GO MNCs process was conducted. To design the experiment, response surface methodology (RSM) was applied. Therefore, the four variables studied were MT concentration (X₁), PS dosage (X₂), Fe₃O₄-GO MNCs dosage (X₃) and reaction time (X₄). These variables and the range studied have been selected based on the literature and the results obtained in preliminary studies (Table 1). Finally, this model was designated on the basis of the high ranking order polynomials in which additional conditions were significant. Based on the MT degradation yield data shown in Table 1, a quadratic model was made using RSM since it was statistically substantial for the response (Y) for MT degradation efficiency.

In terms of coding factors, the empirical RSM model represents the interaction and importance of the variables in the direction of the response. One-factor coefficients represent the effect of only one factor, while two-factor and second-order element coefficients correspond to interactions or quadratic effects between two factors. Therefore, the final empirical formula model of the response according to the coding factor is given by Eq. (9).

The correlation coefficient, R² value, was significant to validate the elaborated model. The R² values for degradation yield of MT is 0.9819. It shows an overall change of 98.19 in the correlation of MT mining production between experimental and predicted values. These high R² values indicate that the predicted response was near the experimental value and the model is proper to connect with the experiment data. Therefore, the R² shows good agreement between experimental data. Fig. 3-(A) and 3-(B) presents the predicted response versus actual response and the normal % probability versus externally studentized residuals for degradation yield of MT using PS/Fe₃O₄-GO MNCs process, respectively. The residuals can be seen to exhibit a normal distribution since all points were along a straight line. The residuals also revealed that the model could not be further improved through changes in the response because the data points were scattered and did not display a “sigmoidal shape” curve. However, the graphs indicated that the model presented in Equation (9) can be considered as the best suggested model for the obtained experimental data to find the optimal degradation performance of MT. Furthermore, low standard deviation values of 2.46 was obtained for the model.

Close

Fig. 3

(A) the predicted response versus actual response (B) the normal % probability versus externally studentized residuals for degradation of MT using PS/Fe₃O₄-GO MNCs process.

Export to PPT

The significance and contribution of independent variables and interactions were determined using graphical Pareto analysis to explain the results. This analysis determines the percentage effect of each independent variable on the dependent variable (MT degradation), based on Equation (10).

Close

Fig. 4

(A) Graphical Pareto analysis of the significant operating parameters effects, (B) Effect of MT concentration on MT degradation yield, (C) Response surface plot of MT concentration & PS dosage, (D) Response surface plot of Fe₃O₄-GO MNCs dosage & Reaction time. (Operating parameters set at their central points of 3.7 mg/L MT concentration, PS dosage of 1.75 mM, pH of 4, Fe₃O₄-GO MNCs dosage of 175 mg/L and Reaction time of 17.5 min).

Export to PPT

The study of degradation yield of MT from groundwater using PS/Fe₃O₄-GO MNCs procedure at different MT concentrations was conducted in the range of 1–10 mg/L. According to Fig. 4-(B), there was a significant decrease in its degradation efficiency in groundwater as the MT concentration increased. In fact, an increase in MT concentration from 1 to 20 mg/L caused reduction of efficiency of the process from 94.7 to 60.8 %. The reduction in MT degradation efficiency at higher concentrations may be due to inhibition of PS reactions at redox-active centres, reduced action of PS and Fe₃O₄-GO MNCs, leading to reduced formation of reactive species, and competition between intermediate and parent compounds in reacting with oxygen reactive radicals.

In fact, since the ${\mathrm{S}\mathrm{O}}_4^{2-}$ radicals appear to be inert and would not be considered a type of pollutant, it should only be limited to a maximum concentration of 250 mg/L, based on the water's taste and odour as determined by the United Protection Agency. In addition to the cost of PS, it is preferable to select the lowest concentration of PS that can result in MT degradation within a given reaction time. The effect of PS concentrations (0.5 to 3 mM) on MT degradation yield was investigated (Fig. 4-(C)). As can be observed in Fig. 4-(C), the degradation yield of MT increased with growing PS concentration. When the concentration of PS increased from 0.5 to 2 mM, the degradation yield of MT increased from 30.4 % to 86.7 %. However, when the PS concentration is higher than 2 mM, the degradation yield of MT is slightly reduced.

The reason may be that ${\mathrm{S}\mathrm{O}}_4^{-\bullet }$ is converted into ${\mathrm{S}}_2{\mathrm{O}}_8^{-\bullet }$ through Eqs. (11) and (12), while the oxidation potential of ${\mathrm{S}}_2{\mathrm{O}}_8^{-\bullet }$ is lower than that of ${{\mathrm{S}\mathrm{O}}_4}^{-\bullet }$ .

Therefore, the optimal PS concentration for degradation yield of MT was found to be 2 mM in the present study.

The effect of Fe₃O₄-GO MNCs dosage on PS on MT degradation performance was investigated. Fig. 4-(D) shows the MT degradation yield as a function of PS concentration and Fe₃O₄-GO MNCs dosage in solution.

As can be seen from Fig. 4-(D), an enhancement of MT degradation yield is observed by increasing Fe₃O₄-GO MNCs dosage from 50 to 200 mg/L had a beneficial effect on the MT degradation efficiency. The reason for this is the increased number of active sites in Fe₃O₄-GO MNCs to react with PS, yielding more reactive sites of ${\mathrm{S}\mathrm{O}}_4^{-\bullet }$ radicals (Eq.13).

However, it is found that as the dosage of Fe₃O₄-GO MNCs had increased from 200 to 300 mg/L, MT degradation yield decreased. However, when the Fe₃O₄-GO MNCs dosage is higher than 200 mg/L, the degradation yield of MT is decreased. A possible explanation for this is (1) elevated ${\mathrm{S}\mathrm{O}}_4^{-\bullet }$ species quenching resulting from the association of ${\mathrm{S}\mathrm{O}}_4^{-\bullet }$ species alone triggered in excess of activator, and (2) increased quenching of ${\mathrm{S}\mathrm{O}}_4^{-\bullet }$ with Fe²⁺ at high activator dose through Eq.14.

Thereupon, the MT degradation yield decreased when the Fe₃O₄-GO MNCs was added further beyond about 200 mg/L. For the control experiment, only 13.5 % MT loss was observed in the absence of Fe₃O₄-GO MNCs, indicating a low oxidation capacity of PS without using an activator.

3.4 Optimization and validation of model

With the objective to maximize the MT degradation yield using PS/Fe₃O₄-GO MNCs process, the variation of operating parameters values presents by the ramps demonstrated in Fig. 5-(A). The desirability of each factor and response is shown, as well as the mixed desirability. Highlighted points illustrate optimal values for each factor or response (moving points horizontally) and how well those goals are achieved (how high the slope is). The complexity desirability of the model is 0.940, remember that the goal of optimization is to find a global solution that satisfies all objectives, desirability is just a mathematical property, not every desirability will reach a value of 1.0. A series of experiments were conducted to test the model to assess the reliability of the optimal conditions predicted by the empirical model. The experimental values were found to deviate less from the predicted values, and these results further support the validity of the model.

Close

Fig. 5

(A) Ramps of numerical optimization of operating parameters values – (B) MT degradation, COD and TOC removal in the optimum condition.

Export to PPT

The optimum conditions predicted for MT degradation yield using PS/Fe₃O₄-GO MNCs process are as the MT concentration of 3.7 mg/L, PS dosage of 2.3 mM, Fe₃O₄-GO MNCs dosage of 224 mg/L, reaction time of 22 min, and pH solution of 4.0. According to the obtained results, the predicted and experimental degradation yield of 95.9 % and 96.5 % achieved in the mentioned optimum condition.

Furthermore, the accuracy of the developed polynomial model was successfully validated by performing different processing runs under a range of operating conditions and is summarized in Table 3. From this, it can be concluded that the experimentally obtained MT degradation yields are reasonably consistent with the yields predicted by the RSM model.

3.5 Degradation kinetic

In the present work, first and second order kinetic model was used to investigate the MT degradation performance using PS/Fe₃O₄-GO MNCs process. The kinetic studies of MT degradation yield were performed in optimum condition. From the kinetic results obtained (Table 4), the correlation coefficient (R²) of the model used was approximately 0.9965, confirming the high ability of model to fit the kinetic data of MT degradation by PS/Fe₃O₄-GO MNCs process. This is also confirmed by the strong agreement between the first-order reaction kinetics model and the experimental data from the studied process.

3.6 Mechanism of MT degradation

3.6.1

3.6.1 Radical scavengers effect

In addition, in order to better explain the MT degradation mechanism and to identify the dominant radicals in the degradation process of PS/Fe₃O₄-GO MNCs, radical quenching experiments were performed, the results of which are shown in Fig. 6- (A). From Fig. 6-(A), it is clearly seen that there was a significant decrease in MT degradation efficiency in the presence of ethanol (EtOH), benzoquinone (BQ), and tertiary butanol (TBA). However, the degradation efficiency was clearly inhibited when TBA, EtOH and BQ were added to the solution separately, and the degradation efficiencies were only 74.1 %, 27.9 % and 30.1 % for BQ, EtOH and TBA, respectively. BQ, EtOH, and TBA are established to determine the impact of ${\mathrm{O}\mathrm{H}}^{\bullet }$ and ${\mathrm{S}\mathrm{O}}_4^{-\bullet }$ radicals at different levels. The second-order rate constants given for the quenching reactions of EtOH and TBA are as follows (Li et al., 2018):

(15)

$$\mathrm{T}\mathrm{B}\mathrm{A}+{\mathrm{O}\mathrm{H}}^{\bullet }3.8-7.6\times {10}^9{\mathrm{M}}^{-1}{\mathrm{S}}^{-1}$$

(16)

$$\mathrm{T}\mathrm{B}\mathrm{A}+{\mathrm{S}\mathrm{O}}_4^{-\bullet }4.0-9.1\times {10}^5{\mathrm{M}}^{-1}{\mathrm{S}}^{-1}$$

(17)

$$\mathrm{E}\mathrm{t}\mathrm{O}\mathrm{H}+{\mathrm{O}\mathrm{H}}^{\bullet }1.2-2.8\times {10}^9{\mathrm{M}}^{-1}{\mathrm{S}}^{-1}$$

(18)

$$\mathrm{E}\mathrm{t}\mathrm{O}\mathrm{H}+{\mathrm{S}\mathrm{O}}_4^{-\bullet }1.6-7.7\times {10}^7{\mathrm{M}}^{-1}{\mathrm{S}}^{-1}$$

Close

Fig. 6

(A) MT degradation yield in the existence of different scavengers, (B) Effect of inorganic ions (Inorganic ions concentration of 0.2 mM), (C) MT degradation in five cycles in the optimum condition.

Export to PPT

Basically, ${\mathrm{O}\mathrm{H}}^{-\bullet }$ interacts with targets via hydrogen abstraction, but the ${\mathrm{S}\mathrm{O}}_4^{-\bullet }$ response is a precise electron transfer mechanism upon activation.. Hence, ${\mathrm{S}\mathrm{O}}_4^{-\bullet }$ is more selective, which is much less of an attractor for TBA. Based on the equations above, it can be observed that ${\mathrm{O}\mathrm{H}}^{\bullet }$ appears to be almost 1000 times more reactive than ${\mathrm{S}\mathrm{O}}_4^{-\bullet }$ in the reaction rates toward TBA.

3.7 Reusability of Fe₃O₄-GO MNCs

The reusability potential of the PS/Fe₃O₄-GO MNCs system was investigated in order to determine the practical applicability of the Fe₃O₄-GO MNCs. For this aim, the prepared Fe₃O₄-GO MNCs were used in five subsequent tests and the removal efficiency was recorded. As depicted in Fig. 6-(C), around 13 % decreases in the removal efficiency of MT achieved by the end of the fifth experiment, which demonstrates the appropriate reusability of the Fe₃O₄-GO MNCs in consecutive experiments. This reduction is explained by poisoning the Fe₃O₄-GO MNCs surface with organic by-products or reducing active sites on the catalyst surface.

3.8 Degradation pathway

The proposed degradation pathway of MT demonstrated in Fig. 7 (Li et al., 2019).

Close

Fig. 7

The proposed degradation pathway of MT demonstrated.

Export to PPT

3.9 Real sample

The physical and chemical properties of the groundwater samples are shown in Table 5. Treatment trials were performed under most favorable conditions acquired from batch runs. A 100 mL groundwater sample was employed for examination.

The amount of MT remaining in the groundwater samples in this study was determined by solvent extraction experiments with dichloromethane (dichloromethane), which is denser than water (1.33 g mL⁻¹). After filtering the final sample, the solution was transferred to a 1 L decanter, followed by three additional extraction steps using 25, 50, and 25 cm³ of solvent dichloromethane, respectively. At each stage, transfer the remaining MT to the solvent phase by shaking the decanter. The solvent phase is then drained into a balloon placed underneath. Once the three extraction stages were completed and the solvent phase was completely removed, the indicated amount of anhydrous sodium sulphate was added to the sample to remove water. Thereafter, the separated organic phase was stored in a fume hood and evaporated to fully recover the extraction solvent. After evaporation, a sample volume (500 µL) was set up for analysis on the HPLC system (Dehghani et al., 2017). From the obtained results, it was found that the initial MT concentration of 0.93 ± 0.23 µg.L^-1 in the groundwater samples decreased to zero under optimal conditions after applying the PS/Fe₃O₄-GO-MNCs method.