Application of Sono–photo-Fenton process for degradation of phenol derivatives in petrochemical wastewater using full factorial design of experiment

Abstract

In this study, the degradation and mineralization of petrochemical wastewater containing phenol and nitrophenol with chemical oxygen demand (COD) at 260 mg/L and total organic carbon (TOC) at 100 mg/L, were investigated by Sono–photo-Fenton process. The full factorial design of experiment was used to explore the influence of operational variables such as pH, initial concentration of hydrogen peroxide and Ferrous ions on the removal of COD. The analysis of variance (ANOVA) showed a high determination coefficient value (R² = 0.9884, $R_{\text{pred}}^{ 2}$ = 0.9713, $R_{\text{adj}}^{ 2}$ = 0.9822) for removal of COD and satisfactory prediction second-order regression model. The graphical response surface plots were employed to determine the optimum conditions. The operational variables were optimized by the model as follows: the [Fe²⁺] = 24 mg/L, pH at 3, and [H₂O₂] = 500 mg/L. The results showed that at the predicted optimum conditions and after 60 min of reaction, the degradation (removal of COD) and mineralization (removal of TOC) were 76.1 and 68.7%, respectively.

Introduction

The wastewater produced from Karoon petrochemical plant in Iran contains ortho-toluidine, nitrobenzene, nitrocresol, phenol derivatives and other aromatic compounds. Definite amounts of aromatic components are lost during a process which holds a wide range of non-biodegradable chemicals that cause environmental pollution. The phenol derivatives are toxic and bio-refractory pollutant in petrochemical wastewater which can be a source of major harm to the environment and human health [1]. Based on the environmental protection agency, the permissible limit of nitrophenols in aqueous solution is 1 mg/L. The remediation of wastewaters polluted with nitrophenols by traditional methods is really difficult, has high operational costs, secondary contamination and a long reaction time; subsequently, phenol derivatives are resistant according to their high solubility and stability in water [2]. Therefore, employing new methods for the treatment of the wastewater containing these pollutants without these problems is essential.

Advanced oxidation processes (AOPs) are effective and environmentally friendly methods that can degrade organic pollutants that are resistant to the conventional treatment methods into simple byproducts and finally mineralize them into carbon dioxide and water [3]. The reactive and nonspecific oxidant, hydroxyl radicals, with high electrochemical oxidation potential were produced by AOPs [4]. An aqueous solution containing benzoic acid was treated by electrocoagulation (EC) and electrochemical Fenton (EF) processes by Sandhwar and Prasad [5]. In the previous work, the treatment of spent caustic in the wastewater of petrochemical industries was investigated by ozonation in alkaline media followed by electrocoagulation process [6]. High-Performance Nanocatalyst for Adsorptive and Photo-assisted Fenton-Like Degradation of Phenol was studied by Gazi et al. [7].

Several studies have been done on combined AOPs that can increase the production of more hydroxyl radicals for degradation of resistant organic contaminants [8, 9]. The ultrasonic and Fenton processes are AOPs that can be suitable treatment methods which have been employed for the removal of various pollutants in the wastewater [10, 11]. The addition of ultrasonic irradiation to the Fenton and photo-Fenton process can be a good choice to overcome the restrictions of mentioned processes. The Sonochemistry includes the chemical effects of producing ultrasound when a sound wave is passed through an aqueous medium. Frequently, as a moment of the hemolytic cleavage of water in the cavitation bubbles, the reactive radical species such as hydroxyl radicals and hydrogen atoms were formed to promote the degradation of pollutant [12]. The use of ultrasonic irradiation can produce vibrational wave energy and shear stresses at the cavitation interface and produce locals with high pressure and temperature; this feature can lead to the improvement of Fenton and photo-Fenton reactions [13]. Thus, the combination of ultrasonic with photo-Fenton process has been employed to increase the efficiency of the mentioned process [14].

In this project, full factorial design(FFD)was used in planning experiments as a statistical technique where several factors are controlled and their effects on each other are explored at three levels [15]. The ability of general full factorial design in the modeling of complex systems makes it a more practical technique than other traditional methods for modeling a multi-variable system [16].

In this work, the effect of operational variables such as pH, initial concentration of ferrous ions and hydrogen peroxide was investigated on the degradation of phenol derivatives by Sono–photo-Fenton process. The important variable and interaction effects between them have been studied by FFD method. The removal of COD (%) was the optimized response and the correlation between the response and significant factors was determined.

Experimental

Material

The diluted wastewater was the effluent of an Iranian petrochemical Company, which was sampled on Mars 2017 with the following properties (Table 1). The features of the real wastewater were not similar in different times and the amounts of factors were not completely fixed. The composition of the real wastewater is highly variable and depends on the process condition. All studied samples were taken at one time and one point and the feedstock samples were kept in the laboratory in a cool place, which were then homogenized and tested; therefore, the amounts of factors in Table 1 were reported fixed.

Table 1 Characteristic of diluted petrochemical wastewater

The ferrous sulfate heptahydrate (FeSO₄·7H₂O) as the source of Fe(II), hydrogen peroxide solution (30% w/w), H₂SO₄ and NaOH are all provided by Merck Company of Germany.

General procedure

The experiments were performed in a glass cylindrical reactor with 1 L of capacity. The light source was a mercury lamp, Philips 15 W (UV-C) at 254 nm, which was positioned horizontally above the reactor. The ultrasonic irradiation was coupled with a transducer in the reactor at a frequency of 20 kHz and equipped with an ultrasonic generator (Chrom Technol. UP-800 Model, 800 W output, and variable power control).

The reactor was prepared with a jacket of water and external circulating flow by a thermostat for regulating temperature at 25 °C. A pH meter, PT-10P Sartorius Instrument from Germany Company was used to adjust the initial pH of the solution.

An aqueous solution containing 10% of sodium sulfite was used to quench the reactions. The wastewater was evaporated and the solid residue was weighted for measuring the dissolved solid particles or total dissolved solids (TDS). The degradation and mineralization of the studied wastewater by Sono–photo-Fenton process were estimated by chemical oxygen demand (COD) and total organic carbon (TOC) tests, respectively. The TOC was tested by Shimadzu TOC-VCSN, equipped with an automatic auto-sampler and platinum-based catalyst. The synthetic air at the rate of 0.15 L min⁻¹ acts as a carrier gas. Before sending the samples to TOC analyzer, the pH was acidified to lower than 4, for releasing the inorganic carbon as CO₂ from the solution. The COD was measured by the HACH’s COD method using a COD reactor from HACH Company and a direct monitoring spectrophotometer (DR/5000U) at 620 nm according to the standard methods [17]. The studied wastewater was obtained from a petrochemical plant in Iran; after primary filtration, it was diluted with distilled water and after dilution, it holds 100 mg/L of nitrophenol and 50 mg/L of phenol. The initial COD and TOC of the wastewater were measured at 270 and 100 mg/L, respectively. The removal percent for COD and TOC was achieved according to the following equations (Eqs. 1, 2).

$$ {\text{Removal of COD}}\,\,\left( {\text{\% }} \right) = \left( {\frac{{[{\text{COD}}]_{0} - [{\text{COD}}]}}{{[{\text{COD}}]_{0} }}} \right) \times 100, $$

(1)

$$ {\text{Removal of }}\,\,{\text{TOC}}\,\, \left( \% \right) = \left( {\frac{{[{\text{TOC }}]_{0} - [{\text{TOC}}]}}{{[{\text{TOC}}]_{0} }}} \right) \times 100, $$

(2)

where $\left[ {\text{TOC}} \right]_{0}$ and $\left[ {\text{COD}} \right]_{0}$ are the concentration of TOC and amount of COD at the start of the reaction and $[ {\text{TOC]}}$ and $[ {\text{COD]}}$ are their quantity at time $t$.

Experimental design

The full factorial design method was used with three independent factors including the concentration of hydrogen peroxide ($C_{\text{HP}}$), ferrous ion ($C_{\text{F}}$), and pH. The experimental design method was applied and the removal percentage of COD was selected as response to get the optimum conditions. Based on the previous studies and preliminary experiments, the ranges of parameters for BBD were selected. The input variables and their levels in the experiment are presented in Table 2.

Table 2 The range and levels of the factors

The general full factorial is a comprehensive design of experiment that all levels of each variable in a test are combined with all levels of every other variable.

In this work, 27 experiments were run wholly, including combination of all levels of other factors. All factorial designs were performed randomly to lessen experimental errors. Then, the F-test analysis of variance with a 95% confidence interval was employed to assess the statistical effect of the significant variables and their interactions.

Data analysis

According to the experimental design, the following model was fitted to the response variable (Y) in the form of a polynomial equation (Eq. 3):

$$ Y = b_{0} + \mathop \sum \nolimits b_{i} x_{i} + \mathop \sum \nolimits \mathop \sum \nolimits b_{ij} x_{i} x_{j} + \mathop \sum \nolimits \mathop \sum \nolimits b_{ii} x_{i}^{2} + \varepsilon , $$

(3)

where $\varepsilon$ is the residual term, $b_{0}$ is a constant, $b_{ij}$ is the linear interaction effect between the input variables, $x_{i}$ and $x_{j}$ ($i = 1, 2\,\, {\text{and}} \,\,3; \,\,\,j = 1, \,\,2\,\, {\text{and}} 3$) are independent variables, $b_{i}$ is the slope of the variable, $b_{ii}$ is the second order of input variable ($x_{i}$). The ANOVA was employed to inspect the significance of each term in the polynomial equation [18]. The MINITAB 17 was used to determine the coefficients of Eq. (3) with Response Surface Methodology (RSM). The experimental design involves 27 runs and the natural values of these factors; the experimental and predicted response values for the removal of COD are displayed in Table 3. In all runs, the time of reaction was 60 min.

Table 3 The 3-factor full factorial design matrix and the response function [removal of COD (%) in 60 min of reaction]

Results and discussion

Statistical analysis

The ANOVA was used to determine the significant effect of the main factors and their interactions on the removal of COD (%). The P value, F ratio, sum of squares and mean square of each factor are presented in Table 3. The importance of the data is refereed by its P value, with values nearer to zero meaning more significance. The P value should be less than or equal to 0.05 to study statistical significance for 95% confidence level. The aims of this section were to define the optimum condition for maximum removal of COD. The steps of the FFD were investigated by many researchers [19, 20].

The correctness of the model is verified in Fig. 1, which compares the experimental values in contradiction of the predicted responses of the model in the removal of COD. These results showed a good agreement between predicted and experimental values. It was detected that the predicted response from the model is in coincidence with the experimental data.

Fig. 1

Comparing the experimental and predicted value for the removal of COD in Sono–photo-Fenton process

ANOVA tests for the removal of COD in Sono–photo-Fenton process

In this study, the effect of three independent variables on the response function was investigated by the FFD and RSM, to obtain the optimum conditions. The mathematical relation between the response and three important variables can be estimated by a quadratic polynomial equation [21]. An empirical relation between the response (removal of COD) and independent variables was obtained and expressed by the following second-order polynomial equations (Eq. 4):

$$ \begin{aligned} {\text{Removal }}\,\,{\text{of }}\,\,{\text{COD }}\,(\% )\, & \, = \, 39.18 + 0.0978X_{{C_{\text{HP}} }} - 2.65X_{{C_{\text{PH}} }} + 1.568X_{{C_{\text{F}} }} - 0.00016X_{{C_{\text{HP}} }}^{2} \\ &\quad - 0.3011X_{{C_{\text{PH}} }}^{2} - 0.02812X_{{C_{\text{F}} }}^{2} + 0.002272X_{{C_{\text{PH}} }} X_{{C_{\text{HP}} }} - 0.000097X_{{C_{\text{HP}} }} X_{{C_{\text{F}} }} - 0.0068 X_{{C_{\text{PH}} }} X_{{C_{\text{F}} }} \\ \end{aligned} $$

(4)

The statistical properties of selected significant model terms are presented in Table 4 to describe the removal efficiency of COD as a function of the studied variables.

Table 4 ANOVA tests for quadratic models in the removal of COD by Sono–photo-Fenton process

According to the analysis, it was obtained that $X_{{C_{\text{HP}} }} , X_{{C_{\text{PH}} }}$, $X_{{C_{\text{F}} }}$,$X_{\text{PH}}^{2}$, $X_{\text{F}}^{2}$, $X_{\text{HP}}^{2}$, and $X_{{C_{\text{HP}} }} X_{{C_{\text{PH}} }}$ were significant terms and other model terms ($X_{{C_{\text{PH}} }} X_{{C_{\text{F}} }}$ and $X_{{C_{\text{HP}} }} X_{{C_{\text{F}} }}$.) are not significant (with a probability value larger than 0.05). These results showed a good agreement between experimental and predicted values for degradation efficiencies. As can be seen, the important reaction parameters were: direct effect of pH (− 2.65) > direct effect of Ferrous (1.568) > second-order effect of pH (− 0.3011) > direct effect of concentration of H₂O₂ (0.0978) > second-order effect of Ferrous (− 0.02812) > interaction between the concentration of the ferrous and pH (− 0.0068) > second-order effect of H₂O₂ and pH (− 0.002272) > second-order effect of H₂O₂ (− 0.00016) > interaction between the concentration of the ferrous and H₂O₂ (− 0.000097).

A list of information for the proposed model such as the coefficients, and the ANOVA for testing the significance of the regression coefficient is offered in Table 4. It can be seen that more than 95% of the data (R² = 98.84%, adjusted R² = 98.22%) can be well predicted by the model, signifying that the terms which were considered in the suggested model are significant enough to make satisfactory predictions. Though, the addition of more terms increases the predictions of the model. Really, the high value of F ratio (160.77) approves the significance of the proposed model.

Main and interaction effect plots

The main effect (mean) plot is suitable for considering data in a designed experiment, with respect to main factors, where the factors are at two or more levels. The main effect plots of three factors in COD removal (%) are shown in Fig. 2. As demonstrated, the pH has a synergistic impact on the response.

Fig. 2

Main effect plots for COD removal (%) in uncoded value for t = 60 min

Influence of different variables on the removal of COD

The results can be achieved as three-dimensional (3D) plot presentations to study the influence of different variables on the response.

Effect of hydrogen peroxide concentration

As it can be seen from Fig. 3, the COD removal percentage was enhanced by an increase in the concentration of hydrogen peroxide based on the following equations (Eqs. 5–8).

$$ {\text{H}}_{2} {\text{O}}_{2 } + {\text{Fe}}^{2 + } \to \,{\text{OH}}^{ \cdot } + {\text{OH}}^{ - } + {\text{Fe}}^{3 + } $$

(5)

$$ {\text{OH}}^{ \cdot } + {\text{Fe}}^{2 + } \to {\text{OH}}^{ - } + {\text{Fe}}^{3 + } \to {\text{FeOH}}^{2 + } $$

(6)

$$ {\text{H}}_{2} {\text{O}}_{2 } \mathop{\longrightarrow}\limits^{)))}2{\text{OH}}^{ \cdot } $$

(7)

$$ {\text{H}}_{2} {\text{O}}_{2 } + hv \to 2{\text{OH}}^{ \cdot } $$

(8)

Fig. 3

Surface of response for the removal efficiency of COD versus the dosage of H₂O₂ (X_CHP) and Ferrous ions (X_CF)

The removal of COD was increased by an increase in the dosage of H₂O₂ from 250 to 500 mg/L, because more OH^· radicals were produced. However, the progress in the removal of COD was not notable or even it was decreasing with increases in the dosages of H₂O₂ from 500 to 750 mg/L. An extra increase in the concentration of hydrogen peroxide affected the reaction of the produced hydroxyl radicals with the additional H₂O₂ molecules to form hydroperoxyl radicals that have a less oxidizing ability according to the following reaction (Eq. 9):

$$ {\text{H}}_{2} {\text{O}}_{2 } + {\text{OH}}^{ \cdot } \to {\text{H}}_{2} {\text{O}} + {\text{HO}}_{2}^{ \cdot } $$

(9)

The hydrogen peroxide acts as free-radical scavenger at high concentrations; so, the removal of COD was decreased at high concentration of hydrogen peroxide (Eq. 10) [22].

$$ {\text{HO}}_{2}^{ \cdot } + {\text{OH}}^{ \cdot } \to {\text{H}}_{2} {\text{O}} + {\text{O}}_{2 } $$

(10)

Also, the auto decomposition of H₂O₂ to water and oxygen and the reaction of H₂O₂ with hydroxyl radicals instead of pollutant molecules were happened [23, 24]. The optimum amount of H₂O₂ offered by Model was 500 mg/L for the removal of COD. So, H₂O₂ should be added at an optimum amount.

Effect of ferrous ion concentration

Different concentrations of Fe²⁺ (from 12 to 36 mg/L) were used to achieve its optimum amounts. The removal efficiency of COD was reduced by an extra increase in the concentration of ferrous ion above 24 mg/L. The reason is that at high dosage of Fe(II), the production of hydroxyl radicals, generally originated from the breakdown of H₂O₂, was so high that many of the hydroxyl radicals were spent through the side reactions before using for the removal of COD (Eqs. 5, 6). Additionally, at high dosages of ferrous ion, the brown turbidity of the solution was occurred; consequently, the absorption of the UV light which is important for photolysis was delayed and the recombination of ${\text{OH}}^{ \cdot }$ radicals was occurred (Eq. 11). This result is in promise with the results of many investigates [24].

$$ {\text{FeOH}}^{2 + } + hv \to {\text{Fe}}^{2 + } + {\text{OH}}^{ \cdot } $$

(11)

Influence of pH

The pH of the solution is an important factor in the degradation of pollutants by Sono–photo-Fenton processes. At pH > 4, the formation of Fe(II) complexes decreases the production of free radicals and also the regeneration of ferrous ion was reduced by the precipitation of ferric oxyhydroxide. At pH around 3, the decomposition of hydrogen peroxide and production of hydroxyl radicals were increased; thus, the availability of the free radicals for the degradation of the pollutants was enhanced. Related results were observed by other researchers [25]. The effect of pH on the removal of COD is presented at Fig. 4. The degradation efficiency reduced considerably at alkaline pH because of the formation of precipitates of Fe(OH)₂ (s) and Fe(OH)₃ (s), which suppressed the production of HO radicals. The optimum pH was found to be 3. The optimum pH was 3.0–3.5 for Fenton reaction [26]. The pH controls the degradation efficiency of the Fenton, by controlling the solubility of Fe(II)/Fe(III), which finally adjusts the creation of HO radicals (Eqs. 12, 13).

$$ {\text{H}}_{2} {\text{O}}_{2 } + {\text{Fe}}^{2 + } + {\text{H}}^{ + } \to {\text{OH}}^{ \cdot } + {\text{H}}_{2} {\text{O}} + {\text{Fe}}^{3 + } $$

(12)

$$ {\text{H}}_{ 2} {\text{O}}_{ 2 } + {\text{Fe}}^{3 + } \leftrightarrow {\text{FeOOH}}^{2 + } + H^{ + } \to {\text{Fe}}^{2 + } + {\text{H}}^{ + } + {\text{HOO}}^{ \cdot } $$

(13)

Fig. 4

Surface of response for the removal efficiency of COD versus the pH (X_pH) and the concentration of H₂O₂ (X_CHP)

The response surface plots for the removal of COD in Sono–photo-Fenton process are revealed in Figs. 3, 4 and 5. From these 3D plots, the simultaneous interaction of the two factors on the responses was showed.

Fig. 5

Surface of response for the removal efficiency of COD versus the dosage of ferrous ions (X_CF) and pH (X_pH)

Response optimization and confirmation

In this study, the main purpose of the optimization was to obtain the optimum conditions and variables for the removal of COD in Sono–photo-Fenton process. The requested goal was introduced as “COD (%) maximum” to obtain largest decomposition yield. Degradation efficiency at optimum values of the process variables is presented in Table 5. Further experimental tests were done in the predicted value conditions and were verified. The results showed that the maximum degradation percent was achieved when the values of each variable were arranged as optimum values, which was in good agreement with the value predicted by the model. It discloses that the method used to optimize conditions in the removal of COD was successful.

Table 5 Degradation efficiency at optimum values of the process variables

At the achieved optimum condition, the COD of wastewater was reduced from 260 to 62 mg/L and the TOC was reduced from 100 to 31.3 mg/L, and about 76.1% of COD and 68.7% of TOC were removed.

Comparative study in Fenton, photo-Fenton and Sono–photo-Fenton process

The ultrasonic and photo-Fenton processes could be considered as appropriate and operative treatment methods that have been frequently used to remove many pollutants in the wastewater [27, 28]. Fenton and photo-Fenton processes are considered as a technology with a good potential for removing pollutant compounds from wastewater at room temperature and pressure [29]. As illustrated in Fig. 6, the removal of COD in optimum condition in Sono–photo-Fenton process, photo-Fenton (by stopping ultrasonic source) and Fenton process (by stopping UV light) were 76.0, 51.5 and 40.7%, respectively. In Fenton process, an acidic mixture of Fe²⁺ and H₂O₂ is used to create powerful OH^· radicals; also, the rapid transformation between Fe²⁺ and Fe³⁺ is very important for the high performance of the process [30]. In spite of this, it should be noted that the poor conversion property and the separation problems of the catalyst from the treated wastewater in the homogeneous Fenton processes, and the precipitation of Fe (OH)₃, depending on the pH, have limited the performance of this process. To overcome these limitations, the introduction of ultrasonic irradiation to the system can be a good choice. The addition of ultrasonic irradiation can lead to the creation, growth and collapse of microbubbles, which results in the local zones with a high temperature and pressure and subsequently, the enhancement of Fenton and photo-Fenton reactions [31].

Fig. 6

The removal of COD in different processes at optimum condition

Though, when both the US and UV are combined with Fenton process (Sono–photo-Fenton process), an important increase in the removal of COD (76.1% in 60 min) was obtained. About 24.6% raise in the degradation under the same time proposes that the more hydroxyl radicals were formed by Sono–photo-Fenton rather than photo-Fenton process. Therefore, a greater amount of degradation was achieved for Sono–photo-Fenton (US + UV + Fenton) compared with the individual processes. This phenomenon may be due to the production of hydroxyl radicals over photo-Fenton (Eqs. 8, 11, 14–16) and Sono–Fenton reactions (Eqs. 17–21):

$$ {\text{Fe}}^{3 + } + {\text{H}}_{2} {\text{O}} + hv \to {\text{Fe}}^{2 + } + {\text{H}}^{ + } + {\text{OH}}^{ \cdot } $$

(14)

$$ {\text{Fe}}^{3 + } + {\text{HO}}_{2}^{ \cdot } \to {\text{Fe}}^{2 + } + {\text{H}}^{ + } + {\text{O}}_{2} $$

(15)

$$ {\text{Fe}}^{2 + } + {\text{HO}}_{2}^{ \cdot } \to {\text{Fe}}^{3 + } + {\text{HO}}_{2}^{ - } $$

(16)

Subsequently, the high efficiency reached by the ultrasound supported in Sono–photo-Fenton process was originated from the production of additional hydroxyl radicals according to the following equations (Eqs. 17–21).

$$ {\text{H}}_{2} {\text{O }}\mathop{\longrightarrow}\limits^{)))}{\text{OH}}^{ \cdot } + {\text{H}}^{ \cdot } $$

(17)

$$ {\text{O}}_{2} \mathop{\longrightarrow}\limits^{)))}2{\text{O}}^{ \cdot } $$

(18)

$$ O^{ \cdot } + {\text{H}}_{2} {\text{O }}\mathop{\longrightarrow}\limits^{)))}2{\text{OH}}^{ \cdot } $$

(19)

Based on the Eqs. (6) and (13), the following reaction (Eqs. 20, 21) can be happened in the presence of ultrasonic wave:

$$ {\text{FeOH}}^{2 + } \mathop{\longrightarrow}\limits^{)))}{\text{Fe}}^{2 + } + {\text{OH}}^{ \cdot } $$

(20)

$$ {\text{FeOOH}}^{2 + } \mathop{\longrightarrow}\limits^{)))}{\text{Fe}}^{2 + } + {\text{HOO}}^{ \cdot } $$

(21)

The produced FeOOH²⁺ in these reactions can be decomposed into Fe²⁺ and hydroperoxyl by sonolysis. The recreated Fe²⁺ could then react with H₂O₂ once more and a closed rotation catalytic mechanism is then customary.

On the other hand, the ultrasound could inhibit the combination of the catalyst elements, thus improving the number of active sites exposed to the degradation reaction [32]. The turbulence effect produced by the ultrasound can accelerate the mass transfer in the solution by improving the sonochemical effect [33].

Based on Eq. (16), the Fe³⁺ was changed to Fe²⁺ and the Fenton reaction was started for the degradation of the phenol derivatives. With continuous irradiation of the Fe³⁺ solution, the regeneration of the catalyst and the production of additional hydroxyl radical were achieved. But, on the time that all of the H₂O₂ is spent by Fe²⁺, the reaction is stopped [34].

Conclusions

In this study, the FFD design of the experiment was used for the removal of COD in petrochemical wastewater by Sono–photo-Fenton process.

The influence of initial concentrations of H₂O₂, Fe²⁺ and pH on the removal of COD has been investigated. The optimum conditions predicted by the model were as follows: the [Fe²⁺] = 24 mg/L, pH at 3, and [H₂O₂] = 500 mg/L. The ANOVA tests were performed to determine the significance of independent variables on the response function. The ANOVA showed a satisfactory prediction second-order regression model and a high determination coefficient value ((R² = 0.9884, $R_{\text{pred}}^{ 2}$  = 0.9713, $R_{\text{adj}}^{ 2}$  = 0.9822 for removal of COD). The mean plot and three-dimensional plots were employed to study the parts of each factor, as well as their interactions on the removal of COD. At the achieved optimum condition, the COD of wastewater was reduced from 260 to 62 mg/L and the TOC was reduced from 100 to 31.3 mg/L and the removal of COD and TOC was 76.1 and 68.7%, respectively.

The study demonstrates that ultrasonic treatment can be a suitable method to improve the Fenton and photo-Fenton degradation methods.

The effect of ultrasonic power and the intensity of the UV lamp should be investigated in future works.