Heterogeneous Fenton-like oxidation of petrochemical wastewater using a magnetically separable catalyst (MNPs@C): process optimization, reaction kinetics and degradation mechanisms

Abstract

The strong tendency of magnetite nanoparticles (MNPs) to agglomerate limits their application in oxidation processes, due to the reduction of surface/volume ratio, dispersion stability and catalytic activity. To solve this problem, we successfully coated MNPs with an average size of 50 nm on an activated carbon surface in order to prepare a magnetic recoverable composite (MNPs@C). It was employed as a heterogeneous catalyst in the Fenton oxidation for petrochemical wastewater (PCW) treatment, due to its high ability in decompose H₂O₂ molecules. XRD, BET, VSM, SEM, TEM and EDX techniques were utilized to determine the catalyst's characteristics. The activity of the catalyst was assessed for the Fenton reaction using COD removal efficiency. Experiments related to the Fenton oxidation process (FOP) were carried out in order for process optimization and evaluation of the degradation kinetics and mechanism. It was observed that both oxidation Fenton and adsorption processes occurred simultaneously in the MNPs@C/H₂O₂ system. The results clearly showed that the organic compounds in the PCW have been degraded by the hydroxyl free radicals (˙OH) released from decomposition of H₂O₂ in the presence of MNPs@C. Under the optimum operating conditions (pH 3.0, 1 g L⁻¹ catalyst and 50 mM H₂O₂), the removal efficiency of COD was found to be 83.5% within 120 min reaction time. More than 65% of COD was experimentally removed after five catalytic cycles, which demonstrates the promising application of the catalyst in the oxidative degradation of organic pollutants. In addition, MNPs@C exhibited low iron leaching (<0.3 g L⁻¹) and stable catalytic activity upon five recycling cycles. The catalyst could be easily recovered and showed high potential for applications in wastewater treatment without secondary pollution. In conclusion, the MNPs@C/H₂O₂ system, as a promising technique, can provide appropriate conditions for the pretreatment of PCW prior to biological processes or as a tertiary treatment for the reuse of effluents.

---

1. Introduction

Nowadays, contamination of natural aqueous systems is a crisis due to rapid industrial development and the resulting major quantity of toxic compounds discharged into the environment. Petrochemical wastewaters (PCW) not only contain a huge load of organic pollutants at high concentrations but also exhibit a wide diversity with respect to their molecular structures. These are characteristically less biodegradable in nature with diverse pollutants containing high concentrations of salt and carbon. The discharge of PCW may cause serious environmental pollution and human health concerns.^1,2 Therefore, an efficient PCW treatment process is required to meet the increasingly stringent discharge standards.

Several treatment methods, including biological, physical, chemical, to name a few, have previously been tested in PCW treatment. However, some of these processes suffer from drawbacks such as partial degradation of the effluent, toxic byproducts and excessive sludge production, energy consumption, and secondary phases generation that impose extra cost in the process.^1,3 Additionally, biological treatment methods have limitations due to poor availability of hydrocarbons to microorganisms resulting from their complex structure and water insoluble nature, especially when salinity is higher in wastewaters.¹ Therefore, to overcome these drawbacks, enhancement of the treatment efficiency of conventional wastewater treatment methods is necessary.

Advanced oxidation processes (AOPs), using various types of oxidants, have been found to be effective at degrading recalcitrant and chemically complicated contaminants. AOPs involve the generation of highly active oxidizing species which attack and decompose organic components and make these processes more efficient, in comparison with other physically-based techniques (e.g. adsorption and flocculation).^4,5 In this regard, Fenton and Fenton-like oxidation consisting of catalyst (Fe²⁺/Fe³⁺) and hydrogen peroxide (H₂O₂) have been reported to be one of the most effective techniques for degrading and mineralizing organic contaminants in wastewater.^6,7 Fenton oxidation process (FOP), according to eqn (1) and (2) below, is one way to produce hydroxide radicals, HO˙, by a reaction between H₂O₂ as an oxidant and Fe²⁺/Fe³⁺ ions as a catalyst:^4,8

Fe²⁺ + H₂O₂ → Fe³⁺ + HO˙ + OH⁻ (1)

Fe³⁺ + H₂O₂ ↔ Fe²⁺ + HOO˙ + H⁺ (2)

The generated free radicals, HO˙, are robust oxidant species reacting unselectively with organic compounds which result in the mineralization of these compounds into inorganic ions, CO_2, and H₂O.^9,10

However, there are certain limitations regarding the application of FOPs. The homogeneous type operates in a narrow range of pH (2–4). Besides, this type is uneconomical and has particular problems regarding the separation and recovery of the ions after the catalyst reaction. This process can also generate secondary pollution (e.g. acid or metal ions) as well as metal hydroxide sludge.¹¹ To overcome these problems, heterogeneous Fenton system can be applied. This method is based on using iron-containing solids (Fe₂O₃, Fe₃O₄, FeO, FeOOH etc.) and/or incorporating Fenton's catalyst onto surfaces of different carriers (e.g. activated carbon, zeolite, clay, multi-walled carbon nanotubes and polymer).^11–13 Among iron oxide minerals, as reported previously, magnetite nanoparticles (MNPs) constitute the most effective heterogeneous Fenton catalyst.^8,14 MNPs can decompose H₂O_2, which results in the formation of HO˙ as mentioned in eqn (1). However, MNPs have a strong tendency to agglomerate due to intra-particle interactions (e.g. van der Waals and intrinsic magnetic interactions), which could decrease the surface/volume ratio of MNPs, and disperse their stability in the solution and, finally, their catalytic activity is reduced as well.^15,16

Powder activated carbon (PAC), as a supporting for Fenton catalysts, can be considered as a promising and economical method due to its wide availability, low cost, high specific surface area, and porosity.^16,17 Not only can the performance of FOP regarding the adsorption of pollutants be increased but also the separation of catalyst from aquatic media can be facilitated by applying PAC. The magnetic composites have considerably been studied on to be used as heterogeneous catalysts in the Fenton and Fenton-like processes because of their highly acceptable reusability and stability, low toxicity, and easy separation. Several studies concluded that MNPs have a high catalyst activity and can effectively decompose H₂O₂ into HO˙ radicals. They have also observed that these composites can improve electron transfer and have higher surface areas as well as active sites in comparison with single magnetic catalysts.^4,18 Furthermore, the magnetic property of these composites helps to easy and rapid separation of catalysts from the reaction solution.

Since most studies have been conducted on the performance of various types of FOP for synthetic wastewater treatment and few studies have been carried out on real wastewater treatment, the performance and ability of these processes for real applications is vague. Therefore, the objective of this paper was to develop a wastewater treatment method to benefit the environmental management of petrochemical industries. MNPs loaded on the powder activated carbon (MNPs@C) were characterized and employed in a heterogeneous catalyst Fenton oxidation system treating PCW. Effects of solution pH, H₂O₂ concentration, and catalyst quantity were evaluated on the FOP performance in a batch system. The performance of FOP in treating defined wastewater was evaluated with respect to chemical oxygen demand (COD). Moreover, the kinetics of COD removal as well as the stability and reusability of the catalyst were assessed.

2. Materials and methods

2.1. MNPs@C preparation and characteristics

To synthesize of the catalyst ferric nitrate (Fe(NO₃)₃·9H₂O, ≥99.0%), nitric acid (HNO₃, 65%) powder activated carbon (PAC) were used as purchased. For the heterogeneous Fenton reaction, hydrogen peroxide solution (H₂O₂, 35%) was used. All chemicals applied were of analytical grade and were provided by Merck Co. (Merck, Darmstadt, Germany). MNPs@C catalyst was synthesized by using a chemical co-precipitation method according to the procedure reported in our previous study.¹⁹ The XRD pattern of PAC, MNPs and MNPs@C were analyzed (Quantachrome, NOVA 2000) using graphite monochromatic copper radiation (Cu Kα, λ = 1.54 Å) at 25 °C. The surface morphology of MNPs@C and MNPs distribution within the PAC was evaluated using a scanning electron microscope (SEM, PHILIPS, XL-30) at 25 keV. The shape and size of MNPs impregnated on the PAC were also assessed by using transmission electron microscopy (TEM, PHILIPS, EM, 208) at 100 keV. An energy dispersive X-ray technique (EDX, PHILIPS, XL-30) was applied to characterize the adsorbent elemental composition. The magnetic property of MNPs@C was investigated employing a vibrating sample magnetometer (VSM, 7400, Lakeshare, USA) at ±10 kOe in room temperature. The physicochemical properties of PAC, MNPs and MNPs@C were analyzed using the Brunaeur, Emmett and Teller (BET, Quantachrome, NOVA 2000) method using N₂ adsorption–desorption isotherms at 77.3 K.

2.2. Wastewater sampling and characterization

The study was conducted on industrial wastewater collected from a petrochemical company in Khuzestan Province, southeast Iran. Wastewater samples were taken in one day and from the exit point of the treatment plant, when the petroleum refinery was operating under normal conditions. Samples were transferred to the laboratory and stored under refrigeration (4 °C) until use. The characteristics of the wastewater used in the oxidation experiments are given in Table 1. The wastewater presents high levels of TDS and a moderate level of COD and TOC content.

Table 1 Properties of PCW used in this study

Parameter	Unit	Range
pH	—	7.0–7.50
COD	mg L^−1	50–1000
BOD5	mg L^−1	10–220
BOD5/COD	—	0.2–0.3
TOC	mg L^−1	250–300
TDS	mg L^−1	22 000–100 000
Conductivity	μs cm^−1	≈100 000
Appearance	—	Light yellow

---

2.3. Batch experiment set-up and procedure

In the present study, all the experiments were performed in a batch comprising the glass flasks filled with 200 mL of the sample at 25 ± 1 °C. After pH-adjusting, a known amount of Fenton reagents (H₂O₂ and MNPs@C) was added to glass flasks and were mixed using a shaker for a certain period of time with an agitation speed of 200 rpm to ensure the ideal mixing of the reagents and the solution. At given time intervals, 2 mL of the sample from the solution was extracted and the aqueous phase was then separated using an external magnet in less than 1 min. Finally, the residual COD was determined using titrimetric method (5220-C; closed-reflux) according to the standard methods.²⁰ The removal percentage of COD was calculated using eqn (3).

The effects of variables such as pH, different catalyst, and H₂O₂ dosages on the COD removal efficiency were assessed. Firstly, the effect of the pH in the range of 2 to 9 was investigated for a period of 240 min at 25 ± 1 °C. Hydrochloric acid 0.1 M (HCl) and sodium hydroxide 0.1 M (NaOH) (i.e. purchased from Merck CO.) were used to adjust the pH of the samples. A pH meter (HACH Sension 4) was employed for pH measurement. After the optimization of pH, the effect of different dosages of catalyst and H₂O₂ were evaluated over a range of 0.1 to 5 g L⁻¹ and 5 to 500 mM, respectively. Afterwards, the kinetics of COD removal by MNPs@C was also studied. A GC-MS (Agilent, USA) was used for the organic compounds analysis. Moreover, reusability of catalyst under obtained optimal conditions was investigated in five consecutive cycles of use. It should be mentioned that the control experiments in which the catalyst was not added were conducted in parallel with the main ones. All experiments were carried out in triplicate, and all results were expressed as the mean value of three measurements. The error bars in the figures were omitted for graphic simplicity except for the cases where they were necessary.

where COD_i and COD_f are the COD values at initial and time t of the Fenton reaction, respectively.

3. Results and discussion

3.1. Characterization of catalyst

The XRD pattern of MNPs@C in the 2θ range of 10–70° shows a broad diffraction peak at 2θ = 25°, belonging to the characteristic reflection of carbon amorphous nature (see Fig. 1(a)). The characteristic peaks for MNPs at 2θ angles of 30.07°, 35.44°, 43.15°, 54.6°, 56.99°, and 62.6°, which can be marked by indices (220), (311), (400), (422), (511), and (440) were based on the JCPDS, no. 19-0629. This confirms the successful synthesis of MNPs. It was also verified through the absence of peaks corresponding to maghemite, ranging from a 2θ angle of 20° to 30°. As a result, the XRD pattern of MNPs was quite similar to that of pure magnetite. For MNPs@C catalyst, all the peaks belonging to both carbon and MNPs were observed, which demonstrated the successful impregnation of MNPs on the carbon surface.

Fig. 1 (a) XRD patterns of PAC, MNPs and MNPs@C, (b) SEM images of PAC, (c) SEM image of MNPs@C and (d) TEM image of MNPs@C.

Fig. 1(b) and (c) show the SEM images and the surface morphology of PAC and MNPs@C, respectively. It indicates appropriate porosity and uniform distribution of the pores for the PAC. The external surface PAC was more porous when compared with MNPs@C. This can be due to the agglomeration and uniform distribution of MNPs on the PAC surface for MNPs@C catalyst. Fig. 1(c) also reveals that the external surface of MNPs@C has irregular clumps cavities, which could provide more reactive sites and a high catalyst ability for the prepared composite. The TEM micrograph of catalyst (Fig. 1(d)) indicated that the MNPs were successfully synthesized with nano-size (50 nm) and a cubic structure, which is consistent with the results of the XRD analysis. The results of EDX analysis (Fig. 2(a)) elucidated the presence of specific elements, including carbon, oxygen, lead, zinc, and iron in the structure of the synthesized catalyst. As seen, the oxygen (4.9%) and iron (16.3%) are the major elements of the impregnating process. The amount of lead and zinc, collectively, was <1 wt%, which represents very minor impurities in the PAC structure.

Fig. 2 EDX (a) and VSM analysis of MNPs@C (b) and magnetically separation performance of PAC, MNPs and MNPs@C from solution using an external magnet (c).

Textural properties (specific surface area, volume, and average diameter of the pores) of PAC, MNPs and MNPs@C are summarized in ESI data Table S-1.† These characteristics were measured using BET method with N₂ adsorption–desorption isotherms at 77.3 K and p/p₀ = 0.99. The specific surface area of PAC decreased from 936 to 671.2 m² g⁻¹ after the coating of MNPs. This 28% decrease can originate from the filling of pores of PAC by MNPs, as reported in the literature.²¹ For MNPs@C, the average size and volume of pores were estimated to be 3.5 nm and 4.87 cm³ g⁻¹ respectively, which according to the IUPAC classification, it can be classified to the mesopores groups.²² As represented in Table S-1,† the specific surface area of MNPs was lower than those of PAC and MNPs@C. The results of N₂ adsorption/desorption reveal type IV isotherm for the MNPs@C as categorized by the IUPAC, indicating that the MNPs@C structure is typically mesoporous.

The maximum magnetic saturation of the MNPs@C catalyst pre- and post-use in the FOP was 7.1 and 6.05 emu g⁻¹, respectively. For MNPs, however, maximum saturation of magnetization of 11.4 emu g⁻¹ was obtained, which was higher than that of MNPs@C pre- and post-use in the FOP. This decrease can be attributed to the presence of non-magnetic PAC in the catalyst texture. A typical S-type hysteresis loop with no residual magnetism or coercivity was found for MNPs@C, suggesting that the MNPs@C were super-paramagnetic. According to Fig. 2(b), after the reaction, there were no remarkable changes in the magnetic properties of MNPs@C, which indicates that the catalyst could be used for several times with only a slight loss of saturation magnetization. We also observed that the catalyst had a good magnetic response when it was placed in a magnetic field (Fig. 2(c)). MNPs@C demonstrated that not only it can be easily and rapidly separated from solutions, but also it can be potentially used as a magnetic catalyst to remove contaminants from the aqueous environment to avoid a secondary pollution.

3.2. Comparison of different processes in COD removal

The removal efficiency of COD by various processes was compared under identical experimental conditions (pH = 3.0 ± 0.3 and T = 25 ± 1 °C) and the results are platted in Fig. 3. Heterogeneous Fenton oxidation of PCW by MNPs@C and H₂O₂ exhibits the highest COD removal (86.2%) compared with that by other processes for 240 min treatments. The COD removal efficiency of 42.7, 49.3, 62.4 and 71.1% was obtained by H₂O₂, PAC, MNPs@C and H₂O₂/PAC, respectively during 240 min reaction time. Low removal of COD by H₂O₂ can be explained by the lower oxidation potential of H₂O₂, compared with the oxidation potentials of HO˙ and HO₂˙ radicals. The adsorption of contaminants on PAC and MNPs@C leads to the removal of COD. Moreover, it was observed that the removal efficiency achieved by PAC and MNPs@C was higher than that of H₂O₂ alone, which is attributed to their similar surface area and pore structures. According to Table S-1,† the specific surface area of PAC composite was higher than that of MNPs@C. However, it was observed that the percentages of COD removed by MNPs@C were higher than the corresponding amounts obtained by PAC. This means that coating MNPs onto PAC contributed to having additional adsorption sites for contaminants. Therefore, it can be concluded that MNPs have a synergistic effect on PAC, which could result in an increase in COD removal by MNPs@C.

Fig. 3 The removal efficiency of COD by different processes (pH: 3.0 ± 0.3, PAC, MNPs and MNPs@C dose: 1 g L⁻¹, H₂O₂: 50 mM and T: 25 ± 1 °C).

As shown in Fig. 3, when the solution was exposed to the combination of H₂O₂ and MNPs@C, COD content significantly decreased to 86.2% after 240 min reaction. This removal percentage was higher than that of H₂O₂ and MNPs@C alone. Under this condition, high removal of COD can be due to the simultaneous occurrence of oxidation and adsorption. This showed that MNPs in the structure of PAC had a synergistic effect on the decomposition of H₂O₂ to produce reactive species for the degradation of organic pollutants. However, in the heterogeneous Fenton reactions investigated in previous studies, the activated carbon and graphite could generate free radicals (e.g. superoxide ion and activate hydrogen peroxide).²³ In this process, H₂O₂ activating ability and catalytic activity increased via MNPs distribution on the PAC surface, which could also avoid the agglomeration of MNPs. These results demonstrated that MNPs@C exhibited a good catalytic activity for the Fenton process. Hence, the COD removal difference in FOP with various materials was mainly related to catalytic oxidation efficiency. In this study, the MNPs@C catalyst and H₂O₂ oxidant were simultaneously employed as a Fenton heterogeneous process for further experiments.

For MNPs@C/H₂O₂ process, it is clear from Fig. 3 that the removal efficiency of COD has increased as a result of increasing the reaction time, especially at initial times of Fenton reaction. It could be attributed to the further produce of hydroxyl free radicals (HO˙) by increasing the reaction time.^14,24 In other word, an increase in the reaction time helps to more interaction between catalyst and oxidant, and subsequently decomposition rate of H₂O₂ increased. The radicals are produced by H₂O₂ decomposition in the presence of MNPs@C catalyst according to the following equations.

Fe²⁺ + H₂O₂ → Fe³⁺ + HO˙ + OH⁻ (4)

3.3. Effect of initial pH

The pH of solution can affect the activity of the oxidant and the substrate as well as the H₂O₂ stability. Several researchers have reported that the solution pH can dramatically influence the Fenton degradation of organic compounds and at higher rates in the region 2.5–4, with an optimum at pH 3.0.^8,25 Fig. 4(a) reveals the effect of various ranges of pH on COD removal in an FOP (i.e. MNPs@C/H₂O₂). The minimum COD content in the solution was observed for the pH value equal to 3.0. The COD removal (%) increased from 77.2 to 85.3 when the initial pH of the solution increased from 2.0 to 3.0, whereas it decreased to 68.4% for pH 9.0 after 240 min of reaction time. Therefore, the value of 3.0 is considered to be the best pH for maximum removal of COD in FOP. At lower pH values (acidic), a more fraction of iron species are dissolved in the solution which can result in an increase in the oxidation rate of contaminants.^5,26 Based on these findings, the significant effect of the pH value on the catalytic activity of Fe²⁺ in the Fenton reaction system can be concluded. This observation agrees with the results reported in the literature for the Fenton degradation of contaminants.^2,8,11 It has previously been reported that the pH values around 2.8 are the optimal values for maximum HO˙ production in the FOP. In a further related study, Wang et al.,²⁵ demonstrated that very low values of pH, where the concentration of H⁺ ions is too high, will slow down the formation of FeOOH²⁺, which consecutively causes the production rates of ferrous ions and hydroxyl radicals to decrease as well. At higher pH values, however, the lower oxidation potential of HO˙ radicals, decomposition of H₂O_2, and deactivation of the catalyst due to the formation of ferric hydroxide complexes can lead to the reduction of removal efficiency.^26,27

Fig. 4 Effect of initial pH solution (a) and catalyst loading (b) on COD removal in the FOP (H₂O₂: 50 mM and T: 25 ± 1 °C; (a) MNPs@C: 1.0 g L⁻¹ and (b) pH: 3.0 ± 0.3).

In addition, results in Fig. 4(b) exhibit an increase in the COD removal efficiency parallel with increasing the reaction time, which could be due to the production of additional hydroxyl free radicals (HO˙).⁴ It is observable that the reduction efficiency of COD by MNPs@C/H₂O₂ system was kinetically faster at initial times of the reaction, and then slightly increased and reached the maximum values at about 60 min. However, the reaction rate was approximately constant after times of 60 min, and no further COD removal increase was observed with further prolonged reaction time, suggesting that the COD removal by MNPs@C/H₂O₂ is a rapid process.

3.4. Effect of catalyst loading

The removal of COD in the MNPs@C/H₂O₂ system was evaluated with various loadings of catalyst, pH: 3.0 ± 0.3, and H₂O₂ concentration of 50 mM within 240 min reaction (Fig. 4(b)). It was observed that the percentage of COD removal increases with an increase in the amount of catalyst. It is clear from Fig. 4(b) that during the FOP, the efficiency of COD removal increased from 55 to 91.75% as the catalyst loading increased from 0.1 to 5 g L⁻¹. This is mainly due to increasing the number of active sites and also improving the decomposition of H₂O₂, which results in generating more HO˙ radicals. Several researchers founded that the higher amounts of heterogeneous catalyst provide additional surface area for the adsorption as well as additional amounts of iron species for the formation of HO˙ radicals.^8,28,29 Moreover, an increase in the catalyst loading can accelerate the decomposition of H₂O₂ to HO˙ due to the presence of more redox-active centers. As shown in Fig. 4(b), since remarkable changes in the removal of COD by FOP were not observed beyond 1 g L⁻¹ loading, we used this amount as the best operating condition of catalyst loading in this study. These results are in agreement with those reported in the literature.²⁷

3.5. Effect of H₂O₂ concentration

In a Fenton system, the oxidant concentration is a major factor that can significantly influence the degradation of organics. In this reaction, H₂O₂ is the source of HO˙ and the major cost for scale-up application. H₂O₂ is directly related to the number of produced HO˙ radicals, so it can play the role of an oxidizing agent in FOP.³⁰ The effects of different concentrations of H₂O₂ (i.e. ranging from 5 to 500 mM) on the COD removal in the FOP was assessed under the optimum conditions (catalyst loading of 1 g L⁻¹ and pH 3.0 ± 0.3) during 240 min reaction time and the results are illustrated in Fig. 5. It was observed that the removal (%) of the COD was a function of H₂O₂ concentration at the given conditions. A rise in initial concentration of H₂O₂ from 5 to 50 mM could enhance COD removal. As shown in Fig. 5, maximum removal of COD was obtained at the concentration of 50 mM H₂O₂. At lower concentrations of H₂O₂ in the solution, an adequate number of HO˙ radicals cannot be produced. This can also contribute to lowering the oxidation rate and subsequently reducing the removal efficiency.³¹ When it is increased, however, the process efficiency is enhanced due to higher concentration of HO˙ radicals. Therefore, in the current study, the optimal initial concentration of H₂O₂ (50 mM) was deemed for the effective oxidative degradation of PCW. It is clear from Fig. 5 that excessive concentrations of H₂O₂ (>50 mM) do not improve the removal efficiency of COD. This can be explained by considering the scavenging effect of HO˙ radicals at the presence of excessive H₂O₂, which results in a decrease in the number of HO˙ radicals in the solution (eqn (5) and (6)).^11,32,33 It was reported previously that a competitive reaction forms between H₂O₂ and the contaminant when there is high H₂O₂ concentration in the solution. Moreover, in this condition, HO˙ radicals react with H₂O₂ to produce HO₂˙ radical, which is less effective in the degradation of the contaminant compared with the effectiveness of HO˙.³⁴ This could, in fact, reduce the rate of degradation. Segura et al.,³⁵ demonstrated that when one of the Fenton reagents such as H₂O₂ and/or Fe²⁺ is added in excess, the organic degradation reaction is inhibited by radical scavenging reactions. In this study, such inhibition was observed at the higher H₂O₂ concentrations of 50 mM.

HO˙ + H₂O₂ → H₂O + HO₂˙/O₂⁻˙ (5)

HO₂˙ + HO˙ → H₂O + O₂ (6)

Fig. 5 Effect of H₂O₂ feed concentration on COD removal in the FOP (pH: 3.0 ± 0.3, MNPs@C: 1.0 g L⁻¹ and T: 25 ± 1 °C).

3.6. Kinetic study

Since the Fenton reaction is complex, it is certainly inconceivable to carry out a detailed kinetic study on different individual reactions that have occurred during the Fenton process. However, in an approximate kinetic study, this is possible through some parameters (e.g., COD and TOC) that represent the overall organic pollutants.³⁶ In recent years, different types of kinetic models for heterogeneous Fenton oxidation reactions have been examined. In this study, the degradation kinetics of COD by MNPs@C/H₂O₂ system was investigated at different dosages of catalyst and H₂O₂ using three models (zero-, first- and second-order models).⁸ These models were employed in order to provide a comparison between the results of model estimation and experimental data. The non-linear equations and constants of the kinetic models of COD removal by MNPs@C/H₂O₂ system are given in ESI data Table S-2,† where C and C₀ are the COD concentrations of the PCW at the time of t (mg L⁻¹), t is degradation time (min), and k is the rate coefficient of kinetic model (min⁻¹).

The calculation of the error deviation is vital in the evaluation of the goodness of the fit of the kinetic models and the description of the degradation phenomena. In this study, to find out the fit quality of any kinetic model, apart from the correlation coefficient (R²), the validities of the models were assessed by three tests: R_adj², the root-mean-square error (RMSE), and the sum of squared errors (SSE). These criteria describe the goodness of fit between the experimental and predicted data. R_adj², RMSE and SSE can be calculated according to Table S-3.† The smaller RMSE and SSE values as well as the R_adj² and R² values close to one suggest the best and the most valid model.

The results from the study of kinetic models showed that the COD removal by MNPs@C/H₂O₂ system can be better described by the pseudo-second order model. As presented in Table S-2,† the regression coefficient (R²) obtained by second-order kinetic model is higher than the value obtained with the zero- and first-order models. Furthermore, the higher R_adj² and lower RMSE and SSE calculated for the pseudo-second-order model demonstrate that the COD elimination using the FOP ranked next to the second-order kinetic model. The non-linear curves of pseudo-second-order kinetic model of COD removal from PCW by MNPs@C/H₂O₂ system for various concentrations of the catalyst and H₂O₂ are platted in Fig. 6(a) and (b). It can be observed that the degradation process follows the pseudo-second-order model with a fairly good consistent. Fig. 6(c) also exhibits the agreement between the experimental data from the studied process and those from the kinetic models under the optimum conditions whereas zero- and first-order models indicated a poor performance to fit the curve.

Fig. 6 Pseudo-second-order kinetic model of COD removal in the FOP for various concentrations of catalyst (a) and H₂O₂ (b) and the coherence between the experimental data from FOP with those from different kinetic models (c). Changes of the values of the rate coefficient of pseudo-second-order kinetic model at various concentrations of catalyst and H₂O₂ (insert of a and b).

According to Table S-3,† the values of the reaction rate constants (k) increased as a result of increasing catalyst loading in the system. When the catalyst loading is enhanced from 0.1 to 5.0 g L⁻¹, the value of k increases from 2.5 × 10⁻⁵ to 1.4 × 10⁻⁴ (min⁻¹). This can be attributed to increasing the surface of fresh iron (initial catalyst addition) which can accelerate the reactions of iron dissolution and H₂O₂ decomposition, as reported in the literature.^11,37 On the other hand, we found out, as expected, that the values of k also increased from 7.2 × 10⁻⁵ to 1.3 × 10⁻⁴ min⁻¹ when the H₂O₂ dose increased from 5 to 200 mM. This phenomenon might be explained by the higher production of hydroxyl radicals from H₂O₂ decomposition when H₂O₂ dose increased. These results are in line with the reports of previously conducted studies for heterogeneous Fenton oxidation of various compounds.^38,39

3.7. Mechanism of COD removal in the MNPs@C/H₂O₂

As previously mentioned, the COD removal was not mainly based on adsorption by MNPs@C or direct oxidation by H₂O₂. Therefore, the effective removal of COD results from using heterogeneous Fenton system (MNPs@C/H₂O₂). In heterogeneous Fenton type process, the reaction between ferrous or ferric ions and hydrogen peroxide takes place on the surface of solid catalyst. Furthermore, this reaction depends on the specific surface area of the catalyst.⁴⁰ In this study, the reaction was possibly initiated by adsorption on the PAC surface. In fact, this can help the degradation process.

There is no doubt that the MNPs can easily adsorb H₂O₂ molecules. Also, the surface properties and chemical compositions of MNPs play a critical role in the peroxidase-like catalysis in the removal of organic pollutants.⁴¹ On the other hand, the MNPs catalyst requires having a large specific surface and high dispersion ability to effectively interact with H₂O₂. Herein, when the initial molecules of H₂O₂ are adsorbed on the surface of catalyst, they are catalyzed by MNPs coated on PAC surface to produce HO˙ and HO₂˙ as described in eqn (7). The HO˙ radicals are then responsible for the degradation and mineralization of organic molecules and converting them into H₂O and CO₂. The possible mechanism of H₂O₂ activation by MNPs@C catalyst and the two routes of pollutants removal (i.e. adsorption onto MNPs@C and degradation by MNPs@C/H₂O₂), are given in Scheme 1. In the following, Fe²⁺ ions, in reaction with H₂O₂, were oxidized to Fe³⁺ ions (eqn (7)) and then regenerated via the reduction of Fe³⁺ by H₂O₂ (eqn (8)). Fe³⁺ species can also produce HO₂˙/, O₂⁻˙, which is less reactive than HO˙ radicals.^11,42 In this proposed mechanism, organic matters are broken down by hydroxyl radicals including HO˙_(ads) on the surface of catalyst and HO˙_(free) in the bulk solution (eqn (9)).

HO˙_(ads)/HO˙_(free) + organic matter → degraded products (9)

Scheme 1 The route of organic pollutants adsorption onto MNPs@C and their degradation by the heterogeneous Fenton oxidation catalyzed with MNPs@C and H₂O₂.

To evaluate the performance of MNPs@C/H₂O₂ system, we analyzed PCW before and after treatment with GC-MS. A wide range of organic compounds were identified in raw PCW as shown in Fig. 6. The main compounds are also listed in Table 2. This analysis exhibited the presence of some organic aromatics (e.g. benzene and toluene compounds) and a wide variety of organic acids (e.g. dichloroacetic acid, pentanoic acid and trichloroacetic acid) as summarized in Table 2. It can be seen that the amount of most organic substrates significantly decreased after treatment by MNPs@C/H₂O₂ system when the two chromatograms were compared with each other. The complete removal of some organic pollutants (i.e. 2, 3, 11, 12 and 13 compounds) was obtained after treatment, indicating the MNPs@C/H₂O₂ system is an effective process for the organic compounds degradation and removal (Fig. 7 and Table 2). Therefore, MNPs@C/H₂O₂ system proves to be a promising technique which can provide proper conditions for the pretreatment of PCW prior to biological process.

Table 2 Main organic pollutants identified in PCW by GC-MS analysis

No.	Retention time (min)	Chemicals	Area (10^6)		Removal (%)
			Untreated	Treated
1	13.37	2-Hydrazino-4,6-dimethylpyrimidine ditms peak 1	6.5	0.553	91.5
2	13.71	4-Hydroxymandelic acid, ethyl ester, di-TMS, 5H-naphtho	6.27	—	100
3	15.16	Benzene, 1-methyl-2-nitro-	32.89	—	100
4	15.84	Methoxyacetic acid, 3-tridecyl ester	32.893	0.976	97.03
5	16.22	Benzene, 1-methyl-3-nitro-	15.96	0.716	95.5
6	16.86	Benzenamine, 2-chloro-6-methyl- $$ o-toluidine, 6-chloro- $$ 2-amino-3-chlorotoluene $$ 2-chloro-6-methylaniline $$ 2-methyl-6-chloroaniline $$ 3-chloro-2-aminotoluene $$ 6-chloro-o-toluidine $$ 6-chloro-o-toluidine (NH2 = 1) $$ 6-chloro-2-toluidine $$ 6-ch	16.838	1.199	92.87
7	17.16	Benzenamine, 3-chloro-4-methyl- $$ p-toluidine, 3-chloro- $$ DKC 1347 $$ 1-amino-3-chloro-4-methylbenzene $$ 2-chloro-4-aminotoluene $$ 3-chloro-p-toluidine $$ 3-chloro-4-methylaniline $$ 3-chloro-4-methylbenzenamine $$ 4-methyl-3-chloroaniline $$ 2-chlor	11.01	4.37	60.24
8	18.85	2-Decen-1-ol, (E)-	1.69	0.712	57.9
9	20.75	2,3-Nonadiene	5.018	0.49	90.16
10	21.5	(1,1-Dodecanediol, diacetate)	5.092	1.38	72.9
11	22.32	2,6-Dichloro-3-methylaniline $$ benzenamine, 2,6-dichloro-3-methyl-	19.78	—	100
12	23.41	Benzene, 2-methyl-1,3-dinitro- $$ toluene, 2,6-dinitro- $$ 2,6-dinitrotoluene $$ 2,6-Dnt $$ 1-methyl-2,6-dinitrobenzene $$ Rcra waste number U106 $$ 2-methyl-1,3-dinitro-benzene	31.28	—	100
13	26.46	Pentanoic acid, 2,2,4-trimethyl-3-carboxyisopropyl, isobutyl ester	7.53	—	100

---

Fig. 7 GC-MS chromatograms of untreated (a) and treated (b) PCW by FOP.

3.8. Effect of MNPs@C reuse times

The deactivation and potential reuse of the catalyst, from a practical point of view, is an important characteristic. A series of batch tests were conducted to evaluate the capability of the MNPs@C used in the present FOP for reuse. To this purpose, the removal percentages of COD was investigated under obtained optimal conditions for at least five consecutive cycles. Herein, between each cycle of use, the catalyst was separated from the solution, dried and was then reused for the next experiment without any modifications. After each recycling, the solution was analyzed in order to determine the concentrations of COD.

According to Fig. 8, the COD removal efficiency by MNPs@C/H₂O₂ system decreased after five use cycles. The COD removal in FOP decreased from 86% in the first cycle to 65% after 5 cycles of reaction. The results of this study emphasize the potentiality of MNPs@C as a reusable catalyst in Fenton oxidation. As compared with first cycle, the values of drop in the COD removal efficiency were found to be 7.2, 18.2, 20.3 and 22% for second, third, fourth and fifth cycles, respectively. This drop for each cycle can be explained by the mass loss of MNPs. In addition, deactivation of the catalyst can also be attributed to the decay of active catalytic sites caused by low amount of leached iron from the catalyst surface. Meanwhile, the difficulty in the complete removal of residual by-products and reactants from the active catalytic sites in the following washing and drying procedures can be a reason for the loss of activity. In conclusion, FOP with MNPs@C is a stable method for treating PCW containing organic compounds.

Fig. 8 Effect of MNPs@C reuse times on COD removal in the FOP under the optimum operation conditions (pH: 3.0 ± 0.3, MNPs@C: 1.0 g L⁻¹, H₂O₂: 50 mM and T: 25 ± 1 °C).

3.9. MNPs@C stability in the FOP

The stability of the catalyst was also assessed through determining the concentrations of the dissolved iron ions in the solution during the five consecutive cycles. The iron concentration in the reaction solution in all studied cycles was found to be <0.3 mg L⁻¹, which is very low (data not shown). This means that the active species were strongly embedded into the mesoporous matrix of carbon through stable bonding. Therefore, this finding confirms that leaching of iron from MNPs@C might not cause metal pollution in the water, even under acidic conditions (pH = 3.0 ± 0.3). These results revealed that MNPs@C catalyst has acceptable catalytic stability and durability. In addition, it can be employed as a promising heterogeneous catalyst to PCW with a very low loss of catalytic activity.

3.10. Economic assessment

Based on the market research we conducted recently, the price of the commercial activated carbon and MNPs ranges from 100 and 65 to 150 and 160 $ kg⁻¹, respectively. While, according to the procedure we applied, the cost of preparation of MNPs@C catalyst was between 60 and 70 $ kg⁻¹, which was lower than the cost of activated carbon and MNPs alone. In addition, it is noteworthy that a part of the capital preparation cost of MNPs@C catalyst in practical applications could significantly be reduced, due to the high operational reusability and good stability. Meanwhile, we observed that MNPs@C is easily recoverable via a magnet, thus avoiding the secondary pollution and production of waste or requirement for filtration and centrifuging, will lead to a cost-effective and environmental-friendly wastewater treatment technique. Though this study was conducted in the laboratory and indicated satisfactory results, further studies should be conducted on a full-scale basis for PCW treatment.

4. Conclusion

MNPs@C magnetic composite was fabricated using chemical co-precipitation method. It was also characterized and used as a catalyst in the presence of H₂O₂ in the Fenton oxidation system. The catalyst showed good magnetic response and could rapidly and completely be recovered by an external magnet. Optimal conditions were obtained to be pH 3.0, 1.0 g L⁻¹ MNPs@C and 50 mM H₂O₂. A COD removal efficiency of more than 65% was still obtained after five consecutive cycles of reaction. It was found that in the FOP, the degradation of organic pollutants was due to the attack of ˙OH produced by the surface-catalyzed decomposition of hydrogen peroxide on the MNPs. The kinetics of COD removal was well fitted to the pseudo-second-order model. As a conclusion, MNPs@C maintained its activity after being reused five times and the leaching of iron from the catalyst was determined to be <0.3 g L⁻¹, suggesting that MNPs@C had good long-term stability. These results revealed that MNPs@C might have promising and effective application in the petrochemical wastewater treatment.

Acknowledgements

The authors are grateful for the financial support provided by the Environmental Technologies Research Center, Ahvaz Jundishapur University of Medical Sciences.

References

1. D. K. Yeruva, S. Jukuri, G. Velvizhi, A. N. Kumar, Y. Swamy and S. V. Mohan, Bioresour. Technol., 2015, 188, 33–42.
2. B. Diya'uddeen, S. R. Pouran, A. A. Aziz and W. Daud, RSC Adv., 2015, 5, 68159–68168.
3. R. Davarnejad, M. Mohammadi and A. F. Ismail, Journal of Water Process Engineering, 2014, 3, 18–25.
4. A. Buthiyappan, A. A. A. Raman and W. M. A. W. Daud, RSC Adv., 2016, 6, 25222–25241.
5. I. Arslan-Alaton, G. Tureli and T. Olmez-Hanci, Photochem. Photobiol. Sci., 2009, 8, 628–638.
6. A. Babuponnusami and K. Muthukumar, Chem. Eng. J., 2012, 183, 1–9.
7. P. Nidheesh, RSC Adv., 2015, 5, 40552–40577.
8. B. Kakavandi, A. Takdastan, N. Jaafarzadeh, M. Azizi, A. Mirzaei and A. Azari, J. Photochem. Photobiol., A, 2016, 314, 178–188.
9. F. Velichkova, C. Julcour-Lebigue, B. Koumanova and H. Delmas, J. Environ. Chem. Eng., 2013, 1, 1214–1222.
10. P. Maharaja, E. Gokul, N. Prabhakaran, S. Karthikeyan, R. Boopathy, S. Swarnalatha and G. Sekaran, RSC Adv., 2016, 6, 4250–4261.
11. N. Jaafarzadeh, B. Kakavandi, A. Takdastan, R. R. Kalantary, M. Azizi and S. Jorfi, RSC Adv., 2015, 5, 84718–84728.
12. L. Luo, Y. Yao, F. Gong, Z. Huang, W. Lu, W. Chen and L. Zhang, RSC Adv., 2016, 6, 47661–47668.
13. A. J. Jafari, B. Kakavandi, R. R. Kalantary, H. Gharibi, A. Asadi, A. Azari, A. A. Babaei and A. Takdastan, Korean J. Chem. Eng., 2016, 1–13,  DOI:10.1007/s11814-016-0155-x.
14. P. Nidheesh, R. Gandhimathi, S. Velmathi and N. Sanjini, RSC Adv., 2014, 4, 5698–5708.
15. S. Shin, H. Yoon and J. Jang, Catal. Commun., 2008, 10, 178–182.
16. B. Kakavandi, M. Jahangiri-rad, M. Rafiee, A. R. Esfahani and A. A. Babaei, Microporous Mesoporous Mater., 2016, 231, 192–206.
17. B. Kakavandi, A. Jonidi Jafari, R. Rezaei Kalantary, S. Nasseri, A. Esrafili, A. Gholizadeh and A. Azari, J. Chem. Technol. Biotechnol., 2016 DOI:10.1002/jctb.4925.
18. L. Zhou, J. Ma, H. Zhang, Y. Shao and Y. Li, Appl. Surf. Sci., 2015, 324, 490–498.
19. B. Kakavandi, A. Jonidi Jafari, R. Rezaei Kalantary, S. Nasseri, A. Ameri and A. Esrafily, Iran. J. Environ. Health Sci. Eng., 2013, 10, 1–9.
20. A. D. Eaton, L. S. Clesceri and A. E. Greenberg, Standard methods for the examination of water and wastewater, Washington, DC, 2005, pp. 20001–23710.
21. A. Azari, B. Kakavandi, R. R. Kalantary, E. Ahmadi, M. Gholami, Z. Torkshavand and M. Azizi, J. Porous Mater., 2015, 22, 1083–1096.
22. L. Huang, Y. Sun, W. Wang, Q. Yue and T. Yang, Chem. Eng. J., 2011, 171, 1446–1453.
23. F. Lücking, H. Köser, M. Jank and A. Ritter, Water Res., 1998, 32, 2607–2614.
24. P. Nidheesh and R. Rajan, RSC Adv., 2016, 6, 5330–5340.
25. A. Wang, J. Qu, H. Liu and J. Ru, Appl. Catal., B, 2008, 84, 393–399.
26. L. Xu and J. Wang, Appl. Catal., B, 2012, 123, 117–126.
27. Y. Kuang, Q. Wang, Z. Chen, M. Megharaj and R. Naidu, J. Colloid Interface Sci., 2013, 410, 67–73.
28. A. S. Giri and A. K. Golder, RSC Adv., 2014, 4, 6738–6745.
29. S. Karthikeyan, C. J. Magthalin, A. Mandal and G. Sekaran, RSC Adv., 2014, 4, 19183–19195.
30. N. Masomboon, C. Ratanatamskul and M.-C. Lu, J. Hazard. Mater., 2010, 176, 92–98.
31. S. Karthikeyan, R. B. Ahamed, M. Velan and G. Sekaran, RSC Adv., 2014, 4, 63354–63366.
32. L. Zhou, H. Zhang, L. Ji, Y. Shao and Y. Li, RSC Adv., 2014, 4, 24900–24908.
33. C. Zheng, X. Cheng, P. Chen, C. Yang, S. Bao, J. Xia, M. Guo and X. Sun, RSC Adv., 2015, 5, 40872–40883.
34. S.-T. Yang, W. Zhang, J. Xie, R. Liao, X. Zhang, B. Yu, R. Wu, X. Liu, H. Li and Z. Guo, RSC Adv., 2015, 5, 5458–5463.
35. Y. Segura, F. Martínez, J. A. Melero and J. L. G. Fierro, Chem. Eng. J., 2015, 269, 298–305.
36. J. Wei, Y. Song, X. Tu, L. Zhao and E. Zhi, Chem. Eng. J., 2013, 218, 319–326.
37. Y. Wang, R. Priambodo, H. Zhang and Y.-H. Huang, RSC Adv., 2015, 5, 45276–45283.
38. Q. Liao, J. Sun and L. Gao, Colloids Surf., A, 2009, 345, 95–100.
39. C. Bao, H. Zhang, L. Zhou, Y. Shao, J. Ma and Q. Wu, RSC Adv., 2015, 5, 72423–72432.
40. E. Garrido-Ramírez, B. Theng and M. Mora, Appl. Clay Sci., 2010, 47, 182–192.
41. C. Wang, H. Liu and Z. Sun, Int. J. Photoenergy, 2012, 2012, 1–10.
42. H. Wang, H. Jiang, S. Wang, W. Shi, J. He, H. Liu and Y. Huang, RSC Adv., 2014, 4, 45809–45815.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra17624k

---