Sonochemical degradation of bisphenol A using persulfate activated by hematite nanoparticles

In this study, hematite nanoparticles (HNPs) were used as the persulfate (PS) activator for the sonocatalytic degradation of bisphenol A (BPA). The physicochemical properties of the synthesized HNPs were investigated using X-ray diffraction (XRD), field emission scanning electron microscopy (FESEM), Fourier transform infrared spectroscopy (FTIR) and transmission electron microscopy (TEM). The effect of different operational parameters (pH, nanoparticle dosage, persulfate concentration, and ultrasonic power) on catalytic activity were evaluated. The BPA degradation rate was improved when ultrasonic (US) irradiation was used simultaneously with HNPs for activating PS. According to the results, 98.94% of BPA (10 mg/L) was degraded within 15 min of reaction time at 4 mM persulfate and 0.01 g/L HNPs under ultrasonic irradiation of 250 W. The degree of mineralization of BPA was measured using chemical oxygen demand (COD), and 36.98% was achieved under optimum conditions. Quenching tests were done using different scavenger compounds; these showed that both hydroxyl and sulfate radicals were reactive species in BPA degradation. According to the results of reusability tests, the degradation efficiency decreased to 86.34%, indicating that HNPs can be recycled several times. All of the anions tested, but mainly hydrogen phosphate, had an inhibitory effect on BPA degradation. The results showed that the US/HNPs/PS process is effective for the degradation of the organic pollutants.


GRAPHICAL ABSTRACT INTRODUCTION
Endocrine disrupting compounds (EDCs) contain a variety of toxic organic compounds (such as alkylphenol azoxylate, polychlorinated bisphenyls, insecticides, steroid sex hormones, polybromin compounds, phatalites, and bisphenol A), which are emerging contaminants in water resources (Staples et al. ; Vandenberg et al. ).
Bisphenol A (BPA; 4,4 0 -(propane-2,2-diyl)diphenol) is widely used as a monomer in the plastics industry to produce materials such as epoxy resins and polycarbonates; it is also used in various consumer products such as food packaging, PVC, and baby bottles. This pollutant is introduced into the environment by point and non-point sources (Staples et al. ; Vandenberg et al. ; Rochester ). Point sources include sewage sludge, landfill, raw sewage, and runoff. Non-point sources are polycarbonate plastics, residual epoxy resins, and synthetic polymers. BPA can penetrate underground water through industrial and agricultural wells (Staples et al. ; Huang et al. ). BPA levels reported in runoff and drinking water were 12 μg/L and 0.1 μg/L, respectively, and in some sources the detected concentrations were higher than 100 mg/L (Staples et al. ; Escalona et al. ). Very low levels of BPA lead to cardiovascular diseases, propagation of human prostate cancer cells, and type 2 diabetes (Mohapatra et al. ).
Different biological and chemical methods have been used to remove BPA from water and wastewater. Advanced oxidation processes (AOPs) are known to be the most effective method for degrading organic pollutants due to the production of reactive species such as hydroxyl radicals ( • OH) that have high oxidation potential (E 0 ¼ 2.8 V). These processes include ozonation, Fenton reaction, photocatalysis, electrooxidation and sonochemical processes ( Jiang et al. ; Escalona et al. ; Huang et al. ).
Nowadays, researchers have focused on sulfate-radicalbased AOPs for the degradation of refractory organic pollutants (Mokhtari et al. ; Huang et al. ). The sulfate radical (SO À 4 ), with a redox potential of 2.5-3.1 V, is one of the most commonly used oxidants and is produced by the activation of peroxymonosulfate (PMS, HSO À 5 ) or persulfate (PS, S 2 O 2À 8 ). The persulfate or peroxydisulfate ion attracts more attention than PMS, because it is cheaper, more stable and more suitable for water/wastewater decontamination (Lu et al. ). Activating agents such as transition metals, heat, ultraviolet (UV), and ultrasound (US), are applied to activate PS to produce sulfate radicals (Equation (1) Among these agents, the transition metals are the most widely used. Iron is the most commonly used metal, as it is more cost-effective than other metals and relatively nontoxic. It can activate persulfate (S 2 O 2À 8 ) to a sulfate radical, as shown in Equation (2) Heterogeneous catalytic processes (HCPs), using ironcontaining solids [hematite (iron(III) oxide, Fe 2 O 3 ), magnetite (Fe 3 O 4 ), etc.], have been considered to be efficient catalysts. HCPs have advantages, including low leaching of transition metals, lack of secondary pollution, and lack of sludge production. Hematite has been widely used as an efficient heterogeneous catalyst in the degradation of different pollutants. It is capable of activating persulfate and improving its oxidation potential (Lin & Zhang ; Kermani et al. ).
Recent researchers have focused on combinative or hybrid techniques. The use of ultrasonic waves can enhance the performance of the hybrid system through persulfate activation and the formation of H • and OH • radicals (Ai et al. ; Lu et al. ). During ultrasonic irradiation, the development, growth and collapse of bubbles provide high pressures and temperatures in the aqueous solution that may directly degrade organic pollutants (Mosleh et al. ). Ultrasound irradiation can enhance the reaction mass transfer in solid-liquid interphases. In addition, ultrasonic irradiation reduces the consumption of chemicals such as oxidants and catalysts (Guo et So far, studies have been carried out on the use of heterogeneous catalysts combined with UV and US irradiation as peroxymonosulfate and persulfate activators for the removal of different pollutants (Gobouri ; Liu et al. ; Tomić et al. ). Hematite nanoparticles (HNPs, α-Fe 2 O 3 ), as heterogeneous activators, can effectively activate persulfate for the production of sulfate free radicals that are strong oxidants for the degradation of pollutants. No studies are available on the sonocatalytic degradation of emerging pollutants (EPs) such as BPA with activation of persulfate by HNPs. In the present study, the activation of persulfate was performed simultaneously by US irradiation and HNPs. The catalytic performance of the US/HNPs/PS system for the degradation of BPA was investigated under different operational parameters (pH, catalyst dose, PS dose, reaction time and pollutant concentration). After the optimization of the process, mineralization and reusability were also examined.

Synthesis of hematite nanoparticles
HNPs were synthesized by the hydrothermal method described by Ning et al. (). Briefly, 2.2 g FeCl 3 .6H 2 O and 2 g urea were added to 50 mL distilled water. 25 mL of ammonium hydroxide, as the precipitating agent, was added dropwise with magnetic stirring. The solution was transferred into a Teflon-lined autoclave and then was heated at 150 C for 12 hours. After cooling at room temperature, the resulting precipitation was washed with distilled water and ethanol several times. Finally, it was calcinated at 450 C for 3 hours.

Characterization instruments
Scanning electron microscopy was performed to investigate the surface morphology of the nanoparticles using a field emission scanning electron microscopy (FESEM). To determine the crystalline structure of the synthesized nanoparticles, the XRD pattern was obtained using X-ray diffraction (XRD). The functional groups of hematite nanoparticles were determined by Fourier transform infrared spectroscopy (FTIR). Iron oxide nanoparticles were observed in electron micrographs obtained with transmission electron microscopy (TEM).

Experimental procedure
The experiments were done in 250 mL flask containing 100 mL of the sample with mechanical stirring at room temperature (25 ± 1 C). The BPA stock solution (500 mg/L) was prepared by dissolving given amounts of BPA in deionized water. At each stage of the experiment, hematite nanoparticles were added to the BPA solution and then a certain amount of sodium persulfate was added to the solution. The solution pH was adjusted by the addition of H 2 SO 4 (0.1 M) or NaOH (0.1 M). The process was performed in different operational conditions including pH (3-11), BPA concentration (10-50 mg/L), persulfate concentration (1-5 mM), hematite nanoparticle dosage (0.005-0.02 g/L), ultrasonic power (100-300 W), and 30 min of reaction time.
The samples were collected during the reaction at different time intervals. A small volume of the solution (5 mL) was taken and each sample was filtered with 0.22-μm syringe filters. 0.5 mL methanol was added as the scavenger of OH • and SO À 4 for terminating the oxidation reaction. After optimizing the process, to determine the dominant reactive species, quenching experiments were carried out in the presence of different scavenger compounds including methanol and tert-butyl alcohol. The mineralization of the US/HNPs/PS system based on chemical oxygen demand (COD) and the reusability of HNPs were also investigated. The kinetic models for BPA degradation by US/HNPs/PS system were examined.

Analytical methods
The pH measurements were done using a pH meter (Eutech 2700). The concentration of BPA was determined by highperformance liquid chromatography (HPLC) (KNAUER, Germany) with ultimate variable wavelength UV detector at 220 nm. The isocratic mobile phase consisted of acetonitrile and water with a ratio of 1:1 (50% Millipore water and 50% acetonitrile) at the flow rate of 1 mL/min À1 using a C18 column (Kromasil 100-5C18, 250 mm × 4.6 mm × 5 μm particle size). The retention time for BPA was 7.16-9.06 min. According to the calibration curve at different levels of BPA (0.1-50 mg/L), the R-squared was 0.9999. The limit of detection (LOD) and the limit of quantification (LOQ) were 0.01 mg/L and 0.1 mg/L, respectively.
The BPA removal efficiency (%) was calculated according to Equation (9): where C 0 and C t are initial and residual BPA concentrations (mg/L) at reaction time t (min) (Huang et al. ). COD was measured based on the standard methods for the examination of water and wastewater using a spectrophotometer (DR5000, HACH).

Characterization of hematite nanoparticles
The FESEM image of HNPs is shown in Figure 1(a), and shows that the HNPs are spherical and have a diameter of less than 100 nm. The approximate crystallite size of nanoparticles (D) was calculated using Scherer's equation (Equation (10)): where λ is the X-ray wavelength (1.540598 Å), β is the full width at half-maximum, and θ is the diffraction angle (Lassoued et al. ). The approximate crystallite size of HNPs was found to be in the range of 65.86-74.80 nm, which indicated that the surface of the hematite nanoparticles was almost porous and provided active sites for improving process efficiency. The size and morphology of the HNPs were specified by TEM imaging (Figure 1(b). The image shows a particle size of 10-25 nm, which indicates that the HNPs easily aggregate to form larger particles or chain structures.
The results of the FTIR analysis of the nanoparticles are shown in Figure 1(c). The FTIR spectra indicate absorption peaks around 450-600 cm À1 , 1,400-1,600 cm À1 and 3,440 cm À1 . Two bands at 471 cm À1 and 585 cm À1 were ascribed to the Fe-O stretching mode of hematite. The absorption peaks at 1,405 cm À1 , 1,622 cm À1 and 3,440 cm À1 were related to the stretching and bending vibration of OH groups.
For characterizing the phase and crystallization, XRD measurement was done for the HNPs (Figure 1(d)). The XRD patterns show sharp peaks at 2θ values of 25, 33.51, 35.96, 41.26, 49.81, 54.36, 58.61, 62.76 and 64.31 that correspond to 012, 104, 110, 113, 024, 116, 018, 214, and 300. The crystalline structure of the HNPs was identified using the standard XRD pattern of hematite (JCPDS Card,. No other crystalline phases of Fe 2 O 3 or impurities were observed based on the XRD pattern, which indicates the pure HNP phase.

Comparison of alternative processes for BPA degradation
The degradation of BPA was performed by different processes under the same experimental conditions. As can be seen (Figure 2), the BPA degradation rate in single systems is lower than binary and ternary systems due to increasing the formation of the reactive radicals ( • OH and SO À 4 ) in hybrid systems (Takdastan et al. ).
The results indicated that the removal efficiency of BPA was low under US irradiation alone (35.86%) and PS alone (53.54%). The removal efficiency using US/PS was 75.38%, whereas the US/HNPs and US/HNPs/PS systems could remove 84.60% and 98.87%, respectively.
Ultrasonic irradiation alone cannot provide a high degradation efficiency for organic contaminants. In the US/PS system, the heat induced by US can accelerate the decomposition of PS and facilitate the pollutant degradation (Lu et al. ). In the presence of Fe 2 O 3 nanoparticles as a catalyst, ultrasonic irradiation provides a larger HNP surface area and the cavitation increases. Persulfate is relatively stable at room temperature, but adding nanoparticles leads to the activation of persulfate. The sulfate radical (SO À 4 ) formed increases the process efficiency (Matzek & Carter ). The presence of a catalyst in the ultrasound system leads to an increase in the intensification of cavitational activity in the reactor and reduces the activation energy. The increase in the intensity of cavitation reduces the collapse pressure for a single cavity and increases the number of cavitation bubbles, thus improving the ultrasonic efficiency (Karim & Shriwastav ). When Fe 2 O 3 is added to the system, fixed cavitation bubbles in the ultrasonic field explode asymmetrically on its surface, increasing the cavitation effect (Liu et al. ). When the catalyst and PS were used together, the degradation rate of BPA increased. This indicates that HNPs have a high capability of activating PS (Jiang et al. ). The combination of US and the HNPs/PS system improved the reaction effectiveness.
In addition to sulfate radicals, hydroxyl radicals ( • OH) are also formed, which result in increased efficiency (Jiang et al. ). OH • radicals are extremely active and have a strong oxidation potential. In the sonochemical reactor, the decomposition of the water molecule takes place by sonochemical waves and produces hydroxyl radicals. In the presence of US irradiation alone, a very small amount of • OH is formed, which leads to low removal efficiency (Wang et al. ). With the addition of the HNPs as a sonocatalyst, the degradation rate is increased as more water molecules are decomposed and more • OH produced as a result of the increase in the number of cavitation bubbles formed on the sonocatalyst surface (Khataee et al. ).

The effect of solution pH
The solution pH plays an important role in oxidation processes because of its effect on the surface properties and the ionization degree of the pollutant and oxidant (Kermani et al. ). The effect of pH (in the range of 3-11) on the degradation of BPA by the US/HNPs/PS process is shown in Figure 3. As can be seen, increasing pH in alkaline conditions led to reduction in degradation efficiency. For pH values of 3.0, 5.0, 7.0, 9.0 and 11.0, the degradation efficiencies at the reaction time of 30 min were 95.74%, 83.6%, 80.95%, 71.1% and 63.18%, respectively. The best decomposition was obtained in acidic conditions (pH: 3.0).
At higher pH values, because of the catalyst deactivation caused by the formation of ferric hydroxide complexes, PS molecules were less activated by the catalyst and BPA degradation was decreased. Similar results have been presented in the other studies for the degradation of various pollutants (Ghauch et    In acidic conditions (pH < 7), sulfate radicals are dominant. They have a higher oxidation potential than hydroxyl radicals, therefore the removal efficiency is higher in acidic pH (Equations (11)-(13)) (Matzek & Carter ).
Acidic conditions : Alkaline conditions: SO À 4 þOH À !H þ þSO 2À 4 þOH (13) The pKa value of BPA (9.6 and 10.2) can be a reason for the increase in degradation efficiency under acidic conditions. At pH ¼ 3.0, BPA exists in its cationic state and becomes positive, which leads to a strong electrostatic repulsion between the positively charged BPA molecules and the negatively charged catalyst (Gültekin & Ince ; Gobouri ). Therefore, pH ¼ 3.0 was considered to be the optimal pH for the next steps of the experiments.

The effect of sodium persulfate concentration
The effect of different concentrations of PS (1-5 mM) on BPA degradation by US/HNPs was investigated. With increasing PS concentration (from 1 to 4 mM) at 15 min intervals, the removal efficiency was increased from 75.21% to 96.2% (Figure 4(a)). At higher PS dosages, more sulfate radicals were formed, resulting in higher degradation rates. These results are consistent with results from other studies (Monteagudo et al. ; Takdastan et al. ). However, the degradation rate was 75.8% at PS 5 mM. This was because of (a) the recombination of SO À 4 and OH • radicals (Equations (14) and (15)); (b) the reaction between excess persulfate and SO À 4 (Equation (16));  (17)) (Kermani et al. ; Monteagudo et al. ).

and (c) unproductive persulfate activation (Equation
The effect of catalyst dosage The optimization of the catalyst dosage is important because of forming a dispersed suspension in combination with organic pollutants. Different dosages of HNPs (0.005-0.02 g/L) were investigated for the removal efficiency of BPA. As can be seen (Figure 4(b)), for dosages of 0.005, 0.007, 0.01 and 0.02 g/L at the reaction time of 15 min, the degradation rates were 67.84%, 69.37%, 96.02% and 70.36%, respectively. According to the results, by the addition of various dosages of HNPs up to 0.01 g/L, the degradation rate of BPA increased, which can be attributed to an increase in the production of hydroxyl and sulfate radicals. At higher dosages, more active sites are provided that lead to the higher rates of PS activation and generation of sulfate radicals (Kermani et al. ). The maximum efficiency (96.2%) was achieved with the catalyst dosage 0.01 g/L.
At doses higher than 0.01 g/L, the degradation rate was decreased. This is attributed to lower exploitation of active sites on the catalyst surface. At higher catalyst dosages, the production of aggregates and subsequent self-binding of catalyst particles lead to the reduction of active sites (Kermani et al. ; Takdastan et al. ).

The effect of ultrasonic power
Figure 4(c) shows the effect of various levels of ultrasonic power on the degradation rate of BPA. As observed, BPA removal increased by increasing the ultrasonic power from 100 W (69.55%) to 300 W (98.02%). During the US irradiation, the tiny gas bubbles in water can form, grow and collapse (Hu & Long ). At higher ultrasonic power, the pulsation and collapse of bubbles increased, causing a higher generation of the cavitation bubbles and reactive radicals. At higher ultrasonic power, more turbulence improved the mass transfer in the solution (Gao et al. ; Monteagudo et al. ).
At the ultrasonic power levels higher than 250 W a very slight increase in BPA degradation efficiency was observed. Above an optimal or critical ultrasonic power amount, a cloud of the cavitation bubbles forms but hardly collapses. In addition, it disperses the formation of sound waves that lead to a little increase in removal efficiency (Camargo-Perea et al. ).
There was no significant difference between the removal efficiencies for the ultrasonic power of 250 W (97.64%) and 300 W (98.02%), so 250 W was selected as the optimum level.

Effect of the BPA initial concentration
The effect of various BPA concentrations on the removal efficiency was studied (Figure 4(d)). It can be seen that the removal efficiency decreased at concentrations higher than 10 mg/L. So for the initial concentration of 50 mg/L at a reaction time of 15 min, the removal efficiency was 48.27%. The maximum removal efficiency (98.64%) was obtained for the BPA concentration of 10 mg/L. The production of less hydroxyl radicals at higher concentrations and subsequently, the decreasing reactions between BPA and OH • radicals results in a decrease of the degradation rate. In addition, at high concentrations, there may occur a competitive reaction for OH • and SO 4 •À radicals between BPA and the intermediates (Li et al. ).

Kinetics for sonocatalytic activity of HNPs
In order to determine the reaction rate, the pseudo-firstorder kinetic model was studied. Based on the results ( Figure 5), under the optimal conditions, the BPA degradation was fitted by Equation (18).
where C 0 and C are the initial concentration and concentration of BPA at time t, t is the reaction time (min) and k obs (the slope of obtained from the plot of LnC/C 0 versus elapsed time) is the rate constant (min À1 ) (Li et al. ).
The results showed that by increasing the initial concentration of, the pseudo-first-order kinetic rate of BPA degradation decreased. This may have been due to the production of less reactive species at the higher concentrations. The constant values of the pseudo-first-order kinetic model of BPA are presented in Table 1. By increasing the initial concentration of BPA from 10 to 50 mg/L, the k obs rate was decreased from 0.2485 to 0.043.

Mineralization
The BPA mineralization (based on COD reduction) in the US/HNPs/PS process was determined under optimized conditions (pH: 3.0, PS concentration: 4 Mm, initial BPA concentration: 10 mg/L, catalyst dosage: 0.01 g/L, US power: 250 W). According to the results (Figure 6), the mineralization degree was 36.98% after 15 min. The organic chemical reaction with radicals leads to partial mineralization. In the study by Kermani et al. (), the mineralization degree was 36% for the degradation of both 2,4-D and MCPA herbicides using HNPs/PS. They stated that low efficiency can be due to the difficulty of completely removing the by-products formed, which requires more oxidation (Kermani et al. ).

Reusability of the catalyst
The reusability of the catalyst is one of the key factors for the development and application of a catalyst at industrial scale.
In order to evaluate the possibility of catalyst reuse, experiments were done for three consecutive cycles under the same conditions. For this purpose, HNPs were separated by a centrifuge and then washed with double distilled water several times and dried after each run. The obtained results are presented in Figure 7. The removal efficiency decreased after three cycles. The findings showed the removal efficiency of 98.64% at the first run. The removal value reached to 86.34% after the third run. This slight decrease could be because of a reduction in the catalyst adsorption capacity and HNP mass loss on the surface of the catalyst (Xu et al. ). Thus, HNPs have good reusability for several runs.

Effect of scavengers
To identify the dominating radical species, quenching experiments were done using two organic compounds. Methanol (MeOH) and tert-butyl alcohol (TBA) have been extensively employed previously as scavengers of SO 4 •À and OH • radicals due to their different reactivity with radicals (MeOH: k SO4 •À þMeOH ¼ 3.2 × 10 6 M À1 s À1 and k OH • þMeOH ¼ 9.7 × 10 8 M À1 s À1 , TBA: k SO4 •À þTBA ¼ 4 × 10 5 M À1 s À1 and k OH In the absence of the scavengers, 98.64% of BPA was removed. According to the results (Figure 8), in the presence of methanol, the degradation rate was reduced to 69.15%. The decrease in removal efficiency in the presence of methanol as a hydroxyl and sulfate radical scavenger confirmed that sulfate radicals might be the active species. In the case of TBA addition, the removal rate was 63.65%. TBA has a good reactivity with hydroxyl radicals and is a specific scavenger for OH • . The decreasing BPA degradation in the presence of TBA indicates that the OH • may be one of the radicals involved (Gu et al. ; Oh et al. ). The results indicated that both SO À 4 and OH • radicals were responsible for BPA degradation using the US/HNPs/PS process.

Effect of anions
The presence of anions in natural aquatic environments affects the catalytic performance of AOPs. Therefore, the effect of anions such as chloride (Cl À ), sulfate (SO 2À 4 ) and hydrogen phosphate (HPO 2À 4 ) were evaluated on the degradation of BPA by the US/HNPs/PS system (Figure 9). In the absence of anions, the removal efficiency was 98.64%. The removal efficiencies were reduced to 83.58%, 61.6% and 51.81% in the presence of Cl À , SO 2À 4 and HPO 2À 4 , respectively. The inhibition of the anions was in order of Cl À < SO 2À 4 < HPO 2À 4 . The anions act as a scavengers of free radicals and form weaker radical species, which is how they reduce the removal efficiency of the pollutant (Golshan et al. ).
The reason for the slight inhibitory effect of the chloride ion is the reaction between the sulfate radical and the chloride ion that leads to the production of the chlorine radical. The generation of active chlorine species during the reactions of free radicals with the chloride ion are based on Equations (19)-(22). The chlorine radicals (Cl • , Cl 2 • ) have lower redox potentials in comparison to sulfate and hydroxyl radicals (Golshan et al. ).
Cl þ Cl À ! Cl À 2 (20) The low inhibiting effect of the sulfate ion (SO 2À 4 ) can be attributed to the scavenging of hydroxyl radicals by the sulfate ions and the formation of the inevitable product of sulfate radicals in this system (Equation (23)  Hydrogen phosphate (HPO 2À 4 ) has a greater inhibitory effect than sulfate and chloride ions due to the production of the hydrogen phosphate radical by electron transfer (Equation (24)). Also, these ions act as sulfate radical scavengers and reduce active sites on the catalyst by blocking the pores (Gao et

CONCLUSION
In this study, the catalytic application of HNPs for persulfate activation in the presence of US was investigated. US irradiation improved the performance of HNPs/PS in BPA degradation. The maximum removal efficiency was achieved at pH 3, PS concentration of 4 mM, catalyst dosage of 0.01 g/L and US power of 250 W at 15 min. Further increase in the ultrasonic power from the optimal amount leads to slight changes in degradation efficiency. In general, in terms of the ultrasonic power, typically optimum levels of ultrasonic power are recommended. The quenching tests indicated that both OH • and SO 4 •À radicals may be the oxidizing species involved in the BPA degradation in the US/HNPs/PS process. The catalyst showed good reusability after three cycles. The mineralization rate under optimum experimental conditions was 36.98%. Results showed an inhibition in the degradation of BPA for all studied anions. Chloride and sulfate played a minor role in the decrease of efficiency in comparison with hydrogen phosphate. The kinetics of the catalytic oxidation of BPA by US/HNPs/PS was in good agreement with the pseudo-first-order model. Hematite nanoparticles (α-Fe 2 O 3 ) are reusable and efficient nanoparticles for the catalytic degradation of BPA.