Removal of Amoxicillin from Aqueous Media by Fenton-like Sonolysis/H2O2 Process Using Zero-Valent Iron Nanoparticles

High concentrations of antibiotics have been identified in aqueous media, which has diminished the quality of water resources. These compounds are usually highly toxic and have low biodegradability, and there have been reports about their mutagenic or carcinogenic effects. The aim of this study was to apply zero-valent iron-oxide nanoparticles in the presence of hydrogen peroxide and the sonolysis process for the removal of the amoxicillin antibiotic from aqueous media. In this study, zero-valent iron nanoparticles were prepared by an iron chloride reduction method in the presence of sodium borohydride (NaBH4), and the obtained nanoparticles were characterized by scanning electron microscopy (SEM), transmission electron microscopy (TEM), X-ray diffraction (XRD), and vibrating-sample magnetometry (VSM). Then, using a Fenton-like process, synthetic wastewater containing 100 to 500 mg/L amoxicillin antibiotic was investigated, and the effects of different parameters, such as the frequency (1 and 2 kHz), contact time (15 to 120 min), the concentration of hydrogen peroxide (0.3%, 0.5%, and 6%), the dose of zero-valent iron nanoparticles (0.05, 0.1, 0.5 g/L), and pH (3, 5, 10) were thoroughly studied. A pH of 3, hydrogen peroxide concentration of 3%, ultrasonic-wave frequency of 130 kHz, zero-valent iron nanoparticles of 0.5 g/L, and contaminant concentration of 100 mg/L were obtained as the optimal conditions of the combined US/H2O2/nZVI process. Under the optimal conditions of the combined process of zero-valent iron nanoparticles and hydrogen peroxide in the presence of ultrasonic waves, a 99.7% removal efficiency of amoxicillin was achieved in 120 min. The results show that the combined US/H2O2/nZVI process could be successfully used to remove environmental contaminants, including antibiotics such as amoxicillin, with a high removal percentage.


Introduction
Drugs and related products have been considered a major concern in recent years and have drawn a lot of attention. Drugs treat or prevent microbial infections in humans and even animals. Pharmaceutical compounds such as antibiotics are detected in aqueous environments. These compounds are present in surface water, groundwater, wastewater, and even drinking water. Pharmaceutical compounds can be released into the environment Fe 0 + 2H + + Ultrasonic → Fe 2+ + H 2 (in acid solution) (1)  (9) Under optimum conditions, the generated ferric ions (Reaction 2) can react with the excess hydrogen peroxide (Reaction 5) or water (Reaction 7). Ferric ions also react with oxidase (Reaction 10) or metallic iron (Reaction 11) when they are converted to ferrous species, and this cycle continues. R • + Fe 3+ → R + + Fe 2+ (10) → 3 Fe 2+ Fe + 2 Fe 3+ (11) The above point clearly indicates that Fenton reactions are limited to the presence of iron cations in the intermediate reaction, as well as the ability of the intermediate reaction to regenerate and continuously produce the solute iron species. Research has shown that ultrasonic waves play a critical role in the dissolution of iron particles in a medium; therefore, the addition of zero-valent iron (for the generation of ferrous and ferric ions) and hydrogen peroxide (for the production of hydroxyl radical ions) and ultrasonic irradiation induce complex redox reactions, which in turn cause the degradation of the antibiotic [22]. AMX + • OH/ • O 2 H/other active radicals → (oxidized AMX) → Harmless species (XH 2 O 2 , intermediates Yco 2 , etc.) (12) In light of the issue of amoxicillin in hospital effluents and water resources and, ultimately, its effects on human health, the goal of the current study was to synthesize zero-valent iron nanoparticles, determine their characteristics, and combine them with hydrogen peroxide along with ultrasonic waves to investigate amoxicillin removal from aqueous solutions.

Chemicals and Reagents
The amoxicillin antibiotic and methanol were obtained from Sigma Aldrich (Burghausen, Germany), and hydrochloric acid was obtained from Scharlau Chemie (Barcelona, Spain). Hydrogen peroxide 30%, sulfuric acid 98%, sodium hydroxide, potassium iodide, sodium thiosulfate, and starch were purchased from CMC (Frankfurt, Germany). Whatman filters with a pore size of 1.5 microns were obtained from Prolabo (Paris, France) to filter the samples.

Synthesis of Zero-Valent Iron Nanoparticles
Zero-valent iron nanoparticles were synthesized according to Ibrahem et al.'s method as follows [23]: Iron nanoparticles were synthesized by mixing a ferric chloride (FeCl 3 ) solution with a solution containing a reducing agent. The ferric chloride solution was prepared by dissolving 0.4 g of ferric chloride in 50 mL of deionized water. Then, in order to prepare the reducing agent solution, 0.6 g of sodium borohydride (NaBH 4 ) was dissolved in 100 mL of deionized water. This solution (containing a reducing agent) was immersed in an ice bath before the ferric chloride solution was added dropwise with a flow rate of 0.13 mL/s using a microtube pump ( Figure 1). It is worth mentioning that when no bubbles are produced for a long time, the reaction is considered complete. The final product was a black sludge, which was separated by a strong magnet. Then, the zero-valent iron nanoparticles were washed several times with distilled water and acetone to remove the residues and impurities. Finally, the synthesized nanoparticles were stored in a desiccator under nitrogen gas (oxygen-free conditions to prevent oxidation).

Characterization of Zero-Valent Iron Nanoparticles
The characteristics, morphology, and size of nZVI were analyzed using a scanning electron microscope (SEM; model KyKy-EM3200) and transmission electron microscope (TEM; model PHILIPS, EM, 100 Kv). The crystal structure of the nanoparticles was also determined by X-ray diffraction (XRD) using an XRD 3100 diffractometer (Eindhoven, Netherlands, Philips model). Moreover, the magnetic properties of the nanoparticles were determined using a 7410 Shore Lake vibrating-sample magnetometer.

Experimental Section
In this study, specific doses of zero-valent iron nanoparticles (0.05, 0.1, and 0.5 g/L), different initial concentrations of AMX (100, 300, and 600 mg/L), and certain concentrations of hydrogen peroxide (0.3%, 1%, and 2%) were added to the reactor. The reaction times were varied (15,30,60,90, and 120 min), and pH was set to three values (3, 5, and 11) [24]. To produce ultrasonic waves, an ultrasonic bath was used, which was capable of generating ultrasonic waves at frequencies of 35 and 130 kHz with a power of 500 W Elma Schmidbauer GmbH, Singen, Germany). Samples containing zero-valent iron nanoparticles were passed through a 0.2 μm filter at the end of the reaction. To attain the calibration curve, solutions with specific concentrations of amoxicillin were injected into the Shimadzu HPLC apparatus (Kyoto, Japan) to be detected, and by calculating the peak area of the resulting peaks, the calibration curve was plotted to determine the unknown concentrations. After determining the residual concentration of amoxicillin in the samples, the removal efficiency of the different conditions, as well as the optimum reaction conditions, were determined. In order to determine the hydrogen peroxide in the output samples and to assay the hydrogen peroxide used in the research, the potassium iodide method was used [25].

Characterization of Zero-Valent Iron Nanoparticles
The characteristics, morphology, and size of nZVI were analyzed using a scanning electron microscope (SEM; model KyKy-EM3200) and transmission electron microscope (TEM; model PHILIPS, EM, 100 Kv). The crystal structure of the nanoparticles was also determined by X-ray diffraction (XRD) using an XRD 3100 diffractometer (Eindhoven, Netherlands, Philips model). Moreover, the magnetic properties of the nanoparticles were determined using a 7410 Shore Lake vibrating-sample magnetometer.

Experimental Section
In this study, specific doses of zero-valent iron nanoparticles (0.05, 0.1, and 0.5 g/L), different initial concentrations of AMX (100, 300, and 600 mg/L), and certain concentrations of hydrogen peroxide (0.3%, 1%, and 2%) were added to the reactor. The reaction times were varied (15,30,60,90, and 120 min), and pH was set to three values (3, 5, and 11) [24]. To produce ultrasonic waves, an ultrasonic bath was used, which was capable of generating ultrasonic waves at frequencies of 35 and 130 kHz with a power of 500 W Elma Schmidbauer GmbH, Singen, Germany). Samples containing zero-valent iron nanoparticles were passed through a 0.2 µm filter at the end of the reaction. To attain the calibration curve, solutions with specific concentrations of amoxicillin were injected into the Shimadzu HPLC apparatus (Kyoto, Japan) to be detected, and by calculating the peak area of the resulting peaks, the calibration curve was plotted to determine the unknown concentrations. After determining the residual concentration of amoxicillin in the samples, the removal efficiency of the different conditions, as well as the optimum reaction conditions, were determined. In order to determine the hydrogen peroxide in the output samples and to assay the hydrogen peroxide used in the research, the potassium iodide method was used [25].
Removal of the interference of hydrogen peroxide in the COD assay: Hydrogen peroxide interferes in the COD test as follows [26].
Although hydrogen peroxide is volatile on its own, its volatility can be reduced in synthetic samples by adding some substances. Consequently, the best way to eliminate the interference of hydrogen peroxide is to measure its residual concentration in the output samples and correct the COD. However, measurements of hydrogen peroxide in the output samples revealed that the residual amount of hydrogen peroxide was not significant. To remove hydrogen peroxide from the environment, different kinds of methods, including the usage of sodium thiosulfate or sodium borohydride, boiling, and incubation in a water bath at a temperature of 45 • C for 24 h, were considered [27].

Analysis of Samples
After the preparation of input and output samples of the reactor (centrifugation and passage through a 0.2 µm filter), the samples were injected into the HPLC device to determine the removal efficiency of amoxicillin. The amoxicillin antibiotic concentration was determined by HPLC (HPLC, Shimadzu, Kyoto, Japan, LC 10 A HPLC) equipped with a UV detector (SPD-10 AV) at 420 nm using a mobile phase with a volume percentage of 95:5 of potassium dihydrogen phosphate and methanol. The peak corresponding to the standard graph obtained from the HPLC device is shown in Figure 2. Equation (16) shows the removal efficiency of AMX: where C i and Co are the initial and residual concentrations of AMX (mg/L).
Molecules 2022, 27, 6308 5 of 14 the interference of hydrogen peroxide is to measure its residual concentration in the output samples and correct the COD. However, measurements of hydrogen peroxide in the output samples revealed that the residual amount of hydrogen peroxide was not significant. To remove hydrogen peroxide from the environment, different kinds of methods, including the usage of sodium thiosulfate or sodium borohydride, boiling, and incubation in a water bath at a temperature of 45 °C for 24 h, were considered [27].

Analysis of Samples
After the preparation of input and output samples of the reactor (centrifugation and passage through a 0.2 μm filter), the samples were injected into the HPLC device to determine the removal efficiency of amoxicillin. The amoxicillin antibiotic concentration was determined by HPLC (HPLC, Shimadzu, Kyoto, Japan, LC 10 A HPLC) equipped with a UV detector (SPD-10 AV) at 420 nm using a mobile phase with a volume percentage of 95:5 of potassium dihydrogen phosphate and methanol. The peak corresponding to the standard graph obtained from the HPLC device is shown in Figure  2. Equation (16) shows the removal efficiency of AMX: where Ci and Co are the initial and residual concentrations of AMX (mg/L). The measurement of the final concentration of the amoxicillin antibiotic was the main criterion for evaluating the process. Test number 5220B in the book of standard methods was used to measure the COD parameter in the input and output samples of the reactor in order to detect the conversion rate and the removal efficiency of the amoxicillin contaminant. All experiments were performed according to the instructions of the Standard Methods for the Examination of Water and Wastewater book [28]. The experiments were repeated three times, and the average results are reported. The measurement of the final concentration of the amoxicillin antibiotic was the main criterion for evaluating the process. Test number 5220B in the book of standard methods was used to measure the COD parameter in the input and output samples of the reactor in order to detect the conversion rate and the removal efficiency of the amoxicillin contaminant. All experiments were performed according to the instructions of the Standard Methods for the Examination of Water and Wastewater book [28]. The experiments were repeated three times, and the average results are reported.  Figure 3a, peaks 42 and 66 indicate the formation of nZVI nanoparticles, which is in accordance with the study by Dong et al. [29]. The VSM analysis was used to study the magnetic behavior of nZVI, which was measured at 65 emu/g, and TEM images revealed that spherical nanoparticles were in a chain-like form with a rough surface. Moreover, the sizes of nanoparticles were 25-50 nm, according to SEM.

Properties of the Synthesized Nanoparticles
Figure 3a-d show the TEM, XRD, VSM, and SEM analysis results of the synthesized nanoparticles. The XRD patterns were drawn for angles of 30-90. According to Figure 3a, peaks 42 and 66 indicate the formation of nZVI nanoparticles, which is in accordance with the study by Dong et al. [29] The VSM analysis was used to study the magnetic behavior of nZVI, which was measured at 65 emu/g, and TEM images revealed that spherical nanoparticles were in a chain-like form with a rough surface. Moreover, the sizes of nanoparticles were 25-50 nm, according to SEM.

Investigation of the Effect of Different Dosages of Iron Nanoparticles
In this study, to evaluate the effect of the zero-valent iron nanoparticle dosage, values of 0.05, 0.1, and 0.5 g/L were selected. The obtained results are presented in Figure  4. Under these conditions, the amoxicillin concentration, hydrogen peroxide concentration, frequency, and pH considered were 100 mg/L, 1%, 130 kHz, and 5, respectively. The results showed that the removal efficiency increased with the increasing dosage of nanoparticles, so the highest efficiency was observed at a dosage of 0.5 g/L.

Investigation of the Effect of Different Dosages of Iron Nanoparticles
In this study, to evaluate the effect of the zero-valent iron nanoparticle dosage, values of 0.05, 0.1, and 0.5 g/L were selected. The obtained results are presented in Figure 4. Under these conditions, the amoxicillin concentration, hydrogen peroxide concentration, frequency, and pH considered were 100 mg/L, 1%, 130 kHz, and 5, respectively. The results showed that the removal efficiency increased with the increasing dosage of nanoparticles, so the highest efficiency was observed at a dosage of 0.5 g/L.

Determination of the Optimum Concentration of Hydrogen Peroxide
Hydrogen peroxide is a parameter that influences the Fenton process. In this study, the effect of hydrogen peroxide on the oxidation of the amoxicillin antibiotic was investigated. In order to achieve the best process efficiency, the contaminant was exposed to different concentrations of hydrogen peroxide (0.05, 1%, and 3%). As shown in Figure 5, the maximum removal efficiency of 95% was obtained at a concentration of 3% and after 120 min.

Determination of the Effect of pH Changes
The effect of initial pH values of 3, 5, and 11 was investigated in the presence of 100 mg/L amoxicillin, 3% hydrogen peroxide, ultrasonic waves of 130 kHz, and 0.5 g/L iron nanoparticles. As shown in Figure 6, the efficiency of the nano-Fenton process decreased as the pH increased, so the maximum removal efficiency was observed at pH = 3.

Determination of the Optimum Concentration of Hydrogen Peroxide
Hydrogen peroxide is a parameter that influences the Fenton process. In this study, the effect of hydrogen peroxide on the oxidation of the amoxicillin antibiotic was investigated. In order to achieve the best process efficiency, the contaminant was exposed to different concentrations of hydrogen peroxide (0.05, 1%, and 3%). As shown in Figure  5, the maximum removal efficiency of 95% was obtained at a concentration of 3% and after 120 min.

Determination of the Effect of pH Changes
The effect of initial pH values of 3, 5, and 11 was investigated in the presence of 100 mg/L amoxicillin, 3% hydrogen peroxide, ultrasonic waves of 130 kHz, and 0.5 g/L iron nanoparticles. As shown in Figure 6, the efficiency of the nano-Fenton process decreased as the pH increased, so the maximum removal efficiency was observed at pH = 3.

Determination of the Optimum Concentration of Hydrogen Peroxide
Hydrogen peroxide is a parameter that influences the Fenton process. In this study, the effect of hydrogen peroxide on the oxidation of the amoxicillin antibiotic was investigated. In order to achieve the best process efficiency, the contaminant was exposed to different concentrations of hydrogen peroxide (0.05, 1%, and 3%). As shown in Figure  5, the maximum removal efficiency of 95% was obtained at a concentration of 3% and after 120 min.

Determination of the Effect of pH Changes
The effect of initial pH values of 3, 5, and 11 was investigated in the presence of 100 mg/L amoxicillin, 3% hydrogen peroxide, ultrasonic waves of 130 kHz, and 0.5 g/L iron nanoparticles. As shown in Figure 6, the efficiency of the nano-Fenton process decreased as the pH increased, so the maximum removal efficiency was observed at pH = 3.   Figure 7 shows the removal efficiency of the amoxicillin antibiotic at two ultrasonic frequencies (35 and 130 kHz) in the presence of 100 mg/L amoxicillin, 3% hydrogen peroxide, and 0.5 g/L iron nanoparticles. The reaction time in these conditions was 0-120 min, and the results indicated that the process showed more efficiency at the 130 kHz frequency.  Figure 7 shows the removal efficiency of the amoxicillin antibiotic at two ultrasonic frequencies (35 and 130 kHz) in the presence of 100 mg/L amoxicillin, 3% hydrogen peroxide, and 0.5 g/L iron nanoparticles. The reaction time in these conditions was 0-120 min, and the results indicated that the process showed more efficiency at the 130 kHz frequency.

Effect of Different Concentrations of Amoxicillin Antibiotic
The effect of different concentrations of amoxicillin was also studied in the presence of nZVI = 0.5 g/L, H2O2 = 3%, pH = 3, and frequency = 130 kHz. The obtained results are shown in Figure 8.

Effect of Different Concentrations of Amoxicillin Antibiotic
The effect of different concentrations of amoxicillin was also studied in the presence of nZVI = 0.5 g/L, H 2 O 2 = 3%, pH = 3, and frequency = 130 kHz. The obtained results are shown in Figure 8.

Determination of the Efficiency of the Process in the Absence of One of the Parameters
In this study, the efficiency of the US/H2O2/nZVI process was investigated. In order to study the effect of the combination of these processes on amoxicillin removal, the efficacy of each of these parameters was examined individually, and the results were compared with those of the combined process. The results are presented in Figure 9. As can be seen in all of the figures, these experiments were performed using US, nZVI, and H2O2 individually and the combination of each of them, and the results were compared with those of the combined method.

Determination of the Efficiency of the Process in the Absence of One of the Parameters
In this study, the efficiency of the US/H 2 O 2 /nZVI process was investigated. In order to study the effect of the combination of these processes on amoxicillin removal, the efficacy of each of these parameters was examined individually, and the results were compared with those of the combined process. The results are presented in Figure 9. As can be seen in all of the figures, these experiments were performed using US, nZVI, and H 2 O 2 individually and the combination of each of them, and the results were compared with those of the combined method.
to study the effect of the combination of these processes on amoxicillin removal, the efficacy of each of these parameters was examined individually, and the results were compared with those of the combined process. The results are presented in Figure 9. As can be seen in all of the figures, these experiments were performed using US, nZVI, and H2O2 individually and the combination of each of them, and the results were compared with those of the combined method.

Determination of the System Performance in COD Removal
The efficiency of the nano-Fenton process in removing the COD of the amoxicillin antibiotic was also investigated. To measure the COD, the reflex method was used according to Experiment No. 5220B of the Standard Methods Book [28]. The obtained results are presented in Figure 10.

Determination of the System Performance in COD Removal
The efficiency of the nano-Fenton process in removing the COD of the amoxicillin antibiotic was also investigated. To measure the COD, the reflex method was used according to Experiment No. 5220B of the Standard Methods Book [28]. The obtained results are presented in Figure 10.

Discussion
Investigation of the effect of different dosages of iron nanoparticles. The highest removal of AMX obtained was 72% using 0.5 g/L zero-valent iron nanoparticles. The increase in the adsorbent surface or its availability to antibiotic molecules, which was obtained by increasing the adsorbent dose, increases the removal percentage. [30] In Ali et al.'s study, the application of a green approach using pomegranate peel coated with zero-valent iron nanoparticles was used for the removal of amoxicillin. They reported that the removal efficiency increased with an increase in the adsorbent dosage (0.25-2.5 g/L) [31].The same results were obtained by Antoine Ghauch et al. for carbamazepine removal using the combined ultrasonic/Fe/H2O2 process [13]. Zhang et al. also found that the system's efficiency increased with the increase in zero-valent iron nanoparticles [16].
Determination of the optimum dose of hydrogen peroxide. In order to achieve the best performance in amoxicillin removal, the contaminant was exposed to different doses of hydrogen peroxide (0.5%, 1%, and 3%). According to the results, the highest removal of 95% was achieved at a concentration of 3% and with the maximum time. Studies have

Discussion
Investigation of the effect of different dosages of iron nanoparticles. The highest removal of AMX obtained was 72% using 0.5 g/L zero-valent iron nanoparticles. The increase in the adsorbent surface or its availability to antibiotic molecules, which was obtained by increasing the adsorbent dose, increases the removal percentage [30]. In Ali et al.'s study, the application of a green approach using pomegranate peel coated with zero-valent iron nanoparticles was used for the removal of amoxicillin. They reported that the removal efficiency increased with an increase in the adsorbent dosage (0.25-2.5 g/L) [31].The same results were obtained by Antoine Ghauch et al. for carbamazepine removal using the combined ultrasonic/Fe/H 2 O 2 process [13]. Zhang et al. also found that the system's efficiency increased with the increase in zero-valent iron nanoparticles [16].
Determination of the optimum dose of hydrogen peroxide. In order to achieve the best performance in amoxicillin removal, the contaminant was exposed to different doses of hydrogen peroxide (0.5%, 1%, and 3%). According to the results, the highest removal of 95% was achieved at a concentration of 3% and with the maximum time. Studies have shown that if excess hydrogen peroxide remains in the medium, the hydrogen peroxide itself reacts with the hydroxyl radicals, which in turn reduces the removal efficiency of the system [32]. Due to the improved efficiency of the system at 3% hydrogen peroxide, it can be confirmed that no excess hydrogen peroxide was present in the medium. The findings of Antoine Ghauch et al. also showed that increasing hydrogen peroxide to optimal concentrations would improve the system's efficiency [13].
Determination of the effect of pH changes. According to the diagrams and curves shown, the amount of contaminant removal increased dramatically at pH = 3 to 99.7%. However, the amount of contaminant removal decreased significantly at pH = 11. These results indicate that pH plays a critical role in system performance. The higher removal efficiency under acidic conditions for amoxicillin can be attributed to the generation of more OH radicals, which can rapidly oxidize the reactive AMX molecule [33]. Zero-valent iron nanoparticles have better solubility in acidic environments compared to alkaline conditions. Under alkaline conditions, zero-valent iron nanoparticles aggregate and agglomerate, which significantly reduces the efficiency of the system as well as the contact between the particles and hydrogen peroxide. Advanced oxidation processes perform better in acidic environments than in alkaline environments [34]. Studies by Catalkaya and Sengul also confirm these results [35].
Determination of the effect of frequency changes on system performance. The system performance was examined at a frequency of 5 kHz, and the results showed that the removal rate of the system was 90% at a frequency of 35 kHz. The findings demonstrate that changing the frequency from 130 kHz to 35 kHz at constant power has no significant effect on system performance; however, in this case, the efficiency of the system was reduced. The obtained results are consistent with the findings of Manousaki et al. [36].
Determination of the efficiency of the system at different concentrations of amoxicillin antibiotic. The results of evaluating the efficiency of the nano-Fenton process at different concentrations of amoxicillin showed that the removal rate was equal to 77.4%, 84.4%, and 99.7% at concentrations of 600 ppm, 300 ppm, and 100 ppm, respectively (Diagram 7). Since the production of hydroxyl radicals will remain constant under constant conditions, they will be consumed by increasing the amount of the contaminant in the radical environment. Therefore, the efficiency of the process will decrease with increasing concentrations of contaminants [37]. The results of Seid-Mohammadi's study showed that the removal efficiency in the combined US/H 2 O 2 /NiO process decreases with the increase in the concentration of antibiotics [38]. Hung-Yee Shu et al. found that the process efficiency decreases with increasing contaminant concentration [39].
Determination of the efficiency of the process in the absence of one of the parameters. In order to evaluate the effect of each factor used in the combined system, under optimum conditions, 100 mg/L of the contaminant was put into contact with ultrasonic waves alone, zero-valent iron nanoparticles alone, hydrogen peroxide alone, and the combination of two parameters in the absence of one of the parameters. The results showed that in the absence of ultrasonic waves, hydrogen peroxide, or zero-valent iron nanoparticles, even under acidic conditions, the system performance was significantly reduced. This difference in the system performance can be clearly observed by comparing the diagrams of US/H 2 O 2 /nZVI and H 2 O 2 /nZVI, which shows the critical role of ultrasonic waves in improving the efficiency of the system. Ultrasonic waves enable better mixing of zero-valent iron nanoparticles and more contact with hydrogen peroxide. In addition, ultrasonic waves increase the contact surface of zero-valent iron nanoparticles with hydrogen peroxide by creating tiny bubbles (cavitation). On the other hand, the absence of hydrogen peroxide in the medium results in the removal of hydroxyl radical sources (although ultrasonic waves are somewhat capable of generating hydroxyl radicals from water molecules, the amount of the produced hydroxyl radical is not enough). The absence of zero-valent iron nanoparticles, which are considered a major source of electrons, which are donated to hy-drogen peroxide and produce hydroxide radicals, can cause the disruption and diminished removal efficiency of the amoxicillin antibiotic. The obtained results confirm the necessity of the presence of each of these parameters in the process (ultrasonic waves, hydrogen peroxide, and zero-valent iron nanoparticles). Determination of the system performance in COD removal. The COD test was performed to evaluate the efficiency of the system in converting the amoxicillin antibiotic to water and carbon dioxide, which was investigated in this study. The COD measurement was performed using the open reflex method according to Experiment No. B5220 of the Standard Methods Book [28]. As shown in the relevant diagram, the system's efficiency in removing COD in the presence of the three parameters used in the system is significantly increased. The findings also illustrate that the system used in this study is not capable of the complete removal of COD from the solution, which indicates that the contaminant may be converted to other substances, and therefore, further investigation is needed. Similar results were obtained in the study by Emad S. Elmolla et al. on COD removal [40].

Conclusions
The aim of this study was to investigate the application of zero-valent iron oxide nanoparticles in the presence of hydrogen peroxide and the sonolysis process in the removal of the amoxicillin antibiotic from aqueous media. Zero-valent iron nanoparticles were successfully synthesized using the reduction method, and the results showed that the maximum system efficiency occurred at lower contaminant concentrations, under acidic conditions, in the presence of 0.5 g/L zero-valent iron nanoparticles and 3% hydrogen peroxide, and with maximum contact time. The results also indicate that the pH of the medium has a significant effect on increasing the system's efficiency in the removal of the amoxicillin antibiotic. The presence of ultrasonic waves also has a significant effect on the system's efficiency, so according to all of the results of this study, the combined US/H 2 O 2 /nZVI process can be considered a promising candidate for the removal of the antibiotic.