Oxidative Degradation of Tetracycline by Magnetite and Persulfate: Performance, Water Matrix Effect, and Reaction Mechanism

This study presents a strategy to remove tetracycline by using magnetite-activated persulfate. Magnetite (Fe3O4) was synthesized at high purity levels—as established via X-ray diffractometry, transmission electron microscopy, and N2 sorption analyses—and tetracycline was degraded within 60 min in the presence of both magnetite and persulfate (K2S2O8), while the use of either substance yielded limited degradation efficiency. The effects of magnetite and persulfate dosage, the initial concentration of tetracycline, and the initial pH on the oxidative degradation of tetracycline were interrogated. The results demonstrate that the efficiency of tetracycline removal increased in line with magnetite and persulfate dosage. However, the reaction rate increased only when increasing the magnetite dosage, not the persulfate dosage. This finding indicates that magnetite serves as a catalyst in converting persulfate species into sulfate radicals. Acidic conditions were favorable for tetracycline degradation. Moreover, the effects of using a water matrix were investigated by using wastewater treatment plant effluent. Comparably lower removal efficiencies were obtained in the effluent than in ultrapure water, most likely due to competitive reactions among the organic and inorganic species in the effluent. Increased concentrations of persulfate also enhanced removal efficiency in the effluent. The tetracycline degradation pathway through the magnetite/persulfate system was identified by using a liquid chromatograph-tandem mass spectrometer. Overall, this study demonstrates that heterogeneous Fenton reactions when using a mixture of magnetite and persulfate have a high potential to control micropollutants in wastewater.


Introduction
Pharmaceuticals such as antibiotics, which are highly water-soluble and hardly biodegradable, have been widely detected in aquatic systems in recent decades, resulting in serious threats to both public health and the natural environment, due not only to their toxicity but also to the prevalence of bacteria that have become antibiotic-resistant as a result of their heavy use [1,2]. Among a number of antibiotics, tetracycline (TC) is one of the most widely used in veterinary medicine and causes many of the environmental problems described above [3,4].
Conventional biological wastewater treatment processes, such as activated sludge, are relatively inefficient at removing TC [5,6]. According to Oulton et al. (2010), the reported removal efficiency of TC by activated sludge varies widely, ranging from 23% to 91%. Advanced oxidation processes (AOPs) have recently been proven efficient for removing non-biodegradable organic substances, including TC [7]. The primary mechanism behind AOPs involves generating highly reactive free radicals and then exploiting these species to degrade organic compounds oxidatively [8]. The Fenton process is a well-known AOP in which hydroxyl radicals are generated via reactions between iron salts and hydrogen peroxide (Fe 2+ /H 2 O 2 ) [9]. However, it has certain limitations, including the inefficient utilization of rapidly formed hydroxyl radicals and the ineffective removal of organic pollutants [10]. AOPs involving sulfate radicals (SO 4 •− ) have gained great attention as a means of overcoming these limitations, since they have redox potentials of 2.5 to 3.1 V, higher than those of hydroxyl radicals (1.8 to 2.7 V). In addition, sulfate radicals live longer than hydroxyl radicals and are more selective at removing pollutants. Their operating pH is also much wider than hydroxyl radicals [11,12]. Persulfate (PS, S 2 O 8 2− ) is stable at room temperature and can be activated by using heat, ultraviolet (UV), and transition metals (e.g., Fe 2+ and Co 2+ ) to form highly reactive sulfate radicals, as described in Equation (1) below [13]: Specifically, the reaction mechanism of PS activated by Fe 2+ , as explained in Equations (2) and (3) [14], is similar to the Fenton process, because the molecular structure of PS is an asymmetrically substituted derivative of hydrogen peroxide [15]: Although the Fe 2+ /PS process has been improved in terms of redox potentials and applicable operating pH when compared to the classic Fenton process, it still has several inherent drawbacks that have limited its broad application in wastewater treatment. First, excess Fe 2+ causes SO 4 •− scavenging, which in turn suppresses oxidation of the target contaminant via the mechanism described in Equation (3). Second, Fe 2+ activation only occurs effectively under acidic pH conditions (2)(3)(4), in order to prevent rapid iron precipitation such as the formation of ferric hydroxy complexes when the pH is above 4. Moreover, at the end of the Fe 2+ /PS process, a large amount of iron sludge is yielded, thereby leading to additional cost in treating it before discharge [16,17].
Heterogeneous Fenton systems that use iron-containing solids (Fe 2 O 3 , Fe 3 O 4 , FeO(OH), etc.) as a source of ferrous iron have been introduced to overcome the limitations of homogeneous Fenton processes [18]. With the use of a heterogeneous Fenton process, the rate of sludge production can be regulated, and the operating pH can be wider than in the homogeneous Fenton process. Magnetite (Fe 3 O 4 ) nanoparticles are considered effective catalysts for heterogeneous Fenton processes, because Fe 2+ of Fe 3 O 4 has octahedral sites that support the decomposition of H 2 O 2 or PS into reactive radicals [19,20]. In addition, Fe 3 O 4 nanoparticles can be easily manufactured, have high biocompatibilities and stabilities, and are easily separated by using an external magnetic field. These attributes make them highly suitable as catalysts in the removal of non-degradable organic substances [21,22]. Table 1 summarizes the experimental conditions used in this research compared with other studies reported in the literature. Heterogeneous Fenton processes that employ Fe 3 O 4 and PS have been reported to degrade several pharmaceutical compounds, such as norfloxacin and sulfamethoxazole [23,24]. The use of nanocomposite materials in which the magnetite was combined with other supporting materials, such as biochar, activated carbon, and chitosan for degradation of tetracycline, was also reported [25][26][27][28]. Several previous works have reported the inhibitory effect of organic and inorganic substances individually [21,25,26], but there is no report on the inhibitory effect of actual wastewater samples on TC degradation by the Fe 3 O 4 /PS system. Thus, the practicality of a heterogeneous Fenton process that uses Fe 3 O 4 and PS in real-world applications has yet to be tested. This study evaluates the capacity of Fe 3 O 4 as a catalyst for activating PS to the extent that it can degrade TC. Fe 3 O 4 nanoparticles were synthesized and subsequently characterized by using nitrogen sorption via Brunauer-Emmett-Teller (BET) analyses, X-ray diffraction (XRD), and transmission electron microscopy (TEM). The removal efficiency of TC under various reaction conditions was investigated. Reaction variables include Fe 3 O 4 and PS concentration as well as the initial pH. The effect of the water matrix was evaluated by comparing removal efficiencies from deionized water and wastewater treatment plant effluent. The TC degradation mechanism was further investigated by analyzing the reaction intermediate, using the liquid chromatograph-tandem mass spectrometer.

Reagent
First, ultrapure water was purged with N 2 gas for 1 h. 10 mL of 3.0 M FeCl 2 ·4H 2 O (in 2.0 M HCl) and 10 mL of 6.0 M FeCl 3 ·6H 2 O (in 2.0 M HCl) solutions were prepared as starting materials. An aqueous solution of ferric chloride and ferrous chloride at a 2:1 molar ratio of Fe 3+ : Fe 2+ was prepared, injected into a 1 L three-necked round flask containing 600 mL deionized water, and heated to 50 to 60 • C with stirring. After 10 min, 200 mL of 2.8 M NH 4 OH was added dropwise at a constant flow rate for 45 min, and the pH of the aqueous solution was raised to 10-11. The resulting solution was then heated and stirred at 50 to 60 • C for one hour and then stirred without heating for an additional hour. The reaction was continuously purged with N 2 gas during all of the synthesis steps described above. The resulting black solid was separated with a neodymium magnet, washed with ultrapure water until it assumed a neutral pH, and then washed three times with methanol. Finally, the product was dried overnight at 50 • C in a vacuum oven and stored in an anaerobic chamber prior to use in experiments.

Fe 3 O 4 Characterization
The crystal structure of Fe 3 O 4 was analyzed by X-ray diffractometer (XRD, D8 AD-VANCE, Bruker, Billerica, MA, USA). Morphology and size were observed using a transmission electron microscope (TEM, JEM-2010, JEOL, Tokyo, Japan). The specific surface areas of the Fe 3 O 4 particles were obtained by applying BET analyses to N 2 sorption data acquired at 77 K and p/p 0 = 0.99 (BELSORP-mini II, MicrotracBEL, Osaka, Japan).

TC Degradation Experiment Using Fe 3 O 4 /PS
Fifty milliliters of 41.6 µM of TC dissolved in ultrapure water was transferred to a glass vial. Fe 3 O 4 and PS were added into the glass vial simultaneously at the start of the experiment; therefore, this study did not consider the equilibrium time for TC adsorption on the catalyst. The solution was continuously mixed using a vertical rotating mixer (VM-80, Miulab, Zhejiang, Hangzhou, China) at a speed of 20 rpm. The batch experiments were conducted by adjusting the initial concentrations of Fe 3 O 4 and PS from 0.2 to 1 g/L and 0.05 to 1 mM, respectively, to investigate the impacts of Fe 3 O 4 and PS concentration on TC removal. To examine the effect of the initial pH on TC removal, it was adjusted by using 0.1 M NaOH or 0.1 M HClO 4 over a range of 3 to 9. As ClO 4 − has little impact on the oxidation process, it is a more appropriate ionic species for adjusting the pH than HCl and H 2 SO 4 [23] . The reaction time for each batch experiment was 60 min, during which a 1-mL sample was collected from each vial at specified time intervals, immediately filtered with a 0.45 µm PES syringe filter, and then quenched by adding 0.1 mL of ethanol to prevent further side reactions prior to analysis [30]. TC concentration over the experimental time was plotted by a pseudo first-order kinetic equation: C t = C 0 ·exp(−k obs ·t) (5) where k is the observed rate constant (min −1 ), C 0 is the initial concentration of TC, and C t is the concentration of TC at time t.
To understand the formation of intermediate product of TC during its reaction with Fe 3 O 4 /PS, a batch experiment was also conducted. TC, PS, and Fe 3 O 4 were added into a glass beaker with 200 mL of ultrapure water, resulting in initial concentrations of 41.6 µM, 1 mM, and 1 g/L respectively. pH was adjusted to 4.7. Samples were taken at 0, 10, 20, 30, 40, 50, and 60 min, filtered with a 0.45 µm PES syringe filter, and 0.1 mL of ethanol was added into a 0.9 mL sample to quench the reaction.
To examine the effect of the water matrix on TC removal by Fe 3 O 4 /PS, batch experiments using ultrapure water or municipal effluent wastewater from a regional wastewater treatment plant (WWTP) were performed and their results compared. The common parameters of the effluent are presented in the supporting information (Table S1).

Analytical Methods
For the quantitation analysis in the degradation experiments, TC concentration was analyzed by high-performance liquid chromatography (HPLC, U-3000, Thermo Fisher, Waltham, MA, USA). TC was separated in chromatography by using a C18 column (Accucore AQ 4.6 × 150 mm, 2.6 µm, Thermo Fisher, Waltham, MA, USA) and was detected at a 280 nm absorbance wavelength. The oven temperature was set at 50 • C. The mixture of 0.1% phosphoric acid/acetonitrile (85:15, v/v) was used as isocratic elution at a flow rate of 1 mL/min. The sample injection volume was 10 µL, while the transformation products of TC were quantified by HPLC (Vanquish Flex, Thermo Fisher Scientific, Waltham, MA, USA) coupled with a triple quadrupole mass spectrometer (TSQ Quantis, Thermo Fisher Scientific, Waltham, MA, USA). Chromatographic separation was performed using a C18 Nanomaterials 2021, 11, 2292 5 of 15 column (4.6 × 50 mm, 3.5 µm, Agilent, Santa Clara, CA, USA). A detailed description of the method can be found in Text S1.
A spectrofluorometer (F-7000 Fluorescence Spectrophotometer, Hitachi, Tokyo, Japan) was used to investigate changes in fluorescence intensity for the samples during the heterogeneous Fenton reaction. 19-0629) [26]. In addition, the positions of the diffraction peaks indicate an inverted spinel structure [31]. The average Fe 3 O 4 particle size can be obtained from the XRD pattern via the Debye-Scherrer formula D = Kλ/βcos(θ), where K is the Scherrer constant (0.89), λ is the X-ray wavelength (1.54 Å), β is full width at the half maximum of a given peak, and θ is the Bragg diffraction angle [17]. Application of the Debye-Scherrer equation to the most intense peak, i.e., the (311) reflection ( Figure 1), establishes an average particle size of 9.59 nm.
Nanomaterials 2021, 11, x FOR PEER REVIEW 5 of 16 of TC were quantified by HPLC (Vanquish Flex, Thermo Fisher Scientific, Waltham, MA,USA) coupled with a triple quadrupole mass spectrometer (TSQ Quantis, Thermo Fisher Scientific, Waltham, MA, USA). Chromatographic separation was performed using a C18 column (4.6 × 50 mm, 3.5 μm, Agilent, Santa Clara, CA, USA). A detailed description of the method can be found in Text S1. A spectrofluorometer (F-7000 Fluorescence Spectrophotometer, Hitachi, Tokyo, Japan) was used to investigate changes in fluorescence intensity for the samples during the heterogeneous Fenton reaction.

Physical and Morphological Characteristics of Fe3O4
XRD analyses were conducted to understand the crystallinity of the synthesized Fe3O4 nanoparticles. The diffraction results, as shown in Figure  19-0629) [26]. In addition, the positions of the diffraction peaks indicate an inverted spinel structure [31]. The average Fe3O4 particle size can be obtained from the XRD pattern via the Debye-Scherrer formula D = Kλ/βcos(θ), where K is the Scherrer constant (0.89), λ is the X-ray wavelength (1.54 Å), β is full width at the half maximum of a given peak, and θ is the Bragg diffraction angle [17]. Application of the Debye-Scherrer equation to the most intense peak, i.e., the (311) reflection ( Figure 1), establishes an average particle size of 9.59 nm. The morphology of the Fe3O4 particles was further identified by TEM, images for which are presented in Figure 2. Most of the Fe3O4 nanoparticles were quasi-spherical in shape with a diameter of 10 nm or less and formed aggregates that ranged in size from several hundreds of nm to several μm. The individual particle size obtained in the TEM analysis matched well with the sizes indicated by the XRD analyses. These results are similar to Kwon et al., who also reported on the properties of Fe3O4 [29]. The specific surface area of Fe3O4 obtained by BET analysis was 93.82 m 2 /g. The morphology of the Fe 3 O 4 particles was further identified by TEM, images for which are presented in Figure 2. Most of the Fe 3 O 4 nanoparticles were quasi-spherical in shape with a diameter of 10 nm or less and formed aggregates that ranged in size from several hundreds of nm to several µm. The individual particle size obtained in the TEM analysis matched well with the sizes indicated by the XRD analyses. These results are similar to Kwon et al., who also reported on the properties of Fe 3 O 4 [29]. The specific surface area of Fe 3 O 4 obtained by BET analysis was 93.82 m 2 /g.

The Catalytic Role of Fe3O4 in TC Degradation
The catalytic activity of Fe3O4 when activating PS and subsequently degrading TC was evaluated by monitoring TC removal in the presence or absence of PS and Fe3O4. As shown in Figure 3, PS alone resulted in only 4.4% TC removal efficiency after a one-hour

The Catalytic Role of Fe 3 O 4 in TC Degradation
The catalytic activity of Fe 3 O 4 when activating PS and subsequently degrading TC was evaluated by monitoring TC removal in the presence or absence of PS and Fe 3 O 4 . As shown in Figure 3, PS alone resulted in only 4.4% TC removal efficiency after a one-hour reaction, while a comparably higher removal efficiency of 49.7% was observed when the only Fe 3 O 4 was used. The dominant removal of TC by Fe 3 O 4 was likely due to adsorption by the Fe 3 O 4 surface, which is in line with the results from some previous similar studies [16,26]. This was further confirmed by the HPLC chromatography results for samples in the batch experiment with Fe 3 O 4 alone, which did not show any intermediate peaks ( Figure S1).

The Catalytic Role of Fe3O4 in TC Degradation
The catalytic activity of Fe3O4 when activating PS and subsequently degrading TC was evaluated by monitoring TC removal in the presence or absence of PS and Fe3O4. As shown in Figure 3, PS alone resulted in only 4.4% TC removal efficiency after a one-hour reaction, while a comparably higher removal efficiency of 49.7% was observed when the only Fe3O4 was used. The dominant removal of TC by Fe3O4 was likely due to adsorption by the Fe3O4 surface, which is in line with the results from some previous similar studies [16,26]. This was further confirmed by the HPLC chromatography results for samples in the batch experiment with Fe3O4 alone, which did not show any intermediate peaks (Figure S1). In comparison to the above results, reactions conducted with both Fe3O4 and PS exhibited much higher TC removal efficiencies of 90.0% in 30 min and 98.1% in 1 h. An HPLC chromatogram of the reaction solution revealed several intermediate peaks, thus establishing that oxidation is involved in TC removal ( Figure S2). This result indicated that the enhanced removal efficiency in Fe3O4/PS was due to the oxidative degradation of TC. The concentration of total organic carbon (TOC) in the reaction solution also decreased by more than 50%, indicating that TC species were oxidatively decomposed by catalytic reactions. The degradation kinetic was further interpreted by the pseudo first-order kinetic. The high linearity (R 2 = 0.9875) of the fitted line indicates that the reaction had pseudo first-order kinetics, which has been widely observed for similar catalytic systems [26]. The kobs for the Fe3O4/PS system was calculated as 0.1038 min −1 . Soluble iron concentration was maintained at less than 0.3 mg/L throughout the experiment for all samples, thereby verifying the high stability of Fe3O4 during the catalytic reaction. In comparison to the above results, reactions conducted with both Fe 3 O 4 and PS exhibited much higher TC removal efficiencies of 90.0% in 30 min and 98.1% in 1 h. An HPLC chromatogram of the reaction solution revealed several intermediate peaks, thus establishing that oxidation is involved in TC removal ( Figure S2). This result indicated that the enhanced removal efficiency in Fe 3 O 4 /PS was due to the oxidative degradation of TC. The concentration of total organic carbon (TOC) in the reaction solution also decreased by more than 50%, indicating that TC species were oxidatively decomposed by catalytic reactions. The degradation kinetic was further interpreted by the pseudo first-order kinetic. The high linearity (R 2 = 0.9875) of the fitted line indicates that the reaction had pseudo first-order kinetics, which has been widely observed for similar catalytic systems [26]. The k obs for the Fe 3 O 4 /PS system was calculated as 0.1038 min −1 . Soluble iron concentration was maintained at less than 0.3 mg/L throughout the experiment for all samples, thereby verifying the high stability of Fe 3 O 4 during the catalytic reaction.

Effect of Fe 3 O 4 Concentration
As discussed in Section 3.2, Fe 3 O 4 was proven as an efficient catalyst for activating PS in order to remove TC. The effect of catalyst dosage on TC removal by Fe 3 O 4 /PS with various initial Fe 3 O 4 concentrations (0.2-2.0 g/L) was studied. Initial concentrations of TC and PS were 41.6 µM and 1 mM, respectively. As presented in Figure 4A, the removal of TC increased when increasing the initial Fe 3 O 4 concentration. For example, when 0.2 g/L, 0.5 g/L, and 1.0 g/L of Fe 3 O 4 were used, the removal efficiencies of TC at a 60-min reaction time were 29.8%, 79.2%, 95.9%, respectively, and full removal of TC within our detection limits was achieved at Fe 3 O 4 concentrations of 1.5 g/L or 2.0 g/L. In Figure 4B, the pseudo first-order rate constants gradually increased from 0.0039 to 0.1309 min −1 along with an increase in initial Fe 3 O 4 concentration. The enhanced kinetic constants in increased Fe 3 O 4 addition were attributed to the adsorption and oxidation. The higher catalyst concentrations accelerate TC decomposition, which can be explained by the greater concentration of active sites at higher Fe 3 O 4 concentrations. Moreover, the large numbers of active sites also enhance the adsorption of TC and PS molecules to Fe 3 O 4 , leading to higher reaction kinetics [32].
ious initial Fe3O4 concentrations (0.2-2.0 g/L) was studied. Initial concentrations of TC and PS were 41.6 μM and 1 mM, respectively. As presented in Figure 4A, the removal of TC increased when increasing the initial Fe3O4 concentration. For example, when 0.2 g/L, 0.5 g/L, and 1.0 g/L of Fe3O4 were used, the removal efficiencies of TC at a 60-min reaction time were 29.8%, 79.2%, 95.9%, respectively, and full removal of TC within our detection limits was achieved at Fe3O4 concentrations of 1.5 g/L or 2.0 g/L. In Figure 4B, the pseudo first-order rate constants gradually increased from 0.0039 to 0.1309 min −1 along with an increase in initial Fe3O4 concentration. The enhanced kinetic constants in increased Fe3O4 addition were attributed to the adsorption and oxidation. The higher catalyst concentrations accelerate TC decomposition, which can be explained by the greater concentration of active sites at higher Fe3O4 concentrations. Moreover, the large numbers of active sites also enhance the adsorption of TC and PS molecules to Fe3O4, leading to higher reaction kinetics [32].

Effect of PS Concentration
Similar to the batch study of the effect of Fe3O4 concentration on TC removal in the above section, the impact of the initial concentration of PS (0 to 1 mM) on TC removal in Fe3O4/PS was also investigated. As presented in Figure 5A, a system without the addition of PS had a TC removal efficiency of 43.5%, but this increased to 63.9% even with a relatively low 0.05 mM PS concentration. Further increasing the PS concentration to 1 mM PS dramatically raised TC decomposition efficiency to 95.9%. This distinct improvement in efficiency at higher PS concentrations possibly originates from the greater reactive radical generation apparent at higher PS concentrations [28,33].

Effect of PS Concentration
Similar to the batch study of the effect of Fe 3 O 4 concentration on TC removal in the above section, the impact of the initial concentration of PS (0 to 1 mM) on TC removal in Fe 3 O 4 /PS was also investigated. As presented in Figure 5A, a system without the addition of PS had a TC removal efficiency of 43.5%, but this increased to 63.9% even with a relatively low 0.05 mM PS concentration. Further increasing the PS concentration to 1 mM PS dramatically raised TC decomposition efficiency to 95.9%. This distinct improvement in efficiency at higher PS concentrations possibly originates from the greater reactive radical generation apparent at higher PS concentrations [28,33]. Based on the kinetic fitting results in Figure 5A, the corresponding rate constants and removal efficiencies of TC under different PS initial concentrations were obtained ( Figure  5B). PS concentration did not significantly affect the rate constant, in comparison to the impactor degree of Fe3O4 detailed in Section 3.3.1. Interestingly, TC removal efficiency after 60 min of reaction increased from 43.5 to 85.2% when PS concentration increased from 0 to 0.2 mM; thereafter, removal efficiency remained stable at above 0.2 mM.
These results clearly indicate that the catalyst concentration determines the reaction rate in the heterogeneous Fenton system. By comparison, PS does not significantly influence the reaction rate, because PS is an oxidizing agent that is decomposed by the catalyst. In addition, excess PS hinders the overall reaction, due to the generation of sulfate anions via scavenging reactions among radicals, as shown in Equations (6) and (7) [12,34]. Based on the kinetic fitting results in Figure 5A, the corresponding rate constants and removal efficiencies of TC under different PS initial concentrations were obtained ( Figure 5B). PS concentration did not significantly affect the rate constant, in comparison to the impactor degree of Fe 3 O 4 detailed in Section 3.3.1. Interestingly, TC removal efficiency after 60 min of reaction increased from 43.5 to 85.2% when PS concentration increased from 0 to 0.2 mM; thereafter, removal efficiency remained stable at above 0.2 mM.
These results clearly indicate that the catalyst concentration determines the reaction rate in the heterogeneous Fenton system. By comparison, PS does not significantly influence the reaction rate, because PS is an oxidizing agent that is decomposed by the catalyst. In addition, excess PS hinders the overall reaction, due to the generation of sulfate anions via scavenging reactions among radicals, as shown in Equations (6) and (7) [12,34].

Effect of Initial pH
The heterogeneous Fenton reaction is influenced by the solution's pH, because the dissolution of ferrous iron (Fe 2+ ) depends on pH. Batch experiments under different pH conditions, ranging from 3.0 to 9.0, were conducted to understand their impact on the removal efficiency of TC by Fe 3 O 4 /PS. As presented in Figure 6, TC removal increased under low pH conditions, and the highest rate constant (0.1530 min −1 ) was obtained at pH 3.0. The rate constant decreased under alkaline conditions, and the lowest rate coefficient value of 0.066 min −1 was observed at pH 9.0. Previous reports demonstrated that when pH > 4.0, the concentration of soluble Fe 2+ ions decreased, due to the formation of Fe 2+ complexes and precipitation; subsequently, the activation rates for PS and SO 4 •− production decreased [26]. In alkaline pH conditions, SO 4 •− radicals converted into •OH radicals, thus serving as scavengers for SO 4 •− via the reactions shown in Equations (8) and (9)   However, the change in the reaction rate was not as dramatic as seen in similar studies [34,37], because, in our study, pH was only adjusted accordingly before starting the batch experiments, unlike the previous literature that used a pH buffer throughout the experiments. As a result, the initial pH values of 3.0, 4.8, 7.2, and 9.1 applied to the batch experiment were changed to final pH values of 3.0, 4.1, 4.2, and 4.5, respectively. No drop in pH was observed for samples with an initially low pH, while samples with higher initial pH values experienced more significant drops in this regard. These decreases in pH can be explained by the production of carboxylic acid products and the decomposition of sulfate salts [38,39].

Effect of the Water Matrix
As mentioned in the introduction, the water matrix effect of actual wastewater on TC removal by Fe3O4/PS system have not been reported; therefore, we examined the effect of the water matrix on the removal of TC by Fe3O4/PS by using municipal effluent wastewater from a local WWTP, while ultrapure water was used in the comparison batch study ( Figure S3).
The rate constants were similar in both batch studies, namely 0.100 min −1 in the municipal effluent and 0.107 min −1 in the ultrapure water, but removal efficiency at 60 min was significantly different at 88.3% and 96.0%, respectively. This clearly demonstrates a key role of the water matrix on the heterogeneous Fenton reaction for TC degradation. However, the change in the reaction rate was not as dramatic as seen in similar studies [34,37], because, in our study, pH was only adjusted accordingly before starting the batch experiments, unlike the previous literature that used a pH buffer throughout the experiments. As a result, the initial pH values of 3.0, 4.8, 7.2, and 9.1 applied to the batch experiment were changed to final pH values of 3.0, 4.1, 4.2, and 4.5, respectively. No drop in pH was observed for samples with an initially low pH, while samples with higher initial pH values experienced more significant drops in this regard. These decreases in pH can be explained by the production of carboxylic acid products and the decomposition of sulfate salts [38,39].

Effect of the Water Matrix
As mentioned in the introduction, the water matrix effect of actual wastewater on TC removal by Fe 3 O 4 /PS system have not been reported; therefore, we examined the effect of the water matrix on the removal of TC by Fe 3 O 4 /PS by using municipal effluent wastewater from a local WWTP, while ultrapure water was used in the comparison batch study ( Figure S3). The rate constants were similar in both batch studies, namely 0.100 min −1 in the municipal effluent and 0.107 min −1 in the ultrapure water, but removal efficiency at 60 min was significantly different at 88.3% and 96.0%, respectively. This clearly demonstrates a key role of the water matrix on the heterogeneous Fenton reaction for TC degradation. The consumption of PS in the municipal effluent was comparatively greater, due to the presence of other organic and inorganic matter, which by comparison did not affect the reaction rate. Additional experiments were conducted to interrogate how Fe 3 O 4 and PS concentration influence TC removal in municipal effluent. The results of these experiments are summarized in Table 2. Overall, TC removal efficiency was lower in the municipal effluent than that in the ultrapure water, thereby suggesting that the generated oxidative radicals were consumed by organic/inorganic matter in the effluent and TC, which were competing with each other. Interestingly, when only the PS was introduced without Fe 3 O 4 , relatively higher TC removal efficiency was observed in the batch experiments treating the municipal effluent (42.5%) than those treating ultrapure water (4.4%). This may have been the result of the presence of metal cations (Mn, Cu, Ni) in the municipal effluent (Table S1) acting as persulfate activators (Equation (10)) [40][41][42]. Li et al. (2016) reported that the very low concentration of Mn 2+ , such as 0.5 mg/L, was enough to produce the hydroxyl radicals in the presence of hydrogen peroxide [43]. Moreover, organic matters, such as quinones and phenols, can also activate PS for subsequent degradation of organic matters, as reported by Fang et al. (2013) and Ahmad et al. (2013) [44,45].
When the initial PS concentration was fixed at 1 mM, increasing Fe 3 O 4 concentrations also led to improvements in TC removal efficiencies, namely 42.5%, 54.2%, and 78.3% at 0 g/L, 0.5 g/L, and 1.0 g/L of Fe 3 O 4 , respectively. A similar trend was observed in ultrapure water. The effect of PS concentration was also probed in municipal effluent. In ultrapure water, increasing the dosage of PS from 0 to 0.2 mM yielded a logarithmic increase in TC degradation efficiency, which plateaued at PS concentrations from 0.2 to 1.0 mM. By comparison, in municipal effluent, TC removal efficiency gradually increased with PS content at low PS concentrations (0 to 0.5 mM), albeit it sharply increased at 1.0 mM. The low enhancement of TC removal efficiency with greater PS concentration is likely the result of TC competing with other organic and inorganic matter in the municipal effluent, which consumes radical species. Sulfate radicals can be scavenged at high-pH, organic carbons, and various anionic species, including Cl − , HCO 3 − , and CO 3 2− that are present in effluent via the reactions in Equation (11) to (13) [40,46]. TOC removal efficiencies were 50 to 17% for ultrapure water and municipal effluent, respectively, indicating that TC degradation was hindered by organic and inorganic matter in the municipal effluent. The consumption of oxidative radicals by such compounds in the municipal effluent resulted in lower TC removal efficiencies: Changes in organic matter with Fe 3 O 4 /PS were further analyzed by using excitationemission matrix (EEM) fluorescence analyses. EEM data collected from TC-spiked municipal effluent ( Figure 7A) show peaks in the 360-380 nm/520-540 nm (excitation/emission) and 230-240 nm/520-540 nm (excitation/emission) regions, corresponding to humic acidlike and fulvic acid-like groups, respectively. After 1 h of the TC degradation reaction ( Figure 7B), overall peak intensity decreased by approximately half, indicating the degradation of organic matter during the Fe 3 O 4 /PS experiments. Moreover, peak intensities in the 230-240 nm/340-400 nm region comparatively increased after the reaction, thus indicating the presence of low-molecular-weight organic substances such as tryptophan-like and other protein-like compounds [47,48]. The EEM matrices were further analyzed by the humification index (HIX), which is the ratio of the integral of the emission spectrum (excited at 255 nm) over the spectral range of 434 to 480 nm to the integral of the emission spectrum over the spectral range of 330 to 346 nm (also excited at 255 nm) [49]. High-molecular-weight organic compounds are characterized by high HIX values. Table 3 summarizes the HIX values obtained for several water samples, municipal effluent, municipal effluent spiked with TC before the Fe3O4/PS experiment, and Fe3O4/PS-treated water. The HIX value was 6.51 for the municipal effluent, and this dramatically increased to 21.53 upon adding TC (41.6 μM). The HIX value for the Fe3O4/PS-treated water was relatively lower at 7.55, which is similar to the HIX value for municipal effluent. Similar trends were observed in the SUVA254 value, which is the ratio of decadal absorbance at 254 nm to the dissolved organic carbon concentration (DOC). Typically, the SUVA254 ranges from 1.2 to 2.6 L/mg·m for secondary effluents [50]. The SUVA254 value for the municipal effluent in this study was 1.51 L/mg·m, which falls within the reported The EEM matrices were further analyzed by the humification index (HIX), which is the ratio of the integral of the emission spectrum (excited at 255 nm) over the spectral range of 434 to 480 nm to the integral of the emission spectrum over the spectral range of 330 to 346 nm (also excited at 255 nm) [49]. High-molecular-weight organic compounds are characterized by high HIX values. Table 3 summarizes the HIX values obtained for several water samples, municipal effluent, municipal effluent spiked with TC before the Fe 3 O 4 /PS experiment, and Fe 3 O 4 /PS-treated water. The HIX value was 6.51 for the municipal effluent, and this dramatically increased to 21.53 upon adding TC (41.6 µM). The HIX value for the Fe 3 O 4 /PS-treated water was relatively lower at 7.55, which is similar to the HIX value for municipal effluent. Similar trends were observed in the SUVA 254 value, which is the ratio of decadal absorbance at 254 nm to the dissolved organic carbon concentration (DOC). Typically, the SUVA 254 ranges from 1.2 to 2.6 L/mg·m for secondary effluents [50]. The SUVA 254 value for the municipal effluent in this study was 1.51 L/mg·m, which falls within the reported range. By comparison, municipal effluent spiked with TC exhibited a relatively greater SUVA 254 of 4.13, which subsequently reduced to 2.20 after Fe 3 O 4 /PS treatment. As the HIX and SUVA 254 are measures of the aromaticity of organic matter, we conclude that the aromatic rings in the TC molecule degraded during the reaction promoted by the Fe 3 O 4 /PS. The reduced aromaticity resulting from the Fe 3 O 4 /PS agrees well with related studies reporting on the mechanism of TC degradation in the Fenton reaction [21,51] .

Transformation Products and Proposed Pathways
Intensity development during the reaction between TC and Fe 3 O 4 /PS for the TC peak eluted at 5.09 min, and six major peaks inferred to the transformation products (TPs) can be seen in total ion chromatography in Figures 8 and S4. The TPs were found at elution times ranging from 10 to 14 min, namely, TP 1(m/z = 344, 10.69 min), TP 2 (m/z = 358, 11.06 min), TP 3 (m/z = 274, 11.77 min), TP 4 (m/z = 290, 12.02 min), TP 5 (m/z = 256, 12.92 min), and TP 6 (m/z = 284, 13.82 min). TC peak intensity significantly decreased over time, while the TP 6 peak intensity increased over time. Unlike TP 6, the intensities of TP 1-5 peaks initially increased but then decreased. These TPs with their predicated molecular structures, and the corresponding mass spectrum intensities, are illustrated in Table S2 and Figure S5. the TP 6 peak intensity increased over time. Unlike TP 6, the intensities of TP 1-5 peaks initially increased but then decreased. These TPs with their predicated molecular structures, and the corresponding mass spectrum intensities, are illustrated in Table S2 and Figure S5. A possible TC degradation pathway is proposed in Figure 9. First, the formation of TP with an m/z = 417 was the result of losing dimethyl amino at C4 of TC (m/z = 445), due to low N-C bond energy [28]. TP with an m/z = 401 formed through the dehydration pathway from C6 of the TP with an m/z = 417 [52]. Through the deamidation reaction, namely, the loss of the acylamino group at C1 of TP with an m/z = 401 [53], the formation of TP 2 (m/z = 358) was observed. Thereafter, TP 1 (m/z = 344) was found, due to the cleavage of the carboatomic ring of TP 2 [54], followed by decarboxylation, which converted TP 1 to the TP with an m/z = 300 [55]. After dimethyl and deformaldehyde process by H addition, TP with an m/z = 279 was obtained, while the formation of TP 6 (m/z = 284) was attributed to TP with an m/z = 279 of decarboxylation and H addition reaction to C-H double bond on benzene ring as well as carboxyl addition. TP 6 was also identified as one of the transformation products for previous references using the photocatalyst to remove TC [56,57]. Unlike the pathway from an m/z = 300 to an m/z = 284, TP 4 (m/z = 290) was formed by adding dimethyl and H to the C-O double bond of the TP with an m/z = 300. Based on TP 4, TP 3 was formed due to its decarboxylation, while TP 5 was obtained through its dimethyl decarboxylation and H addition reaction to a C-H double bond on a benzene ring. When compared to TP 3, TP 3 (a) had the same m/z as TP 3 but a different molecular structure and was directly derived from TC. TP 3 (a) was also found in a previous study using a catalyst regenerated from Fenton sludge containing heavy metal, in order to active persulfate for TC degradation [53]. A possible TC degradation pathway is proposed in Figure 9. First, the formation of TP with an m/z = 417 was the result of losing dimethyl amino at C4 of TC (m/z = 445), due to low N-C bond energy [28]. TP with an m/z = 401 formed through the dehydration pathway from C6 of the TP with an m/z = 417 [52]. Through the deamidation reaction, namely, the loss of the acylamino group at C1 of TP with an m/z = 401 [53], the formation of TP 2 (m/z = 358) was observed. Thereafter, TP 1 (m/z = 344) was found, due to the cleavage of the carboatomic ring of TP 2 [54], followed by decarboxylation, which converted TP 1 to the TP with an m/z = 300 [55]. After dimethyl and deformaldehyde process by H addition, TP with an m/z = 279 was obtained, while the formation of TP 6 (m/z = 284) was attributed to TP with an m/z = 279 of decarboxylation and H addition reaction to C-H double bond on benzene ring as well as carboxyl addition. TP 6 was also identified as one of the transformation products for previous references using the photocatalyst to remove TC [56,57]. Unlike the pathway from an m/z = 300 to an m/z = 284, TP 4 (m/z = 290) was formed by adding dimethyl and H to the C-O double bond of the TP with an m/z = 300. Based on TP 4, TP 3 was formed due to its decarboxylation, while TP 5 was obtained through its dimethyl decarboxylation and H addition reaction to a C-H double bond on a benzene ring. When compared to TP 3, TP 3 (a) had the same m/z as TP 3 but a different molecular structure and was directly derived from TC. TP 3 (a) was also found in a previous study using a catalyst regenerated from Fenton sludge containing heavy metal, in order to active persulfate for TC degradation [53]. Nanomaterials 2021, 11, x FOR PEER REVIEW 13 of 16

Conclusions
This study presents a new understanding of the degradation of TC via a heterogeneous Fenton system using Fe3O4 and persulfate. Fe3O4 was successfully synthesized by using the coprecipitation method under alkaline conditions, and the resulting materials were characterized via XRD, TEM, and BET analyses. The catalytic roles of Fe3O4 for persulfate activation and TC degradation were evaluated by monitoring TC reduction in the presence or absence of persulfate and Fe3O4. TC was removed from solutions by adsorption onto Fe3O4 surfaces, where the TC was then oxidatively degraded by the Fenton reaction. The effects of various reaction conditions, including Fe3O4 concentration, PS concentration, and the initial pH, were further interrogated to understand the reaction mechanism. The kinetics of TC decomposition gradually increased in line with an increase in Fe3O4 concentration, thus indicating the catalytic role of Fe3O4. Increasing the concentration of PS influenced TC removal efficiency but not TC removal kinetics, because persulfate is an oxidizing agent that is decomposed by the catalyst. Excess PS hindered the overall reaction by scavenging sulfate radicals. Lower initial pH conditions enhanced the overall reaction kinetic by increasing Fe 2+ generation. The effect of the water matrix was further investigated by using WWTP effluent. Overall, TC removal efficiency was lower in the effluent versus ultrapure water, due to TC competing with organic and inorganic matter that consumes generated oxidative radicals. DOM analyses, e.g., EEM and SUVA254, revealed that reactions promoted by the HFS break the aromatic ring moieties of TC molecules. The oxidative degradation of TC was nonetheless maintained in the effluent, and TC removal efficiency in WWTP effluent could be increased by raising the concentration of PS. The tetracycline degradation pathway through the magnetite/persulfate system was identified by using a liquid chromatograph-tandem mass spectrometer. Thus, the heterogeneous Fenton process with Fe3O4 and PS appears to have significant potential for use in removing and controlling micropollutants in wastewater.

Supplementary Materials:
The following are available online at www.mdpi.com/xxx/s1, Text S1: HPLC-MS/MS operational parameters for identifying the transformation product of tetracycline removed by magnetite/PS, Table S1: common parameters of municipal effluent wastewater, Table S2: transformation products (TPs) of tetracycline removed by Fe3O4/PS, Figure S1: HPLC peak when TC was removed by only Fe3O4, Figure S2: HPLC peak when TC was removed by Fe3O4/PS, Figure S3 Figure S4: the total ion chromatogram (HPLC-MS/MS) of tetracycline removed by Fe3O4/PS and its transformation products in samples taken over the reaction

Conclusions
This study presents a new understanding of the degradation of TC via a heterogeneous Fenton system using Fe 3 O 4 and persulfate. Fe 3 O 4 was successfully synthesized by using the coprecipitation method under alkaline conditions, and the resulting materials were characterized via XRD, TEM, and BET analyses. The catalytic roles of Fe 3 O 4 for persulfate activation and TC degradation were evaluated by monitoring TC reduction in the presence or absence of persulfate and Fe 3 O 4 . TC was removed from solutions by adsorption onto Fe 3 O 4 surfaces, where the TC was then oxidatively degraded by the Fenton reaction. The effects of various reaction conditions, including Fe 3 O 4 concentration, PS concentration, and the initial pH, were further interrogated to understand the reaction mechanism. The kinetics of TC decomposition gradually increased in line with an increase in Fe 3 O 4 concentration, thus indicating the catalytic role of Fe 3 O 4 . Increasing the concentration of PS influenced TC removal efficiency but not TC removal kinetics, because persulfate is an oxidizing agent that is decomposed by the catalyst. Excess PS hindered the overall reaction by scavenging sulfate radicals. Lower initial pH conditions enhanced the overall reaction kinetic by increasing Fe 2+ generation. The effect of the water matrix was further investigated by using WWTP effluent. Overall, TC removal efficiency was lower in the effluent versus ultrapure water, due to TC competing with organic and inorganic matter that consumes generated oxidative radicals. DOM analyses, e.g., EEM and SUVA 254 , revealed that reactions promoted by the HFS break the aromatic ring moieties of TC molecules. The oxidative degradation of TC was nonetheless maintained in the effluent, and TC removal efficiency in WWTP effluent could be increased by raising the concentration of PS. The tetracycline degradation pathway through the magnetite/persulfate system was identified by using a liquid chromatograph-tandem mass spectrometer. Thus, the heterogeneous Fenton process with Fe 3 O 4 and PS appears to have significant potential for use in removing and controlling micropollutants in wastewater.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/10 .3390/nano11092292/s1, Text S1: HPLC-MS/MS operational parameters for identifying the transformation product of tetracycline removed by magnetite/PS, Table S1: common parameters of municipal effluent wastewater, Table S2: transformation products (TPs) of tetracycline removed by Fe 3 O 4 /PS, Figure S1: HPLC peak when TC was removed by only Fe 3 O 4 , Figure S2: HPLC peak when TC was removed by Fe 3 O 4 /PS, Figure S3 Funding: This study was supported by the Advanced Research Project funded by the SeoulTech (Seoul National University of Science and Technology).

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.