Effect of UV Irradiation on the Structural Variation of Metal Oxide-Silica Nanocomposites for Enhanced Removal of Erythromycin at Neutral pH

: In this study, the effect of UV treatment on the physicochemical properties and structural variation of metal oxide-silica nanocomposites (Mn 2 O 3 -Fe 2 O 3 @SiO 2 ) has been investigated. Based on the results, UV irradiation signiﬁcantly affects the nanocomposite structure, where SiO 2 network reconﬁguration, change in surface OH group density, and surface area were observed. Erythromycin (ERY) has been chosen as a module pollutant to compare the performance of the pristine and UV-treated nanocomposites. The pristine nanocomposite had a high adsorption efﬁciency (99.47%) and photocatalytic activity (99.57%) at neutral pH for ERY in the ﬁrst cycle, and this efﬁciency decreased signiﬁcantly for the multiple cycles. However, different results have been observed for the UV-treated nanocomposite, where it retained its performance for ten consecutive cycles. This enhanced performance is attributed to the structural modiﬁcations after UV exposure, where increased surface area, pore volume, and OH group density resulted in an increased number of the possible mechanisms responsible for the adsorption/oxidation of ERY. Moreover, oxidation of adsorbed molecules by UV light after each cycle can also be another reason for enhanced removal. For the ﬁrst time, the fate of ERY is studied using regenerated nanocomposites after the last cycle. LC/MS/MS results showed that ERY degraded in 20 min, and the produced reaction by-products were adsorbed by nanocomposites. This study could be a foundation research for the practical approaches for the regeneration of nanomaterials and the successful removal of organic pollutants from aquatic environments.


Introduction
In recent decades, dyes, phenols, heavy metals, antibiotics, pesticides, and polycyclic aromatic hydrocarbons are among the most important pollutants that require special management [1]. Antibiotics have been extensively used a considerable amount in medicine and veterinary fields for various purposes [2]. Usually, antibiotics cannot completely metabolize in the body and are excreted via urine and feces and transported to the wastewater treatment plants. Since the conventional treatment processes are not adequate to remove the emerging micropollutants, they eventually enter the aquatic environment and find their way to the groundwater, drinking water, and soil [3]. Residual antibiotics in the aquatic environment can cause toxicity to the organisms and induce the occurrence of antibiotic resistance genes (ARGs) and antibiotic resistance bacteria (ARB), which have become a severe problem worldwide [4]. According to Decision 495/2015 by the European Union, scanning electron microscopy (FE-SEM) and high-resolution transmission electron microscope (HR-TEM). FE-SEM was analyzed using Hitachi S-4300, Japan, where the dried nanocomposites were coated with a gold-platinum alloy by ion-sputtering (E-1048 Hitachi ion sputter). The FE-SEM image of the nanocomposite showed a highly porous surface with quasi spheroid morphology (Figure 1a). The HR-TEM images were taken with a field emission TEM (FE-TEM, JEM-2010F, JEOL, Tokyo, Japan). The two-dimensional (2D) elemental mapping was performed by energy-dispersive X-ray spectroscopy (EDX) (X-Maxn 80 T, Oxford, UK). The presence of the metal oxide particles in the silica matrices was confirmed in the HR-TEM image shown in Figure 1b. The Figure 1b inset shows that the crystallite planes were assigned by estimating the fringe width values on the HR-TEM image and then relating it with the inter-planer spacing values from the XRD. The lattice spacing of 0.26 nm matched with Bragg reflection from (104) and (222) planes of Fe 2 O 3 and Mn 2 O 3 , respectively. The results showed that the MOS nanocomposite consists of uniformly distributed Si, Mn, Fe, and O (Figure 1c-g). The 2D mapping of Fe overlapped with the Mn, which indicates the vicinity of Mn 2 O 3 and Fe 2 O 3 nanoparticles. The EDX data displayed peaks in only the desired atoms, which reveals the absence of impurities during the synthesis and preparation of the nanocomposite. It should be noted that the presence of oxygen in high density can be attributed to adsorbed oxygen. The selected area electron diffraction (SAED) pattern is shown in Figure 1h. Since the silica structure is amorphous, the observed diffraction spots/rings corresponded to metal oxides, which are in accordance with XRD analysis.

Texture and Morphology of the Pristine Nanocomposite
In order to evaluate the successful formation of the desired nanocomposite, the texture and morphology of the pristine nanocomposite were analyzed by field emission scanning electron microscopy (FE-SEM) and high-resolution transmission electron microscope (HR-TEM). FE-SEM was analyzed using Hitachi S-4300, Japan, where the dried nanocomposites were coated with a gold-platinum alloy by ion-sputtering (E-1048 Hitachi ion sputter). The FE-SEM image of the nanocomposite showed a highly porous surface with quasi spheroid morphology (Figure 1a). The HR-TEM images were taken with a field emission TEM (FE-TEM, JEM-2010F, JEOL, Tokyo, Japan). The two-dimensional (2D) elemental mapping was performed by energy-dispersive X-ray spectroscopy (EDX) (X-Maxn 80 T, Oxford, UK). The presence of the metal oxide particles in the silica matrices was confirmed in the HR-TEM image shown in Figure 1b. The Figure 1b inset shows that the crystallite planes were assigned by estimating the fringe width values on the HR-TEM image and then relating it with the inter-planer spacing values from the XRD. The lattice spacing of 0.26 nm matched with Bragg reflection from (104) and (222) planes of Fe2O3 and Mn2O3, respectively. The results showed that the MOS nanocomposite consists of uniformly distributed Si, Mn, Fe, and O (Figure 1c-g). The 2D mapping of Fe overlapped with the Mn, which indicates the vicinity of Mn2O3 and Fe2O3 nanoparticles. The EDX data displayed peaks in only the desired atoms, which reveals the absence of impurities during the synthesis and preparation of the nanocomposite. It should be noted that the presence of oxygen in high density can be attributed to adsorbed oxygen. The selected area electron diffraction (SAED) pattern is shown in Figure 1h. Since the silica structure is amorphous, the observed diffraction spots/rings corresponded to metal oxides, which are in accordance with XRD analysis.

Performance of Pristine and UV-Treated Nanocomposites for Removal of ERY
The pristine and UV-treated nanocomposites were tested for their ability to remove ERY under adsorption and heterogeneous photo-Fenton processes. Figure 2a shows that in the first cycle the pristine nanocomposite exhibited excellent removal performance, about 99.58%, and 99.47% for degradation and adsorption respectively. However, this performance decreased for the last cycles and reached 25.12% for adsorption. On the other hand, the efficiency was 82.24% for the last cycle in the degradation process. The kinetic model for degradation has been discussed in Supplementary Figure S1a. Removal of ERY using UV and UV/H 2 O 2 as blank experiments also has been shown in Supplementary Figure S1b.

XRD:
The change in the nanocomposite structure after the UV irradiation (20 and 180 min) was investigated by XRD ( Figure 3). The diffraction patterns were obtained using the Ultima IV (Rigajku, Tokyo, Japan) X-ray diffractometer (XRD) with Cu Kα and a Ni filter, and the scanning speed was set to 3° min −1 . The XRD patterns showed a broad diffraction peak at 22.34°, attributed to the amorphous SiO2 network [23]. Only some peaks were designated to Fe2O3 and Mn2O3 crystals due to the low content of Fe and Mn. However, the diffraction peak at 2θ of 32.96° is attributed to α-Mn2O3 (the most intense 222 Bragg reflection) (JCPDS No. 24-508). Two distinct diffraction peaks were observed at 2θ, of 32.96° and 35.30° (the most intense Bragg reflection in 104 and 110), which are attributed to the α-Fe2O3 (JCPDS No. 33-0664) [24]. There was an insignificant change in the XRD spectrum of the MOS after UV exposure, with a slight shift of the peaks towards higher 2θ values. Moreover, the intensity of diffraction peaks slightly increased after UV irradiation. However, the UV exposure did not induce crystal modification or any new crystal Possible mechanisms for adsorption: Antibiotics' adsorption onto the nanocomposite surface depends on their electrostatic nature and plays an important role, especially in the low initial concentrations. For ERY, different charge forms of ERY can be expected when pH is higher or lower than the pKa (i.e., 8.8.). Similarly, the charge on the nanocomposite surface differs when pH is greater or less than that of the potential of zero charges of the nanocomposite. Hence, the surface charge of the nanocomposite and also the structure of the ERY should be studied. The zeta potential of the nanocomposite has been shown in Supplementary Figure S1. The zero-point charge is the pH when the surface charge of the nanocomposite is zero. The pH drift method was utilized to determine the pHzpc value in this study. The pHzpc was 6.00 for MOS; therefore, MOS has a positive charge if the pH is <6, while the surface is negatively charged if the pH is >6. Since the pH of the solution is 7.3, the nanocomposite surface is negatively charged. The chemical structure of the ERY is shown in Supplementary Figure S1. ERY includes the erythronolide portion, which contains 14 membered lactone rings and 2 sugar-containing polyhydroxylactones. D-desosamine amino sugar is attached to the C-5 portion of the lactone ring. The other sugar, called L cladinose, is attached to the C-3 portion of the lactone ring firmly through a b-glycosidic bond. The protonation/deprotonation of the ERY is based on their pKa values, which control the adsorption of the ERY on the nanocomposite surface due to electrostatic forces. For pH below the pKa values, ERY contains positively charged groups in its structures, while at higher pH values, it is neutral [17]. Thus, at pH 7.3, ERY has a positive charge. On the other hand, according to the zeta potential results, the nanocomposite's surface has a negative charge in this pH. Considering the above, an electrostatic attraction will occur, which will result in better dispersion of the ERY on the nanocomposite surface, improving its adsorption. Moreover, the high specific surface can also improve its adsorption capability. Another possible reason for the high adsorption of ERY on MOS can be the formation of a hydrogen bond between the hydroxyl groups of Si with the oxygens of ERY molecules [18].
Proposed mechanisms for degradation: The hydroxyl radical ( • OH) generated from H 2 O 2 , surface-bound • OH ( • OH surf ), hole (h + ), and the electron (e − ) superoxide ( • O 2 − ) can be considered as reactive species responsible for ERY degradation. Detailed experiments for the main reactive species have been discussed in the Supplementary Figure S2 [1,17,19].
Regeneration/reusability: Erythromycin removal, including adsorption and degradation, was performed for 10 consecutive cycles, in which after every cycle, the catalyst was washed with DI water, dried in the oven (100 • C), and irradiated with UV-C for 20 min. The regeneration and the reusability experiments were conducted under the same conditions mentioned in Section 3.3, except for the catalyst dosage, which was used at 1 g L −1 to consider the mass loss during the process. The degradation trend indicated enhanced performance and stability of the nanocomposite under the heterogeneous photo-Fenton reaction, which maintained the performance until the last cycle and showed 91.76% removal even after 10 cycles of operation. For the adsorption process, the efficiency of each cycle was also significantly enhanced compared to the pristine nanocomposite. However, it decreased to 51.18% in the last cycle, probably due to some deposition occurring on the nanocomposite surface ( Figure 2b). Table 1 shows the removal efficacy of ERY in this work and in the literature.

XRD:
The change in the nanocomposite structure after the UV irradiation (20 and 180 min) was investigated by XRD ( Figure 3). The diffraction patterns were obtained using the Ultima IV (Rigajku, Tokyo, Japan) X-ray diffractometer (XRD) with Cu Kα and a Ni filter, and the scanning speed was set to 3 • min −1 . The XRD patterns showed a broad diffraction peak at 22.34 • , attributed to the amorphous SiO 2 network [23]. Only some peaks were designated to Fe 2 O 3 and Mn 2 O 3 crystals due to the low content of Fe and Mn. However, the diffraction peak at 2θ of 32.96 • is attributed to α-Mn 2 O 3 (the most intense 222 Bragg reflection) (JCPDS No. 24-508). Two distinct diffraction peaks were observed at 2θ, of 32.96 • and 35.30 • (the most intense Bragg reflection in 104 and 110), which are attributed to the α-Fe 2 O 3 (JCPDS No. 33-0664) [24]. There was an insignificant change in the XRD spectrum of the MOS after UV exposure, with a slight shift of the peaks towards higher 2θ values. Moreover, the intensity of diffraction peaks slightly increased after UV irradiation. However, the UV exposure did not induce crystal modification or any new crystal formation since no new peaks were observed in the XRD patterns after UV irradiation. Similar results have been reported in a study conducted by Wang et al., where the intensity of diffraction peaks in polyethylene (HDPE)/titanium dioxide (TiO 2 ) composites increased after UV irradiation [25]. The crystallite size (D) of nanocomposites was calculated using the Debye-Scherer formula (D = (kλ/βcos θ)), where k is the shape factor (0.9), λ is the X-ray wavelength (0.15418 nm, CuK α ), β is the full width at half maximum (FWHM), and θ is the diffraction angle [26]. The average diameter of the crystallite size was about 21 nm for the pristine sample, which showed a slight increase and became 23 nm for UV-irradiated samples. This change can be due to the slight increase in the intensity of diffraction peaks after UV irradiation. The lattice parameters for α-Fe 2 O 3 were a (5.00 Å) and c (13.6 Å), and for α-Mn 2 O 3 was 9.2 Å, which have not been changed, suggesting that UV light does not affect the lattice parameters.
Catalysts 2022, 12, x FOR PEER REVIEW 6 of 14 using the Debye-Scherer formula (D = (kλ/βcos θ)), where k is the shape factor (0.9), λ is the X-ray wavelength (0.15418 nm, CuKα), β is the full width at half maximum (FWHM), and θ is the diffraction angle [26]. The average diameter of the crystallite size was about 21 nm for the pristine sample, which showed a slight increase and became 23 nm for UVirradiated samples. This change can be due to the slight increase in the intensity of diffraction peaks after UV irradiation. The lattice parameters for α-Fe2O3 were a (5.00 Å) and c (13.6 Å), and for α-Mn2O3 was 9.2 Å, which have not been changed, suggesting that UV light does not affect the lattice parameters.

FTIR:
In order to obtain a better understanding of the structure and composition of the pristine and UV-irradiated samples, the Fourier-transform infrared (FTIR) analysis was carried out in the range of 400-4000 cm -1 (Figure 4). FTIR spectra of samples were recorded using KBr pellets over a FTIR spectrometer (Cary670, USA, California, Agilent). The recording was performed with a single-beam spectrometer with 60 added scans. All of the samples showed peaks at the same location with various intensities. The absorption band is related to antisymmetric stretching of Si-O-Si (1088 cm −1 ), and the symmetric mode of Si-O-Si (810 cm −1 ), stretching vibration of Si-OH (945 cm −1 ), and bending of Si-O (457 cm −1 ) were observed, which implies the SiO2 network [27]. It should be noted that the peak at 945 cm −1 was weakened because the Si-O-Si bond and SiO2 network were formed by dewatering of the silanol group in the calcination process [19]. In the FTIR spectrum of nanocomposites, the band at 3446 cm −1 is due to the surface hydroxyl groups, and the lowintensity band at 1628 cm −1 is attributed to the O-H bending mode of the adsorbed water molecules [28]. Stretching vibration of the M-O bond (metal oxides) in the MO6 octahedron and MO4 tetrahedron indicates the presence of the Si-O bond with metals, which is hidden here due to overlapping with the Si-O band at 457 cm −1 . A low-intensity peak at 1711 cm −1 was ascribed to the carboxylic group stretching vibration of C=O due to the fuels used in this work [19]. A significant change has been observed in the FTIR spectra of UVirradiated samples in all the peaks (Figure 4a). The increased intensity at ∼945 cm −1 (corresponding to Si-OH) is due to the formation of new hydroxyl groups, which was confirmed by the increased intensity of the 3435 cm −1 band shown in Figure 4b. The strained Si-O-Si bonds may have shattered and rearranged as a SiO2 network with higher content

FTIR:
In order to obtain a better understanding of the structure and composition of the pristine and UV-irradiated samples, the Fourier-transform infrared (FTIR) analysis was carried out in the range of 400-4000 cm −1 (Figure 4). FTIR spectra of samples were recorded using KBr pellets over a FTIR spectrometer (Cary670, USA, California, Agilent). The recording was performed with a single-beam spectrometer with 60 added scans. All of the samples showed peaks at the same location with various intensities. The absorption band is related to antisymmetric stretching of Si-O-Si (1088 cm −1 ), and the symmetric mode of Si-O-Si (810 cm −1 ), stretching vibration of Si-OH (945 cm −1 ), and bending of Si-O (457 cm −1 ) were observed, which implies the SiO 2 network [27]. It should be noted that the peak at 945 cm −1 was weakened because the Si-O-Si bond and SiO 2 network were formed by dewatering of the silanol group in the calcination process [19]. In the FTIR spectrum of nanocomposites, the band at 3446 cm −1 is due to the surface hydroxyl groups, and the low-intensity band at 1628 cm −1 is attributed to the O-H bending mode of the adsorbed water molecules [28]. Stretching vibration of the M-O bond (metal oxides) in the MO 6 octahedron and MO 4 tetrahedron indicates the presence of the Si-O bond with metals, which is hidden here due to overlapping with the Si-O band at 457 cm −1 . A low-intensity peak at 1711 cm −1 was ascribed to the carboxylic group stretching vibration of C=O due to the fuels used in this work [19]. A significant change has been observed in the FTIR spectra of UV-irradiated samples in all the peaks (Figure 4a). The increased intensity at ∼945 cm −1 (corresponding to Si-OH) is due to the formation of new hydroxyl groups, which was confirmed by the increased intensity of the 3435 cm −1 band shown in Figure 4b. The strained Si-O-Si bonds may have shattered and rearranged as a SiO 2 network with higher content of OH groups, which can be one of the reasons for enhanced performance of the nanocomposite, especially in the adsorption process since it can increase the number of the hydrogen bonds between the hydroxyl groups of Si with the oxygens of ERY molecules.
It should be noted that this finally brings more ERY molecules in the vicinity of the MOS and increases the chance of degradation along with adsorption.
Catalysts 2022, 12, x FOR PEER REVIEW 7 of 14 of OH groups, which can be one of the reasons for enhanced performance of the nanocomposite, especially in the adsorption process since it can increase the number of the hydrogen bonds between the hydroxyl groups of Si with the oxygens of ERY molecules. It should be noted that this finally brings more ERY molecules in the vicinity of the MOS and increases the chance of degradation along with adsorption. BET: Porosity properties and specific surface area of pristine and UV-irradiated nanocomposites were analyzed by the N2 adsorption-desorption isotherm ( Figure 5). For Brunauer-Emmett-Teller (BET) analysis, the N2 adsorption-desorption isotherm was measured at 77 K in a Gemini series Micromeritics 2360 instrument. Before subjecting the nanocomposites for BET analysis, samples were degassed at 473 K for 2 h with a Micromeritics FlowPrep 060. As shown in Figure 5, the sorption isotherm of all nanocomposites was type IV with an H2(b) hysteresis loop in the relative pressure range of 0.45-0.9, which belongs to the structure of mesoporous materials [29]. Based on the Brunauer-Emmett-Teller method, pristine MOS has a specific surface area (SBET) of 891.58 m 2 /g, which increased to 1025.19 and 1038.67 for 20 and 180 min of UV irradiation. BET: Porosity properties and specific surface area of pristine and UV-irradiated nanocomposites were analyzed by the N 2 adsorption-desorption isotherm ( Figure 5). For Brunauer-Emmett-Teller (BET) analysis, the N 2 adsorption-desorption isotherm was measured at 77 K in a Gemini series Micromeritics 2360 instrument. Before subjecting the nanocomposites for BET analysis, samples were degassed at 473 K for 2 h with a Micromeritics FlowPrep 060. As shown in Figure 5, the sorption isotherm of all nanocomposites was type IV with an H2(b) hysteresis loop in the relative pressure range of 0.45-0.9, which belongs to the structure of mesoporous materials [29]. Based on the Brunauer-Emmett-Teller method, pristine MOS has a specific surface area (S BET ) of 891.58 m 2 /g, which increased to 1025.19 and 1038.67 for 20 and 180 min of UV irradiation. of OH groups, which can be one of the reasons for enhanced performance of the nanocomposite, especially in the adsorption process since it can increase the number of the hydrogen bonds between the hydroxyl groups of Si with the oxygens of ERY molecules. It should be noted that this finally brings more ERY molecules in the vicinity of the MOS and increases the chance of degradation along with adsorption. BET: Porosity properties and specific surface area of pristine and UV-irradiated nanocomposites were analyzed by the N2 adsorption-desorption isotherm ( Figure 5). For Brunauer-Emmett-Teller (BET) analysis, the N2 adsorption-desorption isotherm was measured at 77 K in a Gemini series Micromeritics 2360 instrument. Before subjecting the nanocomposites for BET analysis, samples were degassed at 473 K for 2 h with a Micromeritics FlowPrep 060. As shown in Figure 5, the sorption isotherm of all nanocomposites was type IV with an H2(b) hysteresis loop in the relative pressure range of 0.45-0.9, which belongs to the structure of mesoporous materials [29]. Based on the Brunauer-Emmett-Teller method, pristine MOS has a specific surface area (SBET) of 891.58 m 2 /g, which increased to 1025.19 and 1038.67 for 20 and 180 min of UV irradiation. Parameters such as S BET , pore volume, and average pore diameter have been listed in Table 2. Total pore volume has also slightly increased from 0.98 to 1.039 and 1.059. However, the pore diameter decreased after subjecting the pristine nanocomposite to UV irradiation. In the present case, the reconfiguration of the SiO 2 network occurred in a non-uniform way, leading to a decreased pore diameter and increased surface properties. Since the surface area was high and increased even more after UV exposure, it resulted in successive adsorption-degradation cycles. Another reason for the enhanced performance in the adsorption of ERY is that an increase in pore volume increases the chances of ERY diffusion, which directly favors the adsorption. This diffusion can also affect the degradation process by better scattering the ERY on the active sites that exist on the surface, which finally results in better degradation. Moreover, oxidation of residual organic molecules on the surface and pores of the nanocomposite can also be considered as another reason for the regeneration ability of the UV-treated MOS.

HRXPS:
For the elemental oxidation state and surface chemical composition before and after UV exposure, nanocomposites were examined using X-ray photoelectron spectroscopy (XPS) analysis. A K-alpha XPS instrument (Thermo Scientific Inc., East Grinstead, England, UK) with a monochromatic Al Kα X-ray source was utilized, and the pressure was chosen to be 4.8 × 10 −9 mbar. High-energy resolution scans of the Si 2p, Mn 2p, Fe 2p, and O 1s, of pristine and UV-irradiated nanocomposites and associated components, are shown in Figure 6a-l. It should be noted that the concentration of Fe and Mn was low, which resulted in the absence of peaks in the full spectra in Figure 6a [30].
The fraction of Si-O bonds decreased by 30∼35% for the UV-irradiated samples after 20 and 180 min of UV exposure, respectively. The area under the O-H peak increased by 60∼64% after irradiation (Figure 6e,f), and the increase in the OH density in the FTIR spectrum supports this fact. These results strongly suggested that the SiO 2 network was affected by UV photons and resulted in structural modifications. The decreased area for the Si-O curve implies that part of the Si-O-Si bonds could be broken and generate new Si-OH groups. As mentioned in the FTIR analysis, the generation of the new hydroxyl groups can be involved in the enhanced performance of nanocomposites in removing ERY. It should be noted that enhanced performance after increasing the number of OH groups indicates that surface-bound • OH ( • OH surf ) was one of the main reactive species responsible for the degradation process.
The Fe 2p spectrum of the pristine sample has been shown in Figure 6g, and the two peaks centered at 710.53 eV for Fe 2p 3/2 and 724.55 eV for Fe 2p 1/2 show that Fe exists in the form of Fe 3+ [31].

ESR:
The ESR spectra have been used to detect different oxidation states, metal complexes, and chemical environments of atoms. The ESR spectra at room temperature before and after UV irradiation are shown in Figure 7a-c. It can be seen that the ESR spectra showed a single broad signal in all of the nanocomposites, and no new signal was observed. The g value of 2.02-2.03 shows the combined contribution of Fe as FexOy clusters or Fe2O3 nanoparticles in the silica matrix [33]. It should be noted that isolated Fe 3+ shows the signal at 4.30, and the absence of this signal is proof that isolated Fe 3+ did not form after UV exposure. For the Mn, no ESR signal was found in any of the samples since Mn 3+ is ESR-silent, and it did not change for UV-irradiated samples, suggesting that Mn remained in the same state. It should be noted that other forms of manganese have an intense sextet which is not found here, further proving that Mn stayed in the form of Mn 3+ . These findings support our findings by XPS.

ESR:
The ESR spectra have been used to detect different oxidation states, metal complexes, and chemical environments of atoms. The ESR spectra at room temperature before and after UV irradiation are shown in Figure 7a-c. It can be seen that the ESR spectra showed a single broad signal in all of the nanocomposites, and no new signal was observed. The g value of 2.02-2.03 shows the combined contribution of Fe as Fe x O y clusters or Fe 2 O 3 nanoparticles in the silica matrix [33]. It should be noted that isolated Fe 3+ shows the signal at 4.30, and the absence of this signal is proof that isolated Fe 3+ did not form after UV exposure. For the Mn, no ESR signal was found in any of the samples since Mn 3+ is ESR-silent, and it did not change for UV-irradiated samples, suggesting that Mn remained in the same state. It should be noted that other forms of manganese have an intense sextet which is not found here, further proving that Mn stayed in the form of Mn 3+ . These findings support our findings by XPS.

Degradation By-Products Using the Regenerated Nanocomposite after the Last Cycle
The reaction by-products using pristine nanomaterials have been studied in past decades. However, to the best of our knowledge, there is no study using regenerated nanocomposites after the last cycle. To further study the quality of regenerated nanocomposites after the last cycle, by-products of the heterogeneous photo-Fenton reaction after 20 and 120 min have been studied using LC/MS/MS, and the detailed methodology is shown in Supplementary Table S1. The results imply that after a 20 min irradiation time, ERY was significantly decomposed. This is evident by the decrease in the ERY peak intensity and the observation of new peaks corresponding to ERY photocatalytic products. Figure 8 shows a full-scan mass spectrum initial ERY, infused at 0.5 mL min −1 using a syringe pump. The identification of transformation products focused on the most abundant compounds formed during the first 20 min of the reaction, according to their molecular masses, retention times, and empirical formulas. The MS spectra of the parent ERY reside on its molecular masses [MH] + , ERY (734.4681) [18]. In this work, the two most abundant intermediate s(m/z: 421 and m/z: 290) were observed after 20 min of reaction. However, after 120 min, these peaks disappeared, due to the successful adsorption of by-products by the highly porous nanocomposite used in this study.

Degradation By-Products Using the Regenerated Nanocomposite after the Last Cycle
The reaction by-products using pristine nanomaterials have been studied in past decades. However, to the best of our knowledge, there is no study using regenerated nanocomposites after the last cycle. To further study the quality of regenerated nanocomposites after the last cycle, by-products of the heterogeneous photo-Fenton reaction after 20 and 120 min have been studied using LC/MS/MS, and the detailed methodology is shown in Supplementary Table S1. The results imply that after a 20 min irradiation time, ERY was significantly decomposed. This is evident by the decrease in the ERY peak intensity and the observation of new peaks corresponding to ERY photocatalytic products. Figure 8 shows a full-scan mass spectrum initial ERY, infused at 0.5 mL min −1 using a syringe pump. The identification of transformation products focused on the most abundant compounds formed during the first 20 min of the reaction, according to their molecular masses, retention times, and empirical formulas. The MS spectra of the parent ERY reside on its molecular masses [MH] + , ERY (734.4681) [18]. In this work, the two most abundant intermediate s(m/z: 421 and m/z: 290) were observed after 20 min of reaction. However, after 120 min, these peaks disappeared, due to the successful adsorption of by-products by the highly porous nanocomposite used in this study. Catalysts 2022, 12, x FOR PEER REVIEW 11 of 14 Figure 8. ERY degradation by-products using regenerated nanocomposites after the last cycle.

Synthesis of Metal Oxide Silica Nanocomposites
The Mn2O3-Fe2O3@SiO2 nanocomposites were prepared according to our previously published procedure with slight modifications [19]. In summary, Mn2O3-Fe2O3@SiO2 nanocomposites were synthesized using the sol-gel/combustion method. In a typical synthesis, 1 g of Mn(NO3)2·4H2O and 1.16 g of Fe(NO3)3·9H2O were added to 60 mL of cetyltrimethylammonium chloride (25% H2O) and mixed for 30 min at 40 °C. Then, 2 g of urea (fuel) was added to the solution and mixed for another 30 min. In the next step, 40 mL of TEOS was added dropwise to the solution. The solution was incubated at 60 °C for 24 h. Then, the sample was calcined at 550 °C for 6 h. The obtained metal oxide silica was coded as MOS.

Experimental Procedure
Erythromycin (ERY) was used as a module pollutant to examine the performance of pristine and regenerated nanocomposites under heterogeneous photo-Fenton and adsorption processes. For this purpose, 200 mL of ERY solution (2 mg L −1 ) was poured into a quartz glass, and then 25 mmol L −1 of 28% H2O2 was added to the solution. Then, 100 mg of the nanocomposite was added to the previous solution. No pH adjustments were made, Figure 8. ERY degradation by-products using regenerated nanocomposites after the last cycle.

Synthesis of Metal Oxide Silica Nanocomposites
The Mn 2 O 3 -Fe 2 O 3 @SiO 2 nanocomposites were prepared according to our previously published procedure with slight modifications [19]. In summary, Mn 2 O 3 -Fe 2 O 3 @SiO 2 nanocomposites were synthesized using the sol-gel/combustion method. In a typical synthesis, 1 g of Mn(NO 3 ) 2 ·4H 2 O and 1.16 g of Fe(NO 3 ) 3 ·9H 2 O were added to 60 mL of cetyltrimethylammonium chloride (25% H 2 O) and mixed for 30 min at 40 • C. Then, 2 g of urea (fuel) was added to the solution and mixed for another 30 min. In the next step, 40 mL of TEOS was added dropwise to the solution. The solution was incubated at 60 • C for 24 h. Then, the sample was calcined at 550 • C for 6 h. The obtained metal oxide silica was coded as MOS.

Experimental Procedure
Erythromycin (ERY) was used as a module pollutant to examine the performance of pristine and regenerated nanocomposites under heterogeneous photo-Fenton and adsorption processes. For this purpose, 200 mL of ERY solution (2 mg L −1 ) was poured into a quartz glass, and then 25 mmol L −1 of 28% H 2 O 2 was added to the solution. Then, 100 mg of the nanocomposite was added to the previous solution. No pH adjustments were made, and the pH of the solution was 7.3. Finally, the mixture was irradiated with UV light. The experiments were conducted using a black acrylic reactor with a quartz tube in the middle and four UV lamps parallel to the quartz tube. During the experiments, the reactor temperature was maintained at 25 • C using cooling fans connected to the bottom of the reactor. The UV-C lamp (8 W, λmax = 254 nm, Philips) was used as the irradiation source. Samples were collected at specific time intervals and filtered to eliminate the particles. Adsorption experiments were conducted under the same conditions without light and H 2 O 2 .
Detection and analysis: Erythromycin concentration was analyzed by LC/MS/MS, Thermo Vanquish (Thermo Scientific, Waltham, MA, USA), and the column used was Waters Cortects C18 (2.1 × 100, 1.6 µm). The detailed analysis methodology has been tabulated in Supplementary Materials Table S1.

Conclusions
In this study, the effect of UV treatment on the structural variations and regeneration of Mn 2 O 3 -Fe 2 O 3 @SiO 2 was studied by experimental and spectroscopic analysis. It was evident that there was a significant change in the silica network, hydroxyl groups, and porosity properties of the nanocomposite after UV exposure. The pristine nanocomposite showed an excellent ERY adsorption/degradation performance. The high adsorption efficiency was due to electrostatic attraction, high surface area, and the possibility of hydrogen bonding. The high degradation efficiency was due to the high surface area and reactive species produced in the degradation reaction. However, the pristine nanocomposite could not retain its performance for ten cycles. On the other hand, the UV-irradiated nanocomposite showed enhanced removal efficiency for ten consecutive cycles. FTIR and XPS analysis showed that silica network reconfiguration after UV irradiation with an increased hydroxyl density could increase the possible mechanism involved with adsorption/degradation. Enhanced performance after increasing the OH groups could indicate the vital role of surface-bound • OH ( • OH surf ) in the removal process. BET analysis showed that the surface area was high and even increased after UV exposure, resulting in successive adsorption-degradation cycles. Additionally, increased pore volume may have favored the diffusion of ERY molecules.
For this reason, MOS showed an excellent ERY removal performance even after ten cycles. Moreover, these observations can be due to the oxidation of residual organic molecules on the surface or in the pores of the nanocomposite. For the first time, the fate of ERY was studied using regenerated nanocomposites after the last cycle. LC/MS/MS results suggested that the fabricated nanocomposite is not only reusable but can also be a dual functional nanomaterial, which could break down the parent compound and adsorb the formed reaction by-products. This study could be a foundation research for the practical approaches for the regeneration of nanomaterials and the successful removal of organic pollutants from aquatic environments.

Conflicts of Interest:
The authors declare no conflict of interest.