Dual Approach for Effective Photodegradation and Mineralization of Carbamazepine Using a Ternary ZnFe₂O₄/(P,S)‑g‑C₃N₄/Two-Dimensional TiO₂ Composite

Abstract

This study proposes a dual strategy for the efficient degradation of carbamazepine through combining a photocatalytic reaction with persulfate (PS) activation. To improve the physicochemical properties, two magnetic materials with ratios of 50:50 and 70:30 of ZnFe₂O₄ to (P,S)-g-C₃N₄ were combined with TiO₂ nanosheets prepared through a fluorine ion-free lyophilization method. The presence of (P,S)-g-C₃N₄ in a heterojunction with 2D TiO₂ mitigated rapid charge carrier recombination, while the magnetic properties improved the photocatalyst recyclability and supported PS activation. Surface properties, phase identification, and structural and chemical composition of materials were studied using X-ray diffraction (XRD), transmission electron microscopy (TEM), scanning electron microscopy (SEM), and X-ray photoelectron spectroscopy (XPS). Optical properties of photocatalysts were determined by UV–vis diffuse reflectance spectroscopy (DR/UV–vis) and photoluminescence spectroscopy (PL). The effect of PS addition in the photocatalytic system was correlated with the observed generation of hydroxyl radicals. The highest concentration of hydroxyl radicals was observed during the first 15 min of the process. The mechanism investigation confirmed that superoxide and hydroxyl radicals were the primary contributors in the case of photocatalysis. In contrast, for PS-assisted photocatalysis, the major reactive species, depending on the pH, were hydroxyl radicals and sulfate radicals. Six common transformation intermediates were identified, and a degradation pathway was proposed. Overall, this investigation emphasizes the potential and effectiveness of the PS-assisted photodegradation process using a cost-effective ternary ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ composite for improved degradation of carbamazepine, a persistent organic pollutant frequently present in water.

1. Introduction

One of the greatest global problems facing our world is environmental pollution related to urbanization, mining, and industrialization. Since the publication of Rachel Carson’s book “Silent Spring”, people have been more concerned about the presence and the widespread distribution of persistent pollutants in water, soil, and air. The main contributors to environmental contamination are agricultural, petrochemical, textile, food processing and pharmaceutical industries. − Some of the pollutants from anthropogenic sources, which are persistent and susceptible to bioaccumulation and biomagnification, are distributed in the ecosystem, posing risks to living organisms. Therefore, identifying and monitoring pollutants that pose an environmental risk has consequently been introduced worldwide.

The major water pollutants of emerging concern are mainly classified as pesticides, pharmaceuticals, industrial chemicals, and heavy metals. , Among them, some of the commonly used pharmaceuticals for human and animal health are found in wastewater influents and effluents almost in an unchanged form. Conventional treatment technologies have proven to be ineffective in completely eliminating and removing these active pharmaceutical ingredients. For example, carbamazepine is one of the most frequently detected pharmaceuticals in water bodies and is thus proposed as an anthropogenic marker of water contamination. ,

Considering the pollution of water with pharmaceutical residues, there is a need for the development of more advanced treatment technologies to efficiently remove these compounds from wastewater before they are discharged into the environment. In this regard, improving water treatment technologies is a proactive approach of high importance to increasing water quality and ensuring the sustainability of water cycle management.

The advanced oxidation processes (AOP) based on the in situ generation of the strongest oxidants, hydroxyl radicals and sulfate radicals, have been recognized as a promising approach for wastewater treatment from residues of recalcitrant and emerging organic contaminants. In recent years, ozonation, the Fenton process, and UV irradiation have already been established and started at a technological scale for the treatment of drinking water and water reuse facilities. Nevertheless, ozonation is restricted by the high cost of ozone, while the Fenton reaction usually operates under a limited pH range (3–4) and generates Fe­(OH)₃ precipitation. Therefore, among the various AOP processes, photocatalysis utilizing sunlight is a promising method for environmental remediation.

Although great progress has been achieved in the field of solar-driven photocatalytic degradation of persistent organic pollutants, there are still some challenges concerning (i) photocatalyst design for the particular application, (ii) separation of the semiconductor particles from the post-treatment suspension, (iii) low quantum efficiency associated with charge transfer, and (iv) reactor design enabling efficient light penetration and exposure of the photocatalyst to the light.

The critical properties that determine the activity of photocatalysts are mainly defined by composition and crystal structure. The degradation efficiency is dependent on the morphological properties of the photocatalyst, the absorption capacity of the substrate, and the wavelengths of incident irradiation. Furthermore, parameters such as highly developed specific surface area, particle size, shape, and crystallinity have a significant influence on photocatalytic activity. In recent years, two-dimensional (2D) semiconductors have demonstrated great potential in photodegradation processes, as they provide a larger surface area for enhanced reactant adsorption and reduce charge diffusion paths, resulting in improved separation of photoinduced charge carriers. For example, it has been reported that the thickness of g-C₃N₄ influences both the photophysical properties and the kinetics of charge carriers. Moreover, the hybridization of organic and inorganic components can improve charge separation and reduce electron–hole recombination, thereby prolonging the lifetimes of the photogenerated electrons and holes. Furthermore, the improvements can extend light absorption of the material into the visible range, thereby maximizing the utilization of the solar spectrum. , Since g-C₃N₄ is a visible light semiconductor, hybridization with UV light active semiconductors enables the utilization of these materials in solar light applications. For example, the combination of g-C₃N₄ with wideband gap zinc oxide modified with gold particles formed a hybrid material of ZnO/Au/g-C₃N₄, which led to a high removal efficiency of organic compounds. In another study, the heterocomposite of g-C₃N₄/ZnO resulted in 93% degradation of bisphenol A (BPA), whereas degradation of 78% and 83% was observed in the presence of ZnO and g-C₃N₄ separately.

In this investigation, we used titanium dioxide (TiO₂) as a widely utilized n-type semiconductor for the photodegradation of carbamazepine. To enhance the adhesion of organic molecules and improve overall functionality, we synthesized nanosheet-shaped TiO₂ using the fluorine-free lyophilization method. Furthermore, the hybridization of TiO₂ with the two-dimensional, metal-free semiconductor graphitic carbon nitride (g-C₃N₄), which has a narrower bandgap of 2.7 eV, was applied to enhance the photocatalyst response in visible light. The structure of g-C₃N₄ consists of earth-abundant elements such as carbon and nitrogen. Connections of tri-s-triazine units with planar amino groups contribute to the chemical stability of graphitic carbon nitride. The internal electric field at the interface between g-C₃N₄ and TiO₂ can lower the recombination rate of photogenerated charged carriers.

Addressing the technical challenges is crucial for unlocking the transformative power of photocatalysis in environmental remediation. One of the key bottlenecks hindering the practical application of photocatalysis is the reduced recyclability of the photocatalyst, which limits its continuous use. Therefore, to overcome this limitation, we incorporated spinel zinc ferrite nanoparticles with superparamagnetic properties into the g-C₃N₄ structure, enhancing the reuse and recyclability of the photocatalyst. While photocatalysis shows promise for degrading persistent organic pollutants, it also faces challenges related to the kinetics of the photodegradation process, resulting in longer reaction times, which hinder its full potential. As a solution, combining photocatalysis with other advanced oxidation processes can improve the degradation efficiency of persistent and emerging contaminants. The combination of photocatalysis with ozonation and Fenton/Fenton-like reactions has been studied for its potential to enhance wastewater treatment efficiency. , Photocatalytic ozonation can enhance the mass transfer limitations associated with catalysts and produce a higher concentration of hydroxyl radicals (•OH). However, due to the main drawbacks of ozonation, alternative solutions are being investigated. In recent years, persulfates have attracted significant interest due to their ability to be activated under photocatalytic conditions. Sodium persulfate (PS) acts as a precursor for hydroxyl and sulfate radicals, which possess the highest redox potential (E ₀ = 2.5–3.1 V) among in situ-generated reactive species in water. PS is a soluble compound in water (solubility of 556 g/dm³ at 20 °C) with a relatively high stability. It has a high bond energy of 140 kJ/mol, making it resistant to direct reactions with organic matter. Therefore, efficient activation of PS is crucial. Various methods can be employed for PS activation, including light exposure, heating, ultrasonication, or reduction with transition metals. Also, catalytic materials can be utilized to enhance the activation of PS. In our study, the fabricated heterocomposite of ZnFe₂O₄/g-C₃N₄/2D TiO₂, which incorporates transition metals and carbon-based materials, exhibits improved separation capabilities and shows promise in the activation of persulfates for the degradation of emerging contaminants in water through the persulfate-assisted photocatalysis process.

The mechanism and kinetics of carbamazepine degradation in the persulfate-assisted photocatalysis process were investigated. Particularly, the effect of PS concentration (0.5, 1.5, and 3 mM), amount of ZnFe₂O₄, TiO₂ nanosheets, and (P,S)-g-C₃N₄ in the heterocomposite structure on physicochemical and photocatalytic properties was studied in detail.

2. Experimental Section

2.1. Materials

Iron nitrate nonahydrate (≥98%), zinc nitrate hexahydrate (≥99%), 2-propanol, and all scavenging agents used for quenching experiments were provided by Sigma-Aldrich (Germany). HPLC-grade acetonitrile and orthophosphoric acid (85+%, HPLC grade) were purchased from Merck (Germany). Sodium acetate (ACS pure S.A.) and ethylene glycol (≥99%) were obtained from POCH (Poland) and CHEMPUR (Poland). All chemicals were used as received without further purification. Sodium persulfate (PS) was used as sodium peroxydisulfate (≥98%, Reagent grade) and was purchased from Sigma-Aldrich (Germany). Melamine (≥99%) was obtained from Alfa Aesar whereas thiourea (≥99%) and ammonium phosphate dibasic (NH₄)₂HPO₄ (≥99%) were purchased from Merck. Titanium oxysulfate hydrate (Technical grade) was purchased from Merck. Ammonium solution (25 wt %) and hydrogen peroxide (30 wt %) were obtained from POCH (Poland). Sodium chloride (NaCl, ≥99.8%), sodium nitrate (NaNO₃, ≥99.5%), and sodium sulfate (Na₂SO₄, ≥99%) were purchased from Sigma-Aldrich (Germany). Humic acid was obtained from Merck.

2.2. Synthesis Procedure of Photocatalysts

2.2.1. Synthesis of Zinc Ferrite (ZnFe₂O₄) Nanoparticles

Iron nitrate nonahydrate Fe­(NO₃)₃·9H₂O (4.04 g) and zinc nitrate hexahydrate Zn­(NO₃)₂·6H₂O (1.48 g) were dissolved in ethylene glycol (C₂H₆O₂). Sodium acetate (CH₃COONa; 6.3 g) was added, and the suspension was stirred until it was homogeneously dispersed. After stirring, the suspension was placed into ultrasonication for 2 h, and then it was transferred into a Teflon-lined autoclave at 180 °C for 12 h. After the reaction, the precipitate was collected and washed a few times with anhydrous ethanol and water. The obtained sample was dried at 120 °C for 6 h. The sample was ground into powder and calcined at 500 °C for 2 h at a 5 °C/min rate. A fine dark brown powder with magnetic properties was obtained.

2.2.2. Synthesis of ZnFe₂O₄/(P,S)-g-C₃N₄ Heterocomposite

Two different photocatalysts, denoted as 50:50% w/w and 70:30% w/w ZnFe₂O₄/(P,S)-g-C₃N₄, were obtained to study the effect of the zinc ferrite content on photocatalytic activity and magnetic properties.

First, phosphorus and sulfur-codoped graphitic carbon nitride were obtained through thermal treatment involving thiourea, melamine, and (NH₄)₂HPO₄. The powder was heated at 550 °C for 4 h followed by grinding and heating at 500 °C for 3 h until a fine yellow powder was obtained. In the next step, the magnetic particles were incorporated into the (P,S)-g-C₃N₄ heterocomposite structure to obtain a visible light-active magnetic photocatalyst. In this regard, a defined amount of zinc ferrite was dispersed with doped carbon nitride into 50 cm³ of deionized water (DI) and ultrasonicated for 1 h. After ultrasonication, the suspension was placed into a Teflon-lined reactor and subjected to a hydrothermal reaction at 150 °C for 5 h and at a 5 °C/min rate. The as-obtained sample was dried at 80 °C.

2.2.3. Synthesis of the ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ Heterocomposite

Finally, the heterocomposite ZnFe₂O₄/(P,S)-g-C₃N₄ was combined with 2D TiO₂. In the first step, TiO₂ nanosheets were synthesized through an environmentally sustainable freeze-drying method. At 40 °C, 10 g of titanium oxysulfate hydrate was dissolved in 300 cm³ of deionized water. The solution was cooled in an ice bath until it reached 3 °C. Ammonium hydroxide solution (16%) was added until the pH of 9, and the white precipitate was formed, which was washed several times with deionized water and resuspended in 300 cm³ of deionized water. After adding 60 cm³ of 30% aqueous H₂O₂, pH was decreased from 8 to 2, and the color was changed from white to yellow. The as-obtained gel was aged for 48 h at 3 °C and further lyophilized at −84 °C and a pressure of 0.3 Pa. The dry powder was calcinated at 700 °C for 1 h as the final stage in the formation of TiO₂ nanosheets. Then, the heterocomposite of ZnFe₂O₄/(P,S)-g-C₃N₄ was prepared by adding ZnFe₂O₄/(P,S)-g-C₃N₄ and 2D TiO₂ in 50 cm³ of 2-propanol. After 2h of sonication, 2-propanol was removed by centrifugation, and the sample was dispersed in 50 cm³ of water, placed into a hydrothermal reactor at 150 °C for 5 h and at a 5 °C/min rate, and further dried to obtain the powder material. Obtained powders were noted as ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (5, 10, and 20% w/w), where 5, 10, and 20% are the amounts of ZnFe₂O₄/(P,S)-g-C₃N₄ (70:30) combined with 2D TiO₂.

2.3. Characterization of Photocatalytic Materials

The crystallographic structure and phase identification of photocatalysts were investigated with an X-ray powder diffractometer (Rigaku MiniFlex 600). The spectra were recorded in a 2θ range of 5–80°, at a scan speed of 1°·min^–1 and step of 0.01°. The determination of the crystallite size of the photocatalysts was performed using Debye–Scherrer’s equation:

$$D=K\lambda /\beta \cos \theta$$ 1

where K is a Scherrer constant, λ the wavelength of the X-ray beam, and β and θ represent the full width at half-maximum of the peak and the Bragg angle.

The specific surface area was evaluated by the Brunauer–Emmett–Teller (BET) method using a Micromeritics Gemini V (model 2365) (Norcross, GA, USA). Before adsorption, the synthesized materials were degassed at 200 °C for 2 h in the presence of nitrogen (N₂).

The UV–vis diffuse reflectance spectra were obtained by using a ThermoScientific Evolution 220 spectrophotometer (Waltham, MA, USA) with barium sulfate as a standard reference. The Kubelka–Munk equation [αhν]^1/p = A(hν – E _g) was used to determine the optical bandgap of photocatalysts, where α expresses the optical absorption coefficient, hν is the photon energy, A stands for the proportionality constant, and p is 1/2 for direct transition.

The photoluminescence spectra (PL) were obtained by using a Shimadzu RF-6000 spectrofluorometer (Kyoto, Japan). The excitation source was a 150 W xenon lamp with an excitation wavelength of 315 nm. The Fourier transform infrared spectroscopy (FTIR) was conducted using a Nicolet iS10 (Thermo Fisher Scientific Waltham, MA, USA) spectrometer in the wavelength range of 400–4000 cm^–1 and resolution of 2 cm^–1.

The morphology of selected photocatalysts was analyzed by the transmission electron microscope (TEM) Tecnai 20F X-Twin. Chemical composition was obtained by XPS spectroscopy using a PHI 5000 VersaProbe spectrometer (Scanning ESCA Microprobe ULVAC-PHI). Monochromatic X-rays with energy hν = 1486.6 eV (AlKα) were observed with a power of 25 W of the excitation source. Inspection and high-resolution X-ray photoelectron spectroscopy (XPS) spectra were recorded with the energy matching function 117.4 and 23.5 eV, respectively. The Smart Background Cut function was used to determine the intensity of the individual XPS signals. The recorded spectra in a narrow range of bond energies were fitted using the Gauss/Lorentz function. The analyzed bond energies for individual elements detected on the surface of the tested samples were corrected in relation to the C 1s carbon peak (C–C bond) as an internal standard used during the chemical state determination procedure. Sensitivity coefficients characteristic of a given element and the type of X-ray radiation were used to estimate the quantitative chemical composition of the tested samples.

Electrochemical impedance spectroscopy (EIS) measurements were performed using a Metrohm Autolab potentiostat galvanostat (PGSTAT204; the Netherlands) with 1 M Na₂SO₄ as an electrolyte. The photocatalyst was deposited on the carbon screen-printed electrode with a Ag/AgCl reference electrode (Metrohm Autolab). The diameter of the working electrode area was 4 mm. The AC voltage amplitude was 0.01 V, and the frequency ranged from 0.1 Hz to 100 kHz at 0 V vs OCP (open circuit potential). The Mott–Schottky analysis was performed to determine the flat band (Fb) potential of materials used for the preparation of the ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (10% w/w) composite. EIS data were recorded from anodic toward the cathodic direction for the applied frequency of 1000 Hz in the potential range from 0 to −1.1 V vs Ag/AgCl. The transient photocurrent response of the photocatalysts was recorded at a light-on/off interval of 50 s at 0 V vs Ag under 372 nm LED light illumination.

Magnetic properties of the heterocomposite and zinc ferrite were analyzed using a SQUID magnetometer (Quantum Design MPMS XL7).

A Thermogravimetric Analyzer-2 SF 1100 (Mettler Toledo) was used under the following conditions: nitrogen flow of 10 cm³/min, temperature of 35–650 °C, and heating rate of 10 °C/min.

A leaching analysis of Fe was conducted by using a photometric method. A HACH spectrophotometer (DR 5000) and the LCK 521 method were used to measure the concentration of iron after treatment, to evaluate the stability of the photocatalyst.

2.4. Photocatalytic Activity

Photocatalytic activity of the obtained heterocomposites was studied in a model reaction of carbamazepine (CBZ) degradation. In this regard, a solution of carbamazepine (CBZ, 14 mg·dm ^–3) was placed into a 25 cm³ quartz reactor, and 25 mg of the photocatalyst was added to the solution. A Xe lamp (model 6271H, Oriel, USA) with a water IR filter was used as a UV/vis irradiation source. The UV flux intensity equal to 25 mW·cm^–2 was measured at the border of the reactor at a distance of 40 cm from the irradiation source. During the reaction, a constant airflow of 4 dm³ ·h^–1 was introduced through the suspension, which was stirred and thermostated at 20 °C. Before irradiation, the reaction system was kept in the dark for 30 min to achieve the adsorption–desorption equilibrium. Then, the process was initiated. Photodegradation experiments were conducted for 120 min of irradiation. The degradation efficiency of carbamazepine was monitored using a high-performance liquid chromatography system (HPLC, model Shimadzu LC-6A), combined with a photodiode array detector (SPD-M20A) and C18 column (Phenomenex Gemini 5 μm; 150 × 4.6 mm). Six different products were identified by high-performance liquid chromatography (LC) accompanied by mass spectrometer (MS) in full scan and single ion monitoring modes (SIM).

First, the photodegradation efficiency of CBZ was tested for the samples of ZnFe₂O₄/(P,S)-g-C₃N₄ (50:50% w/w) and (70:30% w/w). Then, the selected photocatalyst was modified with 2D TiO₂ in different loadings of 5, 10, and 20% ZnFe₂O₄/(P,S)-g-C₃N₄ to TiO₂. Finally, the most active heterocomposite was used in the third stage of experiments, which consists of the PS-assisted photodegradation process. For quenching experiments, isopropanol, tert-butyl alcohol (TBA), para- benzoquinone (P-BQ), ammonium oxalate (NH₄)₂C₂O₄, silver nitrate AgNO₃, and sodium azide NaN₃ were used as hydroxyl and sulfate radicals (HO^•, SO₄ ^•–), hydroxyl radicals (HO^•), superoxide radicals (O₂ ^•–), holes (h⁺), electrons (e^–), and singlet oxygen species (¹O₂) scavengers, respectively. The rate constant of CBZ degradation was evaluated according to the pseudo-first-order kinetic model with the following equation:

$$\ln C_0/C=kt$$ 2

where C ₀ and C are the initial concentration and the concentration at a certain reaction time, and k is related to the rate constant, while t is a treatment time in minutes.

The effectiveness of the hybrid oxidation system of photocatalysis and persulfates activation was evaluated via the synergy index:

$$\mathrm{synergy}\mathrm{index}=k(\mathrm{photocatalysis}+\mathrm{PS}3\mathrm{mM})/k(\mathrm{photocatalysis})+k(\mathrm{PS}3\mathrm{mM})$$ 3

where k represents the rate constant of the combined process of photocatalysis and PS activation and the rate constant of separate processes of photocatalysis and PS activation.

3. Results and Discussion

3.1. Characterization of ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ Magnetic Photocatalysts

First, the morphology and surface structure of photocatalysts were studied. Cubic and various crystal shapes of ZnFe₂O₄ with diameters up to 20 nm were observed using transmission electron microscopy and the results are presented in Figure a. 2D TiO₂ appeared to be in the form of nanosheets (Figure b). The average particle size for the ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ composite was about 25–30 nm. Additionally, the interface between materials is presented in Figure S1 in the Supporting Information. Energy-dispersive spectroscopy (EDS) confirmed the presence of Zn, Fe, Ti, C, O, and S in the structures of the photocatalytic material. Based on STEM images and mapping of the selected area, it was noticed that layers of (P,S)-g-C₃N₄ are embedded in 2D TiO₂ and ZnFe₂O₄ structure (see Figure f).

1.

Microscopy analyses of ZnFe₂O₄ (a), FFT of 2D TiO₂ (b), TEM of ZF/CN/2D TiO₂ (10%) (c), particle size distribution (d), and STEM-EDS elemental mapping of the ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (10%) composite (e, f).

As presented in Figure a, the C 1s spectrum consists of three peaks at 284.8, 287.0, and 288.7 eV assigned to C–C, CO, and C–(N)₃ bonds. The observed signals around 398.9 and 400.6 eV corresponded to the sp³ hybridization of C–N (g-C₃N₄) and C–(NH) _x surface amino groups (Figure b). The P 2p region can be deconvoluted into two peaks of P 2 p_3/2 and P 2p_1/2 at 133.3 and 134.3 eV and additionally the Zn 3s peak was observed at 140.1 eV (Figure c). The XPS peaks of Ti 2p_3/2 were identified as Ti⁴⁺ and Ti-Ox with binding energies of 458.8 and 459.0 eV. Assigned peaks of Ti 2p_1/2 were identified at 464.3 and 465.0 eV (Figure e). The O 1s could be deconvoluted into three peaks. The main peak at 529.9 eV is related to oxygen in the TiO₂ crystal lattice (Ti–O–Ti). The less intense peaks at 531.2 and 532.2 eV are assigned to oxygen in P–O/CO and C–O bonds. Characteristic peaks of Zn 2p_3/2 and Zn 2p_1/2 were noticed at 1022.0 and 1045.0 eV, confirming the presence of Zn–O bonds (Figure f). The fitting peaks of 709.8 and 711.5 eV for the Fe 2p_3/2 spectrum in Figure g are ascribed to the Fe–O bond and respective oxidation states of (Fe²⁺) and (Fe³⁺) in tetrahedral and octahedral sites of ZnFe₂O₄.

2.

High-resolution XPS spectra of C 1s (a), N 1s (b), P 2p (c), Ti 2p (d), O 1s (e), Zn 2p (f), and Fe 2p (g) for the ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (10% w/w) composite.

The composition of the crystalline phase was studied by XRD analysis for both as-obtained spinel ferrite and its further modifications of (P,S)-g-C₃N₄, 2D TiO₂, and ZnFe₂O₄/(P,S)-g-C₃N₄/ 2D TiO₂ (5, 10, and 20% w/w) and the results are shown in Figure a. Formation of pure ZnFe₂O₄ was confirmed by the presence of the signals at 2θ = 18.40, 30.06, 35.27, 42.93, 53.25, 56.65, and 62.60°, which can be indexed to the (111), (220), (311), (400), (422), (511), and (440) spinel zinc ferrite diffraction planes according to JCPDS card no. 82–1042. The reflections at 2θ of 25.53, 36.9, 37.7, 48.0, 53.9, and 62.7° corresponded to anatase crystal planes (101), (103), (004), (200), (105), and (204), respectively (JCPDS card no. 21–1272). The diffraction peak at 2θ = 13.2°, corresponding to the diffraction plane of (100) for the (P,S)-g-C₃N₄, was observed only in the single phase sample and vanished for the ZnFe₂O₄/(P,S)-g-C₃N₄ composite. The diffraction signal at 2θ = 27.7° corresponds with the (002) diffraction plane of graphitic carbon nitride and is ascribed to the interlayer stacking structure of the aromatic compound. Identified diffraction peaks of 2D TiO₂ and ZnFe₂O₄ are observed in all samples of ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (5, 10, and 20% w/w) composites. Increasing the amount of ZnFe₂O₄/(P,S)-g-C₃N₄ to 2D TiO₂ from 5% and 10% to 20% resulted in an increased intensity of the characteristic peak of zinc ferrite at 2θ 35.27° and lower intensity of characteristic diffraction peaks of TiO₂ at 2θ = 25.53, 36.9, 37.7, 48.0, 53.9, and 62.7°.

3.

X-ray diffraction patterns of ZnFe₂O₄ (ZF), (P,S)-g-C₃N₄ (CN), 2D TiO₂, and ZF/CN/2D TiO₂ (5%, 10%, 20%) composites (a), DR/UV–vis spectra of ZnFe₂O₄ (ZF), (P,S)-g-C₃N₄ (CN), 2D TiO₂, and ZF/CN/2D TiO₂ (5%, 10%, 20%) composites (b), magnetic hysteresis loop (magnetization M, versus applied magnetic field) of ZnFe₂O₄ and ZnFe₂O₄/(P,S)-g-C₃N₄ (70:30) at 300 K (c), heterocomposites ZF/CN/2D TiO₂ (5, 10, 20 wt %) (d), zero field cooled and field cooled (ZFC/FC) magnetization curves of ZnFe₂O₄ (e), high field magnetization, M, (applied magnetic field of 6 T) versus temperature for ZnFe₂O₄ (f).

As presented in Table S1, the specific surface areas of bare zinc ferrite, (P,S)-g-C₃N₄, and TiO₂ nanosheets were 26.0, 17.4, and 10.9 m²·g^–1, respectively.

The optical absorption properties of the prepared photocatalysts were analyzed by diffuse reflectance UV/vis spectroscopy (DRS), as shown in Figure b. The light absorption edge of 2D TiO₂ was about 400 nm. The adsorption edge of (P,S)-g-C₃N₄ was observed at about 510 nm, corresponding to the bandgap of 2.42 eV. For zinc ferrite, the absorption edge was around 700 nm due to charge transfer between O 2p and Fe 3d. The ternary composites showed optical absorption in both the UV and visible ranges, which suggests their utility for solar-driven photocatalysis. The strong absorption in the UV range with a threshold of about 400 nm for all ZF/CN/2D TiO₂ composites is related to the electronic structure of TiO₂. The light absorption in the 400–600 nm range resulted from the presence of 5, 10, and 20 wt % of ZnFe₂O₄/(P,S)-g-C₃N₄ in the ternary ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ composite. The highest visible light absorption was noticed for the ternary composite containing the highest amount of ZnFe₂O₄/(P,S)-g-C₃N₄ (20 wt %) to 2D TiO₂. The spectra were transformed into the Kubelka–Munk function, and the Tauc transformation was used to determine the optical bandgap energies. As expected, the zinc ferrite sample exhibited a low optical bandgap energy of 1.74 eV, and the optical band gap of TiO₂ was about 3.15 eV (see Figure S2).

The magnetic properties of ZnFe₂O₄, (P,S)-g-C₃N₄, ZnFe₂O₄/(P,S)-g-C₃N₄, 2D TiO₂, and heterocomposites of ZnFe₂O₄/(P,S)-g-C₃N₄/2DTiO₂ (5%), ZnFe₂O₄/(P,S)-g-C₃N₄/2DTiO₂ (10%), and ZnFe₂O₄/(P,S)-g-C₃N₄/2DTiO₂ (20%) were analyzed, and room-temperature hysteresis loops are presented in Figure c–f. The high field magnetization (applied field of 6 T) of ZnFe₂O₄ was 26.8 emu·g^–1 (Table S1). The anhydrous hysteresis loop confirmed the superparamagnetic properties of ZnFe₂O₄ (Figure c). However, (P,S)-g-C₃N₄ and 2D TiO₂ showed diamagnetic properties (Figure S3). Therefore, the combination of ZnFe₂O₄ with (P,S)-g-C₃N₄ gives rise to a decreased magnetization of 19.16 emu·g^–1. A further reduction in magnetization is found in ZF/CN/2D TiO₂ as a consequence of the incorporation of diamagnetic TiO₂, with a slight increase in magnetization with the increase in the amount of zinc ferrite (ZnFe₂O₄) in the heterocomposite. In fact, bulk ZnFe₂O₄ is antiferromagnetic with Néel temperature, T _N = 10 K, displaying reduced magnetic susceptibility at 300 K. ZnFe₂O₄ is a normal spinel, where Fe³⁺ cations antiferromagnetically coupled occupy octahedral positions, and Zn²⁺ cations (nonmagnetic) are preferentially located at the tetrahedral positions. The occurrence of mixed states (i.e., Fe³⁺ cations in both B and A sites) in Zn ferrite nanoparticles leads to ferrimagnetic behavior and characteristic superparamagnetism in the nanoscale regime. The temperature dependence of the ZFC/FC magnetization measured from 5 to 300 K in an applied magnetic field of 100 Oe is presented in Figure e.

The zero field cooled (ZFC) magnetization curve is determined by increasing the temperature of the sample from 5 to 300 K in the presence of a magnetic field, while the field cooled (FC) magnetization curve was analyzed by measuring magnetization and decreasing the temperature in the presence of the magnetic field. From the ZFC/FC curve, the superparamagnetic nature of the ferrite nanoparticles can be inferred. Furthermore, two important parameters, blocking temperature (T _B) and irreversibility temperature (T _irr), can be determined. The blocking temperature (T _B) is the maximum temperature at which a maximum in the ZFC magnetization is obtained, and the irreversibility temperature (T _irr) is determined at the split point of the ZFC and FC curves. From measurement, T _B was estimated to be 136 K and T _irr was estimated to be 170 K. The difference between T _B and T _irr can be explained by the presence of large particles and their impact to magnetization since they unblock at increased temperatures. The occurrence of a certain magnetic frustration at low temperatures can be deduced from the maximum in the FC curve around 50 K. Such a maximum value can also be detected in the temperature dependence of magnetization at a high applied magnetic field (6 T, see Figure f).

The EIS Nyquist plots of (P,S)-g-C₃N₄, 2D TiO₂, ZnFe₂O₄, and ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (10%) are presented in Figure S4. The diameter of the semicircle in the EIS Nyquist plot correlates with the charge transfer resistance and separation efficiency of the photogenerated charge carriers at the photocatalyst/electrolyte. The smallest impedance arc radius was observed for ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (10%) and 2D TiO₂ samples and the highest for pure ZnFe₂O₄. The results were further confirmed by the photocurrent response of the photocatalytic materials (Figure S4b). The low photocurrent signal of ZnFe₂O₄ proved the inefficient transfer and low charge carriers’ separation.

Thermal gravimetry analysis was conducted to evaluate the thermal stability of the photocatalyst and indirectly confirm the presence of (P,S)-g-C₃N₄ in the composite structure. As presented in Figure S5, both composites of ZF/CN and ZF/CN/2D TiO₂ (10%) were stable until 550 °C, and above that temperature, a decrease in the slope was observed, which is attributed to the decomposition of (P,S)-g-C₃N₄. The temperature increase from 500 to 600 °C resulted in almost 27.4% of weight loss, which corresponds to the decomposition amount of (P,S)-g-C₃N₄. According to the literature, the decomposition of the carbon nitride phase occurs in the temperature range of 500 to 600 °C; afterward, the residual amount is attributed to the remaining composition of the material. , Above 600°, a plateau is obtained, indicating the presence of inorganic ZnFe₂O₄ material in a residual amount of about 64.6%). In Figure S5­(b), the TGA analysis of ZF/CN/2D TiO₂ (10%) revealed that about 0.5% weight loss was observed up to 200 °C, typical for moisture loss. From the temperature range of 500 to 600 °C, a weight loss of 2.5% was observed, which can be attributed to decomposition of (P,S)-g-C₃N₄.

3.2. Photocatalytic Degradation of Carbamazepine (CBZ)

The photocatalytic activity of the prepared photocatalytic materials was studied in the reaction of carbamazepine degradation as a model compound under UV–vis light irradiation. Since the photocatalytic degradation of CBZ occurs on the surface of the photocatalyst, it is important to determine the adsorption properties of the material. Adsorption values analyzed after 30 min of adsorption–desorption equilibrium were less than 4%. The photodegradation activity of ZnFe₂O₄ and (P,S)-g-C₃N₄ was about 14% and 93% in 120 min of irradiation. A combination of ZnFe₂O₄ with (P,S)-g-C₃N₄ (70% w/w) increased the degradation rate of CBZ by 2-fold compared to the activity of pure ZnFe₂O₄ (Figure a and Table ).

4.

Photocatalytic degradation of carbamazepine (CBZ; 14 mg·dm^–3), 25 cm^–3, 20 °C. (%) Degradation efficiency of CBZ in the presence of (P,S)-g-C₃N₄, ZnFe₂O₄, ZF/CN (50:50), and ZF/CN (70:30) (a), degradation efficiency of CBZ in the presence of 2D TiO₂, ZF/CN (70:30) and heterocomposites of ZF/CN/2D TiO₂ (10%) (b), and degradation efficiency of CBZ in the presence of persulfate (PS; 0.5, 1.5, 3.0 mM) and combined system of PS + ZF/CN/2D TiO₂ (10%) (c).

1. Photodegradation of Carbamazepine (CBZ; 14 mg·dm^–3).

type of photocatalyst	process	degradation rate constant k (min–1)	TOC removal after 120 min (%)
2D TiO2	photocatalysis	0.121	64
(P,S)-g-C3N4	photocatalysis	0.014	18
ZnFe2O4	photocatalysis	0.001	1
ZF/CN (70:30)	photocatalysis	0.004	8
ZF/CN (50:50)	photocatalysis	0.003	3
ZF/CN/2D TiO2 (5%)	photocatalysis	0.071	39
ZF/CN/2D TiO2 (10%)	photocatalysis	0.176	68
ZF/CN/2D TiO2 (20%)	photocatalysis	0.042	38
peroxydisulfate (PS; 0.5 mM)	AOP	0.014	27
peroxydisulfate (PS; 1.5 mM)	AOP	0.036	39
peroxydisulfate (PS; 3.0 mM)	AOP	0.058	41
(PS; 3.0 mM) + ZF/CN/2D TiO2 (10%)	AOP + photocatalysis	0.370	74

Further modification of ZnFe₂O₄/(P,S)-g-C₃N₄ (70% w/w) with 2D TiO₂ markedly improved the photocatalytic activity (Figure b, Figure S6). The highest photocatalytic activity was found for ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (10%). After only 30 min of the photodegradation process, a significant CBZ degradation of 97% was observed for ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (10%); meanwhile, for 2D TiO₂, it was 87%. The degradation kinetics was studied, and pseudo-first-order was found to fit the kinetic model of CBZ degradation (Figure S6).

A combination of ZnFe₂O₄ and (P,S)-g-C₃N₄ increased the rate constant from 0.001 to 0.004 min^–1, while further modification with 2D TiO₂ increased the degradation rate constant to 0.176 min^–1 for the ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (10%) composite (Table ). For 2D TiO₂, the apparent rate constant was 0.121 min^–1. Also, the apparent rate constants for ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (5%) and ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (20%) composites were lower after the first 30 min of photodegradation than that for ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (10%) (Table ). In the case of ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (5%), containing 95% 2D TiO₂ and 5% ZnFe₂O₄/(P,S)-g-C₃N₄, the photoactivity was quite similar to that of TiO₂ due to an insufficient amount of g-C₃N₄ in the composite structure to improve charge carriers recombination. On the other hand, for the ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (20%) composite containing a higher amount of zinc spinel ferrite, the apparent rate constant was 0.042 min^–1. A higher content of the magnetic phase in the composite structure can lead to a decrease in activity, attributed to the ineffective dispersion of zinc spinel ferrite particles, which aggregate, hindering effective distribution in the reaction medium. In the analysis of total organic carbon (TOC), the removal efficiency was significantly enhanced for the ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (10%) magnetic composite, achieving a 74% effectiveness.

Furthermore, to evaluate the stability and recyclability of ZF/CN/2D TiO₂ (10%), carbamazepine photodegradation was performed in the subsequent cycles of photodegradation, as presented in Figure S7 in the Supporting Information. The results showed that the prepared photocatalyst exhibited good stability and recyclability. The degradation efficiency of carbamazepine after four consecutive cycles was 100%. However, a slight decrease to 96% was observed in the fifth cycle, which was attributed to the loss of a small amount of photocatalyst during the recovery stage. Photometric analyses of the iron concentration in the aqueous media in the presence of ZF, ZF/CN/ZF/CN/2D TiO₂ (10%) confirmed the stability of the composite material. As presented in Figure S8, the small amounts of iron ions released into water were noticed during CBZ degradation in the presence of ZnFe₂O₄, with a concentration of 0.064 mg·dm^–3 Fe after treatment. The stability of ZnFe₂O₄ improved when combined with (P,S)-g-C₃N₄ as a composite (ZF/CN 70:30), resulting in a reduced iron presence in the aqueous phase with a concentration of 0.051 mg·dm^–3. The lowest concentration of Fe in water was noticed for CBZ degradation in the presence of the ZF/CN/2D TiO₂ (10%) composite (0.011 mg·dm^–3), which is below the detection limit.

3.3. Photodegradation of Carbamazepine in the PS-Assisted Photocatalysis Process

In the next step of our study, to increase the efficiency of the photodegradation process, persulfate activation in the advanced oxidation process was investigated. Persulfates (S₂O₈ ^2–) are extensively employed in environmental remediation, namely, in aqueous media and soil, owing to their notable oxidation potential, affordable cost of 0.18 USD/mol, and extended environmental retention half-life time (t _1/2) ranging from 2 to 600 days in groundwaters. In the present study, sodium persulfate was used as a precursor of reactive oxygen species (SO₄ ^•–, HO^•) due to its lower cost, good stability, and high solubility (556 g·dm^–3 at 20 °C) as compared to potassium persulfate as sulfate radicals precursors (K₂S₂O₈, 53 g·cm^–3).

Three different concentrations of PS (0.5, 1.5, 3.0 mM) were used to degrade carbamazepine. Considering only the application of PS for degradation, the highest carbamazepine degradation was observed for 3.0 mM PS in a 120 min treatment time. A significant increase in the CBZ removal was noticed in the first 30 min of the process in the presence of PS, where CBZ removal efficiency increased from 32% to 85%, which is attributed to the activation of the PS by light (Figure c,d). The rate constant of CBZ degradation increased with the increase of PS concentration from 0.014, 0.036, and 0.058 min^–1 for PS (0.5, 1.5, and 3.0 mM), respectively (Table ). In the degradation process using 3.0 mM PS, CBZ is fast converted into oxidation byproducts. Therefore, the mineralization efficiency measured as TOC reduction was less than 41%.

The PS-assisted photocatalysis process yielded more efficient CBZ degradation and mineralization. Within the first 15 min of irradiation, above 99.5% CBZ degradation was achieved through the combined photodegradation process. The addition of PS in the system enhanced the reaction rate, and the degradation rate constant equaled 0.370 min^–1, suggesting the presence of more reactive species, which contributed to the synergy of the hybrid process in the presence of the ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (10%) composite. The mineralization efficiency increased from 68% for the ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (10%) photocatalysis process to 74% for the combined PS-assisted photocatalysis process (Table ).

The addition of PS to the photocatalytic system reduced the pH of water to 6.2. In this regard, the stability of the composite material during the photocatalytic process in combination with PS activation was evaluated by analyzing the leaching of Fe ions, as presented in Figure S8 in the Supporting Information. The performed analysis indicated that the Fe concentration in the solution was below the detection limit, as iron was present in the water at a concentration of approximately 0.03 mg/dm³. In seven consecutive cycles of carbamazepine degradation, the obtained photocatalyst showed good recyclability (Figure S9 in the Supporting Information). The degradation efficiency of CBZ remained stable during the first six cycles. However, with the seventh cycle, it decreased to 92%. The decrease in degradation efficiency can be attributed to mass loss of the sample after treatment. The recovery of the photocatalyst after the last cycle of treatment was 24.11 mg from an initial mass of 25 mg (96% efficiency). To verify if the photocatalyst amount can influence the degradation efficiency, an additional mass of 0.89 mg was added to the recovered photocatalyst portion, and the experiment was conducted under the same conditions. As a result, almost 100% degradation was achieved after 30 min. Furthermore, an additional test was conducted using 24.11 mg of a nonrecycled photocatalyst under the same experimental conditions. The results showed that after 30 min of treatment, 97.5% degradation of carbamazepine was observed. Moreover, the XRD analysis of the sample before and after treatment confirmed that there is no structural change in the photocatalyst, proving stability of the material toward photocorrosion or etching of the surface in the presence of PS and formation of metal salts on the surface of the photocatalyst in the case of the metal leaching process (Figure S9).

3.4. Mechanism and Degradation Pathway of Carbamazepine

Radical quenching experiments were performed to study the reactive species involved in the CBZ degradation. In the photocatalytic system using ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (10%), the predominant oxidative species involved in the photodegradation were superoxide anion radicals (O₂ ^•–). The photodegradation efficiency of CBZ decreased to 65%, 94%, and 96% in the presence of P-BQ (O₂ ^•–), isopropanol (SO₄ ^•–, HO^•) and TBA (HO^•), whereas in the absence of scavengers, complete degradation was observed, as presented in Figure .

5.

Effect of scavengers on the degradation efficiency of CBZ in the photocatalysis process (a) and in the PS-assisted photocatalysis (b).

Furthermore, the scavenging agents were added to the hybrid system of the PS/ZF/CN/2D TiO₂ heterocomposite. Based on the obtained results, the addition of scavengers decreased the photodegradation efficiency of CBZ and prolonged the reaction time from 30 to 120 min. This effect was significantly observed when isopropanol and tert-butyl alcohol (TBA) were added to the reaction system. In the presence of isopropyl alcohol (HO^•, SO₄ ^•–), the degradation efficiency decreased from almost 100% (30 min) to 70.5% (120 min), while in the presence of TBA (HO^•), the degradation efficiency declined to 82% (120 min), see Figure b.

Additionally, the generation of hydroxyl radicals (HO^•) was investigated in the reaction of coumarin conversion to 7-hydroxycoumarin in the presence of ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (10% w/w) and PS + ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (10% w/w) under the same experimental conditions. Generation of hydroxycoumarin (7-OH-coumarin) from coumarin during photocatalytic reaction was confirmed by increased signals of the emission spectra, as presented in Figure a. From quantitative measurements, it was observed that the concentration of 7-hydroxycoumarin (7-OH-coumarin) in a 5 min reaction was determined to be about 95 nM and it increased to 1355 nM in 60 min irradiation time (Figure b).

6.

Generation of hydroxycoumarin (7-OH-coumarin) from coumarin in photocatalysis and PS-assisted photodegradation process (a) and quantitation of hydroxycoumarin concentration as a probe to detect generated ·OH in the photocatalytic and PS-assisted photocatalysis process (b).

The effect of PS addition in the photocatalytic system correlates with the observed generation of hydroxyl radicals. A high concentration of hydroxycoumarin (1807 nM) was noticed in the first 5 min of the reaction compared to the photocatalytic reaction, confirming the fast conversion of sulfate radicals into hydroxyl radicals in the system (Figure b). The highest hydroxyl radical concentration was noticed in the first 15–20 min of the process, corresponding to the fast degradation of CBZ of 99.5%. With the increase of irradiation time to 60 min, the concentration of hydroxycoumarin decreases, which foretells a consumption of sulfate radicals in the photocatalytic system.

Based on the detected intermediates (P1–P6) and other reports, − we proposed the degradation pathway of carbamazepine in three systems: (1) photocatalysis in the presence of ZF/CN/2D TiO₂ (10%), (2) (PS)-AOP process, and (3) (PS)-assisted photocatalysis process in the presence of the ZF/CN/2D TiO₂ (10%) composite. The O₂ ^•–, SO₄ ^•–, and ^•OH radicals attack the CBZ molecule through deamidation, hydroxylation, and e^– transfer and effectively participate in the degradation of carbamazepine.

The photodegradation pathway in the presence of the ZF/CN/2D TiO₂ (10%) composite was assisted by superoxide (O₂ ^•–) and hydroxyl radicals (HO^•), while in the combined system of (ZF/CN/2D TiO₂+ PS), the photodegradation was assisted by coexistence of hydroxyl radicals (HO^•), superoxide anion radicals (O₂ ^•–), and sulfate radicals (SO₄ ^•–). However, the identified products of the photocatalysis process and the PS-assisted photodegradation process were identical. The carbamazepine degradation pathways first involve one electron transfer oxidation leading to a C-centered radical cation, which is stabilized by resonance. The transient carbon–carbon free radical cation subsequently attacked by the O₂ ^•– and SO₄ ^•– radicals may form intermediates P1, P2, and P3 with m/z of 252.09, which remain in dynamic equilibrium, as presented in Figure . Hydroxylation of P1 (10,11-epoxy CBZ) and P2 (10-OH CBZ) led to the formation of 10,11-dihydro-10,11-dihydroxycarbamazepine (m/z 270.4), and its subsequent hydroxylation gives P4 (hydroxy-keto-intermediate, m/z 223.2). Simultaneously, hydroxylation of P2 leads to the formation of two carbonyl groups in product with m/z = 268.09, which is further transformed into P5 by an intramolecular reaction. Another pathway involves hydroxylation and 7 membered-ring contraction with subsequent hydroxylation and the loss of CONH₂ to give a product with m/z 209.2, which is transformed in the presence of ^•OH radicals to acridine (P6) with m/z 179.1 and its further oxidation leads to acridone (m/z 195.6). Detected intermediates P4, P5, and P6 are degraded into smaller molecules and mineralized into CO₂, H₂O, NH₄ ⁺, and other lower molecular weight compounds, which is consistent with high TOC removal efficiency.

7.

Proposed mechanism of carbamazepine degradation in the presence of the ZF/CN/2D TiO₂ (10%) composite and in the PS-assisted photodegradation process.

3.5. Effect of Photocatalyst Amount, Initial pH, and Inorganic Ions in the PS-Assisted Photocatalysis Process

The effect of the photocatalyst amounts of 0.5, 1, and 2 g·dm^–3 was studied in the system of PS-assisted photocatalysis in the primary concentration of CBZ of 14 mg·dm^–3, as shown in Figure a. No significant differences were observed in the degradation efficiency of carbamazepine. However, a slightly higher degradation efficiency and kinetic rate constant was noted with a photocatalyst dosage of 1 g·dm^–3, which is consistent with our previous study. The higher photocatalyst amount does not contribute to greater photon utilization since most of the incident light is already absorbed near the surface. According to the effect of pH, the highest degradation efficiency was observed at both pH 4 and 7, with degradation efficiencies of 96% and 99.5% after 15 min of degradation, as presented in Table S2. At neutral pH, favorable electrostatic interactions between CBZ molecules and the photocatalyst surface, combined with an increased number of reactive oxygen species, promote efficient degradation. In acidic conditions, sulfate radicals (SO₄ ^•–) may enhance the attack on CBZ molecules; however, the lower adsorption of protonated carbamazepine on the positively charged photocatalyst surface may also result in slightly lower degradation efficiency. In alkaline conditions, although the formation of hydroxyl radicals (HO^•) is enhanced, partial charge repulsion and increased recombination can reduce the overall efficiency. The presence of sulfate (SO₄ ^2–), nitrate (NO₃ ^–), carbonate (CO₃ ^2–), chloride (Cl^–) anions, and humic acid (HA) in the concentration 0.1 mM in PS-assisted photocatalysis system was also investigated. The lowest degradation efficiency was observed in the presence of chloride anions (Cl^–) with 64% in 15 min of treatment time and a rate constant of 0.086 min^–1.

8.

Effect of the photocatalyst (a), pH (b), and inorganic anions (c) on the photodegradation of carbamazepine in the PS-assisted photocatalysis system.

The decrease in the degradation of CBZ is attributed to the reaction of sulfate radicals (SO₄ ^•‑) and hydroxyl radicals (HO^•) with chloride ions (Cl^–) to produce less reactive chloride radicals (Cl^•, 2.0 eV), see eqs –. Also, the addition of sulfate ions resulted in a decrease of activity (93%), as presented in Table S3. The presence of carbonate anions (CO₃ ^2–) decreased the degradation efficiency to 94% and the rate constant to 0.184 min. The reaction of carbonate anions with sulfate radicals can lead to the production of carbonate radicals (CO₃ ^•–) with a redox potential of 1.6 V, eq .

$${\mathrm{SO}}_4^{•-}+{\mathrm{Cl}}^{-}\to {\mathrm{SO}}_4^{2-}+{\mathrm{Cl}}^{•}$$ 4

$${\mathrm{Cl}}^{-}+{\mathrm{HO}}^{•}\to {\mathrm{ClOH}}^{•-}$$ 5

$${\mathrm{ClOH}}^{•-}+{\mathrm{H}}^{+}\to {\mathrm{Cl}}^{•}+{\mathrm{H}}_2\mathrm{O}$$ 6

$${\mathrm{SO}}_4^{•-}+{\mathrm{CO}}_3^{2-}\to {\mathrm{SO}}_4^{2-}+{\mathrm{CO}}_3^{•-}$$ 7

The presence of nitrate (NO₃ ^–) and humic acid (HA) exhibited only minor inhibitory effects on the degradation process, with removal efficiencies of 99% and 97%, respectively. The corresponding pseudo-first-order rate constants decreased to 0.308 and 0.238 min^–1 (Table S3).

Finally, the cost of treatment for photocatalysis and (PS)-assisted photocatalysis was evaluated assuming the treatment of 1 m³ effluent with the same operational parameters as the photocatalytic process for laboratory scale use. Both systems were compared concerning cost-effectiveness. The price of energy for nonhousehold consumers in Europe, according to Eurostat, was estimated to be around 0.200 Euro (€)/kWh. From the cost estimation presented in Table , it was found that the combination of photocatalysis with PS activation reduced the cost of treatment from 0.120 and 0.035 €/m³ for photocatalysis to 0.015 €/m³ for the combined process since the treatment time was decreased 2 times for the photocatalysis process and 4 times for the PS-assisted photocatalysis process compared to the PS activation process. Therefore, the PS-assisted photocatalysis process proved to be the most effective with regard to the cost of treatment.

2. Evaluation of the Cost of Treatment for Photocatalysis and PS-Assisted Photocatalysis for Industrial Implementation.

process	time (min)	V (m3)	energy consumption (kWh)	cost of energy [€/m 3 ]	cost of Energy [€/m 3 ]/25 cm 3
PS (3.0 mM)	120	1	0.6	0.120	3 × 10–6
photocatalysis	30	1	0.15	0.035	7.5 × 10–7
PS-assisted photocatalysis	15	1	0.075	0.015	3.75 × 10–7

4. Conclusions

The effect of different amounts of ZF/CN and 2D TiO₂, from 5% to 20% w/w, on magnetic separation, degradation efficiency, and mineralization was investigated. The combined system of photocatalysis and PS activation led to improved photocatalytic activity with a higher degradation rate constant of 0.370 min^–1 than for the photocatalysis process (0.176 min^–1) in the presence of a ZnFe₂O₄/(P,S)-g-C₃N₄/2D TiO₂ (10%) composite. The main reactive species involved in the photodegradation were superoxide anion radicals and hydroxyl radicals, whereas in the hybrid system of PS-assisted photocatalysis, the photodegradation mainly proceeded by e^– transfer and in the presence of hydroxyl radicals, sulfate radicals, and superoxide anion radicals.

The combination of photocatalysis with PS activation reduced the treatment cost by decreasing the treatment time by 2 times for the photocatalysis process and 4 times for the PS-assisted photocatalysis process compared to PS activation. Therefore, the PS-assisted photocatalysis process proved to be the most cost-effective, in terms of treatment time and degradation efficiency.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c08594.

• The manuscript is accompanied by a Supporting Information; The content of Supporting Information consists of: Characterization properties of photocatalyst (Table S1); TEM images of photocatalysts (Figure S1); Tauc’s plot (Figure S2), Hysteresis loops (Figure S3); Electrochemical measurements (Figure S4); TGA measurements (Figure S5); Degradation efficiency (C/Co) (Figure S6); Recyclability test during photocatalysis (Figure S7); Iron concentration in samples after treatment (Figure S8); Recyclability test during photocatalysis assisted PS (Figure S9); Generation of 7-hydroxycoumarin from coumarin during photocatalysis and photocatalysis assisted PS (Figure S10); Influence of pH, photocatalyst dosage (Table S2); Effect of inorganic anions in CBZ degradation during PS-assisted photocatalysis (Table S3) (PDF)

The authors declare no competing financial interest.