g-C3N4@TiO2@Fe3O4 Multifunctional Nanomaterial for Magnetic Solid-Phase Extraction and Photocatalytic Degradation-Based Removal of Trimethoprim and Isoniazid

In this period when environmental pollution has become uncontrollable, the removal of drug active substances reaching the environment and the analysis of drug active substances in different matrix environments are important for both living life and a sustainable environment. Therefore, the production of multifunctional materials that can be used in these two different processes has gained importance in the literature. Based on this thought, in this study, a g-C3N4@TiO2@Fe3O4 multifunctional nanohybrid material was synthesized and used for magnetic solid-phase extraction (MSPE) and photocatalytic degradation of trimethoprim and isoniazid, used together in tuberculosis treatment. All analyses were performed by high-performance liquid chromatography using a diode-array detection (HPLC-DAD) system. The synthesized material was characterized by X-ray diffraction spectroscopy (XRD), Raman spectroscopy, Fourier transform infrared (FTIR) spectroscopy, Brunauer–Emmett–Teller (BET) method, ζ-potential analysis, field-emission scanning electron microscopy (FE-SEM), and energy-dispersive X-ray spectroscopy (EDX). Important analytical parameters for the MSPE method such as the pH value of the sample solution, the volume of the sample solution, the amount of the sorbent, the type and volume of the elution solvent, and extraction time were optimized. The optimized MSPE method was then applied to different environmental waters and pharmaceutical samples. The recovery percentages for these samples were found to be between 95 and 107%. For trimethoprim and isoniazid, the limit of detections (LODs) were 0.055 and 0.145 and the limit of quantifications (LOQs) were 0.167 and 0.439 ng·mL–1, respectively. It was observed that ∼100% of trimethoprim and isoniazid active components were photocatalytically removed from the g-C3N4@TiO2@Fe3O4 nanohybrid material in ∼120 min under UV light.


INTRODUCTION
The combined use of drug active ingredients for the treatment of diseases is important in the fight against diseases. Tuberculosis (TB), caused by Mycobacterium tuberculosis bacteria, one of the most serious diseases of the past years, is known to kill more than 2 million people annually according to the data from the World Health Organization. 1 The fact that the treatment of TB is very difficult and very costly has brought along the use of multiple drugs. In the treatment of TB, the combination of trimethoprim (TRM) and isoniazid (INH) drug active ingredients is used. 2 Trimethoprim is used to treat various bacterial infections of the respiratory, urinary, and gastrointestinal tract, while isoniazid is used as a prodrug in the treatment phase. 3,4 While the analysis of these drug active ingredients in the most frequently used drug forms is important for quality control, the control of their release to the environment after use is important for the sustainability of vitality. Therefore, it is essential in this direction to develop accurate and sensitive analytical methods for the analysis of drug active ingredients used in combination in this way.
Analysis techniques for drug active ingredients include highperformance liquid chromatography (HPLC), 3 gas chromatography−mass spectroscopy (GC−MS), 4 liquid chromatography−mass spectroscopy (LC−MS), 5 and high-performance thin-layer chromatography (HPTLC). 6 However, direct analysis is not always possible due to complications such as the lowest analyte concentration that these devices can measure, higher than the amounts found in the samples studied, and/or the matrix effects of foreign species present in the sample environment. 7,8 Therefore, for the accurate and sensitive analysis of drug active ingredients in the instrumental detection system, a sample preparation method is required to separate drug active ingredients from the matrix environment and bring their concentration to a measurable level. 9 MSPE known as a simple and inexpensive sample preparation method has been successfully applied for the separation and enrichment of drug components. 10,11 The synthesis and use of new generation nanoadsorbents stand out as the most important developments that lead to different perspectives in solid-phase extraction techniques. Since the basic principle in the solidphase extraction method is based on adsorption, the large surface areas of the produced nanomaterials maximize the adsorption capacity for analytes. At the same time, for the photocatalytic removal of the analyzed pollutants, the acquisition of photocatalytic properties during the synthesis of nanoadsorbents is an important issue in the literature. 12 −14 The fact that graphite carbon nitride (g-C 3 N 4 ) has a large surface area and is a photocatalytic active material has led to its frequent use in the literature. 12−14 Titanium dioxide (TiO 2 ) is widely used in photocatalytic degradation studies due to its low bandwidth, very high surface area, and high photocatalytic efficiency. 10,15−17 Despite the above-mentioned advantages of g-C 3 N 4 and TiO 2 NPs, they need to be isolated quickly and easily without agglomeration from the solution environment in solidphase extraction and photocatalytic experiment steps. One of the effective solutions for this is the modification of nanoma-terials with magnetic NPs such as Fe 3 O 4 and α-Fe 2 O 3 . In this way, nanomaterials that gain magnetic properties can be easily isolated from the solution environment at any time by applying an external magnetic field. 17 In the light of this information, a multifunctional nanohybrid material containing g-C 3 N 4 NPs with high adsorption and photocatalytic properties, TiO 2 NPs with high photocatalytic properties, and Fe 3 O 4 NPs with high magnetism was produced. The g-C 3 N 4 @TiO 2 @Fe 3 O 4 multifunctional nanomaterial was synthesized using a simple and green solvothermal synthesis procedure. The synthesized material was successfully used for magnetic solid-phase extraction and photocatalytic degradation of trimethoprim and isoniazid drug active ingredients.

EXPERIMENTAL SECTION
2.1. Materials, Reagents, and Equipment. Analytical grade reagents were used throughout the experimental study. TiO 2 NPs had a particle size of 21 nm; FeCI 3 , CH 3 COONa, urea, and ethylene glycol were obtained from Sigma-Aldrich (St. Louis). All solutions used in HPLC-DAD analyses were of LC grade purity (Sigma-Aldrich, St. Louis). A Milli-Q system deionized water system (Millipore) was used to obtain deionized water (resistivity 18.2 MΩ·cm). 200 mg·L −1 stock solutions of isoniazid and trimethoprim were prepared in methanol. The analysis of trimethoprim and isoniazid were performed by the HPLC-DAD system including a C18 column (150 × 4.6 mm, 5 μm particle size, and 25°C of column temperature). The C18 column was obtained from USEM research and development company (Kayseri, Turkey). A methanol:acetate buffer mobile phase at a flow rate of 1.2 mL· min −1 was used. UV detection after chromatographic separation was performed at different absorption maxima depending on the used probe analyte. A photocatalytic reactor with a 400 W power of a UV lamp (Unitermrm, Turkey) was employed for the photocatalytic degradation experiments. The structure and morphology of the synthesized g-C 3 N 4 NPs and g-C 3 N 4 @ TiO 2 @Fe 3 O 4 NPs were investigated using a field-emission scanning electron microscope (FE-SEM, Gemini 550). The crystallographic structures of g-C 3 N 4 NPs and g-C 3 N 4 @TiO 2 @ Fe 3 O 4 NPs were illuminated by employing a Bruker AXS D8 Xray powder diffractometer with a simple cubic lattice and Cu Kα radiation (λ = 0.15406 nm), and the scan range (2θ) was from 5 to 90°. Raman spectra of TiO 2 NPs, g-C 3 N 4 NPs, and g-C 3 N 4 @ TiO 2 @Fe 3 O 4 NPs were recorded using a WITec alpha 300 M dispersive Raman spectrometer consisting of a He−Ne laser system (excitation wavelength is 532 nm). A Micromeritics Gemini VII BET analyzer was used to find the surface area, pore volume, and pore width for g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs. For this purpose, the BET-N 2 method was used. The ζ-potentials of g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs were measured at different pH values using a Malvern ZEN2600 Zetasizer Nano ZS.
2.2. Fabrication of the g-C 3 N 4 @TiO 2 @Fe 3 O 4 Nanohybrid Material. For the production of g-C 3 N 4 NPs, the calcination method as a simple and fast method proven in the literature was used. 15 First, 1.0 g of urea powder was placed in a porcelain crucible and calcined in a muffle furnace at 550°C at a heating rate of 2.3°C·min −1 in for 4 h, and then cooled at room temperature. At the end of this process, slightly yellowish g-C 3 N 4 NPs were obtained.
For the production of g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs, 0.5 g of g-C 3 N 4 NPs was weighed and ultrasonicated for 20 min in 20 mL of an ethylene glycol solution. Then, 0.5 g of TiO 2 NPs was weighed in a second beaker and then ultrasonicated in 20 mL of ethylene glycol for 20 min. Next, 0.5 g of FeCl 3 and 2 g of CH 3 COONa were weighed in a third beaker and ultrasonicated in 15 mL of ethylene glycol until the homogenization process was completed. These three mixtures obtained were collected in a beaker. After mixing in a magnetic stirrer at 900 rpm for 5 min, it was transferred into a Teflon solvothermal synthesis vessel and subjected to a solvothermal synthesis process in a solvothermal unit at 180°C for 16 h. After 16 h, g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs were collected by a neodymium magnet, and washing with ethanol and deionized water was performed. It was then left to dry at 80°C in an oven.
2.3. Magnetic Solid-Phase Extraction Method. A total of 10 mL of a model solution containing trimethoprim and isoniazid at known concentrations was prepared in a phosphate buffer medium (25 mM, pH 6.0) and 50 mg of g-C 3 N 4 @TiO 2 @ Fe 3 O 4 NPs was added to this solution. The mixture was mixed for 5 min with the help of a vortex to adsorb the drug molecules on the g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs. Finally, g-C 3 N 4 @TiO 2 @ Fe 3 O 4 NPs were isolated from the sample solution by applying an external magnetic field with a neodymium magnet. The resulting waste phase was completely discarded. Next, g-C 3 N 4 @ TiO 2 @Fe 3 O 4 NPs were dispersed in 0.25 mL of ethanol and vortexed for 5 min to desorb drug molecules from g-C 3 N 4 @ TiO 2 @Fe 3 O 4 NPs. The same elution process was performed three times in total, and the eluents were collected in the same tube and the final eluent volume was 0.75 mL. In the last step, the eluent phase was taken using a micropipette, filtered through a 0.22 filter, and analyzed using an HPLC-DAD detection system. For the next MSPE treatment, g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs were washed with 25 mL of water two times. The same MSPE/HPLC-DAD procedure was applied to the standard solutions and blank samples. The MSPE/HPLC-DAD procedure is schematized in Figure 1A.
2.4. Real Sample Applications. Real sample experiments were carried out on two different drug samples commercially available in a tablet form, lake water, seawater, and drinking water samples. Overall, 10 tablets were taken from the drugs in a tablet and homogenized in a mortar. Further, 25 mg of each sample was weighed into a flask and dissolved in 250 mL of methanol. A certain volume of these samples was taken and pure water and buffer solution were added to it, and the developed MSPE/HPLC-DAD procedure was applied. Environmental water samples were filtered through a 0.22 μm filter before use and stored at 4°C.
2.5. Photocatalytic Degradation of Drug Molecules. The photocatalysis performance of g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs was evaluated in 100 mL of a model solution medium containing known concentrations of trimethoprim or isoniazid molecules. For this purpose, 100 mg of g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs was added to 100 mL of a model solution, and the mixture was stirred in a dark environment without light until the adsorption of the drug molecules was complete. Next, the mixture was transferred to a photocatalytic reactor and exposed to a 400 W UV halogen lamp (365 nm wavelength) located in the center of the reactor. To determine the rate of photocatalytic degradation versus time, aliquots of a 1.0 mL sample were taken for analysis at programmed time intervals (10 min/sample) and analyzed using HPLC-DAD. The degradation rate of each drug active species was calculated using the peak areas obtained from the spectra recorded by HPLC-DAD at increasing times. The photocatalytic degradation process of drug molecules is schematized in Figure 1B.

RESULTS AND DISCUSSION
3.1. Characterization of the g-C 3 N 4 @TiO 2 @Fe 3 O 4 Multifunctional Nanomaterial. FE-SEM images of the produced g-C 3 N 4 NPs and g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs at 25 kV at 50 000 times magnification and EDX images at 50 000− 10 000 times magnification are shown in Figure 2. As shown in Figure 2, after calcination at 550°C, the complex leaf structure of g-C 3 N 4 NPs is seen (Figure 2A). In the images in Figure 2B,C, the final forms of the modified g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs are seen. Fe 3 O 4 nanospheres with sizes varying in the range of 25− 120 nm were placed between the foliated structures of g-C 3 N 4 , and these nanospheres were surrounded by both g-C 3 N 4 NPs and TiO 2 NPs during synthesis. As can be understood from the EDX spectrum shown in Figure 2D, all of the species (C, N, O, Fe, Ti) doped in the structure can be seen. The unlabeled peak in the EDX spectrum is due to the gold plating originating from the measurement procedure in the FE-SEM. The FE-SEM and EDX results proved that the g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs were successfully produced.
In the FTIR spectra shown in Figure  In Figure 4, XRD patterns of g-C 3 N 4 NPs, Fe 3 O 4 NPs, TiO 2 NPs, and g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs are shown. An intense peak at 2θ = 12.65 and 27.02°(JCPDS 087-1526) corresponds to graphitic planes of C 3 N 4 NPs. The XRD peak observed at 12.65°, corresponding to the (100) plane, is due to the in-plane structural packing motif of the aromatic segments. The strongest peak at 27.02°, corresponding to the (002) plane, is due to the stacking of the conjugated aromatic system, which characterizes the interlayer d-spacing of g-C 3 N 4 . Characteristic peaks for anatase TiO 2 NPs were confirmed by other studies in the literature. 16 For the XRD spectrum of When the spectrum of g-C 3 N 4 @TiO 2 @ Fe 3 O 4 NPs is examined, it is seen that the characteristic peaks of Fe 3 O 4 NPs, TiO 2 NPs, and g-C 3 N 4 NPs are obtained by decreasing their peak intensities. This is due to the composite structure of g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs NPs.
The Raman spectra of pure Fe 3 O 4 NPs can be characterized by three peaks at 286, 482, and 664 cm −1 . Even if the laser used in the Raman spectrometer is used at very low energy, many false peaks of hematite can be seen in the Raman spectrum of magnetite. Fe 3 O 4 has a spinel structure, giving rise to five Raman bands: three T 2g , one E, and one A 1g ( Figure 5). The data obtained from the Raman spectra of the materials were confirmed by other studies in the literature. 10,16 The BET surface area, pore volume, and pore width values for g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs were found to be 45.5 m 2 ·g −1 , 0.29 cm 3 ·g −1 , and 254.9 Å, respectively. Materials with pore widths between 2 and 50 nm are classified as mesoporous materials. The BET isotherm results showed that mesopore characteristics of g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs are dominant.  Trimethoprim and isoniazid are polar compounds containing acidic and carboxylic groups, so they can be found in a neutral or ionic form. If the pH of the solution is below the pK a value, these compounds are converted to a molecular form. The fact that most drug active species have more than one pK a value is another important factor to be aware of. It is assumed that the adsorption efficiency increases through effective intermolecular interactions between the analyte and the adsorbent. The effect of pH of the sample solution on the adsorption efficiency of the analytes was investigated in the range of 2.0−10.0 on model solutions containing 0.25 μg·mL −1 of trimethoprim and isoniazid. (Figure 6A). The results obtained showed that the extraction efficiency of both active ingredients increases in parallel with the increasing pH value, reaching a maximum at pH 6.0, and the extraction efficiency decreases above pH values. Considering the results, the optimum pH value was selected as 6.0. ζ-Potential analysis was performed to determine the surface charge of the synthesized g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs, and the results are given in Figure S1. It was observed that the surface of the g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs was positively charged when the pH of the aqueous solution was between 2 and 4, and it was negatively charged at higher pH values. It was observed that the surface of the g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs became more negatively charged as the pH increased. The pK a values of trimethoprim and isoniazid are 7.12 and 1.82, respectively. While the driving force for adsorption of trimethoprim was electrostatic interactions, the driving force for adsorption of isoniazid was predicted to be interactions such as van der Waals forces and hydrogen bonding.
3.2.2. Effect of the g-C 3 N 4 @TiO 2 @Fe 3 O 4 NP Amount on MSPE Efficiency. The effect of the g-C 3 N 4 @TiO 2 @Fe 3 O 4 NP amount on the adsorption efficiency of trimethoprim and isoniazid was investigated between 1 and 100 mg ( Figure 6B). The obtained results showed that the increase in the amount of the adsorbent per unit analyte concentration caused an increase in the extraction efficiency. It was observed that the maximum extraction efficiency was reached with the addition of 50 mg of g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs, and there was no significant change in extraction efficiency with more additions. Therefore, 50 mg of g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs was used in subsequent experiments for both drug molecules.
3.2.3. Effect of the Eluent Type and Volume on MSPE Efficiency. In solid-phase extraction methods, the elution of analytes on the adsorbent before analysis is an important step. Since trimethoprim and isoniazid are analytes in an organic form, it is appropriate to use organic solvents in their elutions. For this purpose, methanol (MeOH), ethanol (EtOH), isopropyl alcohol (IPA), acetonitrile (ACN), and acetone were used in the desorption stage of trimethoprim and isoniazid. After the elution step, the analyte concentrations in the final phase were determined by HPLC. Analyses were carried out by preparing suitable standards for each eluent phase. The quantitative recovery values for trimethoprim and isoniazid were obtained using ethanol and methanol, but the highest recoveries were reached with ethanol ( Figure 6C). Therefore, ethanol was used as an eluent for subsequent experiments. It is important to perform the desorption process with the minimum possible volume of eluent to obtain a high enhancement factor, which is an important criterion for obtaining lower detection limits and consuming a low volume of an organic solvent. Desorptions of trimethoprim and isoniazid drug active ingredients adsorbed on g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs were performed with volumes of ethanol varying from 0.5 to 4 mL and sequential extraction ( Figure 6D). The results obtained showed that both active pharmaceutical ingredients were recovered quantitatively with sequential extraction using 0.25 mL of ethanol three times. The eluents were collected in the same tube and the final eluent volume was 0.75 mL.
3.2.4. Effect of the Sample Solution Volume on MSPE Efficiency. The volume of the sample solution to which the method developed in sample preparation techniques can be applied is one of the most important parameters affecting the important analytical performance parameter such as the preconcentration factor, the limit of detection, and the limit of quantification. The developed MSPE method based on g-C 3 N 4 @TiO 2 @Fe 3 O 4 was applied to sample solutions changing from 15 to 100 mL under the optimum experimental conditions. The results obtained showed that the developed MSPE method can be applied up to 75 mL of the sample volume and that the extraction yields for both drug active ingredients at higher volumes were not quantitative ( Figure S2). Therefore, the developed method was applied to real samples with a maximum volume of 75 mL.
3.2.5. Analytical Performance of the MSPE Method. For validation of the MSPE method, the following validation parameters were evaluated: linearity, sensitivity, precision, and preconcentration factor (Table 1). Calibration curves for trimethoprim and isoniazid were obtained after the application of the MSPE method to increasing analyte concentrations. The calculation of the LOD was based on dividing 3 times the standard deviation of the signal of the lowest concentration of the analyte concentration obtained by the slope of the calibration curve, while the LOQ was based on 10 times the standard deviation. The precision of the MSPE method was given as intraday RSD % and interday RSD % calculated by applying the MSPE method to 10 standard solutions containing 0.25 μg·mL −1 trimethoprim and 0.5 μg mL −1 isoniazid. Average recovery values were calculated by applying the MSPE method to 10 standard solutions containing 0.25 μg·mL −1 trimethoprim and 0.5 μg·mL −1 isoniazid. Analytical performance parameters for the MSPE/HPLC-DAD procedure are shown in Table 1.
3.2.6. Adsorption Capacity of g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs. Adsorption capacity is the quantitative definition of the amount of analyte adsorbed per gram amount of the adsorbent by determining the amount of the adsorbent required to quantitatively enrich the analytes in a given solution. In this study, the adsorption capacity of g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs for trimethoprim and isoniazid was studied as another important analytical performance parameter. To find the amount of the analyte adsorbed per gram adsorbent amount, 50 mg of the adsorbent was weighed, and the MSPE method was applied at increasing concentrations of trimethoprim and isoniazid (50, 100, 250, 500, 1000, 2000, and 4000 ppm) in 10 mL of a model solution medium. After the application of the MSPE method, all samples were measured at an appropriate wavelength for each analyte in the HPLC-DAD system, and the amount of mg analyte per gram adsorbent was determined over the peak areas obtained. Adsorption capacity values for trimethoprim and isoniazid were determined as 39.9 and 47.8 mg·g −1 , respectively.
3.2.7. Accuracy of the MSPE Method. Accuracy studies for the optimized MSPE method based on g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs were performed on lake water, seawater, drinking water, and commercially available drug preparations. For this, known amounts of analytes were added to the real samples prepared, as explained in Section 2.4, and then the developed method was applied to these samples. Addition/recovery results are shown in Table 2. The recovery percentages made in the study were found to be between 95 and 107%. The obtained recovery values showed that the developed method can be successfully applied in these matrix media.
3.2.8. Photocatalytic Degradation of Drug Molecules on g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs. Photocatalytic degradation experiments were carried out in a sample solution medium containing drug active species. The volume of the irradiated solution was 150 mL, and the optical beam path was designed in the dark and at a distance of 10 cm from the samples in continuous magnetic stirring. A 400 W curing high-pressure UV halogen lamp was placed in the center of the reactor in a quartz sleeve. Samples in 150 mL of an aqueous phase (trimethoprim and isoniazid with 20 ppm concentration) were exposed to UV light in the presence of a photocatalyst. The time under the illumination of a UV lamp was 200 min for both species, after which 1 mL aliquots were taken for HPLC-DAD analysis at scheduled time intervals (10 min/sample). The concentration and degradation rate of the probe molecules were measured throughout the photocatalytic reaction using HPLC-DAD.
Photocatalytic efficiency was calculated using eq 1 C C decomposition rate (%) (C )/ 100 where C 0 and C t (mg·L −1 ) are the concentrations of probing molecules at the initial stage and the adsorption equilibrium concentration after irradiation at an irradiation time t (min), respectively. In chemical kinetics, when the reaction is first order, the reaction rate is measured based on the concentration of the reactants. So, the rate of the first-order reaction is directly proportional to the concentration of the probe molecules. The kinetics of the photocatalytic degradation rate of the probe molecules was evaluated according to the Langmuir−Hinshelwood kinetic model as given below. 8 The photocatalytic degradation efficiency of g-C 3 N 4 @TiO 2 @ Fe 3 O 4 NPs for the trimethoprim and isoniazid drug active species was analyzed according to the Langmuir−Hinshelwood kinetic model. The results obtained showed that ∼100% of trimethoprim and isoniazid were photocatalytically degraded in ∼120 and 100 min, respectively ( Figure 7A). The graph of the photocatalytic removal of drug molecules versus time and the kinetic study graph are shown in Figure 7B−D. The represented fitted straight lines were plotted with a slope equal to the apparent pseudo-first-order rate constant k (min −1 ).
The UV−vis spectra were recorded for g-C 3 N 4 NPs, TiO 2 NPs, and g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs (Figure 8). It was determined that TiO 2 NPs and g-C 3 N 4 NPs have maximum absorptions at 520 and 324.7 nm wavelengths, respectively. The absorption edges of TiO 2 NPs and g-C 3 N 4 NPs can be associated with the transfer of electrons from the valence band to the conduction band. The ternary nanocomposite g-C 3 N 4 @  TiO 2 @Fe 3 O 4 differed from other individual components in its absorption properties, giving a maximum absorption peak centered around 317.7 nm. This also can be attributed to the electronic transition from the valence band to the conduction band of the ternary composite and may also suggest significant interfacial contact between the g-C 3 N 4 NPs, TiO 2 NPs, and Fe 3 O 4 NP components in the final nanocomposite. Also, mixing the 4s orbital of Ti with the 4s orbital of Fe may result in the formation of the conduction band at low energy. Also, Tauc's plot used to calculate direct band gaps showed that g-C 3 N 4 NPs, TiO 2 NPs, and g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs have direct band gaps of 2.39, 1.43, and 3.08 eV, respectively. These data further confirmed the formation of the intended heterojunction, enabling photocatalysis in the UV−visible light range. The probable mechanism of the photocatalytic reaction of g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs is shown in Figure 9. TiO 2 NPs, from the wide-band-gap semiconductor class, have low UV light capture capacity to form light-induced electrons in the conduction band and accompanying holes in the valence band. g-C 3 N 4 NPs in g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs are the main component that successfully collect UV light and cause the excitation of electrons from the valence band to the conduction Figure 8. UV−vis spectra of g-C 3 N 4 NPs, TiO 2 NPs, and g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs and the corresponding Tauc plot of g-C 3 N 4 NPs, TiO 2 NPs, and g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs.
band. Since the conduction band of TiO 2 has a higher reduction potential, electrons migrate from the conduction band of g-C 3 N 4 to the conduction band of TiO 2 due to light. In addition, however, TiO 2 with the reduction potential of the conduction band higher than that of Fe 3 O 4 NPs paves the way for further transfer of photogenerated electrons to the conduction band of Fe 3 O 4 NPs. Electrons in the conduction band of TiO 2 tend to react easily with adsorbed oxygen (O 2 ) molecules to form superoxide anion radicals ( • O 2 − ). Also, a simultaneous hole transfer occurred from the valence band of TiO 2 to the valence band of g-C 3 N 4 . As a result of the reaction of the holes in the valence band of g-C 3 N 4 with water molecules (H 2 O), hydroxyl radicals ( • OH) are formed. These radicals then react with trimethoprim and isoniazid molecules, causing their degradation. Therefore, modifying a wide-band-gap semiconductor with a narrow-band-gap semiconductor is frequently used in photocatalyst applications. The resulting hybrid photocatalyst not only absorbs the photon in the UV−visible range but also has a suppressed recombination rate of the photogenerated charges. Starting from this point, in this study, while the g-C 3 N 4 region of the g-C 3 N 4 @TiO 2 @Fe 3 O 4 nanohybrid material is the active site that collects light, the TiO 2 and Fe 3 O 4 regions play the main role in the separation of charge carriers. Thus, it led to an increase in the photocatalytic efficiency of the designed photocatalyst.

CONCLUSIONS
The fact that trimethoprim and isoniazid are used together in the treatment of diseases such as tuberculosis makes it important to analyze both of them at the same time and to remove them from environmental waters after use. In this study, which was designed with this awareness, the usability of g-C 3 N 4 @TiO 2 @ Fe 3 O 4 NPs as a magnetic adsorbent in magnetic solid-phase extraction for accurate and sensitive analysis of trimethoprim and isoniazid in different matrix media with the HPLC-DAD technique and as a magnetic photocatalyst for the degradation of trimethoprim and isoniazid in water medium was investigated. Due to its graphene-like structure, g-C 3 N 4 NPs with high adsorption and photocatalytic properties were combined with TiO 2 NPs with high photocatalytic properties and Fe 3 O 4 nanoparticles with strong magnetic properties to produce a new hybrid material offering three different properties for these two different process. The obtained results proved that g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs can be used successfully in the magnetic solid-phase extraction and the photocatalytic removal of trimethoprim and isoniazid.  The authors declare no competing financial interest. This manuscript is a part of Gokhan Sarp's doctoral thesis.

■ ACKNOWLEDGMENTS
The authors are grateful for the financial support of the Unit of the Scientific Research Projects of Erciyes University (TDK-2022-11760). . Plausible photocatalytic degradation mechanism of organic pollutants in g-C 3 N 4 @TiO 2 @Fe 3 O 4 NPs.