Ternary Heterojunction Graphitic Carbon Nitride/Cupric Sulfide/Titanium Dioxide Photoelectrochemical Sensor for Sesamol Quantification and Antioxidant Synergism

Sesamol (SM) is a potent natural antioxidant that can quench free radicals and modulate the cholinergic system in the brain, thereby ameliorating memory and cognitive impairment in Alzheimer’s disease patients. Moreover, the total antioxidant capacity can be amplified by synergistic interactions between different antioxidants. Here, we constructed a ternary heterojunction graphitic carbon nitride/cupric sulfide/titanium dioxide (g-C3N4/CuS/TiO2) photoelectrochemical (PEC) sensor for the quantification of SM and its synergistic interactions with other antioxidants. Crucially, the Schottky barrier in ternary semiconductors considerably enhances electron transfer. The PEC sensor showed a wide linear range for SM detection, ranging from 2 to 1277 μmol L−1, and had a limit of detection of 1.8 μmol L−1. Remarkably, this sensing platform could evaluate the synergism between SM and five typical lipid-soluble antioxidants: tert-butyl hydroquinone, vitamin E, butyl hydroxyanisole, propyl gallate, and butylated hydroxytoluene. Owing to its low redox potential, SM could reduce antioxidant radicals and promote their regeneration, which increased the overall antioxidant performance. The g-C3N4/CuS/TiO2 PEC sensor exhibited high sensitivity, satisfactory selectivity, and stability, and was successfully applied for SM determination in both soybean and peanut oils. The findings of this study provide guidance for the development of nutritional foods, nutrition analysis, and the treatment of diseases caused by free radicals.


Introduction
Many people live unhealthy lifestyles, which often include smoking, drinking, getting insufficient sleep, and having an unbalanced diet, all of which result in the accumulation of a large number of oxidizing substances (reactive oxygen and nitrogen free radicals) in the body. These substances disrupt the redox balance of the body and damage biological macromolecules, organelles, and tissue [1]. This process is called oxidative stress [2]. Antioxidants are crucial for slowing down or preventing harmful oxidation and function by scavenging excess free radicals to combat oxidative stress and maintain redox homeostasis. Natural foods, such as vegetables, fruits, and tea, are rich in antioxidants, and their consumption can lower the risk of some chronic diseases [3]. Thus, there is interest in identifying natural sources of antioxidants and adding them to food formulations to improve human health and treat diseases [4]. Some natural antioxidants, such as vitamins E and C and natural polyphenols (catechins, resveratrol, and tea polyphenols), have received investigated and confirmed using the ternary heterojunction g-C 3 N 4 /CuS/TiO 2 sensor. The findings of this study will aid the development of sensors for food safety, quality control, and food component compatibility testing and will aid the discovery of treatments for diseases caused by free radicals.
Biosensors 2023, 13, x FOR PEER REVIEW 3 of 13 sensing material was then applied for SM detection (Scheme 1). Further, the synergistic effects of SM and other antioxidants, including tert-butyl hydroquinone (TBHQ), vitamin E (VE), butyl hydroxyanisole (BHA), propyl gallate (PG), and butylated hydroxytoluene (BHT), were investigated and confirmed using the ternary heterojunction g-C3N4/CuS/TiO2 sensor. The findings of this study will aid the development of sensors for food safety, quality control, and food component compatibility testing and will aid the discovery of treatments for diseases caused by free radicals.

Synthesis of g-C3N4/CuS/TiO2 Composites
TiO2 nanorods were synthesized using the classic hydrothermal method. Briefly, hydrochloric acid (15 mL), ultrapure water (15 mL), and titanium butoxide (500 μL) were added sequentially to a beaker and stirred for a few minutes. The mixture was transferred to a Teflon-lined autoclave containing a conductive fluorine-doped tin oxide (FTO) electrode. After react on at 150 °C for 5 h, the FTO electrode with TiO2 nanorods was rinsed with deionized water and annealed at 500 °C for 2 h at a heating rate of 2 °C min −1 . Then, copper sulfate solution (0.25 μL, 0.1 mol L −1 ) and sodium thiosulfate solution (4.75 mL, 0.1 mol L −1 ) were added to a Teflon-lined autoclave containing an electrode modified with TiO2 nanorods and reacted at 120 °C for 5 h to obtain a sample, denoted CuS/TiO2. Using the same conditions, different doping amounts of CuS can be obtained by adding different proportions of copper sulfate and sodium thiosulfate solutions.
G-C3N4 was prepared by heating melamine at 550 °C for 2 h at a rate of 2 °C min −1 . The products were washed three times with ultrapure water and ethanol sequentially, and dried at 60 °C. Then, g-C3N4 (10 mg) was added to ultrapure water (20 mL) and ultrasonicated for 3 h. The suspension was then centrifuged at a low speed for 10 min, and the supernatant was collected for further use.

Synthesis of g-C 3 N 4 /CuS/TiO 2 Composites
TiO 2 nanorods were synthesized using the classic hydrothermal method. Briefly, hydrochloric acid (15 mL), ultrapure water (15 mL), and titanium butoxide (500 µL) were added sequentially to a beaker and stirred for a few minutes. The mixture was transferred to a Teflon-lined autoclave containing a conductive fluorine-doped tin oxide (FTO) electrode. After react on at 150 • C for 5 h, the FTO electrode with TiO 2 nanorods was rinsed with deionized water and annealed at 500 • C for 2 h at a heating rate of 2 • C min −1 . Then, copper sulfate solution (0.25 µL, 0.1 mol L −1 ) and sodium thiosulfate solution (4.75 mL, 0.1 mol L −1 ) were added to a Teflon-lined autoclave containing an electrode modified with TiO 2 nanorods and reacted at 120 • C for 5 h to obtain a sample, denoted CuS/TiO 2 . Using the same conditions, different doping amounts of CuS can be obtained by adding different proportions of copper sulfate and sodium thiosulfate solutions. G-C 3 N 4 was prepared by heating melamine at 550 • C for 2 h at a rate of 2 • C min −1 . The products were washed three times with ultrapure water and ethanol sequentially, and dried at 60 • C. Then, g-C 3 N 4 (10 mg) was added to ultrapure water (20 mL) and ultrasonicated for 3 h. The suspension was then centrifuged at a low speed for 10 min, and the supernatant was collected for further use.
To obtain the g-C 3 N 4 /CuS/TiO 2 composite, the supernatant of g-C 3 N 4 was added to a watch glass containing the CuS/TiO 2 FTO electrode and soaked for 2 h at 70 • C. Then, the g-C 3 N 4 /CuS/TiO 2 electrode was dried with nitrogen and calcined at 550 • C for 2 h under nitrogen atmosphere at a heating rate of 2 • C min −1 . Under the same reaction conditions as those used to synthesize the composite materials, different loading amounts of g-C 3 N 4 were obtained by adding different masses of g-C 3 N 4 powder.

Antioxidant Capacity and Synergistic Effects
First, an antioxidant standard solution having a concentration of 10 mmol L −1 was prepared by dissolving SM, GA, CHA, EGC, EGCG, and PC in ethanol. For SM detection, the g-C 3 N 4 /CuS/TiO 2 electrode was fastened to the PEC cell and suspended in an electrolyte comprising a phosphate-buffered saline (PBS) solution (3.0 mL; 0.1 mol L −1 ) of SM, and the photocurrent was measured at least three times at room temperature. The photocurrent was obtained using the formula: ∆I = I sample -I blank , where I sample is the photocurrent in the presence of the target and I blank is the photocurrent in the absence of the target. To investigate the synergistic effect of the SM-based antioxidant systems, 10 mmol L −1 mixtures of SM+VE, SM+BHA, SM+BHT, SM+TBHQ, SM+PG, SM+BHA+BHT+TBHQ+PG, and SM+BHA+BHT+TBHQ+PG+VE were prepared. The molar quantities of all antioxidants in these mixtures were equal. Similarly, VE, BHA, BHT, TBHQ, SM+VE, SM+BHA, SM+BHT, SM+TBHQ, SM+PG, SM+BHA+BHT+TBHQ+PG, and SM+BHA+BHT+TBHQ +PG+VE were added to the detection cell, and the photocurrent was measured. The synergistic effects can be calculated and quantified by comparing the photocurrents of a mixture of antioxidants. Figure 1A shows a schematic of the g-C 3 N 4 /CuS/TiO 2 composite synthesized in three steps. Scanning electron microscopy (SEM) images revealed that TiO 2 grew uniformly on FTO and had a smooth surface ( Figure S1A in Supplementary Materials). However, the surface became rough after the deposition of the CuS nanoparticles ( Figure S1B). Figure 1B shows that the TiO 2 nanorod structure did not change after loading with CuS and g-C 3 N 4 . As shown in Figure 1C, the tops of the nanorods have a flower-like appearance, which should increase the contact area available for reaction and is conducive to the collection and utilization of visible light. The transmission electron microscopy (TEM) image shown in Figure 1D was used to calculate the cross-sectional width of the g-C 3 N 4 /CuS/TiO 2 composites: approximately 127 nm, whereas the length of the nanorods was approximately 500-900 nm. These dimensions should promote the longitudinal migration of photogenerated carriers. As shown in Figure 1E, the CuS nanoparticles were uniformly doped on the surfaces of the TiO 2 nanorods, and the figure shows that the particle size ofCuS was less than 5 nm. Lattice fringes having interplanar spaces of 0.268, 0.338, and 0.264 nm can be clearly seen in Figure 1F, and these can be ascribed to the (200) and (204) crystal planes of anatase TiO 2 and(110) crystal plane of CuS, respectively. Energy-dispersive X-ray spectroscopy ( Figure 1G) measurements revealed the homogeneous distributions of Ti, O, Cu, S, C, and N in the g-C 3 N 4 /CuS/TiO 2 composite, suggesting that the composite had been successfully synthesized.

Photoelectrochemical Properties of the g-C3N4/CuS/TiO2 PEC Sensing Platform
Next, electrochemical impedance spectroscopy (EIS) was performed to determine the conductivities of the prepared samples. As shown in Figure 2B, g-C3N4/CuS/TiO2 yielded the smallest Nyquist semicircle radius, indicating that the composite had the lowest electron transfer resistance and higher conductivity than the other samples. This result can be attributed to the formation of a large number of heterointerfaces between the ternary sem-

Photoelectrochemical Properties of the g-C 3 N 4 /CuS/TiO 2 PEC Sensing Platform
Next, electrochemical impedance spectroscopy (EIS) was performed to determine the conductivities of the prepared samples. As shown in Figure 2B, g-C 3 N 4 /CuS/TiO 2 yielded the smallest Nyquist semicircle radius, indicating that the composite had the lowest electron transfer resistance and higher conductivity than the other samples. This result can be attributed to the formation of a large number of heterointerfaces between the ternary semiconductor materials, which is conducive to charge transport.
The photocurrent response provides information about the optoelectronic properties and detection sensitivity of a material. Therefore, we investigated the photocurrents of the TiO 2 , g-C 3 N 4 , CuS, CuS/TiO 2 , and g-C 3 N 4 /CuS/TiO 2 -modified FTO electrodes ( Figure 2C). Owing to the slow separation of photoinduced electron-hole pairs, the single semiconductor materials (i.e., TiO 2 , g-C 3 N 4 , and CuS) showed relatively low photocurrents both in the presence and absence of antioxidants. In contrast, the CuS/TiO 2 electrode exhibited higher photoelectric activity: its photocurrent increased by 1.49 µA with respect to that of TiO 2 alone, without the addition of the antioxidant. This is because CuS nanoparticles have high absorbance in the visible-light region. However, the photocurrent only increased by 0.32 µA after the addition of 123.46 µmol L −1 SM. Therefore, the sensitivity of the CuS/TiO 2 electrode is low and is not sufficient to detect SM. Thus, further modification ofCuS/TiO 2 by loading with g-C 3 N 4 nanosheets is required. As expected, the photoelectric response and detection sensitivity of the ternary semiconductor composite (g-C 3 N 4 /CuS/TiO 2 ) were significantly improved. Notably, the photocurrent response of the ternary semiconductor composite increased by 0.62 µA in relation to that of the binary semiconductor composite, and the photocurrent was increased by 2.68 µA in the presence of 123.46 µmol L −1 SM, suggesting that the ternary heterojunction PEC sensing platform is suitable for evaluating the antioxidant properties of SM. The significant increase in photocurrent can be attributed to two factors: (1) loading with g-C 3 N 4 nanosheets increased the number of active sites for the reaction and enhanced light absorption, and (2) the heterojunctions in the ternary semiconductor composites promoted the reaction between antioxidants and holes and greatly improved the electron-hole separation.
To determine the optical properties, UV-Vis absorbance spectroscopy was carried out, as shown in Figure 2D. TiO 2 nanorods can only absorb light between 200 and 400 nm because of their wide band gap. However, the band gap was narrowed and more active sites for the reaction were produced after doping with CuS and g-C 3 N 4 , which increased the collection and utilization of light. Therefore, the ternary g-C 3 N 4 /CuS/TiO 2 semiconductor exhibited greater light absorption at 200-800 nm, which is consistent with the EIS results. The ability of a sensor to detect antioxidants is strongly dependent on the band structure of the composites. Therefore, to achieve selective detection, the careful design of the valence band (VB) position of the semiconductor is required. As shown in Figure 2E, the relationship between the Kubelka-Munk function and the photon energy of the samples can be obtained from the UV-Vis absorbance spectra. The VB values ofTiO 2 , CuS/TiO 2 , and g-C 3 N 4 /CuS/TiO 2 are 2.95, 2.30, and 2.50 eV, respectively. The Mott-Schottky (M-S) plots ofg-C 3 N 4 /CuS/TiO 2 semiconductor exhibited positive slopes ( Figure 2F), suggesting that the composites are n-type heterostructures, and the flat-band potential (V fb ) was −0.25 V (vs. Ag/AgCl). In addition, the conduction band (CB) of the n-type semiconductor is 0.1 V smaller than its V fb ; thus, the CB of g-C 3 N 4 /CuS/TiO 2 is −0.35 V (vs. Ag/AgCl).

Optimization of Experimental Conditions
The intrinsic properties of photosensitive materials are the main factors affecting the photocurrent signal, but other factors, such as the degree of doping of the semiconductor material, laser wavelength, and applied voltage, cannot be ignored. Optimization experiments were performed inPBS solution containing 123.46 µmol L −1 SM. As shown in Figure S3A, we investigated the influence of the amount of deposited CuS on the photocurrent. Crucially, an appropriate amount of CuS nanoparticles will increase the photocurrent, but an excessive amount could result in a drastic decrease in photocurrent. Based on the experiments, a doping level of 5% was selected to prepare the CuS/TiO 2 composite. The mass ofg-C 3 N 4 is another key factor that affects the properties of semiconductors. As shown in Figure S3B, g-C 3 N 4 nanosheets provide more active sites for the reaction, and those prepared using a mass of 10 mg generated the highest photocurrent. As shown by the UV-Vis spectra, the g-C 3 N 4 /CuS/TiO 2 exhibited different absorbances in different bands. In particular, as an excitation light source, 630 nm red light has strong penetration. In addition, red light accounts for a large proportion of sunlight, suggesting its suitability for enhancing the sensitivity of the sensor for SM detection ( Figure S3C). As shown in Figure S3D, there is a significant gradient between the semiconductor and FTO electrode in the initial state. When a potential is applied, electrons flow rapidly from the conductor band of the semiconductor to the FTO electrode to generate a photocurrent signal. The current at 0 V was 77.3% of that at +0.08 V, indicating that the PEC sensor has adequate sensitivity for SM detection. Therefore, 0 V was chosen as the working voltage for further experiments.

Antioxidant Assay and Detection Mechanism of the g-C 3 N 4 /CuS/TiO 2 PEC Sensing Platform
To analyze the detection performance of the as-prepared g-C 3 N 4 /CuS/TiO 2 PEC sensor upon 630 nm laser irradiation, the photocurrent signal was measured in PBS solutions (0.1 mol L −1 ) having different concentrations of SM, PG, and the SM+PG mixture under the optimized assay conditions. As shown in Figure 3A (Table S2), the SM detection performance of the g-C 3 N 4 /CuS/TiO 2 PEC sensing platform is excellent. In particular, the proposed sensing platform exhibits a distinctly higher detection sensitivity and a wider linear detection range, which completely satisfy actual sample detection requirements. The photocurrent response of the g-C 3 N 4 /CuS/TiO 2 PEC sensor to PG was lower than that of the SM sensor ( Figure 3B  These results indicate that the photoelectric response from the two-component mixture of antioxidants was significantly greater than the combined photocurrents of individual components at the same concentration. This indicates a synergistic effect between SM and PG, especially at high concentrations. Based on these findings, we investigated the synergistic effects of SM and VE, BHA, BHT, and TBHQ. As shown in Figure S4, SM was mixed in pairs with these four typical antioxidants, and the obtained trends are similar to that of the two-component mixtures of SM+PG. Presumably, an additive effect occurs at low concentrations, and a synergistic effect occurs at high concentrations. It has been reported that long distances suppress electron transport between antioxidants at low concentrations. Therefore, at relatively high concentrations, frequent electron transfer and the coupling with the oxidant may promote the regeneration of antioxidants, thus resulting in a more pronounced synergistic effect [29][30][31]. To understand the sensing principle of our sensor, as well as the observed synergistic effects, the electrochemical properties and photoresponses of the antioxidants were investigated in detail. As shown in Table S1, cyclic voltammetry of the antioxidants at a concentration of 484.812 µmol L −1 was carried out, yielding the following redox potentials: SM (0.373 V) < VE (0.464 V) < TBHQ (0.531 V) < BHA (0.539 V) < BHT (0.550 V) < PG (0.657 V). SM exhibited the smallest redox potential, suggesting its ability to reduce antioxidants having higher redox potentials [29,32]. As shown in Figure 4 and Table S1, the synergistic effect of binary antioxidant mixtures SM+VE, SM+TBHQ, SM+BHA, SM+BHT, and SM+PG resulted in increases of 1.57-, 1.38-, 2.31-, 2.42-, and 2.67-times, respectively, compared to the sum of the single antioxidant. This result is similar to the trend in the redox potential of the antioxidants. However, the synergistic effect of SM+VE was more obvious than that of SM+TBHQ, which is inconsistent with the increasing trend in redox potential. This is because both SM and VE are natural antioxidants, and their molecular structure and electronic properties (nucleophilic, electrophilic, ionization potential, and phenolic bond dissociation energy) are similar, thereby showing better performance [33]. Furthermore, we investigated the antioxidant effects of mixtures of natural antioxidants (SM+VE), four artificial antioxidants (BHA+BHT+TBHQ+PG), and six natural and artificial antioxidants (SM+VE+BHA+BHT+TBHQ+PG). All the samples showed improvements in the photocurrent responses, SM+VE, BHA+BHT+TBHQ+PG, and SM+VE+BHA+BHT+TBHQ+ PG showed increases of 1.57-, 1.64-, and 1.56-times, respectively, in relation to the sum of photoresponses of all single antioxidants. These results indicate that there are synergistic and antagonistic effects between the antioxidant molecules. In particular, there is a negative correlation between complexity of the antioxidant components and the antioxidant effect; thus, it is crucial to select a suitable formula to maximize the antioxidant performance [34]. These data also indicate that natural antioxidants exhibit good antioxidant properties and could be used to replace synthetic antioxidants as food stabilizers. These results indicate that the photoelectric response from the two-component mixture of antioxidants was significantly greater than the combined photocurrents of individual components at the same concentration. This indicates a synergistic effect between SM and PG, especially at high concentrations. Based on these findings, we investigated the synergistic effects of SM and VE, BHA, BHT, and TBHQ. As shown in Figure S4, SM was mixed in pairs with these four typical antioxidants, and the obtained trends are similar to that of the two-component mixtures of SM+PG. Presumably, an additive effect occurs at low concentrations, and a synergistic effect occurs at high concentrations. It has been reported that long distances suppress electron transport between antioxidants at low concentrations. Therefore, at relatively high concentrations, frequent electron transfer and the coupling with the oxidant may promote the regeneration of antioxidants, thus resulting in a more pronounced synergistic effect [29][30][31]. To understand the sensing principle of our sensor, as well as the observed synergistic effects, the electrochemical properties and photoresponses of the antioxidants were investigated in detail. As shown in Table S1, cyclic voltammetry of the antioxidants at a concentration of 484.812 μmol L −1 was carried out, yielding the following redox potentials: SM (0.373 V) < VE (0.464 V) < TBHQ (0.531 V) < BHA (0.539 V) < BHT (0.550 V) < PG (0.657 V). SM exhibited the smallest redox potential, suggesting its ability to reduce antioxidants having higher redox potentials [29,32]. As shown in Figure 4 and Table S1, the synergistic effect of binary antioxidant mixtures SM+VE, SM+TBHQ, SM+BHA, SM+BHT, and SM+PG resulted in increases of 1.57-, 1.38-, 2.31-, 2.42-, and 2.67-times, respectively, compared to the sum of the single antioxidant. This result is similar to the trend in the redox potential of the antioxidants. However, the synergistic effect of SM+VE was more obvious than that of SM+TBHQ, which is inconsistent with the increasing trend in redox potential. This is because both SM and VE are natural antioxidants, and their molecular structure and electronic properties (nucleophilic, electrophilic, ionization potential, and phenolic Based on the above results, we conclude that coupling oxidation plays a key role in the synergism of the studied antioxidants: specifically, it reduces the potential difference between pairs of antioxidants and promotes their regeneration, thus improving the performance of binary mixtures of antioxidants. Additionally, the heterojunctions in the g-C 3 N 4 /CuS/TiO 2 composite promote the reaction between antioxidants and holes, thus improving the separation and transition of electrons and holes (Scheme 1) and amplifying the detectable signal. Therefore, the g-C 3 N 4 /CuS/TiO 2 -modified platform could be applied for the detection and evaluation of the synergistic effect of SM based antioxidants. tively, in relation to the sum of photoresponses of all single antioxidants. These results indicate that there are synergistic and antagonistic effects between the antioxidant molecules. In particular, there is a negative correlation between complexity of the antioxidant components and the antioxidant effect; thus, it is crucial to select a suitable formula to maximize the antioxidant performance [34]. These data also indicate that natural antioxidants exhibit good antioxidant properties and could be used to replace synthetic antioxidants as food stabilizers.
. Based on the above results, we conclude that coupling oxidation plays a key role in the synergism of the studied antioxidants: specifically, it reduces the potential difference between pairs of antioxidants and promotes their regeneration, thus improving the performance of binary mixtures of antioxidants. Additionally, the heterojunctions in the g-C3N4/CuS/TiO2 composite promote the reaction between antioxidants and holes, thus improving the separation and transition of electrons and holes (Scheme 1) and amplifying the detectable signal. Therefore, the g-C3N4/CuS/TiO2-modified platform could be applied for the detection and evaluation of the synergistic effect of SM based antioxidants.

Detection Selectivity Stability and Reusability ofg-C3N4/CuS/TiO2 PEC Sensing Platform
Sensor selectivity depends largely on the band structure of the composite. Specifically, an appropriate VB can oxidize target species with a lower redox potential, but does not react with the sugars and amino acids in food. As shown in Figure 5A, we tested the effect of 18 times of fructose, glucose, sucrose, L-malic acid, L-citric acid, ethanol, L-threonine, L-proline, L-lysine, and L-histidine, as well as 700 times of Na + , K + , Mg 2+ and Ca 2+ on photocurrent based on g-C3N4/CuS/TiO2 PEC sensor with 123.46 μmol L −1 SM. No detectable signals (<10%) corresponding to the interfering agents were detected, suggesting that the PEC sensor hashigh SM detection selectivity. Furthermore, we had tested the photocurrent of real samples (soybean oil and peanut oil) in the same conditions of SM, there

Detection Selectivity Stability and Reusability ofg-C 3 N 4 /CuS/TiO 2 PEC Sensing Platform
Sensor selectivity depends largely on the band structure of the composite. Specifically, an appropriate VB can oxidize target species with a lower redox potential, but does not react with the sugars and amino acids in food. As shown in Figure 5A, we tested the effect of 18 times of fructose, glucose, sucrose, L-malic acid, L-citric acid, ethanol, L-threonine, L-proline, L-lysine, and L-histidine, as well as 700 times of Na + , K + , Mg 2+ and Ca 2+ on photocurrent based on g-C 3 N 4 /CuS/TiO 2 PEC sensor with 123.46 µmol L −1 SM. No detectable signals (<10%) corresponding to the interfering agents were detected, suggesting that the PEC sensor hashigh SM detection selectivity. Furthermore, we had tested the photocurrent of real samples (soybean oil and peanut oil) in the same conditions of SM, there are insignificant change in the photocurrent response, which further verified the feasibility of the sensor in the analysis of antioxidant capacity of food.
Sensor stability and reusability are also important criterions to evaluate the performance ofPEC sensor, especially for practical, long-term detection in real samples. As shown in Figure 5B, the g-C 3 N 4 /CuS/TiO 2 PEC sensor was monitored using amperometry with 500 s of "on-off-on" cycling in the presence of 1270.574 µmol L −1 SM, which showed no obvious decrease in the photoelectric signal. After 15 days, the photocurrent signal was reduced to 96.8% of the initial state, indicating that the sensor has excellent stability and reusability.

SM Detection in Soybean and Peanut Oils
Oils and fats naturally become rancid as a result of oxidation during storage or processing. The addition of synthetic or natural antioxidants can slow down lipid peroxidation and speed up the scavenging of free radicals [11,35,36]. In view of the high selectivity of the g-C 3 N 4 /CuS/TiO 2 PEC sensor platform, the applicability of the sensor for use with real oil samples was investigated using soybean and peanut oils containing various SM concentrations. The corresponding calibration curve is shown in Figure 3A. As listed in Table 1, both soybean and peanut oil samples exhibited high recoveries and low relative standard deviations (RSD; <3%) at different SM concentrations, indicating excellent detection precision and accuracy. Therefore, the natural antioxidant SM showed excellent antioxidant properties and could replace synthetic antioxidants as quality stabilizers for edible oils, which is important for food safety, quality control, and ingredient screening.
Sensor stability and reusability are also important criterions to evaluate the performance ofPEC sensor, especially for practical, long-term detection in real samples. As shown in Figure 5B, the g-C3N4/CuS/TiO2 PEC sensor was monitored using amperometry with 500 s of "on-off-on" cycling in the presence of 1270.574 μmol L −1 SM, which showed no obvious decrease in the photoelectric signal. After 15 days, the photocurrent signal was reduced to 96.8% of the initial state, indicating that the sensor has excellent stability and reusability.

SM Detection in Soybean and Peanut Oils
Oils and fats naturally become rancid as a result of oxidation during storage or processing. The addition of synthetic or natural antioxidants can slow down lipid peroxidation and speed up the scavenging of free radicals [11,35,36]. In view of the high selectivity of the g-C3N4/CuS/TiO2 PEC sensor platform, the applicability of the sensor for use with real oil samples was investigated using soybean and peanut oils containing various SM concentrations. The corresponding calibration curve is shown in Figure 3A. As listed in Table 1, both soybean and peanut oil samples exhibited high recoveries and low relative standard deviations (RSD; <3%) at different SM concentrations, indicating excellent detection precision and accuracy. Therefore, the natural antioxidant SM showed excellent antioxidant properties and could replace synthetic antioxidants as quality stabilizers for edible oils, which is important for food safety, quality control, and ingredient screening.

Conclusions
We developed a novel red-light-mediated signal-on PEC sensor based on a g-C3N4/CuS/TiO2 ternary heterojunction for a highly sensitive analysis of the synergistic

Conclusions
We developed a novel red-light-mediated signal-on PEC sensor based on a g-C 3 N 4 / CuS/TiO 2 ternary heterojunction for a highly sensitive analysis of the synergistic antioxidant effects of SM and other antioxidants. The g-C 3 N 4 /CuS/TiO 2 heterojunction forms a Schottky barrier that facilitates electron transfer. When SM and other antioxidants are present, they scavenge the holes in the VB of g-C 3 N 4 /CuS/TiO 2 , suppressing the recombination of photogenerated charge carriers, thus increasing the PEC photocurrent. The sensor also allowed the rapid evaluation of the synergistic effects of SM and VE, BHA, BHT, TBHQ, and PG. The synergistic antioxidant reaction of multiple antioxidants is mainly driven by coupling oxidation, which enhances the total antioxidant performance by lowering the potential difference between the antioxidants. We found that SM reduces the free radicals of other antioxidants and promotes their regeneration because of its low redox potential. Overall, the g-C 3 N 4 /CuS/TiO 2 PEC sensor exhibited high sensitivity, good selectivity, and stability, and was successfully applied to the detection of SM in soybean and peanut oils. The reported sensor has applications in nutrition analysis and the development of nutritional additives, which could help improve the quality of life of consumers. In addition, the sensor has potential uses in food fraud analysis, as well as medical research.