CdS-Modified TiO2 Nanotubes with Heterojunction Structure: A Photoelectrochemical Sensor for Glutathione

The formation of heterojunction structures can effectively prevent the recombination of photogenerated electron–hole pairs in semiconductors and result in the enhancement of photoelectric properties. Using TiO2 nanotubes (prepared using the hydrothermal-impregnation method) as carriers, CdS-TiO2NTs were fabricated as a photoelectrochemical (PEC) sensor, which can be used under visible light and can exhibit good PEC performance due to the existence of the heterojunction structure. The experimental results show that the prepared CdS-TiO2NTs electrode had a linear response to 2–16 mM glutathione (GSH). The sensor’s sensitivity and detection limit (LOD) were 102.9 µA·mM−1 cm−2 and 27.7 µM, respectively. Moreover, the biosensor had good stability, indicating the potential application of this kind of heterojunction PEC biosensor.


Introduction
Glutathione (GSH) is a free-radical scavenger in the body and plays a primary role in functions such as immune regulation, gene expression regulation, enzyme activity, etc. [1]. Many human diseases can influence GSH concentration, and GSH detection is very important due to its significance in physiological circumstances [2]. Among different biomolecule detection methods, photoelectrochemical (PEC) detection is prominent due to its high response speed, high sensitivity, and simple instruments [3].
PEC sensors' foundations are based on the relation between concentrations of the detected substances and photocurrents. The PEC sensor, which is commonly fabricated by photoactive materials, is firstly photoexcitated to form electron-hole pairs [4][5][6]. During the reaction of photoexcitation species with analytes, the photocurrent is obtained. Thus, a crucial aspect of a prominent PEC sensor is the selecting of an excellent photoelectrode, which will control the sensor's sensitivity, detection limit (LOD), etc. [7].
Nowadays, the electrode materials with photoelectric responses are focused on inorganic materials, such as TiO 2 [8], CdS [9,10], ZnO [11], and CuO [12]. Among them, TiO 2 is most prominent due to its stable chemical properties, high PEC performance, and good biocompatibility [13]. Compared with common powder materials, TiO 2 nanotubes are more prominent for their unique 1D nanostructure, which can shorten the distance between photogenerated carriers and accelerate electron transport [14]. The modifiers and aimed samples can also be highly dispersed across the nanotubes, and result in a quick response. As is common sense, UV rays are harmful to most biological molecules. Due to its big band gap, the pure TiO 2 photoelectrode can only be used under UV light, which limits the utility of TiO 2 as a biosensor [15]. The fast recombination of photogenerated charges is another shortcoming of TiO 2 , which decreases the PEC sensor's sensitivity.
To improve the PEC performance of TiO 2 , such as narrowing its energy band and improving its electron-hole separation efficiency, different metals and nonmetals (C [16], S [17], N [18], Fe [19], Ti [20], Bi [21], and so on) were used as dopants, and precious metals (Au [22], Pt [23,24], Ag [25], and so on) were used as modifications. Doping and modification can decrease the recombination ratio of photoelectrons and photoholes and promote their separation in TiO 2 [26,27]. However, the impurity level formed by element doping may become the recombination site of the photogenerated electrons and holes, and precious metals are expensive and easily lost in the PEC reaction process.
As has been generally reported, constructing heterojunctions with other semiconductors (CdS [8], CuO [28], MoS 2 [29], g-C 3 N 4 [30], SnS 2 [31], and so on) is also an effective way to improve TiO 2 's PEC performance. Among the above-mentioned semiconductors, CdS's band gap is 2.3 eV, which can lead to a higher absorption coefficient under visible-light illumination. However, the unstable properties of CdS limits its practical utility. Forming composite materials with stable TiO 2 is a good solution to this limitation. In our previous work, we have prepared TiO 2 nanotube-supported CdS as a photocatalyst. Compared with pure TiO 2 nanotubes or CdS nanoparticles, the fabricated composite exhibited better photoelectric properties [32]. According to the simulation of density functional theory (DFT), the formation of the heterojunction structure can also be considered the existence of an internal electric field [33]. After the heterojunction structure is formed in the CdS-TiO 2 system, the charge-transfer barrier is decreased, and the photogenerated electrons and holes can be effectively separated, resulting in a faster charge transport rate and lower recombination efficiency. Therefore, the formed CdS-TiO 2 sensors are not only active under visible light but can also lead to better sensing properties being obtained for biomolecules [34,35]. It has been reported that CdS modification can obtain 2.9-fold PEC signal enhancement compared with that of 0.1% for Fe-TiO 2 [36].
Based on the improved PEC properties of the heterojunction structure, a TiO 2 nanotube was modified with powdery CdS to form a CdS-TiO 2 NTs electrode, which was further applied for GSH detection in this paper. Through the characterization, investigation of PEC sensitivity, and stability of the CdS-TiO 2 NTs electrode, the practicality and sensing mechanism of the fabricated CdS-TiO 2 NTs sensor is also discussed.

Synthesis of CdS-TiO 2 NTs
Hydrogen-titanate nanotubes prepared using the hydrothermal method were soaked in titanium sol for 10 min, naturally dried in the air, and calcined at 400 • C for 2 h to obtain TiO 2 NTs [37]. A certain amount of sulfur powder was dissolved in tetrahydrofuran solution. After a quantity of Cd(NO 3 ) 2 ·4H 2 O was slowly added through stirring, the prepared TiO 2 NTs were added. The calculated CdS/TiO 2 molar ratio was 1:1. NaBH 4 solution was slowly added under the condition of an ice water bath and N 2 . After being stirred for 30 min, the solution was ultrasonic treated for 30 min, and then stirred overnight. The samples were successively cleaned by tetrahydrofuran, water, and ethanol, and then treated in a vacuum drying oven to obtain CdS-TiO 2 NTs.

Preparation of CdS-TiO 2 NTs Electrode
20 µL of CdS-TiO 2 NTs solution was dropped on a glassy-carbon electrode. After the dropped solution was dried, the rest of the solution was added, and 15 µL of 3% chitosan solution was dropped. After naturally air drying overnight, the CdS-TiO 2 NTs electrode was obtained.

Photoelectrochemical Testing
The CdS-TiO 2 NTs electrode's electrochemical properties were investigated using the three-electrode system from the Zahner electrochemical workstation. The CdS-TiO 2 NTs electrode, Ag/AgCl electrode, and Pt electrode were used as the working electrode, reference electrode, and auxiliary electrode, respectively. An LED lamp with visible light (429 nm, 35 W/m 2 ) was used as the illumination source, and the pH of electrolytes was maintained at 7.4.

Instrument Model
XRD patterns were obtained using the Rigaku D/MAX-2500 X-ray diffractometer (Rigaku SmartLab, Rigaku Corporation, Tokyo, Japan). Scanning electron microscope (SEM) and energy dispersive spectrometer (EDS) images were observed on JSM-7800F (Japanese electronics, Beijing, China). To obtain TEM and HR-TEM images, JEM-2800 (Japanese electronics, Shanghai, China) and the Talos F200X G2 transmission electron microscope were used, respectively. X-ray photoelectron spectroscopy (Kratos Axis Ultra DLD (Kratos Analytical Ltd., Manchester, UK), Al target, double anode Al/Mg target) recorded the XPS patterns. All the electrochemical experiments were performed on the Zennium electrochemical workstation (Kronach-Gundelsdorf, Germany, bias voltage: 0.5 V).  (Figure 2b), indicating the chemical state Ti 4+ of titanium oxide. As displayed in Figure 2c, the O 1s XPS spectrum showed two peaks, with the one at 531.4 eV corresponding to lattice oxygen, and the other at 532.9eV corresponding to adsorbed oxygen. As shown in Figure 2d, the peaks at BE of 404.7 eV and 411.5 eV corresponded to Cd 3d 5/2 and Cd 3d 3/2 , respectively, indicating the existence of Cd 2+ in CdS-TiO 2 NTs. The high-solution S 2p peaked at 161.2 eV and 162.4 eV, which is evidence for S 2in CdS. In summary, the XPS results in Figure 2a (Figure 2b), indicating the chemical state Ti 4+ of titanium oxide. As displayed in Figure 2c, the O 1s XPS spectrum showed two peaks, with the one at 531.4 eV corresponding to lattice oxygen, and the other at 532.9eV corresponding to adsorbed oxygen. As shown in Figure 2d, the peaks at BE of 404.7 eV and 411.5 eV corresponded to Cd 3d5/2 and Cd 3d3/2, respectively, indicating the existence of Cd 2+ in CdS-TiO2NTs. The high-solution S 2p peaked at 161.2 eV and 162.4 eV, which is evidence for S 2-in CdS. In summary, the XPS results in Figure 2a-e show the coexistence of CdS and TiO2, consistent with the results of TEM and SEM.

Surface Morphology Characterization
In order to investigate the composition of the prepared materials, XRD patterns were performed. Figure 3 shows the XRD patterns of TiO2NTs and CdS-TiO2NTs. Most of the peaks on pattern a in Figure 3 corresponded to that of anatase TiO2 (JCPDS# 21-1271), indicating that the TiO2NTs are mainly anatase type. The peak at 11° corresponded to the (200) crystal plane of H2Ti3O7 (JCPDS # 47-0561), which is the undecomposed hydrogen titanate. As shown on pattern b in Figure 3, CdS diffraction peaks (JCPDS# 10-0454) emerged in CdS-TiO2NTs. The addition of CdS weakened the diffraction peaks of TiO2, which could be attributed to the high concentration of CdS in TiO2. Moreover, the wide diffraction peaks indicate the small particle sizes of CdS and TiO2. By combining the results of the above TEM, SEM, and XPS analyses, the successful loading of CdS on TiO2 nanotubes can be confirmed. In order to investigate the composition of the prepared materials, XRD patterns were performed. Figure 3 shows the XRD patterns of TiO 2 NTs and CdS-TiO 2 NTs. Most of the peaks on pattern a in Figure 3 corresponded to that of anatase TiO 2 (JCPDS# 21-1271), indicating that the TiO 2 NTs are mainly anatase type. The peak at 11 • corresponded to the (200) crystal plane of H 2 Ti 3 O 7 (JCPDS # 47-0561), which is the undecomposed hydrogen titanate. As shown on pattern b in Figure 3, CdS diffraction peaks (JCPDS# 10-0454) emerged in CdS-TiO 2 NTs. The addition of CdS weakened the diffraction peaks of TiO 2 , which could be attributed to the high concentration of CdS in TiO 2 . Moreover, the wide diffraction peaks indicate the small particle sizes of CdS and TiO 2 . By combining the results of the above TEM, SEM, and XPS analyses, the successful loading of CdS on TiO 2 nanotubes can be confirmed.  EIS is commonly used to detect electrode-electrolyte interactions and charge-transfer phenomena [38]. The arc radius in the EIS Nyquist plot reveals the charge-transfer resistance, and charge-transfer ability of the electrode [39]. As shown in Figure 4, the Rs of the samples were almost all the same. However, the Rct for the TiO2NTs and CdS/TiO2NTs were 153 Ω and 102.7 Ω, respectively. Compared with TiO2NTs, the value of the CdS/TiO2NTs was smaller. This could be attributed to the formation of a heterojunction structure in the CdS/TiO2NTs, which resulted in the rapid transmission of charges and reduced charge-transfer resistance. Under light illumination, the effective separation and transmission of photogenerated carriers resulted in the small Rct of CdS/TiO2NTs. In the PEC process, the small Rct can facilitate the transfer of electrons and result in a high photocurrent. EIS is commonly used to detect electrode-electrolyte interactions and charge-transfer phenomena [38]. The arc radius in the EIS Nyquist plot reveals the charge-transfer resistance, and charge-transfer ability of the electrode [39]. As shown in Figure 4, the Rs of the samples were almost all the same. However, the Rct for the TiO 2 NTs and CdS/TiO 2 NTs were 153 Ω and 102.7 Ω, respectively. Compared with TiO 2 NTs, the value of the CdS/TiO 2 NTs was smaller. This could be attributed to the formation of a heterojunction structure in the CdS/TiO 2 NTs, which resulted in the rapid transmission of charges and reduced charge-transfer resistance. Under light illumination, the effective separation and transmission of photogenerated carriers resulted in the small Rct of CdS/TiO 2 NTs. In the PEC process, the small Rct can facilitate the transfer of electrons and result in a high photocurrent.
phenomena [38]. The arc radius in the EIS Nyquist plot reveals the charge-transfer resistance, and charge-transfer ability of the electrode [39]. As shown in Figure 4, the Rs of the samples were almost all the same. However, the Rct for the TiO2NTs and CdS/TiO2NTs were 153 Ω and 102.7 Ω, respectively. Compared with TiO2NTs, the value of the CdS/TiO2NTs was smaller. This could be attributed to the formation of a heterojunction structure in the CdS/TiO2NTs, which resulted in the rapid transmission of charges and reduced charge-transfer resistance. Under light illumination, the effective separation and transmission of photogenerated carriers resulted in the small Rct of CdS/TiO2NTs. In the PEC process, the small Rct can facilitate the transfer of electrons and result in a high photocurrent.

Electrochemical Characterization
The sensitivity of the fabricated CdS-TiO 2 NTs for GSH was monitored under 429 nm LED-light irradiation. Figure 5a shows the photocurrent-time curves of the CdS-TiO 2 NTs in GSH solutions with 18 different concentrations, which ranged from 0.5 mM to 17 mM. Figure 5b displays the relationship between the GSH concentrations and the photocurrents of the CdS-TiO 2 NTs. The photocurrent-GSH concentration showed a linear relationship in the range of 2-16 mM. The regression equation is Y = −20.2X + 354 (R 2 = 0.9953) (Y referring to photocurrent, and X referring to GSH concentration). The detection limit (LOD) of the CdS-TiO 2 NTs biosensor was calculated based on the equation of LOD = 3 Sb/S (Sb is the standard deviation of the blank signal, and S is the sensitivity of the fabricated sensor), which was equal to 27.7 µM.
The stability of the CdS-TiO 2 NTs sensor was also investigated. The CdS-TiO 2 NTs electrode was stored in air at room temperature. Its current response to PBS solution was monitored 9 times under 429 nm LED-light irradiation (bias voltage 0.3 V). As shown in Figure 6, the CdS-TiO 2 NTs maintained at least 89.3% of the initial value during this test, suggesting the relatively high stability of the fabricated CdS-TiO 2 NTs electrode [39].
TiO 2 NTs exhibit no sensitivity to GSH under visible light. However, the CdS-TiO 2 NTs presented good sensitivity under the same condition. Obviously, the modification of CdS endowed the TiO 2 with PEC activity under visible light. Moreover, the heterojunction structure of the CdS-TiO 2 composite induced rapid transmission of the charges. In our previous work, a density functional theory (DFT) simulation was conducted for a similar CdS-TiO 2 composite [40]. As indicated by the simulation results, tunnels formed at the interlayer section of the CdS-TiO 2 composite and facilitated the charge transfer between CdS and TiO 2 . During the PEC process, CdS was firstly excited under visible light. Because the conduction bands (CBs) of CdS are higher than that of TiO 2 , the excited electrons in CdS transferred to the CBs of TiO 2 and generated a photocurrent. in GSH solutions with 18 different concentrations, which ranged from 0.5 mM to 17 mM. Figure 5b displays the relationship between the GSH concentrations and the photocurrents of the CdS-TiO2NTs. The photocurrent-GSH concentration showed a linear relationship in the range of 2-16 mM. The regression equation is Y = −20.2X + 354 (R 2 = 0.9953) (Y referring to photocurrent, and X referring to GSH concentration). The detection limit (LOD) of the CdS-TiO2NTs biosensor was calculated based on the equation of LOD = 3 Sb/S (Sb is the standard deviation of the blank signal, and S is the sensitivity of the fabricated sensor), which was equal to 27.7 µ M. Nanomaterials 2022, 12, x 8 Figure 6, the CdS-TiO2NTs maintained at least 89.3% of the initial value during this suggesting the relatively high stability of the fabricated CdS-TiO2NTs electrode [39]. TiO2NTs exhibit no sensitivity to GSH under visible light. However, the C TiO2NTs presented good sensitivity under the same condition. Obviously, the modi tion of CdS endowed the TiO2 with PEC activity under visible light. Moreover, the he Based on the experiment results and our former simulation results, the PEC detection process for GSH by CdS-TiO 2 NTs electrode biosensor in this work was preliminarily put forward to explain the improved performance of CdS-TiO 2 NTs, as shown in Figure 7. Under the illumination of visible light, electrons are photo-generated in CdS and transferred from the CBs of CdS to the CBs of TiO 2 , reaching the glassy carbon electrode and eventually the counter electrode. Meanwhile, the holes in CdS react with GSH to form GSSG. During this process, the inner self-combination of electrons and holes in both CdS and TiO 2 are decreased, and electrons and holes are accelerated at different parts. Thus, the association of CdS with TiO 2 can result in superior PEC performance.
TiO2NTs presented good sensitivity under the same condition. Obviously, the modifica tion of CdS endowed the TiO2 with PEC activity under visible light. Moreover, the hetero junction structure of the CdS-TiO2 composite induced rapid transmission of the charges In our previous work, a density functional theory (DFT) simulation was conducted for a similar CdS-TiO2 composite [40]. As indicated by the simulation results, tunnels formed at the interlayer section of the CdS-TiO2 composite and facilitated the charge transfer be tween CdS and TiO2. During the PEC process, CdS was firstly excited under visible light Because the conduction bands (CBs) of CdS are higher than that of TiO2, the excited elec trons in CdS transferred to the CBs of TiO2 and generated a photocurrent.
Based on the experiment results and our former simulation results, the PEC detection process for GSH by CdS-TiO2NTs electrode biosensor in this work was preliminarily pu forward to explain the improved performance of CdS-TiO2NTs, as shown in Figure 7. Un der the illumination of visible light, electrons are photo-generated in CdS and transferred from the CBs of CdS to the CBs of TiO2, reaching the glassy carbon electrode and eventu ally the counter electrode. Meanwhile, the holes in CdS react with GSH to form GSSG During this process, the inner self-combination of electrons and holes in both CdS and TiO2 are decreased, and electrons and holes are accelerated at different parts. Thus, the association of CdS with TiO2 can result in superior PEC performance.

Conclusions
In summary, CdS-TiO 2 NTs with a heterojunction structure were successfully designed as a PEC sensor for GSH detection. The CdS-TiO 2 NTs exhibited good PEC properties due to the effective charge transfer between CdS and TiO 2 . The photocurrent signals of the CdS-TiO 2 NTs biosensor were linear to 2-16 mM GSH solution. In addition to GSH, future applications of this PEC sensor to other substances are anticipated.