A Cu2+-doped two-dimensional material-based heterojunction photoelectrode: application for highly sensitive photoelectrochemical detection of hydrogen sulfide

In this work, on the basis of a Cu2+-doped two-dimensional material-based heterojunction photoelectrode, a novel anodic photoelectrochemical (PEC) sensing platform was constructed for highly sensitive detection of endogenous H2S. Briefly, with g-C3N4 and TiO2 as representative materials, the sensor was fabricated by modifying g-C3N4/TiO2 nanorod arrays (NAs) onto the surface of fluorine-doped tin oxide (FTO) and then doping Cu2+ as a CuxS (x = 1, 2) precursor. After the binding of S2− with surface-attached Cu2+, the signal was quenched owing to the in situ generation of CuxS which offers trapping sites to hinder generation of photocurrent signals. Since the photocurrent inhibition was intimately associated with the concentration of S2−, a highly sensitive PEC biosensor was fabricated for H2S detection. More importantly, the proposed sensing platform showed the enormous potential of g-C3N4/TiO2 NAs for further development of PEC bioanalysis, which may serve as a common basis for other semiconductor applications and stimulates the exploration of numerous high-performance nanocomposites.


Introduction
Since hydrogen sulde (H 2 S) was found to be the third endogenously generated gaseous signaling molecule following nitric oxide and carbon monoxide with cytoprotective properties, great attention has been drawn in the eld of clinical diagnostics. [1][2][3][4] In addition, H 2 S has also been known to play a crucial role in a series of physiological processes, including antioxidation, 5 anti-inammation 6 and apoptosis. 7 On the other hand, when the release concentration of H 2 S in the atmosphere is greater than the olfactory perception threshold of 300 ppb, it will harm human health and induce nausea, headaches, and lung irritation. 8 Even chronic, low-level exposures can also lead to irreversible health effects. 9,10 From this point of view, there is an essential demand to develop a reliable and highperformance approach for H 2 S monitoring.
Photoelectrochemical bioanalysis represents an elegant route for highly sensitive detection and exhibits versatile advantages of decreased costs, simple sample preparation, high sensitivity and selectivity, [11][12][13][14][15][16][17][18] which has inspired the rapid development of this eld in recent years. Previously, many PEC analytical methods have been exploited for H 2 S detection. [19][20][21][22][23][24] The most common strategy is in situ generated CdS on the surface of TiO 2 to enhance the photocurrent response. [20][21][22][23][24] But the strategy of in situ sensitization via CdS has its own limitations on only tting for the PEC substrates with low photoelectrical activity. And to the best of our knowledge, few works had been conducted for H 2 S detection with the strategy of in situ quenching.
2D materials have been among the most important research hotspots in the past years for their superlative physical properties and manifold implications in various elds. These materials consist of atomically thin sheets with large specic surface area exhibiting covalent in-plane bonding and weak interlayer and layer-substrate bonding. Besides, 2D materials can not only display improved inherent properties of the bulk materials but also give birth to new properties that the corresponding bulk materials do not possess. [25][26][27] On the other hand, in pursuit of achieving better semiconducting performances, heterostructures comprised by different semiconductors are being considered as favorite schemes as compared to the pure ones. It is believed that such a structure could integrate different properties of the individual semiconductors and thus generate enhanced properties. [28][29][30][31][32] Hence, of particular interest here is the possibility of utilizing ingenious 2D material-based heterojunction for innovative PEC detection of S 2À . We hypothesize that such a PEC platform possesses great potential in improving performance of PEC detection of S 2À . If possible, the great enhancement of light-harvesting efficiency is beneting from the feature of 2D material with large surface area, meanwhile, heterojunction is taken fully advantages of the contribution to the photoinduced charge separation in both the semiconductors and inhibition of the charge recombination, resulting in the improvement in photocurrent generation. 33,34 To verify this hypothesis, with g-C 3 N 4 and TiO 2 as representative materials, herein, we put forward a novel and general PEC sensing platform for highly sensitive detection of H 2 S through modifying FTO substrate with Cu 2+ -doped g-C 3 N 4 /TiO 2 NAs (Scheme 1). In this work, different from the previous strategies for PEC sensors of H 2 S, the obvious photocurrent quenching is appeared upon exposure to S 2À , owing to in situ formed Cu x S (x ¼ 1, 2) has a much lower conduction band edge than g-C 3 N 4 and offers plentiful surface recombination centers. 35,36 Thereby, in the presence of S 2À , the photocurrent intensity is expected to have an evident slip and by monitoring the reduction of photocurrent, we could quantitatively determine the concentration of S 2À .

Materials and reagents
Fluorine-doped tin oxide (FTO) glass substrate with a thickness of 1.1 mm (sheet resistance # 15 U per square) was ordered from South China Science & Technology Co. Ltd. Urea (CO(NH 2 ) 2 ), tetrabutyl titanate (C 16 H 36 O 4 Ti), hydrochloric acid (HCl), acetone (C 3 H 6 O), anhydrous ethanol (C 2 H 5 OH), copper sulfate (CuSO 4 ) and triethanolamine (TEOA) were all purchased from Sinopharm Chemical Reagent Co. Ltd. All other reagents were of analytical grade and used as received. Additionally, all aqueous solutions were prepared with deionized water (DI water, 18 MU cm À1 ), which was obtained from a MilliQ water purication system.

Synthesis of g-C 3 N 4 nanosheets
10.0 g of urea was put into an alumina crucible with a cover, heated with a ramp rate of 10 C min to 550 C in air atmosphere in a muffle furnace, and maintained for 4 h. Aerward, the resulting pale yellow agglomerate was milled into powder by a mortar and the g-C 3 N 4 nanosheets were obtained by liquid exfoliation of the bulk g-C 3 N 4 powder in water. Briey, the bulk g-C 3 N 4 powder was dispersed into 100 mL of distilled water and ultrasonicated for 2 h. The residual unexfoliated bulk g-C 3 N 4 was removed by centrifugation at 4500g. Subsequently, the supernatant was further centrifuged at 8000g, and the obtained precipitation was dried at 70 C in a oven.

Preparation of the biosensor
The g-C 3 N 4 /TiO 2 /FTO was rst prepared as follows using a onestep hydrothermal method. First, 10 mg of as-prepared g-C 3 N 4 powder was homogeneously dispersed in 6 mL of ultrapure water via sonication, and the obtained suspension was mixed with equal volume of concentrated hydrochloric acid. Aer 5 min stirring, 200 mL of tetrabutyl titanate (TBOT) was added into the above suspension and stirred for 30 min. Then the above mixture was transferred into a Teon-lined stainless steel autoclave. Subsequently, pieces of cleaned FTO substrate were placed against the wall of the autoclave with conductive sides facing down. The autoclave was kept in an oven at 150 C for 10 h and then allowed to cool down to room temperature. Finally, the FTO substrates were removed, rinsed with ultrapure water, put into a muffle furnace and annealed at 450 C for 1 h to form TiO 2 /g-C 3 N 4 NAs on the FTO substrate surface. To prepare Cu x S precursor, the obtained g-C 3 N 4 /TiO 2 /FTO was immersed into 1.0 mM CuSO 4 solution with 1 h gently shaking to dope Cu 2+ on electrode surface.

PEC measurement
The as-prepared PEC sensing platform was exposed to different concentration of Na 2 S solution (H 2 S in the aqueous medium) for 10 min. Followed by washing with ultrapure water thoroughly to remove excess S 2À . Aer that, the resulted substrate

Materials characterization
Experimentally, g-C 3 N 4 and g-C 3 N 4 /TiO 2 /FTO were prepared through a thermo-polymerization method 32 and a modied hydrothermal method according to a previous report. 37 The structural and morphology information of the as-prepared g-C 3 N 4 /TiO 2 NAs were characterized by scanning electron microscopy (SEM), transmission electron microscopy (TEM). As exhibited in Fig. 1a, Cu 2+ -doped g-C 3 N 4 /TiO 2 NAs were vertically grown on the surface of FTO and g-C 3 N 4 could not distinctly tell via SEM. As indicated in the le inset of Fig. 1a, white g-C 3 N 4 / TiO 2 lm shows the great affinity to FTO. As shown in the right inset of Fig. 1a, the energy dispersive X-ray spectrum (EDX) of Cu 2+ -doped g-C 3 N 4 /TiO 2 /FTO distinctly verify the existence of Ti, O, C, N and Cu elements. The strong peaks for O and Ti elements were found in the spectrum, due to its abundant amount in the composites. Moreover, the presence of Cu in the spectrum revealed that the doping was successful. Fig. 1b of elemental mapping also indicates a uniform distribution of Ti, O, C, N and Cu elements in the sample. Through the TEM image, a typical planar sheet-like the exfoliated g-C 3 N 4 sample with an irregular shape was displayed in Fig. 1c. In addition, Fig. 1d indicates TiO 2 NRs was fully wrapped by g-C 3 N 4 and the existence of a very close and distinguishable g-C 3 N 4 /TiO 2 interface, revealing the moderate interfacial contact and successful synthesis of g-C 3 N 4 /TiO 2 nanohybrid.
The composition and crystal-phase properties of g-C 3 N 4 / TiO 2 /FTO was identied by X-ray diffraction (XRD). As  the growth of TiO 2 on FTO substrates with high orientation selectivity. No other impurity peaks were detected. As for the g-C 3 N 4 /TiO 2 /FTO (curve d), the feature diffraction at 27.5 could properly match the (002) plane of g-C 3 N 4 (curve a), implying that g-C 3 N 4 was successfully modied on the surface of TiO 2 nanorods. Then as illustrated in Fig. 2b, X-ray photoelectron  Paper spectroscopy (XPS) was applied to study the surface chemical compositions and oxidation states of Cu 2+ -doped g-C 3 N 4 /TiO 2 / FTO before (curve a) and aer (curve b) reaction with S 2À . Curve a veries the presence of C, N, Ti, O and Cu peaks and a new characteristic peak of S 2p emerged, suggesting the presence of sulfur element on the electrode. Fig. 2c reveals the highresolution XPS spectrum of Cu 2p 3/2 , which could be further distributed into three parts located at 932.4, 931.4, and 929.2 eV. As reported in the previous work, 38 the main peak at 931.4 eV of Cu + was produced from the interaction between supercial Cu 2+ and S 2À , indicating the presence of Cu x S (x ¼ 1, 2) on the surface of g-C 3 N 4 /TiO 2 NAs. Besides, the one weak peak at 932.4 eV was attributed to CuS, 39 while the other at 929.2 eV was assigned to the tiny amount of CuO. Combined with above results, Cu x S was newly formed on the electrode surface aer incubation of S 2À . In addition, the electrochemical impedance spectroscopy (EIS) was also employed to analyze the interfacial properties of the modied electrode in 5 mM [Fe(CN) 6 ] 4À/3À (as the redox probe) containing 0.1 M KCl. The impedance spectrum includes a semicircular portion at higher frequencies and a linear portion at lower frequencies. The diameter of the semicircle is equal to the electron transfer resistance (R et ). As shown in Fig. 2d, the R et value of FTO (curve a) was very small, aer coating g-C 3 N 4 /TiO 2 on electrode surface, R et value increased signicantly (curve b), which may be due to the introduction of g-C 3 N 4 /TiO 2 hindered the electron transfer on the electrode. Then, the R et value increased aer the modied electrode treated with 50 nM S 2À (curve c). The reason for the increase in the resistance value were owing to the generation of Cu x S which offered the trapping sites with new energy levels and thus suppressed the electron transfer on the surface of the g-C 3 N 4 /TiO 2 /FTO.

Optimization of experimental conditions
As PEC sensing performance was affected by some factors inuencing the photocurrent response like the length of TiO 2 NAs and the mass of g-C 3 N 4 , the optimal preparation conditions were conducted. The length of TiO 2 NAs could be controlled via hydrothermal reaction time. As revealed in Fig. 3a, the photocurrent intensity was increasing with the increased reaction time of TiO 2 NAs. However, if reaction time is extended to over 10 h, the white lm composed of aligned TiO 2 NAs starts to peel off from the FTO substrate during annealing because the length of TiO 2 NAs was too long to adhere to the FTO substrate consistently. As a result, 10 h is the optimal reaction time associated with maximal photocurrent intensity of TiO 2 NAs. Fig. 3b shows the effects on photocurrent response of g-C 3 N 4 / TiO 2 /FTO prepared with different mass of g-C 3 N 4 . It can be told that g-C 3 N 4 /TiO 2 /FTO produced the highest photocurrent with 10.0 mg g-C 3 N 4 added. As the mass grew, g-C 3 N 4 immobilized on TiO 2 lm gradually increased, thereby leading to wider light  absorption range. With more g-C 3 N 4 added, excessive g-C 3 N 4 immobilization occurred on TiO 2 lm, which offered more surface recombination centers and the resistance to decrease the photocurrent intensity. Thereby, 10.0 mg of g-C 3 N 4 was added into the mixture in subsequent experiments.

Characterization of the PEC biosensor
To evaluate the feasibility of the sensing platform, their PEC behaviors were then characterized by chronoamperometric it curves from the stepwise transient photocurrent responses upon intermittent light irradiation. As shown in Fig. 4a, the photocurrent of bare FTO is negligible (curve a), while g-C 3 N 4 /TiO 2 NAs electrode exhibited the signicantly strong photocurrent (curve b). Compared with g-C 3 N 4 /TiO 2 /FTO, the photocurrent response of Cu 2+ -doped g-C 3 N 4 /TiO 2 /FTO appeared a small decrease (curve c) which resulted from the capture of photoelectrons by Cu 2+ . And as can be seen from curve d, the photocurrent signal of Cu 2+ -doped g-C 3 N 4 /TiO 2 / FTO with the exposure of the electrode to a 50 nM Na 2 S solution caused a noticeable quenching, because of the generation of Cu x S which offered the trapping sites with new energy levels and thus suppressed the electron transfer on the surface of the g-C 3 N 4 /TiO 2 /FTO, as illustrated in Fig. 4b. 40 In addition, the corresponding UV-vis diffuse reectance spectra were also performed to further conrmed the enhanced absorption aer the successful preparation of Cu 2+ -doped g-C 3 N 4 /TiO 2 /FTO, as shown in Fig. 4c. Fig. 4d indicated the signal response of g-C 3 N 4 /TiO 2 /FTO upon irradiation repeated every 10 s. The irradiation process was repeated over 400 s and no obvious variation could be observed, featuring the high stability of the biosensor. Thereby, all of these results demonstrated the feasible fabrication of the PEC sensing platform.

Analytical performance
The substantial photocurrent decrement in the presence of trace amounts of the H 2 S captured in an aqueous medium (for convenience, Na 2 S was used as the source here) demonstrated the suitability of the fabricated biosensor for S 2À and H 2 S determination. Fig. 5a exhibited the decrement of photocurrent aer reaction with various S 2À concentrations. Fig. 5b shows the photocurrent decrement linearly increased with the increasing S 2À concentrations from 1 Â 10 À9 to 5 Â 10 À6 M and the lowest detection limit of S 2À was estimated at 5.8 Â 10 À11 M (S/N ¼ 3), which was comparable to other H 2 S PEC sensors in Table 1. As demonstrated in Table 1, different than the common zero-dimensional and one-dimensional materialbased heterojunction photoelectrode, g-C 3 N 4 /TiO 2 /FTO as the representative two-dimensional material-based heterojunction photoelectrode truly has the relatively better photoelectric conversion efficiency and perfectly ts for H 2 S detection with in situ quenching strategy. The reproducibility of the PEC biosensor was assessed on the basis of the relative standard deviation (RSD) for the intra-assay and interassay precision. The intra-assay precision was obtained by parallel measuring S 2À ve times at concentrations of 10 nM, 50 nM, and 100 nM, which yielded a RSD values of 4.0%, 3.6%, and 5.2%, respectively. The interassay precision was determined by assaying S 2À at the same concentration using ve sensing electrodes prepared under identical conditions, where the RSD values were 5.8%, 4.6%, and 5.4%, respectively. These results indicated the satisfactory precision and reproducibility of this biosensor. To verify the selectivity of the PEC sensor, the common anions and other species potentially coexisting in the solution, including NO 3 À , Cl À , SO 4 2À , CH 3 COO À and HPO 4

2À
were selected for interference test. As displayed in Fig. 5c, the photocurrent response to the interfering ions with the addition of 100-fold excess in comparison with S 2À were very close to the blank test, because of the interfering ions cannot have the reaction with Cu 2+ to in situ generate the substrate insoluble in aqueous solution. Moreover, the photocurrent of the mixture containing S 2À was approximately the same as pure S 2À , indicating that the coexistence of the S 2À with the interfering ions did not have a signicant effect on the photocurrent of the sensing platform. Additionally, the long term stability performance of the designed sensing platform was also evaluated. There was no apparent change of the photocurrent response aer the biosensor was stored at 4 C in a refrigerator for over 1 month, and 94.4% of the initial photocurrent response was maintained aer storage for over 2 months, suggesting the robustness of the as-designed PEC sensor. For the feasibility of practical application, the performance of the sensing platform was tested in human plasma. Different concentrations of mixed anions were then added. As shown in Fig. 5d, the small signal difference between normal human plasma and Tris-HCl solution samples indicated the precision of the sensing platform and the potential for practical applications.

Conclusions
In summary, we successfully designed and fabricated a novel and general PEC sensing platform for highly sensitive H 2 S detection based on Cu 2+ -doped g-C 3 N 4 /TiO 2 NAs heterojunction photoelectrode, in situ formed Cu x S would open a new pathway for the electronÀhole recombination and thus efficiently inhibit the photocurrent generation of the sensing platform. Importantly, the above biosensor was highly sensitive and easy to prepare, manifesting a wide linear response range with S 2À detection limit of 58 pM. This study displayed the desirable potential of g-C 3 N 4 /TiO 2 as the representative 2D materialbased heterojunction in improving performance of PEC detection and was expected to inspire more interests in the implementation of numerous other semiconductor applications. Further work will focus on experimental optimization for better performance.

Conflicts of interest
There are no conicts to declare.