Electronic Structure Engineering of Cu2O Film/ZnO Nanorods Array All-Oxide p-n Heterostructure for Enhanced Photoelectrochemical Property and Self-powered Biosensing Application

We have engineered the electronic structure at the interface between Cu2O and ZnO nanorods (NRs) array, through adjusting the carrier concentration of Cu2O. The electrodeposition of Cu2O at pH 11 acquired the highest carrier concentration, resulting in the largest interfacial electric field between Cu2O and ZnO, which finally led to the highest separation efficiency of photogenerated charge carriers. The optimized Cu2O/ZnO NRs array p-n heterostructures exhibited enhanced PEC performance, such as elevated photocurrent and photoconversion efficiency, as well as excellent sensing performance for the sensitive detection of glutathione (GSH) in PBS buffer even at applied bias of 0 V which made the device self-powered. Besides, the favorable selectivity, high reproducibility and extremely wide detection range, make such heterostructure a promising candidate for PEC biosensing applications, probably for the extended field of PEC water splitting or other solar photovoltaic beacons.

of electrolyte during the electrodeposition synthesis process 27 . So this provides an approach to engineer the electronic structure at the interface of Cu 2 O/ZnO heterostructure. Besides, the non-toxicity, abundance and low cost of copper, as well as the long-term stability associated with oxides, allows for the feasibility of constructing PEC devices with environment-friendly Cu 2 O 28,29 .
In this work, we assembled Cu 2 O/ZnO NRs array all-oxide p-n heterostructure by directly electrodepositing p-type Cu 2 O film onto the vertically oriented n-type ZnO NRs array. As illustrated in Figure 1, the step-wise structure of energy levels constructed in the Cu 2 O/ZnO heterostructure was responsible for the mechanism of PEC biosensor. When Cu 2 O and ZnO came into contact, a p-n junction formed at their interface. There was a favorable energy band alignment for electrons transport from the conduction band (CB) of Cu 2 O to the CB of ZnO, and for holes transport from the valence band (VB) of ZnO to the VB of Cu 2 O. In the presence of illumination, the photoinduced electrons and holes were efficiently separated by the built-in interfacial electric field. The electrons transported through each individual ZnO NR and finally reached the FTO electrode to export the PEC electrical signals. While the holes migrated through Cu 2 O film and subsequently took part in the oxidation of GSH to GSSG at the surface of Cu 2 O. The quasi-1 D nature of ZnO NRs array significantly increased the interface area between ZnO and Cu 2 O, and the introduction of Cu 2 O largely improved the visible light absorption. Simultaneously, the engineered Cu 2 O/ZnO electronic structure at interface and electron transport channels of the individual ZnO NRs both contributed to the efficient separation of photoinduced charges, thus resulting in the enhanced PEC property and ideal biosensing performance for detection of GSH.

Results
Characterization of Cu 2 O/ZnO heterostructure. Figure 2 shows the morphology of Cu 2 O/ZnO NRs array heterostructure prepared by electrodeposition of Cu 2 O at pH 11 for different durations. Independent ZnO NRs, with average length of ,1.8 mm and diameter of ,150 nm, were vertically grown on the FTO surface. For the electrodeposition of Cu 2 O for 10 min (Fig 1A and B), the individual Cu 2 O crystal particles spread all over the top surface of ZnO NRs array, and covered almost half of the surface area. When the electrodeposition time was increased to 20 min ( Fig 1C and D), the Cu 2 O crystal particles started to form a film except for several points where ZnO was still exposed. By comparing the inset of Figure 2A and Figure 2C, the size of the Cu 2 O crystal particles obviously increased. For the samples of 30 min ( Figure 2E and F), the Cu 2 O particles further expand their size, and the Cu 2 O film covered all over the ZnO NRs. By comparing Figure 2B, D and F,the thickness of Cu 2 O film increased from ,0.8 mm to ,1.1 mm along with the growth time. It is worth pointing out that the Cu 2 O crystal particles partially sank into the ZnO NRs array, so that the interface area between Cu 2 O and ZnO was significantly raised. In addition, the longer time ZnO was immersed in the alkaline electrodeposition solution, the more serious etching happened to ZnO NRs. Moreover, it is noticed that with the increase of size and amount of Cu 2 O particles, the vertical ZnO NRs suffered more stress and strain from Cu 2 O particles. However, taking all factors in terms of the PEC property into consideration, we still choose 30 min for electrodeposition of Cu 2 O film. Figure 3A shows the X-ray diffraction (XRD) patterns of the ZnO NRs array and fabricated Cu 2 O/ZnO heterostructure. The diffraction peaks of the FTO substrate were marked with spades. For pristine ZnO NRs array, diffraction peaks were identified to the hexagonal ZnO crystalline phase with a wurtzite structure. After the deposition of Cu 2 O at pH 11 for 30 min, diffraction peaks from Cu 2 O were observed in addition to those from ZnO NRs array, which can be indexed to the cubic phase Cu 2 O (JCPDS 78-2076). No diffraction peaks of impurities were found in the XRD patterns.
X-ray photoelectron spectroscopy (XPS) ( Figure 3B) was performed in order to identify the chemical state of the Cu element in the heterostructure. Since XPS is only capable for surface elemental analysis, the Zn element cannot be observed in the wide scan spectrum. In the Cu 2p core level spectrum, the Cu 2p 3/2 and Cu 2p 1/2 spin-orbital photo-electrons were located at binding energies of 932.6 eV and 952.5 eV, respectively. Such results are in good agreement with the reported values of Cu 2 O 25,30 , thus demonstrating the layer electrodeposited upon ZnO NR array was Cu 2 O rather than Cu or CuO. Figure 4A shows the UV-vis diffuse reflectance spectra (DRS) of pristine ZnO NRs array and Cu 2 O/ZnO heterostructure with varied electrodeposition time of Cu 2 O at pH 11. The pristine ZnO resulted in an obvious absorption below 380 nm, which originated from the band edge of ZnO. And only a little absorption in the visible region can be observed, which might have resulted from the scattering effect of the NRs array structure. However, the absorption spectrum was significantly enhanced after subsequent electrodeposition of Cu 2 O. It is noteworthy that the visible light absorption difference between 10 min sample and 20 min sample was greater than that between 20 min sample and 30 min sample, which was corresponding to the isolated Cu 2 O nanocrystals for 10 min growth time, almost covering Cu 2 O film for 20 min growth time and the full covering Cu 2 O film for 30 min growth time. The strong absorption below 600 nm is attributed to the band edge absorption of the nanocrystaline Cu 2 O film. It suggests that the so-fabricated Cu 2 O/ZnO NRs array heterostructure has a quite broad absorption range from visible to ultraviolet, which well agrees with the solar irradiation. Consequently, good photoelectrochemical property of such nanostructure under white light irradiation is expected. The following PEC measurements were all based on the samples with electrodeposition of Cu 2 O for 30 min.
Electronic structure engineering of Cu 2 O/ZnO heterostructure. Figure 4B shows the current density of pristine ZnO NRs array and Cu 2 O/ZnO heterostructures electrodepostied at various pH values without and with illumination as the function of applied bias potential. For pristine ZnO, the dark current density is on the order of 10 23 mA/cm 2 , and the photocurrent density has a slight increasment. While all samlpes with Cu 2 O/ZnO heterostructure showed pronounced photoresponse under illumination, which can be attributed to the improved visible light absorption and efficient interfacial charge transport of Cu 2 O/ZnO heterostructure. Noticeably, the photoresponse was gradually improved along with the increase of the electrodepostion pH value from 9 to 11. The saturated photocurrent density of Cu 2 O (pH 11)/ZnO was approximately 13.5 times that of pristine ZnO NRs array, 2.3 times  31,32 , which in turn led to the alteration of the built-in potential in Cu 2 O/ZnO hetrostructure 14,15 . As illustrated in Figure 5, the built-in potential V bi is equal to the difference between the Fermi levels of ZnO and Cu 2 O. Consequently, the samples with higher electrodepostion pH value, that is with higher carrier oncentration of Cu 2 O, possess larger builtin potential. When the illumination was applied, the photoinduced electron-hole pairs were separated under the effect of the interfacial electric field. Therefore, Cu 2 O (pH 11)/ZnO got the highest charge separation efficiency, which resulted in the best photoresponse at bias potential of 0V.
Based on the discussion above, Cu 2 O/ZnO heterostructure electrodeposited at pH value of 12 or 13 should perform better photoresponse. However, in such strong alkaline electrodepostion electrolytes, the dissolution and etching of ZnO NRs array is very serious, which would certainly introduce a high concentration of interface states at the Cu 2 O/ZnO heterostructure, thus to a large extent affecting the quality of the heterostructure 10,14 . Moreover, considering the stability of the fabricated devices, pH 11 is the optimal parameter for construction of Cu 2 O/ZnOheterostructure based PEC biosesnsor with ideal performance. So the following PEC measurements were based on the samples acquired at pH 11.
PEC property of Cu 2 O/ZnO heterostructure photoanode. Electrochemical impedence spectroscopy (EIS) was carried out both in the dark and under illumination at bias potential of 0 V vs. Ag/AgCl. It is known that the semicircle in the Nyquist plot at high frequencies is the characteristic of the charge transfer process, and the diameter of the semicircle is equal to the charge transfer resistance (Rct). As shown in Figure 6A, the Rct of Cu 2 O/ZnO heterostructure was apparently much smaller than that of the pristine ZnO NRs array under illumination. Such result indicates that the formation of Cu 2 O/ZnO heterostructure dramatically promoted the interfacial charge transport and the separation efficiency of photoinduced charges under illumination, thus demonstrating the enhanced PEC performance for Cu 2 O/ZnO NRs array heterostructure.
PEC behaviors of pristine ZnO NRs array and Cu 2 O/ZnO heterostructure with and without addition of GSH at bias potential of 0V were also characterized in Figure 6B. No photocurrent was observed in the dark. Upon photoexcitation, the photocurrent composed of three steps: (a) the photocurrent climbed promptly; (b) the photocurrent decreased sharply; (c) the photocurrent gradually reached a stable state. These three stages are corresponding with the photoin-  duced electrons under illumination, the recombination of photoinduced electron-hole pairs, and the balance of generation and recombination process, respectively 25 . The pristine ZnO NRs array photoanode performed an average photocurrent density of 0.097 mA/cm 2 (curve a), whereas the Cu 2 O/ZnO heterostructure showed an average photocurrent density of 1.35 mA/cm 2 (curve b), demonstrating the remarkable improvement of photoelectric conversion efficiency. The formed p-n heterojunction between Cu 2 O and ZnO provided an interfacial electric field, which subsequently contributed to the charge separation efficiency, and finally led to the significantly enhanced photoresponse. Moreover, the elevated photocurrent could be further strengthened with the presence of sacrificial reagent who served as the hole scavenger. Since GSH plays an important role in many biological functions, it was adopted as a model molecule to demonstrate the feasibility of fabricating Cu 2 O/ZnO heterostructure based PEC biosensor. When GSH was added in the blank PBS (200 mM), the average photocurrent density reached 2.91 mA/cm 2 (curve c). The great photocurrent change indicated the fabricated Cu 2 O/ZnO heterostructure is suitable for detecting GSH based on PEC method.
There is no doubt that when a positive bias potential is applied, the photocurrent density will apparently increase due to the more efficient charge separation and longer lifetime of photoinduced elec-tron-hole pairs. Neverthless, in order to realize the self-powered function, plus taking energy conservation and the selectivity of PEC biosensors into consideration, 0 V was selected as bias potential in this work.
Self-powered PEC biosensing of GSH. In order to further verify the sensitive sensing of GSH based on Cu 2 O/ZnO heterostructure, Figure 7 displays the photocurrent response in the presence of GSH with various concentrations at applied bias potential of 0 V (vs. Ag/AgCl). The photocurrent density performed a fine linear relationship with the concentration of GSH from10 to 1000 mM (R 2 5 0.991), which is much wider than that of porous TiO 2 -Pt nanowhisker 33 , flower-like Cu 2 O/ZnO 29 and IrO 2 -Hemin-TiO 2 6 photoanodes based on PEC method, as well as Au nanoparticles@ Si nanowires 34 , Hg/Pd 35 and ordered mesoporous carbon 36 photoanodes based on electrochemical method. The detection limit was estimated as low as 0.42 mM through 3s. The detailed performance comparisons with previously reported literatures are listed in Table S1. Obviously, the proposed Cu 2 O/ZnO NRs array heterostructure based PEC biosensor shows great promise for applications in the monitoring of GSH with high sensitivity and wide linear range.  The selectivity of Cu 2 O/ZnO heterostructure based PEC biosensor was confirmed by measuring the photocurrent response not only for common chemical or biological interferences but also for various metal ions with the concentration of 50 mM, showing much smaller or negligible signals compared to that of GSH with the concentration of 200 mM ( Figure 8A and B). The excellent selectivity is ascribed to the low applied bias potential which minimized the interference from other reductive species 6,37 . Furthermore, the relative standard deviation (R.S.D) of the photoresponse to 200 mM GSH was 5.2% for six successive measurements. The R.S.D for detection of 200 mM GSH with six different Cu 2 O/ZnO based photoanode under the same condition was 6.4%. In addition, the photocurrent density was stable for 2 h in 0.1 M PBS containing 10 mM GSH ( Figure S5), and 91% of the initial photocurrent response was maintained after storing for over half month in the repeating test ( Figure S6). All the results demonstrated the reproducibility and stability of fabricated heterostructures and their practical potential for PEC biosensing application.

Discussion
The enhanced PEC property and excellent PEC biosensing performance was probably attributed to the following five points: (a) The quasi 1-D nature of ZnO NRs array offered large p-n heterojunction interface area when in combination with Cu 2 O film; (b) The introduction of Cu 2 O significantly improved the visible light absorption, so that the solar energy could be utilized more sufficiently; (c) The electronic structure between Cu 2 O and ZnO was engineered thorough adjusting the carrier concentration in Cu 2 O and the thickness of Cu 2 O film. Thereby, the optimal build-in potential was acquired at the Cu 2 O/ZnO interface, resulting in the efficient charge separation; (d) The step-wise energy band structure facilitated the photoinduced electrons to the electrode while promoted the photoinduced holes to accumulate in the VB of Cu 2 O and subsequently to be consumed by participating in the oxidation of GSH. This process further prevented the recombination of photogenerated carriers; (e) The individual and vertical alignment of single crystal ZnO NRs provides transport channels for electrons to rapidly reach FTO electrode.
To conclude, the Cu 2 O/ZnO NRs array all-oxide p-n heterostructure was successfully fabricated, and the electronic structure at the interface between Cu 2 O and ZnO was engineered through adjusting the carrier concentration of Cu 2 O. The results demonstrated the electrodeposition of Cu 2 O at pH 11 acquired the highest carrier concentration, resulting in the largest build-in electric field, which finally led to the highest separation efficiency of photoinduced charge carriers. The optimized heterostructure exhibited enhanced PEC performance and was subsequently adopted for self-powered PEC sensing of GSH. The Cu 2 O/ZnO based PEC biosensor performed a linear range from 10 mM to 1000 mM and an estimated detection limit of 0.42 mM even at bias potential of 0 V. Moreover, the excellent selectivity, reproducibility and stability suggested that such heterostructure is a competitive candidate for advanced PEC biomolecular detection, maybe for the extended field of PEC water splitting or other solar photovoltaic beacons.
Preparation of Cu 2 O/ZnO NRs array based photoanode. The ordered ZnO NRs array was prepared by the hydrothermal growth method on fluorine doped tin oxide (FTO) substrates [38][39][40] . The ZnO seed layer was acquired by spin coating a colloidal solution of Zn(CH 3 COO) 2 .2H 2 O (0.5 M) onto the FTO substrate, and subsequently annealed at 350 uC in air for 30 min. In the hydrothermal process, Zn(NO 3 ) 2 .6H 2 O and HMTA were dissolved in deionized (DI) water with equal concentration of 50 mM. Besides, PEI (5 mM) was dissolved in the solution to enhance the aspect ratio of the NRs array. The substrate coated with the ZnO seed layer was placed in the precursor solution at 90 uC for 3 h, and then rinsed with DI water for three times.
The electrodeposition of Cu 2 O thin film was conducted with a three-electrode system in water bath (40 uC). Aqueous solution containing Cu 2 SO 4 (0.05 M) and LA (0.1 M) was prepared, and the pH value was adjusted to 9.0, 10.0, 11.0 and 12.0 through dropwise adding NaOH (4 M). Afterward, the prepared ZnO NRs array was immersed into the resulting solution under the potential of 20.4 V vs. Ag/AgCl for 10 min, 20 min and 30 min. Finally, the prepared Cu 2 O/ZnO heterostructure was again rinsed with DI water for three times and dried in nitrogen gas. www.nature.com/scientificreports