High-Performance Self-Powered UV Detector Based on SnO2-TiO2 Nanomace Arrays

Photoelectrochemical cell-typed self-powered UV detectors have attracted intensive research interest due to their low cost, simple fabrication process, and fast response. In this paper, SnO2-TiO2 nanomace arrays composed of SnO2 nanotube trunk and TiO2 nanobranches were prepared using soft chemical methods, and an environment-friendly self-powered UV photodetector using this nanostructure as the photoanode was assembled. Due to the synergistic effect of greatly accelerated electron-hole separation, enhanced surface area, and reduced charge recombination provided by SnO2-TiO2 nanomace array, the nanostructured detector displays an excellent performance over that based on bare SnO2 arrays. The impact of the growing time of TiO2 branches on the performance of UV photodetector was systematically studied. The device based on optimized SnO2-TiO2 nanomace arrays exhibits a high responsivity of 0.145 A/W at 365 nm, a fast rising time of 0.037 s, and a decay time of 0.015 s, as well as excellent spectral selectivity. This self-powered photodetector is a promising candidate for high-sensitivity, high-speed UV-detecting application.


Background
Ultraviolet photodetectors (UVPDs) have been widely used in many fields, such as remote control, chemical analysis, water purification, flame detection, early missile plume detection, and secure space-to-space communication [1]. To avoid the use of costly UV pass filters and achieve visibleblind operation, wide bandgap semiconductors have been studied widely for light detecting, especially in the ultraviolet region [2]. In recent decades, nanostructured semiconductors such as nanorods, nanowires, nanotubes, and nanobranches have attracted extensive research interest due to their high surface-to-volume ratio and rationally designed surface morphology [3][4][5][6][7][8][9][10][11][12][13]. Photoelectrochemical cell (PEC)-typed photodetectors assembled with nanostructured semiconductors exhibit a high responsivity and a fast transient response compared with traditional photoconductive semiconductor thin film detectors. As a new and efficient way to fabricate high-performance photodetectors, PEC-based devices can avoid complicated epitaxial processes and expensive single crystal substrates, which is very important for the growing cheaper optoelectronic applications. Therefore, self-powered UVPDs based on PEC device have attracted intensive research interest. Self-powered UVPDs based on PEC structure have been fabricated using a liquid I − /I 3 − redox couple electrolyte [14][15][16][17][18] and a nanocrystalline TiO 2 film [14] or a multilayer TiO 2 nanorod-assembled cloth/nanorod array-based electrode [15]. Impressive performances were observed in these UVPDs. However, liquid I − /I 3 − redox couple electrolyte is not ideal for long-term operation: it is highly corrosive, volatile, and photoreactive, interacting with common metallic components and sealing materials. From this point, water-based electrolytes may be the most safe, most stable, and most environment-friendly electrolyte. Zhang et al. have reported a UV-visible photodetector based on ZnO/ CuO heterojunctions and NaSO 4 aqueous solution, which shows an excellent photodetection performance [19]. TiO 2 has attracted a great deal of attention due to its outstanding physical and chemical properties for water electrolytebased UVPDs. Lee et al. reported a UV detector based on a TiO 2 film/water solid-liquid heterojunction [20], which exhibits high photosensitivity, excellent spectral selectivity and fast response. In order to further enlarge the TiO 2 /electrolyte contact area, Xie et al. fabricated a self-powered PEC photodetector based on TiO 2 nanorod arrays/water UVPD [21]. Until now, water electrolyte-based UVPDs still show a lower photoresponsivity than those using I − /I 3 − redox couple electrolyte. Moreover, low electron mobility of TiO 2 increases the probability of photon-induced electron recombination with the electrolyte. By contrast, SnO 2 possesses a high electron mobility, suggesting a faster diffusion transport of photon-induced electrons to the transparent conductive oxide current collector. Recently, highquality TiO 2 /SnO 2 heterojunction nanostructures have been prepared by different methods for optoelectronic applications [17,22]. Impressive performance has been observed in UVPDs using TiO 2 /SnO 2 branched heterojunctions and SnO 2 mesoporous spheres @ TiO 2 as electrode materials [16,17]. However, all these UVPDs were assembled with disordered nanostructures. It can be expected that if ordered SnO 2 -TiO 2 nanostructure arrays with a high electron transport efficiency are adopted as the photoanode of the UVPDs, a much better photodetecting performance can be obtained.
In this work, ordered SnO 2 -TiO 2 nanomace arrays (STNMAs) was synthesized using soft chemical methods. An environment-friendly self-powered UVPD was assembled using the STNMAs as photoanode and water as electrolyte. The schematic structure of STNMAs/H 2 O UVPD is shown in Fig. 1. STNMAs grown vertically on fluorinedoped tin oxide (FTO) glass were used as the active electrode. The STNMA-based device shows a higher photocurrent density than that of the bare SnO 2 nanotubebased device under UV irradiation. The spectral photosensitivity and response time are characterized to evaluate the potential of the STNMA UVPD. The impact of the growing time of TiO 2 branches on the performance of UV photodetector was also studied. The self-powered UVPD based on optimized STNMAs exhibits a high responsivity of 0.145 A/W, a fast rise time of 0.037 s, and a decay time of 0.015 s, as well as excellent spectral selectivity. Moreover, the electrolyte of this photodetector is water, which is low cost, stable, and environment friendly.

Synthesis of SnO 2 Nanotube Arrays
FTO glass (2 cm × 2 cm) was ultrasonically cleaned with ethanol and deionized water for 15 min respectively and then dried in the air. A 10-nm Sn film was deposited on FTO by thermal evaporation and annealed in air at 550°C for 1 h to form a dense SnO 2 layer. High-quality ZnO nanorod arrays were prepared on the SnO 2 buffered FTO glass by a two-step hydrothermal method. The details could be found in our previous work [23]. SnO 2 shell layer was deposited on the ZnO nanorod array by a liquid phase deposition. FTO covered with ZnO nanorod arrays was immersed in Na 2 SnO 3 aqueous solution at 60°C for 1 h. Then the sample was immersed in 0.01 M dilute hydrochloric acid to remove the ZnO template, and uniform SnO 2 nanotube arrays (SNAs) were obtained.

Synthesis of SnO 2 -TiO 2 Nanomace Arrays
TiO 2 nanobranches were grown on the SnO 2 nanotube trunk by a simple aqueous chemical growth method. The SnO 2 nanotube arrays on FTO glass prepared above were put in an aqueous solution of 0.2 M TiCl 4 at room temperature. In order to achieve different TiO 2 nanobranch length, the deposition was conducted at 6, 12, 18, and 24 h respectively. The resulted STNMAs were thoroughly rinsed with deionized water and then annealed at 450°C for 30 min.

Assemble of the UV Detector
The PEC-typed photodetector was assembled in a similar structure of a dye-sensitized solar cell, as discussed in our previous work [24]. In brief, the obtained STNMAs synthesized on FTO glass were used as the active electrode and a 20-nm-thick Pt film deposited on FTO glass by magnetron sputtering is adopted as the counter electrode. The active electrode (SnO 2 /FTO) and the counter electrode (Pt/FTO) were adhered together face to face with a 60-μm-thick sealing material (SX-1170-60, Solaronix SA, Aubonne, Switzerland). Finally, deionized water was injected into the space between the top and counter electrodes. The effective area of the UV detector was approximately 0.2 cm 2 .

Characterization
The crystal structure of the samples was examined by Xray diffraction (XRD; XD-3, PG Instruments Ltd., Beijing, China) with Cu Kα radiation (λ = 0.154 nm). The surface morphology of the samples was characterized using a field emission scanning electron microscope (FESEM; Hitachi S-4800, Hitachi, Ltd., Chiyoda, Tokyo, Japan) and a transmission electron microscope (TEM; F-20, FEI Company, Hillsboro, OR, USA). The optical transmittance was measured using an UV-visible dual beam spectrophotometer (TU-1900, PG Instruments, Ltd., Beijing, China). A 500-W Xenon lamp (7ILX500, 7Star Optical Instruments Co., Beijing, China) with a monochromator (7ISW30, 7Star Optical Instruments Co.) was used as UV light source to generate monochromatic light for the spectral response characterization. The spectral photoresponse characteristics were obtained by a programmable sourcemeter (2400, Keithley Instruments Inc., Cleveland, OH, USA). The photoresponse switching behavior measurement was obtained by an electrochemical workstation (RST5200, Zhengzhou Shirusi Instrument Technology Co. Ltd., Zhengzhou, China).

Results and Discussion
Morphology of SnO 2 nanotube arrays (SNAs) and STNMAs was examined by a FESEM. As shown in Fig. 2a, ordered SNAs with opened top were grown uniformly on the surface of FTO glass substrate. Further analysis indicates that the nanotubes have a diameter of 50-80 nm and a wall thickness less than 10 nm. The density of nanotubes is typically 30 nanotube/μm 2 . Figure 2b-e illustrates the SnO 2 nanotube arrays immersed in TiCl 4 solution for 6, 12, 18, and 24 h, respectively. It can be clearly seen that the SnO 2 nanotubes grow almost vertically to the FTO substrate and are covered with a large number of TiO 2 nanobranches to form a nanomace structure. The morphology of SNA and STNMA is also checked by TEM. As shown in Fig. 2g, h for the bare SNA and STNMA grown for 18 h, the SnO 2 nanotube has a length of about 500 nm and the TiO2 branches grow tightly on the wall of SnO 2 nanotubes. The morphology of the STNMAs is strongly dependent on the growth time. As the growth time increased, the branches become more numerous and longer. These nanobranches coated on SnO 2 nanotube would greatly enlarge the specific surface area and roughness, which is important for PEC applications. However, once the deposition time reaches 24 h or longer, the branches will form continuous network that greatly suppresses the effective active area, which would result in a decrease of the active area of TiO 2 contacted with the electrolyte. This is confirmed by the reduced performance The transmission spectrum of the FTO glass, SNAs, and STNMAs is shown in Fig. 3a. A sharp absorption edge located at 320 nm can be observed for FTO glass. The absorption edge of SnO 2 nanotube arrays and 6-hgrown STNMAs is similar to that of the FTO glass, but the absorption edge of 12-24-h-grown STNMAs all show an obvious red shift. The transmittance of FTO reaches to zero when the wavelength is shorter than 305 nm, which determines the spectral response edge in the short-wavelength region. The strong light scattering by the TiO 2 nanobranches causes a lower transmittance of all STNMAs than that of FTO and SnO 2 nanotubes in the wavelength range of 400-550 nm. From these transmittance spectra, it can be concluded that only light with the wavelength between 305 and 400 nm can be well absorbed by TiO 2 arrays and contribute to the UV photoresponsivity, which is confirmed in the following spectral response characterization. The spectral responsivity of these photodetectors was measured in the range of 300-550 nm at zero bias, as shown in Fig. 3b. The responsivity is calculated by the following formula: R = I/ AE, where R is the responsivity, I is the measured photocurrent, A is the active area of the photodetector device, and E is the irradiance intensity of the light source, which is measured by a standard light power meter. The device performs as a self-powered photodetector that operates at a nominal zero-applied voltage, with a large photocurrent response under a weak light illumination. As shown in Fig. 3b, the maximum responsivity value for a bare SNA-based UV photodetector is approximately 0.01 A/W at 335 nm, corresponding to an incident photon-to-current conversion efficiency (IPCE) of only 3.7%. Normally, oxygen vacancy can be easily formed in SnO 2 material and cause a high charge recombination. TiO 2 nanobranch deposition on the SNAs can passivate the surface of SnO 2 and reduce the electron-hole recombination. The STNMA-based photodetectors show much better UV photoresponsivity. The peak responsivity of STNMAs grown for 18 h is approximately 0.145 A/W at 365 nm. The corresponding IPCE is higher than 49.2%, which is much higher than other H 2 O-based PEC detectors at this wavelength [20,23,24]. Considering the loss of incident photons caused by the light absorption and scattering of the FTO glass, much higher internal quantum efficiency can be expected. The TiO 2 nanobranches coated on the SnO 2 nanotube arrays greatly increase not only the contact interface area between the STNMAs and the electrolyte but also the light scattering ability, resulting in an enhancement of the photon harvesting efficiency. Furthermore, these ultrathin branches are very effective at transporting holes to the TiO 2 /water interface as most electron-hole pairs are formed within the diffusion length, ultimately minimizing the recombination loss. Moreover, the photoelectrons injected into SnO 2 nanotube from TiO 2 nanobranch reach the FTO collecting electrode rapidly because SnO 2 possesses a higher electron mobility than TiO 2 . When the growth time reaches 24 h or longer, the branches on the nanotube arrays are interconnected. The active area of TiO 2 contacted with the electrolyte decreases. Therefore, an excessively long a b Fig. 3 The UV-visible transmittance spectra and responsivity spectrum of photodetectors. a Spectrum of transmittance for FTO glass substrate, SNAs, and STNMAs with different growth time. b Responsivity spectrum of photodetectors based on SNAs and STNMAs growth time is disadvantageous and leads to a reduced photovoltaic performance of the UV photodetector. To characterize their responsivity to a fast-changing light signal, the photocurrent density-time characteristics of the devices were measured at 0 V bias under an intermittent 365 nm UV-light irradiation with a power of 129 μW/cm 2 . The incident radiation is switched with an on/off interval of 10 s. Five repeated cycles are displayed in Fig. 4a, which indicates that the photocurrent can be reproducibly switched between the "ON" state and the "OFF" state by periodically turning the UV light on and off. When the deposition time of TiO 2 nanobranch is less than 6 h, the photocurrent density is quite low. In this case, only TiO 2 nanoparticles with a high defect density were formed on the surface of SnO 2 nanotube, which would result in a high electron-hole recombination and a poor photoresponse. With the increase of growth time, the crystal quality of the TiO 2 nanobranches was improved and the surface area was greatly increased. Therefore, the photocurrent has a significant increase when the growth time is longer than 6 h and reaches the maximum when the deposition time is 18 h. From the enlarged rising and decaying edges of the photocurrent response curve, the rising time and decay time of the UV detector are approximately 0.037 and 0.015 s (Fig. 4b, c), indicating a rapid photoresponse characteristic. The quantitative criterion for the rising time is the time to reach 90% of the stable photocurrent, and that for decay time is time to reach 1/e (37%) of the original photocurrent. The overall performance of the STNMA-based self-powered UV detector is considerably better than that reported by other works, as compared in Table 1.
Schematic diagram of energy band matching and device working mechanism are shown in Fig. 5. When the incident light travels through FTO glass and reaches the active layer of TiO 2 nanobranches, photons with energy exceeding the TiO 2 bandgap will be absorbed and electrons are excited from the valance band to the conduction band, and electron-hole pairs will be generated thereafter. The built-in potential across the interface works as the driving force to separate the electron-hole pairs. Negative electrons move along from TiO 2 nanobranch to the SnO 2 nanotube and get collected by the FTO electrode. These electrons will easily transfer into the external circuit and return to the Pt layer of the counter electrode since the work function of FTO matches with the conduction band of SnO 2 and TiO 2 .

Conclusions
In summary, we have synthesized SnO 2 -TiO 2 nanomace arrays composed of SnO 2 nanotube trunk and TiO 2 nanobranches using soft chemical methods. A selfpowered UV detector was assembled using this nanostructure as the active electrode and water as the electrolyte. Due to the accelerated electron-hole separation speed by the SnO 2 -TiO 2 core-shell structure, enlarged surface area of TiO 2 nanobranches, and fast electron transport property of SnO 2 nanotube, an excellent performance was obtained in this nanostructured photodetector. For the detector based on the optical STNMAs, a high IPCE up to 49.2% is observed at 365 nm, which is more than 10 times larger than the maximum IPCE of bare SnO 2 nanotube (3.7%). A rapid response time and an excellent spectral selectivity were also obtained in this photodetector. We believe that this SnO 2 -TiO 2 nanomace structure can be extended to other applications based on photoelectrochemical effect, such as dye-sensitized solar cells and photoelectrochemical hydrogen production.