Hydrothermal Growth of an Al-Doped α-Ga2O3 Nanorod Array and Its Application in Self-Powered Solar-Blind UV Photodetection Based on a Photoelectrochemical Cell

Herein, we successfully fabricated an Al-doped α-Ga2O3 nanorod array on FTO using the hydrothermal and post-annealing processes. To the best of our knowledge, it is the first time that an Al-doped α-Ga2O3 nanorod array on FTO has been realized via a much simpler and cheaper way than that based on metal–organic chemical vapor deposition, magnetron sputtering, molecular beam epitaxy, and pulsed laser deposition. And, a self-powered Al-doped α-Ga2O3 nanorod array/FTO photodetector was also realized as a photoanode at 0 V (vs. Ag/AgCl) in a photoelectrochemical (PEC) cell, showing a peak responsivity of 1.46 mA/W at 260 nm. The response speed of the Al-doped device was 0.421 s for rise time, and 0.139 s for decay time under solar-blind UV (260 nm) illumination. Compared with the undoped device, the responsivity of the Al-doped device was ~5.84 times larger, and the response speed was relatively faster. When increasing the biases from 0 V to 1 V, the responsivity, quantum efficiency, and detectivity of the Al-doped device were enhanced from 1.46 mA/W to 2.02 mA/W, from ~0.7% to ~0.96%, and from ~6 × 109 Jones to ~1 × 1010 Jones, respectively, due to the enlarged depletion region. Therefore, Al doping may provide a route to enhance the self-powered photodetection performance of α-Ga2O3 nanorod arrays.


Introduction
Ga 2 O 3 is a promising wide-bandgap oxide semiconductor for potential applications in solar-bind UV photodetection, high-power devices, gas sensors, and transparent conductive oxides [1][2][3][4][5]. Among these applications, solar-bind UV photodetection can be more simply and easily investigated, owing to its inherent solar-bind UV absorption properties, which can be applied in flame detection, UV astronomy and dosimetry, water purification, spaceto-space communication, and missile warnings [1][2][3][4][5]. Therefore, large amounts of research have been performed on the basis of the MSM (metal-semiconductor-metal) structure [3][4][5]. However, the MSM structure always needs external biases to separate the photogenerated carriers. Schottky junctions, p-n junctions, or n-n junctions can realize self-powered photodetection, owing to their built-in electric field [6]. And, such a mode has caught much attention, owing to it working at 0 V without external biases, which possesses the advanced properties of energy conservation and environmental protection [6]. In fact, self-powered photodetectors have been realized among other metal oxide semiconductors such as TiO 2 , ZnO, CuO, and NiO [6,7]. As a reference for Ga 2 O 3 -based photodetectors, constructing α-Ga 2 O 3 nanorod array and Al-doped α-Ga 2 O 3 nanorod array were realized, an undoped GaOOH nanorod array and Al-doped GaOOH nanorod array were grown on FTO at 200°C for 10 h through hydrothermal processes. During the hydrothermal growth of the Aldoped samples, gallium nitrate and aluminum nitrate aqueous solution mixed with a mole ratio of 1:0.1 for Ga and Al, and the solution was poured into the reactor as the precursor solution. For the undoped samples, the precursor solution only contained gallium nitrate aqueous solution. After the as-grown undoped GaOOH nanorod array/FTO and Al-doped GaOOH nanorod array/FTO samples were washed by ultrapure water and blown by dry N 2 gas, they were post-annealed immediately at 550 • C in O 2 (10 sccm) atmosphere for 2 h. Thus, the undoped α-Ga 2 O 3 nanorod array and Al-doped α-Ga 2 O 3 nanorod array were fabricated on FTO substrates. The material properties were investigated by scan electron microscopy (SEM), energy dispersive spectroscopy (EDS), and X-ray diffraction (XRD) with Cu Kα as a radiation source and UV-vis diffuse reflectance absorption spectra.
An undoped α-Ga 2 O 3 nanorod array/FTO and Al-doped α-Ga 2 O 3 nanorod array/FTO structure were immersed into a 0.5 M Na 2 SO 4 aqueous solution, and measured as photoanodes in a PEC cell. The undoped α-Ga 2 O 3 nanorod array/FTO or Al-doped α-Ga 2 O 3 nanorod array/FTO, Pt, Ag/AgCl were, respectively, used as the working electrode, the counter electrode and the reference electrode in a three-electrode system. The self-powered properties, including the dark current, light current and transient photoresponse, were investigated at 0 V (vs. Ag/AgCl) based on an electrochemical workstation. When the undoped α-Ga 2 O 3 nanorod array/FTO or Al-doped α-Ga 2 O 3 nanorod array/FTO, and Pt electrodes were connected with a lock-in amplifier, the spectral responsivity was also collected at 0 V (vs. Ag/AgCl) under illumination of the monochromatic light generated by a 300 W UV-enhanced Xe lamp and a monochromator. Figure 1a,b, respectively, shows a surface morphology picture of an undoped α-Ga 2 O 3 nanorod array/FTO and Al-doped α-Ga 2 O 3 nanorod array/FTO structure by SEM. From the surface morphology images, the Al-doped α-Ga 2 O 3 nanorod in Figure 1b also exhibits a diamond-like shape, which is similar to the α-Ga 2 O 3 nanorod array without doping, as shown in Figure 1a. Therefore, Al doping did not obviously change the surface morphology of the α-Ga 2 O 3 nanorod. As seen in Figure 1c, the EDS measurement obviously indicates the two dominated peaks at 0.52 keV and 1.11 keV for the undoped α-Ga 2 O 3 nanorod array which, respectively, stand for the O and Ga elements. For the Al-doped α-Ga 2 O 3 nanorod array, another EDS peak at 1.49 keV was also observed, corresponding to Al element. The EDS results also indicate an atom ratio of 5.57:1 for Ga: Al in the Al-doped α-Ga 2 O 3 nanorod array. To further evaluate the element distributions in the Al-doped α-Ga 2 O 3 nanorod array, Figure 1d-f are, respectively, the EDS mapping pictures of O, Ga, and Al element distributions in the Al-doped α-Ga 2 O 3 nanorod array, in detail. From Figure 1d-f, O, Ga, and Al elements are uniformly distributed in the Al-doped α-Ga 2 O 3 nanorod array. As a result, Al was successfully and uniformly doped into α-Ga 2 O 3 nanorods.

Results and Discussion
The XRD measurements of an undoped α-Ga 2 O 3 nanorod array/FTO and Al-doped α-Ga 2 O 3 nanorod array/FTO were investigated to reveal the crystal structure, as shown in Figure 2a.  [18,19], respectively. Owing to the smaller ionic radius of Al 3+ (0.0535 nm) than that of Ga 3+ (0.062 nm), the corresponding diffraction peaks of the Al-doped α-Ga 2 O 3 nanorod array slightly shift to a larger angle in comparison with the undoped α-Ga 2 O 3 nanorod array [18,19,24]. Figure 2b is the UV-vis diffuse reflectance absorption spectra of the undoped α-Ga 2 O 3 nanorod array/FTO and Al-doped α-Ga 2 O 3 nanorod array/FTO. After Al doping, a slightly blue-shift property of the Al-doped α-Ga 2 O 3 nanorod array in the solar-blind UV region has been observed in Figure 2b. Based on the absorption properties, optical bandgaps can be estimated by extrapolating the straight-line portion to the hν axis near absorption edge, as shown in the inset of Figure  The XRD measurements of an undoped α-Ga2O3 nanorod array/FTO and Al-doped α-Ga2O3 nanorod array/FTO were investigated to reveal the crystal structure, as shown in Figure 2a. The FTO substrate shows the diffraction peaks at ~26.6°, ~34°, ~37.9°, ~51.7°, ~54.7°, ~61.8° and ~65.9°, which can be attributed to be the SnO2 diffraction of (110), (101), (200), (211), (220), (310) and (301) faces, respectively. For the undoped α-Ga2O3 nanorod array/FTO, ~36.21°, ~50.35, ~63.69° and ~64.99° correspond to the (110), (024), (214) and (300) face diffractions of α-Ga2O3, respectively. Moreover, Al-doped α-Ga2O3 diffraction peaks are located at ~36.25°, ~50.45°, ~63.71°and ~65.07°, which also correspond to the (110), (024), (214), and (300) face diffractions of α-Ga2O3 [18,19], respectively. Owing to the smaller ionic radius of Al 3+ (0.0535 nm) than that of Ga 3+ (0.062 nm), the corresponding diffraction peaks of the Al-doped α-Ga2O3 nanorod array slightly shift to a larger angle in comparison with the undoped α-Ga2O3 nanorod array [18,19,24]. Figure 2b is the UV-vis diffuse reflectance absorption spectra of the undoped α-Ga2O3 nanorod array/FTO and Al-doped α-Ga2O3 nanorod array/FTO. After Al doping, a slightly blue-shift property of the Al-doped α-Ga2O3 nanorod array in the solar-blind UV region has been observed in Figure 2b. Based on the absorption properties, optical bandgaps can be estimated by extrapolating the straight-line portion to the hν axis near absorption edge, as shown in the inset of Figure 2b. The optical bandgap of the undoped α-Ga2O3 nanorod array is around ~4.6 eV, while the optical bandgap of the Al-doped α-Ga2O3 nanorod array is around ~4.8 eV. Due to the large bandgap of Al2O3 than that of Ga2O3, the slightly enlarged optical bandgap may originate from the incorporation of Al into α-Ga2O3. Figure 3a schematically illustrates the measurement setup during UV photodetection, where the undoped α-Ga 2 O 3 nanorod array/FTO or Al-doped α-Ga 2 O 3 nanorod array/FTO were selected as a photoanode in the shown PEC cell. Based on an electrochemical workstation, the three-electrode system additionally consists of Pt (counter electrode) ang Ag/AgCl (reference electrode). A total of 0.5 M Na 2 SO 4 aqueous solution was used as the electrolyte. UV light was illuminated on the surface of the undoped α-Ga 2 O 3 nanorod array/FTO or Al-doped α-Ga 2 O 3 nanorod array/FTO, to evaluate the photoresponse properties. When the spectral responsivity was collected, the undoped α-Ga 2 O 3 nanorod array/FTO or Al-doped α-Ga 2 O 3 nanorod array/FTO, and Pt electrodes were connected with a lock-in amplifier. The detailed mechanism will be explained in the working mechanism part. Figure 3c shows solar-blind UV photoresponse spectra at 0 V (vs. Ag/AgCl). The peak wavelength is~270 nm, with a peak responsivity of~0.25 mA/W for the undoped α-Ga 2 O 3 nanorod array/FTO device, while the peak wavelength is 260 nm with a peak responsivity of~1.46 mA/W for the Al-doped α-Ga 2 O 3 nanorod array/FTO device. Because of the larger optical bandgap of the Al-doped α-Ga 2 O 3 nanorod array, the peak photoresponse shifts to the shorter wavelength, which is consistent with the slight blue-shift property of the Al-doped α-Ga 2 O 3 nanorod array in the solar-blind UV region in Figure 2b. And the enhanced responsivity is also related to Al doping. According to the literature, the photoresponse of Ga 2 O 3 can be improved after Al doping, although further investigations should be conducted to reveal the reason [12][13][14][15][16]. Herein, the Al-doped α-Ga 2 O 3 nanorod array is also~5.84 times larger in peak responsivity than the undoped α-Ga 2 O 3 nanorod array at 0V (vs. Ag/AgCl). The inset of Figure 3c shows the peak responsivity of the undoped α-Ga 2 O 3 nanorod array/FTO device and Al-doped α-Ga 2 O 3 nanorod array/FTO device varying with the applied voltages (vs. Ag/AgCl). The peak responsivity increases from~0.25 mA/W at 0 V to 1 mA/W at 0.9 V for the undoped α-Ga 2 O 3 nanorod array/FTO device. Obviously, the peak responsivity increases from~1.46 mA/W at 0 V to 2.02 mA/W at 1 V for the Al-doped α-Ga 2 O 3 nanorod array/FTO device, which is much larger than that of the undoped device. Based on the responsivity and dark current, the quantum efficiency and detectivity of the Al-doped α-Ga 2 O 3 nanorod array/FTO device can be roughly evaluated [5]. Figure 3d shows the voltage dependences of the calculated quantum efficiency and detectivity in details. The quantum efficiency varies from~0.7% at 0 V (vs. Ag/AgCl) to~0.96% at 1 V (vs. Ag/AgCl), while the detectivity increases from 6 × 10 9 Jones at 0 V (vs. Ag/AgCl) to~1 × 10 10 Jones at 1 V (vs. Ag/AgCl). The enhanced responsivity, quantum efficiency, and detectivity are ascribed from the wider depletion region when positive voltages are added to the FTO electrode.
As shown in Figure 4a, when the photoanode in a PEC cell was 0 V (vs. Ag/AgCl), we, respectively, measured the time-dependent current curves of an undoped α-Ga 2 O 3 nanorod array/FTO under 270 nm light illumination (0.47 mW/cm 2 ) and Al-doped α-Ga 2 O 3 nanorod array/FTO under 260 nm light illumination (0.31 mW/cm 2 ), with 10 s on and 10 s off. Thus, our device shows good stability and repetition at the self-powered photodetection mode in 5 cycles. And the steady light current of the Al-doped α-Ga 2 O 3 nanorod array/FTO is also much larger than that of the undoped α-Ga 2 O 3 nanorod array/FTO. The rise processes of the undoped α-Ga 2 O 3 nanorod array/FTO are clearly longer than those of the Al-doped α-Ga 2 O 3 nanorod array/FTO. When the shutter is off, the light current immediately drops to the initial level for both of the devices. To further evaluate long-period stability and repetition, we also measured a time-dependent current curve of the Al-doped α-Ga 2 O 3 nanorod array/FTO under 260 nm light illumination (0.39 mW/cm 2 ), with 10 s on and 10 s off, at the self-powered photodetection mode in 50 cycles. As shown in Figure 4b, the steady light current of the Al-doped α-Ga 2 O 3 nanorod array/FTO nearly remains the same value at 0 V (vs. Ag/AgCl) under 260 nm illumination (0.39 mW/cm 2 ). By the following equation: I(t) = I s + Ie −t/τ (where I(t) is the light current decay, I s is the steady current, I is the photocurrent, and τ is the rise time or decay time), the rise time and decay time can be fitted during a rise process and a decay process, respectively, in Figure 4a. Figure  Figure 3a schematically illustrates the measurement setup during UV photodetection, where the undoped α-Ga2O3 nanorod array/FTO or Al-doped α-Ga2O3 nanorod array/FTO were selected as a photoanode in the shown PEC cell. Based on an electrochemical workstation, the three-electrode system additionally consists of Pt (counter electrode) ang Ag/AgCl (reference electrode). A total of 0.5 M Na2SO4 aqueous solution was used as the electrolyte. UV light was illuminated on the surface of the undoped α-Ga2O3 nanorod array/FTO or Al-doped α-Ga2O3 nanorod array/FTO, to evaluate the photoresponse properties. When the spectral responsivity was collected, the undoped α-Ga2O3 nanorod array/FTO or Al-doped α-Ga2O3 nanorod array/FTO, and Pt electrodes were connected with a lock-in amplifier. Figure 3a schematically illustrates the measurement setup during UV photodetection, where the undoped α-Ga2O3 nanorod array/FTO or Al-doped α-Ga2O3 nanorod array/FTO were selected as a photoanode in the shown PEC cell. Based on an electrochemical workstation, the three-electrode system additionally consists of Pt (counter electrode) ang Ag/AgCl (reference electrode). A total of 0.5 M Na2SO4 aqueous solution was used as the electrolyte. UV light was illuminated on the surface of the undoped α-Ga2O3 nanorod array/FTO or Al-doped α-Ga2O3 nanorod array/FTO, to evaluate the photoresponse properties. When the spectral responsivity was collected, the undoped α-Ga2O3 nanorod array/FTO or Al-doped α-Ga2O3 nanorod array/FTO, and Pt electrodes were connected with a lock-in amplifier.
(a) (b)  nm illumination for the Al-doped α-Ga2O3 nanorod array/FTO device. Considering the undoped α-Ga2O3 nanorod array/FTO and Al-doped α-Ga2O3 nanorod array/FTO structure as photoanodes, the light current shifting towards the positive direction compared with the negative dark current indicates that positive photocurrent has been generated. Thus, the photogenerated electrons flows from undoped α-Ga2O3 nanorod array or Aldoped α-Ga2O3 nanorod array to FTO in the undoped α-Ga2O3 nanorod array/FTO or Aldoped α-Ga2O3 nanorod array/FTO device at 0V (vs. Ag/AgCl). The detailed mechanism will be explained in the working mechanism part. Figure 3c shows solar-blind UV photoresponse spectra at 0V (vs. Ag/AgCl). The peak wavelength is ~270 nm, with a peak responsivity of ~0.25 mA/W for the undoped α-Ga2O3 nanorod array/FTO device, while the peak wavelength is 260 nm with a peak responsivity of ~1.46 mA/W for the Al-doped α-Ga2O3 nanorod array/FTO device. Because of the larger optical bandgap of the Al-doped α-Ga2O3 nanorod array, the peak photoresponse shifts to the shorter wavelength, which To further explain the working mechanism, Figure 5 schematically depicts energy band diagrams of the undoped α-Ga 2 O 3 nanorod array/FTO device or Al-doped α-Ga 2 O 3 nanorod array/FTO device as a photoanode in the PEC cell. In the dark, a Schottky junction is produced between the nanorod array and Na 2 SO 4 electrolyte when a nanorod array on FTO is immersed into a Na 2 SO 4 aqueous solution in Figure 5a [20]. Thus, a depletion region is generated at the surface of the nanorod array. Owing to the nanorod array/FTO structure selected as the photoanode; positive biases are added to FTO when the PEC cell works as shown in Figure 5a. Therefore, the forward current in our device is just the reverse current as in the traditional Schottky junction. However, photogenerated electron-hole pairs are produced in the depletion region of the undoped α-Ga 2 O 3 nanorod array or Al-doped α-Ga 2 O 3 nanorod array under solar-blind UV light illumination in Figure 5b,c. Under the drift of the built-in electric field, photogenerated electrons move from the undoped α-Ga 2 O 3 nanorod array or Al-doped α-Ga 2 O 3 nanorod array to FTO, while photogenerated holes moves from the undoped α-Ga 2 O 3 nanorod array or Al-doped α-Ga 2 O 3 nanorod array to the electrolyte. Thus, a large positive photocurrent is generated under solar-blind UV light illumination, the direction of which is the same transportation direction as that of the forward dark current. Thus, the light current shifting towards a positive direction has been observed for both the undoped α-Ga 2 O 3 nanorod array/FTO device and Al-doped α-Ga 2 O 3 nanorod array/FTO devices in Figures 3 and 4. It is noteworthy that the larger bandgap of the Al-doped α-Ga 2 O 3 may lift its conduction edge as the incorporation of MgO into ZnO [25]. Therefore, a larger Schottky barrier may exist in the Al-doped α-Ga 2 O 3 nanorod array/FTO PEC cell than that in the undoped α-Ga 2 O 3 nanorod array/FTO PEC cell. As a result, a wider depletion region in the Al-doped α-Ga 2 O 3 nanorod array/FTO PEC cell can improve the separation of more photogenerated electron-hole pairs, and contributes to a larger photocurrent and responsivity. Therefore, the larger photocurrent and responsivity have been demonstrated through doping Al into the α-Ga 2 O 3 nanorod array. Generally, the depletion region of a Schottky junction can be enhanced under reverse biases. Considering the undoped α-Ga 2 O 3 nanorod array/FTO or Al-doped α-Ga 2 O 3 nanorod array/FTO as the photoanode in the PEC cell, such biases are the forward biases as indicated in Figure 5a. As a result, the depletion region becomes wider in Figure 5c under positive biases than that at 0 V in Figure 5b. Therefore, larger photocurrent is formed under positive biases, which is also the reason of the improved responsivity, quantum efficiency, and detectivity when positive voltages are added to the FTO electrode, as shown in Figure 3c,d.  To further explain the working mechanism, Figure 5 schematically depicts energy band diagrams of the undoped α-Ga2O3 nanorod array/FTO device or Al-doped α-Ga2O3 nanorod array/FTO device as a photoanode in the PEC cell. In the dark, a Schottky junction is produced between the nanorod array and Na2SO4 electrolyte when a nanorod array ward biases as indicated in Figure 5a. As a result, the depletion region becomes wider in Figure 5c under positive biases than that at 0 V in Figure 5b. Therefore larger photocurrent is formed under positive biases, which is also the reason of the improved responsivity, quantum efficiency, and detectivity when positive voltages are added to the FTO electrode, as shown in Figure 3c,d.

Conclusions
In this work, using a simple and cheap processes, including hydrothermal and postannealing, we realized an Al-doped α-Ga2O3 nanorod array on FTO. To the best of our knowledge, it is the first time that Al-doped α-Ga2O3 nanorod array has been fabricated on FTO via a much simpler and cheaper way than that based on MOCVD, magnetron sputtering, MBE and PLD. When the Al-doped α-Ga2O3 nanorod array/FTO structure is served as a photoanode in a PEC cell, the Al-doped α-Ga2O3 nanorod array/FTO photo-

Conclusions
In this work, using a simple and cheap processes, including hydrothermal and postannealing, we realized an Al-doped α-Ga 2 O 3 nanorod array on FTO. To the best of our knowledge, it is the first time that Al-doped α-Ga 2 O 3 nanorod array has been fabricated on FTO via a much simpler and cheaper way than that based on MOCVD, magnetron sputtering, MBE and PLD. When the Al-doped α-Ga 2 O 3 nanorod array/FTO structure is served as a photoanode in a PEC cell, the Al-doped α-Ga 2 O 3 nanorod array/FTO photodetector had the self-powered photodetection properties at 0 V (vs. Ag/AgCl) with a peak responsivity of~1.46 mA/W at 260 nm. The rise time was 0.421 s, and the decay time was 0.139 s, under solar-blind UV (260 nm) illumination. Compared with the undoped α-Ga 2 O 3 nanorod array/FTO device, the peak responsivity of the Al-doped device was 5.84 times larger, and the response speed was relatively faster. The reason is that a larger Schottky barrier and wider depletion region may exist in the Al-doped α-Ga 2 O 3 nanorod array/FTO PEC cell than that in the undoped α-Ga 2 O 3 nanorod array/FTO PEC cell. As a result, the wider depletion region in the Al-doped α-Ga 2 O 3 nanorod array/FTO PEC cell can improve the separation of more photogenerated electron-hole pairs, contributing to a larger photocurrent and responsivity. The peak responsivity of the Al-doped α-Ga 2 O 3 nanorod array/FTO device increased to 2.02 mA/W at 1 V (vs. Ag/AgCl). Additionally, with the increase of the positive biases, the enlarged depletion region contributes to the enhanced responsivity, quantum efficiency (from~0.7% at 0 V (vs. Ag/AgCl) to~0.96% at 1 V (vs. Ag/AgCl)), and detectivity (from~6 × 10 9 Jones at 0 V (vs. Ag/AgCl) to~1 × 10 10 Jones at 1 V (vs. Ag/AgCl)). Therefore, doping Al into α-Ga 2 O 3 may provide a route to enhance the self-powered photodetection performances of α-Ga 2 O 3 nanorod arrays.