1. Introduction
Ga
2O
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, space-to-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
2O
3-based photodetectors, constructing Schottky junctions, p–n junctions, or n–n junctions are still the most feasible ways to realize self-powered photodetection.
Currently, element doping, such as Si, Ge, Sn, In, or Al doping, is an important technique to realize the more multifunctional applications mentioned above, due to its controllable technique of modulating the electrical and optical properties [
8,
9,
10,
11,
12,
13,
14,
15,
16,
17]. Unlike other elements, In or Al doping can more effectively realize an extension of the bandgap [
8,
9,
10,
11,
12,
13,
14,
15,
16]. For example, the bandgap of inherent Ga
2O
3 can be extended into smaller values by alloying In with it [
8,
9,
10,
11], and can also be expanded into larger values by alloying Al with it [
12,
13,
14,
15,
16]. Therefore, Al doping is more important for Ga
2O
3 to be applied in solar-bind UV photodetection, due to the fact that its solar-blind UV absorption characterization can be maintained. S.-H. Yuan et al. reported that the peak responsivity of an MSM photodetector at 230 nm, under a bias voltage of 5 V, can reach 53.61 times greater than that of a photodetector without doping any Al content [
13]. And Q. Feng et al. obtained a 10-times-higher photocurrent for an (Al
0.12Ga
0.88)
2O
3 MSM device, compared to an undoped photodetector [
12]. Moreover, a graphene/(AlGa)
2O
3/GaN self-powered device also showed excellent photodetection properties, with a peak responsivity of ∼20 mA/W, a rise time of ∼2 μs, and a decay time of ∼10 ms [
14]. Therefore, Al-doped Ga
2O
3 can bring an improvement in photoresponsivity, both in a non-self-powered photodetection mode and in a self-powered photodetection mode [
12,
13,
14,
15,
16]. According to the literature, many of the recent efforts to dope Al into Ga
2O
3 described above were performed using metal–organic chemical vapor deposition (MOCVD), magnetron sputtering, molecular beam epitaxy (MBE), and pulsed laser deposition (PLD) [
12,
13,
14,
15,
16]. Obviously, these methods are expensive and complicated. However, an α-Ga
2O
3 nanorod array can be easily grown on FTO using the hydrothermal and post-annealing processes, which have the advantages of being more simple and much cheaper procedures [
18,
19,
20]. To the best of our knowledge, such processes have not been applied for Al-doped α-Ga
2O
3 nanorod arrays. More importantly, α-Ga
2O
3 nanorod array/FTO can be employed as the basic structure for realizing self-powered photodetection, due to the fact that a Schottky junction can be formed at the surface between the α-Ga
2O
3 nanorod array and the electrolyte. This is a famous photodetection mode, when the α-Ga
2O
3 nanorod array/FTO structure is immersed into the electrolyte in a photoelectrochemical (PEC) cell [
18,
19,
20,
21,
22,
23]. Therefore, we used an Al-doped α-Ga
2O
3 nanorod array/FTO as the photoanode in a PEC cell to demonstrate the self-powered photodetection properties.
Herein, an Al-doped α-Ga2O3 nanorod array on FTO was achieved using the hydrothermal and post-annealing processes. It is the first time that an Al-doped α-Ga2O3 nanorod array on FTO has been realized using a much simpler and cheaper way. Based on a PEC cell, a self-powered Al-doped α-Ga2O3 nanorod array/FTO photodetector was demonstrated as the photoanode, with a peak responsivity of 1.46 mA/W at 260 nm at 0 V (vs. Ag/AgCl). Moreover, the device had a response speed of 0.421 s for rise time and 0.139 s for decay time under solar-blind UV (260 nm) illumination. The responsivity of the Al-doped device was ~5.84 times larger, and the response speed was relatively faster than the performance of an undoped device, which is due to the fact that a larger Schottky barrier and wider depletion region may exist in the Al-doped α-Ga2O3 nanorod array/FTO device than that in the undoped α-Ga2O3 nanorod array/FTO PEC device. Owing to the enlarged depletion region resulting from increasing the positive biases from 0 V to 1 V, the responsivity, quantum efficiency, and detectivity of the Al-doped device were respectively 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.
3. Results and Discussion
Figure 1a,b, respectively, shows a surface morphology picture of an undoped α-Ga
2O
3 nanorod array/FTO and Al-doped α-Ga
2O
3 nanorod array/FTO structure by SEM. From the surface morphology images, the Al-doped α-Ga
2O
3 nanorod in
Figure 1b also exhibits a diamond-like shape, which is similar to the α-Ga
2O
3 nanorod array without doping, as shown in
Figure 1a. Therefore, Al doping did not obviously change the surface morphology of the α-Ga
2O
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
2O
3 nanorod array which, respectively, stand for the O and Ga elements. For the Al-doped α-Ga
2O
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
2O
3 nanorod array. To further evaluate the element distributions in the Al-doped α-Ga
2O
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
2O
3 nanorod array, in detail. From
Figure 1d–f, O, Ga, and Al elements are uniformly distributed in the Al-doped α-Ga
2O
3 nanorod array. As a result, Al was successfully and uniformly doped into α-Ga
2O
3 nanorods.
The XRD measurements of an undoped α-Ga
2O
3 nanorod array/FTO and Al-doped α-Ga
2O
3 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 SnO
2 diffraction of (110), (101), (200), (211), (220), (310) and (301) faces, respectively. For the undoped α-Ga
2O
3 nanorod array/FTO, ~36.21°, ~50.35, ~63.69° and ~64.99° correspond to the (110), (024), (214) and (300) face diffractions of α-Ga
2O
3, respectively. Moreover, Al-doped α-Ga
2O
3 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-Ga
2O
3 [
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
2O
3 nanorod array slightly shift to a larger angle in comparison with the undoped α-Ga
2O
3 nanorod array [
18,
19,
24].
Figure 2b is the UV-vis diffuse reflectance absorption spectra of the undoped α-Ga
2O
3 nanorod array/FTO and Al-doped α-Ga
2O
3 nanorod array/FTO. After Al doping, a slightly blue-shift property of the Al-doped α-Ga
2O
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 2b. The optical bandgap of the undoped α-Ga
2O
3 nanorod array is around ~4.6 eV, while the optical bandgap of the Al-doped α-Ga
2O
3 nanorod array is around ~4.8 eV. Due to the large bandgap of Al
2O
3 than that of Ga
2O
3, the slightly enlarged optical bandgap may originate from the incorporation of Al into α-Ga
2O
3.
Figure 3a schematically illustrates the measurement setup during UV photodetection, where the undoped α-Ga
2O
3 nanorod array/FTO or Al-doped α-Ga
2O
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
2SO
4 aqueous solution was used as the electrolyte. UV light was illuminated on the surface of the undoped α-Ga
2O
3 nanorod array/FTO or Al-doped α-Ga
2O
3 nanorod array/FTO, to evaluate the photoresponse properties. When the spectral responsivity was collected, the undoped α-Ga
2O
3 nanorod array/FTO or Al-doped α-Ga
2O
3 nanorod array/FTO, and Pt electrodes were connected with a lock-in amplifier.
Figure 3b shows the dark current and light current of an undoped α-Ga
2O
3 nanorod array/FTO device under 270 nm (0.47 mW/cm
2) UV light illumination, and the dark and light currents under 255 nm (0.24 mW/cm
2), 260 nm (0.31 mW/cm
2), and 265 nm (0.39 mW/cm
2) UV light illumination of an Al-doped α-Ga
2O
3 nanorod array/FTO device. The light current increases from −0.17 mA to −0.08 mA at 0 V (vs. Ag/AgCl) under 270 nm illumination for the undoped α-Ga
2O
3 nanorod array/FTO device. Obviously, the light current of the Al-doped α-Ga
2O
3 nanorod array/FTO device increases from −0.18 mA to 0.34 mA at 0 V (vs. Ag/AgCl) under 255 nm illumination, from −0.18 mA to 0.35 mA at 0 V (vs. Ag/AgCl) under 260 nm illumination, and from −0.18 mA to 0.33 mA at 0 V (vs. Ag/AgCl) under 265 nm illumination. And larger net photocurrent was realized after Al was doped into the α-Ga
2O
3 nanorod array. The net short-circuit photocurrent produced at 0 V (vs. Ag/AgCl) indicates that our devices can realize the self-powered photodetection for both the undoped α-Ga
2O
3 nanorod array/FTO device and the Al-doped α-Ga
2O
3 nanorod array/FTO device. Additionally, the open-circuit voltage is estimated to be ~−0.3 V under 255 nm illumination, ~−0.33 V under 260 nm illumination, and ~−0.33 V under 265 nm illumination for the Al-doped α-Ga
2O
3 nanorod array/FTO device. Considering the undoped α-Ga
2O
3 nanorod array/FTO and Al-doped α-Ga
2O
3 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-Ga
2O
3 nanorod array or Al-doped α-Ga
2O
3 nanorod array to FTO in the undoped α-Ga
2O
3 nanorod array/FTO or Al-doped α-Ga
2O
3 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 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
2O
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
2O
3 nanorod array/FTO device. Because of the larger optical bandgap of the Al-doped α-Ga
2O
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
2O
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
2O
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
2O
3 nanorod array is also ~5.84 times larger in peak responsivity than the undoped α-Ga
2O
3 nanorod array at 0 V (vs. Ag/AgCl). The inset of
Figure 3c shows the peak responsivity of the undoped α-Ga
2O
3 nanorod array/FTO device and Al-doped α-Ga
2O
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
2O
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
2O
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
2O
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
2O
3 nanorod array/FTO under 270 nm light illumination (0.47 mW/cm
2) and Al-doped α-Ga
2O
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
2O
3 nanorod array/FTO is also much larger than that of the undoped α-Ga
2O
3 nanorod array/FTO. The rise processes of the undoped α-Ga
2O
3 nanorod array/FTO are clearly longer than those of the Al-doped α-Ga
2O
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
2O
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
2O
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) =
Is +
Ie−t/τ (where
I(
t) is the light current decay,
Is 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 4c is the transient photoresponse of the undoped α-Ga
2O
3 nanorod array/FTO under 270 nm light illumination (0.47 mW/cm
2) and Al-doped α-Ga
2O
3 nanorod array/FTO under 260 nm light illumination (0.31 mW/cm
2) at 0 V (vs. Ag/AgCl) during rise processes. From
Figure 4c, the rise time of the undoped α-Ga
2O
3 nanorod array/FTO device and Al-doped α-Ga
2O
3 nanorod array/FTO device is, respectively, 1.560 s and 0.421 s. The longer rise time is consistent with the slower rise process in
Figure 4a.
Figure 4d is the corresponding transient photoresponse curves during decay processes in
Figure 4a. The decay time of the undoped α-Ga
2O
3 nanorod array/FTO device and Al-doped α-Ga
2O
3 nanorod array/FTO device is fitted to be 0.141 s and 0.139 s, respectively. Unlike the much faster rise time of the Al-doped α-Ga
2O
3 nanorod array/FTO device (
Figure 4c), the decay time is nearly the same for both of the devices.
To further explain the working mechanism,
Figure 5 schematically depicts energy band diagrams of the undoped α-Ga
2O
3 nanorod array/FTO device or Al-doped α-Ga
2O
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
2SO
4 electrolyte when a nanorod array on FTO is immersed into a Na
2SO
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
2O
3 nanorod array or Al-doped α-Ga
2O
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
2O
3 nanorod array or Al-doped α-Ga
2O
3 nanorod array to FTO, while photogenerated holes moves from the undoped α-Ga
2O
3 nanorod array or Al-doped α-Ga
2O
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
2O
3 nanorod array/FTO device and Al-doped α-Ga
2O
3 nanorod array/FTO devices in
Figure 3 and
Figure 4. It is noteworthy that the larger bandgap of the Al-doped α-Ga
2O
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
2O
3 nanorod array/FTO PEC cell than that in the undoped α-Ga
2O
3 nanorod array/FTO PEC cell. As a result, a wider depletion region in the Al-doped α-Ga
2O
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
2O
3 nanorod array. Generally, the depletion region of a Schottky junction can be enhanced under reverse biases. Considering the undoped α-Ga
2O
3 nanorod array/FTO or Al-doped α-Ga
2O
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.