Voltage-Tunable UVC–UVB Dual-Band Metal–Semiconductor–Metal Photodetector Based on Ga 2 O 3 / MgZnO Heterostructure by RF Sputtering

: Dual-band metal–semiconductor–metal (MSM) photodetectors (PDs) with a Ga 2 O 3 / MgZnO heterostructure were fabricated by radio frequency (RF) sputtering, which can detect ultraviolet C (UVC) and ultraviolet B (UVB) bands individually by controlling di ﬀ erent bias voltages. A PD with the annealing temperature of Ga 2 O 3 at 600 ◦ C can improve the crystal quality of Ga 2 O 3 thin ﬁlm and exhibit the least persistent photoconductivity (PPC) e ﬀ ect. However, a PD with the annealing temperature of Ga 2 O 3 at 600 ◦ C cannot achieve a voltage-tunable dual-band characteristic. On the contrary, the PD without annealing can suppress the carriers from the bottom layer of MgZnO thin ﬁlm at a lower bias voltage of 1 V. At this time, the peak responsivity at 250 nm was mainly dominated by the top layer of Ga 2 O 3 thin ﬁlm. Then, as the bias voltage increased to 5 V, the peak detection wavelength shifted from 250 (UVC) to 320 nm (UVB). In addition, the PD with a 25 nm–thick SiO 2 layer inserted between Ga 2 O 3 and MgZnO thin ﬁlm can achieve a broader operating bias voltage range for dual-band applications.


Introduction
For solid-state lighting, ultraviolet (UV) light has unique applications compared with visible light. UV light has been widely used in disinfection, biotechnology, medical science, military, and space technology. According to the different applications, UV light can be divided into three sub-bands: ultraviolet A (UVA) (320-400 nm), ultraviolet B (UVB) (280-320 nm), and ultraviolet C (UVC) (200-280 nm). Wide-bandgap semiconductor materials, such as Ga 2 O 3 , BeZnO, ZnO, Sr 2 Nb 3 O 10 , Al x Ga 1−x N, and Mg x Zn 1−x O, have been generally used to design and fabricate UV photodetectors (PDs) [1][2][3][4][5][6][7][8][9][10][11]. Al x Ga 1−x N based on GaN series materials has been developed for decades and is widely used in near-UV PDs. However, for deep-UV PDs, the Al composition of AlGaN is generally over 35% to modulate the detection wavelength of PDs. This higher Al composition in AlGaN can induce significant leakage current from defects through the path [12][13][14]. Mg x Zn 1−x O, a representative ternary alloy of wide-bandgap semiconductor material, has gained popularity for fabricating UVB and UVC optoelectronic devices due to its energy bandgap, ranging from 3.37 (ZnO) to 7.8 eV (MgO) [15][16][17][18]. However, the crystal structure of Mg x Zn 1−x O can be changed from wurtzite to cubic as the Mg composition is increased. Therefore, it is challenging to grow a single crystalline Mg x Zn 1−x O, especially for wurtzite Mg x Zn 1−x O with more than 36% Mg composition [19]. A Ga 2 O 3 material with an intrinsic wide-bandgap range of 4.2-4.9 eV has been considered as an ideal candidate for fabricating UVC PDs, especially in solar-blind applications [20][21][22][23][24][25][26][27][28][29][30].
The Ga 2 O 3 series heterojunction UV PDs have been previously analyzed. Hung et al. reported a Ga 2 O 3 /AlGaN/GaN heterostructure UV three-band PD, and the absorbed wavelength ranges from UVA to UVC [31]. Nakagomi et al. fabricated a deep-UV PD based on the β-Ga 2 O 3 /GaN heterojunction, and the highest responsivity was located at 240 nm [32]. Zhao et al. designed a solar-blind PD based on a ZnO/Ga 2 O 3 core-shell heterostructured microwire that exhibited high responsivity at 0 V and different absorbed wavelengths, at 0, −2, and 2 V [20]. Despite these findings, the ZnO series material was chosen in studies based on its advantage of a simple process. However, the studies on Ga 2 O 3 /MgZnO heterojunction UV PDs are still unclear. Furthermore, the devices used in the previous studies had a dual-band absorbed wavelength; however, the central absorbed wavelength of PDs could not be changed while the bias voltage was varied. In another previous study, MgZnO/SiO 2 /ZnO heterojunction dual-band UV PDs were successfully grown and fabricated, allowing users to choose the central absorption wavelength from UVB to UVA at different operating bias voltages, and the central absorption wavelength shift can achieve a value of 55 nm [33]. In this study, Ga 2 O 3 /MgZnO heterojunction dual-band UVC and UVB PDs were manufactured by RF sputtering, and the impact of the Ga 2 O 3 annealing temperatures was investigated. In addition, the influence of the SiO 2 insertion layer on the Ga 2 O 3 and MgZnO thin films was also investigated.

Materials and Methods
In this study, acetone, isopropanol, and deionized water were successively used to clean the sapphire substrate using an ultrasonic cleaner (Delta-D150, Taiwan) for 30 min. After the typical cleaning process, a 150 nm-thick MgZnO thin film was grown on the c-plane sapphire substrate, using an RF magnetron sputter system (KD-SPUTTER, Kao Duen Technology Corp., New Taipei City, Taiwan). The MgZnO target had an Mg content of 20%. Ar and oxygen (O 2 ) gases were used as sputtering gases. The Ar flow rate, O 2 flow rate, chamber pressure, and RF power were maintained at 25 sccm, 3 sccm, 10 mTorr, and 80 W, respectively. After growing the MgZnO thin film, the samples were annealed at 700 • C, in a tube furnace (ADVANCE RIKO, RHL-P series P610CP, Yokohama, Japan), in air, for 30 min. Then, parts of the samples were divided to grow a 25 nm-thick SiO 2 on an MgZnO thin film at 300 • C by using a plasma-enhanced chemical vapor deposition (PECVD) system (Oxford Plasmalab System 100, Bristol, UK). This was followed by growing a 370 nm-thick Ga 2 O 3 thin film on the MgZnO or SiO 2 thin film by an RF magnetron sputter system, at room temperature. The Ar flow rate, O 2 flow rate, chamber pressure, and RF power were 98 sccm, 2 sccm, 5 mTorr, and 80 W, respectively. Finally, the Ga 2 O 3 /MgZnO heterojunction structure was annealed at 600 or 800 • C. Figure 1 presents all the proposed designs of Ga 2 O 3 /MgZnO heterojunction PDs. The following two types of structures were used in this study: Ga 2 O 3 with or without annealing temperatures at 600 or 800 • C in air for 30 min (labeled as PD1, PD2, and PD3), and an unannealed Ga 2 O 3 thin film with a 25 nm-thick SiO 2 insertion layer (labeled as PD4). The devices with a metal-semiconductor-metal (MSM) structure consisting of two interdigitated contact electrodes. The length, width, and finger space were 1000, 100, and 200 µm, respectively. A Ti/Au (25 nm/180 nm) metal was subsequently deposited as electrodes, using an electron beam evaporator. Figure 2 presents the cross-sectional line scan of the energy-dispersive X-ray spectroscopy (EDX) analysis for Ga 2 O 3 /MgZnO heterojunction UV PDs with and without a 25 nm-thick SiO 2 insertion layer. Different stacked materials, including Ti, Au, Ga 2 O 3 , MgZnO, Al 2 O 3 , and SiO 2 , were observed. The details of the growth parameter are shown in Figure 1.    Figure 3 presents the photo-and dark-current characteristics of the fabricated MSM PDs, with and without different annealing temperatures of the Ga2O3 thin film, which were measured by using a 300 W Xe arc lamp as a light source. It was found that the dark currents of PD1, PD2, and PD3 were 1.72 × 10 −6 , 3.15 × 10 −8 , and 8.81 × 10 −4 A at 5 V, respectively. The PD2 with an annealing temperature of 600 °C exhibited the lowest dark current, which is attributed to the improvement of the thin film's     Figure 3 presents the photo-and dark-current characteristics of the fabricated MSM PDs, with and without different annealing temperatures of the Ga2O3 thin film, which were measured by using a 300 W Xe arc lamp as a light source. It was found that the dark currents of PD1, PD2, and PD3 were 1.72 × 10 −6 , 3.15 × 10 −8 , and 8.81 × 10 −4 A at 5 V, respectively. The PD2 with an annealing temperature of 600 °C exhibited the lowest dark current, which is attributed to the improvement of the thin film's The dark current characteristics of these PDs were analyzed by using an semiconductor parameter analyzer (Agilent HP-4156C, Santa Clara, CA, USA) with a cascade micro-chamber and cascade DCP 100 series low-noise electrical performance probe. An analysis of the spectral responsivity measurements by a Monochromator (HORIBA/JOBIN YVON/SPEX TRIAX 320 system, Kyoto, Japan) was also performed with a 300 W Xe arc lamp as a light source and a standard synchronous detection scheme.  Figure 3 presents the photo-and dark-current characteristics of the fabricated MSM PDs, with and without different annealing temperatures of the Ga 2 O 3 thin film, which were measured by using a 300 W Xe arc lamp as a light source. It was found that the dark currents of PD1, PD2, and PD3 were 1.72 × 10 −6 , 3.15 × 10 −8 , and 8.81 × 10 −4 A at 5 V, respectively. The PD2 with an annealing temperature of 600 • C exhibited the lowest dark current, which is attributed to the improvement of the thin film's crystal quality after annealing. The grain size of the Ga 2 O 3 thin film increased as the annealing temperature increased for scanning electron microscopy (not shown here). The Ga 2 O 3 thin film deposited by sputtering was originally in an amorphous state. After the annealing process, the amorphous state can change to the polymorphic state, which indicates that more crystal structures can reorganize. Therefore, the number of grain boundaries can decrease as the annealing temperature increases, that is, the dark current of the sample can be reduced at high annealing temperatures due to the reduction of defects and healing of leakage current paths. However, the dark current apparently increased for PD3 with an annealing temperature of Ga 2 O 3 at 800 • C. Ju et al. reported that MgZnO thin films exhibited better crystal quality at an annealing temperature of 750 • C. When the annealing temperature of the MgZnO thin film was raised to 850 • C, the full-width half-maximum (FWHM) of the X-ray diffraction (XRD) signal rapidly increased owing to phase separation [34]. In this study, the annealing temperature of the MgZnO thin film was 700 • C, to obtain a high crystal quality before the growth of the Ga 2 O 3 thin film. However, the Ga 2 O 3 thin film grown on the MgZnO thin film was subsequently annealed at a temperature of 800 • C, which can result in the destruction of the MgZnO crystalline structure. In addition, Chikoidze et al. proposed that the annealing temperature of Ga 2 O 3 thin films above 700 • C in air or oxygen ambient would lead to the creation of oxygen vacancies (V o ) due to the out-diffusion of oxygen atoms from the Ga 2 O 3 thin film. This generation of V o would lead to an additional source of free electrons and an n-type conductivity enhancement [35]. This is why PD3 with an annealing temperature of Ga 2 O 3 at 800 • C had a reverse effect on the dark-current characteristic. In contrast, the photo-to-dark-current ratios (PDCRs) of PD1, PD2, and PD3 at 5 V were 30.75, 132.18, and 1.33, respectively. The PD2 with an annealing temperature of Ga 2 O 3 at 600 • C demonstrated the highest PDCR, owing to the improvement of the Ga 2 O 3 thin film quality and maintenance of the MgZnO crystalline quality at the same time.

The Photo and Dark Current
Coatings 2020, 10, x FOR PEER REVIEW 4 of 10 crystal quality after annealing. The grain size of the Ga2O3 thin film increased as the annealing temperature increased for scanning electron microscopy (not shown here). The Ga2O3 thin film deposited by sputtering was originally in an amorphous state. After the annealing process, the amorphous state can change to the polymorphic state, which indicates that more crystal structures can reorganize. Therefore, the number of grain boundaries can decrease as the annealing temperature increases, that is, the dark current of the sample can be reduced at high annealing temperatures due to the reduction of defects and healing of leakage current paths. However, the dark current apparently increased for PD3 with an annealing temperature of Ga2O3 at 800 °C. Ju et al. reported that MgZnO thin films exhibited better crystal quality at an annealing temperature of 750 °C. When the annealing temperature of the MgZnO thin film was raised to 850 °C, the full-width half-maximum (FWHM) of the X-ray diffraction (XRD) signal rapidly increased owing to phase separation [34]. In this study, the annealing temperature of the MgZnO thin film was 700 °C, to obtain a high crystal quality before the growth of the Ga2O3 thin film. However, the Ga2O3 thin film grown on the MgZnO thin film was subsequently annealed at a temperature of 800 °C, which can result in the destruction of the MgZnO crystalline structure. In addition, Chikoidze et al. proposed that the annealing temperature of Ga2O3 thin films above 700 °C in air or oxygen ambient would lead to the creation of oxygen vacancies (Vo) due to the out-diffusion of oxygen atoms from the Ga2O3 thin film. This generation of Vo would lead to an additional source of free electrons and an n-type conductivity enhancement [35]. This is why PD3 with an annealing temperature of Ga2O3 at 800 °C had a reverse effect on the dark-current characteristic. In contrast, the photo-to-dark-current ratios (PDCRs) of PD1, PD2, and PD3 at 5 V were 30.75, 132.18, and 1.33, respectively. The PD2 with an annealing temperature of Ga2O3 at 600 °C demonstrated the highest PDCR, owing to the improvement of the Ga2O3 thin film quality and maintenance of the MgZnO crystalline quality at the same time.   Figure 4a presents the time-dependent current analysis of the Ga2O3/MgZnO heterojunction UV MSM PDs, with and without different annealing temperatures, using an Agilent HP-4156C semiconductor parameter analyzer, and the excitation light source used a 300 W Xe lamp with an electronic shutter to stop the excitation. Oxide semiconductor materials usually have a persistent photoconductivity (PPC) effect, which is a phenomenon in which carriers are captured by defects,   Figure 4a presents the time-dependent current analysis of the Ga 2 O 3 /MgZnO heterojunction UV MSM PDs, with and without different annealing temperatures, using an Agilent HP-4156C semiconductor parameter analyzer, and the excitation light source used a 300 W Xe lamp with an electronic shutter to stop the excitation. Oxide semiconductor materials usually have a persistent photoconductivity (PPC) effect, which is a phenomenon in which carriers are captured by defects, and they persist for an extended period after the termination of light excitation. Therefore, carriers trapped at defects can be released slowly, and the carrier transport to the electrodes is delayed, which can affect the transient behaviors of the photocurrent in these types of PDs. Namely, the PPC effect is closely related to the defect density of the PDs. As shown in Figure 4a, the photocurrent of PD2 with an annealing temperature of Ga 2 O 3 at 600 • C dropped faster than those of the other two because of the smaller PPC effect. Then, the curves can be fitted by the following equation [36,37]:

Persistent Photoconductivity Effect
where A 1 and A 2 are constants, t is the measured time, and τ 1 and τ 2 are the time constants. The τ 1 of PD1, PD2, and PD3 was calculated to be 5.2, 1.78, and 8.28 s, respectively. Apparently, the PD2 with an annealing temperature of Ga 2 O 3 at 600 • C achieved the best time response characteristic. This is because the defect density of Ga 2 O 3 at the grain boundary can be significantly reduced without damaging the quality of the MgZnO thin film; therefore, the PPC effect has a lesser influence. This time-response characteristic is consistent with the aforementioned result of the dark current. Figure 4b presents the I-V characteristics of PD2 under UV ON/OFF cycles. The PD2 with an annealing temperature of Ga 2 O 3 at 600 • C exhibited sufficient cyclicality under periodic illumination, indicating that the device could be reproduced.
where A1 and A2 are constants, t is the measured time, and τ1 and τ2 are the time constants. The τ1 of PD1, PD2, and PD3 was calculated to be 5.2, 1.78, and 8.28 s, respectively. Apparently, the PD2 with an annealing temperature of Ga2O3 at 600 °C achieved the best time response characteristic. This is because the defect density of Ga2O3 at the grain boundary can be significantly reduced without damaging the quality of the MgZnO thin film; therefore, the PPC effect has a lesser influence. This time-response characteristic is consistent with the aforementioned result of the dark current. Figure  4b presents the I-V characteristics of PD2 under UV ON/OFF cycles. The PD2 with an annealing temperature of Ga2O3 at 600 °C exhibited sufficient cyclicality under periodic illumination, indicating that the device could be reproduced.

Voltage-Tunable the Central Wavelength of Spectrum
Figure 5a-c presents the normalized responsivities of PD1, PD2, and PD3 at different bias voltages, respectively. As shown in Figure 5a, it was found that the wavelengths of peak responsivity of PD1 at bias voltages of 1 and 5 V were approximately 250 and 320 nm, respectively. As the bias voltage of PD1 increased, the peak detection wavelength of PD1 could shift from 250 (UVC) to 320 nm (UVB). The absorbed wavelength of Ga2O3 and MgZnO is 250 and 320 nm, respectively. It is evident that a dual-band PD can be achieved by controlling the bias voltages. This is because, when the bias voltage is small, the electric field of the MSM PD can only be distributed to the shallower thin-film area. At this time, only the electron-hole pairs generated in the Ga2O3 thin film on the top layer of the MSM PD can be swept to the electrodes by the lower electric field. Then, when the bias voltage is increased, the electric field can be distributed to cover the Ga2O3 and MgZnO thin-film area. Moreover, the internal quantum efficiency of the MgZnO thin film was higher than that of the Ga2O3 thin film. Therefore, in the case of the higher bias voltage, the peak responsivity of PD was mainly dominated by the MgZnO thin film. Next, as shown in Figure 5b, it can be observed that the same voltage-tunable dual-band characteristic as PD1 in Figure 5a cannot be achieved. Irrespective of the

Voltage-Tunable the Central Wavelength of Spectrum
Figure 5a-c presents the normalized responsivities of PD1, PD2, and PD3 at different bias voltages, respectively. As shown in Figure 5a, it was found that the wavelengths of peak responsivity of PD1 at bias voltages of 1 and 5 V were approximately 250 and 320 nm, respectively. As the bias voltage of PD1 increased, the peak detection wavelength of PD1 could shift from 250 (UVC) to 320 nm (UVB). The absorbed wavelength of Ga 2 O 3 and MgZnO is 250 and 320 nm, respectively. It is evident that a dual-band PD can be achieved by controlling the bias voltages. This is because, when the bias voltage is small, the electric field of the MSM PD can only be distributed to the shallower thin-film area. At this time, only the electron-hole pairs generated in the Ga 2 O 3 thin film on the top layer of the MSM PD can be swept to the electrodes by the lower electric field. Then, when the bias voltage is increased, the electric field can be distributed to cover the Ga 2 O 3 and MgZnO thin-film area. Moreover, the internal quantum efficiency of the MgZnO thin film was higher than that of the Ga 2 O 3 thin film. Therefore, in the case of the higher bias voltage, the peak responsivity of PD was mainly dominated by the MgZnO thin film. Next, as shown in Figure 5b, it can be observed that the same voltage-tunable dual-band characteristic as PD1 in Figure 5a cannot be achieved. Irrespective of the small or large bias voltage, the peak of the response was mainly dominated by the MgZnO thin film, and it can only detect the UVB band. The relevant description and explanation are shown in the schematic diagram of the energy band in Figure 5d. The left image in Figure 5d is a schematic diagram of the energy band of PD1. Because the Ga 2 O 3 thin film of PD1 without an annealing process has a higher resistance and a lower carrier concentration, a higher potential energy barrier can be generated at the heterojunction of Ga 2 O 3 and MgZnO, to prevent electrons and holes from the bottom layer of the MgZnO thin film at lower bias voltages. Therefore, such a result can assist in achieving a voltage-tunable dual-band PD. However, the right image in Figure 5d is a schematic diagram of the energy band of PD2. Owing to the annealing process, the resistance of the Ga 2 O 3 thin film of PD2 decreases, and the carrier concentration increases, which reduces the potential energy barrier between the heterojunction of Ga 2 O 3 and MgZnO. It is indicated that this lower barrier height cannot effectively keep off the carriers, despite lower bias voltages. This result makes it easy for electrons and holes from the bottom layer of the MgZnO thin film to transit the barrier height and reach the electrodes. Finally, in Figure 5c, the annealing process at 800 • C caused the phase separation of the MgZnO thin film and the increase of oxygen vacancies in the Ga 2 O 3 thin film, as indicated in the previous paragraph. Therefore, PD3 with an annealing temperature of 800 • C exhibited the worst response characteristics and could not achieve dual-band detection.

Extend the Operating Bias Voltage of UVB
Although voltage-tunable UVC-UVB dual-band MSM PDs, which can shift the central detection wavelength to over 70 nm by using different bias voltages, were successfully fabricated, the bias voltages of the PDs for detecting UVC and UVB were too close, limiting the operating voltage range of the devices. Therefore, the structure of PD4 was also designed with a 25 nm-thick SiO2 layer

Extend the Operating Bias Voltage of UVB
Although voltage-tunable UVC-UVB dual-band MSM PDs, which can shift the central detection wavelength to over 70 nm by using different bias voltages, were successfully fabricated, the bias voltages of the PDs for detecting UVC and UVB were too close, limiting the operating voltage range of the devices. Therefore, the structure of PD4 was also designed with a 25 nm-thick SiO 2 layer inserted between Ga 2 O 3 and the MgZnO thin film to conduct this concept. As shown in Figure 6a, the peak responsivity of PD4 with a 25 nm-thick SiO 2 insertion layer was still dominated by the Ga 2 O 3 thin film, to detect the UVC range at a bias voltage of 3 V, despite the bias voltage being as high as 7 V. It was not until the bias voltage reached 24 V that the peak responsivity of PD4 could be turned to be dominated by the MgZnO thin film to detect the UVB range. Considering these results, a 25 nm-thick SiO 2 inserting layer had a significant effect and could achieve a broader operating bias voltage range. Figure 6b presents the schematic energy band diagram of the electron transport mechanism for PD4 with a SiO 2 inserting layer. SiO 2 has a significantly high potential energy barrier compared to the original heterojunction of Ga 2 O 3 /MgZnO. As shown in Figure 6b, there are only two ways for the electron to pass through this high barrier. One is across, by a significantly high bias voltage, and the other is directly through the tunneling effect. This is why PD4 cannot detect the UVB band until the bias condition of 24 V. In addition, owing to the influence of the tunneling effect, the response values of PD4 in the UVB range at a bias voltage of 7 V were higher than those of 3 V. Namely, according to this design concept, a suitable voltage-tunable dual-band photodetector can be fabricated by using different SiO 2 thicknesses, without affecting the quality of the Ga 2 O 3 and MgZnO thin films. Moreover, the performance of this study is compared with state-of-the-art reports, as shown in Table 1. Although dual-band PDs have been reported several times in previous studies, only the design proposed in this study has the ability to turn the central wavelength of the spectrum from UVC to UVB by different bias voltages. according to this design concept, a suitable voltage-tunable dual-band photodetector can be fabricated by using different SiO2 thicknesses, without affecting the quality of the Ga2O3 and MgZnO thin films. Moreover, the performance of this study is compared with state-of-the-art reports, as shown in Table 1. Although dual-band PDs have been reported several times in previous studies, only the design proposed in this study has the ability to turn the central wavelength of the spectrum from UVC to UVB by different bias voltages.

Conclusions
In summary, Ga2O3/MgZnO heterostructure MSM PDs were fabricated by RF sputtering, and these fabricated PDs could detect the UVC and UVB dual-band individually by controlling different bias voltages. Although the crystal quality of the Ga2O3 thin film, the dark current of PD, PDCR of PD, and PCC effect can be improved by an annealing temperature of 600 °C, it cannot achieve a voltage-tunable dual-band characteristic. This was because the PD with an annealing temperature of Ga2O3 at 600 °C had a lower potential energy barrier between the Ga2O3 and MgZnO heterojunctions. Otherwise, the PD without annealing had a higher heterojunction barrier and could suppress the

Conclusions
In summary, Ga 2 O 3 /MgZnO heterostructure MSM PDs were fabricated by RF sputtering, and these fabricated PDs could detect the UVC and UVB dual-band individually by controlling different bias voltages. Although the crystal quality of the Ga 2 O 3 thin film, the dark current of PD, PDCR of PD, and PCC effect can be improved by an annealing temperature of 600 • C, it cannot achieve a voltage-tunable dual-band characteristic. This was because the PD with an annealing temperature of Ga 2 O 3 at 600 • C had a lower potential energy barrier between the Ga 2 O 3 and MgZnO heterojunctions. Otherwise, the PD without annealing had a higher heterojunction barrier and could suppress the carriers from the bottom layer of the MgZnO thin film to cross this potential barrier at a lower bias voltage of 1 V. The peak responsivity was observed at 250 nm and was mainly dominated by the top layer of the Ga 2 O 3 thin film. Next, as the bias voltage increased to 5 V, the peak detection wavelength shifted from UVC to UVB. Furthermore, a 25 nm-thick SiO 2 layer inserted between the Ga 2 O 3 and MgZnO thin film of PD can achieve a broader operating voltage range for dual-band applications owing to the more effective blocking effect.