High Dynamic Range 120 dB Ga₂O₃‐In₂Se₃ Photodetector Array by Ferroelectric Polarization Enhancement

ABSTRACT

Gallium oxide‐based photodetectors have specialized applications in both civilian and military fields due to their excellent spectral selectivity for solar‐blind ultraviolet (SBUV) light. However, gallium oxide photodetectors still face challenges in weak‐light detection capability, which hinders their further commercialization. Here, we devised a Ga₂O₃‐In₂Se₃ single‐crystal van der Waals heterojunction photodetector by using the polarization modulation effect of ferroelectric semiconductors. Under illumination with a 255 nm light, the Ga₂O₃‐In₂Se₃ photodetector achieves an on/off ratio of 712,434 and a detectivity of 1.6 × 10¹⁴ Jones, with a linear dynamic range (LDR) of 120 dB. Furthermore, we fabricated a 5×5 array and all photodetectors exhibit consistent optoelectronic performance, indicating the imaging application potential of the array. This work provides novel insights for ferroelectric‐enhanced solar‐blind ultraviolet detectors.

We devised a Ga₂O₃‐In₂Se₃ single‐crystal van der Waals heterojunction photodetector and 5×5 array. Under illumination with a 255 nm light, the Ga₂O₃‐In₂Se₃ photodetector achieves an on/off ratio of 712,434 and a detectivity of 1.6 × 10¹⁴ Jones, with a linear dynamic range (LDR) of 120 dB.

1. Introduction

Photodetectors play a pivotal role in modern optoelectronics and underpin a broad spectrum of applications ranging from conventional domains such as signal detection [1, 2], imaging [3], quantum information processing [4, 5], telecommunications [6, 7], wearable electronics [8, 9], and space science [10], to emerging interdisciplinary scenarios including photonic nociceptors [11], autonomous driving systems [12], and flame alarm technologies [13], thereby continuously expanding their functional scope and technological relevance. The basic principle of a photodetector is that a semiconductor device converts optical signals (amplitude, frequency) into electrical signals (electrons, voltage) [14, 15]. The solar‐blind ultraviolet (SBUV) photodetector is a device operating in the 200–280 nm wavelength band, featuring extremely low background noise [2]. This is attributed to the near absence of solar‐blind light in the atmosphere, which is predominantly absorbed by the ozone layer. Gallium oxide (Ga₂O₃) represents a new generation of ultra‐wide bandgap semiconductor materials, featuring a bandgap (∼4.8–4.9 eV) significantly broader than silicon carbide and gallium nitride [16]. In recent years, extensive research has been conducted on the growth of Ga₂O₃ thin films, including methods such as chemical vapor deposition (CVD) [17], pulsed laser deposition (PLD) [18], atomic layer deposition (ALD) [19], and metal‐organic chemical vapor deposition (MOCVD) [20]. Among these methods, MOCVD is the preferred method for industrial‐scale production of Ga₂O₃ films, as it enables the fabrication of Ga₂O₃ films with wafer‐scale single‐crystal quality. Along with the maturation of Ga₂O₃ thin‐film growth technologies, significant advances have also been achieved in Ga₂O₃‐based photodetectors. Various device architectures, including metal‐semiconductor‐metal (MSM), Schottky, p‐n junction, and heterojunction photodetectors, have been successfully developed and demonstrated outstanding performance. Although various types of Ga₂O₃ photodetectors have been investigated, challenges such as low‐light detection and ultra‐wide linear dynamic range (LDR) remain unresolved [21], which significantly hinders the commercialization of gallium oxide materials.

Recently, numerous studies have reported that the van der Waals semiconductor material α‐In₂Se₃ exhibits both in‐plane and out‐of‐plane ferroelectricity at room temperature [22], along with excellent ferroelectric controllability, demonstrating significant application potential in both ferroelectric and photodetection fields. On the one hand, since the light absorption edge of In₂Se₃ is determined by its bandgap, it is a narrow‐bandgap semiconductor material. On the other hand, In₂Se₃ exhibits a strong response to irradiation in the visible to near‐infrared spectral range (254–1064 nm); therefore, In₂Se₃ has been widely used in research on broadband detectors [23, 24, 25, 26]. Ga₂O₃, as an emerging n‐type ultra‐wide bandgap semiconductor material, holds significant application potential in fields such as power devices and ultraviolet detectors [27]. Consequently, it is more efficient and reliable than narrow‐bandgap materials like Si and wide‐bandgap materials like SiC [28, 29]. Conventional heterojunction photodetectors primarily rely on static band engineering to enhance device performance. However, due to the relatively weak built‐in electric field and the lack of active tunability, photogenerated electrons and holes are prone to recombination, and carrier mobility is often limited. Traditional heterojunction devices frequently struggle to simultaneously achieve high responsivity, elevated detectivity, and fast response speed [30]. Therefore, constructing a novel device architecture that integrates the synergistic regulation of the built‐in electric field and the ferroelectric polarization field is of great significance for realizing high‐performance solar‐blind ultraviolet photodetectors [31]. The ferroelectric polarization modulation effect of In₂Se₃ can enhance photogenerated carrier separation efficiency [32]. Under P_down conditions, due to the lower built‐in electric field strength/barrier height, more carriers can traverse the built‐in electric field region, resulting in a higher current compared to the original state. In contrast, under the P_up state, the higher built‐in electric field strength/barrier height impedes charge carrier transfer, resulting in lower current [33]. The negative capacitance effect enhances the built‐in electric field, resulting in a significant increase in the Schottky barrier or carrier injection barrier at the metal‐semiconductor interface [34]. This suppresses electron injection from the metal into the conduction band, thereby reducing the dark current induced by thermal excitation. Furthermore, under weak incident light conditions, the photocurrent induced by low‐intensity illumination can be collected and amplified more rapidly and efficiently due to the negative capacitance effect [35, 36].

In this work, we demonstrate a CVD growth and transfer method for fabricating single‐crystal Ga₂O₃‐In₂Se₃ heterojunctions. The intrinsic ferroelectric polarization of In₂Se₃ can be utilized to establish a built‐in polarization field at the heterojunction interface, which promotes the efficient separation and transport of photogenerated carriers while suppressing carrier recombination. Consequently, the dark current is reduced, leading to substantial improvements in device performance, particularly in terms of specific detectivity and responsivity [37]. Thanks to the superior quality of the heterojunction devices, these ferroelectric photodetectors exhibit excellent response under solar‐blind ultraviolet irradiation. Under illumination with a 255 nm light source, the device achieves a responsivity of 250 A/W and a detectivity of 1.6 × 10¹⁴ Jones, with an on/off ratio of 712,434 and a linear dynamic range of 120 dB. In addition, we fabricated a 5×5 array and tested all devices to investigate the performance consistency and stability of this heterojunction photodetector, as well as the imaging capability of the array.

2. Experimental Section

2.1. Growth of Ga₂O₃ Films

The Ga₂O₃ single‐crystal films were grown on 2‐inch sapphire substrate by metal‐organic chemical vapor deposition (MOCVD) using triethylgallium as Ga source. The growth temperature was set at 800°C, with a reaction chamber pressure of 50 Torr. The carrier gases used were nitrogen and oxygen, with flow rates of 200 and 3000 sccm, respectively.

2.2. Growth of In₂Se₃ Films

The In₂Se₃ films were grown using the chemical vapor deposition (CVD) method in a tube furnace. The powder source utilized approximately 100 mg of In₂Se₃ powder and 20 mg of InSe powder, with a mass ratio of 5:1. The powders were thoroughly mixed and evenly spread on a quartz boat. The quartz boat containing the powders was then placed in the heating zone of the tube furnace. Commercial mica (KMg₃AlSi₃O₁₀F₂) was used as the substrate and placed directly on top of the In₂Se₃ and InSe powders. The distance between the two raw materials and the substrate was controlled within a narrow range (approximately 3–4 mm). Prior to the reaction, to prevent the materials from overheating and oxidizing due to air exposure during the reaction, the system is placed in an argon atmosphere for 1 h to purge the air and ensure the system is fully immersed in argon. Subsequently, under argon flow, the temperature in the heating zone is raised to 800°C. Once the target temperature is reached, the system is held at this temperature for 20–30 min. The heating rate of the system is typically 33°C/min. After the holding period, the products are allowed to cool to room temperature inside the furnace. The quartz boat is then slowly removed from the heating zone, resulting in the formation of two‐dimensional α‐In₂Se₃ on the mica substrate.

2.3. Methods for Transferring In₂Se₃ Films

First, a layer of polymethyl methacrylate (PMMA) solution was spin‐coated onto the surface of a mica wafer bearing a two‐dimensional In₂Se₃ film using a spin coater operating at 2000 rpm. Next, the PMMA‐coated sample was placed on a heating plate at 80°C and baked for 5 min to evaporate the organic solvent in the PMMA, thereby curing it into a film. Then, blue masking tape was used to tightly wrap the edges of the sample. The sample was immersed in deionized water at 60°C for 5 min, and the PMMA film containing the sample was peeled off the mica sheet using tweezers, aided by the buoyancy of the water. Next, the PMMA film containing the sample was transferred onto a Ga₂O₃ film and placed on a heating stage at 50°C for 5–10 min. In this manner, the In₂Se₃ film was transferred onto the Ga₂O₃‐coated substrate using the PMMA‐assisted transfer technique. Finally, the sample was placed in a glass petri dish, and a small amount of residual PMMA on the sample was dissolved using an acetone solution.

2.4. Fabrication and Characterization of Photodetectors

Au or Bi electrodes were deposited onto Ga₂O₃‐In₂Se₃ heterojunctions using thermal evaporation. First, a small amount of Au or Bi particles was placed on a tungsten boat, and then the heterojunction, secured with a copper mesh mask, was fixed to the top of the evaporation chamber. Next, the pressure inside the chamber was reduced to 7 × 10^{− 4} Pa using a vacuum pump. The current was then slowly adjusted to gradually increase the temperature of the tungsten boat. Once the gold particles had completely evaporated onto the surface of the heterojunction, the sample was removed, thereby completing the preparation of the photodetector sample. All current‐voltage (I‐V) and current‐time (I‐T) curves were measured using a semiconductor analyzer (Keithley 4200‐SCS) in the vacuum environment of a probe station (MyProber). The light source used for measurement was a 255 nm light‐emitting diode (LED), whose optical power density had been calibrated using an ultraviolet radiometer.

2.5. Characterization of Materials

The Ga₂O₃/In₂Se₃ heterojunction was observed using an optical microscope (Ningbo Sunny, CX40M). The structure of the material was characterized using X‐ray diffraction (XRD) and Raman spectroscopy. Selected‐area electron diffraction (SAED) was employed to investigate the crystal structures of Ga₂O₃ and In₂Se₃. Ultraviolet photoelectron spectroscopy (UPS) and ultraviolet‐visible (UV‐Vis) absorption spectroscopy were utilized to determine the work functions, identify the positions of the Fermi level and valence band, and evaluate the bandgap energies of the materials. The morphology of the Ga₂O₃‐In₂Se₃ heterojunction was investigated using a scanning electron microscope (SEM). Atomic force microscopy (AFM) was used to measure the height difference at the interface of the Ga₂O₃‐In₂Se₃ heterojunction. Kelvin probe force microscopy (KPFM) was employed to investigate the potential difference in the heterojunction region. Piezoresponse force microscopy (PFM) is a powerful technique for determining the ferroelectricity of materials; it clearly reveals experimental observations of robust intralayer ferroelectricity in the Ga₂O₃‐In₂Se₃ thin films. TEM data were collected using an aberration‐corrected transmission electron microscope (JEOL, JEM‐ARM200F).

2.6. Fabrication of 5×5 Array

Following thin‐film growth and transfer, an α‐In₂Se₃ film was successfully transferred onto a sapphire substrate coated with a Ga₂O₃ film. A copper mesh was then placed over the In₂Se₃ surface to serve as a hard mask, defining the channel (approximately 57 µm). Subsequently, an Au electrode was deposited onto the surface of the In₂Se₃ film using a thermal evaporator. Finally, after removing the copper mesh, a 5×5 array of Ga₂O₃‐In₂Se₃ heterojunction devices was obtained.

3. Results and Discussion

Using In₂Se₃‐InSe mixed powders as raw materials, In₂Se₃ thin films were successfully grown on mica substrates via the chemical vapor deposition (CVD) method in a tube furnace. Figure 1a illustrates the growth process, raw materials, and molecular structure. Freshly exfoliated mica was selected as the growth substrate due to its chemical inertness and atomically smooth surface, which allow for migration with a relatively low energy barrier—a critical factor for the epitaxial growth of 2D materials [38, 39]. The large‐area growth of Ga₂O₃ and In₂Se₃ films has made the fabrication of Ga₂O₃‐In₂Se₃ heterojunction array devices possible (Figure S1). Specifically, the schematic structure of a 5×5 Ga₂O₃‐In₂Se₃ array, consisting of Ga₂O₃ films fabricated by the MOCVD method and In₂Se₃ films prepared by the CVD method, is shown in Figure 1b. The 5×5 array is obtained by integrating individual devices, while the heterojunction, combined with Au electrodes deposited via thermal evaporation, forms a single photodetector, as illustrated in Figure 1c. Figure 1d,e shows the crystal structures of α‐In₂Se₃ and Ga₂O₃, respectively. The crystal structure of α‐In₂Se₃ clearly reveals the crystal is stacked in the sequence of Se‐In‐Se‐In‐Se atomic layers [40]. The crystal structure diagram of Ga₂O₃ reveals the coordination of Ga and O atoms, with each Ga atom surrounded by six O atoms. Figure S2 shows the SEM image of the Ga₂O₃/In₂Se₃ heterojunction, in which the surface morphology and the interface boundary between the two materials can be clearly observed. In addition, Figure S3 shows the Raman spectra and XRD patterns of the In₂Se₃ and Ga₂O₃ thin films. The Raman spectrum of the In₂Se₃ film exhibits characteristic peaks at 98, 183, and 203 cm^{− 1}, with a pronounced peak located at 203 cm^{− 1}, which is a typical signature of crystalline In₂Se₃ [41]. The corresponding XRD pattern displays distinct diffraction peaks indexed to the (006), (0010), (0014), and (0016) crystal planes, indicating a highly oriented layered structure. For the Ga₂O₃ film, characteristic Raman peaks are observed at 201, 352, 632, and 767 cm^{− 1}, which are consistent with the reported vibrational modes of monoclinic Ga₂O₃. The XRD diffraction peaks can be assigned to the (−201), (401), and (402) crystal planes, together with several diffraction peaks originating from the Al₂O₃ substrate. The Raman and XRD results of both In₂Se₃ and Ga₂O₃ films are in good agreement with the corresponding standard reference spectra and diffraction data, and no obvious impurity‐related peaks are detected. These results demonstrate the successful growth of high‐quality In₂Se₃ and Ga₂O₃ thin films with excellent crystallinity and phase purity.

FIGURE 1.

Growth process, device fabrication and structural characterization. (a) Schematic diagram of the CVD growth of 2D In₂Se₃. (b) Schematic diagram of a 5×5 Ga₂O₃‐In₂Se₃ array and (c) a single device. (d) and (e) show the crystal structures of Ga₂O₃ and In₂Se₃, respectively. (f) HRTEM image of the Ga₂O₃‐In2Se3 interface. (g) HRTEM image of the Ga₂O₃. (h) SAED patterns of Ga₂O₃ and In₂Se₃. (i) HAADF‐STEM image of Ga₂O₃. (j) HAADF‐STEM image of In₂Se₃. (k) HAADF image and the corresponding elemental mapping of Ga/O/In/Se elements.

Figure 1f shows a high‐resolution transmission electron microscopy (TEM) image of the Ga₂O₃/In₂Se₃ heterojunction, clearly revealing the interface between the two materials and the arrangement of the atomic layers. Figure 1g shows the HRTEM image of Ga₂O₃. To verify the single‐crystalline characteristics of the Ga₂O₃ and In₂Se₃ films, the selected area electron diffraction (SAED) patterns of both materials are presented in Figure 1h. As shown in the figure, both films exhibit well‐defined diffraction spots, indicating their high crystalline quality and single‐crystalline nature. Furthermore, the observed diffraction spots can be indexed to the corresponding crystal planes of Ga₂O₃ and In₂Se₃, providing additional evidence for the successful growth of single‐crystalline films in this work. High‐angle annular dark‐field scanning transmission electron microscopy (HAADF‐STEM) images of In₂Se₃ and Ga₂O₃ are shown in Figures 1i, and 1j respectively. Figure 1i reveals that the interplanar distances between the two Ga₂O₃ crystal planes (with an angle of 63°) are 0.27 and 0.33 nm, corresponding to the (111) and (201) crystal planes of β‐Ga₂O₃, respectively. Similarly, Figure 1j shows that the lattice spacing of In₂Se₃ is 0.976 nm, corresponding to the (0001) crystal plane [42]. The elemental distribution mapping in the Ga₂O₃/In₂Se₃ heterojunction is shown in Figure 1k, with the corresponding elements of Ga, O, In, and Se. Meanwhile, Figure S4 presents the XPS spectra of the In₂Se₃ and Ga₂O₃ films. Distinct In and Se characteristic peaks are observed in the In₂Se₃ spectrum, while no impurity‐related signals are detected, demonstrating the high chemical purity of the as‐grown α‐In₂Se₃ films. For the Ga₂O₃ film, the absence of a high‐binding‐energy shoulder or component associated with oxygen‐vacancy defects suggests a low defect concentration, further confirming its high material purity and superior surface quality.

Figure 2a shows the ferroelectric switching behavior of the Ga₂O₃/In₂Se₃ heterojunction over a direct current (DC) voltage range from −70 V to 70 V [43]. The typical butterfly‐shaped loop in the amplitude‐voltage curve shown in Figure 2b, along with the hysteresis loops and 180° phase shift in Figure 2c, indicates the ferroelectric switching behavior of the Ga₂O₃/In₂Se₃ heterojunction. To investigate the effect of In₂Se₃ on device performance, I–V characteristic curves were measured under different polarization states. As shown in the test results in Figure 2d,e, when the polarization direction is upward, negative polarization charges rapidly accumulate at the interface and combine with holes encountered there, resulting in a shorter depletion layer. Simultaneously, the built‐in electric field in the junction region decreases, allowing carriers to traverse the barrier more easily, thereby increasing the photocurrent. Conversely, when the polarization direction is downward, the depletion layer widens and the barrier height increases, achieving the effect of suppressing dark current. Figure 2f compares the detection sensitivity under different polarization states and the initial state. It is evident from the figure that when the polarization direction is upward and downward, they exert significant enhancement and suppression effects on the device, respectively. Similarly, based on Figure 2g, Kelvin probe force microscopy (KPFM) measurements of the Ga₂O₃‐In₂Se₃ heterojunction revealed a height difference of approximately 52.5 nm, as shown in Figure 2h. Additionally, the potential difference across this heterojunction, as shown in Figure 2i, is approximately 163 mV, clearly indicating the formation of an intrinsic electric field at the junction interface. Figure 2j shows the UPS spectra of the Ga₂O₃ and In₂Se₃ films. The work functions were determined by linearly extrapolating the secondary electron cutoff regions and using the incident photon energy of 21.22 eV. The calculated work functions are 4.78 eV for Ga₂O₃ and 4.28 eV for In₂Se₃. In addition, the valence band onset positions obtained from the UPS spectra indicate that the energy separations between the Fermi level and the valence band maximum are 3.78 and 1.16 eV for Ga₂O₃ and In₂Se₃, respectively. The UV‐Vis absorption spectra of Ga₂O₃ and In₂Se₃ are shown in Figure 2k. Their optical bandgaps were estimated from Tauc plots, with the bandgap of Ga₂O₃ being 4.83 eV and that of In₂Se₃ being 1.30 eV. Combined with the UPS results, these optical measurements can determine the band alignment of Ga₂O₃ and In₂Se₃. Figure 2l shows the band structures of the materials Ga₂O₃ and In₂Se₃, respectively. Figure 2m displays the band diagram of the heterojunction device composed of these two materials under illumination by a light source (255 nm).

FIGURE 2.

Schematic diagram of (a) the PFM measurement. (b) and (c) show the amplitude hysteresis loops and the phase captured at a selected point on the loop, respectively. (d) and (e) show the I–V curves of the device under different bias conditions. (f) shows a comparison of detectivity under different polarization conditions. (g), (h), and (i) show the KPFM images of the Ga₂O₃/In₂Se₃ heterojunction interface, including height and potential profiles. (j) Ultraviolet photoelectron spectroscopy (UPS) spectra of Ga₂O₃ and In₂Se₃. (k) Ultraviolet‐visible (UV‐Vis) absorption spectra of Ga₂O₃ and In₂Se₃. (l) shows the band structure of Ga₂O₃ and In₂Se₃. (m) shows the band diagram of the heterojunction under illumination.

In this study, given that Ga₂O₃ inherently possesses solar‐blind characteristics, while In₂Se₃ can only absorb visible and near‐infrared light and lacks solar‐blind characteristics, we tested the broad‐spectrum response of Ga₂O₃/In₂Se₃, as shown in Figure 3a. We presented the responsivity curve of the heterojunction device tested at wavelengths from 255 to 900 nm to verify whether the heterojunction still exhibits solar‐blind characteristics. As can be seen from the figure, the responsivity of the device is excellent under the illumination of a 255 nm light source. However, when illuminated with light sources of wavelengths greater than 255 nm, the responsivity drops sharply and approaches zero. By comparing the responsivity at 255 nm with that at other wavelengths, the rejection ratio of the device can be obtained, as shown in Figure 3b, which can be calculated using the following formula [44]:

$$R_{255nm}/R_{\lambda >255nm}$$

here, R _255 nm  represents the responsivity of the device under the illumination of a 255 nm light source, and R _{λ > 255 nm} represents the corresponding responsivity of the device under the illumination of light sources with wavelengths greater than 255 nm. From the calculation and the data in the figure, the rejection ratio is as high as 8 × 10⁴ when the incident wavelength exceeds 300 nm. Since the bandgap of Ga₂O₃ is 4.8–4.9 eV, the corresponding cut‐off wavelength is 254 nm. Therefore, 255 nm ultraviolet light can be strongly absorbed by Ga₂O₃. In the heterojunction device, In₂Se₃ mainly plays the roles of constructing an internal built‐in electric field, improving the separation efficiency of carriers, and reducing the recombination rate, rather than participating in light absorption. Therefore, the heterojunction photodetector constructed by combining the two materials still maintains solar‐blind characteristics. As shown in the noise spectrum in Figure 3c, the current noise gradually decreases with increasing frequency and approaches the shot‐noise limit in the high‐frequency region. As shown in Figure 3c, at a bias voltage of −5 V, the shot noise limit is 6×10^{− 15} A/Hz^1/2.

FIGURE 3.

(a) Spectral response of the Ga₂O₃/In₂Se₃ heterojunction photodetector. (b) Rejection ratios of the Ga₂O₃/In₂Se₃ heterojunction photodetector. (c) Noise current spectral of the Ga₂O₃/In₂Se₃ heterojunction photodetector measured under a bias voltage of −5 V. (d) I–V curve of an Au contact heterojunction device under illumination by a 255 nm light source. (e) Responsivity and detectivity of the Au contacts under different bias voltages. (f) The external quantum efficiency (EQE) and photo‐dark current ratio (PDCR). (g) I‐V curve of a Bi‐contact heterojunction device under illumination by a 255 nm light source, along with responsivity, detectivity (h), EQE and PDCR (i) under different bias voltages. (j) Photocurrent and dark current of the Bi‐contact device under illumination by a 255 nm light source. (k) Fitted curve of the linear dynamic range. (l) Photo‐dark current ratio (PDCR).

Au and Bi metal electrodes were deposited respectively on the surface of Ga₂O₃‐In₂Se₃ heterojunctions, and their I‐V curves were measured at different light power densities under 255 nm illumination. In Figure 3d, the test results for the heterojunction with an Au contact show that the device's output current changes significantly as the light power density increases from 60 to 300 µW/cm². Of course, the I‐V curve is not the only metric describing the performance of a photodetector; responsivity and detectivity are also important metrics. Responsivity can be calculated using the following formula [40, 45]:

$$R=\left(I_L-I_D\right)/S·P_L$$

here, R represents the responsivity, I_L is the photocurrent, and I_D is the dark current; S denotes the effective area of the photodetector, and P_L is the power density of the test light source. The detectivity is a key parameter describing the sensitivity of a photodetector. According to the definition of detectivity, the specific detectivity can be expressed as [30, 46]:

$$D^{\ast }=R·\sqrt{S·\mathrm{\Delta }f}/I_n$$

D* represents the detectivity, R represents the responsivity, Δf is the bandwidth, S is the effective area of the device, and I_n is the current noise. Experimental measurements show that the responsivity and detectivity of the Au‐contacted photodetector are as shown in Figure 3e. The responsivity reached 241.98 A/W at a voltage of −10 V, while the detectivity reached 1.6×10¹⁴ Jones at a voltage of −10 V. Furthermore, in addition to responsivity and detectivity, the external quantum efficiency (EQE) and photocurrent on‐off ratio (PDCR) serve as crucial figures of merit for photodetectors. Specifically, the EQE signifies the efficiency of a photodetector in converting incident photons into collected charge carriers, whereas the PDCR gauges the device's capability to respond to optical signals. The mathematical expression for calculating the EQE is given as follows:

$$EQE=R\frac{hc}{q\lambda }\times 100\%$$

In the formula, R represents the response rate, h is the Planck constant, c is the speed of light, q is the elementary charge, and λ represents the wavelength. In Figure 3f, it can be observed that the EQE of this heterojunction photodetector reaches a maximum of 1.17×10⁵% at a voltage of −10 V, while the on/off ratio reaches a maximum of 712434 at a voltage of −8 V. When the device uses Bi electrodes, as shown in Figure 3g, as the light power density increases from 60 to 300 µW/cm², the photocurrent decreases by one order of magnitude compared to the gold‐contacted device, while the dark current increases by one order of magnitude. Similarly, in Figure 3h, the responsivity and detectivity calculated using the formula are 38.1 A/W at −10 V and 2.5 × 10¹³ Jones at −10 V, respectively, showing a slight decrease compared to the Au contact. Likewise, in Figure 3i, the maximum EQE of the tested device was 1.85×10⁴%, while the on/off ratio was 542,890; both performance metrics remain inferior to those achieved with Au contacts. In Figure 3j, the Bi‐contacted device was tested again; when the optical power density was increased from 300 pW/cm² to 3 mW/cm², the device's output current differed little from the previous measurement. The linear dynamic range (LDR) refers to the proportional range during which the device can effectively process and distinguish between the minimum and maximum signals [47]. Based on the data in Figure 3j, the device's LDR was fitted, and the result is shown in Figure 3k. The fitted LDR value is 120 dB; a high linear dynamic range enables the processing of a wider range of signal intensities. The formula for calculating the linear dynamic range (LDR) is as follows:

$$LDR=20lg\frac{P_{max}}{P_{min}}$$

here, P_max is the maximum optical power density, and P_min is the minimum optical power density. Meanwhile, the switching ratio calculated in Figure 3l is 496,880.

Of course, the response speed of heterojunction devices is also a key parameter for evaluating their performance. Figure 4a shows the I–V characteristic curve obtained from experimental tests evaluating the transient photoresponse of the photodetector under UV irradiation. The tests show that when the UV light is turned on, the device current increases rapidly, with a rise time of 89 ms. Conversely, when the UV light is turned off, the device current decreases rapidly, with a fall time of 87 ms. Subsequently, when the device was subjected to this operating condition for an extended period, Figure 4b shows that there was virtually no significant change in the photocurrent relative to the dark current, indicating that the Ga₂O₃/α‐In₂Se₃ heterojunction photodetector exhibits reliable stability at room temperature. To further verify the reliability and reproducibility of the heterojunction devices, five Ga₂O₃‐In₂Se₃ heterojunction photodetectors were randomly selected and characterized under identical measurement conditions. The corresponding results are presented in Table 1 and Figures S5–S7. A comparison of the key performance parameters of the five devices is also summarized in Table 1. The results show a high degree of consistency among the devices, with only minor variations in responsivity, detectivity, and response speed. These findings further demonstrate the excellent reproducibility and stability of the proposed heterojunction engineering strategy.

FIGURE 4.

(a) I–V curve and a magnified view of the response time, measured using a 255 nm light source. (b) Stability test of a Ga₂O₃‐In₂Se₃ heterojunction device at a bias voltage of 1 V.

TABLE 1.

Comparison of the photoelectric performance of five Ga₂O₃/α‐In₂Se₃ heterojunction devices.

Device Code	R (A W− 1)	D* (Jones)	EQE (%)	PDCR	Response time
1	20.5	1.37×1013	104	54923	88 ms
2	194.4	1.29×1014	9.46×104	6194	86 ms
3	29.4	1.96×1013	1.43×104	46692	91 ms
4	207.1	1.38×1014	105	12132	92 ms
5	17.7	1.18×1013	8.64×103	44321	101 ms

Meanwhile, to evaluate the effectiveness of the heterojunction engineering strategy in enhancing device performance, individual Ga₂O₃ and In₂Se₃ photodetectors were fabricated and characterized under 255 nm illumination. The corresponding results are presented in Figure S8. Furthermore, a comprehensive comparison between the Ga₂O₃/α‐In₂Se₃ heterojunction device and the pristine Ga₂O₃ and In₂Se₃ photodetectors is summarized in Table 2. The results clearly demonstrate that the Ga₂O₃/α‐In₂Se₃ heterojunction effectively improves the overall photodetection performance. In particular, the Ga₂O₃/α‐In₂Se₃ (Au) device exhibits a remarkable responsivity of 241.89 A/W, an EQE of 1.17×10⁵%, and a photo‐to‐dark current ratio (PDCR) of 7.12 × 10⁵. These values are significantly higher than those of the pristine Ga₂O₃ photodetector (19.7 A/W, 9.72×10³%, 6.79 × 10⁴) and the pristine In₂Se₃ photodetector (158.7 A/W, 7.72×10⁴%, 123).

TABLE 2.

Comparison of the photodetection performance of the Ga₂O₃/α‐In₂Se₃ heterojunction device with pure Ga₂O₃ and pure In₂Se₃ photodetectors.

Device Type	R (A W− 1)	D* (Jones)	EQE (%)	PDCR	Response time
Ga2O3/In2Se3 (Au)	241.98	1.6×1014	1.17×105	712434	87 ms
Ga2O3/In2Se3 (Bi)	38.1	2.5×1013	1.85×104	542890	92 ms
Ga2O3	19.7	1.33×1013	9.72×103	67947	117 ms
In2Se3	158.7	1.1×1014	7.72×104	123	144 ms

Figure 5 compares the device performance of this work with that of other relevant studies (such as ALD/Pt β‐Ga₂O₃, SeO₂/GaN heterojunctions, and MOCVD/Ni β‐Ga₂O₃). In terms of parameters such as response time [25, 26, 47, 48, 49, 50, 51, 52], responsivity [15, 53, 54, 55, 56, 57, 58, 59], and detectivity [25, 26, 47, 53, 55, 56, 60, 61], the Ga₂O₃/α‐In₂Se₃ heterojunction in this work demonstrates superior performance, with the exception of MoS₂, ALD/Pt β‐Ga₂O₃, In₂Se₃/GaN, and Ga₂O₃/GaN devices. Furthermore, the figure clearly shows that the Ga₂O₃/α‐In₂Se₃ heterojunction device also exhibits superior performance in terms of on‐off ratio.

FIGURE 5.

(a), (b), (c) The comparison of our work with the responsivity, detectivity, and response speed of previously reported photodetectors.

Following the testing and evaluation of the photodetection performance of Ga₂O₃/α‐In₂Se₃ heterojunction devices, we also investigated their role as optical data receivers in optical communications, as shown in Figure 6a. First, the photocurrent is defined as a high level, representing a value of 1, while the dark current is defined as a low level, representing a value of 0. A signal processing device is placed at the front end of the photodetector to process the binary code signal generating specific characters. After the binary code containing In₂Se₃ information is transmitted by a computer‐controlled program, the photodetector can obtain the corresponding current signal, which is then collected by the detection system. The acquired current values are processed by a signal processing device—an oscilloscope—at the back end. As clearly shown in Figure 6b, the device can accurately distinguish the corresponding signals, convert them into American Standard Code for Information Interchange (ASCII), and subsequently display the special character In₂Se₃ with precision.

FIGURE 6.

(a) Schematic diagram of an ultraviolet optical communication application. (b) The ASCII character being transmitted is “In₂Se₃”.

We fabricated a 5×5 array of Ga₂O₃/α‐In₂Se₃ heterojunction photodetectors, as shown in Figure 7a. By moving a pair of probes connected to a semiconductor parameter analyzer, we simultaneously recorded the photocurrent and dark current of the 5×5 array device under a 10 V bias and a light power density of 300 µW/cm². The obtained data are presented in Figure 7b, which demonstrates outstanding consistency in the photocurrent and dark current of the array device and responsivity [62, 63]. Additionally, statistical analyses of the array's responsivity, detectivity, and on‐off ratio were conducted, as shown in Figure 7c–e. These analyses confirm that the array device exhibits excellent performance stability. Verification of the imaging performance of the Ga₂O₃/α‐In₂Se₃ heterojunction array device is presented in Figure 7f,g. In the aforementioned experiments, the light and dark currents of the array device were recorded and presented in the form of a two‐dimensional current comparison diagram, as shown in Figure 7h. A distinct “I”‐shaped pattern is clearly visible in the figure. Although some deviations exist in individual cells, these can be compensated for by increasing the number of pixels or other means. This directly demonstrates the imaging advantages of the Ga₂O₃/α‐In₂Se₃ heterojunction array device.

FIGURE 7.

(a) A 5×5 array of Ga₂O₃/α‐In₂Se₃ heterojunction devices. (b) A statistical plot of the array's photocurrent and dark current. (c), (d), and (e) show statistical plots of the array's responsivity, detectivity, and on/off ratio, respectively. (f) An image of the array in the on state; (g) an image of the array in the off state. (h) The letter “I” was formed by switching the light source on and off.

4. Conclusion

We fabricated a 5×5 array composed of Ga₂O₃‐In₂Se₃ single‐crystal heterojunction photodetectors. Under 255 nm SBUV light irradiation, the Ga₂O₃‐In₂Se₃ photodetector achieved a switching ratio of 712,434, a detection limit of 1.6×10¹⁴ Jones, and a linear dynamic range of 120 dB. We have also conducted research on encoding for UV light communication. Among the 25 devices in the 5×5 array, all photodetectors exhibited consistent photoelectric performance, indicating that the array holds potential for imaging applications. This study provides new insights for the design of ferroelectric‐enhanced solar‐blind ultraviolet detectors.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: smll74595‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.