Bandgap-Tuned Yttrium-Doped Indium Oxide Alloy Thin Films for High-Performance Solar-Blind Ultraviolet Photodetectors

Abstract

Yttrium oxide (Y₂O₃) has emerged as a key material for advanced solar-blind ultraviolet (SBUV) photodetectors, attributable to its large bandgap energy (~5.5 eV), high dielectric constant, excellent silicon compatibility, and robust thermal stability. To precisely tune its optical bandgap for optimal alignment with the intrinsic solar-blind region, this study prepared Y_1.5In_0.5O₃ ternary alloy films via co-sputtering, achieving an optimized bandgap of 4.70 eV. After optimizing the photosensitive layer, we fabricated a self-powered Pt/Y_1.5In_0.5O₃/p-GaN back-to-back heterojunction SBUV photodetector was fabricated based on the optimized photosensitive layer. Under photovoltaic operation (0 V), the resulting device exhibited impressive performance metrics: a narrow spectral response (FWHM ~50 nm), quick rise/decay times of 30 and 75 ms, respectively, and high operational durability (less than 0.8% photocurrent degradation over 100 cycles). The detector also maintained a low noise current level (2.95 × 10⁻¹² A/Hz^1/2 at 1 Hz) and a low noise-equivalent power (NEP) of 4.42 × 10⁻⁹ W/Hz^1/2, indicating high sensitivity to weak optical signals. These results establish Y_xIn_2−xO₃ ternary alloy as a viable material platform for SBUV detection and provide a new design strategy for developing highly sensitive, low-noise and spectrally selective ultraviolet photodetectors.

1. Introduction

Solar-blind ultraviolet (SBUV, 200–280 nm) photodetectors have seen growing interest over the past few years owing to their intrinsic insensitivity to solar radiation and high selectivity toward deep-UV light [1,2,3]. These features are critical for applications in missile tracking, flame detection, secure space communication, and environmental monitoring [4,5,6,7]. Unlike conventional UV detectors that suffer from parasitic visible and near-UV responses, SBUV photodetectors operate without external optical filters, thereby achieving high signal-to-noise ratios and superior detection accuracy under complex illumination environments [7,8,9,10,11]. To realize such devices, the active materials should exhibit wide bandgaps within the solar-blind spectral range (4.4–6.2 eV), together with high responsivity, fast temporal response, exceptional spectral selectivity and high operational durability.

Over the past decade, several wide-bandgap semiconductors have been explored for SBUV detection, including diamond [12,13], AlGaN [14,15], and MgZnO alloys [16,17]. Although these materials have achieved remarkable progress, each suffers from intrinsic limitations. Diamond photodetectors are hindered by high synthesis costs and defect-related issues [18]. AlGaN offers a tunable bandgap but demands high-temperature epitaxial growth and faces lattice mismatch challenges with common substrates, which hinder scalable fabrication [15]. Similarly, MgZnO alloys can extend the absorption edge into the solar-blind region but are prone to phase segregation, poor thermal stability, and limited compositional control [19,20,21]. These challenges underscore the need for alternative material systems that combine bandgap tunability, structural stability, and fabrication compatibility.

Yttrium oxide (Y₂O₃) has recently emerged as a compelling candidate for SBUV photodetectors due to its wide bandgap (~5.5 eV), high dielectric constant, outstanding thermal stability, and chemical robustness [22,23,24,25,26]. Furthermore, Y₂O₃ exhibits excellent compatibility with silicon and III-nitride semiconductors, allowing for versatile integration into existing microelectronic and optoelectronic platforms [27,28,29]. Nevertheless, the fixed bandgap of binary Y₂O₃, while within the solar-blind range, is not optimally aligned with the lower-energy edge of the atmospheric transmission window, potentially limiting photon absorption efficiency and responsivity for the longest usable SBUV wavelengths. Fine-tuning the optical bandgap of Y₂O₃ while preserving its crystalline stability and favorable dielectric properties is essential for achieving efficient solar-blind detection.

Ternary oxide alloys offer a powerful strategy for bandgap engineering, enabling systematic modulation of electronic structure and optical properties of wide-bandgap materials [30]. By introducing cations with smaller bandgaps into the Y₂O₃ lattice, one can achieve continuous bandgap tunability, improved carrier transport, and tailored defect states. Indium oxide (In₂O₃), with a smaller bandgap (~3.6 eV) and a similar cubic structure, is an ideal candidate for alloying with Y₂O₃ [31,32,33,34,35,36]. The underlying mechanism lies in the band alignment: as the Y/In ratio varies, the conduction band minimum shifts significantly while the valence band maximum remains relatively stable, thereby enabling continuous bandgap modulation. The formation of Y_xIn_2−xO₃ ternary alloys can bridge the bandgap between Y₂O₃ and In₂O₃, enabling precise tuning within the solar-blind range while retaining excellent crystallinity and stability. Moreover, such ternary oxide systems offer a promising pathway to overcome the limitations of conventional binary semiconductors by combining the merits of both a wide bandgap and improved electrical transport characteristics.

In this work, we report the controllable synthesis of Y_xIn_2−xO₃ ternary alloy thin films via a co-sputtering process, achieving precise composition and bandgap modulation. Based on the optimized material, a self-powered Pt/Y_xIn_2−xO₃/p-GaN back-to-back heterojunction SBUV photodetector was fabricated. Operating at 0 V bias, the device exhibits rapid response (rise/decay times of 30/75 ms) and excellent stability, with less than 0.8% photocurrent degradation over 100 cycles. Furthermore, the photodetector demonstrates an ultralow noise current of 2.95 × 10⁻¹² A/Hz^1/2 at 1 Hz and a low noise-equivalent power (NEP) of 4.42 × 10⁻⁹ W/Hz^1/2, indicating exceptional sensitivity to weak optical signals. This work validates Y_xIn_2−xO₃ as a robust and tunable material platform for high-performance, low-noise SBUV photodetection and introduces a heterojunction design paradigm that synergizes bandgap-engineered absorption, built-in potential-driven charge separation, and intrinsically low dark current.

2. Materials and Methods

2.1. Substrate Pre-Treatment

To obtain clean substrate surfaces, p-Si, p-GaN, and quartz pieces were ultrasonically cleaned in acetone, ethanol, and ultrapure water for 10 min each, followed by drying with high-purity nitrogen gas.

2.2. Y_xIn_2−xO₃ Film Deposition

Y_xIn_2−xO₃ ternary alloy thin films were fabricated on cleaned substrates using radio-frequency (RF) magnetron co-sputtering. The source materials were high-purity (99.99%) Y₂O₃ and In₂O₃ ceramic targets. To modulate the film stoichiometry, the Y₂O₃ target power was kept constant at 150 W, whereas the In₂O₃ power was varied (20 W, 30 W, 40W). The deposition was carried out for 120 min at ambient temperature, under a constant chamber pressure of 0.5 Pa. The working atmosphere was Ar (20 sccm) and O₂ (5 sccm).

2.3. Photodetector Fabrication

An ultra-thin Pt layer was first deposited onto the Y_xIn_2−xO₃ film surface through a shadow mask by sputtering, serving as the photocarrier collection window (deposition parameters: power 50 W, duration 30 s, working pressure 0.5 Pa, Ar flow rate 30 sccm). Subsequently, a thicker Pt electrode was sputtered onto the ultra-thin Pt layer using a second mask under identical conditions, except for an extended deposition time of 3 min. For the bottom contact, an indium ohmic electrode was bonded to the p-GaN substrate.

2.4. Material and Device Characterization

Surface and cross-sectional morphologies were examined using a ZEISS Sigma 300 field emission scanning electron microscope (SEM, Carl Zeiss, Oberkochen, Germany). Elemental distribution and quantitative composition analysis were performed via energy-dispersive X-ray spectroscopy (EDS) attached to the SEM system. The ultraviolet–visible transmission spectra of the films was recorded using a Shimadzu UV-2600i spectrophotometer (Shimadzu Corporation, Kyoto, Japan), and the optical bandgap was calculated by the Tauc plot method. Three-dimensional surface topography and root-mean-square (RMS) roughness were assessed by atomic force microscopy (AFM, Bruker Dimension Icon, Bruker, Santa Barbara, CA, USA). The spectral responsivity of the fabricated devices was calibrated using a Zolix DSR500 system (Zolix Instruments Co., Ltd., Beijing, China). The optoelectronic properties, including current–voltage (I-V) characteristics and photoresponsivity, were measured with a Keithley 4200A-SCS parameter analyzer (Keithley Instruments, Solon, OH, USA). Low-frequency noise measurements were conducted using a PRIMARIUS FS-Pro semiconductor parameter analyzer (Primarius Technologies, Shanghai, China). Steady-state and time-resolved photoluminescence (TRPL) spectra were acquired using an Edinburgh Instruments FLS1000 spectrophotometer (Edinburgh Instruments, Livingston, UK). Unless otherwise specified, films for structural characterization such as SEM and AFM are deposited on Si substrates, films for UV-Vis characterization are deposited on quartz substrates, and films for constructing SBUV photodetectors and subsequent photoelectric performance tests are fabricated on p-GaN substrates.

3. Results and Discussion

3.1. Precise Bandgap Engineering and Microstructural Analysis

As schematically illustrated in Figure 1a, with the sputtering power of Y₂O₃ fixed at 150 W, a series of Y_xIn_2−xO₃ ternary alloy films with varying compositions was successfully fabricated by regulating the sputtering power of the In₂O₃ target to 20 W, 30 W and 40 W via RF magnetron co-sputtering. The UV-Vis transmission spectra of all films are presented in Figure S1. The optical bandgap values under each condition were derived from the Tauc plot method with the following equation [37], as summarized in Figure 1b.

$${\left(\mathsf{\alpha }\mathrm{h}\mathsf{\nu }\right)}^2 =\mathrm{A}(\mathrm{h}\mathsf{\nu }-\mathrm{Eg})$$

(1)

where α is the absorption coefficient, hν is photon energy, and A is a material-dependent. The bandgap Eg was obtained by the linear fitting of the curves and extrapolating to (αhν)² = 0. Notably, as shown in Figure 1c, the optical bandgap of the Y_xIn_2−xO₃ films exhibits a clear linear increasing trend with rising In₂O₃ sputtering power. This trend indicates that incorporating In₂O₃ effectively modulates the bandgap continuously (the specific mechanism is shown in Figure S2), thereby confirming the feasibility and effectiveness of the co-sputtering technique for precise bandgap tuning in wide-bandgap oxide semiconductors. To achieve efficient detection in the SBUV region, we selected the film deposited with an In₂O₃ sputtering power of 30 W, which possesses a bandgap value of 4.70 eV and an absorption edge located within the SBUV range, as the photosensitive absorption layer. The transmittance of this film in the 200–280 nm wavelength range is between 10.2% and 77.3%, indicating strong absorption characteristics suitable for SBUV photodetector applications. The surface morphology, elemental distribution, and chemical composition of the selected films were systematically investigated using SEM and EDS mapping. The corresponding results are compiled in Figure 1d. SEM images reveal that the film surface is uniform and dense. EDS elemental mapping confirms the homogeneous spatial distribution of Y, In, and O throughout the film. Combined with the quantitative EDS analysis provided in Figure S3, the atomic ratio of Y to In was determined to be approximately 3:1, suggesting the chemical composition can be accurately represented as Y_1.5In_0.5O₃. The cross-sectional SEM image of the Y_1.5In_0.5O₃ film (Figure 1e) further demonstrates its compactness and structural uniformity. The measured thickness is about 150 nm, sufficient to ensure effective absorption of incident SBUV radiation. Furthermore, AFM measurements (Figure 1f) indicate a RMS roughness of only about 1.77 nm, highlighting the excellent surface flatness of the film, which provides an ideal substrate for subsequent deposition of ultra-thin Pt electrodes.

Figure 1. (a) Schematic illustration of the RF magnetron co-sputtering process for fabricating Y_xIn_2−xO₃ thin films. (b) Determination of the optical bandgap for Y_xIn_2−xO₃ films using Tauc plots. (c) Variation in the optical bandgap of Y_xIn_2−xO₃ films as a function of the In₂O₃ sputtering power. (d) Surface SEM image and corresponding EDS elemental mapping of Y, In, and O for the Y_1.5In_0.5O₃ film. (e) Cross-sectional SEM image of the Y_1.5In_0.5O₃ film deposited on a silicon substrate. (f) AFM topographic image of the Y_1.5In_0.5O₃ film.

3.2. Heterojunction Design and Self-Powered Operation

We developed and assembled a Pt/Y_1.5In_0.5O₃/p-GaN back-to-back heterojunction SBUV photodetector using the previously prepared Y_1.5In_0.5O₃ film. The overall structure is schematically shown in Figure 2a, where Y_1.5In_0.5O₃ functions as the photosensitive absorption layer, p-GaN serves as the substrate and forms a heterojunction with Y_1.5In_0.5O₃, the ultra-thin Pt layer functions as the photocarrier collection layer, the thick Pt pad serves as the top electrode, and the In ohmic contact electrode serves as the bottom electrode. An optical micrograph of the device is presented in Figure S4a, from which the area of the ultra-thin Pt window is estimated to be approximately 1.125 mm². Furthermore, the UV-vis transmission spectrum of an identical ultra-thin Pt layer deposited on a quartz substrate is given in Figure S4b, revealing a high transmittance of around 90% over the broad wavelength range of 200–800 nm. These results demonstrate the excellent optical transparency of the Pt layer across the SBUV to visible region, ensuring minimal absorption of light before it reaches the underlying Y_1.5In_0.5O₃ photosensitive layer. Coupled with the intrinsically high electrical conductivity of Pt, the ultra-thin Pt layer not only facilitates efficient transport of photogenerated carriers but also maintains high incident photon flux, thereby synergistically optimizing the detection efficiency. To elucidate the internal energy band alignment and carrier transport mechanism, ultraviolet photoelectron spectroscopy (UPS) was employed to characterize the Y_1.5In_0.5O₃ film. The secondary electron cut-off region and the valence band onset region are shown in Figure 2(b) and Figure 2(c), respectively. From the UPS data, the work function of Y_1.5In_0.5O₃ was determined to be 3.39 eV, and the energy difference between the valence band maximum and the Fermi level was found to be 2.99 eV. By combining the optically derived bandgap from Figure 1b, the complete band structure of Y_1.5In_0.5O₃ was unambiguously established. Figure 2d depicts the initial band alignment among isolated Pt, Y_1.5In_0.5O₃, and p-GaN before contact. Figure 2e illustrates the proposed equilibrium band diagram after the formation of the intimate back-to-back heterojunction. Upon junction formation, Fermi-level alignment induces band bending near both the Pt/Y_1.5In_0.5O₃ interface and the Y_1.5In_0.5O₃/p-GaN interface. Consequently, back-to-back potential barriers emerge, substantially suppressing the injection of majority carriers at zero bias and hence markedly reducing the dark current. Under SBUV illumination at 0 V bias, photogenerated electron–hole pairs are generated within the Y_1.5In_0.5O₃ photosensitive layer. Driven by the built-in electric field at the Pt/Y_1.5In_0.5O₃ Schottky interface, these carriers undergo spatial separation. Specifically, holes migrate toward and are collected by the Pt electrode, whereas electrons must overcome the barrier at the Y_1.5In_0.5O₃/p-GaN interface, eventually being received by the bottom In electrode, completing the conversion of optical signals into electrical output.

Figure 2. (a) Schematic diagram of the device structure of the Y_1.5In_0.5O₃-based SBUV photodetector; (b) Secondary electron cut-off region of the UPS spectrum of the Y_1.5In_0.5O₃ film; (c) Valence band onset region of the UPS spectrum of the Y_1.5In_0.5O₃ film; (d) Equilibrium band diagrams of isolated Pt, Y_1.5In_0.5O₃, and p-GaN; (e) Energy band diagram illustrating the back-to-back heterojunction formed by the intimate contact among Pt, Y_1.5In_0.5O₃, and p-GaN.

3.3. Comprehensive Photodetection Performance and Noise Analysis

Figure 3a presents the current–voltage (I-V) characteristics of the device based on Y_1.5In_0.5O₃ film under both dark conditions and 255 nm ultraviolet illumination. The results demonstrate that at bias voltages of 0.15 V and −2 V, the dark currents of the device are 9.3 × 10⁻¹³ A and 6.2 × 10⁻¹⁰ A, respectively, while the photocurrents are 4.4 × 10⁻⁹ A and 6.5 × 10⁻⁸ A, respectively. The self-biasing behavior of the device in the dark state can be ascribed to the built-in electric field at the Pt/Y_1.5In_0.5O₃ Schottky interface, as depicted in Figure 2e. Moreover, Figure S5 provides direct evidence supporting the existence of this built-in electric field. The photo-to-dark current ratio (PDCR) is defined by the following equation [38]:

$$\mathrm{PDCR} = {\displaystyle \frac{{\mathrm{I}}_{\mathrm{photo}}-{\mathrm{I}}_{\mathrm{dark}}}{{\mathrm{I}}_{\mathrm{dark}}}}$$

(2)

where I_photo represents the current under SBUV irradiation, and I_dark denotes the current in dark conditions. Calculations reveal that at 0.15 V and −2 V bias, the PDCR values are approximately 4.7 × 10³ and 2.4 × 10¹, respectively, yielding a substantial benefit for its use highly sensitive SBUV detection applications. Figure 3b presents the spectral responsivity of the Y_1.5In_0.5O₃ device at 0 V bias, showing a peak responsivity of 0.67 mA/W at 230 nm with a full width at half maximum (FWHM) of ~50 nm. This indicates its high selectivity towards ultraviolet light around 254 nm, which is essential for accurate identification of SBUV signals. Figure S6a shows the photoluminescence (PL) spectrum of the Y_1.5In_0.5O₃ film, indicating that within the excitation wavelength (EX) range of 400–800 nm, the emission wavelength (EM) exhibits a peak at 565 nm. Figure S6b displays the photoluminescence excitation (PLE) spectrum of the Y_1.5In_0.5O₃ film, revealing that excitation in the 240–380 nm range effectively induces emission at 565 nm, with EX = 365 nm being the most efficient excitation wavelength. Figure 3c and Figure S6c show the time-resolved PL (TRPL) decay curves plotted in both linear and logarithmic coordinates. By applying Equations (3) and (4) for fitting and analysis, the average decay lifetime of photogenerated carriers in the Y_1.5In_0.5O₃ film is determined to be 4.67 ms. This prolonged lifetime suggests suppressed recombination of photogenerated carriers in the Y_1.5In_0.5O₃ film, thereby laying a foundation for its promising optoelectronic response performance.

$$\mathrm{Normalized} \mathrm{Intensity}={ \mathrm{A}}_1{\mathrm{e}}^{{\displaystyle \frac{-\mathrm{t}}{\mathsf{\tau }1}}}+{ \mathrm{A}}_2{\mathrm{e}}^{{\displaystyle \frac{-\mathrm{t}}{\mathsf{\tau }2}}}$$

(3)

$${\mathsf{\tau }}_{\mathrm{average} }={\displaystyle \frac{{\mathrm{A}}_1{\mathsf{\tau }}_1^2+{ \mathrm{A}}_2{\mathsf{\tau }}_2^2}{{\mathrm{A}}_1+{ \mathrm{A}}_2}}$$

(4)

Figure 3. (a) Current–voltage (I-V) characteristics of the Y_1.5In_0.5O₃-based device measured under dark conditions and under 255 nm (200 μW) ultraviolet illumination; (b) Spectral responsivity of the Y_1.5In_0.5O₃-based device at 0 V bias; (c) Time-resolved photoluminescence spectrum of the Y_1.5In_0.5O₃ film; (d) I-V characteristics of the Y_1.5In_0.5O₃-based device under 255 nm ultraviolet illumination with varying optical power intensities; (e) Variation in the V_OC and I_SC of the Y_1.5In_0.5O₃-based device with optical power. (f) Noise power spectral density as a function of frequency at different biases.

Figure 3d compares the current-voltage (I–V) characteristics of the Y_1.5In_0.5O₃ device under 255 nm ultraviolet illumination at varying optical power intensities. The results demonstrate a marked increase in photocurrent with rising optical power. The dependence of the open-circuit voltage (V_OC) and short-circuit current (I_SC) on the optical power, extracted from Figure 3d, is summarized in Figure 3e. It is evident that both V_OC and I_SC increase with enhanced optical power. As a key indicator of photovoltaic performance [39,40,41], a high V_OC is essential for achieving high power conversion efficiency in Y_1.5In_0.5O₃-based devices. Through fitting analysis, the exponent describing the power-law dependence of I_SC on optical power is determined to be 1.11, which is close to the ideal value of 1, suggesting highly stable output characteristics under photovoltaic operation [42].

The detection limit of a photodetector is characterized by its NEP, which is defined as the incident optical power required to produce a signal-to-noise ratio of 1 in a 1 Hz bandwidth. This parameter quantifies the sensitivity to weak light signals and is given by:

$$\mathrm{NEP} = {\displaystyle \frac{{\mathrm{I}}_{\mathrm{noise}}}{\mathrm{R}}}$$

(5)

where I_noise denotes the noise current measured in the dark, and R is the responsivity. The noise current I_noise is given by

$${\mathrm{I}}_{\mathrm{noise} }=\sqrt{\int {\mathrm{S}}_{\mathrm{i}}^2\mathrm{df}}$$

(6)

As derived from Figure 3b, the Y_1.5In_0.5O₃ device exhibits a responsivity of 0.34 mA/W under 255 nm illumination. Figure 3e presents the noise power spectral density (S_i-f) characteristics at different biases. At 0 V bias, the measured noise current (I_noise) and the corresponding NEP are 2.95 × 10⁻¹² A/Hz^1/2 and 4.42 × 10⁻⁹ W/Hz^1/2, respectively, indicating its a low noise floor and superior capability for weak-light detection.

Furthermore, the total noise current (I_noise) is primarily governed by four distinct components: shot noise (I_s), thermal noise (I_t), flicker noise (I_1/f), and generation-recombination noise (I_g−r), as given by Equation (7).

$${\mathrm{I}}_{\mathrm{noise} }=\sqrt{{{\mathrm{I}}_{\mathrm{s}}}^2+{{\mathrm{I}}_{\mathrm{t}}}^2+{{\mathrm{I}}_{{\displaystyle \frac{1}{\mathrm{f}}}}}^2+{{\mathrm{I}}_{\mathrm{g}-\mathrm{r}}}^2}$$

(7)

As evidenced by the data in Figure 3e and Figure S7, no discernible noise plateau is observed within the measured frequency range. Instead, the logarithmic noise power spectral density, lg(S_i), decreases linearly with increasing logarithmic frequency, lg(f), indicating that S_i adheres to a power-law relationship with f, expressed as:

$${\mathrm{S}}_{\mathrm{i}}\propto 1/{\mathrm{f}}^{\mathsf{\gamma }}$$

(8)

The extracted flicker noise exponents (γ) at 0 V, 0.5 V, and 1 V are 0.973, 0.975, and 0.991, respectively, strongly indicating that the device noise is predominantly governed by the flicker noise component (I_1/f). Furthermore, as the bias voltage increases from 0 V to 1 V, the overall noise amplitude rises by approximately 1.5 orders of magnitude, suggesting a direct correlation between the noise sources and the carrier transport process.

3.4. Dynamic Response and Operational Durability

To further investigate the photodetection performance of the device, this study measured the multi-cycle current-time (I-T) characteristics of the Y_1.5In_0.5O₃ device under two sets of conditions: (i) illumination with 255 nm ultraviolet light at 200 μW under bias voltages ranging from −0.15 V to 0.15 V, and (ii) at 0 V bias under 255 nm ultraviolet illumination with optical power densities varying from 129 μW to 233 μW. The results are presented in Figure 4(a) and (b), respectively. The increase in photocurrent with increasingly negative bias can be attributed to the built-in electric field at the Pt/Y_1.5In_0.5O₃ Schottky interface, which is consistent with Figure 3a. Under negative bias, which aligns with the direction of the built-in field, the efficient separation of photogenerated carriers is facilitated, resulting in a higher photocurrent, whereas a positive bias produces the opposite effect. The device exhibited reproducible and stable photoresponse behavior under varying bias and illumination conditions, confirming its excellent stability and repeatability. To quantitatively analyze the influence of optical power on the photocurrent at a fixed 0 V bias, the photocurrent from each response cycle under different light intensities was extracted and plotted in Figure S8a, showing high consistency across cycles, with the photocurrent increasing nearly linearly with rising optical power. Similarly, to evaluate the effect of bias voltage on the photocurrent under fixed ultraviolet illumination, the photocurrent from each cycle under different bias voltages was extracted and summarized in Figure S8c. The results reveal a decreasing trend in photocurrent with increasing bias voltage, a behavior consistent with the I–V characteristics shown in Figure 2a. To verify the critical role of the ultrathin Pt layer in carrier collection, we fabricated and tested a control device without the ultrathin Pt layer. Under identical illumination conditions (255 nm, 200 μW), the photocurrent of the control device is only at the pA level (Figure S9). This pronounced difference directly demonstrates that the ultrathin Pt layer enhances carrier extraction efficiency by providing an effective charge-collection pathway, thereby ensuring the superior photoresponse performance of the device.

Figure 4. (a) I-T characteristics of the Y_1.5In_0.5O₃ device under 255 nm ultraviolet illumination (200 μW) at different bias voltages; (b) Multi-cycle I-T curves of the device measured at 0 V bias under 255 nm ultraviolet illumination with different optical power densities; (c) Single-cycle I-T curve of the device, representing a typical cycle selected from (b), where the response time is defined by the time taken for the photocurrent to decay between 90% and 10% of the peak magnitude; (d–f) Long-term performance of the device.

To further analyze the influence of ultraviolet intensity and bias voltage on the response time of the device, a single-cycle I-T curve was extracted from Figure 4b and is presented in Figure 4c. Under 0 V bias and 200 μW, 255 nm ultraviolet illumination, the device exhibits rise and decay times of approximately 30 ms and 75 ms, respectively. This rapid response is attributed to the pyroelectric effect, wherein the transient current spikes generated upon the initiation and termination of illumination effectively reduce the overall response time [43,44,45]. To systematically evaluate the statistical trends of response time with respect to bias voltage and light intensity, the rise and decay times from each response cycle in Figure 4a,b were extracted and are summarized in Figure S8b,d. The results confirm that the response time remains below 110 ms across all measured conditions.

To quantify the operational durability of the device, the I-T characteristics of the Y_1.5In_0.5O₃ device were measured over 100 consecutive cycles at 0 V bias, as shown in Figure 4e. The photocurrent response exhibits remarkable stability and repeatability throughout the test. Representative cycles, specifically the 1st–5th and the 96th–100th, are displayed in Figure 4(d) and (f), respectively. Remarkably, the device exhibits a minimal photocurrent degradation of only 0.8% after 100 cycles. These results collectively indicate that the Y_1.5In_0.5O₃-based device possesses high response speed, excellent durability and repeatability, thereby rendering it a promising candidate for reliable deployment in practical applications.

4. Conclusions

In summary, we have successfully synthesized Y_xIn_2−xO₃ bandgap-engineered ternary alloy thin films via a controllable RF magnetron co-sputtering process. By regulating the sputtering power of the In₂O₃ target, the optical bandgap was precisely engineered, and the Y_1.5In_0.5O₃ film with an optimal bandgap of 4.70 eV was employed as the photoresponse medium for constructing a Pt/Y_1.5In_0.5O₃/p-GaN back-to-back heterojunction SBUV photodetector.

The fabricated device operates efficiently at 0 V bias, demonstrating a combination of high spectral selectivity (FWHM~50 nm), fast response (rise/decay times of 30/75 ms), and excellent operational durability (only 0.8% photocurrent variation over 100 cycles). Moreover, the detector exhibits an ultralow noise current of 2.95 × 10⁻¹² A/Hz^1/2 and a low noise-equivalent power (NEP) of 4.42 × 10⁻⁹ W/Hz^1/2 underscoring its superior sensitivity to weak optical signals.

These results validate Y_xIn_2−xO₃ as a highly promising material platform for high-performance, self-powered SBUV photodetection. The proposed heterojunction architecture, which synergistically combines bandgap-engineered absorption, built-in potential-driven carrier separation, and low dark current, offers a viable and effective design strategy for future high-sensitivity, low-noise, and spectrally selective UV photodetectors.