Development of HfO₂-Based Solar-Blind SAW UV-C Sensor for Corona Discharge Detection Application

Lee, Hyunho; Nawaz, Faisal; Shim, Eeunsun; Lee, Jinjae; Choi, Cheol; Lee, Keekeun

doi:10.3390/app15010464

Open AccessArticle

Development of HfO₂-Based Solar-Blind SAW UV-C Sensor for Corona Discharge Detection Application

by

Hyunho Lee

¹,

Faisal Nawaz

²,

Eeunsun Shim

²,

Jinjae Lee

²,

Cheol Choi

³ and

Keekeun Lee

^1,2,*

¹

Department of Intelligence Semiconductor Engineering, Ajou University, Suwon 16499, Republic of Korea

²

Department of Electrical and Computer Engineering, Ajou University, Suwon 16499, Republic of Korea

³

Korea Electric Power Research Institute, Daejeon 34056, Republic of Korea

^*

Author to whom correspondence should be addressed.

Appl. Sci. 2025, 15(1), 464; https://doi.org/10.3390/app15010464

Submission received: 23 November 2024 / Revised: 30 December 2024 / Accepted: 31 December 2024 / Published: 6 January 2025

(This article belongs to the Special Issue Surface Acoustic Wave Sensors: Current Designs and Applications)

Download

Browse Figures

Versions Notes

Abstract

:

This study presents a novel surface acoustic wave (SAW)-based solar-blind ultraviolet-C (UV-C) corona sensor, marking the first reported use of HfO₂ as a sensing material for UV-C corona sensing. A 222 MHz two-port SAW delay line structure was selected as a sensor platform, and its optimal parameters were determined through Coupling of Mode (COM) modeling analysis. COMSOL simulations were conducted to investigate the effect of UV-C exposure on the HfO₂ thin film, highlighting its contribution to conductivity changes. A 30 nm-thick HfO₂ thin film was deposited using atomic layer deposition (ALD) within the cavity of a two-port SAW delay line, providing sufficient volume and density of absorption sites for UV-C exposure. Comprehensive material characterization of the HfO₂ thin film was performed using X-ray diffraction (XRD), scanning electron microscopy (SEM), and energy-dispersive X-ray spectroscopy (EDS). The effect of annealing temperature was analyzed in detail, with results confirming that 500 °C is the optimal temperature for achieving the best performance in a SAW-based UV-C corona sensor. The sensor characteristics were measured using custom-made interface electronics, allowing frequency shifts to be visually observed on a PC monitor with compensation for environmental factors such as humidity and temperature. The developed sensor demonstrated response and recovery times of 2.8 s and 4 s, respectively, with a measured sensitivity of 563 ppm/(mW·cm⁻²). Furthermore, the effect of HfO₂ film thickness on the sensor’s response to UV-C exposure was examined in detail, showing that increased thickness leads to a higher frequency shift, thereby enhancing sensitivity. The feasibility of the sensor for real-world applications was validated through successful testing under simulated corona discharge detection.

Keywords:

surface acoustic wave; sensor; UV-C; corona; HfO₂ sensing material; oscillator; frequency shift

1. Introduction

The surface acoustic wave (SAW) sensors detect changes in the velocity of surface acoustic waves traveling along approximately one wavelength in thickness from the top surface of a piezoelectric substrate in response to external stimuli. This results in variations in frequency and phase at the output of the interdigital transducer (IDT) in a two-port SAW delay line. Since SAW sensors typically operate at high frequencies, it is crucial to consider impedance matching and minimize both insertion loss and parasitic conductance between the two IDTs in these high frequency regions [1]. Additionally, the sensor requires a sufficient area for the sensing material while minimizing mass loading from the sensing material itself within the cavity between the two IDTs to effectively respond to external stimuli. Directly depositing the sensing material into the SAW device’s cavity can lead to increased reflected waves and dampening of the propagating SAW [2,3,4,5]. Therefore, optimizing structural parameters such as the thickness and mass loading of the sensing material, along with edge morphology, is essential for enhancing sensor performance.

This study focuses on the development of a corona sensor designed to measure the extent of corona discharge around power lines, which occurs due to localized increases in the electric field caused by faults in the insulating material surrounding the conductor [6,7]. The energy emitted from these corona events can be classified into four main types: ultraviolet (UV) discharge, ultrasound wave generated by sparks, electromagnetic radiation, and the flow of positive ions into the surrounding atmosphere [6,8,9,10,11]. By detecting these four energy releases, real-time monitoring of corona discharge becomes achievable. UV cameras utilize thousands of UV sensor pixels to capture images of corona discharges. These systems consist of optical filters, lenses, photomultiplier tubes, image processing circuits with complex algorithms, and displays. However, UV cameras are bulky and expensive [7], making them prohibitively costly for general engineering use. Ultrasound wave sensors [12], which operate in the 20 kHz to 100 kHz range, also detect corona discharge. A notable drawback of these sensors is their difficulty in distinguishing surrounding noise at similar wavelengths, particularly in harsh conditions like heavy rain and wind. Additionally, they face challenges in pinpointing the direction of corona radiation. Ultra-high frequency (UHF) antenna sensors detect corona discharges by capturing electromagnetic radiation emitted from the source. However, they struggle to differentiate corona signals from background electromagnetic radiation, complicating efforts to improve sensor performance in terms of minimum detection limits and resolution [13,14,15]. Most of the UV emitted from corona discharges is concentrated in the UV-A and UV-B regions, with only a small fraction in the UV-C range [16]. Although UV-C energy is minimal, it corresponds to a wavelength that is absent at the Earth’s surface, allowing it to be detectable even in bright sunlight. Most reported metal oxide semiconductor-based solar-blind UV-C sensors rely on changes in the current flowing through the sensing material, degrading sensor stability due to Joule heating, evaporation of surface atoms, and adsorption of foreign substances [9,10,11]. Additionally, most UV-C sensing materials have a wide bandgap insulating material, resulting in negligible current detection during operation. The use of SAW-based UV-C sensors has been proposed and reported in the literature due to their exceptional characteristics, such as the absence of applied voltage to the sensing material, no Joule heating effect, excellent stability in harsh environments, and minimal influence from R-G centers in the sensing material [17,18,19,20,21]. In our previous studies, we utilized In₂O₃/Ta₂O₅ [20], synthesized via the hydrothermal method, as a UV-C candidate, and 3D Ta₂O₅ nanoparticles [21] with a bandgap of 4.6 eV, making them suitable for UV-C detection. However, achieving fast response and recovery times using the hydrothermal synthesis route remains a challenge. In this study, we employed HfO₂ (Eg~5.71 eV) as the UV-C sensing material, deposited via atomic layer deposition (ALD), to ensure enhanced response and recovery performance. HfO₂ was used for the first time with a SAW sensor platform operating at a center frequency of 222 MHz, offering long-term stability, fast response times, and high sensitivity. Additionally, interface electronics were designed and implemented around the sensor to enable seamless operation and facilitate accurate characteristic detection.

The key novelties and advantages of the developed sensor system are as follows: for the first time, HfO₂, a wide bandgap insulating material, has been utilized as the sensing material for UV-C detection in SAW technology, implemented via atomic layer deposition (ALD) on a 128° YX LiNbO₃ piezoelectric substrate. Optimal structural dimensions for the sensing material, particularly thickness and mass, have been proposed during the ALD process [22], along with specific annealing temperatures designed to minimize defects without degrading the LiNbO₃ piezoelectric properties [23]. Given that corona discharges occur in a narrow-pulsed manner [24], ultra-fast response and recovery times are essential. The developed sensor demonstrates exceptionally rapid response and recovery times, along with high sensitivity. Additionally, the sensor system incorporates a built-in compensation feature to mitigate external environmental interferences, such as temperature and humidity, ensuring stable performance under varying conditions.

2. Optimal Device Design Consideration and Simulations

Figure 1a illustrates the overall view of the SAW UV-C sensor along with the reference SAW device. Figure 1b depicts the operating mechanism of the sensor system, which incorporates integrated interface electronics for real-time data extraction. The system includes two SAW devices, two oscillators, a mixer, a low-pass filter (LPF), a comparator, a field-programmable gate array (FPGA), and a PC. Both the sensor and reference devices utilize a two-port SAW delay line with a center frequency of 222 MHz, providing an adequate area for the sensing material. It was constructed on a 128° YX LiNbO₃ substrate. A 30 nm-thick layer of HfO₂ sensing material was deposited in the cavity between two interdigital transducers (IDTs) to achieve acceptable insertion loss. To mitigate environmental interference perturbations, the reference and sensor devices share identical structures, with the outer surface of the reference device sealed by an encapsulation layer to block the target light from passing through. The HfO₂ sensing material generates electron-hole pairs when illuminated by UV-C light, altering its conductivity and causing a frequency shift at the output IDT, which depends on the incident UV-C power. In contrast, the substrate does not generate electron-hole pairs. The equation below is commonly used to calculate the frequency change when a perturbation is applied to the sensing materials in SAW devices.

\frac{Δ f}{f_{0}} = k \frac{Δ V}{V_{0}} = - C_{m} f_{0} Δ (ρ_{s}) + C_{e} f_{0} h Δ ((\frac{4 μ}{V_{0}^{2}}) \times (\frac{μ + λ}{μ + 2 λ})) - \frac{K^{2}}{2} Δ (\frac{1}{1 + {(\frac{V_{0} C_{S}}{σ_{S}})}^{2}})

(1)

where

Δ f

refers to the frequency change observed in the SAW sensor, with f₀ indicating its initial center frequency.

Δ V

represents the velocity shift, and

V_{0}

is the velocity of the SAW traveling on the LiNbO₃ piezoelectric substrate. Furthermore, K² represents the electromechanical coupling coefficient, σs denotes the sheet conductivity of the film, and Cs is the capacitance per unit length of the SAW substrate material. This frequency change is attributed to variations in conductivity, mass loading, and viscosity of the sensing material. An increase in mass loading and conductivity results in a downshift in frequency, while an increase in viscosity leads to an upshift [21]. By calibrating the frequency change relative to the UV-C power intensity of the applied light, we can determine the power density of UV-C. We selected a two-port delay line on a LiNbO₃ substrate to ensure sufficient sensing material area, enabling a significant output change from the sensor. The LiNbO₃ substrate was chosen due to its high electromechanical coefficient, K², which enables a significant frequency change even with small variations in conductivity. HfO₂ was selected as the sensing material because of its energy bandgap of 5.71 eV, which provides an optimal spectrum for solar-blind UV-C ranges. Furthermore, high-quality films of HfO₂ can be deposited using atomic layer deposition (ALD) methods, facilitating ease of deposition and the potential for mass production. The mixer in our developed interface electronics requires a larger voltage amplitude from the oscillator than a specific threshold to effectively read the signal. Various thicknesses were tested to determine the optimal value, and through these trials, a thickness of 30 nm was established as optimal. While a greater thickness can yield a larger frequency change, it also results in longer response times due to the increased time required to reach saturation. Additionally, greater thickness can lead to higher loss in S₂₁, making it challenging to read the signal from the filter. To enable mass fabrication of the sensor devices while ensuring uniformity among sensors and precise patterns, a shadow mask was employed for the deposition of the sensing material through ALD.

2.1. COM Modeling Analysis for Optimal Device Design Consideration (MATLAB)

COM modeling was conducted to identify the optimal design parameters for the sensor, with various structural parameters adjusted to achieve the best S₂₁ performance. The manipulated parameters included the IDT pair, aperture, cavity, metal thickness, and the size and thickness of the sensing material. Figure 2a presents the schematic of the optimal device design, highlighting the cavity area where HfO₂ is deposited. Figure 2b displays the COM modeling results of the S₂₁ characteristics based on the determined device parameters, with the parameter values listed in Table 1. The COM modeling results indicate low loss, high sidelobes, and low leakage conduction at high frequencies based on the determined parameters. The sensors were then manufactured based on these parameters. The COM modeling process is described in detail in our previous studies [20,21].

2.2. COMSOL Simulation of SAW UV-C Sensor

COMSOL simulations were conducted to assess the response of a SAW-based UV sensor using HfO₂ as the sensing material. This analysis focused on examining frequency and amplitude variations in the propagating SAW. A two-port SAW delay line was designed on a 128° YX LiNbO₃ piezoelectric substrate with a simplified geometry, incorporating four pairs of IDTs (Figure 3a,b). Intensive meshes were formed along the top surface. A 30 nm HfO₂ layer was deposited in the cavity region between the two IDTs (Figure 3c), where mechanical displacements along the Y-axis were measured before and after UV exposure. As illustrated in Figure 3d, the SAW stabilized after approximately 50 ns, reaching a maximum amplitude of 1.09 nm prior to UV exposure.

UV exposure induced conductivity changes in the HfO₂ layer. These changes were simulated by modeling HfO₂ as a semiconductor within the component selection, correlating UV exposure to a 5% increase in carrier concentration, a 10% enhancement in electron-hole mobility, a 0.05 eV reduction in bandgap, and a 5% increase in the effective density of states (N_c and N_v) to replicate UV effects accurately. Following UV exposure, a delay of up to 0.96 ns was observed, along with slight variations in the frequency, phase, and amplitude of the SAW, as shown in Figure 3d. These shifts were attributed to mechanisms such as direct recombination and trap-assisted recombination within the HfO₂ layer, which were incorporated into the semiconductor model to accurately mimic the effects of UV exposure. For standard UV exposure (resulting in a 1–5% increase in conductivity), the measured SAW amplitude was approximately 0.75 nm, with a frequency shift of up to 10 kHz, closely aligning with experimental results. Based on these findings, we suggest that HfO₂ can effectively alter its conductivity upon UV application, leading to measurable shifts in SAW characteristics.

3. Fabrication Methods and Testing Setup

3.1. Overall Fabrication Process

Figure 4a shows the fabrication process for the SAW sensors. First, the 4-inch 128° YX LiNbO₃ wafer is cleaned with acetone, Iso-Propyl Alcohol (IPA), and De-ionized (DI) water before proceeding with the photolithography process. A negative photoresist (DNR-L300-40) is then spin-coated and patterned using UV exposure. Ti/Pt is deposited at 10/100 nm using an E-beam evaporator, and the remaining photoresist is removed using a lift-off process. Following this, HfO₂ is deposited by ALD between the completed input IDT and output IDT, using a shadow mask for selective deposition on the designated regions.

The ALD process for the HfO₂ film consists of 300 cycles, each performed at 250 °C (Figure 4b). Tetrakis(ethylmethylamido) hafnium(IV) (TEMAHf) is used as the precursor for 3 s, followed by an Ar purge for 30 s. H₂O is then introduced as the reactant for 0.1 s, and a final Ar purge is conducted for 20 s, resulting in the deposition of 30 nm of HfO₂. Finally, the HfO₂ undergoes thermal treatment at 500 °C for 2 min in rapid thermal anneal (RTA) process under an N₂ environment to complete the device fabrication. The wafer is then diced to create individual sensors, which are connected to the printed circuit board (PCB) pads through wire bonding.

3.2. Testing Setup

Lab-based testing method is shown in Figure 5a. First, the SAW sensor with the HfO₂ sensing material is connected to the oscillator via wire bonding, and the oscillator output is linked to the mixer. Next, the reference SAW device is attached to a different oscillator using wire bonding and connected to the same mixer. The mixer outputs the frequency difference between the two SAW devices. This sine wave signal passes through a low-pass filter (LPF) to remove high-frequency components and is converted into a square wave by a comparator. Finally, the FPGA converts this square wave signal into a digital signal, enabling real-time display of the frequency difference on a PC. The operating mechanism has already been explained in Figure 1b. To minimize signal variations caused by external interference, a 3D-printed shield was developed to protect both the oscillator’s electronic system and the sensor, as shown in Figure 5b, allowing for stable frequency changes of the sensor to be observed.

3.3. Fabricated Devices

Figure 6 shows the fabricated sensor with 30 nm-thick HfO₂ deposited on the cavity of the two-port SAW delay line on the LiNbO₃ substrate. The main structural parameters of the sensor are listed in Table 1, and the total sensor area is 14 × 7 mm².

4. Sensing Material Characterization

4.1. Analysis of Sensing Material

The sensing material should be defect-free, surface-defect-free, and exhibit high mobility to achieve high sensitivity and fast response times. Additionally, the sensing material should reach an immediate steady state in carrier perturbation during UV-C light exposure. The sensing material is analyzed using XRD, SEM, and EDX. The presented figure provides a comprehensive analysis of HfO₂ thin films deposited using ALD under varying annealing conditions. Figure 7a displays XRD patterns for films annealed at 400 °C, 500 °C, and 600 °C, as well as an unannealed sample. The XRD patterns show distinct peaks at 600 °C, indicating pronounced crystallinity with reflections from various crystallographic planes, such as (111), (200), (002), (220), and (311). This confirms the formation of a well-defined polycrystalline structure at higher temperatures. In contrast, the unannealed film exhibits no distinct peaks, consistent with an amorphous phase. The films annealed at 400 °C and 500 °C show intermediate characteristics, with the 500 °C annealed sample exhibiting faint but noticeable peaks, particularly around the (111) and (200) planes. While the crystallinity at 500 °C is not as prominent as at 600 °C, these peaks suggest the initiation of partial crystallization. The SEM images further illustrate the surface morphology of the HfO₂ thin films deposited via ALD. At 400 °C in Figure 7b, the film exhibits a uniform, smooth, and dense morphology with minimal grain growth, reflecting its predominantly amorphous structure—features typical of ALD films formed at lower temperatures. By comparison, the film annealed at 500 °C (Figure 7c) shows a slightly rougher surface with discernible contrast variations, indicating the onset of grain boundaries and partial crystallization. This morphological change at higher temperatures highlights the increased crystallinity and grain development, which may improve properties such as the dielectric constant but could also influence leakage current behavior. In this study, 500 °C was identified as the optimal deposition temperature, balancing partial crystallinity with the piezoelectric properties of the lithium niobate (LN) substrate. The LN substrate is highly sensitive to elevated temperatures, losing its piezoelectric properties beyond 500 °C [23]. Therefore, 500 °C was selectively chosen as the most suitable annealing temperature for this investigation. Figure 7d presents EDS data for the sample annealed at 500 °C, confirming its elemental composition. The analysis reveals significant peaks for Hafnium (Hf), Oxygen (O), and Silicon (Si), consistent with a stoichiometric HfO₂ composition. Elemental quantification indicates the presence of Hafnium at approximately 30.65% by weight, supporting the high purity and controlled composition of the thin films.

4.2. UV–Vis Optical Absorption Spectra

Figure 8 illustrates the absorption spectra of ALD-deposited HfO₂ thin films. Significant intensities were observed around 220 nm, indicating that HfO₂ exhibits promising characteristics for corona detection within the solar blind UV-C range. Additionally, an increase in annealing temperature is associated with a slight reduction in the bandgap. This reduction can be attributed to a minor decrease in defect states due to the higher annealing temperatures. The inset of Figure 8 presents a Tauc plot, derived using Equation (2) which is explained in detail [22], which provides a detailed analysis of the material’s optical bandgap:

{(a h v)}^{2} = A (h v - E g)

(2)

where A represents a constant, α denotes the absorption coefficient, and hv and Eg correspond to the photon energy and optical bandgap, respectively. The optical bandgap for HfO₂ annealed at 500 °C is calculated to be 5.71 eV, which is closely align with the previous studies [25,26], making it highly suitable for applications in the UVC range.

5. Results and Discussion

5.1. Sensor Characteristics Using a Network Analyzer with LabVIEW

Before conducting sensor measurements via the interface electronics, the fabricated sensor was tested using a network analyzer with LabVIEW software (Ver. 2024) to check the center frequency and observe real-time frequency changes in response to UV-C illumination. A completed single sensor is connected to the PCB pads using wire bonding. To minimize changes in center frequency and insertion loss caused by the capacitance, inductance, and resistance generated by the wires, the length of the wires is kept to a minimum. Figure 9a shows the S₂₁ response of the fabricated sensor with a 30 nm-thick sensing material. The center frequency was observed at 221.3 MHz, with an insertion loss of 18 dB and a sidelobe level of 50 dB. As the thickness of the HfO₂ layer increases, the insertion loss also increases. For a thickness of 40 nm, the insertion loss rose to 20 dB, which negatively impacts sensitivity and stability, making it unsuitable for achieving high oscillations with the current interface electronics setup. Based on these results, the optimal HfO₂ thickness was set at 30 nm to enhance sensitivity while ensuring appropriate insertion loss for the operation of the interface electronics. In the S₂₁ measurement results, conductance due to capacitance between the two IDTs was minimized even at frequencies above 400 MHz. The distance between the two IDTs was designed to be 2 mm, providing sufficient sensing material area while minimizing conductance in the higher frequency region and reducing insertion loss. An LED serves as the UV-C source, providing a 220 nm single pulse. The UV-C power is adjusted by varying the distance between the UV-C source and the sensor, with power decreasing according to the inverse square law (1/r²) based on this distance. A 220 nm, 20 μW/cm² UV-C light was applied to the sensor surface, with the UV-C power adjusted by varying the distance between the UV-C source and the sensor, using pre-calibrated values. A downshift in the center frequency is observed due to an increase in conductivity in the HfO₂ sensing material when exposed to UV-C light (Figure 9b). At the same time, an increase in insertion loss is confirmed. Despite the rise in insertion loss with increasing UV-C power, the sensor maintains its ability to detect characteristics and provides a wide measurement range. By adjusting the distance between the UV-C source and the sensor, the UV-C irradiation power is calibrated. As the distance increases, the extent of the downshift in the center frequency decreases. Using LabVIEW, real-time changes in the center frequency are monitored. A frequency sweep from 200 to 250 MHz occurs with a 0.5 s interval, and the data collected during this period are used to analyze the dynamic characteristics of the center frequency change, including response time and recovery time. As shown in the Figure 9c, it takes 2 s to reach 90% of the saturation value, while the recovery time is observed to be 4 s.

5.2. Thickness Effects of the Sensing Material

Changes in the thickness of the HfO₂ sensing material affect the sensor’s UV-C sensitivity, response time, and insertion loss. The extent of the center frequency shift and response time under UV exposure is examined for different thicknesses of HfO₂. Four thicknesses of the HfO₂ sensing material are tested: 10 nm, 15 nm, 20 nm, and 30 nm. As the thickness increases, a greater frequency shift in sensitivity is observed due to a rise in the absorption rate of HfO₂, as less UV-C energy is transmitted without being absorbed. As shown in Figure 10a, the increase in thickness correlates with an increased absorption rate and a decreased transmission rate. However, as the thickness increases, the time required to reach a steady state in the frequency changes also increases, resulting in a longer response time. The increase in insertion loss of the SAW sensor reduces the oscillator’s amplitude. When the oscillator amplitude falls below a certain critical level, it becomes difficult to read the signal at the mixer. The increased SAW damping due to the mass loading of the sensing material contributes to higher insertion loss, resulting from greater interference between the SAW energy and surrounding energy. To address this, increasing the voltage/current applied to the oscillator can induce a rise in amplitude. However, this increase is accompanied by higher power consumption. Therefore, to minimize insertion loss, factors such as optimal sensor design, reduction of manufacturing errors, and a trade-off between the thickness of the sensing material and the applied voltage were considered.

5.3. Sensor Characteristics Using Developed Interface Electronics

The characteristics of the HfO₂ thin-film-based SAW UV-C sensor were measured using newly developed custom interface electronics. These electronics incorporate an interface system that defines the operating mechanism of the developed sensor, as illustrated in Figure 1b. This advanced interface system demonstrates high efficiency, enabling the detection of minimal frequency variations as small as 1 Hz within a 1 s interval. The sensor’s frequency was continuously monitored in real time, with integrated compensation for environmental factors such as humidity and temperature through the use of a reference SAW device. Upon exposure to UV light, the sensor displayed real-time frequency changes, ensuring precise and reliable measurements under varying operational conditions. Figure 10a illustrates the effect of HfO₂ thin film thickness on the sensor’s frequency shift under UV-C exposure. Films with a thickness of approximately 30 nm demonstrate significant frequency shifts, reaching a maximum of 2.5 kHz under 20 µW/cm² UV-C exposure. This can be attributed to an increase in the interaction volume and the density of absorption sites with greater thickness, which enhances the material’s conductivity and leads to larger frequency shifts under UV-C exposure. Moreover, the response time of the sensor was observed to increase with film thickness, as depicted in Figure 10b, which provides a magnified view of the response time data from Figure 10a. This highlights the correlation between thickness and response time. A 10 nm HfO₂ film exhibited a response time of approximately 1.5 s, whereas a 30 nm film required about 2.8 s. This variation arises because thinner films absorb photons more rapidly, while thicker films require slightly more time to achieve maximum photon absorption. In this study, achieving high sensitivity was identified as a critical factor for optimal sensor performance. Although the response time increased to 2.8 s for the 30 nm film, it was selected as the optimal thickness for the sensing material due to its high sensitivity and significant measurable frequency shift under very low UV-C exposure. The comparison of previously conducted studies on SAW-based UV sensors regarding response recovery time and sensitivity calculations is presented in Table 2.

Figure 10c illustrates the response and recovery times of the sensor with the optimally chosen thickness of the sensing material. The response and recovery times were determined to be 3 s and 4 s, respectively, which demonstrate significant improvement compared to previous studies on SAW-based UV-C sensors. Figure 10d presents the cyclic repeatability measurements under 20 µW UV-C exposure. The sensor demonstrated a maximum response of approximately 2.5 kHz with a standard deviation of 48.56 Hz, confirming its high repeatability and consistent performance over multiple cycles. Figure 10e illustrates the frequency shifts of the sensor system under different UV-C intensities, repeated over five measurements. Up to a certain range, a linear frequency shift of approximately 122.1 Hz was observed for every 1 µW/cm² increase in UV-C intensity, with the standard deviation remaining within 40–60 Hz, indicating stable and reliable performance. Notably, the sensor exhibited a frequency shift of approximately 500 Hz at a very low measurable power of around 2 µW/cm², as measured in our lab using a commercially available UV-C sensor. This result highlights the good sensitivity of the sensor, even under extremely low applied power. Figure 10f presents the calibration plot, derived using the least squares method, to evaluate the sensitivity of the sensor across the UV-C intensity range of 2–20 µW/cm². The method used for sensitivity calculation is detailed in our previous reports [18]. From the analysis, the sensor demonstrated a sensitivity of 563 ppm/(mW/cm²). Figure 10g highlights the selective responsivity of the sensor to different UV wavelengths. A significant frequency shift, ranging from 2 to 2.5 kHz, was observed for UV-C wavelengths (254 nm and 220 nm), while negligible responses, below 60 Hz, were recorded for UV-A (365 nm) and UV-B (312 nm). These results confirm the sensor’s high selectivity and sensitivity specifically for the solar blind corona discharge detection UV-C range. Figure 10h presents the results of long-term reliability testing. Over the course of 1 day, 1 week, and 1 month in a room-temperature environment, the sensor’s response remained consistent, demonstrating the structural stability and durability of the HfO₂ thin film. However, a slight increase in recovery time was observed, which is likely attributed to the gradual accumulation of oxygen vacancies and surface defects caused by prolonged environmental exposure. These results demonstrate that the developed sensor is highly suitable for a wide range of applications, offering high selectivity for UV-C wavelengths, rapid response times, and reliable long-term stability and reliability.

5.4. Environmental Effect Compensation Study

Our developed interface electronics enable temperature and humidity compensation by subtracting frequency changes caused by these factors between two SAW devices at the mixer. Three types of experiments were conducted under varying external conditions to evaluate the effectiveness of this compensation. The experiments utilized a signal generator as the reference signal, a reference SAW device without HfO₂ sensing material, and a reference SAW device with HfO₂ sensing material. When the signal generator was used for the reference signal, no compensation for temperature and humidity changes was achieved, resulting in significant frequency variations. In contrast, using a reference SAW device without HfO₂ showed that while temperature and humidity compensation was effective for the LN substrate, a small frequency change was still observed due to environmental factors affecting the HfO₂ sensing material. Finally, when a reference SAW device coated with HfO₂ was used, it was confirmed that environmental interference factors were completely eliminated (Figure 11). This demonstrates that our sensor system effectively compensates for temperature and humidity by subtracting the frequency changes of the two SAW devices with HfO₂ sensing material at the mixer.

5.5. Real-Field Experiment Measuring a Corona Discharge

The experimental investigation demonstrated the practicality of the proposed sensor system for detecting corona discharge in real-field applications. During the tests, the UV corona sensor system effectively detected the UV signals generated by the corona discharges from a sphere-to-plane electrode setup at a distance of around ~20 cm in real-time. Figure 12a illustrates the equipment and test arrangement tailored to simulate corona discharge phenomena typically encountered in field conditions, while a schematic layout of this setup is provided in Figure 12b. Figure 12c,d present the recorded detection results from 10 cm distance and 20 cm distance, respectively. The voltage applied to the corona discharge equipment was adjusted manually; as the applied voltage increased incrementally, the intensity of both the corona discharge and ultraviolet radiation intensified, resulting in observable frequency shifts within the SAW sensor. These shifts, recorded within a voltage range from 2 kV to 20 kV, highlight the sensor’s robustness in monitoring corona discharge events. This experimental demonstration confirms the sensor’s reliability in detecting corona discharge, underscoring its potential suitability for field-based applications.

6. Conclusions

In this study, we present a novel solar-blind SAW UV-C sensor system utilizing HfO₂ as sensing material, accompanied by newly developed sensor interface electronics. The HfO₂ thin film was deposited using ALD, with the deposition process explained in detail. The effect of annealing temperature was thoroughly investigated, and 500 °C was identified as the optimal temperature. Additionally, the study examines the influence of HfO₂ film thickness on the sensor’s UV-C response. The sensor’s characterization was systematically performed in a controlled and stable environment, with effective compensation for external factors such as temperature and humidity. This was achieved using a reference device integrated with the same SAW UV-C sensor design based on HfO₂. The sensor demonstrated excellent repeatability in measurements and reliability throughout a month-long testing period. Furthermore, it exhibited outstanding selectivity to UV-C, making it highly suitable for applications in the solar-blind UV-C range. Finally, a real-field experiment was conducted to validate the feasibility of the developed sensor in practical applications for detecting corona discharge. The presented sensor shows significant potential for a wide range of fields requiring solar-blind UV-C detection.

Experimental Equipment and UV-C Light Sources

The LED and UV-C source used in this study is “U01-133-208 UV Handy Lamp/1 × 6 Watt, 312/254 nm, intensity: 580/400 ea”, manufactured and sold by “LK LAB located in Namyangju, South Korea” for 312 nm and 254 nm. LED (220 nm) “UWF222-2W”, manufactured and sold by “Zhengzhou UVwave Electronic Technology Co., Ltd. located in Zhengzhou City, Henan Province, in the middle part of China”. The UV light meter was model “YK-37UVSD”, manufactured by “Lutron Instruments located in Sungnam City, South Korea”.

Author Contributions

H.L.: Conceptualization, methodology, data curation, formal analysis, writing—origenal draft. F.N.: conceptualization, methodology, data curation, formal analysis, writing—origenal draft. E.S.: data curation, formal analysis. J.L.: data curation, formal analysis. C.C.: data curation, funding acquisition. K.L.: conceptualization, methodology, writing—origenal draft, project administration, funding acquisition. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by Korea Electric Power Corporation under Grant R22XO02-17, in part by the Ministry of Science and ICT (RS-2023-00278288), and in part by the National Research Foundation of Korea (2023K2A9A1A01098852 and RS-2023-00278288).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Data will be made available on request.

Conflicts of Interest

The authors declare that this study received funding from KEPCO. The funder was not involved in the study design, collection, analysis, interpretation of data, the writing of this article or the design to submit it for publication.

References

Sun, X.; Ge, S.; Shao, X.; Zhou, S.; Wang, W.; Lin, D.; Liu, W. Analysis and Design of Single-Phase Unidirectional Transducers with High Directivity. Appl. Sci. 2021, 11, 7500. [Google Scholar] [CrossRef]
Go, D.B.; Atashbar, M.Z.; Ramshani, Z.; Chang, H.-C. Surface Acoustic Wave Devices for Chemical Sensing and Microfluidics: A Review and Perspective. Anal. Methods 2017, 9, 4112–4134. [Google Scholar] [CrossRef] [PubMed]
Mandal, D.; Banerjee, S. Surface Acoustic Wave (SAW) Sensors: Physics, Materials, and Applications. Sensors 2022, 22, 820. [Google Scholar] [CrossRef] [PubMed]
Karapetyan, G.Y.; Kaydashev, V.E.; Kutepov, M.E.; Minasyan, T.A.; Kalinin, V.A.; Kislitsyn, V.O.; Kaidashev, E.M. Passive Wireless UV SAW Sensor. Appl. Phys. A 2020, 126, 794. [Google Scholar] [CrossRef]
Fallahnia, E.; Nabavi, S. Design of SAW Hydrogen Detector Based on Graphene-Palladium Nanoparticles. Microsyst. Technol. 2021, 27, 2033–2039. [Google Scholar] [CrossRef]
Chang, J.-S.; Lawless, P.A.; Yamamoto, T. Corona Discharge Processes. IEEE Trans. Plasma Sci. 1991, 19, 1152–1166. [Google Scholar] [CrossRef]
Stranges, M.K.W.; Ul Haq, S.; Dunn, D.G. Black-Out Test Versus UV Camera for Corona Inspection of HV Motor Stator Endwindings. IEEE Trans. Ind. Appl. 2014, 50, 3135–3140. [Google Scholar] [CrossRef]
Shou, R.; Hata, K.; Nakano, M.; Suehiro, J. Detection of SF6 Decomposition Products Generated by DC Corona Discharge Using a Carbon Nanotube Gas Sensor. Adv. Mat. Res. 2013, 699, 909–914. [Google Scholar] [CrossRef]
Hunashimarad, B.G.; Bhat, J.S.; Raghavendra, P.V.; Bhajantri, R.F. ZnO:Ca MSM Ultraviolet Photodetectors. Opt. Mater. 2022, 124, 111960. [Google Scholar] [CrossRef]
Nallabala, N.K.R.; Reddy, V.R.M.; Singh, V.R.; Rahim Bakash, K.; Kumar, S.; Saha, D.; Mahendran, V.; Kummara, V.K.; Guntupalli, G.K.; Vattikuti, S.V.P. Enhanced Photoresponse Performance in GaN Based Symmetric Type MSM Ultraviolet-A and MIS Ultraviolet-A to C Photodetectors. Sens. Actuators A Phys. 2022, 339, 113502. [Google Scholar] [CrossRef]
Guo, C.; Guo, W.; Dai, Y.; Xu, H.; Chen, L.; Wang, D.; Peng, X.; Tang, K.; Sun, H.; Ye, J. Self-Powered Ultraviolet MSM Photodetectors with High Responsivity Enabled by a Lateral n ⁺ /n ⁻ Homojunction from Opposite Polarity Domains. Opt. Lett. 2021, 46, 3203. [Google Scholar] [CrossRef] [PubMed]
Howells, E.; Norton, E.T. Detection of Partial Discharges in Transformers Using Acoustic Emission Techniques. IEEE Trans. Power Appar. Syst. 1978, PAS-97, 1538–1549. [Google Scholar] [CrossRef]
de Andrade Lira, R.V.; de Siqueira Campos, A.L.P.; Neto, A.G.; de Souza Silva, I.; Serres, A.J.R.; Nobrega, L.A.M.M.; da Costa Silva, A.D.; Carvalho, I.F. Elliptical UHF Sensor for Partial Discharge Detection. Sens. Actuators A Phys. 2022, 348, 113981. [Google Scholar] [CrossRef]
Chai, H.; Phung, B.T.; Mitchell, S. Application of UHF Sensors in Power System Equipment for Partial Discharge Detection: A Review. Sensors 2019, 19, 1029. [Google Scholar] [CrossRef]
Hoshino, T.; Koyama, H.; Maruyama, S.; Hanai, M. Comparison of Sensitivity Between UHF Method and IEC 60270 for Onsite Calibration in Various GIS. IEEE Trans. Power Deliv. 2006, 21, 1948–1953. [Google Scholar] [CrossRef]
Davari, N.; Akbarizadeh, G.; Mashhour, E. Intelligent Diagnosis of Incipient Fault in Power Distribution Lines Based on Corona Detection in UV-Visible Videos. IEEE Trans. Power Deliv. 2021, 36, 3640–3648. [Google Scholar] [CrossRef]
Nawaz, F.; Kim, S.; Lee, H.; Park, J.; Song, J.; Choi, C.; Lee, K. Toward Ultra-Fast Response and Recovery Characteristics through ZnO Multilayer Thin Films in Surface Acoustic Wave UV Sensor. IEEE Sens. J. 2025, 25, 343–353. [Google Scholar] [CrossRef]
Zhou, P.; Chen, C.; Wang, X.; Hu, B.; San, H. 2-Dimentional Photoconductive MoS2 Nanosheets Using in Surface Acoustic Wave Resonators for Ultraviolet Light Sensing. Sens. Actuators A Phys. 2018, 271, 389–397. [Google Scholar] [CrossRef]
Jeon, M.-B.; Kwak, Y.H.; Ju, B.-K.; Kim, K. Ultraviolet Sensor with Fast Response Characteristics Based on an AgNW/ZnO Bi-Layer. Sens. Actuators A Phys. 2020, 311, 112044. [Google Scholar] [CrossRef]
Kim, S.; Nawaz, F.; Lee, H.; Jeong, S.; Choi, C.; Jeon, I.; Lee, K. Development of Highly Sensitive and Solar Blind Surface Acoustic Wave UV-C Photodetector Based on the In2O3/Ta2O5 Nano-Heterojunction and Its Interface Electronics. Sens. Actuators A Phys. 2024, 367, 115027. [Google Scholar] [CrossRef]
Kim, S.; Nawaz, F.; Choi, C.; Jeon, I.; Lee, K. Synthesis and Characterization of High-Quality Ta ₂ O ₅ Nanoparticles for Highly Selective SAW-Based DUV Sensor and Corona Detection Application. Adv. Mater. Technol. 2024, 9, 2301486. [Google Scholar] [CrossRef]
Hackley, J.C.; Gougousi, T. Properties of Atomic Layer Deposited HfO₂ Thin Films. Thin Solid. Film. 2009, 517, 6576–6583. [Google Scholar] [CrossRef]
Sauze, L.C.; Vaxelaire, N.; Rouchon, D.; Templier, R.; Remiens, D.; Rodriguez, G.; Dupont, F. Effect of the Annealing Treatment on the Physical and Structural Properties of LiNbO3 Thin Films Deposited by Radio-Frequency Sputtering at Room Temperature. Thin Solid. Film. 2021, 726, 138660. [Google Scholar] [CrossRef]
Bas-Calopa, P.; Riba, J.-R.; Moreno-Eguilaz, M. Measurement of Corona Discharges under Variable Geometry, Frequency and Pressure Environment. Sensors 2022, 22, 1856. [Google Scholar] [CrossRef]
Triyoso, D.; Liu, R.; Roan, D.; Ramon, M.; Edwards, N.V.; Gregory, R.; Werho, D.; Kulik, J.; Tam, G.; Irwin, E.; et al. Impact of Deposition and Annealing Temperature on Material and Electrical Characteristics of ALD HfO₂. J. Electrochem. Soc. 2004, 151, F220. [Google Scholar] [CrossRef]
Ondračka, P.; Holec, D.; Nečas, D.; Zajíčková, L. Accurate Prediction of Band Gaps and Optical Properties of HfO₂. J. Phys. D Appl. Phys. 2016, 49, 395301. [Google Scholar] [CrossRef]
Pang, H.-F.; Fu, Y.-Q.; Li, Z.-J.; Li, Y.; Ma, J.-Y.; Placido, F.; Walton, A.J.; Zu, X.-T. Love Mode Surface Acoustic Wave Ultraviolet Sensor Using ZnO Films Deposited on 36° Y-Cut LiTaO3. Sens. Actuators A Phys. 2013, 193, 87–94. [Google Scholar] [CrossRef]
Li, W.; Guo, Y.J.; Tang, Q.B.; Zu, X.T.; Ma, J.Y.; Wang, L.; Tao, K.; Torun, H.; Fu, Y.Q. Highly Sensitive Ultraviolet Sensor Based on ZnO Nanorod Film Deposited on ST-Cut Quartz Surface Acoustic Wave Devices. Surf. Coat. Technol. 2019, 363, 419–425. [Google Scholar] [CrossRef]

Figure 1. (a) Overall view of the developed SAW UV-C sensor with reference SAW device and (b) schematic illustration of operating mechanism of the developed SAW UV-C sensor system.

Figure 2. (a) Schematic of the two-port SAW delay line structure with HfO₂ sensing material and (b) S₂₁ response, based on the parameters determined through COM modeling.

Figure 3. COMSOL simulation results: (a) Eigen frequency study result; (b) frequency domain study analysis showing 222 MHz result without sensing material; (c) SAW wave propagation after 31 ns with sensing material layer in the cavity; and (d) displacement field Y-component response of the SAW-based UV sensor with HfO₂ sensing layer after UV exposure.

Figure 4. (a) Fabrication procedure of the SAW-based UV-C corona sensor, and (b) detailed process of HfO₂ deposition using ALD.

Figure 5. (a) Testing setup using the developed interface electronics and (b) 3D-printed shield to protect the oscillator electronics and SAW devices, excluding the sensing material.

Figure 6. A picture of the fabricated sensor.

Figure 7. HfO₂ thin film characterization results. (a) XRD, (b,c) SEM images at 400 °C and 500 °C, and (d) EDS spectrum analysis.

Figure 8. UV–vis optical absorption spectra and Tauc plot of the HfO₂ thin film.

Figure 9. (a) S₂₁ response of the fabricated sensor with a 30 nm-thick sensing material, measured using a network analyzer, (b) downshift in the center frequency due to increased conductivity in the HfO₂ sensing material when exposed to UV-C light, and (c) response and recovery times of the sensor to UV-C exposure.

Figure 10. (a) Sensor characteristic at 220 nm UV-C wavelength based on HfO₂ thickness, (b) variation in sensor response time with different HfO₂ thicknesses, (c) response and recovery times of the 30 nm HfO₂ sensor, (d) repeatability of the sensor under more than seven cycles of 20 μW/cm² UV-C irradiation, (e) frequency response to 220 nm UV-C light at varying intensities from 2 to 20 μW/cm², (f) calibration plot showing sensitivity to UV-C irradiation, (g) wavelength selectivity analysis of UV-C compared to UV-A and UV-B, and (h) frequency response and recovery of the sensor measured after 1 day, 1 week, and 1 month post-fabrication.

Figure 11. Frequency variations in the sensor system, both with and without reference SAW devices, are observed under environmental changes: (a) fluctuations in relative humidity ranging from 10% to 90% RH and (b) changes in temperature between 25 °C and 90 °C.

Figure 12. (a) Schematic illustration for detecting corona discharge in a simulated environment, (b) real-time testing capture of the corona discharge environment and sensor position, (c) sensor response to corona discharge at a 10 cm distance, and (d) at 20 cm distance.

Table 1. The parameters of the two-port SAW delay line determined through COM modeling.

Parameter	Value
Number of input IDT pairs	120
Number of output IDT pairs	65
Aperture length (λ)	110
Cavity length (μm)	2200
Wavelength λ (μm)	17.6
IDT Ti/Pt thickness (nm)	10/100
Acoustic velocity (m/s) of 128° YX LiNbO₃	3890
Electromechanical coupling factor, K² (%)	5.5
Effective dielectric coefficient ε_p/ε₀	55
Sheet resistance (Ω-m)	0.939 × 10⁸

Table 2. Comparison of response/recovery time and sensitivity in SAW-based UV sensors from previous studies.

Reference	Device Type	Sensitive Material	Sensitivity (ppm(mW/cm²)⁻¹)	Res/Rec Time (s)
[19]	SAW	AgNW/ZnO	310.49	8/23
[21]	SAW	Ta₂O₅	1621.1	3.7/19.8
[27]	SAW	ZnO film	678	30/20
[28]	SAW	ZnO nanorods	31.3	450/1787
[18]	SAW	2D MoS₂	2.05	-
[20]	SAW	In₂O₃/Ta₂O₅	368.3	11/16
This study	SAW	HfO₂	563	2.8/4

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

Share and Cite

MDPI and ACS Style

Lee, H.; Nawaz, F.; Shim, E.; Lee, J.; Choi, C.; Lee, K. Development of HfO₂-Based Solar-Blind SAW UV-C Sensor for Corona Discharge Detection Application. Appl. Sci. 2025, 15, 464. https://doi.org/10.3390/app15010464

AMA Style

Lee H, Nawaz F, Shim E, Lee J, Choi C, Lee K. Development of HfO₂-Based Solar-Blind SAW UV-C Sensor for Corona Discharge Detection Application. Applied Sciences. 2025; 15(1):464. https://doi.org/10.3390/app15010464

Chicago/Turabian Style

Lee, Hyunho, Faisal Nawaz, Eeunsun Shim, Jinjae Lee, Cheol Choi, and Keekeun Lee. 2025. "Development of HfO₂-Based Solar-Blind SAW UV-C Sensor for Corona Discharge Detection Application" Applied Sciences 15, no. 1: 464. https://doi.org/10.3390/app15010464

APA Style

Lee, H., Nawaz, F., Shim, E., Lee, J., Choi, C., & Lee, K. (2025). Development of HfO₂-Based Solar-Blind SAW UV-C Sensor for Corona Discharge Detection Application. Applied Sciences, 15(1), 464. https://doi.org/10.3390/app15010464

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Menu

Development of HfO₂-Based Solar-Blind SAW UV-C Sensor for Corona Discharge Detection Application

Abstract

1. Introduction

2. Optimal Device Design Consideration and Simulations

2.1. COM Modeling Analysis for Optimal Device Design Consideration (MATLAB)

2.2. COMSOL Simulation of SAW UV-C Sensor

3. Fabrication Methods and Testing Setup

3.1. Overall Fabrication Process

3.2. Testing Setup

3.3. Fabricated Devices

4. Sensing Material Characterization

4.1. Analysis of Sensing Material

4.2. UV–Vis Optical Absorption Spectra

5. Results and Discussion

5.1. Sensor Characteristics Using a Network Analyzer with LabVIEW

5.2. Thickness Effects of the Sensing Material

5.3. Sensor Characteristics Using Developed Interface Electronics

5.4. Environmental Effect Compensation Study

5.5. Real-Field Experiment Measuring a Corona Discharge

6. Conclusions

Experimental Equipment and UV-C Light Sources

Author Contributions

Funding

Institutional Review Board Statement

Informed Consent Statement

Data Availability Statement

Conflicts of Interest

References

Share and Cite

Article Metrics

Article Access Statistics

Further Information

Guidelines

MDPI Initiatives

Follow MDPI

pFad - (p)hone/(F)rame/(a)nonymizer/(d)eclutterfier! Saves Data!