Improvements in Resistive and Capacitive Switching Behaviors in Ga₂O₃ Memristors via High-Temperature Annealing Process

Abstract

This study investigates the effect of a high-temperature annealing process on the characteristics and performance of a memristor based on a Ag/Ga₂O₃/Pt structure. Through X-ray diffraction analysis, successful phase conversion from amorphous Ga₂O₃ to β-Ga₂O₃ is confirmed, attributed to an increase in grain size and recrystallization induced by annealing. X-ray photoelectron spectroscopy analysis revealed a higher oxygen vacancy in annealed Ga₂O₃ thin films, which is crucial for conductive filament formation and charge transport in memristors. Films with abundant oxygen vacancies exhibit decreased set voltages and increased capacitance in a low-resistive state, enabling easy capacitance control depending on channel presence. In addition, an excellent memory device with a high on/off ratio can be implemented due to the reduction of leakage current due to recrystallization. Therefore, it is possible to manufacture a thin film suitable for a memristor by increasing the oxygen vacancy in the Ga₂O₃ film while improving the overall crystallinity through the annealing process. This study highlights the significance of annealing in modulating capacitance and high-resistive/low-resistive state properties of Ga₂O₃ memristors, contributing to optimizing device design and performance. This study underscores the significance of high-temperature annealing in improving the channel-switching characteristics of Ga₂O₃-based memristors, which is crucial for the development of low-power, high-efficiency memory device.

1. Introduction

Ga₂O₃ is highly valued in electronic and optical devices due to its properties, such as a wide band gap of 4.9 eV, excellent physicochemical properties, and thermal stability [1,2]. In general, Ga₂O₃ has five different atomic structures corresponding to α, β, γ, δ and ε, of which the beta phase is known to be the most stable [3]. Due to the stability of β-Ga₂O₃, various devices, such as capacitors, photodetectors, gas sensors, and resistive memory, are being studied and developed through various processes using the most stable structure of β-Ga₂O₃ [4,5,6,7,8]. With the recent increase in storage capacity, there has been a growing interest in capacitive and resistive memory for data and energy storage applications. Ga₂O₃ stands out as a promising material for such storage because it has a wide band gap, making it suitable as an excellent electrical insulator. Thus, its stability as an insulator suggests its suitability as a dielectric material for capacitors. Additionally, its diverse growth methods offer the advantage of ease of growth, through various techniques. By controlling growth conditions to adjust oxygen vacancies, Ga₂O₃ can also be used as a suitable material for resistive memory applications. In particular, a resistive random-access memory (RRAM) device based on Ga₂O₃ can operate based on resistive switching, enabling the device to freely transition between a high-resistance state (HRS) and a low-resistance state (LRS) [8,9]. RRAM consists of a simple structure, known as a metal-Ga₂O₃-metal structure, which is similar to that of a capacitor. Based on this, a memory capacitor leveraging resistance change was proposed [10]. Oxygen vacancies, an intrinsic defect in these Ga₂O₃-based resistance-dependent memories and capacitors, greatly affect the formation of filament channels, which is an important component of memristor and memcapacitor behaviors, and therefore, the behavior and performance of these devices [11,12]. The research groups are actively studying various aspects of Ga₂O₃ memristor and memcapacitor technology, including analyzing Ga₂O₃ properties, such as electrical conductivity, carrier concentration, and mobility [13,14,15,16]. To date, most research has focused on understanding the underlying mechanisms of resistive switching and improving device performance, durability, and scalability [17,18,19]. To improve these properties, studies are also underway to optimize the quality of insulator thin films using various vacuum deposition methods, such as chemical vapor deposition (CVD), atomic layer deposition, and sputtering [17,18,19,20,21]. Although sputtering does not guarantee the same film quality as ALD and CVD, it can still be a suitable deposition technique for the fabrication of electronic devices that do not require the excellent charge transport properties provided by epitaxial substrates [20]. In addition, an RF sputtering system precisely controls deposition parameters, such as sputtering power, Ar/O₂ gas pressure, and substrate temperature, to enable reproducible growth of Ga₂O₃ films with desired properties. However, it is noted that the quality of the thin films may be relatively lower compared to those grown by CVD [21]. Therefore, oxide semiconductor thin films grown by RF sputtering can be easily applied to devices such as memristors because they easily form conductive filaments, due to intrinsic defects. However, an excess of crystal defects can pose challenges for device properties and reliability. Therefore, strategies such as growth pressure control or the adoption of laminated structure growth have been reported to improve the quality of thin films grown by RF sputtering [22,23]. In addition, the most effective way to improve the film quality is known to be thermal treatment [24]. The effects of recrystallization and defect control resulting from thermal treatment are expected to control the characteristics of high-performance memristors and memcapacitors, offering a straightforward and rapid method of improving thin film quality [24,25].

In the Ga₂O₃-based memristor, the quality of Ga₂O₃ thin film and the presence of oxygen vacancies significantly impact Ga₂O₃ memristor behavior [26], with recrystallization from amorphous to crystalline states leading to reduced leakage current and an increased on/off ratio, making it suitable for high-performance memory devices. In addition, oxide semiconductors with a wide bandgap, such as Ga₂O₃, can develop devices suitable for high-performance capacitors because recrystallization removes deep levels caused by crystal defects, resulting in lower leakage currents and higher-quality insulation properties [27,28]. Additionally, high-temperature annealing enhances the crystalline quality of Ga₂O₃ by reducing defects, and crucially affects the formation of oxygen vacancies. These vacancies are essential for creating and controlling conductive filaments necessary for resistive switching [29]. Unlike other semiconductor devices, Ga₂O₃-based memristors can switch between high-resistivity and low-resistivity states after annealing, although the additional oxygen vacancies created during annealing improve the overall crystallinity of the thin film. This dual role of annealing to improve crystal quality and optimize oxygen vacancies can be an important process in sputter-grown Ga₂O₃. Therefore, optimization of the oxygen vacancy concentration in Ga₂O₃ thin film can affect the performance of the capacitor and memristor, as well as channel formation and dissolution, which affect the resistance of the memristor [30,31]. The formation of conductive channels by controlling oxygen vacancies, which can affect resistance and capacitance, will be able to be controlled based on crystallization through thermal treatment of Ga₂O₃ thin films, which is crucial for the realization of highly efficient devices produced with Ga₂O₃-based memristors and memcapacitors. In this study, Ga₂O₃ thin films grown by RF sputter were annealed at various temperatures to control the crystal defect in the Ga₂O₃ thin films. The relationship between the resistive switch behavior and capacitance along the conductive filament channel of the metal/Ga₂O₃/metal structure was then systematically studied.

2. Materials and Methods

Prior to depositing Ga₂O₃ using RF sputtering, the c-plane sapphire substrate is cleaned using the RCA cleaning. This process involves ultrasonically treating substrates with acetone for 10 min, followed by isopropyl alcohol (IPA) for 10 min, and then deionized water for 10 min to remove organic impurities from the surface. Subsequently, the 50 nm thick bottom electrode of Pt is deposited using RF sputtering, followed by deposition of a 100 nm thick Ga₂O₃ thin film using RF sputtering with a high-purity Ga₂O₃ target. Before deposition, the base pressure of the chamber is pumped down to 8.7 × 10⁻⁷ Torr, with a working pressure of 3 × 10⁻³ Torr, RF power set at 100 W, and growth temperature maintained at room temperature. Following the growth of the Ga₂O₃ thin film, furnace annealing is performed at 400 °C, 600 °C, and 800 °C for 3 min in an air atmosphere. Subsequent to thermal treatment on the Ga₂O₃ thin film, a 50 nm thick Ag electrode is deposited as a top electrode using a thermal evaporator.

The surface structure of the Ga₂O₃ film grown by RF sputtering was characterized using AFM (NanoFocus, my-Scope plus, Seoul, Republic of Korea). High-resolution X-ray diffraction (HR-XRD) ω/2θ scan (PANalytical, X’Pert Pro MRD, Tokyo, Japan) was conducted to investigate the crystalline structures of the Ga₂O₃ film on the bottom metal. X-ray photoelectron spectroscopy (XPS) spectra obtained from the ThermoFisher Scientific (Waltham, MA, USA) NEXSA system were carefully analyzed to detect the presence of oxygen vacancies and to examine the chemical environment of the Ga₂O₃ elements on the surface of the films. The optical band gap of the Ga₂O₃ thin film was determined by measuring its optical absorbance in the UV-visible range (ThermoFisher Scientific, Evolution 300) of 200 nm to 800 nm. Electrical characterization, including leakage current and memristor behavior, was carried out using a Keithley 2614B source meter. The capacitance of the devices was characterized using a 4284A LCR meter, with frequencies ranging from 500 Hz to 1.0 MHz. Through these experimental techniques and characterization methods, we investigated the relationship between the characteristics of the LRS and high-resistance state (HRS) and the capacitance of the Ga₂O₃ memristors. Moreover, we analyzed the effect of furnace annealing temperature on the electrical behavior and performance of Ag/Ga₂O₃/Pt memristors.

3. Results and Discussion

Figure 1a,b shows the results of HR-XRD ω/2θ scans of Ga₂O₃ thin films grown on Pt/sapphire templates at room temperature and after thermal treatment at 800 °C to investigate the crystallinity of the film, respectively. As shown in Figure 1a, Ga₂O₃/Pt/sapphire without thermal treatment, despite the presence of Ga₂O₃ thin film, no peaks related to Ga₂O₃ were observed, only the (111) Pt peak related to the bottom electrode and the (006) Al₂O₃ peak related to the sapphire substrate. It is clearly evident that the Ga₂O₃ thin film grown by sputtering was formed from an amorphous Ga₂O₃ thin film that does not exhibit good crystallinity with a single-crystal orientation due to low growth temperature and plasma damage [32]. For Ga₂O₃/Pt/sapphire annealed at temperatures above 400 °C, only the (111) Pt peak and (006) Al₂O₃ peak were observed in the broad XRD scan with high intensity, as depicted in Figure 1a. However, upon closer examination, in Figure 1b, where the XRD analysis was focused on a lower intensity range, peaks corresponding to (−201) and (−603) planes associated with β-Ga₂O₃ emerged around 18.4° and 59°, respectively. This indicates that the thermal treatment had preferentially oriented the amorphous Ga₂O₃ thin film toward the β-Ga₂O₃ plane, resulting in a significant improvement in crystallinity [33,34]. In particular, the XRD analysis in Figure 1b reveals the evolution of the β-Ga₂O₃ phase in the annealed Ga₂O₃ thin films. At 400 °C, very faint (−201) and (−603) β-Ga₂O₃ peaks were detected around 18.4° and 57°, respectively, suggesting that the β-Ga₂O₃ phase was first formed. Subsequently, as the annealing temperature was increased to 600 °C, the β-Ga₂O₃ phase associated with the (−201) and (−603) planes became more pronounced, with broader peaks near 18.4° and spanning 56° to 61°. Finally, at 800 °C, the intensity of the XRD peaks increased notably, accompanied by a reduction in full width at half maximum, indicative of further crystallization. These observations indicate a progressive enhancement in the formation and crystallinity of the β-Ga₂O₃ phase with increasing annealing temperature.

Figure 1.

(a) HR-XRD ω/2θ broad scan spectra ranging from 15° to 75° for as-grown, 400 °C, 600 °C, and 800 °C annealed-Ga₂O₃/Pt/sapphire, and (b) measurement of ω/2θ spectra focused on the low-XRD-intensity region to identify weak β-Ga₂O₃ (−201) and (−603) peaks.

XPS analysis was conducted on as-grown Ga₂O₃/Pt/sapphire and Ga₂O₃/Pt/sapphire annealed at 800 °C to characterize the defects of Ga and O atoms. Figure 2a,b shows the O1s and Ga3d peaks of as-grown and 800 °C-annealed Ga₂O₃ thin films, respectively. As shown in Figure 2a, the O1s spectrum of Ga₂O₃ demonstrated a satisfactory fitting with two peaks: Ga-O and oxygen vacancies. The Ga-O, located at a binding energy of 530.0 eV, corresponded to the O²⁻ ions surrounded by Ga atoms. Conversely, the Vo peak, positioned at a binding energy of 531.6 eV, was attributed to the presence of oxygen vacancies [35]. Upon thermal treatment at 800 °C, the ratio of Vo peak area (Vo/[Ga-O + Vo]) to the total area of the O1s peak increased from 23% to 28% for the as-grown Ga₂O₃ thin film. This increase was attributed to the volatilization of oxygen atoms with the Ga₂O₃ thin film during high-temperature thermal treatment, leading to an augmentation of oxygen vacancies [36]. Figure 2b shows the Ga3d peak of both as-grown and 800 °C-annealed Ga₂O₃ thin films. The Ga3d peak can be characterized by a Ga³⁺ peak at 20.0 eV and a Ga¹⁺ peak at 18.6 eV. The Ga³⁺ peak corresponds to Ga within Ga₂O₃, while the Ga¹⁺ peak represents Ga within the Ga₂O phase [37]. When comparing the Ga³⁺ and Ga⁺ peaks of the as-grown and 800 °C-annealed Ga₂O₃ thin films, the Ga³⁺ratio (Ga³⁺/[Ga³⁺ + Ga⁺]) in the as-grown Ga₂O₃ showed 83.7%, which decreased to 80.0% after the 800 °C annealing process. Conversely, the Ga⁺ peak ratio (Ga⁺/[Ga³⁺ + Ga⁺]) in the as-grown Ga₂O₃ thin film was 16.3%, which increased to 20.0% following the 800 °C-annealing process. This indicates that the high-temperature annealing process at 800 °C caused the volatilization of oxygen from the Ga₂O₃ thin film, resulting in a reduction of the Ga₂O₃ phase and the formation of some Ga₂O. This finding aligns with the O1s peak analysis of both the as-grown and 800 °C-annealed Ga₂O₃ thin film. Moreover, there was no significant formation of trace oxygen vacancies and, thus, no substantial formation of the Ga₂O phase. Considering the HR-XRD results in Figure 1, this indicates that despite the formation of oxygen vacancies and Ga₂O phase, the high-temperature annealing process can produce β-Ga₂O₃ thin films with relatively good crystal quality from initially amorphous Ga₂O₃ thin films.

Figure 2.

XPS (a) O1s and (b) Ga3d peaks of as-grown and 800 °C-annealed Ga₂O₃ thin films. The O1s were fitted with Ga-O and V_o peaks, and the Ga3d peaks were fitted with Ga³⁺, Ga⁺, and O2s. The black line represents the actual measurement. The green, blue, and pink lines correspond to the spectra fitted to the items shown in the legend of each figure.

As shown in Figure 3, the surface morphology of Ga₂O₃ thin films deposited on the (0001) sapphire was analyzed using AFM. With annealing temperature increasing from 25 °C to 800 °C, the average size of the surface grains of Ga₂O₃ thin films increased from 32.1 nm to 76.7 nm. In addition, as depicted in Figure 3e, the root-mean-square (RMS) roughness of the surface increased from 0.22 nm to 0.68 nm as the annealing temperature increased from 25 °C to 800 °C, respectively. The increase in the size of surface grains and surface roughness of Ga₂O₃ thin films during this high-temperature annealing process can be attributed to the vibration and movement of atoms induced by the high-temperature recrystallization of the Ga₂O₃ structure [38]. The post-nucleation growth of the island structure was attributed to the low surface mobility of atoms adsorbed during the sputtering process. Subsequent recrystallization through the Oswald-ripening effect, like grain growth during the thermal treatment process, is believed to crystallize the initial amorphous Ga₂O₃ into β-Ga₂O₃. This transformation is expected to improve the structural quality and electrical properties of the thin film [39].

Figure 3.

Surface morphologies of (a) as-grown, (b) 400 °C-, (c) 600 °C-, and (d) 800 °C-annealed Ga₂O₃ thin film, measured using AFM. (e) Schematic of Ga₂O₃/Pt structure grown on sapphire substrate. (f) Surface RMS roughness of Ga₂O₃ films as a function of annealing temperature.

Figure 4a shows the Tauc plot of the Ga₂O₃ thin films with different thermal treatments as a function of photon energy. The energy bandgap of Ga₂O₃ films under different thermal treatments can be determined using the equation αhν = B(hν − E_g)^1/2, where α is the absorption coefficient, h is the Planck constant (4.135 ×10⁻¹⁵ eVs), ν denotes the frequency (s⁻¹), B is a constant, and E_g represents the energy band gap (eV) [40]. As the annealing temperature increased, it is apparent that the slopes for high energies above 5.0 eV were nearly identical but shifted towards higher energies. This implies that the optical bandgap of the Ga₂O₃ thin film expands with an increasing in the annealing temperature due to the improvement of crystallinity. More specifically, the optical bandgap energy of the Ga₂O₃ thin film was calculated by extrapolating the linear portion of the curve to the energy axis, as shown in Figure 4b. As the annealing temperature increased from 25 °C to 800 °C, the optical bandgap of the Ga₂O₃ thin film increased from 4.69 eV to 4.90 eV, respectively. This increase in bandgap energy was attributed to the reduction of defects and impurities in the Ga₂O₃ thin film due to the annealing process. It is known that in Ga₂O₃ thin film, the donors are formed by oxygen vacancies and Ga interstitials, while acceptors are formed by Ga vacancies or a Ga-O vacancy pair [41,42,43]. As evident from the XPS O1s and Ga3d analysis depicted in Figure 2, oxygen vacancies generated during a high-temperature annealing process can introduce additional donor levels below the bandgap, resulting in a decrease in the effective bandgap. However, as depicted in Figure 1, during the high-temperature annealing process, when the amorphous Ga₂O₃ thin film transforms into β-Ga₂O₃ thin film, two-dimensional crystal defects, such as dislocations and grain boundaries, in addition to oxygen vacancies, are diminished. This reduction in crystal defects can further diminish the shallow levels around the bandgap. Consequently, the Ga₂O₃ thin film exhibits an energy bandgap closely aligned with the ideal energy bandgap of β-Ga₂O₃, approximately 4.9 eV [44].

Figure 4.

(a) Plot of (αhν)² against photon energy for the Ga₂O₃ films with different annealing temperatures. The inset shows the absorbance spectra of Ga₂O₃ thin films annealed at different temperatures. (b) Energy bandgap of Ga₂O₃ films as a function of the annealing temperature ranging from 25 °C to 800 °C.

Figure 5a shows a schematic of Ag/Ga₂O₃/Pt grown on c-plane sapphire memristor structure. Probe contacts were made to the top electrode, Ag, and the bottom electrode, Pt, respectively, to measure the electrical and memristive characteristics. Figure 5b shows the current (I)–voltage (V) curves obtained under applied bias ranging from −0.1 V to +0.1 V for a Ag/Ga₂O₃/Pt memristor structure containing Ga₂O₃ films annealed at various temperatures. The memristor demonstrated a current of +1.7 μA at 1.0 V. However, following an annealing process at temperatures exceeding 400 °C, the current drastically decreased to only a few tens of pA at +1.0 V. The substantial difference in leakage current was attributed to the significant reduction in crystal defects as the initially as-grown Ga₂O₃ thin film underwent transformation from its amorphous form to β-Ga₂O₃ crystallization through the annealing process [45]. In addition, as shown in Figure 4b, although high leakage current was observed initially due to the increased electron mobility facilitated by numerous crystal defects in the amorphous form near the bandgap of as-grown Ga₂O₃ thin films, the annealing process above 400 °C is believed to have mitigated these crystal defects and enhanced insulating properties by promoting recrystallization from amorphous to β-Ga₂O₃ thin film. This transformation resulted in a lower operating current for the memristors with the Ag/Ga₂O₃/Pt structure.

Figure 5.

(a) Schematic of Ag/Ga₂O₃/Pt memristor structure grown on sapphire substrate. (b) Current (I)–Voltage (V) curves for Ag/Ga₂O₃/Pt memristor. (c) I–V curves of the forming process for HRS-LRS conversion of as-grown Ag/Ga₂O₃/Pt memristor and Ag/Ga₂O₃/Pt memristor annealed at different temperatures. (d) Leakage current at 1.0 V and forming voltage of the Ag/Ga₂O₃/Pt memristor as a function of annealing temperature.

Figure 5c illustrates the I–V curves depicting the forming process of Ag/Ga₂O₃/Pt memristors annealed at different temperatures. The forming process consisted of applying an applied voltage, increasing from 0 V to 5 V, followed by a decrease back down to 0 V. In memristors, the forming process is the initial operation required to form conductive filaments necessary for the transition from the HRS to the LRS [46]. An as-grown Ga₂O₃ memristor without the annealing process can transition from HRS to LRS at voltages as low as about 2.07 V. However, Ga₂O₃ memristors that have been annealed at temperature above 400 °C must undergo the forming process at voltages above 3.5 V. Figure 5d shows that as the annealing temperature of the as-grown Ga₂O₃ memristor increased from 400 °C to 800 °C, the forming voltage increased from 3.5 V to 4.74 V, and the leakage current gradually declined from 21 pA to 14 pA, respectively. This trend can be attributed to the crystallization of the amorphous thin film into Ga₂O₃ thin film as the thermal treatment progresses. The abundance of crystal defects in amorphous, as-grown Ga₂O₃ thin film involves relatively easy migration of electrons [47]. Therefore, the forming process involving the application of high voltage to form conductive filaments is not necessary. However, as the annealing temperature increases, the transformation of the amorphous Ga₂O₃ thin film into a β-Ga₂O₃ thin film with crystalline properties occurs, rendering it challenging for oxygen vacancies and Ag⁺ ions to migrate. Consequently, a forming process is required to create conductive filament channel through which current can flow at high voltages [48].

Figure 6a–d shows the resistive switching (RS) behaviors of bipolar memristors with an Ag/Ga₂O₃/Pt structure composed of as-grown Ga₂O₃, 400 °C-, 600 °C-, and 800 °C-annealed Ga₂O₃ thin films, respectively. The RS curves depict four bias steps: 0.0 V to 5.0 V (step 1), 5.0 V to 0.0 V (step 2), 0.0 V to −5.0 V (step 3), and −5.0 V to 0.0 V (step 4). During the first step (0.0 V to 5.0 V), a set process occurred, and the transition from HRS to LRS occurred due to the formation of conductive filaments. As depicted in Figure 6a, the as-grown Ga₂O₃ memristor exhibited a set voltage from the HRS to the LRS at an average of 3.0 V, which is comparable to the forming voltage shown in Figure 5c. As a result, it is believed that as-grown Ag/Ga₂O₃/Pt memristor is a free-forming device. However, the annealed Ga₂O₃ memristor displayed a set process at less than 1.0 V, significantly lower than the forming voltage after the forming process. The forming voltage of Ga₂O₃ memristors annealed at 400 °C or higher increased with the rise in annealing temperature. However, after the forming process, the set voltage remained almost consistent regardless of the annealing temperature. In Figure 6a, during the reset process from 0.0 V to −5.0 V, the as-grown Ga₂O₃ memristor exhibited a sharp transition to the HRS at an average of −1.1 V. This reset process caused the as-grown Ga₂O₃ memristor to transition from LRS to HRS. However, the Ga₂O₃ memristor subjected to thermal treatment at temperature above 400 °C, regardless of the annealing temperature, demonstrated a reset process starting at −0.4 V. This reset process gradually switched to the HRS until −5.0 V was reached. Subsequently, upon applying an injection voltage from −5.0 V to 0.0 V, the HRSs are observed. Following this, by applying forward and reverse bias, the previously observed set and reset processes were repeatedly demonstrated, showing the RS behaviors. The annealed Ga₂O₃ memristors demonstrated lower set voltages compared to those of the as-grown Ga₂O₃ memristor. As indicated by the XPS results of Figure 2, this is presumed to be attributed to the increase in oxygen vacancies during the annealing process, which enhances the migration of Ag⁺ ions to form readily conductive filaments, thereby resulting in lower set voltages. In general, in metal/Ga₂O₃/metal memristors, the formation mechanism of conductive filaments is attributed to the migration of oxygen vacancies and metal ions [49]. Therefore, it is believed that the high-quality annealed-Ga₂O₃ memristors with a higher concentration of oxygen vacancies facilitate the formation of additional conductive filaments once the initial conductive filaments are formed through the forming process, ultimately leading to lower set voltages [50]. In particular, as-grown Ga₂O₃ memristors exhibit sharp set and reset processes, while annealed Ga₂O₃ memristors exhibit a lower set voltage and a reset process with gradually increasing resistance. This is speculated to be due to the fact that as-grown Ga₂O₃ memristors are formed with one or fewer conductive filament, while annealed Ga₂O₃ memristors are formed with multiple conductive filaments. The sharp set and reset processes observed in as-grown Ga₂O₃ memristor are believed to stem from the highly radical transition between the HRS and LRS in the amorphous Ga₂O₃ thin film. Only a very small number of conductive filaments were formed and broken at the weakest points by Ag⁺ ions, rather than oxygen vacancies, depending on the applied voltage. In contrast, as shown in Figure 1 and Figure 2, the annealed Ga₂O₃ thin films experienced a slight increase in oxygen vacancies [51], but the crystalline defects decreased as the crystallization processed from amorphous to β-Ga₂O₃, resulting in the formation of an excellent thin film. Consequently, it became challenging for initial conductive filaments to form, leading to a high forming voltage, as shown in Figure 5d. However, there remains the possibility that multiple conductive filaments can be formed by crystal defects such as grain boundaries crystallized from β-Ga₂O₃ when high voltage is applied. Therefore, it is speculated that the set process occurs radically at low voltage due to the increase in oxygen vacancies and the generation of multiple conductive filaments by Ag⁺ ions during the annealing process. Conversely, under reverse bias, a gradual reset process occurs as conductive filaments break due to the migration of Ag⁺ ions, and the thermal effect of high voltage application gradually takes effect. This gradual reset process offers greater controllability compared to a rapid reset, making memristors particularly suitable for analog memory applications, such as multi-bit operation, with excellent retention and durability [52]. Figure 6f–i shows the operating current of the HRS and LRS at 0.5 V for memristors with a Ag/Ga₂O₃/Pt structure composed of as-grown Ga₂O₃ and annealed Ga₂O₃ at different temperatures, with 30 measurements of resistive switching ranging from −0.5 V to + 5.0 V. The annealed Ga₂O₃ memristors exhibited larger HRS/LRS ratios compared to the as-grown Ga₂O₃ memristor due to lower leakage current in the HRS state. It is believed that the HRS/LRS ratio increases with rising annealing temperature because the as-grown Ga₂O₃ thin films crystallize from amorphous to β-Ga₂O₃ thin film through a high-temperature annealing process, as shown in Figure 1. This process slightly increases point defects, such as oxygen vacancies, as shown in Figure 2a, but decreases overall crystal defects, resulting in a higher HRS/LRS ratio at higher annealing temperatures [53]. The large HRS/LRS ratio and low set voltage observed in the Ag/Ga₂O₃/Pt memristor, resulting from the high-temperature annealing process for Ga₂O₃, indicate a clear distinction between the on and off states of the memristor and lower power consumption.

Figure 6.

RS behaviors of I–V curves measured for 30 repetitions for (a) as-grown Ag/Ga₂O₃/Pt memristor and annealed Ag/Ga₂O₃/Pt memristors with different temperatures of (b) 400 °C, (c) 600 °C, and (d) 800 °C. (e) The schematic diagram of the formation of conductive filaments in the HRS and LRS of as-grown Ag/Ga₂O₃/Pt memristors, and in the HRS, intermediate resistive state (IRS), and LRS of 800 °C-annealed Ag/Ga₂O₃/Pt memristors. On/off ratio between HRS and LRS for (f) as-grown Ag/Ga₂O₃/Pt memristor and (g) 400 °C-, (h) 600 °C-, (i) 800 °C-annealed Ag/Ga₂O₃/Pt memristors as a function of the measurement cycle.

Figure 7a shows the C-V curves depicting the capacitive switching behavior of the Ag/Ga₂O₃/Pt memristor functioning as a memcapacitor with applied voltage across the set and reset processes, where the set and reset processes represent the low capacitance state to high capacitance state and the high capacitance state to low capacitance state. Initially, the as-grown Ga₂O₃ memcapacitor exhibited a set voltage of 4.5 V and a reset voltage of −7.23 V. However, as the annealing temperature progressed beyond 400 °C, both the set voltage and reset voltage decreased. In particular, at 800 °C, the set voltage and reset voltage were measured to be 3.2 V and −4.8 V, respectively. Furthermore, it was observed that the maximum capacitance of the as-grown Ga₂O₃ memcapacitor significantly increased from 37 nF to 66 nF for the 800 °C-annealed Ga₂O₃ memcapacitor following the set process. This indicates that the as-grown amorphous Ga₂O₃ thin film can be positively affected, in terms of capacitance, by a high-temperature annealing process, thereby improving its charge storage capacity, and improving its performance as a memcapacitor. The observation that the maximum capacitance increases with the effect of the annealing process indicates that the electrical behavior and performance of the Ga₂O₃-based memristors and memcapacitor, which exhibit this capacitive switching behavior, are significantly influenced by the thermal treatment effect. In addition, to calculate the density of trapped charges (N_charge) of memristors that can be used as memcapacitors in Ag/Ga₂O₃/Pt structures with Ga₂O₃ annealed at different annealing temperatures, the following equation was used:

$$N_{charge}=\frac{∆V · C_{LRS}}{qA }$$

where ΔV is the difference between the set voltage and reset voltage, and C_LRS is the capacitance in the LRS, q is the electronic charge (1.602 × 10⁻¹⁹ C), and A is the area of the device [54]. The charge capture density provides information about the density of charges trapped in the Ga₂O₃ thin film present between the Ag and Pt electrodes, which is directly related to the charge storage capacity of the Ag/Ga₂O₃/Pt memcapacitor that can also be used to form a memristor. The as-grown Ga₂O₃ memcapacitor exhibited a ΔV value of 11.5 V, but as the annealing temperature increased from 400 °C to 800 °C, it decreased from 9.6 V to 8.0 V, respectively. The C_LRS value of the as-grown Ga₂O₃ memcapacitor was 22.4 nF. However, it increased significantly from 29.4 nF to 56.8 nF as the annealing temperature increased from 400 °C to 800 °C. As a result, as shown in Figure 7c, the trapped charge density of the as-grown Ga₂O₃ memcapacitor was 5.1 × 10¹⁹ cm⁻², but with the increase in annealing temperature, the trapped charge density increased continuously, reaching 9.1 × 10¹⁹ cm⁻² at 800 °C annealing temperature. This increase in the N_charge value was attributed to the migration of Ag⁺ ions through the defect site, such as oxygen vacancies, and their trapping in high−purity Ga₂O₃ [55]. Therefore, it has a higher capacitance value because of the Ag⁺ ion trap generated at the defective site, which affects the performance improvement of the memristor and capacitor. Based on these results, it is speculated that higher annealing temperatures increase the N_charge and the density of trapped charges, resulting in better charge storage and improved device performance.

Figure 7.

(a) C-V hysteresis curves representing the capacitive switching behavior of Ag/Ga₂O₃/Pt memristors heated at different temperatures into membranes, (b) set and reset voltage variations of Ag/Ga₂O₃/Pt membranes as a function of annealing temperature, and (c) charge trapping density of Ag/Ga₂O₃/Pt membranes as a function of different annealing temperatures.

4. Conclusions

This study demonstrates significant improvements in the thin film characteristics and the memristor through the annealing process of Ga₂O₃ thin films. As a result of XRD measurements, the phase transformation from an annealing-processed Ga₂O₃ thin film to β-Ga₂O₃ was confirmed. It was also confirmed, through absorbance analysis, that the amorphous Ga₂O₃ was changed to β-Ga₂O₃. Through XPS analysis, it was found that the oxygen vacancy density was higher in the annealed Ga₂O₃ thin film. Through the above two results, the leakage current rapidly decreased through recrystallization in the annealing Ga₂O₃ device, and the formation of a conductive film became easy due to the increase in oxygen vacancies, confirming a decrease in set voltage. It is crucial for conductive filament formation and the charge transport path in memristors. A Ga₂O₃ thin film with abundant oxygen vacancies exhibits decreased set voltage and increased capacitance in LRS, attributed to additional carriers such as oxygen vacancies and Ag⁺ ions. This indicates easy capacitance control, depending on channel presence. Based on these results, we suggest that annealing temperatures up to 800 °C improve the performance of Ga₂O₃ memristors. The thermal treatment at this temperature enhances the crystalline quality of the Ga₂O₃ thin films, reducing defects that typically hinder the formation of stable high-resistance and low-resistance states. Consequently, the memristor’s ability to switch between these states is improved, thereby enhancing its performance.

Author Contributions

Conceptualization, S.-N.L.; methodology, H.J.L.; validation, H.J.L. and J.-H.K.; resources, S.-N.L.; data curation, H.J.L. and J.-H.K.; writing—original draft preparation, H.J.L. and S.-N.L.; writing—review and editing, S.-N.L.; visualization, H.J.L. and H.-J.K.; supervision, S.-N.L.; project administration, S.-N.L.; funding acquisition, S.-N.L. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors on request.

Conflicts of Interest

The authors declare no conflicts of interest.

Funding Statement

This research was supported by a Research Program (NRF−2020R1A2C1009630) through the National Research Foundation of Korea (NRF), funded by the Ministry of Education, Science and Technology, Republic of Korea.