Preparation of heterojunction photoelectric sensor and neural synaptic bionic application of heterojunction photoelectric sensor

Abstract

The invention relates to the field of photoelectric material and device physics, and discloses a preparation method of a Ga2O3: Sn / HfOx heterojunction photoelectric sensor, and the method comprises the following steps: S1, cleaning a substrate, S2, preparing a Ga2O3: Sn thin film, S3, growing an electrode, S4, preparing an HfOx thin film, and carrying out neural synapse bionic application of the Ga2O3: Sn / HfOx heterojunction photoelectric sensor. Comprising Ga2O3: Sn / HfOx synaptic plasticity research and Ga2O3: Sn / HfOx photoelectric sensor sensing and storage calculation integrated application research. The Ga2O3: Sn thin film is prepared by adopting a sol-gel method with accurate and controllable doping amount, donor doping with controllable concentration is realized by controlling the mass of the added tin sheet, so that the photocurrent and responsivity of the photoelectric sensor are increased, the HfOx thin film is prepared by adopting a magnetron sputtering method, the photoelectric sensor with a double-layer structure is constructed, the detection rate and responsivity of the device are improved, and the photoelectric sensor is suitable for the field of photoelectric sensors. An ultraviolet signal is used to simulate a plurality of synaptic functions, including EPSC, PPF and empirical learning behaviors. Furthermore, a high-linearity light-enhanced electrical suppression behavior is realized, and a basis is provided for realizing a high-precision image recognition function.

Description

Technical Field

[0001] The present invention relates to the field of optoelectronic materials and device physics, and in particular to a Ga2O3:Sn / HfO x Preparation of heterojunction photoelectric sensors and their synaptic bionic applications. Background Art

[0002] With the continuous development of current fields such as artificial intelligence, we are gradually moving towards an era of information explosion. However, traditional computer neuromorphic network systems face multiple challenges, including limited operating bandwidth, high integration density, and slow information processing speed. These problems not only lead to the lengthy and cumbersome information processing process, serious loss of feature information, but also a significant increase in energy consumption. Therefore, we urgently need innovative technologies to meet these challenges. The human brain has hundreds of millions of sensory synapses and neurons, which can achieve highly parallel information processing, have excellent memory storage capabilities, and have high fault tolerance, low energy consumption, and strong reconfiguration. This makes brain-like neuromorphic devices an ideal choice for processing unstructured information tasks and are favored by researchers. They not only provide us with a new way of information processing, but also a key way to realize applications such as emerging artificial intelligence systems. The key to realizing brain-like neuromorphic devices is to build artificial neural networks, and synaptic bionics is the basis for achieving this goal.

[0003] Photoelectric sensors have the characteristics of high responsiveness, high detection rate and high signal-to-noise ratio, and can show obvious light response even in extremely dark light conditions, showing broad application prospects in the field of neural synaptic bionics. At present, most reported studies focus on the development of neuromorphic sensors working in the visible light range, which are intended to become substitutes for the human visual system. However, the exploration of the day-blind ultraviolet band is also of extremely important value and broad potential application prospects. Day-blind ultraviolet refers to the ultraviolet part with a wavelength between 200nm-300nm, which is strongly absorbed by the earth's atmosphere and usually cannot reach the ground. The research and development of neuromorphic sensors for this special band can not only expand the boundaries of existing light sensing technology, but also play a unique role in fields such as environmental monitoring, aerospace, biomedical imaging and advanced security protection. For this reason, the present invention proposes a Ga2O3:Sn / HfO x Preparation of heterojunction photoelectric sensors and their synaptic bionic applications. Summary of the invention

[0004] The purpose of the present invention is to provide a preparation method of a heterojunction photoelectric sensor and a neural synapse biomimetic method thereof, aiming to prepare Ga2O3:Sn thin film by a sol-gel method with a precisely controllable doping amount, achieve donor doping with a controllable concentration by controlling the mass of the added tin sheet, increase the photocurrent and response of the photoelectric sensor, and prepare HfO by a magnetron sputtering method. x The film was used to construct a double-layer photoelectric sensor, which improved the detection rate and responsiveness of the device and had certain advantages in high-performance weak light detection. Secondly, the effects of different electrodes and different oxygen partial pressures on HfO x Thin film on Ga2O3:Sn / HfO x The effects of photoelectric sensors were studied, and UV signals were used to simulate a variety of synaptic functions, including EPSC, PPF, and experiential learning behaviors. Furthermore, highly linear light-enhanced electro-inhibition behavior was achieved, providing a basis for high-precision image recognition. Finally, the device's sensing, storage, and computing technology for weak ultraviolet light was expanded, so that the recognition, memory, photoelectric information demodulation, and logic operation functions can be completed on one device, making up for the lack of visible light perception in certain specific application scenarios, such as ultraviolet flame detection and fingerprint recognition, and providing a new structure for the development of future neuromorphic computing and artificial vision systems.

[0005] To achieve the above object, the present invention provides the following technical solution: a Ga2O3:Sn / HfO x A method for preparing a heterojunction photoelectric sensor, the method comprising the following steps:

[0006] S1: Substrate cleaning

[0007] The substrate used in this experiment is double-polished c-plane sapphire. The cleanliness of the substrate has a crucial impact on the uniformity of film growth, adhesion and overall performance. Therefore, the substrate must be thoroughly cleaned before spin coating to remove any impurities and contaminants that may affect the quality of the film. The sapphire substrate is treated with acetone, ethanol, and deionized water for 10 minutes in turn. After cleaning, it is blown dry with nitrogen for standby use;

[0008] S2: Ga2O3:Sn thin film preparation

[0009] (1) Sol preparation

[0010] In the experiment, gallium nitrate was used as the gallium source precursor, ethylene glycol was used as the solvent, and triethanolamine was used as the stabilizer. In addition, in order to dope with tin, high-purity metal tin flakes (purity up to 99.99%) were selected. These reagents were mixed in a specific ratio to prepare a sol with a concentration of 0.8 mol / L. Among them, Ga2O3, Ga2O3:Sn1, Ga2O3:Sn2, and Ga2O3:Sn3 represent samples with undoped and tin-doped concentrations of 5%, 10%, and 15%, respectively. This preparation process needs to be carried out at room temperature, stirred for 2 hours to ensure that the components are fully mixed, and then left to stand for 24 hours so that the sol can stabilize and reach a standby state. All reagent information and specific amounts used in the experiment have been listed in detail in Tables 2-1 and 2-2.

[0011] Table 2-1 Experimental reagents

[0012]

[0013] Table 2-2 Dosage of chemicals for preparing Ga2O3:Sn thin film

[0014]

[0015] (2) Spin coating process

[0016] The prepared solution was pipetted and dripped onto the clean sapphire substrate surface to ensure that the sol could evenly cover the entire surface. Then, in order to further improve the uniform coverage of the sol and promote the volatilization of some organic solvents, the sol was spin-coated at a speed of 3000 r / min for 30 seconds, and the acceleration was set to 1000 rpm / s. Finally, a wet film was formed on the substrate surface.

[0017] (3) Drying and heat treatment

[0018] After each layer of film is spin-coated, it is placed on a hot plate and heated at 100°C for 10 minutes to evaporate some of the water and organic solvent on the surface of the substrate and dry the wet film. In order to ensure the continuity and uniformity of the film and achieve sufficient thickness, it is necessary to heat treat at 500°C for 15 minutes after each 4 layers of film are prepared, and the heating rate is controlled to be 10°C / min. During this process, a series of chemical reactions occur in the gel film, and finally a dense tin-doped gallium oxide film is formed and deposited on the substrate. This process is repeated 6 times.

[0019] (4) Post-annealing treatment

[0020] The prepared samples were placed in a tubular annealing furnace and subjected to subsequent annealing treatment at 600 °C for 2 h in an air atmosphere;

[0021] S3: Growing electrodes

[0022] Use vacuum thermal evaporation equipment to evaporate metal electrodes. The electrode materials used are gold (Au) wire or aluminum (Al) wire. The evaporation boat uses a metal tungsten boat. Grow Al (Au) metal electrodes with a thickness of 30nm on the Ga2O3:Sn film with interdigitated electrodes (10 pairs of interdigitated electrodes, each interdigitated with a length of 1000μm, a width of 100μm, and a spacing of 100μm);

[0023] S4: HfO x Thin film preparation

[0024] HkDJ x The target material was provided by Zhongnuo New Materials (Beijing) Technology Co., Ltd. The size and purity of the target material were Φ50.8×5mm and 99.99%. Preparation of HfO by RF magnetron sputtering x The vacuum degree in the chamber is 5.0×10 -6 Torr, ignition pressure 0.05 Torr, growth pressure 7.5×10 -3 Torr, argon flow rate of 18.0 sccm, sputtering power of 60 W, and growth at room temperature.

[0025] A Ga2O3:Sn / HfO x Heterojunction photoelectric sensors for neurosynaptic biomimetic applications, including Ga2O3:Sn / HfO x Synaptic plasticity research and Ga2O3:Sn / HfO x Research on the application of photoelectric sensor sensing, storage and computing

[0026] Preferably, the Ga2O3:Sn / HfO x Synaptic plasticity studies include Ga2O3:Sn / HfO x Photoelectric sensors simulating short-term plasticity of biological synapses with Ga2O3:Sn / HfO x Photoelectric sensors mimic long-term plasticity of biological synapses.

[0027] Preferably, the Ga2O3:Sn / HfO x Photoelectric sensors simulate short-term plasticity of biological synapses: In the human brain, presynaptic neurons and postsynaptic neurons are connected through synapses, which are the link between presynaptic and postsynaptic neurons. The key to simulating synaptic activity is to regulate the connection strength between each neuron unit, so as to simulate the process of signal transmission and learning between neurons through synapses in the biological nervous system. In the synaptic performance test, the reading voltage is 0.5V, the light excitation is the LA-410UV light source and the light pulse is filtered by a filter with a cutoff wavelength of 380nm. A single light pulse acting on the device will trigger a postsynaptic excitation current (EPSC). At the end of the light pulse, the EPSC will reach a peak and then gradually weaken. This EPSC behavior is used to evaluate the connection strength of the synapse.

[0028] Preferably, the Ga2O3:Sn / HfO x Photoelectric sensors simulate long-term plasticity of biological synapses: Biologically, memory can be roughly divided into two types according to its duration: short-term memory and long-term memory. Both short-term memory and long-term memory are closely related to synaptic plasticity. Short-term memory corresponds to the short-term plasticity (Short-Term Plasticity, STP) of neural synapses. Short-term plasticity is a temporary and short-term synaptic plasticity change, which usually lasts from a few seconds to a few minutes. Short-term plasticity makes the connection strength between neurons change rapidly in a short period of time, thereby achieving The storage and processing of short-term memory, long-term memory corresponds to the long-term plasticity (Long-Term Plasticity, LTP) of synapses. Long-term plasticity is a persistent and long-lasting synaptic plasticity change, which usually lasts for several hours or even days. Long-term plasticity enables the connection pattern between neurons to be changed stably over a long period of time, thereby realizing the storage and consolidation of long-term memory. In the experiment, the plasticity changes of biological synapses can be simulated by adjusting the various parameters of the light pulse, and the transition from short-term plasticity to long-term plasticity can be achieved.

[0029] Preferably, the Ga2O3:Sn / HfO x The research on the integrated sensing, storage and computing application of photoelectric sensors includes the simulation of image recognition and memory functions, and the simulation of photoelectric information demodulation and logical operation functions.

[0030] Preferably, the simulation of the image recognition and memory function: the visual system is one of the key ways for organisms to perceive the external environment, acquire and process information. This complex and sophisticated physiological mechanism can capture light signals and convert these light signals into electrical signals through photoreceptor cells in the eyes, which are then transmitted to the brain for further analysis and integration. This ability to acquire and recognize images plays a vital role in our daily lives, helping us to understand the world, understand the environment, and make correct responses and decisions.

[0031] Preferably, the optoelectronic information demodulation: optoelectronic demodulator is the core component of optical communication equipment, and is a device used for optical signal conversion and modulation. In optical communication systems, optoelectronic demodulators are usually used to receive optical signals and convert them into electrical signals for demodulation and processing at the receiving end. It can convert the intensity, frequency or phase information of light waves into corresponding electrical signals, and then amplify, filter and decode them through circuits to finally extract the original information content.

[0032] Preferably, the logic operation function simulation: using the device's ability to integrate and process optical signals and electrical signals and to convert them to each other, a system that can realize complex logic operation functions can be constructed. Boolean logic, as a branch of mathematical logic, was founded by George Boole in the mid-19th century. Its core concept is to use two opposing values, "true" and "false", to describe the state of a proposition. In the digital expression system, "true" corresponds to the value 1, and "false" corresponds to the value 0. This binary logic system constitutes the basic logic unit of digital circuits and computer hardware, is the basis of computer hardware and software design, and is also the basis of logical operations in programming languages. It enables programmers to use simple logical rules to construct complex program structures, combine Boolean logic with the joint action of optoelectronic devices, and realize information processing and transmission through the logical operation function of optoelectronic devices.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] The present invention adopts a sol-gel method with a precisely controllable doping amount to prepare Ga2O3:Sn thin film, and realizes donor doping with controllable concentration by controlling the mass of the added tin sheet, so that the photocurrent and responsiveness of the photoelectric sensor are increased, and the magnetron sputtering method is used to prepare HfO x The film was used to construct a double-layer photoelectric sensor, which improved the detection rate and responsiveness of the device and had certain advantages in high-performance weak light detection. Secondly, the effects of different electrodes and different oxygen partial pressures on HfO x Thin film on Ga2O3:Sn / HfO x Effect of photoelectric sensors.

[0035] The present invention uses ultraviolet signals to simulate a variety of synaptic functions, including EPSC, PPF, and experiential learning behavior. Furthermore, a highly linear light-enhanced electrical inhibition behavior is achieved, providing a basis for realizing high-precision image recognition functions. Finally, the device's sensing, storage, and computing technology for weak ultraviolet light is expanded, so that the recognition, memory, photoelectric information demodulation, and logic operation functions can be completed on one device, making up for the lack of visible light perception in certain specific application scenarios, such as ultraviolet flame detection and fingerprint recognition, and providing a new structure for the future development of neuromorphic computing and artificial vision systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is Ga2O3:Sn3 / Al / HfO of the present invention. x EPSC effect diagram of the device;

[0037] Figure 2 (a) Ga2O3:Sn3 / Al and Ga2O3:Sn3 / Al / HfO of the present invention xDevice (b) Ga2O3:Sn3 / Au and Ga2O3:Sn3 / Au / HfO x Device (c)Ga2O3 / Al / HfO x and Ga2O3:Sn3 / Al / HfO x Comparison of EPSC effects of devices;

[0038] Figure 3 The Ga2O3:Sn / HfO x (a) Pulse facilitation (PPF) of the device; (b) PPF index diagram at different time intervals;

[0039] Figure 4 The Ga2O3:Sn / HfO x The current variation of the device under the sequence of ultraviolet light pulses at different time intervals;

[0040] Figure 5 The Ga2O3:Sn / HfO x The transition diagram from short-term plasticity (STP) to long-term plasticity (LTP) caused by different (a) light intensity (b) light duration (c) number of light pulses;

[0041] Figure 6 is the experiential learning behavior diagram of the present invention;

[0042] Figure 7 is a diagram of the light-enhanced electrical suppression behavior of the present invention;

[0043] Figure 8 The Ga2O3:Sn / HfO x Devices achieve light intensity-dependent image recognition and memory function diagram;

[0044] Fig. 9 The present invention is a) decoding "00", "01", "10" and "11" by light pulses of different light intensities; (b) "NENU" is demodulated according to the eight-bit code of ASCII;

[0045] Fig.10 The Ga2O3:Sn / HfO in the present invention is taken as an example of "1". x Device optoelectronic operation flow chart;

[0046] Fig.11 The Ga2O3:Sn / HfO x A diagram of the device's response to ten different sequences of optical pulses;

[0047] Fig.12 (a) The image of the letter "nenu" used in the test of the present invention; (b) Ga2O3:Sn3 / HfOx The output current of the device corresponds to the result diagram;

[0048] Fig.13 The AND gate logic operation of the present invention simulates Ga2O3:Sn / HfO x The current variation with time of the device under single light pulse stimulation at (a) 0.5V reading voltage, (b) 1V voltage, and (c) 1V reading voltage; (d) AND gate optical logic function schematic and truth table; (e) AND gate logic operation simulation diagram. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] A Ga2O3:Sn / HfO x A method for preparing a heterojunction photoelectric sensor, the method comprising the following steps:

[0051] S1: Substrate cleaning

[0052] The substrate used in this experiment is double-polished c-plane sapphire. The cleanliness of the substrate has a crucial impact on the uniformity of film growth, adhesion and overall performance. Therefore, the substrate must be thoroughly cleaned before spin coating to remove any impurities and contaminants that may affect the quality of the film. The sapphire substrate is treated with acetone, ethanol, and deionized water for 10 minutes in turn. After cleaning, it is blown dry with nitrogen for standby use;

[0053] S2: Ga2O3:Sn thin film preparation

[0054] (1) Sol preparation

[0055] In the experiment, gallium nitrate was used as the gallium source precursor, ethylene glycol was used as the solvent, and triethanolamine was used as the stabilizer. In addition, in order to dope with tin, high-purity metal tin flakes (purity up to 99.99%) were selected. These reagents were mixed in a specific ratio to prepare a sol with a concentration of 0.8 mol / L. Among them, Ga2O3, Ga2O3:Sn1, Ga2O3:Sn2, and Ga2O3:Sn3 represent samples with undoped and tin-doped concentrations of 5%, 10%, and 15%, respectively. This preparation process needs to be carried out at room temperature, stirred for 2 hours to ensure that the components are fully mixed, and then left to stand for 24 hours so that the sol can stabilize and reach a standby state. All reagent information and specific amounts used in the experiment have been listed in detail in Tables 2-1 and 2-2.

[0056] Table 2-1 Experimental reagents

[0057]

[0058] Table 2-2 Dosage of chemicals for preparing Ga2O3:Sn thin film

[0059]

[0060] (2) Spin coating process

[0061] The prepared solution was pipetted and dripped onto the clean sapphire substrate surface to ensure that the sol could evenly cover the entire surface. Then, in order to further improve the uniform coverage of the sol and promote the volatilization of some organic solvents, the sol was spin-coated at a speed of 3000 r / min for 30 seconds, and the acceleration was set to 1000 rpm / s. Finally, a wet film was formed on the substrate surface.

[0062] (3) Drying and heat treatment

[0063] After each layer of film is spin-coated, it is placed on a hot plate and heated at 100°C for 10 minutes to evaporate some of the water and organic solvent on the surface of the substrate and dry the wet film. In order to ensure the continuity and uniformity of the film and achieve sufficient thickness, it is necessary to heat treat at 500°C for 15 minutes after each 4 layers of film are prepared, and the heating rate is controlled to be 10°C / min. During this process, a series of chemical reactions occur in the gel film, and finally a dense tin-doped gallium oxide film is formed and deposited on the substrate. This process is repeated 6 times.

[0064] (4) Post-annealing treatment

[0065] The prepared samples were placed in a tubular annealing furnace and subjected to subsequent annealing treatment at 600 °C for 2 h in an air atmosphere;

[0066] S3: Growing electrodes

[0067] Use vacuum thermal evaporation equipment to evaporate metal electrodes. The electrode materials used are gold (Au) wire or aluminum (Al) wire. The evaporation boat uses a metal tungsten boat. Grow Al (Au) metal electrodes with a thickness of 30nm on the Ga2O3:Sn film with interdigitated electrodes (10 pairs of interdigitated electrodes, each interdigitated with a length of 1000μm, a width of 100μm, and a spacing of 100μm);

[0068] S4: HfO x Thin film preparation

[0069] HkDJ xThe target material was provided by Zhongnuo New Materials (Beijing) Technology Co., Ltd. The size and purity of the target material were Φ50.8×5mm and 99.99%. Preparation of HfO by RF magnetron sputtering x The vacuum degree in the chamber is 5.0×10 -6 Torr, ignition pressure 0.05 Torr, growth pressure 7.5×10 -3 Torr, argon flow rate of 18.0 sccm, sputtering power of 60 W, and growth at room temperature.

[0070] A Ga2O3:Sn / HfO x Heterojunction photoelectric sensors for neurosynaptic biomimetic applications, including Ga2O3:Sn / HfO x Synaptic plasticity research and Ga2O3:Sn / HfO x Research on the integrated sensing, storage and computing application of photoelectric sensors.

[0071] In this embodiment: Ga2O3:Sn / HfO x Synaptic plasticity studies include Ga2O3:Sn / HfO x Photoelectric sensors simulating short-term plasticity of biological synapses with Ga2O3:Sn / HfO x Photoelectric sensors mimic long-term plasticity of biological synapses.

[0072] In this embodiment: Ga2O3:Sn / HfO x Photoelectric sensors simulate short-term plasticity of biological synapses: In the human brain, presynaptic neurons and postsynaptic neurons are connected through synapses, which are the link between presynaptic and postsynaptic neurons. The key to simulating synaptic activity is to regulate the connection strength between each neuron unit, so as to simulate the process of signal transmission and learning between neurons through synapses in the biological nervous system. In the synaptic performance test, the reading voltage is 0.5V, the light excitation is the LA-410UV light source and the light pulse is filtered by a filter with a cutoff wavelength of 380nm. A single light pulse acting on the device will trigger a postsynaptic excitation current (EPSC). At the end of the light pulse, the EPSC will reach a peak and then gradually weaken. This EPSC behavior is used to evaluate the connection strength of the synapse.

[0073] Figure 1 Shows the sandwich structure Ga2O3:Sn3 / Al / HfO x The light intensity applied to the device is 5μW / cm 2 EPSC effect triggered by a single light pulse of 2 s duration. When light is applied, the current rises rapidly, which can be attributed to the photogenerated electron-hole pairs in the device. After the light source is removed, the current drops from 21.74 nA to 19.57 nA within 100 s, realizing a slow relaxation process.

[0074] In order to compare the test results of pure Ga2O3:Sn3 / Al devices under the same conditions, Ga2O3:Sn3 / Al devices and sandwich structure Ga2O3:Sn3 / Al / HfO x The test results of the device are normalized, such as Figure 2 As shown in (a), a sandwich structure of Ga2O3:Sn3 / Al / HfO was found. x The EPSC of the device after 100s of light removal was 79.54% of the peak value, while the EPSC of pure Ga2O3:Sn3 / Al was 36.98% of the peak value, and the EPSC of HfO x The introduction of the layer increases the relaxation time of the device and increases the non-volatility, which is consistent with the results in Chapter 3. Similarly, the same results are obtained in the corresponding Au electrode device, such as Figure 2 (b) as shown. Figure 2 (c) shows the sandwich structure of Ga2O3 / Al / HfO x and Ga2O3:Sn3 / Al / HfO x Compared with the EPSC of the device, the strategy of tin-doped gallium oxide also greatly improves the relaxation time of the device. x The introduction of the layer has a synergistic effect, which greatly improves the EPSC relaxation time of the device.

[0075] The increase in EPSC relaxation time has certain benefits in simulating synaptic plasticity and provides some potential advantages for artificial intelligence and neural computing: 1. Enhanced learning and adaptability: The increase in the EPSC relaxation time of synapses can enhance the learning and adaptability of artificial neural networks, enabling artificial neural networks to better adapt to environmental changes and learn new patterns, thereby improving their performance and efficiency; 2. Increased memory capacity: The increase in EPSC relaxation time causes the persistence of synaptic connections to increase, thereby increasing the memory capacity of artificial neural networks. This feature of enhanced memory capacity enables artificial neural networks to better handle large-scale data and complex tasks; 3. Improved neural network stability: The increase in EPSC relaxation time enables synapses to better maintain the stability of neural networks. This stability helps prevent instability or oscillations in neural networks and improves their reliability in long-term operation. In summary, the increase in EPSC relaxation time helps improve the performance and efficiency of artificial intelligence systems and promote the development of the field of neural computing. In the following experiments, the sandwich structure Ga2O3:Sn3 / Al / HfO was selected. x The device was investigated (hereinafter referred to as Ga2O3:Sn / HfO x device).

[0076] In a synapse, when continuous light pulse stimulation is applied, the amplitude of the excitatory postsynaptic current induced by the second stimulus is significantly larger than the first one. This phenomenon is similar to the pulse pair facilitation (PPF) process in biological synapses. [113-114] The EPSC gain can be measured using the PPF index. Figure 3 (a) is the time interval of 2s and the light power density of 5μW / cm 2 The current change curve under the stimulation of ultraviolet light pulses. The test results show that under the action of two consecutive light pulses, the excitatory postsynaptic current shows an enhanced phenomenon. Figure 3 (b) gives the PPF index at different time intervals. The specific calculation method of the PPF index is as follows:

[0077]

[0078] Among them, A1 and A2 represent the response amplitude of the current after the first and second light pulse stimulation, respectively. It can be observed from the figure that with the increase of the time interval between the two light pulses, the PPF index shows a gradual downward trend. When the pulse interval is 0.5s, the measured PPF index is about 196%, and when the time interval is 120s, the PPF index drops to about 173%. This phenomenon is due to the fact that under the shorter pulse interval, the excitatory postsynaptic current generated by the first light stimulation is more fully retained and superimposed on the current of the second light stimulation, which is consistent with the dependence of PPF on the pulse interval in biological synapses. At a longer time interval, the degree of attenuation of PPF is relatively small; at a shorter time interval, the decrease in PPF is larger. This change better simulates the PPF behavior of biological synapses.

[0079] This PPF function can be extended to the spike rate dependent plasticity (SRDP) function by using spike trains, such as Figure 4 The UV pulse sequence consists of 15 consecutive UV pulses (with a light intensity of 5 μW / cm 2 Here, the pulse frequency is adjusted by changing the interval between two adjacent light pulses from 2s to 15s. By increasing the pulse frequency, a larger EPSC can be obtained, which is similar to the SRDP function of biological synapses.

[0080] In this embodiment: Ga2O3:Sn / HfO xPhotoelectric sensors simulate long-term plasticity of biological synapses: Biologically, memory can be roughly divided into two types according to its duration: short-term memory and long-term memory. Both short-term memory and long-term memory are closely related to synaptic plasticity. Short-term memory corresponds to the short-term plasticity (Short-Term Plasticity, STP) of neural synapses. Short-term plasticity is a temporary and short-term synaptic plasticity change, which usually lasts from a few seconds to a few minutes. Short-term plasticity makes the connection strength between neurons change rapidly in a short period of time, thereby achieving The storage and processing of short-term memory, long-term memory corresponds to the long-term plasticity (Long-Term Plasticity, LTP) of synapses. Long-term plasticity is a persistent and long-lasting synaptic plasticity change, which usually lasts for several hours or even days. Long-term plasticity enables the connection pattern between neurons to be changed stably over a long period of time, thereby realizing the storage and consolidation of long-term memory. In the experiment, the plasticity changes of biological synapses can be simulated by adjusting the various parameters of the light pulse, and the transition from short-term plasticity to long-term plasticity can be achieved.

[0081] Figure 5 The photocurrent responses under different illumination conditions are given. Changes in illumination conditions not only directly affect the peak value of EPSC, but also effectively regulate the relaxation process after the light is removed. Figure 5 (a) is the photocurrent response process under single pulse stimulation with different light intensities. When the light power density is 1 μW / cm 2 After irradiation with light, EPSC gradually decays and approaches the initial state, which represents the short-term plasticity of biological synapses. As the light power density increases, EPSC achieves a higher response amplitude and reaches a stable high current state, which represents the long-term plasticity of biological synapses. This transition can also be achieved by increasing the pulse width of the light pulse or increasing the number of pulses. Figure 5 (b) shows that the optical power is 5 μW / cm 2 As the pulse time increases from 0.5s to 2s, the peak photocurrent level achieved by the device shows a synchronous growth trend. At the same time, the relaxation time will also increase accordingly with the increase of illumination time, achieving the extension of memory time and showing obvious long-term plasticity. Repeated stimulation of synapses will increase the stability of synaptic connections. Under pulse conditions (light pulse width of 2s, light intensity of 5μW / cm 2 ) remains unchanged, different numbers of light pulses are applied to the device to stimulate the device. The experimental results are as follows Figure 5As shown in (c), as the number of light pulses increases, the relaxation time of the device after the light pulses are removed is significantly prolonged. After a period of time, the photocurrent of the device will gradually tend to and stabilize at the equilibrium level. The current level of the device when it reaches the equilibrium state will also increase with the increase in the number of light pulses. This phenomenon reveals the impact of repeated stimulation on device performance, that is, through multiple and continuous stimulation, the short-term plasticity of the device can be transformed into long-term plasticity. In summary, the transformation from short-term plasticity to long-term plasticity can be achieved by adjusting the parameters of the light pulses, such as intensity, time and number. This behavior is quite similar to the learning mechanism of the human brain, just as people consolidate and deepen their memories through continuous and high-intensity learning.

[0082] Learning-experience behaviors, as one of the important learning rules of long-term plasticity, refers to the fact that the learning and receiving process of synapses will be affected by past stimulation experiences. When a synapse receives the same or similar stimulation again after receiving a certain stimulation in the past, it will more easily transmit the corresponding information. In order to realize experiential learning in the device, the ultraviolet light irradiation corresponds to the learning process, and the removal of the ultraviolet light corresponds to the forgetting process, which successfully simulates this behavior. Figure 6 As shown, during the first learning-forgetting process, 15 μW / cm 2 The device was stimulated with light for 10 seconds, and the device reached a memory level of about 238nA. After the light signal was removed, it spontaneously relaxed to an intermediate state. The first forgetting process took about 100 seconds. In the second learning-forgetting process, the device was stimulated with light at the same light power density. Repeated learning behavior only took 1 second to reach the first learning level, and the time to forget to the forgetting benchmark after the first learning increased to about 185 seconds. With repeated learning, the memory trace faded more slowly. When the same light stimulation was performed for the third time, the current response peak value consistent with the previous two learnings was restored in a very short time (0.8 seconds). This not only verified the efficient reproducibility of the learning behavior, but also highlighted the device's ability to quickly and stably regain the learned state under multiple stimulations. After the light signal is removed, the photogenerated electrons that have not been recombined with holes will be accumulated with the photogenerated electrons generated by the next ultraviolet irradiation process, which helps to shorten the memory recovery time when learning again. This is consistent with the process of human learning and cognition of things. In the process of learning and cognition of things, the content that humans first come into contact with is often forgotten. However, when people are exposed to the same content again, they find that it takes less time to reach their previous level of cognition.

[0083] Ga2O3:Sn / HfO xThe ultraviolet light response of the device realizes the light enhancement process of the device. The device will maintain a long current relaxation time after light stimulation. This continuous photoconductivity phenomenon can be erased by applying a reverse voltage pulse, that is, light-enhanced electrical inhibition behavior. Figure 7 As shown in the figure, when the device is in the low resistance state, the device is continuously illuminated with a light intensity of 1 μW / cm 2 , a light pulse with a pulse width of 500ms and a pulse interval of 2s. The current in this process is represented by a black ball. These data show that with the continuous action of the light pulse, the current shows a trend of gradual accumulation and increase, which simulates the process of light enhancement. This light enhancement behavior can be suppressed by applying a negative electric pulse signal (amplitude of -2V, width of 100ms). After completing the 50 light pulse irradiation, 50 negative electric pulses were immediately applied. The current state of this process is represented by a red ball. Obviously, the current gradually decreases with the application of the electric pulse, which simulates the inhibitory behavior of biological neural synapses. In order to reflect the repeatability of the device, 8 consecutive cycles of testing were carried out. Each cycle included the same sequence of light pulses and negative voltage pulses. The current response of the device was relatively consistent, showing that it has good repeatability.

[0084] In this embodiment: Ga2O3:Sn / HfO x The research on the integrated sensing, storage and computing application of photoelectric sensors includes the simulation of image recognition and memory functions, and the simulation of photoelectric information demodulation and logical operation functions.

[0085] In this embodiment: Simulation of image recognition and memory function: The visual system is one of the key ways for organisms to perceive the external environment, acquire and process information. This complex and sophisticated physiological mechanism can capture light signals and convert these light signals into electrical signals through photoreceptor cells in the eyes, which are then transmitted to the brain for further analysis and integration. This ability to acquire and recognize images plays a vital role in our daily lives, helping us to understand the world, understand the environment, and make correct responses and decisions.

[0086] like Figure 8 As shown, two different light intensities (10 μW / cm 2 and 5μW / cm 2 ) light source, for the completely consistent Ga2O3:Sn / HfO xThe device was illuminated to explore its ability to recognize light signals under different lighting conditions. In order to distinguish light signals of different intensities, the device's ability to recognize the letters "L" and "T" was tested separately. When the light signal was input, the device's output current was recorded and these current signals were normalized. According to the principle that the current response intensity corresponds to the color depth, the normalized current value was converted into different color pixel unit outputs. The results showed that the area illuminated by the light signal formed a clear contrast with the unilluminated area, which strongly proved that the device has image recognition function. The input signals of different light intensities resulted in significantly different pixel block image outputs, which further verified the device's sensitivity to light signal intensity. As the light intensity increases, the device's ability to recognize and remember letters will also be stronger, because stronger light signals can more effectively activate the device, produce more obvious electrical responses, and promote more reliable letter recognition and memory processes. It is worth noting that even after 100s of stopping the light, the device can still maintain the memory of the previous light signal stimulation, which is manifested in the continuous output of the retained image information, and the images corresponding to light signals of different intensities can still be clearly distinguished. Based on the above experimental results, it is concluded that the Ga2O3:Sn / HfO x The device exhibits excellent image recognition and memory properties. It can effectively convert perceived light signals of different intensities into electrical signal outputs that are easy to identify and distinguish, and has long-term memory capabilities.

[0087] In this embodiment: Photoelectric information demodulation: Photoelectric demodulator is the core component of optical communication equipment. It is a device used for optical signal conversion and modulation. In optical communication systems, photoelectric demodulators are usually used to receive optical signals and convert them into electrical signals for demodulation and processing at the receiving end. It can convert the intensity, frequency or phase information of light waves into corresponding electrical signals, and then amplify, filter and decode them through circuits to finally extract the original information content.

[0088] 1. Text message demodulation

[0089] Figure 8 Ga2O3:Sn / HfO x The device has the ability to demodulate information related to ultraviolet light of different light intensities. Because the device responds differently to ultraviolet light of different light intensities, it can accurately demodulate and store the two bits of information carried by a single ultraviolet light pulse. Fig. 9 (a) shows that the light intensity is 1 μW / cm 2 , 5μW / cm 2 、10μW / cm 2 、15μW / cm 2The ultraviolet light signal can be demodulated into digital information "00", "01", "10" and "11" and stored. According to the American Standard Code for Information Interchange (ASCII code), "01001110" can be demodulated into the letter "N". x The device demodulates and stores "NENU", such as Fig. 9 (b) as shown.

[0090] 2. Image information demodulation

[0091] Since the device has different characteristics in response to the number and timing of pulses, it is possible to try to recognize numbers from the input image by dividing the image into five parts and stimulating the device with different time sequences of light pulses. In this process, the time sequence of light pulses is passed to the device as an input signal. The device will generate different current responses according to the time sequence of light pulses. Fig.10 As shown in (a), the image of the number "1" is selected as the input signal. The input optical information is represented as a 5×4 image, where white represents 0 and blue represents 1. Each set of optical information is divided into 5 rows, each row containing 4 pixels arranged continuously. The image information division result of the number "1" is shown in Fig.10 (b) The pixel information of each row is converted into an input pulse sequence of 4 time frames, which are then input to the Ga2O3:Sn / HfO x The time frame corresponding to the blue pixel contains the input of the ultraviolet light pulse, while the time frame corresponding to the white pixel does not contain the input of the light pulse. Due to the difference in the response of the device to different light pulses applied in different time frames, specifically, when a light pulse is applied to the device, the EPSC generated by the device becomes larger, and when multiple pulses are applied continuously to the device, the EPSC generated by the device is even larger. After the light is removed for a period of time, the EPSC of the device gradually drops to the initial state, so the input image information is represented by the photoelectric time characteristics of the device. The image information of the number "1" is extracted by jointly processing the time features in 5 input pulse streams. Five exactly the same devices are required. Each device inputs a corresponding pulse stream sequence. The pulse stream sequence is as follows: Fig.10 (c) is shown. After the input pulse sequence, the change in the current response depends on the input light pulse time sequence and can be used to analyze the input. The state of the output is determined by the collective current state of the five devices. Different light pulse time sequence inputs will result in different current states of the device, resulting in different overall states represented by all devices integrated together. In a specific setting, the different states of each device after stimulation can represent different features given in the input image, and then the state of the device is read through pre-set rules to identify the number as the "1" of the original input.

[0092] Fig.11Ga2O3:Sn / HfO x The device responds to 10 different UV light pulse timing sequences, which represent multiple possible combinations of 4 pixels arranged in different orders. The read current after each light pulse sequence represents the final state of the device, and 10 different UV light pulse sequences correspond to 10 different current states. Different light pulse sequences can be distinguished by the current state of the device, thereby accurately identifying the input optical image and obtaining the original input information.

[0093] After obtaining the final current state value that can distinguish the input light pulse sequence, the letters "n", "e", "n" and "u" are recognized to achieve more complex optical input information reading. Fig.12 As shown in (a), the experiment requires five identical devices to implement. The letters are input in the form of 5×4 images. Each image is divided into five parts, and the corresponding light pulse time series is input. Fig.12 (b) shows the experimental results. The final results are composed of 5 Ga2O3:Sn / HfO x The combination of the device's output current values ​​is represented by the test results. It can be seen from the test results that after different light pulse sequences are input into the device, the current combination of the device presents different states. x The device can correctly recognize the input images of these four original images.

[0094] In this embodiment: Logical operation function simulation: By utilizing the device's ability to integrate and process optical signals and electrical signals and their mutual conversion, a system that realizes complex logical operation functions can be constructed. Boolean logic, as a branch of mathematical logic, was founded by George Boole in the mid-19th century. Its core concept is to describe the state of a proposition with two opposing values, "true" and "false". In the digital expression system, "true" corresponds to the value 1, and "false" corresponds to the value 0. This binary logic system constitutes the basic logical unit of digital circuits and computer hardware, is the basis of computer hardware and software design, and is also the basis of logical operations in programming languages. It enables programmers to use simple logical rules to construct complex program structures, combine Boolean logic with the joint action of optoelectronic devices, and realize information processing and transmission through the logical operation function of optoelectronic devices.

[0095] In the proposed logic device, a pair of applied voltage and light pulses form two independent inputs (Opt. and Ele.) of the optoelectronic device to control the output Y. Among them, in the light input terminal Opt., the light pulse stimulation condition of light turning on is recorded as "1", and the light turning off condition is recorded as "0"; at the electrical input terminal, the voltage of 1V is a high-level input (recorded as logic "1"), and the voltage of 0.5V is a low-level input (recorded as logic "0"). When the current level exceeds 30nA, it is set as a valid signal output and outputs logic "1"; any output current below 30nA is output as logic "0". Fig.13 As shown, Ga2O3:Sn / HfO x The device implements the AND gate logic operation. Fig.13 (a) is Ga2O3:Sn / HfO x The device reads the current after being stimulated by a single light pulse at a voltage of 0.5V. After being stimulated by the light pulse, the current increases from about 11.26nA to 21.74nA. After the light is removed, the current gradually returns to the initial state. When a voltage of 1V is applied to the device, its initial dark current is 22.31nA. Fig.13 (b) As shown. After applying a single light pulse stimulus to the device at a reading voltage of 1V, the current increases to 45.04nA, as shown in Fig.13 Based on the above experimental results, this device realizes the "AND" gate logic operation under the joint regulation of light and electricity, which means that the output is "1" only when both input ends are "1". Fig.13 (d) is the signal truth table of the “AND” gate logic circuit; Fig.13 (e) Simulates the relationship between the output signal and input signal of the "AND" gate logic operation.

[0096] The present invention compares the sandwich structure Ga2O3:Sn3 / Al / HfO x The EPSC effect of the device. Experimental results show that the sandwich structure Ga2O3:Sn3 / Al / HfO x The device exhibited the longest EPSC relaxation time, with the EPSC at 100s after the light was removed being 79.54% of the peak value, while the EPSC of the Ga2O3:Sn3 / Al device under the same conditions was only 36.98% of the peak value. This significant difference indicates that HfO x The introduction of the Ga2O3:Sn3 / Al / HfO sandwich structure enhances the memory capacity and non-volatility of the device. Based on the comparison results, the Ga2O3:Sn3 / Al / HfO sandwich structure is selected. xThe device was systematically tested for synaptic bionic functions, and successfully simulated various synaptic functions, including PPF, short-term / long-term plasticity, and experiential learning behavior, by regulating various parameters of light pulses. The device's light-enhanced electrical inhibition behavior was simulated by using the synergistic effect of light pulses and electric pulses; finally, the device's non-volatile characteristics to weak ultraviolet light and excellent photoelectric response effect were used to simulate image recognition and memory functions, demodulate photoelectric information of text and images, and integrate sensing, storage, and computing functions for logical operations.

[0097] The processing results show that the Ga2O3:Sn film can be prepared by the sol-gel method with precisely controllable doping amount. The concentration of donor doping can be controlled by controlling the mass of the added tin sheet, which increases the photocurrent and response of the photoelectric sensor. The HfO x The film was used to construct a double-layer photoelectric sensor, which improved the detection rate and responsiveness of the device and had certain advantages in high-performance weak light detection. Secondly, the effects of different electrodes and different oxygen partial pressures on HfO x Thin film on Ga2O3:Sn / HfO x Effect of photoelectric sensors.

[0098] The ultraviolet signal was used to simulate a variety of synaptic functions, including EPSC, PPF, and experiential learning behavior. Furthermore, a highly linear light-enhanced electrical inhibition behavior was achieved, providing a basis for realizing high-precision image recognition functions. Finally, the device's sensing, storage, and computing technology for weak ultraviolet light was expanded, enabling the recognition and memory of images, optoelectronic information demodulation, and logical operations to be completed on one device, making up for the lack of visible light perception in certain specific application scenarios, such as ultraviolet flame detection and fingerprint recognition, and providing a new structure for the development of future neuromorphic computing and artificial vision systems.

[0099] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0100] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A Ga2O3:Sn / HfO x A method for preparing a heterojunction photoelectric sensor, characterized in that: The method comprises the following steps: S1: Substrate cleaning The substrate used in this experiment is double-polished c-plane sapphire. The cleanliness of the substrate has a crucial impact on the uniformity of film growth, adhesion and overall performance. Therefore, the substrate must be thoroughly cleaned before spin coating to remove any impurities and contaminants that may affect the quality of the film. The sapphire substrate is treated with acetone, ethanol, and deionized water for 10 minutes in turn. After cleaning, it is blown dry with nitrogen for standby use; S2: Ga2O3:Sn thin film preparation (1) Sol preparation In the experiment, gallium nitrate was used as the gallium source precursor, ethylene glycol was used as the solvent, and triethanolamine was used as the stabilizer. In addition, in order to dope with tin, high-purity metal tin flakes (purity up to 99.99%) were selected. These reagents were mixed in a specific ratio to prepare a sol with a concentration of 0.8 mol / L. Among them, Ga2O3, Ga2O3:Sn1, Ga2O3:Sn2, and Ga2O3:Sn3 represent samples with undoped and tin-doped concentrations of 5%, 10%, and 15%, respectively. This preparation process needs to be carried out at room temperature, stirred for 2 hours to ensure that the components are fully mixed, and then left to stand for 24 hours so that the sol can stabilize and reach a standby state. All reagent information and specific amounts used in the experiment have been listed in detail in Tables 2-1 and 2-2. Table 2-1 Experimental reagents Table 2-2 Dosage of chemicals for preparing Ga2O3:Sn thin film (2) Spin coating process The prepared solution was pipetted and dripped onto the clean sapphire substrate surface to ensure that the sol could evenly cover the entire surface. Then, in order to further improve the uniform coverage of the sol and promote the volatilization of some organic solvents, the sol was spin-coated at a speed of 3000 r / min for 30 seconds, and the acceleration was set to 1000 rpm / s. Finally, a wet film was formed on the substrate surface. (3) Drying and heat treatment After each layer of film is spin-coated, it is placed on a hot plate and heated at 100°C for 10 minutes to evaporate some of the water and organic solvent on the surface of the substrate and dry the wet film. In order to ensure the continuity and uniformity of the film and achieve sufficient thickness, it is necessary to heat treat at 500°C for 15 minutes after each 4 layers of film are prepared, and the heating rate is controlled to be 10°C / min. During this process, a series of chemical reactions occur in the gel film, and finally a dense tin-doped gallium oxide film is formed and deposited on the substrate. This process is repeated 6 times. (4) Post-annealing treatment The prepared samples were placed in a tubular annealing furnace and subjected to subsequent annealing treatment at 600 °C for 2 h in an air atmosphere; S3: Growing electrodes Use vacuum thermal evaporation equipment to evaporate metal electrodes. The electrode materials used are gold (Au) wire or aluminum (Al) wire. The evaporation boat uses a metal tungsten boat. Grow Al (Au) metal electrodes with a thickness of 30nm on the Ga2O3:Sn film with interdigitated electrodes (10 pairs of interdigitated electrodes, each interdigitated with a length of 1000μm, a width of 100μm, and a spacing of 100μm); S4: HfO x Thin film preparation HkDJ x The target material was provided by Zhongnuo New Materials (Beijing) Technology Co., Ltd. The size and purity of the target material were Φ50.8×5mm and 99.99%. Preparation of HfO by RF magnetron sputtering x The vacuum degree in the chamber is 5.0×10 -6 Torr, ignition pressure 0.05 Torr, growth pressure 7.5×10 -3 Torr, argon flow rate of 18.0 sccm, sputtering power of 60 W, and growth at room temperature.

2. A Ga2O3:Sn / HfO x Neural synaptic bionic application of heterojunction photoelectric sensor, based on a Ga2O3:Sn / HfO as claimed in claim 1 x A method for preparing a heterojunction photoelectric sensor, characterized in that: Including Ga2O3:Sn / HfO x Synaptic plasticity research and Ga2O3:Sn / HfO x Research on the integrated sensing, storage and computing application of photoelectric sensors.

3. A Ga2O3:Sn / HfO according to claim 2 x The neural synaptic bionic application of heterojunction photoelectric sensor is characterized by: The Ga2O3:Sn / HfO x Synaptic plasticity studies include Ga2O3:Sn / HfO x Photoelectric sensors simulating short-term plasticity of biological synapses with Ga2O3:Sn / HfO x Photoelectric sensors mimic long-term plasticity of biological synapses.

4. Ga2O3:Sn / HfO according to claim 3 x The neural synaptic bionic application of heterojunction photoelectric sensor is characterized by: The Ga2O3:Sn / HfO x Photoelectric sensors simulate short-term plasticity of biological synapses: In the human brain, presynaptic neurons and postsynaptic neurons are connected through synapses, which are the link between presynaptic and postsynaptic neurons. The key to simulating synaptic activity is to regulate the connection strength between each neuron unit, so as to simulate the process of signal transmission and learning between neurons through synapses in the biological nervous system. In the synaptic performance test, the reading voltage is 0.5V, the light excitation is the LA-410UV light source and the light pulse is filtered by a filter with a cutoff wavelength of 380nm. A single light pulse acting on the device will trigger a postsynaptic excitation current (EPSC). At the end of the light pulse, the EPSC will reach a peak and then gradually weaken. This EPSC behavior is used to evaluate the connection strength of the synapse.

5. The neural synaptic bionic application of a Ga2O3:Sn / HfOx heterojunction photoelectric sensor according to claim 3, characterized in that: The Ga2O3:Sn / HfO x Photoelectric sensors simulate long-term plasticity of biological synapses: Biologically, memory can be roughly divided into two types according to its duration: short-term memory and long-term memory. Both short-term memory and long-term memory are closely related to synaptic plasticity. Short-term memory corresponds to the short-term plasticity (Short-Term Plasticity, STP) of neural synapses. Short-term plasticity is a temporary and short-term synaptic plasticity change, which usually lasts from a few seconds to a few minutes. Short-term plasticity makes the connection strength between neurons change rapidly in a short period of time, thereby achieving The storage and processing of short-term memory, long-term memory corresponds to the long-term plasticity (Long-Term Plasticity, LTP) of synapses. Long-term plasticity is a persistent and long-lasting synaptic plasticity change, which usually lasts for several hours or even days. Long-term plasticity enables the connection pattern between neurons to be changed stably over a long period of time, thereby realizing the storage and consolidation of long-term memory. In the experiment, the plasticity changes of biological synapses can be simulated by adjusting the various parameters of the light pulse, and the transition from short-term plasticity to long-term plasticity can be achieved.

6. Ga2O3:Sn / HfO according to claim 2 x The neural synaptic bionic application of heterojunction photoelectric sensor is characterized by: The Ga2O3:Sn / HfO x The research on the integrated sensing, storage and computing application of photoelectric sensors includes the simulation of image recognition and memory functions, and the simulation of photoelectric information demodulation and logical operation functions.

7. A Ga2O3:Sn / HfO according to claim 6 x The neural synaptic bionic application of heterojunction photoelectric sensor is characterized by: Simulation of the image recognition and memory functions: The visual system is one of the key ways for organisms to perceive the external environment, acquire and process information. This complex and sophisticated physiological mechanism can capture light signals and convert these light signals into electrical signals through photoreceptor cells in the eyes, which are then transmitted to the brain for further analysis and integration. This ability to acquire and recognize images plays a vital role in our daily lives, helping us to understand the world, understand the environment, and make correct responses and decisions.

8. Ga2O3:Sn / HfO according to claim 6 x The neural synaptic bionic application of heterojunction photoelectric sensor is characterized by: The photoelectric information demodulation: The photoelectric demodulator is the core component of optical communication equipment. It is a device used for optical signal conversion and modulation. In optical communication systems, the photoelectric demodulator is usually used to receive optical signals and convert them into electrical signals for demodulation and processing at the receiving end. It can convert the intensity, frequency or phase information of the light wave into the corresponding electrical signal, and then amplify, filter and decode it through the circuit to finally extract the original information content.

9. Ga2O3:Sn / HfO according to claim 6 x The neural synaptic bionic application of heterojunction photoelectric sensor is characterized by: The logic operation function simulation: using the device's ability to integrate and process optical and electrical signals and convert them to each other, a system that can realize complex logic operation functions can be constructed. Boolean logic, as a branch of mathematical logic, was founded by George Boole in the mid-19th century. Its core concept is to use two opposing values, "true" and "false", to describe the state of a proposition. In the digital expression system, "true" corresponds to the value 1, and "false" corresponds to the value 0. This binary logic system constitutes the basic logic unit of digital circuits and computer hardware, is the basis of computer hardware and software design, and is also the basis of logical operations in programming languages. It enables programmers to use simple logical rules to construct complex program structures, combine Boolean logic with the joint action of optoelectronic devices, and realize information processing and transmission through the logical operation function of optoelectronic devices.

Images