Composite artificial synapse

Abstract

The invention relates to the technical field of artificial synapses, in particular to a composite artificial synapse which comprises a core functional layer, the core functional layer is a composite material layer, and the composite material layer is formed by mixing a SrAl2O4-based long afterglow material and a NaYF4-based up-conversion material; eu < 2 + > and Dy < 3 + > are doped in the SrAl2O4-based long afterglow material; yb < 3 + > and Tm < 3 + > are doped in the NaYF4-based up-conversion material; under excitation of near-infrared light, the NaYF4-based up-conversion material transmits energy to the SrAl2O4-based long-afterglow material through up-conversion luminescence, so that the SrAl2O4-based long-afterglow material generates long-afterglow luminescence, and a synaptic function is realized; through thermal stimulation, the release process of defect trap electrons in the SrAl2O4-based long afterglow material is regulated and controlled, and thermally induced synaptic plasticity is realized. According to the invention, SrAl2O4: Eu < 2 + >, Dy < 3 + > and NaYF4: Yb < 3 + >, Tm < 3 + > are coupled, so that long afterglow luminescence (LPL) of near-infrared triggering and thermal regulation is realized, and information encryption, handwritten number recognition and mechanical arm gesture control are realized by artificial synapses based on photo-thermal synergistic stimulation.

Description

Technical Field

[0001] This invention relates to the field of artificial synapse technology, and in particular to a composite artificial synapse. Background Technology

[0002] The persistent luminescence of lanthanide ions offers unique opportunities for information storage, optical sensing, and biomimetic computing, possessing excellent photothermal stability, tunable emission wavelength, and long afterglow time. However, their application in neuromorphic applications faces challenges related to near-infrared (NIR) excitation and multimodal modulation.

[0003] Human brain-inspired artificial neural networks (ANNs) are considered promising candidates for next-generation computing systems due to their potential to overcome the limitations of the von Neumann architecture in low power consumption and parallel information processing. The fundamental building blocks of ANNs are synapses and neurons, driving the rapid development of artificial synapses, particularly optical synapses, which offer unique advantages in ultrafast response, low crosstalk, and wide bandwidth coverage. Among various brain-inspired materials, lanthanide-doped inorganic materials with long-persistent luminescence (LPL) are attracting increasing attention because they not only enable the temporal dynamics of biological synapses to persist for seconds to hours after excitation cessation, but also enhance the stability and scalability of memory performance in next-generation neuromorphic computing systems.

[0004] Wang et al. first proposed a method based on LPL SrAl2O4:Eu 2+ ,Dy 3+ A fully optical artificial synapse was developed, excited by ultraviolet (UV) light, achieving a pairwise pulse facilitation (PPF) exponent of 165%. Lee et al. achieved multicolor synaptic triggering by combining three UV-excited luminescent materials and applied LPL-driven reservoir processes to human action recognition, demonstrating the broad prospects of LPL-based synaptic devices. Subsequently, Hao et al. proposed a bimodal artificial synapse utilizing UV-triggered luminescence of a CaSrS:Eu layer and mechanoluminescence of a ZnS:Cu layer. Image recognition accuracy reached 92.5% using different transparency and hardness levels, providing a promising strategy for constructing multimodal artificial synapses based on lanthanide ion LPL. Meanwhile, Dong et al. used only one Li... 0.1 Na 0.9 NbO3:Pr 3+ A bimodal optical synapse was obtained by simultaneously simulating ultraviolet light and force using phosphors. This bimodal optical synapse exhibited excellent performance in hardware-level noise reduction. Although the development of multimodal all-optical synapses based on lanthanide ion LPL is rapid, it is still in its early stages. For example, multimodal all-optical synapses with near-infrared (NIR) excitation need to overcome the problem of the human eye's insensitivity to NIR. In addition, the practical hardware-level applications of lanthanide ion LPL synapses remain to be explored. Summary of the Invention

[0005] In view of the problems existing in the above or prior art, the present invention is proposed.

[0006] Therefore, the object of this invention is to provide a composite artificial synapse.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a composite artificial synapse, comprising, Core functional layer; The core functional layer is a composite material layer, which is composed of a mixture of SrAl2O4-based long afterglow material and NaYF4-based upconversion material. The SrAl2O4-based long afterglow material is doped with Eu. 2+ and Dy 3+ ; The NaYF4-based upconversion material is doped with Yb. 3+ and Tm 3+ ; Under near-infrared light excitation, the NaYF4-based upconversion material transfers energy to the SrAl2O4-based long-afterglow material through upconversion luminescence, enabling the SrAl2O4-based long-afterglow material to produce long-afterglow luminescence and realize synaptic function. Thermo-induced synaptic plasticity is achieved by regulating the release process of defect-trapped electrons in SrAl2O4-based long afterglow materials through thermal stimulation.

[0008] As a preferred embodiment of the composite artificial synapse of the present invention, wherein: the SrAl2O4-based long afterglow material contains Eu 2 + The doping concentration is 6 at%, Dy 3+ The doping level is 4 at%; Yb in the NaYF4-based upconversion material 3+ The doping concentration is 20 at%, Tm 3+ The doping level is 1 at%.

[0009] As a preferred embodiment of the composite artificial synapse of the present invention, the weight ratio of the SrAl2O4-based long afterglow material to the NaYF4-based upconversion material is 2:1.

[0010] As a preferred embodiment of the composite artificial synapse of the present invention, the composite material layer is in the form of a thin film or a particle aggregate.

[0011] As a preferred embodiment of the composite artificial synapse of the present invention, the NaYF4-based upconversion material is synthesized by hydrothermal synthesis.

[0012] This invention also proposes an information encryption system based on composite artificial synapses, comprising, Synaptic array module: an array composed of the aforementioned composite artificial synapses; Excitation module: can selectively output near-infrared light or ultraviolet light, the excitation module is used to excite the synaptic array module to produce long afterglow luminescence; Signal acquisition module: The signal acquisition module is a photoelectric sensor used to acquire the long afterglow emission intensity at different time points after the excitation module stops excitation, and to calculate the synaptic weights; Encoding conversion module: The encoding conversion module converts the synaptic weights obtained by the signal acquisition module into binary code, and then converts the binary code into character information according to the encoding rules; Time-varying control module: Used to control the time interval between the excitation module stopping excitation and the signal acquisition module acquiring the signal, so as to realize the encryption of time-varying information.

[0013] As a preferred embodiment of the information encryption system based on composite artificial synapses of the present invention, when the excitation wavelength output by the excitation module deviates, the encoding conversion module outputs an error character to achieve anti-interference encryption.

[0014] This invention also proposes a handwritten digit recognition system based on composite artificial synapses, comprising, Synaptic weight mapping module: The change in long afterglow luminescence intensity of the composite artificial synapse under pulsed light stimulation is used as the synaptic weight; Artificial neural network modules include an input layer, hidden layers, and an output layer; The number of neurons in the input layer is matched with the number of pixels in the handwritten digit image to be recognized, and is used to receive signals corresponding to the image pixels; The hidden layer contains at least one neuron and is configured with an activation function adapted to the training and recognition of artificial neural networks; The number of neurons in the output layer is matched to the number of categories of the handwritten digits to be recognized; Image input module: Used to import handwritten digit image data and preprocess the image data; Recognition Calculation Module: The forward propagation algorithm is used to calculate the output of the artificial neural network module. The network parameters are trained by combining the cross-entropy loss function and the gradient descent algorithm, and the recognition result of handwritten digits is output. Accuracy calibration module: used to calculate the accuracy of recognition results.

[0015] As a preferred embodiment of the handwritten digit recognition system based on composite artificial synapses of the present invention, wherein: the connection weights between the input layer and the hidden layer of the artificial neural network module, and the connection weights between the hidden layer and the output layer are all provided by the synaptic weight mapping module; The preprocessing includes removing image noise, unifying image grayscale features, and standardizing image pixel values ​​to meet the input requirements of the artificial neural network module.

[0016] This invention also proposes a robotic arm gesture control system based on composite artificial synapses, comprising, Synaptic sensing module: composed of at least one of the aforementioned composite artificial synapses, the number of which corresponds one-to-one with the movable operating parts of the robotic arm, used to sense multimodal stimuli and generate long-persistence light emission signals; Multimodal excitation and detection module: including optical excitation unit, temperature stimulation unit and optical signal detection unit; The photoexcitation unit is used to excite the composite artificial synapse in the synaptic sensing module to generate a long-afterglow luminescence signal. The temperature stimulation unit is used to provide adjustable temperature stimulation to regulate the release process of defect trap electrons in the composite artificial synapse; The optical signal detection unit is used to detect the long-persistence light emission signal generated by the composite artificial synapse and convert it into a processable signal. Signal processing module: Used to receive the signal output by the optical signal detection unit, distinguish between the thermal excitation state and the room temperature excitation state of the composite artificial synapse according to the difference in signal intensity, and output the corresponding state judgment result; Robotic arm drive module: Connects to the power transmission mechanism of the robotic arm. Based on the status judgment result of the signal processing module, it controls the bending or extending movements of the corresponding movable parts of the robotic arm to realize the preset gesture. Thermal stress feedback module: When the robotic arm comes into contact with an external object, if the temperature stimulation unit detects that the object temperature has reached the preset thermal threshold, the composite artificial synapse generates an enhanced long afterglow luminescence signal. After receiving the enhanced signal, the signal processing module controls the robotic arm drive module to perform a retraction protection action.

[0017] The beneficial effects of this invention: by using SrAl2O4:Eu 2+ ,Dy 3+ With NaYF4:Yb 3+ ,Tm 3+ Coupling was used to achieve near-infrared triggered and thermally modulated long afterglow luminescence (LPL). This was achieved using NaYF4:Yb 3+ ,Tm 3+ Energy transfer upconversion luminescence to SrAl2O4:Eu 2+ ,Dy 3+Energy transfer in the excited state enabled near-infrared triggered synaptic function, with paired pulse promotion (PPF) increased to 213% under 980nm excitation. Simultaneously, LPL modulation was achieved by regulating the release of defect-trapped electrons through thermal activation kinetics. Furthermore, artificial synapses based on photothermal synapses achieved information encryption, handwritten digit recognition, and robotic arm gesture control with a recognition accuracy of 91.12%. Additionally, a visual-tactile bimodal synapse was proposed to simulate the physiological response of grasping a hot cup. These findings open the possibility for the application of LPL-based multimodal synapses in brain-inspired photonics, intelligent sensing, and human-computer interaction. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is an artificial visual-tactile perception system based on biological synapses.

[0020] Figure 2 Near-infrared light-triggered long afterglow luminescence in SrAl2O4–NaYF4 composite materials.

[0021] Figure 3 Synaptic plasticity of SrAl2O4–NaYF4 composite material.

[0022] Figure 4 For the luminescence mechanism and thermal trap release behavior.

[0023] Figure 5 For information encryption system and handwritten digit recognition based on SrAl2O4–NaYF4 artificial synapses.

[0024] Figure 6 This study aims to achieve gesture control and thermal simulation of robotic arms based on photothermal dual-mode artificial synapses.

[0025] Figure 7 This is an X-ray diffraction (XRD) pattern.

[0026] Figure 8 For SEM images and corresponding NaYF4:20%Yb 3+ 1%Tm 3+ and SrAl2O4:6%Eu 2+ 4%Dy 3+ EDS mapping.

[0027] Figure 9 SrAl2O4:Eu2+ ,Dy 3+ –NaYF4:20%Yb 3+ ,x%Tm 3+ (x=0.1%,0.5%,1%,3%,5%,15%) Emission spectrum of composite material after 5 seconds of stopping 980 nm excitation.

[0028] Figure 10 The emission spectrum of the SrAl2O4-NaYF4 composite material under seven 980 nm laser pulses is shown.

[0029] Figure 11 SrAl2O4:Eu 2+ ,Dy 3+ The image shows the result after 10 seconds of pre-charging at 365 nanometers.

[0030] Figure 12 An integrated 3×4 array of SrAl2O4-NaYF4 composite.

[0031] Figure 13 The pulse counts were measured for the emission spectrum of the SrAl2O4-NaYF4 composite material under 980 nm excitation.

[0032] Figure 14 Optical images, binary codes, and information of SrAl2O4-NaYF4 composites from a 3×4 array under 365 nm excitation.

[0033] Figure 15 Synaptic plasticity in SrAl2O4-NaYF4 composites simulated under ultraviolet light. Detailed Implementation

[0034] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0035] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0036] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0037] Example 1 is the first embodiment of the present invention. This embodiment provides a composite artificial synapse, including a core functional layer; wherein the core functional layer is a composite material layer, which is composed of a mixture of SrAl2O4-based long afterglow material and NaYF4-based upconversion material; SrAl2O4-based long afterglow material doped with Eu 2+ and Dy 3+ ; NaYF4-based upconversion materials are doped with Yb 3+ and Tm 3+ ; Under 980nm near-infrared light excitation, NaYF4-based upconversion material transfers energy to SrAl2O4-based long-afterglow material through upconversion luminescence, enabling SrAl2O4-based long-afterglow material to produce long-afterglow luminescence (LPL) to achieve synaptic function. The release process of defect-trapped electrons in SrAl2O4-based long afterglow materials can be regulated by thermal stimulation at 303K-503K, thereby achieving thermally induced synaptic plasticity.

[0038] Specifically, Eu in SrAl2O4-based long afterglow materials 2+ The doping concentration is 6 at%, Dy 3+ The doping level is 4 at%; Yb in NaYF4-based upconversion materials 3+ The doping concentration is 20 at%, Tm 3+ The doping level is 1 at%.

[0039] Furthermore, the mixing weight ratio of SrAl2O4-based long afterglow material to NaYF4-based upconversion material is 2:1.

[0040] The composite material layer is either a thin film or a particle aggregate.

[0041] Preferably, NaYF4-based upconversion materials are synthesized using a hydrothermal method.

[0042] Reference Figure 1 , Figure 1 This is an artificial vision-touch perception system based on biological synapses, wherein (a) is SrAl2O4:Eu 2 + ,Dy 3+ and NaYF4:Yb 3+ ,Tm 3+ (a) Schematic diagram of a bimodal synapse triggered by a 980nm thermal pulse in a composite material. (b) Artificial neural network (ANN) for handwritten digit recognition using near-infrared (NIR) triggering of visual synapses. (c) Gesture control of a robotic arm based on a thermally simulated tactile system.

[0043] The human brain is the hub of the central nervous system, responsible for integrating sensory information such as vision, touch, smell, hearing, and taste to achieve memory and learning. Of external information, 80% is obtained through visible light perception via the retina. Tactile reception is another important way to acquire external information, such as temperature, humidity, pain, pressure, and vibration. Inspired by the human multimodal sensory architecture, an artificial vision-touch sensory system (…) has been constructed. Figure 1 Visual perception compensates for the insensitivity of retinal sensors by using near-infrared simulation, enabling the recognition of handwritten digits; while tactile perception controls the bending and gestures of the robotic arm through thermal stimulation.

[0044] The visual and tactile sensory systems communicate via SrAl2O4: 6%Eu 2+ 4% Dy 3+ and NaYF4:20%Yb 3+ 1%Tm 3+ The nanoparticles (SrAl2O4–NaYF4) were mixed in a 2:1 weight ratio. NaYF4: 20% Yb 3+ 1%Tm 3+ The nanoparticles were synthesized via a hydrothermal method, and the composition of SrAl2O4 was 6% Eu. 2+ 4%Dy 3+ Purchased from Shenzhen Looking Long Technology Co., Ltd. (Reference) Figure 7 The X-ray diffraction (XRD) pattern shows that NaYF4:20%Yb 3+ 1%Tm 3+ The nanoparticles are pure hexagonal phase, SrAl2O4:6%Eu 2+ 4%Dy 3+ It is a pure monoclinic phase. Figure 7 In the middle: (a) is NaYF4:20%Yb 3+ 1%Tm 3+ (b) is SrAl2O4:6%Eu 2+ 4%Dy 3+ .

[0045] Reference Figure 8 , Figure 8 SEM (Scanning Electron Microscopy) images and corresponding NaYF4:20%Yb 3+ 1%Tm 3+ and SrAl2O4:6%Eu 2+ 4%Dy 3+ EDS (Energy Dispersive X-ray Spectrometry) mapping, where SrAl2O4:6%Eu 2+ 4%Dy 3+ and NaYF4:20%Yb 3+ 1%Tm3+ Scanning electron microscopy (SEM) images showed a uniform distribution. Furthermore, energy-dispersive spectroscopy (EDS) analysis revealed a chemical composition consistent with the XRD pattern. SEM images showed rod-shaped particles with a length of ~1.3 μm and a diameter of ~320 nm. EDS analysis of Na, Y, F, Yb, and Tm confirmed the presence of the host elements (Na, Y, F) and dopant ions (Yb). 3+ and Tm 3+ The uniform spatial distribution of Yb indicates that 3+ and Tm 3+ Uniformly incorporated into NaYF4. From SrAl2O4:6%Eu 2+ 4%Dy 3+ EDS of Sr, Al, O, Eu, and Dy confirmed the presence of host elements (Sr, Al, O) and dopant ions (Eu). 2+ and Dy 3+ The uniform spatial distribution of Eu indicates that 2+ and Dy 3+ It is uniformly incorporated into SrAl2O4.

[0046] Reference Figure 2 , Figure 2 Near-infrared light-triggered long afterglow luminescence in SrAl2O4–NaYF4 composite materials. (a) SrAl2O4:Eu 2+ ,Dy 3+ Excitation spectrum monitored at 522 nm. (b) SrAl2O4:Eu 2+ ,Dy 3+ NaYF4:Yb 3+ ,Tm 3+ The upconversion emission spectrum of the SrAl2O4–NaYF4 composite material under 980 nm excitation. (c) is the SrAl2O4:Eu 2+ ,Dy 3+ SrAl2O4:Eu 2+ ,Dy 3+ NaYF4:Yb 3+ ,Tm 3+ The amplified emission spectra of the SrAl2O4–NaYF4 composite material under 980 nm excitation are shown. (d) is the emission spectrum of the SrAl2O4–NaYF4 composite material 5 s after 980 nm excitation is stopped. (e) is the emission spectrum of SrAl2O4:Eu 2+ ,Dy 3+Emission spectra under excitation at 452 nm, 477 nm, 647 nm, 698 nm, and 800 nm, respectively. (f) shows the luminescence decay curve of the SrAl2O4–NaYF4 composite material after irradiation at 980 nm for 60 s, with the emission signal collected after a 5 s delay. (g) shows the optical image of the SrAl2O4–NaYF4 composite material after 980 nm excitation was stopped.

[0047] To achieve near-infrared light-triggered artificial synapses, long-afterglow luminescence excited by near-infrared light was first developed in a SrAl2O4–NaYF4 composite material. (SrAl2O4:Eu) 2+ ,Dy 3+ The excitation spectrum exhibits strong absorption in the 420–500 nm range, attributed to Eu. 2+ 4f of ions 7 →4f 6 5d 1 The transition indicates that the material can be efficiently excited by visible blue light ( Figure 2 a). Simultaneously, under 980nm excitation, NaYF4:Yb 3+ ,Tm 3+ The nanoparticles exhibited five emission peaks with center wavelengths at 452 nm, 477 nm, 647 nm, 698 nm, and 800 nm, which are attributed to [various factors]. 1 D2→ 3 F4, 1 G4→ 3 H6, 1 G4→ 3 F4, 3 F3→ 3 H6 and 3 H4→ 3 H6, jump ( Figure 2 b). The significant peaks at 452 nm and 477 nm are related to SrAl2O4:Eu 2+ ,Dy 3+ The excitation windows overlap well, indicating the presence of NaYF4:Yb. 3+ ,Tm 3+ After nanoparticles, SrAl2O4:Eu 2+ ,Dy 3+ It can be excited at 980 nm and exhibits luminescence. The composite material exhibits broadband luminescence of 500–600 nm under 980 nm excitation. Figure 2 c), while SrAl2O4:Eu 2+ ,Dy 3+ The absence of a luminescent band rules out the possibility of direct absorption of 980nm photons. After 5 seconds of cessation of 980nm excitation, the emission spectrum of the SrAl2O4–NaYF4 composite material differs from that of the SrAl2O4:Eu composite material.2+ ,Dy 3+ Broadband emission consistency under 452nm excitation ( Figure 2 (d and e) confirmed the presence of SrAl2O4:Eu in the composite material. 2+ ,Dy 3+ The emission can be simulated using 980nm.

[0048] In addition, refer to Figure 9 , Figure 9 SrAl2O4:Eu 2+ ,Dy 3+ –NaYF4:20%Yb 3+ ,x%Tm 3+ The emission spectra of the composite material (x=0.1%, 0.5%, 1%, 3%, 5%, 15%) after 5 seconds of cessation of 980 nm excitation were obtained by optimizing the NaYF4:20%Yb composition. 3+ ,x%Tm 3+ Tm in nanoparticles 3+ By reducing the doping concentration to 1%, the strongest blue light emission band and SrAl2O4:Eu under 980nm excitation were achieved. 2+ ,Dy 3+ The light emitted.

[0049] When the 980 nm excitation is turned off ( Figure 9 In the emission spectrum collected 5 seconds later, the composite material exhibited a high degree of Eu in the 500-600 nm range. 2+ 4f 6 5d 1 →4f 7 The transition produces long-lasting luminescence. With Tm 3+ Increasing the content from 0.1% to 1% increases emission intensity, but it decreases significantly at higher concentrations (3%, 5%, and 15%). Therefore, the optimal Tm... 3+ The doping is approximately 1%, therefore x = 1%Tm is used. 3+ Further experiments will be conducted.

[0050] To evaluate the long-persistent luminescence capability, the luminescence decay curves of the composite material under 980 nm excitation were measured. The decay curves showed that a weak luminescence intensity remained 105 s after cessation of 980 nm excitation, indicating that the composite material possesses long-persistent luminescence capability. The decay curves were calculated using the formula... Fitting: where , , and Let represent the emission intensity, offset intensity, initial intensity of the i-th component, and corresponding decay time constant at time t, respectively. The fitting parameters are summarized in Table 1.

[0051] Table 1. Fitting parameters for the exponential decay curve of the SrAl2O4-NaYF4 composite material.

[0052]

[0053] The average decay lifetime of the composite material was calculated to be 10⁸ s. SrAl₂O₄:Eu 2+ ,Dy 3+ The long afterglow behavior originates from electron trapping-release processes associated with defect states. Tm 3+ The 1D2 and 1G4 excited states undergo radiative transitions via an upconversion process, pumping energy to Eu via luminescent resonant energy transfer. 2+ The 5d excited state. A portion of the electrons in the 5d excited state are captured by oxygen vacancies and Dy traps through multiphonon relaxation. When the 980 nm excitation is stopped, the captured electrons are released into the Eu region through thermal activation. 2+ The 5d excited state was obtained, producing a long afterglow centered at 522 nm with a lifetime of up to 108 s. Even 60 min after stopping 980 nm excitation, the optical image of the SrAl2O4–NaYF4 composite material was still observable with the naked eye. Figure 2 g), further confirming the long afterglow capability of the composite material. The long afterglow capability triggered by 980nm lays the foundation for realizing artificial synapses in SrAl2O4–NaYF4 composite material under near-infrared light excitation.

[0054] Reference Figure 3 , Figure 3 Synaptic plasticity of SrAl2O4–NaYF4 composites is shown in the following figures: (a) emission spectra of the composite triggered by a single 980 nm pulse with different powers (duration = 0.2 s); (b) emission spectra of the composite triggered by a single 980 nm pulse with different pulse widths (power = 128 mW); (c) dependence of PL intensity on pulse duration at different excitation powers; (d) emission intensity triggered by two consecutive light pulses; and (e) PPF exponent as a function of the time interval between two consecutive light pulses. (f) The transition from STP to LTP by increasing pulse power (duration = 0.1 sec, interval = 0.1 sec). (g) The transition from STP to LTP by increasing the number of pulses (duration = 0.15 sec, interval = 0.15 sec). (h) The learning-experience behavior of the composite material under 980 nm pulses.

[0055] The artificial synaptic plasticity of SrAl2O4–NaYF4 composites was evaluated by varying the duration, power, and number of optical pulses. When triggered by a single 980 nm pulse with a fixed duration of 0.2 seconds (…),… Figure 3 a) A single emission band centered at 522 nm was detected. The emission intensity gradually increased as the pulse power increased from 128 mW to 349 mW. This trend can be attributed to the emission from NaYF4:20%Yb. 3+ 1%Tm 3+ The luminescent resonant energy transfer of nanoparticles enables Eu 2+ The number of pumped electrons in the 5d state increases, and at the same time, SrAl2O4:Eu 2+ ,Dy 3+ Eu 2+ The 5d-4f radiative transition is enhanced with increasing excitation power. Similarly, at a fixed power of 128 mW, a gradual increase in luminescence intensity is also observed with increasing pulse duration. Figure 3 b). The luminescence intensity exhibits an approximately linear growth behavior with increasing excitation power and pulse duration. Figure 3 c). These tunable strengths demonstrate the ability of synapses to respond according to stimulation mechanisms, indicating that the SrAl2O4–NaYF4 composite material exhibits synaptic plasticity under 980 nm excitation.

[0056] PPF was then investigated because it is a key factor in short-term synaptic plasticity (STP) in biological synapses. When two consecutive 980 nm pulses were applied to the composite material with a pulse duration of 0.1 seconds and an interval of 0.15 seconds, two emission peaks were observed. Figure 3 d). The second emission peak is significantly enhanced compared to the first. The PPF is as high as 213%, indicating that our composite material has stronger synaptic plasticity. The PPF exponent gradually decreases with increasing pulse interval ( Figure 3 e). Decrease occurs via a double exponential function. Fitting: where A1 and A2 are the initial facilitated amplitudes, and Δt is the pulse interval. Calculation results are obtained. The relaxation times are 0.063 seconds and 19.9 seconds, respectively. The exponential decay of the PPF highlights its potential for time-dependent information processing.

[0057] The LPL properties of the SrAl2O4–NaYF4 composite material were used to simulate STP, long-term plasticity (LTP), and the transitions between them, properties widely considered fundamental to learning and decision-making processes in biological systems. Under low-power optical stimulation, the LPL intensity of the composite material decayed rapidly, corresponding to STP and short-term memory (STM).

[0058] Reference Figure 10 , Figure 10The emission spectra of the SrAl2O4-NaYF4 composite material under seven 980 nm laser pulses with different excitation powers (pulse duration = 0.1 s, interval = 0.1 s). The inserted image is a magnified emission spectrum within the blurred box. With increasing pulse power, the afterglow decays more slowly and for a longer duration, indicating a transition to LTP and LTM. Figure 3 f and 10). Similarly, as the number of 980nm pulses increases, the LPL intensity gradually increases, and the LPL intensity is enhanced at the same decay time, while the time to recover to the original state is prolonged. Therefore, increasing the number of excitation pulses also induces the transition from STP to LTP ( Figure 3 g). As the number of impulses increases, the enhanced response strength and prolonged memory retention are consistent with the principle that repetitive stimulation promotes memory consolidation in the human brain.

[0059] It is worth noting that this composite material also exhibits synaptic behavior of "learning-forgetting-relearning," such as... Figure 3 As shown in h. During the learning process, a 15-pulse sequence at 980nm accelerated Eu. 2+ Pumping of the 5d excited state and promotion of interaction with oxygen vacancies and Dy 3+ The accumulation of carriers in the dopant-associated traps leads to an increase in the LPL intensity from 345 to 644, corresponding to memory acquisition. Upon cessation of external stimulation, radiative and non-radiative transitions of electrons in the excited state cause a sharp decrease in carrier density and LPL intensity, corresponding to the forgetting process. Subsequently, when the same pulse is applied, the LPL intensity recovers to its original value on the 5th pulse, consistent with the relearning phase. This rapid relearning phase is attributed to the accumulation of carriers in the traps during the learning process. Incomplete release of electrons from the traps inhibits repeated trapping and promotes electron re-entry into the Eu field. 2+ Accumulation in the excited state leads to a rapid increase in luminescence intensity and a shortened relearning time.

[0060] Reference Figure 4 , Figure 4 The light emission mechanism and thermally induced trap release behavior are shown in Figure 1, where (a) is the light emission spectrum at different temperatures after ultraviolet irradiation, and (b) is the light emission spectrum of SrAl2O4:Eu. 2+ ,Dy 3+ Thermoluminescence curves after 10 seconds of pre-irradiation with a 365nm lamp. (c) The change in luminescence intensity at 373K over time after 10 seconds of 365nm excitation. (d) An optical image showing the change in luminescence intensity at 373K over time after 10 seconds of 365nm excitation. (e) A schematic diagram of the microscopic mechanism of the composite material excited by 980nm.

[0061] The long afterglow luminescence of phonons associated with composite materials was investigated in SrAl2O4:Eu 2+ ,Dy3+ The thermal activation trap dynamics in the image. After ultraviolet irradiation, the luminescence intensity first increased from 303 K to 363 K, and then decreased from 363 K to 503 K. Figure 4 a) The increase in intensity is attributed to the rapid release of electrons from the traps due to increased temperature, while the decrease in intensity upon further heating is attributed to the near or complete release of electrons from specific traps. To quantitatively understand SrAl2O4:Eu 2+ ,Dy 3+ Defects in the sample were investigated, and thermoluminescence (TL) measurements were performed after ultraviolet charging. Figure 4 b). Two distinct TL peaks were observed in the temperature range of 300–600 K, located near 352 K and 425 K, respectively. Different TL peaks correspond to different traps, and different thermal activation energies are required to release the trapped electrons. The TL emission spectrum was decomposed by multi-Gaussian fitting, and empirical formulas were used. Estimate SrAl2O4:Eu 2+ ,Dy 3+ The depth of the trap. The value represents the temperature corresponding to the TL peak, and E is the trap depth. The calculated trap depths are 0.704 eV and 0.85 eV, corresponding to TL peaks at 352 K and 425 K, respectively. This engineered trap depth distribution can be attributed to Dy. 3+ Defects caused by dopants and intrinsic defects such as oxygen vacancies. Electrons are emitted through the absorption of ultraviolet phonons or from NaYF4:20%Yb excited at 980 nm. 3+ 1%Tm 3+ Nanoparticles are pumped from the valence band to the conduction band via luminescent resonance energy transfer. Figure 4 (e) Subsequently, a luminescent band centered at 522 nm is generated through radiative transitions, or the electrons are trapped via multiphonon relaxation. These trapped electrons can be thermally released and recombine with holes in the valence band, producing persistent afterglow emission through 5d-4f transitions. This trap-release process is analogous to information storage and delayed release in biological synapses.

[0062] To further elucidate the dynamics of trap release, isothermal TL (twitch velocity) was measured at 373 K. Figure 4 (c and d). The samples were first excited at 365 nm for 10 seconds, then left to stand for 30 seconds, followed by heating to 373 K. The change in luminescence intensity over time was recorded. The results showed that the intensity initially increased rapidly and then gradually decreased, which is consistent with the microscopic mechanism of electron release from specific traps under thermal activation.

[0063] Reference Figure 11 , Figure 11 SrAl2O4:Eu 2+ ,Dy 3+After a 10-second pre-charge at 365 nm, followed by a 30-second rest in darkness, and then heating to 373 K or 423 K, the time-dependent sustained luminescence intensity was observed. Notably, the luminescence intensity was higher and decayed faster at 423 K, indicating that higher temperatures can accelerate the release of trapped electrons. Figure 11 ).

[0064] This thermally activated luminescence provides a new strategy for developing multimodal artificial synaptic devices coupled with temperature sensors.

[0065] Example 2, refer to Figure 5 and Figure 12 This is the second embodiment of the present invention. Unlike the previous embodiment, this embodiment provides an information encryption system based on composite artificial synapses, including: Synaptic array module: An array of composite artificial synapses arranged in 3 rows and 4 columns; see reference. Figure 12 , Figure 12 An integrated 3×4 array of SrAl2O4-NaYF4 composite was used to demonstrate the feasibility of optical information encoding.

[0066] Excitation module: can selectively output 980nm near-infrared light or 365nm ultraviolet light. The excitation module is used to excite the synaptic array module to produce long afterglow emission (LPL). Among them, the pulse width of 980nm near-infrared light is 0.1-0.2s and the excitation power is 304-447mW; the irradiation intensity of 365nm ultraviolet light is 10-20mW / cm² and the pre-irradiation time is 10-15s. Signal acquisition module: The signal acquisition module is equipped with a photoelectric sensor with a detection wavelength of 522nm, which is used to acquire the long afterglow emission (LPL) intensity at different time points after the excitation module stops excitation, and calculate the synaptic weight (synaptic weight = LPL intensity corresponding to the i-th excitation pulse / LPL intensity corresponding to the 1st excitation pulse × 100%). Encoding Conversion Module: The encoding conversion module converts the synaptic weights obtained by the signal acquisition module into binary code, and then converts the binary code into character information according to the ASCII encoding rules; Among them, the mapping rules between pre-stored synaptic weights and binary codes are as follows: synaptic weights of 100%-110% correspond to binary codes "01", 110%-115% correspond to "00", 115%-120% correspond to "10", and 120%-130% correspond to "11". Time-varying control module: Used to control the time interval between the excitation module stopping excitation and the signal acquisition module acquiring the signal. The time interval includes 5s, 10s, 30s, 1h, and 4h to achieve time-varying information encryption.

[0067] Specifically, when the excitation wavelength output by the excitation module deviates from 980nm±5nm or 365nm±5nm, the encoding conversion module outputs an error character to achieve anti-interference encryption.

[0068] Reference Figure 5 , Figure 5 This paper presents an information encryption system based on SrAl2O4–NaYF4 artificial synapses and its application in handwritten digit recognition. (a) shows the synaptic weight modulation using seven pulses in the SrAl2O4–NaYF4 artificial synapse excited by 980nm light at 477mW. (b) is a schematic diagram illustrating the conversion principle between synaptic weights, visual signals, and binary codes. (c) demonstrates the time-varying information encryption system. (d) is a schematic diagram of the artificial neural network (ANN) used for handwritten digit recognition. (e) shows the function of the recognition accuracy of a 28×28 pixel handwritten digit image as a function of the number of training epochs.

[0069] By constructing a 3×4 array of SrAl2O4–NaYF4 composite material ( Figure 12 This study explored time-varying information encryption, in which synaptic weights were used as binary codes. Synaptic weights are defined as... ,in, This is the luminescence intensity triggered by the first 980nm pulse. It is the emission intensity triggered by the i-th 980nm pulse ( Figure 5 a). The relationship between binary code and synaptic weights is as follows: Figure 5 b and Table 2 are shown.

[0070] Table 2 Relationship between the binary code and synaptic weight of SrAl2O4 synapses

[0071] Matrix information can be sent to the receiver, for example The numbers in the matrix represent the number of pulses triggered by the corresponding sample.

[0072] Reference Figure 13 The emission spectra of the SrAl2O4-NaYF4 composite material were obtained from a 3×4 array and measured under 980 nm excitation, with different pulse counts for each.

[0073] When a trigger pulse is applied, based on the synaptic weights of the SrAl2O4 composite material under 980 nm pulse triggering, the binary code matrix will appear as follows 5 seconds after the 980 nm simulation: Matrix information will be converted according to ACSI. Subsequently, as LPL decays, 10 seconds after the near-infrared (NIR) simulation, the binary code matrix will appear as... This generates time-varying information. Finally, 4 hours after the NIR simulation, the time-varying binary code matrix and information will be displayed as follows: and ( Figure 5 c).

[0074] Reference Figure 14 , Figure 14 Optical images, binary codes, and information of SrAl₂O₄-NaYF₄ composite materials from a 3×4 array excited at 365 nm. Furthermore, differences in the wavelength and power of the excitation light can produce erroneous information. Figure 14 Therefore, time-varying information transmission and encryption can be achieved, providing a new strategy for next-generation secure communication, advanced anti-counterfeiting of high-end goods, and intelligent authentication systems for electronic passports and advanced identity verification.

[0075] Example 3, referring to Figure 5 This is the third embodiment of the present invention. Unlike the previous embodiment, this embodiment provides a handwritten digit recognition system based on a composite artificial synapse, comprising: Synaptic weight mapping module: The change in long afterglow emission (LPL) intensity of the composite artificial synapse under pulsed light stimulation is used as the synaptic weight; Artificial neural network modules include an input layer, hidden layers, and an output layer; The number of neurons in the input layer matches the number of pixels in the handwritten digit image to be recognized, and is used to receive the signals corresponding to the image pixels; The hidden layer contains at least one neuron and is configured with an activation function adapted to the training and recognition of artificial neural networks; The number of neurons in the output layer matches the number of categories of the handwritten digits to be recognized; Specifically, the input layer contains 784 neurons, corresponding to a 28×28 pixel handwritten digit image; The hidden layer contains 300 neurons, and the activation function is the ReLU function; The output layer contains 10 neurons, corresponding to the numbers 0-9 respectively; The connection weights between the input layer and the hidden layer, and the connection weights between the hidden layer and the output layer of the artificial neural network module are all provided by the synaptic weight mapping module; Image input module: Used to import handwritten digit image data in MNIST format and to preprocess the image data; Preprocessing includes removing image noise, unifying image grayscale features, and standardizing image pixel values ​​to meet the input requirements of artificial neural network modules; Recognition Calculation Module: The forward propagation algorithm is used to calculate the output of the artificial neural network module. The network parameters are trained by combining the cross-entropy loss function and the gradient descent algorithm, and the recognition result of handwritten digits is output. The gradient descent algorithm has a learning rate of 0.001-0.01 and a batch size of 32-64. Accuracy calibration module: used to statistically analyze the accuracy of recognition results, ensuring that the system's recognition accuracy on the MNIST test set (10,000 samples) is ≥91.12%±0.5%.

[0076] To further verify the visual information processing capabilities of the SrAl2O4–NaYF4 synapse, an artificial neural network (ANN) simulation was performed using the CrossSim cross-impedance simulator and the 28×28 MNIST handwritten digit dataset. The network architecture includes an input layer with 784 neurons, corresponding to 28×28 pixels of a digit image; a hidden layer with 300 neurons; and an output layer with 10 neurons, corresponding to digits 0–9. Figure 5 d). In this model, the photomodulated synaptic weights of the SrAl2O4–NaYF4 synapse are directly mapped to the synaptic weights between the input-hidden layer and the hidden-output layer, and these weights originate from the changes in luminescence intensity under pulsed light stimulation. The recognition accuracy of the SrAl2O4–NaYF4 synapse is as high as 91.12% ( Figure 5 e) indicates that photothermal synapses exhibit excellent recognition capabilities in neuromorphic computing tasks.

[0077] Example 4, refer to Figure 6 This is the fourth embodiment of the present invention. Unlike the previous embodiment, this embodiment provides a robotic arm gesture control system based on a composite artificial synapse, comprising: Synaptic sensing module: It consists of at least one composite artificial synapse. The number of composite artificial synapses corresponds one-to-one with the movable operating parts of the robotic arm. It is used to sense multimodal stimuli and generate long afterglow emission (LPL) signals. Multimodal excitation and detection module: including optical excitation unit, temperature stimulation unit and optical signal detection unit; The photoexcitation unit is used to excite the composite artificial synapses in the synaptic sensing module to generate long afterglow emission (LPL) signals; The temperature stimulation unit is used to provide adjustable temperature stimulation to regulate the release process of defect-trapped electrons in the composite artificial synapse; The optical signal detection unit is used to detect the long afterglow emission (LPL) signal generated by the composite artificial synapse and convert it into a processable signal; Signal processing module: Used to receive the signal output by the optical signal detection unit, distinguish between the thermal excitation state and the room temperature excitation state of the composite artificial synapse according to the difference in signal intensity, and output the corresponding state judgment result; Robotic arm drive module: Connects to the power transmission mechanism of the robotic arm. Based on the status judgment result of the signal processing module, it controls the bending or extending movements of the corresponding movable parts of the robotic arm to realize the preset gesture. Thermal stress feedback module: When the robotic arm comes into contact with an external object, if the temperature stimulation unit detects that the object temperature has reached the preset thermal threshold, the composite artificial synapse generates an enhanced long afterglow luminescence (LPL) signal. After receiving the enhanced signal, the signal processing module controls the robotic arm drive module to perform a retraction protection action.

[0078] Specifically, the synaptic sensing module consists of 5 composite artificial synapses, which correspond one-to-one with the thumb, index finger, middle finger, ring finger, and little finger of the robotic arm. The multimodal excitation and detection module includes a 365nm ultraviolet light excitation unit, a temperature control unit, and five photoelectric sensors; The 365nm ultraviolet excitation unit is used to simultaneously excite the five devices in the synaptic sensing module; the temperature control unit can provide temperature stimulation of 303K-503K with a temperature control accuracy of ±1K; the detection wavelength of the five photoelectric sensors is 522nm, and the detection sensitivity is 0.1-1μW / cm², which are used to detect the LPL intensity of the corresponding devices. The signal processing module is used to convert the LPL signal detected by the photoelectric sensor into an electrical signal, and to determine the excitation state of the corresponding device based on the intensity of the electrical signal. When the device is in an environment with a temperature of 373K or above, if its LPL intensity is ≥30% higher than that in the room temperature environment of 303K, it is determined to be in a thermal excitation state; otherwise, it is determined to be in a room temperature excitation state. Based on the judgment results of the signal processing module, the robotic arm drive module controls the bending or extension of the corresponding fingers, with the finger bending angle ranging from 0 to 90°. Thermal stress feedback module: When the robotic arm comes into contact with an external object, if the temperature control unit detects that the object temperature is ≥373K, it triggers the corresponding device to generate an enhanced LPL signal; after receiving the enhanced signal, the signal processing module controls the robotic arm drive module to perform a retraction action to protect the robotic arm.

[0079] Reference Figure 6 , Figure 6 This paper presents the hand gesture control and thermal simulation of a robotic arm based on photothermal dual-mode artificial synapses. (a) shows the control strategy of the robotic arm based on photothermal dual-mode artificial synapses. (b) shows the thermal stress response with a vision-tactile coordination and feedback system.

[0080] Inspired by the multimodal perception of biological nervous systems, a robotic arm gesture control system was constructed based on SrAl2O4–NaYF4 composite material artificial synapses, integrating visual and tactile information, and combining optical and thermal simulations. Figure 6 a) A photoelectric sensor detects the luminescence of the SrAl2O4–NaYF4 composite material in real time and transmits it to a control chip. This chip displays the photocurrent from the SrAl2O4–NaYF4 composite material in real time, and controls the bending and extension of the five fingers via gears based on the optical signals from the composite material. Considering the small spot size of the 980nm laser and the requirement for simultaneous stimulation of all five components, ultraviolet light is used as the excitation source.

[0081] Reference Figure 15 , Figure 15 Synaptic plasticity in the SrAl2O4-NaYF4 composite material simulated under ultraviolet light. (a) shows the luminescence response triggered by a single 365 nm pulse with different excitation powers (duration = 0.1 s). (b) shows the luminescence response under a single 365 nm pulse with different pulse durations (power = 330 mW). (c) shows the transition from STM to LTM induced by changing the frequency of the light pulse (duration = 0.1 s). (d) shows the luminescence intensity triggered by two consecutive ultraviolet pulses (duration = 0.1 s, interval = 0.1 s). (e) shows the dependence of the PPF exponent as a function of the interval between optical pulses (duration = 0.1 s). (f) shows the learning-forgetting-relearning process of the SrAl2O4-NaYF4 composite material (duration = 0.1 s, interval = 0.3 s).

[0082] The SrAl2O4–NaYF4 composite material achieved optical synaptic plasticity under 365 nm excitation, including short-range facilitation (PPF), short-range enhancement (STP), long-range enhancement (LTP), the transition from STP to LTP, and learning behavior. Figure 15 The composite material was charged under ultraviolet light for 10 seconds and then left to stand for 30 seconds. Samples attached to the thumb, ring finger, and little finger were then transferred to a high-temperature environment of 373 K, while other samples remained at room temperature. The high temperature promoted thermally activated luminescence within the trap, generating a stronger optical signal that controlled the gears and caused the thumb, ring finger, and little finger to bend, thus displaying the robotic arm's "V" sign.

[0083] Furthermore, a thermal stress response with a visual-tactile synapse coordination and feedback system was proposed based on SrAl2O4–NaYF4 composite artificial synapses. The thermal stress response refers to the visual pathway capturing an image through the retina and transmitting it to the visual cortex, prompting the brain to issue a motor command to grasp a cup. Simultaneously, when a finger touches a hot cup, tactile receptors detect the temperature increase and transmit signals to the central nervous system, triggering a protective response and causing the hand to rapidly retract. The optical simulation of the SrAl2O4–NaYF4 synapse is similar to visual object perception, guiding a robotic arm to grasp an object. When the robotic arm contacts the object, the object's high temperature triggers a stronger optical signal from thermally activated light emitted from a trap. This stronger optical signal is fed back to the chip, causing the robotic arm to retract in real time. These dual-mode synaptic characteristics provide a novel design strategy for multimodal sensory neuromorphic devices and show great potential in next-generation intelligent human-computer interaction systems.

[0084] In summary, a photothermal dual-mode artificial synapse was achieved based on the SrAl2O4–NaYF4 composite material. This was achieved by introducing Yb... 3+ To Tm 3+ The upconversion process and Eu 2+ A near-infrared triggered artificial synapse was achieved through 5d-state luminescent resonant energy transfer. This composite material exhibits excellent optical properties, including an average decay lifetime of 108 seconds and a PPF index as high as 213% under 980 nm excitation. Synaptic plasticity was successfully tunable, encompassing STP, LTP, STP-to-LTP transitions, and the learning-forgetting-relearning cycle. Simultaneously, tactile perception was realized using pyroelectric dynamics (TL) and decay dynamics generated by electrons released upon thermal activation in the trap. In a system-level demonstration, handwritten digit recognition was achieved with an accuracy of 91.12% by mapping synaptic weights to an artificial neural network (ANN). Furthermore, a robotic arm gesture control system based on SrAl2O4–NaYF4 composite artificial synapses and a thermal stress response system with visual-tactile coordination and feedback were constructed, integrating visual and tactile information and combining optical and thermal simulations. This provides a novel design concept for multimodal photothermal synapses and offers a path for advancing the frontier development of intelligent human-computer interaction systems with multimodal perception and adaptive feedback.

[0085] In summary, this invention constructs a mixture of NaYF4:Yb 3+ ,Tm 3+ To SrAl2O4:Eu 2+ ,Dy 3+Energy transfer was achieved, realizing a dual-mode optical synapse triggered by near-infrared (NIR). PPF, tunable short-term and long-term memory, and learning behavior were realized using NIR simulations, with the PPF exponent increased to 213%, the highest PPF achieved by a lanthanide-doped LPL-based optical synapse. Long-term plasticity was used to recognize handwritten digits with an accuracy of 91.12%. Furthermore, through SrAl2O4:Eu... 2+ ,Dy 3+ The release of LPL (simulated by thermal and optical pulses) controlled different gestures of the robotic arm. A visual-tactile bimodal synapse was also proposed to simulate the pressure response when grasping a hot object. These findings offer insights into the application of LPL-based multimodal synapses in neuromorphic technologies and human-computer interaction utilizing near-infrared and thermal stimulation.

[0086] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A composite artificial synapse, characterized in that: include, Core functional layer; The core functional layer is a composite material layer, which is composed of a mixture of SrAl2O4-based long afterglow material and NaYF4-based upconversion material. The SrAl2O4-based long afterglow material is doped with Eu. 2+ and Dy 3+ ; The NaYF4-based upconversion material is doped with Yb. 3+ and Tm 3+ ; Under near-infrared light excitation, the NaYF4-based upconversion material transfers energy to the SrAl2O4-based long-afterglow material through upconversion luminescence, enabling the SrAl2O4-based long-afterglow material to produce long-afterglow luminescence and realize synaptic function. Thermo-induced synaptic plasticity is achieved by regulating the release process of defect-trapped electrons in SrAl2O4-based long afterglow materials through thermal stimulation.

2. The composite artificial synapse as described in claim 1, characterized in that: Eu in the SrAl2O4-based long afterglow material 2+ The doping concentration is 6 at%, Dy 3+ The doping level is 4 at%; Yb in the NaYF4-based upconversion material 3+ The doping concentration is 20 at%, Tm 3+ The doping level is 1 at%.

3. The composite artificial synapse as described in claim 1 or 2, characterized in that: The mixing weight ratio of the SrAl2O4-based long afterglow material to the NaYF4-based upconversion material is 2:

1.

4. The composite artificial synapse as described in claim 1, characterized in that: The composite material layer is in the form of a thin film or a particle aggregate.

5. The composite artificial synapse as described in claim 1, characterized in that: NaYF4-based upconversion materials are synthesized using a hydrothermal method.

6. An information encryption system based on composite artificial synapses, characterized in that: include, Synaptic array module: an array composed of composite artificial synapses as described in any one of claims 1-5; Excitation module: can selectively output near-infrared light or ultraviolet light, the excitation module is used to excite the synaptic array module to produce long afterglow luminescence; Signal acquisition module: The signal acquisition module is a photoelectric sensor used to acquire the long afterglow emission intensity at different time points after the excitation module stops excitation, and to calculate the synaptic weights; Encoding conversion module: The encoding conversion module converts the synaptic weights obtained by the signal acquisition module into binary code, and then converts the binary code into character information according to the encoding rules; Time-varying control module: Used to control the time interval between the excitation module stopping excitation and the signal acquisition module acquiring the signal, so as to realize the encryption of time-varying information.

7. The information encryption system based on composite artificial synapses as described in claim 6, characterized in that: When the excitation wavelength output by the excitation module deviates, the encoding conversion module outputs an error character to achieve anti-interference encryption.

8. A handwritten digit recognition system based on composite artificial synapses, characterized in that: include, Synaptic weight mapping module: The change in long afterglow luminescence intensity of the composite artificial synapse as described in any one of claims 1-5 under pulsed light stimulation is used as the synaptic weight; Artificial neural network modules include an input layer, hidden layers, and an output layer; The number of neurons in the input layer is matched with the number of pixels in the handwritten digit image to be recognized, and is used to receive signals corresponding to the image pixels; The hidden layer contains at least one neuron and is configured with an activation function adapted to the training and recognition of artificial neural networks; The number of neurons in the output layer is matched to the number of categories of the handwritten digits to be recognized; Image input module: Used to import handwritten digit image data and preprocess the image data; Recognition Calculation Module: The forward propagation algorithm is used to calculate the output of the artificial neural network module. The network parameters are trained by combining the cross-entropy loss function and the gradient descent algorithm, and the recognition result of handwritten digits is output. Accuracy calibration module: used to calculate the accuracy of recognition results.

9. The handwritten digit recognition system based on composite artificial synapses as described in claim 8, characterized in that: The connection weights between the input layer and the hidden layer, and the connection weights between the hidden layer and the output layer of the artificial neural network module are all provided by the synaptic weight mapping module. The preprocessing includes removing image noise, unifying image grayscale features, and standardizing image pixel values ​​to meet the input requirements of the artificial neural network module.

10. A robotic arm gesture control system based on composite artificial synapses, characterized in that: include, Synaptic sensing module: composed of at least one composite artificial synapse as described in any one of claims 1-5, wherein the number of the composite artificial synapses corresponds one-to-one with the movable operating parts of the robotic arm, and is used to sense multimodal stimuli and generate long-persistence light emission signals; Multimodal excitation and detection module: including optical excitation unit, temperature stimulation unit and optical signal detection unit; The photoexcitation unit is used to excite the composite artificial synapse in the synaptic sensing module to generate a long-afterglow luminescence signal. The temperature stimulation unit is used to provide adjustable temperature stimulation to regulate the release process of defect trap electrons in the composite artificial synapse; The optical signal detection unit is used to detect the long-persistence light emission signal generated by the composite artificial synapse and convert it into a processable signal. Signal processing module: Used to receive the signal output by the optical signal detection unit, distinguish between the thermal excitation state and the room temperature excitation state of the composite artificial synapse according to the difference in signal intensity, and output the corresponding state judgment result; Robotic arm drive module: Connects to the power transmission mechanism of the robotic arm. Based on the status judgment result of the signal processing module, it controls the bending or extending movements of the corresponding movable parts of the robotic arm to realize the preset gesture. Thermal stress feedback module: When the robotic arm comes into contact with an external object, if the temperature stimulation unit detects that the object temperature has reached the preset thermal threshold, the composite artificial synapse generates an enhanced long afterglow luminescence signal. After receiving the enhanced signal, the signal processing module controls the robotic arm drive module to perform a retraction protection action.

Images