A photovoltaic memristor based on a diphenylalanine-based layered two-dimensional metal-organic framework and a preparation method and application thereof

Abstract

This invention discloses a photoelectric memristor based on a layered two-dimensional metal-organic framework (MOF) of diphenylalanine, its fabrication method, and its applications. The photoelectric memristor includes a substrate, two-dimensional layered MOF Ni(FF)₂H₂O nanosheets disposed on the substrate, and a first electrode and a second electrode respectively disposed at the top and bottom of the nanosheets. The device is fabricated using a modified hydrothermal method to obtain Ni(FF)₂H₂O crystals, followed by mechanical exfoliation to obtain nanosheets, and then platinum electrodes are deposited by magnetron sputtering to form a vertical sandwich structure. The device achieves multi-resistivity photoelectric storage and switching by controlling the formation and disruption of hydrogen bonds in the conductive channels within the nanosheets. This invention achieves a performance exceeding 645 Tbit / inch. 2 Ultra-high storage density, greater than 10 3 High switching ratio, exceeding 10 3 Its advantages, such as long-term stability and low power consumption, make it applicable to fields such as neuromorphic computing, artificial neural networks, and vehicle image enhancement and recognition.

Description

Technical Field

[0001] This invention belongs to the field of microelectronic device technology, specifically relating to a photoresistor based on a layered two-dimensional metal-organic framework of diphenylalanine, its fabrication method, and its application. Background Technology

[0002] In recent years, the rapid development of artificial intelligence has led to an exponential increase in the demand for data storage and computing, and traditional memory is gradually encountering physical limits in terms of storage density. Memristors, with their advantages of scalability, high density, low power consumption, high-speed multi-level storage, and compatibility with CMOS processes, are considered one of the key devices for breaking through the bottleneck of storage density.

[0003] To advance the size of memristor memory cells to the sub-nanometer scale and improve storage density, researchers have developed filamentary memristors, which achieve non-volatile storage through the generation or destruction of nanoscale conductive filaments in insulating channels. These devices utilize atomic-level defects themselves as storage sites, significantly increasing storage density. However, due to random bifurcation and lateral diffusion during filament generation, filamentary memristor arrays exhibit significant leakage current, leading to a series of bypass problems and a low on / off ratio, posing a major challenge to the integration of high-density resistive memory technology.

[0004] To retain the atomic-scale size advantage of conductive filaments in disordered media, while reducing leakage current and avoiding the randomness of filament formation, directly constructing crystal frameworks with regular channels has become a promising direction. The pore size, chemical environment, and spatial arrangement of metal-organic frameworks (MOFs) can be customized before synthesis, essentially achieving controllable construction of nanochannels. By selecting different metal centers and organic ligands and adjusting self-assembly conditions, the pore structure serving as conductive channels can be precisely controlled. Theoretically, if suitable organic ligands are selected to insulate adjacent pores of the nanochannels, it is hoped that ultra-high density storage capabilities can be provided while mitigating the impact of leakage current on device performance.

[0005] However, most existing metal-organic framework materials suffer from deliquescence and insufficient stability, making it difficult to meet the requirements of practical device applications. Furthermore, during crystal fabrication, crystallinity issues often affect the uniformity and rigidity of conductive channels, thus hindering the consistency of resistance switching and increasing the difficulty of large-scale integration. Fabricating bulk crystalline materials into layered two-dimensional forms is considered a potential direction for solving the crystallinity problem, but existing two-dimensional metal-organic framework materials still have shortcomings in terms of the controllability of conductive channels, photoelectric response characteristics, and long-term stability.

[0006] Therefore, for the growing demand for high-density, high-integration in-memory computing in the field of artificial intelligence, developing a two-dimensional metal-organic framework material with stable structure, controllable conductive channels, and good photoelectric response characteristics, and using it to construct high-performance memristors, remains an urgent technical problem to be solved in this field. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention provides a photoelectric memristor based on a layered two-dimensional metal-organic framework of diphenylalanine, its fabrication method, and its applications. Through photoelectric modulation of hydrogen bonds in the two-dimensional Ni(FF)₂H₂O metal-organic framework, a speed exceeding 645 Tbit / inch is achieved. 2 Ultra-high storage density, greater than 10 3 High switching ratio and over 10 3 It boasts long-term stability for seconds, along with advantages such as low power consumption, in-memory computing, and environmental sustainability.

[0008] This invention is achieved through the following technical solution:

[0009] A photoresistor based on a layered two-dimensional metal-organic framework containing diphenylalanine, comprising:

[0010] A substrate;

[0011] Two-dimensional layered metal-organic framework nanosheets disposed on the substrate, wherein the chemical formula of the two-dimensional layered metal-organic framework is Ni(FF)2H2O, and FF is diphenylalanine;

[0012] The first electrode is disposed on top of the two-dimensional layered metal-organic framework nanosheet;

[0013] The second electrode is disposed at the bottom of the two-dimensional layered metal-organic framework nanosheet;

[0014] The photoelectric memristor achieves multi-resistivity photoelectric storage and switching by controlling the formation and disruption of hydrogen bonds in the conductive channels within the two-dimensional layered metal-organic framework nanosheets.

[0015] Preferably, the two-dimensional layered metal-organic framework nanosheets have a layered structure stacked along the c-axis, with the layers bonded by van der Waals forces, and a single layer thickness of 1.32 nm; the nanosheets have a C2 polar point group and a P21 chiral space group.

[0016] Preferably, in the two-dimensional layered metal-organic framework, each Ni 2+ In the octahedral coordination geometry, two nitrogen atoms and four oxygen atoms from three different diphenylalanine ligands, along with one water molecule, are coordinated; the framework forms a one-dimensional [Ni(FF)2H2O] extending along the a-axis. n Chains, which are linked in the ab plane by diphenylalanine molecules.

[0017] Preferably, both the first electrode and the second electrode are platinum electrodes, forming a vertical sandwich structure Pt / Ni(FF)2H2O / Pt memristor.

[0018] Preferably, the storage density of the photomemristor is greater than 645 Tbit / inch. 2 High-to-low impedance switching ratio greater than 10 3 The resistive state retention time exceeds 10 3 s; Write operation voltage is 10 V, reset voltage is -15 V, write power is 10 3 A·cm 2 .

[0019] The above-mentioned method for fabricating a photomemristor based on a layered two-dimensional metal-organic framework of diphenylalanine includes the following steps:

[0020] Step 1) Prepare metal-organic framework Ni(FF)2H2O crystals using a modified hydrothermal method;

[0021] Step 2) Ni(FF)2H2O nanosheets were separated from the Ni(FF)2H2O crystals obtained in Step 1) using a mechanical exfoliation method;

[0022] Step 3) Transfer the Ni(FF)2H2O nanosheets obtained in Step 2) onto the surface of a silicon substrate;

[0023] Step 4) Platinum electrodes are deposited on the top and bottom of the Ni(FF)2H2O nanosheets treated in Step 3) by magnetron sputtering to form a vertical sandwich structure Pt / Ni(FF)2H2O / Pt photomemristor.

[0024] Preferably, the modified hydrothermal method in step 1) is as follows: Ni(NO3)2 solution is mixed with diphenylalanine solution, pH is adjusted to 4.3, NaOH solution is added and then transferred to a high-pressure reactor, heated to 65°C at 2°C / min and reacted at a constant temperature for 24 h, and then cooled to room temperature at 0.5°C / min.

[0025] Preferably, the thickness of the Ni(FF)2H2O nanosheets in step 2) is 4.01 nm to 10 μm.

[0026] The above-mentioned photoelectro-memristors based on diphenylalanine layered two-dimensional metal-organic frameworks are applied in neuromorphic computing, artificial neural networks, image enhancement and recognition, or vehicle-mounted intelligent vision systems.

[0027] The above-mentioned preparation method yields a layered two-dimensional metal-organic framework photoelectric memristor with diphenylalanine, which can be used in neuromorphic computing, artificial neural networks, image enhancement and recognition, or vehicle-mounted intelligent vision systems.

[0028] The beneficial effects of this invention are as follows:

[0029] (1) This invention achieves photo-stable switching of resistive states by precisely controlling the formation and destruction of hydrogen bonds in the conductive channels of two-dimensional Ni(FF)2H2O nanosheets, with a storage density exceeding 645 Tbit / inch. 2 It is far superior to traditional wire-type memristors (such as Pt / TaO). x Materials such as Ta, Cu / SiO2 / Pt have a storage density of approximately 50 Tbit / inch. 2 The ultra-high storage density allows a single storage unit to store multiple bits of information, meeting the efficient storage needs of neural network models for large amounts of weights and biases, reducing the number of storage units and lowering storage costs.

[0030] (2) The operating voltage of the photomemristor of the present invention is approximately 10 V, and the write power is approximately 10 V. 3 A·cm 2 This device exhibits low power consumption. Furthermore, it enables an in-memory computing architecture, performing data storage and computation in the same physical location. This avoids the latency and energy waste caused by data movement in traditional von Neumann architectures, significantly improving the efficiency of neural network computation. In addition, the Ni(FF)₂H₂O nanosheets possess excellent interfacial properties, which help improve the response speed and data throughput of the memristor.

[0031] (3) Experimental results show that the photomemristor of the present invention can achieve a speed of over 10 3 Maintains a stable resistance state within seconds with minimal fluctuations; the high-to-low resistance switching ratio is greater than 10. 3 This ensures thorough erasure and accurate reading. Compared to traditional wire-type memristors (with an on / off ratio of approximately 10), this is significantly improved. 2 This invention features a higher on / off ratio, ensuring the integrity and reliability of the data storage system. Using platinum as the source and drain material, it forms a good ohmic contact with Ni(FF)₂H₂O nanosheets, reducing contact resistance and further decreasing energy consumption.

[0032] (4) The conductive channels within the Ni(FF)2H2O nanosheets of this invention are uniformly distributed, resulting in consistent resistive switching effects at the nanoscale. This overcomes the performance inconsistency caused by the random distribution of conductive filaments in traditional filament-type memristors, providing a solid foundation for large-scale integration and expansion. By increasing the number and array density of nanosheets, the storage capacity can be easily expanded to meet the ever-growing data storage demands.

[0033] (5) In addition to data storage, the photoelectromemristor of this invention shows potential applications in neuromorphic computing and artificial neural networks. It can simulate the weight adjustment process of biological synapses, providing a new hardware foundation for brain-like computing and intelligent systems. When combined with convolutional neural networks, it can be used to develop ultra-high precision artificial retinas with facial and color recognition capabilities. Furthermore, this device also has broad application prospects in the fields of sensors and programmable logic circuits.

[0034] (6) The core material Ni(FF)₂H₂O crystal in the photoelectric memristor of this invention is synthesized using a modified hydrothermal method. Because the diphenylalanine molecule itself has a hydrophobic benzene ring side chain, and the Ni(FF)₂H₂O crystal contains coordinated water of crystallization, it is stable under normal temperature and pressure. The modified hydrothermal method reduces the emission of harmful substances and energy consumption, conforming to the concepts of green chemistry and sustainable development. This material has good stability and recyclability, which helps reduce electronic waste and promotes the development of a circular economy.

[0035] (7) The present invention successfully prepared a high-performance Pt / Ni(FF)2H2O / Pt photomemristor, realizing the photoinduced multi-resistance state storage function based on a two-dimensional metal-organic framework. This changed the traditional view that multi-resistance state storage technology is unstable, unreliable and difficult to apply in the existing technology, and provided a new idea for high-density data storage technology. Attached Figure Description

[0036] Figure 1 Here is a schematic diagram and crystallographic characterization of the Ni(FF)₂H₂O metal-organic framework in Example 2: a represents [Ni(FF)₂H₂O] n A schematic diagram of the three-dimensional structure formed by the stacking of chains along the c-axis through inter-chain hydrogen bonds; b is a comparison of experimental and simulated X-ray diffraction patterns of Ni(FF)2H2O nanosheets; c is a comparison of the lattice fringes of the (hk0) crystal plane, the atomic structure of synchrotron X-ray diffraction, and the electron diffraction pattern along the c-axis.

[0037] Figure 2 The following are the electrical performance test graphs of the Pt / Ni(FF)2H2O / Pt photomemristor in Example 3: a) is a typical bipolar IV curve; b) is the IV characteristic curve after 50 cycles; c) is the IV characteristic curve after 10 cycles. 3 High and low resistance state retention characteristics within seconds; d is the high and low resistance state retention performance during 300 consecutive cyclic scans; e is the complex impedance diagram at 0.1 V, 0.3 V, and 0.5 V; f is the relationship between resistance and magnetic field in the Hall effect;

[0038] Figure 3The following are the photoelectric response characteristics of the Pt / Ni(FF)2H2O / Pt photomemristor in Example 3: a) Bipolar IV curves at wavelengths of 447 nm, 532 nm, and 632 nm; b) Resistance retention characteristics under visible light pulses of different wavelengths; c) Photoelectric response curves of -10 V pulse voltage and 447 nm light pulse; d) Effect of different light intensities on resistance retention characteristics under 447 nm light pulse; e) Retention characteristics from low to high resistance state under 447 nm light pulse; f) Arrhenius plot of the proton migration barrier.

[0039] Figure 4 The following are performance diagrams for image recognition and autonomous training of the algorithm-enhanced camera based on the photomemristor of this invention in a tunnel scene in Example 4: a and d are images of the tunnel interior captured by a commercial surveillance camera outside the tunnel; b and e are images of the dimly lit tunnel interior captured by the algorithm-enhanced camera outside the tunnel, and the recognition and distance judgment of vehicles; c is the algorithm-enhanced camera capturing environmental information while performing algorithm training; f is over 10 6 After training for 1 second, the system performs visual enhancement recognition and distance judgment on cars in a dimly lit tunnel; g represents the difference in light sensitivity between the ordinary camera and the algorithm-enhanced camera, and the recovery of recognition accuracy after autonomous training; h represents the change in distance judgment accuracy of the algorithm-enhanced camera and its autonomous recovery; i represents the image recognition correctness, accuracy, power consumption, recall rate, and overall performance after training, autonomous operation, and retraining. Detailed Implementation

[0040] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0041] Unless otherwise specified, the technical means used in the following embodiments are all conventional means well known to those skilled in the art, and the experimental methods without specific conditions are all conventional methods in the art.

[0042] Unless otherwise specified, all materials and reagents used in the following examples are commercially available.

[0043] Example 1: Synthesis of Ni(FF)₂H₂O metal-organic framework crystals

[0044] This embodiment uses a modified hydrothermal method to prepare Ni(FF)₂H₂O metal-organic framework crystals. The specific steps are as follows:

[0045] In a clean centrifuge tube, 21 μL of Ni(NO3)2 solution was added to 5 mL of deionized water, and the mixture was stirred until homogeneous. The pH of the solution was measured to be 4.3. Then, 40 mL of freshly prepared diphenylalanine (FF) solution was added to the mixture, and the mixture was stirred. Next, 15 μL of a 1 mol / L NaOH aqueous solution was added, and the mixture was transferred to an autoclave. The autoclave was heated to 65 °C at a heating rate of 2 °C / min and reacted at this temperature for 24 h. After the reaction was complete, the solution was cooled to room temperature at a rate of 0.5 °C / min. Green crystals up to 10 μm in size were observed in the lower part of the autoclave reaction vessel. The crystals were filtered from the reaction mixture using a filter bag and washed with anhydrous ethanol. Finally, the product was dried in a vacuum oven to obtain Ni(FF)2H2O metal-organic framework microcrystals.

[0046] Even slight changes in solution concentration, polarity, pH value, or temperature can lead to a decrease in crystal yield or the formation of entirely new phases, thereby affecting the crystallinity of the produced Ni(FF)2H2O. The on / off ratio and stability of the memristor are also affected as a result.

[0047] Example 2: Preparation and Characterization of Ni(FF)₂H₂O Nanosheets

[0048] 1. Preparation of Ni(FF)₂H₂O nanosheets

[0049] Ni(FF)₂H₂O nanosheets were isolated from the Ni(FF)₂H₂O crystals prepared in Example 1 using a standard mechanical exfoliation method. Details are as follows:

[0050] First, use a clean scalpel to gently cut several tiny grooves with a spacing of 100~200 μm on the surface of the crystal. Then, place the crystal on a special blue film adhesive tape for mechanical peeling, fold the tape in half so that the block is wrapped by the top and bottom layers of tape, and then slowly tear the tape to complete one peeling. Repeat the folding and tearing operation 5~8 times to peel the block off layer by layer on the tape, forming uniformly distributed Ni(FF)2H2O nanosheets.

[0051] Cut the blue film with the nanosheets attached to it into a suitable size and gently attach it to the pre-cleaned SiO2 / Si substrate. Use a lint-free cotton swab to gently wipe the back of the tape to ensure full contact between the tape and the substrate and remove air bubbles. Let it stand for 2-3 minutes to allow a stable van der Waals contact to form between the material and the substrate. Then, slowly and evenly peel the tape off the substrate to transfer the nanosheets on the tape to the surface of the silicon substrate.

[0052] 2. Characterization of Ni(FF)₂H₂O nanosheets

[0053] The obtained nanosheets were structurally characterized, and the results are as follows: Figure 1As shown, the details are as follows:

[0054] like Figure 1 As shown in Figure a, Ni(FF)₂H₂O forms an extended one-dimensional [Ni(FF)₂H₂O] along the a-axis. n These chains are linked by diphenylalanine molecules in an ab plane, which is approximately 1.32 nm thick. The planes are stacked along the c-axis by van der Waals forces to form a three-dimensional crystal.

[0055] Single-crystal X-ray diffraction (SC-XRD) such as Figure 1 As shown in Figure b, Ni(FF)₂H₂O crystals crystallize in a monoclinic system, possessing a polar point group C₂ and a chiral space group P₂₁. In the octahedral coordination geometry, each Ni ion (Ni 2+ The resistance switching phenomenon in Ni(FF)2H2O is clearly regulated by guest water molecules within the conductive channel. The hydrogen bond network constructed by these water molecules is affected differently by light irradiation of different wavelengths, thus achieving effective control of the resistance state.

[0056] The experimental X-ray diffraction pattern of the three-layer Ni(FF)2H2O nanosheets showed no significant difference from the simulated crystal pattern. Figure 1 (b) indicates that the nanosheets retain the original properties of diphenylalanine and have excellent structural stability and integrity.

[0057] High-resolution transmission electron microscopy characterization of the three-layer Ni(FF)₂H₂O nanosheets, as shown in... Figure 1 As shown in Figure c, the lattice fringes of the (hk0) crystal plane are clearly displayed over a large area, indicating its excellent robustness; its atomic structure is similar to that of bulk crystals, demonstrating the structural integrity and crystallinity of the nanosheets; the electron diffraction pattern of the selected region matches well with the simulation pattern, confirming the integrity and effectiveness of the atomic structure of the nanosheets.

[0058] Example 3: Fabrication and Performance Testing of a Pt / Ni(FF)2H2O / Pt Vertical Sandwich Structure Photomemristor

[0059] 1. Fabrication of photomemristors

[0060] The Ni(FF)₂H₂O nanosheets transferred to the silicon substrate in Example 2 were used as the core storage medium. Platinum metal was deposited at the bottom and top of the nanosheets using magnetron sputtering as the bottom and top electrodes, respectively, forming a vertical sandwich-structured Pt / Ni(FF)₂H₂O / Pt photomemristor. Details are as follows:

[0061] Ni(FF)₂H₂O nanosheets were used as the core storage medium. Photolithography was used to fabricate the photoresist on the sample surface into the shape required for the electrodes. The sample was then placed in a magnetron sputtering vacuum chamber, which was evacuated to a high base vacuum and then purged with high-purity inert argon gas for 10 hours. -1 ~10 -2 The working gas pressure is Pa. A high-voltage electric field is applied between the platinum target and the substrate to cause Ar atoms in the cavity to undergo collisional ionization, forming Ar atoms. + Ions bombard the target surface under the strong acceleration of the cathode electric field, ejecting target atoms from the target surface. The sputtered target atoms are transported to the substrate surface in a vacuum environment, where they nucleate and grow to eventually form a continuous thin film.

[0062] The core material thickness of the fabricated device can be as low as 4.01 nm.

[0063] 2. Electrical performance testing

[0064] The electrical performance of the prepared Pt / Ni(FF)₂H₂O / Pt photoresistors was tested using a Keithley 2400 and Keithley 6517B source meter with external device electrodes. The scan voltage range was -20 to 20 V, the scan step size was 200 mV, and the current limit was 105 mA. The electrical performance of 50 prepared Pt / Ni(FF)₂H₂O / Pt photoresistors was tested. All devices exhibited typical memristor effect and photoresistive switching phenomenon, and the yield rate of devices within the same voltage range was above 90%. The results are as follows... Figure 2 As shown, the details are as follows:

[0065] like Figure 2 As shown in Figure a, due to the hydrogen bond network formed by water molecules in the nanopores facilitating proton transport, the Pt / Ni(FF)2H2O / Pt photomemristor (approximately 1 μm thick) exhibits a typical bipolar IV curve, where the set voltage V0 is... set The reset voltage is 10V. reset The voltage is -15 V, and the high-to-low resistance current ratio is approximately 10. 3 The polarization power required for a write operation is approximately 10. 2 W·cm 2 It demonstrates excellent energy efficiency advantages.

[0066] like Figure 2 As shown in Figure b, the IV characteristic curves of the Pt / Ni(FF)2H2O / Pt photomemristor maintain good consistency after 50 cycles.

[0067] like Figure 2 As shown in c, in a length of 10 3 Within a test period of seconds, the high-resistivity resistor (R)HRS ) and low-resistance state resistance (R LRS The fluctuations are negligible, indicating that the Pt / Ni(FF)2H2O / Pt opto-memristor has excellent long-term stability.

[0068] like Figure 2 As shown in Figure d, the Pt / Ni(FF)2H2O / Pt photomemristor maintained a stability greater than 10 during 300 consecutive read / write cycles. 3 The high switching ratio further confirms its excellent cycle durability.

[0069] These superior properties highlight the advantages of two-dimensional Ni(FF)2H2O materials, including their bipolar switching capability, low energy consumption characteristics, and application potential for long-term enhancement and suppression functions, thus making them stand out from traditional resistive random access memory technology and demonstrating unique competitiveness.

[0070] Figure 2 The image shows the complex impedance spectra of the Pt / Ni(FF)2H2O / Pt photomemristor at 0.1 V, 0.3 V, and 0.5 V (the real part Z' of the total impedance Z represents the resistive part, and the imaginary part Z'' represents the reactive part). The curves showing the relationship between the real part Z' and the imaginary part Z'' are parabolic, indicating that the two-dimensional Ni(FF)2H2O metal-organic framework has nonlinear electrical characteristics.

[0071] Hall effect measurement results as follows Figure 2 As shown in Figure f, the carrier concentration in the conductive channel of the Pt / Ni(FF)2H2O / Pt photomemristor is as high as approximately 4.77 × 10⁻⁶. 15 (Ω·C) -3 .

[0072] 3. Photoresponse performance test

[0073] The resistance switching characteristics of the prepared Pt / Ni(FF)₂H₂O / Pt photomemristor under visible light irradiation at different wavelengths were tested. Semiconductor lasers with wavelengths of 447 nm, 532 nm, and 632 nm from the MDL-III series produced by Changchun New Industries Optoelectronic Technology Co., Ltd. were used. These lasers all have an output power of 1000 mW and a circular Gaussian spot size of 1–3 mm. The results are as follows: Figure 3 As shown, the details are as follows:

[0074] like Figure 3 As shown in Figure a, the Pt / Ni(FF)₂H₂O / Pt photomemristor exhibits bipolar IV characteristics under visible light at different wavelengths of 447 nm, 532 nm, and 632 nm. Compared to 632 nm light, the V under 447 nm visible light is significantly higher. setThe voltage is lower (about 10 V) because the charge carriers of the Pt / Ni(FF)2H2O / Pt photomemristor are protons, while the photogenerated charge carriers are usually electrons. Therefore, the turn-on voltage is higher when irradiated with higher-energy, shorter-wavelength visible light.

[0075] like Figure 3 As shown in Figure b, under illumination by continuous monochromatic light sources at different wavelengths of 447 nm, 532 nm, and 632 nm, the optical power density is approximately 1 W / mm². 2 After applying a reset voltage of -10 V, the laser was re-applied and illuminated perpendicularly onto the device for 0.5 s. The Pt / Ni(FF)2H2O / Pt opto-memristor exhibited good resistance uniformity, demonstrating stability and retention during multi-resistance state switching.

[0076] like Figure 3 As shown in Figure c, after applying a -10 V AC current under a 447 nm optical pulse, the resistance change characteristics exhibit good signal carrying capacity, which means that the Pt / Ni(FF)2H2O / Pt opto-memristor has good application potential in the field of information technology.

[0077] like Figure 3 As shown in Figure d, under a 447 nm light pulse, the light intensity increases from 0 to 0.9 mW / cm². 2 (A total of 64 sets of data) When the light intensity is stronger under the condition of constant wavelength, the resistance of Pt / Ni(FF)2H2O / Pt photomemristor is greater.

[0078] like Figure 3 As shown in Figure e, both the low-resistivity state and the high-resistivity state exhibit good resistance time retention characteristics under a 447 nm light pulse.

[0079] The above experimental results demonstrate that the Pt / Ni(FF)2H2O / Pt photomemristor exhibits good resistance uniformity under illumination of constant wavelength and intensity.

[0080] Figure 3 f is the Arrhenius diagram of the proton migration barrier under different resistance states, indicating that the energy required for proton migration is higher and the proton migration barrier is higher in the low resistance state.

[0081] The Pt / Ni(FF)₂H₂O / Pt vertical sandwich structure of this embodiment reduces the core material thickness to 4.01 nm, a new low among all conductive polymer materials. By alternating the use of Ni(FF)₂H₂O nanochannels as resistive random access memory (RAM) cells, each conductive channel in the core storage region nanosheet of the opto-memristor in this embodiment can be independently addressed, providing a nanopore size of 0.8 nm × 0.8 nm for the resistive switching unit. Combined with the device's multi-resistive-state characteristics (each memory cell can store more than 3 bits of information), its theoretical storage density can reach 645 Tbit / inch. 2 This storage density is far higher than that of traditional filament memristors (such as Pt / TaO). x Materials such as Ta, Cu / SiO2 / Pt have a storage density of approximately 50 Tbit / inch. 2 ).

[0082] Example 4: Simulated Application in Vehicle Image Enhancement and Recognition

[0083] This invention, based on the fabrication of a multi-resistivity photoelectric memory based on Ni(FF)₂H₂O nanosheets, focuses on the application of multi-resistivity storage to meet the high-speed data storage and computing demands brought about by the development of artificial intelligence. Addressing the pain point of current commercial automotive image sensors in image recognition applications where they cannot predict traffic flow conditions in dimly lit tunnels, this invention proposes a method for fabricating a photoelectric memristor based on a layered diphenylalanine metal-organic framework. Simulated intelligent automotive cameras based on this photoelectric memristor help to perform intelligent image recognition within the short adaptation time of the human eye to drastic changes in light intensity, reducing the occurrence of accidents. Traditional filamentary memristors typically suffer from problems such as large leakage current, small on / off ratio, and poor photoelectric control capability. This is because the filaments of filamentary memristors, which utilize ion or defect migration, have random branching and lateral diffusion characteristics, resulting in leakage current that makes it difficult to fabricate usable integrated devices. Specifically:

[0084] The Pt / Ni(FF)₂H₂O / Pt photomemristor prepared in Example 3 was applied to an image enhancement and recognition system simulating an in-vehicle scanning camera. A convolutional neural network was employed, and its performance was evaluated using DeepMind's DM control physics simulation and reinforcement learning system.

[0085] To achieve this goal, an in-situ computing hardware system integrating sensor-based computation was constructed based on an artificial neural network architecture. This neural network performs internal computation through a 4×6 cross-dot array, which also serves as the light input. By applying light of different intensities and wavelengths to different array nodes, the resistivity of individual array nodes is controlled. Based on Kirchhoff's current law, after reading the total current, the neural network algorithm analyzes and solves the weights of the current at each node, thereby obtaining the illumination information (wavelength and intensity) at each array node, completing the input and storage of light information. Based on the obtained light information, the photosensitivity is adjusted to automatically brighten images of dimly lit tunnels and reduce exposure in brightly lit tunnels, resulting in clear tunnel photographs. Furthermore, a convolutional neural network algorithm is used for intelligent object recognition and distance assessment of obstacles and vehicles in the images.

[0086] like Figure 4 As shown in a and d, ordinary commercial cameras cannot effectively identify dimly lit scenes inside tunnels from outside; while algorithm-enhanced cameras using Pt / Ni(FF)2H2O / Pt photomemristors can achieve clear identification under the same conditions. Figure 4 (b, e).

[0087] like Figure 4 As shown in Figure c, the algorithm-enhanced camera can perform autonomous algorithm training based on the obtained images during the scanning and recognition process, demonstrating a strong autonomous correction capability.

[0088] like Figure 4 As shown in Figure f, this algorithm enhances the camera's ability to operate continuously for over 10 hours. 6 It takes seconds and maintains good recognition accuracy.

[0089] like Figure 4 As shown in g and h, after a slight decrease in scanning and recognition accuracy with the increase of the training cycle, the algorithm-enhanced camera can regain 100% recognition accuracy after autonomous training, and the distance judgment error also decreases and recovers, which fully demonstrates the stability and convergence of the model.

[0090] like Figure 4 As shown in Figure i, after training, autonomous operation, and retraining, the algorithm enhances the camera's image recognition accuracy, precision, power consumption, recall rate, and overall performance, maintaining high accuracy while consuming low power.

[0091] The above experimental results fully demonstrate the potential of the device of the present invention in intelligent image recognition systems, especially in fields that require high-density storage and intelligent image enhancement recognition.

[0092] The embodiments described above are only some, not all, of the embodiments of the present invention. The detailed description of the embodiments of the present invention is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments. The scope of protection of the present invention is determined by the scope claimed in the claims. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

Claims

1. A photoresistor based on a layered two-dimensional metal-organic framework containing diphenylalanine, characterized in that, include: A substrate; Two-dimensional layered metal-organic framework nanosheets disposed on the substrate, wherein the chemical formula of the two-dimensional layered metal-organic framework is Ni(FF)2H2O, and FF is diphenylalanine; The first electrode is disposed on top of the two-dimensional layered metal-organic framework nanosheet; The second electrode is disposed at the bottom of the two-dimensional layered metal-organic framework nanosheet; The photoelectric memristor achieves multi-resistivity photoelectric storage and switching by controlling the formation and disruption of hydrogen bonds in the conductive channels within the two-dimensional layered metal-organic framework nanosheets.

2. The photoresistor based on a layered two-dimensional metal-organic framework using diphenylalanine according to claim 1, characterized in that, The two-dimensional layered metal-organic framework nanosheets have a layered structure stacked along the c-axis, with the layers bonded by van der Waals forces, and a single layer thickness of 1.32 nm; the nanosheets have a C2 polar point group and a P21 chiral space group.

3. The photoresistor based on a layered two-dimensional metal-organic framework using diphenylalanine according to claim 1, characterized in that, In the two-dimensional layered metal-organic framework, each Ni 2+ In the octahedral coordination geometry, two nitrogen atoms and four oxygen atoms from three different diphenylalanine ligands, along with one water molecule, are coordinated; the framework forms a one-dimensional [Ni(FF)2H2O] extending along the a-axis. n Chains, which are linked in the ab plane by diphenylalanine molecules.

4. The photoresistor based on a layered two-dimensional metal-organic framework using diphenylalanine according to claim 1, characterized in that, Both the first electrode and the second electrode are platinum electrodes, forming a vertical sandwich structure Pt / Ni(FF)2H2O / Pt memristor.

5. The photoresistor based on a layered two-dimensional metal-organic framework using diphenylalanine according to claim 1, characterized in that, The storage density of the photomemristor is greater than 645 Tbit / inch. 2 High-to-low impedance switching ratio greater than 10 3 The resistive state retention time exceeds 10 3 s; Write operation voltage is 10 V, reset voltage is -15 V, write power is 10 3 A·cm 2 .

6. A method for fabricating a photoresistor based on a layered two-dimensional metal-organic framework of diphenylalanine as described in any one of claims 1-5, characterized in that, Includes the following steps: Step 1) Prepare metal-organic framework Ni(FF)2H2O crystals using a modified hydrothermal method; Step 2) Ni(FF)2H2O nanosheets were separated from the Ni(FF)2H2O crystals obtained in Step 1) using a mechanical exfoliation method; Step 3) Transfer the Ni(FF)2H2O nanosheets obtained in Step 2) onto the surface of a silicon substrate; Step 4) Platinum electrodes are deposited on the top and bottom of the Ni(FF)2H2O nanosheets treated in Step 3) by magnetron sputtering to form a vertical sandwich structure Pt / Ni(FF)2H2O / Pt photomemristor.

7. The preparation method according to claim 6, characterized in that, Step 1) The modified hydrothermal method is as follows: Ni(NO3)2 solution and diphenylalanine solution are mixed, pH is adjusted to 4.3, NaOH solution is added and then transferred to a high-pressure reactor. The mixture is heated to 65℃ at 2℃ / min and reacted at a constant temperature for 24 h. Then it is cooled to room temperature at 0.5℃ / min.

8. The preparation method according to claim 6, characterized in that, Step 2) The thickness of the Ni(FF)2H2O nanosheets is 4.01 nm to 10 μm.

9. The application of the photoelectro-memristor based on a layered two-dimensional metal-organic framework of diphenylalanine as described in any one of claims 1-5 in neuromorphic computing, artificial neural networks, image enhancement and recognition, or vehicle-mounted intelligent vision systems.

10. The application of the photoelectro-memristor based on a layered two-dimensional metal-organic framework of diphenylalanine prepared by any one of the preparation methods described in claims 6-8 in neuromorphic computing, artificial neural networks, image enhancement and recognition, or vehicle-mounted intelligent vision systems.

Images