Bionic neuron memristor and preparation method thereof

Abstract

The application provides a kind of biomimetic neuron memristor and preparation method thereof, it is related to biomimetic neuron technical field, including substrate, parallel electrode and dielectric layer;The parallel electrode and the dielectric layer are all arranged on the substrate;The parallel electrode includes positive electrode layer and negative electrode layer, and the dielectric layer is arranged between the positive electrode layer and the negative electrode layer;The dielectric layer uses semiconductor material containing low activation energy ions in lattice;There are two kinds of low activation energy ions in the lattice in the dielectric layer, and under the driving of electric field, two kinds of low activation energy ions are regulated by electric field, and move to positive and negative two poles respectively, for simulating the membrane potential change process generated by the change of sodium and potassium ion transport in and out of membrane when neuron is stimulated.The application can realize the simulation of neuron membrane potential change, neuron cumulative emission phenomenon and neuron refractory period behavior without building peripheral circuit.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of biomimetic neurons, in particular to a biomimetic neuron memristor and a preparation method thereof. BACKGROUND

[0002] At present, the research on biomimetic neurons is still in its infancy, and the simulation of the function of neurons is also very mechanical. More peripheral circuits need to be built to simulate the function. In 2018, Huang et al. used a device with a W / WO3 / poly(3,4-ethylenedioxythiophene):polystyrenesulfonate / Pt structure to simulate the change of the membrane potential of neurons and the cumulative emission function of neurons by proton migration in the device, which has a great significance for neuromorphic computing. However, in this work, the function needs to be realized with the aid of peripheral circuits, and the cost is relatively large while realizing the function. However, the electrical pulse signal emission of biological neurons is realized by the joint action of sodium ions and potassium ions, and based on this, the neurons also have advanced behaviors such as cumulative emission and refractory period, so it is very important to study the migration rule of multiple ions in the material.

[0003] For the existing robot system, although it can complete part of the work like a person, it is still a product of computer system, and there is an essential difference from the human system. Because of this, the robot does not have a perfect emotional system, and the reaction is not as sensitive as the biological body. Human beings process information through neurons and synapses, and a large number of synapses and neurons work together to complete various thinking and instructions. Therefore, simulating neurons for use in robot systems is of great significance to the development of robot systems. SUMMARY

[0004] The problem solved by the present application is how to realize the simulation of the change of the membrane potential of neurons, the cumulative emission phenomenon of neurons and the refractory behavior of neurons without building peripheral circuits.

[0005] To solve the above problems, the present application provides a biomimetic neuron memristor, which comprises a substrate, parallel electrodes and a dielectric layer; the parallel electrodes and the dielectric layer are arranged on the substrate;

[0006] The parallel electrodes comprise a positive electrode layer and a negative electrode layer, and the dielectric layer is arranged between the positive electrode layer and the negative electrode layer;

[0007] The dielectric layer adopts a semiconductor material containing low-activation-energy ions in a lattice;

[0008] The two low-activation-energy ions in the lattice of the medium layer are regulated by the electric field and move to the positive and negative poles respectively under the drive of the electric field, for simulating the membrane potential change process of the sodium and potassium ion transport in and out of the membrane of the neuron when the neuron is stimulated.

[0009] In the above-mentioned biomimetic neuron memristor, the biomimetic neuron memristor takes electrical excitation as the input signal and can realize the related functions of the neuron without applying a peripheral circuit. By adopting a parallel electrode planar structure, selecting high-conductivity inert metals as the positive electrode layer and the negative electrode layer, and selecting a material containing two low-activation-energy ions in the lattice as the medium layer, the migration of the two low-activation-energy ions in the material is caused by the excitation of the applied voltage, a built-in electric field is generated, and the biomimetic neuron memristor exhibits a voltage opposite to the direction of the applied voltage after the voltage is removed, thereby simulating the membrane potential change process of the neuron. The biomimetic neuron memristor is applied to a neuron filter, brain-like computing, a neural network, and binary conversion.

[0010] Further, under the action of the electric field, the current gradually increases from a negative value to a positive value as the voltage increases, and gradually decreases from a positive value to a negative value as the voltage decreases, for simulating the change of the membrane potential of the neuron when the neuron is stimulated.

[0011] In the above-mentioned biomimetic neuron memristor, after the voltage is applied, the biomimetic neuron memristor generates a sudden current, and the positive and negative ions with small activation energy accumulate on the two sides of the material, so that the same size of voltage stimulation is continuously applied, and it is found that the second applied voltage cannot generate a sudden current, for simulating the absolute refractory period behavior of the neuron. A new biomimetic neuron memristor is used, a threshold size of voltage stimulation is applied for the first time, the biomimetic neuron memristor generates a sudden current, and a voltage stimulation higher than the first size is applied for the second time, the biomimetic neuron memristor still generates a sudden current, for simulating the relative refractory period behavior of the neuron.

[0012] Further, the material of the positive electrode layer and the negative electrode layer adopts inert metal oxide; the inert metal oxide is one or both of Au and Pt.

[0013] In the above-mentioned biomimetic neuron memristor, the positive electrode layer and the negative electrode layer are generated by ultraviolet lithography and electron beam evaporation.

[0014] Further, the medium layer adopts a pure inorganic perovskite material, and the semiconductor perovskite material containing two low-activation-energy ions in the lattice is one of Cs2AgBiCl6, Cs2AgBiBr6, and Cs2AgBiI6.

[0015] In the biomimetic neuron memristor, the preparation method of the medium layer can adopt the method of coating, spin coating, adding anti-solvent, and then evaporating the solvent to obtain a coating layer, or can adopt the method of evaporation to prepare the perovskite medium layer.

[0016] Further, the thickness of the medium layer is 400-500nm.

[0017] Further, an external voltage is applied between the positive electrode layer and the negative electrode layer, and the size and direction of the voltage are adjusted to simulate the change of the membrane potential inside and outside the cell membrane.

[0018] Further, the positive electrode layer, the negative electrode layer and the medium layer are all in direct contact with the substrate, and the material of the substrate adopts a silicon wafer.

[0019] A preparation method of a biomimetic neuron memristor comprises the following steps:

[0020] S1: configuring a precursor solution, dissolving CsBr, BiBr3 and AgBr in dimethyl sulfoxide at a ratio of 2:1:1, heating and stirring at 75 DEG C for 4h, and configuring a 0.4mol / mL transparent solution of 10ml;

[0021] S2: placing the silicon wafer into acetone and alcohol respectively, and ultrasonic cleaning in deionized water for 10h, and then taking out and blowing dry with nitrogen;

[0022] S3: spin coating the cleaned silicon wafer with photoresist, exposing the silicon wafer to ultraviolet light for 3s, and then placing it in a developing solution for 45s;

[0023] S4: placing the developed silicon wafer into an electron beam evaporation deposition system for deposition, the gas pressure is below 1x10-5Pa, first depositing a 5nm Ti electrode, and then depositing an 80nm Au electrode;

[0024] S5: placing the deposited wafer in a photoresist remover for photoresist removal, to obtain parallel electrodes attached to the silicon wafer;

[0025] S6: placing the wafer in an oxygen plasma cleaning machine for oxygen reaction for 20min, taking two drops of the configured precursor solution and dropping it on the wafer, rotating at a speed of 4000r / s for 1min, and adding isopropyl alcohol at the 5s;

[0026] S7: placing the spin-coated wafer on a hot plate and heating at 250 DEG C for 10min.

[0027] The technical scheme adopted by the present application has the following beneficial effects:

[0028] The bionic neuron memristor of the present application adopts a parallel electrode planar structure, selects high conductivity inert metal as the electrode, and selects a material containing two low activation energy ions in the lattice as the dielectric layer. The excitation of an applied voltage causes the migration of two low activation energy ions in the material, resulting in a built-in electric field, so that the bionic neuron memristor exhibits a voltage opposite to the direction of the applied voltage after the voltage is removed, thereby simulating the change of neuron membrane potential, the cumulative emission phenomenon of neurons and the refractory behavior of neurons. BRIEF DESCRIPTION OF DRAWINGS

[0029] Figure 1 A structural schematic diagram of the bionic neuron memristor provided for the first embodiment of the present application is shown in the figure.

[0030] Figure 2 A diagram of the bionic neuron memristor simulating the change of neuron membrane potential provided for the first embodiment of the present application is shown in the figure.

[0031] Figure 3 A diagram of the bionic neuron memristor simulating the cumulative emission behavior of neurons provided for the first embodiment of the present application is shown in the figure.

[0032] Figure 4 A diagram of the bionic neuron memristor simulating the absolute refractory behavior of neurons provided for the first embodiment of the present application is shown in the figure.

[0033] Figure 5 A diagram of the bionic neuron memristor simulating the relative refractory behavior of neurons provided for the first embodiment of the present application is shown in the figure.

[0034] Figure 6 A schematic diagram of biological neuron signal transmission is shown in the figure.

[0035] Figure 7 A schematic diagram of the refractory behavior of biological neurons and the corresponding ion transport behavior is shown in the figure.

[0036] Figure 8 A flow chart of the preparation method of the bionic neuron memristor provided for the second embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0037] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0038] The following are specific embodiments of the present application and further descriptions of the technical solutions of the present application are made with reference to the accompanying drawings, but the present application is not limited to these embodiments.

[0039] Embodiment One

[0040] The present embodiment provides a bionic neuron memristor, as shown in the figure. Figure 1As shown, it comprises a substrate, parallel electrodes and a dielectric layer; the parallel electrodes and the dielectric layer are both arranged on the substrate;

[0041] The parallel electrodes comprise a positive electrode layer and a negative electrode layer, and the dielectric layer is arranged between the positive electrode layer and the negative electrode layer;

[0042] The dielectric layer adopts a semiconductor material containing two low-activation-energy ions in a lattice;

[0043] There are two low-activation-energy ions in the lattice of the dielectric layer, and under the driving of an electric field, the two low-activation-energy ions are regulated by the electric field and move to the positive and negative electrodes respectively, for simulating the membrane potential change process of sodium and potassium ion transport in and out of the membrane of a neuron when the neuron is stimulated.

[0044] Specifically, the biomimetic neuron memristor takes electrical excitation as an input signal and can realize the related functions of a neuron without applying a peripheral circuit. By adopting a parallel electrode planar structure, selecting a high-conductivity inert metal as the positive electrode layer and the negative electrode layer, and selecting a material containing two low-activation-energy ions in a lattice as the dielectric layer, the migration of the two low-activation-energy ions in the material is caused by the excitation of an external voltage, a built-in electric field is generated, and the biomimetic neuron memristor exhibits a voltage opposite to the direction of the applied voltage after the voltage is removed, thereby simulating the membrane potential change process of a neuron.

[0045] The biomimetic neuron memristor is applied to a neuron filter, brain-like computing, a neural network and binary conversion.

[0046] Specifically, the biomimetic neuron memristor is a planar structure combining an inert positive electrode, a semiconductor perovskite material containing two low-activation-energy ions, and an inert negative electrode, and the biomimetic neuron memristor directly takes electrical stimulation as an input signal and belongs to an electrical excitation type. After a voltage source applies a voltage, under the action of an electric field, the two ions with low activation energy in the material lattice move to the electric field direction and the opposite electric field direction respectively, and when the voltage is removed, a built-in electric field is formed due to the accumulation of the two ions on both sides of the material, discharge occurs, a reverse current is generated, and the change of the membrane potential of a neuron is simulated by using this characteristic.

[0047] Referring to Figure 1 Specifically, the biomimetic neuron memristor selects a silicon wafer as a substrate, and there is an oxidized silicon dioxide layer between the substrate and the parallel electrodes and the dielectric layer; Au is selected as the parallel electrodes with a thickness of 80 nm; there is metal Ti with a thickness of 5 nm between the silicon wafer and the Au as a connection; Cs2AgBiBr6 is selected as the dielectric layer with a thickness of 400 nm. The positive electrode layer and the negative electrode layer of the parallel electrodes are connected with a voltage source and an ammeter, and the substrate is grounded, for applying a voltage to the biomimetic neuron memristor for electrical excitation.

[0048] Wherein, the bionic neuron memristor under the action of electric field, with the increase of voltage, the current from the beginning of the negative value gradually increases to positive value; with the decrease of voltage, the current gradually decreases, from positive to negative, for simulating the change of membrane potential when the neuron is stimulated.

[0049] Specifically, the bionic neuron memristor has a threshold voltage of current instant rise, when a bionic neuron memristor is selected, the threshold voltage is continuously applied, and the voltage stimulation is continuously applied for five times, at the fifth time, the neuron is excited, for simulating the cumulative emission process of neuron.

[0050] Specifically, after the bionic neuron memristor is applied with voltage, the bionic neuron memristor produces mutation current, and the positive and negative ions with small activation energy are accumulated on both sides of the material, so that the same size voltage stimulation is continuously applied, and it is found that the second applied voltage cannot produce mutation current, for simulating the absolute refractory period behavior of neuron. A new bionic neuron memristor is used, the threshold size voltage stimulation is applied for the first time, the bionic neuron memristor produces mutation current, and the voltage stimulation higher than the first size is applied for the second time, the bionic neuron memristor still produces mutation current, for simulating the relative refractory period behavior of neuron.

[0051] Referring to Figure 2 A triangular wave voltage is applied to the bionic neuron memristor from small to large and then from large to small, when the voltage gradually increases, the current slowly increases to the threshold size, as the voltage continues to increase, the current rapidly increases beyond the threshold, as the voltage size decreases, the current gradually decreases and appears negative value, and as the voltage decreases to zero, the current slowly returns to the initial value, simulating the change of membrane potential when the neuron is stimulated.

[0052] Referring to Figure 3 A pulse voltage smaller than the threshold voltage is applied to the bionic neuron memristor all the time, and is continuously applied for five times, until the fifth time, the current exceeds the threshold, and the action potential changes.

[0053] Referring to Figure 4 Two triangular wave pulse voltages with the same size are applied to the bionic neuron memristor, and the pulse voltage size is higher than the threshold value of the device, the current of the device exceeds the threshold value under the first pulse, which shows that the neuron emits, and under the same size pulse immediately, the current size is less than the threshold value, and the neuron does not emit, which shows the absolute refractory period behavior of neuron.

[0054] Referring to Figure 5 A voltage larger than the threshold value is first applied to the bionic neuron memristor, the current of the device is higher than the threshold value, simulating the emission of neuron, and then a voltage larger than the first voltage is applied, the current of the device is still higher than the threshold value, simulating the emission of neuron, and the whole simulates the relative refractory period phenomenon of neuron.

[0055] The material of the positive electrode layer and the negative electrode layer is inert metal oxide; the inert metal oxide is one or both of Au and Pt.

[0056] Specifically, the positive electrode layer and the negative electrode layer are generated by ultraviolet lithography and electron beam evaporation.

[0057] The medium layer is made of pure inorganic perovskite material, and the semiconductor perovskite material containing two low-activation-energy ions in the crystal lattice is one of Cs2AgBiCl6, Cs2AgBiBr6, and Cs2AgBiI6.

[0058] Specifically, the medium layer can be prepared by coating, spin coating, adding anti-solvent, and then evaporating the solvent to obtain a coating layer, or by evaporation to prepare a perovskite medium layer.

[0059] As shown in Figure 6 and Figure 7 show the schematic diagram of biological neuron signal transmission and the schematic diagram of biological neuron refractory behavior and corresponding ion transport behavior, respectively; the biological neuron electric pulse signal emission is realized by the joint action of sodium ions and potassium ions.

[0060] Specifically, there are two kinds of ions in the biomimetic neuron memristor, which are regulated by an electric field. Compared with the traditional one kind of ion regulated by an electric field, it can better simulate the related behavior of neurons without the need to design and connect peripheral circuits. The biomimetic neuron memristor can simulate multiple behaviors of neurons, and because the ion regulated by the electric field has a clear activation energy, the device has a clear threshold when the current changes suddenly, better simulating the activity of sodium ion pump, the hyperpolarization process, and the resting potential less than 0 in real neurons. The biomimetic neuron memristor has good air stability, and the test can be repeated multiple times, while the performance of traditional organic perovskite material will continue to decline with the test. The biomimetic neuron memristor has great application potential in the field of robots and neuromorphic computing.

[0061] The thickness of the medium layer is 400-500 nm.

[0062] The external voltage is applied between the positive electrode layer and the negative electrode layer, and the size and direction of the voltage are adjusted to simulate the change of the membrane potential inside and outside the cell membrane.

[0063] The positive electrode layer, the negative electrode layer, and the medium layer are in direct contact with the substrate, and the material of the substrate is a silicon wafer.

[0064] The bionic neuron memristor of the application adopts a parallel electrode planar structure, selects high-conductivity inert metal as the electrode, and selects a material containing two low-activation-energy ions in the crystal lattice as the dielectric layer. The excitation of an external voltage causes the migration of the two low-activation-energy ions in the material, generates a built-in electric field, and makes the bionic neuron memristor exhibit a voltage opposite to the direction of the applied voltage after the voltage is removed, thereby simulating the change of the membrane potential of a neuron, the cumulative emission phenomenon of a neuron, and the refractory period behavior of a neuron.

[0065] Embodiment Two

[0066] The embodiment provides a preparation method of a bionic neuron memristor, as shown in the figure, the method comprises the steps of: Figure 8

[0067] S1: configuring a precursor solution, dissolving CsBr, BiBr3 and AgBr in dimethyl sulfoxide at a ratio of 2:1:1, heating and stirring at 75 DEG C for 4h, and configuring 10ml of a transparent solution of 0.4mol / mL;

[0068] S2: placing the silicon wafer into acetone and alcohol respectively, ultrasonic cleaning in deionized water for 10h, and taking out and blowing dry with nitrogen;

[0069] S3: spin-coating photoresist on the cleaned silicon wafer, exposing the silicon wafer to ultraviolet light for 3s, and then developing in a developing solution for 45s;

[0070] S4: placing the developed silicon wafer into an electron beam evaporation deposition system for deposition, the gas pressure is below 1x10-5Pa, first depositing a 5nm Ti electrode, and then depositing an 80nm Au electrode;

[0071] S5: placing the deposited wafer into a stripping solution for stripping to obtain parallel electrodes attached to the silicon wafer;

[0072] S6: placing the wafer in an oxygen plasma cleaning machine for oxygen reaction for 20min, taking two drops of the configured precursor solution and adding it on the wafer, rotating at a speed of 4000r / s for 1min, and adding isopropyl alcohol at the 5th second;

[0073] S7: placing the spin-coated wafer on a hot plate and heating at 250 DEG C for 10min.

[0074] The bionic neuron memristor is applied to a neuron filter, brain-like computing, a neural network and binary conversion.

[0075] ​The method prepares a biomimetic neuron memristor. The biomimetic neuron memristor selects high-conductivity inert metal as an electrode and selects a material containing two low-activation-energy ions in a crystal lattice as a dielectric layer. By using the excitation of an applied voltage, the migration of the two low-activation-energy ions in the material is caused, a built-in electric field is generated, the biomimetic neuron memristor shows a voltage opposite to the direction of the applied voltage after the voltage is removed, and thus the change of the membrane potential of a neuron, the cumulative emission phenomenon of a neuron and the refractory period behavior of a neuron are simulated.

[0076] Although the present disclosure is disclosed as above, the protection scope of the present disclosure is not limited to this. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present disclosure, and these changes and modifications shall fall within the protection scope of the present disclosure.

Claims

1. A biomimetic neuron memristor, characterized in that, It includes a substrate, a parallel counter electrode, and a dielectric layer; both the parallel counter electrode and the dielectric layer are disposed on the substrate. The parallel counter electrode includes a positive electrode layer and a negative electrode layer, and the dielectric layer is disposed between the positive electrode layer and the negative electrode layer; The dielectric layer is made of a semiconductor material containing low activation energy ions in its crystal lattice; The lattice in the dielectric layer contains two types of low activation energy ions. Under the drive of an electric field, the two types of low activation energy ions are regulated by the electric field and move towards the positive and negative poles respectively, which is used to simulate the membrane potential change process caused by the change in sodium and potassium ion transport across the membrane when a neuron is stimulated. The dielectric layer is made of pure inorganic perovskite material, and the semiconductor perovskite material containing two low activation energy ions in the lattice is one of Cs2AgBiCl6, Cs2AgBiBr6, and Cs2AgBiI6.

2. The biomimetic neuron memristor according to claim 1, characterized in that, Under the influence of an electric field, the current in the bionic neuron memristor gradually increases from a negative value to a positive value as the voltage increases; conversely, as the voltage decreases, the current gradually decreases from a positive value to a negative value, thus simulating the change in membrane potential when a neuron is stimulated.

3. The biomimetic neuron memristor according to claim 1, characterized in that, The positive electrode layer and the negative electrode layer are made of one or both of Au and Pt.

4. The biomimetic neuron memristor according to claim 1, characterized in that, The thickness of the dielectric layer is 400-500 nm.

5. The biomimetic neuron memristor according to claim 1, characterized in that, An external voltage is applied between the positive electrode layer and the negative electrode layer, and the magnitude and direction of the voltage are adjusted to simulate the changes in the membrane potential between the inside and outside of the cell membrane.

6. The biomimetic neuron memristor according to claim 1, characterized in that, The positive electrode layer, the negative electrode layer, and the dielectric layer are all in direct contact with the substrate, and the substrate is made of silicon wafer.

7. The method for fabricating a biomimetic neuron memristor as described in any one of claims 1-6, characterized in that, Including the following steps: S1: Prepare the precursor solution by dissolving CsBr, BiBr3, and AgBr in dimethyl sulfoxide in a ratio of 2:1:1 and heating and stirring at 75°C for 4 hours to prepare 10 ml of a 0.4 mol / mL clear solution. S2: Place the silicon wafer in acetone and alcohol respectively, ultrasonically clean it in deionized water for 10 hours, and then take it out and dry it with nitrogen gas. S3: Spin-coat the cleaned silicon wafer with photoresist, expose the silicon wafer to ultraviolet light for 3 seconds, and then immerse it in the developer for 45 seconds. S4: Place the developed silicon wafer into an electron beam evaporation deposition system for deposition, with a gas pressure of 1x10⁻⁶. -5 Below Pa, first deposit a 5nm Ti electrode, then deposit an 80nm Au electrode; S5: Place the deposited wafer in a resist remover solution to remove the resist, and obtain parallel counter electrodes attached to the silicon wafer; S6: Place the tablet in an oxygen plasma cleaner and react with oxygen for 20 minutes. Then, add two drops of the prepared precursor solution to the tablet and rotate it at 4000 r / s for 1 minute. At the 5th second, add isopropanol. S7: Place the spin-coated film on a hot plate and heat at 250°C for 10 minutes.

Images