Neuromorphic memristor based on thiophene organic polymer and preparation method thereof

Abstract

The invention discloses a neuromorphic memristor based on a thiophene organic polymer and a preparation method of the neuromorphic memristor, and belongs to the field of organic diode memristor preparation. The neuromorphic memristor comprises a top electrode, an organic active layer, a bottom electrode and a substrate from top to bottom, wherein the organic active layer is prepared from a thiophene organic polymer poly-3-dodecyl thiophene PQT-12. According to the invention, the PQT-12 is used as a key material, the plasticity of biological synapses is simulated, efficient information storage and processing are realized, and the PQT-12 is used as an organic active layer and has the characteristics of excellent charge retention capability and low power consumption. The memristor prepared by the invention shows excellent performance in the aspects of switching speed, stability and expandability, has a wide application prospect, and is particularly applied to the fields of neuromorphic calculation, low-power-consumption storage devices, artificial intelligence systems and the like.

Description

Technical Field

[0001] The present invention belongs to the field of preparation of organic diode memristor devices, in particular to a neuromorphic memristor based on thiophene organic polymer and a preparation method thereof. Background Art

[0002] With the continuous advancement of science and technology, information technology and artificial intelligence (AI) technology have flourished and have been widely used in various fields, such as intelligent driving, AI translation, and intelligent manufacturing. However, the computing of emerging technologies such as artificial intelligence that require processing large amounts of data currently uses the traditional von Neumann architecture, which has an inherent energy efficiency bottleneck when performing neural computing. Its computing units and storage units are separated from each other, and data will generate a lot of energy consumption and delays when communicating. This defect limits the development of emerging technologies and determines their upper limit. In contrast, the human brain can achieve efficient information perception, processing, and memory at extremely low power consumption. Therefore, developing devices that simulate the human brain or neural network for computing can break the traditional von Neumann bottleneck and is a feasible method for efficient computing.

[0003] Previous studies have attempted to simulate human brain neuron synapses based on complementary metal oxide semiconductors (CMOS). However, when the device is scaled down according to Moore's law, the leakage power consumption will become very large, thus limiting its operating frequency. In recent years, with the gradual failure of Moore's law, neuromorphic computing has gradually become a technology with great potential. Its system has a high degree of interconnectivity and parallelism, and can process data directly in memory while maintaining low energy consumption. The emerging artificial synaptic devices based on memristors, or neuromorphic memristors, have attracted more and more research interests and attention in simulating brain synaptic plasticity due to their advantages such as low power consumption, high data storage density and fast operation. Internationally, the research on neuromorphic memristors mainly focuses on the selection of materials and the optimization of device performance. Among the many memristor materials, polymer-based memristors are favored due to their excellent flexibility, processability and relatively low production cost.

[0004] As an organic semiconductor material, thiophene organic polymers such as PQT-12 (poly 3-dodecylthiophene) are widely used in organic electronic devices due to their good electrical conductivity and thermal stability. The molecular structure of PQT-12 contains long-chain alkyl substituents, which not only enhances its solubility but also improves its charge mobility, making it have good prospects in the application of neuromorphic memristors. However, there are still relatively few systematic studies on thiophene organic polymers as neuromorphic memristor materials, and they have certain limitations and instabilities. Therefore, exploring the performance characteristics and application potential of thiophene organic polymer-based memristors has important theoretical value and practical significance. Summary of the invention

[0005] The purpose of the present invention is to provide a neuromorphic memristor based on thiophene organic polymer and a preparation method thereof.

[0006] The neuromorphic memristor based on thiophene organic polymer provided by the present invention comprises a top electrode, an organic active layer, a bottom electrode and a substrate from top to bottom, and the organic active layer is prepared from an organic polymer poly 3-dodecylthiophene (PQT-12).

[0007] Furthermore, the preparation method of the organic active layer is: Poly (3-dodecylthiophene) (PQT-12) is dissolved in chloroform to prepare a poly (3-dodecylthiophene) solution, and the prepared solution is spin-coated onto a substrate with a bottom electrode using a solution deposition technique to form a film.

[0008] Furthermore, the concentration of the poly (3-dodecylthiophene) solution is 3 mg / ml, and the thickness of the organic active layer is 50-80 nm.

[0009] Furthermore, the material of the bottom electrode is indium tin oxide, the bottom electrode is a layer of indium tin oxide grown on the surface of the substrate, the bottom electrode is the anode of the neuromorphic memristor, and the thickness of the bottom electrode is 130-150 nm.

[0010] Furthermore, the material of the top electrode is aluminum, the top electrode is the cathode of the neuromorphic memristor, and the thickness of the top electrode is 90-120 nm.

[0011] Furthermore, the material of the substrate is glass.

[0012] The present invention further includes a method for preparing the above-mentioned thiophene organic polymer neuromorphic memristor, characterized by comprising the following steps: Step 1: growing a bottom electrode layer on the surface of the substrate, and cleaning the substrate with the bottom electrode with ethanol, glass cleaning agent, and deionized water ultrasonically in sequence, and drying it with ordinary nitrogen, followed by drying and ultraviolet ozone treatment; Step 2: dissolving poly (3-dodecylthiophene) in chloroform to prepare a poly (3-dodecylthiophene) solution, and evenly spin-coating the poly (3-dodecylthiophene) solution on the surface of the material treated in step 2, followed by annealing in an oven.

[0013] Step 3: Place the substrate with the organic active layer into a vacuum evaporation system, obtain a top electrode on the organic active layer through evaporation treatment, and obtain a neuromorphic memristor based on thiophene organic polymer after the top electrode is cooled.

[0014] Furthermore, in step 1, the ultrasonic cleaning time of ethanol, glass cleaning agent and ultrapure water is 25 to 30 minutes, the ultraviolet ozone treatment time is 10 to 15 minutes, and the drying temperature is set to 120°C.

[0015] Furthermore, in the step 2, the concentration of the poly 3-dodecylthiophene solution is 3 mg / ml solution.

[0016] Furthermore, in the step 2, the poly 3-dodecylthiophene solution is spin-coated on the substrate with the bottom electrode treated in step 1 to form a film with the help of a pipette, and then placed in an oven at 60°C for annealing for 30 minutes; the spin coating process is: first rotate at a low speed of 500 rpm for 9 s, and then rotate at a high speed of 3000 rpm for 30 s until an organic active layer of a set thickness is obtained, and the thickness of the organic active layer is 50nm to 80nm.

[0017] Furthermore, in step 3, during the evaporation process, the vacuum evaporation system is evacuated to a pressure in the chamber lower than 5×10 -4 Pa, the top electrode is evaporated, the evaporation rate of the top electrode is 1-3 Å / s, and the thickness of the top electrode is controlled by a crystal oscillator. After the treatment, the vacuum state is maintained until the top electrode is cooled to room temperature. The cooling time of the top electrode is 20 to 30 minutes.

[0018] Compared with the prior art, the present invention has the following significant advantages: 1) The memristor prepared by the present invention has the characteristics of low switching voltage and low power consumption, which makes it particularly advantageous in low-energy consumption applications. Compared with some inorganic memristors, it can operate with smaller voltage and current, reducing energy consumption and improving the long-term stability and reliability of the device.

[0019] 2) The memristor prepared by the present invention has the characteristics of simple structural design and easy operation, and can be processed by solution deposition method (spin coating) at low cost. This has obvious cost advantages over traditional hard material preparation methods (such as vacuum evaporation) and can provide higher efficiency in large-area, high-throughput production.

[0020] 3) The memristor prepared by the present invention can simulate the behavior of neural synapses and has brain-like computing and neuromorphic processing capabilities. This makes it have application advantages in the fields of artificial intelligence and machine learning. Compared with traditional electronic computing devices, it can efficiently perform neural network tasks and has higher parallel computing capabilities.

[0021] 4) PQT-12 in the organic active layer of the memristor prepared by the present invention is an organic material with good flexibility and is suitable for wearable devices, flexible electronic products and other fields. In addition, thiophene organic polymers are also biocompatible, making the memristor prepared by the present invention have potential in the fields of biomedicine and brain-like computing. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 A schematic diagram of the structure of a thiophene polymer-based neuromorphic memristor prepared by the present invention; Figure 2 A schematic diagram of the forward current-voltage of the thiophene polymer-based neuromorphic memristor prepared in Example 1 at room temperature; Figure 3 A negative current-voltage schematic diagram of a thiophene polymer-based neuromorphic memristor prepared in Example 1 at room temperature; Figure 4 This is a graph showing the change in forward conductance of the thiophene polymer-based neuromorphic memristor prepared in Example 1 at room temperature; Figure 5 This is a graph showing the negative conductance change of the thiophene polymer-based neuromorphic memristor prepared in Example 1 at room temperature; Figure 6 A schematic diagram of simulation of the excitatory postsynaptic current (EPSC) of the neuromorphic memristor based on thiophene polymer prepared in Example 1; Figure 7 A schematic diagram of a double-pulse facilitated synaptic function simulation of a thiophene polymer-based neuromorphic memristor prepared in Example 1; Figure 8 A schematic diagram of a double-pulse inhibition synaptic function simulation of a thiophene polymer-based neuromorphic memristor prepared in Example 1; Fig. 9 A schematic diagram of simulation of the post-tetanic potentiation synaptic function of the thiophene polymer-based neuromorphic memristor prepared in Example 1; Fig.10 A schematic diagram of simulation of enhancement and inhibition of synaptic function of a neuromorphic memristor based on thiophene polymer prepared in Example 1; Fig.11 Schematic diagram of the current response of the thiophene polymer-based neuromorphic memristor prepared in Example 1 to 10 pulses with different amplitudes. DETAILED DESCRIPTION

[0023] The technical solution of the present invention is described in detail below through embodiments, but the protection scope of the present invention is not limited to the embodiments.

[0024] In the examples of the present invention, poly (3-dodecylthiophene) (PQT-12) was purchased from Sigma-Aldrich (Shanghai) Trading Co., Ltd.

[0025] The present invention provides a neuromorphic memristor based on thiophene organic polymer and a preparation method thereof. The structure of the neuromorphic memristor is as follows: Figure 1 As shown, from top to bottom, it includes a top electrode 1, an organic active layer 2, a bottom electrode 3 and a substrate 4, and is a vertical structure as a whole, wherein the organic active layer is prepared from poly 3-dodecylthiophene.

[0026] The thiophene polymer poly 3-dodecylthiophene PQT-12 has a good π-π conjugated structure, which can effectively improve the conductivity of electrons. This enables it to perform electron migration well when used as an organic active layer in a memristor, improving the switching characteristics and response speed of the device. The advantage of using PQT-12 as an organic active layer compared to other thiophene organic polymers is that PQT-12 has a highly conjugated π-electron system, which enhances the charge transfer capability, enabling it to achieve more stable conductivity regulation in memristor devices. Its thiophene conjugated main chain makes the electron mobility inside the material higher, reduces the scattering loss of electrons in the pathway, and thus improves the stability and response speed of the memristor effect. The long alkyl side chain (dodecyl) in the PQT-12 structure gives it good solution processability and thin film forming ability, making it easy to form high-quality thin films through low-cost methods such as spin coating and spray coating. The molecular structure of PQT-12 is conducive to the migration of ions / protons in the material, which is crucial for memristor devices. Compared with PEDOT, the memristive effect of PQT-12 is more stable and non-volatile, while the memristive effect of PEDOT mainly depends on ion doping / dedoping, which is easily affected by the environment, resulting in poor information retention. In the present invention, PQT-12 regulates ion transitions through electric fields and is suitable for refined synaptic function simulation, such as synaptic weight regulation, STDP (pulse time-dependent plasticity), PPF (double pulse facilitation), etc. In addition, PQT-12 is compatible with organic semiconductor manufacturing processes and can be integrated with silicon-based CMOS, while the acidity of PEDOT will corrode metal electrodes and affect the life of integrated circuits. Therefore, in the application of high-performance memristors and neuromorphic computing devices, thiophene organic polymer PQT-12 has more advantages.

[0027] Example 1 (1) Glass was selected as the substrate, and a layer of indium tin oxide was generated on the glass as the bottom electrode with a thickness of 150 nm to form ITO conductive glass. The ITO conductive glass was cleaned with ethanol, glass cleaning agent, and ultrapure water for 30 min in sequence, then dried with ordinary nitrogen and placed in an oven at 120°C for 60 min.

[0028] (2) The dried ITO conductive glass was treated with UV-ozone for 15 min.

[0029] (3) Use a pipette to spin-coat a 3 mg / mL PQT-12 solution dissolved in chloroform on the ITO conductive glass to form a film, and then anneal it in a drying oven at 60 °C for 30 min. The formed PQT-12 layer is the organic active layer. The spin coating process is: first spin at a low speed of 500 rpm for 9 s, then spin at a high speed of 3000 rpm for 30 s, until the thickness of the organic active layer is 60 nm.

[0030] (4) Place the prepared ITO conductive glass with an organic active layer into a vacuum evaporation device and control the pressure in the vacuum chamber to 5×10 -4 Pa, and start evaporating a metal aluminum electrode as the top electrode at a rate of 1-3 Å / s and a thickness of 120 nm.

[0031] (5) After the evaporation experiment is completed, the metal electrode is cooled to room temperature in a vacuum chamber to obtain a neuromorphic memristor based on thiophene organic polymer. The chamber is then opened to take out the neuromorphic memristor and conduct relevant electrical performance tests.

[0032] (6) Performance test results: Figure 2 and Figure 3 , respectively, are the typical current-voltage (I-V) characteristic curves of the neuromorphic memristor under positive voltage scan and negative voltage scan at room temperature. During the electrical test, the aluminum electrode (top electrode) was grounded, and the electrical signal was applied to the ITO electrode (bottom electrode). Under the positive voltage (0-6-0V) scan, the current corresponding to the neuromorphic memristor at 6V gradually increased, which means that the overall conductive state of the neuromorphic memristor gradually increased; conversely, under the negative voltage (0--6-0V) scan, the current corresponding to the neuromorphic memristor at -6V gradually decreased, which means that the overall conductive state of the neuromorphic memristor gradually decreased.

[0033] Figure 4 This is a diagram showing the change in forward conductance of the thiophene polymer-based neuromorphic memristor prepared in Example 1 at room temperature, that is, a schematic diagram of its enhanced synaptic function simulation at room temperature. Figure 5 This is a diagram showing the negative conductance change of the thiophene polymer-based neuromorphic memristor prepared in Example 1 at room temperature, that is, a schematic diagram of its inhibitory synaptic function simulation at room temperature.

[0034] from Figure 3 and Figure 4It can be clearly seen in the figure that under the test of periodic bipolar electrical signals, the neuromorphic memristor has a non-characteristic curve of hysteresis in its electrical characteristics on the IV plane, which is a typical feature of memristors. The hysteresis phenomenon is essentially a manifestation of the memristor simulating the plasticity of neural synapses.

[0035] Figure 4 and Figure 5 This is a schematic diagram of the conductivity change of the neuromorphic memristor based on thiophene organic polymer prepared in Example 1 at room temperature. In a memristor, the conductivity is similar to the synaptic weight. As can be seen from the figure, the change in conductivity is nonlinear, which means that when a voltage is applied, the current does not simply change in proportion to the voltage. As a series of periodic electrical signals are applied to the neuromorphic memristor, the conductivity of the neuromorphic memristor changes, thereby forming a memory effect in the neuromorphic memristor. This nonlinear transmission characteristic enables the memristor to simulate the weight adjustment process of the synapse.

[0036] Figure 6 Schematic diagram of the simulation of the synaptic function of the excitatory postsynaptic current (EPSC) of the neuromorphic memristor of the thiophene organic polymer prepared in Example 1. The generation of EPSC is usually a fast transient process. Under the voltage stimulation of +6V, the current rises sharply and quickly reaches a peak value, and then gradually decreases over time.

[0037] Figure 7 and Figure 8 Schematic diagram of the simulation of the synaptic function of the double pulse facilitation (PPF) and double pulse depression (PPD) of the neuromorphic memristor of thiophene organic polymer prepared in Example 1. In the neural synapse, double pulse facilitation (PPF) and double pulse depression (PPD) respectively represent the effects of two short-time-interval stimulations on synaptic transmission. When the time interval between the two pulses is short (6V, 100ms), the reaction caused by the second pulse is stronger than that of the first pulse. This phenomenon is called facilitation. It usually indicates the enhancement of the postsynaptic response, that is, the enhancement of the synaptic weight. This effect simulates the short-term potentiation (STP) process in the neural synapse; on the contrary, double pulse depression means that when the time interval between the two pulses is short, the reaction of the second pulse is weaker than that of the first pulse. This phenomenon reflects the inhibition of the postsynaptic response and simulates the short-term depression (STD) process in the neural synapse.

[0038] Fig. 9 Schematic diagram of the thiophene organic polymer neuromorphic memristor simulating post-tetanic potentiation (PTP) synaptic function prepared in Example 1. Similar to PPF, post-tetanic potentiation (PTP) refers to the continuous increase in the strength of synaptic weights after high-frequency stimulation. When a continuous positive voltage signal (6V, 100ms) is applied to the neuromorphic memristor, the current continues to increase, forming an upward curve.

[0039] Fig.10 The conductivity change curve of the neuromorphic memristor of thiophene organic polymer prepared in Example 1 under the application of continuous positive and negative pulses, wherein (a) is the conductivity change curve under the application of continuous positive pulses, and (b) is the conductivity change curve under the application of continuous negative pulses. When a positive electrical signal (6V) is applied, the conductivity gradually increases; when a negative electrical signal is applied, the conductivity gradually decreases.

[0040] Fig.11 Schematic diagram of the current response of the neuromorphic memristor of thiophene organic polymer prepared in Example 1 to 10 different pulse amplitudes (6V, 7V, 8V, 9V). In the neuromorphic memristor, the change in the conductivity of the neuromorphic memristor is also affected by the change in the pulse amplitude (voltage intensity). In other words, the amplitude of the pulse determines the plasticity of the memristor and affects the increase and decrease of its conductivity. A series of pulse stimuli with different amplitudes (6V, 7V, 8V, 9V) were applied to the neuromorphic memristor. The results showed that the larger the voltage amplitude, the greater the overall level of synaptic weight.

[0041] In summary, the present invention lays a foundation for exploring neuromorphic memristors based on thiophene organic polymers and promotes the research of neuromorphic memristors. The neuromorphic memristors prepared by the present invention are expected to be widely used in the fields of intelligent hardware, neuromorphic computers, autonomous driving, and robots, and promote the development of future technologies.

[0042] As described above, although the present invention has been shown and described with reference to specific preferred embodiments, it should not be construed as limiting the present invention itself. Various changes may be made to the form and details without departing from the spirit and scope of the present invention.

Claims

1. A neuromorphic memristor based on thiophene organic polymer, characterized in that: The neuromorphic memristor comprises, from top to bottom, a top electrode, an organic active layer, a bottom electrode and a substrate, wherein the organic active layer is made of poly-3-dodecylthiophene.

2. The neuromorphic memristor according to claim 1, characterized in that The preparation method of the organic active layer is: Poly (3-dodecylthiophene) is dissolved in chloroform to prepare a poly (3-dodecylthiophene) solution, and then the poly (3-dodecylthiophene) solution is spin-coated on a substrate with a bottom electrode to form an organic active layer after film formation.

3. The neuromorphic memristor according to claim 2, characterized in that The concentration of the poly 3-dodecylthiophene solution is 3 mg / ml, and the thickness of the organic active layer is 50-80 nm.

4. The neuromorphic memristor according to claim 1, characterized in that The material of the top electrode is metal aluminum; the material of the bottom electrode is indium tin oxide; and the material of the substrate is glass.

5. The neuromorphic memristor according to claim 1, characterized in that The thickness of the top electrode is 90-120 nm; the thickness of the bottom electrode is 130-150 nm.

6. The method for preparing a neuromorphic memristor according to any one of claims 1 to 5, characterized in that: The steps include: Step 1: The substrate with the bottom electrode is cleaned with ethanol, glass cleaning agent, and deionized water ultrasonically in sequence, dried with nitrogen, and then dried and treated with ultraviolet ozone; Step 2: dissolving poly (3-dodecylthiophene) in chloroform to prepare a poly (3-dodecylthiophene) solution, and evenly spin-coating the poly (3-dodecylthiophene) solution on the substrate with the bottom electrode obtained in step 1, and then placing the substrate in an oven for annealing to obtain a substrate with an organic active layer; Step 3: Place the substrate with the organic active layer into a vacuum evaporation system, obtain a top electrode on the organic active layer through evaporation treatment, and obtain a neuromorphic memristor based on thiophene organic polymer after the top electrode is cooled.

7. The preparation method according to claim 6, characterized in that: In the step 2, the poly 3-dodecylthiophene solution is spin-coated on the substrate with the bottom electrode by means of a pipette gun to form a film, and then placed in an oven at 60° C. for annealing for 20 to 30 minutes.

8. The preparation method according to claim 6, characterized in that: In the step 2, the spin coating process is: firstly rotating at a low speed of 500 rpm for 9 s, and then rotating at a high speed of 3000 rpm for 30 s, until an organic active layer of a set thickness is obtained.

9. The preparation method according to claim 6, characterized in that: In the step 1, the ultrasonic cleaning time of ethanol, glass cleaning agent and ultrapure water is 25 to 30 minutes, and the ultraviolet ozone treatment time is 10 to 15 minutes.

10. The preparation method according to claim 6, characterized in that: In the step 3, during the evaporation process, the vacuum evaporation system is evacuated to a pressure in the chamber lower than 5×10 -4 Pa, the top electrode was deposited at a rate of 1-3 Å / s. A crystal oscillator was used to control the thickness of the top electrode. After the treatment, the vacuum state was maintained until the top electrode cooled to room temperature.

Images