Artificial synaptic device based on Nb2C-zinc oxide memristor and preparation method thereof
By forming a heterostructure of the two-dimensional Nb2C layer on the zinc oxide layer, the problem of performance degradation of oxide memristors under long-term operation is solved, the stability and response speed are improved, and it is suitable for high-performance neuromorphic calculations and Mass production.
Patent Information
- Application Number
- CN202411003103.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-25
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2044-07-25
AI Technical Summary
Existing oxide memristors have performance degradation problems in long-term operation or extreme environments, which affects their application and development in the field of brain-like computing.
Using artificial synaptic devices based on Nb2C-zinc oxide, a two-dimensional Nb2C layer is formed on the zinc oxide layer through spin coating or spraying process to form a heterostructure, combining substrate layers such as silicon, glass, silicon carbide and electrodes such as platinum and gold. Material, a memristor device that forms a three-layer structure.
It improves the stability and durability of the device, reduces energy consumption and operating voltage, enhances the response speed and storage density of the device, and is suitable for large-scale production and high-performance neuromorphic calculations.
Smart Images

Figure CN118555906B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of memristors, and in particular relates to an artificial synaptic device based on a Nb2C-zinc oxide memristor and a preparation method thereof. Background Art
[0002] The von Neumann architecture is a typical representative of traditional computer architecture and is a milestone in history. However, the traditional von Neumann architecture has many limitations, including the memory bottleneck problem, that is, the performance bottleneck caused by data and instructions sharing the same bus and storage system; the limitations of the instruction pipeline, such as pipeline conflicts and pipeline stalls; memory access delays, which make the processor wait too long for memory responses; security issues, especially attacks such as buffer overflows; concurrency issues, due to limited parallel processing capabilities; and scalability limitations, which affect the system's performance when processing large-scale data and complex computing tasks. These problems together limit the overall performance and efficiency of computer systems.
[0003] Artificial synapses simulate the dynamic connections and plasticity of biological synapses, allowing computers to self-learn and adapt without prior programming. This adaptability enables neuromorphic computing systems to demonstrate extremely high efficiency and accuracy in areas such as image recognition, natural language processing, and complex decision-making. In addition, the use of artificial synapses reduces the need for data transmission between different hardware components, significantly reducing latency and energy consumption, thereby demonstrating excellent performance when processing large-scale data.
[0004] Memristors, as the fourth basic circuit element besides resistors, capacitors, and inductors, have shown great application potential. They have the advantages of high-speed computing capability, low power consumption, and easy integration, and are perfectly compatible with existing CMOS processes. Memristors are seen as a strong competitor for the next generation of non-volatile storage technology. In addition, the working principle and behavior of memristors are similar to those of biological neural synapses, making them not only suitable for logical operations in non-volatile states, but also capable of performing brain-like neuromorphic computing tasks, opening up new paths for future computing technologies.
[0005] At present, in the field of memristor technology, memristors based on oxide material systems have shown high maturity and application prospects. This type of memristor mainly uses transition metal oxides (such as TiO2, ZnO, etc.) as its resistive functional layer. This material has been widely studied due to its unique resistance change characteristics and excellent physical properties. Transition metal oxide memristors have shown significant advantages over traditional storage technologies due to their high-speed switching, low energy consumption, easy integration with existing semiconductor processes, high storage density, and compatibility with CMOS processes. They are considered to be brain-like computing devices with great development potential.
[0006] However, despite the many attractive advantages of oxide memristors, they still face some challenges in commercialization and widespread application. Among them, the stability and durability of the device is one of the current research focuses. For example, memristors may encounter performance degradation problems after long-term operation or in extreme environments, which may be caused by factors such as material aging and unstable expansion of the conductance channel. This greatly limits the application and development of memristors in the field of brain-like computing. Summary of the invention
[0007] In view of the deficiencies in the prior art, the purpose of the present invention is to provide an artificial synaptic device based on Nb2C-zinc oxide memristor and a preparation method thereof, thereby solving the problems in the prior art.
[0008] The purpose of the present invention can be achieved through the following technical solutions:
[0009] The artificial synaptic device based on Nb2C-zinc oxide memristor includes, from bottom to top, a substrate layer, a bottom electrode layer, a resistive switching functional layer and a top electrode layer; wherein the resistive switching functional layer is a heterostructure composed of a two-dimensional Nb2C layer and a zinc oxide layer, and the resistive switching functional layer is prepared by covering the two-dimensional Nb2C layer on the zinc oxide layer by a spin coating or spray coating process.
[0010] Furthermore, the substrate layer is made of silicon, glass, silicon carbide or gallium nitride.
[0011] Furthermore, the bottom electrode layer is made of platinum, gold or tungsten.
[0012] Furthermore, the top electrode layer is made of silver, copper or iron.
[0013] Furthermore, the thickness of the top electrode layer and the bottom electrode layer is 80-100 nm.
[0014] The method for preparing the artificial synaptic device based on Nb2C-zinc oxide memristor comprises the following steps:
[0015] S1, depositing a bottom electrode on the substrate layer by DC magnetron sputtering or electron beam evaporation to form a bottom electrode layer;
[0016] S2, depositing zinc oxide on the bottom electrode layer by radio frequency magnetron sputtering to form a zinc oxide layer;
[0017] S3, fully stirring and mixing the Nb2C powder and the dimethyl sulfoxide solution to obtain a Nb2C mixed solution; and coating the Nb2C mixed solution on the zinc oxide layer by spin coating or spray coating, and then drying to form a two-dimensional Nb2C layer;
[0018] S4, attaching a mask plate to the two-dimensional Nb2C layer, and depositing a top electrode layer on the two-dimensional Nb2C layer by a DC magnetron sputtering method.
[0019] Furthermore, the top electrode layer pattern is a rectangle of 450 micrometers×450 micrometers.
[0020] Furthermore, the concentration of the Nb2C mixed solution is 10 mg / ml.
[0021] An image recognition system comprises the above-mentioned artificial synaptic device based on Nb2C-zinc oxide memristor.
[0022] Beneficial effects of the present invention:
[0023] 1. The artificial synaptic device based on Nb2C-zinc oxide memristor and its preparation method proposed in the present invention use a new heterostructure of Nb2C-zinc oxide as the functional layer to form a three-layer artificial synaptic device with the top electrode and the bottom electrode. The preparation method has simple operation steps, low energy consumption and high yield. The memristor using Cu and W as electrode materials can enhance the compatibility of the device's back-end process and is suitable for large-scale production.
[0024] 2. The artificial synaptic device based on Nb2C-zinc oxide memristor proposed in the present invention simulates the characteristics of neural synapses well, and is applied to the development of storage-computing integrated structure and high-performance neuromorphic computing, promoting the development of highly integrated and miniaturized brain-like chips.
[0025] 3. As a transition metal carbide, the electronic structure of Nb2C in the present invention shows metallic conductivity. The introduction of Nb2C enables the device to perform well in fast response and low power consumption applications. At the same time, compared with other transition metal carbide materials, Nb2C has better chemical stability in air and water, so the introduction of Nb2C can improve the stability of the device under long-term operation.
[0026] 4. The present invention introduces two-dimensional layered Nb2C and ZnO to form a new heterostructure as the resistive switching functional layer of the memristor. The unique structural characteristics of the two-dimensional material are used to guide the formation of conductive filaments, so that the memristor has better electrical properties, lower operating voltage and lower energy consumption, as well as better stability and durability. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0028] Figure 1 Schematic diagram of the structure and electrical testing of the artificial synaptic device of the present invention;
[0029] Figure 2 This is a diagram of an experiment of simulating synaptic weight enhancement of the artificial synaptic device of the present invention under continuous application of 10 forward voltage scanning cycles;
[0030] Figure 3 This is a diagram of a simulated synaptic weight inhibition experiment of the artificial synaptic device of the present invention under continuous application of 10 negative voltage scanning cycles;
[0031] Figure 4 The characteristic diagram of simulated double pulse fractionation (PPF) of the artificial synaptic device of the present invention;
[0032] Figure 5 The long-term potentiation (LTP) and long-term depression (LTD) characteristics of the artificial synaptic device of the present invention are simulated.
[0033] In the figure: 1-substrate layer, 2-bottom electrode layer, 3-zinc oxide layer, 4-two-dimensional Nb2C layer; 5-top electrode layer. DETAILED DESCRIPTION
[0034] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0035] Example 1
[0036] like Figure 1 As shown, the artificial synaptic device based on Nb2C-zinc oxide memristor includes, from bottom to top, a substrate layer 1, a bottom electrode layer 2, a resistive switching functional layer and a top electrode layer 5; wherein the resistive switching functional layer is a heterostructure composed of a two-dimensional Nb2C layer 4 and a zinc oxide layer 3, and the resistive switching functional layer is prepared by covering the two-dimensional Nb2C layer on the zinc oxide layer by a spin coating or spray coating process;
[0037] Under external stimulation (including electrical signals of different amplitudes, numbers and frequencies), the artificial synaptic device can simulate a variety of biological synaptic characteristics, such as synaptic weight regulation, double pulse alienation, long-term potentiation, long-term depression, etc. It can be used to simulate brain-like neural synapses and realize neuromorphic computing. In addition, the heterostructure uses its unique structural characteristics to regulate the formation of conductive filaments and improve the performance of the device.
[0038] Nb2C is a new type of two-dimensional inorganic compound composed of transition metal carbon / nitride, which has fast carrier migration rate, high chemical stability, excellent photoelectric properties, good mechanical and electrical conductivity. Applying Nb2C materials to oxide memristors can effectively solve the existing problems of oxide memristors. The high conductivity of Nb2C can reduce the operating voltage and energy consumption of memristors; the fast carrier migration speed of Nb2C can increase the switching speed of memristors; its stable structure and high specific surface area can help improve the stability and reliability of memristors and extend their service life; moreover, Nb2C can also help realize multi-level storage and further improve storage density.
[0039] In this embodiment, the substrate layer 1 is made of silicon. In some embodiments, the substrate layer 1 may also be made of glass, silicon carbide, gallium nitride, etc.
[0040] In this embodiment, the material of the bottom electrode layer 2 is platinum. In some embodiments, the material of the bottom electrode layer 2 may also be gold or tungsten.
[0041] In this embodiment, the material of the top electrode layer 5 is silver. In some embodiments, the material of the top electrode layer 5 may also be copper or iron.
[0042] In this embodiment, the thickness of the top electrode layer 5 and the bottom electrode layer 2 are both 100 nm. In some embodiments, the thickness of the top electrode layer 5 and the bottom electrode layer 2 may be 80-100 nm.
[0043] Example 2
[0044] In this embodiment, a method for preparing an artificial synaptic device based on Nb2C-zinc oxide memristor is disclosed to prepare the artificial synaptic device in Example 1; the preparation steps include:
[0045] S1, preparation of bottom electrode layer 2:
[0046] The bottom electrode platinum is deposited on the substrate layer 1 by DC magnetron sputtering, and the thickness of the control electrode is 100 nanometers.
[0047] In this embodiment, the substrate layer 1 is a silicon wafer with a thickness of 500 microns. A platinum film with a thickness of 100 nanometers is plated on the silicon wafer by using a DC magnetron sputtering method. The entire platinum film serves as a bottom electrode.
[0048] S2, preparation of zinc oxide layer 3:
[0049] Depositing zinc oxide on the silicon wafer obtained in S1 by radio frequency magnetron sputtering;
[0050] In this embodiment, a zinc oxide layer 3 with a thickness of 60 nanometers is deposited on the platinum film by using a radio frequency magnetron sputtering method.
[0051] S3, preparing a mixed solution of Nb2C and dimethyl sulfoxide:
[0052] Weigh Nb2C powder and dimethyl sulfoxide solution into a beaker to prepare a 10 mg / ml Nb2C mixed solution, and stir magnetically to mix them evenly. Stir magnetically for 2 hours until the solution and Nb2C are fully mixed; In this embodiment, weigh 50 mg of Nb2C powder and 5 ml of dimethyl sulfoxide solution into a beaker to prepare a 10 mg / ml Nb2C mixed solution; The Nb2C used in this embodiment is Nb2C;
[0053] S4, preparation of two-dimensional Nb2C layer 4:
[0054] Place the silicon wafer coated with the platinum electrode and the zinc oxide layer 3 on a glue machine, set the glue machine speed to 1000 rpm, and set the time to 60 seconds. Use a dropper to absorb the Nb2C solution and evenly drop it on the zinc oxide layer 3, then start the glue machine, and after the machine is finished running, take out the silicon wafer and place it on the heating table, set the temperature to 80 degrees Celsius, and dry it for 20 minutes to obtain a two-dimensional Nb2C layer 4. In some embodiments, the two-dimensional Nb2C layer 4 can also be prepared by spraying;
[0055] S5, preparation of top electrode:
[0056] Silver electrodes are plated on the two-dimensional Nb2C layer, a mask is attached to the two-dimensional Nb2C layer, and the layer is placed in a magnetron sputtering chamber to plate a 100-nanometer silver film to obtain an artificial synaptic device based on Nb2C-zinc oxide memristor.
[0057] The pattern of the top electrode layer 5 is a rectangle of 450 micrometers×450 micrometers. The structure of the device is a silver top electrode, a two-dimensional Nb2C film, a zinc oxide film, and a platinum bottom electrode.
[0058] Example 3
[0059] In this embodiment, the electrical properties of the artificial synaptic device prepared in Example 2 were tested. Figure 1 As shown, during the test, the bottom electrode layer 2 is grounded and the top electrode layer 5 is connected to a positive voltage.
[0060] Figure 2 and Figure 3 The curve reflects the change of synaptic weight simulated by the artificial synaptic device. In which, the device was subjected to 10 consecutive forward voltage scans from 0 to 1.5V, and it was found that the current increased with the increase of the number of scans (e.g. Figure 2 As shown in the figure, the synaptic weight is simulated to increase with the number of scans. The device is subjected to 10 consecutive negative voltage scans from 0 to -1.7 V, and the current is found to decrease with the number of scans (as shown in the figure). Figure 3As shown in Figure 2, it simulates the depression of synaptic weight as the number of scans increases. Therefore, the device can well simulate the change of synaptic weight.
[0061] Figure 4 This reflects the artificial synaptic device's simulated double pulse factorization (PPF) characteristics. When two consecutive 1.5V pulses are applied to the device, the response current of the second pulse is larger than that of the first pulse, which simulates the PPF characteristics well.
[0062] Figure 5 It reflects the artificial synaptic device's simulation of long-term potentiation and long-term depression characteristics. First, 100 consecutive +2V pulses were applied to the device, and the device's conductance increased with the increase in the number of pulses, simulating the long-term potentiation characteristics of biological synapses; then 100 consecutive -2V pulses were applied to the device, and the device's conductance decreased with the increase in the number of pulses, simulating the long-term depression characteristics of biological synapses.
[0063] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.
[0064] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements all fall within the scope of the present invention to be protected.
Claims
1. Based on Nb 2 C-ZnO memristor artificial synaptic device, It is characterized in that From bottom to top, it includes: substrate layer, bottom electrode layer, resistive switching functional layer and top electrode layer; the resistive switching functional layer is a two-dimensional Nb 2 The heterogeneous structure composed of a C layer and a zinc oxide layer, wherein the resistive switching functional layer is formed by using a spin coating or spraying process to coat a two-dimensional Nb 2 The C layer is prepared by covering the zinc oxide layer; The Nb-based 2 The preparation process of the artificial synaptic device of C-zinc oxide memristor includes: S1, depositing a bottom electrode on the substrate layer by DC magnetron sputtering or electron beam evaporation to form a bottom electrode layer; S2, depositing zinc oxide on the bottom electrode layer by radio frequency magnetron sputtering to form a zinc oxide layer; S3, Nb 2 C powder and dimethyl sulfoxide solution were fully stirred and mixed to obtain Nb 2 C mixed solution; and Nb 2 C mixed solution is coated on the zinc oxide layer and then dried to form a two-dimensional Nb 2 C layer; S4, attach the mask plate to the two-dimensional Nb 2 On the C layer, DC magnetron sputtering was used to deposit two-dimensional Nb 2 A top electrode layer is deposited on the C layer.
2. The Nb-based 2 C-ZnO memristor artificial synaptic device, It is characterized in that The substrate layer is made of silicon, glass, silicon carbide or gallium nitride.
3. The Nb-based 2 C-ZnO memristor artificial synaptic device, It is characterized in that The bottom electrode layer is made of platinum, gold or tungsten.
4. The Nb-based 2 C-ZnO memristor artificial synaptic device, It is characterized in that The material of the top electrode layer is silver, copper or iron.
5. The Nb-based 2 C-ZnO memristor artificial synaptic device, It is characterized in that The thickness of the top electrode layer and the bottom electrode layer is 80-100 nm.
6. The Nb-based 2 C-ZnO memristor artificial synaptic device preparation method, It is characterized in that The following steps are involved: S1, depositing a bottom electrode on the substrate layer by DC magnetron sputtering or electron beam evaporation to form a bottom electrode layer; S2, depositing zinc oxide on the bottom electrode layer by radio frequency magnetron sputtering to form a zinc oxide layer; S3, Nb 2 C powder and dimethyl sulfoxide solution were fully stirred and mixed to obtain Nb 2 C mixed solution; and Nb 2 C mixed solution is coated on the zinc oxide layer and then dried to form a two-dimensional Nb 2 C layer; S4, attach the mask plate to the two-dimensional Nb 2 On the C layer, DC magnetron sputtering was used to deposit two-dimensional Nb 2 A top electrode layer is deposited on the C layer.
7. The Nb-based 2 C-ZnO memristor artificial synaptic device preparation method, It is characterized in that The top electrode layer pattern is a rectangle of 450 micrometers×450 micrometers.
8. The Nb-based 2 C-ZnO memristor artificial synaptic device preparation method, It is characterized in that The Nb 2 The concentration of the mixed solution C is 10 mg / ml.
9. An image recognition system, It is characterized in that The Nb-based 2 C-ZnO memristor artificial synaptic device.
