A flexible reconfigurable neuron device and a preparation method thereof

Abstract

The application discloses a flexible reconfigurable neuron device and a preparation method thereof. The flexible reconfigurable neuron device comprises a flexible substrate, a bottom electrode which is an inert metal and is formed on the flexible substrate, a two-dimensional semiconductor nanosheet layer which is formed on the bottom electrode, a high-k hafnium-based oxide dielectric layer which is formed on the two-dimensional semiconductor nanosheet layer, and a top electrode which is an active metal and is formed on the high-k hafnium-based oxide dielectric layer. By applying voltage pulses of different sizes on the top electrode, the process of metal ion migration and accumulation in the top electrode is changed, the strength of the conductive filament in the functional layer is controlled, the switching between the volatile neuron and the non-volatile synapse of the device is realized, and the reconfiguration function is realized in the same device.

Description

Technical Field

[0001] This invention relates to the field of semiconductor technology, and more specifically to a flexible reconfigurable neuron device and its fabrication method. Background Technology

[0002] The human brain is a highly efficient computing system, and neural networks play an indispensable role, successfully sparking a boom in neuromorphic computing. The basic unit in biological neural networks is the neuron, and neurons are connected by synapses. A neural network contains 10... 12 10 neurons and 10 15 Neural synapses play a crucial role in neuromorphic computation. Specifically, synapses generate stable and adjustable weights upon activation by action potentials, enabling high-performance neuromorphic computation. Neurons integrate and process action potential signals from previous levels to fire action potentials at the next level, which is essential for spiking neural network computation. To achieve highly efficient neuromorphic computation similar to that of the human brain, it is necessary to simulate biological synapses and neurons using electronic devices.

[0003] Neuromorphic electronic devices based on memristors have a natural advantage in simulating biological synapses and neurons, with their two ends corresponding to the front and back ends of a synapse, exhibiting structural similarity to biological systems. However, due to the different characteristics of biological synapses and neurons, different material systems and device structures are often required. Simulating biological synapses requires excellent storage performance and continuously adjustable conductance, necessitating sufficiently robust memristor filaments to achieve non-volatile conductance modulation. In contrast, simulating neurons requires volatile threshold transition behavior, requiring weaker connections in the memristor filaments to achieve instantaneous response. Electronic synapses and neurons based on different material systems and device structures face challenges such as current matching and heterogeneous chip integration, limiting the integration and development of neuromorphic computing chips. Constructing reconfigurable neuronal devices using the same device structure and material system can largely solve this problem, enabling multiple functions, improving chip integration, and saving costs, showing great application potential in next-generation low-power neuromorphic computing.

[0004] Flexible electronics offer advantages such as low cost, lightweight, portability, and bend resistance, making them ideal for wearable applications. However, due to the portability requirements of their operating environments, it's necessary to avoid complex circuits and excessive component combinations as much as possible. Implementing multiple functions within a single device can save chip space and improve the utilization and flexibility of individual devices. Therefore, developing reconfigurable flexible electronic devices has become an effective path and a new trend in the next generation of wearable technology. Summary of the Invention

[0005] This invention discloses a flexible reconfigurable neuron device, comprising: a flexible substrate; a bottom electrode, which is an inert metal, formed on the flexible substrate; a two-dimensional semiconductor nanosheet layer formed on the bottom electrode; a high-k hafnium-based oxide dielectric layer formed on the two-dimensional semiconductor nanosheet layer; and a top electrode, which is an active metal, formed on the high-k hafnium-based oxide dielectric layer. By applying voltage pulses of different magnitudes to the top electrode, the migration and accumulation process of metal ions in the top electrode is changed, and the strength of the conductive filaments in the functional layer is controlled, thereby realizing the switching between volatile neurons and non-volatile synapses in the device, and achieving reconfiguration function in the same device.

[0006] In the flexible reconfigurable neuron device of the present invention, preferably, the bottom electrode is Pd, Pt, Al, or Au.

[0007] In the flexible reconfigurable neuron device of the present invention, preferably, the two-dimensional semiconductor nanosheet layer is WS2, MoS2, ReS2, WSe2, MoSe2, ReSe2, or MoTe2.

[0008] In the flexible reconfigurable neuron device of the present invention, preferably, the high-K hafnium-based oxide dielectric layer is HfTaO. x HfAlO x HfSiO x HfTiO x .

[0009] In the flexible reconfigurable neuron device of the present invention, the top electrode is preferably Ag, Cu, or Ti.

[0010] This invention also discloses a method for fabricating a flexible reconfigurable neuronal device, comprising the following steps: forming an inert metal as a bottom electrode on a flexible substrate; forming a two-dimensional semiconductor nanosheet layer on the bottom electrode; forming a high-k hafnium-based oxide dielectric layer on the two-dimensional semiconductor nanosheet layer; forming an active metal as a top electrode on the high-k hafnium-based oxide dielectric layer; and by applying voltage pulses of different magnitudes to the top electrode, changing the migration and accumulation process of metal ions in the top electrode, controlling the strength of conductive filaments in the functional layer, and realizing the switching of the device between volatile neurons and non-volatile synapses, thereby achieving reconfiguration function in the same device.

[0011] In the method for fabricating the flexible reconfigurable neuron device of the present invention, preferably, the bottom electrode is Pd, Pt, Al, or Au.

[0012] In the method for fabricating the flexible reconfigurable neuron device of the present invention, preferably, the two-dimensional semiconductor nanosheet layer is WS2, MoS2, ReS2, WSe2, MoSe2, ReSe2, or MoTe2.

[0013] In the method for fabricating the flexible reconfigurable neuron device of the present invention, preferably, the high-k hafnium-based oxide dielectric layer is HfTaO. x HfAlO x HfSiO x HfTiO x .

[0014] In the method for fabricating the flexible reconfigurable neuron device of the present invention, the top electrode is preferably made of Ag, Cu, and Ti.

[0015] Beneficial effects:

[0016] (1) Using electronic devices to realize the working mode of the human brain, and using memristors to simulate the basic working units in the biological neural network to complete the neural network calculation, providing physical hardware for it, and having advantages such as high efficiency and low power consumption when processing a large amount of irregular data.

[0017] (2) Breaking the traditional working method of electronic devices, by controlling the strength of conductive filaments, the volatile and non-volatile characteristics can be switched in the same device unit to realize the dual simulation of neural synapse and neuron functions. It has obvious advantages in saving chip costs, improving chip utilization, and reducing the difficulty of integrating different functional units on the same chip.

[0018] (3) Flexible reconfigurable neuronal devices are fabricated using a heterostructure of two-dimensional semiconductor nanosheets and high-k hafnium-based oxides, making them more flexible and functionally adaptable. Using the same device structure not only reduces the complexity of the process but also improves the portability of the device, which has great potential for development in wearable applications. Attached Figure Description

[0019] Figure 1 This is a flowchart of a method for fabricating flexible reconfigurable neuronal devices.

[0020] Figures 2-5 This is a schematic diagram of the structure of each stage in the fabrication method of flexible reconfigurable neuronal devices.

[0021] Figure 6 This is a schematic diagram illustrating the operating principle of a flexible reconfigurable neuron device. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions in the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for explaining this invention and are not intended to limit this invention. The described embodiments are merely some embodiments of this invention, not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0023] In the description of this invention, it should be noted that the terms "upper," "lower," "vertical," and "horizontal," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0024] Furthermore, many specific details of the invention, such as the structure, materials, dimensions, processing techniques, and methods of the device, are described below to provide a clearer understanding of the invention. However, as those skilled in the art will understand, the invention may be implemented without adhering to these specific details. Unless specifically indicated below, various parts of the device may be made of materials known to those skilled in the art, or may employ materials with similar functionality developed in the future.

[0025] Figure 1 This is a flowchart of a method for fabricating flexible, reconfigurable neuronal devices. (For example...) Figure 1 As shown, the fabrication method of the flexible reconfigurable neuron device includes the following steps:

[0026] In step S1, a flexible substrate PEN100 is prepared for fabricating flexible reconfigurable neuronal devices. The flexible substrate can also be PET, PI, PDMS, photographic paper, etc.

[0027] In step S2, a Pd layer with a thickness of 50 nm to 200 nm is prepared on the flexible substrate 100 using physical vapor deposition as the bottom electrode 101. Figure 2 As shown. The bottom electrode material can also be inert metals such as Pt, Al, and Au.

[0028] In step S3, WS2 nanosheet solution is spin-coated onto the bottom electrode 101 at a spin speed of 2000 rpm to 4000 rpm for 30 to 120 seconds. Subsequently, it is baked on a hot plate at 60°C to 150°C for 5 to 60 minutes to complete the fabrication of the two-dimensional semiconductor nanosheet layer 102. Figure 3 As shown. Two-dimensional semiconductor nanosheets can also be two-dimensional semiconductor materials such as MoS2, ReS2, MoSe2, ReSe2, and MoTe2.

[0029] In step S4, HfTaO with a thickness of 10 nm to 30 nm is grown on the two-dimensional semiconductor nanosheet layer 102 using atomic layer deposition technology. x Thin film, as a high-k hafnium-based oxide dielectric layer 103, such as Figure 4 As shown. The material for the high-K hafnium-based oxide dielectric layer can also be HfAlO. x HfSiO x HfTiO x wait.

[0030] In step S5, multiple top electrodes 104 with a thickness of 50 nm to 150 nm and spaced apart from each other are fabricated on the high-k hafnium-based oxide dielectric layer 103 using physical vapor deposition, such as... Figure 5 As shown. The top electrode is preferably a square with a side length of 200 μm. The material of the top electrode is preferably an active metal such as Ag, Cu, or Ti.

[0031] like Figure 5 As shown, the flexible reconfigurable neuron device includes: a flexible substrate 100; a bottom electrode 101, which is an inert metal, formed on the flexible substrate 100; a two-dimensional semiconductor nanosheet layer 102 formed on the bottom electrode 101; a high-k hafnium-based oxide dielectric layer 103 formed on the two-dimensional semiconductor nanosheet layer 102; and a top electrode 104, which is an active metal, formed on the high-k hafnium-based oxide dielectric layer 103.

[0032] When a positive voltage is applied to the top electrode 104 of the device, the top electrode Ag104 is oxidized to Ag. + Gradually, the Ag atoms migrate towards the bottom electrode 101, reducing to Ag atoms and accumulating to form a conductive channel, gradually connecting the top electrode 104 and the bottom electrode 101. By applying voltage pulses of different magnitudes to the top electrode 104, the Ag atoms in the top electrode 104 are altered. + The process of ion migration and accumulation controls the strength of conductive filaments in the functional layer, enabling the device to switch between volatile neurons and non-volatile synapses, achieving reconfiguration functions within the same device, and improving the device's flexibility in different application scenarios.

[0033] Specifically, such as Figure 6 As shown, the switching between volatile neurons and non-volatile synapses:

[0034] Neuron-like pattern: When a small positive voltage is applied to the device, a weak conductive channel is formed between the top electrode 104 and the bottom electrode 101, producing a firing function similar to that of a neuron. After the voltage is removed, the conductive channel breaks, and the device returns to a quiescent state.

[0035] Neural synaptic pattern: When a large positive voltage is applied to the device, a strong conductive channel is formed between the top electrode 104 and the bottom electrode 101, generating a weighted function of biological neural synapses. The conductive channel remains stable after the voltage is removed, and the device has the ability to maintain this state over a long period.

[0036] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A flexible reconfigurable neuron device, characterized in that, include: Flexible substrate; The bottom electrode, which is an inert metal, is formed on the flexible substrate; A two-dimensional semiconductor nanosheet layer is formed on the bottom electrode; A high-k hafnium-based oxide dielectric layer is formed on the two-dimensional semiconductor nanosheet layer; The top electrode, which is an active metal, is formed on the high-k hafnium-based oxide dielectric layer. By applying voltage pulses of varying magnitudes to the top electrode, the migration and accumulation of metal ions within the top electrode are altered, thereby controlling the strength of the conductive filaments in the functional layer. This allows the device to switch between volatile neurons and non-volatile synapses, achieving reconfiguration functionality within the same device. The high-k hafnium-based oxide dielectric layer is HfTaO. x HfAlO x HfSiO x HfTiO x , The two-dimensional semiconductor nanosheets are WS2, MoS2, ReS2, WSe2, MoSe2, ReSe2, and MoTe2.

2. The flexible reconfigurable neuron device according to claim 1, characterized in that, The bottom electrode is Pd, Pt, Al, or Au.

3. The flexible reconfigurable neuron device according to claim 1, characterized in that, The top electrode is composed of Ag, Cu, and Ti.

4. A method for fabricating a flexible reconfigurable neuron device, characterized in that, Includes the following steps: An inert metal is formed on a flexible substrate as the bottom electrode; A two-dimensional semiconductor nanosheet layer is formed on the bottom electrode; A high-k hafnium-based oxide dielectric layer is formed on the two-dimensional semiconductor nanosheet layer; An active metal is formed on the high-k hafnium-based oxide dielectric layer as the top electrode. By applying voltage pulses of varying magnitudes to the top electrode, the migration and accumulation of metal ions within the top electrode are altered, thereby controlling the strength of the conductive filaments in the functional layer. This allows the device to switch between volatile neurons and non-volatile synapses, achieving reconfiguration functionality within the same device. The high-k hafnium-based oxide dielectric layer is HfTaO. x HfAlO x HfSiO x HfTiO x , The two-dimensional semiconductor nanosheets are WS2, MoS2, ReS2, WSe2, MoSe2, ReSe2, and MoTe2.

5. The method for fabricating a flexible reconfigurable neuron device according to claim 4, characterized in that, The bottom electrode is Pd, Pt, Al, or Au.

6. The method for fabricating a flexible reconfigurable neuron device according to claim 4, characterized in that, The top electrode is composed of Ag, Cu, and Ti.

Images