Artificial synapse device based on two-dimensional black phosphorus / tin sulfide heterojunction

Through artificial synaptic devices based on two-dimensional black phosphorus/tin sulfide heterojunction, the adsorption-desorption effect of black phosphorus surface state and gate voltage regulation are used to build a neuromorphic network, which solves the low energy efficiency of existing artificial intelligence chips and the low energy efficiency of existing artificial intelligence chips and The complex structure of the problem realizes low-energy consumption neuromorphic computing and high-integrated chip design.

CN119170651APending Publication Date: 2024-12-20SHENZHEN POLYTECHNIC
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411234254.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-03
Publication Date
2024-12-20

AI Technical Summary

Technical Problem

The existing artificial intelligence chips are based on the von Neumann architecture, resulting in low energy efficiency, complex structure, and lack of high parallelism and low power consumption characteristics that simulate the human brain, making it difficult to achieve low energy consumption neuromorphic computing.

Method used

Using artificial synaptic devices based on two-dimensional black phosphorus/tin sulfide heterojunction, the adsorption-desorption effect of the black phosphorus surface state on carriers is used, and the long-term excitation and inhibition of heterojunction is regulated through gate voltage. Characteristics, build a neuromorphic network to achieve linear improvement in synaptic plasticity.

Benefits of technology

It reduces the energy consumption of artificial intelligence computing, simplifies synaptic weight training of neural networks, and is easy to achieve high-integration small-size, low-energy chip design due to the atomic thickness and flat surface of two-dimensional materials.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119170651A_ABST
    Figure CN119170651A_ABST
Patent Text Reader

Abstract

The invention discloses an artificial synapse device based on a two-dimensional black phosphorus / tin sulfide heterojunction, the artificial synapse device comprises a black phosphorus two-dimensional film, a tin sulfide two-dimensional film, a thermal silicon oxide substrate and an electrode, the thermal silicon oxide substrate comprises a silicon dioxide layer and a silicon material layer, the lower surface of the black phosphorus two-dimensional film is in contact with the silicon dioxide layer of the thermal silicon oxide substrate, and the lower surface of the tin sulfide two-dimensional film is in contact with the silicon material layer of the thermal silicon oxide substrate. One part of the tin sulfide two-dimensional thin film is stacked on the black phosphorus two-dimensional thin film, the rest part of the tin sulfide two-dimensional thin film is in contact with a silicon dioxide layer on the surface of the thermal silicon oxide substrate, and the electrodes are divided into a source electrode, a drain electrode and a grid electrode. According to the artificial synaptic device based on the two-dimensional black phosphorus / tin sulfide heterojunction, the synaptic plasticity of the artificial synaptic device is derived from the adsorption-desorption effect of a black phosphorus surface state on carriers, and meanwhile, the linearity of long-time-history excitation and inhibition characteristics of the heterojunction artificial synaptic device is improved through grid voltage; and a neural morphological network is constructed, so that the purpose of reducing the energy consumption of artificial intelligence calculation is achieved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of artificial synaptic devices, and in particular to an artificial synaptic device based on a two-dimensional black phosphorus / tin sulfide heterojunction. Background Art

[0002] In order to reduce computing energy consumption and promote the development of artificial intelligence technology, artificial intelligence chips based on new materials and new physical mechanisms are gradually being used in large quantities in the field of artificial intelligence technology. However, the current mainstream artificial intelligence chips mainly use the von Neumann architecture, which separates the storage path and processing path of data, making this type of artificial intelligence chip face problems such as low energy efficiency and complex structure.

[0003] Considering that neuromorphic chips that simulate the working principle of the human brain have the characteristics of high parallelism, high fault tolerance, storage and computing fusion, and low power consumption, and synapses are the core of information transmission and processing, the development of new artificial synaptic devices is of great significance to the development of artificial intelligence chips. The present invention proposes to use an artificial neural synaptic device based on a two-dimensional material heterojunction, that is, to use the third type band structure of the two-dimensional material heterojunction to regulate the carrier transport characteristics, and further improve the linearity of the device synaptic plasticity. In this way, low-energy neuromorphic computing can be achieved.

[0004] In addition, because the two-dimensional material used has atomic-level thickness, a flat surface and no dangling bonds, the present invention is conducive to realizing a highly integrated, small-size, low-energy artificial intelligence chip. Summary of the invention

[0005] The purpose of the present invention is to provide an artificial synaptic device based on a two-dimensional black phosphorus / tin sulfide heterojunction, the synaptic plasticity of which originates from the adsorption-desorption effect of black phosphorus surface states on carriers. At the same time, the linearity of the long-term excitation and inhibition characteristics of the heterojunction artificial synaptic device is improved by the gate voltage, and a neuromorphic network is constructed to achieve the purpose of reducing the energy consumption of artificial intelligence computing.

[0006] To achieve the above-mentioned purpose, the present invention provides an artificial synaptic device based on a two-dimensional black phosphorus / tin sulfide heterojunction, comprising a black phosphorus two-dimensional film, a tin sulfide two-dimensional film, a thermally oxidized silicon substrate, and electrodes, wherein the thermally oxidized silicon substrate comprises a silicon dioxide layer and a silicon material layer stacked up and down, the lower surface of the black phosphorus two-dimensional film is in contact with the silicon dioxide layer of the thermally oxidized silicon substrate, the tin sulfide two-dimensional film is partially stacked on the black phosphorus two-dimensional film, and the remaining part of the tin sulfide two-dimensional film is in contact with the silicon dioxide layer on the surface of the thermally oxidized silicon substrate to form a black phosphorus / tin sulfide heterojunction, and the electrode is divided into a source electrode, a drain electrode, and a gate electrode.

[0007] Preferably, the gate is connected to a silicon material layer on the back of a thermally oxidized silicon substrate, and the source electrode is grounded.

[0008] Preferably, the drain electrode and the source electrode are connected to the ends of the black phosphorus two-dimensional film and the tin sulfide two-dimensional film, respectively.

[0009] Preferably, the silicon dioxide oxidizes the lower surface layer of the black phosphorus two-dimensional film in direct contact with the silicon dioxide, and the generated phosphorus oxide surface state provides a material basis for the synaptic device to simulate synaptic characteristics.

[0010] Preferably, the loop composed of the gate and source electrodes applies a square wave voltage to simulate a presynaptic pulse voltage signal, and the square wave voltage generates an electric field through the silicon dioxide dielectric layer to act on the black phosphorus / tin sulfide heterojunction to generate a pulse current signal.

[0011] Preferably, the current magnitude collected by the source electrode and drain electrode loop through the black phosphorus / tin sulfide heterojunction is a postsynaptic pulse current signal.

[0012] The advantages and positive effects of the artificial synaptic device based on the two-dimensional black phosphorus / tin sulfide heterojunction described in the present invention are:

[0013] 1. The artificial synaptic device of the present invention reduces the energy consumption of artificial intelligence computing and saves energy; it is easy to perform synaptic weight training of neural networks.

[0014] 2. The black phosphorus and tin sulfide two-dimensional materials of the present invention have no dangling bonds on their surfaces, and there is no difficulty in integration due to lattice mismatch. The two-dimensional materials themselves have atomic-level thickness, making it easy to manufacture nano-sized artificial intelligence chips.

[0015] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 Graphs showing the plasticity test results of the synaptic device of the present invention, wherein (a) is the synaptic plasticity of black phosphorus, (b) is the synaptic plasticity of the black phosphorus / tin sulfide heterojunction, (c) is the relationship between the nonlinearity of the black phosphorus and the black phosphorus / tin sulfide heterojunction in the long-term excitation characteristic region and the gate voltage, and (d) is the relationship between the nonlinearity of the black phosphorus and the black phosphorus / tin sulfide heterojunction in the long-term inhibition characteristic region and the gate voltage;

[0017] Figure 2 This is a schematic diagram of the structure of the artificial synapse device of the present invention;

[0018] Figure 3 : is a cross-sectional view of the crystal structure of black phosphorus (BP) of the present invention, and the middle of the red dotted line is the phosphorus oxide layer;

[0019] Figure 4 This is a simulation curve diagram of the synaptic excitation characteristics of black phosphorus / tin sulfide of the present invention. DETAILED DESCRIPTION

[0020] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0021] Unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0022] Example

[0023] An artificial synaptic device based on a two-dimensional black phosphorus / tin sulfide heterojunction, which is composed of a black phosphorus two-dimensional film, a tin sulfide two-dimensional film, a thermally oxidized silicon substrate, and electrodes; the thermally oxidized silicon substrate includes a silicon dioxide layer and a silicon material layer stacked up and down, the lower surface of the black phosphorus two-dimensional film is in contact with the silicon dioxide layer of the thermally oxidized silicon substrate, and the silicon dioxide is used to oxidize the lower surface of the black phosphorus two-dimensional film in direct contact with it, and the generated phosphorus oxide surface state provides a material basis for the device to simulate synaptic characteristics; the tin sulfide two-dimensional film is partially stacked on the black phosphorus two-dimensional film, and the remaining part is in contact with the silicon oxide layer on the surface of the thermally oxidized silicon substrate , forming a black phosphorus / tin sulfide heterojunction; the electrodes are divided into a source electrode, a drain electrode and a gate, the drain electrode and the source electrode are respectively connected to the ends of the black phosphorus two-dimensional film and the tin sulfide two-dimensional film, and the gate is connected to the silicon material on the back of the thermally oxidized silicon substrate; the source electrode is grounded; a square wave voltage is applied through a loop composed of the gate and the source electrode as a presynaptic pulse voltage signal, and the electric field generated by the square wave voltage through the silicon oxide dielectric layer acts on the black phosphorus / tin sulfide heterojunction to generate a pulse current signal, and the current passing through the black phosphorus / tin sulfide heterojunction is collected through the drain electrode and the source electrode loop as a postsynaptic pulse current signal.

[0024] A thermally oxidized silicon substrate with a 300-nanometer-thick silicon dioxide layer is used; the presynaptic pulse voltage is applied by the gate, and then the postsynaptic pulse current is obtained through the source electrode and the drain electrode. The gate is the Si layer at the bottom of the thermally oxidized silicon substrate, the drain electrode is connected to the black phosphorus material, and the source electrode is connected to the tin sulfide material. Parameters for testing synaptic plasticity: V ds =0.1V, pulse width is 10ms, pulse interval is 250ms.

[0025] like Figure 1 As shown, Figure 1 (a) shows the synaptic plasticity of black phosphorus. Figure 1 (b) shows the synaptic plasticity of black phosphorus / tin sulfide heterojunction. Figure 1 (c) shows the relationship between the nonlinearity of black phosphorus and black phosphorus / tin sulfide heterojunction in the long-term excitation characteristic region and the gate voltage. Figure 1 (d) shows the relationship between the nonlinearity of black phosphorus and black phosphorus / tin sulfide heterojunction in the long-term inhibition characteristic region and the gate voltage. Both artificial synaptic devices based on black phosphorus and black phosphorus / tin sulfide heterojunctions exhibit long-term synaptic excitation and inhibition characteristics, and in the long-term inhibition characteristic region, the nonlinearity of the black phosphorus / tin sulfide heterojunction device is significantly lower than that of the black phosphorus device, which is conducive to the synaptic weight training of neural networks.

[0026] like Figure 2 As shown, the core of the artificial synaptic device consists of a black phosphorus / tin sulfide heterojunction stacked on top of each other; the black phosphorus film is below the tin sulfide film. The black phosphorus film is in direct contact with the silicon dioxide layer, producing an amorphous layer of phosphorus oxide.

[0027] like Figure 3 As shown, Figure 3 This is a cross-sectional diagram of the crystal structure of black phosphorus (BP), with the red dotted line showing the phosphorus oxide layer.

[0028] like Figure 4 As shown in the figure, the synaptic excitation characteristics when a pulse voltage with different time widths and an amplitude of 60 volts is applied through the gate can be found that as the pulse time width increases, the more carriers adsorbed by the phosphorus oxide layer, the larger the postsynaptic current, and the longer the postsynaptic current is maintained.

[0029] Therefore, the present invention adopts the above-mentioned artificial synaptic device based on the two-dimensional black phosphorus / tin sulfide heterojunction. The synaptic plasticity of the artificial synaptic device originates from the adsorption-desorption effect of the black phosphorus surface state on carriers. At the same time, the symmetry of the long-term excitation and inhibition characteristics of the heterojunction artificial synaptic device is improved by the gate voltage to construct a neuromorphic network to achieve the purpose of reducing the energy consumption of artificial intelligence computing.

[0030] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. An artificial synaptic device based on a two-dimensional black phosphorus / tin sulfide heterojunction, Features: It includes a black phosphorus two-dimensional film, a tin sulfide two-dimensional film, a thermally oxidized silicon substrate, and an electrode. The thermally oxidized silicon substrate includes a silicon dioxide layer and a silicon material layer stacked up and down. The lower surface of the black phosphorus two-dimensional film is in contact with the silicon dioxide layer of the thermally oxidized silicon substrate. A part of the tin sulfide two-dimensional film is stacked on the black phosphorus two-dimensional film, and the remaining part of the tin sulfide two-dimensional film is in contact with the silicon dioxide layer on the surface of the thermally oxidized silicon substrate to form a black phosphorus / tin sulfide heterojunction. The electrode is divided into a source electrode, a drain electrode, and a gate electrode.

2. An artificial synaptic device based on a two-dimensional black phosphorus / tin sulfide heterojunction according to claim 1, Features: The gate electrode is connected to the silicon material layer on the back of the thermally oxidized silicon substrate, and the source electrode is grounded.

3. An artificial synaptic device based on a two-dimensional black phosphorus / tin sulfide heterojunction according to claim 1, Features: The drain electrode and the source electrode are respectively connected to the ends of the black phosphorus two-dimensional film and the tin sulfide two-dimensional film.

4. An artificial synaptic device based on a two-dimensional black phosphorus / tin sulfide heterojunction according to claim 1, Features: The silicon dioxide oxidizes the lower surface layer of the black phosphorus two-dimensional film in direct contact with the silicon dioxide, and the generated phosphorus oxide surface state provides a material basis for the synaptic device to simulate synaptic characteristics.

5. An artificial synaptic device based on a two-dimensional black phosphorus / tin sulfide heterojunction according to claim 1, Features: The circuit composed of the gate and source electrodes applies a square wave voltage as a presynaptic pulse voltage signal. The square wave voltage generates an electric field through the silicon dioxide dielectric layer and acts on the black phosphorus / tin sulfide heterojunction to generate a pulse current signal.

6. An artificial synaptic device based on a two-dimensional black phosphorus / tin sulfide heterojunction according to claim 1, Features: The source electrode and drain electrode loop collect the current passing through the black phosphorus / tin sulfide heterojunction as a postsynaptic pulse current signal.