A temperature-aware artificial neuron system based on chalcogenide threshold switching devices

Abstract

The application provides a temperature sensing artificial neuron system based on a chalcogenide threshold switching device, and belongs to the technical field of artificial neural morphology, and comprises a resistor, a chalcogenide threshold switching device and a capacitor; the chalcogenide threshold switching device is used for realizing switching between a high-resistance state and a low-resistance state under different excitations; wherein, the threshold voltage of the chalcogenide threshold switching device decreases with the increase of temperature, and then the amplitude of a pulse signal decreases and the frequency of periodic oscillation increases; the change of the threshold voltage is taken as a response to perceive a temperature signal; an excitation input end is used for providing excitation for the capacitor, controlling the capacitor to discharge through the chalcogenide threshold switching device, or controlling the capacitor to charge through the resistor. The temperature sensing artificial neuron system in the application can convert a temperature signal into a pulse signal, does not need to additionally introduce a temperature sensor, has a compact structure, a simple circuit structure, small leakage current and low power consumption, and is conducive to large-scale integration.

Description

Technical Field

[0001] This invention belongs to the field of artificial neuromorphic technology, and more specifically, relates to a temperature-sensing artificial neuron system based on a chalcogenide threshold switching device. Background Technology

[0002] Biologically inspired artificial neuromorphic computing is considered an effective means to overcome the von Neumann bottleneck. Through the cross-connection of artificial neurons and synapses, it is possible to achieve brain-like computing functions with high energy efficiency, low latency, and high parallelism. Traditional artificial neurons are mainly built based on CMOS circuits, which have complex structures, poor miniaturization, and are not conducive to large-scale integration. In recent years, emerging semiconductor two-terminal devices have great advantages in simulating the functions of biological neurons and building artificial neural network hardware due to their compact structure and simple process.

[0003] Sensory artificial neurons are artificial neurons that convert external analog signals into pulse signals, and are widely used in fields such as electronic skin, bioelectronic interfaces, and smart sensors. Temperature sensing is an important function in the biological sensory nervous system, achieved through the joint efforts of temperature receptors, sensory neurons, and the central nervous system.

[0004] Currently, mainstream temperature-sensing artificial neural network systems are generally composed of electronic components such as resistors, capacitors, and threshold switching devices. Among the many semiconductor devices with threshold switching behavior, metal-insulator transition (MIT) devices are the most commonly used. However, their large milliampere-level leakage current leads to high overall power consumption (10–1000 nJ). They also have the problems of a small applicable temperature range (< metal-insulator transition temperature, typically 70°C) and poor temperature response. Neural network systems built using MIT devices usually require an additional temperature sensor in series, which further reduces the integration density.

[0005] Therefore, finding a threshold switching device with low leakage current and easy fabrication to build a compact, ultra-low power temperature-sensing artificial neuron system has become an urgent problem to be solved. Summary of the Invention

[0006] To address the shortcomings of existing technologies, the present invention aims to provide a temperature-sensing artificial neuron system based on an oxononic threshold switch (OTS). Utilizing the characteristics of Ovonic Threshold Switches (OTS)—low leakage current, ease of fabrication, sensitivity to temperature changes, and wide applicable temperature range—a compact, low-power temperature-sensing artificial neuron system can be constructed. This aims to solve the problems of high leakage current, narrow temperature range, and poor temperature response in existing temperature-sensing artificial neuron systems.

[0007] To achieve the above objectives, the present invention provides a temperature-sensing artificial neuron system based on a chalcogenide threshold switching device, comprising: a resistor, a chalcogenide threshold switching device, and a capacitor;

[0008] The resistor and the chalcogenide threshold switching device are connected in series, and the capacitor and the chalcogenide threshold switching device are connected in parallel; the resistance value of the resistor is greater than the resistance value of the chalcogenide threshold switching device in the low resistance state and less than the resistance value of the chalcogenide threshold switching device in the high resistance state; the resistor and the capacitor form a charging circuit; the chalcogenide threshold switching device and the capacitor form a discharging circuit; the two ends of the resistor are the excitation input terminal and the pulse signal output terminal, respectively.

[0009] A chalcogenide threshold switching device is used to switch between a high-resistivity state and a low-resistivity state under different excitations. Specifically, based on the characteristic that the threshold voltage of the chalcogenide threshold switching device decreases with increasing temperature, thereby reducing the amplitude of the pulse signal and increasing the frequency of periodic oscillation, the temperature signal is sensed by using the change in threshold voltage as a response.

[0010] The excitation input is used to provide excitation to the capacitor, control the capacitor to discharge through the chalcogenide threshold switching device, or control the capacitor to charge through the resistor.

[0011] More specifically, the temperature-sensing artificial neuron system converts temperature signals into pulse signals; and the temperature-sensing artificial neuron system will only output pulse signals when the voltage across the capacitor reaches the threshold voltage of the chalcogenide threshold switching device.

[0012] More preferably, under voltage / current excitation, when the voltage across the capacitor is greater than the threshold voltage of the chalcogenide threshold switching device, the chalcogenide threshold switching device is in a low-resistance state, and the capacitor discharges through the chalcogenide threshold switching device; when the voltage across the capacitor is less than the holding voltage of the chalcogenide threshold switching device, the chalcogenide threshold switching device changes to a high-resistance state, and the capacitor is charged through the resistor.

[0013] More preferably, the sulfide threshold transition device includes a first electrode layer, a second electrode layer, and a functional layer located between the first electrode layer and the second electrode layer;

[0014] The first electrode layer and the second electrode layer are electrodes provided for applying excitation voltage / current to the functional layer. The first electrode layer is used to connect one end of the resistor and the capacitor, and the second electrode layer is used to connect the other end of the capacitor and ground. The functional layer is used to switch between a high-resistance state and a low-resistance state under different voltage / current excitations. Based on the characteristic that the threshold voltage of the functional layer decreases with increasing temperature, thereby reducing the amplitude of the pulse signal and increasing the periodic oscillation frequency, the temperature signal is sensed by using the change in the threshold voltage as a response.

[0015] More preferably, the functional layer of the sulfide threshold switching device adopts a sulfide-based gating material, which is a tellurium-based OTS material doped with one or more elements; or a selenium-based OTS material doped with one or more elements; or a sulfide-based OTS material doped with one or more elements.

[0016] More preferably, both the first electrode layer and the second electrode layer are made of inert electrode materials; wherein, the inert electrode material is one or more of Pt, W, Au, Ru, Al, TiN, TiW, Ta, TaN, IrO2, ITO and IZO.

[0017] More preferably, the capacitor is a fixed capacitor or a variable capacitor with a capacitance value of 1fF to 1μF.

[0018] More preferably, the thickness of the first electrode layer and the second electrode layer is 5 nm to 200 nm; the thickness of the functional layer is 3 nm to 200 nm.

[0019] More preferably, the sensing temperature range of the sulfide threshold switching device is lower than the crystallization temperature of the sulfide gate material.

[0020] More preferably, the sulfide threshold switching device has a nanoampere level leakage current.

[0021] In summary, compared with the prior art, the above-described technical solutions conceived by this invention have the following advantages:

[0022] Beneficial effects:

[0023] This invention provides a temperature-sensing artificial neuron system based on a chalcogenide threshold switching device, comprising a chalcogenide threshold switching device, a resistor, and a capacitor. The chalcogenide threshold switching device has low leakage current (nanoampere level) and is easy to fabricate, which helps reduce the overall power consumption of the neuron system. At the same time, the chalcogenide threshold switching device itself is sensitive to temperature changes and has a wide applicable temperature range (below the crystallization temperature of chalcogenide gating materials, which are generally above 100°C). Therefore, the temperature-sensing artificial neuron system does not require an additional temperature sensor, has a compact structure, and a simple circuit structure, which is conducive to large-scale integration. Attached Figure Description

[0024] Figure 1 This is a voltage-current scan curve of an N-Si2Te3 chalcogenide threshold transition device provided in an embodiment of the present invention;

[0025] Figure 2 This is a voltage-current scan curve of an N-Si2Te3 chalcogenide threshold transition device at different temperatures, provided in an embodiment of the present invention.

[0026] Figure 3This is a schematic diagram of the structure of a temperature-sensing artificial neuron system provided in an embodiment of the present invention;

[0027] Figure 4 This is a response diagram of a temperature-sensing artificial neuron system based on an N-Si2Te3 chalcogenide threshold transition device provided in an embodiment of the present invention under different temperatures and 4V voltage excitation inputs. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0029] To meet the need for developing compact, low-power temperature-sensing artificial neuron systems, this invention provides a temperature-sensing artificial neuron system based on a chalcogenide threshold switching device, comprising: a resistor R S 1. Chalcogenide threshold transition device TS and capacitor C; resistor R S A sulfide threshold switching device TS is connected in series, and a capacitor C is connected in parallel with the sulfide threshold switching device TS; wherein, the sulfide threshold switching device is used to sense the temperature signal and responds with a change in the threshold voltage; resistor R S The capacitor C forms a charging circuit; the sulfide threshold conversion device TS and the capacitor C form a discharging circuit; the resistor R... S The two ends are the excitation input and pulse signal output of the temperature-sensing artificial neuron system, respectively.

[0030] In some embodiments, the chalcogenide threshold switching device can achieve an instantaneous switching from a high-resistivity state to a low-resistivity state; specifically, when the voltage applied to the chalcogenide threshold switching device increases from zero to the threshold voltage V... th During the process, it is in a high-resistivity state; when the voltage applied to the sulfide threshold transition device exceeds the threshold voltage V... th It then enters a low-resistance state; when the voltage decreases and is less than the holding voltage V h Afterwards, the device spontaneously returns to a high-impedance state, wherein the threshold voltage V th Greater than the holding voltage V h .

[0031] In some embodiments, the sulfide threshold transition device includes: a first electrode layer, a second electrode layer, and a functional layer located between the first electrode layer and the second electrode layer;

[0032] In some embodiments, the first electrode layer and the second electrode layer of the sulfide threshold transition device are both made of inert electrode materials, which are one or more of Pt, W, Au, Ru, Al, TiN, TiW, Ta, TaN, IrO2, ITO and IZO, and the thickness of the first electrode layer and the second electrode layer is from 5 nm to 200 nm.

[0033] In some embodiments, the functional layer of the sulfide threshold switching device is made of a sulfide-based gate material, which is a tellurium-based OTS material doped with single or multiple elements such as C-Te, Si-Te, Ge-Te, Al-Te, and In-Te; or a selenium-based OTS material doped with single or multiple elements such as Ge-Se and Ge-As-Se; or a sulfide-based OTS material doped with single or multiple elements such as Ge-S and Ga-S, and the thickness of the functional layer is from 3 nm to 200 nm.

[0034] In some embodiments, resistor R S The resistance value is greater than the resistance value of the sulfide threshold transition device in the low-resistance state and less than the resistance value of the sulfide threshold transition device in the high-resistance state, so as to achieve a suitable voltage division; the capacitor C is a fixed capacitor or a variable capacitor, and its capacitance value is from 1fF to 1μF.

[0035] In some embodiments, the temperature-sensing artificial neuron system can convert temperature signals into pulse signals; and the temperature-sensing artificial neuron system will only output pulse signals when the voltage across capacitor C reaches the threshold voltage of the chalcogenide threshold switching device.

[0036] In some embodiments, the excitation input is voltage excitation or current excitation, the pulse signal is output in the form of periodic oscillation, and the frequency of the periodic oscillation increases with the increase of the sensed temperature within a certain range.

[0037] The above technical solution will be described in detail below with reference to specific embodiments.

[0038] Example 1

[0039] This embodiment provides a sulfide threshold switching device, which includes a first electrode layer, a second electrode layer, and a functional layer located between the first electrode layer and the second electrode layer.

[0040] Specifically, the functional layer adopts a sulfide-based gated material, which is a tellurium-based OTS material doped with single or multiple elements such as C-Te, Si-Te, Ge-Te, Al-Te, and In-Te; or a selenium-based OTS material doped with single or multiple elements such as Ge-Se and Ge-As-Se; or a sulfide-based OTS material doped with single or multiple elements such as Ge-S and Ga-S, and the thickness of the functional layer is from 3 nm to 200 nm.

[0041] Preferably, in this embodiment, the thickness of the functional layer is 50nm; naturally, in other embodiments, the thickness of the functional layer can also be 10nm, 20nm, 30nm, 40nm, 60nm, 70nm, 80nm, 90nm, 100nm or 110nm, etc.

[0042] Specifically, both the first electrode layer and the second electrode layer are made of inert electrode materials, which are one or more of Pt, W, Au, Ru, Al, TiN, TiW, Ta, TaN, IrO2, ITO and IZO, and the thickness of the first electrode layer and the second electrode layer is from 5 nm to 200 nm.

[0043] Preferably, in this embodiment, both the first electrode layer and the second electrode layer are W electrodes, and both have a thickness of 100 nm.

[0044] Figure 1 A voltage-current scan curve of an N-Si2Te3 chalcogenide threshold switching device is provided. The results show that the chalcogenide threshold switching device has an extremely low nanoampere leakage current (5.5 nA), which is beneficial for building an ultra-low power temperature-sensing artificial neuron system.

[0045] Figure 2 This invention provides voltage-current scan curves of an N-Si2Te3 chalcogenide threshold transition device at different temperatures. The threshold voltage V is shown between 0 and 100°C. th The decrease with increasing temperature indicates that the chalcogenide threshold switching device is temperature-sensitive and has a wide applicable temperature range, possessing the ability to sense temperature changes in situ. The temperature-sensing artificial neuron system constructed using this chalcogenide threshold switching device does not require the introduction of an additional temperature sensor, has a compact structure, and is conducive to large-scale integration.

[0046] Example 2

[0047] This embodiment provides a temperature-sensing artificial neuron system based on a chalcogenide threshold transition device. Figure 3 This is a schematic diagram of the structure of a temperature-sensing artificial neuron system, including: resistor R S 1. Chalcogenide threshold switching device TS and capacitor C;

[0048] Specifically, the chalcogenide threshold switching device is used to sense temperature signals and responds with changes in threshold voltage; resistor R S The capacitor C forms a charging circuit; the sulfide threshold conversion device TS and the capacitor C form a discharging circuit; the resistor R... S The two ends are the excitation input and pulse signal output of the temperature-sensing artificial neuron system, respectively.

[0049] Specifically, the resistance R SThe resistance value is greater than the resistance value of the sulfide threshold transition device in the low-resistance state and less than the resistance value of the sulfide threshold transition device in the high-resistance state, so as to achieve a suitable voltage division; the capacitor C is a fixed capacitor or a variable capacitor, and its capacitance value is from 1fF to 1μF.

[0050] Specifically, under a certain voltage excitation input, the temperature-sensing artificial neuron system will respond when the voltage across capacitor C exceeds the threshold voltage V of the chalcogenide threshold switching device. th When the chalcogenide threshold switching device transitions to a low-resistance state, capacitor C discharges through the chalcogenide threshold switching device; when the voltage across capacitor C is less than the holding voltage V of the chalcogenide threshold switching device... h At this time, the chalcogenide threshold transition device switches to a high-resistivity state, and the capacitor C flows through the resistor R. S The device is recharged; this process is repeated to output a periodically oscillating pulse signal.

[0051] Specifically, under a certain voltage excitation input, the threshold voltage V of the chalcogenide threshold switching device in the temperature-sensing artificial neuron system... th The amplitude of the output pulse signal decreases as the temperature increases, thus the frequency of the periodic oscillation increases; the temperature-sensing artificial neuron system thus realizes the function of converting the temperature signal into a pulse signal.

[0052] Figure 4 This is a response diagram of a temperature-sensing artificial neuron system based on an N-Si2Te3 chalcogenide threshold transition device under different temperatures and 4V voltage excitation inputs. Under different temperatures and 4V voltage excitation inputs, the temperature-sensing artificial neuron system outputs periodically oscillating pulse signals. The frequency of the periodic oscillation increases with the increase of the sensed temperature within a certain range, realizing the function of converting temperature signals into pulse signals.

[0053] This invention provides a temperature-sensing artificial neuron system based on a chalcogenide threshold switching device, constructed from the chalcogenide threshold switching device, resistors, and capacitors. The chalcogenide threshold switching device has low leakage current and is easy to fabricate, which helps reduce the overall power consumption of the neuron. Simultaneously, the chalcogenide threshold switching device is sensitive to temperature changes and has a wide applicable temperature range; therefore, the temperature-sensing artificial neuron system does not require an additional temperature sensor, has a compact structure, and a simple circuit structure, which is beneficial for large-scale integration.

[0054] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A temperature-sensing artificial neuron system based on a chalcogenide threshold transition device, characterized in that, include: Resistors, chalcogenide threshold switching devices, and capacitors; The resistor and the sulfide threshold switching device are connected in series, and the capacitor and the sulfide threshold switching device are connected in parallel; the resistance value of the resistor is greater than the resistance value of the sulfide threshold switching device in the low-resistance state and less than the resistance value of the sulfide threshold switching device in the high-resistance state; the resistor and the capacitor form a charging circuit; the sulfide threshold switching device and the capacitor form a discharging circuit; the two ends of the resistor are respectively the excitation input terminal and the pulse signal output terminal. The chalcogenide threshold switching device is used to switch between a high-resistivity state and a low-resistivity state under different excitations; wherein, based on the characteristic that the threshold voltage of the chalcogenide threshold switching device decreases with increasing temperature, thereby reducing the amplitude of the pulse signal and increasing the frequency of periodic oscillation, the change in threshold voltage is used as a response to sense the temperature signal. The excitation input terminal is used to provide excitation to the capacitor, control the capacitor to discharge through the chalcogenide threshold switching device, or control the capacitor to charge through the resistor; The functional layer of the sulfide threshold switching device adopts a sulfide-based gating material, which is a tellurium-based OTS material doped with one or more elements, or a selenium-based OTS material doped with one or more elements, or a sulfide-based OTS material doped with one or more elements. The chalcogenide threshold transition device includes a first electrode layer, a second electrode layer, and a functional layer located between the first electrode layer and the second electrode layer; The first electrode layer and the second electrode layer are electrodes provided for applying excitation voltage / current to the functional layer. The first electrode layer is used to connect one end of the resistor and the capacitor, and the second electrode layer is used to connect the other end of the capacitor and ground. The functional layer is used to switch between a high-resistance state and a low-resistance state under different voltage / current excitations. Based on the characteristic that the threshold voltage of the functional layer decreases with increasing temperature, thereby reducing the amplitude of the pulse signal and increasing the periodic oscillation frequency, the temperature signal is sensed by using the change in the threshold voltage as a response.

2. The temperature-sensing artificial neuron system according to claim 1, characterized in that, The temperature-sensing artificial neuron system converts temperature signals into pulse signals; and the temperature-sensing artificial neuron system will only output pulse signals when the voltage across the capacitor reaches the threshold voltage of the chalcogenide threshold conversion device.

3. The temperature-sensing artificial neuron system according to claim 1 or 2, characterized in that, Under voltage / current excitation, when the voltage across the capacitor is greater than the threshold voltage of the chalcogenide threshold switching device, the chalcogenide threshold switching device is in a low-resistance state, and the capacitor discharges through the chalcogenide threshold switching device; when the voltage across the capacitor is less than the holding voltage of the chalcogenide threshold switching device, the chalcogenide threshold switching device changes to a high-resistance state, and the capacitor is charged through the resistor.

4. The temperature-sensing artificial neuron system according to claim 1, characterized in that, Both the first electrode layer and the second electrode layer are made of inert electrode materials; wherein, the inert electrode materials are one or more of Pt, W, Au, Ru, Al, TiN, TiW, Ta, TaN, IrO2, ITO and IZO.

5. The temperature-sensing artificial neuron system according to claim 1, characterized in that, The capacitor is a fixed capacitor or a variable capacitor, with a capacitance value ranging from 1 fF to 1 μF.

6. The temperature-sensing artificial neuron system according to claim 1, characterized in that, The thickness of the first electrode layer and the second electrode layer is 5 nm to 200 nm; the thickness of the functional layer is 3 nm to 200 nm.

7. The temperature-sensing artificial neuron system according to claim 1, characterized in that, The sensing temperature range of the sulfide threshold switching device is lower than the crystallization temperature of the sulfide-based gating material.

8. The temperature-sensing artificial neuron system according to claim 1 or 2, characterized in that, The chalcogenide threshold switching device has a leakage current in the nanoampere range.

Images