A carbon-based field-effect force-sensitive synaptic device and its preparation method

Abstract

The present invention discloses a carbon-based field-effect force-sensitive synaptic device and a method for preparing the same. The synaptic device comprises an overlapping carbon-based laminated conformal presynaptic membrane and a carbon-based field-effect postsynaptic membrane. The carbon-based laminated conformal presynaptic membrane comprises a transmitter layer, a modified graphene floating gate layer, and a dielectric layer, which are sequentially stacked. The transmitter layer is a hexagonal boron nitride layer or an ion gel layer. The carbon-based laminated conformal presynaptic membrane has a plurality of inverted pyramid structures distributed in an array. A microstructured separation interface is formed between the transmitter layer and the channel layer. When the dielectric layer receives an action potential pulse, the highest point of the transmitter layer contacts the channel layer. The present invention combines artificial synaptic electronics with heterojunction technology. The force-sensitive floating gate is an arrayed micro-separated structure. Based on the carbon-based field effect, the device achieves a high degree of integration of sensing, storage, and computing functions at the synaptic interface in the field of pressure sensing. This not only achieves force-to-electricity signal conversion but also simulates the biological nervous system composed of neurons and synapses.

Description

Technical Field

[0001] The present invention relates to the technical field of microelectronic device technology, and in particular to a force-sensitive flexible synaptic device based on carbon-based field effect and a preparation method thereof. Background Art

[0002] Currently, the development of artificial intelligence (AI) is limited by the "von Neumann bottleneck" of storage. The physical separation of storage and computing units results in high computational effort, high energy consumption, and low timeliness, hindering the development of AI in an era of information explosion. The development of a "sensing, storage, and computing integrated system" can break the bottleneck of separate units in traditional architectures by integrating sensing, storage, and computing units. This will reduce the processing of redundant data, significantly reduce computer power consumption, and accelerate processing timeliness.

[0003] Flexible electronic skin is a new type of flexible sensor system that mimics human skin. By using electronic devices to mimic human skin, this type of flexible sensing system has the characteristics and functions of human skin. It has the characteristics of connectivity, low power consumption, and multi-function, and can be widely used in medical health, intelligent robots and other fields. However, the neuromorphic functions of skin-like sensing systems such as flexible electronic skin are realized through back-end device circuits. In comparison, human skin is a multi-layer structure that not only has receptors to sense environmental changes, but also has synaptically connected neurons to achieve edge storage and processing. Therefore, the scientific and technological challenges facing skin-like neuromorphic tactile sensing materials and devices are: how to achieve a high degree of integration of sensing-storage-computing functions at the flexible device level.

[0004] At present, the patent with the announcement number "CN115241320A" discloses a bionic synaptic transistor and its preparation method and application. The bionic synaptic transistor includes a substrate, a source electrode and a drain electrode, a channel layer, a charge-trapping tunneling layer, a photosensitive layer and a passivation layer stacked in sequence from bottom to top. It can realize the memory function of photoelectric signals and the recognition function of input images, imitating the visual recognition of images by organisms.

[0005] Another patent, CN117525199A, discloses a heterojunction sensing, storage, and computing device and its preparation method. The heterojunction sensing, storage, and computing device comprises a substrate, a back gate electrode, a stack of first / second / third gate dielectric layers, a one-dimensional nanowire, a two-dimensional layered semiconductor material (overlapping the one-dimensional nanowire to form a one-dimensional and two-dimensional heterojunction, serving as a channel layer), a source electrode, and a drain electrode. This device integrates information perception, storage, and computing functions within a single unit, improving the device's photoelectric sensing range and sensitivity and enabling integrated storage and computing applications with information perception capabilities in the visible and near-infrared bands.

[0006] The above-mentioned existing patents have achieved the integrated integration of sensing, storage and computing in devices to a certain extent, but they are all related to the fields of vision and optoelectronics, and the manufacturing process is relatively complex. The application of sensing, storage and computing functions at the pressure sensor device level still needs to be continuously explored. Summary of the Invention

[0007] The technical problem to be solved by the present invention is to provide a carbon-based field-effect force-sensitive synaptic device, which can utilize the charge transfer and storage characteristics of the carbon-based multi-layer heterostructure, and combine the force sensitivity and gate control characteristics of the microstructure electrode to achieve high integration of the sensing-storage-computing functions at the interface of the synaptic device; the technical problem to be solved by the present invention also includes providing a method for preparing a carbon-based field-effect force-sensitive synaptic device, which enables the prepared carbon-based field-effect force-sensitive synaptic device to form a heterojunction stack and a microstructure separation interface.

[0008] In order to solve the above technical problems, in a first aspect, the carbon-based field-effect force-sensitive synaptic device provided by the present invention adopts the following technical solutions:

[0009] A carbon-based field-effect force-sensitive synaptic device comprises an overlapping carbon-based laminated conformal presynaptic membrane and a carbon-based field-effect postsynaptic membrane, wherein the carbon-based field-effect postsynaptic membrane comprises a channel layer, and a source and a drain formed at both ends of the channel layer; the carbon-based laminated conformal presynaptic membrane comprises a transmitter layer, a modified graphene floating gate layer, and a dielectric layer stacked in sequence, wherein the transmitter layer is a hexagonal boron nitride layer or an ion gel layer; the carbon-based laminated conformal presynaptic membrane presents a plurality of inverted pyramid structures distributed in an array, a microstructured separation interface is formed between the transmitter layer and the channel layer, and when the dielectric layer receives an action potential pulse, the highest point of the transmitter layer can contact the channel layer.

[0010] By adopting the above technical solution, the transfer layer, the modified graphene floating gate layer and the dielectric layer form a heterojunction stack. The hexagonal boron nitride layer is used as the transfer layer. The wide bandgap characteristics and large electron affinity of hexagonal boron nitride enable it to form a higher tunneling barrier in the heterostructure, which can effectively curb the occurrence of direct tunneling current, making it possible to transfer charge. At the same time, the transfer layer and the dielectric layer form a stable insulating layer around the modified graphene floating gate layer to effectively isolate the floating gate to prevent the charge stored in the modified graphene floating gate layer from leaking. The charge in the channel layer will enter the modified graphene floating gate layer through the transfer layer under the action of the external electric field and be stored in the modified graphene floating gate layer (see Figure 1After the dielectric layer receives an action potential pulse, the entire carbon-based laminated conformal presynaptic membrane deforms, the highest potential of the transmitter layer contacts the channel layer, and some of the charge accumulated in the modified graphene floating gate layer returns to the channel layer through the transmitter layer and is exported for computational processing, achieving force-to-electricity conversion and the transmission of neurotransmitters (free electrons). The charge accumulated in the modified graphene floating gate layer utilizes the nonlinear relationship between pressure and tunneling current to realize a nonlinear force-sensitive synaptic memory process, ultimately achieving a high degree of integration of sensing-storage-computing functions at the interface of synaptic devices.

[0011] Moreover, the excellent surface quality of hexagonal boron nitride can not only significantly reduce interface defects and thus reduce leakage current, but also ensure the ideal flatness of the heterojunction interface, optimize carrier transport, improve charge transfer efficiency, and enhance the sensitivity of flexible synaptic devices to pressure.

[0012] The ion gel layer is used as the transmitter layer. The polymers that make up the ion gel are mostly colloidal block copolymers with a cross-linked network structure. The ionic liquid in the polymer copolymer will stably exist in the ion gel network structure in the form of anions and cations. However, under the action of an external electric field, the local movement of ions between the polymer chain and the ion coordination site produces the migration of anions and cations, thereby generating an uneven charge distribution inside the ion gel. The anions and cations are stacked at both ends of the dielectric layer interface to form a double layer (see Figure 2 ), so the ion gel can transmit the action potential pulses received by the dielectric layer to the channel layer for computational processing through the movement of charged ions, realizing force-to-electricity conversion and the transmission of neurotransmitters (charged ions). At the same time, the ion gel and the dielectric layer can form a stable insulating layer around the modified graphene floating gate layer, effectively isolating the floating gate to prevent the leakage of charge stored in the modified graphene floating gate layer. The charge accumulated in the modified graphene floating gate layer utilizes the nonlinear relationship between pressure and double-layer ion flow to realize a nonlinear force-sensitive synaptic memory process, ultimately achieving a high degree of integration of sensing-storage-computing functions at the interface of synaptic devices.

[0013] Furthermore, the carbon-based laminated conformal presynaptic membrane, with its multiple arrays of inverted pyramid structures, increases compressibility and stress concentration, enabling a significant change in the distance between the two electrodes under low pressure. This produces a large change in the electrical signal, significantly improving the sensitivity of the pressure sensor. The microstructured separation interface formed between the transmitter layer and the channel layer enhances the non-volatility of the charge in the modified graphene floating gate layer, further strengthening the memory process.

[0014] The present invention prepares a synaptic device by combining artificial synaptic electronics with heterojunction technology. The force-sensitive floating gate of the entire synaptic device is an array-type micro-separation structure, and a variety of carbon-based materials are rationally integrated. Based on the carbon-based field effect, the sensing, storage and computing functions in the pressure sensing field are highly integrated at the synaptic interface. It not only realizes force-to-electric signal conversion, but also has the memory, learning and training functions of neurons. It can simulate the biological nervous system composed of neurons and synapses, so that the synaptic device has the characteristics of parallelism, low power consumption, fault tolerance, self-learning and robustness. It has the potential to overcome the traditional von Neumann bottleneck and create a new paradigm to deal with complex problems such as pattern recognition, image classification, decision-making and associative learning.

[0015] Optionally, a side surface of the dielectric layer away from the modified graphene floating gate layer is covered with a top gate layer, and a material of the top gate layer is a carbon nanocomposite material.

[0016] By adopting the above technical solution, the carbon nanocomposite material enables the top gate layer to convert the pressure exerted on the flexible synaptic device into action potential pulses, thereby driving the entire weighted sum calculation process and realizing parallel processing of distributed pressure.

[0017] Optionally, the material of the dielectric layer is at least one of boron nitride and hafnium oxide.

[0018] By adopting the above technical solution, boron nitride and hafnium oxide both have excellent insulation properties, which can effectively prevent the leakage of charges stored in the modified graphene floating gate layer, thereby ensuring the force-electric conversion process and memory process of the synaptic device.

[0019] Optionally, the channel layer is formed by integrating carbon nanotubes on a flexible substrate.

[0020] By adopting the above technical solution, a flexible substrate is used as the base substrate of the synaptic device, and carbon nanotubes are integrated on the flexible substrate to obtain a channel layer. This makes the channel layer highly sensitive to electrical signals and obtains a flexible synaptic device that is close to bionic human skin.

[0021] Optionally, the preparation method of the ion gel comprises the following steps:

[0022] Polyethylene oxide, lithium perchlorate and methanol in a mass fraction ratio of 0.1:(0.01-0.015):(3-5) are prepared to obtain a mixed solution, the mixed solution is fully stirred, the mixed solution is centrifuged, and the supernatant is collected to obtain an ion gel.

[0023] By adopting the above technical solution, the ion gel prepared above can enable anions and cations to exist stably in the ion gel network structure, while making the migration of anions and cations more sensitive to the external electric field, so as to further enhance the non-volatility of the charge in the modified graphene floating gate layer, strengthen the memory process, and further improve the sensitivity of pressure sensing of the synaptic device.

[0024] In a second aspect, the present invention provides a method for preparing a carbon-based field-effect force-sensitive synaptic device using the following technical solutions:

[0025] A method for preparing a carbon-based field-effect force-sensitive synaptic device comprises the following steps:

[0026] A mold having an array of inverted pyramid-shaped grooves was prepared, and the mold was placed in a reaction chamber, wherein the temperature of the reaction chamber was maintained at 900-1100° C., an H 2 atmosphere with a flow rate of 9-11 sccm, and a vacuum environment of 0.01-0.2 Pa.

[0027] A boroamethane volatile is introduced into the reaction chamber to grow a transfer layer on the inner surface of the mold; a mixed gas is then introduced into the reaction chamber, wherein the mixed gas includes methane, ammonia, and an inert gas at a flow rate ratio of 10:(0.01-3):(190-210), and a modified graphene floating gate layer is grown on the surface of the transfer layer;

[0028] Continue to introduce borohydrazine volatiles into the reaction chamber to deposit a dielectric layer on the modified graphene floating gate layer; remove the mold from the reaction chamber, cast a carbon nanocomposite material on the surface of the dielectric layer, and obtain a top gate layer after curing, thereby preparing a carbon-based laminated conformal presynaptic membrane;

[0029] The carbon-based laminated conformal presynaptic membrane is transferred from the mold to a temporary substrate for storage. When used, the carbon-based laminated conformal presynaptic membrane and the carbon-based field-effect postsynaptic membrane are aligned and assembled, and then the temporary substrate is removed to obtain a carbon-based field-effect force-sensitive synaptic device.

[0030] By adopting the above technical solution and following the above steps, a stable heterojunction stack is formed between the hexagonal boron nitride transporter layer, the modified graphene floating gate layer, and the dielectric layer, and a microstructured separation interface is formed between the transporter layer and the channel layer. Furthermore, by doping the graphene with ammonia, an n-type nitrogen doping treatment is performed, resulting in a modified graphene floating gate layer that is essentially a mixed film of carbon atoms and nitrogen atoms. This improves the mechanical properties of the modified graphene floating gate layer and enables the formation of a higher localized charge, thereby enhancing the pressure sensing sensitivity of the synaptic device.

[0031] Optionally, the operations of depositing the transfer layer and the dielectric layer each include the following steps:

[0032] The reaction cavity is connected to a cavity filled with borazine, and an inert gas with a flow rate of 40-60 sccm is introduced into the reaction cavity from the cavity filled with borazine, so that the borazine volatiles enter the reaction cavity along with the inert gas flow and grow on the inner surface of the mold to form a transfer layer.

[0033] By adopting the above technical solution, by setting a certain flow rate of inert gas through borazine, an appropriate amount of borazine volatiles can be brought into the reaction chamber, so that a hexagonal boron nitride layer of uniform thickness can be formed on the mold.

[0034] Optionally, the number of deposited layers of hexagonal boron nitride in the transfer layer is 1-2 layers, and the number of deposited layers of boron nitride in the dielectric layer is 15-20 layers.

[0035] By adopting the above technical solution, the number of hexagonal boron nitride deposition layers of the dielectric layer is grown to 15-20 layers, which can prevent the top gate layer from being conductive with the modified graphene floating gate layer, thereby causing the modified graphene floating gate layer to lose its ability to store charge and the nonlinear relationship between pressure and ion / electron flow, thereby losing its memory storage function.

[0036] Optionally, the operation of transferring the carbon-based laminated conformal presynaptic membrane to a temporary substrate specifically includes the following steps:

[0037] Polydimethylsiloxane is spin-coated on the top gate layer and dried to obtain a temporary substrate / carbon-based laminated conformal presynaptic membrane / mold combination. The above combination is half-immersed in an etching solution, ensuring that the mold is in contact with the etching solution, and the mold is etched away.

[0038] Optionally, the preparation of the mold includes the following steps:

[0039] The silicon wafer is cleaned to remove surface impurities, and photolithography is performed on the silicon wafer to expose periodic micron-sized square holes in the silicon dioxide layer of the silicon wafer. The exposed square holes in the silicon dioxide layer are then dry-etched. After cleaning, the silicon wafer is wet-etched with potassium hydroxide and isopropyl alcohol to obtain multiple array-distributed inverted pyramid structure grooves on the silicon wafer.

[0040] By adopting the above technical solution, multiple inverted pyramid structures are first positioned by photolithography to expose the corresponding silicon dioxide layer, then the square holes are efficiently etched and deepened by dry etching, and finally the silicon wafer is anisotropically etched with potassium hydroxide to form multiple regular pyramid tip structures.

[0041] In a third aspect, the present invention provides a method for preparing a carbon-based field-effect force-sensitive synaptic device using the following technical solutions:

[0042] A method for preparing a carbon-based field-effect force-sensitive synaptic device comprises the following steps:

[0043] A mold having an array of inverted pyramid-shaped grooves was prepared, and the mold was placed in a reaction chamber, wherein the temperature of the reaction chamber was maintained at 900-1100° C., an H 2 atmosphere with a flow rate of 9-11 sccm, and a vacuum environment of 0.01-0.2 Pa.

[0044] Growing a modified graphene floating gate layer on the inner surface of the mold, and then continuing to grow a dielectric layer on the inner surface of the modified graphene floating gate layer mold; removing the mold from the reaction chamber, pouring a carbon nanocomposite material on the surface of the dielectric layer, and curing to obtain a top gate layer;

[0045] Combining the top gate layer with a temporary substrate, removing the mold, coating the surface of the modified graphene floating gate layer with an ion gel and curing it to form a transmitter layer, thereby obtaining a carbon-based laminated conformal presynaptic membrane;

[0046] During use, the carbon-based laminated conformal presynaptic membrane and the carbon-based field-effect postsynaptic membrane are aligned and assembled, and then the temporary substrate is removed to obtain a carbon-based field-effect force-sensitive synaptic device.

[0047] By adopting the above technical solution and following the above steps, a stable heterojunction stack can be formed between the ion gel-based transmitter layer, the modified graphene floating gate layer and the dielectric layer, thereby combining artificial synaptic electronics with heterojunction technology. The overall structure of the synaptic device adopts a separated design, and rationally integrates multiple carbon-based materials such as carbon nanocomposites, graphene, and carbon nanotubes. Based on the carbon-based field effect, a high degree of integration of sensing, storage, and computing functions at the synaptic interface is achieved in the field of pressure sensing.

[0048] In summary, the present invention includes at least one of the following beneficial technical effects:

[0049] 1. Hexagonal boron nitride or ion gel can form a stable insulating layer with the dielectric layer around the modified graphene floating gate layer, effectively isolating the floating gate to prevent the leakage of charge stored in the modified graphene floating gate layer. The charge accumulated in the modified graphene floating gate layer utilizes the nonlinear relationship between pressure and double layer ion flow or pressure and tunneling electron flow to realize nonlinear force-sensitive synaptic memory process, ultimately achieving a high degree of integration of sensing, storage and computing functions at the interface of synaptic devices.

[0050] 2. The carbon-based laminated conformal presynaptic membrane is arranged in an inverted pyramidal structure, increasing compressibility and stress concentration to achieve a significant change in the distance between the two electrodes under low pressure, thereby generating a large change in the electrical signal and significantly improving the sensitivity of the pressure sensor. The microstructured separation interface between the transmitter layer and the channel layer can enhance the non-volatility of the charge in the modified graphene floating gate layer, further strengthening this memory process.

[0051] 3. Synaptic devices are prepared by combining artificial synaptic electronics with heterojunction technology. The overall structure of the synaptic device adopts a discrete design and rationally integrates a variety of carbon-based materials. Based on the carbon-based field effect in the field of pressure sensing, the sensing, storage and computing functions are highly integrated at the synaptic interface. Not only does it realize force-to-electric signal conversion, but it also requires the memory, learning and training functions of neurons, simulating the biological nervous system composed of neurons and synapses. The synaptic device has the characteristics of parallelism, low power consumption, fault tolerance, self-learning and robustness. It has the potential to overcome the traditional von Neumann bottleneck and create a new paradigm to deal with complex problems such as pattern recognition, image classification, decision-making and associative learning. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a schematic diagram of the physical mechanism of the carbon-based field-effect force-sensitive synaptic device of the present invention (neurotransmitters are electrons).

[0053] Figure 2 It is a schematic diagram of the physical mechanism of the carbon-based field-effect force-sensitive synaptic device of the present invention (neurotransmitters are ions).

[0054] Figure 3 This is a block diagram for realizing the sensing, storage and computing functions of the carbon-based field-effect force-sensitive synaptic device of the present invention.

[0055] Figure 4 It is a schematic structural diagram of the carbon-based field-effect force-sensitive synaptic device of the present invention.

[0056] Figure 5 It is a schematic diagram of the pressure sensing principle of the carbon-based field-effect force-sensitive synaptic device of the present invention.

[0057] Figure 6 This is a flow chart of a method for preparing a carbon-based field-effect force-sensitive synaptic device (the transmitter layer is a hexagonal boron nitride layer) of the present invention.

[0058] Explanation of the accompanying symbols: 1. Carbon-based laminated conformal presynaptic membrane; 11. Transmitter layer; 12. Modified graphene floating gate layer; 13. Dielectric layer; 14. Top gate layer; 2. Carbon-based field-effect postsynaptic membrane; 21. Channel layer; 22. Source; 23. Drain; 3. Flexible substrate. DETAILED DESCRIPTION

[0059] The following is combined with Figure 4 The present invention is described in further detail.

[0060] The embodiment of the present invention discloses a carbon-based field-effect force-sensitive synaptic device, referring to Figure 4 The carbon-based field-effect force-sensitive synaptic device includes a carbon-based laminated common presynaptic membrane 1, a carbon-based field-effect postsynaptic membrane 2 and a flexible substrate 3 which are stacked in sequence.

[0061] The carbon-based postsynaptic field effect membrane 2 includes a channel layer 21, with a source electrode 22 and a drain electrode 23 formed at both ends of the channel layer 21. The channel layer 21 is formed by integrating carbon nanotubes (CNTs) on a flexible substrate 3.

[0062] The carbon-based laminated conformal presynaptic membrane 1 comprises a transmitter layer 11, a modified graphene floating gate layer 12, and a dielectric layer 13, which are stacked in sequence. The transmitter layer 11 is a hexagonal boron nitride layer or an ion gel layer. The entire carbon-based laminated conformal presynaptic membrane 1 has an array of multiple inverted pyramid structures. A microstructured separation interface is formed between the transmitter layer 11 and the channel layer 21. When the carbon-based laminated conformal presynaptic membrane 1 is deformed by force, the highest point of the transmitter layer 11 can contact the channel layer 21. The side of the dielectric layer 13 facing away from the modified graphene floating gate layer 12 is covered with a top gate layer 14. The material of the top gate layer 14 is a carbon nanocomposite material, which is a graphene slurry and / or a carbon nanotube slurry.

[0063] Part 1: The transmitter layer is a hexagonal boron nitride layer

[0064] Example 1:

[0065] A method for preparing a carbon-based field-effect force-sensitive synaptic device comprises the following steps:

[0066] Preparation of mold:

[0067] First, a silicon wafer was ultrasonically cleaned in deionized water, acetone, and alcohol for 15 minutes each to remove surface impurities. A photoresist was then spin-coated on the clean silicon wafer and dried on a hot plate at 100°C for 10 minutes. After drying, the wafer was exposed using a binary exposure machine and developed with a developer to form a square pattern. After photolithography, periodic micron-scale square holes were exposed in the silicon dioxide layer on the silicon wafer.

[0068] First, an RIE dry etcher (model PLASMALAB 133) was used to etch the exposed silicon dioxide using trifluoromethane as the etching gas, effectively etching and deepening the square holes.

[0069] The photoresist on the sample surface was then cleaned with acetone and alcohol. The pyramid structure was obtained by anisotropic etching of the silicon wafer with potassium hydroxide (concentration of 10%). Adding isopropyl alcohol to the etching solution can improve the orderliness of the pyramid structure etching; after rinsing with deionized water and then blowing it dry with a high-purity nitrogen gun, a mold with an array of multiple inverted pyramid structure grooves was obtained.

[0070] Preparation of carbon-based laminated conformal presynaptic membrane:

[0071] The mold was placed in a reaction chamber of a heating tube furnace, and the heating tube furnace was evacuated to 0.1 Pa. The heating tube furnace was heated to 1000° C. within 30 minutes. After the temperature reached 1000° C., it was maintained for 10 minutes in a H2 atmosphere with a flow rate of 10 sccm to stabilize the temperature of the entire heating tube furnace at 1000° C.

[0072] The reaction chamber in the heating tube furnace is connected to a cavity containing borazine. The valve between the cavity containing borazine and the heating tube furnace is opened, and argon gas at a flow rate of 50 seem is introduced into the cavity containing borazine, so that the borazine volatiles enter the reaction cavity along with the inert gas flow. The growth is carried out for 20 minutes to form a transfer layer on the inner surface of the mold.

[0073] The valve between the chamber containing borohydrazine and the heating tube furnace was closed, and a mixed gas comprising methane at a flow rate of 10 sccm, ammonia at 2 sccm, and argon at 200 sccm was introduced into the reaction chamber of the heating tube furnace for 10 minutes to grow a modified graphene floating gate layer on the surface of the transfer layer;

[0074] The valve between the cavity containing borazine and the heating tube furnace was opened, and argon gas at a flow rate of 50 seem was introduced into the cavity containing borazine, so that the borazine volatiles entered the reaction chamber along with the inert gas flow. The dielectric layer was grown on the inner surface of the mold for 300 minutes.

[0075] The mold is taken out of the reaction chamber, and graphene slurry is poured on the surface of the dielectric layer. After heating and curing, a top gate layer is obtained, that is, a carbon-based laminated conformal presynaptic membrane is prepared;

[0076] Polydimethylsiloxane is spin-coated on the top gate layer and dried to obtain a temporary substrate / carbon-based laminated conformal presynaptic membrane / mold combination. The above combination is half-immersed in an etching solution, which is a 10% potassium hydroxide solution, to ensure that the mold is in contact with the etching solution and to corrode and remove the mold.

[0077] Preparation of carbon-based field-effect postsynaptic membrane:

[0078] A carbon nanotube field-effect transistor is prepared on a flexible substrate (PET film) using CMOS integration technology to obtain a channel layer, and a drain and a source are formed on both sides of the channel layer (the gate portion can be played by a carbon-based laminated conformal presynaptic membrane).

[0079] Packaging of synaptic devices:

[0080] The resistive carbon-based laminated conformal presynaptic membrane and the carbon-based field-effect postsynaptic membrane are stored separately. When used, a laminating machine is used to align and package the carbon-based laminated conformal presynaptic membrane and the carbon-based field-effect postsynaptic membrane, and then the polydimethylsiloxane temporary substrate is removed to obtain a carbon-based field-effect force-sensitive synaptic device.

[0081] Example 2:

[0082] A method for preparing a carbon-based field-effect force-sensitive synaptic device is different from that of Example 1 in that argon gas with a flow rate of 40 sccm is introduced into the cavity containing borazine.

[0083] Example 3:

[0084] A method for preparing a carbon-based field-effect force-sensitive synaptic device is different from that of Example 1 in that argon gas with a flow rate of 60 sccm is introduced into the cavity containing borazine.

[0085] Example 4:

[0086] A method for preparing a carbon-based field-effect force-sensitive synaptic device is different from that of Example 1 in that the flow rate ratio of methane, ammonia and inert gas in the mixed gas is 10:0.01:190.

[0087] Example 5:

[0088] A method for preparing a carbon-based field-effect force-sensitive synaptic device is different from that of Example 1 in that the flow rate ratio of methane, ammonia and inert gas in the mixed gas is 10:3:210.

[0089] Comparative Example 1:

[0090] The difference from Example 1 is that the flow rate ratio of methane, ammonia and inert gas in the mixed gas is 10:5:200.

[0091] Comparative Example 2:

[0092] The difference from Example 1 is that the mold having an array of multiple inverted pyramid structure grooves is replaced with a silicon wafer that has not been treated as described above.

[0093] Performance test method:

[0094] The sensitivity of Examples 1-5 and Comparative Example 1 was measured, and the results are shown in Table 1.

[0095] The sensitivity is calculated by the ratio of the relative change in source-drain current to the change in pressure:

[0096]

[0097] ΔI is the relative change of source-drain current;

[0098] ΔP is the ratio of pressure change;

[0099] I o is the source current.

[0100] Table 1:

[0101] <![CDATA[Sensitivity (*10^3 kPa -1 )]]> Example 1 1.2 Example 2 1.1 Example 3 1.0 Example 4 1.1 Example 5 1.0 Comparative Example 1 0.6 Comparative Example 2 0.3

[0102] Combining Examples 1-3 and Table 1, it can be seen that when argon is introduced within the flow rate range of the present invention, the sensitivity of the flexible synaptic device is within 10^3 kPa. -1 The above means that it has excellent pressure sensing sensitivity.

[0103] In combination with Example 1, Examples 4-5 and Comparative Example 1, the mixed gas was introduced within the flow rate ratio range of the present application, and the sensitivity of the flexible synaptic device was 10^3kPa. -1 The above has excellent pressure sensing sensitivity.

[0104] It can be seen from Example 1, Comparative Example 2 and Table 1 that the array-type micro-separation structure between the transmitter layer and the channel layer in the present invention can significantly improve the sensitivity of the flexible synaptic device.

[0105] Part 2: The transmitter layer is an ion gel layer

[0106] Example 6:

[0107] A method for preparing a carbon-based field-effect force-sensitive synaptic device differs from Example 1 in that the preparation process of the carbon-based laminated conformal presynaptic membrane is different, as follows:

[0108] The mold was placed in a reaction chamber of a heating tube furnace, and the heating tube furnace was evacuated to 0.1 Pa. The heating tube furnace was heated to 1000° C. within 30 minutes. After the temperature reached 1000° C., it was maintained for 10 minutes in a H2 atmosphere with a flow rate of 10 sccm to stabilize the temperature of the entire heating tube furnace at 1000° C.

[0109] A mixed gas comprising 10 sccm of methane, 2 sccm of ammonia, and 200 sccm of argon was introduced into a reaction chamber in a heated tubular furnace for 10 minutes to grow a modified graphene floating gate layer on the inner surface of the mold.

[0110] The reaction chamber in the heating tube furnace is connected to a cavity containing borazine. A valve between the cavity containing borazine and the heating tube furnace is opened, and argon gas at a flow rate of 50 seem is introduced into the cavity containing borazine, so that borazine volatiles follow the inert gas flow into the reaction cavity. The dielectric layer is grown on the modified graphene floating gate layer for 300 minutes. The mold is removed from the reaction cavity, and a graphene slurry is poured on the surface of the dielectric layer. The top gate layer is obtained after heating and curing.

[0111] Polydimethylsiloxane is spin-coated on the top gate layer and dried to obtain a temporary substrate / carbon-based laminated conformal presynaptic membrane / mold combination. The above combination is half-immersed in an etching solution to etch and remove the mold to expose the modified graphene floating gate layer. An ion gel is coated on the surface of the modified graphene floating gate layer and cured to form a transmitter layer, thereby obtaining a carbon-based laminated conformal presynaptic membrane.

[0112] The preparation process of ion gel includes the following steps:

[0113] A mixed solution of polyethylene oxide, lithium perchlorate and methanol in a mass fraction ratio of 0.1:0.012:4 was prepared, stirred at 45°C for 3 hours using a magnetic stirrer, and then the mixed solution was centrifuged at high speed using a centrifuge, and the supernatant was collected to obtain an ion gel.

[0114] Example 7:

[0115] A method for preparing a carbon-based field-effect force-sensitive synaptic device is different from that of Example 6 in that the mass fraction ratio of polyethylene oxide, lithium perchlorate and methanol is 0.1:0.012:4.

[0116] Example 8:

[0117] A method for preparing a carbon-based field-effect force-sensitive synaptic device is different from that of Example 6 in that the mass fraction ratio of polyethylene oxide, lithium perchlorate and methanol is 0.1:0.012:4.

[0118] Performance test method:

[0119] Sensitivity calculations were performed on Examples 6-8, and the results are shown in Table 2.

[0120] Table 2:

[0121] <![CDATA[Sensitivity (*10^3 kPa -1 )]]> Example 6 1.0 Example 7 0.9 Example 8 1.1

[0122] Combining Examples 6-8 and Table 1, it can be seen that the sensitivity of the flexible synaptic device prepared by preparing the ion gel within the mass fraction ratio range of the present invention is 0.9*10^3kPa. -1 The above has excellent pressure sensing sensitivity.

[0123] The above are all preferred embodiments of the present invention, and are not intended to limit the scope of protection of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A carbon-based field-effect force-sensitive synaptic device, characterized in that: The present invention comprises a carbon-based laminated conformal presynaptic membrane (1) and a carbon-based field effect postsynaptic membrane (2) which are arranged in an overlapping manner, wherein the carbon-based field effect postsynaptic membrane (2) comprises a channel layer (21), a source electrode (22) and a drain electrode (23) formed at both ends of the channel layer (21); the carbon-based laminated conformal presynaptic membrane (1) comprises a transmitter layer (11), a modified graphene floating gate layer (12) and a dielectric layer (13) which are arranged in a stacked manner, wherein the transmitter layer (11) is a hexagonal boron nitride layer or an ion gel layer, and ... and the carbon-based laminated conformal presynaptic membrane (1) comprises a transmitter layer (11), a modified graphene floating gate layer (12) and a dielectric layer (13) which are arranged in a stacked manner. The base stacked conformal presynaptic membrane (1) presents a plurality of inverted pyramid structures distributed in an array, a microstructure separation interface is formed between the transmitter layer (11) and the channel layer (21), and when the dielectric layer (13) receives an action potential pulse, the highest point of the transmitter layer (11) can contact the channel layer (21); the side surface of the dielectric layer (13) away from the modified graphene floating gate layer (12) is covered with a top gate layer (14), and the material of the top gate layer (14) is a carbon nanocomposite material.

2. The carbon-based field-effect force-sensitive synaptic device according to claim 1, characterized in that: The channel layer (21) is formed by integrating carbon nanotubes on a flexible substrate (3).

3. The carbon-based field-effect force-sensitive synaptic device according to claim 1, characterized in that: The preparation method of the ion gel comprises the following steps: A mixed solution is prepared by mixing polyethylene oxide, lithium perchlorate and methanol in a mass fraction ratio of 0.1:(0.01-0.015):(3-5), fully stirring the mixed solution, centrifuging the mixed solution, and collecting the supernatant to obtain an ion gel.

4. A method for preparing a carbon-based field-effect force-sensitive synaptic device according to any one of claims 1 to 3, characterized in that: The following steps are involved: A mold having an array of inverted pyramid-shaped grooves was prepared, and the mold was placed in a reaction chamber, wherein the temperature of the reaction chamber was maintained at 900-1100° C., an H 2 atmosphere with a flow rate of 9-11 sccm, and a vacuum environment of 0.01-0.2 Pa. A boroamethane volatile is introduced into the reaction chamber to grow a transfer layer (11) on the inner surface of the mold; a mixed gas is then introduced into the reaction chamber, wherein the mixed gas includes methane, ammonia, and an inert gas at a flow rate ratio of 10: (0.01-3): (190-210), and a modified graphene floating gate layer (12) is grown on the surface of the transfer layer (11); Continue to introduce borohydrazine volatiles into the reaction chamber to deposit a dielectric layer (13) on the modified graphene floating gate layer (12); remove the mold from the reaction chamber, cast a carbon nanocomposite material on the surface of the dielectric layer (13), and obtain a top gate layer (14) after curing, thereby preparing a carbon-based laminated conformal presynaptic membrane; The carbon-based laminated conformal presynaptic membrane is transferred from the mold to a temporary substrate for storage. When in use, the carbon-based laminated conformal presynaptic membrane and the carbon-based field effect postsynaptic membrane (2) are aligned and assembled, and then the temporary substrate is removed to obtain a carbon-based field effect force-sensitive synaptic device.

5. The method for preparing a carbon-based field-effect force-sensitive synaptic device according to claim 4, characterized in that: The operations of depositing the transfer layer (11) and the dielectric layer (13) both include the following steps: The reaction cavity is connected to a cavity containing borazine, and an inert gas with a flow rate of 40-60 sccm is introduced from the cavity containing borazine into the reaction cavity, so that the borazine volatiles follow the inert gas flow into the reaction cavity and grow on the inner surface of the mold to form a transfer layer (11).

6. The method for preparing a carbon-based field-effect force-sensitive synaptic device according to claim 4, characterized in that: The number of deposited layers of hexagonal boron nitride in the transfer layer (11) is 1-2 layers, and the number of deposited layers of boron nitride in the dielectric layer (13) is 15-20 layers.

7. The method for preparing a carbon-based field-effect force-sensitive synaptic device according to claim 4, characterized in that: The operation of transferring the carbon-based laminated conformal presynaptic membrane to a temporary substrate specifically comprises the following steps: Polydimethylsiloxane is spin-coated on the top gate layer (14), and after drying, a combination of a temporary substrate / carbon-based laminated conformal presynaptic membrane / mold is obtained. The combination is half-immersed in an etching solution, and the mold is ensured to be in contact with the etching solution, and the mold is etched away.

8. The method for preparing a carbon-based field-effect force-sensitive synaptic device according to claim 4, characterized in that: The preparation of the mold comprises the following steps: The silicon wafer is cleaned to remove surface impurities, and photolithography is performed on the silicon wafer to expose periodic micron-sized square holes in the silicon dioxide layer of the silicon wafer. The exposed square holes in the silicon dioxide layer are then dry-etched. After cleaning, the silicon wafer is wet-etched with potassium hydroxide and isopropyl alcohol to obtain multiple array-distributed inverted pyramid structure grooves on the silicon wafer.

9. A method for preparing a carbon-based field-effect force-sensitive synaptic device according to any one of claims 1 to 3, characterized in that: The following steps are involved: A mold having an array of inverted pyramid-shaped grooves was prepared, and the mold was placed in a reaction chamber, wherein the temperature of the reaction chamber was maintained at 900-1100° C., an H 2 atmosphere with a flow rate of 9-11 sccm, and a vacuum environment of 0.01-0.2 Pa. A modified graphene floating gate layer (12) is grown on the inner surface of the mold, and then a dielectric layer (13) is continuously grown on the inner surface of the modified graphene floating gate layer (12) mold; the mold is removed from the reaction chamber, a carbon nanocomposite material is poured on the surface of the dielectric layer (13), and a top gate layer (14) is obtained after curing; The top gate layer (14) is combined with a temporary substrate, the mold is removed, and an ion gel is coated on the surface of the modified graphene floating gate layer (12) and solidified to form a transmitter layer (11), thereby obtaining a carbon-based laminated conformal presynaptic membrane; During use, the carbon-based laminated conformal presynaptic membrane and the carbon-based field-effect postsynaptic membrane (2) are aligned and assembled, and then the temporary substrate is removed to obtain a carbon-based field-effect force-sensitive synaptic device.

Images