Reconfigurable synapse device, neuromorphic device including reconfigurable synapse device, and manufacturing methods thereof
A reconfigurable synapse device using α-In2Se3 with a Se-vacancy generation region and In oxide layer addresses industrial applicability issues by ensuring compatibility with CMOS processes and improving integration density and mass production.
Patent Information
- Application Number
- KR1020250177475
- Authority / Receiving Office
- KR · KR
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2025-07-04
- Filing Date
- 2025-11-20
- Publication Date
- 2026-11-11
- Estimated Expiration
- 2045-11-20
AI Technical Summary
Existing reconfigurable neuromorphic devices based on two-dimensional materials face challenges in industrial applicability due to process complexity, reproducibility issues, and integration density limitations, particularly when stacking multiple types of two-dimensional materials and using multiple gate electrodes.
A reconfigurable synapse device utilizing a single two-dimensional material, α-In2Se3, with a Se-vacancy generation region and an In oxide layer, enabling ambipolar characteristics and a single-gate structure, fabricated through oxygen plasma treatment, which is compatible with CMOS processes.
The solution allows for a reconfigurable synapse device with improved industrial applicability, enhanced integration density, and mass production capabilities, while maintaining excellent performance and reproducibility.
Smart Images

Figure 112025130433858-PAT00001_ABST
Abstract
Description
Technology Field
[0001] The present invention relates to an electronic device, a method for manufacturing the same, and a device including the electronic device, and more specifically, to a synapse device, a neuromorphic device including the same, and a method for manufacturing the same. Background Technology
[0002] Neuromorphic computing systems are receiving significant attention as a new concept capable of overcoming the limitations of existing von Neumann-based computer systems. Neuromorphic computing implements artificial intelligence operations by hardware-mimicking the human brain. Based on the observation that the human brain performs very complex functions while consuming only about 20W of energy, neuromorphic computing mimics the structure of the human brain itself, enabling it to perform AI operations such as association, reasoning, and recognition—which are far superior to existing von Neumann-based computing—with ultra-low power consumption.
[0003] Neuromorphic systems that enable such neuromorphic computing are composed of numerous neurons (neuronal elements) and synapses (synaptic elements), just like the human brain, and include additional circuits for signal processing and transmission. Synapses must remember the connection strength (weight) based on the correlation of spikes expressed by neurons, and, in some cases, adjust the said connection strength through processes of potentialization and depression. As synaptic elements, RRAM (resistive random access memory) and memristor-based devices have been widely studied, and recently, MOSFET (metal-oxide-semiconductor field-effect transistor)-based synaptic elements are also being researched.
[0004] However, since each AI chip (neuromorphic device) is custom-made for a single application, there is a need to perform reconfiguration operations to process dynamic data distribution across various scenarios; consequently, there are difficulties in fabricating devices capable of performing these operations. Reconfigurable neuromorphic devices enable versatile AI applications by maximizing the reallocation of finite resources to increase functionality rather than complexity. Reconfigurable devices can integrate multiple functions—such as synaptic short-term plasticity (STP), long-term plasticity (LTP), dendrites, and filtering—into a single device in an all-in-one manner, allowing for adaptability to various applications.
[0005] Meanwhile, research is underway on next-generation materials capable of overcoming the limitations of existing devices, and two-dimensional materials (2D materials) are attracting attention as next-generation materials due to their superior properties compared to silicon. Two-dimensional materials are single-layer or half-layer solids in which atoms form a specific crystal structure, and they can exhibit excellent electrical, optical, and mechanical properties. Transition metal dichalcogenides (TMDs) are representative two-dimensional materials, and representative TMD materials include MoS2 (n-type semiconductor), WS2 (n-type semiconductor), and WSe2 (p-type semiconductor). MoS2 is a semiconductor that has a relatively wide direct bandgap of 1.9 eV in its monolayer form and an indirect bandgap of 1.3 eV in its bulk form. Due to features such as a high ON / OFF current ratio resulting from the wide bandgap, high field-effect mobility, and low subthreshold swing (SS), it is considered a promising candidate for low-power electronic devices. In addition, it has recently been reported that two-dimensional materials can be grown over large areas using the metal-organic chemical vapor deposition (MOCVD) method, which may be an essential factor in lowering semiconductor production costs.
[0006] Technologies have been proposed to fabricate reconfigurable neuromorphic devices by applying two-dimensional materials (semiconductor materials). However, reconfigurable neuromorphic devices based on two-dimensional materials developed and reported to date feature a structure in which two or more types of two-dimensional materials are stacked and bonded, and two or more gate electrodes are used. By using two or more types of two-dimensional materials and two gate electrodes, they can be fabricated to modulate ions, charges, phases, etc. However, bonding (stacking) two or more types of two-dimensional materials in multiple layers is impossible to apply to actual industrial processes, is incompatible with CMOS (complementary metal oxide semiconductor) processes, and suffers from poor reproducibility; furthermore, two or more gates can be a critical disadvantage regarding integration density. Therefore, existing reconfigurable neuromorphic devices based on two-dimensional materials may be impossible to apply to actual industries due to circuit complexity and process difficulties.
[0007] Therefore, there may be a need to develop reconfigurable devices (neuromorphic devices) based on two-dimensional materials that possess excellent performance and are applicable to actual industrialization in terms of ease of process, reproducibility, mass producibility, and device / circuit structure. Prior art literature
[0008] Korean Published Patent 10-2018-0052819 The problem to be solved
[0009] The technical problem that the present invention aims to solve is to provide a reconfigurable synapse device with excellent performance and applicability to actual industrialization in terms of ease of process, reproducibility, mass production, and device / circuit structure, by applying a two-dimensional material.
[0010] In addition, the technical problem that the present invention aims to solve is to provide a method for manufacturing the above-mentioned reconfigurable synapse device.
[0011] In addition, the technical problem that the present invention aims to solve is to provide a neuromorphic device comprising the above-described reconfigurable synapse element and a method for manufacturing the same.
[0012] The problems that the present invention aims to solve are not limited to those mentioned above, and other unmentioned problems will be understood by those skilled in the art from the description below. means of solving the problem
[0013] According to one embodiment of the present invention, a reconfigurable synapse device is provided, comprising: a channel layer comprising α-In2Se3, which is a two-dimensional material; a gate electrode for applying an electric field to the channel layer; a source electrode and a drain electrode electrically connected to the channel layer; an In oxide layer disposed between the channel layer and the source electrode and between the channel layer and the drain electrode; and a Se-vacancy generation region provided within the channel layer in contact with the In oxide layer at each of the source electrode and drain electrode sides, wherein a plurality of Se-vacancies are generated; wherein the channel layer has both n-type semiconductor characteristics and p-type semiconductor characteristics, and the synapse device is configured to be reconfigurable with ambipolar characteristics.
[0014] The above In oxide layer may include In2O3.
[0015] The above In oxide layer may have a thickness of about 1 to 3 nm.
[0016] The above Se-vacuum generation region can be placed below the above In oxide layer.
[0017] The above Se-vacancy generation region may be a region where the Se element is reduced by at least about 10% compared to other regions of the channel layer.
[0018] The above channel layer may have a monolayer structure of α-In2Se3 or a multilayer structure in which the monolayer is stacked multiple times.
[0019] The above channel layer can be composed of a single two-dimensional material.
[0020] The above synapse device may have a single-gate structure.
[0021] According to another embodiment of the present invention, a neuromorphic device comprising the aforementioned reconfigurable synapse element is provided.
[0022] According to another embodiment of the present invention, a method for manufacturing a reconfigurable synapse device comprising a gate electrode, a source electrode, and a drain electrode is provided, comprising: a step of forming a channel layer comprising a two-dimensional material α-In2Se3; a step of performing oxygen plasma treatment on a first and second region of the channel layer to form an In oxide layer in the first and second region and forming a Se-vacancy generation region having a plurality of Se-vacancies within the channel layer in contact with the In oxide layer; and a step of forming a source electrode electrically connected to the In oxide layer on the first region of the channel layer and a drain electrode electrically connected to the In oxide layer on the second region of the channel layer; wherein the channel layer has both n-type semiconductor characteristics and p-type semiconductor characteristics, and the synapse device is configured to have ambipolar characteristics and be reconfigurable.
[0023] After forming a mask pattern having openings that expose the first and second regions respectively on the channel layer, the oxygen plasma treatment can be performed on the first and second regions of the channel layer exposed by the mask pattern.
[0024] The above In oxide layer may include In2O3.
[0025] The above Se-vacuum generation region can be placed below the above In oxide layer.
[0026] The above channel layer may have a monolayer structure of α-In2Se3 or a multilayer structure in which the monolayer is stacked multiple times.
[0027] The above channel layer can be composed of a single two-dimensional material.
[0028] The above synapse device may have a single-gate structure.
[0029] According to another embodiment of the present invention, a method for manufacturing a neuromorphic device is provided, comprising the step of manufacturing a reconfigurable synapse device using the method described above. Effects of the invention
[0030] According to embodiments of the present invention, by applying a two-dimensional material, a reconfigurable synapse device having excellent performance and applicability to actual industrialization in terms of ease of process, reproducibility, mass production, and device / circuit structure can be realized. Furthermore, according to embodiments of the present invention, a neuromorphic device including the above-mentioned reconfigurable synapse device can be realized.
[0031] According to one embodiment, ambipolar characteristics can be realized by using a single two-dimensional material (α-In2Se3, a ferroelectric material) in the channel layer and treating it with oxygen plasma. Therefore, in the manufacturing of a reconfigurable synapse device, a process that is applicable to actual industrialization (mass production), compatible with silicon processes such as CMOS processes, and has excellent reproducibility can be realized. Furthermore, according to one embodiment, a synapse device having reconfigurable characteristics can be realized even in a single-gate structure. This can be advantageous for improving integration density and increasing feasibility compared to devices using two or more gates.
[0032] By applying a reconfigurable synapse element according to the embodiments of the present invention, a neuromorphic element and system having excellent performance can be realized.
[0033] However, the effects of the present invention are not limited to the above effects and can be extended in various ways without departing from the technical concept and scope of the present invention. Brief explanation of the drawing
[0034] FIG. 1 is a cross-sectional view showing a reconfigurable synapse element according to one embodiment of the present invention. FIG. 2 is a cross-sectional view showing a reconfigurable synapse element according to another embodiment of the present invention. FIGS. 3a to 3d are cross-sectional views showing a method for manufacturing a reconfigurable synapse device according to one embodiment of the present invention. FIG. 4 is a cross-sectional view illustrating a method for manufacturing a reconfigurable synapse device according to another embodiment of the present invention. FIGS. 5a to 5c are cross-sectional views showing a method for manufacturing a reconfigurable synapse device according to an embodiment of the present invention. Figure 6 is a graph showing the photoluminescence (PL) analysis results for the channel layer when doping by oxygen plasma treatment was not performed and when it was performed. Figures 7 to 9 are graphs showing the results of X-ray photoelectron spectroscopy (XPS) analysis of the channel layer in cases where doping by oxygen plasma treatment was not performed and cases where it was performed. Figure 10 is a graph showing the transfer characteristic of a device (transistor) according to a comparative example. FIG. 11 is a graph showing the transfer characteristic of a reconfigurable synapse device according to one embodiment of the present invention. FIG. 12 is a graph showing the weight update characteristics in the excitatory mode of a reconfigurable synapse device according to one embodiment of the present invention. FIG. 13 is a graph showing the weight update characteristics in the inhibitory mode of a reconfigurable synapse device according to one embodiment of the present invention. FIG. 14 is a graph showing a rough comparison of the channel layer configuration and ON / OFF current ratio of a reconfigurable synapse device according to an embodiment of the present invention and a synapse device based on a conventional two-dimensional material. Specific details for implementing the invention
[0035] Hereinafter, embodiments of the present invention will be described in detail with reference to the attached drawings.
[0036] The embodiments of the present invention described below are provided to more clearly explain the present invention to those skilled in the art, and the scope of the present invention is not limited by the following embodiments, and the following embodiments may be modified in various other forms.
[0037] The terms used herein are for describing specific embodiments and are not intended to limit the invention. Terms used herein in the singular form may include plural forms unless the context clearly indicates otherwise. Additionally, the terms “comprise” and / or “comprising” used herein specify the presence of the mentioned features, steps, numbers, actions, components, elements, and / or groups thereof, and do not exclude the presence or addition of one or more other features, steps, numbers, actions, components, elements, and / or groups thereof. Furthermore, the term “connected” used herein means not only that components are directly connected, but also includes the concept of indirectly connecting components through the interposition of additional components between them.
[0038] Furthermore, when a component is described in this specification as being located "on" another component, this includes not only cases where a component is in contact with another component, but also cases where another component exists between the two components. The term "and / or" as used in this specification includes any one of the listed items and all combinations of one or more thereof. Additionally, terms of degree such as "about" and "substantially" as used in this specification are used to mean a range of numerical values or degrees or approximate values, taking into account inherent manufacturing and material tolerances, and are used to prevent an infringer from unfairly exploiting the disclosures in which precise or absolute figures provided to aid in understanding this specification are mentioned.
[0039] Embodiments of the present invention will be described in detail below with reference to the attached drawings. The sizes or thicknesses of the areas or parts depicted in the attached drawings may be slightly exaggerated for the clarity of the specification and convenience of explanation. Throughout the detailed description, the same reference numerals indicate the same components.
[0040] FIG. 1 is a cross-sectional view showing a reconfigurable synapse element according to one embodiment of the present invention.
[0041] Referring to FIG. 1, a reconfigurable synapse device according to an embodiment of the present invention may include a channel layer (200) comprising α-In2Se3, which is a two-dimensional material (2D material), a gate electrode (100) for applying an electric field to the channel layer (200), and a source electrode (300A) and a drain electrode (300B) electrically connected (contacted) to the channel layer (200).
[0042] A two-dimensional material may be a material in which unit layers are bonded by van der Waals forces. Additionally, the two-dimensional material may be bonded to a predetermined underlying material layer by van der Waals forces. In this respect, the two-dimensional material may be referred to as a 'two-dimensional van der Waals (vdW) material'. The two-dimensional material may have a monolayer structure or a multilayer structure in which the monolayer is stacked multiple times. The monolayer may have a crystal structure and may correspond to a unit layer. The crystal structure of the monolayer may be 2H α-phase (α-In2Se3).
[0043] α-In2Se3 is an In2Se3 material having an α-phase (crystalline phase) and can possess ferroelectricity. α-In2Se3 can be a ferroelectric semiconductor and can be applied as a memory material. α-In2Se3 can basically have n-type semiconductor characteristics. Since α-In2Se3 materials can have a thin two-dimensional layer structure at the molecular layer (atomic layer) level and can effectively provide high switching speed, large on / off ratio, low operating voltage, and low power consumption characteristics from an interface without dangling coupling, they have excellent potential to form artificial synapse devices. Two-dimensional materials such as CuInP2S6 (CIPS), In2Se3, and WTe2, which possess ferroelectricity, can be utilized as analog synapse devices such as ferroelectric field-effect transistors (FeFETs), ferroelectric semiconductor FETs (FeSFETs), and ferroelectric semiconductor junctions (FSJs). In particular, α-In2Se3 can possess both in-plane (IP) and out-of-plane (OOP) ferroelectric polarization at the single-layer level and can have characteristics that allow for the inversion of IP polarization by OOP bias and the inversion of OOP polarization by IP bias at room temperature.
[0044] The channel layer (200) may have a monolayer structure of α-In2Se3 or a multilayer structure in which the monolayer is stacked multiple times. Even if the monolayer (i.e., unit layer) of the two-dimensional material is stacked multiple times, the quantum mechanical properties of the two-dimensional material can be maintained. Therefore, the channel layer (200) can be referred to as a 'two-dimensional material' even when it has a structure in which the monolayer (i.e., unit layer) is stacked multiple times. In the channel layer (200), the monolayer may be stacked several times to tens of times. For example, the channel layer (200) may have a monolayer structure of α-In2Se3 or a structure in which the monolayer of α-In2Se3 is stacked about 2 to 30 times. In other words, the number of monolayers of α-In2Se3 constituting the channel layer (200) may be about 1 to 30 layers. The channel layer (200) may have a monolayer structure of α-In2Se3, a bilayer structure, or a multilayer structure of three or more layers (a structure similar to bulk).
[0045] The channel layer (200) may be composed of a single two-dimensional material. In other words, the type of two-dimensional material included in the channel layer (200) may be one (i.e., α-In2Se3). When two or more types of two-dimensional materials are bonded (stacked) in multiple layers as in the conventional method, this technology has problems and limitations such as being impossible to apply to actual industrial processes, incompatibility with CMOS processes, and poor reproducibility. However, in the embodiment of the present invention, since a single two-dimensional material is applied to the channel layer (200), the problems and limitations of the technology of bonding two or more types of two-dimensional materials in multiple layers can be overcome.
[0046] A reconfigurable synapse device according to an embodiment of the present invention may include an In oxide layer (10) disposed between a channel layer (200) and a source electrode (300A) and between a channel layer (200) and a drain electrode (300B). The In oxide layer (10) disposed between the channel layer (200) and the source electrode (300A) may be referred to as a first In oxide layer, and the In oxide layer (10) disposed between the channel layer (200) and the drain electrode (300B) may be referred to as a second In oxide layer. Additionally, the synapse device may include a Se-vacancy generation region (20) in which a plurality of Se-vacancies are generated, provided within the channel layer (200) in contact with the In oxide layer (10) on each side of the source electrode (300A) and the drain electrode (300B). The Se-vacancy generation region (20) provided within the channel layer (200) in contact with the In oxide layer (10) on the side of the source electrode (300A) can be referred to as the first Se-vacancy generation region, and the Se-vacancy generation region (20) provided within the channel layer (200) in contact with the In oxide layer (10) on the side of the drain electrode (300B) can be referred to as the second Se-vacancy generation region. The Se-vacancy generation region (20) may be a region in which a plurality of Se-vacancies are generated as at least some Se elements are removed from the α-In2Se3 structure. The Se-vacancy generation region (20) can be viewed as part of the channel layer (200). The first In oxide layer may be disposed between the source electrode (300A) and the first Se-vacancy generation region, and the second In oxide layer may be disposed between the drain electrode (300B) and the second Se-vacancy generation region.
[0047] The In oxide layer (10) and the Se-vacancy generation region (20) may be formed by 'oxygen plasma treatment'. The In oxide layer (10) may, for example, contain In2O3 or be composed of In2O3. The In oxide layer (10) may be a semiconductor and may have higher electrical conductivity than α-In2Se3. As a non-limiting example, the thickness of the In oxide layer (10) may be about 1 to 3 nm or about 2 to 3 nm. The In oxide layer (10) may be formed on the surface portion of the channel layer (200) and may be formed to protrude somewhat above the surface of the channel layer (200). Additionally, the In oxide layer (10) may be formed to penetrate somewhat into the interior of the channel layer (200).
[0048] The Se-vacancy region (20) may be placed below the In oxide layer (10). The Se-vacancy region (20) may be in contact with the lower surface of the In oxide layer (10). The Se-vacancy region (20) may be a region where the Se element is reduced by at least about 10% compared to other regions of the channel layer (200). As a non-limiting example, the Se-vacancy region (20) may be a region where the Se element is reduced by about 10–99% compared to other regions of the channel layer (200), a region where it is reduced by about 10–70%, a region where it is reduced by about 10–50%, or a region where it is reduced by about 10–35%. The thickness of the Se-vacancy region (20) may be several Å or more. Depending on the thickness of the channel layer (200) or the conditions of the oxygen plasma treatment, the thickness of the Se-vacancy region (20) may vary. According to one example, a Se-vacancy region (20) may be formed throughout the entire region of the channel layer (200) below the In oxide layer (10), but is not limited thereto.
[0049] The In oxide layer (10) and the Se-vacancy generation region (20) can mutually perform the role of exhibiting p-type semiconductor characteristics in the channel layer (200) and enabling the synapse device to have reconfigurable characteristics. The In oxide layer (10) has relatively high electrical conductivity and can attract and absorb electrons. In the semiconductor channel region in contact with the In oxide layer (10), the concentration of electrons is relatively low, so p-type semiconductor characteristics may appear. The In oxide layer (10) can cause a p-type doping effect in the channel region in contact with it. The Se-vacancy in the Se-vacancy generation region (20) can act as a type of acceptor. Therefore, the Se-vacancy generation region (20) can have p-type semiconductor characteristics. Electrons can be trapped in the Se-vacancy, and a p-type doping effect may appear. The source / drain junction region or the adjacent region may exhibit p-type semiconductor characteristics due to the In oxide layer (10) and the Se-vacancy generation region (20).
[0050] Surface charge transfer doping (SCTD) effects may occur due to the In oxide layer (10) and the Se-vacancy generation region (20). A p-type doping effect may occur in the oxygen plasma-treated source / drain junction region or in the adjacent region, while n-type semiconductor characteristics based on the fundamental properties of α-In2Se3 may be maintained in the remaining channel region. In this regard, p-type doping characteristics may occur as the Fermi level of the channel is lowered in the source / drain junction region or in the adjacent region, and the remaining channel region may have n-type semiconductor characteristics. Therefore, the channel layer (200) may have both n-type semiconductor characteristics and p-type semiconductor characteristics.
[0051] The above synapse device may have ambipolar characteristics and may be configured to be reconfigurable. That is, the above synapse device may possess both n-type transistor characteristics and p-type transistor characteristics, and may operate as an n-type transistor or a p-type transistor by controlling gate voltage conditions, drain voltage conditions, etc. In this respect, the above synapse device may be a reconfigurable device. The above synapse device may be a type of memory transistor. The above memory transistor may be a ferroelectric memory transistor and may be referred to as an FeFET or FeSFET. Furthermore, the above synapse device may be a type of neuromorphic device.
[0052] According to one embodiment, the gate electrode (100) may be a bottom gate electrode. In this case, the gate electrode (100) may be disposed below the channel layer (200). A gate insulating layer (150) may be disposed between the gate electrode (100) and the channel layer (200). The gate electrode (100) may be formed within the substrate or on the substrate. The ratio of the length (width) of the gate electrode (100) to the length (width) of the channel layer (200) is not limited to what is illustrated and may vary. The gate insulating layer (150) may be configured to include at least one of various dielectric materials. Any gate insulating layer material of a general transistor may be applicable to the gate insulating layer (150). In some cases, the synapse element may have a top gate structure rather than a bottom gate structure. In other words, the gate electrode (100) may be disposed above the channel layer (200) rather than below it.
[0053] The above synapse device may have a single gate structure. That is, the above synapse device may be a device that controls the characteristics of the channel layer (200) with a single gate electrode (i.e., 100) rather than multiple gate electrodes. Using two or more gate electrodes in a single transistor can be a fatal disadvantage to integration density and may result in reduced process efficiency and mass production capabilities. However, since the synapse device according to the embodiment of the present invention can implement reconfigurable characteristics using a single gate, it may be significantly advantageous in terms of improved integration density, mass production capabilities, and commercialization.
[0054] Meanwhile, the source electrode (300A) and the drain electrode (300B) may include a conductive material such as a metal or a metal compound. The source electrode (300A) and the drain electrode (300B) may have a single-layer structure or a multi-layer structure. The source electrode (300A) and the drain electrode (300B) may be disposed on different regions of the upper surface of the channel layer (200). Each of the source electrode (300A) and the drain electrode (300B) may be disposed on the corresponding In oxide layer (10).
[0055] FIG. 2 is a cross-sectional view showing a reconfigurable synapse element according to another embodiment of the present invention.
[0056] Referring to FIG. 2, in the reconfigurable synapse device according to the present embodiment, the Se-vacancy generation region (20) may have a relatively thinner thickness than the Se-vacancy generation region (20) of FIG. 1. In FIG. 2, the Se-vacancy generation region (20) may be formed thinner than the thickness of the channel layer (200) region below the In oxide layer (10). Thus, an α-In2Se3 region of the channel layer (200) may exist below the Se-vacancy generation region (20). Depending on the total thickness of the channel layer (200) or the oxygen plasma treatment conditions, the thickness of the Se-vacancy generation region (20) may be controlled. In FIG. 2, the remaining configuration excluding the Se-vacancy generation region (20) may be the same or similar as described in FIG. 1.
[0057] FIGS. 3a to 3d are cross-sectional views showing a method for manufacturing a reconfigurable synapse device according to one embodiment of the present invention.
[0058] Referring to FIG. 3a, a channel layer (205) comprising α-In2Se3, a two-dimensional material, can be formed on a predetermined substructure. The channel layer (205) can be formed by a transfer method or a growth method. If necessary, a patterning process may be performed to define the channel layer (205). α-In2Se3 is an In2Se3 material having an α-phase and may have ferroelectricity. α-In2Se3 may be a semiconductor having ferroelectricity. α-In2Se3 may basically have n-type semiconductor characteristics. The channel layer (205) may have a single-layer structure of α-In2Se3 or a multilayer structure in which the single layer is stacked multiple times. In the channel layer (205), the single layer may be stacked several to tens of times. For example, the number of single layers of α-In2Se3 constituting the channel layer (205) may be about 1 to 30 layers.
[0059] The channel layer (205) may be composed of a single two-dimensional material. In other words, the type of two-dimensional material included in the channel layer (205) may be one (i.e., α-In2Se3). When two or more types of two-dimensional materials are bonded (stacked) in multiple layers as in the past, this technology has problems and limitations such as being impossible to apply to actual industrial processes, incompatibility with CMOS processes, and poor reproducibility. However, in the embodiment of the present invention, since a single two-dimensional material is applied to the channel layer (205), the problems and limitations of the technology of bonding two or more types of two-dimensional materials in multiple layers can be overcome.
[0060] According to one embodiment, a gate electrode (105) may be provided within or on a predetermined substrate, and a channel layer (205) may be formed on the gate electrode (105). In this case, the gate electrode (105) may be a bottom gate electrode. A gate insulating layer (155) may be disposed between the gate electrode (105) and the channel layer (205). A gate insulating layer (155) may be formed on the gate electrode (105), and a channel layer (205) may be formed on the gate insulating layer (155). However, in some cases, a top gate structure instead of a bottom gate structure may be applied.
[0061] Referring to FIG. 3b, a mask pattern (255) having openings that expose first and second regions, respectively, can be formed on a channel layer (205). Here, the first and second regions may be regions where a source electrode and a drain electrode are to be formed. The mask pattern (255) may be formed, for example, from a photoresist (PR) material. The mask pattern (255) can be defined by applying a photoresist (PR) material and then patterning it to expose the regions where the source / drain are to be formed. However, the mask pattern (255) may be formed from a material other than photoresist (PR).
[0062] Referring to FIG. 3c, oxygen plasma treatment can be performed on the first and second regions of the channel layer (205) exposed by the mask pattern (255). By performing oxygen plasma treatment on the first and second regions of the channel layer (205), an In oxide layer (15) can be formed in the first and second regions, and together with this, a Se-vacancy generation region (25) having a plurality of Se-vacancies can be formed within the channel layer (205) in contact with the In oxide layer (15).
[0063] The oxygen plasma treatment described above can be performed, for example, in an ICP-RIE (inductively coupled plasma reactive ion etching) device and in an oxygen atmosphere. During the oxygen plasma treatment, the pressure of the vacuum chamber may be hundreds to thousands of mTorr (e.g., 500 mTorr), the oxygen flow rate may be several to tens of sccm (e.g., 10 sccm), the plasma generation power may be several to tens of W (e.g., 6 W), and the treatment time may be about 1 second to tens of minutes (e.g., 30 seconds). However, the equipment and specific process conditions used for the oxygen plasma treatment are not limited to those described above and may vary in various ways.
[0064] The In oxide layer (15) formed in the first region where the source electrode is to be formed and the Se-vacancy generation region (25) below it may be referred to as the first In oxide layer and the first Se-vacancy generation region, respectively. The In oxide layer (15) formed in the second region where the drain electrode is to be formed and the Se-vacancy generation region (25) below it may be referred to as the second In oxide layer and the second Se-vacancy generation region, respectively.
[0065] The In oxide layer (15) can be formed by oxidation by oxygen plasma. The In oxide layer (15) may, for example, include In2O3 or be composed of In2O3. The In oxide layer (15) may be a semiconductor and may have higher electrical conductivity than α-In2Se3. The In oxide layer (15) may have a thin thickness of about 10 nm or less or about 7 nm or less. As a non-limiting example, the thickness of the In oxide layer (15) may be about 1 to 7 nm or about 2 to 6 nm. The In oxide layer (15) may be formed on the surface portion of the channel layer (205) and may be formed to protrude somewhat above the surface of the channel layer (205). Additionally, the In oxide layer (15) may be formed to penetrate somewhat into the interior of the channel layer (205).
[0066] The Se-vacancy region (25) can be created as some of the Se elements escape (i.e., volatilize) from the channel layer (205) by oxygen plasma treatment. The Se-vacancy region (25) can be formed during the process of forming the In oxide layer (15). The Se-vacancy region (25) can be placed below the In oxide layer (15). The Se-vacancy region (25) can be in contact with the lower surface of the In oxide layer (15). The Se-vacancy region (25) can be a region where the Se element is reduced by at least about 10% compared to other regions of the channel layer (205). As a non-limiting example, the Se-vacancy region (25) may be a region in which the Se element is reduced by about 10 to 99% compared to other regions of the channel layer (205), a region in which it is reduced by about 10 to 70%, a region in which it is reduced by about 10 to 50%, or a region in which it is reduced by about 10 to 35%. The thickness of the Se-vacancy region (25) may be several Å or more. Depending on the thickness of the channel layer (205) or the conditions of the oxygen plasma treatment, the thickness of the Se-vacancy region (25) may vary. According to one example, the Se-vacancy region (25) may be formed over the entire region of the channel layer (205) below the In oxide layer (15), but is not limited thereto.
[0067] The In oxide layer (15) and the Se-vacancy generation region (25) can mutually perform the role of exhibiting p-type semiconductor characteristics in the channel layer (205) and enabling the synapse device to have reconfigurable characteristics. The In oxide layer (15) has relatively high electrical conductivity and can attract and absorb electrons. In the semiconductor channel region in contact with the In oxide layer (15), the concentration of electrons is relatively low, so p-type semiconductor characteristics may appear. The In oxide layer (15) can cause a p-type doping effect in the channel region in contact with it. The Se-vacancy in the Se-vacancy generation region (25) can act as a type of acceptor. Therefore, the Se-vacancy generation region (25) can have p-type semiconductor characteristics. Electrons can be trapped in the Se-vacancy, and a p-type doping effect may appear. The source / drain junction region or the adjacent region may exhibit p-type semiconductor characteristics due to the In oxide layer (15) and the Se-vacancy generation region (25).
[0068] Surface charge transfer doping (SCTD) effects may occur due to the In oxide layer (15) and Se-vacancy generation region (25) formed by oxygen plasma treatment. A p-type doping effect may occur in the oxygen plasma-treated source / drain junction region or in the adjacent region, while n-type semiconductor characteristics based on the fundamental properties of α-In2Se3 may be maintained in the remaining channel region. In this regard, p-type doping characteristics may occur as the Fermi level of the channel is lowered in the source / drain junction region or in the adjacent region, and the remaining channel region may have n-type semiconductor characteristics. Therefore, the channel layer (205) may have both n-type semiconductor characteristics and p-type semiconductor characteristics.
[0069] In embodiments of the present invention, the surface charge transfer doping (SCTD) effect by oxygen plasma treatment can be utilized. SCTD can induce energy band bending by controlling the type and concentration of dopant adsorbed onto the semiconductor surface. Since SCTD is not only non-destructive but also provides a large surface area to volume ratio, it can be a technique advantageous for low-dimensional nanostructures. While doping techniques including chemical solution exposure, high-k-based remote doping, and doping by chemical adsorption have been proposed, doping using oxygen plasma treatment can be very useful because it is highly compatible with current semiconductor technologies such as CMOS processes.
[0070] Referring to FIG. 3d, a source electrode (305A) electrically connected to the In oxide layer (15) can be formed on the first region of the channel layer (205), and a drain electrode (305B) electrically connected to the In oxide layer (15) can be formed on the second region of the channel layer (205). The source electrode (305A) and the drain electrode (305B) can be formed, for example, by a lift-off method, and the mask pattern (255) of FIG. 3c can be used in the lift-off method. After forming a conductive layer for the source / drain electrodes on the device structure of FIG. 3c, the source electrode (305A) and the drain electrode (305B) can be formed by removing the mask pattern (255) and the portion of the conductive layer placed thereon. However, the specific method of forming the source electrode (305A) and the drain electrode (305B) is not limited to the above and may vary depending on the case.
[0071] The source electrode (305A) and the drain electrode (305B) may comprise a conductive material such as a metal or a metal compound. The source electrode (305A) and the drain electrode (305B) may have a single-layer structure or a multi-layer structure. The source electrode (305A) and the drain electrode (305B) may be disposed on different regions of the upper surface of the channel layer (205). Each of the source electrode (305A) and the drain electrode (305B) may be disposed on the corresponding In oxide layer (15).
[0072] The channel layer (205) may have both n-type semiconductor characteristics and p-type semiconductor characteristics, and in this regard, the synapse device may have ambipolar characteristics and may be configured to be reconfigurable. That is, the synapse device may have both n-type transistor characteristics and p-type transistor characteristics, and may operate as an n-type transistor or as a p-type transistor by controlling gate voltage conditions, drain voltage conditions, etc. In this respect, the synapse device may be a reconfigurable device. The synapse device may be a type of memory transistor. The memory transistor may be a ferroelectric memory transistor and may be referred to as an FeFET or FeSFET. In addition, the synapse device may be a type of neuromorphic device.
[0073] The above synapse device may have a single gate structure. That is, the above synapse device may be a device that controls the characteristics of the channel layer (205) with a single gate electrode (i.e., 105) rather than multiple gate electrodes. Using two or more gate electrodes in a single transistor can be a fatal disadvantage to integration density and may result in reduced process efficiency and mass production capabilities. However, since the synapse device according to the embodiment of the present invention can implement reconfigurable characteristics using a single gate, it may be significantly advantageous in terms of improved integration density, mass production capabilities, and commercialization.
[0074] FIG. 4 is a cross-sectional view illustrating a method for manufacturing a reconfigurable synapse device according to another embodiment of the present invention.
[0075] Referring to FIG. 4, the Se-vacancy region (25) may have a relatively thinner thickness than the Se-vacancy region (25) of FIG. 3d. In FIG. 4, the Se-vacancy region (25) may be formed thinner than the thickness of the channel layer (205) region below the In oxide layer (15). Thus, an α-In2Se3 region of the channel layer (205) may exist below the Se-vacancy region (25). Depending on the total thickness of the channel layer (205) or the oxygen plasma treatment conditions, the thickness of the Se-vacancy region (25) may be controlled.
[0076] FIGS. 5a to 5c are cross-sectional views showing a method for manufacturing a reconfigurable synapse device according to an embodiment of the present invention.
[0077] Referring to FIG. 5a, a channel layer (207) comprising α-In2Se3, a two-dimensional material, can be formed on a predetermined substructure. The channel layer (207) may have a single-layer structure of α-In2Se3 or a multilayer structure in which the single layer is stacked multiple times. A gate electrode (107) may be provided within or on a predetermined substrate, a gate insulating layer (157) may be provided on the gate electrode (107), and a channel layer (207) may be formed on the gate insulating layer (157). In this case, the gate electrode (107) may be a bottom gate electrode. However, in some cases, a top gate structure rather than a bottom gate structure may be applied.
[0078] Referring to FIG. 5b, a mask pattern (257) having openings that expose first and second regions, respectively, can be formed on the channel layer (207). Here, the first and second regions may be regions where a source electrode and a drain electrode are to be formed. The mask pattern (257) may be formed, for example, from a photoresist (PR) material.
[0079] Referring to FIG. 5c, oxygen plasma treatment can be performed on the first and second regions of the channel layer (207) exposed by the mask pattern (257 in FIG. 5b). Through the oxygen plasma treatment, an In oxide layer (17) can be formed in the first and second regions, and together with this, a Se-vacancy generation region (27) having a plurality of Se-vacancies can be formed within the channel layer (207) in contact with the In oxide layer (17). Here, the red sphere region represents a Se-vacancy. In FIG. 5c, for convenience, the In oxide layer (17) is not clearly depicted, but it can be seen that a thin In oxide layer (17) is formed by the oxygen plasma treatment.
[0080] Next, a source electrode (307A) electrically connected to the In oxide layer (17) on the first region of the channel layer (207) and a drain electrode (307B) electrically connected to the In oxide layer (17) on the second region of the channel layer (207) can be formed. The source electrode (307A) and the drain electrode (3057) can be formed, for example, by a lift-off method, and the mask pattern (257) of FIG. 5b can be used in the lift-off method. After forming a conductive layer for the source / drain electrodes on the device structure of FIG. 5b, the source electrode (307A) and the drain electrode (307B) can be formed by removing the mask pattern (257) and the portion of the conductive layer placed thereon.
[0081] The source electrode (307A) and the drain electrode (307B) may include a conductive material such as a metal or a metal compound. The source electrode (307A) and the drain electrode (307B) may have a single-layer structure or a multi-layer structure. For example, the source electrode (307A) and the drain electrode (307B) may have a multi-layer structure including a Ti layer and an Au layer disposed on the Ti layer, but are not limited thereto. The metallic material and layer structure applicable to the source electrode (307A) and the drain electrode (307B) may vary in many ways.
[0082] Conventional reconfigurable neuromorphic / memory devices based on two-dimensional materials have a structure in which two or more types of two-dimensional materials are stacked and bonded, and two or more gate electrodes are used. By using two or more types of two-dimensional materials, they can be fabricated to modulate ions, charges, phases, etc., through two gate electrodes. However, bonding (stacking) two or more types of two-dimensional materials in multiple layers is impossible to apply to actual industrial processes, is incompatible with CMOS processes, and suffers from poor reproducibility; furthermore, two or more gates can be a critical disadvantage regarding integration density. Therefore, conventional reconfigurable neuromorphic devices based on two-dimensional materials may be impossible to apply to actual industries due to circuit complexity and process difficulties.
[0083] However, according to an embodiment of the present invention, since bipolar characteristics can be realized by performing oxygen plasma treatment on a channel layer composed of ferroelectric α-In2Se3, a reconfigurable synapse device can be realized even when a single two-dimensional material and a single gate structure are applied. Therefore, according to an embodiment of the present invention, a reconfigurable synapse device and its manufacturing process can be realized that have excellent performance and are applicable to actual industrialization in terms of ease of process, reproducibility, mass production, and device / circuit structure. In particular, oxygen plasma treatment can be well compatible with general CMOS processes and can be easily performed with a simple process.
[0084] According to an embodiment of the present invention, a neuromorphic device and system can be configured by applying a synapse device according to the above-described embodiment. The neuromorphic device may include a CMOS peripheral circuit connected to the synapse device. The CMOS peripheral circuit may include a pre-synaptic neuron circuit and a post-synaptic neuron circuit, etc. As the configuration of a neuromorphic device and system including a synapse device is well known, a detailed description thereof is omitted.
[0085] In addition, according to an embodiment of the present invention, a method for manufacturing a neuromorphic device is provided, comprising the step of manufacturing a reconfigurable synapse device using a method according to the above-described embodiment. The method for manufacturing the reconfigurable synapse device may be as described with reference to FIGS. 3a to 3d, FIGS. 4, and FIGS. 5a to 5c, etc.
[0086] Figure 6 is a graph showing the photoluminescence (PL) analysis results for the channel layer when doping by oxygen plasma treatment is not performed and when doping is performed. The case where doping by oxygen plasma treatment is not performed is indicated as "w / o doping," and the case where doping by oxygen plasma treatment is performed is indicated as "w doping." Here, the channel layer may be a layer composed of α-In2Se3. The doping may refer to the SCTD effect.
[0087] Referring to Figure 6, it can be seen that when doping by oxygen plasma treatment (w doping) is performed, an In2O3-related peak occurs at approximately 2.0 eV. This may indicate that In oxide (In2O3) was formed by oxygen plasma treatment. Additionally, when doping is performed (w doping), a peak corresponding to a B exciton was identified. From the PL analysis results, it can be confirmed that In oxide (In2O3) was formed and that the n-type characteristics changed to ambipolar characteristics.
[0088] Figures 7 to 9 are graphs showing the results of X-ray photoelectron spectroscopy (XPS) analysis of channel layers when doping by oxygen plasma treatment is not performed (w / o doping) and when it is performed (w doping). Figure 7 is for the element In, Figure 8 is for the element Se, and Figure 9 is for the element O.
[0089] Referring to Figures 7 and 8, it can be seen that for In and Se, the XPS peaks were shifted to high binding energies by oxygen plasma treatment. In may have been oxidized to In2O3. In the case of Se, Se vacancies may have been formed as some elements volatilized and were removed. After doping treatment, peaks corresponding to Se vacancies appeared.
[0090] Referring to Fig. 9, a metal-oxygen (MO) bonding-related peak occurred at 529.2 eV in the O1s graph after doping treatment. This may be due to the formation of In2O3.
[0091] From the XPS analysis results of Figures 7 to 9, it can be confirmed that In oxide (In2O3) and Se vacancies are formed by oxygen plasma treatment.
[0092] FIG. 10 is a graph showing the transfer characteristic of a device (transistor) according to a comparative example. Here, the transfer characteristic may refer to the characteristics of an output signal with respect to an input signal. The device (transistor) according to the comparative example may be a device that has not been treated with oxygen plasma. That is, the device (transistor) according to the comparative example includes a channel layer composed of α-In2Se3 having n-type characteristics, and does not include an In oxide layer and a Se-vacancy generation region.
[0093] Referring to FIG. 10, it can be seen that the device (transistor) according to the comparative example, to which oxygen plasma treatment was not applied, exhibits only n-type transistor characteristics. Such a device may not be reconfigurable.
[0094] FIG. 11 is a graph showing the transfer characteristic of a reconfigurable synapse device according to one embodiment of the present invention. The reconfigurable synapse device according to the embodiment is subjected to oxygen plasma treatment and may have a structure as described in FIG. 1.
[0095] Referring to FIG. 11, it can be seen that the synapse device according to the embodiment to which oxygen plasma treatment is applied possesses both n-type transistor characteristics and p-type transistor characteristics. It can be seen that the α-In2Se3-based material layer, which was an n-type semiconductor, has been transformed into a semiconductor possessing both n-type semiconductor characteristics and p-type semiconductor characteristics. Therefore, the ON current of the synapse device can appear at both negative (-) gate voltage and positive (+) gate voltage. The synapse device can have ambipolar characteristics and can be configured to be reconfigurable. By controlling operating conditions such as gate voltage conditions and drain voltage conditions, the synapse device can operate as an n-type transistor or as a p-type transistor.
[0096] FIG. 12 is a graph showing the weight update characteristics in the excitatory mode of a reconfigurable synapse device according to one embodiment of the present invention.
[0097] FIG. 13 is a graph showing the weight update characteristics in the inhibitory mode of a reconfigurable synapse device according to one embodiment of the present invention.
[0098] Referring to FIGS. 12 and 13, it can be seen that using a reconfigurable synapse device according to an embodiment, potentialization and depression operations can be performed in the excitatory mode and inhibitory mode, respectively. The excitatory mode may have n-type transistor characteristics, and the inhibitory mode may have p-type transistor characteristics. Using a single reconfigurable synapse device, n-type transistor characteristics and p-type transistor characteristics can be implemented as needed.
[0099] FIG. 14 is a graph showing a rough comparison of the channel layer configuration and ON / OFF current ratio of a reconfigurable synapse device according to an embodiment of the present invention and a synapse device based on a conventional two-dimensional material. Here, the X-axis (stack layers) may represent the number of two-dimensional materials constituting the channel layer.
[0100] Referring to FIG. 14, the reconfigurable synapse device according to the embodiment can obtain a relatively high level of ON / OFF current ratio characteristics using only a single two-dimensional material. In the case of conventional synapse devices based on two-dimensional materials, two to four different two-dimensional materials were stacked and used. Furthermore, most conventional synapse devices based on two-dimensional materials used dual gates (indicated by DG), and when dual gates (indicated by DG) were not used, complex and difficult chemical processing steps were required. However, according to the embodiment of the present invention, by applying a single two-dimensional material and a single-gate structure while applying a simple oxygen plasma processing step, a reconfigurable synapse device with excellent performance can be realized. Since the ON / OFF current ratio characteristics for various devices in FIG. 14 may not have been obtained under the same measurement conditions, a fair comparison between them may be difficult.
[0101] A reconfigurable synapse device according to an embodiment of the present invention may be an ambipolar device in which an ON current can appear at a negative (-) gate voltage and a positive (+) gate voltage. The reconfigurable synapse device can be used for synaptic operations and memory operations, and can be utilized as a reconfigurable memory device. Although various existing reconfigurable memory devices based on two-dimensional materials have been fabricated, they are difficult to apply to actual semiconductor processes and may be incompatible with existing processes because they are composed of structures in which at least two to as many as four two-dimensional materials are joined, or dual-gate (double-gate) structures with two gate electrodes. According to an embodiment of the present invention, a channel can be constructed using a single two-dimensional material by applying α-In2Se3, a ferroelectric two-dimensional material, as the channel material. Furthermore, a reconfigurable synapse device can be realized with only a simple oxygen plasma treatment, which is compatible with silicon processes compared to existing device structures.
[0102] A reconfigurable synapse device according to an embodiment of the present invention can be applied to a logic switch or memory cell that operates by engineering energy bands through controlling the number of layers of a two-dimensional material. Furthermore, a neuromorphic device including the reconfigurable synapse device may have operating characteristics identical or similar to commercially available artificial intelligence devices. Since the synapse device does not include a capacitor, it enables higher integration compared to conventional memory cells and can be fabricated in a vertical stacked structure. Additionally, the reconfigurable synapse device and the neuromorphic device including it can be usefully applied in various fields of artificial intelligence. By utilizing a reconfigurable bipolar synapse device, multiple artificial intelligence operations can be performed simultaneously, and multiple application operations can be performed.
[0103] According to the embodiments of the present invention described above, by applying a two-dimensional material, a reconfigurable synapse device with excellent performance and applicability to actual industrialization in terms of ease of process, reproducibility, mass production, and device / circuit structure can be realized. Furthermore, according to the embodiments of the present invention, a neuromorphic device including the above-described reconfigurable synapse device can be realized. According to one embodiment, since ambipolar characteristics can be realized through oxygen plasma treatment while using a single two-dimensional material (α-In2Se3, a ferroelectric material) in the channel layer, a process applicable to actual industrialization (mass production), compatible with silicon processes such as CMOS processes, and having excellent reproducibility can be realized in the manufacturing of the reconfigurable synapse device. Additionally, according to one embodiment, since a synapse device capable of having reconfigurable characteristics even in a single-gate structure can be realized, it may be advantageous for improving integration density and increasing feasibility compared to devices using two or more gates. By applying a reconfigurable synapse element according to the embodiments of the present invention, a neuromorphic element and system having excellent performance can be realized.
[0104] This specification discloses preferred embodiments of the present invention. Although specific terms have been used, they are used merely in a general sense to facilitate the explanation of the technical content of the invention and to aid in understanding the invention, and are not intended to limit the scope of the invention. It is obvious to those skilled in the art that, in addition to the embodiments disclosed herein, other variations based on the technical concept of the present invention are possible. Those skilled in the art will understand that the reconfigurable synapse element, the neuromorphic element including the reconfigurable synapse element, and the method of manufacturing the same according to the embodiments described with reference to FIGS. 1 to 9 and FIGS. 11 to 14 can be variously substituted, changed, and modified within the scope of the technical concept of the present invention. Therefore, the scope of the invention should not be determined by the described embodiments but by the technical concept described in the claims. Explanation of the symbols
[0105] * Explanation of symbols for major parts of the drawing * 10, 15, 17: In oxide layer 20, 25, 27 : Se-Public Occurrence Area 100, 105, 107: Gate electrodes 150, 155, 157: Gate insulation layer 200, 205, 207: Channel layer 255, 257 : Mask pattern 300A, 305A, 307A: Source electrode 300B, 305B, 307B: Drain electrode
Claims
Claim 1 A reconfigurable synaptic device comprising: a channel layer comprising α-In2Se3, a two-dimensional material; a gate electrode for applying an electric field to the channel layer; a source electrode and a drain electrode electrically connected to the channel layer; and an In oxide layer disposed between the channel layer and the source electrode and between the channel layer and the drain electrode. A reconfigurable synapse device comprising: a Se-vacancy generation region in which a plurality of Se-vacancies are generated, provided within the channel layer in contact with the In oxide layer at each of the source electrode and the drain electrode sides; wherein a source-side first In oxide layer disposed between the channel layer and the source electrode and a drain-side second In oxide layer disposed between the channel layer and the drain electrode are respectively disposed with the channel layer between the source electrode and the drain electrode in between, and a first Se-vacancy generation region in contact with the first In oxide layer and a second Se-vacancy generation region in contact with the second In oxide layer are provided, wherein the first In oxide layer is disposed between the source electrode and the first Se-vacancy generation region and the second In oxide layer is disposed between the drain electrode and the second Se-vacancy generation region, and the channel layer has both n-type semiconductor characteristics and p-type semiconductor characteristics, and the synapse device is configured to have ambipolar characteristics and be reconfigurable. Claim 2 In claim 1, the In oxide layer is a reconfigurable synapse element comprising In2O3. Claim 3 In claim 1, the In oxide layer is a reconfigurable synapse element having a thickness of 1 to 3 nm. Claim 4 In claim 1, the Se-vacancy generation region is a reconfigurable synapse element disposed below the In oxide layer. Claim 5 A reconfigurable synapse device according to claim 1, wherein the Se-vacancy generation region is a region in which the Se element is reduced by at least 10% compared to other regions of the channel layer. Claim 6 In claim 1, the channel layer is a reconfigurable synapse device having a monolayer structure of α-In2Se3 or a multilayer structure in which the monolayer is stacked multiple times. Claim 7 In claim 1, the channel layer is a reconfigurable synaptic device composed of a single two-dimensional material. Claim 8 In claim 1, the synapse element is a reconfigurable synapse element having a single gate structure. Claim 9 A neuromorphic device comprising a reconfigurable synapse element as described in any one of claims 1 to 8. Claim 10 A method for manufacturing a reconfigurable synapse device comprising a gate electrode, a source electrode, and a drain electrode, comprising: forming a channel layer comprising a two-dimensional material, α-In2Se3; performing oxygen plasma treatment on first and second regions of the channel layer to form an In oxide layer on the first and second regions and forming a Se-vacancy generation region having a plurality of Se-vacancies within the channel layer in contact with the In oxide layer; A method for manufacturing a reconfigurable synapse device comprising the step of forming a source electrode electrically connected to the In oxide layer on the first region of the channel layer and a drain electrode electrically connected to the In oxide layer on the second region of the channel layer; wherein a source-side first In oxide layer disposed between the channel layer and the source electrode and a drain-side second In oxide layer disposed between the channel layer and the drain electrode are respectively disposed with the region of the channel layer between the source electrode and the drain electrode in between, and a first Se-vacancy generation region in contact with the first In oxide layer and a second Se-vacancy generation region in contact with the second In oxide layer are provided, wherein the first In oxide layer is disposed between the source electrode and the first Se-vacancy generation region and the second In oxide layer is disposed between the drain electrode and the second Se-vacancy generation region, and the channel layer has both n-type semiconductor characteristics and p-type semiconductor characteristics, and the synapse device is configured to have ambipolar characteristics and be reconfigurable. Claim 11 A method for manufacturing a reconfigurable synapse device according to claim 10, wherein a mask pattern having an opening that exposes the first and second regions respectively on the channel layer is formed, and then the oxygen plasma treatment is performed on the first and second regions of the channel layer exposed by the mask pattern. Claim 12 In claim 10, a method for manufacturing a reconfigurable synapse device wherein the In oxide layer comprises In2O3. Claim 13 In claim 10, the method for manufacturing a reconfigurable synapse device wherein the Se-vacancy generation region is disposed below the In oxide layer. Claim 14 A method for manufacturing a reconfigurable synapse device according to claim 10, wherein the channel layer has a monolayer structure of α-In2Se3 or a multilayer structure in which the monolayer is stacked multiple times. Claim 15 In claim 10, a method for manufacturing a reconfigurable synaptic device in which the channel layer is composed of a single two-dimensional material. Claim 16 In claim 10, the synapse element is a method for manufacturing a reconfigurable synapse element having a single gate structure. Claim 17 A method for manufacturing a neuromorphic device comprising the step of manufacturing a reconfigurable synapse device using the method described in any one of claims 10 to 16.
Citation Information
Patent Citations
Field-effect transistor capable of enhancement mode operation
KR1020220166910A
Synaptic device, reservoir computing device including the synaptic device, and reservoir computing method using the computing device
KR1020230012964A
Semiconductor device and electronic device including the same
KR1020240011036A
Synapse device and fabrication method of the same
KR1020240066801A
