KR20210110041A - Photo-neuromorphic device, preparing method of the same, and artificial neural network including the same - Google Patents

Abstract

The present application provides an optical neuromorphic element, which includes a lower electrode, a photoactive layer formed on the lower electrode, and an upper electrode formed on the photoactive layer. The lower electrode and the upper electrode cross each other to form a cross-bar structure, and at least one of the lower electrode and the upper electrode includes a transparent electrode.

Description

Optical neuromorphic device, manufacturing method thereof, and artificial neural network including same

The present application relates to an optical neuromorphic device, a manufacturing method thereof, and an artificial neural network including the same.

The brain contains hundreds of billions of nerve cells, that is, neurons, and neurons can learn and memorize information by exchanging signals with other neurons through synapses. When the sum of the synaptic potentials input through the dendrite is greater than the threshold potential, the neuron generates an action potential to transmit a signal to another neuron through the axon. This is called neuron spiking. When spiking of a neuron occurs, the strength of a signal transmitted from a neuron may vary according to the strength of a synapse connection between a neuron and another neuron. That is, when the synaptic connection strength is adjusted, the strength of a signal transmitted to a neuron may vary, and information may be learned and stored therefrom.

The neuromorphic device is constructed by mimicking the information transmission and processing process of such biological nerve cells, stores a synaptic weight, which is a value corresponding to the synaptic connection strength, in memory, and performs signal processing based on the stored weight. can be done That is, the neuromorphic device can be utilized to implement an intelligent system by mimicking the human brain. In particular, recently, starting with machine learning technology, deep learning technology has been introduced, and the use of neuromorphic technology is wide, such as the appearance of Alpha Go, an artificial intelligence developed using this technology. is getting

Neuromorphic devices can process a lot of information with less energy by processing information in parallel similar to the human brain. In this regard, existing neuromorphic devices mimic the behavior of biological synapses in terms of learning and memory by changing the conductivity using electrical signals. However, electrons move in a solid medium, and since the bandwidth is narrow, there is a limit to the speed of information transmission, and heat is generated by collision between electrons, resulting in large energy loss. In addition, it is difficult to implement a large-scale parallel operation such as a brain due to the occurrence of cross-talk. As such, as the problem of neuromorphic devices using electrical signals has emerged, interest in neuromorphic devices using photons is increasing.

Korean Patent Application Laid-Open No. 10-2019-0136291, which is the background technology of the present application, relates to a neuromorphic synaptic device having a multi-level conductance and an operating method thereof, but the disclosed patent discloses a problem of a neuromorphic device using an electrical signal can't solve

The present application provides an optical neuromorphic device, a manufacturing method thereof, and an artificial neural network including the same in order to solve the problems of the prior art.

However, the technical problems to be achieved by the embodiments of the present application are not limited to the technical problems as described above, and other technical problems may exist.

As a technical means for achieving the above technical problem, a first aspect of the present application, a lower electrode; a photoactive layer formed on the lower electrode; and an upper electrode formed on the photoactive layer; including, wherein the lower electrode and the upper electrode cross each other to form a cross-bar structure, and at least one of the lower electrode and the upper electrode includes a transparent electrode. provide the element.

According to the exemplary embodiment of the present application, a photoactive layer may be formed on an intersection between the lower electrode and the upper electrode, but is not limited thereto.

According to an exemplary embodiment of the present disclosure, an array structure may be formed by including two or more of the lower electrode and the upper electrode, respectively, but is not limited thereto.

According to one embodiment of the present application, the photoactive layer is IGZO, IZO. ITZO, HIZO, ITO, InP, InGaP, ZTO, ZnO, TIZO, ZnO-Ga ₂ O ₃ , ZnO-Al ₂ O ₃ , SnO ₂ -Sb ₂ O ₃ , and combinations thereof. may be, but is not limited thereto.

According to one embodiment of the present application, the photoactive layer may include an oxygen vacancy, but is not limited thereto.

According to the exemplary embodiment of the present application, the photoactive layer may have a bandgap of 1 eV to 4 eV, but is not limited thereto.

According to the exemplary embodiment of the present application, an insulating layer may be further included between the lower electrode and the photoactive layer or between the upper electrode and the photoactive layer, but is not limited thereto.

According to one embodiment of the present application, the insulating layer is Al ₂ O ₃ , SiO ₂ , SiON, Si ₃ N ₄ , SiC, ZrO ₂ , AlN, Fe ₂ O ₃ , ZnO, MgF ₂ , Ta ₂ O ₅ , TaO _, BN, BST, PZT, and may include one selected from the group consisting of combinations thereof, but is not limited thereto.

According to one embodiment of the present application, the transparent electrode is ITO, ZITO, ZIO, GIO, ZTO, FTO, AZO, GZO, In ₄ Sn ₃ O ₁₂ , Zn _(1-x) Mg _x O (0≤x≤1 ), and may include one selected from the group consisting of combinations thereof, but is not limited thereto.

According to an exemplary embodiment of the present disclosure, a light source disposed toward the transparent electrode may be further included, but is not limited thereto.

A second aspect of the present application, the steps of forming an insulating layer on the lower electrode; forming a photoactive layer on the insulating layer; and forming, on the photoactive layer, an upper electrode crossing the lower electrode to form a cross-bar structure; It provides a method of manufacturing an optical neuromorphic device comprising a.

According to one embodiment of the present application, the photoactive layer may be formed by a sol-gel method, but is not limited thereto.

According to one embodiment of the present application, the photoactive layer may be formed by heat treatment in a vacuum atmosphere, but is not limited thereto.

A third aspect of the present application provides an artificial neural network including the optical neuromorphic device according to the first aspect of the present application.

The above-described problem solving means are merely exemplary, and should not be construed as limiting the present application. In addition to the exemplary embodiments described above, additional embodiments may exist in the drawings and detailed description.

According to the above-described problem solving means of the present application, the optical neuromorphic device according to the present application may have electrical conductivity by light irradiation. That is, unlike the conventional neuromorphic device, it operates by a photocurrent generated by an optical signal rather than an electrical signal. Since it operates by photocurrent induced by photons rather than electrons, energy efficiency is high by reducing heat generation, and large-scale parallel operations can be processed quickly.

When irradiating light to the photoneuromorphic device according to the present application, holes are selectively trapped, and electrons can ascend to the conduction band, so that resistance is lowered, and oxygen vacancies are ionized. Accordingly, the electrical conductivity of the photoactive layer is increased. As described above, a phenomenon in which resistance is lowered and electrical conductivity is increased by light irradiation can be defined as “learning”.

The optical neuromorphic device according to the present disclosure can “remember and retain” the electrical conductivity “learned” by the previous optical signal even after the optical signal is removed.

The optical neuromorphic device according to the present application simulates the structure of a cranial nerve capable of rapidly processing a large amount of information using a small amount of energy. An artificial neural network using this can be utilized as a core technology of artificial intelligence (AI).

However, the effects obtainable herein are not limited to the above-described effects, and other effects may exist.

1 is a cross-sectional view of an optical neuromorphic device according to an embodiment of the present application.
2 is an optical micrograph of an optical neuromorphic device according to an embodiment of the present application.
3 is a schematic diagram of an array structure using an optical neuromorphic device according to an embodiment of the present application.
4 is a comparative graph of the photoelectric characteristics of the optical neuromorphic device according to Examples and Comparative Examples of the present application.
5 is a graph of photoelectric characteristics of a neuromorphic device according to an embodiment of the present application.
6 is a graph of photoelectric characteristics of a neuromorphic device according to an embodiment of the present application.
7 is a graph of photoelectric characteristics of a neuromorphic device according to an embodiment of the present application.
8 is a graph of photoelectric characteristics of a neuromorphic device according to an embodiment of the present application.
9 is a graph illustrating a memory function of an optical neuromorphic device according to an embodiment of the present application.
10 is a graph illustrating a memory function of an optical neuromorphic device according to an embodiment of the present application.
11 is a graph related to symmetric firing time-based plasticity of an optical neuromorphic device according to an embodiment of the present disclosure.

Hereinafter, embodiments of the present application will be described in detail with reference to the accompanying drawings so that those of ordinary skill in the art to which the present application pertains can easily implement them. However, the present application may be embodied in several different forms and is not limited to the embodiments described herein. And in order to clearly explain the present application in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Throughout this specification, when a part is said to be "connected" with another part, it includes not only the case where it is "directly connected" but also the case where it is "electrically connected" with another element interposed therebetween. do.

Throughout this specification, when it is said that a member is positioned "on", "on", "on", "under", "under", or "under" another member, this means that a member is positioned on the other member. It includes not only the case where they are in contact, but also the case where another member exists between two members.

Throughout this specification, when a part "includes" a certain component, it means that other components may be further included, rather than excluding other components, unless otherwise stated.

As used herein, the terms "about," "substantially," and the like are used in a sense at or close to the numerical value when the manufacturing and material tolerances inherent in the stated meaning are presented, and to aid in the understanding of the present application. It is used to prevent an unconscionable infringer from using the mentioned disclosure in an unreasonable manner. Also, throughout this specification, "step to" or "step to" does not mean "step for".

Throughout this specification, the term "combination of these" included in the expression of the Markush form means one or more mixtures or combinations selected from the group consisting of the components described in the expression of the Markush form, and the components It is meant to include one or more selected from the group consisting of.

Throughout this specification, reference to “A and/or B” means “A, B, or A and B”.

Throughout this specification, "cross-bar structure" means a structure in which another electrode is disposed on one electrode so that bar-shaped electrodes cross each other.

Hereinafter, the optical neuromorphic device of the present application will be described in detail with reference to embodiments, examples, and drawings. However, the present application is not limited to these embodiments and examples and drawings.

As a technical means for achieving the above technical problem, a first aspect of the present application, a lower electrode; a photoactive layer formed on the lower electrode; and an upper electrode formed on the photoactive layer; including, wherein the lower electrode and the upper electrode cross each other to form a cross-bar structure, and at least one of the lower electrode and the upper electrode includes a transparent electrode. provide the element.

1 is a cross-sectional view of an optical neuromorphic device according to an embodiment of the present application.

2 is an optical micrograph of an optical neuromorphic device according to an embodiment of the present application.

3 is a schematic diagram of an array structure using an optical neuromorphic device according to an embodiment of the present application.

1 to 3 , the structure consisting of one lower electrode 100 , one photoactive layer 200 , and one upper electrode 300 may operate as one neuromorphic device, as will be described later. , it is also possible to form an array structure including two or more of the lower electrode 100 and the upper electrode 300, respectively, as shown in FIG. 3, and within the array structure, one lower electrode 100 and one photoactive layer ( 200), and a structure consisting of one upper electrode 300 may operate as one neuromorphic device.

According to the exemplary embodiment of the present application, the photoactive layer 200 may be formed on the intersection of the lower electrode 100 and the upper electrode 300 , but is not limited thereto.

In the photoneuromorphic device according to the present disclosure, the photoactive layer 200 may have electrical conductivity by light irradiation. That is, unlike the conventional neuromorphic device, it operates by a photocurrent generated by an optical signal rather than an electrical signal. Since it operates by photocurrent induced by photons rather than electrons, the amount of heat generated is reduced and energy efficiency is high. In addition, information transmission speed can be improved, and large-scale parallel operations can be quickly performed.

Specifically, by irradiating light to the photoactive layer 200 , as electrons in the photoactive layer 200 move, holes may be generated to generate a photocurrent.

According to the exemplary embodiment of the present application, the photoactive layer 200 may include an oxygen vacancy, but is not limited thereto.

Since the photoactive layer 200 includes the oxygen vacancies, it may have free electrons in the photoactive layer 200 to maintain electrical neutrality. When oxygen vacancies increase, free electrons also increase, which can improve electrical conductivity. On the other hand, since an excessively large number of oxygen vacancies act as an obstacle in the electron movement path, electron mobility may be lowered, and thus the photoactive layer 200 includes oxygen vacancies to an appropriate degree.

When light is irradiated to the photoactive layer 200 , the holes are trapped on the interface, and electrons can move to a conduction band, so that the current of the photoactive layer 200 increases, and resistance this lowers Also, the oxygen vacancies are ionized. Accordingly, the electrical conductivity of the photoactive layer 200 is increased. As described above, a phenomenon in which resistance is lowered and electrical conductivity is increased by light irradiation can be defined as a “learning” step.

Specifically, an interface defect is formed on the interface between the photoactive layer 200 and the upper electrode 300 or the photoactive layer 200 and the insulating layer (not shown) due to a difference in the atomic structure of each layer. By irradiating light to the photoneuromorphic device according to the present application, holes and electrons are generated, and the holes are selectively trapped on the interfacial defect, and the electrons are transferred to a conduction band. Since it can move, the electrical conductivity is improved.

In this regard, the photoactive layer 200 is an n-type semiconductor in which a majority carrier is an electron, and since most of the interfacial defects are present in the vicinity of a valence band on a band gap, the electrons can easily migrate to the conduction band, and the holes can be trapped on the interfacial defects. The detrapment time of the trapped holes may be different depending on the activation energy of the photoactive layer 200 and the size of the bandgap.

According to one embodiment of the present application, the photoactive layer 200 is IGZO, IZO. ITZO, HIZO, ITO, ZTO, ZnO, TIZO, ZnO-Ga ₂ O ₃ , ZnO-Al ₂ O ₃ , SnO ₂ -Sb ₂ O ₃ , and may include one selected from the group consisting of combinations thereof. , but is not limited thereto.

According to the exemplary embodiment of the present application, the photoactive layer 200 may have a bandgap of 1 eV to 4 eV or more, but is not limited thereto. In this regard, when the bandgap is too small, the region where holes can be trapped on the bandgap is quickly saturated, and the untrapped holes recombine with electrons, effectively increasing electrical conductivity. can't do it

The wavelength of the irradiated light may be adjusted according to the size of the band gap of the photoactive layer 200 .

The photoactive layer 200 may maintain the electrical conductivity formed by the previous optical signal even when the optical signal is removed. This can be defined as the step of "remembering and retaining" information.

Preferably, the photoactive layer 200 may be IGZO. IGZO has a bandgap of about 3 eV or more, and the photoneuromorphic device according to the present application uses a material having a rather large bandgap, so that the reverse reaction in which electrons are refilled in holes generated by light irradiation may not easily occur. Accordingly, the "learned" electrical conductivity can be "remembered and maintained".

According to the exemplary embodiment of the present application, an array structure may be formed by including two or more of the lower electrode 100 and the upper electrode 300, respectively, but is not limited thereto.

The optical neuromorphic device according to the present disclosure is a two-terminal device and is suitable for implementing a three-dimensional integrated device by forming an array structure in the crossbar shape. Since the crossbar structure intersects a plurality of rows and columns, parallel operations, for example, vector multiplication operations, can be efficiently processed.

In addition, with such a structure, two or more photoactive layers 200 can be adjacent to each other and share one electrode as shown in FIG. 3 , and each photoactive layer 200 is a pre-synaptic neuron and It can be defined as a post-synaptic neuron. A symmetric spike-timing-dependent plasticity (Symmetric STDP) function may be implemented by such a structure.

Spike-timing-dependent plasticity (STDP) is when a synapse between a neuron and a neuron receives a signal in the brain, and the synaptic weight (strength) changes according to the time difference between a signal and the next signal means to change.

In this regard, in the photoneuromorphic device according to the present application, when light is removed while being irradiated with light, the electrical conductivity is gradually decreased. Then, the final electrical conductivity can be adjusted depending on whether the light is irradiated quickly or late. In the optical neuromorphic device, symmetric firing time-based plasticity (Symmetric STDP) means that in optical signals 1 and 2, which are different optical signals, optical signal 1 is applied and then optical signal 2 is applied. It means that the tendency seen when the order is changed and the tendency seen when the second optical signal is applied and then the first optical signal is applied is the same.

According to one embodiment of the present application, it may further include a light source (not shown) disposed toward the transparent electrode, but is not limited thereto.

The light from the light source may be transmitted through the transparent electrode 100 or 300 to be irradiated onto the photoactive layer 200 .

According to one embodiment of the present application, the light may be irradiated with ultraviolet and visible light, but is not limited thereto. Light having a different wavelength range may be selected according to the band gap of the photoactive layer 200 . Preferably, light having an energy similar to or greater than the band gap of the photoactive layer 200 may be irradiated.

As the bandgap of the photoactive layer 200 increases, the light energy required for the “learning” may increase. As the light energy increases, photocurrent characteristics may be increased.

The light may be irradiated in the form of a pulse, and the pulse may be a square pulse, a sawtooth pulse, a sine pulse, or a ramp pulse, but is not limited thereto.

According to one embodiment of the present application, an insulating layer (not shown) may be further included between the lower electrode 100 and the photoactive layer 200 or between the upper electrode 300 and the photoactive layer 200 . However, the present invention is not limited thereto.

By including the insulating layer, the holes generated by light irradiation can be easily trapped. By including the insulating layer, the interfacial defect may be effectively formed and the amount of holes trapped may be increased.

According to one embodiment of the present application, the insulating layer is Al ₂ O ₃ , SiO ₂ , SiON, Si ₃ N ₄ , SiC, ZrO ₂ , AlN, Fe ₂ O ₃ , ZnO, MgF ₂ , Ta ₂ O ₅ , TaO _, BN, BST, PZT, and may include one selected from the group consisting of combinations thereof, but is not limited thereto.

According to one embodiment of the present application, the transparent electrode is ITO, ZITO, ZIO, GIO, ZTO, FTO, AZO, GZO, In ₄ Sn ₃ O ₁₂ , Zn _(1-x) Mg _x O (0≤x≤1 ), and may include one selected from the group consisting of combinations thereof, but is not limited thereto.

A second aspect of the present application, the steps of forming an insulating layer on the lower electrode; forming a photoactive layer on the insulating layer; and forming, on the photoactive layer, an upper electrode crossing the lower electrode to form a cross-bar structure; It provides a method of manufacturing an optical neuromorphic device comprising a.

With respect to the method of manufacturing an optical neuromorphic device according to the second aspect of the present application, detailed descriptions of parts overlapping with the first aspect of the present application are omitted, but even if the description is omitted, the contents described in the first aspect of the present application are The same can be applied to the second aspect of the present application.

According to one embodiment of the present application, the photoactive layer 200 may be formed by a sol-gel method, but is not limited thereto.

In the photoneuromorphic device according to the present disclosure, photoelectric properties may be expressed differently depending on a manufacturing method of forming the photoactive layer 200 .

There is an advantage that the photoactive layer 200 can be manufactured in a large area by using the sol-gel method.

According to one embodiment of the present application, the photoactive layer 200 may be formed by heat treatment in a vacuum atmosphere, but is not limited thereto.

In the sol-gel process, heat treatment or photo-annealing using extreme ultraviolet rays may be used. Preferably, the photoactive layer 200 capable of easily photoreacting and generating an effective photocurrent may be manufactured by heat treatment in a vacuum atmosphere, but is not limited thereto. Since oxygen supply is limited by the heat treatment in the vacuum atmosphere, the oxygen vacancies may be effectively formed on the photoactive layer 200 .

According to one embodiment of the present application, the lower electrode 100 may be formed on a separate substrate before the step of forming an insulating layer (not shown) on the lower electrode 100 , but is not limited thereto. For example, the substrate may be made of glass, plastic, silicon, or the like.

In the method of manufacturing an optical neuromorphic device according to the present application, the insulating layer (not shown) has an advantage that it can be simply formed through heat treatment without a deposition or coating process.

A third aspect of the present application provides an artificial neural network including the optical neuromorphic device according to the first aspect of the present application.

With respect to the artificial neural network according to the third aspect of the present application, detailed descriptions of parts overlapping with the first and/or second aspects of the present application are omitted, but even if the description is omitted, the first aspect and/or the first aspect of the present application The contents described in the second aspect may be equally applied to the third aspect of the present application.

The optical neuromorphic device according to the present application simulates the structure of a cranial nerve capable of rapidly processing a large amount of information using a small amount of energy. An artificial neural network using this can be utilized as a core technology of artificial intelligence (AI).

The present invention will be described in more detail through the following examples, but the following examples are for illustrative purposes only and are not intended to limit the scope of the present application.

[Example 1]

A 3x3 cm ² , 0.7 t glass substrate on which nothing was deposited was washed. Then, a shadow mask made of a chromium material in the shape of a lower electrode was placed on the substrate, and aluminum was deposited to a thickness of 30 nm using a thermal evaporation device. _{Then, an aluminum oxide (Al 2} O ₃ ) layer was formed by oxidizing the surface of the aluminum by heat treatment at 150° C. for 30 minutes on a hot plate.

Then, in order to proceed with the IGZO sol-gel process, a sol-gel solution was first prepared. Specifically, In(NO ₃ ) ₃ ·xH ₂ O (99.999%): Ga(NO ₃ )-·xH ₂ O (99.9%): Zn(NO ₃ ) ₂ · Purchased from Sigma-Aldrich xH ₂ O (99.999%) precursor (solute) was mixed with 2-methoxyethanol solvent in a molar ratio of 6.8:1.0:2.2 to prepare a solution of 0.250 mol. After mixing the solvent and the solute, the mixture was stirred at room temperature for 24 hours or more so that the solute was dissolved in the solvent.

Before coating the prepared sol-gel solution, the substrate on which the aluminum oxide layer was formed was surface-treated with an extreme ultraviolet cleaner (DUV cleaner) for at least 20 minutes so that the solution could be evenly coated. Then, the solution was sprayed on the substrate and spin coating was performed at 4000 rpm for 30 seconds. After spin coating, pre-bake was performed on a hot plate at 60° C. for 2 minutes, and the temperature was raised in the range of 60° C. to 200° C. for 10 minutes and then lowered back to room temperature to perform heat treatment. After the pre-break was finished, heat treatment was performed at 350° C. for 1 hour in a vacuum chamber to induce metal bonding of the IGZO thin film.

In a photolithography process, IGZO was patterned to a size of ^{150 μm 2} with an LCE etching solution (LCE etchant). As the photoresist, positive photoresist purchased from AZ Electronic Materials was used.

Then, a photolithography lift-off process was performed to pattern the upper electrode. A negative photoresist (AZη 5214E) of AZ Electronic Materials was formed on the patterned IGZO, and 30 nm of ITO was deposited thereon by sputtering equipment (RF-sputter). Then, the lift-off process was performed by immersion in acetone solution.

[Example 2]

It was prepared in the same manner as in Example 1 but subjected to extreme ultraviolet (DUV) treatment instead of heat treatment.

[Comparative example]

It was prepared in the same manner as in Example 1, but was heat-treated in the atmosphere instead of in a vacuum chamber.

[Experimental Example 1]

The photoelectric properties of the optical neuromorphic devices prepared in Examples and Comparative Examples were compared.

4 is a comparative graph of the photoelectric characteristics of the optical neuromorphic device according to Examples and Comparative Examples of the present application.

Through this, it can be confirmed that the photoelectric properties are the best in the case of heat treatment in vacuum {Therma (vac.) in Example 1 and FIG. 4), and when extreme ultraviolet (DUV) treatment is performed instead of heat treatment {Example 2, FIG. 4 of DUV}, the photocurrent and photoreactivity are reduced compared to Example 1, and it can be seen that the photocurrent is very low and there is little reactivity to light in the case of heat treatment in the atmosphere (Comparative Example, Thermal (air) of FIG. 4 ).

[Experimental Example 2]

The photoelectric characteristics according to the wavelength of the irradiated light of the optical neuromorphic device prepared in the above example were compared. The light was irradiated with a width of 100 ms of the light source and 60 light pulses.

5 is a graph of photoelectric characteristics of a neuromorphic device according to an embodiment of the present application.

6 is a graph of photoelectric characteristics of a neuromorphic device according to an embodiment of the present application.

7 is a graph of photoelectric characteristics of a neuromorphic device according to an embodiment of the present application.

8 is a graph of photoelectric characteristics of a neuromorphic device according to an embodiment of the present application.

5 is red light (640 nm), FIG. 6 is green light (520 nm), FIG. 7 is blue light (465 nm), and FIG. 8 is ultraviolet light (400 nm).

5 to 8 , the Y-axis PSC is a current value increased by light irradiation when the current value before light irradiation is set to 0. As can be seen from the graph, when red light and green light having relatively low energy are received, since they have a wavelength that cannot sufficiently induce the generation of electron-hole pairs by light irradiation, a relatively small amount of current increases. However, in the case of relatively high energy blue light, especially UV, a large amount of electron-hole pairs are formed in the photoreactive layer (IGZO layer), resulting in a high PSC current value. Through this, it is possible to control the amount of change in the PSC of the device by adjusting the type of irradiated light (by wavelength vs. each), and it can be seen that it can be applied to various neuromorphic devices such as the degree of memory by using this.

[Experimental Example 3]

In order to confirm the "memory" ability of the optical neuromorphic device manufactured in the above example, a light pulse having a width of 100 ms was applied and then removed to check the photoelectric characteristics according to time.

9 is a graph showing the memory capacity of the optical neuromorphic device according to an embodiment of the present application.

Specifically, in FIG. 9 , it can be confirmed that the short-term memory function is implemented as a result of continuously adding and then removing 10 optical pulses having a frequency of 0.5 Hz.

10 is a graph illustrating a memory function of an optical neuromorphic device according to an embodiment of the present application.

Specifically, in FIG. 10 , it can be confirmed that the long-term memory function is implemented as a result of continuously adding and removing 120 optical pulses having a frequency of 6 Hz.

Referring to FIGS. 9 and 10 , it can be seen that the synaptic connection strength is controlled according to the amount and frequency of light irradiated for the same time. As shown in FIG. 9 , when the frequency of light irradiation is low, the current variation is not maintained high, and as in FIG. 10 , when the frequency of light irradiation is high, it can be confirmed that the current variation is maintained high. Through this, the optical neuromorphic device according to the present application is similar to the principle of short-term memory (STM) and long-term memory (LTM) of a living body when a strong signal of high frequency is given in a short time. It can be seen that the memory is induced.

[Experimental Example 4]

The symmetric firing time-based plasticity (Symmetric STDP) characteristics of the optical neuromorphic device manufactured in the above example were confirmed.

11 is a graph related to symmetric firing time-based plasticity of an optical neuromorphic device according to an embodiment of the present application.

Referring to FIG. 11 , the x-axis is the time interval Δt between optical signal 1 and optical signal 2 applied to the optical neuromorphic device, and it can be confirmed that current flow can be controlled by adjusting Δt. That is, it can be seen that the optical neuromorphic device has a symmetric firing time-based plasticity (Symmetric STDP) characteristic.

The foregoing description of the present application is for illustration, and those of ordinary skill in the art to which the present application pertains will understand that it can be easily modified into other specific forms without changing the technical spirit or essential features of the present application. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive. For example, each component described as a single type may be implemented in a distributed manner, and likewise components described as distributed may be implemented in a combined form.

The scope of the present application is indicated by the following claims rather than the above detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalents should be construed as being included in the scope of the present application.

100: lower electrode
200: photoactive layer
300: upper electrode

Claims (14)

lower electrode;
a photoactive layer formed on the lower electrode; and
an upper electrode formed on the photoactive layer;
including,
The lower electrode and the upper electrode cross each other to form a cross-bar structure,
At least one of the lower electrode and the upper electrode comprises a transparent electrode,
Optical neuromorphic device.
The method of claim 1,
An optical neuromorphic device, wherein a photoactive layer is formed on the intersection of the lower electrode and the upper electrode.
The method of claim 1,
An optical neuromorphic device comprising two or more of each of the lower electrode and the upper electrode to form an array structure.
The method of claim 1,
The photoactive layer is IGZO, IZO. group consisting of ITZO, HIZO, ITO, InP, InGaP, ZTO, ZnO, TIZO, ZnO-Ga ₂ O ₃ , ZnO-Al ₂ O ₃ , SnO ₂ -Sb ₂ O ₃ , a-Si, and combinations thereof Including those selected from, optical neuromorphic device.
The method of claim 1,
The photoactive layer will include an oxygen vacancy (oxygen vacancy), an optical neuromorphic device.
The method of claim 1,
The photoactive layer will have a bandgap of 1 eV to 4 eV or more, an optical neuromorphic device.
The method of claim 1,
The optical neuromorphic device further comprising an insulating layer between the lower electrode and the photoactive layer or between the upper electrode and the photoactive layer.
8. The method of claim 7,
The insulating layer is Al ₂ O ₃ , SiO ₂ , SiON, Si ₃ N ₄ , SiC, ZrO ₂ , AlN, Fe ₂ O ₃ , ZnO, MgF ₂ , Ta ₂ O ₅ , TaO _, BN, BST, PZT, and An optical neuromorphic device comprising one selected from the group consisting of combinations thereof.
The method of claim 1,
The transparent electrode is ITO, ZITO, ZIO, GIO, ZTO, FTO, AZO, GZO, In ₄ Sn ₃ O ₁₂ , Zn ( _1-x) Mg _x O (0≤x≤1), and combinations thereof The optical neuromorphic device comprising one selected from the group consisting of.
The method of claim 1,
The optical neuromorphic device further comprising a light source disposed toward the transparent electrode.
forming a photoactive layer on the lower electrode; and
forming an upper electrode intersecting the lower electrode on the photoactive layer to form a cross-bar structure;
containing,
A method for manufacturing an optical neuromorphic device.
12. The method of claim 11,
The photoactive layer is formed by a sol-gel (sol-gel) method, the method of manufacturing an optical neuromorphic device.
13. The method of claim 12,
The photoactive layer is a method of manufacturing an optical neuromorphic device that is formed by heat treatment in a vacuum atmosphere.
According to any one of claims 1 to 11, comprising the optical neuromorphic device according to any one of claims, an artificial neural network.

Images