AI system

Abstract

To provide a system that creates a net list from circuit diagrams or documents indicative of circuit configurations.SOLUTION: An AI system includes a first electronic instrument. The first electronic instrument includes: an input / output interface; a control unit; and a first conversion unit. The input / output interface is electrically connected to the control unit. The first conversion unit is electrically connected to the control unit. The input / output interface includes a function for transmitting, to the control unit, input data generated by operation of the user. The control unit includes a function for transmitting the input data to the first conversion unit. Therein, the input data means is a circuit diagram in which a circuit configuration is drawn, or a document file indicative of the circuit configuration. The first conversion unit includes a circuit configured with a neural network, and is configured to use the neural network of the first conversion unit to convert the input data to a net list.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] One aspect of the present invention relates to an AI system and a method of operating an AI system.

[0002] Note that one embodiment of the present invention is not limited to the above technical field. The technical field of the invention disclosed in this specification relates to an object, a method, or a manufacturing method. Alternatively, one embodiment of the present invention relates to a process, a machine, a manufacture, or a composition of matter. Therefore, examples of the technical field of one embodiment of the present invention disclosed in this specification more specifically include a semiconductor device, a display device, a liquid crystal display device, a light-emitting device, a power storage device, an imaging device, a storage device, a signal processing device, a processor, an electronic device, a system, a driving method thereof, a manufacturing method thereof, or an inspection method thereof. [Background technology]

[0003] An artificial neural network (hereafter referred to as a neural network) is an information processing system modeled on a neural network. It is expected that the use of neural networks will lead to the realization of computers with higher performance than conventional von Neumann-type computers, and in recent years, various research projects have been conducted to build neural networks on electronic circuits.

[0004] For example, Patent Document 1 discloses a control system that converts the charging characteristics of a secondary battery into image data and uses a convolutional neural network (CNN) to distinguish between normal and abnormal characteristics of the secondary battery from the image data. Also, for example, Patent Document 2 discloses a system that analyzes literature data using a neural network or the like. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] International Publication No. 2019 / 021095 [Patent Document 2] JP 2018-49430 A Summary of the Invention [Problem to be solved by the invention]

[0006] When manufacturing electronic devices, semiconductor devices, semiconductor wafers, etc., for example, their specifications are determined in advance, and circuit diagrams are created based on the specifications. However, even if the specifications are the same, the direction and arrangement of wiring, circuit elements, etc. may be designed differently depending on the creator of the circuit diagram. For this reason, even if the specifications and circuit configuration are the same, the appearance of the circuit diagram may differ.

[0007] In other words, even if the specifications and circuit configuration are the same, there can be many different circuit diagrams depending on how the circuit is represented. For this reason, when performing an image search in a database using image recognition processing such as AI (Artificial Intelligence) with a circuit diagram as an input image, even if a circuit with the same specifications and circuit configuration exists in the database, the circuit in the database may not be output in the image search results because its appearance may differ from the input image.

[0008] An object of one embodiment of the present invention is to provide an AI system that converts an image or document showing a circuit configuration into a netlist.Another object of one embodiment of the present invention is to provide an AI system that can search for a circuit configuration.Another object of one embodiment of the present invention is to provide a novel AI system.Another object of one embodiment of the present invention is to provide an operation method of the novel AI system.

[0009] The problems of one embodiment of the present invention are not limited to the problems listed above. The problems listed above do not preclude the existence of other problems. The other problems are problems not mentioned in this section, which will be described below. Problems not mentioned in this section can be derived by a person skilled in the art from the description in the specification or drawings, and can be appropriately extracted from these descriptions. One embodiment of the present invention solves at least one of the problems listed above and other problems. One embodiment of the present invention does not need to solve all of the problems listed above and other problems. [Means for solving the problem]

[0010] (1) One aspect of the present invention is an AI system having a first electronic device, the first electronic device having an input / output interface, a control unit, and a first conversion unit. The input / output interface is electrically connected to the control unit, and the first conversion unit is electrically connected to the control unit. The input / output interface has a function of transmitting input data generated by a user's operation to the control unit, and the control unit has a function of transmitting the input data to the first conversion unit. The first conversion unit has a circuit in which a neural network is configured, and the first conversion unit has a function of converting the input data into a first netlist by the neural network. The input data is a circuit diagram depicting a circuit configuration, or a document file showing the circuit configuration.

[0011] (2) Alternatively, in one aspect of the present invention, in the configuration of (1) above, the first electronic device may have a first database and a second database. The first database is electrically connected to the control unit, and the second database is electrically connected to the control unit. The first database stores a second netlist, and the second database stores literature data linked to the second netlist. The control unit has a function of searching the first database for a circuit configuration of the first netlist, and a function of reading out the literature data from the second database and outputting it to the input / output interface when the second netlist is found in the search for the circuit configuration of the first netlist.

[0012] (3) Alternatively, in one aspect of the present invention, in the configuration of (1) above, the electronic device may include a second electronic device, the first electronic device may include an external interface, and the second electronic device may include a third database and a fourth database. The third database is electrically connected to the external interface, the fourth database is electrically connected to the external interface, the third database stores a second netlist, and the fourth database stores literature data linked to the second netlist. The control unit has a function of communicating with the second electronic device via the external interface and searching the third database for a circuit configuration of the first netlist, and a function of reading out literature data from the fourth database and outputting the literature data to the input / output interface when the second netlist is found in the third database in the search for the circuit configuration of the first netlist.

[0013] (4) Alternatively, one aspect of the present invention is an AI system having a first electronic device and a second electronic device, the first electronic device having an input / output interface, a control unit, and an external interface, and the second electronic device having a second conversion unit. The input / output interface is electrically connected to the control unit, and the external interface is electrically connected to the control unit and the second conversion unit of the second electronic device. The input / output interface has a function of transmitting input data generated by a user's operation to the control unit, and the control unit has a function of transmitting the input data to the second conversion unit of the second electronic device via the external interface. The second conversion unit has a circuit in which a neural network is configured, the second conversion unit has a function of converting the input data into a first netlist by the neural network, and the control unit has a function of acquiring the first netlist from the second electronic device via the external interface. The input data is a circuit diagram depicting a circuit configuration, or a document file showing the circuit configuration.

[0014] (5) Alternatively, in one aspect of the present invention, in the configuration of (4) above, the second electronic device may have a third database and a fourth database. The third database is electrically connected to the external interface, and the fourth database is electrically connected to the external interface. The third database stores the second netlist, and the fourth database stores literature data linked to the second netlist. The control unit has a function of communicating with the second electronic device via the external interface and searching the third database for a circuit configuration of the first netlist, and a function of reading out literature data from the fourth database and outputting it to the input / output interface when the second netlist is found in the third database in the search for the circuit configuration of the first netlist.

[0015] (6) Alternatively, one aspect of the present invention is a method for operating an AI system having an input / output interface, a control unit, and a first conversion unit. The first conversion unit has a circuit in which a neural network is configured, the input / output interface is electrically connected to the control unit, and the first conversion unit is electrically connected to the control unit. The method for operating the AI ​​system has first to third steps. The first step includes a step of inputting input data created by a user to the control unit, the second step includes a step of converting the input data into a first netlist by the neural network of the first conversion unit, and the third step includes a step of outputting the input data to the input / output interface via the control unit.

[0016] (7) Alternatively, the operating method of (6) above, which is one aspect of the present invention, may include steps 4 to 6. The AI ​​system includes a first database and a second database, the first database being electrically connected to the control unit, and the second database being electrically connected to the control unit. The first database stores a second netlist, and the second database stores literature data linked to the second netlist. The fourth step includes a step of searching the first database for a circuit configuration of the first netlist, the fifth step includes a step of reading the literature data from the second database and outputting it to the input / output interface when the second netlist is found from the first database in the fourth step, and the sixth step includes a step of the control unit outputting information that the first netlist was not found from the first database to the input / output interface when the second netlist is not found from the first database in the fourth step.

[0017] In this specification and the like, a semiconductor device is a device that utilizes semiconductor characteristics, and refers to a circuit including a semiconductor element (transistor, diode, photodiode, etc.), a device having such a circuit, etc. Also, refers to any device that can function by utilizing semiconductor characteristics. For example, an integrated circuit, a chip including an integrated circuit, and an electronic component that houses a chip in a package are examples of semiconductor devices. Also, memory devices, display devices, light-emitting devices, lighting devices, electronic devices, etc. are themselves semiconductor devices and may have semiconductor devices.

[0018] In addition, when it is stated in this specification that X and Y are connected, the following cases are also disclosed in this specification: when X and Y are electrically connected, when X and Y are functionally connected, and when X and Y are directly connected. Therefore, it is not limited to a specific connection relationship, for example, a connection relationship shown in a figure or text, and it is also disclosed in a figure or text other than the connection relationship shown in the figure or text. X and Y are objects (for example, a device, an element, a circuit, a wiring, an electrode, a terminal, a conductive film, a layer, etc.).

[0019] As an example of a case where X and Y are electrically connected, one or more elements (e.g., a switch, a transistor, a capacitive element, an inductor, a resistive element, a diode, a display device, a light-emitting device, a load, etc.) that enable the electrical connection between X and Y can be connected between X and Y. The switch has a function of controlling on / off. In other words, the switch has a function of being in a conductive state (on state) or a non-conductive state (off state) and controls whether or not a current flows.

[0020] As an example of a case where X and Y are functionally connected, one or more circuits that enable the functional connection between X and Y (for example, logic circuits (inverters, NAND circuits, NOR circuits, etc.), signal conversion circuits (digital-analog conversion circuits, analog-digital conversion circuits, gamma correction circuits, etc.), potential level conversion circuits (power supply circuits (boosting circuits, step-down circuits, etc.), level shifter circuits that change the potential level of a signal, etc.), voltage sources, current sources, switching circuits, amplifier circuits (circuits that can increase the signal amplitude or current amount, operational amplifiers, differential amplifier circuits, source follower circuits, buffer circuits, etc.), signal generation circuits, memory circuits, control circuits, etc.) can be connected between X and Y. As an example, even if another circuit is sandwiched between X and Y, if a signal output from X is transmitted to Y, X and Y are considered to be functionally connected.

[0021] In addition, when it is explicitly stated that X and Y are electrically connected, this includes the cases where X and Y are electrically connected (i.e., when they are connected with another element or circuit between them), where X and Y are functionally connected (i.e., when they are functionally connected with another circuit between them), and where X and Y are directly connected (i.e., when they are connected without another element or circuit between them). In other words, when it is explicitly stated that X and Y are electrically connected, this is the same as when it is explicitly stated that they are simply connected.

[0022] Also, for example, it can be expressed as "X, Y, and the source (or first terminal, etc.) and drain (or second terminal, etc.) of the transistor are electrically connected to each other, and are electrically connected in the order of X, the source (or first terminal, etc.) of the transistor, the drain (or second terminal, etc.) of the transistor, and Y." Or, it can be expressed as "The source (or first terminal, etc.) of the transistor is electrically connected to X, the drain (or second terminal, etc.) of the transistor is electrically connected to Y, and X, the source (or first terminal, etc.) of the transistor, the drain (or second terminal, etc.) of the transistor, and Y are electrically connected in this order." Or, it can be expressed as "X is electrically connected to Y via the source (or first terminal, etc.) and drain (or second terminal, etc.) of the transistor, and X, the source (or first terminal, etc.) of the transistor, the drain (or second terminal, etc.) of the transistor, and Y are provided in this connection order." By using expressions similar to these examples to specify the order of connections in a circuit configuration, the source (or first terminal, etc.) and drain (or second terminal, etc.) of a transistor can be distinguished and the technical scope can be determined. Note that these expressions are merely examples and are not limited to these expressions. Here, X and Y are objects (e.g., a device, an element, a circuit, a wiring, an electrode, a terminal, a conductive film, a layer, etc.).

[0023] In addition, even when components that are independent on a circuit diagram are shown as being electrically connected to each other, one component may have the functions of multiple components. For example, when a part of a wiring also functions as an electrode, one conductive film has both the functions of a wiring and an electrode. Therefore, the term "electrical connection" in this specification also includes such a case where one conductive film has the functions of multiple components.

[0024] In addition, in this specification, the term "resistance element" refers to a circuit element, wiring, etc. having a resistance value. Therefore, in this specification, the term "resistance element" refers to wiring having a resistance value, a transistor in which a current flows between the source and drain, a diode, a coil, etc. Therefore, the term "resistance element" can be replaced with terms such as "resistance", "load", and "region having a resistance value", and conversely, the terms "resistance", "load", and "region having a resistance value" can be replaced with terms such as "resistance element". The resistance value can be, for example, preferably 1 mΩ or more and 10 Ω or less, more preferably 5 mΩ or more and 5 Ω or less, and even more preferably 10 mΩ or more and 1 Ω or less. In addition, for example, 1 Ω or more and 1×10 9 It may be set to Ω or less.

[0025] In addition, in this specification, the term "capacitive element" refers to a circuit element having a capacitance value, a region of a wiring having a capacitance value, a parasitic capacitance, a gate capacitance of a transistor, and the like. Therefore, in this specification, the term "capacitive element" refers not only to a circuit element including a pair of electrodes and a dielectric contained between the electrodes, but also to a parasitic capacitance appearing between wirings, a gate capacitance appearing between one of the source or drain of a transistor and a gate, and the like. In addition, the terms "capacitive element", "parasitic capacitance", "gate capacitance", and the like can be replaced with terms such as "capacitance", and conversely, the term "capacitance" can be replaced with terms such as "capacitive element", "parasitic capacitance", and "gate capacitance". In addition, the term "pair of electrodes" in "capacitance" can be replaced with "pair of conductors", "pair of conductive regions", "pair of regions", and the like. The value of the capacitance can be, for example, 0.05 fF or more and 10 pF or less. In addition, it may be, for example, 1 pF or more and 10 μF or less.

[0026] In addition, in this specification, a transistor has three terminals called a gate, a source, and a drain. The gate is a control terminal that controls the conductive state of the transistor. The two terminals that function as a source or a drain are input / output terminals of the transistor. One of the two input / output terminals becomes a source and the other becomes a drain depending on the conductivity type (n-channel type, p-channel type) of the transistor and the level of the potential applied to the three terminals of the transistor. For this reason, in this specification, the terms source and drain can be replaced. In addition, in this specification, when describing the connection relationship of a transistor, the terms "one of the source or drain" (or the first electrode or the first terminal) and "the other of the source or drain" (or the second electrode or the second terminal) are used. Note that, depending on the structure of the transistor, a backgate may be included in addition to the above-mentioned three terminals. In this case, in this specification, one of the gate or the backgate of the transistor may be referred to as the first gate, and the other of the gate or the backgate of the transistor may be referred to as the second gate. Furthermore, in the same transistor, the terms "gate" and "backgate" may be interchangeable. Furthermore, when a transistor has three or more gates, in this specification and the like, the respective gates may be referred to as a first gate, a second gate, a third gate, and so on.

[0027] In this specification and the like, a node can be referred to as a terminal, a wiring, an electrode, a conductive layer, a conductor, an impurity region, etc. depending on a circuit configuration, a device structure, etc. Also, a terminal, a wiring, etc. can be referred to as a node.

[0028] In addition, in this specification and the like, "voltage" and "potential" can be interchanged as appropriate. "Voltage" refers to the potential difference from a reference potential, and if the reference potential is the ground potential (earth potential), for example, "voltage" can be interchanged as "potential." Ground potential does not necessarily mean 0V. Note that potential is relative, and the potential applied to wiring, etc. may change depending on the reference potential.

[0029] Generally, "electric current" is defined as a charge transfer phenomenon (electrical conduction) accompanying the movement of a positively charged body, but the statement "electrical conduction of a positively charged body is occurring" can be rephrased as "electrical conduction of a negatively charged body is occurring in the opposite direction." Therefore, in this specification, unless otherwise specified, "electric current" refers to a charge transfer phenomenon (electrical conduction) accompanying the movement of carriers. The carriers referred to here include electrons, holes, anions, cations, complex ions, etc., and the carriers differ depending on the system through which the current flows (for example, semiconductors, metals, electrolytes, vacuum, etc.). In addition, the "direction of current" in wiring, etc. is the direction in which positive carriers move, and is described as a positive current amount. In other words, the direction in which negative carriers move is the opposite direction to the direction of current, and is expressed as a negative current amount. Therefore, in this specification, etc., unless otherwise specified regarding the positive and negative of the current (or the direction of the current), a statement such as "current flows from element A to element B" can be rephrased as "current flows from element B to element A" etc. Furthermore, statements such as "current is input to element A" can be rephrased as "current is output from element A" or the like.

[0030] In addition, in this specification, the ordinal numbers "first," "second," and "third" are used to avoid confusion of components. Therefore, they do not limit the number of components. Furthermore, they do not limit the order of the components. For example, a component referred to as "first" in one embodiment of this specification may be a component referred to as "second" in another embodiment or in the claims. Also, for example, a component referred to as "first" in one embodiment of this specification may be omitted in another embodiment or in the claims.

[0031] In addition, in this specification, the terms indicating the arrangement, such as "above" and "below", may be used for convenience in order to explain the positional relationship between the components with reference to the drawings. Furthermore, the positional relationship between the components changes as appropriate depending on the direction in which each component is depicted. Therefore, the terms are not limited to those described in the specification, and can be rephrased appropriately depending on the situation. For example, the expression "insulator located on the upper surface of a conductor" can be rephrased as "insulator located on the lower surface of a conductor" by rotating the orientation of the drawing shown by 180 degrees.

[0032] In addition, the terms "above" and "below" do not limit the positional relationship of components to being directly above or below and in direct contact with each other. For example, the expression "electrode B on insulating layer A" does not require that electrode B be formed in direct contact with insulating layer A, and does not exclude the inclusion of other components between insulating layer A and electrode B.

[0033] In addition, in this specification and the like, the terms "film" and "layer" can be interchanged depending on the situation. For example, the term "conductive layer" may be changed to the term "conductive film". Or, for example, the term "insulating film" may be changed to the term "insulating layer". Or, depending on the situation, it is possible to replace the terms "film" and "layer" with other terms without using them. For example, the terms "conductive layer" or "conductive film" may be changed to the term "conductor". Or, for example, the terms "insulating layer" and "insulating film" may be changed to the term "insulating body".

[0034] In addition, the terms "electrode", "wiring", "terminal" and the like in this specification do not limit the functions of these components. For example, "electrode" may be used as a part of "wiring", and vice versa. Furthermore, the terms "electrode" and "wiring" include cases where a plurality of "electrodes" and "wiring" are integrally formed. Furthermore, for example, "terminal" may be used as a part of "wiring" or "electrode", and vice versa. Furthermore, the term "terminal" includes cases where a plurality of "electrodes", "wiring", "terminals", and the like are integrally formed. Therefore, for example, an "electrode" can be a part of a "wiring" or "terminal", and for example, a "terminal" can be a part of a "wiring" or "electrode". Furthermore, the terms "electrode", "wiring", "terminal" and the like may be replaced with terms such as "region" depending on the circumstances.

[0035] In addition, in this specification and the like, terms such as "wiring", "signal line", and "power line" can be interchanged with each other depending on the circumstances. For example, the term "wiring" may be changed to the term "signal line". For example, the term "wiring" may be changed to the term "power line". The opposite is also true, and terms such as "signal line" and "power line" may be changed to the term "wiring". The term "power line" may be changed to the term "signal line". The opposite is also true, and terms such as "signal line" may be changed to the term "power line". The term "potential" applied to the wiring may be changed to the term "signal" depending on the circumstances. The opposite is also true, and terms such as "signal" may be changed to the term "potential".

[0036] In this specification and the like, the impurity of a semiconductor refers to, for example, other than the main component constituting the semiconductor layer. For example, an element with a concentration of less than 0.1 atomic % is an impurity. When an impurity is included, for example, DOS (Density of States) may be formed in the semiconductor, carrier mobility may decrease, crystallinity may decrease, and so on. When the semiconductor is an oxide semiconductor, the impurity that changes the characteristics of the semiconductor may be, for example, a Group 1 element, a Group 2 element, a Group 13 element, a Group 14 element, a Group 15 element, transition metals other than the main component, and in particular, for example, hydrogen (also included in water), lithium, sodium, silicon, boron, phosphorus, carbon, nitrogen, and so on. When the semiconductor is a silicon layer, the impurity that changes the characteristics of the semiconductor may be, for example, oxygen, Group 1 elements other than hydrogen, Group 2 elements, Group 13 elements, Group 15 elements, and so on.

[0037] In this specification and the like, a switch refers to a device that has a function of being in a conductive state (on state) or a non-conductive state (off state) and controlling whether or not a current flows. Alternatively, a switch refers to a device that has a function of selecting and switching a path through which a current flows. As an example, an electrical switch, a mechanical switch, or the like can be used. In other words, the switch is not limited to a specific one as long as it can control a current.

[0038] Examples of electrical switches include transistors (e.g., bipolar transistors, MOS transistors, etc.), diodes (e.g., PN diodes, PIN diodes, Schottky diodes, MIM (Metal Insulator Metal) diodes, MIS (Metal Insulator Semiconductor) diodes, diode-connected transistors, etc.), and logic circuits combining these. When a transistor is used as a switch, the "conductive state" of the transistor refers to a state in which the source electrode and drain electrode of the transistor can be considered to be electrically short-circuited. Also, the "non-conductive state" of the transistor refers to a state in which the source electrode and drain electrode of the transistor can be considered to be electrically cut off. When a transistor is operated simply as a switch, the polarity (conductivity type) of the transistor is not particularly limited.

[0039] An example of a mechanical switch is a switch that uses MEMS (microelectromechanical system) technology. The switch has an electrode that can be moved mechanically, and the movement of the electrode controls whether the switch is conductive or non-conductive.

[0040] In this specification, "parallel" refers to a state in which two straight lines are arranged at an angle of -10° or more and 10° or less. Therefore, it also includes the case of -5° or more and 5° or less. Furthermore, "substantially parallel" or "roughly parallel" refers to a state in which two straight lines are arranged at an angle of -30° or more and 30° or less. Furthermore, "perpendicular" refers to a state in which two straight lines are arranged at an angle of 80° or more and 100° or less. Therefore, it also includes the case of 85° or more and 95° or less. Furthermore, "substantially perpendicular" or "approximately perpendicular" refers to a state in which two straight lines are arranged at an angle of 60° or more and 120° or less. Effect of the Invention

[0041] According to one aspect of the present invention, it is possible to provide an AI system that converts an image or document showing a circuit configuration into a netlist. Alternatively, according to one aspect of the present invention, it is possible to provide an AI system capable of searching for a circuit configuration. Alternatively, according to one aspect of the present invention, it is possible to provide a novel AI system. Alternatively, according to one aspect of the present invention, it is possible to provide an operation method of a novel AI system.

[0042] The effects of one embodiment of the present invention are not limited to the effects listed above. The effects listed above do not preclude the existence of other effects. The other effects are effects not mentioned in this section, which will be described below. Effects not mentioned in this section can be derived by a person skilled in the art from the description in the specification or drawings, and can be appropriately extracted from these descriptions. One embodiment of the present invention has at least one of the effects listed above and other effects. Therefore, one embodiment of the present invention may not have the effects listed above in some cases. [Brief description of the drawings]

[0043] [Figure 1] FIG. 1 is a block diagram showing an example of the system configuration. [Diagram 2] FIG. 2 is a block diagram showing an example of the system configuration. [Diagram 3] FIG. 3 is a flowchart showing an example of the operation of the system. [Figure 4] FIG. 4 is a flowchart showing an example of the operation of the system. [Diagram 5] FIG. 5 is a diagram for explaining the procedure for creating a netlist from a circuit diagram. [Figure 6] FIG. 6 is a diagram for explaining the procedure for creating a netlist from a document file. [Figure 7] FIG. 7 is a block diagram illustrating an example of the operation of the system. [Figure 8] 8A and 8B are diagrams illustrating a hierarchical neural network. [Figure 9]FIG. 9 is a block diagram showing an example of the configuration of an arithmetic circuit. [Figure 10] FIG. 10 is a circuit diagram showing an example of the configuration of a circuit included in the arithmetic circuit. [Figure 11] FIG. 11 is a timing chart showing an example of the operation of the arithmetic circuit. [Figure 12] FIG. 12 is a block diagram showing an example of the configuration of an arithmetic circuit. [Figure 13] FIG. 13 is a circuit diagram showing an example of the configuration of the arithmetic circuit. [Figure 14] FIG. 14 is a circuit diagram showing a configuration example of the arithmetic circuit. [Figure 15] FIG. 15 is a circuit diagram showing an example of the configuration of the arithmetic circuit. [Figure 16] FIG. 16 is a circuit diagram showing an example of an equivalent circuit of the arithmetic circuit of FIG. [Figure 17] FIG. 17 is a block diagram showing an example of the configuration of an arithmetic circuit. [Figure 18] FIG. 18 is a timing chart showing an example of the operation of the arithmetic circuit. [Figure 19] FIG. 19 is a schematic cross-sectional view illustrating the configuration of a semiconductor device. [Figure 20] FIG. 20 is a schematic cross-sectional view illustrating the configuration of a semiconductor device. [Figure 21] 21A to 21C are schematic cross-sectional views illustrating the configuration of a semiconductor device. [Figure 22] 22A and 22B are schematic cross-sectional views illustrating examples of the structure of a transistor. [Diagram 23] FIG. 23 is a schematic cross-sectional view illustrating a configuration example of a semiconductor device. [Figure 24] 24A and 24B are schematic cross-sectional views illustrating examples of the structure of a transistor. [Diagram 25] FIG. 25 is a schematic cross-sectional view illustrating a configuration example of a semiconductor device. [Figure 26] FIG. 26A is a top view showing an example of the configuration of a capacitor, and FIGS. 26B and 26C are cross-sectional perspective views showing the example of the configuration of a capacitor. [Figure 27]FIG. 27A is a top view showing an example of the configuration of a capacitor, FIG. 27B is a cross-sectional view showing the example of the configuration of a capacitor, and FIG. 27C is a cross-sectional perspective view showing the example of the configuration of a capacitor. [Figure 28] FIG. 28A is a diagram for explaining the classification of IGZO crystal structures, FIG. 28B is a diagram for explaining the XRD spectrum of quartz glass, FIG. 28C is a diagram for explaining the XRD spectrum of crystalline IGZO, and FIG. 28D is a diagram for explaining the ultrafine electron beam diffraction pattern of crystalline IGZO. [Figure 29] FIG. 29A is a circuit diagram showing the configuration of a multiplication circuit included in a prototype semiconductor device, and FIG. 29B is an optical microscope photograph of the prototype semiconductor device. [Diagram 30] Figure 30A is a graph showing the source-drain current IDS(VW, VX) of transistor M2 of a multiplication circuit included in a prototype semiconductor device when data equivalent to VW is written into the multiplication circuit and a voltage VX is applied to wiring VX, and Figure 30B is a graph showing the multiplication characteristics of the multiplication circuit included in the prototype semiconductor device calculated from Figure 30A. [Diagram 31] FIG. 31 is a graph showing the temperature dependence of the multiplication characteristics of a multiplication circuit included in a prototype semiconductor device. [Diagram 32] 32A and 32B are graphs showing the change over time in the multiplication characteristics of a multiplication circuit included in a prototype semiconductor device. [Diagram 33] FIG. 33A is a graph showing the multiplication characteristics of a multiplication circuit included in a prototype semiconductor device, and FIG. 33B is a graph showing the degree of variation in the multiplication characteristics when each potential is written to the multiplication circuit included in the prototype semiconductor device. [Diagram 34] FIG. 34 is a graph showing the degree of element-to-element variation in the read current of each of a plurality of multiplier circuits included in a prototype semiconductor device. [Diagram 35] 35A, 35B, 35C, and 35D are graphs showing the degree of element variation in read current in a configuration of multiple multiplication circuits, obtained by Monte Carlo analysis. [Diagram 36]FIG. 36 is a diagram showing an example of a hierarchical artificial neural network model used for calculating the inference accuracy. [Figure 37] FIG. 37 is a circuit diagram illustrating a configuration example of a semiconductor device. [Figure 38] FIG. 38A is a graph showing the result of multiplying the first data by the second data, and FIG. 38B is a graph showing the calculated value according to the number of rows of the memory cell array. [Figure 39] 39A and 39B are histograms showing the variation in the value of the product of the first data and the second data when the variation in transistor characteristics is taken into consideration. [Diagram 40] FIG. 40A is a graph showing the degree of match output from the output layers of a neural network constructed by a circuit simulator and a neural network constructed by a programming language, and FIG. 40B is a graph showing the correlation between the values ​​output from the output layers of a neural network constructed by a circuit simulator and a neural network constructed by a programming language. [Diagram 41] FIG. 41 shows an example of an output waveform from the output layer in a neural network constructed by a circuit simulator. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0044] In an artificial neural network (hereafter referred to as a neural network), the strength of synapses can be changed by providing existing information to the neural network. This process of providing existing information to a neural network and determining the strength of connections is sometimes called "learning."

[0045] In addition, by providing some information to a neural network that has undergone "learning" (with connection strengths determined), it is possible for the network to output new information based on the connection strengths. In this way, the process of outputting new information based on given information and connection strengths in a neural network is sometimes called "inference" or "cognition."

[0046] Examples of neural network models include the Hopfield type and hierarchical type. In particular, a neural network with a multi-layer structure is sometimes called a "deep neural network" (DNN), and machine learning using a deep neural network is sometimes called "deep learning."

[0047] In this specification and the like, the term "metal oxide" refers to an oxide of a metal in a broad sense. Metal oxides are classified into oxide insulators, oxide conductors (including transparent oxide conductors), oxide semiconductors (also referred to as "oxide semiconductors" or simply "OSs"). For example, when a metal oxide is used in an active layer of a transistor, the metal oxide may be referred to as an oxide semiconductor. That is, when a metal oxide can constitute a channel formation region of a transistor having at least one of an amplifying function, a rectifying function, and a switching function, the metal oxide can be referred to as a metal oxide semiconductor, or an OS for short. In addition, when an OS FET or an OS transistor is described, it can be rephrased as a transistor having a metal oxide or an oxide semiconductor.

[0048] In this specification and the like, nitrogen-containing metal oxides may also be collectively referred to as metal oxides. Nitrogen-containing metal oxides may also be referred to as metal oxynitrides.

[0049] In this specification and the like, a configuration shown in each embodiment (or example) can be combined as appropriate with a configuration shown in another embodiment (or example) to form one aspect of the present invention. When multiple configuration examples are shown in one embodiment, the configuration examples can be combined as appropriate with each other.

[0050] In addition, the content (or a part of the content) described in one embodiment (or example) can be applied, combined, or replaced with at least one of another content (or a part of the content) described in that embodiment (or example) and another content (or a part of the content) described in one or more other embodiments (or one or more other examples).

[0051] In addition, the contents described in the embodiments (or examples) refer to the contents described in each embodiment (or example) using various figures, or the contents described using text in the specification.

[0052] Furthermore, a figure (or a part thereof) described in one embodiment (or example) can be combined with another part of that figure, with another figure (or a part thereof) described in that embodiment (or example), and with at least one figure (or a part thereof) described in one or more other embodiment (or one or more other examples) to form even more figures.

[0053] The embodiments (or examples) described in this specification are described with reference to the drawings. However, it is easily understood by those skilled in the art that the embodiments (or examples) can be implemented in many different ways, and that the form and details can be changed in various ways without departing from the spirit and scope of the invention. Therefore, the present invention is not interpreted as being limited to the description of the embodiments (or examples). In the configuration of the invention of the embodiments (or the configuration of the examples), the same reference numerals are used in common between different drawings for the same parts or parts having similar functions, and repeated explanations thereof may be omitted. In addition, in perspective views and the like, the description of some components may be omitted in order to ensure the clarity of the drawings.

[0054] In this specification, when the same symbol is used for multiple elements, particularly when it is necessary to distinguish between them, an identification symbol such as “_1”, "[n]”, "[m,n]”, etc. may be added to the symbol.

[0055] In addition, in the drawings of this specification, the size, layer thickness, or area may be exaggerated for clarity. Therefore, it is not necessarily limited to the scale. Note that the drawings are schematic illustrations of ideal examples, and are not limited to the shapes or values ​​shown in the drawings. For example, it is possible to include variations in signals, voltages, or currents due to noise, or variations in signals, voltages, or currents due to timing deviations.

[0056] In addition, regarding this specification and the like, when In:Ga:Zn = 4:2:3 or in the vicinity thereof, when In is 4 with respect to the total number of atoms, Ga is 1 or more and 3 or less (1 ≤ Ga ≤ 3), and Zn is 2 or more and 4.1 or less (2 ≤ Zn ≤ 4.1). When In:Ga:Zn = 5:1:6 or in the vicinity thereof, when In is 5 with respect to the total number of atoms, Ga is greater than 0.1 and 2 or less (0.1 < Ga ≤ 2), and Zn is 5 or more and 7 or less (5 ≤ Zn ≤ 7). When In:Ga:Zn = 1:1:1 or in the vicinity thereof, when In is 1 with respect to the total number of atoms, Ga is greater than 0.1 and 2 or less (0.1 < Ga ≤ 2), and Zn is greater than 0.1 and 2 or less (0.1 < Zn ≤ 2). When In:Ga:Zn = 5:1:3 or in the vicinity thereof, when In is 5 with respect to the total number of atoms, Ga is 0.5 or more and 1.5 or less (0.5 ≤ Ga ≤ 1.5), and Zn is 2 or more and 4.1 or less (2 ≤ Zn ≤ 4.1). When In:Ga:Zn = 10:1:3 or in the vicinity thereof, when In is 10 with respect to the total number of atoms, Ga is 0.5 or more and 1.5 or less (0.5 ≤ Ga ≤ 1.5), and Zn is 2 or more and 4.1 or less (2 ≤ Zn ≤ 4.1). When In:Zn = 2:1 or in the vicinity thereof, when In is 1 with respect to the total number of atoms, Zn is greater than 0.25 and 0.75 or less (0.25 < Zn ≤ 0.75). When In:Zn = 5:1 or in the vicinity thereof, when In is 1 with respect to the total number of atoms, Zn is greater than 0.12 and 0.25 or less (0.12 < Zn ≤ 0.25). When In:Zn = 10:1 or in the vicinity thereof, when In is 1 with respect to the total number of atoms, Zn is greater than 0.07 and 0.12 or less (0.07 < Zn ≤ 0.12).

[0057] (Embodiment 1) In this embodiment, a system according to an aspect of the present invention will be described.

[0058] Figure 1 shows a system (sometimes called an electronic device) that has the function of converting "circuit diagrams" and "circuit configurations described in the claims" into netlists in AI using a neural network. The system also has the function of searching an existing database using the converted netlist.

[0059] A netlist is data in an electronic circuit or the like that contains connection information for circuit elements, logic circuits, signal conversion circuits, potential level conversion circuits, voltage sources, current sources, switching circuits, amplifier circuits, etc., included in the electronic circuit. Specifically, a netlist is data that describes the connection destinations of terminals of circuit elements, circuits, etc., included in the electronic circuit, and is used in circuit simulators, circuit design software, etc.

[0060] The system SIH shown in Fig. 1 includes an electronic device ED. The electronic device ED includes an input / output interface INTFC, a control unit CTL, a conversion unit PTN, a database DTB1, a database DTB2, and a storage unit MP.

[0061] The input / output interface INTFC is electrically connected to the control unit CTL. When the user uses the system SIH, the input / output interface INTFC has a function of inputting and outputting information between the user and the electronic device ED. Examples of the input / output interface INTFC include display devices such as an organic EL (Electro Luminescence) display and a liquid crystal display, and hardware such as a keyboard and a pointing device (e.g., a mouse). The display device may also have an input device such as a touch panel.

[0062] The memory unit MP is electrically connected to the control unit CTL. The memory unit MP includes a volatile storage device and a non-volatile storage device.

[0063] An example of a volatile storage device is a dynamic random access memory (DRAM). A volatile storage device has a function of temporarily storing data required during a calculation process or while software is running.

[0064] Examples of non-volatile storage devices include HDD (hard disk drive), SSD (solid state drive), optical disk, magnetic tape, etc. In this specification and the like, when optical disk, magnetic tape, etc. are used, a readable and writable device and the optical disk, magnetic tape, etc. are collectively referred to as a non-volatile storage device. A non-volatile storage device has a function of storing, for example, software execution programs, drawings depicting circuit configurations, netlists of circuit configurations, etc.

[0065] The conversion unit PTN is electrically connected to the control unit CTL. The conversion unit PTN has a function of converting a circuit diagram, a document file expressing a circuit in text (e.g., the scope of claims in a patent specification), etc., into a netlist. The conversion unit PTN may be, for example, an arithmetic circuit constituting a neural network. In addition, when a neural network is constituted in the conversion unit PTN, it is assumed that the neural network has already been trained and that the weight coefficients between the neurons contained in the neural network have been determined.

[0066] The database DTB1 is electrically connected to the control unit CTL. The database DTB1 has a function of storing document data such as patent specifications, papers, and materials.

[0067] The database DTB2 is electrically connected to the control unit CTL. The database DTB2 has a function of storing, for example, a netlist of a circuit described in the literature data stored in the database DTB1. The netlist may also include a management number, a management symbol, and the like for linking the netlist with the circuit configuration of the literature data.

[0068] The databases DTB1 and DTB2 may be integrated into one database.

[0069] As described above, the databases DTB1 and DTB2 store literature data, netlists, and the like, and therefore the databases DTB1 and DTB2 may be included in the memory unit MP. In particular, it is preferable that the databases DTB1 and DTB2 are non-volatile storage devices of the memory unit MP.

[0070] 2, the electronic device ED may have an external interface INF. The external interface INF has a function of communicating with an electronic device WSV outside the electronic device ED. Therefore, the external interface INF is electrically connected to the control unit CTL and the electronic device WSV.

[0071] The electronic device WSV may be, for example, an external server, etc. Therefore, it is preferable that the external interface INF is connected to the electronic device WSV via an Internet line or the like.

[0072] The electronic device WSV has, for example, a database WDTB1, a database WDTB2, and a conversion unit WPTN. The database WDTB1 stores literature data, similar to the database DTB1. The database WDTB2 stores netlists of circuits described in the literature data of the database WDTB1, similar to the database DTB2. The conversion unit WPTN has a function of converting circuit diagrams, document files that describe circuits in text, and the like, into netlists, similar to the conversion unit PTN.

[0073] At least one of the database WDTB1, the database WDTB2, and the conversion unit WPTN may function as an external server. By providing at least one of the database WDTB1, the database WDTB2, and the conversion unit WPTN as an external server, it may be possible to increase the scale, storage capacity, calculation capacity, and the like of the provided server. For example, by having the database WDTB1 function as an external server, the database WDTB1 may be able to store more document data than the database DTB1. Also, for example, by having the database WDTB2 function as an external server, the database WDTB2 may be able to store more information such as netlists than the database DTB2. Also, for example, the conversion unit WPTN may be able to have a calculation circuit larger in scale than the conversion unit PTN.

[0074] <<Example 1>> Here, an example of the operation of the system SIH in Fig. 1 will be described. Fig. 3 is a flowchart showing an example of the operation of the system SIH, and the example of the operation of the system SIH has steps STI01 to STI03. In Fig. 3, the start of the example of the operation is written as "START" and the end of the example of the operation is written as "END". In this example of the operation, an operation of converting a document file in which a circuit diagram or a circuit is expressed in text into a netlist will be described.

[0075] Step STI01 includes a step in which a user inputs a circuit diagram or a document file expressing a circuit in text to the control unit CTL of the electronic device ED using the input / output interface INTFC. Examples of means for inputting the circuit diagram by the user include means for creating a circuit diagram using circuit design software, a circuit simulator, paint software, CAD software, etc. Examples of means for inputting the document file by the user include means for creating a document file using document creation software, a text editor, etc. Also, circuit diagrams and document files in the process of being created, created circuit diagrams, document files, etc. may be temporarily stored in the memory unit MP. In this specification, the circuit diagram or document file created in step STI01 is referred to as input data.

[0076] In step STI01, in addition to the created data, circuit diagrams, document files, and the like read from the database DTB1 may be applied as input data.

[0077] Step STI02 includes a step of converting the input data created in step STI01 into a netlist by the conversion unit PTN. Specifically, for example, a user uses the input / output interface INTFC to transmit a signal including input data and an instruction to convert the input data into a netlist to the control unit CTL. Upon receiving the input data and the signal, the control unit CTL transmits the input data to the conversion unit PTN. Upon receiving the input data, the conversion unit PTN converts the input data into a netlist.

[0078] When the input data is a circuit diagram, the method of converting the input data into a netlist is preferably, for example, a method using a convolutional neural network (CNN). When the input data is a document file, the method of converting the input data into a netlist is preferably, for example, a method using a recurrent neural network. Specific examples of the method of converting the input data into a netlist will be described later.

[0079] The converted netlist may be temporarily stored in the memory unit MP.

[0080] Step STI03 includes a step of outputting the netlist converted in step STI02 to a display device or the like included in the input / output interface INTFC. Specifically, for example, the netlist converted in step STI02 is transmitted to a display device or the like included in the input / output interface INTFC via the control unit CTL. Thereafter, the netlist is displayed on the display device or the like, allowing the user to check the contents of the netlist converted from the input data.

[0081] After step STI03 is performed, this operation ends.

[0082] <<Example 2>> Next, an example of operation in the system SIH of Fig. 1, which is different from the flowchart shown in Fig. 3, will be described. Fig. 4 is a flowchart showing an example of operation in the system SIH, which is an example of operation in which steps STI04 to STI08 are further added to the example of operation in Fig. 3. This example of operation will explain a method of searching for a circuit using a converted netlist.

[0083] For steps STI01 to STI03 shown in the flowchart of FIG. 4, the description of steps STI01 to STI03 in the flowchart of FIG. 3 should be referred to.

[0084] Step STI04 includes a step of searching the database DTB2 for a circuit using the netlist converted in step STI02. Specifically, for example, the control unit CTL transmits a signal including an instruction to read the netlist linked to the literature data stored in the database DTB2 to the database DTB2. Then, upon receiving the signal, the database DTB2 reads out the netlist and transmits it to the control unit CTL, and the control unit CTL compares the netlist converted in step STI02 with the netlist included in the database DTB2.

[0085] In addition, the reading of netlists linked to literature data from database DTB2 may target all netlists stored in database DTB2, or conditions may be applied to limit the reading to a portion of netlists stored in database DTB2.

[0086] The search performed in step STI04 may extract a complete match with the netlist converted in step STI02, or may extract a partial match (similar).

[0087] The converted netlist used for the search may be temporarily stored in the memory unit MP.

[0088] Also, the search for the circuit performed in step STI04 may use AI. Specifically, for example, AI may be used to compare the descriptions of the netlist converted in step STI02 and the netlist linked to the literature data stored in database DTB2, calculate a similarity indicating the degree of agreement between the types, numbers, and connection configurations of circuit elements in the two netlists, and output search results in order of the similarity.

[0089] Step STI05 includes a step of judging whether or not the netlist converted in step STI02 is found in database DTB2 in the search of step STI04. The netlist found in database DTB2 here includes a case where the netlist matches completely with the netlist searched in step STI04, and a case where the netlist matches partially. If the judgment shows that the netlist searched in step STI04 is found in database DTB2, the operation proceeds to step STI06. If the judgment shows that the netlist searched in step STI04 is not found in database DTB2, the operation proceeds to step STI07.

[0090] Step STI06 includes a step of reading, from the database DTB1, literature data corresponding to the netlist found in the database DTB2 in the search of step STI04. Specifically, for example, the control unit CTL transmits to the database DTB1 a signal including an instruction to read out literature data corresponding to the netlist retrieved in the search of step STI04 and read out from the database DTB2. Upon receiving the signal, the database DTB1 reads out the literature data and transmits it to the control unit CTL. The control unit CTL transmits the literature data to a display device or the like included in the input / output interface INTFC. Thereafter, the literature data is displayed on the display device or the like, allowing the user to confirm the contents of the literature data.

[0091] The converted netlist used in the search and the document data read out from the database DTB1 may be temporarily stored in the memory unit MP.

[0092] After step STI06 is performed, the operation ends.

[0093] Step STI07 includes a step of outputting a result of the search in step STI04 that the netlist converted in step STI02 was not found in the database DTB2. Specifically, for example, the control unit CTL transmits information to the input / output interface INTFC that the netlist converted in step STI02 was not found in the database DTB2. This allows the user to confirm the search result that the netlist converted in step STI02 was not found in the database DTB2 by displaying the information on the display device or the like.

[0094] Step STI08 includes a step of storing the input data in the database DTB1 and storing the netlist converted in step STI02 in the database DTB2. Specifically, the user uses the input / output interface INTFC to transmit a signal including an instruction to save the input data and the netlist converted in step STI02 to the control unit CTL. By receiving the signal, the control unit CTL transmits the input data and a signal including an instruction to write the input data to the database DTB1 to the database DTB1, and transmits the netlist converted in step STI02 and a signal including an instruction to write the netlist to the database DTB2 to the database DTB2. At this time, the input data transmitted to the database DTB1 and the converted netlist transmitted to the database DTB2 may be transmitted from the storage unit MP in which they are temporarily stored, via the control unit CTL.

[0095] After step STI08 is performed, the operation ends.

[0096] Note that the operation method of one embodiment of the present invention is not limited to the above-mentioned steps STI01 to STI08. In this specification and the like, the processes shown in the flowchart are classified by function and shown as independent steps. However, in actual processing and the like, it is difficult to separate the processes shown in the flowchart by function, and there may be cases where one step is involved in multiple steps, or where one step is involved across multiple steps. Therefore, the processes shown in the flowchart are not limited to the steps described in the specification, and can be appropriately replaced depending on the situation. Specifically, the order of steps can be replaced, steps can be added, deleted, etc. depending on the situation, case, or need. For example, step STI08 may be deleted from this operation example and not performed if the user does not want it.

[0097] The operation method of one aspect of the present invention is not limited to the operation example described in the present embodiment. For example, the operation example of converting input data to a netlist using the conversion unit PTN of the electronic device ED has been described above, but the conversion unit PTN may use the conversion unit WPTN of the electronic device WSV instead. Also, the operation example of searching for a circuit in a netlist using the databases DTB1 and DTB2 of the electronic device ED has been described above, but the databases DTB1 and DTB2 may use the databases WDTB1 and WDTB2 of the electronic device WSV instead.

[0098] In addition, according to one aspect of the present invention, a business model may be provided in which a service for converting input data into a netlist and / or a service for searching for circuits using a netlist is provided for a fee using the electronic device ED and an external electronic device WSV.

[0099] <<Example 1 of how to convert to a netlist>> Next, a method for converting a circuit diagram into a netlist when the input data is a circuit diagram in step STI02 of the above operation example will be described.

[0100] FIG. 5 is a schematic diagram showing the flow of converting a circuit diagram into a netlist.

[0101] Stage PH1 in Fig. 5 shows an example of an image to be input data to the conversion unit PTN in step STI02. A circuit diagram is drawn in the image PIC, and the circuit diagram shows circuit symbols, wiring, and their connection configuration. In some cases, as shown in the image PIC in Fig. 5, the image PIC may include the names of the circuit symbols, characters, symbols, etc. indicating the names of the wiring. Also, the circuit diagram shown in the image PIC in Fig. 5 may not include the names of the circuit symbols, characters, symbols, etc. indicating the names of the wiring.

[0102] Stage PH2 in Fig. 5 shows an example in which the image PIC in stage PH1 is input to the conversion unit PTN, and object area recognition is performed on the image PIC. In the image PIC on the left side of stage PH2 in Fig. 5, circuit symbols and electrical connection parts (corresponding to black circles in the circuit diagram shown in the image PIC) are recognized and are surrounded by dotted lines. Depending on the situation, the object area may also include wiring.

[0103] Examples of the above-mentioned object region recognition method include Objectness, CPMC (Constrained Parametric Min-Cuts), and Object Proposals.

[0104] Next, after the recognition of the object region is completed, image recognition is performed for each object region. This allows the conversion unit PTN to recognize what kind of circuit element the circuit symbol surrounded by the dotted line is. For example, by performing image recognition on the image PIC on the left side of stage PH2 in FIG. 5, the conversion unit PTN can recognize, for example, the circuit symbol surrounded by the thick dashed line as a capacitive element, and can recognize, for example, the circuit symbol surrounded by the thick dashed line as a transistor. By recognizing what circuit element the circuit symbol indicates, the conversion unit PTN can assign to the circuit symbol a name (for example, a letter, an abbreviation, a code, a word, etc.) to be described in the netlist.

[0105] Similarly, by performing image recognition, the conversion unit PTN can recognize electrical connection parts (black circles, etc.) surrounded by dotted lines, which allows the conversion unit PTN to give names (for example, letters, abbreviations, symbols, words, etc.) to be written in a netlist to the electrical connection parts (black circles, etc.).

[0106] One method of performing image recognition is to use AI to have the conversion unit PTN learn circuit symbols as training data in advance, and when an image PIC is given to the conversion unit PTN as input data, the circuit symbols contained in the image PIC are extracted based on the learned circuit symbols.

[0107] In addition, as a method for performing image recognition, for example, a convolutional neural network (CNN) or the like can be used. In addition, when using a convolutional neural network, images of circuit symbols, images of electrical connection parts (black circles, etc.), or parts of these images can be used in advance as filters for the convolutional layer of the convolutional neural network. This makes it possible to calculate the similarity between the circuit symbols, electrical connection parts (black circles, etc.), etc. included in the image PIC and the filter by calculation using the convolutional neural network, and the circuit symbols, electrical connection parts (black circles, etc.), etc. included in the image PIC can be identified from the similarity.

[0108] In addition, when the conversion unit PTN recognizes the connection between the electrical connection part (black circle, etc.) and the circuit symbol, it is preferable to repeatedly perform the recognition of the object area and the image recognition. Specifically, the recognition of the object area and the image recognition are performed once to cause the conversion unit PTN to recognize the electrical connection part (black circle, etc.), and to assign a symbol to the electrical connection part (black circle, etc.) for description in a netlist. Next, the first image recognition is performed to determine in which direction the wiring connected to the electrical connection part (black circle, etc.) extends, and in the second object area recognition, the area is expanded in the direction of the extension, and the wiring and the electrical connection part (black circle, etc.) are recognized as an object area. After that, in the second image recognition, the direction in which the wiring extends is determined again, and from the third time onwards, the recognition of the object area and the image recognition are similarly repeated. This allows the conversion unit PTN to recognize the wiring connected to the electrical connection part (such as a black circle) in accordance with the number of times that the object area recognition and the image recognition are repeated, and ultimately allows the conversion unit PTN to recognize the electrical connection between the circuit symbol and the electrical connection part (such as a black circle). As an example, the image PIC on the left side of stage PH2 in Fig. 5 illustrates an area surrounded by a thick two-dot chain line as the wiring area obtained by repeatedly performing the object area recognition and the image recognition.

[0109] In addition, when the image PIC in step PH1 includes characters and symbols indicating the names of circuit symbols and wiring, those names, characters, symbols, etc. may be recognized together with the circuit symbols and electrical connection parts in the object region recognition stage, as shown in the image PIC on the right side of step PH2 in Fig. 5. This allows those names, characters, symbols, etc. obtained by object region recognition to be linked to the circuit symbols and electrical connection parts that have also been recognized in the object region. In addition, the names, characters, symbols, etc. linked to the circuit symbols and electrical connection parts can be treated as symbols, characters, etc. to be described in the netlist.

[0110] At stage PH3 in FIG. 5, an example is shown in which the connection configuration between the circuit symbols and electrical connection parts (black circles, etc.) recognized by the conversion unit PTN at stage PH2 in FIG. 5 is described in a netlist.

[0111] At the left end of the netlist NTL, the circuit symbols recognized in the image PIC are written as the names CSW of the circuit elements (e.g., letters, abbreviations, codes, words, etc.). For example, Tr[1] and Tr[2] indicate transistors in the circuit diagram depicted in the image PIC, C[1] indicates a capacitive element in the circuit diagram depicted in the image PIC, and EL[1] indicates a light-emitting element in the circuit diagram depicted in the image PIC.

[0112] In addition, in the netlist NTL, the names CNP (e.g., letters, abbreviations, symbols, words, etc.) of electrical connection parts (black circles, etc.) are written to the right of the names CSW of the circuit elements via a space SPC as information indicating the connection configuration of the circuit symbols. When there are multiple terminals of the circuit symbols, it is preferable to provide a space between the names CNP. The order in which the names CNP of the electrical connection parts (black circles, etc.) are written is determined by the terminals of the circuit symbols of the names CSW written in that column. For example, in the netlist NTL, the electrical connections of each terminal of a transistor are written in the order of one of the source or drain, the gate, and the other of the source or drain. For example, in the netlist NTL, the electrical connections of each terminal of a light-emitting element are written in the order of the input terminal, then the output terminal.

[0113] As described above, by performing object area recognition and image recognition, the circuit diagram that is used as input data can be converted into a netlist.

[0114] <<Example 2 of how to convert to a netlist>> Next, a method for converting a circuit diagram into a netlist when input data such as a document file is used in step STI02 of the above operation example will be described.

[0115] FIG. 6 is a schematic diagram showing the flow of converting a document file into a netlist.

[0116] Stage PH4 in Fig. 6 shows an example of a document file as input data to be input to the conversion unit PTN in step STI02. In the document file DOC shown in stage PH4 in Fig. 6, a circuit configuration is described as text as information for conversion to a netlist using the conversion unit PTN.

[0117] The document file DOC expressing the circuit configuration as text may be, for example, a description of the circuit described in a patent specification, the scope of claims accompanying the patent specification, etc. As an example of the conversion method here, the document file DOC is assumed to contain text as shown in Table 1 below.

[0118] [Table 1]

[0119] When the above-mentioned document file DOC is input as input data, the conversion unit PTN performs text analysis on the document file DOC, for example.

[0120] As a method of performing text analysis, for example, it is preferable to use AI to have the conversion unit PTN learn document files (e.g., papers, patent claims described in patent publications, etc.) and netlists corresponding to those document files as training data in advance. This allows the conversion unit PTN to convert from the document file to a netlist based on the learned content when the document file is provided as input data.

[0121] In addition, examples of methods for creating the above-mentioned teaching data include using software to generate multiple document files (e.g., multiple ``claims'' with the same content but different descriptions) from a single netlist.

[0122] In addition, as a method for performing text analysis, for example, a recurrent neural network (RNN) can be used.

[0123] By performing text analysis on the document file DOC, the conversion unit PTN can recognize circuit elements, wiring, or electrical connection points from the circuit configuration shown in the document file DOC. For example, by performing text analysis on the sentence in the first paragraph of the document file DOC, the conversion unit PTN can recognize that the circuit configuration shown in the document file DOC includes a first transistor, a second transistor, a capacitive element, and a light-emitting element. Next, by performing text analysis on each of the sentences in the second, third, fifth, and seventh paragraphs, for example, the conversion unit PTN can recognize that the circuit configuration shown in the document file DOC includes a signal line, a scanning line, a first power line, and a second power line that are electrically connected. Here, the conversion unit PTN names the circuit elements CSW (e.g., letters, abbreviations, symbols, words, etc.), for example, the first transistor as Tr[1], the second transistor as Tr[2], the capacitive element as C[1], and the light-emitting element as EL[1], and names the electrical connection points CNP (e.g., letters, abbreviations, symbols, words, etc.), for example, the signal line, scanning line, first power supply line, and second power supply line as N1, N2, N5, and N6, respectively, thereby being able to describe the netlist NTL of stage PH5 in Figure 6 at this stage.

[0124] Note that the netlist NTL shown in step PH5 of Fig. 6 is assumed to follow the same rules as those for the netlist NTL shown in step PH3 of Fig. 5. Therefore, a space SPC is provided between the name CSW of the circuit element and the name CNP of the electrical connection point.

[0125] Continuing from step PH5, by performing text analysis on the document file DOC, the conversion unit PTN can write a more detailed netlist NTL. For example, by performing text analysis on the sentences in the fourth and eighth paragraphs of the document file DOC, the conversion unit PTN can recognize that the source of the first transistor, the gate of the second transistor, and one of the pair of electrodes of the capacitive element are connected to the same electrical connection point. Here, the conversion unit PTN names the electrical connection point N3.

[0126] Furthermore, by performing text analysis on the sentences in the sixth and ninth paragraphs of the document file DOC, the conversion unit PTN can recognize that the source of the second transistor, the input terminal of the light-emitting element, and the other of the pair of electrodes of the capacitive element are connected to the same electrical connection point. Here, the conversion unit PTN names this electrical connection point N4.

[0127] As described above, by performing text analysis on the document file DOC, determining the circuit elements contained in the circuit configuration described in the document file DOC, and extracting their electrical connections, the conversion unit PTN can write the netlist NTL shown in stage PH6 of Figure 6.

[0128] By using the system described in this embodiment, it is possible to convert a circuit diagram or a document file into a netlist, and to search for the converted netlist from a database. In addition, when technical documents such as magazines, science and engineering books, papers, materials from academic societies and lectures, patent publications, and patent gazettes are stored in the database (assuming that there is no infringement of intellectual property rights such as copyrights, and no violation of laws related to intellectual property rights), a user can use the system to determine whether a circuit diagram or document file created by the user is a new matter. In addition, a user can use the system to determine whether a circuit diagram or document file created by the user is a publicly known matter. In other words, by using the system, a user can more efficiently conduct a prior art search for a circuit diagram or document file created by the user.

[0129] Here, a specific example of searching a netlist using the system SIH of FIG. 1 will be described.

[0130] For example, as shown in FIG. 7, in the system SIH of FIG. 1, consider a case in which database DTB1 stores information PKEDD, information PKPD, information HSCD, and information HSPD, and database DTB2 stores netlists PKEDN, PKPN, netlists HSCN, and netlists HSPN (however, the input / output interface INTFC, control unit CTL, conversion unit PTN, and memory unit MP are omitted from FIG. 7).

[0131] Information PKEDD includes, for example, circuit diagrams and specifications of well-known electronic devices, information PKPD includes, for example, technical content (patent specifications, particularly patent drawings, patent claims, papers, magazines, etc.) in which parties other than the user were involved, information HSCD includes, for example, patent claims for circuits in patent specifications in which the user was involved (regardless of whether or not an application has been filed), and information HSPD includes, for example, drawings of circuits in patent specifications in which the user was involved (regardless of whether or not an application has been filed).

[0132] Moreover, the netlist PKEDN has a netlist corresponding to the circuit diagrams etc. included in the information PKEDD, the netlist PKPN has a netlist corresponding to the patent drawings, claims etc. included in the information PKPD, the netlist HSCN has a netlist corresponding to the claims etc. included in the information HSCD, and the netlist HSPN has a netlist corresponding to the patent drawings etc. included in the information HSPD. In addition, in Fig. 7, thick solid lines are drawn between the netlist PKEDN and the information PKEDD, between the netlist PKPN and the information PKPD, between the netlist HSCN and the information HSCD, and between the netlist HSPN and the information HSPD to represent the link between the netlists and the information.

[0133] Here, consider a case where, as the first search SRC1, one netlist PKEDN corresponding to a circuit diagram of a publicly known electronic device is searched for, with multiple netlists HSCN such as the scope of claims of patent applications in which the user has been involved as the search range. In this case, if a netlist corresponding to the netlist PKEDN is found from the multiple netlists HSCN, and the filing date of the patent corresponding to the found netlist is earlier than the date on which the electronic device was made public, it is possible to find an infringement of the user's patent by the electronic device. In other words, by performing the first search SRC1, it is possible to investigate infringement of the user's patent by publicly known electronic devices.

[0134] In addition, as the second search SRC2, a case is considered in which a netlist HSCN of a patent claim before the filing of a patent application involving the user is searched for within the search ranges of a plurality of netlists PKEDN corresponding to circuit diagrams of publicly known electronic devices and a plurality of netlists PKPN corresponding to technical contents involved by parties other than the user. In this case, if a netlist corresponding to the netlist HSCN is found from the plurality of netlists PKEDN and the plurality of netlists PKPN, it can be determined that the netlist HSCN is publicly known. In other words, by performing the second search SRC2, the novelty of an invention involved by the user can be investigated before the patent application is filed. This may improve the patent validity of the patent application involving the user.

[0135] Also, as the third search SRC3, consider a case where one netlist PKEDN corresponding to a circuit diagram of a publicly known electronic device is searched for by using multiple netlists HSPN such as drawings of patent applications in which the user was involved as the search range. In this case, if a netlist corresponding to the netlist PKEDN is found from the multiple netlists HSPN and the filing date of the patent corresponding to the found netlist is earlier than the date on which the electronic device was publicly known, it can be said that the electronic device may utilize the contents of a patent application in which the user was involved. In other words, by performing the third search SRC3, it is possible to check the degree of similarity between the circuit diagram of a publicly known electronic device and the circuit diagram of the contents of a patent application in which the user was involved.

[0136] In the above, the third search SRC3 is described as searching one netlist PKEDN corresponding to a circuit diagram of a known electronic device, and multiple netlists HSPN such as drawings of a patent application in which the user was involved, but the third search SRC3 may search one netlist HSPN such as drawings of a patent application in which the user was involved, and multiple netlists PKEDN corresponding to circuit diagrams of known electronic devices, as the search range. In this search, the degree of similarity between the circuit diagram of the known electronic device and the circuit diagram of the content of the patent application in which the user was involved can be checked.

[0137] Also, the first search SRC1 and the third search SRC3 may be performed simultaneously.

[0138] Also, the first search SRC 1, the second search SRC 2, and the third search SRC 3 can each use AI. For example, by using AI, the descriptions of the netlist to be searched and the netlist included in the search range can be compared, and the similarity indicating the degree of agreement between the types, numbers, and connection configurations of circuit elements in each netlist can be calculated, and the search results can be output in order of the similarity.

[0139] 1, each of the information PKEDD, information PKPD, information HSCD, and information HSPD stored in the database DTB1 is linked to the netlist stored in the database DTB2. Also, each of the first search SRC1, the second search SRC2, and the third search SRC3 can perform a search without converting the netlist into another file (e.g., a circuit diagram, a document file, etc.), which makes the search easier and increases the search speed.

[0140] Although the above describes a search example using the system SIH in FIG. 1, the same search example as above can also be performed using the system SIH in FIG.

[0141] Incidentally, when learning is performed on the neural network of the conversion unit PTN or the conversion unit WPTN, a large amount of data (sometimes called big data) is required. As a method for preparing a large amount of data, for example, a program for automatically randomly generating a netlist is created, and then image data is created from the netlist using circuit design software, a circuit simulator, or the like. At this time, it is preferable to program the netlist to generate and image data in a series. This allows a set of the netlist and image data of the circuit to be prepared as data for learning. As a method for preparing a large amount of data, for example, a program for automatically randomly generating a netlist is created, and further, a program for creating a document file from the automatically generated netlist is created. At this time, it is preferable to program the netlist to generate and document file in a series. This allows a set of the netlist and document file to be prepared as data for learning.

[0142] Note that this embodiment mode can be appropriately combined with other embodiment modes described in this specification.

[0143] (Embodiment 2) In this embodiment, an example of an arithmetic circuit that performs arithmetic on a neural network, which is used in a system of one embodiment of the present invention, will be described.

[0144] <Hierarchical neural network> First, a hierarchical neural network will be described. As an example, a hierarchical neural network has one input layer, one or more intermediate layers (hidden layers), and one output layer, and is composed of a total of three or more layers. FIG. 8A shows an example of a hierarchical neural network 100, which has a first layer to an Rth layer (where R can be an integer of 4 or more). In particular, the first layer corresponds to the input layer, the Rth layer corresponds to the output layer, and the other layers correspond to intermediate layers. FIG. 8A shows the (k-1)th layer and the kth layer (where k is an integer of 3 or more and R-1 or less) as intermediate layers, and does not show the other intermediate layers.

[0145] Each layer of the neural network 100 has one or more neurons. In FIG. 8A, the first layer is made up of neurons N1 (1) Neuron N p (1) (where p is an integer equal to or greater than 1), and the (k-1)th layer has neurons N1 (k-1) Neuron N m (k-1) (where m is an integer equal to or greater than 1), and the kth layer has neurons N1 (k) Neuron N n (k) (where n is an integer equal to or greater than 1), and the Rth layer has neurons N1 (R) Neuron N q (R) (where q is an integer equal to or greater than 1).

[0146] In addition, in Fig. 8A, neuron N1 (1) , neuron N p (1) , neuron N1 (k-1) , neuron N m (k-1) , neuron N1 (k) , neuron N n (k) , neuron N1 (R) , neuron N q (R)In addition, the (k-1)th layer neuron N i (k-1) (where i is an integer between 1 and m), and the kth layer neuron N j (k) (where j is an integer between 1 and n) are also shown, and other neurons are omitted from the illustration.

[0147] Next, we will explain the transmission of signals from neurons in the previous layer to neurons in the next layer, and the signals input and output to and from each neuron. j (k) We are focusing on:

[0148] Figure 8B shows the kth layer of neurons N j (k) and neuron N j (k) The input signal to neuron N j (k) 4 shows a signal output from the

[0149] Specifically, the (k-1)th layer neuron N1 (k-1) Neuron N m (k-1) The output signals of each (k-1) ~z m (k-1) But neuron N j (k) The output is directed to neuron N j (k) is z1 (k-1) ~z m (k-1) Depending on z j (k) Generate z j (k) is output as an output signal to each neuron in the (k+1)th layer (not shown).

[0150] The degree of signal transmission for a signal input from a neuron in the previous layer to a neuron in the next layer is determined by the connection strength (hereinafter referred to as a weighting coefficient) of the synapse that connects those neurons. In the neural network 100, the signal output from a neuron in the previous layer is multiplied by the corresponding weighting coefficient before being input to a neuron in the next layer. If i is an integer between 1 and m, then the (k-1)th layer neuron N i (k-1) and the kth layer neuron N j (k) The weight coefficient of the synapse between i (k-1) j (k) Then, the kth layer neuron N j (k) The signal input to can be expressed by equation (D1).

[0151]

number

[0152] That is, the (k-1)th layer neuron N1 (k-1) Neuron N m (k-1) From each of these, the k-th layer neuron N j (k) When a signal is transmitted to the (k-1) ~z m (k-1) The weighting coefficients (w1 (k-1) j (k) Or even w m (k-1) j (k) ) is multiplied by the k-th layer neuron N j (k) For w1 (k-1) j (k) z1 (k-1) Or even w m (k-1) j (k) z m (k-1)At this time, the kth layer neuron N j (k) The sum of the signals input to j (k) becomes equation (D2).

[0153]

number

[0154] In addition, the weighting factor w1 (k-1) j (k) Or even w m (k-1) j (k) and the neuron's signal z1 (k-1) ~z m (k-1) The result of the multiplication and accumulation of and can be biased. When the bias is b, equation (D2) can be rewritten as the following equation.

[0155]

number

[0156] Neuron N j (k) u j (k) Depending on j (k) where neuron N j (k) Output signal z from j (k) is defined as follows:

[0157]

number

[0158] The function f(u j (k)) is an activation function in a hierarchical neural network, and can be a step function, a linear ramp function, a sigmoid function, etc. The activation function can be the same for all neurons or different. In addition, the activation functions of neurons can be the same or different for each layer.

[0159] Meanwhile, the signal, weighting coefficient w, or bias b output by the neuron of each layer may be an analog value or a digital value. The digital value may be, for example, a binary value or a ternary value. It may also be a value with a larger number of bits. For example, in the case of an analog value, a linear ramp function, a sigmoid function, or the like may be used as an activation function. In the case of a binary digital value, for example, a step function with an output of -1 or 1, or 0 or 1 may be used. Furthermore, the signal output by the neuron of each layer may be ternary or more. In this case, the activation function may be ternary or more, for example, a step function with an output of -1, 0, or 1, or a step function with an output of 0, 1, or 2 may be used. Furthermore, for example, a step function with an output of -2, -1, 0, 1, or 2 may be used as an activation function that outputs five values. By using digital values ​​for at least one of the signals, weighting coefficients w, and biases b output by neurons in each layer, it is possible to reduce the circuit size, reduce power consumption, and increase the calculation speed, etc. Also, by using analog values ​​for at least one of the signals, weighting coefficients w, and biases b output by neurons in each layer, it is possible to improve the accuracy of calculations.

[0160] In neural network 100, an input signal is input to the first layer (input layer), and each layer from the first layer (input layer) to the last layer (output layer) generates an output signal based on the signal input from the previous layer using formula (D1), formula (D2) (or formula (D3)), and formula (D4), and outputs the output signal to the next layer. The signal output from the last layer (output layer) corresponds to the result of calculation by neural network 100.

[0161] <Arithmetic circuit configuration example 1> Next, an example of a circuit for performing product-sum operations and activation function operations in the above-mentioned neural network 100 will be described.

[0162] Fig. 9 shows a configuration example of the arithmetic circuit MAC1. The arithmetic circuit MAC1 shown in Fig. 9 is a circuit that performs a multiplication and accumulation operation on first data stored in a memory cell described later and input second data, and performs an activation function operation using the result of the multiplication and accumulation operation. Note that the first data and the second data can be, for example, analog data or multi-valued data (discrete data).

[0163] The arithmetic circuit MAC1 has a current source circuit CS, a current mirror circuit CM, a circuit WDD, a circuit WLD, a circuit CLD, a circuit OFST, an activation function circuit ACTV, and a memory cell array CA.

[0164] The memory cell array CA has a memory cell AM[1], a memory cell AM[2], a memory cell AMref[1], and a memory cell AMref[2]. The memory cell AM[1] and the memory cell AM[2] have a role of storing first data, and the memory cell AMref[1] and the memory cell AMref[2] have a function of storing reference data required for performing a multiply-and-accumulate operation. Note that the reference data can be analog data or multi-valued data (discrete data) like the first data and the second data.

[0165] 9, the memory cell array CA has two memory cells arranged in a matrix in the row direction and two memory cells arranged in the column direction, but the memory cell array CA may have three or more memory cells arranged in a matrix in the row direction and three or more memory cells arranged in the column direction. Also, when multiplication is performed instead of product-sum operation, the memory cell array CA may have one memory cell arranged in a matrix in the row direction and two or more memory cells arranged in the column direction.

[0166] The memory cell AM[1], the memory cell AM[2], the memory cell AMref[1], and the memory cell AMref[2] each include a transistor Tr11, a transistor Tr12, and a capacitance C1.

[0167] Note that the transistor Tr11 is preferably an OS transistor. In addition, the channel formation region of the transistor Tr11 is more preferably an oxide containing at least one of indium, an element M (for example, the element M may be one or more elements selected from aluminum, gallium, yttrium, copper, vanadium, beryllium, boron, titanium, iron, nickel, germanium, zirconium, molybdenum, lanthanum, cerium, neodymium, hafnium, tantalum, tungsten, magnesium, etc.), and zinc. It is further preferable that the transistor Tr11 has the structure of the transistor described in the third embodiment.

[0168] By using an OS transistor as the transistor Tr11, the leakage current of the transistor Tr11 can be suppressed, which may result in a product-sum operation circuit with high calculation accuracy. In addition, by using an OS transistor as the transistor Tr11, the leakage current from the retention node to the write word line when the transistor Tr11 is in a non-conducting state can be made very small. In other words, the number of refresh operations of the potential of the retention node can be reduced, which reduces the power consumption of the product-sum operation circuit.

[0169] In addition, by using an OS transistor for the transistor Tr12, the transistor Tr12 can be manufactured simultaneously with the transistor Tr11, which may shorten the manufacturing process of the product-sum operation circuit. The channel formation region of the transistor Tr12 may contain silicon instead of oxide. The silicon may be, for example, amorphous silicon (sometimes called hydrogenated amorphous silicon), microcrystalline silicon, polycrystalline silicon, single crystal silicon, or the like.

[0170] In each of the memory cells AM[1], AM[2], AMref[1], and AMref[2], a first terminal of the transistor Tr11 is electrically connected to a gate of the transistor Tr12. A first terminal of the transistor Tr12 is electrically connected to the wiring VR. A first terminal of the capacitor C1 is electrically connected to the gate of the transistor Tr12.

[0171] In the memory cell AM[1], the second terminal of the transistor Tr11 is electrically connected to the wiring WD, and the gate of the transistor Tr11 is electrically connected to the wiring WL[1]. The second terminal of the transistor Tr12 is electrically connected to the wiring BL, and the second terminal of the capacitor C1 is electrically connected to the wiring CL[1]. In FIG. 9, the connection point between the first terminal of the transistor Tr11, the gate of the transistor Tr12, and the first terminal of the capacitor C1 in the memory cell AM[1] is referred to as a node NM[1]. In addition, the current flowing from the wiring BL to the second terminal of the transistor Tr12 is referred to as I AM[1] Let us assume that.

[0172] In the memory cell AM[2], the second terminal of the transistor Tr11 is electrically connected to the wiring WD, and the gate of the transistor Tr11 is electrically connected to the wiring WL[2]. The second terminal of the transistor Tr12 is electrically connected to the wiring BL, and the second terminal of the capacitor C1 is electrically connected to the wiring CL[2]. In FIG. 9, the connection point between the first terminal of the transistor Tr11, the gate of the transistor Tr12, and the first terminal of the capacitor C1 in the memory cell AM[2] is referred to as a node NM[2]. In addition, the current flowing from the wiring BL to the second terminal of the transistor Tr12 is referred to as I AM[2] Let us assume that.

[0173] In the memory cell AMref[1], the second terminal of the transistor Tr11 is electrically connected to the wiring WDref, and the gate of the transistor Tr11 is electrically connected to the wiring WL[1]. The second terminal of the transistor Tr12 is electrically connected to the wiring BLref, and the second terminal of the capacitor C1 is electrically connected to the wiring CL[1]. In FIG. 9, in the memory cell AMref[1], the connection point between the first terminal of the transistor Tr11, the gate of the transistor Tr12, and the first terminal of the capacitor C1 is defined as a node NMref[1]. In addition, the current flowing from the wiring BLref to the second terminal of the transistor Tr12 is defined as I AMref[1] Let us assume that.

[0174] In the memory cell AMref[2], the second terminal of the transistor Tr11 is electrically connected to the wiring WDref, and the gate of the transistor Tr11 is electrically connected to the wiring WL[2]. The second terminal of the transistor Tr12 is electrically connected to the wiring BLref, and the second terminal of the capacitor C1 is electrically connected to the wiring CL[2]. In FIG. 9, the connection point between the first terminal of the transistor Tr11, the gate of the transistor Tr12, and the first terminal of the capacitor C1 in the memory cell AMref[2] is referred to as a node NMref[2]. In addition, the current flowing from the wiring BLref to the second terminal of the transistor Tr12 is referred to as I AMref[2] Let us assume that.

[0175] The above-mentioned nodes NM[1], NM[2], NMref[1], and NMref[2] function as retention nodes for the respective memory cells.

[0176] The wiring VR is a wiring for passing a current between the first terminal and the second terminal of the transistor Tr12 of each of the memory cells AM[1], AM[2], AMref[1], and AMref[2]. Therefore, the wiring VR functions as a wiring for applying a predetermined potential. Note that in this embodiment, the potential applied by the wiring VR can be a reference potential or a potential lower than the reference potential.

[0177] The current source circuit CS is electrically connected to the wiring BL and the wiring BLref. The current source circuit CS has a function of supplying a current to the wiring BL and the wiring BLref. The current amounts supplied to the wiring BL and the wiring BLref may be different from each other. In this configuration example, the current flowing from the current source circuit CS to the wiring BL is referred to as I C Let I be the current flowing from the current source circuit CS to the wiring BLref. Cref Let us assume that.

[0178] The current mirror circuit CM has a wiring IL and a wiring ILref. The wiring IL is electrically connected to the wiring BL, and in FIG. 9, the connection point of the wiring IL and the wiring BL is illustrated as a node NP. The wiring ILref is electrically connected to the wiring BLref, and in FIG. 9, the connection point of the wiring ILref and the wiring BLref is illustrated as a node NPref. The current mirror circuit CM has a function of discharging a current according to the potential of the node NPref from the node NPref of the wiring BLref to the wiring ILref, and discharging a current of the same amount from the node NP of the wiring BL to the wiring IL. Note that in FIG. 9, the current discharged from the node NP to the wiring IL and the current discharged from the node NPref to the wiring ILref are referred to as I CM In addition, the current flowing from the current mirror circuit CM to the memory cell array CA in the wiring BL is expressed as I B In the wiring BLref, the current flowing from the current mirror circuit CM to the memory cell array CA is I Bref It is written as follows.

[0179] The circuit WDD is electrically connected to the wiring WD and the wiring WDref The circuit WDD has a function of transmitting data to be stored in each memory cell of the memory cell array CA.

[0180] The circuit WLD is electrically connected to the wiring WL[1] and the wiring WL[2]. The circuit WLD has a function of selecting a memory cell to which data is to be written when data is written to a memory cell included in the memory cell array CA.

[0181] The circuit CLD is electrically connected to the wirings CL[1] and CL[2]. The circuit CLD has a function of applying a potential to the second terminal of the capacitor C1 of each memory cell included in the memory cell array CA.

[0182] The circuit OFST is electrically connected to the wiring BL and the wiring OL. The circuit OFST has a function of measuring the amount of current flowing from the wiring BL to the circuit OFST and / or the amount of change in the current flowing from the wiring BL to the circuit OFST. In addition, the circuit OFST has a function of outputting the result of the measurement to the wiring OL. Note that the circuit OFST may be configured to output the result of the measurement directly as a current to the wiring OL, or may be configured to convert the result of the measurement into a voltage and output it to the wiring OL. Note that in FIG. 9, the current flowing from the wiring BL to the circuit OFST is represented as I α He wrote:

[0183] For example, the circuit OFST may have a configuration shown in Fig. 10. In Fig. 10, the circuit OFST has a transistor Tr21, a transistor Tr22, a transistor Tr23, a capacitance C2, and a resistor R1.

[0184] A first terminal of the capacitance C2 is electrically connected to the wiring BL, and a first terminal of the resistor R1 is electrically connected to the wiring BL. A second terminal of the capacitance C2 is electrically connected to a first terminal of the transistor Tr21, and the first terminal of the transistor Tr21 is electrically connected to the gate of the transistor Tr22. A first terminal of the transistor Tr22 is electrically connected to a first terminal of the transistor Tr23, and the first terminal of the transistor Tr23 is electrically connected to the wiring OL. The electrical connection point between the first terminal of the capacitance C2 and the first terminal of the resistor R1 is a node Na, and the electrical connection point between the second terminal of the capacitance C2, the first terminal of the transistor Tr21, and the gate of the transistor Tr22 is a node Nb.

[0185] A second terminal of the resistor R1 is electrically connected to the wiring VrefL. A second terminal of the transistor Tr21 is electrically connected to the wiring VaL, and a gate of the transistor Tr21 is electrically connected to the wiring RST. A second terminal of the transistor Tr22 is electrically connected to the wiring VDDL. A second terminal of the transistor Tr23 is electrically connected to the wiring VSSL, and a gate of the transistor Tr23 is electrically connected to the wiring VbL.

[0186] The wiring VrefL is a wiring that provides a potential Vref, the wiring VaL is a wiring that provides a potential Va, and the wiring VbL is a wiring that provides a potential Vb. The wiring VDDL is a wiring that provides a potential VDD, and the wiring VSSL is a wiring that provides a potential VSS. In particular, in the configuration example of the circuit OFST here, the potential VDD is a high-level potential, and the potential VSS is a low-level potential. The wiring RST is a wiring that provides a potential for switching the transistor Tr21 between a conductive state and a non-conductive state.

[0187] In the circuit OFST shown in FIG. 10, a source follower circuit is configured by the transistor Tr22, the transistor Tr23, the wiring VDDL, the wiring VSSL, and the wiring VbL.

[0188] In the circuit OFST shown in FIG. 10, a current flowing from the wiring BL and a potential according to the resistance of the resistor R1 are applied to the node Na by the resistor R1 and the wiring VrefL.

[0189] An operation example of the circuit OFST shown in Fig. 10 will be described. When a first current (hereinafter referred to as a first current) flows from the wiring BL, a potential according to the first current and the resistance of the resistor R1 is applied to the node Na by the resistor R1 and the wiring VrefL. At this time, the transistor Tr21 is made conductive to apply a potential Va to the node Nb. After that, the transistor Tr21 is made non-conductive.

[0190] Next, when a second current (hereinafter referred to as the second current) flows from the wiring BL, a potential according to the second current and the resistance of the resistor R1 is applied to the node Na by the resistor R1 and the wiring VrefL, in the same manner as when the first current flows. At this time, since the node Nb is in a floating state, the change in the potential of the node Na also changes the potential of the node Nb due to capacitive coupling. The change in the potential of the node Na is represented as ΔV Na When the capacitance coupling coefficient is set to 1, the potential of node Nb is Va+ΔV Na The threshold voltage of the transistor Tr22 is V th When this is done, the potential from the wiring OL is Va+ΔV Na -V th Here, the potential Va is the threshold voltage V th By doing so, the potential ΔV Na can be output.

[0191] Potential ΔV Na is determined according to the change amount from the first current to the second current, the resistance value of the resistor R1, and the potential Vref. Since the resistance value of the resistor R1 and the potential Vref can be known, the potential ΔV Na From this, the amount of change in the current flowing through the wiring BL can be obtained.

[0192] The activation function circuit ACTV is electrically connected to the wiring OL and the wiring NIL. The result of the change in the current measured by the circuit OFST is input to the activation function circuit ACTV via the wiring OL. The activation function circuit ACTV is a circuit that performs a calculation on the result according to a predefined function system. As the function system, for example, a sigmoid function, a tanh function, a softmax function, a ReLU function, a threshold function, etc. can be used, and these functions are applied as activation functions in the neural network.

[0193] <Operation example 1 of the calculation circuit> Next, an example of the operation of the arithmetic circuit MAC1 will be described.

[0194] 11 is a timing chart showing an operation example of the arithmetic circuit MAC1. The timing chart in FIG. 11 shows changes in the potentials of the wiring WL[1], the wiring WL[2], the wiring WD, the wiring WDref, the node NM[1], the node NM[2], the node NMref[1], the node NMref[2], the wiring CL[1], and the wiring CL[2] from time T01 to time T09. B -I α , and current I Bref In particular, the current I B -I α indicates the sum of currents flowing from the wiring BL to the memory cells AM[1] and AM[2] of the memory cell array CA.

[0195] <<From time T01 to time T02>> Between time T01 and time T02, a high-level potential (denoted as High in FIG. 11) is applied to the wiring WL[1], and a low-level potential (denoted as Low in FIG. 11) is applied to the wiring WL[2]. In addition, the wiring WD is V PR -V W[1] A large potential is applied, and the wiring WDref is V higher than the ground potential. PR A large potential is applied to the wiring CL[1] and the wiring CL[2]. Furthermore, a reference potential (denoted as REFP in FIG. 11) is applied to each of the wirings CL[1] and CL[2].

[0196] In addition, the potential V W[1] is a potential corresponding to one of the first data. PR is the potential corresponding to the reference data.

[0197] At this time, a high-level potential is applied to the gates of the transistors Tr11 of the memory cells AM[1] and AMref[1], so that the transistors Tr11 of the memory cells AM[1] and AMref[1] are turned on. Therefore, in the memory cell AM[1], the line WD and the node NM[1] are in a conductive state, so that the potential of the node NM[1] is V PR -V W[1] Similarly, in the memory cell AMref[1], the line WDref and the node NMref[1] are in a conductive state, so that the potential of the node NMref[1] becomes V PR It becomes.

[0198] Here, consider the current flowing from the second terminal to the first terminal of the transistor Tr12 of each of the memory cells AM[1] and AMref[1]. Let I AM[1],0 Then, I AM[1],0 can be expressed by the following formula:

[0199]

number

[0200] k is a constant determined by the channel length, channel width, mobility, and capacitance of the gate insulating film of the transistor Tr12. th is the threshold voltage of transistor Tr12.

[0201] The current flowing from the wiring BLref to the first terminal via the second terminal of the transistor Tr12 of the memory cell AMref[1] is I AMref[1],0 Similarly, when AMref[1],0 can be expressed by the following formula:

[0202]

number

[0203] Since a low-level potential is applied to the gates of the transistors Tr11 in the memory cells AM[2] and AMref[2], the transistors Tr11 in the memory cells AM[2] and AMref[2] are turned off, and no potential is written to the nodes NM[2] and NMref[2].

[0204] <<From time T02 to time T03>> Between time T02 and time T03, a low-level potential is applied to the line WL[1]. At this time, a low-level potential is applied to the gates of the transistors Tr11 of the memory cells AM[1] and AMref[1], so that the transistors Tr11 of the memory cells AM[1] and AMref[1] are turned off.

[0205] In addition, the low-level potential continues to be applied to the line WL[2] from before time T02, so that the transistors Tr11 of the memory cells AM[2] and AMref[2] are kept in the off state from before time T02.

[0206] As described above, transistors Tr11 of memory cell AM[1], memory cell AM[2], memory cell AMref[1], and memory cell AMref[2] are in the off state, so that the potentials of nodes NM[1], NM[2], NMref[1], and NMref[2] are maintained between time T02 and time T03.

[0207] In particular, as described in the explanation of the circuit configuration of the arithmetic circuit MAC1, by applying an OS transistor to each of the transistors Tr11 in the memory cell AM[1], the memory cell AM[2], the memory cell AMref[1], and the memory cell AMref[2], the leakage current flowing between the first terminal and the second terminal of the transistor Tr11 can be reduced, so that the potentials of the nodes NM[1], NM[2], NMref[1], and NMref[2] can be maintained for a long period of time.

[0208] Between time T02 and time T03, the ground potential is applied to the wiring WD and the wiring WDref. Since the transistors Tr11 of the memory cells AM[1], AM[2], AMref[1], and AMref[2] are in the off state, the application of potentials from the wiring WD and the wiring WDref does not rewrite the potentials held at the nodes NM[1], NM[2], NMref[1], and NMref[2].

[0209] <<From time T03 to time T04>> Between time T03 and time T04, a low-level potential is applied to the wiring WL[1], and a high-level potential is applied to the wiring WL[2]. In addition, the wiring WD is V PR -V W[2] A large potential is applied, and the wiring WDref is V higher than the ground potential. PR Furthermore, the reference potential is being applied to the wiring CL[1] and the wiring CL[2] continuously from before time T02.

[0210] In addition, the potential V W[2] is a potential corresponding to one of the first data.

[0211] At this time, a high-level potential is applied to the gates of the transistors Tr11 of the memory cells AM[2] and AMref[2], so that the transistors Tr11 of the memory cells AM[2] and AMref[2] are turned on. Therefore, in the memory cell AM[2], the line WD and the node NM[2] are in a conductive state, so that the potential of the node NM[2] is V PR -V W[2] Similarly, in the memory cell AMref[2], the line WDref and the node NMref[2] are in a conductive state, so that the potential of the node NMref[2] becomes V PR It becomes.

[0212] Here, consider the current flowing from the second terminal to the first terminal of the transistor Tr12 of each of the memory cells AM[2] and AMref[2]. Let I AM[2],0 Then, I AM[2],0 can be expressed by the following formula:

[0213]

number

[0214] The current flowing from the wiring BLref to the first terminal via the second terminal of the transistor Tr12 of the memory cell AMref[2] is I AMref[2],0 Similarly, when I AMref[2],0 can be expressed by the following formula:

[0215]

number

[0216] <<From time T04 to time T05>> Here, the current flowing through the wiring BL and the wiring BLref from time T04 to time T05 will be described.

[0217] A current is supplied to the wiring BLref from the current source circuit CS. In addition, a current is discharged from the wiring BLref by the current mirror circuit CM, the memory cell AMref[1], and the memory cell AMref[2]. The current supplied from the current source circuit CS to the wiring BLref is represented as I Cref The current drawn by the current mirror circuit CM is I CM,0 Then, the following equation holds according to Kirchhoff's law:

[0218]

number

[0219] A current is supplied to the wiring BL from the current source circuit CS. In addition, a current is discharged from the wiring BL by the current mirror circuit CM, the memory cell AM[1], and the memory cell AM[2]. Furthermore, a current also flows from the wiring BL to the circuit OFST. The current supplied from the current source circuit CS to the wiring BL is defined as I C Let I be the current flowing from the wiring BL to the circuit OFST. α,0 Then, the following equation holds according to Kirchhoff's law:

[0220]

number

[0221] <<From time T05 to time T06>> Between time T05 and time T06, the line CL[1] is charged to a potential higher than the reference potential V X[1] At this time, a high potential is applied to the second terminal of the capacitor C1 of each of the memory cell AM[1] and the memory cell AMref[1]. X[1] is applied, the potential of the gate of the transistor Tr12 increases.

[0222] In addition, the potential V x[1] is a potential corresponding to one of the second data.

[0223] The increase in the potential of the gate of the transistor Tr12 is the potential obtained by multiplying the change in the potential of the wiring CL[1] by a capacitance coupling coefficient determined by the configuration of the memory cell. The capacitance coupling coefficient is calculated based on the capacitance of the capacitor C1, the gate capacitance of the transistor Tr12, the parasitic capacitance, etc. In this operation example, to avoid complexity of explanation, the increase in the potential of the wiring CL[1] and the increase in the potential of the gate of the transistor Tr12 are described as having the same value. This is equivalent to the capacitance coupling coefficient of each of the memory cell AM[1] and the memory cell AMref[1] being set to 1.

[0224] Since the capacitive coupling coefficient is set to 1, a potential V X[1] When this is applied, the potentials of the nodes NM[1] and NMref[1] become V X[1] to rise.

[0225] Here, consider the current flowing from the second terminal to the first terminal of the transistor Tr12 of each of the memory cells AM[1] and AMref[1]. Let I AM[1],1 Then, I AM[1],1 can be expressed by the following formula:

[0226]

number

[0227] In other words, the potential V X[1] By applying the voltage Vcc, the current flowing from the line BL to the first terminal of the transistor Tr12 in the memory cell AM[1] via the second terminal is I AM[1],1 -I AM[1],0 (In Figure 11, ΔI AM[1] ) increases.

[0228] Similarly, the current flowing from the wiring BLref to the first terminal via the second terminal of the transistor Tr12 of the memory cell AMref[1] is I AMref[1],1 Then, I AMref[1],1 can be expressed by the following formula:

[0229]

number

[0230] In other words, the potential V X[1] By applying the voltage Vcc, the current flowing from the wiring BLref to the first terminal of the transistor Tr12 of the memory cell AMref[1] via the second terminal is I AMref[1],1 -I AMref[1],0 (In Figure 11, ΔI AMref[1] ) increases.

[0231] Here, the current flowing through the wiring BL and the wiring BLref will be described.

[0232] The current I from the current source circuit CS flows through the wiring BLref in the same manner as between time T04 and time T05. Cref At the same time, a current is discharged to the wiring BLref by the current mirror circuit CM, the memory cell AMref[1], and the memory cell AMref[2]. The current discharged by the current mirror circuit CM in the wiring BLref is defined as I CM,1 Then, the following equation holds according to Kirchhoff's law:

[0233]

number

[0234] The current I from the current source circuit CS flows through the wiring BL in the same manner as between time T04 and time T05. CAt the same time, a current is discharged to the wiring BL by the current mirror circuit CM, the memory cell AM[1], and the memory cell AM[2]. Furthermore, a current also flows from the wiring BL to the circuit OFST. In the wiring BL, the current flowing from the wiring BL to the circuit OFST is defined as I α,1 Then, the following equation holds according to Kirchhoff's law:

[0235]

number

[0236] Between time T04 and time T05, the current I flows from the wiring BL to the circuit OFST. α,0 and the current I flowing from the wiring BL to the circuit OFST between time T05 and time T06. α,1 The difference between and is ΔI α Hereafter, ΔI α is referred to as the differential current in the arithmetic circuit MAC1. α can be expressed as follows using equations (E1) to (E10):

[0237]

number

[0238] <<From time T06 to time T07>> Between time T06 and time T07, the reference potential is applied to the line CL[1]. At this time, the reference potential is applied to the second terminals of the capacitors C1 of the memory cells AM[1] and AMref[1], so that the potentials of the nodes NM[1] and NMref[1] return to the potentials between time T04 and time T05.

[0239] <<From time T07 to time T08>> Between time T07 and time T08, the line CL[1] is charged to a potential higher than the reference potential V X[1] A higher potential is applied, and the line CL[2] is V higher than the reference potential. X[2]At this time, a high potential is applied to the second terminal of the capacitor C1 of each of the memory cell AM[1] and the memory cell AMref[1]. X[1] is applied to the second terminal of the capacitor C1 of each of the memory cells AM[2] and AMref[2], and a potential V X[2] As a result, the potential of the gate of the transistor Tr12 in each of the memory cell AM[1], the memory cell AM[2], the memory cell AMref[1], and the memory cell AMref[2] rises.

[0240] The changes in the potentials of the nodes of the memory cell AM[1] and the memory cell AMref[1] refer to the operations from time T05 to time T06. Similarly, the capacitive coupling coefficients of the memory cells AM[2] and AMref[2] are also explained assuming that each memory cell has a capacitance of 1.

[0241] Since the capacitive coupling coefficient is set to 1, a potential V X[2] When this voltage is applied, the potentials of the nodes NM[2] and NMref[2] become V X[2] to rise.

[0242] Here, consider the current flowing from the second terminal to the first terminal of the transistor Tr12 of each of the memory cells AM[2] and AMref[2]. Let I be the current flowing from the wiring BL to the first terminal via the second terminal of the transistor Tr12 of the memory cell AM[1]. AM[2],1 Then, I AM[2],1 can be expressed by the following formula:

[0243]

number

[0244] In other words, the potential V X[2] By applying the voltage Vcc, the current flowing from the line BL to the first terminal of the transistor Tr12 in the memory cell AM[2] via the second terminal is IAM[2],1 -I AM[2],0 (In Figure 11, ΔI AM[2] ) increases.

[0245] Similarly, the current flowing from the wiring BLref to the first terminal through the second terminal of the transistor Tr12 of the memory cell AMref[2] is I AMref[2],1 Then, I AMref[2],1 can be expressed by the following formula:

[0246]

number

[0247] In other words, the potential V X[2] By applying the voltage Vcc, the current flowing from the wiring BLref to the first terminal of the transistor Tr12 of the memory cell AMref[2] via the second terminal is I AMref[2],1 -I AMref[2],0 (In Figure 11, ΔI AMref[2] ) increases.

[0248] Here, the current flowing through the wiring BL and the wiring BLref will be described.

[0249] The current I from the current source circuit CS flows through the wiring BLref in the same manner as between time T04 and time T05. Cref At the same time, a current is discharged to the wiring BLref by the current mirror circuit CM, the memory cell AMref[1], and the memory cell AMref[2]. The current discharged by the current mirror circuit CM in the wiring BLref is defined as I CM,2 Then, the following equation holds according to Kirchhoff's law:

[0250]

number

[0251] The current I from the current source circuit CS flows through the wiring BL in the same manner as between time T04 and time T05. CAt the same time, a current is discharged to the wiring BL by the current mirror circuit CM, the memory cell AM[1], and the memory cell AM[2]. Furthermore, a current also flows from the wiring BL to the circuit OFST. In the wiring BL, the current flowing from the wiring BL to the circuit OFST is defined as I α,3 Then, the following equation holds according to Kirchhoff's law:

[0252]

number

[0253] Between time T04 and time T05, the current I flows from the wiring BL to the circuit OFST. α,0 and the current I flowing from the wiring BL to the circuit OFST between time T07 and time T08. α,3 The difference between the current and the current difference ΔI α can be expressed as follows using formulas (E1) to (E8) and (E12) to (E15).

[0254]

number

[0255] As shown in equations (E11) and (E16), the differential current ΔI α is a plurality of first data, i.e., potentials V W and a plurality of second data potentials V X In other words, the differential current ΔI α By measuring this with the circuit OFST, the value of the sum of products of the first data and the second data can be obtained.

[0256] <<From time T08 to time T09>> Between time T08 and time T09, the reference potential is applied to the wiring CL[1] and the wiring CL[2]. At this time, the reference potential is applied to the second terminal of the capacitance C1 of each of the memory cell AM[1], the memory cell AM[2], the memory cell AMref[1], and the memory cell AMref[2], so that the potentials of the nodes NM[1], NM[2], NMref[1], and NMref[2] return to the potentials between time T06 and time T07.

[0257] Between time T05 and time T06, V X[1] Between time T07 and time T08, V is applied to the wiring CL[1] and the wiring CL[2]. X[1] , V X[2] In the above example, the potential applied to the wiring CL[1] and the wiring CL[2] may be lower than the reference potential REFP. When a potential lower than the reference potential REFP is applied to the wiring CL[1] and / or the wiring CL[2], the potential of the retention node of the memory cell connected to the wiring CL[1] and / or the wiring CL[2] can be lowered by capacitive coupling. This makes it possible to multiply the first data by the negative value of the second data in a product-sum operation. For example, between time T07 and time T08, when V X[2] Not -V X[2] When the differential current ΔI α can be expressed as the following formula:

[0258]

number

[0259] In this operation example, the memory cell array CA having memory cells arranged in a matrix of 2 rows and 2 columns is treated, but the product-sum operation can be performed in the same manner for a memory cell array with 1 row and 2 or more columns, or a memory cell array with 3 or more rows and 3 or more columns. In this case, the product-sum operation circuit supplies one of the columns with reference data (potential V PR), it is possible to simultaneously perform the sum-of-products operation for the remaining number of columns among the multiple columns. In other words, by increasing the number of columns in the memory cell array, it is possible to provide a semiconductor device that realizes high-speed sum-of-products operation. Also, by increasing the number of rows, it is possible to increase the number of terms to be added in the sum-of-products operation. When the number of rows is increased, the difference current ΔI α can be expressed by the following formula:

[0260]

number

[0261] When the multiplication and accumulation circuit described in this embodiment is applied to the hidden layer, the weight coefficient w s[k]s[k-1] (k) is stored in each memory cell AM of the same column as the first data, and the output signal z s[k-1] (k-1) is applied from the wiring CL of each row (second data), the differential current ΔI α In addition, the sum of products of the first data and the second data can be calculated by using the sum of products to calculate the value of the activation function, and the output signal z s[k] (k) It can be said that:

[0262] In addition, when the multiply-and-accumulate circuit described in this embodiment is applied to the output layer, the weight coefficient w s[L]s[L-1] (L) is stored in each memory cell AM of the same column as the first data, and the output signal z s[L-1] (L-1) is applied from the wiring CL of each row (second data), the differential current ΔI α In addition, the sum of products of the first data and the second data can be calculated by using the sum of products to calculate the value of the activation function, and the value of the activation function is used as a signal to generate the output signal z s[L](L) It can be said that:

[0263] The input layer described in this embodiment may function as a buffer circuit that outputs an input signal to the second layer.

[0264] Incidentally, in the arithmetic circuit described in this embodiment, the number of rows of memory cells AM is the number of neurons in the previous layer. In other words, the number of rows of memory cells AM corresponds to the number of output signals of neurons in the previous layer that are input to one neuron in the next layer. And the number of columns of memory cells AM is the number of neurons in the next layer. In other words, the number of columns of memory cells AM corresponds to the number of output signals output from neurons in the next layer. In other words, the number of rows and columns of the memory cell array of the arithmetic circuit are determined by the number of neurons in the previous layer and the next layer, so it is sufficient to determine and design the number of rows and columns of the memory cell array according to the neural network to be configured.

[0265] The arithmetic circuit described in this embodiment may be changed in configuration depending on the situation. For example, the arithmetic circuit MAC1 shown in Fig. 9 may be changed to the arithmetic circuit MAC1 shown in Fig. 12. The arithmetic circuit MAC1 in Fig. 12 is configured by adding a memory cell AM[1] of the memory cell array CA and a memory cell AMB to the column including the memory cell AM[1], in comparison with the arithmetic circuit MAC1 in Fig. 9.

[0266] The memory cell AMB is electrically connected to a wiring WD, a wiring BL, a wiring WLB, and a wiring CLB. The wiring WLB is electrically connected to a circuit WLD, and the wiring CLB is electrically connected to a circuit CLD.

[0267] In the memory cell AMB, a connection point between the first terminal of the transistor Tr11, the gate of the transistor Tr12, and the first terminal of the capacitor C1 is defined as a node NMB.

[0268] The wiring WLB functions as a wiring that supplies a selection signal from the circuit WLD to the memory cell AMB when writing data to the memory cell AMB. The wiring CLB functions as a wiring that applies a constant potential to the second terminal of the capacitor C1 of the memory cell AMB. The constant potential is preferably a ground potential or a low-level potential.

[0269] As an example of the operation of the arithmetic circuit MAC1 in Fig. 12, for example, in the timing chart of Fig. 11, between time T01 and time T05, the node NMB is held at the ground potential, low level potential, or potential applied by the wiring VR so that the transistor Tr12 of the memory cell AMB is in the off state. Then, in the timing chart of Fig. 11, between time T05 and time T09, an arbitrary current I BIAS A potential V BIAS At this time, I BIAS is expressed by the following formula:

[0270]

number

[0271] In this case, equations (E16) and (E18) can be rewritten as the following equations.

[0272]

number

[0273]

number

[0274] The formulas (E20) and (E21) correspond to an operation that further gives an arbitrary bias to the result of the multiplication and addition operation. In other words, the operation of the formula (D3) can be performed by using the arithmetic circuit MAC1 in FIG. 12. BIASis determined not only by the potential of the node NMB but also by the potential applied by the line CLB. For example, in the timing chart of FIG. 11, between time T01 and time T05, the ground potential is applied to the line CLB so that the transistor Tr12 of the memory cell AMB is in an off state, and between time T05 and time T09, the potential of the line CLB is changed from the ground potential to an arbitrary potential, so that an arbitrary current I BIAS may be made to flow.

[0275] Also, for example, the arithmetic circuit MAC1 shown in Fig. 9 may be changed to an arithmetic circuit MAC1A shown in Fig. 13. The arithmetic circuit MAC1A in Fig. 13 has a circuit CMS combining the current source circuit CS and the current mirror circuit CM in the arithmetic circuit MAC1 in Fig. 9, a circuit OFAC combining the circuit OFST and the activation function circuit ACTV, and a memory cell array CA.

[0276] The circuit CMS includes, for example, a current mirror circuit CM, a current source circuit CS1, a current source circuit CS2, and a switch SW3.

[0277] The current mirror circuit CM includes, for example, a transistor Tr31 and a transistor Tr32. The current source circuit CS1 includes, for example, a transistor Tr33, a capacitor C6, and a switch SW1. The current source circuit CS2 includes, for example, a transistor Tr34, a capacitor C7, and a switch SW2.

[0278] The circuit OFAC includes, for example, a switch SW4 and a resistor RE.

[0279] Note that each of the transistors Tr31 to Tr33 is preferably a p-channel transistor as shown in Fig. 13. Also, the transistor Tr34 is preferably an n-channel transistor as shown in Fig. 13. Each of the transistors Tr31 to Tr34 can be, for example, a Si transistor.

[0280] Unless otherwise specified, the transistors Tr31 to Tr34 are assumed to operate in the saturation region when on, i.e., the gate voltage, source voltage, and drain voltage of each of the above-mentioned transistors are appropriately biased to a voltage range in which the transistors operate in the saturation region.

[0281] Each of the switches SW1 to SW4 can be, for example, an electrical switch, a mechanical switch, etc. In particular, when an electrical switch is used for each of the switches SW1 to SW4, an OS transistor, a Si transistor, or the like can be used as the electrical switch.

[0282] For the memory cell array CA, please refer to the description of the memory cell array CA of the arithmetic circuit MAC1 in Fig. 9. Note that in Fig. 13, the circuits CLD, WDD, and WLD are omitted.

[0283] In the current mirror circuit CM, a first terminal of the transistor Tr31 is electrically connected to the wiring VHE, a second terminal of the transistor Tr31 is electrically connected to the gate of the transistor Tr31 and the wiring BLref, a first terminal of the transistor Tr32 is electrically connected to the wiring VHE, and a second terminal of the transistor Tr32 is electrically connected to a first terminal of the switch SW3 and a first terminal of the switch SW4 of the circuit OFAC.

[0284] In the current source circuit CS1, a first terminal of the transistor Tr33 is electrically connected to the wiring VHE, a second terminal of the transistor Tr33 is electrically connected to a first terminal of the switch SW1, a second terminal of the switch SW3, and the wiring BL, and a gate of the transistor Tr33 is electrically connected to a second terminal of the switch SW1 and a first terminal of the capacitor C6. A second terminal of the capacitor C6 is electrically connected to the wiring VHE.

[0285] In the current source circuit CS2, a first terminal of the transistor Tr34 is electrically connected to the line VLE, ​​a second terminal of the transistor Tr34 is electrically connected to a first terminal of the switch SW2, a first terminal of the switch SW3, and a first terminal of the switch SW4 of the circuit OFAC, and a gate of the transistor Tr34 is electrically connected to a second terminal of the switch SW2 and a first terminal of the capacitor C7. A second terminal of the capacitor C7 is electrically connected to the line VLE.

[0286] In the circuit OFAC, a second terminal of the switch SW4 is electrically connected to a first terminal of the resistor RE, and a second terminal of the resistor RE is electrically connected to the wiring VcL.

[0287] The wiring VHE functions as a wiring that supplies a constant voltage. The constant voltage can be, for example, a high-level potential. The wiring VLE functions as a wiring that supplies a constant voltage. The constant voltage can be, for example, a low-level potential, a ground potential, etc. The wiring VcL functions as a wiring that supplies a constant voltage. The constant voltage can be, for example, a high-level potential, a low-level potential, a ground potential, etc.

[0288] The current mirror circuit CM has a function of supplying a current corresponding to the potential of the second terminal of the transistor Tr31 from the wiring VHE to the second terminal of the transistor Tr31, and also supplying a current from the wiring VHE to the second terminal of the transistor Tr32. Note that, at this time, it is preferable that the amounts of current flowing between the source and drain of the transistor Tr31 and between the source and drain of the transistor Tr32 are equal to each other.

[0289] The resistor RE included in the circuit OFAC has a function of converting a current input to a first terminal of the resistor RE via the switch SW4 into a voltage. That is, the circuit OFAC functions as, for example, a current-voltage conversion circuit.

[0290] Next, a specific example of the operation of the arithmetic circuit MAC1A will be described.

[0291] First, the retention nodes of the memory cells AM[1] and AM[2] included in the memory cell array CA are V PR -V W[1] , V PR -V W[2] In addition, the memory cells AMref[1] and AMref[2] included in the memory cell array CA are both stored at their respective storage nodes. PR It is also assumed that a potential REFP is input to each of the wiring CL[1] and the wiring CL[2].

[0292] At this time, the currents flowing through the transistors Tr12 of the memory cells AM[1], AM[2], AMref[1], and AMref[2] are expressed by equations (E1), (E3), (E2), and (E4).

[0293] Next, as shown in FIG. 14, the switches SW1 and SW2 are turned on, and the switches SW3 and SW4 are turned off.

[0294] Here, when the current flowing through the wiring BL is I3, the current I3 is the sum of the currents flowing between the source and drain of the transistors Tr12 in the memory cells AM[1] and AM[2]. Therefore, from the formulas (E1) and (E3), I3=I AM[1],0 +I AM[2],0 It can be said that:

[0295] In the current source circuit CS1, the switch SW1 is on, so that the transistor Tr33 is diode-connected, and the gate of the transistor Tr33 has a potential according to the current I3, so that the current I3 flows between the source and drain of the transistor Tr33.

[0296] At this time, in the current source circuit CS1, the switch SW1 is turned off, and the potential according to the current I3 at the gate of the transistor Tr33 is held by the capacitor C6, thereby allowing the current source circuit CS1 to fix the amount of current output to the wiring BL to I3.

[0297] On the other hand, when the current flowing through the wiring BLref is I4, the current I4 is the sum of the currents flowing between the source and drain of the transistors Tr12 in the memory cells AMref[1] and AMref[2]. Therefore, from equations (E2) and (E4), I4=I AMref[1],0 +I AMref[2],0 It can be said that:

[0298] As a result, in the current mirror circuit CM, a current I4 flows between the source and drain of the transistor Tr31, which causes a current I4 to also flow between the source and drain of the transistor Tr32.

[0299] The current I4 flowing between the source and drain of the transistor Tr32 flows into the current source circuit CS2. Because the switch SW2 is on, the transistor Tr34 is in a diode-connected configuration. As a result, the gate of the transistor Tr34 has a potential according to the current I4, and the current I4 flows between the source and drain of the transistor Tr34.

[0300] Here, in the current source circuit CS2, by turning off the switch SW2, the potential according to the current I4 of the gate of the transistor Tr34 is held by the capacitor C7, thereby allowing the current source circuit CS2 to fix the amount of current output to the line VLE to I4.

[0301] Next, when the operation of the arithmetic circuit MAC1A changes from FIG. 14 to FIG. 15, the potential of the wiring CL[1] becomes V X[1] +REFP and the potential of the line CL[2] is V X[2] It is assumed that it has changed to +REFP.

[0302] At this time, the currents flowing through the transistors Tr12 of the memory cells AM[1], AM[2], AMref[1], and AMref[2] change to equations (E7), (E12), (E8), and (E13).

[0303] Furthermore, as shown in FIG. 15, the switches SW3 and SW4 are each turned on.

[0304] Here, when the current flowing through the wiring BL is I1, the current I1 is the sum of the currents flowing between the source and drain of the transistors Tr12 in the memory cells AM[1] and AM[2]. Therefore, from equations (E7) and (E12), I1=I AM[1],1 +I AM[2],1 It can be said that:

[0305] In addition, when the current flowing through the wiring BLref is I2, the current I2 is the sum of the currents flowing between the source and drain of the transistors Tr12 in the memory cells AMref[1] and AMref[2]. Therefore, from equations (E8) and (E13), I2=I AMref[1],1 +I AMref[2],1 It can be said that:

[0306] As a result, in the current mirror circuit CM, a current I2 flows between the source and drain of the transistor Tr31, which causes a current I2 to also flow between the source and drain of the transistor Tr32.

[0307] Here, since the switch SW4 of the circuit OFAC is in the on state, a current flows between the circuit OFAC and the circuit CMS. When the current flowing between the first terminal and the second terminal of the switch SW4 is I5, I5 = I1 - I2 - I3 + I4 = 2k(V W[1] V X[1] +V W[2] V X[2] ) and the sum of products can be calculated in the same manner as in equation (E16).

[0308] Moreover, an equivalent circuit of the arithmetic circuit MAC1A in Fig. 15 can be the circuit shown in Fig. 16. The current source CI1 shown in Fig. 16 corresponds to the memory cell AM[1] and the memory cell AM[2] in Fig. 15, the current source CI2 shown in Fig. 16 corresponds to the current source circuit CS1, the current source CI3 shown in Fig. 16 corresponds to the current source circuit CS2, and the current source CI4 shown in Fig. 16 corresponds to the current mirror circuit CM2.

[0309] In addition, the memory cell array CA of the arithmetic circuit MAC1A in FIG. 13 is a memory cell array having memory cells arranged in a matrix of 2 rows and 2 columns, but the product-sum operation can be performed in the same manner for a memory cell array with 1 row and 2 or more columns, or a memory cell array with 3 or more rows and 3 or more columns. In this case, the product-sum operation circuit uses one of the multiple columns as reference data (potential V PR ), it is possible to simultaneously perform the multiplication and accumulation operation for the remaining number of columns. Also, by increasing the number of rows, it is possible to increase the number of terms to be added in the multiplication and accumulation operation. In this case, the current I5 is I5 = 2kΣV, as in equation (E18). W[i] V X[i] It can be expressed as:

[0310] Here, in the circuit OFAC, the current I5 is converted to a voltage by a resistor RE. Although not shown in Fig. 13, the circuit OFAC is a circuit that performs calculations according to a predefined activation function in response to the voltage, so that it can perform calculations of a hierarchical neural network, similar to the calculation circuit MAC1 in Fig. 9.

[0311] <Arithmetic circuit configuration example 2> Next, an example of a circuit for performing product-sum operations and activation function operations, which has a circuit configuration different from that of the arithmetic circuit MAC1 in the above-described neural network 100, will be described.

[0312] Fig. 17 shows a configuration example of the arithmetic circuit MAC2. The arithmetic circuit MAC2 shown in Fig. 17 is a circuit that performs a multiplication and accumulation operation on the first data corresponding to the voltage held in each cell and the input second data, and performs an activation function operation using the result of the multiplication and accumulation operation. Note that the first data and the second data can be, for example, analog data or multi-valued data (discrete data).

[0313] The arithmetic circuit MAC2 has a circuit WCS, a circuit XCS, a circuit WSD, a circuit SWS1, a circuit SWS2, a cell array CA2, and conversion circuits ITRZ[1] through ITRZ[m].

[0314] The cell array CA2 has cells IM[1,1] to IM[m,n] (where m is an integer equal to or greater than 1, and n is an integer equal to or greater than 1) and cells IMref[1] to IMref[m]. Cells IM[1,1] to IM[m,n] have a function of holding a potential corresponding to an amount of current corresponding to first data, and cells IMref[1] to IMref[m] have a function of supplying the held potential and a voltage corresponding to second data required for performing a product-sum operation to signal lines XCL[1] to XCL[m].

[0315] In addition, the cell array CA2 in Figure 17 has n+1 cells arranged in a matrix in the row direction and m cells in the column direction, but the cell array CA2 may also be configured to have two or more cells arranged in a matrix in the row direction and one or more cells arranged in the column direction.

[0316] Cells IM[1,1] to IM[m,n] each have a transistor F1, a transistor F2, and a capacitance C5, and cells IMref[1] to IMref[m] each have a transistor F1m, a transistor F2m, and a capacitance C5m.

[0317] In addition, unless otherwise specified, the transistors F1 and F1m are considered to ultimately operate in a linear region when they are in an on state. That is, the gate voltage, source voltage, and drain voltage of each of the above-mentioned transistors are considered to include a case where they are appropriately biased to a voltage within a range in which they operate in a linear region. However, one embodiment of the present invention is not limited to this. For example, the transistors F1 and F1m may operate in a saturation region when they are in an on state, or may operate in both a linear region and a saturation region.

[0318] Furthermore, unless otherwise specified, the transistors F2 and F2m are assumed to operate in the subthreshold region (i.e., the gate-source voltage of the transistor F2 or the transistor F2m is lower than the threshold voltage). In other words, the gate voltage, source voltage, and drain voltage of each of the above-mentioned transistors are assumed to be appropriately biased to a voltage range in which they operate in the subthreshold region. For this reason, the transistors F2 and F2m are assumed to operate such that an off-current flows between the source and drain.

[0319] Also, like the transistor Tr11, the transistor F1 and / or the transistor F1m are preferably OS transistors. In addition, the channel formation region of the transistor F1 and / or the transistor F1m is more preferably an oxide containing at least one of indium, an element M (for example, the element M may be one or more elements selected from aluminum, gallium, yttrium, copper, vanadium, beryllium, boron, titanium, iron, nickel, germanium, zirconium, molybdenum, lanthanum, cerium, neodymium, hafnium, tantalum, tungsten, magnesium, etc.), and zinc. It is further preferable that the transistor Tr1 and / or the transistor F1m has the structure of the transistor described in the third embodiment.

[0320] By using OS transistors as the transistors F1 and / or F1m, the leakage current of the transistors F1 and / or F1m can be suppressed, which may result in a product-sum operation circuit with high calculation accuracy. Furthermore, by using OS transistors as the transistors F1 and / or F1m, the leakage current from the retention node to the write word line when the transistors F1 and / or F1m are in a non-conducting state can be made very small. In other words, the number of refresh operations of the potential of the retention node can be reduced, which can reduce the power consumption of the product-sum operation circuit.

[0321] In addition, by using an OS transistor for the transistor F2 and / or the transistor F2m, the transistors can operate in a wide current range in the subthreshold region, thereby reducing current consumption. In addition, by using an OS transistor for the transistor F2 and / or the transistor F2m, the transistors can be manufactured simultaneously with the transistor Tr11, which may shorten the manufacturing process of the product-sum operation circuit. In addition, the transistor F2 and / or the transistor F2m may be a transistor containing silicon in the channel formation region. As the silicon, for example, amorphous silicon (sometimes called hydrogenated amorphous silicon), microcrystalline silicon, polycrystalline silicon, single crystal silicon, or the like can be used.

[0322] In each of the cells IM[1,1] to IM[m,n], a first terminal of the transistor F1 is electrically connected to a gate of the transistor F2. A first terminal of the transistor F2 is electrically connected to a wiring VE. A first terminal of the capacitor C5 is electrically connected to the gate of the transistor F2.

[0323] Incidentally, one embodiment of the present invention does not depend on the connection configuration of the back gate of a transistor. In FIG. 17, the back gate is illustrated in the transistor F1 and the transistor F2, and the configuration having the back gate is shown. Although the connection configuration of the back gate is not illustrated in FIG. 17, the electrical connection destination of the back gate can be determined at the design stage. For example, in a transistor having a back gate, the gate and the back gate may be electrically connected to increase the on-current of the transistor. That is, for example, the gate and the back gate of the transistor M2 may be electrically connected. Also, for example, in a transistor having a back gate, in order to change the threshold voltage of the transistor or to reduce the off-current of the transistor, a wiring electrically connected to an external circuit or the like may be provided, and a potential may be applied to the back gate of the transistor by the external circuit or the like. Note that this also applies to the transistor F1m, the transistor F2m, the transistors F3[1] to F3[n] and the transistors F4[1] to F4[n] described later, and further to transistors described not only in FIG. 17 but also in other parts of the specification or transistors illustrated in other drawings.

[0324] The semiconductor device of one embodiment of the present invention does not depend on the structure of a transistor included in the semiconductor device. For example, the transistor F1 and the transistor F2 illustrated in FIG. 17 may have a structure without a backgate, that is, a single-gate structure, as illustrated in FIG. 17. Some of the transistors may have a backgate, and other transistors may not have a backgate. This also applies to the transistors F1m, F2m, the transistors F3[1] to F3[n], and the transistors F4[1] to F4[n] described later, and further to the transistors described in other parts of the specification or illustrated in other drawings, as well as the circuit diagram illustrated in FIG. 17.

[0325] The wiring VE is a wiring for passing a current between the first terminal and the second terminal of each of the transistors F2 of the cells IM[1,1], IM[m,1], IM[1,n], and IM[m,n], and also functions as a wiring for passing a current between the first terminal and the second terminal of each of the transistors F2 of the cells IMref[1] and IMref[m]. As an example, the wiring VE functions as a wiring for supplying a constant voltage. The constant voltage can be, for example, a low-level potential, a ground potential, or the like.

[0326] In the cell IM[1,1], the second terminal of the transistor F1 is electrically connected to the wiring WCL[1], and the gate of the transistor F1 is electrically connected to the wiring WSL[1]. The second terminal of the transistor F2 is electrically connected to the wiring WCL[1], and the second terminal of the capacitor C5 is electrically connected to the wiring XCL[1]. In FIG. 17, the connection point of the first terminal of the transistor F1, the gate of the transistor F2, and the first terminal of the capacitor C5 in the cell IM[1,1] is the node NN[1,1].

[0327] In the cell IM[m,1], the second terminal of the transistor F1 is electrically connected to the wiring WCL[1], and the gate of the transistor F1 is electrically connected to the wiring WSL[m]. The second terminal of the transistor F2 is electrically connected to the wiring WCL[1], and the second terminal of the capacitor C5 is electrically connected to the wiring XCL[m]. In FIG. 17, the connection point of the first terminal of the transistor F1, the gate of the transistor F2, and the first terminal of the capacitor C5 in the cell IM[m,1] is the node NN[m,1].

[0328] In the cell IM[1,n], the second terminal of the transistor F1 is electrically connected to the wiring WCL[n], and the gate of the transistor F1 is electrically connected to the wiring WSL[1]. The second terminal of the transistor F2 is electrically connected to the wiring WCL[n], and the second terminal of the capacitor C5 is electrically connected to the wiring XCL[1]. In FIG. 17, the connection point of the first terminal of the transistor F1, the gate of the transistor F2, and the first terminal of the capacitor C5 in the cell IM[1,n] is the node NN[1,n].

[0329] In the cell IM[m,n], the second terminal of the transistor F1 is electrically connected to the wiring WCL[n], and the gate of the transistor F1 is electrically connected to the wiring WSL[m]. The second terminal of the transistor F2 is electrically connected to the wiring WCL[n], and the second terminal of the capacitor C5 is electrically connected to the wiring XCL[m]. In FIG. 17, the connection point of the first terminal of the transistor F1, the gate of the transistor F2, and the first terminal of the capacitor C5 in the cell IM[m,n] is the node NN[m,n].

[0330] In the cell IMref[1], the second terminal of the transistor F1m is electrically connected to the wiring XCL[1], and the gate of the transistor F1m is electrically connected to the wiring WSL[1]. The second terminal of the transistor F2m is electrically connected to the wiring XCL[1], and the second terminal of the capacitor C5 is electrically connected to the wiring XCL[1]. In FIG. 17, the connection point of the first terminal of the transistor F1m, the gate of the transistor F2m, and the first terminal of the capacitor C5 in the cell IMref[1] is the node NNref[1].

[0331] In the cell IMref[m], the second terminal of the transistor F1m is electrically connected to the wiring XCL[m], and the gate of the transistor F1m is electrically connected to the wiring WSL[m]. The second terminal of the transistor F2m is electrically connected to the wiring XCL[m], and the second terminal of the capacitor C5 is electrically connected to the wiring XCL[m]. In FIG. 17, the connection point of the first terminal of the transistor F1m, the gate of the transistor F2m, and the first terminal of the capacitor C5 in the cell IMref[m] is the node NNref[m].

[0332] The above-mentioned nodes NN[1,1], NN[m,1], NN[1,n], NN[m,n], NNref[1], and NMref[m] function as holding nodes for the respective cells.

[0333] The circuit SWS1 includes transistors F3[1] to F3[n]. A first terminal of the transistor F3[1] is electrically connected to the wiring WCL[1], a second terminal of the transistor F3[1] is electrically connected to the circuit WCS, and a gate of the transistor F3[1] is electrically connected to the wiring SWL1. A first terminal of the transistor F3[m] is electrically connected to the wiring WCL[m], a second terminal of the transistor F3[m] is electrically connected to the circuit WCS, and a gate of the transistor F3[m] is electrically connected to the wiring SWL1.

[0334] The transistors F3[1] to F3[n] are preferably OS transistors, similar to the transistor Tr11. In addition, the channel formation region of the transistor F1 and / or the transistor F1m is more preferably an oxide containing at least one of indium, an element M (for example, the element M may be one or more elements selected from aluminum, gallium, yttrium, copper, vanadium, beryllium, boron, titanium, iron, nickel, germanium, zirconium, molybdenum, lanthanum, cerium, neodymium, hafnium, tantalum, tungsten, magnesium, etc.), and zinc. The transistors F3[1] to F3[n] are more preferably structured as the transistors described in the third embodiment.

[0335] The circuit SWS1 functions as a circuit that switches between a conductive state and a non-conductive state between the circuit WCS and each of the wirings WCL[1] to WCL[n].

[0336] The circuit SWS2 includes transistors F4[1] to F4[n]. A first terminal of the transistor F4[1] is electrically connected to the wiring WCL[1], a second terminal of the transistor F4[1] is electrically connected to the conversion circuit ITRZ[1], and a gate of the transistor F4[1] is electrically connected to the wiring SWL2. A first terminal of the transistor F4[m] is electrically connected to the wiring WCL[m], a second terminal of the transistor F4[m] is electrically connected to the conversion circuit ITRZ[1], and a gate of the transistor F4[m] is electrically connected to the wiring SWL2.

[0337] The transistors F4[1] to F4[n] are preferably OS transistors, similar to the transistor Tr11. In addition, the channel formation region of the transistor F1 and / or the transistor F1m is more preferably an oxide containing at least one of indium, an element M (for example, the element M may be one or more elements selected from aluminum, gallium, yttrium, copper, vanadium, beryllium, boron, titanium, iron, nickel, germanium, zirconium, molybdenum, lanthanum, cerium, neodymium, hafnium, tantalum, tungsten, magnesium, etc.), and zinc. The transistors F4[1] to F4[n] are more preferably structured as the transistors described in the third embodiment.

[0338] The circuit SWS2 functions as a circuit that switches the conductive state and the non-conductive state between the wiring WCL[1] and the circuit ITRZ[1] and between the wiring WCL[n] and the circuit ITRZ[n].

[0339] The circuit WCS has a function of transmitting data to be stored in each cell of the cell array CA2.

[0340] The circuit XCS is electrically connected to the wirings XCL[1] to XCL[m]. The circuit XCS has a function of supplying a current corresponding to the reference data or a current corresponding to the second data to each of the cells IMref[1] to IMref[m] included in the cell array CA2.

[0341] The circuit WSD is electrically connected to the wirings WSL[1] to WSL[m]. When writing first data to a cell in the cell array CA2, the circuit WSD has a function of selecting a memory cell to which data is to be written by transmitting a predetermined signal to the wirings WSL[1] to WSL[m].

[0342] The circuit WSD is electrically connected to the wiring SWL1 and the wiring SWL2. The circuit WSD has a function of bringing the circuit WCS and the cell array CA2 into a conductive state or a non-conductive state by transmitting a predetermined signal to the wiring SWL1, and a function of bringing the conversion circuits ITRZ[1] to ITRZ[m] and the cell array CA2 into a conductive state or a non-conductive state by transmitting a predetermined signal to the wiring SWL2.

[0343] Each of the conversion circuits ITRZ[1] to ITRZ[m] has an input terminal and an output terminal. Each of the conversion circuits ITRZ[1] to ITRZ[m] has a function of converting a current input to the input terminal into a voltage corresponding to the current and outputting the voltage from the output terminal. As an example, the circuit OFST can be applied to each of the conversion circuits ITRZ[1] to ITRZ[m]. Each of the conversion circuits ITRZ[1] to ITRZ[m] may have an activation function circuit ACTV, and may use the converted voltage to perform an activation function calculation and output the result of the calculation to the output terminal.

[0344] <Operation example 2 of the calculation circuit> Next, an example of the operation of the arithmetic circuit MAC2 will be described.

[0345] Fig. 18 shows a timing chart of an operation example of the arithmetic circuit MAC2. The timing chart in Fig. 18 shows the fluctuations in the potentials of the wiring SWL1, wiring SWL2, wiring WSL[i] (i is an integer between 1 and m-1), wiring WSL[i+1], wiring XCL[i], wiring XCL[i+1], node NN[i,j], node NNref[i], node NN[i+1,j], and node NNref[i+1] between time T11 and time T23 and in the vicinity thereof. Furthermore, the timing chart in Fig. 18 shows the amount of current I flowing between the first terminal and the second terminal of the transistor F2 included in the cell IM[i,j]. F2 [i,j] and the amount of current I flowing between the first and second terminals of transistor F2m included in cell IMref[i] F2m[i] and the amount of current I flowing between the first and second terminals of transistor F2 included in cell IM[i+1,j] F2 [i+1,j] and the amount of current I flowing between the first and second terminals of transistor F2m included in cell IMref[i+1] F2m The respective variations of [i+1] and are also shown.

[0346] In this operation example, the potential of the wiring VE is the ground potential GND. Also, before time T11, the transistors F1 included in the cells IM[i,j] and IM[i+1,j], and the transistors F1m included in the cells IMref[i] and IMref[i+1] are turned on, and the potentials of the nodes NN[i,j], NNref[i], NN[i+1,j], and NNref[i+1] are set to the ground potential GND.

[0347] Also, as an initial setting, transistors F1 included in cells IM[1,1] to IM[m,n] and transistors F1m included in cells IMref[1] to IMref[m] are turned on, and the potentials of nodes NN[1,1] to NN[m,n] and nodes NNref[1] to NNref[m] are set to the ground potential GND.

[0348] <<From time T11 to time T12>> Between time T11 and time T12, a high-level potential (denoted as High in FIG. 18) is applied to the wiring SWL1, and a low-level potential (denoted as Low in FIG. 18) is applied to the wiring SWL2. As a result, a high-level potential is applied to the gates of the transistors F3[1] through F3[n], turning the transistors F3[1] through F3[n] on, and a low-level potential is applied to the gates of the transistors F4[1] through F4[n], turning the transistors F4[1] through F4[n] off.

[0349] Also, between time T11 and time T12, a low-level potential is applied to the wiring WSL[i] and the wiring WSL[i+1]. As a result, a low-level potential is applied to the gate of the transistor F1 included in the cells IM[i,1] to IM[i,n] in the i-th row of the cell array CA2 and the gate of the transistor F1m included in the cell IMref[i], and the transistors F1 and F1m are turned off. Also, a low-level potential is applied to the gate of the transistor F1 included in the cells IM[i+1,1] to IM[i+1,n] in the i+1-th row of the cell array CA2 and the gate of the transistor F1m included in the cell IMref[i+1], and the transistors F1 and F1m are turned off.

[0350] Between time T11 and time T12, the ground potential GND is applied to the wiring XCL[i] and the wiring XCL[i+1].

[0351] In addition, between time T11 and time T12, no current flows through the wiring WCL[j], the wiring XCL[i], and the wiring XCL[i+1]. F2 [i,j], I F2m [i]I F2 [i+1,j], I F2m [i+1] is 0.

[0352] <<From time T12 to time T13>> Between time T12 and time T13, a high-level potential is applied to the wiring WSL[i]. As a result, a high-level potential is applied to the gate of the transistor F1 included in the cells IM[i,1] to IM[i,n] in the i-th row of the cell array CA2 and the gate of the transistor F1m included in the cell IMref[i], and the transistors F1 and F1m are turned on. In addition, between time T12 and time T13, a low-level potential is applied to the wirings WSL[1] to WSL[m] other than the wiring WSL[i], and the transistors F1 included in the cells IM[1,1] to IM[m,n] other than the i-th row of the cell array CA2 and the transistors F1m included in the cells IMref[1] to IMref[m] other than the i-th row are turned off.

[0353] Furthermore, a low-level potential is applied to the wirings XCL[1] to XCL[m].

[0354] <<From time T13 to time T14>> Between time T13 and time T14, a current of I0[i,j] flows from the circuit WCS to the cell array CA2 via the transistor F3[j]. At this time, a conductive state is established between the first terminal of the transistor F1 included in the cell IM[i,j] in the i-th row of the cell array CA2 and the wiring WCL[j], and a non-conductive state is established between the first terminal of the transistor F1 included in the cells IM[1,j] to IM[m,j] other than the i-th row of the cell array CA2 and the wiring WCL[j], so that a current of I0[i,j] flows from the wiring WCL[j] to the cell IM[i,j].

[0355] Incidentally, when the transistor F1 included in the cell IM[i,j] is turned on, the transistor F2 included in the cell IM[i,j] is configured as a diode connection. Therefore, when a current flows from the wiring WCL[j] to the cell IM[i,j], the potentials of the gate of the transistor F2 and the second terminal of the transistor F2 become approximately equal. This potential is determined by the amount of current flowing from the wiring WCL[j] to the cell IM[i,j] and the potential of the first terminal of the transistor F2 (GND in this case). In this operation example, when a current of the amount I0[i,j] flows from the wiring WCL[j] to the cell IM[i,j], the potential of the gate of the transistor F2 (node ​​NN[i,j]) becomes V g [i,j]. In other words, the gate-source voltage of transistor F2 is V g [i,j]-GND, and a current of an amount I0[i,j] flows between the first terminal and the second terminal of the transistor F2.

[0356] Here, the threshold voltage of the transistor F2 is V th Then, the amount of current I0[i,j] when the transistor F2 operates in the subthreshold region can be expressed as follows:

[0357]

number

[0358] In addition, I a is V g V th where K is a correction coefficient determined by temperature, device structure, etc.

[0359] In addition, between time T13 and time T14, the current I ref0 At this time, a current of I flows from the line XCL[i] to the cell IMref[i] because the first terminal of the transistor F1m included in the cell IMref[i] is in a conductive state with the line XCL[i].ref0 A current of flows.

[0360] As with cell IM[i,j], when transistor F1m included in cell IMref[i] is turned on, transistor F2m included in cell IMref[i,j] is configured as a diode connection. Therefore, when a current flows from wiring XCL[i] to cell IMref[i], the potentials of the gate of transistor F2m and the second terminal of transistor F2m become approximately equal. This potential is determined by the amount of current flowing from wiring XCL[i] to cell IMref[i] and the potential of the first terminal of transistor F2m (GND in this case), etc. In this operation example, a current amount I ref0 When a current of V flows, the gate of transistor F2 (node ​​NNref[i]) is V gm [i], and the potential of the wiring XCL[i] at this time is also V gm [i]. That is, in transistor F2m, the gate-source voltage is V gm [i]-GND, and a current I flows between the first and second terminals of the transistor F2m. ref0 A current of flows.

[0361] Here, the threshold voltage of the transistor F2m is V thm [i], the amount of current I when transistor F2m operates in the subthreshold region ref0 can be written as the following formula. Note that the correction coefficient K is the same as that of transistor F2 included in cell IM[i,j]. For example, the device structure and size (channel length, channel width) of the transistors are the same. In addition, the correction coefficient K of each transistor varies due to manufacturing variations, but it is assumed that the variation is suppressed to a level that allows the discussion below to be carried out with sufficient accuracy for practical use.

[0362]

number

[0363] Here, the weighting coefficient w[i,j], which is the first data, is defined as follows.

[0364]

number

[0365] Therefore, equation (F1) can be rewritten as follows:

[0366]

number

[0367] <<From time T14 to time T15>> Between time T14 and time T15, a low-level potential is applied to the wiring WSL[i], so that a low-level potential is applied to the gates of the transistors F1 included in the cells IM[i,1] to IM[i,n] in the i-th row of the cell array CA2 and the gate of the transistor F1m included in the cell IMref[i], turning off the transistors F1 and F1m.

[0368] When the transistor F1 included in the cell IM[i,j] is turned off, the capacitance C5 is charged to V g [i,j]-V gm [i] is held. In addition, by turning off the transistor F1 included in the cell IMref[i], the capacitance C5m holds 0, which is the difference between the potential of the gate of the transistor F2m (node ​​NNref[i]) and the potential of the wiring XCL[i]. Note that the potential held by the capacitance C5m may be a potential other than 0 (here, Δ) depending on the transistor characteristics of the transistors F1m and F2m during the operation from time T13 to time T14. However, the following discussion holds if we consider the potential of the node NNref[i] to be a potential obtained by adding Δ to the potential of the wiring XCL[i].

[0369] <<From time T15 to time T16>> Between time T15 and time T16, GND is applied to the line XCL[i]. As a result, the potentials of the nodes NN[i,1] to NN[i,n] change due to the capacitive coupling of the capacitor C5 included in each of the cells IM[i,1] to IM[i,n] in the i-th row, and the potential of the node NNref[i] changes due to the capacitive coupling of the capacitor C5 included in the cell IMref[i].

[0370] The change in potential of nodes NN[i,1] to NN[i,n] is calculated by multiplying the change in potential of line XCL[i] by a capacitive coupling coefficient determined by the configuration of each of cells IM[i,1] to IM[i,n] included in cell array CA2. The capacitive coupling coefficient is calculated from the capacitance of capacitor C5, the gate capacitance of transistor F2, parasitic capacitance, etc. In each of cells IM[i,1] to IM[i,n], when the capacitive coupling coefficient of capacitor C5 is p, the potential of node NN[i,j] of cell IM[i,j] is calculated by multiplying the potential at the time point between time T14 and time T15 by p(V gm [i]-GND) decreases.

[0371] Similarly, when the potential of the wiring XCL[i] changes, the potential of the node NNref[i] also changes due to the capacitive coupling of the capacitance C5m included in the cell IMref[i]. If the capacitive coupling coefficient of the capacitance C5m is p, like the capacitance C5, the potential of the node NNref[i] of the cell IMref[i] changes from the potential at the time point between time T14 and time T15 to p(V gm [i]-GND) decreases.

[0372] As a result, the potential of the node NN[i,j] of the cell IM[i,j] drops, so that the transistor F2 is turned off. Similarly, the potential of the node NNref[i] of the cell IMref[i] drops, so that the transistor F2m is also turned off. Therefore, between time T15 and time T16, F2 [i,j], IF2m Each of [i] will be 0.

[0373] <<From time T16 to time T17>> Between time T16 and time T17, a high-level potential is applied to the wiring WSL[i+1]. As a result, a high-level potential is applied to the gate of the transistor F1 included in the cells IM[i+1,1] to IM[i+1,n] in the i+1th row of the cell array CA2 and the gate of the transistor F1m included in the cell IMref[i+1], and the transistors F1 and F1m are turned on. Also, between time T16 and time T17, a low-level potential is applied to the wirings WSL[1] to WSL[m] except for the wiring WSL[i+1], and the transistors F1 included in the cells IM[1,1] to IM[m,n] other than the i+1th row of the cell array CA2 and the transistors F1m included in the cells IMref[1] to IMref[m] other than the i+1th row of the cell array CA2 are turned off.

[0374] Furthermore, a low-level potential is applied to the wirings XCL[1] to XCL[m].

[0375] <<From time T17 to time T18>> Between time T17 and time T18, a current of I0[i+1,j] flows from the circuit WCS to the cell array CA2 via the transistor F3[j]. At this time, a first terminal of the transistor F1 included in the cell IM[i+1,j] in the i+1th row of the cell array CA2 is in a conductive state between the wiring WCL[j], and a first terminal of the transistor F1 included in the cells IM[1,j] to IM[m,j] other than the i+1th row of the cell array CA2 is in a non-conductive state between the wiring WCL[j], so that a current of I0[i+1,j] flows from the wiring WCL[j] to the cell IM[i+1,j].

[0376] Incidentally, when the transistor F1 included in the cell IM[i+1,j] is turned on, the transistor F2 included in the cell IM[i+1,j] is configured as a diode connection. Therefore, when a current flows from the wiring WCL[j] to the cell IM[i+1,j], the potentials of the gate of the transistor F2 and the second terminal of the transistor F2 become almost equal. The potential is determined by the amount of current flowing from the wiring WCL[j] to the cell IM[i+1,j] and the potential of the first terminal of the transistor F2 (GND in this case). In this operation example, when a current of the amount I0[i+1,j] flows from the wiring WCL[j] to the cell IM[i+1,j], the potential of the gate of the transistor F2 (node ​​NN[i+1,j]) becomes V g [i+1,j]. In other words, the gate-source voltage of transistor F2 is V g [i+1,j]-GND, and a current of I0[i+1,j] flows between the first terminal and the second terminal of the transistor F2.

[0377] Here, the threshold voltage of transistor F2 is V th When the transistor F2 operates in the subthreshold region, the current I0[i+1,j] can be expressed as follows: Note that the correction coefficient is K, which is the same as for the transistor F2 included in the cell IM[i,j] and the transistor F2m included in the cell IMref[i].

[0378]

number

[0379] In addition, between time T17 and time T18, the current I ref0 At this time, a current of I flows from the line XCL[i+1] to the cell IMref[i+1] because the first terminal of the transistor F1m included in the cell IMref[i+1] is in a conductive state. ref0 A current of flows.

[0380] As with cell IM[i+1,j], when transistor F1m included in cell IMref[i+1] is turned on, transistor F2m included in cell IMref[i+1,j] is configured as a diode connection. Therefore, when a current flows from wiring XCL[i+1] to cell IMref[i+1], the potentials of the gate of transistor F2m and the second terminal of transistor F2m become approximately equal. This potential is determined by the amount of current flowing from wiring XCL[i+1] to cell IMref[i+1] and the potential of the first terminal of transistor F2m (GND in this case), etc. In this operation example, a current amount I ref0 As a result of this current flowing, the gate of transistor F2 (node ​​NNref[i+1]) is V gm [i+1], and the potential of the wiring XCL[i+1] at this time is also V gm [i+1]. In other words, in transistor F2m, the gate-source voltage is V gm [i+1]-GND, and a current I ref0 A current of flows.

[0381] Here, the threshold voltage of the transistor F2m is V thm When [i+1,j], the amount of current I when transistor F2m operates in the subthreshold region is ref0 can be written as follows: Note that the correction coefficient K is the same as that of the transistor F2 included in the cell IM[i+1,j].

[0382]

number

[0383] Here, the weighting coefficient w[i+1,j], which is the first data, is defined as follows.

[0384]

number

[0385] Therefore, equation (F5) can be rewritten as

[0386]

number

[0387] <<From time T18 to time T19>> Between time T18 and time T19, a low-level potential is applied to the wiring WSL[i+1], so that a low-level potential is applied to the gates of the transistors F1 included in the cells IM[i+1,1] to IM[i+1,n] in the i-th row of the cell array CA2 and the gate of the transistor F1m included in the cell IMref[i+1], turning off the transistors F1 and F1m.

[0388] When the transistor F1 included in the cell IM[i+1,j] is turned off, the capacitance C5 is charged to V g [i+1,j]-V gm [i+1] is held. In addition, by turning off the transistor F1 included in the cell IMref[i+1], the capacitance C5m holds 0, which is the difference between the potential of the gate of the transistor F2m (node ​​NNref[i+1]) and the potential of the wiring XCL[i+1]. Note that the potential held by C5m may become a potential other than 0 (here, Δ) depending on the transistor characteristics of the transistors F1m and F2m during the operation from time T18 to time T19. However, the following discussion holds if we consider the potential of the node NNref[i] to be a potential obtained by adding Δ to the potential of the wiring XCL[i].

[0389] <<From time T19 to time T20>> Between time T19 and time T20, GND is applied to the line XCL[i+1]. As a result, the potentials of the nodes NN[i,1] to NN[i+1,n] change due to the capacitive coupling of the capacitor C5 included in each of the cells IM[i+1,1] to IM[i+1,n] in the i+1th row, and the potential of the node NNref[i+1] changes due to the capacitive coupling of the capacitor C5 included in the cell IMref[i+1].

[0390] The change in the potential of the nodes NN[i+1,1] to NN[i+1,n] is calculated by multiplying the change in the potential of the wiring XCL[i+1] by a capacitive coupling coefficient determined by the configuration of each of the cells IM[i+1,1] to IM[i+1,n] included in the cell array CA2. The capacitive coupling coefficient is calculated from the capacitance of the capacitor C5, the gate capacitance of the transistor F2, the parasitic capacitance, and the like. When the capacitive coupling coefficient of the capacitor C5 in each of the cells IM[i+1,1] to IM[i+1,n] is p, which is the same as the capacitive coupling coefficient of the capacitor C5 in each of the cells IM[i,1] to IM[i,n], the potential of the node NN[i+1,j] of the cell IM[i+1,j] is calculated by multiplying the change in the potential of the wiring XCL[i+1] by p(V gm [i+1]-GND) decreases.

[0391] Similarly, when the potential of the wiring XCL[i+1] changes, the potential of the node NNref[i+1] also changes due to the capacitive coupling of the capacitance C5m included in the cell IMref[i+1]. If the capacitive coupling coefficient of the capacitance C5m is p, like the capacitance C5, the potential of the node NNref[i+1] of the cell IMref[i+1] changes from the potential at the time point between time T18 and time T19 to p(V gm [i+1]-GND) decreases.

[0392] As a result, the potential of the node NN[i+1,j] of the cell IM[i+1,j] drops, so that the transistor F2 is turned off. Similarly, the potential of the node NNref[i] of the cell IMref[i+1] drops, so that the transistor F2m is also turned off. Therefore, between time T19 and time T20, F2 [i+1,j], I F2m Each of [i+1] is 0.

[0393] <<From time T20 to time T21>> Between time T20 and time T21, a low-level potential is applied to the wiring SWL1, so that a low-level potential is applied to each of the gates of the transistors F3[1] to F3[n], and each of the transistors F3[1] to F3[n] is turned off.

[0394] <<From time T21 to time T22>> A high-level potential is applied to the wiring SWL2 from time T21 to time T22. As a result, a high-level potential is applied to each of the gates of the transistors F4[1] to F4[n], and each of the transistors F4[1] to F4[n] is turned off.

[0395] <<From time T22 to time T23>> Between time T22 and time T23, the current I ref0 x[i] times I ref0 In this operation example, x corresponds to the value of the neuron signal, which is the second data. At this time, the potential of the wiring XCL[i] ranges from 0 to V gm [i] + ΔV[i].

[0396] When the potential of the wiring XCL[i] changes, the potentials of the nodes NN[i,1] to NN[i,n] also change due to capacitive coupling by the capacitor C5 included in each of the cells IM[i,1] to IM[i,n] in the i-th row of the cell array CA2. Therefore, the potential of the node NN[i,j] of the cell IM[i,j] is V g [i,j]+pΔV[i].

[0397] Similarly, when the potential of the wiring XCL[i] changes, the potential of the node NNref[i] also changes due to the capacitive coupling of the capacitance C5m included in the cell IMref[i]. Therefore, the potential of the node NNref[i] of the cell IMref[i] is V gm [i]+pΔV[i].

[0398] As a result, during the period from time T22 to time T23, the current I1[i,j] flowing between the first terminal and the second terminal of the transistor F2 and the current I ref1 [i,j] can be written as follows:

[0399]

number

[0400]

number

[0401] From equations (F9) and (F10), x[i] can be expressed by the following equation.

[0402]

number

[0403] Therefore, equation (F9) can be rewritten as

[0404]

number

[0405] In other words, the current flowing between the first and second terminals of transistor F2 included in cell IM[i,j] is proportional to the product of the weighting coefficient w[i,j], which is the first data, and the value x[i] of the neuron signal, which is the second data.

[0406] In addition, between time T22 and time T23, the current I ref0 x[i+1] is x[i+1] times I ref0 In this operation example, x corresponds to the value of the neuron signal, which is the second data. At this time, the potential of the wire XCL[i+1] changes from 0 to V gm [i+1] + ΔV[i+1].

[0407] When the potential of the wiring XCL[i+1] changes, the potentials of the nodes NN[i+1,1] to NN[i+1,n] also change due to the capacitive coupling of the capacitor C5 included in each of the cells IM[i+1,1] to IM[i+1,n] in the i+1th row of the cell array CA2. Therefore, the potential of the node NN[i+1,j] of the cell IM[i+1,j] is V g [i+1,j]+pΔV[i+1].

[0408] Similarly, when the potential of the wiring XCL[i+1] changes, the potential of the node NNref[i+1] also changes due to the capacitive coupling of the capacitance C5m included in the cell IMref[i+1]. Therefore, the potential of the node NNref[i+1] of the cell IMref[i+1] is V gm [i+1]+pΔV[i+1].

[0409] As a result, during the period from time T22 to time T23, the current I1[i+1,j] flowing between the first terminal and the second terminal of the transistor F2 and the current I ref1 [i+1,j] can be written as follows:

[0410]

number

[0411]

number

[0412] From equations (F13) and (F14), x[i+1] can be expressed by the following equation.

[0413]

number

[0414] Therefore, equation (F13) can be rewritten as

[0415]

number

[0416] In other words, the current flowing between the first and second terminals of transistor F2 included in cell IM[i+1,j] is proportional to the product of the weighting coefficient w[i+1,j], which is the first data, and the value x[i+1] of the neuron signal, which is the second data.

[0417] Here, consider the total current flowing from the conversion circuit ITRZ[j] to the cells IM[i,j] and IM[i+1,j] via the transistor F4[j] and the wiring WCL[j]. Let I S [j], then I S [j] can be expressed by the following equation using equations (F12) and (F16).

[0418]

number

[0419] Therefore, the current output from the conversion circuit ITRZ[j] is proportional to the sum of the products of the first data, the weighting coefficients w[i,j] and w[i+1,j], and the second data, the neuron signal values ​​x[i] and x[i+1].

[0420] In the above operation example, the sum of the currents flowing through the cells IM[i,j] and IM[i+1,j] is treated, but the sum of the currents flowing through the cells IM[1,j] to IM[m,j] may also be treated as a plurality of cells. In this case, the formula (F17) can be rewritten as the following formula.

[0421]

number

[0422] Therefore, even in the case of an arithmetic circuit MAC2 having a cell array CA2 with three or more rows and two or more columns, it is possible to perform a product-sum operation as described above. In this case, the product-sum operation circuit selects one of the multiple columns as a current amount I ref0 , and xI ref0 In other words, by increasing the number of columns in the memory cell array, a semiconductor device that realizes high-speed product-sum calculation can be provided.

[0423] When the multiplication and accumulation circuit described in this embodiment is applied to the hidden layer, the weight coefficient w s[k]s[k-1] (k) is used as the first data, the current amount according to the first data is stored in each cell IM in the same column, and the output signal z s[k-1] (k-1) is used as the second data, and a current corresponding to the second data is passed from the circuit XCS to the wiring XCL of each row, so that the current I SIn addition, the sum of products of the first data and the second data can be calculated by using the sum of products to calculate the value of the activation function, and the output signal z s[k] (k) It can be said that:

[0424] In addition, when the multiply-and-accumulate circuit described in this embodiment is applied to the output layer, the weight coefficient w s[L]s[L-1] (L) is used as the first data, and the current amount according to the first data is stored in each cell IM in the same column, and the output signal z s[L-1] (L-1) is used as the second data, and a current corresponding to the second data is passed from the circuit XCS to the wiring XCL of each row, so that the current I S In addition, the sum of products of the first data and the second data can be calculated by using the sum of products to calculate the value of the activation function, and the value of the activation function is used as a signal to generate the output signal z s[L] (L) It can be said that:

[0425] The input layer described in this embodiment may function as a buffer circuit that outputs an input signal to the second layer.

[0426] In addition, in this embodiment, the transistors included in the arithmetic circuits MAC1 and MAC2 are OS transistors or Si transistors, but one embodiment of the present invention is not limited thereto. The transistors included in the arithmetic circuits MAC1 and MAC2 can be, for example, transistors with a semiconductor such as Ge as an active layer, transistors with a compound semiconductor such as ZnSe, CdS, GaAs, InP, GaN, or SiGe as an active layer, transistors with a carbon nanotube as an active layer, transistors with an organic semiconductor as an active layer, or the like.

[0427] Note that this embodiment mode can be appropriately combined with other embodiment modes described in this specification.

[0428] (Embodiment 3) In this embodiment, a configuration example of the arithmetic circuit described in the above embodiment and a configuration example of a transistor that can be applied to the arithmetic circuit will be described.

[0429] <Example of semiconductor device configuration> 19 includes a transistor 300, a transistor 500, and a capacitor 600. Fig. 21A is a cross-sectional view of the transistor 500 in the channel length direction, Fig. 21B is a cross-sectional view of the transistor 500 in the channel width direction, and Fig. 21C is a cross-sectional view of the transistor 300 in the channel width direction.

[0430] The transistor 500 is a transistor having a metal oxide in a channel formation region (OS transistor). Since the off-state current of the transistor 500 is small, by using the transistor 500 for a semiconductor device, such as a transistor Tr11 of a memory cell array CA included in an arithmetic circuit MAC1, data written therein can be held for a long period of time. In other words, the frequency of refresh operations is low or no refresh operations are required, so that the power consumption of the semiconductor device can be reduced.

[0431] 19, the semiconductor device described in this embodiment includes a transistor 300, a transistor 500, and a capacitor 600. The transistor 500 is provided above the transistor 300, and the capacitor 600 is provided above the transistors 300 and 500. Note that the capacitor 600 can be a capacitance C1 of the memory cell array CA included in the arithmetic circuit MAC1 or a capacitance C2 of the circuit OFST described in the above embodiment.

[0432] The transistor 300 is provided over a substrate 311 and includes a conductor 316, an insulator 315, a semiconductor region 313 formed of part of the substrate 311, and a low-resistance region 314a and a low-resistance region 314b functioning as a source region or a drain region. Note that the transistor 300 can be applied to, for example, a transistor Tr12 of a memory cell array CA included in the arithmetic circuit MAC1 described in the above embodiment.

[0433] As the substrate 311, it is preferable to use a semiconductor substrate (for example, a single crystal substrate or a silicon substrate).

[0434] 21C, the upper surface and the side surface in the channel width direction of the semiconductor region 313 of the transistor 300 are covered with a conductor 316 via an insulator 315. By forming the transistor 300 as a fin type in this manner, the effective channel width is increased, thereby improving the on-characteristics of the transistor 300. In addition, the contribution of the electric field of the gate electrode can be increased, thereby improving the off-characteristics of the transistor 300.

[0435] The transistor 300 may be either a p-channel type or an n-channel type.

[0436] The region where the channel of the semiconductor region 313 is formed, the region nearby, the low resistance region 314a which becomes the source region or the drain region, and the low resistance region 314b preferably contain a semiconductor such as a silicon-based semiconductor, and preferably contain single crystal silicon. Alternatively, they may be formed of a material containing Ge (germanium), SiGe (silicon germanium), GaAs (gallium arsenide), GaAlAs (gallium aluminum arsenide), or the like. A configuration using silicon in which the effective mass is controlled by applying stress to the crystal lattice and changing the lattice spacing may be used. Alternatively, the transistor 300 may be a HEMT (High Electron Mobility Transistor) by using GaAs and GaAlAs, or the like.

[0437] The low resistance region 314a and the low resistance region 314b contain, in addition to the semiconductor material applied to the semiconductor region 313, an element that imparts n-type conductivity, such as arsenic or phosphorus, or an element that imparts p-type conductivity, such as boron.

[0438] The conductor 316 functioning as the gate electrode can be made of a conductive material such as a semiconductor material, such as silicon, containing an element that imparts n-type conductivity, such as arsenic or phosphorus, or an element that imparts p-type conductivity, such as boron, a metal material, an alloy material, or a metal oxide material.

[0439] Since the work function is determined by the material of the conductor, the threshold voltage of the transistor can be adjusted by selecting the material of the conductor. Specifically, it is preferable to use a material such as titanium nitride or tantalum nitride for the conductor. Furthermore, in order to achieve both electrical conductivity and embeddability, it is preferable to use a metal material such as tungsten or aluminum as the conductor in a laminated state, and in particular, it is preferable to use tungsten in terms of heat resistance.

[0440] 19 is just an example and is not limited to the structure, and an appropriate transistor may be used depending on the circuit configuration and driving method. For example, when the semiconductor device is a unipolar circuit including only OS transistors, the structure of the transistor 300 may be the same as that of a transistor 500 including an oxide semiconductor, as shown in FIG 20. Details of the transistor 500 will be described later.

[0441] An insulator 320, an insulator 322, an insulator 324, and an insulator 326 are stacked in this order over the transistor 300.

[0442] As the insulators 320, 322, 324, and 326, for example, silicon oxide, silicon oxynitride, silicon nitride oxide, silicon nitride, aluminum oxide, aluminum oxynitride, aluminum nitride oxide, aluminum nitride, or the like can be used.

[0443] In this specification, silicon oxynitride refers to a material having a higher oxygen content than nitrogen, silicon nitride oxide refers to a material having a higher nitrogen content than oxygen, aluminum oxynitride refers to a material having a higher oxygen content than nitrogen, and aluminum nitride oxide refers to a material having a higher nitrogen content than oxygen.

[0444] The insulator 322 may function as a planarizing film that planarizes steps caused by the transistor 300 provided thereunder. For example, the top surface of the insulator 322 may be planarized by a planarization process using a chemical mechanical polishing (CMP) method or the like to improve the planarity.

[0445] The insulator 324 is preferably a film having a barrier property that prevents diffusion of hydrogen or impurities from the substrate 311 or the transistor 300 to a region where the transistor 500 is provided.

[0446] As an example of a film having a barrier property against hydrogen, for example, silicon nitride formed by a CVD method can be used. Here, when hydrogen diffuses into a semiconductor element having an oxide semiconductor such as the transistor 500, the characteristics of the semiconductor element may deteriorate. Therefore, it is preferable to use a film that suppresses the diffusion of hydrogen between the transistor 500 and the transistor 300. Specifically, the film that suppresses the diffusion of hydrogen is a film that releases a small amount of hydrogen.

[0447] The amount of desorption of hydrogen can be analyzed, for example, by using thermal desorption spectrometry (TDS). For example, the amount of desorption of hydrogen from the insulator 324 is calculated as 10×10 per area of ​​the insulator 324 when the surface temperature of the film is in the range of 50° C. to 500° C. in the TDS analysis. 15 atoms / cm 2 Less than or equal to 5×10 15 atoms / cm 2 The following is acceptable.

[0448] The insulator 326 preferably has a lower dielectric constant than the insulator 324. For example, the relative dielectric constant of the insulator 326 is preferably less than 4, and more preferably less than 3. For example, the relative dielectric constant of the insulator 326 is preferably 0.7 times or less, and more preferably 0.6 times or less, the relative dielectric constant of the insulator 324. By using a material with a low dielectric constant as the interlayer film, the parasitic capacitance generated between wirings can be reduced.

[0449] A conductor 328, a conductor 330, and the like that connect to the capacitor 600 or the transistor 500 are embedded in the insulator 320, the insulator 322, the insulator 324, and the insulator 326. The conductor 328 and the conductor 330 function as a plug or a wiring. In addition, a plurality of conductors that function as a plug or a wiring may be collectively given the same reference symbol. In this specification and the like, a wiring and a plug connected to the wiring may be integrated. That is, there are cases where a part of a conductor functions as a wiring and a case where a part of a conductor functions as a plug.

[0450] As the material for each plug and wiring (conductor 328, conductor 330, etc.), a conductive material such as a metal material, an alloy material, a metal nitride material, or a metal oxide material can be used in a single layer or a laminated layer. It is preferable to use a high melting point material such as tungsten or molybdenum that has both heat resistance and conductivity, and it is preferable to use tungsten. Alternatively, it is preferable to form the wiring from a low resistance conductive material such as aluminum or copper. By using a low resistance conductive material, it is possible to reduce the wiring resistance.

[0451] A wiring layer may be provided over the insulator 326 and the conductor 330. For example, in FIG. 19 , an insulator 350, an insulator 352, and an insulator 354 are stacked in this order. A conductor 356 is formed in the insulator 350, the insulator 352, and the insulator 354. The conductor 356 functions as a plug or a wiring connected to the transistor 300. Note that the conductor 356 can be provided using a material similar to that of the conductor 328 and the conductor 330.

[0452] Note that, for example, the insulator 350 is preferably an insulator having a barrier property against hydrogen, similar to the insulator 324. The conductor 356 preferably includes a conductor having a barrier property against hydrogen. In particular, a conductor having a barrier property against hydrogen is formed in an opening of the insulator 350 having a barrier property against hydrogen. With this structure, the transistor 300 and the transistor 500 can be separated by a barrier layer, and diffusion of hydrogen from the transistor 300 to the transistor 500 can be suppressed.

[0453] Note that, for example, tantalum nitride or the like may be used as the conductor having a barrier property against hydrogen. By stacking tantalum nitride and tungsten having high conductivity, it is possible to suppress diffusion of hydrogen from the transistor 300 while maintaining the conductivity of the wiring. In this case, a structure in which the tantalum nitride layer having a barrier property against hydrogen is in contact with the insulator 350 having a barrier property against hydrogen is preferable.

[0454] A wiring layer may be provided over the insulator 354 and the conductor 356. For example, in FIG. 19, an insulator 360, an insulator 362, and an insulator 364 are stacked in this order. A conductor 366 is formed in the insulator 360, the insulator 362, and the insulator 364. The conductor 366 functions as a plug or a wiring. The conductor 366 can be provided using a material similar to that of the conductor 328 and the conductor 330.

[0455] Note that, for example, the insulator 360 is preferably an insulator having a barrier property against hydrogen, similar to the insulator 324. The conductor 366 preferably includes a conductor having a barrier property against hydrogen. In particular, a conductor having a barrier property against hydrogen is formed in an opening of the insulator 360 having a barrier property against hydrogen. With this structure, the transistor 300 and the transistor 500 can be separated by a barrier layer, and diffusion of hydrogen from the transistor 300 to the transistor 500 can be suppressed.

[0456] A wiring layer may be provided over the insulator 364 and the conductor 366. For example, in FIG. 19, an insulator 370, an insulator 372, and an insulator 374 are stacked in this order. A conductor 376 is formed in the insulator 370, the insulator 372, and the insulator 374. The conductor 376 functions as a plug or a wiring. Note that the conductor 376 can be provided using a material similar to that of the conductor 328 and the conductor 330.

[0457] Note that, for example, the insulator 370 is preferably an insulator having a barrier property against hydrogen, similar to the insulator 324. The conductor 376 preferably includes a conductor having a barrier property against hydrogen. In particular, a conductor having a barrier property against hydrogen is formed in an opening of the insulator 370 having a barrier property against hydrogen. With this structure, the transistor 300 and the transistor 500 can be separated by a barrier layer, and diffusion of hydrogen from the transistor 300 to the transistor 500 can be suppressed.

[0458] A wiring layer may be provided over the insulator 374 and the conductor 376. For example, in FIG. 19, an insulator 380, an insulator 382, ​​and an insulator 384 are stacked in this order. A conductor 386 is formed in the insulator 380, the insulator 382, ​​and the insulator 384. The conductor 386 functions as a plug or a wiring. Note that the conductor 386 can be provided using a material similar to that of the conductor 328 and the conductor 330.

[0459] Note that, for example, the insulator 380 is preferably an insulator having a barrier property against hydrogen, similar to the insulator 324. The conductor 386 preferably includes a conductor having a barrier property against hydrogen. In particular, a conductor having a barrier property against hydrogen is formed in an opening of the insulator 380 having a barrier property against hydrogen. With this structure, the transistor 300 and the transistor 500 can be separated by a barrier layer, and diffusion of hydrogen from the transistor 300 to the transistor 500 can be suppressed.

[0460] In the above, a wiring layer including the conductor 356, a wiring layer including the conductor 366, a wiring layer including the conductor 376, and a wiring layer including the conductor 386 have been described, but the semiconductor device of this embodiment is not limited to this. There may be three or fewer wiring layers similar to the wiring layer including the conductor 356, or there may be five or more wiring layers similar to the wiring layer including the conductor 356.

[0461] An insulator 510, an insulator 512, an insulator 514, and an insulator 516 are stacked in this order over the insulator 384. Any of the insulator 510, the insulator 512, the insulator 514, and the insulator 516 is preferably made of a substance that has a barrier property against oxygen or hydrogen.

[0462] For example, the insulator 510 and the insulator 514 are preferably formed using a film having a barrier property that prevents hydrogen or impurities from diffusing from the substrate 311 or a region where the transistor 300 is provided to a region where the transistor 500 is provided. Therefore, a material similar to that of the insulator 324 can be used.

[0463] As an example of a film having a barrier property against hydrogen, silicon nitride formed by a CVD method can be used. Here, when hydrogen diffuses into a semiconductor element having an oxide semiconductor such as the transistor 500, the characteristics of the semiconductor element may deteriorate. Therefore, it is preferable to use a film that suppresses the diffusion of hydrogen between the transistor 500 and the transistor 300. Specifically, the film that suppresses the diffusion of hydrogen is a film that releases a small amount of hydrogen.

[0464] As a film having a barrier property against hydrogen, for example, the insulator 510 and the insulator 514 are preferably made of a metal oxide such as aluminum oxide, hafnium oxide, or tantalum oxide.

[0465] In particular, aluminum oxide has a high blocking effect of preventing the film from permeating both oxygen and impurities such as hydrogen and moisture, which are factors that cause fluctuations in the electrical characteristics of a transistor. Therefore, aluminum oxide can prevent impurities such as hydrogen and moisture from entering the transistor 500 during and after the transistor manufacturing process. In addition, aluminum oxide can suppress the release of oxygen from the oxide that constitutes the transistor 500. Therefore, aluminum oxide is suitable for use as a protective film for the transistor 500.

[0466] For example, the insulator 512 and the insulator 516 can be formed using a material similar to that of the insulator 320. By using a material with a relatively low dielectric constant for these insulators, parasitic capacitance between wirings can be reduced. For example, the insulator 512 and the insulator 516 can be formed using a silicon oxide film, a silicon oxynitride film, or the like.

[0467] A conductor 518, a conductor constituting the transistor 500 (for example, the conductor 503), and the like are embedded in the insulator 510, the insulator 512, the insulator 514, and the insulator 516. Note that the conductor 518 functions as a plug or a wiring connected to the capacitor 600 or the transistor 300. The conductor 518 can be formed using a material similar to that of the conductor 328 and the conductor 330.

[0468] In particular, the insulator 510 and the conductor 518 in the region in contact with the insulator 514 are preferably conductors having barrier properties against oxygen, hydrogen, and water. With this structure, the transistors 300 and 500 can be separated by a layer having barrier properties against oxygen, hydrogen, and water, and diffusion of hydrogen from the transistor 300 to the transistor 500 can be suppressed.

[0469] Above the insulator 516 is a transistor 500 .

[0470] As shown in Figures 21A and 21B, a transistor 500 has a conductor 503 arranged so as to be embedded in an insulator 514 and an insulator 516, an insulator 520 arranged on the insulator 516 and the conductor 503, an insulator 522 arranged on the insulator 520, an insulator 524 arranged on the insulator 522, an oxide 530a arranged on the insulator 524, an oxide 530b arranged on the oxide 530a, conductors 542a and 542b arranged apart from each other on the oxide 530b, an insulator 580 arranged on the conductors 542a and 542b and having an opening formed therein overlapping between the conductors 542a and 542b, an oxide 530c arranged on the bottom and side surfaces of the opening, an insulator 550 arranged on the formation surface of the oxide 530c, and a conductor 560 arranged on the formation surface of the insulator 550.

[0471] 21A and 21B, it is preferable that an insulator 544 is disposed between the oxide 530a, the oxide 530b, the conductor 542a, and the conductor 542b and the insulator 580. It is preferable that the conductor 560 has a conductor 560a provided inside the insulator 550 and a conductor 560b provided so as to be embedded inside the conductor 560a, as shown in FIGS. 21A and 21B. It is preferable that an insulator 574 is disposed on the insulator 580, the conductor 560, and the insulator 550, as shown in FIGS. 21A and 21B.

[0472] In the following, oxide 530a, oxide 530b, and oxide 530c may be collectively referred to as oxide 530.

[0473] Note that, in the transistor 500, a three-layer structure of the oxide 530a, the oxide 530b, and the oxide 530c is illustrated in the region where a channel is formed and in the vicinity thereof; however, one embodiment of the present invention is not limited thereto. For example, a single layer of the oxide 530b, a two-layer structure of the oxide 530b and the oxide 530a, a two-layer structure of the oxide 530b and the oxide 530c, or a stacked structure of four or more layers may be provided. In addition, in the transistor 500, the conductor 560 is illustrated as having a two-layer structure; however, one embodiment of the present invention is not limited thereto. For example, the conductor 560 may have a single-layer structure or a stacked structure of three or more layers. The transistor 500 illustrated in FIG. 19 and FIG. 21A is merely an example, and the present invention is not limited thereto, and an appropriate transistor may be used depending on the circuit configuration and the driving method.

[0474] Here, the conductor 560 functions as a gate electrode of the transistor, and the conductors 542a and 542b function as a source electrode and a drain electrode, respectively. As described above, the conductor 560 is formed so as to be embedded in the opening of the insulator 580 and in the region sandwiched between the conductors 542a and 542b. The arrangement of the conductors 560, 542a, and 542b is selected in a self-aligned manner with respect to the opening of the insulator 580. That is, in the transistor 500, the gate electrode can be arranged in a self-aligned manner between the source electrode and the drain electrode. Therefore, the conductor 560 can be formed without providing a margin for alignment, so that the area occupied by the transistor 500 can be reduced. This allows the semiconductor device to be miniaturized and highly integrated.

[0475] Furthermore, since the conductor 560 is formed in a self-aligned manner in the region between the conductor 542a and the conductor 542b, the conductor 560 does not have a region overlapping with the conductor 542a or the conductor 542b. This makes it possible to reduce the parasitic capacitance formed between the conductor 560 and the conductor 542a and between the conductor 560 and the conductor 542b. As a result, the switching speed of the transistor 500 can be improved, and high frequency characteristics can be achieved.

[0476] The conductor 560 may function as a first gate (also referred to as a top gate) electrode. The conductor 503 may function as a second gate (also referred to as a bottom gate) electrode. In this case, the threshold voltage of the transistor 500 can be controlled by changing the potential applied to the conductor 503 independently of the potential applied to the conductor 560. In particular, by applying a negative potential to the conductor 503, the threshold voltage of the transistor 500 can be made higher than 0 V, and the off-current can be reduced. Therefore, applying a negative potential to the conductor 503 can reduce the drain current when the potential applied to the conductor 560 is 0 V, compared to when a negative potential is not applied.

[0477] The conductor 503 is disposed so as to overlap the oxide 530 and the conductor 560. In this manner, when a potential is applied to the conductor 560 and the conductor 503, an electric field generated from the conductor 560 and an electric field generated from the conductor 503 are connected, and a channel formation region formed in the oxide 530 can be covered. In this specification and the like, a transistor structure in which a channel formation region is electrically surrounded by the electric fields of the first gate electrode and the second gate electrode is referred to as a surrounded channel (S-channel) structure.

[0478] The conductor 503 has a structure similar to that of the conductor 518, and the conductor 503a is formed in contact with the inner walls of the openings of the insulator 514 and the insulator 516, and the conductor 503b is formed further inward. Note that although the transistor 500 has a structure in which the conductor 503a and the conductor 503b are stacked, one embodiment of the present invention is not limited to this. For example, the conductor 503 may have a single layer structure or a stacked structure of three or more layers.

[0479] Here, the conductor 503a is preferably made of a conductive material having a function of suppressing the diffusion of impurities such as hydrogen atoms, hydrogen molecules, water molecules, and copper atoms (the impurities are less likely to permeate through). Alternatively, it is preferable to use a conductive material having a function of suppressing the diffusion of oxygen (e.g., at least one of oxygen atoms, oxygen molecules, etc.) (the oxygen is less likely to permeate through). Note that in this specification, the function of suppressing the diffusion of impurities or oxygen refers to the function of suppressing the diffusion of any one or all of the impurities or the oxygen.

[0480] For example, the conductor 503a has a function of suppressing the diffusion of oxygen, so that the conductor 503b can be prevented from being oxidized and the conductivity from decreasing.

[0481] In addition, when the conductor 503 also functions as a wiring, it is preferable that the conductor 503b is made of a highly conductive material mainly composed of tungsten, copper, or aluminum. In that case, the conductor 503a is not necessarily provided. Although the conductor 503b is illustrated as a single layer, it may have a laminated structure, for example, a laminate of titanium or titanium nitride and the above-mentioned conductive material.

[0482] The insulators 520, 522, and 524 function as a second gate insulating film.

[0483] Here, the insulator 524 in contact with the oxide 530 preferably contains more oxygen than the oxygen required for the stoichiometric composition. That is, an excess oxygen region is preferably formed in the insulator 524. By providing such an insulator containing excess oxygen in contact with the oxide 530, oxygen vacancies in the oxide 530 can be reduced and the reliability of the transistor 500 can be improved.

[0484] Specifically, it is preferable to use an oxide material from which some oxygen is released by heating as an insulator having an excess oxygen region. An oxide material from which oxygen is released by heating is an oxide material from which the amount of released oxygen converted into oxygen atoms is 1.0×10 18 atoms / cm 3 More than 1.0×10 19 atoms / cm 3 More preferably, 2.0×10 19 atoms / cm 3 or more than 3.0 x 10 20 atoms / cm 3 The oxide film is one having the above-mentioned properties. The surface temperature of the film during the TDS analysis is preferably in the range of 100° C. or higher and 700° C. or lower, or 100° C. or higher and 400° C. or lower.

[0485] In addition, the oxide 530 may be brought into contact with the insulator having the excess oxygen region and subjected to one or more of heat treatment, microwave treatment, and RF treatment. By carrying out such treatment, water or hydrogen in the oxide 530 can be removed. For example, a reaction occurs in the oxide 530 that breaks the bond of VoH, in other words, "V O H→V O A reaction of "H + H" occurs, resulting in dehydrogenation. Some of the hydrogen generated at this time may combine with oxygen to form HO and be removed from the oxide 530 or the insulator near the oxide 530. Some of the hydrogen may also be diffused or captured (also referred to as gettering) in the conductor 542a and the conductor 542b.

[0486] In addition, the microwave treatment is preferably performed using, for example, a device having a power source that generates high-density plasma or a device having a power source that applies RF to the substrate side. For example, high-density oxygen radicals can be generated by using a gas containing oxygen and high-density plasma, and the oxygen radicals generated by the high-density plasma can be efficiently introduced into the oxide 530 or an insulator near the oxide 530 by applying RF to the substrate side. In addition, the pressure of the microwave treatment may be set to 133 Pa or more, preferably 200 Pa or more, and more preferably 400 Pa or more. In addition, for example, oxygen and argon are used as gases to be introduced into the microwave treatment device, and the oxygen flow rate ratio (O2 / (O2+Ar)) is 50% or less, preferably 10% or more and 30% or less.

[0487] In addition, in a manufacturing process of the transistor 500, it is preferable to perform heat treatment while the surface of the oxide 530 is exposed. The heat treatment may be performed, for example, at a temperature of 100° C. or higher and 450° C. or lower, more preferably 350° C. or higher and 400° C. or lower. Note that the heat treatment is performed in an atmosphere of nitrogen gas or an inert gas, or an atmosphere containing an oxidizing gas at 10 ppm or higher, 1% or higher, or 10% or higher. For example, the heat treatment is preferably performed in an oxygen atmosphere. In this way, oxygen is supplied to the oxide 530 to reduce oxygen deficiencies (V O) can be reduced. The heat treatment may be performed under reduced pressure. Alternatively, the heat treatment may be performed in an atmosphere containing 10 ppm or more, 1% or more, or 10% or more of an oxidizing gas in order to compensate for the desorbed oxygen after the heat treatment in a nitrogen gas or inert gas atmosphere. Alternatively, the heat treatment may be performed in an atmosphere containing 10 ppm or more, 1% or more, or 10% or more of an oxidizing gas in order to compensate for the desorbed oxygen. Alternatively, the heat treatment may be performed in an atmosphere containing 10 ppm or more, 1% or more, or 10% or more of an oxidizing gas in succession to a heat treatment in a nitrogen gas or inert gas atmosphere.

[0488] In addition, by subjecting the oxide 530 to an oxygen supplying treatment, the oxygen vacancies in the oxide 530 are repaired by the supplied oxygen. In other words, O In addition, the reaction of the hydrogen remaining in the oxide 530 with the supplied oxygen can be removed as H2O (dehydration). As a result, the hydrogen remaining in the oxide 530 recombines with the oxygen vacancies to form V O The formation of H can be suppressed.

[0489] When the insulator 524 has an excess oxygen region, it is preferable that the insulator 522 has a function of suppressing the diffusion of oxygen (eg, oxygen atoms, oxygen molecules, etc.) (the oxygen is less likely to permeate).

[0490] The insulator 522 preferably has a function of suppressing diffusion of oxygen and impurities, so that oxygen contained in the oxide 530 does not diffuse toward the insulator 520. Furthermore, reaction of the conductor 503 with the insulator 524 or oxygen contained in the oxide 530 can be suppressed.

[0491] The insulator 522 is preferably a single layer or a multilayer insulator containing a so-called high-k material, such as aluminum oxide, hafnium oxide, oxide containing aluminum and hafnium (hafnium aluminate), tantalum oxide, zirconium oxide, lead zirconate titanate (PZT), strontium titanate (SrTiO3), or (Ba,Sr)TiO3 (BST). As transistors become smaller and more highly integrated, problems such as leakage current may occur due to the thinning of the gate insulating film. By using a high-k material for the insulator that functions as the gate insulating film, it is possible to reduce the gate potential during transistor operation while maintaining the physical film thickness.

[0492] In particular, it is preferable to use an insulator containing an oxide of one or both of aluminum and hafnium, which are insulating materials having a function of suppressing the diffusion of impurities and oxygen (the oxygen is unlikely to permeate through them). As an insulator containing an oxide of one or both of aluminum and hafnium, it is preferable to use aluminum oxide, hafnium oxide, an oxide containing aluminum and hafnium (hafnium aluminate), or the like. When the insulator 522 is formed using such a material, the insulator 522 functions as a layer that suppresses release of oxygen from the oxide 530 and the intrusion of impurities such as hydrogen into the oxide 530 from the periphery of the transistor 500.

[0493] Alternatively, for example, aluminum oxide, bismuth oxide, germanium oxide, niobium oxide, silicon oxide, titanium oxide, tungsten oxide, yttrium oxide, or zirconium oxide may be added to these insulators. Alternatively, these insulators may be nitrided. Silicon oxide, silicon oxynitride, or silicon nitride may be stacked on the above insulators.

[0494] In addition, it is preferable that the insulator 520 is thermally stable. For example, silicon oxide and silicon oxynitride are preferable because they are thermally stable. In addition, by combining a high-k insulator with silicon oxide or silicon oxynitride, it is possible to obtain the insulator 520 having a thermally stable laminated structure with a high relative dielectric constant.

[0495] 21A and 21B, the second gate insulating film has a three-layer stack structure including the insulators 520, 522, and 524. However, the second gate insulating film may have a single layer, two layers, or four or more layers. In this case, the second gate insulating film is not limited to a stack structure made of the same material, and may have a stack structure made of different materials.

[0496] In the transistor 500, it is preferable to use a metal oxide functioning as an oxide semiconductor for the oxide 530 including the channel formation region. For example, a metal oxide such as In-M-Zn oxide (wherein the element M is one or more selected from aluminum, gallium, yttrium, copper, vanadium, beryllium, boron, titanium, iron, nickel, germanium, zirconium, molybdenum, lanthanum, cerium, neodymium, hafnium, tantalum, tungsten, magnesium, etc.) may be used as the oxide 530. In particular, it is preferable that the In-M-Zn oxide applicable to the oxide 530 is CAAC-OS (C-Axis Aligned Crystalline Oxide Semiconductor) or CAC-OS (Cloud-Aligned Composite Oxide Semiconductor). In addition, In-Ga oxide, In-Zn oxide, In oxide, etc. may be used as the oxide 530.

[0497] In addition, it is preferable to use a metal oxide with a low carrier concentration for the transistor 500. In order to reduce the carrier concentration of the metal oxide, the impurity concentration in the metal oxide may be reduced to reduce the defect state density. In this specification and the like, a low impurity concentration and a low defect state density are referred to as high purity intrinsic or substantially high purity intrinsic. Note that examples of impurities in the metal oxide include hydrogen, nitrogen, alkali metals, alkaline earth metals, iron, nickel, silicon, and the like.

[0498] In particular, hydrogen contained in the metal oxide reacts with oxygen that bonds with the metal atom to form water, which may cause oxygen vacancies in the metal oxide. In addition, when hydrogen enters an oxygen vacancy in the oxide 530, the oxygen vacancy and hydrogen combine to form V O May form H. V O H functions as a donor and may generate electrons as carriers. In addition, some of the hydrogen may bond with oxygen that bonds with metal atoms to generate electrons as carriers. Therefore, a transistor using a metal oxide containing a large amount of hydrogen is likely to have normally-on characteristics. In addition, hydrogen in a metal oxide is easily mobile due to stress such as heat or an electric field, so if a metal oxide contains a large amount of hydrogen, the reliability of the transistor may be deteriorated. In one embodiment of the present invention, V in the oxide 530 O It is preferable to reduce H as much as possible and make it high-purity intrinsic or substantially high-purity intrinsic. O In order to obtain a metal oxide with a sufficiently reduced amount of H, it is important to remove impurities such as water and hydrogen from the metal oxide (sometimes referred to as dehydration or dehydrogenation treatment) and to supply oxygen to the metal oxide to compensate for the oxygen deficiency (sometimes referred to as oxygen addition treatment). O By using a metal oxide in which impurities such as H are sufficiently reduced for the channel formation region of a transistor, stable electrical characteristics can be obtained.

[0499] A defect in which hydrogen has entered an oxygen vacancy can function as a donor for a metal oxide. However, it is difficult to quantitatively evaluate the defect. Therefore, in a metal oxide, the carrier concentration may be used instead of the donor concentration. Therefore, in this specification, the carrier concentration assuming a state in which no electric field is applied may be used as a parameter of the metal oxide, instead of the donor concentration. In other words, the "carrier concentration" described in this specification may be rephrased as the "donor concentration".

[0500] Therefore, when a metal oxide is used for the oxide 530, it is preferable that hydrogen in the metal oxide is reduced as much as possible. Specifically, in the metal oxide, the hydrogen concentration obtained by secondary ion mass spectrometry (SIMS) is 1×10 20 atoms / cm 3 Less than 1 x 10 19 atoms / cm 3 less than 5×10 18 atoms / cm 3 less than 1×10 18 atoms / cm 3 By using a metal oxide in which impurities such as hydrogen are sufficiently reduced for a channel formation region of a transistor, stable electrical characteristics can be obtained.

[0501] In addition, when a metal oxide is used for the oxide 530, the metal oxide is a semiconductor that has a wide band gap and is intrinsic (also called I-type) or substantially intrinsic, and the carrier concentration of the metal oxide in the channel formation region is 1×10 18 cm -3 Preferably less than 1×10 17 cm -3 More preferably, it is less than 1×10 16 cm -3 More preferably, it is less than 1×10 13 cm -3 More preferably, it is less than 1×10 12 cm -3 The lower limit of the carrier concentration of the metal oxide in the channel formation region is not particularly limited, but is preferably, for example, 1×10 -9 cm -3 It can be said that:

[0502] Furthermore, when a metal oxide is used for the oxide 530, when the conductors 542a and 542b come into contact with the oxide 530, oxygen in the oxide 530 may diffuse to the conductors 542a and 542b, and the conductors 542a and 542b may be oxidized. When the conductors 542a and 542b are oxidized, the conductivity of the conductors 542a and 542b is likely to decrease. The diffusion of oxygen in the oxide 530 to the conductors 542a and 542b can be expressed in other words as the conductors 542a and 542b absorbing the oxygen in the oxide 530.

[0503] Furthermore, oxygen in the oxide 530 may diffuse into the conductor 542a and the conductor 542b, forming a foreign layer between the conductor 542a and the oxide 530b and between the conductor 542b and the oxide 530b. Since the foreign layer contains more oxygen than the conductor 542a and the conductor 542b, the foreign layer is presumed to have insulating properties. In this case, the three-layer structure of the conductor 542a or the conductor 542b, the foreign layer, and the oxide 530b can be regarded as a three-layer structure made of a metal-insulator-semiconductor, and may be called a MIS (Metal-Insulator-Semiconductor) structure or a diode junction structure mainly based on the MIS structure.

[0504] Note that the above-mentioned different layer is not limited to being formed between the conductor 542a and the conductor 542b and the oxide 530b. For example, the different layer may be formed between the conductor 542a and the conductor 542b and the oxide 530c, between the conductor 542a and the conductor 542b and the oxide 530b, or between the conductor 542a and the conductor 542b and the oxide 530c.

[0505] The metal oxide that functions as a channel formation region in the oxide 530 preferably has a band gap of 2 eV or more, preferably 2.5 eV or more. In this manner, by using a metal oxide with a wide band gap, the off-state current of the transistor can be reduced.

[0506] The oxide 530 has the oxide 530a below the oxide 530b, so that the diffusion of impurities from structures formed below the oxide 530a to the oxide 530b can be suppressed. Also, the oxide 530 has the oxide 530c on the oxide 530b, so that the diffusion of impurities from structures formed above the oxide 530c to the oxide 530b can be suppressed.

[0507] The oxide 530 preferably has a laminated structure of a plurality of oxide layers having different atomic ratios of metal atoms. Specifically, in the metal oxide used for the oxide 530a, the atomic ratio of element M among the constituent elements is preferably larger than the atomic ratio of element M among the constituent elements in the metal oxide used for the oxide 530b. In addition, in the metal oxide used for the oxide 530a, the atomic ratio of element M to In is preferably larger than the atomic ratio of element M to In in the metal oxide used for the oxide 530b. In addition, in the metal oxide used for the oxide 530b, the atomic ratio of In to element M is preferably larger than the atomic ratio of In to element M in the metal oxide used for the oxide 530a. In addition, the oxide 530c can use a metal oxide that can be used for the oxide 530a or the oxide 530b.

[0508] Specifically, the oxide 530a may be a metal oxide having an atomic ratio of In, Ga, and Zn of In:Ga:Zn=1:3:4 or 1:1:0.5. The oxide 530b may be a metal oxide having an atomic ratio of In, Ga, and Zn of In:Ga:Zn=4:2:3 or 1:1:1. The oxide 530c may be a metal oxide having an atomic ratio of In, Ga, and Zn of In:Ga:Zn=1:3:4 and an atomic ratio of Ga and Zn of Ga:Zn=2:1 or Ga:Zn=2:5. Specific examples of the oxide 530c having a layered structure include layered structures in which the atomic ratios of In, Ga, and Zn are In:Ga:Zn=4:2:3 and In:Ga:Zn=1:3:4, layered structures in which the atomic ratios of Ga and Zn are Ga:Zn=2:1 and In:Ga:Zn=4:2:3, layered structures in which the atomic ratios of Ga and Zn are Ga:Zn=2:5 and In:Ga:Zn=4:2:3, and layered structures in which gallium oxide and the atomic ratio of In, Ga, and Zn are In:Ga:Zn=4:2:3.

[0509] Furthermore, for example, when the atomic ratio of In to element M in the metal oxide used for oxide 530a is smaller than the atomic ratio of In to element M in the metal oxide used for oxide 530b, an In-Ga-Zn oxide having an atomic ratio of In, Ga, and Zn of In:Ga:Zn=5:1:6 or thereabouts, In:Ga:Zn=5:1:3 or thereabouts, In:Ga:Zn=10:1:3 or thereabouts, or the like, can be used as oxide 530b.

[0510] In addition to the compositions described above, the oxide 530b may be a metal oxide having a composition of In:Zn=2:1, a composition of In:Zn=5:1, a composition of In:Zn=10:1, or a composition close to any one of these.

[0511] It is preferable to combine these oxides 530a, 530b, and 530c while satisfying the above-mentioned atomic ratio relationship. For example, it is preferable to use metal oxides having a composition of In:Ga:Zn=1:3:4 for the oxides 530a and 530c, and metal oxides having a composition of In:Ga:Zn=4:2:3 to 4.1 for the oxides 530b. The above composition indicates the atomic ratio in the oxide formed on the substrate, or the atomic ratio in the sputtering target. In addition, it is preferable to increase the ratio of In in the composition of the oxide 530b, since this can increase the on-current or field effect mobility of the transistor.

[0512] In addition, the energy of the conduction band minimum of the oxide 530a and the oxide 530c is preferably higher than that of the oxide 530b. In other words, the electron affinity of the oxide 530a and the oxide 530c is preferably smaller than that of the oxide 530b.

[0513] Here, the energy level of the conduction band minimum changes gradually at the junction of the oxide 530a, the oxide 530b, and the oxide 530c. In other words, it can be said that the energy level of the conduction band minimum at the junction of the oxide 530a, the oxide 530b, and the oxide 530c changes continuously or is a continuous junction. To achieve this, it is preferable to reduce the defect level density of the mixed layer formed at the interface between the oxide 530a and the oxide 530b and the interface between the oxide 530b and the oxide 530c.

[0514] Specifically, the oxides 530a and 530b, and the oxides 530b and 530c have a common element other than oxygen (as a main component), so that a mixed layer with a low density of defect states can be formed. For example, when the oxide 530b is an In-Ga-Zn oxide, the oxides 530a and 530c may be made of an In-Ga-Zn oxide, a Ga-Zn oxide, or a gallium oxide.

[0515] At this time, the main carrier path is the oxide 530b. By configuring the oxide 530a and the oxide 530c as described above, the defect state density at the interface between the oxide 530a and the oxide 530b and at the interface between the oxide 530b and the oxide 530c can be reduced. Therefore, the effect of interface scattering on carrier conduction is reduced, and the transistor 500 can obtain a high on-state current.

[0516] Conductors 542a and 542b functioning as a source electrode and a drain electrode are provided on the oxide 530b. As the conductors 542a and 542b, it is preferable to use a metal element selected from aluminum, chromium, copper, silver, gold, platinum, tantalum, nickel, titanium, molybdenum, tungsten, hafnium, vanadium, niobium, manganese, magnesium, zirconium, beryllium, indium, ruthenium, iridium, strontium, and lanthanum, an alloy containing the above-mentioned metal element as a component, or an alloy combining the above-mentioned metal elements. For example, it is preferable to use tantalum nitride, titanium nitride, tungsten, a nitride containing titanium and aluminum, a nitride containing tantalum and aluminum, ruthenium oxide, ruthenium nitride, an oxide containing strontium and ruthenium, an oxide containing lanthanum and nickel, or the like. In addition, tantalum nitride, titanium nitride, nitrides containing titanium and aluminum, nitrides containing tantalum and aluminum, ruthenium oxide, ruthenium nitride, oxides containing strontium and ruthenium, and oxides containing lanthanum and nickel are conductive materials that are difficult to oxidize, or materials that maintain conductivity even when oxygen is absorbed, and are therefore preferable.Furthermore, metal nitride films such as tantalum nitride are preferable because they have barrier properties against hydrogen or oxygen.

[0517] 21A and 21B, the conductor 542a and the conductor 542b are shown as a single-layer structure, but may be a laminated structure of two or more layers. For example, a tantalum nitride film and a tungsten film may be laminated. A titanium film and an aluminum film may be laminated. Alternatively, a two-layer structure in which an aluminum film is laminated on a tungsten film, a two-layer structure in which a copper film is laminated on a copper-magnesium-aluminum alloy film, a two-layer structure in which a copper film is laminated on a titanium film, or a two-layer structure in which a copper film is laminated on a tungsten film may be used.

[0518] In addition, there are three-layer structures in which a titanium film or titanium nitride film is laminated on the titanium film or titanium nitride film, an aluminum film or copper film is laminated on the titanium film or titanium nitride film, and a titanium film or titanium nitride film is further formed thereon, and a three-layer structure in which a molybdenum film or molybdenum nitride film is laminated on the molybdenum film or molybdenum nitride film, an aluminum film or copper film is laminated on the molybdenum film or molybdenum nitride film, and a molybdenum film or molybdenum nitride film is further formed thereon, etc. Note that a transparent conductive material containing indium oxide, tin oxide, or zinc oxide may be used.

[0519] 21A, regions 543a and 543b may be formed as low-resistance regions at and near the interface of the oxide 530 with the conductor 542a (conductor 542b). In this case, the region 543a functions as one of the source region and the drain region, and the region 543b functions as the other of the source region and the drain region. A channel formation region is formed in a region sandwiched between the regions 543a and 543b.

[0520] By providing the conductor 542a (conductor 542b) so as to be in contact with the oxide 530, the oxygen concentration in the region 543a (region 543b) may be reduced. Also, a metal compound layer containing a metal contained in the conductor 542a (conductor 542b) and a component of the oxide 530 may be formed in the region 543a (region 543b). In such a case, the carrier concentration in the region 543a (region 543b) increases, and the region 543a (region 543b) becomes a low-resistance region.

[0521] The insulator 544 is provided to cover the conductor 542a and the conductor 542b and suppresses oxidation of the conductor 542a and the conductor 542b. In this case, the insulator 544 may be provided to cover the side surface of the oxide 530 and to be in contact with the insulator 524.

[0522] The insulator 544 can be a metal oxide containing one or more elements selected from hafnium, aluminum, gallium, yttrium, zirconium, tungsten, titanium, tantalum, nickel, germanium, neodymium, lanthanum, magnesium, etc. Alternatively, the insulator 544 can be silicon nitride oxide, silicon nitride, or the like.

[0523] In particular, it is preferable to use, as the insulator 544, an insulator containing an oxide of either or both of aluminum and hafnium, such as aluminum oxide, hafnium oxide, or an oxide containing aluminum and hafnium (hafnium aluminate). In particular, hafnium aluminate has higher heat resistance than a hafnium oxide film. Therefore, it is preferable because it is less likely to crystallize in a heat treatment in a later step. Note that the insulator 544 is not an essential component when the conductor 542a and the conductor 542b are made of a material having oxidation resistance or when the conductivity does not decrease significantly even when oxygen is absorbed. It may be designed appropriately depending on the desired transistor characteristics.

[0524] The insulator 544 can prevent impurities such as water and hydrogen contained in the insulator 580 from diffusing to the oxide 530b through the oxide 530c and the insulator 550. The insulator 580 can also prevent the conductor 560 from being oxidized by excess oxygen contained in the insulator 580.

[0525] The insulator 550 functions as a first gate insulating film. The insulator 550 is preferably disposed in contact with the inside (top and side surfaces) of the oxide 530c. The insulator 550 is preferably formed using an insulator that contains excess oxygen and releases oxygen by heating, similar to the insulator 524 described above.

[0526] Specifically, silicon oxide having excess oxygen, silicon oxynitride, silicon nitride oxide, silicon nitride, silicon oxide to which fluorine is added, silicon oxide to which carbon is added, silicon oxide to which carbon and nitrogen are added, and silicon oxide having vacancies can be used. In particular, silicon oxide and silicon oxynitride are preferable because they are stable against heat.

[0527] By providing an insulator that releases oxygen when heated as the insulator 550 in contact with the top surface of the oxide 530c, oxygen can be effectively supplied from the insulator 550 to the channel formation region of the oxide 530b through the oxide 530c. Similarly to the insulator 524, the concentration of impurities such as water or hydrogen in the insulator 550 is preferably reduced. The thickness of the insulator 550 is preferably 1 nm or more and 20 nm or less.

[0528] Furthermore, in order to efficiently supply excess oxygen contained in the insulator 550 to the oxide 530, a metal oxide may be provided between the insulator 550 and the conductor 560. The metal oxide preferably suppresses oxygen diffusion from the insulator 550 to the conductor 560. By providing a metal oxide that suppresses oxygen diffusion, the diffusion of excess oxygen from the insulator 550 to the conductor 560 is suppressed. That is, a decrease in the amount of excess oxygen supplied to the oxide 530 can be suppressed. Furthermore, oxidation of the conductor 560 due to the excess oxygen can be suppressed. As the metal oxide, a material that can be used for the insulator 544 may be used.

[0529] The insulator 550 may have a laminated structure, similar to the second gate insulating film. As transistors become smaller and more highly integrated, problems such as leakage current may occur due to the thinning of the gate insulating film. Therefore, by making the insulator that functions as the gate insulating film a laminated structure of a high-k material and a thermally stable material, it is possible to reduce the gate potential during transistor operation while maintaining the physical film thickness. In addition, a laminated structure that is thermally stable and has a high relative dielectric constant can be obtained.

[0530] Although the conductor 560 functioning as the first gate electrode is shown as having a two-layer structure in FIGS. 21A and 21B, it may have a single-layer structure or a laminated structure of three or more layers.

[0531] The conductor 560a is preferably made of a conductive material having a function of suppressing the diffusion of impurities such as hydrogen atoms, hydrogen molecules, water molecules, nitrogen atoms, nitrogen molecules, nitrogen oxide molecules (N2O, NO, NO2, etc.), and copper atoms. Alternatively, it is preferably made of a conductive material having a function of suppressing the diffusion of oxygen (for example, at least one of oxygen atoms, oxygen molecules, etc.). When the conductor 560a has a function of suppressing the diffusion of oxygen, it is possible to suppress the conductor 560b from being oxidized by the oxygen contained in the insulator 550 and the conductivity from decreasing. As a conductive material having a function of suppressing the diffusion of oxygen, for example, tantalum, tantalum nitride, ruthenium, or ruthenium oxide is preferably used. In addition, an oxide semiconductor applicable to the oxide 530 can be used as the conductor 560a. In that case, the conductor 560b can be formed by a sputtering method to reduce the electrical resistance value of the conductor 560a and make it a conductor. This can be called an OC (Oxide Conductor) electrode.

[0532] The conductor 560b is preferably made of a conductive material containing tungsten, copper, or aluminum as a main component. Since the conductor 560b also functions as wiring, it is preferable to use a conductor having high conductivity. For example, a conductive material containing tungsten, copper, or aluminum as a main component can be used. The conductor 560b may have a layered structure, for example, a layered structure of titanium or titanium nitride and the above conductive material.

[0533] The insulator 580 is provided on the conductor 542a and the conductor 542b via the insulator 544. The insulator 580 preferably has an excess oxygen region. For example, the insulator 580 preferably has silicon oxide, silicon oxynitride, silicon nitride oxide, silicon nitride, silicon oxide to which fluorine has been added, silicon oxide to which carbon has been added, silicon oxide to which carbon and nitrogen have been added, silicon oxide having voids, or a resin. In particular, silicon oxide and silicon oxynitride are preferable because they are thermally stable. In particular, silicon oxide and silicon oxide having voids are preferable because they allow for easy formation of an excess oxygen region in a later step.

[0534] The insulator 580 preferably has an excess oxygen region. By providing the insulator 580, which releases oxygen when heated, in contact with the oxide 530c, oxygen in the insulator 580 can be efficiently supplied to the oxide 530 through the oxide 530c. Note that the concentration of impurities such as water or hydrogen in the insulator 580 is preferably reduced.

[0535] The opening of the insulator 580 is formed to overlap the region between the conductor 542a and the conductor 542b. As a result, the conductor 560 is formed so as to be embedded in the opening of the insulator 580 and the region sandwiched between the conductor 542a and the conductor 542b.

[0536] In miniaturizing a semiconductor device, it is necessary to shorten the gate length, but it is also necessary to prevent the conductivity of the conductor 560 from decreasing. If the film thickness of the conductor 560 is increased for that purpose, the conductor 560 may have a shape with a high aspect ratio. In this embodiment, the conductor 560 is provided so as to be embedded in the opening of the insulator 580, so that even if the conductor 560 has a shape with a high aspect ratio, the conductor 560 can be formed without collapsing during the process.

[0537] The insulator 574 is preferably provided in contact with a top surface of the insulator 580, a top surface of the conductor 560, and a top surface of the insulator 550. By forming the insulator 574 by a sputtering method, excess oxygen regions can be provided in the insulator 550 and the insulator 580. This allows oxygen to be supplied from the excess oxygen regions into the oxide 530.

[0538] For example, the insulator 574 can be a metal oxide containing one or more elements selected from hafnium, aluminum, gallium, yttrium, zirconium, tungsten, titanium, tantalum, nickel, germanium, magnesium, and the like.

[0539] In particular, aluminum oxide has high barrier properties and can suppress the diffusion of hydrogen and nitrogen even in a thin film of 0.5 nm to 3.0 nm. Therefore, aluminum oxide formed by sputtering can function as a barrier film against impurities such as hydrogen as well as an oxygen source.

[0540] An insulator 581 functioning as an interlayer film is preferably provided over the insulator 574. Like the insulator 524, the insulator 581 preferably has a reduced concentration of impurities such as water or hydrogen.

[0541] Furthermore, conductor 540a and conductor 540b are arranged in openings formed in insulator 581, insulator 574, insulator 580, and insulator 544. Conductor 540a and conductor 540b are provided facing each other with conductor 560 interposed therebetween. Conductor 540a and conductor 540b have the same structure as conductor 546 and conductor 548, which will be described later.

[0542] An insulator 582 is provided over the insulator 581. The insulator 582 is preferably made of a substance that has a barrier property against oxygen and hydrogen. Therefore, the insulator 582 can be made of a material similar to that of the insulator 514. For example, the insulator 582 is preferably made of a metal oxide such as aluminum oxide, hafnium oxide, or tantalum oxide.

[0543] In particular, aluminum oxide has a high blocking effect of preventing the film from permeating both oxygen and impurities such as hydrogen and moisture, which are factors that cause fluctuations in the electrical characteristics of a transistor. Therefore, aluminum oxide can prevent impurities such as hydrogen and moisture from entering the transistor 500 during and after the transistor manufacturing process. In addition, aluminum oxide can suppress the release of oxygen from the oxide that constitutes the transistor 500. Therefore, aluminum oxide is suitable for use as a protective film for the transistor 500.

[0544] An insulator 586 is provided over the insulator 582. The insulator 586 can be formed using a material similar to that of the insulator 320. By using a material with a relatively low dielectric constant for these insulators, parasitic capacitance between wirings can be reduced. For example, a silicon oxide film, a silicon oxynitride film, or the like can be used as the insulator 586.

[0545] Furthermore, conductors 546, conductors 548, etc. are embedded in insulators 520, 522, 524, 544, 580, 574, 581, 582, and 586.

[0546] The conductor 546 and the conductor 548 function as a plug or a wiring connected to the capacitor 600, the transistor 500, or the transistor 300. The conductor 546 and the conductor 548 can be formed using a material similar to that of the conductor 328 and the conductor 330.

[0547] After the transistor 500 is formed, an opening may be formed to surround the transistor 500, and an insulator having a high barrier property against hydrogen or water may be formed to cover the opening. By surrounding the transistor 500 with the insulator having a high barrier property, it is possible to prevent moisture and hydrogen from entering from the outside. Alternatively, a plurality of transistors 500 may be collectively surrounded by an insulator having a high barrier property against hydrogen or water. When an opening is formed to surround the transistor 500, for example, it is preferable to form an opening that reaches the insulator 514 or the insulator 522 and form the insulator having a high barrier property in contact with the insulator 514 or the insulator 522, since this can serve as part of the manufacturing process of the transistor 500. As the insulator having a high barrier property against hydrogen or water, for example, a material similar to that of the insulator 522 may be used.

[0548] Next, a capacitor 600 is provided above the transistor 500. The capacitor 600 includes a conductor 610, a conductor 620, and an insulator 630.

[0549] A conductor 612 may be provided over the conductor 546 and the conductor 548. The conductor 612 functions as a plug or a wiring connected to the transistor 500. The conductor 610 functions as an electrode of the capacitor 600. Note that the conductor 612 and the conductor 610 can be formed at the same time.

[0550] A metal film containing an element selected from molybdenum, titanium, tantalum, tungsten, aluminum, copper, chromium, neodymium, and scandium, or a metal nitride film containing the above-mentioned element (tantalum nitride film, titanium nitride film, molybdenum nitride film, tungsten nitride film), or the like can be used for the conductor 612 and the conductor 610. Alternatively, a conductive material such as indium tin oxide, indium oxide containing tungsten oxide, indium zinc oxide containing tungsten oxide, indium oxide containing titanium oxide, indium tin oxide containing titanium oxide, indium zinc oxide, or indium tin oxide to which silicon oxide is added can also be used.

[0551] 19, the conductor 612 and the conductor 610 have a single-layer structure, but are not limited to this structure and may have a stacked structure of two or more layers. For example, a conductor having barrier properties and a conductor having high adhesion to the conductor having high conductivity may be formed between a conductor having barrier properties and a conductor having high conductivity.

[0552] The conductor 620 is provided so as to overlap with the conductor 610 with the insulator 630 interposed therebetween. Note that the conductor 620 can be made of a conductive material such as a metal material, an alloy material, or a metal oxide material. It is preferable to use a high melting point material such as tungsten or molybdenum that has both heat resistance and conductivity, and it is particularly preferable to use tungsten. When the conductor 620 is formed simultaneously with other structures such as a conductor, a low resistance metal material such as Cu (copper) or Al (aluminum) may be used.

[0553] An insulator 650 is provided over the conductor 620 and the insulator 630. The insulator 650 can be provided using a material similar to that of the insulator 320. The insulator 650 may also function as a planarizing film that covers the uneven shape underneath.

[0554] By using this structure, in a semiconductor device including a transistor having an oxide semiconductor, a change in electrical characteristics can be suppressed and reliability can be improved, or miniaturization or high integration can be achieved in a semiconductor device including a transistor having an oxide semiconductor.

[0555] Next, another structure example of the OS transistor illustrated in Fig. 19 and Fig. 20 will be described. Fig. 22A and Fig. 22B are modifications of the transistor 500 illustrated in Fig. 21A and Fig. 21B, in which Fig. 22A is a cross-sectional view of the transistor 500 in the channel length direction and Fig. 22B is a cross-sectional view of the transistor 500 in the channel width direction. Note that the structures illustrated in Fig. 22A and Fig. 22B can also be applied to other transistors included in the semiconductor device of one embodiment of the present invention, such as the transistor 300.

[0556] 22A and 22B differs from the transistor 500 shown in Figures 21A and 21B in that the transistor 500 includes an insulator 402 and an insulator 404. The transistor 500 also differs from the transistor 500 shown in Figures 21A and 21B in that an insulator 552 is provided in contact with a side surface of the conductor 540a and an insulator 552 is provided in contact with a side surface of the conductor 540b. The transistor 500 also differs from the transistor 500 shown in Figures 21A and 21B in that an insulator 520 is not provided.

[0557] 22A and 22B, the insulator 402 is provided over the insulator 512. Furthermore, the insulator 404 is provided over the insulator 574 and the insulator 402.

[0558] 22A and 22B, the insulator 514, the insulator 516, the insulator 522, the insulator 524, the insulator 544, the insulator 580, and the insulator 574 are provided, and the insulator 404 covers them. That is, the insulator 404 is in contact with the top surface of the insulator 574, the side surface of the insulator 574, the side surface of the insulator 580, the side surface of the insulator 544, the side surface of the insulator 524, the side surface of the insulator 522, the side surface of the insulator 516, the side surface of the insulator 514, and the top surface of the insulator 402. As a result, the oxide 530 and the like are isolated from the outside by the insulator 404 and the insulator 402.

[0559] The insulators 402 and 404 preferably have a high function of suppressing diffusion of hydrogen (for example, at least one of hydrogen atoms, hydrogen molecules, and the like) or water molecules. For example, the insulators 402 and 404 are preferably made of silicon nitride or silicon nitride oxide, which are materials with high hydrogen barrier properties. This can suppress diffusion of hydrogen and the like into the oxide 530, thereby suppressing deterioration of the characteristics of the transistor 500. Therefore, the reliability of the semiconductor device of one embodiment of the present invention can be improved.

[0560] The insulator 552 is provided in contact with the insulator 581, the insulator 404, the insulator 574, the insulator 580, and the insulator 544. The insulator 552 preferably has a function of suppressing diffusion of hydrogen or water molecules. For example, the insulator 552 is preferably made of an insulator having a high hydrogen barrier property, such as silicon nitride, aluminum oxide, or silicon nitride oxide. In particular, silicon nitride is preferably used as the insulator 552 because it has a high hydrogen barrier property. By using a material having a high hydrogen barrier property as the insulator 552, impurities such as water or hydrogen can be suppressed from diffusing from the insulator 580 or the like to the oxide 530 through the conductor 540a and the conductor 540b. Furthermore, oxygen contained in the insulator 580 can be suppressed from being absorbed by the conductor 540a and the conductor 540b. As described above, the reliability of the semiconductor device of one embodiment of the present invention can be improved.

[0561] 23 is a cross-sectional view illustrating a configuration example of a semiconductor device in which the transistor 500 and the transistor 300 have the configurations shown in FIGS. 22A and 22B. An insulator 552 is provided on the side surface of the conductor 546.

[0562] The transistor 500 shown in FIGS. 22A and 22B may have a modified configuration depending on the situation. For example, the transistor 500 shown in FIGS. 22A and 22B may be modified to the transistor shown in FIG. 24 as a modified example. FIG. 24A is a cross-sectional view of the transistor in the channel length direction, and FIG. 24B is a cross-sectional view of the transistor in the channel width direction. The transistor shown in FIGS. 24A and 24B differs from the transistor shown in FIGS. 22A and 22B in that the oxide 530c has a two-layer structure of an oxide 530c1 and an oxide 530c2.

[0563] The oxide 530c1 is in contact with the top surface of the insulator 524, the side surface of the oxide 530a, the top surface and side surface of the oxide 530b, the side surfaces of the conductors 542a and 542b, the side surface of the insulator 544, and the side surface of the insulator 580. The oxide 530c2 is in contact with the insulator 550.

[0564] The oxide 530c1 may be, for example, an In-Zn oxide. The oxide 530c2 may be the same material as that used for the oxide 530c when the oxide 530c has a single-layer structure. For example, the oxide 530c2 may be a metal oxide with an atomic ratio of n:Ga:Zn=1:3:4, Ga:Zn=2:1, or Ga:Zn=2:5.

[0565] By forming the oxide 530c in a two-layer structure of the oxide 530c1 and the oxide 530c2, the on-current of the transistor can be increased compared to when the oxide 530c has a single-layer structure. Therefore, the transistor can be applied as, for example, a power MOS transistor. The oxide 530c of the transistor having the configuration shown in FIG. 21A and FIG. 21B can also have a two-layer structure of the oxide 530c1 and the oxide 530c2.

[0566] 24A and 24B can be used, for example, as the transistor 300 shown in Figures 19 and 20. As described above, the transistor 300 can be used, for example, as the transistor Tr12 in the memory cell array CA included in the arithmetic circuit MAC1 described in the above embodiment. Note that the transistors shown in Figures 24A and 24B can also be used as transistors other than the transistors 300 and 500 included in the semiconductor device of one embodiment of the present invention.

[0567] 25 is a cross-sectional view illustrating a structural example of a semiconductor device in which the transistor 500 has the transistor structure illustrated in FIG. 21A and the transistor 300 has the transistor structure illustrated in FIG. 24A. Note that an insulator 552 is provided on a side surface of the conductor 546, as in the case of FIG. 23. As illustrated in FIG. 25, in the semiconductor device of one embodiment of the present invention, the transistors 300 and 500 are both OS transistors, but the transistors 300 and 500 can have different structures.

[0568] Next, a capacitive element that can be applied to the semiconductor device of FIGS. 19 and 20 will be described.

[0569] Fig. 26A to Fig. 26C show a capacitive element 600A as an example of the capacitive element 600 that can be applied to the semiconductor device shown in Fig. 19. Fig. 26A is a top view of the capacitive element 600A, Fig. 26B is a perspective view showing a cross section of the capacitive element 600A taken along dashed line L3-L4, and Fig. 26C is a perspective view showing a cross section of the capacitive element 600A taken along dashed line W3-L4.

[0570] The conductor 610 functions as one of a pair of electrodes of the capacitor 600A, and the conductor 620 functions as the other of the pair of electrodes of the capacitor 600A. The insulator 630 functions as a dielectric material sandwiched between the pair of electrodes.

[0571] As the insulator 630, for example, silicon oxide, silicon oxynitride, silicon nitride oxide, silicon nitride, aluminum oxide, aluminum oxynitride, aluminum nitride oxide, aluminum nitride, hafnium oxide, hafnium oxynitride, hafnium nitride oxide, hafnium nitride, zirconium oxide, etc. can be used, and can be provided as a stacked layer or a single layer.

[0572] Also, for example, a laminated structure of a material with high dielectric strength, such as silicon oxynitride, and a high dielectric constant (high-k) material may be used for the insulator 630. With this configuration, the capacitive element 600A can ensure sufficient capacitance by having an insulator with high dielectric constant (high-k), and the capacitive element 600A can improve its dielectric strength by having an insulator with high dielectric strength, thereby suppressing electrostatic breakdown of the capacitive element 600A.

[0573] Examples of high dielectric constant (high-k) material insulators (materials with a high relative dielectric constant) include gallium oxide, hafnium oxide, zirconium oxide, oxides having aluminum and hafnium, oxynitrides having aluminum and hafnium, oxides having silicon and hafnium, oxynitrides having silicon and hafnium, and nitrides having silicon and hafnium.

[0574] Alternatively, the insulator 630 may be a single layer or a multilayer insulator containing a high-k material such as aluminum oxide, hafnium oxide, tantalum oxide, zirconium oxide, lead zirconate titanate (PZT), strontium titanate (SrTiO3), or (Ba, Sr)TiO3 (BST). For example, when the insulator 630 is a multilayer, a three-layer stack in which zirconium oxide, aluminum oxide, and zirconium oxide are formed in this order, or a four-layer stack in which zirconium oxide, aluminum oxide, zirconium oxide, and aluminum oxide are formed in this order may be used. Alternatively, the insulator 630 may be a compound containing hafnium and zirconium. As semiconductor devices become finer and more highly integrated, problems such as leakage current in transistors and capacitors may occur due to the thinning of the dielectric material used in the gate insulator and the capacitor. By using a high-k material for the gate insulator and the insulator that functions as the dielectric used in the capacitor, it is possible to reduce the gate potential during transistor operation and ensure the capacitance of the capacitor while maintaining the physical film thickness.

[0575] The capacitor 600 is electrically connected to the conductor 546 and the conductor 548 at the bottom of the conductor 610. The conductor 546 and the conductor 548 function as plugs or wiring for connecting to other circuit elements. In addition, in Figures 26A to 26C, the conductor 546 and the conductor 548 are collectively referred to as the conductor 540.

[0576] Also, in order to clarify the figures, Figures 26A to 26C omit insulator 586 in which conductor 546 and conductor 548 are embedded, and insulator 650 covering conductor 620 and insulator 630.

[0577] 19, 20, and 26A to 26C is a planar type capacitive element, but the shape of the capacitive element is not limited to this. For example, the capacitive element 600 may be a cylindrical capacitive element 600B shown in FIGS. 27A to 27C.

[0578] 27A is a top view of the capacitive element 600B, FIG. 27B is a cross-sectional view of the capacitive element 600B taken along dashed line L3-L4, and FIG. 27C is a perspective view showing the cross-section of the capacitive element 600B taken along dashed line W3-L4.

[0579] In FIG. 27B, the capacitor 600B has an insulator 631 on an insulator 586 in which a conductor 540 is embedded, an insulator 651 having an opening, a conductor 610 that functions as one of a pair of electrodes, and a conductor 620 that functions as the other of the pair of electrodes.

[0580] Also, in FIG. 27C, insulator 586, insulator 650, and insulator 651 are omitted for clarity.

[0581] The insulator 631 can be, for example, a material similar to that of the insulator 586.

[0582] Furthermore, a conductor 611 is embedded in the insulator 631 so as to be electrically connected to the conductor 540. The conductor 611 can be made of, for example, the same material as the conductors 330 and 518.

[0583] The insulator 651 can be, for example, a material similar to that of the insulator 586.

[0584] As described above, the insulator 651 has an opening that overlaps with the conductor 611.

[0585] The conductor 610 is formed on the bottom and side surfaces of the opening. That is, the conductor 610 overlaps with the conductor 611 and is electrically connected to the conductor 611.

[0586] The conductor 610 is formed by forming an opening in the insulator 651 by etching or the like, and then depositing the conductor 610 by sputtering, ALD, or the like. Thereafter, the conductor 610 deposited on the insulator 651 may be removed by chemical mechanical polishing (CMP) or the like, leaving the conductor 610 deposited in the opening.

[0587] The insulator 630 is located on the insulator 651 and on the surface on which the conductor 610 is formed. The insulator 630 functions as a dielectric material sandwiched between a pair of electrodes in the capacitor.

[0588] Conductor 620 is formed on insulator 630 so that the opening of insulator 651 is filled.

[0589] The insulator 650 is formed to cover the insulator 630 and the conductor 620 .

[0590] 27A to 27C can have a higher capacitance value than the planar type capacitive element 600A. Therefore, by using the capacitive element 600B as the capacitors C1 and C2 described in the above embodiment, for example, the voltage between the terminals of the capacitors can be maintained for a long time.

[0591] Note that this embodiment mode can be appropriately combined with other embodiment modes described in this specification.

[0592] (Embodiment 4) In this embodiment, structures of a cloud-aligned composite oxide semiconductor (CAC-OS) and a c-axis aligned crystalline oxide semiconductor (CAAC-OS), which are metal oxides that can be used for the OS transistor described in the above embodiment, will be described.

[0593] <Metal oxide composition> CAC-OS or CAC-metal oxide has a conductive function in a part of the material and an insulating function in a part of the material, and has a function as a semiconductor as a whole. When CAC-OS or CAC-metal oxide is used in an active layer of a transistor, the conductive function is a function of flowing electrons (or holes) as carriers, and the insulating function is a function of not flowing electrons as carriers. By making the conductive function and the insulating function act complementarily, a switching function (on / off function) can be imparted to CAC-OS or CAC-metal oxide. By separating the respective functions in CAC-OS or CAC-metal oxide, both functions can be maximized.

[0594] Moreover, the CAC-OS or CAC-metal oxide has a conductive region and an insulating region. The conductive region has the above-mentioned conductive function, and the insulating region has the above-mentioned insulating function. In addition, in the material, the conductive region and the insulating region may be separated at the nanoparticle level. In addition, the conductive region and the insulating region may be unevenly distributed in the material. In addition, the conductive region may be observed to be connected in a cloud shape with a blurred periphery.

[0595] In addition, in the CAC-OS or CAC-metal oxide, the conductive regions and the insulating regions may each be dispersed in the material with a size of 0.5 nm to 10 nm, preferably 0.5 nm to 3 nm.

[0596] In addition, the CAC-OS or CAC-metal oxide is composed of components having different band gaps. For example, the CAC-OS or CAC-metal oxide is composed of a component having a wide gap due to an insulating region and a component having a narrow gap due to a conductive region. In this configuration, when carriers are caused to flow, the carriers mainly flow in the component having the narrow gap. In addition, the component having the narrow gap acts complementarily on the component having the wide gap, and carriers also flow in the component having the wide gap in conjunction with the component having the narrow gap. Therefore, when the above CAC-OS or CAC-metal oxide is used in the channel formation region of a transistor, a high current driving force in the on state of the transistor, that is, a large on-current and a high field effect mobility can be obtained.

[0597] That is, CAC-OS or CAC-metal oxide can also be called a matrix composite or a metal matrix composite.

[0598] <Metal oxide structure> Oxide semiconductors are classified into single-crystal oxide semiconductors and other non-single-crystal oxide semiconductors, such as CAAC-OS, polycrystalline oxide semiconductors, nanocrystalline oxide semiconductors (nc-OS), amorphous-like oxide semiconductors (a-like OS), and amorphous oxide semiconductors.

[0599] In addition, when focusing on the crystal structure, oxide semiconductors may be classified differently from the above. Here, classification of crystal structures in oxide semiconductors will be described with reference to Fig. 28A. Fig. 28A is a diagram for explaining classification of crystal structures of oxide semiconductors, typically IGZO (metal oxide containing In, Ga, and Zn).

[0600] As shown in FIG. 28A, IGZO is broadly classified into Amorphous, Crystalline, and Crystal. Amorphous includes completely amorphous. Crystalline includes c-axis aligned crystalline (CAAC), nanocrystalline (nc), and cloud-aligned composite (CAC). Crystalline excludes single crystal, poly crystal, and completely amorphous. Crystal includes single crystal and poly crystal.

[0601] The structure in the bold frame shown in FIG. 28A is an intermediate state between Amorphous and Crystal, and belongs to a new boundary region (New crystalline phase). This structure is in the boundary region between Amorphous and Crystal. In other words, this structure can be said to be a structure that is completely different from Amorphous, which is energetically unstable, or Crystal.

[0602] The crystal structure of the film or substrate can be evaluated using an X-ray diffraction (XRD) image. Here, XRD spectra of quartz glass and IGZO (also called crystalline IGZO) having a crystal structure classified as Crystalline are shown in Fig. 28B and Fig. 28C. Fig. 28B shows the XRD spectrum of quartz glass, and Fig. 28C shows the XRD spectrum of crystalline IGZO. The composition of the crystalline IGZO shown in Fig. 28C is in the vicinity of In:Ga:Zn=4:2:3 [atomic ratio]. The thickness of the crystalline IGZO shown in Fig. 28C is 500 nm.

[0603] As shown by the arrows in FIG. 28B, the shape of the peak in the XRD spectrum of silica glass is almost symmetrical. On the other hand, as shown by the arrows in FIG. 28C, the peak in the XRD spectrum of crystalline IGZO is asymmetrical. The asymmetric shape of the peak in the XRD spectrum clearly indicates the presence of crystals. In other words, if the shape of the peak in the XRD spectrum is not symmetrical, it cannot be said to be amorphous. Note that FIG. 28C clearly shows the crystal phase (IGZO crystal phase) at or near 2θ=31°. It is presumed that the reason why the shape of the peak in the XRD spectrum is asymmetrical is due to the crystal phase (microcrystals).

[0604] Specifically, the XRD spectrum of crystalline IGZO shown in FIG. 28C has a peak at or near 2θ=34°. Microcrystals have a peak at or near 2θ=31°. When an oxide semiconductor film is evaluated from an X-ray diffraction image, the width of the spectrum on the lower angle side is wider than the peak at or near 2θ=34°, as shown in FIG. 28C. This suggests that the oxide semiconductor film contains microcrystals having a peak at or near 2θ=31°.

[0605] The crystal structure of the film can be evaluated from the diffraction pattern (also called the nanobeam electron diffraction pattern) observed by nanobeam electron diffraction (NBED). Figure 28D shows the diffraction pattern of an IGZO film formed at room temperature on the substrate. The IGZO film shown in Figure 28D was formed by sputtering using an oxide target with an In:Ga:Zn=1:1:1 [atomic ratio]. In the nanobeam electron diffraction method, electron diffraction was performed with a probe diameter of 1 nm.

[0606] As shown in Figure 28D, a spot-like pattern, not a halo, is observed in the diffraction pattern of the IGZO film formed at room temperature. Therefore, it is presumed that the IGZO film formed at room temperature is in an intermediate state, neither crystalline nor amorphous, and it cannot be concluded that it is in an amorphous state.

[0607] CAAC-OS has a c-axis orientation and a distorted crystal structure in which multiple nanocrystals are connected in the ab-plane direction. The distortion refers to a portion where the lattice orientation changes between a region with a uniform lattice arrangement and a region with a different uniform lattice arrangement in the region where multiple nanocrystals are connected.

[0608] Nanocrystals are basically hexagonal, but may be non-regular hexagonal. In addition, the lattice arrangement may be pentagonal or heptagonal due to the distortion. In CAAC-OS, no clear grain boundary can be confirmed even in the vicinity of the distortion. In other words, it is found that the formation of grain boundaries is suppressed by the distortion of the lattice arrangement. This is thought to be because CAAC-OS can tolerate distortion due to the fact that the arrangement of oxygen atoms in the ab-plane direction is not dense and the bond distance between atoms changes due to the substitution of metal elements.

[0609] A crystal structure in which clear grain boundaries are observed is called polycrystal. The grain boundaries are likely to become recombination centers and capture carriers, causing a decrease in the on-current of a transistor or a decrease in field-effect mobility. Therefore, CAAC-OS in which clear grain boundaries are not observed is one of the crystalline oxides having a crystal structure suitable for a semiconductor layer of a transistor. In order to form a CAAC-OS, a structure containing Zn is preferable. For example, In-Zn oxide and In-Ga-Zn oxide are suitable because they can suppress the occurrence of grain boundaries more than In oxide.

[0610] CAAC-OS also tends to have a layered crystal structure (also referred to as a layered structure) in which a layer containing indium and oxygen (hereinafter, In layer) and a layer containing element M, zinc, and oxygen (hereinafter, (M, Zn) layer) are stacked. Note that indium and element M can be substituted for each other, and when element M in an (M, Zn) layer is substituted for indium, it can also be represented as an (In, M, Zn) layer. When indium in an In layer is substituted for element M, it can also be represented as an (In, M) layer.

[0611] CAAC-OS is an oxide semiconductor with high crystallinity. On the other hand, since no clear crystal grain boundaries can be identified in CAAC-OS, it can be said that a decrease in electron mobility due to crystal grain boundaries is unlikely to occur. In addition, since the crystallinity of an oxide semiconductor can be decreased by the inclusion of impurities or the generation of defects, CAAC-OS can be said to be an oxide semiconductor with few impurities and defects (such as oxygen vacancies). Therefore, an oxide semiconductor having CAAC-OS has stable physical properties. Therefore, an oxide semiconductor having CAAC-OS is resistant to heat and has high reliability. In addition, CAAC-OS is stable against high temperatures (so-called thermal budget) in the manufacturing process. Therefore, the use of CAAC-OS in an OS transistor can increase the degree of freedom in the manufacturing process.

[0612] The nc-OS has periodic atomic arrangement in a microscopic region (for example, a region of 1 nm to 10 nm, particularly a region of 1 nm to 3 nm). In addition, the nc-OS does not show regularity in the crystal orientation between different nanocrystals. Therefore, no orientation is seen throughout the film. Therefore, depending on the analysis method, the nc-OS may be indistinguishable from an a-like OS or an amorphous oxide semiconductor.

[0613] The a-like OS is an oxide semiconductor having a structure between the nc-OS and an amorphous oxide semiconductor. The a-like OS has a void or low-density region. That is, the a-like OS has lower crystallinity than the nc-OS and CAAC-OS.

[0614] Oxide semiconductors have a variety of structures and have different characteristics. The oxide semiconductor of one embodiment of the present invention may include two or more of an amorphous oxide semiconductor, a polycrystalline oxide semiconductor, an a-like OS, an nc-OS, and a CAAC-OS.

[0615] <Transistor Having Oxide Semiconductor> Next, the case where the oxide semiconductor is used for a transistor will be described.

[0616] By using the oxide semiconductor for a transistor, a transistor with high field-effect mobility and high reliability can be realized.

[0617] In addition, it is preferable to use an oxide semiconductor with a low carrier concentration for the transistor. In order to reduce the carrier concentration of the oxide semiconductor film, the impurity concentration in the oxide semiconductor film may be reduced to reduce the density of defect states. In this specification and the like, a low impurity concentration and a low density of defect states may be referred to as highly pure intrinsic or substantially highly pure intrinsic, or may be referred to as intrinsic or substantially intrinsic.

[0618] Furthermore, a highly purified intrinsic or substantially highly purified intrinsic oxide semiconductor film has a low density of defect states, and therefore the density of trap states might also be low.

[0619] In addition, charges trapped in the trap states of an oxide semiconductor take a long time to disappear and may behave as if they are fixed charges. Therefore, a transistor in which a channel formation region is formed in an oxide semiconductor with a high density of trap states may have unstable electrical characteristics.

[0620] Therefore, in order to stabilize the electrical characteristics of a transistor, it is effective to reduce the impurity concentration in the oxide semiconductor. In order to reduce the impurity concentration in the oxide semiconductor, it is preferable to also reduce the impurity concentration in a nearby film. Examples of impurities include hydrogen, nitrogen, alkali metals, alkaline earth metals, iron, nickel, silicon, and the like.

[0621] <Impurities> Here, the influence of each impurity in an oxide semiconductor will be described.

[0622] When an oxide semiconductor contains silicon or carbon, which is one of the group 14 elements, defect levels are formed in the oxide semiconductor. For this reason, the concentrations of silicon or carbon in the oxide semiconductor and those near the interface with the oxide semiconductor (concentrations obtained by secondary ion mass spectrometry (SIMS)) are calculated as follows: 18 atoms / cm 3 Less than or equal to 2 x 10 17 atoms / cm 3 The following applies.

[0623] Furthermore, when an oxide semiconductor contains an alkali metal or an alkaline earth metal, defect levels are formed and carriers are generated in some cases. Therefore, a transistor using an oxide semiconductor containing an alkali metal or an alkaline earth metal is likely to have normally-on characteristics. For this reason, it is preferable to reduce the concentration of the alkali metal or the alkaline earth metal in the oxide semiconductor. Specifically, when the concentration of the alkali metal or the alkaline earth metal in the oxide semiconductor obtained by SIMS is reduced to 1×10 18 atoms / cm 3 Less than or equal to 2 x 10 16 atoms / cm 3 To the following:

[0624] Furthermore, when nitrogen is contained in an oxide semiconductor, electrons serving as carriers are generated, the carrier concentration increases, and the oxide semiconductor is likely to become n-type. As a result, a transistor using an oxide semiconductor containing nitrogen as a semiconductor is likely to have normally-on characteristics. Therefore, it is preferable that the nitrogen content in the oxide semiconductor is reduced as much as possible. For example, the nitrogen concentration in the oxide semiconductor is less than 5×10 19 atoms / cm 3 Less than 5 x 10 18 atoms / cm 3 Less than or equal to 1×10 18 atoms / cm 3 Less than 5×10, more preferably 17 atoms / cm 3 The following applies.

[0625] Furthermore, hydrogen contained in an oxide semiconductor reacts with oxygen bonded to a metal atom to form water, which may form an oxygen vacancy. When hydrogen enters the oxygen vacancy, electrons serving as carriers may be generated. In addition, some of the hydrogen may bond with oxygen bonded to a metal atom to generate electrons serving as carriers. Therefore, a transistor using an oxide semiconductor containing hydrogen is likely to have normally-on characteristics. For this reason, it is preferable to reduce hydrogen in the oxide semiconductor as much as possible. Specifically, when the hydrogen concentration in an oxide semiconductor measured by SIMS is 1×10 20 atoms / cm 3 Less than 1 x 10 19 atoms / cm 3 less than 5×10 18 atoms / cm 3 less than 1×10 18 atoms / cm 3 Less than.

[0626] When an oxide semiconductor in which impurities are sufficiently reduced is used for a channel formation region of a transistor, stable electrical characteristics can be obtained.

[0627] Note that this embodiment mode can be appropriately combined with other embodiment modes described in this specification. EXAMPLES

[0628] In this embodiment, in order to confirm whether the calculation of the product of the first data and the second data is performed properly in the memory cell AM of the arithmetic circuit MAC1 described in the second embodiment, an actual circuit was prototyped and various measurements and calculations were performed.

[0629] <Measurements and Calculations> The multiplication circuit AME shown in FIG. 29A is a part of an arithmetic circuit actually prototyped, and corresponds to the memory cell AM of the arithmetic circuit MAC1 described in the second embodiment. Therefore, the transistor M1 and the capacitance CP included in the multiplication circuit AME correspond to the transistor M1 and the capacitance C1 included in the memory cell AM shown in FIG. 9, respectively. In particular, the transistor M2-1 and the transistor M2-2 included in the multiplication circuit AME correspond to the transistor M2 included in the memory cell AM shown in FIG. 9. That is, the transistor M2-1 and the transistor M2-2 are electrically connected in series, and their gates are electrically connected to each other. In this embodiment, the transistor M2-1 and the transistor M2-2 are collectively referred to as the transistor M2. Also, the wiring VY shown in FIG. 29A corresponds to the wiring BL in FIG. 9, the wiring BW shown in FIG. 29A corresponds to the wiring WD in FIG. 9, the wiring VX shown in FIG. 29A corresponds to the wiring CL in FIG. 9, and the wiring WW shown in FIG. 29A corresponds to the wiring WL in FIG. 9.

[0630] Furthermore, the transistor M1 included in the multiplication circuit AME has a back gate, and the back gate is electrically connected to the wiring BG.

[0631] The transistor M1 is an OS transistor whose channel formation region contains In-Ga-Zn oxide, and the channel length (hereinafter, L length) of the transistor M1 is 0.35 μm and the channel width (hereinafter, W length) is 0.35 μm. The transistor M2 is a Si transistor whose channel formation region contains single crystal silicon, and the L length of the transistors M2-1 and M2-2 is 8 μm and the W length is 0.32 μm.

[0632] 29B is a photograph of the top surface of a cell array CA3 having a prototype multiplication circuit AME taken with an optical microscope. In the cell array CA3, the multiplication circuits AME are arranged in a 9×16 matrix. Each multiplication circuit AME of the cell array CA3 is electrically connected to each of the electrode pads EP1 to EP6. The electrode pad EP1 is electrically connected to the wiring WW, the electrode pad EP2 is electrically connected to the wiring BW, the electrode pad EP3 is electrically connected to the wiring VX, the electrode pad EP4 is electrically connected to the wiring BG, the electrode pad EP5 is electrically connected to the wiring VY, and the electrode pad EP6 is electrically connected to the wiring VR.

[0633] In the multiplication circuit AME, data writing, data holding, and data reading were performed.

[0634] When writing data to the multiplication circuit AME, the following voltages were applied: 5V to the wire WW, 0V to the wire VX, -6V to the wire BG, 3V to the wire VY, and 0V to the wire VR. The data to be written to the node NM is V W The potential was applied from the wiring BW.

[0635] When the data written to the multiplication circuit AME was to be held, voltages were applied to the multiplication circuit AME by setting the wiring WW to 0V, the wiring BW to 0V, the wiring VX to 0V, the wiring BG to -6V, the wiring VY to 0V, and the wiring VR to 0V.

[0636] When reading out the written data from the multiplication circuit AME, the following voltages were applied to the wire WW: 0 V, the wire BW: 0 V, the wire BG: -6 V, the wire VY: 3 V, and the wire VR: 0 V. The potential applied to the wire VX was a potential V in the range of 0 V to 3.0 V inclusive in increments of 0.1 V. X It was decided.

[0637] In the multiplication circuit AME, the potentials applied to the wiring WW, wiring BW, wiring VX, wiring BG, wiring VY, and wiring VR when a write operation, data retention, and read operation are performed are summarized in the table below.

[0638] [Table 2]

[0639] When the data written to the multiplication circuit AME is read, the current flowing from the line VY through the source-drain of the transistor M2 to the line VR is measured.

[0640] FIG. 30A shows the potential V W and potential V X and the source-drain current I DS (V W , V X ) characteristics. From Figure 30A, V X is fixed at an arbitrary potential, V W By increasing I DS (V W , V X ) was increased. W Increasing V is equivalent to shifting the threshold of M2 to the negative side. W is fixed at an arbitrary potential, V X By increasing I DS (V W , V X ) was increased. W is set to 1.5V, and V X When is set to 1.5V, I DS (V W , V X) was estimated to be approximately 1.3 μA.

[0641] Here, V W0 is set to 1.5V, and V X0 If we set it to 1.5V, then I DS (V W0 , V X0 ) = 1.3μA. Furthermore, V W0 The voltage change is ΔV W When I is given DS (V W0 +ΔV W , V X0 ), V X0 The voltage change is ΔV X When I is given DS (V W0 , V X0 +ΔV X ), and V W0 The voltage change is ΔV W Given V X0 The voltage change is ΔV X When I is given DS (V W0 +ΔV W , V X0 +ΔV X ) and ΔI y is defined as follows:

[0642]

number

[0643] In addition, V W0 +ΔV W is the voltage given by the wire BW, so ΔV W The voltage range of is -1V to 1V. X0 +ΔV X is the voltage given by the wire BW, so ΔV X The voltage range is between -1.5V and 1.5V.

[0644] Then, using equations (E1), (E2), (E7), and (E8), ΔI y The calculation is as follows:

[0645]

number

[0646] Here, by setting k to 1 / 2 (normalizing it to an appropriate value), ΔI y and ΔV W and ΔV X The relationship between and can be expressed in FIG. W and ΔV X The differential current ΔI y Therefore, the current flowing between the source and drain of the transistor M2 is determined as I DS (V W0 , V X0 ), I DS (V W0 +ΔV W , V X0 ), I DS (V W0 , V X0 +ΔV X ), I DS (V W0 +ΔV W , V X0 +ΔV X ) and calculate the difference current ΔI y By calculating ΔV W and ΔV X The product of can be found.

[0647] FIG. 31 shows the results of applying 0.5V (ΔV W =-1.0V), 2.5V (ΔV W = 1.0V) and then performing a read operation. Figure 31 shows that, compared to 27°C, the slope at 85°C is 40% higher and at -40°C is 30% lower. The results for 85°C and -40°C show the temperature dependence of saturation mobility. The slope differs depending on the temperature, but each ΔVW ΔI in y and ΔV X The correlation with is 0.989 or higher, and it is believed that the slope can be easily corrected by appropriately normalizing the correlation according to the temperature.

[0648] 32A and 32B are graphs showing the change over time in the amount of current flowing between the source and drain of the transistor M2 after data is held in the node NM. X is set to 1.0 V (the potential of the wire VX is set to 2.5 V), and the measurement conditions for FIG. 32B are ΔV X In each of FIG. 32A and FIG. 32B, the potential held at the node NM is set to 0.5 V (ΔV W =-1.0V), 1.0V (ΔV W =-0.5V), 1.5V (ΔV W =0V), 2.0V (ΔV W =0.5V), 2.5V (ΔV W = 1.0V), the current I flowing between the source and drain of transistor M2 under each condition is y was measured.

[0649] As shown in Figures 32A and 32B, 2 Seconds to 1.0 x 10 5 In the first few seconds, the current I flowing between the source and drain of transistor M2 under each condition is y The result was that there was no change in the 1.0×10 5 Differential current ΔI up to seconds y The rate of change in the 1.08×10 5 Differential current ΔI up to seconds y The rate of change was less than 4%.

[0650] FIG. 33A shows a multiplication circuit AME with a node NM at 0.5 V (ΔV W =-1.0V), 1.0V (ΔV W=-0.5V), 2.0V (ΔV W =0.5V), 2.5V (ΔV W = 1.0V), and then write the differential current ΔI y 33A is a graph showing the multiplication characteristics of the multiplication circuit AME obtained by reading out ΔI y =1.0×V X , ΔI y =0.5×V X , ΔI y =-0.5×V X , ΔI y =-1.0×V X From FIG. 33A, the result of the multiplication characteristic in the multiplication circuit AME is ΔI y =1.0×V X , ΔI y =0.5×V X , ΔI y =-0.5×V X , ΔI y =-1.0×V X It was found that the results roughly correspond to the linear functions of

[0651] FIG. 33B shows that a 0.5V (ΔV W =-1.0V), 1.0V (ΔV W =-0.5V), 2.0V (ΔV W =0.5V), 2.5V (ΔV W = 1.0V) and then performing a read operation. X is set to 1.0 V (the potential of the wiring VX is set to 2.5 V). After the write operation, a read operation is performed to obtain the differential current ΔI y The calculation is repeated 50 times, and the average of the 50 times is the ΔV W Differential current ΔI at y This was set as one set, and each ΔI W About 50 sets of differential current ΔI y Measurements were carried out.

[0652] In FIG. 33B, the horizontal axis represents the differential current ΔI y 33B, the variation in data writing is generally within the range of -0.4% to 0.4%.

[0653] In addition, for the 12 multiplier circuits AME, 0.5V (ΔV W =-1.0V), 1.0V (ΔV W =-0.5V), 2.0V (ΔV W =0.5V), 2.5V (ΔV W = 1.0V) was written, the ΔI read from each multiplier circuit AME y The degree of element variation is shown in FIG. 34. W ΔV at X The dependence of ΔV X Except for the vicinity of =0 (the area greater than -0.02 and smaller than 0.02), ΔI y The element variation of each ΔV W , ΔV X The rate is less than 5%. y ΔV of element variation X The dependence on ΔV W It is shown that there is no dependence on the absolute value of ΔI y The cause of the device variation in is considered to be, for example, the variation in the saturation mobility of the Si transistor, and the drain current in the saturation region of the Si transistor deviating from the gradual approximation (square approximation). meas is (ΔV W ,ΔV X )=(+1,+1), σ meas = 0.023, (ΔV W ,ΔV X )=(+1,-1), σ meas = 0.025, (ΔV W ,ΔV X )=(-1,+1), σ meas = 0.034, (ΔV W ,ΔV X)=(-1,-1), σ meas =0.032.

[0654] Also, ΔI in Figure 34 y In order to confirm the validity of the cause of the element variation, a Monte Carlo analysis was performed and the results are shown in FIGS. 35A to 35D. In FIGS. 35A to 35D, local variations are set, and in the circuit configuration of FIG. 29A, (ΔV W ,ΔV X 35A to 35D show the element variations obtained by simulation in the cases of σ=(+1,+1), (+1,-1), (-1,+1), and (-1,-1). sim is (ΔV W ,ΔV X )=(+1,+1), σ sim = 0.051, (ΔV W ,ΔV X )=(+1,-1), σ sim = 0.038, (ΔV W ,ΔV X )=(-1,+1), σ sim = 0.025, (ΔV W ,ΔV X )=(-1,-1), σ sim = 0.017. The element variation obtained by measurement was σ meas From Figure 34, it is 0.023 to 0.034, so the element variation σ obtained by Monte Carlo analysis is sim It was confirmed that the values ​​were generally consistent, ranging from 0.017 to 0.051.

[0655] Next, the inference accuracy of a three-layer fully connected artificial neural network model was calculated when using an arithmetic circuit having the cell array CA3 of Fig. 29B. The model of the neural network is shown in Fig. 36, and the neural network has an input layer, a middle layer, and an output layer. The input layer has 784 neurons, the middle layer has 100 neurons, and the output layer has 10 neurons.

[0656] The neural network was implemented on a computer using the program language Python, and the weight coefficient was calculated by learning the implemented neural network using the handwritten character data set MNIST. Next, the arithmetic circuit having the cell array CA3 of FIG. 29B was set as the multiplication in the model of FIG. 36, and the weight coefficient was held in each of the multiplication circuits AME of the cell array CA3, and inference was performed. The activation function in the intermediate layer was set as a sigmoid function, and the activation function in the output layer was set as a softmax function. As a result, the inference accuracy of the arithmetic circuit having the cell array CA3 of FIG. 29B was 97.77%. In the model of FIG. 36, the inference accuracy was 97.89% when the operation in the product-sum operation was set as ideal multiplication (the calculation was performed on a computer using the program language Python). Therefore, in the model of FIG. 36, the inference accuracy when the arithmetic circuit of this embodiment was used was almost the same as the inference accuracy when the product-sum operation was set as ideal multiplication. EXAMPLES

[0657] In this embodiment, in order to confirm whether the multiplication and accumulation of the first data and the second data is properly performed in the arithmetic circuit MAC1A described in the second embodiment, various calculations are performed using a circuit simulator.

[0658] First, a circuit configuration for performing various calculations will be described. Fig. 37 shows an example of a circuit configuration obtained by modifying the arithmetic circuit MAC1A shown in Fig. 13. Therefore, in the arithmetic circuit MAC1A shown in Fig. 37, the description of the parts that overlap with the arithmetic circuit MAC1A in Fig. 13 will be omitted.

[0659] The arithmetic circuit MAC1A shown in Fig. 37 has a configuration in which a plurality of columns of memory cells AM of the memory cell array CA of Fig. 13 are arranged, and the circuit CMS of Fig. 13 and the circuit OFAC of Fig. 13 are modified. Moreover, the memory cells AM of the memory cell array CA of Fig. 37 may be arranged in a plurality of rows.

[0660] Since the memory cells AM of the memory cell array CA in Fig. 37 are arranged in multiple columns, the circuit CMS in Fig. 37 has a configuration obtained by modifying the circuit CMS in Fig. 13 according to the memory cells AM arranged in multiple columns. For example, the current mirror circuit CM included in the circuit CMS in Fig. 37 has a transistor Tr32[1] and a transistor Tr32[2] corresponding to the transistor Tr32 in Fig. 13, a current source circuit CS1[1] and a current source circuit CS1[2] corresponding to the current source circuit CS1 in Fig. 13, a current source circuit CS2[1] and a current source circuit CS2[2] corresponding to the current source circuit CS2 in Fig. 13, and a switch SW3[1] and a switch SW3[2] corresponding to the switch SW3 in Fig. 13.

[0661] The transistor Tr32[1], the current source circuit CS1[1], and the current source circuit CS2[1] in Fig. 30 are circuits for performing a multiplication and accumulation operation between the first data held in the memory cell AM[1,1] and the memory cell AM[2,1] arranged in the first column of the memory cell array CA, and the second data input to the memory cell AM[1,1] and the memory cell AM[2,1]. The transistor Tr32[2], the current source circuit CS1[2], and the current source circuit CS2[2] in Fig. 30 are circuits for performing a multiplication and accumulation operation between the first data held in the memory cell AM[1,2] and the memory cell AM[2,2] arranged in the first column of the memory cell array CA, and the second data input to the memory cell AM[1,2] and the memory cell AM[2,2].

[0662] The circuit OFAC also has switches SW4[1] and SW4[2] that correspond to the switch SW4 in FIG. 13, resistors RE[1] and RE[2] that correspond to the resistor RE in FIG. 13, and an operational amplifier OP[1] and an operational amplifier OP[2].

[0663] The first terminal of the switch SW4[1] is electrically connected to the first terminal of the switch SW3[1], and the second terminal of the switch SW4[1] is electrically connected to the first terminal of the resistor RE[1] and the inverting input terminal of the operational amplifier OP[1]. The non-inverting input terminal of the operational amplifier OP[1] is electrically connected to the wiring VdL, and the output terminal of the operational amplifier OP[1] is electrically connected to the second terminal of the resistor RE[1] and the wiring NIL[1]. In other words, the resistor RE[1] and the operational amplifier OP[1] form a current-voltage conversion circuit.

[0664] The switch SW4[2], resistor RE[2], and operational amplifier OP[2] are electrically connected in the same manner as the switch SW4[1], resistor RE[1], and operational amplifier OP[1]. Therefore, a current-voltage conversion circuit is also formed between the resistor RE[2] and the operational amplifier OP[2].

[0665] The wiring VdL functions as a wiring for supplying a constant voltage. In particular, the constant voltage is input as a reference potential of the current-voltage conversion circuit described above.

[0666] The current-voltage conversion circuit, which is composed of the resistor RE[1] and the operational amplifier OP[1], has the function of converting the current I5 generated by the memory cell AM in the first column of the memory cell array CA, the current source circuit CS1[1], the current source circuit CS2[1], and the transistor Tr32[1], into a voltage. The current-voltage conversion circuit, which is composed of the resistor RE[2] and the operational amplifier OP[2], has the function of converting the current I5 generated by the memory cell AM in the second column of the memory cell array CA, the current source circuit CS1[2], the current source circuit CS2[2], and the transistor Tr32[2], into a voltage.

[0667] Here, the memory cell array CA in FIG. 37 was configured with memory cells AM arranged in a matrix of n rows and 1 column, and a product-sum operation was performed on the first data and the second data in the arithmetic circuit MAC1A using a circuit simulator.

[0668] The circuit configurations of the memory cell AM and the memory cell AMref are the same as those of the memory cell AM and the memory cell AMref shown in Fig. 13. The transistor Tr12 and the transistors Tr31 to Tr34 are assumed to be Si transistors, with an L length of 8 µm and a W length of 0.32 µm. The transistor Tr11 is assumed to be an OS transistor, with an L length of 0.35 µm and a W length of 0.35 µm.

[0669] The memory cells AM included in the memory cell array CA of the arithmetic circuit MAC1A input to the circuit simulator were set to memory cells AM[1,1] to AM[25,1], and the wirings CL for inputting the second data were set to wirings CL[1] to CL

[25] (i.e., n=25). Each of the memory cells AM[1,1] to AM[25,1] holds a potential corresponding to "-1" or "+1" as the first data (weighting coefficient), and a potential corresponding to "-1", "0", or "+1" was input to the wirings CL[1] to CL

[25] as the second data (value of the neuron signal).

[0670] 38A is a graph showing the calculation values ​​of the sum-of-products operation un...

Claims

1. A first electronic device is provided. the first electronic device has an input / output interface, a control unit, and a first conversion unit; The input / output interface is electrically connected to the control unit, The first conversion unit is electrically connected to the control unit, the input / output interface has a function of transmitting input data generated by a user's operation to the control unit; The control unit has a function of transmitting the input data to the first conversion unit, The first conversion unit has a circuit in which a neural network is configured, the first conversion unit has a function of converting the input data into a first netlist by the neural network; the input data is a circuit diagram depicting a circuit configuration or a document file depicting the circuit configuration, The first electronic device has a first database and a second database, the first database is electrically connected to the control unit; the second database is electrically connected to the control unit; The first database stores a plurality of second netlists; The second database stores a plurality of literature data items linked to the plurality of second netlists, The control unit is a function of searching the first database for a circuit configuration of the first netlist; a function of, in searching for a circuit configuration of the first netlist, comparing the first netlist with each of the plurality of second netlists, calculating a similarity between the first netlist and each of the plurality of second netlists, and outputting a search result from among the plurality of second netlists with a high similarity to the input / output interface; AI system.

2. A first electronic device, the first electronic device has an input / output interface, a control unit, and a first conversion unit; The input / output interface is electrically connected to the control unit, The first conversion unit is electrically connected to the control unit, the input / output interface has a function of transmitting input data generated by a user's operation to the control unit; The control unit has a function of transmitting the input data to the first conversion unit, The first conversion unit has a circuit in which a neural network is configured, the first conversion unit has a function of converting the input data into a first netlist by the neural network; the input data is a circuit diagram depicting a circuit configuration or a document file depicting the circuit configuration, The first electronic device has a first database and a second database, the first database is electrically connected to the control unit; the second database is electrically connected to the control unit; The first database stores a plurality of second netlists; The second database stores a plurality of literature data items linked to the plurality of second netlists, The control unit is a function of searching the first database for a circuit configuration of the first netlist; a function of, in searching for a circuit configuration of the first netlist, comparing the first netlist with each of the plurality of second netlists, calculating a similarity between the first netlist and each of the plurality of second netlists, and outputting the literature data from the second netlist having the highest similarity among the plurality of second netlists to the input / output interface; AI system.

Images