Information processing system and simulation method

The information processing system addresses the lack of real-time visualization in semiconductor manufacturing by using simulation models to optimize process states and predict failures, enhancing efficiency and reducing costs.

KR103045282B1Active Publication Date: 2026-11-06TOKYO ELECTRON LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
KR1020210018892
Authority / Receiving Office
KR · KR
Patent Type
Patents
Current Assignee / Owner
Priority Date
2020-02-20
Filing Date
2021-02-10
Publication Date
2026-11-06
Estimated Expiration
2041-02-10

AI Technical Summary

Technical Problem

Existing semiconductor manufacturing processes lack real-time visualization and predictive capabilities for process states, leading to inefficiencies and potential failures.

Method used

An information processing system that utilizes a simulation model to generate virtual sensor data and process results, integrating with physical sensors to visualize and optimize the process state in real-time, enabling fault detection and parameter adjustment.

Benefits of technology

Enables real-time visualization and predictive maintenance, reducing operational costs and improving process efficiency by minimizing errors and optimizing parameters.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 112021016998718-PAT00001_ABST
    Figure 112021016998718-PAT00001_ABST
Patent Text Reader

Abstract

The present invention aims to provide a technology for visualizing the process state of a semiconductor manufacturing device during the execution of a process in a semiconductor manufacturing device. The above problem is solved by having an information processing system that executes a simulation of a process state running in a semiconductor manufacturing device using a simulation model of the semiconductor manufacturing device, comprising: a physical sensor data acquisition unit that acquires physical sensor data measured in a semiconductor manufacturing device running a process according to process parameters; a simulation execution unit that executes a simulation based on a simulation model according to process parameters to produce virtual sensor data and virtual process result data; and a display control unit that visualizes the process state of the semiconductor manufacturing device during the execution of a process using the physical sensor data, virtual sensor data, and virtual process result data and displays it on a display unit.
Need to check novelty before this filing date? Find Prior Art

Description

Technology Field

[0001] The present disclosure relates to an information processing system and a simulation method. Background Technology

[0002] Process simulation is used in the field of manufacturing or research and development of semiconductor products. Process simulation can handle various physical phenomena related to a semiconductor process (hereinafter referred to as the process) using a physical model (e.g., see Patent Document 1).

[0003] For example, in process simulations, the process state during execution (e.g., wafer film attachment method, temperature state, gas state, etc.) was estimated from measurement results after the process was executed. Prior art literature

[0004] [Patent Document 1] Japanese Patent Publication No. 2018-125451 The problem to be solved

[0005] The present disclosure provides a technology for visualizing the process state of a semiconductor manufacturing device while a process is being executed in a semiconductor manufacturing device. means of solving the problem

[0006] An information processing system according to one aspect of the present disclosure is an information processing system that executes a simulation of a process state being executed in a semiconductor manufacturing device using a simulation model of the semiconductor manufacturing device, and comprises: a physical sensor data acquisition unit that acquires physical sensor data measured in the semiconductor manufacturing device being executed in the process according to process parameters; a simulation execution unit that executes a simulation according to the simulation model according to the process parameters to produce virtual sensor data and virtual process result data; and a display control unit that visualizes the process state of the semiconductor manufacturing device during the execution of the process using the physical sensor data, the virtual sensor data, and the virtual process result data and displays it on a display unit. Effects of the invention

[0007] According to the present disclosure, the process state of the semiconductor manufacturing device can be visualized while the process is being executed in the semiconductor manufacturing device. Brief explanation of the drawing

[0008] FIG. 1 is a configuration diagram of an example of an information processing system according to the present embodiment. Figure 2 is a hardware configuration diagram of an example of a computer. FIG. 3 is a functional block diagram of an example of an autonomous control controller according to the present embodiment. FIG. 4 is a functional block diagram of an example of an interpretation server according to the present embodiment. FIG. 5 is a flowchart of an example of processing of an information processing system according to the present embodiment. Figure 6 is an image of an example of a temperature evaluation screen. Figure 7 is an example drawing illustrating a screen example of a temperature display screen. Figure 8 is an image of an example of an operation screen for operating a temperature evaluation screen. Figure 9 is a flowchart of an example of the process for editing a simulation model. Figure 10 is a flowchart of an example of a process for performing fault pre-detection. Figure 11 is a flowchart of an example of a process for performing fault detection in advance. FIG. 12 is a schematic diagram of an example illustrating a temperature-related part of a semiconductor manufacturing apparatus. Figure 13 is an explanatory diagram of an example showing the process of creating and updating a temperature-related simulation model. Figure 14 is an explanatory diagram of an example showing the process of creating and updating a temperature-related simulation model. Figure 15 is an explanatory diagram of an example showing temperature categories extracted from the process of creating and updating a temperature-related simulation model. Figure 16 is an explanatory diagram of an example showing the processing of a virtual sensor update. Figure 17 is an explanatory diagram of an example showing the processing of updating a simulation model. FIG. 18 is an explanatory diagram of an example showing the processing of updating a simulation model when the specifications of a semiconductor manufacturing device of a target are changed. Figure 19 is an explanatory diagram of an example showing the processing of pre-detection of maintenance timing. FIG. 20 is an explanatory diagram of an example showing the process of improving the productivity of a semiconductor manufacturing device of a target. FIG. 21 is an explanatory diagram of an example showing the energy saving treatment of a semiconductor manufacturing device of the target. Specific details for implementing the invention

[0009] Hereinafter, a description of a form for carrying out the present invention is provided with reference to the drawings.

[0010] <System Configuration>

[0011] FIG. 1 is a configuration diagram of an example of an information processing system according to the present embodiment. The information processing system (1) shown in FIG. 1 has a semiconductor manufacturing device (10), a conventional sensor (11), an additional sensor (12), an autonomous control controller (13), a device control controller (20), a host computer (22), an external measuring device (24), an interpretation server (26), an AR server (28), a management server (30), and a data lake (32).

[0012] A semiconductor manufacturing device (10), an autonomous control controller (13), a device control controller (20), a host computer (22), an external measuring device (24), an analysis server (26), an AR server (28), and a management server (30) are connected to communicate via a network (40), such as a LAN (Local Area Network).

[0013] The semiconductor manufacturing device (10) is, for example, a heat treatment film deposition device and executes a process according to a control command (process parameter) output from the device control controller (20). The semiconductor manufacturing device (10) is equipped with a plurality of conventional sensors (11). The conventional sensors (11) are examples of physical sensors that measure temperature, pressure, etc. as physical sensor data. The additional sensors (12) are examples of physical sensors that are equipped to verify the accuracy of the virtual sensor data described later. The virtual sensor data is calculated based on a simulation model (physical model) described later. Since the additional sensors (12) are equipped to verify the accuracy of the virtual sensor data calculated based on the simulation model, they do not need to be equipped in the final product shipped to the customer. The additional sensors (12) measure temperature, pressure, etc. as physical sensor data.

[0014] The autonomous control controller (13) is a controller for autonomously controlling the semiconductor manufacturing device (10), and as described below, it performs simulation of the process state running in the semiconductor manufacturing device (10) using a simulation model, optimization of process parameters, pre-detection of failures, pre-detection of maintenance timing, etc. The autonomous control controller (13) is installed in each semiconductor manufacturing device (10). By executing a simulation using the simulation model described below, the autonomous control controller (13) calculates the state of film attachment on the wafer, the temperature of the wafer, the state of gas, plasma, etc., as a result after executing a process according to process parameters. The data (virtual data) calculated by executing a simulation using the simulation model described below includes virtual sensor data and virtual process result data. Virtual sensor data is data output from a virtual sensor. Virtual process result data is data included in the virtual data other than virtual sensor data.

[0015] The device control controller (20) is a controller having a computer configuration for controlling the semiconductor manufacturing device (10). As described below, the device control controller (20) outputs optimized process parameters to the semiconductor manufacturing device (10) as process parameters for controlling the control components of the semiconductor manufacturing device (10).

[0016] A host computer (22) is an example of a man-machine interface (MMI) that receives instructions from a worker regarding a semiconductor manufacturing device (10) and provides information about the semiconductor manufacturing device (10) to the worker.

[0017] The external measuring device (24) is a measuring device that measures the results after the execution of a process according to process parameters, such as a film thickness measuring device, a sheet resistance measuring device, or a particle measuring device. For example, the external measuring device (24) measures the attachment state of a film on a wafer, such as a monitor wafer. Hereinafter, the data measured by the external measuring device (24) is referred to as physical process result data.

[0018] As described below, the analysis server (26) performs data analysis for updating the simulation model, optimizing process parameters, and pre-detecting failures or pre-detecting maintenance times. The AR server (28) uses AR (augmented reality) technology to display the results of a simulation of the process state running in the semiconductor manufacturing device (10) or to display a screen to support tasks such as maintenance.

[0019] The management server (30) manages information regarding multiple semiconductor manufacturing devices (10) of, for example, one or more companies, which is stored in a data storage area such as a data lake (32) [process parameters for executing a process in the semiconductor manufacturing device (10), physical sensor data when the process is executed according to the process parameters, and physical process result data, etc.]. By editing a simulation model based on this information regarding multiple semiconductor manufacturing devices (10) as described below, the base simulation model can be edited (corrected) based on the behavior of the multiple semiconductor manufacturing devices (10). The base of the simulation model is a greatest common divisor model based on the multiple semiconductor manufacturing devices (10).

[0020] The base of the simulation model is edited to suit each semiconductor manufacturing device (10), for example, by deep learning. This editing is performed, for example, whenever the semiconductor manufacturing device (10) is operated, so that the prediction accuracy of the simulation model in each semiconductor manufacturing device (10) improves with each operation.

[0021] In addition, the information processing system (1) of FIG. 1 is an example, and it goes without saying that there are various system configuration examples depending on the use or purpose. The classification of devices such as the semiconductor manufacturing device (10), autonomous control controller (13), device control controller (20), host computer (22), external measuring device (24), analysis server (26), AR server (28), and management server (30) of FIG. 1 is an example.

[0022] For example, the information processing system (1) can be configured in various ways, such as having at least two of the semiconductor manufacturing device (10), autonomous control controller (13), device control controller (20), host computer (22), external measuring device (24), analysis server (26), AR server (28), and management server (30) integrated or divided. Additionally, the analysis server (26) and AR server (28) may be configured to handle multiple semiconductor manufacturing devices (10) in an integrated manner, as in the information processing system (1) of FIG. 1, or may be installed one-to-one with the semiconductor manufacturing device (10). The analysis server (26) and AR server (28) installed one-to-one with the semiconductor manufacturing device (10) can perform processing specialized for the corresponding semiconductor manufacturing device (10).

[0023] <Hardware Configuration>

[0024] The autonomous control controller (13), device control controller (20), host computer (22), interpretation server (26), AR server (28), and management server (30) of the information processing system (1) illustrated in FIG. 1 are realized by a computer with a hardware configuration such as that illustrated in FIG. 2. FIG. 2 is a hardware configuration diagram of an example of a computer.

[0025] The computer (500) of FIG. 2 is equipped with an input device (501), an output device (502), an external I / F (interface) (503), RAM (Random Access Memory) (504), ROM (Read Only Memory) (505), a CPU (Central Processing Unit) (506), a communication I / F (507), and an HDD (Hard Disk Drive) (508), each of which is interconnected by a bus (B). Additionally, the input device (501) and the output device (502) may be connected and used when necessary.

[0026] The input device (501) is a keyboard, mouse, touch panel, etc., and is used by an operator, etc. to input each operation signal. The output device (502) is a display, etc., and displays the processing result by the computer (500). The communication I / F (507) is an interface that connects the computer (500) to a network. The HDD (508) is an example of a non-volatile storage device that stores programs or data.

[0027] The external I / F (503) is an interface with an external device. The computer (500) can read and / or write to a recording medium (503a), such as an SD (Secure Digital) memory card, through the external I / F (503). The ROM (505) is an example of a non-volatile semiconductor memory (memory device) in which programs or data are stored. The RAM (504) is an example of a volatile semiconductor memory (memory device) in which programs or data are temporarily stored.

[0028] The CPU (506) is a computational device that realizes the control or function of the entire computer (500) by reading programs or data from memory devices such as ROM (505) or HDD (508) onto RAM (504) and executing processing.

[0029] The autonomous control controller (13), device control controller (20), host computer (22), analysis server (26), AR server (28), and management server (30) of FIG. 1 can realize various functions through the hardware configuration of the computer (500) of FIG. 2.

[0030] <Function Configuration>

[0031] The autonomous control controller (13) of the information processing system (1) according to the present embodiment is realized, for example, as a functional block of FIG. 3. FIG. 3 is a functional block diagram of an example of an autonomous control controller according to the present embodiment. In addition, the functional block diagram of FIG. 3 omits the illustration of configurations that are unnecessary for the description of the present embodiment.

[0032] The autonomous control controller (13) realizes a physical process result data acquisition unit (100), a physical sensor data acquisition unit (102), a process parameter acquisition unit (104), a database (106), a simulation execution unit (108), a simulation result determination unit (110), a display control unit (112), and a simulation model update unit (114) by executing a program for the autonomous control controller (13). Additionally, the simulation result determination unit (110) is configured to have a fault pre-detection unit (120), a maintenance timing pre-detection unit (122), and a process parameter adjustment unit (124).

[0033] The physical process result data acquisition unit (100) acquires physical process result data, which is the result after executing a process according to process parameters, and stores it in a database (106).

[0034] The physical sensor data acquisition unit (102) acquires physical sensor data measured by the existing sensor (11) and the additional sensor (12) and provides it to the simulation execution unit (108). The process parameter acquisition unit (104) acquires process parameters of a process running in the semiconductor manufacturing device (10) and provides them to the simulation execution unit (108). The database (106) is a data storage area that stores data used by the simulation execution unit (108) for simulation and data used by the analysis server (26) for updating the simulation model or analyzing data.

[0035] The simulation execution unit (108) can calculate the process state of the semiconductor manufacturing device (10) that is running the process by executing a simulation based on a simulation model according to the same process parameters as the semiconductor manufacturing device (10) that is running the process.

[0036] In addition, the simulation model of the semiconductor manufacturing device (10) used by the simulation execution unit (108) for the simulation may, for example, use a 1DCAE simulation model. The 1DCAE simulation model enables evaluation analysis prior to structural design (3DCAE) by representing the entire semiconductor manufacturing device (10) as a functional base in the upstream design, which is functional design.

[0037] The simulation result determination unit (110) performs fault pre-detection by the fault pre-detection unit (120) and maintenance time pre-detection by the maintenance time pre-detection unit (122) as described below by using physical process result data, physical sensor data, virtual process result data, and virtual sensor data based on the same process parameters. In addition, the simulation result determination unit (110) performs optimization of process parameters by the process parameter adjustment unit (124) as described below.

[0038] The display control unit (112) can visualize the process status of the semiconductor manufacturing device (10) in real time and display it on, for example, a host computer (22) by using virtual process result data and virtual sensor data from a simulation executed according to the same process parameters as the semiconductor manufacturing device (10) while executing a process in the semiconductor manufacturing device (10).

[0039] Accordingly, the display control unit (112) can realize a so-called digital twin that reproduces a change in physical space, called the process state of a semiconductor manufacturing device (10) that is executing a process, in cyber space with real-time connectivity. In the digital twin, while executing a process in the semiconductor manufacturing device (10), the process state of the semiconductor manufacturing device (10) can be reproduced in cyber space in real time.

[0040] By utilizing this digital twin environment, the simulation result judgment unit (110) can monitor the process status of the semiconductor manufacturing device (10), and in addition, it can detect failures in advance, detect maintenance timing in advance, and adjust process parameters as described later.

[0041] The simulation model update unit (114) updates the simulation model used by the simulation execution unit (108) for simulating the process state with the simulation model edited by the analysis server (26).

[0042] In addition, the interpretation server (26) of the information processing system (1) according to the present embodiment is realized, for example, as a functional block of FIG. 4. FIG. 4 is a functional block diagram of an example of an interpretation server according to the present embodiment. In addition, the functional block diagram of FIG. 4 omits the illustration of configurations that are unnecessary for the description of the present embodiment.

[0043] The interpretation server (26) realizes the physical data acquisition unit (140), virtual data acquisition unit (142), process parameter acquisition unit (144), simulation model memory unit (146), simulation model editing unit (148), and simulation model update request unit (150) by executing a program for the interpretation server (26).

[0044] The physical data acquisition unit (140) acquires physical sensor data and physical process result data of the semiconductor manufacturing device (10) to be analyzed from the autonomous control controller (13) or the management server (30) as physical data and provides them to the simulation model editing unit (148).

[0045] The virtual data acquisition unit (142) acquires virtual sensor data and virtual process result data of the semiconductor manufacturing device (10) to be analyzed from the autonomous control controller (13) or the management server (30) as virtual (Cyber) data and provides them to the simulation model editing unit (148).

[0046] The process parameter acquisition unit (144) acquires process parameters of the semiconductor manufacturing device (10) to be analyzed from the autonomous control controller (13) or the management server (30) and provides them to the simulation model editing unit (148).

[0047] The simulation model memory unit (146) stores a simulation model for use by the simulation execution unit (108) of the autonomous control controller (13) to simulate the process state of the semiconductor manufacturing device (10). The simulation model editing unit (148) uses the provided physical data, virtual data, and process parameters to edit the simulation model, for example using machine learning, so that the difference between the physical data and virtual data due to the same process parameters becomes smaller (to become an optimal simulation model). Furthermore, the editing of the simulation model does not necessarily have to be performed during normal operation of the semiconductor manufacturing device (10), but can be performed, for example, when a change in physical specifications is made to the semiconductor manufacturing device (10) to be simulated. The simulation model update request unit (150) requests the autonomous control controller (13) to update the edited simulation model.

[0048] <Processing>

[0049] Preparation

[0050] In the information processing system (1) according to the present embodiment, a simulation model edited by, for example, machine learning is used so that the difference between the physical data of a semiconductor manufacturing device (10) that has executed a process according to process parameters and the virtual data of a simulation that has executed the same process parameters as the semiconductor manufacturing device (10) is reduced.

[0051] By using such a simulation model, the information processing system (1) according to the present embodiment ensures the accuracy of the simulation results using the simulation model. Additionally, the editing of the simulation model may be performed based on the elapsed time since the previous editing, the number of times the process is executed, the increase in the difference between physical data and virtual data according to the same process parameters, etc.

[0052] Process Execution

[0053] FIG. 5 is a flowchart of an example of processing of an information processing system according to the present embodiment. According to process parameters output from a device control controller (20), a semiconductor manufacturing device (10) executes a process in step S10. In step S12, an autonomous control controller (13) acquires physical sensor data measured by an existing sensor (11) and an additional sensor (12) from the semiconductor manufacturing device (10) that is executing the process.

[0054] In step S14, the simulation execution unit (108) of the autonomous control controller (13) executes a simulation based on a simulation model according to the same process parameters as the semiconductor manufacturing device (10) that is executing the process, and produces virtual sensor data and virtual process result data.

[0055] In step S16, the display control unit (112) of the autonomous control controller (13) can visualize the process state of the semiconductor manufacturing device (10) that is running the process, for example as in FIG. 6, and display it on the host computer (22) by using physical sensor data of the semiconductor manufacturing device (10) that is running the process, virtual process result data from a simulation executed according to the same process parameters as the semiconductor manufacturing device (10), and virtual sensor data.

[0056] FIG. 6 is an image of an example of a temperature evaluation screen. The temperature evaluation screen (1000) of FIG. 6 is an example of a screen in which a temperature display screen (1002) of a running process and a temperature display screen (1004) that was predicted in advance before the execution of the process are displayed simultaneously.

[0057] In addition, the temperature evaluation screen (1000) of FIG. 6 is an example, and it may be a screen that simultaneously displays a temperature display screen based on the results of a simulation executed according to process parameters before optimization and a temperature display screen based on the results of a simulation executed according to process parameters after optimization. By doing so, the operator can check the degree of improvement resulting from the optimization of process parameters. In addition, the temperature evaluation screen (1000) of FIG. 6 may be a screen that simultaneously displays a temperature display screen of a running process and a predicted future temperature display screen.

[0058] The temperature display screens (1002 and 1004) of FIG. 6 display the temperature, virtual airflow, convection, etc., of each part, such as a wafer, inner tube, outer tube, inner temperature sensor, and outer temperature sensor. The temperature display screens (1002 and 1004) of FIG. 6 display the distribution of temperature or gas concentration in color. The temperature display screens (1002 and 1004) of FIG. 6 can be displayed from 360 degrees or various viewpoints. FIG. 6 illustrates, as an example, a screen example in which the temperature display screens (1002 and 1004) can display necessary locations by cutting out a circular shape. Additionally, unnecessary locations may be omitted from display.

[0059] The temperature display screens (1002 and 1004) of FIG. 6 display the temperature measurement points by physical sensors or virtual sensors as black dots. The temperature display screens (1002 and 1004) of FIG. 6 may also display the temperature at the location clicked by the operator with a mouse or the like (the location where the operator wants to know the temperature). The temperature display screens (1002 and 1004) of FIG. 6 are examples, and may display a temperature display screen in the form shown in FIG. 7, for example. FIG. 7 is an example drawing illustrating a screen example of a temperature display screen.

[0060] In addition, the operation of the temperature display screens (1002 and 1004) displayed on the temperature evaluation screen (1000) of FIG. 6 can be performed from an operation screen (1100) such as that shown in FIG. 8. FIG. 8 is an image of an example of an operation screen for operating the temperature evaluation screen.

[0061] The operation screen (1100) of FIG. 8 can read a specified file, such as a file of past data, by specifying it in the file designation field (1102) and clicking the file reading button, and can display a temperature display screen (1002 and 1004) based on the read file. Additionally, the operation screen (1100) of FIG. 8 is an example in which a past arbitrary time is specified by a frame to display a temperature display screen (1002 and 1004) based on the read file.

[0062] The operation screen (1100) of FIG. 8 may specify a file of data of a semiconductor manufacturing device (10) that is running a process in the file designation field (1102). Additionally, the operation screen (1100) of FIG. 8 illustrates an example screen including a field (1104) for displaying data as a graph, a field (1106) for loading / saving a 3D viewpoint file, and a field (1108) for setting the display / non-display of information.

[0063] Returning to step S18 of FIG. 5, the autonomous control controller (13) repeats the processing of steps S12 to S16 until the process running in the semiconductor manufacturing device (10) is terminated. When the process running in the semiconductor manufacturing device (10) is terminated, the process proceeds from step S18 to the processing of S20, and the simulation result determination unit (110) of the autonomous control controller (13) compares the physical sensor data and virtual sensor data at the same location and time to determine whether the physical sensor data and virtual sensor data at the same location and time are the same.

[0064] If it is determined that the physical sensor data and the virtual sensor data are not identical, the simulation result determination unit (110) performs the process parameter adjustment process of step S22, which optimizes the process parameters so that the result after the process execution desired by the customer can be obtained.

[0065] Additionally, the process parameter adjustment process of step S22 may stop the optimization of process parameters when, for example, the difference between physical sensor data and virtual sensor data at the same location and time exceeds a predetermined threshold, and address the issue by editing the simulation model or maintaining the semiconductor manufacturing device (10).

[0066] Proceeding to step S24, the simulation result determination unit (110) performs fault pre-detection by the fault pre-detection unit (120) and maintenance time pre-detection by the maintenance time pre-detection unit (122) as described below. If it is determined that the physical sensor data and the virtual sensor data are identical, the simulation result determination unit (110) skips the process parameter adjustment processing of step S22 and the pre-detection processing of step S24.

[0067] According to the processing of the flowchart of Fig. 5, for example, until the difference between physical sensor data and virtual sensor data at the same location and time exceeds a predetermined threshold, the process parameters can be optimized to obtain the result after the process execution desired by the customer.

[0068] Meanwhile, if the difference between the physical sensor data and the virtual sensor data at the same location and time exceeds a predetermined threshold, the system can respond by editing the simulation model or by taking measures regarding the fault or maintenance timing detected by the pre-detection processing of step S24, thereby enabling the customer to obtain the desired result after the process execution.

[0069] Editing of the simulation model is performed in the processing order as illustrated in FIG. 9, for example. FIG. 9 is a flowchart of an example of the processing for editing the simulation model. In step S30, the analysis server (26) obtains process parameters of a process executed in the semiconductor manufacturing device (10), physical data of the semiconductor manufacturing device (10) which is the result of the process according to the process parameters, and virtual data calculated based on the simulation model.

[0070] In step S32, the simulation model editing unit (148) of the interpretation server (26) determines whether the difference between the physical sensor data and the virtual sensor data at the same location and time exceeds a predetermined threshold. If it does not exceed the predetermined threshold, the simulation model editing unit (148) skips the processing of steps S34 to S36.

[0071] If a predetermined threshold is exceeded, the simulation model editing unit (148) in step S34 uses the physical data, virtual data, and process parameters acquired in step S30 to edit the simulation model using, for example, machine learning or statistical processing, so that the difference between the physical data and virtual data due to the same process parameters becomes smaller.

[0072] Proceeding to step S36, the simulation model update request unit (150) of the interpretation server (26) requests the update of the simulation model edited in step S34 from the autonomous control controller (13), thereby enabling the simulation model used by the simulation execution unit (108) of the autonomous control controller (13) to be updated.

[0073] In addition, for the base simulation model, physical data, virtual data, and process parameters of multiple semiconductor manufacturing devices (10) are acquired, and the simulation model can be edited using, for example, machine learning or statistical processing.

[0074] The fault pre-detection shown in step S24 of FIG. 5 is executed in the processing sequence shown, for example, in FIG. 10 and FIG. 11. FIG. 10 and FIG. 11 are flowcharts of an example of a process for performing fault pre-detection.

[0075] In step S40, the fault pre-detection unit (120) of the autonomous control controller (13) acquires process parameters of a process executed in the semiconductor manufacturing device (10), physical data of the semiconductor manufacturing device (10) which is the result of the process according to the process parameters, and virtual data for the next and subsequent processes calculated based on a simulation model. Additionally, virtual data for the next and subsequent processes can be acquired by executing the virtual process for the next and subsequent processes based on the simulation model. The virtual data calculated based on the simulation model in this manner may acquire not only the process for the next process but also multiple process portions considering maintenance cycles, etc. In step S42, the fault pre-detection unit (120) determines whether the difference between the physical sensor data and the virtual sensor data at the same location and time exceeds a predetermined threshold value.

[0076] If it does not exceed a predetermined threshold, the fault pre-detection unit (120) skips the fault pre-detection processing of step S44. If it exceeds a predetermined threshold, the fault pre-detection unit (120) pre-detects it as a fault. In step S44, for the pre-detected fault, the fault pre-detection processing in the order shown in FIG. 11 is performed.

[0077] In step S50 of FIG. 11, the fault pre-detection unit (120) determines whether the pre-detection fault is known. Additionally, the determination of whether the pre-detection fault is known may be made using statistical classification methods such as the type of item and feature quantity of physical data and virtual data that exceed a threshold value, or classification methods based on deep learning.

[0078] If the pre-detected fault is known, the fault pre-detection unit (120) searches for the type and cause of the fault from existing fault determination data in step S52. Proceeding to step S54, the fault pre-detection unit (120) calculates the number of processes or the period until it affects manufacturing from the searched type and cause of the fault and past fault history. Proceeding to step S56, the fault pre-detection unit (120) displays the fault pre-detection information and the fault response method, for example, on a host computer (22).

[0079] Accordingly, the operator can respond to a failure at a timing that takes into account the number of processes or the period until it affects manufacturing by checking the failure pre-detection information and failure response method displayed on the host computer (22).

[0080] Meanwhile, if the pre-detected fault is not a known fault, the fault pre-detection unit (120) calculates the trend of future deviation from the data of the deviation amounts up to the present between physical data and virtual data in step S58. Proceeding to step S60, the fault pre-detection unit (120) executes a simulation using a simulation model based on the calculated trend of future deviation amounts to calculate the number of processes or periods affecting manufacturing.

[0081] Proceeding to step S62, the fault pre-detection unit (120) displays fault pre-detection information and the number of processes or period until manufacturing is affected, calculated in step S60, on, for example, a host computer (22).

[0082] Proceeding to step S64, the fault pre-detection unit (120), after determining the cause of the fault, adds the type and cause of the fault entered by the operator as existing fault determination data. Through the processing of step S64, the type of fault whose cause has been determined is determined to be a known fault in step S50.

[0083] The process for pre-detection of failures in FIGS. 10 and 11 determines the difference between physical data and virtual data as a pre-failure state. In a simulation using a simulation model, by calculating virtual process result data that differs from physical sensor data and virtual sensor data, the impact of that difference on the process can be determined. Additionally, by accumulating and verifying the amount of deviation for each process, it is possible to determine how many processes remain until lot-out (which affects production).

[0084] In this way, the information processing system (1) according to the present embodiment can respond to unknown failure patterns by determining the difference between physical data and virtual data as the state prior to failure. In addition, the information processing system (1) according to the present embodiment can also determine the impact on the process of the location where the condition has deteriorated, and depending on the degree of impact, it is possible to operate by delaying the repair time without urgently stopping the semiconductor manufacturing device (10).

[0085] According to the information processing system (1) of the present embodiment, it is possible to detect in advance failures that cannot be predicted from past cases and failures that can be predicted from the difference between virtual sensor data predicted in advance in a simulation model and physical sensor data of a running process. In addition, according to the information processing system (1) of the present embodiment, since it is possible to detect failures in advance, it is possible to accurately determine the maintenance time when maintenance is required, and unnecessary maintenance can be suppressed, thereby reducing running costs.

[0086] For example, by performing a simulation using a simulation model of a semiconductor manufacturing device (10), the cumulative film thickness attached during the process of the semiconductor manufacturing device (10) can be calculated. As a result, the optimal cleaning recipe time for the cumulative film thickness can be calculated, thereby suppressing the consumption of expensive cleaning gas and preventing excessive cutting of film residue or quartz.

[0087] Summary

[0088] According to the information processing system (1) of the present embodiment, the state of the process, such as the attachment state of the film on the wafer of the semiconductor manufacturing device (10) that is executing the process, the temperature of the wafer, gas, plasma, etc., can be calculated by the simulation model of the semiconductor manufacturing device (10) and visualized in real time.

[0089] Accordingly, according to the information processing system (1) of the present embodiment, results equivalent to those of an external measuring device (24) can be obtained in real time without waiting for a measurement by an external measuring device (24) after the process is executed. As a result, after the process is executed, the operator can determine whether the processed wafer is normal, lot out, good, etc., without waiting for a measurement by an external measuring device (24). According to the information processing system (1) of the present embodiment, operating costs can be reduced by reducing the number of monitor wafers or the number of various measurements by an external measuring device (24). In addition, according to the information processing system (1) of the present embodiment, since the state of the process of the semiconductor manufacturing device (10) can be visualized while the process is being executed in the semiconductor manufacturing device (10), physical phenomena occurring on the wafer can be visualized to objectively identify problems and promote development.

[0090] In addition, according to the process parameter adjustment processing by the process parameter adjustment unit (124) of the information processing system (1) according to the present embodiment, the process parameters can be optimized by searching for and changing process parameters for each process based on the simulation model of the semiconductor manufacturing device (10). Furthermore, the simulation based on the simulation model of the semiconductor manufacturing device (10) can also simulate parts related to process performance, such as cumulative film thickness, thereby suppressing the influence of cumulative film thickness and ensuring that the semiconductor manufacturing device (10) can always demonstrate its full potential performance.

[0091] In addition, since the information processing system (1) according to the present embodiment enables automatic input of process parameters, setting process parameters by hand becomes unnecessary, which can reduce the number of processes and eliminate input errors or deviations caused by the operator.

[0092] In addition, the information processing system (1) according to the present embodiment can change the ratio of process-related factors to meet customer requirements by searching for and changing process parameters based on a simulation model of a semiconductor manufacturing device (10). For example, process parameters that reduce power consumption or process parameters that reduce gas consumption can be realized to meet customer requirements.

[0093] In addition, the information processing system (1) according to the present embodiment can realize the best control based on a simulation model by specifying, for example, the film thickness or film quality to be attached on the wafer and limitations (e.g., thermal budget) and by searching for and changing process parameters.

[0094] In addition, the information processing system (1) according to the present embodiment can predict the possibility of failure from the difference between virtual sensor data predicted in advance in a simulation model and physical sensor data of a running process, and can detect failure in advance or detect the timing of maintenance in advance.

[0095] [Example 1]

[0096] Below, an example of a temperature simulation model is described as an example of a 1DCAE simulation model of a semiconductor manufacturing device (10).

[0097] FIG. 12 is a schematic diagram of an example illustrating a temperature-related part of a semiconductor manufacturing apparatus. The semiconductor manufacturing apparatus (10) has a heat treatment furnace capable of accommodating multiple wafers. The heat treatment furnace is a processing vessel for accommodating and processing wafers. The heat treatment furnace is formed into a roughly cylindrical shape using quartz with high heat resistance and has an exhaust port. The heat treatment furnace is configured in a vertical shape extending in the vertical direction. A gas exhaust port is connected to the exhaust port of the heat treatment furnace.

[0098] Heaters are installed on the sides of the heat treatment furnace to surround it, and are configured to heat the wafers contained within the furnace. The bottom of the heat treatment furnace is open to allow for the loading and unloading of wafers. The opening at the bottom of the heat treatment furnace is configured to be opened and closed by a lifting mechanism.

[0099] A wafer boat is installed above the lifting mechanism. The wafer boat is a substrate holder for holding wafers and is configured to hold multiple wafers vertically in a spaced-apart manner.

[0100] For example, in the schematic diagram of FIG. 12, by creating and inputting the heater configuration, reactor configuration, and configuration around the heater as a 1DCAE simulation model, a virtual temperature sensor can be constructed based on a power value, which is an example of input data given to the heater. In addition, to verify the accuracy of the virtual temperature sensor, a physical temperature sensor included in the existing sensor (11) and additional sensor (12) is used.

[0101] For example, examples of existing sensors (11) include an inner temperature sensor, an outer temperature sensor, a heater ambient temperature sensor, a reactor internal temperature sensor, a gas supply pipe heater temperature sensor, a gas exhaust pipe heater temperature sensor, and a lifting mechanism heater temperature sensor. Additionally, examples of additional sensors (12) include a temperature sensor installed on the wafer surface, a temperature sensor installed on the heating pipe and reactor side lifting mechanism surface, and a gas introduction temperature sensor. The additional sensors (12) are not installed in the final product shipped to the customer. The existing sensors (11) are installed in the final product shipped to the customer.

[0102] For example, in the creation and updating of a temperature-related simulation model, processing is performed as shown in FIG. 13. FIG. 13 is an explanatory diagram of an example illustrating the processing of the creation and updating of a temperature-related simulation model.

[0103] The simulation execution unit (108) of the autonomous control controller (13) acquires, as an example of process parameters of a process running in the semiconductor manufacturing device (10), such as the power value of a heater, the setting value of a blower, the gas flow rate and temperature, and the wafer boat lifting position. The simulation execution unit (108) outputs virtual temperature sensor data and virtual process result data by executing a simulation based on the simulation model of the semiconductor manufacturing device (10) according to the process parameters.

[0104] Additionally, the physical temperature sensor, which is the existing sensor (11) and the additional sensor (12), measures the temperature of the measurement point and outputs it as physical temperature sensor data of the measurement point. The database (106) stores the outputted virtual temperature sensor data, virtual process result data, physical temperature sensor data of the measurement point, and previous physical temperature sensor data.

[0105] The simulation result determination unit (110) compares the virtual temperature sensor data and physical temperature sensor data stored in the database (106) for each measurement point and determines whether the virtual temperature sensor data and physical temperature sensor data of the same measurement point stored in the database (106) are the same.

[0106] If it is determined that they are identical, the simulation result determination unit (110) determines that the virtual temperature sensor data is accurate. If it is determined that they are not identical, the simulation result determination unit (110) outputs the physical temperature sensor data and the virtual temperature sensor data for recording.

[0107] Afterwards, the simulation model of the simulation execution unit (108) is updated by manually editing the simulation model while performing data verification, for example, offline.

[0108] In this way, in the creation and updating of the simulation model, the simulation model of the simulation execution unit (108) can be updated as needed while performing data verification offline. Additionally, in the creation and updating of the simulation model, as long as no physical specification change is made to the target semiconductor manufacturing device (10), virtual data and physical data may be compared, and process parameters may be edited according to an algorithm, thereby performing an operation that produces maximum performance for the input data of the specifications.

[0109] In addition, for example, in the creation and updating of a temperature-related simulation model, processing as shown in FIG. 14 may be performed. FIG. 14 is an explanatory diagram of an example showing the processing of the creation and updating of a temperature-related simulation model. For example, in the normal operation phase (during normal operation), by performing processing as shown in FIG. 14, an example is shown that operation is possible in terms of adjusting process parameters as a method other than updating the simulation model.

[0110] The simulation execution unit (108) of the autonomous control controller (13) acquires process parameters of a process running in a semiconductor manufacturing device (10). The simulation execution unit (108) outputs virtual sensor data and virtual process result data by executing a simulation based on a simulation model of the semiconductor manufacturing device (10) according to the acquired process parameters.

[0111] The existing sensor (11) measures a measurement point and outputs it as physical sensor data of the measurement point. The database (106) stores the outputted virtual sensor data, virtual process result data, and physical sensor data of the existing sensor (11). Additionally, the database (106) stores virtual sensor data of the additional sensor (12) previously simulated and virtual sensor data of the additional sensor (12) generated as a result of statistical analysis.

[0112] The simulation result determination unit (110) compares the virtual sensor data with the physical sensor data of the existing sensor (11) to determine whether the virtual sensor data and the physical sensor data of the same measurement point stored in the database (106) are identical. Additionally, the simulation result determination unit (110) may determine whether the virtual sensor data of the same measurement point is identical by using the virtual sensor data of the additional sensor (12) previously simulated and the virtual sensor data of the additional sensor (12) generated as a result of statistical analysis as comparison targets.

[0113] If it is determined that they are not identical, the simulation result determination unit (110) adjusts the process parameters according to the algorithm. If the amount of adjustment of the adjusted process parameters does not exceed a predetermined threshold, the simulation execution unit (108) updates the process parameters with the adjusted process parameters and continues the simulation by the simulation model of the semiconductor manufacturing device (10).

[0114] Additionally, if the amount of adjustment of the process parameter exceeds a threshold value, the simulation execution unit (108) stops the optimization by adjusting the process parameter. If it is determined that the virtual sensor data and physical sensor data of the same measurement point stored in the database (106) are identical, the simulation result determination unit (110) determines that the adjustment of the process parameter is unnecessary and terminates the process without adjusting the process parameter.

[0115] After the process is finished, physical process result data such as film thickness, sheet resistance, etching rate, and particle map data is input from the customer, and is stored in the database (106) as virtual process result data.

[0116] In the normal operation phase where the processing shown in FIG. 14 is performed, virtual data and physical data are compared, and process parameters are edited according to an algorithm, thereby performing an operation that produces maximum performance for the input data of the specifications. In addition, in the normal operation phase where the processing shown in FIG. 14 is performed, a delay of 1 RUN occurs, but as the data stored in the database (106) increases, the precision of machine learning of the loop indicated by the dashed arrow is improved, and the precision of predictive control, such as the adjustment of optimal process parameters, is improved with each repeated RUN.

[0117] FIG. 15 shows temperature categories extracted from the process of creating and updating a temperature-related simulation model shown in FIG. 14. The simulation execution unit (108) of the autonomous control controller (13) acquires process parameters of a process running in a semiconductor manufacturing device (10). The simulation execution unit (108) outputs virtual temperature sensor data and virtual process result data by executing a simulation based on the simulation model of the semiconductor manufacturing device (10) according to the acquired process parameters.

[0118] Additionally, the existing sensor (11) measures the temperature of the measurement point and outputs it as physical temperature sensor data of the measurement point. The database (106) stores the outputted virtual temperature sensor data, virtual process result data, and the physical temperature sensor data of the existing sensor (11).

[0119] Additionally, the database (106) stores previously simulated virtual temperature sensor data. The simulation result determination unit (110) compares the virtual temperature sensor data stored in the database (106) with the physical temperature sensor data of the existing sensor (11) to determine whether the virtual temperature sensor data and the physical temperature sensor data of the same measurement point stored in the database (106) are identical.

[0120] Additionally, the simulation result determination unit (110) may determine whether the virtual temperature sensor data of the same measurement point is the same by comparing the virtual sensor data of the additional sensor (12) previously simulated and the virtual sensor data of the additional sensor (12) generated as a result of statistical analysis.

[0121] If it is determined that they are not identical, the simulation result determination unit (110) adjusts the process parameters according to the algorithm. The simulation execution unit (108) executes a simulation based on the simulation model of the semiconductor manufacturing device (10) according to the adjusted process parameters (temperature parameters). Additionally, if it is determined that they are identical, the simulation result determination unit (110) determines that adjustment of the process parameters is unnecessary and does not adjust the process parameters.

[0122] In the processing illustrated in FIG. 15, optimal process parameters are edited according to an algorithm based on the result of comparing virtual temperature sensor data and physical temperature sensor data. Additionally, the virtual temperature sensor does not necessarily have to correspond one-to-one with the existing sensor (11) and the additional sensor (12). For example, the virtual temperature sensor may be a virtual sensor that outputs virtual temperature sensor data that can be calculated from the physical sensor data of a plurality of existing sensors (11) and additional sensors (12). The virtual temperature sensor is a temperature distribution sensor within a reactor, a temperature distribution sensor within a wafer surface, a temperature distribution sensor between wafer surfaces, or a supply gas temperature sensor, etc.

[0123] In addition, examples of virtual sensors include wafer in-plane / inter-plane temperature sensors, gas pressure distribution sensors within a reactor, gas concentration distribution sensors within a reactor, plasma distribution sensors, gas flow velocity and flow direction sensors within a reactor, supply gas temperature sensors, dew point sensors calculated from thermometers and hygrometers, sensors that predict the total airflow within a reactor using a very small number of measuring instruments, sensors that predict the oxygen concentration distribution within a reactor using a very small number of measuring instruments, sensors for measuring abnormal points due to vibration, and sensors for material degradation due to vibration.

[0124] Optimization of the virtual sensor is performed, for example, as illustrated in FIG. 16. FIG. 16 is an explanatory diagram of an example showing the processing of updating the virtual sensor. FIG. 16 shows an example of a virtual gas concentration sensor, which is an example of a virtual sensor.

[0125] The simulation execution unit (108) of the autonomous control controller (13) acquires process parameters of a process running in a semiconductor manufacturing device (10). The simulation execution unit (108) outputs virtual sensor data and virtual process result data by executing a simulation based on a simulation model of the semiconductor manufacturing device (10) according to the acquired process parameters.

[0126] In addition, a physical temperature sensor and a physical pressure sensor, which are examples of existing sensors (11), measure a measurement point and output the physical temperature sensor data and physical pressure sensor data of the measurement point. The database (106) stores the outputted virtual sensor data, virtual process result data, and physical sensor data of the existing sensors (11). In addition, the database (106) stores previously simulated virtual sensor data and virtual sensor data generated from statistical analysis results.

[0127] The simulation result determination unit (110) compares the virtual sensor data with the physical sensor data of the existing sensor (11) to determine whether the virtual sensor data and the physical sensor data of the same measurement point stored in the database (106) are identical. Additionally, the simulation result determination unit (110) may determine whether the virtual sensor data of the same measurement point is identical by using the virtual sensor data of the additional sensor (12) previously simulated and the virtual sensor data of the additional sensor (12) generated as a result of statistical analysis as comparison targets.

[0128] If it is determined that they are not identical, the simulation result determination unit (110) adjusts the process parameter (gas concentration parameter) according to the algorithm. The simulation execution unit (108) updates the adjusted gas concentration parameter and continues the simulation by the simulation model of the semiconductor manufacturing device (10).

[0129] If it is determined that the virtual sensor data and physical sensor data of the same measurement point stored in the database (106) are identical, the simulation result determination unit (110) determines that adjustment of the gas concentration parameter is unnecessary and terminates the optimization of the virtual sensor.

[0130] In addition, physical phenomena including virtual sensor operation in virtual space are determined by a simulation model. When the specifications of the target semiconductor manufacturing device (10) are changed, the simulation model is basically updated as well. Therefore, when the specifications of the target semiconductor manufacturing device (10) are changed, process parameters related to the virtual sensor are optimized as shown in FIG. 16. FIG. 16 also utilizes virtual sensor data obtained from statistical analysis results.

[0131] In FIG. 16, based on the process results and sensor data related to temperature and pressure, the update of gas concentration parameters is automatically performed according to a machine learning algorithm based on the comparison results between physical data and virtual data. When the physical data and virtual data match, the construction of a virtual sensor regarding the gas concentration of the target semiconductor manufacturing device (10) is completed.

[0132] Optimization of the simulation model can be performed by commissioning in a virtual space, for example, as illustrated in FIG. 17. FIG. 17 is an explanatory diagram of an example showing the processing of updating the simulation model.

[0133] In the processing of FIG. 17, optimization of the simulation model is performed using the simulation model and virtual sensors, and optimization of process parameters is performed in a virtual space based on physical process result data that has already been acquired. In FIG. 17, optimization of the simulation model and optimization of process parameters are performed according to a machine learning algorithm with training data. In addition, the number of commissioning operations in the virtual space may be specified by a setting.

[0134] In addition, updating the simulation model is required when the specifications of the target semiconductor manufacturing device (10) are changed. FIG. 18 is an explanatory diagram of an example showing the processing of updating the simulation model when the specifications of the target semiconductor manufacturing device are changed.

[0135] In the processing of FIG. 18, when the specifications of the target semiconductor manufacturing device (10) are changed, if the simulation model of the target semiconductor manufacturing device (10) is already stored in the simulation model library, the existing simulation model is reused.

[0136] In addition, if a simulation model of the target semiconductor manufacturing device (10) is not already stored in the simulation model library, a new simulation model is created. Verification of the created simulation model is performed based on evaluation process parameters, virtual data resulting from running a simulation according to the evaluation process parameters, and physical data that has already been identified. Furthermore, the newly created simulation model may be installed and registered in the library after its accuracy is confirmed by executing a process using the target semiconductor manufacturing device (10).

[0137] In addition, updating the simulation model is essential when the process specifications change. The process of updating the simulation model when the process specifications change is also the same as Fig. 18, except for some parts.

[0138] For example, in the process of updating a simulation model when the specifications of a process are changed, the verification of the created simulation model is performed based on physical data that has already been clarified. In addition, the newly created simulation model may be installed and registered in a library after its accuracy is confirmed by executing a process using the target semiconductor manufacturing device (10).

[0139] Pre-detection of the maintenance time is performed, for example, as illustrated in FIG. 19. FIG. 19 is an explanatory diagram of an example showing the processing of pre-detection of the maintenance time. The simulation execution unit (108) of the autonomous control controller (13) acquires process parameters of a process running in the semiconductor manufacturing device (10).

[0140] The simulation execution unit (108) outputs virtual sensor data and virtual process result data by executing a simulation based on a simulation model of a semiconductor manufacturing device (10) according to the acquired process parameters.

[0141] The simulation result determination unit (110) compares the virtual sensor data with the physical sensor data of the existing sensor (11) to determine whether the virtual sensor data and the physical sensor data of the same measurement point stored in the database (106) are identical. Additionally, the simulation result determination unit (110) may determine whether the virtual sensor data of the same measurement point is identical by using the virtual sensor data of the additional sensor (12) previously simulated and the virtual sensor data of the additional sensor (12) generated as a result of statistical analysis as comparison targets.

[0142] If it is determined that they are identical, the simulation result determination unit (110) determines that there is no abnormality and terminates the processing. If it is determined that they are not identical, the simulation result determination unit (110) determines whether the difference between the virtual sensor data and the physical sensor data compared exceeds a threshold value.

[0143] If the threshold is not exceeded, the simulation result determination unit (110) performs a life prediction according to a life prediction algorithm and determines whether an alert is required. Additionally, the life prediction algorithm is generated by machine learning or statistical analysis software, etc. If it is determined that an alert is not required, the simulation result determination unit (110) terminates the processing.

[0144] If it is determined that an alarm is required, the simulation result determination unit (110) initiates processing of a pre-maintenance warning. Additionally, if the difference between the virtual sensor data and the physical sensor data compared exceeds a threshold value, the simulation result determination unit (110) initiates processing of a pre-maintenance warning.

[0145] In processing the advance warning for maintenance timing, the parts or modules requiring a warning are identified by the simulation result determination unit (110). The simulation result determination unit (110) identifies preventive maintenance (PM) information for the specific parts or modules requiring a warning from product data or part information, etc. Based on the identified preventive maintenance information, the simulation result determination unit (110) notifies the maintenance timing of the specific parts or modules requiring a warning.

[0146] [Example 2]

[0147] Example 2 utilizes the structure of Example 1 to improve the productivity of the target semiconductor manufacturing device (10), save energy, and achieve a balance between improved productivity and preventive maintenance detection.

[0148] FIG. 20 is an explanatory diagram showing an example of processing for improving the productivity of a semiconductor manufacturing device. FIG. 20 illustrates an example of reducing the time required for wafer transfer as an example of processing for improving the productivity of a semiconductor manufacturing device (10).

[0149] In the semiconductor manufacturing device (10) of the target, for example, to reduce the wafer transfer time by 1% every day, if there are mechanical parts capable of speeding up, the process parameters are optimized to the extent possible to improve productivity. In addition, the limit of productivity improvement can be determined, for example, by determining the limit by a simulation model or by an alarm based on preventive maintenance information.

[0150] For example, in FIG. 20, mechanical parts with a margin for speed-up are searched from the mechanical parts of the target semiconductor manufacturing device (10), and speed-up is sought to the limit of productivity improvement. In this embodiment, since speed-up to the limit of productivity improvement can be sought by, for example, by determining the limit by a simulation model or by an alarm based on preventive maintenance information, it is possible to significantly improve productivity. In addition, it is desirable to make the degree of improvement for productivity improvement (maximization) selectable.

[0151] In addition, in this embodiment, productivity can be improved while achieving maximum operating rate of the target semiconductor manufacturing device (10) by, for example, determining limits by a simulation model or by an alarm based on preventive maintenance information.

[0152] FIG. 21 is an explanatory diagram of an example showing the energy saving process of a semiconductor manufacturing device. In the semiconductor manufacturing device (10) of the target, in order to realize energy saving by reducing the consumption of gas or electricity, the process parameters are optimized to the extent possible.

[0153] For example, in FIG. 21, the process parameters of the target semiconductor manufacturing device (10) are adjusted within a range where the process parameters do not exceed a predetermined threshold value, thereby reducing the consumption of gas or electricity and realizing energy saving. In this way, in this embodiment, energy saving can be realized while achieving maximum operating rate of the target semiconductor manufacturing device (10).

[0154] In addition, by combining the processing of the pre-detection of the maintenance timing of FIG. 19 with the processing of the productivity improvement of the target semiconductor manufacturing device of FIG. 20, it is possible to achieve a balance between productivity improvement and preventive maintenance detection in this embodiment. When FIG. 19 and FIG. 20 are combined, the process parameter optimized for speed-up in FIG. 20 is used as the process parameter of the simulation execution unit (108) of FIG. 19.

[0155] According to the present embodiment, the optimization of process parameters that were previously performed based on the operator's experience or intuition can be automated, and the process status of a process running in a semiconductor manufacturing device (10) can be visualized.

[0156] Although preferred embodiments of the present invention have been described in detail above, the present invention is not limited to the aforementioned embodiments, and various modifications and substitutions can be made to the aforementioned embodiments without departing from the scope of the present invention. For example, in this embodiment, a heat treatment film deposition apparatus was described as an example, but it can also be applied to batch film deposition apparatus such as CVD (Chemical Vapor Deposition), thermal oxidation, and ALD (Atomic Layer Deposition).

Claims

Claim 1 An information processing system that executes a simulation of a process state currently running in a semiconductor manufacturing device using a simulation model of the semiconductor manufacturing device, comprising: a physical sensor data acquisition unit that acquires physical sensor data measured in the semiconductor manufacturing device running the process according to process parameters; a simulation execution unit that executes a simulation based on the simulation model according to the process parameters to produce virtual sensor data and virtual process result data; a simulation result determination unit that compares the physical sensor data and the virtual sensor data at the same location and time to determine whether the physical sensor data and the virtual sensor data at the same location and time are identical; and a display control unit that visualizes the process state of the semiconductor manufacturing device during the execution of the process using the physical sensor data, the virtual sensor data, and the virtual process result data and displays it on a display unit, wherein the display control unit visualizes 1) a predicted process state of the semiconductor manufacturing device based on the virtual sensor data and the virtual process result data produced by the simulation execution unit, and 2) an actual process state of the semiconductor manufacturing device based on the physical sensor data and the physical process result data, simultaneously during the execution of the process, on the display unit as a view of a part of the semiconductor manufacturing device, and each of the views An information processing system for simulating a semiconductor manufacturing device, comprising, after executing the process according to the process parameters, a form in which corresponding measurement points for the virtual process result data and the physical process result data can be compared with each other. Claim 2 An information processing system for simulating a semiconductor manufacturing device, wherein, in claim 1, the simulation result determination unit further comprises a process parameter adjustment unit that adjusts the process parameters so that the physical sensor data and the virtual sensor data approximate each other. Claim 3 An information processing system for simulating a semiconductor manufacturing device, further comprising a simulation model editing unit that creates or updates the simulation model such that, in claim 1 or 2, the physical process result data after executing a process according to the process parameters is approximate to the virtual process result data calculated by the simulation execution unit, and the physical sensor data is approximate to the virtual sensor data. Claim 4 An information processing system for simulating a semiconductor manufacturing device, wherein, in paragraph 3, the simulation model editing unit creates or updates the simulation model using the physical process result data, virtual process result data, physical sensor data, and virtual sensor data of a plurality of the semiconductor manufacturing devices. Claim 5 An information processing system for simulating a semiconductor manufacturing device, wherein, in claim 1 or 2, the simulation result determination unit further comprises a fault pre-detection unit that performs fault pre-detection of the semiconductor manufacturing device based on at least one of the difference between physical process result data after executing a process according to the process parameters and the virtual process result data calculated by the simulation execution unit, or the difference between the physical sensor data and the virtual sensor data. Claim 6 An information processing system for simulating a semiconductor manufacturing device, wherein, in claim 5, the fault pre-detection unit monitors the difference between the physical process result data and the virtual process result data, or the difference between the physical sensor data and the virtual sensor data, and performs fault pre-detection of the semiconductor manufacturing device using a fault pattern learned from the difference between the physical process result data and the virtual process result data, or the difference between the physical sensor data and the virtual sensor data, at the time of a known fault. Claim 7 An information processing system for simulating a semiconductor manufacturing device, wherein, in paragraph 2, the process parameter adjustment unit determines the process parameter that becomes the virtual process result data approximated by the physical process result data so as to become the physical process result data specified by the user. Claim 8 In paragraph 3, the information processing system for simulating a semiconductor manufacturing device further comprises a fault pre-detection unit that performs fault pre-detection of the semiconductor manufacturing device based on at least one of the difference between physical process result data after executing a process according to the process parameters and the virtual process result data calculated by the simulation execution unit, or the difference between the physical sensor data and the virtual sensor data. Claim 9 In claim 4, the information processing system for simulating a semiconductor manufacturing device further comprises a fault pre-detection unit that performs fault pre-detection of the semiconductor manufacturing device based on at least one of the difference between physical process result data after executing a process according to the process parameters and the virtual process result data calculated by the simulation execution unit, or the difference between the physical sensor data and the virtual sensor data. Claim 10 An information processing system, as a simulation method for executing a simulation of a process state running in a semiconductor manufacturing device using a simulation model of the semiconductor manufacturing device, comprises: a process of acquiring physical sensor data measured in the semiconductor manufacturing device running the process according to process parameters; a process of generating virtual sensor data and virtual process result data by executing a simulation by the simulation model according to the process parameters; a process of determining whether the physical sensor data and the virtual sensor data at the same location and time are identical by comparing the physical sensor data and the virtual sensor data at the same location and time; and a process of visualizing the process state of the semiconductor manufacturing device and displaying it on a display unit during the execution of the process using the physical sensor data, the virtual sensor data, and the virtual process result data. The displaying process includes: 1) a predicted process state of the semiconductor manufacturing device based on the virtual sensor data and the virtual process result data, and 2) an actual process state of the semiconductor manufacturing device based on the physical sensor data and the physical process result data, simultaneously during the execution of the process as a view of a part of the semiconductor manufacturing device. Each of the views is, after executing the process according to the process parameters, the virtual process A method for simulating a semiconductor manufacturing device, comprising including result data and corresponding measurement points for the physical process result data in a form that can be compared with each other.

Citation Information

Patent Citations

  • Methods, devices and systems for monitoring, controlling and optimizing processes

    US20020156542A1

  • System and method for using first-principles simulation to characterize a semiconductor manufacturing process

    US8014991B2