Substrate processing apparatus, method of manufacturing semiconductor device, and computer-readable recording medium

Abstract

The present application provides the following structure: acquire sensor data related to a component as an abnormality precursor detection object to generate a normal model, and monitor the state of the device based on the normal model, after maintenance of the component as an abnormality precursor detection object, acquire sensor data and generate a normal model again from the sensor data, and monitor the state of the device based on the normal model to detect a precursor of an abnormality before the device stops abnormally.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to a substrate processing apparatus, a manufacturing method of a semiconductor device, and a precursor detection program. BACKGROUND

[0002] Generally, a substrate processing apparatus that manufactures a semiconductor device by forming a thin film on a substrate such as a wafer is configured with various components such as a vacuum pump that performs vacuum exhaust on a processing chamber, a mass flow controller that controls the flow rate of a reactive gas or the like, an on-off valve, a pressure gauge, a heater that heats the processing chamber, and a conveyance mechanism that conveys the substrate.

[0003] Each of the various components gradually deteriorates and fails with use, and thus needs to be replaced with a new component. As a method of replacement, either of a method in which replacement is performed with a margin before failure occurs, or a method in which a regular replacement period is determined for each component is sometimes employed. In the case of using a component until failure occurs, in this case, all of the substrates processed by the substrate processing apparatus at the time of failure become defective products, and the substrates and the production time at the time of failure sometimes become a loss. In addition, in the case of regular replacement before failure, replacement needs to be performed during a period until failure, that is, every short period, and thus the frequency of replacement of the components increases, and sometimes leads to an increase in the cost of operation.

[0004] In addition, various technologies related to the maintenance of these components are proposed as in Patent Literature 1 or Patent Literature 2, but there still remains a case where an abnormality of a component cannot be detected in advance.

[0005] PRIOR ART DOCUMENTS

[0006] PATENT LITERATURE

[0007] Patent Literature 1: International Publication No. 2016-157402

[0008] Patent Literature 2: International Publication No. 2017-158682 SUMMARY

[0009] PROBLEMS TO BE SOLVED BY THE INVENTION

[0010] An object of the present disclosure is to provide a structure capable of detecting a precursor of an abnormality of a component.

[0011] MEANS FOR SOLVING THE PROBLEMS

[0012] According to one embodiment of the present disclosure, there is provided a structure in which sensor data related to a component that is an object of abnormality precursor detection is acquired to generate a normal model, a state of an apparatus is monitored on the basis of the normal model, after replacement or maintenance of the component that is the object of abnormality precursor detection, the sensor data is acquired and the normal model is generated again on the basis of the sensor data, and a precursor of an abnormality is detected before the apparatus abnormally stops on the basis of the normal model.

[0013] Effects of Invention

[0014] According to the present disclosure, there is provided a technology capable of detecting a precursor of an abnormality of a component. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 is a perspective view showing the schematic structure of a substrate processing apparatus according to one embodiment.

[0016] Figure 2 is a longitudinal sectional view showing the schematic structure of a processing furnace of a substrate processing apparatus according to one embodiment.

[0017] Figure 3 is a block diagram showing the schematic structure of a main control section of a substrate processing apparatus according to one embodiment.

[0018] Figure 4 is a flowchart showing a substrate processing procedure when a substrate processing apparatus according to one embodiment is used as a semiconductor manufacturing apparatus.

[0019] Figure 5 is a block diagram showing a control system of a substrate processing apparatus according to one embodiment.

[0020] Figure 6 is an explanatory diagram of singular spectrum transformation in a control system of a substrate processing apparatus according to one embodiment.

[0021] Figure 7 is a flowchart showing a part of a procedure of a precursor detection process according to a specific example of the third embodiment.

[0022] Figure 8 is a flowchart showing a part of a procedure of a precursor detection process according to a specific example of the fourth embodiment. DETAILED DESCRIPTION

[0023] Hereinafter, a method of manufacturing a semiconductor apparatus, a precursor detection program, and a substrate processing apparatus according to one embodiment of the present disclosure will be described. Furthermore, in the following description, the front direction of a substrate processing apparatus is indicated by an arrow F, the rear direction is indicated by an arrow B, the right direction is indicated by an arrow R, the left direction is indicated by an arrow L, the upper direction is indicated by an arrow U, and the lower direction is indicated by an arrow D. Figure 1 ​

[0024] <Overall configuration of processing apparatus>

[0025] While referring to Figure 1 , Figure 2 , the configuration of the substrate processing apparatus 10 will be described. As shown in Figure 1 , the substrate processing apparatus 10 is provided with a housing 12 composed of a pressure-resistant container. An opening portion provided in a maintainable manner is formed in the front wall of the housing 12, and a pair of front maintenance doors 14 is provided as an access mechanism that opens and closes the opening portion. Further, in this substrate processing apparatus 10, a wafer cassette (substrate container) 18 that houses a substrate (wafer) 16 of silicon or the like described later (see Figure 2 ) is used as a carrier that transports the substrate 16 into and out of the housing 12.

[0026] In the front wall of the housing 12, a wafer cassette loading / unloading port is formed in a manner that communicates the inside and outside of the housing 12. A load port 20 is provided in the wafer cassette loading / unloading port. The load port 20 is configured to load the wafer cassette 18 thereon and to align the wafer cassette 18.

[0027] A rotary wafer cassette rack 22 is provided in the upper portion of the substantially central portion of the housing 12. The rotary wafer cassette rack 22 is configured to store a plurality of wafer cassettes 18 thereon. The rotary wafer cassette rack 22 is provided with a support that is vertically erected and rotates in the horizontal plane, and a plurality of shelves that are supported in a radial manner at each of the upper, middle, and lower stages of the support.

[0028] A wafer cassette transport apparatus 24 is provided between the load port 20 and the rotary wafer cassette rack 22 in the housing 12. The wafer cassette transport apparatus 24 has a wafer cassette elevator 24A that can be raised and lowered while holding the wafer cassette 18, and a wafer cassette transport mechanism 24B. By the continuous operation of the wafer cassette elevator 24A and the wafer cassette transport mechanism 24B, the wafer cassette 18 is transported between the load port 20, the rotary wafer cassette rack 22, and a wafer cassette opener 26 described later.

[0029] In the lower portion of the housing 12, a sub-housing 28 is provided from the substantially central portion to the rear end of the housing 12. A pair of wafer cassette openers 26 that transport the substrate 16 into and out of the sub-housing 28 is provided in the front wall of the sub-housing 28.

[0030] Each wafer cassette opener 26 is provided with a placement table that places the wafer cassette 18, and a lid attachment / detachment mechanism 30 that attaches and detaches the lid of the wafer cassette 18. The wafer cassette opener 26 is configured to open and close the substrate inlet / outlet of the wafer cassette 18 by attaching and detaching the lid of the wafer cassette 18 placed on the placement table by the lid attachment / detachment mechanism 30.

[0031] A transfer chamber 32 is formed within the sub-housing 28, which is fluidly isolated from the space where the wafer cassette transport device 24 and the rotary wafer cassette holder 22 are located. A substrate transfer mechanism 34 is provided in the front region of the transfer chamber 32. The substrate transfer mechanism 34 consists of a substrate transfer device 34A capable of rotating or moving the substrate 16 in a horizontal direction and a substrate transfer device elevator 34B capable of raising and lowering the substrate transfer device 34A.

[0032] The substrate transfer device elevator 34B is located between the right end of the front region of the transfer chamber 32 of the sub-housing 28 and the right end of the housing 12. Additionally, the substrate transfer device 34A includes a clamp (not shown) that serves as a holder for the substrate 16. Through the continuous operation of these substrate transfer device elevators 34B and substrate transfer device 34A, the substrate 16 can be charged and discharging relative to the crystal boat 36, which serves as a substrate holder.

[0033] like Figure 2 As shown, a crystal boat lift 38 is provided inside the sub-shell 28 (transfer chamber 32) to raise and lower the crystal boat 36. The lifting platform of the crystal boat lift 38 is connected to the arm 40, and a cover 42 is horizontally mounted on the arm 40. The cover 42 vertically supports the crystal boat 36 and is configured to close the lower end of the processing furnace 44, which will be described later.

[0034] Mainly composed of Figure 1 The rotary wafer cassette holder 22, wafer cassette conveyor 24, substrate transfer mechanism 34, and wafer boat 36 are shown. Figure 2 The crystal boat elevator 38 shown and the rotating mechanism 46 described later constitute the conveying mechanism for the conveying substrate 16.

[0035] like Figure 1 As shown, a processing furnace 44 is provided above the standby section 50, which houses the crystal boat 36 and keeps it in standby mode. Additionally, a cleaning unit 52 is provided on the left end of the transfer chamber 32, opposite to the side of the substrate transfer device elevator 34B. The cleaning unit 52 is configured to supply a cleaning atmosphere or clean air 52A as an inert gas.

[0036] Furthermore, on the outer periphery of the housing 12 and the sub-housing 28, a plurality of device covers (not shown) are installed as entry mechanisms into the substrate processing apparatus 10. At the ends of the housing 12 and the sub-housing 28 opposite to these device covers, door switches 54 (only the door switch 54 of the housing 12 is shown) are provided as entry sensors.

[0037] Further, a substrate detection sensor 56 that detects placement of the wafer boat 18 on the load port 20 is provided. These switches, sensors, and the like, such as the door switches 54 and the substrate detection sensor 56, are electrically connected to a substrate processing apparatus controller 58 (see FIG. 2) that will be described later as a main control unit. Figure 2 , Figure 3 ).

[0038] As shown in FIG. 1, the substrate processing apparatus 10 includes a housing 12, a load lock chamber 14, a transfer chamber 16, a load lock chamber 18, a transfer chamber 20, and a processing furnace 44. The load lock chamber 14, the transfer chamber 16, the load lock chamber 18, the transfer chamber 20, and the processing furnace 44 are provided in the housing 12. Figure 2 As shown in FIG. 1, the substrate processing apparatus 10 includes a housing 12, a load lock chamber 14, a transfer chamber 16, a load lock chamber 18, a transfer chamber 20, and a processing furnace 44. The load lock chamber 14, the transfer chamber 16, the load lock chamber 18, the transfer chamber 20, and the processing furnace 44 are provided in the housing 12.

[0039] As shown in FIG. 1, the substrate processing apparatus 10 includes a housing 12, a load lock chamber 14, a transfer chamber 16, a load lock chamber 18, a transfer chamber 20, and a processing furnace 44. The load lock chamber 14, the transfer chamber 16, the load lock chamber 18, the transfer chamber 20, and the processing furnace 44 are provided in the housing 12.

[0040] As shown in FIG. 1, the substrate processing apparatus 10 includes a housing 12, a load lock chamber 14, a transfer chamber 16, a load lock chamber 18, a transfer chamber 20, and a processing furnace 44. The load lock chamber 14, the transfer chamber 16, the load lock chamber 18, the transfer chamber 20, and the processing furnace 44 are provided in the housing 12. Figure 2 As shown in FIG. 1, the substrate processing apparatus 10 includes a housing 12, a load lock chamber 14, a transfer chamber 16, a load lock chamber 18, a transfer chamber 20, and a processing furnace 44. The load lock chamber 14, the transfer chamber 16, the load lock chamber 18, the transfer chamber 20, and the processing furnace 44 are provided in the housing 12. Figure 5 As shown in FIG. 1, the substrate processing apparatus 10 includes a housing 12, a load lock chamber 14, a transfer chamber 16, a load lock chamber 18, a transfer chamber 20, and a processing furnace 44. The load lock chamber 14, the transfer chamber 16, the load lock chamber 18, the transfer chamber 20, and the processing furnace 44 are provided in the housing 12.

[0041] <Structure of processing furnace>

[0042] As shown in FIG. 1, the substrate processing apparatus 10 includes a housing 12, a load lock chamber 14, a transfer chamber 16, a load lock chamber 18, a transfer chamber 20, and a processing furnace 44. The load lock chamber 14, the transfer chamber 16, the load lock chamber 18, the transfer chamber 20, and the processing furnace 44 are provided in the housing 12. Figure 2 As shown in FIG. 1, the substrate processing apparatus 10 includes a housing 12, a load lock chamber 14, a transfer chamber 16, a load lock chamber 18, a transfer chamber 20, and a processing furnace 44. The load lock chamber 14, the transfer chamber 16, the load lock chamber 18, the transfer chamber 20, and the processing furnace 44 are provided in the housing 12.

[0043] A cylindrical heater 88 is provided on the outside of the reaction tube 84 so as to surround the side surface of the reaction tube 84. The heater 88 is vertically installed by being supported on a heater base 90.

[0044] Below the outer reaction tube 84B, a cylindrical furnace mouth portion (manifold) 92 is arranged in a concentric manner with the outer reaction tube 84B. The furnace mouth portion 92 is arranged so as to support the lower end portion of the inner reaction tube 84A and the lower end portion of the outer reaction tube 84B, and engages with the lower end portion of the inner reaction tube 84A and the lower end portion of the outer reaction tube 84B, respectively.

[0045] Further, an O-ring 94 as a sealing member is provided between the furnace mouth portion 92 and the outer reaction tube 84B. The furnace mouth portion 92 is supported by the heater base 90, whereby the reaction tube 84 is in a state of being vertically installed. The reaction vessel is formed by the reaction tube 84 and the furnace mouth portion 92.

[0046] The furnace mouth portion 92 is connected with a processing gas nozzle 96A and a purge gas nozzle 96B in a manner of communicating with the processing chamber 86. The processing gas nozzle 96A is connected with the processing gas supply pipe 66A. An unillustrated processing gas supply source or the like is connected with the processing gas supply pipe 66A on the upstream side thereof via the MFC 64A. In addition, the purge gas nozzle 96B is connected with the purge gas supply pipe 66B. An unillustrated purge gas supply source or the like is connected with the purge gas supply pipe 66B on the upstream side thereof via the MFC 64B.

[0047] The furnace mouth portion 92 is connected with the exhaust pipe 68 that exhausts the atmosphere of the processing chamber 86. The exhaust pipe 68 is disposed at the lower end portion of a cylindrical space 98 formed by the gap between the inner reaction tube 84A and the outer reaction tube 84B, and communicates with the cylindrical space 98. On the downstream side of the exhaust pipe 68, a pressure sensor 70, a pressure adjustment portion 72, and a vacuum pump 74 are connected in this order from the upstream side.

[0048] Below the furnace mouth portion 92, a disc-shaped lid body 42 that can hermetically close the lower end opening of the furnace mouth portion 92 is provided, and an O-ring 100 as a sealing member that abuts against the lower end of the furnace mouth portion 92 is provided on the upper surface of the lid body 42.

[0049] On the side of the lid body 42 opposite to the processing chamber 86 in the vicinity of the center portion thereof, a rotation mechanism 46 that rotates the boat 36 is provided. A rotation shaft 102 of the rotation mechanism 46 penetrates the lid body 42 and supports the boat 36 from below. In addition, a rotation motor 46A is built in the rotation mechanism 46, and the rotation mechanism 46 is configured to rotate the rotation shaft 102 thereof by the rotation motor 46A, rotate the boat 36, and thereby rotate the substrate 16.

[0050] The lid 42 is configured to be raised and lowered in the vertical direction by a boat elevator 38 provided outside the reaction tube 84. The boat 36 is configured to be transported to the processing chamber 86 by raising and lowering the lid 42. The rotation motor 46A of the rotation mechanism 46 and the boat elevator 38 are electrically connected to the transport controller 48.

[0051] The boat 36 is configured to arrange and hold a plurality of substrates 16 in a horizontal posture and in a state in which the centers thereof are aligned with each other in a plurality of layers. In addition, a plurality of heat insulating plates 104 in the shape of a circular plate as heat insulating members are arranged in a plurality of layers in a horizontal posture in the lower portion of the boat 36. The boat 36 and the heat insulating plates 104 are, for example, made of a heat resistant material such as quartz or silicon carbide. The heat insulating plates 104 are provided to make it difficult for heat from the heater 88 to be transmitted to the port portion 92.

[0052] In addition, a temperature sensor 106 as a temperature detector is provided in the reaction tube 84. The heater 88 and the temperature sensor 106 are electrically connected to the temperature controller 76.

[0053] <Operation of the substrate processing apparatus>

[0054] Next, with reference to Figure 1 and Figure 2 , a method of forming a thin film on the substrate 16 will be described as one process of a manufacturing process of a semiconductor device. Further, the operation of each portion of the substrate processing apparatus 10 is controlled by the substrate processing apparatus controller 58.

[0055] As shown in Figure 1 , when the wafer cassette 18 is supplied to the load port 20 by an in-process transport device (not shown), the wafer cassette 18 is detected by the substrate detection sensor 56, and the wafer cassette loading / unloading port is opened by a front stopper (not shown). Then, the wafer cassette 18 on the load port 20 is carried into the inside of the housing 12 from the wafer cassette loading / unloading port by the wafer cassette transport device 24.

[0056] The wafer cassette 18 carried into the inside of the housing 12 is automatically transported to the shelf plate of the rotary wafer cassette rack 22 by the wafer cassette transport device 24 and is temporarily stored. Thereafter, the wafer cassette 18 is transferred from the shelf plate to the placement table of one wafer cassette opener 26. Further, the wafer cassette 18 carried into the inside of the housing 12 can be directly transferred to the placement table of the wafer cassette opener 26 by the wafer cassette transport device 24.

[0057] The lid of the wafer cassette 18 placed on the placement table is removed by the lid attachment / detachment mechanism 30, and the substrate inlet / outlet is opened. Thereafter, the substrate 16 (refer to Figure 2) is picked up from the cassette 18 by the gripper of the substrate transfer device 34A, and after the orientation is adjusted by a not-shown slot aligner, is carried into a standby section 50 located at the back of the transfer chamber 32, and is loaded into the boat 36. Then, the substrate transfer device 34A returns to the stage on which the cassette 18 is placed, takes out the next substrate 16 from the cassette 18, and loads it into the boat 36.

[0058] In the loading of the substrates 16 into the boat 36 by the substrate transfer mechanism 34 in the cassette opener 26 on one side (upper or lower), the other cassette 18 is transported from the rotary cassette stand 22 by the cassette transport device 24 on the stage of the cassette opener 26 on the other side (lower or upper). The opening of the cassette 18 by the cassette opener 26 is performed simultaneously with the transfer of the other cassette 18 to the stage.

[0059] When the number of substrates 16 specified in advance are loaded into the boat 36, the lower end of the processing furnace 44 is opened by a not-shown furnace port gate. Then, the boat 36 holding the group of substrates 16 is carried (loaded) into the processing furnace 44 by raising the lid 42 by the boat elevator 38.

[0060] As described above, when the boat 36 holding a plurality of substrates 16 is carried (loaded) into the processing chamber 86 of the processing furnace 44, as shown in Figure 2 the lid 42 becomes a state in which the lower end of the furnace port 92 is sealed by the O-ring 100.

[0061] After that, vacuum evacuation is performed by the vacuum pump 74 to make the processing chamber 86 a desired pressure (degree of vacuum). At this time, feedback control is performed on the pressure adjusting section 72 (valve opening degree) based on the pressure value measured by the pressure sensor 70. In addition, heating is performed by the heater 88 to make the processing chamber 86 a desired temperature. At this time, feedback control is performed on the power supply amount to the heater 88 based on the temperature value detected by the temperature sensor 106. Then, the boat 36 and the substrates 16 are rotated by the rotation mechanism 46.

[0062] Then, the processing gas supplied from a processing gas supply source and controlled to a desired flow rate by the MFC 64A flows in the processing gas supply pipe 66A, and is introduced into the processing chamber 86 from the processing gas nozzle 96A. The introduced processing gas rises in the processing chamber 86, flows out from the upper end opening of the inner reaction pipe 84A to the cylindrical space 98, and is discharged from the exhaust pipe 68. The processing gas contacts the surfaces of the substrates 16 while passing through the processing chamber 86, and at this time, a thin film is deposited on the surfaces of the substrates 16 by thermal reaction.

[0063] When the predetermined processing time elapses, the purge gas supplied from the purge gas supply source and controlled to the desired flow rate by the MFC 64B is supplied to the processing chamber 86, the processing chamber 86 is replaced with the inert gas, and the pressure of the processing chamber 86 is returned to the normal pressure.

[0064] After that, the lid body 42 is lowered by the boat lifter 38 to open the lower end of the furnace mouth portion 92, and the boat 36 holding the processed substrate 16 is carried out (unloaded) from the lower end of the furnace mouth portion 92 to the outside of the reaction tube 84. After that, the processed substrate 16 is taken out (unloaded) from the boat 36 and accommodated in the wafer cassette 18.

[0065] After the unloading, the wafer cassette 18 in which the processed substrate 16 is accommodated is carried out to the outside of the housing 12 in substantially the reverse order of the above-described order, except for the adjustment process in the notch alignment device.

[0066] <Structure of controller for substrate processing apparatus>

[0067] Next, the substrate processing apparatus controller 58 as the main control section will be specifically described with reference to Figure 3

[0068] The substrate processing apparatus controller 58 is mainly composed of an arithmetic control section 108 such as a CPU (Central Processing Unit), a storage section 114 provided with a RAM 110, a ROM 112, and a HDD not shown, an input section 116 such as a mouse or a keyboard, and a display section 118 such as a monitor, and is configured to be able to set each data by the arithmetic control section 108, the storage section 114, the input section 116, and the display section 118.

[0069] The arithmetic control section 108 constitutes the center of the substrate processing apparatus controller 58, executes a control program stored in the ROM 112, and executes a process (for example, a process process as a substrate processing process, or the like) stored in the storage section 114 also constituting a process storage section, in accordance with an instruction from the input section 116.

[0070] The ROM 112 is a recording medium composed of a flash memory, a hard disk, or the like, and stores an operation program of the arithmetic control section 108 that controls the operation of each component (for example, the vacuum pump 74, or the like) of the substrate processing apparatus 10, or the like. In addition, the RAM 110 (memory) functions as a work area (temporary storage section) of the arithmetic control section 108.

[0071] ​Here, the substrate processing procedure (process procedure) is a procedure that defines the processing conditions, processing steps, and the like for processing the substrate 16. In addition, in the procedure file, the set values, transmission timing, and the like that are transmitted to the conveyance controller 48, the temperature controller 76, the pressure controller 78, the gas supply controller 80, and the like are set for each step of the substrate processing procedure.

[0072] The arithmetic control section 108 has a function of controlling the temperature, pressure, flow rate of the processing gas introduced into the processing furnace 44, and the like within the processing furnace 44 to perform predetermined processing on the substrate 16 loaded into the processing furnace 44.

[0073] The conveyance controller 48 is configured to control the conveyance actions of the rotary wafer cassette rack 22, the boat elevator 38, the wafer cassette conveyance device 24, the substrate transfer mechanism 34, the boat 36, and the rotation mechanism 46 that constitute the conveyance mechanism of the substrate 16, respectively.

[0074] In addition, sensors are built in the rotary wafer cassette rack 22, the boat elevator 38, the wafer cassette conveyance device 24, the substrate transfer mechanism 34, the boat 36, and the rotation mechanism 46, respectively. When the sensors indicate predetermined values, abnormal values, and the like, respectively, the substrate processing apparatus controller 58 is notified of the situation. Furthermore, the detection system of the abnormal precursors of the components of the substrate processing apparatus 10 will be described later in detail.

[0075] The storage section 114 is provided with a data storage area 120 that stores various data and the like and a program storage area 122 that stores various programs including the substrate processing procedure. The data storage area 120 stores various parameters related to the procedure file. In addition, the program storage area 122 stores various programs required to control the apparatus including the above-described substrate processing procedure.

[0076] In addition, a touch panel that is not shown is provided to the display section 118 of the substrate processing apparatus controller 58. The touch panel is configured to display an operation screen that accepts the input of operation commands for the above-described substrate conveyance system and substrate processing system. Furthermore, the substrate processing apparatus controller 58 can be a structure that includes at least the display section 118 and the input section 116 like a personal computer, a mobile phone, and the like as an operation terminal (terminal apparatus).

[0077] The temperature controller 76 adjusts the temperature within the processing furnace 44 by controlling the temperature of the heater 88 of the processing furnace 44. Furthermore, when the temperature sensor 106 indicates predetermined values, abnormal values, and the like, the substrate processing apparatus controller 58 is notified of the situation.

[0078] The pressure controller 78 controls the pressure adjustment unit 72 based on the pressure value detected by the pressure sensor 70, so that the pressure in the processing chamber 86 reaches the desired pressure at a desired timing. Furthermore, when the pressure sensor 70 indicates a predetermined value, an abnormal value, etc., the controller 58 of the board processing apparatus is notified of the situation.

[0079] The gas supply controller 80 is configured to control the MFCs 64A and 64B so that the flow rate of the gas supplied to the processing chamber 86 is the desired flow rate at a desired timing. In addition, when the sensors (not shown) provided in the MFCs 64A, 64B, etc., indicate a predetermined value, an abnormal value, etc., the controller 58 of the board processing apparatus is notified of the situation.

[0080] <Substrate Processing Steps>

[0081] Next, use Figure 4 A general description of the substrate processing steps of using the substrate processing apparatus 10 of this embodiment as a semiconductor manufacturing apparatus to process substrates will be provided. This substrate processing step is, for example, a step in a semiconductor device (IC, LSI, etc.) manufacturing method. Furthermore, in the following description, the operation and processing of each component constituting the substrate processing apparatus 10 are controlled by the substrate processing apparatus controller 58.

[0082] Here, an example of forming a film on substrate 16 by alternately supplying a raw material gas (first processing gas) and a reactive gas (second processing gas) to substrate 16 will be described. Furthermore, an example of forming a silicon nitride (SiN) film as a thin film on substrate 16 using hexachlorosilane (Si2Cl6, hereinafter referred to as HCDS) gas as the raw material gas and ammonia (NH3) as the reactive gas will be described below. In addition, for example, a predetermined film can be pre-formed on substrate 16, or a predetermined pattern can be pre-formed on substrate 16 or the predetermined film.

[0083] (Substrate handling process S102)

[0084] First, in the substrate loading process S102, the substrate 16 is loaded into the crystal boat 36 and moved into the processing chamber 86.

[0085] (Film forming process S104)

[0086] In the film formation process S104, the following four steps are performed sequentially to form a thin film on the surface of the substrate 16. In addition, during steps 1 to 4, the substrate 16 is heated to a predetermined temperature using a heater 88.

[0087] [Step 1]

[0088] In Step 1, an unillustrated on-off valve provided in the process gas supply pipe 66A and the pressure adjusting portion 72 (APC valve) provided in the exhaust pipe 68 are opened, and the HCDS gas, the flow of which is regulated by the MFC 64A, is made to pass through the process gas supply pipe 66A. Then, the HCDS gas is supplied from the process gas nozzle 96A to the processing chamber 86, and exhaust gas is exhausted from the exhaust pipe 68. At this time, the pressure of the processing chamber 86 is maintained at a predetermined pressure. Thus, a silicon thin film (Si film) is formed on the surface of the substrate 16.

[0089] [Step 2]

[0090] In Step 2, the on-off valve of the process gas supply pipe 66A is closed to stop the supply of the HCDS gas. The pressure adjusting portion 72 (APC valve) of the exhaust pipe 68 is maintained in an open state, and the processing chamber 86 is exhausted by the vacuum pump 74 to remove the residual gas from the processing chamber 86. In addition, the on-off valve provided in the purge gas supply pipe 66B is opened, and an inert gas such as N2 is supplied to the processing chamber 86 to purge the processing chamber 86, and the residual gas in the processing chamber 86 is exhausted to the outside of the processing chamber 86.

[0091] [Step 3]

[0092] In Step 3, an unillustrated on-off valve provided in the purge gas supply pipe 66B and the pressure adjusting portion 72 (APC valve) provided in the exhaust pipe 68 are opened, and the NH3 gas, the flow of which is regulated by the MFC 64B, is made to pass through the purge gas supply pipe 66B. Then, the NH3 gas is supplied from the purge gas nozzle 96B to the processing chamber 86, and exhaust gas is exhausted from the exhaust pipe 68. At this time, the pressure of the processing chamber 86 is maintained at a predetermined pressure. Thus, the Si film formed on the surface of the substrate 16 by the HCDS gas is subjected to surface reaction with the NH3 gas, and a SiN film is formed on the substrate 16.

[0093] [Step 4]

[0094] In Step 4, the on-off valve of the purge gas supply pipe 66B is closed to stop the supply of the NH3 gas. The pressure adjusting portion 72 (APC valve) of the exhaust pipe 68 is maintained in an open state, and the processing chamber 86 is exhausted by the vacuum pump 74 to remove the residual gas from the processing chamber 86. In addition, an inert gas such as N2 is supplied to the processing chamber 86 to purge the processing chamber 86 again.

[0095] The above Steps 1 to 4 are repeated as one cycle, and a SiN film having a predetermined film thickness is formed on the substrate 16 by repeating the cycle a plurality of times.

[0096] (Substrate unloading process S106)

[0097] In the substrate unloading process S106, the pod 36 on which the substrate 16 on which the SiN film is formed is placed is unloaded from the processing chamber 86.

[0098] <Control system in the present embodiment>

[0099] Next, the control system for detecting a precursor of an abnormality (a precursor of a failure) of each component of the substrate processing apparatus 10 will be described with reference to Figure 5 and Figure 6 The control system for detecting a precursor of an abnormality (a precursor of a failure) of each component of the substrate processing apparatus 10 will be described. Hereinafter, an example in which a thin film is formed on the substrate 16 by the substrate processing apparatus 10 will be described.

[0100] As shown in Figure 5 , the control system is provided with a controller for substrate processing 58 as a main control unit, a precursor detection controller 82 as a precursor detection unit, various sensor groups 124, a data collection unit (hereinafter referred to as DCU) 126, and an edge controller (hereinafter referred to as EC) 128, which are connected to each other by wire or wirelessly.

[0101] The controller for substrate processing 58 is connected to an unillustrated host computer including a customer host computer and an unillustrated operation unit. The operation unit is configured to be able to exchange various data (sensor data and the like) acquired by the controller for substrate processing 58 with the host computer.

[0102] The precursor detection controller 82 acquires sensor data from sensors provided to various components of the substrate processing apparatus 10 to monitor the state of the substrate processing apparatus 10. Specifically, the precursor detection controller 82 calculates a numerical index using data from the various sensor groups 124, compares the numerical index with a threshold value determined in advance, and detects a precursor of an abnormality. Further, the precursor detection controller 82 has a precursor detection program for detecting a precursor of an abnormality based on a variation in sensor data.

[0103] In addition, the precursor detection controller 82 has two systems, a system directly connected to the controller for substrate processing 58 and a system connected to the controller for substrate processing 58 via the DCU 126. Therefore, in a case where a precursor of an abnormality is detected by the precursor detection controller 82, a signal can be output directly to the controller for substrate processing 58 without passing through the DCU 126, an alarm is generated, and information of sensor data of a sensor provided to a component in which the precursor of an abnormality is detected is displayed on a screen of the display unit 118 (refer to Figure 3 ).

[0104] The various sensor group 124 is a sensor (for example, a pressure sensor 70, a temperature sensor 106, and the like) provided to various components provided in the substrate processing apparatus 10, and detects a flow rate, a concentration, a temperature, a humidity (dew point), a pressure, a current, a voltage, a torque, a vibration, a position, a rotation speed, and the like of each component.

[0105] The DCU 126 collects and accumulates data of the various sensor group 124 in the execution of the process recipe. In addition, the EC 128 temporarily acquires sensor data as needed according to the kind of the sensor, and transmits the data to the precursor detection controller 82 after applying a process such as a Fast Fourier Transform (hereinafter referred to as FFT) to the raw data.

[0106] In addition, the various sensor group 124 is divided into a first sensor system 124A and a second sensor system 124B that differ in the transmission path. The first sensor system 124A is a system that acquires raw data in real time at a unit of 0.1 seconds, and transmits the raw data from the first sensor system 124A to the precursor detection controller 82 via the substrate processing apparatus controller 58 and the DCU 126 in real time. This first sensor system 124A includes, for example, a temperature sensor, a pressure sensor, a gas flow rate sensor, and the like.

[0107] On the other hand, the second sensor system 124B is a system that performs a process such as an FFT in the EC 128 to extract only a portion necessary for analysis, and transmits the data in the form of a processed file. The processed data is transmitted from the second sensor system 124B to the precursor detection controller 82 via the EC 128. This second sensor system 124B includes, for example, a vibration sensor and the like.

[0108] In the case where the sensor is a vibration sensor, vibration data is accumulated at a unit of milliseconds, and thus the amount of data becomes large. If the data is directly transmitted to the precursor detection controller 82, the capacity of the storage section of the precursor detection controller 82 is largely consumed. Since the data of the vibration sensor is finally processed by an FFT or the like to be used for analysis, by performing the process in advance in the EC 128, the amount of information can be reduced, and the data is transmitted to the precursor detection controller 82 in a form that is easy to analyze.

[0109] (First Embodiment)

[0110] Hereinafter, a first embodiment of a detection procedure of an abnormal precursor of each component of the substrate processing apparatus 10 using the above-described control system will be described in detail.

[0111] [Calculation of Non-Normality]

[0112] First, a "degree of abnormality" is calculated using values detected by a plurality of sensors directly provided to a component that is a detection target of an abnormality precursor and values detected by sensors of other components that are directly or indirectly affected by the state of the component. In the present embodiment, for example, the degree of abnormality is configured to have a property such that the value of the degree of abnormality substantially increases when the component that is the detection target of an abnormality precursor approaches an abnormal state. In addition, the degree of abnormality can also be configured to have a property such that the value decreases if the component that is the detection target of an abnormality precursor approaches an abnormal state.

[0113] [Raw data constituting the degree of abnormality]

[0114] The sequence of substrate processing is constituted by, for example, a plurality of events having various purposes, such as the carrying-in of the substrate 16 into the processing chamber 86, the evacuation of the processing chamber 86, the temperature increase, the purge using an inert gas, the temperature increase standby, the processing of the substrate 16 (for example, film formation), the replacement of the gas in the processing chamber 86, the return to atmospheric pressure, the carrying-out of the processed substrate 16, and the like. In addition, the above events are an example of a sequence of substrate processing, and there are cases where each event is further finely divided.

[0115] In the present embodiment, not all of the sensor data in the sequence is used, and the values of one or more sensors in one or more specific events among the events are used as raw data for calculating a numerical index, that is, the "degree of abnormality", in the algorithm. In addition, the degree of abnormality value is monitored for each Run (batch processing unit), and a precursor of an abnormality of each component of the substrate processing device 10 is detected. In this way, by using only the data of specific events, the data storage amount can be saved.

[0116] For example, the detection of a precursor of an abnormality of the vacuum pump 74 becomes a state that is easily detected at a timing at which a large load is applied to the vacuum pump 74. The step of reducing the pressure of the processing chamber 86 from atmospheric pressure to a predetermined pressure, that is, the start of evacuation or the pressure zone close to atmospheric pressure several minutes after the start of evacuation corresponds to the timing at which a large load is applied to the vacuum pump 74.

[0117] Specifically, one substrate processing device 10 is responsible for a plurality of processes, and there are cases where different processing processes such as processes having different film formation conditions are mixed and started. Since the raw material gas flows at the time of film formation of the substrate 16, the raw material gas sometimes reacts or thermally decomposes to generate a solid, and the solid sometimes applies a load to the vacuum pump 74, so monitoring in the film formation event is also effective for the detection of a precursor of an abnormality.

[0118] On the other hand, regarding the event of evacuation of the vacuum before the substrate processing, even if the subsequent substrate processing event is different, there are many cases where the event of evacuation of the vacuum is common. That is, even in the case where a plurality of different film formation conditions are started in the same apparatus, by monitoring the state at the start of evacuation in each Run which is common, sensor data is acquired, the same state can be known over time regardless of the content of the substrate processing, and high-precision prediction can be performed.

[0119] [Calculation example of abnormality]

[0120] Here, calculation examples of abnormality in the case of using sensor data of a vibration sensor, and in the case of using sensor data of a sensor other than a vibration sensor (for example, a current sensor, a temperature sensor, an exhaust pressure sensor, torque value data, and current data, etc.).

[0121] First, in the case where sensor data using a vibration sensor (vibration data) is judged for each individual frequency whether there is an abnormality, the following steps are taken.

[0122] (1) Acquire vibration data (raw data) detected by a vibration sensor in the sensor data in a specified step in each step constituting a process recipe.

[0123] (2) Transform the acquired vibration data into a vibration spectrum by processing such as FFT, and extract the frequency (the value is the amplitude (envelope) of the vibration, and in the example case, 500 dimensions) of a predetermined range (for example, 10 to 5000 Hz) of the transformed vibration spectrum at a predetermined frequency interval (for example, every 10 Hz).

[0124] (3) For each frequency extracted, calculate the mean μ and the standard deviation σ of the amplitude of the vibration spectrum using a predetermined number of data (for example, 30 Run amounts) of the process recipe at normal times, assuming that the amplitude at normal times follows a normal distribution N(μ, σ), and use it as a normal model.

[0125] (4) Take the value of (2) after the normal model is generated as an abnormality vector, compare the amplitude value of the normal model with a threshold value decided in advance for each frequency extracted, and in the case where the amplitude value of a predetermined number (for example, m (m ≥ 1) or more) of frequencies deviates from the threshold value, judge that a precursor to an abnormality has occurred (there is an abnormality precursor). In addition, for example, using the mean μ and the standard deviation σ calculated in (3), calculate the threshold value within a range (μ ± 3σ) obtained by adding or subtracting a value of 3 times the standard deviation σ from the mean value μ.

[0126] In addition, in the case where sensor data using a vibration sensor (vibration data) is judged by the sum of the amplitudes of each frequency, the following steps are taken.

[0127] (1) Obtain vibration data (raw data) detected by a vibration sensor from among sensor data in a specified step in each step constituting the process.

[0128] (2) Transform the obtained vibration data into a vibration spectrum by processing such as FFT, and extract frequencies in a predetermined range (for example, 10 to 5000 Hz) of the transformed vibration spectrum at predetermined frequency intervals (for example, every 10 Hz) (the values are amplitudes (envelopes) of the vibration, and in the illustrated case, 500 dimensions).

[0129] (3) Add the sum of the amplitudes of each frequency extracted in total for each Run in normal times (since one sum of the amplitudes is obtained for every 1 Run, if it is 30 Runs, 30 numbers are obtained).

[0130] (4) Calculate the mean μ and the standard deviation σ of the data group from the group of values obtained for each Run, and assume that the sum obtained for each Run follows a normal distribution N(μ, σ), and use it as a normal model.

[0131] (5) Use the value of (3) after the normal model is generated as a degree of abnormality, compare the amplitude value of the normal model with a threshold value decided in advance, and in the case where the amplitude value deviates from the threshold value, determine that there is a sign of abnormality (a sign that an abnormality has occurred). Further, for example, using the mean μ and the standard deviation σ obtained in (3), calculate the threshold value within a range (μ ± 3σ) obtained by adding or subtracting a value of 3 times the standard deviation σ from the mean value μ.

[0132] In addition, in the case where a judgment is made for each basic statistic using sensor data other than vibration data, the following steps are taken.

[0133] (1) Select one or more data from among the basic statistics of the mean, the standard deviation, the N quantile, the maximum value, and the minimum value of the sensor data of the subject event in normal times.

[0134] (2) Obtain the mean μ and the standard deviation σ for each statistic of the selected basic statistics of normal times, and assume that each basic statistic follows a normal distribution. Use this as a normal model of each basic statistic of the sensor.

[0135] (3) Use the value of (1) after the normal model is generated as a degree of abnormality, and in the case where the value of each basic statistic deviates from a predetermined threshold value decided in advance, determine that there is a sign of abnormality. Further, for example, using the mean μ and the standard deviation σ obtained in (2), calculate the threshold value within a range (μ ± 3σ) obtained by adding or subtracting a value of 3 times the standard deviation σ from the mean value μ.

[0136] In addition, as shown in Figure 6 in the case of using sensor data other than vibration data and using singular spectrum transformation for the determination, the following steps are taken. Further, in the following steps, using the partial time series of the Run p with the window width n of the surroundings, 2 data matrices X and Z are generated on the past and current sides. The following steps are the general approach of singular spectrum transformation.

[0137] (1) Respective M-dimensional longitudinal vectors are prepared, and these vectors are longitudinally connected n times from the uppermost S(p-n+1, 1) to the lowermost S(p, M) to form an Mn-dimensional longitudinal vector.

[0138] Sensor data of the time 1, 2,..., M of the target event of the Run p-n+1

[0139] {S(p-n+1, 1), S(p-n+1, 2),..., S(p-n+1, M)}

[0140] ...

[0141] Sensor data of the time 1, 2,..., M of the target event of the Run p-1

[0142] {S(p-1, 1), S(p-1, 2),..., S(p-1, M)}

[0143] Sensor data of the time 1, 2,..., M of the target event of the Run p

[0144] {S(p, 1), S(p, 2),..., S(p, M)}

[0145] (2) Respective M-dimensional longitudinal vectors are prepared, and these vectors are longitudinally connected n times from the uppermost S(p-n+1, 1) to the lowermost S(p, M) to form an Mn-dimensional longitudinal vector (offset to the old 1 Run group compared to (1)).

[0146] Sensor data of the time 1, 2,..., M of the target event of the Run p-n

[0147] {S(p-n, 1), S(p-n, 2),..., S(p-n, M)}

[0148] ...

[0149] Sensor data of the time 1, 2,..., M of the target event of the Run p-2

[0150] {S(p-2, 1), S(p-2, 2),..., S(p-2, M)}

[0151] Sensor data of the time 1, 2,..., M of the target event of the Run p-1

[0152] {S(p-1, 1), S(p-1, 2),..., S(p-1, M)}

[0153] (3) As with (1) and (2) above, K longitudinal vectors sequentially formed are prepared, and a matrix X(p) of Mn x K dimensions is generated by arranging these longitudinal vectors from left to right in order from old to new. As described above, a history matrix for performing singular spectrum transformation is generated.

[0154] (4) As M-dimensional longitudinal vectors, longitudinal vectors formed by longitudinally connecting n of the vectors from the uppermost S(p+L, 1) to the lowermost S(p+L-n+1, M) are prepared. Further, L is set to a positive integer.

[0155] Sensor data of time 1, 2,..., M of the object event of Run p+L

[0156] {S(p+L, 1), S(p+L, 2),..., S(p+L, M)}

[0157] ...

[0158] Sensor data of time 1, 2,..., M of the object event of Run p+L-n+2

[0159] {S(p+L-n+2, 1), S(p+L-n+2, 2),..., S(p+L-n+2, M)}

[0160] Sensor data of time 1, 2,..., M of the object event of Run p+L-n+1

[0161] {S(p+L-n+1, 1), S(p+L-n+1, 2),..., S(p+L-n+1, M)}

[0162] (5) As M-dimensional longitudinal vectors, longitudinal vectors formed by longitudinally connecting n of the vectors from the uppermost S(p+L-1, 1) to the lowermost S(p+L-n, M) (Run group shifted by 1 old compared to (4)) are prepared.

[0163] Sensor data of time 1, 2,..., M of the object event of Run p+L-1

[0164] {S(p+L-1, 1), S(p+L-1, 2),..., S(p+L-1, M)}

[0165] ...

[0166] Sensor data of time 1, 2,..., M of the object event of Run p+L-n+1

[0167] {S(p+L-n+1, 1), S(p+L-n+1, 2),..., S(p+L-n+1, M)}

[0168] Sensor data of time 1, 2,..., M of the object event of run p+L-n

[0169] {S(p+L-n, 1), S(p+L-n, 2),..., S(p+L-n, M)}

[0170] (6) As with the above (4), (5), R longitudinal vectors successively configured are prepared, and a matrix Z(p) of Mn x R dimensions obtained by arranging these longitudinal vectors from left to right from old to new is generated. Thus, a test matrix of singular spectrum transformation is generated.

[0171] (7) Singular value decomposition is performed on the above matrix X(p) and matrix Z(p), and singular spectrum transformation is performed.

[0172] (8) r left singular vectors obtained by singular value decomposition are selected in X(p), and m left singular vectors obtained by singular value decomposition are selected in Z(p), and matrices are configured with U(r), Q(m), respectively, and the maximum singular value of the product of these matrices U(r) T Q(m) is found. This maximum singular value is set as λ (0 ≤ λ ≤ 1), and 1-λ is set as the abnormality degree (variation degree). In the case where this abnormality degree deviates from a predetermined threshold value decided in advance, it is judged that there is an abnormality precursor.

[0173] [Abnormality precursor judgment using abnormality degree]

[0174] In addition, as a judgment method of the presence or absence of an abnormality precursor using the abnormality degree, for example, the following method is considered. Further, in the case where it is judged that there is an abnormality precursor, the substrate processing device controller 58 is notified.

[0175] (1) A method of judging that there is an abnormality precursor in the case where the abnormality degree of at least 1 sensor data deviates from the threshold value.

[0176] (2) A method of judging that there is an abnormality precursor in the case where the abnormality degrees of 2 or more sensor data deviate from the threshold value.

[0177] (3) A method of judging that there is an abnormality precursor in the case where the abnormality degrees of 1 or 2 or more sensor data deviate from the threshold value a predetermined number of times (for example, 3 times).

[0178] (4) A method of judging that there is an abnormality precursor in the case where the abnormality degrees of sensor data other than vibration data deviate from the threshold value continuously a predetermined number of times (for example, 3 times).

[0179] (5) A method that does not judge an abnormality as a sign of an impending event if the abnormality of sensor data other than vibration data deviates from the threshold, but the abnormality of vibration data does not deviate from the threshold.

[0180] (6) A method to determine that there are signs of abnormality when both the abnormality of vibration data and the abnormality of sensor data other than vibration data deviate from the threshold.

[0181] For example, in the methods (2), (5), and (6) above, since multiple sensor data are used to determine abnormal signs, false detections by the sensors can be reduced. Furthermore, the variation in abnormality is not necessarily monotonic; therefore, in the methods (3) and (4) above, false judgments can be reduced when the abnormality value fluctuates around the threshold. In addition, the formula for calculating the abnormality, the threshold, and the program are different for each component and each device, and are pre-assembled into the sign detection controller 82.

[0182] [Display of the analysis screen for abnormal warning signs]

[0183] The analysis screen of abnormality detection can be displayed on the display unit 118 of the controller 58 for the board processing device (see reference). Figure 3 The display shows the abnormality level. Therefore, it is possible to visually observe the progression of abnormality, thresholds, and the number of times the threshold is exceeded, thus enabling the identification of the component's status through abnormality levels.

[0184] [Through the EC]

[0185] Here, on Figure 5 The case of the second sensor system 124B shown, i.e., the case where EC128 is located between the sensor and the precursor detection controller 82, will be explained.

[0186] [Time synchronization]

[0187] The vibration sensor data is transformed in EC128 and then sent to the precursor detection controller 82 in the form of EC128's time. In order to simultaneously use this vibration sensor data and other sensor data with times from DCU126 and the board processing device controller 58 for analysis, the times of both need to be synchronized for analysis. Therefore, EC128, DCU126, and the precursor detection controller 82 periodically take in their times, using the time of the board processing device controller 58 as a reference time, to synchronize their times. Thus, the time synchronization of all components enables accurate analysis.

[0188] Here, vacuum pump 74 (reference) Figure 2 Taking the example of a substrate processing apparatus 10, a method for detecting abnormal signs of components will be specifically explained.

[0189] In the processing chamber 86 of the substrate processing apparatus 10, reaction byproducts of the processing gas accumulate inside, and when the amount, height of the reaction byproducts reaches a certain level, the rotation of the vacuum pump 74 is urgently stopped.

[0190] Here, by continuously monitoring at least one of the sensor data of the current data, temperature data, exhaust pressure data, and vibration data of the vacuum pump 74, and analyzing the change in behavior of these sensor data using a precursor detection program in the precursor detection controller 82, the precursor of the abnormality of the vacuum pump 74 can be detected. In the case where the precursor of the abnormality is detected, the information is transmitted to the substrate processing apparatus controller 58, and the operator is notified to replace or maintain the vacuum pump 74.

[0191] (Second Embodiment)

[0192] Next, the second embodiment of the detection procedure of the abnormality precursor of each component of the substrate processing apparatus 10 using the above-described control system will be described. In addition, the structure of the precursor detection controller 82 and the like, and the abnormality precursor judgment using the abnormality degree are the same as those of the first embodiment.

[0193] [Calculation of Abnormality Degree]

[0194] In the present embodiment, the sensor data at the normal time is learned using the values of a plurality of sensors provided to the component which is the object of the abnormality precursor detection and the values of the sensors of other components which are directly or indirectly affected by the state of the component, and the learned data and the data during operation are used to calculate the "abnormality degree".

[0195] In the present embodiment, for example, it is configured to have a property that the value of the abnormality degree substantially increases when the component which is the object of the abnormality precursor detection approaches an abnormal state. In addition, the abnormality degree can also be configured to have a property that the value decreases if the component which is the object of the abnormality precursor detection approaches an abnormal state.

[0196] Here, the detection method of the abnormality precursor of the component of the substrate processing apparatus 10 will be described taking the vacuum pump 74 (refer to Figure 2 ) as an example.

[0197] Generally, in a state where the processing chamber 86 is evacuated by the vacuum pump 74, the inert gas, film forming gas flows in the vacuum pump 74 to become a state where the load is high, and a state where the abnormality precursor is easily detected. On the other hand, in a state where the processing chamber 86 is not evacuated by the vacuum pump 74, the load of the vacuum pump 74 becomes a state where the load is small, and a state where the abnormality precursor is difficult to detect or the abnormality is difficult to occur. Therefore, in the past, the vacuum pump 74 was monitored in a state where the processing chamber 86 was evacuated.

[0198] On the other hand, in the present embodiment, in the event of a state in which the processing chamber 86 is not being vacuumed by the vacuum pump 74 and there is no substrate 16 in the processing chamber 86, a large amount of gas is intentionally caused to flow to the vacuum pump 74 to increase the load on the vacuum pump 74. Also, by monitoring the current data, vibration data, temperature data, back pressure data, and the like of the vacuum pump 74 in this state, it is easy to detect a precursor of an abnormality.

[0199] In this way, by applying a load to the vacuum pump 74 in a state in which the processing chamber 86 is not being vacuumed, even if the vacuum pump 74 stops when the load is applied, it is possible to prevent damage to the substrate 16. Also, in the case in which the vacuum pump 74 stops to the extent of the load being applied in a state in which the processing chamber 86 is not being vacuumed, it is considered that the vacuum pump 74 is in a state just before a failure occurs. As a result, it is possible to avoid a situation in which the vacuum pump 74 stops in a state in which the processing chamber 86 is being vacuumed, i.e., during substrate processing.

[0200] (Third Embodiment)

[0201] Next, a third embodiment of the procedure for detecting a precursor of an abnormality of each component of the substrate processing apparatus 10 using the above-described control system will be described in detail. Furthermore, the structure of the precursor detection controller 82 and the like, and the abnormality precursor determination using the abnormality degree are the same as in the first and second embodiments.

[0202] In the present embodiment, in the case in which a component that is an abnormality precursor detection object is replaced or maintained, a normal model after the replacement or maintenance is generated, and the substrate processing apparatus 10 is monitored based on this normal model, and an abnormality precursor determination is performed.

[0203] In the present embodiment, replacement or maintenance of a component that is an abnormality precursor detection object is detected automatically or semi-automatically. For example, in the case in which a component that is an abnormality precursor detection object has operation cumulative time information, it is possible to detect replacement of the component using the operation cumulative time information. The operation cumulative time possessed by a component that is an abnormality precursor detection object is generally held in a non-volatile storage medium, and thus the operation time is accumulated until the replacement of the component, and the operation time is reset by the replacement. Therefore, the operation cumulative time possessed by a component that is an abnormality precursor detection object is monitored, and in the case in which the operation cumulative time decreases, it is possible to detect that there is replacement of the component.

[0204] Specifically, the substrate processing apparatus controller 58 transmits the operation cumulative time of a component that is an abnormality precursor detection object to the precursor detection controller 82 at predetermined intervals, and the precursor detection controller determines whether the newly transmitted operation cumulative time is shorter than the operation cumulative time stored previously. In the case in which the determination is affirmative, it is possible to determine that there is replacement of the component.

[0205] In addition, instead of the operation cumulative time information, even in a case where the component that is the abnormality precursor detection object does not have an operation cumulative time, it is possible to detect the component replacement using the work of detaching the signal connector at the time of component replacement. In the work of detaching the signal connector at the time of component replacement, the signal line of the component is opened (disconnected), and thus in a case where the signal line of the component is opened (disconnected), at the time of energization of the signal line next time, the worker is prompted to perform a confirmation input of whether there is replacement or maintenance. For example, if the confirmation input is not performed on the operation screen, it is not possible to start other work. Thus, it is possible to semi-automatically determine that there is replacement of the component.

[0206] In a case where it is determined that there is replacement or maintenance of the component that is the abnormality precursor detection object, as part of the precursor detection processing, the precursor detection controller 82 newly acquires sensor data of the component that is the abnormality precursor detection object to update the normal model. Then, based on the updated normal model, the abnormality degree is calculated. The calculation of the abnormality degree, the precursor detection based on the monitoring abnormality degree value, can be performed similarly to the first and second embodiments.

[0207] Here, a case where the vacuum pump 74 (refer to Figure 2 ) is replaced is taken as an example, and the method of detecting the abnormality precursor of the substrate processing device 10 is described in detail.

[0208] The substrate processing device controller 58 is configured to, until the vacuum pump 74 is replaced, if the operation time is acquired as the sensor data of the vacuum pump 74, accumulate the acquired operation time of the vacuum pump 74, and reset the operation time (operation cumulative time) by replacement. In addition, as shown in Figure 5 , the substrate processing device controller 58 is connected to the precursor detection controller 82, and transmits the operation cumulative time to the precursor detection controller 82 at predetermined intervals. Furthermore, the time (operation cumulative time) obtained by accumulating the operation time of the vacuum pump 74 can also be managed by the precursor detection controller 82 by directly acquiring the operation time from the vacuum pump 74.

[0209] As part of the precursor detection processing, the precursor detection controller 82 acquires the operation cumulative time of the vacuum pump 74 from the substrate processing device controller 58, and calculates the abnormality degree based on the normal model and the operation cumulative time of the vacuum pump 74. Figure 7As shown, the precursor detection controller 82 acquires the operation cumulative time transmitted from the substrate processing apparatus controller 58 (S10), judges whether it is shorter than the operation cumulative time stored previously (S12), in the case of affirmative, judges that there is replacement of the vacuum pump 74, acquires sensor data required for generation of a normal model after the replacement (S14). For example, sensor data of a predetermined number of times of the process (for example, an amount of 30 runs) is acquired. Then, based on the acquired sensor data, a normal model is generated (S15). For example, an average μ, a standard deviation σ are calculated using the sensor data of the predetermined number of times of the process, it is assumed that each of the sensor data at the normal time follows a normal distribution N(μ, σ), and it is used as the normal model. Based on the normal model obtained, the abnormality degree is calculated (S16), and the data of the abnormality degree stored previously is rewritten to the calculated abnormality degree (S17). Then, the substrate processing apparatus 10 is monitored (S18), and the abnormality precursor judgment is performed. As to the calculation of the abnormality degree, the monitoring of the abnormality degree value, the same as the first and second embodiments can be performed.

[0210] According to the present embodiment, after the replacement or the maintenance of the component as the abnormality precursor detection object is performed, the normal model is newly generated, and thus appropriate abnormality precursor detection (detection of a precursor of an abnormality) can be performed. In addition, since the replacement or the maintenance of the component as the abnormality precursor detection object is detected automatically or semi-automatically, the change of the abnormality value of the required monitoring object can be appropriately performed.

[0211] (Fourth Embodiment)

[0212] Next, the fourth embodiment of the detection procedure of the abnormality precursor of each component of the substrate processing apparatus 10 using the above-described control system will be described in detail. Further, the structure of the precursor detection controller 82 and the like, the abnormality precursor judgment using the abnormality degree are the same as those of the first to third embodiments.

[0213] In the present embodiment, in the case where the replacement or the maintenance of the component as the abnormality precursor detection object is performed, before the normal model after the replacement or the maintenance is newly generated, a judgment of whether the normal model is newly generated or the normal model before the replacement or the maintenance is continued to be used is performed. As to the automatic or semi-automatic detection of the replacement or the maintenance of the component as the abnormality precursor detection object, the same as the third embodiment is performed.

[0214] Specifically, in the case where the replacement or the maintenance of the component as the abnormality precursor detection object is judged to exist, the precursor detection controller 82 acquires sensor data of a smaller amount of data required for generation of the normal model. Then, based on the acquired sensor data, a judgment of whether the normal model before the replacement or the maintenance can be continued to be used is performed.

[0215] In a case where it is judged that the normal model before the replacement or the maintenance can be continued to be used, the sensor data required for generating the normal model is not acquired, and the normal model before the replacement or the maintenance is used. Therefore, the calculation of the abnormality degree is not required, and the same abnormality degree value as before the replacement or the maintenance is monitored to perform the precursor detection.

[0216] In a case where it is judged that the normal model before the replacement or the maintenance cannot be continued to be used, the acquisition of the sensor data is further performed to obtain the sensor data required for generating the normal model, and the normal model is newly generated. Then, the abnormality degree is calculated based on the new normal model, and the new abnormality degree value is monitored to perform the precursor detection.

[0217] Here, as a specific example, a case where the vacuum pump 74 (refer to Figure 2 ) is replaced is exemplified, and the method of detecting the abnormality precursor of the substrate processing apparatus 10 is specifically described.

[0218] [Specific Example]

[0219] As shown in Figure 8 , the precursor detection controller 82 acquires the operation cumulative time transmitted from the substrate processing apparatus controller 58 (S30), judges whether it is shorter than the operation cumulative time stored before (S32), judges that there is a replacement of the vacuum pump 74 in a case where it is judged to be affirmative, and acquires the sensor data required for judging whether the normal model before the replacement can be used (judgment-use sensor data) (S33). The data amount of the judgment-use sensor data becomes an amount of sensor data (for example, an amount of 10 Run) that is less than an amount of sensor data (for example, an amount of 30 Run) required for generating the normal model. Then, it is statistically judged whether the distribution of the acquired judgment-use sensor data is equal to the data distribution of the normal model before the replacement, and it is judged whether the normal model before the replacement can be used (S34).

[0220] As one example, the statistical judgment can be performed as follows.

[0221] (1) For the data group before the replacement and the data group after the replacement, normality is judged by Shapiro-Wilk test,

[0222] (2) In F test, it is judged whether the variances of the data group before the replacement and the data group after the replacement are equal,

[0223] (3) Based on the results of (1) and (2) above, a test of the difference between the average values (representative values) is performed by any one of a t-test, a Welch's t test, and a Mann-Whitney U test.

[0224] In a case where the distribution of the acquired sensor data is equal to the data distribution of the normal model before the replacement, it is determined that the normal model before the replacement can be used (Yes), and the acquisition of the sensor data required for generating the normal model is not performed, and the abnormality degree value based on the normal model before the replacement is monitored (S39) to perform the precursor detection.

[0225] In a case where the distribution of the acquired sensor data is not equal to the data distribution of the normal model before the replacement, it is determined that the normal model before the replacement cannot be used (No), and the acquisition of the sensor data is further performed (S35), the sensor data required for generating the normal model is obtained, and the normal model is newly generated (S36). Based on the obtained normal model, the abnormality degree is calculated (S37), and the previously stored data of the abnormality degree is rewritten as the calculated abnormality degree (S38). Then, the substrate processing apparatus 10 is monitored (S39) to perform the abnormality precursor determination. The calculation of the abnormality degree and the monitoring of the abnormality degree value can be performed similarly to the first and second embodiments.

[0226] According to the present embodiment, whether or not the normal model before the replacement or the maintenance of the component can be used is determined by acquiring a smaller amount of sensor data than that required for generating the normal model for the component that is an object of the abnormality precursor detection. Therefore, the time during which the monitoring for the abnormality precursor detection is stopped due to the generation of the normal model can be shortened.

[0227] (Action, Effect)

[0228] According to the above-described embodiments, the substrate processing apparatus 10 has a control system that detects the abnormality precursor of the component, and thus the component can be replaced or maintained at the time point at which the abnormality precursor of the component is detected by the control system. In particular, with regard to the failure precursor detection of the vacuum pump 74, the accuracy of the abnormality precursor can be improved by continuously monitoring the sensor data such as the current data, the temperature data, the exhaust pressure data, and the vibration data of the vacuum pump 74.

[0229] Thus, the replacement or the like can be performed before the component fails, and the replacement frequency can be reduced by using the component until just before the failure. In addition, by preventing the failure during the processing of the substrate, the improvement of the apparatus operation rate, the prevention of the reduction of the yield of the product (substrate 16), and the reduction of the unnecessary maintenance time can be achieved.

[0230] In addition, according to the above-described embodiment, the abnormality precursor detecting controller 82 that detects the abnormality precursor is connected to the substrate processing apparatus controller 58. Therefore, it is possible to acquire and analyze data for a specific substrate processing sequence in which the abnormality precursor is easily detected.

[0231] In addition, even after replacement and maintenance of the component that is the object of the abnormality precursor detection, it is possible to detect a precursor of an abnormality of the component that is the object of the abnormality precursor detection using an appropriate normal model.

[0232] (Other Embodiments)

[0233] The above-described embodiments of the present disclosure have been specifically described, but the present disclosure is not limited to the above-described embodiments, and various modifications can be made without departing from the spirit of the present disclosure.

[0234] For example, in the above-described embodiments, an example in which a thin film is formed on a substrate 16 has been described. However, the present disclosure is not limited to such a mode, and for example, in a case where a thin film or the like formed on the substrate 16 is subjected to an oxidation process, a diffusion process, an annealing process, an etching process, or the like, the present disclosure can be appropriately applied.

[0235] In addition, in the present embodiment, an example in which a thin film is formed using the substrate processing apparatus 10 having the heat-wall-type processing furnace 44 has been described, but the present disclosure is not limited thereto, and can be appropriately applied to a case where a thin film is formed using a substrate processing apparatus having a cold-wall-type processing furnace. Furthermore, in the above-described embodiments, an example in which a thin film is formed using the batch-type substrate processing apparatus 10 that processes a plurality of substrates 16 at a time has been described, but the present disclosure is not limited thereto.

[0236] In addition, the present disclosure is not limited to a semiconductor manufacturing apparatus or the like that processes a semiconductor substrate like the substrate processing apparatus 10 of the above-described embodiments, and can be applied to an LCD (Liquid Crystal Display) manufacturing apparatus that processes a glass substrate.

[0237] Explanation of Reference Numerals

[0238] 10 substrate processing apparatus,

[0239] 16 substrate,

[0240] 58 substrate processing apparatus controller (example of main control section),

[0241] 74 vacuum pump,

[0242] 82 abnormality precursor detecting controller (example of abnormality precursor detecting section),

[0243] 86 processing chamber,

[0244] μ mean value,

[0245] σ standard deviation.

Claims

1. A substrate processing apparatus comprising: a main control unit that controls to execute a process recipe including a plurality of steps to perform a predetermined process on a substrate; and a precursor detection unit that acquires sensor data related to a component that is an abnormality precursor detection object to generate a normal model, and monitors a state of the apparatus based on the normal model, characterized in that the precursor detection unit acquires the sensor data after replacement or maintenance of the component that is the abnormality precursor detection object, and acquires, from the sensor data, vibration data detected by a vibration sensor in a specified step among the steps constituting the process recipe, transforms the acquired vibration data into a vibration spectrum, extracts the transformed vibration spectrum at predetermined frequency intervals, calculates, for each frequency extracted, an average value and a standard deviation of an amplitude of the vibration spectrum using data of a predetermined number of times of the process recipe at normal times, regenerates the normal model using the average value and the standard deviation of the amplitude of the vibration spectrum obtained, and monitors the state of the apparatus based on the normal model to detect a precursor of an abnormality before the apparatus abnormally stops.

2. The substrate processing apparatus according to claim 1, characterized in that the precursor detection unit determines whether the normal model before the replacement or maintenance can be used as a normal model after the replacement or maintenance based on the sensor data that is less in data amount than that required to generate the normal model.

3. The substrate processing apparatus according to claim 2, characterized in that the precursor detection unit uses the normal model before the replacement or maintenance as the normal model after the replacement or maintenance in a case where it is determined that the normal model before the replacement or maintenance can be used as the normal model after the replacement or maintenance, and acquires the sensor data and generates the normal model after the replacement or maintenance based on the sensor data in a case where it is determined that the normal model before the replacement or maintenance cannot be used as the normal model after the replacement or maintenance.

4. The substrate processing apparatus according to claim 1, characterized in that the specified step is a step of reducing a pressure of a processing chamber that processes the substrate from an atmospheric pressure to a predetermined pressure.

5. The substrate processing apparatus according to claim 1, characterized in that the precursor detection unit generates the normal model using the average value and the standard deviation of the amplitude of the vibration spectrum, compares, for the amount of the frequencies extracted, an amplitude value of the normal model with a threshold value decided in advance, and determines that there is a precursor of an abnormality in a case where the amplitude values of a predetermined number or more of frequencies deviate from the threshold value.

6. The substrate processing apparatus according to claim 5, characterized in that the threshold value is calculated within a range obtained by adding or subtracting a value of three times the standard deviation to or from the average value using the average value and the standard deviation.

7. The substrate processing apparatus according to claim 5, characterized in that ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ The precursor detection unit generates an alarm when an abnormal precursor is detected, and displays sensor data of a component that has detected the abnormal precursor on a screen.

8. A method of manufacturing a semiconductor device, having a substrate processing step that performs a process recipe including a plurality of steps to perform a predetermined process on a substrate, characterized by comprising: the substrate processing step has: a step of acquiring sensor data related to a component that is an object of detection of an abnormal precursor after replacement or maintenance of the component that is the object of detection of the abnormal precursor; a step of acquiring, from the collected sensor data, vibration data detected by a vibration sensor among the sensor data in a specified step among the steps constituting the process recipe; a step of transforming the acquired vibration data into a vibration spectrum; a step of extracting, at predetermined frequency intervals, frequencies of a predetermined range of the transformed vibration spectrum, and calculating, for each of the extracted frequencies, an average value and a standard deviation of an amplitude of the vibration spectrum using a predetermined number of amounts of data of the process recipe at normal times; a step of regenerating a normal model using the average value and the standard deviation of the amplitude of the vibration spectrum; and a step of detecting a precursor of an abnormality based on the state of the device before an abnormal stop of the device.

9. A computer-readable recording medium recording a precursor detection program that is executed by a substrate processing device that acquires sensor data related to a component that is an object of detection of an abnormal precursor to generate a normal model, and monitors a state of a device based on the normal model, characterized in that: the precursor detection program causes the substrate processing device to execute: a step of acquiring the sensor data after replacement or maintenance of the component that is the object of detection of the abnormal precursor, and acquiring, from the acquired sensor data, vibration data detected by a vibration sensor among the sensor data in a specified step among the steps constituting a process recipe; a step of transforming the acquired vibration data into a vibration spectrum; a step of extracting, at predetermined frequency intervals, frequencies of a predetermined range of the transformed vibration spectrum, and calculating, for each of the extracted frequencies, an average value and a standard deviation of an amplitude of the vibration spectrum using a predetermined number of amounts of data of the process recipe at normal times; a step of regenerating a normal model using the average value and the standard deviation of the amplitude of the vibration spectrum from the acquired sensor data; and a step of monitoring the state of the device based on the generated normal model, and detecting a precursor of an abnormality of the device.

10. A substrate processing device that acquires sensor data related to a component that is an object of detection of an abnormal precursor to generate a normal model, and monitors a state of a device based on the normal model, characterized in that: the substrate processing device includes a precursor detection unit configured to: acquire the sensor data after replacement or maintenance of the component that is the object of detection of the abnormal precursor, and regenerate the normal model from the sensor data again, monitoring a state of the device based on the normal model, detecting a precursor of abnormality before the device abnormally stops, the precursor detection section judges whether or not the normal model before the replacement or the maintenance can be used as the normal model after the replacement or the maintenance based on the sensor data of which the amount of data required to generate the normal model is less than the amount of data.

11. A computer-readable recording medium recording a precursor detection program which is executed by a substrate processing device that generates a normal model based on sensor data related to a component which is an abnormality precursor detection object and monitors a state of the device based on the normal model, characterized in that, the precursor detection program causes a precursor detection section to execute: after replacement or maintenance of the component which is the abnormality precursor detection object, a step of acquiring the sensor data and generating the normal model again from the sensor data; and a step of monitoring a state of the device based on the normal model, detecting a precursor of abnormality before the device abnormally stops, in the step of detecting the precursor of abnormality, the precursor detection section judges whether or not the normal model before the replacement or the maintenance can be used as the normal model after the replacement or the maintenance based on the sensor data of which the amount of data required to generate the normal model is less than the amount of data.

12. A method of manufacturing a semiconductor device, having: a process of generating a normal model based on sensor data related to a component which is an abnormality precursor detection object and monitoring a state of the device based on the normal model; and a process of processing a substrate, characterized in that, the process of monitoring the state of the device has: a process in which a precursor detection section generates the normal model again from the sensor data after replacement or maintenance of the component which is the abnormality precursor detection object; and the process in which the precursor detection section monitors the state of the device based on the normal model, detects a precursor of abnormality before the device abnormally stops, in the process of detecting the precursor of abnormality, the precursor detection section judges whether or not the normal model before the replacement or the maintenance can be used as the normal model after the replacement or the maintenance based on the sensor data of which the amount of data required to generate the normal model is less than the amount of data.

Images