Parameter correction device, parameter correction method, and computer program

Abstract

The invention provides a parameter correction device, a parameter correction method, and a computer program, which can reduce environmental load and rapidly correct discharge characteristics. The regression model construction unit uses, as an input, a feature quantity difference (dF), which is the difference between a reference feature quantity (F0) of a reference waveform measured using a reference parameter and peripheral feature quantities (F1-Fm) of peripheral waveforms (W1-Wm) measured using peripheral parameters (P1-Pm) obtained by changing the value of a part of the reference parameter. A regression model is constructed by machine learning of training data in which a parameter difference (dP), which is the difference between the reference parameter and the peripheral parameters (P1-Pm), is used as an output. The correction unit inputs, to the regression model, a feature quantity difference (F0 '-F0), which is the difference between a reference feature quantity (F0) of the reference waveform and an object feature quantity (F0') of the object waveform measured using the reference parameter after the reference waveform is measured, and corrects the reference parameter on the basis of a correction quantity output from the regression model.

Description

Technical Field

[0001] The subject matter disclosed in this specification relates to a parameter correction device, a parameter correction method, and a computer program. Background Art

[0002] In the manufacturing process of flat-panel displays, devices called coaters are used. Coating machines are substrate processing devices that use a pump to spray a treatment liquid from a slit nozzle, coating the entire substrate being transported with the treatment liquid. In recent years, with the advancement of product quality, such coaters have been required to apply the treatment liquid in a manner that ensures a uniform film thickness across the entire substrate. For example, in Patent Document 1, the discharge characteristics of the treatment liquid are repeatedly measured to adjust the parameters used to control the pump and achieve optimization.

[0003] That is, the optimization process of Patent Document 1 includes: a simulated ejection process, which ejects the processing liquid onto the outside of the substrate; a ejection characteristic measurement process, which measures the ejection characteristics of the processing liquid in the simulated ejection process; a state quantity derivation process, which derives the state quantity of the measured ejection characteristic deviating from the target characteristic; and a learning process, which performs machine learning on the changes in the state quantity accompanying the parameter change to construct a learning model. Moreover, while the state quantity exceeds the prescribed allowable range, the simulated ejection process, the ejection characteristic measurement process, the state quantity derivation process, and the learning process are repeatedly executed based on the parameter change based on the learning model. When the state quantity enters the allowable range, the last changed parameter is set as the parameter when the processing liquid is ejected in the processing liquid supply process.

[0004] Patent Document 1: Japanese Patent Application Laid-Open No. 2020-040046

[0005] However, even when spraying using optimized control parameters, the spray characteristics may sometimes be disturbed due to wear and degradation of components. In this case, if the control parameters are readjusted as described in Patent Document 1, restarting the process takes time, significantly reducing the yield. Furthermore, the large amount of process fluid consumed may increase the environmental impact. Summary of the Invention

[0006] An object of the present invention is to provide a technology capable of reducing environmental load and quickly correcting discharge characteristics.

[0007] In order to solve the above-mentioned problems, a first method is a parameter correction device for correcting parameters used to control a coating device so that the ejection characteristics measured when the processing liquid is ejected from the nozzle become the target ejection characteristics, comprising: a regression model construction unit, which constructs a regression model that outputs a correction amount for correcting the baseline parameter based on the characteristic amount difference by using machine learning of training data that takes as input the difference between the characteristic amount of the baseline ejection characteristic measured using the baseline parameter and the characteristic amount of the ejection characteristic measured using the peripheral parameter after changing the value of a part of the baseline parameter, i.e., the characteristic amount difference; and a correction unit, which inputs into the regression model the difference between the characteristic amount of the baseline ejection characteristic and the characteristic amount of the ejection characteristic measured using the baseline parameter after the time point when the baseline ejection characteristic is measured, and corrects the baseline parameter based on the correction amount output from the regression model.

[0008] A second aspect is the parameter correction device according to the first aspect, further comprising an abnormality determination unit that determines whether the discharge characteristic has an abnormality, and when the abnormality determination unit determines that the discharge characteristic has an abnormality, the correction unit corrects the reference parameter.

[0009] A third aspect provides the parameter correction device according to the first aspect or the second aspect, wherein the discharge characteristic is a discharge pressure applied to the processing liquid.

[0010] The fourth method is a parameter correction method for correcting parameters used to control a coating device so that the ejection characteristics measured when the processing liquid is ejected become target ejection characteristics, comprising the following steps: constructing a regression model that outputs a correction amount for correcting the baseline parameter based on the characteristic amount difference by using machine learning of training data that uses a difference between a characteristic amount of a baseline ejection characteristic measured using a baseline parameter and a characteristic amount of the ejection characteristic measured using a peripheral parameter after changing the value of a part of the baseline parameter, i.e., a characteristic amount difference, as input, and uses the difference between the baseline parameter and the peripheral parameter, i.e., a parameter difference, as output; and inputting the difference between the characteristic amount of the baseline ejection characteristic and the characteristic amount of the ejection characteristic measured using the baseline parameter after the time point when the baseline ejection characteristic is measured into the regression model, and correcting the baseline parameter based on the correction amount output from the regression model.

[0011] A fifth aspect is a computer program that causes the computer to execute the parameter calibration method of the fourth aspect.

[0012] According to the first to fifth aspects, even if the discharge characteristics measured using the reference parameters are disturbed, the disturbance in the discharge characteristics can be quickly corrected by correcting the reference parameters using the parameter differences calculated using the regression model as the correction amount. This prevents a reduction in the yield of the coating apparatus.

[0013] According to the parameter correction device of the second aspect, even when an abnormality occurs in the discharge characteristics using the reference parameters, the reference parameters can be appropriately corrected.

[0014] According to the parameter correction device of the third aspect, the reference parameter can be corrected based on the discharge pressure. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a diagram schematically showing the overall structure of a coating apparatus according to an embodiment.

[0016] Figure 2 Yes Figure 1 FIG. 1 is a diagram showing the structure of a processing liquid supply mechanism included in the coating apparatus shown.

[0017] Figure 3 Yes Figure 2 Graph showing an example of a movement pattern of the operating disk in the pump shown.

[0018] Figure 4 This is a block diagram showing a configuration example of a control unit.

[0019] Figure 5 This is a diagram showing an example of a discharge pressure waveform.

[0020] Figure 6 This is a diagram for explaining an example of feature amounts.

[0021] Figure 7 It is a diagram for explaining another example of feature amounts.

[0022] Figure 8 This is a diagram showing the flow of the process of constructing a regression model based on the control unit.

[0023] Figure 9 This is a diagram showing the flow of a reference parameter correction process executed by the control unit.

[0024] Figure 10 It is a block diagram schematically showing the functions of the control unit.

[0025] Description of Reference Numerals

[0026] 1: Coating device

[0027] 9: Control unit (parameter correction device)

[0028] 913: Feature calculation unit

[0029] 915: Abnormality Judgment Department

[0030] 917: Correction Department

[0031] 919: Regression Model Construction Department

[0032] 931: Computer Programs

[0033] W: Discharge pressure waveform (discharge characteristics)

[0034] Wst: Reference waveform (reference discharge characteristics)

[0035] Y: regression model DETAILED DESCRIPTION

[0036] The following describes embodiments of the present invention with reference to the accompanying drawings. The structural components described in these embodiments are merely illustrative and are not intended to limit the scope of the present invention to these figures. In the accompanying drawings, the dimensions and numbers of various components may be exaggerated or simplified as necessary for ease of understanding.

[0037] 1. Implementation Method

[0038] Figure 1 This figure schematically illustrates the overall structure of a coating apparatus 1 according to an embodiment. Coating apparatus 1 is a substrate processing apparatus that applies a processing liquid to the upper surface Sf of a substrate S. Coating apparatus 1 functions as a discharge characteristic monitoring device that monitors the discharge characteristics of the processing liquid. Specifically, discharge characteristics are physical quantities related to discharge, specifically, discharge pressure and discharge flow rate. The following description will illustrate the case where the discharge characteristic is discharge pressure.

[0039] The substrate S is, for example, a glass substrate for a liquid crystal display device. Alternatively, the substrate S may be a variety of substrates to be processed for electronic devices, such as semiconductor wafers, glass substrates for photomasks, glass substrates for plasma displays, glass or ceramic substrates for magneto-optical disks, glass substrates for organic EL devices, glass substrates or silicon substrates for solar cells, other flexible substrates, and printed circuit boards. The coating device 1 is, for example, a slit coater.

[0040] exist Figure 1In order to illustrate the configuration relationship of the various elements of the coating device 1, an XYZ coordinate system is defined. The conveying direction of the substrate S is the "X direction". The direction in which the substrate S moves in the X direction (toward the downstream side of the conveying direction) is the +X direction, and the opposite direction (toward the upstream side of the conveying direction) is the -X direction. In addition, the direction orthogonal to the X direction is the Y direction, and the direction orthogonal to the X and Y directions is the Z direction. In the following description, the Z direction is set to the vertical direction, and the X and Y directions are set to the horizontal directions. In the Z direction, the +Z direction is set to the upper direction, and the -Z direction is set to the lower direction. In addition, these directions are not intended to limit the configuration of the coating device.

[0041] The coating apparatus 1 includes, in order from the +X direction, an input conveyor 100, an input transfer unit 2, a floating stage unit 3, an output transfer unit 4, and an output conveyor 110. The input conveyor 100, the input transfer unit 2, the floating stage unit 3, the output transfer unit 4, and the output conveyor 110 form a transport path for the substrate S. The coating apparatus 1 also includes a substrate transport unit 5, a coating mechanism 7, a processing liquid supply mechanism 8, and a control unit 9.

[0042] The substrate S is conveyed to the input conveyor 100 from an apparatus upstream of the coating apparatus 1. The input conveyor 100 includes a roller conveyor 101 and a rotary drive mechanism 102. The rotary drive mechanism 102 rotates the rollers of the roller conveyor 101. The rotation of the rollers of the roller conveyor 101 causes the substrate S to be conveyed downstream (in the +X direction) in a horizontal position. The "horizontal position" refers to a state in which the main surface (the surface with the largest area) of the substrate S is parallel to the horizontal plane (XY plane).

[0043] The input transfer unit 2 includes a roller conveyor 21 and a rotation and lifting drive mechanism 22. The rotation and lifting drive mechanism 22 rotates the rollers of the roller conveyor 21 and raises and lowers the roller conveyor 21. The rotation of the roller conveyor 21 transports the substrate S in a horizontal position downstream (in the +X direction). Furthermore, the raising and lowering of the roller conveyor 21 changes the Z-direction position of the substrate S. The substrate S is transferred from the input conveyor 100 to the floating stage 3 via the input transfer unit 2.

[0044] like Figure 1As shown, the floating workbench portion 3 is roughly flat. The floating workbench portion 3 is divided into three parts along the X direction. The floating workbench portion 3 has an inlet floating workbench 31, a coating workbench 32 and an outlet floating workbench 33 in sequence toward the +X direction. The upper surface of the inlet floating workbench 31, the upper surface of the coating workbench 32 and the upper surface of the outlet floating workbench 33 are on the same plane. The floating workbench portion 3 also has a lifting pin drive mechanism 34, a floating control mechanism 35 and a lifting drive mechanism 36. The lifting pin drive mechanism 34 lifts and lowers a plurality of lifting pins arranged on the inlet floating workbench 31. The floating control mechanism 35 supplies compressed air for floating the substrate S to the inlet floating workbench 31, the coating workbench 32 and the outlet floating workbench 33. The lifting drive mechanism 36 lifts and lowers the outlet floating workbench 33.

[0045] A plurality of ejection holes are arranged in a matrix on the upper surfaces of the inlet flotation stage 31 and the outlet flotation stage 33. These ejection holes eject compressed air supplied by the flotation control mechanism 35. When compressed air is ejected from each ejection hole, the substrate S floats upward relative to the flotation stage 3. As a result, the lower surface Sb of the substrate S separates from the upper surface of the flotation stage 3, and the substrate S is supported in a horizontal position. When the substrate S is in the flotation state, the distance (floatation amount) between the lower surface Sb of the substrate S and the upper surface of the flotation stage 3 is, for example, not less than 10 μm and not more than 500 μm.

[0046] The upper surface of the coating workbench 32 is provided with a spray hole for spraying compressed air supplied from the floating control mechanism 35 and a suction hole for sucking gas. The spray hole and the suction hole are alternately arranged in the X direction and the Y direction. The floating control mechanism 35 controls the amount of compressed air sprayed from the spray hole and the amount of air sucked from the suction hole. Thus, the floating amount of the substrate S relative to the coating workbench 32 is precisely controlled in such a manner that the position in the Z direction of the upper surface Sf of the substrate S above the coating workbench 32 is a specified value. In addition, the floating amount of the substrate S relative to the coating workbench 32 is calculated by the control unit 9 based on the detection result of the sensor 61 or the sensor 62 described later. In addition, the floating amount of the substrate S relative to the coating workbench 32 is preferably adjustable with high precision by airflow control.

[0047] The substrate S loaded onto the floatation stage 3 is propelled in the +X direction by the roller conveyor 21 and is transported to the entrance floatation stage 31. The entrance floatation stage 31, the coating stage 32, and the exit floatation stage 33 support the substrate S in a floating state. For example, the structure described in Japanese Patent No. 5346643 may be used as the floatation stage 3.

[0048] The substrate transport unit 5 is positioned below the floating stage 3. It includes a chuck mechanism 51 and a suction and travel control mechanism 52. The chuck mechanism 51 includes a suction pad (not shown) attached to a suction member. The chuck mechanism 51 supports the substrate S from below by bringing the suction pad into contact with the peripheral edge of the lower surface Sb of the substrate S. The suction and travel control mechanism 52 applies negative pressure to the suction pad, thereby sucking the substrate S onto the suction pad. Furthermore, the suction and travel control mechanism 52 causes the substrate transport unit 5 to reciprocate in the X direction.

[0049] The chuck mechanism 51 holds the substrate S with its lower surface Sb positioned higher than the upper surface of the float stage 3 . With the peripheral edge portion held by the chuck mechanism 51 , the substrate S maintains a horizontal posture due to the buoyancy applied from the float stage 3 .

[0050] like Figure 1 As shown, the coating apparatus 1 includes a sensor 61 for measuring the plate thickness. The sensor 61 is disposed near the roller conveyor 21. The sensor 61 detects the Z-direction position of the upper surface Sf of the substrate S held by the chuck mechanism 51. Furthermore, by positioning the chuck (not shown) without holding the substrate S directly below the sensor 61, the sensor 61 can detect the vertical Z-direction position of the upper surface of the adsorption member, i.e., the adsorption surface.

[0051] The chuck mechanism 51 holds the substrate S loaded onto the float stage 3 and moves in the +X direction. As a result, the substrate S is transported from above the entrance float stage 31, through above the coating stage 32, to above the exit float stage 33. The substrate S then moves from the exit float stage 33 to the output transfer unit 4.

[0052] The output transfer unit 4 moves the substrate S from a position above the exit float table 33 to the output conveyor 110. The output transfer unit 4 includes a roller conveyor 41 and a rotation and elevation drive mechanism 42. The rotation and elevation drive mechanism 42 rotates the roller conveyor 41 and raises and lowers it in the Z direction. The rotation of the rollers of the roller conveyor 41 moves the substrate S in the +X direction. Furthermore, the raising and lowering of the roller conveyor 41 changes the position of the substrate S in the Z direction.

[0053] The output conveyor 110 includes a roller conveyor 111 and a rotation drive mechanism 112. By rotating the rollers of the roller conveyor 111, the output conveyor 110 transports the substrate S in the +X direction and discharges the substrate S out of the coating apparatus 1. The input conveyor 100 and the output conveyor 110 are part of the coating apparatus 1. However, the input conveyor 100 and the output conveyor 110 may also be incorporated into a separate apparatus from the coating apparatus 1.

[0054] The coating mechanism 7 applies the treatment liquid to the upper surface Sf of the substrate S. The coating mechanism 7 is arranged above the conveying path of the substrate S. The coating mechanism 7 has a nozzle 71. The nozzle 71 is a slit nozzle having a slit-shaped nozzle on the lower surface. The nozzle 71 is connected to a positioning mechanism (not shown). The positioning mechanism positions the nozzle 71 at a coating position ( Figure 1 The processing liquid supply mechanism 8 is connected to the nozzle 71. The processing liquid supply mechanism 8 supplies the processing liquid to the nozzle 71, and the processing liquid is ejected from the ejection port arranged on the lower surface of the nozzle 71.

[0055] Figure 2 Yes Figure 1 FIG. 8 is a diagram showing the structure of a treatment liquid supply mechanism 8 of the coating device 1 shown in FIG. The treatment liquid supply mechanism 8 includes a pump 81, a pipe 82, a treatment liquid replenishing unit 83, a pipe 84, an opening and closing valve 85, a pressure sensor 86, and a driving unit 87. The pump 81 is a supply source for supplying the treatment liquid to the nozzle 71, and supplies the treatment liquid by utilizing volume changes. As the pump 81, for example, a bellows-type pump described in Japanese Patent Application Laid-Open No. 10-61558 can be used. Figure 2 As shown, the pump 81 includes a flexible tube 811 that is elastically expandable and contractible in the radial direction. One end of the flexible tube 811 is connected to the processing liquid replenishing unit 83 via a pipe 82 , and the other end of the flexible tube 811 is connected to the nozzle 71 via a pipe 84 .

[0056] Pump 81 includes a bellows 812 that is elastically deformable in the axial direction. Bellows 812 includes a small bellows portion 813, a large bellows portion 814, a pump chamber 815, and a working disc portion 816. Pump chamber 815 is disposed between flexible tube 811 and bellows 812. An incompressible medium is enclosed in pump chamber 815. Working disc portion 816 is connected to drive unit 87.

[0057] The treatment liquid replenishing unit 83 includes a storage tank 831 for storing treatment liquid. The storage tank 831 is connected to the pump 81 via a pipe 82. An on-off valve 833 is installed in the pipe 82. The on-off valve 833 opens and closes according to commands from the control unit 9. When the on-off valve 833 is open, the treatment liquid can be replenished from the storage tank 831 to the flexible tube 811 of the pump 81. When the on-off valve 833 is closed, the supply of treatment liquid from the storage tank 831 to the flexible tube 811 of the pump 81 is restricted.

[0058] The pipe 84 is connected to the output side of the pump 81. An on-off valve 85 is provided on the pipe 84. The on-off valve 85 opens and closes the pipe 84 in response to commands from the control unit 9. The opening and closing of the on-off valve 85 switches the supply of the treatment liquid to the nozzle 71 between on and off. A pressure sensor 86 is provided on the pipe 84. The pressure sensor 86 detects the pressure (discharge pressure) applied to the treatment liquid supplied to the nozzle 71 and outputs a signal indicating the detected pressure value to the control unit 9.

[0059] Figure 3 Yes Figure 2 FIG. 8 is a graph showing an example of a movement pattern of the working disc portion 816 in the pump 81 shown in FIG. Figure 3 In FIG. 8 , the horizontal axis represents time, and the vertical axis represents the moving speed of the working disk portion 816. The driving portion 87 uses the command from the control unit 9 to Figure 3 The movement pattern shown (a pattern indicating a change in the speed of the working disc portion 816 relative to the passage of time) causes the working disc portion 816 to be displaced in the axial direction. The displacement of the working disc portion 816 causes the volume of the inner side of the bellows 812 to change. As a result, the flexible tube 813 expands and contracts in the radial direction to perform a pumping action, and the treatment liquid supplied from the treatment liquid replenishing unit 83 is supplied to the nozzle 71. The movement pattern of the working disc portion 816 is closely related to the ejection characteristics of the treatment liquid ejected from the nozzle 71. Therefore, the ejection pressure waveform indicating the time change of the ejection pressure corresponding to the movement pattern of the working disc portion 816 is measured (refer to Figure 5 As the discharge pressure increases or decreases, the discharge amount (the amount of the processing liquid discharged from the nozzle 71) also increases or decreases. The discharge pressure waveform is an example of discharge characteristics when the nozzle 71 discharges the processing liquid.

[0060] In this embodiment, by adjusting various parameters that determine the movement of the working disk portion 816 (acceleration time, stabilization speed, stabilization speed time, deceleration time, etc.), appropriate optimization processing is performed to make the spray characteristics of the processing liquid sprayed from the nozzle 71 (specifically, the time change of the spray speed (spray pressure)) consistent with or close to the desired target characteristics.

[0061] like Figure 1 and Figure 2 As shown, a sensor 62 is arranged on the nozzle 71 to which the processing liquid is supplied from the processing liquid supply mechanism 8. The sensor 62 detects the height of the substrate S in the Z direction in a non-contact manner. The sensor 62 is electrically connected to the control unit 9. Based on the detection result of the sensor 62, the control unit 9 measures the distance (separation distance) between the floating substrate S and the upper surface of the coating workbench 32. Then, the control unit 9 adjusts the coating position of the nozzle 71 by the positioning mechanism based on the measured separation distance. In addition, as the sensor 62, an optical sensor or an ultrasonic sensor can be applied.

[0062] The coating mechanism 7 includes a nozzle cleaning standby unit 72. The nozzle cleaning standby unit 72 performs prescribed maintenance on the nozzle 71 positioned in the maintenance position. The nozzle cleaning standby unit 72 includes a roller 721, a cleaning portion 722, and a roller groove 723. The nozzle cleaning standby unit 72 cleans the nozzle 71 and forms a liquid accumulation, thereby adjusting the discharge port of the nozzle 71 to a state suitable for coating processing. Furthermore, in order to evaluate the discharge pressure applied to the process liquid, the coating apparatus 1 performs a simulated discharge of the process liquid from the nozzle 71 while the nozzle 71 is positioned in the maintenance position.

[0063] Figure 4 1 is a block diagram showing a structural example of a control unit 9. The control unit 9 controls the actions of the various structural components in the coating device 1. A computer can be used as the control unit 9. The control unit 9 has a processor 91 and a memory 93. The processor has, for example, a CPU (Central Processing Unit). The memory 93 has a temporary storage device such as a RAM (Random Access Memory). In addition, the memory 93 may also have a non-temporary storage device such as an HDD (Hard Disk Drive) or an SSD (Solid State Drive). The memory 93 is connected to the processor 91 via bus wiring.

[0064] The control unit 9 includes a display device 95 for displaying various information and an input device 97 for accepting user input commands. The display device 95 and the input device 97 are connected to the processor 91 via a wiring bus. The display device 95 is, for example, a liquid crystal display. The input device 97 is, for example, a mouse or a keyboard. Furthermore, by including a touch panel in the display device 95, the display device 95 can function as an input device.

[0065] The memory 93 stores a computer program 931. The computer program 931 is provided to the control unit 9 via a recording medium M. Specifically, the recording medium M stores the computer program 931 so that it can be read by the computer, i.e., the control unit 9. Specifically, the recording medium M is a USB (Universal Serial Bus) memory, an optical disc such as a DVD (Digital Versatile Disc), or a magnetic disk.

[0066] The processor 91 executes the computer program 931 to function as a discharge control unit 910 , a discharge pressure measurement unit 911 , a feature value calculation unit 913 , an abnormality determination unit 915 , a correction unit 917 , and a regression model construction unit 919 .

[0067] The discharge control unit 910 controls the operation (supply operation) of the pump 81 that supplies the processing liquid to the nozzle 71 based on pre-set control parameters. In the coating apparatus 1, in order to coat the processing liquid discharged from the nozzle 71 on the upper surface Sf of the substrate S with a uniform film thickness, the control parameters closely related to the discharge pressure waveform are optimized in advance before the start of production (or mass production) of the substrate S, so that the discharge pressure waveform has an ideal shape.

[0068] The control parameter is, for example, a set value that specifies the movement of the work plate portion 816 and can be used Figure 3 The following set values ​​for pump control are used as control parameters to be optimized.

[0069] Stable speed V1

[0070] Acceleration time T1: the time from a stopped state to a stable speed V1

[0071] Stable speed time T2: the time for the stable speed V1 to last

[0072] Stable speed V2

[0073] Deceleration time T3: the time it takes to decelerate from the stable speed V1 to the stable speed V2

[0074] Stable speed time T4: the time for the stable speed V2 to last

[0075] Stable speed V3

[0076] Acceleration time T5: the time it takes to accelerate from stable speed V2 to stable speed V3

[0077] Stable speed time T6: the time for the stable speed V3 to continue

[0078] Stable speed V4

[0079] Deceleration time T7: the time it takes to decelerate from the stable speed V3 to the stable speed V4

[0080] Stable speed time T8: the time for the stable speed V4 to last

[0081] Stable speed V5

[0082] Acceleration time T9: Time to accelerate from stable speed V4 to stable speed V5

[0083] Stable speed time T10: the time for the stable speed V5 to last

[0084] Deceleration time T11: The time it takes to decelerate from the stable speed V5 to a stop state

[0085] return Figure 4The discharge pressure measurement unit 911 measures the discharge pressure. Specifically, the discharge pressure measurement unit 911 acquires the discharge pressure measured by the pressure sensor 86 at a predetermined sampling period. The discharge pressure measurement unit 911 acquires a discharge pressure waveform representing the temporal variation in the discharge pressure applied to the processing liquid during discharge of the processing liquid from the nozzle 71, and appropriately stores discharge pressure data representing the acquired discharge pressure waveform in the memory 93. The discharge pressure data is time-series data representing the pressure measured at each moment.

[0086] Figure 5 : is a diagram showing an example of the discharge pressure waveform W. Figure 5 In the figure, the horizontal axis represents time and the vertical axis represents ejection pressure. Figure 5 As shown, the discharge pressure is measured during a predetermined period from before the start of discharge of the processing liquid from the nozzle 71 to after the end of discharge of the processing liquid from the nozzle 71 .

[0087] exist Figure 5 In the example shown, the discharge pressure at time ta when the treatment liquid starts to be discharged from the nozzle 71 and the discharge pressure at time te when the treatment liquid ends to be discharged from the nozzle 71 are the initial pressure Pi. The pressure at the start of discharge and the pressure at the end of discharge are not always the same as the initial pressure Pi. The discharge period from time ta to time te is divided into a rising period Tab, a transition period Tbc, a stable period Tcd, and a falling period Tde.

[0088] The rising period Tab is the period from time ta, when the processing liquid supply mechanism 8 begins discharging the processing liquid from the nozzle 71, to time tb, when the discharge pressure reaches the target pressure Pt. Time ta is the time when the processing liquid supply mechanism 8 begins moving the operating disk 816. That is, when the processing liquid begins to be discharged from the nozzle 71 at time ta, the discharge pressure increases from the initial pressure Pi to the target pressure Pt during the period from time ta to time tb.

[0089] The transition period Tbc is a period from time tb to time tc after a predetermined vibration damping period has passed. The vibration damping period is a period required for the temporal change in the discharge pressure to stabilize.

[0090] The stable period Tcd is the period from time tc to time td, when the processing liquid supply mechanism 8 begins to reduce the discharge pressure. Time td is the time when the processing liquid supply mechanism 8 begins to decelerate from the target speed of the working disk 816. The processing liquid supply mechanism 8 moves the working disk 816 at a constant speed from time tc to time td, and begins to decelerate the working disk 816 at time td. During the stable period Tcd, the discharge pressure remains essentially stable at the target pressure Pt. However, during the stable period Tcd, the temporal variation in the discharge pressure also includes slight fluctuations, causing the discharge pressure to increase or decrease relative to the target pressure Pt.

[0091] The falling period Tde is the period from time td to time te, when the processing liquid supply mechanism 8 stops discharging the processing liquid from the nozzle 71. That is, time te is the time when the processing liquid supply mechanism 8 stops the operating disk portion 816. That is, the discharge pressure decreases to the initial pressure Pi from time td to time te, and the discharge of the processing liquid from the nozzle 71 stops at time te.

[0092] In the coating apparatus 1, before starting the production (or mass production) of the substrate S, the control parameters are adjusted in advance so that the discharge pressure waveform W becomes the target reference waveform Wst (for example, Figure 5 Specifically, the control parameters are adjusted by repeating a series of processes, including simulated ejection using the control parameters, measurement of the ejection pressure waveform during the simulated ejection, evaluation of the measured ejection pressure waveform, and updating of the control parameters based on the evaluation. For example, the method described in Patent Document 1 can be used to update the control parameters. Hereinafter, the control parameters adjusted so that the ejection pressure waveform matches the reference waveform Wst are referred to as "reference parameters P0."

[0093] When the substrate S is generated, the coating device 1 uses the reference parameter P0 to perform the coating process. That is, the adjusted reference parameter P0 is used by the discharge control unit 910 to control the pump, and the processing liquid is discharged to the substrate S according to the reference waveform Wst. However, even if the coating is performed with the reference parameter P0, due to various reasons such as wear and degradation of components such as the pump 81 over time, or the influence of interference, disturbances (abnormalities) may sometimes occur in the discharge pressure waveform. That is, the discharge pressure waveform may have a shape that is different from the reference waveform Wst. If disturbance occurs in the discharge pressure waveform, it is difficult to perform normal coating. Therefore, the control unit 9 corrects the reference parameter P0, for example, by periodically performing simulated discharges to measure the discharge pressure waveform, so that the measured discharge pressure waveform is close to the original reference waveform Wst. In this way, the control unit 9 functions as a "parameter correction device".

[0094] return Figure 4The feature quantity calculation unit 913 calculates the feature quantity based on the discharge pressure data measured by the discharge pressure measurement unit 911 . Figure 6 This is a diagram illustrating an example of a characteristic quantity. In this example, the degree of overshoot generated during the rise of the discharge pressure waveform W is calculated as a characteristic quantity. Specifically, the characteristic quantity calculation unit 913 calculates the sign (positive or negative) of the secondary differential value Dif2 of the discharge pressure at time t11, when the discharge pressure reaches its maximum value Pmax. The characteristic quantity calculation unit 913 then calculates time t12, when the sign of the secondary differential value Dif2 switches twice from the sign at time t11. Furthermore, the characteristic quantity calculation unit 913 calculates the characteristic quantity of the time variation of the discharge pressure during the initial vibration period T2_s from time t11 to t12.

[0095] For example, the characteristic quantity calculation unit 913 selects the smaller of the minimum discharge pressure value Pmin during the initial vibration period T2_s and the stable pressure Pm (the average discharge pressure value during the stable period T3) as the target pressure Pg. Alternatively, the characteristic quantity calculation unit 913 may calculate the difference between the maximum value Pmax and the target pressure Pg (= Pmax - Pg) as the characteristic quantity. Due to differences in the momentum of the discharge pressure increase, the greater the overshoot of the temporal change in the discharge pressure, the larger the characteristic quantity.

[0096] Figure 7 This figure illustrates another example of a feature quantity. In this example, the stability of the temporal variation of the discharge pressure during the transition period Tbc is calculated as a feature quantity. Specifically, the feature quantity calculation unit 913 calculates the root mean square error (RMSE(P_measure, Pm)) between the discharge pressure during the transition period Tbc and the average of the discharge pressures during the stable period Tcd, i.e., the stable pressure Pm, as the feature quantity. The greater the ringing (oscillation) of the discharge pressure during the transition period Tbc, the larger this feature quantity becomes.

[0097] The feature quantity calculated by the feature quantity calculation unit 913 is not limited to Figure 6 and Figure 7 As the feature amount, for example, several feature amounts described in Japanese Patent Application Laid-Open No. 2022-138109 can be used.

[0098] return Figure 4 The abnormality determination unit 915 determines whether the discharge pressure waveform W measured by the discharge pressure measuring unit 911 is abnormal. As described above, in order to correct the reference parameter P0, the control unit 9 determines whether an abnormality has occurred in the discharge pressure waveform (hereinafter also referred to as "target waveform Wst'") measured by the simulated discharge using the reference parameter P0.

[0099] For example, the abnormality determination unit 915 calculates the abnormality of the object waveform Wst′ and performs abnormality determination based on the abnormality. The abnormality is a value that indicates the degree of deviation from the distribution of the characteristic quantities of the normal ejection pressure waveform. In the calculation of the abnormality, a learned model constructed by taking the characteristic quantities of the ejection pressure waveform as input and the abnormality as output can be used. Such a learned model is constructed by performing machine learning based on the k-nearest neighbor method using a group of characteristic quantities of a plurality of normal ejection pressure waveforms as learning data. The abnormality is preferably the Mahalanobis distance from the nearest k, but may also be the Euclidean distance or the Manhattan distance.

[0100] Alternatively, the abnormality determination unit 915 may determine the presence of an abnormality in the target waveform Wst′ based on the probability of the target waveform Wst′ matching an abnormality. Furthermore, the target waveform Wst′ can be classified into a normal state (i.e., a state close to the reference waveform Wst) and one or more pre-conceptually assumed abnormal states. Therefore, the abnormality determination unit 915 may calculate the probability of matching (hereinafter also referred to as "abnormality probability") for each of several pre-conceptually assumed abnormalities.

[0101] In the calculation of the abnormality probability, a learned model constructed in a manner that uses a characteristic value as input to output the abnormality probability can be used. In the case of constructing such a learned model, first, several characteristic values ​​of normal pressure ejection waveforms (hereinafter also referred to as "normal characteristic values") and a considerable number of characteristic values ​​of ejection pressure waveforms of each type of abnormality are prepared. Then, in order to distinguish between normal and one or more abnormal states, target values ​​for each state are set, and the learning of the model is promoted in a manner that the model outputs a target value corresponding to the input data. In addition, a linear regression model can be used as a basic model. In the case of the learned model obtained in this way, the closer the output value is to a specific target value, the higher the probability of conforming to the abnormality corresponding to the specific target value.

[0102] Alternatively, the abnormality determination unit 915 may calculate both the abnormality degree and the abnormality probability, and determine the presence or absence of an abnormality based on these results.

[0103] The correction unit 917 corrects the reference parameter P0. Furthermore, in the coating apparatus 1, the abnormality determination unit 915 determines whether an abnormality (disturbance) has occurred in the discharge pressure waveform W. Furthermore, if the abnormality determination unit 915 determines that an abnormality has occurred, the correction unit 917 corrects the reference parameter P0. However, the abnormality determination unit 915 may be omitted. In this case, the correction unit 917 can correct the reference parameter P0 based on the measured discharge pressure waveform W regardless of whether an abnormality has occurred.

[0104] As will be described later, the correction unit 917 calculates the difference between the reference feature value F0 of the reference waveform Wst and the feature value F0′ of the target waveform Wst′, i.e., the feature value difference F0′-F0. The correction unit 917 then inputs the feature value difference F0′-F0 into the regression model Y, calculates the correction value, i.e., the parameter difference ΔP, and adds the obtained parameter difference ΔP to the reference parameter P0, thereby obtaining the correction parameter P. C .

[0105] The regression model construction unit 919 constructs the regression model Y. For the construction method of the regression model Y, refer to Figure 8 Provide explanation. Figure 8 1 is a diagram showing the flow of the process of constructing the regression model Y performed by the control unit 9 .

[0106] First, the regression model building unit 919 creates a plurality of peripheral parameters P1, P2, ..., P m (m is the number of peripheral parameters) (Step S11). Peripheral parameters P1 to P m It is a parameter that changes at least one of the values ​​included in the reference parameter P0 by a small amount. For example, when the reference parameter P0 has n values ​​(p 10 、p 20 ,…,p n0 ) (n is the number of parameter values), by changing some or all of the values ​​by a small amount, m peripheral parameters P1 to P1 that are different from each other are created. m .

[0107] Next, the coating device 1 uses the peripheral parameters P1 to P m Simulate the ejection and measure the ejection pressure waveform (step S12). m The corresponding plurality of discharge pressure waveforms are discharge pressure data of the peripheral waveforms W1, W2, ..., Wm.

[0108] Next, the feature quantity calculation unit 913 calculates the reference feature quantity F0 of the reference waveform Wst and the feature quantities of the surrounding waveforms W1 to Wm, namely, the peripheral feature quantities F1, F2, ..., F m (Step S13) Then, the regression model building unit 919 calculates the surrounding parameters P1 to P m For each of them, create differential data sets D1, D2, ..., D m (Step S14). Each differential data set D1 to D m By parameter difference d P and the feature difference d F group composition.

[0109] Parameter difference d PThe reference parameter P0 and the peripheral parameters P1 to P m For example, let the reference parameter P0 be (p 10 、p 20 ,…,p n0 )(n is the number of parameters), set the peripheral parameter P1 to (p 11 、p 21 ,…,p n1 ), as shown in the following formula, the parameter difference d is calculated by subtracting the values ​​of the peripheral parameter P1 from the values ​​of the reference parameter P0. P1 .

[0110] d P1 =(p 10 -p 11 、p 20 -p 21 ,…,p n0 -p n1 )

[0111] In addition, the feature difference d F The reference feature quantity F0 of the reference waveform Wst and the peripheral feature quantities F1 to F0 of the peripheral waveforms W1 to Wm are m Specifically, by calculating the difference between the surrounding feature values ​​F1 to F m For example, let the reference feature quantity F0 be (f 10 、f 20 ,…,f k0 )(k is the number of feature quantities), the surrounding feature quantity F1 is set to (f 11 、f 21 ,…,f k1 ), as shown in the following formula, the feature quantity difference d is calculated by subtracting each feature quantity of the reference feature quantity F0 from each feature quantity of the surrounding feature quantity F1. F1 .

[0112] d F1 =(f 11 -f 10 、f 21 -f 20 ,…,f k1 -f k0 )

[0113] Next, the regression model building unit 919 uses a plurality of difference data sets D1 to D m The regression model Y is constructed by machine learning as training data. That is, the regression model construction unit 919 performs training learning and transforms the difference data sets D1 to D m The characteristic difference d in F1 ~d Fm As input, the differential data set D1~Dm The parameter difference d in P1 ~d Pm As output. For example, a random forest or a neural network can be used as the regression model Y. The learned regression model Y can predict the correction amount (parameter difference) of the reference parameter P0 required to eliminate the feature quantity difference based on the feature quantity difference. The parameters of the regression model Y obtained by machine learning (learned parameters) are appropriately stored in the memory 93.

[0114] <Calibration of reference parameters>

[0115] Next, the process of correcting the reference parameter P0 using the constructed regression model Y will be described. Figure 9 2 is a diagram showing the flow of the correction process executed by the control unit 9 . Figure 9 The calibration process shown is performed, for example, periodically or based on an operator's instruction input during the coating process using the adjusted reference parameter P0.

[0116] First, the control unit 9 performs simulated discharge using the reference parameter P0 and measures the discharge pressure to obtain a target waveform Wst' (step S21). The target waveform Wst' is an example of a discharge characteristic measured after the reference waveform Wst, which is a reference discharge characteristic, is measured.

[0117] Next, the abnormality determination unit 915 of the control unit 9 determines whether the target waveform Wst′ is abnormal (step S22). If, in step S22, it is determined that there is no abnormality in the target waveform Wst′, the control unit 9 ends the correction process. On the other hand, if, in step S22, it is determined that there is an abnormality in the target waveform Wst′, the feature quantity calculation unit 913 calculates the feature quantity of the target waveform Wst′, namely, the target feature quantity F0′ (step S23).

[0118] Next, the correction unit 917 of the control unit 9 uses the object feature quantity F0′ calculated in step S23 to correct the reference parameter P0 (step S24). Specifically, the correction unit 917 calculates the difference between the object feature quantity F0′ and the reference feature quantity F0, that is, the feature quantity difference F0′-F0. The correction unit 917 inputs the calculated feature quantity difference F0′-F0 into the regression model Y to obtain the correction quantity, that is, the parameter difference ΔP. The correction unit 917 calculates the correction parameter P by adding the calculated parameter difference ΔP to the reference parameter P0. C .

[0119] When calculating the correction parameter P C When the correction unit 917 corrects the parameter P C The correction parameter P is stored in the memory 93. Then, the discharge control unit 910 uses the correction parameter P CThe substrate S is subjected to a coating process.

[0120] Figure 10 Schematically shows the function of the control unit 9. Figure 8 As described in the previous section, in the construction phase of the regression model Y, the reference feature value F0 of the reference waveform Wst and the surrounding waveforms W1 to W m The surrounding feature values ​​F1 to F m The feature quantity calculation unit 913 calculates the feature quantity difference d obtained from the delivered feature quantity and sends it to the regression model construction unit 919. F and parameter difference d P The combination of the differential data sets D1~D m Use it as training data for machine learning to build a regression model Y.

[0121] In addition, in the application stage of the regression model building unit 919, as in Figure 9 As described in , the object feature quantity F0′ of the object waveform Wst′ measured using the reference parameter P0 is calculated by the feature quantity calculation unit 913 and delivered to the correction unit 917. The correction unit 917 then inputs the feature quantity difference F0′-F0 between the object feature quantity F0′ and the reference feature quantity F0 into the regression model Y. The correction unit 917 then corrects the reference parameter P0 using the correction quantity, i.e., the parameter difference ΔP, output by the regression model Y, to generate the correction parameter P C .

[0122] Effects

[0123] Even if disturbances occur in the target waveform Wst′ measured using the reference parameter P0, the control unit 9 can correct the disturbances in the target waveform Wst′ by using the parameter difference ΔP calculated using the regression model Y as a correction factor to correct the reference parameter P0. This eliminates the need to readjust the reference parameter P0 during repeated simulated ejection, thereby reducing environmental load and correcting disturbances in the target waveform Wst′. This prevents a decrease in the yield of the coating apparatus 1.

[0124] <2. Modifications>

[0125] As mentioned above, although embodiment was described, this invention is not limited to what was described above, Various deformation|transformation is possible.

[0126] For example, in the above-described embodiment, the parameter correction device is configured by the control unit 9 included in the coating device 1. However, the parameter correction device may be configured by a computer device separate from the coating device 1.

[0127] In the above embodiment, the discharge pressure is measured as the discharge characteristic, but the discharge characteristic is not limited to the discharge pressure. For example, the discharge flow rate of the processing liquid discharged from the nozzle 71 may be measured as the discharge characteristic. In this case, for example, a flow meter may be provided on the pipe 84 to measure the discharge flow rate.

[0128] Although the present invention has been described in detail, the above description is illustrative in all aspects and the present invention is not limited thereto. It should be understood that numerous modifications not shown can be envisioned without departing from the scope of the present invention. The various structures described in the above embodiments and modifications may be appropriately combined or omitted as long as they do not conflict with each other.

Claims

1. A parameter correction device, wherein: Correcting the parameters for controlling the coating device so that the discharge characteristics measured when the treatment liquid is discharged from the nozzle become the target discharge characteristics, have: a regression model construction unit, which constructs a regression model that outputs a correction amount for correcting the baseline parameter based on the feature quantity difference by machine learning using training data having as input a feature quantity of a baseline discharge characteristic measured using a baseline parameter and a feature quantity of the discharge characteristic measured using a peripheral parameter after a portion of the baseline parameter has been changed; and outputs a parameter difference, which is a difference between the baseline parameter and the peripheral parameter; as well as The correction unit inputs a difference between a feature quantity of the reference discharge characteristic and a feature quantity of the discharge characteristic measured using the reference parameter after the reference discharge characteristic is measured into the regression model, and corrects the reference parameter based on the correction quantity output from the regression model.

2. The parameter correction device according to claim 1, wherein: It also includes an abnormality determination unit for determining whether the ejection characteristics are abnormal. The correction unit corrects the reference parameter when the abnormality determination unit determines that the discharge characteristic has an abnormality.

3. The parameter correction device according to claim 1 or 2, wherein: The discharge characteristic is the discharge pressure applied to the processing liquid.

4. A parameter correction method, wherein: Correcting the parameters for controlling the coating device so that the discharge characteristics measured when the treatment liquid is discharged become the target discharge characteristics, The process includes the following steps: By using machine learning that takes as input a feature quantity of a baseline ejection characteristic measured using a baseline parameter and a feature quantity of the ejection characteristic measured using a peripheral parameter after a portion of the baseline parameter has been changed, i.e., a feature quantity difference, and outputs as output a parameter difference, i.e., a difference between the baseline parameter and the peripheral parameter, a regression model is constructed that outputs a correction quantity for correcting the baseline parameter based on the feature quantity difference; as well as The difference between the characteristic amount of the reference discharge characteristic and the characteristic amount of the discharge characteristic measured using the reference parameter after the reference discharge characteristic is measured is input to the regression model, and the reference parameter is corrected based on the correction amount output from the regression model.

5. A computer program, wherein The computer is caused to execute the parameter correction method according to claim 4.

Images