Parameter correction device, parameter correction method, and computer program

Abstract

To provide a technique capable of quickly correcting discharge characteristics while reducing an environmental load.SOLUTION: A regression model construction unit 919 constructs a regression model Y, by utilizing machine learning that uses training data which uses as input a feature difference dF, which is a difference between a reference feature F0 of a reference waveform Wst measured with a reference parameter P0 and the peripheral feature F1 to Fm of the peripheral waveforms W1 to Wm measured with peripheral parameters P1 to Pm obtained by changing some of the values of the reference parameter P0, and which uses as output the parameter difference dP, which is the difference between the reference parameter P0 and the peripheral parameters P1 to Pm. A correction unit 917 inputs a feature difference F0'-F0, which is the difference between the reference feature F0 of the reference waveform Wst and a target feature F0' of a target waveform Wst' measured with the reference parameter P0 after the reference waveform Wst was measured, into the regression model Y, and corrects the reference parameter P0 according to the correction amount output from the regression model Y.SELECTED DRAWING: Figure 10

Description

[Technical Field]

[0001] The subject matter disclosed herein relates to a parameter correction device, a parameter correction method, and a computer program. [Background technology]

[0002] In the manufacturing process of flat panel displays, a device called a coater is used. A coater is a substrate processing device that discharges a processing liquid from a slit nozzle by driving a pump, thereby coating the processing liquid over the entire substrate being transported. With the recent trend toward higher product quality, such coaters are required to apply the processing liquid so that the film thickness of the processing liquid is uniform over the entire substrate. For example, in Patent Document 1, parameters for controlling the pump are adjusted and optimized by repeatedly measuring the discharge characteristics when discharging the processing liquid.

[0003] That is, the optimization process of Patent Document 1 includes a pseudo-discharge process of discharging a treatment liquid onto a surface other than the substrate, a discharge characteristic measurement process of measuring the discharge characteristics of the treatment liquid in the pseudo-discharge process, a state quantity derivation process of deriving a state quantity representing the deviation of the measured discharge characteristics from a target characteristic, and a learning process of constructing a learning model by machine learning the changes in the state quantities resulting from parameter changes. Then, while the state quantities exceed a predetermined allowable range, the parameters are changed based on the learning model, and the pseudo-discharge process, discharge characteristic measurement process, state quantity derivation process, and learning process are repeatedly executed. When the state quantities fall within the allowable range, the last-changed parameters are set as the parameters for discharging the treatment liquid in the treatment liquid supply process. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2020-040046 Summary of the Invention [Problem to be solved by the invention]

[0005] However, even when ejection is performed using optimized control parameters, the ejection characteristics may be disturbed due to the effects of wear and deterioration of parts, etc. In this case, if the control parameters are readjusted as in Patent Document 1, it will take a long time to restart the process, resulting in a significant decrease in yield. In addition, consuming a large amount of processing liquid may increase the environmental load.

[0006] An object of the present invention is to provide a technique that can quickly correct ejection characteristics while reducing the environmental load. [Means for solving the problem]

[0007] In order to solve the above problem, a first aspect is a parameter correction device that corrects parameters for controlling a coating device so that ejection characteristics measured when a treatment liquid is ejected from a nozzle become target ejection characteristics, the parameter correction device including: a regression model construction unit that constructs a regression model from the feature difference by machine learning using training data that receives as input a feature difference that is the difference between a feature amount of a reference ejection characteristic measured with a reference parameter and a feature amount of an ejection characteristic measured with a peripheral parameter obtained by varying a value of the reference parameter, and outputs a parameter difference that is the difference between the reference parameter and the peripheral parameter; and a correction unit that inputs into the regression model a difference between the feature amount of the reference ejection characteristic and a feature amount of an ejection characteristic measured with the reference parameter after the reference ejection characteristic was measured, and corrects the reference parameter according to the correction amount output from the regression model.

[0008] A second aspect is a parameter correction device of the first aspect, further comprising an abnormality judgment unit that judges whether the ejection characteristics are abnormal, and the correction unit corrects the reference parameters when the judgment unit judges that the ejection characteristics are abnormal.

[0009] A third aspect is the parameter correction device according to the first or second aspect, wherein the ejection characteristic is an ejection pressure applied to the treatment liquid.

[0010] A fourth aspect is a parameter correction method for correcting parameters for controlling a coating device so that ejection characteristics measured when a treatment liquid is ejected become target ejection characteristics, the method including: constructing a regression model that outputs a correction amount for correcting the reference parameter from a feature amount difference, which is the difference between a feature amount of a reference ejection characteristic measured with a reference parameter and a feature amount of an ejection characteristic measured with a peripheral parameter obtained by changing a value of a part of the reference parameter, by machine learning using training data that outputs a parameter difference, which is the difference between the reference parameter and the peripheral parameter; and inputting into the regression model the difference between the feature amount of the reference ejection characteristic and a feature amount of an ejection characteristic measured with the reference parameter after the time when the reference ejection characteristic was measured, and correcting the reference parameter according to the correction amount output from the regression model.

[0011] A fifth aspect is a computer program causing a computer to execute the parameter correction method of the fourth aspect. [Effects of the Invention]

[0012] According to the first to fifth aspects, even if a disturbance occurs in the ejection characteristics measured using the reference parameters, the disturbance in the ejection characteristics can be quickly corrected by correcting the reference parameters using the parameter difference calculated using the regression model as the correction amount, thereby preventing a decrease in the yield of the coating device.

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

[0014] According to the parameter correction device of the third aspect, the reference parameters can be corrected based on the discharge pressure. [Brief explanation of the drawings]

[0015] [Figure 1] 1 is a diagram schematically illustrating an overall configuration of a coating apparatus according to an embodiment. [Figure 2] 2 is a diagram showing the configuration of a treatment liquid supply mechanism included in the coating apparatus shown in FIG. 1. FIG. [Figure 3] 3 is a graph showing an example of a movement pattern of an operating disk portion in the pump shown in FIG. 2. [Figure 4] FIG. 2 is a block diagram showing an example of the configuration of a control unit. [Figure 5] FIG. 4 is a diagram showing an example of a discharge pressure waveform. [Figure 6] FIG. 10 is a diagram illustrating an example of a feature amount. [Figure 7] FIG. 10 is a diagram for explaining another example of feature amounts. [Figure 8] FIG. 10 is a diagram showing the flow of a regression model construction process performed by the control unit. [Figure 9] FIG. 10 is a diagram showing a flow of a reference parameter correction process executed by a control unit. [Figure 10] FIG. 2 is a block diagram conceptually showing the functions of a control unit. DETAILED DESCRIPTION OF THE INVENTION

[0016] Hereinafter, an embodiment of the present invention will be described with reference to the accompanying drawings. Note that the components described in the embodiment are merely examples and are not intended to limit the scope of the present invention. In the drawings, the dimensions and numbers of each part may be exaggerated or simplified as necessary to facilitate understanding.

[0017] <1. Embodiment> FIG. 1 is a diagram schematically illustrating the overall configuration of a coating apparatus 1 according to an embodiment. The coating apparatus 1 is a substrate processing apparatus that coats a processing liquid on an upper surface Sf of a substrate S. The coating apparatus 1 functions as a discharge characteristic monitoring device that monitors the discharge characteristics of the processing liquid. The discharge characteristics are specifically physical quantities related to discharge, such as the discharge pressure and the discharge flow rate. In the following description, a case will be described in which the discharge characteristic is the discharge pressure.

[0018] The substrate S is, for example, a glass substrate for a liquid crystal display device. The substrate S may also be a semiconductor wafer, a glass substrate for a photomask, a glass substrate for a plasma display, a glass or ceramic substrate for a magnetic or optical disk, a glass substrate for an organic EL display, a glass or silicon substrate for a solar cell, or any other substrate to be processed for electronic devices such as a flexible substrate or a printed circuit board. The coating device 1 is, for example, a slit coater.

[0019] In FIG. 1, an XYZ coordinate system is defined to explain the positional relationship of each element of the coating apparatus 1. The transport direction of the substrate S is the "X direction." The direction in which the substrate S advances in the X direction (toward downstream in the transport direction) is the +X direction, and the opposite direction (toward upstream in the transport direction) is the -X direction. The direction perpendicular to the X direction is the Y direction, and the direction perpendicular to the X and Y directions is the Z direction. In the following description, the Z direction is the vertical direction, and the X and Y directions are the horizontal directions. In the Z direction, the +Z direction is the upward direction, and the -Z direction is the downward direction. Note that these directions are not intended to limit the arrangement of the coating apparatus.

[0020] The coating apparatus 1 includes, in order in 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 along which the substrate S passes. The coating apparatus 1 also includes a substrate transport unit 5, a coating mechanism 7, a treatment liquid supply mechanism 8, and a control unit 9.

[0021] The substrate S is transported to the input conveyor 100 from a device upstream of the coating device 1. The input conveyor 100 includes a roller conveyor 101 and a rotation drive mechanism 102. The rotation drive mechanism 102 rotates each roller of the roller conveyor 101. Due to the rotation of each roller of the roller conveyor 101, the substrate S is transported downstream (+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).

[0022] The input transfer section 2 is equipped with a roller conveyor 21 and a rotation / lifting drive mechanism 22. The rotation / lifting drive mechanism 22 rotates each roller of the roller conveyor 21 and raises and lowers the roller conveyor 21. The rotation of the roller conveyor 21 transports the substrate S downstream (+X direction) in a horizontal position. The elevation of the roller conveyor 21 also changes the position of the substrate S in the Z direction. The substrate S is transferred from the input conveyor 100 to the floating stage section 3 via the input transfer section 2.

[0023] As shown in FIG. 1 , the floating stage unit 3 is substantially flat. The floating stage unit 3 is divided into three sections along the X direction. The floating stage unit 3 includes, in order along the +X direction, an entrance floating stage 31, a coating stage 32, and an exit floating stage 33. The upper surfaces of the entrance floating stage 31, the coating stage 32, and the exit floating stage 33 are on the same plane. The floating stage unit 3 further includes a lift pin drive mechanism 34, a floating control mechanism 35, and an elevation drive mechanism 36. The lift pin drive mechanism 34 raises and lowers the multiple lift pins arranged on the entrance floating stage 31. The floating control mechanism 35 supplies compressed air to the entrance floating stage 31, the coating stage 32, and the exit floating stage 33 to float the substrate S. The elevation drive mechanism 36 raises and lowers the exit floating stage 33.

[0024] A large number of nozzle holes for ejecting compressed air supplied from the levitation control mechanism 35 are arranged in a matrix on the upper surface of the entrance levitation stage 31 and the upper surface of the exit levitation stage 33. When compressed air is ejected from each nozzle hole, the substrate S is levitated upward relative to the levitation stage part 3. As a result, the lower surface Sb of the substrate S is separated from the upper surface of the levitation stage part 3, and the substrate S is supported in a horizontal position. When the substrate S is in a levitated state, the distance (levitation amount) between the lower surface Sb of the substrate S and the upper surface of the levitation stage part 3 is, for example, 10 μm or more and 500 μm or less.

[0025] The upper surface of the coating stage 32 is provided with jet holes for jetting compressed air supplied from the levitation control mechanism 35 and suction holes for sucking gas. The jet holes and suction holes are arranged alternately in the X and Y directions. The levitation control mechanism 35 controls the amount of compressed air jetted from the jet holes and the amount of air sucked from the suction holes. This precisely controls the amount of levitation of the substrate S relative to the coating stage 32 so that the position in the Z direction of the upper surface Sf of the substrate S passing above the coating stage 32 is a specified value. The amount of levitation of the substrate S relative to the coating stage 32 is calculated by the control unit 9 based on the detection results of a sensor 61 or a sensor 62, which will be described later. The amount of levitation of the substrate S relative to the coating stage 32 is preferably adjustable with high precision by airflow control.

[0026] The substrate S carried into the floating stage unit 3 is imparted with a propulsive force in the +X direction by the roller conveyor 21, and is transported onto the entrance floating stage 31. The entrance floating stage 31, the coating stage 32, and the exit floating stage 33 support the substrate S in a floating state. For example, the configuration described in Japanese Patent No. 5346643 may be adopted as the floating stage unit 3.

[0027] The substrate transport unit 5 is disposed below the floating stage unit 3. The substrate transport unit 5 includes a chuck mechanism 51 and a suction / travel control mechanism 52. The chuck mechanism 51 includes a suction pad (not shown) provided on 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 / travel control mechanism 52 applies negative pressure to the suction pad, thereby suctioning the substrate S to the suction pad. The suction / travel control mechanism 52 also causes the substrate transport unit 5 to travel back and forth in the X direction.

[0028] The chuck mechanism 51 holds the substrate S in a state where the lower surface Sb of the substrate S is positioned higher than the upper surface of the floating stage part 3. With the peripheral edge of the substrate S held by the chuck mechanism 51, the buoyancy applied by the floating stage part 3 keeps the substrate S in a horizontal position.

[0029] 1, the coating device 1 includes a sensor 61 for measuring plate thickness. The sensor 61 is disposed near the roller conveyor 21. The sensor 61 detects the position in the Z direction of the upper surface Sf of the substrate S held by the chuck mechanism 51. Furthermore, a chuck (not shown) that is not holding the substrate S is positioned directly below the sensor 61, so that the sensor 61 can detect the position in the vertical direction Z of the suction surface, which is the upper surface of the suction member.

[0030] The chuck mechanism 51 moves in the +X direction while holding the substrate S that has been carried into the floating stage section 3. As a result, the substrate S is transported from above the entrance floating stage 31, via above the coating stage 32, to above the exit floating stage 33. Then, the substrate S is moved from the exit floating stage 33 to the output transfer section 4.

[0031] The output transfer unit 4 moves the substrate S from a position above the exit floating stage 33 to the output conveyor 110. The output transfer unit 4 includes a roller conveyor 41 and a rotation / lifting drive mechanism 42. The rotation / lifting drive mechanism 42 drives the roller conveyor 41 to rotate and also raises and lowers the roller conveyor 41 in the Z direction. As each roller of the roller conveyor 41 rotates, the substrate S moves in the +X direction. Furthermore, as the roller conveyor 41 rises and falls, the substrate S is displaced in the Z direction.

[0032] The output conveyor 110 includes a roller conveyor 111 and a rotation drive mechanism 112. The output conveyor 110 transports the substrate S in the +X direction by the rotation of each roller of the roller conveyor 111, and delivers the substrate S to the outside 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 be incorporated into a device separate from the coating apparatus 1.

[0033] The coating mechanism 7 coats the upper surface Sf of the substrate S with a processing liquid. The coating mechanism 7 is disposed above the transport path of the substrate S. The coating mechanism 7 has a nozzle 71. The nozzle 71 is a slit nozzle having a slit-shaped outlet on its lower surface. The nozzle 71 is connected to a positioning mechanism (not shown). The positioning mechanism moves the nozzle 71 between a coating position above the coating stage 32 (the position indicated by the solid line in FIG. 1) and a maintenance position, which will be described later. 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, causing the processing liquid to be ejected from an outlet disposed on the lower surface of the nozzle 71.

[0034] FIG. 2 is a diagram showing the configuration of a treatment liquid supply mechanism 8 included in the coating apparatus 1 shown in FIG. 1. The treatment liquid supply mechanism 8 includes a pump 81, a pipe 82, a treatment liquid replenishment unit 83, a pipe 84, an on-off valve 85, a pressure sensor 86, and a drive unit 87. The pump 81 is a supply source for supplying the treatment liquid to the nozzle 71 and supplies the treatment liquid by changing its volume. For example, a bellows-type pump as described in Japanese Patent Application Laid-Open No. 10-61558 can be used as the pump 81. As shown in FIG. 2, the pump 81 has 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 treatment liquid replenishment unit 83 via the pipe 82. The other end of the flexible tube 811 is connected to the nozzle 71 via the pipe 84.

[0035] The pump 81 has a bellows 812 that is elastically deformable in the axial direction. The bellows 812 has a small bellows section 813, a large bellows section 814, a pump chamber 815, and an operating disk section 816. The pump chamber 815 is disposed between the flexible tube 811 and the bellows 812. An incompressible medium is sealed in the pump chamber 815. The operating disk section 816 is connected to the drive section 87.

[0036] The processing liquid replenishment unit 83 has a storage tank 831 that stores the processing liquid. The storage tank 831 is connected to the pump 81 via a pipe 82. An on-off valve 833 is inserted in the pipe 82. The on-off valve 833 opens and closes in response to a command from the control unit 9. When the on-off valve 833 is opened, the processing 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 replenishment of the processing liquid from the storage tank 831 to the flexible tube 811 of the pump 81 is restricted.

[0037] The pipe 84 is connected to the output side of the pump 81. The on-off valve 85 is provided in the pipe 84. The on-off valve 85 opens and closes the pipe 84 in response to commands from the control unit 9. The on-off valve 85 opens and closes the pipe 84, switching between sending and stopping the processing liquid to the nozzle 71. The pressure sensor 86 is provided in the pipe 84. The pressure sensor 86 detects the pressure (discharge pressure) applied to the processing liquid sent to the nozzle 71, and outputs a signal indicating the detected pressure value to the control unit 9.

[0038] FIG. 3 is a graph showing an example of a movement pattern of the actuation disc portion 816 in the pump 81 shown in FIG. 2. In FIG. 3, the horizontal axis represents time, and the vertical axis represents the movement speed of the actuation disc portion 816. In response to a command from the control unit 9, the drive unit 87 displaces the actuation disc portion 816 in the axial direction according to the movement pattern shown in FIG. 3 (a pattern showing the change in the speed of the actuation disc portion 816 over time). The displacement of the actuation disc portion 816 changes the internal volume of the bellows 812. This causes the flexible tube 813 to expand and contract radially, performing a pumping operation, and the processing liquid replenished from the processing liquid replenishment unit 83 is fed toward the nozzle 71. The movement pattern of the actuation disc portion 816 is closely related to the discharge characteristics of the processing liquid discharged from the nozzle 71. Therefore, a discharge pressure waveform (see FIG. 5) showing the time change in the discharge pressure corresponding to the movement pattern of the actuation disc portion 816 is measured. The discharge rate (the amount of treatment liquid discharged from the nozzle 71) also increases or decreases in accordance with the increase or decrease in the discharge pressure. The discharge pressure waveform is an example of the discharge characteristics when the nozzle 71 discharges the treatment liquid.

[0039] In this embodiment, by adjusting various parameters (acceleration time, steady-state velocity, steady-state velocity time, deceleration time, etc.) that define the movement of the actuation disk portion 816, an optimization process is appropriately performed to match or approximate the ejection characteristics of the processing liquid ejected from the nozzle 71 (specifically, the change in ejection velocity (ejection pressure) over time) to the desired target characteristics.

[0040] As shown in FIGS. 1 and 2, a sensor 62 is disposed in 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 stage 32. Then, based on the measured separation distance, the control unit 9 adjusts the coating position of the nozzle 71 using the positioning mechanism. Note that an optical sensor or an ultrasonic sensor can be used as the sensor 62.

[0041] The coating mechanism 7 includes a nozzle cleaning standby unit 72. The nozzle cleaning standby unit 72 performs predetermined maintenance on the nozzle 71 positioned at the maintenance position. The nozzle cleaning standby unit 72 includes a roller 721, a cleaning section 722, and a roller vat 723. The nozzle cleaning standby unit 72 cleans the nozzle 71 and forms a liquid pool, thereby preparing the discharge port of the nozzle 71 for a state suitable for coating processing. Furthermore, in the coating device 1, in order to evaluate the discharge pressure applied to the processing liquid, a simulated discharge is performed in which the processing liquid is discharged from the nozzle 71 while the nozzle 71 is positioned at the maintenance position.

[0042] FIG. 4 is a block diagram showing an example of the configuration of the control unit 9. The control unit 9 controls the operation of each component in the coating apparatus 1. A computer can be used as the control unit 9. The control unit 9 includes a processor 91 and a memory 93. The processor has, for example, a CPU (Central Processing Unit). The memory 93 has a transient storage device such as a RAM (Random Access Memory). The memory 93 may also have a non-transient 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 a bus wiring.

[0043] The control unit 9 has a display device 95 that displays various information and an input device 97 that accepts user command inputs. 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 has, for example, a mouse or a keyboard. Note that the display device 95 may have a touch panel so that it functions as an input device.

[0044] The memory 93 stores a computer program 931. The computer program 931 is provided to the control unit 9 via a recording medium M. That is, the computer program 931 is recorded on the recording medium M so as to be readable by the control unit 9, which is a computer. The recording medium M is specifically a USB (Universal Serial Bus) memory, an optical disk such as a DVD (Digital Versatile Disc), a magnetic disk, or the like.

[0045] The processor 91 executes the computer program 931 to function as a discharge control unit 910 , a discharge pressure measurement unit 911 , a feature calculation unit 913 , an abnormality determination unit 915 , a correction unit 917 , and a regression model construction unit 919 .

[0046] The discharge control unit 910 controls, based on preset control parameters, the operation (supply operation) of the pump 81 that supplies the processing liquid to the nozzle 71. In the coating apparatus 1, in order to coat the processing liquid discharged from the nozzle 71 onto the upper surface Sf of the substrate S with a uniform film thickness, before starting production (or mass production) of the substrate S, the control parameters closely related to the discharge pressure waveform are optimized in advance so that the discharge pressure waveform has an ideal shape.

[0047] The control parameters are, for example, setting values ​​that define the movement of the operating disk portion 816, and the setting values ​​for controlling a plurality of pumps shown in FIG. 3 and below can be used as the control parameters to be optimized. ·Constant speed V1 Acceleration time T1: Time required to accelerate from a stopped state to a steady speed V1 Steady-state speed time T2: Time to maintain steady-state speed V1 Constant speed V2 Acceleration time T3: Time to decelerate from steady speed V1 to steady speed V2 Steady-state speed time T4: Time to maintain steady-state speed V2 ·Constant Speed ​​V3 Acceleration time T5: Time to accelerate from steady speed V2 to steady speed V3 - Steady speed time T6: Time to continue steady speed V3 Constant Speed ​​V4 Acceleration time T7: Time to decelerate from steady speed V3 to steady speed V4 - Steady speed time T8: Time to continue steady speed V4 Constant Speed ​​V5 Acceleration time T9: Time to accelerate from steady speed V4 to steady speed V5 - Steady speed time T10: Time to continue steady speed V5 Deceleration time T11: Time to decelerate from steady speed V5 to a stopped state

[0048] Returning to FIG. 4, the 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 indicating the time change in the discharge pressure applied to the processing liquid during the period in which the processing liquid is discharged from the nozzle 71, and appropriately stores discharge pressure data indicating the acquired discharge pressure waveform in the memory 93. The discharge pressure data is time-series data indicating the pressure measured at each time.

[0049] 5 is a diagram showing an example of a discharge pressure waveform W. In FIG. 5, the horizontal axis represents time, and the vertical axis represents discharge pressure. As shown in FIG. 5, the discharge pressure is measured for a predetermined period from before the nozzle 71 starts discharging the treatment liquid to after the nozzle 71 finishes discharging the treatment liquid.

[0050] 5, the discharge pressure at time ta when discharge of the treatment liquid from the nozzle 71 starts and the discharge pressure at time te when discharge of the treatment liquid from the nozzle 71 ends are the initial pressure Pi. Note that the pressure at the start of discharge and the pressure at the end of discharge do not always match the initial pressure Pi. The discharge period from time ta to time te is divided into a rise period Tab, a transition period Tbc, a steady period Tcd, and a fall period Tde.

[0051] The rise period Tab is the period from time ta when the processing liquid supply mechanism 8 starts 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 starts moving the operating disk portion 816. In other words, when the processing liquid starts to be discharged from the nozzle 71 at time ta, the discharge pressure increases from the initial pressure Pi to the target pressure Pt between time ta and time tb.

[0052] The transition period Tbc is the period from time tb to time tc, which is the time it takes for a predetermined vibration attenuation period to elapse. This vibration attenuation period is the period required for the time change in the discharge pressure to stabilize.

[0053] The steady period Tcd is the period from time tc to time td when the treatment liquid supply mechanism 8 starts to reduce the discharge pressure. Time td is the time when the treatment liquid supply mechanism 8 starts to decelerate the operating disk unit 816 from the target speed. The treatment liquid supply mechanism 8 moves the operating disk unit 816 at a constant speed from time tc to time td, and starts to decelerate the operating disk unit 816 at time td. During the steady period Tcd, the discharge pressure basically stabilizes at the target pressure Pt. However, even during the steady period Tcd, the change in the discharge pressure over time includes minute vibrations, and the discharge pressure may become larger or smaller than the target pressure Pt.

[0054] The falling period Tde is the period from time td to time te when the treatment liquid supply mechanism 8 finishes discharging the treatment liquid from the nozzle 71. Time te is the time when the treatment liquid supply mechanism 8 stops the operating disk unit 816. In other words, the discharge pressure decreases to the initial pressure Pi between time td and time te, and at time te, the discharge of the treatment liquid from the nozzle 71 stops.

[0055] In the coating apparatus 1, before starting 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, the waveform shown by the dashed line in FIG. 5). Specifically, the control parameters are adjusted by repeating a series of processes including pseudo-discharge using the control parameters, measurement of the discharge pressure waveform during the pseudo-discharge, evaluation of the measured discharge pressure waveform, and updating of the control parameters based on the evaluation. To update the control parameters, for example, the method described in Patent Document 1 can be used. Hereinafter, the control parameters adjusted so that the discharge pressure waveform becomes the reference waveform Wst will be referred to as "reference parameters P0."

[0056] When producing a substrate S, the coating apparatus 1 performs a coating process using the reference parameter P0. That is, the discharge control unit 910 controls the pump using the adjusted reference parameter P0, thereby discharging the processing liquid onto the substrate S according to the reference waveform Wst. However, even when coating is performed using the reference parameter P0, a disturbance (abnormality) may occur in the discharge pressure waveform due to various reasons, such as wear or deterioration of parts such as the pump 81 over time or the influence of external disturbances. That is, the discharge pressure waveform may have a shape that is partially different from the reference waveform Wst. If the discharge pressure waveform is disturbed, it becomes difficult to perform normal coating. Therefore, the control unit 9 measures the discharge pressure waveform, for example, by periodically performing pseudo-discharge, and corrects the reference parameter P0 so that the measured discharge pressure waveform approaches the original reference waveform Wst. In this way, the control unit 9 functions as a "parameter correction device."

[0057] Returning to FIG. 4, the feature amount calculation unit 913 calculates a feature amount from the discharge pressure data measured by the discharge pressure measurement unit 911. FIG. 6 is a diagram for explaining an example of a feature amount. In this example, the degree of overshoot that occurs at the rising edge of the discharge pressure waveform W is calculated as the feature amount. Specifically, the feature amount calculation unit 913 calculates the sign (positive or negative) of the second-order differential value Dif2 of the discharge pressure at time t11 when the discharge pressure reaches the maximum value Pmax. Then, the feature amount calculation unit 913 calculates the time t12 when the sign of the second-order differential value Dif2 switches twice from the sign at time t11. Furthermore, the feature amount calculation unit 913 calculates a feature amount of the time change of the discharge pressure during the initial vibration period T2_s from time t11 to t12.

[0058] For example, the feature amount calculation unit 913 selects the smaller of the minimum value Pmin of the discharge pressure during the initial vibration period T2_s and the steady pressure Pm (the average value of the discharge pressure during the steady period T3) as the target pressure Pg.The feature amount calculation unit 913 may then calculate the difference between the maximum value Pmax and the target pressure Pg (=Pmax-Pg) as the feature amount.This feature amount becomes relatively larger as the overshoot of the discharge pressure over time increases due to the momentum of the rise in the discharge pressure.

[0059] FIG. 7 is a diagram illustrating another example of the feature quantity. In this example, the stability of the time change of the discharge pressure during the transition period Tbc is calculated as the 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 steady pressure Pm, which is the average value of the discharge pressure during the steady period Tcd, as the feature quantity. Such a feature quantity becomes relatively larger as the ringing (waveform oscillation) of the discharge pressure during the transition period Tbc increases.

[0060] The feature amounts calculated by the feature amount calculation unit 913 are not limited to the feature amounts described in Fig. 6 and Fig. 7. As the feature amounts, for example, some feature amounts described in JP 2022-138109 A can be adopted.

[0061] 4, the abnormality determination unit 915 determines whether or not the discharge pressure waveform W measured by the discharge pressure measurement unit 911 is abnormal. As described above, in order to correct the reference parameter P0, the control unit 9 determines whether or not an abnormality has occurred in the discharge pressure waveform (hereinafter also referred to as "target waveform Wst'") measured by pseudo-discharge using the reference parameter P0.

[0062] For example, the abnormality determination unit 915 calculates the degree of abnormality of the target waveform Wst' and performs an abnormality determination based on the degree of abnormality. The degree of abnormality is a value indicating the degree of deviation from the distribution of feature quantities of a normal discharge pressure waveform. To calculate the degree of abnormality, a trained model constructed to input feature quantities of the discharge pressure waveform and output the degree of abnormality can be used. Such a trained model is constructed by performing machine learning based on the k-nearest neighbor method using a set of feature quantities of multiple discharge pressure waveforms that are considered normal as training data. The degree of abnormality is preferably the Mahalanobis distance from the k-nearest neighbors, but may also be the Euclidean distance or Manhattan distance.

[0063] The anomaly determination unit 915 may also determine whether or not the target waveform Wst' is abnormal based on the probability that it corresponds to an anomaly. The target waveform Wst' may be classified into a normal state (i.e., a state close to the reference waveform Wst) and one or more types of abnormal states that are assumed in advance. Therefore, the anomaly determination unit 915 may calculate the probability that it corresponds to each of several assumed anomalies (hereinafter also referred to as "anomaly probability").

[0064] To calculate the abnormality probability, a trained model constructed to input feature values ​​and output the abnormality probability can be used. When constructing such a trained model, first, several feature values ​​of normal pressure discharge waveforms (hereinafter referred to as "normal feature values") and a considerable number of feature values ​​of discharge pressure waveforms for each type of abnormality are prepared. Then, in order to distinguish between normal and one or more abnormal states, target values ​​are set for each state, and the model is trained so that it outputs the target value corresponding to the input data. Note that a linear regression model can be used as the base model. In the case of a trained model obtained in this way, the closer the output value is to a specific target value, the higher the probability that it corresponds to the abnormality corresponding to that specific target value.

[0065] The abnormality determination unit 915 may calculate both the abnormality degree and the abnormality probability, and determine whether or not an abnormality exists based on these results.

[0066] The correction unit 917 corrects the reference parameter P0. In the coating apparatus 1, the abnormality determination unit 915 determines whether an abnormality (disturbance) has occurred in the discharge pressure waveform W. 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 may correct the reference parameter P0 based on the measured discharge pressure waveform W, regardless of whether an abnormality has occurred.

[0067] As will be described later, the correction unit 917 calculates a feature difference F0'-F0, which is the difference between a reference feature F0, which is the feature of the reference waveform Wst, and F0', which is the feature of the target waveform Wst'. The correction unit 917 then inputs the feature difference F0'-F0 into the regression model Y to calculate a parameter difference ΔP, which is a correction amount, and adds the obtained parameter difference ΔP to the reference parameter P0 to calculate a correction parameter P C Get.

[0068] The regression model construction unit 919 constructs the regression model Y. The method for constructing the regression model Y will be described with reference to Fig. 8. Fig. 8 is a diagram showing the flow of the process of constructing the regression model Y by the control unit 9.

[0069] First, the regression model construction unit 919 calculates a plurality of peripheral parameters P1, P2, P m (m is the number of peripheral parameters) are created (step S11). m is a parameter obtained by slightly changing at least one of the values ​​included in the reference parameter P0. For example, if the reference parameter P0 has n values ​​(p 10 ,p 20 ,···,p n0 ) (n is the number of parameter values), by slightly changing some or all of the values, m mutually different peripheral parameters P1 to P m is created.

[0070] Next, the coating apparatus 1 calculates the peripheral parameters P1 to P m Then, a simulated discharge is performed using each of the peripheral parameters P1, P2, . . . , P m Discharge pressure data of peripheral waveforms W1, W2, . . . , Wm, which are a plurality of discharge pressure waveforms corresponding to the respective waveforms, are acquired.

[0071] Next, the feature calculation unit 913 calculates the reference feature F0 of the reference waveform Wst and the surrounding feature F1, F2, . . . , F1, which are the feature of the surrounding waveforms W1 to Wm. m Then, the regression model construction unit 919 calculates the peripheral parameters P1 to P m For each differential dataset D1, D2, , D m (Step S14) m is the parameter difference d P and feature difference d F It is composed of a set of

[0072] Parameter difference d P is the reference parameter P0 and each peripheral parameter P1 to P m For example, the reference parameter P0 is (p 10 ,p 20 ,···p n0 ) (n is the number of parameters), and the peripheral parameter P1 is (p 11 ,p 21 ,···,p n1 ), then the parameter difference d P1 is calculated by subtracting each value of the peripheral parameter P1 from each value of the reference parameter P0, as shown in the following equation.

[0073] d P1 =(p 10 -p 11 ,p 20 -p 21 ,···,p n0 -p n1 )

[0074] In addition, the feature difference d F is the reference feature F0 of the reference waveform Wst and the peripheral feature F1 to F2 of the peripheral waveforms W1 to Wm. m Specifically, the difference between the surrounding feature values ​​F1 to F m For example, the reference feature F0 is calculated by subtracting the reference feature F0 from (f 10 ,f 20 ,···,f k0 ) (k is the number of features), and the peripheral feature F1 is (f 11 ,f 21 ,···,f k1 ), then the feature difference d F1 is calculated by subtracting each feature amount of the reference feature amount F0 from each feature amount of the surrounding feature amount F1, as shown in the following equation.

[0075] d F1 =(f 11 -f 10 ,f 21 -f 20 ,···,f k1 -f k0 )

[0076] Next, the regression model construction unit 919 generates a plurality of differential data sets D1 to D m The regression model Y is constructed by machine learning using the difference data sets D1 to D m The feature difference d F1 ~d Fm is used as input, and the differential data sets D1 to D m Parameter difference d P1 ~d Pm A supervised learning is performed in which the output is the regression model Y. For example, a random forest or a neural network can be used as the regression model Y. The trained regression model Y can predict, from the feature difference, the correction amount (parameter difference) of the reference parameter P0 required to eliminate the feature difference. The parameters of the regression model Y obtained by machine learning (trained parameters) are appropriately stored in the memory 93.

[0077] <Correction of reference parameters> Next, a process of correcting the reference parameter P0 using the constructed regression model Y will be described. Fig. 9 is a diagram showing the flow of the correction process executed by the control unit 9. The correction process shown in Fig. 9 is performed, for example, periodically or based on a command input by an operator at the stage of performing the coating process using the adjusted reference parameter P0.

[0078] First, the control unit 9 performs a pseudo-ejection using the reference parameter P0 and measures the ejection pressure to obtain a target waveform Wst' (step S21). The target waveform Wst' is an example of an ejection characteristic measured after the reference waveform Wst, which is a reference ejection characteristic, is measured.

[0079] Next, the abnormality determination section 915 of the control unit 9 determines whether or not the target waveform Wst' is abnormal (step S22). If it is determined in step S22 that the target waveform Wst' is not abnormal, the control unit 9 ends the correction process. On the other hand, if it is determined in step S22 that the target waveform Wst' is abnormal, the feature calculation section 913 calculates a target feature F0', which is a feature of the target waveform Wst' (step S23).

[0080] Next, the correction unit 917 of the control unit 9 corrects the reference parameter P0 using the target feature F0' calculated in step S23 (step S24). Specifically, the correction unit 917 calculates a feature difference F0'-F0, which is the difference between the target feature F0' and the reference feature F0. The correction unit 917 inputs the calculated feature difference F0'-F0 into the regression model Y to obtain a parameter difference ΔP, which is a correction amount. The correction unit 917 adds the calculated parameter difference ΔP to the reference parameter P0 to obtain the correction parameter P C Calculate.

[0081] Correction parameter P C When is calculated, the correction unit 917 calculates the correction parameter P C is stored in the memory 93. After that, the discharge control unit 910 stores the correction parameter P C The coating process is performed on the substrate S using the above.

[0082] 10 is a block diagram conceptually showing the function of the control unit 9. As explained in FIG. 8, in the construction stage of the regression model Y, the reference feature amount F0 of the reference waveform Wst and the peripheral waveforms W1 to W m The peripheral features F1 to F m is calculated by the feature calculation unit 913 and passed to the regression model construction unit 919. Then, the regression model construction unit 919 calculates the feature difference d F and the parameter difference d P The differential data sets D1 to D mA regression model Y is constructed by performing machine learning using the above as training data.

[0083] 9, the target feature F0' of the target waveform Wst' measured with the reference parameter P0 is calculated by the feature calculation unit 913 and passed to the correction unit 917. The correction unit 917 then inputs the feature difference F0'-F0 between the target feature F0' and the reference feature F0 to the regression model Y. The correction unit 917 then corrects the reference parameter P0 using the parameter difference ΔP, which is the correction amount output by the regression model Y, to obtain the correction parameter P C is generated.

[0084] <Effects> According to the control unit 9, even if a disturbance occurs in the target waveform Wst' measured using the reference parameter P0, the disturbance in the target waveform Wst' can be corrected by correcting the reference parameter P0 using the parameter difference ΔP calculated using the regression model Y as the correction amount. In other words, since there is no need to readjust the reference parameter P0 by repeating pseudo-ejection, the disturbance in the target waveform Wst' can be corrected while reducing the environmental load. This makes it possible to avoid a decrease in the yield of the coating apparatus 1.

[0085] <2. Modifications> Although the embodiments have been described above, the present invention is not limited to the above and various modifications are possible.

[0086] For example, in the above embodiment, the parameter correction device is configured as the control unit 9 provided in the coating apparatus 1. However, the parameter correction device may be configured as a computer device separate from the coating apparatus 1.

[0087] 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, the discharge flow rate may be measured by providing a flow meter in the pipe 84, for example.

[0088] Although the present invention has been described in detail, the above description is merely illustrative in all respects and does not limit the present invention. It is understood that countless variations not illustrated can be envisioned without departing from the scope of the present invention. The configurations described in the above embodiments and variations can be combined or omitted as appropriate as long as they are not mutually inconsistent. [Explanation of symbols]

[0089] 1: Coating device 9: Control unit (parameter correction device) 913: Feature calculation unit 915: Abnormality determination section 917: Correction section 919: Regression Model Building Department 931: Computer Program W: Discharge pressure waveform (discharge characteristics) Wst: Standard waveform (standard discharge characteristics) Y: regression model

Claims

1. a parameter correction device that corrects parameters for controlling a coating device so that ejection characteristics measured when a treatment liquid is ejected from a nozzle become target ejection characteristics, a regression model construction unit that constructs a regression model that outputs a correction amount for correcting the reference parameters from the feature amount difference by machine learning using training data that receives as input a feature amount difference that is the difference between a feature amount of a reference ejection characteristic measured using a reference parameter and a feature amount of an ejection characteristic measured using a peripheral parameter obtained by changing a value of part of the reference parameter, and outputs a parameter difference that is the difference between the reference parameter and the peripheral parameter; and a correction unit that inputs a difference between a feature amount of the reference ejection characteristic and a feature amount of an ejection characteristic measured using the reference parameters after the time point at which the reference ejection characteristic was measured into the regression model, and corrects the reference parameters in accordance with the correction amount output from the regression model; A parameter correction device comprising:

2. 2. The parameter correction device according to claim 1, an abnormality determination unit that determines whether the ejection characteristics are abnormal; Furthermore, The correction unit corrects the reference parameters when the determination unit determines that the ejection characteristics are abnormal.

3. 3. The parameter correction device according to claim 1, The parameter correction device, wherein the ejection characteristic is an ejection pressure applied to the processing liquid.

4. A parameter correction method for correcting parameters for controlling a coating device so that ejection characteristics measured when a treatment liquid is ejected become target ejection characteristics, the method comprising: a step of constructing a regression model that outputs a correction amount for correcting the reference parameters from the feature amount difference by machine learning using training data that takes as input a feature amount difference that is the difference between a feature amount of a reference ejection characteristic measured using a reference parameter and a feature amount of an ejection characteristic measured using a peripheral parameter obtained by changing a part of the value of the reference parameter, and outputs a parameter difference that is the difference between the reference parameter and the peripheral parameter; a step of inputting a difference between the feature amount of the reference ejection characteristic and the feature amount of the ejection characteristic measured using the reference parameters after the time point at which the reference ejection characteristic was measured into the regression model, and correcting the reference parameters according to the correction amount output from the regression model; A parameter correction method comprising:

5. A computer program comprising: A computer program causing the computer to execute the parameter correction method according to claim 4.

Images