Pump control parameter adjustment method, computer program, recording medium, spraying and coating device

Through machine learning of the regression model, adjusting the pump control parameters and combining offset optimization, the problem of inconsistent with the injection pressure curve and the target curve is solved, and the consistency and accuracy of coating effect in the injection action are improved.

CN117619592BActive Publication Date: 2024-11-08SCREEN HOLDINGS CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202311091614.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2022-09-01
Filing Date
2023-08-28
Publication Date
2024-11-08
Estimated Expiration
2043-08-28

AI Technical Summary

Technical Problem

In the evaluation and control of the spray pressure, there is a difference between the evaluation and the actual spray operation conditions in the prior art, resulting in inconsistent with the spray pressure curve and the target curve, which affects the coating effect.

Method used

By adjusting the pump control parameters using machine learning using regression model, the injection pressure curve is close to the pre-set target curve, and considering the spray condition differences in the re-adjustment process, the pump control parameters are optimized through offset to actual injection Match the target curve during the action.

Benefits of technology

The matching of the discharge pressure curve and the target curve in the actual discharge operation is achieved, and the consistency and accuracy of the coating effect are improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117619592B_ABST
    Figure CN117619592B_ABST
Patent Text Reader

Abstract

The present invention relates to a method for adjusting pump control parameters, a computer program, a recording medium, a spraying device, and a coating device, which optimizes the pump control parameters used to deliver a treatment liquid from a pump to a nozzle so that characteristics consistent with a target can be obtained in an actual spraying action. The method for adjusting pump control parameters of the present invention includes: an adjustment process, temporarily setting control parameters, delivering a treatment liquid from a pump, obtaining a spraying pressure curve, and adjusting control parameters so that the obtained spraying pressure curve is close to a pre-set target curve; a process of performing a spraying action based on the adjusted control parameters to obtain a spraying pressure curve; a setting process, setting an offset corresponding to the difference between the spraying pressure curve obtained in the adjustment process and the spraying pressure curve obtained in the spraying action; and a readjustment process, readjusting the control parameters so that the curve after adding the offset to the obtained spraying pressure curve is close to the target curve.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to a technique for ejecting a treatment liquid from a nozzle by applying an ejection pressure to the treatment liquid. In addition, as an object to which the treatment liquid is ejected from the nozzle, for example, there are semiconductor substrates, photomask substrates, liquid crystal display substrates, organic EL display substrates, plasma display substrates, FED (Field Emission Display) substrates, optical disk substrates, magnetic disk substrates, and magneto-optical disk substrates. Background Art

[0002] As shown in Japanese Patent Publication No. 2011-005465 (Patent Document 1) and Japanese Patent Publication No. 2020-040046 (Patent Document 2), when the treatment liquid ejected from the nozzle is applied on the substrate, the ejection pressure applied to the treatment liquid has a greater impact on the thickness of the treatment liquid applied on the substrate. Therefore, in Patent Document 1, the waveform of the ejection pressure is divided into multiple intervals, and the slope of the waveform of each interval is used to evaluate whether the ejection pressure is within the allowable range. In addition, in Patent Document 2, for each region such as the rising region and the stable ejection region, the optimization of the control parameters associated with the ejection pressure is sought. That is, in Patent Documents 1 and 2, by evaluating the waveform of the ejection pressure in each of multiple intervals or regions, a high-precision evaluation of the ejection pressure can be performed.

[0003] In the above-mentioned prior art, the evaluation of the ejection pressure is based on the measurement result of the ejection pressure obtained when the nozzle is positioned at a position different from the position in the ejection action of the actual coating substrate (for example, a maintenance position). Therefore, the ejection conditions during the evaluation are not necessarily the same as the ejection conditions during the actual ejection action. In particular, when the difference is large, when the substrate is coated using the control parameters optimized based on the evaluation results, the problem of not necessarily ejecting according to the target characteristics may arise. In this regard, the above-mentioned prior art still has room for improvement. Summary of the invention

[0004] The present invention has been made in view of the above problems, and an object of the present invention is to provide a technology capable of optimizing pump control parameters for delivering a processing liquid from a pump to a nozzle so as to obtain characteristics matching a target in an actual discharge operation.

[0005] One embodiment of the present invention is a method for adjusting a pump control parameter used for performing a spraying action of spraying a processing liquid delivered by a pump from a nozzle toward a substrate, the method comprising: an adjustment step of temporarily setting the pump control parameter, delivering the processing liquid from the pump, acquiring a spray pressure curve representing the variation of the spray pressure applied to the delivered processing liquid over time, and adjusting the pump control parameter by machine learning using a regression model so that the acquired spray pressure curve approaches a preset target curve; a step of operating the pump based on the adjusted pump control parameter, performing the spraying action, and acquiring the spray pressure curve; a step of setting an offset corresponding to the difference between the spray pressure curve acquired in the adjustment step using the adjusted pump control parameter and the spray pressure curve acquired in the spraying action; and a readjustment step of temporarily setting the pump control parameter, delivering the processing liquid from the pump, acquiring the spray pressure curve, and readjusting the pump control parameter by the machine learning so that a curve obtained by adding the offset to the acquired spray pressure curve approaches the target curve.

[0006] In addition, another embodiment of the present invention is a spraying device that performs a spraying operation of spraying a processing liquid onto a substrate, wherein the device comprises: a nozzle that sprays the processing liquid; a pump that delivers the processing liquid to the nozzle; a detection unit that detects the spraying pressure applied to the processing liquid delivered by the pump; and a control unit that controls the pump based on a predetermined pump control parameter. Furthermore, the control unit obtains a spraying pressure curve that indicates the change in the spraying pressure applied to the delivered processing liquid over time based on the detection result of the detection unit, and performs the pump control parameter adjustment method of the above invention to optimize the pump control parameter.

[0007] In the present invention thus constituted, in the adjustment process, the treatment liquid is sent using the temporarily set pump control parameters and the pump control parameters are adjusted based on the detection result of the ejection pressure curve at that time. Thus, the pump control parameters are optimized under the ejection conditions at that time. On the other hand, the pump control parameters thus optimized may not necessarily be the best in the actual ejection operation.

[0008] Therefore, in the present invention, the ejection action is performed using the adjusted pump control parameters, and the difference between the ejection pressure curve obtained at this time and the target curve is calculated as an offset. Then, in the readjustment process, the offset is added to the measured ejection pressure curve. Thus, by evaluating the ejection pressure curve and adjusting the pump control parameters in a state where the offset is added, it is possible to perform optimization that takes into account the difference between the ejection conditions in the adjustment process and the readjustment process and the ejection conditions in the actual ejection action. That is, by performing such a readjustment process, the ejection pressure curve in the actual ejection action can be matched with the target curve.

[0009] As described above, according to the present invention, it is possible to optimize the pump control parameters for delivering the processing liquid from the pump to the nozzle so that characteristics matching the target can be obtained in the actual discharge operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 This is a diagram schematically showing the overall structure of one embodiment of the coating device of the present invention.

[0011] Figure 2 It is a diagram showing the structure of a coating liquid supply mechanism.

[0012] Figure 3 This is a block diagram showing an example of the configuration of a control unit.

[0013] Figure 4 This is a flowchart showing an example of a discharge pressure evaluation method.

[0014] FIG. 5A to FIG. 5B This is a diagram for explaining each period used for evaluating the discharge pressure.

[0015] Figure 6 It is a diagram schematically showing an example of calculation performed by the pressure evaluation unit.

[0016] Figure 7 It is a diagram for explaining evaluation items based on the feature amount Fv1.

[0017] Figure 8 It is a diagram for explaining evaluation items based on the feature amount Fv2.

[0018] Fig. 9 It is a diagram for explaining evaluation items based on the feature amount Fv3.

[0019] FIG. 10A to FIG. 10B It is a diagram for explaining evaluation items based on the feature amount Fv4.

[0020] FIG. 11A to FIG. 11B It is a diagram for explaining evaluation items based on the feature amount Fv5.

[0021] Fig.12 It is a diagram for explaining evaluation items based on the feature amount Fv6.

[0022] Fig.13 It is a diagram for explaining evaluation items based on the feature amount Fv7.

[0023] Fig.14 It is a diagram for explaining evaluation items based on the feature amount Fv8.

[0024] Fig.15 It is a diagram for explaining evaluation items based on the feature amount Fv9.

[0025] Fig.16 This is a diagram for explaining evaluation items based on the feature amount Fv10.

[0026] Fig.17 This is a diagram for explaining evaluation items based on the feature amount Fv11.

[0027] Fig.18 It is a diagram for explaining evaluation items based on the feature amount Fv12.

[0028] Fig.19 This is a diagram for explaining evaluation items based on the feature amount Fv13.

[0029] Fig. 20 It is a diagram for explaining evaluation items based on the feature amount Fv14.

[0030] Fig.21 This is a flowchart showing an example of parameter adjustment processing.

[0031] Fig. 22 This is a diagram showing an example of the offset amount.

[0032] Fig.23 This is a diagram for explaining each period used in a modified example of the evaluation item of the discharge pressure.

[0033] Fig.24 It is a diagram for explaining a modification example of the evaluation items based on the feature amount Fv1_1.

[0034] Description of Reference Numerals

[0035] 1Coating device

[0036] 71 Nozzle (spraying device)

[0037] 81 Pump (spray device)

[0038] 86 Pressure gauge (detection unit)

[0039] 9Control unit (control unit, ejection device)

[0040] 5. Substrate transport unit (relative movement mechanism)

[0041] 721 roller (roller member)

[0042] 97 Discharge Pressure Evaluation Procedure

[0043] 98 parameter adjustment program

[0044] 99 Discharge pressure measurement data (discharge pressure curve)

[0045] M Recording Media

[0046] S substrate

[0047] S102~S105 Adjustment process

[0048] S103, S209: Steps for obtaining the ejection pressure curve

[0049] S211 Process of obtaining the offset

[0050] S102~S105, S207~S210 readjustment process DETAILED DESCRIPTION

[0051] Figure 1 1 is a diagram schematically showing the overall structure of an embodiment of a coating device of the present invention. The coating device 1 is a device for coating a coating material from a Figure 1 A slot coater that applies a coating liquid to the upper surface Sf of a substrate S that is conveyed in a horizontal posture from the left hand side to the right hand side. In addition, in the following figures, in order to clarify the configuration relationship of each part of the device, the conveying direction of the substrate S is set as the "X direction", and the direction from Figure 1 The horizontal direction from the left hand side to the right hand side is called "+X direction", and the opposite direction is called "-X direction". In addition, the direction from the front side to the back side of the device in the horizontal direction Y orthogonal to the X direction is called "-Y direction", and the opposite direction is called "+Y direction". Furthermore, the upward direction and the downward direction in the vertical direction Z are respectively called "+Z direction" and "-Z direction".

[0052] In the coating device 1, an input conveyor 100, an input transfer unit 2, a floating workbench unit 3, an output transfer unit 4, and an output conveyor 110 are arranged in sequence and closely along the conveying direction Dt (+X direction) of the substrate S. As described in detail below, they form a conveying path for the substrate S extending in a substantially horizontal direction. In addition, in the following description, when the positional relationship is expressed in association with the conveying direction Dt of the substrate S, the "upstream side of the conveying direction Dt of the substrate S" is sometimes referred to as the "upstream side", and the "downstream side of the conveying direction Dt of the substrate S" is referred to as the "downstream side". In this example, when observed from a certain reference position, relatively, the (-X) side is equivalent to the "upstream side", and the (+X) side is equivalent to the "downstream side".

[0053] The substrate S to be processed is Figure 1The left-hand side is carried into the input conveyor 100. The input conveyor 100 comprises: a roller conveyor 101; a rotation drive mechanism 102, which rotationally drives the roller conveyor 101, and through the rotation of the roller conveyor 101, the substrate S is transported in a horizontal posture to the downstream side, that is, the (+X) direction. The input transfer section 2 comprises: a roller conveyor 21; a rotation lifting drive mechanism 22, which has the function of rotationally driving the roller conveyor 21 and the function of lifting the roller conveyor 21. By rotating the roller conveyor 21, the substrate S is further transported to the (+X) direction. In addition, by lifting and lowering the roller conveyor 21, the position of the substrate S in the vertical direction Z is changed. Through the input transfer section 2 constructed in this way, the substrate S is transferred from the input conveyor 100 to the floating workbench section 3.

[0054] The floating workbench section 3 is provided with a flat workbench divided into three parts along the conveying direction Dt of the substrate. That is, the floating workbench section 3 is provided with an inlet floating workbench 31, a coating workbench 32 and an outlet floating workbench 33, and the upper surfaces of each of these workbench form a part of the same plane. Furthermore, the floating workbench section 3 has a lifting pin driving mechanism 34, a floating control mechanism 35 and a lifting driving mechanism 36. The lifting pin driving mechanism 34 can lift and lower the lifting pins (not shown) provided on the inlet floating workbench 31. The floating control mechanism 35 can supply compressed air for floating the substrate S to each workbench of the floating workbench section 3. The lifting driving mechanism 36 can lift and lower the outlet floating workbench 33.

[0055] A plurality of ejection holes are provided in a matrix on the upper surfaces of the inlet floating stage 31 and the outlet floating stage 33, respectively. The ejection holes eject compressed air supplied from the floating control mechanism 35, and the substrate S is floated by the buoyancy given by the ejected airflow. In this way, the lower surface Sb of the substrate S is supported in a horizontal posture in a state separated from the upper surface of the stage. The distance between the lower surface Sb of the substrate S and the upper surface of the stage, i.e., the floating amount, can be set, for example, to 10 microns to 500 microns.

[0056] On the other hand, on the upper surface of the coating workbench 32, there are alternately arranged ejection holes for ejecting compressed air and suction holes for sucking air between the lower surface Sb of the substrate S and the upper surface of the workbench. The floating control mechanism 35 precisely controls the distance between the lower surface Sb of the substrate S and the upper surface of the coating workbench 32 by controlling the ejection amount of compressed air from the ejection holes and the suction amount from the suction holes. Thus, the position of the upper surface Sf of the substrate S above the coating workbench 32 in the vertical direction Z is controlled to a specified value. As a specific structure of the floating workbench part 3, for example, the structure described in Japanese Patent No. 5346643 can be applied. In addition, the floating amount on the coating workbench 32 is calculated by the control unit 9 based on the detection results of the sensors 61 and 62 described in detail later, and can be adjusted with high precision through airflow control.

[0057] The substrate S carried into the floating stage section 3 via the input transfer section 2 is given a propulsion force in the (+X) direction by the rotation of the roller conveyor 21, and is carried to 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, but do not have the function of moving the substrate S in the horizontal direction. The substrate S in the floating stage section 3 is carried out by the substrate carrying section 5 arranged below the entrance floating stage 31, the coating stage 32, and the exit floating stage 33.

[0058] The substrate conveying unit 5 includes: a chuck mechanism 51 that supports the substrate S from below by partially contacting the peripheral portion of the lower surface of the substrate S; and an adsorption travel control mechanism 52 that has a function of applying negative pressure to an adsorption pad (not shown) of an adsorption member provided at the upper end of the chuck mechanism 51 to adsorb and hold the substrate S, and a function of reciprocating the chuck mechanism 51 in the X direction. In a state where the chuck mechanism 51 holds the substrate S, the lower surface Sb of the substrate S is located at a position higher than the upper surface of each workbench of the floating workbench unit 3. Therefore, the substrate S is adsorbed and held at the peripheral portion by the chuck mechanism 51, and is maintained in a horizontal posture as a whole by the buoyancy given by the floating workbench unit 3. In addition, in the stage where the lower surface Sb of the substrate S is partially held by the chuck mechanism 51, in order to detect the position of the upper surface of the substrate S in the vertical direction Z, a sensor 61 for measuring the plate thickness is arranged near the roller conveyor 21. Since a chuck (not shown) not holding the substrate S is located directly below the sensor 61 , the sensor 61 can detect the position in the vertical direction Z of the upper surface of the suction member, that is, the suction surface.

[0059] The chuck mechanism 51 holds the substrate S moved from the input transfer unit 2 to the floating stage unit 3. In this state, the chuck mechanism 51 moves in the (+X) direction, and the substrate S is transported from above the entrance floating stage 31 to above the coating stage 32 to above the exit floating stage 33. The transported substrate S is delivered to the output transfer unit 4 disposed on the (+X) side of the exit floating stage 33.

[0060] The output transfer unit 4 includes: a roller conveyor 41; a rotation and lifting drive mechanism 42, which has the function of rotating and driving the roller conveyor 41 and the function of lifting it. By rotating the roller conveyor 41, a propulsion force in the (+X) direction is applied to the substrate S, and the substrate S is further transported along the transport direction Dt. In addition, by lifting and lowering the roller conveyor 41, the position of the substrate S in the vertical direction Z is changed. Through the output transfer unit 4, the substrate S is transferred from the top of the exit floating worktable 33 to the output conveyor 110.

[0061] The output conveyor 110 includes a roller conveyor 111 and a rotation drive mechanism 112 that rotationally drives the roller conveyor 111. By the rotation of the roller conveyor 111, the substrate S is further transported in the (+X) direction and finally discharged from the coating device 1. The input conveyor 100 and the output conveyor 110 can also be provided as a part of the structure of the coating device 1, but can also be separate from the coating device 1. In addition, for example, a substrate discharge mechanism of another unit provided on the upstream side of the coating device 1 can be used as the input conveyor 100. In addition, a substrate receiving mechanism of another unit provided on the downstream side of the coating device 1 can also be used as the output conveyor 110.

[0062] On the conveying path of the substrate S thus conveyed, a coating mechanism 7 is arranged for coating the upper surface Sf of the substrate S with a coating liquid. The coating mechanism 7 includes a slit nozzle (hereinafter referred to as a "nozzle") 71 having a slit-shaped discharge port. In addition, although not shown in the figure, a positioning mechanism is connected to the nozzle 71, and the nozzle 71 is positioned at a coating position ( Figure 1 The coating liquid supply mechanism 8 is connected to the nozzle 71, and the coating liquid is supplied from the coating liquid supply mechanism 8, and the coating liquid is sprayed from the spray port opened downward at the lower part of the nozzle 71. In addition, when "spraying action" is mentioned below, the term refers to the action of spraying the coating liquid from the nozzle 71, especially the action performed when the nozzle 71 is positioned at the coating position.

[0063] Figure 2 is a diagram showing the structure of the coating liquid supply mechanism. Figure 2As shown, the coating liquid supply mechanism 8 includes a pump 81, which is used as a supply source for supplying the coating liquid to the nozzle 71 and supplies the coating liquid by volume change. As the pump 81, for example, a bellows-type pump described in Japanese Patent Publication No. 10-61558 can be used. The pump 81 has a flexible tube 811 that can expand and contract elastically in the radial direction. One end of the flexible tube 811 is connected to the coating liquid replenishing unit 83 through a pipe 82, and the other end is connected to the nozzle 71 through a pipe 84.

[0064] A bellows 812 that is elastically deformable in the axial direction is arranged on the outside of the flexible tube 811. The bellows 812 has a small bellows portion 813 and a large bellows portion 814, and a non-compressible medium is sealed in a pump chamber 815 between the flexible tube 811 and the bellows 812. In addition, a working disk portion 816 is provided between the small bellows portion 813 and the large bellows portion 814. A driving portion 817 is connected to the working disk portion 816. When the driving portion 817 operates according to an instruction from the control unit 9, the working disk portion 816 is displaced in the axial direction in a prescribed movement mode (a mode indicating a change in the speed of the working disk portion 816 over time), so that the volume inside the bellows 812 changes. As a result, the flexible tube 811 expands and contracts in the radial direction to perform a pumping action, and supplies the coating liquid appropriately supplied from the coating liquid replenishing unit 83 to the nozzle 71. Therefore, the movement pattern of the working disk portion 816 is closely related to the discharge characteristics (temporal change in discharge pressure) of the coating liquid discharged from the nozzle 71, and predetermined discharge characteristics can be obtained according to the movement pattern.

[0065] The movement pattern of the working plate portion 816 is realized by the driving portion 817 performing an action based on a pre-set pump control parameter (hereinafter, sometimes referred to as a "control parameter") and moving the working plate portion 816. Therefore, it is the pump control parameter that defines the movement pattern of the working plate portion 816, and if the pump control parameter is changed, the movement pattern of the working plate portion 816 also changes, thereby changing the ejection characteristics of the coating liquid ejected from the nozzle 71.

[0066] The coating liquid replenishing unit 83 has a storage tank 831 for storing the coating liquid. The storage tank 831 is connected to the pump 81 through the pipe 82. In addition, an on-off valve 833 is installed in the pipe 82. The on-off valve 833 is opened according to the replenishment instruction from the control unit 9, and the coating liquid in the storage tank 831 can be replenished into the flexible tube 811 of the pump 81. On the contrary, the on-off valve 833 is closed according to the replenishment stop instruction from the control unit 9, and the coating liquid is replenished from the storage tank 831 to the flexible tube 811 of the pump 81.

[0067] On the output side of the pump 81 ( Figure 2An on-off valve 85 is installed in the piping 84 connected to the nozzle 71 (left-hand side of the nozzle 71), and the on-off valve 85 is opened and closed according to the on-off command from the control unit 9. In this way, the feeding and stopping of the coating liquid to the nozzle 71 can be switched. In addition, a pressure gauge 86 is installed on the piping 84, and the pressure gauge 86 detects the pressure (discharge pressure) of the coating liquid fed to the nozzle 71 and outputs its detection result (pressure value) to the control unit 9.

[0068] In this way, Figure 2 As shown in the figure, a sensor 62 for detecting the floating height of the substrate S is provided on the nozzle 71 to which the coating liquid is supplied from the coating liquid supply mechanism 8. The sensor 62 can measure the separation distance between the floating substrate S and the upper surface of the work table of the coating work table 32, and the control unit 9 controls the positioning mechanism (not shown) according to the detection value, thereby adjusting the position where the nozzle 71 is lowered. In addition, as the sensor 62, an optical sensor, an ultrasonic sensor, etc. can be used.

[0069] return Figure 1 In order to perform prescribed maintenance on the nozzle 71, a nozzle cleaning standby unit 72 is provided on the coating mechanism 7. The nozzle cleaning standby unit 72 mainly includes a roller 721, a cleaning portion 722, a groove portion 723, etc. Then, by performing nozzle cleaning and liquid accumulation formation by these, the nozzle 71 The ejection port is adjusted to a state suitable for the next coating process. Specifically, when the nozzle 71 is positioned at the cleaning position above the nozzle cleaning standby unit 72, the cleaning portion 722 cleans the ejection port at the lower end of the nozzle 71 and its surroundings. In addition, when the nozzle 71 is positioned at the preparatory ejection position directly above the roller 721, the coating liquid is preliminarily ejected from the nozzle 71 to the roller 721, thereby forming an accumulation of liquid around the ejection port. The positions where the nozzle 71 is separated from the coating workbench 32, including these cleaning positions and preparatory ejection positions, are collectively referred to as "maintenance positions".

[0070] In addition, as described later, in this embodiment, the ejection pressure applied to the coating liquid ejected from the nozzle 71 is evaluated, and based on the result, the control parameters of the pump 81 during ejection are optimized. For this purpose, more specifically, in order to measure the ejection pressure during ejection, a simulated ejection of the coating liquid from the nozzle 71 is performed. At this time, the nozzle 71 is positioned at the preliminary ejection position, and the ejection is performed to the roller 721 located directly below.

[0071] Furthermore, the coating device 1 is provided with a control unit 9 ( Figure 3 ). Figure 3 FIG. 1 is a block diagram showing an example of the structure of a control unit. Figure 3As shown in FIG. 1 , the control unit 9 is a computer including a computing unit 91, a storage unit 93, and a UI (User Interface) 95. The computing unit 91 is a processor composed of a CPU (Central Processing Unit) and the like, and by executing a discharge pressure evaluation program 97, a measurement execution unit 911 for executing the measurement of the discharge pressure and a pressure evaluation unit 913 for evaluating the measured discharge pressure are constructed. In addition, by executing a parameter adjustment program 98 by the computing unit 91, a parameter adjustment unit 915 for adjusting the pump control parameter based on the evaluation result of the discharge pressure is constructed.

[0072] The storage unit 93 is a storage device such as a HDD (Hard Disk Drive) or an SSD (Solid State Drive), and stores the discharge pressure evaluation program 97 and discharge pressure measurement data 99 measured as the discharge pressure evaluation program 97 is executed.

[0073] The discharge pressure evaluation program 97 is provided, for example, by a recording medium M provided separately from the control unit 9. The discharge pressure evaluation program 97 is recorded in the recording medium M in a readable manner by a computer (control unit 9). As the recording medium M, for example, a USB (Universal Serial Bus) memory, a memory card, or a storage device of an external server computer can be cited. In addition, the UI 95 has a display for displaying information to the user and an input device for accepting input operations based on the user. As the control unit 9 having such a structure, various computers such as a desktop type, a laptop type, or a tablet type can be used.

[0074] Figure 4 1 is a flowchart showing an example of a method for evaluating the ejection pressure performed based on the ejection pressure evaluation program. In step 101, the nozzle 71 is positioned at a preliminary ejection position directly above the roller 721 of the maintenance unit 72. In this state, the pump 81 is driven by a pre-set control parameter to eject the coating liquid from the nozzle 71 to the roller 721 (simulated ejection) (step S102). Specifically, the measurement execution unit 911 moves the working disk unit 816 based on the control parameters specified in the ejection pressure evaluation program 97, thereby ejecting the coating liquid from the nozzle 71.

[0075] At this time, the output (pressure value) of the pressure gauge 86 is regularly read in a predetermined sampling cycle by the measurement execution unit 911 of the control unit 9. Thus, the measurement execution unit 911 can measure the discharge pressure applied to the coating liquid delivered from the pump 81. The result of measuring the discharge pressure is stored in the storage unit 93 as discharge pressure measurement data 99 corresponding to the time when the discharge of the coating liquid is started. The control unit 9 can obtain a discharge pressure curve representing the change of the discharge pressure over time from the discharge pressure measurement data 99 (step S103).

[0076] Regarding the discharge pressure curve, a target curve required to achieve an ideal discharge state is determined in advance. Specifically, the discharge pressure evaluation program 97 defines control parameters for moving the operating disk portion 816 so that the discharge pressure changes in the following order.

[0077] From the initial pressure Pi to the target pressure Pt which is greater than the initial pressure Pi,

[0078] The discharge pressure is stable at the target pressure Pt,

[0079] The discharge pressure decreases from the target pressure Pt to the initial pressure Pi.

[0080] In step S104, a predetermined offset is added to the acquired ejection pressure curve. The offset is added to compensate for the difference in ejection conditions between the simulated ejection and the ejection operation of ejecting the coating liquid to the substrate S at a position opposite to the substrate S in the parameter adjustment process described later. In other cases, the offset is set to zero or step S104 is skipped.

[0081] In step S105, the pressure evaluation unit 913 evaluates the change of the ejection pressure over time shown in the ejection pressure measurement data 99 according to the specified evaluation items. As described later, this evaluation item extracts a specified feature quantity from the change of the ejection pressure over time shown in the ejection pressure measurement data 99, and evaluates the change of the ejection pressure over time based on the feature quantity. The evaluation result is represented by an evaluation value obtained by quantifying the degree of deviation between the measured ejection pressure curve and the target curve. For various evaluation items prepared to evaluate the change of the ejection pressure over time shown in the ejection pressure measurement data 99, the evaluation values ​​are calculated respectively. Next, each evaluation item for this purpose is described in detail.

[0082] Figure 5A and Figure 5B FIG. 2 is a diagram for explaining each period used in evaluating the discharge pressure. Figure 5A and Figure 5BIn the graph in which the horizontal axis represents time and the vertical axis represents the discharge pressure, the change of the discharge pressure with time (discharge pressure curve) is schematically shown. In addition, the expression of this graph is the same as that in each figure shown later.

[0083] Figure 5A is an example of a discharge pressure curve under an ideal discharge state, and discharge control is performed using this as a target curve. Under an ideal discharge state, after the discharge pressure increases from the initial pressure Pi to the target pressure Pt greater than the initial pressure Pi, it continues to be stable at the target pressure Pt. Then, the discharge pressure decreases from the target pressure Pt to the initial pressure Pi, and discharge ends. On the other hand, in actual operation, Figure 5B As shown, the discharge pressure curve may deviate from the target curve, particularly during the transient period of discharge pressure fluctuation. The degree of deviation of the discharge pressure curve from the target curve is evaluated.

[0084] In this example, the discharge pressure measurement data 99 is acquired from before the coating liquid starts to be discharged from the nozzle 71 to after the coating liquid is discharged from the nozzle 71 (i.e., before and after the entire discharge period Tt). In addition, in this example, the discharge pressure at the time ta when the coating liquid starts to be discharged from the nozzle 71 and the discharge pressure at the time te when the coating liquid is discharged from the nozzle 71 are finished are the initial pressure Pi. However, the pressures at the start and end of discharge may not always be consistent with the initial pressure Pi.

[0085] like Figure 5B As shown, the discharge period Tt can be divided into four periods Ta, Tb, Tc, and Td. The details of the rising period Ta, the transition period Tb, the steady period Tc, and the falling period Td are as follows.

[0086] The rising period Ta is the period from the time ta when the coating liquid supply mechanism 8 starts to discharge the coating liquid from the nozzle 71 (i.e., the time ta when the coating liquid supply mechanism 8 starts to move the operating disk portion 816) to the time tb when the discharge pressure reaches the target pressure Pt. That is, when the coating liquid starts to be discharged from the nozzle 71 at the time ta, the discharge pressure increases from the initial pressure Pi to the target pressure Pt during the period from the time ta to the time tb.

[0087] The transition period Tb 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 of the discharge pressure to stabilize, and is set by, for example, a user input operation on the UI 95 and stored in the storage unit 93 .

[0088] The stable period Tc is a period from time tc to time td when the coating liquid supply mechanism 8 starts to reduce the discharge pressure (i.e., time td when the coating liquid supply mechanism 8 starts to decelerate from the target speed of the operating disk portion 816). That is, the coating liquid supply mechanism 8 moves the operating disk portion 816 at a constant speed during the period from time tc to time td, and starts to decelerate the operating disk portion 816 at time td. In addition, during the stable period Tc, the discharge pressure is basically stable at the target pressure Pt. However, during the stable period Tc, the change in the discharge pressure over time also includes a small vibration, and the discharge pressure becomes larger or smaller than the target pressure Pt.

[0089] The transition period Tb and the stable period Tc constitute a constant voltage period Tbc. That is, the constant voltage period Tbc is a period from time tb to time td.

[0090] The falling period Td is the period from time td to time te when the coating liquid supply mechanism 8 stops spraying the coating liquid from the nozzle 71 (i.e., time te when the coating liquid supply mechanism 8 stops the actuating disk portion 816). That is, the spraying pressure decreases to the initial pressure Pi during the period from time td to time te, and the spraying of the coating liquid from the nozzle 71 stops at time te.

[0091] Figure 6 FIG. 2 is a diagram schematically showing an example of calculation performed by the pressure evaluation unit on the change of the discharge pressure over time. Figure 6 As shown, the pressure evaluation unit 913 calculates the first differential D1 of the change in the discharge pressure over time by differentiating the change in the discharge pressure over time by time. Furthermore, the pressure evaluation unit 913 calculates the second differential D2 of the change in the discharge pressure over time by differentiating the first differential D1 of the change in the discharge pressure over time by time. In addition, the pressure evaluation unit 913 calculates the mean absolute error MAE and the root mean square error RMSE based on the following formulas:

[0092] MAE(α,β)=(1 / n)·(Σ|α-β|),

[0093] RMSE(α,β)=((1 / n)·(Σ(α-β) 2 )) 1 / 2 ,

[0094] n = number of data.

[0095] Figure 7 This is a diagram for explaining evaluation items for evaluating the temporal change in the discharge pressure based on the feature amount Fv1. Figure 7The evaluation item is based on the error (ideal trapezoidal absolute error) between a trapezoidal waveform having an amplitude equivalent to the difference between the average value of the ejection pressure in the stable period Tc (i.e., the stable pressure Pm) and the initial pressure Pi and the ejection pressure measurement data 99 to evaluate the change in the ejection pressure over time shown in the ejection pressure measurement data 99.

[0096] Specifically, a linear regression analysis is performed on the temporal change in the discharge pressure between a predetermined lower reference pressure and a predetermined upper reference pressure greater than the lower reference pressure during the rising period Ta to calculate a rising regression line Lr_R. The rising regression line Lr_R increases linearly from the initial pressure Pi to the steady pressure Pm during the period from time t11 to time t12.

[0097] Similarly, linear regression analysis is performed on the temporal change in discharge pressure between the upper reference pressure and the lower reference pressure in the falling period Td to calculate a falling regression line Lr_F which decreases linearly from the steady pressure Pm to the initial pressure Pi during the period from time t13 to time t14.

[0098] In addition, the lower reference pressure and the upper reference pressure are pressures greater than the initial pressure Pi and less than the target pressure Pt, and are set, for example, by a user input operation on the UI 95 and stored in the storage unit 93. In this example, the lower reference pressure is a pressure obtained by adding a pressure of 20% of the absolute value of the difference between the initial pressure Pi and the target pressure Pt to the initial pressure Pi, and the upper reference pressure is a pressure obtained by adding a pressure of 80% of the absolute value of the difference between the initial pressure Pi and the target pressure Pt to the initial pressure Pi.

[0099] In addition, the start approximate straight line Lr_s is set for the interval from time ta to time t11. The start approximate straight line Lr_s is a straight line whose slope is zero, indicating the initial pressure Pi. That is, the start approximate straight line Lr_s is a straight line connecting the start time (time ta) of the ejection start of the coating liquid from the nozzle 71 to the start time of the rising regression line Lr_R. In addition, depending on the state (slope) of the regression line, sometimes time t11 is earlier than time ta, and sometimes time t12 is later than time tb. As a result, when t11<ta, the start approximate straight line Lr_s is omitted.

[0100] In addition, the end approximate straight line Lr_e is set for the interval from time t14 to time te. The end approximate straight line Lr_e is a straight line whose slope indicating the initial pressure Pi is zero. That is, the end approximate straight line Lr_e is a straight line connecting the end time of the descending regression straight line Lr_F to the end time (time te) of the ejection of the coating liquid from the nozzle 71. In addition, in the case of te<t14, the end approximate straight line Lr_e is omitted.

[0101] Furthermore, a stable straight line Lr_m is set for the interval from time t12 to time t13. The stable straight line Lr_m is a straight line with a slope of zero indicating the stable pressure Pm. That is, the stable straight line Lr_m is a straight line indicating the stable pressure Pm connecting the end time (time t12) of the rising regression line Lr_R and the start time (time t13) of the falling regression line Lr_F.

[0102] In this way, the approximate waveform WF1 composed of the start approximate straight line Lr_s, the rising regression line Lr_R, the stable straight line Lr_m, the falling regression line Lr_F, and the end approximate straight line Lr_e arranged in time series is calculated. Then, the pressure evaluation unit 913 calculates the mean absolute error MAE (ideal trapezoidal absolute error) between the discharge pressure measurement data 99 and the approximate waveform WF1 as the feature value Fv1 during the entire discharge period Tt from time ta to time te. In addition, the pressure evaluation unit 913 normalizes the feature value Fv1 to a range of greater than 0 and less than 2 based on a predetermined threshold value Th1 (for example, 0.05).

[0103] Specifically, the feature value Fv1 is transformed into a standardized feature value Fv1 (ie, evaluation value V1) based on the following formula:

[0104] If Fv1<Th1, then Fv1=0;

[0105] If Fv1≥Th1, then Fv1=(Fv1+1-Th1)×c1.

[0106] Here, the coefficient c1 is a normalization coefficient, and is set in advance to a value (for example, 15) within a range where the feature quantity Fv1 is 2 or less.

[0107] Based on Figure 7 In the evaluation of the feature amount Fv1, when the temporal change of the overall discharge pressure during the discharge period Tt greatly deviates from the ideal shape (ie, trapezoidal shape), a large score (ie, poor evaluation) can be given to the discharge pressure.

[0108] Figure 8 This is a diagram for explaining evaluation items for evaluating the change in discharge pressure over time based on the feature amount Fv2. Figure 8 The evaluation item evaluates the smoothness of the rise of the ejection pressure. Specifically, during the rise period Ta, a curve regression analysis is performed on the change in the ejection pressure between the lower reference pressure P2_1 and the upper reference pressure P2_u which is larger than the lower reference pressure P2_1 over time, and the rise regression curve Nr is calculated. The curve regression analysis is performed by a quadratic curve.

[0109] The lower reference pressure P2_l is set to the initial pressure Pi. On the other hand, the upper reference pressure P2_u is a pressure greater than the lower reference pressure P2_l and less than the target pressure Pt, and is set, for example, based on the user's input operation to the UI95 and stored in the storage unit 93. In this example, the upper reference pressure P2_u is a pressure obtained by adding 20% ​​of the absolute value of the difference between the initial pressure Pi and the target pressure Pt to the initial pressure Pi. The rising regression curve Nr increases from the lower reference pressure P2_l (initial pressure Pi) to the upper reference pressure P2_u from the time t21 to the time t22. In addition, the time t21 coincides with the time ta, and the time t22 is a time later than the time ta and earlier than the time tb.

[0110] In this way, the waveform WF2 composed of the rising regression curve Nr is calculated. Then, the pressure evaluation unit 913 calculates the root mean square error RMSE between the discharge pressure measurement data 99 and the waveform WF2 as the feature quantity Fv2 in the rising initial period Ta_s from time t21 to time t22. In addition, the pressure evaluation unit 913 standardizes the feature quantity Fv2 to a range of greater than 0 and less than 2 based on a predetermined threshold value Th2 (for example, 0.05). Specifically, the feature quantity Fv2 is converted into the standardized feature quantity Fv2 (i.e., the evaluation value V2) based on the following formula:

[0111] If Fv2<Th2, then Fv2=0;

[0112] If Fv2≥Th2, then Fv2=2.

[0113] Based on Figure 8 The characteristic quantity Fv2 is affected by the state before the coating liquid starts to be ejected from the nozzle 71. If the ejection pressure is abnormal just after the ejection starts, the ejection pressure may be rated higher (i.e., poorer). In addition, the curve that can be used in the curve regression analysis is not limited to the quadratic curve, but may be other curves such as an exponential function.

[0114] Fig. 9 This is a diagram for explaining evaluation items for evaluating the temporal change in the discharge pressure based on the feature amount Fv3. Fig. 9 The evaluation item evaluates whether the rising period Ta is controlled within a certain period. Specifically, the pressure evaluation unit 913 calculates the length of the rising period Ta from time ta to time tb required for the ejection pressure to increase from the initial pressure Pi to the target pressure Pt (= tb-ta) as the feature value Fv3. In addition, the pressure evaluation unit 913 standardizes the feature value Fv3 within a range of greater than 0 and less than 1 based on a specified threshold value Th3 (for example, 350ms). Specifically, based on the following formula, the feature value Fv3 is converted into the standardized feature value Fv3 (i.e., the evaluation value V3):

[0115] If Fv3<Th3, then Fv3=0,

[0116] If Fv3≥Th3, then Fv2=1.

[0117] Based on Fig. 9 The evaluation of the feature amount Fv3 can give a larger score (ie, a poorer evaluation) to the discharge pressure that requires a longer time to rise to the target pressure Pt.

[0118] Fig. 10A and Fig. 10B 4 is a diagram for explaining an evaluation item for evaluating the change in the ejection pressure over time based on the feature value Fv4. More specifically, Fig. 10A is a diagram for explaining evaluation items for evaluating the change in discharge pressure over time. Fig. 10B This is a diagram showing an example of the change over time of the discharge pressure determined to be inappropriate by the evaluation based on the feature amount Fv4. Fig. 10A The evaluation item evaluates the presence or absence of abnormality in the rise of the discharge pressure. Specifically, the pressure evaluation unit 913 calculates the first differential D1 of the change in the discharge pressure with time for the rise period Ta from time ta to time tb, and obtains the first differential waveform WF4.

[0119] Then, the pressure evaluation unit 913 obtains the number of times the first-order differential waveform WF4 crosses the predetermined threshold value Th4 during the rising period Ta as the feature value Fv4. Fig. 10A In the example, the first differential waveform WF4 and the threshold Th4 (for example, 0.002) intersect at time t41 and time t42, respectively, and the number of intersections (feature value Fv4) is two times. In addition, the pressure evaluation unit 913 normalizes the feature value Fv4 to a range of greater than 0 and less than 1. Specifically, the feature value Fv4 is converted into the normalized feature value Fv4 (i.e., the evaluation value V4) based on the following formula:

[0120] If Fv4≤2, then Fv4=0

[0121] If Fv4>2, then Fv4=1,

[0122] Based on Fig. 10A In the evaluation of the characteristic quantity Fv4, when a step occurs in the temporal change of the ejection pressure during the rising period Ta (for example, Fig. 10B As shown in FIG. 1 , a large score (i.e., a poor evaluation) can be given to the ejection pressure.

[0123] Fig.11A and Fig. 11B : is a diagram for explaining the evaluation items for evaluating the change of the ejection pressure with time based on the feature value Fv5. More specifically, Fig.11AThis is a diagram for explaining evaluation items for evaluating the change in discharge pressure over time. Fig. 11B This is a diagram showing an example of the change over time of the discharge pressure determined to be inappropriate by the evaluation based on the feature amount Fv5. Fig.11A The evaluation item evaluates the presence or absence of abnormality in the rise of the discharge pressure. Specifically, the pressure evaluation unit 913 calculates the secondary differential D2 of the change in the discharge pressure with time for the rise period Ta from time ta to time tb, and obtains the secondary differential waveform WF5.

[0124] Then, the pressure evaluation unit 913 obtains the number of times the absolute value of the second derivative waveform WF5 crosses the predetermined threshold value Th5 during the rising period Ta as the feature value Fv5. Fig.11A In the example, the absolute value of the secondary differential waveform WF5 crosses the threshold Th5 (for example, 0.0002) at time t51, t52, t53, and t54, respectively, and the number of crossovers (feature value Fv5) is four times. In addition, the pressure evaluation unit 913 standardizes the feature value Fv5 to a range of greater than 0 and less than 1. Specifically, based on the following formula, the feature value Fv5 is converted into a standardized feature value Fv5 (i.e., evaluation value V5):

[0125] If Fv5=4, then Fv5=0,

[0126] If Fv5≠4, then Fv5=1.

[0127] Based on Fig.11A In the evaluation of the characteristic quantity Fv5, when a step occurs in the temporal change of the ejection pressure during the rising period Ta (for example, Fig. 11B As shown in FIG. 1 , a larger score (i.e., a poor evaluation) can be given to the ejection pressure.

[0128] Fig.12 This is a diagram for explaining evaluation items for evaluating the temporal change in the discharge pressure based on the feature amount Fv6. Fig.12 The evaluation item evaluates whether the discharge pressure stalls in the second half of the rise. Specifically, the pressure evaluation unit 913 calculates the second differential D2 of the discharge pressure change with time for the rise period Ta from time ta to time tb, and obtains the second differential waveform WF6.

[0129] Then, the pressure evaluation unit 913 calculates the time T_1st when the secondary differential waveform WF6 is greater than the prescribed positive threshold value (Th5) and the time T_2nd when the secondary differential waveform WF6 is less than the prescribed negative threshold value (-Th5) during the rising period Ta. Here, the positive threshold and the negative threshold have the same absolute value (Th5) and have different signs from each other. The absolute value (Th5) of the positive and negative thresholds is equal to the absolute value of the threshold Th5 used in the evaluation based on the above-mentioned characteristic quantity Fv5. Then, the pressure evaluation unit 913 calculates the ratio of their times (=T_1st / T_2nd) as the characteristic quantity Fv6. Furthermore, the pressure evaluation unit 913 converts the characteristic quantity Fv6 based on the following formula:

[0130] Fv6=|1-Fv6|.

[0131] In addition, the pressure evaluation unit 913 uses a predetermined threshold value Th6 (for example, 0.2) to normalize the feature quantity Fv6 converted in this way to a range of not less than 0 and not more than 2. Specifically, the feature quantity Fv6 is converted into a normalized feature quantity Fv6 (ie, evaluation value V6) based on the following formula:

[0132] If Fv6<Th6, then Fv6=0;

[0133] If Fv6≥Th6, then Fv6=f(Fv6);

[0134] f(γ)=4×γ-0.8.

[0135] In addition, the function f(γ) having γ as a variable is not limited to the example shown here, and can be changed arbitrarily.

[0136] The conveying speed of the substrate S to be coated with the coating liquid does not stall even in the second half of the acceleration period and reaches the target speed. Therefore, it is preferred that the ejection pressure applied to the coating liquid does not stall during the rising period Ta and reaches the target pressure Pt. Fig.12 In the evaluation of the characteristic amount Fv6, when the discharge pressure stalls during the rising period Ta, a large score (that is, a poor evaluation) can be given to the discharge pressure.

[0137] Fig.13 This is a diagram for explaining evaluation items for evaluating the change in discharge pressure over time based on the feature amount Fv7. Fig.13The evaluation item evaluates the sharpness of the change in the ejection pressure over time at the end of the rise. Specifically, a linear regression analysis is performed on the change in the ejection pressure over time between the lower reference pressure P7_1 and the upper reference pressure P7_u which is greater than the lower reference pressure P7_1 during the rise period Ta, and the regression line Lr at the end of the rise is calculated. Here, the lower reference pressure P7_1 is the pressure obtained by adding 80% of the absolute value of the difference between the initial pressure Pi and the target pressure Pt to the initial pressure Pi. The upper reference pressure P7_u is the pressure obtained by adding 90% of the absolute value of the difference between the initial pressure Pi and the target pressure Pt to the initial pressure Pi. The ejection pressure increases from the lower reference pressure P7_1 to the upper reference pressure P7_u during the period from time t71 to time t72.

[0138] The rising end regression line Lr increases with the passage of time and reaches the stable pressure Pm (the average value of the discharge pressure during the stable period Tc) at time t73. In this way, the rising end regression line Lr is set for the interval from time t71 to time t73. Furthermore, the pressure evaluation unit 913 sets an extended setting line Lm with a slope of zero representing the stable pressure Pm during the period from time t73 to time tb. As described above, time tb is the moment when the discharge pressure reaches the target pressure Pt, which is equivalent to the end time of the rising period Ta. That is, the extended setting line Lm is set in a manner extending from the rising end regression line Lr (the end time point) to the end time point of the rising period Ta. In addition, in the case of tb<t73, the extended setting line Lm is omitted.

[0139] In this way, the approximate waveform WF7 composed of the rising end regression straight line Lr and the extended setting straight line Lm arranged in time series is calculated. Then, the pressure evaluation unit 913 calculates the value representing the difference between the discharge pressure measurement data 99 and the approximate waveform WF7 as the feature value Fv7 in the rising end period Ta_e from the time t72 when the discharge pressure becomes 90% of the target pressure Pt to the time tb when it becomes 100%. Specifically, the weighted reference time width Tw is set to t73-t72. And the weighted square root error sum is calculated based on the following formula:

[0140] Fv7=(Σ(P_measure-WF7) 2 ×W) 1 / 2 ,

[0141] P_measure = discharge pressure measurement data 99,

[0142] In the range of time tb≤t73+2×Tw, W=1,

[0143] In the range of time tb>t73+2×Tw, W=w,

[0144] W is a weight coefficient greater than 1, for example, 10.

[0145] In addition, the pressure evaluation unit 913 normalizes the feature value Fv7 to a range of not less than 0 and not more than 2 based on a predetermined threshold value Th7 (for example, 0.6). Specifically, the feature value Fv7 is converted into a normalized feature value Fv7 (ie, evaluation value V7) based on the following formula:

[0146] If Fv7<Th7, then Fv7=0,

[0147] If Fv7≥Th7, then Fv7=Fv7 / c7,

[0148] c7 is an arbitrary positive constant, for example, 1.1.

[0149] Based on Fig.13 In the evaluation of the feature quantity Fv7, when the rising momentum is weak and the change in the discharge pressure over time shows a waveform with rounded corners, a large score (ie, a poor evaluation) can be given to the discharge pressure.

[0150] Fig.14 This is a diagram for explaining evaluation items for evaluating the temporal change in the discharge pressure based on the feature amount Fv8. Fig.14 The evaluation item evaluates the degree of overshoot generated in the rise of the discharge pressure. Specifically, the pressure evaluation unit 913 obtains the sign (positive / negative) of the secondary differential D2 of the discharge pressure at the time t81 when the discharge pressure reaches the maximum value Pmax. Then, the pressure evaluation unit 913 calculates the time t82 when the sign of the secondary differential D2 of the discharge pressure switches twice from the sign at the time t81. Then, the change in the discharge pressure over time in the initial vibration period Tb_s from the time t81 to the time t82 is evaluated.

[0151] Specifically, the minimum value P8min of the change in the ejection pressure over time during the initial vibration period Tb_s is obtained, and the smaller pressure between the stable pressure Pm and the pressure P8min is selected as the target pressure Pg. Then, the feature value Fv8 is calculated based on the difference between the maximum pressure Pmax and the target pressure Pg, that is, the following formula:

[0152] Fv8=Pmax-Pg.

[0153] Furthermore, the pressure evaluation unit 913 uses a predetermined threshold value Th8 (for example, 0.035) to normalize the feature value Fv8 to a range of 0 to 2. Specifically, the feature value Fv8 is converted into a normalized feature value Fv8 (ie, evaluation value V8) based on the following formula:

[0154] If Fv8<Th8, then Fv8=0,

[0155] If Fv8≥Th8, then Fv8=Fv8 / c8,

[0156] c8 is an arbitrary positive constant, for example, 0.12.

[0157] Based on Fig.14 In the evaluation of the feature quantity Fv8, when the temporal change of the discharge pressure shows a strong upward momentum and a large overshoot, a large score (ie, a poor evaluation) can be given to the discharge pressure.

[0158] Fig.15 This is a diagram for explaining evaluation items for evaluating the temporal change in the discharge pressure based on the feature amount Fv9. Fig.15 The evaluation item evaluates the stability of the change of the ejection pressure over time in the transition period Tb. Specifically, the pressure evaluation unit 913 calculates the root mean square error RMSE (P_measure, Pm) as the feature value Fv9 based on the following formula for the ejection pressure in the transition period Tb and the stable pressure Pm which is the average value of the ejection pressure in the stable period Tc:

[0159] Fv9 = RMSE (P_measure, Pm)

[0160] P_measure=discharge pressure measurement data 99.

[0161] Furthermore, the pressure evaluation unit 913 normalizes the feature value Fv9 to a range of not less than 0 and not more than 2. Specifically, the feature value Fv9 is converted into a normalized feature value Fv9 (ie, evaluation value V9) based on the following formula:

[0162] Fv9=Fv9 / c9,

[0163] c9 is an arbitrary positive constant, for example, 0.04.

[0164] Based on Fig.15 In the evaluation of the feature amount Fv9, when the temporal change in the discharge pressure shows ringing in the transition period Tb, a large score (ie, a poor evaluation) can be given to the discharge pressure.

[0165] Fig.16 This is a diagram for explaining evaluation items for evaluating the temporal change in the discharge pressure based on the feature amount Fv10. Fig.16 The evaluation item evaluates the stability of the change of the discharge pressure over time during the constant pressure period Tbc. Specifically, the pressure evaluation unit 913 obtains the maximum value Pmax and the minimum value P10min of the discharge pressure during the constant pressure period Tbc. In addition, the pressure evaluation unit 913 calculates the feature value Fv10 based on the difference between the maximum pressure Pmax and the minimum pressure P10min during the constant pressure period Tbc, that is, the following formula:

[0166] Fv10=Pmax-P10min.

[0167] Furthermore, the pressure evaluation unit 913 uses a threshold value Th10 (for example, 0.12) to normalize the feature value Fv10 to a range of not less than 0 and not more than 2. Specifically, the feature value Fv10 is converted into a normalized feature value Fv10 (ie, evaluation value V10) based on the following formula:

[0168] If Fv10<Th10, then Fv10=0,

[0169] If Fv10≥Th10, then Fv10=Fv10 / Th10.

[0170] Based on Fig.16 In the evaluation of the feature quantity Fv10, when the temporal change of the discharge pressure in the stable period Tc which has a large influence on the film thickness of the coating liquid shows a large deviation, a large score (ie, a poor evaluation) can be given to the discharge pressure.

[0171] Fig.17 This is a diagram for explaining evaluation items for evaluating the temporal change in the discharge pressure based on the feature amount Fv11. Fig.17 The evaluation item evaluates the linearity of the change of the ejection pressure with time during the rising period Ta. Specifically, during the rising period Ta, the pressure rising period Tar from the moment t111 when the ejection pressure becomes the lower reference pressure P11_1 to the moment t113 when the ejection pressure becomes the upper reference pressure P11_u which is greater than the lower reference pressure P11_1 is calculated. Here, the lower reference pressure P11_1 is the pressure obtained by adding 20% ​​of the absolute value of the difference between the initial pressure Pi and the target pressure Pt to the initial pressure Pi. The upper reference pressure P11_u is the pressure obtained by adding 80% of the absolute value of the difference between the initial pressure Pi and the target pressure Pt to the initial pressure Pi. During the period from the moment t111 to the moment t113, the ejection pressure increases from the lower reference pressure P11_1 to the upper reference pressure P11_u.

[0172] Furthermore, the pressure evaluation unit 913 obtains specific data Dm satisfying a predetermined condition from the discharge pressure measurement data 99 between the lower reference pressure P11_1 and the upper reference pressure P11_u. The predetermined condition will be described in detail below.

[0173] From the ejection pressure measurement data 99, one of the measurement data between the lower reference pressure P11_1 and the upper reference pressure P11_u is selected as the candidate data Dc. Then, the regression line obtained by performing a linear regression analysis on the change of the ejection pressure over time between the measurement data D1 representing the lower reference pressure P11_1 and the candidate data Dc is set as the approximate straight line Lr_1. In addition, the regression line obtained by performing a linear regression analysis on the change of the ejection pressure over time between the measurement data Du representing the upper reference pressure P11_u and the candidate data Dc is set as the approximate straight line Lr_2. Furthermore, in the interval between the measurement data D1 and the candidate data Dc, the root mean square error RMSE (P_measure, Lr1) between the change of the ejection pressure over time and the approximate straight line Lr_1 is obtained. Similarly, in the interval between the candidate data Dc and the measurement data Du, the root mean square error RMSE (P_measure, Lr2) between the change of the ejection pressure over time and the approximate straight line Lr_2 is obtained. Here, P_measure represents discharge pressure measurement data 99 .

[0174] Furthermore, their sum is calculated based on the following formula.

[0175] Er=RMSE(P_measure,Lr1)+RMSE(P_measure,Lr2)

[0176] Then, the candidate data Dc in which the sum Er is the smallest among the discharge pressure measurement data 99 is specified as the specific data Dm.

[0177] Then, the feature value Fv11 is calculated based on the following formula, with the slopes of the above-mentioned approximate straight line Lr_1 and the approximate straight line Lr_2 obtained for the specific data Dm being K1 and K2, respectively:

[0178] If K1>K2, then Fv11=1-K1 / K2,

[0179] If K1≤K2, then Fv11=1-K2 / K1.

[0180] In addition, the pressure evaluation unit 913 normalizes the feature quantity Fv11 to a range of not less than 0 and not more than 1 based on a predetermined threshold value Th11 (for example, 0.25). Specifically, the feature quantity Fv11 is converted into a normalized feature quantity Fv11 (ie, evaluation value V11) based on the following formula:

[0181] If Fv11<Th11, then Fv11=0,

[0182] If Fv11≥Th11, then Fv11=f(Fv11),

[0183] f(γ)=4×γ-1.

[0184] In addition, the function f(γ) having γ as a variable is not limited to the example shown here, and can be changed arbitrarily.

[0185] Based on Fig.17 In the evaluation of the feature quantity Fv11, when the discharge pressure during the rising period Ta shows a poor linear time change, a large score (ie, a poor evaluation) can be given to the discharge pressure.

[0186] Fig.18 This is a diagram for explaining evaluation items for evaluating the temporal change in the discharge pressure based on the feature amount Fv12. Fig.18 The evaluation item evaluates the sharpness of the change in the ejection pressure over time at the end of the rise. Specifically, a linear regression analysis is performed on the change in the ejection pressure between the lower reference pressure P12_1 and the upper reference pressure P12_u which is greater than the lower reference pressure P12_1 during the rise period Ta, and the regression line Lr at the end of the rise is calculated. Here, the lower reference pressure P12_1 is the pressure obtained by adding 70% of the absolute value of the difference between the initial pressure Pi and the target pressure Pt to the initial pressure Pi. The upper reference pressure P12_u is the pressure obtained by adding 90% of the absolute value of the difference between the initial pressure Pi and the target pressure Pt to the initial pressure Pi. During the period from time t121 to time t122, the ejection pressure increases from the lower reference pressure P12_1 to the upper reference pressure P12_u.

[0187] The rising end regression line Lr increases with the passage of time and reaches the target pressure Pt at time t123. In this way, the rising end regression line Lr is set for the interval from time t121 to time t123. Furthermore, the pressure evaluation unit 913 sets an extended setting straight line Lm with a slope of zero representing the target pressure Pt during the period from time t123 to time tb. As described above, time tb is the moment when the discharge pressure reaches the target pressure Pt, which is equivalent to the end time of the rising period Ta. That is, the extended setting straight line Lm is set in a manner extending from (the end time point of) the rising end regression line Lr to the end time point of the rising period Ta. In addition, in the case of tb<t123, the extended setting straight line Lm is omitted.

[0188] In this way, the approximate waveform WF12 composed of the rising end regression straight line Lr and the extended setting straight line Lm arranged in time series is calculated. Then, the pressure evaluation unit 913 calculates the value representing the difference between the discharge pressure measurement data 99 and the approximate waveform WF12 as the feature value Fv12 in the rising end period Ta_e from the time t122 ​​when the discharge pressure becomes 90% of the target pressure Pt to the time tb when it becomes 100%. Specifically, the square root error sum is calculated based on the following formula:

[0189] Fv12=(Σ(P_measure-WF12) 2 ) 1 / 2 ,

[0190] P_measure=discharge pressure measurement data 99.

[0191] In addition, the pressure evaluation unit 913 normalizes the feature quantity Fv12 to a range of not less than 0 and not more than 1 based on a predetermined threshold value Th12 (for example, 0.8). Specifically, the feature quantity Fv12 is converted into a normalized feature quantity Fv12 (ie, evaluation value V12) based on the following formula:

[0192] If Fv12<Th12, then Fv12=0,

[0193] If Fv12 ≥ Th12, then Fv12 = f(Fv12),

[0194] f(γ)=5×γ-4.

[0195] In addition, the function f(γ) having γ as a variable is not limited to the example shown here, and can be changed arbitrarily.

[0196] Based on Fig.18 In the evaluation of the characteristic amount Fv12, when the rising momentum is weak and the change in the discharge pressure over time shows a waveform with rounded corners as a whole, a large score (ie, a poor evaluation) can be given to the discharge pressure.

[0197] Fig.19 This is a diagram for explaining evaluation items for evaluating the change in discharge pressure over time based on the feature amount Fv13. Fig.19 The evaluation item evaluates the sharpness of the change in the ejection pressure over time at the end of the rise. Specifically, a linear regression analysis is performed on the change in the ejection pressure over time between the lower reference pressure P13_1 and the upper reference pressure P13_u which is larger than the lower reference pressure P13_1 during the rise period Ta to calculate the regression line Lr at the end of the rise. Here, the lower reference pressure P13_1 is the pressure obtained by adding 90% of the absolute value of the difference between the initial pressure Pi and the target pressure Pt to the initial pressure Pi. The upper reference pressure P13_u is the pressure obtained by adding 95% of the absolute value of the difference between the initial pressure Pi and the target pressure Pt to the initial pressure Pi. During the period from time t131 to time t132, the ejection pressure increases from the lower reference pressure P13_1 to the upper reference pressure P13_u.

[0198] The rising end regression line Lr increases with the passage of time and reaches the target pressure Pt at time t133. In this way, the rising end regression line Lr is set for the interval from time t131 to time t133. Furthermore, the pressure evaluation unit 913 sets an extended setting line Lm with a slope of zero representing the target pressure Pt in the period from time t133 to time tb. As described above, time tb is the moment when the discharge pressure reaches the target pressure Pt, which is equivalent to the end time of the rising period Ta. That is, the extended setting line Lm is set in a manner extending from the rising end regression line Lr (the end time point) to the end time point of the rising period Ta. In addition, in the case of tb<t133, the extended setting line Lm is omitted.

[0199] In this way, the approximate waveform WF13 composed of the rising end regression straight line Lr and the extended setting straight line Lm arranged in time series is calculated. Then, the pressure evaluation unit 913 calculates the value representing the difference between the discharge pressure measurement data 99 and the approximate waveform WF13 as the feature value Fv13 in the rising end period Ta_e from the time t132 when the discharge pressure becomes 95% of the target pressure Pt to the time tb when it becomes 100%. Specifically, the square root error sum is calculated based on the following formula:

[0200] Fv13=(Σ(P_measure-WF13) 2 ) 1 / 2 ,

[0201] P_measure=discharge pressure measurement data 99.

[0202] In addition, the pressure evaluation unit 913 normalizes the feature quantity Fv13 to a range of not less than 0 and not more than 1 based on a predetermined threshold value Th13 (for example, 0.1). Specifically, the feature quantity Fv13 is converted into a normalized feature quantity Fv13 (ie, evaluation value V13) based on the following formula:

[0203] If Fv13<Th13, then Fv13=0,

[0204] If Fv13 ≥ Th13, then Fv13 = f(Fv13),

[0205] f(γ)=(10×γ-1) / 3.

[0206] In addition, the function f(γ) having γ as a variable is not limited to the example shown here, and can be changed arbitrarily.

[0207] Based on Fig.19 In the evaluation of the feature quantity Fv13, when the upward momentum is weak and the change in the discharge pressure over time shows a waveform with rounded corners as a whole, a large score (ie, a poor evaluation) can be given to the discharge pressure.

[0208] Fig. 20 This is a diagram for explaining evaluation items for evaluating the change in discharge pressure over time based on the feature amount Fv14. Fig. 20 The evaluation item evaluates the degree of overshoot generated in the rise of the discharge pressure. Specifically, the pressure evaluation unit 913 obtains the sign (positive / negative) of the secondary differential D2 of the discharge pressure at the time t141 when the discharge pressure reaches the maximum value Pmax. In addition, the pressure evaluation unit 913 calculates the time t142 when the sign of the secondary differential D2 of the discharge pressure switches twice from the sign at the time t141. Then, the change in the discharge pressure over time in the initial vibration period Tb_s from the time t141 to the time t142 is evaluated.

[0209] Specifically, the minimum value P14min of the change in the discharge pressure over time during the initial vibration period Tb_s is obtained. Then, the difference between the maximum pressure Pmax and the stable pressure Pm is calculated based on the following formula:

[0210] OVER=Pmax-Pm,

[0211] The difference between the stable pressure Pm and the minimum pressure P14min is calculated based on the following formula:

[0212] UNDER=Pm-P14min.

[0213] Then, the feature value Fv14 is calculated based on the sum of OVER and UNDER, that is, the following formula:

[0214] Fv14=OVER+UNDER.

[0215] Furthermore, the pressure evaluation unit 913 uses a predetermined threshold value Th14 (for example, 0.01) to normalize the feature value Fv14 to a range of 0 to 1. Specifically, the feature value Fv14 is converted into a normalized feature value Fv14 (ie, evaluation value V14) based on the following formula:

[0216] If Fv14<Th14, then Fv14=0,

[0217] If Fv14 ≥ Th14, then Fv14 = f(Fv14),

[0218] f(γ)=(100×γ-1) / 9.

[0219] Based on Fig. 20 The error from the target pressure Pt is calculated for each of the overshoot point and the backwash point (a phenomenon in which the pressure drops due to the reaction after a sharp rise) in the evaluation of the characteristic quantity Fv14. Therefore, when the change of the ejection pressure over time shows a large overshoot or backwash, a large score (i.e., a poor evaluation) can be given to the ejection pressure.

[0220] The above is a description of the various evaluation items that can be used to evaluate the change of the ejection pressure over time, and the characteristic quantities Fv1 to Fv14 extracted from each evaluation item and the evaluation values ​​V1 to V14. In this way, for each evaluation item, the evaluation values ​​V1 to V14 that quantitatively represent the degree of deviation between the measured ejection pressure curve and the target curve are obtained. In each evaluation value V1 to V14, the closer the measured result is to the target curve, the smaller the score is given. On the other hand, if the deviation from the target curve becomes larger, the score becomes larger. In this way, the degree of deviation between the measured ejection pressure curve and the target curve is quantitatively represented by an evaluation value with a larger score as the deviation becomes larger.

[0221] Next, the optimization process of the pump control parameters based on the evaluation results is described. The evaluation method described so far quantitatively indicates how close (or deviates) the discharge pressure curve is to the target curve when the discharge is performed using the currently set control parameters. In fact, in the discharge operation when the coating liquid is applied to the substrate S, the control parameters need to be set in such a way that the discharge pressure curve is as close to the target curve as possible.

[0222] Therefore, not only the evaluation is performed, but also the control parameters need to be changed and set according to the evaluation results so that the ejection pressure curve approaches the target curve. In the control unit 9, the computing unit 91 executes the parameter adjustment program 98 to adjust the control parameters so that the ejection pressure curve approaches the target curve. In addition, the control parameters may be one or more.

[0223] Fig.21 201 is a flowchart showing an example of parameter adjustment processing performed based on a parameter adjustment program. In step S201, an initialization process is performed. More specifically, in the initialization process, the offset is set to zero. In addition, the pump control parameter is set to an appropriate initial value. The initial value of the control parameter can be prepared in advance as a default value, and, for example, it can also be the set value last used in the past ejection action. In addition, it can also be set to add a random value to these values ​​and intentionally move away from the optimal value.

[0224] In step S202, the above-mentioned discharge pressure evaluation process ( Figure 4 ). That is, the nozzle 71 is positioned at the preliminary ejection position (step S101), simulated ejection using the set control parameters is performed (step S102), and the ejection pressure curve at this time is obtained (step S103). Since the offset is initially set to zero, step S104 is essentially skipped, and the acquired ejection pressure curve is evaluated (step S105) to obtain evaluation values ​​V1 to V14. The scores of each evaluation value V1 to V14 are stored in the storage unit 93 in association with the setting values ​​of the control parameters for obtaining them.

[0225] Then, it is determined whether the scores represented by these evaluation values ​​V1 to V14 are less than a predetermined threshold value (step S203). Here, the evaluation of the scores can also be performed based on threshold values ​​set separately for the evaluation values ​​V1 to V14. Alternatively, the scores of these evaluation values ​​V1 to V14 can be applied to appropriate calculation formulas to obtain a single or several summarized scores, and then compared with the threshold values.

[0226] When the obtained discharge pressure curve is sufficiently close to the target curve, the score becomes smaller. On the other hand, if the discharge pressure curve deviates greatly from the target curve, the score becomes larger. Therefore, if the judgment in step S203 is "no", that is, if the discharge pressure curve deviates greatly from the target curve beyond the allowable range, the control parameter with the smallest score is searched (step S204).

[0227] In the search for control parameters at this time, an appropriate machine learning algorithm can be applied to infer the control parameters that obtain a smaller score based on the relationship between the control parameters set in the past and the scores obtained therefrom. According to the purpose of this embodiment, a machine learning algorithm using a regression model is preferably used, for example, Bayesian optimization method can be appropriately applied.

[0228] Since the Bayesian optimization method is well known, a detailed description is omitted here. However, in the application of this embodiment, based on the relationship between the control parameter and the score obtained previously, the value of the control parameter with a smaller score is searched, and finally the value of the control parameter with a high probability of being inferred to have the smallest score is output.

[0229] The value thus estimated is changed and set as a new control parameter (step S205), and the process returns to step S202 to perform the discharge pressure evaluation process again. The above process is repeated until the score is lower than the threshold value ("yes" in step S203), that is, until the deviation between the discharge pressure curve and the target curve converges within the allowable range. As a result, the discharge pressure curve in the simulated discharge can be shown to be very close to the target curve. That is, the control parameter can be adjusted in such a way that the discharge pressure curve in the simulated discharge is sufficiently close to the target curve.

[0230] In step S206 executed when the score is lower than the threshold, if the prescribed processing end condition is met, the process jumps to step S212 and the setting value of the control parameter at this time is determined as the optimal value. The processing end condition will be described later, but the processing end condition is not met at this time.

[0231] However, the operating conditions of the simulated ejection performed when the nozzle 71 is positioned at the preliminary ejection position and the actual ejection operation performed when the nozzle 71 is positioned at the coating position opposite to the upper surface Sf of the substrate S conveyed on the coating workbench 32 cannot be said to be completely the same. Therefore, the control parameters optimized for the ejection pressure curve in the simulated ejection as described above may not be optimal in the ejection operation to the substrate S.

[0232] In this sense, it is effective to perform an actual ejection operation to obtain an ejection pressure curve and adjust the control parameters. However, in general, in order to optimize the control parameters, it is necessary to obtain the ejection pressure curve multiple times. Therefore, if the post-processing of the coating liquid ejected onto the substrate is considered, it is not realistic to repeatedly perform ejection onto the substrate.

[0233] Therefore, in the parameter adjustment process of this embodiment, after the simulated ejection is performed and the control parameters are adjusted as described above, the ejection action of positioning the nozzle 71 at the coating position and ejecting the coating liquid is performed using the control parameters. Furthermore, if there is a significant difference between the ejection pressure curve obtained as a result and the ejection pressure curve obtained by the simulated ejection, the control parameters are readjusted in consideration of the difference, and finally the ejection pressure curve in the ejection action is brought close to the target curve.

[0234] Specifically, the nozzle 71 is positioned at the coating position (step S207), the pump 81 is operated to discharge using the adjusted control parameters (step S208), and the discharge pressure curve at this time is obtained (step S209). At this time, for example, a dummy substrate can be placed on the coating workbench 32.

[0235] The ejection pressure curve thus obtained is compared with the ejection pressure curve finally obtained by the previous simulated ejection. At this time, it is considered that there is a difference caused by the difference in the operating conditions of the two. Specifically, the difference between the two ejection pressure curves is calculated (step S210), and if the difference is less than a predetermined threshold value ("yes" in step S210), the setting value of the control parameter at this time is determined as the optimal value (step S212).

[0236] On the other hand, when the difference is greater than the threshold value ("No" in step S210), an offset corresponding to the difference is set (step S211), and the offset is stored in the storage unit 93. Then, the process returns to step S202 and the process is repeated. In the discharge pressure evaluation process executed at this time, the offset set in step S210 is added to the discharge pressure curve obtained by performing simulated discharge (step S104), and an evaluation based on the added curve is performed (step S105).

[0237] The difference in operating conditions between the simulated ejection and the ejection action does not affect the ejection pressure curve obtained by the simulated ejection. However, the curve after adding the offset obviously reflects the difference. Therefore, by optimizing the control parameters (steps S202 to S205) by treating the curve after adding the offset as if it were an actual measurement result, the control parameters can be set to eliminate the difference between the simulated ejection and the ejection action. That is, the control parameters set in this way can achieve the operating conditions that can make the ejection pressure curve in the ejection action actually performed by positioning the nozzle 71 at the coating position close to the target curve.

[0238] In the optimization of the control parameters (steps S202 to S205) executed after the offset amount is set (step S211), if the score evaluating the discharge pressure curve obtained during the simulated discharge is lower than the threshold, the processing end condition in step S206 is satisfied.

[0239] Since the discharge pressure curve is evaluated after adding the offset, the pump control parameters at this time do not necessarily meet the conditions for obtaining an ideal discharge pressure curve in the simulated discharge. However, since the offset is set in consideration of the difference between the simulated discharge and the discharge operation, the pump control parameters set based on this can make the discharge pressure curve in the actual discharge operation close to the target curve.

[0240] Therefore, by setting the setting value of the control parameter at this time to the final optimal value (step S212), the ejection characteristics when applying the coating liquid to the substrate S can be as targeted. That is, the original purpose, namely, the optimization of the pump control parameters applied during the ejection action, is achieved. In the series of processes so far, the ejection action of the nozzle 71 ejecting the coating liquid at the coating position is only once. However, the ejection action can also be performed again using the optimized control parameters to re-evaluate the control parameters.

[0241] Fig. 22 : is a diagram showing an example of an offset. In the figure, the dotted line represents an example of a discharge pressure curve obtained by simulating a discharge using the optimized control parameters, and the solid line represents an example of a discharge pressure curve obtained in a discharge action using the same control parameters. Then, the offset obtained from them is represented by a dotted line. As the offset in this case, the difference between the two can be used directly. In addition, in order to eliminate the influence of noise, appropriate filtering processing can also be applied. The filtering processing can be applied to each discharge pressure curve before the difference is obtained, and can also be applied to the obtained difference.

[0242] As described above, in the above embodiment, the discharge pressure curve obtained by performing simulated discharge is evaluated, and when the pump control parameters are optimized based on the result, an offset corresponding to the difference between the discharge pressure curve obtained in the simulated discharge and the discharge pressure curve obtained when the discharge operation is performed by applying the same pump control parameters is added to the actually measured discharge pressure curve. Therefore, in the evaluation of the discharge pressure curve and the optimization process of the pump control parameters based on the evaluation, the difference in the operating conditions during the simulated discharge and during the discharge operation is reflected.

[0243] Therefore, the pump control parameters during the ejection operation are optimized based on the results of the simulated ejection. By using the optimized pump control parameters to perform the ejection operation, in this embodiment, the ejection pressure curve when the coating liquid is actually ejected onto the substrate S can be made to match the target curve. Thus, the substrate S can be coated well with the desired ejection characteristics.

[0244] Here, if Figure 7 As shown, the characteristic amount Fv1 represents the difference between the approximate waveform WF1 (first approximate waveform) that approximates the temporal change of the overall ejection pressure during the ejection period Tt and the temporal change of the overall ejection pressure during the ejection period Tt. In this configuration, the overall ejection pressure during the ejection period Tt can be appropriately evaluated based on the approximate waveform WF1 that represents the temporal change of the overall ejection pressure during the ejection period Tt from the start of ejection of the coating liquid from the nozzle 71 to the end of ejection.

[0245] In particular, the approximate waveform WF1 has:

[0246] The rising regression line Lr_R (rising approximate line) approximates the change in the ejection pressure that increases with time after the ejection of the coating liquid from the nozzle 71 begins, so that the pressure increases linearly with time from the initial pressure Pi (the ejection start pressure) to the stable pressure Pm that is greater than the initial pressure Pi;

[0247] The initial approximate straight line Lr_s is set between the time point (time ta) when the coating liquid starts to be ejected from the nozzle 71 and the rising regression line Lr_R, and represents the initial pressure Pi;

[0248] The descending regression line Lr_F (descending approximate line) is a linear approximation of the change in the discharge pressure that decreases with the passage of time before the discharge of the coating liquid from the nozzle 71 ends, so that the pressure decreases linearly with the passage of time from the stable pressure Pm to the initial pressure Pi (discharge end pressure) that is smaller than the stable pressure Pm.

[0249] The end approximate straight line Lr_e is provided between the descending regression straight line Lr_F and the time point (time te) at which the ejection of the coating liquid from the nozzle 71 ends, and represents the initial pressure Pi (the ejection end pressure); and

[0250] The stable straight line Lr_m connects the rising regression line Lr_R and the falling regression line Lr_F to represent the stable pressure Pm.

[0251] In this configuration, the temporal change in the discharge pressure during the entire discharge period Tt from the start to the end of discharge of the coating liquid from the nozzle 71 is approximated by a trapezoidal waveform, and the discharge pressure during the entire discharge period Tt can be appropriately evaluated.

[0252] In addition, in the second evaluation stage, characteristic quantities Fv2 to Fv10 (second characteristic quantities) of the change in the ejection pressure over time in a period (second period) shorter than the ejection period Tt in the ejection period Tt are extracted, and the change in the ejection pressure over time is evaluated based on the characteristic quantities Fv2 to Fv10 (step S105). In this configuration, the ejection pressure can be evaluated with high accuracy based on the change in the ejection pressure over time in a period shorter than the ejection period Tt from the start to the end of ejection of the coating liquid from the nozzle 71.

[0253] In addition, Figure 8 In the evaluation items shown, the discharge pressure of the predetermined rising initial period Ta_s (second period) from the start of the discharge of the coating liquid from the nozzle 71 is evaluated. During the rising initial period Ta_s, the discharge pressure increases with the passage of time, and the characteristic quantity Fv2 (second characteristic quantity) of the difference between the rising regression curve Nr representing the change of the discharge pressure with time in the rising initial period Ta_s and the change of the discharge pressure with time in the rising initial period Ta_s is extracted. In this structure, the discharge pressure can be evaluated by considering the change of the discharge pressure with time after the coating liquid is just discharged from the nozzle 71.

[0254] In addition, the discharge pressure is evaluated during the rising period Ta (second period) from the start of discharge of the coating liquid from the nozzle 71 until the discharge pressure increases to the target pressure Pt (predetermined pressure). In this configuration, the discharge pressure can be evaluated by considering the temporal change of the discharge pressure during the rising period Ta.

[0255] Specifically, in Fig. 9 Among the evaluation items shown, the length of the rising period Ta is extracted as the feature amount Fv2 (second feature amount). In this configuration, the discharge pressure can be evaluated in consideration of the rising speed of the discharge pressure.

[0256] In addition, Fig. 10A and Fig. 10BIn the evaluation items shown, the number of times the first differential D1 of the ejection pressure change with time crosses a predetermined threshold value Th4 during the rising period Ta is extracted as a feature value Fv4 (second feature value). In this configuration, the ejection pressure can be evaluated by considering the smoothness of the ejection pressure change with time during the rising period.

[0257] In addition, Fig.11A and Fig. 11B Among the evaluation items shown, the number of times the absolute value of the second differential D2 of the change in the ejection pressure over time crosses the predetermined threshold Th5 during the rising period Ta is extracted as the feature value Fv5 (second feature value). In this configuration, the ejection pressure can be evaluated by considering the smoothness of the change in the ejection pressure over time during the rising period Ta.

[0258] In addition, Fig.12 In the evaluation items shown, in the rising period Ta, the ratio of the time T_1st when the second differential D2 of the change of the ejection pressure with time is greater than the predetermined positive threshold value (Th5) and the time T_2nd when the second differential D2 of the change of the ejection pressure with time is less than the negative threshold value (-Th5) having the same absolute value as the positive threshold value is extracted as the feature value Fv6 (second feature value). In this configuration, the ejection pressure can be evaluated by considering the difference in the change of the ejection pressure with time at the beginning and the end of the rising period Ta.

[0259] In addition, Fig.13 In the evaluation items shown, the discharge pressure in the rising end period Ta_e (second period) in which the discharge pressure increases to the target pressure Pt (prescribed pressure) is evaluated. That is, the feature quantity Fv7 (second feature quantity) representing the difference between the approximate waveform WF7 (rising end approximate waveform) which approximates the change of the discharge pressure with time in the rising end period Ta_e and the change of the discharge pressure with time in the rising end period Ta_e is extracted. The approximate waveform WF7 is composed of the rising end period regression straight line Lr (rising end approximate straight line) and the extended straight line Lm. The rising end regression straight line Lr overlaps with the approximate curve obtained by linearly approximating the change of the discharge pressure with time that increases with time in the pressure range (P7_l to P7_u) less than the target pressure Pt, and linearly increases to the stable pressure Pm with the passage of time. The extended straight line Lm extends from the rising end regression straight line Lr (end time point) to the end time point of the rising end period Ta_e, and represents the stable pressure Pm. In this configuration, the discharge pressure can be evaluated in consideration of the degree of stall of the discharge pressure at the end of the rising period Ta.

[0260] In addition, Fig.14Among the evaluation items shown, the discharge pressure in the initial vibration period Tb_s (second period) from the moment (time t81) when the discharge pressure reaches the maximum value Pmax to the moment (time t82) when the second differential D2 of the change in discharge pressure with time crosses zero twice is evaluated. Specifically, the minimum value P8min of the discharge pressure in the initial vibration period Tb_s is obtained, and the difference between the smaller pressure of the minimum value P8min and the stable pressure Pm and the maximum value Pmax of the discharge pressure is extracted as the feature quantity Fv8 (second feature quantity). In this structure, the discharge pressure can be evaluated by taking into account the overshoot of the discharge pressure.

[0261] In addition, Fig.15 In the evaluation items shown, the discharge pressure in the transition period Tb (second period) specified from the time point (time tb) when the discharge pressure exceeds the target pressure Pt (specified pressure) is evaluated. Specifically, the feature Fv9 (second feature) representing the difference between the discharge pressure in the rising period Ta and the stable pressure Pm is extracted. In this structure, the discharge pressure can be evaluated by considering the stability of the discharge pressure after the discharge pressure reaches the target pressure Pt.

[0262] In addition, Fig.16 Among the evaluation items shown, the discharge pressure in the constant pressure period Tbc from the time (time tb) when the discharge pressure exceeds the target pressure Pt (predetermined pressure) to the time (time td) when the discharge pressure starts to decrease toward the end of the discharge of the coating liquid from the nozzle 71 is evaluated. Specifically, the feature amount Fv10 representing the difference between the maximum value Pmax and the minimum value P10min of the discharge pressure in the constant pressure period Tbc is extracted. In this configuration, the discharge pressure can be evaluated in consideration of the stability of the discharge pressure in the constant pressure period Tbc of the discharge pressure.

[0263] In addition, in the third evaluation stage, characteristic quantities Fv11 to Fv14 (third characteristic quantities) of the change in the ejection pressure over time in a period (third period) shorter than the ejection period Tt in the ejection period Tt are extracted, and the change in the ejection pressure over time is evaluated based on the characteristic quantities Fv11 to Fv14 (step S105). In this configuration, the ejection pressure can be evaluated with high accuracy based on the change in the ejection pressure over time in a period shorter than the ejection period Tt from the start of ejection of the coating liquid from the nozzle 71 to the end of ejection.

[0264] In addition, Fig.17In the evaluation items shown, the discharge pressure is evaluated during the pressure rise period Tar (third period) during which the discharge pressure increases from the lower reference pressure P11_1 (lower reference value) to the upper reference pressure P11_u (upper reference value) over time after the coating liquid starts to be discharged from the nozzle 71. Specifically, a measurement value is obtained, which is one of the discharge pressure measurement data 99 between the lower reference pressure P11_1 and the upper reference pressure P11_u, and the sum of the root mean square error between the approximate straight line Lr_1 obtained by linear regression with respect to the discharge pressure change over time in the interval (t111 to t112) between the lower reference pressure P11_1 and one of the measurement data and the discharge pressure change over time, and the root mean square error between the approximate straight line Lr_2 obtained by linear regression with respect to the discharge pressure change over time in the interval (t112 to t113) between one of the measurement data and the upper reference pressure P11_u and the discharge pressure change over time is minimized. Then, the ratio of the slope K1 of the straight line between the lower reference pressure P11_1 and one measurement data to the slope K2 of the straight line between one measurement data and the upper reference pressure P11_u is extracted as the feature value Fv11 (third feature value). In this configuration, the discharge pressure can be evaluated in consideration of the linearity of the increase in discharge pressure.

[0265] In addition, Fig.18 In the evaluation items shown, the discharge pressure in the rising end period Ta_e (third period) in which the discharge pressure increases to the target pressure Pt (prescribed pressure) is evaluated. That is, the feature quantity Fv12 (third feature quantity) representing the difference between the approximate waveform WF12 (rising end approximate waveform) which approximates the change of the discharge pressure in the rising end period Ta_e with time and the change of the discharge pressure in the rising end period Ta_e with time is extracted. Here, the rising end approximate waveform WF12 is composed of the rising end regression straight line Lr and the extended straight line Lm. The rising end regression straight line Lr overlaps with the approximate curve obtained by linearly approximating the change of the discharge pressure with time that increases with time in the pressure range (P12_l to P12_u) smaller than the target pressure Pt, and linearly increases to the target pressure Pt with time. The extended straight line Lm is extended from the rising end regression straight line Lr (end time point) to the end time point of the rising end period Ta_e, and represents the target pressure Pt. In this configuration, the discharge pressure can be evaluated in consideration of the degree of stall of the discharge pressure at the end of the rising period Ta.

[0266] In addition, Fig.19In the evaluation items shown, the discharge pressure in the rising end period Ta_e (third period) in which the discharge pressure increases to the target pressure Pt (prescribed pressure) is evaluated. That is, the feature quantity Fv13 (third feature quantity) representing the difference between the approximate waveform WF13 (rising end approximate waveform) which approximates the change of the discharge pressure in the rising end period Ta_e with time and the change of the discharge pressure in the rising end period Ta_e with time is extracted. Here, the rising end approximate waveform WF13 is composed of the rising end regression straight line Lr and the extended straight line Lm. The rising end regression straight line Lr overlaps with the approximate curve obtained by linearly approximating the change of the discharge pressure with time that increases with time in the pressure range (P13_l to P13_u) smaller than the target pressure Pt, and linearly increases to the target pressure Pt with time. The extended straight line Lm is extended from the rising end regression straight line Lr (end time point) to the end time point of the rising end period Ta_e, and represents the target pressure Pt. In this configuration, the discharge pressure can be evaluated in consideration of the degree of stall of the discharge pressure in the final stage of the rising period Ta.

[0267] In addition, Fig. 20 In the evaluation items shown, the discharge pressure in the initial vibration period Tb_s (third period) from the time point (time t141) when the discharge pressure reaches the maximum value Pmax to the time point (time t142) when the second differential D2 of the change in discharge pressure with time crosses zero twice is evaluated. Specifically, the sum of the value obtained by subtracting the stable pressure Pm from the maximum value Pmax of the discharge pressure and the value obtained by subtracting the minimum value P14min of the discharge pressure in the initial vibration period Tb_s from the stable pressure Pm is extracted as the feature quantity Fv14 (third feature quantity). In this structure, the discharge pressure can be evaluated by taking into account the overshoot of the discharge pressure.

[0268] As described above, the coating device 1 of this embodiment is an embodiment of the "coating device" in the present invention, and can also be said to be an embodiment of the "dispensing device" of the present invention. In a narrow sense, the nozzle 71, the pump 81, the pressure gauge 86 and the control unit 9 function as the "dispensing device" of the present invention. Among them, the pressure gauge 86 is equivalent to the "detection unit" of the present invention, and the control unit 9 is equivalent to the "control unit" of the present invention.

[0269] In the above embodiment, the substrate conveying unit 5 corresponds to the "relative movement mechanism" of the present invention, and the roller 721 corresponds to the "roller member" of the present invention. The discharge pressure measurement data 99 corresponds to the "discharge pressure curve" of the present invention.

[0270] In the above embodiment, steps S102 to S105 are equivalent to the "adjustment process" of the present invention, and steps S102 to S105 and S207 to S210 are equivalent to the "re-adjustment process" of the present invention as a whole. In addition, steps S103 and S209 are equivalent to the "process of obtaining the discharge pressure curve" of the present invention, and step S211 is equivalent to the "process of obtaining the offset amount" of the present invention.

[0271] In addition, the present invention is not limited to the above-mentioned embodiment, and various changes other than the above-mentioned embodiment can be made as long as they do not deviate from the main purpose of the present invention. For example, in the above-mentioned embodiment, the ejection characteristics are measured based on the pressure value detected by the pressure gauge 86 installed on the pipe 84, but the installation position of the pressure gauge 86 is not limited to this, and the installation position is arbitrary as long as it can detect the pressure of the coating liquid delivered to the nozzle 71.

[0272] In addition, in the above-mentioned embodiment, the Bayesian optimization method is applied as a machine learning algorithm for searching for the optimal value of the pump control parameter. This method is applicable to the parameter search of the present invention, but the parameter search method is not limited thereto. That is, it is possible to use various algorithms that estimate the control parameter whose score becomes smaller based on the information storing the relationship between the discretely set control parameter and the score obtained under the condition, so as to achieve the optimization of the control parameter.

[0273] In the above embodiment, the bellows type pump 81 is used, but the type of pump is not limited thereto, and for example, a piston cylinder type pump may be used (for example, Japanese Patent Application Laid-Open No. 2008-101510).

[0274] In addition, in the above-mentioned embodiment, the present invention is applicable to a coating device 1 that supplies a coating liquid to an upper surface Sf of a substrate S while the substrate S is floated, but the applicable objects of the present invention are not limited to this. The present invention can be applied to all substrate processing technologies that supply a processing liquid from a nozzle to an upper surface of a substrate and perform a prescribed processing by conveying a processing liquid to a nozzle.

[0275] In addition, in the calculation Figure 7 When the approximate waveform WF1 is used, a straight line having a slope of zero indicating the target pressure Pt may be used instead of the stable straight line Lr_m.

[0276] Furthermore, it can also replace Figure 7 The characteristic amount Fv1 shown is used to calculate a characteristic amount Fv1_1 described in the following modification, and the temporal change in the discharge pressure is evaluated based on the characteristic amount Fv1_1. Fig.23 is a diagram for explaining each period used in a modified example of the evaluation item of the discharge pressure. Fig.24This is a diagram for explaining a modified example of the evaluation item for evaluating the change in the ejection pressure over time based on the feature value Fv1_1. Fig.23 With the above Figure 5B The common points are marked with corresponding reference numerals and the description is omitted as appropriate. Fig.24 With the above Figure 7 For the differences, the common points are marked with corresponding figure symbols and the description is appropriately omitted.

[0277] like Fig.23 As shown, in a modified example of the evaluation item, a focus period Troi is used. The focus period Troi is a period from time ta to time td. That is, the focus period Troi is composed of a rising period Ta, a transition period Tb, and a stable period Tc, in other words, a rising period Ta and a constant voltage period Tbc.

[0278] like Fig.24 As shown, in this modification, similarly to the case of the above-mentioned feature quantity Fv1, the rising regression line Lr_R is calculated, and the initial approximate line Lr_s is set. Furthermore, during the period from time t12 to time td, a straight line with a slope of zero, that is, a stable straight line Lr_m_1, is set which represents the stable pressure Pm (i.e., the average of the measured values ​​of the discharge pressure during the stable period Tc).

[0279] In this way, the approximate waveform WF1_1 composed of the start approximate straight line Lr_s, the rising regression straight line Lr_R and the stable straight line Lr_m_1 arranged in time series is calculated. In addition, the pressure evaluation unit 913 calculates the mean absolute error MAE between the ejection pressure measurement data 99 and the approximate waveform WF1_1 as the feature value Fv1_1 in the entire period of interest Troi from time ta to time td. In addition, the pressure evaluation unit 913 standardizes the feature value Fv1_1 to a prescribed range based on a prescribed threshold value Th1_1 (for example, 0.05). Specifically, the feature value Fv1_1 is converted into the standardized feature value Fv1_1 (i.e., the evaluation value V1_1) based on the following formula:

[0280] If Fv1_1<Th1_1, then Fv1_1=0,

[0281] If Fv1_1≥Th1_1, then Fv1_1=(Fv1_1+1-Th1_1)×c1_1.

[0282] Here, the upper limit of Fv1_1 is 2×c1_1, and the coefficient c1_1 is a normalization coefficient and is an arbitrary positive constant. In addition, the specific method of normalizing the feature value Fv1_1 is not limited to the example shown here, and can be changed as appropriate.

[0283] Based on Fig.24 In the evaluation of the feature amount Fv1_1, when the temporal change of the ejection pressure in the entire attention period Troi deviates greatly from the ideal shape, a large score (ie, a poor evaluation) can be given to the ejection pressure.

[0284] In this modification, Figure 4 In the evaluation of the measurement result of the ejection pressure shown in FIG. 1 (step S105), the pressure evaluation unit 913 calculates the evaluation value V1_1 based on the result of extracting the feature quantity Fv1_1 from the ejection pressure instead of the feature quantity Fv1. In particular, for the first (I=1) evaluation stage, instead of allocating the feature quantity Fv1_1 based on the ejection pressure, the pressure evaluation unit 913 calculates the evaluation value V1_1 based on the result of extracting the feature quantity Fv1_1 from the ejection pressure instead of the feature quantity Fv1. Figure 7 The evaluation of the feature quantity Fv1 shown is based on the distribution Fig.24 On this basis, the evaluation of the feature quantity Fv1_1 is performed Fig.21 The results of the measurements shown are evaluated. Fig.21 In the table, the feature quantity of the evaluation item when the evaluation stage I is 1 is not the feature quantity Fv1, but Fig.24 The feature quantity Fv1_1 shown.

[0285] In addition, the discharge pressure is measured in the period of interest Troi (main period, evaluation object period) from the start of discharge of the treatment liquid from the nozzle 71, through the discharge pressure rising to the target pressure Pt (prescribed pressure), to the discharge pressure starting to decrease from the target pressure Pt. And, in the first (I=1) evaluation stage, the characteristic quantity Fv1_1 (overall characteristic quantity) of the change of the overall discharge pressure of the period of interest Troi with time is extracted, and the change of the discharge pressure with time is evaluated based on the characteristic quantity Fv1_1. In this way, the appropriateness of the overall discharge pressure of the period of interest Troi that affects the thickness of the treatment liquid applied to the substrate S can be reflected in the evaluation of the discharge pressure.

[0286] Furthermore, in this variation, when the influence of the period of concern Troi on the thickness of the processing liquid applied to the substrate S is particularly large (in other words, when the influence on the period after the period of concern Troi is small), the appropriateness of the overall ejection pressure of the period of concern Troi can be reflected in the evaluation of the ejection pressure.

[0287] In addition, the characteristic quantity Fv1_1 (main characteristic quantity) represents the difference between the approximate waveform WF1_1 (main approximate waveform) which approximates the temporal change of the overall ejection pressure of the period of interest Troi (main period) and the temporal change of the overall ejection pressure of the period of interest Troi. In this configuration, the overall ejection pressure of the period of interest Troi can be appropriately evaluated based on the approximate waveform WF1_1 for the temporal change of the overall ejection pressure of the period of interest Troi.

[0288] In particular, the approximate waveform WF1_1 has:

[0289] The rising regression line Lr_R (rising approximate line) is obtained by linearly approximating the change in the ejection pressure that increases with time after the coating liquid starts to be ejected from the nozzle 71, and increases linearly with time from the initial pressure Pi (ejection start pressure) to a stable pressure Pm that is greater than the initial pressure Pi;

[0290] The initial approximate straight line Lr_s is provided between the start time point (time ta) when the coating liquid starts to be ejected from the nozzle 71 and the ascending regression straight line Lr_R, and represents the initial pressure Pi; and

[0291] The stable straight line Lr_m_1 is provided between the time when the rising regression straight line Lr_R reaches the stable pressure Pm and the end of the attention period Troi (between the time t12 and the time td), and represents the stable pressure Pm.

[0292] In this configuration, the temporal change in the discharge pressure in the entire period of interest Troi is approximated, and the discharge pressure in the entire period of interest Troi can be appropriately evaluated.

[0293] By the way, when using Fig.24 When evaluating the discharge pressure based on the characteristic amount Fv1_1 shown, it is not necessary to measure the discharge pressure in the period after the attention period Troi (that is, the falling period Td) in step S101.

[0294] As described above with reference to the specific embodiments, in the method for adjusting the pump control parameters of the present invention, for example, the Bayesian optimization method can be used as a machine learning method. When the Bayesian optimization method is applied in the present invention, the setting value of the pump control parameter for obtaining a discharge pressure curve closer to the target curve can be efficiently estimated based on the correspondence between the pump control parameters set in the past and the discharge pressure curve obtained therefrom. Therefore, the number of trials for actually performing simulated discharge can be reduced, the pump control parameters can be optimized in a short time, and the amount of consumed treatment fluid can be reduced to reduce the environmental load.

[0295] In the adjustment process, the pump control parameters can be adjusted so that the score obtained by quantifying the degree of deviation between the acquired discharge pressure curve and the target curve becomes smaller. Thus, the pump control parameters can be optimized using various machine learning algorithms that search for parameters with smaller scores.

[0296] In this case, in the adjustment process and the readjustment process, the estimation of the pump control parameters based on the discharge pressure curve and the score and the acquisition of the discharge pressure curve using the estimated pump control parameters may be repeatedly performed. In this way, by repeatedly estimating the pump control parameters and measuring the parameters, the pump control parameters can be made closer to the optimal value. In this case, by efficiently estimating the pump control parameters, the number of executions of the action for acquiring the discharge pressure curve can be reduced.

[0297] In addition, for example, in the readjustment process, the pump control parameters can be readjusted so that the score obtained by quantifying the degree of deviation between the acquired discharge pressure curve and the target curve becomes smaller. By adding the offset to the acquired discharge pressure curve, it is possible to perform an evaluation taking into account the change factors caused by the operating conditions in the discharge action, and by re-adjusting the pump control parameters based on this, it is possible to efficiently find the optimal value of the pump control parameter for obtaining the desired discharge characteristics in the discharge action.

[0298] In addition, in the method for adjusting the pump control parameters of the present invention, for example, the acquisition of the ejection pressure curve in the adjustment process and the readjustment process can also be performed in a state where the nozzle is positioned at a position different from the position of the nozzle in the ejection action. It is preferable to acquire the ejection pressure curve under the same operating conditions as the ejection action in order to find the optimal value of the pump control parameter under the operating conditions, but it is not necessarily realistic if the treatment of the treatment liquid ejected therefor is taken into consideration. In the present invention, even if there is a difference in operating conditions between the adjustment process or the readjustment process and the actual ejection action, the different pump control parameters can be optimized. Therefore, by performing the adjustment process and the readjustment process with the nozzle located at a position different from that during the ejection action, it is possible to facilitate the treatment of the treatment liquid and optimize the pump control parameters.

[0299] In addition, for example, when the difference between the discharge pressure curve obtained by using the pump control parameters adjusted in the adjustment process and the discharge pressure curve obtained in the discharge operation is less than a predetermined threshold value, it can be said that the difference in discharge results caused by the difference in each operation condition is small. In this case, the set value of the pump control parameter at this time can be determined as the optimal value without performing the readjustment process.

[0300] In addition, the present invention can be implemented as a computer program, wherein a computer that controls a spraying device that performs a spraying action of spraying a processing liquid delivered by a pump from a nozzle toward a substrate executes the following steps: an adjustment step of temporarily setting pump control parameters, delivering processing liquid from the pump, obtaining a spray pressure curve that represents the change in the spray pressure given to the delivered processing liquid over time, and adjusting the pump control parameters by machine learning using a regression model so that the obtained spray pressure curve approaches a preset target curve; a step of operating the pump based on the adjusted pump control parameters, performing a spraying action, and obtaining a spray pressure curve; a step of setting an offset corresponding to the difference between the spray pressure curve obtained in the adjustment step using the adjusted pump control parameters and the spray pressure curve obtained in the spraying action; and a readjustment step of temporarily setting the pump control parameters, delivering processing liquid from the pump, obtaining a spray pressure curve, and readjusting the pump control parameters by machine learning so that the curve obtained by adding the offset to the acquired spray pressure curve approaches the target curve.

[0301] Alternatively, it may be a computer program that causes a computer that controls a discharge device that performs a discharge operation that discharges a treatment liquid delivered by a pump from a nozzle toward a substrate to perform the following steps: a step of acquiring a discharge pressure curve that indicates a change in discharge pressure applied to the treatment liquid delivered by the pump over time; a step of adding a predetermined offset to the acquired discharge pressure curve and calculating a score that quantifies the degree of deviation between the added curve and a target curve; and a step of searching for a pump control parameter that minimizes the score by machine learning using a regression model. In this case, the offset can be set, for example, based on the difference between the discharge pressure curve acquired by performing the discharge operation and the discharge pressure curve acquired by positioning the nozzle at a position different from the position of the nozzle during the discharge operation.

[0302] Such a computer program enables the pump control parameter adjustment method of the present invention to be implemented by an existing device that does not have the assumed structure. For example, the present invention can be implemented as a computer-readable recording medium that non-temporarily records such a computer program.

[0303] Furthermore, by further providing a relative movement mechanism in the discharge device of the present invention, the relative movement mechanism moves the nozzle and the substrate relative to each other, thereby enabling the coating device to coat the processing liquid over a wide area on the substrate.

[0304] In this case, a liquid receiving unit is further provided to receive the processing liquid ejected from the nozzle positioned at the predetermined preliminary ejection position before the ejection operation, and the nozzle may also be positioned at the preliminary ejection position during the adjustment process. Furthermore, the liquid receiving unit may also include a roller member disposed directly below the nozzle positioned at the preliminary ejection position. According to such a structure, the processing liquid ejected preliminarily from the nozzle can be recovered without being attached to the substrate.

[0305] The present invention can be applied to all coating technologies in which a processing liquid is supplied to a nozzle and then discharged from the nozzle toward a substrate.

Claims

1. A method for adjusting a pump control parameter for controlling a discharge action of a treatment liquid delivered by a pump from a nozzle positioned at a coating position toward a substrate, wherein: include: an adjustment step of positioning the nozzle at a preliminary ejection position, temporarily setting the pump control parameters, delivering the treatment liquid from the pump, acquiring an ejection pressure curve representing the variation of the ejection pressure applied to the delivered treatment liquid over time, and adjusting the pump control parameters by machine learning using a regression model so that the acquired ejection pressure curve approaches a preset target curve; A step of positioning the nozzle at a coating position, operating the pump based on the pump control parameters adjusted in the adjustment step, performing the ejection action, and acquiring the ejection pressure curve; a step of setting an offset amount corresponding to a difference between the discharge pressure curves respectively acquired at the preliminary discharge position and the coating position using the pump control parameter adjusted in the adjustment step; and A readjustment process is performed to position the nozzle at the preparatory ejection position, temporarily set the pump control parameters, deliver the treatment liquid from the pump, obtain the ejection pressure curve, and readjust the pump control parameters through machine learning so that the curve obtained by adding the offset to the acquired ejection pressure curve is close to the target curve.

2. The method for adjusting the pump control parameters according to claim 1, in, As the machine learning method, the Bayesian optimization method is used.

3. The method for adjusting the pump control parameters according to claim 1 or 2, in, In the adjustment step, the pump control parameter is adjusted so that a score obtained by quantifying the degree of deviation between the acquired discharge pressure curve and the target curve becomes smaller.

4. The method for adjusting the pump control parameters according to claim 3, in, In the adjustment step and the readjustment step, estimation of the pump control parameter based on the discharge pressure curve and the score and acquisition of the discharge pressure curve using the estimated pump control parameter are repeatedly performed.

5. The method for adjusting the pump control parameters according to claim 3, in, In the readjustment step, the pump control parameter is readjusted so that a score obtained by quantifying the degree of deviation between the acquired discharge pressure curve and the target curve and adding the offset amount is reduced.

6. The method for adjusting pump control parameters according to claim 1 or 2, in, The preliminary ejection position and the coating position are different from each other.

7. The method for adjusting pump control parameters according to claim 1 or 2, in, When the difference between the ejection pressure curves respectively obtained at the preliminary ejection position and the coating position using the pump control parameter adjusted by the adjustment process is smaller than a prescribed threshold value, the readjustment process is not performed, and the set value of the pump control parameter at this time is determined as an optimal value.

8. A computer program, in, The following steps are executed by causing a computer that controls a discharge device that performs a discharge operation of discharging a treatment liquid delivered by a pump from a nozzle positioned at a coating position toward a substrate: an adjustment step of positioning the nozzle at a preliminary ejection position, temporarily setting a pump control parameter, delivering the treatment liquid from the pump, obtaining an ejection pressure curve representing a change in ejection pressure applied to the delivered treatment liquid over time, and adjusting the pump control parameter by machine learning using a regression model so that the acquired ejection pressure curve approaches a preset target curve; A step of positioning the nozzle at a coating position, operating the pump based on the pump control parameters adjusted in the adjustment step, performing the ejection action, and acquiring the ejection pressure curve; a step of setting an offset amount corresponding to a difference between the discharge pressure curves respectively acquired at the preliminary discharge position and the coating position using the pump control parameter adjusted in the adjustment step; and A readjustment process is performed to position the nozzle at the preparatory ejection position, temporarily set the pump control parameters, deliver the treatment liquid from the pump, obtain the ejection pressure curve, and readjust the pump control parameters through machine learning so that the curve obtained by adding the offset to the acquired ejection pressure curve is close to the target curve.

9. A computer program, in, The following steps are executed by causing a computer that controls a discharge device that performs a discharge operation of discharging a treatment liquid delivered by a pump from a nozzle positioned at a coating position toward a substrate: A step of acquiring a discharge pressure curve indicating a change over time in the discharge pressure applied to the treatment liquid delivered from the pump in a state where the nozzle is positioned at a preliminary discharge position; A step of adding a predetermined offset to the obtained ejection pressure curve and calculating a score by quantifying the degree of deviation between the added curve and a target curve, wherein the target curve is the ejection pressure curve for achieving an ideal ejection state for the nozzle positioned at the coating position; and By machine learning using a regression model, a process of the pump control parameter in which the score becomes the minimum is searched.

10. The computer program according to claim 9, in, The offset amount is set according to a difference between the ejection pressure curve acquired when the nozzle is positioned at the preliminary ejection position and the ejection pressure curve acquired when the nozzle is positioned at the coating position.

11. A computer readable recording medium, in, The computer program according to any one of claims 8 to 10 is recorded non-temporarily.

12. A spraying device for performing a spraying operation of spraying a processing liquid onto a substrate, in, have: a nozzle for spraying the treatment liquid; a pump, delivering the treatment liquid to the nozzle; a detection unit that detects a discharge pressure applied to the treatment liquid delivered by the pump; as well as a control unit that controls the pump based on a predetermined pump control parameter, The control unit obtains a discharge pressure curve indicating temporal changes in the discharge pressure applied to the sent processing liquid based on the detection result of the detection unit, and executes the pump control parameter adjustment method according to claim 1 or 2 to optimize the pump control parameter.

13. A coating device, in, have: A spraying portion having the same structure as the spraying device according to claim 12; and The relative movement mechanism enables the nozzle and the substrate to move relative to each other.

14. The coating device according to claim 13, in, The device includes a liquid receiving portion for receiving the processing liquid ejected from the nozzle positioned at the preliminary ejection position before the ejection operation.

15. The coating device according to claim 14, in, The liquid receiving portion includes a roller member disposed directly below the nozzle positioned at the preliminary ejection position.