Prediction device and etching control system

Abstract

This invention provides a prediction device and etching control system that can efficiently achieve a complex etching amount distribution in wet etching. [Solution] The prediction device performs processing on a model that represents the relationship between process parameters for controlling the operation of multiple nozzles in wet etching of a substrate and the distribution of etching amounts in the plane of the substrate. The prediction device includes an update unit that optimizes the model parameters using training data based on process parameters and the measured or referenced distribution of etching amounts in the plane of the substrate, a calculation unit that calculates process parameters corresponding to a specified distribution of etching amounts using the updated model, and a providing unit that provides the calculated process parameters to an external party.

Description

Technical Field

[0001] The present disclosure relates to a prediction device and an etching control system.

Background Art

[0002] Conventionally, in semiconductor processing, a developing device and a single wafer cleaning device that discharge a chemical solution such as a developing solution onto a rotating substrate are known (for example, see Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] The present disclosure provides a prediction device and an etching control system that can efficiently realize a complicated etching amount distribution in wet etching.

Means for Solving the Problems

[0005] An etching control system according to an embodiment is a prediction device that performs processing related to a model representing the relationship between a process parameter for controlling the operations of a plurality of nozzles and the distribution of the etching amount in the plane of a substrate in wet etching of the substrate, and includes an update unit that optimizes the parameters of the model with training data based on the process parameter and the measured or referenced distribution of the etching amount in the plane of the substrate, a calculation unit that calculates a process parameter corresponding to a specified etching amount distribution using the model with updated parameters, and a providing unit that provides the calculated process parameter to the outside.

[0006] The etching control system according to the embodiment is an etching control system for dual dispensing, comprising a prediction device and an etching control device, wherein the prediction device updates the parameters of a model using process parameters that define the discharge position, discharge time, and travel speed of a plurality of nozzles, including a first nozzle for discharging rinse onto a rotating substrate and a second nozzle for discharging a chemical solution for wet etching the substrate, and a model that represents the relationship between these parameters and the distribution of etching amount in the plane of the substrate, calculates the process parameters using the model with updated parameters, and provides the calculated process parameters, and the etching control device acquires the process parameters and controls the operation of the first nozzle and the second nozzle.

[0007] The etching control system according to this embodiment comprises a prediction device, a management server, and an etching control device that controls the operation of a plurality of nozzles in wet etching of a substrate, wherein the prediction device calculates process parameters that define the discharge position, discharge time, and travel speed of the plurality of nozzles, and transmits the process parameters to the management server or the Host IF of the etching control device, the management server accepts the input of the process parameters provided by the prediction device via a management application, and the etching control device obtains the process parameters via the Host IF or the management application and controls the plurality of nozzles based on the process parameters.

[0008] According to this disclosure, a complex distribution of etching amounts can be efficiently achieved in wet etching. [Brief explanation of the drawing]

[0009] [Figure 1] Figure 1 shows an example of the configuration of an etching control system. [Figure 2] Figure 2 illustrates the distribution of etching amounts. [Figure 3] Figure 3 is a diagram illustrating the dual dispensing process. [Figure 4] Figure 4 shows an example of the configuration of the prediction device. [Figure 5] Figure 5 shows an example of the configuration of an etching control device. [Figure 6] Figure 6 shows an example of an etching control system. [Figure 7] Figure 7 shows an example of an etching control system. [Figure 8] Figure 8 shows an example of an etching control system. [Figure 9] Figure 9 is a flowchart showing the processing flow of the etching control system. [Figure 10] Figure 10 is a flowchart showing the flow of the model update process and the process parameter prediction process. [Figure 11] Figure 11 is a flowchart showing the flow of the model update process and the process parameter prediction process. [Figure 12] Figure 12 shows an example of a computer running a program. [Modes for carrying out the invention]

[0010] In recent years, with the increasing complexity of semiconductor processing, the performance requirements for wet etching control in single-wafer washing equipment have become more diverse. For example, in addition to the conventional uniform etching profile, there are cases where control capabilities are required to correct and cancel out the amount of residual film generated in previous processes to achieve uniformity.

[0011] Another known technique for controlling the wet etching profile is the swing sequence. The swing sequence is a method in which a nozzle that dispenses the chemical solution moves back and forth in the radial direction of a rotating substrate.

[0012] However, conventional techniques have the problem of being unable to efficiently achieve complex etching distributions in wet etching.

[0013] Here, the etching amount is the depth of etching. Also, the distribution of the etching amount is the etching amount for each position in the radial direction of the substrate (radial position).

[0014] For example, in a conventional swing sequence, since the nozzle continues to move during the rotation of the substrate (wafer), the operations shown in the process recipe become complicated, and it is difficult to meet the need to increase or decrease the etching amount at an arbitrary radial position.

[0015] Note that the process recipe is information that defines the operations of one or more nozzles in a swing sequence.

[0016] Also, for example, in a conventional swing sequence, the method for optimizing the process recipe is not automated, and the optimization of the process recipe is performed based on the engineer's experience and many trials.

[0017] Therefore, a technology that can efficiently realize a complex distribution of etching amounts in wet etching is expected.

[0018] Hereinafter, embodiments of an etching control system and an etching control method will be described in detail based on the drawings. Note that the disclosed technology is not limited by the following embodiments.

[0019] In the embodiment, it is assumed that wet etching called a dual dispense process is performed. In the dual dispense process, two nozzles discharge liquid onto a circular substrate.

[0020] The first of the two nozzles discharges a rinse (e.g., water). Also, the second of the two nozzles discharges a chemical solution (e.g., an etching solution). The chemical solution corrodes the substrate. Also, the rinse dilutes the chemical solution and suppresses the degree of corrosion of the substrate.

[0021] One of the two nozzles dispenses the liquid towards the center of the substrate, while the other dispenses the liquid towards the outer periphery of the substrate. The etching control system moves the positions of the two nozzles according to the process recipe to create the intended etching distribution on the substrate.

[0022] In a dual-dispense process, controllability can be improved by fixing the position of one nozzle in the center of the substrate and moving the position of the other nozzle around the periphery. Furthermore, the dual-dispense process can be used in developing and single-fed washing, among other applications.

[0023] [Configuration of the Embodiment] The configuration of the etching control system will be explained using Figure 1. Figure 1 is a diagram showing an example of the configuration of the etching control system.

[0024] As shown in Figure 1, the etching control system 1 includes a prediction device 10 and an etching control device 20.

[0025] The prediction device 10 updates a model representing the relationship between the etching amount distribution and process parameters, and predicts process parameters using this model. Note that the etching amount may be replaced not only with the reduction in film thickness on a single-film wafer, but also with the etching amount on the device pattern or CD (Critical Dimension).

[0026] Process parameters are information that defines the operation of the two nozzles in a dual-dispensing process. For example, a process recipe is generated based on the process parameters.

[0027] As shown in Figure 2, the etching amount distribution represents the etching amount at each location on the substrate. Figure 2 is a diagram illustrating the etching amount distribution. The horizontal axis of Figure 2 represents the distance from the center of the substrate, i.e., the radial position. The vertical axis of Figure 2 represents the etching amount at each radial position.

[0028] The etching amount distribution may be data that actually shows the etching amount at a certain radius position (for example, every 1 mm), or it may be a parameter used to specify the shape of the curve.

[0029] The model can be any model capable of representing the relationship between the etching amount distribution and the process parameters, such as a regression model. Furthermore, the model is not limited to regression models; it may also be a neural network or similar.

[0030] The etching control device 20 controls the etching apparatus based on process parameters. Specifically, the etching control device 20 controls the operation of the two nozzles in the dual dispense process.

[0031] Here, we will explain the dual dispense process using Figure 3. Figure 3 is a diagram illustrating the dual dispense process.

[0032] As shown in Figure 3, the dual dispensing process in this embodiment includes steps S501, S502, S503, S504, S505, S506, S507, and S508. In the dual dispensing process, all of these steps may be performed, or some steps may be omitted. Process parameters define the operation of the nozzle (dispensing position, dispensing time, speed, etc.) in each step.

[0033] Wafer 61 is a circular (disk-shaped) substrate that is etched in a dual-dispense process. Nozzle 62 is a nozzle that dispenses rinse (e.g., water). Nozzle 63 is a nozzle that dispenses chemical solution (e.g., etching solution).

[0034] In step S501 (Type 3 outer), nozzle 62 dispenses rinse into the center of wafer 61, and nozzle 63 dispenses the chemical solution into the outer periphery of wafer 61. Furthermore, for step S501, the dispensing position and dispensing time of nozzle 63 are defined by process parameters.

[0035] In step S502 (Type 3 scan-in), nozzle 62 dispenses rinse into the center of wafer 61, and nozzle 63 dispenses the chemical solution while moving from the outer edge of wafer 61 to the center. In addition, for step S502, the movement speed of nozzle 63 is defined by process parameters.

[0036] In step S503 (Type 3 inner), nozzle 62 discharges rinse to the center of wafer 61, and nozzle 63 discharges the chemical solution to the outer edge of wafer 61 (however, closer to the center than in step S501 (Type 3 outer)). Furthermore, for step S503, the discharge position and discharge time of nozzle 63 are defined by process parameters.

[0037] In step S504 (Type 2 inner), nozzle 63 dispenses the chemical solution into the center of wafer 61, and nozzle 62 dispenses rinse onto the outer periphery of wafer 61. Furthermore, for step S504, the dispensing position and dispensing time of nozzle 62 are defined by process parameters.

[0038] In step S505 (Type 2 scanout), nozzle 63 dispenses the chemical solution into the center of wafer 61, and nozzle 62 dispenses rinse while moving from the center to the outer edge of wafer 61. Furthermore, for step S505, the movement speed of nozzle 62 is defined by process parameters.

[0039] In step S506 (Type 2 outer), nozzle 63 discharges the chemical solution to the center of wafer 61, and nozzle 62 discharges the rinse to the outer edge of wafer 61 (but further from the center than in step S504 (Type 2 inner)). In addition, for step S506, the discharge position and discharge time of nozzle 62 are defined by process parameters.

[0040] In step S507 (Type 1), the nozzle 63 dispenses the chemical solution into the center of the wafer 61. Furthermore, the dispensing time of the nozzle 63 in step S507 is defined by process parameters.

[0041] In step S508 (Type 1), the nozzle 62 dispenses rinse into the center of the wafer 61.

[0042] In the Type 3 recipe, the amount of etching increases in the area where the chemical solution is discharged by nozzle 63, and the longer the discharge time of nozzle 63 (or the slower the movement speed in scanning), the greater the etching compared to other areas. In the Type 2 recipe, the amount of etching decreases in the area where the rinse is discharged by nozzle 62, and the longer the discharge time of nozzle 62 (or the slower the movement speed in scanning), the greater the etching compared to other areas.

[0043] The start and end positions of movement in the scan (S502, S505) are derived from the discharge positions of the preceding and succeeding steps. Furthermore, the discharge time is derived from the movement speed, start position, and end position. Therefore, for the scan, only the movement speed needs to be defined by the process parameters.

[0044] By adjusting the process parameters of each step as explained in Figure 3, the etching control device 20 can achieve various etching shapes.

[0045] Returning to Figure 1, let's explain the processing flow of the etching control system 1. First, the prediction device 10 updates the model using training data (step S1).

[0046] The training data consists of the shape of the etching amount (calibration curve data) obtained by actually measuring the substrate, combined with the process parameters used at that time. Such a combination of data can be called training data in machine learning. The prediction device 10 can update its model using known learning methods.

[0047] Next, the prediction device 10 predicts the process parameters based on the updated model and target profile (step S2). The target profile is the distribution of a specified etching amount. The process parameters corresponding to the target profile are unknown.

[0048] For example, the target profile is an etching amount distribution that cancels out any residual film on the substrate created in the previous process.

[0049] The prediction device 10 inputs the target profile into the updated model and outputs the process parameters. In other words, in step S2, inference processing is performed using the updated model.

[0050] Next, the prediction device 10 outputs process parameters (step S3). Then, the etching control device 20 performs etching based on the process parameters (step S4).

[0051] Furthermore, the etching amount distribution for training data and target profiles may be measured using a spectrophotometer, scatterometry, SEM (Scanning Electron Microscope), etc.

[0052] The configuration of the prediction device 10 will be explained using Figure 4. Figure 4 is a block diagram showing an example of the configuration of the prediction device according to this embodiment.

[0053] As shown in Figure 4, the prediction device 10 includes an I / F (interface) unit 11, a storage unit 12, and a control unit 13.

[0054] The I / F unit 11 is an interface for exchanging data between other devices. For example, the I / F unit 11 is a NIC (Network Interface Card). Furthermore, the I / F unit 11 may be connected to input / output devices such as a mouse, keyboard, display, and speaker.

[0055] The memory unit 12 is implemented by, for example, semiconductor memory elements such as RAM (Random Access Memory) and flash memory, or storage devices such as hard disks and optical discs.

[0056] The memory unit 12 stores model information 121. Model information 121 is information for constructing a model. Model information 121 is updated by the prediction device 10. For example, model information 121 may be parameters such as regression coefficients in a regression model.

[0057] The control unit 13 is implemented, for example, by a CPU (Central Processing Unit), MPU (Micro Processing Unit), GPU (Graphics Processing Unit), etc., which executes a program stored in an internal memory device using RAM as the working area.

[0058] Furthermore, the control unit 13 may be implemented by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

[0059] The control unit 13 includes a calculation unit 131, an update unit 132, and a provision unit 133. Note that the internal configuration of the control unit 13 is not limited to the configuration described here; other configurations are also acceptable as long as they perform the information processing described later.

[0060] The calculation unit 131 inputs the etching amount distribution into the model constructed based on the model information 121 and calculates process parameters. The calculation unit 131 can input the etching amount distribution included in the training data, or the etching amount distribution specified as the target profile, into the model.

[0061] The update unit 132 updates the model parameters so that the model representing the relationship between process parameters, which are parameters for controlling the operation of multiple nozzles for etching the substrate, and the distribution of etching amount within the plane of the substrate is optimized.

[0062] Here, the update unit 132 updates the model parameters (model information 121) so that the difference between the etching amount distribution of the training data and the etching amount distribution predicted from the process parameters becomes smaller.

[0063] For example, the update unit 132 can update parameters using the least squares method if the model is a regression model, or use the backpropagation method if the model is a neural network.

[0064] The supply unit 133 provides process parameters calculated by the calculation unit 131. The supply unit 133 may output the process parameters to an output device such as a display and a printer, or it may transmit the process parameters to another device including the etching control device 20.

[0065] The configuration of the etching control device 20 will be explained using Figure 5. Figure 5 is a diagram showing an example of the configuration of the etching control device.

[0066] As shown in Figure 5, the etching control device 20 has an I / F unit 21, a storage unit 22, and a control unit 23.

[0067] The I / F unit 21 is an interface for exchanging data between other devices. For example, the I / F unit 21 is a NIC (Network Interface Card). Furthermore, the I / F unit 21 may be connected to input / output devices such as a mouse, keyboard, display, and speaker.

[0068] The memory unit 22 is implemented by, for example, semiconductor memory elements such as RAM and flash memory, or storage devices such as hard disks and optical discs.

[0069] The memory unit 22 stores the process recipe 221. The process recipe 221 is information that defines the operation of each nozzle in the dual dispensing process.

[0070] The control unit 23 is implemented, for example, by a CPU, MPU, GPU, etc., which executes a program stored in an internal memory device using RAM as the working area.

[0071] Furthermore, the control unit 23 may be implemented by an integrated circuit such as an ASIC or FPGA.

[0072] The control unit 23 includes an acquisition unit 231, an update unit 232, and an operation control unit 233. Note that the internal configuration of the control unit 13 is not limited to the configuration described here; other configurations are also acceptable as long as they perform the information processing described later.

[0073] The acquisition unit 231 acquires process parameters provided by the prediction device 10. The specific method for acquiring process parameters by the acquisition unit 231 will be described later.

[0074] The update unit 232 updates the process recipe 221 based on the process parameters acquired by the acquisition unit 231.

[0075] The motion control unit 233 controls the operation of multiple nozzles using process parameters acquired by the acquisition unit 231. The motion control unit 233 performs a dual dispense process by operating the nozzles according to the process recipe.

[0076] Here, using Figures 6, 7, and 8, we will describe several embodiments in which the acquisition method of process parameters by the acquisition unit 231 of the etching control device 20 differs. Figures 6, 7, and 8 are diagrams showing embodiments of the etching control system.

[0077] [Example 1] As shown in Figure 6, the etching control device 20 can acquire process parameters entered by the user. For example, the acquisition unit 231 acquires process parameters entered via an input device provided in the etching control device 20. This allows the etching control device 20 to acquire process parameters more reliably.

[0078] In this case, the etching control device 20 is equipped with a device monitor that has a screen providing a GUI (Graphical User Interface) and an input device such as a keyboard. The user inputs the process parameters output from the prediction device 10 via the input device of the device monitor (step S11).

[0079] Next, the etching control device 20 updates the process recipe based on the process parameters (step S12). Then, the etching control device 20 performs etching according to the updated process recipe 221 (step S13).

[0080] [Example 2] As shown in Figure 7, the etching control device 20 can obtain process parameters via the management server 30.

[0081] The management server 30 provides a management application. The management application has functions for referencing and rewriting process recipes 221 stored in the etching control device 20.

[0082] The acquisition unit 231 acquires process parameters entered via a management application provided by the management server 30. In this way, the etching control device 20 can easily acquire process parameters by utilizing an existing application.

[0083] The terminal device used by the user, such as a PC, corresponds to the prediction device 10. The prediction device 10 connects to the management server 30 via a network and receives a management application. The prediction device 10 runs the management application using a dedicated client application or a web browser. The management application provides a GUI for inputting process parameters.

[0084] The user inputs process parameters via a management application run by the prediction device 10 (step S21). The management server 30 then sends updated process recipe data based on the input process parameters to the etching control device 20 (step S22).

[0085] Next, the etching control device 20 updates the process recipe 221 using the received update data (step S23). Then, the etching control device 20 performs etching according to the updated process recipe 221 (step S24).

[0086] [Example 3] As shown in Figure 8, the etching control device 20 can obtain process parameters via the management server 30 through the Host IF. The supply unit 133 inputs the process parameters provided by the calculation unit 131 to the Host IF connecting the prediction device 10 and the etching control device 20. The acquisition unit 231 acquires the process parameters input to the Host IF. This automates the provision of process parameters.

[0087] The Host IF is an interface for the host controller located in the etching control device 20. The Host IF connects the prediction device 10 and the etching control device 20 via wired or wireless connection, for example, using Ethernet®, Wi-Fi®, LAN, etc.

[0088] The prediction device 10 transmits process parameters to the etching control device 20 via the Host IF (step S31).

[0089] The etching control device 20 receives process parameters as variable parameters (step S32). Then, the etching control device 20 performs etching according to the variable parameters (step S33).

[0090] Variable parameters are parameters used to modify a part of the process recipe 221. However, the process recipe 221 itself, which is stored in the memory unit 22, is not updated by the variable parameters.

[0091] In this embodiment, the etching control device 20 temporarily generates a process recipe modified by the variable parameters and performs etching according to the temporarily generated process recipe.

[0092] [Processing of the embodiment] Figure 9 illustrates the processing flow of the etching control system 1. Figure 9 is a flowchart showing the processing flow of the etching control system.

[0093] First, the prediction device 10 accepts training data as input (step S101). Next, the prediction device 10 updates the model using the training data (step S102).

[0094] Next, the prediction device 10 inputs the target profile into the updated model and predicts the process parameters (step S103). The etching control device 20 performs etching based on the process parameters (step S104).

[0095] The method for obtaining process parameters using the etching control device 20 is as described in the example.

[0096] Here, we will explain the flow of the model update process and process parameter prediction process by the prediction device 10 using Figures 10 and 11. Figures 10 and 11 are flowcharts showing the flow of the model update process and process parameter prediction process.

[0097] The model will output the following 11 process parameters. The model will then represent the relationship between the process parameters, which are the parameters for determining the discharge time, discharge position, and travel speed of the first nozzle that discharges rinse onto the rotating substrate and the second nozzle that discharges the chemical solution that etches the substrate, and the distribution of the etching amount. (1) Central discharge (Type 1) time: Related to the discharge time of nozzle 63 in step S507 (Type 1). (2) Type 2 outer circumference discharge position_1: Related to the discharge position of nozzle 62 in step S504 (Type 2 inner). (3) Type 2 outer discharge time_1: Related to the discharge time of both nozzles in step S504 (Type 2 inner). (4) Type 2 Scan speed: Related to the movement speed of the nozzle 62 in step S505 (Type 2 scan out). (5) Type 2 outer circumference discharge position_2: Related to the discharge position of nozzle 62 in step S506 (Type 2 outer). (6) Type 2 outer discharge time_2: Related to the discharge time of both nozzles in step S506 (Type 2 outer). (7) Type 3 outer circumference discharge position_1: Related to the discharge position of nozzle 63 in step S501 (Type 3 outer). (8) Type 3 outer discharge time_1: Related to the discharge time of both nozzles in step S501 (Type 3 outer). (9) Type 3 Scan speed: Related to the movement speed of the nozzle 63 in step S502 (Type 3 scan in). (10) Type 3 outer circumference discharge position_2: Related to the discharge position of nozzle 63 in step S503 (Type 3 inside). (11) Type 3 outer discharge time_2: Related to the discharge time of both nozzles in step S503 (Type 3 inner).

[0098] Process parameters are given names such as discharge position, discharge time, and speed, but the values ​​of these process parameters do not necessarily directly determine the nozzle's discharge position, discharge time, and speed.

[0099] For example, the process parameter "(2) Type2 outer periphery discharge position_1" is related to the discharge position of the nozzle 62 in step S506 (Type2 outer) in Figure 3, but does not necessarily uniquely determine the discharge position of the nozzle 62.

[0100] For example, the etching control device 20 processes the values ​​of each process parameter as appropriate, determines the units, and generates a process recipe. In this case, the etching control device 20 controls the operation of the nozzle according to the process recipe.

[0101] Figure 10 is a flowchart showing the model update process. First, the prediction device 10 acquires the central etching amount as training data (step S201). Specifically, the prediction device 10 acquires a combination of the central etching amount and the process parameter (actual value) from (1).

[0102] Here, by referring to various etching shapes prepared in advance, the prediction device 10 can update the model using only the etching amount in the central part and the process parameters of (1). If the model is updated using a pre-prepared etching shape, that is, if the shape is not fine-tuned (step S202, No), the prediction device 10 proceeds to step S204.

[0103] On the other hand, if the model is not updated using a pre-prepared etching shape, that is, if the shape is to be fine-tuned (step S202, No), the prediction device 10 proceeds to step S203.

[0104] In shape fitting (step S203), the prediction device 10 updates the model based on the combination of process parameters (actual values) (2), (5), (7), and (10) and the etching amount.

[0105] Furthermore, the prediction device 10 performs outer perimeter discharge position optimization (step S204). Here, the prediction device 10 uses the current model to calculate the process parameters (predicted values) (1) to (11) from the target profile.

[0106] The prediction device 10 then updates the provisional parameter set so that the error between the predicted value calculated in step S204 and the process parameters included in the training data is reduced. The provisional parameter set is a set of provisional parameters determined by the target profile and the process parameters (1) to (11), and is used in steps S206 and S207.

[0107] The prediction device 10 performs discharge time optimization (step S206). Here, the prediction device 10 adjusts the process parameters (2), (4), (5), (7), (9), and (10).

[0108] Furthermore, the prediction device 10 calculates the error between the etching amount distribution derived from the process parameters predicted up to step S206 and the target profile as a model residual (step S207).

[0109] At this point, if the termination condition is met (step S208, Yes), the prediction device 10 outputs the process parameters predicted up to step S206 as the final process parameters and terminates the process. For example, termination conditions include the model residual becoming sufficiently small, or the iteration being performed a certain number of times.

[0110] If the termination condition is not met (step S208, No), the prediction device 10 returns to step S206 and repeats the process.

[0111] In steps S201, S203, and S204, process parameters mainly related to the ejection position are adjusted, and the shape of the etching amount distribution is determined. Furthermore, in step S206, process parameters mainly related to the ejection time are adjusted, and the shape of the etching amount distribution expands and contracts, approaching the target profile.

[0112] The prediction device 10 may perform the update process according to the flow shown in Figure 11. Figure 11 is a flowchart showing the flow of the model update process. In the example in Figure 11, the process branches depending on whether or not fine-tuning of the shape is performed in the first step.

[0113] If the model is updated using a pre-prepared etching shape, that is, if no fine adjustments are made to the shape (step S301, No), the prediction device 10 proceeds to step S304.

[0114] In step S304, during the optimization of the outer edge discharge position, the prediction device 10 uses a model to calculate the process parameters (predicted values) (1) to (11) from the target profile. The prediction device 10 also obtains the central etching amount (step S305).

[0115] On the other hand, if the model is not updated using a pre-prepared etching shape, that is, if the shape is to be fine-tuned (step S301, Yes), the prediction device 10 proceeds to step S302.

[0116] Then, the prediction device 10 performs central etching amount selection and shape fitting (step S302), similar to steps S201 and S203 in Figure 10, and proceeds to step S303.

[0117] Then, in the outer perimeter discharge position optimization step S303, the prediction device 10 uses the model to calculate the process parameters (predicted values) (2) to (11) from the target profile. The prediction device 10 has already obtained the process parameter (1) in step S302.

[0118] Then, the prediction device 10 updates the provisional parameter set (step S306) so that the error between the predicted value calculated in step S304 or S305 and the process parameters included in the training data is reduced. The provisional parameter set is a set of provisional parameters determined by the target profile and the process parameters (1) to (11), and is used in steps S307 and S308.

[0119] The prediction device 10 performs discharge time optimization (step S307). Here, the prediction device 10 adjusts the process parameters (2), (4), (5), (7), (9), and (10).

[0120] Furthermore, the prediction device 10 calculates the error between the etching amount distribution derived from the process parameters predicted up to step S307 and the target profile as a model residual (step S308).

[0121] At this point, if the termination condition is met (step S309, Yes), the prediction device 10 outputs the process parameters predicted up to step S307 as the final process parameters and terminates the process. For example, the termination condition may be that the model residual has become sufficiently small, or that a certain number of iterations have been performed.

[0122] If the termination condition is not met (step S309, No), the prediction device 10 returns to step S206 and repeats the process.

[0123] In steps S301, S302, S303, S304, and S305, process parameters mainly related to the ejection position are adjusted, and the shape of the etching amount distribution is determined. Furthermore, in step S307, process parameters mainly related to the ejection time are adjusted, and the shape of the etching amount distribution expands and contracts, approaching the target profile.

[0124] As described above, the etching control system 1 of the embodiment includes a prediction device 10 and an etching control device 20. The prediction device 10 includes an update unit 132, a calculation unit 131, and a supply unit 133. The update unit 132 updates the model parameters so that the model representing the relationship between process parameters, which are parameters for controlling the operation of multiple nozzles for etching a substrate, and the distribution of etching amount in the plane of the substrate is optimized. The calculation unit 131 uses the model whose parameters have been updated by the update unit 132 to calculate process parameters corresponding to a specified etching amount distribution. The supply unit 133 provides the process parameters calculated by the calculation unit 131. The etching control device 20 includes an acquisition unit 231 and an operation control unit 233. The acquisition unit 231 acquires process parameters. The operation control unit 233 uses the process parameters acquired by the acquisition unit 231 to control the operation of multiple nozzles. As a result, according to the embodiment, a complex etching amount distribution can be efficiently realized in wet etching.

[0125] The embodiments disclosed herein should be considered in all respects to be illustrative and not restrictive. The above embodiments may be omitted, replaced, or modified in various forms without departing from the scope and spirit of the appended claims.

[0126] The various processes described in the above embodiments can be achieved by executing a pre-prepared program on a computer. Therefore, below, an example of a computer that executes a program having the same functions as those in the above embodiments will be described. Figure 12 is a diagram showing an example of a computer that executes a program.

[0127] As shown in Figure 12, the computer 1000 includes a computer 1010 that performs various calculations, an input device 1020 that receives data input, and a monitor 1030. The computer 1000 also includes an interface device 1040 for connecting to various devices and a communication device 1050 for connecting to other information processing devices, etc., by wired or wireless connection. The computer 1000 also includes a RAM 1060 for temporarily storing various information and a storage device 1070. Each of the devices 1010 to 1070 is connected to a bus 1080.

[0128] The storage device 1070 stores programs having the same functions as the calculation unit 131, update unit 132, and provision unit 133 shown in Figure 4. The storage device 1070 also stores model information 121. The input device 1020 receives various types of information, such as operation information, from the user of the computer 1000. The monitor 1030 displays various screens, such as display screens, to the user of the computer 1000. The interface device 1040 is connected to, for example, a printing device. The communication device 1050 is connected to, for example, a network (not shown) and exchanges various types of information with other information processing devices.

[0129] Computer 1010 performs various processes by reading each program stored in the storage device 1070, loading it into the RAM 1060, and executing it. These programs also enable computer 1000 to function as the calculation unit 131, update unit 132, and provision unit 133 shown in Figure 4.

[0130] The above program does not necessarily have to be stored in the storage device 1070. For example, the program may be stored on a storage medium that the computer 1000 can read and execute. Examples of storage media that the computer 1000 can read include portable recording media such as CD-ROMs, DVDs (Digital Versatile Discs), and USB (Universal Serial Bus) memory, semiconductor memory such as flash memory, and hard disk drives. Alternatively, the program may be stored on a device connected to a public network, the internet, or a LAN, and the computer 1000 may read and execute the program from there.

[0131] Here, we have described an example of a computer for realizing the prediction device 10, but the etching control device 20 is also realized by a computer with a similar configuration to the one described here. [Explanation of Symbols]

[0132] 1. Etching control system 10 Prediction device 11, 21 I / F section 12, 22 Storage section 13, 23 Control Unit 20 Etching control device 30 Management Server 61 wafers 62, 63 nozzles 121 Model Information 131 Calculation Department 132 Update Department 133 Provision Department 221 Process Recipes 231 Acquisition Department 232 Update Department 233 Operation Control Unit

Claims

1. A prediction device that performs processing related to a model representing the relationship between process parameters for controlling the operation of multiple nozzles in wet etching of a substrate and the distribution of etching amount in the plane of the substrate, An update unit that optimizes the parameters of the model using training data based on the process parameters and the measured or referenced etching amount distribution within the plane of the substrate, A calculation unit that uses the updated model to calculate the process parameters corresponding to the distribution of a specified etching amount, A supply unit that provides the calculated process parameters to an external party, A prediction device characterized by having the following features.

2. The update unit optimizes the model parameters using the training data based on the process parameters that define the discharge position, discharge time, and movement speed of the plurality of nozzles, and the distribution. The prediction device according to feature 1.

3. An etching control system for dual dispensing, comprising a prediction device and an etching control device, The prediction device uses a model that represents the relationship between process parameters that define the discharge position, discharge time, and travel speed of a plurality of nozzles, including a first nozzle that discharges rinse onto a rotating substrate and a second nozzle that discharges a chemical solution for wet etching the substrate, and the distribution of etching amount in the plane of the substrate, to update the parameters of the model, calculate the process parameters using the model with the updated parameters, and provide the calculated process parameters. The etching control device acquires the process parameters and controls the operation of the first nozzle and the second nozzle. An etching control system characterized by the following:

4. An etching control system comprising a prediction device, a management server, and an etching control device that controls the operation of multiple nozzles in wet etching of a substrate, The prediction device calculates process parameters that define the discharge position, discharge time, and movement speed of the plurality of nozzles, and transmits the process parameters to the management server or the Host IF of the etching control device. The management server accepts the input of the process parameters provided by the prediction device via the management application. The etching control device acquires the process parameters via the Host IF or the management application and controls the plurality of nozzles based on the process parameters. An etching control system characterized by the following:

Images