JP7080804B2 - Processing condition determination method for plasma processing and processing condition determination device for plasma processing - Google Patents

Description

This specification discloses a processing condition determining method for plasma processing and a processing condition determining device for plasma processing.

Conventionally, a processing result estimation model that estimates the processing result based on the monitor output from the sensor that monitors the process amount during plasma processing and the preset prediction formula of the processing result, and the processing result based on the estimation result are the targets. A system including an optimum recipe calculation model for calculating a correction amount of processing conditions so as to be a value has been proposed (see, for example, Patent Document 1). In this system, the influence of disturbance can be suppressed by controlling the plasma processing based on the recipe (processing conditions) generated by the optimal recipe calculation model.

Japanese Unexamined Patent Publication No. 2005-33228

By the way, in actual plasma processing, it may be desired to fix (limit) the processing conditions for some of a plurality of items due to structural restrictions of the plasma processing device, restrictions on operating time, and the like. Even in such a case, it is required to optimize the processing conditions for the remaining items so that the plasma processing can be performed appropriately.

The present disclosure provides a processing condition determination method or a processing condition determining apparatus capable of estimating the processing conditions for the remaining items for appropriately performing plasma processing even if the processing conditions for some items are fixed. The main purpose is to provide.

The present disclosure has taken the following steps to achieve the above-mentioned main objectives.

The first method for determining the processing conditions for plasma processing of the present disclosure is
It is a processing condition determination method for plasma processing that determines the processing conditions required for plasma processing in which plasma is applied to the object to be processed from the head.
Accepts the user to select one of multiple items for which different types of processing conditions are defined.
A model was created by the machine learning method in which some of the above-mentioned multiple items including the selected item were used as input items and the remaining items were used as output items.
Accepting the input of processing conditions for the input items from the user,
Using the created model, the processing conditions for the output items are estimated from the processing conditions for the input items that have been input.
The gist is that.

In the first method for determining processing conditions for plasma processing of the present disclosure, first, the user accepts the selection of one of a plurality of items for which different types of processing conditions are defined, and the user selects one of the plurality of items. A model is created by the machine learning method with some items including the items marked with as input items and the remaining items as output items. Then, the processing condition determination method accepts the input of the processing condition for the input item from the user, and estimates the processing condition for the output item from the processing condition for the input item input using the created model. As a result, even if it becomes necessary to fix (limit) the processing conditions for some items due to structural restrictions on the plasma processing device or restrictions on the operating time, the processing conditions for the remaining items are appropriately set. It is possible to estimate. That is, the user can obtain appropriate processing conditions for the remaining items by selecting an item for which the processing conditions are to be fixed from a plurality of items and inputting the processing conditions for the items.

In the first method for determining processing conditions for plasma processing of the present disclosure, the plurality of items include essential items including the processing target value of the plasma processing and selection items that can be arbitrarily selected by the user. However, the input item has the required item and the item selected by the user among the selection items, and the output item has the item not selected by the user among the selection items. It may be a thing. By doing so, the user selects the item for which the processing condition is to be fixed from the selection items and inputs the processing condition for that item, so that the processing condition for the remaining items for achieving the processing target value can be set. Obtainable. In this case, the required item includes the processing target value, information about the processing object, and information about the processing environment, and the selection item includes the distance between the head and the processing object. It may include the relative moving speed between the head and the processing object.

The second method for determining the processing conditions for plasma processing of the present disclosure is
It is a processing condition determination method for plasma processing that determines the processing conditions required for plasma processing in which plasma is applied to the object to be processed from the head.
A machine learning method is used for each combination of multiple types of input items, using a model in which some items are input items and the remaining items are output items among a plurality of items for which different types of processing conditions are defined. Created by
The user accepts the selection of any one of the plurality of items, and also accepts the input of the processing conditions for the input item including the selected item from the user.
From each of the created models, a model including the selected item in the input item and the item not selected in the output item is extracted, and the input is made using the extracted model. Estimate the processing conditions for the output items from the processing conditions for the input items.
The gist is that.

The second method for determining the processing conditions for plasma processing of the present disclosure is a model in which, first, some of the plurality of items for which different types of processing conditions are defined are set as input items and the remaining items are set as output items. Is created by the machine learning method for each combination of multiple types with different input items. Next, the processing condition determination method accepts the selection of any one of a plurality of items from the user, and also accepts the input of the processing condition for the input item including the selected item from the user. Then, in the processing condition determination method, a model that includes selected items from each created model in the input items and items that are not selected in the output items is extracted, and input is performed using the extracted model. Estimate the processing conditions for output items from the processing conditions for the input items. As a result, even if it becomes necessary to fix (limit) the processing conditions for some items due to structural restrictions on the plasma processing device or restrictions on the operating time, the processing conditions for the remaining items are appropriately set. It is possible to estimate. That is, the user can obtain appropriate processing conditions for the remaining items by selecting an item for which the processing conditions are to be fixed from a plurality of items and inputting the processing conditions for the items.

In addition, instead of the above-mentioned method for determining the processing conditions for the first plasma processing of the present disclosure, a processing condition determining device for the first plasma processing can be used. Further, instead of the above-described method for determining the processing conditions for the second plasma processing of the present disclosure, a processing condition determining device for the second plasma processing can be used.

It is a schematic block diagram of a plasma processing apparatus 10. It is a schematic block diagram of a plasma processing apparatus 10. It is a schematic block diagram of the processing condition determination apparatus 50 for plasma processing of this embodiment. It is explanatory drawing which shows the list of the parameter used for plasma processing. It is a flowchart which shows an example of a model creation process. It is explanatory drawing which shows the state of model creation. It is explanatory drawing which shows the kind of each model. It is a flowchart which shows an example of a process condition determination process. It is a flowchart which shows the model creation process of a modification. It is a flowchart which shows the processing condition determination process of the modification.

A mode for carrying out the present disclosure will be described with reference to the drawings.

1 and 2 are schematic configuration diagrams of the plasma processing apparatus 10. FIG. 3 is a schematic configuration diagram of the processing condition determining device 50 for plasma processing of the present embodiment. The plasma processing apparatus 10 irradiates the work W with plasma to process the surface of the work W. The work W is, for example, a substrate. After the adhesive is applied to the substrate, the components are mounted. The plasma processing apparatus 10 improves the wettability (adhesive strength of the adhesive) of the substrate surface by subjecting the substrate to plasma treatment.

As shown in FIGS. 1 and 2, the plasma processing device 10 includes a head 20, a scanning device 30 that scans the head 20 in the X-axis direction (see FIG. 2), and a control device 40 that controls the entire device (see FIG. 2). ) And.

The head 20 includes a pair of electrodes, a holding member that holds the pair of electrodes so as to project into the reaction chamber, and a nozzle 21 having a spout communicating with the reaction chamber. The head 20 blows the plasma-generated processing gas onto the surface of the work W by applying a voltage to the pair of electrodes and passing the processing gas through the reaction chamber to eject the processing gas from the nozzle 21. As shown in FIG. 2, the scanning device 30 includes a slider 31 and a motor 32 that moves the slider 31 in the X-axis direction. The head 20 is fixed to the slider 31 and moves in the X-axis direction with the movement of the slider 31. The head 20 is configured so that the ejection port (tip portion) of the nozzle 21, the surface (upper surface) of the work W, and the distance L can be adjusted within a range of predetermined upper and lower limit values.

The control device 40 is configured as a microprocessor centered on a CPU, and includes a ROM, a RAM, an input / output interface, a communication interface, and the like in addition to the CPU. A signal or the like from a position sensor that detects the position of the slider 31 in the X-axis direction is input to the control device 40 via the input / output interface. From the control device 40, a control signal to the motor 32, a control signal to the head 20, and the like are output via the input / output interface. Further, the control device 40 communicates with the processing condition determining device 50 of the present embodiment via the communication interface. The control device 40 controls the head 20 according to the processing conditions transmitted from the processing condition determining device 50 to perform plasma processing on the work W.

The processing condition determining device 50 of the present embodiment is known among a plurality of items (parameters) in which different types of processing conditions are determined when plasma processing is performed on the work W using the plasma processing device 10 described above. The item of is used as an input item and the remaining unknown items are used as output items, and the processing conditions for output items are determined from the processing conditions for input items using a model. The processing condition determining device 50 is configured by using an external storage device such as a CPU, ROM, RAM, a hard disk or a flash memory drive (SSD), and a general-purpose computer having various interfaces. A display device 57 such as a liquid crystal display and an input device 58 such as a keyboard and a mouse are connected to the computer. As a functional block, the processing condition determination device 50 includes a selection item reception unit 51, a model creation unit 52, an input parameter reception unit 53, an output parameter estimation unit 54, and plasma processing result data. A storage unit 55 and a model storage unit 56 are provided. The plasma processing result data storage unit 55 is a storage unit that stores the processing conditions (plasma processing result data) used in the plasma processing executed in the past. The model storage unit 56 is a storage unit that stores the model created by the model creation unit 52. Each functional block functions as a unit of hardware such as a computer CPU, ROM, RAM, an external storage device, and various interfaces, and software including an installed program.

Here, the input items are composed of required items and selected items among the selected items that can be arbitrarily selected by the user. The output item consists of selected items that have not been selected by the user.

In the present embodiment, the required items are "adhesive strength", "water contact angle", "surface free energy", "saturated contact angle", "number of irradiations", "heater ON / OFF", and "nozzle type". , "Type of adhesive", "Work base material", "Work additive", "Work melting point", "Work thickness", "Atmospheric temperature", "Atmospheric humidity", "Supply gas temperature", Includes "surrounding gas flow", "surrounding gas air volume", "electrode length", "internal nozzle usage time", "work surface condition", etc. "Water contact angle" and "surface free energy" are indicators of the wettability of the work surface. "Saturate contact angle" indicates the minimum value of the water contact angle. "Number of irradiations" indicates the number of irradiations of plasma. "Heater ON / OFF" indicates the presence or absence of heating of the shield gas used to protect the treated portion. "Nozzle type" indicates the type of the nozzle 21 used for the head 20. "Type of adhesive" indicates the type of adhesive to be applied to the work W after plasma treatment is applied to the work W (substrate). "Work base material" indicates a base material such as aluminum or polypropylene. "Work additive" indicates a material such as glass fiber or rubber. "Atmospheric temperature" indicates the temperature around the work during plasma processing. "Atmospheric humidity" indicates the humidity around the work during plasma processing. The “supply gas temperature” indicates the temperature of the processing gas supplied to the plasma processing apparatus 10. The "surrounding gas flow" indicates the direction of the gas around the work during plasma processing. "Ambient gas air volume" indicates the air volume of the gas around the work during plasma processing. "Electrode length" indicates the length of the electrode inside the device. "Usage time of internal nozzle" indicates the usage time of the internal nozzle of the device. "Work surface condition" indicates the presence or absence of dirt on the work surface and the type of dirt. The above-mentioned essential items are examples, and can be appropriately determined depending on the plasma processing device 10 used and the type of plasma processing to be performed.

In the present embodiment, the selection items include "head movement speed", "head distance", and the like. As shown in FIG. 2, the “head moving speed” indicates the moving speed (scanning speed) V of the head 20 in the X-axis direction during plasma irradiation. As shown in FIG. 2, the “head distance” indicates the distance L between the tip (spout) of the nozzle 21 and the work surface. The selection item described above is an example, and may include, for example, a plasma injection amount.

The selection item receiving unit 51 accepts the selection of the selection item for which the user wants to fix (limit) the processing condition among the selection items via the input device 58. In the present embodiment, the selection type is either "head movement speed" or "head distance", or "head movement speed" or "head distance". There are three cases when not selected.

The model creation unit 52 creates a model for estimating the processing conditions for the output items from the processing conditions for the input items. FIG. 5 is a flowchart showing an example of the model creation process executed by the model creation unit 52. When the model creation process is executed, the model creation unit 52 first accepts the selection of an item by the user from the selection items by the selection item reception unit 51 (step S100). Subsequently, the model creation unit 52 sets the required items and the selected items selected by the user as input items, and sets the selected items not selected by the user as output items. (Step S110). Next, the model creation unit 52 sets the processing conditions for the set input items using the past plasma processing result data stored in the plasma processing result data storage unit 55 (step S120). Further, the model creation unit 52 sets the processing conditions for the set output items using the past plasma processing result data stored in the plasma processing result data storage unit 55 (step S130). Then, the model creation unit 52 creates a model using the machine learning method from the processing conditions for the set input items and the processing conditions for the output items (step S140), and registers the created model in the model storage unit 56. Then (step S150), the model creation process is terminated. As the machine learning method, for example, a well-known neural network can be used. In the present embodiment, for example, when the selection of "head movement speed" among the selection items "head movement speed" and "head distance" is accepted, the model creation unit 52 is as shown in FIG. In addition, the required items ("adhesion strength", "work surface condition", etc.) and "head movement speed" are set as input items, and "head distance" is set as output items, and each of the input items and output items is used. A model is created by machine learning method by giving past plasma processing result data. In the present embodiment, as shown in FIG. 7, the model creation unit 52 sets the “head movement speed” and the “head distance” as output items (unknown items) according to the user's selection, and the “model 1”. Model 2 with "head movement speed" as an input item (known item) and "head distance" as an output item, and model 3 with "head distance" as an input item and "head movement speed" as an output item. Create one of them.

The input parameter receiving unit 53 receives input of processing conditions for input items (required items and selected items selected by the user) via the input device 58. The details of the processing conditions are described above.

The output parameter estimation unit 54 estimates the processing conditions for the output items from the processing conditions for the input items using the model created by the model creation unit 52. FIG. 8 is a flowchart showing an example of the processing condition determination process executed by the output parameter estimation unit 54. First, the output parameter estimation unit 54 receives the input of the processing conditions for the input items by the input parameter reception unit 53 (step S200). Subsequently, the output parameter estimation unit 54 reads out the model created by the model creation unit 52 and stored in the model storage unit 56, and estimates the processing conditions for the output items from the processing conditions for the input items using the model. (Step S210). Then, the output parameter estimation unit 54 displays the processing conditions for the estimated output items on the display device 57, and displays the processing conditions for the input input items and the processing conditions for the estimated output items in the plasma processing device 10. It is transmitted to the control device 40 (step S220). The processing condition determination process is terminated. The user confirms the processing conditions displayed on the display device 57, and adjusts the head 20 such as “head distance”. The control device 40 of the plasma processing device 10 controls the head 20 according to the received processing conditions, and controls the scanning device 30 so that the head 20 moves according to the “moving speed of the head” to perform plasma processing on the work W. To give.

Here, the correspondence between the main elements of the embodiment and the main elements of the present disclosure described in the claims will be described. That is, the head 20 of the present embodiment corresponds to the head, the work W (board) corresponds to the object to be processed, the selection item reception unit 51 corresponds to the item selection reception means, and the model creation unit 52 serves as the model creation means. The input parameter receiving unit 53 corresponds to the condition input receiving means, and the output parameter estimating unit 54 corresponds to the processing condition estimating means.

The processing condition determination device 50 of the present embodiment described above accepts the selection of at least one item from a plurality of items for which different types of processing conditions are defined from the user, and selects the selected item from the plurality of items. Create a model by machine learning method with some items including some items as input items and the remaining items as output items. Then, the processing condition determination device 50 receives the input of the processing condition for the input item from the user, and estimates the processing condition for the output item from the processing condition for the input item input using the created model. As a result, even if the processing condition determining device 50 needs to fix (limit) the processing conditions for some items in advance due to reasons such as structural restrictions of the plasma processing device 10 and restrictions on the operating time. By creating a model corresponding to, it is possible to appropriately estimate the processing conditions for the remaining items using the model.

Further, the processing condition determination device 50 of the present embodiment simply allows the user to select an item whose processing is to be fixed from the selection items, and the required item and the selected item among the selection items are selected as input items. A model with the items that were not performed as output items is created by the machine learning method based on the past plasma processing result data. As a result, the user can obtain unknown processing conditions for plasma processing by a simple operation.

It should be noted that the present disclosure is not limited to the above-described embodiment, and it goes without saying that the present disclosure can be carried out in various embodiments as long as it belongs to the technical scope of the present disclosure.

For example, in the above-described embodiment, the processing condition determination device 50 creates a model in which some items including the selected item among the selected items are input items and the remaining items are output items, and the created model is created. Was used to estimate the processing conditions for output items from the processing conditions for input items. However, the processing condition determination device 50 may estimate the processing conditions for model creation and output items as follows. FIG. 9 is a flowchart showing a model creation process of a modification example executed by the model creation unit 52. FIG. 10 is a flowchart showing a processing condition estimation process of a modified example executed by the output parameter estimation unit 54.

In the model creation process of FIG. 9, the model creation unit 52 first determines the combination target item to be combined to be included in the input item among the selection items (step S300). Subsequently, the model creation unit 52 sets the required items and the combination target items as input items, and sets the selection items other than the combination target items (non-combination target items) as output items (step S310). .. Next, the model creation unit 52 sets the processing conditions for the set input items using the past plasma processing result data stored in the plasma processing result data storage unit 55 (step S320). Further, the model creation unit 52 sets the processing conditions for the set output items by using the past plasma processing result data stored in the plasma processing result data storage unit 55 (step S330). Then, the model creation unit 52 creates a model using the machine learning method from the processing conditions for the set input items and the processing conditions for the output items (step S340), and registers the created model in the model storage unit 56. (Step S350). Next, the model creation unit 52 determines whether or not the model creation is completed for all combinations of input items (step S360). When the model creation unit 52 determines that the model creation for the combination of all the input items has not been completed, the model creation unit 52 determines another combination target item (step S370), returns to step S310, and creates a model. repeat. On the other hand, when the model creation unit 52 determines that the model creation is completed for all the combinations of input items, the model creation process ends. When the "head movement speed" and the "head distance" are selected items, the model creation unit 52 creates the three models of the above-mentioned models 1 to 3 in advance and stores them in the model storage unit 56.

In the processing condition estimation process of FIG. 10, the output parameter estimation unit 54 first accepts the selection of an item by the user from the selection items by the selection item reception unit 51 (step S400). Subsequently, the output parameter estimation unit 54 sets the required items and the selected items selected by the user as input items, and sets the selected items not selected by the user as output items. (Step S410). Next, the output parameter estimation unit 54 receives the input of the processing conditions for the input items by the input parameter reception unit 53 (step S420). Subsequently, the output parameter estimation unit 54 selects a corresponding model based on the set input items and output items, reads them out from the model storage unit 56 (step S430), and uses the read models to read the processing conditions for the input items. The processing conditions for the output items are estimated from (step S440). Then, the output parameter estimation unit 54 displays the processing conditions for the estimated output items on the display device 57, and displays the processing conditions for the input input items and the processing conditions for the estimated output items in the plasma processing device 10. It is transmitted to the control device 40 (step S450). The processing condition determination process is terminated.

In the above-described embodiment, the present disclosure is in the form of a processing condition determining device, but it may be in the form of a processing condition determining method.

The present disclosure can be used in the manufacturing industry of plasma processing equipment and the like.

10 Plasma processing device, 20 heads, 21 nozzles, 30 scanning device, 31 slider, 32 motor, 40 control device, 50 processing condition determination device, 51 selection item reception unit, 52 model creation unit, 53 input parameter reception unit, 54 output Parameter estimation unit, 55 plasma processing result data storage unit, 56 model storage unit, 57 display device, 58 input device, W work.

Claims (6)

It is a processing condition determination method for plasma processing that determines the processing conditions required for plasma processing in which plasma is applied to the object to be processed from the head.
Accepts the user to select one of multiple items for which different types of processing conditions are defined.
A model was created by the machine learning method in which some of the above-mentioned multiple items including the selected item were used as input items and the remaining items were used as output items.
Accepting the input of processing conditions for the input items from the user,
Using the created model, the processing conditions for the output items are estimated from the processing conditions for the input items that have been input.
A method for determining processing conditions for plasma processing. The processing condition determination method for plasma processing according to claim 1.
The plurality of items include essential items including the processing target value of the plasma processing and selection items that can be arbitrarily selected by the user.
The input item includes the required item and an item selected by the user among the selection items.
The output item has an item of the selection items that has not been selected by the user.
A method for determining processing conditions for plasma processing. The processing condition determination method for plasma processing according to claim 2.
The required items include the processing target value, information on the processing object, and information on the processing environment.
The selection items include a distance between the head and the object to be processed, and a relative moving speed between the head and the object to be processed.
A method for determining processing conditions for plasma processing. It is a processing condition determination method for plasma processing that determines the processing conditions required for plasma processing in which plasma is applied to the object to be processed from the head.
A machine learning method is used for each combination of multiple types of input items, using a model in which some items are input items and the remaining items are output items among a plurality of items for which different types of processing conditions are defined. Created by
The user accepts the selection of any one of the plurality of items, and also accepts the input of the processing conditions for the input item including the selected item from the user.
From each of the created models, a model including the selected item in the input item and the item not selected in the output item is extracted, and the input is made using the extracted model. Estimate the processing conditions for the output items from the processing conditions for the input items.
A method for determining processing conditions for plasma processing. It is a processing condition determination device for plasma processing that determines the processing conditions required for plasma processing that irradiates the object to be processed with plasma from the head.
An item selection receiving means that accepts the selection of one of a plurality of items for which different types of processing conditions are defined from the user, and
A model creation means for creating a model by a machine learning method in which some items including the selected item among the plurality of items are input items and the remaining items are output items.
Condition input receiving means for receiving input of processing conditions for the input items from the user,
A processing condition estimation means for estimating the processing conditions for the output items from the processing conditions for the input items input using the created model, and
A processing condition determination device for plasma processing. It is a processing condition determination device for plasma processing that determines the processing conditions required for plasma processing that irradiates the object to be processed with plasma from the head.
A storage means for storing a model in which some items are input items and the remaining items are output items among a plurality of items for which different types of processing conditions are defined, for each combination of a plurality of different types of the input items. When,
A receiving means that accepts the selection of at least one of the plurality of items from the user and also accepts the input of the processing conditions for the input item including the selected item from the user.
From each model stored in the storage means, a model in which the selected item is included in the input item and the item not selected is included in the output item is extracted, and the extracted model is used. A processing condition estimation means for estimating the processing conditions for the output item from the processing conditions for the input input item, and
A processing condition determination device for plasma processing.

Images