TWI740313B - Virtual measurement method, device, and computer readbale storage medium - Google Patents

Abstract

A virtual measurement method includes: obtaining production data of at least one production equipment; using the production data to simultaneously predict prediction data of measured products and unmeasured products through a prediction model, the prediction data including size data of the products. The invention also provides a virtual measuring device and a computer readable storage medium.

Description

Virtual measurement method, device and computer readable storage medium

The invention relates to the field of measurement, in particular to a virtual measurement method, device and computer readable storage medium.

In the production process of semiconductors or panels, it is necessary to measure the film thickness or line width of the processed product in real time to ensure that the process is correct. In the early days, sampling was usually used for measurement, but the manufacturing process was complicated year by year and the accuracy increased sharply, so the frequency of sampling had to be increased to achieve the effect. However, the cost of the measurement machine is expensive, and the automation construction requires space and huge expenditure. Therefore, the cost of the existing measurement method is relatively high.

In view of the above content, it is necessary to propose a virtual measurement method, device and computer-readable storage medium to solve the above-mentioned problems.

The first aspect of the present invention provides a virtual measurement method, which includes: obtaining production data of at least one production device; using the production data to calculate the predicted data of the measured product and the unmeasured product through a prediction model, The forecast data includes size data of the product.

Further, after the prediction data is calculated, the method further includes: obtaining Take the measurement data that has been sampled on the measured product; determine whether the difference between the measurement data and the prediction data is within a preset range; when the difference is not within the preset range, use all Update the prediction model with the production data and the measurement data.

Further, the step of updating the prediction model specifically includes: generating a user interface to display the difference between the measurement data and the prediction data and the preset range; receiving an instruction to update the prediction model; using The production data and the measurement data are used to reconstruct or adjust the prediction model.

Further, after the prediction data is calculated, the method further includes: determining whether the prediction is successful; if the prediction is not successful, sending an alarm message; if the prediction is successful, generating a user interface to display the prediction data.

Further, the method further includes: acquiring the production materials and measurement data of the measured product; using the production materials and the measurement data to establish the prediction model, the prediction model being a machine learning model.

Further, the step of obtaining the production data and the measurement data of the measured product specifically includes: receiving the production data sent by at least one production device and the measurement data sent by at least one measurement device; The production data and the measurement data are extracted, converted and loaded; the production data and the measurement data are stored in an analysis database.

Further, the method further includes: comparing the measurement data of the same product by a plurality of the measurement devices at a predetermined time interval to calibrate the measurement data.

Further, the size information includes the film thickness and line width of the product.

A second aspect of the present invention provides a virtual measurement device. The device includes a processor and a memory. The memory stores a number of computer programs. The processor is used to execute the computer programs stored in the memory to implement the virtual measurement. Method steps.

The third aspect of the present invention provides a computer-readable storage medium, the computer The readable storage medium stores a plurality of instructions, and the plurality of instructions may be executed by one or more processors to implement the steps of the virtual measurement method described above.

The aforementioned virtual measurement method, device and computer-readable storage medium can use a predictive model to calculate the predicted data of the measured product and the unmeasured product, thereby realizing virtual measurement in industrial production and reducing the frequency of random inspections. It saves inspection costs and improves the accuracy and reliability of virtual measurement.

100: Virtual measuring device

200: production device

300: Measuring device

10: memory

20: processor

30: Virtual measurement setting program

40: communication unit

50: display unit

60: input unit

101: Get modules

102: Training Module

103: Prediction Module

104: User Interface Control Module

105: Judgment Module

106: Alarm module

107: Comparison module

Fig. 1 is a schematic diagram of an application environment of a virtual measurement device according to an embodiment of the present invention.

FIG. 2 is a schematic diagram of the architecture of a virtual measurement device according to an embodiment of the present invention.

FIG. 3 is a functional module diagram of a virtual measurement setting program according to an embodiment of the present invention.

Fig. 4 is a flowchart of a virtual measurement method according to an embodiment of the present invention.

The following specific embodiments will further illustrate the present invention in conjunction with the above-mentioned drawings.

In order to make the objectives, technical solutions, and advantages of the present invention clearer, the following further describes the present invention in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, but not used to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work shall fall within the protection scope of the present invention.

It should be further noted that, in this article, the terms "include", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements , And also include other elements not explicitly listed, or elements inherent to the process, method, article, or device. Without more restrictions In this case, the element defined by the sentence "including a..." does not exclude the existence of other same elements in the process, method, article or device that includes the element.

Please refer to FIG. 1, which is a schematic diagram of a preferred embodiment of the operating environment of the virtual measurement device of the present invention. In one embodiment, the virtual measurement device 100 is in communication connection with at least one production device 200 and at least one measurement device 300.

The production device 200 may be a production device used in the production process of semiconductors or panels, such as a set of production machines of the yellow light process, including but not limited to, pre-cleaning machines, photoresist coating machines, pre-baking machines, Exposure machine, developing machine, post-baking machine, etc.; it is understood that the production device can also be other devices, such as a coating machine, a solder paste printer, etc.

The measuring device 300 is used to measure various sizes of products. For example, line width, film thickness, etc. It can be understood that the measured size is not limited to this, and can be set according to requirements. For example, the size information may also include the length, width, and height dimensions and angles of the whole or part of the structure of the product.

Please refer to FIG. 2, which is a schematic diagram of a preferred embodiment of the virtual measurement device 100 of the present invention.

In one embodiment, the virtual measurement device 100 includes a memory 10, a processor 20, and a virtual measurement setting program 30 that is stored in the memory 10 and can be run on the processor 20. When the processor 20 executes the virtual measurement setting program 30, the steps in the embodiment of the virtual measurement method are implemented, such as steps S401 to S409 shown in FIG. 4. Alternatively, when the processor 20 executes the virtual measurement setting program 30, the function of each module in the embodiment of the virtual measurement setting program is realized, such as the modules 101 to 107 in FIG. 3.

The virtual measurement setting program 30 may be divided into one or more modules, and the one or more modules are stored in the memory 10 and executed by the processor 20 to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the virtual measurement setting program 30 in the virtual measurement device 100. For example, the virtual measurement setting program 30 can be divided into the acquisition module 101 in FIG. 3, The training module 102, the prediction module 103, the user interface control module 104, the judgment module 105, the alarm module 106, and the comparison module 107. The specific functions of each module refer to the function of each module in Figure 3 below.

The so-called processor 20 may be a central processing unit (Central Processing Unit, CPU), or other general-purpose processors, digital signal processors (Digital Signal Processors, DSP), and dedicated integrated circuits (Application Specific Integrated Circuits, ASICs). , Ready-made programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor, or the processor 20 may also be any conventional processor, etc. The processor 20 may use various interfaces and buses to connect to various parts of the virtual measurement device 100.

The memory 10 can be used to store the virtual measurement setting program 30 and/or the module, and the processor 20 runs or executes the computer program and/or module stored in the memory 10, and calls the storage The data in the memory 10 realizes various functions of the virtual measurement device 100. The memory 10 may include a high-speed random access memory, and may also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), and a secure digital (Secure Digital, SD) card, flash memory card (Flash Card), at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device.

In one embodiment, the virtual measurement device 100 further includes a communication unit 40, a display unit 50 and an input unit 60. The communication unit 40, the display unit 50, and the input unit 60 are electrically connected to the processor 20, respectively.

The communication unit 40 is used to establish a communication connection with the production device 200, the measurement device 300 or other computer devices in a wired or wireless manner. The communication unit 40 may be a wired communication unit or a wireless communication unit.

The display unit 50 is used to display the processing result of the processor 20. The display unit 50 may include at least one display screen or a touch screen.

The input unit 60 is used to input various information or commands. The input unit 60 may include one or more of a keyboard, a mouse, a touch screen, and the like.

Those skilled in the art can understand that the schematic diagram is only an example of the virtual measurement device 100 and does not constitute a limitation on the virtual measurement device 100. It may include more or less components than those shown in the figure, or combine some components. , Or different components, for example, the virtual measurement device 100 may also include network access equipment, a bus, and so on.

FIG. 3 is a functional module diagram of a preferred embodiment of the virtual measurement setting program of the present invention.

Referring to FIG. 3, the virtual measurement setting program 30 may include an acquisition module 101, a training module 102, a prediction module 103, a user interface control module 104, a judgment module 105, an alarm module 106, and a comparison module. 107. In one embodiment, the aforementioned modules may be programmable software instructions that are stored in the memory 10 and can be invoked and executed by the processor 20. It can be understood that, in other implementation manners, the above-mentioned modules may also be program instructions or firmware that are solidified in the processor 20.

The acquisition module 101 is used to acquire production data and measurement data.

In one embodiment, the acquisition module 101 is used to acquire the production data sent by the production device 200 and the measurement data sent by the measurement device 300.

The production materials include the production parameters of the production device 200. Taking the yellow light process machine as an example, the production parameters include numerical or nominal parameters, and the numerical parameters include temperature, time, voltage, and current related to photoresistance. , Rotation speed, etc., the name type parameters include the code of the pallet, etc.

The measurement data includes the size data of the product produced by the production device 200, the size data includes the film thickness and line width of the product, but is not limited to this, the size data may also include the length and width of the entire or partial structure of the product High size, angle and other information.

In one embodiment, the acquisition module 101 is further configured to receive an instruction to update the prediction model.

The training module 102 is used to establish and update a prediction model based on production data and measurement data. The prediction model may be a statistical model or a machine learning model.

The prediction module 103 is used to calculate the prediction data of the measured product and the unmeasured product by the prediction model based on the real-time production data, and the prediction data includes the size data of the product.

The user interface control module 104 is used to generate a user interface to display the user interface through the display unit 50.

In one embodiment, the user interface control module 104 generates a user interface to display the prediction data.

In one embodiment, the user interface control module 104 is also used to generate a user interface to display the difference between the measurement data and the prediction data and a preset range of the difference.

The judgment module 105 is used for judging whether the difference between the measurement data and the prediction data is within the preset range.

The judgment module 105 is also used to judge whether the prediction data is successfully calculated.

The alarm module 106 is used to send out alarm information when the prediction fails.

The comparison module 107 is used to compare the measurement data of the same product of the multiple measurement devices 300 to calibrate the measurement data.

Fig. 4 is a flowchart of a virtual measurement method in an embodiment of the present invention. According to different needs, the order of the steps in the flowchart can be changed, and some steps can be omitted.

Step S401, using production data and measurement data to establish a prediction model.

In one embodiment, the production materials of the production device 200 are first obtained, and the measurement data obtained by measuring the products produced by the production device 200 with the measurement device 300; and then the production data and the measurement data are used to establish a forecast Model.

The production data and corresponding measurement data can be stored in an analysis database. The analysis database includes a plurality of sample data, and each sample data includes the production data of a production device 200 and the measurement data of the corresponding product.

The production materials may include multiple production parameters of the production device 200. Taking the yellow light process machine as an example, the production parameters include numerical or nominal parameters, and the numerical parameters include temperature, time, voltage, Current, rotation speed, etc., the name type parameters include the code of the tray, etc.; the measurement data can include the film thickness and line width.

For another example, when the production device 200 is a coating machine, its production materials may include one or more of the distance between the target and the substrate, the coating gas concentration, the coating time, the target sputtering speed, and the gear rotation speed; The measurement data can include film thickness and line width.

For another example, when the production device 200 is a solder paste printer, its production data can include parameters such as squeegee pressure, printing speed, demolding speed, and demolding distance; measurement data can include solder paste height, solder paste area, and solder paste volume.

In one embodiment, the step of obtaining production materials and measurement data specifically includes: receiving the production materials issued by at least one of the production devices 200 and the measurement data issued by at least one measurement device 300; The production data and the measurement data are extracted, transformed and loaded (Extract-Transform-Load, ETL); the production data and the measurement data are stored in an analysis database.

The prediction model is a statistical model or a machine learning model, such as a CNN or RNN neural network model. After establishing a prediction model using multiple production data and corresponding measurement data, the test sample data is then logged into the prediction model for testing. When the test results meet the preset requirements, the prediction model can be applied to virtual measurement. It can be understood that after the prediction model is established, as the sample data continues to increase, the prediction model can continue to be updated with new sample data. When building the predictive model, domain knowledge or analyst experience can be added.

In one embodiment, a prediction model can be established for different groups of production devices 200, different measurement targets, and different measurement points, and then aggregated into a predicted value of a product based on the cut products.

Step S402: Obtain production materials.

Specifically, real-time production materials of at least one production device 200 are obtained.

In step S403, the forecast data of the measured product and the forecast data of the unmeasured product are simultaneously calculated by the forecast model.

Specifically, the obtained production data is used to adapt the prediction data of the measured product and the prediction data of the unmeasured product through the prediction model. The prediction data includes size data of the product, and whether the product is qualified can be predicted by the prediction data.

In step S404, it is judged whether the prediction is successful.

If the prediction is successful, then go to step S406; if the prediction fails, then go to step S405 to issue an alarm message.

Specifically, when the production data is not obtained or the prediction data is not successfully calculated, the judgment module 105 judges that the prediction fails, and the alarm module 106 can send the alarm information to the Computer Integrated Manufacturing (CIM) engineer, or Send it to the Manufacturing Execution System (MES) so that the engineer can handle the exception in time. The alarm module 106 can also control the display unit to issue an early warning prompt.

In step S406, a user interface is generated to display the forecast data.

When the prediction is successful, a user interface is generated, and the display unit 50 can display the user interface to display the forecast data for the engineer's reference.

Step S407: Obtain measurement data for random inspection of unmeasured products.

In order to avoid the loss of follow-up production due to incorrect predictions, without increasing the burden on the factory, the measurement data in the random inspection program carried out before the shipment of the unmeasured products is used to verify the forecast data. material. Step S407 is a non-essential step, which can be determined according to the production situation of the factory, such as whether the goods are on-demand or the precision of the product.

Step S408: Determine whether the difference between the measurement data and the prediction data is within a preset range.

The preset range is an allowable error range, which can be set according to requirements. If it is determined that the difference between the measurement data and the prediction data exceeds the preset range, step S409 is entered; if it is determined that the difference between the measurement data and the prediction data is within the prediction If it is within the range, the prediction model can continue to be used, and return to step S402.

In step S409, the prediction model is updated using production data and measurement data.

When updating the prediction model, the original prediction model can be deleted and the new prediction model can be reconstructed based on the original and newly acquired production data and measurement data in the analysis database, or only the original prediction model can be adjusted, such as adjusting The initial statistical model parameters or the total number of layers of the machine learning model (for example, the number of hidden layers) and/or the number of neurons in each layer. After updating the prediction model, return to step S402.

In one embodiment, step S409 specifically includes the following steps.

First, a user interface is generated to display the difference between the measurement data and the prediction data and the preset threshold.

Second, receive instructions to update the prediction model.

Then, using the production data and the measurement data to reconstruct or modify the prediction model.

In other embodiments, step S401 can be omitted, and the virtual measurement can be realized by using the established prediction model.

In other embodiments, steps S404 to S409 may also be omitted.

In other embodiments, the method may further include the step of comparing the measurement data of the same product by a plurality of the measurement devices 300 at predetermined intervals to calibrate the measurement data.

It can be understood that for the same product and the same film layer, multiple measurement data can be obtained through measurement by multiple measurement devices 300, and comparison of multiple measurement data can be used for the factory personnel to calibrate the measurement device 300.

The above-mentioned virtual measurement device 100, method, and computer-readable storage medium can obtain production data of at least one production device; using the production data, a prediction model is used to calculate the predicted data of the measured product and the unmeasured product The prediction data includes the size data of the product; the measurement data for sampling the measured product is obtained. Therefore, the above-mentioned virtual measurement device 100, method, and computer-readable storage medium can realize virtual measurement in industrial production. Measure, improve the quality of measurement with a smaller cost. In addition, the aforementioned virtual measurement device 100, method, and computer-readable storage medium can also determine whether the difference between the measurement data and the prediction data is within a preset range; when the difference exceeds the prediction When setting the scope, update the prediction model using the production data and the measurement data. Therefore, the above-mentioned virtual measurement device 100, method, and computer-readable storage medium can reduce the frequency of random inspections and save inspection costs; it can also monitor forecast data to prevent false predictions from affecting subsequent production, and improve the accuracy of virtual measurement Sex and reliability.

In summary, this publication clearly meets the requirements of a patent for invention, so it filed a patent application in accordance with the law. However, the above are only the preferred embodiments of the present invention, and cannot limit the scope of the patent application in this case. All the equivalent modifications or changes made by persons familiar with the technique of the present application in accordance with the spirit of the present invention shall be covered by the scope of the following patent applications.

Claims (8)

A virtual measurement method includes: obtaining the production data of at least one production device; using the production data to simultaneously calculate the prediction data of the measured product and the prediction data of the unmeasured product through a prediction model, the prediction The data includes the size data of the product; obtain the measurement data for random inspection of the unmeasured products; determine whether the difference between the measurement data and the forecast data is within a preset range; when the difference is not in the expected range When the range is set, updating the prediction model using the production data and the measurement data includes: generating a user interface to display the difference between the measurement data and the prediction data and the prediction Set the scope; receive an instruction to update the prediction model; use the production data and the measurement data to reconstruct or adjust the prediction model. The virtual measurement method according to claim 1, wherein, after the prediction data is calculated, the method further includes: judging whether the prediction is successful; if the prediction is not successful, sending an alarm message; if the prediction is successful, generating a The user interface to display the forecast data. The virtual measurement method according to claim 1, wherein the method further comprises: obtaining the production data and the measurement data of the measured product; using the production data and the measurement data to establish the prediction Model, the prediction model is a statistical model or a machine learning model. The virtual measurement method according to claim 3, wherein the step of obtaining the production data and the measurement data of the measured product includes: receiving the production data and at least one measurement sent by the at least one production device The measurement data sent by the device; Extract, convert and load the production data and the measurement data; store the production data and the measurement data in an analysis database. The virtual measurement method according to claim 4, wherein the method further comprises: comparing measurement data of the same product by a plurality of measurement devices at predetermined intervals to correct the measurement data. The virtual measurement method according to claim 1, wherein the size data includes the film thickness and line width of the product. A virtual measurement device, the device includes a processor and a memory, and a number of computer programs are stored on the memory. When the processor is used to execute the computer programs stored in the memory, any one of request items 1 to 6 is realized. The steps of the virtual measurement method. A computer-readable storage medium, wherein the computer-readable storage medium stores a plurality of instructions, and the plurality of instructions can be executed by one or more processors, so as to realize the requirements described in any one of claims 1 to 6 The steps of the virtual measurement method.

Images