Photoresist consumption prediction method and device, storage medium and equipment

Abstract

The invention discloses a photoresist consumption prediction method and device, a storage medium and equipment. The method comprises the following steps: acquiring the number of wafers fed in a to-be-predicted time length; using a preset polynomial regression model to predict the photoresist consumption in the to-be-predicted duration, and outputting a prediction result; the goodness of fit of the preset polynomial regression model is greater than that of a linear regression model. By adopting the scheme provided by the invention, the accuracy of predicting the consumption of the photoresist can be improved.

Description

Technical Field

[0001] The present invention relates to the field of semiconductor technology, and particularly to a method and device for predicting the consumption of photoresist, a storage medium, and a device. Background Art

[0002] Photo Resist (PR), also known as photoresist, is commonly used in the lithography stage of semiconductor manufacturing to form patterns on wafers. The consumption of photoresist generally needs to be calculated in advance for booking from photoresist manufacturers.

[0003] If the calculated value of photoresist consumption is too small compared with the actual value, it will affect the R & D and production progress of products. If the photoresist consumption is too large compared with the actual value, it will cause the accumulation of photoresist inventory, and may lead to the scrapping and waste of photoresist due to the influence of the photoresist shelf life. Therefore, the prediction of photoresist consumption is very important.

[0004] However, the existing methods for predicting the consumption of photoresist have poor accuracy. Summary of the Invention

[0005] The problem to be solved by the present invention is: how to improve the accuracy of predicting the consumption of photoresist.

[0006] To solve the above problems, an embodiment of the present invention provides a method for predicting the consumption of photoresist, which is characterized by including:

[0007] Obtain the number of wafers loaded within the duration to be predicted;

[0008] Use a preset polynomial regression model to predict the consumption of photoresist within the duration to be predicted, and output a prediction result; the goodness of fit of the preset polynomial regression model is greater than that of the linear regression model.

[0009] In a possible embodiment, the goodness of fit of the preset polynomial regression model is greater than a goodness-of-fit threshold.

[0010] In a possible embodiment, the preset polynomial regression model is the regression model with the highest goodness of fit.

[0011] In a possible embodiment, the preset polynomial regression model is obtained by the following method:

[0012] Collect the historical number of wafers loaded and the corresponding historical consumption of photoresist as training samples;

[0013] Use the collected data to train a number of initial regression models to obtain trained regression models;

[0014] Select the regression model with the highest goodness of fit from the trained regression models as the preset polynomial regression model.

[0015] In a possible embodiment, the preset polynomial regression model is a quadratic polynomial regression model.

[0016] In a possible embodiment, the expression of the quadratic polynomial regression model is:

[0017] Y = a1X2 + b1X + c1

[0018] Wherein, Y represents the consumption of photoresist within a preset time period; X represents the number of wafers put into production per month; a1, b1, and c1 are model parameters.

[0019] In a possible embodiment, the preset polynomial regression model is a cubic polynomial regression model.

[0020] In a possible embodiment, the expression of the cubic polynomial regression model is:

[0021] Y = a2X3 + b2X2 + c2X + d

[0022] Wherein, Y represents the consumption of photoresist within a preset time period; X represents the number of wafers put into production per month; a2, b2, and c2 are model parameters.

[0023] In a possible embodiment, the duration to be predicted is one month.

[0024] The embodiment of the present invention also provides a prediction device for the consumption of photoresist, and the device includes:

[0025] An acquisition unit, adapted to acquire the number of wafers put into production within the duration to be predicted;

[0026] A prediction unit, adapted to predict the consumption of photoresist within the duration to be predicted by using a preset polynomial regression model; the goodness of fit of the preset polynomial regression model is greater than that of the linear regression model;

[0027] An output unit, adapted to output a prediction result.

[0028] The embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored, and the computer program is executed by a processor to implement the steps of any one of the above methods.

[0029] The embodiment of the present invention also provides an electronic device, including a memory and a processor, wherein a computer program capable of running on the processor is stored on the memory, and when the processor runs the computer program, it executes the steps of any one of the above methods.

[0030] Compared with the prior art, the technical solution of the embodiment of the present invention has the following advantages:

[0031] Applying the solution of the present invention, the consumption of photoresist within the to-be-predicted duration is predicted by using a preset polynomial regression model. Since the goodness of fit of the preset polynomial regression model is greater than that of the linear regression model, the accuracy of predicting the consumption of photoresist can be effectively improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 is a flowchart of a method for predicting the consumption of photoresist in an embodiment of the present invention;

[0033] Figure 2 is a flowchart of a method for obtaining a preset polynomial regression model in an embodiment of the present invention;

[0034] Figure 3 is a schematic structural diagram of a device for predicting the consumption of photoresist in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS )

[0035] Currently, when predicting the monthly consumption of photoresist, the commonly used mathematical model is: Y = kX, where Y is the monthly consumption of photoresist, X is the number of wafers put into production per month, and k is the consumption of photoresist per wafer. Among them, k can be estimated through historical data.

[0036] It is found by the inventor through research that the goodness of fit of the above mathematical model is relatively low, so the accuracy of the obtained prediction result is relatively poor.

[0037] To solve this problem, the present invention provides a method for predicting the consumption of photoresist. By applying this method, a preset polynomial regression model is used to predict the consumption of photoresist within the to-be-predicted duration, thereby effectively improving the accuracy of predicting the consumption of photoresist.

[0038] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following detailed description of the specific embodiments of the present invention will be given with reference to the accompanying drawings.

[0039] Referring to Figure 1 , an embodiment of the present invention provides a method for predicting the consumption of photoresist, and the method may include the following steps:

[0040] Step 11, obtaining the number of wafers put into production within the to-be-predicted duration.

[0041] In specific implementation, the to-be-predicted duration can be set according to actual needs. For example, when ordering photoresist from a photoresist manufacturer monthly, the to-be-predicted duration can be one month, thereby predicting the monthly consumption of photoresist. When ordering photoresist from a photoresist manufacturer quarterly, the to-be-predicted duration can be one quarter, thereby predicting the quarterly consumption of photoresist.

[0042] In a specific implementation, the number of wafers put into production within the to-be-predicted time period is obtained, and based on this, the monthly consumption of wafers within the to-be-predicted time period can be predicted. Taking the prediction of the monthly consumption of photoresist as an example, the consumption of photoresist for the next month can be predicted in the current month. At this time, the number of wafers put into production obtained should be the number of wafers put into production in the next month.

[0043] Step 12, use a preset polynomial regression model to predict the consumption of photoresist within the to-be-predicted time period and output a prediction result; the goodness of fit of the preset polynomial regression model is greater than that of the linear regression model.

[0044] In a specific implementation, the preset polynomial regression model is a type of regression model. Through the research of the inventor, it is found that the polynomial regression model has a higher goodness of fit and more accurate prediction compared to the linear regression model.

[0045] In a specific implementation, the goodness of fit refers to the degree of fitting of the regression line to the observed values. The statistic for measuring the goodness of fit is the coefficient of determination (also known as the coefficient of determination) R². The maximum value of R² is 1. The closer the value of R² is to 1, the better the fitting degree of the regression line to the observed values; conversely, the smaller the value of R², the worse the fitting degree of the regression line to the observed values.

[0046] In a specific implementation, multiple linear regression models can be trained. Each regression model is trained with the number of wafers put into production as the independent variable and the consumption of photoresist as the dependent variable. From the multiple trained linear regression models, select the regression model whose goodness of fit is greater than the goodness-of-fit threshold, that is, the regression model with an R² value greater than the preset goodness-of-fit threshold, as the photoresist consumption prediction model.

[0047] In an embodiment of the present invention, the regression model with the highest goodness of fit can be selected as the photoresist consumption prediction model.

[0048] Through the research of the inventor, it is found that the goodness of fit of the polynomial regression model is greater than that of the linear regression model

[0049] In an embodiment of the present invention, the preset polynomial regression model is a quadratic polynomial regression model. Specifically, the expression of the quadratic polynomial regression model is:

[0050] Y = a1X

[0052] ,

[0051] , , 2 , + b1X + c1 (1)

[0051] Wherein, Y represents the consumption of photoresist within a preset time period; X represents the number of wafers put into production per month; a1, b1, and c1 are model parameters.

[0052] In another embodiment of the present invention, the preset polynomial regression model is a cubic polynomial regression model. Specifically, the expression of the cubic polynomial regression model is:

[0053] Y = a2X 3 + b2X 2 + c2X + d

[0054] Wherein, Y represents the consumption of photoresist within a preset time period; X represents the number of wafers put into production per month; a2, b2, and c2 are model parameters.

[0055] Subsequently, after obtaining the number of wafers put into production within the to-be-predicted time period, substituting it into formula (1) or (2), the corresponding consumption of photoresist can be obtained.

[0056] Using the polynomial regression model as the prediction model for the consumption of photoresist can more accurately predict the consumption of photoresist, thereby avoiding affecting the R & D and production progress of products, or avoiding the waste of photoresist caused by the accumulation of photoresist inventory.

[0057] Figure 2 It is a schematic diagram of the method for obtaining the preset polynomial regression model in an embodiment of the present invention. Referring to Figure 2 , the method may include:

[0058] Step 21, collecting the historical number of wafers put into production and the corresponding historical consumption of photoresist as training samples.

[0059] In a specific implementation, when the preset polynomial regression model is used to predict the consumption of photoresist for the to-be-predicted time period, the number of wafers put into production within the historical to-be-predicted time period and the corresponding historical consumption of photoresist can be collected for model training.

[0060] For example, when the polynomial regression model is used to predict the monthly consumption of photoresist in May 2025, the monthly consumption of photoresist before May 2025 and the number of wafers put into production per month before May 2025 can be collected for model training.

[0061] Step 22, using the collected data to train a number of initial regression models to obtain the trained regression models.

[0062] In a specific implementation, the number of initial regression models may include: logarithmic function model, exponential function model, sine function model, cosine function model, hyperbolic function model, etc. Set initial parameters for each initial regression model, and use the collected data to train the initial regression models. Compare the training results each time with the corresponding historical consumption of photoresist, so as to continuously optimize the model until the model parameters are stable.

[0063] Step 23: Select the regression model with the highest goodness of fit from the trained regression models as the preset polynomial regression model.

[0064] In specific implementation, the goodness of fit of each trained regression model can be calculated, that is, the coefficient of determination R² of each trained regression model is calculated, and the regression model with the largest coefficient of determination R² is selected as the preset polynomial regression model.

[0065] As can be seen from the above, in the solution of the present invention, a polynomial regression model is used to predict the consumption of photoresist within the to-be-predicted duration. Since this polynomial regression model has a high goodness of fit, it can more accurately predict the consumption of photoresist. Moreover, this polynomial regression model is simple and practical, and has stronger practical applicability.

[0066] To enable those skilled in the art to better understand and implement the present invention, the corresponding device, test system, electronic device and computer-readable storage medium of the above method are described in detail below.

[0067] Referring to Figure 3 , an embodiment of the present invention further provides a prediction device 30 for the consumption of photoresist. The device 30 may include: an acquisition unit 31, a prediction unit 32, and an output unit 33. Among them:

[0068] The acquisition unit 31 is adapted to acquire the number of wafers loaded within the to-be-predicted duration;

[0069] The prediction unit 32 is adapted to predict the consumption of photoresist within the to-be-predicted duration by using the preset polynomial regression model; the goodness of fit of the preset polynomial regression model is greater than that of the linear regression model;

[0070] The output unit 33 is adapted to output the prediction result.

[0071] In some embodiments, the device 30 may further include a training unit 34. The training unit 34 may collect the historical number of wafers loaded and the corresponding historical consumption of photoresist as training samples, and use the collected data to train a number of initial regression models to obtain trained regression models, and finally select the regression model with the highest goodness of fit from the trained regression models as the preset polynomial regression model.

[0072] In some embodiments, the preset polynomial regression model may be a quadratic polynomial regression model.

[0073] In other embodiments, the preset polynomial regression model may be a cubic polynomial regression model.

[0074] Regarding the obtaining unit 31, the prediction unit 32, and the output unit 33, specific implementation can refer to the above description of steps 11 to 13, and will not be elaborated here.

[0075] By using the photoresist consumption prediction device 30 in the embodiment of the present invention, the photoresist consumption can be accurately predicted. Moreover, the polynomial regression model is simple and practical, and has stronger practical applicability.

[0076] The embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored. The computer program is executed by a processor to implement the steps of any of the above methods.

[0077] In specific implementation, the computer-readable storage medium may include: ROM, RAM, disk, or optical disc, etc.

[0078] The embodiment of the present invention also provides an electronic device. The electronic device includes a memory and a processor. A computer program capable of running on the processor is stored on the memory. When the processor runs the computer program, it executes the steps of any of the above methods.

[0079] Regarding each device and product described in the above embodiments and the respective modules / units included therein, they can be software modules / units, hardware modules / units, or partially software modules / units and partially hardware modules / units. For example, for each device and product applied to or integrated into a chip, the respective modules / units included therein can all be implemented in a hardware manner such as circuits. Or, at least some of the modules / units can be implemented in the form of software programs, which run on a processor integrated inside the chip, and the remaining (if any) part of the modules / units can be implemented in a hardware manner such as circuits; for each device and product applied to or integrated into a chip module, the respective modules / units included therein can all be implemented in a hardware manner such as circuits. Different modules / units can be located in the same component (such as a chip, a circuit module, etc.) or different components of the chip module. Or, at least some of the modules / units can be implemented in the form of software programs, which run on a processor integrated inside the chip module, and the remaining (if any) part of the modules / units can be implemented in a hardware manner such as circuits; for each device and product applied to or integrated into a terminal, the respective modules / units included therein can all be implemented in a hardware manner such as circuits. Different modules / units can be located in the same component (such as a chip, a circuit module, etc.) or different components inside the terminal. Or, at least some of the modules / units can be implemented in the form of software programs, which run on a processor integrated inside the terminal, and the remaining (if any) part of the modules / units can be implemented in a hardware manner such as circuits.

[0080] Although the present invention is disclosed as above, the present invention is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the scope defined by the claims.

Claims

1. A method for predicting the consumption of photoresist, characterized in that, Including: Obtaining the number of wafers loaded within the to-be-predicted duration; Predicting the consumption amount of photoresist within the to-be-predicted duration by using a preset polynomial regression model, and outputting a prediction result; The goodness of fit of the preset polynomial regression model is greater than that of the linear regression model.

2. The method for predicting the consumption of photoresist according to claim 1, wherein The goodness of fit of the preset polynomial regression model is greater than a goodness-of-fit threshold.

3. The method for predicting the consumption of photoresist according to claim 1, wherein, The preset polynomial regression model is the regression model with the highest goodness of fit.

4. The method for predicting the consumption of photoresist according to claim 3, wherein The preset polynomial regression model is obtained by the following method: Collecting the historical number of wafers loaded and the corresponding historical consumption amount of photoresist as training samples; Training a number of initial regression models by using the collected data to obtain trained regression models; Selecting the regression model with the highest goodness of fit from the trained regression models as the preset polynomial regression model.

5. The method for predicting the consumption of photoresist according to claim 2 or 3, characterized in that, The preset polynomial regression model is a quadratic polynomial regression model.

6. The method for predicting the consumption of photoresist according to claim 5, characterized in that, The expression of the quadratic polynomial regression model is: Y = a1X 2 + b1X + c1 Wherein, Y represents the consumption amount of photoresist within a preset duration; X represents the number of wafers loaded per month; a1, b1 and c1 are model parameters.

7. The method for predicting the consumption of photoresist according to claim 2 or 3, characterized in that The preset polynomial regression model is a cubic polynomial regression model.

8. The method for predicting the consumption of photoresist according to claim 7, wherein, The expression of the cubic polynomial regression model is: Y = a2X 3 + b2X 2 + c2X + d Wherein, Y represents the consumption amount of photoresist within a preset duration; X represents the number of wafers loaded per month; a2, b2 and c2 are model parameters.

9. The method for predicting the consumption of photoresist according to claim 1, wherein, The to-be-predicted duration is one month.

10. A prediction device for the consumption amount of photoresist, characterized in that, Including: An obtaining unit, adapted to obtain the number of wafers loaded within the to-be-predicted duration; A predicting unit, adapted to predict the consumption amount of photoresist within the to-be-predicted duration by using a preset polynomial regression model; The goodness of fit of the preset polynomial regression model is greater than that of the linear regression model; An output unit, adapted to output a prediction result.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer program is executed by a processor to implement the steps of the method according to any one of claims 1 to 9.

12. An electronic device, comprising a memory and a processor, wherein a computer program capable of running on the processor is stored on the memory, characterized in that, When the processor runs the computer program, it executes the steps of the method according to any one of claims 1 to 9.

Images