CN-120874545-B - Wet process parameter prediction method and system for solar cell manufacturing

Abstract

The application provides a wet process parameter prediction method and a system for solar cell manufacturing, which belong to the technical field of parameter tuning, wherein the wet process parameter prediction method for solar cell manufacturing comprises the steps of dividing the wet process into a pretreatment process, a purification process, an etching process and a post-treatment process according to the process of the wet process; the method comprises the steps of respectively creating a pretreatment prediction model, a purification prediction model, an etching prediction model and a post-treatment prediction model, obtaining moderate value functions, obtaining preset technological parameters in a wet process, wherein the technological parameters comprise a silicon wafer type, a purification liquid parameter, an etching liquid parameter and a cleaning liquid parameter, inputting the moderate value functions and the preset technological parameters into the technological parameter prediction model, obtaining an initial population according to the preset technological parameters, guiding the evolution of the initial population according to the moderate value functions, and searching to obtain optimized technological parameters. The stability and consistency of the wet process in the solar cell manufacturing process are remarkably improved.

Inventors

• WANG SHIXIAN
• DAI JUN
• XU BINBIN
• GUAN YIJUN

Assignees

• 元能微电子科技南通有限公司

Dates

Publication Date

20260508

Application Date

20250715

Claims (5)

1. 1. The wet process parameter prediction method for solar cell manufacturing is characterized by comprising the following steps of: Dividing the wet process into a pretreatment process, a purification process, an etching process and a post-treatment process according to the process of the wet process; Respectively creating a pretreatment prediction model corresponding to the pretreatment process, a purification prediction model corresponding to the purification process, an etching prediction model corresponding to the etching process and a post-treatment prediction model corresponding to the post-treatment process; Obtaining preset technological parameters in a wet process, wherein the technological parameters comprise a silicon wafer type, a purified liquid parameter, an etching liquid parameter and a cleaning liquid parameter, and obtaining an initial population according to the technological parameters; Inputting the silicon chip type and the cleaning liquid parameters into the pretreatment prediction model to output the cleanliness prediction data and the cleaning liquid residue prediction data of the silicon chip surface; inputting the purified liquid parameters, the cleanliness prediction data of the surface of the silicon wafer and the cleaning liquid residue prediction data into the purified prediction model to output the predicted purity of the silicon wafer; inputting the predicted purity of the silicon wafer and the etching liquid parameters into the etching prediction model to output etching prediction depth and etching residual prediction data; Inputting the cleaning fluid parameters, the etching predicted depth and the etching residual predicted data into a post-processing prediction model to output the predicted cleanliness of the silicon wafer product and the predicted flatness of the silicon wafer product; Acquiring a first weight of an etching prediction depth, a second weight of a silicon wafer product prediction cleanliness and a third weight of a silicon wafer product prediction flatness; According to the first weight, the second weight and the third weight, weighting and summing the etching predicted depth, the silicon wafer product predicted cleanliness and the silicon wafer product predicted flatness to obtain a moderate value; and guiding the initial population to evolve according to the moderate value so as to search and obtain optimized technological parameters.
2. 2. The method of claim 1, wherein said deriving an initial population from said process parameters comprises: respectively obtaining the purification liquid parameter, the etching liquid parameter and the limit interval of the cleaning liquid parameter; obtaining the number of individuals of the initial population; And generating initial individuals of the purified liquid parameter, the etching liquid parameter and the cleaning liquid parameter at equal intervals in the limiting interval according to the number of the individuals, so as to obtain an initial population.
3. 3. The method according to claim 2, wherein creating a pretreatment prediction model corresponding to the pretreatment process, a purification prediction model corresponding to the purification process, an etching prediction model corresponding to the etching process, and a post-treatment prediction model corresponding to the post-treatment process, respectively, comprises: The method comprises the steps of constructing a training data set, wherein the training data set comprises process parameters, and cleaning degree measurement data, cleaning liquid residue measurement data, silicon wafer measurement purity, etching measurement depth, etching residue measurement data, silicon wafer product measurement cleaning degree and silicon wafer product measurement flatness of a silicon wafer surface corresponding to the process parameters; Training a first preset neural network model according to the training data set to obtain the preprocessing prediction model; Training a second preset neural network model according to the training data set to obtain the purification prediction model; training a third preset neural network model according to the training data set to obtain the etching prediction model; And training a fourth preset neural network model according to the training data set to obtain the post-processing prediction model.
4. 4. The method for predicting wet process parameters for solar cell fabrication of claim 3, Training a first preset neural network model according to the training data set to obtain the preprocessing prediction model, wherein the training data set comprises the following steps: Taking the cleanliness measurement data and the cleaning fluid residue prediction parameter in the training data set as a first training label, and taking the silicon wafer type and the cleaning fluid parameter in the process parameter as a first training sample; Inputting the first training sample into the first preset neural network model to obtain the cleaning fluid parameter prediction data and the cleaning fluid parameter prediction parameters; Based on the first training label, the cleaning fluid parameter prediction data and the cleaning fluid parameter prediction parameters, carrying out iterative updating on the first preset neural network model until an iterative termination condition is reached, and obtaining the preprocessing prediction model; training a second preset neural network model according to the training data set to obtain the purification prediction model, wherein the training comprises the following steps: Taking the silicon wafer measurement purity in the training data set as a second training label, and taking cleaning liquid parameter prediction data, cleaning liquid parameter prediction parameters and purified liquid parameters in the process parameters as second training samples; inputting the second training sample into the second preset neural network model to obtain the predicted purity of the silicon wafer; Based on the second training label and the silicon chip prediction purity, carrying out iterative updating on the second preset neural network model until reaching an iterative termination condition to obtain the purification prediction model; Training a third preset neural network model according to the training data set to obtain the etching prediction model, wherein the training comprises the following steps: Taking the cleanliness measurement data in the training data set as a third training label, and taking the predicted purity of the silicon wafer and etching liquid parameters in the process parameters as a third training sample; inputting the third training sample into the third preset neural network model to obtain the etching prediction depth and etching residue prediction data; based on the third training label, the etching prediction depth and the etching residual prediction data, carrying out iterative updating on the third preset neural network model until reaching an iterative termination condition to obtain the etching prediction model; training a fourth preset neural network model according to the training data set to obtain the post-processing prediction model, wherein the training comprises the following steps: Taking the silicon wafer product measurement cleanliness and the silicon wafer product measurement flatness in the training data set as a fourth training label, and taking the etching prediction depth, the etching residue prediction data and cleaning liquid parameters in the process parameters as a fourth training sample; Inputting the fourth training sample into the fourth preset neural network model to obtain the predicted cleanliness and the predicted flatness of the silicon wafer product; And based on the fourth training label, the predicted cleanliness of the silicon chip product and the predicted flatness of the silicon chip product, carrying out iterative updating on the fourth preset neural network model until reaching an iterative termination condition to obtain the post-processing prediction model.
5. 5. A storage medium storing a computer program to be loaded by a processor for performing the steps of a wet process parameter prediction method for solar cell manufacturing according to any one of claims 1 to 4.

Description

Wet process parameter prediction method and system for solar cell manufacturing Technical Field The application relates to the technical field of parameter tuning, in particular to a wet process parameter prediction method and a system for solar cell manufacturing. Background In the field of solar cell manufacturing, a wet process is performed through a plurality of key links such as cleaning, purifying, etching and subsequent chemical treatment of a silicon wafer. The wet etching process reacts with the surface of the silicon wafer through a specific chemical solution, so that redundant materials are precisely removed, and the required circuit pattern and structure are formed. However, minor fluctuations in process parameters in wet processes can have a significant impact on the properties of the final product. Therefore, how to accurately predict and control wet process parameters to ensure the consistency and stability of the solar cell manufacturing process is a current urgent problem to be solved. Disclosure of Invention Aiming at the defects existing in the prior art, the application provides a wet process parameter prediction method and a system for solar cell manufacturing. In a first aspect, the present application provides a method for predicting wet process parameters for solar cell fabrication, including: Dividing the wet process into a pretreatment process, a purification process, an etching process and a post-treatment process according to the process of the wet process; Respectively creating a pretreatment prediction model corresponding to the pretreatment process, a purification prediction model corresponding to the purification process, an etching prediction model corresponding to the etching process and a post-treatment prediction model corresponding to the post-treatment process, and obtaining a moderate function according to the pretreatment prediction model, the purification prediction model, the etching prediction model and the post-treatment prediction model; Obtaining each preset technological parameter in a wet process, wherein the technological parameters comprise a silicon wafer type, a purified liquid parameter, an etching liquid parameter and a cleaning liquid parameter; inputting the moderate value function and each process parameter into a process parameter prediction model, obtaining an initial population according to the process parameters, and guiding the initial population to evolve according to the moderate value function so as to search and obtain optimized process parameters. Optionally, the obtaining the initial population according to the process parameters includes: respectively obtaining the purification liquid parameter, the etching liquid parameter and the limit interval of the cleaning liquid parameter; obtaining the number of individuals of the initial population; And generating initial individuals of the purified liquid parameter, the etching liquid parameter and the cleaning liquid parameter at equal intervals in the limiting interval according to the number of the individuals, so as to obtain an initial population. Optionally, the guiding the initial population to evolve according to the fitness function to search for optimized process parameters includes: Inputting the initial population into the moderate value function to output a finished product prediction quality parameter; And guiding the initial population to evolve according to the predicted quality parameters of the finished product so as to search and obtain optimized technological parameters. Optionally, said inputting said initial population into said fitness function to output a product prediction quality parameter comprises: inputting each initial individual in the initial population to the moderate value function, wherein the initial individual comprises a plurality of purification liquid parameters, a plurality of etching liquid parameters and a plurality of cleaning liquid parameters; Inputting the silicon chip type and the cleaning liquid parameters into the pretreatment prediction model to output the cleanliness prediction data and the cleaning liquid residue prediction data of the silicon chip surface; inputting the purified liquid parameters, the cleanliness prediction data of the surface of the silicon wafer and the cleaning liquid residue prediction data into the purified prediction model to output the predicted purity of the silicon wafer; inputting the predicted purity of the silicon wafer and the etching liquid parameters into the etching prediction model to output etching prediction depth and etching residual prediction data; Inputting the cleaning fluid parameters, the etching predicted depth and the etching residual predicted data into a post-processing prediction model to output the predicted cleanliness of the silicon wafer product and the predicted flatness of the silicon wafer product; And obtaining a finished product prediction quality parameter according to the etching predi