CO2 dry-type cleaning system based on wafer carrying device

Abstract

The invention discloses a CO2 dry-type cleaning system based on a wafer carrying device, which relates to the technical field of semiconductor cleaning and comprises an image collection module, an image analysis module, a pollution migration coefficient acquisition module, an optimal cleaning duration acquisition module and a real-time cleaning duration determination module. The technical problems that the diffusion and removal degree of pollutants on the surface of the wafer cannot be evaluated through CO2 dry cleaning, and the cleaning duration is correspondingly optimized according to the diffusion conditions of the pollutants on the wafer substrates with different pollution degrees are solved, and the pollution transformation coefficient and the pollution migration coefficient of the wafer substrates of different pollution groups are analyzed. The optimal cleaning duration is accurately selected, the problem of low efficiency caused by the adoption of unified cleaning duration in a traditional method is solved, it is ensured that pollutants are completely removed, and migration of the pollutants in the cleaning process is avoided.

Description

Technical Field

[0001] The present invention belongs to the technical field of semiconductor cleaning, and in particular is a CO2 dry cleaning system based on a wafer handling device. Background Art

[0002] With the rapid development of semiconductor manufacturing technology, the cleaning precision requirements for wafer substrates are becoming increasingly higher. Wafer substrates are subjected to multiple handling operations during the manufacturing process, especially under the influence of wafer handling equipment. Surface contaminants on wafer substrates seriously affect the performance and yield of semiconductor devices.

[0003] Currently, traditional cleaning technologies, such as wet cleaning and plasma cleaning, can effectively remove some contaminants, but these technologies have some shortcomings. Wet cleaning relies on water and chemical solvents for cleaning, which can cause secondary contamination during use. Plasma cleaning uses high-energy plasma to remove contaminants, but high-energy plasma can cause irreversible damage to the wafer surface. To overcome these problems, CO2 dry cleaning technology has gradually become a focus of research and application. CO2 dry cleaning technology uses compressed carbon dioxide gas to spray to remove contaminants from the wafer surface.

[0004] However, existing CO2 dry cleaning methods typically use a fixed cleaning time. Because wafer substrates with different levels of contamination require different cleaning times to achieve optimal cleaning results, existing technologies lack an effective mechanism to optimize the cleaning time, resulting in a failure to maximize cleaning results.

[0005] At the same time, the existing CO2 dry cleaning technology generally cleans contaminants on the wafer surface through gas jets. However, due to the dynamic characteristics of the airflow, the diffusion process of the contaminants has not been effectively controlled. Even if the contaminants are cleaned, they may diffuse again to other areas of the surface during the cleaning process, resulting in the migration of contaminants and incomplete removal of contamination on the wafer surface. The existing CO2 dry cleaning cannot evaluate the diffusion and removal degree of contaminants on the wafer surface, and cannot optimize the cleaning time accordingly according to the diffusion of contaminants on wafer substrates with different degrees of contamination, resulting in low cleaning efficiency or incomplete cleaning, affecting the cleaning effect and subsequent processes. Based on this, a CO2 dry cleaning system based on a wafer handling device is proposed. Summary of the Invention

[0006] The purpose of the present invention is to provide a CO2 dry cleaning system based on a wafer handling device, which solves the technical problem that the existing CO2 dry cleaning cannot evaluate the diffusion and removal degree of pollutants on the wafer surface, and optimizes the cleaning time accordingly according to the diffusion of pollutants on wafer substrates with different degrees of contamination.

[0007] A CO2 dry cleaning system based on a wafer handling device, comprising:

[0008] An image collection module divides the wafer substrates into pollution groups and collects images of the wafer substrates in different pollution groups before and after cleaning at different preset cleaning times;

[0009] An image analysis module is used to analyze the difference between the contamination levels in the images of wafer substrates in different contamination groups before and after cleaning at different preset cleaning times, thereby obtaining the contamination conversion coefficients corresponding to the different cleaning times for each contamination group;

[0010] The pollution migration coefficient acquisition module is used to analyze the images of wafer substrates in different pollution groups before and after cleaning at different preset cleaning times, and then obtain the pollution migration coefficient corresponding to each pollution group at different cleaning times;

[0011] The optimal cleaning time acquisition module analyzes the pollution conversion coefficient and pollution migration coefficient corresponding to each pollution group at different cleaning times to obtain the optimal cleaning time corresponding to each pollution component;

[0012] The real-time cleaning duration determination module uses the optimal cleaning duration of the contamination group corresponding to the wafer substrate to be cleaned as the real-time cleaning duration of the wafer substrate to be cleaned and performs a cleaning operation on the wafer substrate to be cleaned for a corresponding duration.

[0013] As a further solution of the present invention, the specific method of grouping the contamination of the wafer substrate is as follows:

[0014] First, the wafer substrate is evenly divided into multiple sub-region blocks, and each sub-region block is numbered in order from left to right and from top to bottom, and then the block numbers corresponding to each sub-region block on the wafer substrate are obtained. The number of blocks marked as contaminated blocks in each wafer substrate is obtained, and the corresponding contamination degree of each wafer substrate is obtained according to the number of contaminated blocks. The wafer substrates are contaminated and grouped according to the contamination degree. The contamination groups are divided into the first contamination group, the second contamination group and the third contamination group. The specific marking method of the contaminated blocks is: measure the roughness of each sub-region block on the surface of the wafer substrate, mark the blocks with a roughness greater than the preset threshold Y3 as contaminated blocks, and do not do any processing on the others.

[0015] As a further solution of the present invention, a specific method for obtaining the contamination degree corresponding to each wafer substrate according to the number of contaminated blocks is as follows:

[0016] The product of the number of contaminated blocks and the area of a single block is taken as the contaminated area, the ratio of the contaminated area to the area of the wafer substrate is taken as the corresponding contamination degree of the wafer substrate, the wafer substrates with a contamination degree greater than the preset threshold value Y1 are taken as the first contamination group, the wafer substrates with a contamination degree less than or equal to the preset threshold value Y1 and greater than or equal to the preset threshold value Y2 are taken as the second contamination group, and the wafer substrates with a contamination degree less than the preset threshold value Y1 are taken as the third contamination group, and the preset threshold value Y1 is greater than the preset threshold value Y2.

[0017] As a further solution of the present invention, the specific method of obtaining the pollution conversion coefficient corresponding to each pollution group at different cleaning times is as follows:

[0018] S1: First, randomly select one of the preset cleaning times as the target cleaning time;

[0019] S2: The absolute value of the difference between the contamination levels of each wafer substrate in the first contamination group before and after cleaning at the target cleaning time is used as the contamination level difference corresponding to each substrate in the first contamination group at the target cleaning time, and the standard deviation of the contamination level difference is used as the contamination conversion coefficient X1 corresponding to the first contamination group at the target cleaning time;

[0020] S3: Repeating the same analysis method used to analyze the before-and-after images of the wafer substrates in the first contamination group at different preset cleaning times, analyze the before-and-after images of the wafer substrates in the second and third contamination groups at different preset cleaning times, thereby obtaining the contamination conversion coefficients X2 and X3 corresponding to the second and third contamination groups at the target cleaning times, respectively;

[0021] S4: Repeat steps S1-S3 to obtain the pollution conversion coefficients X1t, X2t and X3t corresponding to different cleaning times for each pollution group, where t refers to different cleaning times.

[0022] As a further solution of the present invention, the specific method for obtaining the pollution migration coefficient corresponding to each pollution group at different cleaning times is as follows:

[0023] S01: First, randomly select one of the preset cleaning times as the cleaning time for analysis;

[0024] S02: Selecting the first contamination group as the analysis contamination group, and selecting a wafer substrate from the analysis contamination group as the analysis substrate;

[0025] The center point of the analysis substrate is used as the reference point and a two-dimensional coordinate system is set on the analysis substrate with it as the origin. The calibration point coordinates of each contaminated block of the analysis substrate are obtained from the image of the analysis substrate before cleaning. The distance LAe between the calibration points of each contaminated block in the image of the analysis substrate before cleaning and the reference point are calculated, and the discrete value EA of the distance LAe is calculated, where e refers to each contaminated block in the image of the analysis substrate before cleaning. The calibration point coordinates of each contaminated block of the analysis substrate are then obtained from the image of the analysis substrate after cleaning. The distance LBr between the calibration points of each contaminated block in the image of the analysis substrate after cleaning and the reference point is obtained by using a distance calculation formula, and the discrete value EB of the distance LBr is obtained, where r refers to each contaminated block in the image of the analysis substrate after cleaning. The absolute value of the difference between the pollution dispersion coefficient EA of the image of the analysis substrate before cleaning and the pollution dispersion coefficient EB of the image after cleaning is used as the pollution dispersion coefficient ER1 of the analysis substrate under the analysis cleaning time.

[0026] S03: Repeat step S02 to analyze the images of each wafer substrate in the first contamination group before and after cleaning one by one, thereby obtaining the pollution dispersion coefficient of each wafer substrate in the first contamination group under the analyzed cleaning time, and taking the average of the maximum and minimum values of the pollution dispersion coefficient as the pollution migration coefficient QY11 of the first contamination group under the analyzed cleaning time;

[0027] S04: Repeat steps S01-S03 to analyze the pre- and post-cleaning images of wafer substrates of different pollution groups at different preset cleaning times, and then obtain the pollution migration coefficients QY1t, QY2t, and QY3t corresponding to each pollution group at different cleaning times.

[0028] As a further solution of the present invention, the specific method for obtaining the optimal cleaning time corresponding to each pollution component is:

[0029] Obtaining the cleaning time TA corresponding to the minimum pollution conversion coefficient among the pollution conversion coefficients of the first pollution group under different cleaning time periods, and simultaneously obtaining the cleaning time TB corresponding to the minimum pollution migration coefficient among the pollution migration coefficients of the first pollution group under different cleaning time periods;

[0030] When TB is greater than or equal to TA, the time corresponding to TB is used as the optimal cleaning time ZJ1 corresponding to the first pollution group; when TB is less than TA, the time corresponding to TA is used as the optimal cleaning time ZJ1 corresponding to the first pollution group; the same method is used to analyze the pollution conversion coefficient and pollution migration coefficient corresponding to each pollution group at different cleaning times, and then obtain the optimal cleaning time ZJ1, ZJ2 and ZJ3 corresponding to each pollution group.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] (1) The present invention uses a pollution conversion coefficient to reflect the change in pollution degree of wafer substrates with different pollution groups before and after cleaning at different cleaning times. The smaller the pollution conversion coefficient, the better the corresponding cleaning effect, and the larger the pollution conversion coefficient, the incomplete removal of pollutants.

[0033] (2) The present invention uses the contamination migration coefficient to reflect the distribution change of contaminants in the substrate surface space before and after cleaning of wafer substrates of different contamination groups. The smaller the contamination migration coefficient, the less contaminant migration or the complete removal, and the better the cleaning uniformity. The larger the contamination migration coefficient, the more serious the contaminant migration.

[0034] (3) The present invention accurately selects the optimal cleaning time by analyzing the pollution conversion coefficient and pollution migration coefficient of wafer substrates in different pollution groups, thereby avoiding the inefficiency problem of using a unified cleaning time in traditional methods. The calculation of the pollution conversion coefficient and the pollution migration coefficient can effectively optimize the cleaning process, ensure the complete removal of pollutants, ensure the maximization of the cleaning effect, and avoid the problems of pollutant migration and secondary pollution. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 Schematic diagram of the system framework structure of the present invention;

[0036] Figure 2 It is a schematic structural diagram of the process of grouping contamination of wafer substrates according to the present invention. DETAILED DESCRIPTION

[0037] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0038] Example 1: Please refer to Figure 1-Figure 2 , the present application provides a CO2 dry cleaning system based on a wafer handling device, comprising;

[0039] An image collection module is configured to set multiple preset cleaning times and multiple wafer substrates of the same specifications, group the multiple wafer substrates of the same specifications into pollution groups, perform cleaning operations of different preset cleaning times on the wafer substrates in different pollution groups, and acquire images of the wafer substrates before and after cleaning, thereby completing the collection of images of the wafer substrates in different pollution groups before and after cleaning at different preset cleaning times;

[0040] The specific method of contamination grouping for multiple wafer substrates of the same specifications is as follows:

[0041] First, the wafer substrate is evenly divided into multiple sub-region blocks, and each sub-region block is numbered in order from left to right and from top to bottom. Then, the block number i corresponding to each sub-region block on the wafer substrate is obtained, where i refers to a different sub-region block, i = 1, 2, ..., a, a refers to the total number of sub-region blocks in the wafer substrate, a is a positive integer, and a satisfies a ≥ 2;

[0042] Obtaining the number of blocks marked as contaminated blocks in each wafer substrate, and obtaining the contamination degree corresponding to each wafer substrate according to the number of contaminated blocks, and grouping the wafer substrates according to the contamination degree, and the contamination groups are divided into a first contamination group, a second contamination group, and a third contamination group;

[0043] The specific method for marking contaminated blocks is as follows: since contaminants usually affect the surface roughness, as they often cause surface unevenness and increase surface roughness, the roughness of each sub-region block on the surface of the wafer substrate is measured using equipment such as an atomic force microscope (AFM), an optical profilometer, or a scanning electron microscope (SEM). Blocks with a roughness greater than a preset threshold value Y3 are marked as contaminated blocks, and the others are not processed in any way. The specific preset threshold value Y3 is formulated by relevant personnel based on actual needs. The above technologies for obtaining the roughness of each sub-region block on the substrate surface are all existing and mature technologies and are therefore not described in detail here.

[0044] The specific method of obtaining the contamination degree corresponding to each wafer substrate according to the number of contaminated blocks is to multiply the number of contaminated blocks by the area of each block as the contaminated area, and to use the ratio of the contaminated area to the area of the wafer substrate as the contamination degree corresponding to the wafer substrate;

[0045] The block area is the ratio of the wafer substrate area to the total number of sub-region blocks a, and the area of each sub-region block is equal;

[0046] The wafer substrates with a contamination level greater than a preset threshold value Y1 are classified as a first contamination group, the wafer substrates with a contamination level less than or equal to the preset threshold value Y1 and greater than or equal to the preset threshold value Y2 are classified as a second contamination group, and the wafer substrates with a contamination level less than the preset threshold value Y1 are classified as a third contamination group. The specific values of the preset threshold value Y1 and the preset threshold value Y2 are formulated by relevant personnel based on actual needs, and the preset threshold value Y1 is greater than the preset threshold value Y2.

[0047] The pollution level of the first pollution group is greater than that of the second pollution group, and the pollution level of the second pollution group is greater than that of the third pollution group;

[0048] It should be noted that multiple wafer substrates are randomly selected. In addition to the same specifications, the contamination parts are also random. At the same time, the contamination level of the wafer substrates in the first contamination group is greater than that of the wafer substrates in the second contamination group, and the contamination level of the wafer substrates in the second contamination group is greater than that of the wafer substrates in the third contamination group. At the same time, it is assumed that the number of wafer substrates in the first contamination group, the second contamination group and the third contamination group is greater than 2.

[0049] An image analysis module is used to analyze the images of wafer substrates in different pollution groups before and after cleaning at different preset cleaning times, and then obtain the pollution conversion coefficients corresponding to each pollution group at different cleaning times;

[0050] The difference between the contamination levels of wafer substrates in different contamination groups before and after cleaning at different preset cleaning times is analyzed to obtain the contamination conversion coefficients corresponding to the wafer substrates in different contamination groups at different preset cleaning times. The specific method is as follows:

[0051] S1: First, analyze the pre- and post-cleaning images of the wafer substrates in the first contamination group at different preset cleaning times; first, randomly select one of the different preset cleaning times as the target cleaning time;

[0052] S2: Obtain the absolute value of the difference in contamination level before and after cleaning for each wafer substrate in the first contamination group at the target cleaning time, and use the absolute value as the contamination level difference Wj corresponding to each substrate in the first contamination group at the target cleaning time, where j refers to different wafer substrates in the first contamination group, j=1, 2, ..., b, b refers to the total number of wafer substrates in the first contamination group, b is a positive integer, and b satisfies b>2;

[0053] The standard deviation of the pollution degree difference Wj corresponding to each wafer substrate in the first pollution group under the target cleaning time is used as the pollution conversion coefficient X1 corresponding to the first pollution group under the target cleaning time;

[0054] S3: Repeating the same analysis method used to analyze the before-and-after images of the wafer substrates in the first contamination group at different preset cleaning times, analyze the before-and-after images of the wafer substrates in the second and third contamination groups at different preset cleaning times, thereby obtaining the contamination conversion coefficients X2 and X3 corresponding to the second and third contamination groups at the target cleaning times, respectively;

[0055] S4: Repeat steps S1-S3 to obtain the pollution conversion coefficients X1t, X2t, and X3t corresponding to different cleaning times for each pollution group, where t represents different cleaning times, t=1, 2, ..., c, c represents the total number of cleaning times, c is a positive integer, and c satisfies c≥2;

[0056] The pollution conversion coefficient reflects the change in pollution degree of wafer substrates with different pollution groups before and after cleaning at different cleaning times. The smaller the pollution conversion coefficient, the better the corresponding cleaning effect, and the larger the pollution conversion coefficient, the incomplete removal of pollutants.

[0057] The pollution migration coefficient acquisition module is used to analyze the images of wafer substrates in different pollution groups before and after cleaning at different preset cleaning times, and then obtain the pollution migration coefficient corresponding to each pollution group at different cleaning times. The specific method is as follows:

[0058] S01: First, randomly select one of the preset cleaning times as the cleaning time for analysis;

[0059] S02: Selecting the first contamination group as the analysis contamination group, and selecting a wafer substrate from the analysis contamination group as the analysis substrate;

[0060] A two-dimensional coordinate system is set on the analysis substrate with the center point of the analysis substrate as the origin, and the center point of the analysis substrate is used as the reference point. The center points of each sub-region block of the analysis substrate are used as calibration points of each sub-region block, and the calibration point coordinates Be(BXe, BYe) of each contaminated block of the analysis substrate are obtained from the image of the analysis substrate before cleaning. The distance LAe between the calibration points of each contaminated block in the image of the analysis substrate before cleaning and the reference point is calculated using a distance calculation formula, and the discrete value EA of the distance LAe is obtained using a discrete value calculation formula, where e refers to each contaminated block in the image of the analysis substrate before cleaning;

[0061] The specific method for calculating and obtaining the distance LAe between the calibration points and the reference points of each contaminated block in the image of the analyzed substrate before cleaning is as follows:

[0062] Distance calculation formula: Calculate and obtain the distance LAe between the calibration points and the reference points of each contaminated area in the image of the substrate before cleaning;

[0063] Then, the coordinates of the calibration points Cr(CXr, CYr) of each contaminated area of the analysis substrate are obtained from the image of the analysis substrate after cleaning. The distance LBr between the calibration points of each contaminated area in the image of the analysis substrate after cleaning and the reference point is calculated using the distance calculation formula. The discrete value EB of the distance LBr is obtained using the discrete value calculation formula, where r refers to each contaminated area in the image of the analysis substrate after cleaning.

[0064] The specific method for calculating and analyzing the distance LBr between the calibration points and the reference points of each contaminated area in the image after substrate cleaning is as follows:

[0065] Distance calculation formula: Calculate and obtain the distance LB between the calibration points of each contaminated area in the image after the substrate is cleaned and the reference point;

[0066] The absolute value of the difference between the pollution dispersion coefficient EA of the image of the analysis substrate before cleaning and the pollution dispersion coefficient EB of the image after cleaning is used as the pollution dispersion coefficient ER1 of the analysis substrate under the analysis cleaning time;

[0067] S03: Repeat step S02 to analyze the images before and after cleaning of each wafer substrate in the first contamination group one by one, thereby obtaining the pollution dispersion coefficient ERj of each wafer substrate in the first contamination group at the analyzed cleaning time, and taking the average of the maximum and minimum pollution dispersion coefficients of each wafer substrate in the first contamination group as the pollution migration coefficient QY11 of the first contamination group at the analyzed cleaning time;

[0068] S04: Repeat steps S01-S03 to analyze the images of wafer substrates in different pollution groups before and after cleaning at different preset cleaning times, thereby obtaining the pollution migration coefficients QY1t, QY2t, and QY3t corresponding to each pollution group at different cleaning times;

[0069] The contamination migration coefficient reflects the distribution changes of contaminants in the substrate surface space before and after cleaning for wafer substrates with different contamination groups. The smaller the contamination migration coefficient, the less contaminant migration or the complete removal, and the better the cleaning uniformity. The larger the contamination migration coefficient, the more serious the contaminant migration.

[0070] The optimal cleaning time acquisition module analyzes the pollution conversion coefficient and pollution migration coefficient corresponding to each pollution group at different cleaning times to obtain the optimal cleaning time corresponding to each pollution component. The specific method is as follows:

[0071] Obtaining the cleaning time TA corresponding to the minimum pollution conversion coefficient among the pollution conversion coefficients of the first pollution group under different cleaning time periods, and simultaneously obtaining the cleaning time TB corresponding to the minimum pollution migration coefficient among the pollution migration coefficients of the first pollution group under different cleaning time periods;

[0072] When TB is greater than or equal to TA, the duration corresponding to TB is used as the optimal cleaning duration ZJ1 corresponding to the first pollution group; when TB is less than TA, the duration corresponding to TA is used as the optimal cleaning duration ZJ1 corresponding to the first pollution group;

[0073] The same method is used to analyze the pollution conversion coefficient and pollution migration coefficient corresponding to each pollution group at different cleaning times, and then the optimal cleaning time ZJ1, ZJ2 and ZJ3 corresponding to each pollution group are obtained.

[0074] The real-time cleaning duration determination module determines and outputs the corresponding real-time cleaning duration based on the optimal cleaning duration of the contamination group corresponding to the wafer substrate to be cleaned, and controls the cleaning system of the wafer handling device to perform a cleaning operation of the corresponding duration on the wafer substrate to be cleaned according to the real-time cleaning duration. The specific method is as follows;

[0075] That is, first, the contamination group corresponding to the wafer substrate to be cleaned is determined based on the image of the wafer substrate to be cleaned before cleaning, and the optimal cleaning time of the contamination group corresponding to the wafer substrate to be cleaned is used as the real-time cleaning time of the wafer substrate to be cleaned. According to the real-time cleaning time, the cleaning system of the wafer transport device is controlled to perform a cleaning operation on the wafer substrate to be cleaned for a corresponding time.

[0076] By analyzing the pollution conversion coefficient and pollution migration coefficient of wafer substrates in different pollution groups, the optimal cleaning time is accurately selected, avoiding the inefficiency of using a unified cleaning time in traditional methods. It can provide personalized cleaning solutions for wafer substrates with different pollution levels. The calculation of the pollution conversion coefficient and pollution migration coefficient can effectively optimize the cleaning process, ensure the complete removal of pollutants, ensure the maximization of cleaning effects, and avoid the problems of pollutant migration and secondary pollution.

[0077] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters and thresholds in the formulas are set by technicians in this field according to actual conditions.

[0078] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A CO2 dry cleaning system based on a wafer handling device, characterized in that: include; An image collection module divides the wafer substrates into pollution groups and collects images of the wafer substrates in different pollution groups before and after cleaning at different preset cleaning times; An image analysis module is used to analyze the difference between the contamination levels in the images of wafer substrates in different contamination groups before and after cleaning at different preset cleaning times, thereby obtaining the contamination conversion coefficients corresponding to the different cleaning times for each contamination group; The pollution migration coefficient acquisition module is used to analyze the images of wafer substrates in different pollution groups before and after cleaning at different preset cleaning times, and then obtain the pollution migration coefficient corresponding to each pollution group at different cleaning times; The optimal cleaning time acquisition module analyzes the pollution conversion coefficient and pollution migration coefficient corresponding to each pollution group at different cleaning times to obtain the optimal cleaning time corresponding to each pollution component; The real-time cleaning duration determination module uses the optimal cleaning duration of the contamination group corresponding to the wafer substrate to be cleaned as the real-time cleaning duration of the wafer substrate to be cleaned and performs a cleaning operation on the wafer substrate to be cleaned for a corresponding duration.

2. A CO2 dry cleaning system based on a wafer handling device according to claim 1, characterized in that: The specific method of grouping contamination of wafer substrates is as follows: First, the wafer substrate is evenly divided into multiple sub-region blocks, and each sub-region block is numbered in order from left to right and from top to bottom, so as to obtain the block numbers corresponding to each sub-region block on the wafer substrate, obtain the number of blocks marked as contaminated blocks in each wafer substrate, and obtain the corresponding contamination degree of each wafer substrate according to the number of contaminated blocks, and the wafer substrates with a contamination degree greater than a preset threshold value Y1 are regarded as the first contamination group, the wafer substrates with a contamination degree less than or equal to the preset threshold value Y1 and greater than or equal to the preset threshold value Y2 are regarded as the second contamination group, and the wafer substrates with a contamination degree less than the preset threshold value Y1 are regarded as the third contamination group, and the preset threshold value Y1 is greater than the preset threshold value Y2.

3. A CO2 dry cleaning system based on a wafer handling device according to claim 2, characterized in that: The specific method for obtaining the contamination degree corresponding to each wafer substrate according to the number of contaminated blocks is as follows: The product of the number of contaminated blocks and the area of a single block is taken as the contaminated area, and the ratio of the contaminated area to the wafer substrate area is taken as the corresponding contamination degree of the wafer substrate.

4. A CO2 dry cleaning system based on a wafer handling device according to claim 2, characterized in that: The specific method of obtaining the pollution conversion coefficient corresponding to each pollution group at different cleaning times is as follows: S1: First, randomly select one of the preset cleaning times as the target cleaning time; S2: The absolute value of the difference between the contamination levels of each wafer substrate in the first contamination group before and after cleaning at the target cleaning time is used as the contamination level difference corresponding to each substrate in the first contamination group at the target cleaning time, and the standard deviation of the contamination level difference is used as the contamination conversion coefficient X1 corresponding to the first contamination group at the target cleaning time; S3: Repeating the same analysis method used to analyze the before-and-after images of the wafer substrates in the first contamination group at different preset cleaning times, analyze the before-and-after images of the wafer substrates in the second and third contamination groups at different preset cleaning times, thereby obtaining the contamination conversion coefficients X2 and X3 corresponding to the second and third contamination groups at the target cleaning times, respectively; S4: Repeat steps S1-S3 to obtain the pollution conversion coefficients X1t, X2t and X3t corresponding to different cleaning times for each pollution group, where t refers to different cleaning times.

5. A CO2 dry cleaning system based on a wafer handling device according to claim 4, characterized in that: The specific method for obtaining the pollution migration coefficient corresponding to each pollution group at different cleaning times is as follows: S01: First, randomly select one of the preset cleaning times as the cleaning time for analysis; S02: Selecting the first contamination group as the analysis contamination group, and selecting a wafer substrate from the analysis contamination group as the analysis substrate; The center point of the analysis substrate is used as the reference point and a two-dimensional coordinate system is set on the analysis substrate with it as the origin. The calibration point coordinates of each contaminated block of the analysis substrate are obtained from the image of the analysis substrate before cleaning. The distance LAe between the calibration points of each contaminated block in the image of the analysis substrate before cleaning and the reference point are calculated, and the discrete value EA of the distance LAe is calculated, where e refers to each contaminated block in the image of the analysis substrate before cleaning. The calibration point coordinates of each contaminated block of the analysis substrate are then obtained from the image of the analysis substrate after cleaning. The distance LBr between the calibration points of each contaminated block in the image of the analysis substrate after cleaning and the reference point is obtained by using a distance calculation formula, and the discrete value EB of the distance LBr is obtained, where r refers to each contaminated block in the image of the analysis substrate after cleaning. The absolute value of the difference between the pollution dispersion coefficient EA of the image of the analysis substrate before cleaning and the pollution dispersion coefficient EB of the image after cleaning is used as the pollution dispersion coefficient ER1 of the analysis substrate under the analysis cleaning time. S03: Repeat step S02 to analyze the images of each wafer substrate in the first contamination group before and after cleaning one by one, thereby obtaining the pollution dispersion coefficient of each wafer substrate in the first contamination group under the analyzed cleaning time, and taking the average of the maximum and minimum values of the pollution dispersion coefficient as the pollution migration coefficient QY11 of the first contamination group under the analyzed cleaning time; S04: Repeat steps S01-S03 to analyze the pre- and post-cleaning images of wafer substrates of different pollution groups at different preset cleaning times, and then obtain the pollution migration coefficients QY1t, QY2t, and QY3t corresponding to each pollution group at different cleaning times.

6. A CO2 dry cleaning system based on a wafer handling device according to claim 5, characterized in that: The specific method for obtaining the optimal cleaning time corresponding to each pollution component is: Obtaining the cleaning time TA corresponding to the minimum pollution conversion coefficient among the pollution conversion coefficients of the first pollution group under different cleaning time periods, and simultaneously obtaining the cleaning time TB corresponding to the minimum pollution migration coefficient among the pollution migration coefficients of the first pollution group under different cleaning time periods; When TB is greater than or equal to TA, the time corresponding to TB is used as the optimal cleaning time ZJ1 corresponding to the first pollution group; when TB is less than TA, the time corresponding to TA is used as the optimal cleaning time ZJ1 corresponding to the first pollution group; the same method is used to analyze the pollution conversion coefficient and pollution migration coefficient corresponding to each pollution group at different cleaning times, and then obtain the optimal cleaning time ZJ1, ZJ2 and ZJ3 corresponding to each pollution group.

7. The CO2 dry cleaning system based on a wafer handling device according to claim 2, characterized in that: The specific marking method for polluted blocks is: The roughness of each sub-region block on the surface of the wafer substrate is measured, and the blocks with a roughness greater than a preset threshold value Y3 are marked as contaminated blocks, and no processing is performed on the others.

8. The CO2 dry cleaning system based on a wafer handling device according to claim 3, characterized in that: The pollution level of the first pollution group is greater than that of the second pollution group, and the pollution level of the second pollution group is greater than that of the third pollution group.

Images