A wafer surface cleaning monitoring method and system

Abstract

The present invention provides a wafer surface cleaning monitoring method and system, and relates to the field of wafer cleaning monitoring technology. The method comprises: determining cleaning control parameters; synchronously generating a monitoring parameter sequence; obtaining cleaning monitoring data; based on the cleaning monitoring data, performing trajectory deviation analysis and prediction through a trajectory deviation operation module to obtain cleaning deviation information; performing cleaning probability distribution calculation based on the cleaning deviation information and cleaning control parameters; performing monitoring data feedback based on the cleaning probability distribution calculation result, and generating monitoring control guidance information, thereby solving the technical problem that the existing wafer cleaning monitoring methods mostly monitor the position of the wafer and the abnormality of the cleaning equipment, lack analysis and correction of the cleaning trajectory, and thus lead to poor monitoring effect, thereby achieving the technical effect of improving the wafer surface cleaning monitoring effect, performing feedback correction in time, and improving the cleaning quality.

Description

Technical Field

[0001] The present disclosure relates to the technical field of wafer cleaning monitoring, and in particular to a wafer surface cleaning monitoring method and system. Background Art

[0002] Wafers are important components for manufacturing integrated circuits. The electronic components and circuit connections in integrated circuits are relatively fine. Therefore, if there are dust, metal particles and other pollutants on the surface of the wafers used for integrated circuit production, it will cause damage to the internal circuits of the integrated circuits and affect their use. Therefore, it is of great significance to monitor the wafer surface cleaning process.

[0003] Existing wafer cleaning monitoring methods have a technical problem that most of them monitor the position of the wafer and the abnormalities of the cleaning equipment, but lack analysis and correction of the cleaning trajectory, which leads to poor monitoring results. Summary of the invention

[0004] The present invention provides a wafer surface cleaning monitoring method and system to solve the technical problem that the existing wafer cleaning monitoring methods mostly monitor the position of the wafer and the abnormality of the cleaning equipment, lack analysis and correction of the cleaning trajectory, and thus lead to poor monitoring effect.

[0005] According to a first aspect of the present disclosure, a wafer surface cleaning monitoring method is provided, including: identifying cleaning targets on the wafer surface, locating distribution positions, and determining cleaning control parameters; using the cleaning control parameters to control parameters of cleaning enforcement equipment, and synchronously generating a monitoring parameter sequence; collecting monitoring data of the wafer cleaning process, extracting cleaning trajectory tracking parameters from the monitoring data, and obtaining cleaning monitoring data; constructing a trajectory deviation operation module using the monitoring parameter sequence and cleaning control parameters, and based on the cleaning monitoring data, performing trajectory deviation analysis and prediction through the trajectory deviation operation module to obtain cleaning deviation information; calculating the cleaning probability distribution based on the cleaning deviation information and cleaning control parameters; providing monitoring data feedback based on the cleaning probability distribution calculation result, and producing monitoring control guidance information.

[0006] According to a second aspect of the present disclosure, a wafer surface cleaning monitoring system is provided, comprising: a cleaning target identification unit, the cleaning target identification unit is used to identify the cleaning target on the wafer surface, locate the distribution position, and determine the cleaning control parameters; a parameter control unit, the parameter control unit is used to use the cleaning control parameters to control the parameters of the cleaning enforcement equipment, and synchronously generate a monitoring parameter sequence; a cleaning trajectory tracking unit, the cleaning trajectory tracking unit is used to collect monitoring data of the wafer cleaning process, extract cleaning trajectory tracking parameters from the monitoring data, and obtain cleaning monitoring data; a trajectory deviation analysis unit, the trajectory deviation analysis unit is used to construct a trajectory deviation operation module using the monitoring parameter sequence and the cleaning control parameters, and based on the cleaning monitoring data, perform trajectory deviation analysis and prediction through the trajectory deviation operation module to obtain cleaning deviation information; a cleaning probability distribution calculation unit, the cleaning probability distribution calculation unit is used to calculate the cleaning probability distribution based on the cleaning deviation information and the cleaning control parameters; a monitoring data feedback unit, the monitoring data feedback unit is used to feedback the monitoring data based on the cleaning probability distribution calculation result, and produce monitoring control guidance information.

[0007] According to the present disclosure, one or more technical solutions are adopted to achieve the following beneficial effects:

[0008] Cleaning targets are identified on the wafer surface, their distribution positions are located, cleaning control parameters are determined, cleaning enforcement equipment is controlled using cleaning control parameters, and a monitoring parameter sequence is generated simultaneously to collect monitoring data of the wafer cleaning process. Cleaning trajectory tracking parameters are extracted from the monitoring data to obtain cleaning monitoring data. A trajectory deviation operation module is constructed using the monitoring parameter sequence and cleaning control parameters. Based on the cleaning monitoring data, trajectory deviation analysis and prediction are performed through the trajectory deviation operation module to obtain cleaning deviation information. Cleaning probability distribution calculation is performed based on the cleaning deviation information and cleaning control parameters. Monitoring data feedback is performed based on the cleaning probability distribution calculation results, and monitoring and control guidance information is produced. By performing deviation analysis on the cleaning trajectory and then performing cleaning probability analysis, the technical effect of improving wafer surface cleaning monitoring effect, timely feedback correction, and improving cleaning quality is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the prior art description are briefly introduced below. The drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without creative work.

[0010] Figure 1A schematic diagram of a process flow of a wafer surface cleaning monitoring method provided by an embodiment of the present invention;

[0011] Figure 2 A schematic structural diagram of a wafer surface cleaning monitoring system provided by an embodiment of the present invention.

[0012] Explanation of the reference numerals: cleaning target identification unit 11 , parameter control unit 12 , cleaning trajectory tracking unit 13 , trajectory deviation analysis unit 14 , cleaning probability distribution calculation unit 15 , monitoring data feedback unit 16 . DETAILED DESCRIPTION

[0013] In order to make the purpose, technical solution and advantages of the present invention more obvious, the exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments of the present invention, and it should be understood that the present invention is not limited to the exemplary embodiments described here.

[0014] The terms used in the specification are used to describe the embodiments, rather than to limit the present invention. As used in the specification, the singular terms "a", "an" and "the" are intended to also include plural forms, unless the context clearly indicates otherwise. When used in the specification, the terms "comprise" and / or "include" specify the presence of steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other steps, operations, elements, components and / or groups thereof.

[0015] Unless otherwise defined, all terms (including technical and scientific terms) used in this specification shall have the same meaning as commonly understood by those skilled in the art to which the present invention belongs. Terms, such as those defined in commonly used dictionaries, should not be interpreted in an idealized or overly formal sense unless explicitly defined herein. Throughout the specification, the same reference numerals represent the same elements.

[0016] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for display, data for analysis, etc.) involved in the present invention are all information and data authorized by the user or fully authorized by all parties.

[0017] Embodiment 1

[0018] Figure 1 A wafer surface cleaning monitoring method is provided in an embodiment of the present application, and the method comprises:

[0019] Identify cleaning targets on the wafer surface, locate distribution positions, and determine cleaning control parameters;

[0020] In a preferred embodiment, it also includes:

[0021] Obtain wafer surface cleaning target recognition results, including cleaning target location and cleaning target type; cluster the cleaning target location based on the cleaning target type, and use the clustering result to generate a label to identify the cleaning target location; determine the cleaning parameters according to the type clustering result of the cleaning target location; determine the cleaning control trajectory of the cleaning target location based on the identification, establish a correlation between the cleaning control trajectory and the cleaning parameters, and obtain the cleaning control parameters.

[0022] The cleaning target is identified on the wafer surface. The cleaning target is the contaminant on the wafer surface, such as metal, organic matter and granular particle residues. Specifically, the cleaning target identification model based on the neural network can be constructed and trained using the historical wafer surface image and the corresponding historical contaminant type and historical contaminant distribution position. The input of the cleaning target identification model is the wafer surface image, and the output is the contaminant type and contaminant distribution position. Then, the wafer surface is imaged, and the image acquisition result is input into the cleaning target identification model to obtain the corresponding contaminant type and contaminant position. The contaminant type is used as the cleaning target type, and the contaminant position is used as the cleaning target location. The cleaning target identification result includes the cleaning target location and the cleaning target type. Further, the cleaning target location is clustered based on the cleaning target type, that is, the same type of contaminants are clustered into one type of clustering result, and multiple types of clustering results can be obtained to form a clustering result. The cleaning target location is identified by generating a label using the clustering result. In short, the cleaning target location of different types of clustering results is marked with different labels to facilitate subsequent cleaning control.

[0023] According to the type clustering results of cleaning target positioning, the cleaning parameters are determined. The cleaning parameters include cleaning water pressure, water flow rate and other cleaning parameters corresponding to the cleaning enforcement equipment used to clean the wafer surface (equipment for cleaning the wafer surface, including a nozzle, and the cleaning liquid flowing out of the nozzle is used to clean the wafer surface). In other words, different types of cleaning targets have different adsorption forces on the wafer surface, and require different cleaning water pressures, water flow rates, etc. Specifically, historical wafer surface cleaning record data can be collected and acquired, and the historical cleaning parameters corresponding to different types of cleaning targets can be extracted from them to form a cleaning parameter database. Then, based on the type clustering results of cleaning target positioning, the types of cleaning targets are matched in the cleaning parameter database to obtain the corresponding cleaning parameters.

[0024] The cleaning control trajectory is determined based on the identification of the cleaning target. In simple terms, the cleaning target is identified by generating a label using the clustering result. The cleaning control parameters corresponding to each type of clustering result are the same. The same type of cleaning target is connected at the shortest distance according to the identification, and multiple trajectories of different types of cleaning target are obtained. The multiple trajectories are connected together as the cleaning control trajectory. At the same time, each trajectory in the cleaning control trajectory has a corresponding cleaning parameter, thereby establishing an association between the cleaning control trajectory and the cleaning parameter. The cleaning control trajectory and the association with the cleaning parameter are used as the cleaning control parameter, so that the cleaning trajectory of the nozzle of the cleaning enforcement equipment is controlled while the nozzle is controlled to spray cleaning liquid according to the corresponding cleaning parameter to clean the wafer surface.

[0025] The cleaning control parameters are used to control the parameters of the cleaning law enforcement equipment and to simultaneously generate a monitoring parameter sequence;

[0026] Using the cleaning control parameters to control the parameters of the cleaning enforcement equipment is to control the nozzle of the cleaning enforcement equipment to move according to the cleaning control trajectory in the cleaning control parameters. At the same time, according to the cleaning control trajectory and the correlation relationship with the cleaning parameters, the nozzle is controlled to spray cleaning liquid to clean the wafer surface according to the cleaning parameters during the movement, and a monitoring parameter sequence is generated synchronously. The monitoring parameter sequence refers to a sequence of parameters that need to be monitored after the cleaning enforcement equipment is parameter controlled and the cleaning enforcement equipment is cleaned according to the corresponding cleaning control parameters in different time periods. In simple terms, the above steps obtain multiple trajectories for different types of cleaning target positioning, connect the multiple trajectories together as the cleaning control trajectory, and at the same time, each trajectory in the cleaning control trajectory has a corresponding cleaning parameter, thereby establishing the correlation relationship between the cleaning control trajectory and the cleaning parameters. Then, according to the multiple trajectories, one type of cleaning target is positioned as a monitoring target, and the monitoring target time sequence can be obtained by arranging them in sequence, and then the range of the multiple trajectories corresponding to each type of cleaning target positioning is used as the monitoring target range. The monitoring target time sequence and the monitoring target range constitute the monitoring parameter sequence.

[0027] Collect monitoring data of the wafer cleaning process, extract cleaning trajectory tracking parameters from the monitoring data, and obtain cleaning monitoring data;

[0028] In a preferred embodiment, it also includes:

[0029] The wafer cleaning process is imaged by monitoring equipment to obtain image acquisition information, including the acquisition point position; the aiming nozzle parameter information of the monitoring equipment is obtained, wherein the aiming nozzle parameter information includes the nozzle position, the nozzle operating parameters, and the cleaning target; the pre-aiming distance and the center point are determined according to the acquisition position, the nozzle position, and the cleaning target; the spray contact area is identified through the image acquisition information to determine the cleaning monitoring area, and the pre-aiming distance, the center point, and the cleaning monitoring area are used as the cleaning monitoring data.

[0030] Collect monitoring data of the wafer cleaning process, extract cleaning trajectory tracking parameters from the monitoring data, and obtain cleaning monitoring data. The specific process is as follows:

[0031] The wafer cleaning process is imaged by monitoring equipment (image acquisition equipment, such as cameras, cameras, CCD image sensors, etc.) to obtain image acquisition information. The image acquisition information includes the location of the acquisition point, that is, the location of the pollutants, that is, the cleaning target, during the wafer cleaning process. The aiming nozzle parameter information of the monitoring equipment is obtained. The aiming nozzle parameter information refers to the nozzle parameter information of the nozzle on the cleaning enforcement equipment. That is, the surface of the wafer is cleaned by spraying cleaning liquid through the nozzle. The aiming nozzle parameter information includes the nozzle position, nozzle working parameters, and cleaning target. The nozzle position is the location of the nozzle, and the nozzle working parameters are the corresponding cleaning water pressure and other parameters. The cleaning target is the pollutants on the wafer being cleaned by the nozzle, which can be directly acquired through the monitoring equipment. According to the acquisition position, nozzle position, and cleaning target, the pre-aiming distance and center point are determined. The pre-aiming distance refers to the distance from the nozzle position to the acquisition position, and the center point refers to the nozzle center point of the nozzle position. The nozzle is actually a mechanical arm that can move, so the center point is also the center point of the nozzle length.

[0032] The spray contact area is identified through image acquisition information to determine the cleaning monitoring area. The cleaning monitoring area can be understood as the contact area between the cleaning liquid sprayed from the nozzle and the wafer surface. The edge of the spray contact surface can be identified based on the image acquisition information, and the contact area is calculated as the cleaning monitoring area based on the edge identification result. The preview distance, center point, and cleaning monitoring area are used as the cleaning monitoring data for subsequent cleaning deviation analysis, thereby improving the cleaning quality.

[0033] A trajectory deviation calculation module is constructed using the monitoring parameter sequence and the cleaning control parameters. Based on the cleaning monitoring data, the trajectory deviation calculation module is used to perform trajectory deviation analysis and prediction to obtain cleaning deviation information.

[0034] In a preferred embodiment, it also includes:

[0035] According to the cleaning control parameters, the control preset trajectory is determined; according to the monitoring parameter sequence, the monitoring target timing and the monitoring target range are determined; according to the monitoring target timing, the corresponding cleaning control parameters are determined, and based on the cleaning control parameters, the cleaning nozzle width and the nozzle working parameters are extracted, the cleaning coverage range is predicted, and the width cleaning area tolerance threshold is calculated based on the cleaning coverage range and the monitoring target range to determine the cleaning deviation tolerance threshold; based on the pure tracking algorithm, the underlying operation logic of the trajectory deviation operation module is constructed, the control preset trajectory is used as the reference trajectory, and the cleaning deviation tolerance threshold is used as the output correction condition to construct the trajectory deviation operation module.

[0036] In a preferred embodiment, it also includes:

[0037] According to the cleaning monitoring data, the first image of the acquisition sequence in the image acquisition information is obtained; the preview distance and the center point are read, and a target circle is drawn with the center point as the center and the preview distance as the radius, and the intersection group of the target circle and the preset control trajectory is used as the preview pixel point; the actual cleaning movement direction of the first image is obtained, and the angle between the preview pixel point and the actual cleaning movement direction is used as the deviation angle; according to the deviation angle and the preview distance, the driving curvature is determined, and based on the driving curvature, the first actual turning angle is obtained; the second image in the image acquisition information is extracted, the second image and the first image are adjacent images in the acquisition sequence, and the second actual turning angle is calculated according to the second image; according to the first actual turning angle and the second actual turning angle, a tracking trajectory is generated, and the trajectory deviation information is determined using the tracking trajectory and the reference trajectory; based on the cleaning monitoring area, the deviation data of the trajectory deviation information is corrected using the cleaning deviation tolerance threshold to determine the cleaning deviation information.

[0038] The trajectory deviation calculation module is constructed using the monitoring parameter sequence and cleaning control parameters. The specific process is as follows:

[0039] The cleaning control parameters include a cleaning control trajectory and a correlation between the cleaning control trajectory and the cleaning parameters, and the cleaning control trajectory is extracted as a control preset trajectory. According to the monitoring parameter sequence, the monitoring target timing and the monitoring target range are extracted. The monitoring target timing refers to a type of cleaning target positioning, and each type of cleaning target positioning has a corresponding cleaning control parameter, thereby obtaining the cleaning control parameter corresponding to the monitoring target timing, and based on the cleaning control parameter, extracting the cleaning nozzle width and nozzle operating parameters (cleaning water pressure, flow rate, etc.). Specifically, the wafer surface cleaning can be performed according to history so that the historical nozzle width, historical nozzle operating parameters and the washing nozzle width, nozzle operating parameters are compared for similarity, and the cleaning liquid coverage corresponding to the same historical nozzle width and historical nozzle operating parameters is obtained as the predicted cleaning coverage. Based on the cleaning coverage range and the monitoring target range, the width cleaning area tolerance threshold is calculated to determine the cleaning deviation tolerance threshold. That is to say, after the cleaning liquid is sprayed onto the wafer surface, diffusion will occur. As long as the diffused cleaning liquid can diffuse to the monitoring target range, the cleaning trajectory is considered qualified and no adjustment is required. Based on this, the width difference between the cleaning coverage range and the monitoring target range is obtained as the width cleaning area tolerance threshold. The underlying operation logic of the trajectory deviation operation module is constructed based on the pure tracking algorithm, and the control preset trajectory is used as the reference trajectory, and the cleaning deviation tolerance threshold is used as the output correction condition. That is to say, only when the deviation between the actual cleaning trajectory and the control preset trajectory exceeds the cleaning deviation tolerance threshold, it is considered that the trajectory is deviated, and then the trajectory is corrected, thereby constructing the trajectory deviation operation module. Analyzing the trajectory diffusion of the cleaning liquid is convenient for improving the accuracy of wafer surface cleaning monitoring.

[0040] Further based on the cleaning monitoring data, the trajectory deviation analysis and prediction is performed through the trajectory deviation calculation module to obtain the cleaning deviation information. The specific process is as follows:

[0041] According to the cleaning monitoring data, the first image of the acquisition sequence in the image acquisition information is obtained, that is, an image in the image acquisition information corresponding to the acquisition position, nozzle position, and cleaning target in the cleaning monitoring data. Further, the preview distance and the center point are read in the first image. The definitions of the preview distance and the center point have been explained in the previous steps. Here, the preview distance is represented by the number of pixels in the first image. With the center point as the center and the preview distance as the radius, a target circle is drawn, and the intersection group of the target circle and the preset control trajectory is used as the preview pixel point. The actual cleaning movement direction of the first image, that is, the movement direction of the nozzle, is obtained, and the angle between the preview pixel point and the actual cleaning movement direction is used as the deviation angle. According to the deviation angle and the preview distance, the travel curvature is determined, and based on the travel curvature, the first actual turning angle is obtained, as follows: , , further, can be simplified to ,in, is the running curvature, is the first actual turning angle, It is the length of the nozzle, which is determined according to the actual situation. is the preview distance, is the deviation angle.

[0042] Further extract the second image in the image acquisition information, the second image and the first image are adjacent images in the acquisition time sequence, and adopt the method of obtaining the first actual angle, and calculate the second actual angle according to the second image. Connect the first actual angle and the second actual angle to form a tracking trajectory, compare the tracking trajectory with the reference trajectory, and obtain the deviation between the two as the trajectory deviation information. Based on the cleaning monitoring area, the deviation data of the trajectory deviation information is corrected using the cleaning deviation tolerance threshold to determine the cleaning deviation information. In simple terms, if the deviation range of the trajectory deviation information is within the cleaning deviation tolerance threshold, and the trajectory range of the tracking trajectory is within the cleaning monitoring area, it is considered that there is no cleaning deviation. On the contrary, if the deviation range of the trajectory deviation information is not within the cleaning deviation tolerance threshold, and the trajectory range of the tracking trajectory is not within the cleaning monitoring area, the difference between the trajectory deviation information and the cleaning deviation tolerance threshold is obtained as the cleaning deviation information. In this way, real-time tracking of wafer surface cleaning is achieved, and the cleaning deviation information is obtained by technology, so as to adjust the cleaning trajectory and improve the wafer cleaning quality.

[0043] Calculating the cleaning probability distribution based on the cleaning deviation information and the cleaning control parameters;

[0044] In a preferred embodiment, it also includes:

[0045] Based on the cleaning deviation information and cleaning control parameters, a cleaning range intensity analysis is performed to determine the cleaning center point, cleaning range, and cleaning intensity distribution; a cleaning intensity distribution map is constructed based on the cleaning center point, cleaning range, and cleaning intensity distribution; based on the monitoring parameter sequence, the cleaning intensity distribution map is correspondingly connected according to the monitoring time sequence to determine a wafer cleaning intensity distribution overlay map; the cleaning intensity in the wafer cleaning intensity distribution overlay map is used to perform probability prediction to obtain the cleaning probability of each area.

[0046] In a preferred embodiment, it also includes:

[0047] A wafer cleaning target library is constructed, and a wafer cleaning intensity test is performed based on the wafer cleaning target library to determine the test data of each cleaning target; the intensity-cleaning probability relationship is fitted using the test data of each cleaning target, wherein the intensity-cleaning probability relationship includes a primary direct probability, a secondary indirect probability, and a superimposed probability coefficient of the primary direct probability and the secondary indirect probability; based on the probability relationship fitting result, a prediction operation module is constructed to perform probability prediction on the cleaning intensity.

[0048] The cleaning probability distribution calculation is performed based on the cleaning deviation information and the cleaning control parameters. The specific process is as follows:

[0049] Based on the cleaning deviation information and cleaning control parameters, a cleaning range strength analysis is performed to determine the cleaning center point, cleaning range, and cleaning strength distribution. The cleaning center point is the center point of the coverage formed by the cleaning liquid sprayed from the nozzle on the wafer surface. Generally, if the nozzle is perpendicular to the wafer surface and the wafer surface is a plane, then the center point is the intersection of the straight line perpendicular to the wafer surface through the center point of the nozzle of the nozzle and the wafer surface. The cleaning range is the coverage formed by the cleaning liquid sprayed from the nozzle on the wafer surface. The cleaning control trajectory in the cleaning control parameters can be corrected according to the cleaning deviation information, and the corrected trajectory range is used as the cleaning range. The cleaning strength is the flushing strength formed by the cleaning liquid on the wafer surface. Generally, the closer the position is to the cleaning center point, the greater the cleaning strength. Specifically, a cleaning test can be performed according to the cleaning control parameters, or the cleaning strength of each position on the wafer surface corresponding to the same cleaning control parameters in the history can be collected as the cleaning strength distribution. The cleaning force distribution map is constructed based on the cleaning center point, cleaning range, and cleaning force distribution, that is, the wafer surface image is used as the basic map, and the cleaning center point, cleaning range, and cleaning force distribution are marked on the basic map. The result is the cleaning force distribution map, wherein different types of cleaning target positioning in the cleaning control parameters correspond to different cleaning force distribution maps. Based on the monitoring parameter sequence, the cleaning force distribution maps are correspondingly connected according to the monitoring time sequence to obtain a wafer cleaning force distribution superposition map, that is, after different cleaning force distribution maps are superimposed together, there may be overlapping cleaning positions, and there may also be positions that are not cleaned.

[0050] The cleaning intensity in the wafer cleaning intensity distribution overlay map is used to perform probability prediction to obtain the cleaning probability of each area. Specifically, a wafer cleaning target library is first constructed, and the wafer cleaning target library includes the positions of all cleaning targets in the aforementioned cleaning target positioning. Based on the wafer cleaning target library, a wafer cleaning intensity test is performed to determine the test data of each cleaning target. The test data of each cleaning target can be understood as historical cleaning record data, that is, the historical cleaning record data identical to each cleaning target is obtained as the test data of each cleaning target. The test data of each cleaning target includes the historical cleaning center point, the historical cleaning range, the historical cleaning intensity distribution, and the historical cleaning effect. Based on this, a historical cleaning intensity distribution map is constructed. Based on the monitoring parameter sequence, the historical cleaning intensity distribution map is correspondingly connected according to the monitoring time sequence to obtain multiple historical wafer cleaning intensity distribution superposition maps, and then the intensity-cleaning probability relationship is fitted, wherein the intensity-cleaning probability relationship includes a direct probability, a secondary indirect probability, and a superposition probability coefficient of the direct probability and the secondary indirect probability. The direct probability refers to the probability that a cleaning target is cleaned only once and the cleaning intensity is large, that is, the cleaning effect is good. Specifically, the ratio of the number of times the historical cleaning effect after one cleaning corresponding to the position reaches the expected cleaning effect to the total number of test data of each cleaning target can be counted as the direct probability. The secondary indirect probability refers to the situation that a cleaning target does not achieve the expected cleaning effect after one cleaning, but when its adjacent cleaning target is cleaned, the cleaning liquid will diffuse to the cleaning target, so that the cleaning effect of the cleaning target reaches the expected cleaning effect. The ratio of the number of times the historical cleaning effect after the second cleaning corresponding to the position reaches the expected cleaning effect to the total number of test data of each cleaning target is used as the secondary indirect probability. The superposition probability coefficient of the first direct probability and the second indirect probability refers to the probability that a cleaning target can be cleaned twice after one cleaning. For the cleaning targets that have been cleaned twice, the first direct probability and the second indirect probability are added together as the superposition probability coefficient of the first direct probability and the second indirect probability. Based on the probability relationship fitting results, a prediction operation module is constructed. The prediction operation module is a functional module for probability prediction of cleaning intensity. That is to say, based on the first direct probability, second indirect probability and the superimposed probability coefficient of the first direct probability and the second indirect probability corresponding to each cleaning target in the intensity-cleaning probability relationship, the cleaning probability distribution calculation is performed on the cleaning target, that is, the probability of each cleaning target being cleaned properly is predicted. That is to say, the probability relationship fitting results are stored in the prediction operation module, and then the positions of each cleaning target in the cleaning target positioning are matched with each cleaning target in the probability relationship fitting results to obtain the corresponding first direct probability, second indirect probability and first direct probability and second indirect probability. The sum of the three probabilities can be calculated as the cleaning probability of each cleaning target, thereby obtaining the cleaning probability distribution calculation results of each cleaning target.This enables cleaning probability analysis of different cleaning targets on the wafer surface and cleaning monitoring of the wafer surface, making it easier to make timely corrections to locations with poor cleaning effects and improving cleaning quality.

[0051] Based on the cleaning probability distribution calculation results, monitoring data feedback is provided and monitoring control guidance information is generated.

[0052] Based on the calculation results of the cleaning probability distribution, monitoring data feedback is performed and production monitoring control guidance information is produced. Production monitoring control guidance information refers to feedback correction information for cleaning control parameters. For example, if the cleaning probability of a cleaning target in the calculation results of the cleaning probability distribution is less than or equal to the preset probability, the preset probability is set by technicians in this field based on historical experience, such as 80%. Then, according to the cleaning target, the cleaning control trajectory in the cleaning control parameters is modified so that the cleaning control trajectory approaches the cleaning target. At the same time, the cleaning parameters are adjusted, such as increasing the cleaning water pressure, etc., as production monitoring control guidance information. Subsequently, the wafer surface cleaning can be adjusted according to the production monitoring control guidance information to improve the cleaning effect.

[0053] Based on the above analysis, it can be seen that one or more technical solutions provided by the present disclosure can achieve the following beneficial effects:

[0054] Cleaning targets are identified on the wafer surface, their distribution positions are located, cleaning control parameters are determined, cleaning enforcement equipment is controlled using cleaning control parameters, and a monitoring parameter sequence is generated simultaneously to collect monitoring data of the wafer cleaning process. Cleaning trajectory tracking parameters are extracted from the monitoring data to obtain cleaning monitoring data. A trajectory deviation operation module is constructed using the monitoring parameter sequence and cleaning control parameters. Based on the cleaning monitoring data, trajectory deviation analysis and prediction are performed through the trajectory deviation operation module to obtain cleaning deviation information. Cleaning probability distribution calculation is performed based on the cleaning deviation information and cleaning control parameters. Monitoring data feedback is performed based on the cleaning probability distribution calculation results, and monitoring and control guidance information is produced. By performing deviation analysis on the cleaning trajectory and then performing cleaning probability analysis, the technical effect of improving wafer surface cleaning monitoring effect, timely feedback correction, and improving cleaning quality is achieved.

[0055] Embodiment 2

[0056] Based on the same inventive concept as the wafer surface cleaning monitoring method in the above-mentioned embodiment, Figure 2 As shown, the present application also provides a wafer surface cleaning monitoring system, the system comprising:

[0057] A cleaning target identification unit 11, wherein the cleaning target identification unit 11 is used to identify cleaning targets on the wafer surface, locate distribution positions, and determine cleaning control parameters;

[0058] A parameter control unit 12, the parameter control unit 12 is used to use the cleaning control parameters to perform parameter control on the cleaning law enforcement equipment and synchronously generate a monitoring parameter sequence;

[0059] A cleaning trajectory tracking unit 13, wherein the cleaning trajectory tracking unit 13 is used to collect monitoring data of the wafer cleaning process, extract cleaning trajectory tracking parameters from the monitoring data, and obtain cleaning monitoring data;

[0060] A trajectory deviation analysis unit 14, wherein the trajectory deviation analysis unit 14 is used to construct a trajectory deviation operation module using a monitoring parameter sequence and a cleaning control parameter, and perform trajectory deviation analysis and prediction through the trajectory deviation operation module based on cleaning monitoring data to obtain cleaning deviation information;

[0061] A cleaning probability distribution calculation unit 15, wherein the cleaning probability distribution calculation unit 15 is used to perform cleaning probability distribution calculation based on the cleaning deviation information and cleaning control parameters;

[0062] The monitoring data feedback unit 16 is used to perform monitoring data feedback based on the cleaning probability distribution calculation result and generate monitoring control guidance information.

[0063] Furthermore, the cleaning trajectory tracking unit 13 is also used for:

[0064] Capture images of the wafer cleaning process through monitoring equipment to obtain image acquisition information, including acquisition point locations;

[0065] Obtaining aiming nozzle parameter information of the monitoring device, wherein the aiming nozzle parameter information includes nozzle position, nozzle operating parameters, and cleaning target;

[0066] Determine the preview distance and center point according to the collection position, nozzle position, and cleaning target;

[0067] The spray contact area is identified through image acquisition information to determine the cleaning monitoring area, and the preview distance, center point, and cleaning monitoring area are used as the cleaning monitoring data.

[0068] Furthermore, the trajectory deviation analysis unit 14 is also used for:

[0069] Determining a preset control trajectory according to the cleaning control parameters;

[0070] Determine the monitoring target timing and monitoring target range according to the monitoring parameter sequence;

[0071] Determine the corresponding cleaning control parameters according to the monitoring target time sequence, extract the cleaning nozzle width and nozzle working parameters based on the cleaning control parameters, predict the cleaning coverage range, and calculate the width cleaning area tolerance threshold based on the cleaning coverage range and the monitoring target range to determine the cleaning deviation tolerance threshold;

[0072] The underlying operation logic of the trajectory deviation operation module is constructed based on the pure tracking algorithm, the control preset trajectory is used as the reference trajectory, and the cleaning deviation tolerance threshold is used as the output correction condition to construct the trajectory deviation operation module.

[0073] Furthermore, the trajectory deviation analysis unit 14 is also used for:

[0074] According to the cleaning monitoring data, a first image of the acquisition time sequence in the image acquisition information is acquired;

[0075] Read the preview distance and the center point, draw a target circle with the center point as the center and the preview distance as the radius, and use the intersection group of the target circle and the preset control trajectory as the preview pixel point;

[0076] The actual cleaning movement direction of the first image is obtained, and the angle between the preview pixel point and the actual cleaning movement direction is used as the deviation angle;

[0077] Determine a travel curvature according to the deviation angle and the preview distance, and obtain a first actual turning angle based on the travel curvature;

[0078] Extracting a second image from the image acquisition information, where the second image and the first image are adjacent images in an acquisition time sequence, and obtaining a second actual rotation angle by calculation according to the second image;

[0079] generating a tracking trajectory according to the first actual rotation angle and the second actual rotation angle, and determining trajectory deviation information using the tracking trajectory and the reference trajectory;

[0080] Based on the cleaning monitoring area, the cleaning deviation tolerance threshold is used to correct the deviation data of the trajectory deviation information to determine the cleaning deviation information.

[0081] Furthermore, the cleaning probability distribution calculation unit 15 is also used for:

[0082] Based on the cleaning deviation information and cleaning control parameters, a cleaning range strength analysis is performed to determine the cleaning center point, cleaning range, and cleaning strength distribution;

[0083] Constructing a cleaning intensity distribution map based on the cleaning center point, cleaning range, and cleaning intensity distribution;

[0084] Based on the monitoring parameter sequence, the cleaning intensity distribution diagrams are correspondingly connected according to the monitoring time sequence to determine a wafer cleaning intensity distribution overlay diagram;

[0085] The cleaning intensity in the wafer cleaning intensity distribution overlay diagram is used to perform probability prediction to obtain the cleaning probability of each area.

[0086] Furthermore, the cleaning probability distribution calculation unit 15 is also used for:

[0087] Constructing a wafer cleaning target library, and performing a wafer cleaning intensity test based on the wafer cleaning target library to determine test data for each cleaning target;

[0088] Using the test data of each cleaning target to fit the intensity-cleaning probability relationship, wherein the intensity-cleaning probability relationship includes a primary direct probability, a secondary indirect probability, and a superimposed probability coefficient of the primary direct probability and the secondary indirect probability;

[0089] Based on the probability relationship fitting results, a prediction operation module is constructed to perform probability prediction of the cleaning intensity.

[0090] Furthermore, the cleaning target identification unit 11 is also used for:

[0091] Obtain wafer surface cleaning target recognition results, including cleaning target location and cleaning target type;

[0092] Clustering the cleaning target locations based on the cleaning target types, and using clustering results to generate labels to identify the cleaning target locations;

[0093] Determine the cleaning parameters according to the type clustering results of cleaning target positioning;

[0094] The cleaning control trajectory is determined by locating the cleaning target based on the identifier, and an association relationship between the cleaning control trajectory and the cleaning parameters is established to obtain the cleaning control parameters.

[0095] The specific example of a wafer surface cleaning monitoring method in the aforementioned embodiment 1 is also applicable to a wafer surface cleaning monitoring system in the present embodiment. Through the aforementioned detailed description of a wafer surface cleaning monitoring method, a person skilled in the art can clearly understand a wafer surface cleaning monitoring system in the present embodiment, so for the sake of brevity of the specification, it will not be described in detail here.

[0096] It should be understood that various forms of the processes shown above can be used to reorder, add or delete steps, as long as the expected results of the technical solutions disclosed in the present invention can be achieved, and this document does not limit them here.

[0097] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A wafer surface cleaning monitoring method, characterized in that: include: Identify cleaning targets on the wafer surface, locate distribution positions, and determine cleaning control parameters; The cleaning execution device is parameter controlled by using the cleaning control parameters, and a monitoring parameter sequence is generated synchronously, wherein the monitoring parameter sequence refers to a sequence of parameters that need to be monitored after the cleaning execution device is parameter controlled and the cleaning execution device is cleaned according to the corresponding cleaning control parameters in different time periods; Collect monitoring data of the wafer cleaning process, extract cleaning trajectory tracking parameters from the monitoring data, and obtain cleaning monitoring data; A trajectory deviation calculation module is constructed using the monitoring parameter sequence and the cleaning control parameters. Based on the cleaning monitoring data, the trajectory deviation calculation module is used to perform trajectory deviation analysis and prediction to obtain cleaning deviation information. Calculating the cleaning probability distribution based on the cleaning deviation information and the cleaning control parameters; Provide monitoring data feedback based on the cleaning probability distribution calculation results and generate monitoring control guidance information; The collecting of monitoring data of the wafer cleaning process, extracting cleaning trajectory tracking parameters from the monitoring data, and obtaining cleaning monitoring data include: Capture images of the wafer cleaning process through monitoring equipment to obtain image acquisition information, including acquisition point locations; Obtaining aiming nozzle parameter information of the monitoring device, wherein the aiming nozzle parameter information includes nozzle position, nozzle operating parameters, and cleaning target; Determine the preview distance and center point according to the collection point position, nozzle position, and cleaning target; The spray contact area is identified through image acquisition information to determine the cleaning monitoring area, and the preview distance, center point, and cleaning monitoring area are used as the cleaning monitoring data.

2. The method according to claim 1, characterized in that The method of constructing a trajectory deviation operation module using the monitoring parameter sequence and the cleaning control parameters includes: Determining a preset control trajectory according to the cleaning control parameters; Determine the monitoring target timing and monitoring target range according to the monitoring parameter sequence; Determine the corresponding cleaning control parameters according to the monitoring target time sequence, extract the cleaning nozzle width and nozzle working parameters based on the cleaning control parameters, predict the cleaning coverage range, and calculate the width cleaning area tolerance threshold based on the cleaning coverage range and the monitoring target range to determine the cleaning deviation tolerance threshold; The underlying operation logic of the trajectory deviation operation module is constructed based on the pure tracking algorithm, the control preset trajectory is used as the reference trajectory, and the cleaning deviation tolerance threshold is used as the output correction condition to construct the trajectory deviation operation module.

3. The method according to claim 2, characterized in that Based on the cleaning monitoring data, the trajectory deviation calculation module is used to perform trajectory deviation analysis and prediction to obtain cleaning deviation information, including: According to the cleaning monitoring data, a first image of the acquisition time sequence in the image acquisition information is acquired; Read the preview distance and the center point, draw a target circle with the center point as the center and the preview distance as the radius, and use the intersection group of the target circle and the preset control trajectory as the preview pixel point; The actual cleaning movement direction of the first image is obtained, and the angle between the preview pixel point and the actual cleaning movement direction is used as the deviation angle; Determine a travel curvature according to the deviation angle and the preview distance, and obtain a first actual turning angle based on the travel curvature; Extracting a second image from the image acquisition information, where the second image and the first image are adjacent images in an acquisition time sequence, and obtaining a second actual rotation angle by calculation according to the second image; generating a tracking trajectory according to the first actual rotation angle and the second actual rotation angle, and determining trajectory deviation information using the tracking trajectory and the reference trajectory; Based on the cleaning monitoring area, the cleaning deviation tolerance threshold is used to correct the deviation data of the trajectory deviation information to determine the cleaning deviation information.

4. The method according to claim 1, characterized in that The cleaning probability distribution calculation is performed based on the cleaning deviation information and the cleaning control parameters, including: Based on the cleaning deviation information and cleaning control parameters, a cleaning range strength analysis is performed to determine the cleaning center point, cleaning range, and cleaning strength distribution; Constructing a cleaning intensity distribution map based on the cleaning center point, cleaning range, and cleaning intensity distribution; Based on the monitoring parameter sequence, the cleaning intensity distribution diagrams are correspondingly connected according to the monitoring time sequence to determine a wafer cleaning intensity distribution overlay diagram; The cleaning intensity in the wafer cleaning intensity distribution overlay diagram is used to perform probability prediction to obtain the cleaning probability of each area.

5. The method according to claim 4, characterized in that Probabilistic prediction is performed using the cleaning intensity in the wafer cleaning intensity distribution overlay diagram, including: Constructing a wafer cleaning target library, and performing a wafer cleaning intensity test based on the wafer cleaning target library to determine test data for each cleaning target; Using the test data of each cleaning target to fit the intensity-cleaning probability relationship, wherein the intensity-cleaning probability relationship includes a primary direct probability, a secondary indirect probability, and a superimposed probability coefficient of the primary direct probability and the secondary indirect probability; Based on the probability relationship fitting results, a prediction operation module is constructed to perform probability prediction of the cleaning intensity.

6. The method according to claim 1, characterized in that The cleaning target identification, distribution position positioning, and cleaning control parameter determination on the wafer surface include: Obtain wafer surface cleaning target recognition results, including cleaning target location and cleaning target type; Clustering the cleaning target locations based on the cleaning target types, and using clustering results to generate labels to identify the cleaning target locations; Determine the cleaning parameters according to the type clustering results of cleaning target positioning; The cleaning control trajectory is determined by locating the cleaning target based on the identifier, and an association relationship between the cleaning control trajectory and the cleaning parameters is established to obtain the cleaning control parameters.

7. A wafer surface cleaning monitoring system, characterized in that: The system is used to perform the steps of any one of the wafer surface cleaning monitoring methods according to claims 1 to 6, comprising: A cleaning target identification unit, which is used to identify cleaning targets on the wafer surface, locate distribution positions, and determine cleaning control parameters; A parameter control unit, the parameter control unit is used to perform parameter control on the cleaning execution device using the cleaning control parameters, and synchronously generate a monitoring parameter sequence, the monitoring parameter sequence refers to a sequence of parameters that need to be monitored after the cleaning execution device is parameter controlled and the cleaning execution device is cleaned according to the corresponding cleaning control parameters in different time periods; A cleaning trajectory tracking unit, wherein the cleaning trajectory tracking unit is used to collect monitoring data of a wafer cleaning process, extract cleaning trajectory tracking parameters from the monitoring data, and obtain cleaning monitoring data; A trajectory deviation analysis unit, wherein the trajectory deviation analysis unit is used to construct a trajectory deviation operation module using a monitoring parameter sequence and a cleaning control parameter, and to perform trajectory deviation analysis and prediction through the trajectory deviation operation module based on cleaning monitoring data to obtain cleaning deviation information; A cleaning probability distribution calculation unit, the cleaning probability distribution calculation unit is used to perform cleaning probability distribution calculation based on the cleaning deviation information and cleaning control parameters; A monitoring data feedback unit is used to perform monitoring data feedback based on the cleaning probability distribution calculation result and generate monitoring control guidance information.

Images