High-precision dynamic spectrum modulation method and system for semiconductor wafer detection light source

Abstract

The invention relates to the technical field of wafer spectrum detection, and discloses a high-precision dynamic spectrum modulation method and system for a semiconductor wafer detection light source. A detection light source module is driven to carry out initialized irradiation on a wafer to be detected, initial detection information is collected, the initial detection information is matched according to a database, and a region division detection scheme is formulated; applying corresponding spectral irradiation according to a region division detection scheme, acquiring wafer data in an initial state, performing adaptive evaluation on the detection data to adaptively regulate and control the detection scheme, fusing the detection data in the initial state and the detection data in the regulation and control state, analyzing the detection progress, and performing cyclic regulation and control until the progress meets a threshold value. The dynamic spectrum modulation ensures that each area of the surface of the wafer is accurately detected, the detection scheme is evaluated and adjusted in real time, the detection coverage rate and accuracy are improved, the detection efficiency and accuracy are improved through data fusion and progress analysis, manual intervention is reduced, and the problem that in the prior art, the leakage detection phenomenon is likely to occur due to static light source irradiation is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of wafer spectrum detection, and in particular to a high-precision dynamic spectrum modulation method and system for a semiconductor wafer detection light source. Background Art

[0002] Spectral inspection can not only effectively detect defects on the wafer surface, but also provide real-time feedback for subsequent manufacturing processes. However, as the semiconductor industry's requirements for wafer quality continue to increase, traditional inspection methods face some bottlenecks in terms of accuracy and efficiency.

[0003] In traditional wafer inspection, a static light source is often used. This method cannot adapt to the complex tiny defects and uneven spectral characteristics of the wafer surface, resulting in missed detection and false detection. Summary of the invention

[0004] The purpose of the present invention is to provide a high-precision dynamic spectrum modulation method and system for a semiconductor wafer inspection light source, aiming to solve the problem of missed detection easily occurring in static light source illumination in the prior art.

[0005] The present invention is implemented in this way. In a first aspect, the present invention provides a high-precision dynamic spectrum modulation method for a semiconductor wafer inspection light source, comprising: The detection light source module is driven to perform initial light source irradiation and feedback information collection on the wafer to be tested, so as to obtain initial detection information of the wafer to be tested; Performing wafer information matching on the initial detection information of the wafer to be tested according to a preset database to obtain a region division detection scheme corresponding to the wafer to be tested; Applying corresponding spectral illumination to each position area of ​​the wafer to be tested according to the area division detection scheme, and collecting detection data of the wafer to be tested to obtain wafer detection data of the initial state of the wafer to be tested; Performing an evaluation process on the detection adaptability of the wafer detection data of the wafer to be tested, and adaptively regulating and executing the area division detection scheme according to the evaluation result, so as to obtain the wafer detection data of the regulated state of the wafer to be tested; Data fusion analysis is performed on the wafer detection data of the initial state and the control state of the wafer to be tested to obtain the detection progress characteristics of the wafer to be tested, and the adaptive control of the area division detection scheme and the data collection of the wafer detection data are repeated until the detection progress characteristics meet the specified threshold value to complete the spectral detection of the wafer to be tested.

[0006] In a second aspect, the present invention provides a high-precision dynamic spectrum modulation system for a semiconductor wafer inspection light source, which is used to implement a high-precision dynamic spectrum modulation method for a semiconductor wafer inspection light source as described in any one of the first aspects, including: An information acquisition module is used to drive the detection light source module to perform initial light source irradiation on the wafer to be tested and collect feedback information to obtain initial detection information of the wafer to be tested; A scheme planning module is used to match the initial detection information of the wafer to be tested with wafer information according to a preset database to obtain a regional division detection scheme corresponding to the wafer to be tested; An initial detection module, used to apply corresponding spectral illumination to each position area of ​​the wafer to be tested according to the area division detection scheme, and collect detection data of the wafer to be tested to obtain wafer detection data of the initial state of the wafer to be tested; A control detection module, used for evaluating the detection adaptability of the wafer detection data of the wafer to be tested, and adaptively controlling and executing the area division detection scheme according to the evaluation result, so as to obtain the wafer detection data of the control state of the wafer to be tested; A cyclic detection module is used to perform data fusion analysis on the wafer detection data of the initial state and the control state of the wafer to be tested to obtain the detection progress characteristics of the wafer to be tested, and cyclically perform adaptive control of the area division detection scheme and data collection of the wafer detection data until the detection progress characteristics meet the specified threshold value to complete the spectral detection of the wafer to be tested.

[0007] The present invention provides a high-precision dynamic spectrum modulation method for a semiconductor wafer inspection light source, which has the following beneficial effects: The present invention drives the detection light source module to perform initialization irradiation on the wafer to be tested, collects initial detection information, matches the initial detection information according to a database, formulates a regional division detection plan, applies corresponding spectral irradiation according to the regional division detection plan, collects detection data, obtains wafer data in an initial state, performs adaptability evaluation on the detection data, and adaptively adjusts the detection plan according to the evaluation result, fuses the detection data of the initial and regulated states, analyzes the detection progress, cyclically adjusts the control until the progress meets the threshold, completes the spectral detection, and ensures that each area on the wafer surface is accurately detected by dynamic spectral modulation. The detection plan is evaluated and adjusted in real time to improve the detection coverage and accuracy. Data fusion and progress analysis improve the efficiency and accuracy of detection, reduce manual intervention, and solve the problem of missed detection easily caused by static light source irradiation in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1It is a schematic diagram of the steps of a high-precision dynamic spectrum modulation method for a semiconductor wafer inspection light source provided by an embodiment of the present invention; Figure 2 It is a structural schematic diagram of a high-precision dynamic spectrum modulation system for a semiconductor wafer inspection light source provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0009] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0010] The implementation of the present invention is described in detail below in conjunction with specific embodiments.

[0011] Reference Figure 1 , Figure 2 As shown, a preferred embodiment of the present invention is provided.

[0012] In a first aspect, the present invention provides a high-precision dynamic spectrum modulation method for a semiconductor wafer inspection light source, comprising: S1: driving the detection light source module to perform initial light source irradiation and feedback information collection on the wafer to be tested, so as to obtain initial detection information of the wafer to be tested; S2: performing wafer information matching on the initial detection information of the wafer to be tested according to a preset database to obtain a region division detection scheme corresponding to the wafer to be tested; S3: applying corresponding spectral illumination to each position area of ​​the wafer to be tested according to the area division detection scheme, and collecting detection data of the wafer to be tested to obtain wafer detection data of the initial state of the wafer to be tested; S4: performing an evaluation process on the wafer detection data of the wafer to be tested for detection adaptability, and adaptively regulating and executing the area division detection scheme according to the evaluation result, so as to obtain wafer detection data of the regulated state of the wafer to be tested; S5: Perform data fusion analysis on the wafer detection data of the initial state and the control state of the wafer to be tested to obtain the detection progress characteristics of the wafer to be tested, and cyclically perform adaptive control of the area division detection scheme and data collection of the wafer detection data until the detection progress characteristics meet the specified threshold to complete the spectral detection of the wafer to be tested.

[0013] Specifically, in step S1 of the embodiment provided by the present invention, the driving detection light source module is started, and the wafer to be tested is initially irradiated with a specific light source intensity and wavelength range, and a suitable spectral range is selected, usually a band covering possible defects or special spectral characteristics of the wafer to be tested is selected, and the light source is evenly distributed on the surface of the wafer to ensure that the entire wafer surface is illuminated. This irradiation is usually full coverage, and the light source intensity does not need to be too high at this stage to avoid damaging the wafer or over-excitation.

[0014] More specifically, after the light source is irradiated, the detection system uses sensors (such as spectral sensors, photoelectric detectors, etc.) to collect light signals reflected or transmitted from the wafer surface. The collected signals include data such as light intensity, spectral characteristics, and reflectivity. The feedback information includes the reflection characteristics or absorption characteristics of different areas on the wafer surface. These signals will be converted into preliminary detection data for subsequent analysis.

[0015] More specifically, based on the feedback signal data, the system generates preliminary detection information of the wafer to be tested, including spectral characteristics, reflection intensity, etc. This information provides basic data for subsequent area division, matching and analysis.

[0016] It can be understood that through light source irradiation and feedback information collection, the overall spectral characteristics and reflective properties of the wafer surface can be obtained, providing necessary data support for subsequent wafer status assessment and regional division detection plan formulation. The initial light source irradiation covers the entire wafer, which can ensure that data collection in no area is missed. The uniformity of light source irradiation and high-precision collection of feedback signals ensure the accuracy of preliminary detection information and provide high-quality data support for subsequent regional division and adaptive regulation.

[0017] Specifically, in step S2 of the embodiment provided by the present invention, the system first reads and loads the initial detection information of the wafer to be tested, including feedback data after light source irradiation, reflection spectrum, surface characteristics, etc. This information is obtained based on the initialization light source irradiation and feedback information collection stage.

[0018] More specifically, the system matches the initially collected inspection information with a preset wafer database. The preset database contains information such as spectral characteristics, known defect patterns, surface scratches, and contamination of different types of wafers. By comparing the initial inspection information with the template data in the database, the system can identify the basic type of wafer, surface state, and possible defect characteristics.

[0019] More specifically, the system analyzes the initial detection information of the wafer to be tested based on a matching algorithm (such as similarity calculation, machine learning algorithm or deep learning model) to identify key features on the wafer surface. Features may include local thickness differences on the wafer surface, changes in spectral reflectivity, and potential defect areas (such as cracks, contaminants, surface unevenness, etc.). Based on the matching results, the system divides the wafer into multiple detection areas. The basis for the division of these areas includes: the spatial distribution of the reflective characteristics of the wafer surface, possible defect areas, preset accuracy requirements, etc.

[0020] More specifically, each detection area corresponds to a specific spectral illumination parameter, including spectral band, light source intensity, scanning mode, etc. These parameters will be used for subsequent spectral illumination and data collection. The system formulates different detection strategies for each area according to the regional division detection scheme. For example, some areas require more refined spectral illumination or more frequent data collection, while some areas can adopt low-precision, low-frequency detection strategies.

[0021] More specifically, after initially generating the regional division detection scheme, the system will also optimize and adjust according to the complexity of the wafer surface and the actual detection needs. For example, some special areas may need to be dynamically adjusted and refined according to the current detection data. The system uses an adaptive algorithm to optimize the regional division detection scheme, and fine-tunes the division of the detection area according to the preliminary detection results to improve the accuracy of the regional division. Based on the optimized regional division, the system finally generates a complete detection scheme suitable for the wafer to be tested. The scheme includes specific regional division, detection mode, light source irradiation parameters, acquisition strategy, etc.

[0022] It can be understood that through database matching and feature analysis, the wafer can be accurately divided into regions, and regions with different characteristics and states can be assigned to different detection areas. Based on the characteristics of different wafers, customized detection plans can be generated. The spectral illumination and detection strategies of each area can accurately match its surface characteristics and possible defect types, thereby improving the accuracy of detection.

[0023] Specifically, in step S3 of the embodiment provided by the present invention, based on the area division detection scheme generated in the previous step, the system clarifies the specific spectral irradiation scheme for each area. Each area requires different spectral wavelength, irradiation intensity, irradiation time and other parameters. These parameters are customized according to the surface characteristics of the wafer and known defect types. For each area, the spectral irradiation scheme will be optimized according to the characteristics of the area to ensure that the most comprehensive and accurate detection data can be obtained.

[0024] More specifically, according to the area division and illumination scheme, the system drives the light source module to perform spectral illumination on each position area of ​​the wafer to be tested in turn. At this time, the wavelength, intensity, and illumination method (such as direct or scattered) of the light source are adjusted according to the needs of the area. After the illumination of each area is completed, the reverse spectral data will be collected according to the reflected or transmitted light signal to capture the detailed changes on the wafer surface in the area. Under the illumination of the light source, the light reflected back from the wafer surface will be collected by high-precision sensors (such as spectral detectors, CCD cameras, photoelectric detectors, etc.). The system will collect the detection data of each area one by one according to the regionalized design. The detection data includes reflectivity, spectral intensity, surface morphological characteristics, defect signals, etc. The collected data of each area will be stored and associated with the lighting conditions of the area.

[0025] More specifically, the system compares the inspection data of each area with the known database model to preliminarily determine the state of the wafer. Based on the collected data of all areas, the system summarizes and generates the initial state inspection data of the wafer to be tested. These data comprehensively reflect the surface condition of the wafer, including potential defects, surface unevenness, etc. The initial inspection data usually includes the spectral reflectivity, defect location, contaminated area and other information of each area of ​​the wafer, and can generate relevant visual images or reports as needed for subsequent analysis or processing.

[0026] It can be understood that by applying different spectral illumination schemes to different areas of the wafer and collecting feedback information, the system can obtain the surface state of the wafer in detail and accurately, especially detect subtle defects and unevenness. According to the specific characteristics of the wafer and the known database model, the regional division detection scheme can finely optimize the spectral illumination scheme of each area to improve the accuracy and efficiency of detection. Through efficient data collection and processing methods, the system can quickly identify the status of the wafer, including potential defects or surface problems, which helps to improve quality control and wafer detection accuracy in the production process.

[0027] Specifically, in step S4 of the embodiment provided by the present invention, the system first performs a detailed analysis on the collected wafer inspection data, and evaluates the quality and accuracy of the data by performing statistical analysis on the inspection results of each area (such as reflectivity, spectral data, and surface features). Based on known standard values ​​or preset database information, the system automatically determines whether the current inspection data meets the expected accuracy requirements. If there are abnormalities in the inspection data (for example, the spectral data is too noisy, the reflection in certain areas is abnormal, or defects cannot be distinguished), the system will identify these deviations and evaluate their adaptability.

[0028] More specifically, by comparing and analyzing the detection data of each area, the system can find out which areas have poor data quality and specifically determine whether it is caused by insufficient light source illumination, insufficient sensor accuracy, surface contamination or other environmental factors. If the data in certain areas cannot meet the detection requirements or does not meet the expected characteristics, the system will mark them as "problem areas" or "areas requiring adjustment."

[0029] It should be noted that there are several ways to deal with these problem areas, including: adjusting the detection plan for subsequent areas based on the problem area and detecting the problem area in the next round of detection, or adjusting the detection plan for the problem area in real time to complete the detection of the problem area.

[0030] More specifically, for unsuitable areas, the system will dynamically adjust the area division and illumination parameters, for example, increasing the intensity of the light source, changing the wavelength of the light source, adjusting the illumination angle or extending the illumination time, etc., to improve the detection accuracy of the area. If the detection effect of some areas is poor, different detection modes may be used, such as using different sensors, using more accurate spectral analysis instruments, or changing the data collection method (such as multiple scans, different illumination angles, etc.). In some cases, the system will readjust the regional division based on the preliminary detection results, merge or refine the area, so as to conduct more accurate detection for specific problems.

[0031] More specifically, after completing the adaptive control of the detection scheme, the system will re-execute the adjusted detection scheme, that is, re-inspect the wafer to be tested under the adjusted area division, light source illumination parameters and detection mode, collect new data, process and analyze the new data and compare them with the preliminary detection data to ensure that the adjusted scheme can effectively improve the accuracy and integrity of the data.

[0032] More specifically, the system will generate controlled wafer inspection data, which will include more accurate inspection results of each area after adaptive control, covering the complete state of the wafer. The wafer inspection data in the controlled state is more accurate than the initial data and can reflect the actual situation on the surface and inside of the wafer, especially in identifying defects, surface unevenness or other minor changes, and can provide more detailed analysis results.

[0033] It can be understood that through adaptive control, the system can respond to changes in the surface conditions of different wafers and optimize the detection process, thereby improving the accuracy of the detection results. In complex environments or when the surface characteristics of the wafers change significantly, the detection scheme can be dynamically adjusted to reduce the impact of environmental interference or instrument accuracy issues.

[0034] Specifically, in step S5 of the embodiment provided by the present invention, the initial state of the wafer to be tested (data before regulation) is fused with the data after regulation. Specifically, these data include spectral information, reflection intensity, surface morphology, etc. of different regions. Weighted averaging, maximum likelihood estimation (MLE) or data fusion algorithm (such as Kalman filtering, Bayesian inference, etc.) are used to merge the detection data from different regions and time points to form a global and comprehensive detection data model, and detection progress features are extracted from the fused data, which are usually indicators related to detection completion, such as: detection data collection coverage of each region, data accuracy, correction progress of problem areas, etc.

[0035] More specifically, based on the extracted progress features, the system evaluates the current progress of the inspection. The evaluation criteria for the inspection progress features usually include: completeness of the coverage area: whether all the inspection data of each area has been collected, data quality: whether the collected data meets the accuracy requirements, whether there are large deviations or anomalies, and control effect: whether the spectral irradiation adjustment for the non-compliant areas is effective, and whether the original inspection problems have been improved.

[0036] More specifically, the system sets a prescribed progress threshold, usually based on factors such as completeness of testing, data quality, and time, to ensure that all necessary tests are completed within the specified time and with the required accuracy.

[0037] More specifically, based on the current inspection progress characteristics, the system determines whether the set progress threshold has been reached. If not, the system will further optimize the area division and spectral illumination scheme based on the analysis results of the previous data. For example, it can increase the illumination intensity in certain areas, adjust the acquisition frequency, or use different inspection modes (such as high-resolution acquisition). The system continues to collect data on the unfinished areas of the wafer to be tested and updates the inspection data based on the optimized inspection scheme. The system will provide real-time feedback on the quality of the current data collection, and refine and optimize the scheme based on the real-time feedback to ensure the efficiency and accuracy of the inspection.

[0038] More specifically, when the detection progress characteristics meet the specified threshold, it means that all areas have successfully collected data with sufficient accuracy and have performed necessary adaptive adjustments to achieve the expected detection effect. At this point, the system has completed the spectral detection of the wafer and can generate a final test report or subsequent analysis data.

[0039] It is understandable that by dynamically adjusting the area division and spectral illumination scheme, the system can adaptively adjust to the characteristics of different wafer surfaces and optimize the detection process. Through data fusion and progress feature evaluation, the system can quickly identify weak links in the detection, adjust the scheme in time, reduce repeated collection and invalid operations, thereby improving the overall detection efficiency and data accuracy. Adaptive cycle control and data collection ensure that after each adjustment, the detection result is closer to the exact value, ultimately ensuring the comprehensiveness and accuracy of wafer detection and meeting the high standards in industrial production.

[0040] The present invention provides a high-precision dynamic spectrum modulation method for a semiconductor wafer inspection light source, which has the following beneficial effects: The present invention drives the detection light source module to perform initialization irradiation on the wafer to be tested, collects initial detection information, matches the initial detection information according to a database, formulates a regional division detection plan, applies corresponding spectral irradiation according to the regional division detection plan, collects detection data, obtains wafer data in an initial state, performs adaptability evaluation on the detection data, and adaptively adjusts the detection plan according to the evaluation result, fuses the detection data of the initial and regulated states, analyzes the detection progress, cyclically adjusts the control until the progress meets the threshold, completes the spectral detection, and ensures that each area on the wafer surface is accurately detected by dynamic spectral modulation. The detection plan is evaluated and adjusted in real time to improve the detection coverage and accuracy. Data fusion and progress analysis improve the efficiency and accuracy of detection, reduce manual intervention, and solve the problem of missed detection easily caused by static light source irradiation in the prior art.

[0041] Preferably, the step of driving the detection light source module to perform initial light source irradiation and feedback information collection on the wafer to be tested to obtain initial detection information of the wafer to be tested includes: S11: initializing parameter configuration of the detection light source module so that the detection light source module is ready to perform light source irradiation work; S12: driving the detection light source module to perform multi-angle visible light irradiation and reflected light information collection on the wafer to be tested, obtaining reflected light information corresponding to the multi-angle visible light irradiation of the wafer to be tested, and performing multi-angle feedback information combination on the reflected light information corresponding to the multi-angle visible light irradiation, so as to obtain wafer specification information of the wafer to be tested; S13: driving the detection light source module to perform infrared light irradiation and reflected light information collection on the wafer to be tested, obtaining reflected light information of the wafer to be tested corresponding to the infrared light irradiation, and extracting information feedback content of the wafer surface area and the wafer internal layer level of the reflected light information corresponding to the infrared light irradiation, so as to obtain wafer layout information of the wafer to be tested; S14: The wafer specification information and the wafer layout information together constitute the initial detection information.

[0042] Specifically, to ensure that the light source module can correctly perform the light source illumination task, the detection light source module needs to be initialized first. The system selects the appropriate light source type (such as visible light, infrared light) and the corresponding wavelength according to the detection requirements of the wafer to be tested. Different light source wavelengths are suitable for detecting different characteristics of the wafer. For example, visible light is used for detecting surface features, and infrared light is used for detecting internal structures. By adjusting the intensity, illumination angle and illumination range of the light source, it is ensured that all areas of the wafer to be tested are covered, and excessive or uneven illumination is avoided. The scanning mode (such as static illumination or dynamic scanning) and scanning speed are configured to ensure the stability and consistency of the light source illumination during the detection process.

[0043] More specifically, multi-angle visible light illumination is used to obtain surface information of the wafer in order to evaluate its surface morphology and appearance characteristics. The detection light source module illuminates the surface of the wafer to be tested at multiple different angles. Different illumination angles can help detect the reflective characteristics of the wafer surface, thereby revealing surface unevenness, defects or other structural features. Optical sensors are used to collect light signals reflected from the wafer surface, including reflection intensity, spectral distribution and other data. Through these reflected light information, the roughness of the wafer surface, coating uniformity and other defects can be inferred.

[0044] More specifically, the reflected light information at different angles is synthesized, denoised and optimized to form comprehensive feedback information. This information helps the system evaluate the wafer's specifications, including surface geometric features such as diameter, thickness, and edge morphology. By analyzing the changes in the reflected light information and combining it with a preset database or model, the wafer's specifications (such as size, shape, surface roughness, etc.) can be inferred.

[0045] More specifically, infrared light irradiation helps detect the internal structure and layer information of the wafer, and is particularly suitable for evaluating defects inside the wafer or the distribution of materials at different layers. The light source module is driven to irradiate with infrared light, which can penetrate the surface of the wafer and detect information about the internal layers of the wafer, such as interlayer contact, stress distribution, etc. The infrared light signal reflected from the inside of the wafer is collected by the infrared sensor. These reflected light signals provide multi-dimensional information on the surface and internal layers of the wafer, such as the absorption and reflectivity of the material, and the presence of internal defects. The infrared reflected light information is analyzed to extract information about the surface area and internal structure of the wafer. Specifically, different areas of the wafer (for example, surface coating, inner layer material, defective area, etc.) and the layout of the wafer (such as whether there are internal cracks, the thickness of the material layer, etc.) can be identified through changes in the reflected signal.

[0046] More specifically, the specification information and layout information from visible light and infrared light detection are combined to form complete initial detection information. The multi-angle reflection data from visible light irradiation provides basic information about the appearance, size and surface characteristics of the wafer. The reflection information from infrared light irradiation reveals the internal structure of the wafer, the distribution of the surface coating, and possible defects or hierarchical structures. The wafer specification information is integrated with the wafer layout information, combining the advantages of both to form complete initial detection information. This can fully describe the external morphology and internal structural characteristics of the wafer, helping with subsequent further analysis and decision-making.

[0047] More specifically, the initial inspection information includes wafer specification information (such as size, shape, surface roughness, etc.) and wafer layout information (such as internal material layers, defect areas, etc.), which provide the basis for subsequent fine inspection, defect location and performance evaluation.

[0048] It is understandable that through the combination of visible light and infrared light, the system can simultaneously obtain the surface and internal structure information of the wafer, thereby improving the comprehensiveness of detection. Visible light is suitable for evaluating the appearance and surface features of the wafer, while infrared light can penetrate deep into the wafer to reveal its hierarchical structure and potential defects. Through multi-angle light source illumination and information fusion, it can minimize errors caused by factors such as illumination angle and uneven surface reflection, and improve data reliability and accuracy. The system can extract detailed information on the wafer surface and internal areas based on the reflected light information of different wavelengths. Through this multi-level information feedback, the system can better understand the overall layout of the wafer and help with subsequent defect diagnosis and quality control.

[0049] Preferably, the step of performing wafer information matching on the initial detection information of the wafer to be tested according to a preset database to obtain a region division detection scheme corresponding to the wafer to be tested includes: S21: Retrieving several pre-entered standard wafer information from a preset database, and performing information similarity matching on the initial detection information of the wafer to be tested according to the several standard wafer information, so as to obtain similarity characteristics between the wafer to be tested and the several standard wafer information; S22: When the similarity characteristics between the wafer to be tested and a certain standard wafer information reach a preset standard, the wafer pointed to by the standard wafer information whose similarity characteristics reach the preset standard is used as the target wafer, and the area division detection scheme of the target wafer is retrieved from the preset database as the area division detection scheme of the wafer to be tested; S23: When the similarity characteristics between the wafer to be tested and any standard wafer information do not meet the preset standard, the wafer executed by the standard wafer information with the highest similarity characteristics is used as the reference wafer, and the area division detection scheme of the reference wafer is retrieved from the preset database as the reference detection scheme; S24: performing wafer difference analysis on the wafer to be tested and the reference wafer according to the initial detection information of the wafer to be tested and the standard wafer information of the reference wafer to obtain wafer difference characteristics, and making difference adjustments to the reference detection scheme according to the wafer difference characteristics to obtain a regional division detection scheme for the wafer to be tested.

[0050] Specifically, pre-entered standard wafer information is extracted from the database to provide a reference benchmark for the initial detection information of the wafer to be tested. The system retrieves multiple entered standard wafer information from the preset database based on the basic information of the wafer to be tested (such as specifications, surface features, layout, etc.). These standard wafer information may include different types of wafers, such as wafers with different sizes, different surface conditions, and different material levels.

[0051] More specifically, the standard wafer information extracted from the database will be used as a candidate for subsequent matching and analysis. By comparing the similarity of the initial detection information of the wafer to be tested and the standard wafer, the most matching standard wafer is determined, and key information such as its specifications, surface features, and wafer layout are extracted from the initial detection information of the wafer to be tested. The initial detection information of the wafer to be tested is compared with the relevant information of each standard wafer using a similarity matching algorithm (such as cosine similarity, Euclidean distance, etc.) to obtain the similarity characteristics of the wafer to be tested and the standard wafer. According to the similarity results, the standard wafer information that is most similar to the wafer to be tested is selected. If the similarity of a certain standard wafer reaches the preset standard, the standard wafer is considered to be the target wafer.

[0052] More specifically, by finding the standard wafer with the highest similarity, the corresponding regional division detection scheme is obtained to guide the subsequent detection of the wafer to be tested. If the information similarity characteristics of the wafer to be tested and a certain standard wafer meet the preset standard, the standard wafer is used as the target wafer, and its regional division detection scheme is obtained, and the regional division detection scheme of the target wafer is retrieved from the preset database. The scheme usually includes different regional divisions of the wafer, key detection areas, detection methods for different areas, etc. The regional division detection scheme of the target wafer is applied to the wafer to be tested as a preliminary regional division detection scheme.

[0053] More specifically, when the similarity between the wafer to be tested and any standard wafer does not reach the preset standard, the standard wafer with the highest similarity is selected as the reference wafer, and the area division detection plan of the reference wafer is retrieved from the database as the preliminary detection plan for the wafer to be tested.

[0054] More specifically, a difference analysis is performed based on the initial detection information of the wafer to be tested and the standard wafer information of the reference wafer. The focus of the difference analysis is on the differences between the two in terms of size, shape, surface features, material hierarchy, etc. Wafer difference features are extracted from the difference analysis, such as size differences, surface defect differences, material hierarchy differences, etc., and the reference detection plan is adjusted based on the difference features. For example, if the size of the wafer to be tested is different from that of the reference wafer, the ratio of regional division may need to be adjusted; if there are different defects on the surface, it may be necessary to increase detection points or change the detection method in a targeted manner to obtain the regional division detection plan for the wafer to be tested: based on the adjusted plan, the final regional division detection plan for the wafer to be tested is obtained.

[0055] It is understandable that through similarity matching and difference analysis, the regional division detection scheme can be customized according to the specific characteristics of the wafer to be tested. Whether it is a high match with the standard wafer or a lack of complete match, a corresponding detection scheme can be provided to achieve personalized and accurate detection. Based on the standard wafer information in the preset database, the system can quickly and accurately generate a regional division detection scheme for the wafer to be tested, reduce manual intervention and subjective judgment, and improve detection efficiency and accuracy of results.

[0056] Preferably, the steps of applying corresponding spectral illumination to each position area of ​​the wafer to be tested according to the area division detection scheme, and collecting detection data of the wafer to be tested to obtain wafer detection data of the initial state of the wafer to be tested include: S31: performing digital twin simulation on the wafer to be tested according to the area division detection scheme to obtain a digital twin model of the wafer to be tested; S32: performing disassembly processing on the execution phase of the area division detection scheme to obtain several detection execution phases of the area division detection scheme, and performing labeling processing on the information feedback area of ​​the digital twin model of the wafer to be tested according to each detection execution phase to obtain an information feedback labeling set of the digital twin model of the wafer to be tested; wherein the detection execution phase includes several position areas on the wafer to be tested where spectral irradiation and information collection are performed simultaneously and corresponding spectral irradiation schemes, and the information feedback labeling set includes information feedback labels corresponding to each position area, and the information feedback labeling is used to feed back the collected information to the corresponding model area on the digital twin model of the wafer to be tested; S33: configuring module parameters of the detection light source module according to the first detection execution stage, so as to activate the corresponding light source interface in the detection light source module and adjust the direction of the illumination port in the detection light source module, so that the light source interface is in the working mode and the illumination port is in the calibration mode; S34: the light source interface in working mode provides an illumination spectrum corresponding to the detection execution stage, and the illumination port in calibration mode controls an illumination direction of the illumination spectrum so as to apply the illumination spectrum to a plurality of position areas on the wafer to be tested corresponding to the detection execution stage; S35: collecting detection data for the wafer to be tested, and feeding back the collected detection data to a corresponding model area on the digital twin model of the wafer to be tested according to the information feedback annotation set, so as to obtain wafer detection data of an initial state of the wafer to be tested.

[0057] Specifically, a digital twin model of the wafer to be tested is created to provide virtualization support for subsequent spectral irradiation and data acquisition. According to the specific detection requirements of the wafer to be tested, digital twin simulation is performed according to the predetermined area division detection plan. Through simulation, a high-precision digital twin model is obtained, covering the physical characteristics of the wafer and information on its different regions. The model output is used for subsequent feedback annotation and data analysis.

[0058] More specifically, the detection scheme is disassembled into stages, and the overall task is divided into multiple execution stages. Each stage covers a specific area and spectral illumination scheme. According to the disassembled stages, different areas in the digital twin model are annotated with information feedback. The annotation set of each area will be used to indicate which areas will be spectrally illuminated and data collected in each detection stage.

[0059] More specifically, according to the initial detection execution stage sequence, the detection light source module is parameterized, the light source interface in the light source module is activated, and its illumination mode is configured, and the illumination port in the light source module is adjusted so that it is correctly aligned with the corresponding area of ​​the wafer to be tested, ensuring the accuracy and consistency of the spectral illumination.

[0060] More specifically, at each detection execution stage, the initial state data of the wafer is obtained through precise spectral illumination and information collection. The light source interface in working mode provides corresponding spectral illumination. According to the requirements of each detection stage, a specific position area is illuminated. The illumination port is in calibration mode and the illumination direction is adjusted to ensure that the spectrum can accurately cover the selected area.

[0061] More specifically, through the feedback mechanism, the model and data collection are adjusted in real time to ensure the efficiency and accuracy of the detection process. According to the collected detection data, the information feedback annotation set in the digital twin model is updated in real time to ensure that the collected data at each stage is consistent with the virtual model.

[0062] Preferably, the steps of evaluating the detection adaptability of the wafer detection data of the wafer to be tested, and adaptively regulating and executing the area division detection scheme according to the evaluation result to obtain the wafer detection data of the regulated state of the wafer to be tested include: S41: analyzing the completeness of information feedback on the model area on the digital twin model of the wafer to be tested that has received the test data information feedback, so as to obtain the integrity adaptation characteristics of the area division test scheme; S42: analyzing the fineness of information feedback on the model area on the digital twin model of the wafer to be tested that has received the test data information feedback, so as to obtain the fineness adaptation characteristics of the area division test scheme; S43: The integrity adaptation feature and the fineness adaptation feature of the region division detection scheme are used together as an evaluation result of the detection adaptability of the region division detection scheme in the detection execution phase; S44: marking the detection execution stage of the next execution order in the area division detection scheme as the execution stage to be regulated, and performing adaptive regulation processing on the execution stage to be regulated according to the integrity adaptation feature and the fineness adaptation feature, so as to obtain the execution stage to be regulated that completes the adaptive regulation; S45: adjusting the information feedback annotation set accordingly according to the execution stage to be regulated to obtain the information feedback annotation set corresponding to the adaptive regulation; S46: configuring module parameters of the detection light source module according to the execution stage to be regulated, so as to activate the corresponding light source interface in the detection light source module and adjust the direction of the illumination port in the detection light source module, so that the light source interface is in the working mode and the illumination port is in the calibration mode; S47: the light source interface in working mode provides an illumination spectrum corresponding to the execution stage to be regulated, and the illumination port in calibration mode controls an illumination direction of the illumination spectrum so as to apply the illumination spectrum to a plurality of position areas on the wafer to be tested corresponding to the execution stage to be regulated; S48: Collecting detection data for the wafer to be tested, and feeding back the collected detection data to the corresponding model area on the digital twin model of the wafer to be tested according to the information feedback annotation set, so as to obtain wafer detection data of the control state of the wafer to be tested.

[0063] Specifically, the information feedback completeness analysis is performed on the model area that receives the test data feedback in the digital twin model of the wafer to be tested, and the integrity of the test plan is evaluated to ensure that all key information is fully fed back and the accuracy and comprehensiveness of the test data are guaranteed.

[0064] More specifically, information feedback fineness analysis is performed on the model area that receives inspection data feedback in the digital twin model of the wafer to be tested. Through fineness analysis, the resolution of the inspection data is evaluated to ensure that slight changes or anomalies can be identified during the inspection process, thereby improving inspection accuracy.

[0065] More specifically, the integrity adaptation features and the precision adaptation features are used together as the adaptability evaluation results of the detection execution phase to comprehensively evaluate the adaptability of the detection plan, provide a basis for the next execution decision, and ensure that the detection plan can be adjusted according to different situations.

[0066] More specifically, the execution stage to be regulated is marked as the execution stage to be regulated, and the stage is adaptively regulated according to the integrity adaptation characteristics and the fineness adaptation characteristics. According to the evaluation results, the detection execution stage is dynamically adjusted to optimize the resource allocation and operation strategy during the detection process.

[0067] More specifically, according to the execution stage after adaptive regulation, the information feedback annotation set is adjusted accordingly to ensure that the feedback annotation can match the new detection requirements and regulation status, making the detection process more targeted.

[0068] More specifically, according to the execution stage to be controlled, the parameter configuration of the detection light source module is adjusted, the corresponding light source interface is activated and the illumination port is debugged. By precisely controlling the working mode and illumination direction of the light source, the illumination accuracy of the detection spectrum is improved to ensure that each area of ​​the wafer to be tested can be properly illuminated.

[0069] More specifically, inspection data is collected on the wafer to be tested, and a set of annotations is made based on the information feedback, and the data is fed back to the digital twin model. Through real-time data feedback, it is ensured that the control status of the wafer in the digital twin model can accurately reflect the current inspection status, providing a basis for subsequent decision-making.

[0070] It is understandable that by evaluating the completeness and precision of information feedback, the accuracy of wafer detection can be effectively improved. The adaptive control function can be dynamically adjusted according to the feedback from the detection stage, making the detection process more targeted and flexible, and accurately controlling the light source interface and irradiation direction to ensure that the spectrum can effectively irradiate all positions of the wafer, improve the uniformity of lighting and the effectiveness of detection, adjust the state of the model area through real-time feedback, enhance the real-time and reliability of the model, and provide accurate data support for subsequent decision-making.

[0071] Preferably, the step of performing adaptive control processing on the execution stage to be regulated according to the integrity adaptation feature and the fineness adaptation feature to obtain the execution stage to be regulated in which adaptive control is completed includes: S441: analyzing the control correction tendency and control correction degree of the to-be-controlled execution stage according to the integrity adaptation feature, so as to obtain the to-be-controlled spatial characteristics fed back by the integrity adaptation feature; S442: analyzing the control correction tendency and the control correction degree of the to-be-controlled execution stage according to the fineness adaptation feature, so as to obtain the to-be-controlled spectral characteristics fed back by the fineness adaptation feature; S443: taking the spatial features to be regulated and the spectral features to be regulated as regulation constraints, and adaptively regulating the execution parameters of the execution stage to be regulated according to the regulation constraints. In the process of adaptive regulation, simulating the effect of the parameter regulation plan of the execution stage to be regulated, and supervising the effect of the simulation according to the regulation constraints, so as to iterate the parameter regulation plan until the simulation effect satisfies the regulation constraints, and determining the parameter regulation plan at this time as the actual regulation plan; S444: modifying the to-be-regulated execution stage according to the actual regulation scheme to obtain the to-be-regulated execution stage that completes the adaptive regulation.

[0072] Specifically, according to the integrity adaptation characteristics, the control correction tendency and degree of the control execution stage are analyzed to obtain the feedback of the space to be controlled. Through the feedback of the integrity adaptation characteristics, the space that needs to be optimized in this stage (such as the detection area, analysis depth, etc.) is judged, and the direction and intensity of the control are analyzed. This can help identify which areas need more detection data or higher-precision measurements.

[0073] More specifically, according to the fineness adaptation feature, the control correction tendency and degree of the control execution stage are analyzed to obtain the feedback of the spectral control feature. Through the feedback of the fineness adaptation feature, it is evaluated whether the spectral resolution needs to be improved in the control execution stage, and the detection of the surface or internal features of the wafer is further refined. This helps to ensure that even small changes and details can be effectively captured.

[0074] More specifically, the spatial features to be controlled are combined with the spectral features to be controlled to form control constraints, and the control execution phase is adaptively controlled based on these constraints. Reasonable control constraints are set in combination with the spatial and spectral characteristics to ensure that both the integrity of the spatial range and the accuracy of the spectral analysis can be met during adaptive control, thereby comprehensively improving the adaptability and accuracy of the detection process.

[0075] More specifically, in the adaptive control process, the effect of the parameter control plan in the control execution phase is simulated, and the simulation effect is supervised according to the control constraints. The effect of the plan is simulated, and the impact of the adjusted parameters on the test results is predicted to ensure that the effect after adaptive control meets expectations and avoid errors or inadaptability problems in actual execution. The simulation and supervision links can provide opportunities to discover potential problems in advance.

[0076] More specifically, according to the simulation results, the parameter control plan is iterated until the simulation results meet the control constraints. The parameter control plan at this time is the actual control plan. Through continuous iteration, the accuracy and effectiveness of the control plan are ensured to avoid error accumulation during execution. This process can improve the accuracy and effect of the final control plan and ensure that the detection can achieve optimal performance under different conditions.

[0077] More specifically, the control execution stage is modified according to the actual control scheme, and finally the control execution stage of the adaptive control is obtained. Through the modified scheme, it is ensured that the control execution stage can fully adapt to the changing conditions and execute the optimal detection parameters. The final control scheme can not only meet the detection needs, but also improve the adaptability and flexibility of the overall system.

[0078] It is understandable that through dynamic and fine-grained control and optimization, the wafer inspection process can be adaptively adjusted according to real-time feedback, thereby improving the accuracy and efficiency of inspection. Through simulation, iteration and fine constraints, it is ensured that each execution stage can be completed with the optimal configuration, thereby improving the intelligence level and adaptability of the entire inspection system.

[0079] Preferably, data fusion analysis is performed on the wafer detection data of the initial state and the control state of the wafer to be tested to obtain the detection progress characteristics of the wafer to be tested, and the adaptive control of the area division detection scheme and the data collection of the wafer detection data are cyclically performed until the detection progress characteristics meet the specified threshold value to complete the spectral detection of the wafer to be tested. The steps include: S51: performing data fusion analysis on the wafer detection data of the initial state and the control state of the wafer to be tested to obtain the detection progress characteristics of the wafer to be tested; wherein the detection progress characteristics include spatial progress and detail progress; S52: judging the detection progress feature according to a specified threshold value to obtain a spatial progress judgment result and a detailed progress judgment result in the detection progress feature; S53: If the spatial progress judgment result shows that the wafer to be tested has not completed the detection of all position areas, continue to adaptively control the area division detection scheme and collect detection data of the wafer to be tested; S54: If the detailed progress judgment result shows that the wafer to be tested has a position area where the detection details are insufficient, then continue to adaptively control the area division detection scheme and collect detection data of the wafer to be tested; S55: When the detection progress characteristic meets the specified threshold, the spectrum detection of the wafer to be tested is completed.

[0080] Specifically, data fusion analysis is performed on the wafer inspection data in the initial state and the control state of the wafer to be tested to generate inspection progress characteristics, including spatial progress (i.e., whether all areas to be inspected have been covered) and detail progress (i.e., whether the inspection details are sufficient and whether minor changes or anomalies have been identified). Through data fusion, more comprehensive and accurate inspection progress information can be obtained, and the comprehensiveness and precision of wafer inspection can be monitored in real time. This analysis provides a basis for subsequent regulation, so that the progress of each link can be effectively evaluated.

[0081] More specifically, the spatial progress and detailed progress are judged according to the specified thresholds to evaluate whether they meet the preset standards. If the inspection results show that some areas of the wafer have not been inspected, adaptive control and data collection will continue. If the inspection shows that certain locations lack sufficient details or depth (such as failure to detect minor changes or defects), the inspection plan will continue to be adjusted for additional collection. By judging these two progress characteristics, it is possible to accurately identify which areas or details are not fully covered during the inspection process, and adjust the inspection strategy in a timely manner to ensure comprehensive and accurate inspection.

[0082] More specifically, according to the results of spatial progress and detailed progress judgment, the regional division detection plan is dynamically adjusted, and data collection is performed until the detection progress meets the threshold. Adaptive regulation and data collection make the detection process not only continuous but also flexible, and the detection strategy can be adjusted according to real-time feedback. This can avoid missing key areas or details, optimize resource allocation, and improve detection efficiency and accuracy.

[0083] More specifically, when both the spatial progress and the detailed progress meet the specified thresholds, the wafer inspection is judged to be completed, and finally the spectral inspection is completed, ensuring that all areas and details that need to be inspected have been fully inspected and the inspection progress meets the requirements, and finally obtaining accurate inspection results to ensure the quality and reliability of spectral inspection.

[0084] It is understandable that through adaptive control and data fusion, the detection coverage and precision are ensured, and the omission of key areas or minor changes is avoided. Real-time progress tracking and dynamic control make the detection process more efficient. There is no need to over-detect completed areas or details. The scheme and collection strategy can be adjusted according to the real-time feedback of the detection, which improves the adaptability and flexibility of the entire detection process. Through this adaptive feedback mechanism, it is ultimately ensured that the spectral detection of the wafer meets the requirements of high precision and high quality.

[0085] Reference Figure 2 As shown, in a second aspect, the present invention provides a high-precision dynamic spectrum modulation system for a semiconductor wafer inspection light source, which is used to implement a high-precision dynamic spectrum modulation method for a semiconductor wafer inspection light source as described in any one of the first aspects, including: An information acquisition module is used to drive the detection light source module to perform initial light source irradiation on the wafer to be tested and collect feedback information to obtain initial detection information of the wafer to be tested; A scheme planning module is used to match the initial detection information of the wafer to be tested with wafer information according to a preset database to obtain a regional division detection scheme corresponding to the wafer to be tested; An initial detection module, used to apply corresponding spectral illumination to each position area of ​​the wafer to be tested according to the area division detection scheme, and collect detection data of the wafer to be tested to obtain wafer detection data of the initial state of the wafer to be tested; A control detection module, used for evaluating the detection adaptability of the wafer detection data of the wafer to be tested, and adaptively controlling and executing the area division detection scheme according to the evaluation result, so as to obtain the wafer detection data of the control state of the wafer to be tested; A cyclic detection module is used to perform data fusion analysis on the wafer detection data of the initial state and the control state of the wafer to be tested to obtain the detection progress characteristics of the wafer to be tested, and cyclically perform adaptive control of the area division detection scheme and data collection of the wafer detection data until the detection progress characteristics meet the specified threshold value to complete the spectral detection of the wafer to be tested.

[0086] In this embodiment, for the specific implementation of each module in the above system embodiment, please refer to the above method embodiment, which will not be repeated here.

[0087] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A high-precision dynamic spectrum modulation method for semiconductor wafer inspection light source, characterized in that: include: The detection light source module is driven to perform initial light source irradiation and feedback information collection on the wafer to be tested, so as to obtain initial detection information of the wafer to be tested; Performing wafer information matching on the initial detection information of the wafer to be tested according to a preset database to obtain a region division detection scheme corresponding to the wafer to be tested; Applying corresponding spectral illumination to each position area of ​​the wafer to be tested according to the area division detection scheme, and collecting detection data of the wafer to be tested to obtain wafer detection data of the initial state of the wafer to be tested; Performing an evaluation process on the detection adaptability of the wafer detection data of the wafer to be tested, and adaptively regulating and executing the area division detection scheme according to the evaluation result, so as to obtain the wafer detection data of the regulated state of the wafer to be tested; Data fusion analysis is performed on the wafer detection data of the initial state and the control state of the wafer to be tested to obtain the detection progress characteristics of the wafer to be tested, and the adaptive control of the area division detection scheme and the data collection of the wafer detection data are repeated until the detection progress characteristics meet the specified threshold value to complete the spectral detection of the wafer to be tested.

2. The high-precision dynamic spectrum modulation method for semiconductor wafer inspection light source according to claim 1, characterized in that: The steps of driving the detection light source module to perform initial light source irradiation and feedback information collection on the wafer to be tested to obtain initial detection information of the wafer to be tested include: Initializing parameter configuration of the detection light source module so that the detection light source module is ready to perform light source illumination work; Driving the detection light source module to perform multi-angle visible light irradiation and reflected light information collection on the wafer to be tested, obtaining reflected light information corresponding to the multi-angle visible light irradiation of the wafer to be tested, and performing multi-angle feedback information combination on the reflected light information corresponding to the multi-angle visible light irradiation, so as to obtain wafer specification information of the wafer to be tested; The detection light source module is driven to perform infrared light irradiation and reflected light information collection on the wafer to be tested, so as to obtain reflected light information of the wafer to be tested corresponding to the infrared light irradiation, and information feedback content of the wafer surface area and the wafer internal layer level is extracted from the reflected light information corresponding to the infrared light irradiation, so as to obtain wafer layout information of the wafer to be tested; The wafer specification information and the wafer layout information together constitute the initial detection information.

3. The high-precision dynamic spectrum modulation method for semiconductor wafer inspection light source according to claim 1, characterized in that: The steps of performing wafer information matching on the initial detection information of the wafer to be tested according to a preset database to obtain a region division detection scheme corresponding to the wafer to be tested include: Retrieving several pre-entered standard wafer information from a preset database, and performing information similarity matching on the initial detection information of the wafer to be tested according to the several standard wafer information, so as to obtain similarity characteristics between the wafer to be tested and the several standard wafer information; When the similarity characteristics between the wafer to be tested and a certain standard wafer information reach a preset standard, the wafer pointed to by the standard wafer information whose similarity characteristics reach the preset standard is used as the target wafer, and the area division detection scheme of the target wafer is retrieved from the preset database as the area division detection scheme of the wafer to be tested; When the similarity characteristics between the wafer to be tested and any standard wafer information do not meet the preset standard, the wafer executed by the standard wafer information with the highest similarity characteristics is used as the reference wafer, and the regional division detection scheme of the reference wafer is retrieved from the preset database as the reference detection scheme; Wafer difference analysis is performed on the wafer to be tested and the reference wafer according to the initial detection information of the wafer to be tested and the standard wafer information of the reference wafer to obtain wafer difference characteristics, and the reference detection scheme is difference-adjusted according to the wafer difference characteristics to obtain a regional division detection scheme for the wafer to be tested.

4. The high-precision dynamic spectrum modulation method for semiconductor wafer inspection light source according to claim 1, characterized in that: The steps of applying corresponding spectral illumination to each position area of ​​the wafer to be tested according to the area division detection scheme, and collecting detection data of the wafer to be tested to obtain wafer detection data of the initial state of the wafer to be tested include: Performing digital twin simulation on the wafer to be tested according to the area division detection scheme to obtain a digital twin model of the wafer to be tested; The area division detection scheme is disassembled in the execution stage to obtain several detection execution stages of the area division detection scheme, and the digital twin model of the wafer to be tested is labeled with information feedback areas according to each detection execution stage to obtain an information feedback annotation set of the digital twin model of the wafer to be tested; wherein the detection execution stage includes several position areas on the wafer to be tested where spectral irradiation and information collection are performed simultaneously and the corresponding spectral irradiation schemes, and the information feedback annotation set includes information feedback annotations corresponding to each position area, and the information feedback annotations are used to feed back the collected information to the corresponding model area on the digital twin model of the wafer to be tested; According to the first detection execution stage, module parameters of the detection light source module are configured to activate the corresponding light source interface in the detection light source module and adjust the direction of the illumination port in the detection light source module so that the light source interface is in the working mode and the illumination port is in the calibration mode; The light source interface in working mode provides an illumination spectrum corresponding to the detection execution stage, and the illumination port in calibration mode controls the illumination direction of the illumination spectrum so as to apply the illumination spectrum to a plurality of position areas on the wafer to be tested corresponding to the detection execution stage; Detection data is collected for the wafer to be tested, and the collected detection data is fed back to the corresponding model area on the digital twin model of the wafer to be tested according to the information feedback annotation set, so as to obtain wafer detection data of the initial state of the wafer to be tested.

5. The high-precision dynamic spectrum modulation method for semiconductor wafer inspection light source according to claim 4, characterized in that: The steps of evaluating the detection adaptability of the wafer detection data of the wafer to be tested, and adaptively regulating and executing the area division detection scheme according to the evaluation result to obtain the wafer detection data of the regulated state of the wafer to be tested include: Analyzing the completeness of information feedback on the model area of ​​the digital twin model of the wafer to be tested that has received the test data information feedback, so as to obtain the integrity adaptation characteristics of the area division test scheme; Performing information feedback precision analysis on the model area on the digital twin model of the wafer to be tested that has received the test data information feedback, so as to obtain the precision adaptation characteristics of the area division test scheme; The integrity adaptation feature and the fineness adaptation feature of the region division detection scheme are used together as the evaluation result of the detection adaptability of the detection execution phase of the region division detection scheme; Marking the detection execution stage of the next execution order in the area division detection scheme as the execution stage to be regulated, and performing adaptive regulation processing on the execution stage to be regulated according to the integrity adaptation feature and the fineness adaptation feature to obtain the execution stage to be regulated that completes the adaptive regulation; According to the execution stage to be regulated, the information feedback annotation set is adjusted accordingly to obtain the information feedback annotation set corresponding to the adaptive regulation; According to the execution stage to be regulated, module parameters of the detection light source module are configured to activate the corresponding light source interface in the detection light source module and adjust the direction of the illumination port in the detection light source module, so that the light source interface is in a working mode and the illumination port is in a calibration mode; The light source interface in working mode provides an illumination spectrum corresponding to the execution stage to be regulated, and the illumination port in calibration mode controls the illumination direction of the illumination spectrum so as to apply the illumination spectrum to a plurality of position areas on the wafer to be tested corresponding to the execution stage to be regulated; Detection data is collected for the wafer to be tested, and the collected detection data is fed back to the corresponding model area on the digital twin model of the wafer to be tested according to the information feedback annotation set, so as to obtain wafer detection data of the control state of the wafer to be tested.

6. The high-precision dynamic spectrum modulation method for semiconductor wafer inspection light source according to claim 5, characterized in that: The step of performing adaptive control processing on the execution stage to be regulated according to the integrity adaptation feature and the fineness adaptation feature to obtain the execution stage to be regulated that completes the adaptive control comprises: Analyzing the control correction tendency and control correction degree of the execution stage to be controlled according to the integrity adaptation feature, so as to obtain the space to be controlled feature fed back by the integrity adaptation feature; Analyzing the control correction tendency and the control correction degree of the execution stage to be controlled according to the fineness adaptation feature, so as to obtain the spectral control feature to be controlled fed back by the fineness adaptation feature; The spatial feature to be regulated and the spectral feature to be regulated are used as regulation constraints, and adaptive regulation of the execution parameters of the execution stage to be regulated is performed according to the regulation constraints. In the process of adaptive regulation, the effect of the parameter regulation plan of the execution stage to be regulated is simulated, and the effect of the simulation is supervised according to the regulation constraints, so as to iterate the parameter regulation plan until the simulation effect satisfies the regulation constraints, and the parameter regulation plan at this time is determined as the actual regulation plan; The scheme of the execution stage to be regulated is modified according to the actual regulation scheme to obtain the execution stage to be regulated that completes the adaptive regulation.

7. The high-precision dynamic spectrum modulation method for semiconductor wafer inspection light source according to claim 1, characterized in that: The steps of performing data fusion analysis on the wafer detection data of the initial state and the control state of the wafer to be tested to obtain the detection progress characteristics of the wafer to be tested, and cyclically performing the adaptive control of the area division detection scheme and the data collection of the wafer detection data until the detection progress characteristics meet the specified threshold value to complete the spectral detection of the wafer to be tested include: Performing data fusion analysis on wafer detection data of the initial state and the control state of the wafer to be tested to obtain detection progress characteristics of the wafer to be tested; wherein the detection progress characteristics include spatial progress and detail progress; The detection progress feature is judged according to a specified threshold value to obtain a spatial progress judgment result and a detailed progress judgment result in the detection progress feature; If the spatial progress judgment result shows that the wafer to be tested has not completed the detection of all position areas, then continue to adaptively control the area division detection scheme and collect detection data of the wafer to be tested; If the detailed progress judgment result shows that the wafer to be tested has a position area where the detection details are insufficient, then continue to adaptively control the area division detection scheme and collect detection data of the wafer to be tested; When the detection progress characteristic meets the specified threshold, the spectrum detection of the wafer to be tested is completed.

8. A high-precision dynamic spectrum modulation system for semiconductor wafer inspection light source, characterized in that: A high-precision dynamic spectrum modulation method for realizing a semiconductor wafer inspection light source as described in any one of claims 1 to 7, comprising: An information acquisition module is used to drive the detection light source module to perform initial light source irradiation on the wafer to be tested and collect feedback information to obtain initial detection information of the wafer to be tested; A scheme planning module is used to match the initial detection information of the wafer to be tested with wafer information according to a preset database to obtain a regional division detection scheme corresponding to the wafer to be tested; An initial detection module, used to apply corresponding spectral illumination to each position area of ​​the wafer to be tested according to the area division detection scheme, and collect detection data of the wafer to be tested to obtain wafer detection data of the initial state of the wafer to be tested; A control detection module, used for evaluating the detection adaptability of the wafer detection data of the wafer to be tested, and adaptively controlling and executing the area division detection scheme according to the evaluation result, so as to obtain the wafer detection data of the control state of the wafer to be tested; A cyclic detection module is used to perform data fusion analysis on the wafer detection data of the initial state and the control state of the wafer to be tested to obtain the detection progress characteristics of the wafer to be tested, and cyclically perform adaptive control of the area division detection scheme and data collection of the wafer detection data until the detection progress characteristics meet the specified threshold value to complete the spectral detection of the wafer to be tested.

Images