A detection system and method capable of accurately detecting the porosity defect rate of heavily doped single crystals

Abstract

This invention discloses a detection system and method for accurately detecting the porosity defect rate of heavily doped single crystals, relating to the field of image detection technology. The system acquires and preprocesses image data; detects high-brightness points based on pixel brightness values; determines center and edge points using distance ratios; and identifies the pore outline by judging whether the edge points form a circumferential distribution based on angle. Upon successful identification, a defect alert is issued. This invention significantly improves detection efficiency by replacing manual inch-by-inch observation with vacuum adsorption and automatic scanning; it eliminates subjective errors by using brightness, distance, and angle thresholds for quantitative judgment; it verifies the ring structure using the cosine theorem and 360-degree fault-tolerant comparison to accurately match the pore morphology; and it simultaneously marks defects, stores reports, and triggers alarms, achieving traceable quality management and real-time intervention.

Description

Technical Field

[0001] This invention relates to the field of image detection technology, specifically to a detection system and method that can accurately detect the porosity defect rate of heavily doped single crystals. Background Technology

[0002] Heavy-doped single-crystal silicon wafers are the core substrate material for power devices and epitaxial wafer manufacturing. Their surface quality plays a decisive role in the performance and yield of the final chip. Among them, porosity defects mainly originate from the entrainment of gas during crystal growth. After polishing, they exhibit a unique structural feature of surface openings and subcutaneous cavities. Under specific oblique illumination conditions, this defect will form a unique optical reaction. The edge of the porosity protrudes due to material extrusion, forming a bright outer ring. At the same time, the subcutaneous cavities reflect or scatter light, forming a noticeable bright spot at the center of the defect. This type of defect is a key defect type that causes epitaxial layer penetration defects and subsequent device failures. Currently, the visual inspection of such defects still relies on manual experience inspection. That is, quality inspectors who have been trained for a long time and have rich experience in defect identification observe the wafer inch by inch with the naked eye under a microscope or high-magnification optical imaging system. This is not only inefficient, but the inspection results are also easily affected by the visual fatigue of the quality inspectors, resulting in missed or false detections. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a detection system and method that can accurately detect the porosity defect rate of heavily doped single crystals. This solves the problem that current detection methods, which rely on manual experience, are not only inefficient but also depend entirely on the subjective judgment of quality inspectors. This makes it easy to miss or misdetect composite features such as a bright ring around the periphery and a bright spot in the center due to eye fatigue or experience differences, making it difficult to meet the requirements of high-precision mass production.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a method for accurately detecting the porosity defect rate of heavily doped single crystals, comprising the following specific steps: Step 1: Acquire image data and perform preprocessing. The image data includes pixel coordinates and pixel brightness values; Step 2: Detect high-brightness points based on pixel brightness values, and determine the center point and edge points based on the distance ratio between high-brightness points. Determine whether the edge points exhibit a circumferential distribution based on the angle formed by the center point and edge points, thereby identifying the porosity contour, and then proceed to Step 3; otherwise, the detection ends; Step 3: Issue a porosity defect warning.

[0005] Furthermore, the specific detection method for the high-brightness point is as follows: a high-brightness threshold is preset, and the brightness value of any pixel is compared with the high-brightness threshold. If the brightness value of the pixel is within the range of the high-brightness threshold, the pixel is marked as a high-brightness point; if the brightness value of the pixel is outside the range of the high-brightness threshold, the comparison continues.

[0006] Furthermore, the method for determining the center point and the edge point is as follows: by calculating the spacing between the high-brightness points and normalizing and filtering out equidistant point pairs, counting the number of occurrences of each point, taking the point with the most occurrences as the center point, and then the other high-brightness point paired with this center point is the edge point.

[0007] Furthermore, the method of calculating the spacing between high-brightness points and normalizing and filtering out equidistant point pairs is as follows: calculate the distance between each pair of high-brightness points based on the pixel coordinates to obtain the corresponding distance value; preset a distance error threshold, count the total number of distance values ​​to obtain the distance quantity, sum all distance values ​​and divide by the distance quantity to obtain the average distance value, then divide each distance value by the average distance value to obtain the corresponding distance ratio, and compare this ratio with the distance error threshold to filter out distance values ​​that meet the error range, and use the ratio relationship to identify distance pairs with equal or approximately equal circular radii; at the same time, record the high-brightness point A and high-brightness point B corresponding to each distance value, and save the point pair information that meets the conditions.

[0008] Furthermore, the method of counting the occurrence frequency of each point and taking the point with the most occurrences as the center point is as follows: count the occurrence frequency of these high-brightness points according to the pixel coordinates, count the number of times each high-brightness point appears in the point pair that meets the conditions; and use a quick sorting algorithm to sort the frequencies in descending order, and select the high-brightness point with the most occurrences as the center point.

[0009] Furthermore, the distance value is expressed as follows: Let the pixel coordinates of the highlight point M be... The pixel coordinates of the high-brightness point N are The distance value is obtained by calculating the pixel coordinates of the high-brightness point M and the pixel coordinates of the high-brightness point N using the Euclidean distance formula. ;in, Indicates the distance value. This represents the x-coordinate of the pixel coordinates of the highlight M. This represents the x-coordinate of the pixel coordinates of the highlight point N. The ordinate represents the pixel coordinates of the highlight point M. The ordinate represents the pixel coordinates of the highlighted point N.

[0010] Furthermore, the method for verifying whether edge points form a complete ring or a ring-like structure by means of angular closure conditions is as follows: by combining the center point with multiple edge points, the sum of the angles formed by the center point and each edge point in each combination is calculated, and a preset angle tolerance threshold is set, which is the sum of 360 degrees and positive or negative k, where k is a manually set tolerance value. If the sum of the angles formed is within the angle tolerance threshold range, then these edge points are determined to form a complete ring or a ring-like structure; otherwise, they are determined not to form a ring.

[0011] Furthermore, the method of combining the center point with multiple edge points and calculating the sum of the angles formed by the center point and each edge point in each combination is as follows: Mark the center point as center point c, and randomly mark three edge points, namely edge point a, edge point b, and edge point d. The distance between center point c and edge point a is denoted as distance value A, and the distance between center point c and edge point b is denoted as distance value B. The distance between edge point a and edge point b is calculated using the Euclidean distance formula and denoted as distance value C. The angle at center point c is then calculated using the cosine theorem. Similarly, calculate separately The included angle and The included angle Add these three included angles to obtain a set of total angle values ​​corresponding to the edge points. Repeat the above process until all edge point combinations are included in the calculation to obtain several total angle values. By traversing all edge point combinations, ensure that the included angle formed by each group of adjacent edge points and the center point is included in the evaluation.

[0012] Furthermore, the included angle The calculation method is as follows: Among them, for Taking the inverse cosine gives the included angle. , This represents the distance between the center point c and the edge point a. This represents the distance between center point c and edge point b. This represents the distance between edge point a and edge point b.

[0013] A detection system capable of accurately detecting the porosity defect rate of heavily doped single crystals includes the following specific modules: an image acquisition module: acquiring image data and performing preprocessing; the image data includes pixel coordinates and pixel brightness values; a porosity recognition module: detecting high-brightness points based on pixel brightness values, determining center points and edge points based on the distance ratio between high-brightness points, determining whether edge points exhibit a circumferential distribution based on the angle formed by the center point and edge points, identifying the porosity outline, and then executing a defect warning module; otherwise, the detection ends; and a defect warning module: issuing a porosity defect warning.

[0014] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: 1. By fixing heavily doped single-crystal silicon wafers onto a vacuum adsorption stage, and combining this with a precision two-dimensional moving platform to automatically position the area to be inspected, and continuously acquiring high-resolution images by a CMOS camera, the traditional method of manually moving and observing inch by inch under a microscope is replaced, significantly improving the inspection efficiency.

[0015] 2. By setting high brightness threshold, distance error threshold and angle tolerance threshold, the brightness of pixels, distance between pixels and angle closure conditions are quantitatively compared and automatically judged. This avoids the omission or false detection of the composite features of peripheral bright ring and central bright spot caused by visual fatigue and experience difference in manual quality inspection, and ensures the objectivity and consistency of the test results.

[0016] 3. The angle between the center point and the edge point is calculated using the cosine theorem. By accumulating the sum of multiple adjacent angles and comparing it with the 360-degree fault tolerance threshold, it is verified whether the edge point forms a complete ring or a ring-like structure. This accurately matches the unique morphological features of the pore surface opening and subcutaneous cavity, effectively eliminating non-pore interference.

[0017] 4. The identified pore contours are image-annotated, pixel coordinates are recorded, a defect detection report is generated and stored in a local database to form a traceable quality record; at the same time, an audible and visual alarm is triggered to realize real-time on-site intervention and provide data support for subsequent process improvement and yield analysis.

[0018] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0019] Figure 1 This is a flowchart of the method for detecting the porosity defect rate of heavily doped single crystals according to the present invention.

[0020] Figure 2 This is a structural diagram of the detection system for detecting the porosity defect rate of heavily doped single crystals according to the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus.

[0023] Example 1: like Figure 1 As shown, this embodiment of the invention provides a method for accurately detecting the porosity defect rate of heavily doped single crystals, comprising the following specific steps: Step 1: The heavily doped single-crystal silicon wafer is fixed on a vacuum adsorption stage to ensure its stable position during the inspection process and avoid movement or vibration that could affect imaging accuracy. The area to be inspected is moved to the underside of the microscope objective by controlling a precision two-dimensional moving platform, achieving precise positioning and automatic scanning of the area under test. The angle and intensity of the light source are adjusted to achieve oblique illumination. The optical characteristics of oblique illumination are used to highlight the morphological features of the bright ring around the periphery and the bright spot in the center of the pore defects. The CMOS camera continuously acquires high-resolution image data after macro focusing. The optical image is converted into a digital signal for subsequent algorithm analysis. The image data is then subjected to noise reduction processing to improve the quality of the image data. The image data includes pixel coordinates and pixel brightness values.

[0024] Step 2: High-brightness points are detected based on pixel brightness values ​​to initially screen out luminous areas belonging to pore defects. The center point and edge point are determined based on the distance ratio between high-brightness points. That is, the center point and edge point of the pore are separated from the high-brightness points using geometric features. Based on the angle formed by the center point and edge point, it is determined whether the edge point presents a circumferential distribution to identify the pore outline. That is, the edge point is verified to form a complete ring or quasi-ring by the angle closure condition to eliminate non-pore interference. Then, proceed to Step 3. Otherwise, the detection ends.

[0025] Step 3: Mark the identified pore outlines on the image to visually mark the defect location for easy manual verification. Record the pixel coordinates and quantify the defect location for spatial information for subsequent statistical analysis. Generate a defect detection report and store it in the local database to form a traceable quality record, supporting yield analysis and process improvement. At the same time, trigger an audible and visual alarm to prompt operators to handle the issue and notify on-site personnel in real time to intervene in a timely manner to prevent defective wafers from flowing into the next process. Then, the inspection ends.

[0026] Example 2 differs from Example 1 in that: The specific detection method for high-brightness spots is as follows: A preset high-brightness threshold is established by manually identifying and recording the brightness values ​​of pixels in the image that appear bright due to pore defects during the experiment. This set of records is then used to construct a threshold range for automatic identification by manually calibrating and collecting the brightness feature values ​​of pore defects. The brightness value of any pixel is compared with the high-brightness threshold. If the pixel brightness value is within the high-brightness threshold range, it is marked as a high-brightness point. This high-brightness threshold is used to compare the pixel brightness with the high-brightness threshold and filter out candidate points whose brightness matches the pore characteristics. If the pixel brightness value is outside the high-brightness threshold range, the comparison continues.

[0027] The method for determining the center point and edge points is as follows: By calculating the spacing between high-brightness points and normalizing them to select equidistant point pairs, counting the frequency of each point, and taking the point with the most frequent occurrence as the center point, the other high-brightness point paired with this center point is the edge point.

[0028] The method for selecting equidistant point pairs by calculating the spacing between high-brightness points and normalizing them is as follows: The distances between all pairs of high-brightness points are calculated based on pixel coordinates to obtain corresponding distance values. This is used to establish spatial relationships between high-brightness points, providing basic data for subsequent distance-based geometric analysis. A preset distance error threshold is set, and the total number of distance values ​​is counted to obtain the distance quantity. The sum of all distance values ​​is then divided by the distance quantity to obtain the average distance. Each distance value is then divided by the average distance to obtain the corresponding distance ratio. This ratio is compared with the distance error threshold to filter out distance values ​​that meet the error range. The ratio relationship is used to identify distance pairs with equal or approximately equal circular radii, because in reality, pores are not necessarily regular circles, and the distance error threshold is used for error tolerance. At the same time, the high-brightness points A and B corresponding to each distance value are recorded, and the point pair information that meets the conditions is saved for subsequent tracking and statistical analysis of the association frequency of each high-brightness point.

[0029] The method for counting the frequency of each point and selecting the point with the most occurrences as the center point is as follows: The frequency of occurrence of these high-brightness points is counted based on pixel coordinates. By counting the number of times each high-brightness point appears in a pair of points that meet the conditions, its probability as a candidate center point is reflected. A quick sorting algorithm is used to sort the frequencies in descending order, and the high-brightness point with the most occurrences is selected as the center point. The point with the highest frequency often forms a stable distance ratio relationship with multiple other points, which conforms to the geometric feature that the center point of the stomata is connected to multiple surrounding edge points.

[0030] The distance value is expressed as follows: Let the pixel coordinates of the high-brightness point M be... The pixel coordinates of the high-brightness point N are The distance value is obtained by calculating the pixel coordinates of the high-brightness point M and the pixel coordinates of the high-brightness point N using the Euclidean distance formula. ; in, Indicates the distance value. This represents the x-coordinate of the pixel coordinates of the highlight M. This represents the x-coordinate of the pixel coordinates of the highlight point N. The ordinate represents the pixel coordinates of the highlight point M. The ordinate represents the pixel coordinates of the highlighted point N.

[0031] The method to verify whether edge points form a complete ring or ring-like structure using the angular closure condition is as follows: By combining the center point with multiple edge points, the sum of the angles formed by the center point and each edge point in each combination is calculated. A preset angle tolerance threshold is set, which is the sum of 360 degrees and ±k, where k is a manually set tolerance value. This tolerance mechanism is introduced to adapt to the irregular circular features of actual pores caused by crystal growth or polishing processes, thereby improving the practicality of the detection. If the sum of the angles formed is within the angle tolerance threshold range, these edge points are determined to form a complete ring or quasi-ring; otherwise, they are determined not to form a ring.

[0032] The method for calculating the sum of the angles formed by the center point and each edge point in each combination is as follows: The center point is marked as center point c, and three edge points are randomly marked as edge point a, edge point b, and edge point d. This establishes the basic object for the geometric analysis of porosity defects. The spatial relationship between the center point and the edge points is transformed into a computable triangular structure. The distance between center point c and edge point a is denoted as distance value A, and the distance between center point c and edge point b is denoted as distance value B. The distance between edge point a and edge point b is calculated using the Euclidean distance formula and denoted as distance value C. The included angle at center point c is calculated using the law of cosines. Similarly, calculate separately The included angle and The included angle Add the three included angles to obtain a set of total angle values ​​corresponding to the edge points. By accumulating the sum of three adjacent included angles, we can initially verify whether these edge points form a complete circumference distribution around the center point. Repeat the above process until all edge point combinations are included in the calculation and several total angle values ​​are obtained. By traversing all edge point combinations, we can ensure that the included angle formed by each set of adjacent edge points and the center point is included in the evaluation, so as to avoid missing local defects.

[0033] included angle The calculation method is as follows: ; Among them, for Taking the inverse cosine gives the included angle. , This represents the distance between the center point c and the edge point a. This represents the distance between center point c and edge point b. This represents the distance between edge point a and edge point b.

[0034] Example 3: like Figure 2 As shown: A detection system capable of accurately detecting the porosity defect rate of heavily doped single crystals includes the following specific modules: Image acquisition module: Acquires image data and performs preprocessing. Image data includes pixel coordinates and pixel brightness values. The pore recognition module detects high-brightness points based on pixel brightness values, determines the center point and edge points based on the distance ratio between high-brightness points, determines whether the edge points are distributed in a circle based on the angle formed by the center point and edge points, identifies the pore outline, and then executes the defect prompting module; otherwise, the detection ends. Defect alert module: Issues alerts for porosity defects.

[0035] The preferred embodiments disclosed in this invention are merely illustrative examples of feasible implementation methods and are not intended to exhaustively cover all technical details of the invention, nor do they constitute a limitation on the scope of protection of this invention. In practical applications, those skilled in the art can make appropriate adjustments, combinations, or substitutions to the methods or systems described in these embodiments based on specific production conditions, equipment configurations, and process requirements, without departing from the core concept of this invention. For example, the acquisition method, data processing algorithm, control threshold, or specific implementation form of the execution unit can all be reasonably modified according to the actual situation.

[0036] Furthermore, the technical concepts disclosed in this invention have universal extensibility and adaptability. They are not only applicable to the specific scenarios described in the embodiments, but can also be applied in similar technical fields or related industrial processes through analogy, transplantation, or improvement. Any technical solution formed by making logically equivalent substitutions, reasonable adjustments to the order of steps, or recombination of module functions based on the principles, ideas, or framework disclosed in this specification should be considered to fall within the spirit and scope of this invention.

[0037] It should be further clarified that the specific descriptions and drawings in the patent documents are for the purpose of assisting in understanding the present invention only, and their details should not be interpreted as limitations on the claims. The true scope of protection of the present invention should be determined by the content of the claims recorded in the authorized text, and should cover all equivalent technical solutions that comply with the provisions of patent law under these claims. Any implementation method that has the same or similar function and achieves similar effects through reasonable changes in technical means under the guidance of the concept of the present invention falls within the scope of protection sought by the present invention.

[0038] Therefore, the descriptions in this specification are merely illustrative. Any adjustments to implementation methods, equivalent substitutions of technical features, or further applications based on the concept of this invention, as long as they do not depart from the overall technical approach described in this invention, should be included within the scope of protection of this invention. We encourage those skilled in the art to innovate and optimize based on their understanding of the core of this invention and in conjunction with specific practices, so as to jointly promote the progress and development of related technologies.

Claims

1. A method for accurately detecting the porosity defect rate of heavily doped single crystals, characterized in that: The specific steps include the following: Step 1: Acquire image data and perform preprocessing. Image data includes pixel coordinates and pixel brightness values. Step 2: Detect high-brightness points based on pixel brightness values, and determine the center point and edge points based on the distance ratio between high-brightness points. Determine whether the edge points are distributed in a circle based on the angle formed by the center point and edge points, identify the pore outline, and then proceed to Step 3; otherwise, end the detection. Step 3: Issue a porosity defect warning.

2. The method for accurately detecting the porosity defect rate of heavily doped single crystals according to claim 1, characterized in that: The specific detection method for the high-brightness points is as follows: A preset highlight threshold is set. The brightness value of any pixel is compared with the highlight threshold. If the brightness value of a pixel is within the range of the highlight threshold, the pixel is marked as a highlight. If the brightness value of a pixel is outside the range of the highlight threshold, the comparison continues.

3. The method for accurately detecting the porosity defect rate of heavily doped single crystals according to claim 2, characterized in that: The method for determining the center point and the edge point is as follows: By calculating the spacing between high-brightness points and normalizing them to select equidistant point pairs, counting the frequency of each point, and taking the point with the most frequent occurrence as the center point, the other high-brightness point paired with this center point is the edge point.

4. The method for accurately detecting the porosity defect rate of heavily doped single crystals according to claim 3, characterized in that: The method of calculating the spacing between high-brightness points and normalizing to filter out equidistant point pairs is as follows: Calculate the distance between each pair of all highlight points based on the pixel coordinates to obtain the corresponding distance value; A preset distance error threshold is set, the total number of distance values ​​is counted to obtain the distance quantity, and the sum of all distance values ​​is divided by the distance quantity to obtain the distance mean. Then, each distance value is divided by the distance mean to obtain the corresponding distance ratio, and this ratio is compared with the distance error threshold to filter out distance values ​​that meet the error range. The ratio relationship is used to identify distance pairs with equal or approximately equal circular radii. At the same time, the high point A and high point B corresponding to each distance value are recorded, and the point pair information that meets the conditions is saved.

5. The method for accurately detecting the porosity defect rate of heavily doped single crystals according to claim 4, characterized in that: The method for counting the occurrences of each point and selecting the point with the most occurrences as the center point is as follows: The frequency of occurrence of these high-brightness points is counted based on pixel coordinates, by counting the number of times each high-brightness point appears in a pair of points that meet the conditions; The frequency was sorted in descending order using a quicksort algorithm, and the most frequent highlight was selected as the center point.

6. The method for accurately detecting the porosity defect rate of heavily doped single crystals according to claim 5, characterized in that: The distance value is expressed in the following manner: Let the pixel coordinates of the high-brightness point M be... The pixel coordinates of the high-brightness point N are ; The distance value is obtained by calculating the pixel coordinates of the high-brightness point M and the pixel coordinates of the high-brightness point N using the Euclidean distance formula. ； in, Indicates the distance value. This represents the x-coordinate of the pixel coordinates of the highlight M. This represents the x-coordinate of the pixel coordinates of the highlight point N. The ordinate represents the pixel coordinates of the highlight point M. The ordinate represents the pixel coordinates of the highlighted point N.

7. The method for accurately detecting the porosity defect rate of heavily doped single crystals according to claim 6, characterized in that: The method for verifying whether edge points form a complete ring or ring-like structure using the angular closure condition is as follows: By combining the center point with multiple edge points, the sum of the angles formed by the center point and each edge point in each combination is calculated. A preset angle tolerance threshold is set, which is the sum of 360 degrees and ±k, where k is a manually set tolerance value. If the sum of the angles formed is within the angle tolerance threshold range, these edge points are determined to form a complete ring or a ring-like structure; otherwise, they are determined not to form a ring.

8. The method for accurately detecting the porosity defect rate of heavily doped single crystals according to claim 7, characterized in that: The method of calculating the sum of the angles formed by the center point and each edge point in each combination by combining the center point with multiple edge points is as follows: Mark the center point as center point c, and randomly mark three edge points as edge point a, edge point b, and edge point d. The distance between center point c and edge point a is denoted as distance A, and the distance between center point c and edge point b is denoted as distance B. Calculate the distance between edge point a and edge point b using the Euclidean distance formula, and denot it as distance C. Use the law of cosines to find the angle between them at center point c. Similarly, calculate separately The included angle and The included angle Add these three included angles to obtain a set of total angle values ​​corresponding to the edge points. Repeat the above process until all edge point combinations are included in the calculation to obtain several total angle values. By traversing all edge point combinations, ensure that the included angle formed by each group of adjacent edge points and the center point is included in the evaluation.

9. The method for accurately detecting the porosity defect rate of heavily doped single crystals according to claim 8, characterized in that: The included angle The calculation method is as follows: ； Among them, for The angle is obtained by taking the inverse cosine. , This represents the distance between the center point c and the edge point a. This represents the distance between center point c and edge point b. This represents the distance between edge point a and edge point b.

10. A detection system capable of accurately detecting the porosity defect rate of heavily doped single crystals, used to implement the detection method for accurately detecting the porosity defect rate of heavily doped single crystals as described in any one of claims 1-9, characterized in that, The detection system that can accurately detect the porosity defect rate of heavily doped single crystals includes: Image acquisition module: Acquires image data and performs preprocessing. Image data includes pixel coordinates and pixel brightness values. The pore recognition module detects high-brightness points based on pixel brightness values, determines the center point and edge points based on the distance ratio between high-brightness points, determines whether the edge points are distributed in a circle based on the angle formed by the center point and edge points, identifies the pore outline, and then executes the defect prompting module; otherwise, the detection ends. Defect alert module: Issues alerts for porosity defects.

Images